mechanical stress enhances cd9 expression in … · mechanical stress enhances cd9 expression in...

Mechanical stress enhances CD9 expression in cultured podocytes

A. Blumenthal,1 J. Giebel,1 G. Warsow,1 L. Li,2 R. Ummanni,3 S. Schordan,1 E. Schordan,1 P. Klemm,1

N. Gretz,4 K. Endlich,1 and N. Endlich1

1Department of Anatomy and Cell Biology, Universitätsmedizin Greifswald, Greifswald, Germany; 2Department of Pathology,University Medical Center Göttingen, Göttingen, Germany; 3Center for Chemical Biology, CSIR-Indian Institute of ChemicalTechnology, Hyderabad, India; and 4Medical Faculty Mannheim, Medical Research Center, University of Heidelberg,Mannheim, Germany

Submitted 8 April 2014; accepted in final form 3 December 2014

Blumenthal A, Giebel J, Warsow G, Li L, Ummanni R, SchordanS, Schordan E, Klemm P, Gretz N, Endlich K, Endlich N. Mechanicalstress enhances CD9 expression in cultured podocytes. Am J PhysiolRenal Physiol 308: F602–F613, 2015. First published December 10,2014; doi:10.1152/ajprenal.00190.2014.—Elevated glomerular pressurerepresents a high risk for the development of severe kidney diseasesand causes an increase in mechanical load to podocytes. In this study,we investigated whether mechanical stress alters gene expression incultured podocytes using gene arrays. We found that tetraspanin CD9is significantly upregulated in cultured podocytes after mechanicalstress. The differential expression of CD9 was confirmed by RT-PCRand Western blotting under stretched and unstretched conditions.Furthermore, mechanical stress resulted in a relocalization of CD9. Toget an insight into the functional role of CD9, podocytes weretransfected with pEGFP-CD9. The expression of CD9 induced theformation of substratum-attached thin arborized protrusions. Ca2�

depletion revealed that podocytes overexpressing CD9 possess alteredadhesive properties in contrast to the control transfected cells. Finally,elevated CD9 expression increased migration of podocytes in a woundassay. In summary, our results suggest that upregulation of CD9 mayplay an important role in podocyte morphology, adhesion, and migra-tion.

podocytes; tetraspanins; mechnical stress; CD9

PODOCYTES ARE THE MAJOR TARGET in many glomerular diseaseslike glomerular hypertension, diabetic nephropathy, and focalsegmental glomerulosclerosis (FSGS). These diseases are hall-marked by proteinuria finally resulting in end-stage renaldisease (ESRD).

Podocytes cover the outer aspect of the glomerular capillar-ies by their foot processes interdigitating with their neighbor-ing cells. Beside their complex three-dimensional (3D) mor-phology, podocytes are characterized by a specific cell-cellcontact, the slit diaphragm. Since podocytes are postmitotic,lost podocytes cannot be replaced, which will finally result inareas with uncovered glomerular basement membrane (GBM).It is obvious that the attachment of podocytes to the GBM isessential for proper blood filtration. Glomerular diseases in-volve the alterations of podocyte morphology or attachment tothe GBM and are supposed to be aggravated by increasedmechanical load to podocytes (15).

The present study is based on gene array experiments thatwere done to identify and evaluate differentially regulatedgenes in cultured podocytes after mechanical challenge. Wefound that besides �3-integrin, the tetraspanin CD9 is signifi-

cantly upregulated in stretched podocytes. In fact, the attach-ment of podocytes to the GBM is predominantly mediated by�3�1-integrins. These transmembrane proteins are a main partof focal contacts that are linked to the actin cytoskeleton (26)that is exclusively expressed in the foot processes. Notably, apodocyte-specific ablation of �3�1 resulted in a severe renalphenotype (13, 23). In addition, another transmembrane pro-tein, the tetraspanin CD151, plays an important role in theattachment of podocytes to the GBM as CD151�/� micedeveloped, depending on the genetic background, proteinuriaand glomerular nephropathy (2, 33). Moreover, increasedblood pressure in CD151�/� mice induced by DOCA-salttreatment significantly aggravated the development of ne-phropathy (32). Furthermore, in humans a rare frame-shiftmutation in CD151 causes hereditary nephritis (21). Therefore,one might suggest that a reduction of mechanical stress topodocytes could ameliorate the progression of CD151 knock-out-induced nephropathy also in humans. In consideration ofthis background, the upregulation of CD9 rather than CD151 instretched podocytes was an interesting and unexpected finding.

CD9 belongs to the tetraspanin family of transmembraneproteins that contribute to the structural organization of theplasma membrane and are not only important for adhesion butinfluence also the morphology and motility of cells. Prominentmembers, beside CD151 (TSPAN 24) and CD9 (TSPAN 29),are CD37 (TSPAN 26), CD63 (TSPAN 30), and CD81(TSPAN 27). Unlike many other tetraspanins, CD9 associateswith CD151 and CD81 as well as �3�1- or �6�1-integrins inso-called tetraspanin-enriched microdomains (TEMs) in thecell membrane (4). Previously, in the human kidney CD9 wasidentified in the distal convoluted tubuli, collecting ducts, andin mesangial as well as endothelial cells of the glomerulus (27,37). Moreover, mouse podocytes express CD9 mRNA (7, 9).However, nothing is known about the localization and functionof CD9 either in unstretched or in stretched podocytes.

In the present study, we studied CD9, which is significantlyupregulated after 3-day exposure to mechanical stress withregard to localization and function in cultured podocytes.

MATERIALS AND METHODS

Cell culture. Conditionally immortalized podocytes (SVI; CLSCell Line Service, Eppelheim, Germany), L929 mouse skin fibro-blasts, and mouse mesangial cells were handled as previouslydescribed (6). For the isolation of kidneys, glomeruli, and primarypodocytes, mice (BL6) were anesthetized and euthanized accord-ing to the German animal protection law. Glomeruli and primarypodocytes were isolated by the sieving technique that was de-scribed previously (17).

Address for reprint requests and other correspondence: J. Giebel, Dept. ofAnatomy and Cell Biology, Universitätsmedizin Greifswald, Friedrich-Loef-fler-Str. 23c, 17487 Greifswald, Germany (e-mail: [email protected]).

Am J Physiol Renal Physiol 308: F602–F613, 2015.First published December 10, 2014; doi:10.1152/ajprenal.00190.2014.

