Inhibition of basic fibroblast growth factor-induced growth promotion ...

THE JOURNALOF BIOLOGICAL CHEMISTRY

Vol. 268, No. 32, Issue of November 15,pp. 24215-24222.1993 Printed in U.S.A.

Q 1993 by The American Society for Biochemistry and Molecular Biology, Inc.

Inhibition of Basic Fibroblast Growth Factor-induced Growth Promotion byOverexpression of Syndecan-1” (Received for publication, May 3, 1993, and in revised form, July 8, 1993)

Markku Mali$#, Klaus Elenius$#, Heini M. MiettinenSG, and Markku Jalkanengli From the $Department of Medical Biochemistry, Uniuersity of Turku, Kiinamyllynkatu 10, SF-20520 Turku and the STurku Centre for Biotechnology, P. 0.Box 123, SF-20521 Turku, Finland

The expressionof syndecan-1, the prototype member tions of growth factors, of their receptors, and of matrixof the cell surface proteoglycan family, follows mor- binding cell surface molecules may orchestrate, together with phogenetic rather than histological boundaries during genetic programs, the events that result in early embryogenorgan formation. As a heparan sulfate-containingcell esis. surface molecule,syndecan-1 can simultaneously bind Members of the heparin-binding growth factor family exvarious components of the extracellular matrix and hibit expression patterns that are typical for molecules inmembers of the heparin-binding growth factors. In- volved in regulation of growth and differentiationduring early deed, syndecan-1 may act as a co-receptor for basic fibroblast growth factor (bFGF) (Salmivirta, M., development (Burgess and Maciag, 1989). Binding of basic Heino, J., and Jalkanen,M. (1992)J.Biol. Chern. 267, fibroblast growth factor (bFGF) to itstyrosine kinase receptor 17606-17610),allowing the growth factor to bind the requires prior binding to heparin, to free heparan sulfate, or tyrosine kinase bFGF receptor. We have studied the to heparan sulfateproteoglycan at thecell surface (Rapraeger et al., 1991; Yayon et al., 1991). Earlier studieshave suggested role of syndecan-1in growth factor response growby ing 3T3 cells transfected with syndecan-l in the pres- that binding of bFGF to heparan sulfate might protect the ence of bFGF. The enhanced expression of syndecan-1 growth factor from degradation (Saksela et al., 1988). It has caused down-regulation of bFGF-induced cell prolif- further been proposed that heparan sulfate proteoglycans eration and, at thesame time, enhancement of cell could provide a reservoir of bFGF atthe cell surface, a matrix interactions. Thus, the induced expression of reservoir which might otherwise be released either by protethe heparansulfate co-receptor (syndecan-1)may provide a mechanismto restrict FGF action and modulate olysis of the proteoglycans or by partial degradation of heparan sulfate chains (Saksela andRifkin, 1990). cell-matrix interactions to maintain co-ordinated Members of the syndecan family have conserved transmemgrowth of cells during organ formation. brane and cytoplasmic domains but heterogeneous ectodomains containingdifferentially organized SerGly-duplexes for A fundamental question in developmental biology is how heparan sulfate attachment, and they may be viewed as cancell growth is co-ordinated and spatiotemporally regulated. didate molecules for heparan sulfate co-receptors of FGF Numerous growth factors and theirreceptors have been iden- (Bernfield et al., 1992; Jalkanen et al., 1993). In addition to tified, and it generally is appreciated that theirspatiotemporal bFGF (Kiefer et al., 1990; Elenius et al., 1992), syndecan-1 expression plays a major role in growth regulation. Just as binds to many ECM molecules (e.g. fibronectin, fibrillar colgrowth promotion during the induction of cell differentiation lagens, thrombospondin, tenascin)via its heparansulfate side and organogenesis is necessary, so also is restriction of growth. chains (see Jalkanen et al., 1991). Furthermore, syndecan-1 Defects in the regulation of growth factor function could lead simultaneously binds to components of ECM and bFGF (Salto abnormal growth and carcinogenesis. In addition, growth mivirta et al., 1992), suggesting that syndecan-like molecules stimulation is known to depend on the microenvironment can direct FGF-growth promotion to cell-cell and cell-matrix around cells, suggesting that cells also require positional sig- interfaces. nals which theycanobtain by receptors for extracellular Syndecan-1 is localized in morphologically active mesenmatrix (ECM)’ molecules. Therefore, the co-ordinated func- chymal cells during development of tooth (Thesleff et al., 1988), limb (Solursh et al., 1990), and kidney (Vainio et al., * This work was supported by The Finnish Academy, The Finnish 1989). Syndecan-1 is also simultaneously expressed with one Cancer Union, The Technology Development Centre (TEKES), The United States Public Health Service (DE09399-011, The Emil Aal- of the FGF members (int-2) in the condensing mesenchyme tonen Foundation, The Paulo Foundation, The Research and Science during tooth development (Thesleff et al., 1988; Wilkinson et Foundation of Farmos, and The Duodecim Foundation. The costs of al., 1989). Interestingly, Elenius et al. (1992) have shown that publication of this article were defrayed in part by the payment of bFGF and transforming growth factor 0 (TGF-P), especially page charges. This article must therefore be hereby marked “aduer- when administered at the same time, increased the surface tisenent” in accordance with 18 U.S.C. Section 1734 solely to indicate expression of syndecan-1 in NIH 3T3 cells, which resulted in this fact. ll To whom correspondence should be addressed Turku Centre for enhanced binding of bFGF to cells. To further investigate the Biotechnology, Tykistokatu 6,Biocity, P. 0. Box 123, SF-20521 possible function of syndecan-1 in mesenchymal cells, we Turku, Finland. Tel.: +358-21-633-8601; Fax: +358-21-633-8000. transfected NIH 3T3 cells with constructs expressing mouse The abbreviations used are: ECM, extracellular matrix; bFGF, syndecan-1. Unexpectedly, cells expressing about 10-fold basic fibroblast growth factor;BSA, bovine serum albumin;FCS, fetal more syndecan-1 than control cells were found to downcalf serum; FGF, fibroblast growth factor; HBDFN, heparin-binding domain of fibronectin; ”‘I-dU, ”‘iododeoxyuridine; PBS, phosphate- regulate bFGF-induced cell proliferation. These results sugbuffered saline; TGF-@,transforming growth factor-@;DMEM, Dul- gest that induced expression of syndecan-1 may, in fact, restrict the effect of FGF in developing mesenchyme during becco’s modified Eagle’s medium; EGF, epidermal growthfactor; PCR, polymerase chain reaction. organogenesis and tissue regeneration. This mechanism could

