Abstract Humanized mouse models are essential for evaluating the engraftment capacity and genetic integrity of gene-modified hematopoietic stem and progenitor cells (HSPCs). Here, we compared two widely used xenotransplantation platforms, NSG and NBSGW mice, in the context of lentiviral vector (LVV) transduction and CRISPR/Cas9-mediated gene correction. HSPCs harboring high LVV copy numbers exhibited engraftment deficits in NSG mice that were not observed in NBSGW mice. This discrepancy highlights the potential for the NBSGW model to mask safety liabilities of LVV-modified products due to its higher overall levels of human chimerism. In contrast, CRISPR/Cas9 editing with a single-stranded oligodeoxynucleotide donor yielded comparable correction rates in both models, even across decreasing input cell doses, demonstrating that long-term repopulating hematopoietic stem cells (HSCs) retain equivalent engraftment capacity in each strain. Single-cell RNA-sequencing revealed distinct progenitor populations that were markedly under-represented in the NSG model but preserved in NBSGW recipients, emphasizing the greater capacity of NBSGW mice to better support multilineage human hematopoiesis. Together, these findings establish that both NSG and NBSGW mice are suitable for assessing long-term engraftment and gene modification outcomes in human HSPCs. However, the significantly higher percentage of human cell chimerism in the NBSGW model may obscure cell populations with engraftment deficits. Careful selection of in vivo models is therefore critical for rigorous preclinical evaluation of gene therapy products prior to clinical translation.
Kuo et al. (Wed,) studied this question.