Large Scale Analysis of Foamy Virus Vector Integration Sites in Human CD34+ Cells

• Published in: Blood Vol. 104; no. 11; p. 496
• Main Authors: Trobridge, Grant D., Miller, Daniel G., Jacobs, Michael A., Allen, James M., Olson, Erik, Kiem, Hans-Peter, Kaul, Rajinder, Russell, David W.
• Format: Journal Article
• Language: English
• Published: Elsevier Inc 16.11.2004
• ISSN: 0006-4971; 1528-0020
• DOI: 10.1182/blood.V104.11.496.496

The ability of retroviruses to efficiently integrate into the host cell's genome has led to their use as gene delivery vehicles for gene therapy. However, integration in the genome can have adverse effects as observed in a gene therapy trial for X-linked SCID using an oncoretroviral vector. Recent studies have shown that an oncoretroviral vector integrated preferentially near transcription start sites and that a lentiviral vector integrated preferentially within genes. Foamy viruses are integrating retroviruses with many properties that distinguish them from onco- or lentiviruses, perhaps the most important characteristic for gene therapy being that they are non-pathogenic. We previously showed that foamy vectors efficiently transduce CD34+ SCID mouse-repopulating cells (SRCs) from human mobilized peripheral blood, demonstrating their potential for hematopoietic stem cell gene therapy. We present here the first large-scale analysis of foamy vector integration sites. Integration sites were determined by infecting human CD34+ cells or normal fibroblasts with a foamy vector carrying a bacterial origin of replication, then rescuing plasmids containing vector provirus-genomic junction sites in bacteria, and sequencing the foamy vector LTR-human genome junctions. Over 1900 unique integration sites in human CD34+ cells and 1000 unique sites in normal human fibroblasts were mapped using the human genome database. The foamy vector did not integrate preferentially into genes. The percentage of integration sites within Refseq genes in human CD34+ cells, human fibroblasts and randomly generated sites was 29, 23, and 32% respectively. Foamy vectors showed only a slight preference for integration within 1 kb 5′ or 3′ of Refseq transcription start sites. In summary, our data show that foamy vectors have a distinct integration site profile relative to oncoretroviral and lentiviral vectors. Future studies will be required to determine if the unique integration site preference of foamy vectors translates into a reduced risk for oncogenesis in gene therapy applications.