24215

24216

Syndecan-1and bFGF Response

restrict the effect of growth factor without necessarily eliminating the factor itself. Moreover, stimulatedsyndecan-1 expression could promote cell binding to ECM molecules, thus at the same time participating in the formation of cellular organization and tissue architecture.

for 1h with 30 pg/ml primary antibody 281-2 (monocIonal antibody against mouse syndecan-1 ectodomain; Jalkanen et al., 1985). Fluorescein-conjugated rabbit anti-mouse immunoglobulins (Dakopatts) was used as a secondary antibody. After the cells were washed with PBS, the coverslips were mounted, using Glysergel (Dakopatts), and viewed under epifluorescence. Quuntitation of Syndecan-1-Quantitation of syndecan-1 at the cell surfaces and itsaccumulation in the medium wasmeasured using MATERIALS ANDMETHODS radioiodinated 281-2 as described earlier (Elenius et al., 1992). Cell Culture and Growth Factor Treatment-NIH 3T3 fibroblasts For Western blot experiments, cell surface syndecan-1 ectodomain were routinely cultured in bicarbonate-buffered Dulbecco's modified was size-fractionated by SDS-polyacrylamide gel electrophoresis (graEagle's medium (DMEM, Life Technologies, Inc.) containing 5% dient 2-22.5%) and transferred onto Zeta-Probe membrane. Syndeheat inactivated fetal calf serum (FCS, Life Technologies, Inc.). In can-1 antigen was thereafter detected with radioiodinated 281-2, as experiments where dexamethasone induction of the RSV-MMTV- described by Elenius et al. (1992). LTR promoter (Rous sarcoma virus-mouse mammary tumor virusRNA isolation and Northern blot analysis were carried out accordlong terminal repeat) was involved, the culture medium was supple- ing to Mali et al. (1990). Filters were probed with multiprime-labeled mented with 1 p M dexamethasone. In other experiments to test the pM-4, the partial cDNA of syndecan-1 (Saunderset al., 1989). bFGF response, 50,000 cells were transferred into six-well culture Cell Binding Assay-A 96-well culture plate (Nunc) was coated plates or10,000 cells into 24-well cultureplates(Greiner). The with the heparin-binding domain of fibronectin (HBDFN) (a-chyfollowing day, medium was removed, and thecells were incubated for motrypsin 40-kDa proteolytic fragment; Calbiochem;5 pg/ml) in PBS 24 h in fresh medium containing 2% carboxymethyl-Sephadex-eluted for 2 h at 37 "C. The plates were then blocked with 1%BSA in PBS FCS (Vogel et al., 1978). To these cultures, 20 ng/ml recombinant for 1 h at 37 "C and washed with PBS. Cultured cells (as described human bFGF (Promega or Peprotech), 1 ng/ml aFGF (acidic FGF, above) were detached, using 0.5 mM EDTA-PBS, centrifuged, and Peprotech), or 15 ng/ml EGF (epidermal growth factor, Sigma) was resuspended into DMEM. Then, 50,000 cells were transferred into added for an additional 24-h culture period. Next, 0.25 pCi/ml wells and incubated for 1h at 37 "C. Next, the cells were washedwith 126i~dodeoxyuridine (lZ5I-dU,New England Nuclear) was added to the DMEM, fixed in 2% paraformaldehyde, stained with 5% crystal cells. Four hours later, cells were washed with phosphate-buffered violet, and washed with distilled water. The culture plates were saline (PBS) and solubilized in 1 M NaOH. Radioactivity was meas- allowed to airdry, after which the cells were suspended in 10%acetic ured using a gamma counter (LKB). acid and spectrophotometrically measured at 600 nM. To inhibit cell Expression Vectors and Transfections-A stretch of 743 nucleotides binding, heparin was added to themedium a t 100 pg/ml. Nonspecific from the 5' end of mouse syndecan-1 was cloned by polymerase chain binding was measured by replacing HBDFN with BSA. reaction (PCR), using primers that had been generated according to Iz5I-bFGFBinding Assay-Binding of lz5I-bFGFonto cell surfaces the published sequence (Saunders et al.., 1989). DNAwas synthesized was estimated essentially as previously described by Nugent and by reverse transcription of total RNA isolated from normal murine Edelman (1992). bFGF was radioiodinated by the chloramine-T mammary gland epithelial cells using a primer corresponding to a method (Stahli et al., 1983).Cells were transferred into 24-well culture sequence immediately 3' from the desired amplification sequence (5'- plates and cultured in DMEM containing 5% FCS. The following day GGCTGATGG-3'). The DNA was further amplified by PCR (Saiki the culture medium was replaced