Herpesvirus entry mediator is a serotype specific determinant of pathogenesis in ocular herpes

Abstract

Infection with herpes simplex virus type 1 (HSV-1) and HSV-2 is initiated by viral glycoprotein D (gD) binding to a receptor on the host cell. Two receptors, herpesvirus entry mediator (HVEM) and nectin-1, mediate entry in murine models of HSV-1 and HSV-2. HVEM is dispensable for HSV-2 infection of the vagina and brain, but is required for WT pathogenesis of HSV-1 infection of the cornea. By challenging WT and HVEM KO mice with multiple strains of HSV-1 and HSV-2, we demonstrate that without HVEM, all HSV-1 strains tested do not replicate well in the cornea and infection does not result in severe symptoms, as observed in WT mice. In contrast, all HSV-2 strains tested had no requirement for HVEM to replicate to WT levels in the cornea and still cause severe disease. These findings imply that HSV-2 does not require HVEM to cause disease regardless of route of entry, but HVEM must be present for HSV-1 to cause full pathogenesis in the eye. These findings uncover a unique role for HVEM in mediating HSV-1 infection in an area innervated by the trigeminal ganglion and may explain why the presence of HVEM can lead to severe inflammation in the cornea. Thus, the dependence on HVEM is a dividing point between HSV-1 and HSV-2 that evolved to infect areas innervated by different sensory ganglia.

Herpes simplex virus type 1 (HSV-1) and HSV-2 are common human pathogens with over 50% of United States adults seropositive for HSV-1 and over 15% seropositive for HSV-2 (1). Typically, HSV-1 is responsible for oral infections and HSV-2 is the cause of genital infections, but this division is becoming less distinct (2–4). Severe outcomes caused by both serotypes include encephalitis, stromal keratitis, and meningitis (4, 5). Importantly, HSV-1 remains a significant cause of ocular morbidity and blindness despite the availability of the antiviral drug, acyclovir (6, 7).

Herpes infection begins when the viral envelope fuses with a host cell membrane. This process is dependent on envelope glycoprotein D (gD) binding to one of its receptors. The gD-receptor binding event triggers glycoproteins gB, gH, and gL to execute fusion by a yet unknown mechanism (8). The known gD receptors include: herpesvirus entry mediator (HVEM) (9), nectin-1 (10), nectin-2 (11), and 3-O-sulfated-heparan sulfate (3-O-S-HS) (12). Of these receptors, HVEM and nectin-1 have proven important in in vivo murine models of HSV infection (13–15). HVEM is a member of the TNF receptor superfamily and participates in a wide variety of cellular and physiological functions, including apoptosis (16) and immune cell signaling (17). Nectin-1 is a member of the Ig superfamily and functions in cell-cell adhesion (18).

The importance of HVEM and nectin-1 in mediating HSV infection in vivo was demonstrated by the use of KO mice. HSV-2 vaginal infection was completely prevented in HVEM/nectin-1 double-KO mice. Interestingly, the absence of HVEM alone was not protective in the vaginal model of HSV-2 infection, and similar pathology and viral spread was observed in WT and HVEM KO mice. HSV-2 caused an attenuated disease in nectin-1 KO mice and less virus was able to spread to the nervous system compared with HVEM KO and WT mice (13). Nectin-1 was shown to be essential for HSV-2 infection of the CNS by direct cranial inoculation, but HVEM was dispensable for infection via this route (14). In contrast to the vaginal and cranial model of HSV-2 infection, we previously published that both HVEM and nectin-1 are important for pathogenesis of HSV-1 infection of the cornea (15). No infection of the cornea was observed in HVEM/nectin-1 double-KO mice, but reduced viral titers were observed in eye swabs and tissues of HVEM KO and nectin-1 KO compared with WT mice. Importantly, only WT mice developed symptoms of HSV-1 infection including periocular lesions and symptoms of CNS infection. Similar to the vaginal HSV-2 model, the absence of nectin-1 had a greater impact on the virus’s ability to spread to the nervous system than HVEM.

