Overexpression of DDX43 Mediates MEK Inhibitor Resistance Through RAS Upregulation in Uveal Melanoma Cells

Published OnlineFirst June 4, 2014; DOI: 10.1158/1535-7163.MCT-14-0095

Molecular Cancer Cancer Biology and Signal Transduction Therapeutics

Overexpression of DDX43 Mediates MEK Inhibitor Resistance through RAS Upregulation in Uveal Melanoma Cells

Grazia Ambrosini1, Raya Khanin2, Richard D. Carvajal1, and Gary K. Schwartz3

Abstract The majority of uveal melanomas carry oncogenic mutations in the G proteins GNAQ and GNA11, with consequent activation of the MAPK pathway. Selective MEK inhibitors, such as selumetinib, have shown clinical benefit in uveal melanoma. However, mechanisms of drug resistance limit their efficacy in some patients. Analysis of MEK inhibitor–resistant uveal melanoma cell lines revealed the induction of RAS protein expression and activity. This effect was mediated by the RNA helicase DDX43, which was remarkably overexpressed in these cells. Depletion of DDX43 in MEK inhibitor–resistant cells decreased RAS proteins and inhibited ERK and AKT pathways. On the contrary, ectopic expression of DDX43 in parental uveal melanoma cells induced RAS protein levels and rendered cells resistant to MEK inhibition. Similar to DDX43 depletion, downregulation of KRAS, HRAS, and NRAS inhibited downstream pathways in the resistant cells, overcoming mutant GNAQ signaling. We also analyzed the expression of DDX43 in liver metastases of patients with uveal melanoma by RT-PCR, and found a significant overexpression of DDX43 in patients who did not benefit from selumetinib therapy. In conclusion, DDX43 induces RAS protein expression and signaling, mediating a novel mechanism of MEK inhibitor resistance. The detection of DDX43 in patients with uveal melanoma could lead to more targeted therapies for this disease. Mol Cancer Ther; 13(8); 2073–80. 2014 AACR.

Introduction the small-molecule MEK inhibitor, selumetinib, can inhib- Uveal melanoma accounts for approximately 5% of all it pERK and cyclin D1, resulting in decreased viability of melanomas. Most arise from melanocytes within the uveal uveal melanoma cell lines (11). Furthermore, in patients tract, which consists of the iris, ciliary body, and choroid with uveal melanoma, selumetinib treatment can result in of the eye (1). This type of melanoma has a very high tumor shrinkage, and the sustained inhibition of pERK tendency to metastasize to the liver (2). As opposed to may be predictive of benefit (12). Recently, our group cutaneus melanoma, BRAF and RAS mutations are absent reported the results of a randomized phase II clinical trial or extremely rare in uveal melanoma (3). Rather, activat- of selumetinib versus temozolomide in patients with ing mutations in G-protein a-subunits, GNAQ or GNA11, metastatic uveal melanoma (13). These results indicated have been detected in nearly 80% of primary uveal mel- an improvement in the response rate and a significant anomas (4–6), and mediate the activation of the MAPK median progression-free survival. Thus, selumetinib is pathway (4). The prognosis of patients with metastatic the first drug in medical oncology to ever show a clinical benefit in patients with this disease. Despite these prom- uveal melanoma is poor with a median 1-year survival de novo rate of less than 30% (7, 8). Nearly all patients with uveal ising results, " " and "acquired" resistance to this melanoma die due to metastatic disease and there are drug and other MEK inhibitors remains a major clinical limited therapeutic options (9, 10). We have reported that problem (14). For example, resistance to selumetinib has been described in colorectal cancer cells carrying BRAF and RAS mutations, where resistance is mediated by the Authors' Affiliations: 1Jennifer Goodman Linn Laboratory of New Drug amplification of the driving oncogene (15, 16). KRAS Development, Department of Medicine; 2Department of Bioinformatics, overexpression has also been implicated in resistance to Memorial Sloan-Kettering Cancer Center; and 3Department of Medicine, Columbia University Medical Center, New York, New York MEK inhibition in colon carcinoma (17). RAS activation has been described as a mechanism of acquired resistance Note: Supplementary data for this article are available at Molecular Cancer Therapeutics Online (http://mct.aacrjournals.org/). to BRAF inhibitors in cutaneus melanoma with BRAF mutation, due to receptor tyrosine kinase (RTK) activation Corresponding Author: Grazia Ambrosini, Columbia University, The Her- bert Irving Comprehensive Cancer Center, 1130 St. Nicholas Ave, Room (18) or NF1 loss (19). In uveal melanoma with GNAQ/ 207, New York, NY 10032. Phone: 212-851-4905; Fax: 212-851-4906; GNA11 mutations, the mechanisms of resistance to MEK E-mail: [email protected] inhibitors are unknown and understanding this process doi: 10.1158/1535-7163.MCT-14-0095 will prove critical to develop more effective therapies for 2014 American Association for Cancer Research. the treatment of this disease.

www.aacrjournals.org 2073

Ambrosini et al.

