Individualized strategies to target specific mechanisms of disease in malignant melanoma patients displaying unique mutational signatures

INTRODUCTION

Melanoma is a form of cancer whose incidence is rising each year in the developed world, and is second to leukemia in terms of loss of years of potential life from cancers [1]. Despite recent improvements in mortality rates, current deaths from melanoma are estimated to comprise 85% of all cancers affecting the skin. This is corroborated by the poor survival associated with melanoma when diagnosed at an advanced stage [2]. Therefore, the development of effective therapies is a major challenge in this field.

Molecular diagnostics of cancer have proved that targeted therapies can be effective in many cancer settings, as measured by the recent improvement in cancer survival statistics (World Cancer Statistics 2008; ICD-10 C18–21). The use of EGFR inhibitors in lung cancer [3, 4] and Imatinib in chronic myeloid leukemia (CML) patients [5] are two relevant examples. Targeted therapies in melanoma are mostly directed towards inhibiting MAPK-ERK1/2 signaling (MAPK hereafter), [6]. Mutational analyses have recently enabled the detection of up to 50% of malignant melanomas carrying an activating mutation in BRAF [7], and these can now be treated with specific B-RAF inhibitors [8]. In the clinic, this targeted approach, even when used in combination with MEK inhibitors, is of limited benefit to patient survival and, after a period, the cancer reappears aggressively [9–11].

From a molecular perspective, data from Next Generation Sequencing (NGS) show that more mutated genes than initially expected participate in tumorigenesis, including that of melanoma [12–14]. This involves a dynamic process of subclonal competition that eventually dictates multifactorial clinical resistance to B-RAF inhibitors, which is dependent on reactivation of MAPK signaling or other proliferative and/or pro-survival pathways [15–17].

Taking advantage of available melanoma NGS data, we characterized biopsies from advanced melanoma patients and cell lines by studying the presence of somatic mutations in a selected group of genes. We thereby detected unique signatures of mutated genes that are potentially associated with specific inhibitors, and explored the effects of case-specific combinations of the latter ex vivo and in vivo. Guided by individual mutational profiles, tailored combinations of inhibitors simultaneously targeting MAPK-dependent and MAPK-independent signaling were very efficient at inhibiting aberrant melanoma growth assessed in multiple cell lines, and xenografted tumors and biopsies grown in mice. Thus, specific mutational signatures could guide the design of personalized therapies based on the use of specific combinations of drugs that target case-specific pathogenic signaling mechanisms.

RESULTS

A targeted approach to characterizing the mutational status of lesions of advanced melanoma patients

To better understand the molecular character of specific melanoma lesions, we set up a targeted mutational study followed by functional analyses (described in Supplementary Figure 1). The genomic design of this study focused on the coding regions of a specific group of 217 genes that had previously been shown to be mutated in melanoma and selected mainly on the basis of their relevance in melanoma and their association with inhibitors of potential clinical use (see Materials and Methods for further explanation). To test this approach our selection of genes was compared in silico with the whole genome/exome sequencing (WGS/WES respectively) data already available for 11 advanced melanoma cell lines and 158 human melanomas (see Materials and Methods, [13, 14, 18, 19]). This comparison revealed an average of 3.74 mutated genes that can participate in multiple targetable signaling pathways, including PLC, MAPK, RTKs (receptors with tyrosine kinase activity), PI3K-mTOR and JAK-STAT (Figure 1 and Supplementary Table I). These results prompted us to study advanced melanoma cases (Breslow index ≥4 mm or metastasis) in 18 clinically characterized patients (clinical characteristics summarized in Supplementary Table II) using a targeted primary ultrasequencing approach, followed by secondary validation analysis (see Materials and Methods for further details). By these methods, an average of 3.4 mutated genes were identified in 11 of the 18 patients, enabled the detection of lesion-specific genes such as BRAF, RAC1, KRAS, HGF and MAPK7, amongst others. Interestingly, there was a wide range of mutation frequencies and combinations, which perhaps reflects the rich and heterogeneous microclonal composition expected in melanoma tumors (400X average depth/mutation; Table I) [20]. Furthermore, actionable mutations such as BRAF^V600E that can guide targeted therapy (using B-RAF inhibitors) were detected in the same melanoma alongside other mutated genes that may also guide therapy (Table I). It is significant that mutations in four patients could not be validated due to limitations of the tissue sample (see Materials and Methods), and that no mutations were identified in three other patients. Thus, this targeted approach could be adopted to identify genomic alterations affecting one or several genes. These may be explored as potential targets for therapy in specific cases of melanoma.

