Immunodominant Epitopes in Herpes Simplex Virus Type 2 Glycoprotein D Are Recognized by CD4 Lymphocytes from Both HSV-1 and HSV-2 Seropositive Subjects This information is current as of September 28, 2021. Min Kim, Janette Taylor, John Sidney, Zorka Mikloska, Neil Bodsworth, Katerina Lagios, Heather Dunckley, Karen Byth-Wilson, Martine Denis, Robert Finlayson, Rajiv Khanna, Alessandro Sette and Anthony L. Cunningham

J Immunol 2008; 181:6604-6615; ; Downloaded from doi: 10.4049/jimmunol.181.9.6604 http://www.jimmunol.org/content/181/9/6604 http://www.jimmunol.org/ References This article cites 41 articles, 19 of which you can access for free at: http://www.jimmunol.org/content/181/9/6604.full#ref-list-1

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists by guest on September 28, 2021

• Fast Publication! 4 weeks from acceptance to publication

*average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2008 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Immunodominant Epitopes in Herpes Simplex Virus Type 2 Glycoprotein D Are Recognized by CD4 Lymphocytes from Both HSV-1 and HSV-2 Seropositive Subjects1

Min Kim,* Janette Taylor,* John Sidney,† Zorka Mikloska,* Neil Bodsworth,‡ Katerina Lagios,§ Heather Dunckley,¶ Karen Byth-Wilson,ʈ Martine Denis,# Robert Finlayson,‡ Rajiv Khanna,** Alessandro Sette,† and Anthony L. Cunningham2*

In human recurrent cutaneous herpes simplex, there is a sequential infiltrate of CD4 and then CD8 lymphocytes into lesions. CD4 lymphocytes are the major producers of the key cytokine IFN-␥ in lesions. They recognize mainly structural proteins and espe- cially glycoproteins D and B (gD and gB) when restimulated in vitro. Recent human vaccine trials using recombinant gD showed partial protection of HSV seronegative women against genital herpes disease and also, in placebo recipients, showed protection by Downloaded from prior HSV1 infection. In this study, we have defined immunodominant peptide epitopes recognized by 8 HSV1؉ and/or 16 HSV2؉ patients using 51Cr-release cytotoxicity and IFN-␥ ELISPOT assays. Using a set of 39 overlapping 20-mer peptides, more than six immunodominant epitopes were defined in gD2 (two to six peptide epitopes were recognized for each subject). Further fine mapping of these responses for 4 of the 20-mers, using a panel of 9 internal 12-mers for each 20-mers, combined with MHC II typing and also direct in vitro binding assay of these peptides to individual DR molecules, showed more than one epitope per 20-mers and promiscuous binding of individual 20-mers and 12-mers to multiple DR types. All four 20-mer peptides were http://www.jimmunol.org/ cross-recognized by both HSV1؉/HSV2؊ and HSV1؊/HSV2؉ subjects, but the sites of recognition differed within the 20- mers where their sequences were divergent. This work provides a basis for CD4 lymphocyte cross-recognition of gD2 and possibly cross-protection observed in previous clinical studies and in vaccine trials. The Journal of Immunology, 2008, 181: 6604–6615.

rimary genital herpes occurs in individuals without preex- and recurrent HSV infections, in recovery from infection, and in isting Abs who acquire HSV1 or HSV2 de novo. Recur- restricting HSV spread in the nervous system (3–5). They are P rent episodes of genital herpes occur despite the presence recruited to sites of productive HSV infection or reactivation in by guest on September 28, 2021 of circulating neutralizing antiviral Abs (1). Such episodes of in- the dorsal root ganglion and skin (3). In skin, the immunore- fections are of a shorter duration and less severity than primary active cells responsible for controlling the transmitted HSV in- infection. Preexisting immunity to HSV1 reduces the severity of clude the normal constituents of the squamous epidermis, ker- genital herpes caused by HSV2 infections (2). After virus reacti- atinocytes and Langerhans cells, and infiltrating cells: first vates from latency in the neurons of the dorsal root ganglion, it is predominantly monocytes/macrophages and CD4 lymphocytes, transported anterogradely to the axon terminus and then transmit- and later predominantly CD8 lymphocytes, as shown by immu- ted to the epidermal keratinocytes. A sequence of viral and immu- nohistochemistry and direct T cell cloning from lesions (6, 7). nologic interactions occurring both in the dorsal root ganglion and Infection of epidermal keratinocytes induces the secretion of a the recurrent herpetic lesion follows (3). sequence of chemokines and cytokines, which is reflected in the In humans and/or murine models, HSV-specific CD4 and whole lesion in vivo, that is, first IFN-␣ and ␤-chemokines and CD8 T lymphocytes play a central role in controlling primary then IL-12 followed by IL-1 and IL-6 (8). The ␤-chemokines may assist in chemotaxis of monocytes, CD4, and CD8 lym- phocytes into lesions. IFN-␣ and IFN-␥ synergize to inhibit *Centre for Virus Research, Westmead Millennium Institute, Westmead, New South Wales and University of Sydney, New South Wales, Australia; †La Jolla infection of keratinocytes after transmission from axon termini Institute for Allergy and Immunology, La Jolla, CA 92037; ‡Taylor Square Pri- (9). HSV1 or HSV2 down-regulates MHC class I expression by vate Clinic, Darlinghurst, New South Wales, Australia; §Parramatta Sexual Health epidermal keratinocytes, and this is reversed by IFN-␥ mainly Clinic, Parramatta, New South Wales, Australia; ¶Australian Red Cross Blood Service, Sydney, New South Wales, Australia; ʈWestmead Hospital, Westmead, secreted by CD4 lymphocytes infiltrating the lesion (10–12). New South Wales, Australia; #GlaxoSmithKline Biologicals, Rixensart, Belgium; The CD8 lymphocytes do not recognize the infected keratino- and **Queensland Institute of Medical Research, Herston, Queensland, Australia cytes until MHC I is restored on their surface by IFN-␥ secreted Received for publication May 20, 2008. Accepted for publication August 16, 2008. by CD4 lymphocytes. Both CD4 and CD8 CTLs have been The costs of publication of this article were defrayed in part by the payment of page isolated from genital lesions ex vivo and shown to have cyto- charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. toxic activity (13). The CD8 lymphocyte infiltrate appears to 1 This work was supported by the National Health and Medical Research Council of correlate with virus eradication from the skin (14). CD4 CTL Australia (Project Grant 253684 and Program Grant 358399). were also shown to recognize HSV2 tegument proteins espe- 2 Address correspondence and reprint request to Prof. Anthony L. Cunningham, cially VP16 and VP22 (13). These CD4 CTL probably act early, Westmead Millennium Institute, Darcy Road, Westmead, New South Wales 2145, and CD8 CTL late, in controlling HSV (3, 6). Australia. E-mail address: [email protected] Previous work from our laboratory has shown that both hu- Copyright © 2008 by The American Association of Immunologists, Inc. 0022-1767/08/$2.00 man CD4 and CD8 T lymphocytes recognize IFN-␥-stimulated www.jimmunol.org The Journal of Immunology 6605

HSV1-infected keratinocytes. Using vaccinia virus recombinants Herpes simplex viruses and peptides expressing HSV2 proteins and blood CD4 lymphocytes restimu- HSV2 strain 186 was grown and titered in Vero cells for subsequent use as lated in vitro, we showed that CD8 T lymphocytes recognized control. immediate early/early proteins, whereas CD4 T lymphocytes rec- Initial screening was performed using a library of gD2 20-mer peptides ognized late HSV1 or HSV2 structural proteins, especially glyco- provided by Dr. Martine Denis of SmithKline Beecham (SKB). Recombi- protein D (gD)3 (8, 15, 16), complementing earlier studies by nant gD2 was also kindly provided by SKB. Each of these 39 peptides was 20 aa long and had 10 residue overlaps. They were dissolved in DMSO to Zarling et al. demonstrating that gD can stimulate human CD4 a final volume of 2 mg/ml and stored at Ϫ80°C. helper cells (17). Parallel studies in mice have also showed CD4 For the second screening, the selected peptides based on the result of lymphocyte specificity for gD (18, 19). the first screening were truncated into nine serial 12-mers within the Successful trials of gD of HSV2 (gD2) immunization in mice four most frequently recognized 20-mer peptides (see peptides 2, 24, 30, and 34 in Fig. 1). The selected immunodominant 20-mers and serial and guinea pigs (20) preceded human trials of immunization with 12-mers were produced by Mimotopes. They had an 11-aa sequence over- recombinant gD2 vaccine mixed with the adjuvants alum and lap with adjacent peptides. The peptides were dissolved in DMSO at a deacylated monophosphoryl lipid A. The latter were shown to sub- concentration of 10 mM and stored at Ϫ80°C. stantially induce protection (Ͼ70%) (2) against genital herpes dis- ease in HSV1 and HSV2 seronegative but not in HSV1 seropos- Preparation of HSV2-specific CD4 effector T lymphocytes itive women. Prior natural HSV1 infection reduced development PBMC prepared by Ficoll-Hypaque gradient were stimulated with UV- of HSV2 genital herpes disease (2). gD2 has also been shown to inactivated HSV2 (12) and then cultured in RPMI 1640 (Invitrogen) sup- induce IFN-␥ secretion from the PBMC of similarly immunized plemented with 10% FCS (Invitrogen), 2 mM glutamine (Sigma-Aldrich, RF10), and 20 U/ml recombinant IL-2 (Roche). CD4 effector T lympho- patients when stimulated in vitro (21). Downloaded from cytes were enriched by CD8 beads (Miltenyi Biotec) immediately before In view of the importance of gD2 as an immunogen, the objec- the cytotoxicity experiment. The efficacy of CD8 T cell depletion was tive of this study was to identify the immunodominant peptides of checked routinely by flow cytometry using anti Leu 2a ϩ 2b Ab (BD gD2 recognized by bulk human CD4 lymphocytes in most HSV2 Biosciences) and showed Ͻ1% of CD8 T cell contamination. Peptide- seropositive subjects by screening a gD2 peptide library. We also specific CD4 effector T lymphocytes were, if necessary, restimulated with gamma-irradiated (5000 rads) and peptide-sensitized (2 ␮g/ml for1hat determined their MHC II restriction and whether such peptides ϩ 37°C) autologous PBMC.

