|Home | About | Journals | Submit | Contact Us | Français|
Survivin expression in melanoma is inversely correlated with patient survival. Transgenic mice harboring melanocyte-specific overexpression of survivin exhibit increased susceptibility to UV-induced melanoma and metastatic progression. To understand the mechanistic basis for metastatic progression, we investigated the effects of Survivin on the motility of human melanocytes and melanoma cells. We found that Survivin overexpression enhanced migration on fibronectin and invasion through Matrigel, whereas Survivin knockdown under sub-apoptotic conditions blocked migration and invasion. In melanocytes, Survivin overexpression activated the Akt and MAPK pathways. Akt phosphorylation was required for Survivin-enhanced migration and invasion, whereas Erk phosphorylation was required only for enhanced invasion. In both melanocytes and melanoma cells, Survivin overexpression was associated with upregulation of α5 integrin (fibronectin receptor component), the antibody-mediated blockade or siRNA targeting of which blocked Survivin-enhanced migration. Knockdown of α5 integrin did not affect Akt activation, but inhibition of Akt phosphorylation prevented α5 integrin upregulation elicited by Survivin overexpression. Together, our results showed that Survivin enhanced the migration and invasion of melanocytic cells and suggested that Survivin may promote melanoma metastasis by supporting Akt-dependent upregulation of α5 integrin.
Cell migration plays a central role in biological processes such as embryonic development and wound healing (1). In cancer, aberrant cell migration is involved in tumor progression and metastasis as malignant cells hijack the migratory machinery and acquire the ability to move and invade other tissues (2). The propensity to metastasize is particularly high in melanoma, a potentially fatal form of cancer arising from melanocytes in the skin (3). The probability of metastasis in melanoma directly relates to the (Breslow) depth of tumor invasion, which inversely correlates with patient survival (3). In early lesions, melanoma cells exhibit radial growth phase, which is characterized primarily by lateral migration. More advanced lesions display vertical growth phase, characterized by deeper invasion and increased risk of metastasis. Thus, deregulated cell migration is a key feature of melanoma progression, and required for metastasis.
Several pathways important in melanoma migration and invasion have been previously characterized. First, the mitogen-activated protein kinase (MAPK) pathway is activated in most melanomas (4), with roughly two-thirds of tumors harboring activating mutations in BRAF (5), and constitutive MAPK activation in immortalized melanocytes confers tumorigenicity (6). Second, Akt activation is found in 70% of melanomas (7) and mediates melanoma cell migration, invasion and metastasis in various model systems (8–10). In particular, activation of Akt has also been shown to convert a melanoma cell line from radial to vertical growth phenotype (11).
Survivin is an Inhibitor of Apoptosis Protein (IAP) that has been characterized both as a regulator of cell division and cell death (12). As a chromosomal passenger protein, it is required for proper chromosomal alignment during mitosis (13). Survivin protects cells against both caspase-dependent and caspase-independent apoptosis (14), and interference with Survivin function in melanoma cells leads to spontaneous apoptosis (15) and impairs tumor growth (16). Survivin is upregulated during melanocyte transformation (17), and several studies have shown an inverse correlation between Survivin expression in advanced melanoma and patient survival (18, 19). We have previously shown that transgenic expression of Survivin in murine melanocytes promotes UV-induced melanoma tumor formation and metastasis (20). The metastatic phenotype in these mice was unexpected, as previous mouse tumor models based on Survivin expression in uroepithelium (21) and keratinocytes (22) did not reveal metastasis.
Here we describe an alternate role for Survivin in tumor progression, namely its promotion of cellular motility. We show that expression of Survivin in melanocytes, and its over-expression in melanoma cells, enhances cell migration and invasion. The pro-migratory effect in Survivin-expressing cells is mediated by activation of Akt and upregulation of α5 integrin, a component of the fibronectin receptor (23).
