African green monkey (Vero) cells were maintained in complete Dulbecco’s Modified Eagle Medium (DMEM) (Gibco) supplemented with 10% Fetal Calf Serum (FBS) (ThermoFisher). VC2 was constructed as described previously (30). Briefly, the VC2 recombinant virus was constructed utilizing the two-step Double-Red Recombination protocol using the HSV-1(F) viral genome cloned as a bacterial artificial chromosome (BAC). The virus was cultivated in Vero cells. HSV-1((McKrae) was a gift by the late Dr. James Hill (Louisiana State University Health Sciences, Center, New Orleans, LA).

Female Balb/CJ mice (8-10-week-old) were purchased from Jackson Laboratories, (Bar Harbor, ME USA) and were housed in the Louisiana State University School of Veterinary Medicine (LSU-SVM) ABSL2 facility. A prime-boost vaccination strategy was used. For prime, 100µl of vaccine (10⁷ PFU in DMEM) were injected intramuscularly into the right hind leg followed by a booster dose into the left hind leg 21 days later. Mock vaccinated animals received PBS. All animals were challenged 21 days or 8 months after the last vaccination with a lethal dose (10⁶/eye) of HSV-1 (McKrae). For challenge, animals were anesthetized, and a linear partial epithelial corneal debridement was performed with a 27G needle before 10µl of HSV-1 (McKrae) was applied to the ocular surface. Animals were observed daily for clinical signs of disease and euthanized as described in the IACUC euthanasia criteria.

Ocular scoring of mice was performed by Dr. Andrew Lewing, a board-certified veterinary ophthalmologist (ACL) according to a modified established ocular disease scoring system (45). Briefly, this scoring system provides objective categorization of corneal opacity, corneal neovascularization, corneal ulceration and ocular discharge [see Table 1 for the scoring system]. All examinations were performed following induction of a light plane of anesthesia using inhaled isoflurane in oxygen using a handheld biomicroscope (Kowa, SL-17). Normal animals were assigned a score of 0 per eye, and animals with ocular disease were assigned a score of up to 8 per eye, for a maximum possible score of 16 per animal at each time point. All animals were confirmed to be normal with a score of 0 for each eye prior to challenge and were scored again 5 days post infection.

The whole mouse eye was collected in PBS following euthanasia, mince and incubated with collagenase in HBSS buffer for 2 hours. The homogenized solution was passed through a 70µm filter to prepare a single cell solution. Mandibular lymph nodes (mLN) were collected and processed through a 70µm filter to prepare a single cell solution. A pre-titrated antibody mixture was incubated for 30 minutes at 4°C, washed and fixed with 2% paraformaldehyde. The next day, samples were analyzed using the BD FACS-Aria equipment and data was processed using FCS Express 7. The anti-mouse antibodies used for flow cytometry were; CD45-APC Vio770, Ly6G-PE CF594, CD3-PerCP-ef710, CD4-FITC, CD8a-BV650, γδTCR-APC, MHCii-BV711, CD11C-PE. CD64-BV605, CD49b-BV421, CD19-BV786, CD44-BV711 and CD62L-BV421. Gating strategy is presented in Supplementary Figure S1 .

Following euthanasia, eyes were immediately fixed using 10% formalin for 3 days and processed in the Histology Core Facility of LSU-SVM. For IFM and RNAScope, 5µm thick Formalin-Fixed Paraffin-Embedded (FFPE) sections were prepared on glass slides. To detect viral replication, the HSV-1 UL-48 RNA was used as the target gene with the Probe-V-HSV-1-UL48-C3 purchased from Advanced Cell Diagnostics. The RNAScope assay was performed according to ACDBio guidelines using the Opal 620 dye (Akoya Biosciences) as the substrate. Following the RNAScope assay, the slides were blocked with 10%FBS and incubated overnight with rabbit polyclonal anti-HSV-1 (Dako). The following day slides were washed and anti-rabbit FITC was used as the secondary antibody. Next, the background was reduced with the TrueView auto fluorescence kit (Vectorlabs). After adding mounting media with DAPI, slides were visualized with a Zeiss observer Z1 inverse microscope.