These unexpected results suggest that the HSV receptor requirements may depend on the route of infection or the viral serotype being studied. To address the possibility that the presence of HVEM may impact the pathogenesis of HSV-1 and HSV-2 differently, we challenged both WT and HVEM KO mice with three strains of HSV-1 and HSV-2 to compare viral replication and pathology. Interestingly, our results indicate that unlike HSV-1, HSV-2 does not require HVEM to produce disease in cornea infections.

Results

HSV-2 Replicates to Similar Levels in WT and HVEM KO Mice After Corneal Infection.

To determine if the absence of HVEM impacts the ability of HSV-2 to infect the cornea we challenged adult male WT and HVEM KO mice via cornea scarification with HSV-2 strain G (HSV-2/G), a laboratory strain originally isolated from a genital lesion (19). Eye swabs were collected on days 1, 2, and 3 postinfection and viral titer was measured via a standard plaque assay. To assess whether or not the absence of HVEM would impact viral spread to adjacent tissue and the nervous system mice were killed on day 3 postinfection, periocular skin (POS), trigeminal ganglia (TG), and brains were removed, homogenized, and titered as detailed in Materials and Methods. Unlike experiments with HSV-1, viral titers from HVEM KO eye swabs were significantly higher than those from WT mice only on day 1 postinfection (Fig. 1A). There was no significant difference between genotypes on days 2 or 3 postinfection. Additionally, viral titers from POS, TG, and brains were not significantly different between the two groups (Fig. 1B). This result is in contrast to what we observed in our previous study using HSV-1 strain 17 (HSV-1/17) (15) where viral titers from HVEM KO eye swabs were significantly lower than WT mice on days 1, 3, and 5 postinfection. To ensure that this phenomenon was not strain-specific, we repeated the experiment with HSV-2 strain 186 (HSV-2/186), which was originally isolated from a genital lesion (20). The results from the HSV-2/186 experiment were remarkably similar to the results from the HSV-2/G experiment. There was not a significant difference between viral titers from WT eye swabs and HVEM KO eye swabs after HSV-2/186 infection (Fig. 1C). Similarly, differences in viral titer from POS, TG, and brains of WT and HVEM KO mice infected with HSV-2/186 were not significant (Fig. 1D).

Fig. 1.

HSV-2 replication in corneas and tissue of WT and HVEM KO mice. Mice were inoculated with 2 × 10⁶ pfu per eye as detailed in Materials and Methods. Eye swabs were collected on days 1, 2, and 3 postinfection and titered. Animals were killed on day 3 postinfection and tissues were removed, homogenized and titered. Individual samples (unfilled shapes) and averages (filled shapes) from WT (circles and solid line) and HVEM KO (squares and dotted line) are shown (n ≥ 5 mice for each group). Asterisks represent a P value of less than 0.05 calculated using a two-tailed Student t test. (A) Eye swabs from mice infected with HSV-2 strain G. (B) Tissue titers from mice infected with HSV-2 strain G on day 3 postinfection. (C) Eye swabs from mice infected with HSV-2 strain 186. (D) Tissue titers from mice infected with HSV-2 strain 186 on day 3 postinfection.

HSV-1 Produces Lower Viral Titers in HVEM KO Mice After Corneal Infection.

Our previous study demonstrated that HSV-1/17 infection was attenuated in HVEM KO mice after corneal inoculation (15). Given that two HSV-2 strains showed no difference between WT and HVEM KO, mice we wanted to confirm that observations using HSV-1/17 were not strain-specific. We then challenged WT and HVEM KO mice with HSV-1 strain F (HSV-1/F), a common laboratory strain originally isolated from an oral lesion (19), and HSV-1 strain KOS (HSV-1/KOS), a common laboratory strain originally isolated from a lip lesion (21). After infection with HSV-1/F fewer pfu were recovered from eye swabs of HVEM KO mice compared with WT on average (Fig. 2A). This difference was statistically significant (P < 0.05) on days 1 and 2 postinfection, but not on day 3 postinfection. However, viral titers from POS and TG recovered on day 3 postinfection from the HVEM KO group were significantly lower than those from the WT group (Fig. 2B). HVEM KO mice infected with HSV-1/KOS produced lower viral titers in eye swabs compared with WT controls on average (Fig. 2C). This difference was statistically significant on days 2 and 3 postinfection. Titers from POS and TG from HVEM KO mice infected with HSV-1/KOS were also significantly lower on average than POS and TG from WT mice (Fig. 2D). These results from HSV-1/F and HSV-1/KOS experiments were similar to those obtained previously using HSV-1/17 where eye swab and tissue titers from HVEM KO mice were significantly lower than those from WT mice (15), confirming that HVEM contributes to HSV-1 infection of the eye.

