^hInGeTox: a human-based in vitro platform to evaluate lentivirus/host interactions that contribute to genotoxicity

Lentivirus vectors are effective for treatment of genetic disease. However, safety associated with vector related genotoxicity is of concern and currently available models are not reliably predictive of safety in humans. We have developed ^hInGeTox as the first human in vitro platform that uses induced pluripotent stem cells and their hepatocyte like cell derivatives to better understand vector-host interactions that relate vectors to their potential genotoxicity. Using lentiviral vectors carrying the eGFP expression cassette under SFFV promoter activity, that only differ by their LTR and SIN configuration, we characterised vector host interactions potentially implicated in genotoxicity. To do this, lentiviral infected cells were subjected to an array of assays and data from these was used for multi-omics analyses of vector effects on cells at early and late harvest time points. Data on the integration sites of lentiviral vectors in cancer genes and differential expression levels of these genes, showed that both vector configurations are capable of activating cancer genes. Through IS tracking in bulk infected cell populations, we also saw an increase in the viral sequence count in cancer genes present over time which were differentially regulated. RNASeq also showed each vector had potential to generate fusion transcripts with the human genome suggestive of gene splicing or vector mediated readthrough from the internal SFFV promoter. Initially, after infection, both vector configurations were associated with differential expression of genes associated cytokine production, however, after culturing over time there were differences in differential expression in cells infected by each LV. This was marked in particular by the expression of genes involved in the response to DNA damage in cells transduced by the SIN vector, suggesting effects likely to prevent tumour development, in contrast to the expression of genes involved in methylation, characteristic of tumour development, in cells transduced by the LTR vector. Both sets of lentiviral infected cells were also found associated with differential expression of MECOM and LMO2 genes known to be associated with clonal dominance, supporting their potential genotoxicity. Alignment of transcriptomic signatures from iPSC and HLC infected cultures with known cancer gene signatures showed the LTR vector with a higher cancer score than the SIN vector over time in iPSC and also in HLC, which further suggests higher genotoxic potential by the LTR configuration lentivirus. By application of ^hInGeTox to cells infected with LV at the pre-clinical stage of development, we hope that ^hInGeTox can act as a useful pre-clinical tool to identify lentivirus-host interactions that may be considered contributory to genotoxicity to improve safer lentiviral vector design for gene therapy.