1931-857X/15 Copyright © 2015 the American Physiological Society http://www.ajprenal.orgF602

by 10.220.33.3 on April 12, 2017

http://ajprenal.physiology.org/D

ownloaded from

mailto:[email protected]

http://ajprenal.physiology.org/

Immunocytochemistry. Cells were grown on glass coverslips coatedwith collagen IV for 48 h in RPMI at 37 and 38°C, respectively, andfixed in 2% paraformaldehyde (PFA) for 10 min at room temperature(RT) or in acetone/methanol (50/50) for 20 min at �20°C, respec-tively. PFA-fixed cells were permeabilized with 0.1% Triton X-100(in PBS) for 3 min. Cells were washed with PBS and blocked withblocking solution (PBS, 2% fetal bovine serum, 2% bovine serumfraction V, 0.2% fish gelatin) for 1 h at RT. Primary antibodies wereincubated for 90 min at RT. The following antibodies were used: ratanti-CD9, mouse anti-paxillin (both from BD Biosciences, San Jose,CA), mouse anti-�-actinin-4, and mouse anti-vinculin (both fromSigma-Aldrich, Steinheim, Germany). F-actin was visualized usingAlexa Fluor 488-phalloidin or Alexa Fluor 546-phalloidin (Invitrogen,Life Technologies) when fixed with PFA or by goat anti-�-actin (I-19)antibody (Santa Cruz Biotechnology, Santa Cruz, CA) when fixedwith methanol/acetone. After a washing step with PBS (3 � 5 min),cells were incubated with secondary antibodies for 60 min at RT.Antigen-antibody complexes were visualized with Cy2- or Cy3-conjugated secondary antibodies (all Dianova, Hamburg, Germany).Cells were washed (3 � 5 min in PBS) and embedded in mountingmedium (40 ml PBS, 10 g Mowiol, and 20 ml glycerol). Images weretaken using a Leica TCS SP5 confocal laser-scanning microscope(Leica Microsystems, Wetzlar, Germany).

The colocalization between CD9 and various cytoskeletal proteinswas quantified using the ImageJ “colocalization” plugin. Thresholdswere set by the “intermodes” method. Colocalized regions weremeasured in pixels and related to the total cell area.

Immunohistochemistry. Kidneys for immunohistochemical experi-ments were taken from total body-perfused mice. All procedures inanimals were performed in accordance with national animal protec-tion guidelines that conform with the National Institutes of HealthGuide for the Care and Use of Laboratory Animals and were ap-proved by the local governmental authorities. Mice were anesthetized(ip) with a mixture of Hostaket (1.2 ml/kg body weight) and Rompun(0.8 ml/kg body weight). After an abdominal section, a cannula wasinserted into the abdominal aorta (from retrograde), below the branchof the renal arteries. With a 3% paraformaldehyde-PBS fixative, themouse was directly perfused for 3 min (30 ml/min). Then, the kidneyswere removed, dehydrated, and embedded into paraffin by standardprocedures. With a Leica SM2000R microtom, 4-�m-thin sectionswere prepared. After rehydration, sections were unmasked incitrate buffer (0.1 M, pH 6.0) by heating for 5 min in a pressurecooker. The subsequent staining procedure was as described forimmunocytochemistry. The following primary antibodies wereused: mouse anti-human CD9 (DCS, Hamburg, Germany) andguinea pig anti-mouse nephrin (Progen, Heidelberg, Germany).Primary antibodies were detected by Alexa Fluor 488-conjugateddonkey anti-mouse or by Cy3-conjugated donkey anti-guinea pigsecondary antibody (Dianova), respectively. Nuclei were stainedwith Hoechst 33342 (Sigma-Aldrich).

Western blotting. Confluent cells were trypsinized, washed twice inPBS, and lysed (lysis buffer: 50 mM octylglucoside, 50 mM Tris, 150mM NaCl, 10 mM CaCl2, 1 mM MgCl2, 1 mM PMFS). Afterincubation (20 min, 4°C, under agitation), the lysate was centrifuged(10.000 g for 20 min at 4°C). The protein-containing supernatant(20–30 �g/lane) was separated using a 4–20% Mini-PROTEAN TGXGel (Bio-Rad Laboratories, München, Germany) and transferred to anitrocellulose membrane. Blocking step (5% low-fat milk, 60 min)was followed by incubation with a primary antibody (rat anti-CD9,BD Bioscience). Detection of bound antibodies was done with analkaline phosphatase-labeled secondary antibody (rabbit anti-rat AP,Acris, Hiddenhausen, Germany) and a phosphatase conjugate sub-strate kit (Bio-Rad Laboratories).

RT-PCR. RNA isolation was performed using TRI reagent (Sigma-Aldrich) according to the manufacturer’s protocol. Reverse transcrip-tion was done with 1 �g denatured and DnaseI-treated RNA (20 �lRT reaction), 2 �l oligo(dT)12–18 primer (Invitrogen Life Technolo-

gies), and 1 �l MuLV reverse transcriptase (50 U/�l; Applied Bio-systems Roche, Foster City, CA). Then, 2.5 �l cDNA (50 �l PCR)was analyzed in the presence of 1 �l Pwo-polymerase (1 U/�l; PeqlabBiotechnologie, Erlangen, Germany) and 20 pmol specific sense andantisense primers. The following primers were purchased from Invit-rogen Life Technologies: mouse CD9 (GenBank accession no.NC_000072.4), forward 5=-GCT ACT CGA GCC ATG CCG GTCAAA GGA GGT AGC-3=, reverse 5=-CAC TTG GTA CCG ACCATT TCT CGG CTC CTG CG-3=, 681-bp product size; and mouse�-actin (GenBank accession no. NM_007393), forward 5=-AGC CATGTA CGT AGC CAT CC-3=, reverse 5=-CTC TCA GCT GTG GTGGTG AA-3=, 228-bp product size. After denaturation for 3 min at95°C, 33 cycles at 95°C for 30 s, 60°C for 45 s, and 72°C for 2 minwere performed by an additional incubation at 72°C for 5 min.Controls were done by running PCR without reverse transcriptase orcDNA, respectively. PCR products were further analyzed by 1%agarose gel electrophoresis.

Mechanical stress. Mechanical stress was applied as reportedpreviously (17). Differentiated podocytes were grown (3 days) in aBioFlex six-well cell culture plate (Bioflex; Flexcell International,Hillsborough, NC) coated with collagen IV. The cell culture plate wasassembled on a manifold connected to the stretch apparatus (CLS,Cell Line Service). Podocytes were stretched (1–3 days) with a cyclefrequency of 0.5 Hz and a maximum linear strain of 5%. Three wellswere not exposed to mechanical stress and served as controls.

Gene array. For gene array analysis, podocytes were cultured asalready described. After 3-day mechanical stress, podocyte RNA wasisolated and hybridized with GeneChip Mouse Genome 430 2.0Arrays from Affymetrix (Santa Clara, CA). The gene expression datawere uploaded on the Gene Expression Omnibus (GEO) website(GSExxxx). The following data analysis steps were performed withJMP Genomics 4.0 software (Cary, NC). Mapping of array probeidentifiers to genes was done using custom CDF files for Entrez Geneidentifiers, v.17.0 from Brainarray.org. Background correction wasperformed following the robust multichip average (RMA) method; thedata were quantile normalized and log2 transformed. Calculation ofdifferential expression was done with the ANOVA method of JMPGenomics. The final gene expression data can be found in thesupplemental material (all supplemental material for this article isavailable on the journal website).