Effects of specific targeted therapy guided by mutational signature

To explore how to use mutational data to design targeted therapies based on specific mutational characterizations, the functional effects of specific combination therapies were studied in advanced melanoma cell lines with known mutational profiles (Supplementary Table III). Taking A375 advanced melanoma cells as an example, we detected and validated mutations in BRAF, FGFR2 and mTOR that could reasonably be expected to associate with Vemurafenib (BRAFi (V), hereafter), Vargatef (FGFR2i (Va)) and Everolimus (mTORi (E)). Exponentially growing A375 cells were incubated with increasing concentrations of each inhibitor. This caused a concentration-dependent reduction in cell proliferation from which the IC₅₀ of each inhibitor was calculated (Figure 2A and Supplementary Table III). These concentrations were used for subsequent experiments. Next, the mechanistic effects of treatment with each inhibitor (using IC₅₀ values in each case) were analyzed in A375 cells that had been serum-starved to provoke the inhibition of the intended mutation-associated downstream signaling. These were assessed by western blot using P-ERK1/2, P-p38 and P-S6 antibodies (Figure 2B).

To discover more about the biological effects of multiple combinations of these inhibitors on proliferation, A375 cells were incubated with IC₅₀ concentrations of BRAFi, FGFR2i and mTORi in single, double or triple combinations (blue, green and red lines, respectively, in Figure 2C). The combinatorial treatments were more effective at reducing melanoma cell growth than the monotherapies. The triple combination was the most efficient, and had no non-specific cytotoxic effects (Figure 2C and 2E). These results were confirmed using DNA synthesis as an alternative read-out (Figure 2D and 2E). Thus, under these conditions, a combination of inhibitors guided by a specific mutational signature, simultaneously targeted multiple signaling mechanisms controlling the growth of A375 cells. Analyzing the mechanistic effects of these drug combinations on their associated signaling pathways in this system, showed that treatment with BRAFi inhibited MAPK signaling. However, treatment of A375 cells with the inhibitors mTORi and FGFR2i, alone or in combination (E+Va), had no such effect (Figure 2F and 2G), despite being very effective at inhibiting cell proliferation and DNA synthesis (Figures 2C, and 2E). Thus, using genetically defined inhibitors in this system we can specifically target a combination of MAPK-dependent (V) and MAPK-independent (E+Va) signaling mechanisms that control the malignant growth of A375 melanoma cells. This observation was not confined to these cells and more examples of specific mutational signatures guiding effective combinatorial therapies comprising MAPK-dependent and MAPK-independent mechanisms in other human advanced melanoma cell models are shown in Supplementary Figures 2, 3 and 5.

Increased effects of targeted therapy against an appropriate mutational background

As part of a heterogeneous network of aberrant intracellular signaling, multiple deregulated pathways can participate in the mechanistic control of melanoma growth (Figure 1). We examined whether a combination therapy designed for a specific mutational signature could be more effective when used against a genetically appropriate background. A group of melanoma cell lines harboring unique mutational signatures (Supplementary Table III) was treated in parallel with the genetically defined inhibitors for A375 cells, BRAFi, FGFR2i and mTORi, alone or in combination. In general, each treatment was more effective in A375 cells than in the other melanoma cell lines, with the possible exception of mTORi used alone (Figure 3A). The triple drug combination treatment (V+E+Va) simultaneously targeting MAPK-dependent and MAPK-independent proliferation mechanisms produced greater inhibition of A375 cell growth than the others (Figure 3A and 3B). Likewise, other specific treatments based on the combination of genetically defined inhibitors in other melanoma cell lines showed a stronger effect than that of A375 cells (Supplementary Figure 4A and 4B). More examples of specific mutational signatures guiding more effective combination therapies when used against an appropriate mutational background are shown in Figure 5E and Supplementary Figure 5C. In summary, based on specific mutational signatures, a specific treatment consisting of a combination of genetically defined inhibitors may have stronger anti-melanoma activity when used against an appropriate genomic background.

In vivo effects of a targeted therapy that combines MAPK-ERK-dependent and MAPK-ERK-independent mechanisms of inhibition

To study the in vivo effects of targeted therapy oriented by a specific mutational profile, xenografted tumors from A375 melanoma cells were generated in BALB/C mice (nu^−/−/nu^−/−). Once grown to a volume of approximately 100 mm³, tumors were assigned to four comparable groups and treated daily with vehicle, a MAPK-dependent inhibitor (BRAFi), a MAPK-independent combination of inhibitors (FGFR2i+ mTORi), or a triple combination of the latter (BRAFi+FGFR2i+mTORi). As shown in Figure 4A and 4B, both treatments used independently reduced tumor growth to a similar extent. However, the triple combination (V+Va+E) proved most effective at reducing melanoma growth in this system. Once the experiment was finished, the remaining growth potential of these tumors was characterized by studying Ki67 and the mitotic index in tumor sections. As might be expected, a marked decrease in both proliferation markers in those tumors treated with the combination of MAPK-dependent and MAPK-independent inhibitors (Figure 4C-4G) was observed, thereby confirming in vivo our previous findings in cultured cells (Figures 2D, 2E, and 5).

A pre-clinical example of targeted therapy guided by a specific mutational signature in patient 17

These findings were challenged by integrating the study of a freshly resected biopsy from patient 17 (Table I and Supplementary Table II) in the working pipeline (illustrated in Figure 5A and Supplementary Figure 1). First, a fragment of the biopsy was characterized which enabled the detection of somatic mutations in the BRAF and MAPK7 (ERK5 hereafter) genes (Table I and Figure 5A), which were associated with specific inhibitors like Vemurafenib (BRAFi (V)) and XMD-8–92 (ERK5i (X)).

