Extracellular matrix and nuclear localization of βig‐h3 in human bladder smooth muscle and fibroblast cells

ExtracellularMatrixandNuclearLocalizationofbig-h3inHumanBladderSmoothMuscleandFibroblastCellsPaulC.Billings,*DavidJ.Herrick,UmbertoKucich,BeatriceN.Engelsberg,WilliamR.Abrams,EdwardJ.Macarak,JoelRosenbloom,andPamelaS.HowardDepartmentofAnatomyandHistology,UniversityofPennsylvania,SchoolofDentalMedicine,Philadelphia,Pennsylvania19104Abstract Theextracellularmatrix(ECM)playsanessentialroleinbladderstructureandfunction.Inthisstudy,expressionofbig-h3,arecentlyidentifiedextracellularmatrixprotein,wasinvestigatedinhumanbladdertissue,andhumanbladdersmooth-muscle(SMC)andfibroblastcellsinvitro.SMCssecretedgreaterthanthreetimesthelevelofthisproteincomparedwithfibroblasts.Therelativelevelsofbig-h3mRNAinthetwocelltypesreflectedtheproteinexpression.ImmunohistochemicalanalysisdemonstratedproteindepositionintheECMaswellascytoplasmiclocalizationand,unexpectedly,nuclei.Anti-big-h3antibodiesalsostainedthematrixsurroundingthedetrusorSMCsandnucleiofbladderfibroblasts,SMCs,andurotheliuminintactbladdertissue.Westernblotanalysesofmediumandmatrixfractionsobtainedfromcellsinvitrorevealedproteinof;7074kDa,whereasnuclearextractscontaineda65-kDareactiveproteinband.WeproposethatalthoughthisproteinisastructuralcomponentofbladderECM,itsnuclearlocalizationsuggeststhatithasotherregulatoryand/orstructuralfunctions.J.Cell.Biochem.79:261273,2000. 2000Wiley-Liss,Inc.Keywords:extracellularmatrix;big-h3;bladder;localizationThebladderisadynamicorganwhosefunc-tionistostoreandperiodicallyeliminateurinefromthebody.Thefillingphaseofthemicturi-tioncycleischaracterizedbyagradualin-creaseinvolumeofurinewithlittleincreaseinpressure.Oncefull,thebladderwallexperi-encesasharpincreaseinpressure,resultingincontractionofthedetrusorSMCthatinitiatebladderemptyingShapiroandLepor,1995.Properbladderfunctionrequirespreciseinter-actionsbetweentheECMandresidentstromalandmusclecells.TheECMconsistsofawidearrayofsecretedproteins,includingcollagens,laminins,fi-bronectin,elastin,andproteoglycansreviewedinAumailleyandGayraud,1998.TheECMcomponentsinteractwitheachother,formingahighlyorganizednetworkthatmaintainstis-suearchitectureandprovidesasupportstruc-tureforresidentcellularcomponents.TypesIandIIIcollagencomprisethemajorcollagensfoundinthebladderwall,theirprimarysourcesbeingSMCfromthedetrusorlayerandfibroblastsinthelaminapropriaEwaltetal.,1992;Baskinetal.,1993;Coplenetal.,1994;Deveaudetal.,1998.TypeIVcollagenisfoundasaninvestment,surroundingindividualSMCwithinthemuscularismucosaanddetrusorlayerDeveaudetal.,1998andamajorcom-ponentoftheurothelialbasementmembraneWilsonetal.,1996.Elastinandtheelastin-associatedmicrofibrillarproteins,fibrillin-1andMAGP,arealsoassociatedwithbothcom-partmentsofthebladderwall,andarethoughttocontributeflexibilityandelasticityKooetal.,1998.Theproperarrangementofthesematrixproteinswithoneanotherandwiththecellularcomponentsofthebladderwallallcon-tributetonormalorganfunction:gradualex-pansionofthebladderwallonfillingandrapidcontractionandrecoilduringmicturitionMacarakandHoward,1997;Changetal.,1998.QuantitativechangesinbladderECMAbbreviationsused:ECM,extracellularmatrix;PBS,phosphate-bufferedsaline;PBST,PBScontaining0.1%Tween20;SMC,smooth-musclecells;TGF-b1,transform-inggrowthfactor-b1.Grantsponsor:NationalInstitutesofHealth;Grantnum-bers:DK52220,DK48215,DK45419,andDK54987.