Normal human melanocytes were propagated from discarded foreskins (24). Human metastatic melanoma cell lines LOX, YUSAC2 (YU2), and YUGEN8 (YU8) are previously described (25), and HTB-66 and SK-MEL-28 (SK-28) were from American Type Culture Collection. The human primary melanoma cell lines WM35 and WM793 were kindly provided by Meenhard Herlyn (Wistar Institute). The 4C7 human melanoma cells, which express wild-type human Survivin in a tetracycline (tet)-regulated (tet-off) manner, are previously described (26). All cells were grown at 37 °C in 5% CO2 humidified air.
Polyhema (Sigma Chemical Co.) was dissolved in 95% ethanol (120 mg per mL) overnight at 65 °C, diluted 1:10 in 95% ethanol, then added to 6-well plates (0.5 mL per well) and allowed to dry overnight at 37 °C. Melanocytes were added in growth medium and incubated for 48 h. Apoptotic cells were detected by propidium iodide (PI, Sigma) staining and flow cytometry as previously described (14).
Cells were added to transwell polycarbonate membranes (#3422, Costar) after coating the lower surfaces with 0.5 mL recombinant human fibronectin (10 μg per mL, Invitrogen) in PBS overnight at 37 °C. To block cellular proliferation during the assay, cells were first cultured with mitomycin C (10 μg per mL, Sigma) for 2 h. After washing, cells were resuspended in 100 μL of serum-free medium and added to the upper surface of transwells. Each lower chamber contained 0.5 mL serum-containing media. After incubation at 37 °C (optimal time determined for each cell line), non-migratory cells on upper membrane surfaces were removed using a cotton-tip applicator. Cells that migrated through to the lower membrane surface were fixed with 4% paraformaldehyde in PBS for 15 min and then stained with 4′-6-diamidino-2-phenylindole (DAPI, 1 μg per mL, Sigma) in PBS for 5 min at room temperature. Migrating cells were quantified by manually counting 5 representative high-powered fields, using a Nikon Eclipse TE300 fluorescence microscope. For some experiments, inhibitors of PI3K (wortmanin, Sigma), Akt (Akt VIII, Calbiochem) and MAPK (u0126, PD98059; Cell Signaling Technology) were also added at non-cytotoxic concentrations. For antibody blocking experiments, cells were resuspended in serum-free media containing 10 μg per mL antibody against α5 integrin (IIA1, BD Biosciences), α5β1 integrin (P1D6, Covance), α6 integrin (MAB13501, R&D Systems) or control mouse IgG (Sigma), and then seeded onto the upper surface of the Transwell chamber. After incubation for 3 h at 37 °C to allow cell adherence to the membrane, upper chambers were gently washed with PBS to remove antibodies and replenished with 100 μL serum-free media. For wound healing assay, a linear scratch was made using a sterile pipette tip on semi-confluent adherent cells, and indicated antibodies were added for 3 h.
The Matrigel matrix in BD BioCoat invasion chambers (354480, Discovery Labware) was rehydrated with serum-free medium for 2 h at 37 °C in a humidified incubator. To block cellular proliferation during the assay, cells were first cultured with mitomycin C as above. Cells were added in 0.5 mL serum-free medium to each upper chamber, and 0.75 mL culture medium was added to each lower chamber as a chemo-attractant. After incubation at 37 °C (optimal time determined for each cell line), non-invading cells were removed from upper membrane surfaces and invading cells were quantified on lower membrane surfaces as described above.
Melanocytes or melanoma cells were grown to 90% confluence and then infected with a control adenovirus expressing GFP (pAd-GFP) or a Survivin expressing GFP-tagged adenovirus (pAd-Survivin-GFP or pAd-Surv) (27), kindly provided by Dario Altieri (Univ. of Massachusetts). After 6 h, the culture medium was changed to remove excess virus.
