Regulatory T Cells in HIV-Infected Immunological Nonresponders Are Increased in Blood but Depleted in Lymphoid Tissue and Predict Immunological Reconstitution

Abstract

Background: 

HIV-infected immunological nonresponders fail to immune reconstitute despite optimal treatment. We hypothesized that regulatory T cells (Tregs) are involved in immunological reconstitution. Tregs and Treg subpopulations were measured in blood and Foxp3⁺ cells in lymphoid tissue, and the impact of Tregs on immunological reconstitution was determined.

Methods: 

HIV-infected individuals on combination antiretroviral therapy for a minimum of 2 years were included. The study population included 14 immunological nonresponders (INR; CD4 T-cell count <200 cells/μL), 33 intermediate responders (CD4 T-cell count 200–500 cells/μL), 30 responders (CD4 T-cell count >500 cells/μL), and 34 healthy controls. Tregs, Treg subpopulations, and intracellular staining for interleukin 10 in peripheral blood were measured using flow cytometry. Foxp3⁺ cells in lymphoid tissue were evaluated using immunolabeling. The CD4 T-cell count was determined at inclusion and after 1 year of follow-up.

Results: 

INR displayed high percentage of Tregs and activated Tregs in peripheral blood accompanied by a high percentage of Tregs expressing interleukin 10, whereas numbers of Foxp3⁺ cells in lymphoid tissue were low. In contrast, responders resembled healthy controls. Finally, in INR, high level of Tregs in blood and Foxp3⁺ cells in lymphoid tissue were associated with higher level of immunological reconstitution after 1 year of follow-up.

Conclusions: 

In conclusion, altered distribution of Tregs was found in INR. Interestingly, high level of Tregs predicted higher level of immunological reconstitution suggesting a role for Tregs in immunological reconstitution.

In Brief

Supplemental Digital Content is Available in the Text.

INTRODUCTION

Treatment of HIV-infected individuals with combination antiretroviral therapy (cART) results in suppression of viral replication and increasing CD4 T-cell count. However, in immunological nonresponders (INR) CD4 T-cell counts do not increase sufficiently.^1–3 Immunological nonresponse is associated with high age, low CD4 nadir, long duration of HIV infection, and co-infection with hepatitis C virus.^4–9 However, these factors do not provide a full explanation, suggesting immunological factors in the host contribute.

The role of regulatory T cells (Tregs) in patients with immunological nonresponse is poorly described. Tregs play a role in sustaining tolerance to self-antigens, exerting their function through cell-to-cell contact, or by cytokines, primarily interleukin (IL)-10 and transforming growth factor β.¹¹ Furthermore, Tregs have anti-inflammatory properties resulting in downregulation of immune activation.¹² Immune activation is a hallmark of HIV infection and a strong predictor of disease progression.¹³ Thus, Tregs may be involved in the pathogenesis of HIV infection. In HIV-infected individuals, elevated Tregs in peripheral blood have been found in several studies although conflicting results exist as well.^14–18 A redistribution of Tregs between lymphoid tissue and blood has been suggested as an explanation of these conflicting results.^19–21 Furthermore, a subdivision of Tregs into activated, resting, and nonsuppressive Tregs, according to suppressive capabilities and expression of Foxp3 and CD45RA was recently described.²² So far, Treg subpopulations have not been described in HIV-infected INR.

Increased risk of both AIDS and non-AIDS mortality in individuals with immunological nonresponse is widely accepted,^7,23,24 demonstrating the importance of delineating the cause of poor immunological reconstitution. This study was designed to investigate Tregs in immunological nonresponders. Tregs including Treg subpopulations were measured in peripheral blood in 3 groups of HIV-infected patients on cART with various levels of immunological reconstitution and in healthy controls. To assess the distribution of Tregs between blood and lymphoid tissue, the number of Foxp3⁺ cells was determined in lymphoid tissue. Finally, to determine the potential impact of Tregs on immunological reconstitution, the CD4 T-cell counts in patients were measured at inclusion in the study and after 1 year of follow-up.

