A temperature-sensitive and less immunogenic Sendai virus for efficient gene editing

Abstract

The therapeutic potential of gene editing technologies hinges on the development of safe and effective delivery methods. In this study, we developed a temperature-sensitive and less immunogenic Sendai virus (ts SeV) as a novel delivery vector for CRISPR-Cas9 and for efficient gene editing in sensitive human cell types with limited induction of an innate immune response. ts SeV demonstrates high transduction efficiency in human CD34⁺ hematopoietic stem and progenitor cells (HSPCs) including transduction of the CD34⁺/CD38^-/CD45RA^-/CD90⁺(Thy1⁺)/CD49f^high stem cell enriched subpopulation. The frequency of CCR5 editing exceeded 90% and bi-allelic CCR5 editing exceeded 70% resulting in significant inhibition of HIV-1 infection in primary human CD14⁺ monocytes. These results demonstrate the potential of the ts SeV platform as a safe, efficient, and flexible addition to the current gene-editing tool delivery methods, which may help further expand the possibilities in personalized medicine and the treatment of genetic disorders.

Importance: Gene editing has the potential to be a powerful tool for the treatment of human diseases including HIV, β-thalassemias, and sickle cell disease. Recent advances have begun to overcome one of the major limiting factors of this technology, namely delivery of the CRISPR-Cas9 gene editing machinery, by utilizing viral vectors. However, gene editing therapies have yet to be implemented due to inherent risks associated with the DNA viral vectors typically used for delivery. As an alternative strategy, we have developed an RNA-based Sendai virus CRISPR-Cas9 delivery vector that does not integrate into the genome, is temperature sensitive, and does not induce a significant host interferon response. This recombinant SeV successfully delivered CRISPR-Cas9 in primary human CD14+ monocytes ex vivo resulting in a high level of CCR5 editing and inhibition of HIV infection.

Keywords: CCR5; CRISPR/Cas9; HIV; Paramyxoviridae; Sendai virus; gene editing; hematopoietic stem and progenitor cells; monocytes; viral vector.

Fig 1

Sendai virus incorporating Cas9 and a guide RNA flanked by self-cleaving ribozymes contains mutations in P and L that impart a temperature-sensitive phenotype. (A) Shown is the Sendai virus genome containing SeV genes N (nucleoprotein), P (phosphoprotein), V, C, M (matrix), F (fusion protein), HN (attachment protein), and L (large RNA-dependent RNA polymerase). An eGFP-P2A-Cas9 cassette (5.1 kb) was inserted between N and P, and a guide RNA flanked by self-cleaving ribozymes (rbz 1 and 2) (0.2 kb total) was inserted between P and M (see Materials and Methods for further details). Mutations were made in both P and L in order to impart a temperature-sensitive phenotype. (B) 293T cells are infected at 32°C for 2 days by either wild type (WT) or temperature sensitive (TS) SeV-Cas9 then shifted to 37°C. Shown is the total GFP measured over 7 weeks, performed in triplicate. The dotted line indicates the limit of detection as determined by the mean of mock-infected cells. (C) As in (B), supernatant is taken and used to infect Vero-CCL81 cells and the titer is calculated in infectious units per mL. Experiment performed in duplicate. (D) 293T cells are infected at 32°C for 2 days then shifted to 37°C (black) until the time point indicated (cyan). Experiment performed in technical triplicate and biological duplicate. (E) Both wt and ts SeV-Cas9 containing a gRNA targeting mCherry are used to infect 293 FLP-mCherry cells. eGFP positivity indicates successful transduction of the SeV-Cas9 and mCherry indicates a lack of indels and subsequent knockout of the mCherry gene. Performed in triplicate. Editing compared using Welch’s t test. (ns, not significant; **, P < 0.01; ***, P < 0.005, and ****, P < 0.0001).

Fig 2

The impact of mutations in P and L on Sendai virus ISG stimulation. (A) and (B) Wt and ts Sendai virus was used to infect 293T cells at 34C for 2 days then shifted to 37°C. (A) RIG-1 transcripts or (B) IFIT1 transcripts were determined by RT-qPCR and normalized to HPRT transcripts. (C) and (D) Wt and ts Sendai virus was used to infect monocyte-derived macrophages from three different donors at 34°C for 24 h and RIG-I (C) or IFIT1 (D) transcripts were determined by RT-qPCR as above. (E) and (F) Sendai virus containing the temperature-sensitive mutations in P only, L only, or wild type was used to infect 293T cells at multiple MOIs. Each MOI was performed in triplicate and relative SeV genomes compared against the fold induction of (E) RIG-I or (F) IFIT1 were measured at 48 hpi. (G) and (H) Both WT and Pmut SeV were used to infect BSR-T7s at five different MOIs and both (G) relative full-length copies and (H) relative DVG copies are measured. (I) The ratio of relative SeV DVG (H) and full-length genome copies (G) for wt SeV and Pmut SeV 3 dpi in BSR-T7 at five different MOIs. (J) 293Tflp cells were infected with wt SeV and Pmut SeV produced in BSR-T7 cells (MOI 0.08–10) at an MOI of 0.2, 1, or 5 for 3 days. The ratio of DVG to full-length genome copies was determined by RT-qPCR. (K) IFIT1 transcripts relative to HRPT transcripts were determined by RT-qPCR in 293Tflp cells infected with wt SeV or Pmut SeV at an MOI of 0.2–5 from viral stocks produced in BSR-T7 (MOI 0.08–10). Error bars indicate SEM, and all comparisons done using Welch’s t test. (ns, not significant; **, P < 0.01; ***, P < 0.005, and ****, P < 0.0001). Additionally, experiments in monocyte-derived macrophages were performed in three different donors.