Real-time PCR. RNA isolation and reverse transcription wereconducted as described above (RT-PCR). cDNA was diluted 1:5 forreal-time PCR. Total PCR volume for each cDNA was 20 �l,containing 10 �l of Platinum Green qPCR SuperMix-UDG (Invitro-gen Life Technologies), 2 �l diluted cDNA, 2 �l mix of primers (200

Table 1. Changes in tetraspanin and integrin expression instretched and unstretched podocytes as revealed by genearray

DescriptionLSMeanStretch

LSMeanUnstretched Change, %

TetraspaninsCd9 CD9 antigen 10.33 9.73 52Cd37 CD37 antigen 4.47 4.55 �6Cd53 CD53 antigen 3.57 3.72 �11Cd81 CD81 antigen 11.26 11.27 �1Cd82 CD82 antigen 4.81 4.9 �6Cd151 CD151 antigen 9.65 9.42 17

IntegrinsItga1 Integrin-�1 7.75 7.91 �12Itga2 Integrin-�2 4.37 4.2 13Itga3 Integrin-�3 7.84 7.28 47Itga5 Integrin-�5 5.1 4.92 13Itga6 Integrin-�6 5.87 7.14 �59Itgb1 Integrin-�1 9.51 9.38 9Itgb4 Integrin-�4 4.11 4.01 7

F603CD9 IN PODOCYTES

AJP-Renal Physiol • doi:10.1152/ajprenal.00190.2014 • www.ajprenal.org

by 10.220.33.3 on April 12, 2017


ownloaded from


ng of each primer, sense and antisense), and 1 �l �20 BSA. Thefollowing primers were purchased from Invitrogen Life Technologies:mouse GAPDH (GenBank accession no. BC096440), forward 5=-ACC CAG AAG ACT GTG GAT GG-3=, reverse 5=-CAC ATT GGGGGT AGG AAC AC-3=, 171-bp product size; mouse B2m (GenBankaccession no. NM_009735.3), forward 5=-CGG TGA CCC TGG TCTTTC TG-3=, reverse 5=-GAT TTC AAT GTG AGG CGG GTG-3=,160-bp product size; and mouse CD9 (GenBank accession no.NC_000072.4), forward 5=-TGC AGT GCT TGC TAT TGG AC-3=,reverse 5=-GGC GAA TAT CAC CAA GAG GA-3=, 220-bp productsize. PCR was done in a LightCycler Nano (Roche Diagnostics,Indianapolis, IN), starting with a denaturation step, 10 min at 95°C,followed by 40 cycles at 95°C for 20 s, annealing at 60°C for 20 s, andelongation at 72°C for 20 s. Samples were normalized against thehousekeeping genes GAPDH and B2m for quantification. In additionto that, specificity of PCR products was analyzed by 1% agarose gelelectrophoresis.

Generation and transfection of plasmids. For generation of theCD9 expression plasmid, total RNA of differentiated SVI podocyteswas prepared using TRI reagent (Sigma-Aldrich) according to themanufacturer’s protocol. cDNA was synthesized by using MuLVreverse transcriptase (Applied Biosystems Roche). The coding se-

quence of CD9 was amplified by standard PCR using Pwo-polymer-ase (Peqlab Biotechnology, Erlangen, Germany) and specific primers[primers border start- and stop-codon and contain interfaces for KpnI(reverse primer) and XhoI (forward primer), CD9 for 5=-GCT ACTCGA GCC ATG CCG GTC AAA GGA GGT AGC-3= and CD9 rev5=-CAC TTG GTA CCG ACC ATT TCT CGG CTC CTG CG-3=].Primers were purchased from Invitrogen. The PCR product wascloned into the pEGFP-N3 expression vector [enhanced green fluo-rescent protein (EGFP); BD Biosciences]. To verify the recombinantsequences to be free of PCR errors, cloned products were controlledby sequence analysis (LGC Genomics, Berlin, Germany).

For visualization of filopodia, we used a pEGFP-fascin S39Aplasmid encoding the nonphosphorylable, not phosphorylated mutant(kindly provided by Dr. J. Adams, Cleveland Clinic Foundation,Cleveland, OH). For transfection, cells were seeded overnight on glasscoverslips coated with collagen IV. In one tissue culture dish (35 � 10mm), four glass coverslips were transfected simultaneously with 2.5�g of plasmid DNA using jetPEI transfection reagent (Polyplustransfection, Illkirch, France) according to the manufacturer’s proto-col. After 12–24 h, cells were used for live cell imaging.

Live imaging. For live imaging, cells were seeded on glass cover-slips that were attached to the bottom of a custom-built plexiglas

C DPP

CD

9

CP CP

E

β -ac

tin

RT-PCR

CD

9 +R

T

500 bp

500 bp

CP PP

A

F G

WB – CD9B

kidney glomeruli200 bp

200 bp

CD

9 -R

T

H2O β-ac

tin

CD

9 +R

T

CD

9 -R

T

H2O

Kid

ney

Glo

mer

uli

CP

37

25

kDa

Fig. 1. Detection of CD9 by RT-PCR, real-time PCR, Western blotting (WB), and im-munofluorescence. By RT-PCR, a 681 bpproduct for CD9 was detected in wholekidney extract, isolated glomeruli, culturedpodocytes (CP), as well as in primary podo-cytes (PP). �-Actin (228 bp) was amplified asa housekeeping gene. Control PCR withoutRT (CD9-RT) or H2O instead of cDNA(H2O) showed no signal (A). WB analysis ofCD9 in the kidney, isolated glomeruli, andCP reveals an expected band at 24 kDa (B).In confluent podocytes, CD9 is localized allover the cell membrane in a homogenouspattern. An accumulation of CD9 is seen atcell-cell contacts (C; fixation with 2% PFA).In nonconfluent podocytes, CD9 is distrib-uted all over the cell membrane. Moreover,thin arborized protrusions (TAPs) are posi-tive for CD9 (D; fixation with methanol/acetone). In PP, the localization of CD9 issimilar to that in CP (E; fixation with 2%PFA). Real-time PCR shows a significantincrease of CD9 in stretched CP comparedwith unstretched cells. Normalization of CD9to GAPDH (F) and B2m (G) indicates asignificant increase of 20 or 50%, respec-tively; n � 9. Scale bars � 25 �m. *P 0.05, ***P 0.001.

F604 CD9 IN PODOCYTES


by 10.220.33.3 on April 12, 2017


ownloaded from


chamber filled with modified RPMI-1640 medium (w/o phenol redsupplemented with 10% FBS). The chamber was sealed and fixed ona Leica DMI 6000B fluorescence microscope (Leica Microsystems).Chamber and objective were constantly warmed by an air streamincubator to 37°C. Images were taken with an integrated digitalcamera and processed with Openlab software (Improvision PerkinEl-mer, Waltham, MA).

