Tissue-resident memory CD8 T cells (CD8 TRM) are critical for maintaining barrier immunity. CD8 TRM have been mainly studied in the skin, lung and gut, with recent studies suggesting that the signals that control tissue residence and phenotype are highly tissue dependent. We examined the T cell compartment in healthy human cervicovaginal tissue (CVT) and found that most CD8 T cells were granzyme B+ and TCF-1. To address if this phenotype is driven by CVT tissue residence, we used a mouse model to control for environmental factors. Using localized and systemic infection models, we found that CD8 TRM in the mouse CVT gradually acquired a granzyme B+, TCF-1 phenotype as seen in human CVT. In contrast to CD8 TRM in the gut, these CD8 TRM were not stably maintained regardless of the initial infection route, which led to reductions in local immunity. Our data show that residence in the CVT is sufficient to progressively shape the size and function of its CD8 TRM compartment.

Postinfection or immunization, naive CD8 T cells differentiate into several major populations of memory T cells with distinct functions and trafficking patterns. Tissue-resident memory CD8 T cells (CD8 TRM) are a major subset defined by the fact that they do not recirculate via blood and lymph, instead remaining in the tissue sites where they were initially seeded by effector T cells during the primary response (15). Because of their privileged location, CD8 TRM are uniquely poised to respond to subsequent encounters with their cognate Ag by directly killing infected cells (6, 7), proliferating to boost the local CD8 T cell pool (8, 9), and activating and recruiting additional immune cells from both resident and circulating populations (10, 11). Many factors influence the formation of TRM, including the route of priming, exposure to Ag and inflammation, and the type of tissue in which residency is established. For example, the presence of local Ag boosts the size and function of TRM populations within the matched tissue site (2, 10, 12, 13). Similarly, local inflammation enhances the number and function of TRM in an Ag-independent manner (14, 15), and cytokine cues such as TGF-β, IL-15, and CCL5 appear to directly regulate TRM formation and maintenance (1619).

Despite the importance of local exposure to Ag and inflammation in TRM development, strong evidence also exists that TRM can broadly seed distal tissues after both systemic and local immunization or infection. Systemic immunization and infection models are routinely used in mice to generate memory T cell populations across many body sites (2024). Likewise, localized infections such as HSV and vaccinia virus seed CD8 TRM at distal tissue sites, albeit to a much lower extent than at the primary site of infection (12, 13, 25, 26). Together, these studies suggest that TRM can develop within diverse tissues across a range of levels of inflammation and Ag availability, although it is currently unclear how the combination of these variables may ultimately tune the characteristics of the resulting TRM population.

The tissue of residence itself also impacts the phenotype of CD8 TRM populations. Skin-resident CD8 T cells that express the putative residency markers CD69 and integrin αE (CD103) respond to local HSV infection and rely on TGF-β and IL-15 signals for their development (2, 2729). CD8 TRM in other organs may differ in their phenotype or developmental requirements. For example, CD69 and CD103 are not necessarily expressed by TRM in the uterus or pancreas (23) but may be required by TRM in other organs, such as the salivary gland, lung, or kidney (3032). Likewise, some organs such as the small intestine maintain TRM that constitutively express granzyme B in the absence of Ag or rechallenge (6, 33), a phenotype which has not been observed as robustly in other tissue sites, including the lung and uterus (34, 35).

Pathogens that infect mucosal barrier surfaces represent a significant global health burden. The cervicovaginal tissue (CVT) has broad importance in understanding the pathogenesis and vaccinology of sexually transmitted infections (STIs). Compared with memory CD8 T cells in the upper female reproductive tract and other mucosal and lymphoid tissues, the phenotypic and functional characteristics of memory CD8 T cells in the CVT remain relatively understudied. We recently reported that the CD8 T cell compartment in human CVT includes a subset of CD8 T cells that robustly express granzyme B (36). It is unclear to what extent the distinctive features we observed in human CVT may be driven by recurring local infections, exposure to inflammatory cues, or a result of signals intrinsically associated with the cervicovaginal microenvironment.

In this article, we report that CD8 T cells isolated from healthy human CVT lacked expression of the memory-associated transcription factor TCF-1. To determine if this characteristic was driven by residence in the tissue itself or a consequence of local exposure to antigenic insults or inflammation in human CVT, we used multiple immunization strategies to assess memory CD8 T cell differentiation and maintenance in the mouse CVT. We found that memory CD8 T cells in the mouse CVT at later memory time points closely resembled the CD8 T cells found in human CVT regardless of the initial immunization route and early phenotypic differences. Memory CD8 T cell numbers were stably maintained in the periphery and gut but gradually declined in the CVT over 5 mo postimmunization, which was associated with an initial delay in viral control upon HSV vaginal challenge. We conclude that residence in the CVT is sufficient to alter the canonical differentiation and maintenance program of memory CD8 T cells.

C57BL/6J and B6.PL-Thy1a/CyJ mice were purchased from The Jackson Laboratory and maintained in specific pathogen–free conditions at the Fred Hutchinson Cancer Research Center (FHCRC). All mice used in experiments were female and 8–12 wk of age. Experiments were approved by the FHCRC Institutional Animal Care and Use Committee.

Women recruited for this study (n = 13) consented to vaginal biopsy and blood sampling. Eligibility criteria included aged >18 to <45 y, normal Papanicolaou smear within the past 3–5 y, not menopausal, hepatitis C negative, and no report of active genital tract irritation or infection. Informed written consent was obtained from all participants. The study and procedures were approved by the FHCRC Institutional Review Board.

Mice were injected s.c. with 2 mg of medroxyprogesterone acetate (MDPA) injectable suspension (Greenstone) 7 d before vaginal infections. To control for any effects of estrus cycle phase on the CVT immune compartment, mice were also treated with MDPA 7 d prior to any memory timepoints at which the CVT was processed to generate single-cell suspensions.

Seven days after MDPA treatment, mice were infected intravaginally with 2 × 104 PFU wild-type HSV-2 (strain 186 syn+) or 1.88 × 105 PFU of thymidine kinase-negative HSV-2 186 kpn to the vaginal canal in a 10-μl volume (37). Postinfection with wild-type HSV-2, mice were monitored daily, and clinical disease progression was scored as follows: 0, no sign; 1, slight genital erythema; 2, moderate genital erythema and edema; 3, significant genital inflammation with visible lesion; 4, hind leg paralysis or other severe condition requiring euthanasia; and 5, moribund or dead. Mice were euthanized if they were moribund or showed signs of severe disease, including hind leg weakness, hind leg paralysis, or hunched posture. Viral titer was determined by plaque assay on Vero cells (American Type Culture Collection) or by PCR as previously described (38).

Listeria monocytogenes strains were generated to recombinantly express and secrete OVA with or without the HSV-2 glycoprotein B (gB)–derived peptide SSIEFARL according to previously described methods using the pPL2 vector to achieve integration into the bacterial genome (39, 40). Note, this LM-gB strain is distinct from a previously published LM-gB strain, which used the pAM401 plasmid for expression of the gB epitope (41). Naive B6 mice received 4000 CFU of LM-gB i.v. via tail vein injection.

To discriminate CD8 T cells in the vasculature from those located within tissues, we used an intravascular labeling technique as previously described (42). To assess cell proliferation, mice were injected with 2 mg of BrdU i.p. on day 1 and then administered BrdU for the following 7 d via drinking water at a concentration of 0.8 mg/ml.

Mice were injected i.p. with 500 ng of anti-Thy1.1 Ab (Clone 19E12, Bio X Cell) to deplete Thy1.1-expressing cells from the circulation and peripheral lymphoid tissues.

To prepare single-cell suspensions from mouse tissue, cervicovaginal tracts were digested with 2 mg/ml collagenase D (Roche) and 1.5 mg/ml DNase I (Roche) for 30 min at 37°C.

Human biopsy samples were trimmed to 2 mm2 pieces and digested with collagenase II (700 U/ml, Sigma-Aldrich) and DNase I (1 U/ml, Sigma-Aldrich) for two subsequent 30-min digestions at 37°C as previously described (43).

Cells were incubated in LIVE/DEAD fixable amine-reactive viability dye (Invitrogen), blocked for Fc binding (clone 2.4G2 for mice or TruStain [BioLegend] for human), and then stained with tetramers and Abs.

The following fluorochrome-conjugated Abs were used to stain mouse cells for flow cytometry: anti-CD8 (clone 53-6.7, PerCP-eFluor710), anti-CD4 (clone GK1.5, allophycocyanin-eFluor780), anti-CD45.1 (clone A20, Brilliant UV 395), anti-CD45.2 (clone 104, Brilliant UV 737), anti-Thy1.1 (clone HIS51, FITC or SuperBright645), anti-CD44 (clone IM7, allophycocyanin or AlexaFluor700), anti-CD69 (clone H1.2F3, PE-Dazzle594 or PE-Cy7), anti-CD103 (clone M290, Brilliant Violet 786), anti–granzyme B (clone GB11, Pacific Blue), anti–TCF-1 (clone C63D9, PE or PE-Cy7), and anti-BrdU (clone Bu20A, allophycocyanin). Anti-CD8, anti-CD4, anti-Thy1.1, and anti-BrdU were purchased from eBioscience. Anti-CD45.1, anti-CD45.2, anti-CD44, and anti-CD103 were purchased from BD. Anti-CD69 and anti–granzyme B Abs were purchased from BioLegend. Anti–TCF-1 Ab was purchased from Cell Signaling Technologies.

The following fluorochrome-conjugated Abs were used to stain human cells for flow cytometry: anti-CD3 (clone UCHT1, Brilliant UV 661), anti-CD45 (clone HI30, Brilliant UV 805), anti-CD8 (clone RPA-T8, Brilliant UV 496), anti-CD4 (clone RPA-T4, Brilliant Violet 605), anti-CD45RA (clone HI100, allophycocyanin-H7), anti-CCR7 (clone G043H7, AlexaFluor488), anti–granzyme B (clone GB11, AlexaFluor700), anti-CD103 (clone Ber-ACT8, Brilliant Violet 750), and anti–TCF-1 (clone C63D9, PE). Anti-CD3, anti-CD45, anti-CD4, anti-CD8, anti-CD45RA, anti-CD103, and anti–granzyme B were purchased from BD. Anti-CCR7 Ab was purchased from BioLegend.

