Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 7-Day Trial for You or Your Team.

Learn More →

Thrombospondin Mediates Calcium Mobilization in Fibroblasts via Its Arg-Gly-Asp and Carboxyl-terminal Domains

Thrombospondin Mediates Calcium Mobilization in Fibroblasts via Its Arg-Gly-Asp and... THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 270, No. 40, Issue of October 6, pp. 23747–23753, 1995 © 1995 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in U.S.A. Thrombospondin Mediates Calcium Mobilization in Fibroblasts via Its Arg-Gly-Asp and Carboxyl-terminal Domains* (Received for publication, February 17, 1995, and in revised form, July 18, 1995) Peter W. Tsao and Shaker A. Mousa‡ From the DuPont Merck Pharmaceutical Company, Cardiovascular Diseases Division, Wilmington, Delaware 19880-0400 Recently there has been major interest in the integrin avb3 Thrombospondin is a matrix glycoprotein found in various cells that can modulate cell attachment, migra- because of its role in angiogenesis and apoptosis (20, 21). An tion, and proliferation. We now show that intact soluble initial step in these processes often involves the binding of a thrombospondin causes a transient [Ca ] increase in matrix protein to avb3. There has been extensive work on the IMR-90 fibroblasts. This [Ca ] increase is mediated role of TSP on cell attachment involving avb3; however, little is partly by the RGD-containing domain of throm- known about the signal transduction events (4, 7, 18). Increas- bospondin that binds to the integrin avb3 as demon- ing evidence from various groups clearly shows that the role of strated by inhibitor studies using anti-avb3 antibody integrins extends beyond the binding of ligands or matrix pro- and RGD-containing peptides. A non-RGD and non-avb3 teins (22–24). The binding of various matrix proteins to their component of this [Ca ] increase is mediated by the receptors, especially integrins, often triggers different signals carboxyl-terminal domain of thrombospondin through such as tyrosine phosphorylation and changes in intracellular an unidentified receptor on fibroblasts as shown by the pH or Ca (25–27). These outside-in signaling events serve as antibody to the carboxyl-terminal of thrombospondin, important regulatory mechanisms to mediate subsequent mo- C6.7. In addition, the carboxyl-terminal derived peptide, 21 lecular events such as modulation of integrin affinity (28, 29) or RFYVVMWK, also triggers [Ca ] increase in ;35% of 21 21 gene transcription (30). fibroblasts. Both EGTA and Ni block the entire [Ca ] In this study, we investigated the role of TSP in triggering increase indicating that this is due to an influx of extra- Ca mobilization in a lung fibroblast cell line, IMR-90. In cellular Ca . B6H12, an antibody to the integrin-asso- addition, we also used known inhibitors of both the cell binding ciated protein, blocks this [Ca ] increase by 50%, sug- domains and the receptors of TSP to elucidate the structural gesting that some of the Ca might be entering through basis of TSP-induced transmembrane signaling. The findings an integrin-associated calcium channel. The current suggest that intact TSP can mediate Ca mobilization via findings demonstrate that multiple domains on throm- bospondin can trigger signal transduction events by in- both the RGD and carboxyl-terminal domains. creasing [Ca ] through their interactions with differ- EXPERIMENTAL PROCEDURES ent cell receptors. Materials—Human platelet-derived thrombospondin, TSP-1, was from Haematologic Technologies Inc. (Essex Junction, VT). Fura-2/AM ester was from Molecular Probes. Echistatin and the RFYVVMWK Thrombospondin (TSP) is a 450-kDa trimeric matrix glyco- peptide were purchased from Bachem (King of Prussia, PA). Peptides protein. It was first described as an a-granular protein secreted GRGDSP, GRGESP, and CSVTCG were synthesized by Dr. Ram by platelets upon thrombin activation. TSP is also synthesized Seetharam of the DuPont Merck Pharmaceutical Co. Monoclonal anti- bodies used in this study include anti-avb3 (LM609, Chemicon, CA), by fibroblasts, endothelial cells, smooth muscle cells, mono- anti-avb5 (P1F6, Becton Dickinson), and anti-CD36 (FA6.152, Immu- cytes, macrophages, osteoblasts, and tumor cells (1–6). TSP notech, ME). Anti-TSP antibodies, C6.7 and A4.1, were from Life Tech- has been shown to modulate cell attachment, migration, and nologies, Inc. Anti-integrin-associated protein antibody, B6H12, was a proliferation and has an important role in wound healing, gift from Dr. Eric Brown (Washington University, St. Louis) (31). All hemostasis, and angiogenesis (2, 3, 5, 7–10). Like other matrix antibodies were purified IgG protein except C6.7 and A4.1 which were proteins such as fibrinogen and fibronectin, TSP possesses in ascites form. Lung fibroblast cell line IMR-90 (American Type Cul- ture Collection) was grown in Dulbecco’s modified Eagle’s medium with multiple cell binding domains (6, 11). Four known cell binding 10% fetal bovine serum (Life Technologies, Inc.). To minimize the effect domains have been described for thrombospondin: (i) the hep- of cell passage, cells were used between passages 12 and 15 and were in arin binding domain at the amino-terminal (12), (ii) the type I culture 4–6 days before each experiment. repeats containing the CSVTCG sequence (13), (iii) the Arg- Cell Surface Expression of Thrombospondin Receptors—Expression Gly-Asp (RGD) sequence in the type III calcium binding repeat of thrombospondin receptors on IMR-90 was determined by indirect (7), and (iv) the carboxyl-terminal cell binding domain (14, 15). immunofluorescence and analyzed by flow cytometry. IMR-90 were removed from culture flasks using trypsin/EDTA. Cells were washed Each of these domains was found to bind to distinct cell surface and resuspended in Ca -containing phosphate-buffered saline. They receptors: (i) heparan sulfate proteoglycans (16), (ii) CD36 or were then incubated with saturating amounts of primary antibodies GPIV (17), (iii) integrin avb3 (7, 18), and (iv) the 105/80-kDa followed by phycoerythrin-conjugated goat anti-mouse IgG F(ab9) receptor (4) or the 52-kDa receptor (19), respectively. (Chemicon). Labeled cells were analyzed on a FACScan flow cytometer (Becton Dickinson). Indirect Immunofluorescence Microscopy—IMR-90 fibroblasts were * The costs of publication of this article were defrayed in part by the cultured in chamber slides using the same conditions described above. payment of page charges. This article must therefore be hereby marked For immunofluorescence staining, cells were washed with phosphate- “advertisement” in accordance with 18 U.S.C. Section 1734 solely to buffered saline with 0.15% bovine serum albumin and fixed in 4% indicate this fact. paraformaldehyde/phosphate-buffered saline for 15 min. The slides ‡ To whom correspondence should be addressed: The DuPont Merck were incubated overnight at 4 °C with the primary antibody, LM609 (2 Pharmaceutical Co., The Experimental Station, E400/3456, Wilming- mg/ml), followed by a 1:400 dilution of Cy3-conjugated goat anti-mouse ton, DE 19880-0400. Tel.: 302-695-8418; Fax: 302-695-4083. The abbreviations used are: TSP, thrombospondin; IAP, integrin- IgG (Jackson ImmunoResearch Laboratories, West Grove, PA) for2hat associated protein. 4 °C. To induce integrin clustering, fibroblasts were pretreated with This is an Open Access article under the CC BY license. 21 23748 Thrombospondin Mediates Ca Mobilization in Fibroblasts TSP for 1–2 min and fixed immediately with paraformaldehyde fol- lowed by antibody staining. Cells were visualized using a 633/N.A. 1.4 oil immersion objective on a Zeiss Axiovert microscope (Carl Zeiss). Fluorescence images were collected through a dichroic mirror and 580 nm long pass filters onto a CCD camera (Zeiss ZVS-47DEC). Images were stored and processed on a Macintosh workstation using Oncor Image software (Oncor Imaging, Gaithersburg, MD) and Adobe Photoshop v.3. Intracellular Calcium [Ca ] Measurements—Intracellular calcium in single cells was measured using Fura-2 ratio imaging. Briefly, IMR-90 fibroblasts were grown on fibronectin-coated glass coverslips and loaded with 2 mM Fura/AM ester for1hat37 °C. Cells were then washed three times with buffer containing 130 mM NaCl, 5 mM KCl, 1 mM MgCl , 1.8 mM CaCl ,10mM glucose, and 20 mM HEPES, at pH 7.4. 