Gene Editing with a Novel Binary Retroviral Vector System to Reduce HLA Class I Allogenicity

Background: HLA polymorphisms restrict availability of suitable donors, complicate logistics of procuring and transporting matched organs, and necessitate life-long immunosuppression.  We previously showed that lentiviral vectors delivering short hairpin RNAs can knock down HLA expression in established human cell lines and primary cells, allowing evasion from immune recognition by alloreactive T cells. Now, gene editing with Clustered Regularly Interspaced Short Palindromic Repeats (CRISPR) can be used to permanently knock out HLA genes and reduce allogenicity. Here we describe a novel binary retroviral vector system, combining a retroviral replicating vector (RRV) which actively spreads through dividing cells but has limited payload capacity, with a retroviral non-replicating vector (RNV) which is incapable of cell-to-cell spread but has high transgene capacity, to enable co-delivery of CRISPR/Cas9 for gene editing to reduce HLA expression.
Materials & methods: RNV was engineered to co-express the Gal4/VP16 synthetic transcription factor, CRISPR-associated gene editing enzyme Cas9, and Green Fluorescent Protein (GFP). RRV was engineered to co-express β2-microglobulin (β2M)-targeted single-guide RNA (sgRNA) and Red Fluorescent Protein (RFP), and its promoter sequence was replaced with GAL4 binding sites. After initial vector transduction at multiplicity of infection (MOI) 0.01 in established and primary human cell lines, flow cytometry was used to evaluate transduction levels by GFP and RFP expression, and to evaluate loss of cell surface HLA expression by anti-β2M antibody staining.
Results: Initial co-transduction levels were confirmed to be ~1% as expected based on MOI. Subsequently, increasing GFP(+) / RFP(+) dual fluorescence was observed over time both in vitro and in vivo, confirming co-replication of both RRV and RNV and concomitant transgene expression from both vectors. Gene editing by CRISPR/Cas9 was also confirmed by flow cytometry, with >50% reduction of β2M observed in co-transduced cells.
Conclusions: Gene editing to knock out β2M, which is generally required for cell surface expression of Class I HLA, can be achieved by co-transduction with binary RRV/RNV even after initially low transduction levels. Previously, we showed that shRNA targeting the Class I α chain could reduce HLA expression in an allele-specific manner, and gene editing can achieve comparable specificity with the use of allele-targeted CRISPR sgRNA. While efficient gene transfer to entire solid organs remains a technical hurdle, application of this strategy can be envisioned for ex vivo transduction of cellular transplants such as hematopoietic stem/progenitor cells and pancreatic islet cells.  By engineering the allograft rather than immunosuppressing the recipient, this approach represents a fundamental shift in the concept of achieving histocompatibility.