MDS1/EVI1 and PRDM16 deregulation in hematopoiesis after experimental and clinical gene transfer

Abstract

Comprehensive integration site analysis in clinical gene therapy has shown that therapeutic vector integrations can lead to clonal selection and even malignant transformation by transcriptional activation of cellular proto-oncogenes. Assessing the clonal dynamics of gene corrected cells revealed an expansion of hematopoiesis marked with unique retroviral insertion sites mainly in the loci of MDS1/EVI1, PRDM16 and SETBP1 in two patients treated for correction of chronic granulomatous disease (CGD) 5 months after gene therapy. This cell population remained stable for additional 1.5 years. At later time points, both individuals developed a myelodysplastic syndrome (MDS) driven by gene-corrected cell clones carrying integrations in the MDS1/EVI1 locus. In this thesis, large scale integration site mapping using non-restrictive and highly sensitive linear amplification mediated polymerase chain reaction (LAM-PCR) focused on the gamma-retroviral integration monitoring within the two gene loci MDS1/EVI1 and PRDM16 in Wiskott - Aldrich syndrome (WAS) clinical gene therapy. Both WAS patients experienced substantial improvement in their clinical conditions, showing a polyclonal repopulation of the hematopoietic system after reinfusion of autologous ex vivo transduced hematopoietic progenitor cells. Although MDS1/EVI1 and PRDM16 were scored as common integration sites (CIS), and MDS1/EVI1 and PRDM16 containing clones were mostly restricted to the myeloid compartment of reconstituted hematopoiesis, no signs of clonal expansion related to these two genes were observed in both patients at 2 years follow up. We hypothesized that a concerted sustained or transient overexpression of MDS1/EVI1 and PRDM16 as pro-myelocytic key regulating transcription factors might be used to substantially influence proliferating and differentiation capacity of hematopoietic stem and progenitor cells. Therefore, murine Mds1/Evi1 and Prdm16 were cloned and expressed either by integration-proficient (LV) or integration-deficient lentiviral vectors (IDLV). Long-term Mds1/Evi1 and Prdm16 expression in hematopoietic progenitors driven by LV was achieved, as well transient expression by IDLV. However, no significant clonal expansion was observed after experimental overexpression of Mds1/Evi1 and Prdm16, as transduced cells became apoptotic after expression. To further assess the biosafety of IDLVs for clinical gene therapy we performed genome wide large-scale IDLV integration analysis in vitro, using conventional and non-restrictive LAM-PCR, as well as newly established two-directional LAM-PCR. This first ever large scale IDLV integration analyses yielded more than 800 unique, mappable IDLV ISs in vitro. Our data revealed a genome-wide close to random integration pattern without any preference for gene coding regions. The results also provided direct molecular evidence that the background integration of D64V mutant IDLVs is not mediated by residual catalytic activity of the mutant integrase. The risk of insertional mutagenesis events mediated by IDLV is highly minimized compared to their integration proficient counterparts and it provides one of the potentially safest tools for efficient gene delivery for clinical applications. The utilization of IDLVs for transient overexpressing is versatile, but warrants careful evaluation of potential toxic effects, as Mds1/EvI1 and Prdm16 as chosen transgenes to expand hematopoietic progenitors have shown.