2 2 The coverslip was mounted on the stage of a Nikon Diaphot microscope using an open coverslip chamber (Warner Instruments, Hamden, CT), and the cells (;10–20 cells per field) were visualized using a Nikon Fluor 403/N.A. 1.3 oil immersion objective. Dual excitation was pro- vided from a 150-watt xenon arc lamp through a Ludl filter wheel system with band pass filters at 340 and 380 nm. Epifluorescence images were captured by an intensified CCD camera (QX-100, Quan- tex), and emitted light was collected through a dichroic mirror and a 520 nm long pass filter. The system was operated by a Dell 386 Dimen- sion computer using the Image1/FL software (Universal Imaging Corp, West Chester, PA) for both data acquisition and analysis. The 340/380 FIG.1. Thrombospondin-induced transient increase in [Ca ] ratio was calculated and exported to a Microsoft Excel spreadsheet, and in IMR-90 lung fibroblasts. TSP-induced increase in [Ca ] in a intracellular calcium was estimated using the equation: concentration-dependent manner. Representative [Ca ] tracings of single cell recordings from IMR-90. R 2 R S min f2 [Ca ] 5 K 3 (Eq. 1) i d R 2 R S max b2 R was determined by 2–4 mM ionomycin (Calbiochem) and R by 10 max min mM EGTA. S and S were determined by procedures published by f2 b2 Grynkiewicz et al. (32). All single-cell fluorescence recordings were done at room temperature with background subtracted. Background was determined using a field on the coverslip with no cells. For each exper- iment, the responses of 5–10 individual cells were averaged. In inhibi- tion experiments, peptides or antibodies were preincubated with cells for 2–5 min before addition of TSP. Increases in [Ca ] due to TSP varied among different cell preparations in the range of 91–569 nM.To normalize for this interexperimental difference, % control or % inhibi- tion was calculated from the TSP response of each experiment. Cell morphology was monitored by phase contrast microscopy. None of the antagonists or antibodies used in these experiments significantly af- fected the morphology of these cells within the time period of the experiment. RESULTS Addition of TSP to IMR-90 cells caused a concentration-de- pendent transient rise in [Ca ] ranging from 55-;1000 nM in .90% of the cells (Fig. 1). This [Ca ] increase peaked within 10–30 s and returned to resting levels after 60–180 s depend- ing on the concentration of TSP (Fig. 1). With 0.08 mM TSP, the mean increase in [Ca ] was 220 6 34 nM (n 5 20) above the resting level of 93 6 8nM (n 5 18) (Fig. 2A). This increase in [Ca ] was eliminated completely in the absence of extracel- lular Ca by the addition of 5 mM EGTA (Fig. 2B). It is possible that the blocking of the [Ca ] increase is due to EGTA’s effect on Ca -dependent ligand binding to TSP recep- tors rather than depletion of the source of Ca available for transmembrane fluxes (4, 33). Chelation of Ca has been shown to dissociate the integrin heterodimer and alter the FIG.2. Thrombospondin-induced transient increase in [Ca ] tertiary structure of TSP (7, 18, 34, 35). This, in turn, could in IMR-90 is dependent on extracellular Ca . Representative indirectly affect ligand binding and thus the ability to trigger 21 [Ca ] tracings of recordings from cells stimulated with 0.08 mM TSP 21 21 an increase in [Ca ] . To verify that the attenuated Ca was (A) and pretreatment with 5 mM EGTA (B)or10mM Ni (C) before TSP addition. Note the inhibition of this increase in [Ca ] by EGTA or due to inhibition of Ca influx, cells were pretreated with 21 21 Ni pretreatment. Representative tracings of single cell recordings Ni , which is known to block Ca fluxes through many types from at least four separate experiments. Assay was carried out as 21 21 21 of Ca channels. Ni at 10 mM abolished the rise in [Ca ] described under “Experimental Procedures.” induced by TSP similar to that of EGTA (Fig. 2C). Neither 21 21 EGTA nor Ni affected the baseline [Ca ] under these con- ditions. These results indicate that the increase in [Ca ] from and multiple cell surface receptors. To address the issue of TSP is dependent on extracellular Ca fluxes into the cells. structural specificity, experiments were carried out to study Role of av Integrins in Mediating TSP-induced [Ca ] In- the effects of known inhibitors on these cell binding domains crease—TSP is known to have multiple cell binding domains and receptors. One of the major cell surface receptors for TSP 21 Thrombospondin Mediates Ca Mobilization in Fibroblasts 23749 FIG.3. Cell surface expression of thrombospondin receptors in IMR-90 fibroblasts by flow cytometry. Cells were gated by for- ward and side light scatter, and 5,000 gated events were collected. IMR-90 expressed both avb3 and avb5 but not CD36 on their cell surface. Assay was carried out as described under “Experimental Procedures.” belongs to the av integrin family (6, 7). Previous studies have shown that other soluble ligands to the av integrin were able to raise [Ca ] (36, 37). The IMR-90 cells express both avb3 and avb5 on their cell surface as evident from the flow cytometry data (Fig. 3). To investigate the possibility that the TSP-in- duced increase in [Ca ] was mediated via the av integrin, the effect of an antibody to avb3 (LM609) was studied (38). LM609, FIG.4. [Ca ] tracings showing the effect of various anti-av antibodies on the thrombospondin-induced rise in [Ca ] . Cells at 4 and 10 mg/ml, inhibited TSP-induced increases in [Ca ] were pretreated with 4 mg/ml LM609 (anti-avb3) (A), 4 mg/ml B6H12 by 54 and 63%, respectively (Figs. 4A and 5). (anti-IAP) (B), or 10 mg/ml P1F6 (anti-avb5) (C) before addition of 0.08 Since this Ca increase was apparently due to a calcium mM TSP. Representative tracings of single cell recordings from at least influx, we investigated further whether it was via the integrin- four separate experiments. Note the partial inhibition of TSP-induced increases in [Ca ] with LM609 and B6H12 but not with P1F6. Assay associated protein (IAP) which has been shown to mediate i 21 was carried out as described under “Experimental Procedures.” integrin-regulated Ca flux into cells (39). We tested the effect of a recently described antibody to IAP, B6H12 (31), on TSP- induced increases in [Ca ] . B6H12, at 4 and 10 mg/ml, inhib- ited TSP-induced increases in [Ca ] by 42% and 50%, respec- tively (Figs. 4B and 5). On the other hand, the TSP-induced increase in [Ca ] was not inhibited by the control antibody to avb5 (Figs. 4C and 5). In addition, LM609 and B6H12 do not block non-TSP-, i.e. thrombin, mediated increases in [Ca ] , which indicates they specifically block avb3 integrin-mediated 21 2 Ca flux. None of these monoclonal antibodies raised the baseline [Ca ] at the concentrations tested. Earlier studies have reported that integrin clustering is very important in signal transduction (25, 40, 41). Here we investi- gated the effect of TSP and inhibitors of the TSP-avb3 inter- action on integrin clustering. As shown in Fig. 6, TSP induces significant integrin clustering in IMR-90. This is not seen in cells that were not exposed to TSP or those that were pre- treated with LM609, suggesting a specific interaction between the integrin and the matrix protein. Role of the RGD Domain in TSP-induced[Ca ] Increase— Many of the integrins, including avb3, recognize the RGD FIG.5. Effect of anti-av antibodies on the thrombospondin- tripeptide sequence as a binding motif. The RGD domain is induced rise in [Ca ] . Antibodies LM609 and B6H12 (4 and 10 found in the type III calcium binding repeats and has been mg/ml) partially inhibited TSP-induced increases in [Ca ] . P1F6 at a similar concentration has minimal effect. Mean 6 S.E. (n 5 4). Assay shown to support cell attachment. Two RGD-containing pep- was carried out as described under “Experimental Procedures.” tides, GRGDSP and echistatin, were used to test the RGD dependence of the TSP-induced increase in [Ca ] . At 0.8 mM concentration of GRGESP only inhibited 7% (Fig. 7, A and B). GRGDSP, the increase is inhibited by 48%, whereas the same Echistatin, which is known to be a more potent integrin antag- onist (42), inhibited the increase in [Ca ] by 72% at 0.8 mM Data not shown. (Fig. 7C). Both GRGDSP and echistatin inhibit the TSP-in- 21 23750 Thrombospondin Mediates Ca Mobilization in Fibroblasts FIG.7. Effect of RGD-containing peptides on the throm- bospondin-induced rise in [Ca ] . Cells were pretreated with 0.8 mM GRGDSP (A), 0.8 mM GRGESP (B), or 0.8 mM echistatin (C) before the addition of 0.08 mM TSP. Note the partial inhibition of TSP-induced increases in [Ca ] with GRGDSP and echistatin but not with the control inactive peptide GRGESP. However, GRGDSP at this concen- tration also increases [Ca ] . Representative tracings of single cell recordings from at least four separate experiments. Assay was carried out as described under “Experimental Procedures.” [Ca ] signal when TSP is added. However, this is unlikely since subsequent stimulation with other agonists, such as thrombin, was able to elicit another transient increase in 21 2 [Ca ] . This response to GRGDSP was RGD-specific as the control peptide, GRGESP, did not induce an increase in [Ca ] . Also, echistatin, by itself, demonstrated a similar increase in [Ca ] in 50% of the cells at 5 mM, but no effect was seen at 1 mM. Role of Non-RGD Cell Binding Domains in [Ca ] In- crease—Since neither anti-avb3 nor RGD-containing peptides FIG.6. Immunolocalization of avb3 integrin on IMR-90 using can completely inhibit TSP-induced rises in [Ca ] , it is pos- indirect fluorescence. Fluorescence images showing very diffuse