Elevated effector cell sensitivity to Treg-cell suppression that is not associated with reduced Th17-cell expression distinguishes HIV+ asymptomatic subjects from progressors



Suppression mediated by Treg cells is a balance between Treg-cell suppressive potency versus sensitivity of effector cells to Treg-cell suppression. We assessed if this balance, along with Treg-cell number relative to the Treg-cell counter-regulatory cytokine IL-17, differs between asymptomatic HIV+ subjects versus those who progress onto disease. Cross-over studies comparing Treg-cell potency, measured by effector cell proliferation or IFN-γ expression, from HIV-infected versus control subjects to suppress the proliferation of allogeneic control effector cells demonstrated increased sensitivity of CD4+CD25 effector cells from asymptomatic HIV+ subjects to suppression, rather than an increase in the suppressive potential of their CD4+CD25+ Treg cells. In contrast, HIV+ progressors did not differ from controls in Treg-cell potency or effector cell sensitivity to Treg-cell suppression. Both CD4+CD25+Foxp3+ Treg and effector IL-17 absolute cell numbers were significantly lower in all HIV+ subjects tested and not restored by antiviral therapy. Thus, these novel data suggest that elevated Treg-cell-mediated suppression due to increased sensitivity of effectors to Treg cells may be a natural host response in chronic asymptomatic HIV infection, which is lost as disease progresses and that this feature of CD25 effector cells is not inextricably linked to reduced production of the Treg-cell counter-regulatory cytokine IL-17.


Treg cells are a subset of CD4+ T lymphocytes that can potently negatively regulate immune responses. Treg cells can restrain the vigour of diverse antigen-specific responses in humans and consequently have been associated with the inability to clear infection of some pathogens 1–3. However, in HIV infection, Treg cells appear to play opposing roles, contingent on disease stage. In acute HIV-1 infection, the presence of Treg cells is hypothesised to dampen protective antiviral responses 4–7, while in the chronic phase their presence may be protective by limiting damaging immune activation 8–14. Assessing the significance of Treg cells in HIV infection therefore requires a systematic analysis of both Treg-cell function and number.

The emerging consensus from several laboratories is that Treg cells with suppressor potential can be detected in all stages of HIV disease 8, 12, 15. However, qualitative aspects of Treg-cell function in HIV infection remain poorly characterised. Specifically, it remains largely unknown whether HIV infection alters Treg-cell suppressive potential or alters effector cell sensitivity to Treg-cell suppression. Our laboratory previously reported enhanced Treg-cell-mediated suppression in treatment-naive chronically HIV-1-infected asymptomatic patients compared to healthy controls 15. Kinter et al. 13 also highlighted Treg cells from lymph nodes of chronically HIV-1-infected patients to be significantly more potent at suppressing HIV-1-specific CD8 cytolytic activity than peripheral blood Treg cells 13. In our previous study 15 we went on to demonstrate for the first time that the net increase in Treg-cell-mediated suppressor potential in asymptomatic HIV+ subjects was due to increased sensitivity of effector cells to be suppressed, rather than an increase in the potency of their Treg cells to mediate suppression, emphasising the importance of assessing Treg-cell function in the context of both the Treg and effector cell simultaneously. This study extends these observations and probes Treg cell quality in HIV+ progressors prior to and after Highly Active Anti-Retroviral Therapy (HAART) initiation.

In addition to impacting quality, HIV infection is known to alter Treg cell quantity. Several studies, including ours, report a decline in absolute Treg-cell number in chronic HIV infection 8, 11, 15. Some studies show Treg-cell frequency to be elevated in HIV infection 16, 17, but this discrepancy may reflect CD4+ T-cell count disparity in HIV+ subjects. A systematic longitudinal analysis of Treg-cell absolute number in HIV+ progressors prior to and after HAART initiation is therefore warranted. Furthermore, the importance of examining Treg-cell quantity in the context of the Treg-cell counter-regulatory cytokine, IL-17 18, 19, is increasingly being recognised. Studies in nonhuman primate models of lentiviral infection and in HIV-infected human individuals highlight pathogenic infection to be associated with loss of Th17 cells 19–23. IL-17 serves to maintain the integrity of the mucosal barrier. Loss of Th17 cells may permit microbial translocation across the gastrointestinal mucosa and thereby promote immune activation driven by bacterial lipopolyscaacharide, which is associated with disease progression 20, 24, 25. In this manuscript we provide novel insight into both qualitative and quantitative aspects of Treg cells in chronic HIV infection. We demonstrate that increased sensitivity of effector cells to Treg-cell mediated suppression is a feature of asymptomatic HIV-1 infected patients, but not patients who have progressed onto therapy; that this function is not inextricably linked to reduced expression of the counter-regulatory IL-17 cytokine and that reduced Treg and IL-17 numbers is a feature of chronic HIV infection that is not restored by up to 12 months of antiviral therapy.


