Asymptomatic HLA-A*02:01–Restricted Epitopes from Herpes Simplex Virus Glycoprotein B Preferentially Recall Polyfunctional CD8⁺ T Cells from Seropositive Asymptomatic Individuals and Protect HLA Transgenic Mice against Ocular Herpes

Human study population

During the last decade (i.e., January 2003–January 2013) we have screened a total of 525 individuals for HSV-1 and HSV-2 seropositivity. Among these, a cohort of 207 immunocompetent individuals, with an age range of 18–65 y (median, 32 y), who were seropositive for HSV-1 were enrolled in the current study. Three hundred eighty-five individuals were white, 140 were non-white (African, Asian, Hispanic, and others), 274 were females, and 251 were males. All patients were negative for HIV, hepatitis B virus, and had no history of immunodeficiency. Two hundred eighteen patients were HSV-1– or HSV-1/HSV-2–seropositive, among whom 208 patients were healthy and ASYMP (individuals who have never had any recurrent herpes disease, ocular, genital, or elsewhere, based on their self-report and physician examination; even a single episode of any herpetic disease in their life will exclude the individual from this group). The other 10 remaining patients were defined as HSV-1–seropositive SYMP who suffered frequent and severe recurrent oral and/or orofacial lesions, with two patients having had clinically well-documented HSK. Signs of recurrent disease in SYMP patients were defined as herpetic lid lesions, herpetic conjunctivitis, dendritic or geographic keratitis, stromal keratitis, and iritis consistent with HSK, with one or more episodes per year for the past 2 y. However, at the time of blood collection, SYMP patients have no recurrent disease (other than corneal scarring) and have had no recurrences during the past 30 d. They have no ocular disease other than HSK, have no history of recurrent genital herpes, are HSV-1–seropositive and HSV-2–seronegative. Because the spectrum of recurrent ocular herpes disease is wide, our emphasis is mainly on the number of recurrent episodes, not on the severity of the recurrent disease. For simplicity, just the extreme cases of SYMP and ASYMP patients are used. Thus, no attempt was made to assign specific T cell epitopes to specific severity of recurrent lesions. Patients are also excluded when they 1) have an active ocular (or elsewhere) herpetic lesion, or had one in the past 30 d; 2) are seropositive for HSV-2; 3) are pregnant or breastfeeding; or 4) ever took acyclovir or related antiviral drugs or any immunosuppressive drugs. SYMP and ASYMP groups were matched for age, gender, serological status, and race. Sixty-nine healthy control individuals were seronegative for both HSV-1 and HSV-2 and had no history of ocular HSK, genital lesions, or orofacial herpes disease. All subjects were enrolled at the University of California Irvine under approved Institutional Review Board–approved protocols (nos. 2003–3111 and 2009–6963). Written informed consent was received from all participants prior to inclusion in the study.

Bioinformatics analyses

HSV-1 gB open reading frames used in this study were from strain 17 (National Center for Biotechnology Information, accession no. NC-001806). Candidate HLA-A*02:01–restricted epitopes were identified using previously described software from the National Institutes of Health Bioinformatics and Molecular Analysis Section (Washington, DC; http://bimas.dcrt.nih.gov/molbio/hla_bind/) and the SYFPEITHI algorithm (http://www.syfpeithi.de/) (3). Potential cleavage sites for human proteasome were identified using NetChop 3.0 (http://www.cbs.dtu.dk/services/NetChop/) (3). MHC Pathway (http://www.mhc-pathway.net) was also employed in this screening (3).

Peptide synthesis

Based on the bioinformatics analysis, 10 putative HLA-A*02:01–binding peptides from gB with high estimated t_1/2 of dissociation were synthesized by Magenex (San Diego, CA) on a 9050 Pep Synthesizer using solid-phase peptide synthesis and standard 9-fluorenylmethoxycarbonyl technology (PE Applied Biosystems, Foster City, CA). The purity of peptides was between 75 and 96%, as determined by reversed-phase HPLC (Vydac C18) and mass spectroscopy (Voyager MALDI-TOF system). Stock solutions were made at 1 mg/ml in 10% DMSO in PBS. All peptides were aliquoted and stored at −20°C until assayed.

Binding with soluble HLA-A*02:01 molecules

Quantitative assays to measure binding of peptides to soluble HLA-A*02:01 molecules are based on inhibition of binding of a radiolabeled standard peptide, as recently described (28). Briefly, 1–10 nM radiolabeled peptide was coincubated with 1 M–1 nM purified MHC and 1–3 µM human β₂-microglobulin. After 2 d, binding of radiolabeled peptide to MHC class I molecules was determined by capturing MHC/peptide complexes on Greiner Lumitrac 600 microplates coated with W6/32 Ab and measuring bound counts per minutes using a TopCount microscintillation counter. Concentration of peptide yielding 50% inhibition of binding of radiolabeled probe peptide (IC₅₀) was then calculated.

