CSPG4 Enhanced Melanoma Motility and Growth Requires a Cysteine in the Core Protein Transmembrane Domain

Abstract

CSPG4 is a cell surface proteoglycan that enhances malignant potential in melanoma and several other tumor types. CSPG4 functions as a transmembrane scaffold in melanoma cells to activate oncogenic signaling pathways such as focal adhesion kinase (FAK) and extracellular signal regulated kinases 1,2 (Erk1,2), that control motility, invasion and anchorage independent growth. Here, we demonstrate that CSPG4 promotes directional motility and anchorage independent growth of melanoma cells by organizing and positioning a signaling complex containing activated FAK to lipid rafts within the plasma membrane of migrating cells. This FAK-containing signal transduction platform, which consists of syntenin-1, active Src and caveolin-1 requires the cytoplasmic domain of CSPG4 for assembly. Enhanced directional motility promoted by this complex also requires a CSPG4 transmembrane cysteine residue C2230. Substituting C2230 with alanine (CSPG4^C2230A) still permits assembly of the signaling complex, however Src remains in an inactive state. CSPG4^C2230A also fails to promote anchorage independent growth and activation of Erk 1,2. Therapies that target the transmembrane domain of CSPG4 could be a novel strategy for limiting progression by disrupting its function as a compartmentalized motogenic and growth promoting oncogenic signaling node.

INTRODUCTION

CSPG4/NG2 (Chondroitin Sulfate Proteoglycan 4/NG2) is a tumor cell surface proteoglycan that functions as a transmembrane scaffold to activate intracellular signaling pathways. Numerous studies indicate that CSPG4 expression in melanomas and other tumors promotes a malignant tumor cell phenotype as well as resistance to therapy [1–8]. CSPG4 is considered a potential clinical target in melanoma. It is expressed in multiple subtypes of primary melanomas as well as metastases and it has been used a marker to isolate circulating melanoma cells [9–13]. CSGP4 is also highly expressed in progenitor cell populations, and CSPG4 expression is associated with cancer cell stem-like properties [14]. Although CSPG4 is thus a recognized factor in melanoma progression, the mechanism(s) for its oncogenic functions are not completely understood. CSPG4 is not an oncogene in the traditional sense, but rather functions as a co-receptor to support the activation of oncogenic signaling components such as FAK (focal adhesion kinase) and/or BRAF^V600E/MEK/Erk 1,2) which regulate oncogenic pathways that control tumor cell migration, invasion and proliferation.

Structurally, CSPG4 consists of an N-linked type I transmembrane 280 kDa glycoprotein and a 450 kDa chondroitin sulfate proteoglycan component expressed on the plasma membrane of cells. Its protein core performs scaffolding functions for the formation of key signaling complexes that enhance melanoma progression [3,4]. Cell culture analyses show that CSPG4 is located on the tips of leading edge filopodia of migrating cells [15] which is an optimal location for an initial point of cell contact with the extracellular matrix (ECM). CSPG4 associates both directly and indirectly with β1 integrins to enhance cell adhesion and motility [3,4,16]. CSPG4-regulated signaling in radial growth phase melanoma cells also promotes EMT-like phenotypic changes and it promotes tumorigenic potential [17].

CSPG4/NG2 has several key structural domains that mediate its scaffolding/signaling functions. In addition to an association with integrins, the extracellular region of the core protein binds to extracellular matrix (ECM) proteins such as fibronectin, and types I and VI collagen [4,16,18]. It also binds growth factors such as PDGF-AA and bFGF [3,4,16]. The cytoplasmic tail of the core protein contains a carboxyl terminal PDZ-domain binding site that docks one or more PDZ motif-containing proteins and two threonine phosphorylation sites [4,5,19,20]. Furthermore, the CSPG4 core protein contains a transmembrane cysteine (C2230). Cysteine residues within transmembrane or juxta-membrane domains are important for proper localization and function of certain plasma membrane associated proteins [21].

The motility and growth-promoting functions of CSPG4 are dependent on its interaction with the progression associated protein syntenin-1 (mda9) [22,23]. While syntenin-1 promotes melanoma progression by complexing with activated Src to promote FAK phosphorylation (pFAK) [22,23], mechanisms by which this complex is delivered to FAK are incompletely understood. In the present study we demonstrate that CSPG4 forms a complex with syntenin-1/ activated Src/caveolin which functions to activate FAK and stimulate directed cell migration. Additionally, a transmembrane cysteine residue (C2230) in the CSPG4 core protein is also required for CSPG4 enhanced directional migration and for constitutive activation of the BRAF^V600E/pMEK/pErk 1,2) pathway. While CSPG4^C2230A still associates with a complex of syntenin/Src/caveolin, Src within this complex remains in an inactive state and fails to enhance FAK or Erk 1,2 activation. Thus, selective targeting of the transmembrane cysteine in CSPG4 could be an effective and novel strategy for inhibiting oncogenic signaling by the proteoglycan.

