Subclinical Herpes Simplex Virus Type 1 Infections Provide Site-Specific Resistance to an Unrelated Pathogen

This article is featured in In This Issue, p.1379

Viral infections of skin or mucosa generally leave a resident memory cell population behind at the site of infection (1). HSV-1 is a common cause of corneal infectious keratitis in which infections of the cornea leave behind a pathogenic immune infiltrate that reduces clarity (2). Ocular HSV-1 pathologies of the cornea can be divided into two general types. The less common, but far more visually debilitating, of the two are lesions of the corneal stroma, termed herpes stromal keratitis (HSK). These lesions are immunologically mediated, can occur in the absence of detectable replicating virus, and can cause permanent visual impairment due to scarring (3). In humans, HSK is mediated by a pathogenic, possibly HSV-1–specific, T cell population (4, 5). More common than HSK is epithelial keratitis. Epithelial keratitis involves pathognomonic dendritic or geographic-shaped lesions in the corneal epithelium, where virus replication and destruction of corneal epithelial cells occurs (6). These lesions generally resolve without permanent visual impairment (7). The long-term subclinical effects of epithelial keratitis on the immune responses of the cornea need to be investigated, because epithelial disease is often the first step toward blinding HSK.

In the murine model, primary HSV-1 infection of corneas results in epithelial lesions that resemble the dendritic or geographic herpes epithelial keratitis seen in humans (2). In mice, as in humans, epithelial lesions do not necessarily progress to HSK (7, 8). Factors contributing to the different pathological outcomes of HSV-1 infections include the initial infectious dose and genotype of both the host and the virus. Indeed, murine HSV infections can be manipulated experimentally to favor certain disease states by varying these factors. For instance, corneal infections of BALB/c mice with the KOS strain of HSV-1 results in sequential development of epithelial lesions followed by HSK, whereas similar infections of C57BL/6 mice cause epithelial lesions without progression to HSK (9, 10). Following resolution of HSV-1 KOS–induced epithelial keratitis, the corneas of the majority of C57BL/6 mice lack any visually detectable signs of infection or inflammation. The immunopathology of HSK is well defined (9–12), but little is known about what, if any, long-term effects the immune response to acute epithelial keratitis has on the corneal microenvironment and its susceptibility to subsequent infections.

It is increasingly apparent that an individual’s health is not maintained simply by the absence of infection, but by good immunological management of infections. Intracellular pathogens induce site-specific immunity in nonlymphoid tissue (13, 14). We hypothesized that the local memory response to acute herpes epithelial keratitis may be beneficial to the host in modulating subsequent infections or trauma. This hypothesis was intuited because of the coevolution of HSV-1 and humanity (15), the ubiquitous nature of HSV-1 infections, and the lack of effective herd immunity (16, 17). Our hypothesis does have some precedent, because chronic γ herpes virus infections induce a systemic inflammation that provides intrinsic resistance to subsequent novel immunological challenges (18), and acute viral infections induce tissue-resident memory (Trm) CD8⁺ T cells that, when activated, provide intrinsic resistance to unrelated pathogens (19). In addition, α herpes infections, such as HSV-1 and HSV-2, are known to leave an immunological footprint at the site of infection (14, 20, 21).

We demonstrate that mouse corneas with resolved HSV-1 epithelial lesions that do not progress to HSK nonetheless maintain a prolonged and significant subclinical immunological response. Unlike other viral infections in the skin or lungs that leave a predominantly CD8⁺-dominated Trm infiltrate (13, 14), resolved corneal infections leave a CD4⁺ T cell–dominated infiltrate and elevated levels of cytokine and chemokine expression but almost no CD8⁺ T cells. We show that the effects of the subclinical inflammation on the health of the cornea are, at best, a double-edged sword. The subclinical infiltrate of the cornea renders it more susceptible to subsequent trauma-induced inflammation but, at the same time, more resistant to subsequent infection with an unrelated pathogen.

Female wild-type C57BL/6 mice were purchased from the Jackson Laboratory (Bar Harbor, ME) and housed in the University of Pittsburgh vivarium. Mice were allowed ≥1 wk of acclamation prior to use in experiments. The use of animals was in accordance with the Association for Research in Vision and Ophthalmology Statement for the Use of Animals in Ophthalmic and Vision Research, and all procedures were approved by the University of Pittsburgh Institutional Animal Care and Use Committee.

The HSV-1 strains KOS (22) (a generous gift from Dr. D. Coen) and HSV-1 RE (23) (a generous gift from Dr. R. Lausch) were maintained as low-passage stocks. Strain identity was confirmed by sequencing the viral DNA at key sites, including a marker diagnostic for KOS (24) and the US9 regions (see GenBank: KF498959.1). Pseudorabies virus (PRV) used in this study was generated using a Becker-strain bacterial artificial chromosome (a kind gift from Lynn Enquist, Princeton University). A thymidine kinase–knockout PRV was generated by insertion of a kanamycin-resistance cassette and monomeric DsRed fluorescent protein (pMred-kan-in) into the thymidine kinase open reading frame, and the resulting sequence, containing the mRed gene, in frame with thymidine kinase was amplified using the following primers: forward, 5′-GCGGCAACCTGGTGGTGGCCTCGCTGGACCCGGACGAGCACATGGCCTCCTCCGAGGACGTCAT-3′ and reverse, 5′-TCAGGTAGCGCGACGTGTTGACCAGCATGGCGTAGACGTTTTATTAGGCGCCGGTGGAGTGGCGG-3′. The kanamycin-resistance cassette was subsequently resolved using red-mediated recombination and virus was then obtained from bacterial artificial chromosome–transfected Vero cells. PRV was subsequently propagated to high titer in PK15 epithelial cells and purified.

