Herpes simplex virus-1 KOS-63 strain is virulent and causes titer-dependent corneal nerve damage and keratitis

To investigate the acute clinical, immunological, and corneal nerve changes following corneal HSV-1 KOS-63 strain inoculation. Corneas of C57BL/6 mice were inoculated with either low dose (Ld) or high dose (Hd) HSV-1 KOS-63 or culture medium. Clinical evaluation was conducted up to 7 days post inoculation (dpi). Viral titers were assessed by standard plaque assay. Excised corneas were stained for CD45 and beta-III tubulin. Corneal flow cytometry was performed to assess changes in leukocyte subpopulations. Corneal sensation was measured using a Cochet-Bonnet esthesiometer. Naïve, sham-infected (post scarification), and McKrae-infected C57BL/6 corneas served as two negative and positive controls, respectively. Compared to Ld infected mice, Hd HSV-1 KOS-63 demonstrated higher incidence of corneal opacity (1.5 ×) and neovascularization (2.6 × ; p < 0.05). At 7 dpi Hd infected mice showed more severe corneal opacity (2.23 vs. 0.87; p = 0.0003), neovascularization (6.00 vs. 0.75; p < 0.0001), and blepharitis (3.11 vs. 2.06; p = 0.001) compared to the Ld group. At 3 dpi epitheliopathy was significantly larger in the Hd group (23.59% vs. 3.44%; p = 0.001). Similarly, corneal opacity was significantly higher in Hd McKrae-infected corneas as compared with Ld McKrae-infected corneas at 3 and 5 dpi. No significant corneal opacity, neovascularization, blepharitis, and epitheliopathy were observed in naïve or sham-infected mice. Higher viral titers were detected in corneas (1 and 3 dpi) and trigeminal ganglia (TG) (3 and 5 dpi) in Hd versus Ld KOS-63 groups (p < 0.05). Leukocyte density showed a gradual increase over time from 1 to 7 dpi in both KOS-63 and McKrae-infected corneas. Corneal flow cytometric analysis (3 dpi) demonstrated a higher percentage of Gr-1 + (71.6 vs. 26.3) and CD11b + (90.6 vs. 41.1) cells in Hd versus Ld KOS-63 groups. Corneal nerve density significantly decreased in both Hd KOS-63 and Hd McKrae infected corneas in comparison with naïve and sham-infected corneas. At 3 dpi corneal nerve density was lower in the Hd versus Ld KOS-63 groups (16.79 vs. 57.41 mm/mm2; p = 0.004). Corneal sensation decreased accordingly at 5 and 7 dpi in both Ld and Hd KOS-63-infected mice. Corneal inoculation with HSV-1 KOS-63 strain shows acute keratitis and nerve degeneration in a dose-dependent fashion, demonstrating virulence of this strain.


Materials and methods
Animals. Six-to-twleve-week-old male and female C57BL/6 mice (Charles River Laboratories, Inc, Wilmington, MA) were used in this study. All the procedures were approved by Institutional Animal Care and Use Committee of Tufts Medical Center and the Schepens Eye Research Institute. Our study was compliant with Association for Research in Vision and Ophthalmology statement for the use of animals in ophthalmic and visual research. were used in this study. Vero cells (African green monkey kidney cells, kindly provided by Dr. Judy Lieberman, Boston Children's Hospital, Boston, MA) were cultured in RPMI 1640 medium with L-glutamine (Corning/ Mediatech, Inc. Manassas, VA) on 6-well plates. After a monolayer was formed, the cells were infected with the virus and incubated for 48 h at 37 °C. Virus-infected Vero cells were incubated with gentleMACS dissociator (Miltenyi Biotec GmbH, Bergisch Gladbach, Germany) to extract the virus from the infected cells. High-speed centrifugation (40,000 G for 30 min at 4 °C) was applied to pellet the virus. The virus pellet was re-suspended in RPMI 1640 as above. The prepared virus was titrated using a standard plaque assay 10 and aliquots of the virus were kept at − 80 °C for further use.