with medium containing 2% caret al., 1988) using the following pair of primers: 5"AAGGAATTboxy-methyl-Sephadex-eluted-FCS, and cells were cultured for an CACTCCGCGGGAGAGGTGCGG-3" (EcoRI-site underlined) and additional 24 h. Cells were washed three times with blocking medium F-GGAGCCGAGGTCTCATGGAGGCCAGGTT-3'(corresponding 0.05% BSA in DMEM and subsequently incubated with 10 ng/ml of to nucleotides 716-743 immediately 3' from a unique SphI site). The '"1-bFGF (25,000 counts/min/ng) in blocking medium for 2.5 h at reactions were carried out using a RNA PCR-kit anda DNA thermal 4 "C. After incubation cells were washed three times with PBS at cycler (both from Perkin Elmer).The PCR product was digested with 4 "C.Bound '"I-bFGF was extracted by washing for 5 s with 1 ml of EcoRI and SphI andligated into pGEM3Z (Promega) for sequencing 2 M NaC1, pH 7.4, and after that two times for 1 min with 500 pl of (Sanger et al., 1977). Purified 5' insert was then ligated with purified 2 M NaC1, 20 mM sodium acetate, pH 4.0. The '"I-bFGF extracted at 3' insert (SphI-EcoRI fragment; Saunders et al., 1989) as a three- pH 7.4 represents mainly bFGF bound to cell surface heparan sulfate piece ligation into the EcoRI site of Bluescript SK+. The derived proteoglycan, and the radioactivity extracted at pH 4.0 represents plasmid was named pBS-msyn. The pBS-msyn construct was digested bFGF bound to a complex composed of heparan sulfate proteoglycan with XbaI and SalI (polylinker sites around the EcoRI site), and the and tyrosine kinase FGF receptor (Nugentand Edelman, 1992). derived insert of mouse syndecan-1 cDNA was ligated into theNheI/ Extracted radioactivity was measured using a gamma counter (LKB). SalI site of pMAMneo (Clontech). This plasmid was named Nonspecific binding was estimated by adding 1,000-foldexcess of cold pMAMneo-msyn. The construction of a control transfection vector bFGF with '"I-bFGF into incubation medium. Nonspecific binding pMAMneo-hGH was as described earlier (Leppa et al., 1992). The EcoRI-digested full-length mouse syndecan-1 insert was also was approximately 15 or 30% of total Iz5I-bFGFbound when cells ligated into the EcoRI site of a vector pBGS2 containing a SRa were washed at pH 7.4 or 4.0, respectively. Cell numbers were estipromoter (Takebe et al., 1988) and a neomycin-resistance cassette (a mated by counting cells from parallelly cultured wells. gift from Bruce L. Granger, Montana State University, Bozeman, RESULTS MT). Thederived plasmid was named pBGS-msyn. All transfection plasmids were isolated by the CsCl density gradient Transfection of NIH 3T3 Cells with Syndecan-l-Fullmethod. Before transfections, pMAMneo constructs were linearized length cDNA of mouse syndecan-1 was generated using the with MluI and pBGS-msyn was linearized with XbaI. Thereafter, the plasmids were extracted in a chloroform-phenol solution and precip- PCR method and a partial cDNA clone of mouse syndecan-1 (pM-4;Saunders et al., 1989).Two different transfection itated with ethanol. Transfections were performed using lipofectin (Bethesda Research vectors were used in this study. The full-length syndecan-1 Laboratories) liposome transfection according to the manufacturer's cDNAwas cloned into pMAMneo expression vector. The instructions. After selection for 2 weeks with Geneticin (G418, 750 derived plasmid was named pMAMneo-msyn. The pMAMneo pg/ml, Sigma), surviving clones were isolated from growth plates vector containsa dexamethasone-inducible RSV-MMTVusing cloning cylinders. Expression of syndecan-1 was detected by LTR promoter and a neomycin resistance gene for stable immunocytochemistry, and those cells expressing high amounts of syndecan-1 were selected for further studies. Transfected stock cells transfection (see Leppa et al., 1992).As a control transfection, we used a pMAMneo-hGH construct, containing exons four were routinely cultured with Geneticin (250 pg/ml). Fluorescence Microscopy-Cells on coverslips were fixed for 10 min and five from human growth hormone gene (Leppa et al., in ice-cold methanol. After washing with PBS, the cells were prein- 1992). To exclude the possibility that dexamethasone treatcubated for 15 min in blocking medium (PBS supplemented with ment could influence our results, we made an additional 0.2% bovine serum albumin (BSA)). The cells were then incubated H. M. Miettinen, B. L. Granger, and M. Jalkanen, manuscript in preparation.