HSV-1 infections were carried out using wild-type female C57BL/6 mice between 6 and 12 wk of age. Briefly, the mice were anesthetized by i.p. injection of a mixture of ketamine and xylazine (Henry Schein Animal Health, Dublin, OH) at doses of 100 and 11 mg/kg of body weight, respectively. Subsequently, the corneas of anesthetized mice were scarified using a 30-gauge needle. The scratches on the corneal epithelium were deep enough to reach, but not penetrate, the basement membrane. Three microliters of PBS (mock infection) or PBS containing 1 × 10⁵ PFU of HSV-1 strains KOS or RE or PRV virus lacking thymidine kinase was applied directly to the scarified corneas. The effects of anesthesia were reversed by injecting mice with 10 μg of atipamezole hydrochloride (Henry Schein Animal Health).

Syngeneic (C57BL/6) orthotopic corneal transplantations were performed on age-matched groups of mice. HSV-1 KOS–infected recipient corneas were examined prior to transplantation, and rare mice with neovascularization and/or opacity scores ≥ 1 in any area of the cornea (see scoring below) were excluded from the study. Donor corneal buttons (2 mm diameter) were excised from noninfected C57BL/6 mice and transplanted onto the recipient corneal graft bed. The recipient mice were anesthetized with a mixture of ketamine and xylazine (Henry Schein Animal Health), as described above for corneal infections. The cornea graft was secured with eight interrupted 11-0 nylon sutures (Sharpoint, Black Nylon Monofilament, 5”/13 cm, 0.1 metric, 11-0; Surgical Specialties, Reading, PA). Sutures were removed from the graft at 7 d posttransplant.

Viral replication was assessed by eye swab at various days postinfection (dpi). A sterile tube was filled with 500 μl of sterile HBSS or PBS. A sterile surgical spear (cat. no. 008680; Beaver-Visitec, Sydney, Australia) was soaked in the sterile buffer, and the surface of the cornea was brushed with the swab 20 times. The swab was returned to the tube and stored at −80°C. The number of viral PFU per swab was determined via plaque assay. African green monkey kidney cells (Vero cells; ATCC CLL-81), grown to confluence in a sterile 48-well tissue culture plate (Becton Dickinson, Franklin Lakes, NJ), were used to enumerate the number of PFU per swab. All excess medium was removed from the cell layer. The fluid contents of the swab tubes were diluted in sterile PBS or HBSS and applied to the cell layer in a volume of 100 μl. Each swab was plated in triplicate. The plates were incubated for 1 h before a methylcellulose-containing medium was laid over the cells to prevent viral spread. The infected plates were incubated for 48 h. The medium was removed, and the cells were fixed with 10% Buffered Formalin Phosphate (cat. no. SF100-4; Fisher Scientific, Fair Lawn, NJ) and stained with 1% Gentian Violet (RICCA CHEMICAL, Arlington, TX). The plaques per well were counted, and the numbers of PFU per swab were calculated. The same protocol was used for assessing PRV and HSV-1 replication in the cornea.

Epithelial lesions, indicating viral replication, were assessed 2 dpi with HSV-1 or PRV using Fluorescein sodium (Akorn, Lake Forest, IL) in sterile PBS at 0.1 mg/ml. Fluorescein drops were applied to the eyes, and the lesion was visualized under a blue light using an Olympus SZX16 Stereo Microscope (Olympus America, Center Valley, PA). Lesions were scored on a five-point scale; 0 = no visible fluorescein uptake in the cornea, 1 = punctate lesions in the cornea, 2 = dendritic lesions in the cornea, 3 = geographic lesions and dendritic lesions taking up <25% of the corneal area, and 4 = dendritic and geographic lesion taking up >25% of the corneal surface.

Corneal pathology was scored for three symptoms: neovascularization, corneal opacity, and corneal sensation. Corneal neovascularization was examined using a stereoscope (Model Z30 L; Cambridge Instruments, Somerville, MA) and was scored in each quadrant from 1 to 4 to give a total measurement on a 16-point scale: 0 = no vessels visible, 1 = vessels extending into the corneal bed, 2 = vessels extending 25–50% toward the corneal center, 3 = vessels extending >50% of the way toward the corneal center but not reaching the center of the cornea, and 4 = vessels extending to the central cornea. Opacity was also scored, by visualization of each quadrant, as follows: 0 = clear, 0.5 = small isolated imperfections or regions of translucence, 1 = minimal translucence or haziness extending over region but allowing visualization of the iris, 1.5 = translucence or haziness extending over the quadrant with small punctate regions of opacity, 2 = cloudiness and regional opacity allowing only heavily distorted visualization of the iris, 2.5 = opacity extending over the entire quadrant with punctate regional translucence allowing very limited and strongly obscured visualization of the iris, 3 = complete stromal opacity with no visualization of the iris structures, 3.5 = complete stromal opacity associated with associated pathology, such a granuloma or ulcer, and 4 = perforation of the cornea. Hypoesthesia was measured regionally on a five-point scale, with a score of five representing complete sensitivity and score of zero representing a cornea completely insensitive to topical physical stimuli. Using a sterile plastic probe, the cornea was touched in four peripheral regions and central region. If the mouse responds to the touch of the probe, that region of the cornea is scored as positive.