Cells were transfected at 70% – 80% confluency in 6-well plates with oligonucleotides targeting α5 integrin, Akt1/2 or Survivin (sc-29372, sc-43609 or sc-29499, respectively; Santa Cruz Biotechnology) or a control scrambled sequence (17). The siRNA duplexes (10 μM stocks; 8 μL for α5 integrin, 6 μL for Akt or Survivin) were diluted in 100 μL transfection medium (sc-36868, Santa Cruz Biotechnology), then combined with 100 μL transfection medium containing transfection reagent (sc-29528, 8 μL for α5 integrin and 4 μL for Akt or Survivin; Santa Cruz Biotechnology) for each well. After incubating 30 min at room temperature, the transfection mixture was added drop-wise to cells in 0.8 mL transfection medium per well. Plates were incubated at 37 °C for 7 h, then 1 mL of 2X normal growth medium was added to each well.
Trypsinized cells were washed, lysed, and then electrophoresed and transferred to membranes as described (17). After blocking, membranes were incubated for 1 h with primary antibodies (see Supplementary methods) dissolved in PBS containing 0.1% Tween and 5% non-fat milk. After staining with species-appropriate secondary horseradish peroxidase-conjugated antibodies (PerkinElmer Life Sciences), protein bands were visualized by enhanced chemiluminescence (PerkinElmer Life Sciences) and autoradiography.
Analyses were performed with Prism 3.0 software (GraphPad). Data derived from multiple determinations were subjected to two-sided t tests. P values ≤ 0.05were considered statistically significant.
To investigate potential roles of Survivin in promoting melanoma metastasis, we first examined whether its well-established capacity to confer apoptotic protection (12) could also impart resistance to detachment-induced apoptosis (anoikis). Melanocytes must acquire resistance to anoikis to survive loss of contact with keratinocytes and basement membrane which necessarily occurs as an early event in melanoma metastasis. For this purpose, melanocytes were infected with control or Survivin-expressing viruses (Supplementary Fig. S1A), and then seeded onto Polyhema-coated plates to prevent attachment (Supplementary Fig. S1B). As shown in Supplementary Fig. S1C and S1D, the apoptotic fraction was only reduced by 10–15% in Survivin-expressing melanocytes. Although we have previously shown that Survivin expression protects melanocytes against UV-induced apoptosis (14, 20), this modest protection against anoikis suggests that Survivin promotion of metastasis may occur through alternate mechanisms.
Given the importance of migration and invasion in melanoma progression and metastasis, we next investigated whether Survivin expression could support either of these activities in human melanocytes or melanoma cells. Survivin was expressed in melanocytes and over-expressed in two melanoma cell lines (LOX, YU8) (Fig. 1A, left), and migrating cells were visualized (Supplementary Fig. 2A). We observed that Survivin expression enhanced cellular migration in melanocytes (Fig. 1A, middle) and melanoma cells (Fig. 1A, right). Similarly, by Matrigel assay, we found that Survivin-expressing melanocytes demonstrated markedly increased invasive activity compared to control cells (Fig. 1B, left). This enhanced invasive capacity was recapitulated in melanoma cell lines (LOX, WM35) over-expressing Survivin (Fig. 1B, right). Thus, Survivin expression in human melanocytes and melanoma cells is associated with enhanced capacity for migration and invasion.
We noted that the effect of Survivin expression on melanocyte invasion was more pronounced than that on migration (Fig. 1A, middle and Fig. 1B, left), and wondered whether this could reflect differential response to various extracellular matrix proteins used as substrates in the migration and invasion assays. The substrate in the migration assay was fibronectin while the Matrigel (according to the manufacturer) in the invasion assay is comprised of collagen IV and laminin I. Therefore, we re-examined the migratory activities of Survivin-expressing melanocytes in the context of each of these substrates. While Survivin did not enhance melanocyte migration in the absence of substrate, we did observe Survivin-enhanced migration on fibronectin and laminin (Fig. 1C). While melanocyte migration was increased on collagen, no difference was seen for Survivin-expressing cells on that substrate (Fig. 1C). On the other hand, while Survivin enhanced migration of YU2 melanoma cells on fibronectin (see below) there was no increased migration of Survivin-over-expressing YU2 cells on laminin (not shown).