MATERIALS AND METHODS

Cell lines

WM1552C radial growth phase, WM1341D vertical growth phase and 1205Lu metastatic human melanoma cells were generously provided by Dr. Meenhard Herlyn (The Wistar Institute, University of Pennsylvania, Philadelphia, PA). Cells were cultured as previously described [17]. The WM1552C^CSPG4 wild type, mutant, and mock stable transfectants were generated as described previously and maintained in medium supplemented with 0.25 mg/ml G418.

DNA constructs

The full-length (CSPG4^WT) and cytoplasmic domain-truncated CSPG4 variant (CSPG4^ΔCD) were generated as described previously [17,24]. CSPG4^ΔPDZ was generated by PCR-site-directed mutagenesis (primers 5’CTGCCCTTAAGAATGGCTAGTACTGGGTGTGAGG 3’, 5’CCTCACACCCAGTACTAGCCATTCTTAAGGGCAG 3’) modifying threonine 2252 to a stop codon. CSPG4^C2230A was generated by site directed mutagenesis, converting cysteine 2230 to alanine (primers 5’ AATCATCCCCATGGCCCTGGTACTTC 3’, 5’GAAGTACCAGGGCCATGGGGATGATG 3’). Mutagenesis was carried out using the QuikChange™ Site-Directed Mutagenesis Kit (Stratagene) using the manufacturer’s suggested protocol. All DNA constructs were verified by sequencing at the Biomedical Genomics Center at the University of Minnesota.

shRNA and siRNA

Lentivirus vector shRNA against CSPG4 that expresses via a doxycycline-inducible promoter in the pTRIPZ vector (mature sequence= ATGCCTTCTAGACTGGAGG) was purchased from Thermo Fisher Scientific. Lentivirus was generated using the Trans-Lentiviral shRNA Packaging System according to manufacturer instructions (Thermo Fisher Scientific). Briefly, expression and packaging vectors were expressed in HEK293T cells. The supernatant was collected 48 and 96 hours after transfection, lentiviral particles were concentrated via ultracentrifugation, and resuspended in phosphate-buffered saline. After lentiviral infection, cell lines stably expressing the pTRIPZ vector were established via selection with 0.5 μg/mL puromycin (Sigma-Aldrich). The anti-CSPG4 siRNA was designed by our lab as previously described [17]. Syntenin-specific siRNA was purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Non-silencing negative control siRNA was purchased from Qiagen.

Antibodies and reagents

The anti-CSPG4 monoclonal antibody 9.2.27 utilized for western blot, immunoprecipitation and flow cytometry in these studies was provided by Dr. Ralph Reisfeld (The Scripps Research Institute, La Jolla, CA). Other antibodies used for IP and western blot were purchased from the indicated companies: Tubulin specific antibody from Oncogene Research Products (Pasadena, CA); anti-β1 integrin, phosphorylated Erk 1,2 (pErk 1,2), total Erk 1,2, Caveolin-1, phosphorylated-Src (pSrc)(pY416), pSrc(pY526), phosphorylated FAK(pY397) (pFAK) and FAK specific antibodies from Cell Signaling Technology, Inc (Boston, MA). Anti-Src and syntenin specific antibodies were from Santa Cruz Biotechnology; normal mouse monoclonal IgG_2a, normal rabbit IgG and goat anti-mouse FC from ICN Pharmaceuticals (Aurora, OH); peroxidase-conjugated goat anti-mouse and goat anti-rabbit secondary antibodies from Jackson ImmunoResearch Laboratories (West Grove, PA). Methyl-β-cyclodextran was purchased from Santa Cruz Biotechnology. Human plasma fibronectin was purified as previously described [25].

