The HERV-K(HML-2) Transcriptome in Non-diseased Tissue
Human endogenous retrovirus (HERV) transcripts are known to be highly expressed in cancers, yet their activity in non-diseased tissue is largely unknown. Using the GTEx RNA-seq dataset from normal tissue sampled at autopsy, we characterized individual expression of the recent HERV-K(HML-2) provirus group across 13,000 different samples of 54 different tissues from 948 individuals. HML-2 transcripts could be identified in every tissue sampled, and were elevated in the cerebellum, pituitary, testis, and thyroid. 37 different individual proviruses were expressed in one or more tissues, representing all three LTR5 subgroups. 9 proviruses were identified as having LTR-driven transcription, 7 of which belonged to the most recent LTR5HS subgroup. Proviruses of different subgroups displayed a bias in tissue expression which may be associated with differences in transcription factor binding sites in their LTRs. Provirus expression was greater in evolutionarily older proviruses with an earliest shared ancestor of gorilla or older. HML-2 expression was significantly affected by biological sex in one tissue, while age and timing of death (Hardy score) had little effect. Proviruses containing intact gag, pro and env ORFs were expressed in the dataset, with almost every tissue measured potentially expressing at least one intact ORF (gag).