4. In Vitro and In Vivo Analysis of Feline Immunodeficiency Virus-Based Lentiviral Vector Integration

Abstract

Retroviral vector-mediated gene transfer holds promise in the treatment of a variety of genetic diseases. A distinguishing feature of retroviral vectors is their ability to integrate into the host genome, resulting in sustained transgene expression. However, with the ability to integrate comes a risk of insertional mutagenesis. Viral integration mapping studies provide an important first step in understanding the biology and fate of the vector DNA in host cells and are directly applicable to gene transfer using retroviral vectors. Using a sensitive nested PCR-based genome walking technique, we analyzed feline immunodeficiency virus (FIV)-based lentiviral vector integration events in vitro in FIV-transduced human HepG2 hepatoma cells and in vivo in the liver of mice receiving systemic injection of FIV vector. For in vitro mapping studies, we analyzed 869 sequences, which represented 211 distinct FIV integration sites. We also compared the FIV integration preferences with published results for HIV and MLV vectors. We found that ~75% of FIV integration occurred in a RefSeq gene and intron regions were preferred over exons. Of the 159 RefSeq genes with FIV integration, 139 were represented on a HepG2 microarray data set (Stanford University). The median expression level for these 139 genes was 138 and was about 3.8-fold higher than that of all the genes on the array (median expression levels for all genes on the array was 36). These data indicate that like HIV and MLV, FIV integration favors actively transcribed genes. Furthermore, similar to HIV but distinct from MLV, FIV integrated throughout the entire length of the transcriptional units, and regions around the transcription start sites were not the favored target for FIV integration. Although no specific sequences in the host genome were identified as FIV integration sites, FIV integration sites appear to be surrounded by A/T rich regions. Interestingly, FIV integrations did not coincide with the hot spots described previously for HIV integration in a human T cell line. For in vivo mapping studies, we analyzed FIV integration events in the liver of FIV-injected mice at day 5 and day 21 post gene transfer. To date, ~30 clones have been sequenced and 8 unique FIV integration sites have been identified. Five of them occurred in a gene (4 in introns and 1 in exon). The remaining 3 FIV integrations were in non-coding regions far away from transcriptional start sites. Similar mapping studies in additional animals are currently ongoing. These data suggest that although FIV integration preference is more similar to HIV than to MLV, certain differences may exist between FIV and HIV vectors. In addition, results may vary depending on the target cell type and the proliferation status of the transduced cells. These findings provide important background for understanding the host-vector interactions and stress the need for developing more site-restricted or insulated lentiviral vectors for gene transfer applications.