Gene editing for latent herpes simplex virus infection reduces viral load and shedding in vivo – Nature.com

Posted: May 18, 2024 at 2:42 am

Meganuclease therapy reduces ganglionic viral load after ocular or genital HSV infection

We previously evaluated several AAV serotypes for delivery of meganucleases to latently infected mice, and found the best results with AAV-Rh10, followed by AAV8 and AAV18. To further improve efficacy, we tested additional neurotropic AAV serotypes, including AAV7, AAV9, AAV-DJ, and AAV-DJ/810,11, for delivery of the anti-HSV1 meganuclease, m5, at a dose of 1012 AAV genomes (vg) per mouse (Fig.S1a), using our model of orofacial HSV disease. Both AAV9 and AAV-Dj/8 were superior to 1012 vg AAV-Rh10, the best of our previously used serotypes8, showing HSV reductions in superior cervical ganglia (SCG) of 95% (p=105) and 90% (p=0.018), respectively, relative to untreated controls (Fig.S1b), comparing favorably with the 65% reduction previously obtained with m5 alone delivered with AAV-Rh10 8. Similarly, AAV9 and AAV-Dj/8 showed better activity than AAV-Rh10 in trigeminal ganglia (TG), with HSV load reductions of 48% (p=0.07) and 41% (p=0.5), respectively (Fig.S1b), compared with our prior observation of no detectable reduction using AAV-Rh10 delivering m58. The route of AAV administration (retro-orbital vein vs. intradermally into the whisker pad) did not have any detectable impact on either AAV transduction or gene editing efficiencies (Fig.S1eg).

We previously demonstrated that gene editing of HSV could be increased by using combinations of AAV serotypes for meganuclease delivery, rather than a single AAV serotype, a finding we ascribed to the heterogeneity of neuronal subsets within HSV-infected ganglia8. We therefore evaluated gene editing with the anti-HSV1 meganuclease m5, which cleaves a sequence in the UL19 gene coding for the major capsid protein VP512, when delivered using single AAV serotypes vs. combinations of AAV9, AAV-Dj/8, and AAV-Rh10 (Fig.S2a) administered as a total dose of 1012 vg per mouse. In agreement with our previous results, combinations of AAV serotypes led to robust HSV gene editing, with the triple combination of AAV9, AAV-Dj/8, and AAV-Rh10 showing especially strong reductions in HSV loads and mutagenesis of residual HSV across both SCG and TG (Fig.S2bg).

While orofacial infections with HSV are extremely common, genital infections, which lead to latent infection of dorsal root ganglia (DRG), also represent a major cause of morbidity. We therefore established latent genital infections in mice by intravaginal inoculation with HSV-1 after treatment with Depo Provera, which synchronizes the estrus cycle and increases HSV infection13. Infected mice were treated with a total dose of 31012 vg of the AAV9, AAV-Dj/8, and AAV-Rh10 combination delivering two HSV1-specific meganucleases simultaneously (m5 along with m8, which targets a sequence in the UL30 gene coding for the catalytic subunit of the viral DNA polymerase12. In parallel, we tested the same AAV combination against latent orofacial HSV infection as described above (Fig.1a, b). Remarkably, efficacy in the vaginal model of infection was the highest we have observed to date, with a 97.7% reduction in HSV viral load in DRG (Fig.1c). This compared favorably with the orofacial infection group treated in parallel, in which (in agreement with our previous studies) we observed robust gene editing with significant reductions of ganglionic HSV loads of 89% in SCG and 61% in TG (Fig.1d, e).

a, b Experimental timeline of (a) vaginal or (b) ocular infection and meganuclease therapy. RO, retroorbital; TV, tail vein. c HSV loads in DRGs from control (n=7) and dual meganuclease-treated (n=4) mice vaginally infected with HSV-1; p=0.001. d. HSV loads in SCGs and TGs from control (n=10) and dual meganuclease-treated (n=10) mice ocularly infected with HSV-1; p=0.0046 and 0.0034 for SCG and TG, respectively. e Gene editing at the m5 target site of residual virus quantified by T7E1 assay in SCG and TG from dual meganuclease-treated mice (n=10). Each graph shows individual and mean values with standard deviation, percent decrease of HSV loads in treated mice compared to control mice and statistical analysis (unpaired one-tailed Mann-Whitney test with **p<0.01). AAV loads are shown in Supplemental Fig.9a, b. Source data are provided as a Source Data file.

