Acquired Resistance to BRAF Inhibition Can Confer Cross-Resistance to Combined BRAF/MEK Inhibition

ORIGINAL ARTICLE Acquired Resistance to BRAF Inhibition Can Confer Cross-Resistance to Combined BRAF/MEK Inhibition Kavitha Gowrishankar 1, Stephanie Snoyman 1, Gulietta M. Pupo 1, Therese M. Becker 1, Richard F. Kefford 1 and Helen Rizos 1 Aberrant activation of the BRAF kinase occurs in B6% of melanomas, and although BRAF inhibitors have shown significant early clinical success, acquired resistance occurs in most patients. Resistance to chronic BRAF inhibition often involves reactivation of mitogen-activated protein kinase (MAPK) signaling, and the combined targeting of BRAF and its downstream target MAPK/ERK kinase (MEK) may delay or overcome resistance. To investigate the efficacy of combination BRAF and MEK inhibition, we generated melanoma cell clones resistant to the BRAF inhibitor GSK2118436. These BRAF inhibitor resistant sublines acquired resistance through several distinct mechanisms, including the acquisition of activating N-RAS mutations and increased accumulation of COT1. These alterations uniformly promoted MAPK reactivation and most conferred resistance to MEK inhibition and to the concurrent inhibition of BRAF and MEK. These data indicate that melanoma tumors are likely to develop heterogeneous mechanisms of resistance, many of which will confer resistance to multiple MAPK inhibitory therapies. Journal of Investigative Dermatology (212) 132, 185 1859; doi:1.138/jid.212.63; published online 22 March 212 INTRODUCTION Constitutively activating mutations in the serine/threonine kinase BRAF are found in a large proportion of metastatic melanomas (Davies et al., 22; Long et al., 211). The majority of these mutations result in a single amino-acid substitution of valine by glutamic acid at amino acid 6, and this leads to a 5-fold increase in the kinase activity of BRAF (Davies et al., 22; Wan et al., 24). As a result, most melanomas display a dependency on the RAF/MEK/ERK mitogen-activated protein kinase (MAPK) pathway, as shown by early clinical success with the RAF-targeted inhibitors Vemurafenib (PLX432) and GSK2118436 (Flaherty et al., 21; Kefford et al., 21). Despite the marked initial responses to BRAF inhibitors, tumor regrowth occurs in most patients (Kefford et al., 21), with a median progression-free survival of 5.3 months (Flaherty et al., 21; Chapman et al., 211). Resistance to 1 Westmead Institute for Cancer Research, University of Sydney at Westmead Millennium Institute and Melanoma Institute Australia, Westmead Hospital, Westmead, New South Wales, Australia Correspondence: Helen Rizos, Westmead Institute for Cancer Research, University of Sydney at Westmead Millennium Institute, Westmead Hospital, Westmead, New South Wales 2145, Australia. E-mail: helen.rizos@sydney.edu.au Abbreviations: DR, drug resistant; ERK, extracellular signal regulated kinase; IC 5, half-maximal inhibitory concentration; MAPK, mitogen-activated protein kinase; MEK, MAPK/ERK kinase; mtor, mammalian target of rapamycin; PDGFRb, platelet-derived growth factor receptor-b; p-erk, phosphorylated extracellular signal regulated kinase; 3K, phosphatidylinositol 3-kinase; PTEN, phosphatase and tensin homolog; prb, retinoblastoma protein; shrna, short hairpin RNA Received 16 August 211; revised 26 January 212; accepted 3 January 212; published online 22 March 212 chronic BRAF inhibition involves reactivation of MAPK signaling and/or induction of prosurvival signals. For instance, a subset of BRAF inhibitor resistant melanoma tumors restored MAPK activity by expressing truncated forms of BRAF or by elevating the expression of the serine/threonine kinases BRAF, CRAF, or COT1 (Montagut et al., 28; Corcoran et al., 21; Johannessen et al., 21; Poulikakos et al., 211). Similarly, activating mutations in N-RAS (Q61K/ R) and MEK1 (C121S) triggered phosphorylated extracellular signal regulated kinase (p-erk) signaling in BRAF inhibitor - resistant cell lines and clinical samples (Nazarian et al., 21; Wagle et al., 211). Moreover, the persistent expression of the receptor tyrosine kinases platelet-derived growth factor receptor-b (PDGFRb) and IGF-1R conferred BRAF inhibitor resistance in a MAPK-independent manner that may involve enhanced phosphatidylinositol 3-kinase (3K)/AKT pathway activation (Nazarian et al., 21; Villanueva et al., 21; Shi et al., 211). In line with these data, PTEN (phosphatase and tensin homolog) loss also activated 3K/AKT signaling and conferred partial resistance to BRAF inhibition in BRAF V6E - positive melanoma cell lines (Paraiso et al., 211). Importantly, recent studies indicate that MAPK reactivation can modulate sensitivity to MAPK/ERK kinase (MEK) inhibition. For instance, melanoma cell lines harboring oncogenic N-RAS were sensitive to MEK inhibitors (Montagut et al., 28; Nazarian et al., 21), whereas tumor cells expressing mutant MEK1 or COT1 were refractory to MEK inhibition and maintained ERK phosphorylation in the presence of these inhibitors (Emery et al., 29; Johannessen et al., 21). There are currently limited data available on the activity of combining BRAF and MEK inhibitors, although the concurrent use of these inhibitors overcame resistance to 185 Journal of Investigative Dermatology (212), Volume 132 & 212 The Society for Investigative Dermatology

