HSV-specific T cells persistently localize to the female reproductive tract in humans.

Female reproductive tract T cells recognize specific HSV peptides.

Female reproductive tract T cells readily expand in vitro with standard conditions.

The majority of human T cells reside in mucosal barrier tissues outside the vasculature, and have a motility and adhesion molecule profile limiting re-entry into the circulation (1). These cells, termed tissue-resident memory (T_RM) lymphocytes, are heterogeneous and include conventional T cells with hypervariable CDR3 sequences. Some innate lymphoid cells with conserved TCR variable gene usage and CDR3 amino acid (AA) domains such as mucosal-associated invariant T (MAIT) and invariant natural killer T (iNKT) cells also have memory characteristics and are included in the broad category of tissue-resident lymphocytes (2). T_RM can recognize pathogens with high sensitivity and specificity. Emplacement of cells elicited by vaccination in barrier tissues is an aspirational goal for vaccines (3).

Herpes simplex virus (HSV) infection of mice provides a valuable T_RM model. After recovery from primary cutaneous or female reproductive tract (FRT) infection, the natural attenuation of HSV in mice prevents ganglionic reactivation and recurrent epithelial replication. Skin and FRT T_RM persist without re-exposure to antigen (4). Skin CD8+ T_RM specific for an immunodominant MHC-I-restricted peptide localize to the epidermis in the region of primary infection, and mediate local protection from exogenous re-challenge. HSV-specific CD4 T cells occur mostly in the dermis after recovery from infection and are not as restricted to the site of primary infection (5, 6). In the FRT, after priming by non-lethal primary infection, protection against challenge is mediated by local CD4 T cells (7). HSV-specific CD8 T cells can also protect mice if pulled from the circulation to establish FRT T_RM (3).

Human HSV T_RM studies are important because HSV pathogenesis in the natural host differs from that observed in experimental animals. Acute lethality upon primary infection is only observed in persons with profound global immunodeficiency, or variations in genes most closely associated with innate immunity (8). In naïve mice, in contrast, FRT inoculation with wild-type HSV-2 is almost always fatal (9). Infected persons have periodic epithelial recurrences, with HSV-2 DNA detectable in the genital tract a median of 17% of days, in many cases without lesions or symptoms (10). These recurrences are initiated by viral reactivation in sensory ganglia, which are also a relevant local battlefield involving antigen-specific CD4 and CD8 T cells in both mice and humans (11, 12). In mice, spontaneous or even immune deficiency-triggered recurrences are very unusual and seldom progress to lytic viral replication in the periphery. Overall, periodic re-exposure to HSV antigen in humans could condition skin and FRT T_RM in ways that differ from the single priming episode in murine models.

We previously found that HSV-specific CD4 and CD8 T cells can be detected in cervical specimens, even during receipt of suppressive antiviral medication (13, 14). Genital skin biopsies also revealed preferential deposition of CD8α+ cells at the DEJ of healed lesion skin compared to normal skin. Laser capture microdissection (LCM) and immunohistology showed that these cells had low chemokine receptor and mobility-associated mRNA expression and expressed cytolytic and antiviral cytokine transcripts (15). Some DEJ CD8α+ cells stained in situ with fluorescent multimer reagents comprised of HSV peptides and subject-matched HLA class I, identifying them as antigen-specific CD8+ T_RM. TRCB CDR3 sequences from LCM-captured DEJ CD8α+ cells from serial biopsies over several years revealed persistent oligoclonal expansions consistent with the localization of clonally expanded T_RM (15). In the present report, we build on these previous findings by more precisely measuring the frequency of HSV-2-reactive CD4 and CD8 T cells directly ex vivo in FRT specimens obtained at times when HSV is absent, and by determining their antigenic breadth and fine specificity.