Mice generally show little if any spontaneous HSV reactivation, with minimal to no viral shedding at peripheral sites, limiting their utility in cure studies. The BET (Bromo and Extra-Terminal domain) bromodomain inhibitor JQ1 was reported to reactivate latent HSV in vitro in primary neuronal cultures, and HSV could be detected in the eyes of about one-third of latently infected mice treated with JQ114. To evaluate the utility of JQ1 for our cure work, we extended these studies to determine the quantitative kinetics of HSV shedding after JQ1 therapy.

A single intraperitoneal (IP) injection of JQ1 (50mg/kg) given to latently infected mice (Fig.S3a) led to detectable shedding from the eyes of 56% (5/9) of animals, compared with 0/9 animals treated with vehicle alone (Fig.S3b, c). Viral shedding was transient, peaking at 2 days post JQ1, with maximal viral loads ranging from about 102 to almost 106 copies/swab (Fig.S3c). A direct comparison suggested that JQ1 may be a more powerful reactivation stimulus for HSV than hyperthermic stress (HS)15, which in our hands led to detectable virus shedding in less than 20% of animals (2/12 HS vs 4/10 JQ1), with peak shedding viral loads two logs lower than after JQ1 treatment (Fig.S3df). Sequential treatment with JQ1 at one-week intervals led to repeated shedding episodes with similar kinetics as observed above (Fig.S4ae). Over the course of three sequential JQ1 reactivations (Fig.S4e), shedding from individual mice was stochastic; 10/12 (83%) mice shed detectable virus at least once, but only 1/12 (8%) shed after all three treatments, while 4/12 (33%) and 5/12 (42%) shed only after two or one of the three treatments, respectively (Fig.S4e and TableS1). Shedding was typically unilateral (only detected in one eye), despite the initial inoculation being to both eyes (unilateral shedding was observed in 33/37 (89%) of events). The side of shedding in one episode was not predictive of the side of future shedding events (TableS1). Importantly for cure studies, repeated weekly reactivation of virus with JQ1 up to 7 weekly injections did not change ganglionic viral loads compared with control animals (Fig.S4g, h).

The ability to reproducibly induce HSV reactivation and shedding with JQ1 allowed us to investigate the relationship between ganglionic viral load reduction using meganucleases and subsequent viral shedding at peripheral sites. Latently infected mice were treated as above using the AAV9, AAV-Dj/8, and AAV-Rh10 combination delivering HSV1-specific meganucleases m5 and m8 at a total dose of 31012 vg, or left untreated as controls. One month later, mice were administered JQ1, and eye swabs were collected daily for 4 days (Fig.2a). Consistent with our previous results, ganglionic tissues from treated mice showed a 98% and 42% reduction in mean viral loads in SCG and TG, respectively, when compared to control untreated animals (Fig.2b, c) and gene editing in the remaining viral genomes (Fig.2d). After JQ1 administration, only 3/10 (30%) of the dual meganuclease-treated mice had detectable virus in eye swabs, compared with 5/10 (50%) of control untreated animals (Fig.2e, f). The mean titer of HSV in positive eye swabs was 3104 copies/ml in the meganuclease-treated animals, compared with 1.2105 copies/ml in the control animals. Area under the curve analysis (AUC) demonstrated a 95% reduction (p=0.15) in total viral shedding in treated vs. control animals (Fig.2g). In a separate experiment performed similarly (Fig.3a), ganglionic tissues from treated mice showed 97% reduction in mean latent HSV genomes in both SCG and TG when compared to control untreated animals (Fig.3b, c) and gene editing in the remaining genomes (Fig.3d). While 3/8 mice from the control group shed virus with a mean viral titer of 8.2105 copies/ml, 0/8 meganuclease-treated mice had detectable shedding, representing a 100% decrease in total virus shed (p=0.10, Fig.3eg).