dis- tribution of avb3 integrin on the surface of IMR-90 (A). With pretreat- sible that there may be other components that mediate the ment of 0.17 mM TSP for 2 min (B), cells showed significant integrin 21 Ca signaling in addition to the integrin avb3. To identify the clustering as indicated by the white arrows. A 2-min preincubation with structural basis for this non-avb3 component of the intracellu- 20 mg/ml LM609, which prevents the interaction of avb3 with TSP, lar signaling, a series of experiments was carried out using showed significantly less clustering (C). Cells incubated with isotype match antibody did not stain (not shown). The distribution of avb3 was antibodies and peptides directed to either the different cell observed using LM609 followed by Cy3-conjugated goat anti-mouse IgG binding domains or the receptor(s) of TSP. as described under “Experimental Procedures” (magnification, 3 1000). One of the potential candidates is CD36, which is the recep- tor for the type I repeat of TSP (17). This surface glycoprotein duced [Ca ] increase in a concentration-dependent manner CD36 has been shown to mediate signal transduction in plate- (Fig. 8). lets by increasing [Ca ] (43). However, in this system, it is 21 21 In addition to blocking the TSP-induced increase in [Ca ] , unlikely that CD36 mediates the increase in [Ca ] as IMR-90 i i GRGDSP by itself, at concentration .0.8 mM, caused a tran- do not express CD36 on their cell surface (Fig. 3). In addition, sient increase in [Ca ] in .90% of the cells. It is possible that the peptide CSVTCG, which corresponds to the CD36 binding the depletion of available Ca due to this GRGDSP-induced motif on TSP and has been shown to inhibit cell adhesion (13), 21 21 increase in [Ca ] might be responsible for the attenuation of did not have any effect on TSP-induced rises in [Ca ] (Fig. 9). i i 21 Thrombospondin Mediates Ca Mobilization in Fibroblasts 23751 FIG. 10. [Ca ] tracing showing the effect of thrombospondin carboxyl-terminal peptide on [Ca ] in IMR-90 cells. The carbox- yl-terminal peptide, RFYVVMWK (100 mM), increases [Ca ] in IMR- 90. Representative tracings of single cell recordings from at least four separate experiments. Assay was carried out as described under “Ex- perimental Procedures.” domain, the TSP-induced increase in [Ca ] is inhibited by FIG.8. Dose-response relationship of the inhibitory effect of 50%. This suggests that part of the Ca signal generated by RGD-containing peptides. The effect of echistatin (l), GRGDSP (E), TSP may be due to the binding of the carboxyl-terminal domain or GRGESP (3) on TSP-induced increases in [Ca ] . Echistatin showed to its receptor (Fig. 9). This effect is apparently specific to the more potent inhibition of TSP-induced increases in [Ca ] than carboxyl-terminal binding domain since antibody A4.1 (10, 44), GRGDSP with a shift of 3–4 log concentrations to the left. Each data directed against the amino-terminal half of the central stalk point is the mean of at least four experiments with S.E. # 10%. Assay was carried out as described under “Experimental Procedures.” region of TSP, did not block the [Ca ] increase. Nonspecific antibody did not have any significant effect on [Ca ] increase. Interestingly, when both LM609 (10 mg/ml) and C6.7 (1:100) were used together, no synergistic or additive effect was seen. It is unclear, at this point, if this is due to a lack of cooperativity of the two sites. Nevertheless, more direct evidence of the role of the carboxyl-terminal of TSP came from studies with the carboxyl-terminal domain peptide RFYVVMWK (45). This pep- tide at 100 mM triggered a transient increase in [Ca ] that reached $2 mM in 35% of the cells (Fig. 10). This suggests that the carboxyl-terminal cell binding domain may mediate part of the TSP-induced rise in [Ca ] . Altogether, these data indicate that intact TSP can mediate an increase in [Ca ] in IMR-90 fibroblasts via at least two pathways: the RGD domain binding to the avb3 integrin and the carboxyl-terminal domain possibly binding to an unidentified receptor. DISCUSSION In this study we have shown that TSP, a matrix protein which binds to multiple cell surface receptors, increases [Ca ] in IMR-90 fibroblasts. This increase, which is partially medi- FIG.9. Effect of inhibitors against the non-RGD domains of ated by the RGD domain of TSP binding to the integrin avb3, thrombospondin on thrombospondin-induced rise in [Ca ] . is not limited to IMR-90, as we have also observed this in Antibodies to TSP cell binding domains, C6.7 (1:100), A4.1 (1:100), endothelial cells. We have also shown that known inhibitors of heparin (100 mg/ml), and control mouse IgG (50 mg/ml) were preincu- avb3, GRGDSP peptide and echistatin, in addition to blocking bated with TSP for 30 min before addition to cells. Peptide CSVTCG TSP-induced signaling, are capable of triggering an increase in (0.1 mM) was added directly to the IMR-90 cells. Of these inhibitors only the antibody to the carboxyl-terminal domain of TSP, C6.7, is effective [Ca ] by themselves. This suggests that RGD-containing pep- in inhibiting TSP-induced increase in [Ca ] . Assay was carried out as tides can function as both competitive antagonists and partial described under “Experimental Procedures.” agonists for their receptors. This had been reported previously 21 by several other groups; however, our threshold concentrations CSVTCG by itself had no effect on baseline [Ca ] . Likewise, are much higher than those reported elsewhere (36, 37). One the heparin binding domain of TSP also did not seem to play a 21 possible explanation is that high concentrations of peptides or role in generating this [Ca ] increase because heparin, at a antibodies could induce clustering of integrins, which could concentration (100 mg/ml) that had been shown to effectively activate the integrins and trigger transmembrane Ca flux block cell adhesion (15), did not inhibit TSP-induced rises in 21 not normally seen at the lower concentrations (26, 46). This [Ca ] (Fig. 9). clustering is seen in the present study when avb3 integrins Another possibility is the carboxyl-terminal cell binding do- interact with TSP. These findings illustrate that integrin clus- main of TSP which has been described to support non-RGD-de- pendent cell adhesion in melanoma cell lines (14, 15). When TSP was preincubated with C6.7, an antibody against this P. W. Tsao and S. A. Mousa, unpublished observations. 21 23752 Thrombospondin Mediates Ca Mobilization in Fibroblasts tering, presumably caused by multimeric interactions between With intact TSP, only 1 mM is required to trigger similar in- the ligand and its receptors, is important in transmembrane creases in [Ca ] . The main reason for this disparity in thresh- signaling (41). This is further supported by the observation old concentrations is the multivalent and higher affinity bind- ing of the multimeric matrix protein TSP versus the that RGD peptide can produce an increase in [Ca ] (.1mM) in kidney epithelial cells when coupled to beads but not when it monovalent and lower affinity binding of the peptide fragments (41, 50). In addition to having high affinity receptors, multim- is in the soluble monomeric form (29). In addition, earlier b2 integrin (CD11b/CD18) demonstrate that an- eric ligands are more effective in inducing receptor clustering studies with as shown in the present study; together, these have been shown tibodies to the a and b can induce Ca signal only if they are to work synergistically in transmembrane signaling (41). cross-linked (40). In addition, TSP might potentially mediate signaling via Another confounding factor is that this ligand-induced inte- other receptors. One such TSP receptor is CD36 which was not grin signaling may be dependent on the type of cell preparation studied here due to the lack of expression on IMR-90 fibro- being used. Sjaastad et al. (29) have reported that soluble RGD blasts. There is evidence from the literature that CD36 is (;0.3 mM) was not able to produce an increase in [Ca ] in capable of mediating signal transduction by increasing [Ca ] kidney epithelial cells, whereas two other groups both observed i in platelets and triggering oxidative bursts in monocytes (43, an increase in osteoclasts at a similar concentration (36, 37). 