Corneal HSV-1 inoculation and study groups. Mice were anesthetized with intraperitoneal injections of ketamine (100 mg/kg) and xylazine (10 mg/kg) mixture. A drop of 0.5% proparacaine hydrochloride (Akorn, Lake Forest, IL) was applied on the right eye to obtain local anesthesia. Right corneas were then scarified in a 5 × 5 grid pattern with a 30-gauge needle. Five microliter (µl) of RPMI containing 2 × 10 4 (Low dose, Ld) or 2 × 10 6 (High dose, Hd) plaque forming unit (PFU) of HSV-1 were then applied on the scarified corneas. Naïve and sham-infected mice were used as negative controls throughout the study. Sham-infected mice were scarified similar to infected mice and corneas were inoculated with 5 µl of RPMI. HSV-1 McKrae infected mice served as positive controls in this study. One dose of buprenorphine (0.1 mg/kg) was injected subcutaneously right after the corneal scarification for pain control. The studies were carried out in compliance with the ARRIVE guidelines 35 . Clinical grading. Mice were anesthetized at days 1, 3, 5, and 7 post-inoculation (dpi) with a ketamine/xylazine mixture and evaluated for blepharitis, corneal opacity, neovascularization, and epitheliopathy. Blepharitis was scored as 0, no puffiness of eyelids; 1, noticeably puffy lids; 2, puffy lids plus moderate crusting; 3, eyelids 50% swollen shut with severe crusting; and 4, eyelids totally shut 9 . Corneal opacity was scored using slit-lamp bio-microscopy, using a modified scale 27: 0, no opacity; 1, epithelial involvement of < 50%; 1.5, epithelial involvement of > 50% of the cornea; 2, stromal involvement (haze) of < 50% of the cornea and iris visible; 2.5, stromal involvement of > 50% of the cornea and iris visible; 3, stromal involvement of < 50%, iris partially invisible; 3.5, stromal involvement of > 50%, iris partially invisible; and 4, totally opaque cornea with invisible iris. Corneal neovascularization was graded from 0 to 12, based on the number of clock hours (with every 30° considered as one clock hour) that new vessels, extending from the limbus toward the cornea, occupied 36  www.nature.com/scientificreports/ epitheliopathy we used a previously described method 37 . In brief, a 1.0 mg fluorescein sodium ophthalmic strip (FUL-GLO, Akorn, Lake Forest, IL) was diluted in 1 ml 1 × PBS (Life Technologies, Carlsbad, CA) and 5 µl of 0.1% fluorescein solution was applied on the ocular surface. After 15 s, the cornea was washed with PBS 2-3 times and using a cotton tip applicator, the excess fluid gently removed after each wash to prevent pooling of fluorescein. Photos were obtained by a Canon digital camera (PowerShot A2200), mounted on top of the slit-lamp binocular (SL-1E, Topcon, Oakland, NJ), using white light (prior to fluorescein staining) and cobalt blue light. Images were then analyzed with ImageJ software 38 (public domain Java image processor, NIH, Bethesda, MD; https ://image j.nih.gov/ij/) to calculate the percentage of epithelial defect. For this purpose, the area of epitheliopathy, defined as the distinct fluorescein stained areas (including dendrites), was selected using polygon selection tool and measured first. Then the epitheliopathy area was divided by the area of the whole cornea, measured by selection of the whole cornea using oval selection tool. Clinical grading was performed by 3 masked observers.
Viral titration of corneas and trigeminal ganglia. We pooled and homogenized 2 corneas or 2 ipsilateral TGs from 2 infected mice in 300 µl of RPMI using gentle MACS dissociator for each experiment. A standard plaque assay 10  Confocal microscopy and image analysis. The whole-mounted corneas were imaged using a confocal Leica TCS SP5 (Leica Microsystems, Wetzlar, Germany). Central and peripheral areas of each cornea were assessed separately, with the central area defined as the area within 0.6 mm of the corneal center, and the periphery defined as being within a 0.9-1.5-mm radial distance from the center. Three different non-overlapping fields from the peripheral cornea and one central field were imaged for each cornea. At least three different corneas were stained and imaged per group. Stacked images were analysed off-line with IMARIS software version 8.0 (Bitplane AG, Zurich, Switzerland) to calculate the number of CD45 (pan-leukocyte marker) positive cells. The number of cells in peripheral corneal fields was then averaged. Corneal nerves were traced using NeuronJ 39,40 (http://www.image scien ce.org/meije ring/softw are/neuro nj/), a free semi-automatic software. Nerve density was reported as mm/mm 2 .
Flow cytometry. At  www.nature.com/scientificreports/ micro-filament length required to measure a blink reflex was recorded. The procedure was repeated three times to ensure reproducibility and average length was used for analysis.

Statistical analysis.