single agents in the setting of MEK mutations and COT1 overexpression (Emery et al., 29; Johannessen et al., 21). In this report, we demonstrate that distinct mechanisms of acquired resistance to BRAF inhibition can originate from a single melanoma tumor cell. These resistance drivers included the acquisition of activating N-RAS mutations and increased accumulation of COT1. These distinct alterations uniformly reactivated MAPK signaling and typically conferred cross-resistance to MEK inhibitors and the combination of MEK and BRAF inhibitors. These data indicate that melanoma tumors from a single patient are likely to develop heterogeneous mechanisms of resistance, many of which will confer resistance to multiple MAPK inhibitory therapies. RESULTS Generation of melanoma cell clones with acquired resistance to the RAF kinase inhibitor GSK2118436 To examine acquired resistance to the potent and highly selective BRAF kinase inhibitor, GSK2118436 (), we generated resistance in the human melanoma - derived cell line (Jiang et al., 21). This cell line harbors activating mutations in BRAF (V6E), EGFR (P753S), and CDK4 (R24C), and is exquisitely sensitive to BRAF inhibition (half-maximal inhibitory concentration (IC 5 ) 6nM; Figure 1a). As expected, cells responded to increasing concentrations of by undergoing a G1 cell cycle arrest followed by potent apoptosis (Figure 1b and c). The BRAF inhibitor induced proliferative arrest was associated with activation of the retinoblastoma protein (prb; i.e, loss of a Relative viability (percentage of control) c 1 8 6 4 2 2 4 6 8 1 (nm) nm 1 nm 9± 3± 8±1 88±2 14±1 Annexin V-APC 78±9 b d nm 1nM Time (hours) ±1 62±2 29± 9±2 7±1 1±1 4±1 48±3 1 nm 1 4 24 48 72 p-erk Y24 Total ERK p27 Kip1 Cyclin D1 p-prb S87/811 Total prb Figure 1. Analysis of the BRAF inhibitor sensitive melanoma cell line. (a) Viability curve for cell line treated with increasing concentrations of for 72 hours (relative to DMSO-treated controls; mean±sd; n ¼ 3). (b) Cell cycle distribution of cells treated with either DMSO ( nm) or (1 nm) for 72 hours. (c) Dual-color flow cytometric Annexin V analysis for apoptosis of cells treated with DMSO ( nm) or (1 nm) for 72 hours. APC, allophycocyanin;, propidium iodide. (d) cells were treated with 1 nm for increasing duration. The effects on extracellular signal regulated kinase (ERK) activation and cell cycle regulators were determined by immunoblotting. hyperphosphorylated and total prb), accumulation of the CDK inhibitor p27 Kip1, and loss of cyclin D1 expression. Importantly, activation of prb and cell cycle arrest was preceded by decreases in p-erk that were detectable within 1 hour and up to 24 hours after exposure to (Figure 1d). At 5 weeks after continuously exposing cells to 1 nm, single cell derived drug-resistant clones emerged at a frequency of B1in1 5 cells. Five of these drugresistant (DR) clones (, 4, 6, 8, and 9) were expanded and characterized. All five DR clones retained the activating BRAF, EGFR, and CDK4 mutations found in the parental line (data not shown) and showed strong resistance to (Figure 2a) with IC 5 not reached at 1, nm (data not shown). In all resistant DR clones, MAPK reactivation was confirmed using protein and gene expression analyses; p-erk expression was maintained in the presence and absence of the drug (Figure 2b), and the resistant sublines demonstrated a MEK/ERK transcriptome signature indicative of persistent MEK/ERK activation (Dry et al., 21; Nazarian et al., 21; Figure 2c). This gene set included many known transcription targets of ERK signaling, including MYC, ETV5, and the negative feedback regulators DUSP4/6 and SPRY1/2. As a result, the DR clones responded to BRAF inhibition with a continued, although somewhat slower, rate of proliferation (data not shown), with little evidence of -induced cell death (Figure 2d). Notably, although DR clones survived in response to BRAF inhibition, they all showed evidence of reduced S-phase entry that was associated with p27 Kip1 accumulation and reduced levels of cyclin D1 (Figure 2b). Analyses for mutations in the candidate oncogenes H-RAS, K-RAS, N-RAS, BRAF, and KIT revealed that and sublines carried the N-RAS Q61H -activating mutation at a frequency of B3%. This mutation was not detected in the parental cell line. Concordant with these data, unsupervised clustering of differential gene expression profiles showed that and share characteristics and group away from the parental and other DR clones (Supplementary Figure S1 online). We also noted that PTEN expression was retained, CRAF levels were not substantially elevated (Figure 3a), and truncated variant forms of BRAF were not detected in the DR sublines (Supplementary Figure S2 online). Furthermore, although AKT was activated (increased p-akt) and PDGFRb expression was elevated in the and sublines (Figure 3a), these clones did not show activation-associated PDGFRb phosphorylation in a phospho-rtk array or western immunoblotting (detailed analyses of are shown in Figure 3b), and they did not display an increased PDGFRbspecific gene signature (Wu et al., 28; Nazarian et al., 21; Supplementary Figure S3 online). Finally, the concurrent inhibition of MAPK with 3K/mammalian target of rapamycin (mtor; BEZ235) or PDGFRb (imatinib mesylate) did not trigger cell death (Figure 3c). Collectively, these data suggest that neither PDGFRb nor the 3K pathway mediate BRAF inhibitor resistance in the subline. The cell clone displayed elevated expression of the COT1 kinase (Figure 3a), and in accordance with a recent report www.jidonline.org 1851

