Regulatory T‐cells in acute dengue viral infection

Summary

Although regulatory T‐cells (T_regs) have been shown to be expanded in acute dengue, their role in pathogenesis and their relationship to clinical disease severity and extent of viraemia have not been fully evaluated. The frequency of T_regs was assessed in 56 adult patients with acute dengue by determining the proportion of forkhead box protein 3 (FoxP3) expressing CD4⁺ CD25⁺T‐cells (FoxP3⁺ cells). Dengue virus (DENV) viral loads were measured by quantitative real‐time polymerase chain reaction (PCR) and DENV‐specific T‐cell responses were measured by ex‐vivo interferon (IFN)‐γ enzyme‐linked immunospot (ELISPOT) assays to overlapping peptide pools of DENV‐NS3, NS1 and NS5. CD45RA and CCR4 were used to phenotype different subsets of T‐cells and their suppressive potential was assessed by their expression of cytotoxic T lymphocyte‐antigen 4 (CTLA‐4) and Fas. While the frequency of FoxP3⁺ cells in patients was significantly higher (P < 0·0001) when compared to healthy individuals, they did not show any relationship with clinical disease severity or the degree of viraemia. The frequency of FoxP3⁺ cells did not correlate with either ex‐vivo IFN‐γ DENV‐NS3‐, NS5‐ or NS1‐specific T‐cell responses. FoxP3⁺ cells of patients with acute dengue were predominantly CD45RA⁺ FoxP3^low, followed by CD45RA‐FoxP3^low, with only a small proportion of FoxP3⁺ cells being of the highly suppressive effector Treg subtype. Expression of CCR4 was also low in the majority of T‐cells, with only CCR4 only being expressed at high levels in the effector T_reg population. Therefore, although FoxP3⁺ cells are expanded in acute dengue, they predominantly consist of naive T_regs, with poor suppressive capacity.

Abbreviations

DENV

dengue virus

DF

dengue fever

DHF

dengue haemorrhagic fever

nT_regs

natural thymic‐derived T_regs

T_regs

regulatory T‐cells

Introduction

Dengue viral infections are one of the most rapidly emerging mosquito‐borne viral infections in the world. Approximately 3·97 billion individuals are at risk of infection, and the virus infects aproximately 390 million individuals annually.1 Seventy per cent of these infections occur in Asia,1 in resource‐poor settings, which are hardly equipped to handle the increasing number of dengue infections each year. Although there is a now a licensed dengue vaccine, its efficacy for some serotypes appears to be poor and the World Health Organization (WHO) has not recommended its use in children < 9 years due to safety concerns.2 Therefore, in order to develop safer and more immunogenic vaccines it is essential have a more detailed understanding of the immune responses to the dengue virus (DENV).

Although very comprehensive analyses of T‐cell responses to the DENV in those with acute infection and in those with past infection have been performed,3, 4, 5, 6, 7, 8, 9, 10 there still appears to be a debate regarding the role of T‐cells in dengue. It has been proposed that DENV‐specific cross‐reactive T‐cells to previously infecting DENV serotypes contribute to severe clinical disease by the production of proinflammatory cytokines.8, 9, 11, 12 However, more recent studies have shown that the presence of highly polyfunctional T‐cells directed against protective human leucocyte antigen (HLA) alleles was associated with milder clinical disease.3, 4, 5, 13 In addition, no differences were found in the frequency and functionality of cross‐reactive T‐cells in those who were naturally infected with the DENV and had asymptomatic or severe clinical disease.6 In acute infection, the frequency of DENV‐specific T‐cells was found to be higher in those with milder clinical disease14 and also found to express cutaneous leucocyte antigen (CLA), which supports a protective role of T‐cells in dengue.15 Although DENV‐specific T‐cells are absent or present at very low frequency in peripheral blood in those with acute infection,13 T‐cells in general have been shown to be highly activated.16, 17