Whole mouse eyes were collected after euthanasia and immediately frozen using OCT in liquid nitrogen. Samples were stored at -80°C until processing. For microscopy, 8µm thick frozen sections were prepared on glass slides using a cryostat. Tissue sections were fixed briefly for 1 minute using 2% paraformaldehyde at room temperature (RT). After blocking with 5% FBS for 1 hour, slides were incubated with fluorophore-conjugated primary antibody overnight at 40°C. Slides were then washed, fixed and mounting media was added. Images were captured with a Zeiss observer Z1 microscope. The antibodies used for IFM were: Anti-γδTCR-AF488 (GL3), anti-LY6G-AF594 (1A8) and Alexa Fluor 594 anti-mouse CD31 (MEEC13.3) (Biolegend. Inc.) and LYVE-1 AlexaFluor488 (ALY7) (Thermofisher, Inc).

BrdU labeling reagent (Invitrogen, cat# 000103) was injected (1ml/Kg) intraperitoneally (IP) on the day before challenge and administered every day following infection for 4 days. At day 5, animals were euthanized, cells were stained and fixed as mentioned above. Cells were then permeabilized using permeabilization buffer (eBioscience) and incorporated BrdU was stained using Rb-anti-BrdU primary antibodies and aRb-AlexaFluor488 secondary antibodies. Data was captured using a BD FACS Aria Flow cytometer and calculated as a percentage of cells positive for BrdU within a specific population.

At 5 days post infection (dpi), viral shedding on the ocular surface was determined using plaque assay. For virus collection, a sterile cotton swab soaked in DMEM was swabbed gently on the ocular surface and collected in a 1.5 ml Eppendorf tube containing DMEM. Samples were stored at -80°C. For virus quantification, a viral plaque assay was performed using Vero cells. Vero monolayers were prepared in 12 well tissue culture plates and incubated with 10-fold dilutions of each sample at room temperature (RT) for 1 hour with shaking. Cells were subsequently washed with complete medium and incubated with complete medium with 1% methyl cellulose for 3 days at 37°C with 5% CO₂ for viral plaque formation. Next, plates were washed, fixed using 10% formalin and stained with crystal violet. Viral plaques were counted using a light microscope and calculated according to the dilution factor.

Mouse tears were collected using a sterile cotton swab and placed in a 1.5 ml Eppendorf tube containing DMEM. Each sample represents a pool of 5 animals. Approximately 100 PFU of HSV-1 (McKrae) was mixed with 0.5 ml DMEM containing tears and incubated at 37°C for 1 hour for neutralization. After incubation, the mixture was placed on Vero monolayers on 12 well plates and incubated at room temperature with shaking. The plates were subsequently washed and incubated with complete medium with 1% methyl cellulose for 3 days at 37⁰C with 5% CO₂ for viral plaque formation.

Following challenge with HSV-1(McKrae), TGs were collected at the time of euthanasia and kept frozen until analysis. On the day of analysis, TGs were thawed, and total DNA was collected using the Qiagen Blood and Tissue kit per manufacturer’s instructions. HSV-1 glycoprotein D (gD) was used as the target gene for quantification using the following primer-probe mixtures purchased from IDT; gD FP – 5^’-GTCCGGAAACAACCCTACAA-3^’, gD RP – 5^’-GCATTCGGTGTACTCCATGA-3^’, and qPCR Probe – PrimeTime 5’ 6-FAM™/ZEN™/3’ 5^’-TTGGTTTCGGATGGGAGGCAACT-3^’ IB^®FQ. For qPCR, the Prime time Gene Expression Master Mix (IDT) was used according to the manufacturer’s instructions and the reaction was run using the 7900HT Fast Real-Time PCR System with the 384-Well Block Module. TGs from naïve animals were used to set cut-off values and the gD G-block (IDT) was used to create a standard curve.

To reduce T cell infiltration in the mouse eyes following infection, FTY720 (Millipore-Sigma) was applied topically to ocular surfaces. FTY720 was dissolved in water at 10 mg/mL. One drop (approx. 10 μl) of this solution was then applied twice daily to the ocular surface one day before infection and continued until 10 DPI. Uninfected naïve mice were treated similarly as controls.