(Johannessen et al., 21) short hairpin RNA (shrna)- mediated depletion of COT1 cooperated with BRAF inhibition to suppress viability, without affecting the sensitivity of the parental line (Figure 3d). Relative viability (percentage of control) 16 14 12 1 8 6 4 2 2 4 6 8 (nm) 1 + + + + + + p-erk Y24 Total ERK p27 kipl Cyclin D1 p-prb S87/811 Total prb + + + + + + MYC-6556 ETV5-145519 DUSP4-26541 ST3GAL5-389382 DUSP6-526288 MAFF-143138 SPRED2-31975 SSH1-34686 SPRY1-676669 SPRY1-331692 SPRY2-659575 DLX1-342672 ST3GAL5-52643 ETV5-77746 SPRED1-5777 MYC-627646 SH2B3-65631 PLAT-37144 RND3-421524 UCN2-422437 IER3-119367 SLC2A1-642731 CCND1-58261 ±1 62±2 29± 1±1 1 nm BRAFi 9±2 7±1 4±1 48±3 15. 12. 9. 6. 3.. Mean MEK/ERK activation signature expression 1 9 8 7 1 nm BRAFi 48±4 29± 21±6 65±8 16±2 18±5 ± 58±2 29±1 12±1 ± 69±6 14±1 9±1 52±1 35±1 14±3 61±6 16±1 16±1 61±2 23± 18± 76±7 17± 9± 61±4 26± 13±1 ± 78±2 14±1 8±1 Counts Annexin V-APC 1852 Journal of Investigative Dermatology (212), Volume 132

We then compared the gene expression profiles of the -derived resistant sublines with a signature predictive of MEK resistance, which appears to be driven upstream of RAF and independently of RAS or 3K pathway mutations (Dry et al., 21). Although only 6 of the 13 genes in this published resistance signature mapped to the Illumina platform, these genes showed significantly increased expression in the closely related - and -resistant sublines (Supplementary Figure S4 online). These data indicate that distinct mechanisms originating from a single melanoma cell line mediate MAPK reactivation and resistance to BRAF inhibition. Activating N-RAS reactivates MAPK and confers resistance to BRAF inhibition Activating N-RAS mutations (Q61K and Q61R) have previously been identified in two biopsy samples derived from BRAF inhibitor (PLX432) resistant melanoma metastases from a single patient (Nazarian et al., 21). To thoroughly examine the role of oncogenic N-RAS in regulating tumor growth and conferring resistance to BRAF inhibition, we initially silenced the expression of N-RAS using highly specific shrna molecules (Figure 4a). The suppression of N-RAS in the N-RAS wild-type and N-RAS mutant and sublines did not result in proliferative arrest + + + + + + p-akt S473 Total AKT PTEN PDGFRβ IGF-1R 1 2 1 2 1: p-egfr 2: p-pdgfrβ p-pdgfrβ Y751 Total PDGFRβ p-erk Y24 COT1 CRAF MelMS % Of cells in sub-g1 6 5 4 3 2 1 MAPK inhibition MAPK inhibition +imatinib MAPK inhibition +BEZ 235 BEZ235 Imatinib Total c-kit p-c-kit Y719 Total AKT p-akt S473 % Of cells in sub-g1 3 2 1 COT1 shrna + + COT1 shrna + + + + COT1 Figure 3. BRAF inhibitor resistant isogenic sublines display distinct mechanisms of acquired resistance. (a) parental and BRAF inhibitor resistant sublines were treated with DMSO ( ) or 1 nm ( þ ) for 24 hours, and the effects on known drug-resistance regulators were determined by immunoblotting. (b) Phosphorylation of platelet-derived growth factor receptor-b (PDGFRb) was analyzed in the parent and subline using phospho-rtk antibody arrays (left panel) and western immunoblotting (right panel). The positions of phosphorylated EGFR (positive in and ) and phosphorylated PDGFRb (negative in and ) are indicated. DR, drug resistant. (c) Impact of mitogen-activated protein kinase (MAPK) inhibition (1 nm and 5 nm GSK-MEKi) either alone or in combination with receptor tyrosine kinase inhibition (1 nm imatinib mesylate) or phosphatidylinositol 3-kinase (3K)/mammalian target of rapamycin (mtor) inhibition (1 mm BEZ235) on cell survival at 72 hours after treatment. The inhibitory activity of imatinib and BEZ235 on c-kit and AKT activation was confirmed in the c-kit mutant (c-kit W557 K558 ) melanoma cell line, MelMS. (d) and were transfected with COT1 short hairpin RNA (shrna; þ ) or control ( ) shrna. At 24 hours after transfection, cells were treated with DMSO ( ) or 1 nm ( þ ) for an additional 72 hours. COT1 silencing (right panel) and the effects on cell death (left panel) were measured 72 hours after drug addition. Figure 2. Mitogen-activated protein kinase (MAPK) reactivation in cells with acquired resistance to. (a) Viability curves of the parental and isogenic melanoma sublines treated with the indicated concentrations for 72 hours (relative to DMSO-treated controls; mean±sd; n ¼ 3). DR, drug resistant. (b) parental cells and BRAF inhibitor resistant sublines were treated with DMSO ( ) or 1 nm ( þ ) for 24 hours, and the effects on extracellular signal regulated kinase (ERK) activation and cell cycle regulators were determined by immunoblotting. (c; left panel) Heat map for MEK/ERK activation signature in each of the cell lines treated with DMSO ( ) or 1 nm ( þ ) for 24 hours. Color scale, log2-transformed expression (red, high; green, low) for each gene (row) normalized by the mean of all samples. The probe ID is shown after each gene symbol. (c; right panel) Histograms of mean log2-transformed expression of all transcripts included in the MEK/ERK activation signature (see left panel). MEK, MAPK/ERK kinase. (d) effects on cell cycle distribution and apoptosis. APC, allophycocyanin;, propidium iodide. www.jidonline.org 1853

