Research ArticleDengue

Virus-specific T lymphocytes home to the skin during natural dengue infection

See allHide authors and affiliations

Science Translational Medicine  11 Mar 2015:
Vol. 7, Issue 278, pp. 278ra35
DOI: 10.1126/scitranslmed.aaa0526

Getting under dengue virus’s skin

Dengue virus infection is transmitted by mosquitoes, suggesting that a vaccine targeting the immune response to the skin could have protective effects. Rivino et al. examined individuals with natural dengue infection to determine whether skin-mediated immunity indeed contributes to the fight against dengue. They found that dengue infection induced highly activated CD8+ T cells that express the skin-homing marker cutaneous lymphocyte-associated antigen (CLA). CLA expression by these cells correlated with their ability to traffic to the skin during dengue infection. Notably, CLA was not up-regulated in bystander HCMV-restricted cells in these individuals, suggesting that the skin-targeted homing is pathogen-specific. These data support a role for skin-directed immunosurveillance against dengue and reinforce a skin-targeted vaccine strategy for this virus.

Abstract

Dengue, which is the most prevalent mosquito-borne viral disease afflicting human populations, causes a spectrum of clinical symptoms that include fever, muscle and joint pain, maculopapular skin rash, and hemorrhagic manifestations. Patients infected with dengue develop a broad antigen-specific T lymphocyte response, but the phenotype and functional properties of these cells are only partially understood. We show that natural infection induces dengue-specific CD8+ T lymphocytes that are highly activated and proliferating, exhibit antiviral effector functions, and express CXCR3, CCR5, and the skin-homing marker cutaneous lymphocyte-associated antigen (CLA). In the same patients, bystander human cytomegalovirus –specific CD8+ T cells are also activated during acute dengue infection but do not express the same tissue-homing phenotype. We show that CLA expression by circulating dengue-specific CD4+ and CD8+ T cells correlates with their in vivo ability to traffic to the skin during dengue infection. The juxtaposition of dengue-specific T cells with virus-permissive cell types at sites of possible dengue exposure represents a previously uncharacterized form of immune surveillance for this virus. These findings suggest that vaccination strategies may need to induce dengue-specific T cells with similar homing properties to provide durable protection against dengue viruses.

INTRODUCTION

Dengue virus (DENV) is a mosquito-borne flavivirus estimated to cause up to 390 million infections per year, of which 100 million cases are symptomatic (1). Infection causes a wide spectrum of clinical manifestations, from a self-limiting febrile illness termed dengue fever (DF) to more severe life-threatening dengue hemorrhagic fever (DHF) and dengue shock syndromes (DSSs) that are characterized by plasma leakage and hemorrhagic manifestations. A broad T cell response targeting all viral proteins is elicited during natural dengue infection, with the nonstructural proteins NS3 and NS5 being the major targets (2, 3). The role and functionality of virus-specific T cells during acute dengue infection are still not fully understood (4). In acute pediatric dengue infections, virus-specific CD8+ T cells were found to be highly activated, apoptotic, and to have impaired recognition of the serotype of the secondary infecting virus (5). A study from the same group reported a correlation between the magnitude of the CD8+ T cell response and disease severity (3), but this was not observed in other studies (6, 7). However, an increasing number of studies in both mice (8, 9) and humans (6, 10) are providing evidence of a protective role of dengue-specific CD4+ and CD8+ T cells from severe disease. A recent study reported that dengue-specific CD8+ T cells restricted in human leukocyte antigen (HLA) molecules that are associated with decreased disease susceptibility display higher magnitude and polyfunctional responses, suggesting that a vigorous CD8+ T cell response is associated with protection from severe disease (10). The strong evidence of the protective role of T cells during dengue infection suggests that a dengue vaccine should also target the T cell component of the antiviral response to provide durable protection across all dengue serotypes. Thus, a detailed understanding of the differentiation and specificity of dengue-specific lymphocytes that underlies natural dengue immunity is needed.

The protective capacity of virus-specific lymphocytes depends largely on the route of immunization (11). It is known that the context in which T cells initially encounter their cognate antigen determines the coordinated expression of adhesion molecules and chemokine receptors, which provide specific anatomical address codes for the responding cells that will govern their tissue localization. Hence, tissue-derived antigen-presenting cells within skin or gut imprint the expression of the E-selectin ligand, cutaneous lymphocyte-associated antigen (CLA), or the integrin α4β7 [binding to mucosal addressin cell adhesion molecule–1 (MADCAM-1)] and CCR9 (binding to CCL25), which are characteristic of skin- or gut-homing T cells, respectively (12, 13).