51). However, these two studies utilized monoclonal antibodies Even within the same cell type the data are not consistent. In against CD36 to trigger the signaling events, and, therefore, it that regard, echistatin was shown to induce increases in is uncertain if intact TSP or its fragments will produce similar [Ca ] in rat osteoclasts by one group but not by the other results via CD36. group (36, 37). This discrepancy indicates the complexity of In summary, we have demonstrated that TSP can trigger ligand/integrin-mediated signaling. transient increases in [Ca ] in IMR-90 fibroblasts by increas- Ligand-induced signaling through integrins does not follow ing Ca influx. The response is inhibitable by specific TSP the paradigm for traditional signaling receptors. In general, receptor antibodies indicating that this is a receptor-mediated integrins differ from the known signaling receptors in their effect. While we have shown this to occur primarily via the short cytoplasmic domain, lack of kinase activity, and lack of RGD and carboxyl-terminal domains of TSP and their corre- direct interaction with G-proteins (23). In this study we have 21 sponding receptors, we cannot rule out the possibility of the shown that the [Ca ] signal triggered by the soluble matrix other TSP receptors as signaling molecules in other cellular protein, TSP, via the integrin avb3, is dependent on extracel- 21 systems. The data from the present study, together with pre- lular calcium. The present data using EGTA or Ni indicates vious reports, indicate that TSP can support cell attachment that the increase in [Ca ] is due primarily to an influx of 21 21 and signaling through its different cell binding domains (4, 7, extracellular Ca , possibly through some type of Ca chan- 13, 14, 18, 45). These observations confirm that matrix pro- nel. Additional evidence supporting this hypothesis was dem- teins, with their multiple cell binding domains, can interact onstrated by our data with the antibody to IAP, B6H12, which with corresponding cell surface receptors to transduce signals was shown to block ;50% of [Ca ] increase. IAP, which has across the cell membrane. These outside-in signaling events been shown to be closely associated with avb3 (31), was re- will be important in understanding the role TSP plays in cell cently shown by Schwartz et al. (39) to be required specifically attachment, migration, and proliferation. for integrin-regulated Ca influx. This suggests that IAP may function as an integrin-associated calcium channel. Moreover, Acknowledgments—We thank Dr. Robert C. Newton for critically reading the manuscript, Dr. Ram Seetharam for peptide synthesis, Dr. purified integrin aIIbb3 was reported to function as a calcium Eric J. Brown for his gift of B6H12, Pamela Kidd for her expert tech- channel when reconstituted into liposomes (47). This suggests nical assistance, and William Lorelli for his editorial assistance. that the integrin itself can act as a conduit for transmembrane 21 21 Ca flux that is distinct from the classical Ca channels. REFERENCES Taken together, it is possible that the avb3 integrin or its 1. Clezardin, P., Jouishomme, H., Chavassieux, P., and Marie, P. J. (1989) Eur. J. Biochem. 181, 721–726 associated proteins, upon ligand binding, can be activated and 2. Dameron, K. M., Volpert, O. V., Tainsky, M. A., and Bouck, N. (1994) function as a Ca channel for signal transduction. Science 265, 1582–1584 3. Taraboletti, G., Roberts, D., Liotta, L. A., and Giavazzi, R. (1990) J. Cell Biol. The present data also show that there are other components 111, 765–772 of the transmembrane Ca flux that are non-RGD and non- 4. Yabkowitz, R., and Dixit, V. M. (1991) Cancer Res. 51, 3648–3656 avb3-dependent. This notion comes from the inability of the 5. Majack, R. A., Goodman, L. V., and Dixit, V. M. (1988) J. Cell Biol. 106, 415–422 synthetic RGD-peptide and LM609 to completely inhibit TSP- 6. Frazier, W. A. (1991) Curr. Opin. Cell Biol. 3, 792–799 induced rises in [Ca ] and is verified by the observation that 7. Lawler, J., Weinstein, R., and Hynes, R. O. (1988) J. Cell Biol. 107, 2351–2361 8. Nicosia, R. F., and Tuszynski, G. P. (1994) J. Cell Biol. 124, 183–193 an antibody to carboxyl-terminal domain, C6.7, is also able to 9. Savill, J., Hogg, N., Ren, Y., and Haslett, C. (1992) J. Clin. Invest. 90, inhibit 50% of the Ca signal. The carboxyl-terminal domain 1513–1522 10. Good, D. J., Polverini, P. J., Rastinejad, F., Le Beau, M. M., Lemons, R. S., of TSP has been reported to support chemotaxis and cell mi- Frazier, W. A., and Bouck, N. P. (1990) Proc. Natl. Acad. Sci. U. S. A. 87, gration (48, 49). Further studies using the carboxyl-terminal 6624–6628 domain peptide, RFYVVMWK, indicate that the carboxyl-ter- 11. Yamada, K. M. (1991) J. Biol. Chem. 266, 12809–12812 12. Dardik, R., and Lahav, J. (1987) Eur. J. Biochem. 168, 347–355 minal domain of TSP is also capable of triggering Ca in- 13. Tuszynski, G. P., Rothman, V. L., Deutch, A. H., Hamilton, B. K., and Eyal, J. creases by itself. Since this Ca increase is observed only in a (1992) J. Cell Biol. 116, 209–217 14. Kosfeld, M. D., Pavlopoulos, T. V., and Frazier, W. A. (1991) J. Biol. Chem. subpopulation of cells, this receptor might not be present on all 266, 24257–24259 cells. Two possible receptors for the carboxyl-terminal binding 15. Kosfeld, M. D., and Frazier, W. A. (1992) J. Biol. Chem. 267, 16230–16236 16. Roberts, D. D. (1988) Cancer Res. 48, 6785–6793 domain have so far been identified, an 80/105-kDa complex on 17. Asch, A. S., Barnwell, J., Silverstein, R. L., and Nachman, R. L. (1987) J. Clin. carcinoma cells (4) and a 52-kDa domain on K-562 cells (19). Invest. 79, 1054–1061 Even if either of these receptors exists in the IMR-90, it is 18. Sun, X., Skorstengaard, K., and Mosher, D. F. (1992) J. Cell Biol. 118, 693–701 19. Gao, A.-G., and Frazier, W. A. (1994) J. Biol. Chem. 269, 29650–29657 unclear if they are responsible for the binding of the carboxyl- 20. Brooks, P. C., Montgomery, A. M. P., Rosenfield, M., Reisfeld, R. A., Hu, T., terminal TSP fragment and the subsequent transduction of the Klier, G., and Cheresh, D. A. (1994) Cell 79, 1157–1164 21. Montgomery, A. M. P., Reisfeld, R. A., and Cheresh, D. A. (1994) Proc. Natl. signal across the cell membrane. The carboxyl-terminal peptide Acad. Sci. U. S. A. 91, 8856–8860 and GRGDSP trigger increases in [Ca ] in the millimolar 22. Hynes, R. O. (1992) Cell 69, 11–25 range at peptide concentrations of 0.1 and 0.8 mM, respectively. 23. Juliano, R. L., and Haskill, S. (1993) J. Cell Biol. 120, 577–585 21 Thrombospondin Mediates Ca Mobilization in Fibroblasts 23753 24. Clark, E. A., Shattil, S. J., and Brugge, J. S. (1994) Trends Biol. Sci. 19, G. A. (1994) Am. J. Physiol. 266, C376–C381 464–469 38. Cheresh, D. A. (1987) Proc. Natl. Acad. Sci. U. S. A. 84, 6471–6475 25. Kornberg, L. J., Earp, H. S., Turner, C. E., Prockop, C., and Juliano, R. L. 39. Schwartz, M., Brown, E. J., and Fazeli, B. (1993) J. Biol. Chem. 268, (1991) Proc. Natl. Acad. Sci. U. S. A. 88, 8392–8396 19931–19934 26. Schwartz, M. A., Ingber, D. E., Lawrence, M., Springer, T. A., and Lechene, C. 40. Ng-Sikorski, J., Andersson, R., Patarroyo, M., and Andersson, T. (1991) (1991) Exp. Cell. Res. 195, 533–535 Exp. Cell Res. 195, 504–508 27. Schwartz, M. A. (1993) J. Cell Biol. 120, 1003–1010 41. Miyamoto, S., Akiyama, S. K., and Yamada, K. M. (1995) Science 267, 883–885 28. O’Toole, T. E., Mandelman, D., Forsyth, J., Shattil, S. J., Plow, E. F., and 42. Gan, Z.-R., Gould, R. J., Jacobs, J. W., Friedman, P. A., and Polokoff, M. A. Ginsberg, M. H. (1991) Science 254, 845–847 (1988) J. Biol. Chem. 263, 19827–19832 29. Sjaastad, M. D., Angres, B., Lewis, R. S., and Nelson, W. J. (1994) Proc. Natl. 43. Alessio, M., Greco, N. J., Primo, L., Ghigo, D., Bosia, A., Tandon, N. N., Acad. Sci. U. S. A. 91, 8214–8218 Ockenhouse, C. F., Jamieson, G. A., and Malavasi, F. (1993) Blood 82, 30. Lin, T. H., Yurochko, A., Kornberg, L., Morris, J., Walker, J. J., Haskill, S., and 3637–3647 Juliano, R. L. (1994) J. Cell Biol. 126, 1585–1593 44. Prater, C. A., Plotkin, J., Jaye, D., and Frazier, W. A. (1991) J. Cell Biol. 112, 31. Brown, E., Hooper, L., Ho, T., and Gresham, H. (1990) J. Cell Biol. 111, 1031–1040 2785–2794 45. Kosfeld, M. D., and Frazier, W. A. (1993) J. Biol. Chem. 268, 8808–8814 32. Grynkiewicz, G., Poenie, M., and Tsien, R. Y. (1985) J. Biol. Chem. 260, 46. Ingber, D. E., Prusty, D., Frangioni, J. J., Cragoe, E. J., Jr., Lechene, C. P., and 3440–3450 Schwartz, M. A. (1990) J. Cell Biol. 110, 1803–1811 33. Smith, J. W., Piotrowicz, R. S., and Mathis, D. (1994) J. Biol. Chem. 269, 47. Rybak, M. E. M., and Renzulli, L. A. (1989) J. Biol. Chem. 264, 14617–14620 960–967 48. Yabkowitz, R., Mansfield, P. J., Dixit, V. M., and Suchard, S. J. (1993) Cancer 34. Steiner, B., Cousot, D., Trzeciak, A., Gillessen, D., and Hadva´ry, P. (1989) Res. 53, 378–387 J. Biol. Chem. 264, 13102–13108 49. Yabkowitz, R., Mansfield, P. J., Ryan, U. S., and Suchard, S. J. (1993) J. Cell. 35. Tsao, P. W., Bozarth, J. M., Jackson, S. A., Forsythe, M. S., Flint, S. K., and Physiol. 157, 24 Mousa, S. A. (1995) Thromb. Res. 77, 543–556 36. Shankar, G., Davison, I., Helfrich, M. H., Mason, W. T., and Horton, M. A. 50. Welder, C. A., Lee, D. H. S., and Takei, F. (1993) J. Immunol. 150, 2203–2210 (1993) J. Cell Sci. 105, 61–68 51. Schu¨epp, B. J., Pfister, H., Clemetson, K. J., Silverstein, R. L., and Jungi, T. W. 37. Zimolo, Z., Wesolowski, G., Tanaka, H., Hyman, J. L., Hoyer, J. R., and Rodan, (1991) Biochem. Biophys. Res. Commun. 175, 263–270 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Biological Chemistry Unpaywall