Statistical analysis was performed with Prism version 6.01 (GraphPad Software, Inc., La Jolla, CA). We used chi square test to compare incidence percentage of corneal opacity, blepharitis, and neovascularisation between two different groups of infected mice. Student's t-test was used to compare means between the two groups and one-way ANOVA, followed by Tukey's test used to compare the means among three or more groups. Results were shown as mean ± standard error of the mean (SEM) and p values of < 0.05 were considered as statistically significant.

Results
Titer-dependent incidence and clinical severity in mice infected with Ld (2 × 10 4 PFU) versus Hd (2 × 10 6 PFU) HSV-1 KOS-63. Ten to 17 mice at each time point were evaluated for clinical signs of the disease in either 2 × 10 4 PFU (Ld) or 2 × 10 6 PFU (Hd) KOS-63-infected mice. At the end of the follow-up period (7 dpi), the incidence of blepharitis did not show a significant difference between the two doses of the virus (93.7% and 94.4% in Ld and Hd KOS-63-infected mice, respectively, p = 0.93). However, the incidence of mice, which developed corneal opacity was significantly higher in the Hd KOS-63-infected mice (88.2%) as compared to the Ld KOS-63-infected mice (56.2%), p = 0.03. The incidence of corneal neovascularisation was also significantly higher in the Hd KOS-63-infected mice (81.8%), as compared to the Ld KOS-63-infected mice (31.2%), p = 0.009. Representative white light and cobalt blue filtered images for comparison of corneal opacity and epitheliopathy in the Ld and Hd HSV-1 KOS-63 infected corneas, as well as naïve and sham-infected mice at different time points are illustrated in Fig. 1. Blepharitis was not present at 1 dpi in any of the groups; however, at 3 dpi blepharitis scores were 0.18 ± 0.10 and 0.62 ± 0.12 in the Ld-and Hd-infected mice, respectively (p = 0.68). At 7 dpi, blepharitis scores increased significantly to 2.06 ± 0.26 and 3.11 ± 0.25 in the Ld-and Hd-infected mice, respectively (p = 0.001). Blepharitis scores showed an increasing trend from 1 to 7 dpi (ANOVA: p < 0.0001) ( Fig. 2A).

Immune cell infiltration in acute herpes simplex keratitis induced by
To further analyze the subpopulation of infiltrated cells in corneas of HSV-1 KOS-63 strain infected mice, we used flow cytometry. The percentage of CD45 + cells in the cornea increased to 4.2% and 8.3% in Ld and Hd groups, respectively, compared to naïve corneas (2.1%). We then gated on CD45 + cells and compared the percentage of other immune cell subpopulations between naïve and infected mice (Fig. 5). While Gr-1 + cells Blepharitis scores. At 1 dpi no blepharitis was observed but it increased gradually in both Hd and Ld infected mice. Blepharitis score was significantly higher in the Hd infected mice compared to Ld infected mice at 7 dpi (**p = 0.001). No blepharitis was observed in naïve and sham infected mice (1-7 dps), 1 dps is presented as representative. (B) Corneal opacity scores are demonstrated in both Ld and Hd infected mice. Opacity scores were higher in the Hd infected mice as compared to Ld infected mice and was significant at 5 dpi (*p = 0.04) and 7 dpi (**p = 0.003). No corneal opacity was observed in naïve and sham infected mice (1-7 dps), 1 dps is presented as representative. (C) Corneal neovascularization scores according to the clock hours area of vessel coverage. Neovascularization was more significant in the Hd infected mice compared to Ld infected mice at 7 dpi ( ǂp < 0.0001). No corneal neovascularization was observed in naïve and sham-infected mice (1-7 dps), 1 dps is presented as representative. (D) Percentage of corneal epithelial defect was higher in the Hd infected mice compared to Ld infected mice in all time points and was significant at 3 dpi (**p = 0.001). No corneal opacity was observed in naïve mice. Minor epithelial defect remained after corneal scarification only at 1 dps, which was not statistically significant compared to naïve mice. Bars are showing mean ± SEM. ANOVA: p < 0.0001. ¥, demonstrates the statistical significance (p < 0.05) between the marked column and naïve and sham-infected. p values are calculated by one-way ANOVA followed by Tukey's multiple comparison test.    