During the blood meal of an infected mosquito, DENV enters through the skin, infects local dendritic cells (14, 15), and travels through the bloodstream to the draining lymph nodes where it activates virus-specific cells. The cell types and organs hosting DENV replication and the roles played by the virus and/or by the host immune response in driving organ pathology have still not been fully elucidated. Autopsy reports from patients deceased after acute dengue infection have described the presence of dengue viral RNA in hepatocytes as well as in macrophages in the skin, spleen, and lymph nodes (16). DENV was detected in the skin and lymph nodes of dengue-infected rhesus monkeys (17) in human cadaveric skin explants (18) and in skin biopsies from areas of skin rash in an individual receiving a live attenuated experimental dengue vaccine (19). Accordingly, skin manifestations, gastrointestinal complications (nausea, vomiting, and diarrhea), and liver enlargement accompanied by elevated blood levels of alanine aminotransferase (an enzyme found in the liver) are commonly observed during acute dengue (20). Skin manifestations include a maculopapular rash that occurs in a large proportion of dengue patients (50 to 82%) on the third or fourth day from fever onset and/or petechiae, a mild hemorrhagic skin manifestation observed toward the end of the febrile period (21). Petechiae is commonly seen in patients with DHF and DSS but more rarely in patients with DF.

Here, we show that during acute dengue infection, DENV-specific CD8+ T cells are activated, proliferate, display antiviral effector functions, and express the skin-homing marker CLA. Expression of CLA by circulating DENV-specific T cells correlates with their capacity to traffic to the skin, because virus-specific T cells were enriched in the skin as compared to the peripheral blood of dengue patients. The enrichment of DENV-specific T cells at sites of possible virus entry highlights a potential role of these cells in immune surveillance during subsequent dengue infections and suggests that vaccination may need to elicit dengue-specific T cells with similar properties to provide durable protection against DENVs.

RESULTS

Dengue-specific CD8+ T cells are activated, proliferate, and display antiviral effector function during acute dengue

The activation status and functionality of virus-specific CD8+ T cells were assessed in a cohort of adult patients during acute, postfebrile, and convalescent phases of dengue infection (details listed in table S1). Dengue-specific CD8+ T cells from peripheral blood of HLA-A*1101+ dengue patients were analyzed by using HLA class I pentamers that bind T cells specific for the HLA-A*1101–restricted NS3 27 peptide (5). Because DENV 2 accounted for most of the infections in Singapore at the time of patient recruitment, we used the DENV 2 variant of the NS3 27 peptide (sequence: GTSGSPIIDK). We observed that the percentages of circulating NS3 27–specific CD8+ T cells vary greatly among patients (from 0.05 to 7% of CD8+ T cells) and generally peak during acute dengue (Fig. 1, A and B). During the acute phase of dengue, most NS3 27–specific T cells were highly activated (CD38+DR+) and proliferating (Ki67+), whereas activation levels progressively diminished and returned to the baseline levels observed during convalescence (Fig. 1, C to E). The activation and proliferation of pentamer CD8+ T cells followed similar kinetics but was more modest than that of pentamer+ cells. As described previously, most dengue patients from our cohort are undergoing secondary infections, and no significant differences in terms of dengue-specific T cell responses were observed between primary and secondary infections (2).

Fig. 1. Dengue-specific CD8+ T cells are activated, proliferating, and produce antiviral effector molecules during acute dengue.

(A) Dengue-specific CD8+ T cell populations targeting the HLA-A*1101–restricted DENV NS3 27 epitope were identified ex vivo from patient PBMCs by using a specific HLA-A*1101 pentamer. Plots are gated on live CD3+ cells. (B) Percentages of pentamer+ cells in the total CD8+ T cell population are summarized. Each patient is represented by a distinct symbol. (C to E) Ex vivo expression of the activation and proliferation markers CD38, HLA-DR, and Ki67 was assessed by ICS. Plots are gated on live CD3+ CD8+ pentamer+ or pentamer cells. Results from one representative patient (C) or seven patients (D and E) are shown. (F and G) Production of IFN-γ, TNF-α, and CD107a by pentamer+ cells was assessed by ICS after stimulating PBMCs ex vivo with the NS3 27 peptide. Plots for one representative donor (F) and for seven patients (G) are shown. Statistics were calculated using a Kruskal-Wallis test, followed by a nonparametric Mann-Whitney test. *P < 0.05, **P < 0.01, ***P < 0.001. A, PF, and C indicate acute, postfebrile, and convalescent time points, respectively. The gating strategy and isotype controls are shown in fig. S1.

