Thrombin receptor and RhoA mediate cell proliferation through integrins and cysteine-rich protein 61

Abstract

A subset of G-protein coupled receptors (GPCRs), including the thrombin receptor (PAR1), elicits mitogenic responses. Thrombin also activates Ras homolog gene family member A (RhoA) and activating protein (AP-1) -mediated gene expression in 1321N1 astrocytoma cells, whereas the nonmitogenic agonist carbachol does not. Transcriptomic analysis was used to explore differential gene induction by these agonists and revealed that the matricellular protein cysteine-rich 61 (Cyr61/CCN1) is selectively induced by thrombin. The ability of GPCR agonists to induce Cyr61 parallels their ability to activate RhoA; agonist-stimulated Cyr61 expression is inhibited by C3 toxin. When Cyr61 is down-regulated using short interfering RNA (siRNA) or short-hairpin RNA (shRNA), thrombin-induced DNA synthesis is significantly attenuated. When Cyr61 expression is induced, it appears in the extracellular compartment and on the cell surface. Extracellular Cyr61 interacts with α₅, α₆, and β₁ integrins on these cells, and monoclonal antibodies directed against α₅ and β₁ integrins inhibit thrombin-induced DNA synthesis. Functional blockade of Cyr61 with soluble heparin or anti-Cyr61 antibodies also inhibits thrombin-induced DNA synthesis. Thus Cyr61 is a highly inducible, secreted extracellular factor through which GPCR and RhoA signaling pathways engage integrins that contribute to GPCR-mediated proliferation.—Walsh, C. T., Radeff-Huang, J., Matteo, R., Hsiao, A., Subramaniam, S., Stupack, D., and Brown, J. H. Thrombin receptor and RhoA mediate cell proliferation through integrins and cysteine-rich protein 61.

Thrombin is a potent mitogen for a variety of cell types, including certain malignant cancer cells (1, 2). The mechanism by which thrombin promotes mitogenic responses has yet to be fully elucidated. Thrombin elicits cellular responses through the activation of G-protein coupled protease-activated receptors (PARs), which can signal through G_12/13, G_q, and G_i to initiate a downstream effect (3, 4). We utilized 1321N1 astrocytoma cells derived from glioblastoma multiforme to study the mechanism by which thrombin induces DNA synthesis. 1321N1 cells express PAR1 and M3 muscarinic receptors (mAChR), which we have shown to initiate overlapping signals, including phosphoinositide hydrolysis, Ca²⁺ mobilization, and protein kinase C activation (5). However, only PAR1 (not mAChR) stimulation activates G_12/13 and the low-molecular-weight GTPase Ras homolog gene family member A (RhoA), induces cell rounding, elicits c-Jun/AP-1 mediated gene expression, and stimulates DNA synthesis (5,6,7,8,9).

RhoA activation is linked to cytoskeletal responses by mechanisms that have been extensively documented (8, 10,11,12). RhoA signaling has also been implicated in aberrant cell growth and transformation (13,14,15), and its upstream regulators, including G_12/13 and various Rho GEFs, are known oncogenes (16,17,18). Several RhoA-dependent mechanisms that contribute to cell proliferation have been described (13,14,15,16, 19, 20), including regulation of transcription factors such as AP-1, which may induce genes involved in cell proliferation (13, 14, 20).

We utilized 1321N1 cells and the differential signaling properties of the endogenous PAR1 and M3 mAChR to explore mechanisms by which RhoA activation enables G-protein coupled receptors (GPCRs) to effect changes in gene expression that promote cell proliferation. Transcriptomic analysis was used to compare genes that were induced downstream of PAR1 and mAChR activation. This microarray analysis identified matricellular protein cysteine-rich 61 (Cyr61/CCN1) as the gene most robustly and selectively induced by thrombin, but not by carbachol. Cyr61 is a highly regulated matricellular protein associated with the extracellular matrix, where it can bind and signal through cell-surface integrins (21,22,23). Like other members of the 6-member family of CCN proteins, Cyr61 regulates a wide range of cellular processes, including proliferation, adhesion, survival, migration, and differentiation (21, 23,24,25,26). Certain glioblastoma cell lines express high levels of Cyr61 (27), and PAR1 activation has been linked to increased cell invasion through Cyr61 (28, 29).

Several GPCR ligands, including sphingosine-1-phosphate (S1P), lysophosphatidic acid (LPA), and prostaglandin F2α induce Cyr61 expression (30,31,32,33,34). Thrombin action on PAR-1 receptors has also been shown by Pendurthi et al. (35, 36) to increase Cyr61 expression in human and murine fibroblasts. Several studies implicate RhoA and the transcription factor AP-1 in regulation of Cyr61 expression (30, 32, 33, 37). We therefore hypothesized that the induction of Cyr61 could play a critical role in GPCR- and RhoA-mediated mitogenic responses. Cyr61 signals through cell surface integrins (24, 38), including the α₆β₁, α_vβ₃, and α_vβ₅ integrins (39,40,41). The possibility that Cyr61, up-regulated in response to GPCR activation, could signal through integrin engagement led us to further hypothesize that Cyr61 might provide a link between GPCR and integrin signal transduction pathways.