Attempts to boost local T cells by therapeutic vaccination of humans have had measurable clinical success as quantified by reduction in genital lesions and HSV shedding (16, 17). However, it is not known if the bivalent approach used, containing two of the more than 75 proteins in the HSV-2 genome, contains the optimal viral antigens. Strategies to determine T cell specificity from minute biological specimens typically rely on in vitro expansion prior to interrogation of the relevant proteome (18). The complex HSV-2 genome and HLA diversity, present challenges to both in situ or direct ex vivo approaches as well as interrogation of T cell lines after expansion. Some data indicate that T_RM can proliferate in vivo (19), but the capacity of these cells to expand in vitro for detailed studies is poorly understood, as T_RM have specific metabolic and cytokine requirements (20). For example, lipids are important for human skin T_RM energy metabolism (21). To explore the hypothesis that some HSV-specific T_RM can expand in vitro, we started with a dendritic cell (DC)-based activation-induced marker (AIM) step to both measure and enrich virus-reactive skin T cells, based on a method used for blood (18). T_RM were then interrogated for fine specificity with HSV genome-wide tools tailored to CD8 and CD4 T cell phenotypes.

Six HSV-2 seropositive women with a median age or 43 (range, 38 to 62) with a history of recurrent genital herpes were enrolled in a prospective biopsy protocol ( Table 1 ). Participants had clinical histories of genital HSV for a median of 19 years prior to specimen collection (range, 1 to 33 years). The median annual number of self-reported recurrences of genital herpes was 2 (range, 1-6), typical for immune competent persons (10). Each participant collected daily genital swab specimens for shedding studies as described (10) and each swab was used for HSV-specific DNA PCR. The median viral DNA detection rate of 22.3% (range, 4.4% to 55.7%) was similar to the median rate of 17% of days determined from large natural history studies (44). Genital swabs at six of the eight times of specimen collection were negative for HSV DNA by PCR, with data not available at two times of biopsy specimen collection.

This report focuses on the human FRT as an interface between HSV and the acquired immune response. For the cervix, we used a combination of biopsies and cytobrush samples to study both deeper and more superficial local T cell responses to this pathogen. In the context of intermittent HSV-2 antigen exposure in the natural host, we find that HSV-2-specific CD8 and CD4 T cells are abundant, retain the ability to recognize antigen by up-regulating activation markers ex vivo, and briskly proliferate in vitro. A broad spectrum of viral antigens and epitopes are recognized, often with high functional avidity.

The work in this report indicates that T_RM can recognize a relatively wide array of HSV-2 antigens. CD8 T cell antigens recognized by the skin and FRT-derived T cells in this report ranged from transcription factors (ICP4 encoded by RS1) expressed immediately after lytic replication starts, to abundant structural proteins such as the major capsid protein VP1 encoded by UL19, expressed late in the viral life cycle. The spectrum of CD4 T cell antigens was also quite wide, especially when the cervical cytobrush-derived T cells are included. We have previously detected HSV-2-specific CD8 T cells in cytobrush specimens (53) and proteome-wide tools could be applied to extend specificity determination to these samples. Within-subject diversity is hard to estimate given the limited anatomic samples we can access, but it is remarkable that up to six discrete CD8 T cell epitopes could be detected within a cervical biopsy as for subject 15018, and up to 14 HSV-2 antigens were recognized by CD4 T cells from a cervical cytobrush from subject 13931. Previously, we resorted to T cell cloning from biopsies or cytobrush samples and typically used traditional genomic HSV-2 DNA molecular libraries to discovery the fine specificity of tissue-derived HSV-specific T cells (33, 35, 54). Throughput is greatly enhanced using polyclonal TCL and virtual libraries of HSV ORFs, at the possible expense of missing rare responses. Previous measures of blood or tissue T cell epitope diversity for HSV-1 and HSV-2 indicated recognition of 20 or fewer epitopes/specimen (18, 37, 55, 56), fairly close to our current data.

Our co-transfection workflow determines HLA restriction in parallel to antigenic specificity for CD8 T cells. For CD4 T cells, established pathways for determining restricting HLA loci and alleles (57) are available. For HSV-2 UL49 AA 49-57, a known HLA B*07:02 restricted CD8 T cell epitopes we also noted recognition by CD4 T cells for subject 9149 ( Figure 4 ). Submission of the HSV-2 sequence to algorithms (43) for the relevant subject’s HLA class I and II alleles does not reveal any internal peptide sequence/allele combinations with high predicted binding affinity.