Growth in soft agar

A layer of 1% agarose in normal growth media was pipetted into six well plates and allowed to solidify. Cells were suspended in 6.75 ml regular growth media at 5000 cells/ml and incubated for 15 min at 37°C. 750 μl of 2% agarose was then added to the tubes, mixed thoroughly by pipetting, and 2 ml of cell suspension was pipetted into triplicate wells. Plates were placed at 4°C for 15 minutes to facilitate rapid polymerization of the agarose, the wells overlayed with 2 ml growth media and incubated at 37°C/5% CO₂ for 17 days. Media were replaced every three days. Colonies were counted in five random fields/well, and data are shown as the average number of colonies from five fields /well from triplicate wells, +/− s.e.m from replicate experiments.

Migration Assays

Six well plates were coated with fibronectin (5μg/ml) and allowed to dry. Cells were plated at high density (5×10⁵/ml, 70 μl/well) in the wells of tissue culture inserts (Ibidi, Woodstock ON, CA) placed in coated 6 well culture plates and allowed to adhere overnight. Inserts were removed the following day, the wells washed with medium two times to remove loose cells. Images of the area between the wells were collected using a 10x objective at 0, 24, 48, and 72 hours and the cell free area quantitated by tracing the open wound area using Adobe Photoshop™. Bars represent the percentage change in cell free area between the 0 and indicated assay time point, from triplicate wells, +/− s.e.m. from replicate experiments.

Digital Analysis of Time Lapse Videos of Cell Migration

Melanoma cells were grown to confluence in fibronectin coated 12.5 cm² tissue culture flask before scratching with a 1mL micropipette. After scratch wounding, cells recovered for 1 hour at 37°C and 5% CO₂ before live cell imaging was performed. The flask was then sealed and fastened to the heated stage of a Nikon TE-300 microscope equipped with a Hamamatsu digital camera and maintained at 37°C for 18–24 hrs. Time lapse micro-photographic images were taken every 15 minutes over the time indicated using a 10X Nikon air objective lens with Hoffman modulation optics. Analysis of the time lapse cell images was performed using the manual tracking module within the Nikon Elements AR 3.10 software package. Values for distance from origin were plotted against time using the tracking module graphing option. Data are shown for a minimum of 40 cells for each cell line/experiment +/− SD.

Detergent Resistant Membrane Isolates/Sucrose gradient

Cell lysate fractionation was performed using a Caveolae/Rafts Isolation Kit (Sigma), following the manufacturer’s suggested protocol. Briefly, adherent melanoma cells were lysed in 1% triton-x-100 buffer and loaded onto a gradient of Sigma Optiprep Density Medium. Gradients were centrifuged for 4 hours at ~200,000g, and 1 ml fractions removed from the column, top to bottom. All steps were carried out at 4°C. An equal portion of fractions 3–8 was immunoprecipitated with anti-CSPG4 antibody 9.2.27 and analyzed for associated proteins by western blot. Caveolae/Rafts are expected to separate within fractions ~3–5.

Immunoblot and Immunoprecipitation

Immunoblot was performed using standard techniques as described previously [17] For immunoprecipitation, cells were lysed on ice by addition of IP buffer (20 mM Tris-HCl, pH7.5, 150 mM NaCl, 1 mM Na₂EDTA, 1 mM EGTA, 1% Triton, 2.5 mM sodium pyrophosphate, 1 mM β-glycerophosphate, 1 mM NaSO₄, 1 μg/ml leupeptin, 1 mM PMSF) and the insoluble materials were removed by centrifugation. For co-immunoprecipitation experiments, we replaced the 1% Triton -×-100 with 1% Nonidet P-40. Lysates were pre-cleared with protein A/G Sepharose beads (Amersham Pharmacia, Piscataway, NY) for 30 min at 4°C. Antibodies were incubated with the lysates overnight at 4°C, and the immune complexes collected by incubation with protein A/G-Sepharose beads for 1 h at 4°C. Immune complexes were washed three times with lysis buffer at 4°C and the associated proteins resolved on SDS-PAGE followed by immunoblotting with the indicated antibodies.

Flow cytometry

Cells were released in PBS/5mM EDTA solution and washed 2 times with FACS buffer (RPMI media supplemented with 1% goat serum and 5mM HEPES). Cells were incubated with the indicated primary antibody for 45 minutes at 4°C, washed 3 times with FACS buffer, and then incubated with goat anti-mouse phycoerythrin-conjugated secondary antibody for 30 minutes at 4°C. Cells were analyzed on either a BD FACSCalibur (CSPG4 staining) or a BD Biosciences Accuri C6 flow cytometry system (β1 integrin staining).