N-RAS shrna + + + + + + + + + + + + p-erk Y24 Total ERK N-RAS % Of cells in sub-g1 6 4 2 N-RAS shrna + + + + + N-RAS Q61K + + p-erk Y24 Total ERK 3 p27 Kip1 Cyclin D1 p-prb S87/811 N-RAS Q61K % cells in sub-g1 2 1 + N-RAS Q61K Figure 4. Oncogenic N-RAS confers BRAF inhibitor resistance. (a) -derived resistant cell lines were transduced with N-RAS no. 1 short hairpin RNA (shrna; þ ) or control ( ) shrna. At 5 days after transduction, cells were treated with DMSO ( ) or 1 nm ( þ ) for an additional 72 hours. The effects on phosphorylated extracellular signal regulated kinase (p-erk) accumulation were detectable 24 hours after exposure to (left panel), and the downstream effects on cell death were measured 72 hours after drug addition (right panel). DR, drug resistant. (b) parental cell lines were lentivirally transduced with N-RAS Q61K or copgfp control, and after 1 days cells were treated with DMSO ( ) or 1 nm ( þ ). The effects on p-erk, cell cycle regulators, and N-RAS accumulation (left panel) and apoptosis (right panel) are shown. (data not shown) or apoptosis in the absence of the BRAF inhibitor, indicating that these resistant clones did not depend on N-RAS signaling for growth (Figure 4a). In the presence of BRAF inhibitor, however, the stable suppression of N-RAS expression suppressed ERK activation, reinstated sensitivity to, and resulted in potent cell death in the N-RAS mutant and sublines (Figure 4a and Supplementary Figure S5 online). This effect was highly specific, as suppression of N-RAS expression in the parental and cells (both wild type for N-RAS) had no effect on sensitivity (Figure 4a and Supplementary Figure S5 online). Conversely, the stable overexpression of MYCtagged N-RAS Q61K promoted accumulation of p-erk and conferred BRAF inhibitor resistance in the parental cell line (Figure 4b). Oncogenic N-RAS dampens sensitivity to MEK inhibition As shown in Figure 3a, activated N-RAS did not activate the 3K/AKT pathway in the and sublines. Consequently, we predicted that melanoma cells expressing oncogenic N-RAS signaled predominantly via the reactivated MAPK pathway and would retain sensitivity to MEK inhibition. Surprisingly, we found that the pattern of sensitivity to MEK inhibition, using the allosteric MEK inhibitor GSK112212 (GSK-MEKi; Gilmartin et al., 211), was significantly diminished in most BRAF inhibitor resistant DR sublines: GSK-MEKi IC 5 o2nm for parental cells (Figure 5a). Critically, the level of resistance to MEK inhibition paralleled the unsupervised clustering of our DR clones in genome-wide, differential expression patterns (Supplementary Figure S1 online). Specifically, both N-RAS mutant and clones showed diminished MEK inhibitor sensitivity; the clone (expresses COT1 and clusters closely with the parental line in gene profiling; Supplementary Figure S1 online) was sensitive to MEK inhibition and showed effective suppression of ERK phosphorylation (Figure 5b). The and clones, which are distinct from the other DR clones but share a common MEK resistance signature (Supplementary Figure S4 online), were highly resistant to MEK inhibition (Figure 5). Importantly, all resistant sublines showed minimal cell death in response to MEK inhibition but displayed a degree of S-phase inhibition that was associated with reduced cyclin D1 (Figure 5b and c). Finally, we also showed that the stable overexpression of MYC-tagged N-RAS Q61K conferred MEK inhibitor resistance in the parental cell line (Figure 5d). Dual inhibition of MEK and BRAF can restore partial sensitivity of melanoma cells with activating N-RAS and BRAF mutations parental cells were exquisitely sensitive to BRAF inhibition, and the combination of BRAF and MEK inhibition enhanced cell death in these cells (Figure 6). Similarly, in the majority of DR clones, targeting the MAPK pathway at both the BRAF and MEK nodes increased the percentage of cells 1854 Journal of Investigative Dermatology (212), Volume 132

a Relative viability (percentage of control) 14 12 1 8 6 4 2 1 2 3 4 5 GSK-MEKi (nm) b GSK-MEKi + + + + + + p-erk Y24 Total ERK p27 Kip1 Cyclin D1 p-prb S87/811 Total prb d % cells in sub-g1 4 3 2 1 +GSK-MEKi N-RAS Q16K c ±1 62±2 29± 1±1 38±7 5nM MEKi 9±1 7± 3±1 5 nm MEKi ± 48±4 29± 21±6 58±2 29±1 12±1 52±1 35±1 63±2 15±2 14±3 21±3 61±2 23± 18± 61±4 26± 13±1 3± 11±3 11±4 23±1 14±12 55±5 17±2 27±4 83±1 11±1 6± 57±6 2±7 2±2 75±4 16±2 9±1 Counts Annexin V-APC Figure 5. Acquired resistance to BRAF inhibition can dampen MAPK/ERK kinase (MEK) inhibitor sensitivity. (a) Viability curves of the parental and isogenic melanoma sublines treated with the indicated GSK-MEKi concentrations for 72 hours (relative to DMSO-treated controls; mean±sd; n ¼ 2). DR, drug resistant. (b) parental cells and BRAF inhibitor resistant sublines were