Samples were acquired on a FACSymphony instrument (BD) and gated as outlined (Supplemental Fig. 1A–C). To estimate cell counts, 2 × 104 AccuCheck Counting Beads (Thermo Fisher Scientific) were added to each sample immediately before acquisition.

FIGURE 1.

Human cervicovaginal-resident CD8 T cells lack expression of TCF-1. (A) Gating strategy to identify memory CD8 T cells in blood and CVT from healthy STI- and HSV-2–negative women. (B) Relative abundance of CD69 and CD103 subsets in blood and CVT. (C) Representative flow plot and abundance of granzyme B+ cells within the memory CD8 population in blood and CVT. (D) Representative flow plot and abundance of CD103 and granzyme B coexpression in blood and CVT. (E) Representative TCF-1 staining and TCF-1+ frequency from CD103+ subset of CD8 T cells from the CVT compared with memory CD8 T cells in matched blood. Flow plots from (A) and (C)–(E) are from one representative participant. Data from (B)–(E) represent pooled results from study participants (n = 13). Each dot in (C)–(E) represents an individual participant. Error bars represent mean ± SD. The p value in (E) was calculated via paired t test. Exact p values are given for all comparisons.

FIGURE 1.

Human cervicovaginal-resident CD8 T cells lack expression of TCF-1. (A) Gating strategy to identify memory CD8 T cells in blood and CVT from healthy STI- and HSV-2–negative women. (B) Relative abundance of CD69 and CD103 subsets in blood and CVT. (C) Representative flow plot and abundance of granzyme B+ cells within the memory CD8 population in blood and CVT. (D) Representative flow plot and abundance of CD103 and granzyme B coexpression in blood and CVT. (E) Representative TCF-1 staining and TCF-1+ frequency from CD103+ subset of CD8 T cells from the CVT compared with memory CD8 T cells in matched blood. Flow plots from (A) and (C)–(E) are from one representative participant. Data from (B)–(E) represent pooled results from study participants (n = 13). Each dot in (C)–(E) represents an individual participant. Error bars represent mean ± SD. The p value in (E) was calculated via paired t test. Exact p values are given for all comparisons.

Close modal

Sample means between two unpaired groups were compared by t test. Sample means between three or more unpaired groups were compared by ordinary one-way ANOVA with Tukey multiple comparisons test. Sample means between three paired groups were compared by repeated measures ANOVA with Greenhouse–Geisser correction. Sample means between three or more groups with multiple group factors were analyzed via two-way ANOVA with Sidak posttests. Linear relationships between variables were evaluated with least-squares regression. Inflammation score trajectories were compared by calculating the area under the curve and associated SEM for each group and then comparing these values via t test. Error bars represent standard deviations. Exact p values are given for all results in which p < 0.05. Analysis was performed using Prism 7 (GraphPad Software).

In our model, susceptible cells are infected at rate βVS by free HSV-2 virus. Productively HSV-2–infected cells have a clearance that is intrinsic or mediated by a humoral response with rate δII or a clearance mediated by HSV-specific CD8 T cells with killing rate kEI. Free virus is produced at a rate pI and cleared at rate cV. HSV-specific CD8 T cells are cleared at rate δEE and proliferate in the presence of infected cells at a maximum rate ωEI. We also assumed the proliferation of HSV-specific CD8 T cells is constrained when infected, and effector cells grow over saturation levels I50 and E50, respectively. Under these assumptions, the model has the following form (see schematics in (Fig. 6D):dSdt=βVSdIdt=βVSδIIkEIdVdt=pIcV(1)dEdt=ωEI(1+II50)(1+EE50)δEE.

The basic reproductive ratio, R0, which represents the number of secondary infections produced by an HSV-infected cell when introduced into a susceptible population, was calculated as

R0=βpT0c(δI+kE0)

When R0 < 1, viremia is controlled.

We performed 100 rounds of model fits to observations from all individual animals for each group with different initial parameter guesses in the optimization algorithm. For each round, we reparameterized the model in function of R0 and used a nonlinear least-squares approach using the differential evolution and the L-BFGS-B algorithms in R (R Development Core Team) to fit the model and to estimate parameters: V0, pT0, E0, R0, δ, c, k, ω, δE, I50, and E50. Because mice were inoculated with ∼2.3 × 106 HSV genomes, we set that number as the upper limit for V0. We also constrained the value estimate E0 to the maximum and minimum HSV-specific CD8 T cell observations at the moment of virus challenge. Finally, we ensured that R0 > 1 in all fitting rounds. In all simulations, t = 0 represented the time of the HSV-2 vaginal challenge. From the best fits, we computed the predicted time of viral clearance as the time when viral load crossed the detection limit for the PCR assay of 4615.4 genomes per vaginal wash.

We repeated the fits by exploring different possibilities for constraining HSV-specific CD8 T cell growth. Specifically, we explored a model in which CD8 growth is constrained only by a saturation level in HSV-infected cells, that is, dEdt=ωEI(1+II50)δEE, or only by a saturation for the number of effector cells, that is, dEdt=ωEI(1+EE50)δEE. To determine the best and most parsimonious model, we computed the Akaike Information Criteria (AIC), AIC=nlnSSE¯+2m, in which m is the number of parameters estimated, n the number of data points in each group, and SSE¯, is the average sum of squares error of all the 100 model-fit rounds in each group. We assumed a model had similar support from the data if the difference between its AIC and the best model (lowest) AIC was less than two. We selected and showed results from the model with the lowest AIC.

To study the phenotypic profile of CD8 T cells within human CVT from healthy women without any genital infections, we isolated lymphocytes from human vaginal biopsies via enzymatic digestion and observed that the majority of memory T cells in this tissue site expressed the residency markers CD69 and CD103 (Fig. 1A, 1B). A large proportion of these CD8 T cells expressed granzyme B ex vivo (Fig. 1C), and the majority of granzyme B+ cells coexpressed CD103 (Fig. 1D), suggesting that they were potentially resident within the tissue. We observed that the memory CD8 T cell compartment in human CVT lacked expression of TCF-1 (Fig. 1E), a transcription factor associated with self-renewal potential in CD8 T cells (44). These data indicated that human memory CD8 T cells within the CVT may possess functional and proliferative characteristics distinct from such cells in other locations.

We next wanted to address if this human CVT CD8 T cell phenotype was attributable to ongoing basal levels of antigenic insult or inflammation in the tissue or might represent more generalized characteristics of CD8 T cells in CVT. To stringently control for external environmental factors, we sought to generate a comparable population of memory CD8 T cells in mouse CVT. We adoptively transferred naive gBT-I CD8 T cells specific to the SSIEFARL epitope from HSV and primed these cells by vaginal infection with nonlethal thymidine-kinase–deficient HSV-2 (HSV-2 TK-) (Fig. 2A). One month following infection, HSV gB tetramer+ CD8 T cells in the CVT expressed CD69, CD103, and granzyme B in similar proportions to those observed in human CVT (Fig. 2B, 2C). In addition, we found that these CD8 T cells largely lacked expression of TCF-1, especially compared with gB tetramer+ cells in the vaginal draining lymph nodes (dLN) (Fig. 2C). Given that these mice had received a vaginal infection, we wondered if this granzyme B+ TCF-1 phenotype might be a result of ongoing inflammatory responses in the CVT. Indeed, analysis of H&E-stained sections of CVT from infected mice revealed that clusters of inflammatory cells remained in the CVT lamina propria as long as 22 d postinfection (Fig. 2D), suggesting that inflammation was not resolved by this time point. Thus, our data suggested that prolonged local tissue inflammation could be driving this distinct memory CD8 T cell phenotype within the CVT in mice.

FIGURE 2.

Mouse cervicovaginal-resident CD8 T cells induced by vaginal immunization mirror the phenotype observed in human samples and lack TCF-1. (A) Experiment schematic to induce CVT-resident CD8 T cells in mice via vaginal infection with 1.88 × 105 PFU of thymidine kinase-negative HSV-2 186 kpn. (B) HSV-specific cells were identified in CVT via staining with an H-2Kb SSIEFARL tetramer conjugated to allophycocyanin. Left, Representative flow plots gated on CD8 T cells. Right, Relative abundance of CD69 and CD103 subsets within gB tetramer+ CD8 T cells in CVT. (C) Representative flow plot and abundance of granzyme B and TCF-1 staining in gB tetramer+ cells in CVT and dLN. (D) Representative images of H&E-stained sections of CVT at 10× objective and average lamina propria inflammation scores. Arrow indicates coalescing clusters of inflammatory cells with follicular organization. Arrowhead indicates clusters of inflammatory cells within the mucosal epithelium. Flow plots and images from (B)–(E) are from one representative mouse. Data from (B) and (C) show representative results from two independent experiments. Each dot in (E) represents inflammation score averaged from one to three mice per time point. Error bars represent mean ± SD. The p values in (C) were calculated via paired t test. Exact p values are given for all comparisons.

FIGURE 2.

Mouse cervicovaginal-resident CD8 T cells induced by vaginal immunization mirror the phenotype observed in human samples and lack TCF-1. (A) Experiment schematic to induce CVT-resident CD8 T cells in mice via vaginal infection with 1.88 × 105 PFU of thymidine kinase-negative HSV-2 186 kpn. (B) HSV-specific cells were identified in CVT via staining with an H-2Kb SSIEFARL tetramer conjugated to allophycocyanin. Left, Representative flow plots gated on CD8 T cells. Right, Relative abundance of CD69 and CD103 subsets within gB tetramer+ CD8 T cells in CVT. (C) Representative flow plot and abundance of granzyme B and TCF-1 staining in gB tetramer+ cells in CVT and dLN. (D) Representative images of H&E-stained sections of CVT at 10× objective and average lamina propria inflammation scores. Arrow indicates coalescing clusters of inflammatory cells with follicular organization. Arrowhead indicates clusters of inflammatory cells within the mucosal epithelium. Flow plots and images from (B)–(E) are from one representative mouse. Data from (B) and (C) show representative results from two independent experiments. Each dot in (E) represents inflammation score averaged from one to three mice per time point. Error bars represent mean ± SD. The p values in (C) were calculated via paired t test. Exact p values are given for all comparisons.