Treatment of podocytes with cytochalasin D or EGTA. Podocyteswere washed with PBS and incubated with cytochalasin D (200 nM,45 min). Cells were fixed with acetone/methanol (50/50) for 20 min at�20°C. CD9 and F-actin were visualized by immunocytochemistry.Images were taken using a Leica TCS SP5 confocal laser-scanningmicroscope (Leica Microsystems). Transfected cells were incubatedwith EGTA (1 mM, 20 min) before RPMI-1640 medium (w/o phenolred supplemented with 10% FBS) was readded. Cells were analyzedby live cell imaging.

Migration assay. Podocytes were cultured as described before.With a forceps tip, wounds of 0.2 mm width were scratched into theconfluent cell layer of pEGFP-CD9-transfected cells. Cells wereincubated 16 h at 38°C and 5% CO2.

Statistical analysis. All data are given as means � SE analyzed byan unpaired t-test with repeated measurements (n). Differences weredetermined significant at a P value 0.05.

RESULTS

CD9 is upregulated by mechanical stress. Since culturedpodocytes were shown to be mechanosensitive (15, 17, 18, 35),we used this model to explore the genes that are regulated bymechanical stress. Therefore, we conducted gene arrays ofstretched (3 days) and unstretched podocytes. Interestingly, wefound that the tetraspanin CD9 was upregulated (52%) bymechanical stress whereas most of the tetraspanins remainedunaffected. Moreover, among integrins the �3-subunit wasupregulated by 47% whereas the �6-subunit was downregu-lated by 59% after 3 days of mechanical stress (Table 1).

CD9 is expressed in primary and cultured podocytes. SinceCD9 was significantly upregulated in stretched podocytes, westudied the endogenous expression of CD9 in whole kidney,isolated glomeruli, primary podocytes, as well as in culturedpodocytes. RT-PCRs with specific primers spanning five in-trons indicated bands for the predicted amplified region at 681bp in all samples tested (Fig. 1A). Primers for the housekeepinggene �-actin (228 bp) were used as a positive control. Further-more, Western blots displayed a band for CD9 at the expectedsize of 24 kDa in whole kidney protein, isolated glomeruli,and cultured podocytes (Fig. 1B). Immunofluorescence re-vealed that CD9 was homogenously distributed over the cellmembrane (Fig. 1, C and D) with an accumulation at cell-cellcontacts in cultured podocytes (Fig. 1, C and E). Moreover, insingle podocytes that were not in contact with neighboringcells, CD9 was also present in membrane processes (Fig. 1D).The distribution of CD9 in primary podocytes was very similarto that in cultured podocytes (Fig. 1E).

CD9 is localized in glomerular podocytes by immunofluorescence.To confirm the localization of CD9 in podocytes, we subjectedparaffin-embedded sections to CD9 and nephrin immunofluo-rescence. CD9 was localized to glomerular cells (Fig. 2).Double staining of CD9 and podocyte-specific nephrin re-vealed colocalization, indicating podocytes as CD9-expressingcells (Fig. 2, C–E). Moreover, CD9 was found in endothelialcells (Fig. 2E).

Mechanical stress changes mRNA expression and localiza-tion of CD9 in cultured podocytes. By real-time PCR, we founda significant increase in CD9 mRNA already 1 day aftermechanical stress. The expression of CD9 mRNA was in-creased by 20% when normalized against the housekeepinggene GAPDH and 50% when normalized against B2m as ahousekeeping gene, respectively (Fig. 1, F and G).

CD9 Nephrin merge

magnification * C magnification # C negative control

A B C

FD E

*

#Fig. 2. Detection of CD9 and nephrin inmouse kidney sections by immunofluores-cence. CD9 is localized in glomerular cells(A). Immunostaining for nephrin identifiespodocytes (B). The overlay shows that CD9is expressed in podocytes (C and magnifica-tions D and E, arrowheads). In contrast tonephrin, CD9 is also located in the endothe-lium of glomerular capillaries (E, arrow).Negative controls without primary antibodiesdepict a weak background staining (of mouseCy2 and guinea pig Cy3) in tubular cellswhile the glomerulus is negative (F). Scalebars � 25 �m (A–C) and 50 �m (F). Nucleiwere stained with Hoechst 33342.



by 10.220.33.3 on April 12, 2017


ownloaded from


Since CD9 is known to colocalize with CD151 and �3/�1-integrins that are associated with actin, we investigated whetherthere is a relationship to the actin cytoskeleton and actin-associ-ated proteins. In fact, CD9 colocalized with F-actin in a punctatepattern at cell-cell contacts in cultured podocytes (Fig. 3, A–C).

The ImageJ colocalization plugin displayed a colocalization areaof 1.03% (Fig. 3N). On the other hand, a lesser colocalizationscore was evident for the focal adhesion proteins vinculin (0.03%)and paxillin (0.27%) and the actin-associated protein �-actinin-4(0.42%), as depicted in Fig. 3, G–I and N.

K

unst

retc

hed BA

D

stre

tche

d

CD9 F-actin

E F

CD9/F-actin

C

CD9/vinculin CD9/paxillin CD9/α-actinin-4

unst

retc

hed

stre

tche

d

G H I

L M

N

Col

ocal

izat

ion

area

[%]

Fig. 3. Colocalization of CD9 with F-actin and actin-associated proteins in stretched and unstretched podocytes. In unstretched podocytes, CD9 is expressed allover the cell membrane and accumulates at cellular contacts (A). F-actin fibers (stained red with Alexa Fluor 546-phalloidin) were orientated in parallel andspanning through the cell. An accumulation of actin is seen at cellular contacts (B). The overlay reveals that CD9 colocalizes with F-actin predominantly at sitesof cell-cell contacts (C). In stretched podocytes, CD9 reorganizes into clusters while cell-cell contacts disappeared (D). F-actin is reorganized into actin-richcenters (ARCs) from which actin fibers radiate (E). CD9 accumulations are located within and around the ARCs (F). In unstretched podocytes, CD9 shows aminor colocalization with the focal adhesion proteins vinculin and paxillin and the actin-associated protein �-actinin-4 (red; G–I). In stretched podocytes, nocolocalization was found with paxillin and vinculin (K and L). In contrast, �-actinin-4 is partially colocalized with CD9 in cell protrusions (M). Additionally,graphs show the fluorescence intensity of both fluorophores at boxed regions. Quantitative colocalization studies (n � �5 cells) were done using ImageJ (N).The colocalization area in unstretched cells was 1.03% for CD9/F-actin, 0.42% for CD9/a-actinin-4, 0.27% for CD9/paxillin, and 0.03% for CD9/vinculin. Afterstretch, the colocalization area increased for CD9/F-actin up to 1.74% and for �-actinin-4 up to 1.42%. For paxillin, the colocalization area was reduced to 0.08%,and for vinculin (0.04%) no changes were observed. A significant change was confirmed for �-actinin-4 (P � 0.03; N). Scale bars � 25 �m (A–C and G–I) and12.5 �m (D–F and K–M).