*Correspondenceto:PaulBillings,DepartmentofAnat-omyandHistology,SchoolofDentalMedicine,UniversityofPennsylvania,4010LocustStreet,Philadelphia,PA19104.E-mail:Received25January,2000;Accepted7April2000JournalofCellularBiochemistry79:261273(2000)2000Wiley-Liss,Iposition as well as alterations in the ar-rangement of matrix proteins can contribute toloss of bladder function and ultimately maylead to renal impairment Lin and McConnell,1995; Shapiro and Lepor, 1995; Macarak andHoward, 1997. Although significant progresshas been made in identifying and characteriz-ing the major components of the bladder ECM,comparatively little is known about the “bridg-ing molecules” that interconnect specific com-ponents of the ECM with each other and resi-dent cells. These proteins may serve animportant function by transferring tensionfrom cells to other matrix components.The ECM protein, big-h3, was originally iso-lated from fetal bovine nuchal ligament by re-ductive saline extraction (designated MP78Gibson et al., 1989) and was hypothesized tobe a microfibrillar protein. More recently,big-h3 was cloned from TGFb1-stimulatedA549 cells Skonier et al., 1992. The nascentprotein contains a secretory signal sequence(residues 123), four homologous internal do-mains, and a cell attachment (RGD) siteSkonier et al., 1992, 1994. The porcine homo-logue has been purified from a fiber-rich frac-tion of pig cartilage and designated RGD-CAPHashimoto et al., 1997. The porcine-derivedprotein binds collagens I, II, and IV. In addi-tion to being closely related to human big-h3 atthe amino acid level, RGD-CAP shows 45%identity with rat osteoblast-specific factor-2and 22% identity with Drosophila fasciclin IHashimoto et al., 1997. Although the physio-logical function of big-h3 is not known, it hasbeen suggested that it interconnects differentmatrix components with each other and resi-dent cells Gibson et al., 1989, 1997, and con-sequently, may serve as a bifunctional linkerprotein.In bovine tissue, big-h3 was found associatedwith collagen fibers in developing nuchal liga-ment, aorta, lung, and mature cornea Gibsonet al., 1997. Immunoreactive material wasalso present in capsule and tubule basementmembranes of developing kidney and reticularfibers in fetal spleen. The staining pattern wassimilar to that observed for type VI collagenGibson et al., 1997. These results indicatethat the protein is a widely expressed compo-nent of the ECM in several organ systems andmay be closely associated with other matrixmacromolecules.In the course of studies directed at elucidat-ing big-h3 function, we observed that cells ob-tained from human bladder tissue express rel-atively high amounts of this protein. In thisreport, we have investigated big-h3 productionby human bladder fibroblast and SMC and itsdeposition in tissue.MATERIALS AND METHODSChemicals and ReagentsBiotinylated goat anti-rabbit IgG and avidinDTexas Red conjugate were purchased fromVector Laboratories (Burlingame, CA). Goatanti-rabbit IgG horseradish peroxidase (HRP)and 4-chloro-1-naphthol were purchased fromSigma (St. Louis, MO). OCT tissue freezingcompound was obtained from Sakura Finetek(Torrance, CA). Sodium dodecyl sulfate (SDS)polyacrylamide gradient minigels were pur-chased from Novex (San Diego, CA). UDP-14C-galactose (300 mCi/mmol) was obtainedfrom NEN (Boston, MA).Cell CulturePrimary human bladder SMCs and fibro-blast cells were isolated as described Baskin etal., 1993; Coplen et al., 1994 and grown inDulbeccos Modified Eagle Medium and HamsF-12 medium (1:1) containing 10% fetal bovineserum and penicillin/streptomycin in an atmo-sphere of 5% CO2in air at 37C. HeLa (humancervical carcinoma) and K562 (human myelog-enous leukemia) cells were grown in RPMI con-taining 5% fetal bovine serum.Purification of