Fig 3

SeV-Cas9 can deliver a diversity of guides and can utilize novel guide strategies. (A) The gRNA cassette in SeV-Cas9 contains two ribozymes, both necessary for efficient downstream editing. Using RNA structure, we visualize and calculate the probability of proper stem-loop formation required for ribozyme cleaved. (B) We compare the indel frequency measured by Sanger sequencing against the predicted ribozyme efficiency as calculated using information from RNA structure. Significance shows that slope does not equal 0. (C) The gRNA cassette in SeV-Cas9 capable of delivering two separate gRNAs by flanking both with two ribozymes each, separated by a GGS linker. (D) Comparing the single guide systems targeting CCR5 or HPRT and the two-guide system targeting both in 293Ts, H441s, and Huh7 cells. Experiment performed in triplicate and indels calculated via Sanger sequencing and Synthego ICE analysis.

Fig 4

Efficient CD34⁺ HSPC transduction by the ts rSeV-Cas9-CCR5. (A) Representative flow cytometry data of human fetal liver and G-CSF peripheral blood mobilized CD34⁺ HSPC infected with ts rSeV-Cas9-CCR5 at MOI 5 and 10 at 34°C. Flow cytometry showed >90% transduction (EGFP⁺) relative to mock-infected cells at 3 dpi. (B) Efficient transduction in the rare human CD34⁺/CD38⁻/Thy1⁺/CD49f ^high HSC enriched subpopulation. Percent eGFP (95.9%, green histogram) was determined relative to mock (gray histogram)-infected cells. (C) and (D) CD34⁺ HSPC transduction by ts rSeV-Cas9-CCR5 yielded >90% transduction across all MOIs greater than 1 tested, both in (C) fetal liver (FL) CD34⁺ HSPCs and (D) mobilized peripheral blood (mPB) CD34⁺ HSPCs. Transduction efficiency in these experiments likely reflects additional infection from virus replicating at permissive temperature.

Fig 5

Editing efficiency in CD34⁺ HSPCs and the effect on hematopoietic differentiation. (A) Fetal liver (FL) CD34⁺ HSPCs or mobilized peripheral blood (mPB) CD34⁺ HSPCs were infected at multiple MOIs, with percent indels calculated via Synthego ICE analysis. (B) Downstream colony differentiation after ts SeV-Cas9 transduced mPB or FL CD34⁺ HSPCs at an MOI of 10. We measure CFU-E: CFU erythroid; CFU-G; CFU granulocytes; CFU-GM: CFU granulocytes and macrophages; CFU-GEMM: CFU granulocyte, erythrocyte, monocyte, megakaryocyte; BFU-E: Burst-forming unit-erythroid, CFU-M: Colony Forming Unit-monocytes. Error bars are SD. (C) Raw colony counts, performed as described in (B). (D) Mobilized peripheral blood CD34⁺ HSPCs and (E) fetal liver CD34⁺ HSPCs are infected by ts SeV Cas9 containing a guide targeting CCR5 or mCherry at an MOI of 10. Percent indels calculated via Illumina sequencing (see Materials and Methods). Top five off-target sites predicted by the CRISPR design tool (crispr.mit.edu).

Fig 6

CCR5 editing of primary CD14⁺ monocytes with ts-SeV-Cas9 limits infection with HIV. (A) ts SeV-Cas9 mediated editing efficiency of CCR5 in CD14⁺ monocytes was determined at 10 and 18 dpi with ts-SeV-Cas9. Cells infected with a ts-SeV-Cas9 mCherry targeting virus were used as a negative control. Two samples per condition were measured. (B) Percentage of HIV infected (mCherry positive) CD14⁺ primary human macrophages at 18 days post-infection with 200 pg/mL HIV. Prior to HIV infection, macrophages were mock treated or infected with the indicated ts SeV-Cas9 at an MOI of 10 at 34°C for 3 days before being shifted to 37°C for 7 days. (C) An HIV growth curve of ts-SeV-Cas9-infected MDMs measuring the accumulation of P24 in the supernatant. Samples were collected at 4, 10, 16, and 19 dpi and the cumulative level of P24 in each sample was calculated.