Suppression of Prostate Cancer Pathogenesis Using an MDA-9/Syntenin (SDCBP) PDZ1 Small-Molecule Inhibitor Free

Despite significant success in clinically managing localized prostate cancer through surgical, radiation and chemotherapeutic approaches, metastatic prostate cancer remains essentially incurable (1). Numerous molecular events, including intracellular-regulated and/or extracellular tumor microenvironment-mediated are driving forces in controlling prostate cancer progression (1, 2). Consequently, understanding disease pathogenesis and developing rationally targeted therapies are mandatory to develop treatments that are potentially curative for advanced prostate cancer.

Melanoma differentiation associated gene-9 (mda-9) Syndecan-binding protein (SDCBP; refs. 3, 4), also known as Syntenin-1 (5), cloned by us in 1996, has recently achieved considerable interest for its central pathogenic role in multiple diverse cancers (6–10). MDA-9/Syntenin expression positively correlates with prostate cancer progression (11), suggesting that this gene/protein might provide a suitable therapeutic target. MDA-9/Syntenin physically interacts with Insulin-like growth factor receptor (IGF-1R) activating the STAT3 pathway, and turning on a plethora of downstream effector proteins that facilitate invasion and angiogenesis. The relevance of IGF-1R (12) and STAT3 (13) activation in prostate cancer progression are well established. Hence, our recent work defined a novel molecular pathway whereby MDA-9/Syntenin plays a decisive role in IGF-1R–mediated STAT3 activation in the context of prostate cancer (11).

Functionally, MDA-9/Syntenin is an adaptor protein that physically interacts with selective binding partners such as IGF-1R (11), c-Src (14), EGFR (15), TGF-βR (16), and TGF-β (17), thereby stimulating downstream cancer-context–specific signaling pathways. These interactions result in induction of metastasis-related phenotypes including invasion (14, 18, 19), migration (18), angiogenesis (20) and epithelial–mesenchymal transition (16, 17). In addition to its role in cancer, the consequences of MDA-9/Syntenin and its' binding partners in different pathological conditions is also well documented (21–23). With few exceptions [Ubiquitin (24), TGF-βR (16)] two highly conserved PDZ domains (PSD95/SAP90, DLGA, and ZO-1), 113-275 amino acids, represent nodal hubs for binding partners (14, 15). In these contexts, designing small molecules that specifically target the PDZ domains, thereby disrupting interactions or altering signaling, would in principle have enormous potential in blocking MDA-9/Syntenin-mediated phenotypes, such as invasion, migration, and angiogenesis.

There is significant interest in targeting the PDZ class of proteins with some initial successes using small peptides, natural products, and small molecules (25). These initial accomplishments are noteworthy and suggested that the PDZ domains may in fact be “druggable.” However, there were several challenges associated with targeting PDZ domains for therapeutic purposes, because they are involved in multiple protein-protein interactions, some of which may be essential for normal cellular functions. Nonetheless, we noted that although there are over 150 PDZ domains in the human genome discovered thus far, their binding surfaces are likely distinct and with different substrate specificities (26). We have confirmed the hypothesis that the PDZ domains are in fact “druggable” and by using a combination of Fragment- and NMR-based drug discovery (FBDD) approaches we derived novel initial pharmacological tools targeting MDA-9/Syntenin (27). These strategies enabled the identification and initial optimization of a small-molecule capable of targeting and antagonizing MDA-9/Syntenin in vitro, in cell cultures through its PDZ1 domain (PDZ1i; ref. 27). Interestingly, these small molecules do not target the PDZ2 domain of MDA-9/Syntenin, demonstrating that despite the similar global fold, these domains tend to have fairly distinct binding surfaces (27). These findings are intriguing and form the basis for developing new classes of PDZ inhibitors and the approaches would be paradigm shifting and pave the way in the future for an entire new field of research for targeting other therapeutically relevant PDZs.

In the present study, we evaluated the effect of inhibiting MDA-9/Syntenin functions with PDZ1i on prostate cancer cell invasion, downstream biochemical changes, tumor angiogenesis, and tumor cell retention in the lungs, and formation of lung and bone metastases. Our studies demonstrate that PDZ1i blocks key interactions between IGF-1R and MDA-9/Syntenin thereby resulting in inhibition of STAT3 and concomitant blocking of production of angiogenic factors and matrix degrading enzymes, culminating in an inhibition of metastasis. These findings suggest that PDZ1i may represent a useful small-molecule inhibitor capable of reducing prostate cancer pathogenicity.

