Protein Kinase D3 Sensitizes RAF Inhibitor RAF265 in Melanoma Cells by Preventing Reactivation of MAPK Signaling

RAS–RAF mitogen-activated protein kinase (MAPK) signaling cascade plays a central role in the regulation of cell proliferation and survival, whereas the deregulation of this pathway frequently occurs in human cancers (1–3). As mutations in RAS or BRAF occur in more than 30% of human cancers, these proteins are very attractive therapeutic targets in many cancer types. Among them, BRAF mutations occur in about 7% of human cancers, with highest prevalence in melanomas (66%) and thyroid (35%–70%) tumors (4, 5). Interestingly, 80% of all BRAF mutations are concentrated on a single substitution of glutamic acid for valine (V600E) within the kinase domain (4). Compared with BRAF, mutations in the 2 RAF isoforms, ARAF and CRAF, are rarely found in human cancers, which is likely due to lower basal kinase activities (6, 7). All 3 RAS isoforms (KRAS, NRAS, and HRAS) are found mutationally activated in 30% of all human cancers, with highest prevalence in pancreas (90%), colon (50%), thyroid (50%), and lung (30%) cancers (1, 2). Although the RAS oncogene has been studied for more than 3 decades, there is no drug on the market that sufficiently inhibits RAS, despite extensive efforts to inhibit activated RAS with low molecular weight inhibitors (3, 8). As an alternative therapeutic strategy, RAF and MAP/ERK kinase (MEK) inhibitors have been developed to inhibit the pathway downstream of RAS, and only 1 RAF inhibitor Sorafineb has been approved by Food and Drug Administration and several inhibitors are still undergoing clinical trials (8). However, clinical responses from these reagents are not as effective or durable as expected, drug resistance frequently occurs in tumors treated with RAF or MEK inhibitors (2, 8). PLX4032 seems to be an effective RAF inhibitor in malignant melanoma with an overall response rate of 81%, but responsive time from patients ranged from 2 to more than 18 months and this could limit the long-term efficacy of the drug (9). A recent report suggested that upregulation of N-RAS and other RTK signals such as platelet-derived growth factor receptor β (PDGFRβ) are responsible for the acquired resistance to PLX4032 (10). Potential mechanisms of resistant to RAF or MEK inhibitors in RAS or BRAF mutant cancers can be attributed to either coactivation of parallel or downstream survival pathways prior to drug treatments or compensatory activation of alternative survival pathways upon drug administration (11–22). In either situation, combinatorial inhibition of multiple survival pathways is required to achieve potent antitumor effects.

RAS signaling pathway is more complex than a linear RAS–RAF–MEK signaling cascade (1). Activated RAS protein can interact with more than 20 effectors including RAF, phosphatidylinositol 3-kinases (PI3K), RAC, RAL, and phospholipase C epsilon to regulate cell proliferation, survival, and differentiation (2). There are multiple feedback loops known to activate downstream of RAS which are differentially regulated depending on the genetic background and tumor lineage being studied (23, 24). These feedback loops can lead to compensatory activation of parallel survival pathways upon drug treatments, and tumor cells are flexible at utilizing the signal pathways for growth and transformation, resulting in rapid drug resistance. Clinical reports show that activation of the MAPK pathway was induced via S6K–PI3K–RAS signaling in tumor samples from patients treated with RAD001, an inhibitor of PI3K pathway (17). In neuroendocrine tumor cells, inhibitor of RAF strongly induced AKT phosphorylation reflecting an activation of PI3K pathway, and inhibitors of PI3K pathways induced ERK phosphorylation indicating MAPK pathway activation (25). Dual targeting of both PI3K and MAPK signaling pathways showed more potent antitumor effect than single treatment alone (17, 21, 25).