Scoring of corneal grafts was done in much the same manner, with minor variations that took into consideration the specific nature of the graft. For judging vascularization following grafting, the eye was visually divided in four quadrants, and each quadrant was scored on a four point scale: 0 = no vessels visible, 1 = vessels extend into the corneal bed, 2 = vessels extending up to graft interface, 3 = vessels extending into the graft but not reaching the center of the graft, and 4 = vessels extending to the central graft. Assessment of opacity was limited to the graft only; scores reflect opacity in the graft and not the supporting corneal bed. Graft scoring used a 16-point system as described above. Images of corneas in live mice were acquired using an Olympus SZX16 Stereo Microscope and an Olympus DP80 digital camera (both from Olympus America).

Total RNA from corneas was obtained using an RNeasy Kit, as per the manufacturer’s instructions (QIAGEN, Germantown, MD). The relative levels of specific mRNAs were assessed using a custom designed NanoString array (NanoString Technologies, Seattle, WA) or by real-time PCR. NanoString analysis was performed by applying purified RNA directly to the NanoString chip, as per the manufacturer’s instructions. Relative gene expression was normalized to a panel of housekeeping genes (Table I). Preparation of cDNA synthesis was performed on 2 μg of total RNA prepared from corneal homogenates, using a high-capacity cDNA Reverse Transcriptase Kit (Applied Biosystems). Real-time PCR primers were derived and quantified using a FAM-labeled MGB TaqMan probe for each gene (Applied Biosystems). The probes recognized CXCL10 (Mm_00445255_ml), CXCL9 (Mm_0043946_ml), CCL5 (Mm_01302427_ml), and histone H1 (Mm_00469312_ml).

Viral DNA copy number was assessed with a real-time PCR protocol using a labeled probe [5′-(FAM)TCCGGACCACTTTTC(NFQ)-3′] that recognizes the gene for glycoprotein H. The reaction was carried out using TaqMan Universal PCR Master Mix. Because HSV-1 contains a single copy of glycoprotein H, we compared the cycle threshold values to that of a plasmid standard and then used linear regression to calculate copy number. Briefly, the trigeminal ganglia (TGs) were placed into a DNA-preserving tissue digestion buffer (QIAGEN). The tissue was digested with proteinase K, and DNA was isolated using a QIAGEN DNeasy kit, as per the manufacturer’s instructions. Mathematical and statistical analyses were performed using Applied Biosystems StepOne software (Thermo Fisher Scientific, Waltham, MA) and GraphPad Prism software (GraphPad, La Jolla, CA). Latency-associated transcript (LAT) was quantified using a TaqMan RT PCR assay, as described above, and LAT forward primer 5′−GCATAGAGAGCCAGGCACAAAA-3′ and reverse primer 5′-ACGTACTCCAAGAAGGCATGTG-3′. These and the probe were described and validated previously (25).

For systemic CD4 depletions, mice were injected i.p. with 100 μg of anti-CD4 clone GK1.5 (Bio X Cell, Lebanon, NH) starting 2 d prior to infection. Injections of anti-CD4 (100 μg) were repeated at 1 and 4 dpi and then every 6 d until the time of sacrifice or PRV infection. Local CD4 depletions were carried out via subconjuctival injections using a Hamilton syringe (Hamilton, Reno, NV). Local CD4 depletions (30 μg of clone GK1.5 in a volume of 20–40 μl of sterile PBS) were performed 3 d before PRV infection. Depletion of macrophages was achieved using a mixture of clodronate liposomes (Molecular Cell Biology and Immunology, Amsterdam, the Netherlands) and anti-Ly6C Ab (Monts-1 clone; Bio X Cell). Then 20 μl of liposomes was diluted 1:1 with a sterile PBS mixture containing 70 μg of anti-Ly6C for total volume of 40 μl. The mixture was administered subconjunctivally via a single injection using a Hamilton syringe. We observed no leakage of this volume, and the efficacy of the depletions was confirmed by flow cytometry.