were also recognized by HSV1 subjects. Furthermore, we corre- http://www.jimmunol.org/ lated these empirically defined CD4-MHC II-restricted epitopes Preparation of target cells for cytotoxicity assays with those predicted by the algorithm TEPITOPE (22). B-lymphoblastoid cell lines (LCLs) were established by EBV transfor- Previous similar human studies in the literature have been lim- mation of peripheral B cells as described previously (8) and used as ited to those defining a single peptide or a preliminary scan of gD target cells for 51Cr-release assay, as well as the HLA-DR and HLA-DQ of HSV1 (gD1) with large peptides using older insensitive T cell blocking assay. PHA blasts prepared as described previously (15) were proliferation assays, defining relatively few epitopes. MHC II re- used to examine the specificity of gD2 peptide-specific CD4 effector T cells. For each of gD2 peptides tested, 104 LCLs were sensitized with striction or HSV1/2 cross-reactivity was not examined (23, 24). 3 ␮g/ml of peptide and 1 ␮Ci of sterile 51Cr (Amersham Pharmacia These studies provide an empirical basis for cross-reactive and Biotech) added for 90 min at 37°C in 5% CO2. After sensitization, the possibly cross-protective epitopes between gD1 and gD2 sus- cells were washed three times by centrifugation for 7 min at 270 ϫ g by guest on September 28, 2021 pected from the vaccine studies. A vaccine effective against both with lukewarm RF10 before coculture with effector cells. The two pos- genital HSV1 and HSV2 infection and disease is required in view itive control tubes included target cells sensitized with the same con- centration of gD2 Ag and target cells infected with 10 PFU/tube of of the recent increasing incidence of genital HSV1 disease, espe- HSV2 instead of peptide. The controls were cell control (nonsensitized cially in adolescents (25). targets with effectors), spontaneous release (targets with no effectors), and total release control tubes containing only target cells labeled with 51 Materials and Methods Cr sodium solution. Patients and HSV type-specific serotyping Chromium release cytotoxicity assays Ϫ ϩ Blood was obtained from 16 HSV2 seropositive (HSV1 /HSV2 or 4 ϩ ϩ A total of 10 peptide-sensitized target cells were cocultured with CD4 HSV1 /HSV2 ) patients usually 1–4 mo after recurrences of genital her- ϩ Ϫ T cell effectors in each well of a Lumaplate (PerkinElmer Life and pes and 8 patients who were only HSV1 seropositive (HSV1 /HSV2 ), ϩ Analytical Sciences) at E:T ratios of 5:1 and 25:1 for 14 h at 37°C in usually 1–4 mo after recurrence of oral herpes. In 9 HSV 2 patients 5% CO in triplicate. The plates were prepared for analysis on a Pack- follow-up bleeds were taken at 6–8 mo and in 3 a further 12–15 mo. None 2 ard TopCount gamma counter (24, 29). The amount of 51Cr released had an episode of recurrent genital herpes within the previous 4 mo. In- was quantified as cpm using a gamma counter, and the percentage of formed consent was obtained from all the blood donors, and the study was specific cytotoxic activity was calculated using the following equation: approved by the Western Sydney Area Health Service Research and specific lysis ϭ [(experimental release Ϫ spontaneous release) ϫ 100]/ Ethics Committee. HSV2 type-specific serology was performed by (total release Ϫ spontaneous release). SEs of experimental cpm (trip- ELISA (Focus Diagnostics) and confirmed by Western blot. HSV1- licates) were Ͻ3%. Differences between the percentage of specific 51Cr specific serology was determined by Western blot (26, 27). release obtained with peptides or controls were assessed for statistical Ͻ Molecular MHC II typing significance by Student’s t test, with p 0.05 indicating recognition of peptide epitopes. HLA-DRB1 generic level typing was performed by PCR sequence-spe- cific oligonucleotide typing based on the method described in the 11th IFN-␥ ELISPOT assays International Histocompatibility Workshop (28). Sequencing-based typ- ing with an in-house method was used for HLA-DQB1 and to resolve To examine the immune response of CD4 T lymphocytes to truncated any ambiguities in HLA types involving HLA-DRB1*03, *08, *11, gD2 peptides, CD8 lymphocytes were depleted from isolated PBMC *12, *13, or *14 alleles. PCR product was sequenced in both forward using Miltenyi CD8 microbeads (Miltenyi Biotec) according to the ␥ and reverse directions using a BigDye Terminator kit (Applied Biosys- manufacturer’s instruction. IFN- production was measured as the im- ␮ tems) on an ABI 3100 sequencer (Applied Biosystems). Sequence data mune response following stimulating CD8-depleted PBMC with 10 M were analyzed using MatchTools sequencing analysis software (Ap- peptide by ELISPOT assay as described below. plied Biosystems). A Millipore plate with Immunobilon-P polyvinylidene difluoride mem- brane was coated with purified IFN-␥ capture Ab (1D1K, Mabtech) to a final concentration of 5 ␮g/ml in sterile PBS. The plate was washed three 3 Abbreviations used in this paper: gD, glycoprotein D; gD1, gD of HSV1; gD2, gD times with sterile PBS and blocked with RF10 for 2 h. After washing the of HSV2; LCL, lymphoblastoid cell line. plate three times with PBS, 5–7 ϫ 104 CD8 lymphocyte-depleted PBMC 6606 HSV-2 gD EPITOPES RECOGNIZED BY HSV1/2ϩ SUBJECTS were added to each well in 100 ␮l of RF10 supplemented with 10 ng/ml Table I. gD2 peptide recognition by CD4 lymphocytes of HSV1- or IL-12. Peptides, UV-inactivated HSV1 and HSV2, and PHA were added at HSV2-seropositive patients and correlation of MHC II type: recognition a final concentration of 10 ␮M, 0.5 multiplicity of infection, and 0.5 ␮g/ml, of 20-mer peptides by CD4 lymphocytes of 12 HSV2-seropositive respectively. After incubating the cells for 40 h at 37°C, the plate was patients with recurrent genital herpes a washed three times with PBS and then three times with PBS containing 0.05% Tween 20 (PBST). Biotinylated IFN-␥ detection Ab (Mabtech, ␮ Peptides Recognized 7-B6-1) diluted to 1 g/ml in PBST containing 1% BSA was added to each Patient HLA Type (Highest to Lowest) well. The plate was incubated for2hatroom temperature and washed six times with PBST. Streptavidin-alkaline phosphatase enzyme conjugate 1 DRB1*01, 11, DQB1*05, 07 10 ϭ34 Ͼ (3)4(5), (Bio-Rad) diluted in 1/1000 was added to each well. After incubating for 25 ϭ 26 Ͼ 27 45 min at room temperature, the plate was washed three times with PBST, 2 DRB1*01, 04, DQB1*05, 08 34 Ͼ 22, 2 and then four times with PBS. BCIP (5-bromo-4-chloro-3-indolyl phos- 3 DRB1*07, 07, DQB1*02, 02 12 Ͼ 24 ϭ 30 phate)/NBT (Bio-Rad) was added as substrate according to the manufac- 4 DRB1*04, 11, DQB1*07, 08 2 (1), (33)34(35) turer’s instructions followed by incubation for ϳ5 min at room temperature 5 DRB1*01, 15, DQB1*05, 06 35 (34)/36 Ͼ 2 in the dark. The reaction was stopped by extensive rinsing of the plate 6 DRB1*01, 15, DQB1*06, 10 1, 2, 3, 4 under running water with underdrain removed. The plate was dried over- 7 DRB1*01, 10, DQB1*02, 05 34 Ͼ 2 night in the dark. The spots were counted after scanned by KS ELISPOT 8 DRB1*04, 07, DQB1*07, 07 24 Ͼ 30 Ͼ 34 system (Zeiss) and counted manually (30). 9 DRB1*01, 10, DQB1*01, 07 26 Ͼ 10 Ͼ 4 10 DRB1*04, 04, DQB1*08, 10 24 Ͼ 30 Ͼ 12 In vitro HLA-DR peptide binding assays 11 DRB1*01, 12, DQB1*01, 07 33, 34 Ͼ 24 Ͼ 1, 2 12 DRB1*01, 15, DQB1*03, 07 1, 2, 3 Ͼ 35, 36 Frequently targeted peptides were tested for in vitro binding to 10 common a HLA-DR molecules. HLA-DR molecules were purified, and binding as- Peptide 1 ϭ aa 1–20; peptide 2, aa 11–30; peptide 3, aa 21–40, and so forth. Downloaded from says were performed essentially as previously described (31). Purified hu- Peptide x Ͼ peptide y indicates that peptide x was more immunostimulatory than ϭ man HLA-DR molecules were incubated with unlabeled gD peptides and peptide y. Peptide x peptide y indicates that peptide x exhibited the same immu- 0.1–1 nM 125I-radiolabeled probe peptides for 48 h. MHC binding of the nogenicity as peptide y. radiolabeled peptide was determined by capturing MHC peptide complexes on LB3.1 (anti-HLA-DR) Ab-coated Lumitrac 600 plates (Greiner Bio- One) and measuring bound cpm using the TopCount (Packard Instrument) then their blood was screened by ELISA and/or Western blot for microscintillation counter. The binding data were analyzed and IC50 (nano- HSV1 and HSV2 Abs. In this study, all patients with genital herpes http://www.jimmunol.org/ molar) was determined as previously described (31, 32). were HSV2-seropositive and all patients with oral herpes were T cell epitope prediction HSV1-seropositive. In studies of HSV1/2 cross-reactive T cell epitopes, only HSV1ϩ/HSV2Ϫ or HSV1Ϫ/HSV2ϩ patients were The whole gD2 sequence was loaded into the peptide prediction software used; after screening, HSV1ϩ/HSV2ϩ patients were excluded to (TEPITOPE) to predict promiscuous epitopes. The prediction threshold was set at 5% and all the available MHC II molecules were selected to avoid ambiguity in interpretation of HSV1/2 cross-reactive match with predicted epitopes (33). epitopes (Tables I and II).