Whole mouse eyes were collected following euthanasia and immediately frozen using liquid nitrogen. On the day of the assay the eyes were pulverized using a mortar and pestle while frozen. The resulted homogenized tissue was weighed and dissolved in Tris-based lysis buffer (Thermofisher) supplemented with protease inhibitor for total protein extraction. For detecting cytokine and chemokine, the Cytokine & Chemokine Convenience 26-Plex Mouse ProcartaPlex™ Panel 1 (Thermofisher) was used according to manufactures instructions and data was acquired using the Bioplex200 equipment.

Statistical analysis was performed using GraphPad Prism 9 software. Survival analysis was performed using the log-rank test. For analysis between three groups one-way ANOVA and Kruskal-Wallis test were performed. To compare results between two groups, the Mann-Whitney test was utilized. The statistical significance level was set at p = 0.05.

The VC2 vaccine strain specifies the amino-terminal deletion of 39 amino acids of glycoprotein K (gK), which has been shown to prevent entry into neuronal axons as well as fusion of the virus with cellular plasma membranes, while the virus replicates efficiently because it can enter epithelial and other cells through endocytosis (21–23). To confirm that the VC2 vaccine strain that contains the gK and the UL20 amino terminal deletions cannot infect neuronal endings, travel to the TG and establish latency, mouse corneas of naïve mice were infected with 10⁶ PFU per eye after mild-scarification with either the HSV-1(F) parental wild-type virus or the VC2 vaccine strain. Both the parental HSV-1(F) and VC2 viruses were avirulent, since none of the mice succumbed to the infection. Two weeks post infection, the amount of viral DNA in the TGs was quantified by quantitative PCR. HSV-1(F) but not VC2 viral DNA was detected in ganglionic tissues indicating that VC2 was unable to reach the TGs and establish latency ( Figure 1A ).

Previously, we reported that VC2 intramuscular vaccination of mice generates robust protection against lethal ocular HSV-1 (McKrae) challenge (18). To assess whether this protective immune response is virus replication-dependent and sustained over time, we immunized mice with VC2 or Ultraviolet (UV)-inactivated VC2 and challenged the mice ocularly with HSV-1 (McKrae) at 21 days or 8 months after the booster immunization. VC2 vaccinated mice were fully protected at both time points ( Figures 1B, C ), and there were no apparent ocular and/or systemic clinical disease symptoms observed ( Figure 1D ). In contrast, mice immunized with the UV-inactivated VC2 succumbed to the HSV-1 (McKrae) within 5 DPI and significant ocular damage was noted characterized by ocular inflammation and cornea damage ( Figures 1D, E ). Determination of the relative number of viral genomes in ganglionic tissues by qPCR revealed that a significantly higher level of viral DNA was present in the TG of UV-VC2 vaccinated animals compared to those vaccinated with VC2 ( Figure 1F ). This data suggests IM immunization with live attenuated VC2 generates a robust protection compared to vaccination with inactivated VC2 virions.

Previously, we showed that VC2 prime-boost intramuscular immunization induces a strong systemic neutralizing antibody response (18). To assess the contribution of neutralizing antibody in the observed ocular protection against HSV-1 (McKrae) infection, tears from vaccinated mice with either the VC2 vaccine, the UV-inactivated VC2, or mock-vaccinated mice prior to challenge were tested for the presence of neutralizing antibody using a plaque reduction neutralizing assay. Tears from VC2 and VC2-inactivated vaccinated animals did not have higher neutralizing ability compared to mock-vaccinated animals ( Figure 2A ), suggesting the absence of strong neutralizing antibody activity at ocular surfaces. Detection of viral antigens on FFPE sections by indirect immunofluorescence of ocular tissues with anti-HSV-1 polyclonal antibody revealed the presence of viral antigens on ocular surfaces at 2- and 5 DPI in all groups of animals ( Figure 2B , green). In addition, transcription of the HSV-1 UL48 gene was detected in all ocular tissues using the RNAScope assay ( Figure 2B , red dots). These results indicated productive infection and viral replication in all groups for at least 5DPI. However, at 5DPI viral shedding in VC2-vaccinated animals were undetectable compared to the mock and UV-VC2-vaccinated mice ( Figure 2C ). This data suggests that despite productive infection and ocular viral replication, the VC2-vaccinated animals experienced robust ocular mucosal responses that significantly reduced viral infection and resultant immunopathogenesis in the absence of neutralizing activity.

