Journalclub – Programm · system (RAS) ﬁnally resulting in several angiotensin peptides, in particular angiotensin II (AngII), which acts via the AngII receptors type 1 and 2

Journalclub – Programm

Veranstaltungsort/-zeit: Hörsaal Anatomie, donnerstags 18:30 - 19:30 Uhr Leitung: Herr Prof. von Bohlen und Halbach sowie Herr Prof. Schwertz

Bitte bei Verhinderung am genannten Termin selbständig einen Tauschpartner organisieren und die Leiter, die Stipendiaten sowie Frau Halle ([email protected]) zeitnah per Email informieren.

Termin Name Titel der zu präsentierenden Publikation 22.08.2013 Lukas Vogel

Seite: Physiology of the (pro)renin receptor: Wnt of change?

Wiebke Malenke Seite:

Reduced Endoglin Activity Limits Cardiac Fibrosis and Improves Survival in Heart Failure

19.09.2013

Leif Koschützke Seite:

Abnormal Development of Dendritic Spines in FMR1 Knock-Out Mice

10.10.2013 Jonas A. Scheiber Seite:

Bone morphogenetic protein -4 and -5 in pancreatic cancer – Novel bidirectional players

24.10.2013

Theresa Ascholl Seite:

Beneficial Effects of Trypsin Inhibitors Derived from a Spider Venom Peptide in L-Arginine-Induced Severe Acute Pancreatitis in Mice

07.11.2013

Georg Laage Seite:

Neutrophil extracellular trap cell death requires both autophagy and superoxide generation

21.11.2013 Daniel Seifert Seite:

Shear stress induces iNOS expression in cultured smooth muscle cells: role of oxidative stress

05.12.2013 Dr. Deborah Janowitz Seite:

Relationships Between Gray Matter, Body Maß Index, and Waist Circumference in Healthy Adults

19.12.2013 Dr. Matthias Grothe Seite:

Effects of acute relapses on neuropsychological status in multiple sclerosis patients

09.01.2014 Ulrike Thorack Seite:

Inhibition of endogenous reverse transcriptase antagonizes human tumor growth

23.01.2014 Eva-Maria Böcker Seite:

Disruption of an EHMT1-Associated Chromatin-Modification Module Causes Intellectual Disability

20.02.2014 Tilmann Peter Seite:

The Role of Transforming Growth Factor-B–Mediated Tumor-Stroma Interactions in Prostate Cancer Progression: An Integrative Approach

27.02.2014 Philipp Heumann Seite:

Metabotropic Regulation of RhoA/Rho-Associated Kinase by L-type Ca2_ Channels

06.03.2014 Kim Rouven Liedtke Seite:

ROS implication in a new antitumor strategy based on non-thermal plasma

20.03.2014 Alexandra Welz Seite:

Activation of Protease-Activated Receptor 2 Induces VEGF Independently of HIF-1

03.04.2014 Johannes Dick Seite:

Anti-opsonic properties of staphylokinase

Lukas Vogel

Physiology of the (pro)renin receptor: Wnt of change? Quelle: 2010 International Society of Nephrology, Kidney International (2010) 78, 246–256 Betreuer: Prof. Jörg Peters (Physiologie) Was bedeutet mir das Thema persönlich? Der Review weist auf verschiedene Funktionen des (P)RR hin und ist damit eine Vorarbeit für mein Promotionsthema. Worauf kommt es mir bei diesem Thema am meisten an? Durch die Rolle des Wnt- Pathways in der Zellpolarisierung während Embryogenese, könnte der (P)RR in adulter Neurogenese ebenfalls mit Polarisierungsvorgängen und dem Zytoskelett in Verbindung stehen. Was fasziniert mich selbst am Thema am meisten? Die adulte Neurogenese als Bestandteil von Lernprozessen. Was gefällt mir am Thema weniger? Durch die verschiedenen Funktionen des (P)RR, mögliche Überlappungen und Proteininteraktionen ist die Aufgabe einzelner Genprodukte schwer nachzuvollziehen.

06/06/2013 1

Physiology of the (pro)renin receptor:Wnt of change?Gabin Sihn1, Anthony Rousselle1, Larissa Vilianovitch1, Celine Burckle1 and Michael Bader1

1Max-Delbruck-Center for Molecular Medicine (MDC), Berlin-Buch, Germany

The (pro)renin receptor is a protein that binds prorenin and

renin in tissues, leading to their activation and, at the same

time, to the initiation of intracellular signaling. The activation

of local renin–angiotensin systems may play an important

role in tissue damage induced by cardiovascular diseases and

diabetes. However, (pro)renin receptor is also called ATP6ap2

because it has been shown to be associated with vacuolar

Hþ -ATPase involvement in vesicular acidification and

signaling in cells. Notably, lack of the protein in vertebrates

leads to developmental alterations and early embryonic

lethality probably as a result of the recently discovered role

of the (pro)renin receptor and the vacuolar Hþ -ATPase in

Wnt signaling. This review summarizes the current findings

about these two functions of (pro)renin receptor/ATP6ap2

pointing out the possible links between both.

Kidney International (2010) 78, 246–256; doi:10.1038/ki.2010.151;

published online 26 May 2010

KEYWORDS: end-organ damage; rennin; V-ATPase; Wnt signaling

Until recently, the only function assigned to renin and itsprecursor prorenin (after activation) was the cleavage ofangiotensinogen as a first step in the renin–angiotensinsystem (RAS) finally resulting in several angiotensin peptides,in particular angiotensin II (AngII), which acts via the AngIIreceptors type 1 and 2 (AT1 and AT2). This activity is notonly observed in plasma but also locally in tissues formingthe basis for tissular RAS,1,2 which are of prime importancefor organ damage in cardiovascular diseases and diabetes.However, since renin is hardly expressed outside the kidneyit remained unclear how the enzyme reaches these sites.The discovery of the (pro)renin receptor ((P)RR/ATP6ap2)has provided a possible solution for this problem.3 Reninand prorenin (henceforth summarized as (pro)renin) bind(P)RR/ATP6ap2 in organs where (pro)renin is normally notgenerated leading to enzymatic activation and to intracellularsignaling (Figure 1).3–5 These data were ending the postulateof prorenin as a precursor devoid of any activity and werefeeding the concept that (pro)renin can have angiotensin-independent actions. These actions are of particular impor-tance in light of the fact that (pro)renin levels in plasma andtissues rise drastically after treatment with blockers of theRAS such as AT1 antagonists, angiotensin-converting enzyme(ACE) inhibitors and, in particular, renin inhibitors.6,7

However, (P)RR/ATP6ap2 had been discovered before as aprotein associated with the vacuolar Hþ -ATPase (V-ATPase)in adrenal chromaffin cells8 and recent data mainly fromfish and frog models revealed essential functions in cellularphysiology and signaling, in particular Wnt signaling, whichare independent of the RAS9,10 (http://zfin.org). Consequently,the question arose, which is the ‘real’ function of the protein,(pro)renin binding or V-ATPase interaction, or are bothfunctions interconnected. This review will try to providea comprehensive description of the current knowledgeabout (P)RR/ATP6ap2 and its interaction with (pro)reninand V-ATPase, which, however, does not yet allow a finalanswer to this question.

PRORENIN AND RENIN

Renin constitutes a crucial step within the RAS due to itsunique ability to proteolytically transform the precursorangiotensinogen into angiotensin I (AngI), which is subse-quently modified into various vasoactive peptides. Prorenin,the inactive form of renin, does not show any enzymatic

r e v i e w http://www.kidney-international.org

& 2010 International Society of Nephrology

Received 10 February 2010; revised 12 March 2010; accepted 23 March

2010; published online 26 May 2010

Correspondence: Michael Bader, Max-Delbruck Center for Molecular

Medicine, Robert-Rossle-Str. 10, D-13125 Berlin-Buch, Germany.

E-mail: [email protected]

246 Kidney International (2010) 78, 246–25606/06/2013 2

http://dx.doi.org/10.1038/ki.2010.151

http://zfin.org

http://www.kidney-international.org

mailto:[email protected]

activity due to a 43-amino acid prosegment, which covers theactive site cleft. Prorenin can be activated in a proteolytic or anon-proteolytic way. Proteolytic activation takes place mostlyin the juxtaglomerular apparatus and is characterized by theremoval of the prosegment. This process is irreversible andinvolves ill-defined proconvertases. Non-proteolytic activa-tion is a conformational change resulting in the unfoldingof the prosegment without any cleavage. Although thisreversible open conformation has only been achieved in vitroby low pH or cold temperature, it suggested thatnon-proteolytic activation is possible but not yet definedin vivo.11,12

In physiological conditions, the plasma prorenin/reninratio is around 10/113 and chronic stimulation of thejuxtaglomerular apparatus decreases this ratio.14 This isexplained by the fact that prorenin is constitutively secretedwhereas renin is stored in secretory vesicles within thejuxtaglomerular apparatus in the kidney until release.

After bilateral nephrectomy, renin disappears and onlyprorenin persists in blood.15 Thus, the kidney seems to be theonly renin-secreting organ and, more importantly, proreninis also secreted by other organs. Organs such as reproductive

tract, eye, adrenal and submandibular gland have beendescribed as extrarenal prorenin sources.16

Given the production of AngII at tissue level such as heartand vascular wall in the absence of local renin expression,17 ithas been assumed that (pro)renin is recruited and locallyactivated in these tissues by an unknown protein. Succes-sively, three different receptors for (pro)renin were described.First, an intracellular renin-binding protein emerged,18 butits genetic deletion did not have any cardiovascular effects.19

Then, the mannose-6-phosphate/insulin-like growth factor IIreceptor received particular attention. It binds and inter-nalizes (pro)renin but contributes to its clearance and doesnot result in AngI formation or intracellular signaling.20–24

Finally, (P)RR/ATP6ap2 was discovered and became the mostpromising candidate.

(PRO)RENIN RECEPTOR

The full-length (pro)renin receptor ((P)RR/ATP6ap2) wasdiscovered in vitro in human mesangial cells and clonedin 2002.3 The human ATP6AP2 gene located on theX chromosome at locus p11.4, codes for a protein with350-amino acids and a predicted mass of around 37 kDa and

‘Opened’ prorenin

‘Closed’ prorenin

(P)RR

(P)RR receptor

Soluble (P)RR

M8-9 fragment

H+-ATPaseLRP6

Frizzled

AT-receptor

HRPAOG

AngII

AngIACE

PLZF

PLZF

MAPK(p38, ERK1/2,

JNK)

ER

Nucleus

Golgi Exocyticvesicle

Vesicular andextracellularacidification

Signalingendosome

Signalingendosome

β-Catenin

β-Arrestins?

?

?

H+

H+ H+

H+

(–) (P)RR(+) PI3K– p85α

(+) PAI-1(+) TGFβ1(+) COX2

(+) En2(+) Axin2

Transcriptional effects

Wnt

Figure 1 | Functions of (P)RR/ATP6ap2. For details see text. ACE, angiotensin-converting enzyme; AOG, angiotensinogen; ER, endoplasmicreticulum; HRP, handle region peptide; MAPK, mitogen-activated protein kinase; PLZF, promyelocytic leukemia zinc finger.

Kidney International (2010) 78, 246–256 247

G Sihn et al.: Update on the (pro)renin receptor r e v i e w

06/06/2013 3

can be divided into four different domains: an N-terminalsignal peptide, an extracellular domain binding (pro)renin, asingle transmembrane domain and a short cytoplasmicdomain.25 Although prominently expressed in the brain,(P)RR/ATP6ap2 has also been reported in various organssuch as the heart, liver, kidney, muscle, adrenal gland,pancreas, ovary, and placenta (Figure 2).3 In addition, micro-array data suggest an ubiquitous expression of its mRNA.26

In the kidney, (P)RR/ATP6ap2 expression has been detectedin mesangial cells,3 podocytes,27,28 and intercalated cells.29

Initially, (P)RR/ATP6ap2 was reported to bind equallyboth renin and prorenin in vitro without internaliza-tion.3,30,31 However, Batenburg et al.32 described differentialbinding properties in vitro (KDa values of 7 and 20 nM forprorenin and renin, respectively), arguing that prorenin maybe the main endogenous agonist in vivo. The binding ofprorenin to the receptor has two important consequences.First, prorenin undergoes non-proteolytic activation, displaysfull enzymatic activity and initiates AngII-dependent effects(Figure 1).3,30,32 Unexpectedly, different studies using variousRAS inhibitors described that (P)RR/ATP6ap2 activationtriggers intracellular pathways in an AngII-independentmanner (Figure 1). Indeed, the mitogen-activated proteinkinase (MAPK) ERK1/2 is activated in cardiomyocytes,33

vascular smooth muscle cells,34 mesangial cells,4,35 andmonocytes.36 This results in the upregulation of pro-fibroticgenes such as transforming growth factor b1,4,5,35,37 plasmino-gen-activator inhibitor-1,4,5,31,35,38 fibronectin,5 collagen 1,5

and induction of proliferation in mesangial cells.4 In kidneytissues, (P)RR/ATP6ap2 upregulates via ERK1/2 stimulationinflammatory mediators such as cyclooxygenase-2 (COX2),37,39

interleukin-1 (IL-1b),39,40 and tumor necrosis factor a.40

Moreover, (P)RR/ATP6ap2 induces ERK1/2 activation inendothelial cells and in the retinal pigment epithelium, whereit increases the expression of monocyte chemotactic protein,vascular endothelial growth factor and again collagen 1.41–44

However, the physiological relevance of these ERK1/2 activa-tion data has to be confirmed, since they have been obtainedwith supraphysiological concentrations of (pro)renin and oftenwithout loss-of-function controls for (P)RR/ATP6ap2, whichare necessary to support its direct interactions with (pro)renin.

Such controls were included, when the activation ofthe promyelocytic leukemia zinc finger transcription factor(PLZF) by (P)RR/ATP6ap2 (Figure 1) was first described incardiac tissue.45,46 PLZF downregulates (P)RR/ATP6ap2expression in a negative feedback loop and, thereby, preventsexcessive receptor activation by high prorenin levels.45

However, this mechanism discovered in cell culture has alsoto be taken with caution since Krebs et al.47 have described thatan increase in renal (pro)renin upregulates (P)RR/ATP6ap2arguing for a positive feedback. However, Feldman et al.7 haveshown that extreme upregulation of plasma renin in renin-inhibitor-treated rats even downregulates (P)RR/ATP6ap2expression in the kidney.

The subcellular localization of (P)RR/ATP6ap2 is unusualfor a receptor. Only a minor amount of the protein wasfound on the plasma membrane, the majority is located onintracellular vesicles.33,45 In addition, a furin cleavage site onthe extracellular part of (P)RR/ATP6ap248 allows the release ofa 28 kDa fragment during passage through the Golgi apparatus(Figure 1), which could be detected in the supernatant of cellsexpressing the protein.49 We have speculated that (pro)reninmight bind to this soluble form of (P)RR/ATP6ap2 and actthrough an unknown receptor on cells (Figure 1).50 The cleavageleaves behind a short protein consisting of the transmembraneand the intracellular domains, which corresponds to theM8-9 fragment of (P)RR/ATP6ap2 found to be associatedwith V-ATPase8 (see (P)RR/ATP6ap2 and V-ATPase).

PRORENIN, (P)RR/ATP6ap2 AND PATHOLOGYProrenin

Although the RAS has a prominent role in the pathogenesisof organ damage induced by diabetes and hypertension, thereis to date no clinical data ascertaining the relevance ofprorenin and (P)RR/ATP6ap2 in RAS-associated pathologies.The question of a potential role of prorenin in pathology wasraised for the first time by the description that humandiabetic patients display 3 to 7-fold higher prorenin/reninratios in the blood.51–53 Experimental evidence was providedby Veniant et al., who reported that transgenic rats with atargeted prorenin expression in the liver and bearing highplasma prorenin levels, develop glomerulosclerosis, cardiachypertrophy, and vascular damage at the age of 20 weeks,independently of any hypertension.54 That such a phenotypeis due to a local activation of prorenin within tissues wassupported by the report that prorenin overexpression in miceresults in elevated AngI contents in the heart, without anyactivation of the circulating RAS.55 Interestingly, the expres-sion of a non-cleavable prorenin still resulted in AngIproduction within tissues, indicating that proteolytic con-version into renin is not a prerequisite for local actions ofprorenin.56 Taken together, these data support non-proteo-lytic activation of prorenin possibly by (P)RR/ATP6ap2 asa contributing factor for tissue damage.

However, several contradicting data urge us to reconsiderthese findings. High prorenin levels are also found duringpregnancy in the maternal plasma (where they can rise by

(P)RR

BrainHeart

Aorta

Muscle

Kidney

Adrenal glandTestis

Ovary

Intestine

β-Actin

Figure 2 | Expression of (P)RR/ATP6ap2 in different rat organs.Expression of (P)RR/ATP6ap2 was assessed in different organsof male and female Sprague–Dawley rats by RNase protectionassay using each 20 mg total RNA with specific probes for rat(P)RR/ATP6ap2 and b-actin.

248 Kidney International (2010) 78, 246–256

r e v i e w G Sihn et al.: Update on the (pro)renin receptor

06/06/2013 4

a factor of 10)57–59 in the absence of any obvious tissuedamage. Moreover, Campbell et al.60 re-evaluated thephenotype of the same transgenic line of rats overexpressingprorenin described by Veniant et al.54 On the contrary to thefirst report, Campbell et al.60 observed only modest renallesions and myocardial fibrosis after 6 months of age. Inaddition, these animals developed hypertension at 3 to 12months of age, which likely caused the slight renal andmyocardial phenotypes. These results undermined proreninas a risk factor for tissue damage. Although the authors werenot able to explain the obvious discrepancy, this revisedphenotype was in accordance with other recent reportspublished by different groups. Peters et al.61 reported that the200-fold increased plasma prorenin levels in a rat modelbearing an inducible renin transgene was not sufficient toinduce glomerulosclerosis or cardiac fibrosis, despite amoderate myocardial hypertrophy and a slight increase inmarkers for hypertrophy, fibrosis, and oxidative stress in theheart.62 Moreover, Mercure et al.63 reported that proreninoverexpression in transgenic mice yielded only mild hyper-tension, myocardial hypertrophy, and albuminuria withoutcardiac fibrosis or renal injury. They even discussed beneficialactions of prorenin, because in one line of their transgenicmice tissue damage is lower than expected from the level ofblood pressure. High (pro)renin levels are also the con-sequence of any pharmacological intervention in the RASincluding the use of ACE or renin inhibitors or AT1antagonists.6,7 However, no cardiovascular damage has everbeen described to be induced by these treatments. In theopposite, these drugs are among the most efficient for amultitude of cardiovascular and metabolic diseases to avoidtissue damage. In conclusion, the importance of (pro)reninin organ damage remains controversial. It may not besufficient to elicit tissue damage per se, but require a specificpathological environment to do so.

(P)RR/ATP6ap2

What about (P)RR/ATP6ap2? Several reports have describeda modulation of (P)RR/ATP6ap2 expression under patho-logical conditions in vitro and in vivo, suggestive of itsinvolvement in tissue damage. Notably, (P)RR/ATP6ap2expression has been shown to be upregulated in the kidneyof streptozotocin-induced diabetic rats,40,64 as well as inmesangial cells upon a 14-day incubation with high glucoseconcentrations in vitro.39 In addition, increased (P)RR/ATP6ap2 expression has been reported in the remnantkidney of nephrectomized rats,65,66 in the heart and kidney ofrats with congestive heart failure,67 in the heart ofspontaneously hypertensive rats,68 and in atherosclerotictissues.69 However, clinical and experimental reports depict-ing its real impact in pathology are scarce. Recently, aJapanese cohort study has revealed a significant associationbetween a polymorphism in the (P)RR/ATP6ap2 gene andblood pressure in humans.70 This observation concurs withreports from our group and others showing that theoverexpression of human (P)RR/ATP6ap2 in the rat leads

to cardiovascular alterations. We observed that a targeted(P)RR/ATP6ap2 overexpression in the vasculature leads tochronic hypertension after 6 months of age,48 and Kaneshiroet al.71 reported that a ubiquitous overexpression leads toglomerulosclerosis in an AngII-independent manner. How-ever, these symptoms were mild and their onset was late inlife despite an early and drastic overexpression of thetransgene in both models not quite supporting an importantrole of (P)RR/ATP6ap2 in local prorenin activation andcardiovascular pathology.

However, the most interesting clinical observation camefrom the description by Ramser et al.72 that a mutation in the(P)RR/ATP6ap2 gene (leading to a reduction of 50% of thefunctional protein) does not result in a cardiovascularphenotype but leads to mental retardation. This is to datethe only report of a (P)RR/ATP6ap2 deficiency in mammals,and it suggests important functions in brain developmentand/or neurophysiology (see (P)RR/ATP6ap2 and V-ATPase).

Nevertheless, the lack of a gene-deficient mouse modelmakes it difficult to assess in a specific manner the relevanceof (P)RR/ATP6ap2 in pathology. In addition, the lack ofinformation on (P)RR/ATP6ap2 structure (X-ray crystal-lography has not been performed yet) hampers the develop-ment of specific antagonists. To study the effects of (P)RR/ATP6ap2 in pathology, Ichihara et al.73 have employed apeptide of 10 amino acids named Handle Region decoyPeptide (HRP) as a blocker of the prorenin/(P)RR/ATP6ap2interaction.

(P)RR/ATP6ap2 INHIBITION AS THERAPEUTIC CONCEPT

HRP was designed on the basis of the prorenin ‘handleregion’ sequence described by Suzuki et al.12 as a decoy forprorenin (but not renin) interacting with (P)RR/ATP6ap2(Figure 1). In a series of seminal experiments, they describedthe beneficial effects of this peptide on kidney damageassociated with experimental diabetes. In streptozotocin-induced diabetic rats, a long-term treatment with HRP wasable to prevent the development of proteinuria andglomerulosclerosis, independent of glycemia and couldreverse already established diabetic nephropathy.73,74 Ichiharaet al.75 also used diabetic mice deficient for the AT1a receptorand treated with an ACE inhibitor. Although insensitive toAngII, these mice still developed glomerulosclerosis, whichwas consistent with observations that RAS blockers areunable to completely prevent end stage organ damageassociated with diabetes.76–78 HRP could efficiently blockglomerulosclerosis in this model. These data were consistentwith a critical role of the (pro)renin/(P)RR/ATP6ap2interaction in diabetic nephropathy.

Matavelli et al.40 recently confirmed the beneficial effectsof HRP in diabetic nephropathy. In addition to a reducedalbuminuria, they observed a significant reduction in theamounts of pro-inflammatory cytokines such as tumornecrosis factor a and IL-1b in the kidney, drawing theconclusion that the (pro)renin/(P)RR/ATP6ap2 axis promotesdiabetic nephropathy by enhancing renal inflammation.



06/06/2013 5

Interestingly, they could not confirm the data of Ichiharaet al.73 showing that HRP treatment in diabetic rats normal-ized AngII levels in the kidney.

The efficiency of HRP was further shown in otherpathologies including cardiac fibrosis in a genetic model ofhypertension (spontaneously hypertensive rats)68 as well asischemia-induced retinal neovascularization and ocularinflammation in endotoxin-induced uveitis.79,80

However, several studies evaluating the efficiency of theHRP in vitro and in vivo have failed to reproduce suchimpressive results. Susic et al.81 could only confirm a mildbeneficial effect of HRP on cardiac hypertrophy in sponta-neously hypertensive rats. HRP was inefficient in blocking thedevelopment of albuminuria and cardiac hypertrophy intransgenic mice with high circulating prorenin levels,although used at concentrations 10 times higher than thoseused by Ichihara et al.73 Krebs et al.82 could not prevent renaldamage by HRP in the Goldblatt hypertension model. Feldtet al.36 could not show that HRP was blocking (pro)reninbinding to (P)RR/ATP6ap2 in monocytes in vitro, norcould it block ERK1/2 signaling induced by (pro)reninin monocytes and vascular smooth muscle cells in vitro.83

More interestingly, they could not block mortality andnephrosclerosis with HRP in hypertension models such asdouble-transgenic rats overexpressing human renin andangiotensinogen83 and Goldblatt rats,84 whereas the renininhibitor aliskiren could. It can be argued that suchhypertensive models not only display high prorenin levelsbut also high renin and high plasma AngII levels andtherefore are not suited to study the actions of HRP, whichhas been designed to target prorenin only. However, thesedata suggest that the efficiency of HRP as a therapeuticcompound for renovascular damage needs further examina-tion, and raises the question whether this efficiency may bepathology-dependent, with a better outcome in diabetes.

Moreover, although the capacity of HRP to competitivelyblock the binding of (pro)renin to (P)RR/ATP6ap2 has beendescribed in vitro,12,85 it cannot be excluded that this peptidemay have other targets in vivo. In particular, it is most likelyprocessed into smaller peptide metabolites in the circulationwith unpredictable and possibly beneficial effects on othercardiovascular and inflammatory mediators. Confirming thereal specificity of HRP in vivo is a crucial point to ascertainthe roles described for (P)RR/ATP6ap2 in pathology.

(P)RR/ATP6ap2 AND V-ATPase(P)RR/ATP6ap2 as V-ATPase subunit

Despite the efforts to understand the role of the (P)RR/ATP6ap2 related to the RAS, we have to keep in mind that(P)RR/ATP6ap2 was initially discovered not as a (pro)renin-binding protein. In 1998, Ludwig et al. described a truncatedform (M8-9) of (P)RR/ATP6ap2 composed of the C-terminalpart and the transmembrane domain, which co-purified onnative polyacrylamide gels with the V-ATPase in chromaffincells of the adrenal medulla and was later named ATP6ap2 forvacuolar Hþ -ATPase-associated protein 2.8 They erroneously

assigned an in-frame ATG immediately upstream of the shortfragment on the chromosome as start codon for the proteinnot considering the possibility of proteolytic cleavage. Whenfull-length (P)RR/ATP6ap2 was discovered in 2002 byNguyen et al.3 the identity of the two proteins was notimmediately obvious leading to two different names for thesame protein.

(P)RR/ATP6ap2 is not only found in vertebrates but alsoin the worm Caenorhabditis elegans and the insect Drosophila,which do not possess any RAS. Interestingly, vertebrates andinvertebrates share a conserved sequence in the transmem-brane and cytosolic domains corresponding to the V-ATPase-associated M8-9 fragment, whereas the sequence coding forthe extracellular domain ((pro)renin-binding domain) isconserved only in vertebrates.48,50 Based on these findings itwas postulated that (P)RR/ATP6ap2 has emerged from thefusion of two genes in vertebrates: one gene encoding theN-terminal part, which binds (pro)renin and anotherone coding for the V-ATPase-associated part with an essentialrole in cell survival.86 Alternatively, the N-terminal part of aV-ATPase-associated protein may have acquired (pro)renin-binding properties during evolution.50 As already statedabove, both parts of the protein are linked by a furin cleavagesite, which may allow their separation and functionalindependence.

Localizations and functions of V-ATPase

The V-ATPase is a multiprotein complex composed of aperipheral V1 domain (eight subunits) responsible for ATPhydrolysis, a V0 domain (six subunits) responsible for protontranslocation and two accessory subunits (Ac45 and (P)RR/ATP6ap2).8,87 The complexity is even higher given that somesubunits have different isoforms, the expression of which istissue-specific and may be involved in the intracellularlocalization of the pump. V-ATPases are mainly withinintracellular membranes and regulate the pH of intracellularcompartments such as endosomes, lysosomes, synapticvesicles, and melanosomes among others.87,88 The resultingacidic environment is crucial for different biological processesincluding uncoupling of internalized ligand-receptor com-plexes, recycling of receptors, processing and degradation ofproteins through pH-dependent enzymes, intracellular traf-ficking and coupled transport of small molecules.89 Inaddition, V-ATPases are present on the plasma membraneof specific cell types where they have a role, for example, inurinary acidification and bone resorption.90 The localizationof (P)RR/ATP6ap2 mainly in intracellular membranes and onlya minor part at the plasma membrane33,45 concurs with thisdistribution of V-ATPase. Moreover, sequence analysis predictedmotifs in the cytosolic domain of (P)RR/ATP6ap2 targeting theprotein to distinct intracellular vesicle compartments.48

Experimental and clinical studies have emphasized thecritical roles of V-ATPases in physiology and pathophysio-logy. Table 1 summarizes genetic defects for V-ATPasesubunits and the large range of their associated phenotypesin different animal models and humans. In the mouse, gene



06/06/2013 6

Table 1 | Summary of genetic alterations in V-ATPase subunits and their associated phenotypes

Subunit Alteration Organism/cell line Phenotype Reference

V1 peripheral domainA morpho. Zebrafish Impaired acid secretion and ion balance 117

A mut. Neurospora Progeny not viable 118

A mut. Yeast Mislocalization of vacuolar proteins 119

B mut. Yeast Mislocalization of vacuolar proteins 119

B KO Yeast Conditional lethality 120

B KO Drosophila Lethal phenotype at larval stage 121

B1 mut. Human Distal renal tubular acidosis with hearing loss 122,123

B1 mut. Human Distal renal tubular acidosis with preserved hearing 124

B1 KO Mouse Severe metabolic acidosis after oral acid loading 125

B1 mut. Rat IMCD cells Impaired proton pump assembly and trafficking 126

C dsRNA Caenorhabditis elegans Impaired ovulation and embryogenesis 127

C1 siRNA Mouse BMMs Defective F-actin ring formation and osteoclastacidification activity

128

E mut. Yeast Impaired V-ATPase function 129

E1 mut. Zebrafish Oculocutaneous albinism, small head and eyes,CNS necrosis, and embryonic lethality

10,97

F mut. Zebrafish Oculocutaneous albinism, small head and eyes,CNS necrosis, and embryonic lethality

10,97

H mut. Zebrafish Oculocutaneous albinism, small head and eyes,CNS necrosis, and embryonic lethality

10,97

V0 membrane-embedded domaina mut. C. elegans Blockade of apical secretion of exosomes 130

a1 morpho. Zebrafish Impaired formation of phagolysosomes andclearance of apoptotic neurons

131

a1 mut. Drosophila neurons Blockade in synaptic vesicle fusion with thepresynaptic membrane

132

a2 mut. Human Defective protein glycosylation and cutis laxa 133

a3 KO Mouse pancreatic b-cells Defective insulin secretion 134

a3 KO Mouse Severe osteopetrosis due to loss of extracellularacidification

135

a3 mut. Human Deficient bone resorption and osteopetrosis 136–138

a3 mut. Mouse Deficient bone resorption and osteopetrosis 139

a3 siRNA Rat osteoclasts Decreased acidification and osteoclast inactivation 140

a4 mut. Human Distal renal tubular acidosis with hearing loss 123

c siRNA Rat BMMs Decreased osteoclast differentiation and boneresorption in vitro

141

c KO Mouse Disruption of the Golgi apparatus and embryonicmortality

142,143

c siRNA Human carcinoma in mice Decreased tumor invasion and metastasis 144

c mut. Zebrafish oculocutaneous albinism, retinal defects,and embryonic lethality

10,97

c mut. Yeast Mislocalization of vacuolar proteins 119

c KO Yeast Conditional lethality 120

c dsRNA C. elegans Impaired ovulation and embryogenesis 127

c’’ dsRNA C. elegans Impaired ovulation and embryogenesis 127

d KO Mouse Embryonic mortality 145

d1 mut. Zebrafish Oculocutaneous albinism, small head and eyes,CNS necrosis, and embryonic lethality

10,97

d2 KO Mouse Impaired osteoclast fusion and increased boneformation

146

d2 shRNA Mouse BMMs Impaired osteoclast differentiation and extracellularacidification

147

Accessory subunitsAc45 mut. Zebrafish Oculocutaneous albinism, small head and eyes,

CNS necrosis and embryonic lethality

10,97

Ac45 KO Mouse Embryonic mortality 103

Ac45 mut. Mouse BMMs and RAW264.7 cells Altered association with V0 complex and decreasedbone resorption

148

ATP6ap2 mut. Zebrafish Oculocutaneous albinism, small head and eyes,CNS necrosis and embryonic lethality

10

ATP6ap2 mut. Xenopus Defects in melanocytes and eye pigmentation,small head, shortened tail

9

ATP6ap2 mut. Human X-linked mental retardation and epilepsy 72

Abbreviations: BMMs, bone marrow-derived monocytes; CNS, central nervous system; dsRNA, double-stranded RNA; IMCD, inner medullary collecting duct; KO, knockout;morpho., morpholino antisense oligonucleotide injection; mut., mutation; shRNA, small hairpin RNA; siRNA, small interfering RNA.



06/06/2013 7

deletions of subunit c, d and the accessory protein Ac45 arelethal. In contrast, mutations in other subunits lead tovarious defects such as renal acidosis, cutis laxa (loose skin)and osteopetrosis suggesting that some subunits havedifferent roles and localizations. Indeed, during osteoclastdifferentiation, V-ATPases are relocated to the membraneand permit an efficient extracellular acidification for boneresorption.91 Like other V-ATPase subunits, the (P)RR/ATP6ap2/ATP6ap2 has been described to be upregulatedduring this differentiation in vitro but its role in osteoclastfunction is yet unknown.92 V-ATPases and endosomalacidification are also involved in the entry process of virusesinto cells93 and (P)RR/ATP6ap2, among other V-ATPasesubunits, has recently been discovered to be pivotal forinfluenza virus infection.94,95

Interestingly, the only yet described human (P)RR/ATP6ap2 mutation results in the deletion of exon 4 [D4-(P)RR/ATP6ap2] and is associated with X-linked mentalretardation (XLMR) and epilepsy.72 This may be explained bya defect in the function of synaptic vesicles due todysfunctional acidification because only one copy of the V-ATPase, including (P)RR/ATP6ap2, is present per vesicle.88

Thus, the full functionality of this single V-ATPase moleculeis critical to supply the driving force for vesicle trafficking,neurotransmitter uptake, and exocytosis and the lack of(P)RR/ATP6ap2 may impair neurotransmission. Moreover,the intracellular signaling triggered by (P)RR/ATP6ap2 after(pro)renin binding may be important for disease etiologybecause immortalized lymphocytes of one patient withXLMR did not show any ERK1/2 activation in the presenceof renin.72 Furthermore, Contrepas et al.96 showed that theD4-(P)RR mutant affects the trafficking of this receptor tothe neurite tips in vitro. However, further studies arewarranted to clarify the pathogenesis of the (P)RR/ATP6ap2-dependent form of XLMR.

(P)RR/ATP6ap2 AND V-ATPase IN EMBRYONICDEVELOPMENT

Actually, the most relevant data for a crucial role of (P)RR/ATP6ap2 as a V-ATPase accessory subunit have emerged fromstudies with lower vertebrates such as zebrafish and Xenopusfrogs. Indeed, zebrafish with insertional mutations in genescoding for V-ATPase subunits and (P)RR/ATP6ap2 share acommon embryonic phenotype10,97 (http://zfin.org). At day2 of development, all mutants including the ones for (P)RR/ATP6ap2 displayed abnormal body and eye pigmentation. Atday 5, all V-ATPase mutants developed multiple organ defects(head, liver, gut, heart) and necrosis in the central nervoussystem finally resulting in embryonic lethality. Accordingly,when we knocked down (P)RR/ATP6ap2 in zebrafishembryos using morpholino antisense oligonucleotide injec-tions we also obtained abnormal pigmentation in the eyes,head, and trunk and early lethality (G Sihn, unpublisheddata). The pigmentation phenotypes may reflect a defect inmelanosome acidification and maturation because V-ATPaseis essential in the first steps of this process.98 The defects in

the nervous system may result from dysfunctional synapticvesicles as already discussed above for the XLMR due to the(P)RR/ATP6ap2 mutation.

However, very recent data from Xenopus embryos andcultured cells provide another intriguing explanation forthese phenotypes. Cruciat et al.9 characterized V-ATPase and(P)RR/ATP6ap2 as essential components of Wnt signaling,which is crucial for several processes in embryonic develop-ment including neural patterning99,100 and melanocytedifferentiation.101 They show that phosphorylation of theWnt coreceptor LRP6, and thereby activation of intracellularWnt signaling, is depending on its sequestration in vesicles,which are acidified by V-ATPase action (Figure 1). Theauthors provided evidence that the pivotal link between LRP6and V-ATPase is formed by (P)RR/ATP6ap2 (Figure 1). Thisprocess is independent of (pro)renin9 and, therefore,inhibitors of its interaction with (P)RR/ATP6ap2 (such asHRP) should not interfere with Wnt-signaling. If however infuture, novel drugs become available, which inhibit (P)RR/ATP6ap2 action, their influence on Wnt signaling need to becritically assessed to avoid serious side effects.

In accordance with the lethal phenotype observed in(P)RR/ATP6ap2-deficient zebrafish and Xenopus frogs, ourmultiple attempts to generate (P)RR/ATP6ap2 knockoutmice also failed and only one low-level chimeric animal wasobtained without germ line transmission of the mutation(Figure 3). Interestingly, for the V-ATPase accessory subunitAc45, which has also been shown to co-purify with the

Injectedblastocysts

473 16/31 30 1 0

Pregnantfosters

Pups ChimerasKnockout

mice

Figure 3 | Attempts to generate (P)RR/ATP6ap2-knockoutmice. (a) Number of pregnant mice, pups, and chimeras obtainedafter blastocyst injection of (P)RR/ATP6ap2-knockout ES cells(RST307 from BayGenomics, http://baygenomics.ucsf.edu) andre-implantation into foster mothers (C57Bl/6); no knockout mousewas obtained subsequent to the breeding of the only chimera.(b) Aspect of the only chimera, displaying very low chimerismin the fur (arrowhead). (c) As comparison, a control chimeraobtained with non-modified ES cells with similar proceduresdisplaying 70–80% chimerism and germ line transmission.



06/06/2013 8

http://zfin.org

http://baygenomics.ucsf.edu

V-ATPase in bovine adrenal chromaffin granules,8,102 thegeneration of knockout animals led to the same result: onlyone low-chimeric mouse.103 Both accessory proteins areencoded by the X-chromosome and, thus, embryonic stem(ES) cells (which are male) with a targeted mutation lack therespective proteins completely. Such cells can obviously notparticipate in the development of a full organism and even aminor contribution of them leads to the death of chimeras.These findings are also in line with knockout studies on otherV-ATPase subunits (Table 1) that point to an essential role forthe V-ATPase during embryonic development, which mayinvolve Wnt signaling according to the data of Cruciat et al.9

Furthermore, b-catenin-knockout ES cells, which completelylack canonical Wnt signaling, also do not allow embryonicdevelopment when injected into host blastocysts.104

(P)RR/ATP6ap2, V-ATPase and Wnt signaling in kidney

How do V-ATPase functions relate to (P)RR/ATP6ap2 in thekidney? The first evidence of a functional link between the(P)RR/ATP6ap2 and the V-ATPase in this organ has recentlybeen provided by Advani et al.29 First, combining in situhybridization, immunohistochemistry and electron micro-scopy, they have mapped the expression of the (P)RR/ATP6ap2 in the distal nephron and collecting ducts. There,(P)RR/ATP6ap2 was predominantly expressed in microvilli atthe apical surface of A-type intercalated cells where it co-localized with the V-ATPase. In addition, in culturedcollecting duct cells, they described that not only (P)RR/ATP6ap2 downregulation by siRNA but also bafilomycin, aselective V-ATPase inhibitor attenuated the increase in ERK1/2 phosphorylation induced by either renin or prorenin. Theseresults show that in the presence of AT1 and AT2 antagonists(that is, AngII-independent), (pro)renin binds (P)RR/ATP6ap2and activates V-ATPases followed by ERK1/2 activation.

Many questions are raised by this study. Given that theantibody they used targets the C terminus of the (P)RR/ATP6ap2, it is not clear whether the main form at themembrane is the full protein or the furin-cleaved M8-9fragment initially described by Ludwig et al.8 Moreover,(P)RR/ATP6ap2 is located on the luminal membrane of theintercalated cells and, thus, can only interact with (pro)reninin the urine. Indeed, prorenin and renin have been found inthe urine of animals and humans,105,106 but it is not clearwhether under physiological conditions the urinary concen-trations reach sufficient levels to initiate (P)RR/ATP6ap2signaling. (Pro)renin is, however, also produced in neighboringprincipal cells and may induce a paracrine interaction.107,108

Wnt signaling is a fundamental molecular pathway inkidney organogenesis and physiology associated with numer-ous kidney diseases such as cancer, fibrosis, cystic disease,renal failure, and diabetic nephropathy.109,110 Recent studiesshowed that aberrant activation of the Wnt-signaling path-way promotes renal fibrosis, glomerulosclerosis, and protei-nuria.111,112 In contrast, inhibition of this pathway was alsoassociated with kidney dysfunction and cystic ciliopa-thy.113,114 It is predictable that (P)RR/ATP6ap2, based on

its pivotal function in Wnt signaling,9 is also implicatedin such processes. Further studies with cell-type specificknockout mouse models are warranted to clarify this issue.

CONCLUSIONS

Can one formulate a unifying hypothesis about the inter-connection between (pro)renin, (P)RR/ATP6ap2, V-ATPaseand Wnt signaling? Possibly, (P)RR/ATP6ap2 is a moleculewhose main function is the correct targeting and/orfunctional assistance of V-ATPase in specific vesicles andmembrane compartments such as synaptic vesicles, melano-somes or the specialized plasma membranes of osteoclastsand A-intercalated cells. It may also be important for theformation and/or function of so called ‘signaling endo-somes’.115 In these compartments, receptor/ligand complexesare sequestered, acidified, and acquire new competences. Inthe Wnt pathway, their acidic environment is essential forLRP6 phosphorylation and intracellular signal transductionto activate b-catenin.9 G-protein-coupled receptors, forexample, AT1, need to be sequestered into such endosomesto interact with b-arrestins and to initiate mitogenic signalingincluding ERK1/2 phosphorylation (Figure 1).116 It isunknown whether acidification of the vesicle by V-ATPaseis also required for this process. These functions of (P)RR/ATP6ap2 and V-ATPase are most likely independent of(pro)renin.9 Nevertheless, (pro)renin may bind to (P)RR/ATP6ap2 in signaling endosomes, which could explain itsnon-proteolytic activation (in the acid environment) and alsothe observed slow-onset ERK1/2 phosphorylation, if oneassumes that the (pro)renin/(P)RR/ATP6ap2 complex cansomehow activate the b-arrestin-MAPK pathway (Figure 1).At this time, a lot of this hypothesis is pure speculation andneeds to be verified by experimental approaches. Besidescultured cells and lower vertebrate models, cell-type specific(P)RR/ATP6ap2 knockout mice will be instrumental to thispurpose.

DISCLOSUREAll the authors declared no competing interests.

REFERENCES1. Paul M, Poyan Mehr A, Kreutz R. Physiology of local renin-angiotensin

systems. Physiol Rev 2006; 86: 747–803.2. Bader M. Tissue renin-angiotensin-aldosterone systems: targets for

pharmacological therapy. Annu Rev Pharmacol Toxicol 2010; 50:

439–465.3. Nguyen G, Delarue F, Burckle C et al. Pivotal role of the renin/prorenin

receptor in angiotensin II production and cellular responses to renin.

J Clin Invest 2002; 109: 1417–1427.4. Huang Y, Noble NA, Zhang J et al. Renin-stimulated TGF-beta1

expression is regulated by a mitogen-activated protein kinase in

mesangial cells. Kidney Int 2007; 72: 45–52.5. Huang Y, Wongamorntham S, Kasting J et al. Renin increases mesangial

cell transforming growth factor-beta1 and matrix proteins through

receptor-mediated, angiotensin II-independent mechanisms. Kidney Int

2006; 69: 105–113.6. Bader M, Ganten D. Regulation of renin: new evidence from cultured

cells and genetically modified mice. J Mol Med 2000; 78: 130–139.7. Feldman DL, Jin L, Xuan H et al. Effects of aliskiren on blood

pressure, albuminuria, and (pro)renin receptor expression in diabetic

TG(mRen-2)27 rats. Hypertension 2008; 52: 130–136.



06/06/2013 9

8. Ludwig J, Kerscher S, Brandt U et al. Identification and characterizationof a novel 9.2-kDa membrane sector-associated protein of vacuolarproton-ATPase from chromaffin granules. J Biol Chem 1998; 273:10939–10947.

9. Cruciat CM, Ohkawara B, Acebron SP et al. Requirement of proreninreceptor and vacuolar H+-ATPase-mediated acidification for Wntsignaling. Science 2010; 327: 459–463.

10. Amsterdam A, Nissen RM, Sun Z et al. Identification of 315 genesessential for early zebrafish development. Proc Natl Acad Sci USA 2004;101: 12792–12797.

11. Pitarresi TM, Rubattu S, Heinrikson R et al. Reversible cryoactivation ofrecombinant human prorenin. J Biol Chem 1992; 267: 11753–11759.

12. Suzuki F, Hayakawa M, Nakagawa T et al. Human prorenin has ‘gate andhandle’ regions for its non-proteolytic activation. J Biol Chem 2003; 278:22217–22222.

13. Danser AH, Derkx FH, Schalekamp MA et al. Determinants ofinterindividual variation of renin and prorenin concentrations: evidencefor a sexual dimorphism of (pro)renin levels in humans. J Hypertens1998; 16: 853–862.

14. Danser AH, Deinum J. Renin, prorenin and the putative (pro)reninreceptor. Hypertension 2005; 46: 1069–1076.

15. Krop M, de Bruyn JH, Derkx FH et al. Renin and prorenin disappearancein humans post-nephrectomy: evidence for binding? Front Biosci 2008;13: 3931–3939.

16. Krop M, Danser AH. Circulating versus tissue renin-angiotensin system:on the origin of (pro)renin. Curr Hypertens Rep 2008; 10: 112–118.

17. van Kesteren CA, Saris JJ, Dekkers DH et al. Cultured neonatal rat cardiacmyocytes and fibroblasts do not synthesize renin or angiotensinogen:evidence for stretch-induced cardiomyocyte hypertrophy independentof angiotensin II. Cardiovasc Res 1999; 43: 148–156.

18. Maru I, Ohta Y, Murata K et al. Molecular cloning and identificationof N-acyl-D-glucosamine 2-epimerase from porcine kidney as arenin-binding protein. J Biol Chem 1996; 271: 16294–16299.

19. Schmitz C, Gotthardt M, Hinderlich S et al. Normal blood pressure andplasma renin activity in mice lacking the renin-binding protein, a cellularrenin inhibitor. J Biol Chem 2000; 275: 15357–15362.

20. Saris JJ, Derkx FH, De Bruin RJ et al. High-affinity prorenin binding tocardiac man-6-P/IGF-II receptors precedes proteolytic activation torenin. Am J Physiol Heart Circ Physiol 2001; 280: H1706–H1715.

21. Saris JJ, Derkx FH, Lamers JM et al. Cardiomyocytes bind and activatenative human prorenin : role of soluble mannose 6-phosphatereceptors. Hypertension 2001; 37: 710–715.

22. Saris JJ, van den Eijnden MM, Lamers JM et al. Prorenin-inducedmyocyte proliferation: no role for intracellular angiotensin II.Hypertension 2002; 39: 573–577.

23. van den Eijnden MM, Saris JJ, de Bruin RJ et al. Prorenin accumulationand activation in human endothelial cells: importance of mannose6-phosphate receptors. Arterioscler Thromb Vasc Biol 2001; 21:911–916.

24. van Kesteren CA, Danser AH, Derkx FH et al. Mannose 6-phosphatereceptor-mediated internalization and activation of prorenin bycardiac cells. Hypertension 1997; 30: 1389–1396.

25. Nguyen G, Contrepas A. Physiology and pharmacology of the (pro)reninreceptor. Curr Opin Pharmacol 2008; 8: 127–132.

26. Su AI, Wiltshire T, Batalov S et al. A gene atlas of the mouse and humanprotein-encoding transcriptomes. Proc Natl Acad Sci USA 2004; 101:6062–6067.

27. Ichihara A, Kaneshiro Y, Takemitsu T et al. The (pro)renin receptor andthe kidney. Semin Nephrol 2007; 27: 524–528.

28. Sakoda M, Ichihara A, Kurauchi-Mito A et al. Aliskiren inhibitsintracellular angiotensin II levels Without Affecting (Pro)renin ReceptorSignals in Human Podocytes. Am J Hypertens 2010; 23: 575–580.

29. Advani A, Kelly DJ, Cox AJ et al. The (Pro)renin receptor: site-specific andfunctional linkage to the vacuolar H+-ATPase in the kidney. Hypertension2009; 54: 261–269.

30. Nabi AH, Kageshima A, Uddin MN et al. Binding properties of ratprorenin and renin to the recombinant rat renin/prorenin receptorprepared by a baculovirus expression system. Int J Mol Med 2006; 18:483–488.

31. Nguyen G, Delarue F, Berrou J et al. Specific receptor binding of renin onhuman mesangial cells in culture increases plasminogen activatorinhibitor-1 antigen. Kidney Int 1996; 50: 1897–1903.

32. Batenburg WW, Krop M, Garrelds IM et al. Prorenin is the endogenousagonist of the (pro)renin receptor. Binding kinetics of renin and proreninin rat vascular smooth muscle cells overexpressing the human(pro)renin receptor. J Hypertens 2007; 25: 2441–2453.

33. Saris JJ, t Hoen PA, Garrelds IM et al. Prorenin induces intracellularsignaling in cardiomyocytes independently of angiotensin II.Hypertension 2006; 48: 564–571.

34. Sakoda M, Ichihara A, Kaneshiro Y et al. (Pro)renin receptor-mediatedactivation of mitogen-activated protein kinases in human vascularsmooth muscle cells. Hypertens Res 2007; 30: 1139–1146.

35. Melnyk RA, Tam J, Boie Y et al. Renin and prorenin activate pathwaysimplicated in organ damage in human mesangial cells independentof angiotensin II production. Am J Nephrol 2009; 30: 232–243.

36. Feldt S, Batenburg WW, Mazak I et al. Prorenin and renin-inducedextracellular signal-regulated kinase 1/2 activation in monocytes isnot blocked by aliskiren or the handle-region peptide. Hypertension2008; 51: 682–688.

37. Kaneshiro Y, Ichihara A, Takemitsu T et al. Increased expression ofcyclooxygenase-2 in the renal cortex of human prorenin receptorgene-transgenic rats. Kidney Int 2006; 70: 641–646.

38. Zhang J, Noble NA, Border WA et al. Receptor-dependent proreninactivation and induction of PAI-1 expression in vascular smooth musclecells. Am J Physiol Endocrinol Metab 2008; 295: E810–E819.

39. Huang J, Siragy HM. Glucose promotes the production of interleukine-1beta and cyclooxygenase-2 in mesangial cells via enhanced (Pro)reninreceptor expression. Endocrinology 2009; 150: 5557–5565.

40. Matavelli LC, Huang J, Siragy HM. (Pro)renin receptor contributes todiabetic nephropathy through enhancing renal inflammation. Clin ExpPharmacol Physiol 2009; 37: 277–282.

41. Alcazar O, Cousins SW, Striker GE et al. (Pro)renin receptor is expressedin human retinal pigment epithelium and participates in extracellularmatrix remodeling. Exp Eye Res 2009; 89: 638–647.

42. Satofuka S, Ichihara A, Nagai N et al. (Pro)renin receptor promoteschoroidal neovascularization by activating its signal transductionand tissue renin-angiotensin system. Am J Pathol 2008; 173:1911–1918.

43. Satofuka S, Ichihara A, Nagai N et al. (Pro)renin receptor-mediated signaltransduction and tissue renin-angiotensin system contribute todiabetes-induced retinal inflammation. Diabetes 2009; 58: 1625–1633.

44. Uraoka M, Ikeda K, Nakagawa Y et al. Prorenin induces ERK activation inendothelial cells to enhance neovascularization independently of therenin-angiotensin system. Biochem Biophys Res Commun 2009; 390:1202–1207.

45. Schefe JH, Menk M, Reinemund J et al. A novel signal transductioncascade involving direct physical interaction of the renin/proreninreceptor with the transcription factor promyelocytic zinc finger protein.Circ Res 2006; 99: 1355–1366.

46. Schefe JH, Neumann C, Goebel M et al. Prorenin engages the (pro)reninreceptor like renin and both ligand activities are unopposed by aliskiren.J Hypertens 2008; 26: 1787–1794.

47. Krebs C, Hamming I, Sadaghiani S et al. Antihypertensive therapyupregulates renin and (pro)renin receptor in the clipped kidney ofGoldblatt hypertensive rats. Kidney Int 2007; 72: 725–730.

48. Burckle C, Bader M. Prorenin and its ancient receptor. Hypertension 2006;48: 549–551.

49. Cousin C, Bracquart D, Contrepas A et al. Soluble form of the (pro)reninreceptor generated by intracellular cleavage by furin is secreted inplasma. Hypertension 2009; 53: 1077–1082.

50. Bader M. The second life of the (pro)renin receptor. J Renin AngiotensinAldosterone Syst 2007; 8: 205–208.

51. Deinum J, Ronn B, Mathiesen E et al. Increase in serum proreninprecedes onset of microalbuminuria in patients with insulin-dependentdiabetes mellitus. Diabetologia 1999; 42: 1006–1010.

52. Luetscher JA, Kraemer FB, Wilson DM et al. Increased plasma inactiverenin in diabetes mellitus. A marker of microvascular complications. NEngl J Med 1985; 312: 1412–1417.

53. Stankovic AR, Fisher ND, Hollenberg NK. Prorenin and angiotensin-dependent renal vasoconstriction in type 1 and type 2 diabetes. J AmSoc Nephrol 2006; 17: 3293–3299.

54. Veniant M, Menard J, Bruneval P et al. Vascular damage withouthypertension in transgenic rats expressing prorenin exclusively in theliver. J Clin Invest 1996; 98: 1966–1970.

55. Prescott G, Silversides DW, Reudelhuber TL. Tissue activity of circulatingprorenin. Am J Hypertens 2002; 15: 280–285.

56. Methot D, Silversides DW, Reudelhuber TL. In vivo enzymatic assayreveals catalytic activity of the human renin precursor in tissues. Circ Res1999; 84: 1067–1072.

57. Hsueh WA, Luetscher JA, Carlson EJ et al. Changes in active and inactiverenin throughout pregnancy. J Clin Endocrinol Metab 1982; 54:1010–1016.



06/06/2013 10

58. Sealey JE, Wilson M, Morganti AA et al. Changes in active and inactiverenin throughout normal pregnancy. Clin Exp Hypertens A 1982; 4:2373–2384.

59. Skinner SL, Cran EJ, Gibson R et al. Angiotensins I and II, active and, reninsubstrate, renin activity, and angiotensinase in human liquor amnii andplasma. Am J Obstet Gynecol 1975; 121: 626–630.

60. Campbell DJ, Karam H, Menard J et al. Prorenin contributes toangiotensin peptide formation in transgenic rats with rat proreninexpression targeted to the liver. Hypertension 2009; 54: 1248–1253.

61. Peters B, Grisk O, Becher B et al. Dose-dependent titration of proreninand blood pressure in Cyp1a1ren-2 transgenic rats: absence of prorenin-induced glomerulosclerosis. J Hypertens 2008; 26: 102–109.

62. Peters J, Schluter T, Riegel T et al. Lack of cardiac fibrosis in a new modelof high prorenin hyperaldosteronism. Am J Physiol Heart Circ Physiol2009; 297: H1845–H1852.

63. Mercure C, Prescott G, Lacombe MJ et al. Chronic increases in circulatingprorenin are not associated with renal or cardiac pathologies.Hypertension 2009; 53: 1062–1069.

64. Siragy HM, Huang J. Renal (pro)renin receptor upregulation in diabeticrats through enhanced angiotensin AT1 receptor and NADPH oxidaseactivity. Exp Physiol 2008; 93: 709–714.

65. Freundlich M, Quiroz Y, Zhang Z et al. Suppression of renin-angiotensingene expression in the kidney by paricalcitol. Kidney Int 2008; 74:1394–1402.

66. Hirose T, Mori N, Totsune K et al. Increased expression of (pro)reninreceptor in the remnant kidneys of 5/6 nephrectomized rats. Regul Pept2010; 159: 93–99.

67. Hirose T, Mori N, Totsune K et al. Gene expression of (pro)renin receptoris upregulated in hearts and kidneys of rats with congestive heartfailure. Peptides 2009; 30: 2316–2322.

68. Ichihara A, Kaneshiro Y, Takemitsu T et al. Nonproteolytic activation ofprorenin contributes to development of cardiac fibrosis in genetichypertension. Hypertension 2006; 47: 894–900.

69. Kaschina E, Scholz H, Steckelings UM et al. Transition fromatherosclerosis to aortic aneurysm in humans coincides with anincreased expression of RAS components. Atherosclerosis 2009; 205:396–403.

70. Hirose T, Hashimoto M, Totsune K et al. Association of (pro)reninreceptor gene polymorphism with blood pressure in Japanese men: theOhasama study. Am J Hypertens 2009; 22: 294–299.

71. Kaneshiro Y, Ichihara A, Sakoda M et al. Slowly progressive, angiotensinII-independent glomerulosclerosis in human (pro)renin receptor-transgenic rats. J Am Soc Nephrol 2007; 18: 1789–1795.

72. Ramser J, Abidi FE, Burckle CA et al. A unique exonic splice enhancermutation in a family with X-linked mental retardation and epilepsypoints to a novel role of the renin receptor. Hum Mol Genet 2005; 14:1019–1027.

73. Ichihara A, Hayashi M, Kaneshiro Y et al. Inhibition of diabeticnephropathy by a decoy peptide corresponding to the ‘handle’ regionfor nonproteolytic activation of prorenin. J Clin Invest 2004; 114:1128–1135.

74. Takahashi H, Ichihara A, Kaneshiro Y et al. Regression of nephropathydeveloped in diabetes by (Pro)renin receptor blockade. J Am SocNephrol 2007; 18: 2054–2061.

75. Ichihara A, Suzuki F, Nakagawa T et al. Prorenin receptor blockadeinhibits development of glomerulosclerosis in diabetic angiotensin IItype 1a receptor-deficient mice. J Am Soc Nephrol 2006; 17: 1950–1961.

76. de Cavanagh EM, Inserra F, Toblli J et al. Enalapril attenuates oxidativestress in diabetic rats. Hypertension 2001; 38: 1130–1136.

77. Onozato ML, Tojo A, Goto A et al. Oxidative stress and nitric oxidesynthase in rat diabetic nephropathy: effects of ACEI and ARB. Kidney Int2002; 61: 186–194.

78. Parving HH, Persson F, Lewis JB et al. Aliskiren combined with losartan intype 2 diabetes and nephropathy. N Engl J Med 2008; 358: 2433–2446.

79. Satofuka S, Ichihara A, Nagai N et al. Role of nonproteolytically activatedprorenin in pathologic, but not physiologic, retinal neovascularization.Invest Ophthalmol Vis Sci 2007; 48: 422–429.

80. Satofuka S, Ichihara A, Nagai N et al. Suppression of ocular inflammationin endotoxin-induced uveitis by inhibiting nonproteolytic activation ofprorenin. Invest Ophthalmol Vis Sci 2006; 47: 2686–2692.

81. Susic D, Zhou X, Frohlich ED et al. Cardiovascular effects of proreninblockade in genetically spontaneously hypertensive rats on normal andhigh-salt diet. Am J Physiol Heart Circ Physiol 2008; 295: H1117–H1121.

82. Krebs C, Weber M, Steinmetz O et al. Effect of (pro)renin receptorinhibition by a decoy peptide on renal damage in the clipped kidney ofGoldblatt rats. Kidney Int 2008; 74: 823–824.

83. Feldt S, Maschke U, Dechend R et al. The putative (pro)renin receptorblocker HRP fails to prevent (pro)renin signaling. J Am Soc Nephrol 2008;19: 743–748.

84. Muller DN, Klanke B, Feldt S et al. (Pro)renin receptor peptide inhibitor‘handle-region’ peptide does not affect hypertensive nephrosclerosis inGoldblatt rats. Hypertension 2008; 51: 676–681.

85. Nabi AH, Biswas KB, Nakagawa T et al. ‘Decoy peptide’ region(RIFLKRMPSI) of prorenin prosegment plays a crucial role in proreninbinding to the (pro)renin receptor. Int J Mol Med 2009; 24: 83–89.

86. Nguyen G, Danser AH. Prorenin and (pro)renin receptor: a review ofavailable data from in vitro studies and experimental models in rodents.Exp Physiol 2008; 93: 557–563.

87. Forgac M. Vacuolar ATPases: rotary proton pumps in physiology andpathophysiology. Nat Rev Mol Cell Biol 2007; 8: 917–929.

88. Takamori S, Holt M, Stenius K et al. Molecular anatomy of a traffickingorganelle. Cell 2006; 127: 831–846.

89. Marshansky V, Futai M. The V-type H+-ATPase in vesicular trafficking:targeting, regulation and function. Curr Opin Cell Biol 2008; 20: 415–426.

90. Hinton A, Bond S, Forgac M. V-ATPase functions in normal and diseaseprocesses. Pflugers Arch 2009; 457: 589–598.

91. Toyomura T, Murata Y, Yamamoto A et al. From lysosomes to the plasmamembrane: localization of vacuolar-type H+ -ATPase with the a3 isoformduring osteoclast differentiation. J Biol Chem 2003; 278: 22023–22030.

92. Nomiyama H, Egami K, Wada N et al. Identification of genes differentiallyexpressed in osteoclast-like cells. J Interferon Cytokine Res 2005; 25:227–231.

93. Guinea R, Carrasco L. Requirement for vacuolar proton-ATPase activityduring entry of influenza virus into cells. J Virol 1995; 69: 2306–2312.

94. Karlas A, Machuy N, Shin Y et al. Genome-wide RNAi screen identifieshuman host factors crucial for influenza virus replication. Nature 2010;463: 818–822.

95. Konig R, Stertz S, Zhou Y et al. Human host factors required for influenzavirus replication. Nature 2010; 463: 813–817.

96. Contrepas A, Walker J, Koulakoff A et al. A role of the (pro)renin receptorin neuronal cell differentiation. Am J Physiol Regul Integr Comp Physiol2009; 297: R250–R257.

97. Nuckels RJ, Ng A, Darland T et al. The vacuolar-ATPase complexregulates retinoblast proliferation and survival, photoreceptormorphogenesis, and pigmentation in the zebrafish eye. InvestOphthalmol Vis Sci 2009; 50: 893–905.

98. Tabata H, Kawamura N, Sun-Wada GH et al. Vacuolar-type H(+)-ATPasewith the a3 isoform is the proton pump on premature melanosomes.Cell Tissue Res 2008; 332: 447–460.

99. Ille F, Sommer L. Wnt signaling: multiple functions in neuraldevelopment. Cell Mol Life Sci 2005; 62: 1100–1108.

100. Malaterre J, Ramsay RG, Mantamadiotis T. Wnt-Frizzled signalling andthe many paths to neural development and adult brain homeostasis.Front Biosci 2007; 12: 492–506.

101. Dunn KJ, Brady M, Ochsenbauer-Jambor C et al. WNT1 and WNT3apromote expansion of melanocytes through distinct modes of action.Pigment Cell Res 2005; 18: 167–180.

102. Supek F, Supekova L, Mandiyan S et al. A novel accessory subunit forvacuolar H(+)-ATPase from chromaffin granules. J Biol Chem 1994; 269:24102–24106.

103. Schoonderwoert VT, Martens GJ. Targeted disruption of the mouse geneencoding the V-ATPase accessory subunit Ac45. Mol Membr Biol 2002;19: 67–71.

104. Huelsken J, Vogel R, Brinkmann V et al. Requirement for beta-catenin inanterior-posterior axis formation in mice. J Cell Biol 2000; 148: 567–578.

105. Nielsen AH, Hermann KL, Mazanti I et al. Urinary excretion of inactiverenin during blockade of the renal tubular protein reabsorption withlysine. J Hypertens 1989; 7: 77–82.

106. Yukimura T, Miura K, Matsushima Y et al. Urinary excretion of renin andits biochemical properties in dogs. Hypertension 1984; 6: 837–842.

107. Prieto-Carrasquero MC, Botros FT, Kobori H et al. Collecting duct renin: amajor player in angiotensin II-dependent hypertension. J Am SocHypertens 2009; 3: 96–104.

108. Kang JJ, Toma I, Sipos A et al. The collecting duct is the major source ofprorenin in diabetes. Hypertension 2008; 51: 1597–1604.

109. Schmidt-Ott KM, Barasch J. WNT/beta-catenin signaling in nephronprogenitors and their epithelial progeny. Kidney Int 2008; 74:1004–1008.

110. Pulkkinen K, Murugan S, Vainio S. Wnt signaling in kidney developmentand disease. Organogenesis 2008; 4: 55–59.

111. He W, Dai C, Li Y et al. Wnt/beta-catenin signaling promotes renalinterstitial fibrosis. J Am Soc Nephrol 2009; 20: 765–776.



06/06/2013 11

112. Dai C, Stolz DB, Kiss LP et al. Wnt/beta-catenin signaling promotespodocyte dysfunction and albuminuria. J Am Soc Nephrol 2009; 20:1997–2008.

113. Lin CL, Wang JY, Ko JY et al. Dickkopf-1 promotes hyperglycemia-induced accumulation of mesangial matrix and renal dysfunction. J AmSoc Nephrol 2010; 21: 124–135.

114. Lancaster MA, Louie CM, Silhavy JL et al. Impaired Wnt-beta-cateninsignaling disrupts adult renal homeostasis and leads to cystic kidneyciliopathy. Nat Med 2009; 15: 1046–1054.

115. Scita G, Di Fiore PP. The endocytic matrix. Nature 2010; 463: 464–473.116. Kovacs JJ, Hara MR, Davenport CL et al. Arrestin development: emerging

roles for beta-arrestins in developmental signaling pathways. Dev Cell2009; 17: 443–458.

117. Horng JL, Lin LY, Huang CJ et al. Knockdown of V-ATPase subunit A(atp6v1a) impairs acid secretion and ion balance in zebrafish (Daniorerio). Am J Physiol Regul Integr Comp Physiol 2007; 292: R2068–R2076.

118. Ferea TL, Bowman BJ. The vacuolar ATPase of Neurospora crassa isindispensable: inactivation of the vma-1 gene by repeat-induced pointmutation. Genetics 1996; 143: 147–154.

119. Klionsky DJ, Nelson H, Nelson N et al. Mutations in the yeast vacuolarATPase result in the mislocalization of vacuolar proteins. J Exp Biol 1992;172: 83–92.

120. Nelson H, Nelson N. Disruption of genes encoding subunits of yeastvacuolar H(+)-ATPase causes conditional lethality. Proc Natl Acad Sci USA1990; 87: 3503–3507.

121. Davies SA, Goodwin SF, Kelly DC et al. Analysis and inactivation ofvha55, the gene encoding the vacuolar ATPase B-subunit in Drosophilamelanogaster reveals a larval lethal phenotype. J Biol Chem 1996; 271:30677–30684.

122. Karet FE, Finberg KE, Nelson RD et al. Mutations in the gene encoding B1subunit of H+-ATPase cause renal tubular acidosis with sensorineuraldeafness. Nat Genet 1999; 21: 84–90.

123. Stover EH, Borthwick KJ, Bavalia C et al. Novel ATP6V1B1 and ATP6V0A4mutations in autosomal recessive distal renal tubular acidosis with newevidence for hearing loss. J Med Genet 2002; 39: 796–803.

124. Smith AN, Skaug J, Choate KA et al. Mutations in ATP6N1B, encoding anew kidney vacuolar proton pump 116-kD subunit, cause recessivedistal renal tubular acidosis with preserved hearing. Nat Genet 2000; 26:71–75.

125. Finberg KE, Wagner CA, Bailey MA et al. The B1-subunit of the H(+)ATPase is required for maximal urinary acidification. Proc Natl Acad SciUSA 2005; 102: 13616–13621.

126. Yang Q, Li G, Singh SK et al. Vacuolar H+ -ATPase B1 subunit mutationsthat cause inherited distal renal tubular acidosis affect proton pumpassembly and trafficking in inner medullary collecting duct cells. J AmSoc Nephrol 2006; 17: 1858–1866.

127. Oka T, Futai M. Requirement of V-ATPase for ovulation andembryogenesis in Caenorhabditis elegans. J Biol Chem 2000; 275:29556–29561.

128. Feng S, Deng L, Chen W et al. Atp6v1c1 is an essential component of theosteoclast proton pump and in F-actin ring formation in osteoclasts.Biochem J 2009; 417: 195–203.

129. Lu M, Vergara S, Zhang L et al. The amino-terminal domain of the Esubunit of vacuolar H(+)-ATPase (V-ATPase) interacts with the H subunitand is required for V-ATPase function. J Biol Chem 2002; 277:38409–38415.

130. Liegeois S, Benedetto A, Garnier JM et al. The V0-ATPase mediates apicalsecretion of exosomes containing Hedgehog-related proteins inCaenorhabditis elegans. J Cell Biol 2006; 173: 949–961.

131. Peri F, Nusslein-Volhard C. Live imaging of neuronal degradation bymicroglia reveals a role for v0-ATPase a1 in phagosomal fusion in vivo.Cell 2008; 133: 916–927.

132. Hiesinger PR, Fayyazuddin A, Mehta SQ et al. The v-ATPase V0 subunita1 is required for a late step in synaptic vesicle exocytosis in Drosophila.Cell 2005; 121: 607–620.

133. Kornak U, Reynders E, Dimopoulou A et al. Impaired glycosylation andcutis laxa caused by mutations in the vesicular H+-ATPase subunitATP6V0A2. Nat Genet 2008; 40: 32–34.

134. Sun-Wada GH, Toyomura T, Murata Y et al. The a3 isoform of V-ATPaseregulates insulin secretion from pancreatic beta-cells. J Cell Sci 2006;119: 4531–4540.

135. Li YP, Chen W, Liang Y et al. Atp6i-deficient mice exhibit severeosteopetrosis due to loss of osteoclast-mediated extracellularacidification. Nat Genet 1999; 23: 447–451.

136. Frattini A, Orchard PJ, Sobacchi C et al. Defects in TCIRG1 subunit of thevacuolar proton pump are responsible for a subset of human autosomalrecessive osteopetrosis. Nat Genet 2000; 25: 343–346.

137. Scimeca JC, Quincey D, Parrinello H et al. Novel mutations in the TCIRG1gene encoding the a3 subunit of the vacuolar proton pump in patientsaffected by infantile malignant osteopetrosis. Hum Mutat 2003; 21:151–157.

138. Susani L, Pangrazio A, Sobacchi C et al. TCIRG1-dependent recessiveosteopetrosis: mutation analysis, functional identification of thesplicing defects, and in vitro rescue by U1 snRNA. Hum Mutat 2004;24: 225–235.

139. Scimeca JC, Franchi A, Trojani C et al. The gene encoding the mousehomologue of the human osteoclast-specific 116-kDa V-ATPase subunitbears a deletion in osteosclerotic (oc/oc) mutants. Bone 2000; 26:207–213.

140. Hu Y, Nyman J, Muhonen P et al. Inhibition of the osteoclast V-ATPaseby small interfering RNAs. FEBS Lett 2005; 579: 4937–4942.

141. Laitala-Leinonen T, Lowik C, Papapoulos S et al. Inhibition ofintravacuolar acidification by antisense RNA decreases osteoclastdifferentiation and bone resorption in vitro. J Cell Sci 1999; 112(Part 21):3657–3666.

142. Inoue H, Noumi T, Nagata M et al. Targeted disruption of the geneencoding the proteolipid subunit of mouse vacuolar H(+)-ATPase leadsto early embryonic lethality. Biochim Biophys Acta 1999; 1413: 130–138.

143. Sun-Wada G, Murata Y, Yamamoto A et al. Acidic endomembraneorganelles are required for mouse postimplantation development. DevBiol 2000; 228: 315–325.

144. Lu X, Qin W, Li J et al. The growth and metastasis of humanhepatocellular carcinoma xenografts are inhibited by small interferingRNA targeting to the subunit ATP6L of proton pump. Cancer Res 2005;65: 6843–6849.

145. Miura GI, Froelick GJ, Marsh DJ et al. The d subunit of the vacuolarATPase (Atp6d) is essential for embryonic development. Transgenic Res2003; 12: 131–133.

146. Lee SH, Rho J, Jeong D et al. v-ATPase V0 subunit d2-deficient miceexhibit impaired osteoclast fusion and increased bone formation. NatMed 2006; 12: 1403–1409.

147. Wu H, Xu G, Li YP. Atp6v0d2 is an essential component of theosteoclast-specific proton pump that mediates extracellular acidificationin bone resorption. J Bone Miner Res 2009; 24: 871–885.

148. Feng H, Cheng T, Pavlos NJ et al. Cytoplasmic terminus of vacuolar typeproton pump accessory subunit Ac45 is required for proper interactionwith V(0) domain subunits and efficient osteoclastic bone resorption.J Biol Chem 2008; 283: 13194–13204.



06/06/2013 12

Wiebke Malenke

Reduced Endoglin Activity Limits Cardiac Fibrosis and Improves Survival in Heart Failure Quelle: Circulation, 2012 Betreuer: Prof. Uwe Lendeckel (Medizinische Biochemie und Molekularbiologie) Was bedeutet mir dieses Thema persönlich? Im Rahmen meiner Promotionsarbeit beschäftige ich mich mit dem TGF-ß-Signalweg im Herzen bei Vorhofflimmern. TGF-ß ist u. A. verantwortlich für kardiales Remodelling (z. B. Fibrose). Endoglin ist eine Komponente dieses Signalweges und in sofern sehr interessant, weil es auf zweierlei Wege (als Korezeptor und als gelöste Form) das TGF-Signalling beeinflusst. Worauf kommt es bei diesem Thema am meisten an? Das Paper zeigt, dass Grundlagenforschung klinische Relevanz haben kann, indem potentielle neue therapeutische Ansätze aufzeigt werden. Außerdem liefert dieses Paper einen anschaulichen Einstieg in das Thema TGF-Signalweg, obwohl mir bewusst ist, dass es nur einen kleinen Teil darstellt. Was fasziniert mich selbst am Thema am meisten? Der Signalweg ist recht komplex und am Herzen wenig erforscht. Es ist unvorstellbar, wie viel im Herzen moduliert und reguliert wird durch z. B. Vorhofflimmern. Die Effekte sind sowohl mikroskopisch, als auch makroskopisch sichtbar (z. B. Fibrose), aber es meist nicht bekannt, welche molekularen Mechanismen dahinter stecken.

06/06/2013 13

Reduced Endoglin Activity Limits Cardiac Fibrosis andImproves Survival in Heart Failure

Navin K. Kapur, MD; Szuhuei Wilson, MS; Adil A. Yunis, BSc; Xiaoying Qiao, PhD;Emily Mackey, BA; Vikram Paruchuri, MD; Corey Baker, BSc; Mark J. Aronovitz, MS;

S. Ananth Karumanchi, MD; Michelle Letarte, PhD; David A. Kass, MD;Michael E. Mendelsohn, MD; Richard H. Karas, MD, PhD

Background—Heart failure is a major cause of morbidity and mortality worldwide. The ubiquitously expressed cytokine

transforming growth factor-b1 (TGFb1) promotes cardiac fibrosis, an important component of progressive heart failure.

Membrane-associated endoglin is a coreceptor for TGFb1 signaling and has been studied in vascular remodeling and

preeclampsia. We hypothesized that reduced endoglin expression may limit cardiac fibrosis in heart failure.

Methods and Results—We first report that endoglin expression is increased in the left ventricle of human subjects with

heart failure and determined that endoglin is required for TGFb1 signaling in human cardiac fibroblasts using

neutralizing antibodies and an siRNA approach. We further identified that reduced endoglin expression attenuates

cardiac fibrosis, preserves left ventricular function, and improves survival in a mouse model of pressure-overload–

induced heart failure. Prior studies have shown that the extracellular domain of endoglin can be cleaved and released

into the circulation as soluble endoglin, which disrupts TGFb1 signaling in endothelium. We now demonstrate that

soluble endoglin limits TGFb1 signaling and type I collagen synthesis in cardiac fibroblasts and further show that

soluble endoglin treatment attenuates cardiac fibrosis in an in vivo model of heart failure.

Conclusion—Our results identify endoglin as a critical component of TGFb1 signaling in the cardiac fibroblast and show

that targeting endoglin attenuates cardiac fibrosis, thereby providing a potentially novel therapeutic approach for

individuals with heart failure. (Circulation. 2012;125:2728-2738.)

Key Words: fibrosis n heart failure n ventricular remodeling

Heart failure is a major cause of morbidity and mortality

that affects .24 million individuals worldwide.1–3 Re-

gardless of the injurious mechanism, a decline in left ventric-

ular (LV) function increases LV pressure and activates

several signaling cascades that promote cardiomyocyte hy-

pertrophy and cardiac fibrosis, a process known as cardiac

remodeling. At each phase of cardiac remodeling, from acute

load to compensatory hypertrophy, various signaling cas-

cades are implicated.4 Among these, transforming growth

factor-b1 (TGFb1) is a profibrogenic cytokine that contrib-

utes to multiple fibroproliferative disorders, including cardiac

fibrosis associated with heart failure.5 In response to angio-

tensin II, TGFb1 expression is increased, converts fibroblasts

into myofibroblasts, and generates extracellular matrix pro-

teins such as type I collagen.6 Excess collagen deposition

exaggerates mechanical stiffness of the LV, impairs myocyte

contractility, disrupts electric coupling, and worsens tissue hyp-

oxia,4 thereby promoting heart failure. Given its central role in

stimulating fibrosis, TGFb1 has been nonselectively targeted in

heart failure models using multiple approaches, none of which

has produced clearly beneficial therapeutic effects.7,8

Editorial see p 2689Clinical Perspective on p 2738

TGFb1 signals through a heteromeric receptor complex

comprising a type II ligand binding receptor in association with

a type I activin-like kinase (ALK) signaling receptor. Once

activated by TGFb1, this receptor complex triggers phosphory-

lation of downstream effector proteins known as Smads (canon-

ical pathway) or mitogen-activated protein kinases (noncanoni-

cal pathway). Specifically, TGFb1-induced phosphorylation of

Smad-2/3 promotes type I collagen synthesis and fibrosis.5,6

Endoglin (CD105) is a 180-kDa homodimeric glycoprotein that

serves as a coreceptor for TGFb1 signaling. Over the past 2

decades, several lines of evidence have suggested that endoglin

Received November 14, 2011; accepted March 27, 2012.From the Molecular Cardiology Research Institute, Tufts Medical Center and Tufts University School of Medicine, Boston, MA (N.K.K., S.W., A.A.Y., X.Q.,

E.M., V.P., C.B., M.J.A., M.E.M., R.H.K.); Center for Vascular Biology, Departments of Medicine, Obstetrics and Gynecology, Surgery, and Pathology, BethIsrael Deaconess Medical Center and Harvard Medical School, Boston, MA (A.K.); Molecular Structure and Function Program, Hospital for Sick Children, andThe Heart and Stroke Foundation Richard Lewar Centre of Excellence, University of Toronto, Toronto, ON, Canada (M.L.); and Department of Medicine, JohnsHopkins University Medical Institutions, Baltimore, MD (D.A.K.). Dr Mendelsohn is currently at Merck and Co Inc, Rahway, NJ.

The online-only Data Supplement is available with this article at http://circ.ahajournals.org/lookup/suppl/doi:10.1161/CIRCULATIONAHA.

111.080002/-/DC1.

Correspondence to Navin K. Kapur, MD, Molecular Cardiology Research Institute, Tufts Medical Center, 800 Washington St, Box 80, Boston, MA02111. E-mail [email protected]© 2012 American Heart Association, Inc.

Circulation is available at http://circ.ahajournals.org DOI: 10.1161/CIRCULATIONAHA.111.080002

at Universitaet Greifswald on December 19, 2012http://circ.ahajournals.org/Downloaded from

06/06/2013 14

plays a critical role in vascular remodeling. First, loss-of-

function mutations in human endoglin result in the autosomal

dominant vascular dysplastic syndrome hereditary hemorrhagic

telangiectasia type 1, characterized by endoglin haploinsuffi-

ciency and visceral arteriovenous malformations.9 Second,

endoglin-null mice die at embryonic day 10.5 as a result of

impaired cardiovascular development and extraembryonic an-

giogenesis.10 However, endoglin-heterozygous mice (Eng1/2)

are viable, have reduced levels of endoglin, and have a pheno-

type that recapitulates that of hereditary hemorrhagic telangiec-

tasia type 1.11 The role of endoglin as a modulator of TGFb1

signaling in heart failure, where fibrosis plays a major role, has

not been explored to date.

The extracellular domain of endoglin can be proteolytically

cleaved by matrix metalloproteinase-14 and circulates as

soluble endoglin (sEng).12 We recently demonstrated that

levels of sEng correlate with clinical measures of heart

failure, including LV end-diastolic pressure and New York

Heart Association classification.13 From these observations,

we hypothesized that impaired function of the TGFb1 core-

ceptor endoglin limits TGFb1-induced collagen synthesis and

cardiac fibrosis, thereby identifying endoglin as a potentially

novel therapeutic target in heart failure. To explore this

hypothesis, we used a model of pressure-overload–induced

heart failure in Eng1/2 mice.

Methods

ReagentsWe purchased recombinant human sEng (1–587 amino acids corre-sponding to the extracellular domain of endoglin; R&D Systems) andrecombinant human TGFb1 (Sigma). Mouse monoclonal antibodiesto human endoglin (SC-73934), human type I collagen (SC-80497),human DDR-2 (SC-81707), and fibroblast marker (ER-TR7) werepurchased from Santa Cruz. A polyclonal antibody to the N-terminalregion of human endoglin (SC-19790) was purchased from SantaCruz. Goat polyclonal antibodies against mouse endoglin and type Icollagen were purchased from R&D Systems (BAF1320) and SantaCruz (SC-25974), respectively. Polyclonal antibodies to human andmouse phosphorylated Smad (pSmad)-2/3 (AB-3849), pSmad-1(06–702), and phosphorylated extracellular regulated kinase (ERK)-1/2 (05–797) were purchased from Millipore; polyclonal antibodiesto human and mouse total Smad-2/3 (3102), total Smad-1 (9743),and total ERK (9102) were purchased from Cell Signaling. Rabbitpolyclonal antibodies to mouse calcineurin were purchased from CellSignaling. A rat monoclonal antibody to mouse CD31 was purchasedfrom Pharmingen (01951A). ELISA kits for human and mouse sEngand TGFb1 were purchased from R&D Systems.

Human LV Tissue SamplingViable LV free wall tissue was obtained from human subjects withheart failure (n520) referred for LV assist device (LVAD) place-ment (HeartMate; Thoratec Corp, Pleasanton, CA). In 7 subjects, anadditional LV sample was obtained after LVAD support at the timeof cardiac transplantation. Control LV tissue was obtained from theNational Disease Research Interchange. All tissue was immediatelyfrozen in liquid nitrogen and stored at 280°C until further process-ing as described below. All surgical procedures and tissue harvestingwere performed in concordance with the National Institutes of Healthand Tufts University Institutional Review Board guidelines.

Mouse Model of Pressure-Overload–InducedHeart FailureAnimals were treated in compliance with the Guide for the Care and

Use of Laboratory Animals (National Academy of Science), and

protocols were approved by the Tufts Medical Center InstitutionalAnimal Care and Use Committee. Adult male 14- to 16-week-oldC57BL/6 wild-type (WT) and Eng1/2 mice underwent thoracicaortic constriction (TAC) as previously described.14,15 At 2, 4, and 10weeks after TAC, mice were euthanized and tissue was obtained forfurther analysis by real-time polymerase chain reaction, immuno-blotting, histology, and ELISA according to the manufacturer’sinstructions. Eng1/2 mice were generously provided by Dr MichelleLetarte, University of Toronto.

Physiological Characterization In VivoTransthoracic echocardiography and pressure-volume loop analyseswere performed on mice as previously described.14–16

TGFb1-Induced Type I Collagen Expression inCardiac FibroblastsThe Tufts Medical Center Institutional Review Board approved thecollection of human tissue for cell culture. Human cardiac fibroblasts(hCFs) were isolated from myocardial tissue harvested duringcardiac surgery at Tufts Medical Center, and mouse cardiac fibro-blasts (mCFs) were isolated from WT and Eng1/2 mice andstimulated with TGFb1 for analysis as previously described.16,17

Loss-of-Function Studies in hCFsFor neutralizing antibody studies, hCFs were pretreated with 0.5mg/mL of either an antibody to endoglin or control IgG isotype for24 hours in fibroblast basal medium without supplementation beforestimulation with TGFb1 (10 ng/mL). After 24 hours, cells wereharvested for real-time polymerase chain reaction and Western blotanalysis. For endoglin silencing experiments, 50 mmol/L siRNAstock (Ambion) was diluted to 1.0 nmol/L in Optimem (Invitrogen)and combined with 2 mL lipofectamine (Invitrogen) diluted in 98 mLOptimem. After 20 minutes of incubation, cells were exposed tohuman endoglin siRNA (Ambion; 145527), scrambled siRNA (neg-ative control; Ambion; 4390844), or GAPDH siRNA (positivecontrol; Ambion; 4390850). At various times between 24 and 48hours after transfection, cells were harvested for analysis.

Recombinant sEng Inhibition of Type I CollagenSynthesis In VitroSerum-starved hCFs were treated for 24 hours with recombinanthuman sEng, stimulated with TGFb1, and harvested for furtheranalysis.

Overexpression of Human sEng and Full-LengthEndoglin In VitroFor conditioned media studies, COS-1 (American Type CultureCollection) cells were transfected with adenovirus overexpressinghuman sEng (AdhsEng; generously provided by Dr S.A. Karuman-chi) or adenovirus with no transgene for 24 hours. Human sEnglevels in conditioned media were confirmed by ELISA. Conditionedmedia was then transferred into 12-well dishes containing serum-starved hCFs and then stimulated with TGFb1. For overexpressionstudies, hCFs were transfected with adenovirus expressing full-length endoglin (generously provided by Dr Calvin P. Vary),stimulated with TGFb1, and harvested for analysis.

sEng Inhibits Pressure-Overload–InducedCardiac FibrosisAdult male 14- to 16-week old C57BL/6 mice received intravenousinjections of adenovirus with no transgene or AdhsEng 1 day beforeTAC. Serum levels of human and mouse sEng were quantified byELISA. Four weeks after TAC, mice were euthanized and tissue wasobtained for further analysis.

Real-Time Quantitative PolymeraseChain ReactionFor all cell-based real-time polymerase chain reaction experiments,total RNA was extracted directly with Trizol (Invitrogen) and

Kapur et al Endoglin and Heart Failure 2729


06/06/2013 15

converted to cDNA with a High Capacity cDNA Reverse Transcrip-tion Kit (Applied Biosystems). For all real-time polymerase chainreaction experiments, samples were quantified in triplicate using 40cycles performed at 94°C for 30 seconds, 60°C for 45 seconds, and72°C for 45 seconds with an ABI Prism 7900 Sequence DetectionSystem using appropriate primers (see the online-only Data Supple-ment) as described.14–16

Immunoblot Analysis (Western)Total protein was extracted and quantified from tissue homogenatesor cultured cells as described.14–16 Immunoblot analysis was thenperformed as previously described using antibodies for human andmouse targeted proteins.

Histological Quantification of CardiacHypertrophy and FibrosisLV collagen abundance was quantified by picrosirius red staining asdescribed.18,19 Cardiomyocyte cross-sectional area and capillarydensity were quantified as described.14

Statistical AnalysisResults are presented as mean6SD. Intergroup comparisons weremade with 2-factor ANOVA. Two-way ANOVA was performed toexamine the effects of cardiac unloading by an LVAD and time onendoglin expression. Repeated-measures ANOVA was used asneeded to account for time. All multiple comparisons with a controlgroup were performed with the Dunnett method. Kaplan-Meieranalysis with log-rank testing was used for survival analysis. Allstatistical analyses were performed with SigmaStat version 3.1(Systat Software, Inc). An a level of P,0.05 was considered toindicate a significant effect or between-group difference.

Results

LV Endoglin Expression Is Increased in HumanHeart FailureTo determine whether endoglin expression is increased in

patients with heart failure, LV samples were obtained from

individuals with end-stage heart failure referred for surgical

implantation of an LVAD. An additional LV sample was

obtained in 7 subjects at the time of cardiac transplantation

after LVAD support to examine the effect of hemodynamic

unloading on endoglin expression. Compared with subjects

without heart failure, endoglin expression was increased in

the failing LV at the time of LVAD implantation (Figure 1A)

and reduced back to control levels after LVAD support

(Figure 1B). These findings indicate an association between

cardiac pressure overload and LV endoglin expression in

heart failure. To determine what cardiac cell types express

endoglin, cardiomyocytes, fibroblasts, and endothelial cells

were next isolated from WT mouse LVs. Endoglin was

expressed by cardiac fibroblasts and endothelium but not by

cardiac myocytes (Figure 1C and 1D).

Increased Membrane-Associated and CirculatingEndoglin Expression in Heart FailureTo explore the functional role of endoglin in heart failure, we

studied Eng1/2 mice. Compared with WT, LV endoglin

expression was lower in Eng1/2 mice (Figure 2A). We then

used the well-established mouse model of LV pressure

overload induced by TAC followed by tissue characterization

at 2, 4, and 10 weeks. In WT mice, compared with sham-

operated controls, LV endoglin mRNA was increased within

2 weeks and remained elevated at 4 and 10 weeks after TAC

(Figure 2B). LV endoglin protein expression was similarly

increased after 2 and 4 weeks of heart failure and returned to

normal levels by 10 weeks in WT mice (data not shown). No

change in endoglin levels was observed in the aorta distal to

the site of TAC ligature (Figure IA in the online-only Data

Supplement), suggesting a direct effect of cardiac pressure

overload on endoglin expression. Serum levels of sEng were

also elevated across all time points of pressure-overload–

induced heart failure (Figure 2C). TAC also increased LV

Figure 1. Increased endoglin expressionin the failing human left ventricle (LV). A,Endoglin expression in LV tissue fromhuman subjects without heart failure(non-HF) and with heart failure beforeinsertion of a LV assist device (pre-LVAD; *P,0.05 vs non-HF). B, LVendoglin expression in 7 subjects before(pre-LVAD) and after (post-LVAD) LVADsupport (*P,0.05 vs pre-LVAD). Repre-sentative Western blots are shownabove quantification graphs. C, Repre-sentative Western blot of endoglinexpression (relative to GAPDH) in iso-lated mouse endothelium, cardiomyo-cytes, and cardiac fibroblasts. D,Endoglin mRNA expression by isolatedmouse cardiomyocytes and cardiacfibroblasts (***P,0.001 between groups).

2730 Circulation June 5, 2012


06/06/2013 16

endoglin mRNA (Figure 2B) and circulating sEng levels

(Figure 2C) in the Eng1/2 mice, but levels were significantly

reduced compared with WT mice at each time point.

Reduced Endoglin Expression Preserves LVFunction and Promotes Survival in Heart FailureWe next examined the functional impact of reduced endoglin

expression in heart failure. Eng1/2 mice demonstrated pre-

served cardiac function and improved survival (88% versus

50%, respectively; P50.01) compared with WT mice after

TAC (Figure 2D–2F and the Table). Consistent with these

observations, WT mice manifest reduced total body weight at

both 4 and 10 weeks after TAC, whereas Eng1/2mice did not

(Table 1). Compared with baseline values, lung weights were

increased in both mouse groups after TAC; however, lung

weights were lower after 10 weeks of TAC in Eng1/2 mice

compared with WT mice (Table 1). After 4 weeks of TAC,

LV systolic and LV end-diastolic pressures were increased in

both WT and Eng1/2 mice compared with the respective

sham-operated controls. Although TAC-induced LV systolic

pressure was higher in Eng1/2 mice compared with WT mice

at both 4 and 10 weeks, LV end-diastolic pressure was not

different between WT and Eng1/2mice at any time point. LV

contractility (dP/dtmax) also decreased in a time-dependent

manner in WT mice but remained unchanged across all

phases of cardiac pressure overload in Eng1/2 mice (the

Table). Echocardiography demonstrated a significant time-

dependent TAC-induced reduction in LV fractional shortening

in WT but not in Eng1/2 mice. After 10 weeks of TAC, WT

mice demonstrated greater LV chamber dilatation and lower

posterior wall thickness compared with Eng1/2 mice (the

Table). These findings suggest that despite identical degrees of

LV pressure overload, reduced endoglin expression in the

Eng1/2 mice preserved LV function and improved survival.

Reduced Endoglin Expression PreservesTAC-Induced Cardiomyocyte Hypertrophy andPromotes Myocardial CapillarityTo study the mechanism underlying improved survival in

Eng1/2mice, we first examined changes in cardiac hypertrophy.

Across all time points, fold changes in LV mass normalized to

tibia length were similar between Eng1/2 and WT mice (the

Table). Cardiomyocyte cross-sectional area was also increased

to a similar degree in both WT and Eng1/2 mice after TAC

(Figure 3A and 3B). Expression of fetal genes and proteins such

as b-myosin heavy chain, SERCA, and calcineurin demon-

strated similar patterns in both WT and Eng1/2 mice after 4

weeks (Figure 3C-3E) and 10 weeks (Figure IIA and IIB in the

online-only Data Supplement) of TAC. Next, we observed a

significant increase in the capillary-to-cardiomyocyte ratio and

total capillary density in both WT and Eng1/2 mice after 4

weeks of TAC. Both measures of myocardial capillarity were

higher in Eng1/2 compared with WT mice after TAC (Figure

3F–3H). Taken together, these data support that TAC-induced

LV hypertrophy is preserved whereas myocardial capillarity is

enhanced in Eng1/2 mice compared with WT mice.

Reduced Endoglin Expression AttenuatesCardiac FibrosisCompared with controls, TAC induced a time-dependent

increase in LV collagen deposition in WT but not Eng1/2

mice (Figure 4A and 4B). Type I collagen mRNA expression

(Figure 4C) and protein expression (Figure 4D) were simi-

Figure 2. Reduced endoglin expression preserves cardiac function and survival in heart failure. A, Endoglin expression in left ventricular(LV) protein lysates from wild-type (WT) and Eng1/2 mice (n53 per group). B, LV endoglin mRNA expression in WT mice after thoracicaortic constriction (TAC; n56 per group). C, Levels of circulating soluble endoglin (sEng) in Eng1/2 mice vs WT (n56 per group). D,M-mode echocardiography in WT mice compared with Eng1/2 mice after TAC. E, Representative LV pressure-volume loops after 4weeks of TAC in Eng1/2 and WT mice. F, Kaplan-Meier survival curves in Eng1/2 and WT mice after TAC (n518 per group). *P,0.05vs WT-sham; †P,0.05 vs Eng1/2-sham; ‡P,0.05 vs WT at the corresponding time point.



06/06/2013 17

larly increased in WT mice after TAC. Eng1/2 mice exhib-

ited a small increase in type I collagen mRNA after 4 weeks

(Figure 4C) and 10 weeks (Figure IIIA in the online-only

Data Supplement) and a modest increase in type I collagen

protein expression after 10 weeks of heart failure (Figure 4D).

At all time points, type I collagen mRNA and protein were

lower in Eng1/2 mice compared with WT mice. Importantly,

similar increases in TGFb1 mRNA (Figure 4E) and active

TGFb1 protein (Figure IIIB in the online-only Data Supple-

ment) were observed in the LV of both WT and Eng1/2 mice

after 4 weeks of TAC. Levels of the downstream target of

TGFb1 signaling, pSmad-2/3, were increased by 4 and 10

weeks of TAC (Figure 4F and Figure IIIC in the online-only

Data Supplement) in WT mice. However, despite similar

increases in LV TGFb1 levels in Eng1/2 mice, TAC did not

induce a significant increase in pSmad-2/3. Consistent with this

impairment of TGFb1 signaling in Eng1/2 mice, expression of

plasminogen activator inhibitor-1 was increased in WT but not

Eng1/2 mice after 4 weeks of TAC (Figure IIID). To confirm

that reduced endoglin expression by Eng1/2 mice limits

TGFb1-induced cardiac fibrosis, mCFs were isolated from the

LV of WT and Eng1/2. TGFb1 simulation induced type I

collagen mRNA expression in WT-mCF but not Eng1/2-mCF

(Figure IIIE in the online-only Data Supplement). These data

identify that reduced endoglin expression attenuates TGFb1

signal transduction and limits cardiac fibrosis. These in vivo

observations led us to explore the dependence of TGFb1

profibrotic signaling on endoglin expression in vitro.