The studies presented here demonstrate that thrombin and other GPCR ligands act through RhoA and c-Jun/AP-1 to regulate Cyr61 expression. Remarkably, inhibition of Cyr61 expression with short interfering RNA(siRNA) or short-hairpin RNA (shRNA) significantly attenuates thrombin-mediated DNA synthesis. We further show that Cyr61 is secreted from and binds to 1321N1 cells in an integrin- and heparin-dependent manner and that blocking extracellular Cyr61 with heparin, anti-Cyr61 antibodies, or integrin antibodies inhibits thrombin-induced DNA synthesis. These data reveal a novel mechanism linking GPCR and integrin signaling and provide a pathway by which GPCRs that activate RhoA can positively affect cell proliferation.

MATERIALS AND METHODS

Materials

The pGEX-2T vector encoding a glutathione S-transferase (GST) -fusion protein containing amino acids 7–89 of rhotekin, which is the RhoA-binding domain, was provided by M. Schwartz (Departments of Microbiology and Biomedical Engineering, Cardiovascular Research Center and Mellon Urological Cancer Research Institute, University of Virginia, Charlottesville, VA, USA) (42). C3 exoenzyme was purchased from Upstate (Charlottesville, VA, USA). Human alpha thrombin was from Enzyme Research Laboratories (South Bend, IN, USA). LPA and S1P were purchased from Avanti Polar Lipids (Alabaster, AL, USA). Histamine dihydrochloride, carbamoylcholine chloride (carbachol), and heparin agarose beads were purchased from Sigma Chemical Company (St. Louis, MO, USA). Lyophilized recombinant human Cyr61 was purchased from Cell Sciences (Canton, MA, USA) and was reconstituted in PBS (pH 7.4) per manufacturer’s instructions. SP600125 and heparin sodium salt were purchased from Calbiochem (Gibbstown, NJ, USA). Anti-Cyr61, anti-actin, and anti-RhoA antibodies used for immunoblotting were purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Functionally inhibitory Cyr61 antibodies were a kind gift from Dr. Lester Lau (University of Illinois College of Medicine, Chicago, IL, USA).

Flow cytometry analysis

For blocking studies and flow cytometry analysis, antibody to integrins α₅β₁, α₆β₁, α_vβ₃, or α_vβ₅ (Millipore, Billerica, MA, USA) was suspended (10 μg/ml) in 2% FBS/PBS (pH 7.6) for 1 h on ice. Cells were washed 3 times by centrifugation, and phycoerythrin-conjugated secondary anti-mouse (Southern Biotech, Birmingham, Al, USA) was added in 2% FBS/PBS (1:40) for a further 30 min on ice. Cells were then washed 3 times and analyzed on a FACSCalibur cytometer (Becton Dickinson, Franklin Lakes, NJ, USA).

Cell culture

1321N1 human astrocytoma cells, originally isolated from primary cultures of a human cerebral glioblastoma multiforme, were cultured in Dulbecco modified Eagle medium supplemented with 5% FBS, 100 U/ml penicillin, 100 μg/ml streptomycin, and 4 mM l-glutamine in a 37°C, 10% CO₂ humidified environment, as described elsewhere (8). Cells were grown to confluence on 6- or 10-cm plates and serum-starved for 24 h prior to the start of each experiment. A 12-h starve was used in the siRNA experiments in order to maintain adequate siRNA-mediated Cyr61 knockdown.

DNA microarray analysis

Cells were cultured as described above and treated with vehicle, 0.5 U/ml thrombin, or 500 μM carbachol and harvested at 30 min, 3 h, 6 h, or 12 h following vehicle or agonist addition. RNA was collected, purified via RNeasy Mini kit (Qiagen, Valencia CA, USA), and hybridized to Affymetrix human genome U133A microarray chips using standard manufacturer’s protocols (Affymetrix, Santa Clara, CA, USA). Summary measures of gene expression were obtained using Affymetrix GCOS 1.1 software and analyzed using the VAMPIRE microarray analysis framework (43, 44). All 22,283 probes were used throughout the analysis. Stable variance models were constructed for each of the experimental conditions at each time point (n=2). Differentially expressed probes were identified using the unpaired VAMPIRE significance test (43) with a 2-sided, Bonferroni-corrected threshold of α_Bonf = 0.05. The VAMPIRE statistical test is a Bayesian statistical method that computes a model-based estimate of noise at each level of gene expression. This estimate was then used to assess the significance of apparent differences in gene expression between 2 experimental conditions.