DDX43, also called HAGE, is a member of the DEAD- Immunoblotting and RAS-GTP pull-down assays box family of ATP-dependent RNA helicases. It was The cells and the biopsy tissues were lysed in RIPA first identified as a cancer/testis antigen, and it is highly buffer supplemented with protease inhibitor cocktail expressed in many tumor types compared with normal tablets (Roche Diagnostics) and 1 mmol/L NaVO3. Total tissues (20). In particular, DDX43 is overexpressed in protein concentration of the lysates was measured by BCA 50% of acute myeloid leukemias (21), and its expression assay (Bio-Rad), and equal amounts of protein were was associated with advanced disease and poor prog- loaded on 4% to 12% PAGE gels (Invitrogen). Polyviny- nosis (22). These helicases browse RNA molecules and lidene difluoride membranes were blocked with 5% non- promote the dissociation of the RNA from ribonucleo- fat dried milk in PBS buffer containing 0.1% Tween-20 proteins (23). In this way, they support processes such (PBST) and probed with antibody for pERK, ERK, pAKT, as gene transcription and posttranscriptional regula- pan-AKT, pRB, c-Jun, cyclin D1, tubulin (Cell Signaling tion, including pre-mRNA splicing, translation initia- Technology), DDX43 (Abcam), KRAS, HRAS (Abnova), tion/elongation, and RNA degradation (24–26). Their and NRAS (Santa Cruz Biotechnology). GTP-bound RAS altered expression levels have been also implicated in was measured using the RAF RAS-binding domain (RBD) tumor initiation, progression, and maintenance (27). pull-down and Detection Kit (Thermo Scientific), as Recently, DDX43 was found highly expressed in mela- instructed by the manufacturers. noma (28), and it promoted protein expression of NRAS by unwinding of mRNA secondary structures, and RNAi-mediated gene knockdown enhancing NRAS translation (29). Here, we report that DDX43-1 and -2 siRNA were from OriGene. siRNA the expression of DDX43 is elevated in uveal melanoma against GNAQ (sc-35429), KRAS (sc-35731), NRAS (sc- cells with acquired resistance to selumetinib and in 36004), HRAS (sc-29340), and control siRNA (sc-37007) patients with metastatic uveal melanoma who did not were purchased from Santa Cruz Biotechnology. They benefit from selumetinib treatment. Furthermore, we were transfected using Lipofectamine RNAiMAX reagent demonstrated that DDX43 regulates RAS protein (Life Technologies). The human DDX43 cDNA clone in expression and AKT activation. Selumetinib-resistant pCMV6-XL5 was obtained from OriGene. cells became sensitive to AKT inhibition, suggesting alternative strategies for the treatment of patients with Tissue sample collection and qRT-PCR uveal melanoma with high DDX43 expression and resis- Matched tumor biopsies were collected from patients tance to MEK inhibitors. with metastatic uveal melanoma (clinicaltrials.gov, # NCT01143402). The protocol was approved by the Institu- Materials and Methods tional Review Board of Memorial Sloan-Kettering Cancer Center (New York, NY), and all patients signed informed Cell lines and reagents consent forms. Total RNA was extracted from flash-frozen The GNAQ (Q209P)-mutant cell lines, Omm1.3 and specimens and cell lines using TRIzol (Life Technologies) Mel270, were kindly provided by Dr. Bruce Ksander followed by DNase digestion and RNeasy (Qiagen) col- (Harvard Medical School, Boston, MA) in 2009. The cells umn purification following the manufacturer’s protocol. have been sequenced for the presence of mutations in The RNA integrity was verified using an Agilent Bioana- codons 209 (exon 5) and 183 (exon 4) of GNAQ and lyzer 2100 (Agilent). Of note, 1 mg of total RNA was reverse GNA11. A karyotype test was also performed for the transcribed using the Thermoscript RT-PCR system (Life cell line in 2012, confirming the authenticity of the cell Technologies). The resultant cDNA was used in RT-PCR lines as reported previously (30). The resistant cell lines, reactions using an iCycler (Bio-Rad) with predesigned Rs-Omm13 and Rs-Mel270, were derived from the cor- TaqMan Gene expression Assays for DDX43 and GAPDH responding cell lines after at least 3 weeks of exposure to genes (Life Technologies). Triplicate C values were aver- escalatingdosesupto1mmol/L of selumetinib. Cells t aged, the targets were interpolated from the standard were cultured in RPMI medium supplemented with curves and normalized to GAPDH. The results are shown 10% FBS, 100 U/mL penicillin and 100 mg/mL strepto- as DDX43/GAPDH ratio and analyzed with the Wilcoxon mycin, and maintained at 37C in 5% CO2. Cells were Mann–Whitney test for two-sided P value. treated with selumetinib (AZD6244, ARRY-142866, AstraZeneca) and MK2206 (supplied by Merck; ref. 31), Results PD0325901 (32), and GSK1120212 (Selleck Chemicals; ref. 33). GNAQ-mutant cell lines with acquired resistance to selumetinib display differential expression of RAS Cell viability assays and DDX43 and deregulated MAPK/AKT pathways Cells were plated in 96-well plates, and treated with the To explore mechanisms of drug resistance, we gener- indicated concentrations of drugs. Viability was assessed ated MEK inhibitor–resistant uveal melanoma cell lines. after 4 days of treatment using the Cell Counting Kit 8 The GNAQ-mutant cell lines, Omm1.3 and Mel270, which from Dojindo Molecular Technologies according to the are sensitive to selumetinib treatment (34), were chroni- manufacturer’s instructions. Cell viability is expressed as cally exposed to increasing doses of selumetinib. The a percentage of untreated cells. resulting cell lines, Rs-Omm1.3 and Rs-Mel270, showed