Related to specificity and phenotype, virus-specific CD4 and CD8 T cells may function optimally in close proximity to one another. For example, we observed that culture-expanded CD8 T_RM expressed minimal IL-2 while CD4 T cells expressed this cytokine in abundance ( Figure 2B ). Our work disclosed coordinated local CD4 and CD8 virus-specific T cell responses in many specimens. Other properties, such as the affinity of T_RM TCRs for peptide-HLA, could also be vital, with the reasonable hypothesis that high T_RM affinity would lead to earlier detection of pathogen and thus better control. Affinity requires biochemical assays to measure, but can be approximated by functional avidity, as in this report. We observed quite high functional avidity for polyclonal TCL specific for peptides in the tegument proteins encoded by UL46 and UL49 ( Figure 5 ), with strong recognition of 1 pg/ml viral peptide. It is likely that the epitope-specific response is heterogeneous at the TCR clonotype level, as we observed in studies of T cells specific for a tumor virus (58), and that functional affinities vary with TCR sequence. In a murine HSV model with localization of virus-specific CD8 T cells to latently infected ganglia, ganglionic T_RM had higher functional avidity than T cells specific for the same HSV epitope found in the lung, a non-infected site (59). In a murine model of cleared vaccinia infection, functional avidity varied between intestinal CD8 T_RM specific between intraepithelial and lamina propria layers (60). Future work will be required to determine if there is a hierarchy of functional avidity between blood and tissue compartments. While we did not study enough subjects to suggest nominating a specific HSV-2 ORF as vaccine candidate, it should be noted that chimeric multi—partial ORF and multi-epitope vaccines are increasingly entering clinical trials for both infection and cancer, including tuberculosis and melanoma (61, 62) and COVID-19 (clinicaltrials.gov NCT04776317). HSV epitope knowledge from the present and prior genital tissue-based studies ((63) combined with blood and neuronal ganglia studies (11) could allow a similar customized antigen to design a T cell-directed HSV vaccine immunogen.

Our studies suggest that the lymphocytic infiltrate seen after HSV infection is comprised at least in part by antigen-specific cells. After recovery from attenuated HSV infection, memory lymphoid clusters containing HSV-reactive CD4 T cells occur submucosally in murine FRT (7) and skin (64). In humans, a leukocytic infiltrate occurs in the endocervix in response to HSV (65), and CD4 and CD8 T cells infiltrate the dermis during and after infection (66). Less is known about the ectocervix. HSV-2 antigen remains in the epithelium regardless of anatomic location in skin vs. mucosal surfaces, and does not penetrate the basement membrane in the immunocompetent host (66). Recently, a plasma cell infiltrate in acute and healed HSV-2 genital skin lesions and local HSV-specific IgG have been reported (67). While the fine specificity of the local B lymphocytes is not yet known, it is possible that local antigen-specific T cells cooperate with B lymphocytes for antibody production

These studies are motivated by the unmet need for preventative and therapeutic HSV vaccines. Therapeutic and preventative vaccine clinical trials have used HSV-2-based candidates ranging in complexity from peptides to whole virus (68). Partial success has been demonstrated with preventative vaccine candidates that trigger CD4 T cells and antibody in humans (69–72), and data that passive antibody transfer can protect mice in an FcR-dependent manner against genital HSV-2 challenge (73, 74) has led to renewed interest in antibody-based vaccine prevention. The success of recombinant zoster vaccine (RZV) for recurrent infection by varicella zoster virus (VZV), a pathogen that shares tropism, sequence, and T cell epitopes with HSV (34), has re-invigorated therapeutic HSV vaccine efforts as well. In this regard, a bivalent gD2/ICP4 HSV-2 protein-based candidate showed efficacy as measured by genital HSV shedding and lesion rates (16, 17). The mechanisms of action of RZV and gD2/ICP4 remain unknown but are presumed to involve antibodies and/or CD4 T cells (75–78). To date, prevalent, durable, and quantitatively significant boosts of CD8 T cells have not been demonstrated in human HSV vaccine immunotherapy trials. We and others (79) hypothesize that HSV-2-specific CD4 and CD8 T cells at the site of recurrent infection are important for containing virus, and that ultimately, boosting the levels or enhancing the phenotype of peripheral virus-specific T cells could lead to clinical and virologic reduction of recurrence infection.