Close modal

To determine if prolonged tissue inflammation drives this CVT memory T cell phenotype, we used a systemic immunization approach to prime gBT-I cells in the absence of significant vaginal inflammation. We generated a recombinant strain of Listeria monocytogenes that expresses the HSV-derived SSIEFARL peptide (LM-gB) and immunized mice i.v. (Fig. 3A). We confirmed that this immunization strategy did not result in detectable levels of vaginal inflammation compared with uninfected mice and caused significantly less inflammation than observed in mice vaginally immunized with HSV-2 TK- (Fig. 3B,). In addition, immunized mice mounted a robust systemic HSV-specific CD8 T cell response, with gB-specific T cells making up an average of 36.4% of CD8 T cells in the blood by 1 wk after immunization (Fig. 3C). At 1 mo after immunization, we could identify these cells in multiple body sites, including the CVT and vaginal dLN (Fig. 3D). Again, CD103+ gB-specific CD8 T cells in the CVT lacked expression of TCF-1, replicating the phenotype observed in both human vaginal tissue and CVT of vaginally immunized mice (Fig. 3E), despite the lack of tissue inflammation.

FIGURE 3.

The cervicovaginal CD8 T cell memory compartment of LM-gB–immunized mice is poorly maintained after immunization. (A) Schematic of experiment to assess phenotype and maintenance of HSV-specific T cell compartment of mice immunized with 4000 CFU of LM-gB i.v. (B) Representative image of H&E-stained sections of CVT at 10× objective and average lamina propria inflammation scores. (C and D) gB tetramer+ CD8 T cell abundance or count in blood, spleen, dLN, and CVT 1 mo after immunization. (E) Representative TCF-1 staining and quantification of TCF-1 expression within CD103+ and CD103 subsets of gB tetramer+ CD8 T cells in CVT and dLN. (F) Mice were treated with BrdU from day 28–35 after LM-gB immunization (2 mg i.p. on day 28, followed by 0.8 mg/ml in drinking water from day 28–35). Representative flow plot and graph showing abundance of BrdU incorporation by TCF-1 subset. (G) Count of gB tet+ cells in spleen, dLN, SI LP, and CVT 1 or 5 mo after LM-gB immunization. Images and flow plots from (B) and (F) are from one representative mouse. Flow plot in (E) is from 10 concatenated CVT samples. Data in (B)–(G) are representative of two to five independent experiments. Each dot in (B) represents inflammation score averaged from three to five mice per time point. Each dot in (C) represents the mean value of 20 mice from a representative experiment. Each dot in (D)–(G) represents an individual mouse. Error bars represent mean ± SD. The p value in (B) was calculated via t test comparing the areas under each curve. The p values in (E) were calculated with repeated measures ANOVA with Greenhouse–Geisser correction. The p values in (G) were calculated via t test using log10-transformed values. Exact p values are given for all values <0.05.

FIGURE 3.

The cervicovaginal CD8 T cell memory compartment of LM-gB–immunized mice is poorly maintained after immunization. (A) Schematic of experiment to assess phenotype and maintenance of HSV-specific T cell compartment of mice immunized with 4000 CFU of LM-gB i.v. (B) Representative image of H&E-stained sections of CVT at 10× objective and average lamina propria inflammation scores. (C and D) gB tetramer+ CD8 T cell abundance or count in blood, spleen, dLN, and CVT 1 mo after immunization. (E) Representative TCF-1 staining and quantification of TCF-1 expression within CD103+ and CD103 subsets of gB tetramer+ CD8 T cells in CVT and dLN. (F) Mice were treated with BrdU from day 28–35 after LM-gB immunization (2 mg i.p. on day 28, followed by 0.8 mg/ml in drinking water from day 28–35). Representative flow plot and graph showing abundance of BrdU incorporation by TCF-1 subset. (G) Count of gB tet+ cells in spleen, dLN, SI LP, and CVT 1 or 5 mo after LM-gB immunization. Images and flow plots from (B) and (F) are from one representative mouse. Flow plot in (E) is from 10 concatenated CVT samples. Data in (B)–(G) are representative of two to five independent experiments. Each dot in (B) represents inflammation score averaged from three to five mice per time point. Each dot in (C) represents the mean value of 20 mice from a representative experiment. Each dot in (D)–(G) represents an individual mouse. Error bars represent mean ± SD. The p value in (B) was calculated via t test comparing the areas under each curve. The p values in (E) were calculated with repeated measures ANOVA with Greenhouse–Geisser correction. The p values in (G) were calculated via t test using log10-transformed values. Exact p values are given for all values <0.05.

Close modal

Previous reports have shown that CD8 TRM populations in the brain (45) and lung (46, 47) lack expression of TCF-1 transcript and protein, respectively. We explored whether this phenotype was consistent among CD8 T cells isolated from distinct regions of the female reproductive tract and compared CD8 T cell numbers and phenotype in the CVT to uterus and ovaries (Supplemental Fig. 2A–F). Briefly, gBT-I T cells in the uterus and ovaries differed phenotypically from CVT gBT-I T cells. gBT-I T cells from the uterus and ovaries were predominantly CD103, and more than half of the gBT-I population in the ovaries expressed TCF-1 1 mo after LM-gB immunization (Supplemental Fig. 2B, 2F).

Given that CD103+ CD8 T cells in CVT lack substantial expression of TCF-1, we next asked whether this memory compartment was impaired in self-renewal and would undergo gradual decay. We immunized mice with LM-gB and waited for 1 mo followed by provision of BrdU for 7 d. Animals were then euthanized to measure BrdU incorporation and infer proliferative self-renewal. To facilitate interpretation of the BrdU incorporation data, we chose a 7-d labeling period to minimize confounding issues, such as peripheral recruitment and phenotypic conversion of memory cells. We found that BrdU incorporation occurred primarily among the TCF-1+ gB-specific population in the CVT and more minimally among the TCF-1 gB-specific T cell population (Fig. 3F). The overall rates of BrdU incorporation we observed were comparable to those previously reported in the female reproductive tract after LCMV infection (48). Despite this evidence of proliferation, the CVT compartment underwent a 3-fold loss in number in the 5 mo following immunization, with some mice having gB-specific CD8 T cell numbers below a reliable limit of detection at the latest timepoints (Fig. 3G, Supplemental Fig. 3A, 3B). Meanwhile, the number of gB-specific CD8 T cells in the spleen, CVT dLN, and small intestine lamina propria (SI LP) beyond 1 mo after immunization appeared stable.

Given that gB-specific CD8 T cells in the CVT continued to decline in number after T cell contraction had concluded in other tissues, we next examined the expression of canonical markers of tissue residency and effector function. To reliably determine what proportion of the gB-specific CD8 T cells in each organ were located within the tissue, we performed intravascular Ab labeling (42). We assessed whether gB-specific CD8 T cells expressed CD69 and CD103. We found that the frequency of CD103+ cells within the gB-specific CD8 T cell subset increased from 20.6 to 54.4% between 1 and 5 mo after LM-gB immunization (Fig. 4A, 4B). Whereas the i.v.-label+ fraction of gB-specific CD8 T cells in the CVT remained almost completely CD103 for the entire duration of the experiment, the shift from a predominantly CD103 population to a predominantly CD103+ population was especially clear within the i.v.-label fraction of CVT gB-specific CD8 T cells (Fig. 4B) and was mainly attributable to a relative loss of CD103 cells from the CVT over time (Fig. 4C). Conversely, the equivalent i.v. label fraction of gB-specific CD8 T cells in the SI LP did not undergo a population-level shift toward becoming CD103+ (Fig. 4D). Finally, the total CD69+ CD103+ population in the CVT was stably maintained over time (Fig. 4C), which could be due to conversion of other subsets to a CD69+ CD103+ phenotype or the ability to self-renew at a rate sufficient for stable maintenance.

FIGURE 4.

Most HSV-specific CD8 T cells remaining in the CVT by 5 mo after immunization express CD69 and CD103. (A) Representative staining showing CD69 and CD103 expression on HSV-specific CD8 T cells in the CVT 1 and 5 mo after immunization. (B and D) Stacked bar plots showing relative abundance of each of four subsets of HSV-specific CD8 T cells based on CD69 and CD103 expression in the i.v. label+ and i.v. label fractions of the CVT or SI LP. (C) Total i.v. label count of each of four subsets of HSV-specific CD8 T cells based on CD69 and CD103 expression 1 wk, 1 mo, 3 mo, and 5 mo after immunization. Data in (B)–(D) are representative of at least two independent experiments per time point. Each dot in (C) represents an individual mouse. Error bars represent mean ± SD. The p values in (C) were calculated via ordinary one-way ANOVA with Tukey post hoc test using log10-transformed values. Exact p values are given for all values <0.05.

FIGURE 4.

Most HSV-specific CD8 T cells remaining in the CVT by 5 mo after immunization express CD69 and CD103. (A) Representative staining showing CD69 and CD103 expression on HSV-specific CD8 T cells in the CVT 1 and 5 mo after immunization. (B and D) Stacked bar plots showing relative abundance of each of four subsets of HSV-specific CD8 T cells based on CD69 and CD103 expression in the i.v. label+ and i.v. label fractions of the CVT or SI LP. (C) Total i.v. label count of each of four subsets of HSV-specific CD8 T cells based on CD69 and CD103 expression 1 wk, 1 mo, 3 mo, and 5 mo after immunization. Data in (B)–(D) are representative of at least two independent experiments per time point. Each dot in (C) represents an individual mouse. Error bars represent mean ± SD. The p values in (C) were calculated via ordinary one-way ANOVA with Tukey post hoc test using log10-transformed values. Exact p values are given for all values <0.05.

