The Role of a Glycoprotein K (gK) CD8⁺ T-Cell Epitope of Herpes Simplex Virus on Virus Replication and Pathogenicity

Abstract

purpose. The authors recently reported that a recombinant HSV-1 expressing two extra copies of glycoprotein K (gK) exacerbated corneal scarring (CS) in mice. The authors also identified a peptide, STVVLITAYGLVLVW, within the signal sequence of gK as an immunodominant gK T-cell–stimulatory region both in vitro and in vivo and identified a highly conserved potential CD8⁺ T-cell epitope (ITAYGLVL) within the peptide. In this study, the effect of giving this octamer (8mer) as an eye drop 1 hour before ocular infection with HSV-1 was investigated.

methods. Naive mice and rabbits received the gK 8mer or control peptides as eye drops and were then ocularly infected with HSV-1. Virus replication in the eye and trigeminal ganglia (TG), survival, CS, and relative amounts of gB, gK, CD4, CD8, IFN-γ, and granzyme A/B transcripts were determined in the cornea and TG of infected animals at various times after infection. The effect of the gK 8mer was also analyzed in immunized HLA transgenic mice.

results. The gK 8mer resulted in a short-term significant increase in virus replication in the eyes of BALB/c mice, C57BL/6 mice, and NZW rabbits. gK 8mer treatment also increased viral neurovirulence and viral induced CS in ocularly infected mice. Moreover, in HSV-infected humanized HLA-A*0201 transgenic mice, the gK 8mer epitope induced strong IFN-γ-producing cytotoxic CD8⁺ T-cell responses, as assessed by CD107a/b expression and IFN-γ ELISAs.

conclusions. gK 8mer induced CD8⁺ T-cell responses were unlikely to occur soon enough to account for increased virus replication on day 1 after infection. In contrast, the data are consistent with CD8⁺ T cells being involved in the appearance of CS at late times after infection. In addition, the gK peptide may affect viral replication and innate immune responses through other undefined mechanisms.

Recurrent infection of HSV-1 is the leading cause of infectious corneal blindness in the United States. ^{1 2 3} Periodic ocular reactivation can produce permanent visual disability ranging in severity from blepharitis, conjunctivitis, and dendritic keratitis to disciform stromal edema and necrotizing stromal keratitis. ^{1 4 5} Although the specific immune response leading to corneal scarring (CS) remains an area of controversy, it is well established that HSV-1 induces CS, and hence subsequent HSV-1-induced corneal blindness is due to an immune response to the virus. ^{6 7 8}  

Glycoprotein K (gK) is one of 11 known HSV-1 glycoproteins. ^{9 10 11} It plays a crucial role in virus-induced cell fusion, ^{12 13 14 15} virion morphogenesis, and virus egress. ^{16 17 18} Previously, we demonstrated that immunization of mice with gK but not with any other HSV-1 glycoprotein, significantly exacerbates CS and facial dermatitis after ocular HSV-1 infection. ^{19 20} The exacerbated CS is independent of mouse or virus strains. ²¹ In addition, transfer of IgG from gK-vaccinated mice to naive mice, but not the transfer of total peripheral blood immune cells, results in similar exacerbated CS after ocular HSV-1 infection. ²¹ Of interest, depletion of CD8-expressing cells prevents the gK vaccination from exacerbating CS after ocular infection. ²² Thus, the gK-induced exacerbation of CS appears to require both anti-gK-IgG and CD8-positive cells that do not have to be primed with gK before ocular HSV-1 infection. 

Recently, using a panel of 33 overlapping gK peptides spanning the full-length of gK and four recombinant gK proteins representing various regions of gK, we mapped an immunodominant in vitro and in vivo CD4⁺ and CD8⁺ T-cell-stimulatory region of gK. ²³ This immunodominant region STVVLITAYGLVLVW, is located within the signal sequence of gK and is highly conserved between and among HSV-1 and HSV-2 strains. ^{9 24 25} More recently we showed that compared with wild-type HSV-1, an HSV-1 mutant that expresses three copies of gK instead of one (an extra copy of the gK open reading frame [ORF] was inserted into the mutant under control of each of the 2 LAT promoters), promotes more severe eye disease and dermatitis in ocularly infected BALB/c and C57BL/6 mice. ²⁶ In contrast, HSV-1 recombinant viruses lacking the gK gene induced little or no eye disease in ocularly infected mice compared with their parental wild-type strain McKrae. ²⁷  

Since overexpressing gK produces increased ocular pathogenesis, we thought it would be interesting to see whether the immunodominant gK peptide would produce similar effects. Giving the STVVLITAYGLVLVW peptide as eye drops 1 hour before ocular infection with HSV-1 exacerbated CS in mice. Using prediction algorithms, we had identified an octamer (8mer) peptide within this immunodominant sequence that is highly conserved among and within HSV-1 and HSV-2 strains. ^{9 23} Giving this 8mer as an eye drop 1 hour before ocular infection resulted in (1) a significant increase in virus replication in the eyes of mice on day 1 postinfection (PI) and in rabbit eyes on days 1 to 4 PI; (2) increased virus pathogenicity (CS and death) in mice; (3) elevated levels of gB, gK, IFN-γ, CD8, GzmA, and GzmB mRNAs in infected mouse corneas; and (4) increased levels of viral gK mRNA, but not viral gB mRNA, in TG of latently infected rabbits. In addition, in HSV-1 immunized HLA-A*0201 transgenic mice, the gK 8mer induced strong cytotoxic CD8⁺ T-cell activity and IFN-γ production as assessed by CD107a/b expression and IFN-γ ELISAs (ELISpot; BD-Pharmingen, San Diego, CA). Taken together, these results suggest that similar to overexpression of full length gK, this gK 8mer, given in conjunction with ocular HSV-1 infection can exacerbate HSV-1-induced CS. In addition, this gK 8mer may be a pathogenic CD8⁺ T-cell epitope that somehow increases HSV pathogenicity. 

