283. In Vivo Selection of MGMT(P140K) Gene Modified Hematopoietic Cells in the Nonhuman Primate Unmasks a Dormant Pool of Repopulating Clones with Progenitor Cell Ontology

Abstract

Previously published gene modified cell transplantation studies in nonhuman primate models have described several key features of hematopoietic reconstitution (Kim, S. et al., 2014 and Wu, C. et al., 2014). These studies applied either retrovirus integration site analysis (ISA) or DNA barcode sequencing (DBS) to identify hematopoietic clones in bone marrow or peripheral blood cells. However, these distinct methods detected long-term clones at different time points (1 year after transplant by ISA; ~3 months after transplant by DBS), and neither evaluated reconstitution in the context of selective advantage, which is often the case in gene therapy. Here, we tracked tens of thousands of unique clones in 8 pigtail macaques for up to 10 years following myeloablative transplantation with autologous, lentivirus (LV) gene-modified CD34+ cells. Seven animals received cells gene modified with the P140K mutant methylguanine methyltransferase transgene, conferring resistance to O6- benzylguanine (O6BG) and bis-chloroethylnitrosourea (BCNU) chemotherapy. In two animals, MGMT(P140K)-expressing LVs were DNA barcoded, permitting simultaneous ISA and DBS tracking. Gene marking in peripheral blood cells ranged from 2.3% to 66%. Direct comparison of ISA and DBS demonstrated that abundant clones (&gt 1% of sequence reads) are readily detected by both methods, but DBS captures up to 2-fold more lower abundance clones. Before in vivo selection with O6BG/BCNU, we observed a cell dose-dependent, successive pattern of hematopoietic reconstitution by ISA analysis, with short-term clones declining within 100 days after transplantation. Long-term clones were observed as early as 1 month after transplant. In the first year after transplant, persistent clones ranged from 8% to 54% of clones detected at a &gt 1% frequency, and remained stable in the absence of selective pressure. Importantly, when O6BG/BCNU was administered we observed novel clonal patterns, which directly correlated with transplanted cell dose and time of chemotherapy administration after transplant. In all animals, chemotherapy induced emergence of previously undetected clones. In animals receiving cell doses exceeding 35×106 CD34+ cells/kg (n = 2), chemotherapy more than 1 year after transplant induced a completely novel clonal repertoire. Gene ontology analysis of integration loci among early, long-term and dormant clonal populations identified the greatest functional overlap between early and dormant pools. These data suggest that some short-term repopulating clones revert to a dormant phase within the first year after transplant. Additionally, these data indicate that transplant of excess CD34+ cell numbers results in early dormancy of a large proportion of early repopulating clones. Together, these findings suggest that previous estimates of short- and long-term clonal frequency are an underestimate of true graft repopulation potential.