To address the antiviral effector capacity of NS3 27–specific T cells, ex vivo isolated peripheral blood mononuclear cells (PBMCs) were stimulated with or without NS3 27 peptide, and production of interferon-γ (IFN-γ), tumor necrosis factor–α (TNF-α), and CD107a was assessed by intracellular cytokine staining (ICS). IFN-γ production and CD107a expression (an indirect measure of cytotoxicity) by pentamer+ cells in response to their specific peptide were similar across all time points (Fig. 1, F and G). Similar results were obtained by using HLA-A*1101 pentamers presenting the DENV 1 and DENV 3/4 variants of the antigenic peptide (fig. S2). In contrast, TNF-α production was low during acute dengue but increased during postfebrile and convalescence phases. Our data suggest that CD8+ T cells specific for the DENV NS3 27 peptide mount an efficient antiviral Tc1-type effector response, because they are activated, proliferating, and produce antiviral effector molecules.

Dengue-specific CD8+ T cells are highly differentiated Tc1 effector cells with skin-homing potential during acute dengue

The combination of chemokine receptors and tissue-homing adhesion molecules expressed by antigen-specific T cells is imprinted during their initial encounter with antigen and provides clues as to their migratory capacity (12). We characterized the phenotype of dengue-specific CD8+ T cells in the peripheral blood of HLA-A*1101+ dengue patients by using NS3 27 pentamers and antibodies directed against a panel of phenotypic/homing T cell markers. Expression profiles of NS3 27 pentamer+ and pentamer cells are shown for one representative patient in Fig. 2 (A and B) and for seven patients in Fig. 3 (and table S2). During acute dengue, we observed that most NS3 27–specific peripheral blood CD8+ T cells express CLA, which is associated with homing to the skin (22). CLA expression was highest in pentamer+ cells during acute dengue as compared to pentamer and pentamer+ cells from subsequent time points (Fig. 3A). In contrast, the gut-homing markers CCR9 and α4β7 were not expressed at all or only at lower levels (Fig. 3, C and D). During acute DENV, specific effector T cells expressed CXCR3 and CCR5 (Fig. 3, E and F), which typically support T cell homing to peripheral tissues during inflammation (23). Expression of CXCR6 was also increased in pentamer+ cells during acute dengue (Fig. 3B). Expression of this marker was detected on a subset of skin-homing T cells in humans (24), as well as on T cells homing to the inflamed liver (25). Combined expression of CCR7 and CD45RA showed that cells specific for the NS3 27 peptide are highly differentiated effector or memory cells of the T effector memory (TEM) or T effector memory RA reexpressing subsets (TEMRA) (Fig. 3, G and H). In conclusion, we show that circulating DENV-specific CD8+ T cells during acute dengue are highly differentiated CXCR3+CCR5+ cells endowed with a skin-homing phenotype.

Fig. 2. Phenotypic characterization of tissue-homing, inflammatory, and memory markers expressed by dengue NS3 27–specific CD8+ T cells.

The ex vivo expression of tissue-homing, inflammatory, and memory markers by NS3 27–specific pentamer+ cells was assessed in PBMCs from dengue patients during three time points of disease. Results from one representative dengue patient are shown. (A and B) Plots are gated on live CD3+ CD8+ cells (A) and NS3 27 pentamer+ or pentamer cells (B, left and right panels, respectively). The gating strategy and isotype controls are shown in fig. S3.

Fig. 3. Dengue NS3 27–specific CD8+ T cells are highly differentiated Tc1 cells endowed with a skin-homing phenotype during acute dengue.

(A to H) The ex vivo expression of tissue-homing, inflammatory, and memory markers by dengue NS3 27–specific pentamer+ and pentamer cells is shown for seven patients. Percentages are calculated on CD3+ CD8+ pentamer+ or pentamer cells. Statistics were calculated using a Kruskal-Wallis test, followed by a nonparametric Mann-Whitney test. (G and H) Percentage of pentamer+ or pentamer cells within the different subsets identified by expression of CCR7 and CD45RA. Naïve: CCR7+CD45RA+; T central memory (TCM): CCR7+CD45RA; T effector memory (TEM): CCR7CD45RA; and T effector memory RA reexpressing (TEMRA): CCR7CD45RA+. Statistics were calculated between each subset and the TCM subset from the same time point. “Δ” indicates P < 0.05 and is calculated between pentamer+ and pentamer TEMRA cells from the postfebrile phase.

Human cytomegalovirus–specific CD8+ T cells are activated during acute dengue but lack expression of the skin-homing marker CLA

Because the skin-homing phenotype of NS3 27 pentamer+ cells during acute dengue was very striking, we chose to further characterize the specificity of CLA expression and determine whether it correlated with homing to the skin. First, we addressed whether the skin-homing marker CLA is induced upon activation on all CD8+ T cells irrespective of their antigen specificity. To do this, we analyzed CLA expression on activated CD8+ T cells specific for an antigen not derived from DENV in the same dengue patient cohort. A previous report showed that during acute dengue infection, there is bystander activation of CD8+ T cells specific for the human cytomegalovirus (HCMV) (26). We thus analyzed the activation status and CLA expression level of HCMV-specific CD8+ T cells from the peripheral blood of dengue patients by using pentamers specific for the HLA-A*0201–restricted NLV peptide from HCMV (sequence: NLVPMVATV). Plots for one representative patient (Fig. 4A) and for six patients are shown (Fig. 4, B and C, and table S3). CD38 and HLA-DR expression were increased on HCMV-specific CD8+ T cells during acute dengue as compared to the subsequent time points. In contrast, CLA expression on NLV-specific CD8+ T cells was low (~6% of cells) and similar across all time points, suggesting that CLA expression is not simply activation-induced.