To assess the extent of cell-mediated immunity at ocular surfaces following infection with HSV-1 (McKrae), immune cellular infiltration was evaluated at 2-, 5- and 9 DPI. No detectable differences were noted for macrophages, dendritic cells and NK cells infiltration at any time points ( Figures 3A–C ). A large neutrophil influx was noted following infection in mock-vaccinated animals at 5DPI ( Figure 1E ). In contrast, VC2 vaccinated animals showed significantly lower neutrophil counts at 5DPI ( Figures 3D–F ). A significantly higher T cell infiltration was noted in VC2, but not in mock-vaccinated mice at 5DPI ( Figures 3G–I ). Further analysis revealed that the majority of the infiltrating T cells in all groups of mice expressed the gamma-delta (γδ) TCR, while a relatively small population of CD4+, CD8+ and double negative TCR T cells were also present ( Figures 3J–M ). These results show that γδT cells were the dominant population of immune cells in vaccinated animals, while neutrophils were the major immune cell infiltrate in mock-vaccinated animals at 5DPI. Non-parametric Spearman correlation among all groups of mice revealed a negative correlation between γδT cells and neutrophil accumulation in corneas at 5DPI among all groups ( Figure 3N ). The infiltrating γδT cell numbers peaked at 9 DPI and eventually dropped to basal levels at 25 DPI (after the resolution of ocular pathogenesis ( Figure 3K ). Importantly, the UV-VC2 vaccinated group of mice had lower γδT cell accumulation and high neutrophil accumulation at 5DPI in comparison to the VC2-vaccinated mice that correlated with the observed elevated ocular disease scores in UV-VC2 versus VC2-vaccinated mice ( Figures 1E , 3E, H ).

To address whether the presence of γδT cells following infection in VC2-vaccinated animals is the result of local proliferation or infiltration, a flow cytometry analysis for Ki67, a marker for cellular proliferation, was performed. Following infection, there was no significant increase of Ki67⁺ γδT cells in VC2-vaccinated animals ( Figure 3O ) suggesting that the presence of γδT cell is due to the increased infiltration rather than local proliferation of resident γδT cells. In addition, γδT cells were detected by indirect immunofluorescence on OCT sections of ocular tissues in the corneal stroma of vaccinated animals with little to no neutrophil presence ( Figure 3P , right-most panel). In contrast, a high number of neutrophils were detected in the corneal stroma and epithelium in mock-vaccinated animals, while low numbers of γδT cells were also present ( Figure 3P , left-most panel). Taken together, this data suggests VC2 vaccinated animals recruit γδT cells following infection at the ocular surface and that increased neutrophil migration is prevented.

HSV-1 infection induces neovascularization following infection (46), and infiltrating immune cells may prolong neovascularization by secreting pro-angiogenic growth factors. This neovascularization is likely to be the source of increased levels of corneal neutrophils (47). To assess whether the VC2-vaccination affects neovascularization, OCT sections were stained for LYVE-1, a marker of lymph-angiogenesis (48) and CD31, a marker of angiogenesis (49). HSV-1 infection in mock-vaccinated animals had high levels of CD31 expression ( Figure 4 , left panel) as also described previously (46, 49). In contrast, VC2-vaccinated animals exhibited strong reactivity with the anti-LYVE-1 antibody, while little or no CD31 expression was detectable in the cornea ( Figure 4 , right panel). This data suggests that neovascularization is not responsible for the higher infiltrating γδT cells in VC2 immunized mice.

To address the status of immune response, we used a 26 plex immunoassay to detect Th1/Th2Th9/Th17Th22 and Treg associated cytokines and chemokines in homogenized eye samples at 5 DPI. The VC2 and UV-VC2 vaccinated animals exhibited a unique cytokine and chemokine expression profile compared to the mock-vaccinated animals. Specifically, the VC2 and UV-VC2 immunized animals had significantly higher IL-4 and IL-22 levels in the eye ( Figures 5A, B ). In contrast, the pro-inflammatory cytokine IL-5 and chemokine Gro-alpha/KC, IP-10 and MCP-1 was detected at lower levels in VC2 and UV-VC2 immunized animals compared to mock-vaccinated animals ( Figures 5C–F ). It has been reported that these pro-inflammatory cytokines are associated with tissue damage, neutrophil accumulation and increased severity of HSV-1 infection (47, 50). Overall, this data suggests VC2 immunization generates a unique adaptive response that reduces pro-inflammatory signals and that this reduction is associated with infiltrating γδT cells.

