3. Development of a Cas9 Protein Delivery System with Lentiviral Vectors for RNA-Guided Genome Editing

Gene correction is an ideal gene therapy strategy for hereditary disease, including sickle cell disease. Recently, the CRISPR/Cas9 system was developed to allow site-specific DNA breakage, which can enhance homologous recombination-based gene correction with template DNA. However, efficient delivery remains crucial for successful genome editing. In this work, we sought to develop a new gene/protein delivery system with lentiviral vectors for efficient genome editing. We first used an HIV-1 based lentiviral vector system to deliver the CRISPR/Cas9 system including guide RNAs and the Cas9 endonuclease since efficient delivery in various cells, including hematopoietic stem cells, has previously been established. However, using this system, the guide RNA/Cas9-transduced cells continuously express the Cas9 endonuclease, increasing the likelihood of off target effects. In addition, the large size of the Cas9 DNA (4.3kb) limits efficient lentiviral packaging, likely reducing transduction efficiency. Therefore, we sought to develop a Cas9 protein delivery system with lentiviral particles that did not integrate the Cas9 DNA. Cyclosphilin A (CypA) has a function to bind to lentiviral capsids; thus, we hypothesized that fusion proteins between Cas9 and CypA could be packaged in lentiviral particles to allow delivery. We designed two Cas9/CypA fusion proteins: “CypA to Cas9” and “Cas9 to CypA”, and we prepared the lentiviral vectors encoding a GFP-targeting guide RNA with the fusion proteins. We transduced a GFP+ stable cell line with GFP-targeting guide RNA vector containing Cas9 fusion proteins. At 14 days after transduction, GFP-positivity (%GFP) was reduced with both Cas9 protein delivery vectors (48-53%, p<0.01), as compared to a guide RNA alone vector control (83%) and no transduction control (83%). The disruption of GFP was comparable to a guide RNA/Cas9 integrating vector (40%). These data suggest that Cas9/CypA fusion proteins can be delivered with lentiviral particles, and the Cas9 fusion proteins have an endonuclease function to efficiently induce a GFP DNA break. To model DNA correction, we then designed a Cas9 protein delivery non-integrating lentiviral vector encoding both GFP-targeting guide RNA and YFP gene template, which contains all essential components for GFP to YFP gene correction in one vector. Conversion of GFP to YFP would thus model gene correction. Silent mutations in the target site in YFP template were required to produce the gene correction vector. We transduced a GFP+ stable cell line with the gene correction vector using the Cas9/CypA fusion protein or Cas9 protein alone control (without CypA fusion). We observed a significant reduction of GFP positivity (20-24% GFP, p<0.01) along with high rates of conversion to YFP positivity (29-30%, p<0.01) among all Cas9 protein delivery vectors, even with Cas9 protein alone control, as compared to no Cas9 control (YFP 4.9%) and no transduction control (YFP 4.3%). The GFP to YFP gene correction was confirmed by DNA sequencing. These data suggest that Cas9 protein alone can be delivered with lentiviral particles and the Cas9 protein delivery system (with or without CypA fusion) allows for efficient one-time gene correction with non-integrating vectors encoding both guide RNA and template. Additionally, we transferred guide RNA and YFP template again 6 days after Cas9 protein delivery; however, no increase of %YFP was observed, suggesting that Cas9 function was lost over the short term (<6 days). In summary, we developed Cas9 protein delivery system with lentiviral vectors, which resulted in efficient GFP gene breakage. The Css9 protein delivery system allowed for efficient one-time GFP to YFP gene correction with non-integrating lentiviral vector encoding both guide RNA and template DNA. Our findings improve the prospects for efficient delivery and safe genome editing.