Despite the intensive research efforts on RAS signaling, our understanding of its regulation is still limited. To identify potential modulators of RAS signaling pathways and to uncover the molecular mechanisms underlying resistance to RAF or MEK inhibitors, we conducted a siRNA screen in combination with a RAF inhibitor (RAF265) to identify genes and/or pathways that sensitize to RAF265 treatment in a BRAF (V600E) mutant melanoma cell line (A2058) which is insensitive to RAF265-induced cell death. By using this approach, we identified protein kinase D3 (PRKD3) that when knocked down could enhance cell killing by RAF265 in A2058. PRKD3 is 1 of 3 members in the protein kinase D (PKD) family which includes PRKD1 and PRKD2. The physiologic functions of PRKD3 are not well understood. Similar to PRKD1 and PRKD2, PRKD3 has been shown to function in protein trafficking and as a histone deacetylase kinase (26, 27). One report showed that PRKD3 enhanced CCK-mediated pancreatic amylase secretion via MEK–ERK–RSK signaling and this process was activated by GI hormone (28). Another report showed that PRKD3 expression levels were elevated in human prostate cancers compared with normal tissues; PRKD3 overexpression activated AKT and ERK in prostate cancer cell lines and promoted cell growth and survival (29). Essential role of PRKD3 in prostate cancer cells is likely due to regulation of MAPK signaling (29). By using A2058 and A375 (RAF265 sensitive cell line) as melanoma cellular models, we showed that PRKD3 inhibition cooperates with RAF265 to prevent the reactivation of MAPK signaling pathway, induce PARP cleavage and caspase activity, interrupt cell-cycle progression, and inhibit colony formation. Finally, we showed that the PRKD3 inhibition sensitizes with multiple RAF and MEK inhibitors in a panel of melanoma cell lines, suggesting that PRKD3 functionally interacts with the MAPK signaling pathway. Thus, PRKD3 provides a potential cancer target to develop effective therapeutic strategies to overcome or prevent drug resistance from RAF or MEK inhibitors.

The kinome siRNA SMARTpool library directed against 779 kinases was purchased from Dharmacon (catalogue no. G-003500). RNAi screens were prepared with 2 compound doses for 12 plates in A2058 cells (6 plates, in duplicate). Briefly, 4 μL of 206 nmol/L siRNAs in serum-free medium were stamped into each well, 0.03 μL of DharmaFECT 1 in 4 μL serum-free medium was mixed and added into each well followed by 30-minute incubation at room temperature to form a lipid/siRNA complex. Then 1,500 cells in 25 μL complete medium were loaded on top of siRNA–lipid complex. Final concentration of siRNA for each reaction was 25 nmol/L. The cells were incubated at 37°C with 5% CO₂. At 24 hours after siRNA transfection, 5 μL of RAF265 was added into each well to make the final concentration of RAF265 at 0.4 μmol/L. Cell viability was analyzed at 72 hours post-RAF265 treatment by using CellTiter-Glo (CTG) Assay (Promega), and data were acquired by using an Envision (PerkinElmer). From the primary screen, the CTG value was normalized by 1-dimensional (1D) normalization scheme based on plate median of screened 384-well plates. The formula for calculating the 1D-normalized value (denoted as x_1D) is x_1D = log x′_1D, where x′_1D = CTG/median_plate CTG. The normalized Z score (NZ) was calculated for every x_1D. The formula used for calculating NZ_1D is NZ_1D = [x_1D − median (x_1D)]/MAD(x_1D). MAD (x_1D) is the median absolute deviation of x_1D and is equal to 1.4826 × median (|x_1D − median (x_1D)|).

Additional Materials and Methods are shown in the Supplementary Data.