Experimental timeline of ocular infection, meganuclease treatment and viral reactivations with JQ1. a Experiment 1 (n=10 per group). HSV loads in SCGs (b: p=0.0057) and TGs (c). Percent decrease of HSV loads in treated mice compared to control mice and statistical analysis (unpaired one-tailed MannWhitney test with ns: not significant, **p<0.01) are indicated. Gene editing at the m5 target site of residual virus quantified by T7E1 assay in SCG and TG from dual meganuclease-treated mice (d). HSV titers in eye swabs collected daily from day 1 to 4 post JQ1 reactivation from control (e) and dual meganuclease-treated (f) infected mice. Panels 2i-k show data for both SCG and both TG from each mouse. Area under the curve (AUC) analysis (g) with p value (unpaired one-tailed MannWhitney test). AAV loads are shown in Supplemental Fig.9c, d. Each graph shows individual and mean values with standard deviation. Source data are provided as a Source Data file.

Experimental timeline of ocular infection, meganuclease treatment and viral reactivations with JQ1. a Experiment 2 (n=8 per group). HSV loads in SCGs (b) and TGs (c). Percent decrease of HSV loads in treated mice compared to control mice and statistical analysis (unpaired one-tailed Mann-Whitney test with ns: not significant, **p<0.01) are indicated. Gene editing at the m5 target site of residual virus quantified by T7E1 assay in SCG and TG from dual meganuclease-treated mice (d). HSV titers in eye swabs collected daily from day 1 to 4 post JQ1 reactivation from control (e) and dual meganuclease-treated (f) infected mice. Panels 2b-d show data for both SCG and both TG from each mouse. Area under the curve (AUC) analysis (g) with p value (unpaired one-tailed MannWhitney test). AAV loads are shown in Supplemental Fig.9e, f. Each graph shows individual and mean values with standard deviation. Source data are provided as a Source Data file.