Second, freshly isolated melanocytes (MELANOMA17 cells hereafter) were also inspected for the presence of mutations of BRAF and ERK5 (Figure 5B) and for the expression of well-known melanoma markers such as S100A and MCSP (Supplementary Figure 6A and 6B). Once characterized, MELANOMA17 cells were incubated with increasing concentrations of BRAFi and ERK5i, and the IC₅₀ of each was calculated (Supplementary Table III). Treatment with BRAFi and ERK5i inhibited B-RAF and ERK5-dependent signaling, assessed by western blot, in starved MELANOMA17 cells (Figure 5C). Simultaneous treatment with both inhibitors was more effective at reducing MELANOMA17 cell proliferation than either inhibitor alone (Figure 5D). Furthermore, and consistent with our previous observations in multiple cell lines, the combination treatment (BRAFi+ERK5i) was also more effective in cells with an appropriate mutational signature (MELANOMA17 cells) when compared with a panel of other melanoma cell lines (Figure 5E), with the possible exception of A375 cells, which were highly sensitive to treatment with the BRAFi dose used (compare the IC₅₀ values for MELANOMA17 with those of A375 cells in Supplementary Table III). This combination therapy also consisted of MAPK-dependent (BRAFi) and MAPK-independent (ERK5i) mechanisms, which successfully suppressed the aberrant growth of MELANOMA17 cells (Figure 5F).

Third, another fragment of the biopsy was implanted in NSG mice and allowed to grow until four tumor-comparable groups of mice could be established. The groups of mice were then treated with vehicle, BRAFi, ERK5i, or a combination of the two (V+X). When used separately, the two inhibitors were able to suppress xenografted MELANOMA17 tumor growth to a similar extent, but growth inhibition was more effective when used in combination (Figure 5G-5I).

Thus, it is possible to study the effects of personalized therapies, guided by targeted mutational profiling of advanced melanoma patients, in pre-clinical ex vivo and in vivo models using freshly resected material from each lesion.

DISCUSSION

Metastatic melanoma provides an instructive example of the development of rationalized therapies guided by molecular diagnostics. Mechanistically, targeted therapy mainly involves the inhibition of the MAPK signaling pathway by using BRAF or MEK inhibitors, alone or in combination [21], as suggested by: A) activating mutations in BRAF and NRAS oncogenes in 48% and 15% of all diagnosed melanomas, respectively [22]; and B) multiple MAPK reactivation mechanisms that confer resistance to BRAF inhibitors [23–25]. This has improved the clinical management of those patients with mutated BRAF, whereby targeted inhibition of aberrant MAPK signaling can increase their OS by up to 11.4 months, although, from a different perspective, it still offers a limited benefit to these patients [10, 26, 27]. To explain this, we can hypothesize that, as part of an intricate network of transforming mechanisms in melanoma, this disease simultaneously uses multiple oncogenic mechanisms, such as, for example, PI3K, MET and GNAQ [28, 29], that, along with MAPK signaling, act as mechanistic drivers of the disease and promote progression or resistance to therapy. In this regard, data from genetically defined melanoma models now show that a rational combination of MAPK and PI3K inhibitors can improve the effects of therapy when used against specific genomic backgrounds [30]. Moreover, patients with specific mutations gained greater benefits when treated with immunotherapy [31, 32]. Thus, better characterization of advanced melanoma lesions could improve our ability to treat this disease through the use of specific therapies that simultaneously target multiple signaling mechanisms.

From a molecular perspective, melanoma is a very heterogeneous disease in which up to 1, 500 somatic mutations may be harbored in the coding exons of a single lesion [13]. This work studies mutations in 217 genes previously shown to be mutated in melanoma [13, 14, 18, 19, 33, 34] in silico by comparing them with mutations in 11 cell lines and 158 human melanomas, and ex vivo by characterizing 18 lesions from advanced melanoma patients. Under all conditions, genes like BRAF, RAC1, FGFR2 and IL6R were mutated at varying frequencies, occurring as part of unique mutational signatures comprising specific combinations of mutated genes that have the potential to participate in multiple signaling pathways and to be associated with specific inhibitors. Functionally, the effects of combination therapies guided by specific mutational signatures were analyzed in multiple melanoma cells harboring unique mutational signatures. Those treatments that simultaneously targeted MAPK-dependent and MAPK-independent signaling were most effective at reducing melanoma growth both ex vivo and in vivo. These observations can be aligned with work from other laboratories, showing that to promote transformation in melanocytes, aberrant MAPK-signaling elicited by BRAF or NRAS oncoproteins requires the active collaboration of other oncogenes, such as PI3K, RAL-GDS, GNAS or C-MYC, that can participate in alternative signaling pathways [35–38]. Thus, a combination of genetically defined inhibitors targeting multiple signaling pathways could be more effective against specific cases of malignant melanoma. We might expect targeting well-known melanoma mechanisms to affect the growth of melanoma cells in general, and this can indeed be observed in our data. Nevertheless, combination therapies guided by specific mutational signatures were most effective when used against an appropriate mutational background. Furthermore, different BRAF-mutated cell lines, each with an individual mutational signature, had different sensitivities to BRAF inhibition (Supplementary Table III). Finally, our data strongly suggest that combining several mutationally selected inhibitors can specifically block important mechanisms that participate in the control of aberrant melanoma cell growth and in the finely tuned cellular decision to activate DNA synthesis (Figures 2B, 2C, 2E, 4A, 4D and 4E). This rules out the possibility that the results were a consequence of nonspecific cytotoxic activities.