RAF265 is an orally bioavailable small molecule that is known to inhibit CRAF, wild-type BRAF, mutant BRAF (V600E), and VEGF receptor 2 (VEGFR-2) and is currently in phase I clinical trials for advanced melanoma (30, 31). As other RAF or MEK inhibitors, RAF265 showed less efficacy in tumors with RAS mutations compared with tumors with a BRAF mutation (31). RAF265 also shows resistance in some tumors bearing the BRAF V600E mutation such as the melanoma cell line A2058. By using A2058 as the cellular model, we conducted a synthetic lethal siRNA screen to identify genetic sensitizers of RAF265. The siRNA SMARTpool kinome library of 779 kinases was used for the screen. We chose to run our screen in the presence of 0.4 μmol/L RAF265 where 20% growth inhibition was achieved, leaving a suitable window for additional, siRNA-enhanced cellular toxicity. As positive technical controls, we used siRNAs targeting Polo-like kinases 1 (PLK1), which is a regulator of mitosis and, when depleted, results in cell lethality (32). BRAF siRNAs were used as the biological positive controls and luciferase (LUC) siRNAs were used as the negative controls. Duplicates were used for each siRNA and NZ were calculated. As shown in Figure 1A and B, control LUC siRNA centered at NZ of 0 and PLK1 siRNAs have a NZ ranging from −17 to −28, indicating a robust signal to noise in the screens. We considered a primary hit, a gene whose knockdown has a minimal effect on cell growth (NZ < 1.5) in the absence of RAF265 but enhances cell killing (NZ > 4.5) in the presence of a low dose of RAF265 (Fig. 1C). Twelve primary hits that met our predefined criteria were identified including MGC5601, C9ORF96, ALPK1, CDC7, PRKD3, PDGFRB, PAPSS2, PIK3R4, RAGE, RPS6KB2, DAPK1, and MAPK11 (Fig. 1C). To further evaluate the sensitizers, we generated dose curves of RAF265 for all the hits with SMARTpool siRNAs and observed significant IC₅₀ shifts with 4 hits; ALPK1, MAPK11, PRKD3, and PIK3R4 (Fig. 1D). The mRNA knockdown efficiencies of the 4 target genes were confirmed by quantitative real time PCR (RT-PCR; Fig. 1E).

To prioritize hits to follow up and to minimize off-target effects when using a SMARTpool siRNA approach, we tested individual siRNAs of the 4 hits for sensitization with RAF265. On the basis of the mRNA knockdown levels of the target genes, we selected 2 potent siRNAs from the 4 individual siRNAs for ALPK1, MAPK11, PIK3R4, and PRKD3 (left panels, Fig. 2A–D) and measured sensitization over a RAF265 dose response (right panels, Fig. 2A–D). Among them, PRKD3 was the only gene in which multiple independent siRNAs resulted in a significant IC₅₀ shift for RAF265 and distinguished itself as the best sensitizer identified from this study (Fig. 2D). To confirm the sensitization observed between PRKD3 siRNAs and RAF265, we established 3 independent stable cell lines with inducible PRKD3 short-hairpin RNAs (shRNA) expressed in A2058 cells. The sequences of PRKD3 shRNAs were different from the PRKD3 siRNAs used in the screen. The 3 PRKD3 shRNAs when induced with DOX resulted in 2- to 8-fold IC₅₀ shifts for RAF265 and the sensitization correlated with protein knockdown efficiencies of PRKD3 shRNAs (Fig. 2E). For the control condition without DOX induction, knockdown of PRKD3 protein levels were not observed, and the sensitization was not significant (Fig. 2E). We showed that knockdown of PRKD3 by using multiple siRNAs and shRNAs enhances the cell killing effects of RAF265 in the resistant melanoma cell line A2058.