AAV-vectored therapies are generally considered safe. Nevertheless, dose-limiting liver toxicity has been observed after AAV administration in humans, non-human primates, and mice, typically at doses of 21014 vg/kg or greater. The 31012 vg/animal dose (~11014 vg/kg) used in the experiments described in Figs.13 approached the level associated with liver toxicity in previous studies. Across multiple studies we observed that 7/70 (10%) animals treated with the 31012 vg/animal dose exhibited clinical signs consistent with hepatotoxicity, including weight change, bloating, and general health decline. Hepatotoxicity was confirmed in these animals by subsequent histopathological evaluation (Fig.S5 and TableS2). We therefore evaluated lower total doses of triple AAV serotype/dual meganuclease therapy (0.6, 1.2, or 1.81012 vg/animal or 1.8, 3.6, or 5.41013 vg/kg) for their tolerability and effects on viral load and JQ1-induced HSV shedding (Fig.4a). These doses showed substantially improved tolerability, both clinically and upon histopathological examination and quantification of the number of inflammatory cell foci (ICF) in livers (Fig.S6a). Dose-dependent reductions in ganglionic HSV loads were observed across the three treatment groups compared to controls, ranging from 69% and 47% in SCG and TG, respectively, at the 0.61012 dose to 94% and 73% at the 1.81012 dose (Fig.4b, c). To evaluate the effect of these reduced doses on HSV shedding, treated mice were subjected to three weekly rounds of JQ1 administration, followed by eye swabbing as described above. While the percentage of dual meganuclease-treated animals shedding virus after the first JQ1 reactivation was not reduced compared with the control mice, it was substantially lower than controls at all doses by the third JQ1 reactivation (0% (0/12), 8% (1/12) and 0% (0/12) for 0.6, 1.2, and 1.81012 vg/mouse groups, respectively, versus 18% (2/11) in the control group) (Fig.4fi). This finding that may relate to the two additional weeks available for meganuclease expression and gene editing activity by the time of the third JQ1 reactivation. Consistent with this interpretation, the reduction in total viral shedding, as determined by AUC analysis, appeared to become more complete over time, with up to a 97100% reduction in all three groups by the final JQ1 reactivation (Fig.4jl). The efficacy of reduced-dose dual meganuclease therapy (1.81012 vg) was confirmed in a separate experiment (Fig.5a), showing a significant decrease in ganglionic viral loads in both SCG and TG (Fig.5b, c). In this experiment, 7 of 12 control animals showed detectable viral shedding after JQ1 reactivation, compared with only 1 of 12 animals treated with AAV-meganuclease therapy, (Fig.5d, e) and reduction in total viral shedding, as determined by AUC analysis (Fig.5f, g). While none of the treated mice exhibited any clinical signs of hepatotoxicity, we did observe higher numbers of ICF in liver of treated animals receiving the 1.81012 vg dose compared to control mice (Fig.S6b). Histologic analysis of H&E stained TG sections from both control and treated animals revealed subtle evidence of neuronal injury, manifesting as neuronal degeneration, necrosis, and axonopathy. The scores grading prevalence and severity of the microscopic changes were higher in treated animals compared to control mice (Fig.S7 and TableS3). However, no mice in either the control or experimental group showed detectable signs of neuropathy.

a Experimental timeline of ocular infection, meganuclease treatment and viral reactivations with JQ1. b, c HSV loads and d, e, AAV loads in SCGs (b; p=0.0016, 0.0012, and <0.0001 for 0.6, 1.2, and 1.81012, respectively, d) and TGs (c; p=0.068, 0.0025 and 0.0016 for 0.6, 1.2, and 1.81012, respectively, e) in control infected mice (n=11) and infected mice treated with dual therapy delivered with 0.6 (n=12), 1.2 (n=12), or 1.8 (n=12)1012 total vg AAV dose. Percent decrease of HSV loads in treated mice compared to control mice and statistical analysis (ordinary one-way Anova, multiple comparisons with ns: not significant, **p<0.01, ****p<0.0001) are indicated. fi Virus titers in eye swabs collected daily from day 1 to 4 after each weekly JQ1 reactivation from control infected mice (f) and infected mice treated with dual therapy delivered with 0.6 (g), 1.2 (h), or 1.8 (i) 1012 total vg AAV dose. jl Area under the curve (AUC) analysis of virus shedding after first (j), second (k), and third (l) JQ1 reactivation from control infected mice (n=11) and infected mice treated with dual therapy delivered with 0.6 (n=12), 1.2 (n=12) or 1.8 (n=12)1012 total vg AAV dose. p values (unpaired, ordinary one-way Anova, with multiple comparisons) compared virus shedding between treatment groups and the control group. Each graph shows individual and mean values with standard deviation. Source data are provided as a Source Data file.