CD25^high CD4⁺ regulatory T‐cells (T_regs) were found to be expanded in those with acute dengue, and these cells were found to suppress the proliferative capacity of DENV‐specific T‐cells in vitro.18 Therefore, it was speculated that expansion of CD25^high CD4⁺ T_regs in acute dengue could counteract the immunopathology that occurred due to cross‐reactive T‐cells.18 In a more recent study, which characterized T_regs on the basis of the presence of forkhead box protein 3 (FoxP3), it was shown that T_regs were indeed expanded in acute dengue and were significantly more expanded in milder clinical disease.19 However, in this study it was observed that the frequency of T_regs was significantly higher in those experiencing a primary dengue infection when compared to those experiencing secondary dengue infections.19 As cross‐reactive DENV‐specific T‐cells are not present in those with a primary dengue infection, the expansion of T_regs is unlikely to suppress DENV‐specific cross‐reactive T‐cells. Also, given the emerging evidence which supports the protective role of DENV‐specific T‐cells in acute dengue, the role of T_regs in acute dengue should be investigated further.

Human T_regs are known to be a heterogeneous population, which differ in their suppressive and proliferative capacity.20, 21, 22 Naive T_regs, which express CD45RA, have shown to be a natural type of T_regs, with different proliferative capacity and tissue homing characteristics compared to CD45RO‐expressing T_regs.20 In addition, based on the expression of CD45RA and the expression intensity of FoxP3, T_regs can also be divided into activated T‐cells transiently expressing FoxP3 (CD45RA⁻ FoxP3^low) and thymic‐derived natural T_regs, which are CD45RA⁺ FoxP3^low, and the highly suppressive effector T_regs CD4⁺ FoxP3^high.23, 24 Therefore, in order to characterize the role of FoxP3‐expressing T‐cells in acute dengue, it would be important to determine the phenotype of T_regs, and also to determine if the expansion of FoxP3‐expressing T‐cells was due to the expansion of natural or highly suppressive T_regs or if it is an expansion of activated effector T‐cells, which transiently express FoxP3 but lack any suppressive capabilities.

In this study we have investigated the relationship of the frequency of FoxP3‐expressing T‐cells with clinical disease severity, viraemia and DENV‐specific T‐cell responses in a cohort of adult patients with acute dengue. We then further characterized the phenotype of FoxP3‐expressing T‐cells that are expanded in patients with acute dengue, based on their expression of CD45RA and C‐C chemokine receptor type 4 (CCR4), and further evaluated their suppressive potential by investigation of expression of CTLA‐4 and Fas.

Methods

Fifty‐three adult patients [mean age 28·98, standard deviation (SD) ± 12·3] with acute confirmed dengue infection were recruited following informed written consent from the National Infectious Diseases Hospital in Sri Lanka. The first blood sample was collected during days 5–9 of illness and a convalescent blood sample was obtained from 10 patients 30–45 days following the onset of illness (the remaining patients did not attend for follow‐up, which is a common event in this setting with limited health‐care resources). The mean duration of illness at the time of obtaining the first sample was 6·8 (SD ± 1·4) days. Ultrasound scans were performed to determine the presence of fluid leakage in pleural and peritoneal cavities. Full blood counts and liver transaminase levels were performed serially throughout the illness. Clinical disease severity was classified according to the 2011 WHO dengue diagnostic criteria.25 Accordingly, patients with ultrasound scan evidence of plasma leakage (those who had pleural effusions or ascites) were classified as having dengue haemorrhagic fever (DHF). Shock was defined as having cold clammy skin, along with a narrowing of pulse pressure of ≤ 20 mmHg.

Based on the above criteria, 27 patients were classified as having DF (dengue fever) and 26 patients as having DHF. Twelve healthy individuals, who were seropositive for dengue and age‐matched, were also recruited from the staff of University of Sri Jayewardenepura as healthy controls for the analysis of regulatory T‐cells.

Ethics statement

The study was approved by the Ethical Review Committee of the University of Sri Jayewardenepura. All patients were recruited post written consent.