Fig. 2.

HSV-1 replication in corneas and tissue of WT and HVEM KO mice. Mice were inoculated and eye swabs and tissue samples were collected as described in Fig. 1. Samples, averages, and genotypes are represented as in Fig. 1 (n ≥ 5 mice for each group). Asterisks represent a P value of less than 0.05 calculated using a two-tailed Student t test. (A) Eye swabs from mice infected with HSV-1 strain F. (B) Tissue titers from mice infected with HSV-1 strain F on day 3 postinfection. (C) Eye swabs from mice infected with HSV-1 strain KOS. (D) Tissue titers from mice infected with HSV-1 strain KOS on day 3 postinfection.

HSV-2 Causes Severe Disease in Both WT and HVEM KO Mice Following Cornea Infection.

Previously, HVEM KO mice were shown to be resistant to symptoms of disease (periocular lesions, neurological morbidity, and death) caused by HSV-1/17 in addition to producing lower viral titers in eye swabs and tissue titers (15). Given that no significant difference was observed in viral titers from HVEM KO and WT animals infected with HSV-2 (Fig. 1), we wanted to investigate the possibility that the absence of HVEM would not protect mice from symptoms of HSV-2 infection. WT and HVEM KO mice were infected with HSV-2 strain 333 (HSV-2/333), originally isolated from a genital lesion and kept at low passage (22), and followed for at least 21 d. All mice (both WT and HVEM KO) developed extra ocular lesions by day 8 postinfection (Fig. 3A). The severity of the lesions was similar between the WT and HVEM KO groups and there was no significant difference between the average maximum lesion score achieved (Fig. 3 B and C). Similarly, both WT and HVEM KO mice developed symptoms of neurological infection including limb weakness, weight loss, hunched posture, ruffled fur, and unsteady gait. The course of these symptoms was similar between mouse genotypes, there was no significant difference in the average maximum neurological score achieved, and no significant difference in the percent body mass lost (Fig. 3 D–F). Some mice from each genotype ultimately succumbed to the infection or developed such severe neurological symptoms or weight loss that they had to be killed in accordance with our animal protocol (Fig. 3G). These findings were in contrast to previous observations where HVEM KO mice after HSV-1/17 infection did not develop lesions, neurological symptoms, or die from the infection. Interestingly, lower viral titers from eye swabs from HVEM KO mice were observed and this difference was significant at some time points (Fig. 3H). There was no difference in tissue titers on day 3 postinfection (Fig. 3I). On repeated experiments, variable results from eye swab titers were obtained. In some studies no significant difference in eye swab titer was observed but in others the difference was significant. However, weight loss, mortality, lesion scores, and neurological scores remained consistent with no significant differences between the genotypes.

Fig. 3.