by 10.220.33.3 on April 12, 2017


ownloaded from


Mechanical stress resulted in changes of cellular morphol-ogy and induced a complete reorganization of the actin cyto-skeleton in cultured podocytes according to our previous stud-ies (17). Instead of a transversal orientation in unstretchedpodocytes, actin fibers were orientated radially, converginginto actin-rich centers (Fig. 3, E and F). Interestingly, local-ization of CD9 was also affected by mechanical stress in thatCD9 was localized in cluster-like structures besides the distri-bution all over the cell membrane (Fig. 3D). Some of theseclusters were located beside or directly inside the actin-richcenters (Fig. 3F). Quantification studies revealed an increase inthe colocalized area (CD9 and F-actin) of up to 1.74% (Fig.3N). Vinculin was found in a radiate pattern around the CD9clusters (Fig. 3K), and paxillin was accumulated in spot-likeclusters at the margin of cultured podocytes (Fig. 3L). Whilecolocalization of vinculin and CD9 remained unchanged(0.04%) after mechanical stress, colocalization of paxillin andCD9 decreased to 0.08% (Fig. 3N). Moreover, �-actinin-4 waslocalized radial to the CD9 clusters as well as to some extentat the margin of the cell (Fig. 3M). The colocalization withCD9 bared a significant (P � 0.03) increase of up to 1.42%compared with unstretched conditions (Fig. 3N).

CD9 localization is influenced by the actin cytoskeleton. Tofind out whether localization of CD9 is linked to the actincytoskeleton, podocytes were treated with cytochalasin D. Intreated cells, the actin filaments disappeared and reorganizationinto spotlike structures was observed. In contrast, CD9 reor-ganized into more netlike structures and cell protrusions dis-appeared. Immunofluorescence of CD9 combined with AlexaFluor 546-phalloidin for staining of F-actin revealed a colocal-ization of CD9 with actin only in some parts of the cell, asshown in Fig. 4.

CD9 expression affects morphology of cultured podocytes.In transfected podocytes, EGFP-CD9 was localized all over thecell membrane, with the strongest expression at the cell-cellcontacts (Fig. 5). In addition, transfected cells developed thinarborized protrusions called TAPs (Fig. 5A) that were previ-

ously described in CD151-expressing podocytes (6). To eval-uate whether the development of TAPs is specific to podocytes,we transfected L929 fibroblasts and mouse mesangial cellswith pEGFP-CD9. As both cell types displayed membraneprotrusions, which were not extended and arborized, they weredifferent from TAPs. Furthermore, to exclude that TAPs rep-resent filopodia, we transfected podocytes with pEGFP-fascinthat is well known to induce filopodia (14). As expected,EGFP-fascin transfection led to the formation of filopodia.These protrusions, however, were clearly distinguished fromCD9-induced TAPs (Fig. 5E). Furthermore, the dynamic be-havior of filopodia was faster than that of TAPs (data notshown). Control podocytes transfected with an empty pEGFPvector did not generate protrusions (Fig. 5F).

Further characterization by live cell imaging of pEGFP-CD9-transfected podocytes showed that TAPs had a maximumdiameter of 300 nm. Cells were classified into three categorieson the basis of the number and branching of membraneprocesses as described previously (6). Briefly, category 0contained cells with 0–15 unbranched processes. Cells with15–50 fine, unbranched protrusions or 1–10 TAPs belonged tocategory I. Cells in category II showed �50 unbranchedprocesses or �10 TAPs. In a comparison of TAP formation ofpEGFP-CD9-transfected cells with control transfected cells, itturned out that 62% of the pEGFP-transfected but only 23% ofthe pEGFP-CD9-transfected cells belonged to category 0.Classification into category I occurred for 37% of control and64% of pEGFP-CD9-transfected cells. Accordingly, by only1% of pEGFP-transfected cells and 13% of pEGFP-CD9-transfected cells, category II revealed striking differences (Fig.6). For categorization, only cells with strong fluorescence werechosen. Thus TAP formation cannot be correlated to transfec-tion levels of podocytes.

By time-lapse microscopy, we found that single TAPs ofpEGFP-CD9-transfected podocytes exhibited different motili-ties. While some TAPs were stationary or moved slowly anddid not change their morphology, others were highly dynamic

A B CCD9 F-actin CD9/F-actin

D FE

Con

trol

Cyt

ocha

lasi

n D

Fig. 4. Colocalization of CD9 and F-actinbefore and after treatment with cytochalasinD. Control cells showed the homogenous dis-tribution of CD9 and characteristic mem-brane processes (A). F-actin is organized intostress fibers (B). Colocalization of CD9 withF-actin filaments is evident at the margin ofthe cell (C). After cytochalasin D treatment,CD9 was accumulated in a netlike structure(D), and F-actin was reorganized in asterlikestructures while cell protrusions disappeared(E). Colocalization of CD9 with F-actin wasobserved only to some extent (F, arrows).Scale bars � 50 �m.



by 10.220.33.3 on April 12, 2017


ownloaded from


in developing new processes that disappeared within a fewminutes (movie 1, supplemental material).

CD9 expression induces migration in podocytes. As CD9localization is dependent on the actin cytoskeleton, we exam-ined whether CD9 expression influences actin-dependent dy-namic processes. Therefore, we studied the migration dynam-ics of pEGFP-CD9-transfected podocytes by using a woundassay. We observed that transfected podocytes show a signif-icantly enhanced migration (P 0.05) compared with pEGFP-transfected control cells (Fig. 7).

Calcium-dependent dynamics of CD9-expressing podocytes.We further studied the dynamics of CD9-expressing podocytesafter depletion and readdition of calcium. After treatment withthe calcium chelator EGTA, we observed a contraction to someextent in control cells (expressing only EGFP) while somebroader cell protrusions were sustained, showing an accumu-lation of EGFP at their tips (Fig. 8A). On the other hand,EGTA-challenged EGFP-CD9-expressing cells rounded up toa minimum while a number of very thin CD9-positive mem-brane processes remained (Fig. 8B). After calcium reconstitu-

E

CP + pEGFP-CD9

BCP + pEGFP-CD9

A

L929 + pEGFP-CD9

C

CP + pEGFP-fascin

FCP + pEGFP

D

Mesangial cell + pEGFP-CD9

Fig. 5. Transfection experiments with pEGFP-CD9,pEGFP-fascin, and pEGFP (EGFP, enhanced green flu-orescent protein). In podocytes (CP), EGFP-CD9 islocalized all over the cell membrane (A and B), atcell-cell contacts (asterisk and insert in B), and in TAPs(star and insert in A). In L929 fibroblasts (C) and mousemesangial cells (D), EGFP-CD9 is distributed similarlybut does not induce formation of TAPs. Transfection ofCP with pEGFP-fascin involves the formation of filop-odia but not TAPs (E). Podocytes transfected with anempty pEGFP-vector do not show formation of TAPs(F). Scale bars: � 25 �m (A, B, D, and E) and 12.5 �m(C and F).



by 10.220.33.3 on April 12, 2017


ownloaded from


tion, control cells were not able to respread (Fig. 8A). Thusthey revealed only a minor change in morphology. pEGFP-CD9-transfected podocytes, however, started to spread alongthe fine CD9-positive processes (TAPs) until they had reachedtheir initial cell size and morphology, respectively (Fig. 8B).