Starting with freshly resected material from a metastatic lesion (patient 17) and trying to match the timing with the clinic, a validated mutational profile was obtained within two weeks of resection. This data enabled the study of a combination therapy based on inhibitors with MAPK-dependent (BRAFi) and MAPK-independent (ERK5i) mechanisms of action in isolated MELANOMA17 cells and in xenografted tumors grown in mice. These gave the best results when combinatorial approaches were used. Of course, this study provides just one example of what targeted characterization of specific lesions might offer by way of diagnostic possibilities for human melanoma in the near future. Considering its potential applicability in routine clinical practice, this approach would require several limitations to be overcome. This would entail: 1) establishing efficient protocols to collect, manipulate and characterize specific lesions that are representative of the various steps of the disease; 2) managing the toxicity due to drug combinations; and 3) dealing with tumor heterogeneity and interactions with the immune system that may be responsible for the eventual resistance acquired after combination treatments. However, there is much scope for studying novel strategies for targeted therapy following a molecular rationale, particularly in a disease like advanced melanoma that offers a limited prospect of survival to patients who are suffering from it (Supplementary Figure 7).

In summary, by adopting targeted approaches we can envisage working with specific signatures of mutated genes that can: 1) help characterize individual lesions in advanced melanoma patients; 2) guide the use of specific inhibitors rationally combined in individualized therapies to target case-specific mechanisms of melanocytic transformation. In this work, a rational combination of genetically defined inhibitors simultaneously targeting MAPK-dependent and MAPK-independent signaling mechanisms showed improved biological outcomes with respect to the malignant growth of specific advanced melanomas.

MATERIALS AND METHODS

Cells and reagents for tissue culture

Eight human advanced melanoma cell lines were used. A375 (CRL-1619™ ), SK-MEL-28 (HTB-72™), SK-MEL-2 (HTB-68™), MALME-3M (HTB-64™), MEWO (HTB-65™) and HT-144 (HTB-63™) cells were obtained from the American Type Cell Collection (ATCC, Rockville, MD). MEL-JUSO (ACC 74) was obtained from the German Collection of Microorganisms and Cell Cultures (DSMZ, Braunschweig, Germany). Genomic data from these cells, including those of the somatic mutations detected in this study, are publicly available at the Broad-Novartis Cancer Cell Line Encyclopedia website (CCLE: http://www.broadinstitute.org). MELANOMA17 cells were established from a primary biopsy sample, as explained in the Supplementary Methods. Commercial cell lines were cultured as recommended by ATCC or DSMZ and incubated with inhibitors, as described in the Supplementary Methods.

Cell proliferation and DNA synthesis assays

Cells growing exponentially to approximately 50% confluence in T96 well plates were incubated with the specific inhibitors while keeping the total amount of DMSO (0.5%) constant. Cellular proliferation was evaluated using AlamarBlue reagent (Life Technologies) and colorimetric changes were quantified using the Synergy™ HTX Multi-Mode Microplate Reader (Biotek). To assess the effects on DNA synthesis, cells were grown in a Millicell EZ SLIDE 8-well glass (Merck Millipore, PEZGS0816) and after treatment with specific inhibitors were incubated for a further 2 h with Click-iT^® EdU (Alexa Fluor^® 594 Imaging Kit; Life Technologies, C10339). Immediately afterwards, cells were prepared for microscopy following the manufacturer’s specifications (see Supplementary Methods for further explanation). Cell images were captured with a Nikon A1R confocal microscope with Plan Apo 10x/0.45NA and Plan ApoVC 60x/1.40NA objectives.

Genomic DNA samples

Matched tumoral and non-tumoral material was obtained from 18 patients diagnosed with advanced melanoma and who were being monitored by the Oncology Department of the Hospital Universitario Marqués de Valdecilla (HUMV; see clinical characteristics in Table I). Tumoral DNA samples were obtained from freshly frozen tissue samples taken at the time of diagnosis, and matched non-tumoral DNA was extracted from saliva or peripheral blood neutrophils. We designed an intra-subject observational study of patients diagnosed with advanced melanoma and with a Breslow index of ≥4 mm, and with a loco-regional infiltration in lymph nodes or presenting distant metastasis. The study, the patient information sheet, and the informed consent form were approved by the Ethics Committee of the HUMV.

