We recently identified the immunoglobulin-CAM CD155/PVR (the poliovirus receptor) as a regulator of cancer invasiveness and glioma migration, but the mechanism through which CD155/PVR controls these processes is unknown. Here, we show that expression of CD155/PVR in rat glioma cells that normally lack this protein enhances their dispersal both in vitro and on primary brain tissue. CD155/PVR expression also reduced substrate adhesion, cell spreading, focal adhesion density, and the number of actin stress fibers in a substrate-dependent manner. Furthermore, we found that expression of CD155/PVR increased Src/focal adhesion kinase signaling in a substrate-dependent manner, enhancing the adhesion-induced activation of paxillin and p130Cas in cells adhering to vitronectin. Conversely, depletion of endogenous CD155/PVR from human glioma cells inhibited their migration, increased cell spreading, and down-regulated the same signaling pathway. These findings implicate CD155/PVR as a regulator of adhesion signaling and suggest a pathway through which glioma and other cancer cells may acquire a dispersive phenotype.
In glioblastoma, the most aggressive form of brain cancer, tumor cells disperse so extensively that current treatment approaches such as surgical resection or radiation therapy have little effect in checking progression (1, 2). This is a significant problem and it is recognized as a major priority to better understand glioma dispersal and to develop therapies that limit this process (2). A first step in this development is the identification of proteins that function aberrantly in dispersal-relevant processes such as cell migration, which is frequently deregulated in cancer cells (3–6).
Cell migration requires a coordinated orchestration of complex events including polarization, protrusion, adhesion, de-adhesion, and retraction (4). Upon encountering a migration-promoting agent, a cell becomes polarized and will extend protrusions in the direction of migration. These protrusions can take on two forms: large, broad lamellipodia and spike-like filopodia. Both forms are driven by actin polymerization and are stabilized by adhering to the extracellular matrix (ECM) or adjacent cells via transmembrane receptors linked to the actin cytoskeleton through focal adhesion complexes. These adhesion points provide a foundation for traction forces that enable the cell to move forward. Disassembly of adhesion points and retraction at the rear of the cell completes the process, enabling net forward migration (7).
Many cell-surface proteins have been implicated in the regulation of cell migration. Growth factor receptors receive environmental cues and can initiate signaling cascades resulting in polarization and directional migration (6). Adhesion molecules such as integrins, cadherins, and immunoglobulin family proteins can regulate adhesion and de-adhesion between a cell and its neighbors or the ECM and can also contribute to polarization and directional motility in response to the ECM (8–10). Adhesion molecules connect to signaling cascades that regulate actin and microtubule dynamics and ultimately control cell behavior (8, 10). The expression of adhesion molecules is frequently deregulated in cancer, resulting in changes to downstream signaling pathways and enhanced motility (5).
We previously reported the identification of CD155/PVR (the poliovirus receptor), as a regulator of cancer cell invasiveness through a functional proteomic screen (11). CD155/PVR is a member of the immunoglobulin family of cell adhesion molecules, first identified based on its ability to mediate the binding of poliovirus to host cells (12). Little is known, however, regarding its endogenous function in cells. We previously found that CD155/PVR was highly expressed in both U87 human glioma cells and primary glioblastoma tumor tissue, and that inactivation of CD155/PVR reduced cell migration in vitro (11). These findings suggested a novel role for CD155/PVR in regulating motility and prompted us to explore its contribution to glioma dispersal in more detail.
Here, we confirm the role of CD155/PVR in regulating glioma dispersal, both in vitro and on primary brain tissue. We report that expression of CD155/PVR in rodent glioma cells that normally lack the protein results in increased motility and substrate-dependent reductions in adhesion and cell spreading. Conversely, we show that depletion of CD155/PVR from human glioma cells that express high levels of the protein results in decreased migration and increased cell spreading. Furthermore, we show that CD155/PVR can modulate an adhesion-induced Src/focal adhesion kinase (FAK)/paxillin/p130Cas signaling pathway to destabilize focal adhesions at the cell-substrate interface.