Figure  However, the central corneal nerve density significantly decreased at 3 and 7 dpi for the Ld and for all timepoints for the Hd KOS-63-infected mice, as compared to sham-infected (p < 0.05) (Fig. 6B). As early as 3 dpi, the central corneal nerve density was significantly lower in the Hd compared with Ld KOS-63-infected mice (16.79 ± 3.17 vs. 57.41 ± 14.64 mm/mm 2 , respectively; p = 0.004) (Fig. 6B). Peripheral corneal nerves further decreased significantly at 7 dpi in the Ld KOS-63-infected mice (30.87 ± 3.61 mm/mm 2 ) compared to shaminfected (p = 0.005). Morevoer, in the Hd KOS-63-infected corneas, peripheral corneal nerves decreased significantly at 3 dpi (33.48 ± 12.99 mm/mm 2 ) and 7 dpi (12.79 ± 8.55 mm/mm 2 ) compared to sham-infected (p = 0.01 and p < 0.0001, respectively) (Fig. 6C). Interestingly, there was no significant difference in peripheral corneal nerve density between the Ld and Hd KOS-63-infected groups (Fig. 6C). However, we did observe a linear decrease in the central and peripheral corneal nerve densities during the 7 days follow-up period in both Ld and Hd KOS-63-infected mice (ANOVA: p < 0.0001). Similar to Hd KOS-63-infected corneas, central corneal nerve density was significantly reduced at 1 dpi in Hd McKrae-infected corneas (62.20 ± 4.32 mm/mm 2 ), as compared to naïve and sham-infected controls. (p = 0.0003 and p = 0.02, respectively ( Supplementary Fig. 4). Additionally, corneal esthesiometry confirmed significantly decreased corneal nerve sensitivity, starting from 5 dpi in both Ld and Hd HSV-1 KOS-63-infected corneas. At 5 dpi corneal nerve sensitivity was significantly lower in the Hd-infected corneas, as compared to Ld-infected corneas (filament length of 0.77 ± 0.37 cm vs. 4.23 ± 0.26 cm; p < 0.0001). Similarly, at 7 dpi, corneal nerve sensitivity was significatly lower in the Hd-infected corneas as compared to Ld-infected corneas (filament length of 0.0 ± 0.0 cm vs. 1.73 ± 0.61 cm; p = 0.007) (Fig. 7). www.nature.com/scientificreports/

Discussion
Herein, we demonstrate that the HSV-1 KOS-63 strain can cause a titer-dependant keratitis in C57BL/6 mice when applied topically after corneal scarification. Although the HSV-1 KOS cannot infect the central nervous system and cause encephalitis when applied peripherally, through corneal inoculation 9,21,25 , we show that it can infect the corneal nerves as part of the peripheral nervous system, and subsequently migrate to the TGs. This is further demonstrated by corneal nerve degeneration observed in our studies. While development of stromal keratitis following inoculation with KOS strain in C57BL/6 mice has been recently reported, the presence of the virus in the TGs or corneal leukocyte and nerve alterations have not been previously assessed 41 . However, in contrast, Nguyen et al. 32 did not observe stromal keratitis following ocular inoculation of HSV-1 KOS on scarified C57BL/6 mice. We show that Hd KOS-63 strain not only increases the frequency of corneal opacity and neovascularization, but also results in increased severity of ocular manifestations as compared to Ld of virus. Similarly, using McKrae strain as a positive control we observed a titer-dependent keratitis. Consistent with our study, higher doses of HSV-1 strain RE demonstrated more severe corneal neovascularization in BALB/c mice at 15 dpi 42 . Moreover, Zheng et al. 43 demonstrated that 2.5 × 10 7 PFU of HSV-1 McKrae caused significantly higher corneal opacity scores than a 100 times lower dose. However, albeit higher dose of HSV-1 KOS-63 (5 × 10 4 PFU vs. 2.5 × 10 4 PFU) increased the incidence of herpetic stromal keratitis in BALB/c mice, it did not cause more severe keratitis 27 . Corneal infection using other HSV-1 strains on mice (CJ394, OD4, 994, 401, 394) 9,18 or rabbits (RE Figure 6. Corneal nerve density in the center and periphery of naïve and sham-infected corneas as well as Ld and Hd HSV-1 KOS-63-infected corneas. (A) Sample confocal micrographs of whole mount corneas stained with NL637 conjugated beta III tubulin antibody. There is a non-significant corneal nerve loss in the sham control group compared with naïve corneas. Sample central and peripheral corneal nerve tracing in naïve and sham-infected mice is demonstrated using NeuronJ software. Diminished corneal nerves in the infected corneas were observed at 1, 3, and 7 dpi especially in the central cornea of Hd infected mice. (B) Corneal nerve density (mm/mm 2 ) was calculated with NeuronJ software after tracing the nerves in whole mount confocal micrographs. (C) Corneal nerve density (mm/mm 2 ) in the peripheral cornea was calculated the same as central cornea. Each group is demonstrating the average of at least 3 whole mount stained corneas. Naïve and sham-infected central and peripheral corneal nerves were not significantly different although there was a trend towards more corneal nerve loss in the sham-infected group compared to the naïve mice corneas. ANOVA: p < 0.0001; *p = 0.01; **p = 0.004 was calculated by one-way ANOVA followed by Tukey's multiple comparison test. ¥, demonstrates the statistical significance (p < 0.05) between the marked column and the sham-infected controls. www.nature.com/scientificreports/ and McKrae) 22 did not demonstrate any titer-dependant severity of the disease. Corneal epithelial defects at 3 dpi were at its highest for both Hd and Ld HSV-1 KOS-63-infected corneas; this is in accordance with a significant rise of inflammatory cells (CD45 + cells) in the central cornea at 3 dpi, as compared to naïve corneas. Significantly larger epithelial defects were observed in the Hd HSV-1 KOS-63 infected corneas as compared to the Ld group.
