CRISPR-Cas9-mediated genome editing delivered by a single AAV9 vector inhibits HSV-1 reactivation in a latent rabbit keratitis model

Abstract

Herpes simples virus 1 (HSV-1) keratitis is a major cause of blindness globally. During primary infection, HSV-1 travels to the trigeminal ganglia and establishes lifelong latency. Although some treatments can reduce symptom severity and recurrence, there is no cure for HSV-1 keratitis. We used CRISPR-Cas9 to co-target gene sequences encoding two essential HSV-1 proteins, ICP0 and ICP27, as a potential therapy for HSV-1 keratitis. In HSV-1-infected Vero cells, the HSV-1 viral load and titer were significantly reduced by plasmid transfection or AAV2 vector transduction expressing Cas9 nuclease from Staphylococcus aureus (SaCas9) and paired guide RNAs (gRNAs). Off-target assessment showed minimal off-target editing activity from the selected gRNAs. We then tested our CRISPR-Cas9 gene editing approach in a latent rabbit model of HSV-1 keratitis. Corneal scarification with all-in-one AAV8(Y733F)-SaCas9 or AAV9-SaCas9 vector reduced viral shedding by over 50%. Interestingly, intravenous administration of the same AAV9-SaCas9 vector eliminated viral shedding in 92% of treated eyes. In addition, treated trigeminal ganglia showed a reduction in HSV-1 DNA and RNA expression. Our results support the utility of single-dose AAV9 all-in-one CRISPR-Cas9 gene editing as a safe and effective strategy for treating HSV-1 keratitis.

Keywords: AAV9; CRISPR-Cas9; HSV-1; HSV-1 keratitis; gene editing; latency; rabbit; reactivation; viral shedding.

Graphical abstract

Figure 1

Different combinations of gRNAs targeting ICP0 and ICP27 reduce HSV-1 viral load and titer in Vero cells (A) Effects of transient plasmid delivery of CRISPR-SaCas9 payload on HSV-1 viral load, which was measured by ddPCR of viral DNA probing for UL28 and normalized to the telomerase reverse transcriptase (TERT) gene DNA sequence. (B) Effects of transient plasmid delivery of CRISPR-SaCas9 payload on HSV-1 viral titer, which was measured by plaque assay using supernatants from the same samples as in (A). In (A) and (B), four biological replicates were normalized to their respective internal controls and then plotted. (C) Schematic illustration of the SaCas9-g2g1 expression construct cloned in an AAV plasmid. (D) Effects of AAV2 transduction of CRISPR-SaCas9 payload on HSV-1 viral load. The MOIs of AAV2 vectors are 2E+4, 1E+5, and 5E+5 vg/cell. (E) Effects of AAV2 transduction of CRISPR-SaCas9 payload on HSV-1 viral titer. In (D) and (E), two biological replicates were normalized to their respective internal controls and then plotted. Results are presented as mean + SEM. Statistical significance was determined by two-tailed Student’s t test: ∗∗p < 0.01; ∗∗∗p < 0.001; ∗∗∗∗p < 0.0001. Note some data points are zero and are not shown on the plots with logarithmic scale.

Figure 2

Off-target analysis of ICP0g2 and ICP27g1 (A) GUIDE-seq analysis of ICP0g2. (B) GUIDE-seq analysis of ICP27g1. In (A) and (B), sequences shown are the top 38 recovered sites based on UMI count. Underline indicates the PAM of the SaCas9 nuclease. Dots indicate matches with the cognate spacer sequence. Colored letters indicate mismatches with the cognate spacer sequence. Dashes indicate truncated sequences compared with the cognate spacer sequence. Small font letters indicate the inserted nucleotides compared with the cognate spacer sequence. (C) Quantification of editing activity by SaCas9 programmed with ICP0g2 at the top 16 GUIDE-seq and in silico nominated off-target sites in the HEK293FT reporter cells. (D) Quantification of editing activity by SaCas9 programmed with ICP27g1 at the on-target site and top 6 GUIDE-seq and in silico nominated off-target sites in the HEK293FT reporter cells. Results were obtained from three independent experiments and presented as mean + SEM. Statistical significance is determined by two-tailed Student’s t test: ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001.

Figure 3

AAV-SaCas9-g2g1 vectors delivered by corneal scarification reduce viral shedding in the latent rabbit keratitis model (A) Schematic representation of experimental design and timeline of ocular infection with HSV-1 and treatment with AAV-SaCas9-g2g1 vectors delivered by corneal scarification. (B) HSV-1 virus shedding frequency in tear swabs presented as the percentage of infectious viral positive swabs per eye. Results are presented as individual points with mean. Statistical significance is determined by two-tailed Student’s t test: ∗p < 0.05.

Figure 4

AAV-SaCas9-g2g1 vector delivered by i.v. administration eliminates viral shedding and reduces viral load in the latent rabbit keratitis model (A) HSV-1 virus shedding frequency in tear swabs presented as the percentage of infectious viral positive swabs per eye. The eye IDs are shown in the table on the right. Circles: right eyes (OD), squares: left eyes (OS). The two eyes from the same rabbit are represented with the same color. The right eye of Rabbit-2 from the buffer group did not receive TCIE due to cornea opacity, and the shedding data were not included for this eye. Results are presented as individual points with mean. Statistical significance is determined by two-tailed Student’s t test: ∗∗p < 0.01. (B) HSV-1 viral load in the rabbit TG as measured by dPCR of viral DNA probing for the HSV-1 LAT sequence and normalized to the rabbit TERT DNA. (C) The level of LAT RNA in the rabbit TG as measured by RT-dPCR. The level of LAT RNA was normalized to the level of rabbit HPRT1 mRNA.