We analyzed pERK and pAKT levels after RAF265 treatment with or without PRKD3 knockdown in A2058 cells. Efficient PRKD3 knockdown was achieved by siRNA transfection as shown by Western blot (Fig. 3B). RAF265 treatments resulted in the reduction of pERK levels in a dose-dependent manner at 2 hours post-RAF265 treatment (Fig. 3A). However, at 72 hours post-RAF265 treatment, both pERK and pAKT levels were upregulated as shown with control siRNA treatments, pERK levels were increased 1.6- to 4.2-fold and pAKT levels were increased more than 2-fold (Fig. 3B). After PRKD3 siRNA treatments, pERK levels were reduced to 0.2 to 0.8, which corresponds to a 2- to 20-fold change compared with the condition without PRKD3 siRNA transfection (Fig. 3B). The pAKT levels were reduced to 0.4 to 0.6, which corresponds to a 4- to 6-fold change compared with the condition without PRKD3 siRNA transfection (Fig. 3B). In our system, total ERK and AKT levels were not affected by compound treatment or PRKD3 knockdown (Fig. 3B). Similar results were obtained with PRKD3 shRNA transfection (Supplementary Fig. S1). Collectively, we showed that PRKD3 knockdown prevented the reactivation of pAKT and pERK induced by RAF265 treatments in A2058 cells.

To further validate PRKD3 in modulating MAPK signaling, pERK and pAKT levels were analyzed after MEK inhibitor U0126 treatment, with and without PRKD3 knockdown. Sensitization of the MEK inhibitor U0126 was observed after PRKD3 knockdown (Supplementary Fig. S2A). The pERK levels were reduced at 2 hours and were upregulated at 72 hours post-U0126 treatment (Supplementary Fig. S2B and C). After PRKD3 siRNA transfection in combination with U0126, the pERK levels were reduced, in particular at high doses of U0126 (Supplementary Fig. S2C). Even though pAKT levels were slightly reduced with PRKD3 siRNA transfection in combination with U0126 treatment, pAKT levels relative to total AKT levels were not affected (Supplementary Fig. S2C). We have shown that both RAF and MEK inhibitors tested alone were unable to induce a sustained inhibition of pERK at 72 hours after compound treatment. In contrast, we were able to restore pERK inhibition in combination with PRKD3 siRNAs. In addition, MEK and pMEK levels were found to be reduced either by PRKD3 siRNA treatment or by RAF265 treatment alone and further reduced after PRKD3 siRNAs in combination with RAF265 (Fig. 3B).

Because elevated CRAF protein levels contributed to resistance to RAF inhibition in a subset of BRAF mutant tumor cells, we tested whether CRAF levels were altered by PRKD3 knockdown (15). We observed increased phospho-CRAF (Ser338) after RAF265 treatments (Fig. 3B), which has been reported for other RAF inhibitors (33). However, PRKD3 siRNA mediated knockdown had no effect on pCRAF or CRAF levels, indicating PRKD3 affects pERK levels in a CRAF-independent manner. We also used genetic approaches to validate the sensitization between PRKD3 inhibition and BRAF but not CRAF. PRKD3 siRNAs in combination with BRAF siRNA treatments resulted in a greater growth inhibition compared with any of the single treatment (Fig. 3C). However, neither the combination of PRKD3 and CRAF siRNA treatments nor the combination of BRAF and CRAF siRNAs showed potent growth inhibition (Fig. 3C). The knockdown efficiencies of PRKD3, BRAF, and CRAF siRNAs were validated by Western blots (Fig. 3C).

To further understand the mechanism of PRKD3 knockdown–induced RAF265 sensitization, we analyzed apoptosis markers, PARP cleavage and caspase activity. RAF265 alone was not able to induce detectable PARP cleavage or caspase activity in A2058 cells (Figs. 3B and 4A). Similar observations were reported for MEK inhibitor AZD6244 which was shown to induce cell-cycle arrest and growth inhibition rather than apoptosis in melanoma cell lines and xenograft models (34). PRKD3 siRNA in combination with RAF265 treatments induced significant PARP cleavage and increased caspase activities, indicating PRKD3 knockdown sensitized RAF265 to induce apoptosis (Figs. 3B and 4A).