SDS-PAGE and Western blotting

Following treatment with agonists and/or inhibitors, cells were washed and lysed in buffer containing 20 mM Tris HCl (pH 7.6), 250 mM NaCl, 3 mM EDTA, 3 mM EGTA, 20 mM β-glycerophosphate, 0.5% Nonidet P-40, 1 mM phenylmethylsulfonylfluoride, 10 μg/ml aprotinin, 1 mM disodium 4-nitrophenylphosphate, 100 μM sodium orthovanadate, and 10 μg/ml leupeptin. Lysates were clarified by centrifugation, 4× Laemmli buffer was added, and samples were boiled for 5 min. Proteins were resolved via SDS-PAGE analysis, and membranes were probed with primary antibodies overnight at 4°C. All primary antibodies were diluted 1:1000 and secondary immunoglobulin G (IgG) -horseradish peroxidase at 1:4000 in 3% BSA in Tris-buffered saline containing 0.1% Tween 20. Proteins were visualized using enhanced chemiluminescence and quantitated using gel documentation software (Alpha Innotech Corp., San Leandro, CA, USA).

Quantitative polymerase chain reaction (PCR)

Total RNA was extracted from treated cells, and cDNA was generated using the Superscript III First Strand Synthesis System for reverse transcriptase PCR (RT-PCR) (Invitrogen, Carlsbad, CA, USA), according to the manufacturer’s instructions. TaqMan gene expression assays (Applied Biosystems, Foster City, CA, USA) were used to amplify human Cyr61 (Hs00155479_m1) or human α-actin (4236315E-0504006) in conjunction with the Platinum Quantitative PCR SuperMix-UDG kit (Invitrogen) and analyzed by the Opticon 2 and Opticon Monitor software (MJ Research, Waltham, MA, USA).

RhoA activation assay

The assay for activated RhoA was preformed as described elsewhere (42, 45). Cells were grown to confluence on 10 cm plates, serum-starved for 24 h, and treated with vehicle or agonist for 3.5 min.

siRNA nucleofection

Cells were grown to confluence in 10-cm dishes, harvested, and counted. Cells were nucleofected with 19-nucleotide siRNAs (Ambion, Austin, TX, USA) using the Amaxa Nucleofector system (Amaxa, Gaithersburg, MD, USA) per manufacturer’s instructions using solution V and program T-16. Silencer predesigned siRNAs targeting human Cyr61 and the Silencer Negative Control #1 siRNA were purchased from Ambion. The Cyr61 siRNA sequence used in the final study was 5′-GGUGGAGUUGACGAGAAACtt-3′. 1321N1 cells were allowed to attach for 12 h, then were starved for 12 h prior to thrombin addition. Efficacy of Cyr61 knockdown was assessed at 12 h; [³H]-thymidine incorporation was assessed at 24 h.

shRNAs, lentivirus production, and infection of the 1321N1 cells

Cyr61 shRNAs and scrambled controls were purchased from Open Biosystems (Huntsville, AL, USA). The Cyr61 sequence chosen for this study was sense, 5′-GCAAACAGAAATCAGGTGTTT-3′ and antisense 5′-AAACACCTGATTTCTGTTTGC-3′. Recombinant lentiviruses were produced by cotransfecting 293T cells with pCMVΔ8.2Rvpr (46), pCMV-VSV-G, and the shRNA/plKO vector using previously established methods (47). Transfections were carried out using TransIT LT-1 (Mirus Bio, Madison, WI, USA). Virus was harvested at 48 and 72 h post-transfection, and infections were carried out in the presence of 6 μg/mL of protamine sulfate (Sigma). Following transfection, cells were selected with 5 μg/mL puromycin and cultured as above. Cells were plated, starved for 24 h, then stimulated with thrombin for 12 h. The expression of Cyr61 was measured, and the stable clone with the highest knockdown efficiency was used for the study. For studies measuring [³H]-thymidine incorporation, stable clones of shRNA expressing cells were grown to confluence in 10-cm dishes, harvested, and seeded into 24-well plates. Cells were then starved for 24 h prior to the addition of thrombin.

[³H]-thymidine incorporation assay

Assays were performed as described previously. 1321N1 cells were grown to confluence in 10-cm dishes, harvested, and plated in 24-well plates (19, 45). For studies involving functionally inhibitory anti-integrin antibodies (i.e., Fig. 6C), cells were seeded into plates coated with collagen (10 μg/ml). Cells were starved for 24 h prior to the addition of heparin, anti-Cyr61, or anti-integrin antibodies and incubated for 1 h prior to the addition of agonists. Cells were labeled with 0.5 μCi [³H]-thymidine per well during the last 4 h of incubation, and incorporation was assessed by scintillation counting.