Fig. 4. HCMV-specific CD8+ T cells targeting the HLA-A*0201–restricted NLV peptide are activated during acute dengue but do not express the skin-homing marker CLA.

(A) CD8+ T cells specific for the NLV peptide were identified ex vivo from PBMCs of dengue patients during three time points of disease by using a specific HLA-A*0201 pentamer. (B and C) The percentage of CD38+HLA-DR+ and CLA+ cells was assessed on pentamer+ cells. Plots for one representative patient (A) and results for six patients are shown (B and C). Plots are gated on CD3+ (top panel) and CD3+, CD8+, and pentamer+ cells (bottom panels). Percentages are calculated on CD3+ CD8+ pentamer+ cells. (D) CLA expression on HCMV pentamer+ cells and DENV NS3 27 pentamer+ cells in an HLA-A*1101+/A*0201+ patient. Plots are gated on live CD3+ CD8+ pentamer cells (top panels), DENV NS3 27 pentamer+ cells (middle panels), or HCMV pentamer+ cells (bottom panels). (E) CLA expression on DENV NS3 27 pentamer+, HCMV NLV pentamer+, or CD8+ pentamer cells is summarized for six patients. Statistics were calculated using a Kruskal-Wallis test, followed by a nonparametric Mann-Whitney test. The gating strategy and isotype controls are shown in fig. S4.

CLA expression on DENV- and HCMV-specific CD8+ T cells was directly compared in an HLA-A*1101+/A*0201+ dengue patient. During acute dengue, CLA expression was increased in DENV-specific but not in HCMV-specific CD8+ T cells (Fig. 4D). Pentamer CD8+ T cells displaying an activated (CD38+HLA-DR+) or highly activated (CD38hiHLA-DRhi) phenotype were enriched in CLA+ cells, suggesting that a proportion of these may be dengue-specific (Fig. 4, D and E). However, the decreased percentage of CLA+ cells within activated pentamer cells compared to that of DENV pentamer+ cells further confirms the presence of at least some bystander T cell activation.

In conclusion, our results show that during acute dengue infection, CLA expression is induced in DENV-specific but not in activated HCMV-specific T cells. This suggests that CLA is not induced simply as a consequence of T cell activation but that additional signals are required for its expression.

The skin-homing receptor CLA is expressed by dengue-specific CD4+ and CD8+ T cells in patients affected by DF and DHF

To investigate whether the skin-homing phenotype was limited to a particular epitope-specific population (specific for NS3 27) or if it also held true for CD8+ T cells specific for other dengue epitopes, we characterized the skin-homing potential of T cells specific for NS3 and NS5. Ex vivo isolated PBMCs from acute dengue patients were stimulated with or without a mixture of NS3/NS5 peptide pools, and responding T cells were stained for expression of the skin- and gut-homing receptors CLA and CCR9. Plots from one representative patient are shown in Fig. 5A. CLA and CCR9 expression are shown for cells that did not respond to NS3/NS5 (cytokine) and for those that produced IFN-γ and TNF-α after stimulation with the DENV peptides (cytokine+). CLA expression on dengue-specific CD4+ and CD8+ T cells was respectively three- or sixfold higher than that on the non–dengue-specific counterpart (Fig. 5B). In the population that did not show cytokine responses, the frequencies of CLA+ cells among total CD4+ and CD8+ T cells were similar to those reported for healthy individuals (~20 and ~11%, respectively) (27). In contrast, CLA+ cells represented about 60% of total dengue peptide–responding CD4+ and CD8+ T cells. Expression of the gut-homing receptor CCR9 did not differ between dengue-specific and nonspecific T cells (Fig. 5, A and C).

Fig. 5. Dengue-specific CD4+ and CD8+ T cells targeting NS3 and NS5 peptides express the skin-homing receptor CLA during acute dengue.