Statistical analysis of type-specific peptide epitopes Definition of gD2 peptide epitopes by human CD4 lymphocytes by guest on September 28, 2021 The statistical software package SPSS for Windows version 14 was used to The entire sequence of gD2 and the key peptide 20-mer epitopes analyze the data. For each of the four peptides (2, 24, 30, and 34), the CD4 identified in this study are shown in Fig. 1a. The amino acid se- lymphocyte responses by IFN-␥ production to nine internal peptides were quences of the key 20-mer peptides of gD2 are compared with available on each subject. The distributions of these results clearly departed those of gD1 in Fig. 1b. from normality for each peptide. We therefore decided to compute the rank scores (one to nine) of the internal peptides for each peptide within each Reactivity of CD4 lymphocytes from HSV2-seropositive sub- subject. Statistical analyses were performed separately for each peptide jects with a history of a recurrent genital herpes was tested against using these rank scores. the full panel of 20-mer peptides using the 51Cr-release cytotox- There were four subjects positive only for HSV2 and eight subjects icity assay and confirmed by IFN-␥ ELISPOT assays. Initially au- positive only for HSV1. Repeated-measures ANOVA was used to test tologous LCLs from 12 patients with recurrent genital herpes were whether there was a significant interaction between the effect of internal peptide number (one to nine) and HSV status (HSV1 only or HSV2 only) sensitized with each peptide of the entire gD2 library and tested on the rank scores for each peptide. Internal peptide number was treated as against patients CD4 lymphocytes in a bulk cytotoxicity assay at a within-subject factor and HSV status as a between-subject factor. Sta- E:T ratios of 5:1 and 25:1. Patient CD4 lymphocytes were tested tistical tests of interaction typically require larger sample sizes to achieve against each peptide individually but, for logistical reasons, usu- adequate power compared with tests of main effects. Since there were only a total of 12 subjects available, a p-value for interaction of Ͻ0.1 was ally in two overlapping sets of 20 or 19 consecutive 20-mer pep- considered to be statistically significant. tides as odd- or even-numbered peptides: 1, 3,…, 33 or 2, 4,…, 34 For each peptide, the amino acid sequences of the nine internal peptides per bleed for each patient. were compared between HSV1 and HSV2. In this way internal peptides Experimental results for E:T ratios of 5:1 and 25:1 were found with at least two major differences in (unlike) amino acids were identified. not to be significantly different ( p Ͼ 0.5) from each other (data not If a statistically significant interaction was detected between the effects of internal peptide number (one to nine) and HSV status, a Mann-Whitney U shown), so only the results for selected experiments with an E:T test was then used to compare the rank scores between the HSV groups for ratio of 5:1 are presented. In the 12 patients initially tested (shown those internal peptides identified as having major amino acid discrepancies in Table I) and another 4 (in Table II), there was CD4 lymphocyte between HSV1 and HSV2. Two-tailed tests with a 5% significance level recognition of at least some gD2 peptides varying from two to six were used for these comparisons. of the 20-mer peptides tested in one batch. The peptides 2, 24, 30, and 34 were the most frequently recognized even-numbered pep- Results tides, followed by peptides 10, 12, and 26 (Fig. 2 and Table I), and Subjects this was confirmed by IFN-␥ ELISPOT. Where one of these even- Overall, 16 patients with a history of genital herpes and 8 patients numbered peptides was recognized, either or both flanking odd- with oral herpes were studied in two stages: general screening for numbered peptides were often recognized (e.g., peptides 1 and 3 gD2 20-mer peptide recognition in 12 patients with genital herpes for peptide 2 or peptides 33 and 35 for peptide 34; Table I). For and then specific studies in 4 with genital herpes and 8 with oral follow-up studies at 6–8 mo in nine subjects and at 12–15 mo in herpes. A history of recent or current active lesions was noted and three subjects after the first bleed, CD4 cytokine responses to the The Journal of Immunology 6607

Table II. gD2 peptide recognition by CD4 lymphocytes of HSV1- or HSV2-seropositive patients and correlation of MHC II type: gD2 12-mer peptide recognition by HSV1- and HSV2-seropositive patientsa

Patient HSV Serotype HLA Type Recognized gD2 Peptides (Highest to Lowest)a Parental 20-mer Peptide

13 HSV1Ϫ/2ϩ DRB1*04011, 13021, 2-5, 2-4, 2-3, 2-2, 2-1, 2 2 DQB1*0302, 0609 24 30 30 34-1, 34-3, 34-2 34 14 HSV1Ϫ/2ϩ DRB1*13011, 1305, 2-2, 2 2 DQB1*0301, 0603 24-4, 24-5 24 30-6, 30-2, 30-4, 30-5, 30-7 30 34-9 34 15 HSV1Ϫ/2ϩ DRB1*0404, 15011, 2-3, 2-5, 2, 2-9 2 DQB1*0302, 0602 24-4, 24-1, 24, 24-3 24 30-6 30 34-3 34 16 HSV1Ϫ/2ϩ DRB1*03011, 04011, 2 DQB1*0201, 0301 24 30 34-3, 34-2, 34-6 34 17 HSV1ϩ/2Ϫ DRB1*0404, 13011, 2 DQB1*0302, 0603 24-2, 24-3, 24-4 24 Downloaded from 30-4, 30-3 30 34 18 HSV1ϩ/2Ϫ DRB1*0404, 15011, 2 DQB1*0302, 0602 24 30-5, 30, 30-6, 30-3 30 34-7 34 ϩ Ϫ

19 HSV1 /2 DRB1*03011, 0406, 2-3 2 http://www.jimmunol.org/ DQB1*0201, 0302 24-2, 24-4, 24-9 24 30-7, 30-5 30 34-6, 34-5, 34-3, 34-7 34 20 HSV1ϩ/2Ϫ DRB1*11011, 11041, 2-3 2 DQB1*0301, 0301 24-4, 24-1, 24-5, 24, 24-3, 24-2 24 30-5 30 34-7, 34-8 34 21 HSV1ϩ/2Ϫ DRB1*01011, 07011, 2-3 2 DQB1*0202, 0501 24 30 34 by guest on September 28, 2021 22 HSV1ϩ/2Ϫ DRB1*04011, 04011, 24-3 2 DQB1*0301, 0301 24 34 23 HSV1ϩ/2Ϫ DRB1*07011, 15011, 2-5, 2-2, 2-1, 2-4, 2-7, 2 2 DQB1*0303, 0602 24-1, 24-3, 24-7 24 30 34-6 34 24 HSV1ϩ/2Ϫ DRB1*01011, 04051, 2-2 2 DQB1*0202, 0501 24-3, 24-4 24 30-4, 30-1 30 34

a 34-1 refers to the first of the nine 12-mers within 20-mer peptide 34.

peptides declined to minimal (by 51Cr-release assay), although get cells. Fig. 3 shows a representative experiment for good responses to gD2 and whole HSV2 Ag were retained (data peptide 12. not shown). MHC II specificity and its nature were determined by incubation of peptide-sensitized target cells with anti-HLA-DR and anti- Definition of peptide epitopes within the native viral protein HLA-DQ Abs. As shown in Table III, CD4 lymphocyte recogni- gD2 and their MHC II restriction tion of peptides 1 and 33 was ablated by anti-HLA-DR Abs. Pep- To confirm that these peptides truly contained epitopes recog- tide 24-4 appeared to be restricted only by HLA-DR (Table III). nized by gD2-specific CD4 lymphocytes, autologous effector T However, with peptide 30-5 there was inhibition by both anti- cell lines from three HSV2-seropositive patients were restimu- HLA-DR and anti-HLA-DQ Abs to similar degrees (data not lated in vitro with peptides 2 or 12 and the effectors were tested shown). MHC II typing of all patients and the predominant pep- initially by 51Cr-release assay against peptide-sensitized target tides recognized by these patients are shown in Tables I and II. autologous PHA blasts and, as controls, target cells infected with HSV2, and recombinant gD2-incubated targets. These Fine mapping of the minimal epitopes within peptides 2, 24, 30, CD4 effector T lymphocytes when stimulated by peptides 2 or and 34 12 showed high levels of specific activity only against either Serial 12-mers within the 20-mers 2, 24, 30, and 34, selected as the peptide 2- or 12-sensitized target cells, respectively, and, in most frequent immunodominant epitopes, were used for fine map- each case, also against gD2-sensitized and HSV2-infected tar- ping of CD4 lymphocyte epitopes. 6608 HSV-2 gD EPITOPES RECOGNIZED BY HSV1/2ϩ SUBJECTS Downloaded from http://www.jimmunol.org/