To assess the role of endogenous Survivin in melanoma cell migration and invasion, Survivin expression was down-regulated in several melanoma cell lines using RNAi. Although RNAi-mediated depletion of Survivin was found to induce apoptosis in pancreatic cancer (28) and glioma cells (29), others showed its partial depletion by gene targeting in melanoma cells only sensitized to drug-induced apoptosis but was not sufficient for induction of apoptosis (30). Similarly, we found in both LOX and YU8 melanoma cells that RNAi-mediated depletion of Survivin (Fig. 2A and B, left) was not associated with significant apoptosis over a 72-h period (Fig. 2A and B, middle). Importantly, this provided a window for investigating the effect of Survivin depletion on migration and invasion without the potential confounding effect of cells undergoing apoptosis during the assay. Migrating (Supplementary Fig. 2B) and invading cells were quantitated following transfection with control or Survivin-specific RNAi. For both melanoma cell lines, we observed that reduced Survivin levels significantly decreased capacity for both migration (Fig. 2A and B, right) and invasion (Fig. 2C). Taken together, these data demonstrate that endogenous Survivin expression is required for melanoma cell migration and invasion.
We next asked whether Survivin expression in melanocytes could promote an epithelial-to-mesenchymal transition (EMT). The EMT refers to phenotypic and genetic changes characteristic of cells that have escaped from epithelia and are primed for metastasis; conventional markers of EMT include loss of E-cadherin, increased expression of fibronectin and vimentin, and nuclear translocation of β-catenin (31). As shown in Supplementary Fig. S3, Survivin-expressing melanocytes did not exhibit changes in these markers. In addition, Survivin over-expression in 4C7 melanoma cells was also not associated with alteration of any of these markers. Thus while Survivin promotes melanocyte migration and invasion, its expression is not sufficient to mediate acquisition of EMT.
Two signaling pathways known to be important in melanoma migration include the MAPK and phosphatidylinositol-3 kinase/Akt pathways (32). Survivin expression in melanocytes was associated with activation of both Akt and MAPK, as reflected by increased levels of phosphorylated Akt and Erk species, respectively (Fig. 3A). We used specific inhibitors of these pathways (32) to assess independently their role in Survivin-enhanced migration. Since these inhibitors were found to be toxic for melanocytes (not shown), we performed these experiments only in melanoma cells. Survivin over-expression was associated with increased levels of phospho-Akt in both YU2 (Fig. 3B) and LOX cells (not shown), while Erk was constitutively phosphorylated. Addition of Akt VIII or u0126 efficiently blocked phosphorylation of Akt and Erk, respectively, in both YU2 (Fig. 3B) and LOX cells (not shown). In migration assays, inhibition of Akt phosphorylation blocked Survivin-enhanced migration of both YU2 and LOX cells (Fig. 3C). Similar inhibitory effects on Survivin-enhanced migration were seen using the PI3K inhibitor wortmanin (Supplementary Fig. S4B). On the other hand, u0126-mediated inhibition of Erk phosphorylation did not affect Survivin-enhanced migration in either melanoma line (Fig. 3C). Similarly, another Erk inhibitor (PD98059) did not block Survivin-enhanced migration in either line (Supplementary Fig. S4C, left). Inhibition of either Akt or Erk, however, was sufficient to block Survivin-enhanced invasion of both YU2 and LOX cells (Fig. 3D). Depletion of Akt in LOX cells by siRNA (Supplementary Fig. S5A) impaired both Survivin-enhanced migration (Supplementary Fig. S5B, left) and invasion (Supplementary Fig. S5B, left). Inhibition of Survivin-enhanced invasion was also seen in both cell lines treated with Erk inhibitor PD98059 (Supplementary Fig. S4C, right). Thus Survivin promotion of melanoma cell migration requires Akt activation (but is independent of Erk activation), while enhanced invasion requires both Akt and Erk activation.