HSV-2/333 replication and symptoms in WT and HVEM KO mice. Mice were inoculated as described in Fig 1. Eye swabs were collected on days 1, 3, 5, and 7 postinfection. Some mice from each group were killed on day 3 postinfection and indicated tissues were removed, homogenized, and titered. Each mouse was weighed daily beginning on the day of inoculation and each mouse was given a lesion score (0–5, 5 being the most severe) and neurological disease score (0–5, 5 being the most severe) each day. Asterisks represent a P value of less than 0.05 calculated using a two-tailed Student t test. Error bars (when present) are one SD from the mean. Groups are represented as described in Fig. 1. Results from a representative experiment are displayed (n = 5 mice per group). (A) Percent of mice without lesions. (B) Average lesion score for each group on days 1 through 14 postinfection. (C) Average maximum lesion score achieved in the first 14 days postinfection. (D) Average maximum weight loss as a percent of starting weight. (E) Average maximum neurological disease score achieved in the first 14 d postinfection. (F) Average neurologic disease score for each group on days 1 through 14 postinfection. Mice that reached a score of 5 and had to be killed were scored as a 5 for the remainder of the experiment. (G) Survival curve. (H) Eye swab titers from WT and HVEM KO mice. (I) Tissue sample titers from day 3 postinfection.

HVEM KO Mice Are Resistant to Severe Disease Caused by HSV-1 Strain McKrae.

HSV-2 strains tend to cause more severe disease in animal models than HSV-1 strains (23, 24), which could explain why HVEM KO mice were not resistant to clinical disease caused by HSV-2/333. Therefore, we challenged HVEM KO and WT mice with the highly virulent HSV-1 strain McKrae (HSV-1/McKrae), originally isolated from a patient with ocular herpes (25, 26), and monitored them for at least 21 d. Similar to other HSV-1 strains tested, titers of eye swabs from HVEM KO mice were significantly lower than those from WT mice (Fig. 4A). HSV-1/McKrae infection led to the death of some (two of six) of the WT mice but none of the HVEM KO mice (Fig. 4B). Unlike HVEM KO animals challenged with HSV-2/333, HVEM KO mice inoculated with HSV-1/McKrae developed an attenuated disease compared with WT controls. All WT mice developed lesions after infection with HSV-1/McKrae, whereas only two of six HVEM KO animals did (Fig. 4D). The average maximum lesion score achieved was lower for HVEM KO mice, but this was not statistically significant (P = 0.08) (Fig. 4E). The average lesion score was lower for the HVEM KO group on each observation day (Fig. 4F). The most striking difference between the groups was in symptoms of neurological disease and weight loss. HVEM KO mice lost significantly less weight than WT (Fig. 4C), and exhibited almost no signs of neurological disease (Fig. 4 G and H). On the other hand, WT mice infected with HSV-1/McKrae exhibited neurological disease symptoms of a similar severity to WT mice infected with HSV-2/333 (Figs. 3 E and F, and 4 G and H).

Fig. 4.

HSV-1/McKrae replication and symptoms in WT and HVEM KO mice. Mice were inoculated as described in Fig 1. Eye swabs were collected on days 1, 3, 5, and 7 postinfection. Each mouse was weighed every day beginning on the day of inoculation and each mouse was given a lesion score (0–5, 5 being the most severe) and neurological disease score (0–5, 5 being the most severe) each day. Asterisks represent a P value of less than 0.05 calculated using a two-tailed Student t test. Error bars (when present) are one SD from the mean. Groups are represented as described in Fig. 1. (n = 6 per group). (A) Eye swab titers from WT and HVEM KO mice. (B) Survival curve. (C) Average maximum weight loss as a percentage of starting weight. (D) Percentage of mice without lesions. (E) Average maximum lesion score achieved in the first 14 d postinfection. (F) Average lesion score for each group on days 1 through 14 postinfection. (G) Average maximum neurological disease score achieved in the first 14 d postinfection. (H) Average neurologic disease score for each group on days 1 through 14 postinfection. Mice that reached a score of 5 and had to be killed were scored as a 5 for the remainder of the experiment.

HSV-1 Replicates to Similar Levels in the Vaginas of WT and HVEM KO Mice.

To test whether or not HVEM is required for WT infection in other models of HSV-1 infection, WT and HVEM KO mice were challenged intravaginally with HSV-1/F and HSV-1/17. Animals were monitored for symptoms of infection and vaginal titers were determined. HSV-1/F caused no symptoms of disease in either WT or HVEM KO mice and HSV-1/17 caused only mild vaginal redness that was similar in both genotypes. Vaginal titers were not statistically different between WT and HVEM KO mice with either strain tested, indicating that there is no dependence on HVEM for HSV-1 to replicate in the murine vagina (Fig. 5).