Impaired CD25 effector cell proliferative function prior in pre-HAART HIV+ progressors

Assessing Treg-cell function is contingent on robust proliferation and cytokine expression by effector cells following TCR ligation. This function is known to be compromised in HIV-1-infected individuals 26, 27. Longitudinal analysis of effector cell proliferative capacity from chronically HIV-1-infected progressor patients prior to the initiation of HAART (Prog. pre-HAART) as well as at 1 and 4 months post-HAART showed a significant reduction in anti-CD3/CD28-induced proliferation, with evidence of gradual recovery by 12 months post-HAART initiation. On the other hand, effector cells from chronically HIV-1-infected untreated subjects proliferated as efficiently as that of controls (Fig. 1A). Consequently, suppression of autologous effector cells could only be reliably measured in chronic untreated and the 12 month post-HAART progressor groups. Figure 1C and D and Supporting Information Fig. 1A and B show that autologous suppression in 12 month HAART patients, tested at two effector:Treg-cell ratios, 1:0.125 and 1:0.06, respectively, were not significantly different to that of controls. In contrast, Fig. 1B shows autologous suppression in chronic untreated patients to be significantly elevated compared to that of controls (mean±SD 70.53%±29.36 in controls versus 89.27%±14.35 in patients, p=0.0104), confirming similar observations in a larger cohort of chronic untreated HIV+ subjects 15.

Figure 1.

Effector cell proliferation and autologous Treg-cell suppressor function in chronic HIV-1 infection. Treg-cell-mediated suppression assays were conducted by culturing 2.5×103 CD4+CD25 effector T cells with varying ratios of CD4+CD25+ Treg cells in triplicate with CD3/28 beads at a bead:cell ratio of 2:1 for 5 days. Proliferation was measured by tritiated thymidine incorporation as mean CPM and percentage suppression was calculated as described in the Materials and methods. (A) Comparative proliferation of CD4+CD25 effector cells (in the absence of Treg cells) from control and HIV+ subjects. Data show mean CPM in CD3/28 stimulated cultures with background un-stimulated cultures having a mean CPM of <150 (data not shown). Data are representative of mean+SD of ten individuals per group. (B) Comparison of autologous suppression of chronically untreated patients versus controls at effector:Treg-cell ratio of 1:0.06. (C and D) Comparison of autologous suppression of progressor patients 12 months post-HAART therapy versus controls at effector:Treg cell ratio of 1:0.125 (C) and 1:0.06 (D). Data are representative of mean+SD of six individuals per group. p-Values were calculated using Mann–Whitney U t test.

Equivalent suppression by Treg cells from asymptomatic subjects and pre-HAART progressors

We performed allogeneic cross-over suppression assays, which we 15 and others 28, 29, have previously used to compare the quality of Treg-cell potency with that of effector cell susceptibility to Treg-cell mediated suppression. Effector cells from allogeneic controls were used as targets. First, we demonstrate that the potency of Treg-cell-mediated suppression was similar when allogeneic or autologous effector cells were used in a suppression assay (Supporting Information Fig. 3A). Next, we compared Treg cells from chronic untreated HIV+ subjects with that of controls and demonstrate as previously reported 15 Treg-cell potency to be similar in these two groups, Fig. 2A. HIV-1-infected progressors prior to and after antiviral therapy were next tested using the same assay. Interestingly, despite effector cells from progressors pre-HAART being impaired (Fig. 1A), we observed that Treg cells from this patient group prior to and longitudinally after HAART initiation retained the capacity to suppress at the same level as Treg cells isolated from controls tested in parallel (Fig. 2B and C, and Supporting Information Fig. 2A–C).