The substrates on which we observed Survivin promotion of melanocyte migration – namely fibronectin and laminin (Fig. 1C), are recognized by αβ heterodimeric integrin receptors. Thus we considered the possibility that Survivin may specifically promote migration and invasion through upregulation on melanocytic cells of one or more particular integrin receptor chains involved in binding to these matrix proteins. We examined in control and Survivin-expressing melanocytes the protein expression of potentialα and β integrin chains known to contribute to fibronectin (α5β1,αvβ3) and laminin (α1β1, α2β1, α3β1, α6β1, α6β4, αvβ3) binding (23). Survivin expression led to upregulation of α1 and α5 integrins, with the latter being more robust, while expression of the other integrins was unaffected or slightly decreased by Survivin expression (Fig. 4A).
The α5β1 heterodimer constitutes the predominant receptor for fibronectin (23). We observed a striking increase in α5 integrin protein levels in melanocytes at 24 h following forced Survivin expression (Fig. 4B). Melanocytes constitutively express β1 integrin, the protein levels of which were not affected by Survivin expression (Fig. 4B). Next we investigated a similar correlation between Survivin and α5 integrin expression in melanoma cells. We screened a small panel of primary and metastatic melanoma cell lines for α5 integrin expression. The α5 integrin was expressed constitutively in 4 of 6 lines, with higher levels in metastatic vs. localized melanoma lines (Fig. 4C, left). In 5 of 6 lines, α5 integrin was upregulated in the context of Survivin over-expression, while one line (SK-28) did not express α5 integrin even after Survivin over-expression (Fig. 4C, left). Thus the connection observed between Survivin and α5 integrin expression in melanocytes is largely conserved in melanoma cells. By contrast, levels of α1 integrin were not increased by Survivin over-expression in two melanoma lines (Fig. 4C, right).
To test whether a functional correlation exists between expression of Survivin and upregulation of α5 integrin, we asked whether α5 integrin is required for Survivin-enhanced motility. First, migration of Survivin-expressing melanocytes on fibronectin was examined in the presence of two different α5 integrin-specific blocking antibodies: one reactive with α5 integrin (IIA1), and another (P1D6) that blocks interaction of the α5β1 heterodimer with its ligand fibronectin. As shown in Fig. 4D top left, the presence of either α5-blocking antibody (but not control IgG or α6-specific antibody) abrogated the effect of Survivin on melanocyte migration. While anti-α5 antibody inhibited migration of Survivin-expressing melanocytes on fibronectin, it was without effect when laminin (not a ligand for α5 integrin) was used as the substrate (Fig. 4D, top right). Similar inhibitory effects of these antibodies were observed on Survivin-enhanced migration of LOX cells in a “scratch” (wound healing) assay without substrate (Fig. 4D, bottom).
As a second test of the requirement for α5 integrin in Survivin-enhanced motility, we examined the effect of α5 integrin gene knockdown. These experiments were performed in melanoma cells which are more amenable to transfection of siRNA than melanocytes. Levels of α5 integrin were reduced in Survivin over-expressing YU2 melanoma cells following transfection of specific siRNA (Fig. 5A, left). Depletion of α5 integrin abrogated Survivin-enhanced migration of these cells (Fig. 5A, right). Similar results were seen in α5 integrin-depleted LOX cells (Fig. 5B). Despite these effects on migration, we found that knockdown of α5 integrin in either YU2 or LOX cells did not inhibit Survivin-enhanced invasion (Fig. 5C). Similarly, α5 integrin blocking antibodies did not impair Survivin-enhanced invasion of LOX cells (Supplementary Fig. S6). These data suggest that α5 integrin is required for Survivin-enhanced melanocyte and melanoma cell migration, but not invasion of melanoma cells.
