Molecular Medicine Israel

Genomic deletions explain the generation of alternative BRAF isoforms conferring resistance to MAPK inhibitors in melanoma

Highlights

  • Alternative BRAF isoforms (altBRAFs), related to resistance, are caused by genomic deletions
  • altBRAFs in BRAF inhibitor-resistant models are not generated by alternative splicing
  • altBRAFs also occur in treatment-naive tumors and regardless of BRAF mutational status

Summary

Resistance to MAPK inhibitors (MAPKi), the main cause of relapse in BRAF-mutant melanoma, is associated with the production of alternative BRAF mRNA isoforms (altBRAFs) in up to 30% of patients receiving BRAF inhibitor monotherapy. These altBRAFs have been described as being generated by alternative pre-mRNA splicing, and splicing modulation has been proposed as a therapeutic strategy to overcome resistance. In contrast, we report that altBRAFs are generated through genomic deletions. Using different in vitro models of altBRAF-mediated melanoma resistance, we demonstrate the production of altBRAFs exclusively from the BRAF V600E allele, correlating with corresponding genomic deletions. Genomic deletions are also detected in tumor samples from melanoma and breast cancer patients expressing altBRAFs. Along with the identification of altBRAFs in BRAF wild-type and in MAPKi-naive melanoma samples, our results represent a major shift in our understanding of mechanisms leading to the generation of BRAF transcripts variants associated with resistance in melanoma.

Introduction

Alternative splicing (AS) of mRNA precursors is a prevalent mechanism of gene expression regulation in complex organisms by which differential use of exons and introns greatly contributes to proteome diversity and physiological regulation.1 Transcriptome-wide analyses have revealed frequent dysregulation of AS in cancer by multiple mechanisms.2,3 These findings opened the door to explore splicing modulation as a therapeutic strategy in cancer.4,5,6,7

A paradigmatic and compelling example of the clinical impact of differential exon usage is found in the B-Raf proto-oncogene (BRAF), a serine/threonine kinase of remarkable importance in cancer biology.8 Initial enthusiasm fueled by unprecedented responses to BRAF and mitogen-activated protein kinase kinase MEK inhibitors (BRAFi and MEKi) in advanced BRAFV600-mutant melanoma9,10,11,12,13 was tempered by evidence of relapse within the first year in ∼50% of patients, due to the acquisition of resistance.14,15,16 Resistance to BRAFi typically involves reactivation of the MAPK (mitogen-activated protein kinase) pathway (also known as RAS/RAF/MEK/ERK pathway). Although a variety of mechanisms have been reported to contribute to BRAFi resistance in melanoma,15,16,17,18,19,20,21,22,23,24,25,26,27 in up to 30% of patients receiving BRAFi, monotherapy resistance has been attributed to the production of alternative BRAF mRNA isoforms (altBRAFs).8,16,17,18,19,20,21,22,23,24,25 First described in BRAFV600-mutant melanoma, altBRAFs have been identified in other BRAFV600-mutant tumor types treated with BRAFi ± MEKi, such as lung adenocarcinomas.28 The resulting BRAF protein isoforms typically lack the RAS-binding domain and evade BRAFi due to their tendency to dimerize also in a RAS-independent manner, thus reactivating the MAPK pathway.17,26,29 altBRAFs were first described in in vitro models generated by exposure of a BRAFV600E melanoma cell line (SKMEL-293) to the BRAFi vemurafenib.17 

Some of the resistant sublines were found to express a BRAF isoform lacking exons 4–8 (BRAF3-9), and because array comparative genomic hybridization (CGH) results did not find evidence of intragenic deletions, the production of BRAF3-9 was attributed to AS.17 A variety of altBRAFs have been subsequently described, lacking diverse combinations of exons between exon 2 and exon 10 (BRAF1-11, BRAF2-11, BRAF3-11, and BRAF1-9) both in cell lines and in patients’ tumors, which again have been attributed to AS of BRAF pre-mRNA.14,18,19,20,21,22,23,27 Samples expressing altBRAFs were found to have acquired resistance to MAPK inhibitors (MAPKi) and harbored BRAFV600 mutations, namely T1799A, located in exon 15 and leading to V600E substitution. These findings, together with the identification of a mutation in intron 8 supposedly responsible for BRAF3-9 production and RNA splicing-mediated BRAFi resistance,30 led Salton et al. to propose modulation of pre-mRNA splicing as a potential therapeutic strategy to overcome resistance to BRAFi in melanoma.30 Indeed, detailed understanding of the molecular mechanisms involved can guide the development of novel therapeutic approaches to overcome resistance to MAPKi, which remains a medical priority.

Here, using a variety of transcriptome datasets, we demonstrate the presence of altBRAFs regardless of the BRAFV600 mutational status and previous exposure to MAPKi. Strikingly, using a collection of resistant melanoma cell lines expressing altBRAFs, as well as patients’ tumor samples, we provide conclusive evidence that the generation of altBRAF is due to intragenic genomic deletions, and thus not the result of choices between alternative competing splice sites. These findings are relevant to understand the mechanisms of drug resistance and for potential therapeutic developments to prevent the otherwise frequent disease relapses.

Results

altBRAF mRNA isoforms can be detected in non-V600 BRAF-mutant tumors and in MAPKi-naive samples

To expand previous studies of BRAF mRNA isoforms16,18,23,27 to larger cohorts of melanoma samples, we used VAST-TOOLS (Vertebrate Alternative Splicing and Transcription Tools)31,32 to identify and quantify mRNA read counts supporting every possible exon-exon junction (EEJ) combination across the BRAF gene, with a minimum coverage of 5 supporting reads in at least 1 sample. We collected a total of 270 publicly available melanoma RNA sequencing (RNA-seq) samples (see STAR Methods), from both patient tumors and cell lines (119 and 151 samples, respectively). Patients’ tumor samples included 57 primary melanomas and 62 metastatic lesions. Most samples were BRAF mutant (200/270, 74%) and 83 samples (44 from patients’ tumors and 39 from cell lines) were described as resistant to BRAFi ± MEKi (Table S1). In addition to all canonical EEJ, we detected reads supporting several alternative, noncanonical EEJ (Figure 1A), most of which were of uncertain biological significance. Importantly, reads supporting alternative EEJ of BRAF1-11 and BRAF3-9 isoforms, which were previously related to MAPKi resistance,17,19 were also detected (Table S2). We identified 7 samples that expressed BRAF1-11 or BRAF3-9. Four of them were reported as BRAF mutant and were exposed to MAPKi (only 3 could be considered resistant) (Figures S1A–S1D). Strikingly, we also identified altBRAFs in 3 BRAF-wild-type primary melanomas, and consequently highly unlikely to have been exposed to MAPKi….

Sign up for our Newsletter