Materials and Methods

Viruses and Cell Lines

Rabbit skin (RS) cells used for preparation of virus stocks and culturing of mouse and rabbit tear films were grown in Eagle’s minimum essential medium (MEM) supplemented with 5% FCS. McKrae, a stromal disease causing a neurovirulent type of HSV-1 was used as the ocular challenge virus. KOS, which when given peripherally does not kill mice and does not produce stromal disease, was used as a live virus vaccine. Culture medium, supplements, and FCS were purchased from Invitrogen (Carlsbad, CA). 

Mice and Rabbits

Inbred BALB/c and C57BL/6 (female, 6 weeks old) mice were obtained from The Jackson Laboratory (Bar Harbor, ME). H-2 class I-negative HLA-A*0201 transgenic mice (referred to as HLA-A*0201) were developed by Francois Lemonier at Pasteur Institute ²⁸ and bred at the University of California Irvine (UCI) . Female New Zealand White rabbits were 8- to 10-weeks old at the time of infection. All animals were maintained in standard germ-free housing conditions at Cedars-Sinai Medical Center (CSMC) and UCI vivariums, with the approval of the Institutional Animal Care and Use Committees. Animals were managed in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

Peptide Synthesis

Based on our previous mapping study and prediction algorithms, we identified an octamer within the signal sequence of HSV-1 gK. ²³ Peptide 2, corresponding to gK amino acids 11-25 (STVVLITAYGLVLVW), the octamer or 8mer within this peptide (ITAYGLVL), and a scrambled peptide (S-8mer) containing the same amino acids as the 8mer (GILAYLVT) were synthesized by the Brain Research Center (University of British Columbia, Vancouver, BC Canada). Peptide 1, corresponding to gK amino acids 1-15 (MLAVRSLQHLSTVVL), and peptide 3, corresponding to gK amino acids 20-35 (LVLVWYTVFGASPLH), were synthesized by Mimotopes (San Diego, CA) and were used as the control in some experiments. Peptides corresponding to portions of VSV-G (YTDIEMNRLGK) and HSV-1 gD (KQPELAPEDPED) were also used as the control in some experiments and were purchased from Sigma-Aldrich (St. Louis, MO). The purity of the synthesized peptide was at least 95% except for peptides 1 and 3, which had a purity of >70%. All peptides and BSA were dissolved in DMSO at a concentration of 50 μg/μL and stored at −20°C. 

Peptides Administered as Eye Drops

Mouse and rabbit eyes received 100 μg of peptide or BSA as an eye drop in 2 μL of DMSO 1 hour before ocular infection. This dose and time were chosen, as they had produced the greatest effect on virus replication in a pilot experiment in mice. 

Infection of Mice and Rabbits

For ocular infections, mice, and rabbits were bilaterally infected in both eyes without scarification or anesthesia by placing as eye drops 2 × 10⁵ PFU of HSV-1 strain McKrae in 2 or 25 μL of tissue culture medium, respectively. The eyelid was closed and gently rubbed for 30 seconds. For determining LD₅₀, mice were similarly ocularly infected with 2 × 10⁵, 2 × 10⁴, 2 × 10³, or 2 × 10² PFU of HSV-1 strain McKrae. 

Eye Disease

The severity of HSV-1-induced CS in infected mice was assessed by a slit lamp biomicroscope examination performed by investigators blinded to the treatment regimen of the mice. CS was scored by using a standard 0 to 4 scale (0, no disease; 1, 25%; 2, 50%; 3, 75%; and 4, 100% staining or involvement). The eyes were examined for CS on day 28 PI. 

Replication and Clearance of HSV-1 from the Eye

Eyes were swabbed once daily on days 1 to 5 with a dacron swab (type 1; Spectrum Laboratories, Los Angeles, CA). The swab was transferred to a 12 × 75-mm culturing tube containing 1 mL of medium, frozen, and thawed, and the virus titers were determined by standard plaque assays on RS cells. 

Immunization

Mice were immunized three times IP at 2-week intervals with baculovirus-expressed gK, as we previously described, ¹¹ or once IP with 1 × 10⁶ PFU of avirulent live HSV-1 strain KOS. HLA-A*0201 mice were immunized three times at 10-day intervals with 1 × 10⁶ PFU of the avirulent HSV-1 strain KOS or mock immunized on the same schedules with medium alone. HLA-A*0201 immunized or mock immunized mice were euthanatized 21 days after the final injection and spleens were harvested. 