Canonical and Noncanonical Signaling inEndoglin-Deficient MiceWe next studied the expression of the TGFb type I receptors

ALK1 and ALK5 in hCFs from fresh tissue samples obtained

during cardiac surgery. We observed lower ALK1 mRNA

expression in hCFs compared with human umbilical vein endo-

thelial cells, whereas ALK5 expression was similar in both cell

types (Figure IVA in the online-only Data Supplement). Next,

TGFb1 stimulation increased pSmad-2/3 expression with no

change in pSmad-1 expression (Figure IVB in the online-only

Data Supplement). Compared with sham-operated controls, LV

pSmad-1 expression was increased in WT mice after 4 weeks of

TAC, whereas no change was observed in Eng1/2 mice (Figure

IVC in the online-only Data Supplement). Immunostaining

confirmed increased nuclear accumulation of pSmad-2/3 and

less accumulation of pSmad-1 in WT mice after TAC. In

contrast, no significant increase in nuclear accumulation of

pSmad-2/3 or pSmad-1 was observed in Eng1/2mice after TAC

(Figure IVD–IVG in the online-only Data Supplement).

Next, we studied phosphorylation of the TGFb1 nonca-

nonical signaling effector ERK. With an siRNA approach,

reduced endoglin expression attenuated TGFb1-induced

phosphorylation of ERK-1/2 in hCFs in vitro (Figure IVH in

the online-only Data Supplement). Consistent with this ob-

servation, phosphorylated ERK-1/2 expression was increased

in WT mice but not Eng1/2 mice after 4 weeks of TAC

(Figure 4I). These findings suggest that reduced endoglin

expression limits TGFb1 signaling via Smad-2/3 and ERK.

Membrane-Associated Endoglin Is Required forTGFb1-Induced Type I Collagen SynthesisTo explore this topic, the role of endoglin was examined by use

of a loss-of-function approach in hCFs. Compared with an

isotype control antibody, pretreatment of hCFs with an antibody

to endoglin attenuated TGFb1-induced type I collagen mRNA

and protein expression (Figure 5A). Similarly, silencing en-

Table. Characterization of Heart Failure in Wild-Type and

Eng1/2 Mice

TAC

Sham 4 wk 10 wk

Mass

Total body weight, g

WT 3563.9 2963.1* 2662.2†

Eng1/2 3461.2 3464.6‡ 3563.7‡

LV weight, mg/tibia

length, mm

WT 761 1161* 1562†

Eng1/2 560.7‡ 961*‡ 1161‡

Total lung weight, mg/

Tibia length, mm

WT 960.7 2461* 2361†

Eng1/2 860.3‡ 1961* 1561†‡

Hemodynamic data

Heart rate, bpm

WT 535645 553640 573693

Eng1/2 549632 522649 547632

LV end-systolic pressure,

mm Hg

WT 114612 130618 98642

Eng1/2 99615 157619*‡ 167614*‡

LV end-diastolic pressure,

mm Hg

WT 1164 3167* 2367*

Eng1/2 965 24612* 2564*

dP/dtmax

WT 900061263 518461092* 395961980*

Eng1/2 814161183 712762615‡ 70156712‡

dP/dtmin

WT 828861446 561461195* 395461955*

Eng1/2 801561563 728162204‡ 73366907‡

Echocardiographic data

LV end-diastolic

diameter, mm

WT 2.460.4 3.960.4* 4.561.8*

Eng1/2 1.860.1‡ 3.460.6* 2.660.03*‡

Fractional shortening, %

WT 7269 4264* 18615*

Eng1/2 5768‡ 4067 67610‡

Posterior wall

thickness, mm

WT 0.960.2 1.460.3* 1.360.3*

Eng1/2 0.560.04‡ 1.360.2* 2.360.1*‡

TAC indicates thoracic aortic constriction; WT, wild-type; and LV, left

ventricular.

*P,0.05, 4 weeks of TAC versus before TAC.

†P,0.05, 10 weeks of TAC versus before TAC.

‡P,0.05, Eng1/2 versus WT.



06/06/2013 18

doglin expression significantly reduced TGFb1-induced type I

collagen mRNA and protein expression, plasminogen activator

inhibitor-1 expression (Figure 5B and 5C), and connective tissue

growth factor in hCFs (Figure VA in the online-only Data

Supplement). These findings identified endoglin as a necessary

component for TGFb1 signaling in hCFs.

sEng Antagonizes TGFb1 Signaling inCardiac FibroblastsPrevious studies have suggested that sEng attenuates TGFb1

signaling in endothelium.20 We next explored whether sEng

modulates cardiac fibroblast function. We first treated hCFs

with recombinant human sEng and observed a dose-

dependent decrease in TGFb1-induced type I collagen ex-

pression (Figure 5D and 5E). To confirm the role of sEng as

a negative modulator of TGFb1 activity, we transfected

COS-1 cells with AdhsEng and confirmed a dose-dependent

increase in the level of sEng in conditioned culture media

(Figure VB in the online-only Data Supplement). Similar to

treatment with recombinant human sEng, treatment of hCFs

with conditioned media from AdhsEng-transfected COS-1

cells also inhibited TGFb1-induced type I collagen and

pSmad-2/3 expression (Figure 5F).

Next, we explored the effect of overexpressing full-length

endoglin using an adenovirus-mediated approach (adenovirus

expressing full-length endoglin) in hCFs and paradoxically

observed a reduction in TGFb1-induced type 1 collagen

expression (Figure VC in the online-only Data Supplement).

To study this further, we measured increased levels of sEng in

hCFs transfected with adenovirus expressing full-length en-

doglin (Figure VD in the online-only Data Supplement).

These findings implicated sEng as a negative feedback

mechanism that downregulates TGFb1 activity in hCFs.

Figure 3. Reduced endoglin expression does not affect cardiac hypertrophy and is associated with increased myocardial capillarity. A,Representative histological staining and (B) bar graph of left ventricular (LV) cardiomyocyte cross-sectional area in both wild-type (WT)and Eng1/2 mice after 4 weeks of thoracic aortic constriction (TAC). C and D, LV b-myosin heavy chain and SERCA mRNA and (E) cal-cineurin protein expression in WT and Eng1/2 mice after 4 weeks of TAC (n56 per group). F, Representative immunostaining of LVmyocardial capillaries (CD311) in Eng1/2 and WT mice after 4 weeks of TAC. G and H, Bar graphs quantifying LV myocardial capillarityafter 4 weeks of TAC in WT and Eng1/2 mice (n56 per group). *P,0.05 vs sham; †P,0.05 vs WT TAC.



06/06/2013 19

sEng Attenuates Cardiac Fibrosis inPressure-Overload–Induced Heart FailureTo explore whether sEng limits cardiac fibrosis in vivo, WT

mice received intravenous injections of AdhsEng, which in-

creased circulating levels of human sEng (Figure 6A). Com-

pared with controls, treatment with AdhsEng significantly re-

duced cardiac fibrosis (Figure 6B) and LV type I collagen

expression (Figure 6C) after 4 weeks of TAC. No significant

change in LV contractility (data not shown) was observed during

this subacute phase of LV pressure overload. These findings

Figure 4. Reduced cardiac fibrosis in Eng1/2 mice after pressure-overload–induced heart failure. A, Representative immunostaining ofleft ventricular (LV) type I collagen in wild-type (WT) and Eng1/2 mice after TAC. B, Quantification of LV fibrosis in Eng1/2 and WT afterthoracic aortic constriction (TAC; n56 per group). C, Levels of LV type I collagen mRNA and (D) protein expression in Eng1/2 and WTmice after TAC (n56 per group). E, Levels of transforming growth factor-b1 (TGFb1) mRNA and (F) pSmad-2/3 protein expression inboth WT and Eng1/2 mice after 4 weeks of TAC. pSmad-2/3 is expressed relative to total Smad-3 in the bar graph (n56 per group)and representative Western blot. *P,0.05 vs sham; †P,0.05 vs WT TAC at each corresponding time point.



06/06/2013 20

support that sEng blocks TGFb1 signaling and limits cardiac

fibrosis in pressure-overload–induced heart failure.

DiscussionOur central finding is that endoglin is required for TGFb1

signaling in hCFs and that selectively inhibiting TGFb1

signaling by reducing endoglin activity attenuates cardiac

fibrosis and improves survival in a mouse model of heart

failure. In contrast to the functional role of endoglin in

promoting TGFb1 signaling, sEng limits TGFb1 signaling,

type I collagen synthesis, and ultimately cardiac fibrosis

(Figure 7).

Our findings have several important clinical implications.

First, previous studies of nonselective TGFb1 blockade have

produced mixed results in heart failure. The ability to selec-

tively modulate TGFb1 activity by limiting the TGFb1

coreceptor endoglin offers a potentially novel approach to

managing heart failure. Second, studies involving endoglin

have focused on vascular remodeling with minimal data

exploring the role of endoglin in heart failure. However,

endoglin has been shown to mediate collagen synthesis

induced by angiotensin II and TGFb1 in rat cardiac fibro-

blasts.21,22 Our findings confirm that endoglin is an important

component of cardiac remodeling both in human LV tissue

and cardiac fibroblasts and in a mouse model. Third, recent

publications have focused on the role of sEng in preeclamp-

sia, vascular function, and renal fibrosis. We recently identi-

fied sEng as a biomarker of heart failure12 and now describe

a functional role for sEng as an autocrine antagonist of

TGFb1 signaling in heart failure. These findings provide new

insight into the regulation of TGFb1 activity in heart failure

and further identify sEng as a potentially novel therapeutic

approach to limit cardiac fibrosis.

We first identified that cardiac levels of endoglin are

increased in subjects with end-stage heart failure referred for

mechanical LVAD support and that endoglin levels are

reduced by mechanical LV unloading. Previous studies have

shown increased cardiac expression of profibrogenic genes,

including TGFb1, in patients with end-stage heart failure

referred for LVAD support.23,24 In both studies, higher levels

of TGFb1 were observed in subjects requiring LVAD support

until cardiac transplantation, suggesting that TGFb1 may

limit myocardial recovery by promoting tissue fibrosis. We

now identify endoglin as a potentially important target to

limit TGFb1 activity in heart failure.

Next, by confirming that endoglin is expressed by cardiac

fibroblasts, we studied the effect of reduced endoglin expres-

sion on cardiac fibrosis in a murine model of heart failure. We

first observed increased endoglin expression in the LV with

no change in abdominal aortic expression in WT mice

subjected to TAC. Importantly, 2 isoforms of membrane-

associated endoglin exist, namely the long and the less

abundant short isoforms.25 We confirmed that long endoglin

is the dominantly expressed isoform in the mouse LV (data

not shown). Consistent with our human observations, we

identified that cardiac pressure overload increases endoglin

expression in heart failure. Although TAC induced a similar

Figure 5. Endoglin and soluble endoglin (sEng) modulate transforming growth factor-b1 (TGFb1) activity. A, Type I collagen mRNA (bargraph) and protein expression (representative Western blot) in human cardiac fibroblasts (hCFs) after treatment with a neutralizing anti-endoglin antibody. B and C, Type I collagen mRNA (bar graph) and protein expression (representative Western blot) and plasminogen activa-tor inhibitor-1 (PAI-1) mRNA expression in hCFs after silencing endoglin expression. D and E, Type I collagen protein (quantification of West-ern blot shown in bar graph) and mRNA expression after treatment with recombinant human sEng (RhsEng). F, Type I collagen proteinsynthesis (bar graph and Western blot) and pSmad-2/3 expression (Western blot) after treatment with conditioned media from COS-1 cellstransfected with adenovirus overexpressing human sEng (AdhsEng). *P,0.05 vs control; †P,0.05 vs TGFb1-stimulated controls.



06/06/2013 21

pattern of endoglin expression in Eng1/2 and WT mice,

levels of endoglin expression were significantly lower in the

Eng1/2 mice throughout. Reduced endoglin expression im-

proved survival in this model of pressure-overload–induced

heart failure while preserving cardiomyocyte hypertrophy,

modestly increasing myocardial capillarity, and significantly

reducing cardiac fibrosis. Collectively, these changes were

more consistent with adaptive as opposed to maladaptive

cardiac remodeling26; however, the most dramatic observa-

tion in this model was the nearly complete attenuation of

cardiac fibrosis in the pressure-overloaded myocardium.

We then confirmed the dependence of TGFb1 signaling on

endoglin expression in hCFs using loss-of-function ap-

proaches. Furthermore, treatment with sEng either as a

recombinant protein or by adenoviral overexpression of sEng

or full-length endoglin in vitro mirrored the phenotype of the

reduced endoglin levels in the Eng1/2 mice, suggesting that

sEng also limits TGFb1 signaling and type I collagen

synthesis. Several prior studies have shown that overexpress-

ing endoglin in rat myoblasts and mouse fibrosarcoma cell

lines limits TGFb1-induced collagen expression.27–30 In the

context of our findings, these observations may highlight

important differences between cell types and species with

regards to the biological activity of endoglin. Furthermore,

these reports may be consistent with our gain-of-function

observations in hCFs and could suggest that transfecting

endoglin into stable cell lines also increases levels of soluble

endoglin in vitro, thereby attenuating TGFb1 activity. Fi-

nally, we studied a potential role of sEng in pressure-

overload–induced heart failure and observed reduced cardiac

fibrosis in mice treated with an adenovirus overexpressing

human sEng.

Previous studies have highlighted the critical role that

TGFb1 signaling plays in cardiac remodeling and heart

failure.5,31,32 Benefits of blocking TGFb1 activity such as

improving diastolic function in hypertensive heart dis-

Figure 6. Overexpressing human solubleendoglin (AdhsEng) limits cardiac fibrosisin vivo. A, Circulating serum levels ofhuman and mouse sEng in wild-type(WT) mice after treatment with AdhsEng(n53 mice per group). B, Representativehistology of left ventricular (LV) collagenexpression and (C) type I collagenmRNA expression in WT mice treatedwith AdhsEng and adenovirus with notransgene (AdNull).

Figure 7. Reduced endoglin activity lim-its cardiac fibrosis by disrupting trans-forming growth factor-b1 (TGFb1) signal-ing. Left, Endoglin expression is requiredfor TGFb1-induced type I collagen syn-thesis and cardiac fibrosis. Middle,Reduced endoglin expression in Eng1/2

mice attenuates TGFb1-induced pSmad-2/3, type I collagen expression, and car-diac fibrosis. Right, Potential mecha-nisms by which soluble endoglin (sEng)interrupts TGFb1-signaling: (1) disruptedligand binding, (2) dimerizing withendoglin, and (3) stimulating inhibitorypathways.



06/06/2013 22

ease7,33,34 have been reported, but in other models of heart

failure, this increased mortality after induction of ischemic

heart failure.8,35 Given the potential for adverse effects of

nonselective TGFb1 blockade, targeting specific aspects of

the TGFb1 signaling cascade may yield better outcomes. Our

findings support this concept in that Eng1/2 mice demon-

strated reduced phosphorylation of Smad-2/3 in association

with LV hypertrophy and limited cardiac fibrosis after TAC.

Because endoglin is highly expressed in cardiac fibroblasts,

modulating endoglin expression in these cells may selectively

influence fibrosis without affecting hypertrophy. This com-

bination results in sustained LV contractility and improved

survival despite chronic pressure-overload–induced heart

failure.

The role of sEng in heart failure also remains poorly under-

stood. First, the mechanism underlying increased sEng expres-

sion in heart failure is unknown. Proteolytic cleavage of sEng

from endoglin may occur both locally in cardiac tissue or

systemically because levels of matrix metalloproteinase-14 are

known to be elevated in heart failure.11,16,36,37 Second, the

mechanism by which elevated sEng levels interrupt TGFb1

signaling remains poorly characterized. Several possibilities

exist (Figure 7). First, endoglin may modulate signaling via

several TGFb family ligands,38,39 and sEng could serve as a

ligand trap for TGFb1 or other ligands, including bone morpho-

metric protein. However, recent studies indicate that bone

morphometric protein-9 and -10 may be the only ligands that

bind to sEng with high affinity.40 Second, sEng may promote

alternate signaling pathways that indirectly inhibit TGFb1 sig-

naling such as bone morphometric protein-7.41 Finally, release of

sEng by ectodomain shedding may render the receptor nonfunc-

tional, thereby further limiting TGFb1 activity in heart failure.

The present study has several limitations. First, we used a

mouse model with reduced total body expression of endoglin

as opposed to changes in cardiac-restricted expression. Sec-

ond, as a result of the technical challenges of sustaining

elevated levels of exogenous sEng with an adenoviral ap-

proach, we were unable to examine whether sEng improves

cardiac function in a longer-term model of heart failure.

ConclusionsTGFb1 is a powerful cytokine that governs the development

of cardiomyocyte hypertrophy and cardiac fibrosis in heart

failure. Therapies designed to nonselectively block TGFb1

activity in heart failure have failed to demonstrate clear

benefit. We now demonstrate the important functional role of

endoglin in heart failure by specifically showing that en-

doglin facilitates whereas sEng attenuates TGFb1-mediated

cardiac fibrosis and further that reduced endoglin expression

can limit cardiac fibrosis, preserve cardiac function, and

improve survival in pressure-overload–induced heart failure.

Our studies support that targeting endoglin provides a poten-

tially unique and novel therapeutic approach for individuals

with heart failure.

Source of FundingDr Kapur is supported by a grant from the National Institutes of

Health (K08HL094909–03).

DisclosuresDr Karumanchi and Dr Letarte are listed as co-inventors on patents

held by the Beth Israel Deaconess Medical Center for the use of

soluble endoglin in diagnosis and therapy of preeclampsia.

References1. Lloyd-Jones D, Adams R, Carnethon M, De Simone G, Ferguson TB,

Flegal K, Ford E, Furie K, Go A, Greenlund K, Haase N, Hailpern S, Ho

M, Howard V, Kissela B, Kittner S, Lackland D, Lisabeth L, Marelli A,

McDermott M, Meigs J, Mozaffarian D, Nichol G, O’Donnell C, Roger

V, Rosamond W, Sacco R, Sorlie P, Stafford R, Steinberger J, Thom T,

Wasserthiel-Smoller S, Wong N, Wylie-Rosett J, Hong Y; American

Heart Association Statistics Committee and Stroke Statistics Subcom-

mittee. Heart disease and stroke statistics—2009 update: a report from the

American Heart Association Statistics Committee and Stroke Statistics

Subcommittee. Circulation. 2009;119:e21–e181.

2. Cowie MR, Mosterd A, Wood DA, Deckers JW, Poole-Wilson PA,

Sutton GC, Grobbee DE. The epidemiology of heart failure. Eur Heart J.

1997;18:208–225.

3. Davies MK, Hobbs FDR, Davis RC, Kenkre JE, Roalfe AK, Wosornu D,

Lancashire RJ. Prevalence of left ventricular systolic dysfunction and

heart failure in the Echographic Heart of England Screening Study: a

population based study. Lancet. 2001;358:439–444.

4. Berk BC, Fujiwara K, Lehoux S. ECM remodeling in hypertensive heart

disease. J Clin Invest. 2007;117:568–575.

5. Leask A. TGFbeta, cardiac fibroblasts, and the fibrotic response. Car-

diovasc Res. 2007;74:207–212.

6. Massague J. TGF-beta signal transduction. Annu Rev Biochem. 1998;67:

753–791.

7. Kuwahara F, Kai H, Tokuda K, Kai M, Takeshita A, Egashira K,

Imaizumi T. Transforming growth factor-beta function blocking prevents

myocardial fibrosis and diastolic dysfunction in pressure-overloaded rats.

Circulation. 2002;106:130–135.

8. Frantz S, Hu K, Adamek A, Wolf J, Sallam A, Maier SK, Lonning S, Ling

H, Ertl G, Bauersachs J. Transforming growth factor beta inhibition

increases mortality and left ventricular dilatation after myocardial

infarction. Basic Res Cardiol. 2008;103:485–492.

9. McAllister KA, Grogg KM, Johnson DW, Gallione CJ, Baldwin MA,

Jackson CE, Helmbold EA, Markel DS, McKinnon WC, Murrell J.

Endoglin, a TGF-beta binding protein of endothelial cells, is the gene for

hereditary haemorrhagic telangiectasia type 1. Nat Genet. 1994;8:

345–351.

10. Li DY, Sorensen LK, Brooke BS, Urness LD, Davis EC, Taylor DG,

Boak BB, Wendel DP. Defective angiogenesis in mice lacking endoglin.

Science. 1999;284:1534–1537.

11. Bourdeau A, Dumont DJ, Letarte M. A murine model of hereditary

hemorrhagic telangiectasia. J Clin Invest. 1999;104:1343–1351.

12. Hawinkels LJ, Kuiper P, Wiercinska E, Verspaget HW, Liu Z, Pardali E,

Sier CF, ten Dijke P. Matrix metalloproteinase-14 (MT1-MMP)-mediated

endoglin shedding inhibits tumor angiogenesis. Cancer Res. 2010;70:

4141–4150.

13. Kapur NK, Heffernan KS, Yunis AA, Parpos P, Kiernan MS, Sahas-

rabudhe NA, Kimmelstiel CD, Kass DA, Karas RH, Mendelsohn ME.

Usefulness of soluble endoglin as a non-invasive measure of left ventric-

ular filling pressure in heart failure. Am J Cardiol. 2010;106:1770–1776.

14. Chintalgattu V, Ai D, Langley RR, Zhang J, Bankson JA, Shih TL, Reddy

AK, Coombes KR, Daher IN, Pati S, Patel SS, Pocius JS, Taffet GE, Buja

LM, Entman ML, Khakoo AY. Cardiomyocyte PDGFR-beta signaling is

an essential component of the mouse cardiac response to load-induced

stress. J Clin Invest. 2010;120:472–484.

15. Patten RD, Pourati I, Aronovitz MJ, Alsheikh-Ali A, Eder S, Force T,

Mendelsohn ME, Karas RH. 17 Beta-estradiol differentially affects left

ventricular and cardiomyocyte hypertrophy following myocardial

infarction and pressure overload. J Card Fail. 2008;14:245–253.

16. Kapur NK, Deming CB, Kapur S, Bian C, Champion HC, Donahue JK,

Kass DA, Rade JJ. Hemodynamic modulation of endocardial thrombore-

sistance. Circulation. 2007;115:67–75.

17. Neuss M, Regitz-Zagrosek V, Hildebrandt A, Fleck E. Isolation and

characterisation of human cardiac fibroblasts from explanted adult hearts.

Cell Tissue Res. 1996;286:145–153.

18. Bozkurt B, Kribbs SB, Clubb FJ Jr, Michael LH, Didenko VV, Hornsby

PJ, Seta Y, Oral H, Spinale FG, Mann DL. Pathophysiologically relevant

concentrations of tumor necrosis factor-a promote progressive left ven-



06/06/2013 23

tricular dysfunction and remodeling in rats. Circulation. 1998;97:

1382–1391.

19. Georgescu SP, Aronovitz MJ, Iovanna JL, Patten RD, Kyriakis JM,

Goruppi S. Decreased metalloprotease 9 induction, cardiac fibrosis and

higher autophagy after pressure overload in mice lacking the transcrip-

tional regulator p8. Am J Physiol Cell Physiol. 2011;301:C1046–C1056.

20. Walshe TE, Saint-Geniez M, Maharaj AS, Sekiyama E, Maldonado AE,

D’Amore PA. TG. F-beta is required for vascular barrier function, endo-

thelial survival and homeostasis of the adult microvasculature. PLoS One.

2009;4:e5149.

21. Chen K, Mehta JL, Li D, Joseph L, Joseph J. Transforming growth factor

beta receptor endoglin is expressed in cardiac fibroblasts and modulates

profibrogenic actions of angiotensin II. Circ Res. 2004;95:1167–1173.

22. Shyu KG, Wang BW, Chen WJ, Kuan P, Hung CR. Mechanism of the

inhibitory effect of atorvastatin on endoglin expression induced by trans-

forming growth factor-beta1 in cultured cardiac fibroblasts. Eur J Heart

Fail. 2010;12:219–226.

23. Caruso R, Caselli C, Boroni C, Campolo J, Milazzo F, Cabiati M, Russo

C, Parolini M, Giannessi D, Frigerio M, Parodi O. Relationship between

myocardial redox state and matrix metalloproteinase activity in patients

on left ventricular assist device support. Circ J. 2011;75:2387–2396.

24. Felkin LE, Lara-Pezzi E, George R, Yacoub MH, Birks EJ, Barton PJ.

Expression of extracellular matrix genes during myocardial recovery

from heart failure after left ventricular assist device support. J Heart Lung

Transplant. 2009;28:117–122.

25. Blanco FJ, Bernabeu C. Alternative splicing factor or splicing factor-2

plays a key role in intron retention of the endoglin gene during endothelial

senescence. Aging Cell. 2011;10:896–907.

26. Weeks KL, McMullen JR. The athlete’s heart vs. the failing heart: can

signaling explain the two distinct outcomes? Physiology (Bethesda).

2011;26:97–105.

27. Obreo J, Dıez-Marques L, Lamas S, Duwell A, Eleno N, Bernabeu C,

Pandiella A, Lopez-Novoa JM, Rodrıguez-Barbero A. Endoglin

expression regulates basal and TGF-beta1-induced extracellular matrix

synthesis in cultured L6E9 myoblasts. Cell Physiol Biochem. 2004;14:

301–310.

28. Rodrıguez-Barbero A, Obreo J, Alvarez-Munoz P, Pandiella A, Bernabeu

C, Lopez-Novoa JM. Endoglin modulation of TGF-beta1-induced

collagen synthesis is dependent on ERK1/2 MAPK activation. Cell

Physiol Biochem. 2006;18:135–142.

29. Velasco S, Alvarez-Munoz P, Pericacho M, Dijke PT, Bernabeu C,

Lopez-Novoa JM, Rodrıguez-Barbero A. L- and S-endoglin differentially

modulate TGFbeta1 signaling mediated by ALK1 and ALK5 in L6E9

myoblasts. J Cell Sci. 2008;121:913–919.

30. Diez-Marques L, Ortega-Velazquez R, Langa C, Rodriguez-Barbero A,

Lopez-Novoa JM, Lamas S, Bernabeu C. Expression of endoglin in

human mesangial cells: modulation of extracellular matrix synthesis.

Biochim Biophys Acta. 2002;1587:36–44.

31. Rosenkranz S. TGF-beta1 and angiotensin networking in cardiac

remodeling. Cardiovasc Res. 2004;63:423–432.

32. Haq S, Choukroun G, Lim H, Tymitz KM, del Monte F, Gwathmey J,

Grazette L, Michael A, Hajjar R, Force T, Molkentin JD. Differential

activation of signal transduction pathways in human hearts with hyper-

trophy versus advanced heart failure. Circulation. 2001;103:670–677.

33. Wang J, Xu N, Feng X, Hou N, Zhang J, Cheng X, Chen Y, Zhang Y,

Yang X. Targeted disruption of Smad4 in cardiomyocytes results in

cardiac hypertrophy and heart failure. Circ Res. 2005;97:821–828.

34. Divakaran V, Adrogue J, Ishiyama M, Entman ML, Haudek S, Sivasu-

bramanian N, Mann DL. Adaptive and maladaptive effects of SMAD3

signaling in the adult heart after hemodynamic pressure overloading. Circ

Heart Fail. 2009;2:633–642.

35. Ikeuchi M, Tsutsui H, Shiomi T, Matsusaka H, Matsushima S, Wen J,

Kubota T, Takeshita A. Inhibition of TGF-beta signaling exacerbates

early cardiac dysfunction but prevents late remodeling after infarction.

Cardiovasc Res. 2004;64:526–535.

36. Peterson JT, Li H, Dillon L, Bryant JW. Evolution of matrix metallo-

protease and tissue inhibitor expression during heart failure progression in

the infarcted rat. Cardiovasc Res. 2000;46:307–315.

37. Barsoum IB, Renaud SJ, Graham CH. Glyceryl trinitrate inhibits hypox-

ia-induced release of soluble fms-like tyrosine kinase-1 and endoglin

from placental tissues. Am J Pathol. 2011;178:2888–2896.

38. Barbara NP, Wrana JL, Letarte M. Endoglin is an accessory protein that

interacts with the signaling receptor complex of multiple members of the

transforming growth factor-beta superfamily. J Biol Chem. 1999;274:

584–594.

39. Venkatesha S, Toporsian M, Lam C, Hanai J, Mammoto T, Kim YM,

Bdolah Y, Lim KH, Yuan HT, Libermann TA, Stillman IE, Roberts D,

D’Amore PA, Epstein FH, Sellke FW, Romero R, Sukhatme VP, Letarte

M, Karumanchi SA. Soluble endoglin contributes to the pathogenesis of

preeclampsia. Nat Med. 2006;12:642–649.

40. Castonguay R, Werner ED, Matthews RG, Presman E, Mulivor AW,

Solban N, Sako D, Pearsall RS, Underwood KW, Seehra J, Kumar R,

Grinberg AV. Soluble endoglin specifically binds BMP9/BMP10 via its

orphan domain, inhibits blood vessel formation and suppresses tumor

growth. J Biol Chem. 2011;286:30034–30046.

41. Zeisberg M, Hanai J, Sugimoto H, Mammoto T, Charytan D, Strutz F,

Kalluri R. BMP-7 counteracts TGF-beta1-induced epithelial-to-

mesenchymal transition and reverses chronic renal injury. Nat Med.

2003;9:964–968.

CLINICAL PERSPECTIVEHeart failure is a major cause of global mortality. Transforming growth factor-b1 (TGFb1) is a cytokine that promotes

cardiac fibrosis in heart failure. Endoglin is a coreceptor that regulates TGFb1 signaling via downstream effector proteins

known as Smads (canonical pathway) or mitogen-activated protein kinases (noncanonical pathway). The extracellular

domain of endoglin can be cleaved into the circulation as soluble endoglin (sEng), which may serve as a natural antagonist

to TGFb1 activity. We now report that endoglin expression is increased in failing human left ventricular tissue and in a

murine model of thoracic aortic constriction–induced heart failure. Using the endoglin haploinsufficient mouse model, we

observed improved survival, limited cardiac fibrosis, and enhanced myocardial capillarity after thoracic aortic constriction.

To study the role of endoglin in vitro, loss-of-function studies demonstrated the dependence of TGFb1 activity on endoglin

expression in human cardiac fibroblasts. Paradoxically, adenovirus-mediated overexpression of full-length endoglin also blocked

TGFb1-induced collagen synthesis. Further study showed that levels of sEng were elevated in the conditioned media after

treatment with the adenovirus, thereby implicating sEng as a negative regulator of TGFb1 activity. This observation was

confirmed by adenovirus-mediated overexpression of human sEng or treatment with recombinant human sEng in vitro. To begin

exploring the utility of sEng as an antifibrotic approach in vivo, treatment with adenovirus-mediated overexpression of human

sEng attenuated cardiac fibrosis in wild-type mice after thoracic aortic constriction. Together, these data identify endoglin as an

important component of cardiac remodeling and a potentially novel target of therapy in heart failure.



06/06/2013 24

�

�

�

�

�

�

�

�

��

�

�

�

�

�

�

�

�

�


06/06/2013 25

� ��

��

�� !"�� "�#�$%&��'(�()�*��+��,-.��

�� /��β012#��3��#��!"��4.0�

�� (�5��6�� "�#��$%&��'(+()�*��+��,-.��

�� 7��"�� (�5��6�� "�#��

3�� *�� *�� "8 β�� 9� 5��6�� "�#�

$�,-.��#��!��:��5��0/.7��!"&��4.0�

�� (� 5��6�� "�#�� ;�� <��0�� !"� �� 4.0� ��

$�,-.�� "8 :�0�� "��

�:�:�� !"��4.0��$%&��'(�()�*��+��&��'(�()�*��!"�"�#��

�� 9��*�� =0��=0)��# ��

2>��#�� 3�� !�� :�� 5�� "8 :�� 0/.7� �� 0��

��# &��*��0/.7��0�&��*��#��!��

:��5��0�� !"��4.0�� 9�5��6�� "�#��;��

<�� *�� 0/.7��0��!"��4.0�

�� 9�5��6�� "�#�� 08��*�� 0/.7��

��0��!"��4.0�� 9�5��6�� "�#��2��!��:��5��=0

�./��*��=��# ��5��"8 :��:��

��!��:��5��=0�./��*��=��

�� !"��4.0��9�5��6�� "�#��$%&�'(�()�*��

�� )�� "8 β�0��

#"8 ��# ��3��#?�� 5��


06/06/2013 26

��0��*�� #��!��

:�� 5�� "8 β�0�� "�� *��

8�<;2� �� # � �� 5�� 0�� ;�� *��

�� # �� 5�� 0��$%&��'(�()�*��+��&�

�'(�()�*��"8 β�0��

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�


06/06/2013 27

��

�

�

�

�/��

�

�

�

�

�� /��

�

�

�

�

�

�

�

�


06/06/2013 28

�� 7�

�

�

�


06/06/2013 29

�� 9�

�

�

�

�

�� 9��

�

�

�

�

�

�

�

�

�

�

�

�


06/06/2013 30

�� )�

�

�

�

�

�

�

�

��3�

�

�

�

�

�

�

�


06/06/2013 31

��":�� "0<#��<��@�

�

�


06/06/2013 32

Leif Koschützke

Abnormal Development of Dendritic Spines in FMR1 Knock-Out Mice Quelle: J Neuroscience 2001 Betreuer: Prof. Oliver von Bohlen und Halbach (Anatomie) Was bedeutet mir dieses Thema persönlich? Ich interessiere mich sehr für das Lernen, insbesondere für die Mechanismen, die dahinterstecken. Die zu identifizieren hat sich als zum Teil äußerst kompliziert herausgestellt. In dieser Hinsicht können jedoch Erb-krankheiten, die zu mentaler Retardierung führen, hilfreich sein. An jenen kann untersucht werden, welche Auswirkungen Gendefekte auf Lernprozesse und die Entwicklung des Gehirns haben. Hierfür lassen sich insbesondere die Untersuchung der Neuroanatomie und, auf zellulärer Ebene, die Spinearchitektur heran-ziehen. Über defekte Gene konnten bislang einige Proteine ausfindig gemacht werden, die essentiell für die Ausbildung eines normal lernenden Gehirns sind. Worauf kommt es bei diesem Thema am meisten an? Da es hierbei um eine der ersten Veröffentlichungen handelt, ist es sicherlich am wichtigsten zu Beginn die Übersicht der Autoren zu dem Thema zu verstehen. Auch ist die Methodik äußerst interessant. Was fasziniert mich selbst am Thema am meisten? Die grundlegende Forschung interessiert mich hier besonders. Auch sehe ich dieses Paper als idealen Ein-stieg in mein Promotionsthema, welches sich mit einem Gendefekt und den Konsequenzen beschäftigt. Es ist sehr nützlich zu verstehen, dass sich die Forschung zunächst auf x-chromosomale mentale Retardie-rung konzentrierte und erst später, wie in meinem Fall, auf Gene, die auf anderen Chromosomen verortet sind, fokussierte. Was gefällt mir am Thema weniger? Insgesamt gibt es wenig an dem vorliegenden Paper zu kritisieren. Dazu gehört vielleicht ein etwas zu groß gewählter Scan-Abstand bei der Suche nach Spines und eine unorthodoxe statistische Methode. Auf Näheres soll im Vortrag und der Diskussion eingegangen werden.

06/06/2013 33

Abnormal Development of Dendritic Spines in FMR1Knock-Out Mice

Esther A. Nimchinsky, Adam M. Oberlander, and Karel Svoboda

Howard Hughes Medical Institute, Cold Spring Harbor Laboratory, Cold Spring Harbor, New York 11724

Fragile X syndrome is caused by a mutation in the FMR1 geneleading to absence of the fragile X mental retardation protein(FMRP). Reports that patients and adult FMR1 knock-out micehave abnormally long dendritic spines of increased densitysuggested that the disorder might involve abnormal spine de-velopment. Because spine length, density, and motility changedramatically in the first postnatal weeks, we analyzed theseproperties in mutant mice and littermate controls at 1, 2, and 4weeks of age. To label neurons, a viral vector carrying theenhanced green fluorescent protein gene was injected into thebarrel cortex. Layer V neurons were imaged on a two-photonlaser scanning microscope in fixed tissue sections. Analysis of.16,000 spines showed clear developmental patterns. Be-tween 1 and 4 weeks of age, spine density increased 2.5-fold,and mean spine length decreased by 17% in normal animals.

Early during cortical synaptogenesis, pyramidal cells in mutantmice had longer spines than controls. At 1 week, spine lengthwas 28% greater in mutants than in controls. At 2 weeks, thisdifference was 10%, and at 4 weeks only 3%. Similarly, spinedensity was 33% greater in mutants than in controls at 1 weekof age. At 2 or 4 weeks of age, differences were not detectable.The spine abnormality was not detected in neocortical organo-typic cultures. The transient nature of the spine abnormality inthe intact animal suggests that FMRP might play a role in thenormal process of dendritic spine growth in coordination withthe experience-dependent development of cortical circuits.

Key words: fragile X; FMRP; dendritic spine; critical period;somatosensory cortex; development; mental retardation;two-photon

Mental retardation is a component of many syndromes. Some arecaused by large-scale chromosomal abnormalities; others arecaused by physical, infectious, or biochemical insults sustainedearly in brain development. The fragile X syndrome is remark-able because it is a mental retardation syndrome caused by amutation in a single gene, FMR1. It is characterized by a constel-lation of signs in addition to the cognitive deficit, includingmacro-orchidism, certain facial features, and abnormalities inattention and short-term memory (Schapiro et al., 1995; de Vrieset al., 1998). The mutation giving rise to the syndrome, a CGGrepeat expansion in the 59 untranslated region of the FMR1 gene,interferes with transcription, and patients do not have measurableamounts of the gene product, fragile X mental retardation protein(FMRP) (Pieretti et al., 1991). This protein appears to act as anRNA binding protein (Feng et al., 1997; Brown et al., 1998) andis localized to neurons, and to dendrites in particular (Devys etal., 1993; Verheij et al., 1993; Feng et al., 1997). Interestingly,FMRP mRNA is found in dendrites, and the expression of FMRPis increased by activation of metabotropic glutamate receptors(Weiler et al., 1997), linking FMRP to synaptic function. Re-markably little is known about the neuropathologic features ofthis disorder. One finding reported both in patients (Rudelli etal., 1985; Hinton et al., 1991; Irwin et al., 2001) and in adult FMR1

knock-out mice (Comery et al., 1997) is an abnormality in den-dritic spines, which were described as being unusually long, sim-ilar to the spines observed in cortical circuits during development(Dailey and Smith, 1996; Fiala et al., 1998; Lendvai et al., 2000),and of increased density. These observations have led some tospeculate that the absence of FMRP causes a defect in spinematuration and pruning (Comery et al., 1997). This hypothesishas not, however, been tested directly in the intact brain. Anothercharacteristic of spines, necessarily neglected in postmortem mor-phologic studies, is their motility, which is believed, in part, torepresent postsynaptic participation in synapse formation (Daileyand Smith, 1996). Spine motility has been shown to be sensitive tosensory deprivation (Lendvai et al., 2000) and is greatest early inpostnatal life (Dunaevsky et al., 1999; Lendvai et al., 2000).Because an abnormality in spine motility early in postnatal lifecould affect synaptogenesis, it is of interest to determine whetherthe absence of FMRP affects normal spine motility. Here we showthat dendritic spines in the intact brains of FMR1 knock-out miceare abnormal early in postnatal life. These abnormalities in thesomatosensory cortex are most pronounced during the period ofgreatest synaptogenesis in that region (White et al., 1997) andsubside largely by the end of the first postnatal month. Spines incomparable neurons maintained in vitro showed no such abnor-malities and were normally motile. These observations supportthe notion that FMRP plays a role in synaptogenesis in the intactbrain.

MATERIALS AND METHODSMale FMR1 knock-out mice of the FVB strain and wild-type (wt)littermate controls were used for the present study. Genotypes weredetermined by PCR analysis of DNA extracted from tail samples takenbefore perfusion or culture preparation. The primers used were the sameas those outlined in the original article describing these animals (Dutch–

Received Feb. 5, 2001; revised April 12, 2001; accepted April 19, 2001.This work was supported by the FRAXA, Mather, and Pew Foundations and by

National Institutes of Health Grants AA05518 and NS38259. We thank BarryJ. Burbach and Karen Greenwood for expert technical assistance, Jodi Koblentz forPCR analysis, Drs. Edward Stern and Gerald Latter for statistical assistance, andDrs. Patrick R. Hof, Kristen M. Harris, Karen Zito, and Joshua Trachtenberg forhelpful discussion.

Correspondence should be addressed to Esther A. Nimchinsky, Howard HughesMedical Institute, Cold Spring Harbor Laboratory, One Bungtown Road, ColdSpring Harbor, NY 11724. E-mail: [email protected] © 2001 Society for Neuroscience 0270-6474/01/215139-08$15.00/0

The Journal of Neuroscience, July 15, 2001, 21(14):5139–5146

06/06/2013 34

Belgian Fragile X Consortium, 1994). All experimental protocols wereconducted according to National Institutes of Health guidelines foranimal research and were approved by the Institutional Animal Care andUse Committee at Cold Spring Harbor Laboratory.

Mice (6, 13, and 27 d of age; n 5 5 for each group) were administeredintracortical injections of enhanced green fluorescent protein (EGFP) ina replication-deficient Sindbis expression vector (Invitrogen, San Diego,CA). Briefly, mice were anesthetized with a mixture of ketamine andxylazine (22 and 15 mg/kg, i.m., respectively). A small burr hole wasdrilled in the skull overlying the posteromedial barrel field using stereo-tactic coordinates determined previously for each age on the basis ofcytochrome oxidase staining, and 50–100 nl of virion-rich supernatantwas delivered using a Picospritzer (General Valve Co, Fairfield, NJ) at adepth of 300–500 mm. Injection volume was titrated to yield sufficientnumbers of well separated neurons within the injected barrel and itsneighbors (Chen et al., 2000). The number of neurons labeled withcomparable injections decreased with postnatal age as described previ-ously (Chen et al., 2000). The skin was closed and the animal wasreturned to its cage. After 24 hr, the animal was anesthetized again andperfused transcardially with cold 4% paraformaldehyde. The brain wasremoved and post-fixed for an additional 6–12 hr. The brain was blockedand cut into 200-mm-thick sections in the coronal plane with a Vibratome(Technical Products International Inc., St Louis, MO). Labeled neuronswere identified and imaged using a custom-built two-photon laser scan-ning microscope (2PLSM) based on an Olympus Fluoview laser scanningmicroscope (Olympus America, Inc., Melville, NY). The light source wasa Ti–Sapphire laser (Tsunami; SpectraPhysics, Mountain View, CA)running at a wavelength of ;910 nm (repetition frequency, 80 MHz;pulse length, 150 fsec). The average power delivered to the backfocalplane of the objective (403; numerical aperature, 0.8) varied dependingon the imaging depth (range, 30–150 mW). Laser power was adjusted sothat additional power failed to reveal previously undetected protrusions.Fluorescence was detected in whole-field detection mode with a photo-multiplier tube (Hamamatsu Corp., Bridgewater, NJ). Dendritic seg-ments of different orders were sampled in a systematic random mannerthroughout the dendritic tree.

For in vitro studies, neocortical slice cultures were prepared from2-d-old mice (n 5 5 animals for each group) according to the method ofStoppini et al. (1991). Slices containing the posteromedial barrel fieldwere selected. The EGFP gene was introduced using biolistic genetransfer (Lo et al., 1994) (Bio-Rad, Hercules, CA) after 3 d in vitro(DIV), and layer V neurons were imaged 2 d later (7 d from the date ofbirth). Slice cultures were placed into a chamber and were continuallyperfused with oxygenated artificial CSF consisting of (in mM): 119 NaCl,2.5 KCl, 26.2 NaHCO3, 1.0 NaHPO4, 11 glucose, 2.5 CaCl2, and 1.3MgCl2; cultures were maintained at 32°C. Time-lapse imaging of system-atically sampled dendritic segments consisted of collecting image stacksevery 2 min for a total of 22 min, yielding 10 time intervals per segment.

Protrusions were analyzed using custom software written in Interac-tive Data Language (Research Support Instruments, Inc., Boulder, CO).Image stacks consisting of 10–50 sections spaced 1 mm apart weregenerated, and protrusion lengths were obtained by measuring the pro-jection of the protrusion from its tip to the point at which it meets thedendritic shaft. When the identity of a protrusion was unclear (forinstance, whether it was a single spine, or two, or an axonal profile), theindividual z images were reviewed. No effort was made to analyze spinesemerging below or above the dendrite. The reason for this is that theresolution of the imaging system, as with all optical systems, was lower inthe z direction than in the x and y directions. Because only the longestspines would reliably be detected along the optical axis, the analysis wasrestricted to spines that could be detected in the x- and y-axes. For spinedensity measurements, the sampling unit was the individual dendriticsegment imaged. As a measure of motility, absolute length differencesbetween subsequent time points were calculated; we refer to the meandifference over 10 intervals as “mean motility” (Chen et al., 2000;Lendvai et al., 2000). The “range of motility” refers to the total excursionof the spine over the imaging time period and is the difference betweenthe maximum and minimum lengths of the spine. “Proportion persistent”was defined as the proportion of the total number of protrusions thatwere present throughout the 22 min imaging session (i.e., with a lengththat never measured 0 mm). To test for interactions among postnatal agesand genotype and to control for variability at different sampling levels,spine lengths and densities in mutant and control groups were analyzedusing a nested ANOVA procedure (SAS Institute, Cary, NC). Thisprocedure was performed on second-order apical dendrites, which rep-

resented the largest single dataset. Length distributions were comparedusing the Kolmogorov–Smirnov two-sample test. For purposes of illus-tration, however, values for each genotype were pooled. Throughout thestudy, injections, perfusions, sectioning, imaging, and spine measure-ments were performed blind to genotype.

RESULTSOne day after intracortical injection of the viral vector containingthe EGFP gene, neurons were well labeled. The tissue sectionswere reminiscent of Golgi stains, in that a small subset of neuronswas fully labeled (Fig. 1A,B). The dendrites of mutant animalsshowed no gross abnormalities when compared with controls.

Figure 1. Two-photon laser scanning micrographs showing morphologyof and spine development in EGFP-expressing layer V pyramidal neuronsin the somatosensory cortex of wild-type (lef t) and mutant (right) micetransfected in vivo before fixation. Note the presence of long protrusionsat 1 week, particularly in the mutant mice (arrows). Scale bars: A, B, 50mm; C–H, 8 mm.

5140 J. Neurosci., July 15, 2001, 21(14):5139–5146 Nimchinsky et al. • Spine Development in FMR1 Knock-Out Mice

06/06/2013 35

Analysis was restricted to layer V pyramidal neurons in theposterior medial barrel field without evidence of truncated apicalor basal dendritic trees. Typically four to seven neurons wereanalyzed in each animal.

Analysis of the morphologic properties of dendritic spines inmutant and control animals during the first postnatal monthrevealed clear developmental trends. Spine length changed dur-ing this period. At 1 week, the overall mean spine length was 20%greater than at 4 weeks. Consistent with other accounts, spinedensity also changed, increasing 2.5-fold between 1 and 4 weeksin normal animals. These developmental changes were evident toa different degree in the mutant animals as well. The differencesbetween the genotypes depended on the postnatal age. At 1 week,the overall mean spine length was 28% greater in the mutantanimals than in the wild-type animals ( p , 0.0001; n 5 1806 and2034 spines for wild-type and mutant animals, respectively). Thisdifference dropped to 10% at 2 weeks (n 5 3452 and 3224 spinesfor wild-type and mutant animals, respectively) and only 3% at 4weeks (n 5 2202 and 3490 spines for wild-type and mutant

animals, respectively) (Fig. 2A–C). At 1 week, the distribution ofspine lengths differed significantly between the two genotypes(Kolmogorov–Smirnov two-sample test; p , 0.001), with mutantshaving fewer short spines (0.5–1.5 mm) and more medium-to-longspines (2.5–5 mm) (Figs. 1C,D, 2D). When categorized by den-dritic order, the spines at most dendritic orders were significantlylonger in the mutant animals compared with the wild-type ani-mals, with the exception of spines on tertiary basal dendrites, forwhich there was no significant difference between the groups (Fig.2G). At 2 weeks, the difference between the length distributionsin the two groups was smaller but still highly significant ( p ,0.001) (Fig. 2E). These differences were accounted for almostcompletely by differences localized to secondary apical and toprimary and tertiary basal dendrites (Fig. 2H). At 1 month, therewere no significant differences overall either in the length distri-bution or at any dendritic orders except for tertiary basal den-drites, for which spines in the mutant animals were significantlyshorter than in the wild-type animals (Fig. 2F,I). A nestedANOVA procedure performed on second-order apical dendrites,

Figure 2. Developmental spine length changes in FMR1 knock-out mice and littermate controls. A, D, G, 1 week; B, E, H, 2 weeks; C, F, I, 4 weeks.A–C, Mean spine length in mutant and control mice. D–F, Cumulative frequency distribution of spine lengths. G–I, Mean spine length at differentbranching levels in the dendritic tree. In D–F, black areas indicate wild type and white areas indicate mutant. In G–I, black bars indicate wild type andgray bars indicate mutant. ap1–3, Primary through tertiary apical dendrites; bas1–3, primary through tertiary basal dendrites. Error bars indicate SEM.*p , 0.05; **p , 0.001; ***p , 0.0001.

Nimchinsky et al. • Spine Development in FMR1 Knock-Out Mice J. Neurosci., July 15, 2001, 21(14):5139–5146 5141

06/06/2013 36

the largest dataset, confirmed these findings, showing a significantinteraction between age and genotype ( p , 0.01) despite consid-erable cell-to-cell variability within animals.

Spine density also varied developmentally with genotype. At 1week, mean spine density in the mutant animals was 33% higherthan in the wild-type animals (n 5 168 and 177 dendritic seg-ments for wt and mutant animals, respectively), and this differ-ence was evident at all sampled dendritic levels (Fig. 3A,D). At 2and 4 weeks there was no significant difference in spine density (2weeks, n 5 176 and 170 dendritic segments for wt and mutantanimals, respectively; 4 weeks, n 5 92 and 133 dendritic segmentsfor wt and mutant animals, respectively) (Fig. 3B,C). Interest-ingly, in secondary apical dendrites, spine density was elevated at1 week, normal at 2 weeks, and then significantly elevated again at4 weeks (Fig. 3D–F). Again, a nested ANOVA procedure per-formed on second-order apical dendrites confirmed a highly sig-nificant interaction between age and genotype ( p , 0.001). Thechanges in spine length and density over time in normal andmutant animals are summarized in Figure 4.

To study the morphologic properties of neurons in the absenceof sensory input, neocortical slice cultures were prepared from2-d-old mouse pups and analyzed at 1 week after the date of birth,corresponding to the time point showing the greatest differencesin the intact brain. Neurons labeled using biolistic gene transferwith EGFP were analyzed for spine length, density, and motility.Layer V neurons with typical pyramidal morphology were chosenfor analysis (Fig. 5A–D). The morphologies of these neurons weresomewhat different from those transfected in the intact brain, inthat apical dendrites reached layer I less frequently. Unlike theresults described for neurons labeled in vivo (Figs. 1–3), spine

lengths and densities in this preparation did not differ significantlybetween the two genotypes (Fig. 6A,C). In separate experiments,hippocampal organotypic cultures (prepared at postnatal day 7)were transfected either biolistically or by injection of the sameSindbis viral vector used in the experiments in the intact brainand imaged after 7 DIV. No significant difference in spine densitywas detected between neurons transfected with the two methods(virus, 0.72 6 0.07; biolistic, 0.73 6 0.035; p . 0.05), indicatingthat the transfection method was not responsible for the lack of amorphologic phenotype in vitro.

Analysis of spine motility in neocortical cultures also showedno significant differences between mutants and controls. Al-though protrusions were quite motile (Figs. 5E, 6E–G), the meanmotility per 2 min time interval, the rate of turnover, and thelengths over which spines ranged were not significantly differentin the mutant animals compared with the wild-type animals (Fig.6E–G).

DISCUSSIONHistologic studies of brains of patients with fragile X syndromeare scarce and can be summarized by noting that the brains aregrossly normal, with apparently normal neuronal densities. Theonly reported abnormalities have been a decrease in synapselength (Rudelli et al., 1985) and abnormally long and thin den-dritic spines (Rudelli et al., 1985; Hinton et al., 1991; Irwin et al.,2001). There is good reason to look at spines for neuropathology,because several mental retardation disorders are characterized byabnormalities in dendritic spines (Marin-Padilla, 1972; Purpura,1975). It is therefore of great interest that a Golgi study of theoccipital cortex of FMR1 knock-out mice at 16 weeks found the

Figure 3. Developmental spine density changes in FMR1 knock-out mice and littermate controls. A, D, 1 week; B, E, 2 weeks; C, F, 4 weeks. A–C, Meanspine density in mutant and control mice. D–F, Mean spine density at different branching levels in the dendritic tree. Means are shown only for primaryand secondary dendrites, because the number of tertiary dendrites sampled was too inconsistent to provide reliable estimates. Black bars, Wild type; graybars, mutant. ap1–2, Primary through tertiary apical dendrites; bas1–2, primary through tertiary basal dendrites. Error bars indicate SEM. *p , 0.05;**p , 0.001; ***p , 0.0001.


06/06/2013 37

spines there to be ;20% longer and up to 50% more denselydistributed than in wild-type controls (Comery et al., 1997).

Our present findings confirm increases in spine length anddensity in the knock-out mouse. However, we measured increasedspine density only transiently at 1 week, with no detectabledifferences at 2 and 4 weeks. Similarly, increases in spine lengthwere greatest at 1 week and were barely detectable by 1 month ofpostnatal life, much less pronounced than in the 16-week-oldanimals described previously (Comery et al., 1997). Several ex-planations for this discrepancy are possible. First, the previousstudy was in the occipital cortex rather than in primary somato-sensory cortex, and dendritic maturation may differ sufficiently inthese two cortical areas to account for this difference. However,two studies of dendritic spine maturation in the rat showed acomparable time course of dendritic spine development in thesetwo areas (Wise et al., 1979; Miller, 1981). Thus, it is unlikely thatregional differences would account for the discrepancies in spinedensity between the two FMR1 studies.

Second, because transfection efficiency decreases with postna-tal age (Chen et al., 2000), the possibility exists that the decreasein the magnitude of the phenotype is attributable to a selectivedecrease of labeling of only certain cell populations (for instance,neurons with abnormal spines). Although we have no reason toexpect such an effect, we cannot conclusively rule it out in theabsence of methods, such as electron microscopy, that do notdepend on the labeling of neuronal subpopulations. However, if

the labeling of abnormal cells decreased with age, we wouldexpect the cell-to-cell variability in mutant animals to be greaterat 1 week than at 4 weeks. In this study, there is a very slightincrease in variability between 1 and 4 weeks (results of ANOVA;data not shown).

Third, there is evidence that spine density normally decreasesafter 1 month postnatally (Wise et al., 1979; Miller, 1981). Thusthe results of the previous study might reflect an abnormalityduring a later phase of synapse pruning. Our methods effectivelypreclude our resolving this issue, because viral infection efficiencyis low after 6 weeks (Chen et al., 2000). This explanation wouldsupport a continuing role for FMRP in the production and main-tenance of appropriate synaptic connections beyond develop-ment. In fact, although spine density was normal on average at 4weeks, it was elevated selectively in second-order apical den-drites, after having been normal at this dendritic level at 2 weeks.This change could therefore represent the beginning of a trendcontinuing into adulthood and might have evolved into the dif-ference detected in the 16-week-old animals. Finally, method-ological differences must be considered, because it is difficult tocompare the neuronal populations labeled with the Golgi stainwith those labeled virally as in the present study, or to compareneurons analyzed with an eyepiece reticule with those imagedusing laser scanning microscopy.

The use of optical rather than electron microscopic methodspresents certain advantages, including the ability to sample very

Figure 4. A, C, Summary of changes inspine length (A) and density (C) inwild-type mice during the first postnatalmonth. n 5 5 animals in each group.Error bars indicate SEM. B, D, Sum-mary of changes in spine length ( B) anddensity (D) in mutant mice during thefirst postnatal month. The value foreach parameter is normalized to thewild-type value. Error bars are com-puted as the sum of the SEM for wild-type and mutant animals. n 5 5 animalsin each group.


06/06/2013 38

large numbers of spines in each animal in an efficient manner. Butan important limitation of optical studies is that the fluorescenceintensity, and by extension, the detectability, of small structuressuch as spines is proportional to their volume. Thus the smallestspines may go undetected using our method. Therefore, thegreater spine density observed in mutant animals at 1 weekcannot be clearly dissociated from differences in spine shape andlength observed at the same time. At 2 and 4 weeks, density isapparently normal, whereas spine length remains abnormal. How-ever, the large variance in the density measurements might stillobscure an abnormality in spine density in more mature tissue. It

is important to stress that our study makes no effort to describethe absolute lengths and densities of dendritic spines; such mea-surements require other methods. The purpose was to detectdifferences between genotypes and among postnatal ages. None-theless we note the close agreement between the spine densitymeasurements reported here at 4 weeks and those reported pre-viously in a study using the Golgi method in the adult animal(Comery et al., 1997). In both of these studies, it is understoodthat the density values obtained are underestimates of the actualspine density.

During the first postnatal week, dendritic spines are sparselydistributed and somewhat longer than in the adult (Dailey et al.,1996; Fiala et al., 1998; Lendvai et al., 2000; the present study).Within the next 3 weeks, they increase in density and becomesomewhat shorter, with fewer very long protrusions. The devel-opmental change in spine density has been described previously(Juraska and Fifkova, 1979; Juraska, 1982; Petit et al., 1988), butalthough it is commonly accepted that spine length decreases withpostnatal age, the present study is to our knowledge the firstsystematic description of this phenomenon. The normal generalpattern of development is followed in the FMR1 knock-out mice,but the increase in density is stunted and the decrease in length isexaggerated by the abnormal values early in postnatal life. Nev-ertheless, general synaptogenesis and remodeling of spines doestake place in these animals, and although the abnormality is notgrossly disruptive, its subtlety might point to a role for FMRP infine adjustments of synaptic connections.

The present study also demonstrates that pyramidal neurons invitro that are prepared from FMR1 knock-out mice do not havespines with abnormal length, density, or motility, despite theabnormalities in spine length and density in the intact brain. Theabsence of a comparable difference in 1-week-old neocorticalcultures invites several interpretations. First, the method of trans-fection in the intact brain differed from that used in vitro. How-ever, no difference in spine density was found between neuronstransfected in vitro with the Sindbis viral construct and thosetransfected biolistically. In addition, even if the labeling methoddiffered, this could not explain the differences observed betweengenotypes. Second, neurons in organotypic cultures 1 week afterthe date of birth might not be comparable with neurons in theintact brain 1 week after the date of birth. Similarly, it is possiblethat our inability to detect a difference between mutant andwild-type animals can be attributed to the presence or absence ofsome factor in the culture medium. Although neurons in organo-typic cultures mature roughly comparably to those in the intactbrain (Caesar et al., 1989; Caeser and Schuz, 1992; Muller et al.,1993), a more subtle effect remains a possibility. Notably, a recentstudy in cultured hippocampal neurons failed to detect differencesin spine length between mutant and wild-type animals, and in-stead found morphologic and electrophysiological evidence forpresynaptic defects (Braun and Segal, 2000). This finding suggeststhe unmasking of a defect in the initial stages of synaptogenesis,which must take place in dissociated cultured neurons after plat-ing but might not be detected in organotypic cultures, in whichthe vast majority of connections are already present when theculture is prepared.

The absence of an abnormal phenotype in organotypic culturesprepared from mutant mice, together with the strong develop-mental dependence of the abnormality in the intact brain, sug-gests that FMRP might not function simply to maintain dendriticspines within certain limits of density and length; its function maybe coordinated with normal patterned activity. In this context, it

Figure 5. Two-photon laser scanning micrographs of biolistically trans-fected layer V neurons in neocortical organotypic cultures at 1 week fromthe date of birth. A, C, Wild type. B, D, Mutant. A, B, Low-magnificationprojections showing pyramidal morphology. C, D, High-magnificationprojections of dendritic segments. Note the increased presence of longprotrusions in this preparation. Scale bar: A, B, 50 mm; C, D, 4 mm. E,Spine motility in neocortical cultures imaged using 2PLSM at 1 week fromthe date of birth. Micrographs taken at 2 min intervals show typicalchanges in spine morphology. One protrusion (arrow) changes shapeseveral times during this interval, branching, extending, and retracting.Another (arrowhead) undergoes changes in the morphology of its head.Others change relatively little during this interval. This example wastaken from a wild-type animal. Numbers indicate time in minutes. Scalebar, 2 mm.


06/06/2013 39

is significant that FMRP expression is increased by afferentstimulation (Todd and Mack, 2000) and that there is evidence forlocal translation of FMRP in response to metabotropic glutamatereceptor stimulation (Weiler and Greenough, 1993; Weiler et al.,1997). The coordination of expression with synaptic activity isconsistent with a phenotype that is most severe during synapto-genesis, as shown in the present study. This raises the possibilitythat the function served by FMRP early in life is related to theproper establishment of synaptic connections in response to syn-aptic activity. If this process is disturbed by the absence of thisprotein, this disruption could be reflected at that time by anabnormal spine density and morphology, and later by cognitivedefects resulting from improperly established connections.

To our knowledge, this is the first suggestion of a link betweena developmentally restricted structural or functional abnormalityand a mental retardation syndrome. This observation may haveimplications for genetic therapeutic approaches, because it wouldindicate that intervention must be administered as early as pos-sible in brain development to avoid long-lasting abnormalities inbrain function.

REFERENCESBraun K, Segal M (2000) FMRP involvement in formation of synapses

among cultured hippocampal neurons. Cereb Cortex 10:1045–1052.Brown V, Small K, Lakkis L, Feng Y, Gunter C, Wilkinson KD, Warren

ST (1998) Purified recombinant Fmrp exhibits selective RNA binding

as an intrinsic property of the fragile X mental retardation protein.J Biol Chem 15521–15527.

Caeser M, Schuz A (1992) Maturation of neurons in neocortical slicecultures: a light and electron microscopic study on in situ and in vitromaterial. J Hirnforsch 33:429–443.

Caesar M, Bonhoeffer T, Bolz J (1989) Cellular organization and devel-opment of slice cultures from rat visual cortex. Exp Brain Res77:234–244.

Chen B, Lendvai B, Nimchinsky EA, Burbach B, Fox K, Svoboda K(2000) Imaging high-resolution structure of GFP-expressing neuronsin neocortex in vivo. Learn Mem 7:433–441.

Comery TA, Harris JB, Willems PJ, Oostra BA, Irwin SA, Weiler IJ,Greenough WT (1997) Abnormal dendritic spines in fragile X knock-out mice: maturation and pruning deficits. Proc Natl Acad Sci USA94:5401–5404.

Dailey ME, Smith SJ (1996) The dynamics of dendritic structure indeveloping hippocampal slices. J Neurosci 16:2983–2994.

de Vries BBA, Halley DJJ, Oostra BA, Niermejjer MF (1998) Thefragile X syndrome. J Med Genet 35:579–589.

Devys D, Lutz Y, Rouyer N, Bellocq JP, Mandel JL (1993) The FMR-1protein is cytoplasmic, most abundant in neurons, and appears normalin carriers of a fragile X premutation. Nat Genet 4:335–340.

Dunaevsky A, Tashiro A, Majewska A, Mason C, Yuste R (1999) De-velopmental regulation of spine motility in the mammalian centralnervous system. Proc Natl Acad Sci USA 96:13438–13443.

Dutch–Belgian Fragile X Consortium (1994) Fmr1 knockout mice: amodel to study fragile X mental retardation. Cell 78:23–33.

Feng Y, Gutekunst CA, Eberhart DE, Yi H, Warren ST, Hersch SM(1997) Fragile X mental retardation protein: nucleocytoplasmic shut-tling and association with somatodendritic ribosomes. J Neurosci17:1539–1547.

Fiala JC, Feinberg M, Popov V, Harris KM (1998) Synaptogenesis viadendritic filopodia in developing hippocampal area CA1. J Neurosci18:8900–8911.

Figure 6. Spine length and density inneocortical cultures analyzed 1 week fromthe date of birth. A, Mean spine length;n 5 1754 and 1667 spines for wt andmutant animals, respectively. B, Spinelength at different levels in the dendritictree. C, Mean spine density; n 5 247 and255 dendritic segments for wt and mutant,respectively. D, Spine density at differentlevels in the dendritic tree. E–G, Spinemotility in neocortical cultures analyzed 1week from the date of birth. E, Meanmotility per 2 min time interval. F, Meanlength range over which spines vary over a22 min time interval. G, Proportion-persistent spines for wild-type and mu-tant animals. Black, Wild type; gray, mu-tant. ap1–3, Primary through tertiaryapical dendrites; bas1–3, primary throughtertiary basal dendrites. Error bars indi-cate SEM. *p , 0.05; **p , 0.001; ***p ,0.0001.


06/06/2013 40

Hinton VJ, Brown WT, Wisniewski K, Rudelli RD (1991) Analysis ofneocortex in three males with the fragile X syndrome. Am J Med Genet41:289–294.

Irwin SA, Patel B, Idupulapati M, Harris JB, Crisostomo RA, Larsen BP,Kooy F, Willems PJ, Cras P, Kozlowski PB, Swain RA, Weiler IJ,Greenough WT (2001) Abnormal dendritic spine characteristics inthe temporal and visual cortices of patients with fragile-X syndrome: aquantitative examination. Am J Med Genet 98:161–167.

Juraska JM (1982) The development of pyramidal neurons after eyeopening in the visual cortex of hooded rats: a quantitative study.J Comp Neurol 212:208–213.

Juraska JM, Fifkova E (1979) A Golgi study of the early postnataldevelopment of the visual cortex of the hooded rat. J Comp Neurol183:247–256.

Lendvai B, Stern EA, Chen B, Svoboda K (2000) Experience-dependentplasticity of dendritic spines in the developing rat barrel cortex in vivo.Nature 404:876–881.

Lo DC, McAllister AK, Katz LC (1994) Neuronal transfection in brainslices using particle-mediated gene transfer. Neuron 13:1263–1268.

Marin-Padilla M (1972) Structural abnormalities of the cerebral cortexin human chromosomal aberrations: a Golgi study. Brain Res44:625–629.

Miller M (1981) Maturation of rat visual cortex. I. A quantitative studyof Golgi-impregnated pyramidal neurons. J Neurocytol 10:859–878.

Muller D, Buchs PA, Stoppini L (1993) Time course of synaptic devel-opment in hippocampal organotypic cultures. Brain Res Dev Brain Res71:93–100.

Petit TL, LeBoutillier FC, Gregorio A, Libstug H (1988) The pattern ofdendritic development in the cerebral cortex of the rat. Brain Res DevBrain Res 41:209–219.

Pieretti M, Zhang F, Fu YH, Warren ST, Oostra BA, Caskey CT, NelsonDL (1991) Absence of expression of the FMR-1 gene in fragile Xsyndrome. Cell 66:817–822.

Purpura DP (1975) Dendritic differentiation in human cerebral cortex:normal and aberrant developmental patterns. Adv Neurol 12:91–116.

Rudelli RD, Brown WT, Wisniewski K, Jenkins EC, Larue-Kamionowska M, Connell F, Wisniewski HM (1985) Adult fragile Xsyndrome: clinico-neuropathologic findings. Acta Neuropathol67:289–295.

Schapiro MB, Murphy DGM, Hagerman RJ, Azari NP, Alexander GE,Miezejeski CM, Hinton VJ, Horwitz B, Haxby JV, Kumar A, White B,Grady CL (1995) Adult fragile X syndrome: neuropsychology, brainanatomy, and metabolism. Am J Med Genet 60:480–493.

Stoppini L, Buchs PA, Muller D (1991) A simple method for organotypiccultures of nervous tissue. J Neurosci Methods 37:173–182.

Todd PK, Mack KJ (2000) Sensory stimulation increases cortical expres-sion of the fragile X mental retardation protein in vivo. Mol Brain Res80:17–25.

Verheij C, Bakker CE, de Graaf E, Keulemans J, Willemsen R, VerkerkAJMH, Galjaard H, Reuser AJJ, Hoogeveen AT, Oostra BA (1993)Characterization and localization of the FMR-1 gene product associ-ated with fragile X syndrome. Nature 363:722–724.

Weiler IJ, Greenough WT (1993) Metabotropic glutamate receptorstrigger postsynaptic protein synthesis. Proc Natl Acad Sci USA90:7168–7171.

Weiler IJ, Irwin SA, Klintsova AY, Spencer CM, Brazelton AD, Mi-yashiro K, Comery TA, Patel B, Eberwine J, Greenough WT (1997)Fragile X mental retardation protein is translated near synapses inresponse to neurotransmitter activation. Proc Natl Acad Sci USA94:5395–5400.

White EL, Weinfeld D, Lev DL (1997) A survey of morphogenesisduring the early postnatal period in PMBSF barrels of mouse Sm1cortex with emphasis on barrel D4. Somatosens Mot Res 14:34–55.

Wise SP, Fleshman JW, Jones EG (1979) Maturation of pyramidal cellform in relation to developing afferent and efferent connections of ratsomatic sensory cortex. Neuroscience 4:1275–1297.


06/06/2013 41

Jonas A. Scheiber

Bone morphogenetic protein -4 and -5 in pancreatic cancer – Novel bidirectional players Quelle: Experimental Cell Research 317 (2011) 2136 - 2146 Betreuer: Prof. Julia Mayerle (Innere Medizin A) Was bedeutet mir das Thema persönlich? TGF-beta Signaling war Thema meiner Promotion und somit erstmaliger Berührungspunkt mit der Grundlagenforschung Worauf kommt es mir bei diesem Thema am meisten an? TGF-beta Signaling stellt ein unglaublich versatiles und komplexes System mit Implikationen für eine Vielzahl von Erkrankungen dar. Was fasziniert mich selbst am Thema am meisten? Das Ausmaß der Komplexität. Was gefällt mir am Thema weniger? Letztlich auch das Ausmaß der Komplexität.

06/06/2013 42

E X P E R I M E N T A L C E L L R E S E A R C H 3 1 7 ( 2 0 1 1 ) 2 1 3 6 – 2 1 4 6

ava i l ab l e a t www.sc i enced i r ec t . com

www.e l sev i e r . com/ loca te /yexc r

Research Article

Bone morphogenetic protein −4 and −5 in pancreaticcancer—Novel bidirectional players

Siru Virtanen 1, Emma-Leena Alarmo 1, Saana Sandström, Minna Ampuja, Anne Kallioniemi⁎

Laboratory of Cancer Genetics, Institute of Medical Technology, University of Tampere and Tampere University Hospital, Finland

A R T I C L E I N F O R M A T I O N

⁎ Corresponding author at: Institute of MedicalE-mail addresses: [email protected] (S

[email protected] (M. Ampuja), anne.kallionieAbbreviations: BMP, bone morphogenetic pr

1 These authors have equal contribution to th

0014-4827/$ – see front matter © 2011 Elseviedoi:10.1016/j.yexcr.2011.06.001

06/06/2013

A B S T R A C T

Article Chronology:

Received 14 January 2011Revised version received 8 June 2011Accepted 9 June 2011Available online 17 June 2011

Bone morphogenetic proteins (BMPs) are multifunctional signaling molecules that have gainedincreasing interest in cancer research. To obtain a systematic view on BMP signaling in pancreaticcancer we first determined the mRNA expression levels of seven BMP ligands (BMP2–BMP8) andsix BMP specific receptors in pancreatic cancer cell lines and normal pancreatic tissue. BMPreceptor expressionwas seen in all cancer and normal samples. Low expression levels of BMP5 andBMP8 were detected in cancer cells compared to the normal samples, whereas BMP4 expressionwas elevated in 25% of the cases. The impact of BMP4 and BMP5 signaling on cell phenotype wasthen evaluated in five pancreatic cancer cell lines. Both ligands suppressed the growth of three celllines (up to 79% decrease in BMP4-treated PANC-1 cells), mainly due to cell cycle changes. BMP4and BMP5 concurrently increased cell migration and invasion (maximally a 10.8-fold increase in

invaded BMP4-treated PANC-1 cells). The phenotypic changes were typically associated with theactivation of the canonical SMAD pathway, although such activation was not observed in thePANC-1 cells. Taken together, BMP4 and BMP5 simultaneously inhibit the growth and promotemigration and invasion of the same pancreatic cells and thus exhibit a biphasic role with bothdetrimental and beneficial functions in pancreatic cancer progression.

© 2011 Elsevier Inc. All rights reserved.

Keywords:

Bone morphogenetic proteinPancreatic cancerProliferationMigrationInvasion

Introduction

Bone morphogenetic proteins (BMPs) are signaling moleculesoriginally identified based on their ability to form bone [1,2]. Theyconstitute a subfamily of secreted signaling ligands belonging tothe transforming growth factor-β (TGF-β) superfamily [3]. Inhumans over 20 BMPs have been identified and based on theirsequence similarity they can be divided into distinct subgroups[3]. BMPs signal through two types of transmembrane serine–threonine kinase receptors, known as type I and type II receptors[4,5]. Upon ligand binding, activated BMP receptor complex

Technology, FIN-33014 Un. Virtanen), [email protected] (A. Kallioniemi).otein; SMAD, Sma- and Mais work.

r Inc. All rights reserved.

43

phosphorylates receptor-activated SMAD proteins (R-SMADs),SMAD1, -5, and -8 [6–8] which then form a heteromeric complexwith SMAD4 (Co-SMAD), translocate to the nucleus and regulatethe transcription of target genes [9]. This signaling pathway isreferred to as the canonical SMAD pathway. Conversely, the BMPscan activate other pathways such as mitogen-activated proteinkinases (MAPK), ERK1/2 and/or p38 [10–12].

At present it is widely acknowledged that the functions ofBMP ligands go beyond their role in osteogenesis. BMPs have beenshown to regulate tissue development and diverse biologicalprocesses such as cell proliferation, differentiation, and apoptosis

iversity of Tampere, Finland. Fax: +358 3 [email protected] (E.-L. Alarmo), [email protected] (S. Sandström),

d-related protein; MAPK, mitogen-activated protein kinase.

http://dx.doi.org/10.1016/j.yexcr.2011.06.001







2137E X P E R I M E N T A L C E L L R E S E A R C H 3 1 7 ( 2 0 1 1 ) 2 1 3 6 – 2 1 4 6

[13,14]. Because of these versatile and essential functions, BMPshave now become of great interest in cancer research. Thereare undoubtedly differences in the expression of BMP familymembers between normal tissue and cancer cells in many cancertypes [15–17]. Another strong evidence for the role of BMPsignaling in cancer came when it was found that mutations inthe BMP receptor BMPR1A cause a rare inherited colorectal cancerpredisposition syndrome known as familial juvenile polyposis[18,19]. Despite of these findings the actual biological role ofBMPs in tumorigenesis is still mostly uncovered. Functional data islimited and the results from different studies are partly contro-versial. Antitumorigenic functions have been described for ex-ample in prostate and colon cancer [20–23], whereas other studiesin prostate cancer and those in breast cancer report protumori-genic or metastasis promoting effects [24–26].

Recent observations suggest that alterations in BMP signalingmay play a role also in pancreatic cancer progression. It has beenshown that the mRNA levels of BMPR1A, BMPR2 and BMP2 are up-regulated in pancreatic cancer [27], however the molecular andcellular consequences of aberrant BMP signaling have beeninvestigated in very small extent [27–30]. Thus far comprehensiveBMP ligand and receptor gene expression analysis has not beenperformed, neither do we have sufficient functional data to createa profound understanding of this complicated signaling pathwayin pancreatic cancer development. Therefore, our interest was firstto determine the gene expression profiles of BMP ligands (BMP2–BMP8) and all BMP specific receptors (ACVR1, BMPR1A, BMPR1B,ACVR2A, ACVR2B, and BMPR2) in normal pancreatic tissues and in alarge panel of pancreatic cancer cell lines. Secondly, we wanted tofurther study the potential role of BMP signaling in pancreaticcancer pathogenesis by examining the capacity of BMP4 and BMP5to modulate growth, migration and invasion of pancreatic cancercell lines.

Materials and methods

Cell culture and reagents

Total of 16 established pancreatic cancer cell lines were used inthis study. Thirteen of them (AsPC-1, BxPC-3, Capan-1, Capan-2,CFPAC-1, HPAC, HPAF-II, Hs 700T, Hs 766T, MIA PaCa-2, PANC-1,SU.86.86, and SW 1990) were obtained from the American TypeCulture Collection (ATCC, Manassas, VA) and three (DanG, Hup-T3,and Hup-T4) from the German Collection of Microorganisms andCell Cultures (Braunschweig, Germany). Cells were cultured underrecommended conditions. Commercially available normal pancre-atic total RNA samples were obtained from Ambion (Austin, TX,USA), two different samples from BioChain Institute (Hayward, CA,USA) and one from Clontech (Mountain View, CA, USA). BMP4 andBMP5 were obtained from R&D Systems (Minneapolis, MN).

Quantitative reverse transcription—polymerase chainreaction (qRT-PCR)

RNA was isolated from cells using the RNeasy Mini Kit (Qiagen,Valencia, CA). RNA was reverse transcribed to cDNA by usingSuperScriptTM III First-Strand Synthesis System for RT-PCR(Invitrogen, Carlsbad, CA) with random hexamers (50 ng/μl) asdescribed [31]. The qRT-PCR was performed using the LightCycler

06/06/2013 44

equipment (Roche, Mannheim, Germany) as described [32].Transcript levels were determined using the LightCycler FastStartDNA Master Hybridization probe kit (Roche) and gene specificprimers and hybridization probes (TIB MOLBIOL, Berlin, Germany;Supplementary Table 1). The expression levels were normalizedagainst the reference gene TBP (TATA box binding protein), whichwas selected from a panel of ten different housekeeping genes(G6PD, PPIA, GAPD, TBP, B2M, GUSB, PBGD, HPRT, ACTB, and PGK1)since it had the most consistent expression levels across thesample set examined in this study.

Immunohistochemistry in primary tumors

BMP4 protein expression was studied in seven pancreatic ductaladenocarcinoma samples using immunohistochemistry. Immu-nostaining was performed using automated Ventana Benchmarkstaining system according to the manufacturer's instructions(Ventana Medical Systems, Tuscon, AZ). BMP4 protein wasdetected with mouse anti-BMP4 monoclonal antibody (1:10 dilu-tion, Chemicon Millipore, Temecula, CA) and ultraView UniversalDAB Detection Kit (Ventana Medical Systems) and the slides werecounterstained with hematoxylin–eosin. The entire tissue corewas evaluated and BMP4 staining pattern classified as negative,low or strong staining.

Cell proliferation assay

Cells were plated (day 0) overnight in 24-well plates and sub-sequently incubated in 10% serum medium containing eitherBMP4 or BMP5 ligand (250 ng/ml, unless otherwise indicated)or an equivalent volume of vehicle (4 mM HCl, 0.1% BSA). Due tovery high proliferation rate of MIA PaCa-2 cells, which led toovergrowth during the experimental period, they were culturedin 1% FBS and these culture conditions were also used in all thesubsequent functional experiments. Fresh medium with BMPligand or vehicle was added every third day. Cell number wascounted 3 (day 4) and 6 (day 7) days after the first addition of theligand or vehicle using Z1 Coulter Particle Counter (BeckmanCoulter, Fullerton, CA). All experiments were performed in 3–6replicates and were repeated at least twice.

Cell cycle and apoptosis analysis

Cell cycle and apoptosis analyses were performed after 48 h afterthe first addition of BMP4, BMP5 or vehicle in a 6-well plate. Bothfloating and adherent cells were collected and each assay wasperformed in triplicate and repeated at least twice. Cell cycleanalysis was performed by propidium iodide (PI) staining and theapoptosis assay was done using the Annexin V-FITC apoptosisdetection kit (Calbiochem, La Jolla, CA) as described [33]. Cell cycledistribution was analyzed with the ModFit LT Version 3.1 program(Verity software house, USA) and apoptosis with EXPO32 ADCVersion 1.2 analysis software (Beckman Coulter) according to themanufacturer's instructions.

Migration and invasion assay

Migration and invasion was studied using BD Bioscience cellculture inserts (8.0 μmpore size) and BD Biocoat Matrigel Invasionchambers (8.0 μm pore size, BD Biosciences, Bedford, MA) as

2138 E X P E R I M E N T A L C E L L R E S E A R C H 3 1 7 ( 2 0 1 1 ) 2 1 3 6 – 2 1 4 6

described [24]. Briefly, cells were treated with BMP4, BMP5 orvehicle for 72 h after which 0.1–2.0×105 cells, depending on thecell line, were transferred in serum free medium to the uppermigration or invasion chamber. In the lower chamber standard cellculture medium with 10% FBS was added, and cells were allowedto migrate or invade for 22 h at 37 °C. PANC-1 cells were allowedto migrate only for 3 h and MIA PaCa-2 cells for 6 h to ascertainthat the effects on cell migration and invasion were not caused bydifferences in cell proliferation. The fixed and stained migrationand invasion membranes were scanned with Aperio ScanScope®XT (software version 9; Aperio Technologies, USA) and analyzedwith ImageJ software [34] as described [24].

Western blotting analysis

To examine the effects of BMP stimulation on pancreatic cancercell on molecular level, cells were collected after 6 hour incubationwith BMP4, BMP5 or vehicle. Protein extraction, SDS-PAGE gelelectrophoresis and blotting were done as previously described[35]. The following primary antibodies were used: rabbit Phospho-SMAD 1/5/8, rabbit phospho-p44/42 MAP kinase, rabbit phospho-p38 MAP Kinase, rabbit p38 MAP Kinase, rabbit p44/42 MAPKinase, and rabbit SMAD5 (1:1000 dilution, Cell Signaling Tech-nology, Inc., Beverly, MA). The latter three antibodies were used asloading controls. Blots were incubated with primary antibodiesovernight at +4 °C and proteins visualized using BM Chemilumi-nescenceWestern Blotting Kit (Mouse/Rabbit) (Roche DiagnosticsGmbH) as described [24]. Western films were scanned and thesignal intensity for each band was determined using ImageJsoftware [34].

Statistical analysis

The Mann–Whitney test was used to statistically compare themedians of the expression levels in the cancer cell line and normalsample group, as well as the BMP-treated and control (vehicle-treated) groups. A P value<0.05 was considered statisticallysignificant.

Results

Expression of BMP ligands and receptors in normal pancreasand pancreatic cancer cells

We determined the relative expression values of BMP type I andtype II receptors ACVR1, BMPR1A, BMPR1B, ACVR2A, ACVR2B, andBMPR2 in 16 pancreatic cancer cell lines and in 4 normal pancreaticsamples using qRT-PCR. As seen in Fig. 1A, detectable expression ofall six BMP specific receptors was seen in all pancreatic cancer celllines as well as in normal pancreas samples. The range in receptorexpression levels waswide both in the cancer cell line group and inthe normal sample group, and did not permit classification of thesamples into distinct groups. Interestingly, the median expressionlevel of all the studied receptors, except for BMPR2, was lower incancerous cell lines compared to normal cells. This difference wasstatistically significant for ACVR1 (P=0.034) and reached close tosignificance also for ACVR2A (P=0.065) and BMPR1B (P=0.081).Notably the expression levels of BMPR1B were somewhat lower

06/06/2013 45

than those of the other receptors. The individual relativeexpression level for each receptor is shown in SupplementaryTable 2.

We next characterized the mRNA expression of the seven BMPligands (BMP2–BMP8) in the same set of cancer cell lines andnormal pancreatic samples (Fig. 1B, Supplementary Table 2). Ingeneral, the expression level of BMP2 was considerably highercompared to the other studied ligands, both in the cancer cellline group and the normal sample group. A slightly higher BMP2median expression level was revealed in the normal sample groupcompared to the cancer cell line group, but the difference was notstatistically significant. BMP3, BMP6 and BMP7 were expressed atvery low levels both in cancer cell lines and in normal pancreaticsamples, and their expression did not notably differ between thesetwo groups. Of note, each of these ligands were highly expressedin few cell lines, for example BMP3 in Hup-T4 cells, BMP6 in Hs700T and AsPC-1 cells, and BMP7 in HPAF-II and Hs 700T cells(Supplementary Table 2). However, for the majority of the cancercell lines the expression of these ligands was similar to thatdetected in the normal sample group.

BMP4, BMP5 and BMP8 were distinct from the other studiedligands due to their different expression patterns between thecancer cell line and the normal sample groups (Fig. 1B). Interest-ingly, BMP4 was expressed at very low level or below detection inall normal samples whereas a subset (4/16) of the cancer cell lineshad highly elevated expression. HPAF-II was one of the cancer celllineswith very high expression of BMP4 (Fig. 1C). Analysis of BMP4protein expression levels in primary pancreatic tumors demon-strated a similar pattern with five out of seven cases havingnegative or low expression and the remaining two cases showingstrong expression. On the contrary, there was a dramatic andstatistically significant (P=0.013) decrease in the relative expres-sion level of BMP5 in cancer cell lines compared to the normalpancreas samples. In fact, PANC-1 was the only cell line in whichBMP5 expression was detected (Fig. 1C). BMP8 has two verysimilar transcript variants, BMP8A and BMP8B, which were studiedhere together as BMP8. Similar to BMP5, the median expressionlevel of BMP8 in the cancer cell lines was significantly lower(P=0.021) than that of the normal pancreas samples.

BMP4 and BMP5 inhibit the proliferation of humanpancreatic cancer cell lines

As indicated above our results show that BMP4, BMP5 and BMP8mRNAs are expressed differently in pancreatic cancer cell linescompared to normal pancreatic samples. In order to determine thepossible functional significance of this aberrant expression, weevaluated the effects of BMP4 and BMP5 on the proliferation ofpancreatic cancer cell lines. BMP8 was not included in this studybecause we were unable to separate the two variants' (BMP8A andBMP8B) expressions profiles from each other.

The proliferation response to BMP4 and BMP5 treatment wasinvestigated in a panel of five pancreatic cancer cell lineswhich hadeither low or no endogenous BMP4 and BMP5 expression (AsPC-1,SU.86.86 and MIA PaCa-2) or high endogenous BMP4 (HPAF-II)or BMP5 (PANC-1) expression (Fig. 1C). BMP5 has not been usedearlier in pancreatic cancer studies, thus we tested different ligandconcentrations (50 ng/ml, 100 ng/ml and 250 ng/ml) to find themost effective dose and counted the cells after 3 and 6 days of firstaddition of BMP5 or vehicle. TheMIA PaCa-2 cells were selected for

ACVR1

50

100

150

200

0

rela

tive

expr

essi

on

100

200

300

400ACVR2A

BMPR1B

ACVR2B

50

100

150

200

0

BMPR1A

50

100

150

200

0

BMPR2

p=0.034

BMP2

0

2500

100

200

300

400

500

2540

BMP3

0

50

100

150180200

BMP4

BMP5

0

50

100

150BMP6

50

100

150360

0

380BMP7

0

50

100

150170

190

BMP8

0

50

100

150

50

100

150

0

BMP4BMP5

Pancreatic cancer cell lines Normal pancreas

p=0.013

p=0.021

A

50

100

150

200

0

0

50

100

150

200

0

rela

tive

expr

essi

onre

lativ

e ex

pres

sion

rela

tive

expr

essi

onre

lativ

e ex

pres

sion

B

0

10

100

50

AsPC-1 HPAF-II MIA

PaCa-2

PANC-1 SU.86.86

C

rela

tive

expr

essi

on

Fig. 1 – BMP ligand and BMP specific receptor mRNA expression. A, mRNA expression of six BMP specific receptors and B, seven BMPligands were determined in sixteen pancreatic cancer cell lines (■) and four normal pancreatic samples (Δ) by qRT-PCR. Medianvalue of relative expression is indicated by a horizontal line. C, BMP4 and BMP5 expression levels in five selected cell lines (AsPC-1,HPAF-II, MIA PaCa-2, PANC-1, and SU.86.86).


06/06/2013 46

cells

/ml x

105

0.5

1.0

1.5

2.0

day 0 day 7day 4

1.0

3.0

2.0

4.0ce

lls/m

l x 1

05

day 0 day7 day 4

BMP4

BMP5

Vehicle

250 ng/ml

100 ng/ml

50 ng/ml

Vehicle

250 ng/ml

100 ng/ml

A

PANC-1B

100

day 4 day 7

**

****

**50

0

%


this analysis since they do not show any endogenous expression.The most pronounced response was obtained with 250 ng/mldose (Fig. 2A). For BMP4, doses of 100 ng/ml and 250 ng/ml weretested (Fig. 2A), and 250 ng/ml was selected for subsequentanalyses.

BMP4 treatment led to growth inhibition in three out of fivestudied pancreatic cancer cell lines as compared to vehicle-treatedcells. Most distinct statistically significant decrease in cell numberwas seen in PANC-1 cells (on average 54% on day 4 and 79% onday 7) and in MIA PaCa-2 cells (37% on day 4 and 30% on day 7)(Fig. 2B). The anti-proliferative response was slightly smaller, butstatistically significant (P<0.01) in SU.86.86 cells (21% on day 4and 22% on day 7, data not shown). HPAF-II and AsPC-1 cells didnot respond to BMP4 treatment with apparent growth stimulationor inhibition. In a similar fashion, BMP5 addition to pancreaticcancer cells resulted in a significant growth suppressing effect inPANC-1 (on average a 22% decrease in cell number on day 4 and36% on day 7), HPAF-II (16% on day 4 and 36% on day 7), and MIAPaCa-2 (26% on day 4) compared to vehicle-treated cells (Fig. 2B).SU.86.86 and AsPC-1 cells did not show any consistent growthresponse to BMP5 treatment (data not shown).

Decreased cell growth is mainly mediated by cell cyclealterations

To more carefully investigate the growth regulating action ofBMP4 and BMP5 in these cancer cell lines, we analyzed their effectson cell cycle distribution after 48 h of stimulation. In the PANC-1cell line, an increased fraction of cells in G1-phase of the cell cyclewas detected after BMP4 (71% vs. 55%, P=0.002) and BMP5treatment (61% vs. 55%, P=0.002) compared to vehicle-treatedcells (Fig. 3A, Table 1). Similarly, the amount of MIA PaCa-2 cells inthe G1-phase was increased after BMP4 treatment (75% vs. 67%,P=0.002) and BMP5 treatment (72% vs. 67%, P=0.002) (Table 1).These changes were accompanied by a concomitant decrease ofcells in S-phase. No changes in cell cycle profile were observed inHPAF-II cells after BMP5 treatment (Table 1). BMP4-treatedSU.86.86 cells had a small but significant difference (50% vs.47%) in the fraction of cells in G1-phase compared to control-treated cells, correlating well with their less dramatic decrease incell growth (Table 1).

Fig. 2 – BMP4 and BMP5 decrease pancreatic cancer cell growth.A, MIA PaCa-2 cells were treated with different doses of BMP4(100 and 250 ng/ml) or BMP5 (50, 100 and 250 ng/ml) orvehicle in medium containing 1% FBS. Cell count wasdetermined 3 days (day 4) and 6 days (day 7) after the firstaddition of the ligand. Day 0 denotes the number of cellsseeded. The mean cell count of 3–6 replicates with error bars(SD) is shown. B, PANC-1, MIA PaCa-2 and HPAF-II cells weretreated with 250 ng/ml BMP4, 250 ng/ml BMP5 or vehicle.BMP-treated and vehicle-treated cells were counted at day 4and day 7. The fraction (%) of BMP-treated cells as compared tovehicle-treated cells is shown. Mean value of 6 replicates withSD from a representative experiment is shown. TheBMP-treated group was statistically compared to avehicle-treated group, *P<0.05, **P<0.005.

06/06/2013 47

To examine the possible effects of BMP4 and BMP5 onapoptosis we analyzed the cell viability by annexin V-FITC/PIstaining after 48 h of BMP treatment. BMP4, which had the mostdramatic effect on PANC-1 cell growth, clearly increased thenumber of apoptotic PANC-1 cells (Fig. 3B) whereas BMP5 had noeffect. In the other four cell lines, BMP4 or BMP5 ligand treatmentsdid not induce any consistent changes on programmed cell death(data not shown).

MIA PaCa-2

HPAF-II

day 4 day 7

day 4 day 7

BMP4Vehicle BMP5

***

*****

100

50

0

%

100

50

0

%

PI200 200 200

500

G1:55%

S:33%

G2:12%

Vehicle BMP4 BMP5

G1:71%

S:19%

G2:10%

600

G1:61%

S:27%

G2:12%

600

A

Eve

nts

10

6

2

Apo

ptot

ic c

ells

(%

) **

BMP4Vehicle

BMP5

B

Fig. 3 – BMP induces changes in cell cycle and apoptosis. A, Cell cycle histograms of BMP4 (250 ng/ml), BMP5 (250 ng/ml) orvehicle-treated PANC-1 cells at 48 h. A representative example withmean values of six replicates is shown. B, Apoptosis analysis wasdone 48 h after the first addition of BMP4 (250 ng/ml) or BMP5 (250 ng/ml) to PANC-1 cells. A representative example of sixreplicates with error bars (SD) is shown. **P<0.005.


BMP4 and BMP5 promote migration and invasion ofpancreatic cancer cells

To assess whether BMP4 and BMP5 treatments induce transcrip-tional changes that influence the ability of the cells to migrate and

Table 1 – Distribution of PANC-1, MIA PaCa-2, HPAF-II, andSU.86.86 cells in cell cycle. Cells were stimulated with250 ng/ml BMP4, 250 ng/ml BMP5 or vehicle for 48 h, PIstained and analyzed with flow cytometry. Percentage(mean and SD of six replicates) of cells in each cell cyclephase is shown from a representative experiment. Asterisks,*P<0.05, **P<0.005.

Cell line/treatment G1 S G2

PANC-1BMP4 71±2.5** 19±3.2** 10±0.8**BMP5 61±1.4** 27±1.7** 12±0.5Vehicle 55±1.6 33±1.3 12±0.6

MIA PaCa-2BMP4 75±0.7** 17±0.7** 8±0.4**BMP5 72±0.8** 19±0.7** 9±0.4*Vehicle 67±1.0 22±0.8 11±0.6

HPAF-IIBMP4 46±1.8 17±0.9 37±2.1BMP5 46±1.2 19±1.8 35±1.9Vehicle 47±2.9 17±0.9 36±3.0

SU.86.86BMP4 50±0.5** 19±0.4** 31±0.6Vehicle 47±0.4 22±0.8 31±0.6

06/06/2013 48

invade, AsPC-1, HPAF-II, MIA PaCa-2, PANC-1, and SU.86.86 cellswere first treated with BMP4, BMP5 or vehicle for 72 h and thensubjected to migration and invasion assays. BMP4 treatmentclearly increased the migration of PANC-1, MIA PaCa-2, and HPAF-II cells (6.0-, 2.4-, and 2.9-fold as compared to vehicle-treated cells,Figs. 4A–B). Similarly, the migration of BMP5-treated PANC-1 andHPAF-II cells increased significantly (1.8- and 3.0-fold, respective-ly, Figs. 4A–B). For MIA PaCa-2 cells, there was a clear tendency forincreasedmigration after BMP5 treatment but this change was notstatistically significant. Furthermore, the enhanced migration wasreflected in increased invasion through Matrigel in all these cellsafter BMP4 or BMP5 treatment with the exception of BMP5-treated PANC-1 cells. The most dramatic increase (10.8-fold) wasdetected in the amount of invaded PANC-1 cells after BMP4treatment (Fig. 4A).