treated with DMSO ( ) or5nm GSK-MEKi ( þ ) for 24 hours, and the effects on extracellular signal regulated kinase (ERK) activation and cell cycle regulators were determined by immunoblotting. (c) GSK-MEKi effects on cell cycle distribution and apoptosis. APC, allophycocyanin;, propidium iodide. (d) parental cell lines were lentivirally transduced with N-RAS Q61K or copgfp control, and after 1 days cells were treated with DMSO ( ) or5nm GSK-MEKi ( þ ) for 72 hours. Percentage of cells in sub-g1 fraction is shown. undergoing cell death, although this increase never reached the level of apoptosis seen in the parental cells (Figure 6b). For example, in response to the combined inhibition of BRAF and MEK, the N-RAS mutant and sublines displayed a sub-g1 fraction of B2% compared with over 6% sub-g1 cells in the parent melanoma line (Figure 6a and b). Moreover, ectopic expression of N-RAS Q61K in the cells diminished the efficacy of the combination of www.jidonline.org 1855

5nM MEKi + 1 nm BRAFi 5nM MEKi + 1 nm BRAFi ±1 62±2 81±2 29± 7± 1±1 4±1 62±3 48±4 29± 21±6 23±9 52±2 18±2 29±1 ± 58±2 84±1 29±1 1±1 12±1 6±1 52±1 35±1 14±3 2 2± 59±1 18±1 23±3 61±2 23± 18± 32±1 5±7 32±7 18±2 61±4 26± 13±1 4±1 74±1 17±1 9±1 % Of cells in sub-g1 7 6 5 4 3 2 1 Counts 1 nm 5nM GSK-MEKi 5 1 15 FL3-Area Annexin V-APC 1 nm 4 + 5 nm GSK-MEKi +GSK-B-RAFi 3 +GSK-MEKi 2 1 N-RAS Q61K % cells in sub-g1 +GSK-MEKi + + + + + + p-erk Y24 Total ERK p27 Kip1 Cyclin D1 p-prb S87/811 Total prb Figure 6. Combined BRAF and MAPK/ERK kinase (MEK) inhibition does not restore sensitivity to mitogen-activated protein kinase (MAPK) inhibition in BRAF inhibitor resistant cells. (a) The effects of (1 nm) and GSK-MEKi (5 nm) on cell cycle distribution and apoptosis of parental and derived sublines at 72 hours after treatment. APC, allophycocyanin;, propidium iodide. (b) Activity of combination and single MAPK inhibitors in promoting cell death in parental and drug resistant (DR) sublines at 72 hours after drug treatment. (c) parental cell lines were lentivirally transduced with N-RAS Q61K or copgfp control, and after 1 days cells were treated with DMSO ( ) or 1 nm and 5 nm GSK-MEKi ( þ ) for 72 hours. Percentage of cells in sub- G1 fraction is shown. (d) parental cells and BRAF inhibitor resistant sublines were treated with DMSO ( ) or 1 nm and 5 nm GSK-MEKi ( þ ) for 24 hours, and the effects on extracellular signal regulated kinase (ERK) activation and cell cycle regulators were determined by immunoblotting. MEK and BRAF inhibition (Figure 6c). Finally, the highly resistant subline remained refractory to the simultaneous inhibition of MEK and BRAF. In line with these data, the combination of BRAF and MEK inhibition did not significantly inhibit MAPK signaling in the DR clones, although the reduced levels of cyclin D1 and p-prb (both markers of cell cycle inhibition) were comparable in the DR clones and parent (Figure 6d). DISCUSSION The prevalent acquisition of resistance to BRAF inhibitors is a major problem in clinical oncology and reflects the capacity of melanomas to either circumvent their dependence on MAPK activity or, more commonly, restore MAPK signaling. For instance, increased expression of the IGF-1R or PDGFRb receptor tyrosine kinases activates the compensatory 3K/ AKT survival pathway (Nazarian et al., 21; Villanueva et al., 21). Similarly, alterations that facilitate ERK activation, such as increased abundance of CRAF or COT1, activating mutations in MEK or N-RAS, and amplification or truncation of BRAF, have been identified in vivo and in vitro (Montagut et al., 28; Corcoran et al., 21; Johannessen et al., 21; Nazarian et al., 21; Poulikakos et al., 211; Wagle et al., 211). In this report, we show that MAPK 1856 Journal of Investigative Dermatology (212), Volume 132

reactivation is the predominant mechanism of acquired BRAF inhibitor resistance, with all resistant melanoma sublines displaying reactivation and often enhanced MAPK signaling in the presence of the BRAF inhibitor GSK2118436. The restoration of MAPK activity in the presence of BRAF inhibitor, and the observation that tumor regression requires near-complete cessation of ERK activity, suggests that the combination of inhibitors targeting multiple nodes in the MAPK pathway may provide a more durable and prolonged clinical response (Bollag et al., 21). Consistent with this hypothesis, combined MEK and BRAF inhibition overcame resistance of colorectal cancer cells to MEK or BRAF inhibitors used singly and was more effective in parental cells compared with either agent alone (Corcoran et al., 21). Moreover, drug trials testing concurrent therapy with BRAF and MEK inhibitors are underway in patients with mutant BRAF metastatic melanoma (Infante et al., 211). In this report, the potency of combined MEK and BRAF inhibition was assessed in a series of melanoma sublines with resistance to BRAF inhibition. These DR sublines were generated from a single parental line; although they displayed distinct mechanisms of resistance, all showed increased MAPK signal pathway output in the presence of. The majority of DR clones showed weaker sensitivity to MEK inhibition alone and to the combination of MEK and BRAF inhibitors when