Figure 1.

Microarray, mRNA, and protein changes in Cyr61 demonstrate selective regulation by thrombin vs. carbachol. 1321N1 cells were treated with vehicle, 500 μM carbachol (CRB), or 0.5 U/ml thrombin (THR) for the times indicated. A) Cells were lysed and the mRNA was collected and hybridized to Affymetrix human genome U133A microarray chips using standard manufacturer’s protocols. Gene expression data was obtained using Affymetrix GCOS 1.1 software and interpreted with the VAMPIRE microarray analysis framework. Bars represent mean + sd of Cyr61 expression in thousands (values in arbitrary units). Values shown above bars represent the fold increase of thrombin vs. carbachol. B) Cells were lysed and mRNA levels were analyzed via quantitative PCR as described in Materials and Methods. Values shown are means + se of 3 independent experiments. C) Cells were lysed and subjected to SDS-PAGE followed by immunoblotting for Cyr61 or actin; a representative blot is shown.

Figure 2.

GPCR agonist-induced RhoA activation correlates with ability to induce Cyr61. 1321N1 cells were stimulated with vehicle (CTL), 500 μM carbachol (CRB), 10 or 100 μM histamine (HIST), 1 or 10 μM LPA; 1 or 10 μM S1P; or 0.5 U/ml thrombin (THR). A) Top: reactions were stopped at 3.5 min, and RhoA activity was assessed as described in Materials and Methods. The blot shows active RhoA (pulled down by rhotekin) and total RhoA visualized using an anti-RhoA antibody. Bottom: cells were stimulated as above, and cells were lysed after 3 h and subjected to SDS-PAGE followed by immunoblotting with anti-Cyr61 or antiactin antibodies. B) Mean densitometric values for active vs. total RhoA from 3 independent experiments are shown. C) Mean densitometric values for Cyr61 normalized by actin for 4 independent experiments. Values are means + se; *P < 0.001 vs. vehicle control.

Figure 3.

Agonist-induced Cyr61 expression is dependent on RhoA and c-Jun activation. A, B) 1321N1 cells were treated with 6 μg/ml C3 exoenzyme or vehicle overnight prior to the addition of vehicle, 0.5 U/ml thrombin (THR) (A), or 10 μM LPA or S1P (B). Cells were lysed at either 3 h (for LPA and S1P; B) or 12 h (for THR; A) following agonist addition and subjected to SDS-PAGE followed by immunoblotting with anti-Cyr61 or antiactin antibodies. C) 1321N1 cells were treated with 10 μM SP600125 or vehicle for 1 h before the addition of 0.5 U/ml thrombin. Cells were lysed 12 h following thrombin addition and subjected to SDS-PAGE. Bar graphs show densitometric results from 3 separate independent experiments using C3 (A) or SP600125 (C), presented as means + se; *P < 0.05 vs. thrombin alone. Representative blots are shown.

Figure 4.

Cyr61 siRNA and shRNA inhibit thrombin-stimulated DNA synthesis. 1321N1 cells were nucleofected with 10 ng of scrambled siRNA (Scram) or Cyr61 siRNA using the Amaxa system, or with infected Cyr61 or scrambled (Scram) shRNA containing lentivirus, as described in Materials and Methods. A, B) Cells were stimulated with thrombin (0.5 U/ml), and the expression of Cyr61 protein was measured after 12 h. Representative blots are shown. C, D) Cells nucleofected with Scrambled or Cyr61 siRNA (C) or shRNA (D) were starved and stimulated with 0.5 U/ml thrombin (THR) or vehicle for 24 h. DNA synthesis was assessed by [³H]-thymidine incorporation. Data represent percentages of maximal stimulation; values are means + se of 3 separate experiments carried out in triplicate. *P < 0.001 vs. Scram siRNA + THR; **P < 0.01 vs. Scram shRNA + THR.

Figure 5.

Cyr61 detection in the extracellular space. A) 1321N1 cells were seeded, starved for 24 h, and 5 U/ml heparin and 50 μl of heparin beads were added to the medium. Cells were then stimulated with vehicle (CTL), thrombin peptide TFLLRN (THRp), S1P, or thrombin (THR) at the dose indicated. The beads were collected after 12 h and washed; bound proteins were resolved by SDS-PAGE and immunoblotted using an anti-Cyr61 antibody. B) Cells were starved and then 5 U/ml heparin plus 50 μl of heparin beads was added prior to 12 h stimulation with 0.5 U/ml thrombin or vehicle. Cells were lysed and analyzed as in Fig. 1C; experiments were carried out at least 3 times, representative blots are shown.

Figure 6.