Statistical analysis

All statistical analysis was performed with Graphpad PRISM 6 (Graphpad Software, San Diego, CA). Data comparing more than two groups was analyzed by one-way analysis of variance utilizing Bonferroni’s adjustment for multiple comparisons. Comparison of two independent samples was done utilizing two-tailed Student’s t test. p<0.01 was considered statistically significant.

RESULTS:

CSPG4 enhances growth, migration and β1 integrin function by assembling a signaling complex of integrin, activated FAK, src, syntenin and caveolin.

Radial growth phase melanoma cells were first evaluated for both a migratory and growth phenotype (Figure 1). Since these cells lack endogenous CSPG4 expression, they were stably transfected with vectors containing wild type CSPG4 (CSPG4^WT), or two truncated CSPG4 constructs that either lack the cytoplasmic domain (CSPG4^ΔCD) or lack the carboxyl terminal PDZ motif binding domain (CSPG4^ΔPDZ) (Supplementary Figure 1). The surface expression of these constructs was evaluated by flow cytometry and the results show similar levels of surface expression (Supplementary Figure 2).

Figure 1. Truncating the CSPG4 cytoplasmic domain inhibits tumor cell migration and growth.

(A) WM1552C cells stably expressing the indicated CSPG4 variants (mock = vector control) were grown to confluence before scratching. The various colored lines represent time lapse digital tracking of an individual cell over the 24-hour period, demonstrating the relative directional motility of each cell. (B) Graphical representation of the average velocity of the individual cells tracked in A. Error bars represent SD, n= minimum of 40 cells/cell line. *p <0.01 compared to WM1552C/CSPG4^WT. (C) Anchorage independent growth assay of the indicated WM1552C transfected cell lines. *p<0.001 vs. WM1552C/CSPG4^WT.

Melanoma cell motility was evaluated using a scratch wound assay on fibronectin-coated surfaces (Figure 1). CSPG4^WT cells exhibited a robust migratory phenotype characterized by a significantly enhanced migration velocity and directional persistence compared to mock transfectants, which were essentially non-motile under these conditions (Figures 1A and 1B). Cells expressing structural mutants of CSPG4 also showed intermediate increases in migration velocities compared to mock transfectants, but in contrast, their migration consisted of a random walk which lacked the directional persistence associated with expression of CSPG4^WT. As we have previously shown [17], melanoma cells expressing CSPG4^WT also exhibit high levels of anchorage independent growth compared to mock transfectants and this requires expression of an intact CSPG4 core protein (Figure 1C). This CSPG4 stimulated growth is depends on the sustained activation of Erk 1,2 leading to increased localization of pErk 1,2 within the nucleus [17].

Previous studies have also linked CSPG4 to enhanced integrin functions, characterized by enhanced cell adhesion, spreading and integrin dependent motility [3–5,17,24,26]. Among other findings these studies have demonstrated the CSPG4^WT is spatially co-distributed with integrin and chondroitin sulfate glycosaminoglycan modification of the extracellular domain interacts with, and promotes, an activated conformation of the α4 integrin subunit [27]. In the current studies, immunoprecipitates of CSPG4^WT from cell extracts include β1 integrin, which was not detected in immunoprecipitates from cells expressing either CSPG4^ΔCD or CSPG4^ΔPDZ (Figure 2A). Integrin mediated FAK phosphorylation (pY397) is also enhanced by CSPG4^WT without a change in the expression level of FAK (Figure 2B). Increased FAK activation was not observed in cells expressing either CSPG4^ΔCD or CSPG4^ΔPDZ (Figure 2B). Both FAK and CSPG4 co-precipitate with Src in WM1341D cells expressing endogenous CSPG4 (Figure 2C). Experiments were next conducted to determine if syntenin, which binds to the PDZ domain binding site of CSPG4 [19] is part of the signaling complex and if it is functionally important for CSPG4^WT enhanced migration and growth (Figure 3). Syntenin co-precipitated with CSPG4^WT and this required the PDZ binding site in the carboxyl terminal end of the protein (Figure 3A and 3B). Syntenin expression was also required for CSPG4^WT to co-precipitate with β1 integrin (Figure 3C) and loss of syntenin expression inhibited CSPG4^WT enhanced cell motility and growth (Figure 3D and 3E). The results demonstrate that syntenin is part of the CSPG4 signaling complex and required for the activated motility and growth functions of CSPG4 in melanoma cells.

Figure 2. CSPG4/β1 integrin complex formation and enhanced FAK phosphorylation.