Serology

Dengue antibody assays were performed using a commercial capture immunoglobulin (Ig)M and IgG enzyme‐linked immunosorbent assay (ELISA) (Panbio, Brisbane, Australia).26, 27 Patients with only dengue‐specific IgM antibody was categorized as having a primary dengue infection, while patients with both dengue‐specific IgM and IgG were categorized under secondary dengue infection.25, 26

Flow cytometry for identification of regulatory T‐cells and phenotyping

Freshly isolated peripheral blood mononuclear cells (PBMCs) were used in all experiments, and staining for flow cytometry was performed immediately following PBMC separation. For identification of FoxP3‐expressing CD4⁺ T‐cells, PBMCs were washed and surface staining was carried out with anti‐CD25 phycoerythrin (PE), anti‐CD4 peridinin chlorophyll (PerCP) anti‐CD3 allophycocyanin (APC), all purchased from Biolegend (San Diego, California). The cells were then permeabilized and stained with anti‐FoxP3 fluorescein isothiocyanate (FITC) in the FoxP3 staining buffer according to the manufacturer's instructions. The T_regs were identified following gating of initially CD3⁺ and CD4⁺ T‐cells and then gating on CD4⁺ T‐cells which expressed FoxP3. Cells were acquired on a Partec Cyflow Cube 6 and analysed with de‐novo FCS Express version 4.

In order to phenotype the T_regs and determine expression of CTLA‐4, we used anti‐CD4 Pacific blue, anti‐CD25 PE, anti‐FoxP3 FITC, anti‐CD45RA APC, anti‐CTLA‐4 APC, anti‐CD95 BV605 and anti‐CCR4 BV605, all purchased from Biolegend (San Diego, California). All FoxP3 staining was performed in FoxP3 staining buffer and cells were acquired on the Guava easy Cyte 12HT flow cytometer and analysed using de‐novo FCS Express version 4.

Qualitative and quantitative assessment of viral loads

The infecting DENV was serotyped and the viral loads quantified as previously described using a multiplex quantitative real‐time PCR.28 RNA was extracted from serum samples using QIAamp Viral RNA Mini Kit (Qiagen, Valencia, CA), according to the manufacturer's protocol. Multiplex quantitative real‐time PCR was performed as previously described using the CDC real‐time PCR assay for detection of the dengue virus,29 and modified to quantify the DENV. Oligonucleotide primers and a dual‐labelled probe for DEN 1, 2, 3 and 4 serotypes were used (Life Technologies, Delhi, India) based on published sequences.29 In order to quantify viruses, standard curves of DENV serotypes were generated as previously described by Fernando et al.28

Peptides

The peptide arrays spanning DENV NS1 (DENV‐2 Singapore/S275/1990, NS1 protein NR‐2751), NS3 (DENV‐3, Philippines/H87/1956, NS3 protein, NR‐2754) and NS5 proteins (DENV‐2, New Guinea C (NGC), NS5 protein, NR‐2746) were obtained from the NIH Biodefense and Emerging Infections Research Resource Repository (NIAID, NIH, Bethesda, MD). The DENV NS3 peptide array consisted of 105 4‐17 mers peptides. NS1 and NS5 proteins were comprised of 60 and 156 peptides, respectively. The peptides were reconstituted as previously described.30 NS1, NS3 and NS5 peptides were pooled to represent the DENV‐ NS1, NS3 and NS5 proteins separately, and all the peptides of these proteins were also pooled to represent a ‘DENV‐all’ pool of peptides.

Ex‐vivo ELISPOT assay

Ex‐vivo IFN‐γ ELISPOT assays were carried out as previously discussed using fresh PBMCs obtained from 74 patients and 11 healthy individuals.6 DENV‐NS3, NS1, NS5 and the DENV‐ALL peptide pool of overlapping peptides were added at a final concentration of 10 μm, as previously described.11, 31 All peptides were tested in duplicate. Phytohaemagglutinin (PHA) was always included as a positive control, and media alone with the PBMCs was included as a negative control. The spots were enumerated using an automated ELISPOT reader (AID GmbH, Strasberg, Germany). Background (cells plus media) was subtracted and data expressed as number of spot‐forming units (SFU) per 10⁶ PBMC.