2074 Mol Cancer Ther; 13(8) August 2014 Molecular Cancer Therapeutics

DDX43 Induces RAS Signaling and Drug Resistance

Figure 1. Selumetinib-resistant cells exhibit RAS and DDX43 upregulation and increased ERK/AKT pathway activation. A, after continuous exposure to selumetinib, the cell lines Rs-Omm1.3 and Rs-Mel270 became resistant to the drug compared with their parental cells, Omm1.3 and Mel270. Cell viability on day 4 was calculated as percentage of untreated controls. Each point is a mean SD. B, immunoblot analysis of parental and selumetinib-resistant Omm1.3 (left) and Mel270 (right) cells treated with increasing concentrations of selumetinib for 24 hours. Both selumetinib-resistant cell lines Rs-Omm1.3 and Rs-Mel270 showed sustained expression of cyclin D1, pRB, c-Jun, pAKT, and RAS independently of treatment. C, whole-cell lysates were subjected to pull-down assays with GST-bound CRAF RAS-binding domain (RBD) for active RAS. The cells were treated with selumetinib for 24 hours or treated and washed afterwards and kept in culture for 24 hours (ND, no drug). The active RAS pull-down (top bands) and total lysates (bottom bands) were immunoblotted with a pan-RAS antibody. D, parental and selumetinib-resistant cells were analyzed for DDX43 expression by immunoblotting, before and after selumetinib treatments. E, real-time PCR was performed using gene-speciﬁc primers and the values were normalized to GAPDH as housekeeping gene. Relative mRNA levels of DDX43 in parental and selumetinib-resistant cell lines are shown. Each experiment was performed in triplicate. Bars, SD.

that IC50 were at least 10-fold higher than their parental We next analyzed the expression of RAS in all the cell cells, Omm1.3 and Mel270, in cell viability assays (Fig. lines. The parental cells expressed low levels of RAS, 1A). The cells also showed cross resistance to the MEK which slightly increased with higher doses of selumetinib. inhibitors PD0325901 and GSK1120212 (Supplementary In contrast, RAS protein levels were elevated in the resis- Fig. S1). We then examined the expression levels and tant cells regardless of treatments (Fig. 1B). phosphorylation status of different components of the To test whether RAS protein expression corresponded ERK pathway after MEK inhibition. In parental cells to increased activity, we used RAF-RBD pull-down Omm1.3 and Mel270, increasing doses of selumetinib for assays. In the parental Omm1.3 and Mel270 cells, basal 24 hours caused a decrease in pERK, cyclin D1, and pRb RAS expression and activity were relatively low and (Fig. 1B), whereas c-Jun and pAKT were slightly temporarily induced by selumetinib without changes in increased. In Rs-Omm1.3 and Rs-Mel270, pERK was also RAS expression (Fig. 1C). This activation of RAS is pos- inhibited by the treatments (Fig. 1B), even though the sibly mediated by the downregulation of the Sprouty signal seems to rebound over time (Supplementary Fig. proteins by ERK inhibition, as reported (35). We have S2). Furthermore, the downstream signaling was deregu- also shown that GNAQ-mutant cells have elevated lated in these cells, with sustained expression of cyclin D1, expression of Spry2 that is downregulated by selumetinib pRb, and high expression of c-Jun and pAKT, indepen- (34). In contrast, both RAS expression and activity were dently of treatments (Fig. 1B). elevated in Rs-Omm1.3 and Rs-Mel270 cell lines, and

www.aacrjournals.org Mol Cancer Ther; 13(8) August 2014 2075

Ambrosini et al.

AB 120 Omm1.3 (V) Omm1.3 DDX43 100 Mel270 (V) 80 Mel270 DDX43

% Viability 40

0 0 0.25 0.5 12 Selumetinib [mmol/L] CDE 120 Ctr- 100 Ctr+Selum 80 siDDX43- siDDX43+Selum 60

% Viability 40

0 Rs-Omm1.3 Rs-Mel270

Figure 2. DDX43 mediates the expression of RAS and resistance to selumetinib. A, parental uveal melanoma (UM) cell lines Omm1.3 and Mel270 were transfected with DDX43 cDNA or an empty vector (V). Cell lysates were subjected to Western blot analysis for expression of DDX43, KRAS, HRAS, NRAS, pERK, and pAKT. B, the cell line Omm1.3 and Mel270 overexpressing DDX43 are less sensitive to increasing doses of selumetinib, after 4 days of treatments. Graph reports the mean SD of three independent experiments. C, DDX43 downregulation decreases RAS protein expression/activity and restores MEK sensitivity. siRNA-mediated knockdown of DDX43 (þ) and control siRNA () in the MEK inhibitor–resistant cell lines decreases KRAS, HRAS, NRAS and downstream signaling molecules pERK, pAKT, and c-Jun. D, pull-down assays for RAS activity in cells after DDX43 downregulation. Total RAS, DDX43 levels, and tubulin are also shown. E, cell viability of Rs-Omm1.3 and Rs-Mel270 cells was measured after DDX43 knockdown and 3 days of selumetinib treatments. Bars are the mean of three experiments, expressed as percentage of untreated cells, SD. , P < 0.001 and #, P < 0.002 for comparison of DDX43 siRNA with and without selumetinib treatment.