Figure 2.

Treg cells from all chronic HIV+ subjects suppress effector cell proliferation. Allogeneic cross-over assays were performed, using a standard healthy control effector cell population co-cultured in triplicate with varying ratios of effector:Treg cells (E:T) as indicated. CD4+CD25+ Treg cells from longitudinal samples of HIV+ progressors pre- and post-initiation of HAART, chronic untreated patients and healthy controls were tested. (A) Comparison of suppression by Treg cells from controls (n=20) and chronic untreated patients (n=33) tested at varying E:T ratios is shown. (B and C) Comparison of suppression by Treg cells from controls and longitudinal samples from patients pre- and post-HAART tested at E:T ratio of 1:0.125 (B) and 1:0.06 (C). Data are representative of mean±SD of six individuals per group. p-Values calculated using Mann–Whitney U t test and Kruskal–Wallis multiple comparisons test.

Treg cells from pre-HAART HIV+ progressors suppress IFN-γ and IL-2 production

To further confirm that Treg-cell potency is preserved in HIV+ progressors, we assessed the potency of Treg cells to suppress the effector cytokines IFN-γ and IL-2. Representative data of IL-2 and IFN-γ suppression in the presence of Treg cells is shown in Fig. 3A. First, we confirmed potency of suppression to be similar when autologous versus allogeneic effectors were compared using single IFN-γ+ cells as a read-out (Supporting Information Fig. 3B). Next, Treg cells from progressors pre- and post-HAART were assessed for suppressive potential of single IL-2 (Fig. 3B), single IFN-γ (Fig. 3C) and IFN-γ/IL-2 double positive (Fig. 3D) from effectors of controls. Figure 3B–D confirm data presented in Fig. 2B and C that Treg-cell potency is similar to that of controls, as measured by suppression of both IFN-γ and IL-2 effector cytokine expression. Taken together, data in Fig. 2 and 3 confirm Treg-cell potency to be preserved similar to that of controls in chronic HIV infection irrespective of disease stage and therapy status.

Figure 3.

Treg cells from all chronic HIV+ subjects suppress IFN-γ and IL-2 expression. IL-2 and IFN-γ production was measured by a standard intracytoplasmic cytokine-staining assay. 2×104 CD4+CD25 effector cells from healthy controls were activated for 16 h with 2:1 (bead:cell ratio) CD3/CD28 mitogenic beads in the presence of brefeldin A, in the presence or absence of varying ratios of Treg cells isolated from patients or controls as indicated. (A) Representative FACS plots depicting gating strategy for single IFN-γ, single IL-2 and double IL-2/IFN-γ+ cells is shown following activation of effector cells in the absence or presence of varying Treg-cell ratios from a control donor. (B–D) Comparison of suppression of single IL-2 (B), single IFN-γ (C), and double IL-2/IFN-γ (D) positive effectors cells by Treg cells from HIV-1-infected subjects pre- and post-HAART versus Treg cells from controls. Data are representative of mean±SD of four individuals per group. p-Values calculated using Mann–Whitney U t test and Kruskal–Wallis test.

Effector cells from asymptomatic HIV+ subjects are highly sensitive to Treg-cell mediated suppression

Previously we reported that effector cells from chronic untreated HIV-1-infected subjects were more sensitive to Treg-cell mediated suppression than effector cells from controls 15. To extend this analysis to cells from HIV+ progressor pre- and post-HAART, we elected to utilise IFN-γ expression as a readout of effector cell function, which we and others have previously reported to be preserved in HIV-1-infected patients 26, 30. A suppression assay based on proliferation could not be utilised as effector cells from HIV-1-infected progressors have impaired proliferative capability (Fig. 1A). We first confirmed in Fig. 4A the frequency of single IFN-γ+ cells to be similar in controls, chronic untreated and progressor subjects. Next Treg cells isolated from a group of controls were each tested for their ability to suppress effectors from progressors, chronic untreated subjects and allogeneic controls in parallel. Figure 4B demonstrates a significant increase in the sensitivity of effectors from chronic untreated to allogeneic Treg-cell mediated suppression where 6/8 subjects tested had a higher mean suppression (mean percentage suppression 68.37%±19.79) than that of controls (36.81%±24.43). On the other hand, the majority of progressor pre-HAART tested (5/8) had similar allogeneic suppression to that of controls and the overall mean suppression levels did not differ between controls and progressors pre-HAART (Fig. 4B). Taken together, these data highlight increased sensitivity of effector cells to Treg-cell mediated suppression to be a distinguishing feature of chronic untreated versus progressor pre-HAART HIV-1-infected subjects.