control by constructing a transfection vector (pBGS-msyn) in which mouse syndecan-1 was brought under SRa-promoter (see "Materialsand Methods"). All theseconstructs were

24217

Syndecan-1 and bFGF Response transfected into NIH 3T3 cells using liposome transfection. After Geneticin selection, separate colonies were cloned and analyzed by immunofluorescence microscopy, using monoclonal antibody 281-2 against mouse syndecan-1 (Jalkanenet al., 1985). pMAMneo-msyn and pBGS-msyn clones expressing high levels of syndecan-1 were designated SynlO.2 and Synl.5, respectively. In addition, one of the several pMAMneo-hGH-transfected cell clones was chosen for use and designated hGH4.4. In the presence of 1 p~ dexamethasone, intense syndecan-1 staining was observed in Syn10.2 cells (Fig. 1F).In the absence of dexamethasone, only weak expression could be seen (Fig. 1E).The hGH4.4 clone (Fig. 1, C and D )and wild type cells (Fig. 1,A and B)were negative, irrespective of dexamethasone treatment. Synl.5 cells containing the SRa-promoter also exhibited bright staining(Fig. 1H) when compared to wild type cells (Fig. IC). Comparable results were observed by Northern analysis of the transfected cell lines (Fig. 2). The Synl.5 cells and dexamethasone-induced SynlO.2 cells contained the expected transfectedsyndecan-1 mRNA species (3 and 4 kilobases, respectively; Fig. 2B, lane 6; Fig. 2 A , lane 4 ) . These mRNA species were not detected in the wild type cells (Fig. 2B, lane 5), uninduced cells (Fig. 2.4, lane 3 ) or control transfected cells (Fig. 2 A , lanes 1 and 2 ) . Transfection Induces High Expression of Syndecan-I at the Cell Surface and Increases Shedding into the Culture Mediurn-Cell surface syndecan-1 content is easily measured by radioimmunoassay of syndecan-1 ectodomain released by mild trypsin treatment (Jalkanen etal., 1987). This assay can also be used to estimate the level of shedding of syndecan-1 by 3T3 cells (Elenius et al., 1992). Dexamethasone treatment of syndecan-l-transfected cells (SynlO.2 clone) induced high levels of syndecan-1 expression at the cell surface (Fig. 3A) as well as shedding into theculture medium (Fig. 3B). On the otherhand, when SynlO.2 cell cultures were deprived of dexamethasone, low levels of syndecan-1 were detected (Fig. 3). With controlhGH4.4 cells, low levels of syndecan-1 were detected, regardless of dexamethasone treatment (Fig. 3). The Synl.5 cells expressed high levels of syndecan-1, both at the cell surfaces (Fig. 3A) and in the culture medium (Fig. 3B). By scanning densitometry, the syndecan-1 expression at the surface of Synl.5 cells and dexamethasone-induced SynlO.2 cells was found to be about 10-fold when compared to control cells. It was clearly seen from these experiments that the enhanced syndecan-1 expression at the cell surface also resulted in increased accumulation of syndecan-1 ectodomain in theculture medium. TOestimate the molecular weight of cell surface syndecan1, the molecules were size-separated by SDS-polyacrylamide gel electrophoresis and thenWestern-blotted. In both hGH4.4 and Syn10.2 cell preparations, a faint smearof about 350 kDa (Fig. 44)was observed. A syndecan-1 molecule of this size is typical for mesenchymal cells (Salmivirta et al., 1991; Elenius et al., 1992). When expression of syndecan-1 was induced by dexamethasone in Synl0.2 cells, a smear corresponding to a proteoglycan of about 250 kDa was detected (Fig. 4A). A syndecan-1 molecule of smaller size was also observed in other pMAMneo-msyn transfected clones in the presence of dexamethasone (SynlO.9 in Fig. 44). The Synl.5 cell lineexpressed a syndecan-1 molecule that was comparable in size (350 kDa) to theendogenous syndecan-1 (Fig 4B). Enhanced Expression of Syndecan-1 Promotes 3T3 Cell Binding to the Heparin-binding Domain of Fibronectin-Isolatedsyndecan-1 molecules have been shown to bind, via heparan sulfateside chains, to theheparin-binding domain of fibronectin(Saundersand Bernfield, 1988; Elenius et al.,

1990). To demonstrate that the cell surface syndecan-1 in transfected 3T3 cells was functional, we tested the ability of transfected cells to bind to HBDFN. The SynlO.2 cells CUIturedinthe presence of dexamethasone (expressing high syndecan-1 levels) showed increased binding to HBDFN when compared to the binding of wild type or control transfected hGH4.4 cells (low syndecan-1 levels, Fig. 5).When the SynlO.2 cells were cultured withoutdexamethasone (low syndecan-1 levels), they did not bindto HBDFN (Fig. 5). Binding of all these cells to HBDFN was inhibited by supplementing the incubation medium with heparin (100 pg/ml). Similar to SynlO.2 cells, Synl.5 cells showed increased binding to HBDFN. Further, this binding was inhibited by heparin (data not shown). Enhanced Syndecan-1 Expression Down-regulates bFGFinduced Cell Proliferation-Recent experimental results from several laboratories support the view that syndecan-like molecules may participate in the formation of bFGF-receptor complexes (Rapraeger et d., 1991; Yayon et d., 1991; Olwin and Rapraeger, 1992; Kan et al., 1993). To investigate this phenomenon with our transfectants, cell proliferation was measured by counting incorporation of 'T-dU, 24 h after bFGF (20 ng/ml)additionin the dexamethasone-treated Synl0.2, hGH4.4, and wild type cell cultures. Increased proliferation was observed with the wild type and hGH4.4 cells, but the syndecan-l-positive Syn10.2 cells did not increase their DNA synthesis under bFGF influence (Fig. SA). When SynlO.2 cells were grown in the absence of dexamethasone, however, bFGF treatment stimulated Syn10.2 cell proliferation (Fig. 6A). To rule out the possibility that dexamethasone treatment itself might have an influence on these results, Synl.5 cells unexposed to dexamethasone were cultured with bFGF. No enhancement of proliferation in bFGF-treated Synl.5 cells was observed (Fig. 6B). Similar kinds of results were also obtained when the number of cells was counted (data not shown). Maximal proliferation of wild type 3T3 cells, in response to incubation with bFGF, was observed at 20 ng/ml, as shown in Fig. 6C. The high levels of syndecan-1 at thesurfaces of transfected 3T3cells could conceivably have disturbed the assembly of growth factor-receptor complexes. Therefore, we tested transfected cells that had been grown in the presence of high growth factor concentration. As seen in Fig. 6C,bFGF at concentrations as high as 100 ng/ml did not promote cell growth, indicating that 3T3 cells transfected with syndecan-1 had become unresponsive to bFGF. Syndecan-l-transfected cells were also unresponsive to aFGF, another heparin-bindinggrowth factor (Fig. 7A). However, EGF, which does not interact with heparan sulfate, generated a proliferative response both in syndecan-1-transfected Synl.5 cells and in wild type cells (Fig. 7B). Enhanced syndecan-1 expression reduced bFGF response but raised the question of altered bFGF-binding to the surfaces of syndecan-1-transfected 3T3 cells. This was studied by incubating Synl.5 and wild type cells with ''9-bFGF. No significant differences were observed between these cells as both low affinity (2 M NaC1, pH 7.4) and high affinity (2 M NaC1, 20 mM sodium acetate, pH 4.0) binding sites (Nugent and Edelman, 1992) were found to be present at similar densities (Fig. 8). DISCUSSION

