299. Insertion Site Analysis and In Vivo Selection

Abstract

Listen

Translate article

Leukemias arising in 2 patients treated with gene therapy for X-SCID may be viewed to result from a collaboration between a transcription factor, abberantly expressed due to insertional mutagenesis, and a constitutively active, receptor-initiated growth signal. We are undertaking a systematic comparison of retrovirus integration sites in transduced hemopoietic cells prior to and following in vivo selection. The system for selection uses a conditional derivative of the thrombopoietin receptor, F36Vmpl, that transmits a growth signal in the presence of an artificial ligand called a chemical inducer of dimerization (CID). Our initial studies have focused on comparing insertion sites identified by Southern analysis with those identified by LAM-PCR. A cohort of 11 mice was transplanted with marrow cells transduced with a bicistronic vector encoding green fluorescent protein (GFP) and F36Vmpl. Marrow aspirates were obtained at baseline and following administration of 3 monthly cycles of CID (AP20187 10mg/kg IP daily for 3 days). 10 of 11 mice exhibited strong responses to CID treatment, with the percentage of GFP positive red cells rising from <20% at baseline to 90% following CID exposure. Following selection, marrow cells from these mice were then transplanted 1:1 into 11 secondary recipients which in turn received 3 additional monthly courses of CID. No GFP-positive cells were detected in 3 of the secondary recipients, and CID administration in these mice failed to elicit a response. Responses to CID treatment were observed in the remaining 8 mice, and in 3 secondary recipients the responses were the same or better than their donors. We compared retorviral insertion sites in donor-recipient pairs by Southern analysis and LAM-PCR. 4 of 11 donor-recipient pairs displayed a common banding pattern by Southern analysis following HindIII digestion, including the 3 pairs in which the hematological response of the secondary recipient equaled or bettered that of their donors. The same 4 donor-recipient pairs were the only ones found to have common proviral insertion sites as identified by LAM-PCR. However, in only one pair did the number of unique insertions identified by LAM-PCR equal that identified by Southern (4 insertions each). In each of the remaining 3 pairs only a single common insertion was identifed by LAM-PCR, whereas Southern detected 3 shared insertions in two pairs, and 5 shared insertions in the remaining pair. Our findings to date demonstrate that traditional LAM-PCR detects only a fraction of insertion sites that are sufficiently abundant to be detectable by Southern blot. As expected, LAM-PCR also identified multiple unique insertions in animals that had no detectable signal by Southern. In addition to modifying our LAM-PCR method to reveal abundant, yet undetected insertion sites, we are extending our analysis to single progenitor colonies taken from mice prior to and following F36Vmpl-based in vivo selection.