Fig. 5.

HSV-1 replication in vaginas of WT and HVEM KO mice. Mice were inoculated with 6.0 × 10⁵ pfu of HSV-1/F in 20 μL (A) or 5.2 × 10⁷ pfu of HSV-1/17 in 20 μL (B). Samples, averages, and genotypes are represented as in Fig. 1 (n ≥ 5 mice for each group). (A) Vaginal wash titers from WT and HVEM KO mice infected with HSV-1/F. (B) Vaginal swab titers from WT and HVEM KO mice infected with HSV-1/17.

Discussion

These studies compared replication and pathogenesis of HSV-1 and HSV-2 strains in the presence and absence of the HSV receptor HVEM. The data presented demonstrate that HSV-1, regardless of strain, does not replicate in the cornea as well, nor cause severe disease without HVEM. This conclusion is based on previously published results (15) using HSV-1/17 in combination with the current studies that used HSV-1/KOS, HSV-1/F, and HSV-1/McKrae. In each case viral titers from eye swabs and tissue samples were significantly lower in HVEM KO mice compared with WT mice. Additionally, HVEM KO mice did not develop severe disease after HSV-1 challenge, whereas WT mice did. This finding is evident because only WT mice died from the infection; infected HVEM KO mice lost less weight, exhibited negligible neurological symptoms, and not all HVEM KO animals developed lesions. These results point to something common and essential about HSV-1 strains in their dependence on HVEM to cause ocular disease because: although the four strains tested vary widely in their virulence, a similar attenuation in HVEM KO mice was obtained in all cases (25, 27, 28). Interestingly, we also tested two HSV-1 strains in vaginal infection and found no dependence on HVEM. Both HSV-1/F and HSV-1/17 replicated to similar levels in WT and HVEM KO mice. These data indicate that HVEM is uniquely required for HSV-1 to replicate to WT levels in the cornea and that HSV-1 does not require HVEM for productive infection in the vagina.

Unlike HSV-1, our results indicate that even in the absence of HVEM, HSV-2 can replicate and cause disease that is very similar to that observed in WT animals after corneal inoculation. This finding is based on the observation that there was no significant differences in viral titer from eye swabs between WT and HVEM KO groups from two HSV-2 strains tested, and when followed for several weeks WT and HVEM KO mice showed similar clinical symptoms, including percentage of mice with lesions, lesion scores, neurological disease scores, and weight loss. These results mirror those found in two other models of HSV-2 infection using receptor deficient mice. After both vaginal and direct cranial infection with HSV-2, WT and HVEM KO mice were indistinguishable with regard to spread of virus, disease symptoms, and mortality (13, 14). Therefore, it appears in murine models of infection HVEM does not play a significant role in HSV-2 disease pathogenesis.

There is no straightforward explanation for the discrepancy between the outcome of HSV-1 and HSV-2 infection in HVEM KO mice after corneal infection. The primary interaction between the virus and HVEM is thought to be through gD binding to initiate entry. A simple explanation would be that HSV-1 gD has a greater preference for HVEM than HSV-2 gD, but previously published work does not support this conclusion. Both viruses use HVEM and nectin-1 for entry in a variety of in vitro and in vivo settings and the region of gD responsible for HVEM binding is identical at the amino acid level between viruses except for position 7 (HSV-1 contains an alanine and HSV-2 contains a proline) (29–31). Studies focused on the differences in receptor use between the two serotypes have not demonstrated a profound difference in the ability to use HVEM (31). Serotype-specific receptor use has been reported for the receptors 3-O-S-HS and nectin-2; HSV-1 can use 3-O-S-HS as a receptor, but not nectin-2, whereas HSV-2 can use nectin-2, but not 3-O-S-HS (29). These differences are unlikely to account for our findings because HVEM/nectin-1 double-KO mice cannot be infected with HSV-1 or HSV-2 regardless of route tested, suggesting that 3-O-S-HS is not important in murine models, and murine nectin-2 is not a receptor for either HSV-1 or HSV-2 (32). Although untested, the possibility remains that HVEM is more important for cell-to-cell spread during in vivo eye infections.