In this paper, we present evidence suggesting that syndecan-1 plays not only a positive but also a negative regulatory role for cell growth. When syndecan-1 or a related molecule is expressed at low levels, itcansupport the activity of heparin-binding growth factors. However, at higher concen-

24218

Syndecan-1 and bFGF Response

FIG.1. Immunofluorescencestaining of syndecan-1-transfected cell clones. Methanol-fixed cells were first incubated with a monoclonal antibody against mouse syndecan-1 (mAb 281-2), and subsequently with fluorescein-conjugated secondary antibody. Panels A, C, E, G, and H depict cell staining after culture in the absence of dexamethasone. Panels B, D, and F depict cell staining after culture in the presence of 1 p M dexamethasone. Panels A and B, wild type cells; panels C and D, transfection control cells transfected with pMAMneo-hGH4.4; panels E and F, pMAMneo-msyn-transfected SynlO.2 cells; panel G, wild type cells; p a w l H , pBGS-msyn-transfected Synl.5 cells.

trations, the excess of heparan sulfate becomes suppressive for bFGF. Thus, the membersof the syndecan family may be versatile cell surface molecules regulating cell growth during earlydevelopmentorduringregeneration of tissues when growth is largely promoted by locally active growth factors. The findings that syndecan-1 expression coincides with cell growth during development support the concept that syndecan-1 mightbe involved in regulationof growth factor activity. In developing tooth mesenchyme, syndecan-1 expression is associated with a heparin-binding growth factor, the int-2 proto-oncogene product(Wilkinson et al., 1989).Theen-

hanced cell proliferation is, indeed, localized to the same site of the developing tooth mesenchyme (Vainio and Thesleff, 1992). A similar mechanism defining regulated proliferation may also be operative duringlimb development, as both bFGF and syndecan-1 havebeen localized to the proliferatingmesenchyme in chicken embryos (Joseph-Silversteinet al., 1989; Solursh et al., 1990). Moreover,during cutaneouswound healing, the synthesis of syndecan-1 is enhancedby proliferating keratinocytes of the wound edge and by endothelial cells of growing capillaries (Elenius et al., 1991). One explanation for syndecan-1 expression coinciding with growth factor expres-

and Syndecan-1

A

B

1 2 3 4

bFGF Response

24219

5 6

200 97 ca

FIG. 2. N o r t h e r n blot analysis of syndecan-1-transfected cell clones. Total RNAs (IO pg) from hGH4.4 (panel A, lanes I and 2 ) , SynlO.2 (panel A, lanes 3 and 4 ) , from wild type (panel E , lane 5),and from Synl.5 cells (panel B, lane 6 ) were size-fractionated and blotted onto a Genescreen FM membrane. The membrane was hybridized with a 32P-labeledpartial cDNA probe for mouse syndecan1 (pM-4). BeforeRNAisolationcellswere culturedeitherinthe presence (panel A, lanes 2 and 4 ) or in the absence(panel A, lanes I and 3; panel B , lanes 5 and 6 ) of 1PM dexamethasone. Arrows indicate the sizes of wild type syndecan-1 mRNAspecies, 2.6 and 3.4 kilobases.

A cell surface

Dex

46 30

!

14

FIG. 4. Western blot analysis of cell surface syndecan-1. Cell surface material, released by trypsin from hGH4.4 ( A ) ,SynlO.2 ( A ) ,Syn10.9 ( A ) , Synl.5 ( R ) ,and wild type ( WT,R ) cells, was sizeseparated using SDS-polyacrylamide gel electrophoresis, and blotted onto Zeta-Probe membrane. Syndecan-1was visualized using radioiodinated monoclonal antibody 281-2 and autoradiography. In panel A, hGH4.4D, SynlO.ZD, and SynlO.9Dcells were cultured in the presence of 1 p~ dexamethasone. Molecular massstandardsare shown between panels (kDa).

hCli4.4

Syn 1.5

0.7

Syn 10.2

WT

0.6-

T

WT

B medium

FIG. 3. The amount of syndecan-1 core protein at the cell surface and in the medium. Cell surface material (corresponding to 200,000 cells)released by trypsintreatment ( A ) and medium samples (corresponding to 100,000 cells) ( E ) were slot-blotted on a Zeta-Probe membrane. Syndecan-1 was detected using radioiodinated monoclonal antibody 281-2 andautoradiography. SynlO.2cells treated with 1PM dexamethasone (Dex) showed enhanced syndecan1 expression, both at the cell surface ( A ) and in the culture medium ( B ) . No enhancement of syndecan-1 expression was seen in control hGH4.4 cells. Synl.5 cells also showed increased expression at the cell surface ( A ) and in the culture medium ( B ) when compared to wild type ( W T )cells.

sion is that syndecan-1 expression might be enhanced by growth factors, e.g. by exposing mesenchymal cells to TGF-P and bFGF (Elenius et al., 1992). Additionally integrins, another family of cell adhesion molecules (Hynes, 1992), are known to modify theirexpressionundertheinfluence of cytokines(SantalaandHeino, 1991). Thissuggeststhat growth factors areheavily involvedin the regulation of matrix receptor expression and, thus, responsible for changes in cell adhesion duringgrowth promotion. It has been generally accepted that heparan sulfate is required for FGF action (Rapraeger et al., 1991; Yayon et al., 1991). If cells are made deficient in heparan sulfate, bFGF cannot bind to them, even though the cells have tyrosine kinase FGF receptors (Rapraeger et al., 1991; Yayon et al., 1991). Recent results strongly support the concept that the binding of bFGF to both receptors (the tyrosine kinase FGF receptor and the heparan sulfate co-receptor) must be simultaneous to induce signal (Nugent and Edelman, 1992; Olwin