Human studies, while challenging, are of interest because HSV pathogenesis varies greatly between species. Non-human primates range from hyper-susceptible (80) to mostly resistant (81). Outbred rats (82), guinea pigs (83), and rabbits (84) can recover and have recurrences, but immunologic tools are limited. Inbred mice differ for HSV susceptibility per genetic background (85) and require hormonal treatment for primary FRT infection (86). Surviving animals have a coordinated humoral, CD4 and CD8 T cell response. HSV-specific T_RM are readily observed in the skin, FRT, and sensory ganglia in mice. Sophisticated tools allow precise localization, quantitation, expression profiling, and live cell monitoring during immune surveillance (87–90). However, HSV does not recur to the level of peripheral culture-positive viral reactivation outside of profound epigenetic manipulations (91). Therefore, explant, re-infection, and parabiosis models are used to study T_RM functionality. In contrast, HSV reactivates in essentially all infected persons (10). For immune competent persons with antibodies to HSV-2 and a clinical history of genital herpes, the population studied in this report, viral DNA is detected in anogenital swabs by PCR on a median of 17-20% of days (10).

The properties associated with effective control of recurrent HSV infection by T_RM are unknown. Mathematical models indicate that T_RM density threshold may be critical to determining if the outcome of neuronal delivery of HSV from axon termini is a brief period of low level viral shedding or exponential growth of virus in a clinically evident recurrent lesion episode (92). Cytokine released by T_RM such as IFN-γ are thought to mediate paracrine effects to induce a net anti-viral state in the vicinity (93). Localization near sites of HSV replication may also be critical. Our human in situ work has stained occasional CD8 T cells at the dermo-epidermal junction (DEJ) in skin biopsies from sites of previous HSV recurrences with HSV-specific peptide-HLA reagents (15).

This study has several limitations. We used moDC as APC because of their ability to cross-present to CD8 T cells and to also process and present antigen to CD4 T cells. Discrete DC subsets prime CD8 and CD4 T cells in murine HSV models (94, 95) and human moDC can present HSV antigen (96), but further research is required to determine the importance of DC in the context of human recurrent HSV infection. We used a strain of HSV-2 deleted in UL41 to re-stimulate HSV-2-specific T cells ex vivo, such that T cell responses to the protein product of UL41 would be missed. UL41 encodes an RNAse with complex influences on HSV replication, pathogenesis and immune evasion (97–100); biases in the spectrum of T cells recalled by presentation of UL41-deleted HeLa cells are therefore possible. The UL41 deletant is based on HSV-2 strain 333. This has a small number of predicted amino acid changes from HSV-2 strain 186 used for whole virus readouts and as the source for HSV-2 proteins. Each subject was infected with yet a 3^rd strain of HSV-2 that could also vary in amino acid sequence, possibly contributing to strain-specific responses (101). Overall, however, coding differences between HSV-2 strains are relatively minor (102, 103).

In addition, we studied a limited number of subjects and supplemented biopsies with cervical cytobrush specimens, a site related to, but immunologic distinct from the ectocervix. Only women were studied in this research restricted to the FRT; similar studies of male-specific tissues could be envisaged but are limited by the male genital biopsies. It could be argued that the T cells studied in this report are not challenged to persist in the periphery in the long-term absence of antigen. The tools created in the current study should enable future work to fully define the phenotype of FRT and skin HSV-specific T cells in humans, and compare and contrast them to T cells left behind by other pathogens that are permanently cleared, such as the related virus VZV. With regards to the requirement for antigen for maintenance of local virus-specific cells, we have previously shown that HSV-2-specific CD4 cells are enriched in cervical cytobrush specimens (13) and genital skin biopsies (66) during treatment with the antiviral drug acyclovir, a treatment (104) that reduces antigen.