Higher corneal viral titer in the Hd KOS-63 infected group may, in part, explain the increased severity of clinical signs observed in this group, as compared to Ld KOS-63 infected group. Similarly, mice that developed herpetic stromal keratitis (HSK), using RE strain, showed higher viral titer in their tear films (although not significant) compared to those without HSK 44 .
We demonstrate increased number of infiltrating leukocyte (CD45 + cells) in corneas of both KOS-63 infected groups. We have previously shown an increased influx of CD45 + cells in the cornea after HSV-1 McKrae infection 10,45 . Here we show that although KOS-63-infected corneas have higher number of CD45 + cells at 3 and 5 dpi compared to McKrae infected mice, their density is lower at 7 dpi, suggesting an earlier, but less severe inflammatory response with KOS-63 infection. Interestingly, analysis of corneal leukocyte subpopulations is different between Hd and Ld KOS-63-infected mice. CD11b + cells are suggested to be involved in causing blepharitis in HSV-1 infected mice 8 . Therefore, a higher percentage of CD11b + cells in the Hd KOS-63-infected mice might explain the more severe blepharitis in the Hd as compared to the Ld KOS-63-infected group. Neutrophils are a source of vascular endothelial growth factor (VEGF)-A and have a crucial role in inducing angiogenesis following HSV-1 ocular infections 46 . Interestingly, we show a larger infiltration of CD45 + Gr-1 + cells into the cornea of the Hd KOS-63-infected mice (71.6%) as compared to the Ld-infected group (26.3%), which can explain the higher neovascularization score in the former group. NK cells and macrophages play a role in the control and clearance of the virus from the cornea 6 . Percentages of both cells increase at 3 dpi as compared to naïve mice. CD11c + DCs are believed to control the virus replication in the cornea indirectly by increasing migration of NK cells to the center of cornea 47 and by systemic dissemination of the virus to the TGs, thus decreasing the virus titer in the cornea 10 . Interestingly, we have recently demonstrated that depletion of CD11c + cells resulted in more severe acute HSV keratitis 10 . The current study shows that Hd-infected mice have a lower frequency of CD11c + cells among CD45 + immune cells (28.7%) as compared to Ld-infected mice (74.1%) at 3 dpi, which can partly explain the higher viral titer in the cornea and more sever keratitis in the Hd-infected mice. It is plausible that in the presence of a lower viral load, DCs can increase and protect the cornea from severe damage, while a higher viral load could prevent increased DC density and therefore cause more severe clinical symptoms.
