Vector insertional mutagenesis drives accelerated hematopoietic stem cell aging and acquisition of somatic mutations in vivo

Hematopoietic Stem and Progenitor Cell Gene Therapy (HSPC-GT) uses retroviral vectors such as self-inactivating (SIN) lentiviral vectors (LVs) to insert therapeutic genes into target genomes. Once reinfused, these modified HSPCs fully reconstitute hematopoiesis, maintain lifelong blood cell homeostasis, and provide therapeutic benefits. However, vector integration may inadvertently activate oncogenes, potentially causing malignancies, as seen in an LV-based trial for adrenoleukodystrophy. Moreover, ectopic oncogene activation in HSPCs induces oncogene-induced senescence (OIS), leading to proliferation arrest and the secretion of pro-inflammatory cytokines, a process known as the senescence-associated secretory phenotype (SASP). Transplantation of oncogene-expressing HSPCs in a xenogeneic model resulted in senescent cell accumulation and chronic inflammation, which triggered myeloid-restricted hematopoiesis and eventually led to bone marrow failure. These findings raise concerns that insertional mutagenesis in HSPC-GT could produce similar adverse effects. To address this issue, we monitored the hematopoiesis of wild-type (WT) mice transplanted with WT HSPCs transduced with either a genotoxic LV carrying the Spleen Focus Forming Virus promoter in its long terminal repeats (LV.SF) or a safer SIN LV with the PGK promoter in an internal position (SINLV.PGK). We applied the same approach to HSPCs from Cdkn2a-/- mice, which lack the key senescence cell cycle inhibitors p16INK4A and p19ARF. Blood composition was monitored every four weeks (up to 2.5 years), and vector integration sites (IS) were retrieved from purified B, T, and myeloid cells for clonal tracking. The genotoxic LV.SF accelerated myeloid tumor onset in mice transplanted with Cdkn2a-/- cells (p < 0.0001) but not in WT mice. However, WT/LV.SF mice exhibited marked myeloid skewing and a dramatic reduction in clonality across all hematopoietic lineages over time, suggesting accelerated aging compared to WT SINLV.PGK mice. Insertions activating the Braf oncogene were prevalent in tumor-infiltrated tissues of the Cdkn2a-/- group and, to a lesser extent, in WT mice. Given the link between OIS, chronic inflammation, and somatic mutations, we analyzed genomic sequences flanking each IS to assess mutation accumulation using a novel Mutation Index (MI), normalized by clone number and coverage. Analysis of over 200,000 IS covering more than 9 Gb of genomic data revealed significantly higher MI in both LV.SF groups than in LV.PGK groups (p < 0.001), with myeloid clones accumulating more mutations than B and T cells. This effect was exacerbated in Cdkn2a/LV.SF mice, indicating that the absence of proto-oncogene activation barriers leads to increased somatic mutations. Whole-genome sequencing confirmed a higher mutation rate in LV.SF groups compared to LV.PGK, validating that no biases were present in the flanking region analysis. Finally, a Bayesian model developed to trace the dynamics of clonal and subclonal evolution based on IS and mutations, confirmed greater mutational diversity and complexity, increased mutation burden and accelerated clonal expansion in LV.SF groups. Overall, our data show that random insertional activation of oncogenes in vivo induces accelerated stem cell aging, clonal hematopoiesis, and heightened somatic mutation burden due to persistent senescent cell accumulation in tissues and justifies further exploration of this underappreciated aspect of vector genotoxicity.