compared with the parental clone, indicating that acquiring resistance to BRAF inhibition commonly confers cross-resistance to MEK inhibitors. The activation of N-RAS conferred potent resistance to, but significantly diminished the sensitivity of melanoma clones to GSK-MEKi alone, or the combination of MEK and BRAF inhibitors. These data strongly suggest that activation of N-RAS maintains MAPK signaling in the presence of BRAF inhibitor, and that additional MEKindependent N-RAS effectors contribute to MEK and BRAF inhibitor resistance. It is known that mutant N-RAS uses CRAF to signal to MEK and ERK, thus bypassing inhibited BRAF (Dumazet al., 26; Jaiswal et al., 29), and that N-RAS can signal via multiple pathways including the 3K/AKT cascade (Smalley, 21). Nevertheless, there are discrepancies in the literature regarding the role of activated RAS in selectively sensitizing cancer cells to MEK inhibition. In one study, BRAF inhibitor resistant melanoma cell lines with an acquired N- RAS Q61K mutation were sensitive to MEK inhibition in the presence of the RAF inhibitor, Vemurafenib/PLX432 (Nazarian et al., 21). Conversely, N-RAS mutation status did not predict MEK inhibitor sensitivity in melanoma cell lines (Solit et al., 26), and MEK inhibitors show only modest clinical activity in patients with RAS-mutant tumors (Nissan and Solit, 211). It seems likely that the impact of mutant RAS on MEK inhibitor responses reflects its expression and activity and ultimately the network of activated N-RAS-dependent effectors. This is in agreement with a recent report demonstrating that K-RAS 13D -mutant HCT116 colorectal cancer cells became resistant to MEK inhibition upon amplification of the driving K-RAS 13D oncogene (Little et al., 211). The and sublines showed constitutive reactivation of MAPK signaling, elevated levels of PDGFRb, and activation of AKT. There was no evidence of activationassociated PDGFRb phosphorylation, and thus this receptor tyrosine kinase did not contribute to AKT activity or BRAF inhibitor resistance in these clones. Signaling via AKT was also not mediating resistance, as the combined inhibition of MAPK (BRAF þ MEK), 3K, and mtor did not trigger apoptosis in the and clones. In contrast, melanoma cells with PDGFRb activation did not show significant reactivation of MAPK signaling (Nazarian et al., 21), and the concurrent inhibition of the MAPK and 3K/AKT/mTOR pathways overcame resistance to the BRAF inhibitor, Vemurafenib (Shi et al., 211). The dual targeting of BRAF and the 3K network appears effective when BRAF inhibition leads to increased AKT phosphorylation, a compensatory prosurvival mechanism not observed in any of our DR clones (Atefi et al., 211; Sanchez- Hernandez et al., 212). The subline differed significantly in retaining some sensitivity to MEK inhibition, in the absence and presence of BRAF inhibitor. This presumably reflects the fact that accumulates elevated levels of the COT1 kinase, which activates ERK through MEK-dependent and -independent mechanisms (Johannessen et al., 21). Our study confirms that many perturbations restoring MAPK activation have the potential to cause resistance to MEK inhibition and to the concurrent inhibition of BRAF and MEK. These data are in agreement with a recent report showing a high degree of cross-resistance to BRAF and MEK inhibitors in BRAF inhibitor resistant melanoma cell lines generated in vitro (Ribas et al., 21). Furthermore, experimental and computational modeling of the MAPK pathway showed that switching activated ERK from BRAF to CRAF is accompanied by increased resistance to MEK inhibition (Sturm et al., 21). The clinical significance of these findings will require validation, and it remains unclear whether initial combination therapy with MEK and RAF inhibitors yields improved efficacy over single agents in terms of the frequency and duration of response. What is becoming increasingly evident, however, is that individualized combinations of highly targeted agents affecting multiple signaling pathways will be required to overcome the complex and heterogeneous resistance networks in melanoma metastases. MATERIALS AND METHODS Cell culture and compounds and MelMS melanoma cells were obtained from Professor P Hersey (Newcastle, NSW, Australia). Cells were grown in DMEM with 1% fetal bovine serum and glutamine (Gibco BRL, Carlsbad, CA). Cells were cultured in a 37 1C incubator with 5% CO 2. Stocks of and GSK-MEKi (GlaxoSmithKline, Collegeville, PA), BEZ235, and imatinib mesylate (Selleck Chemicals, Houston, TX) were prepared in DMSO. Cell authentication was confirmed using the AmpFlSTR Identifiler PCR Amplification Kit from Applied Biosystems (Mulgrave, Victoria, Australia), and using the Gene- Marker V1.91 software (SoftGenetics, State College, PA). www.jidonline.org 1857