Quantitative cytokine assays

Quantitative ELISAs for interleukin (IL)‐23 (Biolegend, San Diego, California), (IL‐17 (Biolegend), IL‐10 (Mabtech, Nacka Strand, Sweden), transforming growth factor (TGF)‐β (Mabtech, Nacka Strand) and IL‐2 (Mabtech, Nacka Strand) were performed in plasma according to the manufacturer's instructions.

Statistical analysis

prism version 6 was used for statistical analysis. As the data were not normally distributed, differences in means were compared using the Mann–Whitney U‐test (two‐tailed). Spearman's rank order correlation coefficient was used to evaluate the correlation between variables. To determine the differences in the proportion of FoxP3⁺ T‐cells in acute and convalescent samples, Wilcoxon's matched‐pairs signed‐rank test (two‐tailed) was used.

Results

The clinical and laboratory features are presented in Table 1. Five patients developed shock, defined as cold clammy skin, along with a narrowing of pulse pressure of ≤ 20 mmHg and were noted to have a reduced urine output. Mild bleeding manifestations such as epistaxis and gum bleeds were seen in four patients. FoxP3‐expressing T‐cells were identified by gating on CD4⁺ T‐cells and then on FoxP3‐ and CD25‐expressing T‐cells (Fig. 1a).

Table 1.

Clinical and laboratory features of those with dengue fever (DF) and dengue haemorrhagic fever (DHF)

Clinical finding	DF (n = 29)	DHF (n = 27)
Hepatomegaly	2 (6·9%)	6 (22·2%)
Pleural effusion or ascites	0	9 (33·3%)
Rise of haematocrit of > 20%	0	27 (100%)
Shock	0	5 (18·5%)
Bleeding manifestations	0	4 (14·8%)
Platelet counts
< 25 000	3 (10·3%)	20 (74·1%)
25 000–49 000	9 (31·0%)	3 (11·1%)
50 000–99 000	15 (51·7%)	4 (14·8%)
> 100 000	2 (6·90%)	0
Lowest lymphocyte count (cells/mm3)
< 750	20 (68·97%)	21 (77·78%)
750–1500	8 (27·59%)	4 (14·81%)
> 1500	1 (3·45%)	2 (7·41%)

Figure 1.

Frequency of forkhead box protein 3 (FoxP3)‐expressing CD4⁺ T‐cells. (a) PBMC were isolated from patients with acute dengue (n = 53) and healthy controls (n = 12) and expression of CD3, CD4, CD25 and FoxP3 were analysed by flow cytometry. Representative figure shown. (b) Peripheral blood mononuclear cells (PBMCs) were isolated from patients with acute dengue (n = 53) and healthy controls (n = 12) and expression of CD3, CD4, CD25 and FoxP3 were analysed by flow cytometry. Frequency of FoxP3⁺ cells in patients with dengue haemorrhagic fever (DHF) (n = 27), dengue fever (DF) (n = 26), healthy individuals (n = 12) and all patients with acute dengue (n = 53). (c) Frequency of FoxP3⁺ cells in patients with acute dengue (n = 9) during the acute phase of the illness and during the convalescent phase. The bars represent the median and interquartile range. *P < 0·05, ***P < 0·001.

Frequency of FoxP3⁺ T‐cells in patients with acute dengue infection

It was previously shown that T_regs were significantly expanded in patients with acute dengue infection and were believed to suppress DENV‐specific T‐cell responses.18 Therefore, we proceeded to determine the association of expansion of FoxP3‐expressing T‐cells with clinical disease severity and DENV‐specific T‐cell responses.