(A and B) Total PBMCs from acute dengue patients were stimulated with or without NS3 and NS5 peptides and analyzed for production of IFN-γ and TNF-α and for expression of CLA and CCR9. Cells that did not respond to NS3/NS5 peptides are indicated as “cytokine” or non–DENV-specific, and those that did respond are indicated as “cytokine+” cells or DENV-specific. (A) Results from one representative patient. CLA (B) and CCR9 (C) expression on DENV-specific or non–DENV-specific CD4+ (empty circle) or CD8+ (filled circle) T cells is shown for 44 patients. (D and E) Patients from (B) were classified according to disease severity in DF with (WS) or without (NWS) warning signs or DHF cases (DF NWS, n = 23; DF WS, n = 16; and DHF, n = 5), and percentages of dengue-specific or non–dengue-specific CD4+ or CD8+ T cells expressing CLA were analyzed. w/o, without peptide. Statistics were calculated using a Kruskal-Wallis test, followed by a nonparametric Mann-Whitney test. The gating strategy and isotype controls are shown in fig. S5.

Next, we addressed whether T cell expression of CLA could be observed in patients with different disease severities. The patients from Fig. 5B were classified on the basis of the 2009 World Health Organization (WHO) classification into patients affected by DF, DHF, or its “warning signs” (table S4) (20). CLA expression was similarly increased in dengue-specific CD4+ and CD8+ T cells of patients with DF with or without warning signs or DHF, as compared to the non–dengue-specific counterpart (Fig. 4, D and E). The total frequencies of antigen-specific CD4+ and CD8+ T cell populations also did not correlate with disease severity (fig. S6).

Dengue-specific T cells are present in the skin of dengue patients

CLA expression by dengue-specific T cells suggests that either these cells are primed in the skin or they encounter their specific antigen in the context of skin-derived dendritic cells in the draining lymph nodes. We thus investigated whether dengue-specific T cells could be detected in the skin of patients during acute dengue infection and at convalescence. Skin blisters were raised on the forearm of two healthy donors, four acute patients, and one convalescent dengue patient (table S5). Blisters were formed by applying a prolonged negative pressure with a clinical-grade suction pump attached to a suction cup placed on the skin of the patients, as described previously (28). The prolonged negative pressure causes the splitting of the epidermis from the dermis at the lamina lucida level and the formation of a unilocular blister. The blister was protected overnight, and the following day, the fluid contained inside the skin blister was aspirated and microcentrifuged to pellet the cellular contents. Cells collected from the skin blisters or from peripheral blood of each patient were plated at equal numbers in the presence or absence of NS3/NS5 peptides and tested for their ability to produce IFN-γ by enzyme-linked immunospot (ELISPOT) assay. Results from two healthy donors and one representative acute dengue patient are shown in Fig. 6A. Virus-specific T cells were significantly enriched in skin blister samples as compared to the peripheral blood counterpart of all five dengue patients tested (P = 0.0078). In contrast, dengue-specific T cells were undetectable in the skin and in the peripheral blood of healthy donors (Fig. 6, A and B). Our results demonstrate that during acute dengue, CLA expression by circulating dengue-specific T cells correlates with their in vivo ability to traffic to the skin.

Fig. 6. Dengue-specific T cells are present in the skin of dengue patients.

IFN-γ production was measured by ELISPOT in cells isolated from skin blister fluid (skin) or from peripheral blood (PBMCs) of healthy controls and dengue patients. Equal number of cells was incubated overnight with or without NS3/NS5 peptide pools or with anti-CD3/28 antibodies. (A) Results for two healthy donors and one representative acute dengue patient. The number of spot-forming cells (SFC) detected in each well is indicated. (B) Data from four acute patients (dengue 1 to 4) and one convalescent (dengue 5) dengue patient are summarized. The number of SFC relative to 105 cells is indicated. “//” indicates that the number of SFC was too high to be counted (>600). Statistics were calculated by using a nonparametric Wilcoxon matched-pairs signed rank test.

DISCUSSION

The phenotype, functionality, and tissue-homing properties of virus-specific T lymphocytes during dengue infection have been poorly characterized. Here, we show that circulating dengue-specific CD4+ and CD8+ T cells express the skin-homing molecule CLA and are able to home to the skin during acute and convalescent dengue infection. Virus-specific T cells isolated from the skin of dengue patients produced IFN-γ after peptide stimulation, suggesting that these cells are capable of exerting antiviral functions in vivo.

In humans, 98% of CLA+ T cells reside in the skin under resting conditions (24). CLA expression has been shown in both mouse and human studies to specifically drive T cell homing to the skin (12, 29) and to be induced during antigenic priming by antigen-presenting cells in cutaneous lymphoid organs or in the skin itself (12, 30). Recent studies showed that dendritic cells in human skin support dengue infection (14). In dengue-infected rhesus monkeys, DENV was present systemically in the skin only after the termination of viremia (from day 6 after infection) in primary-infected animals, whereas it was detectable at earlier time points during secondary infection (17). In humans infected with DENV, maculopapular rashes appear just before the termination of viremia (day 3/4 from fever onset, which coincides with days 6 to 8 after infection). In other viral diseases, maculopapular rashes were shown to contain viral particles (31). The presence of a maculopapular rash was found to be associated with decreased probability of developing severe disease in a child cohort, suggesting that maculopapular rashes may also correlate with the presence of an antiviral immune response (32). In the current study, dengue-specific T cells were detected systemically in the skin of four acute patients and one convalescent dengue patient, and the frequencies at which these cells were present were higher than that of their peripheral blood counterpart. All four acute patients had experienced a skin rash, but this had disappeared by the time the skin blister was performed. The small sample size dictated by the difficulty of recruiting dengue patient volunteers for the skin blisters represents a limitation of this study. A higher number of cases would allow to investigate whether a correlation exists between the presence of dengue-specific T cells in the skin, a maculopapular rash, and disease severity.