Materials and Methods
Cells, reagents, and antibodies. All reagents were obtained from Invitrogen (Carlsbad, CA) unless otherwise indicated. 9L and CNS-1 rat glioma cells, described elsewhere (13), were obtained from American Type Culture Collection (Manassas, VA) and cultured in DMEM (9L) or RPMI (CNS-1) supplemented with 10% FCS (Hyclone, Logan, UT) and 100 units/mL penicillin/streptomycin. A172 human glioma cells, described elsewhere (14), were obtained from American Type Culture Collection and cultured in DMEM supplemented with 10% FCS and 100 units/mL penicillin/streptomycin. Purified vitronectin was obtained from Promega (Madison, WI); purified fibronectin and collagen-1 were obtained from BD Biosciences (San Jose, CA); bovine serum albumin (BSA) was from Sigma (St. Louis, MO); anti-CD155/PVR (clone D171), used for immunocytochemistry, was obtained from Lab Vision (Fremont, CA); anti-CD155, used for immunoblotting, was a kind gift from Dr. Eckard Wimmer; antivinculin (clone 7F9) was from Chemicon (Temecula, CA); AlexaFluor 594 phalloidin and AlexaFluor 488 goat anti-mouse IgG secondary antibody were from Molecular Probes (Eugene, OR); anti-phospho-Src (Tyr416), anti-phospho-paxillin (Tyr118), anti-phospho-p130Cas (Tyr165), anti-phospho-CrkII (Tyr221), anti-phospho-ERK1/2 (Thr202/Tyr204), and anti-phospho-Akt (Ser473) were obtained from Cell Signaling Technology (Beverly, MA); anti-phospho-FAK (Tyr397, clone 14) was from BD Biosciences; anti-β-actin (clone AC-15) was from Sigma. Peroxidase-conjugated antimouse and antirabbit IgG secondary antibodies were obtained from Cell Signaling Technology.
Expression constructs, small interfering RNA, and transfection. A full-length wild-type human CD155/PVR cDNA (accession no. BC015542) in the pCMV-SPORT6 expression vector was purchased from Invitrogen and confirmed by sequencing. The enhanced green fluorescent protein expression vector pEGFP-N1 was obtained from BD Biosciences. Expression constructs were transfected into 9L cells using LipofectAMINE 2000 following the manufacturer's instructions using 0.5 μg DNA per 35 mm dish. Cells were incubated with DNA in OptiMEM for 6 hours after which time normal growth medium was added. Cells were then incubated for 24 to 48 hours to achieve robust expression. Control cells were subjected to identical handling but did not receive DNA. We previously reported the development of a small interfering RNA duplex that efficiently knocked down CD155/PVR protein expression (11). Briefly, a double-stranded siRNA oligonucleotide targeting CD155/PVR (5′-CAA-CUU-UAA-UCU-GCA-ACG-UdTdT-3′) was chemically synthesized (Dharmacon, Lafayette, CO) and transfected into A172 cells using OligofectAMINE following the manufacturer's instructions using 200 nmol/L siRNA per 35 mm dish. Cells were incubated with siRNA in OptiMEM for 6 hours after which time normal growth medium was added. Cells were then incubated for 72 hours to achieve >80% knockdown of CD155/PVR. Control cells were transfected with a scrambled siRNA oligonucleotide at matching concentrations.
Adhesion signaling assay and immunoblotting. Mock- or CD155-transfected cells were grown for 48 hours in 100 mm tissue culture dishes. Cells were detached with Versene and 1 × 106 cells plated onto 100 mm culture dishes previously coated with vitronectin (2 μg/mL), fibronectin (2 μg/mL), or PBS and blocked overnight with 1% BSA in PBS. Cells were allowed to adhere at 37°C for 90 minutes and then washed with room temperature PBS and incubated in NP40 lysis buffer [0.2% NP40, 20 mmol/L Tris (pH 7.5), 150 mmol/L NaCl, 10% glycerol] supplemented with protease inhibitors (Roche, Nutley, NJ) and phosphatase inhibitors (Sigma) for 30 minutes at 4°C. Lysates were collected by scraping, cleared by centrifugation for 15 minutes at 10,000 × g and quantified using the DC Protein Assay (Bio-Rad, Hercules, CA). Lysate (30 μg) was separated on a 10% SDS-PAGE gel and transferred to a nitrocellulose membrane. Membranes were blocked in 5% nonfat dry milk in PBS and probed with primary antibody diluted in 5% BSA in PBS overnight. Antibody binding was detected with peroxidase-conjugated secondary antibodies and visualized using enhanced chemiluminescence substrate (Perkin-Elmer, Boston, MA).