NucleiCultured SMC or fibroblasts were trypsinizedand washed from their flasks with culture mediacontaining 10% fetal calf serum. The cells werepelleted by centrifugation and resuspended inice-cold sucrose buffer I (0.32 M sucrose, 3 mMCaCl2, 2 mM Mg-acetate, 0.1 mM EDTA, 10 mMTris (pH 8), 1 mM dithiothreitol DTT, and 0.1%Triton X-100). Next, cells were lysed with aDounce homogenizer, and cell breakage was pe-riodically assessed by microscopic examination ofthe homogenate. When lysis was complete, thehomogenate was mixed with sucrose buffer II(2.2 M sucrose, 5 mM Mg-acetate, 0.1 mM EDTA,10 mM Tris (pH 8) and 1 mM DTT) and layeredovera2Msucrose cushion. The nuclei werepelleted by centrifugation (30,000g, 45 min, at4C) and resuspended in glycerol storage buffer262 Billings et al.(40% glycerol, 5 mM MgCl2and 0.1 mM EDTA).For analysis, freshly purified nuclei were spreadon glass slides, fixed for 10 min with PBS con-taining 1.5% formalin, and subsequently incu-bated with propidium iodide, which specificallystains DNA.Marker EnzymesGalactosyltransferase activity (Golgi markerenzyme) was determined using the modifiedprocedure of Rens-Domiano and Roth 1989.Briefly, reactions contained extract (2050 mgprotein) in 25 mM HEPES (pH 7), 1 mM DTT,0.5% Triton X-100, 40 mM MnCl, 2 mM ATP,20 mM N-acetylglucosamine, and 2 mM UDP-14C-galactose (NEN, specific activity 300 mCi/mmol) in a total volume of 150 ml. The reac-tions were incubated for1hat37C andterminated by the addition of 50 ml of 0.2 MEDTA and passed over 0.5 ml of Dowex 1 resin(chloride form). The radiolabeled product 14C-lactosamine was eluted with 1 ml of water andcounted in a liquid scintillation counter.Cytochrome C reductase activity (ER/microsomal marker enzyme) was assayed asdescribed Phillips and Langdon, 1962, exceptthat reactions were performed in 25 mM Tris(pH 7.5), 300 mM NaCl, containing 50 mM2,6-dichloroindophenol, and 50 mM NADPH. Theoxidation of NADPH was monitored at 340 nm(E3405 6.22 mM21cm21) Gromer et al., 1998.Antibody Production and PurificationThree synthetic peptides corresponding to dif-ferent regions of human big-h3 (Table I) weresynthesized and purified by high-pressure liquidchromatography (BioSynthesis, Lewisville, TX).Each peptide was coupled to maleimideactivated-KLH (Pierce, Rockford, IL) at a ratio of1 mg peptide/mg KLH. Antibodies were raised inrabbits by Cocalico Biologicals (Reamstown, PA)using the following immunization schedule: aninitial injection of 250 mg, followed 3 weeks laterby three biweekly injections of 100 mg. For affin-ity purification of antibodies, individual peptideswere immobilized on Sulfolink coupling gel(Pierce, Rockford, IL). The IgG fractions werepassed over the column, and the column waswashed with PBS (;10 column vol) until proteinfree. Bound antibodies were eluted with 100 mMglycine (pH 2.5), immediately neutralized withTris base, and adjusted to a concentration of 0.4mg/ml before storage at 220C Harlow andLane, 1988; Chao et al., 1996.Enzyme-Linked Immunosorbent Assaysbig-h3 levels in conditioned medium were de-termined by direct enzyme-linked immunosor-bent assay (ELISA) Kucich et al., 1998. Briefly,microtiter plates were coated with knownamounts of free peptide or conditioned mediumdiluted in 100 mM Na-carbonate buffer (pH9.6). The plates were next incubated withAb1077 (diluted 1:50,000 in PBST) for1hatroom temperature, washed three times withPBST, and incubated with anti-rabbit-HRP (di-luted 1:2,000 in PBST) for 60 min at 37C. Theplates were washed three times with PBST,incubated with peroxidase substrate O-phenyl-ethylenediamine (1 mg/ml) in 50 mM citratebuffer (pH 4.6), 0.01% H2O2for 30 min andread at 450 nm. Standard curves, generated bycoating the plates with known amounts of im-munizing peptide (ALPPRERSRLL-C), wereused to quantitate big-h3 present in media.Western Blot AnalysisProteins were resolved on 8%16% polyacryl-amide gradient gels under reducing conditionsand transferred to nitrocellulose membranesChao et al., 1996. After transfer, the mem-branes were blocked in PBST and 5% nonfat drymilk and subsequently incubated with anti-big-h3 antibodies diluted 1:1,000 (0.4 mg/ml) inPBST, 0.1% bovine serum albumin. Bound anti-body was detected with goat anti-rabbit HRP-conjugate, using chloro-naphthol as substrate.ImmunohistochemistryHuman bladder tissue was embedded inOCT freezing compound and stored at 280C.Five- to 7-mm frozen sections were cut andplaced on albumin-coated slides. The tissuewas fixed for 5 min in 0.4% formaldehyde/PBS,followed by treatment with two 5-min rinses ofTABLE I. Human Peptide Sequences Used toDevelop Antibodies Against big-h3Antibody Peptide sequenceResidueposition1186 IGTNRKYFTNCKQWYQRKIC 55741073aTQLYTDRTEKLRPEMEG-C 1181341077aALPPRERSRLL-C 549559aA Cys residue was added to the carboxy terminus to fa-cilitate coupling to KLH.263big-h3 Synthesis by Bladder Cells100 mM NH4Cl/PBS. The tissue was thentreated with 6 M guanidine-HCl followed by100 mM iodoacetamide to unmask epitopes inthe extracellular matrix Gibson et al., 1997and was subsequently blocked with PBST con-taining bovine serum albumin (BSA). The tis-sue was incubated with affinity-purified pri-mary antibody overnight at 9C (1:20 dilution),washed and incubated with biotinylated anti-rabbit IgG (1:200 dilution), followed by avidinDTexas Red (1:125 dilution). To assess nu-clear localization with antibody Ab 1073, fro-zen sections of human and mouse bladder tis-sue were treated directly with the affinity-purified primary antibody (1:20 dilution) in theabsence of guanidine-HCl/iodoacetamide treat-ment.Cells (2 3 105) were seeded in sterile eight-well Lab Tek chamber slides (Nunc, Naper-ville, IL) and grown to confluence. The cellswere washed twice with PBS, fixed with PBScontaining 1.5% formalin for 10 min at 20C,washed with PBS, and permeabilized with PBScontaining 0.1% Triton X-100 and 1% BSA.Next, cells were incubated with primary anti-body (diluted 1:400 in PBST, 1% BSA) at 4Covernight, washed three times with PBST, andsubsequently incubated with a goat anti-rabbitIgG-rhodamine conjugate (diluted 1:500 inPBST, 0.1% BSA) for1hat20C. The slideswere washed, coverslipped, and examined witha Zeiss fluorescent microscope equipped withepifluorescence optics. Control slides weretreated in an identical manner, but with theaddition of 10 mg/ml immunizing peptide toblock primary antibody binding.Northern Blot AnalysisTotal cytoplasmic RNA was extracted frombladder cells with guanidinium thiocyanate/phenol-chloroform as described Chomczynskiand Sacchi, 1987. RNA was size fractionatedon 1% agarose-formaldehyde gels and trans-ferred to a Zeta-Probe membrane (Bio Rad,Hercules, CA). The filters were hybridized witha 2.1-kb human big-h3 cDNA (GenBank Acces-sion Number: M77349) probe (106cpm/ml hy-bridization mix) labeled with 32P using aReady-To-Go DNA labeling kit (AmershamPharmacia Biotech, Piscataway, NJ) to a spe-cific activity of 0.51 3 109cpm/mg. RNA load-ing and transfer were evaluated by probingwith a 1.4-kb glyceraldehyde phosphate dehy-drogenase (GAPDH) cDNA probe Kucich etal., 1997. Equivalent loading and transferwere also verified by quantitative image anal-ysis