(A) Western blot of β1 integrin from extracts of adherent cells in normal growth medium immunoprecipitated with CSPG4 antibody 9.2.27 or isotype-matched normal mouse IgG_2a as indicated. (B) Western analysis of phosphorylated FAK in extracts of cells expressing full length CSPG4 or truncated variants CSPG4^ΔPDZ and CSPG4^ΔCD. (C) Endogenous CSPG4 and FAK co-immunoprecipitation with anti-Src antibody in parent WM1552C (CSPG4 negative) and WM1341D (CSPG4 positive) cell lines. Immunoprecipitation with normal rabbit IgG (nrIgG) was included as a negative control.

Figure 3. CSPG4 enhanced melanoma cell growth and migration requires syntenin association.

(A) Co-immunoprecipitation of syntenin with CSPG4 in the indicated cell lines. (B) CSPG4 and syntenin fail to co-precipitate in WM1552C cells expressing truncated variants CSPG4^ΔPDZ or CSPG4^ΔCD. (C) Co-immunoprecipitation of β1 Integrin with anti-CSPG4 antibody in WM1552C/CSPG4^WT cells following knockdown of syntenin via siRNA. (D) Migration assay of fibronectin-adherent WM1552C/CSPG4^WT cells following knockdown of syntenin via siRNA. Cells were transfected with syntenin siRNA 24 hours prior to plating in migration assay. *p<0.01 vs. control siRNA treatment. (E) Anchorage-independent growth assay of WM1552C/CSPG4^WT cells transfected 24 hours prior to plating with syntenin siRNA. *p<0.001 vs. control siRNA treatment.

Transmembrane cysteine residue in the core protein is required for targeting the CSPG4 signaling complex to enhance cell migration and growth.

Since CSPG4^WT contains a transmembrane cysteine at residue 2230, the importance of this residue on CSPG4 stimulated growth and motility was evaluated by replacing C2230 with alanine (CSPG4^C2230A). Melanoma cells stably expressed transfected CSPG4^C2230A at surface levels equivalent to those observed with CSPG4^WT (Supplementary Figure 2) and the surface level of β1 integrin was virtually identical in these two cell lines (Supplementary Figure 3).

Cells expressing CSPG4^C2230A did not exhibit enhanced migration or enhanced integrin-mediated phosphorylation of FAK (Figure 4A and 4B). Our previous work has also demonstrated that CSPG4 is required for promoting anchorage independent growth by sustaining constitutive activation of Erk 1,2 in suspended cells which express BRAF^V600E. Cells expressing CSPG4^C2230A did not demonstrate increased anchorage independent growth (Figure 4C) or sustain activation of Erk 1,2 over minimal levels observed in the mock transfectants (Figure 4D). This indicates that the transmembrane cysteine is required both for enhancing cell migration and for enabling the Erk 1,2 activating function in cells expressing the oncogenic BRAF^V660E mutation, as we have shown previously [24].

Figure 4. Transmembrane cysteine C2230 in CSPG4 is critical for enhanced cell motility and growth.

(A) Inset-Digital tracking of individual cells over the 24-hour period, demonstrating the relative directional motility of each cell. Graph -The indicated cell lines were analyzed by time-lapse digital photography as in 1A. The bars represent the average cell velocity of individually tracked cells over the 24-hour assay period, +/− S.D. (*p <0.01 compared to WM1552C/CSPG4^WT) (B) Cells were serum starved for 48 hours and extracts analyzed for FAK activation (pFAK^Y397) by western blot. (C) Anchorage independent growth assay of the indicated WM1552C transfected cell lines. *=p<0.001 vs. WM1552C/CSPG4^WT. (D) Extracts from the indicated cell lines that were either adherent (0 hr) or held in suspension for 24 hours were analyzed for total Erk1,2 and phosphorylated ERK1,2 (pErk1,2) by immunoblot.

Activation of Src and FAK within the CSPG4 signaling complex requires association with lipid rafts and the transmembrane cysteine in the core protein.