Given the requirement of Akt activation and importance of α5 integrin for Survivin promotion of migration (Fig. 3C, ,4D),4D), we examined the relationship between Akt activation and α5 integrin in the context of Survivin expression. In Survivin-over-expressing cells, knockdown of α5 integrin was not associated with significant changes in levels of phosphorylated Akt (Fig. 5A,B). Conversely, addition of PI3K inhibitor abrogated Survivin-mediated upregulation of α5 integrin in both LOX (Fig. 5D, left) and YU2 (not shown) cells. On the other hand, addition of Erk inhibitor u0126 did not interfere with Survivin-mediated upregulation of α5 integrin in LOX (Fig. 5D, right) or YU2 (not shown) cells. Thus α5 integrin appears to be downstream of PI3K/Akt, and its upregulation following Survivin over-expression requires Akt but not Erk activation.
Survivin is widely expressed in cancer, and thought to promote tumor development and progression through two previously characterized functions: apoptotic inhibition and mitotic chromosomal alignment (12). Indeed, multiple studies from our laboratory and others have shown that inhibition of Survivin in both malignant cells in vitro (15, 33) and tumors in vivo (16, 26) precipitates spontaneous apoptosis and mitotic defects. In this study, we demonstrate a third capability of Survivin – promotion of cellular motility. We show that Survivin is both necessary and sufficient to enhance migration and invasion in normal human melanocyte and melanoma cell lines. For all the assays, cells were treated with mitomycin C to inhibit mitosis, and thus preclude any effects that may have resulted from differences in cell division associated with changes in Survivin expression. The observed effects on cellular motility also appear to be independent of Survivin anti-apoptotic function, as both migration and invasion were significantly compromised in melanoma cells depleted of Survivin under conditions not affecting apoptosis. Thus Survivin capacity to promote cellular migration and invasion appears to represent a novel function distinct from its other established activities.
The capacity of Survivin to promote melanocyte migration on fibronectin and laminin, but not collagen, suggested the possibility that Survivin expression may differentially regulate expression of various integrins known to interact with these particular substrates. We decided to focus on the functional role of α5 integrin since we were able consistently to demonstrate dramatic upregulation of this integrin at the protein level. Given constitutive expression of β1 integrin in melanocytes, induction of α5 integrin should constitute a functional (α5β1) fibronectin receptor. The correlation of expression between Survivin and α5 integrin is conserved in both normal melanocytes and several melanoma cell lines. Our finding that α5 integrin is required for Survivin-enhanced melanocyte and melanoma cell migration is consistent with its described role in enhancing cell motility and cancer metastasis (34) and some studies suggesting its importance in melanoma development and metastasis. Expression of α5 integrin has been demonstrated in both primary and metastatic melanocytic lesions (35), and in a mouse model, inhibition of α5 integrin effectively blocked formation of pulmonary metastases (36).
In contrast to its role in migration shown here, we found that α5 integrin is not required for Survivin-enhanced melanoma cell invasion. Although α1 integrin is also upregulated by Survivin in melanocytes, it was not similarly upregulated in melanoma cells. We have examined several melanoma cell lines by qRT-PCR to see whether Survivin upregulates other integrins known to be important for melanoma invasion such as αv and β3 (37, 38), and did not find significant upregulation of expression of these integrin chains with Survivin over-expression (J.A.M., T.L., and D.G., unpublished observations). Expression of the αvβ3 heterodimer has been associated with transition from radial to vertical growth phase melanoma (39). An underlying mechanism may involve interaction of αvβ3 with matrix metalloproteinase-2 (40), which facilitates subsequent degradation of basement membrane (41). It is possible that other integrins are upregulated in Survivin-expressing melanoma cells that allow interaction with matrix proteins other than fibronectin and underlie promotion of cellular invasion. Upregulation of other integrins may also explain our finding that Survivin enhances melanocyte migration on other substrates such as laminin, which was not blocked by antibodies against α5 integrin. However, we did not observe Survivin-enhanced melanoma cell migration on laminin.
