Differential use of Alternative Transcripts in Breast Cancer Subsets
Identifiers: SRA: SRP014830
Microarray based analysis of transcriptional profiles and genomic sequencing of tumor samples have greatly improved the biological understanding and molecular classification of breast cancer. However, certain aspects of the transcriptome, including characterization of differential expression of splice variants, are not well captured by these methods. We report results from single-end Illumina-based RNA-sequencing of 53 primary breast cancers and 6 "normal" breast samples from mammoplasty patients, using both 29 bp reads (on 29 tumor samples) and 100 bp reads (on 24 tumor samples and 6 normals). Novel as well as widely used computational approaches were used to identify differentially expressed splice variants of known RefSeq genes. This analysis identified a number of variants differentially spliced in subsets of breast tumor samples. Some of these splice variants were specific to either ER+ or ER- breast cancers, including the prostate cancer oncogene TPD52/PrLZ and intronic-start variants of genes IQCG and ACOX2, both of which have been implicated as potentially oncogenic in other cancers. We also found differential splicing in subsets of breast cancer samples which did not associate with ER or HER2 status, including a novel short-form Lactoferrin variant similar to dLTF, and the secreted protein NESP55 at the biallelically imprinted GNAS locus. The expression of several variants were verified in both tumor samples and breast cancer derived cell lines using RT-PCR. These data demonstrate that differential use of transcript variants may play an important role in the phenotype of breast cancer subtypes, and can be robustly identified by RNA sequencing analysis.