This study gives evidence of the systemic distribution of an antigen-specific T cell population in the skin during an acute viral infection. Previous studies in humans have shown the preferential accumulation of antigen-specific T cells around the site of infection by skin-tropic viruses such as in lesions caused by herpes simplex virus 2 infection (33, 34) or around the site of purified protein derivative injection in Bacille Calmette-Guérin–vaccinated individuals (35). The relationship between T cells in the skin and blood in humans is currently unclear. In mice, two populations of antigen-specific T cells have been described in the skin: circulating antigen-specific cells that traffic between the skin and blood and skin-resident cells that are preferentially detected at the injection site but can also be found systemically. Skin-resident cells are unable to exit the skin where they represent an important first line of defense during reinfection (36).

It is unclear why CLA+ dengue-specific T cells were absent or present in lower numbers in the blood of postfebrile and convalescent dengue patients. It is possible that these cells are retained in the skin for immune surveillance upon reinfection or, alternatively, DENV-specific CLA+ cells could lose CLA expression after reencounter of their specific antigen in a different context. This has been shown to occur in mice where T cell homing properties can be reprogrammed during subsequent antigenic encounters in different environments (30). The preliminary finding that dengue-specific T cells can be detected in the skin of a convalescent patient may favor the first hypothesis.

In contrast to previous studies in a Thai child cohort (5), we do not observe functional impairment of dengue-specific T cells during acute dengue, at least on the basis of their in vivo proliferative capacity and ex vivo capacity to produce IFN-γ and express CD107a after peptide stimulation. The discrepancies between the two studies may reflect the fundamental differences in the characteristics of the patient cohorts used for the two studies. First, the Thai study involves child cohorts infected with DENV 1 as compared to adult cohorts infected mainly with DENV 2 in our study. The ethnicities of the two cohorts are also different, and this may result in differences in terms of immunodominance of the NS3 27 epitope, which influences the extent to which NS3 27–specific T cell populations will be functionally impaired in vivo (37).

Previous studies suggested that most activated T cells during acute dengue are dengue-specific and that there is minimal contribution of bystander cells to the T cell response. This was based on the observation that during acute dengue, flu-specific CD8+ T cells were not activated (7). However, Sandalova et al. showed that during acute viral infections (hepatitis B virus and dengue), CD8+ T cells specific for persistent herpes viruses such as HCMV were activated and proliferating, whereas those specific for an unrelated nonpersistent virus such as flu were not (26). Here, we provide further evidence of the presence of bystander T cell activation during acute dengue. The absence of CLA expression by HCMV-specific cells demonstrates that this tissue-homing marker is not induced simply as a consequence of T cell activation and suggests that additional yet unknown factors are required for its up-regulation. The cytokine environment may play a role, because it was shown that CLA could be up-regulated in vitro by transforming growth factor–β and IL-12, whereas IL-4 was found to down-regulate it (29).

In conclusion, here, we show that natural dengue infection induces skin-tropic DENV-specific T cells that may be involved in immune surveillance upon subsequent infections. Because the protective efficacy of virus-specific T cells is also determined by their capacity to home to the site of infection, these findings need to be taken into account when designing new vaccines.

MATERIALS AND METHODS

Study design

The patients were recruited as part of two studies, one designed to characterize the host response to dengue infection in clinically well-defined dengue-infected individuals (tables S1 to S4). This study recruits consenting individuals of at least 21 years of age with clinically confirmed dengue status. Virus-specific T cells identified in these individuals were used in the experiments described in Figs. 1 to 5 to support the aim of the study. The other study was designed to understand the role of CD8+ T cells in the pathogenesis of dengue, with a focus on the development of dermatological manifestations (table S5). This study aims to recruit consenting healthy and clinically confirmed dengue-infected individuals of at least 21 years of age. Individuals from this study were used in Fig. 6 to understand the role of dengue-specific T cells in the skin of the dengue patients. Both studies are of an exploratory nature for which no previous data exist for any form of sample size justifications. No blinding or randomization was performed.