remained active even when selumetinib was removed for in uveal melanoma cells, and found a remarkable 24 hours (Fig. 1C). These results show that the selumeti- upregulation of DDX43 in the MEK inhibitor–resistant nib-resistant cells have sustained activation of the RAS, cells, compared with low or no expression in their ERK, and AKT pathways. parental cells (Fig. 1D). Treatments with selumetinib To determine whether the increase of RAS protein did not further increase DDX43 expression either with corresponded to higher RAS mRNA levels, we per- increasing doses (Fig. 1D) or over time (Supplementary formed RT-PCR from parental and MEK inhibitor–resis- Fig. S2). The mRNA expression of DDX43 was analyzed tant cells with or without treatment. We did not ﬁnd any by RT-PCR in all the cell lines, showing a 9-fold and difference in RNA expression for KRAS or NRAS (data 7.6-fold increase in Rs-Omm1.3 and Rs-Mel270 cells, not shown). Rs-Omm1.3 and Rs-Mel270 cells were compared with the respective parental cell lines sequenced for MEK, KRAS, NRAS, HRAS mutations (Fig. 1E). and these were negative. FISH analysis was performed toexcludegeneampliﬁcationoftheoncogeneGNAQor DDX43 mediates the expression of RAS and RAS proteins. Assays to detect activation of RTK were resistance to selumetinib also negative. To determine whether DDX43 was involved in MEK It has been reported that the DEAD-box RNA heli- inhibitor resistance, DDX43 was overexpressed in the case antigen DDX43 regulates NRAS protein expres- selumetinib-sensitive uveal melanoma parental cell lines, sion through unwinding of duplex RNA in melanoma Omm1.3 and Mel270. In these cells, DDX43 induced the cells (29). Thus, we analyzed the expression of DDX43 protein levels of KRAS, HRAS, NRAS, and increased

2076 Mol Cancer Ther; 13(8) August 2014 Molecular Cancer Therapeutics

DDX43 Induces RAS Signaling and Drug Resistance

Figure 3. RAS proteins regulate cell signaling in MEK inhibitor–resistant cells. A, selumetinib-resistant Rs-Omm1.3 (left) and Rs-Mel270 (right) cells were transfected with siRNA for KRAS, HRAS, and NRAS and analyzed for downstream signaling, that is, pMEK, pERK, pAKT, and c-Jun by immunoblotting. B, the cells were also transfected with GNAQ siRNA and analyzed for downstream signaling. C, KRAS and HRAS are necessary for survival of MEK inhibitor–resistant uveal melanoma (UM) cells. Rs-Omm1.3 and Rs-Mel270 cells depleted of each RAS protein or GNAQ were tested in cell viability assays after 72 hours of siRNA transfection. , P < 0.0001 and , P < 0.001 for comparison of cells transfected with siKRAS and siHRAS versus control siRNA, for both cell lines. pAKT (Fig. 2A). Surprisingly, the levels of pERK did not responded to a reduction in RAS activity in both MEK change, possibly because the pathway is regulated by inhibitor–resistant cell lines (Fig. 2D). Most importantly, mutant GNAQ through PKC activation, as reported the depletion of DDX43 caused inhibition of cell viability (36). Nevertheless, higher expression of RAS has impor- of the MEK inhibitor–resistant cells, and this effect was tant implications as it regulates multiple pathways, further enhanced by the treatment with selumetinib including the PI3K/AKT pathway (37, 38). Next, we (Fig. 2E), suggesting that DDX43 is required for MEK tested whether DDX43 affected the response to selume- inhibitor resistance. tinib. DDX43-expressing uveal melanoma cells were more To conﬁrm that the effects of DDX43 depletion are resistant to selumetinib treatment compared with cells mediated by RAS downregulation, each RAS protein transfected with an empty vector (V) in viability assays was silenced by gene-speciﬁc siRNA. KRAS and HRAS (Fig. 2B), suggesting that DDX43 may mediate mechan- knockdown caused downregulation of pMEK, pERK, isms of resistance to MEK inhibition through RAS and and pAKT in both cell lines (Fig. 3A), whereas NRAS AKT activation. didnot.c-JunwasdownregulatedbyKRASandNRAS, We also evaluated the effect of DDX43 depletion in the whereas HRAS regulated expression of c-Jun only in Rs- MEK inhibitor–resistant cells by using two siRNA Mel270. We have reported that GNAQ silencing inhibits sequences (siDDX43-1, Fig. 2C; and siDDX43-2, Supple- pERK and pAKT, and decreases cell viability of the mentary Fig. S3). Silencing of DDX43 corresponded to parental GNAQ-mutant uveal melanoma cells (31). We decreased protein expression of KRAS, NRAS, HRAS and also tested the contribution of mutant GNAQ on down- the downstream effectors pERK, pAKT, and c-Jun in both stream signaling pathways in the resistant cells. GNAQ resistant cell lines (Fig. 2C). DDX43 depletion did not downregulation induced minimal effects on down- change KRAS or NRAS mRNA levels (Supplementary stream pathways (Fig. 3B). Similar to DDX43 down- Fig. S4) as previously reported (29), suggesting that regulation, KRAS and HRAS depletion decreased cell DDX43 regulates RAS posttranscriptionally. The decrease viability of the selumetinib-resistant cell lines, whereas in RAS protein levels induced by DDX43 depletion cor- NRAS had partial effects (Fig. 3C). GNAQ depletion did

www.aacrjournals.org Mol Cancer Ther; 13(8) August 2014 2077

Ambrosini et al.