WT D

hGH4.4 D

SynlO.2 D

WT

hGH4.4

SynlO.2

FIG. 5. Enhanced expression of syndecan-1 promotes cell binding to the heparin-binding domain of fibronectin. Cultured wild type ( W T ) ,hGH4.4, and SynlO.2 cells were plated on HBDFNcoated wells; and after 1 h of incubation, the non-adherentcells were washed away,and the remaining adherent cells were fixed and stained with crystal-violet. Stained cells were suspended in acetic acid and measured spectrophotometrically. The letter D after the nameof the cell clone indicates that cells were precultured in the presence of 1 p~ dexamethasone. All binding to HBDFN was totally inhibited by heparin. Nonspecific bindingto BSA was subtracted from initial results. Means and standard deviations of fourparallel wells are presented.

and Rapraeger, 1992). Furthermore, Kan et al. (1993) have recentlysuggested that binding of heparan sulfate to the second Ig-like domain of the tyrosine kinase FGF receptor is also required for FGF response. Thus, themolecules involved seem to form a ternary complex of a heparan sulfate proteoglycan, a tyrosine kinase FGF receptor, and an FGFligand. We elicited high amounts of syndecan-1 expression by a dexamethasone-inducibleRSV-MMTV-LTRpromoteras well as by a constitutively active SRa-promoterto investigate the role of syndecan-1inthe growth factoraction.The achieved syndecan-1 expression levels were comparable to those previously observed in 3T3 cells exposed to bFGF and TGF-P (Eleniuset al., 1992), indicating that the enhancement of syndecan-1 expression in transfected cells was within physiological limits. Because dexamethasone was found to downregulate basal proliferation rateof 3T3 cells (Fig. 6A, SynlO.2 versus SynlO.2D) and could thus have interferedwith growth

24220

Syndecan-1and bFGF Response CPM 12000

T

A

8000 FIG. 6. Enhanced expression of syndecan-1 down-regulates cell proliferation induced by bFGF. Cell proliferation was measured by lZ6I-dU incorporation (counts/min,y axis). Both 4000 a pMAMneo-msyn clone SynlO.2 ( A ) and a pBGS-msyn clone Synl.5 ( B and C ) were tested. In panels A and B, black bars indicate no bFGF treatment, and 0 hatched bars, bFGF (20 ng/ml) treatof ment. Means and standard deviations three parallel samples are presented. In p a d A , wild type ( W T ) ,hGH4.4, and CPM SynlO.2 cells withthe letter D after their symbols were cultured in the presence of 1 W M dexamethasone. Other results shown in this figure were obtained with- 20000 out dexamethasone treatment. Panel C shows the effect of different concentrations of bFGF on cell proliferation of wild type (open circles) and Synl.5 15000 (closed circles) cells. The bFGF concentration is shown on the x axis, and lZ6IdU incorporationof cells is shown on the 10000

WT D

hGH4.4 D

SynlO.2 D

SynlO.2

CPM

2woJBl

20000

15000

10000

y axis. 5000

5000 0

0

WT

Synl.5

0

4200

60

80

110200

[bFGF]nglml

regulation of 3T3 cells the results were controlled using SRa promoter that does not need dexamethasone activation. Another reason to generate constitutively expressing transfectants was that dexamethasone-induced Syn10.2 cells synthesized less glycosylated syndecan-1 molecules than wild type or Synl.5 cells (Fig. 4). It is not clear why dexamethasoneinduced cells synthesized a smaller syndecan-1 species, but this phenomenon could be related to a finding that glucocorticoids regulate the synthesis of heparan sulfate (Saarni, 1977). However, reduced glycosylation of syndecan-1 did not seem to interfere with the binding of Synl0.2 cells to HBDFN (Fig. 5). The syndecan-1-transfected cells did not exhibit enhanced growth; rather, they became totally unresponsive to bFGF growth stimulation. However, bFGF treatment did stimulate cell proliferation whenSynlO.2 cells were cultured in the absence of dexamethasone, indicating that syndecan-l-transfected cells did, in fact, have a functional tyrosinekinase FGF receptor. Stimulation of cell proliferation was not as high as that observed in control cells, possibly because some SynlO.2 cells still expressed syndecan-1, even in the absence of dexamethasone (Fig. 1E).These results clearly suggest that high amounts of syndecan-1 can disturb the effect of bFGF. Also acidic FGF had lost its ability to induce cell growth in syndecan-1 transfectedcells (Fig. 7 A ) ,which is not surprising as both aFGF and bFGF are known to bind heparan sulfateand FGF tyrosinekinase receptors in 3T3cells. Interestingly, 3T3 cells which showed high levels of syndecan-1 expression responded to EGF almost as well as wild type cells (Fig. 7B). This suggests a rather specific suppression of FGF action by syndecan-1 overexpression.

No significant differences in “‘1-bFGF binding between syndecan-1 expressing and wild type cells were observed, indicating that both cells expressed complexes composed of cell surface heparan sulfate proteoglycans and FGF tyrosine kinase receptors (Fig. 8). It is somewhat surprising that the enhancement of syndecan-1 expression did not result in the increase of bFGF-binding to heparan sulfate-binding sites. One reason for this could be that the increase of syndecan-1 expression is not sufficient to demonstrate differences in the total content of heparan sulfate at the cell surface. This is supported by findings that total uptake of 35S04into cell surface proteoglycan pool was only 2-fold higher in Synl.5 cells than in wild type cells (datanot shown). A second possibility could be that the overexpressed syndecan-1 contains regions responsible for bFGF binding (Turnbull et al., 1992) at a lower frequency than theheparan sulfate chains at the surface of wild type cells. Recently, Nurcombe and coworkers (1993) have reported that such changes exist among the neuronal, heparan sulfate-containing proteoglycans. There could be several other mechanism by which downregulation of bFGF could occur. One possibility involves increased shedding of syndecan-1 ectodomain into culture medium. This excess of syndecan-1 could interact with bFGF, thereby rendering it unavailable for further interaction with the tyrosine kinase FGF receptors. Supporting thisconcept is another proteoglycan, decorin, which can capture TGF-/3 to ECM; it is proposed that this is the mechanism by which decorin blocks TGF-@ action (Yamaguchi et al., 1990). Syndecan-1-containing conditioned Synl.5 medium was able to down-regulate bFGF-induced wild type cell proliferation to