Stabilization of HLA-A*02:01 on class I-HLA–transfected B × T hybrid cell lines (T₂ cell line)

To determine whether synthetic peptides could stabilize HLA-A*02:01 molecule expression on the T₂ cell surface, peptide-inducing HLA-A*02:01 upregulation on T₂ cells was examined according to a previously described protocol (3). T₂ cells (3 × 10⁵/well) were incubated with different concentrations of individual gB peptide (as indicated in Fig. 2B) in 48-well plates for 18 h at 26°C. Cells were then incubated at 37°C for 3 h in the presence of human β₂-microglobulin (1 µg/ml) and BD GolgiStop (5 µg/ml) to block cell surface expression of newly synthesized HLA-A*02:01 molecules. The cells were washed with FACS buffer (1% BSA and 0.1% sodium azide in PBS) and stained with anti-HLA-A2.1–specific mAb BB7.2 (BD Pharmingen, San Diego, CA) at 4°C for 30 min. After incubation, the cells were washed with FACS buffer, fixed with 1% paraformaldehyde in PBS, and analyzed by flow cytometry using a BD LSR II (Becton Dickinson, Mountain View, CA). The acquired data, including mean fluorescence intensity (MFI), were analyzed with a FlowJo software version 9.5.2 (Tree Star). Percentage MFI increase was calculated as [(MFI with the given peptide − MFI without peptide)/(MFI without peptide)] × 100. Each experiment was performed three times, and means ± SD were calculated.

HLA typing

HLA-A2 subtyping was performed using a commercial sequence-specific primer kit (SSPR1-A2; One Lambda, Canoga Park, CA) following the manufacturer’s instructions (29). Briefly, genomic DNA extracted from PBMCs of HSV-seropositive SYMP and ASYMP individuals was analyzed using a Tecan DNA workstation from a 96-well plate with 2 µl volume per well, as previously described (29). The yield and purity of each DNA sample were tested using a UV spectrophotometer. The integrity of DNA samples was ascertained by electrophoresis on agarose gel. Each DNA sample was then subjected to multiple small volume PCR reactions using primers specific to areas of the genome surrounding the single point mutations associated with each allele. Only primers that matched the specific sequence of a particular allele would amplify a product. The PCR products were subsequently electrophoresed on a 2.5% agarose gel with ethidium bromide, and the pattern of amplicon generation was analyzed using HLA Fusion software (One Lambda). Additionally, the HLA-A2 status was confirmed by staining PBMCs with 2 µl anti–HLA-A2 mAb BB7.2 (BD Pharmingen) at 4°C for 30 min. The cells were washed, acquired on a BD LSR II, and analyzed using FlowJo software version 9.5.2 (Tree Star).

PBMC isolation

Healthy individuals (negative for HIV, hepatitis B virus, and with or without any HSV infection history) were recruited at the University of California Irvine Institute for Clinical and Translational Science. Between 40 and 100 ml blood was drawn into yellow-top Vacutainer tubes (Becton Dickinson). The serum was isolated and stored at −80°C for detection of anti–HSV-1 and anti–HSV-2 Abs, as we previously described (9). PBMCs were isolated by gradient centrifugation using leukocyte separation medium (Cellgro, Manassas, VA). The cells were washed in PBS and resuspended in complete culture medium consisting of RPMI 1640 medium containing 10% FBS (Gemini Bio-Products, Woodland, CA) supplemented with 1× penicillin/l-glutamine/streptomycin, 1× sodium pyruvate, 1× nonessential amino acids, and 50 µM 2-ME (Life Technologies, Rockville, MD). Aliquots of freshly isolated PBMCs were also cryopreserved in 90% FCS and 10% DMSO in liquid nitrogen for future testing.

T cell proliferation assay

CD8⁺ T cell proliferation was measured using a CFSE assay as we recently described (2). Briefly, PBMCs were labeled with CFSE (2 µM) and incubated for 5 d with or without individual gB peptide (10 µg/ml). As a positive control, 2 µg/ml PHA was used to stimulate T cells for 3 d. The cells were then washed and stained with PE-conjugated mAbs specific to human CD8 molecules (clone HIT8A; BD Pharmingen). The numbers of dividing CD8⁺ T cells per 300,000 total cells were analyzed by FACS. Their absolute number was calculated using the following formula: number of events in CD8⁺/CFSE⁺ cells × number of events in gated lymphocytes/number of total events acquired.

Flow cytometry analysis

For each stimulation condition, at least 500,000 total events were acquired on a BD LSR II, and data analysis was performed using FlowJo version 9.5.2 (Tree Star). PBMCs were analyzed by flow cytometry after staining with fluorochrome-conjugated human specific mAbs. FITC-conjugated CD8 (clone HIT8A) and FITC-conjugated HLA-A2 (clone BB7.2) were purchased from BD Pharmingen. PE-conjugated gB peptide/tetramer complexes were gifted by the National Institutes of Health Tetramer Facility. A total of 10⁶ PBMCs were stained in PBS containing 1% BSA and 0.1% sodium azide (FACS buffer) for 45 min at 4°C followed by three washes in FACS buffer and fixed in 1% paraformaldehyde. The gating strategy was similar to that used in our previous work (3). Briefly, we gated on single cells, dump⁻ cells, viable cells (aqua blue⁻), lymphocytes, CD3⁺ cells, and CD8⁺ cells before finally gating on functional cells. Reported data have been corrected for background based on the negative (no peptide) control where appropriate, and only responses with a total frequency >0.10% of total CD8⁺ T cells (after background subtraction) were considered to be positive responses.