Our data shows that both central and peripheral subbasal corneal nerve density decreases gradually over the first 7 days after inoculation, with more severe loss in the central cornea and in Hd KOS-63-infected mice. This has been in concordance with significant decreased corneal sensation at 5 and 7 dpi. Corneal nerve damage following HSV infection has been well demonstrated in clinical and preclinical studies 7,10,15,16,48,49   www.nature.com/scientificreports/ reduction of corneal nerve fibers in BALB/c mice at 28 dpi with the HSV-1 KOS strain 16 , albeit without quantification of nerve density. Furthermore, quantification methods for nerve density have been qualitative in the latter two studies and are thus not comparable to the quantitative method used herein. The exact cause of denervation has not been elucidated yet, although direct nerve damage by the virus and viral-induced inflammation could be leading causes 10,16 . Based on the current study, higher corneal nerve loss (at 3 dpi) in the Hd-infected group compared to the Ld group can be explained either by a higher corneal viral titer and/or by a higher number of PMNs and NK cells in the Hd-infected mice compared to the Ld-infected mice. On the other hand, more severe nerve loss in the Hd-infected mice as compared to Ld-infected mice may explain more severe neovascularization and immune infiltration into the cornea and not necessarily direct effect of the virus. We have shown an inverse correlation between corneal nerve density and immune cell infiltration in the cornea in clinical and pre-clinical studies 50 . Tepelus et al. also found similar inverse correlation in patients with dry eye disease 51 . Interestingly, Yun et al. 16 concluded that corneal nerve damage can be maintained even without significant corneal inflammation after HSV-1 infection. These studies assessed nerve loss from different perspectives and still the cause of nerve damage in HSV keratitis remains controversial. It has been previously reported that the HSV-1 KOS is less virulent than strain 17 and McKrae and has mutated US9 and US8A genes as compared with strain 17 52,53 . Different variants of KOS have been used in laboratories up to date, including KOS-63, KOS-79, and KOS1.1 (KOS (M)) 21,54 . In addition, different mutant derivatives of HSV-1 KOS have been used to study the virulence, latency, and reactivation of HSV-1 55-57 . Bowen et al. 54 demonstrated that despite more than 99.2% similarity among different passages of a same virus (HSV-1 KOS-63), there are small genomic differences. Specifically, the HSV-1 KOS-63 strain that was used for the first time by Dix et al. 21 is different in 3 genes (proteins), gB (UL27), VP1/2 (UL36), and RR (UL39), as compared to the HSV-1 KOS-63 from Kinchington group (which is used in this study) 54 . UL36 and UL39 proteins were shown previously to influence virulence and replication of HSV-1 virus, respectively 58 . Thus, different sub-strains, mutants, and maybe different passages of HSV-1 KOS used by different laboratories may provide additional explanation for variable clinical scores and the controversy in the reported literature.
Corneal wounding may result in activation of resident leukocytes and infiltration of leukocytes to the cornea 59 . Therefore, in order to eliminate the possibility that scarification creates a confounding factor, we used shaminfected mice as an additional control group, in which scarified corneas were inoculated with PBS. We demonstrate that superficial corneal scarification, which is much less extensive than corneal wounds described by Stepp et al. 59 do not lead to any significant corneal disease, immune cell infiltration, or corneal nerve damage. Similarly, BenMohamed et al. 12 and Yuan et al. 60 did not show any significant eye disease and corneal neovascularization post scarification, respectively. Prior studies have shown that HSV-1 McKrae can infect the murine corneas without scarification, although this has not been shown with the KOS strain 12,29 , thus requiring scarification to study the Ld and Hd KOS-63 strain effects. However, in this study in order to exclude any confounding factor we performed scarification prior to corneal inoculation with HSV-1 KOS-63 or HSV-1 McKrae stains. Given that the primary objective of the study was to compare the Hd and Ld HSV-1 KOS-63 corneal infection, these results were compared to the more virulent McKrae strain that was used as a positive control. We used a modified corneal opacity scoring system, with a more detailed scale than mild, moderate and severe opacity, which has been widely used previously. Furthermore, we used clock hours to report clinical neovascularization. Although this method is sensitive to differentiate subtle changes at different time points post infection but it cannot distinguish the extent of neovascularization. Thus, we chose this scoring system, as our goal was to assess virulence at the acute phase of the disease and to tease out minor changes. Moreover, since we followed infected mice for only 7 days to assess the acute phase of HSV keratitis, we cannot extend our dose-dependent findings to the chronic phase of infection. In addition, the plaque assay that we used for virus titration has the detection limit of 10 PFU.
In conclusion, our results are indicative of ocular virulence of HSV-1 KOS-63 in a dose-dependent fashion when applied on scarified corneas of C57BL/6 mice. Moreover, we showed that HSV-1 KOS-63 dose can affect incidence, viral titers in the cornea and TG, immune cell infiltration to the cornea, and finally corneal nerve damage at different time points pi. Therefore, the choice of a suitable viral strain and viral dose, based on the aims of the study, is of great importance. However, whether the viral load or different immune cell infiltrates (or both) are responsible for more sever clinical manifestations and corneal nerve damage in the Hd versus Ld infected mice is not clear.