Integrin α₅β₁ mediates adhesion to Cyr61 and thrombin-stimulated DNA synthesis in 1321N1. A, B) Adhesion assays were carried out on 1321N1 cells as described in Materials and Methods. A) Cells were allowed to attach to 0–20 μg/ml Cyr61 in the presence or absence of heparin (10 U/ml) and/or integrin antibodies (10 μg/ml); values are mean ± sd number of attached cells. B) Cells were allowed to attach to 5 μg/ml Cyr61 in the presence of integrin antibodies (10 μg/ml) or EDTA (10 mM); data are plotted as percentage of control; bars represent mean + se number of attached cells. *P < 0.01, **P < 0.001 vs. control. C) 1321N1 cells were seeded, starved, and treated with integrin antibodies (10 μg/ml) 30 min prior to stimulation with 0.5 U/ml thrombin (THR) or vehicle for 24 h. DNA synthesis was assessed by [³H]-thymidine incorporation as described in Materials and Methods. Data are plotted as percentage of maximal thrombin response; values are means + se. †P < 0.001 vs. thrombin alone. All data represent at least 3 separate experiments carried out in triplicate.

Detection of Cyr61 in the culture medium

1321N1 cells were seeded in 6-cm dishes and starved for 24 h. Prior to agonist addition, 50 μl of heparin agarose beads and 1 U/ml of soluble heparin were added to the culture medium. Cyr61 binds heparin; therefore, the soluble heparin acts to scavenge Cyr61 from the cell surface and from the culture medium while the heparin-agarose beads act as a sink to store the Cyr61 for analysis via immunoblot. Cells were then stimulated with thrombin (0.5 U/ml), and the beads and medium were collected after 12 h. The beads and bound proteins were centrifuged and washed 3 times in cold PBS, followed by elution via boiling in Laemmli buffer. Proteins were resolved by SDS-PAGE as described above.

Cell adhesion assay

Forty-eight-well tissue culture plates were incubated with 100 μl of PBS (pH 8.0) containing 0–20 μg/ml Cyr61 for 6 h at 4°C. Plates were then washed once with PBS (pH 7.6) and blocked with a solution of 3% BSA/PBS (pH 7.6) for 1 h at 37°C. 1321N1 cells were added (2.5×10⁵ cells/ml, 200 μl/well) in the presence or absence of antibodies to integrin (10 μg/ml, Millipore) or heparin (10 U/ml) and allowed to attach for 40 min at 37°C. Unbound cells were aspirated, and cells remaining adherent were stained and fixed with 0.1% crystal violet (Sigma) in PBS (pH 7.6) and 20% methanol solution. Unbound dye was removed by washing with water, and the plate was dried overnight. Cell-bound dye was reconstituted in methanol and quantified by absorbance at 600 nm.

Statistical analysis

Statistical significance was determined using ANOVA followed by the Tukey post hoc test. A value of P < 0.05 was considered significant.

RESULTS

The molecular basis for the mitogenic effect of thrombin has not been fully elucidated. To understand how thrombin promotes mitogenic signaling, we performed a microarray-based search to discover downstream genes selectively regulated by thrombin vs. carbachol, a ligand that induces many similar downstream signals but is not mitogenic (5, 8, 48). 1321N1 astrocytoma cells, derived from a human cerebral glioblastoma multiforme, were treated with vehicle, thrombin, or carbachol, harvested at 30 min, 3 h, 6 h, or 12 h after agonist addition, and gene expression patterns were compared with an unpaired VAMPIRE statistical test (43). A number of genes showed significantly different responses to thrombin and carbachol, with the number of significant genes increasing with time. The gene that was the most selectively and robustly upregulated by thrombin (vs. carbachol) at every time point examined was the CCN family member cyr61 (P<0.0001 at 30 min; <1×10⁻⁹ at 3 h; <1×10⁻¹⁵ at 6 h; and <1×10⁻¹⁶ at 12 h; Fig. 1A). CTGF, another CCN family member was also highly upregulated at later times (P<1×10–16 at 6 and 12 h). Both the CCN family members have been shown to exhibit pleiotropic functions, regulating processes such as cell differentiation, migration, and adhesion. The present study focused on examining the cyr61 gene which, due to its both early and sustained expression, was an intriguing potential player in thrombin-stimulated cell proliferation.

We used quantitative PCR to further confirm that thrombin regulates Cyr61 mRNA expression. Cyr61 mRNA levels were significantly increased (P<0.001) as early as 30 min after thrombin addition and continued to increase over a 12-h time course (Fig. 1B). In contrast, carbachol caused a modest increase in Cyr61 mRNA at 30 min (P>0.05) that diminished and was not sustained over time (Fig. 1B). Immunoblots for Cyr61 confirmed that the increase in Cyr61 message was translated into significant increases in Cyr61 protein expression over a time course similar to that seen for mRNA (Fig. 1C). Averaged values from experiments like those shown in Fig. 1C demonstrated that thrombin-induced Cyr61 protein expression is significantly increased (P<0.001) by 3 h. These data are consistent with the classification of cyr61 as an early response gene and suggest that Cyr61 expression may serve as an important early causal event, as well as a more sustained mediator of thrombin signaling.

