707. Retroviral Insertion Site Analysis in Rhesus Macaques Transplanted with CD34+ Hematopoietic Stem Cells Transduced with a Simian Immunodeficiency Virus-Based Lentiviral Vector

Abstract

Two children involved in the successful retroviral gene therapy clinical trial of SCID-X1 (2 year survival: 100%, FFTF 60%) developed a malignant T-cell lymphoproliferation 34 months after treatment. In both cases, vector integration into the vicinity of exon 1 of the protooncogene LMO2 occured (Hacein-Bey-Abina, 2003). The potential for unwanted side effects from retroviral insertional mutagenesis is now a focus of biosafety considerations in gene transfer using integrating vector systems. Several groups could show that gene coding areas, in particular transcription start sites, are preferred insertion loci (Schroder, 2002; Wu, 2003). While these studies mapped a great number of integration sites with statistical significance, they were limited to in vitro gene transfer into nonhematopoietic human cell lines. As T cells do not grow in these cases unless they express the gamma c chain, we reasoned that there may be selection pressure on the retrovirus integration site distribution. Using a highly sensitive and specific linear amplification mediated PCR (LAM-PCR), we analyzed DNA samples from 8 patients enrolled in the X-SCID gene therapy trial. We could perform successive insertion site analysis and subsequently identify more than 400 integration sites after transplantation. In all patient samples examined, a stable polyclonal profile demonstrated polyclonal hematopoietic repopulation. To distinguish the influence of required transgene expression on retrovirus integration site distribution, we analyzed insertion sites in patient CD34+ cells prior to autologous transplantation, shortly after ex vivo transduction. More than 100 integration sites present prior to transplantation could be identified for comparative pre-/post-tranplantation analysis. No obvious difference in insertion site distribution could be detected between pre- and post-transplant samples form a first analysis. However, our data indicate that X-SCID correction, which depends on transgene expression, favors chromosomal locations associated with active transciption. We conclude that the distribution of retrovirus integration sites can be assessed by random sampling in a clinical trial setting and may differ significantly from previously published assumptions, especially in settings of selective advantage related to transgene expression.