Close modal

We next wanted to determine if systemic immunization resulted in constitutive granzyme B expression by CD8 T cells in the CVT, as we observed after vaginal immunization (Fig. 2) and as described for tissue-resident CD8 T cells in the SI LP, which express granzyme B in the absence of Ag re-exposure (6, 49). Few gB-specific CD8 T cells in the CVT expressed granzyme B at 1 mo after LM-gB immunization, but the granzyme B+ population substantially increased at later time points (Fig. 5A). In contrast, a subset of SI LP gB-specific CD8 T cells expressed granzyme B by 1 mo after immunization, and expression remained stable thereafter (Fig. 5B). As expected, granzyme B expression was rarely observed among gB-specific CD8 T cells in the spleen or dLN, regardless of the memory timepoint (Fig. 5B).

FIGURE 5.

After systemic immunization, cervicovaginal-resident CD8 T cells progressively upregulate granzyme B. (A) Representative staining showing granzyme B expression by HSV-specific CD8 T cells in the CVT 1 and 5 mo after immunization. (B) Frequency of granzyme B+ cells among HSV-specific CD8 T cells in spleen, dLN, CVT, and SI LP at 1, 3, and 5 mo postimmunization. (C) Number of granzyme B–expressing cells among HSV-specific CD8 T cells at time points after LM-gB immunization. (D) Concatenated flow files from 10 mice showing CD103 and granzyme B coexpression among the HSV-specific CD8b i.v. label populations in the CVT and dLN 5 mo after LM-gB immunization. (E) Frequency of CD69 and CD103 expression within granzyme B+ HSV-specific CD8 T cells in the CVT and SI LP 1, 3, and 5 mo after immunization. Data in (A)–(E) are representative of at least two independent experiments. Each dot or triangle in (B) and (C) represents an individual mouse. Error bars represent mean ± SD. The p values in (B) and (C) were calculated via ordinary one-way ANOVA with Tukey post hoc test using log10-transformed values for numbers and untransformed values for percentages. Exact p values are given for all values <0.05.

FIGURE 5.

After systemic immunization, cervicovaginal-resident CD8 T cells progressively upregulate granzyme B. (A) Representative staining showing granzyme B expression by HSV-specific CD8 T cells in the CVT 1 and 5 mo after immunization. (B) Frequency of granzyme B+ cells among HSV-specific CD8 T cells in spleen, dLN, CVT, and SI LP at 1, 3, and 5 mo postimmunization. (C) Number of granzyme B–expressing cells among HSV-specific CD8 T cells at time points after LM-gB immunization. (D) Concatenated flow files from 10 mice showing CD103 and granzyme B coexpression among the HSV-specific CD8b i.v. label populations in the CVT and dLN 5 mo after LM-gB immunization. (E) Frequency of CD69 and CD103 expression within granzyme B+ HSV-specific CD8 T cells in the CVT and SI LP 1, 3, and 5 mo after immunization. Data in (A)–(E) are representative of at least two independent experiments. Each dot or triangle in (B) and (C) represents an individual mouse. Error bars represent mean ± SD. The p values in (B) and (C) were calculated via ordinary one-way ANOVA with Tukey post hoc test using log10-transformed values for numbers and untransformed values for percentages. Exact p values are given for all values <0.05.

Close modal

We next examined the phenotypic identity of these granzyme B–expressing cells. By 3 mo postimmunization, granzyme B+ gB-specific CD8 T cells in the CVT were almost exclusively CD103+ (Fig. 5B–E), resembling our results in human samples (Fig. 1D) and suggesting that these two dynamically expressed markers were demarcating the same cellular population. These data suggest that immunization-induced gB-specific memory CD8 T cells differentiate asynchronously in the CVT relative to the gut mucosa or peripheral lymphoid organs.

Given the decline in the number of gB-specific CD8 T cells in the CVT combined with their shift toward CD103 and granzyme B expression, we next evaluated how these changes affected the immunoprotective response to vaginal HSV-2 challenge. Because LM-gB does not prime HSV-specific CD4 T or B cell responses, this immunization system allowed us to specifically assess the protective effect of CD8 T cells in the absence of other Ag-specific responses. We presumed that the most rapid antiviral responses would likely be mediated by tissue-resident gB-specific CD8 T cells, whereas circulating gB-specific CD8 T cells would contribute to later responses. We challenged immunized mice at either 1 mo (“Early Memory”) or 4 mo (“Late Memory”) after immunization (Fig. 6A). A similar proportion of mice in each group survived lethal challenge (Fig. 6B), demonstrating that gB-specific memory CD8 T cells were sufficient to confer protection against HSV-2 lethality in a subset of immunized mice. However, mice challenged earlier after immunization began to show evidence of faster viral clearance starting as early as 3.5 d after challenge (Fig. 6C) and contained a larger T cell population in the first 2.5 d after challenge (Fig. 6D and Supplemental Fig. 3C). We hypothesized that this was due to quantitative and qualitative differences between the early and late CD8 T cell compartment in the CVT.

FIGURE 6.

LM-gB–immunized mice are protected against severe disease after lethal intravaginal challenge with HSV-2, but the efficacy of the immune response wanes over time. (A) Experiment schematic to compare protective efficacy of LM-gB immunization 1 and 4 mo after administration. (B) Survival after vaginal HSV-2 challenge of naive and LM-gB–immunized mice. (C) Vaginal washes were collected after HSV-2 challenge, and HSV-2 titer was determined by plaque assay. Pink dots represent mice that survived challenge, as shown in (B). (D) Number of HSV-specific CD8 T cells (top) and HSV genomes (bottom) in the CVT with modeling estimates overlaid in blue. (E) Schematic of mathematical model relating CVT-infiltrating CD8 T cell immune response to HSV-2 viral expansion kinetics. S and I represent the number of susceptible and infected cells per CVT, respectively; E represents the number of gB tet+ CD8 T cells per CVT; and V represents the number of HSV genomes per vaginal wash. (F) Model estimates of basic reproductive ratios (number of new cells infected by one infected cell when introduced into a pool of susceptible cells) for each group of mice. (G) Scatter plot of areas under curves from (D) and corresponding predicted day of viral clearance. Dotted line represents a fitted line from quadratic regression. Data in (B), (C), and (D) are pooled from at least two independent experiments (n = 8–14 per group in B). Each dot, square, or triangle in (B), (C), and (D) represents data from an individual mouse. Each dot, square, or triangle in (F) and (G) represents one model estimate. Error bars in (C) represent mean ± SD. The p values in (C) were calculated via unpaired t test using log10-transformed values. The p values in (F) were calculated via pairwise-corrected Mann-Whitney test using Bonferroni correction. The p and rs value in (G) were calculated via Spearman rank correlation. Exact p values are given for all values <0.05.

FIGURE 6.

LM-gB–immunized mice are protected against severe disease after lethal intravaginal challenge with HSV-2, but the efficacy of the immune response wanes over time. (A) Experiment schematic to compare protective efficacy of LM-gB immunization 1 and 4 mo after administration. (B) Survival after vaginal HSV-2 challenge of naive and LM-gB–immunized mice. (C) Vaginal washes were collected after HSV-2 challenge, and HSV-2 titer was determined by plaque assay. Pink dots represent mice that survived challenge, as shown in (B). (D) Number of HSV-specific CD8 T cells (top) and HSV genomes (bottom) in the CVT with modeling estimates overlaid in blue. (E) Schematic of mathematical model relating CVT-infiltrating CD8 T cell immune response to HSV-2 viral expansion kinetics. S and I represent the number of susceptible and infected cells per CVT, respectively; E represents the number of gB tet+ CD8 T cells per CVT; and V represents the number of HSV genomes per vaginal wash. (F) Model estimates of basic reproductive ratios (number of new cells infected by one infected cell when introduced into a pool of susceptible cells) for each group of mice. (G) Scatter plot of areas under curves from (D) and corresponding predicted day of viral clearance. Dotted line represents a fitted line from quadratic regression. Data in (B), (C), and (D) are pooled from at least two independent experiments (n = 8–14 per group in B). Each dot, square, or triangle in (B), (C), and (D) represents data from an individual mouse. Each dot, square, or triangle in (F) and (G) represents one model estimate. Error bars in (C) represent mean ± SD. The p values in (C) were calculated via unpaired t test using log10-transformed values. The p values in (F) were calculated via pairwise-corrected Mann-Whitney test using Bonferroni correction. The p and rs value in (G) were calculated via Spearman rank correlation. Exact p values are given for all values <0.05.

Close modal

To better understand the relationship between CVT gB-specific CD8 T cells and viral control, we built upon our previously described approaches to model the acute T cell response to HSV-2 infection (50, 51) and used the mathematical model in Eq. 1 to characterize the different gB-specific T cell response kinetics between the early and late memory groups. In this model, susceptible cells in the CVT are infected by free HSV-2, allowing them to produce new virus or be killed by HSV-specific CD8 T cells (Fig. 6E). We performed 100 model-fit rounds to all observations simultaneously for each group for experimentally determined CVT gB-specific CD8 T cell counts and viral load and overlaid these predictions onto our experimental data (Fig. 6D,). In naive mice, gB-specific CD8 T cells were absent from or rare in the CVT at 0–3.5 d after challenge but were abundant by 6.5 d after challenge (Fig. 6D, top left), corresponding with the entry of effector T cells into the CVT following primary infection. In mice challenged at either early or late time points, gB-specific CD8 T cells increased dramatically in number between 0 and 2.5 d after HSV-2 infection, and HSV viral titer fell by 6.5 postchallenge (Fig. 6D). The best model fits predicted that naive mice had the highest basic reproductive number, meaning that they had the highest number of new cellular infections predicted to derive from a single HSV-infected cell in a susceptible pool (Fig. 6F). In addition, the greater areas under fitted gB tet+ T cell curves (shown in (Fig. 6D, top row) were associated with earlier viral clearance (Fig. 6G). These data suggest the numeric and phenotypic changes we observed within the CVT memory T cell compartment were correlated with a decrease in the efficacy of the antiviral response in the first days after challenge.