In contrast, BMP4- or BMP5-treated AsPC-1 and SU.86.86 cellsdid not exhibit significant changes in their migration or invasioncapability as compared with vehicle-treated cells (data notshown). In conclusion, BMP4 and BMP5 treatment exhibited asimultaneous anti-proliferative response and induction of cellmigration and invasion phenotype in a subset of pancreatic cancercell lines.

BMP treatment can activate SMAD pathway in pancreaticcancer

Finally, we wanted to evaluate whether the phenotypic changesobserved after BMP4 and BMP5 stimulation were caused byactivation of the canonical SMAD pathway or through the MAPkinases p38 or ERK1/2. Phosphorylation levels of SMAD1/5/8, p38and ERK1/2 were determined after a 6 h treatment with BMP4,

Fig. 5 – BMP4 and BMP5 activate the SMAD pathway inpancreatic cancer cells. Activation of SMAD1/5/8 wasdetermined in five pancreatic cell lines after six hour treatmentwith BMP4 (250 ng/ml), BMP5 (250 ng/ml) or vehicle using

Fig. 4 – BMP4 and BMP5 increase the migration and invasion of pancreatic cancer cell lines. Cells were treated with BMP4 (250 ng/ml), BMP5 (250 ng/ml) or vehicle for 72 h and then subjected tomigration and invasion assay. The total area of migrated or invadedcells (in pixels) was determined from four images per membrane. A, The mean total area of migrated (upper chart) and invaded(lower chart) cells from six replicate membranes is shown with SD. The BMP-treated group was statistically compared to thevehicle-treated group, *P<0.05, **P<0.005. B, Representative images of BMP-treated and vehicle-treated migrated PANC-1, MIAPaca-2 and HPAF-II cells (20× objective) are shown.


BMP5 or vehicle. Clear induction of SMAD1/5/8 phosphorylationwas observed in AsPC-1 and MIA PaCa-2 cells after BMP4 andBMP5 treatment (Fig. 5). BMP5 treatment also increasedSMAD1/5/8 phosphorylation in HPAF-II cells whereas BMP4 didnot (Fig. 5). In contrast, BMP4 treatment increased the phosphor-ylation level of SMAD1/5/8 in SU.86.86 cells while BMP5 had noinfluence (Fig. 5). No changes were detected in the activationof the canonical SMAD pathway after either BMP4 or BMP5treatment in PANC-1 cells (Fig. 5). Interestingly, with theexception of AsPC-1, some baseline SMAD1/5/8 phosphorylationwas detected in all of the studied cell lines. ERK1/2MAP kinasewasnot affected by BMP4 or BMP5 treatment in any of the cell lines(data not shown). However, it has to be noted that all cellsexhibited very high basal levels of activated ERK1/2 which made itdifficult to assess the possible additional phosphorylation induc-tion of this protein. No differences in the phosphorylation levels ofp38 were detected with the exception of AsPC-1 showing a smallincrease in the p38 phosphorylation levels after BMP4 treatment.

Western blot. The band intensities were quantitated by ImageJsoftware and the phosphorylated SMAD1/5/8 values werenormalized against the unphosphorylated SMAD5 expressionlevels. Fold change (FC) values under each lane indicate thenormalized P-SMAD1/5/8 intensities in BMP4 or BMP5-treatedsamples as compared to corresponding vehicle control.

Discussion

In recent years, BMPs and BMP signaling have drawn muchattention in cancer research. There is a growing body of studies

06/06/2013 49

image of Fig.�4


showing the association between BMP pathway and differentcancers, but the functional data is still very limited, especially inpancreatic cancer. In this study we took a unique approach by firstperforming a comprehensive expression survey of all six BMPspecific receptors as well as seven BMP ligands (BMP2–BMP8) in alarge panel of sixteen pancreatic cancer cell lines and four normalpancreatic samples to identify appropriate model systems forsubsequent evaluation of the functional role of BMPs. This is thefirst time BMP receptors and ligands have been studied to thisextent in pancreatic cancer.

Our results show that all BMP specific receptors are expressedin pancreatic cancer cell lines thus demonstrating that this part ofthe BMP signaling cascade is indeed functional. The medianexpression level of all BMP receptors, except for BMPR2, was lowerin cancerous pancreatic cell lines compared to normal cells.However, only for ACVR1 this difference was statistically signifi-cant. This common trend seems, however, interesting and mightsuggest down-regulation of BMP receptors in pancreatic cancer. Infact, loss of BMP receptor expression has been previously shown tocorrelate with high tumor grade in prostate cancer [36]. Theexpression levels of BMP receptors BMPR1A, BMPR1B and BMPR2have earlier been studied by Northern blot analysis [27] and RT-PCR [28] in pancreatic cancer. There is a good concordance in thereceptor expression patterns when one compares the samepancreatic cancer cell lines studied here and in these two previousreports. However, both studies [27,28] reported elevated BMPR1Aand BMPR2 mRNA levels in primary pancreatic tumors. One mustkeep in mind that primary tumors contain multiple cell typeswhereas cancer cell lines are homogeneous in their cellular origin.More importantly, cancer cell lines are typically derived frommoreadvanced disease and thus it is conceivable that the receptorexpression levels change during tumor progression.

Here we also show that the relative expression levels of mostBMP ligands did not differ remarkably between pancreatic cancercell lines and normal samples. BMP3 and BMP6, the expressionof which has not been previously studied in pancreatic cancer, aswell as BMP7 were expressed at very low levels both in the greatmajority of cancer cell lines and in all normal pancreatic samples.BMP2 expression was in an earlier study shown to be highlyincreased in pancreatic cancer compared to normal pancreas [27].In a like manner, we observed generally high level BMP2 ex-pression but we did not detect a difference between normal andcancer samples. These results are likely to be due to different studymaterials, especially different normal samples, since data fromfour pancreatic cancer cell lines included in both studies are highlyconcordant. In addition, the assay methods (Northern analysisvs. qRT-PCR) used in these studies have quite different detectionsensitivities which might influence the results. The expressionlevels of BMP5 and BMP8 have not been studied earlier inpancreatic cancer and only to a very small extent in other cancers.Here we demonstrated that BMP5 and BMP8 are present in cancercell lines in significantly lower levels compared to normal samples.They thus appear to be silenced in pancreatic cancer and mighthave an antitumorigenic role. Earlier studies have shown BMP8overexpression in lung cancer cell lines and in breast cancer[31,37]. BMP5 transcripts have been detected at very low levelsin breast cancer cell lines and in primary tumors [31]. Down-regulation of BMP5 expression has also been seen in the adreno-cortical carcinoma cells whereas overexpression was detected insquamous cell carcinoma [38,39]. Finally, the relative expression

06/06/2013 50

level of BMP4was low both in normal pancreas samples and in themajority of the pancreatic cancer cell lines. However, 25% of thecell lines displayed highly elevated BMP4 levels. This up-regulationwas also reported by Gordon et al. [28] in their analysis of publiclyavailable microarray data. Furthermore, previous studies in breastcancer and malignant melanoma have reported high BMP4expression [31,40], and in colorectal cancer the expression ofBMP4 has been demonstrated to increase during tumor progres-sion [41].

It is important to note that comparisons between pancreaticcancer cell lines and RNA isolated from normal pancreas are notstraightforward. The cell lines originate from exocrine pancreaswhereas normal samples contain both exocrine and endocrinetissue. However, as the exocrine part represents over 90% ofthe mass of the pancreas, normal tissue RNA is indeed a validreference. To further address this issue, we compared our resultswith microarray based expression data derived from both nor-mal pancreas and primary tumors (GeneSapiens, http://www.genesapiens.org/; Oncomine, https://www.oncomine.org/). Asexpected, this comparison revealed highly similar expressionpatterns in our four normal samples and the normal pancreassamples in the database. More importantly, the expressionpatterns in the cell lines were highly similar to those seen inprimary pancreatic tumors. In particular, the expression levels ofBMP5 were very low or missing in primary pancreatic tumorsand BMP4 showed generally low expression with a small subset ofprimary tumors having elevated expression compared to normalpancreas. These patterns correspond extremelywell to our cell lineexpression data. Hence this study provides useful information onBMP and BMP receptor expression and a resource of cell lines forfurther studies of BMP function in pancreatic cancer.

BMPs have been shown to influence cancer cell phenotype inmany tumor types but in pancreatic cancer their biological role hasnot been adequately investigated. No clear association betweenBMP ligand or receptor expression and pancreatic cancer cell linecharacteristics, such as cell adhesion, migration, invasion, angio-genesis or tumorigenicity [42], was detected. Neither did the BMPstatus associate with the origin of the cell line (primary tumorvs metastasis) or the mutational patterns of KRAS, p53, p16 andSMAD4, genes most commonly altered in pancreatic cancer[42,43]. However, BMP4 expression might be related to a dif-ferentiation status of tumors since the majority of the cell linesderived from well-differentiated tumors [42] did express BMP4.To study the effects of BMPs in different cellular contexts, weevaluated the functional role of BMP4 and BMP5 in a panel offive pancreatic cancer cell lines with no, low or high endogenousexpression. In general, both BMP4 and BMP5 reduced the growthof pancreatic cancer cells. A marked growth inhibition wasobserved in three out of five pancreatic cancer cell lines (MIAPaCa-2, SU.86.86 and PANC-1) after BMP4 treatment. This result isin agreement with earlier observations in other cancers [35,44,45].BMP5 created similar but less dramatic growth reduction in threecell lines (HPAF-II, MIA PaCa-2 and PANC-1). This effect wasindependent of the endogenous expression level, since it wasobserved in both cell lines with high expression (PANC-1) andno expression (MIA PaCa-2 and HPAF-II). Reduced proliferationafter BMP5 treatment has been previously reported in myelomacells [46].

Consistent with the decreased proliferation, cell cycle analysisshowed an increased fraction of cells in G1-phase and concomitant

http://www.genesapiens.org/

http://www.genesapiens.org/

https://www.oncomine.org/


decrease in S-phase after BMP4 and BMP5 stimulation in MIAPaCa-2 and PANC-1 cells. In BMP4-treated SU.86.86 cells, a smallerdecrease in the S-phase fraction was seen, being consistent witha less dramatic reduction in cell number. In addition, BMP4possessed pro-apoptotic activity in PANC-1 cells, thus furthercontributing to the observed reduction in cell growth. We wereunable to show cytostasis or pro-apoptotic activity in HPAF-II cellsafter BMP5 treatment even though we did see a difference in theirgrowth repeatedly. Thus we do not have conclusive evidenceon the mechanisms of the reduced cell growth in these cells.However, it can be noted that the HPAF-II cells had a high numberof apoptotic cells also in control-treated cells (data not shown)which might have masked the pro-apoptotic effect of the BMP5treatment. Taken together, our data imply that the anti-proliferativeeffects of BMP4 and BMP5 in pancreatic cancer cells are generallydue to cytostasis.

The contribution of BMP5 to pancreatic cancer cell invasivenesshas not been investigated before, and in case of BMP4 it hasbeen studied only in two cell lines (PANC-1 and AsPc-1) [28]. Ourresults extend these findings by demonstrating that also BMP5 isable to increase the migration and invasiveness of pancreaticcancer cells. Importantly we now show that HPAF-II, MIA PaCa-2and PANC-1 pancreatic cancer cell lines with a clear growthinhibitory effect after BMP treatment demonstrated a parallelincrease in cell migration and invasion. In contrast, AsPC-1 andSU.86.86 cells that were less sensitive to BMP4 or BMP5 inducedgrowth arrest did not display distinct changes in their migration/invasion ability. These data indicate that some cells are indeedmore sensitive to BMP stimulation, and more significantly thatBMP4 and BMP5 have functions that are both detrimental andbeneficial to the pancreatic cancer progression.

There are several factors that might contribute to the cell linespecific phenotypes observed in our study. Interestingly theendogenous BMP4 or BMP5 expression level did not seem toaffect the results. Another interesting note is that BMP5 decreasedthe proliferation of HPAF-II cells whereas these cells wereresistant to BMP4 proliferation controlling signals, and similarlySU.86.86 responded to BMP4 but not to BMP5 treatment. Thusthere are distinct differences in the actions of BMPs and moreimportantly different cell lines show diverse responses to thesame BMP ligand. On a molecular level, all cell lines but PANC-1showed SMAD1/5/8 phosporylation after BMP4 and/or BMP5treatment but no MAP kinase activation, regardless of whetherthey showed proliferation or migration response or not. Thesefindings argue that the BMP signal is typically transmitted throughthe canonical route. However, it is important to note that therewas no direct association between the activation of the canonicalBMP signaling pathway and a specific phenotypic response. Thisfact is exemplified by the AsPC-1 cell that showed markedphosphorylation of SMAD1/5/8 after both BMP4 and BMP5treatment but no changes in cell growth or migration/invasion.All of the studied cell lines have wild type SMAD4, although in thecase of AsPC-1 reports are divergent [42,43]. Thus this issue doesnot explain the differences in phenotypes. Taken together, it isconceivable that BMP4 and BMP5 signaling differs in pancreaticcancer, perhaps through BMP receptor usage or the use of com-pletely separate signaling pathways and therefore it is importantnot to see them as functional equivalents. This feature of BMPsignaling underlies its exceptionally pleiotropic and multifunc-tional nature.

06/06/2013 51

One of the main findings of this study was the discovery ofthe dualistic functions of BMP4 and BMP5. They simultaneouslyacted as antiproliferative factors inhibiting cell growth mainlyvia induction of cell cycle arrest and as prometastatic factorsthrough stimulation of cell migration and invasion within thesame pancreatic cancer cells. Similar dualistic effects were recentlyreported after BMP4 treatment in breast cancer [35]. Furthermore,the well-known family member TGF-β has an established dualisticrole in cancer development where it acts as a tumor suppressorduring the first steps of tumorigenesis, but later stimulatesmetastasis formation [47]. It is in fact possible that cells havemetastatic capability and at the same time are characterized byreduced proliferative activity. Anjomshoaa et al. recently reportedthat both metastasizing primaries and liver metastases havereduced proliferative activity relative to non-metastatic colorectalcancer cells [48]. These data agree well with our results wherepancreatic cancer cells respond to BMP signals by growing slowerbut having increased ability to migrate and invade. Furthermore,it is important to note that not all pancreatic cell lines respondedto BMP4 and BMP5 treatment. In similar fashion, BMP7 actionsin prostate cancer were dependent on the cell phenotype [49].These observations clearly suggest that BMPs have diverse rolesin tumor progression evenwithin the same tumor type. The effectsof BMPs on cancer cells need to be further studied with specialattention to cell type and environment where they are expressed,in order to achieve deeper understanding of BMP signaling, itsconsequences and the context in which it acts.

Supplementary materials related to this article can be foundonline at doi:10.1016/j.yexcr.2011.06.001.

Conflict of interest

The authors declare no conflict of interest.

Acknowledgments

This work was supported by the Academy of Finland grant no.122440, the Sigrid Juselius Foundation, TheMedical Research Fundof the Tampere University Hospital, the Finnish Cancer Organiza-tions, and the Finnish Cultural Foundation. We also thank KatiRouhento for skilful technical assistance.

R E F E R E N C E S

[1] A. Reddi, Bone morphogenetic proteins: an unconventionalapproach to isolation of first mammalian morphogens, CytokineGrowth Factor Rev. 8 (1997) 11–20.

[2] J. Wozney, Overview of bone morphogenetic proteins, Spine(Phila Pa 1976) 27 (2002) S2–S8.

[3] M. Kawabata, T. Imamura, K. Miyazono, Signal transduction bybone morphogenetic proteins, Cytokine Growth Factor Rev. 9(1998) 49–61.

[4] B.B. Koenig, J.S. Cook, D.H. Wolsing, J. Ting, J.P. Tiesman, P.E.Correa, C.A. Olson, A.L. Pecquet, F. Ventura, R.A. Grant, et al.,Characterization and cloning of a receptor forBMP-2 and BMP-4 from NIH 3T3 cells, Mol. Cell. Biol. 14 (1994)5961–5974.



[5] C.H. Heldin, K. Miyazono, P. ten Dijke, TGF-beta signalling fromcell membrane to nucleus through SMAD proteins, Nature 390(1997) 465–471.

[6] M. Kretzschmar, J. Doody, J. Massague, Opposing BMP and EGFsignalling pathways converge on the TGF-beta family mediatorSmad1, Nature 389 (1997) 618–622.

[7] R. Nishimura, Y. Kato, D. Chen, S.E. Harris, G.R. Mundy, T. Yoneda,Smad5 and DPC4 are key molecules in mediatingBMP-2-induced osteoblastic differentiation of the pluripotentmesenchymal precursor cell line C2C12, J. Biol. Chem. 273 (1998)1872–1879.

[8] S. Kawai, C. Faucheu, S. Gallea, S. Spinella-Jaegle, A. Atfi, R. Baron,S.R. Roman, Mouse smad8 phosphorylation downstream of BMPreceptors ALK-2, ALK-3, and ALK-6 induces its association withSmad4 and transcriptional activity, Biochem. Biophys. Res.Commun. 271 (2000) 682–687.

[9] G. Lagna, A. Hata, A. Hemmati-Brivanlou, J. Massague, Partnershipbetween DPC4 and SMAD proteins in TGF-beta signallingpathways, Nature 383 (1996) 832–836.

[10] R. Derynck, Y. Zhang, Smad-dependent and Smad-independentpathways in TGF-beta family signalling, Nature 425 (2003)577–584.

[11] A. Nohe, E. Keating, P. Knaus, N. Petersen, Signal transduction ofbone morphogenetic protein receptors, Cell. Signal. 16 (2004)291–299.

[12] D. Javelaud, A. Mauviel, Crosstalk mechanisms between themitogen-activated protein kinase pathways and Smad signalingdownstream of TGF-beta: implications for carcinogenesis,Oncogene 24 (2005) 5742–5750.

[13] H.L. Chen, D.M. Panchision, Concise review: bone morphogeneticprotein pleiotropism in neural stem cells and theirderivatives—alternative pathways, convergent signals, Stem Cells25 (2007) 63–68.

[14] B.L.M. Hogan, Bone morphogenetic proteins: multifunctionalregulators of vertebrate development, Gene Dev. 10 (1996)1580–1594.

[15] J.P. Thawani, A.C. Wang, K.D. Than, C.Y. Lin, F. La Marca, P. Park,Bone morphogenetic proteins and cancer: review of theliterature, Neurosurgery 66 (2010) 233–246 (discussion 246).

[16] A. Singh, R.J. Morris, The Yin and Yang of bone morphogeneticproteins in cancer, Cytokine Growth Factor Rev. 21 (2010)299–313.

[17] E.-L. Alarmo, A. Kallioniemi, Bone morphogenetic proteins inbreast cancer: dual role in tumourigenesis? Endocr. Relat. Cancer17 (2010) R123–R139.

[18] J. Howe, J. Bair, M. Sayed, M. Anderson, F. Mitros, G. Petersen, V.Velculescu, G. Traverso, B. Vogelstein, Germline mutations of thegene encoding bone morphogenetic protein receptor 1A injuvenile polyposis, Nat. Genet. 28 (2001) 184–187.

[19] X. Zhou, K. Woodford-Richens, R. Lehtonen, K. Kurose, M. Aldred,H. Hampel, V. Launonen, S. Virta, R. Pilarski, R. Salovaara, W.Bodmer, B. Conrad, M. Dunlop, S. Hodgson, T. Iwama, H. Järvinen,I. Kellokumpu, J. Kim, B. Leggett, D. Markie, J. Mecklin, K. Neale, R.Phillips, J. Piris, P. Rozen, R. Houlston, L. Aaltonen, I. Tomlinson, C.Eng, Germline mutations in BMPR1A/ALK3 cause a subset of casesof juvenile polyposis syndrome and of Cowden andBannayan–Riley–Ruvalcaba syndromes, Am. J. Hum. Genet. 69(2001) 704–711.

[20] E. Corey, R.L. Vessella, Bone morphogenetic proteins and prostatecancer: evolving complexities, J. Urol. 178 (2007) 750–751.

[21] E.T. Keller, J. Zhang, C.R. Cooper, P.C. Smith, L.K. McCauley, K.J.Pienta, R.S. Taichman, Prostate carcinoma skeletal metastases:cross-talk between tumor and bone, Cancer Metastasis Rev. 20(2001) 333–349.

[22] C.J. Logothetis, S.H. Lin, Osteoblasts in prostate cancer metastasisto bone, Nat. Rev. Cancer 5 (2005) 21–28.

[23] L.L. Kodach, S.A. Bleuming, A.R. Musler, M.P. Peppelenbosch, D.W.Hommes, G.R. van den Brink, C.J. van Noesel, G.J. Offerhaus, J.C.Hardwick, The bone morphogenetic protein pathway is active in

06/06/2013 52

human colon adenomas and inactivated in colorectal cancer,Cancer 112 (2008) 300–306.

[24] E.-L. Alarmo, J. Parssinen, J.M. Ketolainen, K. Savinainen, R. Karhu,A. Kallioniemi, BMP7 influences proliferation, migration, andinvasion of breast cancer cells, Cancer Lett. 275 (2009) 35–43.

[25] Y. Katsuno, A. Hanyu, H. Kanda, Y. Ishikawa, F. Akiyama, T. Iwase,E. Ogata, S. Ehata, K. Miyazono, T. Imamura, Bone morphogeneticprotein signaling enhances invasion and bone metastasis ofbreast cancer cells through Smad pathway, Oncogene 27 (2008)6322–6333.

[26] B.T. Feeley, L. Krenek, N. Liu, W.K. Hsu, S.C. Gamradt, E.M.Schwarz, J. Huard, J.R. Lieberman, Overexpression of noggininhibits BMP-mediated growth of osteolytic prostate cancerlesions, Bone 38 (2006) 154–166.

[27] J. Kleeff, H. Maruyama, T. Ishiwata, H. Sawhney, H. Friess, M.W.Buchler, M. Korc, Bone morphogenetic protein 2 exerts diverseeffects on cell growth in vitro and is expressed in humanpancreatic cancer in vivo, Gastroenterology 116 (1999)1202–1216.

[28] K.J. Gordon, K.C. Kirkbride, T. How, G.C. Blobe, Bonemorphogenetic proteins induce pancreatic cancer cellinvasiveness through a Smad1-dependent mechanism thatinvolves matrix metalloproteinase-2, Carcinogenesis 30 (2009)238–248.

[29] N. Bardeesy, K. Cheng, J. Berger, G. Chu, J. Pahler, P. Olson, A. Hezel,J. Horner, G. Lauwers, D. Hanahan, R. DePinho, Smad4 isdispensable for normal pancreas development yet critical inprogression and tumor biology of pancreas cancer, Genes Dev. 20(2006) 3130–3146.

[30] S. Hamada, K. Satoh, M. Hirota, K. Kimura, A. Kanno, A. Masamune,T. Shimosegawa, Bone morphogenetic protein 4 inducesepithelial–mesenchymal transition through MSX2 induction onpancreatic cancer cell line, J. Cell. Physiol. 213 (2007) 768–774.

[31] E.-L. Alarmo, T. Kuukasjarvi, R. Karhu, A. Kallioniemi, Acomprehensive expression survey of bone morphogeneticproteins in breast cancer highlights the importance of BMP4 andBMP7, Breast Cancer Res. Treat. 103 (2007) 239–246.

[32] J. Parssinen, T. Kuukasjarvi, R. Karhu, A. Kallioniemi, High-levelamplification at 17q23 leads to coordinated overexpression ofmultiple adjacent genes in breast cancer, Br. J. Cancer 96 (2007)1258–1264.

[33] J. Parssinen, E.-L. Alarmo, R. Karhu, A. Kallioniemi, PPM1Dsilencing by RNA interference inhibits proliferation and inducesapoptosis in breast cancer cell lines with wild-type p53, CancerGenet. Cytogenet. 182 (2008) 33–39.

[34] W.S. Rasband, ImageJ, U.S. National Institutes of Health, Bethesda,Maryland, USA, 1997–2009 http://rsb.info.nih.gov/ij/.

[35] J.M. Ketolainen, E.-L. Alarmo, V.J. Tuominen, A. Kallioniemi,Parallel inhibition of cell growth and induction of cell migrationand invasion in breast cancer cells by bone morphogeneticprotein 4, Breast Cancer Res. Treat. 124 (2010) 377–386.

[36] I. Kim, D. Lee, H. Ahn, H. Tokunaga,W. Song, L. Devereaux, D. Jin, T.Sampath, R. Morton, Expression of bone morphogeneticprotein receptors type-IA, -IB and -II correlates with tumor gradein human prostate cancer tissues, Cancer Res. 60 (2000)2840–2844.

[37] L.J. Henderson, B.P. Coe, E.H. Lee, L. Girard, A.F. Gazdar, J.D. Minna,S. Lam, C. MacAulay, W.L. Lam, Genomic and gene expressionprofiling of minute alterations of chromosome arm 1p insmall-cell lung carcinoma cells, Br. J. Cancer 92 (2005)1553–1560.

[38] Y. Jin, G.L. Tipoe, E.C. Liong, T.Y. Lau, P.C. Fung, K.M. Leung,Overexpression of BMP-2/4, -5 and BMPR-IA associated withmalignancy of oral epithelium, Oral Oncol. 37 (2001) 225–233.

[39] I.K. Johnsen, R. Kappler, C.J. Auernhammer, F. Beuschlein,Bone morphogenetic proteins 2 and 5 are down-regulatedin adrenocortical carcinoma and modulate adrenal cellproliferation and steroidogenesis, Cancer Res. 69 (2009)5784–5792.

http://rsb.info.nih.gov/ij/


[40] T. Rothhammer, F. Bataille, T. Spruss, G. Eissner, A.K. Bosserhoff,Functional implication of BMP4 expression on angiogenesis inmalignant melanoma, Oncogene 26 (2007) 4158–4170.

[41] H. Deng, R. Makizumi, T.S. Ravikumar, H. Dong, W. Yang, W.L.Yang, Bone morphogenetic protein-4 is overexpressed in colonicadenocarcinomas and promotes migration and invasion ofHCT116 cells, Exp. Cell Res. 313 (2007) 1033–1044.

[42] E.L. Deer, J. Gonzalez-Hernandez, J.D. Coursen, J.E. Shea, J. Ngatia,C.L. Scaife, M.A. Firpo, S.J. Mulvihill, Phenotype and genotype ofpancreatic cancer cell lines, Pancreas 39 (2010) 425–435.

[43] P.S. Moore, B. Sipos, S. Orlandini, C. Sorio, F.X. Real, N.R. Lemoine,T. Gress, C. Bassi, G. Kloppel, H. Kalthoff, H. Ungefroren, M. Lohr, A.Scarpa, Genetic profile of 22 pancreatic carcinoma cell lines.Analysis of K-ras, p53, p16 and DPC4/Smad4, Virchows Arch. 439(2001) 798–802.

[44] S. Buckley, W. Shi, B. Driscoll, A. Ferrario, K. Anderson, D.Warburton, BMP4 signaling induces senescence and modulatesthe oncogenic phenotype of A549 lung adenocarcinoma cells, Am.J. Physiol. Lung Cell. Mol. Physiol. 286 (2004) L81–L86.

06/06/2013 53

[45] S. Piccirillo, B. Reynolds, N. Zanetti, G. Lamorte, E. Binda, G. Broggi,H. Brem, A. Olivi, F. Dimeco, A. Vescovi, Bone morphogeneticproteins inhibit the tumorigenic potential of human braintumour-initiating cells, Nature 444 (2006) 761–765.

[46] T.B. Ro, R.U. Holt, A.T. Brenne, H. Hjorth-Hansen, A. Waage, O.Hjertner, A. Sundan, M. Borset, Bone morphogenetic protein-5, -6and -7 inhibit growth and induce apoptosis in human myelomacells, Oncogene 23 (2004) 3024–3032.

[47] J. Massagué, R. Gomis, The logic of TGFbeta signaling, FEBS Lett.580 (2006) 2811–2820.

[48] A. Anjomshoaa, S. Nasri, B. Humar, J.L. McCall, A. Chatterjee, H.S.Yoon, L. McNoe, M.A. Black, A.E. Reeve, Slow proliferation as abiological feature of colorectal cancer metastasis, Br. J. Cancer 101(2009) 822–828.

[49] C. Morrissey, L.G. Brown, T.E. Pitts, R.L. Vessella, E. Corey, Bonemorphogenetic protein 7 is expressed in prostate cancermetastases and its effects on prostate tumor cells depend on cellphenotype and the tumor microenvironment, Neoplasia 12(2010) 192–205.

Theresa Ascholl

Beneficial Effects of Trypsin Inhibitors Derived from a Spider Venom Peptide in L-Arginine-Induced Severe Acute Pancreatitis in Mice Quelle: PLOS ONE | 1 April 2013 | Volume 8 | Issue 4 | e61049 Betreuer: Dr. Frank Ulrich Weiß (Innere Medizin A) Worauf kommt es mir bei diesem Thema am meisten an? Die Publikation befasst sich - wie auch ich in meiner Forschung – mit einer Form der Pankreatitis. Zwar handelt es sich hierbei nicht um die chronische, sondern um die akute Pankreatitis. Dieser Unterschied lässt das Thema für mich aber keinesfalls weniger interessant erscheinen. Mir ist es wichtig, meinen Blick nicht nur auf die chronische Erkrankungsform zu beschränken, sondern ihn auch auf andere Formen zu richten. Was bedeutet mir das Thema persönlich? Ich habe dieses Paper vor allem deshalb gewählt, weil es sich mit einer recht außergewöhnlichen Behand-lungsmethode befasst und die Idee sowie die Experimente anschaulich gestaltet sind, sodass es auch für meine Kommilitonen von Interesse sein dürfte. Es kommt mir bei dieser Publikation zum einen auf die Methoden an, die mir und evtl. auch anderen von Nutzen sein könnten, da ich mich selbst auch mit Muta-tionen befasse und zum anderen auf die Projektidee selbst, die dazu motiviert, über den Tellerrand zu schauen und auch außergewöhnlich erscheinende Methoden in seine Überlegungen mit einzubeziehen. Was fasziniert mich selbst am Thema am meisten? Mich fasziniert besonders, dass es der Forschungsgruppe gelungen ist, einen potenten Trypsininhibitor zu erstellen, mit dessen Entwicklung es in der Vergangenheit immer wieder zu Problemen kam, und gleich-zeitig dessen Neurotoxizität zu reduzieren. Außerdem hat das Protein HWTI mitsamt seinem Aufbau und seiner zwei unterschiedlichen sowie voneinander unabhängigen Funktionen mein besonderes Interesse geweckt. Weiterhin ergibt sich aus dem Paper die faszinierende Frage, ob es tatsächlich möglich wäre, einem Spinnengift eine heilende Funktion für die akute Pankreatitis im Menschen zuzuschreiben und somit die Bedeutung der Vogelspinne für den Menschen grundlegend zu verändern. Was gefällt mir am Thema weniger? Ich hätte mir gewünscht, in diesem Paper mehr über weitere Anwendungsmöglichkeiten der Spinne Ornithoctonus huwena in der Forschung zu erfahren. Es wäre schön gewesen, hätte man einen kleinen Überblick über andere verwendbare Proteine der Spinne, vielleicht auch außerhalb der Pankreatitis – Forschung, bekommen.

06/06/2013 54

Beneficial Effects of Trypsin Inhibitors Derived froma Spider Venom Peptide in L-Arginine-Induced SevereAcute Pancreatitis in MiceWeiwen Ning1., Yongjun Wang2., Fan Zhang1, Hengyun Wang1, Fan Wang1, Xiaojuan Wang1,

Huaxin Tang1, Songping Liang1, Xiaoliu Shi2*, Zhonghua Liu1*

1College of Life Sciences, Hunan Normal University, Changsha, Hunan, China, 2Department of Digestion, the Second Xiangya Hospital, Central South University,

Changsha, Hunan, China

Abstract

HWTI is a 55-residue protein isolated from the venom of the spider Ornithoctonus huwena. It is a potent trypsin inhibitor anda moderate voltage-gated potassium channel blocker. Here, we designed and expressed two HWTI mutants, HWTI-mut1and HWTI-mut2, in which the potassium channel inhibitory activity was reduced while the trypsin inhibitory activity of thewild type form (approximately 5 EPU/mg) was retained. Animal studies showed that these mutants were less toxic thanHWTI. The effects of HWTI and HWTI-mut1 were examined in a mouse model of acute pancreatitis induced byintraperitoneal injection of a large dose of L-arginine (4 mg/kg, twice). Serum amylase and serum lipase activities wereassessed, and pathological sections of the pancreas were examined. Treatment with HWTI and HWTI-mut1 significantlyreduced serum amylase and lipase levels in a dose dependent manner. Compared with the control group, at 4 mg/kg, HWTIsignificantly reduced serum amylase level by 47% and serum lipase level by 73%, while HWTI-mut1 significantly reducedserum amylase level by 59% and serum lipase level by 72%. Moreover, HWTI and HWTI-mut1 effectively protected thepancreas from acinar cell damage and inflammatory cell infiltration. The trypsin inhibitory potency and lower neurotoxicityof HWTI-mut1 suggest that it could potentially be developed as a drug for the treatment of acute pancreatitis with few sideeffects.

Citation: Ning W, Wang Y, Zhang F, Wang H, Wang F, et al. (2013) Beneficial Effects of Trypsin Inhibitors Derived from a Spider Venom Peptide in L-Arginine-Induced Severe Acute Pancreatitis in Mice. PLoS ONE 8(4): e61049. doi:10.1371/journal.pone.0061049

Editor: Henrik Einwaechter, Klinikum rechts der Isar der TU Munchen, Germany

Received September 26, 2012; Accepted March 5, 2013; Published April 15, 2013

Copyright: � 2013 Ning et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permitsunrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Funding: This work was supported by grants from the National 973 Program of China (No. 2010CB529800), National Nature Science Foundation of China(No.31071091; No.30971570), and through funding received by Dr. Songping Liang from Xiamen Bioway Biotech Co., Ltd. The funders had no role in study design,data collection and analysis, decision to publish, or preparation of the manuscript.

Competing Interests: Co-author Dr Songping Liang was the holder of a Chinese patent (Proteinase inhibitor from the spider selenocosmia huwena) (ZL 2003 10110608.1) in 2006. This patent covers the sequence of HWTI (the protein studied in this paper) and its trypsin inhibitory activity. This patent was transferred toXiamen Bioway Biotech Co., Ltd. Dr. Songping Liang obtained funding from Xiamen Bioway Biotech Co., Ltd for this research. There are no further patents,products in development or marketed products to declare. This does not alter the authors’ adherence to all the PLOS ONE policies on sharing data and materials.

* E-mail: [email protected] (XS); [email protected] (ZL)

. These authors contributed equally to this work.

Introduction

Severe acute pancreatitis (SAP) is an inflammatory process of

the pancreas that causes acinar death, local complications or even

multiple organ failure. The hospital mortality rate associated with

severe acute pancreatitis is very high (20%–30%) despite the

availability of advanced treatment modalities for this disease.

Although the precise pathogenic mechanism of severe acute

pancreatitis remains unclear, the generally accepted theory is that

it is initiated by the intra-pancreatic activation of proteases. The

activation of pancreatic enzymes results in the digestion of local

acinar cells, which induces the release of more enzymes, triggering

local and systemic inflammatory responses [1–4].

The activation of trypsinogen plays a key role in the progress of

SAP [5]. Trypsin is produced in the pancreas as an inactive

proenzyme, trypsinogen, a small proportion of which is activated

in acinar cells and then inhibited by the pancreatic secretory

trypsin inhibitor (PSTI). In parallel, the protease activated

receptor-2 (PAR-2), located on the surface of acinar and duct

cells, mediates a negative feedback loop that inhibits pancreatic

trypsin secretion when the extracellular concentration of trypsin

increases [6]. An excessive activation of trypsinogen to trypsin may

disturb this balance and contribute to the development of

pancreatitis. Activated trypsin may lead to cell damage, which

can in turn trigger the release of more trypsin. Trypsin activates

kallikrein, a serine protease that liberates kinins such as bradykinin

and kallidin from their precursor kininogens. These kinins can

increase vascular permeability and induce vasodilatation and

neutrophil accumulation. Moreover, the release of trypsin and

other toxic factors into the systemic circulation can lead to

cardiovascular and pulmonary collapse [6].

Given the critical role of trypsin in the pathogenesis of SAP,

protease inhibitors have been considered as a potential treatment

for SAP. Indeed, the earliest clinical application of trypsin

inhibitors for the treatment of acute pancreatitis was reported in

the 1960s [7],and in the 1980s, Japanese researchers demonstrated

the safety and effectiveness of trypsin inhibitors for the treatment

of acute pancreatitis [7]. Recent studies have shown that trypsin

PLOS ONE | www.plosone.org 1 April 2013 | Volume 8 | Issue 4 | e6104906/06/2013 55

inhibitors suppress vascular smooth muscle contraction by

inhibiting calcium influx [8], increase blood flow to the pancreas,

protect vascular endothelial cells from damage by oxygen radicals

[9], and effectively reduce the release of inflammation factors [10].

However, the clinical efficacy of antiproteases is still a matter of

controversy. The use of aprotinin, a broad spectrum Kunitz type

protease inhibitor with strong action against trypsin, chymotrypsin

and kallikrein, for the treatment of acute pancreatitis has been

studied for the last 50 years with disappointing results [7].

However, some researchers argue that most of the studies

addressing the use of aprotinin for the treatment of acute

pancreatitis have not been conducted adequately, have inappro-

priate end-points, and most importantly, have not attained

adequate plasma and peritoneal levels of aprotinin to produce

sufficient inhibitory activity. These researchers proposed that

a well-powered study with adequate aprotinin dosing may clarify

its clinical benefit in severe acute pancreatitis [7].

HWTI is a bifunctional protein that possesses inhibitory activity

against trypsin and blocks voltage-gated potassium channels. It

was isolated from the venom of the spider Ornithoctonus huwena. It

consists of 55 amino acids with 6 cysteine residues forming 3 pairs

of disulfide bonds, and belongs to the BPTI/Kunitz-type serine

protease inhibitor family. HWTI is a very strong trypsin inhibitor

with an inhibitory potential approximately 30-fold higher than

that of aprotinin. However, unlike aprotinin, HWTI is also

a moderate blocker of voltage-gated potassium channels including

Kv1.1, 1.2 and 1.3. A structure-function relationship study

indicated that the binding sites corresponding to the two functions

of HWTI are independent from each other on the surface of the

protein [11]. Therefore, it is possible to reduce the inhibitory

action of HWTI on potassium channels by mutating specific

residues without affecting the trypsin inhibitory activity of the

protein. Moreover, because HWTI is a stronger trypsin inhibitor

than aprotinin, the mutated analogues could be developed as

potential agents for the treatment of SAP with less neurotoxicity.

In this study, two mutants of HWTI were designed and expressed

using a yeast expression system. The protective effect of the

recombinant proteins against SAP was further evaluated in

a mouse SAP model induced by intraperitoneal injection of a large

dose of L-arginine (L-Arg).

Materials and Methods

MaterialsEscherichia coli strain TOP10 was used for production of

plasmids. Saccharomyces cerevisiae strain S78 (Leu2, Ura3, Rep4)

and vector pVT102U/a were used for expression of HWTI and its

mutants. Uracil-deficient YSD and YPD media were used for

propagation of yeast transformants. Enzymes used for DNA

manipulation were purchased from MBI Fermentas (USA). The

gel extraction mini kit was from QIAGEN. A DNA ladder was

purchased from Takara. Tryptone and yeast extract were from

Oxoid. Proteose peptone was obtained from Sangon (Shanghai,

China), and the carrier DNA was from Clontech. All other

reagents were from Sigma, and all chemicals were of analytical

reagent grade. Male Wistar rats weighing 220 to 250 g were used.

Kunming mice (of both sexes, each half) with a body weight of

2062 g were used for assessing toxicity and pharmacodynamics.

All animal experiments were approved by the Animal Welfare

Committee of Hunan Normal University.

Primer Design and Construction of PlasmidsThe cDNA sequence of HWTI was obtained by the RACE

method. PCR primers were designed based on the sequence of

HWTI [11,12] and synthesized by GenScrip. HWTI-forward (59-

CGTCTAGATAAGAGAATAGATACATGCCGTTTGCCC-

39) was used as sense primer with an XbaI restriction site, HWTI-

reverse (59- CCGAAGCTTTTATGCTTTTGCA-

CATCTTTTC-39) was used as antisense primer with a HindIII

restriction site. The amplified products were cloned into the

expression plasmid pVT102U/a after being digested with XbaI

and HindIII. PCR primers corresponding to HWTI-mut1 (for-

ward: 59-CGTCTAGATAAGAGAATAGATACATGCCTT-

GAACCCTCTGACACTGGGAG-39, reverse: 59-CCGAAGCT-

TATGCTTTTGCACATCTTTTCATG-39) and those for

HWTI-mut2 (forward: 59-CGTCTAGATAAGAGAATAGATA-

CATGCCTTGAACCC-39, reverse: 59-CCGAAGCT-

TATGCTTTTGCACATCTTTTCATG -39) were designed ac-

cording to the sequence of pVT102U/a-HWTI and synthesized

by GenScrip. The amplified target fragments were obtained by

inverse-PCR using pVT102U/a-HWTI-R25A as a template.

Conditions for PCR were as follows: 94uC for 5 min; 9uC for

1 min; 63uC for 1 min. 72uC for 1 min, 27 cycles, and 72uC for

1 min. Reverse-PCR was used to amplify the target fragment from

the HWTI primer [13].

Expression of Recombinant HWTI and its MutantsPlasmids confirmed by DNA sequencing were transformed into

S. cerevisiae S78 by using the LiCl method [13]. One positive

transformant on the YSD plate was cultured in 25 mL of YSD

liquid media for 24 hours and then transferred to 750 ml of YSD

liquid media at a dilution of 1:30 for 3–4 days of cultivation.

Fermentation liquor was mixed with 1 M NaAc (pH=4.2) to

reach a final concentration of NaAc of 0.1 M. The mixture was

centrifuged at 11,000 rpm for 20 min, and the supernatant was

collected for further purification.

Purification of Recombinant HWTI and its MutantsThe culture supernatant was filtered through a 0.45 mM filter

membrane and applied onto a CM-Sepharose column (3630 cm)

previously equilibrated with 0.1 M NaAc. After the column was

balanced with 0.1 M NaAc with 10 column volumes, the

chromatography was developed by a stepwise elution with

increasing concentrations of NaCl (0.1, 0.2, 0.5 and 1 M) in

0.1 M NaAc. The recombinant HWTI and its mutants were

eluted at 0.5 M NaCl. The corresponding fractions were collected,

lyophilized and desalted by reverse-phase HPLC (RP-HPLC). The

molecular weights of HWTI and its mutants were determined by

MALDI-TOF MS (Bruker, UltrafleXtreme MALDI-TOF-TOF).

Analysis of the Trypsin Inhibitory Activities of HWTI andits MutantsTrypsin inhibitory activity was assessed using two methods. The

first method is Dixon method [14]. Our previous study showed

that HWTI is a competitive inhibitor of trypsin [11]. The equation

(1) governing this relationship is given below:

1

v~

KM:½I�

VMAX:½S�:Ki

z1

VMAX1z

KM

½S�

� �ð1Þ

Plots were prepared of the reciprocal of rate of metabolite

formation (1/v) versus inhibitor concentration [I] at each substrate

concentration [S]. Velocity (v) was measured according to the

method of Erlanger et al [15] as modified by Benjakul et al [16]

using Na-Benzoyl-DL-arginine 4-nitroanilide Hydrochloride

(BAPNA) as a substrate. Inhibitors (HWTI, HWTI-mut1 or

Beneficial Effects of Trypsin Inhibitors in SAP


HWTI-mut2) with appropriate dilutions in 100 mL buffer solution

and 10 mL 0.6 mg/mL trypsin (10 mL buffer for control) were

added to 96-well plates. According to our preliminary tests, the

concentrations of HWTI were 25, 50, 75, 100 and 200 nM, while

the concentrations of the two mutants were 2.5, 5.0, 7.5, 10 and

20 nM. The inhibitor and trypsin mixture was preincubated for

15 min at 30uC, and then 30 mL substrate (the final concentration

was 0.2 mM or 0.5 mM) was added into the mixture. After

incubation for 25 min at 30uC, 10 mL of 60% acetic acid (v/v) was

added to terminate the reaction. The absorbance of reaction

mixture was read at 410 nm and v was then calculated using the

following formula (2):

v~A{A0

8800|60|27ð2Þ

where 8800 (M216cm21) is the coefficient of p-nitroaniline; A and

A0 are absorbance at 410 nm of the sample and the control,

respectively. The resulting straight lines were analyzed by linear

fitting Sigmaplot. Estimates of Ki were obtained by simultaneously

solving two sets of equations sequentially. The X axis value of

intersection of these pairs of lines represents the value of -Ki.

The second method is an acid-base titration method described

in European Pharmacopoeia 5.0 [17]. The inhibition activity of

HWTI and the mutants were determined by measuring its

inhibitory action on a solution of trypsin of known activity. Briefly,

titration with 0.1 M NaOH neutralized the acid released by

trypsin hydrolyzing the substrate (benzoylarginine ethyl ester

hydrochloride) to maintain the solution at pH 8.060.1. The

temperature of the solution was maintained at 2560.1uC, the

reaction was continued for 6 min and the volume of NaOH added

was recorded every 1 min. The dose of the trypsin in test group

was two-fold of that in control group, and our experiments

indicated that 0.6 mg/mL of an inhibitor (HWTI, HWTI-mut1 or

HWTI-mut2) could just inhibit approximate half of trypsin in test

group, that is, the residual trypsin activity of the test group was

similar with that in control because the titration curves of the three

proteins were observed to approximately overlap with that of

control group. The numbers of millilitres of 0.1 M NaOH used

per second in the test group (n2 mL) and in the control group (n1mL) were determined, respectively, and therefore the inhibitory

activity in European Pharmacopoeia Units (EPU) per milligram

was calculated from the expression (3):

activity EPUð Þ~ (2n1{2n2)|4000|f)

Wð3Þ

Figure 1. Structure and sequence comparison of HWTI and BPTI. (A) Amino acid sequences of HWTI and BPTI and the two mutants. Thedisulfide bonds are indicated above the sequences. Gray shading indicates identical residues between HWTI and BPTI. The potassium channelinhibition-related residues in HWTI are labeled in green, while the corresponding residues in BPTI are labeled in red. In the two BPTI mutants, themutated residues are indicated in red. (B) 3D structures of HWTI (PDB code, 2JOT) and BPTI (PDB code, 1LD5). The residues corresponding to theinhibitory activity against trypsin and potassium channels are indicated.doi:10.1371/journal.pone.0061049.g001



where 4000 is coefficient, f is calibration factor of NaOH (f was

0.217 in this study), and W is weight of an inhibitor (mg).

Potassium Channel Inhibitory Activity of HWTI and itsMutantsPotassium channel activity was assessed on rat dorsal ganglia

(DRG) neurons using the whole cell patch clamp technique. Rat

DRG neurons were rapidly dissociated and maintained in a short-

term primary culture and inhibition of potassium channels on rat

DRG neurons was examined according to the method described

by Yuan et al. [11]. Experimental data were acquired and

analyzed by the program Pulsefit 8.0 (HEKA). All data are

presented as means 6 standard error.

Figure 2. Purification of HWTI and the two mutants. Fractions collected from the CM-Sepharose column were applied to a RP-HPLC columnequilibrated with 0.1% trifluoroacetic acid for further purification at a flow rate of 2 mL/min (buffer B, 0.1% trifluoroacetic acid in acetonitrile) with anincrease from 20% to 30% over 13 minutes. The molecular mass of the purified protein was measured by MALDI-TOF mass spectrometry. Theretention time of HWTI was 19.23 min. (A), and the measured MW was 6,173.25 (B). For HWTI-mut1, the retention time was 21 min (C), and themeasured MW was 6,089.52 (D). The retention time of HWTI-mut2 was 19.5 min (E), and the measured MW was 6,138.12 (F).doi:10.1371/journal.pone.0061049.g002



Intracerebroventricular (i.c.v) Injection96 mice(48 males and 48 females)weighing 18-22 g were

randomly divided into 12 groups:control, HWTI (110, 165, 250,

375 and 562.5 mg/kg), HWTI-mut1 (110, 165, 250, 375, 562.5

and 25,000 mg/kg). Mice were anesthetized with ethyl ether before

injection. A 5 mm65 mm incision was made at the midpoint

between ears and the cranium was exposed. On one side of the

intersection of sagittal sutures and lambdoidal sutures, a hole was

made with a micro injector and 15 mL of solution was injected into

the 4th ventricle. The animal’s behaviors were observed for 48

hours after the injection [18,19].

Evaluation of Therapeutic Effects in a Mouse SAP ModelMouse SAP was induced by intraperitoneal injection of a large

dose of L-arginine (Arg) as described previously [18,19]. Briefly,

130 male mice weighing 22-26 g were fed with a standard diet and

housed in a standard shoebox cage at 20–25uC with a 12 hours

dark/light cycle. A solution of L-Arg (8%) was prepared in normal

saline and the pH was adjusted to 7.0. The solution was

administered intraperitoneally at a dose of 4 g/kg. 1 hour after

the first injection, another injection was executed at the same dose.

Animals were then returned to the cages and allowed free access to

food and water. 6 hours after the first injection, animals were

randomly divided into 12 groups: 0.25 mg/kg (62), 0.5 mg/

kg(62), 1 mg/kg(62), 2 mg/kg(62), 4 mg/kg(62), normal and

control. HWTI and HWTI-mut1 were dissolved in normal saline

and diluted to the required concentrations (0.25–4 mg/kg). A

volume of 0.2 mL of HWTI or HWTI-mut1 solutions of different

concentrations was administered intraperitoneally 8, 24 and 48

hours after the first injection of L-Arg. The control group received

a sham injection of saline. 72 hours after the injection of L-Arg,

blood was collected by using retro-orbital bleeding method, and

centrifuged at 3,000 rpm for 10 min at 4uC.

The serum obtained was used for measurement of amylase and

lipase activity. The serum amylase activity was measured using

starch-iodine method described by Caraway [20,21,22]. Briefly,

10 mL serum sample (H2O for control) was incubated with 500 mLreagent 1 (starch) for 7.5 min at 37uC, 500 mL reagent 2 (iodine),

2.5 mL H2O were immediately added to the reaction solution and

the amount of color development was determined by measuring

the absorbance at 660 nm. The serum amylase activity was

determined by the difference of absorbance between control and

sample. The serum lipase activity was measured using turbidi-

metric assay described by Shihabi ZK et al [23]. 25 mL sample

(saline for control) and 2 mL substrate buffer were mixed and the

absorbance of the mixture was measured at 420 nm 30 seconds

later. Thereafter, the mixture was incubated in 37uC for 10 min

and measured at 420 nm. The serum lipase activity was de-

termined by difference between first and second absorbance.

The pancreas was rapidly removed en bloc, fixed in formalin,

embedded in paraffin, sectioned, stained with hematoxylin (H) and

eosin (E), and evaluated by a pathologist unaware of the sample

identity. 3–5 randomly selected pancreas sections were scored for

pancreas edema, inflammatory cell infiltration, pancreas hemor-

rhage and necrosis according to quantitative method described in

the references [24–26]. The scoring criterion for edema and

inflammatory cell infiltration: 0, absent; 1, mild; 2, moderate; 3,

severe. The scoring criterion for hemorrhage: 0, absent; 3, less

than 2 foci; 5, 3–5 foci; 7, more than 5 foci. The scoring criterion

for necrosis: 0, absent; 3, less than 5%; 5, 5220%; 7, more than

20%. The sum of the scores represented severity of overall

pancreatic damage.

Figure 3. Effect of HWTI and the twomutants on trypsin. Estimation of Ki of HWTI (A), HWTI-mut1 (B) and HWTI-mut2 (C) by Dixon plots. Y axisrepresents reciprocal of velocity (1/v) and X axis represents concentrations of HWTI/HWTI-mut1/HWTI-mut2. The concentration of substrate was0.2 mM (open circle) or 0.5 mM (closed circle). (D) The titration curves of HWTI/HWTI-mut1/HWTI-mut2. The concentration of HWTI and the twomutants was 0.6 mg/mL. Each measurement was conducted in triplicate. Data are expressed as mean6SE (n = 3). Inset, showing the trypsin inhibitoryactivity (expressed in EPU/mg) of the three proteins.doi:10.1371/journal.pone.0061049.g003



Statistical AnalysisData are expressed as means 6 SE. Data analysis was done by

applying unpaired two-tailed Student’s t-test with significance

being assigned to P values,0.05.

Results

Expression and Purification of Recombinant HWTI and itsMutantsHWTI and aprotinin are Kunitz-type serine protease inhibitors.

They share an identical disulfide linkage pattern and therefore

have the same structural motif, namely the Kunitz type motif

(Fig. 1). However, unlike aprotinin, HWTI can also block voltage-

gated potassium channels in addition to its trypsin inhibitory

activity. In our previous study [11], we demonstrated that both

Arg5 and Leu6 are key residues for HWTI binding to potassium

channels, whereas Lys14, which is located in a side chain

protruding from the molecular surface, plays a critical role in

the trypsin inhibitory function of HWTI. These active sites are in

separate and independent locations on the surface of HWTI

(Fig. 1), indicating that the substitution of one active site should not

affect the activity of the other. Aprotinin, which is able to inhibit

trypsin but not potassium channels, shares the key residue Lys19

for trypsin inhibition with HWTI, while Arg5 and Leu6 in HWTI

correspond to Leu10 and Glu11 in aprotinin (Fig. 1). This suggests

that the potassium channel inhibitory activity of HWTI could be

reduced by changing Arg5 and Leu6 to Leu5 and Glu6,

respectively. In addition, our previous study showed that replacing

Arg10 by Thr enhanced the trypsin inhibitory activity of HWTI

[11]. We therefore constructed two mutants with the aim of

Figure 4. Effect of HWTI and the two mutants on potassium channel activity in rat DRG cells. Potassium currents were recorded bywhole-cell patch-clamp with rapidly dissociated DRG neurons. (A) Currents were recorded before and after the application of 10 mM HWTI. (B)Currents measured in the presence of HWTI were converted into percent of control currents and plotted versus the log[HWTI], with an IC50 of3.92 mM. (C) Currents were recorded before and after the application of 10 mM HWTI-mut1. (D) Currents measured in the presence of HWTI-mut1were converted into percent of control currents and plotted versus the log[HWTI-mut1]. (E) Currents were recorded before and after the applicationof 10 mM HWTI-mut2. (F) Currents measured in the presence of HWTI-mut2 were converted into percent of control currents and plotted versus thelog[HWTI-mut2].doi:10.1371/journal.pone.0061049.g004



generating a stronger trypsin inhibitor with lower potassium

channel inhibition activity. Arg5, Leu6 and Arg10 were replaced

by Leu5, Glu6 and The10, respectively, generating HWTI-mut1,

while in HWTI-mut2, only Arg5 and Leu6 were replaced by Leu5

and Glu6, respectively.

The recombinant HWTI and its mutants were expressed by

using the vector pVT102U in S. cerevisiae strain S78. The

expression vectors (pVT102U/a-HWTI, -HWTI-mut1 and -

HWTI-mut2) were constructed and confirmed by DNA sequenc-

ing, and then transformed into S. cerevisiae S78. After the yeast were

cultured for 3 days in YPD media, the recombinant HWTI,

HWTI-mut1 and HWTI-mut2 were recovered in the supernatant,

as determined by SDS-PAGE. The supernatant was first subjected

to cationic chromatography and the fraction eluted by 0.5 M

NaCl contained the target product as determined by MALDI-

TOF MS. This fraction was further purified by RP-HPLC. As

shown in Fig. 2, the retention time of the target product was

approximately 20 min. The measured monoisotopic masses of

HWTI, HWTI-mut1, and HWTI-mut2 were 6,173.25, 6,089.52,

and 6,138.12 Da, respectively (Fig. 2), which is consistent with the

masses estimated from their sequences. The yields of the three

recombinant proteins were 9.3 mg/L for HWTI, 6.7 mg/L for

HWTI-mut1, and 2.1 mg/L for HWTI-mut2. The yield of

HWTI-mut2 was much lower than the yields of HWTI and

HWTI-mut1.

Trypsin Inhibitory Activity of HWTI and its MutantsThe trypsin inhibitory activities of the recombinant proteins

were assessed by spectrophotometry and titration based methods.

In the spectrophotometry method, inhibitory constants (Ki) of

HWTI and the two mutants were determined by using Dixon plots

in order to compare their inhibitory abilities on trypsin. As shown

in Fig. 3, Ki values of the three proteins were 6.3e28 M for

HWTI (Fig. 3A), 5.9e29 M for HWTI-mut1 (Fig. 3B) and

1.0e28 M for HWTI-mut2 (Fig. 3C). HWTI-mut1 showed

slightly higher trypsin inhibitory activity than HWTI and

HWTI-mut2. Next, the trypsin inhibitory activities of the

recombinant proteins were further determined by using a titration

method that was previously described for aprotinin [17]. In this

method, inhibitory activity is calculated from the difference

between the initial and residual activity of trypsin and expressed

as European Pharmacopoeia Units (EPU). One EPU inhibits 50

percent of the enzymatic activity of 2 microkatals of trypsin.

Fig. 3D shows the titration curves of the three proteins, from

which the volume (n1) of 0.1 M NaOH used per second in the test

group was determined to be 1.72e23 mL for HWTI,

Figure 5. Effect of HWTI and HWTI-mut1 on the reduction of serum amylase and lipase levels. Severe acute pancreatitis in mice wasinduced by injecting 4 g/kg L-arginine (8%) twice with a 1 hour interval. HWTI and HWTI-mut1 were administered intraperitoneally 8, 24, and 48hours after the first injection of L-arginine. Mice were sacrificed at 72 hours and serum was collected for the measurement of amylase and lipaselevels. (A) HWTI and HWTI-mut1 significantly reduced serum amylase levels at a dose of 1 mg/kg. At the maximal dose (4 mg/kg), the amylase levelwas nearly half of that of the control group. HWTI-mut1 was more effective at reducing amylase levels in the 4 mg/kg group than HWTI. (B) HWTI andHWTI-mut1 significantly reduced serum lipase levels at a dose of 2 mg/kg. In the 4 mg/kg group, serum lipase level was nearly 1/3 of that of thecontrol group. Data are shown in IU/L and expressed as mean+SE (n = 8–12). * denotes p,0.05 compared with the control group; ** denotes p,0.01compared with the control group; # denotes p,0.05 between lined groups.doi:10.1371/journal.pone.0061049.g005



Figure 6. Protective effects of HWTI and HWTI-mut1 against damage to the pancreas. Pancreatic samples were obtained 72 hours afterthe induction of SAP. Formalin-fixed, paraffin-embedded sections of the pancreas were stained with hematoxylin and eosin. (A) Photomicrograph ofthe normal pancreas. (B) Mice were injected with saline after the induction of SAP. Sections show extensive acinar cell damage, edema and evidentleukocyte infiltration. (C) Treatment with HWTI reduced the symptoms of acinar cell damage, leukocyte infiltration and edema. Normal acinar cellarchitecture was seen in the 4 mg/kg group. (D) Injection of HWTI-mut1 had a greater protective effect on the pancreas than HWTI. The 2 mg/kg and4 mg/kg groups showed no significant damage to acinar cells. Magnification: 2006.doi:10.1371/journal.pone.0061049.g006



1.44e23 mL for HWTI-mut1, or 1.48e23 mL for HWTI-mut2;

the volume (n2) of 0.1 M NaOH used per second in the control

group was 1.58e23 mL. Therefore, the trypsin inhibitory

activities were calculated to be 4.18 EPU/mg for HWTI, 5.01

EPU/mg for HWTI-mut1 and 4.88 EPU/mg for HWTI-mut2

(Fig. 3D, inset). Again, the inhibitory ability of HWTI-mut1 was

slightly higher than those of HWTI and HWTI-mut2. Taken

together, our results indicate that mutation of the residues in the

potassium binding site of HWTI has no effect on its trypsin

inhibitory activity. Moreover, a more potent trypsin inhibitor was

successfully designed and expressed.

Comparison of the Potassium Channel InhibitoryActivities of HWTI and the Two MutantsThe potassium channel inhibitory activities of the three proteins

were measured in rat DRG neurons by using the whole-cell patch

technique [19]. As shown in Fig. 4, the three proteins inhibited

potassium channels in rat DRG neurons in a dose-dependent

manner, but the inhibitory activities of HWTI-mut1 and HWTI-

mut2 were much lower than that of HWTI. Consistent with our

previous study, approximately 60% of the total potassium current

was inhibited by HWTI at the maximal concentration of 10 mM(Fig. 4A), and the IC50 value was 3.92 mM (Fig. 4B). However,

HWTI-mut1 or HWTI-mut2 at a dose of 10 mM inhibited less

than 20% of the total potassium currents (Fig. 4C and 4E).

Because the percent inhibition by the mutant proteins was below

50% even at the maximal concentration of 10 mM, their IC50

values could not be determined (Fig. 4D and 4F). However, the

potassium channel inhibitory activities of both mutants were

considerably lower than that of HWTI. These results confirmed

that we successfully designed and expressed trypsin inhibitors with

reduced potassium channel inhibitory activity.

The neurotoxic properties of the recombinant proteins were

examined by intracerebroventicular injection into mice. One

minute after injection of HWTI, mice showed signs of hysteria or

panic as observed in frantic behavior and flipping. Within 48

hours, HWTI even resulted in 27.5% to 75% death of mice with

the increase of doses (1102562.5 mg/kg) and the median lethal

dose (LD50) was calculated to be 217.15 mg/kg by using the

modified Kaber method. On the contrary, HWTI-mut1 showed

considerably lower neurotoxicity than HWTI, as determined by

the absence of the symptoms described for HWTI even after

injection with a maximal dose of 25 mg/kg HWTI-mut1 (40-fold

higher than the maximal dose of HWTI) and no death of mice

were observed. These results indicated that the mutations

introduced successfully reduced the neurotoxicity of HWTI.

Protective Effect of HWTI and HWTI-mut1 in a MouseModel of SAPIntraperitoneal injection of a large dose of L-arginine is known

to induce SAP in mice. This method was first established by

Mizunuma et al. in 1984 and later improved to evaluate

treatments for SAP [27–29]. In the present study, we applied this

method to examine the protective effect of HWTI and HWTI-

mut1. Mice became sluggish and sleepy after the administration of

the first dose of L-Arg (4 g/kg in saline), and the symptoms

became more apparent after the second dose. Most studies rely on

serum amylase and lipase levels, two diagnostic indexes for acute

Figure 7. Histological scores. The scores include graded assessment of pancreas edema (A), inflammatory cell infiltration (B), hemorrhage (C),Score of necrosis (D). (E) Overall assessment of pancreas damage: sum of scores described above. Data are expressed as mean+SE (n = 3–5). * denotesp,0.05 compared with the control group; ** denotes p,0.01 compared with the control group.doi:10.1371/journal.pone.0061049.g007



pancreatitis, and acinar cell injury to assess the severity of

pancreatitis. As compared with the normal group, SAP mice

showed higher serum amylase and lipase levels and increased

acinar cell damage.

Injection of L-Arg significantly increased serum amylase levels

from 3,177 U/L to 6,144 U/L (Fig. 5A). Treatment of these mice

with HWTI and HWTI-mut1 decreased the serum amylase levels

in a dose dependent manner. For HWTI, a significant decrease of

serum amylase level was observed at a HWTI dose of 0.5 mg/kg

(5,187 U/L, p,0.05). Serum amylase levels continued to decrease

at a dose of 1 mg/kg (3,849 U/L, p,0.01) HWTI and reached

values comparable to those of the normal group at the maximum

dose of 4 mg/kg (3,243 U/L, p,0.01). The same dose-effect

relationship was observed after treatment with HWTI-mut1. A

significant decrease of serum amylase level was observed with

HWTI-mut1 at a dose of 1 mg/kg (3,688 U/L, p,0.05), and the

serum amylase level was 2,316 U/L (p,0.01) at 4 mg/kg HWTI.

In contrast to HWTI, HWTI-mut1 had no obvious effect at

0.5 mg/kg, but was more effective in reducing the serum amylase

level at 4 mg/kg (p,0.05).

The experimental induction of pancreatitis caused a significant

rise in serum lipase compared to the normal group (from 128 U/L

to 886 U/L) (Fig. 5B), and treatment with HWTI and HWTI-

mut1 decreased the serum lipase level, especially at doses of 2 mg/

kg (590 U/L and 382 U/L, respectively, p,0.01) and 4 mg/kg

(382 U/L and 235 U/L, respectively, p,0.01). However, there

were no significant differences in the reduction of serum lipase

levels between HWTI and HWTI-mut1 in all five groups.

Histological examination of sections of the pancreas confirmed

the protective effect of the recombinant proteins against SAP

(Fig. 6). Images of sections from the normal group show the typical

architecture of the pancreas in the normal mouse (Fig. 6A). On the

contrary, nearly all of the acinar cells were damaged and evident

inflammatory cell infiltration was observed in the control group

(Fig. 6B). Sections from mice treated with HWTI (Fig. 6C) and

HWTI-mut1 (Fig. 6D) showed different histology depending on

the treatment dose. Mice treated with low doses (lower than 1 mg/

kg) showed severe acinar cell necrosis, prominent neutrophil

infiltration and massive oedema and hemorrhage. These signs of

acinar cell damage were significantly reduced with higher dose of

2 mg/kg, and at a dose of 4 mg/kg, no significant difference was

detected between SAP and normal mice. Using the histological

scoring method, pancreas edema, inflammatory cell infiltration,

pancreas hemorrhage and acinar cell necrosis of sections of the

pancreas were graded, respectively (Fig. 7A, B, C and D). Total

histological scores were calculated and used to evaluate the overall

tissue injury (Fig. 7E). In agreement with the observation above,

the control group presented the highest total histological score of

9.7561.50, indicating the most severe of SAP. However, the

treatment of HWTI and HWTI-mut1 resulted in decreased

histological scores significantly. In the mice treated with high doses

(2 and 4 mg/kg), the scores for oedema, inflammatory cell

infiltration, hemorrhage and necrosis were all reduced significant-

ly, representing lower total histological scores (at 2 mg/kg,

9.6660.66 for HWTI, 10.0060.63 for HWTI-mut1; at 4 mg/

kg, 8.9560.62 for HWTI, 9.0060.32 for HWTI-mut1) (Fig. 7E).

There were no significant differences between HWTI and HWTI-

mut1.

Discussion

The present study aimed to assess the effect of mutations in

HWTI on abolishing the potassium channel inhibitory activity of

the protein and examine the therapeutic effect of HWTI and its

mutant forms in mice with induced SAP. Two mutants, HWTI-

mut1 and HWTI-mut2, were successfully designed and expressed

in yeast. HWTI-mut1 could be expressed at a high level similar to

HWTI. As expected, the potassium channel inhibitory activity was

reduced in both mutants, which therefore showed lower neuro-

toxicity without impairment of their trypsin inhibitory function.

Furthermore, HWTI-mut1 showed higher trypsin inhibitory

activity than HWTI. Significantly, HWTI and HWTI-mut1 were

effective in the treatment of mice in which SAP was induced by

intraperitoneal injection of large doses of L-Arg. Both proteins

effectively decreased the symptoms of SAP and prevented damage

to acinar cells. Taken together, our results indicate that the potent

inhibitory action and lower neurotoxicity of HWTI-mut1 make it

an ideal candidate for the development of anti-acute pancreatitis

drugs.

Extensive research efforts have been devoted to evaluating the

effects of protease inhibitors in animal models of SAP. In most

studies, inhibitors of trypsin and other related proteases exhibited

a desirable protective effect when used before the onset of SAP.

Tetsuya Hirano et al. evaluated the effect of the trypsin inhibitor

urinastatin in a model of experimental acute pancreatitis induced

by a supramaximal dose of cerulein. Urinastatin was administered

2 hours before, during or 1 hour after cerulein infusion. The

authors reported that the administration of urinastatin before and

during cerulein infusion may suppress the pathogenesis and

evolution of pancreatitis [30]. In an acute pancreatitis model

induced by pancreaticobiliary duct obstruction with cerulein

stimulation and systemic hypotension in the rat, Tetsuya Hirano

and Tadao Manabe further confirmed that preoperatively

administered urinastatin was more protective than that used

postoperatively [31]. More recently, in a study conducted by Yun

Ju Jo et al., a human leukocyte elastase inhibitor, recombinant

guamerin, significantly reduced the severity of cerulein-induced

pancreatitis in rats when administered 30 min before the induction

of pancreatitis [32]. In other studies, N-acetylcysteine (a strong

antioxidant) and the cathepsin B inhibitor CA-074me adminis-

tered preoperatively demonstrated beneficial effects in experimen-

tally induced acute pancreatitis [33,34]. Studies have shown that

transient but high levels of trypsinogen activation are observed

very early in most experimental models [35]. Therefore, treatment

with protease inhibitors preoperatively can efficiently inhibit

trypsinogen activation and reduce the severity of acute pancrea-

titis. This might explain why pretreatment can yield better

protective effects as indicated above.

In clinical practice, pretreatment with protease inhibitors is

actually impractical. Therefore, experts on pancreatitis in Japan

recommend administering protease inhibitors as soon as the

diagnosis of acute pancreatitis is confirmed [36]. In the present

study, we evaluated the protective effect of the trypsin inhibitors

(HWTI and HWTI-mut1) administered after the onset of SAP

induced by L-arginine injection in mice. Administration of high

doses of L-arginine, which is used as a relatively easy and

reproducible model of pancreatitis, leads to significantly increased

plasma amylase, pancreatic MPO activity, trypsin activation, and

histological changes resembling acute pancreatitis in humans

[27,28]. Trypsin activity increased significantly after administra-

tion of L-arginine and continued increasing until 120 hours [28].

The prolonged increase in trypsin activity might contribute to the

severity of the damage observed in this model, as serum amylase

and lipase levels were increased significantly at 6 hours and

reached a maximum at 72 hours, indicating that injury to the

pancreas starts early and develops gradually [28]. This observation

is in contrast to the transient pattern of trypsinogen activation in

the cerulein-induced model of pancreatitis. More recently, Gaiser



et al. developed an excellent mouse model of acute pancreatitis in

which active trypsin can be conditionally expressed within

pancreatic acinar cells using a tamoxifen-inducible genetic

construct [1,35]. In this model, intra-acinar trypsinogen activation

was prolonged over a long time, which was also observed in the L-

arginine-induced model. This study also clearly and elegantly

established that intra-acinar trypsin can induce acute pancreatitis

by itself when the intrinsic protective mechanisms are overloaded

[1,35]. These findings suggest that suppression of the prolonged

trypsinogen activation by inhibitor treatment may decrease the

severity of injury even after the onset of acute pancreatitis. In the

present study, the trypsin inhibitors were administered intraper-

itoneally 8, 24 and 48 hours after the first injection of L-Arg and

their effect was assessed after 72 hours. This not only provided

adequate plasma levels of the inhibitors but also ensured that the

treatment was administered after the diagnosis of acute pancre-

atitis was confirmed. Our results showed that HWTI and HWTI-

mut1 were effective in the treatment of L-Arg induced acute

pancreatitis in mice, which was likely due to their capacity to

inhibit trypsin activity and therefore prevent damage to acinar

cells.

It has been widely accepted that intra-acinar trypsinogen

activation is sufficient to induce acute pancreatitis [1,35].

However, increasing evidence indicates that trypsinogen activation

is not the only factor determining the severity of pancreatitis.

Compared with the original ‘‘trypsin paradigm’’, the newer

‘‘multifaceted paradigm’’ has greatly broadened our understand-

ing of acute pancreatitis. In the ‘‘multifaceted paradigm’’, trypsin

activity is one of several interrelated mechanisms that are

simultaneously activated when acinar cells are injured [37–39].

Our results showing the beneficial effects of the recombinant

trypsin inhibitors on L-arginine-induced acute pancreatitis raise

the question of whether these agents can act by mechanisms other

than trypsin inhibition and whether they can be effective to treat

acute pancreatitis in other models. This issue warrants further

study and may provide clues to further our understanding of the

incidence, development and treatment of acute pancreatitis.

Author Contributions

Conceived and designed the experiments: ZL XS SL. Performed the

experiments: WN YW FZ HW FW XW HT. Analyzed the data: WN YW

ZL XS. Contributed reagents/materials/analysis tools: ZL XS SL. Wrote

the paper: WN ZL XS.

References

1. Gaiser S, Daniluk J, Liu Y, Tsou L, Chu J, et al. (2011) Intracellular activation of

trypsinogen in transgenic mice induces acute but not chronic pancreatitis. Gut

60: 1379–1388.

2. Zhang Z, Wang Y, Dong M, Cui J, Rong D, et al. (2012) Oxymatrine

Ameliorates L-Arginine-Induced Acute Pancreatitis in Rats. Inflammation 35:

605–613.

3. Li W, Nakagawa T, Koyama N, Wang X, Jin J, et al. (2006) Down-regulation of

trypsinogen expression is associated with growth retardation in alpha1, 6-

fucosyltransferase-deficient mice: attenuation of proteinase-activated receptor 2

activity. Glycobiology 16: 1007–1019.

4. Nakamura H, Tashiro M, Asaumi H, Nomiyama Y, Kaku M, et al. (2008)

Increased expression of Smad6 deteriorates murine acute experimental

pancreatitis in two models. Gut 57: 788–798.

5. Qiao SF, Lu TJ, Sun JB, Li F (2005) Alterations of intestinal immune function

and regulatory effects of L-arginine in experimental severe acute pancreatitis

rats. World J Gastroenterol 11: 6216–6218.

6. Hirota M, Ohmuraya M, Baba H (2006) The role of trypsin, trypsin inhibitor,

and trypsin receptor in the onset and aggravation of pancreatitis. J Gastroenterol

41: 832–836.

7. Smith M, Kocher HM, Hunt BJ (2010) Aprotinin in severe acute pancreatitis.

Int J Clin Pract 64: 84–92.

8. Kanayama N, Maehara K, She L, Belayet HM, Khatun S, et al. (1998) Urinary

trypsin inhibitor suppresses vascular smooth muscle contraction by inhibition of

Ca2+ influx. Biochim Biophys Acta 1381: 139–146.

9. Smedly LA, Tonnesen MG, Sandhaus RA, Haslett C, Guthrie LA, et al. (1986)

Neutrophil-mediated injury to endothelial cells.Enhancement by endotoxin and

essential role of neutrophil elastase. J Clin Invest 77: 1233–1243.

10. Inoue K, Takano H, Shimada A, Yanagisawa R, Sakurai M, et al. (2005)

Urinary trypsin inhibitor protects against systemic inflammation induced by

lipopolysaccharide. Mol Pharmacol 67: 673–680.

11. Yuan CH, He QY, Peng K, Diao JB, Jiang LP, et al. (2008) Discovery of

a distinct superfamily of Kunitz-type toxin (KTT) from tarantulas. PLoS One 3:

e3414.

12. Liang S (2004) An overview of peptide toxins from the venom of the Chinese

bird spider Selenocosmia huwena Wang [ =Ornithoctonus huwena (Wang)].

Toxicon 43: 575–585.

13. Wu JJ, He LL, Zhou Z, Chi CW (2002) Gene expression, mutation, and

structure-function relationship of scorpion toxin BmP05 active on SK(Ca)

channels. Biochemistry 41: 2844–2849.

14. Kakkar T, Boxenbaum H, Mayersohn M (1999) Estimation of Ki in

a competitive enzyme-inhibition model: comparisons among three methods of

data analysis. Drug Metab Dispos 27: 756–762.

15. Erlanger BF, Kokowsky N, Cohen W (1961) The preparation and properties of

two new chromogenic substrates of trypsin. Arch Biochem Biophys 95: 271–278.

16. Benjakul S, Visessanguan W, Thummaratwasik P (1999) Isolation and

characterization of trypsin inhibitors from some Thai legume seeds. J Food

Biochem 24: 107–127.

17. Aprotinin. In:European Pharmacopoeia 5.0 monograph A:1015–1016.

18. Meiri N, Ghelardini C, Tesco G, Galeotti N, Dahl D, et al. (1997) Reversible

antisense inhibition of Shaker-like Kv1.1 potassium channel expression impairs

associative memory in mouse and rat. Proc Natl Acad Sci U S A 94: 4430–4434.

19. Nathan JD, Romac J, Peng RY, Peyton M, Macdonald RJ, et al. (2005)Transgenic expression of pancreatic secretory trypsin inhibitor-I ameliorates

secretagogue-induced pancreatitis in mice. Gastroenterology 128: 717–727.

20. Caraway WT (1959) A stable starch substrate for the determination of amylase

in serum and other body fluids. Am J Clin Pathol. 32: 97–99.

21. Gitlitz PH, Frings CS (1976) Interferences with the starch-iodine assay for serum

amylase activity, and effects of hyperlipemia. Clin Chem 22: 2006–2009.

22. Jacewicz D, Dabrowska A, Wyrzykowski D, Pranczk J, Wozniak M, et al. (2010)A novel biosensor for evaluation of apoptotic or necrotic effects of nitrogen

dioxide during acute pancreatitis in rat. Sensors. 10: 280–291.

23. Shihabi ZK, Bishop C (1971) Simplified turbidimetric assay for lipase activity.

Clin Chem. 17: 1150–1153.

24. Spormann H, Sokolowski A, Letko G (1989) Effect of temporary ischemia upon

development and histological patterns of acute pancreatitis in the rat. Pathol Res

Pract. 184: 507–513.

25. Nakamichi I, Habtezion A, Zhong B, Contag CH, Butcher EC, et al. (2005)

Hemin-activated macrophages home to the pancreas and protect from acute

pancreatitis via heme oxygenase-1 induction. J Clin Invest. 115: 3007–3014.

26. Laukkarinen JM, Van Acker GJ, Weiss ER, Steer ML, Perides G (2007) Amouse model of acute biliary pancreatitis induced by retrograde pancreatic duct

infusion of Na-taurocholate. Gut 56: 1590–1598.

27. Tani S, Itoh H, Okabayashi Y, Nakamura T, Fujii M, et al. (1990) New model of

acute necrotizing pancreatitis induced by excessive doses of arginine in rats. DigDis Sci 35: 367–374.

28. Dawra R, Sharif R, Phillips P, Dudeja V, Dhaulakhandi D, et al. (2007)

Development of a new mouse model of acute pancreatitis induced by

administration of L-arginine. Am J Physiol Gastrointest Liver Physiol 292:G1009–G1018.

29. Biczo G, Hegyi P, Berczi S, Dosa S, Hracsko Z, et al. (2010) Inhibition of

arginase activity ameliorates L-arginine-induced acute pancreatitis in rats.

Pancreas 39: 868–874.

30. Hirano T, Manabe T, Tobe T (1993) Effect of urinary trypsin inhibitor on

pancreatic cellular and lysosomal fragility in cerulein-induced acute pancreatitis

in rats. Dig Dis Sci 38: 660–664.

31. Hirano T, Manabe T (1993) Human urinary trypsin inhibitor, urinastatin,prevents pancreatic injuries induced by pancreaticobiliary duct obstruction with

cerulein stimulation and systemic hypotension in the rat. Arch Surg 128: 1322–

1329.

32. Jo YJ, Choi HS, Jun DW, Lee OY, Kang JS, et al. (2008) The effects of a newhuman leukocyte elastase inhibitor (recombinant guamerin) on cerulein-induced

pancreatitis in rats. Int Immunopharmacol 8: 959–966.

33. Van Acker GJ, Saluja AK, Bhagat L, Singh VP, Song AM, et al. (2002)

Cathepsin B inhibition prevents trypsinogen activation and reduces pancreatitisseverity. Am J Physiol Gastrointest Liver Physiol 283: G794–G800.

34. Yagci G, Gul H, Simsek A, Buyukdogan V, Onguru O, et al. (2004) Beneficial

effects of N-acetylcysteine on sodium taurocholate-induced pancreatitis in rats.

J Gastroenterol 39: 268–276.



35. Sah RP, Saluja AK (2011) Trypsinogen activation in acute and chronic

pancreatitis: is it a prerequisite? Gut 60: 1305–1307.36. Kitagawa M, Hayakawa T (2007) Antiproteases in the treatment of acute

pancreatitis. JOP 8: 518–525.

37. Sah RP, Garg P, Saluja AK (2012) Pathogenic mechanisms of acute pancreatitis.Curr Opin Gastroenterol 28: 507–515.

38. Dawra R, Sah RP, Dudeja V, Rishi L, Talukdar R, et al. (2011) Intra-acinar

trypsinogen activation mediates early stages of pancreatic injury but not

inflammation in mice with acute pancreatitis. Gastroenterology 141: 2210–2217.

39. Ji B, Logsdon CD (2011) Digesting new information about the role of trypsin in

pancreatitis. Gastroenterology 141: 1972–1975.



Georg Laage

Neutrophil extracellular trap cell death requires both autophagy and superoxide generation Quelle: Cell Research (2011) 21:290-304 Betreuer: Prof. Alexander Dressel (Neurologie) Georg Laage: „Ich denke, dass es einen fundierten Überblick über NETose allgemein bietet und dabei auch Verknüpfungen zur Autophagie und zum Burst berücksichtigt, die auch im besonderen für mein Thema interessant sind.“ Was beudeutet mir das Thema persönlich? Das Paper befasst sich mit grundlegenden Voraussetzungen für die Bildung von Neutrophil extracellular traps. Dabei ist bekannt, dass bei verschiedenen Erkrankungen die NETose von neutrophilen Granulozyten vermindert ist. Das nun lässt etwaige Hypothesen zu, die die Abnahme der NETose in solchen Fällen nun möglichen, greifbaren Ursachen zuordnen. Eben diese möglichen Ursachen für geringere NETs-Bildung werden in diesem Paper behandelt. Worauf kommt es mir bei diesem Thema am meisten an? Am meisten kommt es mir auf den Punkt an, der entscheidet, wann ein NETs geformt wird und wann die Zelle in einen anderen Zelltod geht, also was diesen teils überschneidenden Weg spaltet. Was fasziniert mich selbst am Thema am meisten? In diesem Punkt trifft die Grundlagen Forschung zur NETose auf mögliche Anwendung bei der Erklärung ihrer Abnahme bei verschiedenen Erkrankungen. Dabei sind beide Themen so aktuell und ungeklärt, dass der Raum für Spekulationen und Ideen immens ist. Was gefällt mir am Thema weniger? An der Publikation selbst finde ich schade, dass die unterschiedlichen Zelltodformen nicht aufgeschlüsselt wurden, was erheblich zum Verständnis des Zusammenhangs zwischen NETose und Apoptose beige-tragen hätte. Umso bedauerlicher ist, dass das Material den Experimentatoren zur Verfügung stand.

06/06/2013 67

NETosis requires autophagy and superoxide generation

290

npg

Cell Research | Vol 21 No 2 | February 2011

ORIGINAL ARTICLE

Neutrophil extracellular trap cell death requires bothautophagy and superoxide generationQuinten Remijsen1, 2, 3, Tom Vanden Berghe2, 3, Ellen Wirawan2, 3, Bob Asselbergh4, Eef Parthoens4, 5,Riet De Rycke4, Sam Noppen4, Michel Delforge6, Jean Willems1, *, Peter Vandenabeele2, 3, *

1IRC - Laboratory of Biochemistry, Department of Medicine, KU Leuven Campus Kortrijk, Kortrijk, Belgium; 2Molecular Signal-ing and Cell Death Unit, Department for Molecular Biomedical Research, VIB, Ghent, Belgium; 3Molecular Signaling and Cell Death Unit, Department of Biomedical Molecular Biology, Ghent University, Ghent, Belgium; 4Microscopy Core Facility, Depart-ment for Molecular Biomedical Research, VIB, Ghent, Belgium; 5Department of Biomedical Molecular Biology, Ghent University, Ghent, Belgium; 6University Hospital Gasthuisberg, Department of Hematology, Leuven, Belgium

*These two authors share equally senior authorship to this work.Correspondence: Peter Vandenabeelea, Jean Willemsb

aTel: +32-0-9-33-13760, Fax: +32-0-9-33-13609E-mail: [email protected]: +32-56-246225, Fax: +32-56-246997E-mail: [email protected] 24 June 2010; revised 28 July 2010; accepted 1 September 2010; published online 9 November 2010

npgCell Research (2011) 21:290-304.© 2011 IBCB, SIBS, CAS All rights reserved 1001-0602/11 $ 32.00 www.nature.com/cr

Neutrophil extracellular traps (NETs) are extracellular chromatin structures that can trap and degrade microbes. They arise from neutrophils that have activated a cell death program called NET cell death, or NETosis. Activation of NETosis has been shown to involve NADPH oxidase activity, disintegration of the nuclear envelope and most gran-ule membranes, decondensation of nuclear chromatin and formation of NETs. We report that in phorbol myristate acetate (PMA)-stimulated neutrophils, intracellular chromatin decondensation and NET formation follow autophagy and superoxide production, both of which are required to mediate PMA-induced NETosis and occur independently of each other. Neutrophils from patients with chronic granulomatous disease, which lack NADPH oxidase activity, still exhibit PMA-induced autophagy. Conversely, PMA-induced NADPH oxidase activity is not affected by pharmaco-logical inhibition of autophagy. Interestingly, inhibition of either autophagy or NADPH oxidase prevents intracellular chromatin decondensation, which is essential for NETosis and NET formation, and results in cell death characterized by hallmarks of apoptosis. These results indicate that apoptosis might function as a backup program for NETosis when autophagy or NADPH oxidase activity is prevented.Keywords: neutrophil extracellular trap; granulocyte; chronic granulomatous disease; superoxide; autophagy; live cell imag-ingCell Research (2011) 21: 290-304. doi:10.1038/cr.2010.150; published online 9 November 2010

Introduction

Neutrophils are short lived but nevertheless very abun-dant phagocytic leukocytes that form a vital first line of defense against invading pathogens. Neutrophil contents are potentially harmful to the host and the release and activation of their microbicidal arsenal is strongly con-trolled by strict regulation of degranulation and superox-

ide production, as well as by regulation of their life span and recruitment into tissues. In the absence of immuno-logical challenge, differentiated neutrophils are commit-ted to undergo caspase-dependent apoptosis within 24 to 48 h after their emigration from the bone marrow [1]. During infection, the rate of programmed cell death can be further modulated by various exogenous and endog-enous stimuli that either extend or shorten the neutrophil life span [2].

In 2004, a novel type of cell death was reported, neu-trophil extracellular trap (NET) cell death [3], which is also named NETosis [4]. Activation of NETosis has been shown to involve NADPH oxidase (Nox2)-mediated oxi-dative burst [5] and disintegration of the nuclear enve-lope and most granule membranes, which together result in massive vacuolization [6], intracellular decondensa-tion of nuclear chromatin [7] and eventually formation of

06/06/2013 68

www.cell-research.com | Cell Research

Quinten Remijsen et al.291

npg

NETs [3]. An essential role in the regulation of NETosis is ascribed to phagocyte Nox2, a highly regulated mem-brane-associated multiprotein complex producing large amounts of superoxide that lead to an oxidative burst [5]. Phorbol myristate acetate (PMA) induces NETosis, but this is prevented by inhibition of NADPH oxidase activity (by diphenylene iodonium, DPI, [6]) and by its absence (as in patients with chronic granulomatous dis-ease (CGD), caused by a congenital defect in Nox2 sub-units [8]). Also typical during NETosis is the generation of many vesicles before plasma membrane rupture [6]. These vesicles have a double phospholipid bilayer and are believed to originate from the nuclear envelope [6, 9], which disintegrates during NET cell death. Finally, but still before plasma membrane permeabilization, nuclear chromatin decondenses and mixes with the contents of the granules; this is essential for formation of functional NETs [7, 10]. Permeabilization of the neutrophil plasma membrane releases these chromatin structures, which are loaded with concentrated antimicrobial molecules, such as lactoferrin, BPI, LL-37 and histones [9, 11]. These structures can trap and possibly even kill microorgan-isms. Consequently, neutrophils can exert an antimicro-bial effect beyond their life span [9].

Data collected over the past 5 years demonstrate the in vivo occurrence of NETosis in different clinical settings such as appendicitis [3], necrotizing fasciitis [12], pneu-monia [13], sepsis [14], leishmaniasis [15] and small vessel vasculitis (SVV) [16], suggesting a pathophysio-logical relevance in these conditions. Recently, the kinet-ics of in vivo NET formation in murine lungs in response to Aspergillus infection was monitored [17]. The results showed that NETs are formed during the early stages of infection. In addition to the reported induction of NETo-sis formation by bacteria [3, 6, 12, 13], fungi [17-19] and protozoa [15], NETosis has also been shown to be induced by LPS-activated platelets [14] and by antineu-trophil autoantibodies isolated from patients with SVV [16], whereas impaired degradation of NETs has been as-sociated with systemic lupus erythematosus as well [20].

In spite of the accumulating evidence for the physi-ological relevance of NETs, the interrelations of the different subcellular events in NETosis remain elusive. Therefore, a comparative study of neutrophils using high-resolution live-cell imaging was set up to analyze the potential interplay between reactive oxygen species (ROS) generation, mitochondrial membrane potential, intracellular chromatin decondensation and several morphological features, such as massive vacuolization. Moreover, the functional contribution of these subcel-lular events to NETosis was studied by using pharmaco-logical inhibitors and cells from CGD patients.

Results

PMA induces typical features of NETosis, which differs both biochemically and morphologically from apoptosis and necrosis

Phorbol myristate acetate stimulation of human neu-trophils resulted in the formation of NETs (Supplementary information, Figure S1), as reported previously [3, 6, 16, 18, 21]. In contrast to constitutive or anti-Fas-induced neutrophil apoptosis, PMA-induced NETosis was in-sensitive to benzyloxycarbonyl-Val-Ala-Asp(Ome)-fluoromethylketone (zVAD-fmk) (Supplementary infor-mation, Figure S2A). This confirms a previous report that excluded the involvement of caspases in NETosis [6]. This confirmation is further supported by the ab-sence of DEVD-ase activity after PMA stimulation (Supplementary information, Figure S2B). Inhibition of programmed necrosis by pretreatment with Nec-1, an inhibitor of RIP1 kinase activity [22], also did not affect the kinetics of PMA-induced NETosis (Supplementary information, Figure S2C). Live cell imaging of healthy neutrophils stimulated with 100 nM PMA reveals a cell death program characterized by immediate cell flatten-ing and increased adherence, followed by loss of mito-chondrial membrane potential and induction of massive vacuolization within approximately 30 min (Figure 1A; Supplementary information, Video S1). Vacuolization is observed for up to 90-130 min after PMA stimulation, until the nuclear envelope disintegrates and nuclear chro-matin decondenses, which allows it to mix with the cyto-plasmic content. During all that time, plasma membrane integrity is preserved. Within 40 min later, the plasma membrane permeabilizes and decondensed chromatin is released. Figure 1B shows the mean percentage of cells undergoing these different subcellular events from a kinetic analysis of 150 cells from four independent experiments. To examine whether loss of mitochondrial membrane potential is sufficient to induce NETosis or possibly affects PMA-induced NETosis, we treated neu-trophils with the protonophore carbonyl cyanide 3-chlo-rophenylhydrazone (CCCP) and the complex III inhibitor antimycin A. Both CCCP and antimycin A are known to dissipate the mitochondrial membrane potential in neu-trophils [23]. Incubation with these agents alone did not induce NET formation and did not affect PMA-induced cell death kinetics (Supplementary information, Figure S3), which suggests that mitochondrial depolarization is not a crucial mediator of DNA decondensation. In this regard, it is noteworthy that neutrophils depend on gly-colysis instead of oxidative phosphorylation to meet their energy demands [23, 24]. This is in agreement with the profound decrease in the number of mitochondria during

06/06/2013 69


292

npg


differentiation of precursor cells into neutrophils [25].

PMA-induced chromatin decondensation requires Nox2 activity

Stimulation with PMA is known to activate Nox2, which generates superoxide (Supplementary information, Figure S4A). This is of relevance because ROS genera-tion has been reported to be an absolute requisite for

the induction of NETosis [6]. However, when NADPH oxidase activity is inhibited by DPI (Supplementary in-formation, Figure S4B) or is absent due to a genetic de-fect in Nox2 subunits (Supplementary information, Fig-ure S4C and S4D), PMA still induces cell death, albeit with delayed kinetics as compared to normal neutrophils (Figure 2A; Supplementary information, Figure S5). Unstimulated neutrophils from CGD patients or healthy

Figure 1 PMA-induced NETosis is characterized, chronologically, by cell flattening and adherence, a drop in mitochondrial membrane potential, vacuolization and intracellular chromatin decondensation. (A) Isolated neutrophils (15 × 104 PMN per ml) were monitored by live cell imaging for four parameters: morphology using differential interface contrast (DIC), mitochon-drial potential using TMRM (red), cell death using the cell-impermeable DNA dye Sytox Green (green) and chromatin decon-densation using the cell-permeable DNA marker Hoechst 33342 (blue). These were performed in a humidified atmosphere containing 5% CO2 at 37 °C. Cells were stimulated with 100 nM PMA and monitored every min for up to 300 min. Important time points are shown in min. Vacuoles are indicated by arrows, and scale bars represent 10 µm. Results are representative of at least four independent experiments. (B) Kinetic analysis of 150 cells from four independent experiments. Shown is the mean ± SD of the percentage of cells undergoing subcellular events associated with cell death (loss of mitochondrial poten-tial, vacuolization, intracellular chromatin decondensation and plasma membrane permeabilization).

06/06/2013 70



npg

volunteers, whether or not pretreated with DPI, die spon-taneously from apoptosis [26], but more slowly than in PMA-induced NETosis (Figure 2A). Live cell imaging of the subcellular events described above revealed that fol-

lowing PMA stimulation, CGD neutrophils, like normal neutrophils, undergo massive vacuolization and lose their mitochondrial potential quickly (Figures 2B and 1A). However, no intracellular chromatin decondensation was

06/06/2013 71


294

npg


observed in CGD neutrophils at any time between PMA stimulation and plasma membrane disintegration (Figure 2B and 2C). The massive vacuolization without disinte-gration of the nuclear envelope seen in CGD neutrophils after PMA stimulation (Figure 2B and 2C) disfavors the proposed causative link between the disintegration of the nuclear envelope and consequent chromatin decondensa-tion on the one hand, and the process of vacuolization on the other hand [6, 9]. Our observations, therefore, in-dicate that mechanisms other than nuclear envelope dis-integration might contribute to the formation of vesicles and vacuoles with double membranes.

PMA-induced vacuolization involves autophagy and oc-curs independently of Nox2 activity

We hypothesized that the numerous vesicles might originate from induction of autophagy because many cell types exhibit massive vacuolization during autophagy-as-sociated cell death [27, 28]. In agreement with the mas-sive vacuolization on PMA stimulation, which was ob-served by live cell imaging (Figure 1; Supplementary in-formation, Video S1), transmission electron microscopic analysis of neutrophils revealed the presence of vesicles with double membranes (Figure 3A). These vesicles are not observed in unstimulated neutrophils, which are char-acterized by multilobulated nuclei and different types of granules, all of which are surrounded by a single phos-pholipid bilayer. Within 15 min after PMA stimulation, premature autophagosome structures are observed engulf-ing abundant granules and ribosomes (Figure 3A). After fusion of autophagosomes with endosomes/lysosomes, the autophagic cargo is degraded in the autophagolyso-some, which consequently becomes less electron dense (Figure 3B). About 80 min after PMA stimulation, a dif-ferent type of vesicle with a double phospholipid bilayer appears. The high electron density of these vesicles sug-gests that they are not autophagosomes (Figure 3A) but

might be remnants of the nuclear envelope, as proposed by Fuchs et al. [6]. At this stage, the number of classic granules decrease. Finally, 120 min after stimulation, the nuclear chromatin and granule contents are mixed, and classic granules are no longer observed (Figure 3A). To further characterize these double-membrane vesicles as autophagic vesicles, we monitored the localization of LC3, a marker of autophagy [29]. During the early stages of autophagy induction, LC3 translocates from the cy-tosol to newly formed autophagosomes. Indeed, within 15 min of PMA stimulation, homogeneously dispersed LC3 shows a punctate distribution pattern that is typical of autophagy. Interestingly, translocation of endogenous LC3 is observed in both normal and CGD neutrophils (Figure 3C; Supplementary information, Videos S2-S5). These results fit with the PMA-induced vacuolization of CGD neutrophils observed by live cell imaging (Figure 2B and 2C) and indicate that PMA induces autophagy in a superoxide-independent manner. We also measured the autophagic flux, which is important for distinguish-ing increased autophagosome formation from impaired degradation/maturation. Pretreatment with the vacuolar H+-ATPase inhibitor bafilomycin A1 prevents lysosomal acidification [30] and results in the accumulation of au-tophagosomes and consequent protection of LC3 from degradation. Stimulation of CGD or normal neutrophils with PMA for 30 min resulted in a decrease in LC3 lev-els, and bafilomycin A1 pretreatment significantly pre-vented PMA-induced LC3 degradation (Supplementary information, Figure S6). These results, together with the observation of mature autophagolysosomes by EM (Fig-ure 3A), indicate that PMA induces the formation and maturation of autophagosomes.