A detail description of synthesis protocol and properties were previously described (27). Additional details presented in the “Supplementary Documents” section.

All cells except ARCaP with its metastatic variant ARCaP-M, were obtained from the ATCC and maintained in culture as per ATCC recommendations. ARCaP and ARCaP-M cells were obtained from Novicure Biotechnology (Birmingham, AL) and maintained in media as recommended by the provider. Primary immortal prostate epithelial cells (RWPE-1) were purchased from the ATCC. HUVEC (Human Umbilical Vein Endothelial Cells) were obtained from Lonza. All cell lines were routinely checked for Mycoplasma contamination by using commercial kits. The majority of experiments used RWPE-1, DU-145, and ARCaP-M cells. All of these cell lines were purchased recently (within the last 3 years) and were strictly maintained as recommended by the manufacturer. PC-3ML-Luc cells were obtained from Dr. M.G. Pomper (Johns Hopkins Medical Institutions, Baltimore, MD) and maintained as previously described (28). The androgen-refractory mouse prostate cancer cell line RM1 was provided by Dr. T.C. Thompson (Baylor College of Medicine, Houston, TX) and was maintained in DMEM as previously described. The metastatic capacity of stable luciferase expressing RM1 cells (RM1-Luc) has been reported previously (29).

MDA-9/Syntenin (SDCBP) antibody was obtained from Abonova Inc. (Taiwan). Phospho-IGF-1R (Tyr1135), IGF-1R, Phospho-Src (Tyr416), Src, Phospho-FAK (Tyr397), FAK, Phospho-STAT3 (Tyr 705), and STAT3 antibodies were purchased from Cell Signaling Technology. β-Actin and EF-1α were from Sigma-Aldrich and EMD Milipore, respectively, and were used as loading controls for the different experiments. All reagents for cell cultures, including media and serum, were purchased from Thermo Fisher Scientific. Recombinant IGFBP-2 was purchased from R & D Biosystems.

Gelatin zymography was used to determine the gelatinolytic activity of MMP-2 and MMP-9 in conditioned media collected from in vitro cell cultures. The experimental protocol is described in the “Supplementary Document” section.

For qPCR, total RNAs were extracted using miRNeasy kits (Qiagen) as recommended by the manufacturer and cDNA was prepared as previously described (20). Quantitative qPCR was performed using an ABI ViiA7 fast real-time PCR system and TaqMan gene expression assays according to the manufacturer's protocol (Applied Biosystems).

Various vector constructs used in this study were either cloned by our group (mda-9, shmda-9; REF. 20) or obtained from commercial vendors. As described previously (20), cells were transfected with an expression vector producing luciferase and selected for neomycin resistance for approximately 2 weeks. Individual colonies (clones) were picked and analyzed for luciferase expression.