and Syndecan-1

bFGF Response

24221

some extent (30-50%) but not totally (data not shown). On the other hand, we failed to detect any stimulation of cell growth with bFGF, even at very high concentrations (up to 4001000 ng/ml, Fig. 6C and data not shown), in syndecan-lpositive cell clones. Secondly, overexpression of syndecan-1 e 300 proteoglycan could inhibit cell proliferation intracellularly in a similar way as described for heparin. Interestingly, heparin 8 has been speculated to down-regulate a protein kinase c200 dependent pathway also involved in bFGF action (Wright et al., 1989). A third possibility is based on speculation that loo binding of bFGF to heparansulfate proteoglycan maybe required for clustering of the tyrosine kinase FGF-receptors on the cell surface (Ruoslahti and Yamaguchi, 1991). In such 0 7 a case, there could simply be too many heparan sulfate proWT Synl.5 teoglycans, compared to the tyrosine kinase FGF-receptors, in oursyndecan-1-transfected cells. The cluster formation .tW and dimerization (Ornitz et al., 1992) of the receptors would, thus, become disturbed,as would phosphorylation of the tyrosine kinase FGF receptor. It is also possible that as a EGF result of high syndecan-1 expression, some unknown factorb) - 300 e is(are) induced, and that such factor(s) could interfere with C receptor activity. 0 200 It is possible that during carcinogenesis suppression of u0 syndecan-1 expression is one of the critical events leading to 8 the non-restricted growth of some malignant cells. For example, UV light-induced skin cancers exhibit decreased syn100 decan-1content,and dedifferentiation of keratinocytes is associated with low amounts of syndecan-1 (Inki et al., 1991, 1992). When cultured in the presence of steroids, a hormon0 ally regulated mouse mammary tumor cell line (S115) exhibits WT Synl.5 transformed phenotype, malignant growth, and down-regulaFIG. 7. The effect of syndecan- 1 transfection on the proliftion of syndecan-1 expression (Leppa et al., 1991).When S115 erative response induced by aFGF and EGF. The response of wild type ( W T ) and Synl.5 cells to 1 ng/ml of aFGF (A) and to 15 cells are transfected with syndecan-1, however, they show ng/ml of EGF ( B )was tested. Cells were cultured either in the absence normal growth and epithelial-like morphology, regardless of (black bars) or in the presence (hatchedbars) of growth factors. The steroid treatment (Leppa et al., 1992). It is possible that amount of"'I-dU incorporation of control cells (no growth factor enhanced expression of syndecan-1 can restrict FGF activity treatment) is expressed as loo%, and theother values are corrected in syndecan-1-transfected S115 cells because a FGF-like according to this. Means and standard deviations of three parallel growth factor has been reported to be responsible for transsamples are shown. formed phenotype of S115 cells (Tanaka et al., 1990). As described above, syndecan-1 expression is enhanced not only by several condensing mesenchymes but also by regenerating tissues. Perhaps at these sites, the high amounts of syndecan-1 at cell surfaces suppress growth factor influence and prevent furtherinduction of cell growth initially triggered by reciprocal tissue contacts. In addition to binding growth factors, syndecan-1 is able to modulate cell-matrix interactions. This was observed with our syndecan-1-transfected cells, as they increased their binding to HBDFN. From the findings presentedherein, we propose a novel regulatory mechanism by which syndecan-mediated changes in cell matrix interactions combined with syndecan-associated changes in growth factor activities might control and regulate normal cell growth and development. 500

-

c)

T

aFGF

I

-

-

I

P T

c)

Synl.5

WT

FIG. 8. Binding of laaI-bFGFto syndecan-1-transfected and wild type cells. Synl.5 and wild type ( W T ) cells were cultured in 24-well plates and incubated with "'I-bFGF. After incubation, "'1bFGF was extracted with 2 M NaC1, pH 7.4, detaching mainly "'1bFGF bound to heparan sulfate (black burs). Subsequently, cells were washed with 2 M NaCl, 20mM sodium acetate, pH 4.0,eluting '"IbFGF bound to heparan sulfate-FGF tyrosine kinase receptor complexes (hatchedburs). Nonspecific binding was estimated by adding a 1000-fold excess of cold bFGF with radioiodinated bFGF, and nonspecific binding was subtracted from the initial results. Means and standarddeviations of four parallel samples are presented.

Acknowledgments-We thank Taina Kalevo and Liisa Peltonen for skillful technical assistance. The pBGS expression vector was a kind gift from Bruce L. Granger. The language of this manuscript was revised by Jeri L. Hill, Ph.D. REFERENCES Bernfield, M., Kokenyesi, R., Kato, M., Hinkes, M. T., Spring, J., Gallo, R. L., and Lose, E. J. (1992) Anno. Reu. Cell Biol. 8 , 365-393 Burgess, H. W., and Maciag, T. (1989) Annu. Rev. Biochern. 68,575-606 Elenius, K., Salmivirta, M., Inki, P., Mali, M., and Jalkanen, M (1990) J. Biol. Chem. 266.17837-17843 Elenius, 2.:Vainio, S., Laato, M., Salmivirta, M., Thesleff, I., and Jalkanen, M. (1991) J. Cell Biol. 114,585-595 Elenius, K., Miiitta, A,, Salmivirta, M., and Jalkanen, M. (1992) J. Biol. Chern. 267.6435-6441 Hynes;R. (1992) Cell 69, 11-25 I