Lymphocyte Isolation

The mice were euthanatized and the spleens were individually harvested and single-cell suspensions prepared for CD8⁺ T-cell-based assays, as we previously described. ²⁹  

CD107 Cytotoxicity Assay

To detect gK- and gD-peptide–specific cytolytic CD8⁺ T-cells in freshly isolated spleen cells, we used the CD107a/b cytotoxicity assay, as previously described. ^{29 30} Briefly, on the day of the assay, spleen cells were incubated with 10 μg/mL of gK-8mer or gD (KQPELAPEDPED) peptide or with 1 μg/mL of anti-CD3 at 37°C for 5 to 6 hours in the presence of a protein transport inhibitor (GolgiStop; BD Biosciences, San Jose, CA) and 10 μL of CD107a-FITC and CD107b-FITC. At the end of the incubation period, the cells were harvested into separate tubes and washed with flow cytometry buffer (FACS; BD Biosciences) and then stained with PE-conjugated anti-mouse CD8 for 30 minutes. The cells were washed again and analyzed with the buffer. 

IFN-γ ELISA

T-cell responses were measured by IFN-γ production in peptide-stimulated spleen cells using in vitro and ex vivo ELISAs (ELISpot; BD-Pharmingen). (1) In vitro ELISA: 2 × 10⁶ cells were prestimulated with 5 PFU/cell of heat-inactivated HSV-1 (strain McKrae) for 6 days. After stimulation, activated cells were harvested and washed, and IFN-γ production was measured by the ELISA, as described for the ex vivo assay. (2) Ex vivo ELISA: To minimize artifacts that might arise from in vitro restimulation of CD8⁺ T cells, we performed an ex vivo experiment using freshly isolated, spleen-derived T-cells directly without prior in vitro stimulation. Briefly, 3 × 10⁵ cells were stimulated ex vivo with 10 μg/mL of individual gK 8mer or gD peptide for 24 hours in IFN-γ ELISA plates coated with anti-mouse IFN-γ capture Ab in a humidified incubator at 37°C with 5% CO₂. The spot-forming cells were developed as described by the manufacturer (IFN- γ-ELISpot KIT; BD-Pharmingen) and counted by stereoscopic microscope. 

RNA Extraction, cDNA Synthesis, and RT-PCR

Tissue processing of corneas and TG, total RNA extraction, and RNA yield were performed as we have described elsewhere. ^{26 31} Briefly, 1000 ng of total RNA was reverse-transcribed with random hexamer primers and murine leukemia virus (MuLV) reverse transcriptase (High Capacity cDNA Reverse Transcription Kit; Applied Biosystems [ABI], Foster City, CA), in accordance with the manufacturer’s recommendations. 

The differences in expression levels of gK, gB, LAT (stable 2kb LAT intron), CD4, CD8, IFN-γ, granzyme A (GzmA), and granzyme B (GzmB) were evaluated using custom-made gene-expression primers (TaqMan; ABI), described later. Relative copy numbers for the gK, gB, and LAT transcripts were calculated by using standard curves generated from the plasmids pGEM-gK1040, pAc-gB1, and pGEM5317, respectively. In all experiments, GAPDH was used for normalization of transcripts. 

Expression levels of the various transcripts were evaluated with commercial gene-expression assays (TaqMan; ABI) with optimized primer and probe concentrations. Primer probe sets consisted of two unlabeled PCR primers and the FAM dye–labeled probe (TaqMan MGB; ABI) formulated into a single mixture. In addition, all cellular amplicons included an intron–exon junction to eliminate signal from genomic DNA contamination. The assays for cellular transcripts were (1) CD4 (assay IMm00442754_m1), amplicon length, 72 bp; (2) CD8 (α chain; assay Mm01182108_m1), amplicon length, 67 bp; (3) GAPDH (assay Mm999999.15_G1), amplicon length, 107 bp; (4) GzmA (assay Mm00439190_m1), amplicon length, 77 bp; (5) GzmB (assay Mm00442834_m1) amplicon length, 95 bp; and (6) IFN-γ (assay Mm00801778_m1; all from ABI) amplicon length, 101 bp. 

Expression levels of HSV-1 gK, gB, and LAT were similarly evaluated by using custom-made gene-expression assays (TaqMan; ABI). The gB primers and probe were forward primer, 5′-AACGCGACGCACATCAAG-3′; reverse primer, 5′-CTGGTACGCGATCAGAAAGC-3′; and probe, 5′-FAM-CAGCCGCAGTACTACC-3′. The gK primers and probe were forward primer, 5′-GGCCACCTACCTCTTGAACTAC-3′; reverse primer, 5′-CAGGCGGGTAATTTTCGTGTAG-3′; and probe, 5′-FAM-CAGGCCGCATCGTATC-3′. The LAT primers and probe were forward primer, 5′-GGGTGGGCTCGTGTTACAG-3′; reverse primer, 5′-GGACGGGTAAGTAACAGAAGTCTCTA-3′; and probe, 5′-FAM-ACACCAGCCCGTTCTT-3′. The amplicon lengths for gB, gK, and LAT were 72, 82, and 81 bp, respectively. 