All in vivo experiments were performed in accordance with IACUC approved protocols. To determine the effect of PDZ1i on tumor cell retention in the lungs, we inoculated cohorts of mice (n = 5, each group) via tail vein injection with either vehicle (DMSO) or test compound (50 μmol/L) pre-treated ARCaP-M-Luc (1 × 10⁶ cells in 100 μL saline) metastatic prostate cancer cells. Luciferase activity was monitored for differential cellular clearance from the lungs between 15 minutes and 5 hours by Bioluminescence imaging (Xenogen in vivo imaging (IVIS) system (Caliper Life Sciences, Inc.). For the lung experimental metastasis model, a total of 5 × 10⁵ ARCaP-M-Luc cells were injected (in 100 μL PBS) by intravenous tail vein injection. Treatment began 12 hours after prostate cancer cell injection. PDZ1i was given at a dose of 30 mg/kg body weight in solution containing DMSO. Drugs were delivered every alternate day for the first three weeks (total 9 injections). Mice were periodically observed for any signs of toxicity. Mice were kept until euthanized as recommended by IACUC. In another experiment, PC-3ML-Luc cells were injected by the intracardiac route to develop bone-metastases. Mice were divided into two groups (n = 8), control and PDZ1i. Vehicle or drug were administered on alternate days for 9 doses within the first three weeks from the day of implantation. Incidence of metastasis was monitored using Bioluminescence imaging, as described previously (30). Two sets of experiments were conducted using the murine-derived prostate cancer cell line RM1-Luc (29). In the first experiment, 1 × 10⁵ RM1-Luc cells were injected by intracardiac route in C57BL/6 mice to develop lung metastases. Similar to the athymic nude mouse study, experimental mice received only three doses of PDZ1i within the first week of treatment. Mice were euthanized on day 9 and the lungs were collected for comparing metastases. In the second experiment, tumor cells were implanted by intracardiac route, divided into two groups, control and PDZ1i and treated as described above. A cohort of 20 mice were used in this experiment. Mice were maintained until they required euthanasia. In an additional experiment, Hi-myc mice (ref. 31; a prostate cancer spontaneous transgenic mouse model) were injected intraperitoneally with PDZ1i starting at 2 months of age and treatment continued for a subsequent 3 weeks (total 9 injections). Mice were kept until they reached 6-months of age. Prostates were removed, photographed, weighed, and processed for paraffin sectioning. Immunohistochemistry was done as previously described (20) with the indicated antibodies.

Co-Immunoprecipitation was performed as described previously (9, 14) using a kit from Pierce (Pierce Biotechnology).

Boyden chamber assays were used to investigate the invasive properties of cancer cells (19, 20). Briefly, cells were pretreated with PDZ1i or DMSO and plated on the upper chamber. After 18 hours, invasive cells in the lower chamber were photographed and analyzed.

Tube formation and chorioallantoic membrane (CAM) assays were performed as described previously (20). Briefly, tumor-derived conditioned media were collected after 24 hours of treatment (either DMSO or PDZ1i) and concentrated. Equal amounts of protein (50 μg) containing conditioned media were mixed with basal media and incubated with HuVEC cells on Matrigel layers. Photographs were taken after 6 hours. In CAM assays, DMSO- or PDZ1i-treated tumor cell-derived conditioned media were implanted on the CAM of 8-day-old fertilized eggs. Photographs were taken on day 12, 4 days after the addition of conditioned media.

Statistical significance analysis was performed using the Student t test in comparison with corresponding controls. Probability values <0.05 were considered statistically significant. Survival curves were analyzed using Cox proportional hazards survival regression using GraphPad Prism.

The chemical structure of PDZ1i was presented (Fig. 1A). To determine the effect of PDZ1i on cell growth and proliferation, long-term (3 week) colony formation (clonal) assays were performed using a series of prostate cancer cell lines and normal immortal prostate epithelial cells (RWPE-1). No significant toxicity was evident in the long-term clonal assay in either prostate cancer or RWPE-l cells after exposure to 50 μmol/L PDZ1i (Fig. 1B). Further confirmation of a lack of overt toxicity or dramatic effects on short term growth of early passage immortal prostate epithelial cells was documented using MTT assays after exposure to 50 μmol/L PDZ1i (Fig. 1C). In addition, only minimal growth inhibition was evident in MTT assays when early passage immortal prostate epithelial cells were treated with 75 or 100 μmol/L PDZ1i (Fig. 1C).

Initially, to assess the direct effect of PDZ1i on mda-9/syntenin–mediated invasion, we used genetically modified (transiently) mda-9/syntenin overexpressing normal immortal prostate epithelial cells (RWPE-1; referred to as RWPE-1 mda-9), which originally expressed minimal levels of MDA-9/Syntenin and minimal to zero invasive ability (Fig. 1D). RWPE-1 mda-9 cells displayed increased invasion; however, this enhancement in invasion was significantly attenuated when cells were treated with PDZ1i (25 and 50 μmol/L; Fig. 1D). We have also tested the anti-invasive effect on prostate cancer cells, including DU-145, ARCaP and its metastatic variant ARCaP-M (Fig. 1E). ARCaP is the parental cell line from which ARCaP-M was derived, consisting of a heterogeneous cell population with both mesenchymal and epithelial phenotypes. All of these prostate cancer cell lines have similar levels of MDA-9/Syntenin (9) and demonstrated comparable inhibition of invasion when treated with PDZ1i.