Figure 4.

Sensitivity of effector cells to Treg-cell-mediated suppression. 2×104 CD4+CD25 effector cells from HIV+ subjects or healthy controls were tested for sensitivity to suppression by Treg cells from healthy controls at an E:T ratio of 1:0.1 using single IFN-γ+ cells as a readout in a standard 16-h intracytoplasmic cytokine staining assay. (A) Mean frequency of single IFN-γ+ cells by effector cells in the absence of Treg cells in controls, chronic untreated and pre-HAART progressors following CD3/28 stimulation is shown (N=6 subjects per group). (B) Comparison of susceptibility of effector cells from eight controls versus eight HIV+ chronic untreated versus nine HIV+ pre-HAART progressors by a panel of Treg cells from controls is shown. Each point represents suppression level of one individual and the horizontal line indicates mean values. p-Values calculated using Mann–Whitney U t test and Kruskal–Wallis test.

Absolute numbers of Treg cells decline in both progressive and asymptomatic HIV-1 infection and is not restored by HAART

Given the significant heterogeneity in absolute CD4+ T-cell counts across the patients groups studied (see Supporting Information Tables 1–3), we calculated absolute Treg-cell numbers based on CD4+CD25+FoxP3+ expression. Relative to controls, a consistent and significant reduction in Treg-cell absolute numbers was noted across all HIV-1-infected patients tested, which did not recover by 8–12 months following HAART initiation (Fig. 5A). The decline in absolute Treg-cell numbers correlated positively with CD4+ T-cell count (Fig. 5B), but not with virus load (Supporting Information Fig. 4). These data suggest that the loss in absolute Treg-cell numbers occurs at the same rate as total CD4+ T-cell loss, and is not subject to selective depletion, in accordance with other reports 8, 11.

Figure 5.

Decline in absolute numbers of Treg cells and IL-17+ cells in chronic HIV-1 infection. PBMCs from controls or HIV+ subjects as indicated were stained for Treg-cell markers or for IL-17 and enumerated as described in the Materials and methods. Absolute Treg and IL-17+ cell numbers were calculated based on patient CD4+ T-cell count at the time of sampling. (A and C) Mean+SD of absolute numbers of CD4+CD25+Foxp3+ Treg cells (A) and IL-17+ cells (C). p-Values calculated using Mann–Whitney U t test and Kruskal–Wallis test. (B and D) Correlation between absolute number of CD4+CD25+Foxp3+ (B) or IL-17+ (D) cells versus total CD4+ cell counts. p-Values calculated using Spearman's correlation. (E) Comparison of Treg cells:Th17 cell ratio in chronic untreated and pre-HAART progressors versus controls. p-Value determined by Mann–Whitney U test.

IL-17 expression is significantly impaired in HIV-1 infection and not restored by HAART

Effector cells are a major source of IL-17 production and known to have a reciprocal developmental pathway to Treg cells, impacting Treg-cell frequency and function 31. We therefore explored if increased sensitivity of effector cells from asymptomatic chronically HIV-1-infected patients to Treg-cell mediated suppression may be attributed to reduced IL-17 expression by these cells. Figure 5C demonstrates a significant decline in absolute IL-17+ cell number across all HIV+ subjects tested with little recovery even at 8–12 months post-HAART initiation, with HIV+ progressors pre-HAART expressing the lowest level of IL-17. Absolute IL-17+ cell number, like absolute Treg-cell number, correlated positively with CD4+ cell count (Fig. 5D), but not virus loads (data not shown). To explore if the observed decline in both Treg-cell and IL-17+ cell numbers occurred proportionally, we compared Treg:IL-17+ cell ratios in controls, HIV+ asymptomatic and HIV+ progressors prior to HAART therapy. Consistent with others 19, we noted the mean Treg:IL-17+ cell ratio in controls to be ∼13:1. This ratio remained unaltered in HIV-1-infected chronic untreated patients (Fig. 5E). In accordance with a greater fall in IL-17+ cell numbers in progressors compared to chronic untreated subjects (Fig. 5C), we observed a trend for an increase in the mean Treg:IL-17+ cell ratio in this group, which was 34:1 versus a ratio of 13:1 in controls; however, this difference did not reach statistical significance (Fig. 5E). These data highlight a significant reduction in effector IL-17 expression in both HIV+ chronic untreated and progressor patients and therefore cannot explain why effector cells from chronic untreated, but not progressors, are more sensitive to Treg-cell-mediated suppression.