a Experimental timeline of ocular infection, meganuclease treatment and viral reactivations with JQ1. b HSV loads in SCGs (b; p=0.0012, and <0.0001 for CTRL vs AAV/MN no JQ1 and for CTRL vs AAV/MN 2x JQ1, respectively), and TGs (c; p=0.0089, and 0.0293 for CTRL vs AAV/MN no JQ1 and for CTRL vs AAV/MN 2x JQ1, respectively) of control infected mice either unreactivated (CTRL no JQ1) or reactivated (CTRL 2x JQ1) and infected mice treated with dual therapy delivered with 1.81012 total AAV dose either unreactivated (AAV/MN no JQ1) or reactivated (AAV/MN 2x JQ1), with n=12 per group. Percent decrease of HSV loads in treated mice compared to control mice and statistical analysis (unpaired one-tailed Mann-Whitney test with *p<0.05; **p<0.01, ****p<0.0001) are indicated. d, e Virus titers in eye swabs collected daily from day 1 to 4 after two weekly JQ1 reactivations (red arrows) from control infected mice (d) and infected mice treated with dual therapy delivered with 1.81012 total AAV dose (e). f, g Area under the curve (AUC) analysis after the first (f), and the second (g) JQ1 reactivation from control infected mice either unreactivated (CTRL no JQ1) or reactivated (CTRL 2x JQ1) and infected mice treated with dual therapy delivered with 1.81012 total AAV dose either unreactivated (AAV/MN no JQ1) or reactivated (AAV/MN 2x JQ1), with n=12 per group. p values (unpaired one-tailed MannWhitney test) are indicated. Each graph shows individual and mean values with standard deviation. AAV loads are shown in Supplemental Fig.9gh. Source data are provided as a Source Data file.

As noted above, genital HSV infection is a major cause of morbidity in humans. We therefore evaluated the reduced-dose dual meganuclease therapy (total dose of 1.81012 vg/animal) in vaginally infected mice (Fig.6a). In agreement with our previous results, the reduced-dose therapy led to a 78.8% (p=0.02) to 95.6% (p=0.006) reduction in latent virus genomes in DRGs (Fig.6b).

a Experimental timeline of intravaginal HSV-1 infection, meganuclease treatment and viral reactivations with JQ1. b HSV loads in DRGs from control infected mice reactivated with 3 weekly JQ1 injections and infected mice treated with dual therapy unreactivated, or reactivated with 3 weekly JQ1 injections with n=8 per group; p=0.0055, and 0.0198 (ordinary one-way Anova, multiple comparisons) for CTRL+JQ1 vs AAV/MN no JQ1 and for CTRL+JQ1 vs AAV/MN+JQ1, respectively. c, d HSV titers in vaginal swabs collected daily from day 1 to 4 post JQ1 injections (red arrows) from control (c) and dual meganuclease-treated (d) infected mice. Area under the curve (AUC) analysis after the first (e), second (f), and third (g) JQ1 reactivation from control infected mice (n=8) and infected mice treated with dual therapy, both reactivated with 3 weekly JQ1 injections (n=8). p values (unpaired one-tailed MannWhitney test) are indicated. Each graph shows individual and mean values with standard deviation. The AAV viral loads are shown in Supplemental Fig.9i. Source data are provided as a Source Data file.

We then sought to evaluate whether JQ1 could induce HSV shedding in the genital infection model, as we previously observed in the ocular infection model. Over 3 sequential JQ1 reactivations, only 2 of 8 control animals (and 1 of 8 AAV/meganuclease-treated animals) shed detectable virus, a rate lower than the 4050% reactivation we typically observe after ocular infection (Fig.6c, d). The apparently lower rate of reactivation seen in the vaginal model compared to the ocular model may be due to lower levels of ganglionic HSV loads in the DRG (102103 vg/106 cells in DRG, Fig.6b vs 104105 vg/106 cells in SCG or TG, Fig.5b, c). While this lower reactivation rate prohibited meaningful statistical analysis, the observation that 2 out of the 8 control mice shed virus over 2 to 3 sequential days, while only 1 of the 8 AAV-treated mice shed virus, on a single day and at a substantially lower level, is qualitatively in agreement with our observations after ocular infection (Fig.6cg).