Enrichment library design, preparation, sequencing and variant calling

Genomic DNA samples were processed using the Qubit^® dsDNA BR Assay Kit (Life Technologies) and quantified using Qubit 2.0 apparatus (Life Technologies). The DNA enrichment library was prepared using a specifically designed HaloPlex Target Enrichment System Kit for this melanoma study (Design ID: 00912-1339502780, Agilent Technologies) following the manufacturer’s instructions. The design focused on the coding regions of a group 217 genes known to be mutated in melanoma, and which were selected because they were: A) genes of known relevance in melanoma, including BRAF, NRAS [7], and EGFR [18, 19]; B) genes that may be associated with pharmacological inhibitors of potential clinical use, such as FGFR2, KIT and ERBB4 [18, 20, 21]; and C) genes that may be involved in chromatin architecture (ARID1A and DNMT3A [14]), intracellular signaling (MEK1 [22]), or transcription (NFATC2 [22]). Briefly, 400 ng of genomic DNA was digested with the specific cocktail of restriction enzymes provided in the kit. Digested DNA was then hybridized to a probe for target enrichment, indexed and captured. Each DNA was then amplified by PCR at Tm = 60°C, for 18 cycles, using a Herculase II Fusion Enzyme kit (Agilent Technologies). Next, amplified target libraries were purified using an Agencourt AMPure XP Kit (Beckman Coulter Genomics), following the manufacturer’s guidelines, and quantified with Qubit 2.0 apparatus (Life Technologies), using the Qubit^® dsDNA HS Assay Kit (Life Technologies). They were also analyzed in parallel by capillary electrophoresis in a 2100 Bioanalyzer (Agilent Technologies), using High Sensitivity DNA reagents and chip Kits (Agilent Technologies). Libraries were sequenced at the Instituto de Medicina Genómica (IMEGEN, Valencia University, Spain) with a MiSeq Personal Sequencer (Illumina). The process of somatic mutation identification described in the Supplementary Methods.

Somatic mutation identification

Sequencing data were aligned against the human reference genome (hg19) using the BWA aligner [39]. The alignment was refined using SAMTOOLS fixmate (PMID: 19505943) and PICARD TOOLS cleanSam tools (http://broadinstitute.github.io/picard.). Local realignment of insertions and deletions (indels) was then performed using the GATK suite [40] before final sorting and indexing. The RAMSES application (PMID: 24296945), written in-house, was used to detect nucleotide substitutions. Small indels were identified using Pindel [41] in paired tumor-normal mode. For greater specificity, only simple insertion and deletion events of fewer than 10 bp were selected. An in-house perl script filter was used to extract high-quality indels: considering the high sequence coverage obtained in these samples, only those indels with a minimum coverage of 20 reads in both tumor and normal samples, and with a minimum frequency of 10% of the reads and a minimum of five independent reads supporting the event in the tumor sample, and with no evidence in the normal sample, were considered. All potential somatic mutations were filtered using the dbSNP132 and 1000 Genomes Project mutation databases and the functional consequence at the protein level was annotated according to the Ensembl database using an in-house perl script based on the Ensembl database API.

Validation analysis

Genomic DNA was amplified using the specific oligonucleotides described in Supplementary Table IV. All amplicons from the same patient were mixed in a tube and each sample was quantified by Qubit 2.0 (Life Technologies), using the Qubit^® dsDNA BR Assay Kit (Life Technologies). MELANOMA17 cells were monitored by Sanger sequencing for the presence of mutations in BRAF and MAPK7 (see the supplementary methods for further details).

Statistics

Unless otherwise specified, all experiments were done in independent triplicates and all numerical data were summarized as the average of the values ± the standard error of the mean (SEM) using GraphPad PRISM. Levels of statistical significance are indicated as follows: *p < 0.05; **p < 0.01; ***p < 0.001.

Western blot

Cells growing exponentially at approximately 70% confluence were treated under the desired conditions. Cells were starved overnight (unless otherwise stated), treated with the appropriate inhibitor and lysed as described in [42]. Whole cell lysates were subjected to acrylamide SDS-PAGE, using standard procedures, then transferred onto a nitrocellulose support membrane (Immobilon, Millipore) and western blotted. The primary and secondary antibodies and the data collection method are described in the Supplementary Methods.

Mice and reagents for in vivo studies

BALB/c Nude mice CAnN.Cg-Foxn1nu/Crl (Charles River) were injected with 6 × 10⁶ A375 melanoma cells in the subcutaneous dorsal area. Approximately one week later, the tumor reached a volume of about 100 mm³, at which point mice were assigned to four tumor size-comparable groups of 12 animals and treated as described in the Supplementary Methods.