To determine if CVT-specific changes in the memory T cell compartment occurred in a tissue-autonomous manner, we depleted Thy1.1+ gBT-I T cells from the blood and lymphoid tissues of LM-gB–immunized mice 1 mo after immunization (Fig. 7A), as previously described (10). This method depleted nearly all splenic gBT-I cells but left the CVT gBT-I compartment intact (Fig. 7B, 7C). We observed that the CVT gBT-I CD8 T cell population in depleted mice declined by the 3 mo time point compared with the undepleted control (Fig. 7D). The gBT-I CD69 CD103 population was noticeably diminished in the CVT of depleted animals already at the 1 mo time point (Fig. 7E, 7F), suggesting that this population is controlled by continual peripheral reseeding. The CD103+ population in the CVT increased proportionally 3 mo following depletion (Fig. 7F). Granzyme B expression increased between 1 and 3 mo regardless of depletion status (Fig. 7G), suggesting that this process occurs without new input from the periphery. The frequency of TCF-1+ gBT-I T cells in the CVT was modestly decreased at the 3 mo time point following depletion (Fig. 7H), which is in line with the observed proportional increase of CD103+ gBT-I T cells. Finally, we challenged depleted and undepleted mice with HSV-2 (Fig. 7I). We observed a more-pronounced increase in gBT-I T cell numbers in the undepleted animals as early as day 2.5 (Fig. 7J). These data suggest that peripheral recruitment of Ag-specific memory T cells occurs rapidly in the CVT and at timepoints typically associated with TRM-mediated in situ responses.

FIGURE 7.

Differentiation and decay of the CVT-resident CD8 T cell compartment are tissue intrinsic. (A) Schematic of experiment to test effect of depleting circulating gBT-I cells. (B) Example flow staining to identify gBT-I population within CD8 T cells from spleen and CVT of depleted or undepleted mice 1 mo after immunization. (C and D) Number of gBT-I cells recovered from spleen or CVT of mice at indicated time points after immunization. (E) Relative abundance of CD69 and CD103 subsets within i.v. label gBT-I cells in CVT at indicated time points. (FH) Frequency of CD103, granzyme B, and TCF-1 expression among gBT-I cells. (I) Schematic of experiment to test effect of depletion on kinetics of CVT T cell expansion. (J) Number of gBT-I cells and endogenous gB tetramer+ cells in the CVT at indicated time points after HSV-2 challenge, with Thy1.1 depletion (green) or without Thy1.1 depletion (black). Data in (C)–(G) are representative of two independent experiments. Data in (J) derive from one experiment. Flow plots in (B) are derived from two representative mice. Each dot in (C)–(J) represents an individual mouse. Error bars represent mean ± SD. The p values in (C), (E), (F), and (G) were calculated via unpaired t test comparing 1 to 3 mo within or across each depletion group. The p values in (I) were calculated via two-way ANOVA and Sidak posttest comparing depleted to undepleted samples within each timepoint. Exact p values are given for all values <0.05.

FIGURE 7.

Differentiation and decay of the CVT-resident CD8 T cell compartment are tissue intrinsic. (A) Schematic of experiment to test effect of depleting circulating gBT-I cells. (B) Example flow staining to identify gBT-I population within CD8 T cells from spleen and CVT of depleted or undepleted mice 1 mo after immunization. (C and D) Number of gBT-I cells recovered from spleen or CVT of mice at indicated time points after immunization. (E) Relative abundance of CD69 and CD103 subsets within i.v. label gBT-I cells in CVT at indicated time points. (FH) Frequency of CD103, granzyme B, and TCF-1 expression among gBT-I cells. (I) Schematic of experiment to test effect of depletion on kinetics of CVT T cell expansion. (J) Number of gBT-I cells and endogenous gB tetramer+ cells in the CVT at indicated time points after HSV-2 challenge, with Thy1.1 depletion (green) or without Thy1.1 depletion (black). Data in (C)–(G) are representative of two independent experiments. Data in (J) derive from one experiment. Flow plots in (B) are derived from two representative mice. Each dot in (C)–(J) represents an individual mouse. Error bars represent mean ± SD. The p values in (C), (E), (F), and (G) were calculated via unpaired t test comparing 1 to 3 mo within or across each depletion group. The p values in (I) were calculated via two-way ANOVA and Sidak posttest comparing depleted to undepleted samples within each timepoint. Exact p values are given for all values <0.05.

Close modal

Finally, we wished to determine whether site-matched immunization would result in improved maintenance of the CVT memory CD8 T cell compartment. We compared the maintenance of HSV-specific CD8 T cells in the CVT of mice immunized vaginally with HSV-2 TK- or i.v. with LM-gB (Fig. 8A). We observed that vaginal immunization with HSV-2 TK- did not result in a larger memory population of HSV-specific CD8 T cells in the CVT than systemic immunization with LM-gB (Fig. 8B). In addition, the gB-specific population induced by vaginal immunization was relatively stably maintained in the spleen and SI LP but underwent decay in the CVT between 1 and 3 mo after immunization (Fig. 8B), highlighting that the decay occurs independently of the immunization route.

FIGURE 8.

Decay of the CVT-resident CD8 T cell compartment occurs regardless of priming infection and route. (A) Schematic of experiment to assess effect of vaginal immunization on longevity of memory CD8 T cells in CVT. (B) Number of gB tetramer+ cells recovered from CVT, SI LP, or spleen of mice at indicated time points after immunization. Data are representative of two independent experiments. Each dot in (B) represents an individual mouse. Error bars represent mean ± SD. The p values in (B) were calculated via unpaired t test comparing 1 to 3 mo within each immunization group. Exact p values are given for all values <0.05.

FIGURE 8.

Decay of the CVT-resident CD8 T cell compartment occurs regardless of priming infection and route. (A) Schematic of experiment to assess effect of vaginal immunization on longevity of memory CD8 T cells in CVT. (B) Number of gB tetramer+ cells recovered from CVT, SI LP, or spleen of mice at indicated time points after immunization. Data are representative of two independent experiments. Each dot in (B) represents an individual mouse. Error bars represent mean ± SD. The p values in (B) were calculated via unpaired t test comparing 1 to 3 mo within each immunization group. Exact p values are given for all values <0.05.

Close modal

In this article, we report that CVT microenvironment plays a dynamic role in shaping the numeric, phenotypic, and functional characteristics of the tissue memory CD8 T cells, which thereby impacts their immunoprotective potential. The cervicovaginal mucosa is distinct from other parts of the female reproductive tract and other mucosal tissues in terms of epithelium type and physiology. Specifically, the vagina and ectocervix are composed of type II mucosa, and the endocervix and uterus are composed of type I mucosa, with the cervical transformation zone marking the area where the simple columnar epithelial cells of the endocervix meet the stratified squamous epithelial cells of the ectocervix and vagina. In contrast, the gut and the lung, other mucosal tissues commonly studied in the case of CD8 T cell memory, are type I mucosal surfaces (52). In addition to these differences in the structure of the epithelium, there are differences in the complexity and makeup of the microbiota at different mucosal tissue sites (53), which may also contribute to tissue-specific environmental differences that could affect memory CD8 T cell differentiation and function. Vaginal-resident CD8 T cells in healthy women did not express TCF-1 (Fig. 1), which could either simply indicate recent activation of these T cells (recent infection) or a memory population with an impaired ability to self-renew. Importantly, given that vaginal CD8 T cells from multiple donors uniformly lacked TCF-1 expression, it suggested that a tissue-driven phenotype was more likely than a recent infection in a cohort of healthy donors with no current genital infections. Using a mouse model, we similarly found that both vaginal (Fig. 2) and systemic (Fig. 3) immunization resulted in a granzyme B+, TCF-1low phenotype among CD8 T cells in the mouse CVT. These data demonstrate that memory CD8 T cell phenotype is driven by immunization and tissue microenvironment and appears conserved between the mouse and human CVT.

Over the course of 5 mo following immunization, Ag-specific memory CD8 T cells in the CVT numerically declined and gradually acquired expression of CD103 and granzyme B. Conversely, HSV-specific CD8 T cells that were circulating (spleen, lymph node) or resident in the small intestine were numerically and phenotypically stable over the same time course. This decline in the CVT TRM compartment occurred at a similar rate even when circulating gBT-I T cells were selectively depleted, suggesting that peripheral T cells only minimally reseed CVT under homeostatic conditions (Fig. 7). These results allow for two possible interpretations: 1) the CVT TRM compartment is short-lived compared with that in other organs or 2) a small proportion of the T cells present in the CVT 1 mo after immunization are stably maintained, bona fide TRM, while the remaining memory T cells either slowly exit the tissue over time or undergo programmed cell death. Ultimately, fate-mapping studies will be required to determine the relationship of the CD103 to CD103+ T cell population in the CVT. Importantly, the notion that a TRM population is short-lived is not without precedent. Several reports indicate that tissue-resident memory T cells are short-lived in the lung (5456) with a pulse chase experiment suggesting that their half-life is only ∼5 d (57). It has been proposed that the short-lived nature of tissue-resident memory T cells in the lung is related to the tissue environment (5860). Overall, the role of tissue microenvironment appears critical for providing signals that are important for resident-memory formation and long-term survival (60). Thus, tissue-resident memory T cells can be autonomously maintained and long-lived as shown in studies examining skin and gut but also short-lived and dependent on peripheral reseeding or continuous Ag encounter (54, 56, 61, 62). Our data presented in this article are in line with a model in which the memory T cell compartment in the CVT falls in the middle of this longevity spectrum, and at least one memory population (CD69 CD103) depends on peripheral reseeding.

