Accumulation of long-term transcriptionally active integrated retroviral vectors in active promoters and enhancers

Most retroviruses preferentially integrate into certain genomic locations and, as a result, their genome-wide integration patterns are non-random. We investigate the epigenetic landscape of integrated retroviral vectors and correlate it with the long-term stability of proviral transcription. Retroviral vectors derived from the avian sarcoma/leukosis virus expressing the GFP reporter were used to transduce the human myeloid lymphoblastoma cell line K562. Because of efficient silencing of avian retrovirus in mammalian cells, only ∼3% of established clones displayed stable proviral expression. We analyzed the vector integration sites in non-selected cells and in clones selected for the GFP expression. This selection led to overrepresentation of proviruses integrated in active transcription units, with particular accumulation in promoter-proximal areas. In parallel, we investigated the integration of vectors equipped with an anti-silencing CpG island core sequence. Such modification increased the frequency of stably expressing proviruses by one order. The modified vectors are also overrepresented in active transcription units, but stably expressed in distal parts of transcriptional units further away from promoters with marked accumulation in enhancers. These results suggest that integrated retroviruses subject to gradual epigenetic silencing during long-term cultivation. Among most genomic compartments, however, active promoters and enhancers protect the adjacent retroviruses from transcriptional silencing.