Spleens from naive female C57BL/6 mice were harvested, triturated into a single-cell suspension, and depleted of CD4⁺ T cells using an EASY-SEP PE selection kit and anti-TCRβ (clone H57-597) and anti-CD4 (clone RM4-5), both conjugated to PE. The depleted splenocytes were infected with HSV-1 KOS or PRV at a multiplicity of infection of one and incubated for 8 h at 37°C in RPMI 1640 supplemented with 10% FBS (Atlanta Biologicals). Spleens from female C57BL/6 mice infected with PRV Becker or HSV-1 KOS (as described above) were harvested at 9 or 10 dpi and elutriated into a single-cell suspension, and the cells were counted. CD4⁺ T cells were enriched using an EasySep Mouse CD4⁺ T Cell Enrichment Kit. T cell enrichment was assessed by flow cytometry; cells were stained with PE-Cy7–conjugated anti-CD4 (RM4-5), FITC-conjugated anti-CD8α (53-6.7), allophycocyanin-Cy7–conjugated anti-CD11b (M1/70), and anti-CD44 conjugated to V450 (clone IM7). All Abs were purchased from BD Biosciences (San Diego CA). CD4⁺ T cell purity ranged from 85 to 93%. The enriched CD4⁺ T cells were placed into culture with the infected or uninfected splenocytes for 5 h in the presence of brefeldin A. The cells were treated with anti-mouse CD16/CD32 (Fcγ III/II receptor; 2.4G2; BD Pharmingen, San Diego, CA) to prevent nonspecific Ab binding and then stained with vital dye (LIVE/DEAD Fixable Aqua; Molecular Probes, Eugene, OR) and Abs for the surface markers listed above for 30 min at room temperature in PBS. After staining, the cells were fixed with BD Cytofix/Cytoperm for 20 min at 4°C (cat. no. 554722) and then permeabilized with BD Perm/Wash (cat. no. 554723; both from BD Biosciences). The cells were stained intracellularly with allophycocyanin Rat Anti-Mouse IFN-γ (cat. no. 554413; BD Biosciences). The cells were analyzed by flow cytometry using a live-cell gate.

Female C57BL/6 mice, infected or not with HSV-1 KOS, were infected via the cornea with 1 × 10⁵ PFU of PRV. Mice received an i.p. injection of 100 μg of BrdU (Sigma, St. Louis, MO) at 9 dpi. Four hours postinjection, the mice were anesthetized and exsanguinated. TGs were harvested and incubated for 1 h in 0.2 U Liberase TM (cat. no. 5401127001; Roche). TGs were washed and passed through a 40-μm nylon mesh filter. Draining lymph node (DLN) cells were mechanically worked into single-cell suspensions and passed through a 40-μm nylon mesh filter. Cells were treated with anti-mouse CD16/CD32 (Fcγ III/II receptor; 2.4G2; BD Pharmingen) to prevent nonspecific Ab binding and then stained for various leukocyte surface markers for 30 min at 4°C. BrdU staining was carried out using a BrdU staining and fixation kit (BD Pharmingen), as per the manufacturer’s instructions, with the supplementation of DNase I (Sigma). Staining was assessed on a flow cytometer (FACSAria; BD Biosciences), and analyses were carried out using FlowJo software (TreeStar, Ashland, OR). Lymphocytes from DLNs were gated by size prior to analysis of subpopulations. Cells in TGs were gated on CD45 expression prior to analysis of subpopulations.

Corneas were incubated for 1 h at 37°C in 0.4 U Liberase TM (cat. no. 5401127001; Roche) in a volume of 100 μl of buffer, then triturated with a 200 μl pipette tip and passed through a 40 μM nylon mesh filter. Corneal cells were washed and stained for analysis via flow cytometry. Various panels of Abs were used dependent on the experiment. Corneal cells were treated with anti-mouse CD16/CD32 (Fcγ III/II receptor; 2.4G2; BD Pharmingen) to prevent nonspecific Ab binding and then stained for various leukocyte surface markers for 30 min at 4°C. The following Abs were purchased from BD Pharmingen: PE-conjugated anti–GR-1 (RB6-8C5), FITC-conjugated, PE-Cy7–conjugated, or Alexa Fluor 647–conjugated anti-CD4 (RM4-5), PerCP-conjugated anti-CD45 (30-F11), allophycocyanin- or FITC-conjugated anti-CD8α (53-6.7), and eFluor 450–conjugated anti-CD11b (M1/70). Anti-F4/80 (clone BMB) conjugated to PE-Cy7 was purchased from eBioscience (San Diego, CA). Anti-TCRβ (clone H57-597) conjugated to Alexa Fluor 647 was purchased from BioLegend (San Diego, CA). For analysis of surface markers only, cells were fixed in 1% paraformaldehyde (Electron Microscopy Sciences, Fort Washington, PA), staining was assessed on a flow cytometer (FACSAria; BD Bioscience), and analyses were carried out using FlowJo software (TreeStar). For transcription factor staining, cells were fixed following a surface staining protocol using a Transcription Factor Buffer Set (cat. no. 562725; BD Pharmingen). T-bet was stained with BV421 conjugated to O4-46 clone and Foxp3 was stained using PE conjugated to MF23 clone also obtained from BD Pharmingen.

As described previously, mouse corneas were dissected, fixed, and processed for whole-mount immunohistochemistry staining. Briefly, nerve fibers were stained with primary Abs that included rabbit polyclonal anti-βIII tubulin (cat. no. ab18207) and chicken polyclonal anti-tyrosine hydroxylase (TH; cat. no. 76442) (both from Abcam, Cambridge, MA) or mouse mAb to calcitonin gene-related peptide (CGRP; cat. no. ab81887; Lot GR158523-12). Secondary Abs included Alexa Fluor 488 goat anti-rabbit IgG (H+L) (cat. no. GR233725-3; Abcam), Alexa Fluor 546 goat anti-chicken IgG (H+L) (cat. no. 1618409; Life Technologies, Grand Island, NY), and Alexa Fluor 647 goat anti-mouse IgG1 (cat. no. A-11040, Invitrogen). The corneas were also stained with DAPI (Sigma).