Tetramer/gB peptide complexes staining

Fresh PBMCs were analyzed for the frequency of CD8⁺ T cells recognizing the gB peptide/tetramer complexes, as we previously described (12, 30). The cells were incubated with gB peptide/tetramer complex for 30–45 min at 37°C. The cell preparations were then washed with FACS buffer and stained with FITC-conjugated anti-human CD8 mAb (BD Pharmingen). The cells were washed and fixed with 1% paraformaldehyde in PBS. The cells were then acquired on a BD LSR II and data were analyzed using FlowJo version 9.5.2 (Tree Star).

IFN-γ-ELISPOT assays

gB-specific IFN-γ–producing CD8⁺ T cells were characterized using ELISPOT. T cell stimulation was measured by IFN-γ production in peptide-stimulated PBMCs using a BD IFN-γ-ELISPOT kit (BD Pharmingen). Briefly, 5 × 10⁵ PBMCs were stimulated with 20 µM individual gB peptides for 5 d. Then, activated PBMCs were harvested, washed, and restimulated with the gB peptides for 24 h in IFN-γ-ELISPOT plates (Millipore) that had been previously coated with anti-human 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 (BD IFN-γ-ELISPOT kit; BD Pharmingen) and counted under stereoscopic microscope. Average spot counts for duplicate wells were calculated and background from wells with cells in medium only was subtracted.

Multiplex cytokine array

Fresh PBMCs (5 × 10⁵ cells) were stimulated in a 96-well round-bottom plate with or without gB peptides for 24, 48, or 96 h. Supernatants were collected after 24, 48, or 96 h of stimulation. Production of six different cytokines (IL-2, IL-6, IL-8, IL-17, IFN-γ, and TNF-α) was assayed using multiplex cytokine arrays (BioLegend) per the manufacturer’s protocols. Samples were acquired on a Labscan 100 analyzer (Luminex) using Bio-Plex manager 6.0 software (Bio-Rad). Background levels were determined from nonstimulated PBMCs.

CD107 cytotoxicity assay

To detect cytolytic CD8⁺ T cells recognizing gB peptides in freshly activated and in vitro–activated PBMCs, we performed CD107a/b cytotoxicity assay. Betts and colleagues (31, 32) have described the CD107 assay as an alternative cytotoxicity assay that is able to address some of the shortcomings of the ⁵¹Cr-release assay. The CD107 assay was performed as described (33, 34) with a few modifications. On the day of the assay, nonstimulated or in vitro gB peptide-stimulated PBMCs were incubated at 37°C for 6 h in a 96-well plate with BD GolgiStop (BD Biosciences), costimulatory anti-CD28 and anti-CD49d Abs (1 µg/ml), and 10 µl CD107a-FITC and CD107b-FITC. At the end of the incubation period the cells were harvested into separate tubes and washed twice with FACS buffer then stained with PE-conjugated anti-human CD8 for 30 min at 4°C. The cells were then washed again, fixed, and 500,000 total events were acquired on a BD LSR II, and data analysis was performed using FlowJo version 9.5.2 (Tree Star).

HLA-A*02:01 transgenic mice

HLA-A*02:01 transgenic mice provided by Dr. Lemonier (Pasteur Institute) were bred at the University of California Irvine. These mice represent the F₁ generation resulting from a cross between HLA-A*02:01/K^b transgenic mice (expressing a chimeric gene consisting of the 1 and 2 domains of HLA-A*02:01 and the 3 domain of H-2K^b) created on the BALB/c genetic background. Genotype of the HLA transgenic mice used in this study was confirmed as HLA-A*02:01, the most common A*02 subtype, supporting that the immunogenic SYMP versus ASYM peptide epitopes reported in this study are likely presented by the HLA-A*02:01 molecule. All animal studies were conducted in facilities approved by the Association for Assessment and Accreditation of Laboratory Animal Care and according to Institutional Animal Care and Use Committee–approved animal protocol (no. 202–2372). All studies have been approved by the University of California Irvine review Institutional Animal Care and Use Committee.

Virus production

HSV-1 (strain McKrae) was used in this study (3) and was grown and titrated on rabbit skin (RS) cells. UV-inactivated HSV-1 was generated as we previously described (1). HSV inactivation was confirmed by the inability to produce plaques when tested on RS cells.