Quantitative real-time PCR was performed with a sequence-detection system (Prism 7900HT; ABI) in 384-well plates, as we described previously. ^{26 31} Real-time PCR was performed in triplicate for each tissue sample. The threshold cycle (C_t) values, which represents the PCR cycle at which there is a noticeable increase in the reporter fluorescence above baseline, were determined with sequence-detection system software (SDS 2.2 Software; ABI). In each experiment, an estimated relative copy number of each target gene was calculated with standard curves generated from plasmids containing the gene of interest: pGem-gK, ¹¹ pAc-gB1, ³² and pGem-5317. Briefly, each plasmid DNA template was serially diluted 10-fold, such that 5 μL contained from 10³ to 10¹¹ copies of the desired gene, then subjected to PCR (TaqMan; BD Biosciences) with the same set of primers as the test samples. By comparing the normalized threshold cycle of each sample with the threshold cycle of the standards, we determined the copy number for each reaction. 

Statistical Analysis

Protective parameters were analyzed by Student’s t-test and the Fisher exact test (Instat; GraphPad, San Diego, CA). Results were considered to be statistically significant at P < 0.05. LD₅₀ was determined by a second software program (StatPlus; AnalystSoft, Vancouver, BC, Canada). 

Results

Increased Virus Replication in Tears of gK Peptide–Treated Mice and Rabbits

Peptide 2 (STVVLITAYGLVLVW), an immunodominant CD4⁺ and CD8⁺ T-cell-stimulatory region of gK in vitro and in vivo, ²³ was given to BALB/c mice (100 μg/eye as an eye drop), 1 hour before ocular infection with HSV-1 strain McKrae, as described in Materials and Methods. Control mice received gK peptide 1 or peptide 3, neither of which induces T-cell activity in vitro or in vivo ²³ ; or an HSV-1 gD peptide, or a VSV-G peptide, as described in Materials and Methods. Peptide 2, but not peptides 1 or 3, significantly increased virus replication in eyes of infected mice on day 1 PI (Fig. 1A ; P < 0.001, Student’s t-test compared with the various control peptides). However, by day 2 PI, no significant differences were detected between peptide 2 and the control (not shown). 

Using prediction algorithms and overlapping peptides 1 and 3, we identified an 8mer (ITAYGLVL) ²³ within peptide 2 that is identical in HSV-1 (I₁₆-L₂₃) and HSV-2 (I₁₇-L₂₃). ^{9 24 25} To determine whether the increased virus replication induced by peptide 2 was associated with this 8mer, an experiment identical with that just described was performed with 100 μg of the 8mer instead of peptide 2. BSA, which in pilot experiments had no effect on HSV-1 replication and resulted in virus titers similar to that of the various peptide controls and the scrambled 8mer (S-8mer) described in Materials and Methods, were used as the control. Similar to peptide 2, on day 1 PI, BALB/c mice that received the 8mer as an eye drop had significantly more virus in their eyes than did the controls (Fig. 1B) . To determine whether the increased virus replication in 8mer-treated eyes was mouse strain specific, the experiment was repeated in C57BL/6 mice. Again, compared with control mice, the 8mer-treated C57BL/6 mice had significantly more virus in their eyes on day 1 PI (Fig. 1B) . Thus, the effect of the 8mer did not appear to be mouse strain specific. In other experiments mice were given the same dose of the 8mer (100 μg/eye) four times: 1 hour before infection and also at 24, 48, and 72 hours PI. No significant differences were observed in mice that received four doses versus one dose of the 8mer (data not shown). 

We previously showed that immunization of mice with full-length baculovirus expressed HSV-1 gK increases HSV-1 replication in the eye and HSV-1-induced eye disease. ¹¹ In addition, people with HSK tend to have elevated antibody titers to gK. ³³ To determine whether the 8mer treatment would effect virus replication in gK vaccinated mice, we immunized BALB/c mice with baculovirus-expressed gK (three times at 2-week intervals) or with the live avirulent HSV-1 strain KOS (once). Three weeks after the final immunization, the mice were treated with the 8mer as described earlier and infected. gK- and KOS-immunized mice that received the 8mer had significantly more virus in their eyes on day 1 PI than the corresponding control mice (Fig. 1C) . Note that the scales for the gK- and KOS-immunized plots are different. Consistent with gK vaccination causing increased HSV-1-induced corneal disease and KOS vaccination providing protection, the gK-vaccinated mice treated with the 8mer had approximately eight times more virus in their eyes than did the KOS-vaccinated mice treated with the 8mer. Thus, the 8mer treatment appeared to increase virus replication on day 1 PI, even in immunized mice. 

To determine whether the 8mer treatment would also increase virus replication in rabbit eyes, 15 rabbits were treated with 100 μg/eye of the 8mer, with the same method as was used for the mice, and infected 1 hour later with 2 × 10⁵ PFU/eye of HSV-1 McKrae. Control rabbits received BSA before ocular infection. Infectious virus was titered in eye swabs collected daily. On days 1 to 4 PI, the 8mer treatment resulted in significantly more virus being detected in the rabbits’ eyes than in the control eyes (Fig. 1D) . Thus, the 8mer treatment also increased virus titers in rabbit eyes, but in rabbits, the effect of the 8mer treatment appeared to last for 4 days rather than just 1 day. 

