Assessing the interplay of RNA-fusion events with cancer driver mutations in melanoma patients

Abstract

Introduction: Treatment options for cutaneous melanoma in recent years have undergone major improvement, primarily due to development of targeted and immunotherapies. Still patient heterogeneity remains a challenge and studies looking for predictors of response by profiling DNA-level somatic alterations in large-scale cohorts have failed to identify consensus biomarkers. In the present study we examine RNA-fusion events to complement other genomic alterations. We hypothesize that this will assist better clinical decision-making (targeted or immunotherapy) in individual patients. Materials and Methods: We performed whole exome sequencing (WES) and complimentary RNA-seq of 28 tumor tissue samples from 28 patients. Of these, ten were wild type for BRAF or NRAS hotspot mutations. We profiled RNA-seq datasets for fusion events. Briefly, after quality-based trimming, fastq reads were aligned using STAR aligner to detect chimeric transcripts. Initial candidates were annotated and filtered for spurious calls using STAR-Fusion and FusionAnnotator. Results and Discussion: We detect two in-frame BRAF fusions, one in BRAF V600E (RABEP1-BRAF) and other in a BRAF kinase-dead mutation (AGKBRAF) sample. As in previous reports, BRAF is the 3prime partner in both the cases with its first 8 and 7 exons deleted respectively. The exon loss may increase BRAF kinase domain dimerization. AGK is a (lipid) kinase that is previously reported as a BRAF fusion partner in melanoma. We also detect an in-frame G12D KRAS fusion (CCDC91-KRAS) in a patient lacking BRAF/ NRAS hotspot mutations. This fusion retains exon 2 and beyond of KRAS, and effectively constitutes the entire protein coding region of G12D KRAS. Conclusion: BRAF fusions are reported not to respond to BRAF inhibitors, so the V600E BRAF fusion is contra-indicated for these therapies. Thus, it is crucial to combine somatic alteration and fusion events as part of the standard NGS-based profiling for melanoma patients. We are further analysing additional fusion candidates detected, and examining the individual contributions when hotspot somatic mutation and gene fusion events are combined.