Inhibition of either Nox2 activity or autophagy prevents PMA-induced NETosis and stimulates apoptosis

To further examine the role of PMA-induced autophagy

Figure 2 Absence of Nox2 activity blocks PMA-induced intracellular chromatin decondensation and delays cell death, where-as vacuolization and the drop in mitochondrial membrane potential are unaffected. (A) Normal and CGD neutrophils (2 × 105) were incubated with or without 10 µM DPI for 30 min, and either left unstimulated or were stimulated with 100 nM PMA for the indicated periods in the presence of the cell-impermeable DNA dye Sytox Green (50 nM). Data are expressed as percentage of maximal Sytox Green fluorescence ± SD (n = 3) as a function of time. *P < 0.01 and **P < 0.001, as compared to PMA-stimulated normal neutrophils. (B) Isolated neutrophils obtained from a CGD patient were examined by live cell imaging for different parameters in a humidified atmosphere containing 5% CO2 at 37 °C. Left panel: morphology using DIC, mitochon-drial potential using TMRM (red), and chromatin decondensation using the cell-permeable DNA marker Hoechst 33342 (blue). Right panel: morphology using DIC, chromatin decondensation using the cell-permeable DNA marker Hoechst 33342 (blue) and cell death using the cell-impermeable DNA dye PI (green). Cells were stimulated with 100 nM PMA and monitored every min for up to 400 min. Important time points are shown in min. Vacuoles are indicated by arrows, and scale bars represent 10 µm. (C) Kinetic analysis of 85 cells from three independent experiments. Shown is the mean ± SD of the percentage of cells undergoing subcellular events associated with cell death (loss of mitochondrial potential, vacuolization, intracellular chromatin decondensation and plasma membrane permeabilization).

06/06/2013 72



npg

Figure 3 PMA-induced vacuolization is preceded by typical features of autophagy. (A) Human neutrophils (3 × 105 PMN per ml) were left unstimulated or were stimulated with 100 nM PMA for 15, 30, 80 or 120 min. Cells were then fixed and examined by transmission electron microscopy (TEM). (B) Zoom in on TEM picture of human neutrophils stimulated with 100 nM PMA for 30 min. De novo formation of isolation membranes (i) (1). Engulfment of cytoplasmic content (2), such as granules or ribo-somes (arrows). Generation of early immature autophagic vacuoles (Avi) (3) characterized by a double phospholipid bilayer (arrowheads). Degradation of vacuolar content on fusion with endosomes/lysosomes in late degradative autophagic vacuole (Avd) (4). (C) PMA stimulation induces recruitment of LC3 to autophagosomes. Normal neutrophils and neutrophils from CGD patients were left unstimulated or were stimulated for 15 min with 100 nM PMA and then fixed, permeabilized and stained for LC3 and for nuclear chromatin, and analyzed by confocal microscopy. DIC and fluorescent images are shown. Scale bars represent 10 µm. Representative results of three independent experiments are shown. In (A and B) the scale bars for full cell images indicate 1 µm and for magnified cell areas 100 nm. Nuclear lobi are indicated (N).

06/06/2013 73


296

npg


in NETosis, we used the PI3K inhibitor wortmannin, which inhibits autophagy [31, 32]. Wortmannin treatment alone did not affect the kinetics (Figure 4C) or morpho-logical changes (Supplementary information, Video S6) of spontaneous neutrophil apoptosis at early time points, as reported previously [26]. As confirmed by live cell imaging, freshly isolated neutrophils pretreated with 100 nM wortmannin no longer show massive vacuolization on PMA stimulation, which further underscores the idea that the massive vacuolization is linked to autophagy induction (Figure 4; Supplementary information, Video S7). Pretreatment with 100 nM wortmannin did not af-fect PMA-induced superoxide production, as shown by live cell imaging using 5-(and-6)-carboxy-2′,7′-dichlorofluorescein diacetate (DCFDA), an ROS-sen-sitive probe (Figure 4A), and as measured by lucigenin chemiluminescence (Figure 4B). Importantly, although Nox2 activity remained intact when autophagy was phar-macologically blocked (Figure 4A), PMA did not induce chromatin decondensation (Figure 4A; Supplementary information, Video S7), but the cell death rate was not affected (Figure 4C). In contrast, Nox2 inhibition led to retardation of the cell death process, as previously mentioned (Figure 2A). However, when we analyzed the cell death process in more detail, we observed that inhibition of either Nox2 or autophagy led to activation of caspases (as measured by DEVDase activity) and to the appearance of morphological features of apoptosis, such as membrane blebbing (Figure 5; Supplementary information, Videos S8 and S9). These events did not occur at any time in normal neutrophils stimulated with PMA (Figure 1; Supplementary information, Figure S1 and Video S1). In conclusion, these results show that in-hibition of either autophagy or Nox2 prevents chromatin decondensation, which is essential for NETosis and NET formation [7], and results in a type of cell death charac-terized by hallmarks of apoptosis. These results indicate that apoptosis might function as a backup program for NETosis when autophagy or Nox2 activity is prevented.

Discussion

NETosis is characterized by decondensation of intrac-ellular chromatin and disintegration of the nuclear enve-lope, which allows mixing of the chromatin with cationic antimicrobial molecules originating from the granules. When the plasma membrane finally permeabilizes, the microbicidal extracellular traps are released. In this ar-ticle we demonstrate that neither inhibition of caspases by zVAD-fmk [6] nor inhibition of RIP1 kinases by necrostatin-1 affects NETosis. Thus, NETosis should be considered a distinct cell death program. It was recently

shown that decondensation of nuclear chromatin preced-ing NET formation is mediated by histone citrullination [7, 10]. The inability of CGD neutrophils to generate NETs in response to PMA indicates that Nox2 activity is essential for the intracellular chromatin decondensation that precedes NET formation [6], but this was not dem-onstrated directly. We demonstrate that Nox2 activity is indeed required for chromatin decondensation in intact neutrophils. We also observed that PMA-induced NETo-sis is associated with induction of autophagy, despite the view that autophagy might not occur in neutrophils [33]. Autophagy is a well-conserved, essential, intracellular degradation process known to regulate protein and organ-elle turnover in many cells [32]. In addition, autophagy has been implicated in cell death as well [28]. As for neutrophil cell death, von Gunten et al. [34] originally observed autophagosome-like structures during nonapop-totic cell death induced by anti-Siglec9 antibodies after priming of neutrophils with proinflammatory cytokines such as GM-CSF, IFN-α and IFN-γ. Autophagy was re-cently confirmed in murine neutrophils by two indepen-dent groups as well as in human neutrophils [35, 36]. Of interest, during the preparation of this article, Mitroulis et al. [37] also reported that autophagy can be observed in human neutrophils in response to PMA stimulation. However, these reports did not investigate autophagy in the context of neutrophil cell death [35-37]. We investi-gated the extent to which autophagy depends on super-oxide production, and whether autophagy induction is implicated in NETosis. Superoxide is the major type of ROS known to stimulate autophagy [38]. Surprisingly, we observed that PMA induces autophagy in CGD neu-trophils, which are devoid of Nox2 activity. These results demonstrate that PMA-induced autophagy does not re-quire Nox2 activity per se.

We next examined the extent to which induced au-tophagy affects events essential for NETosis, such as Nox2 activity and intracellular chromatin decondensa-tion. Pharmacological inhibition of autophagy by wort-mannin prevented the induction of vacuolization but did not affect superoxide production. This inhibition of autophagy markedly prevented intracellular chromatin decondensation without affecting Nox2 activity. This demonstrates that Nox2 activity is necessary but not suf-ficient for the intracellular chromatin decondensation im-plicated in NET formation. Moreover, induced autophagy alone, which is observed in CGD neutrophils lacking Nox2 activity, is also not sufficient to induce intracellular chromatin decondensation. These results demonstrate that a combination of autophagy and superoxide produc-tion is necessary for the induction of intracellular chro-matin decondensation during PMA-induced NETosis.

06/06/2013 74



npg

Figure 4 Inhibition of autophagy by wortmannin blocks vacuolization and intracellular chromatin decondensation, but genera-tion of ROS is unaffected. (A) Human neutrophils were incubated with the ROS probe DCFDA (green) and the cell-permeable DNA dye Hoechst 33342 (blue) for 30 min at 37 °C in a humidified atmosphere containing 5% CO2. Subsequently, medium containing the cell-impermeable DNA marker PI (red) and wortmannin (100 nM) was refreshed. Cells were stimulated with 100 nM PMA and monitored every min by live cell imaging. Important time points are shown in min. Representative results of three independent experiments are shown. The scale bars represent 10 µm. (B) Cells were either untreated or pretreated with wortmannin (100 nM) for 30 min, and then stimulated with 100 nM PMA. Superoxide production was determined by en-hanced chemiluminescence. Mean maximal superoxide production is expressed as percentage of control ± SD (n = 3). (C) In the presence of 50 nM Sytox Green, 2 × 105 neutrophils were incubated with or without wortmannin (100 nM) for 30 min and were either left unstimulated or stimulated with 100 nM PMA for the indicated periods. Cell death was detected by measuring the fluorescence of the cell-impermeable DNA dye Sytox Green. Data are expressed as percentage of maximal Sytox Green fluorescence ± SD (n = 3) as a function of time. *P < 0.05 and **P < 0.01, as compared to PMA-stimulated normal neutrophils.

06/06/2013 75


298

npg


Because intracellular chromatin decondensation is crucial for NETosis, we wondered whether its absence affects the cell death program itself. Interestingly, when either autophagy or Nox2 activity was blocked, PMA induced a type of cell death characterized by membrane blebbing and caspase activity, which are features of apoptosis. Consequently, when intracellular chromatin

decondensation during NETosis was prevented, an alter-native apoptotic cell death pathway came into play. This pathway also eventually resulted in chromatin deconden-sation, but only after plasma membrane permeabilization during post-apoptotic secondary necrosis. Therefore, the chromatin decondensation that follows spontaneous apoptosis of neutrophils ex vivo [26] should not be con-

Figure 5 PMA induces features of apoptosis when Nox2 or autophagy is inhibited. (A) Live cell images of neutrophils isolated from a CGD patient and stimulated with 100 nM PMA. Morphology was examined using DIC and chromatin decondensation using the cell-permeable DNA marker Hoechst 33342 (blue). Cells were monitored every min for up to 300 min, and impor-tant time points are shown in min. PMA stimulation of CGD neutrophils induced membrane blebbing (arrows). (B) Live cell images of isolated neutrophils pretreated with wortmannin (100 nM) for 30 min. Morphology was examined using DIC and chromatin decondensation using the cell-permeable DNA marker Hoechst 33342 (blue). PMA-induced membrane blebbing (arrows) in normal neutrophils pretreated with wortmannin. (C) Normal and CGD neutrophils were incubated with or without DPI (10 µM) for 30 min. Samples were collected (white bars) and the remainder was further stimulated with 100 nM PMA for 3 h (gray bars). Protease activity against the caspase substrate DEVD-amc was determined. Data are expressed as the mean ∆F/min ± SD (n = 3). (D) Normal neutrophils were incubated for 30 min with 10 µM DPI or 100 nM wortmannin (wort) or left untreated. Subsequently, cells were further stimulated for 1 h (light gray bars) or 3 h (dark gray bars) with PMA (100 nM) or with agonistic anti-Fas antibody (250 ng/ml) and analyzed for DEVDase activity. Data are expressed as the mean ∆F/min ± SD (n = 3). *P < 0.01 and **P < 0.001, as compared to PMA-stimulated neutrophils.

06/06/2013 76



npg

fused with proper NET formation during NETosis.Our results demonstrate that both autophagy and Nox2

activity are needed to activate NETosis fully but they might also prevent apoptosis. Indeed, Hampton and co-workers [39-41] have repeatedly shown the ability of activated Nox2 to inhibit caspases. So, we were surprised to observe membrane blebbing and caspase activity in wortmannin-treated neutrophils stimulated with PMA, because NADPH oxidase activity remained unaffected. We propose that both Nox2 activity and autophagy are required to inactivate caspases. Indeed, stimulated CGD neutrophils lacking NADPH oxidase activity still under-go autophagy, show caspase activity and die by apopto-sis. In addition, when autophagy is blocked in stimulated normal neutrophils, caspase activity and apoptosis be-come apparent. These results are in line with the induc-tion of apoptosis observed in mammalian cell lines when starvation-induced autophagy was inhibited [42]. Both prosurvival and prodeath functions have been ascribed to autophagy [43]. A criterion to claim autophagic cell death is to observe decreased cell death when autophagy is inhibited [43]. When neutrophil autophagy was phar-macologically inhibited, we observed impaired intracel-lular chromatin decondensation, but the neutrophils con-tinued to die though with features of apoptosis. This sug-gests that autophagy is required for the initial phase of NETosis (DNA decondensation) rather than for the cell death process itself. Our results using necrostatin-1 [22] also demonstrate that caspase-independent cell death me-diated by RIP1 kinase activity is not implicated in PMA-induced NETosis, which excludes its role in this type of neutrophil cell death.

Bacterial challenge is known to induce Nox2 activa-tion [27], but it can also induce autophagy [35, 36]. It was recently shown that engagement of Toll-like receptor and Fc receptor signaling during phagocytosis induces autophagy in primary neutrophils [36]. Moreover, in line with our results, when primary neutrophils are chal-lenged with Staphylococcus, caspases are inactivated [41] and NETs are formed [6].

A previous observation of CGD neutrophils generating NETs in response to glucose oxidase (GO)-derived ROS, but not to PMA, suggested that ROS alone might be suf-ficient to induce NETosis [7]. We therefore examined H2O2-induced death of human neutrophils. In this regard, it should be noted that direct stimulation with H2O2 is more transient than the continuous production of ROS by GO. Consequently, it is conceivable that direct stimula-tion with H2O2 mimics the Nox2-dependent oxidative burst more closely than GO-derived ROS production. In addition, the GO used to induce NETosis was produced in Aspergillus niger [6]. Because Aspergillus species

have been shown to induce NETosis [18], we cannot rule out the possibility that nonmammalian compounds in GO solutions provided the additional triggers necessary to induce NETosis during GO stimulation. Notewor-thy, it has been shown that H2O2 is not the major ROS regulating autophagy [38], and that it might not even induce autophagy [44]. In agreement with these results, we could not detect any noteworthy autophagy in H2O2-stimulated neutrophils (Supplementary information, Figure S7). Live cell imaging revealed that H2O2 did not induce NETosis, as shown by the absence of cell flatten-ing, adherence, massive vacuolization and intracellular chromatin decondensation in intact cells. At lower mil-limolar concentrations well below 8 mM, H2O2 induced an apoptotic cell death morphology (Supplementary in-formation, Figure S7A and Video S10), whereas at con-centrations above 8 mM, which are not physiological, it induced a necrotic cell death morphology (Supplementary information, Figure S7B and Video S11). These results further support our previous conclusion that ROS gen-eration is insufficient to induce intracellular chromatin decondensation and subsequent NET formation. A non-sufficient role of ROS in NETosis is further supported by the inability of formyl-methionyl-leucyl-phenylalanine (fMLP, a potent inducer of Nox2 activity [26]) to induce NETosis (unpublished observations), presumably due to activation of signaling pathways known to inhibit apop-tosis and autophagy [1, 31, 45]. In addition, stimulation of neonate neutrophils with PMA results in ROS produc-tion but not in NET formation, in contrast to PMA stimu-lation of adult neutrophils [21].

Of interest, LPS and IL-8 have been reported to in-duce NET formation [3]. In contrast, we only observed a delay of apoptosis by stimulation with either LPS [26] or IL-8 (Supplementary information, Video S12). These results are in line with those of others who also did not observe NET formation in response to LPS [14]. Instead, our observations and those of others suggest that LPS and IL-8 protect neutrophils against apoptosis [1, 26]. Of interest, both LPS and IL-8 sensitize Nox2 activity by a process called priming, but they cannot induce Nox2 activity directly [46]. This is relevant, as Nox2 activity is essential for NETosis, as evidenced by the inability of CGD neutrophils to undergo NETosis [6]. The inability of neutrophils from CGD patients to undergo NETosis was recently correlated with their sensitivity to infection [18]. In this regard, it is noteworthy that many other an-timicrobial ROS-dependent processes are also prevented in CGD [47]. Recently, stimulation of neutrophils with compounds that sensitize/prime Nox2 without activating it was also shown to result in the extracellular release of chromatin from viable neutrophils [48]. This chromatin

06/06/2013 77


300

npg


was shown to originate exclusively from mitochondria, which is unusual given the low numbers of mitochondria in neutrophils. These results suggest that functional ex-tracellular traps might originate from the activation of a wide range of different signaling pathways and cellular processes.

In conclusion, we demonstrate that intracellular chro-matin decondensation and NET formation in PMA-stimulated neutrophils are secondary to autophagy and superoxide production, both of which are required but insufficient on their own to mediate NETosis (Figure 6). Moreover, both processes are required to inhibit the activation of caspases and consequent apoptotic cell death (Figure 6). The combined autophagy- and Nox2-dependent chromatin decondensation in intact neutro-phils and the inhibition of caspases might contribute to the generation of functional NETs. It is conceivable that as the interval between intracellular chromatin deconden-

sation and plasma membrane permeabilization lengthens, more intragranular antimicrobial molecules are absorbed by the accessible and voluminous chromatin. In addition, exposure of nuclear histones to the proteolytic activities outside the nucleus might even generate more potent antimicrobial cleavage products, as has been described in fish and amphibians [11], and perhaps the absence of caspase activity is essential in this process. We plan to functionally and biochemically analyze the antimicrobial properties of chromatin released from primary neutro-phils after stimulation of different cell death programs in neutrophils.


Reagents and antibodiesAntimycin A, bafilomycin A1, CCCP, DPI, H2O2, LPS, PMA,

wortmannin, PMSF, Pefabloc-SC and propidium iodide (PI) were

Figure 6 (A) PMA-induced NETosis requires both autophagy and superoxide production, which trigger the intracellular chro-matin decondensation preceding NET formation. (B) Pharmacological (DPI) or genetic (CGD) inactivation of Nox2 activity prevented PMA-induced superoxide production, but not PMA-induced autophagy. No intracellular chromatin decondensation occurred, and despite massive vacuolization, cells finally underwent cell death characterized by features of apoptosis. (C) Pharmacological (wortmannin) inhibition of PMA-induced autophagy did not interfere with PMA-induced superoxide produc-tion. Nevertheless, intracellular chromatin decondensation did not occur, and cells underwent cell death characterized by fea-tures of apoptosis.

06/06/2013 78



npg

purchased from Sigma-Aldrich (St Louis, MO, USA). Poly-l-lysine was purchased from R&D Systems (Abington, UK). zVAD-fmk was from Bachem (Bubendorg, Switzerland), and Annexin V-Alexa 488, Sytox Green (SG), Hoechst blue 33342, tetrameth-ylrhodamine (TMRM) and DCFDA were from Molecular Probes (Carlsbad, OR, USA). Anti-human Fas antibody (clone 2R2) was purchased from Cell Diagnostica (Münster, Germany). Monoclo-nal anti-gp91phox antibody (54.1) was purchased from Santa Cruz (Heidelberg, Germany). LC3 antibody with a higher specificity against LC3-II than for LC3-I was purchased from Nanotools (Teningen, Germany). Goat anti-mouse Alexa 488 Ab was pur-chased from Molecular Probes-Invitrogen. HRP-conjugated anti-tubulin was purchased from Abcam (Cambridge, UK). IL-8 was a kind gift from Dr Paul Proost (Rega Institute, Leuven, Belgium).

PMN isolationBlood samples were collected from healthy volunteers and

two CDG patients after obtaining informed consent. Both CDG patients have a defect in gp91phox. Neither patient suffered from an acute infection at the time blood was drawn. CGD diagnosis was confirmed by measurement of superoxide production after stimula-tion with 100 nM PMA or 0.5 µM fMLP.

Neutrophils were purified by dextran sedimentation, followed by purification on Ficoll-Paque (Pharmacia, Sweden) and hypoton-ic lysis of contaminating red blood cells [26]. Isolated neutrophils were suspended in RPMI 1640 (Sigma-Aldrich) supplemented with 4% heat-inactivated fetal calf serum (FCS), 1% penicillin/streptomycin and 0.1% gentamycin, and incubated at 37 °C in a humidified atmosphere containing 5% CO2.

Caspase cleavage assayAc-DEVD-amc cleavage was measured by a fluorimetric as-

say modified as described previously [49]. Neutrophils (5 × 106 PMN per ml) were incubated for various durations, after which they were centrifuged at 500× g for 5 min. Cell pellets were re-suspended in 140 µl ice-cold caspase lysis buffer (CLB) (1% NP-40, 10 mM Tris-HCl (pH 7.4), 200 mM NaCl, 5 mM EDTA, 10% glycerol, 1 mM oxidized glutathione, 1 mM PMSF, 2 mM DFP, 0.3 mM aprotinin, 1 mM leupeptin). Protein concentration was then determined by the BCA method. A sample containing 15 µg proteins was suspended in a final volume of 150 µl in assay buffer (CLB supplemented with 10 mM DTT and 50 µM DEVD-amc). The release of fluorescent 7-amino-4-methylcoumarin was mea-sured every 2 min for 1 h in a fluorometer (PerSeptive Biosystems, Cambridge, UK) at an excitation wavelength of 370 nm and an emission wavelength of 445 nm. The increase in fluorescence was linear. Data are expressed as the increase in fluorescence per min-ute (∆fluorescence/min).

Cell death assayHuman neutrophils were seeded at 2 × 105 PMN per ml in

black flat-bottomed 96-well plates (Greiner, Nurtingen, Germany) in HEPES-buffered RPMI 1640 medium (Sigma-Aldrich) supple-mented with 4% FCS, 1% penicillin/streptomycin, 0.1% gen-tamycin and 50 nM SG. After incubation for 30 min at 37 °C in 5% CO2 with or without different inhibitors, cells were stimulated with 100 nM PMA, 250 ng/ml anti-Fas antibody or various con-centrations of H2O2, after which the plates were sealed and loaded in a fluorescence Infinite M1000 plate reader (Tecan, Mechelen,

Belgium) and incubated at 37 °C. SG fluorescence was measured every 10 min for various periods. Cells treated with 0.2% Triton X-100 were used as control and experiments were performed in triplicate.

NADPH oxidase activitySuperoxide production was measured as lucigenin-amplified

chemiluminescence using a Biolumat 9505 apparatus (Berthold, Germany) as described [26]. After different treatments, PMN were diluted to 8 × 105 PMN per ml in 250 µl RPMI medium supple-mented with 0.5 mg/ml lucigenin. Then, 50 µl PBS containing 600 nM PMA was added and the kinetics of superoxide production was measured at 37 °C for up to 60 min. As the kinetics of activation was the same for all reaction mixtures, we used the heights of the peaks to express the results as percentages of control.

Live cell imagingNeutrophils were diluted to 15 × 104 PMN per ml in an eight-

chambered 1.0 borosilicate coverglass system (Nalge Nunc Inter-national, Rochester, NY, USA) coated with 0.1% poly-l-lysine and incubated for 30 min at room temperature (RT) in the presence of 3 µg/ml PI, 1:500 Annexin V-Alexa 488 solution (ANN), 10 µg/ml Hoechst 33342 solution (HB), 50 nM TMRM, 1 µM DCFDA or 50 nM SG. Cells were imaged using an Application Solution Multi-Dimensional Workstation (AS-MDW) equipped with a DM IRE2 microscope with a PIFOC P-Piezo element-driven HCX PL APO 63×/1.3, a 75-W Xenon burner (with monochromator) set at 2 mW and a 12-bit CoolSNAP HQ camera (Leica Microsystems, Wetzlar, Germany). Cell morphology was observed by using differential interference contrast (DIC). Hoechst 33342 was excited at 380 nm and emission was detected using a BP340/80/FT400/LP425 filter cube. ANN, DCFDA and SG were excited at 490, 480 and 490 nm, respectively, and emission was detected using a BP470/40/FT500/BP525/50 filter cube. PI and TMRM were excited at 533 and 540 nm, respectively, and emission was detected using a BP515-560/FT580/LP590 filter cube. Phototoxicity and photobleaching were prevented by minimizing the exposure time for fluorescence excitation (< 100 ms) and setting the camera at gain 2 and 2 × 2 binning. Cells were monitored for various times, and multiple im-age stacks were captured every minute, unless indicated otherwise. Different focal planes were set at 1-µm intervals to prevent loss of focus of nonadherent cells. Poly-Lys treatment did not affect the kinetics and morphology of induced cell death of neutrophils.

From each image stack, maximum intensity projections (for PI, ANN, HB, TMRM, DCFDA, SG) and autofocus images (for DIC) were made for each time point by using a script developed in house for ImageJ 1.31i public domain imaging software. 3D deconvolution (iterative restoration based on calculated PSFs) was performed on Image sequences of PI, ANN, HB, TMRM, DCFDA and SG using the Volocity software 5.2.0 (PerkinElmer, Coventry, UK). Subsequent montages of the Multi-tiff time series and three-channel overlays were made in ImageJ 1.31i.

Confocal microscopyIsolated human neutrophils were analyzed by confocal micros-

copy as previously described [50]. In brief, human neutrophils on uncoated glass coverslips were placed in 24-well plates for 30 min at RT. Cells were left unstimulated or were stimulated with 100 nM PMA for various periods up to 60 min at 37 °C in 5% CO2.

06/06/2013 79


302

npg


Subsequently, cells were fixed with 4% paraformaldehyde for 15 min at RT. They were permeabilized with acetone at −20 °C for 5 min. After rehydration with PBS at RT, cells were incubated with blocking buffer (PBS supplemented with 5 mg/ml BSA + 0.5 mg/ml NaN3 + 10% FCS) overnight at 4 °C. They were incubated with or without primary anti-LC3 antibody (10 µg/ml) for 1 h at RT. After 10 washing steps with PAB (PBS supplemented with 5 mg/ml BSA + 0.5 mg/ml NaN3), cells were incubated with 1:500 sec-ondary goat anti-mouse antibody coupled to Alexa 488 for 1 h at RT, and DAPI was added at 10 µg/ml for another 10 min. Then, cells were washed 10 times with PAB, followed by a final rinse with deionized water, after which they were mounted with gelva-tol and sealed with nail polish. Confocal images were captured with a Leica Sp5 AOBS confocal microscope (Leica, Mannheim, Germany). Images were taken by using a 63× HCX PL Apo 1.4 oil objective. DAPI was excited with a UV diode laser at 405 nm and Alexa 488 by using the 488 line of a Multi Argon laser. Z-sections were made at the resolution limit, in this case 0.118 nm, to pro-duce a high-resolution stack suitable for 3D deconvolution and re-construction. 3D iterative restoration based on measured PSFs and 3D reconstruction was performed using the Volocity software 5.2.0 (PerkinElmer). NET cell death features and kinetics were observed in the presence and the absence of fluorescent dyes. Consequently, the observed PMA-induced cell death was not due to phototoxic-ity.

Electron microscopyFreshly purified neutrophils were allowed to adhere to glass

coverslips in RPMI 1640 containing 4% FCS. After incubation for 30 min at 37 °C in 5% CO2, cells were stimulated with 100 nM PMA or left untreated. Neutrophils were then fixed in 0.1 M Na cacodylate buffer containing 4% formaldehyde and 2.5% glutaral-dehyde for 3 h at 4 °C and overnight in new fixative at 4 °C. After washing with 0.1 M Na cacodylate buffer at 4 °C, neutrophils were post-fixed in 1% OsO4 with 1.5% K3Fe(CN)6 in 0.1 M Na cacodylate buffer (pH 7.2). Samples were dehydrated through a graded ethanol series, and a bulk staining with 2% uranyl acetate was included at the 50% ethanol step. The coverslips were then embedded on molds containing pure Spurr’s resin and left to solid-ify at 70 °C. Ultrathin sections made with a Leica EM UC6 ultra microtome were post-stained in a Leica EM AC20 for 40 min in uranyl acetate at 20 °C and for 10 min in lead stain at 20 °C. Grids were viewed with a 1010 transmission electron microscope (JEOL, Japan) operating at 80 kV.

Western blottingFreshly isolated neutrophils (5 × 105 PMN per ml) from either

healthy volunteers or CGD patients were pelleted and resuspended in lysis buffer containing 1% Triton X-100, one tablet of Com-plete Mini protease inhibitor cocktail per 5 ml lysis buffer and supplemented with 1 mM PMSF and 1 mM Pefabloc-SC. After in-cubating the lysate for 15 min on ice, cell debris was removed by centrifugation. Subsequently, 30 µg proteins was separated in 8% SDS-PAGE. Western blots were analyzed by overnight incubation with 1:1 500 anti-gp91phox antibody and 1:6 000 anti-tubulin conju-gated to HRP, followed by incubation with HRP-coupled second-ary antibody at 1:7 000.

StatisticsThe results were analyzed with a two-way ANOVA and a

Bonferroni post-test (GraphPad Prism version 5, San Diego, CA, USA). Significance was accepted at P < 0.05, unless indicated oth-erwise.

Acknowledgments

Special thanks for the enduring help with blood withdrawal go to Dr Eva De Wash and Els Bergé (KUL campus Kortrijk, Depart-ment of Medicine), Dienst Bloedafname of Dr Erik De Logi (Uni-versity Hospital Ghent), Dr Henk Louagie (AZ St-Lucas, Ghent) and Dr Rudy Mortier (IDEWE, Ghent). We thank Gabriel Stelle (UBA, Buenos Aires) for the support. We thank Dr Chris Guerin (DMBR-VIB, Ghent) for critical reading and helpful suggestions and Dr Amin Bredan (DMBR-VIB, Ghent) for editing the article. This research has been supported by Flanders Institute for Bio-technology (VIB) and various grants. European grants: FP6 Apop-Train, MRTN-CT-035624; FP7 EC RTD Integrated Project, Apo-Sys, FP7-200767; Belgian grants: Interuniversity Attraction Poles, IAP 6/18; Flemish grants: Fonds Wetenschappelijke Onderzoek Vlaanderen, 3G.0218.06; Ghent University grants: 01GC0205 and GROUP-ID consortium from the UGent MRP initiative. QR and TVB hold a postdoc mandate position from the ‘Fonds voor Wetenschappelijk Onderzoek’. PV is holder of a Methusalem grant (BOF09/01M00709) from the Flemish Government.

References

1 Luo HR, Loison F. Constitutive neutrophil apoptosis: mecha-nisms and regulation. Am J Hematol 2008; 83:288-295.

2 Maianski NA, Maianski AN, Kuijpers TW, Roos D. Apoptosis of neutrophils. Acta Haematol 2004; 111:56-66.

3 Brinkmann V, Reichard U, Goosmann C et al. Neutrophil ex-tracellular traps kill bacteria. Science 2004; 303:1532-1535.

4 Steinberg BE, Grinstein S. Unconventional roles of the NA-DPH oxidase: signaling, ion homeostasis, and cell death. Sci STKE 2007; 2007:e11.

5 Babior BM. NADPH oxidase: an update. Blood 1999; 93:1464-1476.

6 Fuchs TA, Abed U, Goosmann C et al. Novel cell death pro-gram leads to neutrophil extracellular traps. J Cell Biol 2007; 176:231-241.

7 Wang Y, Li M, Stadler S et al. Histone hypercitrullination me-diates chromatin decondensation and neutrophil extracellular trap formation. J Cell Biol 2009; 184:205-213.

8 van den Berg JM, van Koppen E, Ahlin A et al. Chronic granulomatous disease: the European experience. PLoS ONE 2009; 4:e5234.

9 Brinkmann V, Zychlinsky A. Beneficial suicide: why neutro-phils die to make NETs. Nat Rev Microbiol 2007; 5:577-582.

10 Neeli I, Khan SN, Radic M. Histone deimination as a re-sponse to inflammatory stimuli in neutrophils. J Immunol 2008; 180:1895-1902.

11 Parseghian MH, Luhrs KA. Beyond the walls of the nucleus: the role of histones in cellular signaling and innate immunity. Biochem Cell Biol 2006; 84:589-604.

12 Buchanan JT, Simpson AJ, Aziz RK et al. DNase expression allows the pathogen group A Streptococcus to escape killing in neutrophil extracellular traps. Curr Biol 2006; 16:396-400.

13 Beiter K, Wartha F, Albiger B, Normark S, Zychlinsky A,

06/06/2013 80



npg

Henriques-Normark B. An endonuclease allows Streptococ-cus pneumoniae to escape from neutrophil extracellular traps. Curr Biol 2006; 16:401-407.

14 Clark SR, Ma AC, Tavener SA et al. Platelet TLR4 activates neutrophil extracellular traps to ensnare bacteria in septic blood. Nat Med 2007; 13:463-469.

15 Guimaraes-Costa AB, Nascimento MT, Froment GS et al. Leishmania amazonensis promastigotes induce and are killed by neutrophil extracellular traps. Proc Natl Acad Sci USA 2009; 106:6748-6753.

16 Kessenbrock K, Krumbholz M, Schonermarck U et al. Net-ting neutrophils in autoimmune small-vessel vasculitis. Nat Med 2009; 15:623-625.

17 Bruns S, Kniemeyer O, Hasenberg M et al. Production of extracellular traps against Aspergillus fumigatus in vitro and in infected lung tissue is dependent on invading neutrophils and influenced by hydrophobin RodA. PLoS Pathog 2010; 6:e1000873.

18 Bianchi M, Hakkim A, Brinkmann V et al. Restoration of NET formation by gene therapy in CGD controls aspergillo-sis. Blood 2009; 114:2619-2622.

19 Urban CF, Reichard U, Brinkmann V, Zychlinsky A. Neu-trophil extracellular traps capture and kill Candida albicans yeast and hyphal forms. Cell Microbiol 2006; 8:668-676.

20 Hakkim A, Furnrohr BG, Amann K et al. Impairment of neu-trophil extracellular trap degradation is associated with lupus nephritis. Proc Natl Acad Sci USA 2010; 107:9813-9818.

21 Yost CC, Cody MJ, Harris ES et al. Impaired neutrophil ex-tracellular trap (NET) formation: a novel innate immune defi-ciency of human neonates. Blood 2009; 113:6419-6427.

22 Degterev A, Hitomi J, Germscheid M et al. Identification of RIP1 kinase as a specific cellular target of necrostatins. Nat Chem Biol 2008; 4:313-321.

23 van Raam BJ, Sluiter W, de Wit E, Roos D, Verhoeven AJ, Kuijpers TW. Mitochondrial membrane potential in hu-man neutrophils is maintained by complex III activity in the absence of supercomplex organisation. PLoS ONE 2008; 3:e2013.

24 Borregaard N, Herlin T. Energy metabolism of human neutro-phils during phagocytosis. J Clin Invest 1982; 70:550-557.

25 Maianski NA, Geissler J, Srinivasula SM, Alnemri ES, Roos D, Kuijpers TW. Functional characterization of mitochondria in neutrophils: a role restricted to apoptosis. Cell Death Differ 2004; 11:143-153.

26 Remijsen Q, Vanden Berghe T, Parthoens E, Asselbergh B, Vandenabeele P, Willems J. Inhibition of spontaneous neutro-phil apoptosis by parabutoporin acts independently of NA-DPH oxidase inhibition but by lipid raft-dependent stimula-tion of Akt. J Leukoc Biol 2009; 85:497-507.

27 Kourtis N, Tavernarakis N. Autophagy and cell death in mod-el organisms. Cell Death Differ 2009; 16:21-30.

28 Kroemer G, Galluzzi L, Vandenabeele P et al. Classification of cell death: recommendations of the Nomenclature Commit-tee on Cell Death 2009. Cell Death Differ 2009; 16:3-11.

29 Kabeya Y, Mizushima N, Yamamoto A, Oshitani-Okamoto S, Ohsumi Y, Yoshimori T. LC3, GABARAP and GATE16 localize to autophagosomal membrane depending on form-II formation. J Cell Sci 2004; 117:2805-2812.

30 Yamamoto A, Tagawa Y, Yoshimori T, Moriyama Y, Masaki R,

Tashiro Y. Bafilomycin A1 prevents maturation of autophagic vacuoles by inhibiting fusion between autophagosomes and lysosomes in rat hepatoma cell line, H-4-II-E cells. Cell Struct Funct 1998; 23:33-42.

31 Eskelinen EL. New insights into the mechanisms of mac-roautophagy in mammalian cells. Int Rev Cell Mol Biol 2008; 266:207-247.

32 Levine B, Kroemer G. Autophagy in the pathogenesis of dis-ease. Cell 2008; 132:27-42.

33 Yousefi S, Simon HU. Autophagy in cells of the blood. Bio-chim Biophys Acta 2009; 1793:1461-1464.

34 von Gunten S, Yousefi S, Seitz M et al. Siglec-9 transduces apoptotic and nonapoptotic death signals into neutrophils de-pending on the proinflammatory cytokine environment. Blood 2005; 106:1423-1431.

35 Beertsen W, Willenborg M, Everts V et al. Impaired pha-gosomal maturation in neutrophils leads to periodontitis in lysosomal-associated membrane protein-2 knockout mice. J Immunol 2008; 180:475-482.

36 Huang J, Canadien V, Lam GY et al. Activation of antibacte-rial autophagy by NADPH oxidases. Proc Natl Acad Sci USA 2009; 106:6226-6231.

37 Mitroulis I, Kourtzelis I, Kambas K et al. Regulation of the autophagic machinery in human neutrophils. Eur J Immunol. 2010; 40:1461-1472.

38 Chen Y, Azad MB, Gibson SB. Superoxide is the major reac-tive oxygen species regulating autophagy. Cell Death Differ 2009; 16:1040-1052.

39 Fadeel B, Ahlin A, Henter JI, Orrenius S, Hampton MB. In-volvement of caspases in neutrophil apoptosis: regulation by reactive oxygen species. Blood 1998; 92:4808-4818.

40 Hampton MB, Stamenkovic I, Winterbourn CC. Interaction with substrate sensitises caspase-3 to inactivation by hydro-gen peroxide. FEBS Lett 2002; 517:229-232.

41 Wilkie RP, Vissers MC, Dragunow M, Hampton MB. A functional NADPH oxidase prevents caspase involvement in the clearance of phagocytic neutrophils. Infect Immun 2007; 75:3256-3263.

42 Boya P, Gonzalez-Polo RA, Casares N et al. Inhibition of macroautophagy triggers apoptosis. Mol Cell Biol 2005; 25:1025-1040.

43 Eisenberg-Lerner A, Bialik S, Simon HU, Kimchi A. Life and death partners: apoptosis, autophagy and the cross-talk be-tween them. Cell Death Differ 2009; 16:966-975.

44 Scherz-Shouval R, Shvets E, Fass E, Shorer H, Gil L, Ela-zar Z. Reactive oxygen species are essential for autophagy and specifically regulate the activity of Atg4. EMBO J 2007; 26:1749-1760.

45 Petiot A, Ogier-Denis E, Blommaart EF, Meijer AJ, Codogno P. Distinct classes of phosphatidylinositol 3′-kinases are in-volved in signaling pathways that control macroautophagy in HT-29 cells. J Biol Chem 2000; 275:992-998.

46 Sheppard FR, Kelher MR, Moore EE, McLaughlin NJ, Baner-jee A, Silliman CC. Structural organization of the neutrophil NADPH oxidase: phosphorylation and translocation during priming and activation. J Leukoc Biol 2005; 78:1025-1042.

47 Remijsen Q, Vandenabeele P, Willems J, Kuijpers TW. Recon-stitution of protection against Aspergillus infection in chronic granulomatous disease (CGD). Blood 2009; 114:3497.

06/06/2013 81


304

npg


48 Yousefi S, Mihalache C, Kozlowski E, Schmid I, Simon HU. Viable neutrophils release mitochondrial DNA to form neu-trophil extracellular traps. Cell Death Differ 2009; 16:1438-1444.

49 Vanden Berghe T, Kalai M, van Loo G, Declercq W, Vandena-

beele P. Disruption of HSP90 function reverts tumor necrosis factor-induced necrosis to apoptosis. J Biol Chem 2003; 278:5622-5629.

50 Allen LA. Immunofluorescence and confocal microscopy of neutrophils. Methods Mol Biol 2007; 412:273-287.

(Supplementary information is linked to the online version of the paper on the Cell Research website.)

06/06/2013 82

Daniel Seifert

Shear stress induces iNOS expression in cultured smooth muscle cells: role of oxidative stress Quelle: Am J Physiol Cell Physiol 279: C1880–C1888, 2000 Betreuer: Dr. Raila Busch (Innere Medizin B) Was bedeutet mir das Thema persönlich? Die Publikation mit dem Thema „Shear stress induces iNOS expression in cultured smooth muscle cells: role of oxidative stress” liefert mir Anregungen für meine Doktorarbeit. Einerseits wird bereits im Titel die Verbindung zu meiner Arbeit erwähnt, andererseits werden hier Methoden beschrieben, mit denen ich teil-weise arbeiten werde. Aus diesem Grund ist diese Arbeit für mich persönlich von großem Wert. Worauf kommt es mir bei diesem Thema am meisten an? Neben der Methodik ist es mir bei der Publikation mit dem Thema „Shear stress induces iNOS expression in cultured smooth muscle cells: role of oxidative stress” wichtig, dass die Ergebnisse und der Zusammen-hang zur klinischen Problematik nachvollziehbar erläutert wird. So konnte in den hier beschriebenen Ver-suchen die in vivo Beobachtungen aus Muskelzellen bestätigt werden, dass diese Zellen in Abhängigkeit des Scherstresses verstärkt Stickstoffmonoxid auf Grund der stärkeren Genexpression von iNOs bilden. Was fasziniert mich selbst am Thema am meisten? Mit dieser Arbeit tragen die Verfasser ein wenig zur Aufklärung des Pathomechanismus der überschießen-den Neointima-bildung bei. Dies ist eine relativ häufig auftretende Nebenwirkung von Stentimplantationen, welche in Folge der hohen und steigenden Prävalenz kardiovaskulärer Erkrankungen an Bedeutung ge-winnt. Dieser klinische Bezug führt zu meiner Faszination an dem Thema. Was gefällt mir am Thema weniger? Im Laufe dieser Arbeit wird Bezug zu LPS genommen. Zwar ist es gut, dass die Verfasser dieses als mög-liche Fehlerquelle ausschließen können allerdings werden weitere Stimulatoren der Genexpression von iNOS nur erwähnt und nicht wie das LPS näher untersucht.

06/06/2013 83

279:C1880-C1888, 2000. ;Am J Physiol Cell Physiol Jean-Baptiste MichelWilly Gosgnach, David Messika-Zeitoun, Walter Gonzalez, Monique Philipe andmuscle cells: role of oxidative stressShear stress induces iNOS expression in cultured smooth

You might find this additional info useful...

36 articles, 24 of which you can access for free at: This article citeshttp://ajpcell.physiology.org/content/279/6/C1880.full#ref-list-1

13 other HighWire-hosted articles: This article has been cited by http://ajpcell.physiology.org/content/279/6/C1880#cited-by

including high resolution figures, can be found at: Updated information and serviceshttp://ajpcell.physiology.org/content/279/6/C1880.full

found at: can beAmerican Journal of Physiology - Cell Physiology about Additional material and information

http://www.the-aps.org/publications/ajpcell

This information is current as of May 31, 2013.

0363-6143, ESSN: 1522-1563. Visit our website at http://www.the-aps.org/. Rockville Pike, Bethesda MD 20814-3991. Copyright © 2000 the American Physiological Society. ISSN:molecular physiology. It is published 12 times a year (monthly) by the American Physiological Society, 9650

is dedicated to innovative approaches to the study of cell andAmerican Journal of Physiology - Cell Physiology

at Universitaetsbibliothek G

on May 31, 2013

http://ajpcell.physiology.org/D

ownloaded from

06/06/2013 84

http://ajpcell.physiology.org/content/279/6/C1880#cited-by

http://ajpcell.physiology.org/

Shear stress induces iNOS expression in culturedsmooth muscle cells: role of oxidative stress

WILLY GOSGNACH, DAVID MESSIKA-ZEITOUN, WALTER GONZALEZ,MONIQUE PHILIPE, AND JEAN-BAPTISTE MICHELInstitut National de la Sante et de la Recherche Medicale Unit 460,Centre Hospitalier Universitaire Xavier Bichat, 75870 Paris cedex 18, FranceReceived 8 June 2000; accepted in fina form 14 July 2000

Gosgnach, Willy, David Messika-Zeitoun, WalterGonzalez, Monique Philipe, and Jean-Baptiste Michel.Shear stress induces iNOS expression in cultured smoothmuscle cells: role of oxidative stress. Am J Physiol CellPhysiol 279: C1880–C1888, 2000.—After deendothelializa-tion, the most luminal smooth muscle cells of the neointimaare in contact with blood flo and express inducible nitricoxide synthase (iNOS) in vivo. We hypothesized that shearstress may be a stimulus for this iNOS overexpression. Wehave thus submitted smooth muscle cells to laminar shearand measured the iNOS expression. Shear stress (20 dyn/cm2) induced iNOS mRNA and protein expression, whereasbrain NOS mRNA expression was decreased. Conversely,nitrite production was increased. This production wasblocked by a selective iNOS inhibitor. Pyrrolidine dithiocar-bamate, an antioxidant molecule, and BXT-51072, a glutha-tion peroxidase mimic, both inhibited the shear-inducediNOS expression. Shear stress also increased the expressionof both membrane subunits of NADPH oxidase p22phox andMox-1. Shear stress activated the redox-sensitive nucleartranslocation of the transcription nuclear factor-kB (NF-kB)and stimulated the degradation of both cytosolic inhibitorskB a and b. These results show that shear stress can induceiNOS expression and nitrite production in smooth musclecells and suggest that this regulation is probably mediated byoxidative stress-induced NF-kB activation.

nuclear factor-kB; nitric oxide synthase; NADPH oxidase

THE VASCULAR WALL of large arteries is organized intothree different functional and structural compart-ments. This specifi organization (compartmentaliza-tion of the vascular wall) modulates the response of thevascular wall to different stimuli. Smooth muscle cellsconstitute the media layer, whereas endothelial cellsconstitute the intima and physiologically perceive theshear stress at the interface between the flowin bloodand the fixe arterial wall. In these endothelial cells,shear stress regulates the expression of numerousgenes [angiotensin converting enzyme, platelet-de-rived growth factor, vascular cell adhesion molecule-1(VCAM-1), etc.] (see Ref. 17 for review). In particular,increased shear stress intensity upregulates endothe-lial nitric oxide synthase (NOS) activity and leads to

nitric oxide (NO) overproduction in endothelial cells(23, 24). In response to deendothelialization, smoothmuscle cells migrate from the media to the lumen andform the neointima. During this process, the mostluminal smooth muscle cells are in close contact withthe blood flo and thus sense the shear stress. There-fore, smooth muscle cells undergo the shear stress thatcould modify the pattern of protein expression withinthe most luminal smooth muscle cells. In vivo, a gra-dient of inducible protein expression including induc-ible NO synthase (iNOS) (12) is usually observed fromthe most luminal to the deeper part of the neointimalproliferation (28). We have recently shown that shearstress induces angiotensin converting enzyme expres-sion in smooth muscle cells (9). It would permit theshift of a constitutive endothelial function to smoothmuscle cells in an intimal position. In this way, it hasbeen demonstrated that iNOS expression is induced inthe most luminal smooth muscle cells of the intimallayer in vivo (12).The formation of the neointima is inhibited by high

flo rate (35), and this phenomenon is partially medi-ated by NO production (6). Administration of L-argi-nine, the NOS substrate, was associated with a re-duced neointimal hyperplasia (11) independent of anyendothelial process (32), suggesting that NO may beproduced by smooth muscle cells. Given that smoothmuscle cells are exposed to flo in neointima forma-tion, we hypothesized that shear could participate iniNOS induction and NO production in smooth musclecells. The antiproliferative effects of NO on smoothmuscle cells and its ability to induce expansive remod-eling would be of interest to limit intimal proliferationand lumen stenosis in different pathophysiological sit-uations.A key component of the induction of iNOS is the

nuclear factor-kB (NF-kB). NF-kB is a redox-sensitivefactor that is activated by the cytosolic release of theinhibitor kB (IkB) proteins and the translocation of theactive p50/p65 heterodimer to the nucleus. Increase inthe production of radical oxygen species is a commonpathway to a wide variety of NF-kB inducers (1). Al-

Address for reprint requests and other correspondence: W. Gos-gnach, Institut National de la Sante et de la Recherche Medicaleu460, CHU X. Bichat, 16, rue H. Huchard, 75870 Paris cedex 18,France (E-mail: [email protected]).

The costs of publication of this article were defrayed in part by thepayment of page charges. The article must therefore be herebymarked ‘‘advertisement’’ in accordance with 18 U.S.C. Section 1734solely to indicate this fact.

Am J Physiol Cell Physiol279: C1880–C1888, 2000.

0363-6143/00 $5.00 Copyright © 2000 the American Physiological Society http://www.ajpcell.orgC1880


on May 31, 2013


ownloaded from

06/06/2013 85


though several lines of evidence suggest that shearstress is an inducer of NF-kB activation in endothelialcells (4, 19), there is no data available about the induc-tion of NF-kB in smooth muscle cells submitted toshear stress. The aim of our present study was toevaluate whether shear stress regulates the expressionof iNOS in smooth muscle cells and whether the redox-sensitive factor NF-kB was involved in this regulation.

MATERIALS AND METHODS

Animals. Normotensive male Wistar rats (160–180 g) wereobtained from Iffa Credo (Labresle, France). The procedurefollowed for the care and euthanasia of the studied animalswas in accordance with the European Community standardson the care and use of laboratory animals (Ministere del’Agriculture, France; authorization no. 00577).

Cell isolation and culture. The smooth muscle cells fromthe aortic media were isolated and cultured as describedpreviously (2) and were used at passage 3. The purity of thecultures was assessed by using morphological and immuno-histological criteria. Smooth muscle cells were characterizedwith antibodies raised against smooth muscle cell a-actin (8).

Shear stress device. The cells were seeded on a rectangularplastic (cell culture-treated) plate previously coated withcollagen (0.1% in hydroxy chloride, Sigma). Collagen waschosen as a substrate to increase the adherence forces of thecells submitted to shear stress. Cells were used 2 to 4 daysafter reaching confluenceCells were then exposed to a fluid-impose shear stress

with the use of a parallel plate channel flo device derivedfrom the one described by Levesque and Nerem (15). The cellculture flo chamber was designed to provide a steady,uniform laminar flow It was positioned in a closed continu-ous flo loop. The flo loop consisted of an elevated reservoirthat provided the required pressure drop across the chamberand a roller pump to return the outflo from the collectingreservoir back to the feeding reservoir. The flo chamber andthe entire apparatus were sonicated and sterilized beforeeach experiment to avoid lipopolysaccharide (LPS) contami-nation. The upper reservoir was fille with 350 ml of Dulbec-co’s modifie Eagle’s medium (DMEM, Biomedia) at 37°C(equilibrated with 95% air–5% CO2) supplemented with 10%heat-inactivated fetal calf serum (Biomedia), 20 mM HEPES(Life Technologies), and 1% antibiotic-antimycotic solution(Sigma), with its pH, temperature, and flo rate monitoredcontinuously. The experimental surface of shear-subjectedcells was 18 cm2. As a control, unstressed cells were seeded inthe same conditions without insertion in the flo chamberdevice and were incubated with shear stress-conditionedmedium to test LPS contamination. After being submitted toshear, the cell plates were removed from the flo chambersunder sterile conditions.

Shear stress values. The shear stress intensity Y (dyn/cm2)was calculated as follows (15): Y 5 6 mQ/ph2 where m is themedia viscosity (DMEM 0.0084 6 0.08 poised at 37°C), Q theflo rate (ml/s), p the flo path (1.8 cm), and h the gap heightover the cell layer (0.025 cm). The viscosity of the media andthe flo chamber cross section were constant in the device.To change the intensity of shear stress, the induction florate had to be modified For pharmacological experiments,cells were preincubated with the blocker for 15 min beforebeing submitted to shear stress. The experiments were per-formed with the same concentration of compound in themedium.

Endotoxin measurement. The level of endotoxins (i.e., LPS)was measured in the supernatant of control cells, in thesupernatant conditioned by a 20-dyn/cm2 shear stress, and inthe freshly prepared culture medium using a commercialsemiquantitative assay kit (Sigma) (22). One hundred micro-liters of the sample were incubated for 1 h at 37°C with 100ml of amoebocyte lysate from Limulus polyphemus. The re-sults were evaluated by comparison of the gelation of thesample and compared with a standard curve from 400 to 0.06endotoxin units (EU)/ml of commercial endotoxin solution(Sigma).

Nitrite production. To measure nitrite production, each cellplate was put into 10 ml of DMEMwithout phenol red and 20mM HEPES with or without 2.1025 M of a selective iNOSinhibitor [L-N6(1-iminoethyl)lysine (L-NIL)] (20) for 4 h. Themedium was then removed and nitrites were detected withthe use of a fluorescen assay (18). Nitrite production wasmeasured with the use of a fluorescen assay as describedpreviously (18). In this assay, nitrites, a degradation productof NO, interact with a nonfluorescen substrate (diaminon-aphthalene, Fluka) to form a fluorescen component that isdetectable at 450 nm (1-[H]naphthotriazole). A standardcurve from 10 nM to 10 mM was made using a commercialnitrite solution (Merck) and treated in the same conditions totest the proportionality of the method. One hundred microli-ters of diaminonaphthalene (0.05 mg/ml in HCl 0.62 N) wasadded to 1 ml of the sample. After 10 min of incubation in adark room, 50 ml of NaOH 2.8 N was added to stop thereaction. Fluorescence was read in a spectrofluoromete (Hi-tachi F2000). The cells were scraped off and the total proteinswere assayed using the protein assay system (Bio-Rad). Re-sults were expressed as nanomoles of nitrites per milligramof proteins. Nitrite levels have been shown to reflec .75% ofthe total NO produced by vascular smooth muscle cells (31).

RT-PCR. For RT-PCR analysis, cells were scraped fromeach plate into 1 ml of TRIzol solution (GIBCO BRL). TotalRNA was prepared using the manufacturer’s instructions.One microgram of total mRNA was reverse transcriptedusing an oligo (dT) primer. PCR amplificatio of glyceralde-hyde-3-phosphate dehydrogenase (GAPDH), S14, iNOS,

Table 1. Nucleotide sequence of primers used for PCR with expected size of amplicons

mRNA Sense Primer Antisense Primer Length, bp

GAPDH 59-GTGAAGGTCGGAGTCAACG-39 59-GGTGAAGACGCCAGTGGAC-39 299S14 59-ATCAAACTCCGGGCCACAGG-39 59-CGAGTGCTGTCAGAGGGGATG-39 127iNOS 59-TGCTTTGTGCGGAGTGTCAGT-39 59-CGGACCATCTCCTGCATTTCT-39 227bNOS 59-CTGGCTCAACAGAATACAGGCT-39 59-GCAGTGTACAGCTCTCTGAAGA-39 293p22phox 59-GGAGTGCTCATCTGTCTGCTG-39 59-GTTGGTAGGTGGCTGCTTGAT-39 306Mox-1 59-TCCCTTTACTCTGACCTCTG-39 59-CATAAGAAAACCCCCACCAC-39 578

PCR was carried out to detect mRNAs for rat GAPDH, S14, iNOS, bNOS, p22phox, and Mox-1. GAPDH, glyceraldehyde-3-phosphatedehydrogenase; iNOS, inducible nitric oxide synthase; bNOS, brain nitric oxide synthase.

C1881SHEAR STRESS INDUCES INOS EXPRESSION


on May 31, 2013


ownloaded from

06/06/2013 86


brain NOS (bNOS), p22phox, and Mox-1 mRNA were per-formed using the primers presented in Table 1.

33P-radiolabeled primers were added to the PCR mixture,and the PCR products were separated on an 8% acrylamide/N,N9-dihydroxyethylene-bis-acrylamide gel in 13 Tris-bo-rate-EDTA buffer. After ethidium bromide staining, the gelwas dried and radioactivity was counted using an instantimager. iNOS, bNOS, p22phox, and Mox-1 mRNA levels werenormalized to GAPDH mRNA, and results were expressed inarbitrary units.

Western blot. Cells were scraped off into lysis buffer thatcontained protease inhibitors for Western blot experimentsand total protein measurement. Cell proteins were dena-tured using Laemmli reagent.For the IkB Western blot, 25 mg of the protein samples

were electrophoresed in a 4% SDS-PAGE gel for 15 min at100 V and then in a 12% SDS-PAGE gel for 45 min at 200 V.The samples were transferred to a polyvinylidene fluorid(PVDF) membrane at 300 mA for 1 h. After being blocked for1 h with 5% fat dry milk, the membrane was then incubatedwith monoclonal antibodies against a- and b-isoforms of IkB(Tebu) diluted 1:3,000 in 1% PBS-BSA. The membrane waswashed four times with PBS-Tween, 0.5% fat dry milk, andincubated with the second antibody against mouse IgG la-beled with peroxidase diluted 1:2,000 in 1% PBS-BSA. Themembrane was washed again four times, and the proteinbands were visualized by enhanced chemiluminescence (Am-ersham).For iNOS detection, 50 mg of the protein sample were

electrophoresed in a 4% SDS-PAGE stacking gel and an 8%SDS-PAGE running gel. Samples were transferred to aPVDF membrane at 300 mV for 2 h, and, after being blockedfor 1 h with 1% BSA, incubated overnight at 4°C with apolyclonal antibody against iNOS (Transduction Laborato-ries) diluted 1:500 in 1% PBS-BSA. The incubation with thesecond antibody and the revelation were performed using thesame protocol as for IkB Western blot.

Electrophoretic mobility shift assay. Nuclear proteins fromcells were prepared as previously described (26). Gel-shiftassays were performed with a commercial kit according tothe manufacturer’s instructions (Promega). The NF-kB oligo-nucleotide probe used (59-AGT TGA GGG GAC TTT CCCAGG C-39) was labeled with [g-32P]ATP by using T4 polynu-cleotide kinase. Nuclear proteins (15 mg) were incubated for20 min with the labeled probe and migrated in a 4% poly-acrylamide gel. The specificit of the binding reaction wasdetermined by coincubating duplicate samples with 100-foldmolar excess of unlabeled oligonucleotide probe (competi-tion).

Pharmacology signaling. To test the role of NF-kB inshear-induced iNOS overexpression, we incubated smoothmuscle cells with 1025 M pyrrolidine dithiocarbamate(PDTC), an NF-kB inhibitor (19), for 1 h. The cells were thensubmitted to a 20-dyn/cm2 shear stress for 24 h. To test therole of oxidative stress in the observed phenomenon, thesame experiments were performed with 1025 M of BXT-50172 (21), a potent antioxidant that mimics the gluthation-peroxidase activity.

Statistical method. Results were expressed as means 6SE. Significanc was estimated by analysis of variance andthe Bonferroni test or by the Student’s t-test. P , 0.05 wasconsidered significant

RESULTS

Cell viability and endotoxin measurement. Shear-stressed cells in primary culture for 24 h is a difficul

experimental condition that can lead to cell sufferingand death. To estimate the cell density, we assayed thelevels of total proteins in both control and treatedplates. A 24-h exposure to shear stress (20 dyn/cm2) didnot significantl change the cell density on plates sub-mitted to shear compared with the control, since theamount of total proteins was not modifie in the platestreated with L-NIL and shear stress (Fig. 1A). Further-more, cell integrity was tested by estimating both theGAPDH and S14 (2 housekeeping genes) mRNA levelsby RT-PCR. Neither shear stress nor pharmacological

Fig. 1. Shear stress did not modify cell viability. A: effect of shearstress and inducible nitric oxide synthase (iNOS) inhibition on totalprotein expression in cultured smooth muscle cells. Cells were sub-mitted to 24 h of a 20-dyn/cm2 shear stress in culture medium (shearstress) or in culture medium 1 2.1026 M of L-N6(1-iminoethyl)lysine(L-NIL), a selective inhibitor of iNOS activity (shear stress 1 L-NIL).Control cells were cultured in the same conditions without insertioninto the flo chamber. OD, optical density. B: effect of shear stressand different inhibitors on glyceraldehyde-3-phosphate dehydroge-nase (GAPDH) and S14 mRNA expression in cultured smooth musclecells. Cells were submitted to a 20-dyn/cm2 shear stress with orwithout inhibitors for 24 h. Pyrrolidine dithiocarbamate (PDTC) isan inhibitor of the nuclear factor-kB (NF-kB) system, and BXT-51072 is an antioxidant with a gluthation peroxidase mimickingactivity. The GAPDH and S14 mRNA levels were estimated byRT-PCR from 1 mg of total mRNA. cpm, Counts/min.

C1882 SHEAR STRESS INDUCES INOS EXPRESSION


on May 31, 2013


ownloaded from

06/06/2013 87


treatments influence the GAPDH nor the S14 mRNAsignal (Fig. 1B).Because iNOS expression and NF-kB activation can

be induced by LPS in smooth muscle cells, we verifiethat no LPS contamination was present in the mediumconditioned by a 20-dyn/cm2 shear stress for 24 h.Exposure of the cells to shear stress did not modify thelevel of endotoxin in the conditioned medium, com-pared with the supernatant of the control cells. Thislevel is lower than 4 EU/ml and comparable to the levelestimated in the unused commercially available cul-ture medium.

Shear stress induced the expression of iNOS. Theexpression of iNOS was dependent on the shear stressintensity (Fig. 2A) and its duration (Fig. 2B). TheiNOS/GAPDH mRNA level was 0.02 6 0.004 in controlcells. No induction was observed at 10 dyn/cm2 for 24 h(0.05 6 0.02). A shear value of 20 dyn/cm2 for 24 hinduced a significan response (3.42 6 0.95, P , 0.001).Six hours of a 20-dyn/cm2 shear rate was insufficien toinduce any iNOS expression (0.02 6 0.006). After a20-dyn/cm2 shear stress for 24 h, the level of iNOS

protein was also significantl increased in smooth mus-cle cells (1170 6 73%, P , 0.05; Fig. 2C).Because smooth muscle cells could also express the

brain isoform of the NOS, we tested the effect of shearstress on this isoform mRNA expression. A 24-h shearstress at 20 dyn/cm2 reduced the bNOS mRNA level insmooth muscle cells (from 7.8 6 1.4 in controls to 0.5 60.12, P , 0.0001; Fig. 3A).

Shear stress-induced nitrite production. We esti-mated the production of NO of the stressed smoothmuscle cells by measuring the amount of nitrite, itsmain metabolite. The nitrite production was en-hanced in the medium of smooth muscle cells sub-mitted to a 20-dyn/cm2 shear stress for 24 h. Nitriteproduction in control cells was 0.26 6 0.017 nmol/mgprotein for 4 h, whereas stressed cells synthesized0.69 6 0.09 nmol/mg protein for 4 h (P , 0.001). Thisincrease in nitrite production was inhibited by 2 31025 M L-NIL, a selective inhibitor for iNOS (0.37 60.08 nmol/mg protein for 4 h, P , 0.001), confirminthat this production is mainly due to iNOS expres-sion (Fig. 3B).

Fig. 2. iNOS expression increased in response to shearstress. iNOS mRNA expression corresponded to theratio between the iNOS and the GAPDH PCR productradioactivity. Similar results were obtained when S14mRNAs were used as a reference (data not shown). A:dependency on shear intensity of iNOS mRNA expres-sion in cultured smooth muscle cells. Cells were sub-mitted to different levels of shear stress for 24 h. ***P ,0.001. B: iNOS mRNA expression of cultured smoothmuscle cells submitted to different durations of expo-sure to shear stress (20 dyn/cm2). ***P , 0.001. C: iNOsprotein level in cultured smooth muscle cells submittedto shear stress (20 dyn/cm2 for 24 h). Western blot gel(top) and the densitometric quantificatio of the bands(bottom). *P , 0.05. NS, not significant



on May 31, 2013


ownloaded from

06/06/2013 88


Role of NF-kB and oxidative stress on shear stress-induced iNOS expression. To test the implication thatNF-kB plays a role in the shear stress-induced iNOSexpression, we incubated smooth muscle cells withPDTC, an NF-kB blocker, and an antioxidant drug.PDTC at the concentration of 1025 M completely pre-vented the induction of iNOS by 24-h shear stress at 20dyn/cm2 (0.01 6 0.003 vs. 3.4 6 0.9, P , 0.001; Fig. 4A).Furthermore, BXT-50172 (1025 M), an antioxidant

that mimics gluthation peroxidase activity, suppressedthe shear-induced iNOS induction in smooth musclecells (Fig. 4B). These results suggest that oxidativestress and NF-kB are involved in the response ofsmooth muscle cells to shear stress. We then tested theeffect of shear stress on both NF-kB translocation tothe nucleus and the expression of its IkB. In smoothmuscle cells submitted to shear stress, the amount ofNF-kB translocated to the nucleus was increased by400 6 30% (P , 0.001; Fig. 5A). The coincubation of thesample with an excess of unlabeled oligonucleotideprobe suppressed the gel shift, which showed the spec-ificit of the binding (Fig. 5A, lane competition). Fur-thermore, in the cytoplasm of the stressed cells, IkB a-and b-proteins were reduced, respectively, by 86.56 610.3% and 64.21 6 10.2% of control values (P , 0.001;Fig. 5B).

Shear stress increased the NADPH oxidase expres-sion. The enzyme implicated in the production of reac-tive oxygen species and thus in generation of oxidativestress in smooth muscle cells is the NADPH oxidase.Because our experiments with BXT-51072 have sug-gested that oxidative stress was implicated in the re-sponse to shear stress, we estimated the expression ofthe two active membrane-associated subunits of theNADPH oxidase: p22phox and Mox-1. Shear stress (20dyn/cm2, 24 h) significantl increased the p22phox andMox-1 mRNA levels in smooth muscle cells (from5.47 6 0.12 in the control cells to 7.03 6 0.38 in thestressed ones, P , 0.01 for p22phox; and from 0.331 60.04 in the control cells to 0.879 6 0.19 in the stressedones, P , 0.05 for Mox-1; Fig. 6).

DISCUSSION

The present study shows that shear stress increasedboth iNOS mRNA and protein expression in rat aorticsmooth muscle cells. This induction occurred only afterprolonged exposure (24 h) to shear stress and wasaccompanied by a rise in nitrite production in theconditioned medium. This nitrite overproduction(1160 6 26%) was more important than that which weobserved in smooth muscle cells treated for 24 h with40 mg/ml LPS (i.e., 20,000 EU/ml) plus 100 U/ml inter-feron-g (175 6 22%; data not shown), suggesting thatthis production was physiologically relevant.A preferential inhibitor of iNOS blocked this shear

stress-induced nitrite production. Because this inhibi-tor is 30-fold more specifi to iNOS than to bNOS (20),it is unlikely that the observed NO overproduction alsoresulted from the activity of the brain isoform. Fur-thermore, shear stress significantl decreased bNOSmRNA levels, confirmin that bNOS was not involvedin the observed phenomenon. The decrease in bNOSexpression was not due to cell death during the exper-iment because the amounts of total proteins and house-keeping gene expression (GAPDH and S14) were notsignificantl diminished in stressed plates. These dataconfir the observation reported by Papadaki and co-workers (25). In their study, they showed that twostages in nitrite production rate can be discerned in

Fig. 3. Role of the brain isoform of NOS (bNOS) in shear stress-induced nitrite production. bNOS mRNA expression corresponded tothe ratio between the bNOS and the GAPDH PCR product radioac-tivity. A: decrease in bNOS mRNA level in cultured smooth musclecells submitted to a 20-dyn/cm2 shear stress for 24 h. ***P , 0.001.B: effect of shear stress on nitrite production in cultured smoothmuscle cells. Cells were submitted to a 20-dyn/cm2 shear stress for24 h. To test the role of iNOS in this induction, we preincubated thecells with 2.1026 M of NIL, a selective inhibitor of iNOS activity.**P , 0.01.



on May 31, 2013


ownloaded from

06/06/2013 89


cultured smooth muscle cells exposed to flow In thefirs stage (in the firs hour of exposure to shear stress),nitrite production was quickly increased because of anupregulation of the calcium/calmodulin-sensitivebNOS activity. In the second stage (from 1 to 24 h ofexposure to high shear stress levels), shear stress in-duced a stable production of nitrites that was notsensitive to a calmodulin inhibitor (25). These datasuggest that the mechanism implicated in this NOoverproduction in the late stage (enzyme expression) isdifferent from that implicated in the firs stage (en-zyme activity). Our results confir that although theinitial burst in nitrite production in stressed cells coulddepend on the activity of constitutive NOS, a prolonged

exposure to a physiological level of shear stress inducesthe expression of the iNOS. Its activity is not directlyregulated and produces a large amount of NO (13, 24).A small quantity of protein is able to produce significant levels of NO, and this induction leads to a stableand long-term production of NO (12).This expression can be regulated by numerous extra-

cellular factors, including LPS from bacteria (10, 31)that could be present in our apparatus and then inter-fere with the effect of shear stress. The amount of LPSin the medium conditioned by a 20-dyn/cm2 shearstress for 24 h was not different from that which wasmeasured in the supernatant of the control cells andwas lower than 4 EU/ml. Furthermore, 10 dyn/cm2

Fig. 4. Role of NF-kB and oxidativestress in iNOS expression. iNOSmRNA expression corresponded to theratio between the iNOS and theGAPDH PCR product radioactivity. A:inhibition of shear stress-inducediNOS expression by an NF-kB blockerin cultured smooth muscle cells. Cellswere submitted to a 20-dyn/cm2 shearstress for 24 h in the presence or ab-sence of 1025 M PDTC. ***P , 0.001.B: inhibition of shear stress-inducediNOS expression by an antioxidant incultured smooth muscle cells. Cellswere submitted to a 20-dyn/cm2 shearstress for 24 h in the presence or ab-sence of 1025 M BXT-51072, an antiox-idant that mimics the gluthation per-oxidase activity. ***P , 0.001. **P ,0.01.

Fig. 5. Effect of shear stress on NF-kBactivation. A: the electrophoretic mobilityshift assay gel (top); the densitometricquantificatio of the NF-kB nuclear level(bottom). The competition lane corre-sponded to coincubation of the shearedsample with an excess of unlabeled oligo-nucleotide probe. B: Western blot of inhib-itors kB (IkB) on SDS-PAGE gel (top).Densitometric quantificatio of IkB levelsof cultured smooth muscle cells exposed toshear stress compared with control cells(bottom). ***P , 0.001.



on May 31, 2013


ownloaded from

06/06/2013 90


shear stress and passive transfer of shear-conditionedmedium failed to induce any change in iNOS expres-sion. Thus LPS contamination could not be responsiblefor the induction we observed.Our results clearly demonstrate that shear stress is

a potent inductor of iNOS expression in smooth musclecells. They agree with the in vivo data reported by Yanand coworkers (38) showing that iNOS expression isinduced in the innermost layers of the neointima in ratcarotid arteries after balloon-induced deendothelializa-tion and intimal smooth muscle cell migration. In re-sponse to mechanical injury, smooth muscle cells mi-grate in the intima and proliferate in contact with theblood flo and are thus submitted to shear stress (;20dyn/cm2) (38). The in vivo expression of the enzyme islocalized in the smooth muscle cells that are closest tothe lumen, suggesting that the blood flo contact couldbe necessary for the induction of iNOS expression.Furthermore, in this study, the phenomenon is de-layed, occurring only 24 h after deendothelialization.The intensity of the applied shear stress and the delaynecessary for iNOS induction in smooth muscle cellscorrespond to what we have observed in vitro and areconsistent with an important role of shear stress iniNOS overexpression in vivo. Recently, Fukuda andcoworkers (7) have demonstrated that iNOS expressionis involved in the media layer of shear-induced cerebralaneurysm in rats. Their results strongly suggest thatthe increase in wall shear stress is responsible for theiNOS induction because the reduction of shear atten-uated the iNOS immunoreactivity in the artery.The cytokine-induced iNOS expression is under the

control of the NF-kB (33). Because shear stress is ableto induce NF-kB activation in endothelial cells (14), wewondered whether shear stress-induced iNOS expres-sion in smooth muscle cells could be mediated byNF-kB activation. One part of NF-kB activation is dueto the phosphorylation of IkB by IkB kinase. IkB kinaseactivity is increased by shear stress in endothelial cells

(4) and leads to the dissociation of the cytosolic IkB-NF-kB complex that is accompanied by the degrada-tion of IkB (37). Thus the degradation of IkB is consid-ered to be a marker of NF-kB translocation. Ourresults show that shear stress caused a significandecrease in the cytosolic IkB a and b levels and anincrease in NF-kB translocation to the nucleus ofsmooth muscle cells. Moreover, PDTC, a potent inhib-itor of NF-kB translocation that acts by scavenging theintracellular reactive oxygen intermediates (30), abol-ished the iNOS induction in response to shear stress.Thus the effect of shear stress on iNOS expression was,at least in part, mediated by NF-kB.Translocated NF-kB is able to bind both the pro-

moter sequence (called shear stress responsive ele-ment) that is present in many genes in which expres-sion is modulated by shear stress (29) and an NF-kB-specifi responsive element. The rat iNOS genepromoter has been cloned (30). No shear stress respon-sive element has been found, but several functionalNF-kB responsive elements have been described andare implicated in the cytokine-induced iNOS expres-sion. Thus shear stress-induced iNOS expressionseems to be due, more in part, to the binding of NF-kBon its specifi responsive element in the iNOS genepromoter than to a direct shear-dependent response.These data fi well with the delayed character of theresponse.The intracellular signaling pathway mechanisms

that lead to NF-kB activation in response to shear arenot yet identified However, regulation of the IkB-NF-kB system is considered to be, in part, under thedependence of the redox state of the cell. Shear stressis able to increase superoxide anion production in en-dothelial cells by increasing NADPH oxidase activity(5). A similar phenomenon could occur in stressedsmooth muscle cells. It has been shown that rat aorticsmooth muscle cells use a NADH/NADPH oxidase togenerate superoxide anions (36). This enzyme is com-

Fig. 6. Effect of shear stress on p22phoxand Mox-1 mRNA levels in smoothmuscle cells. Cells were submitted to a20-dyn/cm2 shear stress for 24 h inculture medium. Control cells werecultured in the same conditions with-out insertion in the flo chamber.iNOS mRNA expression correspondedto the ratio between the iNOS andthe GAPDH PCR product. **P , 0.01.*P , 0.05.



on May 31, 2013


ownloaded from

06/06/2013 91


posed of two membrane-bound [p22phox (36) and Mox-1(34)] and three cytosolic subunits (p40, p47, and p67) insmooth muscle cells. The two membrane-associatedsubunits (p22phox and Mox-1) have been shown to beimportant for free radical production. Transfection ofsmooth muscle cells with antisense-p22phox inhibitsangiotensin II-stimulated O2

2z production, showingthat p22phox is essential in free radical synthesis inthese cells (36). Recently, Mox-1 has been described asthe active subunit of the NADH/NADPH oxidase insmooth muscle cells (34). In our experiments, 24 h of a20-dyn/cm2 shear stress increased both p22phox andMox-1 mRNA expression in smooth muscle cells. Be-cause the transcriptional regulation of these genes hasnot yet been described, the mechanism involved in thisinduction remains to be determined. These results sug-gest that oxidative stress is increased in smooth mus-cle cells submitted to shear stress. We then tested theeffect of BXT-51072, an efficien antioxidant withgluthation peroxidase mimicking activity (21), on theiNOS expression. We have recently shown that BXT-51072 inhibits oxidative stress-induced VCAM-1 ex-pression in endothelial cells (26). Because 1025 M ofBXT-51072 abolished the shear stress-induced iNOSexpression, we can conclude that oxidative stress isimplicated in the delayed response of smooth musclecells to shear stress.Our results clearly demonstrate that shear stress is

one of the inducers of iNOS expression in aortic smoothmuscle cells. This conclusion fit with numerous stud-ies showing that flo inhibits neointima formationafter angioplasty (6, 16) and that NO plays a crucialrole in this phenomenon (6). Nevertheless, as NO pro-duction appeared to be accompanied by an increase infree radical production, it is likely that a part of thesynthesized NO was converted into peroxynitrites (3).Because peroxynitrites have been shown to lead to celldisturbances, as well as to the development of athero-sclerotic lesions (27), it is difficul to know whether ornot iNOS induction in intimal smooth muscle cells isbeneficialIn conclusion, our study shows that shear stress

increases oxidative stress in smooth muscle cells andinduces NF-kB-dependent gene expression such asiNOS. Therefore, shear stress could be responsible forluminal iNOS expression in the neointima in vivo (38).Conversely, our data suggest that in the response ofsmooth muscle cells to arterial wall injury, the shearstress induced by blood flo could be an importantstimulus that follows their intimal migration and pro-liferation. Nevertheless, further experiments need tobe performed to better understand the direct or indi-rect signaling pathways that lead to the increase inoxidative stress in response to shear stress.

REFERENCES

1. Baeuerle P. IkB-NF-kB structures: at the interface of inflammation control. Cell 95: 729–731, 1998.

2. Battle T, Arnal JF, Challah M, and Michel JB. Selectiveisolation of rat aortic wall layers and their cell types in culture:application to converting enzyme activity measurement. TissueCell 26: 943–955, 1994.

3. Beckman JS, Beckman TW, Chen J, Marshall PA, andFreeman BA. Apparent hydroxyl radical production by per-oxynitrite: implications for endothelial injury from nitric oxideand superoxide. Proc Natl Acad Sci USA 87: 1620–1624, 1990.

4. Bhullar I, Li Y, Miao H, Zandi E, Kim M, Shyy JJ, andChien S. Fluid shear stress activation of IkB kinase is integrin-dependent. J Biol Chem 273: 30544–30549, 1998.

5. De Keulenaer GW, Chappell DC, Ishizaka N, Nerem RM,Alexander RW, and Griendling KK. Oscillatory and steadylaminar shear stress differentially affect human endothelial re-dox state: role of a superoxide-producing NADH oxidase. CircRes 82: 1094–1101, 1998.

6. Ellenby MI, Ernst CB, Carretero OA, and Scicli AG. Role ofnitric oxide in the effect of blood flo on neointima formation. JVasc Surg 23: 314–322, 1996.

7. Fukuda S, Hashimoto N, Naritomi H, Nagata I, Nozaki K,Kondo S, Kurino M, and Kikuchi H. Prevention of rat cere-bral aneurysm formation by inhibition of nitric oxide synthase.Circulation 101: 2532–2538, 2000.

8. Gabbiani G, Schmid E, Winter S, Chaponnier C, De Chas-tonnay C, Vandekerkhove J, Webert K, and Franck WW.Vascular smooth muscle cells differ from other smooth musclecells: predominance of vimentin filament and a specifi a-typeactin. Proc Natl Acad Sci USA 78: 298–302, 1981.

9. Gosgnach W, Challah M, Coulet F, Michel J, and Battle T.Shear stress induces angiotensin converting enzyme expressionin cultured smooth muscle cells: possible involvement of BFGF.Cardiovasc Res 45: 486–492, 2000.

10. Gross SS and Levi R. Tetrahydrobiopterin synthesis. An ab-solute requirement for cytokine-induced nitric oxide generationby vascular smooth muscle. J Biol Chem 267: 25722–25729,1992.

11. Hamon M, Vallet B, Bauters C, Wernert N, McFadden EP,Lablanche JM, Dupuis B, and Bertrand ME. Long-term oraladministration of L-arginine reduces intimal thickening andenhances neoendothelium-dependent acetylcholine-induced re-laxation after arterial injury. Circulation 90: 1357–1362, 1994.

12. Hansson G, Geng YJ, Holm Y, Hardhammar P, WennmalmA, and Jennische E. Arterial smooth muscle cells express nitricoxide synthase in response to endothelial injury. J Exp Med 180:733–738, 1994.

13. Kibbe M, Billiar T, and Tzeng E. Nitric oxide synthase genetransfer to the vessel wall. Curr Opin Nephrol Hypertens 8:75–81, 1999.

14. Lan Q, Mercurius K, and Davies P. Stimulation of transcrip-tion factors NF-kB and AP 1 in endothelial cells subjected toshear stress. Biochem Biophys Res Commun 201: 950–956, 1994.

15. Levesque MJ and Nerem RM. The elongation and orientationof cultured endothelial cells in response to shear stress. AMSEJ Biochem Eng 107: 341–347, 1985.

16. Major T, Overhiser R, and Panek R. Evidence for NO in-volvement in regulating vascular reactivity in balloon-injuredrat carotid arteries. Am J Physiol Heart Circ Physiol 269: H988–H996, 1995.

17. Malek MA and Izumo S. Control of endothelial cell geneexpression by flow J Biomech 28: 1515–1528, 1995.

18. Misko TP, Schilling RJ, Salvemini D, Moore WM, andCurrie MG. A fluorometri assay for the measurement of nitritein biological samples. Anal Biochem 214: 11–13, 1993.

19. Moellering D, McAndrew J, Jo H, and Darley-Usmar VM.Effects of pyrrolidine dithiocarbamate on endothelial cells: pro-tection against oxidative stress. Free Radic Biol Med 26: 1138–1145, 1999.

20. Moore W, Weber R, Jerome G, Tjoeng F, Misko T, andCurrie M. L-N6-(1-iminoethyl) lysine: a selective inhibitor ofinducible nitric oxide synthase. J Biol Chem 37: 3886–3888,1994.

21. Moutet M, d’Alessio P, Malette P, Devaux V, andChaudiere J. Glutathione peroxidase mimics prevent TNFa-and neutrophil-induced endothelial alterations. Free Radic BiolMed 25: 270–281, 1998.

22. Nadan R and Brown D. An improved in vitro pyrogen test todetect picograms of endotoxin contamination in intravenous



on May 31, 2013


ownloaded from

06/06/2013 92


fluid using limulus amoebocyte lysate. J Lab Clin Med 89:910–918, 1977.

23. Nadaud S, Philippe M, Arnal JF, Michel JB, and SoubrierF. Sustained increase in aortic endothelial nitric oxide synthaseexpression in vivo in a model of chronic high blood flow Circ Res79: 857–863, 1996.

24. Noris M, Morigi M, Donadelli R, Aiello S, Foppolo M,Todeschini M, Orisio S, Remuzzi G, and Remuzzi A. Nitricoxide synthesis by cultured endothelial cells is modulated byflo conditions. Circ Res 76: 536–543, 1995.

25. Papadaki M, Tilton RG, Eskin SG, and McIntire LV. Nitricoxide production by cultured human aortic smooth muscle cells:stimulation by flui flow Am J Physiol Heart Circ Physiol 274:H616–H626, 1998.

26. Pueyo M, Gonzalez W, Nicoletti A, Savoie F, Arnal J, andMichel J. Angiotensin II stimulates endothelial VCAM-1 ex-pression via NF-kB activation induced by intracellular oxidativestress. Arterioscler Thromb Vasc Biol 20: 645–651, 2000.

27. Radomski MW and Salas E. Nitric oxide—biological mediator,modulator and factor of injury: its role in the pathogenesis ofatherosclerosis. Atherosclerosis 118: 69–80, 1995.

28. Rakugi H, Jacob H, Krieger J, Ingelfinger J, and Pratt R.Vascular injury induces angiotensinogen gene expression in themedia and neointima. Circulation 87: 283–290, 1993.

29. Resnick N and Gimbrone MA Jr. Hemodynamic forces arecomplex regulators of endothelial gene expression. FASEB J 9:874–882, 1995.

30. Schrek R, Meier B, Mannel DN, Droge W, and Baeuerle A.Dithiocarbamates as potent inhibitors of nuclear factor kB acti-vation in intact cells. J Exp Med 175: 1181–1194, 1992.

31. Scott-Burden T, Schini VB, Elizando E, Junquero DC, andVanhoutte PM. Platelet-derivated growth factor suppressesand fibroblas growth factor enhances cytokine-induced produc-tion of nitric oxide by cultured smooth muscle cells. Circ Res 71:1088–1100, 1992.

32. Six I, Van Belle E, Bordet R, Corseaux D, Callebert J,Dupuis B, Bauters C, and Bertrand ME. L-Arginine andL-NAME have no effects on the reendothelialization process afterballoon injury. Cardiovasc Res 43: 731–738, 1999.

33. Spink J, Cohen J, and Evans T. The cytokine responsivevascular smooth muscle cell enhancer of inducible nitric oxidesynthase. J Biol Chem 270: 29541–29547, 1995.

34. Suh YA, Arnold RS, Lassegue B, Shi J, Xu X, Sorescu D,Chung AB, Griendling KK, and Lambeth JD. Cell transfor-mation by the superoxide-generating oxidase Mox-1. Nature 401:79–82, 1999.

35. Tronc F, Wassef M, Esposito B, Henrion D, Glagov S, andTedgui A. Role of NO in flow-induce remodeling of the rabbitcommon carotid artery. Arterioscler Thromb Vasc Biol 16: 1256–1262, 1996.

36. Ushio-Fukai M, Zafari AM, Fukui T, Ishizaka N, andGriendling KK. p22phox is a critical component of the superox-ide-generating NADH-NADPH oxidase system and regulatesangiotensin II-induced hypertrophy in vascular smooth musclecells. J Biol Chem 271: 23317–23321, 1996.

37. Verma I and Stevenson J. IkB kinase: beginning, not the end.Proc Natl Acad Sci USA 94: 11758–11760, 1997.

38. Yan ZQ, Yokota T, Zhang W, and Hansson GK. Expression ofinducible nitric oxide synthase inhibits platelet adhesion andrestores blood flo in the injured artery. Circ Res 79: 38–44,1996.



on May 31, 2013


ownloaded from

06/06/2013 93


Dr. Deborah Janowitz

Relationships Between Gray Matter, Body Maß Index, and Waist Circumference in Healthy Adults Quelle: Human Brain Mapping 000:000–000 (2012) 2012 Wiley Periodicals, Inc. Betreuer: Prof. Hans Jörgen Grabe (Psychiatrie und Psychotherapie) Worauf kommt es mir bei diesem Thema am meisten an? Bei dem Thema kommt es darauf an: In dem Paper haben Kurth et al. herausgefunden, dass die graue Substanz des Gehirns mit Übergewicht negativ korreliert (Methode: VBM). Die Untersuchungsergebnisse wurden für BMI und Hüftumfang gesehen, allerdings scheint die Messung des Hüftumfangs sensitiver zu sein. Was bedeutet mir das Thema persönlich? Das Thema bedeutet mir persönlich: ein großes aufregendes Rätsel. Wie kommt es, dass "Löcher" im Kopf entstehen, wenn man übergewichtig ist? Oder umgekehrt? Wie kann man diese Frage lösen? Was fasziniert mich selbst am Thema am meisten? An dem Paper faszinieren mich die klare Struktur, sehr gute Diskussionspunkte, eine deutliche Darstellung und eine sehr ähnliche Fragestellung- wie die, an der ich gerade arbeite. Mich fasziniert an dem Thema die multidisziplinäre Arbeit (Mathematiker, Radiologen, Epidemiologen, Neurologen, Psychiater, Neuropsychologen, -physiologen) und die klinische Relevanz- die Therapie von Demenzerkrankungen wie z. B. Alzheimer und vielen metabolischen Erkrankungen durch Überwicht sind medizinisch bedeutsam. An den Daten/Fragestellungen des Papers orientiere ich teilweise meine derzeitige Forschung. Was gefällt mir am Thema weniger? An dem Paper gefällt mir weniger, dass es sehr viele Ausschlusskriterien für die Teilnahme der Probanden gab, aber dies wird diskutiert und könnte dann an den Untersuchungen, die wir gerade durchführen ein Vorteil unserer Studie sein. Die Studie hat nur eine relativ kleine Fallzahl, aber ist eine gute Vorarbeit. Schwierig finde ich insgesamt, dass die Verteilungsmuster der Atrophien bisher noch schwer zu deuten sind. Es gibt ein "Muster", das aber noch nicht klar von der Forschung interpretiert ist.

06/06/2013 94

r Human Brain Mapping 000:000–000 (2012) r

Relationships Between Gray Matter, Body MassIndex, and Waist Circumference in Healthy Adults

Florian Kurth,1* Jennifer G. Levitt,1 Owen R. Phillips,2 Eileen Luders,2

Roger P. Woods,2 John C. Mazziotta,2 Arthur W. Toga,2

and Katherine L. Narr2

1Department of Psychiatry and Biobehavioral Sciences, UCLA School of Medicine,Los Angeles, California

2Department of Neurology, UCLA School of Medicine, Los Angeles, California

r r

Abstract: Obesity and overweight are often defined by the body mass index (BMI), which associateswith metabolic and cardiovascular disease, and possibly with dementia as well as variations in brainvolume. However, body fat distribution and abdominal obesity (as measured by waist circumference)is more strongly correlated with cardiovascular and metabolic risk than is BMI. While prior studieshave revealed negative associations between gray matter tissue volumes and BMI, the relationshipwith respect to waist circumference remains largely unexplored. We therefore investigated the effectsof both BMI and waist circumference on local gray matter volumes in a group of 115 healthy subjectsscreened to exclude physical or mental disorders that might affect the central nervous system. Resultsrevealed significant negative correlations for both BMI and waist circumference where regional graymatter effects were largest within the hypothalamus and further encompassed prefrontal, anterior tem-poral and inferior parietal cortices, and the cerebellum. However, associations were more widespreadand pronounced for waist circumference than BMI. Follow-up analyses showed that these relationshipsdiffered significantly across gender. While associations were similar for both BMI and waist circumfer-ence for males, females showed more extensive correlations for waist circumference. Our observationssuggest that waist circumference is a more sensitive indicator than BMI, particularly in females, forpotentially determining the adverse effects of obesity and overweight on the brain and associated risksto health. Hum Brain Mapp 00:000–000, 2012. VC 2012 Wiley Periodicals, Inc.

Keywords: obesity; BMI; VBM; hypothalamus; morphometry; brain; sex differences

r r

Contract grant sponsor: Human Brain Project; Contract grantnumbers: P20-MHDA52176, 5P01-EB001955; Contract grantsponsor: NIH/National Center for Research Resources; Contractgrant number: P41 RR013642; Contract grant sponsors: BrainMapping Medical Research Organization; Brain Mapping SupportFoundation; Pierson-Lovelace Foundation; The AhmansonFoundation; William M. and Linda R. Dietel Philanthropic Fundat the Northern Piedmont Community Foundation; TamkinFoundation; Jennifer Jones-Simon Foundation; Capital GroupCompanies Charitable Foundation; Robson Family and NorthstarFund.

*Correspondence to: Florian Kurth, Division of Child and Adoles-cent Psychiatry at UCLA, 760 Westwood Plaza 47-417, LosAngeles, CA 90024-1759, USA. E-mail: [email protected]

Received for publication 2 May 2011; Revised 19 September 2011;Accepted 27 November 2011

DOI: 10.1002/hbm.22021Published online in Wiley Online Library (wileyonlinelibrary.com).

VC 2012 Wiley Periodicals, Inc.

06/06/2013 95

INTRODUCTION

Overweight and obesity represent a major public healthconcern; � 1.6 billion and 400 million people, respectivelymeet criteria for these conditions worldwide [Abelson andKennedy, 2004; WHO, 2006]. Body mass index (BMI) andclearly defined BMI cut-offs [WHO, 2000] are commonlyused as the standard for measuring and diagnosing over-weight and obesity. This highly reproducible measure isshown to correlate with metabolic or cardiovascular dis-ease and other health risks [Field et al., 2001; Wang andBeydoun, 2007; Wilson et al., 2002]. However, BMI doesnot reflect fat distribution within the body and a shift inbody fat distribution toward abdominal obesity appears tobe more closely correlated with secondary adverse effectsthan BMI [Dalton et al., 2003; de Koning et al., 2007; Zhuet al., 2004]. Moreover, initial evidence suggests that waistcircumference, which estimates abdominal fat moredirectly, is increasing at greater rate than can be attribut-able to increases in BMI in US populations [Walls et al.,2010]. This is particularly salient given the apparent higherpredictive value of waist circumference for determiningsecondary adverse effects of obesity.

The condition of overweight and obesity appears toaffect most organ systems, including the central nervoussystem [Gustafson et al., 2003]. While a growing body ofliterature indicates that a higher BMI is associated withlower global or local brain volumes [Gustafson et al., 2004;Pannacciulli et al., 2006; Raji et al., 2010; Taki et al., 2008;Walther et al., 2010; Ward et al., 2005], evidence concern-ing the effects of fat distribution—e.g., as measured bywaist circumference—on brain volume is extremely sparse.Notwithstanding, one recent study showed that wholebrain volume reductions are more closely associated withvisceral adiposity than overall body weight [Debette et al.,2010]. In addition, this prior study measured temporalhorn ventricular volume and reported a trend to indicaterelationships with higher BMI/waist circumference thoughno further local measures of gray matter volume wereassessed. Thus the effects of body fat distribution on localbrain volume and potential differences with respect toBMI remain largely unknown.

In sum, the reported higher predictive value of waistcircumference for adverse health effects, proportionallylarger population increases in waist circumference relativeto BMI, and the findings by Debette et al. for overall brainvolume suggest that waist circumference may constitute amedically more meaningful variable than BMI alone. Fur-thermore, in spite of the lack of prior data concerning therelationships between waist circumference and variationsin local brain tissue volume, the above observations pro-vide leverage to hypothesize that waist circumference maybe a more relevant predictor of the effects of overweightand obesity on local brain volume. To determine and clar-ify the effects of BMI and waist circumference on regionalgray matter volume we thus analyzed the high-resolutionwhole brain images of 115 carefully screened healthy sub-

jects using partial volume tissue segmentation and highdimensional warping for optimized voxel-based mor-phometry. Further, since differential effects have beenreported for males and females as well as with respect toage, at least with respect to BMI-related influences onbrain structure [Luchsinger and Gustafson, 2009; Takiet al., 2008], follow-up analyses were additionally per-formed to examine the modulating effects of these varia-bles in the current study.