Cysteine residues within transmembrane domains of proteins can impact their function and/or localization within specific lipid microdomains of plasma membranes [21]. Removal of membrane cholesterol using methyl-β-cyclodextrin almost completely inhibited CSPG4^WT stimulated FAK phosphorylation, but it had no inhibitory effect on the levels of CSPG4, β1 integrin, FAK, Src or syntenin (Figure 5A). This demonstrates a requirement for CSPG4 signaling complexes to be associated with lipid rafts to promote integrin-mediated phosphorylation of FAK. The failure of CSPG4^C2230A to stimulate FAK and Erk 1,2 signaling pathways suggested that the transmembrane cysteine in the core protein is required for assembly of the CSPG4/syntenin/Src/caveolin signaling platform. To further determine the composition of localized CSPG4 signaling complexes, density gradient centrifugation was performed on cold Triton-X resistant extracts of melanoma cells expressing CSPG4^WT. Immunoprecipitation of the density gradient fractions demonstrated that a CSPG4/syntenin/Src/caveolin complex was detected in fractions 5 and 6. (Figures 5C). As expected, CSPG4^ΔCD immunoprecipitates from fractions 5, 6 did not contain syntenin or Src or caveolin (data not shown), indicating the cytoplasmic domain of the protein is required for assembly of this complex. These results show that CSPG4^WT forms a membrane associated multi-molecular signaling complex with syntenin, Src and caveolin-1 that requires the cytoplasmic domain of the core protein for assembly. Unexpectedly, the results from performing the same experiment with CSPG4^C2230A expressing cells demonstrate that a CSPG4^C2230A syntenin/Src/caveolin signaling complex still forms (Figure 5B), although it migrates in these gradients with a slightly higher buoyant density (fractions 6 and 7) than that observed for CSPG4^WT (Figure 5C).

Figure 5. Proper lipid raft localization of transmembrane cysteine C2230 in CSPG4 is essential for CSPG4-mediated Src activation.

(A) Following 24-hour induction of CSPG4 shRNA expression, 1205Lu cells that stably express a doxycycline-inducible CSPG4 shRNA were harvested (0 hr) or subsequently treated with methyl-β-cyclodextrin (10μM) for 3 hours. Extracts from these cells were isolated and analyzed by western blot to determine the expression levels of the indicated proteins. (B) and (C) Sucrose buoyant density fraction samples were collected and a portion of each fraction from the top (#3) to bottom (#8) of the density gradient from WM1552C/CSPG4^C2230A (B) and WM1552C/CSPG4^WT (C) were then immunoprecipitated using anti-CSPG4 antibody, electrophoresed and probed as indicated. WCL = whole cell lysate. (D) Extracts from the indicated cell lines were immunoprecipitated with anti-CSPG4 antibody. Western blots of electrophoresed samples were probed for protein expression with the indicated antibodies. Whole cell lysates from each cell line are shown for comparison.

We next asked whether Src could be inactive in the complexes associated with CSPG4^C2230A. Src activation requires the removal of a carboxyl terminal tyrosine phosphate (pY⁵²⁶) from the protein which promotes a conformational change necessary for kinase auto activation [28]. Immunoprecipitation of CSPG4 demonstrated that although CSPG4^WTand CSPG4^C2230A both promoted the formation of a syntenin/Src/caveolin-1 complex, these complexes differed in the activated state of Src (Figure 5D). CSPG4^WT immunoprecipitates contained Src in the activated state as shown by the detection of pY⁴¹⁶ in the active site of the kinase (Figure 5D). By contrast, Src associated with CSPG4^C2230A was inactive as evidenced by the lack of pY⁴¹⁶ and the presence of the inhibitory carboxyl pY⁵²⁶. It is important to note that expression of CSPG4^C2230A did not cause a global change in the activation state of total cellular Src as evidenced by similar ratios of activated to inactive Src within total cell extracts (Figure 5D). This demonstrates that it is the pool of CSPG4-associated Src that fails to be activated, suggesting that Src activation occurs separately from assembly of the complex in specific subcellular domains to promote CSPG4 stimulated oncogenic functions.

DISCUSSION:

There have been extensive studies identifying adhesion related signaling pathways important for promoting a metastatic phenotype in melanoma [29–31]. Despite the numerous insights gained by these studies, how microdomains within the plasma membrane regulate subcellular distribution and activation of oncogenic signaling components within these microdomains is not well understood. Previous studies from our laboratory and others have documented the importance of CSPG4 as a transmembrane scaffold that interacts with, and activates, multiple adhesion and growth factor-related oncogenic pathways [4]. The present study demonstrates a novel mechanism by which CSPG4 functions in lipid rafts to promote directional motility and tumor cell growth by assembling and positioning a molecular complex of syntenin-1, activated Src and caveolin (summarized in Figure 6).

Figure 6: CSPG4 functions as a transmembrane scaffold and oncogenic signaling node.