Cell morphology and immunocytochemistry. After transfection, cells were detached with Versene and 1 × 104 cells were deposited on glass coverslips that had been previously coated with 2 μg/mL vitronectin, fibronectin, or collagen-1 and blocked with 1% BSA in PBS. Cells were incubated for 2 hours at 37°C, fixed with 4% paraformaldehyde in PBS, permeabilized with 0.1% Triton X-100 and blocked with 1% BSA in PBS. Cells were incubated with primary antibody for 1 hour at room temperature and washed extensively in PBS. Appropriate species-specific secondary antibodies conjugated to AlexaFluor 488 or 594 were used to visualize antibody staining.
Substrate adhesion assay. Cells were detached with Versene and 2 × 104 cells were deposited into replicate wells of a microtiter plate that had been previously coated with 2 μg/mL vitronectin or collagen-1 and blocked with 1% BSA in PBS. Cells were incubated for 90 minutes at 37°C, briefly shaken at 200 rpm, washed twice with medium, and fixed with 4% paraformaldehyde in PBS. Attached cells were stained for 1 hour with crystal violet (5 mg/mL in 2% ethanol), washed extensively with water, and allowed to dry overnight. Stained cells were solubilized in 50 μL of 2% SDS and absorbance was measured on a SpectraFluor Plus automated plate reader (Tecan, Research Triangle Park, NC).
Transwell migration assay. Migration assays with 9L and A172 cells were done as described previously (11). Briefly, cells were transfected with cDNA for 24 hours, or siRNA for 72 hours, prior to the assay and maintained at subconfluence under normal growth conditions. On the day of the assay, cells were labeled with cell tracker orange (Molecular Probes), and 5 × l04 cells were loaded onto replicate 8 micron, 96-well Fluoroblok membranes (BD Biosciences) with 5% serum chemoattractant in the bottom chamber. The membrane apparatus was incubated at 37°C and cell migration to the underside of each light-impermeable membrane was quantified at various time points using an automated plate reader to measure fluorescence, and was confirmed by visualization using an inverted fluorescence microscope. In a subset of experiments, the membrane surface was coated with 2 μg/mL of vitronectin for 2 hours at room temperature, washed, and blocked with 1% BSA in PBS before loading cells.
Cell proliferation assay. Cell proliferation was measured using the CellTiter 96 AQueous Assay (Promega) following the manufacturer's instructions. Briefly, 2,000 cells from various treatment groups were plated in replicate wells of a 96-well plate and grown overnight. CellTiter 96 AQ reagents were added to a subset of wells at 0, 24, 48, and 72 hours, incubated for 2 hours at 37°C, and absorbance was quantified on an automated plate reader.