Previous report showed that cyclin D1, a cell-cycle regulator for G₁ phase entry, was reduced after RAF265 treatments (31). Cyclin D1 was analyzed in A2058 cells in the presence or absence of PRKD3 siRNAs. Consistent with previous reports, RAF265 treatments resulted in a reduction of cyclin D1, and the reduction was dose dependent (Fig. 3B). PRKD3 siRNA alone did not affect cyclin D1 expression, but PRKD3 siRNA in combination with RAF265 enhanced the reduction of cyclin D1, suggesting an interruption of cell-cycle progression (Fig. 3B). Consistent with this observation, cell-cycle profiles showed that PRKD3 siRNAs or 0.5 μmol/L RAF265 alone partially blocked G₂–M progression, but PRKD3 siRNAs in combination with 0.5 μmol/L RAF265 completely blocked G₂–M progression, result in a 0% of cells in G₂–M (Fig. 4B, Supplementary Fig. S3). It was shown that PRKD3 was phosphorylated at its activation loops during mitosis, suggesting that PRKD3 activity is regulated in a cell-cycle dependent manner (35). Although PRKD3 siRNAs or 0.5 μmol/L RAF265 alone resulted in a reduction of G₀–G₁ cell numbers, we did not observe cooperative effects in reducing G₀–G₁ cell numbers (Fig. 4B, Supplementary Fig. S3). It is possible that the cooperative effects of PRKD3 inhibition and RAF265 in reduction of cyclin D1 is a consequence from blocking of G₂–M entry.

Finally, the sensitization of PRKD3 knockdown and RAF265 was tested in a colony formation growth assay. PRKD3 siRNAs or RAF265 (0.5 and 2 μmol/L) alone partially inhibited the colony formation of A2058 cells, but PRKD3 siRNAs in combination with 0.5 μmol/L or 2 μmol/L RAF265 completely inhibit A2058 cell growth in colonies (Fig. 4C). Taken together, we have shown that PRKD3 knockdown mediates RAF265 sensitization to prevent the reactivation of MAPK signaling, induce apoptosis markers, reduce cell-cycle progression, and induce tumor cell growth inhibition at high density in plastic and low density colony formation assays in A2058 cells.

We analyzed a RAF265 sensitive melanoma cell line A375, which also harbors the BRAF (V600E) mutation. A375 cells are sensitive to RAF265 with an IC₅₀ of 0.3 μmol/L as shown by control LUC siRNAs transfected cells (Fig. 5A). After PRKD3 siRNA transfection, the IC₅₀ for RAF265 was shifted from 0.3 μmol/L to 0.16 μmol/L for both PRKD3 siRNAs (Fig. 5A). Next, colony formation assays were done. PRKD3 siRNAs or RAF265 (0.05 μmol) alone only partially inhibited A375 cell growth in colony formation, but PRKD3 siRNAs in combination with 0.05 μmol/L RAF265 completely inhibited cell growth in colony formation (Fig. 5B).

Similar to A2058 cells, pERK was reactivated after 72 hours of RAF265 treatment and reduced after the combination with PRKD3 knockdown (Fig. 5D). pMEK and MEK levels were reduced after RAF265 treatment, and further reduced after combination with PRKD3 knockdown (Fig. 5D). In A375, RAF265 resulted in a reduction of cyclin D1, and the reduction was dose dependent (Fig. 5C and D). PRKD3 siRNAs alone did not affect cyclin D1 expression, but PRKD3 siRNA in combination with RAF265 enhance the reduction of cyclin D1 (Fig. 5D). PRKD3 siRNA transfection or RAF265 treatments alone in A375 were not sufficient to induce detectable PARP cleavage, whereas PRKD3 siRNA transfection in combination with RAF265 treatments resulted in significant induction of PARP cleavage (Fig. 5C and D). We did not see changes on total CRAF after RAF265 treatment in A375 cells (Fig. 5D). Instead, we observed decreased phospho-CRAF (Ser338) after RAF265 treatments (Fig. 5D), it is likely that CRAF is an efficacy target of RAF265 in the A375 cell line. PRKD3 siRNA mediated knockdown had no effect on pCRAF or CRAF levels, indicating PRKD3 affects pERK levels in a CRAF-independent manner. In A375, PRKD3 siRNAs in combination with BRAF siRNA treatments resulted in a greater growth inhibition compared with any of the single treatments (Fig. 5E). However, neither the combination of PRKD3 and CRAF siRNA treatments nor the combination of BRAF and CRAF siRNAs showed additive growth inhibition (Fig. 5E). The knockdown efficiencies of PRKD3, BRAF, and CRAF siRNAs were validated by Western blots (Fig. 5E). The effects on levels of pERK, pMEK, cyclin D1, and PARP cleavage, as well as colony formation, after PRKD3 knockdown in combination with RAF265 are similar in A375 and A2058 cells, indicating this mechanism is shared by different tumor cells from the same lineage or with the same genetic background BRAF (V600E). These data also suggest that single agent sensitivity to RAF265 is not a requirement for combination PRKD3 knockdown–induced cell killing.