who were treated on the phase II study of selumetinib. A We analyzed biopsies from 14 patients, 6 of whom had partial response or stable disease for 16 weeks, deﬁned as Omm1.3 Rs-Omm1.3MeI270 Rs-MeI270 – + – + – + – + AKTi "responders" (R) according to RECIST, and 8 patients who had disease progression before 16 weeks with no evidence pAKT of radiological response, deﬁned as "nonresponders" (NR). AKT The "responder" patients had low expression of DDX43 Tubulin (Fig. 5A). In contrast, DDX43 was highly expressed in the "nonresponder" group, and there was a statistically signif- B P ¼ 120 icant association with poor outcome ( 0.045). The tissues Omm1.3 from six representative patients were also analyzed by 100 Rs-Omm1.3 immunoblotting for protein expression. The patients who Mel270 80 responded to therapy had complete inhibition of pERK Rs-Mel270 with selumetinib treatment, whereas the "nonresponders" 60 showed none or minimal inhibition (Fig. 5B). The protein

% Viability 40 levels of DDX43 in biopsy tissues were low in "responders," whereas it was highly expressed in "nonresponders." Sim- 20 ilarly, levels of total KRAS and NRAS were elevated in the 0 "nonresponder" group, while HRAS was not detectable in 0 0.6 1.2 2.5 5 these specimens. AKTi [mmol/L] Discussion Figure 4. Inhibition of AKT decreases cell viability of MEK inhibitor– The RAS/RAF/MEK/ERK pathway is often activated resistant cells. A, parental and selumetinib-resistant uveal melanoma by genetic alterations in upstream signaling molecules, (UM) cells were treated with 1.5 mmol/L of MK2206 (AKTi) for 24 hours and analyzed by immunoblotting for target inhibition of pAKT. B, parental and such as RAS and BRAF in cutaneous melanoma, and resistant cells were exposed to increasing concentrations of AKTi (0–5 GNAQ/11 in uveal melanoma (4, 5). Hyperactivation of mmol/L) and analyzed in viability assays after 4 days of treatment. Graph shows the mean of three independent experiments SD. A 4.5 n = 14 4 P = 0.045 not affect cell viability of Rs-Omm1.3 and Rs-Mel270 3.5 cells (Fig. 3C), further demonstrating that RAS proteins 3 become the dominant mediators of cell signaling in 2.5 these cells. 2 1.5 AKT inhibition inhibits cell viability of selumetinib- 1 Median DDX43 expression resistant GNAQ-mutant cells 0.5 Elevated DDX43 and RAS expression mediated the 0 activation of downstream survival pathways, especially RNR pAKT (Figs. 1B and 2A). Thus, we assessed the effects of B AKT inhibition in uveal melanoma cell lines. The AKT inhibitor MK2206 (AKTi) inhibited AKT phosphorylation in parental and MEK inhibitor–resistant cells (Fig. 4A). In cell proliferation assays, AKT inhibition had minimal effects in the parental cells, Omm1.3 and Mel270 (Fig. 4B), while it greatly inhibited cell viability of the resistant cell lines, Rs-Omm1.3 and Rs-Mel270 (Fig. 4B), suggesting that in these cells the DDX43/RAS- mediated activation of AKT is necessary for cell survival, and its inhibition may be a means of overcoming MEK inhibitor resistance. Figure 5. DDX43 mRNA and protein levels are higher in liver biopsies of "responder" versus "nonresponder" patients. A, DDX43 expression was analyzed by RT-PCR in biopsies of 14 patients (6 "responders" and 8 DDX43 is highly expressed in liver metastasis of "nonresponders"). Triplicate values were normalized with GAPDH as patients with uveal melanoma housekeeping gene, and reported as a box plot showing signiﬁcant To determine the relevance of DDX43 expression in association with poor outcome in patients with uveal melanoma treated with selumetinib. B, liver biopsies from six representative patients (P) patients with uveal melanoma, we assessed the expression before () and after (þ) selumetinib treatment were analyzed by of DDX43 by RT-PCR in liver metastasis of patients with immunoblotting with antibodies for pERK, DDX43, KRAS, NRAS, uveal melanoma carrying GNAQ or GNA11 mutations and tubulin.