Pharmacological growth inhibition assays Cultured cells were seeded into 96-well plates (1, cells per well) and 24 hours after seeding serial dilutions of the relevant compound prepared in media were added. Cells were incubated for 72 hours, and cell viability was measured using the Cell proliferation Aqueous MTS assay (Promega, Wisconsin, MD) on a VICTOR 2 Multilabel counter (Perkin Elmer, Waltham, MA). Viability was calculated as a percentage of control (DMSO-treated cells) after background subtraction. A minimum of two independent viability assays each performed in triplicate was carried out for each cell line and drug combination. Cell cycle and apoptosis analysis Adherent and floating cells were combined and cell cycle and apoptosis analyses were performed as previously described (Gallagher et al., 25). DNA extraction and genotyping Genomic DNA was extracted from 1 2 1 6 cells using the Wizard SV Genomic DNA purification system (Promega). cells were genotyped as part of the Sequenom OncoCarta Assay Panel v1. (San Diego, CA). RNA extraction and microarray gene expression analysis Total RNA was extracted as described previously (Scurr et al., 21). Gene expression analysis was performed using the Sentrix Human- Ref-6 v.4. Expression BeadChip (Illumina, San Diego, CA) and BeadStation system from Illumina according to the manufacturer s instructions. Protein detection Total cellular proteins were extracted at 4 1C using RIPA lysis buffer containing protease and phosphatase inhibitors (Roche, Basel, Switzerland). Proteins (4 mg) were resolved on 12% SDS-PAGE and transferred to Immobilon-P membranes (Millipore, Bedford, MA). Phospho-RTK arrays were performed according to the manufacturer s instructions (Human Phospho-RTK Array Kit; R&D Systems, Minneapolis, MN). Western blots were probed with antibodies against p-erk (E4; Santa Cruz Biotechnology, Santa Cruz, CA), total ERK (137F5; Cell Signaling, Danvers, MA), BRAF (L12G7 and 55C6, Cell Signaling), prb S87/811 (Cell Signaling), totalprb (G3-245, Becton Dickinson, Franklin Lakes, NJ), b-actin (AC-74; Sigma-Aldrich), p27 Kip1 (Becton Dickinson), p16 INK4a (JC8, Santa Cruz Biotechnology), cyclin D1 (G124-326, Becton Dickinson), CRAF (E1, Santa Cruz Biotechnology), N-RAS (F155, Santa Cruz Biotechnology), pakt S473 (736E11, Santa Cruz Biotechnology), total AKT (11E7, Cell Signaling), PTEN (A2B1, Santa Cruz Biotechnology), COT/Tpl2 (Cell Signaling), IGF-1R (Cell Signaling), ppdgfrb Y751 (88H8, Cell Signaling), and PDGFRb (C82A3, Cell Signaling). Lentiviral transductions Lentiviruses were produced in HEK293T cells as described previously (Haferkamp et al., 29). Cells were infected using a multiplicity of infection of 5 to provide an efficiency of infection above 9%. For N-RAS silencing experiments, cells were transduced with control-psih-copgfp or N-RAS-pSIH-copGFP for 5 days. Inhibitors or DMSO were added and cells were harvested 72 hours after inhibitor treatment for cell cycle and western blot analyses. The shrna no.1 and 2 sequences targeting N-RAS correspond to nucleotides 567 586 and 741 759 (NM_2524.3), respectively. The COT1 shrna sequence corresponds to nucleotides 1826 1846 (NM_524.2), and the non-silencing negative control shrna did not show complete homology to any known human transcript and had the following sequence: 5 -TTAGAGGCGAGCAAGACTA-3. CONFLICT OF INTEREST The authors state no conflict of interest. ACKNOWLEDGMENTS We are grateful to Tona Gilmer and Sylvie Laquerre (GlaxoSmithKline) for providing and GSK-MEKi. We thank Suzanah Philipsz, Carina Fung, and Jason Todd for excellent technical assistance. This work is supported by program grant 6334 and project grants of the National Health and Medical Research Council of Australia (NHMRC) and an infrastructure grant to Westmead Millennium Institute by the Health Department of NSW through Sydney West Area Health Service. Westmead Institute for Cancer Research is the recipient of capital grant funding from the Australian Cancer Research Foundation. HR is a recipient of a Cancer Institute New South Wales Research Fellowship and an NHMRC Senior Research Fellowship. The support of the Melanoma Foundation of the University of Sydney and colleagues from the Melanoma Institute Australia (incorporating the Sydney Melanoma Unit) and the Department of Anatomical Pathology at the Royal Prince Alfred Hospital are also gratefully acknowledged. SUPPLEMENTARY MATERIAL Supplementary material is linked to the online version of the paper at http:// www.nature.com/jid REFERENCES Atefi M, von Euw E, Attar N et al. (211) Reversing melanoma cross-resistance to BRAF and MEK inhibitors by co-targeting the AKT/mTOR pathway. PLoS One 6:e28973 Bollag G, Hirth P, Tsai J et al. (21) Clinical efficacy of a RAF inhibitor needs broad target blockade in BRAF-mutant melanoma. Nature 467:596 9 Chapman PB, Hauschild A, Robert C et al. (211) Improved survival with vemurafenib in melanoma with BRAF V6E mutation. N Engl J Med 364:257 16 Corcoran RB, Dias-Santagata D, Bergethon K et al. (21) BRAF gene amplification can promote acquired resistance to MEK inhibitors in cancer cells harboring the BRAF V6E mutation. Sci Signal 3:ra84 Davies H, Bignell GR, Cox C et al. (22) Mutations of the BRAF gene in human cancer. Nature 417:949 54 Dry JR, Pavey S, Pratilas CA et al. (21) Transcriptional pathway signatures predict MEK addiction and response to selumetinib (AZD6244). Cancer Res 7:2264 73 Dumaz N, Hayward R, Martin J et al. (26) In melanoma, RAS mutations are accompanied by switching signaling from BRAF to CRAF and disrupted cyclic AMP signaling. Cancer Res 66:9483 91 Emery