To better understand whether the sensitization between PRKD3 inhibition and RAF265 in A2058 and A375 cells is shared with other RAF or MEK inhibitors across cell lines of various lineage and genetic background, we tested additional RAF and MEK inhibitors in a panel of 12 cell lines (Fig. 6 and Supplementary Fig. S4). In addition to RAF265, we analyzed another RAF inhibitor PLX4032, and 2 MEK inhibitors, U0126 and PD0325901. We tested 12 cell lines including 6 melanoma cell lines carrying BRAF (V600E) mutation (RPMI7951, IGR39, A2058, A375, SKMEL5, and WM115), 4 cell lines bearing KRAS mutations of various lineages (PANC1, A549, SW620, and DU145), 1 nonmelanoma cell line with BRAF (V600E) mutation (SKHEP1, liver cancer), and 1 cell line which is wild type for both BRAF and RAS (G402, kidney cancer; Fig. 6).

For both RAF inhibitors RAF265 and PLX4032, 6 of 6 melanoma cell lines tested showed sensitization with both PRKD3 siRNAs (Fig. 6). Among the lines tested, 3 cell lines RPMI7951, IGR39, and A2058 were resistant to RAF265 and PLX4032 (Fig. 6A and C) when 2 μmol/L of compound was used in combination with PRKD3 siRNAs (Fig. 6B and D). For 3 sensitive cell lines A375, SKMEL5, and WM115, 0.3 μmol/L of compound was used in the sensitization assays (Fig. 6B and D). Among the cell lines tested, the sensitization between PRKD3 knockdown and RAF inhibitors was only observed in the melanoma cell lines tested, with only 1 exception in which PRKD3 siRNAs sensitized with RAF265 in colon cancer cell line SW620. It could be that additional genetic lesions outside of the MAPK signaling pathway are necessary for PRKD3 sensitization in nonmelanoma lineages.

For both MEK inhibitors, the sensitization was mainly observed with melanoma cell lines (Supplementary Fig. S4). For MEK inhibitor U0126, sensitization with both PRKD3 siRNAs was observed in 5 of 6 melanoma cell lines showed (Supplementary Fig. S4B), with 1 exception that no sensitization was found with SKMEL5. For MEK inhibitor PD0325901, 4 of 6 melanoma cell lines exhibited sensitization with both PRKD3 siRNAs, but sensitization was not found with melanoma cells IGR39 and WM115 (Supplementary Fig. S4D). In addition, PRKD3 siRNAs sensitized with PD0325901 in 2 nonmelanoma cell lines, SW620 and G402. Collectively, we showed that PRKD3 inhibition sensitizes with RAF and MEK inhibitors in multiple melanoma cell lines. Therefore, it seems that the sensitization between PRKD3 inhibition and RAF or MEK inhibitors is in melanoma lineage selective.