FIGURE 2. gD2 peptides recognized by CD4 lymphocytes from two HSV2-seropositive patients with recurrent genital herpes. CD4 lympho- cytes were enriched by negative selection as outlined in Materials and Methods and stimulated with a UV-inactivated HSV2 Ag. Target cells were LCLs incubated with each of the individual peptides or gD2 and with 51Cr. Exogenous recombinant gD2 was used as the positive control. Effectors FIGURE 1. The amino acid sequences of gD2, as well as key gD1 and and targets were mixed in a ratio of 5:1. Each peptide was tested in trip- gD2 epitopes for induction of CD4 lymphocyte responses. Each 20-mer licate, and histograms represent means. Dashed line represents mean of no peptide analog had a 10-aa overlap with adjacent peptides. Nine 12-mers

peptide controls and 3 SDs. Key peptide recognition was later checked by by guest on September 28, 2021 were synthesized from each 20-mer of peptides 2, 24, 30, and 34 for fine IFN-␥ ELISPOT (see Materials and Methods). Two representative results mapping. Each 12-mer overlapped by 11 aa with adjacent 12-mer peptides (A and B, patient 7 and 8 in Table I, respectively) are shown. (a). The differences in amino acid sequences between HSV1 (strain 17) and HSV2 (clinical isolate 356.2038) for key peptides tested are shown in b. MHC II alleles (22, 33) (e.g., peptide epitopes within the 20-mer peptide 2 for HLA DRB1*0101, *0301, *0402, *0701, *1101/ In some patients definition of minimal epitopes was clear, as 1104/1106, *1305, and *1501/1502). In some cases, there was a shown in Fig. 4A, where there is an increasing recognition of pep- good correlation between predicted and empirically determined tides 2.1–2.4 and no recognition thereafter within peptide 2 (pa- epitopes (e.g., within peptides 2 and 10) whereas in others the tient 13, Fig. 4A). However, in other patients it was very difficult to define minimal epitopes within the 20-mers (patient 14, Fig. 4B). Furthermore, as for the 20-mers it was impossible to assign recognition of individual 12-mers to specific MHC II (HLA-DR) alleles. Indeed, comparisons of the patterns of recognition between patients of different HLA-DR and HLA-DQ types indicate cross- recognition of specific 12-mer and 20-mer peptide epitopes across different MHC II alleles. In completely asymptomatic HSV1-seropositive patients, re- sponses to the 20- and 12-mer peptides were lower and variable, with only 33% positive responding to six major immunodominant peptides (data not shown).

Comparison of empirically defined epitopes with those predicted by the TEPITOPE algorithm According to the TEPITOPE algorithm, there was a very high FIGURE 3. Verification of specificity of peptide-specific T cell lines. A CD4 (cytotoxic) lymphocyte cell line from patient 3 was restimulated with density of predicted epitopes in gD2 for the most common MHC peptide 12 through two cycles and then specificity of the cell lines was II alleles, whether the predictive levels were set at 2% or 5% (e.g., tested against autologous target PHA blasts with a range of peptides in- four epitopes recognized by at least 10 alleles were predicted cluding the peptide stimulator (12), the overlapping flanking peptides and, within peptide 2 when set at the 5% level) (data not shown). The as positive controls, target cells incubated with gD2 or infected with HSV2. algorithm also predicted cross-recognition and binding of different The E:T ratio was 20:1. Experiments were conducted in triplicate. The epitopes within 12-mer or 20-mer peptides according to different representative results of three experiments are shown. The Journal of Immunology 6609

Table III. HLA-DR/DQ specificity of HSV2 peptide recognition by lymphocytes responded to these peptides at lower levels than those HSV-infected patients by patients who were HSV1Ϫ/HSV2ϩ, whereas in other cases CD4 lymphocyte responses were as strong. The comparison of the % Inhibition of CD4 Lymphocyte Response to Peptides four key selected 20-mer peptides and their serial overlapping 12- Measured as 51Cr-Release Assay or IFN-␥ ELISA mers for cross-reactivity between HSV2 and HSV1 are shown in Peptides No Blocking Anti-HLA-DR Anti-HLA-DQ Fig. 6. All were recognized by HSV1ϩ/HSV2Ϫ patients. Because 20-mers there was promiscuous recognition of peptide epitopes across dif- 1088Ϯ 6% Ϫ8 Ϯ 6% ferent MHC II alleles, it was possible to conduct a comparative 33 0 90 Ϯ 7% 10 Ϯ 6% quantitative study of recognition of the nine internal serial 12-mers Ϯ Ϫ Ϯ 35 0 77 6% 11 7% for each 20-mer peptides 2, 24, 30, and 34 by CD4 lymphocytes 12-mers ϩ Ϫ Ϫ ϩ 24-4 0 64 Ϯ 16% 0 from eight HSV1 /HSV2 and four HSV1 /HSV2 patients. Controls As shown in Fig. 6, the patterns of recognition were very similar gD2 0 38 Ϯ 5% 15 Ϯ 5% between CD4 lymphocytes of HSV1ϩ and HSV2ϩ patients for Ϯ Ϯ HSV2 0 33 4% 12 4% peptide 30 and mostly similar for peptides 2, 24, and 34. Peptide LCLs were preincubated with either anti-HLA-DR, anti HLA-DQ (BD Pharmin- 34 showed the greatest differences in predominant recognition of ␮ gen) or isotype control antibodies for 30 min at a final concentration of 50 g/ml prior internal peptides at the N- and C-terminal regions for HSV1ϩ and to peptide pulse for 2 h. 20-mer peptides 1, 33, and 35, control recombinant gD2, and ϩ HSV2 were pulsed onto LCLs and incubated with CD4 T lymphocytes of patient 4 HSV2 patients. For peptides 2 and 24, recognition of the 12-mers and then tested in the 51Cr-release assay. The 12-mer peptide 24-4 was pulsed onto in the amino half of the 20-mer was dominant. LCLs and incubated with CD4 lymphocytes of patient 20. The supernatant was tested Downloaded from for IFN-␥ by ELISA (Endogen) according to the manufacturer’s instruction. Exper- Peptides 30 and 24 have the greatest degree of homology be- iments were performed in triplicate, and results expressed as means Ϯ SD. tween the HSV1 and HSV2 sequences (Fig. 1). The slight differ- ences in responses to some of the 12-mers in peptides 2 and 34 correlation was poor (e.g., peptides 24 and 34 (data not shown)). may be explained by unlike amino acid substitutions at positions 4, This promiscuous cross-recognition and density of predicted 14, and 15 in peptide 2 and positions 3, 6, 8, and 10 in peptide 34, epitopes was consistent with the in vitro binding data (Table IV) especially prolines at positions 3 and 10. No statistically significant http://www.jimmunol.org/ and the CD4 T cell responses by IFN-␥ ELISPOT (Fig. 5). interaction between the effects of internal peptide number and HSV status was found for peptide 30 ( p ϭ 0.239), nor were there Cross-reactive recognition of HSV2 peptides by CD4 any major amino acid differences between HSV1 and HSV2 in this lymphocytes for HSV1-seropositive subjects 20-mer. As was therefore anticipated, none of the Mann-Whitney Several of the peptides were cross-recognized by patients who tests of the rank scores by HSV group were statistically significant were HSV1ϩ/HSV2Ϫ (Fig. 4, C and D). In some cases, their CD4 for any of the nine internal peptides. by guest on September 28, 2021

FIGURE 4. Recognition of gD2 12-mer peptides by CD4 lymphocytes of HSV1ϩ and HSV2ϩ patients. IFN-␥ production by CD4 lymphocytes of two HSV1Ϫ/HSV2ϩ patients (patient 13 and 14; A and B, respectively) and two HSV1ϩ/HSV2Ϫ patients (patient 19 and 20; C and D, respectively) was measured by IFN-␥ ELISPOT after stimulation with gD2 peptides. The dashed line indicates the threshold for recognition, which was 3 SDs above the mean value of nonstimulated CD4 lymphocyte response (as in Fig. 2). 6610 HSV-2 gD EPITOPES RECOGNIZED BY HSV1/2ϩ SUBJECTS

Table IV. HLA-DR peptide binding assays on the gD2 peptidesa

HLA

DRB1* DRB1* DRB1* DRB1* DRB1* DRB1* DRB1* DRB1* DRB1* DRB3* Peptide No. Sequence 0101 0301 0401 0404 0405 0701 1101 1302 1501 0101