Brain slice dispersal assay. The use of organotypic brain slice culture to model glioma dispersal has previously been described in detail (15). All studies were approved by the Yale University Institutional Animal Care and Use Committee. Briefly, brains from postnatal day 2 CD-1 mice (Charles River Laboratories, Wilmington, MA) were obtained by rapid decapitation of anesthetized pups and removal of the brain into ice-cold HBSS containing penicillin, streptomycin, and amphotericin B. The meninges were removed and 350 μm coronal slices prepared using a McIlwain Tissue Chopper vibratome (Brinkmann Instruments. Westbury, NY). Brain slices were dissociated in HBSS and placed on the porous membrane of a 0.4 mm Millicell-CM insert (Millipore, Billerica, MA) in a 35 mm culture dish containing 800 μL of medium (50% Neurobasal-A, 25% HBSS, 20 mmol/L HEPES, 10% FCS, supplemented with B-27, G5, penicillin, streptomycin, and amphotericin B) below the insert. All media above the insert were removed. Brain slice cultures were incubated at 37°C, 5% CO2 in a humid incubator. A single-cell suspension of 9L or CNS-1 cells was prepared and transiently transfected with 0.5 μg full-length CD155/PVR cDNA and 0.5 μg green fluorescent protein (GFP) cDNA, or GFP alone, using LipofectAMINE 2000. The cell suspension was seeded into six-well culture plates coated with a thin layer of sterile 1% GPG agarose (American Bioanalytical, Natick, MA) dissolved in growth medium and incubated at 37°C for 24 hours to encourage the formation of tumor cell spheroids. Using a dissecting microscope, one spheroid was removed from the culture using a pulled glass capillary pipette and gently deposited onto individual brain slices. The initial size of the tumor spheroid was recorded by fluorescence microscopy of the GFP-transfected cells and measured using NIH Image (NIH, Bethesda, MD). The total area of cell migration was subsequently measured at 24, 48, and 72 hours.
CD155/PVR expression enhances glioma cell migration. We addressed the role of CD155/PVR in glioma migration and dispersal by expressing it in 9L rat glioma cells that normally lack the protein, as CD155/PVR expression is restricted to primates (16, 17). 9L cells are nondispersive in primary brain tissue (13) and provide a convenient cell system to study the effects of exogenous CD155/PVR expression. Expression of CD155/PVR by 9L cells after transfection was confirmed by immunocytochemistry and Western blot (Fig. 1A,, inset). 9L cells expressing CD155/PVR exhibited increased transwell migration towards a serum chemoattractant compared with mock-transfected controls at both 6 and 24 hours after plating (Fig. 1A). This difference was not due to changes in cell proliferation as the growth kinetics of 9L cells were unaffected by CD155/PVR expression over 72 hours (Fig. 1B). These results indicate that CD155/PVR expression is sufficient to up-regulate 9L cell migration in vitro. CD155/PVR has been reported elsewhere to interact with the ECM protein vitronectin (18), but the functional significance of this interaction has been unclear. To explore this link, we tested the effects of adding a thin layer of vitronectin to the top of each transwell membrane and found that this enhanced the ability of CD155/PVR to potentiate transwell migration over 12 hours (Fig. 1C). This result suggested a substrate dependence for CD155/PVR function.
CD155/PVR expression reduces substrate adhesion. Given the effects of CD155/PVR on cell migration, we were curious as to whether it might also play a role in substrate adhesion. We allowed mock- or CD155-transfected 9L cells to adhere to microtiter plate wells coated with vitronectin or collagen, and then challenged them to remain attached through several washes. Mock-transfected 9L cells adhered readily to vitronectin-coated wells but adhered poorly to collagen-coated wells. Expression of CD155/PVR in these cells markedly reduced their adhesion to vitronectin (Fig. 2A). In contrast, CD155/PVR did not alter the ability of 9L cells to adhere to collagen. These results suggest a substrate-dependent role for CD155/PVR in regulating adhesion.
CD155/PVR expression reduces cell spreading and destabilizes focal adhesions. We further explored the substrate dependence of CD155/PVR function by assessing the behavior of mock- or CD155-transfected 9L cells on glass coverslips coated with purified ECM proteins. In CD155-transfected 9L cells adhering to vitronectin, CD155/PVR localized to the cell membrane and was recruited to peripheral sites of active membrane ruffling (Fig. 2B), a finding consistent with our previous observations of endogenous CD155/PVR localization in a number of human cancer cell lines (11). Two hours after seeding, 9L cells expressing CD155/PVR were much smaller and less spread on a vitronectin substrate than controls (Fig. 2C). Mock-transfected cells developed many large focal adhesions (Fig. 2C) and a dense meshwork of actin stress fibers (Fig. 2D). In contrast, these structures were seldom observed in 9L cells expressing CD155/PVR (Fig. 2C and D). These observations were consistently made when cells were attached to either vitronectin or fibronectin (data not shown) but not in cells adhering to collagen (Fig. 2C). On collagen, 9L cells exhibited similar morphology and similar focal adhesion density in the presence or absence of CD155/PVR. Taken together, these results suggest a substrate-dependent role for CD155/PVR in regulating adhesion dynamics and, in particular, the formation of mature focal adhesions linked to the actin cytoskeleton.