The stochastic nature of clinical HSV reactivation16, recapitulated when induced by JQ1 in mice (Fig.S4 and TableS1), makes evaluation of viral shedding extremely resource-intensive. Practical constraints, including the number of animals that can be housed and studied at the same time, along with the extended duration of each study (~3 months), limited the statistical power of our individual experiments. We therefore performed a meta-analysis of data from all experiments presented above (Figs.16), combining evidence from infection sites (orofacial or genital), thus comparing 174 swabs from AAV/meganuclease-treated animals to 99 swabs from experimentally-matched controls. The primary endpoint was viral shedding, expressed either as a binary variable (equal to 1 for samples in which HSV was detected and 0 otherwise) or the log10-transformed AUC for quantitative viral shedding. The experiments depicted in Figs.16 represent all of the shedding studies with the dual meganuclease/triple AAV therapy we have performed as of this writing, and each suggests a strong and consistent trend toward a substantial reduction in viral shedding after AAV/meganuclease therapy. Across all studies, the proportion of swabs with detectable HSV was 48% lower among AAV/meganuclease-treated animals compared to controls. The meta-analysis confirmed that animals receiving AAV/meganuclease therapy had a statistically significant reduction in viral shedding (OR=0.41, p=0.010, by generalized linear mixed models, GLMM).

We then asked whether dose or duration of meganuclease therapy was associated with the probability of viral shedding (expressed as a binary variable) or the quantity of viral shedding (expressed as the log10-transformed AUC). Overall, the probability of viral shedding significantly decreased with the dose of AAV/meganuclease (OR=0.66; p=0.023, GLMM), and also with the duration of meganuclease therapy (OR=0.42; p<0.001, GLMM) in treated animals compared to controls. The data further indicate that overall, the quantity of virus shed (AUC) significantly decreased with the AAV/meganuclease dose at a rate of -0.36 log10 copy-days per 1012 increase in dose (LMM; p=0.028), and also with the duration of meganuclease therapy, at a rate of 0.48 log10 copy-days per additional week after treatment (LMM; p=0.017). No significant association was detected between the log10-transformed AUC and the interaction between time and dose (LMM; p=0.59).

The studies described above were performed using a triple AAV serotype/dual meganuclease approach, resulting in each animal receiving a total of 6 unique vectors (3 serotypes2 meganucleases). Clinical translation of such a complex regimen could raise manufacturing and quality control issues. We therefore sought to simplify AAV/meganuclease therapy to reduce the complexity of our therapeutic regimen. We took advantage of the dual cutting meganuclease m4, which recognizes a sequence in the duplicated gene ICP0 in the HSV-1 genome and was previously shown to induce significant decrease of latent viral loads in ganglia of latently infected mice8. Latently infected mice were administered a total dose of 51011 vg of either the combination of AAV9, AAV-Dj/8, and AAV-Rh10, or each single AAV serotype delivering the HSV1-specific meganuclease m4 (Fig.7a). Consistent with the results using the lower dose of 6 x 1011 of the triple AAV-dual MN therapy (Fig.4), ganglionic tissues from treated mice with the triple AAV combination delivering m4 showed a 73.9% (p<0.0001) and 43.7% (p=0.014) reduction in mean viral loads in SCG and TG, respectively, when compared to control untreated animals. When m4 was delivered using single AAV serotypes, the data confirmed that AAV9 on its own could recapitulate the viral load decrease seen with the triple AAV serotype combination, with 77.8% (p<0.0001) and 49% (p=0.0046) reduction in mean viral loads in SCG and TG, respectively (Fig.7b, c). Furthermore, mice having received AAV9 alone showed the lowest levels of liver inflammation of any of the groups, similar to those in the control liver (Fig.7d). At this reduced dose, regardless of the AAV serotype combination used, no detectable neurotoxicity was observed compared to the control animals (Fig.7e, f).