CM, Vijayendran KG, Zipser MC et al. (29) MEK1 mutations confer resistance to MEK and B-RAF inhibition. Proc Natl Acad Sci USA 16:2411 6 Flaherty KT, Puzanov I, Kim KB et al. (21) Inhibition of mutated, activated BRAF in metastatic melanoma. N Engl J Med 363:89 19 Gallagher S, Kefford RF, Rizos H (25) Enforced expression of p14arf induces p53-dependent cell cycle arrest but not apoptosis. Cell Cycle 4:465 72 Gilmartin AG, Bleam MR, Groy A et al. (211) GSK112212 (JTP-7457) is an inhibitor of MEK activity and activation with favorable pharmacokinetic properties for sustained in vivo pathway inhibition. Clin Cancer Res 17:989 1 Haferkamp S, Scurr LL, Becker TM et al. (29) Oncogene-induced senescence does not require the p16(ink4a) or p14arf melanoma tumor suppressors. J Invest Dermatol 128:1983 91 1858 Journal of Investigative Dermatology (212), Volume 132

Infante JR, Falchook GS, Lawrence DP et al. (211) Phase I/II study to assess safety, pharmacokinetics, and efficacy of the oral MEK 1/2 inhibitor GSK112212 (GSK212) dosed in combination with the oral BRAF inhibitor GSK2118436 (GSK436). Proceedings of the Conference Phase I/II study to assess safety, pharmacokinetics, and efficacy of the oral MEK 1/2 inhibitor GSK112212 (GSK212) dosed in combination with the oral BRAF inhibitor GSK2118436 (GSK436); 2 6 April. Chicago, IL Jaiswal BS, Janakiraman V, Kljavin NM et al. (29) Combined targeting of BRAF and CRAF or BRAF and 3K effector pathways is required for efficacy in NRAS mutant tumors. PLoS ONE 4:e5717 Jiang CC, Lai F, Thorne RF et al. (21) MEK-independent survival of B-RAFV6E melanoma cells selected for resistance to apoptosis induced by the RAF inhibitor PLX472. Clin Cancer Res 17:721 3 Johannessen CM, Boehm JS, Kim SY et al. (21) COT drives resistance to RAF inhibition through MAP kinase pathway reactivation. Nature 468:968 72 Kefford R, Arkenau H-T, Brown MP et al. (21) Selective inhibition of oncogenic BRAF V6E/K/D by GSK2118436: evidence of clinical activity in subjects with metastatic melanoma. Pigment Cell & Melanoma Research. Proceedings of the Conference Selective inhibition of oncogenic BRAF V6E/K/D by GSK2118436: Evidence of clinical activity in subjects with metastatic melanoma; 4 7 November, Sydney, Australia Little AS, Balmanno K, Sale MJ et al. (211) Amplification of the driving oncogene, KRAS or BRAF, underpins acquired resistance to MEK1/2 inhibitors in colorectal cancer cells. Sci Signal 4:ra17 Long GV, Menzies AM, Nagrial AM et al. (211) Prognostic and clinicopathologic associations of oncogenic BRAF in metastatic melanoma. J Clin Oncol 22:22 Montagut C, Sharma SV, Shioda T et al. (28) Elevated CRAF as a potential mechanism of acquired resistance to BRAF inhibition in melanoma. Cancer Res 68:4853 61 Nazarian R, Shi H, Wang Q et al. (21) Melanomas acquire resistance to B- RAF(V6E) inhibition by RTK or N-RAS upregulation. Nature 468:973 7 Nissan MH, Solit DB (211) The SWOT of BRAF inhibition in melanoma: RAF inhibitors, MEK inhibitors or both? Curr Oncol Rep 13:479 87 Paraiso KH, Xiang Y, Rebecca VW et al. (211) PTEN loss confers BRAF inhibitor resistance to melanoma cells through the suppression of BIM expression. Cancer Res 71:275 6 Poulikakos, Persaud Y, Janakiraman M et al. (211) RAF inhibitor resistance is mediated by dimerization of aberrantly spliced BRAF(V6E). Nature 48:387 9 Ribas A, von Euw E, Attar N et al. (21) Reversal of cross-resistance to BRAF and MEK inhibitors by targeting the AKT/mTOR pathway. Pigment Cell & Melanoma Research. Proceedings of the Conference Reversal of crossresistance to BRAF and MEK inhibitors by targeting the AKT/mTOR pathway 21; 4 7 November, Sydney, Australia Sanchez-Hernandez I, Baquero P, Calleros L et al. (212) Dual inhibition of (V6E)BRAF and the 3K/AKT/mTOR pathway cooperates to induce apoptosis in melanoma cells through a MEK-independent mechanism. Cancer Lett 314:244 55 Scurr LL, Pupo GM, Becker TM et al. (21) IGFBP7 is not required for B-RAFinduced melanocyte senescence. Cell 141:717 27 Shi H, Kong X, Ribas A et al. (211) Combinatorial treatments that overcome PDGFRbeta-driven resistance of melanoma cells to V6EB-RAF inhibition. Cancer Res 71:567 74 Smalley KS (21) Understanding melanoma signaling networks as the basis for molecular targeted therapy. J Invest Dermatol 13:28 37 Solit DB, Garraway LA, Pratilas CA et al. (26) BRAF mutation predicts sensitivity to MEK inhibition. Nature 439:358 62 Sturm OE, Orton R, Grindlay J et al. (21) The mammalian MAPK/ERK pathway exhibits properties of a negative feedback amplifier. Sci Signal 3:ra9 Villanueva J, Vultur A, Lee JT et al. (21) Acquired resistance to BRAF inhibitors mediated by a RAF kinase switch in melanoma can be overcome by cotargeting MEK and IGF-1R/3K. Cancer Cell 18:683 95 Wagle N, Emery C, Berger MF et al. (211) Dissecting therapeutic resistance to RAF inhibition in melanoma by tumor genomic profiling. J Clin Oncol 7:7 Wan PT, Garnett MJ, Roe SM et al. (24) Mechanism of activation of the RAF-ERK signaling pathway by oncogenic mutations of B-RAF. Cell 116:855 67 Wu E, Palmer N, Tian Z et al. (28) Comprehensive dissection of PDGF- PDGFR signaling pathways in PDGFR genetically defined cells. PLoS One 3:e3794 www.jidonline.org 1859