MATERIAL AND METHODS

Subjects

Participants included 115 healthy Caucasian adults (54M/61 F, see Table I) initially recruited for inclusion in theinternational consortium for brain mapping (ICBM) data-base [Mazziotta et al., 1995]. All subjects received exten-sive medical and neurological examination to minimizethe inclusion of subjects with any medical disorder thatcould possibly affect brain structure or function [Mazziottaet al., 2009]. That is, any history of medical, neurological,and psychiatric illnesses, any trauma leading to loss ofconsciousness, concussion, spinal cord or peripheral nerveinjury, and with very few exceptions (e.g., history of anti-biotics) any medication (prescription or over the counter)led to exclusion from the study. Furthermore, a clinicalexam ensured that study participants had a blood pressureof <140/90 mm Hg, a pulse of <100 and >50 (withoutirregularities), and no signs of chronic or acute systemicdisorder, as well as no abnormal neurological signs and aMini Mental score of � 28. Overweight and obesity asdefined by BMI cutoffs were not exclusion criteria. Of allsubjects, 11 qualified as obese with a BMI � 30 and 31qualified as overweight with a BMI between 25 and 30.The extensive screening procedure resulted in an ex-tremely healthy sample without neurological or psychiatricdisorders, no high blood pressure or evidence for chronicor acute diseases as assessed by a physical examination,known history of diabetes mellitus or lipid disorders, andno use of prescription, over the counter or illicit drugswith the exception of occasional use for disease prevention[for details see Mazziotta et al., 2009]. However, withrespect to obesity it is important to note that the screeningdid not include specific laboratory testing for occult diabe-tes mellitus or lipid disorder. Thus, even if no evidenceemerged from the medical history and the physical

TABLE I. Demographic data

Mean � std Range

Age (years) 45.17 � 15.45 18–80Height (m) 1.71 � 0.1 1.47–1.96Weight (kg) 73.06 � 14.76 44–137BMI (kg m�2) 25.02 � 4.13 18.18–42.37Waist circumference (m) 0.87 � 0.12 0.67–1.32

r Kurth et al. r

r 2 r

06/06/2013 96

examination, it is possible that an undiagnosed metabolicsyndrome in a given subject might have remained unrec-ognized. As part of the study procedures, weight, height,and waist circumference were measured and recorded.The BMI was calculated for each subject by dividing theweight in kilogram by the squared height in meters.

Image Acquisition

For each subject, five T1-weighted images were sequen-tially acquired on a Siemens Sonata 1.5T Scanner atUCLA Brain Mapping Center using a high-resolution T1-weighted MP-RAGE (magnetization-prepared rapid-acqui-sition gradient echo) sequence with the following parame-ters: repetition time ¼ 1,900 ms; echo time ¼ 4.38 ms; flipangle ¼ 15�; 160 contiguous 1-mm sagittal slices; field ofview ¼ 256 � 256 mm2; matrix size ¼ 256 � 256; voxelsize ¼ 1.0 � 1.0 � 1.0 mm3. All five images per subjectwere realigned and averaged to gain a higher signal tonoise ratio.

Image Preprocessing

Individual T1-weighted structural images were proc-essed with SPM8 (http://www.fil.ion.ucl.ac.uk/spm) andthe VBM8 toolbox (http://dbm.neuro.uni-jena.de/vbm.html—described in detail in http://dbm.neuro.uni-jena.de/vbm8/VBM8-Manual.pdf). In short, all imageswere bias-field corrected and segmented into gray matter(GM), white matter (WM), and cerebrospinal fluid (CSF).The segmentation algorithm accounted for partial volumeeffects [Tohka et al., 2004] using adaptive maximum a pos-teriori estimations [Rajapakse et al., 1997] and a hiddenMarkov random field model [Cuadra et al., 2005]. Usingthe diffeomorphic registration algorithm described by Ash-burner [Ashburner, 2007], the individual GM and WM seg-ments in native-space were then nonlinearly normalized tothe DARTEL-Template supplied with the VBM8 toolbox(see http://dbm.neuro.uni-jena.de). The voxel values ofthe normalized tissue segments were then multiplied(‘‘modulated’’) with the nonlinear component of the Jaco-bian determinant, which was derived from the aforemen-tioned normalization step. The resulting gray mattersegments thus preserve the local gray matter volume asidentified in native space corrected for total brain size[Buckner et al., 2004]. A quality check was performed usingtools from the SPM Toolbox and individual visual assess-ment, which yielded no artifacts or failed segmentation andnormalization of the data. Prior to statistical analysis thesegments were smoothed with an 8 � 8 � 8 mm3 FWHMkernel.

Statistical Analysis

Effects of BMI and waist circumference on local graymatter volume were tested using multiple regression anal-

yses in a general linear model framework. Because BMIand waist circumference are highly correlated (Pearson’s r¼ 0.7, P ¼ 3.6 � 10�18), they were tested in two differentmodels, one for BMI and one for waist circumference[Poline, 2007]. Age and gender were included as covari-ates. In addition, we tested for the possible modulatingeffects of age and gender by examining interactionsbetween males (n ¼ 54) and females (n ¼ 61) and age. Inthe presence of a significant interaction, correlations wereexamined within each group separately. All results werecorrected for multiple comparisons by controlling the falsediscovery rate (FDR) at q ¼ 0.05. Additionally, a clusterextent threshold was applied to account only for clustersthat exceeded the expected number of voxels per clusteras calculated according to random field theory [Worsleyet al., 1996].

RESULTS

For both BMI and waist circumference, widespread neg-ative correlations with gray matter were observedthroughout the brain. These negative correlations werelocated in the vicinity of the hypothalamus, the frontaland temporal lobes as well as the cerebellum. Significantcorrelations for waist circumference were more spatiallyexpansive than those observed for BMI (Fig. 1). In thefrontal lobe, the effects of both measures comprised theorbitofrontal cortex, frontal pole, the anterior parts of infe-rior and superior frontal gyri, as well as dorsomesial pre-frontal cortices. Within the temporal lobe, reductions inthe hippocampi and the middle and superior temporalgyri were observed. In the cerebellum both the vermis andwidespread parts of the cerebellar hemispheres wereaffected. Regardless of the measure of interest (BMI or waistcircumference), the global maximum (i.e., the smallest P-value) was located in the hypothalamus (see Figs. 2 and 3,and Tables II and III). Additional less significant regionswere also observed in the insula, the globus pallidus,and the inferior parietal lobe including the angular andsupramarginal gyri (for waist circumference only).Throughout the brain, no positive correlations betweenBMI or waist circumference and gray matter volumewere observed.

In addition, we tested for possible age and gender inter-actions to determine if these variables may influence theobserved correlations with BMI and waist circumferencein this healthy sample. While significant interactions forage were not detected in any brain region, significantinteractions for gender for both BMI and waist circumfer-ences were observed (maps not shown). Correlations werethus subsequently examined within males and femalesseparately. For BMI, follow-up analyses (Fig. 4) withinfemales revealed significant negative correlations withlocal gray matter within one cluster in the cerebellumonly. In males BMI was negatively correlated with localgray matter across more extensive and distributed brain

r Gray Matter, BMI, and Waist Circumference r

r 3 r

06/06/2013 97

Figure 1.

Negative correlations with BMI (left) and waist circumference (right) overlaid onto the semi-

transparent mean template created from the whole study population (P < 0.05, FDR corrected

for multiple comparisons). Negative correlations with waist circumference are more widespread

but comprise of similar areas as negative correlations with BMI. [Color figure can be viewed in

the online issue, which is available at wileyonlinelibrary.com.]

Figure 2.

Negative correlations with BMI overlaid onto the mean template created from the whole study popu-

lation (FDR corrected for multiple comparisons), depicted in neurological convention. In particular, pre-

frontal, hypothalamic, cerebellar, and temporal regions show a decrease of gray matter with increasing

BMI. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

r Kurth et al. r

r 4 r

06/06/2013 98

regions. These regions comprised of prefrontal, mesialfrontal and orbitofrontal, insular and temporal cortices,and also included more spatially localized hypothalamic,hippocampal, parietal, occipital, and cerebellar regions.The most pronounced effect (i.e., the region exhibiting thesmallest P-value) was located in the right mesial frontalcortex.

In contrast to BMI, for waist circumference negative cor-relations were much more expansive when examined infemales separately. Significance was observed in the hypo-thalamus, widespread cerebellar regions, and smaller pari-etal, frontal, right parahippocampal, and occipital regions,with the most pronounced effects observed in the left cere-bellum (smallest P-value). For males, negative correlationswith waist circumference were relatively similar to thoseseen for BMI, although associations became less pro-nounced within frontal regions and more pronouncedwithin the cerebellum and hypothalamus. The overallsmallest P-value was located in left mid-insula cortex. Pos-itive correlations between local gray matter volume andBMI/waist circumference were absent in both males andfemales.

DISCUSSION

In the present analysis we observed widespread reduc-tions of gray matter volume in association with BMI andwaist circumference in bilateral prefrontal cortex, anteriortemporal cortex, inferior parietal cortex and cerebellar andmidbrain hypothalamic regions. Waist circumference,which estimates abdominal fat, was associated with morespatially extensive reductions in gray matter volume thanBMI. Together with prior findings indicating that a fat dis-tribution biased toward abdominal adiposity more accu-rately predicts the risk of cardiovascular and metabolicdiseases [Dalton et al., 2003; Zhu et al., 2004], our observa-tions suggest that abdominal obesity is more closely linkedwith regional gray matter variations than is BMI and thatassociations with waist circumference may be of greatermedical relevance.

BMI has been used as the standard to measure anddefine obesity given that it is considered relatively accu-rate, reproducible, accessible, and inexpensive. These char-acteristics have thus led to the more frequent use of BMIrather than of other measures of body composition, such

Figure 3.

Negative correlations with waist circumference overlaid onto the mean template created from the

whole study population (FDR corrected for multiple comparisons) in neurological convention. The

decrease of gray matter with increasing waist circumference is more extended than with BMI, but

also most pronounced in prefrontal, hypothalamic, cerebellar, and temporopolar regions. [Color

figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]


r 5 r

06/06/2013 99

as hydrostatic weighing, bioelectrical impedance, or devi-ces such as the BOD POD, even though they may estimatethe percentage of body fat more precisely. However, whileBMI allows adjustment of body weight to height, it cannotdistinguish between tissue content and distribution. Thiscan lead to bias, as muscles have a higher specific weightthan fat. A relative increase in muscular mass can there-fore lead to an overestimation of BMI or vice versa. Still,while the aforementioned measures of body compositioncan overcome this drawback, none gives an estimate of thefat distribution toward abdominal obesity, which appearsto be more closely correlated with secondary adverse con-sequences to health than BMI [Dalton et al., 2003; de Kon-ing et al., 2007; Zhu et al., 2004].

In contrast to other measures, waist circumference yieldsan estimate of abdominal adiposity and is less sensitive tochanges in muscle mass. Our observations of more wide-

spread negative correlations for gray matter volume withwaist circumference than with BMI may thus indicate twodifferent but probably complementary conclusions: On theone hand, a shift in body fat distribution toward the abdo-men may be a critical part of the already observed associa-tion between BMI and gray matter volume. Abdominaladiposity would then likely be a more sensitive measurefor indicating possible neurological risks such as dementiaor other neurodegenerative diseases associated with obe-sity [Doherty et al., 2008; Gustafson et al., 2003]. On theother hand the measured waist circumference might beless biased by variations in tissue content, which would atleast partly explain the more extensive and pronouncedcorrelations with regional gray matter reductions. In eithercase, overall BMI appears a less sensitive measure withwhich to investigate the effects of overweight and obesityon the brain. However, it remains the defining measure to

TABLE II. Maxima of significant clusters negatively correlating with BMI, listed by cluster

Maximum(anatomic location)

Extent(mm3)

Significance MNI

P (FDR) T X Y Z

Hypothalamus 1,333 0.0174 5.1453 3 3 1.50.0225 3.5118 6 �3 �9

Left caudate head 0.0218 3.5793 �7.5 13.5 0Left superior frontal gyrus 9,400 0.0174 4.8610 �24 24 51Left mid frontal gyrus 0.0174 4.7594 �39 52.5 �10.5Right frontal pole 0.0174 4.5808 9 63 �21Right superior temporal sulcus 1,356 0.0174 4.7986 60 3 �18Right mid temporal gyrus 0.0242 3.3667 54 10.5 �34.5Right superior temporal gyrus 0.0249 3.3229 57 9 �12Right cerebellar hemisphere 10,540 0.0174 4.5589 18 �66 �30Cerebellar vermis 0.0174 4.1783 �1.5 �72 �24Left cerebellar hemisphere 0.0174 4.1176 25.5 �64.5 �36Left superior temporal sulcus 982 0.0174 4.3029 �54 9 �27Left mid temporal gyrus 0.0192 3.8222 �54 �7.5 �40.5

0.0286 3.1445 �63 �6 �21Left inferior frontal gyrus 620 0.0187 3.8795 �49.5 10.5 27Right mid temporal gyrus 234 0.0192 3.8209 55.5 �52.5 6

0.0355 2.8813 51 �57 15Left supramarginal gyrus 1,221 0.0198 3.7747 �54 �34.5 25.5

0.0212 3.6709 �45 �36 24Left superior temporal gyrus 0.0200 3.7564 �48 �40.5 16.5Left fusiform gyrus 298 0.0211 3.6894 �37.5 �43.5 �16.5

0.0389 2.7812 �37.5 �33 �24Left intraparietal sulcus 208 0.0212 3.6565 �33 �52.5 36Right supramarginal gyrus 528 0.0215 3.6067 63 �27 22.5

0.0276 3.1864 57 �25.5 30Right postcentral gyrus 0.0222 3.5254 64.5 �15 22.5Left hippocampus 250 0.0224 3.5190 �21 �33 �12Left inferior frontal gyrus 827 0.0245 3.3443 �28.5 24 �15Left anterior basal insula 0.0257 3.2829 �33 16.5 �13.5Left basal insula 0.0274 3.2014 �39 3 �15Right mid cingulate gyrus 198 0.0360 2.8654 9 �4.5 40.5

0.0372 2.8293 12 �9 46.50.0445 2.6334 9 �19.5 40.5

P ¼ corrected P-value; T ¼ T-value; X,Y,Z ¼ stereotactic coordinates in MNI space.

r Kurth et al. r

r 6 r

06/06/2013 100

differentiate weight status into ‘‘normal,’’ ‘‘overweight,’’and ‘‘obese’’ categories.

Prior studies commonly report a widespread negativecorrelation between BMI and regional brain [Ho et al.,2010; Raji et al., 2010] and gray matter volumes [Pannac-ciulli et al., 2006; Taki et al., 2008; Walther et al., 2010].This replication of results between different studies cer-tainly provides some confidence in their validity. The pres-ent findings of associations between BMI and diminishedlocal gray matter volumes match well with the observa-tions frequently reported in literature. To our knowledge,very few studies have investigated the relation betweenwaist circumference and brain volume. However, onereport confirms stronger associations between waist cir-cumference and diminished whole brain volume than BMIand diminished whole brain volume [Debette et al., 2010].Furthermore, a trend for larger temporal horn volumeswas observed, which were interpreted to reflect potentiallylinked reductions of hippocampal volume. In the presentstudy, we show that there are stronger associationsbetween local gray matter volumes and waist circumfer-ence than between BMI and local gray matter volumes—which is consistent with the findings of Debette et al.[2010]. In addition, the present study demonstrates moredirectly the negative association between waist circumfer-ence and hippocampal volume (see Table III), which wasimplicated in the prior study.

In agreement with prior studies showing differentialeffects in men and women with respect to BMI–gray mat-ter relationships [Taki et al., 2008], the present study simi-larly found associations to be modulated by gender.Specifically, our results showed relatively prominent BMIeffects within males, which were remained similar forwaist circumference. In contrast, effects of BMI in femaleswere subtle though more pronounced and regionally ex-pansive for waist circumference. These findings matchwell the results of Taki et al. [2008], who observed effectsof BMI in males, but not in females. Further, results are inline with observations focused on such relationships withrespect to cardiovascular risk. That is, in females waist cir-cumference alone explained the observed cardiovascularrisk; in males a combination of both waist circumferenceand BMI was optimal for risk prediction [Zhu et al., 2004].These results might suggest that BMI/waist circumferenceeffects on the brain may also indicate cardiovascular risk(at least in healthy subjects). This assumption, however,warrants further investigation.

Several different mechanisms may account for theobserved association between diminished gray matter andhigh BMI/waist circumference. For example, growing evi-dence points to associations between overweight and obe-sity with genetic factors [Maes et al., 1997; Meyre et al.,2009; Willer et al., 2009], while other data supports theinfluence of the signaling properties of the hormone leptin

TABLE III. Maxima of significant clusters negatively correlating with waist circumference listed by cluster

Maximum (anatomiclocation)

Extent(mm3)

Significance MNI

P (FDR) T X Y Z

Hypothalamus 76,992 0.0013 5.6049 6 �3 �6Left mid frontal gyrus 0.0013 5.4619 �24 22.5 52.5Hypothalamus 0.0013 5.3409 �4.5 �1.5 �6This cluster extends over frontal pole, orbitofrontal cortex, inferior frontal cortex, superior frontal cortex, insulae, right inferior parietal

cortex, pallidum, hippocampi, fusiform gyri, and cerebellumRight superior temporal sulcus 1,697 0.0013 5.1171 58.5 7.5 �25.5Right mid temporal gyrus 0.0046 3.6726 55.5 3 �39

0.0108 3.1697 61.5 �13.5 �33Left superior temporal sulcus 2,800 0.0013 4.8653 �60 �13.5 �15Left mid temporal gyrus 0.0015 4.6759 �63 �4.5 �24

0.0015 4.6315 �58.5 3 �24Left inferior frontal gyrus 4,207 0.0016 4.5379 �52.5 9 24Left supramarginal gyrus 0.0017 4.4851 �49.5 �34.5 31.5

0.0024 4.2287 �43.5 0 15Right cuneus 1,394 0.0017 4.5245 6 �78 36

0.0127 3.0715 6 �90 19.5Left cuneus 0.0048 3.6539 �4.5 �81 33Left angular gyrus 474 0.0034 3.8998 �39 �58.5 36Left mid occipital gyrus 430 0.0129 3.0649 �34.5 �79.5 27

0.0325 2.4737 �24 �81 18Right lingual gyrus 451 0.0205 2.7723 15 �58.5 7.5Right calcarine gyrus 0.0219 2.7279 10.5 �66 6

P ¼ corrected P-value; T ¼ T-value; X,Y,Z ¼ stereotactic coordinates in MNI space. Note that the first cluster encompasses large partsof the brain that do not show their own maxima despite being identified as significant (descriptively listed below).


r 7 r

06/06/2013 101

[Moraes et al., 2009; Scarpace and Zhang, 2009] and/orpoints to behavior as a possible mediator [Benoit et al.,2010; Cohen, 2008; Stoeckel et al., 2008]. However, none ofthese factors alone can be shown as causal. Rather, theyseem linked and to interact leading to the outcome ofoverweight and obesity [Chisholm et al., 1998].

In the present study we included well-screened healthyadults with a focus on Caucasian subjects to exclude thepossibility that variations in body composition or brainstructure amongst different racial or ethnic groups couldinfluence findings [Camhi et al., 2010]. Although we can-not completely rule out a covert diabetes mellitus or lipiddisorder that may be linked to obesity, we can exclude thepossibility that these diseases were known, medicated, orhad already led to clinical sequella that would have beenevident during history taking and physical examination. Itis therefore rather unlikely that possible confounds suchas the presence of a metabolic syndrome can explain theobserved effects.

It is important to note that the rigorous screening,although facilitating investigation of the effects of obesityon the brain independent of secondary disorders, may alsobe viewed a potential limitation. As reported previously[Mazziotta et al., 2009], only 10.7% of all subjects who con-sidered themselves healthy and willing to participate, sat-isfied the inclusion criteria. Given that obesity is linkedwith several secondary disorders [Field et al., 2001; Wangand Beydoun, 2007; Wilson et al., 2002], and patients withthese related disorders have been excluded from thisstudy, the current sample might not be strictly representa-tive of the average obese population (especially in theupper age ranges, where links between obesity and dis-ease risk become stronger).

CONCLUSION

Results from this study support that regional reductionsin gray matter are associated with obesity and point to a

Figure 4.

Negative correlations between BMI/waist circumference and local gray matter volume for male and

female groups (FDR corrected for multiple comparisons). In males, correlations with respect to

BMI/waist circumference are relatively similar. In females, BMI is negatively correlated in the left cer-

ebellum only, while waist circumference is negatively correlated with numerous additional regions.

[Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

r Kurth et al. r

r 8 r

06/06/2013 102

stronger association between abdominal obesity anddiminished gray matter than with BMI, particularly infemales. This may suggest that abdominal obesity is amore relevant risk factor for cardiovascular or metabolicdisease, and potentially also for neurodegenerative disor-ders. However, future studies may clarify the impact of fatdistribution versus the impact of tissue content on thebrain to establish a more predictive measure than BMI.Subsequent studies may also thus investigate the interac-tions between genes, hormones and behavior toward obe-sity-related changes in brain structure more effectively. Inview of the high prevalence of obesity [WHO, 2006],defined to have reached epidemic proportions, a betterunderstanding of the links with central nervous systemstructure and function could have major implications forhealthcare and prevention.

REFERENCES

Abelson P, Kennedy D (2004): The obesity epidemic. Science304:1413.

Anan F, Masaki T, Shimomura T, Fujiki M, Umeno Y, Eshima N,Saikawa T, Yoshimatsu H. (2010): Abdominal visceral fat accu-mulation is associated with hippocampus volume in nonde-mentia patients with type 2 diabetes mellitus. Neuroimage49:57–62.

Ashburner J (2007): A fast diffeomorphic image registration algo-rithm. Neuroimage 38:95–113.

Benoit SC, Davis JF, Davidson TL (2010): Learned and cognitivecontrols of food intake. Brain Res 1350:71–76.

Buckner RL, Head D, Parker J, Fotenos AF, Marcus D, Morris JC,Snyder AZ (2004): A unified approach for morphometric andfunctional data analysis in young, old, and demented adultsusing automated atlas-based head size normalization: Reliabil-ity and validation against manual measurement of total intra-cranial volume. Neuroimage 23:724–738.

Camhi SM, Bray GA, Bouchard C, Greenway FL, Johnson WD,Newton RL, Ravussin E, Ryan DH, Smith SR, Katzmarzyk PT(2011): The relationship of waist circumference and BMI to vis-ceral, subcutaneous, and total body fat: Sex and race differen-ces. Obesity (Silver Spring) 19:402–408.

Chisholm DJ, Samaras K, Markovic T, Carey D, Lapsys N, Camp-bell LV (1998): Obesity: Genes, glands or gluttony? ReprodFertil Dev 10:49–53.

Cohen DA (2008): Neurophysiological pathways to obesity: Belowawareness and beyond individual control. Diabetes 57:1768–1773.

Cuadra MB, Cammoun L, Butz T, Cuisenaire O, Thiran JP (2005):Comparison and validation of tissue modelization and statisti-cal classification methods in T1-weighted MR brain images.IEEE Trans Med Imaging 24:1548–1565.

Dalton M, Cameron AJ, Zimmet PZ, Shaw JE, Jolley D, DunstanDW, Welborn TA (2003): Waist circumference, waist-hip ratio andbody mass index and their correlation with cardiovascular diseaserisk factors in Australian adults. J Intern Med 254:555–563.

de Koning L, Merchant AT, Pogue J, Anand SS (2007): Waist cir-cumference and waist-to-hip ratio as predictors of cardiovascu-lar events: Meta-regression analysis of prospective studies. EurHeart J 28:850–856.

Debette S, Beiser A, Hoffmann U, Decarli C, O’Donnell CJ, Mas-saro JM, Au R, Himali JJ, Wolf PA, Fox CS, Seshadri S (2010):

Visceral fat is associated with lower brain volume in healthymiddle-aged adults. Ann Neurol 68:136–144.

Doherty GH, Oldreive C, Harvey J (2008): Neuroprotective actionsof leptin on central and peripheral neurons in vitro. Neuro-science 154:1297–1307.

Field AE, Coakley EH, Must A, Spadano JL, Laird N, Dietz WH,Rimm E, Colditz GA (2001): Impact of overweight on the riskof developing common chronic diseases during a 10-year pe-riod. Arch Intern Med 161:1581–1586.

Gustafson D, Rothenberg E, Blennow K, Steen B, Skoog I (2003):An 18-year follow-up of overweight and risk of Alzheimer dis-ease. Arch Intern Med 163:1524–1528.

Gustafson D, Lissner L, Bengtsson C, Bjorkelund C, Skoog I(2004): A 24-year follow-up of body mass index and cerebralatrophy. Neurology 63:1876–1881.

Ho AJ, Stein JL, Hua X, Lee S, Hibar DP, Leow AD, Dinov ID,Toga AW, Saykin AJ, Shen L, et al. (2010): A commonly car-ried allele of the obesity-related FTO gene is associated withreduced brain volume in the healthy elderly. Proc Natl AcadSci USA 107:8404–8409.

Luchsinger JA, Gustafson DR (2009): Adiposity and Alzheimer’sdisease. Curr Opin Clin Nutr Metab Care 12:15–21.

Maes HH, Neale MC, Eaves LJ (1997): Genetic and environmentalfactors in relative body weight and human adiposity. BehavGenet 27:325–351.

Mazziotta JC, Toga AW, Evans A, Fox P, Lancaster J (1995): Aprobabilistic atlas of the human brain: Theory and rationale forits development. The International Consortium for Brain Map-ping (ICBM). Neuroimage 2:89–101.

Mazziotta JC, Woods R, Iacoboni M, Sicotte N, Yaden K, Tran M,Bean C, Kaplan J, Toga AW (2009): The myth of the normal,average human brain—The ICBM experience: (1) Subjectscreening and eligibility. Neuroimage 44:914–922.

Meyre D, Delplanque J, Chevre JC, Lecoeur C, Lobbens S, GallinaS, Durand E, Vatin V, Degraeve F, Proenca C, et al. (2009): Ge-nome-wide association study for early-onset and morbid adultobesity identifies three new risk loci in European populations.Nat Genet 41:157–159.

Moraes JC, Coope A, Morari J, Cintra DE, Roman EA, Pauli JR,Romanatto T, Carvalheira JB, Oliveira AL, Saad MJ, et al.(2009): High-fat diet induces apoptosis of hypothalamic neu-rons. PLoS One 4:e5045.

Pannacciulli N, Del Parigi A, Chen K, Le DS, Reiman EM, Tatar-anni PA (2006): Brain abnormalities in human obesity: Avoxel-based morphometric study. Neuroimage 31:1419–1425.

Poline JKF, Pallier C, Penny WD (2007): Contrasts and classical in-ference. In: Friston KAJT, Kiebel SJ, Nichols TE, Penny WD,editors. Statistical Parametric Mapping. London: Elsevier. pp126–139.

Rajapakse JC, Giedd JN, Rapoport JL (1997): Statistical approachto segmentation of single-channel cerebral MR images. IEEETrans Med Imaging 16:176–186.

Raji CA, Ho AJ, Parikshak NN, Becker JT, Lopez OL, Kuller LH,Hua X, Leow AD, Toga AW, Thompson PM (2010): Brainstructure and obesity. Hum Brain Mapp 31:353–364.

Scarpace PJ, Zhang Y (2009): Leptin resistance: A prediposing fac-tor for diet-induced obesity. Am J Physiol Regul Integr CompPhysiol 296:R493–R500.

Stoeckel LE, Weller RE, Cook EW III, Twieg DB, Knowlton RC,Cox JE (2008): Widespread reward-system activation in obesewomen in response to pictures of high-calorie foods. Neuro-image 41:636–647.


r 9 r

06/06/2013 103

Taki Y, Kinomura S, Sato K, Inoue K, Goto R, Okada K, Uchida S,Kawashima R, Fukuda H (2008): Relationship between bodymass index and gray matter volume in 1,428 healthy individu-als. Obesity (Silver Spring) 16:119–124.

Tohka J, Zijdenbos A, Evans A (2004): Fast and robust parameterestimation for statistical partial volume models in brain MRI.Neuroimage 23:84–97.

Walls HL, Stevenson CE, Mannan HR, Abdullah A, Reid CM,McNeil JJ, Peeters A (2011): Comparing trends in BMI andwaist circumference. Obesity (Silver Spring) 19:216–219.

Walther K, Birdsill AC, Glisky EL, Ryan L (2010): Structural braindifferences and cognitive functioning related to body massindex in older females. Hum Brain Mapp 31:1052–1064.

Wang Y, Beydoun MA (2007): The obesity epidemic in the UnitedStates—Gender, age, socioeconomic, racial/ethnic, and geo-graphic characteristics: A systematic review and meta-regres-sion analysis. Epidemiol Rev 29:6–28.

Ward MA, Carlsson CM, Trivedi MA, Sager MA, Johnson SC(2005): The effect of body mass index on global brain volumein middle-aged adults: A cross sectional study. BMC Neurol5:23.

WHO (2000): Obesity: Preventing and managing the global epi-demic. Report of a WHO consultation. World Health OrganTech Rep Ser 894:i–xii, 1–253.

WHO.2006. Obesity and Overweight. Available at: http://www.who.int/mediacentre/factsheets/fs311/en/index.html.

Willer CJ, Speliotes EK, Loos RJ, Li S, Lindgren CM, Heid IM,Berndt SI, Elliott AL, Jackson AU, Lamina C, et al. (2009): Sixnew loci associated with body mass index highlight a neuronalinfluence on body weight regulation. Nat Genet 41:25–34.

Wilson PW, D’Agostino RB, Sullivan L, Parise H, Kannel WB(2002): Overweight and obesity as determinants of cardiovas-cular risk: The Framingham experience. Arch Intern Med162:1867–1872.

Worsley KJ, Marrett S, Neelin P, Vandal AC, Friston KJ, EvansAC (1996): A unified statistical approach for determining sig-nificant signals in images of cerebral activation. Hum BrainMapp 4:58–73.

Zhu S, Heshka S, Wang Z, Shen W, Allison DB, Ross R, Heyms-field SB (2004): Combination of BMI and waist circumferencefor identifying cardiovascular risk factors in whites. Obes Res12:633–645.

r Kurth et al. r

r 10 r

06/06/2013 104

Dr. Matthias Grothe

Effects of acute relapses on neuropsychological status in multiple sclerosis patients Quelle: J Neurol. 2011 Sep;258(9):1603-8 Betreuer: Prof. Christof Kessler (Neurologie) Worauf kommt es mir bei diesem Thema am meisten an? Der Artikel stellt zum ersten Mal systematisch dar, dass es bei einem akuten Krankheitsschub bei RRMS zu einer Verschlechterung der kognitiven Fähigkeit kommt. Was bedeutet mir das Thema persönlich? Diese neuropsychologischen Beeinträchtigungen zu erforschen, das Umfeld und die Betroffenen aber auch darüber zu informieren sowie auch über rehabilitative Ansätze zu informieren. Was fasziniert mich selbst am Thema am meisten? Systematischer Effekt des Schubes? der Erkrankung? einer generalisierten Erkrankung? das Fehlen konkreter Studien macht die Ergebnisse sehr interessant und spekulativ. Aber erstmalig wurden diese Effekte gezeigt. Was gefällt mir am Thema weniger? Die Umsetzung der Studie und die Dokumentation für die Publikation. Es bleiben sehr viele Faktoren nicht genannt, nicht dokumentiert. Insofern auch ein Beispiel auch für eine Studie mit interessanten Ergebnissen, die sehr anfechtbar sind...

06/06/2013 105

ORIGINAL COMMUNICATION

Effects of acute relapses on neuropsychological statusin multiple sclerosis patients

S. A. Morrow • S. Jurgensen • F. Forrestal •

Frederick E. Munchauer • R. H. B. Benedict

Received: 17 December 2010 / Revised: 7 February 2011 / Accepted: 21 February 2011

� Springer-Verlag 2011

Abstract Little is known about neuropsychological status

changes in multiple sclerosis (MS) patients experiencing a

relapse. The Symbol Digit Modalities Test (SDMT) and

MS Neuropsychological Screening Questionnaire (MSNQ)

are brief measures of cognitive performance and self-

reported status, respectively. We retrospectively identified

relapses in subjects participating in the 48-week open-

label, safety-extension study of natalizumab (STRATA) to

determine if changes in cognitive ability occurred during

acute relapses. SDMT and MSNQ were administered prior

to infusions. We analyzed SDMT and MSNQ scores pre-

and post-relapse in 53 MS patients with relapses (cases)

and 115 MS patients without relapses (controls) matched

on age, gender, baseline SDMT and time from study ini-

tiation. ANOVA and GLM were used to compare cases

versus controls overall, and stratified by EDSS cerebral

functional status (cFS) scores. SDMT change pre- to post-

relapse in cases was significantly lower than difference

between similar time points in controls (p = 0.003). When

comparing visit 2 (two visits pre-relapse) to visit 1 (first

visit post-relapse), MSNQ change was significantly dif-

ferent between cases and controls (p = 0.012). For

cFS B 1, the change in SDMT was significantly different

between cases and controls but not for cFS C 2. These

results confirm the involvement of cognitive function

during some MS relapses suggesting the SDMT or MSNQ

can be used to identify transitory neuropsychological status

changes and cognitive relapses.

Keywords Multiple sclerosis � Relapse � Cognition

Introduction

Multiple sclerosis (MS) most commonly presents in the

relapsing-remitting form (RRMS) of the disease, charac-

terized by episodes of subacute worsening of neurological

function due to an acute inflammatory demyelinating lesion,

separated by periods of relative clinical quiescence [1].

Cognitive impairment affects 40–65% of patients [2] and

may occur as early as the first demyelinating event [3].

Deterioration in cognitive status caused by an acute

inflammatory cerebral lesion is known to occur, yet there are

few studies documenting cognitive decline during relapses.

The Safety of Tysabri Re-dosing and Treatment

(STRATA) study was a multi-national, open-label, safety-

extension study of natalizumab, in which MS subjects had

monthly assessments associated with the natalizumab

infusion. The Symbol Digit Modalities Test (SDMT) [4]

and MS Neuropsychological Screening Questionnaire

(MSNQ) [5], brief measures of cognitive performance and

self-reported status respectively, were administered at each

assessment as a screen for worsening cognitive status. We

previously reported on the SDMT and MSNQ findings for

the patients who completed 48 weeks (n = 660).

In the present study, we retrospectively looked for evi-

dence of patients reporting a functional change or wors-

ening leading to further assessment by a neurologist.

S. A. Morrow � R. H. B. Benedict

Jacobs Neurological Institute,

State University of New York at Buffalo,

Buffalo, NY, USA

S. Jurgensen � F. Forrestal � F. E. Munchauer

Biogen Idec, Inc., Cambridge, MA, USA

R. H. B. Benedict (&)

Buffalo General Hospital, 100 High Street,

Buffalo, NY 14203, USA

e-mail: [email protected]

123

J Neurol

DOI 10.1007/s00415-011-5975-3

06/06/2013 106

Relapses were thus identified, documented and treated

clinically. The data collected for the STRATA study pro-

vided a suitable setting to compare neuropsychological

(NP) tests before and after a relapse with repeat evaluations

in a large ethnically diverse population, administered by

non-neuropsychologists across multiple clinical settings.

Thus, we endeavored to determine retrospectively if acute

relapses are accompanied by changes in measures of cog-

nition and neuropsychological status, by analyzing data

from the monthly administrations of the SDMT and MSNQ

before and after a relapse compared to subjects who did not

experience a relapse to determine if cognitive status

declines during clinical relapses.

Methods

Subjects

The data presented here are from 660 MS subjects who

completed all 48 weeks of the STRATA study, as descri-

bed in our previous publications [6, 7]. STRATA was an

open-label, multicenter, extension study to evaluate the

safety and tolerability of natalizumab following re-initia-

tion of dosing after completion of Phase III efficacy trials

(AFFIRM and SENTINEL). Participants were treated

monthly in 164 clinics in 22 countries worldwide. The total

cohort consisted of 1094 MS patients originally treated

with either placebo or natalizumab. Subjects were excluded

if the investigator considered the subject to be immuno-

compromised, had a history of persistent anti-natalizumab

antibodies, had a history of malignancy, organ transplants,

or a clinically significant infection within 30 days of the

screening visit, or had discontinued natalizumab during the

Phase III trial due to an allergic reaction or major adverse

event. Additionally, females who were breastfeeding or

pregnant were excluded. The tests were administered in 14

languages with each infusion and followed for 48 weeks.

Relapses were identified and treated at the discretion of

the subject’s own neurologist, not involved in the STRATA

study. There were 53 patients with confirmed relapses

(cases). The mean age was 44.0 ± 7.1 years, 35 (66.0%)

were female and the median Expanded Disability Status

Scale (EDSS) [8] prior to relapse of 3.0 (range 1.0–7.5).

Just over half (30, 56.6%) received corticosteroid treatment

for this relapse.

The relapsing patients were matched to 115 stable MS

patients (controls), based on time from study initiation, age

(±5 years), gender and SDMT (±3 points). Mean age was

42.3 ± 7.2 years, 75 (65.2%) were female and the median

EDSS was 2.5 (range 0.0–7.0). There was no significant

difference on EDSS categories between cases and controls

(p = 0.717), or on cerebral functional system (FS) score

(p = 0.777) (Table 1). For the analyses data in each con-

trol were taken from study visits equivalent to the pre- and

post-relapse time points in the corresponding matched case.

Tests and procedures

Both the SDMT and MSNQ were administered prior to

initiating the natalizumab infusion at each visit. The SDMT

is a measure of processing speed in which subjects had 90 s

to voice numbers as rapidly as possible that were associ-

ated with target symbols based on a grid printed at the top

of the stimulus page [4, 9]. A higher score indicates better

cognitive function. The MSNQ is a paper and pencil test

designed to screen for cognitive impairment in MS subjects

and includes 15 statements concerning cognitive problems

in daily life. Patients designated a number for each state-

ment from 0 (never, does not occur) to 4 (very often, very

disruptive) and a composite score is calculated. A higher

score indicates the subjective perception of poorer cogni-

tive function [5]. Both tests are reliable when administered

on a monthly basis [6]. SDMT instructions and MSNQ

were translated into each language if not already available.

The EDSS was administered and scored in the standard

manner, including all function system (FS) scores (visual,

brainstem, pyramidal, sensory, cerebellar, bowel/bladder,

cerebral). The cerebral functional system (cFS) score is

Table 1 Pre-relapse demographics and disability measures

Cases Controls p value

Age (years)

Mean ± SD 44.0 ± 7.1 42.3 ± 7.2 0.387*

Gender

N (%) female 35 (66.0%) 75 (65.2%) 0.917?

EDSS prior to relapse

Median (range) 3.0 (1.0–7.5) 2.5 (0.0–7.0) 0.717?

Cerebral FS prior to relapse

Median (range) 0.0 (0.0–2.0) 0.0 (0.0–2.0) 0.777?

SDMT 2 visits prior to relapse

Mean ± SD 52.3 ± 11.6 52.5 ± 11.3 0.937*

SDMT 1 visit prior to relapse

Mean ± SD 53.6 ± 10.9 53.4 ± 11.6 0.895*

MSNQ 1 visit prior to relapse

Mean ± SD 15.0 ± 11.2 13.0 ± 10.8 0.259*

MSNQ 2 visits prior to relapse

Mean ± SD 14.1 ± 10.6 13.3 ± 11.0 0.666*

Corticosteroid treatment

N (%) 37 (56.9%) N/A

SD standard deviation; EDSS expanded disability status scale; FSfunction scale; SDMT Symbol Digit Modalities Test; MSNQ Multiple

Sclerosis Neuropsychological Screening Questionnaire

* ANOVA, ? Chi-square

J Neurol

123

06/06/2013 107

based on patient reported and the rater’s perception of

fatigue, depression and decreased mentation [8]. A cFS

score of 0.0 indicates no difficulty with fatigue, depression

or decreased mentation. A score of 1 indicates a mild

degree of mood or fatigue problem, while a score of 2

indicates moderate fatigue or decrease in mentation. Scores

of 3 or 4 represent moderate and severe decreases in

mentation, respectively, while a 5 indicates frank dementia.


Univariate analysis of variance (ANOVA) was used to

compare mean differences between cases and controls and

changes within each of these groups at the pre- and post-

relapse time points. General linear models (GLM) (i.e.

mixed factor ANOVA) were employed to compare pre- and

post-relapse scores between cases and controls on the

SDMT and MSNQ. We modeled the analysis for the entire

sample and also by cFS.

Results

Data obtained prior to relapse are shown in Table 1.

Relapsing subjects had an SDMT score of 53.6 ± 10.9 at

the visit immediately preceding the relapse, while the

MNSQ pre-relapse was 15.0 ± 11.2. In controls the mean

scores at the time point equivalent to pre-relapse were

SDMT of 53.4 ± 11.6, and MSNQ 13.0 ± 10.8 which

were not significantly different from cases (p = 0.658,

p = 0.112 respectively).

Comparison of the change between the visits immedi-

ately preceding and following the relapse showed a sig-

nificant difference on the SDMT (-1.2 vs. ?1.3,

p = 0.003) in cases versus controls while the MSNQ

change exhibited a trend in the expected direction (?0.6 vs.

-0.5, p = 0.211). Comparison of the change from two

visits prior to the relapse to the first visit following the

relapse showed a significant difference in the change

between the cases and controls on the MSNQ (?1.6 vs.

-0.9, p = 0.012).

Generalized linear models (GLM) were constructed for

both the SDMT and MSNQ with scores from the two [2]

visits preceding and three [3] visits following the con-

firmed relapse compared to controls (Fig. 1). For the

SDMT, there was no between group effect (p = 0.668)

indicating the SDMT scores were not significantly different

in the cases and controls but there was a significant inter-

action (p = 0.029) indicating a significant change over

time. As shown in Fig. 1a, there is an improvement on the

SDMT score over time in both groups, with a worsening in

the relapsing group (cases) only at the assessment proximal

to the relapse.

For the MSNQ, there was no between group effect

(p = 0.221) but there was a trend towards an interaction

(p = 0.090). As shown in Fig. 1b, the MSNQ in the control

group improves steadily over time, while the relapsing

group (cases) demonstrates a worsening in the score prior

to the diagnosis of the relapse, followed by a trend towards

an improving score after the first visit post-relapse. In

cases, the MSNQ score at the second visit post-relapse

reaches the baseline score two assessments prior to the

relapse.

The distribution of cFS scores, from the EDSS per-

formed prior to the relapse, was not significant different

between cases and controls (v2df¼2 = 0.504, p = 0.777).

For cases, 73.5% (n = 39) had a cFS B 1 compared to

18.9% (n = 10) with a cFS C 2 (7.5% not reported), and

for controls 77.4% had a cFS B 1 compared to 26.5% with

a cFS [ 2 (6.1% not reported). Limiting the comparison to

cases and controls with a pre-relapse cFS C 2, there was a

trend towards significance on the SDMT (-1.2 vs. ?1.9,

p = 0.085) and the change on the MSNQ was not

Fig. 1 SDMT (a) and MSNQ (b) scores before (pre) and after (post)

the documented relapse in the cases (n = 53) and at equivalent time

points in controls (n = 115). a SMDT Symbol Digit Modalities

Test, b MSNQ Multiple Sclerosis Neuropsychological Screening

Questionnaire

J Neurol

123

06/06/2013 108

significantly different (?1.5 vs. ?0.1, p = 0.568) in cases

compared to controls over the relapse period. Similarly,

when limiting the analysis to cFS B 1, a significant dif-

ference was found on the SDMT (-1.4 vs. ?1.2,

p = 0.012) but not on the MSNQ (?1.1 vs. -0.6,

p = 0.108). Again, if the change on the MSNQ was

examined from two visits prior to the relapse in those with

a cFS B 1, there was a significant difference between cases

and controls (?2.2 vs. -1.2, p = 0.001).

Discussion

Cognition has always been suspected to be vulnerable to

the effect of a new inflammatory lesion; this is the first

study, to our knowledge, to examine changes in cognitive

ability before, during and after an acute relapse. We have

shown worsening on a performance measure of cognitive

processing speed and attention as well as a self-reported

evaluation of cognitive function during an acute relapse

that resolves over time, presumably as the inflammation

remits and the relapse resolves.

Cognitive dysfunction is clinically relevant in the MS

population, being a primary predictor of employability and

health related quality of life [10–12]. ‘‘Cognitive relapses’’

are known to occur, yet there are few studies documenting

the identification and the consequences of this type of

relapse or the expected recovery. Studies addressing other

primary endpoints have found improvement on mean

Paced Auditory Serial Addition Test (PASAT) [13] scores,

the cognitive component of the Multiple Sclerosis Func-

tional Composite (MSFC), of 2.5–9 points as MS subjects

recover from a relapse [14, 15]. Yet, these studies were

addressing the MSFC as a whole and did not have a

comparison group to control for practice effects, which are

well known to be significant and occur with repeated

administrations of the PASAT [16, 17]. Only two studies

specifically addressed whether other cognitive domains

were affected during an acute relapse by comparing the

scores on NP tests in MS subjects during a documented

relapse, in one study, or 5 weeks after an episode of optic

neuritis in the other, to normal controls; expectedly, deficits

were found in processing speed and memory [18, 19].

However, whether the MS subjects were cognitively

impaired due to the relapse directly or due to chronic

effects of MS was not addressed, as NP scores before or

after the relapse were not available. In our study we found

that MS patients experiencing a relapse worsened on the

SDMT compared to matched MS controls not experiencing

an acute relapse, indicating that cognition can be affected

during an acute relapse.

Interestingly, change in cognitive function was not

necessarily the primary symptom or sign leading to the

diagnosis of an acute relapse in the subjects included in this

study. All subjects were identified as experiencing a relapse

by his/her primary neurologist, based on neurological signs

or symptoms and changes on the EDSS score. Thus, the

changes noted in this study on the SDMT and MSNQ were

in addition to other domains involved in the relapse pre-

sentation. The fact that cognition was not necessarily the

primary complaint or problem for the MS patient may

explain the relatively small difference (-1.0 vs. ?1.2 in

cases and controls respectively) in SDMT scores during the

relapse. Focusing only on those MS subjects with cognitive

changes may lead to a better measure of the magnitude of

change and recovery expected during a ‘‘cognitive

relapse’’. Such a study is underway at our center.

The SDMT and MSNQ were used in this study to

monitor for changes in cognition that might identify the

onset of Progressive Multifocal Leukoencephalopathy

(PML). No cases of PML occurred during the 48 weeks

when these assessments were obtained in the STRATA

study. However, the use of these tests was also ideal to

identify deterioration in cognitive function during relapses.

The SDMT is a measure of information processing speed

and attention which, along with working memory, is the

cognitive domain most commonly affected in MS. It has

been shown to be a sensitive test for both screening and

diagnosis of cognitive impairment in MS [20–22]. Indeed,

it was found to be the most sensitive test in both the

Minimal Assessment of Cognitive Function In MS battery

(MACFIMS) and Rao’s Brief Repeatable Battery of Neu-

ropsychological Tests (BRB) [23]. Furthermore, recent

evidence supports the SDMT as an alternative to the PA-

SAT in the MSFC [24, 25] and both the SDMT and the

MSNQ have been found to be reliable and valid when

administered monthly [6, 7].

Another interesting observation in this study is the

worsening on the subjective rating scale regarding cogni-

tive difficulties, the MSNQ, apparently prior to the onset of

the relapse. Although there is currently no literature

addressing this issue, many MS patients report a prodrome

of non-specific symptoms prior to the onset of physical

symptoms associated with an acute relapse. Further docu-

mentation of this phenomenon is needed.

The fact that worsening of cognitive function can be

documented during a relapse may contribute to defining a

clinically meaningful change in cognition in MS. In con-

trast to measures of physical function, the magnitude of

change on objective tests of cognitive function that have

clinical relevance is not yet well defined. We recently used

a change in vocational status to explore the definition of a

clinically meaningful change, as work disability is known

to have an impact on quality of life and self-image in MS

patients [10, 26]. A decrease on the SDMT of 4.1 occurred

in those who were employed at the first assessment but

J Neurol

123

06/06/2013 109

unemployed at the second, while those who did not have a

change in vocational status remained stable [27]. Thus, a

decrease (worsening) on the SDMT of 4.0 could be con-

sidered clinically meaningful. However, a change in

vocational status is a dramatic indication that there has

been a worsening in an MS patient’s functional status, and

thus may not account for more subtle, but significant and

meaningful, worsening. Quantifying the change on objec-

tive tests during an acute demyelinating event, specifically

when there are cognitive complaints as part of the clinical

presentation, could lead to a better understanding of clin-

ically meaningful change.

Certainly, this study has limitations. Our analysis was

retrospective in nature, and thus many aspects could not be

controlled, such as the variable time between the onset of

the relapse and the administrations of the SDMT and

MSNQ, although the monthly assessments in this study

provided for a relatively small window. Additionally, the

effect of corticosteroids on cognitive tests has not been

fully elucidated. Two studies in MS patients found a

worsening in declarative memory with no effect on short

term memory, learning or attention [28, 29]. However,

another study found worsening on tests of both immediate

and delayed recall on memory tests. All three studies found

these deficits to be reversible before re-assessment in

90 days. None of the studies addressed processing speed

with tests such as the SDMT; furthermore, corticosteroids

are known to cause mood changes, and thus recent

administration of corticosteroids may affect the MSNQ,

which is known to be confounded by depression [5, 30],

although recovery of MSNQ post-relapse appeared to be

evident in our cohort, most of whom were treated with

steroids. Additionally, all subjects in this analysis were

treated with natalizumab; it is not known what, if any,

effect natalizumab has on the severity of relapses compared

with other disease modifying medications. Further, there is

evidence suggesting disease modifying therapies for MS,

including natalizumab, may improve cognitive function

over time [31]. Another weakness of this study is the small

sample size, especially for the cFS subgroup analysis and

the lack information regarding the symptoms and signs

leading to the identification of the relapse. Finally, it is

possible that there were lingering effects of the relapse at

the first assessment post-relapse, confounding our ability to

measure the recovery, including a change in visual function

which could affect the accuracy of our test scores.

Despite these limitations, this study demonstrates that

cognitive function can be involved in relapses. Further

studies need to be done to better refine the approach to the

identification and diagnosis of cognitive relapses and the

tools needed to measure the change that can occur in

cognitive functioning and the expected recovery, with

corticosteroid treatment.

Acknowledgments The authors would like to thank the STRATA

patients, investigators and steering committee for their participation

in this study.

Conflict of interest None.

References

1. Lublin FD, Reingold SC (1996) Defining the clinical course of

multiple sclerosis: results of an international survey. National

Multiple Sclerosis Society (USA) Advisory Committee on Clin-

ical Trials of New Agents in Multiple Sclerosis. Neurology

46:907–911 (see comment)

2. Rao SM, Leo GJ, Bernardin L, Unverzagt F (1991) Cognitive

dysfunction in multiple sclerosis. I. Frequency, patterns, and

prediction. Neurology 41:685–691

3. Feuillet L, Reuter F, Audoin B et al (2007) Early cognitive

impairment in patients with clinically isolated syndrome sug-

gestive of multiple sclerosis. Mult Scler 13:124–127

4. Smith A (1982) Symbol Digit Modalities Test. Manual. Western

Psychological Services, Los Angeles

5. Benedict RHB, Cox D, Thompson LL, Foley FW, Weinstock-

Guttman B, Munschauer F (2004) Reliable screening for neuro-

psychological impairment in MS. Mult Scler 10:675–678

6. Benedict RH, Duquin JA, Jurgensen S et al (2008) Repeated

assessment of neuropsychological deficits in multiple sclerosis

using the Symbol Digit Modalities Test and the MS Neuropsy-

chological Screening Questionnaire. Mult Scler 14:940–946

7. Morrow SA, O’Connor PW, Polman CH, Goodman AD, Kappos

L, Lublin FD, Rudick RA, Jurgensen S, Paes D, Forrestal F,

Benedict RHB (2010) Evaluation of the Symbol Digit Modalities

Test (SDMT) and MS Neuropsychological Screening Question-

naire (MSNQ) in Natalizumab treated MS patients over

48 weeks. Mult Scler 16(11):1385–1392

8. Kurtzke JF (1983) Rating neurologic impairment in multiple

sclerosis: an expanded disability status scale (EDSS). Neurology

33:1444–1452

9. Rao SM (1990) A manual for the Brief Repeatable Battery of

Neuropsychological Tests in Multiple Sclerosis. Medical College

of Wisconscin, Milwaukee

10. Patti F, Pozzilli C, Montanari E et al (2007) Effects of education

level and employment status on HRQoL in early relapsing-

remitting multiple sclerosis. Mult Scler 13:783–791

11. Amato MP, Ponziani G, Rossi F, Liedl CL, Stefanile C, Rossi L

(2001) Quality of life in multiple sclerosis: the impact of

depression, fatigue and disability. Mult Scler 7:340–344

12. Benedict RHB, Wahlig E, Bakshi R et al (2005) Predicting

quality of life in multiple sclerosis: accounting for physical dis-

ability, fatigue, cognition, mood disorder, personality, and

behavior change. J Neurol Sci 231:29–34

13. Gronwall DM (1977) Paced auditory serial-addition task: a

measure of recovery from concussion. Percept Mot Skills

44:367–373

14. Ozakbas S, Cagiran I, Ormeci B, Idiman E (2004) Correlations

between multiple sclerosis functional composite, expanded dis-

ability status scale and health-related quality of life during and

after treatment of relapses in patients with multiple sclerosis.

J Neurol Sci 218:3–7

15. Patzold T, Schwengelbeck M, Ossege LM, Malin JP, Sindern E

(2002) Changes of the MS functional composite and EDSS

during and after treatment of relapses with methylprednisolone

in patients with multiple sclerosis. Acta Neurol Scand

105:164–168

J Neurol

123

06/06/2013 110

16. Barker-Collo SL (2005) Within session practice effects on the

PASAT in clients with multiple sclerosis. Arch Clin Neuropsy-

chol 20:145–152

17. Bever C, Grattan L, Panitch HS, Johnson KP (1995) The brief

repeatable battery of neuropsychological tests for multiple scle-

rosis: a preliminary serial study. Mult Scler 1:165–169

18. Foong J, Rozewicz L, Quaghebeur G, Thompson AJ, Miller

DH, Ron MA (1998) Neuropsychological deficits in multiple

sclerosis after acute relapse. J Neurol Neurosurg Psychiatry

64:529–532

19. Feinstein A, Youl B, Ron M (1992) Acute optic neuritis. A

cognitive and magnetic resonance imaging study. Brain 115(Pt

5):1403–1415

20. Deloire MS, Bonnet MC, Salort E et al (2006) How to detect

cognitive dysfunction at early stages of multiple sclerosis? Mult

Scler 12:445–452

21. Parmenter BA, Weinstock-Guttman B, Garg N, Munschauer F,

Benedict RH (2007) Screening for cognitive impairment in

multiple sclerosis using the Symbol Digit Modalities Test. Mult

Scler 13:52–57

22. Benedict RHB, Cookfair D, Gavett R et al (2006) Validity of the

Minimal Assessment of Cognitive Function in Multiple Sclerosis

(MACFIMS). J Int Neuropsychol Soc 12:549–558

23. Strober L, Englert J, Munschauer F, Weinstock-Guttman B, Rao

S, Benedict R (2009) Sensitivity of conventional memory tests in

multiple sclerosis: comparing the Rao Brief Repeatable Neuro-

psychological Battery and the Minimal Assessment of Cognitive

Function in MS. Mult Scler 15:1077–1084

24. Brochet B, Deloire MS, Bonnet M et al (2008) Should SDMT

substitute for PASAT in MSFC? A 5-year longitudinal study.

Mult Scler 14:1242–1249

25. Drake AS, Weinstock-Guttman B, Morrow SA, Hojnacki D,

Munschauer FE, Benedict RH (2010) Psychometrics and nor-

mative data for the multiple sclerosis functional composite:

replacing the PASAT with the Symbol Digit Modalities Test.

Mult Scler 16:228–237

26. Smith MM, Arnett PA (2005) Factors related to employment

status changes in individuals with multiple sclerosis. Mult Scler

11:602–609

27. Morrow SA, Drake A, Zivadinov R, Munschauer F, Weinstock-

Guttman B, Benedict RHB (2010) Predicting loss of employment

over three years in multiple sclerosis: clinically meaningful cog-

nitive decline. The clinical neuropsychologist 24(7):1131–1145

28. Uttner I, Muller S, Zinser C et al (2005) Reversible impaired

memory induced by pulsed methylprednisolone in patients with

MS. Neurology 64:1971–1973

29. Brunner R, Schaefer D, Hess K, Parzer P, Resch F, Schwab S

(2005) Effect of corticosteroids on short-term and long-term

memory. Neurology 64:335–337

30. Benedict RHB, Munschauer FE, Linn R, Miller C, Foley FW,

Jacobs LD (2003) Screening for multiple sclerosis cognitive

impairment using a self-administered 15-item questionnaire. Mult

Scler 9:95–101

31. Lyros E, Messinis L, Papageorgiou SG, Papathanasopoulos P

(2010) Cognitive dysfunction in multiple sclerosis: the effect of

pharmacological interventions. Int Rev Psychiatry;22:35–42

J Neurol

123

06/06/2013 111

Ulrike Thorack

Inhibition of endogenous reverse transcriptase antagonizes human tumor growth Quelle: Oncogene (2005) 24, 3923–3931 Betreuer: Prof. Hansjörg Schwertz (Immunologie & Transfusionsmedizin) Was bedeutet mir das Thema persönlich? Die hier angesprochenen Grundlagen sind die Basis meines Promotionsthemas. LINE-1 ist auch in Thrombozyten vorhanden. Über die dortige Funktion ist allerdings wenig bekannt. Das Paper beleuchtet die Auswirkungen von LINE-1 Produkten in Bezug auf Krebszellen, was sehr spannend, auch für meine eigene Arbeit, ist. LINE-1 war für mich, vor der Doktorarbeit, völlig unbekannt. In diesem Paper wird ein sehr guter Überblick über dessen Wirkungsausmaße gegeben. Worauf kommt es mir bei diesem Thema am meisten an? Mir kommt es am meisten darauf an, dass in diesem Text erste Grundlagen des LINE-1 erläutert werden, welches auch in Thrombozyten eine Rolle spielt. Was fasziniert mich an dem Thema am meisten? Beeindruckend finde ich, dass in diesem Paper gezeigt wird, dass LINE-1 eine Rolle bei der Krebsentstehung spielt. Die Proteine, welche daraus translatiert werden, steigern die Proliferation von Tumorzellen. Eine Hemmung, durch reverse Transkriptase Hemmstoffe, kann diese starke Zellteilung herunterregulieren und zu einer Differenzierung führen. Dies ist ein wichtiger Aspekt und vielleicht eine Möglichkeit der Krebstherapie. Was gefällt mir am Thema weniger? Ich persönlich hätte mir nähere Erläuterungen zur klinischen Umsetzbarkeit der aufgeführten Ergebnisse gewünscht.

06/06/2013 112

Inhibition of endogenous reverse transcriptase antagonizes human

tumor growth

Ilaria Sciamanna1,7, Matteo Landriscina2,7,8, Carmine Pittoggi1, Michela Quirino2, CristinaMearelli1, Rosanna Beraldi3, Elisabetta Mattei4, Annalucia Serafino5, Alessandra Cassano2,Paola Sinibaldi-Vallebona6, Enrico Garaci6, Carlo Barone2 and Corrado Spadafora*,1

1Istituto Superiore di Sanita, Rome, Italy; 2Medical Oncology Unit, Catholic University, Rome, Italy; 3Department of Pediatrics,Obstetrics and Reproductive Medicine, University of Siena, Italy; 4CNR Institute of Molecular Biology and Pathology, Rome, Italy;5CNR Institute of Neurobiology and Molecular Medicine, Rome, Italy; 6Department of Experimental Medicine and BiochemicalSciences, University ‘Tor Vergata’, Rome, Italy

Undifferentiated cells and embryos express high levels ofendogenous non-telomerase reverse transcriptase (RT) ofretroposon/retroviral origin. We previously found that RTinhibitors modulate cell growth and differentiation inseveral cell lines. We have now sought to establish whetherhigh levels of RT activity are directly linked to celltransformation. To address this possibility, we haveemployed two different approaches to inhibit RT activityin melanoma and prostate carcinoma cell lines: pharma-cological inhibition by two characterized RT inhibitors,nevirapine and efavirenz, and downregulation of expres-sion of RT-encoding LINE-1 elements by RNA inter-ference (RNAi). Both treatments reduced proliferation,induced morphological differentiation and reprogrammedgene expression. These features are reversible upondiscontinuation of the anti-RT treatment, suggesting thatRT contributes to an epigenetic level of control. Mostimportantly, inhibition of RT activity in vivo antagonizedtumor growth in animal experiments. Moreover, pretreat-ment with RT inhibitors attenuated the tumorigenicphenotype of prostate carcinoma cells inoculated in nudemice. Based on these data, the endogenous RT can beregarded as an epigenetic regulator of cell differentiationand proliferation and may represent a novel target incancer therapy.Oncogene (2005) 24, 3923–3931. doi:10.1038/sj.onc.1208562Published online 4 April 2005

Keywords: endogenous reverse transcriptase; proliferation;differentiation; tumor growth; RNAi; anticancer therapy

Introduction

A striking finding emerging from the recent sequencingof the human genome is that retrotransposable elements,such as long interspersed elements (LINEs), Alu andendogenous retroviruses (ERVs) make up some 45% ofhuman DNA (Deininger et al., 2003). All classes ofretroelements, but the Alu family, are endowed with areverse transcriptase (RT)-coding gene, which enablesthem to retrotranspose autonomously. The presence andfunction of retroelements in the genome has longpuzzled biologists. The lack of any obvious cellularfunction initially suggested that these elements, and theRT-coding sequences harbored therein, were mereevolutionary remnants and inspired the concept of ‘junkDNA’. More recently, the realization that retrotrans-posons can reshape the genome and contribute tomodulation of gene expression has led to reconsiderthat hypothesis. Growing evidence indicate that RT-coding genes are expressed at low levels, if at all, indifferentiated nonpathological tissues; in contrast, highexpression is distinctive of germ cells (Kiessling et al.,1989; Giordano et al., 2000), embryos (Poznanski andCalarco, 1991; Packer et al., 1993), embryonic tissues(Mwenda, 1993), and undifferentiated and transformedcells (Deragon et al., 1990; Martin, 1991; Martin andBranciforte, 1993), suggesting that levels of RT expres-sion are linked to the proliferative potential of thecell. In addition, RT gene activity is upregulated bya variety of stimuli acting at the genome-wide level,for example, cellular stress (Hagan et al., 2003),heat shock, cycloheximide, adenovirus infection (Liand Schmid, 2001), genotoxic agents (Hagan and Rudin,2002), and DNA base analogs (Khan et al., 2001).Unscheduled activity of retrotransposons and ERVs isimplicated in a variety of diseases, including cancer(Friedlander and Patarca, 1999; Ostertag and Kazazian,2001). Conversely, inactivation of specific RT-encodingelements using antisense oliogonucleotides or ribozymesinhibited proliferation of human (Kuo et al., 1998) andmurine (Crone et al., 1999) cell lines. The questionremains unanswered as to whether retroelements are to

Received 15 October 2004; revised 21 December 2004; accepted 27January 2005; published online 4 April 2005

*Correspondence: C Spadafora, Biology and Animal Sciences, IstitutoSuperiore di Sanita (Italian National Health Insitute), Viale ReginaElena 299, Via del Castro Laurenziano 25, Rome 00161, Italy;E-mail: [email protected] authors contributed equally to this work8Current address: Clinical Oncology, Department of Internal Medi-cine, University of Foggia, Italy

Oncogene (2005) 24, 3923–3931& 2005 Nature Publishing Group All rights reserved 0950-9232/05 $30.00

www.nature.com/onc

06/06/2013 113

be regarded as endmarkers, or causative triggers, inprocesses associated with cell proliferation and function.We showed previously that non-nucleosidic RT

inhibitors directed against the HIV-encoded RT in factalso inhibit the endogenous RT activity present in earlyembryos (Pittoggi et al., 2003) and undifferentiated cells(Mangiacasale et al., 2003). These RT inhibitorsdisplayed powerful effects in nondifferentiated, ordedifferentiated cells: they inhibited cleavage in murineearly embryos, yielding developmental arrest before theblastocyst stage (Pittoggi et al., 2003), and, moreover,reduced proliferation and facilitated the onset of differ-entiation in murine and human cell lines (Mangiacasaleet al., 2003). These data suggest that the endogenouscellular RT plays a physiological role in cell prolifera-tion and differentiation. Furthermore, there appears tobe a requirement for regulated RT activity in differ-entiated cells. Here we have taken a step further andhave asked whether the endogenous RT plays a directrole in proliferation and differentiation of transformedcells, and, if so, whether modulating RT activity incancer cells might represent a novel approach to inhibittumor growth.

Results

RT inhibitors reversibly reduce cell proliferation inhuman transformed cell lines

In previous work, we reported that the RT inhibitornevirapine, widely used in anti-HIV therapy, blocks theenzymatic activity of endogenous RTs in noninfectedproliferating cells, as revealed using a highly sensitiveRT–PCR based in vitro assay (Pyra et al., 1994), andconcomitantly reduces the rate of proliferation (Man-giacasale et al., 2003). Here we set out to investigate theresponse of human transformed cell lines to prolongedexposure to RT inhibitors. Two well-characterized RTinhibitors, that is, nevirapine and efavirenz, were used.Cells from A-375 melanoma, PC3 prostate carcinomaand TVM-A12 primary melanoma-derived lines werepassaged, counted and replated every 96 h with con-tinuous drug readdition (or DMSO alone in controlcultures) for at least five 96 h-cycles. As shown inFigure 1a, both inhibitors effectively reduce cell growthin all cell lines, with a stable effect during prolongedexposure. Growth inhibition was reversible: when RTinhibitors were removed, proliferation was resumed at acomparable rate to controls within one or two 96 h-cycles. Readdition of the drugs inhibited again prolif-eration in all cell lines. Thus, the reduction of cellgrowth associated with RT inhibition is not inherited asa permanent change through cell division.We next asked whether either RT inhibitor induced

cell death in A-375 or PC3 cell lines. Combined stainingwith PI to reveal permeable necrotic cells, DAPI tovisualize apoptotic nuclei, and DiOC6(3) to monitor theloss of mitochondrial transmembrane potential, re-vealed no significant induction of cell death by eitherRT inhibitor; what low ratio was recorded (15% at most

after 72 h of exposure to either drug) was largelyaccounted for by apoptosis (data not shown). Thus,neither drug has significant nonspecific toxicity. We next

drug

b

cell

cycl

e ph

ase

0

1020304050607080

NEV EFV EFVDMSO

A-375

removal

removal

0

10

20

30

40

50

DMSO

PC3

cell

cycl

e ph

ase

re-addition removal

removal

020406080

100120

re-addition

PC3ce

ll c

ount

sce

ll c

ount

sce

ll c

ount

s

a

drug 020406080

100120

A-375

drug 020406080

100120140160 TVM-A12

re-addition removal

DMSONEVEFV

140

NEV

NEV NEV EFVEFV

G0/G1

S

G2/M

- + + - + + - + + - + + - + + - - - - - - - + + - + +

- + +

- + +

- + + - + + - + + - + + - - - - - - - + + - + +

- + + - + + - + + - - - - - - -++ -++ -++

Figure 1 Inhibition of proliferation by anti-RT drugs. (a) Cellgrowth in cultures treated with DMSO (control), nevirapine (NEV)and efavirenz (EFV). Cells were counted and re-plated every 96 hfor five cycles. Cells were then cultured in inhibitor-free medium(two cycles). RT inhibitors were then re-added for two cycles.Counted cells are expressed as the % of controls, taken as 100.Values represent pooled data from three experiments. (b) Cell cycleprofiles in the presence of RT inhibitors for four 96 h-cycles andafter drug removal

Endogenous reverse transcriptase in tumor growthI Sciamanna et al

3924

Oncogene 06/06/2013 114

sought to establish whether reduced cell growth ratherreflected the induction of cell cycle delay. BiparametricFACS analysis was employed to measure the DNAcontent (revealed by PI) and DNA replication (by BrdUincorporation) after four 96 h-cycles of exposure to RTinhibitors. This depicted significantly altered cell cycleprofile in anti-RT treated cultures, with an increasedproportion of G0/G1 BrdU-negative cells, that wasespecially pronounced in A-375 cell cultures (Figure 1b).Removal of the drugs re-established the original cellcycle profile and abolished the G1 delay.

Nevirapine induces morphological differentiation andmodulates gene expression in transformed cell lines

Melanomas are resistant to most therapeutic treatments:thus, it was relevant to determine whether RT inhibitorsinduced differentiation concomitant with reduced cellgrowth. We first examined A-375 melanoma cells, whichacquire a typical dendritic-like phenotype in response tocertain inducers of differentiation (Sauane et al., 2003).As shown in Figure 2A, morphological differentiation,revealed by cell shape, dendritic-like extensions and

increased adhesion, became evident within 4–5 daysof exposure to nevirapine (d) or efavirenz (g), comparedto DMSO-treated controls (a). By scanning electronmicroscopy (SEM), A-375 cells cultured with nevirapine(e) and efavirenz (h) become flattened compared tountreated controls (b) and exhibit elongated dendriteextensions that adhere tightly to the substrate. Confocalmicroscopy after a-tubulin immunofluorescence alsorevealed the reorganization of microtubule arraysthroughout the length of outgrowing dendrites in RT-inhibited cells (f, i), different from controls (c), in whichshort microtubules concentrate around the nucleatingcenters. Nevirapine treatment induced similar changesin TVM-A12 primary cells derived from melanoma(Figure 2B): untreated cells have a spindle-shapedmorphology by phase contrast (a) and SEM (b);nevirapine-treated TVM-A12 cells formed instead typi-cal branched dendrites (d, e) and displayed well-organized, elongated microtubule arrays (f), comparedto untreated cells (c). Significant morphological changeswere also induced in PC3 prostate carcinoma cells uponexposure to nevirapine (Figure 3c and d) and efavirenz(e, f) compared to controls (a, b). The microtubulenetwork was reorganized, with the appearance offusiform extensions protruding from the cell periphery,particularly in response to nevirapine.The induction of morphological differentiation sug-

gests that critical regulatory genes are modulated inresponse to the RT inhibitory treatment. This wasinvestigated in semiquantitative RT–PCR analysis of

DM

SOD

MSO

Nev

irap

ine

Efa

vire

nz

Aa b c

d e f

g h i

a b c

d e f

B

Nev

irap

ine

Figure 2 Morphological differentiation of melanoma cells in thepresence of RT inhibitors. (A) A-375 cell line cultured in DMSO-(a–c), nevirapine- (d–f) or efavirenz- (g–i). Cultures were observedby phase-contrast microscopy after Wright Giemsa staining (leftcolumn), SEM (middle column) and confocal microscopy (rightcolumn) after a-tubulin (green) and PI staining of nuclei (red). (B)Primary melanoma-derived TVM-A12 cells. DMSO- (a–c) andnevirapine-treated (d–f) cells under phase contrast (left column),SEM (middle column), and confocal microscopy (right column).Bar, 20 mm

Figure 3 Morphological differentiation of PC3 prostate carci-noma cells by RT inhibitors. DMSO- (a, b), nevirapine- (c, d) andefavirenz- (e, f) treated PC3 cells under phase-contrast microscopy(a, c, e), and fluorescence microscopy (b, d, f) after a-tubulin (green)and DAPI staining of nuclei (blue). Bars, 10 mm


3925

Oncogene06/06/2013 115

cultures treated with DMSO only, or nevirapine orefavirenz for four cycles. In A-375 melanoma cells, wefocussed on a set of four genes: the E-cadherin gene,involved in cell–cell adhesion and expressed in differen-tiated but not in tumor cells (Hsu et al., 2000); and thec-myc, bcl-2 (Utikal et al., 2002) and cyclin D1 (Sauteret al., 2002) genes, which are directly implicated inmelanoma cell proliferation and tumor growth. Resultsin Figure 4a indicate that the E-cadherin gene ismarkedly upregulated, whereas c-myc, bcl-2 and ccnd1genes are downregulated, in RT-inhibited A-375 cul-tures compared to controls. One exception was recordedfor efavirenz, which failed to downregulate ccnd1expression. We also analysed PC3 prostate carcinomacells and selected two marker genes of differentiatedprostate epithelia, that is, the prostate-specific antigenPS-A (Lilja, 2003) and androgen receptor (AR) (Linjaet al., 2001) genes. Neither of these genes is expressedin untreated cultures, yet both genes were induced inresponse to RT inhibitors (Figure 4b). Again, theexpression of all genes returned to the original levelwhen the inhibitors were removed. Thus, RT inhibitorydrugs modulate the expression of critical genes intransformed cells, consistent with the induction ofdifferentiation, yet this reprogramming is reversibleand is abolished when RT-inhibition is released.

RNA interference (RNAi) against RT-encoding LINE-1elements reduces proliferation and promotesdifferentiation in melanoma cells

At this point, we wanted to ascertain unambiguouslywhether reduction of cell growth and induction ofdifferentiation by pharmaceutical RT inhibitors arespecifically attributable to the inhibition of the cellular

RT. To address this question, RNAi experiments weredesigned to target specifically LINE-1 element familiesthat are most abundantly expressed in human cells(Brouha et al., 2003, and Supplementary information).Double-stranded RNA oligonucleotides directed againstLINE-1 ORF1 (L1-i, Figure 5a), or the lamin A/C gene(lam), or representing a noninterfering sequence (n.i.),

bcl-2

efvctr nev nev/r efv/rb

efv nev/r efv/ra

ctr nev

gapdh

ccnd1

E-cad

PS-A

AR

gapdh

c-myc

Figure 4 RT inhibitors modulate gene expression in A-375 (a) andPC3 (b) cell lines. RNA extracted from cells treated with DMSO(ctr), nevirapine (nev) or efavirenz (efv), and after removal ofnevirapine (nev/r) or efavirenz (efv/r), was amplified by RT–PCR,blotted and hybridized with internal oligonucleotides

e

f

n.i.

L1-i

020

40

60

80

100

48h 72h

cell

grow

th (

%)

a

21 bp L1-i oligo

ORF25, UTR ORF1 InterORF 3

, UTR

c-myc 45 (± 10,7)

EN-L1

19 (±9,3)

100Gapdh

ORF2

cyc D1 60 (± 8,7)

α-tub

c

22 (±11,8)ORF1

n.i. L1-i % L1-i/n.i.

n.i. L1-i

b 24h 48h 72h 72h

1 2

ORF1 lamin

gapdh

3 5 6 7 8 9 10

gapdh

n.i. L1 n.i. L1 n.i. L1 lam n.i . L1 lam RNAi

d

4

n.i.

L1-i

Figure 5 RNAi to LINE-1 induces morphological differentiation,reduces proliferation and modulates gene expression in A-375 cells.(a) Structure of a full-length LINE-1 element. The position of thesiRNA oligonucleotide L1-i is indicated. Arrowheads indicate thepositions of primer pairs used for RT–PCR analysis. (b) RT–PCRanalysis of LINE-1-ORF1 (upper panel) and GAPDH (lowerpanel) in RNA extracted from cells 24, 48 and 72 h after trans-fection of L1-i (lanes 2, 4 and 6) or noninterfering (n.i., lanes 1, 3and 5), or siRNA to the lamin A/C gene (lam, lane 7). Lanes 8, 9and 10 show RT–PCR reactions of lamin (upper panel) or GAPDH(lower panel) using RNA from cells transfected with noninterfer-ing, LINE-1 and lamin siRNAs, respectively. (c). Gene expressionpatterns after semiquantitative RT–PCR in A-375 cells 72 h aftertransfection with noninterfering or L1-i oligonucleotides. Quanti-tative variations (expressed as the % of signal in L1-i to signal inn.i.-transfected cultures) represent the mean and s.d. from threeindependent experiments. (d) Immunoprecipitated RT protein fromL1-i and n.i.-transfected cultures using L1-EN antibody. Analiquot of total extract was analysed by Western blot usinga-tubulin to control equal protein input in the immunoprecipi-tation. (e) Phase-contrast microscopy of A-375 cultures transfectedwith n.i. and L1-i oligonucleotides for 72 h. (f) Cell growth aftertransfection with n.i. and L1-i oligonucleotides. Counted cells areexpressed as the % of n.i. controls, taken as 100


3926

Oncogene 06/06/2013 116

were transfected in A-375 cells. Semiquantitative RT–PCR analysis 24, 48 and 72 h after transfection indicatedthat LINE-1 expression is progressively reduced inLINE-1-interfered cultures (Figure 5b, lanes 2, 4 and6) compared to controls treated with noninterfering(lanes 1, 3 and 5), or lamin RNAi (lane 7) in which laminexpression is abolished (lane 10). Conversely, lamin geneexpression was not affected in LINE-1-interfered cells(lane 9). Neither LINE-1 nor lamin RNAi influencedGAPDH expression. Thus, our RNAi conditions targetspecifically LINE-1 expression. At 72 h after L1-itransfection, expression of both ORF1 and ORF2 wasreduced by almost 80% compared to cells transfectedwith noninterfering oligonucleotide (Figure 5c). Toascertain that the RT protein product was consistentlydownregulated in L1-i interfered cells, we made useof a recently developed antibody against LINE-1 ORF2-encoded RT (termed EN-L1, Ergun et al., 2004).Immunoprecipitation and Western blot assays indicatethat RT protein levels 72 h after transfection aresignificantly reduced in L1-i cells (Figure 5d). Remark-ably, L1-i interfered cultures, in which the RT proteinwas downregulated, developed a typical differentiatedmorphology (Figure 5e), concomitant with reduced cellgrowth (Figure 5f), compared to cells transfected withnoninterfering oligonucleotide. Furthermore, LINE-1interference induced downregulation of expression ofc-myc and cyclin D1 genes, but not of GAPDH(Figure 5c). These effects are comparable to thoseobtained with pharmacological RT inhibition.

RT inhibitors reduce the growth of human tumorxenografts in athymic nude mice

Given that proliferation and differentiation can bemodulated by RT inhibition in transformed cells, wenext asked whether RT inhibitors also affect tumorgrowth in vivo. Tumorigenic cell lines selected for theseexperiments include A-375 and PC3 lines, as well asHT29 colon and H69 small cell lung carcinoma lines,which also showed reduced cell growth in response toRT inhibitors (Mangiacasale et al., 2003, and datanot shown). Cells were inoculated subcutaneously in thelimb of athymic nude mice. Animals were then subjectedto treatment with efavirenz, because this drug showed ahigher in vivo effectiveness than nevirapine in prelimin-ary assays. The optimal dose (20mg/kg body weight)was determined in dose-response experiments testing 4–40mg/kg of the drug. The efavirenz treatment provedsafe in all animal groups, with no animal death orexplicit signs of toxicity in any of the groups – althoughthe group treated with 40mg/kg showed a significantdecrease of body weight in more than 60% of animals.Figure 6 shows the recorded curves of tumor growth inmice untreated (red) or treated with efavirenz, startingone day (purple), or 1 week (yellow), after tumorinoculation. Tumor growth was markedly reduced intreated compared to untreated animals for all xenografttypes, and tumor progression was antagonized withcomparable effectiveness regardless of the timing of thetreatment start, despite of differences in the initial tumor

size. The growth curves of PC3- and HT29-derivedtumors in animals treated from day one after inocula-tion, but subjected to treatment discontinuation afterday 15 (green curves), indicate that the inhibition oftumor growth requires continuous RT inhibition in vivo.

Efavirenz-treated PC3 cells exhibit reducedtumorigenicity in vivo

Finally, we asked whether pretreatment of transformedcells with efavirenz modifies the tumorigenic potential ofderived xenografts. PC3 prostate cancer cells werecultured with efavirenz for two 96 h cycles, a time thatwas sufficient for induction of the PS-A and AR genes incultured cells (Figure 4b), and subsequently inoculated

Tum

or v

olum

e (c

m3 )

Tum

or v

olum

e (c

m3 )

Tum

or v

olum

e (c

m3 )

Tum

or v

olum

e (c

m3 )

8 10 12 14 16 18 2000.20.40.60.811.21.41.6

0

0.2

0.4

0.6

0.8

1

1.2

16 18 20 22 24 26 28 30

days

H69

PC3

HT29

6 9 12 15 18 21 24 27 30 330

0.5

1

1.5

2

2.5

3

00.20.40.60.811.21.41.61.8

6 9 12 15 18 21 24 27 30 33 36 39 42 45

A375

Figure 6 Efavirenz treatment reduces human tumor growth innude mice. The growth of tumors formed by the indicated cell lineswas monitored in untreated animals (red) and in animals treatedwith efavirenz 1 day (purple) or 1 week (yellow) after inoculation.Green curves show the growth of PC3- and H69-derived tumors inanimals treated starting 1 day after inoculation and subjected totreatment discontinuation after 14 days. Curves show the meanvalue of tumor size in groups of five animals


3927

Oncogene06/06/2013 117

in nude mice. Untreated cells were inoculated in parallelbatches of animals. Efavirenz-pretreated, or untreated,PC3 cell xenografts were then either continuouslytreated with efavirenz in vivo or were left untreated.Figure 7a shows the rate of tumor growth in theseexperiments: untreated PC3 cells developed fast-growingtumors in all animals. In contrast, efavirenz-pretreatedPC3 cells showed a significantly reduced tumor-formingability in vivo, and xenografts grew more slowly.Figure 7b shows the incidence of tumors in animalsmodels exposed to different treatments: while 100% ofinoculated animals developed aggressive tumors usinguntreated PC3 cells, efavirenz-pretreated cells developedslowly-growing xenografts in 65% of the inoculatedanimals. Moreover, only 40% of the animals inoculatedwith pretreated cells and further treated with efavirenzin vivo developed a tumor at all, and in that case thegrowth curve was flat. Thus, pretreatment of cells withanti-RT drugs before inoculation attenuates the tumori-genic potential of transformed cells.

Discussion

This work highlights three unexpected aspects of thehuman genome that have implications for cancer: first,LINE-1 elements are identified as active componentsof a mechanism involved in control of cell differen-tiation and proliferation; second, RNAi-dependentinactivation of LINE-1 elements, or pharmacologicalinhibition of the endogenous RT activity which they

encode, can restore control of these traits in transformedcells; third, inhibitors of RT reduce tumor growth inanimal models in vivo.The RT inhibitors used here, nevirapine and efavir-

enz, share a common biochemical mechanism of actionby binding the hydrophobic pocket in the p66 subunitof RT enzymes (Di Marzo Veronese et al., 1986; Renet al., 2001). Though being designed to target the HIV-encoded RT, nevirapine proved able to inhibit theendogenous retrotranscriptase activity present in non-infected cells (Mangiacasale et al., 2003) in a highlysensitive in vitro assay (Pyra et al., 1994). We now showthat both drugs reduce proliferation of transformedcells, largely independent of cell death, but associatedwith G1 delay or arrest. Concomitant with this, RTinhibitors induce morphological differentiation of trans-formed cells. The induction of differentiation is rapid,different from phenotypic changes elicited by inhibitorsof the telomerase-associated RT (TERT), which requirelong treatment times (120 days) (Damm et al., 2001).Furthermore, we never observed the reorganization ofactin stress fibers or focal adhesion sites typical ofsenescent cells. The absence of senescence-specificmodifications, and the rapid induction of differentia-tion, indicate that the RT inhibitors do not target TERTand induce a low-proliferating differentiated phenotyperather than senescence.That these effects are specifically associated with RT

inhibition was further demonstrated in RNAi experi-ments targeting active LINE-1 retroposon familiesaccounting for 84% of the overall retrotranspositioncapability in human cells (Brouha et al., 2003). In A-375cells, our RNAi conditions specifically downregulateexpression of LINE-1 ORF1 and ORF2 by some 80%,suggesting that the biologically active LINE-1 subgroupwas effectively silenced. Consistently, RT protein levelswere also downregulated in LINE-1-interfered cells.RNAi to RT encoding LINE-1 elements inducedmorphological, proliferative and gene-profiling changesthat are virtually indistinguishable from those causedby pharmacological RT inhibitors. The similarity ofphenotypes induced by independent approaches indi-cates that inhibition of LINE-1 expression, or of RTactivity, is sufficient to delay proliferation and promotedifferentiation. These observations rule out that non-specific side effects of the drugs cause the observedcell phenotypes and highlight the specificity of the roleof RT.Consistent with growth reduction and induction of

differentiation, RT inhibition caused the reprogram-ming of gene expression: this implicates the endogenousRT in modulation of expression of genes that promotethe transition from highly proliferating, transformedphenotypes to low proliferating, differentiated pheno-types, suggesting that genome function is the ultimatetarget of pharmaceutical or RNAi-dependent inhibitionof RT activity. Changes in gene expression are notinherited through cell division, but are reversible whenRT inhibition is released. The reversibility of examinedfeatures after release of the inhibition suggest thatLINE-1 encoded RT is part of an epigenetic mechanism

Figure 7 Reduced tumorigenicity of PC3 cells pretreated withefavirenz. (a) Growth of tumors formed by untreated or efavirenzpretreated cells injected in mice that were not treated or were post-treated with efavirenz in vivo. (b) The outcome of PC3-derivedxenografts after the indicated treatments for 30 days (n¼ 20animals/group)


3928

Oncogene 06/06/2013 118

that can modulate gene expression and contributes tothe molecular mechanisms underlying cell proliferationand differentiation.A relevant finding in this study is the ability of RT

inhibitors to reduce tumor growth in nude miceinoculated with four human xenograft models in vivo.Tumor growth was inhibited as long as the animalswere supplied with RT inhibitor, yet was resumed ondiscontinuation of the treatment, as observed in celllines, consistent with an epigenetic role of the endogen-ous RT in tumor growth. These data illustrate thepromising cytostatic ability of RT inhibitors in cancertreatment. Furthermore, in vitro pretreatment of PC3prostate carcinoma cells with efavirenz attenuates theirtumorigenicity in vivo. Thus, the activation of differ-entiation markers and reduced proliferation associatedwith RT inhibition are part of a large-scale reprogram-ming that can attenuate the malignant phenotype oftransformed cells in vivo.Growing data indicate that epigenetic changes can

reprogram tumor cells and convert the transformedphenotype into a ‘normal’ non pathological state(Lotem and Sachs, 2002; Li et al., 2003). Epigeneticreprogramming can bypass the genetic alterations thatoriginally caused the malignant transformation in avariety of tumors (Lotem and Sachs, 2002). Therefore,epigenetic regulatory factors are viewed as valuable,worth-challenging targets in tumor therapy (Egger et al.,2004). Retrotransposons can contribute to heterochro-matin formation in fission yeast (Schramke and Allshire,2003). Though such a mechanism has not yet beenproved in higher eukaryotes, unpublished results in ourlaboratory suggest that LINE-1-encoded RT is impli-cated in nuclear reorganization and positioning ofspecific genes. Importantly in this respect, the intra-nuclear position of genes is directly linked to theirexpression (Osborne et al., 2004 and references herein).Of relevance to the present study is the observation

that – while many tested compounds targeting the‘epigenome’ have generally proven toxic and/or chemi-cally unstable – both nevirapine and efavirenz have beenused in AIDS treatment for many years. Thus, theprospect of using these RT inhibitors in cancer therapywould have obvious advantages given their epidemio-logical record of generally good tolerance to continuedadministration. In retrospect, epidemiological evidenceindicate that Kaposi’s sarcoma (Portsmouth et al., 2003)and other AIDS-related cancers (Tirelli and Bernardi,2001) have a reduced incidence in patients treated withhighly active antiretroviral therapy (HAART): whilethis is generally viewed as a reflection of the improvedimmune reaction in treated patients, it may also suggesta direct inhibitory effect of HAART on the endogenousRT activity in tumor cells.

Materials and methods

Cell cultures

Human A-375 melanoma (ATCC-CRL-1619), TVM-A12primary melanoma-derived (Melino et al., 1993), HT29

adenocarcinoma (ATCC HTB-38), H69 small-cell-lung carci-noma (SCLC) (ATCC HTB119), and PC3 prostate carcinoma(ATCC CRL-1435) cell lines were seeded in six-well plates at adensity of 104 to 5� 104 cells/well and cultured in DMEM orRPMI 1640 medium with 10% fetal bovine serum. Nevirapineand efavirenz were purified from commercially availableViramune (Boehringer-Ingelheim) and Sustiva (Bristol-MyersSquibb) as described (Pittoggi et al., 2003). The drugs weremade 350 and 15mM in dimethyl sulfoxide (DMSO, Sigma-Aldrich), respectively, and added to cells 5 h after seeding; thesame DMSO volume (0.2% final concentration) was added tocontrols. Fresh RT inhibitor-containing medium was changedevery 48 h. Cells were harvested every 96 h, counted in aBurker chamber (two countings/sample) and replated at thesame density.

Cell cycle and cell death analysis

BrdU (20 mM) was added to the cultures during the last 30minbefore harvesting. Harvested cells were then treated with anti-BrdU antibody and propidium iodide (PI) and subjected tobiparametric analysis of the DNA content and BrdUincorporation in a FACStar Plus flow-cytometer (Beckton-Dickinson). Cell death was assessed by microscopy aftercombined staining with DAPI (nuclear morphology); PI (cellpermeability); and 3,3 dihexyl-oxacarbocyanine [DiOC6(3)], afluorescent probe for mitochondrial transmembrane potential.

Indirect immunofluorescence and confocal microscopy

Cell preparations were fixed with 4% para-formaldehyde for10min and permeabilized in 0.2% Triton-X100 in PBS for5min. Mouse monoclonal anti-bovine a-tubulin (MolecularProbes, A-11126) was revealed by Alexa Fluor 488-conjugatedsecondary antibody (Molecular Probes, A-11001) in A-375 andTVM-A12 cells and FITC-conjugated secondary antibody(Jackson Immunoresearch, cat 115-095-068) in PC3 cells.Nuclei were stained either with 2 mg/ml PI in the presenceof 0.1mg/ml ribonuclease A or with 0.1 mg/ml DAPI. Sampleswere imaged under a confocal Leica TCS 4D microscopeequipped with an argon/kripton laser. Confocal sections weretaken at 0.5–1 mm intervals.

Scanning electron microscopy (SEM)

Cells were fixed in 2.5%. glutharaldehyde in 0.1M Millonig’sphosphate buffer (MPB). After washing, cells were postfixedwith 1% OsO4 (1 h, 41C) in MPB and dehydrated usingincreasing acetone concentrations. Samples were critical-pointdried using liquid CO2 and sputter-coated with gold beforeexamination on a Stereoscan 240 scanning electron microscope(Cambridge Instr., Cambridge, UK).

Semiquantitative RT–PCR

RNA extraction and treatment with RNase-free DNase I wereas described (Pittoggi et al., 2003). cDNAs were synthesizedusing 300 ng of RNA, oligo (dT) and the Thermoscript system(Invitrogen). Reaction mixtures (1/25) were amplified using thePlatinum Taq DNA Polymerase kit (Invitrogen) and 30 pmolof oligonucleotides (MWGBiotech, Ebersberg, Germany; seeSupplementary information) in an initial 2-min step at 941C,followed by cycles of 30 s at 941C, 30 s at 58–621C, 1min at721C. Each oligo pair was used in sequential amplificationseries with increasing numbers (25–40) of cycles. PCR productswere electrophoresed, transferred to membranes and hybri-dized for 16 h at 421C with [32P]g-ATP end-labeled internaloligonucleotides. The intensity of the amplification signal was


3929

Oncogene06/06/2013 119

measured by densitometry in at least three independentexperiments for each gene and normalized to the GAPDHsignal in the same experiment.

RNA interference

Four 21-nt double-stranded siRNA oligonucleotides encom-passing region 1367–2056 in LINE-1 were designed to targetthe consensus sequence of active LINE-1 elements describedby Brouha et al. (2003). The siRNA oligonucleotide targetingbases 2035–2056 (L1-i) was most effective and was used in allexperiments. Control cells were treated with noninterferingoligonucleotide (n.i.), 30-fluorescein-conjugated to monitortransfection efficiency, or with specific siRNA against thelamin A/C gene. All siRNAs were synthesized by Qiagen USA.A-375 melanoma cells were transfected using RNAiFecttransfection reagent (Qiagen) and 300 nM of siRNA. Detailsfor siRNA assays are in Supplementary information.

Western blot and immunoprecipitation

At 72 h after siRNA transfection A-374 cells were harvestedin PBS with 0.1mM. PMSF and lysed in lysis buffer (50mM

Tris-HCl, pH 8.1, 10mM EDTA, 1% SDS) supplemented withprotease inhibitors (1 mM pepstatin, 1mM leupeptin, 0.1mM

PMSF). After centrifugation at 12 000 r.p.m., 41C, 15min, theprotein concentration in the supernatants was determinedusing a Coomassie colorimetric assay (Pierce). Samples (20 mg)were loaded on NuPAGE Novex 10% Bis-Tris gel (Invitro-gen), transferred onto membranes and verified by Westernimmunoblotting using monoclonal anti-a-tubulin (Sigma,T5168) and HRP-conjugated secondary antibody (BIORAD,170-6516). In total, 500mg of protein extract were thenprecleared using 75ml of protein A-Agarose-50% Slurry beads(Upstate Biotechnology) for 30min at 41C. After centrifuga-tion (12 000 r.p.m., 41C), supernatants were incubated over-night (41C, with continuous rotation) with chicken polyclonal

anti-EN-L1 (Ergun et al., 2004), kindly given by GeraldSchumann (Paul-Ehrlich-Institut, Langen, Germany). Pre-cleared extracts were then incubated with 60 ml of beads (1 h,41C) with rotation. After removal of the supernatants, proteinswere eluted from the beads in 1% SDS, 0.1M NaHCO3

and precipitated with 10 volumes of acetone. Pellets wereresuspended in TE buffer and loaded on NuPAGE Novex10% Bis-Tris gel as above. Western blot analysis was carriedout using chicken polyclonal anti-EN-L1 (1 : 40 dilution) anddonkey HRP-conjugated anti-chicken IgY (Jackson Immuno-research Laboratories, 703-035-155).

Tumor xenografts and treatment of animals

Athymic nude mice (5 weeks old) (Harlan, Italy), kept inaccordance with the European Union guidelines, wereinoculated subcutaneously with A-375 melanoma (4� 106),H-69 (107), PC3 (2� 106) and HT-29 (106) cells in 100ml PBS.Mice were subcutaneously injected daily five days a week withEfavirenz (20mg/kg) using a 4mg/ml stock in DMSO freshlydiluted 1 : 1 with physiological solution. Controls were injectedwith 50% DMSO. Treatment started 1 day or 1 week aftertumor implant, and, where indicated, was discontinued after14 days. Tumor growth was monitored every other day bycaliper measurements; volumes were calculated using theformula length�width� height� 0.52 (Umekita et al., 1996).

Acknowledgements

We are indebted with Dr Gerald Schumann for kindlyproviding the antibody against LINE1-encoded RT. We arealso grateful to Dr A Mai for drug purification and to Drs RMangiacasale and S Rutella for cell cycle analysis. This workwas supported by Istituto Superiore di Sanita (Grant C3H3‘Role of endogenous Reverse Transcriptase in tumor growthand embryo differentiation’).

References

Brouha B, Schustak J, Badge RM, Lutz-Prigge S, Farley AH,Moran JV and Kazazian Jr HH. (2003). Proc. Natl. Acad.Sci. USA, 100, 5280–5285.

Crone TM, Schalles SL, Benedict CM, Pan W, Ren L, LoySE, Isom H and Clawson GA. (1999). Hepathology, 29,

1114–1123.Damm K, Hemmann U, Garin-Chesa P, Hauel N,Kauffmann I, Priepke H, Niestroj C, Daiber C, EnekelB, Guilliard B, Lauritsch I, Muller E, Pascolo E, SauterG, Pantic M, Martens UM, Wenz C, Lingner J, KrautN, Rettig WJ and Schnapp A. (2001). EMBO J., 20,

6958–6968.Deininger PL, Moran JV, Batzer MA and Kazazian HH.(2003). Curr. Opin. Gen. Dev., 13, 651–658.

Deragon JM, Sinnett D and Labuda D. (1990). EMBO J., 9,

3363–3368.Di Marzo Veronese F, Copeland TD, De Vico AL, RahmanR, Oroszlan S, Gallo RC and Sarngadharan MG. (1986).Science, 231, 1289–1291.

Egger G, Liang G, Aparicio A and Jones PA. (2004). Nature,429, 457–463.

Ergun S, Buschmann C, Heukeshoven J, Dammann K,Schnieders F, Lauke H, Chalajour F, Nerbil K, StratlingWH and Schumann GG. (2004). J. Biol. Chem., 279,27753–27763.

Friedlander A and Patarca R. (1999). Crit. Rev. Oncogenesis,10, 129–159.

Giordano R, Magnano AR, Zaccagnini G, Pittoggi C,Moscufo N, Lorenzini R and Spadafora C. (2000). J. CellBiol., 148, 1107–1113.

Hagan CR and Rudin CM. (2002). Am. J. Pharmacogenomics.,2, 25–36.