Virulence of 8mer in Mice

To determine whether the 8mer treatment altered viral virulence, groups of 10 to 100 BALB/c mice were treated with the 8mer as described earlier and then infected ocularly in both eyes with 10-fold serial dilutions of McKrae virus ranging from 2 × 10² to 2 × 10⁵ PFU/mouse eye (Table 1) . The 50% lethal dose (LD₅₀), calculated as described in Materials and Methods, was 22,125 PFU/eye in S-8mer- and BSA-treated control mice and only 5448 PFU/eye for the 8mer-treated mice. Thus, the 8mer treatment appeared to increase viral virulence by approximately fourfold as judged by LD₅₀. Because C57BL/6 mice are naturally more resistant to ocular HSV-1 infection (Table 1) , it was not possible to calculate the LD₅₀ in this mouse strain. However, 13% (2/15) of C57BL/6 mice that were treated with the 8mer before infection died. This is the first time in our experience that any C57BL/6 mice died after ocular HSV-1 infection, suggesting that the gK 8mer treatment increased HSV-1-related death in C57BL/6 mice as well as BALB/c mice. 

Effect of gK 8mer on CS in Infected Mice

To determine the effect of the gK 8mer on eye disease, on day 30 PI, we determined the severity of CS in C57BL/6 mice that survived infection with 2 × 10⁵ PFU/eye of HSV-1 McKrae from the experiment shown in Table 1 . Because very few BALB/c mice survived infection with 2 × 10⁵ PFU/eye of HSV-1 McKrae, the severity of CS was determined in surviving BALB/c mice pooled from various experiments performed as part of these studies (Table 1) . A total of 9 of 100 gK 8mer–treated BALB/c mice survived ocular infection with 2 × 10⁵ PFU/eye McKrae, whereas 12 of 55 control-treated BALB/c mice survived. The 8mer-treated BALB/c and C57BL/6 mice both had significantly higher CS scores compared to the corresponding control-treated mice (Fig. 2) . The increased CS was more apparent in C57BL/6 mice because they usually develop very little HSV-1-induced eye disease. 

Effect of 8mer on Expression of Viral and Cellular Transcripts in Corneas of Ocularly Infected Mice

BALB/c mice were treated with the 8mer or the BSA control, infected as just described, and euthanatized on the days indicated. Their corneas were harvested and the total RNA isolated as described in Materials and Methods. RT-PCR was performed to quantitate mRNA levels of gB, gK, CD8, IFN-γ, CD4, GzmA, and GzmB. GAPDH mRNA was used as an internal control. The amount of gB mRNA (Fig. 3A)and gK mRNA (Fig. 3B)appeared higher in the 8mer-treated eyes on day 1 PI but not on days 2, 5, or 30 PI. This result was consistent with the higher virus replication recorded on day 1 PI, but not at later times. 

On days 1 and 2 PI, but not day 5 PI, the 8mer-treated corneas had elevated levels of CD8 mRNA (Fig. 3C) . There also appeared to be elevated IFN-γ mRNA in these corneas on day 1 PI (Fig. 3D) . Both GzmA (Fig. 3F)and GzmB (Fig. 3G)mRNA levels appeared higher in corneas from the 8mer-treated mice on all days examined. In contrast, there did not appear to be any significant differences in the level of CD4 mRNA in corneas from the 8mer versus BSA-treated mice (Fig. 3E) . Thus, the 8mer treatment directly or indirectly increased expression of some transcripts in the cornea after ocular HSV-1 infection. The specific cell types responsible remain to be determined. 

Effect of 8mer Treatment on gB, gK, and LAT Transcripts in TGs of Latently Infected Mice and Rabbits

We previously detected gB and gK transcripts in TGs of latently infected mice after ocular infection with an HSV-1 mutant that over expresses gK. ²⁶ Thus, it was of interest to determine the effect of the 8mer treatment on gB, gK, and LAT expression during latency. BALB/c mice, C57BL/6 mice, and rabbits were treated with the 8mer or BSA control 1 hour before infection with 2 × 10⁵ PFU/eye of HSV-1 McKrae straom. TG from latently infected animals were isolated and RT-PCR was performed on total RNA to determine the relative levels of gK, gB, and the 2-kb LAT intron (LAT-2Kb). 

In the BALB/c mice, as expected, since LAT is the only HSV-1 gene reproducibly shown to be abundantly expressed during neuronal latency, 2kb LAT RNA was detected in both the 8mer- and BSA-treated BALB/c mice (Fig. 4A ; BALB/c mice), whereas gB and gK mRNAs were not detected (Fig. 4A ; BALB/c mice). Although more 2Kb LAT RNA may have been detected in the 8mer-treated mice, the difference was not significant. Results with C57BL/6 mice were similar (Fig. 4A ; C57BL/6 mice), except that lower levels of LAT were detected (the x-axis scales are different). Although the reason for this is not known, it may involve the fact that C57BL/6 mice are more resistant to HSV-1 infection than are BALB/c mice. 

As expected, the stable 2-kb LAT intron but not gB mRNA was detected in all TG from latently infected rabbits, regardless of whether the rabbits received the 8mer or the BSA control (Fig. 4B) . Surprisingly, gK mRNA was detected in TG from five of the seven rabbits treated with the 8mer and in one of the six TG from the BSA control rabbits (Fig. 4B) . Since spontaneous reactivation occurs in rabbits at a rate of approximately 10%, detection of gK mRNA may be due to viral reactivation from latency. If so, it suggests that the gK 8mer treatment results in increased HSV-1 reactivation. However, such reactivation would not explain why gB mRNA was not detected, and the reason for the presence of gK, but not gB, mRNA remains to be determined. 