DISCUSSION

Hypertension is one of the major risks for the developmentof severe renal diseases. Elevated glomerular pressure may beresponsible for an increase in mechanical load to podocytes(15). Mechanical stress involves alterations of podocyte mor-phology, e.g., reduction or loss of foot processes and finallydetachment from the glomerular basement membrane, leadingto end-stage renal failure. In previous studies, we have dem-onstrated that cultured podocytes are mechanosensitive and

respond to mechanical stress by reorganization of the actincytoskeleton, resulting in generation of elongated membraneprocesses (15–17, 35). In the present study, we have investi-gated the genes that are involved in the adaptation of culturedpodocytes to mechanical stress.

We found (by gene array as well as by real-time PCR) thetetraspanin CD9 to be significantly upregulated during me-chanical stress, suggesting an important function in stretchedpodocytes. Previously, it was described that CD9 is localized inthe kidney mainly in distal tubuli, collecting ducts, and inglomeruli (27, 37). Furthermore, it was shown by two inde-pendent groups that podocytes express CD9 mRNA (7, 9).However, there is only sparse information about the localiza-tion and function of CD9 in podocytes. In our cultured podo-cytes, CD9 was found at cellular contacts, which is in agree-

Cat 0 Cat I Cat IIpE

GFP

pEG

FPC

D9

62%

23%

37%

64%

1%

13%

Num

ber o

f cel

ls, %

Fig. 6. Categorization of pEGFP- and pEGFP-CD9-transfected podocytes on the basis of membrane processes. For better visualization of the membraneprocesses, fluorescence intensity was shown as “false colors.” Cells were classified into 3 categories. Sixty-two percent of control but only 23% ofpEGFP-CD9-transfected cells belong to 0. In category I, there were 37% of pEGFP-transfected but up to 64% of pEGFP-CD9-transfected podocytes. One percentof control and 13% of pEGFP-CD9-transfected podocytes were found in category II (n � �150 of 3 independent experiments). Scale bars � 25 �m. The resultsare summarized in the graph at the bottom of the figure. **P 0.01, ***P 0.001.



by 10.220.33.3 on April 12, 2017


ownloaded from


ment with the observation that CD9 is localized at cell-cellcontacts of colon carcinoma cells and of keratinocytes, respec-tively (28, 39).

CD9 belongs to the tetraspanin family of transmembraneproteins and is associated with other tetraspanins like CD81,CD63, and CD151 in a variety of cell types (5, 12, 38). CD9 isknown to play an essential role during sperm-egg fusion (19),cell migration, and tumor progression. Like other tetraspanins,

CD9 associates with integrins and different signaling mole-cules like TNF-� and HB-EGF (1, 4) as well as other tetras-panins. Recently, we have shown that CD9 colocalizes toCD151 in podocytes, indicating their close relationship (6).Since there seems to be an association of both tetraspanins inTEMs, there is increasing evidence that CD9 could modulateCD151 function (11, 12). In this context, it is interesting thatCD151 plays an important role in the development of kidney

0h pEGFP 0h pEGFP-CD9

12h pEGFP 12h pEGFP-CD9

BA

DC

E

% o

f wou

nd c

losu

reFig. 7. Wound assay of podocytes transfected withpEGFP-vector or pEGFP-CD9. EGFP-CD9-ex-pressing cells show a faster migration (12 h afterscratch) into the wound than podocytes transfectedwith the pEGFP-vector. Photographs were takenafter fixation with 2% PFA. The diagram repre-sents wound closure [%] of pEGFP- and pEGFP-CD9-transfected cells (B). Values are means � SE(n � �150 of 3 independent experiments). Scalebar � 50 �m. **P 0.01.



by 10.220.33.3 on April 12, 2017


ownloaded from


diseases. Thus loss of CD151 is related to a strong kidneyphenotype due to podocyte foot process effacement, resultingin proteinuria and focal glomerulosclerosis (FSGS) (2, 32, 33).

In the present study, we have further found that mechanicalstress induces the redistribution of CD9 as well as actin thatwas described previously (16, 17). As the treatment by cy-tochalasin D also involves a redistribution of CD9, this sug-gests a close interaction between CD9 and the actin cytoskel-eton. This is underlined by the observation that motility andgeneration of protrusions are influenced by CD9 expression.However, immunoprecipitation assays are needed to confirmthis hypothesis.

A number of studies demonstrate a correlation of high CD9levels and a low metastatic potential of cancer cells that isattributed to reduced migratory potential (25, 36, 40). Consis-tently, in breast cancer cells CD9 depletion resulted in anenhanced migratory phenotype due to an impaired localizationof talin to focal adhesions (29). In Chinese hamster ovary cells,however, CD9 enhances phosphorylation of the phosphoinosi-tol 3-kinase substrate AKT and thus improves migration (22).This is in accordance with our findings on EGFP-CD9-express-ing podocytes. Furthermore, it was shown that collagen IVinduced cell migration through CD9 in MDA-MB-231 breastcancer cells while there was no effect in nontumorigenicepithelial cells (MCF10A, MCF12A) (10). Moreover, CD9was noted to induce the migration of monocytes (34). Taken

together, these results indicate that the effect of CD9 seems tobe cell type dependent.

Most prominent, however, is the observation that pEGFP-CD9-transfected podocytes developed TAPs. TAP formationhas been already described after transfection with pEGFP-CD151 (6) and seems to be unique for podocytes since tetras-panins did not induce TAPs in other cell types like fibroblastsand mesangial cells. By time-lapse microscopy, we demon-strated static and highly dynamic TAPs. These TAPs persistedafter calcium depletion of podocytes and served as tracks forrespreading after calcium readdition. Interestingly, junctionsbetween keratinocytes changed to a zipper of filopodia underlow-calcium conditions. Moreover, these filopodia showed anaccumulation of the tetraspanins CD9, CD81, and CD151 (28).