Immunization of “humanized” HLA transgenic mice with SYMP and ASYMP peptide epitopes and ocular herpes challenge

Three groups of age-matched female HLA-A*02:01 transgenic mice (n = 10 each) were immunized s.c. with the ASYMP CD8⁺ T cell human epitopes (gB_342–350 and gB_561–569) delivered with the CD4⁺ T cell PADRE epitope emulsified in CpG₁₈₂₆ adjuvant (ASYMP), with the SYMP CD8⁺ T cell human epitopes (gB_183–191 and gB_441–449) delivered with the CD4⁺ T cell PADRE epitope emulsified in CpG₁₈₂₆ adjuvant (SYMP), or with the CpG₁₈₂₆ adjuvant alone (mock) on days 0 and day 21. All immunizations were carried out with 100 µM each peptide. A preliminary experiment was conducted to determine the LD₅₀ of strain McKrae in naive HLA-A*02:01 transgenic mice following ocular challenge. Two × LD₅₀ was then used in peptide- and mock-immunized mice to determine the protective efficacy of SYMP and ASYMP epitopes against lethal ocular herpes infection and disease. Two weeks after the final immunization, mice received an ocular HSV-1 challenge with 2 × 10⁵ PFU (strain McKrae). The corneas were inoculated, without scarification, with the virus in a 4 µl tissue culture medium placed gently and topically on the corneas of immunized and control mice, as previously described (3). Control mice were inoculated using mock samples of virus.

Monitoring of ocular herpes infection and disease

Animals were examined for signs of ocular disease by slit lamp. Clinical assessments were made immediately before inoculation and on days 1, 4, 7, 10, 14, and 21 thereafter. The examination was performed by investigators blinded to the treatment regimen of the mice and scored according to a standard 0–4 scale: 0, no disease; 1, 25%; 2, 50%; 3, 75%; and 4, 100% staining, as previously described (3). To quantify replication and clearance of HSV-1 from the eyes, mice were swabbed daily with moist, type 1 calcium alginate swabs. Swabs were placed in 1.0 ml titration media (Media 199, 2% penicillin/streptomycin, 2% newborn calf serum) and frozen at −80°C until titrated on RS cell monolayers as described previously (3). Mice were also examined for survival in a window of 30 d after challenge, as we previously described (3).

Statistical analyses

Data for each assay were compared by ANOVA and a Student t test using GraphPad Prism 5 software (GraphPad Software, San Diego, CA). Differences between the groups were identified by ANOVA and multiple comparison procedures, as we previously described (35). Data are expressed as the means ± SD. Results were considered to be statistically significant at p < 0.05.

In silico prediction of potential HLA-A*02:01–restricted T cell epitopes from HSV-1 gB Ag

We have searched the deduced HSV-1 (strain 17) gB amino acid sequence for potential HLA-A*02:01–binding regions using BIMAS, SYFPEITHI, and MAPPP predictive computational algorithms. Based on these analyses, we selected 10 potential peptide epitopes with high predicted affinity to the HLA-A*02:01 molecule, a haplotype that is represented in >50% of the world human population, irrespective of gender and ethnicity (2) (Table I). Nine of the 10 gB peptide epitopes shared the HLA-A*02:01–binding motifs: leucine, isoleucine, or methionine at the second position and a valine, leucine, isoleucine, or threonine at the ninth position. As confirmed by the MHCPred computational algorithm, all 10 gB peptides that bear putative antigenic and immunogenic HLA-A*02:01–binding CD8⁺ T cell epitopes are more readily accessible to proteolysis, an event that precedes T cell epitope presentation in association with HLA molecules (2, 3). The 10 potential epitopes are not confined to a particular region of gB (Fig. 1). One of the 10 predicted epitopes, gB_17–25, is located within the gB signal sequence. Seven belong to the external N-terminal ectodomain portion of gB (gB_161–169, gB_183–191, gB_286–294, gB_342–350, gB_343–351, gB_441–449, and gB_447–455). Two are adjacent to the hydrophobic membrane transmembrane anchor domain (gB_561–569 and gB_675–683). None of the predicted peptides is within the transmembrane or the C-terminal intracellular domain and none belongs to known glycosylated regions of gB protein.

Six of 10 potential gB epitope peptides bind with high affinity to HLA-A*02:01 and stabilize its expression on the surface of target cells

Peptides corresponding to the 10 highest probability gB epitopes, described above, were synthesized and the binding affinity of each peptide to soluble purified HLA-A*02:01 molecules was tested using a quantitative assay based on competitive inhibition of the binding of a radiolabeled reference peptide (2, 3). Six of 10 gB peptide epitopes bound with high affinity to the soluble HLA-A*02:01 molecules with a K_d of <100 nM (Fig. 2A). Specifically, gB_17–25, gB_161–169, gB_441–449, gB_447–455, and gB_561–569 peptides bound to HLA-A*02:01 molecules with a K_d of <25 nM. The gB_675–683 peptide bound to HLA-A*02:01 with a K_d of ~95 nM. In contrast, two peptides, gB_286–294 and gB_343–351, bound to HLA-A*02:01 with 100 nM > K_d < 500 nM (intermediate binding affinities). The remaining two peptides, gB_183–191 and gB_342–350, bound to HLAA*02:01 with a K_d of >500 nM (low binding affinities).