2078 Mol Cancer Ther; 13(8) August 2014 Molecular Cancer Therapeutics

DDX43 Induces RAS Signaling and Drug Resistance

this signaling cascade results in dysregulated cell prolif- long-term treatments. However, DDX43 seemed to eration and malignant transformation. Targeting this mediate intrinsic resistance to selumetinib in patients pathway with MEK inhibitors has shown activity in dif- with uveal melanoma with elevated expression at base- ferent types of cancers (39, 40), including uveal melanoma line. It was reported that DDX43 is highly expressed in (13). However, resistance to MEK inhibitors is common chronic myeloid leukemia due to gene demethylation and undermines the efficacy of these treatments (41, 42). (22), and correlated with poorer prognosis in terms of Resistance to MEK inhibitors has been extensively cytogenetic response to drug treatments. Similar described in cutaneus melanoma with BRAF and RAS changes in DDX43 gene methylation may occur in uveal mutations (43). Also in colon carcinoma, MEK inhibition melanoma. leads to gene amplification of RAS and activation of DDX43 has been also identified as a tumor-specific gene downstream pathways (16). Activated RAS has been expressed in sarcoma (20), and it has been associated with linked to MEK inhibitor resistance as it drives the activa- poor clinical outcome in patients with breast cancer (45). tion of both the PI3K-AKT and MEK-ERK pathways (37). Recently, several studies have provided new insights on Here, we found that in uveal melanoma cell lines, the RNA helicases and their emerging roles in cell signaling. sustained inhibition of MEK leads to increased expression For example, DDX3, another member of this family, was of DDX43, and consequent induction of RAS and down- found to regulate the Wnt–b-catenin network in a stream pathways. Furthermore, using RT-PCR, we found genome-wide siRNA screen in human embryonic kidney that DDX43 is overexpressed in patients who did not cells (46). Another report showed DDX5 as a regulator of respond to selumetinib treatment, confirming the in vitro alternative splicing of HRAS (47). findings. Altered expression of DDX43 has been reported Our results provide evidence that DDX43 increased the in human cancers (27, 44). In particular, DDX43 was expression and activation of RAS in uveal melanoma cells required for cell proliferation of malignant melanoma- with GNAQ mutation, and mediated resistance to MEK initiating cells by promoting the expression of NRAS inhibitors. through RNA unwinding (29). However, this is the first Finally, these findings would indicate that DDX43 may time that DDX43 has been linked to MEK inhibitor resis- have predictive value in determining who will most tance through the regulation of RAS. Interestingly, it was benefit from a MEK inhibitor therapy in this disease. reported that MEK inhibitor–resistant colon carcinoma cells showed increased KRAS expression, which repre- fl sented a potential mechanism of MEK inhibitor resistance Disclosure of Potential Con icts of Interest R.D. Carvajal is a consultant/advisory board member for AstraZeneca. (17). However, the modalities of KRAS protein induction No potential conflicts of interest were disclosed by the other authors. remained to be established. Recently, Lito and colleagues have reported that RAF inhibitors cause relief of negative feedback with activation Authors' Contributions Conception and design: G. Ambrosini, R.D. Carvajal, G.K. Schwartz of RAS and rebound of ERK activity, without changes in Development of methodology: G. Ambrosini, G.K. Schwartz RAS expression (35). We have shown a similar activation of Acquisition of data (provided animals, acquired and managed patients, RAS, which was induced by MEK inhibition in the parental provided facilities, etc.): R.D. Carvajal, G.K. Schwartz Analysis and interpretation of data (e.g., statistical analysis, biostatis- cells. However, this effect was sustained in uveal melano- tics, computational analysis): G. Ambrosini, R. Khanin, R.D. Carvajal, G. ma selumetinib-resistant cells, where the expression of K. Schwartz Writing, review, and/or revision of the manuscript: G. Ambrosini, R.D. both DDX43 and RAS was elevated. DDX43 silencing Carvajal, G.K. Schwartz decreased RAS proteins and downstream signaling, i.e., Administrative, technical, or material support (i.e., reporting or orga- pERK and pAKT, making DDX43 a mediator of MEK nizing data, constructing databases): G. Ambrosini, R.D. Carvajal, G.K. Schwartz resistance that could represent a class effect to all MEK Study supervision: G. Ambrosini, R.D. Carvajal, G.K. Schwartz inhibitors. Little and colleagues reported an increase in c- Jun in colorectal cancer cells with acquired resistance to selumetinib (16). Indeed, c-Jun was highly expressed in Grant Support G. Ambrosini, Cycle for Survival (Philanthropy). MEK inhibitor–resistant cells, and it could be downregu- The costs of publication of this article were defrayed in part by the lated by knockdown of DDX43 and RAS, but not GNAQ, payment of page charges. This article must therefore be hereby marked further confirming the importance of RAS signaling in advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. these cells. In our cells, DDX43 mediated "acquired" resistance to Received February 3, 2014; revised May 5, 2014; accepted May 16, 2014; the MEK inhibitor, as its expression increased after published OnlineFirst June 4, 2014.

References 1. Strickland D, Lee JA. Melanomas of eye: stability of rates. Am J 3. Zuidervaart W, van Nieuwpoort F, Stark M, Dijkman R, Packer L, Epidemiol 1981;113:700–02. Borgstein AM, et al. Activation of the MAPK pathway is a common 2. Singh AD, Bergman L, Seregard S. Uveal melanoma: epidemiologic event in uveal melanomas although it rarely occurs through mutation of aspects. Ophthalmol Clin North Am 2005;18:75–84. BRAF or RAS. British J Cancer 2005;92:2032–8.

www.aacrjournals.org Mol Cancer Ther; 13(8) August 2014 2079

Ambrosini et al.