It is noteworthy that we did not observe strong TCF-1 expression in the gut TRM cells either, and yet these cells were stably maintained over time. This could either suggest that TCF-1 is not required for self-renewal of TRM or imply the existence of other compensatory mechanisms to maintain the TRM compartment in certain mucosal tissues. TCF-1 activity is thought to be regulated by β-catenin and the Wnt signaling pathway. However, β-catenin does not regulate the phenotype of memory T cells (63) and β-catenin–deficient T cells mount normal recall responses (64). Because β-catenin is sufficient, but not necessary, to activate TCF-1 (65), a β-catenin–independent pathway may control TCF-1 activity in tertiary tissues with distinct cell fate outcomes compared with β-catenin–dependent activation. A recent study provides some genetic evidence that TCF-1 negatively regulates CD103 expression in the lung tissue (47), indicating the need to further study and dissect the role of TCF-1 in controlling differentiation and maintenance of TRM cells.

Understanding the relationship between TRM in mucosal tissues and the periphery is highly relevant because memory T cell characteristics and frequencies in the blood are used as a benchmark in clinical studies to assess immune responses to vaccines and establish correlates of protection (6669). Our data indicate that there is a disconnect between memory CD8 T cell frequency, phenotype, and stability in the periphery and the CVT, which was driven by tissue residence, whereas local inflammation appeared to have limited effects. Although we did not determine what factor was causally associated with the reduction in protection, the numerical and phenotypical changes of the memory CD8 T cell compartment 1 mo versus 5 mo postimmunization ultimately influenced the rapidity of the immune response to vaginal viral challenge (Fig. 6). This reduction in rapidity may be of particular relevance to the success of a vaccine-induced memory T cell response within the CVT, as the host-pathogen events occurring within this early window likely result in the ultimate success or failure of the protective response (70), and even our observed rapid recruitment of peripheral memory T cells to CVT may be too late to prevent establishment of infection.

Barrier immunity in the CVT remains poorly understood but is of particular interest because STIs have a significant impact on global health. We currently lack effective vaccines to prevent the majority of these infections, including HIV, HSV, and bacterial STIs. Our data demonstrate that following systemic immunization, animals were partially protected against HSV-2 lethality by gB-specific memory CD8 T cells (Fig. 6). Of note, in the HSV-2 infection model, viral titer in the vaginal lumen reflects protection but may not necessarily predict survival, which depends on viral access to the nervous system. Importantly, we observed this protective effect despite the lack of HSV-specific CD4 T or B cell memory, elements of the immune response that are often considered crucial for protection against HSV-2 (7174). In line with this observation, results from participants with recurrent genital HSV-2 infection indeed suggest that CD8 T cell activity within lesions is associated with viral control (29, 50, 75, 76). It is possible that a new focus on vaccine strategies that elicit memory CD8 T cells within the genital mucosa, the site of first exposure to STIs, will increase vaccine efficacy. However, successful design of these approaches, not just in regards to adjuvant and immunization route but also boosting intervals to optimize protection, relies on understanding how memory CD8 T cells are maintained within these genital tract mucosal tissues.

In summary, we demonstrate that the CVT tissue environment controls memory CD8 T cell differentiation and maintenance, which occur in a tissue-autonomous manner and differ across mucosal tissue compartments. The CD8 TRM compartment in the CVT declines over time with a concomitant decrease in the ability to mediate rapid Ag-specific protection.

We thank Kristin Weakly for excellent technical assistance and Sean M. Hughes for assistance with statistical analyses.

This work was supported by National Institute of Allergy and Infectious Diseases, National Institutes of Health (NIH) Grants R01AI123323 (M.P.), R01 AI121129 (J.T.S., J.M.L., and M.P.), R01 AI131914 (J.M.L.), R01 AI141435 (J.M.L.), and T32 AI07140, the Doug and Maggie Walker Fellowship (A.S.W.-D.), NIH Grant T32 AI007509, and Division of Intramural Research, National Institute of Allergy and Infectious Diseases Grant F31 AI140514 (V.A.D.). F.M. is an International Society for Advancement of Cytometry scholar.

The online version of this article contains supplemental material.