In a subsequent experiment, peptides corresponding to the 10 gB epitopes were used to perform a stabilization assay of HLA-A*02:01 molecules on the cell membrane of T₂ cells (a class I-HLA–transfected B × T hybrid cell line). This assay used a mAb specific to a folded structure of HLA-A*02:01 to estimate in a FACS assay the relative amount of HLA-A*02:01 molecules retained on the surface of T₂ cells following incubation with a potential peptide epitope (3). The amount of empty HLA-A*02:01 molecules retained on the surface of T₂ cells is normally at a low level. Each gB peptide was tested individually at four descending concentrations: 20, 10, 5, and 2.5 µM. As shown in Fig. 2B, 5 of 10 peptides, gB_161–169, gB_441–449, gB_447–455, gB_561–569, and gB_675–683, significantly increased the level of HLA-A*02:01 molecules detectable by FACS on the surface of T₂ cells in a dose-dependent manner, suggesting a high affinity of these peptide to HLA-A*02:01 molecules (p < 0.005). The gB_17–25 peptide demonstrated a moderate stabilization ability of HLA-A*02:01 molecules on the surface of T₂ cells, indicating only a medium affinity of this peptide for HLA-A*02:01 molecules. In contrast, the remaining four peptides produced no significant stabilization of HLA-A*02:01 molecules on the surface of T₂ cells. Overall, the six high binding affinity peptides found in Fig. 2A appeared to stabilize HLA-A*02:01 molecules on the cell surface of T₂ cells, as shown in Fig. 2B.

Collectively, these results indicate that, under these conditions, gB_17–25, gB_161–169, gB_441–449, gB_447–455, gB_561–569, and gB_675–683 bind with high affinity to HLA-A*02:01 and stabilize its expression on the surface of target cells. Because binding affinity of an epitope peptide may not always reflect its efficient presentation to T cells or predict its ability to stimulate T cells, we next used a combination of several immunological assays to assess the ability of each of the 10 gB peptides to recall functional HSV-specific CD8⁺ T cells in HSV-1–seropositive individuals.

Frequent IFN-γ–producing CD8⁺ T cells, specific to gB_17–25, gB_183–191, gB_342–350, gB_441–449, and gB_561–569 epitopes, detected in HLA-A*02:01–positive, HSV-seropositive individuals

CD8⁺ T cell responses specific to each gB peptide epitope described above were studied in HLA-A*02:01–positive, HSV-seropositive individuals. HLA-A*02:01–positive, HSV-seronegative individuals were used as controls. The characteristics of our study population with respect to gender, age, HSV-1 and HSV-2 seropositivity, and HLA-A*02:01 frequency distribution are detailed in Materials and Methods. Briefly, during a period of 10 y, 525 immunocompetent individuals, with an age range of 18–65 y (median, 32 y) who were seropositive or seronegative for HSV-1 and/or HSV-2, were enrolled in the study. Among HSV-seropositive individuals, 207 were HSV-1–seropositive and HSV-2–seronegative, 238 were HSV-2–seropositive and HSV-1–seronegative, and 11 individuals were both HSV-1– and HSV-2–seropositive. Because HSV-1 is the principal cause of ocular herpetic disease, this study focused solely on individuals who are HSV-1–seropositive and HSV-2–seronegative.

We first determined the frequency of CD8⁺ T cells specific to each of the 10 gB peptides in HSV-1–seropositive individuals using peptide/PE-labeled HLA-A*02:01 tetramers together with FITC-conjugated mAb specific to human CD8⁺ T cells (Fig. 3A). After peptide in vitro stimulation, we detected significantly higher percentages of tetramer⁺ CD8⁺ T cells specific to 5 of 10 peptides: gB_17–25, gB_183–191, gB_342–350, gB_441–449, and gB_561–569 in PBMCs of 10 HLA-A*02:01–positive, HSV-1–seropositive individuals compared with PBMCs of 10 HLA-A*02:01–positive, HSV-seronegative individuals (p < 0.005; Fig. 3A, 3B). The highest frequencies of tetramer⁺ CD8⁺ T cells, detected with or without in vitro expansion, were recorded against gB_183–191, gB_342–350, and gB_561–569 peptide epitopes (5.5–8.9%). Despite repeated attempts, with and without in vitro expansions, the remaining five gB peptide/HLA-A*02:01 tetramers consistently detected a low and nonsignificant frequency of CD8⁺ T cells in HLA-A*02:01–positive, HSV-seropositive individuals.