Understanding precisely how Treg-cell function may be altered in HIV-infected subjects is of importance in determining if this essential subset represents a reasonable target for immune-based therapy in HIV infection, and if such therapy would be appropriate at all stages of HIV disease. This question is particularly pertinent in HIV infection where Treg cells may play opposing roles, being associated with detrimental outcome in the acute stage by suppressing HIV-specific adaptive immune responses 4–7; indeed in vitro HIV infection has been shown to induce Treg cells 32, 33, but beneficial in the chronic stage by controlling excessive immune activation 8, 11, 12, 34, 35. This study was designed to provide fresh insight into this issue by utilising an experimental approach that we 15 and others 28, 29 have previously used to dissect Treg-cell potency from effector cell sensitivity to Treg-mediated suppression. Furthermore, our optimised suppression assay importantly takes into account the dynamic nature of Treg-cell function, which is critically linked to Treg-cell purity (Supporting Information Fig. 5), signal strength, Treg:effector cell ratio (see 36, 37), and cell density (see Supporting Information Fig. 7), thereby rendering our assay highly sensitive. In so doing, we highlight three novel aspects of Treg-cell function in chronic HIV infection that is discussed below.

It is well known that HIV infection impairs CD4+ T-cell proliferative function, especially in progressors 38–41, which we confirm (Fig. 1A). Consequently it is not possible to conduct an autologous suppression assay using cells from this patient group. We circumvented this limitation by using an allogeneic system 15, 28, 29. This assay enables the potency of Treg cells from different HIV-1-infected groups to be compared by assessing their ability to suppress effector cells from healthy controls. Conversely, effector cells from different patient cohorts can be compared for their sensitivity to be suppressed by Treg cells isolated from controls. Using this assay, we provide unequivocal evidence that CD4+CD25+FoxP3+ Treg-cell potency in all chronic HIV+ subjects tested is comparable to controls tested in parallel, irrespective of their CD4+ T-cell count, virus load, disease stage or therapy status, using either a proliferation assay or an IFN-γ intracellular staining (ICS) assay as a readout. The mechanism for the selective loss of effector cell proliferative capacity, but not Treg cell-suppressive potential, is presently unclear, especially as Treg cells appear to be more readily infected than activated effector cells 15, 42, 43. The implication is that lower IL-2 expression, a hallmark of HIV infection 26, 27, accounts for loss in effector cell proliferation, without impacting the sensitivity of these cells to Treg-cell mediated suppression. This notion is supported by other data showing Treg suppression to be preserved in chronic HIV+ subjects and Simian Immunodeficiency Virus (SIV) models, despite a fall in CD4+ T-cell count 4, 6, 8, 13, 14, 36. Furthermore, the preservation of Treg-cell potency in HIV infection is interesting, as Treg cells are known to critically rely on IL-2 for expansion and function 44, 45 and may reflect threshold differences in IL-2 requirement for Treg and effector cell function.

The second important aspect of this study is the observation that effector cell sensitivity to Treg-cell mediated suppression, using IFN-γ as a readout, is elevated only in chronic untreated HIV+ subjects but not progressor pre- and post-HAART. A previous report by Kinter et al. 13 also highlighted elevated suppression in lymph node Treg cells compared to peripheral blood, but did not establish if this is due to increased potency of patients Treg cells and/or an increased sensitivity of effector cells to Treg-cell suppression. A key question that arises from our data is whether increased effector cell sensitivity to Treg-cell suppression is linked to reduced IL-17 expression. Treg cell development is intimately linked to the counter-regulatory pro-inflammatory cytokine, IL-17, with Treg cells being negatively regulated by Th17 cells 31, 46. Evidence that this cannot be the sole explanation is provided. We demonstrate that effector cells from both chronic untreated and pre-HAART progressors are severely impaired in IL-17 expression. Indeed, progressors have significantly fewer IL-17+ cells than chronic untreated patients. Yet effector cells from progressors are not more sensitive to Treg-cell suppression relative to controls, highlighting that loss in IL-17 expression is not inextricably linked to increased effector cell sensitivity to Treg-cell-mediated suppression. Whether the dramatic loss of circulating IL-17+CD4+ T cells results in IL-17 paucity in vivo is not known and may well be compensated by IL-17 produced by iNKT or γδ T cells 47. On-going studies aim to elucidate the mechanisms of increased effector cell sensitivity to Treg-cell mediated suppression beyond IL-17 expression and whether contact-dependent suppression noted in control cultures (Supporting Information Fig. 6) is also preserved in cells form HIV+ subjects.