We examined the course of HSV-1 infections in C57BL/6 mice infected with the RE strain, which causes HSK, and the KOS strain, which does not. Both virus strains have a disrupted US9 gene (P.R. Kinchington, unpublished observations), and the two strains cause similar corneal epithelial lesions (Fig. 1A, 1B) and HSK in BALB/c mice (10). Moreover, side-by-side infections of C57BL/6 mice with RE and KOS strains of HSV-1 revealed similar levels of viral replication in the eye (Fig. 1A–C). Replicating virus was not detectable in the tear film beyond 6 dpi (data not shown). The viruses established similar latency in TGs, as measured by the number of viral genome copies and expression of LAT (Fig. 1G, 1H). However, the RE strain induced clinically apparent HSK in the majority of the mice, whereas the KOS strain did not (Fig. 1D, 1E). KOS infection induced a partial reduction in corneal sensitivity (as assessed by blink reflex) (Fig. 1F), in contrast to the rapid and complete loss of sensitivity in RE-infected corneas (Fig. 1F). Consistent with our previous work, only the complete loss of sensitivity observed in RE-infected corneas was associated with the development of severe corneal opacity (Fig. 1E, 1F). Because our goal is to characterize the long-term consequences of a subclinical HSV-1–induced inflammation, we subsequently focused on KOS-infected mice that developed transient epithelial keratitis without progression to HSK.

The leukocytic infiltrate was assessed after HSV-1 KOS clearance (9 and 34 dpi) in C57BL/6 corneas that lacked overt pathology (Fig. 2A, 2B). The leukocytic infiltrate in noninfected corneas was negligible (Fig. 2C). The infiltrate was dominated by CD4⁺ T cells and CD11b⁺ myeloid cells at 9 and 34 dpi; although the number of CD4⁺ T cells did not change significantly between the time points (Fig. 1D), the ratio of CD4⁺/CD11b⁺ cells increased (Fig. 2A, 2B, data not shown). CD11b⁺ cells present at 34 dpi were predominantly F4/80⁺ macrophages and lacked a F4/80⁻C11b^hiGr-1^hi neutrophil population (staining not shown). At 9 dpi, the CD4⁺ population expressed more T-bet⁺ cells, both as a percentage of the population and in terms of mean fluorescence intensity, than was expressed by corneal CD4⁺ T cells at 34 dpi (Fig. 2E–H). Foxp3⁺ regulatory T cells accounted for ∼10% of the CD4⁺ T cell infiltrate at 9 and 34 dpi (Fig. 2E, 2F).

Consistent with the changes in T-bet expression (26), we observed that the levels of the inflammatory cytokine IFN-γ were not elevated above that found in naive mice past 15 dpi (Fig. 2I, one-way ANOVA comparing dpi, followed by the Dunn posttest). However, chemokines, such as CCL5, CCL2, CCL22, CXCL10, and CXCL9, were continually expressed at significantly higher levels in corneas previously infected with HSV-1 (Fig. 2I). Direct analysis of expression at 34 dpi by NanoString (Table I) and RT-PCR found that the above chemokines were expressed at levels significantly greater than in naive mice but that inflammatory cytokines, such as IL-17 and IFN-γ, were not (Fig. 2I, data not shown). Corneal HSV-1 infections are known to change corneal nerve phenotype in addition to function (9, 10, 12). Thus, the corneas were also examined for qualitative changes in the nerves innervating KOS- or RE-infected corneas (Supplemental Fig. 1). As reported previously (10), noninfected corneas contain only sensory nerves (CGRP⁺ neurons), and corneas with severe HSK contain sympathetic (TH⁺), but not sensory (CGRP⁺), nerves. Interestingly, KOS-infected corneas with subclinical inflammation possessed sensory and sympathetic neurons (Supplemental Fig. 1C).

We tested the hypothesis that the subclinical inflammation in HSV-1 KOS–infected C57BL/6 mouse corneas would increase corneal sensitivity to clinical inflammation following surgical trauma. To test this, orthotopic syngeneic C57BL/6 corneal grafts were placed on normal corneal beds or corneal beds with subclinical inflammation 28–34 d after HSV-1 KOS infection. All grafts received a temporary tarsorrhaphy to protect the cornea until the sutures were removed from the graft 7 d after transplant. Examination of the grafts immediately after removal of the tarsorrhaphy revealed significantly more inflammation in HSV-1–infected corneas compared with noninfected corneas (Fig. 3A, 3B). Graft opacity was monitored through 40 d after transplant. The level of opacity was significantly higher in grafts placed on HSV-1 KOS–infected corneal beds with subclinical inflammation compared with those placed on noninfected beds (Fig. 3C, 3D), although neither achieved a level of opacity that would be considered rejection in most studies (27–29).