Hagan CR, Sheffield RF and Rudin CM. (2003). Nat. Genet.,35, 219–220.

Hsu MY, Meier FE, Nesbit M, Hsu JY, Van Belle P, Elder DEand Herlyn M. (2000). Am. J. Pathol., 156, 1515–1525.

Khan AS, Muller J and Sears JF. (2001). Virus Res., 79,

39–45.Kiessling AA, Crowell R and Fox C. (1989). Proc. Acad. Natl.

Sci. USA, 86, 5109–5113.Kuo KW, Sheu HM, Huang YS and Leung WC. (1998).

Biochem. Biophys. Res. Commun., 253, 566–570.Li L, Connelly MC, Wetmore C, Curran T and Morgan JI.(2003). Cancer Res., 63, 2733–2736.

Li TH and Schmid CW. (2001). Gene, 276, 135–141.Lilja H. (2003). Urology, 62 (Suppl 1), 27–33.Linja MJ, Savinainem KJ, Saramaki OR, Tammela TL,Vessella RL and Visakorpi T. (2001). Cancer Res., 61,

3550–3555.Lotem J and Sachs L. (2002). Cancer Biol., 12, 339–346.


3930

Oncogene 06/06/2013 120

Mangiacasale R, Pittoggi C, Sciamanna I, Careddu A, MatteiE, Lorenzini R, Travaglini L, Landriscina M, Barone C,Nervi C, Lavia P and Spadafora C. (2003). Oncogene, 22,

2750–2761.Martin SL. (1991). Mol. Cell. Biol., 11, 4804–4807.Martin SL and Branciforte D. (1993). Mol. Cell. Biol., 13,5383–5392.

Melino G, Sinibaldi Vallebona P, D’Atri S, Annichiarico-Petruzzelli M, Rasi G, Catani MV, Tartaglia MI, Vernole P,Spagnoli LG, Finazzi-Agro A and Garaci E. (1993). Clin.Chem. Enzyme Comms., 6, 105–119.

Mwenda JM. (1993). Cell. Mol. Biol., 39, 317–328.Osborne CS, Chalakova L, Brown KE, Carter D, Horton A,Debrand E, Goyenechea B, Mitchell JA, Lopes S, Reik Wand Fraser P. (2004). Nat. Genet., 36, 1065–1071.

Ostertag EM and Kazazian Jr HH. (2001). Annu. Rev. Genet.,35, 501–538.

Packer AI, Manova K and Bacharova RF. (1993). Dev. Biol.,157, 281–283.

Pittoggi C, Sciamanna I, Mattei E, Beraldi R, Lobascio AM,Mai A, Quaglia MG, Lorenzini R and Spadafora C. (2003).Mol. Reprod. Dev., 66, 225–236.

Pyra H, Boni J and Schupbach J. (1994). Proc. Natl. Acad. Sci.USA, 91, 1544–1548.

Portsmouth S, Stebbing J, Gill J, Mandalia S, Bower M, NelsonM, Bower M and Gazzard B. (2003). AIDS, 17, F17–F22.

Poznanski AA and Calarco PG. (1991). Dev. Biol., 143,

271–281.Ren J, Nichols C, Bird L, Chamberlain P, Weaver K, Short S,Stuart DI and Stammers DK. (2001). J. Mol. Biol., 312,

795–805.Sauane M, Gopalkrishnan RV, Sarkar D, Su ZZ, LebedevaIV, Dent P, Pestka S and Fisher PB. (2003). Cyt. GrowthFactor Rev., 14, 35–51.

Sauter ER, Yeo UC, Von Stemm A, Zhu W, Litwin S,Tichansky DS, Pistritto G, Nesbit M, Pinkel D, Herl M andBastian BC. (2002). Cancer Res., 62, 3200–3206.

Schramke V and Allshire R. (2003). Science, 301, 1069–1074.Tirelli U and Bernardi D. (2001). Eur. J. Cancer., 37,1320–1324.

Umekita Y, Hiipakka RA, Koknotis JM and Shutsung L.(1996). Proc. Natl. Acad. Sci. USA, 93, 11802–11807.

Utikal J, Leiter U, Udart M, Kaskel P, Peter RU and KrahnGM. (2002). Cancer Invest., 20, 914–921.

Supplementary Information accompanies the paper on Oncogene website (http://www.nature.com/onc)


3931

Oncogene06/06/2013 121

Eva-Maria Böcker

Disruption of an EHMT1-Associated Chromatin-Modification Module Causes Intellectual Disability Quelle: The American Journal of Human Genetics, 2012 Betreuer: Prof. Andreas W. Kuß, Dr. Lars R. Jensen (Humangenetik) Was bedeutet mir das Thema persönlich? Da sich mein Promotionsthema mit mentaler Retardierung beschäftigt, ist es für mich natürlich interessant, mich auch mit anderen genetischen Ursachen für geistige Behinderung auseinanderzusetzen. Anhand dieses Themas wird deutlich, dass ganz unterschiedliche Gene letztendlich denselben Signalweg nutzen, sodass vielleicht durch die allmähliche Identifizierung der mit geistiger Behinderung assoziierten Gene eine Grundlage für eine Therapie dieser Erkrankung eröffnet werden kann. Worauf kommt es mir bei diesem Thema am meisten an? Bei der Auswahl des Themas war mir wichtig, dass der Leser (zumindest grobe) Einblicke in die Vorgehensweise der humangenetischen Forschung bekommt. Zudem zeigt das vorliegende Paper, dass es in Gestalt von Drosophila neben Maus und Ratte ein weiteres Tiermodell für neuronale Prozesse im Menschen gibt, das aber häufig unterschätzt wird. Was fasziniert mich selbst am Thema am meisten? Meiner Meinung nach ist dieses Thema aus mehreren Gründen faszinierend. Es zeigt die Vielseitigkeit des Krankheitsbildes der mentalen Retardierung, die besonders als syndromale Form mit diversen zusätzlichen Symptomen assoziiert sein kann. Genauso spannend ist auch die Feststellung, dass einerseits verschiedene Mutationen in demselben Gen ganz unterschiedliche Erkrankungen zur Folge haben können und andererseits eine Krankheit durch verschiedene Gendefekte verursacht sein kann. Außerdem zeigt dieses Thema sehr schön den Zusammenhang zwischen genetischen und epigenetischen Faktoren. Was gefällt mir am Thema weniger? Leider steht das Thema nicht in einem direkten Zusammenhang mit meinem Forschungsthema. Dennoch handelt es sich hier meines Erachtens um eine sehr spannende Thematik!

06/06/2013 122

ARTICLE

Disruption of an EHMT1-Associated Chromatin-Modification Module Causes Intellectual Disability

Tjitske Kleefstra,1,2,3,10,* Jamie M. Kramer,1,3,4,10 Kornelia Neveling,1,2,3 Marjolein H. Willemsen,1,2,3

Tom S. Koemans,1,3,4 Lisenka E.L.M. Vissers,1,2,3 Willemijn Wissink-Lindhout,1 Michaela Fenckova,1,3,4

Willem M.R. van den Akker,1,3,4 Nael Nadif Kasri,3,4,5 Willy M. Nillesen,1 Trine Prescott,6

Robin D. Clark,7 Koenraad Devriendt,8 Jeroen van Reeuwijk,1,3 Arjan P.M. de Brouwer,1,3,4

Christian Gilissen,1,2,3 Huiqing Zhou,1,3,9 Han G. Brunner,1,2,3 Joris A. Veltman,1,2,3

Annette Schenck,1,3,4,10,* and Hans van Bokhoven1,3,4,10

Intellectual disability (ID) disorders are genetically and phenotypically highly heterogeneous and present a major challenge in clinical

genetics andmedicine. Althoughmany genes involved in ID have been identified, the etiology is unknown inmost affected individuals.

Moreover, the function ofmost genes associatedwith ID remains poorly characterized. Evidence is accumulating that the control of gene

transcription through epigenetic modification of chromatin structure in neurons has an important role in cognitive processes and in the

etiology of ID. However, our understanding of the keymolecular players andmechanisms in this process is highly fragmentary. Here, we

identify a chromatin-modification module that underlies a recognizable form of ID, the Kleefstra syndrome phenotypic spectrum (KSS).

In a cohort of KSS individuals without mutations in EHMT1 (the only gene known to be disrupted in KSS until now), we identified

de novo mutations in four genes, MBD5, MLL3, SMARCB1, and NR1I3, all of which encode epigenetic regulators. Using Drosophila,

we demonstrate that MBD5, MLL3, and NR1I3 cooperate with EHMT1, whereas SMARCB1 is known to directly interact with MLL3.

We propose a highly conserved epigenetic network that underlies cognition in health and disease. This network should allow the design

of strategies to treat the growing group of ID pathologies that are caused by epigenetic defects.

Introduction

Intellectual disability (ID) disorders affect about 1%–3% of

the western population and are genetically and phenotyp-

ically highly heterogeneous. Mutations in more than 400

genes have been identified, yet the genetic cause remains

unknown in the majority of individuals with ID.1 The

genetic and phenotypic heterogeneity of IDmakes concep-

tualizing strategies for treatment difficult. However,

a number of genes that cause ID appear to act in common

molecular and cellular pathways, raising the possibility

that a group of genetically heterogeneous individuals

with ID could be treated if a common molecular etiology

were targeted.2 One of the emerging mechanisms that

appears to be important in ID is the regulation of neuronal

function through the epigenetic control of gene transcrip-

tion.3 Epigenetic modifications are important factors in

the regulation of cognition in animal models, and several

well-characterized ID disorders, such as Rett syndrome

(MIM 312750), Angelman syndrome (MIM 105830), and

Fragile X syndrome (MIM 300624), have a known epige-

netic origin.3,4 Here, we reveal a chromatin-modification

module that underlies another group of ID disorders with

1Department of Human Genetics, Radboud University Nijmegen Medical Ce

Genetic and Metabolic Diseases, Radboud University Nijmegen Medical Cent

for Molecular Life Sciences, Radboud University Nijmegen Medical Centre, P.

Brain, Cognition, and Behaviour, Radboud University Nijmegen Medical Cen

Cognitive Neuroscience, Radboud University Nijmegen Medical Centre, P.O.

Genetics, Oslo University Hospital, P.O. Box 4950, Nydalen, Oslo N-0424, No

University School of Medicine, Loma Linda, CA 92354, USA; 8Centre for Hum9Department of Molecular Developmental Biology, Faculty of Science, Radbou10These authors contributed equally to this work

*Correspondence: [email protected] (T.K.), [email protected] (A.

DOI 10.1016/j.ajhg.2012.05.003. �2012 by The American Society of Human

The06/06/2013 123

an epigenetic origin, the Kleefstra syndrome phenotypic

spectrum (KSS [see Subjects and Methods]).

The core phenotype of Kleefstra syndrome (MIM

610253) comprises ID, childhood hypotonia, and distinc-

tive facial features.5–7 KSS can be caused by haploinsuffi-

ciency of EHMT1, which encodes a histonemethyltransfer-

ase capable of histone 3 lysine 9 dimethylation (H3K9me2)

in euchromatic regions of the genome.8,9 In our KSS

cohort, about 25% of individuals have EHMT1 loss-of-

function mutations.3 We hypothesized that the ‘‘EHMT1-

negative’’ individuals have mutations in genes that share

a biological function with EHMT1. Here, we report de

novo mutations in four functionally related genes in four

individuals with KSS, indicating that this subclass of ID is

caused by disruption of a common epigenetic module.

Subjects and Methods

SubjectsWe have collected a clinically defined cohort of individuals with

the core features of Kleefstra syndrome and have identified

EHMT1 loss-of-function mutations in 25% of these individuals.3

Despite the clinical similarities in our cohort, there is also

ntre, P.O. Box 9101, 6500 HB Nijmegen, The Netherlands; 2Institute for

re, P.O. Box 9101, 6500 HB Nijmegen, The Netherlands; 3Nijmegen Centre

O. Box 9101, 6500 HB Nijmegen, The Netherlands; 4Donders Institute for

tre, P.O. Box 9104, 6500 HE Nijmegen, The Netherlands; 5Department of

Box 9101, 6500 HB Nijmegen, The Netherlands; 6Department of Medical

rway; 7Division of Medical Genetics, Department of Pediatrics, Loma Linda

an Genetics, University of Leuven, P.O. Box 602, 3000 Leuven, Belgium;

d University Nijmegen

S.)

Genetics. All rights reserved.

American Journal of Human Genetics 91, 73–82, July 13, 2012 73



http://dx.doi.org/10.1016/j.ajhg.2012.05.003

Figure 1. Clinical Photographs and Mutation Information(A–D) Photographs of individuals with KSS and chromatogramscomparing individuals with KSS to parents and/or siblings indicatede novo occurrence for the four probably pathogenic mutationsidentified. Mutations are highlighted in yellow.(A) KS113 with MLL3 mutation c.4441C>T. Reverse-strandsequences of individual KS113, the mother, and two healthysisters are shown. Note the midface hypoplasia, synophrys,upward slant of palpebral fissures, and everted lower lip.(B) KS47 with SMARCB1mutation c.110G>A, which is not presentin the parental DNA. Note the midface hypoplasia, upward slantof the eyes, and tongue protrusion.(C) KS78 with MBD5 mutation c.150del, which is not present inthe parental DNA. A reverse-strand sequence is shown. Note thesynophrys, upward slant of palpebral fissures, upturned nosewith broad tip, full lips, and everted lower lip.(D) KS220withNR1I3mutation c.740T>C, which is not present inthe parental DNA. Reverse-strand sequences are shown. Note themidface hypoplasia, short upturned nose, everted lower lip, andpointed chin.(E) Four KSS individuals (KS1, KS2, KS245, and KS21) with previ-ously published6,7,10 EHMT1 defects show a close resemblance offacial characteristics to the four individuals in (A)–(D).

74 The American Journal of Human Genetics 91, 73–82, July 13, 201206/06/2013 124

heterogeneity for certain features (e.g., renal anomalies and

hearing loss) and significant clinical overlap with other related

syndromes, such as Smith-Magenis syndrome (MIM 182290).

Therefore, we refer to this group of related phenotypes as KSS.

From this study population, we selected nine cases who were

found to be negative for EHMT1 defects and who were previously

screened for pathogenic copy-number variants (CNVs) by high-

density microarray platforms (Affymetrix 250K microarray equiv-

alentorhigher resolution). Informed consentwasobtained fromall

individuals and their parents after approval for the research project

was obtained from the institutional ethical board at Radboud

University Nijmegen Medical Centre, Commissie Mensgebonden

Onderzoek Regio Arnhem-Nijmegen (NL36191.091.11). In this

study, we identified de novo mutations in four individuals

(Figure 1). General information on medical and developmental

characteristics of these four individuals is given below, and their

clinical characteristics are compared with those of KSS individuals

with EHMT1mutations in Table 1.

Individual KS113 was born with a normal birth weight after

a normal pregnancy and delivery to nonconsanguineous healthy

parents. She walked independently at the age of 19 months but

never developed speech. She gradually developed problematic

behavior with periods of hyperactivity and aggressiveness. Her

total intelligence quotient (TIQ) was 35. At the age of 15 years,

she had a height of 148 cm (�2.5 standard deviations [SDs]),

a weight of 41 kg (0 SDs), and an occipitofrontal circumference

(OFC) of 52 cm (�2 SDs). Additional examinations, including

a metabolic screen of blood and urine and DNA analysis of RAI1,

revealed no abnormalities.

Individual KS47, the second child of healthy nonconsanguine-

ous parents, was born at 36 weeks of gestation after a normal preg-

nancy with a birth weight of 2,400 g, a length of 49 cm, and head

circumference of 33.5 cm. Neonatally, Down syndrome (MIM

190685) was suspected. At the age of 2.5 years, she was shunted

for hydrocephalus. At that age, she was not able to sit and did

not speak any words. Subsequently, she required multiple neuro-

surgical procedures for shunt problems because she had an abnor-

mally high production of cerebrospinal fluid as a result of

a chorioid plexus anomaly. A partial plexectomy was required,

and she has since been clinically stable. She had impaired visual

function, which was probably at least partially central in nature.

She is myopic (approximately �7 diopters bilaterally). She is

a generally sociable person who is interested in her surroundings.

At the age of 9.5 years, her height was 144 cm (0 SDs), her weight

was 31 kg (�1 SD), and her head circumference was 52.4 cm

(0 SDs). No abnormalities were detected on a cardiac ultrasound,

a DNA analysis of UBE3A, a Southern blot for 15q11-q13, or stan-

dard cytogenetic karyotyping in both blood and fibroblasts.

Individual KS78 was born to healthy nonconsanguineous

parents. He has three normal siblings, and his 8-month-old

brother had laryngomalacia and macrocephaly. Vaginal delivery

was induced at 42 weeks of gestation because of a postdate preg-

nancy. His birth weight was 3,700 g. He had apnea and stridor

with laryngomalacia. He walked at 17 months and spoke only

single words until after age 3. He was treated with daily growth-

hormone injections for one year between the ages of 12 and

13 years, and this treatment resulted in amodest response of about

4 cm. Since then, he has grown only about 1 cm. He had a grand

mal seizure between the ages of 2 and 3, but since then, he has had

staring spells and mild electroencephalography (EEG) changes.

His TIQ was 69 at the age of 12 years. His behavior is characterized

by anxiety, a high pain tolerance, stereotypic and self-injurious

Table 1. Clinical Characteristics in Individuals with DifferentGenetic Causes of KSS

Symptoms KS113 KS47 KS78 KS220EHMT1Defects

Sex female female male female

Intellectual disability þ þ þ þ 100%

Childhood hypotonia þ þ þ þ 100%

Microcephaly þ � � � 20%

Short stature þ � þ þ 20%

Overweight � � � � 45%

Brachycephaly þ þ þ � 40%

Midface hypoplasia þ þ þ þ 80%

Coarse facies þ þ þ � 50%

Hypertelorism þ þ þ þ 30%

Synophrys þ þ þ � 60%

Arched eyebrows � � þ � 30%

Short nose � þ þ þ 45%

Anteverted nostrils � þ þ þ 25%

Macroglossia(protruding tongue)

� þ � þ 40%

Tented and cupid-bowedupper lip

þ þ þ þ 25%

Thick and evertedlower lip

þ � þ þ 25%

Pointed chin þ � � þ 25%

Dysplastic ear helices þ � � � 50%

Brachydactyly � þ � � 15%

Cardiac anomaly � � � � 45%

Renal anomaly � � � � 15%

Behavioral problems þ � þ þ 75%

Hearing loss(sensorineural)

� � � � 15%

Seizures � � þ � 25%

behaviors such as hand biting, handwringing, and spasmodic self-

hugging, and bouncing when excited. Several nights a week, he

has sleep problems that include 2–3 hr of wakefulness in the

middle of the night. He had aggressive outbursts for which he

received Abilify (15 mg/day), which reduced the frequency

and intensity of the outbursts. At the age of 16 years, he had a

short stature (�2 SDs) and macrocephaly (þ2 SDs). Additional

examinations, including a brain magnetic resonance image

(MRI), a metabolic screen of blood and urine, and DNA analysis

of RAI1, revealed no abnormalities.

Individual KS220 was the third child of healthy, unrelated

parents and was born at term after an uneventful pregnancy

with a birth weight of 3,300 g, a length of 51 cm, and a head

circumference of 36 cm. At the age of 5 months, she developed

a pyelonephritis. Her development was severely delayed: At the

age of 19 months, her mental development was that of a

12-month-old, and at 4.5 years of age, her development was that

of a 2.8-year-old and she attended a special school. Autism spec-

The06/06/2013 125

trum disorder was diagnosed, and she had sleeping difficulties.

When she was 5 years and 11 months old, a clinical examination

revealed that she had a weight of 15.4 kg (0 SDs), a height of 99 cm

(�3.5 SDs), and an OFC of 50 cm (�0.5 SDs). Additional investiga-

tions comprising EEG, a brain MRI (at the age 7 of months), and

abdominal and cardiac ultrasounds revealed no abnormalities.

A screen for metabolic abnormalities and methylation analysis

for Angelman and Prader-Willi (MIM 176270) syndromes were

normal.

Next-Generation SequencingTargeted sequencing was performed on five KSS individuals. The

enriched genes were selected from the AmiGO Gene Ontology

database (see Web Resources) and comprised all genes annotated

with the GO term ‘‘chromatin modification’’ and EHMT1 interac-

tors known from the STRING database (seeWeb Resources). A total

of 316 genes (Table S1, available online) were targeted on a 385K

sequence capture array (Roche NimbleGen). The array design

comprised all coding exons, including surrounding sequences to

cover the splice sites. In total, the design included 4,658 targets

comprising 1,429,871 bp. Sequence capture and posthybridiza-

tion ligation-mediated PCR were done according to the manufac-

turer’s (Roche NimbleGen) instructions with the Titanium

optimized protocol. The amplified captured samples were used

as input for emulsion PCR (emPCR) amplification and subsequent

sequencing with the use of a Roche 454 GS FLX Titanium

sequencer. Data analysis was done with Roche Newbler software

(v.2.3) and the human genome build hg18 (NCBI Build 36.1).

Mapping and coverage statistics were extracted with custom soft-

ware (Table S2). Sequence variations were automatically detected

during mapping, and they were annotated with known RefSeq

genes (UCSC hg19, see Web Resources) and SNP information

(dbSNP129) with the use of in-house analysis software.11

A SOLiD optimized SureSelect Human All Exon Kit (50 Mb,

~21,000 genes; Agilent Technologies) and 3 mg genomic DNA

were used for exome sequencing on four trios of parents and their

affected children. Library preparation was performed as described

previously.12 To allow formultiplexing, we used posthybridization

sample barcodes (Agilent Technologies). Enriched exome libraries

were pooled in equimolar sets of four on the basis of a combined

library concentration of 0.7 pM. Subsequently, the obtained

pool was used for emPCR and bead preparation with the EZbead

system (Life Technologies) and for subsequent sequencing with

a SOLiD 4 system (Life Technologies). Mapping and variant calling

were performed as described previously (Table S3).13

For both targeted and exome sequencing, variants and indels

were selected with the use of strict quality-control settings,

including the presence of at least four (unique) variant reads and

at least 15% variant reads. On average, 2,474 and 21,895 variants

per proband were annotated for targeted sequencing and exome

sequencing, respectively (Table 2).11,12 For variant prioritization,

all nongenic, intronic (other than canonical splice sites), and

synonymous variants were excluded, resulting in an average of

94 (targeted) and 5,596 (exome) variants per sample (Table 2).

Next, all variants from either dbSNP (dbSNP129 [targeted] or

dbSNP132 [exome]) or our in-house database were excluded,

reducing the number of variants to an average of 9 (targeted)

and 167 (exome) (Table 2). On the basis of the knowledge that

KSS is an autosomal-dominant disease caused by de novo muta-

tions, an autosomal-dominant model of inheritance was used for

further prioritization, and all inherited variants except for the

X-linked changes were excluded. X-linked changes were only


Table 2. Prioritization of Detected Variants in Targeted and Exome NGS

Sample

Targeted NGS Family-Based Exome Sequencing

KS113 KS47 KS35 KS94 KS129 Trio KS78 Trio KS220 Trio KS53 Trio KS49

High-confidence variant calls 2,495 2,452 2,403 2,530 2,488 24,604 19,610 20,858 22,511

After exclusion of nongenic, intronic,and synonymous variants

90 98 81 88 102 6,244 5,182 5,424 5,638

After exclusion of known variants 7 14 7 9 10 158 112 299 98

After exclusion of inherited variants � � � � � 4 5 9 3

Confirmed by Sanger sequencing 5 6 3 5 4 1 3 0 0

De novo mutations 1 1 0 0 0 1 3 0 0

The following abbreviation is used: NGS, next-generation sequencing.

excluded when they were present in the paternal DNA. For the

trios examined by exome sequencing, inherited variants were

excluded by the selection of variants that were only present in

the child but not the parents; this resulted in an average number

of five potential de novo variants per proband. Of these, one

variant, on average, was confirmed to be de novo by Sanger

sequencing. For targeted sequencing, the absence of parental

next-generation sequencing (NGS) data required validation by

Sanger sequencing for all variants. Confirmed variants (Tables S4

and S5) were subsequently analyzed in the parental DNA for the

determination of the mode of inheritance. On average, five vari-

ants were confirmed by Sanger sequencing; of these five, zero to

one variant per sample could be proven to be de novo (Table 2).

To further explore the pathogenicity of the de novo variants, we

evaluated data from in silico resources. These comprised the

genomic evolutionary conservation score (phyloP) as well as Poly-

Phen-2 and SIFT (see Web Resources), which predict the impact of

amino acid substitutions on protein structure and function.

Haplotype Analysis with Short-Tandem-Repeat

MarkersPrimers to amplify polymorphic short-tandem-repeat markers in

7q36weredesignedwith thePrimer3program(seeWebResources).

AnM13 tail was added to the 50 and 30 ends of the primers. Markers

were amplified with an M13 forward primer labeled with one of

the fluorophores—FAM, VIC, NED, or ROX—at the 50 end and

a M13 reverse primer with a 50-GTTTCTT-30 added to its 50 end to

reduce tailing. Markers and primer sequences used for haplotype

analysis are shown in Table S6. Final PCR products were mixed

with eight volumes of formamide and half a volume of Genescan

500 (�250) LIZ Size Standard and were analyzed with the ABI

PRISM 3730 DNA analyzer (Applied Biosystems). The results were

evaluated by Genemapper (Applied Biosystems).

Genetic Interaction StudiesWe performed genetic interaction studies on Drosophila by using

the UAS/Gal4 ectopic expression system14 to induce overexpres-

sion or RNAi-mediated knockdown15 of gene orthologs implicated

in KSS. Gene orthologs were identified by the reverse BLAST

method16 and analysis of the treefam database.17 For MLL3,

SMARCB1, and MBD5, clear orthologs (trr, snr1, and sba, respec-

tively) were identified. NR1l3 is formally orthologous to HR96

because it is monophyletic in phylogenetic analyses.17 However,

it is quite unlikely that the two proteins that they encode are func-

tionally equivalent. HR96 has acquired an internal repeat,18 which


is not seen in NR1I3. This results in a lack of homology in the

middle section of the proteins. In contrast, the ecdysone receptor

(EcR) is homologous to NR1I3 across the entire length of the

protein and has a much higher similarity than does HR96. Protein

BLASTof NR1I3 against theDrosophila genome gives EcR as the top

hit in Drosophila melanogaster. We therefore conclude, on the basis

of domain composition and amino acid similarity, that EcR is the

best candidate to be used for functional studies.

Fly stocks were obtained from the BloomingtonDrosophila Stock

Center (Indiana University) and the Vienna Drosophila RNAi

Center (Institute for Molecular Pathology).15 For a complete list

of publically available stocks that were used, see Table S7. UAS-

EHMT flies were described previously.19 For genetic interaction

studies, females of the genotypes MS1096-Gal4 and MS1096-

Gal4/FM7d; UAS-EHMT were crossed to fly stocks indicated in

Table S7. Wing phenotypes were examined in female progeny.

Results

Identification of De Novo Mutations in Epigenetic

Regulators in Four Individuals with KSS

To identify mutations in our EHMT1-negative cohort of

individuals with KSS (Table 1, Subjects and Methods), we

used two strategies. Five individuals were analyzed by a tar-

geted approach in which genes involved in chromatin

modification were sequenced. Because the targeted

approach could limit the possibility of mutation discovery,

four other individuals and their healthy parents were

analyzed by whole-exome sequencing. The retrieved vari-

ants were filtered as described previously so that potential

disease-causing mutations could be identified.12 All vari-

ants remaining after filtering were analyzed by Sanger

sequencing, which revealed six de novo mutations in

four individuals (Tables S4 and S5). Of these six mutations,

none were observed in our in-house exome-sequencing

database (450 individuals). One, identified in POF1B

on the X chromosome, was found once in the 8,760 alleles

reported in the National Heart, Lung, and Blood Institute

(NHLBI) database (Web Resources), which contains data

from more than 5,300 exomes (>10,600 alleles), leaving

five unique de novo changes identified in this study

(Table 3).

Table 3. List of Unique De novo Variantsa Identified by NGS

Gene Individual Sex

RefSeqAccessionNumber

cDNAChange

ProteinChange

PhyloPScore

PolyPhen-2Prediction

SIFTPrediction Protein Function

MLL3 KS113 female NM_170606.2 c.4441C>T p.Arg1481* 6.2 damaging deleterious trimethylates histone H3 atlysine 4 (H3K4me3); centralcomponent of the ASC-2complex

SMARCB1 KS47 female NM_003073.3 c.110G>A p.Arg37His 6.88 probablydamaging

deleterious member of the SWI/SNF familyof ATP-dependent chromatin-remodeling complexes

MBD5 KS78 male NM_018328.4 c.150del p.Thr52Hisfs*31 3.64 � deleterious contains a methyl-bindingdomain that is requiredfor localization to chromatin

MTMR9 KS220 female NM_015458.3 c.310T>G p.Ser104Ala 2.09 benign tolerated myotubularin-related proteinthat is atypical becauseit has no dual-specificityphosphatase domain

NR1I3 KS220 female NM_001077482.2 c.740T>C p.Phe247Ser 3.61 probablydamaging

deleterious nuclear hormone receptor thataffects chromatin structurethrough recruitment ofchromatin-modifyingcomplexes

The following abbreviations are used: NGS, next-generation sequencing; and SWI/SNF, switch/sucrose nonfermentable.aNucleotide positions of variants are based on GRCh37/hg19.

Targeted NGS of individual KS113 revealed a nonsense

mutation (c.4441 C>T [p.Arg1481*]) in the myeloid/

lymphoid or mixed-lineage leukemia 3 gene MLL3

(NM_170606.2) (Figure 1A, Table 3, and Figure S1A). The

father of individual KS113 was deceased, but we could

confirm the absence of this mutation in the mother and

two healthy sisters, who each carried the same paternal

haplotype at the MLL3 locus as KS113 (Figure S2). This

suggests that the mutation occurred as a de novo event.

A single SMARCB1 missense mutation (c.110G>A

[p.Arg37His]; NM_003073.3), predicted to be deleterious,

was detected by targeted NGS in individual KS47 and

was absent in both parents (Figure 1B, Table 3, and

Figure S1B). Interestingly, mutations in this gene as well

as in several other genes that encode proteins of the

SWI/SNF complex were recently reported to cause Coffin

Siris syndrome (CSS [MIM 135900]).20,21 The core pheno-

type of CSS is characterized by ID, coarse facial features,

microcephaly, and a hypoplastic nail of the fifth finger

and/or toe, but the phenotype is also heterogeneous.

Germline mutations in SMARCB1 have also been reported

to predispose to familial schwannomatosis (MIM 162091)

and meningiomas (MIM 607174), but not to ID.22–24

However, these tumor-associated mutations are typically

loss-of-function alleles, and complete loss of SMARCB1

expression is seen in tumors after inactivation of the

second allele.25 In contrast, fibroblasts from individual

KS47 equally expressed wild-type and mutant alleles

(Figure S3), suggesting that substitution of the highly

conserved arginine by histidine in individual KS47 might

cause altered protein function rather than loss of function.

The SMARCB1 mutations identified in CSS are either

missense or in-frame deletions, which suggests that the

The06/06/2013 127

CSS phenotypemight also result from altered protein func-

tion.20 However, the CSS-causing mutations are located at

theC-terminusof theprotein in, or verynear, the conserved

SNF5 domain, whereas the mutation that we identified

is close to the N-terminus of the protein and is not

located in or near a conserved domain. The different loca-

tions of these mutations within SMARCB1 might explain

the differences in phenotype; however, more mutations

will need to be identified before a true genotype-phenotype

correlation can be made. All together, these data suggest

that, depending on the nature of the mutation, altered

SMARCB1 function can lead to diverse forms of ID.

Individual KS78 contains a MBD5 frameshift mutation

(c.150del [p.Thr52Hisfs*31]; NM_018328.4) that results

in a premature stop codon (Figure 1C, Table 3, and

Figure S1C). Deletions encompassing MBD5, as well as

intragenic MBD5 deletions, have previously been identi-

fied in cases with a phenotype reminiscent of Smith-

Magenis syndrome, which is characterized by ID, facial

dysmorphisms, epilepsy, and behavioral problems.26–28

Individual KS78 with the MBD5 frameshift mutation

shows a striking phenotypic overlap with these individ-

uals, underscoring the previously observed phenotypic

similarity between KSS and Smith-Magenis syndrome.5

Analysis of individual KS220 (Figure 1D) revealed three

de novo mutations (Table S5): a splice mutation (c.1318-

1G>C) affecting POF1B, a missense mutation (c.310T>G

[p.Ser104Ala]; NM_015458.3) affecting MTMR9, and

a missense mutation (c.740T>C [p.Phe247Ser]) affecting

NR1I3. The POF1B mutation is found in the NHLBI data-

base and is unlikely to cause KSS because homozygous

mutations in this gene have been reported to cause

premature ovarian failure (MIM 300604) and because


heterozygous carriers have no apparent phenotype.29 This

leaves two unique de novo variants, one in MTMR9 and

one inNR1I3 (Table 3). The amino acid change p.Ser104Ala

caused by theMTMR9mutation is considered to be benign

by PolyPhen-2 and tolerated by SIFT. Intragenic SNPs in

MTMR9 have been associated with obesity (MIM

606641).30 The p.Phe247Ser substitution caused by the

NR1I3 (NM_001077482.2) mutation (Figure 1D and

Figure S1D) is predicted to have a damaging effect on

protein function by both SIFT and PolyPhen-2.

In contrast to their function in neurodevelopment, the

biochemical functions of the genes with de novo muta-

tions are well characterized. MLL3 trimethylates histone

H3 at lysine 4 (H3K4me3) and is a central component of

the activating signal cointegrator-2 (ASC-2) complex

ASCOM, which acts as a transcriptional coactivator for

nuclear hormone receptors.31,32 SMARCB1 is a member

of the switch/sucrose nonfermentable (SWI/SNF) family

of ATP-dependent chromatin-remodeling complexes,

which affect transcription by destabilizing histone-DNA

interactions and altering nucleosome positions.33 MBD5

and its paralog MBD6 contain a methyl-binding domain,

which is required for their localization to chromatin.34

MTMR9, encoded by one of two genes with a unique de

novo mutation in individual KS220, contains a domain

that putatively interacts with SET domains, one of which

is present in EHMT1. However, MTMR9 is cytoplasmic,

and no functional relationship with chromatin regulators

has been reported.35 The other mutated gene in individual

KS220, NR1I3, has been extensively studied in liver

metabolism, but this gene is also expressed in the

brain.36,37 NR1I3 is a nuclear hormone receptor that affects

chromatin structure through recruitment of chromatin-

modifying complexes. Given the predicted damaging

function of the NR1I3 DNA variant (in contrast to the

benign prediction of the MTMR9 mutation) and the

epigenetic role of the encoded protein, we considered

this mutation to be the most likely cause of KSS in indi-

vidual KS220.

Subsequently, we sequencedMBD5, NR1I3, andMLL3 in

a cohort of 50 additional individuals with KSS. No patho-

genic sequence changes were identified, which most likely

reflects the genetic heterogeneity in KSS.

Identification of Functional Connections between

EHMT1, MLL3, SMARCB1, NR1I3, and MBD5

To address the functional relevance of the five identified

genes (MLL3, SMARCB1, MBD5, NR1I3, and MTMR9)

with de novo mutations and to further investigate their

biological relationship with EHMT1, we conducted

genetic interaction studies in Drosophila. We carried out

our experiments by modulating gene expression in the

Drosophila wing, a well-established system for genetic

interaction studies.38 Overexpression of Drosophila EHMT

in the wing consistently causes extra veins in defined

regions of the wing (Figures 2A and 2B). We tested whether

and how genetic manipulation ofMBD5,MLL3, SMARCB1,


NR1l3, and MTMR9 homologs (sba, trr, snr1, EcR, and

CG5026, respectively [see Subjects and Methods]) could

modulate this EHMT-induced wing phenotype. Expression

of sba/MBD5 alone in the wing induced mild ectopic

wing vein formation with about 50% penetrance

(Figure 2C). When sba/MBD5 was overexpressed together

with EHMT, the EHMT phenotype was strongly enhanced,

and this resulted in a consistent disruption of vein

patterning and a strong increase in ectopic vein formation

(Figure 2D). Thus, EHMT and sba induce similar mild vein

phenotypes in the fly wing, and the strongly increased

phenotypic severity elicited by their co-overexpression

indicates that the two genes genetically interact in a syner-

gistic manner. For the MLL3 ortholog trr, we observed

a striking genetic interaction with EHMT upon RNA inter-

ference (RNAi)-induced knockdown. Trr knockdown alone

in the fly wing causes a mild phenotype consisting of

a mild loss of wing veins and mild upward curvature of

the wing (Figure 2E). Combining trr/MLL3 knockdown

with EHMT overexpression resulted in fully penetrant

pupal lethality resulting from necrosis of the entire devel-

oping wing tissue (Figure 2F). This dramatic compound

phenotype indicates an antagonistic relationship between

trr/MLL3 and EHMT given that knockdown of trr/MLL3

dramatically enhances the EHMT overexpression pheno-

type. We also attempted to examine interaction of

EHMT with snr1/SMARCB1. However, knockdown of

snr1/SMARCB1 in the wing causes an extremely severe

phenotype on its own, which precludes conclusions about

potential genetic interactions with EHMT. Finally, we

investigated genetic interactions with EcR/NR1l3 and

CG5026/MTMR9, which both contain de novo missense

mutations in individual KS220 (Table 3). Quantification

of ectopic wing vein formation (Figure 2G) revealed no

effect of CG5026/MTMR9 knockdown on the EHMT over-

expression phenotype (Figure 2H), consistent with the

notion that the MTMR9 mutation does not cause KSS. In

sharp contrast, the EHMT overexpression phenotype was

almost completely rescued by heterozygous loss-of-func-

tion mutations in EcR/NR1I3 (Figure 2H), suggesting that

EHMT requires EcR/NR1I3 for its activity. Moreover, over-

expression of EcR enhanced EHMT-induced ectopic vein

formation (Figure 2H), providing strong evidence of

a synergistic relationship between EHMT and EcR/NR1I3.

These data further suggest that mutations in NR1I3, not

MTMR9, cause KSS.

Discussion

We have identified four genes—MLL3, SMARCB1, MBD5,

and NR1I3—with de novo mutations in individuals with

severe ID and with additional clinical features that closely

resemble those caused by EHMT1 defects. For three of

these genes (MLL3, MBD5, and NR1I3), we were able to

provide evidence of functional cooperation with EHMT1

by using genetic interaction experiments in Drosophila

Figure 2. Drosophila Orthologs of MBD5, MLL3, and NR1I3 Interact Genetically with EHMT(A) The morphology of the wild-type Drosophila wing is defined by five longitudinal veins (L1–L5) and the anterior and posterior crossveins (a-cv and p-cv).(B) Tissue-specific overexpression of UAS-EHMT in the Drosophilawing with the use ofMS1096-Gal4 causes ectopic wing vein formationbetween L2 and L3 in 91% of wings and between the p-cv and L5 in 88% of wings (arrows).(C) Expression of sba/MBD5 with UAS-sba in the fly wing induced mild ectopic wing vein formation posterior to L5 with about 50%penetrance (arrow).(D) In combination with UAS-EHMT, this phenotype was severely enhanced, resulting in a highly consistent disruption of normal L5formation and a massive increase in ectopic vein formation between L2 and L3 in all wings examined (arrows).(E) RNAi-mediated knockdown of trr/MLL3 by induced expression of an inverted repeat (IR) producing double-stranded RNA homolo-gous to trr (UAS-trrIR) caused mild loss of wing vein L5 and a slight upward curvature of the wing.(F) In combination with UAS-EHMT, UAS-trrIR induced pupal lethality caused by the formation of black necrotic tissue in the developingwing (arrowhead). Identical results were obtained with two individual UAS-trrIR lines (Table S7). Data is shown for UAS-trrIR1.(G) EHMT-induced ectopic wing vein formation between L2 and L3 is variable in severity and can be quantified accordingly into wild-type, mild, medium, and strong.(H) The effect of UAS-EHMT expression on ectopic vein formation in this region is rescued by RNAi-mediated knockdown of EHMTwiththe use of UAS-EHMTIR1. Similar results were obtained with two other EHMT RNAi constructs (Table S7). In contrast, CG5026/MTMR9knockdown had no effect on the EHMT-induced phenotype, as observed with three individual UAS-CG5026IR lines (Table S7). Loss-of-function mutations in EcR were able to rescue EHMT-mediated ectopic vein formation, indicating that EcR is required for thisEHMT-induced phenotype. Similar data were obtained with the EcRQ50st allele, the EcRM554fs allele, and two EcR RNAi lines (Table S7).Overexpression of EcR isoform A caused very mild induction of ectopic vein formation. However, in combination with UAS-EHMT,UAS-EcR-A strongly enhanced EHMT-induced ectopic vein formation. Similar results were obtained by overexpression of the otherEcR isoforms, B1 and B2.

(Figures 2 and 3). However, the severe phenotype of the

SMARCB1-mutant fly precluded testing for genetic inter-

actions. The established genetic interactions appear to be

very specific and thus indicative of a true biological rela-

tionship given that a number of other chromatin-related

molecules and transcription factors such as E(z), esc,

Mier1/CG1620, and FoxG/slp2 did not genetically interact

with EHMT in our assay (data not shown). Furthermore,

no interaction was found with MTMR9, one of three

de novo mutated genes in individual KSS220. Drosophila

genetic interaction studies with established disease genes

thus provide an efficient and, in our opinion, urgently

required method of discriminating between rare or even

unique benign DNA variants and causative mutations in

the NGS era.

In addition to the genetic interactions established

here, complementary evidence of molecular interactions

The06/06/2013 129

between MLL3/trr, SMARCB1/snr1, and NR1I3/EcR is

available. MLL3, a critical subunit of the ASCOM coacti-

vator complex,28 and SMARCB1, a core component of

an ATPase-dependent SWI/SNF chromatin-remodeling

complex,33 are important mediators of epigenetic regula-

tion in association with nuclear-receptor transactivation

and directly interact to mediate crosstalk between the

two complexes in which they reside.32,39 NR1I3 was iden-

tified as one of the nuclear receptors that directly associates

with the ASCOM complex.39 SMARCB1 and NR1I3 have

both been shown to interact with nuclear receptor core-

pressor 1 (NCOR1).40,41 In addition, the closest Drosophila

homolog of NR1I3, EcR, was shown to interact both phys-

ically and genetically with the Drosophila ortholog of

MLL3, trr, indicating that the mechanisms of nuclear-

receptor-mediated transcriptional activation are conserved

between mammals and flies.42 These data indicate that


Figure 3. An Epigenetic Network Underlying KSSFunctional studies indicate that genes implicated in KSS occur ina common chromatin-regulating module. This evidence comesfrom investigation of direct protein-protein interactions (solidlines) and from genetic interaction studies with Drosophila mela-nogaster (dashed lines). Green dashed lines indicate a synergisticinteraction, and red dashed lines indicate an antagonistic interac-tion. It has been demonstrated in this study that Drosophila EHMTinteracts genetically with sba/MBD5, trr/MLL3, and EcR/NR1I3(GI1, GI2, and GI3, respectively). Previously, genetic and physicalinteractions between trr and EcR (GI4 and DPI1, respectively), aswell as physical association between SMARCB1 and MLL3(DPI2), have been demonstrated.32,42

MLL3, SMARCB1, and NR1I3 cooperate with each other in

the regulation of gene transcription (Figure 3).

The molecular and here demonstrated functional rela-

tionships between EHMT1 and the other four epigenetic

regulators are mirrored by the phenotypic similarities

among individuals containing a de novo mutation in

any of the five different genes. This leads us to propose

a chromatin-modification module, defined by synergistic

and antagonistic interactions, that underlies KSS (Figure 3).

Further genetic studies in KSS individuals with intact

EHMT1, MLL3, SMARCB1, MBD5, and NR1I3 will most

likely reveal additional members of this module. Extension

of the KSS chromatin-modification module might also be

achieved by the consideration of genes implicated in over-

lapping phenotypes. This is exemplified in this study

through the identification of a mutation in MBD5, a gene

that is also disrupted in individuals with 2q23.1-deletion

syndrome (MIM 156200), which is reminiscent of Smith-

Magenis syndrome.27 Smith-Magenis syndrome and KSS

indeed have remarkably overlapping features.6 Most cases

of Smith-Magenis syndrome are caused by haploinsuffi-

ciency of RAI1, which encodes a transcription factor

that acts in conjunction with chromatin-remodeling

complexes.43,44 Finally, other proteins might be added to

the module on the basis of established functional relation-

ships despite the fact that mutations in the corresponding

genes give rise to disorders with little clinic overlap. This

was exemplified in this study through the identification

of amissensemutation in SMARCB1, which is alsomutated

in Coffin-Sirus syndrome, another genetically and pheno-

typically heterogeneous form of syndromic ID. Other

examples include the histone demethylase JARID1C and

the transcriptional regulator MED12, which both contain

mutations in nonsyndromic ID or ID syndromes without


obvious overlapping features.45,46 However, these epige-

netic regulators both have molecular connections to the

RE1-silencing transcription factor (REST; also known as

NRSF [neuron-restrictive silencer factor]) and to the

EHMT1 paralog G9a/EHMT2.47,48 It is therefore tempting

to speculate that a broader spectrum of ID conditions is

linked to the core KSS chromatin module.

The identification of a chromatin-modification module

underlying ID is of particular interest given recent

evidence showing that epigenetic processes are important

in acute cognitive functioning.4 Adult rescue of cognitive

deficits has been accomplished in several animal ID

models,42–51 including EHMT-mutant flies,19 raising hope

that cognition can be improved postnatally. Further dissec-

tion of the KSS chromatin-regulating module through

mechanistic studies and continued elucidation of its

genetic etiology might allow for fundamental insights

and for the identification of drugs that improve cognition

in this group of genetically heterogeneous but phenotypi-

cally similar individuals. In this respect, we note that

a number of the ID-associated epigenetic regulators, such

as MLL3 and SMARCB1, have also been linked to

cancer.23,52 Research into the associated tumorigenic path-

ways and their applicable drugs might be relevant to neu-

rodevelopmental pathways as well, as was recently shown

for topoisomerase inhibitors in a mouse model for Angel-

man syndrome.53

In summary, the work presented here defines and

characterizes an epigenetic module underlying human

cognitive disorders and underscores the importance of

tight epigenetic control mechanisms in higher brain

function. In addition, the identification of an EHMT-associ-

ated epigeneticmodule, including antagonistic players that

could serve as potential drug targets, is a step toward devel-

oping a strategy to correct for cognitive deficits associated

with this genetically heterogeneous group of ID disorders.

Supplemental Data

Supplemental Data include three figures and seven tables and can

be found with this article online at http://www.cell.com/AJHG.

Acknowledgments

We thank all family members who participated in this study. We

would also like to thank the Bloomington Drosophila Stock Center

and the Vienna Drosophila RNAi Center for providing fly stocks.

We are grateful to Martijn Huynen for bioinformatics analyses

and discussions. This project was supported by the EU FP7

Large-Scale Integrating Project Genetic and Epigenetic Networks

in Cognitive Dysfunction (241995 to A.S. and H.v.B.), the EU

FP7 project TECHGENE (grant agreement 223143 to J.A.V.), the

grant from Hersenstichting Nederland (2009 [1]-122 to T. Kleef-

stra), the ZonMw (Zorg Onderzoek Nederland) Clinical Fellowship

(990700365 to T. Kleefstra), and a Netherlands Organization for

Scientific Research Vidi grant (917-96-346) to A.S. Next-genera-

tion-sequencing experiments were financially supported by the

Department of Human Genetics, Nijmegen.

http://www.cell.com/AJHG

Received: February 27, 2012

Revised: April 10, 2012

Accepted: May 14, 2012

Published online: June 21, 2012

Web Resources

The URLs for data presented herein are as follows:

AmiGO Gene Ontology database, http://amigo.geneontology.org/

cgi-bin/amigo/go.cgi

NHLBI Exome Variant Server, http://snp.gs.washington.edu/EVS/

Online Mendelian Inheritance in Man (OMIM), http://www.

omim.org

PolyPhen-2, http://genetics.bwh.harvard.edu/pph2/

Primer3, http://frodo.wi.mit.edu/

SIFT, http://sift.jcvi.org/

STRING database, http://string.embl.de/

UCSC Genome Bioinformatics, http://www.genome.ucsc.edu/

References

1. Ropers, H.H. (2010). Genetics of early onset cognitive impair-

ment. Annu. Rev. Genomics Hum. Genet. 11, 161–187.

2. van Bokhoven, H. (2011). Genetic and epigenetic networks in

intellectual disabilities. Annu. Rev. Genet. 45, 81–104.

3. van Bokhoven, H., and Kramer, J.M. (2010). Disruption of the

epigenetic code: An emerging mechanism in mental retarda-

tion. Neurobiol. Dis. 39, 3–12.

4. Day, J.J., and Sweatt, J.D. (2011). Epigenetic mechanisms in

cognition. Neuron 70, 813–829.

5. Stewart, D.R., and Kleefstra, T. (2007). The chromosome 9q

subtelomere deletion syndrome. Am. J. Med. Genet. C. Semin.

Med. Genet. 145C, 383–392.

6. Kleefstra, T., Brunner, H.G., Amiel, J., Oudakker, A.R., Nillesen,

W.M., Magee, A., Genevieve, D., Cormier-Daire, V., van Esch,

H., Fryns, J.P., et al. (2006). Loss-of-function mutations in

euchromatin histone methyl transferase 1 (EHMT1) cause

the 9q34 subtelomeric deletion syndrome. Am. J. Hum. Genet.

79, 370–377.

7. Kleefstra, T., van Zelst-Stams, W.A., Nillesen, W.M., Cormier-

Daire, V., Houge, G., Foulds, N., van Dooren, M., Willemsen,

M.H., Pfundt, R., Turner, A., et al. (2009). Further clinical

and molecular delineation of the 9q subtelomeric deletion

syndrome supports a major contribution of EHMT1 haploin-

sufficiency to the core phenotype. J. Med. Genet. 46, 598–606.

8. Tachibana, M., Ueda, J., Fukuda, M., Takeda, N., Ohta, T.,

Iwanari, H., Sakihama, T., Kodama, T., Hamakubo, T., and

Shinkai, Y. (2005). Histone methyltransferases G9a and GLP

form heteromeric complexes and are both crucial for methyl-

ation of euchromatin at H3-K9. Genes Dev. 19, 815–826.

9. Collins, R.E., Northrop, J.P., Horton, J.R., Lee, D.Y., Zhang, X.,

Stallcup, M.R., and Cheng, X. (2008). The ankyrin repeats

of G9a and GLP histone methyltransferases are mono- and

dimethyllysine binding modules. Nat. Struct. Mol. Biol. 15,

245–250.

10. Willemsen, M.H., Vulto-van Silfhout, A.T., Nillesen, W.M.,

Wissink-Lindhout, W.M., van Bokhoven, H., Philip, N.,

Berry-Kravis, E.M., Kini, U., van Ravenswaaij-Arts, C.M., Delle

Chiaie, B., et al. (2012). Update on Kleefstra Syndrome. Mol.

Syndromol. 2, 202–212.

The06/06/2013 131

11. Hoischen, A., Gilissen, C., Arts, P., Wieskamp, N., van der

Vliet, W., Vermeer, S., Steehouwer, M., de Vries, P., Meijer,

R., Seiqueros, J., et al. (2010). Massively parallel sequencing

of ataxia genes after array-based enrichment. Hum. Mutat.

31, 494–499.

12. Vissers, L.E., de Ligt, J., Gilissen, C., Janssen, I., Steehouwer,

M., de Vries, P., van Lier, B., Arts, P.,Wieskamp, N., del Rosario,

M., et al. (2010). A de novo paradigm for mental retardation.

Nat. Genet. 42, 1109–1112.

13. Bredrup, C., Saunier, S., Oud,M.M., Fiskerstrand, T., Hoischen,

A., Brackman, D., Leh, S.M., Midtbø, M., Filhol, E., Bole-

Feysot, C., et al. (2011). Ciliopathies with skeletal anomalies

and renal insufficiency due to mutations in the IFT-A gene

WDR19. Am. J. Hum. Genet. 89, 634–643.

14. Brand, A.H., and Perrimon, N. (1993). Targeted gene expres-

sion as a means of altering cell fates and generating dominant

phenotypes. Development 118, 401–415.

15. Dietzl, G., Chen, D., Schnorrer, F., Su, K.C., Barinova, Y.,

Fellner, M., Gasser, B., Kinsey, K., Oppel, S., Scheiblauer, S.,

et al. (2007). A genome-wide transgenic RNAi library for condi-

tional gene inactivation in Drosophila. Nature 448, 151–156.

16. Inlow, J.K., and Restifo, L.L. (2004). Molecular and compara-

tive genetics of mental retardation. Genetics 166, 835–881.

17. Li, H., Coghlan, A., Ruan, J., Coin, L.J., Heriche, J.K., Osmoth-

erly, L., Li, R., Liu, T., Zhang, Z., Bolund, L., et al. (2006).

TreeFam: A curated database of phylogenetic trees of animal

gene families. Nucleic Acids Res. 34 (Database issue), D572–

D580.

18. Schultz, J., Milpetz, F., Bork, P., and Ponting, C.P. (1998).

SMART, a simple modular architecture research tool: Identifi-

cation of signaling domains. Proc. Natl. Acad. Sci. USA 95,

5857–5864.

19. Kramer, J.M., Kochinke, K., Oortveld, M.A., Marks, H., Kramer,

D., de Jong, E.K., Asztalos, Z., Westwood, J.T., Stunnenberg,

H.G., Sokolowski, M.B., et al. (2011). Epigenetic regulation

of learning and memory by Drosophila EHMT/G9a. PLoS

Biol. 9, e1000569.

20. Tsurusaki, Y.,Okamoto,N.,Ohashi,H., Kosho, T., Imai, Y.,Hibi-

Ko, Y., Kaname, T., Naritomi, K., Kawame, H., Wakui, K., et al.

(2012). Mutations affecting components of the SWI/SNF

complex causeCoffin-Siris syndrome.Nat.Genet.44, 376–378.

21. Santen, G.W., Aten, E., Sun, Y., Almomani, R., Gilissen, C.,

Nielsen, M., Kant, S.G., Snoeck, I.N., Peeters, E.A., Hilhorst-

Hofstee, Y., et al. (2012). Mutations in SWI/SNF chromatin

remodeling complex gene ARID1B cause Coffin-Siris syn-

drome. Nat. Genet. 44, 379–380.

22. Hadfield, K.D., Newman, W.G., Bowers, N.L., Wallace, A.,

Bolger, C., Colley, A., McCann, E., Trump, D., Prescott, T.,

and Evans, D.G. (2008). Molecular characterisation of

SMARCB1 and NF2 in familial and sporadic schwannomato-

sis. J. Med. Genet. 45, 332–339.

23. Christiaans, I., Kenter, S.B., Brink, H.C., van Os, T.A., Baas, F.,

van den Munckhof, P., Kidd, A.M., and Hulsebos, T.J. (2011).

Germline SMARCB1 mutation and somatic NF2 mutations

in familial multiple meningiomas. J. Med. Genet. 48, 93–97.

24. Rousseau, G., Noguchi, T., Bourdon, V., Sobol, H., and

Olschwang, S. (2011). SMARCB1/INI1 germlinemutations con-

tribute to10%of sporadic schwannomatosis. BMCNeurol.11, 9.

25. Eaton, K.W., Tooke, L.S., Wainwright, L.M., Judkins, A.R., and

Biegel, J.A. (2011). Spectrum of SMARCB1/INI1 mutations in

familial and sporadic rhabdoid tumors. Pediatr. Blood Cancer

56, 7–15.


http://amigo.geneontology.org/cgi-bin/amigo/go.cgi

http://amigo.geneontology.org/cgi-bin/amigo/go.cgi

http://snp.gs.washington.edu/EVS/

http://www.omim.org

http://www.omim.org

http://genetics.bwh.harvard.edu/pph2/

http://frodo.wi.mit.edu/

http://sift.jcvi.org/

http://string.embl.de/

http://www.genome.ucsc.edu/

26. Jaillard, S.,Dubourg,C.,Gerard-Blanluet,M.,Delahaye,A., Pas-

quier, L., Dupont, C., Henry, C., Tabet, A.C., Lucas, J., Aboura,

A., et al. (2009). 2q23.1 microdeletion identified by array

comparative genomic hybridisation: An emerging phenotype

with Angelman-like features? J. Med. Genet. 46, 847–855.

27. van Bon, B.W., Koolen, D.A., Brueton, L., McMullan, D., Lich-

tenbelt, K.D., Ades, L.C., Peters, G., Gibson, K., Moloney, S.,

Novara, F., et al. (2010). The 2q23.1 microdeletion syndrome:

Clinical and behavioural phenotype. Eur. J. Hum. Genet. 18,

163–170.

28. Talkowski, M.E., Mullegama, S.V., Rosenfeld, J.A., van Bon,

B.W., Shen, Y., Repnikova, E.A., Gastier-Foster, J., Thrush,

D.L., Kathiresan, S., Ruderfer, D.M., et al. (2011). Assessment

of 2q23.1 microdeletion syndrome implicates MBD5 as

a single causal locus of intellectual disability, epilepsy, and

autism spectrum disorder. Am. J. Hum. Genet. 89, 551–563.

29. Lacombe, A., Lee, H., Zahed, L., Choucair, M., Muller, J.M.,

Nelson, S.F., Salameh, W., and Vilain, E. (2006). Disruption of

POF1B binding to nonmuscle actin filaments is associated with

premature ovarian failure. Am. J. Hum. Genet. 79, 113–119.

30. Yanagiya, T., Tanabe, A., Iida, A., Saito, S., Sekine, A., Takaha-

shi, A., Tsunoda, T., Kamohara, S., Nakata, Y., Kotani, K., et al.

(2007). Association of single-nucleotide polymorphisms in

MTMR9 gene with obesity. Hum. Mol. Genet. 16, 3017–3026.

31. Goo, Y.H., Sohn, Y.C., Kim, D.H., Kim, S.W., Kang, M.J., Jung,

D.J., Kwak, E., Barlev, N.A., Berger, S.L., Chow, V.T., et al.

(2003). Activating signal cointegrator 2 belongs to a novel

steady-state complex that contains a subset of trithorax group

proteins. Mol. Cell. Biol. 23, 140–149.

32. Lee, S., Kim, D.H., Goo, Y.H., Lee, Y.C., Lee, S.K., and Lee, J.W.

(2009). Crucial roles for interactions between MLL3/4 and

INI1 in nuclear receptor transactivation. Mol. Endocrinol.

23, 610–619.

33. Wilson, B.G., and Roberts, C.W. (2011). SWI/SNF nucleosome

remodellers and cancer. Nat. Rev. Cancer 11, 481–492.

34. Laget, S., Joulie, M., Le Masson, F., Sasai, N., Christians, E.,

Pradhan, S., Roberts, R.J., and Defossez, P.A. (2010). The

human proteins MBD5 and MBD6 associate with heterochro-

matin but they do not bind methylated DNA. PLoS ONE 5,

e11982.

35. Laporte, J., Bedez, F., Bolino, A., and Mandel, J.L. (2003).

Myotubularins, a large disease-associated family of cooperat-

ing catalytically active and inactive phosphoinositides phos-

phatases. Hum. Mol. Genet. 12 (Spec No 2), R285–R292.

36. Nishimura, M., Naito, S., and Yokoi, T. (2004). Tissue-specific

mRNA expression profiles of human nuclear receptor subfam-

ilies. Drug Metab. Pharmacokinet. 19, 135–149.

37. Dutheil, F., Dauchy, S., Diry, M., Sazdovitch, V., Cloarec, O.,

Mellottee, L., Bieche, I., Ingelman-Sundberg, M., Flinois, J.P.,

de Waziers, I., et al. (2009). Xenobiotic-metabolizing enzymes

and transporters in the normal human brain: Regional and

cellular mapping as a basis for putative roles in cerebral func-

tion. Drug Metab. Dispos. 37, 1528–1538.

38. Bier, E. (2005). Drosophila, the golden bug, emerges as a tool

for human genetics. Nat. Rev. Genet. 6, 9–23.

39. Choi, E., Lee, S., Yeom, S.Y., Kim, G.H., Lee, J.W., and Kim,

S.W. (2005). Characterization of activating signal cointegra-

tor-2 as a novel transcriptional coactivator of the xenobiotic

nuclear receptor constitutive androstane receptor. Mol. Endo-

crinol. 19, 1711–1719.


40. Underhill, C., Qutob, M.S., Yee, S.P., and Torchia, J. (2000). A

novel nuclear receptor corepressor complex, N-CoR, contains

components of the mammalian SWI/SNF complex and the

corepressor KAP-1. J. Biol. Chem. 275, 40463–40470.

41. Jyrkkarinne, J., Makinen, J., Gynther, J., Savolainen, H., Poso,

A., and Honkakoski, P. (2003). Molecular determinants of

steroid inhibition for the mouse constitutive androstane

receptor. J. Med. Chem. 46, 4687–4695.

42. Sedkov, Y., Cho, E., Petruk, S., Cherbas, L., Smith, S.T., Jones,

R.S., Cherbas, P., Canaani, E., Jaynes, J.B., and Mazo, A.

(2003). Methylation at lysine 4 of histone H3 in ecdysone-

dependent development of Drosophila. Nature 426, 78–83.

43. Bi,W., Saifi, G.M., Shaw, C.J.,Walz, K., Fonseca, P.,Wilson,M.,

Potocki, L., and Lupski, J.R. (2004). Mutations of RAI1, a PHD-

containing protein, in nondeletion patients with Smith-

Magenis syndrome. Hum. Genet. 115, 515–524.

44. Darvekar, S., Johnsen, S.S., Eriksen, A.B., Johansen, T., and

Sjøttem, E. (2012). Identification of two independent nucleo-

some-binding domains in the transcriptional co-activator

SPBP. Biochem. J. 442, 65–75.

45. Jensen, L.R., Amende, M., Gurok, U., Moser, B., Gimmel, V.,

Tzschach, A., Janecke, A.R., Tariverdian, G., Chelly, J., Fryns,

J.P., et al. (2005). Mutations in the JARID1C gene, which is

involved in transcriptional regulation and chromatin remod-

eling, cause X-linked mental retardation. Am. J. Hum. Genet.

76, 227–236.

46. Risheg, H., Graham, J.M., Jr., Clark, R.D., Rogers, R.C., Opitz,

J.M., Moeschler, J.B., Peiffer, A.P., May, M., Joseph, S.M., Jones,

J.R., et al. (2007). A recurrent mutation in MED12 leading to

R961W causes Opitz-Kaveggia syndrome. Nat. Genet. 39,

451–453.

47. Ding, N., Zhou, H., Esteve, P.O., Chin, H.G., Kim, S., Xu, X.,

Joseph, S.M., Friez, M.J., Schwartz, C.E., Pradhan, S., and

Boyer, T.G. (2008). Mediator links epigenetic silencing of

neuronal gene expression with x-linked mental retardation.

Mol. Cell 31, 347–359.

48. Tahiliani, M., Mei, P., Fang, R., Leonor, T., Rutenberg, M.,

Shimizu, F., Li, J., Rao, A., and Shi, Y. (2007). The histone

H3K4 demethylase SMCX links REST target genes to X-linked

mental retardation. Nature 447, 601–605.

49. Yan, Q.J., Rammal, M., Tranfaglia, M., and Bauchwitz, R.P.

(2005). Suppression of two major Fragile X Syndrome mouse

model phenotypes by the mGluR5 antagonist MPEP. Neuro-

pharmacology 49, 1053–1066.

50. McBride, S.M., Choi, C.H., Wang, Y., Liebelt, D., Braunstein,

E., Ferreiro, D., Sehgal, A., Siwicki, K.K., Dockendorff, T.C.,

Nguyen, H.T., et al. (2005). Pharmacological rescue of synaptic

plasticity, courtship behavior, and mushroom body defects

in a Drosophila model of fragile X syndrome. Neuron 45,

753–764.

51. Cobb, S., Guy, J., and Bird, A. (2010). Reversibility of func-

tional deficits in experimental models of Rett syndrome. Bio-

chem. Soc. Trans. 38, 498–506.

52. Gui, Y., Guo, G., Huang, Y., Hu, X., Tang, A., Gao, S., Wu, R.,

Chen, C., Li, X., Zhou, L., et al. (2011). Frequent mutations

of chromatin remodeling genes in transitional cell carcinoma

of the bladder. Nat. Genet. 43, 875–878.

53. Huang, H.S., Allen, J.A., Mabb, A.M., King, I.F., Miriyala, J.,

Taylor-Blake, B., Sciaky, N., Dutton, J.W., Jr., Lee, H.M.,

Chen, X., et al. (2012). Topoisomerase inhibitors unsilence

the dormant allele of Ube3a in neurons. Nature 481, 185–189.

Tilmann Peter

The Role of Transforming Growth Factor-B–Mediated Tumor-Stroma Interactions in Prostate Cancer Progression: An Integrative Approach Quelle: Cancer Res 2009;69:7111-7120. Published Online First August 25, 2009 Betreuer: Prof. Martin Burchardt (Urologie) Was bedeutet mir das Thema persönlich? Das Thema steht in engem Zusammenhang zu meinem Promotionsthema und ermöglicht mir ein besseres Verständnis der Progression und Proliferation des Prostatakarzinoms. Hierbei werden nicht nur einzelne Zellen und ihre Signalkaskaden behandelt, sondern die Interaktion der Tumorzellen mit dem umliegenden Stroma stehen im Mittelpunkt der Betrachtung. Worauf kommt es mir am meisten an? Wichtig war mir, ein Thema zu finden, welches nicht nur von persönlichem Interesse ist, sondern meine Kommilitonen ebenfalls begeistern könnte. Das Modell, welches die Autoren beim PCa nutzen, hat bereits bei anderen Tumorerkrankungen Anwendung erfahren. Dementsprechend regt es sicher innerhalb des Journal-Clubs zur Diskussion an. Was fasziniert mich an diesem Thema? Was mich fasziniert ist die Idee, ein mathematisches 3 - D Modell auf den Tumor und sein umliegendes Stroma zu übertragen. Dies mag durchaus wichtige Aufschlüsse geben, die man nicht erwartet hätte. Was gefällt mir an dem Thema weniger? Im Moment kann ich nicht sagen, was mir an diesem Thema nicht gefällt.

06/06/2013 133

2009;69:7111-7120. Published OnlineFirst August 25, 2009.Cancer Res David Basanta, Douglas W. Strand, Ralf B. Lukner, et al. An Integrative ApproachTumor-Stroma Interactions in Prostate Cancer Progression:

Mediated−βThe Role of Transforming Growth Factor-

Updated version

10.1158/0008-5472.CAN-08-3957doi:

Access the most recent version of this article at:

Material

Supplementary

lhttp://cancerres.aacrjournals.org/content/suppl/2010/05/19/0008-5472.CAN-08-3957.DC2.htmmlhttp://cancerres.aacrjournals.org/content/suppl/2009/08/03/0008-5472.CAN-08-3957.DC1.htAccess the most recent supplemental material at:

Cited Articles

http://cancerres.aacrjournals.org/content/69/17/7111.full.html#ref-list-1

This article cites by 47 articles, 15 of which you can access for free at:

Citing articles

http://cancerres.aacrjournals.org/content/69/17/7111.full.html#related-urls

This article has been cited by 4 HighWire-hosted articles. Access the articles at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected] at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

Permissions

[email protected] at

To request permission to re-use all or part of this article, contact the AACR Publications

Research. on May 29, 2013. © 2009 American Association for Cancercancerres.aacrjournals.org Downloaded from

Published OnlineFirst August 25, 2009; DOI: 10.1158/0008-5472.CAN-08-3957

06/06/2013 134

http://cancerres.aacrjournals.org/lookup/doi/10.1158/0008-5472.CAN-08-3957

http://cancerres.aacrjournals.org/content/suppl/2009/08/03/0008-5472.CAN-08-3957.DC1.html http://cancerres.aacrjournals.org/content/suppl/2010/05/19/0008-5472.CAN-08-3957.DC2.html

http://cancerres.aacrjournals.org/content/69/17/7111.full.html#ref-list-1

http://cancerres.aacrjournals.org/content/69/17/7111.full.html#related-urls

http://cancerres.aacrjournals.org/cgi/alerts



http://cancerres.aacrjournals.org/

The Role of Transforming Growth Factor-B–Mediated

Tumor-Stroma Interactions in Prostate Cancer

Progression: An Integrative Approach

David Basanta,1Douglas W. Strand,

3Ralf B. Lukner,

6Omar E. Franco,

3David E. Cliffel,

5

Gustavo E. Ayala,6Simon W. Hayward,

2,3,4and Alexander R.A. Anderson

1

1Integrated Mathematical Oncology, H. Lee Moffitt Cancer Center and Research Institute, Tampa, Florida; Departments of 2Cancer Biologyand 3Urologic Surgery and 4The Vanderbilt-Ingram Cancer Center, Vanderbilt University Medical Center; 5Department of Chemistry,Vanderbilt University, Nashville, Tennessee; and 6Baylor College of Medicine, Houston, Texas

Abstract

We have implemented a hybrid cellular automata model basedon the structure of human prostate that recapitulates keyinteractions in nascent tumor foci between tumor cells andadjacent stroma. Model simulations show how stochasticinteractions between tumor cells and stroma may lead to astructural suppression of tumor growth, modest proliferation,or unopposed tumor growth. The model incorporates keyaspects of prostate tumor progression, including transforminggrowth factor-B (TGF-B), matrix-degrading enzyme activity,and stromal activation. It also examines the importance ofTGF-B during tumor progression and the role of stromal celldensity in regulating tumor growth. The validity of one of thekey predictions of the model about the effect of epithelial TGF-B production on glandular stability was tested in vivo . Theseexperimental results confirmed the ability of the model togenerate testable biological predictions in addition toproviding new avenues of experimental interest. This workunderscores the need for more pathologically representativemodels to cooperatively drive computational and biologicalmodeling, which together could eventually lead to moreaccurate diagnoses and treatments of prostate cancer. [CancerRes 2009;69(17):7111–20]

Introduction

Cellular automata (CA) models have been used to examinethree-dimensional growth as well as the response to radiationtreatment of brain tumors, such as glioblastoma multiforme (1, 2),with results that agreed well with clinical imaging data.Increasingly, CA models have been applied to more complextumor growth and treatment problems (3, 4). For example,computational models are capable of both selecting effectivetreatments and screening out ineffective treatments for lethalmelanomas and glioblastoma multiforme (5).Computational models of avascular and vascular tumors have

been extensively reported and reviewed previously (6, 7). Suchmodels have been used to show how harsh growth conditions, suchas hypoxia, can promote evolution of invasive tumors, and theresults have been related in a preliminary manner to in vitro and

in vivo tumor growth experiments (8). Tumor growth modelsreported in the literature also consider the role of nutrients,oxygen, necrosis at low pH, aggregation kinetics in the growth,genetic and phenotypical evolution, and morphology of thegrowing tumor (7, 9). Discrete cell simulations are typicallyvalidated by carefully controlled in vitro experiments (10) becausethey can be designed to effectively collect the vast amount ofinformation associated with computational models.CAmodels are based on discrete representation of space and time

on a lattice where only neighboring lattice elements interact. Onevariant of the CA method is the hybrid discrete continuous CA(HCA), where continuous equations and discrete elements aresolved together on the same grid. Several other tumor types havebeen accurately modeled using such methods. HCA models havebeen extensively described in the literature (8, 11–14). Althoughmost of the previously described models neglect intracellularfeatures, they are well suited to study biological phenomena at thecellular level due to the simplicity of the logic, flexibility of thetechnique, and the ability to integrate multiple interacting variablesacross a range of spatial scales.The prostate is a glandular sexual accessory organ composed of

acinar ducts lined with luminal secretory epithelium surrounded bya layer of basal epithelial cells. These epithelial acini areencompassed by a stromal compartment composed predominantlyof layers of smooth muscle (see Fig. 1A). The principal cellularcomponent of the prostate is illustrated in Fig. 1B .Prostatic adenocarcinoma is the second most common cause of

male cancer deaths in the Western world (15). Early discriminationbetween relatively benign lesions and highly aggressive prostateadenocarcinomas is critical for identifying those patients thatrequire aggressive treatment while avoiding overtreating patientswho would otherwise suffer no ill effects from their tumor.In the adult prostate, paracine cross-talk between the epithelial

and surrounding stromal tissues maintains homeostasis (16, 17).The smooth muscle and all other stromal cell types are separatedfrom the glandular acini by a collagen- and laminin-rich basementmembrane that provides positional information contributing to themaintenance of tissue architecture and differentiation throughcellular signaling and structural constraint (18).The loss of homeostatic interactions between organ tissues in

disease has been partially attributed to a breakdown of thepositional information established during development, whichincludes the loss of the basement membrane (19, 20) and analteration of the density and type of extracellular matrix (ECM;ref. 21). This matrix is produced by an ever-expanding populationof myofibroblasts (22). Although there is evidence that stronglyimplicates the role of the basement membrane and the stromal

Note: Supplementary data for this article are available at Cancer Research Online(http://cancerres.aacrjournals.org/).

Requests for reprints: David Basanta, Integrated Mathematical Oncology, H. LeeMoffitt Cancer Center and Research Institute, 12902 Magnolia Drive, Tampa, FL 33612.Phone: 813-745-6433; Fax: 813-745-6433; E-mail: [email protected].

I2009 American Association for Cancer Research.doi:10.1158/0008-5472.CAN-08-3957

www.aacrjournals.org 7111 Cancer Res 2009; 69: (17). September 1, 2009

Systems Biology and Emerging Technologies



06/06/2013 135


Figure 1. A, H&E-stained cross-sections of human prostate at medium magnification showing, from left to right, normal prostate tissue, PIN, and high-grade prostaticadenocarcinoma. In the center panel, PINs are premalignant consisting of atypical luminal epithelial cells that may develop into the invasive cancer shown in theright panel. Prostate cancer is characterized by small disorganized glands without basal cells invading into the surrounding stroma. B, model domain with its keycell types. The basal cells produce TGF-h and help maintain homeostasis. Luminal cells consume TGF-h and can become tumorigenic. The stromal cells occupylocations outside the acini and produce TGF-h in response to the TGF-h once it reaches a certain concentration. C, image with the initial configuration. The simulationdomain is made of three glands, equally spaced and surrounded by stromal cells.

Cancer Research

Cancer Res 2009; 69: (17). September 1, 2009 7112 www.aacrjournals.org



06/06/2013 136


microenvironment on glandular integrity in tumor progression,the conflicting data and the vast number of factors involved in theregulation of these components limit our understanding of themultiple steps by which prostate tumors grow and invadesurrounding tissues.The effect of transforming growth factor-h (TGF-h) on cancer

progression has been the target of much research and debate (23).TGF-h normally inhibits the proliferation of epithelia throughinduction of the cell cycle inhibitors p15 and p21 (24). Thedetermination of whether TGF-h will induce cytostasis orapoptosis in normal epithelia depends on the intensity of theirproliferative activity in addition to poorly understood micro-environmental determinants (23, 25). Stromal production of TGF-hby prostate carcinoma–associated fibroblasts has been shown toincrease the growth and invasiveness of initiated prostate epithelia(26); however, the effect of epithelial TGF-h production fromorganized prostate glands is still unclear. Moreover, systemicinhibitors of TGF-h have yielded conflicting data in therapeutictrials (27), underscoring the need for further analysis of the complexroles of TGF-h at different stages of neoplastic progression.The TGF-h family of cytokines is highly pleiotropic. Some of

their functions have been accurately modeled computationally,including roles in vascular remodeling, hyperplasia, and woundrepair (28). Continuum ordinary differential equation tissue growthmodels of TGF-h–mediated stromal-epithelial interactions wereshown to have good qualitative agreement with experimentalbiological results (29).Prostate tumors have previously been investigated mathemati-

cally (30–34), mainly focusing on the role of prostate-specificantigen and androgen levels in progression. Although insightful,these models were purely continuous and nonspatial and did notconsider the role of stroma. In this article, we model prostatetumorigenesis using a HCA model that integrates five different cellspecies (discrete, individuals) with three different microenviron-mental chemical species (continuous, concentrations), all of whichare thought to play key roles in prostate cancer. Using this model,we have investigated the importance of TGF-h in driving prostatecancer progression. In particular, we examined how TGF-hregulates tumor-stroma interactions in tissues with tumor cellspossessing different degrees of organization (as measured in termsof glandular morphology). In addition, following a predictiongenerated by the model about the role of TGF-h in organizedglandular epithelia, we have provided previously unexaminedpathologic and experimental support for the ability of the modelto appropriately reflect the physiologic role of TGF-h at a specificstage of tumorigenesis.


HCA models are mathematical tools in which discrete entities interact

with continuous ones. In the model, cells are the discrete entitiesrepresented as points on a two-dimensional grid (2 mm � 2 mm slice of

a three-dimensional prostate). This grid structure has been used in other

cancer-related CA models (7, 35, 36) and has proved to be very efficient.

This grid hosts three microenvironmental variables, which are treated ascontinuous concentrations: TGF-h, matrix-degrading enzyme (MDE)

expression, and membrane/ECM. The discrete cells are designated as basal

epithelial, luminal epithelial, motile stroma (representing bone marrow–derived cells such as macrophages), static stroma ( fibromuscular stroma),

and tumor cells. Figure 1B shows how cells are initially arranged: the

simulated section of prostate contains three glands (Supplementary Fig. S1),

each following the structure shown in Fig. 1B and arranged along an axis

going from the upper-left to the lower-right corner of the domain (as shownin Supplementary Fig. S1). The space outside the glands is either empty or

occupied by static and motile stromal cells. Static cells can represent

muscle or fibroblastic lineages, whereas motile cells are predominantly

modeled on bone marrow–derived monocyte/macrophage lineages.Figure 2 shows flowcharts that describe the behavior of the different cell

types. A detailed explanation of these flowcharts can be found in the

Supplementary Appendix. The microenvironmental variables of the model

are TGF-h, MDE expression, and membrane/ECM concentrations (repre-senting both the ECM, which is present in the mesenchyme and is made of

elements such as collagen, fibronectin, laminin, and vitronectin, and the

epithelial basement membrane). The dynamics of these microenvironmen-

tal variables are defined by three partial differential equations that describehow each of them evolves in space and time. The dynamics of TGF-h (Th)

are as follows:

which shows that TGF-h is produced by basal and cancer cells as well as bymotile stromal cells in proportion to the local TGF-h concentration. It alsoshows that TGF-h is consumed by luminal and motile stroma cells. TGF-halso binds to the ECM at a rate that depends on the local concentration of

TGF-h and also that there is some natural decay of the ligand.MDEs (E) are produced by tumor cells (at rate k ; ref. 37), diffuse (at rate

DE), and are depleted as they degrade the ECM and the basement

membrane (l).

Basement membrane/ECM (M) is produced by basal cells (depending on

the current local concentration of ECM ensuring the density never exceeds

the maximum m0) and motile stroma (depending on rate aF, scaled by thelocal concentration of TGF-h; ref. 38). Finally, the ECM gets degraded by the

MDEs at a rate l :

These microenvironmental factors (equations above), coupled with the

discrete cells (defined by the cellular life cycle flowcharts), represent anevolving dynamic hybrid model of the prostate in which tumor progression

emerges from the interactions of all the key elements. Cell behavior results

from the interactions with other cells in competition for space and via the

microenvironmental factors as shown in Fig. 2F . It can be seen that TGF-his a key factor that determines the survival and growth of both tumor and

normal epithelial cells as well as the behavior of stromal cells. The balance

of TGF-h, MDEs, and membrane production modulates the competition ofthe different cellular species.

Time is discrete and is measured in time steps of 24 h. At each time step,

cells follow their life cycle (Fig. 2). The timescale affecting the micro-

environmental variables is assumed to be much faster (�100) than that ofthe cells to accommodate the different timescales affecting processes at the

cellular and molecular levels.

The parameters used to characterize the model are assigned the values

shown in Table 1. Many of the values used are estimates, denoted by anasterisk (39–41), which are biologically plausible but not in every case

experimentally measured. Because the parameters need to be non-

dimensionalized, the results of the nondimensionalization (with respectto time, which is taken to have cell cycles of 24 h, and space of 2 mm � 2

Integrated Investigation of Prostate Cancer Progression




06/06/2013 137


Cancer Research




06/06/2013 138


mm) are also given. These parameters have the property of making thesystem homeostatic before cancer initiation.

Therefore, under normal circumstances (in the absence of tumor cells),

there is a natural homeostatic state for all the variables with equilibrium

concentrations of TGF-h, ECM, and MDE. These represent the initialconditions for each of these microenvironmental variables (Supplementary

Fig. S1).

Results

Because the composition of the stromal compartment isstatistically correlated with prostate disease recurrence (42), therole of stroma on tumors at different stages of malignancy(determined by TGF-h and MDE production) was studied. Each ofthe simulations was allowed to run for 20,000 time steps,corresponding to f55 years in real time. After 10 time steps, sixbasal epithelial cells ( four in the central acinus and one on each ofthe two remaining acini) become abnormal cells initiatingtumorigenesis. The simulations test different stromal configura-tions in which the proportion of stroma in the domain ranges from20% to 40% of the total space. The stromal configurationsconsidered are (motile, nonmotile) as follows: (a) high proportionof motile and nonmotile stroma (40%, 40%), (b) high proportion ofmotile stroma (40%, 10%), (c) high proportion of nonmotile stroma(10%, 40%), and (d) low proportion of motile and nonmotile stroma(10%, 10%).Sample simulations. Figure 3 (and Supplementary Movies)

shows the three main outcomes resulting from the simulations.Figure 3A shows an example of a simulation in which the tumorbreaks out from the acini and grows into the surrounding stroma.Initially, the tumor cells fill the lumen inside the gland (Fig. 3A).Eventually, the basement membrane starts to degrade and TGF-hbegins to leak from the gland and attract motile stroma (Fig. 3B).At the end of the simulation, the tumor has taken a significantportion of the domain but its growth is constrained by the motilestroma responding to the TGF-h gradient (Fig. 3C). As a result ofperiodic boundary conditions, tumor cells growing on the upperboundary appear on the bottom; however, this has no effect on thesimulation outcome. Figure 3B shows a simulation in which tumorcells quickly break from the gland and grow unopposed. Finally,Fig. 3C shows tumor cells producing excessive quantities of MDE,which leads to early breakdown of the basement membrane andresulting leakage of the TGF-h, without which the tumor cells die.Varying the microenvironment. To further investigate the

three distinct outcomes the model can produce, we examinedthe effect of varying each of the three microenvironmentalparameters. Due to the limitations of our parameter estimates,varying them over 2 orders of magnitude also allows us to testthe robustness of the model outcomes as well as the sensitivity(see Fig. 3D). Figure 3D (I–IV ) shows results.MDE production has a dominant effect on the outcome of the

simulations, and results show that too little MDE production leavesthe tumor confined to the gland, whereas too much means that the

basement membrane breaks down (and consequently, TGF-h spillsout of the gland) too early, depriving the tumor cells of the requiredminimum of TGF-h level to survive (Fig. 3). This hints that there isan optimal range of MDE production that would allow the tumorto escape from the gland and invade the surrounding tissue. Thisparticular range is parameter dependent, modulated by the rate atwhich the basal membrane is degraded by tumor cells and repairedby basal and stromal cells as well as the susceptibility of the tumorcells to TGF-h concentration.Contrary to our expectations, tumor TGF-h production does not

modify significantly the simulation outcomes. Stromal configura-tion, on the other hand, has a more significant effect on tumorprogression. Prostates with a high proportion of stromal cells areless likely to be taken over by a tumor than those with a lowerdensity of stromal cells. The role of stroma in tumor promotionbecomes critical after the tumor has escaped from the gland, and itis also at this point when TGF-h, which mediates the interactionsbetween stroma and tumor, plays a more dominant role. In fact,tumor cell TGF-h production from within contained glands seemsto inhibit tumor progression due to the recruitment of matrix-producing motile stroma. This often leads to an increaseddeposition of stromal-produced ECM (see Fig. 3A and compareSupplementary Figs. S2 and S3 with different amounts of stromalcells). To ground this in pathologic reality, human prostaticintraepithelial neoplasia (PIN) biopsies were immunostained forTGF-h and counterstained with Masson’s trichrome to determinewhether an adjacent stromal matrix deposition could be observed.Figure 4 shows very little to no TGF-h immunoreactivity in normalhuman prostate glandular epithelium (Fig. 4A) compared with highimmunoreactivity in some glandular PIN foci (Fig. 4B). Further-more, trichrome staining of these sections revealed that the stromaadjacent TGF-h–expressing PIN foci had increased collagenproduction, confirming the HCA model outcome.Given the potential implications, it was important to determine

whether the prediction that tumor cell TGF-h production does notsignificantly alter the simulation outcome was biologically realistic.To test this, constitutively active TGF-h1 was overexpressed in ahuman prostate epithelial cell line previously shown to be capableof malignant transformation (BPH1; refs. 43, 44). Importantly, thiscell line is able to form acini when recombined with inductive raturogenital mesenchyme (rUGM; ref. 44), allowing us to determinewhether the expression of TGF-h might result in enhancedtumorigenicity.A hemagglutinin (HA)-tagged, constitutively active TGF-h1

previously generated and characterized (45) was expressed inBPH1 cells (see Materials and Methods; Fig. 5A). To confirm itsfunctionality, conditioned media collected from parental BPH1 andBPH1TGFh1 cells were placed on serum-starved human prostatefibroblasts expressing a red fluorescent protein (RFP) TGF-hreporter. Results show an activation of TGF-h signaling byBPH1TGFh1 cell conditioned media, but not the parental control(Fig. 5B). To test our simulation prediction, we recombined each

Figure 2. Flowcharts with the life cycle of the different cell phenotypes. A, tumor cells produce TGF-h (which they need for proliferation) and MDEs. They proliferatewhen there is enough TGF-h and have reached maturity. An insufficient local concentration of TGF-h triggers cell death. B, basal cells replace missing luminal cellsand produce TGF-h to suppress proliferation. C, normal luminal cells consume TGF-h and die when they are surrounded by tumor cells. D, nonmotile stroma cellsconsume TGF-h but will produce more when the local concentration exceeds threshold. E, motile stromal cells are constantly moving chemotactically toward thedirection of the TGF-h gradient. F, key interactions between the cell types and biologically active substances. Solid lines, reduction in concentration by consumption,killing, or inhibition of production at the end with the small red cross bar; dashed lines, increases in concentration by promotion of production or proliferation at theend(s) with the arrow and + symbol. The double arrow between tumor cells and TGF-h represents TGF-h production by the tumor cells that in turn results in tumor cellproliferation by TGF-h.





06/06/2013 139


Figure 3. A , example of a simulation in which the tumors break out from the glands and start growing in the mesenchyme. The domain contains 40% ofmotile and 40% of nonmotile stroma. Figures show three snapshots of the simulation. After f3 mo, each of the glands is entirely occupied by a tumor. After 43 y,two of the tumors have managed to break out from the gland. At the end of the simulation (after f54 y), the three tumors have merged into a single massthat seems posed to take over the entire simulation space, although the pattern of growth seems to be channeled by the stroma. B , example of a simulationin which the tumor takes over the entire prostate. In this simulation, the production of TGF-h is relatively low and the proportion of motile stroma as that ofnonmotile stroma is 10%. After f43 y, the tumor has taken over the entire prostate, degraded entirely the membrane and TGF-h and MDEs can be foundeverywhere in high quantities. C , example of simulation in which the three tumors are initiated, grow, and die out before they manage to break out from theglands. In this simulation the proportion of static stroma is the same as nonmotile: 10%. However, tumor cells now produce MDEs at a significantly higherrate than the other simulations. After 60 d, the central gland has almost been taken over by the tumor. After 1 mo, the central tumor has produced enough MDEsto degrade the basal membrane, although the local concentration of TGF-h is remarkably lower than in comparable simulations. Both tumor cells and TGF-h canbe seen spilling out of the gland into the surrounding stroma. After f3 y, the simulation shows a situation similar to that at the beginning. D , graphicalrepresentation of the results of several simulations using different stromal configurations. Each color represents a different outcome. Green , simulations in whichall the tumor cells die; yellow , those in which the tumor does not break out from the gland; red , those in which the tumor does. I , simulations in which thespace has 40% of static and 40% of motile stroma. II , simulations with 40% of motile and 10% of static stroma; III , simulations with 10% of motile and 40% ofstatic stroma; IV , simulations with 20% of motile and 20% of static stroma.

Cancer Research




06/06/2013 140


epithelial cell line with rUGM for 6 weeks under the kidneycapsule of severe combined immunodeficient mice, a technique

previously applied to the study of prostate development and

tumor formation (44). Kidney capsule grafting of tissuerecombination with inductive rUGM results shows that TGF-h1production decreased the overall size of the tissue recombina-

tion graft by 3- to 4-fold (Fig. 5C). Histologically, trichrome(Fig. 5D) and smooth muscle (Fig. 5E) staining of these grafts

revealed that BPH1TGFh1 grafts show a decrease in smooth

muscle and an increase in periacinar collagen production over

controls, consistent with changes seen around premalignanthuman lesions (Fig. 4). Moreover, histochemical identification of

basement membrane by periodic acid-Schiff (PAS) staining

shows that TGF-h1 production by initiated prostate epitheliadoes not affect glandular containment (Fig. 6).

Discussion

This article introduces a mathematical model that recapitulatessome key elements of tumor progression, such as epithelialinvasion and stromal reorganization. The aim of the model is toprovide a basic platform from which to study the role of TGF-h andstroma in prostate tumor progression from nontumorigenichomeostasis to preneoplasia to malignancy. To be useful,mathematical models must generate hypotheses that can be testedin biological systems and compared with clinical observations and,further, must be amenable to testing and generating a deeperunderstanding of biological phenomena.PIN was first defined 20 years ago and is considered to be the

most common premalignant lesion in prostate cancer (46). PIN canbe made to regress in humans (47) and this process can also bedocumented in experimental animal models (48); however, it is

Figure 4. Human PIN foci expressing TGF-h showaltered adjacent stroma. Compared with normalprostates (A ), PIN foci that express TGF-h (B ) showdecreased periacinar smooth muscle with increasedcollagen deposition as shown by TGF-h (circle ) andtrichrome (arrow ) staining.

Table 1. Model parameters

Parameter Value Normalized value

DTh 2 � 10�9 cm2/s* 0.02

m0 1.050 g/cm3 0.1

aT,B 230 � 10�12 g/d/cells* 0.01aT,C 1.8 � 10�9 cm3/d/cell* 1 (0.1, 10)

c

aE,C 1 � 10�9 0.0001 (0.01, 1 � 10�6)c

cT,S 2.7 � cm3/d/cells* 0.1Th,S,0 2.6 � 10�8 cm3/d/cells* 1

cT,M 9.5 � 10�2 cm3/g/d* 1 � 10�6

c¶T,S 1.35 � 10�2 cm3/g/d* 0.01

s 24 h 1DE 10�9 cm2/s 0.00054

DT 2 � 10�9 cm2/s 0.000108

DN 10�9 cm2/s 0.00054

Stroma (10%-10%),c(40%-10%),

c(10%-40%),

c(40%-40%)

c

* Estimated values.cA series of parameters for which the model was tested.





06/06/2013 141


unclear if such regression happens spontaneously in patients. Thus,although PIN can progress to cancer, this is not an inevitableprogression and the conditions that promote or suppress such aprogression are presently unknown.Under basal conditions, the model presented here is homeo-

static. The disruption of homeostasis as a result of the appearanceof tumor cells can result in one of four different scenariosdepending on the phenotype of the tumor cells as well as thestromal microenvironment. These scenarios (Fig. 3) are as follows:(a) nascent tumors that develop slowly and remain at a stage thatrecapitulates aspects of the premalignant lesion PIN, (b) tumorsthat break out of their acinar containment and invade locally butdo not outgrow the other cell types, (c) tumors that develop rapidlyand take over the prostate (Fig. 4B), and (d) tumors that developquickly but are self-limiting, dying out before progression beyondPIN (Fig. 4C). These results show both the potential of the model tocapture distinct outcomes and the reproduction of the diversity oftimescales at which different prostate tumors are known todevelop.This mathematical model generated several hypotheses. In our

simulated prostate environment, before tumor breakout, TGF-hproduction by tumor cells seems to have little effect on theoutcome of the progression (see Fig. 2), which was corroboratedexperimentally (see Fig. 5). TGF-h in this model does recruit motilestromal cells, which structurally inhibit glandular breakdownthrough matrix production, which was also corroborated byexamination of clinical samples of human PIN foci expressingTGF-h (see Fig. 4). This ineffectiveness persists regardless of eitherthe proportion of stromal subtypes or the rates of production ofMDEs by the tumor cells. Similar to biological assumptions (23, 49),the effects of TGF-h in this model are more likely to be tumorigeniconce tumor cells emerge from a contained PIN-like state as

determined by the effects of the TGF-h gradient on tumor cellgrowth and stromal recruitment (Fig. 3A). This is consistent withpreviously described biological observations (26).A second hypothesis generated by the model concerns the role of

tumor cell MDE production. The results show that there is a criticalrange of values of MDE production that are optimal for tumorprogression. If tumor cells produce MDEs at a suboptimal rate, thetime to break out and even the chances of the tumor progressingfrom PIN are affected negatively (Fig. 3). Alternatively, excessiveproduction of MDEs might work against the tumor and simulationresults suggest that too much MDE production could, undercertain conditions, lead to tumor cell death. Degrading thebasement membrane too quickly may result in TGF-h diffusingout of the glandular acinus, leaving the tumor cells in the glandwith insufficient growth factors to survive (Supplementary Fig. S4).Therefore, it is a prediction of this model that the appropriateresources (e.g., TGF-h) essential for invasion and survival of tumorcells outside of the gland are available once the basementmembrane is breached. The specific values that characterize thiscritical range of MDE production will depend on the rest of theparameters of the model, such as TGF-h production and diffusion,or the rate at which MDE is used to degrade the basal membrane.There are several implications coming from these hypotheses.

The first suggests that, despite the central role that TGF-h playedin the conception of the model (Fig. 2F), the level of production bytumor cells does not affect the initial premalignant stages oftumorigenesis. This hypothesis does not address the role of TGF-honce the tumor progresses beyond containment, where the role ofTGF-h seems to be more protumorigenic. A different hypothesis,which could have far-reaching implications, is the role of excessiveMDE production by tumor cells in halting tumor progression,which, at present, is untested.

Figure 5. Functional TGF-h overexpression by contained prostate epithelialglands. A, HA-tagged, constitutively active TGF-h1 was constitutively expressedin BPH1 human prostate epithelia. B, conditioned medium from BPH1TGFh1

(right ), but not BPH1 (left ), cells in culture activated a RFP TGF-h reporterin serum-starved human prostate fibroblasts. C, kidney capsule tissuerecombination grafting shows a decrease in graft size when rUGM is recombinedwith BPH1TGFh1 (right ) versus parental control (left ). Trichrome (D ) andsmooth muscle a actin (E) staining reveals predominantly smooth musclesurrounding BPH1 acini (left) versus increased periacinar collagen depositionin BPH1-TGF-h grafts (right ).

Cancer Research




06/06/2013 142


Finally, an interesting insight provided by the model is the dualrole of structural constraints in tumor progression. The glandularstructure of the prostate both delays and protects tumor growth.Although this arrangement limits the growth in the early stages, itprovides a haven in which tumor cells can grow until they reach acritical mass. The role of the stromal configuration has also beenshown to be important before and after the tumor breaks out fromthe gland. In those cases in which the proportion of stroma is high,degradation of the basement membrane is hindered by thepresence of motile stroma. Furthermore, when the membrane isbreached, the stroma helps to constrain tumor expansion. Thus,the model hints that the abundance of stromal cells, both motile(that block tumor growth and produce ECM) and static (thatamplify the TGF-h signal and provide a clearer gradient for motilestroma), hinders tumor growth.A key value of designing biologically relevant mathematical

models is the clarification of deficiencies in biological modeling. Atpresent, there is no valid biological model of the effect of stromalheterogeneity on human prostate glandular stability or tumorprogression. As such, adequate experimental validation and modelparameterization of the effects of variable stromal compositionsunder different MDE conditions remains unexamined. However,attempts at developing such a model are ongoing, which willprovide a biological tool for examining both mathematical andpathologic predictions about the effects of stromal heterogeneityon tumor progression.