2 AALLVVAVGLRVVCAKYALA 18 708 1335 1169 2900 107 — 245 96 — 2.1 AALLVVAVGLRV 16 — 2897 — — 4.6 — 14 3212 — 2.2 ALLVVAVGLRVV 3.8 2005 — — — 6.8 — 30 869 — 2.3 LLVVAVGLRVVC 27 — 801 553 657 70 — 535 242 — 2.4 LVVAVGLRVVCA 102 ————895 — — 4779 — 2.5 VVAVGLRVVCAK 91 — — 3570 1988 874 — 4243 1393 — 2.6 VAVGLRVVCAKY 191 — — 1420 874 1072 — 2563 495 — 2.7 AVGLRVVCAKYA 98 4930 — 4080 3074 50 — 2037 251 — 2.8 VGLRVVCAKYAL 61 — — 1513 754 22 — 2426 148 — 2.9 GLRVVCAKYALA 19 1638 541 893 907 67 4461 3353 124 — 24 KAYQQGVTVDSIGMLPRFTP 77 14 57 1664 1989 177 3562 526 — 86 24.1 KAYQQGVTVDSI 50 842 73 4328 1018 498 — — — 4031 24.2 AYQQGVTVDSIG 471 — 212 — — 2719 ———— 24.3 YQQGVTVDSIGM — — 4410 — — 389 ———— 24.4 QQGVTVDSIGML — 365 — — — 299 — 4640 — 262

24.5 QGVTVDSIGMLP — 549 — — — 4747 — — — 580 Downloaded from 24.6 GVTVDSIGMLPR — 108 — — — 792 — 4642 — 657 24.7 VTVDSIGMLPRF 4089 829 4787 — — 1603 — — — 1257 24.8 TVDSIGMLPRFT —————1688 — — — 2805 24.9 VDSIGMLPRFTP —————————— 30 PELVPEDPEDSALLEDPAGT 219 2992 4709 — — — — — 160 1855 30.1 PELVPEDPEDSA 1192 — — — — — — — 3089 — 30.2 ELVPEDPEDSAL —————————— http://www.jimmunol.org/ 30.3 LVPEDPEDSALL —————————— 30.4 VPEDPEDSALLE —————————— 30.5 PEDPEDSALLED —————————— 30.6 EDPEDSALLEDP —————————— 30.7 DPEDSALLEDPA — — — — — — — — — 1616 30.8 PEDSALLEDPAG — — — — — — — — — 2049 30.9 EDSALLEDPAGT —————————— 34 HAPAAPANPGLIIGALAGST 27 — 589 ——487 — 363 3814 — 34.1 HAPAAPANPGLI 3551 ————————— 34.2 APAAPANPGLII 1267 ——————642 ——

34.3 PAAPANPGLIIG 796 — — — — — — 1187 — — by guest on September 28, 2021 34.4 AAPANPGLIIGA 495 ————2775 — 452 1500 — 34.5 APANPGLIIGAL 789 ————4850 — 1329 4714 — 34.6 PANPGLIIGALA 202 — 3699 — — 1558 — 1668 3440 — 34.7 ANPGLIIGALAG 224 ————2508 ———— 34.8 NPGLIIGALAGS 38 ————2105 — 1391 — — 34.9 PGLIIGALAGST 58 — 3463 — — 1110 — 1036 — — a Ͼ Ͻ A dash indicates IC50 of 5000 nM. Significant affinity thresholds of 1000 are shown in boldface type.

However, peptides 2, 24, and 34 demonstrated a statistically some internal peptides, there was a statistically significant inter- significant interaction between the effects of internal peptide num- action between the effects of internal peptide number (1 to 9) and ber and HSV1/2 status ( p-value for interaction 0.098, 0.057, HSV status (HSV1 only or HSV2 only) on the rank scores. Con- 0.068, respectively). For peptide 2, there were major amino acid sidering the total number of internal peptides (36), for the nine in differences between internal peptide 4 and, to a lesser extent, for which major amino acid differences were identified, statistically internal peptide 3 of HSV1 and HSV2. Mann-Whitney tests of the significant differences in 12-mer recognition according to HSV1/2 rank scores by HSV1/2 status showed a statistically significant were detected in four. Among the 27 remaining internal peptides in difference in recognition by HSV1 and HSV2 patients for internal which no major amino acid difference was identified, there were peptide 4 ( p ϭ 0.048) but not for internal peptide 3 ( p ϭ 0.461). only two internal peptides with a statistically significant difference For peptide 24, there was no amino acid difference between between the rank scores by HSV1/2 status (internal peptides 2 and internal peptides 2–8 of HSV1 and HSV2. Nevertheless, Mann- 6 for peptide 24). Thus, HSV1/2 sequence comparisons correlated Whitney tests of the rank scores by HSV1/2 status showed a sta- with observed 12-mer peptide recognition (4 of 9 vs 2 of 27, p ϭ tistically significant difference for internal peptides 2 ( p ϭ 0.048) 0.014, Fisher’s exact test), providing validation for the experimen- and6(p ϭ 0.048) but not for the other internal peptides. tal results. For peptide 34, there were significant amino acid differences between internal peptides 2 and 3 of HSV1 and HSV2 while a Binding of peptides to HLA-DR molecules proline was present at the second amino acid of internal peptide 7. We next assessed whether the set of epitopes identified showed Mann-Whitney tests of the rank scores by HSV group showed promiscuous HLA-DR binding affinity. Each of the 20-mer statistically significant differences for all of these internal peptides epitopes was tested for its capacity to bind to a panel of 10 com- 2(p ϭ 0.048), 3 ( p ϭ 0.028), and 7 ( p ϭ 0.016). mon DR molecules. As shown in Table IV, peptides 2 and 24 were In summary, for the two 20-mer peptides in which major amino found to be degenerate binders, binding 50% or more of the mol- Ͻ acid differences between HSV1 and HSV2 were identified for ecules tested with high affinity (IC50 of 1000 nM). These The Journal of Immunology 6611 Downloaded from

FIGURE 5. gD2 peptides recognized by HSV-infected patients. Each bar indicates the recognized 12-mer or 20-mer peptide. http://www.jimmunol.org/ epitopes also bound several other specificities with intermediate DRB1*0701, and DRB1*1501 was observed with truncations in- affinity (IC50 1000–5000 nM). Peptide 34 was less degenerate, but corporating either the N- or C-terminal regions of the peptide. In still bound 4 of the 10 DR molecules tested with high affinity. By most other cases, optimal binding appeared to be more concen- contrast, peptide 30 bound only DRB1*0101 and DRB1*1501. trated in one region of the 20-mer peptide. However, because the The peptide 35 20-mer, which was less immunogenic than the binding groove of the DR molecule allows for multiple core align- other four, also bound to DRB1*0101, *0404, *0701, and *1302 ments even in the context of a 12-mer peptide, the possibility that (unpublished observation). each of these regions represents multiple closely nested epitopes Analysis of the pattern of binding of the corresponding 12-mer cannot be excluded. truncations of each epitope was also undertaken. The data, shown Taken together with the Ab blocking data, these binding data by guest on September 28, 2021 in Table IV, revealed in several cases that binding of the 20-mer support DR restriction in the majority of cases for peptides 2, 24, peptide may be achieved by use of more than one core region. For and 34 and suggest that binding in several contexts may involve example, in the case of peptide 2, binding to DRB1*0101, more than one core region. These data also suggest that, at least in

FIGURE 6. Immune response profiles of HSV1ϩ or HSV2ϩ patient to nine serial gD2 12-mer peptides within immunodominant 20-mers (peptides 2, 24, 30, and 34). The Mann-Whitney U test has been performed to obtain y values, which have been represented as estimated marginal means (Materials and Methods). Triangle symbols represent HSV1Ϫ/HSV2ϩ patients and square symbols represent HSV1ϩ/HSV2Ϫ patients. 6612 HSV-2 gD EPITOPES RECOGNIZED BY HSV1/2ϩ SUBJECTS some cases, peptide 30 is DQ restricted (recently confirmed by In some instances, differences were observed between the bind- HLA-DQ binding assay, data not shown). In most cases, there was ing assays and the CD4 lymphocyte response in that a peptide a good correlation between peptide binding to HLA-DR and T cell binding to a specific DR type may be recognized in one patient responses to the peptides in subjects of similar HLA-DR types expressing that molecule, but not in another. These differential (e.g., 8 of 10 HLA DRB1*0101-positive subjects recognize responses may reflect the unique T cell repertoire of each donor, or epitopes in peptide 2). differences in peptide processing or differences in fine specificity of HLA-DR types. The TEPITOPE program, which predicts MHC II-specific pep- Discussion tide epitopes within a protein, suggested possible reasons for the In this work, we have demonstrated that CD4 lymphocytes of all difficulty in defining individual 20-mer epitopes in some patients HSV2-positive patients with genital herpes recognized at least two (22). At the most stringent levels, several overlapping or clustered to six of an overlapping library of peptide 20-mers spanning the MHC II-restricted epitopes were predicted to occur (within 20- whole glycoprotein gD2 sequence, including the leader sequence. mers) at various places within the gD2 molecule. They were also Bulk CD4 lymphocytes were used to identify specific epitopes in predicted to be promiscuously recognized by a number of MHC II an effort to minimize the bias induced by using T cell clones. alleles (33). Thus, our T cell response and HLA-DR binding data However this did reduce the magnitude of response, requiring the were mostly consistent with that predicted by the TEPITOPE al- ␥ 51 use of IFN- ELISPOT or Cr-release assay for maximum sen- gorithm in terms of densely clustered epitopes within 20-mers and sitivity. The latter was a valid screening assay, as previously gD- their promiscuous recognition. Nevertheless, there was an incom- specific CD4 lymphocytes have shown to be cytolytic by this assay plete correlation of empirical and predicted epitopes (e.g., the con- Downloaded from (12). This is consistent with the relatively low frequency of 0.2– sistent T cell recognition and binding of epitopes in 330–350 re- 0.4% of gD peptide-specific CD4 lymphocytes in vivo shown by gions by patients with a broad range of HLA-DR alleles). ␥ intracellular cytokine staining for IFN- (34, 35). This low fre- TEPITOPE is clearly one of the best of the MHC II algorithms and quency of responder in HSV-seropositive patients contrasts with may be useful to quickly identify some epitopes within a given those in the other herpesvirus infections, CMV (1.2%), and EBV ϩ protein. However, it does not identify all of them and clearly the (35). In nine HSV2 patients with recurrent genital herpes, their results of the algorithms should be checked experimentally. http://www.jimmunol.org/ CD4 lymphocyte responses (by 51Cr-release assay) to gD2 pep- Cross-recognition of each of four major 20-mer peptide epitopes tides but not whole gD2 or HSV2 declined to undetectable when (peptides 2, 24, 30, and 34) and of the nine serial overlapping re-tested 6–8 and 12–15 mo later. The kinetics of these enhanced minimal T cell epitopes within each 20-mer was examined in pa- memory CD4 lymphocyte responses to gD2 peptides soon after tients who were seropositive only for HSV1 or for HSV2. All four clinical recurrences show similarities to the kinetics of IFN-␥ re- 20-mer peptides were cross-recognized. Because of the promiscu- sponses from PBMC after recurrences of orolabial herpes (6). ity of recognition of the serial internal 12-mer epitopes across dif- However, the detection of IFN-␥ secretion by CD4 lymphocytes ferent HLA-DR alleles, consensus recognition patterns among responding to all peptides in some asymptomatic patients indicates