Our data on the loss of both Treg-cell and IL-17+ subsets extend other observations 18–25, 48. Both Treg-cell and IL-17 numbers correlate with CD4+ T-cell numbers, indicating that these cells are lost as part of the overall decline in CD4+ T-cell count (Fig. 5). Whether the greater loss of IL-17 cells in progressors (Fig. 5C) 19 is indicative of these cells being preferentially targeted over and above Treg cells by HIV 22, 49 or relates to other indirect mechanisms remains to be elucidated. Interestingly, HAART clearly restores effector CD4+ T-cell proliferative capacity (Fig. 1A), but not Treg or IL-17 cell numbers (Fig. 5). Kolte et al. 16 reported increased Treg-cell numbers 5 years after HAART initiation. However, similar to our study, Gaardbo et al. 17 report that Treg cell absolute numbers are significantly reduced prior to HAART, and remain the same at 24 wk following therapy. The failure to restore Treg and IL-17 numbers may reflect inefficient CD4+ T-cell recovery despite efficient virus load control or relate to selective recovery of some but not all CD4+ T-cell subsets following antiviral therapy 50, 51.

In conclusion, our data support the contention that Treg-cell function is preserved despite a significant decline in number across all groups of chronic HIV subjects tested and that effector cells from chronic asymptomatic HIV+ subjects, but not untreated progressors, are rendered more sensitive to suppression relative to controls. Our contention is that elevated sensitivity of effector to Treg-cell suppression may compensate for a reduction in Treg-cell number and reflect a natural host response in the chronic phase of HIV infection that is lost as patients' progress to disease. A reduction in Treg-cell number with no compensatory increase in effector cell sensitivity to Treg-cell suppression would effectively reduce the net homeostatic control exerted by Treg cells. In turn this may contribute to T-cell activation, which is a hallmark of disease progression 30, 52, 53, thereby impacting HIV pathogenesis.

Materials and methods


Subjects were volunteers with HIV infection who attended the outpatient clinic at St Thomas' Hospital, London. A total of 33 treatment naive HIV+ progressors were examined (Supporting Information Table 1). The clinical criteria for selection of patients into this group included evidence of declining CD4+ T-cell count of 500 cells/mm3 in 3 years or a CD4+ T-cell count of <250 cells/mm3 within 3 years of known sero-conversion who were about to initiate therapy. Patients were not selected on viral load (VL). Subjects included were 29 males, four females, with a median age of 38.69 (25–67), 4 median years of infection (<1–17), a median CD4+ count of 240.2 (51–336) and median VL of 101 669 (45≥500 000). For longitudinal studies, these patients were sampled prior to and 1, 4, 8–12 months post-initiation of HAART (Supporting Information Table 2). In addition, 31 chronically infected asymptomatic treatment-naive HIV+ subjects were studied (Supporting Information Table 3). Chronic untreated patients were identified as being treatment naïve with a stable CD4+ count above 300, as measured on at least two occasions (from time of diagnosis and at 6–12 monthly intervals) prior to sampling, not requiring therapy. This group had a median age of 37.87 (26–53), eight of which were females and 23 were males, with a median CD4+ T-cell count of 672.5 (277–1439) and median VL of 17 451 (<50–18 779) and 5.5 (1–16) median years of infection. Control HIV sero-negative blood samples were purchased from the National Blood Transplantation Service at St George's Hospital Tooting, UK, and tested in parallel with samples from HIV+ subjects. For controls subjects, where information was available the intention was to match the patients as closely as possible in terms of age, and to attempt to match in terms of gender if possible. Although all recruited patients were studied, not all subjects could be analysed for all parameters included in this study, which was linked to blood volume, yield and CD4+ T-cell count at the time of experimentation.