CD155/PVR expression enhances adhesion signaling. Given the observed effect of CD155/PVR on focal adhesions, we hypothesized that CD155/PVR might alter the dynamics of focal adhesion complexes such that cells could not develop mature, stable adhesions. CD155/PVR might inhibit the formation of focal adhesions or enhance their turnover. Upon integrin binding to substrate proteins in the ECM, FAK and Src are recruited into developing focal adhesions. FAK is autophosphorylated at Tyr397, exposing an SH2 binding site that recruits and activates Src via a conformational change (19). After Src phosphorylates FAK, the active Src/FAK complex recruits and activates numerous downstream targets including paxillin and p130Cas (20). Activation of paxillin has been shown to be required for the disassembly of focal adhesions (21) and efficient motility (22). Upon activation, both paxillin and p130Cas have been shown to recruit and activate Crk, which can subsequently recruit Dock and activate Rac, eventually leading to increased membrane ruffling and lamellipodial extension, processes required for efficient migration (23, 24).
We asked whether expression of CD155/PVR in 9L cells might alter the activation of focal adhesion complex proteins after substrate adhesion. Mock- or CD155-transfected 9L cells were challenged to adhere to tissue culture dishes coated with ECM substrate proteins for 90 minutes before lysates were prepared. Immunoblot analysis revealed that expression of CD155/PVR resulted in increased phosphorylation of Src, FAK, paxillin, p130Cas, and Crk after adhesion to vitronectin (Fig. 3). The phosphorylation of other signaling proteins including ERK1/2 and Akt were not affected by CD155/PVR expression. For all signaling proteins analyzed, the total levels of each protein were unchanged (data not shown). Similar results were observed at 30 minutes after adhesion and in cells adhering to fibronectin (data not shown). In 9L cells adhering to uncoated tissue culture plastic, we observed much smaller changes in protein phosphorylation due to CD155/PVR expression, with the exception of paxillin, which continued to exhibit a large increase in phosphorylation after adhesion in cells expressing CD155/PVR (Fig. 3). These findings implicate CD155/PVR in modulating a signaling pathway known to control focal adhesion turnover and migration (20). Although these findings suggest a substrate-specific requirement for enhanced adhesion-induced activation of Src, FAK, p130Cas, and Crk, they also suggest a substrate-independent link between CD155/PVR and paxillin.
CD155/PVR expression enhances glioma dispersal on primary brain tissue. To establish the physiologic importance of CD155/PVR in regulating glioma dispersal, we employed an ex vivo brain slice assay. This assay more closely mimics in vivo dispersal because glioma cells must migrate on primary brain tissue (15). Mock- and CD155-transfected 9L cells were cotransfected with GFP and grown in agar to generate spheroids. These spheroids were deposited onto coronal brain sections prepared from postnatal day 2 CD-1 mice as described in Materials and Methods. The radius of dispersal from each spheroid was measured at 0, 24, 48, and 72 hours. Mock-transfected 9L control cells, which are noninvasive in vivo (13), did not disperse far beyond their origin in this assay. In contrast, CD155-transfected 9L cells migrated extensively from their origin (Fig. 4A and B). The extent of CD155-expressing 9L cell dispersal was nearly equivalent to that observed with wild-type rat CNS-1 glioma cells (Fig. 4A), which disperse readily in vivo (13). These results indicate that CD155/PVR expression is sufficient to convert a cell that is nondispersive on primary brain tissue into one that readily migrates.