HSV-1 KOS–infected corneas with subclinical inflammation, the contralateral noninfected cornea, or corneas of noninfected mice were infected with 1 × 10⁵ PFU of PRV. Eyes were swabbed at 1 and 4 d postchallenge, and shed virus was quantified in a viral plaque assay. One day after PRV infection, PRV titers in swabs of previously HSV-1–infected, but clinically uninflamed, corneas were significantly lower than titers in the contralateral noninfected corneas of the same mice or in corneas of mice that were not infected previously (Fig. 4A). Fluorescein staining revealed significantly (p = 0.003) smaller PRV-induced corneal epithelial lesions in corneas previously infected with HSV-1 compared with corneas that were not infected previously (Supplemental Fig. 1D). PRV titers in noninfected corneas contralateral to HSV-1–infected corneas were not significantly different from those in mice that were not infected previously (Fig. 4A). Thus, protection from PRV is not systemic but rather is restricted to the microenvironment of the previously HSV-1–infected, but clinically uninflamed, cornea. Shedding of PRV was undetectable in all treatment groups at 4 and 6 dpi (data not shown). Prior to PRV-1 challenge, the CD45⁺ corneal infiltrate in HSV-1–infected corneas with subclinical inflammation was dominated by CD4⁺ T cells (Fig. 1B); however, at 24 h after PRV infection, the immune infiltrate was dominated by CD11b⁺ cells (Fig. 4B, 4C). The corneal CD11b⁺ population at 24 h post-PRV infection was a roughly 50/50 mix of neutrophils and CD11b⁺ macrophages (Fig. 4C). The mean number of CD45⁺ cells was consistently greater in corneas previously infected with HSV-1 compared with corneas not infected with HSV-1 (Fig. 4B, 4C). This increase was due, in part, to the presence of a CD4⁺ T cell population in HSV-1–infected corneas that was absent from noninfected corneas prior to and after PRV challenge (Fig. 4B, 4C). The size of the overall CD11b⁺ population and the neutrophil population tended to be higher in HSV-1–infected mice, although the difference was not statistically significant in the majority of experiments (Fig. 4C). Mice infected with the attenuated PRV virus developed ocular pathology characterized by corneal hypoesthesia, opacity, and vascularization of the cornea. Interestingly, the severity of that pathology was significantly reduced in mouse corneas that were previously infected with HSV-1 (Supplemental Fig. 1D, 1F).

The fact that CD4⁺ T cells are the dominant immune cell population in the corneas previously infected with HSV-1 KOS and with subclinical inflammation suggested their possible involvement in enhanced PRV clearance. We initially tested whether CD4⁺ T cells are required to maintain the elevated chemokine levels in HSV-1–infected corneas with subclinical inflammation. CD4⁺ T cells were depleted systemically by i.p. injections of anti-CD4 Ab, starting 2 d prior to HSV-1 infection and continuing at weekly intervals through 34 dpi. Flow cytometry demonstrated that there were no CD4⁺ T cells in the corneas of depleted mice at 34 d following HSV-1 infection (data not shown). We found that systemic CD4⁺ T cell depletion dramatically reduced mRNA levels of multiple chemokines in infected corneas that lacked clinically observable inflammation (Fig. 5A). Moreover, CD4⁺ T cell depletion eliminated the enhanced PRV clearance in corneas previously infected with HSV-1 (Fig. 5B).

Because CD4⁺ T cell depletion reduced chemokine levels in HSV-1–infected corneas prior to PRV infection, we proposed that the reduced PRV clearance would be associated with a reduced immune infiltrate in HSV-1–infected corneas of mice depleted of CD4⁺ T cells. The CD4-depleted mice had a CD8⁺ T cell infiltrate in the cornea that was not seen in previously uninfected mice or in HSV-1–infected CD4-replete mice (Fig. 5C). The infiltration of CD8⁺ T cells into the corneas of CD4-depleted mice infected with HSV-1 was described previously (30). Surprisingly, the HSV-1–infected corneas of CD4⁺ T cell–depleted mice showed a higher CD45⁺ infiltrate 24 h after PRV infection compared with HSV-1–infected corneas that were not depleted of CD4⁺ T cells (Fig. 5D).

Because CD4⁺ T cells were required for enhanced PRV clearance in HSV-1–infected corneas, we considered that the enhanced clearance might reflect Ag cross-reactivity between PRV-specific and HSV-1–specific CD4⁺ T cells. However, ex vivo stimulation of T cells from HSV-1– and PRV-infected mice did not reveal cross-reactivity. Thus, CD4⁺ T cells from HSV-1–infected mice responded to HSV-1–infected, but not PRV-infected, APCs, whereas T cells from PRV-infected mice responded to PRV-infected, but not HSV-1–infected, APCs (Fig. 5E).

After CD4⁺ T cells, the next most abundant immune population in HSV-1 KOS–infected corneas is CD11b⁺F4/80⁺Gr-1^Neg-Int macrophages (Fig. 2, data not shown). We considered that CD4⁺ T cells might mediate nonspecific clearance of PRV indirectly by priming a memory macrophage response. To test whether tissue-resident macrophages are required for enhanced PRV clearance, we depleted macrophages from HSV-1–infected corneas prior to PRV infection. Mice with HSV-1–induced subclinical corneal inflammation received subconjuctival clodronate liposomes and anti-Ly6C Ab (Monts-1 clone) or PBS liposomes alone 7 d before PRV infection; corneal infiltrates were assessed 24 h after PRV infection. Clodronate liposome treatment effectively depleted macrophages from HSV-1–infected corneas prior to PRV infection (Fig. 6A). However, depleting macrophages that were resident in the HSV-1–infected corneas prior to PRV infection did not influence the overall CD45⁺ population or the CD11b⁺ population in corneas 24 h after PRV infection (Fig. 6B), and it did not influence PRV clearance from the cornea (Fig. 6C).