a Experimental timeline of ocular infection and meganuclease therapy. b, c HSV loads in SCGs (b; p<0.0001) and TGs (c; p=0.0046 for AAV9 and 0.0142 for 9-Dj/8-Rh10) from infected control and infected mice treated with m4 delivered by retro-orbital (RO)) injections of 51011 vg total of the single or triple combinations of AAV9, -Dj/8 and -Rh10. Percent decrease of HSV loads in treated mice (n=10 per group) compared to control mice (n=10) and statistical analysis (Ordinary one-way Anova, multiple comparisons with *p<0.05; **p<0.01, ****p<0.0001; ns: not significant). d Inflammatory cell foci (ICF) in liver sections from either HSV-infected control mice (n=10), or mice treated with m4 delivered using AAV single or triple combinations of AAV9, -Dj/8, and -Rh10 (n=10 per group); p=0.0009 for Rh10. e, f Severity scores of axonopathy (e) and inflammation (f) in TG from infected control mice (n=3 TG) and infected mice treated with m4 delivered using single or triple combinations of AAV9, -Dj/8 and -Rh10 (n=3 TG per group) and statistical analysis (Ordinary one-way Anova, multiple comparisons with ns: not significant; ***p<0.001). Each graph shows individual and mean values with standard deviation. The AAV viral loads are shown in Supplemental Fig.9jl. Source data are provided as a Source Data file.

To confirm that a simplified regimen composed of AAV9-m4 was also able to reduce peripheral virus shedding, latently infected mice were treated as above using AAV9 delivering either HSV1-specific meganuclease m4 or a catalytically inactive version (m4i) at a dose of 11012 vg. One month later, mice were subjected to two weekly rounds of JQ1 administration, followed by daily eye swabbing for 3 days (Fig.8a, b). A decrease of ganglionic viral loads of 89.6% (p<0.0001) and 69% (p=0.03) in SCG and TG respectively, was observed in m4-treated mice but not in mice treated with the inactive form of the meganuclease m4i (Fig.8c, d). Furthermore, 6 out of 9 control mice and 6 out of 10 m4i-treated mice shed virus after JQ1 reactivations, while only 3 out of 10 m4-treated mice had detectable virus shedding after reactivation (Fig.8ei). These data demonstrate that our simplified regimen can substantially reduce ganglionic viral loads, with an associated decrease in virus shedding after reactivation, and that these effects are dependent on an active enzyme and not on AAV itself. In this experiment, mice treated with m4 had slightly higher levels of liver ICF and TG axonopathy, but not more TG inflammation, compared to control mice (Fig.8jl).

a Experimental timeline of ocular HSV-1 infection, meganuclease treatment and viral reactivations with JQ1. b Schematic of active m4 and inactive m4i meganuclease. c-d, HSV loads in both SCGs (c; p<0.0001 for m4) and both TGs (d; p=0.003 for m4) from control infected mice and infected mice treated the active m4 or inactive m4i (n=10 per group). Percent decrease of HSV loads in treated mice compared to control mice and statistical analysis (unpaired one-tailed Mann-Whitney test with *p<0.05; ****p<0.0001; ns: not significant) are indicated. eg Virus titers in eye swabs collected at day 1 to 3 after each JQ1 reactivation from control infected mice (e) and infected mice treated with active m4 (f) or inactive m4i (g). h, i Area under the curve (AUC) analysis of virus shedding after first (h), and second (i) JQ1 reactivation from control infected mice and infected mice treated with active m4 or inactive m4i (n=10 per group). p values (unpaired one-tailed MannWhitney test) are indicated. j Inflammatory cell foci (ICF) in liver sections from either HSV-infected control mice, mice treated with active m4 or inactive m4i (n=10 per group); p=0.0234 for m4. k, l Severity scores of axonopathy (k; p=0.0007 for m4) and inflammation (l) in TG from HSV-infected control mice, mice treated with active m4 or inactive m4i (n=10 per group) with statistical analysis (Ordinary one-way Anova, multiple comparisons with ns: not significant; *p<0.05; ***p<0.001). Each graph shows individual and mean values with standard deviation. AAV viral loads are shown in Supplemental Fig.9mo. Source data are provided as a Source Data file.

