Ancient human endogenous retroviruses contribute to genetic evolution and regulate cancer cell type-specific gene expression.

Human endogenous retroviruses (HERVs), a type of transposable elements (TEs), play crucial roles in human placental morphogenesis, immune response, and cancer progression. Emerging evidence suggests that TEs have been a rich source of regulatory elements in the human genome, but little is known about the global impact of HERVs on transcriptional networks in cancer. Using genome-wide approaches, we show that HERVs are composed primarily of three ancient superfamilies, ERVL-MaLR, ERVL, and ERV1. This analysis suggests that the integration of exonic, intronic, and intergenic HERVs as well as human or Hominidae gene-specific HERVs contribute to human genomic innovation. HERVs exonized in genes are mainly located in the 3' untranslated region (UTR) or 3' end and participate in basic biological processes. Active HERVs are located mainly in intronic and intergenic regions and tend to function as enhancers and contribute to cancer cell type-specific gene expression. More importantly, HERVs may also define chromatin topologically associating domain (TAD) and loop boundaries in a cell type-specific manner. Taken together, these findings reveal that ancient HERV elements are a source of diverse regulatory sequences, including 3' UTRs, 5' UTRs, promoters, and enhancers, and they contribute to genetic innovation and cancer cell type-specific gene expression, highlighting the previously underestimated importance of these elements.