Thrombospondin Mediates Calcium Mobilization in Fibroblasts via Its Arg-Gly-Asp and Carboxyl-terminal Domains

Tsao, Peter W.; Mousa, Shaker A.
Journal of Biological Chemistry – Oct 1, 1995
Download PDF
7 pages

Loading next page...
 
/lp/unpaywall/thrombospondin-mediates-calcium-mobilization-in-fibroblasts-via-its-MpQlhqJUZo

References

References for this paper are not available at this time. We will be adding them shortly, thank you for your patience.

Publisher
Unpaywall
ISSN
0021-9258
DOI
10.1074/jbc.270.40.23747
Publisher site
See Article on Publisher Site

Abstract

THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 270, No. 40, Issue of October 6, pp. 23747–23753, 1995 © 1995 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in U.S.A. Thrombospondin Mediates Calcium Mobilization in Fibroblasts via Its Arg-Gly-Asp and Carboxyl-terminal Domains* (Received for publication, February 17, 1995, and in revised form, July 18, 1995) Peter W. Tsao and Shaker A. Mousa‡ From the DuPont Merck Pharmaceutical Company, Cardiovascular Diseases Division, Wilmington, Delaware 19880-0400 Recently there has been major interest in the integrin avb3 Thrombospondin is a matrix glycoprotein found in various cells that can modulate cell attachment, migra- because of its role in angiogenesis and apoptosis (20, 21). An tion, and proliferation. We now show that intact soluble initial step in these processes often involves the binding of a thrombospondin causes a transient [Ca ] increase in matrix protein to avb3. There has been extensive work on the IMR-90 fibroblasts. This [Ca ] increase is mediated role of TSP on cell attachment involving avb3; however, little is partly by the RGD-containing domain of throm- known about the signal transduction events (4, 7, 18). Increas- bospondin that binds to the integrin avb3 as demon- ing evidence from various groups clearly shows that the role of strated by inhibitor studies using anti-avb3 antibody integrins extends beyond the binding of ligands or matrix pro- and RGD-containing peptides. A non-RGD and non-avb3 teins (22–24). The binding of various matrix proteins to their component of this [Ca ] increase is mediated by the receptors, especially integrins, often triggers different signals carboxyl-terminal domain of thrombospondin through such as tyrosine phosphorylation and changes in intracellular an unidentified receptor on fibroblasts as shown by the pH or Ca (25–27). These outside-in signaling events serve as antibody to the carboxyl-terminal of thrombospondin, important regulatory mechanisms to mediate subsequent mo- C6.7. In addition, the carboxyl-terminal derived peptide, 21 lecular events such as modulation of integrin affinity (28, 29) or RFYVVMWK, also triggers [Ca ] increase in ;35% of 21 21 gene transcription (30). fibroblasts. Both EGTA and Ni block the entire [Ca ] In this study, we investigated the role of TSP in triggering increase indicating that this is due to an influx of extra- Ca mobilization in a lung fibroblast cell line, IMR-90. In cellular Ca . B6H12, an antibody to the integrin-asso- addition, we also used known inhibitors of both the cell binding ciated protein, blocks this [Ca ] increase by 50%, sug- domains and the receptors of TSP to elucidate the structural gesting that some of the Ca might be entering through basis of TSP-induced transmembrane signaling. The findings an integrin-associated calcium channel. The current suggest that intact TSP can mediate Ca mobilization via findings demonstrate that multiple domains on throm- bospondin can trigger signal transduction events by in- both the RGD and carboxyl-terminal domains. creasing [Ca ] through their interactions with differ- EXPERIMENTAL PROCEDURES ent cell receptors. Materials—Human platelet-derived thrombospondin, TSP-1, was from Haematologic Technologies Inc. (Essex Junction, VT). Fura-2/AM ester was from Molecular Probes. Echistatin and the RFYVVMWK Thrombospondin (TSP) is a 450-kDa trimeric matrix glyco- peptide were purchased from Bachem (King of Prussia, PA). Peptides protein. It was first described as an a-granular protein secreted GRGDSP, GRGESP, and CSVTCG were synthesized by Dr. Ram by platelets upon thrombin activation. TSP is also synthesized Seetharam of the DuPont Merck Pharmaceutical Co. Monoclonal anti- bodies used in this study include anti-avb3 (LM609, Chemicon, CA), by fibroblasts, endothelial cells, smooth muscle cells, mono- anti-avb5 (P1F6, Becton Dickinson), and anti-CD36 (FA6.152, Immu- cytes, macrophages, osteoblasts, and tumor cells (1–6). TSP notech, ME). Anti-TSP antibodies, C6.7 and A4.1, were from Life Tech- has been shown to modulate cell attachment, migration, and nologies, Inc. Anti-integrin-associated protein antibody, B6H12, was a proliferation and has an important role in wound healing, gift from Dr. Eric Brown (Washington University, St. Louis) (31). All hemostasis, and angiogenesis (2, 3, 5, 7–10). Like other matrix antibodies were purified IgG protein except C6.7 and A4.1 which were proteins such as fibrinogen and fibronectin, TSP possesses in ascites form. Lung fibroblast cell line IMR-90 (American Type Cul- ture Collection) was grown in Dulbecco’s modified Eagle’s medium with multiple cell binding domains (6, 11). Four known cell binding 10% fetal bovine serum (Life Technologies, Inc.). To minimize the effect domains have been described for thrombospondin: (i) the hep- of cell passage, cells were used between passages 12 and 15 and were in arin binding domain at the amino-terminal (12), (ii) the type I culture 4–6 days before each experiment. repeats containing the CSVTCG sequence (13), (iii) the Arg- Cell Surface Expression of Thrombospondin Receptors—Expression Gly-Asp (RGD) sequence in the type III calcium binding repeat of thrombospondin receptors on IMR-90 was determined by indirect (7), and (iv) the carboxyl-terminal cell binding domain (14, 15). immunofluorescence and analyzed by flow cytometry. IMR-90 were removed from culture flasks using trypsin/EDTA. Cells were washed Each of these domains was found to bind to distinct cell surface and resuspended in Ca -containing phosphate-buffered saline. They receptors: (i) heparan sulfate proteoglycans (16), (ii) CD36 or were then incubated with saturating amounts of primary antibodies GPIV (17), (iii) integrin avb3 (7, 18), and (iv) the 105/80-kDa followed by phycoerythrin-conjugated goat anti-mouse IgG F(ab9) receptor (4) or the 52-kDa receptor (19), respectively. (Chemicon). Labeled cells were analyzed on a FACScan flow cytometer (Becton Dickinson). Indirect Immunofluorescence Microscopy—IMR-90 fibroblasts were * The costs of publication of this article were defrayed in part by the cultured in chamber slides using the same conditions described above. payment of page charges. This article must therefore be hereby marked For immunofluorescence staining, cells were washed with phosphate- “advertisement” in accordance with 18 U.S.C. Section 1734 solely to buffered saline with 0.15% bovine serum albumin and fixed in 4% indicate this fact. paraformaldehyde/phosphate-buffered saline for 15 min. The slides ‡ To whom correspondence should be addressed: The DuPont Merck were incubated overnight at 4 °C with the primary antibody, LM609 (2 Pharmaceutical Co., The Experimental Station, E400/3456, Wilming- mg/ml), followed by a 1:400 dilution of Cy3-conjugated goat anti-mouse ton, DE 19880-0400. Tel.: 302-695-8418; Fax: 302-695-4083. The abbreviations used are: TSP, thrombospondin; IAP, integrin- IgG (Jackson ImmunoResearch Laboratories, West Grove, PA) for2hat associated protein. 4 °C. To induce integrin clustering, fibroblasts were pretreated with This is an Open Access article under the CC BY license. 21 23748 Thrombospondin Mediates Ca Mobilization in Fibroblasts TSP for 1–2 min and fixed immediately with paraformaldehyde fol- lowed by antibody staining. Cells were visualized using a 633/N.A. 1.4 oil immersion objective on a Zeiss Axiovert microscope (Carl Zeiss). Fluorescence images were collected through a dichroic mirror and 580 nm long pass filters onto a CCD camera (Zeiss ZVS-47DEC). Images were stored and processed on a Macintosh workstation using Oncor Image software (Oncor Imaging, Gaithersburg, MD) and Adobe Photoshop v.3. Intracellular Calcium [Ca ] Measurements—Intracellular calcium in single cells was measured using Fura-2 ratio imaging. Briefly, IMR-90 fibroblasts were grown on fibronectin-coated glass coverslips and loaded with 2 mM Fura/AM ester for1hat37 °C. Cells were then washed three times with buffer containing 130 mM NaCl, 5 mM KCl, 1 mM MgCl , 1.8 mM CaCl ,10mM glucose, and 20 mM HEPES, at pH 7.4. 