Polyclonality of Retroviral Insertion Sites Following Successful Retroviral-Mediated Gene Transfer of CD18 in Canine Leukocyte Adhesion Deficiency.

Canine leukocyte adhesion deficiency (CLAD) represents a large-animal model for the human disease leukocyte adhesion deficiency, type 1 (LAD-1). Both CLAD and LAD are characterized by life-threatening bacterial infections and are due to mutations in the leukocyte integrin CD18. We recently reported successful gene therapy in six dogs with CLAD following infusion of autologous CD34+ bone marrow hematopoietic stem cells transduced with the retroviral vector PG13/MSCV-cCD18. The level of CD18+ gene corrected leukocytes ranged from 1.3% to 8.4% at one year post-transplant in the six CLAD dogs. We continue to monitor four of the six dogs for long-term follow-up. To date no gene transfer-related adverse events have been observed in any of the dogs. However, verification of polyclonality of insertion sites in the gene-corrected cells and characterization of the genomic sites of retroviral integration remain critical safety issues. To this end, we have demonstrated the polyclonality of contributing retrovirus-transduced cells in all six dogs using linear-amplification mediated PCR (LAM-PCR) with DNA obtained from peripheral blood leukocytes up to 2 years following infusion of gene-corrected cells. Multiple clones were shown to contribute to hematopoiesis; no predominant clones emerged over time. Retroviral insertion sites in the six dogs were identified using ligation-mediated PCR (LM-PCR) followed by sequencing of the genomic sequence adjacent to the integrated viral 3′ LTR. Mapping of insertion sites was facilitated by the recent completion of the annotated dog genome sequence. To date, we have identified 341 unique retroviral insertion sites. Analysis of these insertion sites indicated that 46% of the sites are located within gene-coding regions of the genome (introns or exons). Another 35% of the sites were located within 50 kb of the transcription start site of a gene-coding sequence. These observations match previous reports of preferential integration of MLV retroviral vectors near coding sequences. The insertion sites were well-distributed across all chromosomal locations, with no evidence for insertional hotspots or preferential growth of clones containing common insertion sites: only six pairs of insertion sites were located within 5 kb of each other. Further analysis was performed on a sub-group of 294 genes that had insertion sites located either in them or within 50 kb of the transcription start site. 19 of these genes (6.5%) were present in either or both of two cancer gene databases (Sanger Institute Cancer Gene Census Table, Mouse Retrovirus Tagged Cancer Gene Database). Functional annotation of the 294 genes revealed some groups of interest: 17% of the genes were involved in phosphorylation, 4% were proto-oncogenes, 12% encoded zinc-finger proteins, 23% were involved in signal transduction, 2% encoded transforming proteins, 5% were involved in cell proliferation, 8% were involved in transcriptional regulation and 4% encoded protein kinases. We are continuing to sequence and characterize a large number of retroviral insertion sites from the gene-corrected CLAD dogs. Analysis of these insertion sites should provide considerable insight into the risk of genotoxicity through insertional mutagenesis in a disease-specific large-animal model of a human hematopoietic disease.