The frequency of FoxP3⁺ T‐cells was expressed as the proportion of FoxP3‐expressing CD4⁺ T‐cells in the CD4⁺ T‐cell population (% of FoxP3⁺ cells). The frequency of FoxP3⁺ cells in patients with acute dengue [median = 1·67% of FoxP3⁺ cells, interquartile range (IQR) = 1·1–2·5] was significantly higher (P < 0·0001) when compared to healthy individuals (median = 0·6% of FoxP3⁺ cells, IQR = 0·4–0·7). The frequency was significantly higher in patients with DHF (P < 0·0001) and in patients with DF (P < 0·0001) when compared to healthy individuals. Although the frequency of FoxP3⁺ cells was higher in those with DF (median = 1·96% of FoxP3⁺ cells, IQR = 1·3–2·8) when compared to those with DHF (median = 1·44% of FoxP3⁺ cells, IQR = 1·0–1·9), this was not significant (P = 0·11) (Fig. 1b). The frequencies of FoxP3⁺ cells were also similar (P = 0·87) in patients who developed shock (median = 1·67% of FoxP3⁺ cells, IQR = 0·8–2·35) when compared to those who did not (median = 1·41% of FoxP3⁺ cells, IQR = 0·97–1·96). The frequency of FoxP3⁺ cells did not correlate with the platelet counts (Spearman's r = 0·05, P = 0·7), aspartate transaminase levels (Spearman's r = 0·08, P = 0·6) or with alanine transaminase levels (Spearman's r = 0·08, P = 0·5).

As FoxP3⁺ cells were found to be expanded in acute infection, we sought to determine if the expansion was limited to acute infection. Therefore, we determined the frequency of FoxP3⁺ cells in nine of the patients 30–45 days after the onset of illness. The frequency of FoxP3⁺ cells significantly reduced (P = 0·02) 30–45 days after onset of illness in these patients (median = 0·56% of FoxP3⁺ cells, IQR = 0·41–1·1) and the frequencies in patients were similar to those of healthy individuals (median = 0·64% of FoxP3⁺ cells, IQR = 0·44–0·71) (Fig. 1c).

As we found that FoxP3 cells were significantly expanded in acute dengue, we determined if such expansion was associated with relevant cytokines, such as IL‐10, TGF‐β, IL‐23, IL‐2 and IL‐17A. Although serum TGF‐β levels were significantly (P < 0·0001) lower in patients (median = 151·2, IQR = 120·0–164·1 pg/ml) when compared to healthy individuals (median = 360·8, IQR = 257·1–437·9 pg/ml) (Fig. 2a), the level of TGF‐β did not associate with the frequency of FoxP3⁺ cells (Spearman's r = –0·3, P = 0·17) (Fig. 2b). However, TGF‐β levels were significantly higher (P = 0·03) in those with DHF when compared to those with DF. IL‐17A levels were similar in patients (median = 7·23, IQR = 4·9–9·2 pg/ml) when compared to healthy individuals (median = 5·67, IQR = 5·26–8·4 pg.ml). The IL‐17 levels did not correlate with the frequency of FoxP3⁺ cells (Spearman's r = −0·058, P = 0·77) or with any other laboratory parameter. The serum IL‐10 levels were significantly higher (P < 0·001) in patients (median = 33·7, IQR = 11·4–54·6 pg/ml) when compared to healthy individuals (median = 0, IQR = 0–0 pg/ml), as described in our previous studies.7, 14, 32 However, IL‐10 levels did not correlate with the frequency of FoxP3⁺ cells (Spearman's r = −0·19, P = 0·81). Although we assessed plasma IL‐23 and IL‐2 levels in patients, these cytokines were below the level of detection in both patients and healthy individuals.

Figure 2.

Transforming growth factor (TGF)‐β levels in patients with acute dengue. TGF‐β was measured in plasma of patients with acute dengue (n = 53) and healthy controls (n = 12). (a) TGF‐β levels in plasma of patients with acute dengue, healthy controls, patients with DF (n = 26) and DHF (n = 27). (b) Correlation of plasma TGF‐β levels with the frequency of forkhead box protein 3 (FoxP3)‐expressing CD4⁺ T‐cells. The bars represent the median and interquartile range. *P < 0·05, ***P < 0·001.