Primary and acquired resistance to RAF and MEK inhibitors has been linked to reactivation of the MAPK pathway and PI3K pathway. Rebound of pERK levels has been observed for RAF inhibitor such as PLX4032 and MEK inhibitor like PD0325901 (12, 15, 36). Compensatory activation AKT has been reported both for the RAF inhibitor RAF265 and MEK inhibitor AZD6244 (AstraZeneca; refs. 20, 25, 37, 38). Here, we show that inhibition of PRKD3 sensitizes with RAF265 in a resistant melanoma cell line (A2058) by preventing rebound of pERK and pAKT. Our data support a compound-induced resistance mechanism in which the reactivation of the MAPK pathway and PI3K pathway act coordinately to promote cell survival when RAF and/or MEK are inhibited. We observed that pAKT was reduced when PRKD3 was knocked down together with RAF265 in A2058 cells, but not in A375, a difference that could be attributable to the presence of the PTEN deletion in the A2058 cells (20).

The mechanism of how PRKD3 interacts with pERK and pAKT remains to be clarified. PKCϵ has been shown to phosphorylate and regulate PRKD3 activity in prostate cancer cells (29). We tested whether PKCϵ knockdown can phenocopy the PRKD3 knockdown in sensitizing RAF inhibitor. However, we did not observe any sensitization effects (data not shown). Among the 3 PKD proteins, PRKD1 is the most extensively studied and has been implicated in a broad range of cellular process and can be activated by a variety of regulatory peptides (39–41). It has been shown that PRKD1 phosphorylates RIN1 and releases it from competing with RAF for binding to RAS, thus resulting in an activation of the RAF–MEK–ERK pathway (42). It will be interesting to investigate whether PRKD3 shares any function with PRKD1 such as RIN1 phosphorylation. We have tested whether PRKD1 or RIN1 siRNAs can sensitize RAF265 in killing A2058. We did not observe any sensitization effects (data not shown). In melanoma cells, PRKD2 protein expression was undetectable and thus it was not tested in our sensitization studies. Even though PRKD members share some cellular functions, they can be differentially expressed and exert different functions (28). Further investigation of PRKD3 specific cellular function(s) will be required to better understand how PRKD3 ablation augments RAF265 activity in melanoma. One limitation of our screen is that RAF265 can inhibit multiple RAF isoforms, a variety kinases, and VEGFR (43). Similar to Sorafenib, the efficacy RAF265 in vivo may not be due exclusively to the activity on BRAF (V600E) (44). Thus, the sensitization seen with PRKD3 knockdown and RAF265 could be because of a combinatorial effect(s) with other targets in addition to RAF isoforms and VEGFR. The lack of selectivity of RAF265 complicates the interpretation of PRKD3 knockdown sensitization and confounds the ability to clearly define a mechanism of action at this time.

This study is the first demonstration that PRKD3 inhibition can sensitize with RAF or MEK inhibitors in BRAF (V600E) melanoma cells. Our current data support a model in which PRKD3 could prevent reactivation of MAPK signaling caused by RAF or MEK inhibitors and sensitize with these inhibitors to kill resistant tumor cells. PRKD3 is a potentially druggable kinase because there are known inhibitors of this class of enzymes (44, 45). It is not clear whether the kinase activity or a potential scaffold activity of PRKD3 is required for sensitization with the RAF or MEK inhibitor(s). Although our understanding of how PRKD3 modulates pERK and pAKT in response to RAF265 requires further study, PRKD3 provides a synthetic lethal target opportunity to develop effective combination therapy to overcome drug-induced resistance caused by RAF or MEK inhibitors.

M. Labow and L.A. Gaither are shareholders of Novartis. The other authors disclosed no potential conflicts of interest.

We thank S. Zhao and C. Mickanin for screen reagent preparation, and C. Voliva, D. Stuart, L. De Parseval, T. Nagel, and P. Finan for advice and discussion.

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.