Interactions between gD and HVEM may impact the course of the infection apart from allowing the virus to enter cells. As a member of the TNF receptor superfamily, HVEM binds to several natural ligands including LIGHT, LTα, B- and T-lymphocyte attenuator (BTLA), and CD160 and can either provide immune stimulatory or inhibitory signals, depending on which cell expresses HVEM and which ligand it binds (33). Structural studies have shown that the contact residues between HVEM and BTLA are very similar to gD contact points, suggesting that gD is a viral mimic of BTLA (16), which could enable gD to prevent HVEM from interacting with LIGHT, BTLA, and CD160, allowing gD to act as a multifunctional inhibitor of the host derived HVEM ligands and resulting in diverse outcomes because of the complexity of HVEM signaling (34). In line with this idea, HVEM-gD interactions can inhibit apoptosis, enhance NF-κB activation in cells expressing HVEM, and decrease the expression of HVEM on the surface (16, 35). In vivo evidence for the importance of gD-HVEM interactions in signaling also exists. WT mice challenged vaginally with a mutant HSV-2 unable to bind to HVEM led to greater chemokine production early after infection than mice challenged with a WT control virus, suggesting that the virus specifically targets HVEM to alter the immune response in vaginal infections (36). Early events triggered by gD binding to HVEM could favor viral replication and may be more important for the pathogenesis of HSV-1 than HSV-2 in the cornea. A previous study comparing the ability of HSV-1 and HSV-2 strains to cause herpes virus stromal keratitis found that HSV-2 expresses significantly lower levels of gD and gB on the surface of corneal epithelial cells after infection than HSV-1 (37). Perhaps HSV-1 relies on this abundance of gD on the cell surface to alter HVEM function in the cornea and HSV-2 uses alternative strategies.

In summary, we have shown that HSV-1 strains cause only an attenuated infection in the cornea in the absence of HVEM. This infection is characterized by reduced viral titers from eye swabs and tissues, fewer animals developing periocular lesions, and absence of severe symptoms, including neurological morbidity and death. In contrast, the absence of HVEM does not prevent neurological symptoms and death after HSV-2 infection and does not reduce viral titer in two of three strains tested. Although the lower titers observed in HSV-2/333 and slightly lower disease scores may indicate that HVEM has some impact on HSV-2 pathogenesis, it is certainly not to the extent seen in HSV-1 infections. The difference between HSV-1 and HSV-2 dependence on HVEM is unlikely to be purely because of increased virulence of HSV-2 strains, because HSV-1/McKrae and HSV-2/333 caused similar mortality and disease scores in WT mice, but HSV-1/McKrae was attenuated in HVEM KO mice. Given that gD sequences do not vary greatly in clinical isolates (30, 38), and that simple point mutations can selectively eliminate HVEM binding, it is reasonable to assume that gD binding to HVEM remains an important part of the viral life cycle in its natural host having been selected by the cohabitation of HSV-1 with humans. This viral–host interaction appears to be acutely important during HSV-1 infection of the cornea given that HVEM is dispensable for HSV-2 infection of the cornea. This finding is not entirely surprising given that the cornea and the vagina are extraordinarily different tissues and HSV-1 has evolved to infect areas innervated by the TG. Therefore, it is possible that the expression of HVEM in different tissues is one factor that led to the evolutionary split between HSV-1 and HSV-2. Previously, why HSV-1 and HSV-2 even bound to HVEM was puzzling because HVEM only appeared to be important in the absence of nectin-1. Here we have uncovered a unique situation in which HVEM is crucial to pathogenesis. Our present studies provide a model system to investigate why HSV uses two unique receptors and how the relatively small differences between HSV-1 and HSV-2 can lead to different disease outcomes depending on the route of entry. Future studies should focus on defining how gD binding to HVEM influences pathogenesis as both an entry route and potential immune modulator in the different routes of infection. The outcome of such studies may provide means to treat HSV infections of the eye.