To obtain a tractable model, simplifications were unavoidable.For example, all tumor cells share the same phenotype. Nonethe-less, tumors are microcosms of evolution in which several tumorphenotypes coexist as a result of microenvironmental heterogene-ity and a complex evolutionary process. Furthermore, it is knownthat some types of stromal cells, such as fibroblasts, can alter theirphenotype in response to microenvironmental changes producedby a tumor. It is likely that these evolutionary dynamics wouldaffect tumor progression. Therefore, it is our goal to explore theeffect of the stromal-tumor coevolution in prostate tumorprogression.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.

Acknowledgments

Received 10/23/08; revised 5/19/09; accepted 6/17/09; published OnlineFirst 8/25/09.Grant support: NIH/National Cancer Institute grants of the Microenvironmental

Influences in Cancer (2T32 CA009592-21A1), TMEN (1U54 CA126505 and 1U54 CA126568), and ICBP (5U54 CA113007) programs.The costs of publication of this article were defrayed in part by the payment of page

charges. This article must therefore be hereby marked advertisement in accordancewith 18 U.S.C. Section 1734 solely to indicate this fact.We thank the VICBC workshop, in particular Lourdes Estrada, for facilitating and

motivating this integrative cancer research project.

Figure 6. BPH1-TGF-h glands are contained by basement membranes. PAS staining in normal human prostate (A) and tissue recombination grafts (B) of BPH1 (left )and BPH1-TGF-h (right ) shows glandular containment by basement membranes.

References

1. Kansal AR, Torquato V, Harsh GR, Chiocca EA,Deisboeck TS. Cellular automaton of idealized braintumor growth dynamics. Biosystems 2000;55:119–27.2. Zacharaki EI, Stamatakos GS, Nikita KS, Uzunoglu NK.Simulating growth dynamics and radiation response ofavascular tumour spheroids—model validation in thecase of an EMT6/Ro multicellular spheroid. ComputMethods Programs Biomed 2004;76:193–206.3. Anderson ARA, Chaplain MA. Continuous and discretemathematical models of tumor-induced angiogenesis.Bull Math Biol 1998;60:857–99.4. Chaplain MA, McDougall SR, Anderson AR. Mathe-matical modeling of tumor-induced angiogenesis. AnnuRev Biomed Eng 2006;8:233–57.5. Cappuccio A, Elishmereni M, Agur Z. Cancer immu-notherapy by interleukin-21: potential treatment strat-

egies evaluated in a mathematical model. Cancer Res2006;66:7293–300.6. Roose T, Chapman S, Maini P. Mathematical models ofavascular tumour growth. SIAM Rev 2007;49:179–208.7. Gerlee P, Anderson ARA. An evolutionary hybridcellular automaton model of solid tumour growth. JTB2007;246:583–603.8. Anderson ARA, Weaver A, Cummings P, Quaranta V.Tumor morphology and phenotypic evolution driven byselective pressure from the microenvironment. Cell2006;127:905–15.9. Gatenby R, Smallbone K, Maini P, et al. Cellularadaptations to hypoxia and acidosis during somaticevolution of breast cancer. Br J Cancer 2007;97:646–53.10. Song H, Lacks D, Enmon R, Jain S. Monte Carlosimulation of LNCaP human prostate cancer cellaggregation in liquid overlay culture. BiotechnologyProgress 2003;19:1742–9.

11. Chaplain M, Anderson ARA. Mathematical modelling.Simulation and prediction of tumour induced angiogen-esis. Inv Met 1997;16:222–34.12. Patel A, Gawlinski E, Lemieux S, Gatenby R. A cellularautomaton model of early tumor growth and invasion. JTheor Biol 2001;213:315–31.13. Anderson ARA, Pitcairn A. Application of the hybriddiscrete-continuum technique. In: Alt W, editor. Polymerand cell Dynamics. Boston: Birkhauser; 2003.14. Dormann S, Deutsch A. Modeling of self-organizedavascular tumor growth with a hybrid cellular autom-aton. In Silico Biol 2002;2:393–406.15. Harnden P, Naylor B, Shelley MD, Clements H,Coles B, Mason MD. The clinical management ofpatients with a small volume of prostatic cancer onbiopsy: what are the risks of progression? Asystematic review and meta-analysis. Cancer 2008;112:971–81.





06/06/2013 143


16. Hayward SW, Rosen MA, Cunha GR. Stromal-epithelial interactions in the normal and neoplasticprostate. Br J Urol 1997;79 Suppl 2:18–26.17. Cunha GR, Hayward SW, Wang YZ. Role of stroma incarcinogenesis of the prostate. Differentiation 2002;70:473–85.18. Bissell MJ, Hall HG, Parry G. How does theextracellular matrix direct gene expression? J TheorBiol 1982;99:31–68.19. Boudreau N, Werb Z, Bissell MJ. Suppression ofapoptosis by basement membrane requires three-dimensional tissue organization and withdrawal fromthe cell cycle. Proc Natl Acad Sci U S A 1996;93:3509–13.20. Liotta L, Rao C, Wewer U. Biochemical interactions oftumor cells with the basement membrane. Annu RevBiochem 1988;55:1037–57.21. Paszek MJ, Zahir N, Johnson KR, et al. Tensionalhomeostasis and the malignant phenotype. Cancer Cell2005;8:241–54.22. Tuxhorn JA, Ayala GE, Smith MJ, Smith VC, DangTD, Rowley DR. Reactive stroma in human prostatecancer: induction of myofibroblast phenotype andextracellular matrix remodeling. Clin Cancer Res2002;8:2912–23.23. Massague J. TGFh in cancer. Cell 2008;134:215–30.24. Massague J, Blain SW, Lo RS. TGFh signaling ingrowth control, cancer, and heritable disorders. Cell2000;103:295–309.25. Derynck R, Akhurst RJ, Balmain A. TGF-h signalingin tumor suppression and cancer progression. Nat Genet2001;29:117–29.26. Ao M, Franco OE, Park D, Raman D, Williams K,Hayward SW. Cross-talk between paracrine-actingcytokine and chemokine pathways promotes malignan-cy in benign human prostatic epithelium. Cancer Res2007;67:4244–53.27. Dumont N, Arteaga CL. A kinase-inactive type IITGFh receptor impairs BMP signaling in human breast

cancer cells. Biochem Biophys Res Commun 2003;301:108–12.28. Budu-Grajdeanu P, Schugart RC, Friedman A,Valentine C, Agarwal AK, Rovin BH. A mathematicalmodel of venous neointimal hyperplasia formation.Theor Biol Med Model 2008;5:2.29. Michelson S, Leith J. Autocrine and paracrine growthfactors in tumor growth: a mathematical model. BullMath Biol 1991;53:639–56.30. Vollmer RT, Humphrey PA. Tumor volume inprostate cancer and serum prostate-specific antigen.Analysis from a kinetic viewpoint. Am J Clin Pathol2003;119:80–9.31. Jackson TL. A mathematical investigation of themultiple pathways to recurrent prostate cancer: compar-ison with experimental data. Neoplasia 2004;6:697–704.32. Jackson TL. A mathematical model of prostate tumorgrowth and androgen-independent relapse. DiscreteContin Dynam Sys 2004;4:187–202.33. Swanson KR, Harpold HL, True LD. Prostate-specificantigen: a clinical and mathematical conundrum. Am JClin Pathol 2006;125:331–3.34. Swanson KR, True LD, Murray JD. On the use ofquantitative modeling to help understand prostate-specific antigen dynamics and other medical problems.Am J Clin Pathol 2003;119:14–7.35. Anderson ARA. A hybrid mathematical model ofsolid tumour invasion: the importance of cell adhesion.Math Med Biol 2005;22:163–86.36. Deutsch A, Dormann S. Cellular automaton model-ling of biological pattern formation: characterization,applications and analysis. Boston: Birkhauser; 2005.37. Phillips JL, Hayward SW, Wang Y, et al. Theconsequences of chromosomal aneuploidy on geneexpression profiles in a cell line model for prostatecarcinogenesis. Cancer Res 2001;61:8143–9.38. Wynn TA. Cellular and molecular mechanisms offibrosis. J Pathol 2008;214:199–210.39. Thorne RG, Lakkaraju A, Rodriguez-Boulan E,

Nicholson C. In vivo diffusion of lactoferrin in brainextracellular space is regulated by interactions withheparan sulfate. Proc Natl Acad Sci U S A 2008;105:8416–21.40. Wier ML, Edidin M. Effects of cell density andextracellular matrix on the lateral diffusion of majorhistocompatibility antigens in cultured fibroblasts. J CellBiol 1986;103:215–22.41. Gall WE, Edelman GM. Lateral diffusion of surfacemolecules in animal cells and tissues. Science 1981;213:903–5.42. Ayala G, Tuxhorn JA, Wheeler TM, et al. Reactivestroma as a predictor of biochemical-free recurrence inprostate cancer. Clin Cancer Res 2003;9:4792–801.43. Hayward SW, Wang Y, Cao M, et al. Malignanttransformation in a nontumorigenic human prostaticepithelial cell line. Cancer Res 2001;61:8135–42.44. Hayward SW, Haughney PC, Rosen MA, et al.Interactions between adult human prostatic epitheliumand rat urogenital sinus mesenchyme in a tissuerecombination model. Differentiation 1998;63:131–40.45. Wolfraim LA, Alkemade GM, Alex B, Sharpe S, ParksWT, Letterio JJ. Development and application of fullyfunctional epitope-tagged forms of transforming growthfactor-h. J Immunol Met 2002;266:7–18.46. DeMarzo AM, Nelson WG, Isaacs WB, Epstein JI.Pathological and molecular aspects of prostate cancer.Lancet 2003;361:955–64.47. Kang TY, Nichols P, Skinner E, et al. Functionalheterogeneity of prostatic intraepithelial neoplasia: theduration of hormonal therapy influences the response.BJU Int 2007;99:1024–7.48. Narayanan BA, Narayanan NK, Pittman B, Reddy BS.Regression of mouse prostatic intraepithelial neoplasiaby nonsteroidal anti-inflammatory drugs in the trans-genic adenocarcinoma mouse prostate model. ClinCancer Res 2004;10:7727–37.49. Bierie B, Moses HL. TGF-h and cancer. CytokineGrowth Factor Rev 2006;17:29–40.

Cancer Research




06/06/2013 144


Philipp Heumann

Metabotropic Regulation of RhoA/Rho-Associated Kinase by L-type Ca2_ Channels Quelle: Circulation Research is available at http://circres.ahajournals.org DOI: 10.1161/CIRCRESAHA.111.240127 Betreuer: Prof. Olaf Grisk (Physiologie) Was bedeutet mir dieses Thema persönlich? Die Aktivität des RhoA/ROCK-Signalkaskade ist entscheidend für Regulation und Aufrechterhaltung der renalen Gefäßfunktion. Durch die Rho-Kinase wird das Ausmaß der Calciumsensitivierung der glatten Gefäßmuskulatur reguliert und dadurch die Kontraktionsbereitschaft der Gefäßmuskelzellen beeinflusst. Hinzu kommt, dass die Rho-Kinase eine Reihe von Zytoskelettproteinen als Substrat für die Phosphorylierung nutzt und somit einen Einfluss auf die Proliferation, Migration und Differenzierung der glatten Muskelzellen hat. Des Weiteren ist sie an Remodellierungsvorgängen der Gefäßwände beteiligt. Eine chronische Aktivierung des Signalweges ist z.B. mit der Entstehung der arteriellen Hypertonie, chronischen Herzinsuffizienz, Leberzirrhose und dem nephrotischen Syndrom assoziiert. Die Prävalenz der genannten Krankheiten unterstreicht die Notwendigkeit, tiefere Einsichten in die Aktivierung, Regulation und Inhibitionsmöglichkeiten der Signalkaskade zu erhalten. Die Erkenntnisse aus diesen Studien könnten dann zu einer Etablierung neuartiger Therapieoptionen führen. Die RhoA/ROCK Signalkaskade beschäftigt mich während meiner Promotionsarbeit. Ein grundlegendes Verständnis zu diesem Signalweg sollte für jeden angehenden Mediziner interessant sein. Worauf kommt es bei diesem Thema am meisten an? Die Kontraktionskraft der glatten Gefäßmuskulatur kann sowohl durch eine gesteigerte intrazelluläre Calciumkonzentration (via Gq/11) als auch durch eine gesteigerte Calciumsensitivierung (via G12/13) beeinflusst werden. Für letztere Signalkaskade spielt die Rho-Kinase (ROCK) eine entscheidende Rolle. Mir ist es wichtig zu zeigen, dass eine konsequente Trennung dieser beiden Signalwege wahrscheinlich nicht sinnvoll ist. Es besteht viel mehr eine Interaktion (Crosstalk) zwischen beiden Signalkaskaden. Die Kernaussage des Papers ist, dass eine Depolarisations-induzierte Kontraktion nicht nur durch einen erhöhten Calciumeinstrom via L-Typ-Ca2+-Kanäle generiert wird, sondern dass eben dieser Kanal bzw. das durch ihn einströmende Calcium eine direkte / indirekte Wirkung auf die Aktivität des kleinen G-Protein RhoA hat. Was fasziniert mich selbst am Thema am meisten? Die Komplexität und Vielseitigkeit der Wirkungen des Neurotransmitters Noradrenalin – vermittelt durch das sympathische Nervensystem – interessieren mich besonders. Da Noradrenalin als Agonist am α1-Adreno-rezeptor seine Wirkungen unter anderen auch über die nachgeschaltete RhoA/ROCK-Signalkaskade verm-ittelt, soll dieses Paper dazu dienen weitere Informationen zu den pleiotropen Effekten von Noradrenalin zu erhalten und zu verstehen. Hier besteht die Notwendigkeit, weitere durch Noradrenalin-unabhängige Akti-vierungsmöglichkeiten des RhoA-Proteins zu identifizieren. Das Paper beschreibt eine weitere Möglichkeit.

06/06/2013 145

Was gefällt mir am Thema weniger? Ich habe mir das Paper selbstständig ausgewählt und meine gesamte Promotion wird sich mit dieser The-matik befassen. Die Komplexität und Vielseitigkeit dieses Themas macht das Projekt so interessant und ist für mich eine besondere Herausforderung. Ich habe – seit ich an diesem Forschungsprojekt arbeite – sehr viel Neues und Interessante erfahren. Die Frage nach dem, was mir nicht gefällt, kann ich nicht beantwor-ten und das sollte in meinen Augen auch so sein.

06/06/2013 146

Metabotropic Regulation of RhoA/Rho-Associated Kinase byL-type Ca2� Channels

New Mechanism for Depolarization-Evoked MammalianArterial Contraction

Miguel Fernandez-Tenorio, Cristina Porras-Gonzalez, Antonio Castellano, Alberto del Valle-Rodríguez,Jose Lopez-Barneo, Juan Urena

Background: Sustained vascular smooth muscle contraction is mediated by extracellular Ca2� influx through L-typevoltage-gated Ca2� channels (VGCC) and RhoA/Rho-associated kinase (ROCK)-dependent Ca2� sensitization of thecontractile machinery. VGCC activation can also trigger an ion-independent metabotropic pathway that involvesG-protein/phospholipase C activation, inositol 1,4,5-trisphosphate synthesis, and Ca2� release from the sarcoplasmicreticulum (calcium channel-induced Ca2� release). We have studied the functional role of calcium channel-inducedCa2� release and the inter-relations between Ca2� channel and RhoA/ROCK activation.

Methods and Results: We have used normal and genetically modified animals to study single myocyte electrophys-iology and fluorimetry as well as cytosolic Ca2� and diameter in intact arteries. These analyses were complementedwith measurement of tension and RhoA activity in normal and reversibly permeabilized arterial rings. We have foundthat, unexpectedly, L-type Ca2� channel activation and subsequent metabotropic Ca2� release from sarcoplasmicreticulum participate in depolarization-evoked RhoA/ROCK activity and sustained arterial contraction. We show thatthese phenomena do not depend on the change in the membrane potential itself, or the mere release of Ca2� from thesarcoplasmic reticulum, but they require the simultaneous activation of VGCC and the downstream metabotropicpathway with concomitant Ca2� release. During protracted depolarizations, refilling of the stores by a residualextracellular Ca2� influx through VGCC helps maintaining RhoA activity and sustained arterial contraction.

Conclusions: These findings reveal that calcium channel-induced Ca2� release has a major role in tonic vascularsmooth muscle contractility because it links membrane depolarization and Ca2� channel activation with metabotropicCa2� release and sensitization (RhoA/ROCK stimulation). (Circ Res. 2011;108:1348-1357.)

Key Words: calcium channels � sarcoplasmic reticulum � vascular smooth muscle � vasoconstriction

Cardiovascular diseases secondary to alterations in thearterial diameter constitute an important cause of mor-

bidity and mortality in humans; however, the mechanismsthat determine vascular smooth muscle (VSM) contractilityare not fully understood. Myocyte contraction is triggered bythe increase of cytosolic Ca2� concentration ([Ca2�]i) becauseof Ca2� influx through ion channels in the plasmalemma, Ca2�

release from the sarcoplasmic reticulum (SR), or both. Besidesexerting direct control of myosin light chain phosphorylation,cytosolic Ca2� also influences VSM contraction through en-hancement of the sensitivity of the contractile apparatus, aprocess called Ca2� sensitization that is mediated by the smallmonomeric G-protein RhoA and its target Rho-associated kinase(ROCK). The RhoA/ROCK pathway, presumed to be involved

in the pathogenesis of major vascular disorders,1,2 is turned onby excitatory agonists through G-protein-coupled receptor stim-ulation,3–5 and it also seems to be activated by membranedepolarization.6–10 Nonetheless, the relationships between mem-brane voltage changes, cellular Ca2� regulation, and RhoAactivation remain obscure. We have hypothesized that activationof voltage-dependent Ca2� channels and RhoA could sharesome common mechanisms.

Ca2� influx through voltage-gated L-type Ca2� channels(VGCC) is a major event in VSM excitation– contractioncoupling. Depolarization is also known to induce Ca2�

release from the SR independently of extracellular Ca2�

influx (depolarization-induced Ca2� release) by means of ametabotropic pathway that involves G-protein/phospho-

Original received January 3, 2011; revision received April 5, 2011; accepted April 7, 2011. In March 2011, the average time from submission to firstdecision for all original research papers submitted to Circulation Research was 13.2 days.

From the Instituto de Biomedicina de Sevilla and Departamento de Fisiología Medica y Biofísica (M.F.T., C.P.G., A.C., J.L.B., J.U.), HospitalUniversitario Virgen del Rocío/CSIC/Universidad de Sevilla, Seville, Spain; Department of Biology (A.d.V.R.), Center for Developmental Genetics, NewYork University, NY.

Correspondence to Dr Juan Urena and Dr Jose Lopez-Barneo, Instituto de Biomedicina de Sevilla, Edificio IBiS, Campus Hospital Universitario Virgendel Rocío, Avenida Manuel Siurot s/n, E-41013, Sevilla, Spain. E-mail [email protected] [email protected]

© 2011 American Heart Association, Inc.

Circulation Research is available at http://circres.ahajournals.org DOI: 10.1161/CIRCRESAHA.111.240127

1348

06/06/2013 147

lipase C (PLC) activation and inositol 1,4,5-trisphosphate(InsP3) synthesis.11–14Although other membrane proteins in thereceptor/G-protein/PLC cascade have been proposed to be volt-age-sensitive,15–17 it has been shown that L-type Ca2� channelsare necessary for depolarization-induced Ca2� release in vascu-lar14,18–20 and skeletal21 muscle. Therefore, plasmalemmal Ca2�

channels in VSM have been suggested to act as voltage sensorsthat mediate ion-independent coupling between membrane de-polarization and SR Ca2� release (Ca2� channel-induced Ca2�

release [CCICR]).14,20 The precise physiological role of CCICRand its inter-relations with the RhoA/ROCK pathway are as yetunexplored. Herein, we show that, unexpectedly, CCICR isnecessary for sustained VSM contraction and that maintenanceof arterial tone requires Ca2� channel-dependent RhoA activa-tion. Our data suggest that sarcolemmal VGCC not only mediatethe Ca2� influx that triggers contraction but also metabotropi-cally regulate SR Ca2� release and RhoA activation to sensitizethe contractile machinery. These novel findings suggest thatCa2� channels have a role in the regulation of VSM tonus muchbroader than previously thought and provide a comprehensiveview of myocyte contractility. This work helps identify mecha-nisms responsible for hypertension as well as arterial (eg,coronary or cerebral) vasospasms.

Subjects and MethodsAnimal ModelsRats and mice were anesthetized with ketamine (80 mg/kg intramuscu-lar). Rabbits were anesthetized with sodium pentobarbital (75 mg/kgintravenous). Cav1.2 knockout animals were generated at the Institut furPharmakologie und Toxikologie, Munchen, Germany.22,20 All of theexperimental procedures were approved by the Institutional AnimalCare and Use Committee and were conducted in conformity withinstitutional guidelines.

Cytosolic Ca2� Measurements and ElectrophysiologicalRecordings in Isolated MyocytesCytosolic [Ca2�] and patch clamp recordings from dispersed myo-cytes were performed as previously described.14

Confocal Microscopy and ImmunocytochemistryImmunocytochemical studies were performed on fixed cells follow-ing standard protocols. For details, see Methods in the online DataSupplements.

Measurement of Contractility in Arterial Rings and RhoAActivation AssayContractility of arterial rings was studied as previously described.18

RhoA activity was determined from quickly frozen arterial rings

(immersed in liquid nitrogen) using the colorimetric G-LISA RhoAactivation assay biochemistry kit (Cytoskeleton, Denver, CO). Fordetails, see Methods in the online Data Supplements.

Simultaneous Measurements of Fluorescenceand ContractilityBasilar arteries (12–14 mm) were incubated in Hank’s containingFura 2 AM 1.4 �mol/L for 1 hour at room temperature. Experimentswere performed as previously described.19

Reversible PermeabilizationReversible permeabilization was achieved by a modification ofprevious methods.23,24 For details and solutions, see Methods in theOnline Data Supplement at http://circres.ahajournals.org.

Statistical AnalysisData are expressed as mean�standard error and the statistical signifi-cance was estimated using the Student t test. Values of P�0.05 wereconsidered significant.

ResultsSustained Depolarization-Evoked VascularMyocyte Contraction Requires Metabotropic Ca2�

Release From the SRTo investigate the functional role of metabotropic Ca2� releasefrom the SR evoked by VGCC activation (CCICR),14,20,25 basilararterial rings were stimulated by exposure to an external solutioncontaining 70 mmol/L K�, which depolarizes isolated current-clamped arterial myocytes.14 The high K� solution shifted themembrane potential of basilar VSM cells from �50 mV to��20 mV. Depolarization with 70K triggered an isometriccontraction that showed an initial fast and transient (phasic)component, followed by a slowly increasing (tonic) component,which was generally maintained until the end of the stimulus(Figure 1A; Online Figure IA and B). These phasic and toniccomponents were also observed in contractions evoked byapplication of ATP, a vasoactive agent that induces activation ofVGCC26,18 (Figure 1B). Treatment of the arteries with cyclopia-zonic acid (CPA), an SR Ca2� ATPase inhibitor that depletesCa2� stores without effects on VGCC (online Figure IIA and B),selectively reduced the tonic component of contractions evokedby either high K� or ATP (Figure 1A and 1B). Similar results(decrease of the sustained phase of arterial contraction) wereobserved with two other drugs (thapsigargin and ryanodine) thatalso reduce Ca2� storage in SR (Online Figure IIC and IID).These data indicated that the sustained component ofdepolarization-evoked VSM contraction depends on Ca2� re-lease from intracellular stores, possibly as a consequence ofCCICR. CCICR is inhibited by organic VGCC antagonists;14,18

hence, we tested the effect of nifedipine on depolarization-evoked contractions. Nifedipine was more potent to inhibit thetonic than the phasic components of contractions. Nifedipine (10nmol/L) produced a marked reduction of the high K-evoked orATP-evoked tonic contraction leaving intact the phasic period.At higher concentrations of nifedipine (100 nmol/L), the twophases of contraction were abolished (Figure 1C and 1D; OnlineFigure IIE and IIF). These results suggest that VSM cellcontraction elicited by cell depolarization depends on bothionotropic (Ca2� entry) and metabotropic (CCICR) actions ofplasmalemmal Ca2� channels.

To test whether depolarization-evoked sustained VSM con-traction does involve metabotropically induced Ca2� release,

Non-standard Abbreviations and Acronyms

CCICR calcium channel-induced Ca2� release

CPA cyclopiazonic acid

FPL FPL64176

InsP3 inositol 1,4,5-trisphosphate

PLC phospholipase C

ROCK Rho-associated kinase

SR sarcoplasmic reticulum

VGCC L-type voltage-gated Ca2� channels

VSM vascular smooth muscle

Fernandez-Tenorio et al CCICR and RhoA/ROCK in Vascular Smooth Muscle 1349

06/06/2013 148

arterial rings were exposed to U73122, an inhibitor of PLC.14

This treatment left unaffected the phasic component of the70K-evoked contraction but markedly reduced the tonic com-ponent. U73343, a structurally closely related compound thatdoes not inhibit PLC, did not significantly affect VSM contrac-

tility (Figure 2A). Similar results were obtained in basilar arterialrings contracted with ATP and treated with U73122 (Figure 2B).The effect of U73122 was mediated by changes in [Ca2�]i,because in isolated myocytes U73122 decreased the sustainedphase of the 70K-induced Ca2� signal, with no significant effect

Figure 1. Effect of cyclopiazonic acid(CPA) and nifedipine (Nif) on contrac-tion in rat basilar arterial rings. A, Iso-metric contraction induced by70 mmol/L external K� solution (70K)with a phasic component of 7.13�0.23mN, reached at �15 seconds, and atonic component with a tension of7.88�0.3 mN (determined at 6 minutes,n�45). Application for 30 to 40 minutesof 10 �mol/L CPA selectively reducedthe maintained phase of the 70K-evokedcontraction (22%�7.3%; n�6; P�0.05),leaving the phasic component unaffected(1%�3.5% potentiation; n�6; P�0.05).B, Representative traces of isometricforce developed by arterial rings inresponse to ATP (25 �mol/L). The fasttransient component was not signifi-cantly affected in arterial rings treatedwith 10 �mol/L CPA for 20 minutes. Theslow component, measured 30 secondsafter the peak, was markedly reduced(66%�17.6%; n�6; P�0.05). C, Nifedi-pine (10 nmol/L) selectively inhibited thetonic component (57.9%�5.94%; n�11;P�0.01), leaving the phasic componentunaffected (4.26%�2.6% inhibition;n�11; P�0.05). Nifedipine (100 nmol/L)markedly inhibited both phasic andtonic components of contraction

(67.52%�8.45%; n�7; P�0.01; and 91.55%�2.16%; n�7; P�0.01, respectively). D, Nifedipine (10 nmol/L) selectively inhibited theslow component of the ATP-evoked contraction measured 2 minutes after the peak (3.7% for the fast component; n�8; P�0.05; 69%for the slow component; n�8; P�0.01). Nifedipine (100 nmol/L) markedly reduced both components of the ATP-elicited contraction(77% fast component; n�8; P�0.01; 86% slow component; n�8; P�0.01).

Figure 2. Participation of phospholipase C andG-protein in the tonic phase of depolarization-evoked contractions in rat basilar arteries. A,Treatment of arterial rings for 20 to 40 minuteswith U73122 (1 �mol/L) left the phasic componentof the 70K-evoked contraction unaffected(1%�15% inhibition; n�6; P�0.05) and signifi-cantly reduced the amplitude of the tonic compo-nent (38%�12% inhibition; n�6; P�0.01). U73343(1 �mol/L) did not significantly affect the phasic ortonic components (1.2%�1.1% inhibition; n�8;P�0.05; 3.9%�2.4% potentiation; n�8; P�0.05,respectively, measurement at 20 minutes). B,Bathing arterial rings with U73122 (1 �mol/L) for10 minutes reduced the tonic component of theATP-induced contraction measured 30 secondsafter the peak (52% inhibition; n�6; P�0.01),whereas the phasic component was not signifi-cantly affected (17% inhibition; n�6; P�0.05). Cand D, Effect of GDP�S (10 mmol/L) on contrac-tions induced by depolarization (70K) and caffeine(10 mmol/L) in reversibly permeabilized arteries.Values are presented as mean�SEM *P�0.05.

1350 Circulation Research May 27, 2011

06/06/2013 149

on the transient component (Online Figure IIIA through IIID).Further support for the notion that the G-protein/PLC pathwayparticipates in depolarization-induced tonic contractions camefrom experiments on intact arteries in which myocytes had beenreversibly permeabilized in the presence of GDP�S to inhibitG-protein activation.14,23,24 In these arteries, the tonic componentof the 70K-evoked VSM contraction was also selectively re-duced, whereas the phasic component of contraction and themechanical signal induced by caffeine (a drug that directly elicitsCa2� release from the SR) were unaltered (Figure 2C and 2D).Together, these results further supported the notion that stimu-lation of the G-protein/PLC cascade subsequent to membranedepolarization and Ca2� channel activation leads to the SR Ca2�

release necessary for maintained VSM contraction.

Sustained Depolarization-Evoked Contraction RequiresG-Protein-Dependent and InsP3-DependentRhoA/ROCK ActivationIn VSM, a Ca2� sensitization mechanism mediated by theRhoA/ROCK pathway seems to participate in the depolarization-evoked contraction; however, the underlying mechanisms arenot well-known.6–10 We tested in immunostained isolatedmyocytes the redistribution of activated RhoA and its dis-placement toward the plasma membrane in response to highK�-induced depolarization (Figure 3A through 3C).9,27 Be-cause ROCK is a RhoA downstream effector, we studied in

intact pressurized arteries the effect of Y27632, a ROCKinhibitor,1 on the cytosolic Ca2� and mechanical responses tohigh K�. ROCK blockade did not modify the Ca2� signalelicited by depolarization but selectively inhibited the toniccomponent of the contraction (Figure 3D). Qualitatively similarresults (inhibition of the tonic phase of contraction) wereobtained in arterial rings treated with HA1077, another ROCKinhibitor structurally unrelated to Y276321 (Online Figure IIIEand IIIF). These experimental observations further suggestedthat the depolarization-evoked sustained phase of contractionrequires, at least in part, RhoA/ROCK activation.

To show the parallel metabotropic regulation of toniccontraction and RhoA activation, we performed experimentsin reversibly permeabilized arteries (Figure 2C and 2D) thatwere frozen at the end of the depolarization pulse (�10minutes after exposure to high K�) to directly measureRhoA activity. Application of GDP�S and heparin (inhib-itors of G-protein activation and InsP3 receptors, respec-tively) selectively reduced the tonic component of the70K-evoked contractions and also decreased RhoA activityinduced by depolarization (Figure 4; Online Figure IC).These observations supported the view that metabotropicCa2� release from the SR elicited by membrane depolar-ization maintains tonic contractions through activation ofthe RhoA/ROCK pathway.

Figure 3. Cytosolic distribution of acti-vated RhoA in isolated myocytes andeffect of the Rho-associated kinase(ROCK) inhibitor Y27632 on 70K-induced contraction in rat basilarartery. A and B, Confocal immunofluo-rescent images of RhoA distribution(RhoA, green; cell nucleus, red) in a sin-gle basilar smooth muscle cell at rest (A)and after stimulation with 70K (B). Scalebars�7 �m. C, Averaged peripheral:cy-tosolic (Fm/Fc) RhoA ratios for two cen-tral z-sections in control and 70K-induced depolarization (n�12). D, Effectof the ROCK inhibitor Y27632(10 �mol/L) on cytoplasmic Ca2� anddiameter (36.6%�9.27% inhibition oftonic component; n�4; P�0.05) in anintact artery depolarized with 70K. Val-ues are presented as mean�SEM*P�0.05.


06/06/2013 150

Essential Role of L-type Ca2� Channel Activation andCa2�Channel-Induced Ca2� Release in Depolarization-EvokedRhoA Activation and Sustained Myocyte ContractionBecause depolarization-evoked tonic myocyte contraction isparalleled by metabotropic RhoA/ROCK activation, we tested tosee whether blockade of tonic contractility with nifedipine(Figure 1C and 1D) also altered RhoA activity. As expected,70K-induced contractions and RhoA activation in arterialrings were fully prevented by nifedipine (Figure 5A, top andbottom), thus suggesting that Ca2� channel activation isrequired to increase RhoA activity. The critical role of VGCCactivation (independently of membrane depolarization) inRhoA regulation was further supported by two sets ofexperiments. First, the application of FPL64176 (FPL), aselective L-type Ca2� channel agonist with minor effects onthe resting membrane potential,14,18 induced a contraction thatwas fully blocked by nifedipine (Figure 5B and 5C, top

panels; Online Figure ID and IE). RhoA activity, whichincreased by FPL in parallel with the mechanical response,was also almost abolished by nifedipine application (Figure5B and 5C, bottom panels). Second, in a conditional Cav1.2(L-type) Ca2� channel knockout mice model,22,20 contractileresponses to 70K and FPL were greatly reduced in parallelwith inhibition of RhoA activity (Figure 6). In these geneti-cally modified animals, contractions induced by applicationof either phenylephrine or caffeine (because of Ca2� releasefrom InsP3 or ryanodine-sensitive stores, respectively)28,29

were unaffected, thus indicating that VGCC deficiency didnot alter other physiological features of the VSM cells. TheRhoA/ROCK pathway was maintained functional in Cav1.2null myocytes because it was activated by phenylephrine withsimilar potency as in control cells (Figure 6D). Moreover, thesustained contraction evoked by phenylephrine was blockedby pharmacological ROCK inhibition (Online Figure IV).

Figure 4. Participation ofG-proteins and inositol 1,4,5-trisphosphate (InsP3) recep-tors in the sustained compo-nent of the depolarization-induced contraction andRhoA activation in rabbit fem-oral arteries. A, Effect ofGDP�S (10 mmol/L) on contrac-tion induced by depolarization(70K) and caffeine (10 mmol/L)in reversibly permeabilized arter-ies. B, Quantitative summary ofrelative isometric force devel-oped in response to the differ-ent stimuli, 70K or caffeine(n�11). C, Effect of GDP�S onRhoA activation (relative to rest-ing condition in the absence ofany stimulus) during 70K-induced depolarization (n�5). D,Effect of heparin (10 mg/mL) oncontractions induced by 70K

and caffeine (10 mmol/L) in reversibly permeabilized arteries. E, Quantitative summary of isometric force developed in response to the differ-ent stimulus (70K, n�17; caffeine, n�14). F, Effect of heparin on RhoA activation during the 70K-induced depolarization (relative to restingconditions). The vertical arrows indicate the time at which 70K-evoked force was measured and arterial rings were frozen (F) for posteriormeasurement of RhoA activity (�10 minutes after 70K application). Values are presented as mean�SEM. *P�0.05, **P�0.01.

Figure 5. Effect of nifedipine on 70K-induced and FPL64176 (FPL)-inducedcontraction and RhoA activation inrabbit basilar artery. A, Statisticalanalysis of isometric contraction (top)and RhoA activity (bottom) measured10 minutes after stimulation inducedby 70K in the presence or absence ofnifedipine (1 �mol/L, n�3). RhoA activ-ity is normalized to basal condition(control). B, Time-dependent changesin contraction (top) and RhoA activa-tion (bottom) in basilar arterial ringsstimulated with FPL (0.5 �mol/L). Thevertical arrows indicate the time atwhich arterial rings were frozen (F) forposterior measurement of RhoA activ-ity (2 minutes, n�3; 10 minutes, n�10;30 minutes, n�3). C, Statistical analy-sis of isometric contraction (top) andRhoA activity (bottom, measured at 10minutes after stimulation) induced by

FPL (0.5 �mol/L) in the presence (n�32) or absence (n�3) of 1 �mol/L nifedipine. Values are presented as mean�SEM. **P�0.01.


06/06/2013 151

To distinguish between the effects of Ca2� channel acti-vation (and the downstream metabotropic pathway) and Ca2�

entry/release on RhoA activation, we performed experimentsin arteries bathed in Ca-free solutions (0Ca plus 1 mmol/LEGTA; estimated [Ca2�]I �100 nmol/L) to prevent anytransmembrane Ca2� influx. In these conditions, caffeineinduced a transient Ca2� release signal that triggered arterialcontraction (Online Figure VA). These effects of caffeinedecreased progressively as the SR stores were depleted ofCa2� (Figure 7A). RhoA activity measured 10 minutes aftercaffeine application was similar to control values (Figure 7Band 7C), thus indicating that caffeine-induced Ca2� releaseand contraction occur without involvement of the RhoA/ROCK pathway. Pharmacological inhibition of ROCK didnot produce any effect on the mechanical responses tocaffeine (Online Figure VB). Application of a membrane-permeant InsP3 derivative also failed to induce contractionand RhoA activation in arterial rings (Online Figure VI).Therefore, SR Ca2� release per se (bypassing Ca2� channel-induced metabotropic activation) does not seem to be suffi-cient to activate RhoA. In contrast to caffeine, FPL (that alsoinduced a transient Ca2� signal and arterial contraction;online Figure VC) elicited a marked RhoA activation thatdecreased with time as the SR was depleted of Ca2� (Figure7D). On exposure to FPL, RhoA activity increased to amaximal value that started to decrease 10 minutes aftertreatment (Figure 7E and 7F). These findings contrast withthe results shown before in the presence of extracellularCa2�, a condition in which FPL induced sustained contrac-tion and RhoA activation (Figure 5B). In accord with theseobservations, FPL-induced contraction was inhibited by phar-macological blockade of ROCK (Online Figure VD).

It is known that phosphorylation of myosin light chain inKCl-stimulated myocytes is highly sensitive to [Ca2�]i;8

hence, we sought to determine whether changes in extracel-lular [Ca2�] can modify RhoA activation. Experiments wereperformed to compare the data described before in arteriesbathed in 0Ca plus EGTA solutions (estimated [Ca2�] �100

nmol/L) with preparations exposed to 0Ca solutions withoutEGTA added (estimated [Ca2�] �7 �mol/L) for more than60 minutes. In this last solution, both caffeine-evoked con-traction and FPL-induced RhoA activation remained similarto the values observed in control conditions, before applica-tion of any stimulus (Figures 8A, 8B, 7A, and 7D). FPLelicited a sustained contraction paralleled by an increase inRhoA activity that, after reaching a maximal value, remainedalmost constant or declined slowly during treatment (Figures8C, 8D, 7E, and 7F). FPL-induced sustained contraction wasdecreased by pharmacological ROCK inhibition, whereas thecytosolic Ca2� signal was not affected (Online Figure VE).Interestingly, the level of FPL-evoked RhoA activation, after10 minutes of treatment, was not significantly different inarterial rings bathed in either �7 �mol/L (136%�8%; n�8;Figure 8D) or 2.5 mmol/L (143%�6%; n�10; Figure 5B)Ca2� (P�0.18), thus suggesting that VSM cells have a highlyeffective Ca2� uptake system that can maintain refilling theintracellular stores even when extracellular [Ca2�] is reducedto �mol/L levels.

Altogether, these data suggest that depolarization-evokedsustained RhoA activation and myocyte contraction do notdepend on the change in the membrane potential per se or thesimple release of Ca2� from the SR, but they require thesimultaneous activation of VGCC and the downstream stim-ulation of a metabotropic pathway leading to InsP3 synthesisand Ca2� release. Therefore, CCICR has a major role in tonicVSM contractility because it links Ca2� channel activationwith metabotropic Ca2� release and sensitization (RhoA/ROCK stimulation).

DiscussionIn this article, we describe that besides their classical ionotropicrole responsible for Ca2� entry and the Ca2�/calmodulin-dependent activation of myosin light chain kinase, L-type Ca2�

channels have other unexpected functions required for mainte-nance of depolarization-evoked contraction in VSM cells. Weshow that L-type Ca2� channel activation constitutes the key

Figure 6. Genetic ablation of L-typeCa2� channels reduces depolarizationor FPL64176 (FPL)-evoked contractionand RhoA activity in mouse aorta. Rep-resentative traces of contractions evokedby 70K (A) or FPL (B) in Cav1.2 wild-typeand Cav1.2-deficient mice. C, Quantitativesummary of isometric forces generated inarterial rings in response to either 70K(n�16), FPL (5 �mol/L, n�10), phenyleph-rine (1 �mol/L, n�12), and caffeine(10 mmol/L, n�28). Note that caffeine-evoked and phenylephrine-evoked con-tractions were not reduced in Cav1.2�/�

mice. D, Quantitative summary of RhoAactivity (normalized to resting condition)generated in arterial rings in response toeither 70K (n�3), FPL (5 �mol/L, n�3),phenylephrine (1 �mol/L, n�3) and caf-feine (10 mmol/L, n�3). Note that RhoAactivity induced by phenylephrine was notreduced in Cav1.2�/� mice. Values arepresented as mean�SEM. *P�0.05,**P�0.01.


06/06/2013 152

event that triggers depolarization-dependent metabotropic Ca2�

release from SR and smooth muscle sensitization throughactivation of the RhoA/ROCK pathway. It had been describedbefore that in the absence of extracellular Ca2� influx, Ca2�

channels of VSM cells act as voltage sensors that couplemembrane depolarization to InsP3 synthesis and Ca2� releasefrom SR (CCICR).14,25,20 We now show that in normalconditions, CCICR mediates Ca2� release and sensitizationof the contractile apparatus necessary to generate sustainedVSM contractions.

Maintenance of Depolarization-Evoked Vascular MyocyteContraction Requires Functional Ca2� Channels andMetabotropic Ca2� Release From the SRIn agreement with previous studies on other arteries,7,9 highK�-evoked contraction in basilar arterial rings exhibited a

time course characterized by an early transient phase fol-lowed by a tonic sustained component. Transient and sus-tained phases in vascular myocyte contraction were alsoobserved after application of agonists that activate membraneVGCC.18 In this article, we provide strong evidence indicat-ing that, besides transmembrane Ca2� influx, the sustainedphase of contraction also depends on Ca2� release frominternal stores. Therefore, CCICR, a mechanism described inarteries bathed without external Ca2�,14 seems to have amajor functional role in arterial contraction under normalconditions. Exposure to high KCl is often used as a tool tobypass G-protein–coupled receptor stimulation and activatesmooth muscle by changing the K� equilibrium potential andclamping membrane potential above resting level.10 How-ever, it is well-established that in several cell types (includingVSM cells) membrane depolarization can induce direct acti-vation of the G-protein-PLC cascade, leading to InsP3 syn-thesis and Ca2� release. In vascular14,25,20 as well as inskeletal muscle cells,21 a great deal of experimental evidenceexists supporting the view that VGCC are the voltage sensorslinking membrane depolarization to G-protein activation,InsP3 synthesis, and Ca2� release. In bronchial smoothmuscle17 and megakaryocytes lacking VGCC,16 other mem-brane proteins have been proposed to also have intrinsicvoltage sensitivity.16

In the present set of experiments, the role of CCICR on VSMcontraction, studied before in isolated myocytes, has beeninvestigated in arterial rings and reversibly “permeabilized”arteries.23,24 These last preparations better-represent the invivo physiological conditions yet permit pharmacologicalmodifications of the extracellular and intracellular media. Thedata show that sustained depolarization-evoked vascular con-traction depends on G-protein/PLC/InsP3-mediated Ca2� re-lease from the SR, a metabotropic pathway triggered byVGCC activation. We find particularly conclusive the exper-iments performed using the conditional L-type Ca2� channelknockout mouse, in which we have recently shown the strictdependence of depolarization-evoked Ca2� release on themaintenance of functional Ca2� channels.20 In this animalmodel, protracted VSM contraction induced by either mem-brane depolarization (exposure to high K�) or pharmacolog-ical Ca2� channel activation (application of FPL) was re-duced in parallel with the decrease of channel expression.However, contractions induced by either caffeine, whichdirectly activates Ca2� release from stores,29,30 or phenyleph-rine, which activates the G-protein/PLC cascade andROCK,28 were unaltered. Therefore, it seems that VGCChave two different roles in the contractile response to depo-larization. At first instance, membrane depolarization causesmassive Ca2� entry, myosin light chain kinase activation, anda phasic contraction.31 During maintained depolarization,channels undergo a conformational change (switching be-tween inactive and open states) that triggers the metabotropiccascade, resulting in Ca2� release from SR and tonic con-traction. Ca2� channel inactivation and the displacement ofthe membrane potential toward the Ca2� equilibrium poten-tial drastically reduce transmembrane Ca2� influx. However,a small steady-state Ca2� channel activity and Ca2� currenthave been described in sustained depolarized myocytes.32,33

Figure 7. Ca2�-dependence of the FPL64176 (FPL)-evokedcontraction in rabbit basilar artery. A, Statistical analysis ofthe effect of 10 mmol/L caffeine on isometric force at differenttimes during depletion of intracellular calcium stores (n�3). Caf-feine was applied after bathing arterial rings in 0Ca � EGTA(1 mmol/L) solution during the indicated time. Caffeine-evokedcontraction (B) and RhoA activation (C) in arterial rings bathed ina 0Ca � EGTA solution (caffeine 10 mmol/L, n�3). The verticalarrow indicates the time (10 minutes after caffeine application) atwhich arterial rings were frozen (F) for posterior measurement ofRhoA activity. D, Statistical analysis of the effect of FPL onRhoA activation at different times during depletion of intracellu-lar calcium stores (n�4). Experimental conditions similar to (A).E, Representative trace of FPL-evoked contraction (2 minutes,5.34�0.44 mN, n�16; 10 minutes, 0.2�0.8 mN, n�15; 30 min-utes, 0.25�0.2 mN, n�15; 60 minutes, 0.1�0.5, n�15). F, Sta-tistical analysis of RhoA activation in response to FPL(0.5 �mol/L) in arterial rings bathed with 0Ca � EGTA externalsolution (n�3). Vertical arrows indicate the time at which arterialrings were frozen (F) for posterior measurement of RhoA activ-ity. Values are presented as mean�SEM. **P�0.01.


06/06/2013 153

Therefore, it is possible that a residual transmembrane Ca2�

influx, without effect on the contractile machinery itself,contributes to refilling of peripherally located SR, in whichmembranes of the plasmalemma and the superficial SR comeclose together.34 Accumulation of Ca2� in the SR would thenpermit subsequent release of enough Ca2� to activate orsensitize the contractile apparatus. This proposal explainswhy both the phasic and tonic components of contractiondepend on extracellular Ca2�. The higher sensitivity of thetonic phase of contraction to nifedipine could be explainedbecause it is known that this drug has more affinity for theinactivated state of the Ca2� channels than for channels thatare in the closed or open states.35 Because before stimulationchannels are mainly in the closed state, a higher dose of theantagonist is required to block the phasic component of thecontraction.

Sustained Depolarization-Evoked Contraction RequiresL-type Ca2� Channel-Dependent MetabotropicRhoA/ROCK ActivationIn accord with previous observations,9,27 we have shown inimmunostained isolated myocytes the redistribution of acti-vated RhoA and its displacement toward the plasma mem-brane after high K� induced depolarization. In intact arteries,pharmacological inhibition of ROCK (a downstream effectorof RhoA) selectively inhibited the tonic phase of contractionwithout altering the Ca2� signal. Therefore, the data sug-gested that besides metabotropic Ca2� release, activation ofthe RhoA/ROCK pathway is necessary for tonic myocytecontraction. In reversibly permeabilized arteries, RhoA activ-ity was found increased in sustained depolarized arteries, andboth contraction and RhoA activation decreased in parallelafter G-protein blockade with GDP�S or inhibition of InsP3

Figure 8. Ca2�-dependence of the caffeine-evoked and FPL64176 (FPL)-evoked contractions in rabbit basilar arterial rings. A,Statistical analysis of the caffeine (10 mmol/L, n�5) evoked contraction at different times in 0Ca solution. B, Statistical analysis of theFPL-evoked (0.5 �mol/L) RhoA activation at different times in 0Ca solution (n�9 for t�0; n�3 for t�60 minutes). C, Representativetrace of the time course of the FPL-evoked contraction (2 minutes, 19.18�1.12 mN, n�17; 10 minutes, 26.73�1.23 mN, n�35; 30 min-utes, 24.92�0.72 mN, n�16). D, Statistical analysis of the time course of FPL-evoked RhoA activation in arterial rings bathed in nomi-nally 0 external Ca2� (n�3 for 2 and 30 minutes; n�8 for 10 minutes). Vertical arrows indicate the time at which arterial rings were fro-zen (F) for posterior measurement of RhoA activity. Values are presented as mean�SEM. E, Model illustrating the multiple roles ofL-type Ca2� channel in the initiation and maintenance of arterial contraction. Membrane depolarization induces opening of voltage-gated Ca2� channels and a sudden increase of cytosolic [Ca2�] because of massive Ca2� influx. The increase of cytosolic [Ca2�] acti-vates the calmodulin-dependent myosin light chain kinase to overcome the activity of myosin light chain phosphatase, thus generatingthe phasic component of contractions. During a maintained depolarization, activated Ca2� channels induce a metabotropic cascadeleading to inositol trisphosphate synthesis and Ca2� release from sarcoplasmic reticulum (SR). These events activate the RhoA/Rho-associated kinase sensitization pathway and myosin light chain phosphatase inhibition, maintaining the contraction. During maintaineddepolarizations, a residual Ca2� current might contribute to refilling of SR Ca2� store.


06/06/2013 154

receptors with heparin. Hence, high K�-induced sustainedcontraction requires G-protein-mediated RhoA activation andCa2� sensitization, a phenomenon that was previouslythought to be restricted only to receptor-mediated forms ofsmooth muscle activation.3

Because RhoA/Rho kinase is activated by G-protein–coupled receptor stimulation,4,5 the apparent high K�-induced activation of RhoA/ROCK could, in fact, be attrib-utable to activation of a G-protein–coupled receptor withintrinsic voltage sensitivity rather than to membrane depolar-ization.36 This possibility was discarded because FPL, a Ca2�

channel agonist that can activate Ca2� channels withoutaffecting membrane potential,14,18 induced arterial contrac-tion and RhoA activation in a qualitatively similar way to thechanges in the same parameters evoked by KCl. Moreover, inconditional Cav1.2 knockout mice, the absence of FPL-induced or depolarization-induced L-type Ca2� channelsactivation was paralleled by the lack of RhoA activation andcontraction. In this animal model, contraction induced bycaffeine, which bypassed plasma membrane depolarization,was unchanged. Similarly, RhoA activation induced by phen-ylephrine, which is G-protein-mediated, was also unaltered.Therefore, these data strongly suggest that functional L-typeCa2� channels are required for the high K�- evoked or FPL-evoked RhoA activation. This novel role for VGCC on Ca2�

sensitization in VSM helps explain the KCl-evoked Rho/ROCKactivation profusely described in the literature and whose under-lying mechanisms were unknown.6–10 Our experimental dataalso indicate that InsP3-mediated Ca2� release is necessary fordepolarization-evoked or FPL-evoked RhoA activation, becauseit was markedly decreased in arterial rings in which the internalCa2� stores had been previously depleted. Interestingly,caffeine-induced Ca2� release did not induce RhoA activation,thus suggesting that SR Ca2� release per se, bypassing activationof Ca2� channels and the downstream metabotropic pathway, isnot sufficient to induced RhoA activation. Therefore, RhoA/ROCK activation and VSM sensitization requires both Ca2�

channel activation and CCICR from SR.In conclusion, the findings reported here suggest that CCICR

links Ca2� channel activation with RhoA/ROCK stimulation.They indicate that depolarization-evoked sustained RhoA acti-vation and myocyte contraction do not depend on the change inthe membrane potential itself or the mere release of Ca2� fromthe SR, but they require the simultaneous activation of VGCCand the downstream stimulation of a metabotropic pathway,leading to InsP3 synthesis and Ca2� release. An explanatorysummary of the ionotropic and metabotropic functions of VGCCregulating VSM contraction is given in Figure 8E. Our resultscould have a wide functional relevance for the pathogenesis ofvasospasms mediated by vasoactive agents, such as noradrena-lin, endothelin, or ATP, that can activate VGCC.18,37,38 Becausesustained VSM depolarization and VGCC activation mediatenumerous pathophysiological processes,39–42 the data could alsohelp to optimize therapeutic treatment for clinical conditionssuch as hypertension, angina pectoris, and cardiac arrhythmias,in which Ca2� channels antagonist are recommended.43

Sources of FundingThis work was supported by grant PI060137 and Red RECAVA ofthe Spanish Ministry of Health, and by the “Proyecto de Excelencia

(P08-CTS-03530)” of the “Consejería de Innovacion y Ciencia de laJunta de Andalucía and European Union.” Support was also providedby the Marcelino Botín Foundation and the Spanish Ministry ofScience and Innovation.

AcknowledgmentsThe authors thank Dr Franz Hofmann (Munich Technical University)for providing the Cav1.2 knockout mice and Dr Konstantin Levitsky(Instituto de Biomedicina de Sevilla) for technical help withconfocal microscopy.

DisclosuresNone.

References1. Uehata M, Ishizaki T, Satoh H, Ono T, Kawahara T, Morishita T,

Tamakawa H, Yamagami K, Inui J, Maekawa M, Narumiya S. Calciumsensitization of smooth muscle mediated by a Rho-associated proteinkinase in hypertension. Nature. 1997;389:990–994.

2. Masumoto A, Mohri M, Shimokawa H, Urakami L, Usui M, Takeshita A.Suppression of coronary artery spasm by the Rho-kinase inhibitor fasudilin patients with vasospastic angina. Circulation. 2002;105:1545–1547.

3. Kitazawa T, Gaylinn BD, Denney GH, Somlyo AP. G-protein-mediatedCa2� sensitization of smooth muscle contraction through myosin lightchain phosphorylation. J Biol Chem. 1991;266:1708–1715.

4. Gohla A, Schultz G, Offermanns S. Role for G(12)/G(13) in agonist-inducedvascular smooth muscle cell contraction. Circ Res. 2000;87:221–227.

5. Chikumi H, Vazquez-Prado J, Servitja JM, Miyazaki H, Gutkind JS.Potent activation of RhoA by Galpha q and Gq-coupled receptors. J BiolChem. 2002;277:27130–27134.

6. Yanagisawa T, Okada Y. KCl depolarization increases Ca2� sensitivity ofcontractile elements in coronary arterial smooth muscle. Am J Physiol.1994;267:H614–H621.

7. Mita M, Yanagihara H, Hishinuma S, Saito M, Walsh MP. Membranedepolarization-induced contraction of rat caudal arterial smooth muscleinvolves Rho-associated kinase. Biochem J. 2002;364:431–440.

8. Sakurada S, Takuwa N, Sugimoto N, Wang Y, Seto M, Sasaki Y, TakuwaY. Ca2�-dependent activation of Rho and Rho kinase in membranedepolarization-induced and receptor stimulation-induced vascular smoothmuscle contraction. Circ Res. 2003;93:548–556.

9. Urban NH, Berg KM, Ratz PH. K� depolarization induces RhoA kinasetranslocation to caveolae and Ca2� sensitization of arterial muscle. Am JPhysiol. 2003;285:C1377–C1385.

10. Ratz PH, Berg KM, Urban NH, Miner AS. Regulation of smooth musclecalcium sensitivity: KCl as a calcium-sensitizing stimulus. Am J Physiol.2005;288:C769–C783.

11. Best L, Bolton TB. Depolarization of guinea-pig visceral smooth musclecauses hydrolysis of inositol phospholipids. NS Arch Pharmacol. 1986;333:78–82.

12. Itoh T, Seki N, Suzuki S, Ito S, Kajikuri J, Kuriyama H. Membranehyperpolarization inhibits agonist-induced synthesis of inositol 1,4,5-trisphosphate in rabbit mesenteric artery. J Physiol. 1992;451:307–328.

13. Ganitkevich VYa, Isenberg G. Membrane potential modulates inositol1,4,5-trisphosphate-mediated Ca2� transients in guinea-pig coronarymyocytes. J Physiol. 1993;470:35–44.

14. del Valle-Rodríguez A, Lopez-Barneo J, Urena J. Ca2� channel-sarcoplasmic reticulum coupling: a mechanism of arterial myocyte con-traction without Ca2� influx. EMBO J. 2003;22:4337–4345.

15. Ben-Chaim Y, Tour O, Dascal N, Parnas I, Parnas H. The M2 muscarinicG-protein-coupled receptor is voltage-sensitive. J Biol Chem. 2003;278:22482–22491.

16. Mahaut-Smith MP, Martinez-Pinna J, Gurung IS. A role for membranepotential in regulating GPCRs? Trends Pharmacol Sci. 2008;29:421–429.

17. Liu QH, Zheng YM, Korde AS, Yadav VR, Rathore R, Wess J, WangYX. Membrane depolarization causes a direct activation of G protein-coupled receptors leading to local Ca2� release in smooth muscle. ProcNatl Acad Sci U S A. 2009;106:11418–11423.

18. Del Valle-Rodríguez A, Calderon E, Ruiz M, Ordonez A, Lopez-BarneoJ, Urena J. Metabotropic Ca(2�) channel-induced Ca(2�) release andATP-dependent facilitation of arterial myocyte contraction. Proc NatlAcad Sci U S A. 2006;103:4316–4321.

19. Calderon-Sanchez E, Fernandez-Tenorio M, Ordonez A, Lopez-Barneo J,Urena J. Hypoxia inhibits vasoconstriction induced by metabotropic Ca2�


06/06/2013 155

channel-induced Ca2� release in mammalian coronary arteries. Car-diovasc Res. 2009;82:115–124.

20. Fernandez-Tenorio M, Gonzalez-Rodríguez P, Porras C, Castellano A,Moosmang S, Hofmann F, Urena J, Lopez-Barneo J. Genetic ablation ofL-type Ca2� channels abolishes depolarization-induced Ca2� release inarterial smooth muscle. Circ Res. 2010;106:1285–1289.

21. Araya R, Liberona JL, Cardenas JC, Riveros N, Estrada M, Powell JA,Carrasco MA, Jaimovich E. Dihydropyridine receptors as voltage sensorsfor a depolarization-evoked, IP3R-mediated, slow calcium signal inskeletal muscle cells. J Gen Physiol. 2003;121:3–16.

22. Moosmang S, Schulla V, Welling A, Feil R, Feil S, Wegener JW, HofmannF, Klugbauer N. Dominant role of smooth muscle L-type calcium channelCav1.2 for blood pressure regulation. EMBO J. 2003;22:6027–6034.

23. Kobayashi S, Kitazawa T, Somlyo AV, Somlyo AP. Cytosolic heparininhibits muscarinic and alpha-adrenergic Ca2� release in smooth muscle.Physiological role of inositol 1,4,5-trisphosphate in pharmacomechanicalcoupling. J Biol Chem. 1989;264:17997–18004.

24. Lesh RE, Somlyo AP, Owens GK, Somlyo AV. Reversible permeabilization.A novel technique for the intracellular introduction of antisense oligodeoxy-nucleotides into intact smooth muscle. Circ Res. 1995;77:220–230.

25. Urena J, del Valle-Rodríguez A, Lopez-Barneo J. Metabotropic Ca2�

channel-induced calcium release in vascular smooth muscle. CellCalcium. 2007;42:513–520.

26. Katsuragi T, Usune S, Furukawa T. Antagonism by nifedipine of con-traction and Ca2�-influx evoked by ATP in guinea-pig urinary bladder.Br J Pharmacol. 1990;100:370–374.

27. Gong MC, Fujihara H, Somlyo AV, Somlyo AP. Translocation of rhoAassociated with Ca2� sensitization of smooth muscle. J Biol Chem.1997;272:10704–10709.

28. Berridge MJ. Inositol trisphosphate and calcium signalling. Nature. 1993;361:315–325.

29. Kobayashi S, Kanaide H, Nakamura M. Complete overlap of caffeine- andK� depolarization-sensitive intracellular calcium storage site in cultured ratarterial smooth muscle cells. J Biol Chem. 1986;261:15709–15713.

30. Pozzan T, Rizzuto R, Volpe P, Meldolesi J. Molecular and cellularphysiology of intracellular calcium stores. Physiol Rev. 1994;74:595–636.

31. Himpens B, Matthijs G, Somlyo AV, Butler TM, Somlyo AP. Cyto-plasmic free calcium, myosin light chain phosphorylation, and force inphasic and tonic smooth muscle. J Gen Physiol. 1988;92:713–729.

32. Langton PD, Standen NB. Calcium currents elicited by voltages steps andsteady voltages in myocytes isolated from the basilar artery. J Physiol.1993;469:535–548.

33. Fleischmann BK, Murray RK, Kotlikoff MI. Voltage window for sus-tained elevation of cytosolic calcium in smooth muscle cells. Proc NatAcad Sci U S A. 1994;91:11914–11918.

34. Shmygol A, Wray S. Functional architecture of the SR calcium store inuterine smooth muscle. Cell Calcium. 2004;35:501–508.

35. Bean BP. Nitrendipine block of cardiac calcium channels: high-affinitybinding to the inactivated state. Proc Natl Acad Sci U S A. 1984;81:6388–6392.

36. Ben-Chaim Y, Chanda B, Dascal N, Bezanilla F, Parnas I, Parnas H.Movement of ‘gating charge’ is coupled to ligand binding in a G-protein-coupled receptor. Nature. 2006;444:106–109.

37. Nelson MT, Standen NB, Brayden JE, Worley JF III. Noradrenalinecontracts arteries by activating voltage-dependent calcium channels.Nature. 1988;336:382–385.

38. Goto K, Kasuya Y, Matsuki N, Takuwa Y, Kurihara H, Ishikawa T,Kimura S, Yanagisawa M, Masaki T. Endothelin activates the dihydro-pyridine-sensitive, voltaje-dependent Ca2� channel in vascular smoothmuscle. Proc Natl Acad Sci U S A. 1989;86:3915–3918.

39. Harder DR, Dernbach P, Waters A. Possible cellular mechanism forcerebral vasospasm after experimental subarachnoid hemorrhage in thedog. J Clin Invest. 1987;80:875–880.

40. Martens JR, Gelband CH. Alterations in rat interlobar artery membranepotential and K� channels in genetic and nongenetic hypertension. CircRes. 1996;79:295–301.

41. Simard JM, Li X, Tewari K. Increase in functional Ca2� channels incerebral smooth muscle with renal hypertension. Circ Res. 1998;82:1330–1337.

42. Pesic A, Madden JA, Pesic M, Rusch NJ. High blood pressureupregulates arterial L-type Ca2� channels: is membrane depolarizationthe signal? Circ Res. 2004;94:e97–e104.

43. Abernethy DR, Schwartz JB. Calcium-antagonist drugs. N Engl J Med.1999;341:1447–1457.

Novelty and Significance

What Is Known?

● L-type Ca2� channels constitute an important pathway for extracel-lular Ca2� influx and vascular smooth muscle contraction.

● In the absence of extracellular Ca2�, Ca2� channels in vascularmyocytes act as voltage sensors that couple membrane depolar-ization to G-protein/PLC/InsP3 synthesis and Ca2� release fromthe sarcoplasmic reticulum (SR) (calcium channel-induced cal-cium release [CCICR]). Both agonist stimulation and membranedepolarization can evoke Ca2�-dependent RhoA/Rho kinase acti-vation and arterial contraction.

What New Information Does This Article Contribute?

● Depolarization or agonist (ATP)-evoked sustained arterial contractionrequires metabotropic Ca2� release from the SR.

● L-type Ca2� channel activation and metabotropic Ca2�channel-inducedCa2� release play an essential role in depolarization-evoked RhoA/Rho kinase activation and sustained myocyte contraction.

● Depolarization-evoked sustained RhoA activation does not depend onthe change in membrane potential itself or the mere release ofCa2� from the SR, but it requires the simultaneous activation ofvoltage-gated calcium channels (VGCC) and the downstreamstimulation of a metabotropic pathway, leading to InsP3 synthesisand Ca2� release.

Although a wide number of investigations have revealed theimportant contribution of L-type Ca2� channels and RhoA/Rhokinase in the pathophysiology of vascular smooth muscle, theinter-relations between both elements are unknown. We showfor the first time to our knowledge that the depolarization-evoked sustained contraction in these cells is mediated, at leastin part, by Ca2� channel-induced metabotropic Ca2� releasefrom the SR. We also demonstrate that sarcolemmal Ca2�

channels not only mediate the Ca2� influx that triggers con-traction but also metabotropically regulate SR Ca2� release andRhoA/Rho kinase activation responsible for sustained contrac-tion. These novel findings assign to L-type channels a role in theregulation of vascular smooth muscle tonus that is muchbroader than previously thought and help in the understanding ofthe role of Ca2� channels and the RhoA/Rho kinase sensitizationpathway in sustained contraction. Our results could have specialrelevance for understanding the pathogenesis of vasospasmsmediated by vasoactive agents that can activate VGCC, such asnoradrenalin, endothelin, or ATP. Because sustained vascularsmooth muscle depolarization and VGCC activation participate innumerous pathophysiological processes, our data could alsohelp to optimize therapeutic treatment for clinical conditionssuch as hypertension, angina pectoris, and cardiac arrhythmias,in which Ca2� channels antagonist are recommended.


06/06/2013 156

Kim Rouven Liedtke

ROS implication in a new antitumor strategy based on non-thermal plasma Quelle: Int. J. Cancer: 130, 2185–2194 (2012) VC 2011 UICC Betreuer: PD Dr. Wolfram von Bernstorff (Viszeral-, Thorax- und Gefäßchirurgie) Was bedeutet mir das Thema persönlich? Das in diesem Paper behandelte Thema spricht einige Aspekte meiner eigenen Forschungsarbeit an. Daher ist es ein wichtiger Baustein und teilweise auch Grundlage für weiterführende Versuche. Worauf kommt es mir bei diesem Thema am meisten an? Im Besonderen kommt es mir darauf an die wissenschaftlichen Methoden der Arbeit mit Plasma zu vermit-teln. In dieser Hinsicht bietet das Paper eine sehr gute Plattform, da sowohl in vitro als auch in vivo Maß-nahmen behandelt werden. Weiterhin war es mir wichtig, dass es sich um ein medizinisch und nicht physi-kalisch oder chemisch dominiertes Thema handelt, welches ich meiner Einschätzung nach in diesem Paper gefunden habe. Was fasziniert mich selbst am Thema am meisten? Am Faszinierendsten ist es für mich, dass in der Literatur mittlerweile von einer neuen Strategie und einem Paradigmenwechsel in der Anti-Tumor-Therapie gesprochen wird. Also sind das Thema und die Forschung daran tatsächlich wichtig und vielversprechend, sowie anwendungsorientiert. Ich denke insbesondere der letzte Punkt und die traurige Tatsache, dass so viele Menschen früher oder später mit dem Gebiet der Onkologie – sei es direkt oder auch indirekt – in Berührung kommen, macht es `massenkompatibel´. Was gefällt mir am Thema weniger? Leider hat man den Eindruck, dass in diesem Paper bei der Vielzahl von verwendeten Methoden keine so wirklich konsequent zu Ende gedacht und geführt wurde. Beispielsweise wird es versäumt, aus der Mög-lichkeit der indirekten Plasmaanwendung die richtigen Schlüsse zu ziehen, das Potential zu erkennen und dann eine Arbeitshypothese auszuarbeiten. Somit bleibt das Paper insgesamt relativ oberflächlich. Anders-herum formuliert kann man jedoch sagen, dass dieses Paper einen guten Einstieg in die Thematik vermit-telt und dabei einen recht breit gefächerten Überblick bietet.

06/06/2013 157

ROS implication in a new antitumor strategy based onnon-thermal plasma

Marc Vandamme1,2,3, Eric Robert2, Stephanie Lerondel3, Vanessa Sarron2, Delphine Ries2, Sebastien Dozias2,

Julien Sobilo3, David Gosset4, Claudine Kieda4, Brigitte Legrain5, Jean-Michel Pouvesle2 and Alain Le Pape3,6

1 GERMITEC SAS, 30 rue Mozart, 92110 CLICHY, France2 GREMI UMR-6606 CNRS, Universite d’Orleans, 14 rue d’Issoudun - BP 6744, 45067 ORLEANS cedex 2, France3 TAAM-CIPA, UPS44 CNRS, 3B rue de la Ferollerie, 45071 ORLEANS cedex 2, France4 Plateforme de Cytometrie et d’imagerie cellulaire, Centre de Biophysique Moleculaire, UPR4301 CNRS, rue Charles Sadron,

45071 ORLEANS cedex 2, France5 NOVAXIA, ZA Petit Four, 41220 Saint Laurent Nouan6 Inserm U618, Universite Francois Rabelais, Tours, France

Non-thermal plasma (NTP) is generated by ionizing neutral gas molecules/atoms leading to a highly reactive gas at ambient

temperature containing excited molecules, reactive species and generating transient electric fields. Given its potential to

interact with tissue or cells without a significant temperature increase, NTP appears as a promising approach for the

treatment of various diseases including cancer. The aim of our study was to evaluate the interest of NTP both in vitro and in

vivo. To this end, we evaluated the antitumor activity of NTP in vitro on two human cancer cell lines (glioblastoma U87MG and

colorectal carcinoma HCT-116). Our data showed that NTP generated a large amount of reactive oxygen species (ROS), leading

to the formation of DNA damages. This resulted in a multiphase cell cycle arrest and a subsequent apoptosis induction. In

addition, in vivo experiments on U87MG bearing mice showed that NTP induced a reduction of bioluminescence and tumor

volume as compared to nontreated mice. An induction of apoptosis was also observed together with an accumulation of cells

in S phase of the cell cycle suggesting an arrest of tumor proliferation. In conclusion, we demonstrated here that the

potential of NTP to generate ROS renders this strategy particularly promising in the context of tumor treatment.

Plasma, considered as the fourth state of the matter, has al-ready a broad range of applications in industry1 and in medi-cine.2,3 Recently, the development of a new kind of plasmadevices generating non-thermal plasma (NTP) has extendedtheir potential applications especially in biology and medi-cine.4–6 NTP with a temperature less than 40�C at the pointof treatment is a partially ionized media generated by excita-

tion of a gas mixture in a discharge reactor. It contains elec-trons, positive/negative ions, radicals, various excited mole-cules, energetic photons (UV) and generates transient electricfield. Given these interesting properties, potential applicationsare blood coagulation,7 skin decontamination without signifi-cant skin damages,8 wound healing,9 and tumor treatment.10

The dose of NTP delivered is an important parameter toinduce biological responses in tissue and cells.8 Indeed, lowdose of plasma (<1 J cm�2) is able to induce inactivation ofbacteria and proliferation of cells,11,12 while higher dose (>7J cm�2) can induce apoptosis of tumor cells including mela-noma, breast cancer cells and hepatocellular carcinoma.13–17

Sensenig et al. and Kim et al. have suggested that DNA dam-ages and reactive species generated by plasma could be themain causes of this effect.14,18 In a recent work, Kalghatgiet al. showed that a low dose of NTP enhances endothelialcell proliferation due to the reactive oxygen species (ROS)generated by NTP mediated FGF-2 release.11 On non-tumori-genic breast epithelial cell line, NTP was also recentlydescribed to induce DNA damage leading to apoptosis due tothe formation of intracellular ROS.19 ROS are potentiallyharmful on cellular metabolism by affecting cell functionswith a direct effect on cell development, growth, survival aswell as tumorigenesis.20 As described for NTP, the effect ofROS on cells is dose-dependent. While low doses of ROSinduce mutagenesis and cell proliferation, high levels not

Key words: ROS, DNA damages, cancer therapy, non-thermal plasma,

xenograft model

Abbreviations: BLI: bioluminescence imaging; BrdU:

bromodeoxyuridine; DHE: dihydroethidium; DNA: deoxyribonucleic

acid; FE-DBD: floating electrode dielectric barrier discharge; FITC:

fluorescein isothiocyanate; FLI: fluorescence imaging; H2DCFDA: 2, 7-

dichlorodihydrofluorescein diacetate; MTT: 3-(4,5-dimethylthiazol-2-

yl)-2,5-diphenyltetrazolium bromide; NAC: N-Acetyl-L-cysteine; NTP:

non-thermal plasma; ROS: reactive oxygen species; RS: reactive species

Additional Supporting Information may be found in the online

version of this article.

DOI: 10.1002/ijc.26252

History: Received 26 Feb 2011; Accepted 30 May 2011; Online 23

Jun 2011

Correspondence to: Marc Vandamme, GREMI UMR 6606, 14 rue

d’Issoudun - BP6744, 45067 ORLEANS cedex 2, France,

Tel.: þ332-3849-4531, Fax: þ3-323-841-7154,

E-mail: [email protected]

Can

cerTherapy

Int. J. Cancer: 130, 2185–2194 (2012) VC 2011 UICC

International Journal of Cancer

IJC

06/06/2013 158

only inhibit cell proliferation, but also induce a high cyto-toxic effect to the cell and can lead to apoptosis of a widerange of tumors.20–22

Radiotherapy mechanisms are also based on the formationin targeted cells of ROS including superoxide, hydrogen per-oxide and free hydroxyl, that are able to induce lethal DNAdamages which activate cell cycle checkpoints and initiate sig-naling cascades leading to cell death.23,24 NTP, thanks to itsability to generate in situ ROS at the vicinity of the tumor,appears thus as a good candidate for cancer treatment.Indeed, recent development in plasma sources allows thepropagation of cold plasma in small capillary as plasma jet,25

so opening perspective to treat such tumor types as colo-rectal, lung or pancreas tumor with a flexible micro plasmaendoscope (Fig. 1). A preliminary survival study has shownan antitumor activity of plasma floating electrode dielectricbarrier discharge (FE-DBD) in vivo with a significant lifespanincrease in mice bearing U87MG malignant glioma.26 Thisplasma is generated in the gap between an insulated highvoltage electrode and the skin or the tissue (floating elec-trode) (Fig. 1). Very few studies have described mechanismsimplied in the effect of NTP on tumor cells and further char-acterization of global effect of plasma on cells is needed.

The aim of this work was to evaluate antitumor effect ofNTP in vitro along with mechanisms underlying massive celldeath induction. We first documented the in vitro antitumoractivity of plasma using MTT and bioluminescence assay.Then, we considered the role of ROS in this antitumor activ-ity as well as formation of DNA strand breaks, cell cyclemodifications and apoptosis induction.

On U87MG glioma bearing mice, we evaluated effect ofNTP on tumor volume and the consequences of treatmenton cell proliferation, cell cycle and apoptosis induction. Forsuch a goal, bioluminescence imaging (BLI), an imaging mo-dality dependent on cell metabolism and proliferation, is aunique resource when exploited in association with cellularbiomarkers.

Material and MethodsCell culture

U87MG-Luc2 and HCT-116-Luc2 Bioware Ultra celllines were purchased from CaliperV

R

(Caliper Life Sciences,Roissy, France) and maintained in a humidified incubatorcontaining 5% CO2 at 37�C. Cells were cultured in MEMsupplemented with 10% fetal bovine serum, 1% glutamineand 1% penicillin-streptomycin (10,000 U/mL). Both celllines are stably transfected with firefly luciferase gene, thusallowing BLI.

Animals and tumors

Swiss nude female mice, 4 weeks of age, were purchased fromCharles RiverVR (Saint-Germain-sur-l’Arbresle, France) andacclimated for 1 week in the laboratory before experimentation.Animals were housed in plastic cages inside a controlled venti-lated rack with free access to water and food ad libidum. Allexperiments were performed in accordance with national animalcare guidelines (EC directive 86/609/CEE, French decree no 87-848). Tumor xenografts, plasma treatment and BLI were carriedout under general anesthesia obtained with 2% isoflurane in air(AerraneV

R

, Maurepas, France). Tumor xenografts were achievedby subcutaneous injection of tumor cells suspension (106 cells in0.1 mL 0.9% NaCl) into the hind legs. To follow tumor growth,tumor volume (V in mm3) was measured with a caliper and wascalculated as V ¼ (length � width2)/2.

Plasma treatment

Plasma was generated with a FE-DBD by applying a micro-second pulse voltage of 23 kV. Plasma is generated in openair in the gap fixed at 2 mm between the insulated electrode(one mm thick glass) and the sample. Discharge power den-sity of our DBD was 0.52 watts at 2000 Hz, thus a 15s treat-ment using the 0.78 cm2 insulated reactor lead to a dose of10 J/cm2. More detailed characteristics of plasma sources arepresented in Supporting Information.

Figure 1. Non-thermal plasmas. NTP generated with DBD system (left panel). NTP generated at the extremity of a very small capillary so

called ‘‘plasma gun’’ (right panel), plasma expands over 12 cm in a 4 mm inner diameter glass capillary and then in ambient air along

about 2 cm.

Can

cerTherapy

2186 ROS and antitumor effect of non-thermal plasma

Int. J. Cancer: 130, 2185–2194 (2012) VC 2011 UICC06/06/2013 159

In in vitro experiments, U87MG and HCT-116 cells(1x105) were seeded in 24-well plates 24 hr before plasmatreatment. Direct plasma treatment of cells seeded in 24-wellplates was performed in open air, 2 mm upper the mediumof each well containing adherent cells and 500 lL of medium(Supporting Information Fig. S1). For indirect treatment,plasma was applied in a well containing 500 lL mediumonly (without cells). Then, treated medium was immediatelyand carefully dropped in wells containing cells. During thesetwo exposure conditions, the increase in temperature of themedium was observed to be less than 1�C and no significantpH modification was observed (data not shown).

For in vivo studies, when tumors reached a volume of 1256 50 mm3 (D0), mice were randomly assigned into twogroups. In the CTRL group, mice were not treated while in NTPgroup, mice received plasma treatment. Each group of treatmentincluded 16 mice. Tumor treatment was performed daily at 200Hz during 6 min for five consecutive days (corresponding to120 J/cm2/day), this treatment dose was previously determinedwith a tolerance study.26 During all treatment procedure anes-thetized mice were placed on a temperature regulated silver plateand plasma reactor was positioned at a distance of two milli-meters above the tumor (Supporting Information Fig. S1). After24 hr of the last day treatment, all tumors were excised and halfof them were used for cytometry analyses (eight tumors pergroup) and the others tumors were used for histology analyses(eight tumors per group).

Bioluminescence

Tumor growth of treated and nontreated mice was monitoredby tumor volume measurement and BLI. BLI is an imaging mo-dality allowing the evaluation of very early stages of antitumoreffect prior to physical reduction of the tumor and biolumines-cence intensity is closely dependent on the tumor activity.

In vivo BLI: BLI imaging of mice was performed beforethe first treatment (D0) then during treatment course (at D3)and 24 hr after the end of treatment (at D5) using IVISLumina and analyzed with Living Image software (CaliperLife Sciences, Roissy, France). To this end, animals were in-traperitoneally injected with 100 mg/kg luciferin potassiumsalt (Promega, Paris, France) and imaging was acquired 6min after substrate injection under general anesthesia in thedark box of a high sensitivity CCD camera cooled to �90�C(IVIS Lumina II, Caliper Life Sciences, Roissy, France). Ac-quisition settings (binning and duration) were set up depend-ing upon tumor activity at the time of the acquisition andimages were acquired and analyzed using Living Image soft-ware (Caliper Life Sciences, Roissy, France).

Cell growth assay

HCT-116 and U87 cell number and cell growth inhibition byNTP were determined by trypan blue exclusion assay, by BLIand by measuring 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-tetrazolium bromide (MTT) dye absorbance by living cells.Cells were treated as described and incubated for additional

24 hr. In vitro BLI: 300 lg/mL luciferin potassium salt(Promega, Paris, France) was added in the well plate and theBLI was recorded after 5 min incubation at 37�C using IVISLumina (Caliper). BLI intensity of each well in photon/secwas normalized to nontreated cells. Concerning MTT assay,MTT solution (2.5 mg/mL in PBS) was added to each welland cells were incubated for 4 hr. Formazan crystals resultingfrom MTT reduction were dissolved by the addition of 10%SDS in DMSO/acetic acid solution per well. The relativequantity of formazan products formed in each well wasdetected by reading absorbance at 570 nm.

ROS inhibition

To inhibit ROS, N-acetyl-L-cysteine (NAC, Sigma-Aldrich,Lyon, France), an intracellular ROS scavenger, was used.Cells were pretreated with 4 mM NAC for 30 min priortreatment by NTP.

Detection of ROS levels in the medium

ROS such as H2O2,lOH and ONOOl were detected using

an oxidation-sensitive fluorescent probe dye 2, 7-dichlorodi-hydrofluorescein diacetate (H2DCFDA, Sigma-Aldrich, Lyon,France, Ex/Em ¼ 495/529 nm). As H2DCFDA is poorlyselective for O2

l-, dihydroethidium (DHE, Sigma-Aldrich,Lyon, France, Ex/Em ¼ 518/605 nm), which is highly selec-tive for O2

l-, was used for its detection (detailed in Support-ing Information). Fluorescence was acquired using IVISLumina with the suitable filters and analyzed with LivingImage software (Caliper Life Sciences, Roissy, France).

Apoptosis detection

In vitro, cell apoptosis was measured 24 hr after NTP by flowcytometry using an annexin V-FITC apoptosis detection kit(R&D Systems, Lille, France). Non adherent and adherentcells were collected 24 hr after treatment. Using an AnnexinV-FITC Apoptosis Detection Kit (R&D Systems, Lille, France),cells were stained with Annexin V-FITC and propidium iodide (PI)according to the manufacturer’s instructions. For each measure-ment, at least 20,000 cells were analyzed by flow cytometry using aFACSort (Becton andDickinson, Le pont de Claix, France).

In vivo, 24 hr after the end of the treatment, tumors wereexcised and apoptosis indexes were determined by immuno-histochemistry detection of cleaved caspase 3 with SignalStainCleaved Caspase-3 IHC kit (Cell Signaling, Saint QuentinYvelines, France) according to the manufacturer. Theseexperiments were realized under double blind analysis in col-laboration with Novaxia, Saint Laurent Nouan, France, acompany specialized in histology-pathology. The percentageof labeled tumor cells was evaluated in adjacent areas of thetumor mass free from necrosis. Additional details are givenin Supporting Information.

Cell cycle analysis and cell proliferation

Cell cycle analysis was performed 24 hr after NTP by com-bined propidium iodide and bromodeoxyuridine (BrdU,

Can

cerTherapy

Vandamme et al. 2187


Sigma-Aldrich, Lyon, France). Briefly, after methanol fixationand Triton X-100 permeabilization, cells were incubated for30 min with fluorescein isothiocyanate (FITC)-conjugatedanti-BrdU antibody (7 lg/mL, Millipore, Molsheim, France). AnIgG1 mouse secondary FITC-conjugated (Millipore, Molsheim,France) was used as negative control for non-specific fluores-cence. As a positive control, we used cells treated by topotecan(0.2lM). Cells were finally incubated with 5 lg/mL propidiumiodide (Sigma-Aldrich, Lyon, France) prior to analysis using aFACSort (Becton and Dickinson, Le pont de Claix, France).Additional details are given in Supporting Information.

Cell cycle distribution in vivo was determined by DNAcontent analysis after propidium iodide staining as describedearlier.27,28 Briefly, tumor samples were incubated with tryp-sin (0.03 mg/mL) and were dissociated using a Potter pestle.Samples were then treated with RNAse A (0.1 mg/mL) andincubated with propidium iodide (0.4 mg/mL) for 30 min inthe dark. Then samples were filtered on a nylon mesh priorto analysis using a FACSort (Becton and Dickinson, Le pontde Claix, France).

DNA damages

To detect DNA damages after NTP treatment, assessment ofcH2AX immunofluorescence (IF) by flow cytometry was per-formed 1 hr after treatment using cH2AX phosphorylationAssay Kit (Millipore, Molsheim, France) in accordance withmanufacturer protocol. Briefly, after ethanol fixation and sap-onin cell permeabilization, histone H2AX phosphorylated atserine 139 was detected by the addition of the antiphospho-Histone cH2AX, FITC conjugate. As a negative control fornonspecific binding an IgG1 mouse secondary FITC-conju-gated (Millipore, Molsheim, France) was used. As a positivecontrol, we used cells treated by topotecan (0.2lM). Thedegree of cH2AX phosphorylation was analysis using a FAC-Sort (Becton and Dickinson, Le pont de Claix, France).


Unless otherwise noted, data were reported as a mean 6 SEM.Mann-Whitney test was used to evaluate the statistical signifi-cance of the results. DL50 were determined according to theHill slope method (GraphPad Prism 5.0, La Jolla, CA, USA).Differences were considered significant at p values < 0.05. Forall cytometry analyses, RAW data are presented in SupportingInformation Figure S2.

ResultsNTP presents a major antitumor effect in vitro

We first evaluated the antitumor potential of NTP on twodifferent cell lines in vitro using cell viability (cell counts)and metabolism assays (MTT and BLI). As described in Ta-ble 1, 24 hr after treatment, NTP induced a major decreasein cell number as compared to CTRL. In U87MG cell line, a10 J/cm2 treatment induced a 46% significant decrease in liv-ing cells (p ¼ 0.04) associated to 312% significant increase indead cells (p ¼ 0.02). Using 20 J/cm2, almost all cells weredead after 24 hr. These results are in accordance with

Table 1. Cell number after NTP treatment

U87 HCT-116

Groups Living cells Dead cells Living cells Dead cells

CTRL 81,000 6 5,300 6,250 6 144 118,000 6 6,831 17,000 6 1,915

10 J/cm2 44,000 6 2,000 * 19,500 6 3,700 * 56,333 6 4,240 ** 26,000 6 2,582 *

20 J/cm2 3,000 6 1,700 * 34,000 6 3,200 * 12,222 6 1,899 ** 57,111 6 3,068 *

Data are presented as mean 6 SEM.*p < 0.05; **p < 0.01 vs. CTRL.

Figure 2. Antitumor responses to NTP treatments. U87MG (a) and

HCT-116 (b) cells were treated by increasing doses of NTP. Cell

viability was determined 24 hr after treatment by both MTT (l)

and BLI (n) assays. Cell viability of NTP treated cells was

normalized to untreated cells.

Can

cerTherapy



metabolism assays, which showed a similar decrease in cellviability for the two cell lines with both MTT and BLI (Figs.2a, 2b and Supporting Information Fig. S3). The DL50 was�9 J/cm2 in U87MG and �8 J/cm2 in HCT-116, respectively.To understand which components of plasma are involved inthis effect, further investigations were realized using bothMTT and BLI. Considering the concordance between thesetwo tests, only BLI data are presented.

Plasma generated species in the culture medium are the

main cause of cell death

To identify plasma components implicated in cell death,direct and indirect treatments were performed. For U87MG,10 and 20 J/cm2 direct plasma treatment induced a 40% and96% decrease in cell viability, respectively (p ¼ 0.009 and p¼ 0.01). As shown in Figure 3a, a similar decrease wasobserved for both direct and indirect treatment. The sameresults were observed using HCT-116 cells (Supporting Infor-mation Fig. S4a). To confirm the implication of the NTPtreated culture medium in this effect, a quantification of ROSproduction was performed with an oxidation-sensitive fluo-rescent probe dye. In Figure 3d, using H2O2 specific dye(H2DCFA), 10 or 20 J/cm2 NTP induced a 5 fold and a 7fold increase in levels in the medium, respectively (p ¼ 0.002and p ¼ 0.002). Normalization with H2O2 standard concen-

trations revealed that plasma generated an equivalent concen-tration of 600 lM H2O2 for 10 J/cm2 and 900 lM H2O2 for20 J/cm2 treatments.

A second dye with a high specificity for O2� (DHE) showed

an increase in ROS in the both group of treatment (Fig. 3c);20% for 10 J/cm2 (p ¼ 0.02) and 52% for 20 J/cm2 (p ¼ 0.002).To confirm the implication of ROS, cells were pre incubatedwith intracellular ROS scavenger NAC, followed by NTP treat-ment (Fig. 3b). After 24 hr, CTRL cells and CTRL cells withNAC alone did not evidence any cell viability modification(data not shown). In NAC pretreated groups, no significantdecrease in cell viability was observed for both NTP doses, ascompared to CTRL. Similar results were observed in HCT-116cell line (Supporting Information Fig. S4b).

NTP induces a cell cycle arrest and an apoptosis induction

To determine the effect of plasma on cell cycle distribution, aniodide propidium staining was performed 24 hr after plasmatreatment. In U87MG cells, plasma induced a significantdecrease of cells in G0/G1 phase with a significant increase incells in S-phase for 10 and 20 J/cm2 groups as compared toCTRL (Fig. 4a). Concerning G2/M phase, a significant increasewas observed only with 10 J/cm2 of plasma (p ¼ 0.005).

In HCT-116 cells, plasma induced also a significantdecrease in G0/G1 cells for both doses (p ¼ 0.002) while a

Figure 3. ROS implication in the antitumor effect of NTP. (a) A direct and indirect treatment was performed on U87MG and cell viability was

determined by BLI imaging and was normalized to non-treated cells (CTRL). (b) U87MG cells were pretreated with NAC (4 mM), a ROS

scavenger. Cells with or without NAC were directly treated by plasma and viability was determined by BLI. (c) ROS such as H2O2, was

detected using an oxidation-sensitive fluorescent probe dye 2,7-dichlorodihydrofluorescein diacetate (H2DCFDA), fluorescence levels were

expressed as normalized fluorescence intensity. (d) A second dye Dihydroethidium (DHE), which is highly selective for O2l- was used,

fluorescence levels were expressed as normalized fluorescence intensity. *p < 0.05; **p < 0.01.

Can

cerTherapy



significant increase in cells in S phase was observed for 20 J/cm2 only (p ¼ 0.002; Supporting Information Fig. S5a).

A significant increase in cells in G2/M phase was observedfor both treated groups as compared to CTRL (p ¼ 0.005).These results suggested an accumulation of cells resultingfrom a cell cycle arrest, so a cell proliferation assay withBrdU incorporation was performed.

In U87MG (Fig. 4b), NTP induced a 92% and 97%decrease in BrdU incorporation for 10 and 20 J/cm2, respec-tively (p ¼ 0.005). Similar results were observed in HCT-116cell line (Supporting Information Fig. S5b).

To determine whether the decrease in cell proliferationand the cell cycle arrest lead to apoptosis induction, anannexin V staining was performed 24 hr after the treatment.As presented in Figure 4c, plasma treatment induced apopto-sis induction in both cell lines. In U87MG cells, both dosesresulted in 3 and 12 fold increase as compared to CTRLgroup, respectively (p ¼ 0.002). In HCT-116, this apoptosisinduction was observed for both treated groups (SupportingInformation Fig. S6a). For both cell lines, a significantincrease in late apoptosis (Annexin Vþ/PIþ) was also

observed. Moreover, the presence of cells in late apoptosiswas further confirmed by the observed induction of capase3/caspase 7 activity 24 hr after plasma application (SupportingInformation Fig. S7).

NTP induces DNA damages

To determine whether plasma treatment induced DNA dam-ages, assessment of cH2AX IF by flow cytometry was per-formed 1 hr after treatment. In both cell lines, a majorincrease in cH2AX IF was observed after 10 or 20 J/cm2 ofNTP. For U87MG (Fig. 4d), NTP induced a significantincrease of 20 and 40% of cH2AX IF in 10 or 20 J/cm2

group, respectively (p ¼ 0.002; p ¼ 0.0007). In HCT-116(Supporting Information Fig. S6b), this increase was 41 and91% for 10 and 20 J/cm2 of NTP, respectively (p ¼ 0.004; p¼ 0.0007).

NTP exhibits a significant antitumor effect in vivo

Plasma was applied each day during 5 consecutive days onsubcutaneously U87MG-Luc grafted tumors. Tumors wereincluded in treatment groups when they reached 125 6 50

Figure 4. Cellular effects of NTP on U87MG cells. U87MG cells were NTP treated using two different doses (10 J/cm2 and 20 J/cm2). (a) Cell

cycle distribution: 24 hr after treatment, cell cycle distribution was assessed by flow cytometry after propidium iodide staining. (b) Cell

proliferation was evaluated by incorporation of BrdU into S phase by flow cytometry. (c) Cell apoptosis was measured 24 hr after NTP

treatment by annexin V-propidium iodide labeling. (d) DNA damages induction: assessment of cH2AX immunofluorescence (IF) by flow

cytometry was performed 1 hr after treatment. cH2AX IF was normalized to nontreated cells. *p < 0.05; **p < 0.01; ***p < 0.001.

Can

cerTherapy



mm3. Mean tumor volumes were similar between the twogroups: at D0, mean tumor volumes 6 standard error inCTRL and treated mice were 129 6 13 mm3 and 124 6 14mm3, respectively. To determined effect of NTP on tumorcell proliferation and metabolism, BLI was performed before(D0), during (D3) and after (D5) treatment course (Figs. 5aand 5b). In non-treated and treated mice groups, BLI activityshowed 3 fold increase between D0 and D3 (p < 0.0001).From D3 to D5 a twofold increase was observed for theCTRL group only (p ¼ 0.02), while in plasma treated tumors,BLI intensity remained stable over the D3-D5 period.

Tumor volume measurements performed 24 hr after theend of the treatment course (Fig. 5c) revealed a significantlower tumor volume in the treated group (233 mm3) as com-pared to CTRL (400 mm3), (p ¼ 0.008).

NTP induces a cell cycle arrest and apoptosis in tumor

To determine whether plasma induced apoptosis in thesetumors, a caspase 3 immunostaining was performed 24 hr af-ter the last day of treatment (Figs. 5d and 5e) resulting in asignificant 3 fold increase in caspase-3 positive cells (p ¼0.008). Positive cells for caspase 3 immunostaining

Figure 5. In vivo evaluation of NTP antitumor activity. When tumor reached 150 6 50 mm3, mice were randomly assigned into two groups:

control (CTRL) and plasma, eight mice per group. Plasma treatment was delivered each day for five consecutive days (6 min, 200 Hz). Mice

in both groups were sacrificed 24 hr after the last treatment. (a) BLI imaging was performed before the first treatment (D0), during

treatment course (D3) and 24h after the end of treatment protocol (D5). Tumor BLI was normalized to signal at D0. (b) Representative BLI

imaging of CTRL and NTP treated mice at D5. (c) Tumor volume was determined using a caliper 24h after the last day of treatment (D5). (d)

Apoptosis indexes were determined by immunohistochemical detection of cleaved caspase 3. (e) Representative cleaved caspase 3

immunostaining obtained in CTRL and NTP-treated tumors. Magnification x400. (f) Cell cycle distribution was assessed by flow cytometry

after propidium iodide staining. *,# p < 0.05; *** p < 0.001.Can

cerTherapy



represented 4.5 and 13.5% of tumor cells for CTRL andtreated tumors, respectively. Moreover, caspase 3 positivecells were homogeneously distributed in sagittal sectionsfrom treated tumors.

Consequences of plasma on cell cycle distribution weredetermined 24 hr after the last day of treatment (Fig. 5f). Asobserved in vitro, propidium iodide staining showed a signifi-cant decrease in G0/G1 cells in tumor after treatment ascompared to CTRL (p ¼ 0.022). Cells in G0/G1 phasedecrease from 81% to 73% after NTP and this effect wasassociated with a significant 1.5 fold increase in cells in Sphase (p ¼ 0.002) from 13 to 20% in CTRL and Plasmagroups, respectively.