CD155/PVR depletion inhibits the migration of human glioma cells. To test whether our findings were predictive of the role of endogenous CD155/PVR that is frequently expressed by human cancer cells, we used the A172 human glioma cell line that expresses high levels of CD155/PVR. We previously reported the generation of siRNA duplexes specifically targeting CD155/PVR (11) and found that these could efficiently deplete CD155/PVR from A172 cells (Fig. 5A,, inset). Knockdown of CD155/PVR by RNA interference in A172 cells resulted in reduced transwell migration compared with cells transfected with a scrambled siRNA control at 6 and 24 hours after seeding (Fig. 5A). This result is consistent with our previous observation that knockdown of CD155/PVR by fluorophore-assisted light inactivation reduced the transwell migration of U87MG glioma cells (11). The observed inhibition of A172 migration was not due to changes in proliferation as depletion of CD155/PVR did not alter the growth kinetics of these cells over 72 hours (Fig. 5B). These results confirm a role for CD155/PVR in enhancing the migration of human glioma cells.
CD155/PVR depletion increases cell spreading and reduces adhesion signaling. Depletion of CD155/PVR also affected the morphology of A172 human glioma cells. On vitronectin, normal A172 cells typically take on a stellate morphology after 2 hours, with several points of active membrane ruffling stretching the cell in multiple directions (Fig. 5C). CD155/PVR localizes to these sites of active ruffling (Fig. 5C,, insets). In contrast, CD155-depleted cells were more spread and exhibited a much rounder and more uniform morphology that lacked focal areas of membrane ruffling (Fig. 5C), suggesting that CD155/PVR might normally function to disrupt substrate adhesion in these cells.
CD155-depleted A172 cells were also found to show reduced activation of Src, paxillin, p130Cas, and Crk compared with control cells after adhesion to vitronectin (Fig. 5D). The activation states of Erk and Akt were unaffected by depletion of CD155/PVR. The total levels of each signaling protein examined were unchanged (data not shown). In contrast to our observation with expression of exogenous CD155/PVR in 9L cells, FAK phosphorylation at Tyr397 did not seem to be significantly affected by depletion of CD155/PVR in these human glioma cells. Similar results were observed using another human glioma cell line, U87MG (data not shown). This may suggest that CD155/PVR functions downstream of FAK. Residual CD155/PVR may be sufficient for FAK phosphorylation at this site or phosphorylation may be maintained by other proteins in these cells. Importantly, CD155/PVR depletion consistently inhibited the adhesion-induced activation of two established targets of the Src/FAK signaling complex, paxillin and p130Cas.
Together, these results implicate CD155/PVR in regulating the migration of human glioma cells and suggest that this is accomplished by the modulation of an adhesion-induced Src/FAK/paxillin/p130Cas signaling pathway known to regulate focal adhesion turnover and membrane ruffling (20). Taken together with our earlier observations using exogenous expression of CD155/PVR in 9L cells, these results strongly implicate CD155/PVR in glioma dispersal and suggest a mechanism through which this may be accomplished.
Our findings suggest that up-regulation of CD155/PVR expression may be one mechanism through which cancer cells can acquire an invasive phenotype. These results show that CD155/PVR can alter adhesion dynamics, resulting in fewer mature focal adhesions in cells adhering to vitronectin or fibronectin substrates, and suggest that this is achieved through modulation of a signaling pathway implicated in the turnover of focal adhesions (20). Cell migration is a highly regulated process that is required for many aspects of normal and disease biology (3–6). In normal cells, the continuous formation and disassembly of focal adhesions is a tightly controlled, dynamic process essential for migration (21), and the deregulation of this process has been linked to tumor invasion (25). Migration plays a key role in the invasion and dispersal of tumor cells, and proteins that regulate migration are often up-regulated in these cells (5). We have now established that CD155/PVR can promote the dispersal of glioma cells both in vitro and in primary brain tissue. Given that we initially identified CD155/PVR in a screen for fibrosarcoma invasiveness (11), and that CD155/PVR is frequently expressed at high levels in a variety of cancers (11, 26), it seems that this protein may also play a critical role in the migration and invasion of other cancers. Our findings link cellular behavior to molecular signaling events and provide mechanistic insight into the function of CD155/PVR.
