Background:  Endogenous retroviruses (ERV) are present in all higher eukaryotes, where they represent a large fraction of the genomes (8% in humans, divided into about 100 families). They are remnants of ancestral infections by bona fide retroviruses that have infected the germline and have thereafter been transmitted in a Mendelian manner. Conversely, they are a reservoir from which new infectious retroviruses can arise via recombination. The similarity between ERV and present-day infectious retroviruses is remarkable, in terms of genome organization, encoded proteins (with even a functional homolog of the HIV Rev protein for the HERV-K family), structures of the virus-like particles, and to some extent replicative cycles (albeit in a few cases strictly intracellular). Some ERV genes are likely to play a physiological role, associated with the cell–cell fusogenic activity of their encoded envelope proteins:  in both primates and rodents 2 such genes (so-called syncytins) have been co-opted by the host—and conserved over >30 millions years of evolution—for a role in placentation and syncytiotrophoblast formation. Yet, as a rule, ERV are silent, due to cellular repression mechanisms, preventing deleterious insertional mutagenesis:  these include chromatin alteration (eg, CpG methylation), homology-dependent gene silencing, RNA interference, and cytosine deaminase activity of cellular APOBEC proteins, which also restrict present-day infectious retroviruses. However, in a few cases (essentially in tumors), ERV become transcriptionally active and, as such, are likely to produce effects similar to those triggered by exogenous infectious retroviruses:  these effects include insertional mutagenesis, and direct effects of the viral proteins themselves, especially the envelope protein, which exerts immunosuppressive activity. In this respect, we have shown in model mouse tumors (melanoma, neuroblastoma) that the latter effect contributes to tumor cell growth, via inhibition of tumor immune surveillance.

Conclusions:  Although most ERV are defective, due to genetic drift, the molecular biology of functional elements that could be identified in the human and murine genomes will be presented. Overall lessons that ERV teach us on the pathological mechanisms of infectious retroviruses—and vice versa—will be discussed.