Several studies suggested that γδT cells may contain memory populations like αβT cells and can undergo memory-like expansion following antigen recognition (51–56). To address the possibility of whether infiltrating γδT cells represent an HSV-1 specific memory population, we performed a BrdU proliferation experiment in-vivo. Both mock and VC2-vaccinated animals were administered 1 ml concentrated BrdU per 100 g body weight via the intraperitoneal route one day before infection and continued every day. Following euthanasia, both eyes were removed and mLN cells were stained for BrdU positive cells. VC2-vaccinated animals exhibited marked incorporation (red) of BrdU in B (CD19), CD4 and CD8 T cells in mLN at 5DPI compared to mock-vaccinated (gray) animals, suggesting a pre-existing memory population for these cells ( Figure 6A ). However, there was no significant difference detected for BrdU-positive γδT cells in mLN tissues ( Figure 6A , right panel), indicating the absence of a memory population. In addition, BrdU positive γδT cells in the vaccinated animals did not have a higher frequency of T central memory (TCM CD44+CD62L) as BrdU positive CD4+ and CD8+ T cells ( Figure 6B ) suggesting the absence of TCM in the proliferating γδT cell population. Further, there was no difference in BrdU-positive γδT cell in the ocular mucosa tissues following infection and only a small percentage of γδT were BrdU-positive in both mock and VC2-vaccinated groups ( Figure 6C ) suggesting infiltration of pre-existing rather than newly proliferative γδT cells in the eye following infection. Taken together, our data strongly suggest that the infiltrating γδT cells in vaccinated animals were not HSV-1 specific/experienced memory populations, but rather non-specific γδT cells recruited in the infected cornea in vaccinated animals with increased frequency.

Although it has been reported that γδT cells have an important role on the mucosal surface (57–60), the role of the γδT cells in vaccine-induced protection against herpes ocular immunopathogenesis and specifically herpes keratitis has not been investigated. The association of γδT cell infiltration ( Figure 3K ) and lower ocular damage ( Figure 1E ) suggest a functional role of this infiltrating population on corneal pathology following infection. This raises the question of whether the presence of this population is necessary for the control of HSV-1 induced keratitis in vaccinated animals. Unfortunately, there is neither a γδT cell KO animal in the Balb/C background nor an appropriate depleting antibody to remove this cell type from systemic and peripheral circulation to study the specific role of γδT cells following HSV-1 challenge. To circumvent this issue, we used FTY720 (Fingolimod, Sigma), an FDA-approved drug for immune suppression, which inhibits lymphocyte egress from both thymus and secondary lymphoid organs (61). Following infection in vaccinated animals, 1 drop (approximately 10 μl) of FTY720 (10mg/ml) was applied twice daily topically to the mouse eyes to prevent T cell infiltration. Because FTY720 was applied directly to the ocular surface, it was unlikely to alter cellular migration in other tissues. As expected, the FTY720 treatment lowered γδT cell infiltration at 5 DPI ( Figure 7A ). At the same time, FTY720-treated vaccinated animals exhibited a significant increase in neutrophil infiltration ( Figure 7B ) and concomitant increased ocular disease score ( Figure 7C ) compared to the PBS-treated and vaccinated animals. To confirm that FTY720 treatment did not increase ocular scores, naïve animals were also treated with FTY720 and showed minimal ocular damage in the absence of infection ( Figure 7C , green). Further, there was no substantial increase in viral shedding in VC2-vaccinated animals compared to PBS-treated animals (data not shown) suggesting that the observed ocular damage is not due to persistent viral replication, but rather the result of increased neutrophil infiltration. The ocular damage persisted (gross observations) as long as the FTY720 treatment continued, and animals recovered quickly after FTY720 treatment termination (data not shown). Increased neutrophil infiltration ( Figure 7D ) and neovascularization ( Figure 7E ) was also detected more frequently after the challenge in FTY720-treated vaccinated animals. Although FTY720 treatment following challenge in vaccinated animals increased ocular pathogenesis, it did not have any effect on survival (data not shown), as well as on viral loads in the TG tissues ( Figure 7F ) suggesting that the role of γδT cell accumulation is limited to control of ocular immunopathogenesis induced by HSV-1 infection.