Materials and Methods

Ethics Statement.

This study was performed in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. The protocol was approved by the Committee on the Ethics of Animal Experiments of the Northwestern University (Protocol Number: 2012–1728). Where noted, all procedures were performed under ketamine/xylazine anesthesia and every effort was made to minimize suffering.

Cells and Viruses.

HSV-1 strain McKrae was generously given to us by James Hill (Louisiana State University School of Medicine, New Orleans, LA). All other viral strains came from laboratory stocks originally compiled by Patricia Spear (Northwestern University Feinberg School of Medicine, Chicago, IL). Virus was propagated in Vero cells cultured in DMEM with 1% FBS. When the entire monolayer showed cytopathic effect the cell culture media and cells were collected, sonicated, and frozen at −80 °C. The stock was aliquoted, titered, and used for infection of mice. Standard plaque titrations were performed on Vero cells.

Animal Procedures.

Animals were cared for and procedures were performed in accordance with institutional and National Institutes of Health guidelines and approved by the Animal Care and Use Committee at Northwestern University. Mice were kept in a specific pathogen-free environment until infection and then were transferred to a containment facility. WT C57BL/6 from Jackson Labs (WT) and Tnfrsf14^−/− (HVEM KO) (39) 8- to 15-wk-old male mice were used in the studies. Mice were anesthetized with an intraperitoneal injection of a mixture of ketamine and xylazine (20:3) and checked for the absence of a foot-pad reflex before inoculation. For cornea infections, each cornea was lightly abraded 10 times in a cross-hatched pattern with a 25-gauge needle, and 2 × 10⁶ pfu of HSV in 5 μL was dropped onto each cornea (40). Mice were weighed and monitored for clinical symptoms of infection each day beginning on the day of infection. Each mouse was given a lesion score from 0 to 5 (0 = no lesions; 1 = small area of broken skin < 0.5 cm; 2 = area of broken skin between 0.5 cm and 1 cm; 3 = broken skin, bleeding, scabbing, or pustules; 4 = broken skin > 1 cm with multiple pustules or scabbing; and 5 = severe scabbing or bleeding with pustules) and a neurological score from 0 to 5 (0 = no symptoms; 1 = ruffled fur, hunched posture, but can easily be made to move around; 2 = hunched and slow to move; 3 = hunched, some movement, and labored breathing; 4 = hunched, labored breathing, and little or no movement; 5 = moribund or dead). Eye swabs were collected by lightly anesthetizing the mice with isoflurane, gently proptosing each eye, and wiping a sterile cotton swab three times around the eye in a circular motion and twice across the center of the cornea in an “X” pattern, as previously described (15, 40, 41). The swabs were placed in 1 mL of DMEM containing 5% (vol/vol) FBS, 1% gentamicin, 1% ciprofloxacin, and 1% amphotericin B (referred to as DMER) and stored at −80 °C until titered. Before titering, the swabs were thawed and vigorously vortexed for 30 s. Some mice from each genotype were killed on day 3 postinoculation. POS biopsies were taken using a sterile biopsy punch (Sklar Instruments), and trigeminal ganglia and brains were removed and placed in 1 mL of DMER. These tissues were homogenized, sonicated, and stored at −80 °C until titered (15, 40, 41). Before titering, brain samples were spun at low speed to remove debris. For vaginal infections, anesthetized mice were infected with 20 μL of the indicated titer of virus intravaginally, placed on their backs, and allowed to recover from the anesthesia. Vaginal washes were collected by gently pipetting 20 μL of PBS-GCS (PBS with 1% heat inactivated calf serum and 1% glucose) into and out of the vagina three times. Vaginal swabs were collected by gently inserting a foam-tipped applicator (Puritan Medical Supplies) into the vagina and rotating it 360° three times. The swab was then placed in a vial of DMER as described for the eye swabs above. The washes and swabs were titered on Veros, as described.