In the studies described above, systemic CD4⁺ T cell depletion reduced chemokine levels and eliminated enhanced PRV clearance in HSV-1–infected corneas with subclinical inflammation (Fig. 4). It was not clear whether CD4⁺ T cells were required locally in the cornea or mediated these effects systemically. To address this, C57BL/6 mice were infected with HSV-1 KOS, and, at 28–34 dpi, HSV-1–infected mice and noninfected control mice were locally depleted of CD4⁺ T cells by subconjunctival injection of anti-CD4 mAb or were mock depleted. Three days later the corneas were excised, and flow cytometry showed that dispersed cells lacked detectable CD4⁺ or CD8⁺ T cells (Fig. 7A, 7B).

RNA was extracted from HSV-1–infected corneas 3 d after local CD4⁺ T cell depletion or mock depletion and was analyzed for CXCL9 and CCL5 transcripts by quantitative RT-PCR (qRT-PCR). The levels of CXCL9 and CCL5 mRNA were not reduced in HSV-1–infected corneas 3 d after local CD4⁺ T cell depletion (Fig. 7C). Corneas were infected with PRV 3 d after CD4⁺ T cell depletion or mock depletion of HSV-1–infected or noninfected corneas. Twenty-four hours after PRV infection, dispersed corneal cells were analyzed by flow cytometry (Fig. 7D–F). The CD4⁺ T cell–depleted HSV-1–infected corneas exhibited CD11b^Hi, Gr-1^Hi neutrophil and CD11b^Hi, Gr-1^low macrophage populations that were comparable to the mock-depleted corneas (Fig. 7D–F), but they lacked detectable CD4⁺ or CD8⁺ T cells (Fig. 7E). Local depletion of CD4⁺ T cells eliminated the enhanced clearance of PRV from HSV-1–infected corneas (Fig. 7G).

The cornea is unique among mammalian tissue in that it is continually exposed to the environment, it is not vascularized, and its function requires transparency. Indeed, the cornea has developed a degree of immune privilege in an apparent attempt to prevent loss of vision due to vascularization and opacity (31). However, certain types of infectious and noninfectious trauma can overcome the corneal immune privilege, leading to a clinically inflamed cornea. An interesting example is HSV-1, wherein the clinical outcome of infection varies depending on the genetics of the virus and the host. For instance, corneas of C57BL/6 mice infected with the RE strain of HSV-1 develop a transient epithelial lesion followed by the severe vascularization, leukocytic infiltration, and edema that characterize HSK. In contrast, infection of the same mice strain with the same dose of the KOS strain of HSV-1 causes only an initial epithelial lesion with no clinically detectable inflammatory sequelae (Fig. 1). This phenomenon appears to have a counterpart in humans, because many people periodically shed virus to the cornea asymptomatically (32) and even develop epithelial disease without progressing to HSK (7, 8, 33). In fact, these relatively asymptomatic infections are sufficiently common to make one question whether they might provide some advantage to the host.

In this study, we examined the long-term changes to the microenvironment of the cornea resulting from HSV-1 KOS infections. We showed that these infections, although not causing clinically detectable opacity and vascularization, nevertheless result in significant long-term changes to the cornea. We showed that this includes a persistent infiltrate composed predominantly of CD4⁺ T cells, macrophages, and elevated chemokines. Because these corneal changes are subclinical, they would not appear to compromise vision. We observed and other investigators reported that subclinical HSV-1 corneal infections protect from subsequent infections with more virulent strains of HSV-1 and that protection is superior to immunity provided by i.p. or s.c. infection (34). Moreover, protection is less effective or absent in MHC class II–knockout mice (35). Thus, it is reasonable to conclude that protection is due, at least in part, to residual CD4⁺ T cells and macrophages in HSV-1–infected corneas with subclinical inflammation. We posed two additional questions regarding the subclinical inflammation in corneas following HSV-1 infection: does it protect the cornea from unrelated pathogens expressing heterologous Ags, and does it reduce the threshold for clinical inflammation induced by corneal trauma? We answer both questions in the affirmative.

We first used a syngeneic corneal transplantation model to address the second question. We showed that syngeneic corneal grafts (that normally induce only very mild and transient corneal inflammation) induced more prolonged and robust inflammation when placed on HSV-1 KOS–infected corneal beds with subclinical inflammation at 28–34 dpi. Our current data agree with our previous report, in that opacity in syngeneic corneal grafts placed on previously HSV-1 KOS–infected corneas with subclinical inflammation did not reach a sustained opacity score of 12 (opacity of 3 in each quadrant of the graft) that would be considered a rejected or failed corneal graft (28). However, the inflammation in syngeneic grafts placed on previously HSV-1–infected corneal beds was significant and would be unacceptable in a human graft.