Abbreviations used in this article

AIC

Akaike Information Criteria

CD8 TRM

tissue-resident memory CD8 T cell

CVT

cervicovaginal tissue

dLN

draining lymph node

FHCRC

Fred Hutchinson Cancer Research Center

gB

glycoprotein B

HSV-2 TK-

thymidine-kinase–deficient HSV-2

MDPA

medroxyprogesterone acetate

SI LP

small intestine lamina propria

STI

sexually transmitted infection

1.
Wakim
L. M.
,
J.
Waithman
,
N.
van Rooijen
,
W. R.
Heath
,
F. R.
Carbone
.
2008
.
Dendritic cell-induced memory T cell activation in nonlymphoid tissues.
Science
319
:
198
202
.
2.
Gebhardt
T.
,
L. M.
Wakim
,
L.
Eidsmo
,
P. C.
Reading
,
W. R.
Heath
,
F. R.
Carbone
.
2009
.
Memory T cells in nonlymphoid tissue that provide enhanced local immunity during infection with herpes simplex virus.
Nat. Immunol.
10
:
524
530
.
3.
Masopust
D.
,
D.
Choo
,
V.
Vezys
,
E. J.
Wherry
,
J.
Duraiswamy
,
R.
Akondy
,
J.
Wang
,
K. A.
Casey
,
D. L.
Barber
,
K. S.
Kawamura
, et al
2010
.
Dynamic T cell migration program provides resident memory within intestinal epithelium.
J. Exp. Med.
207
:
553
564
.
4.
Szabo
P. A.
,
M.
Miron
,
D. L.
Farber
.
2019
.
Location, location, location: tissue resident memory T cells in mice and humans.
Sci. Immunol.
4
:
eaas9673
.
5.
Masopust
D.
,
A. G.
Soerens
.
2019
.
Tissue-resident T cells and other resident leukocytes.
Annu. Rev. Immunol.
37
:
521
546
.
6.
Casey
K. A.
,
K. A.
Fraser
,
J. M.
Schenkel
,
A.
Moran
,
M. C.
Abt
,
L. K.
Beura
,
P. J.
Lucas
,
D.
Artis
,
E. J.
Wherry
,
K.
Hogquist
, et al
2012
.
Antigen-independent differentiation and maintenance of effector-like resident memory T cells in tissues.
J. Immunol.
188
:
4866
4875
.
7.
Steinbach
K.
,
I.
Vincenti
,
M.
Kreutzfeldt
,
N.
Page
,
A.
Muschaweckh
,
I.
Wagner
,
I.
Drexler
,
D.
Pinschewer
,
T.
Korn
,
D.
Merkler
.
2016
.
Brain-resident memory T cells represent an autonomous cytotoxic barrier to viral infection.
J. Exp. Med.
213
:
1571
1587
.
8.
Beura
L. K.
,
J. S.
Mitchell
,
E. A.
Thompson
,
J. M.
Schenkel
,
J.
Mohammed
,
S.
Wijeyesinghe
,
R.
Fonseca
,
B. J.
Burbach
,
H. D.
Hickman
,
V.
Vezys
, et al
2018
.
Intravital mucosal imaging of CD8+ resident memory T cells shows tissue-autonomous recall responses that amplify secondary memory.
Nat. Immunol.
19
:
173
182
.
9.
Park
S. L.
,
A.
Zaid
,
J. L.
Hor
,
S. N.
Christo
,
J. E.
Prier
,
B.
Davies
,
Y. O.
Alexandre
,
J. L.
Gregory
,
T. A.
Russell
,
T.
Gebhardt
, et al
2018
.
Local proliferation maintains a stable pool of tissue-resident memory T cells after antiviral recall responses.
Nat. Immunol.
19
:
183
191
.
10.
Schenkel
J. M.
,
K. A.
Fraser
,
V.
Vezys
,
D.
Masopust
.
2013
.
Sensing and alarm function of resident memory CD8+ T cells. [Published erratum appears in 2013 Nat. Immunol. 14: 876.]
Nat. Immunol.
14
:
509
513
.
11.
Ariotti
S.
,
M. A.
Hogenbirk
,
F. E.
Dijkgraaf
,
L. L.
Visser
,
M. E.
Hoekstra
,
J. Y.
Song
,
H.
Jacobs
,
J. B.
Haanen
,
T. N.
Schumacher
.
2014
.
T cell memory. Skin-resident memory CD8+ T cells trigger a state of tissue-wide pathogen alert.
Science
346
:
101
105
.
12.
Khan
T. N.
,
J. L.
Mooster
,
A. M.
Kilgore
,
J. F.
Osborn
,
J. C.
Nolz
.
2016
.
Local antigen in nonlymphoid tissue promotes resident memory CD8+ T cell formation during viral infection.
J. Exp. Med.
213
:
951
966
.
13.
Davies
B.
,
J. E.
Prier
,
C. M.
Jones
,
T.
Gebhardt
,
F. R.
Carbone
,
L. K.
Mackay
.
2017
.
Cutting edge: tissue-resident memory T cells generated by multiple immunizations or localized deposition provide enhanced immunity.
J. Immunol.
198
:
2233
2237
.
14.
Mackay
L. K.
,
A. T.
Stock
,
J. Z.
Ma
,
C. M.
Jones
,
S. J.
Kent
,
S. N.
Mueller
,
W. R.
Heath
,
F. R.
Carbone
,
T.
Gebhardt
.
2012
.
Long-lived epithelial immunity by tissue-resident memory T (TRM) cells in the absence of persisting local antigen presentation.
Proc. Natl. Acad. Sci. USA
109
:
7037
7042
.
15.
Shin
H.
,
A.
Iwasaki
.
2012
.
A vaccine strategy that protects against genital herpes by establishing local memory T cells.
Nature
491
:
463
467
.
16.
Zhang
N.
,
M. J.
Bevan
.
2013
.
Transforming growth factor-β signaling controls the formation and maintenance of gut-resident memory T cells by regulating migration and retention.
Immunity
39
:
687
696
.
17.
Iijima
N.
,
A.
Iwasaki
.
2014
.
T cell memory. A local macrophage chemokine network sustains protective tissue-resident memory CD4 T cells.
Science
346
:
93
98
.
18.
Mackay
L. K.
,
E.
Wynne-Jones
,
D.
Freestone
,
D. G.
Pellicci
,
L. A.
Mielke
,
D. M.
Newman
,
A.
Braun
,
F.
Masson
,
A.
Kallies
,
G. T.
Belz
,
F. R.
Carbone
.
2015
.
T-box transcription factors combine with the cytokines TGF-β and IL-15 to control tissue-resident memory T cell fate.
Immunity
43
:
1101
1111
.
19.
Bergsbaken
T.
,
M. J.
Bevan
,
P. J.
Fink
.
2017
.
Local inflammatory cues regulate differentiation and persistence of CD8+ tissue-resident memory T cells.
Cell Rep.
19
:
114
124
.
20.
Reinhardt
R. L.
,
A.
Khoruts
,
R.
Merica
,
T.
Zell
,
M. K.
Jenkins
.
2001
.
Visualizing the generation of memory CD4 T cells in the whole body.
Nature
410
:
101
105
.
21.
Pope
C.
,
S. K.
Kim
,
A.
Marzo
,
D.
Masopust
,
K.
Williams
,
J.
Jiang
,
H.
Shen
,
L.
Lefrancois
.
2001
.
Organ-specific regulation of the CD8 T cell response to Listeria monocytogenes infection.
J. Immunol.
166
:
3402
3409
.
22.
Masopust
D.
,
V.
Vezys
,
A. L.
Marzo
,
L.
Lefrançois
.
2001
.
Preferential localization of effector memory cells in nonlymphoid tissue.
Science
291
:
2413
2417
.
23.
Steinert
E. M.
,
J. M.
Schenkel
,
K. A.
Fraser
,
L. K.
Beura
,
L. S.
Manlove
,
B. Z.
Igyártó
,
P. J.
Southern
,
D.
Masopust
.
2015
.
Quantifying memory CD8 T cells reveals regionalization of immunosurveillance.
Cell
161
:
737
749
.
24.
Milner
J. J.
,
C.
Toma
,
B.
Yu
,
K.
Zhang
,
K.
Omilusik
,
A. T.
Phan
,
D.
Wang
,
A. J.
Getzler
,
T.
Nguyen
,
S.
Crotty
, et al
2017
.
Runx3 programs CD8+ T cell residency in non-lymphoid tissues and tumours. [Published erratum appears in 2018 Nature 554: 392.]
Nature
552
:
253
257
.
25.
Masopust
D.
,
V.
Vezys
,
E. J.
Usherwood
,
L. S.
Cauley
,
S.
Olson
,
A. L.
Marzo
,
R. L.
Ward
,
D. L.
Woodland
,
L.
Lefrancois
.
2004
.
Activated primary and memory CD8 T cells migrate to nonlymphoid tissues regardless of site of activation or tissue of origin.
J. Immunol.
172
:
4875
4882
.
26.
Jiang
X.
,
R. A.
Clark
,
L.
Liu
,
A. J.
Wagers
,
R. C.
Fuhlbrigge
,
T. S.
Kupper
.
2012
.
Skin infection generates non-migratory memory CD8+ T(RM) cells providing global skin immunity.
Nature
483
:
227
231
.
27.
Mackay
L. K.
,
A.
Rahimpour
,
J. Z.
Ma
,
N.
Collins
,
A. T.
Stock
,
M. L.
Hafon
,
J.
Vega-Ramos
,
P.
Lauzurica
,
S. N.
Mueller
,
T.
Stefanovic
, et al
2013
.
The developmental pathway for CD103(+)CD8+ tissue-resident memory T cells of skin.
Nat. Immunol.
14
:
1294
1301
.
28.
Watanabe
R.
,
A.
Gehad
,
C.
Yang
,
L. L.
Scott
,
J. E.
Teague
,
C.
Schlapbach
,
C. P.
Elco
,
V.
Huang
,
T. R.
Matos
,
T. S.
Kupper
,
R. A.
Clark
.
2015
.
Human skin is protected by four functionally and phenotypically discrete populations of resident and recirculating memory T cells.
Sci. Transl. Med.
7
:
279ra39
.
29.
Zhu
J.
,
T.
Peng
,
C.
Johnston
,
K.
Phasouk
,
A. S.
Kask
,
A.
Klock
,
L.
Jin
,
K.
Diem
,
D. M.
Koelle
,
A.
Wald
, et al
2013
.
Immune surveillance by CD8αα+ skin-resident T cells in human herpes virus infection. [Published erratum appears in 2013 Nature 500: 242.]
Nature
497
:
494
497
.
30.
Hofmann
M.
,
H.
Pircher
.
2011
.
E-cadherin promotes accumulation of a unique memory CD8 T-cell population in murine salivary glands.
Proc. Natl. Acad. Sci. USA
108
:
16741
16746
.
31.
Takamura
S.
,
H.
Yagi
,
Y.
Hakata
,
C.
Motozono
,
S. R.
McMaster
,
T.
Masumoto
,
M.
Fujisawa
,
T.
Chikaishi
,
J.
Komeda
,
J.
Itoh
, et al
2016
.
Specific niches for lung-resident memory CD8+ T cells at the site of tissue regeneration enable CD69-independent maintenance.
J. Exp. Med.
213
:
3057
3073
.
32.
Walsh
D. A.
,
H.
Borges da Silva
,
L. K.
Beura
,
C.
Peng
,
S. E.
Hamilton
,
D.
Masopust
,
S. C.
Jameson
.
2019
.
The functional requirement for CD69 in establishment of resident memory CD8(+) T cells varies with tissue location.
J. Immunol.
203
:
946
955
.
33.
Kim
S. K.
,
D. S.
Reed
,
W. R.
Heath
,
F.
Carbone
,
L.
Lefrancois
.
1997
.
Activation and migration of CD8 T cells in the intestinal mucosa.
J. Immunol.
159
:
4295
4306
.
34.
Piet
B.
,
G. J.
de Bree
,
B. S.
Smids-Dierdorp
,
C. M.
van der Loos
,
E. B.
Remmerswaal
,
J. H.
von der Thüsen
,
J. M.
van Haarst
,
J. P.
Eerenberg
,
A.
ten Brinke
,
W.
van der Bij
, et al
2011
.
CD8+ T cells with an intraepithelial phenotype upregulate cytotoxic function upon influenza infection in human lung.
J. Clin. Invest.
121
:
2254
2263
.
35.
Beura
L. K.
,
N. J.
Fares-Frederickson
,
E. M.
Steinert
,
M. C.
Scott
,
E. A.
Thompson
,
K. A.
Fraser
,
J. M.
Schenkel
,
V.
Vezys
,
D.
Masopust
.
2019
.
CD4+ resident memory T cells dominate immunosurveillance and orchestrate local recall responses.