Because the frequency of CD8⁺ T cells specific to a given epitope may not always reflect their functionality, we next determined the ability of each of the 10 gB epitope peptides to recall proliferative and IFN-γ–producing epitope-specific CD8⁺ T cells in HLA-A*02:01–positive, HSV-1–seropositive individuals. HLA-A*02:01–positive, HSV-seronegative individuals were used as negative controls. CD8⁺ T cells were isolated from fresh peripheral blood of each group and stimulated for 5 d with individual gB peptides, and the number of epitope-specific IFN-γ–producing CD8⁺ T cells was determined by a 24-h ELISPOT assay. Our previous analysis revealed that 10 or more spots per 10⁵ CD8⁺ T cells gave a 99.5% probability for defining a positive response and 10 spot-forming cells per 10⁵ CD8⁺ T cells was therefore used as the cutoff point (1, 9). Significant IFN-γ–producing CD8⁺ T cell responses specific to 6 of 10 peptides, gB_17–25, gB_161–169, gB_183–191, gB_342–350, gB_441–449, and gB_561–569 were detected in HLA-A*02:01–positive, HSV-1–seropositive individuals (p < 0.005; Fig. 3C, 3E). In comparison, <10% of individuals showed positive T cell responses to the remaining four peptides (gB_286–294, gB_343–351, gB_447–455, and gB_675–683). Of note, significant IFN-γ–producing CD8⁺ T cell response specific to the gB_161–169 epitope was detected regardless of the low frequency of CD8⁺ T cells detected in HLA-A*02:01–positive, HSV-seropositive individuals (Fig. 3A, 3B). No significant IFN-γ–producing CD8⁺ T cell responses specific to any of the 10 gB epitopes were detected in HLA-A*02:01–positive, HSV-seronegative healthy controls.

Because lack of IFN-γ production by T cells may not always reflect lack of a T cell response (12, 35), we also studied the proliferative responses of CD8⁺ T cells from HSV-1–seropositive individuals labeled with CFSE and restimulated in vitro with each of the 10 gB peptides. The cells were stained with anti-human CD8-PE, and the number of dividing CD8⁺ T cells among a total number of 300,000 cells was determined by FACS, as we previously described (3). Briefly, gated CFSE⁺CD8⁺ T cells from HLAA*02:01–positive, HSV-1–seropositive individuals were examined for proliferation. As shown in Fig. 3D, gB_17–25, gB_161–169, gB_183–191, gB_286–294, gB_342–350, gB_441–449, and gB_561–569 peptide epitopes all induced significant CD8⁺ T cell proliferative responses in HLA-A*02:01–positive, HSV-1–seropositive individuals (p < 0.005). All but gB_286–294 epitopes were positive in the tetramer and ELISPOT assay above. As expected, no proliferative HSV-gB epitope-specific CD8⁺ T cell responses were detected in HLA-A*02:01–positive, HSV-seronegative controls (not shown).

Altogether, the HLA binding and the three immunological assays described above (tetramer frequency, IFN-γ production, and CFSE proliferation) point to gB_17–25, gB_183–191, gB_342–350, gB_441–449, and gB_561–569 as functional immunodominant gB epitopes that specifically recall frequent CD8⁺ T cells in HLA-A*02:01–positive, HSV-1–seropositive individuals.

Frequent IFN-γ–producing CD8⁺ T cells specific to immunodominant gB_342–350 and gB_561–569 epitopes detected in asymptomatic individuals

We next asked whether there was a difference in the frequency and function of CD8⁺ T cells specific to gB_17–25, gB_183–191, gB_342–350, gB_441–449, and gB_561–569 epitopes between HSV-seropositive SYMP and ASYMP individuals. The characteristics of the SYMP and ASYMP study populations used in this study with regard to HSV seropositivity and disease are described in detail in Materials and Methods. Briefly, among the HSV-1–seropositive population, 208 individuals are clinically healthy with no apparent, or history of, ocular, genital, or labial herpes lesions, and these are designated as ASYMP (no history of recurrent HSV disease). Ten patients were identified as HSV-1–seropositive, HSV-2–seronegative SYMP who had a clinically well-documented history of HSV-1 recurrent ocular disease (rHSK) such as herpetic lid lesions, herpetic conjunctivitis, dendritic or geographic keratitis, stromal keratitis, and iritis consistent with rHSK, with one or more episodes per year for the past 2 y.

We first compared the ability of all 10 potential gB epitopes to recall functional CD8⁺ T cells in 10 HLA-A*02:01–positive, HSV-1–seropositive ASYMP and 10 HLA-A*02:01–positive, HSV-1–seropositive SYMP individuals. CD8⁺ T cells, isolated from fresh peripheral blood in each group, were stimulated for 5 d with individual gB peptides and the number of gB-epitope–specific IFN-γ–producing CD8⁺ T cells was determined in a 24-h ELISPOT assay. As shown in Fig. 4A, 2 of 10 peptides, gB_342–350 and gB_561–569, induced significantly higher IFN-γ–producing CD8⁺ T cells from 10 sequentially studied HLA-A*02:01–positive, HSV-1–seropositive ASYMP healthy individuals as compared with 10 SYMP individuals, suggesting that those two epitopes are ASYMP epitopes (p < 0.005). In contrast, gB_183–191 and gB_441–449 peptides appeared to induce significantly higher IFN-γ–producing CD8⁺ T cells preferentially from SYMP individuals compared with ASYMP individuals, suggesting that those two epitopes are SYMP epitopes (p < 0.005). The gB_17–25 and gB_286–294 peptides appeared to induce significantly higher IFN-γ–producing CD8⁺ T cells in both SYMP and ASYMP individuals.