Lytic Activity of gK 8mer-Specific CD8⁺ T Cells from Spleens of Mice Immunized with HSV-1 Strain KOS

To determine whether the gK 8mer could also affect human T cells, we used humanized HLA-A*0201 transgenic mice in this study. The transgenic mice were immunized IP with 1 × 10⁶ PFU/mouse of live avirulent KOS strain HSV-1 and euthanatized 21 days after the third immunization. Spleen cells were prepared as described in Materials and Methods. Live spleen cells were restimulated in vitro with the gK 8mer or gD peptide, and CD107a/b expression was examined by flow cytometry on gated CD8⁺ T-cells (Fig. 5A) . CD107a and CD107b are lysosomal-associated membrane glycoproteins that surround the core of the lytic granules in cytotoxic T cells. ^{34 35} On TCR engagement and stimulation by antigens in association with MHC molecules, CD107a/b is exposed on the cell membrane of cytotoxic T cells. Thus, surface expression of CD107a/b on cytotoxic T lymphocytes (CTLs) can be used as a direct assay for epitope-specific CTL response. ³⁰ The gK 8mer induced significant CD107a/b expression on CD8⁺ T-cells from KOS but not mock immunized HLA-A*0201 transgenic mice, indicating functional cytolytic activity in the spleen directed against the gK 8mer but not the gD peptide (Fig. 5A) . As expected, the gD peptide was negative because the gD peptide is not thought to induce CTL activity. ³⁶ As a positive control, spleen T cells from both KOS and mock immunized HLA-A*0201 transgenic mice had similar CD107a/b expression after stimulation with anti-mouse CD3 (Fig. 5A , right panel). 

gK Peptide-Specific IFN-γ-Producing CD8⁺ T-Cells Detected in Immunized Mice

To minimize artifacts that might arise from in vitro restimulation of CD8⁺ T cells, we performed an ex vivo experiment using freshly isolated spleen-derived T-cells directly. In the ex vivo experiment, both the gK 8mer and the gD peptide elicited strong IFN-γ production in cells from KOS-immunized mice (Fig. 5B) . As expected, when the cells were prestimulated for 6 days in vitro, the IFN-γ response was further increased (Fig. 5C) . Also as expected, spleen cells from mock immunized mice produced significantly less IFN-γ (Figs. 5B 5C ; P < 0.0001). Spleen T cells from KOS and mock-immunized mice had similar levels of IFN-γ expression after stimulation with PHA (Figs. 5B 5C ; right panels). The results shown in Figure 5A 5B 5Cindicate that immunization of humanized HLA-A*0201 transgenic mice with whole HSV-1 induced a CD8 T-cell response directed against the gK 8mer. This 8mer therefore appears to represent a gK epitope that can be recognized by humanized HLA-A*0201 T-cells and, by extension, human CD8 T-cells. 

Discussion

We previously showed that immunization with expressed gK significantly exacerbates CS after ocular HSV-1 infection. ^{19 20} This exacerbation appears to involve anti-gK IgG, because purified IgG from gK-immunized mice injected IP into naïve mice produced the same exacerbated CS after ocular HSV-1 infection. One or more CD8⁺ cell types was also involved because depletion of CD8⁺ cells with anti-CD8 antibody prevented the exacerbation. Consistent with the above, we also previously found that ocular infection of naive mice with an HSV-1 mutant expressing three copies of gK (HSV-gK ³ ) rather than one copy resulted in exacerbated CS. ²⁶  

One of our goals has been to map the region of gK that is involved in the above phenomenon. We speculated that giving a large amount of a gK peptide as an eye drop shortly before ocular HSV-1 infection might simulate ocular infection with the HSV-gK ³ mutant virus that overexpresses gK and allow us to map the region of gK involved in exacerbation of CS. We initially used a 15-amino-acid peptide, STVVLITAYGLVLVW, located within the gK signal sequence, ³⁷ that we previously showed is an immunodominant CD4⁺ and CD8⁺ T-cell stimulatory region. ²³ We then investigated an eight-amino-acid sequence (gK 8mer) within the gK 15mer that is well conserved within and among HSV-1 and -2 strains. ^{9 24 25} As reported herein, the 8mer exacerbated CS in mice, strongly suggesting that this region of gK is involved in gK exacerbation of CS. Furthermore, the increased levels of CD8, IFN-γ, GzmA, and GzmB mRNAs in 8mer-treated BALB/c and C57BL/6 mice, suggests that the gK 8mer may be a CD8⁺ T-cell activator. 