So far, podocyte process formation was inducible by theRho-kinase inhibitor Y-27632 and forskolin (20). However,morphology and dynamics of these protrusions are differentfrom TAPs induced by CD9 or CD151, respectively. It isnoteworthy that tetraspanins are known to affect the morphol-ogy of cells. For example, the tetraspanin CD82 regulates themorphogenesis of microvilli by affecting the cortical actincytoskeleton organization and the curvature of plasma mem-brane, respectively (3). Furthermore, CD9 was found in mi-crovilli of eggs as well as in pseudopods of activated platelets(8, 31), and Kai/CD82-expressing cells offered elongated cel-lular tails and reduced lamellipodia (24). Moreover, a knockout

0‘ 5‘ 20‘ 35‘ 70‘ 120‘

0‘ 5‘ 20‘ 35‘ 70‘ 120‘A

B

Fig. 8. Dynamics of pEGFP- and pEGFP-CD9-transfected podocytes after removal and readdition of Ca2�. After treatment with EGTA, pEGFP-transfectedpodocytes showed a minimal contraction (A) while pEGFP-CD9-transfected podocytes were characterized by a strong contraction and remaining thin protrusions.Time-lapse microscopy reveals a slight spreading of pEGFP-transfected cells after readdition of calcium (A). The pEGFP-CD9-transfected cells, however,showed a strong contraction and remaining thin protrusions that were used as tracks for cell respreading (B). Pictures are presented in inverted colors. Widthof 1 picture represents 25 �m.



by 10.220.33.3 on April 12, 2017


ownloaded from


of tetraspanin CD81 led to impaired actin protrusions anddiminished lamellipodia spreading (30). This underlines therole of tetraspanins in generation of membrane protrusions asseen in our study.

In summary, our results suggest that the tetraspanin CD9,which is upregulated by mechanical stress, plays an essentialrole in the development of TAPs and seems to be crucial for themaintenance of podocyte morphology in an actin-dependentway.

ACKNOWLEDGMENTS

We thank Dr. J. Adams, Cleveland Clinic Foundation (Cleveland, OH) forkindly providing pEGFP-fascin S39A plasmids.

GRANTS

This study was supported by intramural funds from the Universitätsmedizinof the Ernst-Moritz-Arndt-Universität (Department of Cardiovascular Medi-cine, Forschungsverbund Molekulare Medizin), by grants from the EuropeanUnion (FP7-REGPOT-2010-1, EnVision, Grant 264143). This work is alsopart of the research project Greifswald Approach to Individualized Medicine(GANI_MED). The GANI_MED consortium is funded by the Federal Ministryof Education and Research and the Ministry of Cultural Affairs of the FederalState of Mecklenburg-West Pomerania (03IS2061A).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

Author contributions: A.B., J.G., K.E., and N.E. provided conception anddesign of research; A.B., L.L., R.U., S.S., and E.S. performed experiments;A.B., G.W., P.K., N.G., K.E., and N.E. analyzed data; A.B., K.E., and N.E.interpreted results of experiments; A.B. prepared figures; A.B., J.G., and N.E.drafted manuscript; A.B., J.G., K.E., and N.E. edited and revised manuscript;J.G., K.E., and N.E. approved final version of manuscript.

REFERENCES

1. Arduise C, Abache T, Li L, Billard M, Chabanon A, Ludwig A,Mauduit P, Boucheix C, Rubinstein E, Le Naour F. Tetraspaninsregulate ADAM10-mediated cleavage of TNF-alpha and epidermalgrowth factor. J Immunol 181: 7002–7013, 2008.

2. Baleato RM, Guthrie PL, Gubler MC, Ashman LK, Roselli S. Deletionof CD151 results in a strain-dependent glomerular disease due to severealterations of the glomerular basement membrane. Am J Pathol 173:927–937, 2008.

3. Bari R, Guo Q, Xia B, Zhang YH, Giesert EE, Levy S, Zheng JJ,Zhang XA. Tetraspanins regulate the protrusive activities of cell mem-brane. Biochem Biophys Res Commun 415: 619–626, 2011.

4. Berditchevski F. Complexes of tetraspanins with integrins: more thanmeets the eye. J Cell Sci 114: 4143–4151, 2001.

5. Berditchevski F, Gilbert E, Griffiths MR, Fitter S, Ashman L, JennerSJ. Analysis of the CD151-alpha3beta1 integrin and CD151-tetraspanininteractions by mutagenesis. J Biol Chem 276: 41165–41174, 2001.

6. Blumenthal A, Giebel J, Ummanni R, Schluter R, Endlich K, EndlichN. Morphology and migration of podocytes are affected by CD151 levels.Am J Physiol Renal Physiol 302: F1265–F1277, 2012.

7. Boerries M, Grahammer F, Eiselein S, Buck M, Meyer C, Goedel M,Bechtel W, Zschiedrich S, Pfeifer D, Laloe D, Arrondel C, GoncalvesS, Kruger M, Harvey SJ, Busch H, Dengjel J, Huber TB. Molecularfingerprinting of the podocyte reveals novel gene and protein regulatorynetworks. Kidney Int 83: 1052–1064, 2013.

8. Brisson C, Azorsa DO, Jennings LK, Moog S, Cazenave JP, Lanza F.Co-localization of CD9 and GPIIb-IIIa (alpha IIb beta 3 integrin) onactivated platelet pseudopods and alpha-granule membranes. Histochem J29: 153–165, 1997.

9. Brunskill EW, Georgas K, Rumballe B, Little MH, Potter SS. Definingthe molecular character of the developing and adult kidney podocyte.PLoS One 6: e24640, 2011.

10. Castro-Sanchez L, Soto-Guzman A, Navarro-Tito N, Martinez-Orozco R, Salazar EP. Native type IV collagen induces cell migration

through a CD9 and DDR1-dependent pathway in MDA-MB-231 breastcancer cells. Eur J Cell Biol 89: 843–852, 2010.

11. Charrin S, Manie S, Billard M, Ashman L, Gerlier D, Boucheix C,Rubinstein E. Multiple levels of interactions within the tetraspanin web.Biochem Biophys Res Commun 304: 107–112, 2003.

12. Charrin S, Manie S, Oualid M, Billard M, Boucheix C, Rubinstein E.Differential stability of tetraspanin/tetraspanin interactions: role of palmi-toylation. FEBS Lett 516: 139–144, 2002.

13. Chen CA, Hwang JC, Guh JY, Chang JM, Lai YH, Chen HC. Reducedpodocyte expression of alpha3beta1 integrins and podocyte depletion inpatients with primary focal segmental glomerulosclerosis and chronicPAN-treated rats. J Lab Clin Med 147: 74–82, 2006.

14. Cohan CS, Welnhofer EA, Zhao L, Matsumura F, Yamashiro S. Roleof the actin bundling protein fascin in growth cone morphogenesis:localization in filopodia and lamellipodia. Cell Motil Cytoskeleton 48:109–120, 2001.