Our recent work documented a physical interaction between MDA-9/Syntenin and IGF-1R following stimulation with exogenous IGFBP-2 and the resulting significant impact on prostate cancer invasion (11). We also experimentally documented the potential involvement of the first PDZ domain of MDA-9/Syntenin in mediating this interaction (11). On the basis of these considerations, we explored the effect of PDZ1i on MDA-9/Syntenin and IGF-1R interactions. DU-145 and ARCaP-M cells were treated with PDZ1i for 6 hours and cell lysates were subjected to co-IP analysis to determine potential physical interactions between MDA-9/Syntenin and IGF-1R. The same samples were also analyzed for MDA-9/Syntenin and Src interactions confirming that PDZ1i selectively inhibits MDA-9/Syntenin/IGF-1R, but not MDA-9/Src interactions, which occur through the PDZ2 domain, providing further evidence for specificity of PDZ1i (Fig. 2A). Previous studies suggest a role for both PDZ domains of MDA-9/Syntenin in Src interaction, with binding potentially occurring through the PDZ2 domain and then co-interaction with the PDZ1 domain (14). Both proteins, IGF-1R (red) and MDA-9/Syntenin (green) were co-localized (yellow spots in the merged image) in the plasma membrane upon stimulation with IGFBP-2. Treatment of PDZ1i visibly reduced this association and substantiated the ability of PDZ1i to block the physical interaction between MDA-9/Syntenin and IGF-1R (Fig. 2B). Because, this interaction affected STAT3 activation, we investigated whether PDZ1i could downregulate IGF-1R and STAT3 activity in both a dose- and time-dependent manner (Fig. 3A and B). Exogenous stimulation of IGFBP-2 upregulated phosphorylated IGF-1R and STAT3 (active state) in both prostate cancer cells without changing the total corresponding proteins, which was suppressed when cells were pre-treated with PDZ1i. The IGFBP-2–mediated activation and suppression by PDZ1i was sustained for at least 2 hours in both cell lines, which is sufficient to provide signals to regulate downstream invasion-related gene(s) expression. Although PDZ1i did not affect MDA-9/Syntenin/Src interactions, Src activity was reduced (Fig. 3A and B), suggesting that IGF-1R might also play a role in Src activation.

Previous studies indicate that IL-6 can activate STAT3 through IGF-1R in prostate cancer (13). In agreement with that study, we also observed IGF-1R activation in IL-6–treated samples (Supplementary Fig. S1); however, this effect was not suppressed by PDZ1i, suggesting specificity for IGFBP-2/MDA-9/IGF-1R–mediated STAT3 activation. Additional support for this conclusion comes from experiments where IGFBP-2 and IL-6 were used in combination either in the presence or absence of PDZ1i. As predicted, PDZ1i only downregulated IGFBP-2–induced STAT3 activation, but not when IGFBP-2 and IL-6 were used in combination (Fig. 3C). PDZ1i might directly or indirectly down regulate multiple receptor tyrosine kinases, which is currently under investigation. Finally, consistent with our previous observations, PDZ1i downregulated both MMP-2 and MMP-9 (32), at both the protein and activity level as documented by Western blotting and zymography, respectively (Fig. 3D).

MDA-9/Syntenin non-autonomously (through secretion of several angiogenic factors, including IGFBP-2 and IL-8) regulates tumor angiogenesis in melanoma cells (20). In this study, we explored the potential impact of PDZ1i on tumor-derived angiogenesis in prostate cancer. The experimental strategy is outlined in Supplementary Fig. S2A. Data analysis from an antibody-based array highlighted a number of defined pro-angiogenic proteins that are downregulated by PDZ1i (Fig. 4A and B). A complete list of proteins that have been tested is provided in Supplementary Fig. S2. A robust downregulation (>50%) of several pro-anagiogenic proteins, for example, Angiogenin, Angiopoetin-2, Amphiregulin, FGF-basic, GM-CSF, and IL-1β, was evident in the PDZ1i-treated samples. In addition, IGFBP-2, VEGF-A, and CXCL16 also displayed considerable downregulation in PDZ1i-treated samples. The reduction of the anti-angiogenic factor endostanin is surprising and unanticipated and requires further study. Because VEGF-A is a potent angiogenic factor and a downstream target of STAT3 (33), we also scrutinized the expression pattern of VEGF-A at an mRNA level in different prostate cancer cells, including mda-9/syntenin stably expressing RWPE-1 cells (Fig. 4C). MDA-9/Syntenin mediates regulation of CXCL16, TSP-1, uPA (34) and IGFBP-2 (20) and this study supports the specificity of PDZ1i to MDA-9/Syntenin action. Finally, both in vitro tube formation and in vivo CAM assays confirmed the anti-angiogenic properties of PDZ1i-treated tumor cell-derived conditioned media (Fig. 4D).