CSPG4 expression in tumor cells activates multiple oncogenic signaling pathways as a result of its ECM mediated scaffolding function (reviewed in [3,4,16,38]. These pathways include stimulation of Rho family GTPases to promote cytoskeletal reorganization, activation of FAK and downstream pathways to activate integrin mediated signaling, altering the transcriptome to promote mesenchymal transition and interacting with growth factor receptors to stimulate Erk 1,2 and related pathways. CSPG4 function within cholesterol rich lipid rafts requires complexing with syntenin, activated src and caveolin. Mutation of the transmembrane cysteine inhibits CSPG4 mediated FAK activation and Erk 1,2 activation. Details are described in the text.

In melanoma, CSPG4 promotes tumor cell migration by interacting with integrins and promoting FAK activation [3,4,16]. CSPG4 also is required to promote the constitutive activation of the Erk 1,2 pathway in suspended BRAF^V600E expressing cells, linking it directly to a well-recognized and clinically targeted oncogenic signaling pathway [4,24]. Studies have linked CSPG4 to the function of different β1 integrins, demonstrating it can functionally impact adhesion related pathways in the presence of multiple ECM components [3,4,16]. Phosphorylation of two distinct threonine residues by PKCα (T2256) or pErk 1,2 (T2314) on the NG2 homologue stimulate enhances motility and growth, respectively [20]. Finally, several PDZ domain containing proteins such as MUPP1 and syntenin-1 interact with CSPG4, and in the case of syntenin-1 it functions to promote tumor cell motility [19].

Although the importance of these structural domains/residues in the CSPG4 (and the rat homologue NG2) core protein have been well documented, the relevance of the transmembrane domain in promoting its function has not previously been examined. CSPG4/NG2 contain a single cysteine residue (C2230) within the transmembrane domain that our current studies show is critically important for enhancing directional migration. As is the case with the mutation/deletion of the cytoplasmic domain (CSPG4^ΔCD or CSPG4^ΔPDZ), CSPG4 in which cysteine 2230 has been substituted with alanine (CSPG4^C2230A) fails to promote migration, growth and FAK activation. Furthermore, methyl-β-cyclodextrin inhibits CSPG4 stimulated FAK phosphorylation, indicating that CSPG4 function requires its association with lipid rafts in the plasma membrane. Thus, the transmembrane cysteine may localize the CSPG4 core protein to lipid microdomains either by forming disulfide linked intermolecular complexes or being palmitoylated, which would impact core protein/membrane lipid associations (Figure 6) [21].

Intracellular assembly of the CSPG4/syntenin-1/Src/caveolin-1/ complex and activation of Src within this complex are independently regulated. Src activation requires the removal of the inhibitory tyrosine phosphate (pY526) which promotes a conformational change in Src to promote its activation [32]. It is noteworthy that while the syntenin-1/Src/caveolin-1 signaling complex still forms with CSPG4^C2230A, Src within this complex retains the inhibitory pY526, is thus inactive and not available to dock with and/or activate pFAK [32]. Although multiple explanations are possible, the data are consistent with a model in which one or more tyrosine phosphatases function to activate CSPG4-associated Src at the leading edge of migrating cells. One such candidate is the Src activating tyrosine phosphatase (PTPα), which promotes integrin-stimulated FAK autophosphorylation and enhances cell spreading and migration [32]. This impact on Src activation is yet an additional explanation for the functional importance of the transmembrane cysteine of the CSPG4 core protein.

Formation of a CSPG4/syntenin-1/Src/caveolin-1 complex and proper localization in the leading lamellae of migrating cells is required for maintaining a polarized morphology, enhancing FAK activation and promoting directional migration. Although topographical organization and density of ECM can impact directional motility [33], the fibronectin substrate in this study is isotropic. Thus, CSPG4 stimulated directional migration is dependent on intrinsic effects of the proteoglycan on the intracellular machinery and polarized morphology of migrating cells [33]. Since CSPG4 is expressed on the tips of filopodia it is well-positioned to initially bind fibronectin and stabilize membrane protrusions at the leading edge of migrating cells. CSPG4 can impact on the activation of small Rho family GTPases such as cdc42 and Rac1 are implicated in maintaining cell polarity and regulating cytoskeletal reorganization (Figure 6) [26,34].