We then extended our analysis to assess the frequency of circulating CD8⁺ T cells, specific to each of the five immunodominant gB epitopes described above (gB_17–25, gB_183–191, gB_342–350, gB_441–449, and gB_561–569), in PBMCs isolated from five ASYMP and five SYMP individuals. The subdominant epitope gB_675–683 was used as control. To obtain an objective enumeration of gB epitope-specific CD8⁺ T cells, each tetramer was tested at three or four dilutions and the numbers (instead of percentage) of epitope-specific CD8⁺ T cells per 100,000 T cells were determined. Significantly more tetramer⁺CD8⁺ T cells specific to gB_342–350 and gB_561–569 epitopes were detected in ASYMP individuals, confirming these as ASYMP epitopes (p < 0.005; Fig. 4B, 4C). In contrast, significantly higher numbers of tetramer⁺CD8⁺ T cells specific to gB_183–191 and gB_441–449 epitopes were detected in SYMP individuals (p < 0.005; Fig. 4B, 4C), suggesting these as SYMP epitopes.

Asymptomatic gB epitope-specific CD8⁺ T cells displayed concurrent polyfunctional activities and were capable of recognizing naturally processed epitopes on HSV-1–infected target cells

We examined the cytotoxic function of CD8⁺ T cells from SYMP and ASYMP individuals. CD107a and CD107b are lysosomal-associated membrane glycoproteins that surround the core of the lytic granules in cytotoxic T cells (CTLs) (31, 32). Upon TCR engagement and stimulation by Ags in association with MHC molecules, CD107a/b are exposed on the cell membrane of cytotoxic T cells. Thus, the level of CD107a/b expression on the surface of CTLs is used as a direct assay for the epitope-specific CTL response (31). To assess whether gB epitope-specific CD8⁺ T cells display CTL activity, fresh PBMC-derived CD8⁺ T cell lines were generated from HLA-A*02:01–positive ASYMP and HLA-A*02:01–positive SYMP individuals following stimulation in vitro with individual ASYMP (gB_342–350 and gB_561–569) or SYMP (gB_183–191 and gB_441–449) peptides. The cytotoxicity of each of the four CD8⁺ T cell lines was measured against autologous target monocyte-derived dendritic cells either left uninfected (mock) or infected with UV-inactivated HSV-1, with a vaccinia virus expressing gB (VVgB), or with a control vaccinia virus expressing glycoprotein D (VVgD) by detecting the level of CD107a/b expression by FACS on gated CD8⁺ T cells. As shown in Fig. 5A, a high percentage of ASYMP gB_342–350 and gB_561–569 epitope-specific CD8⁺ T cells from healthy HLA-A*02:01–positive, HSV-seropositive ASYMP individuals expressed significant levels of CD107a/b (percentage of CD107a/b/CD8⁺ T cells) following incubation with either HSV-1– or VVgB-infected target monocyte-derived dendritic cells. In contrast, very few SYMP gB_183–191 and gB_441–449 epitope-specific CD8⁺ T cells upregulated CD107a/b after incubation with HSV-1– or VVgB-infected target cells. As expected, no significant percentage of SYMP or ASYMP epitope-specific CD8⁺ T cells upregulated CD107a/b after incubation with mock-infected or VVgD-infected target cells. The results indicate that ASYMP epitope-induced CD8⁺ T cells have cytotoxic activity against HSV-1–infected cells and are able to specifically recognize endogenously processed gB epitopes from both HSV-1– and VVgB-infected target cells. There was no CTL response against any peptides in individuals that were seronegative for HSV regardless of whether they were HLA-A*02:01–positive or HLA-A*02:01–negative (data not shown).