DiscussionIn this study, we evaluated the antitumor potential of NTPon two different cell lines and determined some of theunderlying biological mechanisms. Metabolism assays showedthat NTP had a significant antitumor activity on U87MGand HCT-116 cell lines with an IC50 of �9 J/cm2 and �8 J/cm2, respectively, corresponding to �12s and �13.5s expo-sure duration. In these two cell lines, a dose >20 J/cm2 ledto about 100% cell death. These results were in accordancewith previous published studies which showed an antitumoractivity of NTP on various cell lines including melanoma,colorectal and hepatocellular carcinoma.13–19 In our condi-tions, plasma is generated in ambient air and main activespecies in the gas are reactive species (RS) such as OHl,H2O2, N2*, NO and O2

l�. To identify components involvedin the in vitro antitumor effect, direct and indirect modalitiesof treatment were performed. In the indirect configuration,cells were exposed to the NTP treated medium only and notto the other plasma components such as UV and electricfield. These results and the use of NAC suggest that RS werethe main agent involved in this effect. Using H2O2 standardconcentrations and an oxidant probe dye, an equivalent max-imum concentration of 600 lM H2O2 after 10 J/cm2 of NTPwas observed. This dye is not specific for H2O2 only andDL50 of U87MG after H2O2 determined by BLI and MTTwas 350 lM (data not shown) suggesting that equivalentconcentration of H2O2 was overestimated and other RS arepresent in the medium.

After NTP treatment, a cell cycle arrest in S and G2/Mphase was observed, confirmed with BrdU incorporationwhich showed a very low proliferation rate. These data areconsistent with those of Yan et al. reporting an in vitro cellcycle arrest associated to a modification of cyclin levels witha plasma jet treatment.16 This cell cycle arrest was similar tothe multiphase cell cycle arrest observed after ROS exposurewith high doses of H2O2.

29,30 Dose dependent exposure ofROS leads to cell death,20–22 as observed with NTP. Our dataare in strong correlation with those from others studieswhich reported the effects of NTP on various tumorigenicand non-tumorigenic mammalian cell lines,13–19 thus suggest-

ing a common mechanism of action of NTP, whatever celltype considered.

Radiation therapy, a major antitumor modality that alsoimplies ROS, is able to induce lethal DNA damages whichactivate cell cycle checkpoints and initiate signaling cascadesleading to cell death by generating ROS in cell’s DNA vicin-ity. With NTP, we observed a similar result with the forma-tion of DNA damages in treated cells, leading to cell cyclearrest and finally to cell death with both early and late apo-ptosis. DNA damages are observed as soon as 1 hr after thetreatment, whereas a minimal 3 hr delay is required for phos-phorylation of cH2AX histone triggered by apoptosis associ-ated DNA fragmentation.31 This suggests that the signaldetected was a direct consequence of the treatment and didnot correspond to strand breaks associated with apoptosisinduction and subsequent DNA fragmentation. Phosphoryla-tion of cH2AX on Ser139 after DNA damage can be medi-ated by ATM, ATR and DNA-PK. In a recent work onMCF10A cell line, an induction of cH2AX-phosphorylationwas observed after plasma treatment and the authors showedthat ATM is not the primary mediator of H2AX phosphoryl-ation on Ser139 as observed after ionizing radiation.19 AfterNTP treatment, phosphorylation of H2AX occurs primarilythrough ATR and authors suggest that this activation is theconsequences of stalled replication forks formation afterNTP. Interaction of plasma with DNA needs further investi-gation to understand which kind of DNA damage was prefer-entially induced after treatment.

Given these encouraging results in vitro, we further inves-tigated the potential antitumor properties of NTP treatmentin human malignant glioma xenografted onto nude mice. Inthe present study, as compared to nontreated mice, plasmaexposure induced a significant inhibition of tumor growth(�40%) at the end of the treatment. This effect is confirmedby BLI, a gene expression imaging modality closely depend-ent upon aerobic metabolism, which showed a stabilizationof tumor activity during treatment course (D3-D5). Thesefindings are consistent with a previous work from our lab,showing a 60% lifespan increase in mice bearing U87MGxenografts.10 Given the high chemo and radioresistance ofthe model used here, a significant antitumor effect on suchan aggressive tumor model suggests a high efficacy of ourapproach and evidences the interest of this strategy for can-cer therapy. Higher treatment doses should be of valuable in-terest to enhance this effect more especially that a recent invitro study has also shown an important discrepancy of cellsensitivity between tumor and non-tumor cells.32

The in vivo tumor volume and proliferation stabilizationobtained could be linked to changes in cell cycle distributionwith a decrease in G0/G1 phase and an increase in cells in Sphase. Accumulation of tumor cells in S phase caused byplasma may be mediated by DNA strand break formation,even if this point remains to be proven in vivo. Moreover, ashistologically evidenced, apoptosis induction occurs in thewhole tumor. Even if there is no direct evidence that plasma

Can

cerTherapy



penetrate skin, these data suggest that plasma componentsact either by penetrating in the tissue either by inducingROS releases in the tumor. We can also hypothesize that,as described after ionizing radiation treatment, a cell deathinduced by bystander effect could explain the effect ofplasma inside the tumor.33 Even if free radicals only pene-trate to a reduced extent into the tissues, this property isof particular interest in the perspective of a direct in situapplication of NTP for antitumor treatment, avoiding severesystemic side effects. The other components of plasma,such as UV or heat generation, might also play a role inplasma antitumor effect, even if the available data seemonly reflect a minor contribution in this effect. Indeed UVpenetration in the skin is limited to a few lm and ourtreatment conditions only generate a temperature increaseof about to 3–4�C during a few minutes.26 However, thecombination of effects of the different plasma componentsmight result in a synergistic global effect. Indeed, mecha-nisms of interaction between plasma and cells and plasma

skin penetration remain to be further documented since allplasma components could induce membrane damages,changes in intracellular signaling pathways and exhibit cyto-toxic properties.

Taken together, our results suggest that NTP could be anew strategy against cancer cells, more especially when usedin combination with new generation antitumor drugs. Indeed,recently, Kim et al have demonstrated that association ofNTP to EGF-R targeted therapy resulted in a synergisticeffect.15 The in vivo good tolerability and the effectiveness ofNTP towards tumor cells together with the ability to deliverNTP via a small capillary open new interesting perspectivesfor loco-regional or in situ applications. Lung and colorectaltumors or dysplasia using plasma endoscopic applications arecurrently our first therapeutic targets.

AcknowledgementsThe authors are grateful to the Region Centre (APR ‘‘PlasMed’’), Germitec,Conseil Regional du Centre, CNRS and CG45.

References

1. Bogaerts A, Neyts E, Gijbels R, van derMullen J. Gas discharge plasmas and theirapplications. Spectrochim Acta Part B 2002;57:609–58.

2. Zenker M. Argon plasma coagulation. GMSKrankenhhyg Interdiszip 2008;3:Doc15.

3. Moisan M, Barbeau J, Moreau S, PelletierJ, Tabrizian M, Yahia LH. Low-temperature sterilization using gas plasmas:a review of the experiments and ananalysis of the inactivation mechanisms.Int J Pharm 2001;226:1–21.

4. Kong MG, Kroesen G, Morfill G, NosenkoT, Shimizu T, van Dijk J, Zimmermann JL.Plasma medicine: an introductory review.New J Phys 2009;11:1–35.

5. Laroussi M. Low-temperature plasmas formedicine? IEEE Trans Plasma Sci 2009;37:714–25.

6. Stoffels E, Sakiyama Y, Graves DB. Coldatmospheric plasma: charged species andtheir interactions with cells and tissues.IEEE Trans Plasma Sci 2008;36:1441–57.

7. Kalghatgi SU, Fridman G, Cooper M,Nagaraj G, Peddinghaus M,Balasubramanian M, Vasilets VN, GutsolAF, Fridman A, Friedman G. Mechanismof blood coagulation by nonthermalatmospheric pressure dielectric barrierdischarge plasma. IEEE Trans Plasma Sci2007;35:1559–66.

8. Fridman G, Friedman G, Gutsol A,Shekhter AB, Vasilets VN, Fridman A.Applied plasma medicine. Plasma ProcessPolym 2008;5:503–33.

9. Isbary G, Morfill G, Schmidt HU, GeorgiM, Ramrath K, Heinlin J, Karrer S,Landthaler M, Shimizu T, Steffes B, BunkW, Monetti R, et al. A first prospective

randomized controlled trial to decreasebacterial load using cold atmospheric argonplasma on chronic wounds in patients. Br JDermatol 2010;163:78–82.

10. Vandamme M, Robert E, Dozias S, Sobilo J,Lerondel S, Le Pape A, Pouvesle JM.Response of human glioma U87 Xenograftedon mice to non thermal plasma treatment.Plasma Med 2011;1:27–43.

11. Kalghatgi S, Friedman G, Fridman A,Clyne AM. Endothelial cell proliferation isenhanced by low dose non-thermal plasmathrough fibroblast growth factor-2 release.Ann Biomed Eng 2010;38:748–57.

12. Laroussi M. Low temperature plasma-basedsterilization: overview and state-of-the-art.plasma process. Polym 2005;2:391–400.

13. Fridman G, Shereshevsky A, Jost MM,Brooks AD, Fridman A, Gutsol A, VasiletsV, Friedman G. Floating electrodedielectric barrier discharge plasma in airpromoting apoptotic behavior inmelanoma skin cancer cell lines. PlasmaChem Plasma Process 2007;27:163–76.

14. Kim CH, Bahn JH, Lee SH, Kim GY, JunSI, Lee K, Baek SJ. Induction of cell growtharrest by atmospheric non-thermal plasmain colorectal cancer cells. J Biotechnol 2010;150:530–38.

15. Kim GC, Kim GJ, Park SR, Jeon SM, SeoHJ, Iza F, Lee JK. Air plasma coupled withantibody-conjugated nanoparticles: a newweapon against cancer. J Phys D: Appl Phys2009;42:5.

16. Yan X, Zou F, Zhao S, Lu X, He G, XiongZ, Xiong Q, Zhao Q, Deng P, Hunag J,Yang G. On the mechanism of plasmainducing cell apoptosis. IEEE Trans PlasmaSci 2010;38:2451–7.

17. Zhang XH, Li MJ, Zhou RL, Feng KC,Yang SZ. Ablation of liver cancer cells invitro by a plasma needle. Appl Phys Lett2008;93:1–3.

18. Sensenig R, Kalghatgi S, Cerchar E,Fridman G, Shereshevsky A, Torabi B,Arjunan KP, Podolsky E, Fridman A,Friedman G, Azizkhan-Clifford J, BrooksAD. Non-thermal plasma inducesapoptosis in melanoma cells via productionof intracellular reactive oxygen species.Ann Biomed Eng 2011;39:674–87.

19. Kalghatgi S, Kelly CM, Cerchar E, TorabiB, Alekseev O, Fridman A, Friedman G,Azizkhan-Clifford J. Effects of non-thermalplasma on mammalian cells. PLoS One2011;6:e16270.

20. Dreher D, Junod AF. Role of oxygen freeradicals in cancer development. Eur JCancer 1996;32A:30–8.

21. Fang J, Nakamura H, Iyer AK. Tumor-targeted induction of oxystress for cancertherapy. J Drug Target 2007;15:475–86.

22. Simizu S, Takada M, Umezawa K, ImotoM. Requirement of caspase-3(-like)protease-mediated hydrogen peroxideproduction for apoptosis induced byvarious anticancer drugs. J Biol Chem 1998;273:26900–7.

23. Fei P, El-Deiry WS. P53 and radiationresponses. Oncogene 2003;22:5774–83.

24. Hall EJ. Radiobiology for theradiobiologist, 5th edn. Philadelphia:Lippincott Williams Et Wilkins, 2000.

25. Robert E, Barbosa E, Dozias S, VandammeM, Cachoncinlle C, Viladrosa R, PouvesleJM. Experimental study of a compactnanosecond plasma gun. Plasma ProcessPolym 2009;6:795–802.

Can

cerTherapy



26. Vandamme M, Robert E, Pesnel S, BarbosaE, Dozias S, Sobilo J, Lerondel S, Le PapeA, Pouvesle JM. Antitumor effect ofplasma treatment on U87 gliomaxenografts: preliminary results. plasmaprocess. Polym 2010;7:264–73.

27. Vindelov LL, Christensen IJ. A review oftechniques and results obtained in onelaboratory by an integrated system ofmethods designed for routine clinical flowcytometric DNA analysis. Cytometry 1990;11:753–70.

28. Labussiere M, Pinel S, Vandamme M,Plenat F, Chastagner P. Radiosensitizingproperties of bortezomib depend on

therapeutic schedule. Int J Radiat OncolBiol Phys 2011;79:892–900.

29. Barnouin K, Dubuisson ML, Child ES,Fernandez de Mattos S, Glassford J,Medema RH, Mann DJ, Lam EW. H2O2induces a transient multi-phase cell cyclearrest in mouse fibroblasts throughmodulating cyclin D and p21Cip1expression. J Biol Chem 2002;277:13761–70.

30. Chen QM, Bartholomew JC, Campisi J,Acosta M, Reagan JD, Ames BN.Molecular analysis of H2O2-inducedsenescent-like growth arrest in normalhuman fibroblasts: p53 and Rb control G1

arrest but not cell replication. Biochem J1998;332 (Part 1):43–50.

31. Huang X, Darzynkiewicz Z. Cytometricassessment of histone H2AXphosphorylation: a reporter of DNAdamage. Methods Mol Biol 2006;314:73–80.

32. Georgescu N, Lupu AR. Tumoral andnormal cells treatment with high-voltagepulsed cold atmospheric plasma jets. IEEETrans Plasma Sci 2010;38:1949–55.

33. Rzeszowska-Wolny J, Przybyszewski WM,Widel M. Ionizing radiation-inducedbystander effects, potential targets formodulation of radiotherapy. Eur JPharmacol 2009;625:156–64.

Can

cerTherapy



Alexandra Welz

Activation of Protease-Activated Receptor 2 Induces VEGF Independently of HIF-1 Quelle: PLOS ONE | www.plosone.org 1 September 2012 | Volume 7 | Issue 9 | e46087 Betreuer: Prof. Bernhard Rauch (Allgemeine Pharmakologie) Da ich mich in meiner Doktorarbeit ebenfalls mit dem Protease-aktivierten Rezeptor 2 beschäftige, fand ich es sehr interessant zu sehen, wie viele verschieden Funktionen dieser Rezeptor ausübt. Dieses Paper war eines der ersten, das nicht die Entzündungsfunktionen des PAR2 charakterisierte, sondern proangiogenetische Funktionen des PAR2 darstellt. Besonders gut an diesem Paper fand ich den Schreibstil. Der Methodenteil ist sehr anschaulich und aus-führlich erklärt, was es einem „Neuling“ in der Forschung erleichtert, die Methodik nachzuvollziehen. Zudem fand ich es erstaunlich, dass in dem Paper tatsächlich angemerkt wurde, dass ein anderes Ergeb-nis als das Gefundene erwartet wurde. Sonst erhält man immer den Eindruck, dass die Forscher genau das bestätigen, was sie auch erwartet haben. Nicht so gut gefallen hat mir, dass die Abbildungen und Diagramme so schlecht in den Text eingearbeitet sind. Außerdem ist das Design einiger Diagramme nicht unbedingt nachahmenswert und generell nicht einheitlich.

06/06/2013 168

Activation of Protease-Activated Receptor 2 InducesVEGF Independently of HIF-1Jeppe Grøndahl Rasmussen1,2, Simone Elkjær Riis2, Ole Frøbert3, Sufang Yang2,4, Jens Kastrup5,

Vladimir Zachar2, Ulf Simonsen1, Trine Fink2*

1Department of Biomedicine, Pulmonary and Cardiovascular Pharmacology, Aarhus University, Aarhus, Denmark, 2 Laboratory for Stem Cell Research, Aalborg University,

Aalborg, Denmark, 3Department of Cardiology, Orebro University Hospital, Orebro, Sweden, 4Animal Reproduction Institute, Guangxi University, Nanning, China,

5Cardiac Stem Cell Laboratory, The Heart Centre, Rigshospitalet, Copenhagen University Hospital, Copenhagen, Denmark

Abstract

Background: Human adipose stem cells (hASCs) can promote angiogenesis through secretion of proangiogenic factorssuch as vascular endothelial growth factor (VEGF). In other cell types, it has been shown that induction of VEGF is mediatedby both protease activated receptor 2 (PAR2) and hypoxia inducible factor 1(HIF-1). The present study hypothesized thatPAR2 stimulation through activation of kinase signaling cascades lead to induction of HIF-1 and secretion of VEGF.

Methodology/Principal Findings: Immunohistochemistry revealed the expression of PAR2 receptors on the surface ofhASCs. Blocking the PAR2 receptors with a specific antibody prior to trypsin treatment showed these receptors are involvedin trypsin-evoked increase in VEGF secretion from hASCs. Blocking with specific kinase inhibitors suggested that thatactivation of MEK/ERK and PI3-kinase/Akt pathways are involved in trypsin-eveoked induction of VEGF. The effect of thetrypsin treatment on the transcription of VEGF peaked at 6 hours after the treatment and was comparable to the activationobserved after keeping hASCs for 24 hours at 1% oxygen. In contrast to hypoxia, trypsin alone failed to induce HIF-1measured with ELISA, while the combination of trypsin and hypoxia had an additive effect on both VEGF transcription andsecretion, results which were confirmed by Western blot.

Conclusion: In hASCs trypsin and hypoxia induce VEGF expression through separate pathways.

Citation: Rasmussen JG, Riis SE, Frøbert O, Yang S, Kastrup J, et al. (2012) Activation of Protease-Activated Receptor 2 Induces VEGF Independently of HIF-1. PLoSONE 7(9): e46087. doi:10.1371/journal.pone.0046087

Editor: Fadi N. Salloum, Virginia Commonwealth University Medical center, United States of America

Received November 23, 2011; Accepted August 28, 2012; Published September 25, 2012

Copyright: � 2012 Rasmussen et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permitsunrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Funding: This work was supported by grants from the Danish Heart Foundation (08-4-R64-A2020-B584-22460 to J.G.R., 07-4-B584-A1453-22364 to J.G.R., and 10-04-R78-A2925-22599 to J.G.R.); Brødrene Hartmanns Foundation; Frimodt-Heineke Foundation; Torben and Alice Frimodt Foundation; Købmand Sven Hansen andhustru Ina Hansens Foundation; and the Faculty of Health Science, Aarhus University (to J.G.R.). The funders had no role in study design, data collection andanalysis, decision to publish, or preparation of the manuscript.

Competing Interests: The authors have declared that no competing interests exist.

* E-mail: [email protected]

Introduction

The transplantation of human adipose-derived stem cells

(hASCs) to induce angiogenesis is increasingly recognised as

a therapeutic modality in the treatment of ischemic disease [1,2,3].

In a previous study, we found that both hypoxic culture as well as

treatment with trypsin increases the pro-angiogenic potential of

hASCs [4]. The angiogenic effect induced by hASCs is mainly

paracrine, exerted through cytokines, such as the vascular

endothelial growth factor (VEGF) [5]. Hence, there is great

interest in attempting to increase VEGF expression in order to

optimise the effect of transplanted mesenchymal stem cells [6–8].

VEGF has been shown to be induced both by activation of

protease activated receptor 2 (PAR2) signalling and by the

transcription factor hypoxia inducible factor 1 (HIF-1) [9,10,11].

PAR2 is a G-protein coupled receptor that is activated by

proteolytic cleavage of a tethered ligand, and is known to be

activated by trypsin [12,13,14]. Previous studies have found that

different kinase cascades are implicated in PAR2 signaling

[9,15,16]. Thus, PAR2 was found to activate both the PI3K/

Akt and MEK/ERK pathways in GI epithelial cells [17], mainly

the Rho/ROCK pathway in lung epithelial cells [18], and only

the MEK/ERK pathway in glioblastoma cells [19]. PAR2 is not

expressed in all tissues [20], and so far it is unclear whether PAR2

are expressed in mesenchymal stem cells.

In contrast, HIF-1 has so far been found in most cell types and

tissues. HIF-1 is a master regulator in oxygen homeostasis and

drives the expression of a plethora of genes involved in metabolism

and angiogenesis, including VEGF. HIF-1 is a heterodimer

comprised of the subunits HIF-1a and the aryl hydrocarbon

receptor nuclear translocator (ARNT). In normoxic conditions

HIF-1a is continuously degraded. In hypoxia, however, HIF-1a is

stabilized, and dimerizes with ARNT to form HIF-1 [21].

Interestingly, is has recently been shown, that even in normoxia

activation of PI3K and ERK pathways may stabilize HIF-1a thus

leading to induction of VEGF [22,23]. Moreover, that hypoxia

and PAR2 activation may act synergistically in the promotion of

angiogenesis and that there could be possible crosstalk between the

protease-activated and the hypoxia-activated pathways [24,25].

Therefore, we hypothesized that PAR2 stimulation through

activation of kinase signaling cascades may lead to induction of

HIF-1 and secretion of VEGF. To address the hypothesis we

PLOS ONE | www.plosone.org 1 September 2012 | Volume 7 | Issue 9 | e4608706/06/2013 169

examined in hASCs the expression and the effect of stimulating

and blocking PAR2 receptors on VEGF, inhibitors of Rho kinase

(ROCK), PI3K, and MEK were applied and phosphorylation of

the downstream kinases and VEGF induction was examined.

Finally, the interaction of PAR2 activation and hypoxia on VEGF

and HIF-1 activation was investigated.

Methods

DonorsThis study conforms to the Declaration of Helsinki. All patients

gave written informed consent and the clinical protocol was

approved by the regional Committee on Biomedical Research

Ethics of Northern Jutland, Denmark (project no. VN 2005/54).

The adipose tissue was obtained during elective liposuction from

three healthy patients without cardiovascular disease and not

receiving any medication. The patients were one male and two

females aged 42, 58, and 52 years, respectively. The hASC cell

lines (ASC12, 21, and 23) were established as described previously,

[26] and have been characterized extensively [27,28].

Cell CultureThe hASCs were cultured in a-MEM medium with Glutamax

(Invitrogen, Taastrup, Denmark), 10% fetal bovine serum (FBS)

(Europa Bioproducts Ltd., Cambridge, United Kingdom),

100 IU/ml penicillin, 100 mg/ml streptomycin, and 50 mg/ml

gentamicin (all from Invitrogen). All experiments were performed

with cells from all three patients using cells at passages two to five.

FACS AnalysisSubconfluent cultures of hASCs at P2 were harvested and

strained through a 70 mm filter before they were resuspended in

2% FBS and 0.1% sodium azide in PBS. The incubation with

a particular antibody-fluorophore complex took place at 4uC for

30 min, after which the cells were fixed in 1% formaldehyde.

Antibodies were pre-labeled with Alexa 488, R-Phycoerythrin,

and Alexa 647 fluorophores using the Zenon labeling technology

(Invitrogen). Primary antibodies were directed against CD24,

CD29, CD73, CD117, CD133, CD146, CD200, CD271, HLA-

DR (all from Abcam, Cambridge, UK), CD34, CD44, CD45,

CD105, HLA-ABC (all from Dako, Glostrup, Denmark), CD90,

ABCG2 (both from Santa Cruz Biotechnology, Heidelberg,

Germany), and Stro-1 (Millipore Chemicon, Copenhagen, Den-

mark). Validity of analysis was supported by testing for viability

using a LIVE/DEAD Reduced Biohazard Viability/Cytotoxicity

Kit (Invitrogen). At least 104 events were recorded per sample on

a FACSCanto flow cytometer (BD Bioscience, Brøndby, Den-

mark). Data were analysed using FlowJo 7.2 software package

(TreeStar, Ashland, OR).

Immunofluorescence AssayFor detection of PAR2, hASCs were seeded at 36103 per cm2

in 8-well glass CultureSlide (Falcon; BD Bioscience) and main-

tained in 20% oxygen for three days before they were fixed in 4%

formaldehyde for 30 min. The cells were then reacted with 100-

fold diluted mouse monoclonal antibody specific for PAR2

(SAM11; Santa Cruz Biotechnology) or matching isotype control

(Dako) at 4uC overnight. The epitope sites were visualised with

Alexa 488 conjugated goat anti-mouse IgG and the nuclei were

counterstained with 1:1,000 Hoechst 33342 (both from Invitogen)

at 37uC for two hours. Microscopy was done using an Axio

Observer Z1 (Zeiss, Gottingen, Germany) and image acquisition

and processing with the aid of AxioVision 4.7 software (Zeiss).

Treatment with Varying Concentrations of TrypsinThe cultures were passaged using a standard trypsinisation

procedure based on a mixture of 0.125% trypsin and 0.01%

EDTA for 5 min at 37uC. Based on manufacturers information

regarding trypsin activity (3,600 BAEE U/mg), the 0.125% trypsin

was estimated to have a concentration of 9,000 nM, using a 1 U/

ml,2 nM conversion. For the studies on effect of trypsin

concentration, the hASCs were seeded in six-well plates (Costar)

at 36103 cells per cm2, cultured for three days, and subsequently

treated by either 9 nM, 90 nM, or 9,000 nM trypsin for 4 min,

after which the trypsin was neutralized through addition of fresh

media. Twelve hours after the trypsin treatment mRNA was

harvested for transcriptional analysis. For all other experiments

9,000 nM of trypsin was used.

Varying Length of Incubation after Trypsin TreatmentTo determine the optimal length of incubation after trypsin

treatment, the hASCs were seeded in six-well plates (Costar) at

86103 cells per cm2, cultured for three days, and subsequently

treated by trypsin for 4 min at 37uC, after which the trypsin was

neutralized through addition of fresh media. Two, six, and

12 hours after the trypsin treatment, mRNA was harvested for

transcriptional analysis. For all other experiments, cells were

incubated for six hours after trypsin treatment unless otherwise

stated.

Treatment by Trypsin and HypoxiaNormoxic cultures were performed in a standard incubator in

an atmosphere containing 5% CO2 and 20% O2. Hypoxic culture

experiments were performed in a hypoxic workstation (Xvivo;

BioSpherix, Redfield, NY) at 37uC in an atmosphere containing

5% CO2 balanced with nitrogen to reach oxygen concentration of

1%. The workstation allowed incubation and manipulation of the

cells at continuous hypoxic conditions.

For the studies on VEGF induction, the hASCs were seeded in

six-well plates (Costar) at 86103 cells per cm2, cultured for three

days, and subsequently treated by either trypsin, hypoxia or

Table 1. Primer sequences for primers used in quantitative real-time RT-PCR analysis.

Gene Forward Primer Sequence Reverse Primer Sequence

PPIA 59 TCC TGG CAT CTT GTC CAT G 39 59 CCA TCC AAC CAC TCA GTC TTG 39

YWHAZ 59 ACT TTT GGT ACA TTG TGG CTT CAA 39 59 CCG CCA GGA CAA ACC AGT AT 39

VEGF 59 CGA TTC AAG TGG GGA ATG G 39 59 CAT TGA TCC GGG TTT TAT CC 39

PAR2 59 TCC TCA CTG GAA AAC TGA CC 39 59 GGA AAA GAA AGA CCC ACA GG 39

doi:10.1371/journal.pone.0046087.t001

PAR2 Induces VEGF Independently of HIF-1


a combination of both. For the trypsin-only and treatment, cells

were treated with trypsin for 4 min at 37uC, after which the

trypsin was neutralized through addition of fresh media. Six hours

after the trypsin treatment, either mRNA was harvested for

transcriptional analysis or the media was changed and the cells

incubated an additional 12 hours, after which the media was

collected for analysis of secretion of VEGF. For the hypoxic

treatment, cells were transferred to 1% oxygen for 24 hours after

which either mRNA was harvested or the media was changed and

the cells incubated an additional 12 hours at 1% oxygen, for the

analysis of the VEGF secretion. For the trypsin-hypoxia combi-

nation, the cells were cultured for 18 hours in hypoxia, treated

with trypsin and refed with neutralizing media as described above,

all taking place under hypoxic conditions. After 6 hours the cells

were either harvested for transcriptional analysis or incubated in

fresh media for 12 hours under hypoxia for analysis of VEGF

secretion.

For studies on HIF-1a stabilisation by trypsin and hypoxia, the

cells were seeded in T75 cell culture flasks (Costar) at a density of

86103 cells per cm2 and otherwise treated as described above. In

addition, samples were harvested for analysis of trypsin effect 4

and 12 hours after the treatment.

Figure 1. Immunophenotypical analysis of hASC lines at passage 2. (A) Representative distributions of positive markers expressed on theASC12 cells are presented. (B) Surface markers profile was obtained as an average from ASC12, 21, and 23 lines.doi:10.1371/journal.pone.0046087.g001



Blocking of PAR2Cells were seeded in 12-well plates (Costar) at a density of

86103 cells per cm2 and after three days the blocking of PAR2 was

assayed using PAR2 (SAM11) (Santa Cruz Biotechnology)

antibody at 25 mg/ml both before and during trypsin exposure

for a total of 15 min. Cultures exposed to trypsin were used as

positive controls and cultures exposed to PBS were used to

normalise to the steady-state VEGF gene expression. After

replenishing with complete growth medium, the cultures were

allowed to recover for additional 12 hours, and finally processed

for total RNA to enable real-time RT-PCR analysis.

Inhibition of Kinase ActivityThe cultures were initiated at a density of 86103 cells per cm2

and allowed to progress for three days, when the trypsin exposure

was carried out. The selected kinase inhibitors, including 100 nM

Wortmannin and 50 mM LY294002 (both phosphoinositide 3-

kinase (PI3K)), 40 mM U0126 (mitogen activated protein kinase

MEK), and 10 mM Y27632 (Rho-associated protein kinase

(ROCK)), were allowed to take effect from one to two hours

prior to enzymatic exposure. After supplementing with complete

medium, the cultures were processed for protein analysis by

immunoblotting and total RNA within 15 min and 12 hours,

respectively.

ImmunoblottingThe cells were lysed in 50 mM Tris, 150 mM NaCl,

2 mM EDTA, 0.5% NP40, 0.5 mM phenylmethylsulfonyl fluo-

ride, 50 mM NaF, 0.04% b-mercaptoethanol, 1 mM Na3VO4

and 1 Complete Mini protease inhibitor cocktail tablet (Roche

Diagnostics, Hvidovre, Denmark) per 10 ml total lysis solution.

Protein concentration was measured in the lysates using the BCA

Protein Assay Kit (Thermo Fisher Scientific, Rockford, IL).

Samples amounting to 40 to 60 mg of total protein were heat

denatured, separated by SDS-PAGE using a 10% gel, and

transferred to nitrocellulose or PVDF membranes using the iBlot

transfer equipment (Invitrogen). Membranes were incubated at 4C

overnight with primary antibodies including rabbit polyclonal to

HIF-1a (Abcam) diluted 1:500, mouse monoclonal to b-actin

(Sigma-Aldrich, Brøndby, Denmark) diluted 1:10.000, rabbit

monoclonals to Erk1/2, phospho-Erk1/2, Akt, and phospho-Akt

(all from Cell Signaling Technology, Danvers, MA) diluted 1:500,

and finally mouse monoclonal to PAR2 (SAM11; Santa Cruz

Biotechnology) diluted 1:100. PBS containing 5% skimmed milk

and 1% Tween 20 (Sigma-Aldrich) was used for all dilutions.

Membranes were then incubated at room temperature for 1 hour

with HRP-conjugated secondary polyclonal antibodies rabbit anti-

mouse (Dako, Glostrup, Denmark) or goat anti-rabbit (Santa Cruz

Biotechnology) diluted 1:10.000 and 1:2.500, respectively. The

target proteins were visualised using chromogenic deposition

(tetramethylbenzidine; Sigma-Aldrich) or enhanced chemilumi-

nescence (Amersham ECL Plus; GE Healthcare Europe, Brøndby,

Denmark). Signal acquisition in the latter case was accomplished

in a Kodak Image Station 4000 mm Pro (Carestream Health

Denmark, Skovlunde, Denmark).

Real-time RT-PCRTotal RNA was isolated using the NucleoSpin RNA XS kit

(Macherey-Nagel, Duren, Germany) and the purity and concen-

tration were determined spectrophotometrically (Nanodrop;

Thermo Science, Wilmington, DE). Reverse transcription was

Figure 2. Expression of PAR2 in hASCs and its association withtranscriptional activation of VEGF. (A) Detection of PAR2 in hASClines by immunoblotting. (B) Expression of PAR2 by indirect immuno-fluorescence using SAM11 antibody. Representative pattern as detectedon the surface of ASC12 cells is presented. (C) SAM11 antibody blockedtrypsin-induced PAR2 activation, measured using real-time RT-PCR todetermine VEGF expression levels 12 hours after trypsin exposure(n = 15). Expression levels were corrected for basal VEGF activity inhASCs cultured at 20% oxygen and normalised to the levels induced bytrypsin. Values are represented as the mean and SEM. Scale barindicates 200 mm. Abbreviations: PAR2, protease-activated receptor 2;

VEGF, vascular endothelial growth factor; Ctrl, control (NIH 3T3 cells).doi:10.1371/journal.pone.0046087.g002



Figure 3. Trypsin-activated PAR2 intracellular signaling in hASCs. (A) Schematic rendition of signal-transduction pathways linking PAR2 andVEGF. (B) The effect of specific kinase inhibitors on suppressing trypsin-induced VEGF activation after 5 min trypsin exposure was assessed by real-time RT-PCR (n = 6). Expression levels were normalized to the levels induced by trypsin (Ctrl). (C) The effect of PI3K and Mek inhibitors on



performed with the iScript cDNA synthesis kit (Bio-Rad,

Copenhagen, Denmark). All primers (Table 1) were designed

using the open source software Primer3 and produced by DNA

Technology (Aarhus, Denmark). The amplification reactions were

performed on a My-Cycler real-time PCR system (Bio-Rad) in

a final volume of 25 ml containing 5 pmol of each primer, 0.25 ml

cDNA using the SYBR Green PCR supermix (Bio-Rad). The

thermal cycling protocol involved initial denaturation at 95uC for

3 min and was followed by 40 cycles of denaturation at 95uC for

10 sec and primer annealing and elongation for 30 sec at

predetermined optimal temperature. To test for the specificity of

the product, a melt curve function of the IQ5 Optical System

Software 2.1 (Bio-Rad) was invoked. A four-fold serially diluted

standard curve derived from a pool of all the cDNA samples was

used to calculate relative expression of each gene. For normal-

isation purposes, the geometric mean of two reference genes,

cyclophilin A (PPIA) and tyrosine 3/tryptophan 5-monooxygenase

activation protein (YWHAZ), was used [29].

Quantitation of Secreted VEGFFor all treatments, after the medium was harvested the cells

were lysed in 0.02% sodium dodecyl sulfate. The levels of VEGF

were assessed using a commercially available ELISA VEGF kit

(R&D Systems Europe, Abingdon, United Kingdom) and the cell

numbers were determined using a PicoGreen dsDNA Quantita-

tion Kit (Invitrogen). The fluorescence was measured using

a Wallac 1420 Victor Multilabel Counter (PerkinElmer, Hvidovre,

Denmark) with excitation and emission at 485 nm and 535 nm,

respectively. The secretion of VEGF was normalised with respect

to cell numbers.

Quantitation of HIF-1To investigate the influence of trypsin treatment on the

activation of HIF-1, the ASCs were seeded in duplicates in T75

CellstarH Tissue Culture Flasks (Greiner Bio-One, Cat.

no. 658175) and incubated under standard culturing conditions

for three days. Half of the cells were transferred to an incubator in

the BioSpherix glove box to receive hypoxic treatment. 6 h before

the hypoxic treatment was completed a group of cells from each

incubator system were treated with trypsin. After completion of

the hypoxic treatment, all cells were lysed and nuclear proteins

isolated using a Nuclear Extract Kit (Active Motif, Cat. no. 40010)

and the concentration of nuclear extract determined using a Pierce

BCA Protein Assay Kit (Thermo Scientific, Cat. no. 23227)

according to manufacturer’s instructions. Nuclear extract was kept

at –80uC.

The HIF-1 activation was measured in technical replicates using

a TransAM HIF-1 kit (Active Motif, Cat. no. 47096). For each

sample 20 mg was used per well. The absorbance was measured

using a Wallac 1420 Victor Multilabel Counter (PerkinElmer,

Hvidovre, Denmark) at a wavelength of 450 nm.

StatisticsWhen comparing more than two groups a one-way analysis of

variance (ANOVA) with Bonferroni’s post hoc test was used and

when comparing two samples, a Student’s t-test was used

(SigmaPlot 11.0; Systat Software, Erkrath, Germany). When

necessary, data were logarithmic transformed to display normal

distribution. Data are presented as mean 6 standard error of the

mean (SEM). Statistical significance was assigned to p values

,0.05. The experiments were done using all three cell lines and

when quantitative parameters were assessed the experiments were

repeated at least twice and replicates were involved.

Results

Characterization of hASCsFACS analyses of all three cell lines demonstrated that the

hASC cell lines were positive for the surface markers CD29,

CD44, CD73, CD90, CD105, and HLA-ABC and negative for

CD24, CD34, CD45, CD117, CD133, CD146, CD200, CD271,

ABCG2, HLA-DR, and Stro-1 (Fig. 1A and 1B). Moreover,

a narrow distribution of the marker values indicated a phenotypical

relatedness of the cell lines. In addition, the particular cell lines

displayed plastic adherence and have previously been demon-

strated to have a multilineage potential. [28,30] These character-

istics apply with the general accepted definitions of ASCs [31,32].

Role of PAR2 in Trypsin-mediated Upregulation of VEGFInitially, we investigated whether PAR2 is expressed by hASCs.

Immunoblot analysis revealed the presence of the PAR2 protein

(Fig. 2A) and, furthermore, immunofluorescence indicated that

PAR2 was localised at the plasma membrane (Fig. 2B). After

having confirmed the expression of PAR2, we aimed to link its

activation to upregulation of VEGF. To that end, we carried out

experiments where PAR2 was blocked by applying a specific

antibody. Blocking of PAR2 receptors resulted in a significant

suppression of trypsin-induced VEGF transcriptional upregulation

(Fig. 2C).

Identification of Molecular Mechanisms Underlying PAR2Mediated VEGF Induction

In order to identify molecular mechanisms downstream of

PAR2, we explored involvement of the kinase cascades PI3K,

MAPK, and ROCK that are known to be more or less involved in

PAR2-specific signal transduction in other cell lines (Fig. 3A). The

blocking experiments revealed that the transcriptional activation of

VEGF in hASCs in response to trypsin could be attenuated by

blocking the PI3K and MAPK pathways while blocking the

ROCK pathway had only minor effect (Fig. 3B). This finding was

further confirmed by immunoblotting experiments, demonstrating

that trypsin-dependent phosphorylation of Akt and Erk1/2 took

place, and that this effect was blocked by specific inhibitors of

PI3K and MEK (Fig. 3C).

Induction of VEGF in Response to Trypsin and HypoxiaTo determine the optimal conditions of trypsin treatment, first

several concentrations of trypsin were tested. When VEGF

expression levels were normalized to a PBS exposed control,

9 nM, 90 nM, and 9,000 nM trypsin upregulated VEGF 1.12,

1.21, and 1.52 times respectively with only the 9,000 nM trypsin

resulting in a significant upregulation (p,0.05). (Data not shown).

To determine when the expression of VEGF peaked after

trypsin treatment, cells were incubated for 2, 6, and 12 hours after

trypsin treatment, after which VEGF mRNA levels were de-

termined. In addition, to assess the combined effect of trypsin and

phosphorylation of Akt and Erk1/2, respectively, as a result of 5 min trypsin exposure was determined by immunoblotting. PI3K and Mek inhibitorswere added 2 hours prior to trypsin exposure. Cells after a 4-day culture at 20% oxygen were used as controls (Ctrl). Representative data obtainedfrom ASC12 cells are presented. Values are represented as the mean and SEM. Abbreviations: PAR2, protease-activated receptor 2; VEGF, vascularendothelial growth factor; Ctrl, control.doi:10.1371/journal.pone.0046087.g003



Figure 4. Effect of trypsin and culture at 1% oxygen on VEGF gene expression and secretion by hASCs. (A) The effect of trypsinexposure followed by various lengths of incubation alone or in combination with hypoxic exposure on VEGF gene expression was determined by real-time RT-PCR (n = 6–20). The transcriptional response was normalized to the geometric mean of reference genes PPIA and YWHAZ. Values arerepresented as the mean and SEM. For all time points the VEGF expression was significantly higher after trypsin treatment than in cells not treatedwith trypsin and also for all time points, the VEGF expression in hypoxia was significantly higher than the corresponding normoxic value (p,0.05) (B)Comparative analysis of the effect of trypsin or hypoxic exposure alone and in combination on VEGF secretion in the culture media was measured byELISA (n = 12). Values are represented as the mean and SEM. Asterisks denote statistical difference between this and all other groups (p,0.05).Abbreviations: VEGF, vascular endothelial growth factor.doi:10.1371/journal.pone.0046087.g004



hypoxia this series of experiments was performed both at ambient

oxygen and under hypoxic conditions. The data demonstrate the

expression of VEGF peaks 6 hours after trypsin treatment (Fig. 4A,

red bars) and that for all time points hypoxia has an additional

effect (Fig. 4A, blue bars).

The additive effect of trypsin and hypoxia was also evident

when looking at the amount of secreted VEGF into the media.

Both trypsin and hypoxic treatments alone led to significantly

higher levels of secreted VEGF than control conditions, and the

combination of the two resulted in the highest level of VEGF

secretion (Fig. 4B).

Effect of Trypsin and Hypoxia on HIF-1 InductionHIF-1a is responsible for VEGF upregulation in hypoxia,

therefore we investigated the possibility that trypsin-induced

upregulation of VEGF was effectuated through stabilisation of

HIF-1a. Cells were either subjected to trypsin, hypoxia or

a combination, after which HIF-1 activation/stabilization was

determined by ELISA (Fig. 5A). Both hypoxia and trypsin +hypoxia led to significantly increased levels of HIF-1. However,

treatment with trypsin failed to increase HIF-1 compared to

untreated cells. The cells in this first experiment were cultured for

6 hours after trypsin exposure. Thus, to exclude the possibility that

HIF-1a could have been stabilised in short term in the period

preceding or following this time point. cells were exposed to

trypsin and cultured for 4 or 12 hours prior to analysis of HIF-1

levels by immunoblotting (Fig. 5B). Cells cultured for 24 hours in

hypoxic conditions were used as a positive control. As shown in the

panels in figure 5B, for all cell lines HIF-1a was stabilised in 1%

oxygen culture and HIF-1 levels in trypsin-treated cells both 4 and

12 hours after trypsin treatment remained negligible. Cells

Figure 5. Stabilisation of HIF-1 in hASCs after trypsin and 1% oxygen exposure alone or in combination. (A) HIF-1 activation/stabilization during 6 hours culture after trypsin exposure in combination with 24 hours in hypoxic/normoxic conditions was analysed by ELISA. Allcells were harvested in situ. Values are represented as the mean and SEM (n= 12). Asterisks denote statistical difference between this and all othergroups (p,0.05). (B) Analysis of HIF-1a induction at 4 and 12 hours following 5 min trypsin exposure was done by immunoblotting. All cells wereharvested in situ. HIF-1a positive controls are ASCs subjected to 48 hours of 1% oxygen.doi:10.1371/journal.pone.0046087.g005



receiving no treatment had HIF-1 levels comparable to the cells

treated with trypsin (data not shown).

Discussion

In this study, we have found that PAR2 is expressed on the

surface of hASCs, that trypsin induces the expression of VEGF via

PAR2 activation and subsequent activation of PI3K and MEK

signalling pathways, and that this mechanism is independent of

HIF-1.

We are the first to demonstrate the presence of PAR2 on

hASCs, and to document that in hASCs the PAR2 activation leads

to downstream activation of PI3K and MAPK pathways, and that

the Rho/ROCK pathway is minimally involved. This finding

underscores the interesting observation that PAR2 activation has

very distinct down-stream effects depending on the cell type, as in

some cell types only the MAPK or the Rho/ROCK pathways are

activated [9,18,19].

Furthermore, the additive effect of hypoxia and trypsin

treatment on VEGF induction has not been described pre-

viously either. Several other groups have shown that activation

of PI3K and MAPK pathways may lead to either stabilization

or increased expression of HIF-1 [33,34]. Therefore, we

speculated that the trypsin-induced expression of VEGF may

be mediated via HIF-1 involvement. However, as we demon-

strate here for the first time in ASCs, trypsin-treatment alone

failed to stabilize HIF-1a. Therefore, it can be concluded that

the hypoxia-induced and trypsin-induced VEGF transcription

occurs through separate pathways.

The effect of hASCs transplantation in the treatment of

ischemic disease is believed to be mainly paracrine, with VEGF

playing a key role [1,2,5]. Therefore, it is an important discovery

that trypsin through PAR2 induces a level of VEGF upregulation

in hASCs comparable to that seen in hASCs cultured at 1%

oxygen. In addition, the present investigation, demonstrating

synergistic effects of hypoxia and trypsin on VEGF induction, may

lay the basis for a novel paradigm for preconditioning of the

hASCs by a combination of hypoxic culture and exposure to

trypsin prior to their clinical application. We expect that an

optimised implementation of exposure to a combination of

hypoxia and trypsin will form the basis for a highly efficient

preconditioning which will impact cellular therapies for ischemic

disorders. The clinical implications of these findings are significant

as ischemic disease is a major burden in the health care system and

the deleterious effects of ischemia could be ameliorated by

inducing angiogenesis in the ischemic tissue using ASCs with

potentiated VEGF secretion.

Acknowledgments

The authors wish to thank Helle Skjødt Møller for competent laboratory

assistance.

Author Contributions

Conceived and designed the experiments: JGR SR OF SY JK VZ US TF.

Performed the experiments: JGR SR SY. Analyzed the data: JGR SR SY

VZ US TF. Contributed reagents/materials/analysis tools: JGR SY VZ

US. Wrote the paper: JGR SR OF SY JK VZ US TF.

References

1. Bai X, Alt E (2010) Myocardial regeneration potential of adipose tissue-derived

stem cells. Biochemical and biophysical research communications 401: 321–326.

2. Miranville A, Heeschen C, Sengenes C, Curat CA, Busse R, et al. (2004)Improvement of postnatal neovascularization by human adipose tissue-derived

stem cells. Circulation 110: 349–355.

3. Zachar V, Duroux M, Emmersen J, Rasmussen JG, Pennisi CP, et al. (2011)Hypoxia and adipose-derived stem cell-based tissue regeneration and engineer-

ing. Expert opinion on biological therapy 11: 775–786.

4. Rasmussen JG, Frøbert O, Pilgaard L, Kastrup J, Simonsen U, et al. (2011)Prolonged hypoxic culture and trypsinization increase the pro-angiogenic

potential of human adipose tissue-derived stem cells. Cytotherapy 13: 318–328.

5. Sadat S, Gehmert S, Song Y-H, Yen Y, Bai X, et al. (2007) The cardioprotectiveeffect of mesenchymal stem cells is mediated by IGF-I and VEGF. Biochemical

and biophysical research communications 363: 674–679.

6. Rehman J, Traktuev D, Li J, Merfeld-Clauss S, Temm-Grove CJ, et al. (2004)Secretion of angiogenic and antiapoptotic factors by human adipose stromal

cells. Circulation 109: 1292–1298.

7. Herrmann JL, Wang Y, Abarbanell AM, Weil BR, Tan J, et al. (2010)Preconditioning mesenchymal stem cells with transforming growth factor-alpha

improves mesenchymal stem cell-mediated cardioprotection. Shock (Augusta,Ga) 33: 24–30.

8. Wang X, Zhao T, Huang W, Wang T, Qian J, et al. (2009) Hsp20-engineered

mesenchymal stem cells are resistant to oxidative stress via enhanced activationof Akt and increased secretion of growth factors. Stem cells (Dayton, Ohio) 27:

3021–3031.

9. Liu Y, Mueller BM (2006) Protease-activated receptor-2 regulates vascularendothelial growth factor expression in MDA-MB-231 cells via MAPK

pathways. Biochemical and biophysical research communications 344: 1263–1270.

10. Zhu T, Sennlaub F, Beauchamp MH, Fan L, Joyal JS, et al. (2006)

Proangiogenic effects of protease-activated receptor 2 are tumor necrosisfactor-alpha and consecutively Tie2 dependent. Arteriosclerosis, thrombosis, and

vascular biology 26: 744–750.

11. Forsythe JA, Jiang BH, Iyer NV, Agani F, Leung SW, et al. (1996) Activation ofvascular endothelial growth factor gene transcription by hypoxia-inducible factor

1. Molecular and cellular biology 16: 4604–4613.

12. Nystedt S, Emilsson K, Wahlestedt C, Sundelin J (1994) Molecular cloning ofa potential proteinase activated receptor. Proceedings of the National Academy

of Sciences of the United States of America 91: 9208–9212.

13. Dery O, Thoma MS, Wong H, Grady EF, Bunnett NW (1999) Trafficking ofproteinase-activated receptor-2 and beta-arrestin-1 tagged with green fluorescent

protein. beta-Arrestin-dependent endocytosis of a proteinase receptor. TheJournal of biological chemistry 274: 18524–18535.

14. Vu TK, Hung DT, Wheaton VI, Coughlin SR (1991) Molecular cloning of

a functional thrombin receptor reveals a novel proteolytic mechanism of

receptor activation. Cell 64: 1057–1068.

15. Greenberg DL, Mize GJ, Takayama TK (2003) Protease-activated receptor

mediated RhoA signaling and cytoskeletal reorganization in LNCaP cells.

Biochemistry 42: 702–709.

16. Kawabata A, Saifeddine M, Al-Ani B, Leblond L, Hollenberg MD (1999)

Evaluation of proteinase-activated receptor-1 (PAR1) agonists and antagonists

using a cultured cell receptor desensitization assay: activation of PAR2 by PAR1-

targeted ligands. The Journal of pharmacology and experimental therapeutics

288: 358–370.

17. Tanaka Y, Sekiguchi F, Hong H, Kawabata A (2008) PAR2 triggers IL-8 release

via MEK/ERK and PI3-kinase/Akt pathways in GI epithelial cells. Biochemical

and biophysical research communications 377: 622–626.

18. Yagi Y, Otani H, Ando S, Oshiro A, Kawai K, et al. (2006) Involvement of Rho

signaling in PAR2-mediated regulation of neutrophil adhesion to lung epithelial

cells. European journal of pharmacology 536: 19–27.

19. Dutra-Oliveira A, Monteiro RQ, Mariano-Oliveira A (2012) Protease-activated

receptor-2 (PAR2) mediates VEGF production through the ERK1/2 pathway in

human glioblastoma cell lines. Biochemical and biophysical research commu-

nications 421: 221–227.

20. Camerer E, Barker A, Duong DN, Ganesan R, Kataoka H, et al. (2010) Local

protease signaling contributes to neural tube closure in the mouse embryo.

Developmental cell 18: 25–38.

21. Brahimi-Horn MC, Pouyssegur J (2009) HIF at a glance. Journal of cell science

122: 1055–1057.

22. Leung K-W, Ng H-M, Tang MKS, Wong CCK, Wong RNS, et al. (2011)

Ginsenoside-Rg1 mediates a hypoxia-independent upregulation of hypoxia-

inducible factor-1a to promote angiogenesis. Angiogenesis 14: 515–522.

23. Jing Y, Liu L-Z, Jiang Y, Zhu Y, Guo NL, et al. (2012) Cadmium increases HIF-

1 and VEGF expression through ROS, ERK, and AKT signaling pathways and

induces malignant transformation of human bronchial epithelial cells. Toxico-

logical sciences: an official journal of the Society of Toxicology 125: 10–19.

24. Uusitalo-Jarvinen H, Kurokawa T, Mueller BM, Andrade-Gordon P, Fried-

lander M, et al. (2007) Role of protease activated receptor 1 and 2 signaling in

hypoxia-induced angiogenesis. Arteriosclerosis, thrombosis, and vascular biology

27: 1456–1462.

25. Svensson KJ, Kucharzewska P, Christianson HC, Skold S, Lofstedt T, et al.

(2011) Hypoxia triggers a proangiogenic pathway involving cancer cell

microvesicles and PAR-2-mediated heparin-binding EGF signaling in endothe-

lial cells. Proceedings of the National Academy of Sciences of the United States

of America 108: 13147–13152.



26. Pilgaard L, Lund P, Rasmussen JG, Fink T, Zachar V (2008) Comparative

analysis of highly defined proteases for the isolation of adipose tissue-derived

stem cells. Regenerative medicine 3: 705–715.

27. Fink T, Rasmussen JG, Lund P, Pilgaard L, Soballe K, et al. (2011) Isolation and

expansion of adipose-derived stem cells for tissue engineering. Frontiers in

bioscience (Elite edition) 3: 256–263.

28. Pilgaard L, Lund P, Duroux M, Fink T, Ulrich-Vinther M, et al. (2009) Effect of

oxygen concentration, culture format and donor variability on in vitro

chondrogenesis of human adipose tissue-derived stem cells. Regenerative

medicine 4: 539–548.

29. Fink T, Lund P, Pilgaard L, Rasmussen JG, Duroux M, et al. (2008) Instability

of standard PCR reference genes in adipose-derived stem cells during

propagation, differentiation and hypoxic exposure. BMC molecular biology 9:

98.

30. Fink T, Zachar V (2011) Adipogenic differentiation of human mesenchymal

stem cells. Methods in molecular biology (Clifton, NJ) 698: 243–251.

31. Dominici M, Le Blanc K, Mueller I, Slaper-Cortenbach I, Marini F, et al. (2006)

Minimal criteria for defining multipotent mesenchymal stromal cells. TheInternational Society for Cellular Therapy position statement. Cytotherapy 8:

315–317.

32. Zuk PA, Zhu M, Ashjian P, De Ugarte DA, Huang JI, et al. (2002) Humanadipose tissue is a source of multipotent stem cells. Molecular biology of the cell

13: 4279–4295.33. Zhong H, Chiles K, Feldser D, Laughner E, Hanrahan C, et al. (2000)

Modulation of hypoxia-inducible factor 1alpha expression by the epidermal

growth factor/phosphatidylinositol 3-kinase/PTEN/AKT/FRAP pathway inhuman prostate cancer cells: implications for tumor angiogenesis and

therapeutics. Cancer research 60: 1541–1545.34. Fukuda R, Hirota K, Fan F, Jung YD, Ellis LM, et al. (2002) Insulin-like growth

factor 1 induces hypoxia-inducible factor 1-mediated vascular endothelialgrowth factor expression, which is dependent on MAP kinase and phosphati-

dylinositol 3-kinase signaling in colon cancer cells. The Journal of biological

chemistry 277: 38205–38211.



Johannes Dick

Anti-opsonic properties of staphylokinase Quelle: Microbes and Infection 7 (2005) 476–484, Elsevier Betreuer: Prof. Barbara Bröker (Immunologie) Was bedeutet mir das Thema persönlich? Ich selbst beschäftige mich mit Staphylokokken in einem immunologischen Kontext. Immunevasions-mechanismen von S. aureus zu verstehen ist außerdem möglicherweise entscheidend, um wirksame Präventions- und Heilungsstrategien zu entwickeln. Worauf kommt es mir bei diesem Thema am meisten an? Es ist wichtig zu verstehen, dass S. aureus auf komplexe Art und Weise das Immunsystem des Menschen beeinflussen kann. Nur so wird klar, warum die Entwicklung eines Impfstoffs sich derart schwierig gestaltet. Außerdem bieten die Interaktionen zwischen Wirt und Erreger Erklärungsmöglichkeiten, warum S. aureus sowohl Kommensale als auch gefährlicher humanpathogener Erreger sein kann. Was fasziniert mich selbst am Thema am meisten? Das Projekt zeigt, wie erfolgreich S. aureus sich an den Menschen adaptiert hat. Körpereigene Proteine gegen das menschliche Immunsystem einzusetzen ist eine sehr elegante Vorgehensweise, die ein hohes Maß an Anpassung an einen ganz bestimmten Wirt erfordert. Was gefällt mir am Thema weniger? Die Relevanz der gefundenen Ergebnisse in vivo bleibt unklar. Da sich die Forschung explizit auf humane Enzyme bezieht, sind in vivo Studien kaum durchführbar, da kein Tiermodell verfügbar ist.

06/06/2013 179

Original article

Anti-opsonic properties of staphylokinase

S.H.M. Rooijakkers *, W.J.B. van Wamel, M. Ruyken, K.P.M. van Kessel, J.A.G. van Strijp

Eijkman Winkler Institute, University Medical Center (UMCU) G04-614, Heidelberglaan 100, 3584 CX Utrecht, The Netherlands

Received 7 May 2004; accepted 10 December 2004

Available online 26 February 2005

Abstract

Recently we described a novel bacteriophage-encoded pathogenicity island in Staphylococcus aureus that harbors a number of virulencefactors that are all involved in the evasion of innate immunity. Here we describe a mechanism by which staphylokinase (SAK), frequentlypresent on this pathogenicity island, interferes with innate immune defenses: SAK is anti-opsonic. By activating human plasminogen (PLG)into plasmin (PL) at the bacterial surface, it creates bacterium-bound serine protease activity that leads to degradation of two major opsonins:human immunoglobulin G (IgG) and human C3b. Incubation of opsonized bacteria with PLG and SAK resulted in removal of anti-staphylococcal IgGs and C3b from the bacterial surface. In phagocytosis assays this proved to be a very efficient mechanism to reduce theopsonic activity of human IgG and serum. The fact that SAK activates human PLG at the bacterial surface and removes IgG as well as C3bmakes this protein a unique anti-opsonic molecule.© 2005 Elsevier SAS. All rights reserved.

Keywords: Complement; Phagocytosis; Antibodies; Bacterial proteins

1. Introduction

Staphylococcus aureus is an important human pathogen.Infections due to this organism can range from minor woundinfections to severe sepsis [1]. The first line of human defenseagainst bacterial infections is the innate immune system. Uponentry of bacteria, complement is rapidly activated, leading tothe generation of, for instance, C5a. Local C5a and formy-lated peptides of bacterial origin recruit neutrophils from theblood vessels. Neutrophils then recognize bacteria, leadingto phagocytosis and killing. Phagocytosis only occurs whenbacteria are covered with bridging molecules (opsonins). Oneof the major opsonins is human immunoglobulin G (IgG),

consisting of a conserved carboxy-terminal region (Fc) andan amino-terminal end that contains the variable region defin-ing the antigen recognition site [2]. Via its Fc domain, IgG isbound by Fc receptors on the surface of phagocytes. The otherclass of opsonins is formed by C3 cleavage products. Theinteraction of C1q with bacterium-bound IgG on the one hand,and mannose binding lectins (MBL) or ficolins with bacterialpolysaccharides, on the other hand, initiates complement acti-vation via the classical and lectin pathway, respectively. Bothclassical and lectin pathway activation lead to cleavage ofC4 and C2, resulting in the generation of covalently boundC4b2a complexes. These complexes function as C3 conver-tase enzymes that produce opsonic C3b [3,4]. C3b covalentlybound to microbial surfaces is converted to C3bi as a resultof enzymatic cleavage by factor I in the a-chain [4]. Neutro-phils recognize C3b and C3bi via complement receptors 1 and3, resulting in effective phagocytosis.

Recently, a new chemotaxis inhibitory protein of S. aureus(CHIPS) was found in the supernatants of staphylococci [5,6].CHIPS was the first staphylococcal protein shown to beinvolved in the inhibition of neutrophil chemotaxis towardsthe site of infection by specifically blocking the binding ofbacterial formylated peptides and C5a to their receptors. The

Abbreviations: agr, accessory gene regulator; CHIPS, chemotaxis inhi-bitory protein of Staphylococcus aureus; Fc, conserved carboxy-terminalregion of IgG; GFP, green fluorescent protein; IgG, immunoglobulin G;MASP, MBL-associated serine protease; MBL, mannose binding lectins;MFI, mean fluorescence intensity; PA, plasminogen activator; PL, plasmin;PLG, plasminogen; SAK, staphylokinase; SaPI5, Staphylococcus aureuspathogenicity island 5; SEA, staphylococcal enterotoxin A.

* Corresponding author. Tel.: +31 30 250 6525; fax: +31 30 254 1770.E-mail address: [email protected] (S.H.M. Rooijakkers).

Microbes and Infection 7 (2005) 476–484

www.elsevier.com/locate/micinf

1286-4579/$ - see front matter © 2005 Elsevier SAS. All rights reserved.doi:10.1016/j.micinf.2004.12.014

06/06/2013 180

gene encoding CHIPS (chp) is located on a new S. aureuspathogenicity island 5 (SaPI5) (unpublished). SaPI5 also car-ries the genes for staphylokinase (sak) and staphylococcalenterotoxin A (sea). Staphylokinase (SAK) is an extra-cellular protein that is synthesized during the late exponen-tial growth phase and positively regulated by the accessorygene regulator (agr) [7]. SAK belongs to a group of bacterialplasminogen (PLG) activators (PA) that are not enzymesthemselves but form 1:1 complexes with human PLG andplasmin (PL). These complexes efficiently activate PLG intoits active form, PL [8]. PL is a potent serine protease that isknown for its capacity to degrade fibrin clots and extra-cellular matrix components [9]. S. aureus expresses severalPLG-binding molecules that immobilize PLG to the bacte-rial surface. Bacterium-bound PLG can be activated by SAK,resulting in generation of surface-associated PL [10]. Thisenzymatic activity is retained in the presence of the potentphysiological inhibitor, a2-anti-PL. SAK-mediated PLG-conversion at the staphylococcal surface has been implicatedwith a role in degradation of extra-cellular matrix proteinssurrounding the bacterium. Until now, little evidence of SAKas a staphylococcal virulence factor exists [11]. From the lit-erature, we know that PL is able to hydrolyze human IgG atthe hinge region of the molecule, cleaving the heavy chain ata peptide bond COOH-terminal to lysine 222 [2]. Further-more, while setting up a PLG-ELISA, Harpel et al. [12] acci-dentally found that PLG can also be bound to human IgG andthat PL cleaves immobilized IgG in a physiological environ-ment. This and experiments with PA-treated plasma resultingin cleavage of fluid-phase C3bi to C3c, C3d, g and C3dprompted us to investigate whether SAK-activated PLG alsodegrades opsonins in general [13]. Since the formation of PLby SAK occurs at the bacterial surface, this could create anideal mechanism, from a microbe’s point of view, to modu-late opsonization and thus increase bacterial survival.We dem-onstrate that SAK-activated PLG removes IgG as well as C3bfrom the bacterial surface, leading to impaired phagocytosis.

2. Materials and methods

2.1. Bacterial strains and plasmids

S. aureus Wood, S. aureus Cowan and S. aureus CowanEMS were obtained from the American Type Culture Collec-tion (ATCC, Rockville, MD). S. aureus Newman was a kindgift from Professor T.J. Foster, Moyne Institute, Trinity Col-lege, Dublin, Ireland. S. aureus RN6390 and RN4220 were akind gift from A.L. Cheung, Dartmouth Medical School,Hanover, NH. S. aureus CAPD399 was isolated from a con-tinuous ambulatory peritoneal dialysis (CAPD) patient at theUMC-Utrecht after informed consent. Escherichia coliDH10b was used as a host strain for construction of recom-binant plasmids. S. aureus was cultured overnight at 37 °Con blood agar (Beckton Dickinson, Sparks, MD, USA). Whenneeded for selection purposes, bacteria were grown at 37 °C

in Luria broth supplemented with agar (15 g/l) and 50 µg/mlcarbenicillin (Sigma, St. Louis, MO, USA) or 10 µg/mlchloramphenicol (Sigma). pTrcHISB vector was obtainedfrom Invitrogen, Paisley, UK. pACL1484 was also a kind giftfrom A.L. Cheung.

2.2. Production and purification of recombinantstaphylokinase (rSAK)

sak was cloned and expressed in E. coli according to themethods described for isolation of CHIPS in De Haas et al.[6]. PCR primers used were sak1: GTAAGTGCATCAAGT-TCATTC and sak2: GGAATTCTTATTTCTTTTCTATAACcontaining an EcoRI site (skewed sequence). HIS-tagged SAKwas purified from bacterial cell lysates by nickel (Invitrogen)affinity chromatography. The histidine tag was removed withenterokinase (Invitrogen), and rSAK was obtained after a sec-ond nickel column passage.

2.3. Distribution of anti-SAK Abs in human sera

Normal human serum was obtained after pooling the seraof 10 healthy laboratory workers. Anti-SAK antibodies weredetermined in normal serum, and serum samples wereobtained from 120 blood bank donors after informed con-sent. Microtiter plates (Greiner immunoplates, Fricken-hausen, Germany) were coated with 1 µg/ml rSAK inphosphate-buffered saline (PBS) and left overnight at 4 °C.Between all subsequent incubations wells were washed threetimes with PBS containing 0.05% Tween-20 (Sigma). Thewells were blocked with 4% bovine serum albumin (BSA,Sigma) in PBS with 0.05% Tween-20 and incubated withserial dilutions of the individual sera in PBS with 1% BSAfor 1 h on ice.Anti-SAK antibodies were detected with horse-radish peroxidase (HRP)-conjugated rabbit anti-human IgG(Southern, Birmingham, AL, USA) and visualized with0.1 mg/ml tetra-methyl-benzidine substrate (TMB, Sigma)and 0.15 mg/ml urea hydrogenperoxide (Sigma) in 0.1 Macetate buffer. The reaction was stopped with 1 N H2SO4 andoptical density (OD) was measured at 450 nm with a micro-plate reader (BioRad, Richmond, CA, USA).

2.4. SAK activity in staphylococcal supernatants

Presence of sak in S. aureus strains was determined byPCR. Overnight cultures of S. aureus Wood, S. aureus Cowan,S. aureus Cowan EMS and S. aureus Newman were inocu-lated in 3 ml Todd–Hewitt broth (Difco, Becton Dickinson)to achieve an initial OD660 of 0.08. Cultures were allowed togrow for 6 h, and bacteria were pelleted at different timepoints. Collected supernatants were frozen at –20 °C. To deter-mine the activity of excreted SAK in staphylococcal super-natants, supernatants were incubated with 60 µg/ml humanGlu-PLG (Calbiochem, EMD Biosciences, Darmstadt, Ger-many) and the PL-specific substrate H-Val-Leu-Lys-

477S.H.M. Rooijakkers et al. / Microbes and Infection 7 (2005) 476–484

06/06/2013 181

paranitroanilide (S-2251) (Bachem, Bubendorf, Switzer-land) (3.8 mg/ml, pH 7.8) at 37 °C for 1 h. Digestion of thesubstrate by PL resulted in the release of paranitroanalide thatwas measured at OD405. rSAK was used as a positive control.S. aureus Cowan EMS (sak negative) was used as a negativecontrol.

2.5. IgG and C3b deposition on S. aureus

To wells of 96-well vinyl assay plates (Costar,Acton, MA,USA), 25 × 106 S. aureus (Cowan EMS, staphylococcal pro-tein A and sak-negative strain) in PBS were added, incubatedfor 1 h at 37 °C and then left overnight at 4 °C. Between allsubsequent incubations wells were washed three times withPBS containing 0.05% Tween-20. The wells were blockedwith 4% BSA in PBS with 0.05% Tween-20 and incubatedwith serial dilutions of normal human serum in PBS contain-ing 1% BSA. To analyze surface-bound IgG, bacteria wereincubated in serum for 30 min on ice. For C3b deposition atthe bacterial surface, serum incubation was performed for20 min at 37 °C. Then, bacteria were incubated for 2 h at37 °C with PLG (60 µg/ml), rSAK (5 µg/ml) or PLG + rSAKin PBS-Tween 0.05% with 1% BSA. Bound Abs weredetected with HRP-conjugated rabbit anti-human IgG (Fcregion specific) (Pel-Freez, Rogers, AR, USA). Bound C3bwas detected by incubation with rabbit anti-human C3c (Nor-dic, Tilburg, The Netherlands) followed by HRP-conjugatedgoat anti-rabbit IgG (Southern). Thereafter, reactions werevisualized and measured as for ELISA quantifications above.To allow comparisons across multiple experiments for IgGbinding, OD450 values of the untreated group of the firstexperiment were assigned as the fixed values for untreatedbacteria.

To analyze C3b by flow cytometry, 1 × 108 S. aureus (SAK-producing versus non-producing) were incubated with 50%serum in PBS in reaction tubes for 20 min at 37 °C. Bacteriawere washed twice and incubated with PLG (60 µg/ml), rSAK(5 µg/ml) or PLG + rSAK in PBS for 1 h at 37 °C. Bacteriawere washed twice and incubated with FITC-labeled anti-C3c monoclonal antibody (WM1,ATCC, Rockville, MD [14])for 30 min on ice. After washing, surface-bound C3b wasmeasured by flow cytometry. Alternatively, opsonized bacte-ria were incubated with the PL-substrate (S-2251) for 1 h at37 °C to examine surface-associated PL activity.

2.6. C3b digestion

Human C3 was purified from human plasma according toLundwall and Eggertsen [15] and subsequently digested bytrypsin into C3a and C3b [16]. C3b (6 µg) was incubated withPLG (6 µg) and/or rSAK (0.5 µg) for 1.5 h at 37 °C in a totalvolume of 100 µl PBS. Incubation mixtures were analyzedon a 7.5% SDS-PAGE after 5 min incubation at 100 °C inreducing sample buffer. Proteins were transferred toImmobilon-P (Millipore, Bedford, MA, USA). Blots were

blocked in PBS-BSA 4% in 0.05% Tween, incubated for 1 hwith rabbit anti-human C3c followed by HRP-conjugated goatanti-rabbit IgG and enhanced chemiluminescence (Amer-sham).

2.7. Phagocytosis of S. aureus

S. aureus strain Cowan EMS was FITC labeled by incu-bating an exponential growth culture with 100 µg/ml FITC(Sigma) for 1 h at 37 °C in 0.1 M carbonate buffer, pH 9.6.Free FITC was removed by washing bacteria twice. Alterna-tively, we used heat-killed WVW189, an S. aureus strainwhich constantly expresses green fluorescent protein (GFP).WVW189 is RN6390 containing pACL1484 with a phagerepressor promoter cloned in its EcoRI/XbaI site, upstreamof GFPuvr. The phage repressor promoter was obtained byperforming a PCR on S. aureus CAPD399 DNA with theprimers rep1:GCTCTAGACCGTTTGATAACTTCATAATand rep2:CGGAATTCCTTGTTTTGAATCAAGTCA. Hu-man neutrophils were purified on a Ficoll (Amersham)/Histopaque (density 1.119; Sigma) gradient, as described pre-viously [17], using heparinized whole blood from a singledonor. Phagocytosis was performed for 15 min at 37 °C inpolystyrene round-bottom tubes (Falcon, Becton Dickinson)under vigorous shaking. Reactions were stopped by fixingthe samples in 100 µl of 1% paraformaldehyde. Samples wereevaluated by flow cytometry. Uptake of the FITC-labeled orGFP-producing S. aureus was evaluated by determining themean fluorescence intensity (MFI) of 10,000 neutrophils. Toallow comparisons across multiple experiments, MFI valuesof the untreated group of the first experiment were assignedas the fixed values for untreated bacteria.

In order to evaluate Fc receptor-mediated phagocytosis,human IgG was purified from normal human serum by pas-sage over a Protein G-Sepharose column (Amersham) accord-ing to the manufacturers’ instructions, after which IgG wasdialyzed against PBS. 15 × 106 staphylococci (GFP) wereincubated with different concentrations of human IgG inRPMI (RPMI-1640 25 mM HEPES with L-glutamine; Bio-whittaker, Cambrex, Verviers, Belgium) for 15 min at 37 °Cand washed. Bacteria were then incubated with 50 µl RPMI,60 µg/ml PLG with or without rSAK (5 µg/ml) for 1 h at37 °C. A total of 0.3 × 106 labeled neutrophils (PKH26 CellLinker (Sigma), according to the manufacturers’ instruc-tions) in 50 µl RPMI was added in a ratio of 50 bacteria to1 neutrophil.

For the evaluation of Ig-independent phagocytosis, humanserum was depleted of IgG and IgM by passage of 6 ml ofserum through a Protein G-Sepharose column (5 ml, Amer-sham) and subsequently through Protein L-Sepharose (5 ml,Sigma). Depletion of serum was checked by anti-IgG or IgMELISA. 4 × 106 FITC-labeled bacteria were incubated withdepleted serum for 20 min at 37 °C, washed once and incu-bated with RPMI, PLG with or without rSAK for 1 h at 37 °Cas described above. Human neutrophils were added in a ratioof 16 bacteria to 1 neutrophil.

478 S.H.M. Rooijakkers et al. / Microbes and Infection 7 (2005) 476–484

06/06/2013 182

Finally, staphylococci were pre-opsonized in 5% humanplasma for 20 min at 37 °C, washed once and incubated with10% heat-inactivated plasma in the presence of rSAK for 1 hat 37 °C. Phagocytosis was performed as described above.

3. Results

3.1. Expression and purification of rSAK

sak of S. aureus Newman was cloned and expressed inE. coli. Histidine-tagged SAK was purified by nickel affinitychromatography. The histidine tag was removed and rSAKwas obtained after a second nickel column passage. In a typi-cal run, we could isolate 4 mg from 1 l bacterial culture.

3.2. SAK is produced in vivo

To illustrate that SAK is produced in vivo, we measuredthe pre-existing anti-SAK antibodies in the human popula-tion. Fig. 1 shows that anti-SAK antibodies were detected in120 normal human sera, indicating that most people have beenpre-exposed to SAK. There is no homogenous distribution of

anti-SAK antibodies, one group of people has a high anti-body response against SAK, and a smaller group has a lowertiter against SAK.

3.3. SAK activity in staphylococcal supernatants

S. aureus Wood, S. aureus Cowan and S. aureus Newmanwere positive for sak by PCR, S. aureus Cowan EMS wasnegative. SAK activity in staphylococcal supernatants wasmeasured by incubating supernatants with PLG and a sub-strate specific for PL activity (S-2251). SAK productionstarted during post-exponential phase and was more pro-nounced during stationary growth phase, as could be expectedfrom an agr-regulated exotoxin. The amount of SAK in bac-terial supernatants was estimated by reference to dilutions ofrSAK. No PL activity was detected in supernatants of CowanEMS. In 3-ml growth cultures of the sak-positive strains,detected SAK levels ranged from 1-10 µg/ml (data not shown).We therefore used a concentration of 5 µg/ml SAK in thesubsequent experiments.

3.4. PLG + rSAK cleave human IgG bound to S. aureus

Harpel et al. [12] have illustrated by SDS-PAGE that incu-bation of human IgG with PL results in hydrolysis of IgG atthe hinge region. To study the protease activity ofPLG + rSAK towards IgG at the surface of bacteria, S. aureuswas coated onto immunoplates, and human IgG was boundby incubation with human sera on ice. Subsequent incuba-tion of bacteria with both PLG and rSAK resulted in a sig-nificant decrease in IgG Fc parts at the bacterial surface(P < 0.05 for 0.001–100% serum. Fig. 2). PLG and rSAKalone had no effect on S. aureus-bound IgG.

3.5. Fc receptor-mediated phagocytosis is inhibited byPLG + rSAK

Human IgG can be directly recognized by Fc receptors onhuman neutrophils [18]. Therefore, Fc receptor-mediatedphagocytosis was performed to study the effect of IgG cleav-age by PLG + rSAK at the bacterial surface. Fig. 3 shows thephagocytosis of S. aureus WVW189 when opsonized withpurified human IgG. At different concentrations of IgG,Antibody-dependent phagocytosis is significantly decreasedwhen opsonized bacteria were incubated with PLG + rSAK(P < 0.05), while PLG alone did not have this effect. Also,incubation with rSAK alone did not influence phagocytosis(data not shown).

3.6. Human C3b is cleaved by PLG + rSAK

Seya et al. [13] reported that treatment of human plasmawith urokinase (one of the major physiological PLG activa-tors) resulted in fluid-phase cleavage of human C3b and C3biinto its derivatives C3d and C3c. We studied this protease

Fig. 1. Distribution of anti-SAK antibodies in the healthy population.Anti-SAK titers were determined by testing sera of 120 healthy individualsby ELISA. Anti-SAK titer was defined as the logarithmic value of the dilu-tion factor that gave an OD450 of 0.3 after subtraction from background.Dots present the titer of each individual determined in one experiment. Nor-mal human serum represents the titer of a serum pool obtained from 10 heal-thy laboratory workers. All tested sera are shown to contain pre-existingantibodies against SAK, with a mean titer of 1.6. There is no homogeneousdistribution of anti-SAK antibodies, but antibody response is divided into alarge group of high responders and a smaller group of low responders.


06/06/2013 183

activity more thoroughly for a complex of PLG and rSAK byincubating purified C3b with PLG + rSAK in fluid-phase.Western blot analysis with anti-C3c Antibodies revealed thatincubation of C3b with a combination of PLG + rSAK leadsto digestion of the a-chain (115 kDa) and the b-chain (75 kDa)of C3b into two fragments of 60- and a 20-kDa product(Fig. 4).

3.7. PLG + rSAK cleave human C3b that is bound to S.aureus

Human C3b is a major opsonin recognized by CR1 recep-tors on neutrophils [19]. After formation of a C3 convertase,C3b is activated and covalently bound to the bacterial sur-face. To study whether PLG + rSAK could degrade bacterium-bound C3b, S. aureus was coated onto wells of 96-well vinylassay plates and incubated with human serum for 20 min at37 °C. Subsequent incubation with both PLG and rSAKresulted in a dramatic decrease of C3b molecules at the sur-face (P < 0.05 at 1–100% serum. Fig. 5A) compared to incu-bation with buffer or PLG or rSAK alone.

Analyzing C3b deposition on S. aureus using a higher celldensity and FITC-labeled anti-C3c, we could demonstrate thatendogenously produced SAK also leads to a decrease insurface-bound C3b molecules in the presence of PLG(Fig. 5B. P < 0.05). Fig. 5C indicates that C3b modulation isa result of PL activity at the bacterial surface.

Fig. 2. Human IgG binding to S. aureus.Bacteria were coated onto microtiter plates and IgG was bound by incuba-ting bacteria with various concentrations of serum. Washed bacteria weretreated with PBS (open squares), 60 µg/ml PLG (open circles), 5 µg/ml rSAK(open triangles) or PLG + rSAK (solid squares) at 37 °C for 2 h. The pre-sence of human IgG was determined by ELISA using Fc-specific anti-human IgG. Presence of IgG is expressed as the average OD450 of threeexperiments. PLG + rSAK incubation resulted in a significant decrease inbacterium-bound IgG compared to incubation with PLG or rSAK alone(P < 0.05 for 0.001–100% serum. Student’s t-test). Error bars indicate ±S.E.M.

Fig. 3. The influence of PLG + rSAK on Fc receptor-mediated phagocyto-sis.S. aureus (GFP) was incubated with 0–25 µg/ml purified human IgG. Bac-teria were then treated with PBS (open squares), 60 µg/ml PLG without (opencircles) or with rSAK (5 µg/ml) (solid squares) at 37 °C for 1 h. Bacteriawere added to freshly isolated human neutrophils, and phagocytosis wasallowed for 15 min at 37 °C. IgG-dependent phagocytosis is presented as theaverage MFI of three experiments using neutrophils of different blood donors.Incubation of IgG-opsonized bacteria with PLG + rSAK gave a significantdecrease in phagocytosis at all IgG concentrations (P < 0.05. Student’s t-test).Error bars indicate ± S.E.M.

Fig. 4. Detection of human C3b.Purified human C3b (6 µg) was incubated with PLG (6 µg) with or withoutrSAK (0.5 µg) for 1.5 h at 37 °C. Proteins were separated by SDS-PAGE ona 7.5% gel under reducing conditions and C3b was detected by immunoblot-ting with HRP-conjugated anti-human C3c. Lanes present (1) C3b; (2)C3b + PLG; (3) C3b+rSAK; (4) C3b + PLG + rSAK. Incubation of humanC3b with PLG + rSAK resulted in digestion of the a- and b-chain of C3binto 60- and 20-kDa C3 split products.


06/06/2013 184

3.8. PLG + rSAK prevents complement-drivenphagocytosis

Next to the classical route of complement activation initi-ated by human IgG or IgM, the complement system can beactivated by MBL or ficolins (lectin pathway) finally leadingto C3b and C3bi deposition at the bacterial surface. The above-mentioned experiments suggested that PLG and rSAKtogether can inhibit bacterial uptake by preventing recogni-tion via complement receptors 1 and 3. To investigate this,human serum was depleted of IgG and IgM and used toopsonize FITC-labeled S. aureus for 20 min. IgG and IgMwere depleted to abolish effects of PLG + rSAK on Fcreceptor-mediated phagocytosis (Fig. 3).We observed a stronginhibition of complement-driven phagocytosis by treatment

of opsonized bacteria with PLG + rSAK (P < 0.05 for 1–30%serum. Fig. 6).

3.9. Anti-phagocytic activity of rSAK is not regulated byplasmin inhibitors

The main PL inhibitor in circulation, a2-anti-PL, inhibitsformation of PL by PLG + SAK. Numerous studies haveshown that these inhibitory actions are completely abolishedwhen PLG or PLG-SAK complexes are bound to a target likefibrin or bacterial cells [10]. To study the effects of PL regu-lators on the anti-phagocytic properties of SAK, we incu-bated pre-opsonized bacteria with rSAK in the presence ofcomplement-inactivated plasma as a source of PLG and anti-PL. Fig. 7 shows that even in the presence of PL inhibitors,rSAK significantly reduced the phagocytosis (P < 0.05).

Fig. 5. C3b deposition on S. aureus.A. Bacteria were coated onto microtiter plates and incubated with normal human serum for 20 min at 37 °C. Washed bacteria were treated with PBS (opensquares), 60 µg/ml PLG (open circles), 5 µg/ml rSAK (open triangles) or PLG + rSAK (solid squares) at 37 °C for 2 h. C3b was detected by ELISA usinganti-human-C3c antibodies. At serum concentrations of 1–100% incubation of bacteria with PLG + rSAK led to a marked decrease in C3b molecules on thebacterial surface (P < 0.05. Student’s t-test). B–C. A total of 1 × 108 S. aureus (SAK-producing (SAK+) versus non-producing (SAK–)) were incubated in 50%serum for 20 min at 37 °C. Washed bacteria were treated with PLG (60 µg/ml), rSAK (5 µg/ml) or PLG + rSAK for 1 h at 37 °C. C3b deposition was measuredby flow cytometry using FITC-labeled anti-C3c monoclonal antibodies (B). Alternatively, opsonized bacteria were incubated with the PL substrate (S-2251) for1 h at 37 °C to examine surface-associated PL activity (C). C3b deposition is expressed as the average OD450 of three experiments. Error bars indicate ± SEM.


06/06/2013 185

4. Discussion

Our study demonstrates that PL, formed by the conversionof PLG by SAK, cleaves human IgG as well as human C3bfrom the bacterial surface, leading to impaired phagocytosisby human neutrophils. PL inhibitors could not regulate the

anti-phagocytic effects by PLG and SAK, probably due tothe surface location of the PL. Furthermore, analysis of humansera illustrates that SAK is expressed in vivo, since most indi-viduals have naturally occurring antibodies against SAK. Kill-ing and removal of invading pathogens is the primary task ofprofessional phagocytes [20]. They recognize invaders directlyor with the help of specific antibodies and/or the complementsystem [3]. For S. aureus, a wide spectrum of antibodies,directed against multiple epitopes, can activate the classicalpathway of complement or provide direct recognition via theirFc region [21]. MBL complexed with MBL-associated serineproteases (MASPs) have been shown to play an importantrole in activating the complement system in an antibody- andC1-independent way after binding complex microbial sugarson S. aureus [4]. Both the classical and lectin pathway ofcomplement activation result, via C4 and C2 cleavage prod-ucts, in the formation of C3b clusters at the surface of a patho-gen that are recognized by specific receptors on phagocytes[3]. From a bacterial point of view, modulation of opsonicmolecules like antibodies and complement-split productsseems an ideal strategy to increase bacterial survival.

Bacteria have developed several mechanisms to preventopsonization via antibodies or complement activation. Forexample, several streptococci express molecules that attracthost fluid-phase complement regulator factor H or C4b bind-ing protein to their surfaces and thereby suppress amplifica-tion of complement by the alternative or classical pathwayactivation [22]. Furthermore, several bacteria and viruses areknown to express molecules that bind IgG [23,24]. Forexample, staphylococcal protein A and streptococcal proteinG bind Fc parts of IgG and thereby prevent direct recognitionvia Fc receptors of phagocytes. Another well-known strategyis proteolytic cleavage of Abs, which was earlier described instreptococci and other bacteria [25,26]. SAK-mediated cleav-age of IgG also prevents Fc receptor-mediated phagocytosis.Moreover it seems likely that SAK-mediated cleavage of IgGalso leads to an impaired activation of the classical pathwayof complement, because earlier studies showed that PL cleavesIgG at position Lys 222. This results in removal of the entireFc fragment, including the glycosylation site (Asn 297) nec-essary for recognition by C1q [2]. Next to the ability to cleaveIgG, we observed that PL activity created by PLG and SAKled to digestion of C3b in fluid-phase and also to a decreasein C3b molecules at the staphylococcal surface. We showedthat phagocytosis in the absence of antibodies, predomi-nantly accomplished via recognition of C3b and C3bi, waseffectively inhibited by PLG + rSAK. PLG + rSAK cleavedC3b in both the a-chain as well as in the b-chain.

In flow cytometric experiments measuring C3b deposi-tion, we demonstrated that endogenously produced SAK alsomodulates opsonization. However, in phagocytosis andELISA experiments, we could not use this system due to tech-nical restrictions in buffer constitution and kinetics. Usingpurified rSAK and PLG, the exact specificity and biologicalimportance of the anti-opsonic properties of SAK could be

Fig. 6. Influence of PLG + rSAK on complement-driven phagocytosis.Complement-driven phagocytosis was performed by incubating FITC-labeled S. aureus with normal human serum that was depleted of IgG andIgM. Opsonization was performed for 20 min at 37 °C. Washed bacteriawere treated with PBS (open squares), 60 µg/ml PLG (open circles), 5 µg/mlrSAK (open triangles) or PLG + rSAK (solid squares) for 1 h at 37 °C. Bac-teria were incubated with human neutrophils for 15 min at 37 °C. Phagocy-tosis was quantified by flow cytometry and expressed as the mean fluores-cence of 10,000 neutrophils. Complement-driven phagocytosis is expressedas the average MFI obtained in three experiments using neutrophils of dif-ferent blood donors. (P < 0.05. Student’s t-test). Error bars indicate ± S.E.M.

Fig. 7. Influence of rSAK on phagocytosis in the presence of plasma.FITC-labeled S. aureus was opsonized in 5% human plasma for 20 min at37 °C. Washed bacteria were then incubated with buffer (RPMI) or 10 µg/mlrSAK in the presence of 10% complement-inactivated plasma for 1 h at 37 °C.Bacteria were incubated with human neutrophils for 15 min at 37 °C. Pha-gocytosis was quantified by flow cytometry and expressed as the mean fluo-rescence of 10,000 neutrophils. Phagocytosis is expressed as the averageMFI obtained in three experiments using neutrophils of different blooddonors. (P < 0.05. Student’s t-test). Error bars indicate ± S.E.M.


06/06/2013 186

demonstrated. Earlier studies have shown that a number ofbacteria and viruses can interact with human PLG, an impor-tant protein of our coagulation system that is present in serumand tissue fluids at high concentrations [11]. A Yersinia pes-tis surface protease cleaves PLG at the same site as humanPLG activators do and has been shown to play an importantrole in invasion of Y. pestis in vivo [27]. Other bacteria like S.aureus have been described to be able to bind and activatePLG at their surface, creating surface-bound PL activity [8].S. aureus expresses PLG receptors and excretes SAK, lead-ing to the acquisition of surface-localized PL activity that can-not be inhibited by physiological inhibitors [10]. Active PLat the microbial surface has been implicated with an increasedinvasive potential, but evidence is lacking [11]. Others recentlysuggested that SAK has a role in staphylococcal virulence bydirectly destroying defensins without interacting with PLG[28]. Whether the most important role for SAK in vivo shouldbe ascribed to its anti-opsonic properties or to its invasivecharacter remains to be elucidated.

We found SAK to be present on SaPI5 in S. aureus thatharbors a number of virulence factors that are all involved inthe evasion of innate immunity (unpublished). The gene forSAK, present in 85% of clinical strains, is embedded betweenthe genes for CHIPS and SEA. These two virulence factorsmodulate the function of different chemokine receptors andthereby prevent early migration of neutrophils and mono-cytes towards the site of an infection [6,29]. Data presentedin this article suggest an important role for SAK in modula-tion of the immune response, making SAK the third factorlocated on SaPI5 that modulates innate immunity. To provethat SAK is an important virulence factor, in vivo studies willhave to be performed. However, Gladysheva et al. [30]recently described that the ability of bacterial PLG activatorsto cleave PLG of different animals is restricted. SAK couldonly activate PLG of humans, dogs and baboons. It is inter-esting to see that other proteins encoded by SaPI5 are alsovery specific to humans. For example, CHIPS was shown tospecifically block formylated peptide receptors and C5a recep-tors on human neutrophils [6]. This human specificity seri-ously limits the possibilities for studying the role of SAK as avirulence factor in vivo.

The above observations and the fact that both IgG and C3bplay a key role in opsonization of bacteria led us to concludethat SAK is an important anti-opsonic molecule.

References

[1] K.B. Crossley, G.L. Archer, The staphylococci in human disease,Churchill Livingston, New York, NY, 1997.

[2] D. Stanworth, M. Turner, Immunochemical analysis of human andrabbit immunoglobulins and their subunits, in: Handbook of Experi-mental Immunology, Backwell Scientific Publications, London, 1986,pp. 121–1246.

[3] B. Beutler, Innate immunity: an overview, Mol. Immunol. 40 (12)(2004) 845–859.

[4] O. Neth, D.L. Jack, M. Johnson, N.J. Klein, M.W. Turner, Enhance-ment of complement activation and opsonophagocytosis by com-plexes of mannose-binding lectin with mannose-binding lectin-associated serine protease after binding to Staphylococcus aureus, J.Immunol. 169 (8) (2002) 4430–4436.

[5] K.E. Veldkamp, H.C. Heezius, J. Verhoef, J.A. Van Strijp, K.P. VanKessel, Modulation of neutrophil chemokine receptors by Staphylo-coccus aureus supernate, Infect. Immun. 68 (10) (2000) 5908–5913.

[6] C.J. De Haas, K.E. Veldkamp, A. Peschel, F. Weerkamp, W.J. VanWamel, E.C. Heezius, M.J. Poppelier, K.P. Van Kessel, J.A. VanStrijp, Chemotaxis inhibitory protein of Staphylococcus aureus, abacterial antiinflammatory agent, J. Exp. Med. 199 (5) (2004) 687–695.

[7] S. Arvidson, K. Tegmark, Regulation of virulence determinants inStaphylococcus aureus, Int. J. Med. Microbiol 291 (2) (2001) 159–170.

[8] M.A. Parry, X.C. Zhang, I. Bode, Molecular mechanisms of plasmi-nogen activation: bacterial cofactors provide clues, Trends Biochem.Sci. 25 (2) (2000) 53–59.

[9] R.W. Stephens, A. Vaheri, Plasminogen, Oxford University Press,New York, 1993.

[10] T. Molkanen, J. Tyynela, J. Helin, N. Kalkkinen, P. Kuusela,Enhanced activation of bound plasminogen on Staphylococcus aureusby staphylokinase, FEBS Lett. 517 (1–3) (2002) 72–78.

[11] K. Lahteenmaki, P. Kuusela, T.K. Korhonen, Bacterial plasminogenactivators and receptors, FEMS Microbiol. Rev. 25 (5) (2001) 531–552.

[12] P.C. Harpel, R. Sullivan, T.S. Chang, Binding and activation of plas-minogen on immobilized immunoglobulin G. Identification of theplasmin-derived Fab as the plasminogen-binding fragment, J. Biol.Chem. 264 (1989) 616–624.

[13] T. Seya, S. Nagasawa, M. Matsukura, H. Hasegawa, J.P. Atkinson,Generation of C3d, g and C3d by urokinase-treated plasma in associa-tion with fibrinolysis, Complement 2 (1985) 165–174.

[14] K.E. Veldkamp, K.P. Van Kessel, J. Verhoef, J.A. Van Strijp, Staphy-lococcal culture supernates stimulate human phagocytes, Inflamma-tion 21 (1997) 541–551.

[15] A. Lundwall, G. Eggertsen, Isolation of human complement factorsC3, C5 and H, J. Immunol. Methods 81 (1985) 147–160.

[16] T.A. Gaither, C.H. Hammer, M.M. Frank, Studies of the molecularmechanisms of C3b inactivation and a simplified assay of beta 1H andthe C3b inactivator (C3bINA), J. Immunol. 123 (1979) 1195–1204.

[17] A. Troelstra, B.N. Giepmans, K.P. Van Kessel, H.S. Lichenstein,J. Verhoef, J.A. Van Strijp, Dual effects of soluble CD14 on LPSpriming of neutrophils, J. Leukocyte Biol. 61 (1997) 173–178.

[18] T.W. Huizinga, F. Van Kemenade, L. Koenderman, K.M. Dolman,A.E. Dem Borne, P.A. Tetteroo, et al., The 40-kDa Fc gamma receptor(FcRII) on human neutrophils is essential for the IgG-induced respi-ratory burst and IgG-induced phagocytosis, J. Immunol. 142 (1989)2365–2369.

[19] H. Rosen, S.K. Law, The leukocyte cell surface receptor(s) for theIC3b product of complement, Curr. Top. Microbiol. Immunol. 153(1990) 99–122.

[20] P. Sansonetti, Phagocytosis of bacterial pathogens: implications in thehost response, Semin. Immunol. 13 (6) (2001) 381–390.

[21] G.J. Arlaud, C. Gaboriaud, N.M. Thielens, V. Rossi, Structural biol-ogy of C1, Biochem. Soc. Trans. 30 (Pt. 6) (2002) 1001–1006.

[22] H. Jarva, T.S. Jokiranta, R. Wurzner, S. Meri, Complement resistancemechanisms of streptococci, Mol. Immunol. 40 (2–4) (2003) 95–107.

[23] J.J. Langone, Protein A of Staphylococcus aureus and related immu-noglobulin receptors produced by streptococci and pneumonococci,Adv. Immunol. 32 (1982) 157–252.


06/06/2013 187

[24] T.L. Chapman, I. You, I.M. Joseph, P.J. Bjorkman, S.L. Morrison,M. Raghavan, Characterization of the interaction between the herpessimplex virus type I Fc receptor and immunoglobulin G, J. Biol.Chem. 274 (1999) 6911–6919.

[25] U. Pawel-Rammingen, L. Bjorck, IdeS and SpeB: immunoglobulin-degrading cysteine proteinases of Streptococcus pyogenes, Curr.Opin. Microbiol. 6 (1) (2003) 50–55.

[26] S.J. Kornfeld, A.G. Plaut, Secretory immunity and the bacterial IgAproteases, Rev. Infect. Dis. 3 (1981) 521–534.

[27] O.A. Sodeinde, Y.V. Subrahmanyam, K. Stark, T. Quan, Y. Bao,J.D. Goguen, A surface protease and the invasive character of plague,Science 258 (1992) 1004–1007.

[28] T. Jin, M. Bokarewa, T. Foster, J. Mitchell, J. Higgins, A. Tarkowski,Staphylococcus aureus resists human defensins by production ofstaphylokinase, a novel bacterial evasion mechanism, J. Immunol.172 (2) (2004) 1169–1176.

[29] R. Rahimpour, G. Mitchell, M.H. Khandaker, C. Kong, B. Singh,L. Xu, A. Ochi, R.D. Feldman, J.G. Pickering, B.M. Gill, D.J. Kelvin,Bacterial superantigens induce down-modulation of CC chemokineresponsiveness in human monocytes via an alternative chemokineligand-independent mechanism, J. Immunol. 162 (2) () (1999) 299–307.

[30] I.P. Gladysheva, R.B. Turner, I.Y. Sazonova, L. Liu, G.L. Reed,Coevolutionary patterns in plasminogen activation, Proc. Natl. Acad.Sci. USA 100 (16) (2003) 9168–9172.


06/06/2013 188

Documents

Journalclub – Programm · system (RAS) ﬁnally resulting in several angiotensin peptides, in particular angiotensin II (AngII), which acts via the AngII receptors type 1 and 2