Integrins promote cell migration and invasion through interactions with extracellular matrix proteins that not only enhance adhesion, but also transmit intracellular signals (42). Signaling pathways known to promote integrin-mediated migration include activation of focal adhesion kinase and Src (43), MAPK (44) and Akt (45), which are also associated with deregulated cell migration in cancer (46, 47). We found activation of both MAPK and Akt pathways to be associated with Survivin expression in melanocytes; in melanoma cells, Survivin over-expression activates Akt while the MAPK pathway is constitutively activated. In our system, Akt activation is required for both Survivin-enhanced melanoma cell migration and invasion, while MAPK (Erk) activation is required only for invasion. While knockdown of α5 integrin did not affect Akt phosphorylation, inhibition of Akt blocked the upregulation of α5 integrin following Survivin over-expression. By contrast, Erk inhibition did not interfere with Survivin-induced upregulation of α5 integrin. Our data thus suggest a pathway whereby Survivin promotes melanoma cell migration via Akt-dependent expression of α5 integrin. A role for Akt in the Survivin pathway is consistent with the established role of Akt in melanoma cell migration, invasion and metastasis (8–11). The mechanism by which Survivin activates Akt is not presently clear, however Guha et al. (48) showed that PTEN suppresses Survivin expression. Thus it is possible that Survivin may regulate Akt phosphorylation through negative feedback on PTEN. Indeed, we have seen in melanocytes and two melanoma cell lines that Survivin expression results in downregulation of PTEN (J.A.M. and D.G., unpublished observations). Akt is known to regulate integrin expression/activation by various mechanisms (49, 50). With respect to increased expression of α5 integrin in Survivin-expressing cells, in preliminary studies in melanoma cells we have found that Survivin increases nuclear concentration of C/EBP and NF-1, which are the transcription factors known to regulate α5 integrin expression (51, 52) (J.A.M., T.L., and D.G., unpublished observations).
We recognize the likelihood that Survivin promotes melanoma metastasis by multiple mechanisms. Clearly apoptosis inhibition is important in cancer (53), and likely plays a role in both early and later steps in tumor progression. Our finding, however, that Survivin confers only modest protection against melanocyte anoikis suggests its role as a cytoprotectant may be more critical in metastatic cells where it is expressed at higher levels (25) rather than during the initial process of melanocyte escape from the epithelium. On the other hand, its demonstrated role here in promoting migration and invasion suggests a novel basis for its critical role in early tumor development. While conventional chemotherapy and Survivin-targeted therapies (30, 54) have been aimed at inducing apoptosis in tumor cells in patients with advanced disease, blocking Survivin at earlier stages may thwart melanoma cell motility and allow targeting of localized disease to prevent invasion and subsequent metastasis.
Since original submission of this manuscript, Mehrotra et al. (55) reported that Survivin enhances invasion and metastasis of breast cancer cells by upregulating fibronectin and activating FAK and Src kinases. While this provides separate validation of a role for Survivin in promoting metastasis and motility, it appears that the pathway we have identified in melanoma cells appears to be distinct in several respects. First, we have not observed upregulation of fibronectin (Supplementary Fig. S3) or phosphorylation of FAK or Src in cells over-expressing Survivin (J.A.M. and D.G., unpublished observations). Second, these authors reported that Survivin-enhanced motility does not require Akt activation or upregulation of integrins (55).
We thank Meenhard Herlyn for the WM35 and WM793 cells, and Dario Altieri for the pAd-GFP and pAd-Surv adenoviruses. This work was supported in part by NIH grant AR050102, the Department of Dermatology, and the Huntsman Cancer Foundation. We acknowledge use of the DNA/peptide and Flow cytometry core facilities supported by P30 CA042014 awarded to Huntsman Cancer Institute.
The authors declare no conflicts of interest.