We did not phenotype biopsy- or cytobrush-derived cells for T_RM markers such as CD69 or CD103 ex vivo, and refrained from staining for these markers after T cell expansion due to uncertainty about the effect of in vitro culture on their expression. Several lines of evidence suggest the cells studied in this report are T_RM. Specimens were obtained at times of negative HSV PCR, a surrogate for the absence of HSV antigen. In our previous studies comparing cervical biopsy-derived CD4 T cells with and without expression of the T_RM marker CD103, a higher proportion of CD103(+) than of CD103(-) cells showed IFN-γ responses to whole HSV-2 antigen. Both the CD103(+) and CD103(-) fractions had higher IFN-γ expression than did blood CD4 T cells (14). Expression of T_RM markers such as CD103 and CD69 is heterogeneous (105, 106). Recently, we reported that cervical biopsy-derived T cells obtained at times of negative HSV PCR express high levels of mRNAs encoding CD69 and the integrin subunits of CD103. Moreover, these cells have considerable TRB CDR3 overlap with blood HSV-2-specific CD8 T cells, implying that they are HSV-2-specific (107). Future studies could combine single cell phenotyping at the protein or mRNA levels with TCR sequencing and determination of virus specificity by expressing TCRs in reporter cells, a process we have piloted for blood-origin HSV-specific T cells (34). Theoretically, FRT-derived cells could be fractionated per T_RM marker expression after isolation from biopsies or cytobrush specimens and prior to antigen exposure, but this would likely be technically challenging given the recovery of low cell numbers from these specimens.

It is possible that the HeLa cells used to load HSV-2 antigen into moDC could re-stimulate HPV-reactive T cells. HeLa express HPV18 E6 and E7 (108). Figure 1 is an example in which stimulation of PBMC with moDC-mock-HeLa led to non-zero, CD137-expressing CD4 and CD8 T cells, consistent with our previous publication (18). HPV-specific T cells are rare: in vitro elicitation of HPV E6 or E7-reactive T cells from PBMC uses repetitive cycles of re-stimulation with high concentrations of peptides (109), and in vivo vaccinations use many cycles of peptide-loaded DC (110). It would be possible to query our HSV-reactive T cell lines with HPV expression constructs (111) to detect HPV-specific reactivity in our system.

We did not directly compare the abundance of HSV-specific T cells in PBMC and cervical specimens from the same subjects. In a previous study comparing direct ex vivo PBMC CD137-based AIM with IFN-γ ICS, the abundance of CD8 T cells reactive with HSV-1-loaded autologous moDC was under 1% (18). In a separate study of 67 HSV-2-infected persons, including serial sampling, with ex vivo ICS to detect IFN-γ, IL-2, or TNF-α, the median net abundance of CD4 T cells expressing any one or more cytokine in response to whole killed HSV-2 was 0.56% (112). The observed high abundances of HSV-2-reactive T cells in some of our genital biopsy specimens is evidence for local quantitative enrichment, which could be reinforced by study of simultaneously collected blood and local specimens.

In conclusion, the human FRT is richly and persistently infiltrated with HSV-2-specific CD8 and CD4 T cells in immunocompetent persons with a clinical history of recurrent genital herpes. These virus-specific effectors are readily detectable at both the ectocervix and in genital skin. Responses are typically poly-specific and have high functional avidity. Ultimately, measurement of the functional importance of epithelial T_RM in the control of recurrent HSV infection awaits clinical trials of vaccines capable of emplacing and increasing T_RM. The technologies developed for this study should enable safe and efficient T_RM assays based on clinically acceptable biopsies or cervical cytobrush specimens to be used in clinical trials of next-generation vaccines (113) being evaluated for elicitation of anatomically relevant T cell responses.

Presented in part at the 42nd International Herpesvirus Workshop, Ghent, Belgium, July 2017, abstract 1.05.

The original contributions presented in the study are included in the article/ Supplementary Material . Further inquiries can be directed to the corresponding author.

The studies involving human participants were reviewed and approved by University of Washington Institutional Review Board. The patients/participants provided their written informed consent to participate in this study.

DK, CP, and LC conceived the study. DK wrote the paper and created data visualizations. CJ, DV, and AW collected the specimens. LD, LCJ, LJ, JZ, CP, MH and KL optimized tissue preparations and performed experiments. SS managed data. All authors contributed to the article and approved the submitted version.

NIH grants AI030731 (DMK, CJ, AW), R01 AI042528 (JZ and LC), R01 AI134878 (JZ and LC), P30 CA015704 (core facilities), R01 AI091701 (CP), R21 AI083418 (CP), NIH contract 75N93019C00063 (DK).

DK, CP, and LC are co-inventors of patents concerning HSV vaccines owned and managed by their institutions. DMK receives research funding from Sanofi Pasteur concerning testing samples from a clinical trial of an HSV vaccine.

The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.