We have shown that vaccination of mice with gK provides no protection against ocular challenge and in fact results in higher virus titers in the eye and exacerbation of eye disease after ocular challenge with HSV-1 strain McKrae. ^{19 20} In contrast, vaccination with the avirulent KOS strain of HSV-1 provides significant protection against virus replication and eye disease. Thus, as seen in Figure 1C , it was expected that gK 8mer treatment would have less effect in KOS-vaccinated mice than in gK-vaccinated mice. This result is presumably due to the partial protection afforded by KOS vaccination. In the course of these studies we found that immunization of humanized HLA-A*0201 transgenic mice with the avirulent HSV-1 strain KOS induced gK 8mer–specific IFN-γ-producing cytotoxic CD8⁺ T-cells. This finding suggests that in addition to increasing virus replication in the eye on day 1 PI and exacerbating CS, the gK 8mer is a promiscuous immunodominant epitope that would be recognized and responded to by humans with the HLA-A*0201 haplotype (i.e., >50% of people regardless of race or ethnicity). One immediate practical lesson from this result is that some herpes virus epitopes may lead to pathogenic rather than protective responses. It has been shown that pathogenic peptides are involved in autoimmunity ranging from uveitis to Alzheimer disease. ^{38 39 40 41 42 43 44 45} Logically, it would be best to exclude HSV-1 pathogenic epitopes from any candidate herpes vaccine. Thus, for any epitope based vaccine it is important to screen all the epitopes so that only protective epitopes and no potentially pathogenic epitopes are included. 

CD8⁺ T cells mediate transient herpes stromal keratitis in CD4-deficient mice. ⁴⁶ CD8⁺ T cells are also actively involved in autoimmune diseases. ^{47 48 49 50} The higher CD8 transcript in the cornea of mice that received the gK 8mer correlated with higher GzmA and B but not CD4 transcripts. Since cytotoxic lymphocytes kill target cells by receptor-triggered exocytosis of preformed secretory granules, ^{51 52} higher GzmA and B in corneas of gK 8mer-treated mice may be involved with more severe CS. Excessive accumulation of activated CD8⁺ T cells also increases mortality in perforin-deficient mice. ^{53 54} Furthermore, corneal buttons from patients with HSK have more HSV-reactive CD8⁺ T cells than HSV-reactive CD4⁺ T cells. ⁵⁵ In addition, we showed that in patients with HSK, increased eye disease correlates with increased immune response to gK. ³³ These findings, especially when combined with our finding that the gK 8mer is a CD8⁺ T-cell epitope, are consistent with the gK 8mer causing increased CS after ocular HSV-1 infection in mice via a CD8⁺ T-cell–mediated pathway. 

However, the gK peptide may have affected transient enhancement of viral replication by unknown mechanisms, which are most likely unrelated to the CD8⁺ T-cell responses. This initial enhancement of viral replication may have partially contributed to exacerbation of neurovirulence, since increased viral load was probably transmitted to mice ganglia. However, CS is an immunopathogenic phenomenon occurring at a late time after infection, and thus it is most likely due to the observed enhancement of anti-gK CD8⁺ T-cell immune responses. 

An alternative explanation is that elevation of the various CD8⁺ T-cell–related transcripts reported in this study infiltration of the cornea by other cells. Elevated mRNAs for GzmA and -B and IFN-γ could result from infiltration of NK cells, whereas elevated CD8 and IFN-γ transcripts could be due to CD8α⁺ DCs. In fact, as mentioned earlier, it is unlikely that infiltration of the cornea by CD8⁺ T-cells could occur fast enough to account for the rapid elevation of CD8 and IFN-γ transcripts seen within 24 hours of gK 8mer treatment and infection or the increased virus titers seen 24 hours PI. Another possibility is that the early effects of the gK 8mer seen at 24 hours PI (increased virus titers, elevated CD8 and IFN-γ transcripts) are distinct from those involved in the exacerbated CS which occurs later (after day 12 PI) when sufficient time has elapsed for induction of both gK 8mer antibody and gK 8mer CD8⁺ T-cell responses. Finally, another explanation is that both the innate and induced arms of the immune response are concurrently stimulated by the gK 8mer, with the innate responses producing the rapid effects and the induced arm (specifically CD8⁺ T cells) producing the later effects (CS). 

In summary, we have shown that the gK 8mer given as an eye drop 1 hour before ocular HSV-1 infection exacerbates CS in mice and that this gK 8mer is a promiscuous CD8⁺ T-cell epitope that would be recognized by more than 50% of humans with the HLA-A*0201 haplotype. Whether the gK 8mer exacerbates CS in our mouse model by functioning as a “pathogenic” CD8 epitope remains to be determined. 

 Supported by NIH Grant EY13615 (HG).

 Submitted for publication October 1, 2008; revised December 3, 2008, and January 5, 2009; accepted March 30, 2009.