We have shown that thrombin-stimulated DNA synthesis in 1321N1 cells is RhoA-dependent (9). The literature shows RhoA signaling is linked to Cyr61 expression (30, 32, 33). We used a panel of agonists that signal through GPCRs other than the mAChR and PAR1 to test whether if RhoA activation correlated with Cyr61 expression. Histamine, lysophosphatidic acid, and sphingosine-1-phosphate, as well as carbachol and thrombin, were tested. The data in Fig. 2 show that LPA, S1P, and thrombin all significantly activated RhoA and significantly up-regulated Cyr61 expression, whereas carbachol and histamine had no effect on RhoA, nor did they stimulate Cyr61 expression (Fig. 2). These data suggest that a connection between RhoA and Cyr61 expression is broadly applicable to GPCR signaling rather than specific to PAR1 activation and that Cyr61 might serve as a downstream mediator of RhoA signaling.

To directly demonstrate that RhoA mediates thrombin-induced Cyr61 expression, we utilized the C3 toxin from Clostridium botulinum, which ADP ribosylates and inactivates RhoA proteins (49). C3 treatment prior to thrombin stimulation significantly blocked Cyr61 expression (Fig. 3A), implicating RhoA signaling in thrombin-mediated Cyr61 expression. Treatment with C3 also markedly attenuated LPA- and S1P-induced Cyr61, confirming the importance of RhoA in GPCR-mediated Cyr61 production (Fig. 3B). The Cyr61 promoter contains an AP-1 consensus sequence (30, 50), and RhoA signaling has shown to be involved in the activation of c-Jun NH₂-terminal kinase (JNK) and in AP-1-mediated gene expression (20). To determine whether thrombin-mediated activation of Jun/AP-1 plays a role in Cyr61 expression, we blocked c-Jun activation by JNK with the pharmacological inhibitor SP600125. Inhibition of c-Jun activation markedly decreased thrombin-stimulated Cyr61 protein expression (Fig. 3C). Together these data suggest the Cyr61 expression is dependent on both RhoA and c-Jun activation.

Mitogenesis is a hallmark response to thrombin receptor activation. To determine whether production of Cyr61 contributes to thrombin-induced DNA synthesis in 1321N1 cells, we used siRNA to down-regulate Cyr61 expression. siRNA targeted to Cyr61 caused significant loss of thrombin-stimulated Cyr61 protein expression; nonspecific scrambled siRNA did not (Fig. 4A). Inhibition of Cyr61 expression with siRNA decreased the amount of thrombin-stimulated [³H]-thymidine incorporation by 48 ± 2% (mean±se, P<0.001) relative to cells transfected with scrambled siRNA (Fig. 4C). To confirm the siRNA results, we generated stable 1321N1 clones expressing Cyr61 shRNA. These clones exhibited a decrease in Cyr61 protein expression similar to that of siRNA transfected cells (Fig. 4B) and showed a 38 ± 9% (mean±se, P<0.01) decrease in thrombin-stimulated [³H]-thymidine incorporation relative to cells infected with scrambled shRNA (Fig. 4D).

Cyr61 contains a secretory signal peptide and performs the majority of its functions in the extracellular space (24, 26). Cyr61 is a heparin-binding protein. Detection of extracellular Cyr61 was facilitated by the addition of heparin (both soluble and bound to agarose beads) to compete for Cyr61 on the cell surface, in the medium, and bound to the extracellular matrix. This method of adding heparin to compete for cell-surface bound Cyr61 was utilized previously in studies from the Lau laboratory (22). Cyr61 levels in the medium were markedly increased after 12 h stimulation with these GPCR agonists (Fig. 5A). These experiments also revealed that addition of heparin (soluble and bound to beads) to the medium decreased the amount of Cyr61 detected in the whole cell lysate (i.e., that associated with the cell surface; Fig. 5B). This observation suggests that a significant component of the released Cyr61 is bound to the cell surface.