To assess the potential therapeutic efficacy of PDZ1i we conducted six independent sets of in vivo experiments. First, stable luciferase expressing ARCaP-M (ARCaP-M-Luc) cells were pre-treated with PDZ1i and injected intravenously into animals to determine the effect of this MDA-9/Syntenin inhibitor in modulating retention and adhesion of metastatic tumor cells in the lungs. As shown in Fig. 5A, mice inoculated with PDZ1i pre-treated cells were present at lower concentrations than in control groups in the lungs and cleared more rapidly from the lungs (within 2 hours post-inoculation). This suggests that pre-treatment with PDZ1i might alter the adhesion ability of potentially metastatic tumor cells, which in principle, ultimately have a direct effect on development of metastatic lesions at secondary sites. Moreover, animals injected with PDZ1i-treated ARCaP-M-Luc cells survived longer than control animals treated with vehicle without the small-molecule PDZ inhibitor (Fig. 5A). In the second set of experiments (Fig. 5B), an equal number of cells was implanted into animals using an intravenous tail vein route. Experimental groups received 30 mg/kg body weight of PDZ1i intraperitoneally as described in the figure legend. As predicted, the treated groups developed significantly fewer lesions as compared with control groups, as detected by bioluminescence imaging (BLI), ultimately resulting in prolonged survival (Fig. 5B). To investigate the potential role of PDZ1i in inhibiting bone metastasis, PC-3ML-Luc cells were implanted by intracardiac route in animals (n = 8) and treated with vehicle or PDZ1i. Six out of eight mice from the control group developed metastases in the left or right femur. In contrast, no femur metastases were detected in the PDZ1i-treated group. In addition, a lower incidence of lung (7 out of 8 vs. 5 out of 8) metastases were evident in animals that received PDZ1i.

In the fourth and fifth set of experiments, two cohorts of C57BL/6 mice were injected through intracardiac route with RM1-Luc cells. The first cohort consisted of 5 mice from each group (vehicle and PDZ1i) and were euthanized after 9 days to evaluate tumor burden in the lungs. The second cohort with 10 mice in each group were used to develop Kaplan–Meier survival curves. Similar results as seen in the nude mice experiments were evident, that is, PDZ1i significantly reduced tumor burden in the lungs (Fig. 6A) and significantly enhanced survival (Fig. 6B). In the sixth experimental approach, 8-week-old Hi-myc male mice were treated intraperitoneally with either PDZ1i (30 mg/kg) or vehicle 9 times for the first three weeks. Mice were then maintained until 6-months of age, when adenocarcinomas fully developed in this transgenic animal model. In addition, to understand drug-mediated molecular changes, 48 hours before sacrifice, a single additional PDZ1i treatment was given. Both the size and weight of prostates collected from the PDZ1i-treated groups were significantly less than the control groups (Fig. 6C). H&E staining of prostate sections (Fig. 6D) indicated that histologically the prostate from the control (vehicle-treated) group showed a significant progression to adenocarcinoma, which was not evident in the treated groups (representative photomicrographs are presented). Immunostaining for the expression of pIGF-1R and pSTAT3, two downstream effectors of MDA-9/Syntenin, were also significantly reduced in PDZ1i-treated animals (Fig. 6E), validating the in vitro observations in vivo. A schematic diagram describing potential targets that are affected by PDZ1i is presented in Fig. 6F.