We found that CD155/PVR could enhance the adhesion-induced activation of Src, FAK, paxillin, and p130Cas, all central effectors of adhesion signaling. Upon integrin-mediated substrate adhesion, an activated Src/FAK complex activates paxillin and p130Cas (20). Paxillin is required for the disassembly of focal adhesions (17) and for efficient motility (18). Both paxillin and p130Cas have been shown to recruit and activate Crk (24, 27), which can subsequently recruit the guanine exchange factor Dock180 and activate Rac, resulting in increased membrane ruffling and lamellipodial extension, processes required for efficient migration (20, 28, 29). We determined that enhanced activation of this pathway by CD155/PVR correlated with fewer mature focal adhesions, reduced cell spreading, reduced substrate adhesion, and increased cell migration both in vitro and on primary brain tissue.
How might CD155/PVR regulate the Src/FAK/paxillin/p130Cas pathway? Given the substrate-dependence of the observed CD155/PVR effects, we speculate that CD155/PVR could exist in a tripartite complex with an ECM protein such as vitronectin and an integrin α/β pair (Fig. 6). CD155/PVR has been shown to interact with vitronectin (18) and we have observed a colocalization of CD155/PVR with αv-integrin in cancer cells (11). As such, vitronectin could act as an extracellular scaffold to bring CD155/PVR, integrins, and their associated proteins into close proximity. Once recruited to a nascent focal adhesion, CD155/PVR could interact directly with members of the Src/FAK/paxillin/p130Cas signaling cascade or could recruit other proteins into focal complexes.
Our finding that depletion of CD155/PVR from human glioma cells reduced the adhesion-induced activation of Src, paxillin, and p130Cas, but not FAK, favors a link between CD155/PVR and Src. Recently, the protein phosphatase Shp2 has been reported to interact with CD155/PVR (30), raising the possibility that Shp2 recruited to a focal complex could alter the phosphorylation state of complex members. For example, dephosphorylation of Src at Tyr527 is required to relieve autoinhibition and permit its maximal activation (31). Recruitment of Shp2 by CD155/PVR into a developing focal adhesion could increase Src activity by relieving repression and subsequently lead to enhanced activation of Src/FAK targets including paxillin and p130Cas, ultimately resulting in increased motility.
Alternatively, CD155/PVR could interact with downstream focal complex proteins such as paxillin, an idea supported by our observation that paxillin phosphorylation was highly up-regulated by CD155/PVR expression in a manner independent of substrate availability (Fig. 3). Although we have been unable to coimmunoprecipitate paxillin or Shp2 with CD155/PVR, the possibility of an indirect association remains. Enhanced activation of paxillin or increased recruitment of paxillin into nascent focal adhesions would likely result in increased focal adhesion turnover and enhanced motility (21, 22). Future work will seek to define the interactions between CD155/PVR and focal complex members to better understand how CD155/PVR regulates adhesion signaling.
In summary, we have now extended our earlier studies by demonstrating a key role for CD155/PVR in regulating glioma dispersal. In addition, we have identified a novel mechanism through which CD155/PVR can regulate cancer cell migration. We propose that CD155/PVR may control migration in a substrate-dependent manner by decreasing the stability of focal adhesions at the cell-substrate interface, resulting in cells that are less adherent and more motile. CD155/PVR is up-regulated in many cancers (11, 26, 32). Thus, up-regulation of CD155/PVR expression may be a common pathway through which cancer cells can acquire a more invasive or dispersive phenotype, a requirement for cancer progression (33). CD155/PVR is not normally expressed in the adult brain but has been detected at high levels in glioma patient tissue by immunohistochemistry (11) and microarray analyses (34). Given its accessibility as a surface protein and its differential tissue expression, CD155/PVR may have potential as a therapeutic target for drugs to limit the dispersal of brain tumor cells or the invasion of other cancer cell types.
Grant support: National Cancer Institute grant (CA-81668; D.G. Jay), Goldhirsh Foundation grant (D.G. Jay), and a training grant from the NIH (DK-07542; K.E. Sloan).
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
We thank R. Hauptschein, L. Liscum, and S. Schonhoff for helpful discussions, and E. Wimmer for kindly providing the polyclonal antibody to CD155/PVR.