Integrins are cell-surface receptors for Cyr61 (21, 24). We identified the cell-surface integrins present on 1321N1 cells via FACS and found high expression of α₃ and β₁; moderate expression of α₂, α₅, and β₅; and low or absent expression of β₃ integrins (data not shown). To explore the role of integrins in Cyr61 binding to 1321N1 cells, we examined cell attachment to surfaces coated with increasing amounts of Cyr61. As shown in Fig. 6A, there was a concentration-dependent increase in cell attachment to Cyr61 (0.1–20 μg/ml). Addition of anti-β₁ integrin antibodies inhibited cell attachment to Cyr61 by ∼30% (Fig. 6A), whereas heparin inhibited cell adhesion by ∼50% and the combination of both led to ∼80% inhibition of adhesion. These data suggest that there are both integrin- and heparin-mediated components of cell adhesion to Cyr61 (Fig. 6A). To further investigate the specificity of the integrin-mediated binding component, we measured adhesion of cells to Cyr61 (5 μg/ml) in the presence of integrin-specific inhibitory antibodies (Fig. 6B). Antibodies to α₅, α₆, and β₁ significantly inhibited adhesion, whereas antibodies α₃, α₂, β₃, or β₅ did not (Fig. 6B and data not shown). Thus α₅β₁ and α₆β₁ are the integrins that mediate 1321N1 cell attachment to Cyr61.

In light of the involvement of β₁ integrins in Cyr61 interaction with 1321N1 cells, we examined the ability of integrin blockade (via specific antibodies) to affect thrombin-mediated DNA synthesis. An antibody that blocks all β₁ integrin function inhibited thrombin-induced [³H]-thymidine incorporation by 68 ± 4% (mean±se, P<0.001) and blocking antibodies to α₅ integrins resulted in 48 ± 10% (mean±se, P<0.001) inhibition of proliferation (Fig. 6C). Antibodies to integrins not found to mediate attachment of 1321N1 cells to Cyr61 (e.g., β₃, β₅) did not affect thrombin-stimulated cell proliferation (Fig. 6C). Interestingly antibodies blocking α₆ failed to inhibit DNA synthesis, suggesting that α₆β₁ integrins, while capable of binding to Cyr61, do not participate in mitogenic signaling pathways. These data suggests that the integrin that is present on 1321N1 cells and interacts with Cyr61 to participate in thrombin-induced DNA synthesis is α₅β₁.

Finally, to directly demonstrate a role for extracellular Cyr61 as a mediator of thrombin-induced DNA synthesis, we examined the effects of heparin and anti-Cyr61 antibodies on thrombin-stimulated [³H]-thymidine incorporation. Addition of soluble heparin (0.01–10 U/ml) prior to the addition of thrombin caused a dose-dependent inhibition of thrombin-stimulated DNA synthesis (Fig. 7A). Addition of 10 U/ml heparin did not significantly inhibit either FBS or basic fibroblast growth factor-induced DNA synthesis (Fig. 7A), indicating that heparin is not a nonspecific inhibitor of cell proliferation. We also examined the ability of a functionally inhibitory anti-Cyr61 antibody, a gift from Dr. Lester Lau (University of Illinois College of Medicine, Chicago, IL, USA) (51), to affect thrombin-induced DNA synthesis. The anti-Cyr61 antibody (1–20 μg/ml) blocked thrombin-stimulated [³H]-thymidine incorporation with greatest inhibition of thrombin stimulation by 2 μg/ml antibody (Fig. 7B).

Figure 7.

Heparin and anti-Cyr61 antibodies inhibit thrombin-stimulated DNA synthesis. 1321N1 cells were seeded, starved and treated with 0–10 U/ml heparin (A) or 0–20 μg/ml anti-Cyr61 antibody or 20 μg/ml IgG control (B), then stimulated with 0.5 U/ml thrombin or vehicle for 24 h. DNA synthesis was assessed by [³H]-thymidine incorporation as described in Materials and Methods. Data plotted are fold stimulation; values are means + se of 3 separate experiments carried out in triplicate. *P < 0.001 vs. stimulation with thrombin alone.

DISCUSSION

Activation of GPCRs that couple to G_12/13 and RhoA signaling pathways can lead to aberrant cell growth (13,14,15,16,17,18,19,20, 52), and ligands for these receptors are increased under pathophysiological conditions. For example, thrombin is generated at sites of tissue injury and in the tumor microenvironment (2), and levels of LPA and S1P are also altered in and regulate responses of tumor cells (53,54,55). PAR1 expression is also up-regulated in malignant or invasive cancer cells and displays dysregulated signaling (28, 56,57,58,59). Thus, GPCRs linked to RhoA signaling pathways are likely to play a prominent role in cancer and in other forms of pathophysiological cell growth (1, 60). How RhoA activation contributes to altered cell growth has not, however, been fully clarified.

The possibility that Cyr61 induction is a downstream target for GPCR activation of RhoA was suggested by studies from the Chaqour laboratory demonstrating involvement of RhoA/Rho kinase signaling in Cyr61 transactivation by stretch and S1P (30, 37). Work from other laboratories further demonstrated that S1P, as well as LPA and PGF2α, can induce Cyr61, in several cases via RhoA signaling (31,32,33,34). The studies presented here demonstrate that RhoA and Cyr61 gene expression are coordinately regulated in response to thrombin, LPA, and S1P but not other GPCR ligands, and that RhoA signaling pathways mediate agonist-induced Cyr61 expression in astrocytoma cells. The link between RhoA and Cyr61 gene expression could involve a variety of transcription factors, as RhoA signaling pathways have been reported to regulate SRF, NFκB, and Jun/AP-1 (14, 20, 61, 62). Our findings using the Jun kinase inhibitor SP600125 indicate that Cyr61 expression is dependent on c-Jun activation, suggesting that in 1321N1 cells, AP-1 is a critical mediator of thrombin-induced Cyr61.