Relationship between FoxP3⁺ cells, viraemia and DENV‐specific T‐cell responses

Due to the limitation in the number of cells, we carried out ex‐vivo ELISPOTS and DENV viral loads in only 25 patients with acute dengue. Eight of these patients (eight of 25) had DF and 17 of 25 had DHF based on the WHO 2011 guidelines. We did not observe any correlation between the expansion of FoxP3⁺ cells with the viral loads (Spearman's r = −0·31, P = 0·15). The frequency of FoxP3⁺ cells did not correlate with either ex‐vivo DENV‐NS3‐, NS5‐ or NS1‐specific T‐cell responses.

Only 10 of 25 patients had ex‐vivo ELISPOT responses to DENV‐NS3 peptides of > 50 SFU/1 million PBMCs (a positive response to DENV‐NS3). There was no difference (P = 0·6) in the frequency of FoxP3⁺ cells in those who had a T‐cell response to DENV‐NS3 of > 50 SFU/1 million PBMCs (median = 2·2% of FoxP3⁺ cells, IQR = 1·4–2·8) when compared to those with a high frequency of IFN‐γ T‐cell responses to DENV‐NS3 (median = 2·7% of FoxP3⁺ cells, IQR = 0·8–4·1). Sixteen of 25 of patients had no response to DENV‐NS5 peptides (frequency of ex‐vivo IFN‐γ ELISPOT responses 0 SFU/1 million PBMCs). Again, there were no significant differences in the frequency of FoxP3⁺ cells in those who had no responses to DENV‐NS5 peptides compared to those who had some production.

Phenotypical analysis of FoxP3⁺ cells in acute dengue

FoxP3‐expressing CD4⁺ T‐cells can be categorized as natural thymic‐derived T_regs (nT_regs), highly suppressive T_regs (effector T_regs) and activated T‐cells transiently expressing FoxP3 (non‐T_regs), which are not suppressive, based on the expression of CD45RA and staining intensity of FoxP3.23, 24 While nT_regs and effector T_regs are subsets of regulatory T‐cells, activated T‐cells transiently expressing FoxP3 (CD45RA‐FoxP3^low) have shown to produce proinflammatory cytokines and do not possess any regulatory functions.24 The gating strategies to determine CD45RA‐expressing FoxP3^low/high cells are shown in Fig. 3a. We found that the FoxP3‐expressing T‐cells of patients with acute dengue were predominantly CD45RA⁺ FoxP3^low (median = 45·3, IQR = 23·9–53·1 of the total FoxP3⁺ cell population), followed by CD45RA‐FoxP3^low (median = 17·3, IQR = 9·2–41·5 of the total FoxP3⁺ cell population) (Fig. 3b). Only a small proportion of FoxP3⁺  cells were of the highly suppressive effector T_reg population, which were characterized by CD45RA⁻, FoxP3^high (median = 9·8, IQR = 4·9–14·6 of the total FoxP3⁺ cell population).

Figure 3.

Forkhead box protein 3 (FoxP3)‐expressing T‐cell subtypes in patients with acute dengue. Peripheral blood mononuclear cells (PBMCs) were isolated from patients with acute dengue (n = 10) and expression of CD45RA, FoxP3 and CCR4 were assessed. (a) Representative figure of identification of T‐cell subtypes based on their expression of CD45RA and FoxP3. nT_regs were identified as those which are CD45RA⁺FoxP^low (i), effector T‐cells which are CD45RA⁻ FoxP3^low due to transient expression of FoxP3 (ii) and effector T_regs as those which are CD45RA⁻ FoxP3^high (iii). (b) Proportion of FoxP3‐expressing T‐cell subtypes and effector T‐cells (nT_regs) in patients with acute dengue (n = 10) based on their expression of CD45RA and FoxP3. (c) Proportion of FoxP3⁺ CD4⁺ T‐cell subtypes expressing CCR4 in patients with acute dengue (n = 10) The bars represent the median and interquartile range. *P < 0·05, **P < 0·01.

