Landscape of BRAF transcript variants in human cancer

The BRAFV600E mutant kinase is widely studied as a cancer driver and therapeutic target. However, the regulation of BRAF gene expression remains largely unexplored. Here in this thesis, I investigate how the annotation of the biologically relevant protein-coding BRAF-ref and BRAF-X1 isoforms has evolved in public databases, and I faced the challenges posed by the discrimination and quantification of these BRAF isoforms, providing a detailed discussion and comparison of cutting-edge in silico methods such as short-read sequencing, long-read sequencing, and single-cell sequencing, which represent the state-of-the-art approaches for isoform identification and quantification. Furthermore, I investigate the clinical implications of the BRAF-X1/BRAF-ref transcript ratio, examining how this ratio could influence clinical outcomes. Currently, BRAF-ref corresponds to NM_004333.6 (NCBI), ENST00000646891.2/BRAF-220 (Ensembl), MANE select (MANE), PRINCIPAL:4 (APPRIS), and TCONS_00952581 (FLIBase). BRAF-X1 corresponds to NM_001354609.2 (NCBI), ENST00000496384.7/BRAF-204 (Ensembl), ALTERNATIVE BRAF isoform (APPRIS) and TCONS_00952571 (FLIBase). I built IsoWorm, a bioinformatic pipeline specifically tailored to discriminate and quantify BRAF isoforms and I employed it to analyze more than 500 cancer cell lines and 700 cancer tissue samples. Thanks to the FLIBase database, I also re-analyzed TCGA data on more than 9000 cancer tissue samples and 600 adjacent to tumoral tissue samples, and additionally, I examined more than 2,600 normal tissue samples from the GTEx database. I integrated also additional modules into IsoWorm, enabling me to analyze differents types of short RNA-seq data, such as Quant 3’ sequencing and single-cell sequencing data. This permits me a more detailed investigation into the differences between BRAF isoforms. I consistently found that BRAF-X1 (now BRAF-204) is the most abundant BRAF isoform in human cancer and in normal tissues samples, it is 1.5 to 75 (in cancer) and 4 to 64 times (in normal samples) more expressed than BRAF-ref (now BRAF-220). Crucially, I identified how in papillary renal cell carcinoma (KIRP) an extreme BRAF-204/BRAF-220 ratio, together with the tumor stages and metastasis, higher the risk of the patient’s overall survival. Finally, thanks to the application of in silico structural biology techniques such as modeling with AlphaFold and molecular dynamics simulations, I investigated the structural differences between the C-termini of BRAF-220 and BRAF-204. Preliminary data suggest that the C-terminus of BRAF-220 is more stable compared to that of BRAF-204. This structural stability could differentiate the two isoforms in terms of their biological roles, stability, and functionality. Together with the experimental characterization undertaken over the years, our in silico analyses ultimately establish BRAF as a mix of isoforms, with BRAF-204 being much more expressed than BRAF-220. These findings prompt a systematic benchmarking of the isoforms regarding molecular mechanisms, biological activities, and clinical relevance.