We further demonstrate that residual subclinical inflammation in corneas previously infected with HSV-1 provides a degree of innate resistance to the antigenically unrelated pathogen PRV. The enhanced PRV clearance in corneas with subclinical inflammation was not dependent on pre-existing macrophages within the HSV-1–infected cornea, because clearance was not altered by a protocol that depleted resident macrophages but allowed macrophages to infiltrate after PRV infection (Fig. 6). Thus, although infiltrating macrophages are likely involved in PRV clearance, tissue-resident macrophages present in the corneas previously infected with HSV-1 are not required. However, the enhanced innate resistance to PRV was abrogated by local depletion of CD4⁺ T cells 3 d before PRV infection (Fig. 7). This depletion strategy completely eliminated CD4⁺ T cells from HSV-1–infected corneas 24 h after PRV infection, suggesting that the CD4⁺ T cells that persist in the cornea after HSV-1 KOS infections actively provide intrinsic resistance to subsequent PRV infection. Although a possible explanation is that CD4⁺ T cells that infiltrate the cornea after HSV-1 infection recognize epitopes that are shared by PRV, we and other investigators were unable to show such cross-reactivity in CD4⁺ T cells obtained from HSV-1–infected mice, as demonstrated previously (36–39). Although the relatively small number of CD4⁺ T cells that persist in HSV-1 KOS–infected corneas precluded a direct test of cross-reactivity, our data are inconsistent with antigenic cross-reactivity as the mechanism of enhanced PRV resistance in corneas previously infected with HSV-1. Thus, CD4⁺ T cells that persist in corneas with subclinical inflammation following HSV-1 infection provide innate resistance to unrelated pathogens, likely through interaction with pre-existing or infiltrating macrophages.

We (9) and other investigators (40, 41) demonstrated that infections of BALB/c mice with HSV-1 KOS in the eye or the ear can cause loss of sensation proximal to the site of infection. In the HSK mouse model, pathology is preceded by a complete loss of corneal sensory nerves and blink reflex and is maintained by corneal stromal hyperinnervation by sympathetic nerves (9, 10). In this study, we show that HSV-1 KOS infections of C57BL/6 mice that do not lead to clinically detectable HSK qualitatively change corneal innervation. KOS-infected corneas contain sympathetic and sensory nerves (Supplemental Fig. 1). These changes were reproducibly seen in 14 HSV-1 KOS–infected corneas. We concluded that they were induced by HSV-1, because we never see sympathetic nerves in noninfected corneas (9, 10). In other mouse models, it was demonstrated that nerve damage alone is sufficient to increase inflammation in the cornea and rescind corneal immune privilege (42). It remains to be determined whether these modest changes in corneal innervation and sensitivity following HSV-1 KOS infection of C57BL/6 mice result in or contribute to the subclinical inflammation that persists in the corneas.

Although not the focus of this study, we examined the PRV-induced corneal pathology following infection of HSV-1 KOS–infected and noninfected corneas. Corneas not previously challenged with HSV-1 went on to develop opacity, significant loss of corneal blink reflex, and vascularization, whereas those that were previously infected with HSV-1 KOS showed greatly diminished pathologies (Supplemental Fig. 1E, 1F). Data ancillary to our focus on the corneal microenvironment indicated that mice infected previously with HSV-1 KOS also had smaller adaptive immune responses to PRV, as measured by the number of activated T cells in the DLNs and the total number of proliferating T cells in the TGs and DLNs at 9 d post-PRV infection (Supplemental Fig. 1G–K, data not shown).

Evidence that viral infections can confer innate protection against unrelated pathogens was presented previously. Chronic infections with a murine γ herpes virus induced a memory macrophage population that provided the mouse with an intrinsic resistance to unrelated pathogens. In that same study, it was tested whether HSV-1 could provide innate resistant novel infections. Unlike in our current study, HSV-1 was introduced via i.p. injection, and an unrelated pathogen was introduced through the airway. The investigators were unable to show any nonspecific protective effects of a previous HSV-1 infection against the unrelated pathogen (18). However, they speculated that local HSV-1 infections might provide a site-specific resistance to unrelated pathogens. Another study demonstrated that Trm CD8⁺ T cells remaining in a tissue following acute viral infection can provide intrinsic resistance to heterologous pathogens. However, in that study, the unrelated pathogen was administered along with cognate Ag recognized by the CD8⁺ Trm cells (19). To our knowledge, we are the first to demonstrate a local intrinsic protection mediated by CD4⁺ T cells in response to an apparently subclinical viral infection. The clinical relevance of our findings is underscored by the fact that the protection was observed in the context of a common human infection. Unfortunately, this innate protection comes at a cost, because HSV-1 KOS–infected corneas are also subject to increased inflammation when exposed to nonspecific stimuli, such as surgical trauma.

We thank the Department of Ophthalmology administrative staff. Lori Young, Katie Kozak, and Alice Lang were indispensable in communicating with vendors, regulatory agencies, and other academic departments both in and outside the University of Pittsburgh. We acknowledge Moira L. Geary for assistance in surgeries, animal handling, and the evaluation of clinical disease. Dr. Stephen Harvey assisted in the preparation of RNA and worked closely with the University of Pittsburgh Genomic Research Core. Finally, we thank the University of Pittsburgh Genomic Research Core for assistance with the NanoString analysis.

The authors have no financial conflicts of interest.