Cyr61, like all CCN family members, is processed as a secreted protein (26). We demonstrate that extracellular (cell and matrix associated) Cyr61 is increased in 1321N1 cells treated with thrombin, the PAR1 activating peptide, and S1P. Cyr61 can elicit downstream responses through a myriad of pathways initiated through its interaction with integrin receptors (27, 63, 64). Whether integrin activation contributes to the growth stimulatory effects of thrombin or other GPCRs has not, to our knowledge, been previously considered. Our finding that thrombin and other ligands that activate RhoA lead to marked and selective up-regulation of Cyr61 and that Cyr61 is released and binds to the cell suggested that this protein could contribute to thrombin signaling by engaging integrin-mediated pathways.

The study presented here provides several lines of experimental evidence supporting the conclusion that Cyr61 serves as an integrin-dependent mediator of thrombin-induced DNA synthesis. Specifically, we demonstrate that thrombin-induced DNA synthesis is impaired following knockdown of Cyr61 expression with siRNA or shRNA, by addition of heparin to remove Cyr61 from its heparin binding sites, by addition of functionally inhibitory Cyr61 antibodies, and by addition of monoclonal antibodies directed against the cellular integrins through which Cyr61 binds.

Recombinant Cyr61 protein is not generally able to induce mitogenesis but rather appears to act by mechanisms that converge with growth factor signaling pathways (23). Accordingly, we suggest that Cyr61 acts to complement or synergize with signals that thrombin initiates directly through PAR1 activation. In particular, Cyr61 and integrin activation could serve to prolong the cellular actions of thrombin. Both the induction of Cyr61 shown here, and the induction of c-jun and activation of MAP kinase by thrombin shown in our previous work and that of others, are sustained or biphasic responses (refs. 6, 65 and unpublished data). The basis for the prolonged actions of thrombin is not known but autocrine or paracrine effects mediated through Cyr61 activity and integrin activated pathways could be involved. A recently described paracrine effect of Cyr61 is the induction of matrix metalloproteinases in stromal fibroblasts following thrombin-mediated Cyr61 up-regulation in breast cancer cells (29). Autocrine actions of Cyr61 could induce transcriptional up-regulation of chemokine receptors or integrins, or integrin activation of reactive oxygen species, Akt, extracellular regulated kinase, or eicosanoid signaling pathways (27, 63, 66,67,68,69).

The best characterized integrin-mediating Cyr61 binding and signaling is α_vβ₃(41, 70,71,72); however, this integrin is not present on 1321N1 cells. Most recently, other integrins, including α₆β₁, have also been shown to bind Cyr61 (70, 73, 74). The α₆β₁ and the α₅β₁ integrins both appear to be involved in 1321N1 cell adhesion to Cyr61. We find, however, that only the α₅β₁ integrin is critical for the thrombin-induced mitogenic response. The α₅β₁ integrin has not, to our knowledge, been previously identified as a mediator of Cyr61 signaling. Integrin α₅β₁ has, on the other hand, been shown to be involved in cell proliferation via activation of downstream mediators, in particular focal adhesion kinase, a response that is robustly activated by thrombin stimulation in 1321N1 cells (unpublished data). Thus, we hypothesize that α₅β₁ integrin ligation by Cyr61 can lead to activation of intracellular kinase cascades, which converge with those initiated directly by PAR1 activation, permitting thrombin to serve as an efficacious mitogen.

Cyr61 expression has been widely associated with tumorgenicity (75,76,77). Indeed, Cyr61 is a negative prognostic factor for survival in glioblastoma, likely due to the autocrine stimulatory effects it has on cell proliferation and survival (75). A number of glioblastoma cell lines express high levels of Cyr61 (27) and it is tempting to speculate that this may in part underlie the aggressive nature of these tumors. The striking capacity of thrombin, LPA, and S1P to induce Cyr61 suggests that high levels of these ligands imposed on cells that may also have an enhanced capacity to signal to RhoA activation (e.g., up-regulation of G_12/13 or mutations in any of several Rho GAPs/GEFs isolated as oncogenes) would also induce chronic high level Cyr61 expression. We suggest that therapies that disrupt this signaling axis could augment existing approaches for treating glioblastoma, and that Cyr61 may represent a new therapeutic target of interest for disrupting the growth stimulatory effects of GPCR activation.