Fresh tissue from patient 17, who had been diagnosed with metastatic melanoma (Table I and Figure 5), was minced and xeno-injected into NOD.Cg-Prkdcscid Il2rgtm1Wjl/SzJ mice (commonly known as NOD scid gamma (NSG) mice) (Charles River). Briefly, the animals were anesthetized using ketamine (75 mg/kg) and medetomidine (1.0 mg/kg) and a piece of tumor was inserted in the subcutaneous dorsal area through a small incision in the skin and allowed to grow. Next, mice were sacrificed (as described in supplementary methods) and tumors were collected and minced into pieces of about 2 mm³ and reimplanted into the experimental group of mice. When these mice had grown tumors of an approximate volume of 100 mm³, they were distributed among four groups of six mice, each with comparable tumor volumes and treatments were started, as described in Figure 5 and in the Supplementary Methods.

ACKNOWLEDGMENTS

We are indebted to the patients who have contributed to this study. We especially thank Dr. Fidel Madrazo, José Revert and Carolina Santa Cruz from IDIVAL, and the staff members of the Biobank and the Pathology Service at HUMV, for their exceptional work in sample collection and organization. We would also like to thank the Santander Super-Computation Service for their support. We would also like to acknowledge Dr. Piero Crespo and Dr. Javier León (IBBTEC, Santander, Spain), and Dr. J. Silvio Gutkind and Dr. Xiaodong Feng (NIH, Bethesda, MD, USA) for helpful discussions and comments about this work.

FUNDING

This work was supported by grants from the ISCIII, co-financed by the European Union (FEDER) (PI12/00357) and a Ramón and Cajal research program from MINECO (RYC-2013-14097) to JPV, and from the ISCIII (RTIC; RD06/0020/0107, RD012/0036/0060) and the Asociación Española Contra el Cáncer (AECC) to MAP. The work was also supported by a research award from LUCHAMOS POR LA VIDA to JPV and AGC. I.V. is supported by the Ramón and Cajal research program.

CONFLICTS OF INTEREST

All authors declare no conflict of interest except MAP.

MAP has the following conflicts of interest: Takeda-advisory board. Novartis, Amgen and Roche: Speaker bureau.

REFERENCES

1. Tran TT, Schulman J, Fisher DE. UV and pigmentation: molecular mechanisms and social controversies. Pigment Cell Melanoma Res. 2008; 21:509–16.

2. Erdei E, Torres SM. A new understanding in the epidemiology of melanoma. Expert Rev Anticancer Ther. 2010; 10:1811–23.

3. Mitsudomi T, Yatabe Y. Epidermal growth factor receptor in relation to tumor development: EGFR gene and cancer. FEBS J. 2010; 277:301–8.

4. Maemondo M, Inoue A, Kobayashi K, Sugawara S, Oizumi S, Isobe H, Gemma A, Harada M, Yoshizawa H, Kinoshita I, Fujita Y, Okinaga S, Hirano H, et al. Gefitinib or chemotherapy for non-small-cell lung cancer with mutated EGFR. N Engl J Med. 2010; 362:2380–8.

5. Druker BJ, Talpaz M, Resta DJ, Peng B, Buchdunger E, Ford JM, Lydon NB, Kantarjian H, Capdeville R, Ohno-Jones S, Sawyers CL. Efficacy and safety of a specific inhibitor of the BCR-ABL tyrosine kinase in chronic myeloid leukemia. N Engl J Med. 2001; 344:1031–7.

6. Kwong LN, Davies MA. Targeted therapy for melanoma: rational combinatorial approaches. Oncogene. 2014; 33:1–9.

7. Curtin JA, Fridlyand J, Kageshita T, Patel HN, Busam KJ, Kutzner H, Cho KH, Aiba S, Brocker EB, LeBoit PE, Pinkel D, Bastian BC. Distinct sets of genetic alterations in melanoma. N Engl J Med. 2005; 353:2135–47.

8. Chapman PB, Hauschild A, Robert C, Haanen JB, Ascierto P, Larkin J, Dummer R, Garbe C, Testori A, Maio M, Hogg D, Lorigan P, Lebbe C, et al. Improved survival with vemurafenib in melanoma with BRAF V600E mutation. N Engl J Med. 2011; 364:2507–16.

9. Flaherty KT, Puzanov I, Kim KB, Ribas A, McArthur GA, Sosman JA, O’Dwyer PJ, Lee RJ, Grippo JF, Nolop K, Chapman PB. Inhibition of mutated, activated BRAF in metastatic melanoma. N Engl J Med. 2010; 363:809–19.

10. Sosman JA, Kim KB, Schuchter L, Gonzalez R, Pavlick AC, Weber JS, McArthur GA, Hutson TE, Moschos SJ, Flaherty KT, Hersey P, Kefford R, Lawrence D, et al. Survival in BRAF V600-mutant advanced melanoma treated with vemurafenib. N Engl J Med. 2012; 366:707–14.

11. Shi H, Moriceau G, Kong X, Lee MK, Lee H, Koya RC, Ng C, Chodon T, Scolyer RA, Dahlman KB, Sosman JA, Kefford RF, Long GV, et al. Melanoma whole-exome sequencing identifies (V600E)B-RAF amplification-mediated acquired B-RAF inhibitor resistance. Nat Commun. 2012; 3:724.