Genomic interrogation using whole-genome sequencing, RNA sequencing and single-nucleotide polymorphism profiling of primary and metastatic prostate cancer has provided an unparalleled opportunity to identify potential genetic targets that may afford “precision medicine” approaches for enhancing therapeutic outcomes (35–37). A major component of cancer progression relates to tumor heterogeneity, which may be the driving force in defining why only 10% of prostate cancer cases progress to lethality (38). On the basis of accumulating data, various pathways are being targeted to effect beneficial outcomes in patients with metastatic prostate cancer, including tumor vasculature, androgen receptors, IGF-1 and IL-6 signaling, and cytoprotective chaperones (39, 40). However, therapy of advanced prostate cancer, particularly when metastasis to bone has occurred, still remains an unattainable objective. Developing effective and potentially curative strategies to treat advanced prostate cancer requires a detailed understanding of molecular mechanisms of prostate cancer pathogenesis, including causative genetic changes and molecular determinants of tumor heterogeneity that result in target-site specific metastatic clonal populations. We presently demonstrate that MDA-9/Syntenin (SDCBP) provides a molecular target for developing small molecules that can intervene in the invasive and metastatic properties of prostate cancer both in vitro and in vivo in preclinical animal models. Small molecules specifically targeting interactions between the PDZ1 domain of MDA-9/Syntenin and IGF-1R have been developed (27, 41) and may provide a novel approach for managing prostate cancer.

MDA-9/Syntenin (SDCBP) is an adaptor scaffold protein that elicits its diverse functions by physically interacting with subsets of unique proteins in different regions of the cell (6, 8, 10). We recently identified IGF-1R as a new MDA-9/Syntenin binding partner and are defining the consequences of this interaction in disease progression (11). Considering the importance of these interactions in STAT3 activation, which has a decisive impact on prostate cancer progression as shown in our and other studies (13, 42, 43), we sought to identify small molecule inhibitors that could specifically interrupt MDA-9/Syntenin/IGF-1R interactions thereby affecting STAT3 activation and prostate cancer invasion. Using a fragment-based drug-discovery approach (FBDD) guided by NMR spectroscopy in solution, we previously confirmed the hypothesis that the PDZ domains of MDA-9/Syntenin are “druggable” and amenable to producing selective small molecules that could bind and potentially disrupt protein interactions required for biological activity of MDA-9/Syntenin (27). This strategy resulted in the identification of a PDZ1-specific binding small molecule, PDZ1i, that in micromolar doses could modify the ability of IGF-1R to bind to MDA-9/Syntenin (Fig. 2). Since our approach did not generate small molecules that specifically bound to the PDZ2 of MDA-9/Syntenin (27, 41), we anticipated and confirmed that in prostate cancer cells PDZ1i did not modify MDA-9/c-Src binding. This observation, despite the considerable homology (29% sequence identity and 48% sequence homology) between the PDZ1 and PDZ2 domains of MDA-9/Syntenin, supports the specificity of the PDZ1i. However, in previous studies we demonstrated that both PDZ domains of MDA-9/Syntenin are critical for interacting with c-Src (14), and we now show that PDZ1i downregulated Src activation, thereby negatively influencing further downstream signaling, including p38 and NF-κB activation in different prostate cancer cells (Fig. 2). c-Src can be activated in multiple ways (44, 45) either by autophosphorylation or through various tyrosine kinases, including IGF-1R (45, 46). In prostate cancer cells, PDZ1i-mediated downregulation of c-Src activation is partially IGF-1R dependent. Further studies are warranted to understand precisely how MDA-9/Syntenin chooses specific partners to complex with under normal physiological conditions as well as in specific disease states such as cancer.

Because of the clinical significance of IGF-1R in prostate cancer (47), diverse therapeutic approaches have been tested that target this protein, including human monoclonal antibodies and small molecules to inhibit IGF-1R activity through distinct targeting approaches, for example, IGF-1R ligands, blocking ligand binding, inhibiting enzymatic activity, etc. Although direct IGF-R1 targeting has significant potential, these strategies need to be optimized and critically evaluated in vivo due to ubiquitous expression and the key roles of IGF-1R in different physiological contexts. As documented presently, our PDZ1i small molecule rather than targeting IGF-1R directly, inhibits IGF-1R activity by perturbing MDA-9/Syntenin/IGF-1R interactions, which may be restricted to its' role in cancer cells. In addition, recent studies demonstrate that MDA-9/Syntenin (SDCBP) knockout mice are viable and in these mice tumor-supporting inflammation is inhibited and melanoma metastasis is suppressed (48). Thus, the activity of MDA-9/Syntenin is dispensable for normal cellular functions and MDA-9/Syntenin-targeted molecules are more specific to neoplastic cells, an important prerequisite for drug development. In addition, a functional role for MDA-9/Syntenin in mediating glioma cancer stem cell growth and survival has been demonstrated recently which further supports the involvement of this gene in regulating cancer phenotypes and its potential as a target for therapeutic intervention (49).