J. Exp. Med.
216
:
1214
1229
.
36.
Pattacini
L.
,
A.
Woodward Davis
,
J.
Czartoski
,
F.
Mair
,
S.
Presnell
,
S. M.
Hughes
,
O.
Hyrien
,
G. M.
Lentz
,
A. C.
Kirby
,
M. F.
Fialkow
, et al
2019
.
A pro-inflammatory CD8+ T-cell subset patrols the cervicovaginal tract.
Mucosal Immunol.
12
:
1118
1129
.
37.
Parr
M. B.
,
L.
Kepple
,
M. R.
McDermott
,
M. D.
Drew
,
J. J.
Bozzola
,
E. L.
Parr
.
1994
.
A mouse model for studies of mucosal immunity to vaginal infection by herpes simplex virus type 2.
Lab. Invest.
70
:
369
380
.
38.
Jerome
K. R.
,
M. L.
Huang
,
A.
Wald
,
S.
Selke
,
L.
Corey
.
2002
.
Quantitative stability of DNA after extended storage of clinical specimens as determined by real-time PCR.
J. Clin. Microbiol.
40
:
2609
2611
.
39.
Lauer
P.
,
M. Y.
Chow
,
M. J.
Loessner
,
D. A.
Portnoy
,
R.
Calendar
.
2002
.
Construction, characterization, and use of two Listeria monocytogenes site-specific phage integration vectors. [Published erratum appears in 2003 J. Bacteriol. 185: 1484.]
J. Bacteriol.
184
:
4177
4186
.
40.
Zehn
D.
,
S. Y.
Lee
,
M. J.
Bevan
.
2009
.
Complete but curtailed T-cell response to very low-affinity antigen.
Nature
458
:
211
214
.
41.
Orr
M. T.
,
N. N.
Orgun
,
C. B.
Wilson
,
S. S.
Way
.
2007
.
Cutting edge: recombinant Listeria monocytogenes expressing a single immune-dominant peptide confers protective immunity to herpes simplex virus-1 infection.
J. Immunol.
178
:
4731
4735
.
42.
Galkina
E.
,
J.
Thatte
,
V.
Dabak
,
M. B.
Williams
,
K.
Ley
,
T. J.
Braciale
.
2005
.
Preferential migration of effector CD8+ T cells into the interstitium of the normal lung.
J. Clin. Invest.
115
:
3473
3483
.
43.
McKinnon
L. R.
,
S. M.
Hughes
,
S. C.
De Rosa
,
J. A.
Martinson
,
J.
Plants
,
K. E.
Brady
,
P. P.
Gumbi
,
D. J.
Adams
,
L.
Vojtech
,
C. G.
Galloway
, et al
2014
.
Optimizing viable leukocyte sampling from the female genital tract for clinical trials: an international multi-site study.
PLoS One
9
:
e85675
.
44.
Zhou
X.
,
S.
Yu
,
D. M.
Zhao
,
J. T.
Harty
,
V. P.
Badovinac
,
H. H.
Xue
.
2010
.
Differentiation and persistence of memory CD8(+) T cells depend on T cell factor 1.
Immunity
33
:
229
240
.
45.
Wakim
L. M.
,
A.
Woodward-Davis
,
R.
Liu
,
Y.
Hu
,
J.
Villadangos
,
G.
Smyth
,
M. J.
Bevan
.
2012
.
The molecular signature of tissue resident memory CD8 T cells isolated from the brain.
J. Immunol.
189
:
3462
3471
.
46.
Behr
F. M.
,
N. A. M.
Kragten
,
T. H.
Wesselink
,
B.
Nota
,
R. A. W.
van Lier
,
D.
Amsen
,
R.
Stark
,
P.
Hombrink
,
K. P. J. M.
van Gisbergen
.
2019
.
Blimp-1 rather than hobit drives the formation of tissue-resident memory CD8+ T cells in the lungs.
Front. Immunol.
10
:
400
.
47.
Wu
J.
,
A.
Madi
,
A.
Mieg
,
A.
Hotz-Wagenblatt
,
N.
Weisshaar
,
S.
Ma
,
K.
Mohr
,
T.
Schlimbach
,
M.
Hering
,
H.
Borgers
,
G.
Cui
.
2020
.
T cell factor 1 suppresses CD103+ lung tissue-resident memory T cell development.
Cell Rep.
31
:
107484
.
48.
Schenkel
J. M.
,
K. A.
Fraser
,
K. A.
Casey
,
L. K.
Beura
,
K. E.
Pauken
,
V.
Vezys
,
D.
Masopust
.
2016
.
IL-15-independent maintenance of tissue-resident and boosted effector memory CD8 T cells.
J. Immunol.
196
:
3920
3926
.
49.
Masopust
D.
,
V.
Vezys
,
E. J.
Wherry
,
D. L.
Barber
,
R.
Ahmed
.
2006
.
Cutting edge: gut microenvironment promotes differentiation of a unique memory CD8 T cell population.
J. Immunol.
176
:
2079
2083
.
50.
Schiffer
J. T.
,
L.
Abu-Raddad
,
K. E.
Mark
,
J.
Zhu
,
S.
Selke
,
D. M.
Koelle
,
A.
Wald
,
L.
Corey
.
2010
.
Mucosal host immune response predicts the severity and duration of herpes simplex virus-2 genital tract shedding episodes.
Proc. Natl. Acad. Sci. USA
107
:
18973
18978
.
51.
Schiffer
J. T.
,
D.
Swan
,
R.
Al Sallaq
,
A.
Magaret
,
C.
Johnston
,
K. E.
Mark
,
S.
Selke
,
N.
Ocbamichael
,
S.
Kuntz
,
J.
Zhu
, et al
2013
.
Rapid localized spread and immunologic containment define Herpes simplex virus-2 reactivation in the human genital tract.
eLife
2
:
e00288
.
52.
Iwasaki
A.
2010
.
Antiviral immune responses in the genital tract: clues for vaccines.
Nat. Rev. Immunol.
10
:
699
711
.
53.
Anahtar
M. N.
,
D. B.
Gootenberg
,
C. M.
Mitchell
,
D. S.
Kwon
.
2018
.
Cervicovaginal microbiota and reproductive health: the virtue of simplicity.
Cell Host Microbe
23
:
159
168
.
54.
Wu
T.
,
Y.
Hu
,
Y. T.
Lee
,
K. R.
Bouchard
,
A.
Benechet
,
K.
Khanna
,
L. S.
Cauley
.
2014
.
Lung-resident memory CD8 T cells (TRM) are indispensable for optimal cross-protection against pulmonary virus infection.
J. Leukoc. Biol.
95
:
215
224
.
55.
Takamura
S.
,
S.
Kato
,
C.
Motozono
,
T.
Shimaoka
,
S.
Ueha
,
K.
Matsuo
,
K.
Miyauchi
,
T.
Masumoto
,
A.
Katsushima
,
T.
Nakayama
, et al
2019
.
Interstitial-resident memory CD8+ T cells sustain frontline epithelial memory in the lung.
J. Exp. Med.
216
:
2736
2747
.
56.
van de Wall
S.
,
V. P.
Badovinac
,
J. T.
Harty
.
2021
.
Influenza-specific lung-resident memory CD8+ T cells.
Cold Spring Harb. Perspect. Biol.
13
:
a037978
.
57.
Slütter
B.
,
N.
Van Braeckel-Budimir
,
G.
Abboud
,
S. M.
Varga
,
S.
Salek-Ardakani
,
J. T.
Harty
.
2017
.
Dynamics of influenza-induced lung-resident memory T cells underlie waning heterosubtypic immunity.
Sci. Immunol.
2
:
eaag2031
.
58.
Hayward
S. L.
,
C. D.
Scharer
,
E. K.
Cartwright
,
S.
Takamura
,
Z. T.
Li
,
J. M.
Boss
,
J. E.
Kohlmeier
.
2020
.
Environmental cues regulate epigenetic reprogramming of airway-resident memory CD8+ T cells.
Nat. Immunol.
21
:
309
320
.
59.
Van Braeckel-Budimir
N.
,
J. T.
Harty
.
2017
.
Influenza-induced lung Trm: not all memories last forever.
Immunol. Cell Biol.
95
:
651
655
.
60.
Liu
Y.
,
C.
Ma
,
N.
Zhang
.
2018
.
Tissue-specific control of tissue-resident memory T cells.
Crit. Rev. Immunol.
38
:
79
103
.
61.
Ely
K. H.
,
T.
Cookenham
,
A. D.
Roberts
,
D. L.
Woodland
.
2006
.
Memory T cell populations in the lung airways are maintained by continual recruitment.
J. Immunol.
176
:
537
543
.
62.
Uddbäck
I.
,
E. K.
Cartwright
,
A. S.
Schøller
,
A. N.
Wein
,
S. L.
Hayward
,
J.
Lobby
,
S.
Takamura
,
A. R.
Thomsen
,
J. E.
Kohlmeier
,
J. P.
Christensen
.
2021
.
Long-term maintenance of lung resident memory T cells is mediated by persistent antigen.
Mucosal Immunol.
14
:
92
99
.
63.
Driessens
G.
,
Y.
Zheng
,
T. F.
Gajewski
.
2010
.
Beta-catenin does not regulate memory T cell phenotype.
Nat. Med.
16
:
513
514
,
author reply 514–515
.
64.
Prlic
M.
,
M. J.
Bevan
.
2011
.
Cutting edge: beta-catenin is dispensable for T cell effector differentiation, memory formation, and recall responses.
J. Immunol.
187
:
1542
1546
.
65.
Tiemessen
M. M.
,
M. R.
Baert
,
L.
Kok
,
M. C.
van Eggermond
,
P. J.
van den Elsen
,
R.
Arens
,
F. J.
Staal
.
2014
.
T cell factor 1 represses CD8+ effector T cell formation and function.
J. Immunol.
193
:
5480
5487
.
66.
Lewinsohn
D. A.
,
D. M.
Lewinsohn
,
T. J.
Scriba
.
2017
.
Polyfunctional CD4+ T cells as targets for tuberculosis vaccination.
Front. Immunol.
8
:
1262
.
67.
Streeck
H.
2016
.
Designing optimal HIV-vaccine T-cell responses.
Curr. Opin. HIV AIDS
11
:
593
600
.
68.
Stephenson
K. E.
,
H. T.
D’Couto
,
D. H.
Barouch
.
2016
.
New concepts in HIV-1 vaccine development.
Curr. Opin. Immunol.
41
:
39
46
.
69.
Koup
R. A.
,
B. S.
Graham
,
D. C.
Douek
.
2011
.
The quest for a T cell-based immune correlate of protection against HIV: a story of trials and errors.
Nat. Rev. Immunol.
11
:
65
70
.
70.
Haase
A. T.
2010
.
Targeting early infection to prevent HIV-1 mucosal transmission.
Nature
464
:
217
223
.
71.
Milligan
G. N.
,
D. I.
Bernstein
,
N.
Bourne
.
1998
.
T lymphocytes are required for protection of the vaginal mucosae and sensory ganglia of immune mice against reinfection with herpes simplex virus type 2.
J. Immunol.
160
:
6093
6100
.
72.
Morrison
L. A.
,
X. J.
Da Costa
,
D. M.
Knipe
.
1998
.
Influence of mucosal and parenteral immunization with a replication-defective mutant of HSV-2 on immune responses and protection from genital challenge.
Virology
243
:
178
187
.
73.
Dudley
K. L.
,
N.
Bourne
,
G. N.
Milligan
.
2000
.
Immune protection against HSV-2 in B-cell-deficient mice.
Virology
270
:
454
463
.
74.
Petro
C. D.
,
B.
Weinrick
,
N.
Khajoueinejad
,
C.
Burn
,
R.
Sellers
,
W. R.
Jacobs
Jr.
,
B. C.
Herold
.
2016
.
HSV-2 ΔgD elicits FcγR-effector antibodies that protect against clinical isolates.
JCI Insight
1
:
e88529
.
75.
Peng
T.
,
J.
Zhu
,
K.
Phasouk
,
D. M.
Koelle
,
A.
Wald
,
L.
Corey
.
2012
.
An effector phenotype of CD8+ T cells at the junction epithelium during clinical quiescence of herpes simplex virus 2 infection.
J. Virol.
86
:
10587
10596
.
76.
Zhu
J.
,
D. M.
Koelle
,
J.
Cao
,
J.
Vazquez
,
M. L.
Huang
,
F.
Hladik
,
A.
Wald
,
L.
Corey
.
2007
.
Virus-specific CD8+ T cells accumulate near sensory nerve endings in genital skin during subclinical HSV-2 reactivation.
J. Exp. Med.
204
:
595
603
.

The authors have no financial conflicts of interest.

Supplementary data