~

~~

~

~~~

Syndecan-1 and bFGF Response

24222

Inki, P., Stenback, F., Talve, L., and Jalkanen, M. (1991) Am J. Pathol. 1 3 9 , 1333-1340

Inki., P.,Kujari, H., and Jalkanen, M. (1992) Lab. Inuest. 66,314-323 Jalkanen, M., Nguyen, H., Rapraeger, A., Kurn, N., and Bernfield, M. (1985) J. Cell BWl. 101,976-984 Jalkanen, M., Rapraeger, A., Saunders, S., and Bernfield, M. (1987) J.Cell Biol. 105,3087-3096

Jalkanen, M., Jalkanen, S., and Bernfield, M. (1991) in Receptors for Extracellulor Matrix (McDonald, J. A., and Mecham, R. P., eds) pp. 1-37, Academic Press, SanDiego Jalkanen, M., Elenius, K., and Rapraeger, A. (1993) Trends Glycoscience GlycotechnoL 6,107-120 Joseph-Silverstein, J., Consigli, S. A., Lyser, K. M., and Ver, P. C . (1989) J. CeU BWl. 108,2459-2466 Kan. M.. Wann. F.. Xu. J.. Grahb. J. W.. Hou. J.. and McKeehan. W. (1993) ~, .

Si&&

~I

I

~~I

I

I

26e~1sis-lisl'

103-106

""

~~~

~

1 0 7 , 743-751

Salmivirta, M., Heino, J., and Jalkanen, M. (1992) J. Biol. Chem. 2 6 7 , 1760617mn

Sa&i;F.,

Nicklen, S., and Coulson, S. R. (1977) Pmc. Natl. Acad. Sei. U. S. A .

74,5463-5471

Santala, P., and Heino, J.(1991) J. Biol. Chem. 266,23505-23509 Saunders, S., and Bernfield, M. (1988) J. Cell Biol. 106,423-430 Saunders, S., Jalkanen, M., O'Farrell, S., and Bernfield, M. (1989) J. Cell Biol. 107,1199-1205

Solursh, M., Reiter, R. S., Jensen, K. L., Kato, M., and Bernfield, M. (1990) Deu. Biol. 140,83-92 Stiihli, C., Miggiano, V., Stocker, J., Staehelin,R., Haring, P., and Takacs, B. (1983) Methods EnzymoL 92,242-253 Tanaka, A., Matsumoto, K., Nishizawa, Y., Lu, J., Yamanishi, H., Maeyama, M.. Nonomura. N.. Uchida. N.. and Sato. B. (1990) J.Steroid Biochem. Mol. Bidl. 37,23:29 ' Takebe, Y., Seiki, M., Fu'isawa, J. I., Hoy, P., Yokota, K., Arai, K.-I., Yoshida, M., and Arai, N. (1988j MOL Cell. Biol. 8 , 466-472 Thesleff, I., Jalkanen, M., Valnlo, S., and Bernfield, M. (1988) Deu. Biol. 1 2 9 , ~~

Kiefer, M. C., Stephens, J. C., Crawford, K., Okino, K., and Barr, P. J. (1990) -Pmc. .. Natl. Acad. Sci. IJ. - . S. A. 87.6985-6989 . , ...~ ~... Lappi, S., Harkonen, P., and Jalkanen,M. (1991) Cell Reg. 2 , l - 1 1 Leppa, S., Mali, M., Miettinen, H. M., and Jalkanen, M. (1992) Proc. Natl. A d . Sci. U.S. A. 89,932-936 Mali, M., Jaakkola, P., Arvilommi, A.-M., and Jalkanen, M. (1990) J. Biol. C k m . 265,6884-6889 Nugent, M. A., and Edelman,E. R. (1992) Biochemistry 31,8876-8883 Nurcombe, V., Ford, M. D., Wildschut, J. A., Bartlet, P.F. (1993) Science 2 6 0 , ~

Saksela, O., Moscatelli, D., Sommer, A., and Rifkin, D.B. (1988) J. Cell Biol.

~~

Olwin, B. B., and Rapraeger, A. (1992) J. Cell Biol. 118,621-639 Ornitz, D. M., Ya on, A., Flanagan, J. G., Svahn, C. M., Levi, E., and Leder, P. (1992) Mol. %lL BWL 12,240-247 Rapraeger, A. C., Krufka, A., and Olwin, B. B. (1991) Science 262,1705-1708 Ruoslahti, E., and Yamaguchi, Y. (1991) Cell 64.867-869 Saarni, H. (1977) BWckm. Phnrmocol.27,1029-1032 Saiki, R. K.,Gelfand, D. H., Stoffel, S., Scharf, S. J., Higuchi, R., Horn, G. T., Mullis, K. B., and Erlich, H. A. (1988) ScleFe 239,487-491 Saksela, O., and Rifkin, D.B. (1990) J. Cell Bwl. 1 1 0 , 767-775

I

I

.

I

565-572

Turnbull J. E. Fernig D. G., Ke, Y., Wilkinson, M. C., and Gallagher, J. T. (1992) >. Bioi. Chem.' 267,10337-10341 Vainio, S., Lehtonen, E., Jalkanen, M., Bernfield, M., and Saxen, L. (1989) Deu. Bwl. 134,382-391 Vainio, S., and Thesleff, I. (1992) Differentiation 50.97-105 Vogel, A., Raines, E., Kariya, B., Rivest, M.-J., and Ross, R. (1978) Pmc. Natl. Acad. S C ~U. . S. A. 76,2810-2814 Wilkinson, D. G., Bhatt, S., and McMahon, A. P. (1989) Development 1 0 6 , 131-136

Wright, T. C., Jr., Pukac, L. A., Castellot, J. J., Jr., Karnovsky, M. J., Levine, R. A,, Kim-Park, H.-Y., and Campisi, J. (1989) Proc. NQtl.A d . Sci. U. S. A. 86,3199-3203

Yamaguchi. Y., Mann, D. M., and Ruoslahti, E.(1990) Nature 346,281-284 Yayon, A,, Klagsbrun, M., Esko, J. D., Leder, P., and Ornitz, D.M. (1991) Cell 64,841-848