15. Endlich N, Endlich K. The challenge and response of podocytes toglomerular hypertension. Semin Nephrol 32: 327–341, 2012.

16. Endlich N, Endlich K. Stretch, tension and adhesion—adaptive mecha-nisms of the actin cytoskeleton in podocytes. Eur J Cell Biol 85: 229–234,2006.

17. Endlich N, Kress KR, Reiser J, Uttenweiler D, Kriz W, Mundel P,Endlich K. Podocytes respond to mechanical stress in vitro. J Am SocNephrol 12: 413–422, 2001.

18. Endlich N, Sunohara M, Nietfeld W, Wolski EW, Schiwek D, Kranzlin B,Gretz N, Kriz W, Eickhoff H, Endlich K. Analysis of differential geneexpression in stretched podocytes: osteopontin enhances adaptation ofpodocytes to mechanical stress. FASEB J 16: 1850–1852, 2002.

19. Evans JP. Sperm-egg interaction. Annu Rev Physiol 74: 477–502, 2012.20. Ichimura K, Kurihara H, Sakai T. Actin filament organization of foot

processes in rat podocytes. J Histochem Cytochem 51: 1589–1600, 2003.21. Karamatic Crew V, Burton N, Kagan A, Green CA, Levene C, Flinter

F, Brady RL, Daniels G, Anstee DJ. CD151, the first member of thetetraspanin (TM4) superfamily detected on erythrocytes, is essential forthe correct assembly of human basement membranes in kidney and skin.Blood 104: 2217–2223, 2004.

22. Kotha J, Longhurst C, Appling W, Jennings LK. Tetraspanin CD9regulates beta 1 integrin activation and enhances cell motility to fibronec-tin via a PI-3 kinase-dependent pathway. Exp Cell Res 314: 1811–1822,2008.

23. Kreidberg JA, Donovan MJ, Goldstein SL, Rennke H, Shepherd K,Jones RC, Jaenisch R. Alpha 3 beta 1 integrin has a crucial role in kidneyand lung organogenesis. Development 122: 3537–3547, 1996.

24. Liu WM, Zhang F, Moshiach S, Zhou B, Huang C, Srinivasan K,Khurana S, Zheng Y, Lahti JM, Zhang XA. Tetraspanin CD82 inhibitsprotrusion and retraction in cell movement by attenuating the plasmamembrane-dependent actin organization. PLoS One 7: e51797, 2012.

25. Miyake M, Nakano K, Itoi SI, Koh T, Taki T. Motility-related protein-1(MRP-1/CD9) reduction as a factor of poor prognosis in breast cancer.Cancer Res 56: 1244–1249, 1996.

26. Mundel P, Shankland SJ. Podocyte biology and response to injury. J AmSoc Nephrol 13: 3005–3015, 2002.

27. Nakamura Y, Handa K, Iwamoto R, Tsukamoto T, Takahasi M,Mekada E. Immunohistochemical distribution of CD9, heparin bindingepidermal growth factor-like growth factor, and integrin alpha3beta1 innormal human tissues. J Histochem Cytochem 49: 439–444, 2001.

28. Penas PF, Garcia-Diez A, Sanchez-Madrid F, Yanez-Mo M. Tetraspa-nins are localized at motility-related structures and involved in normalhuman keratinocyte wound healing migration. J Invest Dermatol 114:1126–1135, 2000.

29. Powner D, Kopp PM, Monkley SJ, Critchley DR, Berditchevski F.Tetraspanin CD9 in cell migration. Biochem Soc Trans 39: 563–567, 2011.

30. Quast T, Eppler F, Semmling V, Schild C, Homsi Y, Levy S, Lang T,Kurts C, Kolanus W. CD81 is essential for the formation of membraneprotrusions and regulates Rac1-activation in adhesion-dependent immunecell migration. Blood 118: 1818–1827, 2011.

31. Runge KE, Evans JE, He ZY, Gupta S, McDonald KL, Stahlberg H,Primakoff P, Myles DG. Oocyte CD9 is enriched on the microvillarmembrane and required for normal microvillar shape and distribution. DevBiol 304: 317–325, 2007.

32. Sachs N, Claessen N, Aten J, Kreft M, Teske GJ, Koeman A, ZuurbierCJ, Janssen H, Sonnenberg A. Blood pressure influences end-stage renaldisease of Cd151 knockout mice. J Clin Invest 122: 348–358, 2012.



by 10.220.33.3 on April 12, 2017


ownloaded from


33. Sachs N, Kreft M, van den Bergh Weerman MA, Beynon AJ, PetersTA, Weening JJ, Sonnenberg A. Kidney failure in mice lacking thetetraspanin CD151. J Cell Biol 175: 33–39, 2006.

34. Schenk GJ, Dijkstra S, van het Hof AJ, van der Pol SM, Drexhage JA,van der Valk P, Reijerkerk A, van Horssen J, de Vries HE. Roles forHB-EGF and CD9 in multiple sclerosis. Glia 61: 1890–1905, 2013.

35. Schordan S, Schordan E, Endlich K, Endlich N. �V-Integrins mediatethe mechanoprotective action of osteopontin in podocytes. Am J PhysiolRenal Physiol 300: F119–F132, 2011.

36. Setoguchi T, Kikuchi H, Yamamoto M, Baba M, Ohta M, Kamiya K,Tanaka T, Baba S, Goto-Inoue N, Setou M, Sasaki T, Mori H,Sugimura H, Konno H. Microarray analysis identifies versican and CD9as potent prognostic markers in gastric gastrointestinal stromal tumors.Cancer Sci 102: 883–889, 2011.

37. Sincock PM, Mayrhofer G, Ashman LK. Localization of the transmem-brane 4 superfamily (TM4SF) member PETA-3 (CD151) in normal humantissues: comparison with CD9, CD63, and {alpha}5ss1 integrin. J His-tochem Cytochem 45: 515–526, 1997.

38. Terada N, Baracskay K, Kinter M, Melrose S, Brophy PJ, BoucheixC, Bjartmar C, Kidd G, Trapp BD. The tetraspanin protein, CD9, isexpressed by progenitor cells committed to oligodendrogenesis and islinked to beta1 integrin, CD81, and Tspan-2. Glia 40: 350–359, 2002.

39. Yanez-Mo M, Tejedor R, Rousselle P, Sanchez -Madrid F. Tetraspa-nins in intercellular adhesion of polarized epithelial cells: spatial andfunctional relationship to integrins and cadherins. J Cell Sci 114: 577–587,2001.

40. Zoller M. Tetraspanins: push and pull in suppressing and promotingmetastasis. Nat Rev Cancer 9: 40–55, 2009.



by 10.220.33.3 on April 12, 2017


ownloaded from


mechanical stress enhances cd9 expression in … · mechanical stress enhances cd9 expression in...

Documents