Caveolin-1 is also a component of this CSPG4 molecular complex based on its co-precipitation with CSPG4^WT from density gradient fractions. Caveolin-1 was originally included in these analyses for the purposes of marking plasma membrane containing fractions and it was unexpected that caveolin-1 would associate with the CSPG4 complex. Caveolin-1 performs multiple functions in plasma membranes related in part to its association with lipid rafts [35] and increased levels of caveolin-1 are associated with increased melanoma cell motility and metastasis [36–38]. Caveolin-1 functions in promoting metastasis by stimulating α5β1 integrin function and regulating the Src-dependent activation of oncogenic pathways including Rac1/FAK and Erk 1,2 [36].

CSPG4 also promotes sustained activation of Erk 1,2 and anchorage independent growth of melanoma cells. We have shown that sustained Erk 1,2 activation can be detected for up to 72 hours when the cells are cultured in suspension [24]. CSPG4 is required to enhance growth, and this requires both the PDZ domain binding motif and syntenin-1, indicating that the associated signaling complex is essential for this effect. The results suggest that surface expression of CSPG4 can maintain a least a minimal cytoskeletal organization in suspended cells which is important for promoting activation of Erk 1,2 oncogenic signaling pathways. It has been shown that NG2, the rat homologue of CSPG4, co-distributes with β1 integrin within domains of the plasma membrane not in contact with an ECM substrate [20]. Thus, anchorage-independent surface clustering of integrin and a CSPG4 signaling complex could be sufficient to maintain an organized cytoskeleton and activation of signaling pathways important for promoting survival/growth in circulating CSPG4 positive melanoma cells [10,13,39].

In summary, CSPG4 is not an oncogene per se but it functions as an organizer or cofactor for the assembly and activation of oncogenic pathways including syntenin-1/Src and BRAF^V600E [3,4,16,17] within specific plasma membrane lipid raft microdomains (Figure 6). In adherent and directionally migrating cells CSPG4 serves to cluster integrins and deliver syntenin-1 and active Src to leading lamellae where it can potentiate FAK activation. The large size and complex structure of CSPG4 make it difficult to target with small molecules, although efforts to target it immunologically have either been reported or are underway [3]. Despite its large size, targeting the specific region of the transmembrane domain could effectively inhibit enhanced activation of both FAK and Erk 1,2, both of which are key CSPG4-linked oncogenic pathways. Transmembrane targeting peptides have been developed that can inhibit Erb2 dimerization, activation and breast tumor growth and metastasis [40]. By analogy, a transmembrane targeting peptide that prevents proper membrane localization of CSPG4 into membrane microdomains could be effective at disrupting its function as an oncogenic signaling node. This could be utilized as a part of a combination therapy for patients with advanced melanoma and for patients harboring other CSPG4 positive tumors [3,4,16].

Supplementary Material

Supplementary Figure 1. Graphical representation of CSPG4 mutant constructs and surface expression of CSPG4 variants in stable transfected WM1552C.

CSPG4^WT – wild type CSPG4. CSPG4^ΔCD –threonine 2252 was converted to a stop codon, leaving the five membrane-proximal amino acids in the cytoplasmic domain (arginine-lysine-arginine-asparagine-lysine) intact. CSPG4^ΔPDZ – converted glutamine 2319 to a stop codon, eliminating the four-amino acid PDZ-binding domain. CSPG4^C2230A – converted the transmembrane cysteine 2230 to alanine.

Supplementary Figure 2. CSPG4 wildtype and mutants display similar expression levels in the stable transfected WM1552C cell lines.

Flow cytometry was performed using CSPG4 antibody 9.2.27 and normal mouse IgG_2a as a negative control. All transfected CSPG4 variants show similar levels of stable cell surface expression in the WM1552C cell line.

Supplementary figure 3. β1 integrin surface expression level is unchanged in WM1552C/CSPG4^WT and WM1552C/CSPG4^C2230A cell lines.

WM1552C cells lines expressing either CSPG4^WT (red line) or CSPG4^C2230A (blue line) were assayed by flow cytometry for β1 integrin surface expression. Unstained cells and cells stained with normal mouse IgG₁ were included as controls for non-specific antibody binding.

SIGNIFICANCE:

The present study demonstrates that lipid raft localized CSPG4 stimulates motility and anchorage independent growth by delivering to the plasma membrane a complex of syntenin, active Src and caveolin-1 to promote FAK and Erk 1,2 activation. This supports a highly novel model in which properly localized CSPG4 within specific lipid microdomains is necessary for CSPG4-associated Src and FAK activation. Thus, targeting the transmembrane domain to inhibit proper membrane localization could be exploited to disrupt CSPG4 function in lieu of targeting multiple structural domains of the proteoglycan.