Alternative promoters in CpG depleted regions pervasively account for epigenetic misregulation of cancer transcriptomes

Abstract

AbstractEukaryotic genes are regulated by multiple alternative promoters with distinct expression patterns. In cancer, alternative promoters are pervasively utilized, but our understanding of the mechanism of activation and how their regulatory architecture differs from reference promoters remains elusive. We analyzed 100 CAGE-seq libraries from HCC patients and annotated 4083 alternative promoters in 2926 multi-promoter genes that are known genes involved in hepatocarcinogenesis. Many alternative promoters are undetected in the normal liver. We find that multi-promoter genes are enriched among genes downregulated in the tumor, highlighting alternative promoters’ impact in global transcription changes in cancer. Alternative promoters are depleted for CpG islands, have narrow nucleosome depleted regions, and are enriched for sharp promoters as well as tissue-specific transcription factors. Alternative promoters have high DNA methylation levels around transcription start sites. Tumor cells globally lose DNA methylation, but there exists a hierarchical retention of intragenic DNA methylation, which is dictated by the genomic CG content. As such, intragenic CG-poor regions lose methylation, while CG-rich regions retain it, a phenomenon caused by differential binding of H3K36me3, DNMT3B, TET1 and SETD2. Thus, the selective loss of DNA methylation in CG-poor regions opens the chromatin and makes these regions accessible for transcription. Upon transcription factors availability, alternative transcription can pervasively occur in cancer. These results provide a framework for understanding the importance of alternative promoters in controlling the tumor transcriptomes, highlighting their architecture and role in regulatory mechanism(s).