4. Van Raamsdonk CD, Bezrookove V, Green G, Bauer J, Gaugler L, 25. Linder P, Jankowsky E. From unwinding to clamping - the DEAD box O'Brien JM, et al. Frequent somatic mutations of GNAQ in uveal RNA helicase family. Nature Rev Molec Cell Biol 2011;12:505–16. melanoma and blue naevi. Nature 2009;457:599–602. 26. Jarmoskaite I, Russell R. DEAD-box proteins as RNA helicases and 5. Van Raamsdonk CD, Griewank KG, Crosby MB, Garrido MC, Vemula chaperones. WIREs RNA 2011;2:135–52. S, Wiesner T, et al. Mutations in GNA11 in uveal melanoma. The New 27. Robert F, Pelletier J. Perturbations of RNA helicases in cancer. WIREs England J Med 2010;363:2191–9. RNA 2013;4:333–49. 6. Onken MD, Worley LA, Long MD, Duan S, Council ML, Bowcock AM, 28. MathieuMG,KnightsAJ,PawelecG,RileyCL,WernetD,Lemonnier et al. Oncogenic mutations in GNAQ occur early in uveal melanoma. FA, et al. HAGE, a cancer/testis antigen with potential for melanoma Investig Ophthalmol Visual Sci 2008;49:5230–4. immunotherapy: identification of several MHC class I/II HAGE- 7. Gragoudas ES, Egan KM, Seddon JM, Glynn RJ, Walsh SM, Finn SM, derived immunogenic peptides. Cancer Immunol Immunother 2007; et al. Survival of patients with metastases from uveal melanoma. 56:1885–95. Ophthalmology 1991;98:383–9;. 29. Linley AJ, Mathieu MG, Miles AK, Rees RC, McArdle SE, Regad T. The 8. Kath R, Hayungs J, Bornfeld N, Sauerwein W, Hoffken K, Seeber S. helicase HAGE expressed by malignant melanoma-initiating cells is Prognosis and treatment of disseminated uveal melanoma. Cancer required for tumor cell proliferation in vivo. J Biol Chem 2012;287: 1993;72:2219–23. 13633–43. 9. Patel M, Smyth E, Chapman PB, Wolchok JD, Schwartz GK, Abramson 30. Griewank KG, Yu X, Khalili J, Sozen MM, Stempke-Hale K, Bernatchez DH, et al. Therapeutic implications of the emerging molecular biology C, et al. Genetic and molecular characterization of uveal melanoma cell of uveal melanoma. Clin Cancer Res 2011;17:2087–100. lines. Pigment Cell Melanoma Res 2012;25:182–7. 10. Yonekawa Y, Kim IK. Epidemiology and management of uveal mela- 31. Ambrosini G, Musi E, Ho AL, de Stanchina E, Schwartz GK. Inhibition of noma. Hemat/Oncol Clin North Am 2012;26:1169–84. mutant GNAQ signaling in uveal melanoma induces AMPK-dependent 11. Ambrosini G, Pratilas CA, Qin LX, Tadi M, Surriga O, Carvajal RD, et al. autophagic cell death. Mol Cancer Ther 2013;12:768–76. Identification of unique MEK-dependent genes in GNAQ mutant uveal 32. Thompson N, Lyons J. Recent progress in targeting the Raf/MEK/ERK melanoma involved in cell growth, tumor cell invasion, and MEK- pathway with inhibitors in cancer drug discovery. Curr Opinion Phar- resistance. Clin Cancer Res 2012;18:1–10. macol 2005;5:350–6. 12. Carvajal RD AG, Wolchok JD, Chapman PB, Dickson MA, D'Angelo 33. Gilmartin AG, Bleam MR, Groy A, Moss KG, Minthorn EA, Kulkarni SG, SP, et al. Pharmacodynamic activity of selumetinib to predict radio- et al. GSK1120212 (JTP-74057) is an inhibitor of MEK activity and graphic response in advanced uveal melanoma. J Clin Oncol 2012;30 activation with favorable pharmacokinetic properties for sustained in (suppl; abstr 8598). vivo pathway inhibition. Clin Cancer Res 2011;17:989–1000. 13. Carvajal RD SJ, Quevedo F, Milhem MM, Joshua AM, Kudchadkar RR, 34. Ambrosini G, Pratilas CA, Qin LX, Tadi M, Surriga O, Carvajal RD, et al. et al. Phase II study of selumetinib (sel) versus temozolomide (TMZ) in Identification of unique MEK-dependent genes in GNAQ mutant uveal gnaq/Gna11 (Gq/11) mutant (mut) uveal melanoma (UM). J Clin Oncol melanoma involved in cell growth, tumor cell invasion, and MEK 2013;31 (suppl; abstr CRA9003). resistance. Clin Cancer Res 2012;18:3552–61. 14. Little AS, Balmanno K, Sale MJ, Smith PD, Cook SJ. Tumour cell 35. Lito P, Pratilas CA, Joseph EW, Tadi M, Halilovic E, Zubrowski M, et al. responses to MEK1/2 inhibitors: acquired resistance and pathway Relief of profound feedback inhibition of mitogenic signaling by RAF remodelling. Biochem Society Trans 2012;40:73–8. inhibitors attenuates their activity in BRAFV600E melanomas. Cancer 15. Corcoran RB, Dias-Santagata D, Bergethon K, Iafrate AJ, Settleman J, Cell 2012;22:668–82. Engelman JA. BRAF gene amplification can promote acquired resis- 36. Chen X, Wu Q, Tan L, Porter D, Jager MJ, Emery C, et al. Combined tance to MEK inhibitors in cancer cells harboring the BRAF V600E PKC and MEK inhibition in uveal melanoma with GNAQ and GNA11 mutation. Science Signal 2010;3:ra84. mutations. Oncogene 2013 Oct 21. [Epub ahead of print]. 