2 2 The coverslip was mounted on the stage of a Nikon Diaphot microscope using an open coverslip chamber (Warner Instruments, Hamden, CT), and the cells (;10–20 cells per field) were visualized using a Nikon Fluor 403/N.A. 1.3 oil immersion objective. Dual excitation was pro- vided from a 150-watt xenon arc lamp through a Ludl filter wheel system with band pass filters at 340 and 380 nm. Epifluorescence images were captured by an intensified CCD camera (QX-100, Quan- tex), and emitted light was collected through a dichroic mirror and a 520 nm long pass filter. The system was operated by a Dell 386 Dimen- sion computer using the Image1/FL software (Universal Imaging Corp, West Chester, PA) for both data acquisition and analysis. The 340/380 FIG.1. Thrombospondin-induced transient increase in [Ca ] ratio was calculated and exported to a Microsoft Excel spreadsheet, and in IMR-90 lung fibroblasts. TSP-induced increase in [Ca ] in a intracellular calcium was estimated using the equation: concentration-dependent manner. Representative [Ca ] tracings of single cell recordings from IMR-90. R 2 R S min f2 [Ca ] 5 K 3 (Eq. 1) i d R 2 R S max b2 R was determined by 2–4 mM ionomycin (Calbiochem) and R by 10 max min mM EGTA. S and S were determined by procedures published by f2 b2 Grynkiewicz et al. (32). All single-cell fluorescence recordings were done at room temperature with background subtracted. Background was determined using a field on the coverslip with no cells. For each exper- iment, the responses of 5–10 individual cells were averaged. In inhibi- tion experiments, peptides or antibodies were preincubated with cells for 2–5 min before addition of TSP. Increases in [Ca ] due to TSP varied among different cell preparations in the range of 91–569 nM.To normalize for this interexperimental difference, % control or % inhibi- tion was calculated from the TSP response of each experiment. Cell morphology was monitored by phase contrast microscopy. None of the antagonists or antibodies used in these experiments significantly af- fected the morphology of these cells within the time period of the experiment. RESULTS Addition of TSP to IMR-90 cells caused a concentration-de- pendent transient rise in [Ca ] ranging from 55-;1000 nM in .90% of the cells (Fig. 1). This [Ca ] increase peaked within 10–30 s and returned to resting levels after 60–180 s depend- ing on the concentration of TSP (Fig. 1). With 0.08 mM TSP, the mean increase in [Ca ] was 220 6 34 nM (n 5 20) above the resting level of 93 6 8nM (n 5 18) (Fig. 2A). This increase in [Ca ] was eliminated completely in the absence of extracel- lular Ca by the addition of 5 mM EGTA (Fig. 2B). It is possible that the blocking of the [Ca ] increase is due to EGTA’s effect on Ca -dependent ligand binding to TSP recep- tors rather than depletion of the source of Ca available for transmembrane fluxes (4, 33). Chelation of Ca has been shown to dissociate the integrin heterodimer and alter the FIG.2. Thrombospondin-induced transient increase in [Ca ] tertiary structure of TSP (7, 18, 34, 35). This, in turn, could in IMR-90 is dependent on extracellular Ca . Representative indirectly affect ligand binding and thus the ability to trigger 21 [Ca ] tracings of recordings from cells stimulated with 0.08 mM TSP 21 21 an increase in [Ca ] . To verify that the attenuated Ca was (A) and pretreatment with 5 mM EGTA (B)or10mM Ni (C) before TSP addition. Note the inhibition of this increase in [Ca ] by EGTA or due to inhibition of Ca influx, cells were pretreated with 21 21 Ni pretreatment. Representative tracings of single cell recordings Ni , which is known to block Ca fluxes through many types from at least four separate experiments. Assay was carried out as 21 21 21 of Ca channels. Ni at 10 mM abolished the rise in [Ca ] described under “Experimental Procedures.” induced by TSP similar to that of EGTA (Fig. 2C). Neither 21 21 EGTA nor Ni affected the baseline [Ca ] under these con- ditions. These results indicate that the increase in [Ca ] from and multiple cell surface receptors. To address the issue of TSP is dependent on extracellular Ca fluxes into the cells. structural specificity, experiments were carried out to study Role of av Integrins in Mediating TSP-induced [Ca ] In- the effects of known inhibitors on these cell binding domains crease—TSP is known to have multiple cell binding domains and receptors. One of the major cell surface receptors for TSP 21 Thrombospondin Mediates Ca Mobilization in Fibroblasts 23749 FIG.3. Cell surface expression of thrombospondin receptors in IMR-90 fibroblasts by flow cytometry. Cells were gated by for- ward and side light scatter, and 5,000 gated events were collected. IMR-90 expressed both avb3 and avb5 but not CD36 on their cell surface. Assay was carried out as described under “Experimental Procedures.” belongs to the av integrin family (6, 7). Previous studies have shown that other soluble ligands to the av integrin were able to raise [Ca ] (36, 37). The IMR-90 cells express both avb3 and avb5 on their cell surface as evident from the flow cytometry data (Fig. 3). To investigate the possibility that the TSP-in- duced increase in [Ca ] was mediated via the av integrin, the effect of an antibody to avb3 (LM609) was studied (38). LM609, FIG.4. [Ca ] tracings showing the effect of various anti-av antibodies on the thrombospondin-induced rise in [Ca ] . Cells at 4 and 10 mg/ml, inhibited TSP-induced increases in [Ca ] were pretreated with 4 mg/ml LM609 (anti-avb3) (A), 4 mg/ml B6H12 by 54 and 63%, respectively (Figs. 4A and 5). (anti-IAP) (B), or 10 mg/ml P1F6 (anti-avb5) (C) before addition of 0.08 Since this Ca increase was apparently due to a calcium mM TSP. Representative tracings of single cell recordings from at least influx, we investigated further whether it was via the integrin- four separate experiments. Note the partial inhibition of TSP-induced increases in [Ca ] with LM609 and B6H12 but not with P1F6. Assay associated protein (IAP) which has been shown to mediate i 21 was carried out as described under “Experimental Procedures.” integrin-regulated Ca flux into cells (39). We tested the effect of a recently described antibody to IAP, B6H12 (31), on TSP- induced increases in [Ca ] . B6H12, at 4 and 10 mg/ml, inhib- ited TSP-induced increases in [Ca ] by 42% and 50%, respec- tively (Figs. 4B and 5). On the other hand, the TSP-induced increase in [Ca ] was not inhibited by the control antibody to avb5 (Figs. 4C and 5). In addition, LM609 and B6H12 do not block non-TSP-, i.e. thrombin, mediated increases in [Ca ] , which indicates they specifically block avb3 integrin-mediated 21 2 Ca flux. None of these monoclonal antibodies raised the baseline [Ca ] at the concentrations tested. Earlier studies have reported that integrin clustering is very important in signal transduction (25, 40, 41). Here we investi- gated the effect of TSP and inhibitors of the TSP-avb3 inter- action on integrin clustering. As shown in Fig. 6, TSP induces significant integrin clustering in IMR-90. This is not seen in cells that were not exposed to TSP or those that were pre- treated with LM609, suggesting a specific interaction between the integrin and the matrix protein. Role of the RGD Domain in TSP-induced[Ca ] Increase— Many of the integrins, including avb3, recognize the RGD FIG.5. Effect of anti-av antibodies on the thrombospondin- tripeptide sequence as a binding motif. The RGD domain is induced rise in [Ca ] . Antibodies LM609 and B6H12 (4 and 10 found in the type III calcium binding repeats and has been mg/ml) partially inhibited TSP-induced increases in [Ca ] . P1F6 at a similar concentration has minimal effect. Mean 6 S.E. (n 5 4). Assay shown to support cell attachment. Two RGD-containing pep- was carried out as described under “Experimental Procedures.” tides, GRGDSP and echistatin, were used to test the RGD dependence of the TSP-induced increase in [Ca ] . At 0.8 mM concentration of GRGESP only inhibited 7% (Fig. 7, A and B). GRGDSP, the increase is inhibited by 48%, whereas the same Echistatin, which is known to be a more potent integrin antag- onist (42), inhibited the increase in [Ca ] by 72% at 0.8 mM Data not shown. (Fig. 7C). Both GRGDSP and echistatin inhibit the TSP-in- 21 23750 Thrombospondin Mediates Ca Mobilization in Fibroblasts FIG.7. Effect of RGD-containing peptides on the throm- bospondin-induced rise in [Ca ] . Cells were pretreated with 0.8 mM GRGDSP (A), 0.8 mM GRGESP (B), or 0.8 mM echistatin (C) before the addition of 0.08 mM TSP. Note the partial inhibition of TSP-induced increases in [Ca ] with GRGDSP and echistatin but not with the control inactive peptide GRGESP. However, GRGDSP at this concen- tration also increases [Ca ] . Representative tracings of single cell recordings from at least four separate experiments. Assay was carried out as described under “Experimental Procedures.” [Ca ] signal when TSP is added. However, this is unlikely since subsequent stimulation with other agonists, such as thrombin, was able to elicit another transient increase in 21 2 [Ca ] . This response to GRGDSP was RGD-specific as the control peptide, GRGESP, did not induce an increase in [Ca ] . Also, echistatin, by itself, demonstrated a similar increase in [Ca ] in 50% of the cells at 5 mM, but no effect was seen at 1 mM. Role of Non-RGD Cell Binding Domains in [Ca ] In- crease—Since neither anti-avb3 nor RGD-containing peptides FIG.6. Immunolocalization of avb3 integrin on IMR-90 using can