of ethidium bromide staining of ribosomalRNA, within the blots themselves. The phos-phorimages of the filters were digitized (Storm840, Molecular Dynamics, Sunnyvale, CA), andan identical rectangle was drawn around theprobe-specific signal for each lane of the North-ern blot. The relative response was quantifiedby integration of the observed pixel densitywithin each rectangle (Image-Quant V5.1 soft-ware; Molecular Dynamics, Sunnyvale, CA)yielding a volume measurement that reflectsmRNA content.RESULTSExpression of big-h3 by Cultured Bladder SMCand FibroblastsThe first series of experiments assessedbig-h3 gene expression in several cell types.Total RNA was extracted from primary humanbladder SMC and fibroblasts, as well as HeLa,and K562 cells and analyzed on Northern blots.Both bladder fibroblasts and SMCs expressed abig-h3 transcript of ;3.4 kb, whereas HeLaand K562 cells did not (Fig. 1). Bladder SMCexpressed higher levels of big-h3 (approximate-ly twofold) than fibroblasts. We have observedcomparable levels of expression in primary hu-man lung fibroblasts and SMCs (data notshown), demonstrating that these cell typesexpress big-h3 RNA in other organ systems, aswell.To assess big-h3 protein production by blad-der wall cells, peptide antibodies were gener-ated against the N- and C-terminal regions ofthe protein (see Materials and Methods sec-tion). SMC and fibroblasts were grown to nearconfluence 8 6 0.5 3 106cells/T75 flask, themedium was changed and, at 24-h intervals,medium samples were collected and theamount of big-h3 was quantitated by directELISA. SMC synthesized approximately three-five times the amount of protein compared withfibroblasts (Fig. 2).The results presented above (Figs. 1 and 2)demonstrate that bladder cells in vitro ex-pressed big-h3 that was secreted into the cul-ture medium. To determine protein depositionin the cell-layer/matrix compartment, bladdercells were seeded in slide chambers, fixed andincubated with anti-big-h3 antibodies. Whentreated with the C-terminal antibody (Ab264 Billings et al.1077), fibrous ECM stained brightly (Fig. 3A,C)in cultures of both SMC and fibroblasts. Inaddition, intracellular cytoplasmic stainingwas observed surrounding the nucleus, consis-tent with rough endoplasmic reticulum andGolgi patterns (Fig. 3B). Very fine punctatestaining of nuclei was observed with antibodyAb1077 (Fig. 3B,C).Surprisingly, a different staining patternwas obtained with Ab 1073 (directed at theN-terminal region of the protein). Bladder cellstreated with this antibody exhibited prominentnuclear staining, whereas the nucleolar re-gions were devoid of staining (Fig. 4). Thisstaining was not associated with the outer nu-clear membrane as judged by focusing aboveand below the plane of the cell monolayer. Nu-clear staining with Ab1073 had a fine punctate(stippled/spotted) appearance and was more in-tense than that observed with Ab 1077. Occa-sionally, very fine staining of the ECM was alsoobserved with Ab 1073. In addition, nuclearstaining was also observed when cells wereincubated with a third peptide antibody, Ab1186 (data not shown). To verify antibody spec-ificity, cells were treated with: 1) preimmuneserum, 2) primary antibodies that were prein-cubated with the corresponding immunizingpeptide before exposure to cells, or 3) secondaryantibody alone. Under these conditions, immu-nostaining of cells was consistently negative(Figs. 3D and 4C).Our immunohistochemical studies revealedthat big-h3 produced by bladder cells localizedto both intracellular and extracellular loca-tions. To verify the presence o