16. Little AS, Balmanno K, Sale MJ, Newman S, Dry JR, Hampson M, et al. 37. Karnoub AE, Weinberg RA. Ras oncogenes: split personalities. Nature Amplification of the driving oncogene, KRAS or BRAF, underpins Rev Mol Cell Biol 2008;9:517–31. acquired resistance to MEK1/2 inhibitors in colorectal cancer cells. 38. Rodriguez-Viciana P, Warne PH, Dhand R, Vanhaesebroeck B, Gout I, Science Signal 2011;4:ra17. Fry MJ, et al. Phosphatidylinositol-3-OH kinase as a direct target of 17. Wang Y, Van Becelaere K, Jiang P, Przybranowski S, Omer C, Sebolt- Ras. Nature 1994;370:527–32. Leopold J. A role for K-ras in conferring resistance to the MEK inhibitor, 39. Neuzillet C, Tijeras-Raballand A, de Mestier L, Cros J, Faivre S, CI-1040. Neoplasia 2005;7:336–47. Raymond E. MEK in cancer and cancer therapy. Pharmacol Ther 2014; 18. Nazarian R, Shi H, Wang Q, Kong X, Koya RC, Lee H, et al. Melanomas 141:160–71. acquire resistance to B-RAF(V600E) inhibition by RTK or N-RAS 40. Friday BB, Adjei AA. Advances in targeting the Ras/Raf/MEK/Erk upregulation. Nature 2010;468:973–7. mitogen-activated protein kinase cascade with MEK inhibitors for 19. Whittaker SR, Theurillat JP, Van Allen E, Wagle N, Hsiao J, Cowley GS, cancer therapy. Clin Cancer Res 2008;14:342–6. et al. A genome-scale RNA interference screen implicates NF1 loss in 41. Little AS, Smith PD, Cook SJ. Mechanisms of acquired resistance to resistance to RAF inhibition. Cancer Dis 2013;3:350–62. ERK1/2 pathway inhibitors. Oncogene 2013;32:1207–15. 20. Martelange V, De Smet C, De Plaen E, Lurquin C, Boon T. Identification 42. Poulikakos PI, Solit DB. Resistance to MEK inhibitors: should we co- on a human sarcoma of two new genes with tumor-specific expres- target upstream? Sci Signal 2011;4:pe16. sion. Cancer Res 2000;60:3848–55. 43. Das Thakur M, Stuart DD. The evolution of melanoma resistance 21. Adams SP, Sahota SS, Mijovic A, Czepulkowski B, Padua RA, Mufti reveals therapeutic opportunities. Cancer Res 2013;73:6106–10. GJ, et al. Frequent expression of HAGE in presentation chronic myeloid 44. Abdelhaleem M. Over-expression of RNA helicases in cancer. Anti- leukaemias. Leukemia 2002;16:2238–42. cancer Res 2004;24:3951–3. 22. Roman-Gomez J, Jimenez-Velasco A, Agirre X, Castillejo JA, Navarro 45. Chan S. A-FTMA, McArdle S, Moseley P, Johnson C, Andrew R. HAGE G, San Jose-Eneriz E, et al. Epigenetic regulation of human cancer/ (DDX43) protein expression as an independent biomarker of poor testis antigen gene, HAGE, in chronic myeloid leukemia. Haematolo- clinical outcome of breast cancer (BC) and potential as a therapeutic gica 2007;92:153–62. target for ER-negative BC. J Clinical Oncol 2012;30 (suppl; abstr 1013). 23. Fuller-Pace FV. DExD/H box RNA helicases: multifunctional proteins 46. Cruciat CM, Dolde C, de Groot RE, Ohkawara B, Reinhard C, Korswa- with important roles in transcriptional regulation. Nucleic Acids Res gen HC, et al. RNA helicase DDX3 is a regulatory subunit of casein 2006;34:4206–15. kinase 1 in Wnt-beta-catenin signaling. Science 2013;339:1436–41. 24. Glisovic T, Bachorik JL, Yong J, Dreyfuss G. RNA-binding proteins 47. Camats M, Guil S, Kokolo M, Bach-Elias M. P68 RNA helicase (DDX5) and post-transcriptional gene regulation. FEBS Lett 2008;582: alters activity of cis- and trans-acting factors of the alternative splicing 1977–86. of H-Ras. PloS ONE 2008;3:e2926.

2080 Mol Cancer Ther; 13(8) August 2014 Molecular Cancer Therapeutics

Overexpression of DDX43 Mediates MEK Inhibitor Resistance through RAS Upregulation in Uveal Melanoma Cells

Grazia Ambrosini, Raya Khanin, Richard D. Carvajal, et al.

Mol Cancer Ther 2014;13:2073-2080. Published OnlineFirst June 4, 2014.

Updated version Access the most recent version of this article at: doi:10.1158/1535-7163.MCT-14-0095

Supplementary Access the most recent supplemental material at: Material http://mct.aacrjournals.org/content/suppl/2014/06/04/1535-7163.MCT-14-0095.DC1

Cited articles This article cites 43 articles, 15 of which you can access for free at: http://mct.aacrjournals.org/content/13/8/2073.full#ref-list-1

Citing articles This article has been cited by 1 HighWire-hosted articles. Access the articles at: http://mct.aacrjournals.org/content/13/8/2073.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at Subscriptions [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://mct.aacrjournals.org/content/13/8/2073. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.