12. Stransky N, Egloff AM, Tward AD, Kostic AD, Cibulskis K, Sivachenko A, Kryukov GV, Lawrence MS, Sougnez C, McKenna A, Shefler E, Ramos AH, Stojanov P, et al. The mutational landscape of head and neck squamous cell carcinoma. Science. 2011; 333:1157–60.

13. Berger MF, Hodis E, Heffernan TP, Deribe YL, Lawrence MS, Protopopov A, Ivanova E, Watson IR, Nickerson E, Ghosh P, Zhang H, Zeid R, Ren X, et al. Melanoma genome sequencing reveals frequent PREX2 mutations. Nature. 2012; 485:502–6.

14. Stark MS, Woods SL, Gartside MG, Bonazzi VF, Dutton-Regester K, Aoude LG, Chow D, Sereduk C, Niemi NM, Tang N, Ellis JJ, Reid J, Zismann V, et al. Frequent somatic mutations in MAP3K5 and MAP3K9 in metastatic melanoma identified by exome sequencing. Nat Genet. 2012; 44:165–9.

15. Campbell PJ, Stephens PJ, Pleasance ED, O’Meara S, Li H, Santarius T, Stebbings LA, Leroy C, Edkins S, Hardy C, Teague JW, Menzies A, Goodhead I, et al. Identification of somatically acquired rearrangements in cancer using genome-wide massively parallel paired-end sequencing. Nat Genet. 2008; 40:722–9.

16. Campbell PJ, Pleasance ED, Stephens PJ, Dicks E, Rance R, Goodhead I, Follows GA, Green AR, Futreal PA, Stratton MR. Subclonal phylogenetic structures in cancer revealed by ultra-deep sequencing. Proc Natl Acad Sci U S A. 2008; 105:13081–6.

17. Shi H, Hugo W, Kong X, Hong A, Koya RC, Moriceau G, Chodon T, Guo R, Johnson DB, Dahlman KB, Kelley MC, Kefford RF, Chmielowski B, et al. Acquired resistance and clonal evolution in melanoma during BRAF inhibitor therapy. Cancer Discov. 2014; 4:80–93.

18. Hodis E, Watson IR, Kryukov GV, Arold ST, Imielinski M, Theurillat JP, Nickerson E, Auclair D, Li L, Place C, Dicara D, Ramos AH, Lawrence MS, et al. A landscape of driver mutations in melanoma. Cell. 2012; 150:251–63.

19. Nikolaev SI, Rimoldi D, Iseli C, Valsesia A, Robyr D, Gehrig C, Harshman K, Guipponi M, Bukach O, Zoete V, Michielin O, Muehlethaler K, Speiser D, et al. Exome sequencing identifies recurrent somatic MAP2K1 and MAP2K2 mutations in melanoma. Nat Genet. 2012; 44:133–9.

20. Quintana E, Shackleton M, Foster HR, Fullen DR, Sabel MS, Johnson TM, Morrison SJ. Phenotypic heterogeneity among tumorigenic melanoma cells from patients that is reversible and not hierarchically organized. Cancer Cell. 2010; 18:510–23.

21. Salama AK, Flaherty KT. BRAF in melanoma: current strategies and future directions. Clin Cancer Res. 2013; 19:4326–34.

22. Colombino M, Capone M, Lissia A, Cossu A, Rubino C, De Giorgi V, Massi D, Fonsatti E, Staibano S, Nappi O, Pagani E, Casula M, Manca A, et al. BRAF/NRAS mutation frequencies among primary tumors and metastases in patients with melanoma. J Clin Oncol. 2012; 30:2522–9.

23. Wagle N, Van Allen EM, Treacy DJ, Frederick DT, Cooper ZA, Taylor-Weiner A, Rosenberg M, Goetz EM, Sullivan RJ, Farlow DN, Friedrich DC, Anderka K, Perrin D, et al. MAP kinase pathway alterations in BRAF-mutant melanoma patients with acquired resistance to combined RAF/MEK inhibition. Cancer Discov. 2014; 4:61–8.

24. Van Allen EM, Wagle N, Sucker A, Treacy DJ, Johannessen CM, Goetz EM, Place CS, Taylor-Weiner A, Whittaker S, Kryukov GV, Hodis E, Rosenberg M, McKenna A, et al. The genetic landscape of clinical resistance to RAF inhibition in metastatic melanoma. Cancer Discov. 2014; 4:94–109.

25. Holderfield M, Deuker MM, McCormick F, McMahon M. Targeting RAF kinases for cancer therapy: BRAF-mutated melanoma and beyond. Nat Rev Cancer. 2014; 14:455–67.

26. Flaherty KT, Infante JR, Daud A, Gonzalez R, Kefford RF, Sosman J, Hamid O, Schuchter L, Cebon J, Ibrahim N, Kudchadkar R, Burris HA 3rd, Falchook G, et al. Combined BRAF, and MEK inhibition in melanoma with BRAF V600 mutations. N Engl J Med. 2012; 367:1694–703.