 Disclosure: K.R. Mott, None; A.A. Chentoufi, None; D. Carpenter, None; L. BenMohamed, None; S.L. Wechsler, None; H. Ghiasi, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Effect of gK peptides on virus replication in eyes of mice and rabbits. (A) Virus titers in tears of BALB/c mouse treated with gK peptide 2. One hour before ocular infection with 2 × 10⁵ PFU/eye of HSV-1 strain McKrae, BALB/c mice received 100 μg/eye of gK peptide 2 in eye drops. Control mice received 100 μg/eye of gK peptides 1, gK peptide 3, gD peptide, or VSV-G peptide. Tear swabs were collected from infected eyes on day 1 after ocular infection and virus titers were determined by standard plaque assays. Each point represents the mean titer from 30 eyes from two separate experiments. (B) Virus titers in tears of BALB/c and C57BL/6 mice treated with the gK 8mer. All procedures were as in (A) except that the experimental group was gK 8mer and controls were BSA and the S-8mer. For the BALB/c mice, each point represents the mean of 110 eyes from seven separate experiments. For the C57BL/6 mice, each point represents the mean of 70 eyes from three separate experiments. (C) Virus titers in tears of immunized mice treated with gK 8mer. BALB/c mice were immunized three times with baculovirus expressed gK or once with live avirulent HSV-1 strain KOS. Three weeks after the last immunization, the mice were treated with gK 8mer, BSA, or S-8mer and ocularly infected. Each point represents the mean titer of 30 eyes. (D) Virus titers in tears of rabbits treated with gK 8mer. As in (B) and (C), New Zealand White rabbits were treated with the 8mer (or BSA control) and ocularly infected 1 hour later. Each point represents the mean titer from 30 eyes.

Table 1.

 

Survival of BALB/c and C57BL/6 Mice Following Ocular Application of the 8mer^*

Virus	BALB/c, †									C57BL/6
		2 × 105	2 × 104	2 × 103	2 × 102	LD50	2 × 105
8mer	9/100 (9%)	0/10 (0%)	4/10 (40%)	8/10 (80%)	5,448	13/15 (87%)
BSA	12/55 (22%)	2/10 (20%)	7/10 (70%)	10/10 (100%)	22,125	15/15 (100%)
S-8mer	0/10 (0%)	2/10 (20%)	7/10 (70%)	10/10 (100%)	22,125	10/10 (100%)

*  Mice were treated with 100 μg of 8mer, BSA, or S-8mer per eye and 1 hour later were ocularly infected with different PFU/eye of HSV-1 strain McKrae. Survival was determined 30 days after infection.

†  Survival for virus dose of 2 × 10², 2 × 10³, and 2 × 10⁴ were repeated twice, while the survival results for 2 × 10⁵ PFU/eye is based on seven seperate experiments.

Figure 2.

 

Effect of 8mer on eye disease in infected mice. CS in surviving BALB/c or C57BL/6 mice that received 8mer or BSA was determined 30 days after ocular infection. For BALB/c mice the CS scores represents the average ± SEM of 18 eyes in the 8mer group and 24 eyes in the BSA group. For C57BL/6 mice, the CS scores represents the average ± SEM of 26 eyes for the 8mer and 30 eyes for BSA.

Figure 3.

 

Estimated transcript levels in mouse corneas. Individual corneas from ocularly infected 8mer or BSA-treated BALB/c mice were isolated at various times PI, RNA prepared, and RT-PCR performed on total RNA. GAPDH was used as an endogenous control to normalize relative RNA levels. The gB and gK RNA levels are shown as copy numbers. The RNA levels for CD8, IFN-γ, CD4, and GzmA and -B are shown as normalized expression compared with uninfected control mice (x-fold change). Each bar represents the mean ± SEM of results in five mice. (A) gB mRNA; (B) gK mRNA; (C) CD8 mRNA; (D) IFN-γ mRNA; (E) CD4 mRNA; (F) GzmA mRNA; and (G) GzmB mRNA.

Figure 4.

 

Detection of HSV-1 transcripts in trigeminal ganglia (TG) during latency. RT-PCR was performed on RNA from TG isolated 30 days PI, using GAPDH as an endogenous control as above. (A) Mice. TG from individual mice were combined. Each bar represents the mean ± SEM of five mice. (B) Rabbits. RT-PCR was performed on one half of each TG. Each bar represents the mean ± SEM. A rabbit was considered positive for LAT or gK if one of the TGs was positive (numbers above each bar indicate the number of positive rabbits/total rabbits).

Figure 5.

 

Immunogenicity of gK 8mer in HLA-A*0201 Tg mice after immunization with HSV-1 strain KOS. (A) HLA-A*0201 Tg mice were immunized three times with KOS. Spleen CD8⁺ T cells were isolated 21 days after the third immunization and exposed to 10 μg/mL of gK-8mer, gD peptide or 1μg/mL of anti-CD3 at 37°C for 5 to 6 hours in the presence of protein transport inhibitor and 10 μL of CD107a-FITC and CD107b-FITC. The cells were harvested into separate tubes, washed, stained with PE-conjugated anti-mouse CD8 for 30 minutes, washed, and analyzed by flow cytometry. Left: CD107 expression in spleen cells from HLA-A*0201 Tg mice stimulated ex vivo with the gK 8mer or gD peptide. Right: CD107 expression in spleen cells stimulated ex vivo with anti-CD3. Error bars indicate SEM. (B) Spleen cells isolated directly ex vivo were stimulated as indicated and gK-8mer and gD-peptide-specific IFN-γ-producing T-cells were determined by ELISA. (C) IFN-γ produced by spleen T cells was assessed after 6 days of in vitro stimulation with heat inactivated HSV-1 followed by a 24 hour restimulation with gK-8mer or gD peptide. ELISAs were performed in duplicate for each experiment. Each bar represents the mean results from three mice. The spots were developed and calculated as spot-forming cells = [(mean number of spots in the presence of antigen) − (mean number of spots in the absence of stimulation)]. Error bars, SEM.