eight HSV1-seropositive and four HSV2-seropositive patients be- by guest on September 28, 2021 the importance of also considering asymptomatic oral shedding in came apparent. For one of these 20-mers (peptide 30), the pattern boosting responses. Identification of key 20-mer peptide epitopes of recognition along the 20-mer was very similar by HSV1- or (shown in Table I) was confirmed by both 51Cr-release assays HSV2-seropositive patients, reflecting the sequence homology in using LCLs as targets and ELISPOT assays, and the specificity of gD2 peptide-specific T cell lines was verified by 51Cr-release assay those regions of D1 and D2. In another two (peptides 2 and 34), using PHA blasts as target cells to reduce nonspecific lysis. there was similarity in recognition in one but not in another region Serially truncated 12-mers within the initially identified 20-mer of the 20-mer. These differences in the two 20-mers were corre- peptide epitope were used to define the minimal T cell epitopes in lated with more marked differences in HSV1/2 sequence homology most patients, but in others this was impossible, suggesting clus- (Fig. 1). In peptide 24, there was a slight difference in recognition tering or overlapping of more than one epitope in the 20-mer. It patterns for HSV1 and HSV2, but this was not obvious from in- was also difficult to assign minimal 12-mer peptides to an individ- spection of sequences. Differences in binding of the 12-mer pep- ual HLA-DR or HLA-DQ allele. However, in three patients tested tide to the key binding pockets of different HLA alleles or to the for peptides with HLA-DR- or HLA-DQ-blocking Abs, they were TCR might account for such differences. Nevertheless, overall sta- HLA-DR restricted and all appeared to be recognized in the con- tistical modeling suggested a high degree of correlation between text of several HLA-DR alleles. This was clearer in some of the sequences predicted and observed CD4 lymphocyte responses. patients who were homozygous for the HLA-DR state. However, Our results provide a broader picture of the HSV structural pep- in one patient recognizing peptide 30-5, anti-HLA-DR and anti- tides recognized by human and CD4 lymphocytes than hitherto HLA-DQ Abs were equally inhibitory, suggesting there were also published. In humans using HSV2-specific clones from genital le- HLA-DQ-specific peptides within gD2. sions, Koelle et al. identified HSV type-specific epitopes in the Peptide-MHC binding data showed that peptides 2 and 24, and tegument proteins VP16, VP22, and dUTPase (UL49 and UL50), to a lesser degree peptide 34, bind multiple DR types with high as well as UL21, by expression cloning, thus extending the range affinity. The peptide binding data also support the proposition that of CD4 epitopes beyond gD, gB, and gH (13, 16). in some cases there may be multiple epitopes within the 20-mers In mice, a recent thorough study determined four new immu- studied. This is especially apparent in the case of peptide 2, where nodominant regions in gD, recognized by three murine strains of binding of both N- and C-terminal-truncated peptides was ob- different MHC II types, using 12 regions predicted to be immu- served to DRB1*0101, DRB1*0701, and DRB1*1501. The pres- nodominant from the TEPITOPE algorithm as their starting point ence of overlapping epitopes obscures identification of the exact (36). Allowing for their numbering system of mature gD (which specificity of recognition for each HLA-DR allele. However, while excludes the 25-aa leader sequence of immature gD that was in- the exact minimal epitope has not been elucidated in each context, cluded in our studies) and that their studies focused on gD1 rather we think that the present data will be of value to those studying T than gD2, there was still relatively little overlap with the four cell recognition of gD2 and the correlates of HSV immunity. major 20-mer peptide epitopes found in our study, perhaps not The Journal of Immunology 6613 surprising in view of the species difference. Nevertheless, this mu- The predominant restriction of gD peptide recognition by CD4 rine and our human study showed the presence of multiple lymphocytes according to HLA-DR and broad recognition across epitopes scattered throughout the gD epitope and which can be different HLA-DR alleles can be added to the short list of other cross-recognized by humans or mice of different MHC types. Im- published studies of CD4 lymphocyte recognition of pathogens, portantly, these epitopes induced Th1 rather than the Th2 re- including Plasmodium falciparum, hepatitis C, and HIV (32, sponses of some other gD epitopes, and when the Th1 epitopes 38–40). were used to immunize mice, they protected against viral chal- The high density of CD4 lymphocyte epitopes and their cross- lenge. The major gD epitopes in our study were detected by cy- recognition by different HLA-DR alleles is consistent with our totoxicity and IFN-␥ ELISPOT assays, indicating they were also previously published studies, which showed that glycoprotein D Th1 epitopes and may be useful as immunogens in future. was recognized by all 18 patients studied using gD recombinant An earlier murine study also showed a marked diversity and vaccinia viruses (16). Presumably these results underlie the high lack of immunodominance of peptide epitopes within gD1 recog- rate of protection (73–74%) of HSV1- and HSV2-seronegative nized by mice after infection but a single immunodominant epitope women against genital herpes disease in two trials of a gD2 vac- after 240–260 aa after immunization with gD (18, 19), which was cine with a deacylated monophosphoryl lipid A adjuvant, which confirmed by another study that employed immunization with gD1 induces CD4 lymphocyte responses oriented toward a Th1 pattern molecules lacking the C-terminal 93 aa (37). Brynestad et al. (29) of cytokine production, including IFN-␥ (2). In this study all pa- also showed immunogenicity of peptide 1–23 amino acids of gly- tients tested developed T lymphocyte reactivity to the vaccine coprotein D1 that was later enhanced by palmitic acid conjugation (G. Dubin, M. Denis, A. L. Cunningham, S. Spruance, L. Stan-

(36). berry, D. I. Bernstein, A. Mindel, S. Sacks, S. Tyring, F. Y. Aoki, Downloaded from Only two reports have defined gD epitopes recognized by hu- and M. Slaoui, unpublished results). The vaccines produced by man T lymphocytes. Neither addressed HLA restriction of the hu- Chiron and GSK that differed in their adjuvants and in their effi- man CD4 lymphocyte responses, HLA-DR binding, or HSV1/2 cacy despite both inducing neutralizing Ab has been widely inter- cross-reactivity (using acceptable modern type-specific serology). preted to mean that the Th1 CD4 lymphocyte response induced by Damhof et al. used overlapping synthetic peptides from gD1 to the GSK vaccine is likely to be highly significant (2, 41). Further-