27. Robert C, Karaszewska B, Schachter J, Rutkowski P, Mackiewicz A, Stroiakovski D, Lichinitser M, Dummer R, Grange F, Mortier L, Chiarion-Sileni V, Drucis K, Krajsova I, et al. Improved Overall Survival in Melanoma with Combined Dabrafenib and Trametinib. N Engl J Med. 2014.

28. Turajlic S, Furney SJ, Stamp G, Rana S, Ricken G, Oduko Y, Saturno G, Springer C, Hayes A, Gore M, Larkin J, Marais R. Whole-genome sequencing reveals complex mechanisms of intrinsic resistance to BRAF inhibition. Ann Oncol. 2014; 25:959–67.

29. Straussman R, Morikawa T, Shee K, Barzily-Rokni M, Qian ZR, Du J, Davis A, Mongare MM, Gould J, Frederick DT, Cooper ZA, Chapman PB, Solit DB, et al. Tumour micro-environment elicits innate resistance to RAF inhibitors through HGF secretion. Nature. 2012; 487:500–4.

30. Deuker MM, Marsh Durban V, Phillips WA, McMahon M. PI3’-Kinase Inhibition Forestalls the Onset of MEK1/2 Inhibitor Resistance in BRAF-Mutated Melanoma. Cancer Discov. 2014.

31. Snyder A, Makarov V, Merghoub T, Yuan J, Zaretsky JM, Desrichard A, Walsh LA, Postow MA, Wong P, Ho TS, Hollmann TJ, Bruggeman C, Kannan K, et al. Genetic basis for clinical response to CTLA-4 blockade in melanoma. N Engl J Med. 2014; 371:2189–99.

32. Hu-Lieskovan S, Mok S, Homet Moreno B, Tsoi J, Robert L, Goedert L, Pinheiro EM, Koya RC, Graeber TG, Comin-Anduix B, Ribas A. Improved antitumor activity of immunotherapy with BRAF and MEK inhibitors in BRAFV600E melanoma. Sci Transl Med. 2015; 7:29ra41.

33. Prickett TD, Agrawal NS, Wei X, Yates KE, Lin JC, Wunderlich JR, Cronin JC, Cruz P, Rosenberg SA, Samuels Y. Analysis of the tyrosine kinome in melanoma reveals recurrent mutations in ERBB4. Nat Genet. 2009; 41:1127–32.

34. Wei X, Walia V, Lin JC, Teer JK, Prickett TD, Gartner J, Davis S, Stemke-Hale K, Davies MA, Gershenwald JE, Robinson W, Robinson S, Rosenberg SA, et al. Exome sequencing identifies GRIN2A as frequently mutated in melanoma. Nat Genet. 2011; 43:442–6.

35. Mishra PJ, Ha L, Rieker J, Sviderskaya EV, Bennett DC, Oberst MD, Kelly K, Merlino G. Dissection of RAS downstream pathways in melanomagenesis: a role for Ral in transformation. Oncogene. 2010; 29:2449–56.

36. Michaloglou C, Vredeveld LC, Soengas MS, Denoyelle C, Kuilman T, van der Horst CM, Majoor DM, Shay JW, Mooi WJ, Peeper DS. BRAFE600-associated senescence-like cell cycle arrest of human naevi. Nature. 2005; 436:720–4.

37. Vredeveld LC, Possik PA, Smit MA, Meissl K, Michaloglou C, Horlings HM, Ajouaou A, Kortman PC, Dankort D, McMahon M, Mooi WJ, Peeper DS. Abrogation of BRAFV600E-induced senescence by PI3K pathway activation contributes to melanomagenesis. Genes Dev. 2012; 26:1055–69.

38. Johannessen CM, Johnson LA, Piccioni F, Townes A, Frederick DT, Donahue MK, Narayan R, Flaherty KT, Wargo JA, Root DE, Garraway LA. A melanocyte lineage program confers resistance to MAP kinase pathway inhibition. Nature. 2013; 504:138–42.

39. Li H, Durbin R. Fast and accurate short read alignment with Burrows-Wheeler transform. Bioinformatics. 2009; 25:1754–60.

40. McKenna A, Hanna M, Banks E, Sivachenko A, Cibulskis K, Kernytsky A, Garimella K, Altshuler D, Gabriel S, Daly M, DePristo MA. The Genome Analysis Toolkit: a MapReduce framework for analyzing next-generation DNA sequencing data. Genome Res. 2010; 20:1297–303.

41. Ye K, Schulz MH, Long Q, Apweiler R, Ning Z. Pindel: a pattern growth approach to detect break points of large deletions and medium sized insertions from paired-end short reads. Bioinformatics. 2009; 25:2865–71.

42. Chiariello M, Vaque JP, Crespo P, Gutkind JS. Activation of Ras and Rho GTPases and MAP Kinases by G-protein-coupled receptors. Methods Mol Biol. 2010; 661:137–50.