Ethics statement

Human peripheral blood and skin blister samples were obtained after written informed consent from all participants. The study was conducted in accordance with the Declaration of Helsinki and approved by the Singapore National Healthcare Group ethical review board (DSRB 2008/00293 and DSRB 2013/00209). The details of the dengue patients recruited for this study are summarized in tables S1 to S6.

Dengue diagnosis

Dengue diagnosis was confirmed on the basis of either detection of DENV RNA by reverse transcription polymerase chain reaction or of NS1 antigen by enzyme-linked immunosorbent assay (Bio-Rad). Some patients with positive immunoglobulin M (IgM) and IgG acute serologies (Panbio Dengue Duo Cassette) were also included if they fulfilled the WHO (20) criteria for probable DENV, as described previously (2). Healthy individuals with no clinical history of DENV infection were recruited for the skin study from laboratory staff.

PBMC isolation

Blood samples were collected from patients with confirmed DENV diagnosis in EDTA vacutainer tubes at three different time points from fever onset. PBMCs were isolated from peripheral blood by Ficoll gradient purification and cryopreserved. Cells were thawed on the day of the experiment and were used directly for the ex vivo assays.

Suction blister induction

With skin suction chambers (Medical Engineering, Royal Free Hospital) and a clinical suction pump, suction blisters were raised on the forearm of dengue patients, as described previously (28). A negative pressure of 25 to 40 kPa (200 to 300 mmHg) below atmospheric pressure was applied to the skin for 2 to 4 hours until a unilocular blister was formed. The blister was then protected overnight with a rigid adhesive dressing, and 18 to 24 hours later, the fluid inside the blister was aspirated and microcentrifuged at 650g (3000 rpm) for 4 min. The supernatant was removed and stored at −80°C. The cellular pellet was resuspended in 500 μl of AIM-V supplemented with 2% AB human serum until analyzed.

Synthetic peptides

Peptides from the NS3 and NS5 proteins were designed on the basis of the sequence of DENV 2 virus (strain D2/SG/05K4155DK1/2005) and were purchased from Mimotopes. The peptides consist of 300 15-mer peptides overlapping by 10 amino acids, 122 peptides for the NS3 protein and 178 for the NS5 protein. The purity of the peptides was above 80%, and their composition was confirmed by mass spectrometry analysis. All peptides were dissolved in dimethyl sulfoxide (DMSO) at a concentration of 40 mg/ml, and intermediate working dilutions were performed in RPMI supplemented with penicillin (200 U/ml) and streptomycin (200 μg/ml).

Flow cytometric analysis

Staining was performed on thawed, ex vivo isolated PBMCs from HLA-typed patients. Cells were stained on ice for 30 min with phycoerythin (PE)–labeled HLA-A*1101–NS3 27 or HLA-A*0201–NLV pentamers (ProImmune). NS3 27 pentamers from DENV 2 were used for experiments in Fig. 1, whereas a mixture of pentamers for the DENV 1, 2, and 3/4 variants was used for experiments in Figs. 3 and 4. Cells were washed and subsequently stained at 4°C for 20 min with a combination of the following surface antibodies: anti-CD3 V500, anti-CD8 PerCP (peridinin chlorophyll protein), anti–HLA-DR PE-Cy7, anti-CD38 APC (allophycocyanin) and anti-CLA Pacific Blue (BioLegend) or anti-CD3 PC7 (Beckman Coulter), anti-CD8 Pacific Blue, anti-CXCR3 PerCP Cy 5.5 (BioLegend), anti-CXCR6 Alexa Fluor 647 (BioLegend), anti-CCR5 fluorescein isothiocyanate (FITC), anti-CLA FITC (BioLegend), anti-CCR9 Alexa Fluor 647 (BioLegend), anti-β7 FITC (eBioscience), anti-α4 biotin (eBioscience) followed by a Qdot 800 streptavidin conjugate (Invitrogen), anti-CCR7 (R&D) followed by a secondary goat anti-mouse IgG2a APC antibody (Jackson ImmunoResearch), and anti-CD45RA biotin antibody followed by a Qdot 800 streptavidin conjugate.

For Ki67 staining, cells were fixed and permeabilized using the Cytofix/Cytoperm kit (BD Pharmingen) and stained intracellularly with anti-Ki67 FITC antibodies on ice for 30 min. Samples were acquired on a Fortessa FACScan. All antibodies were purchased from BD Pharmingen unless otherwise stated.

Intracellular cytokine staining

PBMCs were thawed the day before performing the experiment, resuspended in AIM-V with 2% AB human serum, and incubated overnight at 37°C. The following day, cells were stimulated or not with overlapping peptide pools derived from the DENV NS3 and NS5 proteins (at a final concentration of 1 μg/ml) for 6 hours at 37°C in the presence of monensin (used according to the manufacturer’s indication, BioLegend) and of brefeldin A (5 μg/ml) (Sigma-Aldrich), added for the last 4 hours of stimulation. Cells were then washed and stained with the NIR viability stain (Invitrogen), followed by anti-CD3 PC7, anti-CD8 PerCP, anti-CLA FITC, and anti-CCR9 APC (BioLegend) antibodies. ICS was performed with anti–IFN-γ V450 and anti–TNF-α PE antibodies.

