Transcriptome sequencing of human breast cancer reveals aberrant intronic transcription in amplicons and dysregulation of alternative splicing with major therapeutic implications.

Advances in genomic and transcriptome sequencing are revealing the massive scale of previously unrecognised alterations occurring during neoplastic transformation. Breast cancers are genetically and phenotypically heterogeneous. Each of the three major subtypes [ERBB2 amplified, estrogen receptor (ESR)-positive and triple-negative] poses diagnostic and therapeutic challenges. Here we show, using high-resolution next-generation transcriptome sequencing, that in all three breast cancer subtypes, but not matched controls, there was significant overexpression of transcripts from intronic and untranslated regions in addition to exons from specific genes, particularly amplified oncogenes and hormone receptors. For key genes ERBB2 and ESR1, we demonstrate that overexpression is linked to the production of highly modified and truncated splice variants in tumours, but not controls, correlated with tumour subtype. Translation of these tumour-specific splice variants generates truncated proteins with altered subcellular locations and functions, modifying the phenotype, affecting tumour biology, and targeted antitumour therapies. In contrast, tumour suppressors TP53, BRCA1/2 and NF1 did not show intronic overexpression or truncated splice variants in cancers. These findings emphasize the detection of intronic as well as exonic changes in the transcriptional landscapes of cancers have profound therapeutic implications.