completely inhibit TSP-induced rises in [Ca ] , it is pos- indirect fluorescence. Fluorescence images showing very diffuse dis- tribution of avb3 integrin on the surface of IMR-90 (A). With pretreat- sible that there may be other components that mediate the ment of 0.17 mM TSP for 2 min (B), cells showed significant integrin 21 Ca signaling in addition to the integrin avb3. To identify the clustering as indicated by the white arrows. A 2-min preincubation with structural basis for this non-avb3 component of the intracellu- 20 mg/ml LM609, which prevents the interaction of avb3 with TSP, lar signaling, a series of experiments was carried out using showed significantly less clustering (C). Cells incubated with isotype match antibody did not stain (not shown). The distribution of avb3 was antibodies and peptides directed to either the different cell observed using LM609 followed by Cy3-conjugated goat anti-mouse IgG binding domains or the receptor(s) of TSP. as described under “Experimental Procedures” (magnification, 3 1000). One of the potential candidates is CD36, which is the recep- tor for the type I repeat of TSP (17). This surface glycoprotein duced [Ca ] increase in a concentration-dependent manner CD36 has been shown to mediate signal transduction in plate- (Fig. 8). lets by increasing [Ca ] (43). However, in this system, it is 21 21 In addition to blocking the TSP-induced increase in [Ca ] , unlikely that CD36 mediates the increase in [Ca ] as IMR-90 i i GRGDSP by itself, at concentration .0.8 mM, caused a tran- do not express CD36 on their cell surface (Fig. 3). In addition, sient increase in [Ca ] in .90% of the cells. It is possible that the peptide CSVTCG, which corresponds to the CD36 binding the depletion of available Ca due to this GRGDSP-induced motif on TSP and has been shown to inhibit cell adhesion (13), 21 21 increase in [Ca ] might be responsible for the attenuation of did not have any effect on TSP-induced rises in [Ca ] (Fig. 9). i i 21 Thrombospondin Mediates Ca Mobilization in Fibroblasts 23751 FIG. 10. [Ca ] tracing showing the effect of thrombospondin carboxyl-terminal peptide on [Ca ] in IMR-90 cells. The carbox- yl-terminal peptide, RFYVVMWK (100 mM), increases [Ca ] in IMR- 90. Representative tracings of single cell recordings from at least four separate experiments. Assay was carried out as described under “Ex- perimental Procedures.” domain, the TSP-induced increase in [Ca ] is inhibited by FIG.8. Dose-response relationship of the inhibitory effect of 50%. This suggests that part of the Ca signal generated by RGD-containing peptides. The effect of echistatin (l), GRGDSP (E), TSP may be due to the binding of the carboxyl-terminal domain or GRGESP (3) on TSP-induced increases in [Ca ] . Echistatin showed to its receptor (Fig. 9). This effect is apparently specific to the more potent inhibition of TSP-induced increases in [Ca ] than carboxyl-terminal binding domain since antibody A4.1 (10, 44), GRGDSP with a shift of 3–4 log concentrations to the left. Each data directed against the amino-terminal half of the central stalk point is the mean of at least four experiments with S.E. # 10%. Assay was carried out as described under “Experimental Procedures.” region of TSP, did not block the [Ca ] increase. Nonspecific antibody did not have any significant effect on [Ca ] increase. Interestingly, when both LM609 (10 mg/ml) and C6.7 (1:100) were used together, no synergistic or additive effect was seen. It is unclear, at this point, if this is due to a lack of cooperativity of the two sites. Nevertheless, more direct evidence of the role of the carboxyl-terminal of TSP came from studies with the carboxyl-terminal domain peptide RFYVVMWK (45). This pep- tide at 100 mM triggered a transient increase in [Ca ] that reached $2 mM in 35% of the cells (Fig. 10). This suggests that the carboxyl-terminal cell binding domain may mediate part of the TSP-induced rise in [Ca ] . Altogether, these data indicate that intact TSP can mediate an increase in [Ca ] in IMR-90 fibroblasts via at least two pathways: the RGD domain binding to the avb3 integrin and the carboxyl-terminal domain possibly binding to an unidentified receptor. DISCUSSION In this study we have shown that TSP, a matrix protein which binds to multiple cell surface receptors, increases [Ca ] in IMR-90 fibroblasts. This increase, which is partially medi- FIG.9. Effect of inhibitors against the non-RGD domains of ated by the RGD domain of TSP binding to the integrin avb3, thrombospondin on thrombospondin-induced rise in [Ca ] . is not limited to IMR-90, as we have also observed this in Antibodies to TSP cell binding domains, C6.7 (1:100), A4.1 (1:100), endothelial cells. We have also shown that known inhibitors of heparin (100 mg/ml), and control mouse IgG (50 mg/ml) were preincu- avb3, GRGDSP peptide and echistatin, in addition to blocking bated with TSP for 30 min before addition to cells. Peptide CSVTCG TSP-induced signaling, are capable of triggering an increase in (0.1 mM) was added directly to the IMR-90 cells. Of these inhibitors only the antibody to the carboxyl-terminal domain of TSP, C6.7, is effective [Ca ] by themselves. This suggests that RGD-containing pep- in inhibiting TSP-induced increase in [Ca ] . Assay was carried out as tides can function as both competitive antagonists and partial described under “Experimental Procedures.” agonists for their receptors. This had been reported previously 21 by several other groups; however, our threshold concentrations CSVTCG by itself had no effect on baseline [Ca ] . Likewise, are much higher than those reported elsewhere (36, 37). One the heparin binding domain of TSP also did not seem to play a 21 possible explanation is that high concentrations of peptides or role in generating this [Ca ] increase because heparin, at a antibodies could induce clustering of integrins, which could concentration (100 mg/ml) that had been shown to effectively activate the integrins and trigger transmembrane Ca flux block cell adhesion (15), did not inhibit TSP-induced rises in 21 not normally seen at the lower concentrations (26, 46). This [Ca ] (Fig. 9). clustering is seen in the present study when avb3 integrins Another possibility is the carboxyl-terminal cell binding do- interact with TSP. These findings illustrate that integrin clus- main of TSP which has been described to support non-RGD-de- pendent cell adhesion in melanoma cell lines (14, 15). When TSP was preincubated with C6.7, an antibody against this P. W. Tsao and S. A. Mousa, unpublished observations. 21 23752 Thrombospondin Mediates Ca Mobilization in Fibroblasts tering, presumably caused by multimeric interactions between With intact TSP, only 1 mM is required to trigger similar in- the ligand and its receptors, is important in transmembrane creases in [Ca ] . The main reason for this disparity in thresh- signaling (41). This is further supported by the observation old concentrations is the multivalent and higher affinity bind- ing of the multimeric matrix protein TSP versus the that RGD peptide can produce an increase in [Ca ] (.1mM) in kidney epithelial cells when coupled to beads but not when it monovalent and lower affinity binding of the peptide fragments (41, 50). In addition to having high affinity receptors, multim- is in the soluble monomeric form (29). In addition, earlier b2 integrin (CD11b/CD18) demonstrate that an- eric ligands are more effective in inducing receptor clustering studies with as shown in the present study; together, these have been shown tibodies to the a and b can induce Ca signal only if they are to work synergistically in transmembrane signaling (41). cross-linked (40). In addition, TSP might potentially mediate signaling via Another confounding factor is that this ligand-induced inte- other receptors. One such TSP receptor is CD36 which was not grin signaling may be dependent on the type of cell preparation studied here due to the lack of expression on IMR-90 fibro- being used. Sjaastad et al. (29) have reported that soluble RGD blasts. There is evidence from the literature that CD36 is (;0.3 mM) was not able to produce an increase in [Ca ] in capable of mediating signal transduction by increasing [Ca ] kidney epithelial cells, whereas two other groups both observed i in platelets and triggering oxidative bursts in monocytes (43, an increase in osteoclasts at a similar concentration (36, 37). 