For detection of cytokine production by NS3 27 pentamer+ cells, thawed ex vivo isolated PBMCs were either left unstimulated or stimulated with the NS3 27 peptide (1 μg/ml) for 5 hours at 37°C in the presence of brefeldin A (10 μg/ml). To assess degranulation, CD107a FITC antibody was added to the cells at the beginning of the stimulation. After the stimulation, cells were stained with the HLA-A*1101–NS3 27 pentamers and then surface-stained with anti-CD3 PC7 and anti-CD8 PerCP antibodies. ICS was performed with anti–IFN-γ V450 and TNF-α APC. Samples were acquired on a Fortessa FACScan. Antibodies were purchased from BD Pharmingen unless otherwise stated.

IFN-γ ELISPOT assay

ELISPOT assays for the detection of IFN-γ–producing cells were performed as described previously (2). An equal number of cells recovered from the skin blisters or from the peripheral blood of dengue patients/healthy donors was incubated overnight with or without pooled peptides from the NS3 and NS5 proteins (1 μg/ml) or with anti-CD3/28 beads (Invitrogen) as a positive control. For the negative controls, DMSO was added to media at the same concentration present in the peptide pools.

Statistical analyses

All statistics were calculated with GraphPad Prism version 5 and conformed to the journal policy. For Figs. 1 and 3 to 5, statistics were calculated by using a Kruskal-Wallis test, followed by a nonparametric Mann-Whitney test, whereas for Fig. 6, a nonparametric Wilcoxon matched-pairs signed rank test was used. Exact two-tailed P values are listed in table S7. *P < 0.05, **P < 0.01, ***P < 0.001.

SUPPLEMENTARY MATERIALS

www.sciencetranslationalmedicine.org/cgi/content/full/7/278/278ra35/DC1

Fig. S1. Gating strategy and isotype controls for Fig. 1.

Fig. S2. Cytokine production of CD8+ T cells specific for the HLA-A*1101–restricted NS3 27 variants from DENV serotypes 1 and 3/4.

Fig. S3. Gating strategy and isotype controls for Fig. 2.

Fig. S4. Gating strategy and isotype controls for Fig. 4.

Fig. S5. Gating strategy and isotype controls for Fig. 5.

Fig. S6. Frequencies of DENV-specific CD4+ and CD8+ T cells in the blood of dengue patients.

Table S1. Details of patients included in the experiments illustrated in Fig. 1.

Table S2. Details of patients included in the experiments illustrated in Fig. 3.

Table S3. Details of patients included in the experiments illustrated in Fig. 4.

Table S4. Details of patients included in the experiments illustrated in Fig. 5.

Tables S5. Details of patients/donors included in the skin blister study illustrated in Fig. 6.

Tables S6. Source data for all graphs with n < 20.

Tables S7. Exact P values for all graphs.

REFERENCES AND NOTES

Acknowledgments: We thank B. Lee from the Bioinformatics Platform at the Singapore Immunology Network, Agency for Science, Technology and Research, for his help with statistical analyses and the research nurses from Communicable Disease Centre, Tan Tock Seng Hospital, for their efforts in recruiting the dengue patients for this study. Funding: This work was supported by the National Research Foundation [National University of Singapore (NUS)–Hebrew University of Jerusalem–Campus for Research Excellence and Technological Enterprise: R182-005-172-281; NUS Office of the Deputy President (Research and Technology): R182-000-192-133 and National Medical Research Council (NMRC): STOP-Dengue Translational and Clinical Research grant R-182-003-220-275]. L.R. is currently supported by NMRC/STaR/013/2012) (to A.B.). Author contributions: L.R. and P.A.M. conceived and designed the study and wrote the paper; P.A.M. and D.M.K. provided funding; L.R. and E.A.K. performed the experiments and analyzed the data; C.T.T. provided all technical support; T.-L.T., V.C.H.G., D.C.L., and Y.S.L. coordinated the clinical aspects of the study; A.N.A. provided expertise and helped set up the skin blister study; B.J.H. provided HLA typing of the patients; A.B. contributed to setting up of the study and provided funding for L.R.; and D.M.K., N.R.J.G., and A.W.-S. contributed to the writing of the manuscript. Competing interests: A.N.A. has served as a consultant for GlaxoSmithKline and Ono Pharmaceuticals. The other authors have no conflicting financial interests.
View Abstract

Stay Connected to Science Translational Medicine

Navigate This Article