examine the response by T lymphocyte proliferation from 10 more, the gD2 vaccine afforded no additional protection against http://www.jimmunol.org/ healthy HSV seropositive patients and found marked differences in disease to those women who were already HSV1-seropositive. individual responses to peptides. They defined broad immunodom- Such data strongly suggested that T lymphocyte- (especially CD4 inant regions of 1–54, 110–214 and 290–314 of mature gD mol- lymphocyte and Th1 cytokine) mediated immunity was responsi- ecules (23). As in mice, this study emphasized the diversity and ble for the protection against disease in the HSV-seronegative lack of immunodominance in human recognition of gD peptides women and also that there are cross-reactive (and cross-protective) also shown in our studies. DeFreitas et al. examined the response epitopes in gD1 and gD2. Furthermore, in a separate analysis of of human peripheral blood T lymphocytes to the mature gD1 pep- the placebo recipients, HSV1ϩ patients appear to be better pro- tide amino acids 1–23, indicating immunogenicity particularly in tected against genital herpes disease than those who are HSV1- the N-terminal half of the peptide (24). However, there were also seronegative (2). There are other data to suggest that HSV1-sero- by guest on September 28, 2021 responses induced in HSV-seronegative patients who could have positive patients have milder initial genital herpes disease to been nonspecific or, less likely, due to primary immunization. support such a hypothesis (42). This study was designed to provide There are similarities and differences between our results and the first data to examine and possibly support these hypotheses. these human and murine studies. We examined the whole of the Herein, we have clearly demonstrated that there are indeed such gD2 (rather than gD1) molecule, including the leader sequence cross-reactive epitopes in gD2 and that they are actively recog- (resulting in a 25-aa shift in numbering systems between the stud- nized by patients naturally infected with HSV1, inducing a Th1 ies) and the transmembrane region of gD2. We reasoned that gD2 response. In some peptides, this was because of their sequence was clinically a more important molecule and secondly that the similarities. In others, there were two different epitopes within the leader and transmembrane sequences might still be important in one 20-mer, each recognized by either HSV1ϩ or HSV2ϩ subjects. generating CD4 lymphocyte responses, especially following up- This suggests that a gD2/deacylated monophosphoryl lipid A vac- take of apoptotic or necrotic debris by APCs (although usually cine might also be effective against genital HSV1 disease in inducing an MHC I-restricted response). In support of this hypoth- HSV1Ϫ/HSV2Ϫ women or even a future childhood vaccine esis (and the TEPITOPE algorithm), an in vitro binding assay in- against severe diseases caused by both HSV1 and HSV2. dicated the presence of potential broadly recognized MHC II However, recent studies of serial infection by certain immuno- epitopes in the region 10–30 of the immature gD molecule. Fur- logically related viral species in animal models show that reacti- thermore, the serial internal 12-mer peptide studies showed rec- vation of cross-reactive CD8 lymphocyte memory responses by ognition by HSV1-seropositive patients in regions of peptide 2 the second infection can alter the normally naive CD8 lymphocyte overlapping with those detected by DeFreitas et al. (24). Peptide responses to the second virus by heterologous immunity in various 34 (aa 331–350) in our study includes the transmembrane region, ways, such as enhancement, inhibition, or immunopathology. Al- which was not tested by Damhof et al. (23). This region showed though only cross-protection has been observed in natural serial the greatest difference in recognition between HSV1ϩ and HSV2ϩ infection with HSV1 and HSV2 (for disease), these unpredictable patients with clear separation of epitopes in the N- and C-terminal T cell responses indicate the need for appropriate clinical and im- regions. The detection of broad human responses (across HLA-DR munologic observations when such anti-HSV1/2 vaccine strategies alleles) to peptides 24 (aa 231–250) and 30 (aa 291–310) in our are used in trials. However, our results do show a marked contrast study, and not in theirs, probably reflects the enhanced sensitivity between the broadly recognized MHC II-restricted gD2 epitopes of ELISPOT assay over T lymphocyte proliferation for CD4 lym- and the variable but usually narrow spectrum of recognition of phocyte responses. These responses were confirmed with the serial MHC I-restricted peptide epitopes (43). 12-mer studies, which also showed peptide 34 to be the most di- These results lay the foundation for future studies in naturally vergent in recognition by HSV1- and HSV2-seropositive patients infected patients and those immunized against HSV and for future of the four key peptides epitopes. development of vaccines. For example, future studies defining 6614 HSV-2 gD EPITOPES RECOGNIZED BY HSV1/2ϩ SUBJECTS

HSV1/2 type-specific and/or type common epitopes for both CD4 10. Hill, A., P. Jugovic, I. York, G. Russ, J. Bennink, J. Yewdell, H. Ploegh, and and CD8 lymphocytes in key HSV immunogenic proteins other D. Johnson. 1995. Herpes simplex virus turns off the TAP to evade host immu- nity. Nature 375: 411–415. than gD (e.g., tegument proteins) should assist in designing more 11. Koelle, D. M., M. A. Tigges, R. L. Burke, F. W. Symington, S. R. Riddell, effective vaccines, especially in men and HSV1-seropositive H. Abbo, and L. Corey. 1993. Herpes simplex virus infection of human fibro- blasts and keratinocytes inhibits recognition by cloned CD8ϩ cytotoxic T lym- women. Lessons from murine studies, demonstrating potentially phocytes. J. Clin. Invest. 91: 961–968. deleterious Th2-inducing gD epitopes, should also be applied to 12. Mikloska, Z., M. Ruckholdt, I. Ghadiminejad, H. Dunckley, M. Denis, and such human studies. The definition of a number of (promiscuous) A. L. Cunningham. 2000. Monophosphoryl lipid A and QS21 increase CD8 T lymphocyte cytotoxicity to herpes simplex virus-2 infected cell proteins 4 and 27 HLA-DR epitopes will also enable the construction of MHC II through IFN-␥ and IL-12 production. J. Immunol. 164: 5167–5176. tetramers (although this is often difficult and variable results can be 13. Koelle, D. M., J. M. Frank, M. L. Johnson, and W. W. Kwok. 1998. Recognition obtained). of herpes simplex virus type 2 tegument proteins by CD4 T cells infiltrating human genital herpes lesions. J. Virol. 72: 7476–7483. These studies have shown the important epitopes recognized 14. Koelle, D. M., C. M. Posavad, G. R. Barnum, M. L. Johnson, J. M. Frank, and soon after recurrent diseases. However, as a follow-up to these L. Corey. 1998. Clearance of HSV-2 from recurrent genital lesions correlates studies, carefully planned, longitudinal, well-controlled clinical with infiltration of HSV-specific cytotoxic T lymphocytes. J. Clin. Invest. 101: 1500–1508. studies are required to test the importance of CD4 lymphocytes in 15. Mikloska, Z., A. M. Kesson, M. E. Penfold, and A. L. Cunningham. 1996. Herpes general and to each of these epitopes in particular in the control of simplex virus protein targets for CD4 and CD8 lymphocyte cytotoxicity in cul- initial symptomatic HSV1/2 disease. Such studies require prospec- tured epidermal keratinocytes treated with interferon-␥. J. Infect. Dis. 173: 7–17. 16. Mikloska, Z., and A. L. Cunningham. 1998. Herpes simplex virus type 1 glyco- tive lymphocyte collections and epitope-specific studies during proteins gB, gC and gD are major targets for CD4 T-lymphocyte cytotoxicity in primary disease or asymptomatic seroconversion, probably during HLA-DR expressing human epidermal keratinocytes. J. Gen. Virol. 79: 353–361. placebo-controlled vaccine studies in seronegative partners of pa- 17. Zarling, J. M., P. A. Moran, L. A. Lasky, and B. Moss. 1986. Herpes simplex Downloaded from virus (HSV)-specific human T-cell clones recognize HSV glycoprotein D ex- tients with genital herpes. Furthermore, responses to HSV1/2 non- pressed by a recombinant vaccinia virus. J. Virol. 59: 506–509. cross-reactive vs cross-reactive epitopes needed to be tested during 18. Yamashita, K., and E. Heber-Katz. 1989. Lack of immunodominance in the T cell first HSV2 infections in HSV1-seropositive subjects. response to herpes simplex virus glycoprotein D after administration of infectious virus. J. Exp. Med. 170: 997–1002. In general, these studies extend the relatively small number of 19. Heber-Katz, E., S. Valentine, B. Dietzschold, and C. Burns-Purzycki. 1988. pathogens for which MHC II restricted promiscuous epitope rec- Overlapping T cell antigenic sites on a synthetic peptide fragment from herpes simplex virus glycoprotein D, the degenerate MHC restriction elicited, and func- ognition by CD4 lymphocytes, and they provide a mechanism and http://www.jimmunol.org/ tional evidence for antigen-Ia interaction. J. Exp. Med. 167: 275–287. a rationale for the use of safe viral protein vaccine candidates, 20. Bourne, N., F. J. Bravo, M. Francotte, D. I. Bernstein, M. G. Myers, M. Slaoui, especially where CD4 lymphocyte responses are important. Pro- and L. R. Stanberry. 2003. Herpes simplex virus (HSV) type 2 glycoprotein D teins with dense numbers of epitopes each binding to multiple DR subunit vaccines and protection against genital HSV-1 or HSV-2 disease in guinea pigs. J. Infect. Dis. 187: 542–549. and DQ types should be useful. The results also provide further 21. Bastin, C., P. Hermand, M. Francotte, N. Garcon, M. Slaoui, and P. Pala. 1994. data on discrepancies between murine and human CD4 responses Synergistic association of adjuvants QS21 and MPL for induction of cytolytic T-lymphocytes and T-helper responses to recombinant protein antigens. In Pro- and data to compare with predictive algorithms, and they empha- gram and Abstracts of the 34th Interscience Conference on Antimicrobial Agents size the importance of developing peptide-DQ binding assays. and Chemotherapy of the American Society for Microbiology. American Society for Microbiology, Washington, DC, p. 26 (abstract G41).