It has also been shown that the highly suppressive, terminally differentiated effector type of T_regs predominantly expressed CCR4.21 In order to further characterize the T_reg subtypes, we evaluated the expression of CCR4 in nT_regs, effector T_regs and the non‐T_reg population (Fig. 3c). As described previously,21 we found that CCR4 expression was highest in the effector T_reg population (median = 78·9, IQR = 45·3–86·7 of the effector T_reg population) and expression of CCR4 was lower in the nT_regs (median = 20·04, IQR = 11·2–29·8 of the nT_reg population) and in non‐T_regs (median = 20·2, IQR = 9·7–36·8 of the non‐T_reg population) (Fig. 3c).

CTLA‐4 is one of the most important co‐stimulatory molecules expressed by T_regs and is essential for their function.33, 34 Therefore, we sought to determine CTLA‐4 expression on the T_reg population in patients with acute dengue. CTLA‐4 expression was significantly higher (P = 0·002) in the FoxP3⁺ cell population (median = 51·1, IQR = 41·6–62·3 of the total FoxP3⁺ cell population) when compared to the CD4⁺ T‐cell population (median = 37·4, IQR = 31·9–55·9 of the total population of CD4⁺ T‐cells) (Fig. 4a). Fas/FasL‐mediated apoptosis is another important mechanism by which T_regs exert their function.35, 36 Therefore, we also sought to determine Fas expression by the FoxP3⁺ cell population in patients with acute dengue. We found that Fas (CD95) was indeed expressed significantly more highly (P = 0·002) in FoxP3⁺ cells (median = 57·5, IQR = 41·9–73·9 of the total FoxP3⁺ cell population) when compared to the CD4⁺ T‐cell population (median = 20·2, IQR = 15·5–26·1 of the total CD4⁺ T‐cell population (Fig. 4b).

Figure 4.

Expression of cytotoxic T lymphocyte‐antigen 4 (CTLA‐4) and Fas by forkhead box protein 3 (FoxP3)‐expressing T‐cells in patients with acute dengue. Peripheral blood mononuclear cells (PBMCs) were isolated from patients with acute dengue (n = 10) and expression of CTLA‐4 and Fas were assessed by flow cytometry. (a) Expression of CTLA‐4 in the total CD4⁺ T‐cell population and the FoxP3⁺ T‐cell population in patients with acute dengue (n = 10). (b) Expression of Fas (CD95) in the total CD4⁺ T‐cell population and the FoxP3⁺ T‐cell population in patients with acute dengue (n = 10). **P < 0·01.

Discussion

It was previously reported that T_regs expand in acute dengue, possibly to suppress the highly cross‐reactive DENV‐specific T‐cells which are thought to contribute to disease pathogenesis.18 Another study also found that T_regs were expanded in acute dengue, especially in those with milder dengue when compared to those with moderate clinical disease.19 In this study, we also found that the frequency of FoxP3‐expressing CD4⁺ T‐cells (FoxP3⁺ cells) significantly expanded in acute dengue, which returned to the frequencies observed in healthy adult individuals after convalescence. However, expansion of FoxP3⁺ cells was not associated with clinical disease severity, the extent of viraemia, the frequency of DENV‐NS3, NS1 or the overall DENV‐specific T‐cell responses or with any laboratory markers of disease severity, such as platelet counts or levels of liver transaminases. Although the frequency of FoxP3⁺ cells was higher in patients with DF when compared to those with DHF, this was not significant. In addition, the expansion of the FoxP3⁺ cells did not appear to associate with DENV‐specific T‐cell responses, which were assessed by e‐vivo IFN‐γ ELISPOT responses to DENV‐NS1, NS3 and NS5 peptides pooled together and also to the pool consisting of peptides for all three DENV proteins.