Cell separation

Peripheral blood mononuclear cells (PBMCs) were isolated by density gradient centrifugation (Lymphoprep: Axis-Shield PoC AS, Oslo, Norway). CD4+CD45RO+CD25 effector and CD4+CD45RO+CD25+ Treg-cell populations were isolated using Dynabeads T regulatory cell isolation kit (Invitrogen, Paisley, UK) as described previously 15. Purity of isolated fractions was confirmed by immunostaining to be >95% for effector and Treg populations (Supporting Information Fig. 5).

Suppression of effector cell proliferation

All assays were carried out in RPMI-1640 Glutamax 25 mM HEPES media (Invitrogen), 10% human AB serum (Lonza, Sweden), and supplemented with 20 μg/mL Gentamycin (Sigma-Aldrich, UK) as described previously 15 by co-cultuting 2.5×103 effector cells, with at least two ratios of Treg cells. Cells were stimulated with Dynal anti-human CD3/CD28 coated magnetic beads (bead: effector cell ratio, 2:1) (Invitrogen) for 5 days. Each well received 0.5 μCi of (3 H)-thymidine (Perkin Elmer, UK) for the last 16 h of culture.

Suppression of effector cytokine expression

As described previously 15 2×104 effector cells were cultured with varying ratios of Treg cells and stimulated with 2:1 (bead:effector cell) Dynal anti-human CD3/CD28 coated magnetic beads. After the addition of Brefeldin A (Sigma-Aldrich) cultures were maintained for 16 h before ICS for IFN-γ (PE-IFN-γ) and Interleukin-2 (APC-IL-2, both BD Pharmingen, UK) or appropriate isotype control mAbs.

IL-17 detection

PBMCs were cultured with 10 ng/mL phorbol 12-myristate 13-acetate (PMA), 1 μg/mL Ionomycin and 5 μg/mL Brefeldin A (all Sigma-Aldrich) for 4 h. CD4+ T cells were identified as CD3+CD8 by surface staining. Intracellular IL-17 (FITC labelled IL-17A, eBioscience, San Dieago, CA, USA) and IFN-γ (PE-labelled, BD Pharmingen) cytokines were measured using a fixation and permeabilisation kit 15 in a standard ICS assay. Absolute IL-17 numbers were determined as the percentage of cells staining positive for IL-17 secretion multiplied by the absolute CD4+ T-cell count of the patient/control at the time of sampling.

Enumerating Treg cells

FoxP3 expression was determined using anti-human FoxP3 staining set (Clone PCH101, eBioscience). Briefly, cells were surface stained with FITC-labelled CD4+ (clone SK3 BD Pharmingen) and PE-labelled CD25 (clone MEM-181, AbD Serotec, Oxford, UK). Cells were then washed and fix/permeabilised and stained using Fix/Permeabilisation Foxp3 staining kit for FoxP3 or the appropriate isotype control antibody 15. Absolute numbers of Treg cells were determined as the percentage of cells staining for defined Treg cell markers multiplied by the absolute CD4+ T-cell count of the patient/control at the time of sampling.

Statistical analysis

Statistical analysis was performed using Graphpad PRISM software (Graphpad Prism, version 4, CA, USA). Unpaired multiple comparison tests were performed using non-parametric Kruskal–Wallis test. Paired analysis was performed using Student's t test. p-Values of 0.05 and below were considered statistically significant.


G.T. was supported by an educational grant from Gilead Pharmaceuticals. The authors acknowledge financial support from the Department of Health via the National Institute for Health Research (NIHR) comprehensive Biomedical Research Centre award to Guy's & St Thomas' NHS Foundation Trust in partnership with King's College London and King's College Hospital NHS Foundation Trust. We thank our patients for their active participation in the study. The author contributions were as follows: Experiments were conceived and designed by G.T., B.P. and A.V. and performed by G.T. Data were analysed by G.T. and A.V. The manuscript was prepared by G.T., A.V. and B.P.

Conflict of interest: The authors declare no financial or commercial conflict of interest.