Across our studies, we observed that ~10% of the animals treated with a high dose of AAV/meganuclease (231012 vg/animal, or approximately 691013 vg/kg) exhibited clinical signs consistent with hepatotoxicity, including weight change, bloating, and general health decline. When lower doses were evaluated, we observed substantially improved tolerability, both clinically and upon histopathological examination and ICF quantification. To further reduce hepatoxicity, we evaluated the use of neuron-specific promoters (Calmodulin Kinase II (CamKII) and human Synapsin (hSyn)) combined with the CMV enhancer17, to test the hypothesis that limiting enzyme expression to neuronal tissues would decrease or perhaps prevent liver toxicity (Fig.9a). We found that latently infected mice treated with a high dose (21012 vg) of AAV9-E/CamKII-m4 or AAV9-E/hSyn-m4 did not show any clinical signs of hepatotoxicity (weight change, general health decline, or ICF), in contrast to mice treated with 21012 vg AAV9-CBh-m4 (Fig.9b, c). Moreover, while liver inflammation increased over time in AAV9-CBh-m4-treated mice, it remained low in AAV9-E/CamKII-m4 (Fig.S11d). Intriguingly, histopathologic signs of neurotoxicity in TG from AAV9-E/CamKII-m4 or AAV9-E/hSyn-m4 treated mice were also similar to those in control mice, while they were significantly higher in TG from AAV9-CBh-m4 treated mice (Fig.9d, e). A decrease of ganglionic viral loads of 67.9% (p=0.07) and 70.4% (p=0.05) in SCG and TG respectively, was observed in AAV9-E/CamKII-m4-treated mice but not in mice treated with the AAV9-E/hSyn-m4 (Fig.9f, g). Assessment of m4 expression in neuronal tissues at different times post administration of either AAV9-CBh-m4 or AAV9-E/CamKII-m4 showed that the m4 expression increased over time but was in general slightly lower in tissues from AAV9-E/CamKII-m4-treated mice compared with AAV9-CBh-m4-treated mice (Fig.S11a, b). This may explain the slightly lesser degree of viral load reduction in AAV9- E/CamKII-m4-treated mice compared with AAV9- CBh-m4-treated mice (Fig.9f, g). We conclude that the AAV9-E/CamKII-m4 regimen retains efficacy and shows improved tolerability compared to AAV9-CBh-m4.

a Experimental timeline of ocular infection and meganuclease therapy. b Average weight change of infected control mice (n=13) or HSV-infected mice treated with m4 expressed from either the ubiquitous CBh promoter, or the neuronal promoters E/CamKII or E/hSyn (n=12 per group). c Inflammatory cell foci (ICF) in liver sections from either HSV-infected control mice (n=13), or mice treated with m4 expressed from either the CBh, E/CamKII or E/hSyn promoter (n=12 per group); p=0.00455 for CBh-m4. d, e Severity scores of inflammation (d; p=<0.0001 for CBh-m4) and axonopathy (e; p=<0.0001 for CBh-m4) in TG from infected control mice (n=10) and infected mice treated with m4 expressed from either the CBh, E/CamKII or E/hSyn promoter (n=12 per group) with statistical analysis (Ordinary one-way Anova, multiple comparisons with ns: not significant; *p<0.05; ****p<0.0001). f, g HSV loads in SCGs (f) and TGs (g) from infected control (n=10) and infected mice treated with m4 expressed from either the CBh, E/CamKII or E/hSyn promoter (n=12 per group). Percent decrease of HSV loads in treated mice compared to control mice and statistical analysis (Ordinary one-way Anova, multiple comparisons with p values; ns: not significant). Each graph shows individual and mean values with standard deviation. The AAV viral loads are shown in Supplemental Figure10ac. Source data are provided as a Source Data file.

Continued here:
Gene editing for latent herpes simplex virus infection reduces viral load and shedding in vivo - Nature.com

Related Posts