In summary, this study confirms that the PDZ1 domain of MDA-9/Syntenin is “druggable” and it is possible to develop inhibitors that are unique to this domain of MDA-9/Syntenin and affect its interactions with specific proteins without binding or altering protein–protein interactions to the PDZ2 domain of this protein. Although it has been 20 years since MDA-9/Syntenin was cloned, only recently has this gene attracted major attention due to its ubiquitous expression in different cancers, and through investigations by multiple independent research laboratories defining unique functions in cancer cells (6, 8, 27). In principle, with appropriate advancement this small molecule (or other potential inhibitors targeting either PDZ1, PDZ2 or both PDZ domains of MDA-9/Syntenin) could be scaffolds for developing drugs that would be candidates for optimizing “personalized medicine” approaches to treat prostate cancer and other cancers. The role of MDA-9/Syntenin in several physiological processes (6, 50), including exosome biogenesis and pre-synaptic synapses needs to be considered critically in defining molecules that will have selective cancer-specific activity. MDA-9/Syntenin functions principally through interactions with partner proteins and these interactions are influenced by multiple factors including phosphorylation status of ligands, disulfide bond formation, auto-inhibition, competitive binding and allosteric interactions (50). In these contexts, defining MDA-9/Syntenin interacting partners that are disease-specific and generating small-molecule drugs that selectively impact on these interactions are essential to exploit important signaling axes in developing cancer-selective therapies.

P.B. Fisher is a co-founder, has ownership interest and was a consultant in Cancer Targeting Systems, Inc. Virginia Commonwealth University, Johns Hopkins University and Columbia University have ownership interest in CTS. P.B. Fisher is a co-founder and has ownership interest in InVaMet Therapeutics, Inc. (IVMT). Virginia Commonwealth University and the Sanford Burnham Prebys Medical Discovery Institute have ownership interest in IVMT. S.K. Das is the P.I. of an SRA provided by IVMT to Virginia Commonwealth University. No potential conflicts of interest were disclosed by the other authors.

Conception and design: S.K. Das, P.B. Fisher

Development of methodology: S.K. Das, P.B. Fisher

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): S.K. Das, T.P. Kegelman, A.K. Pradhan, P. Bhoopathi, S. Talukdar, S. Maji, L. Emdad

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): S.K. Das, L. Emdad, P.B. Fisher

Writing, review, and/or revision of the manuscript: S.K. Das, D. Sarkar, P.B. Fisher

Study supervision: P.B. Fisher

We thank Drs. Maurizio Pellecchia, University of California Riverside, and Mehmet Kahraman, Ludwig Institute for Cancer Research, UCSD, for valuable discussions and Drs. Maurizio Pellecchia, Bainan Wu, Surya K. De, Angela Purves and Jun Wei for preparing and providing specific reagents used in this study. The present research was supported in part by funding from NIH grants P50 CA058236 (to P.B. Fisher) and NCI Cancer Center Support Grant to VCU Massey Cancer Center P30 CA016059 (to P.B. Fisher and D. Sarkar), the National Foundation for Cancer Research (NFCR; to P.B. Fisher), the Human and Molecular Genetics Enhancement Fund (to S.K. Das and L. Emdad), VCU Massey Cancer Center (MCC) developmental funds (to P.B. Fisher) and VCU Institute of Molecular Medicine (VIMM) developmental funds (to P.B. Fisher, S.K. Das, and L. Emdad). P.B. Fisher holds the Thelma Newmeyer Corman Chair in Cancer Research in the MCC. D. Sarkar is the Harrison Foundation Distinguished Professor in Cancer Research in the MCC.

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.