Some of the suppressive functions of T_regs are known to be mediated by cytokines such as IL‐10 and TGF‐β produced by T_regs.37 TGF‐β is also known to be important in the generation of T_regs in the periphery.37 However, neither IL‐10 nor TGF‐β levels correlated with the frequency of FoxP3⁺ cells. In addition, the TGF‐β levels were significantly lower in patients with acute dengue compared to levels seen in healthy individuals. Human platelets are thought to be one of the main sources of TGF‐β , which is stored in their granules.38 As all patients with acute dengue had some degree of thrombocytopenia, this could have resulted in the lower TGF‐β levels compared to healthy individuals. However, although patients with DHF had significantly lower platelet counts when compared to patients with DF, their plasma TGF‐β levels were significantly higher. As platelets of patients with acute dengue have shown to be highly activated,39 this could have possibly contributed to higher TGF‐β levels in DHF due to increase disease severity.

In all previous studies, although the frequency of T_regs was determined, phenotyping was not performed to determine if they were effector or natural T_regs or activated T‐cells transiently expressing low levels of FoxP3. Although Luhn et al. showed that CD4⁺CD25⁺ T‐cells isolated from patients following in‐vitro activation and expansion suppressed DENV‐specific T‐cells in ex‐vivo ELISPOT assays,18 it would be useful to determine their function and phenotype ex vivo. It has been shown that not all FoxP3‐expressing CD4⁺ CD25⁺ T‐cells are regulatory T‐cells, and those which are CD45RA⁻ FoxP3^low cells are, in fact, activated T‐cells transiently expressing FoxP3 and do not possess any suppressive capabilities (non‐T_reg cells).23, 24 These non‐T_regs cells have been shown to produce proinflammatory cytokines rather than immunosuppressive cytokines.24 Highly suppressive, terminally differentiated effector T_regs are shown to be CD45RA⁻ FoxP3^high and have also been shown to express high levels of CCR4.21, 23 T_regs from CCR4^−/− mice have been shown to have reduced suppressive capacity,22 which showed that CCR4 was important for the suppressive effects of T_regs and is highly expressed on effector T_regs. We found that the majority of FoxP3⁺ cells in patients with acute dengue were of the natural, thymic‐derived T_reg subtype and did not express CCR4. A significant proportion (median = 17·3, IQR = 9·2–41·5 of FoxP3‐expressing T‐cells) were found to be effector T‐cells (non‐T_regs), which are likely to be activated T‐cells transiently expressing FoxP3. These effector T‐cells had low expression of CCR4. Only a minor proportion of the FoxP3⁺ cells were effector T_regs (median = 9·8, IQR = 4·9–14·6 of the total FoxP3⁺ cell population), and a large proportion of these T‐cells expressed CCR4 (median 78·9%). Therefore, although FoxP3 cells are expanded in acute dengue, the majority of such FoxP3⁺ cells appear to be naive, natural T_regs.

T_regs exert their function in part by producing immunosuppressive cytokines, but also by suppressing the activity of T‐cells through a CTLA‐4‐dependent effect.33, 34 In addition, they induce apoptosis of activated T‐cells through Fas/FasL pathways.35, 36 We also found that the expression of CTLA‐4 and Fas was significantly higher in FoxP3⁺ cell patients with acute dengue when compared to the total CD4⁺ T‐cell population. However, the predominant population that is seen in acute dengue are natural thymic‐derived T_regs, which require activation in order to exert their suppressive action. Therefore, it is not clear if the expression of CTLA‐4 and Fas by these subset of T_regs alone would enable them to suppress DENV‐specific T‐cells in vivo. Mechanisms underlying the suppressive ability of these thymic derived T_regs in acute dengue warrants further investigation.

In summary, although FoxP3⁺ cells are expanded in acute dengue, they did not have any relationship with clinical disease severity, DENV‐specific T‐cell function or viraemia. The FoxP3‐expressing T‐cell population predominantly consisted of naive, thymic‐derived natural T_regs followed by activated T‐cells, which are transiently known to express FoxP3. The proportion of highly suppressive effector T_regs, which express CCR4, were present at a very low frequency. Therefore, although FoxP3⁺ cells are expanded in acute dengue, they predominantly consist of naive T_regs with poor suppressive capacity.

Disclosures

None.