51). However, these two studies utilized monoclonal antibodies Even within the same cell type the data are not consistent. In against CD36 to trigger the signaling events, and, therefore, it that regard, echistatin was shown to induce increases in is uncertain if intact TSP or its fragments will produce similar [Ca ] in rat osteoclasts by one group but not by the other results via CD36. group (36, 37). This discrepancy indicates the complexity of In summary, we have demonstrated that TSP can trigger ligand/integrin-mediated signaling. transient increases in [Ca ] in IMR-90 fibroblasts by increas- Ligand-induced signaling through integrins does not follow ing Ca influx. The response is inhibitable by specific TSP the paradigm for traditional signaling receptors. In general, receptor antibodies indicating that this is a receptor-mediated integrins differ from the known signaling receptors in their effect. While we have shown this to occur primarily via the short cytoplasmic domain, lack of kinase activity, and lack of RGD and carboxyl-terminal domains of TSP and their corre- direct interaction with G-proteins (23). In this study we have 21 sponding receptors, we cannot rule out the possibility of the shown that the [Ca ] signal triggered by the soluble matrix other TSP receptors as signaling molecules in other cellular protein, TSP, via the integrin avb3, is dependent on extracel- 21 systems. The data from the present study, together with pre- lular calcium. The present data using EGTA or Ni indicates vious reports, indicate that TSP can support cell attachment that the increase in [Ca ] is due primarily to an influx of 21 21 and signaling through its different cell binding domains (4, 7, extracellular Ca , possibly through some type of Ca chan- 13, 14, 18, 45). These observations confirm that matrix pro- nel. Additional evidence supporting this hypothesis was dem- teins, with their multiple cell binding domains, can interact onstrated by our data with the antibody to IAP, B6H12, which with corresponding cell surface receptors to transduce signals was shown to block ;50% of [Ca ] increase. IAP, which has across the cell membrane. These outside-in signaling events been shown to be closely associated with avb3 (31), was re- will be important in understanding the role TSP plays in cell cently shown by Schwartz et al. (39) to be required specifically attachment, migration, and proliferation. for integrin-regulated Ca influx. This suggests that IAP may function as an integrin-associated calcium channel. Moreover, Acknowledgments—We thank Dr. Robert C. Newton for critically reading the manuscript, Dr. Ram Seetharam for peptide synthesis, Dr. purified integrin aIIbb3 was reported to function as a calcium Eric J. Brown for his gift of B6H12, Pamela Kidd for her expert tech- channel when reconstituted into liposomes (47). This suggests nical assistance, and William Lorelli for his editorial assistance. that the integrin itself can act as a conduit for transmembrane 21 21 Ca flux that is distinct from the classical Ca channels. REFERENCES Taken together, it is possible that the avb3 integrin or its 1. Clezardin, P., Jouishomme, H., Chavassieux, P., and Marie, P. J. (1989) Eur. J. Biochem. 181, 721–726 associated proteins, upon ligand binding, can be activated and 2. Dameron, K. M., Volpert, O. V., Tainsky, M. A., and Bouck, N. (1994) function as a Ca channel for signal transduction. Science 265, 1582–1584 3. Taraboletti, G., Roberts, D., Liotta, L. A., and Giavazzi, R. (1990) J. Cell Biol. The present data also show that there are other components 111, 765–772 of the transmembrane Ca flux that are non-RGD and non- 4. Yabkowitz, R., and Dixit, V. M. (1991) Cancer Res. 51, 3648–3656 avb3-dependent. This notion comes from the inability of the 5. Majack, R. A., Goodman, L. V., and Dixit, V. M. (1988) J. Cell Biol. 106, 415–422 synthetic RGD-peptide and LM609 to completely inhibit TSP- 6. Frazier, W. A. (1991) Curr. Opin. Cell Biol. 3, 792–799 induced rises in [Ca ] and is verified by the observation that 7. Lawler, J., Weinstein, R., and Hynes, R. O. (1988) J. Cell Biol. 107, 2351–2361 8. Nicosia, R. F., and Tuszynski, G. P. (1994) J. Cell Biol. 124, 183–193 an antibody to carboxyl-terminal domain, C6.7, is also able to 9. Savill, J., Hogg, N., Ren, Y., and Haslett, C. (1992) J. Clin. Invest. 90, inhibit 50% of the Ca signal. The carboxyl-terminal domain 1513–1522 10. Good, D. J., Polverini, P. J., Rastinejad, F., Le Beau, M. M., Lemons, R. S., of TSP has been reported to support chemotaxis and cell mi- Frazier, W. A., and Bouck, N. P. (1990) Proc. Natl. Acad. Sci. U. S. A. 87, gration (48, 49). Further studies using the carboxyl-terminal 6624–6628 domain peptide, RFYVVMWK, indicate that the carboxyl-ter- 11. Yamada, K. M. (1991) J. Biol. Chem. 266, 12809–12812 12. Dardik, R., and Lahav, J. (1987) Eur. J. Biochem. 168, 347–355 minal domain of TSP is also capable of triggering Ca in- 13. Tuszynski, G. P., Rothman, V. L., Deutch, A. H., Hamilton, B. K., and Eyal, J. creases by itself. Since this Ca increase is observed only in a (1992) J. Cell Biol. 116, 209–217 14. Kosfeld, M. D., Pavlopoulos, T. V., and Frazier, W. A. (1991) J. Biol. Chem. subpopulation of cells, this receptor might not be present on all 266, 24257–24259 cells. Two possible receptors for the carboxyl-terminal binding 15. Kosfeld, M. D., and Frazier, W. A. (1992) J. Biol. Chem. 267, 16230–16236 16. Roberts, D. D. (1988) Cancer Res. 48, 6785–6793 domain have so far been identified, an 80/105-kDa complex on 17. Asch, A. S., Barnwell, J., Silverstein, R. L., and Nachman, R. L. (1987) J. Clin. carcinoma cells (4) and a 52-kDa domain on K-562 cells (19). Invest. 79, 1054–1061 Even if either of these receptors exists in the IMR-90, it is 18. Sun, X., Skorstengaard, K., and Mosher, D. F. (1992) J. Cell Biol. 118, 693–701 19. Gao, A.-G., and Frazier, W. A. (1994) J. Biol. Chem. 269, 29650–29657 unclear if they are responsible for the binding of the carboxyl- 20. Brooks, P. C., Montgomery, A. M. P., Rosenfield, M., Reisfeld, R. A., Hu, T., terminal TSP fragment and the subsequent transduction of the Klier, G., and Cheresh, D. A. (1994) Cell 79, 1157–1164 21. Montgomery, A. M. P., Reisfeld, R. A., and Cheresh, D. A. (1994) Proc. Natl. signal across the cell membrane. The carboxyl-terminal peptide Acad. Sci. U. S. A. 91, 8856–8860 and GRGDSP trigger increases in [Ca ] in the millimolar 22. Hynes, R. O. (1992) Cell 69, 11–25 range at peptide concentrations of 0.1 and 0.8 mM, respectively. 23. Juliano, R. L., and Haskill, S. (1993) J. Cell Biol. 120, 577–585 21 Thrombospondin Mediates Ca Mobilization in Fibroblasts 23753 24. Clark, E. A., Shattil, S. J., and Brugge, J. S. (1994) Trends Biol. Sci. 19, G. A. (1994) Am. J. Physiol. 266, C376–C381 464–469 38. Cheresh, D. A. (1987) Proc. Natl. Acad. Sci. U. S. A. 84, 6471–6475 25. Kornberg, L. J., Earp, H. S., Turner, C. E., Prockop, C., and Juliano, R. L. 39. Schwartz, M., Brown, E. J., and Fazeli, B. (1993) J. Biol. Chem. 268, (1991) Proc. Natl. Acad. Sci. U. S. A. 88, 8392–8396 19931–19934 26. Schwartz, M. A., Ingber, D. E., Lawrence, M., Springer, T. A., and Lechene, C. 40. Ng-Sikorski, J., Andersson, R., Patarroyo, M., and Andersson, T. (1991) (1991) Exp. Cell. Res. 195, 533–535 Exp. Cell Res. 195, 504–508 27. Schwartz, M. A. (1993) J. Cell Biol. 120, 1003–1010 41. Miyamoto, S., Akiyama, S. K., and Yamada, K. M. (1995) Science 267, 883–885 28. O’Toole, T. E., Mandelman, D., Forsyth, J., Shattil, S. J., Plow, E. F., and 42. Gan, Z.-R., Gould, R. J., Jacobs, J. W., Friedman, P. A., and Polokoff, M. A. Ginsberg, M. H. (1991) Science 254, 845–847 (1988) J. Biol. Chem. 263, 19827–19832 29. Sjaastad, M. D., Angres, B., Lewis, R. S., and Nelson, W. J. (1994) Proc. Natl. 43. Alessio, M., Greco, N. J., Primo, L., Ghigo, D., Bosia, A., Tandon, N. N., Acad. Sci. U. S. A. 91, 8214–8218 Ockenhouse, C. F., Jamieson, G. A., and Malavasi, F. (1993) Blood 82, 30. Lin, T. H., Yurochko, A., Kornberg, L., Morris, J., Walker, J. J., Haskill, S., and 3637–3647 Juliano, R. L. (1994) J. Cell Biol. 126, 1585–1593 44. Prater, C. A., Plotkin, J., Jaye, D., and Frazier, W. A. (1991) J. Cell Biol. 112, 31. Brown, E., Hooper, L., Ho, T., and Gresham, H. (1990) J. Cell Biol. 111, 1031–1040 2785–2794 45. Kosfeld, M. D., and Frazier, W. A. (1993) J. Biol. Chem. 268, 8808–8814 32. Grynkiewicz, G., Poenie, M., and Tsien, R. Y. (1985) J. Biol. Chem. 260, 46. Ingber, D. E., Prusty, D., Frangioni, J. J., Cragoe, E. J., Jr., Lechene, C. P., and 3440–3450 Schwartz, M. A. (1990) J. Cell Biol. 110, 1803–1811 33. Smith, J. W., Piotrowicz, R. S., and Mathis, D. (1994) J. Biol. Chem. 269, 47. Rybak, M. E. M., and Renzulli, L. A. (1989) J. Biol. Chem. 264, 14617–14620 960–967 48. Yabkowitz, R., Mansfield, P. J., Dixit, V. M., and Suchard, S. J. (1993) Cancer 34. Steiner, B., Cousot, D., Trzeciak, A., Gillessen, D., and Hadva´ry, P. (1989) Res. 53, 378–387 J. Biol. Chem. 264, 13102–13108 49. Yabkowitz, R., Mansfield, P. J., Ryan, U. S., and Suchard, S. J. (1993) J. Cell. 35. Tsao, P. W., Bozarth, J. M., Jackson, S. A., Forsythe, M. S., Flint, S. K., and Physiol. 157, 24 Mousa, S. A. (1995) Thromb. Res. 77, 543–556 36. Shankar, G., Davison, I., Helfrich, M. H., Mason, W. T., and Horton, M. A. 50. Welder, C. A., Lee, D. H. S., and Takei, F. (1993) J. Immunol. 150, 2203–2210 (1993) J. Cell Sci. 105, 61–68 51. Schu¨epp, B. J., Pfister, H., Clemetson, K. J., Silverstein, R. L., and Jungi, T. W. 37. Zimolo, Z., Wesolowski, G., Tanaka, H., Hyman, J. L., Hoyer, J. R., and Rodan, (1991) Biochem. Biophys. Res. Commun. 175, 263–270

Journal

Journal of Biological Chemistry – Unpaywall

Published: Oct 1, 1995

There are no references for this article.