The levels of six different inflammatory cytokines (IL-2, IL-6, IL-8, IL-17, IFN-γ, and TNF-α) produced by CD8⁺ T cells from ASYMP versus SYMP individuals following in vitro restimulation with UV-inactivated HSV-1 (strain McKrae) were compared by the Luminex microbeads system. As shown in Fig. 5B, CD8⁺ T cells from SYMP patients produced high levels of IL-6, IL-8, and IL-17 (black-filled portion of the pie chart), consistent with differentiated inflammatory T cells. In contrast, T cells from ASYMP individuals produced more of the effector cytokines IL-2, IFN-γ, and TNF-α (black-filled portion of the pie chart). Fig. 5C summarizes the percentage of ASYMP and SYMP individuals who showed positive results for one or several CD8⁺ T cell functions following in vitro restimulation with either ASYMP or SYMP gB peptides. Overall, 95% of ASYMP individuals had HSV-specific CD8⁺ T cells with three to five functions, indicating their ability to maintain greater frequencies and display concurrent polyfunctional activities: 1) production of high levels of lytic granules upon TCR engagement (CD107a/b cytotoxic activity); 2) expression of IFN-γ (ELISPOT); 3) high proliferation (CFSE); 4) production of high levels of IL-2, IFN-γ, and TNF-α effector cytokines (or proinflammatory cytokines in the case of SYMP individuals); and 5) tetramer frequency (p < 0.005). In contrast, only 32% of SYMP individuals had HSV-specific CD8⁺ T cells with three to five functions whereas most had just one function, suggesting that CD8⁺ T cells from SYMP patients tended to be monofunctional and produced more inflammatory cytokines (p < 0.005). Finally, similar levels of HLA-A*02:01 expression were detected by FACS on the surface of APCs (both dendritic cells and macrophages) derived from SYMP and ASYMP individuals (Fig. 5D). This result suggests that the apparent polyfunctionality of CD8⁺ T cells detected in ASYMP individuals is not a result of a high level of expression of HLA-A0201 molecules on their APCs, but rather to an intrinsic multifunctionality of CD8⁺ T cells specific to ASYMP epitopes.

Collectively, the results indicate that 1) a particular breadth and specificity of virus-specific CD8⁺ T cell responses is associated with control of herpes disease, and 2) HSV-specific CD8⁺ T cells from ASYMP individuals displayed concurrent polyfunctional activities whereas CD8⁺ T cells from SYMP patients tended to be monofunctional.

Immunization with asymptomatic epitopes induced a CD8⁺ T cell–dependent protective immunity against ocular herpes in “humanized” HLA-A*02:01 transgenic mice

Because the in vitro antigenicity of CD8⁺ T cell epitopes may not always translate into functional in vivo immunogenicity, we next set up a preclinical vaccine trial using SYMP and ASYMP epitopes in a novel humanized HLA-A*02:01 transgenic mouse model.

To evaluate whether immunization with ASYMP CD8⁺ T cell epitopes confer any protection against ocular herpes, groups of susceptible HLA-A*02:01 transgenic humanized mice (n = 10 mice/group, BALB/c genetic background) were immunized s.c. twice, 21 d apart with the two ASYMP CD8⁺ T cell epitopes gB_342–350 and gB_561–569 (group 1) or with SYMP epitopes, gB_183–191 and gB_441–449 (group 2). These were delivered together with the CD4⁺ T helper PADRE epitope and emulsified in CpG₁₈₂₆ adjuvant. As negative control, mock-immunized mice received adjuvant alone (group 3). Two weeks after the second and final immunization, animals from all groups received an ocular HSV-1 challenge (2 × 10⁵ PFU, McKrae strain). Of note, the sequences of both SYMP and ASYMP epitopes are highly conserved between HSV-1 and HSV-2 strains (Table II); however, no significant homology exists between the amino acid sequences of the 10 HSV-1 gB T cell epitopes studied and the gB amino acid sequences of varicella zoster virus, EBV, and CMV (Table II). The pathology clinical scores observed in the ASYMP group were significantly lower than those observed in the SYMP group and the mock-immunized group (p = 0.001 for all; Fig. 6A). Furthermore, significantly lower viral loads were detected on day 7 postinfection in eye swabs of the ASYMP group compared with the SYMP and mock-immunized groups (p = 0.01; Fig. 6B). All animals in the ASYMP group survived lethal infection (100%) compared with only 40% survival in the SYMP group and 0% in the mock-immunized group (p < 0.005; Fig. 6C). Overall, there was a positive correlation between survival and the number of ASYMP CD8⁺ T cells detected in the draining lymph node (Fig. 6D, p < 0.005, r² = 0.7996, 95% confidence interval). Similar to what has been observed in ASYMP individuals, cultured CD8⁺ T cells from ASYMP-immunized mice had significantly higher proportions of cells with simultaneous expression of three to five functions, whereas SYMP-immunized mice had few or no detectable cells expressing more than three functions simultaneously (75.5 and 17.5%, respectively; p = 0.005; not shown). To verify the involvement of CD8⁺ T cells in the observed protection induced following immunization with the ASYMP CD8⁺ T cell epitopes, we performed an in vivo depletion of either CD4⁺ or CD8⁺ T cells in immunized mice before virus challenge using specific mAbs. As shown in Fig. 6E, depletion of CD8⁺ T cells, but not of CD4⁺ T cells, significantly abrogated protection against death induced by immunization with the ASYMP CD8⁺ T cell epitopes (p < 0.005), suggesting that CD8⁺ T cells, but not CD4⁺ T cells, were required for protection against lethal ocular herpes.

Altogether, these results indicate that immunization with ASYMP CD8⁺ T cell epitopes, but not with SYMP epitopes, decreased ocular herpes disease, decreased virus replication, and protected against lethal ocular herpes in susceptible HLA transgenic mice.