22. Bian, H., J. F. Reidhaar-Olson, and J. Hammer. 2003. The use of bioinformatics by guest on September 28, 2021 Acknowledgments for identifying class II-restricted T-cell epitopes. Methods 29: 299–309. We thank the staff of Taylor Square Private Clinic and Parramatta Sexual 23. Damhof, R. A., J. W. Drijfhout, A. J. Scheffer, J. B. Wilterdink, G. W. Welling, Health Clinic for their great help with patient samples. We are also grateful and S. Welling-Wester. 1993. T cell responses to synthetic peptides of herpes simplex virus type 1 glycoprotein D in naturally infected individuals. Arch. Virol. to Claire Wolczak and Brenda Wilson for their assistance with typing, and 130: 187–193. Heather Donaghy for proofreading. 24. DeFreitas, E. C., B. Dietzschold, and H. Koprowski. 1985. Human T-lymphocyte response in vitro to synthetic peptides of herpes simplex virus glycoprotein D. Proc. Natl. Acad. Sci. USA 82: 3425–3429. Disclosures 25. Lafferty, W. E. 2002. The changing epidemiology of HSV-1 and HSV-2 and The authors have no financial conflicts of interest. implications for serological testing. Herpes 9: 51–55. 26. Ho, D. W., P. R. Field, W. L. Irving, D. R. Packham, and A. L. Cunningham. 1993. Detection of immunoglobulin M antibodies to glycoprotein G-2 by West- References ern blot (immunoblot) for diagnosis of initial herpes simplex virus type 2 genital 1. Reeves, W. C., L. Corey, H. G. Adams, L. A. Vontver, and K. K. Holmes. 1981. infections. J. Clin. Microbiol. 31: 3157–3164. Risk of recurrence after first episodes of genital herpes: relation to HSV type and 27. Cunningham, A. L., R. Taylor, J. Taylor, C. Marks, J. Shaw, and A. Mindel. antibody response. N. Engl. J. Med. 305: 315–319. 2006. Prevalence of infection with herpes simplex virus types 1 and 2 in Aus- 2. Stanberry, L. R., S. L. Spruance, A. L. Cunningham, D. I. Bernstein, A. Mindel, tralia: a nationwide population based survey. Sex. Transm. Infect. 82: 164–168. S. Sacks, S. Tyring, F. Y. Aoki, M. Slaoui, M. Denis, et al. 2002. Glycoprotein- 28. Kimura, A., and T. Sasazuki. 1992. Eleventh International Hisocompatability D-adjuvant vaccine to prevent genital herpes. N. Engl. J. Med. 347: 1652–1661. Workshop Reference Protocol for the HLA-DNA typing technique. I: HLA 1991. 3. Cunningham, A. L., R. J. Diefenbach, M. Miranda-Saksena, L. Bosnjak, M. Kim, K. Tsuji, M. Aizawa, and T. Sasazuki, eds. Oxford Univeristy Press, Oxford, pp. C. Jones, and M. W. Douglas. 2006. The cycle of human herpes simplex virus 397–419. infection: virus transport and immune control. J. Infect. Dis. 194 (Suppl. 1): 29. Brynestad, K., B. Babbit, L. Huang, and B. T. Rouse. 1990. Influence of peptide S11–S18. acylation, liposome incorporation, and synthetic immunomodulators on the im- 4. Manickan, E., and B. T. Rouse. 1995. Roles of different T-cell subsets in control munogenicity of a 1–23 peptide of glycoprotein D of herpes simplex virus: im- of herpes simplex virus infection determined by using T-cell-deficient mouse- plications for subunit vaccines. J. Virol. 64: 680–685. models. J. Virol. 69: 8178–8179. 30. Wilkinson, J., A. Cope, J. Gill, D. Bourboulia, P. Hayes, N. Imami, T. Kubo, 5. Divito, S., T. L. Cherpes, and R. L. Hendricks. 2006. A triple entente: virus, A. Marcelin, V. Calvez, R. Weiss, et al. 2002. Identification of Kaposi’s sarcoma- neurons, and CD8ϩ T cells maintain HSV-1 latency. Immunol. Res. 36: 119–126. associated herpesvirus (KSHV)-specific cytotoxic T-lymphocyte epitopes and 6. Cunningham, A. L., R. R. Turner, A. C. Miller, M. F. Para, and T. C. Merigan. evaluation of reconstitution of KSHV-specific responses in human immunodefi- 1985. Evolution of recurrent herpes simplex lesions: an immunohistologic study. ciency virus type 1-infected patients receiving highly active antiretroviral ther- J. Clin. Invest. 75: 226–233. apy. J. Virol. 76: 2634–2640. 7. Koelle, D. M., H. Abbo, A. Peck, K. Ziegweid, and L. Corey. 1994. Direct 31. Southwood, S., J. Sidney, A. Kondo, M. F. del Guercio, E. Appella, S. Hoffman, recovery of herpes simplex virus (HSV)-specific T lymphocyte clones from re- R. T. Kubo, R. W. Chesnut, H. M. Grey, and A. Sette. 1998. Several common current genital HSV-2 lesions. J. Infect. Dis. 169: 956–961. HLA-DR types share largely overlapping peptide binding repertoires. J. Immu- 8. Mikloska, Z., V. A. Danis, S. Adams, A. R. Lloyd, D. L. Adrian, and nol. 160: 3363–3373. A. L. Cunningham. 1998. In vivo production of cytokines and ␤ (C-C) chemo- 32. Wilson, C. C., B. Palmer, S. Southwood, J. Sidney, Y. Higashimoto, E. Appella, kines in human recurrent herpes simplex lesions: do herpes simplex virus-in- R. Chesnut, A. Sette, and B. D. Livingston. 2001. Identification and antigenicity fected keratinocytes contribute to their production? J. Infect. Dis. 177: 827–838. of broadly cross-reactive and conserved human immunodeficiency virus type 9. Mikloska, Z., and A. L. Cunningham. 2001. ␣ and ␥ interferons inhibit herpes 1-derived helper T-lymphocyte epitopes. J. Virol. 75: 4195–4207. simplex virus type 1 infection and spread in epidermal cells after axonal trans- 33. Bian, H., and J. Hammer. 2004. Discovery of promiscuous HLA-II-restricted T mission. J. Virol. 75: 11821–11826. cell epitopes with TEPITOPE. Methods 34: 468–475. The Journal of Immunology 6615

34. Gonzalez, J. C., W. W. Kwok, A. Wald, C. L. McClurkan, J. Huang, and 39. Schulze zur Wiesch, J., G. M. Lauer, C. L. Day, A. Y. Kim, K. Ouchi, D. M. Koelle. 2005. Expression of cutaneous lymphocyte-associated antigen and J. E. Duncan, A. G. Wurcel, J. Timm, A. M. Jones, B. Mothe, et al. 2005. Broad E-selectin ligand by circulating human memory CD4ϩ T lymphocytes specific for repertoire of the CD4ϩ Th cell response in spontaneously controlled hepatitis C herpes simplex virus type 2. J. Infect. Dis. 191: 243–254. virus infection includes dominant and highly promiscuous epitopes. J. Immunol. 35. Asanuma, H., M. Sharp, H. T. Maecker, V. C. Maino, and A. M. Arvin. 2000. 175: 3603–3613. Frequencies of memory T cells specific for varicella-zoster virus, herpes simplex 40. Elkington, R., N. H. Shoukry, S. Walker, T. Crough, C. Fazou, A. Kaur, virus, and cytomegalovirus by intracellular detection of cytokine expression. C. M. Walker, and R. Khanna. 2004. Cross-reactive recognition of human and J. Infect. Dis. 181: 859–866. primate cytomegalovirus sequences by human CD4 cytotoxic T lymphocytes specific for glycoprotein B and H. Eur. J. Immunol. 34: 3216–3226. 36. BenMohamed, L., G. Bertrand, C. D. McNamara, H. Gras-Masse, J. Hammer, 41. Corey, L., A. G. Langenberg, R. Ashley, R. E. Sekulovich, A. E. Izu, S. L. Wechsler, and A. B. Nesburn. 2003. Identification of novel immunodom- ϩ J. M. Douglas, Jr., H. H. Handsfield, T. Warren, L. Marr, S. Tyring, et al. 1999. inant CD4 Th1-type T-cell peptide epitopes from herpes simplex virus glyco- Recombinant glycoprotein vaccine for the prevention of genital HSV-2 infection: protein D that confer protective immunity. J. Virol. 77: 9463–9473. two randomized controlled trials: Chiron HSV Vaccine Study Group. J. Am. Med. 37. Grammer, S. F., A. Sette, S. Colon, L. Walker, and R. Chesnut. 1990. Identifi- Assoc. 282: 331–340. cation of an HSV-1/HSV-2 cross-reactive T cell determinant. J. Immunol. 145: 42. Langenberg, A. G., L. Corey, R. L. Ashley, W. P. Leong, and S. E. Straus. 1999. 2249–2253. A prospective study of new infections with herpes simplex virus type 1 and type 38. Doolan, D. L., S. Southwood, R. Chesnut, E. Appella, E. Gomez, A. Richards, 2: Chiron HSV Vaccine Study Group. N. Engl. J. Med. 341: 1432–1438. Y. I. Higashimoto, A. Maewal, J. Sidney, R. A. Gramzinski, et al. 2000. HLA- 43. Hosken, N., P. McGowan, A. Meier, D. M. Koelle, P. Sleath, F. Wagener, DR-promiscuous T cell epitopes from Plasmodium falciparum pre-erythrocytic- M. Elliott, K. Grabstein, C. Posavad, and L. Corey. 2006. Diversity of the CD8ϩ stage antigens restricted by multiple HLA class II alleles. J. Immunol. 165: T-cell response to herpes simplex virus type 2 proteins among persons with 1123–1137. genital herpes. J. Virol. 80: 5509–5515. Downloaded from http://www.jimmunol.org/ by guest on September 28, 2021