METHODS

Study Design

A total of 77 HIV-infected patients and 34 healthy controls were included. All patients had been on cART for a minimum of 2 years, had CD4 nadir <250 cells per microliter (Fig. 1A), and HIV RNA was ≤20 copies per milliliter for at least 2 years before inclusion. Clinical characteristics are presented in Table 1. The patients were divided into 3 groups: (1) immunological nonresponders (INR; CD4 T-cell counts at inclusion in this study <200 cells/μL, n = 14), (2) intermediate responders (IR; CD4 T-cell counts 200–500 cells/μL, n = 33), and (3) responders (RES; CD4 T-cell counts >500 cells/μL, n = 30). Study subjects also participated in a study of Tc17 cells.²⁵ Exclusion criteria were acute or chronic infections other than HIV, including hepatitis B and C, malignancies, immunosuppressive treatment, and pregnancy. Patients were enrolled from the Departments of Infectious Diseases, University Hospitals of Copenhagen, Rigshospitalet and Hvidovre Hospital, Copenhagen, Denmark. Healthy subjects were recruited among hospital staff and matched on age, sex, and ethnicity. Informed consent was obtained from all participants after written and oral information. The study was performed in accordance with the ethical guidelines of the 1975 Declaration of Helsinki and was approved by the Local Ethical Committee (H-2-2009-089) and the Danish Data Protection Agency.

FIGURE 1: A, Increase in CD4 T-cell count from nadir over inclusion in the study to 1 year of follow-up in INR, IR, and RES. Red point indicates inclusion in the study. B–E, Gating strategy of regulatory T cells (Tregs) and subpopulations. Representative plots illustrating gating strategies of (B) CD4+ cells, (C) CD4+CD25+CD127low cells, (D) Tregs (CD4+CD25+ CD127lowFoxp3+), and (E) Tregs subpopulations: (1) activated Tregs (CD4+CD25+CD127lowCD45RA-Foxp3high), (2) resting Tregs (CD4+CD25+CD127lowCD45RA+Foxp3low), and (3) nonsuppressive Tregs (CD4+CD25+CD127lowCD45RA-Foxp3low). (F) Immunolabeling of Foxp3+ cells in a tonsil biopsy from an HIV-infected patient at ×10 magnification. The arrow points to a submucosal hotspot with increased number of Foxp+ cells in the interfollicular area of the lymphoid tissue.

TABLE 1: Clinical Characteristics of the Full Study Population Including Healthy Controls, Responders (CD4 Count >500 Cells/μL), Intermediate Responders (CD4 Count 200–500 Cells/μL) and Immunological Nonresponders (CD4+ Cell Count <200 Cells/μL), and of Those Having a Tonsil Biopsy Made

Blood Analysis

Ethylenediamine tetraacetic acid stabilized blood was used to obtain a full blood count and for flow cytometry. CD4 T-cell counts in patients were measured as a routine analysis before initiation of cART, at inclusion in this study, and after 1 year of follow-up (Fig. 1A). Plasma HIV-RNA levels were measured as described.²⁶ The detection threshold for HIV-RNA was 20 copies per milliliter. Heparinized blood was used to obtain peripheral blood mononuclear cells (PBMC). PBMC were isolated by means of density gradient centrifugation, and freshly isolated PBMC were used for analysis of Tregs and Treg subpopulations by flow cytometry. Frozen PBMC were used for intracellular staining of IL-10.

Flow Cytometry

Within the CD4⁺ and CD8⁺ cell compartment, recently activated (CD69⁺) and chronically activated (CD38⁺HLA-DR⁺) cells were determined (gating strategy added as Supplemental Digital Content,http://links.lww.com/QAI/A520). In brief, 100 μL of ethylenediamine tetraacetic acid blood was incubated with fluorescent dye–conjugated monoclonal antibodies at room temperature for 20 minutes. Erythrocytes were lysed with 2 mL of lysing solution (Becton Dickinson, Franklin Lakes, NJ) at room temperature for 20 minutes, and the samples were washed and resuspended in fluorescence-activated cell sorting flow (BD).^27,28

Tregs (CD4⁺CD25⁺CD127^lowFoxp3+), activated Tregs (CD4⁺CD25⁺CD127^lowFoxp3^highCD45RA), resting Tregs (CD4⁺CD25⁺CD127^lowFoxp3^lowCD45RA⁺), and nonsuppressive Tregs (CD4⁺CD25⁺CD127^lowFoxp3^lowCD45RA⁻) were determined.²⁸ Treg subpopulations are given as percentages of Tregs. Gating strategy is shown in Figure 1. PBMC were incubated with relevant surface marker antibodies for 20 minutes, followed by fixation and permeabilization (Human Foxp3 Buffer Set; BD), and incubated with antibodies against intracellular Foxp3 for 30 minutes. Intracellular staining for IL-10 in Tregs was performed in PBMC from 18 participants (INR = 6, RES = 6, and controls = 6) after stimulation with phytohemagglutinin (1 μg/μL⁻¹) for 24 hours according to manufacturer's instructions (Transcription Factor Buffet Set, BD). Percentage Tregs expressing IL-10 were determined. IL-10 gating was performed using fluorescence minus 1 (gating strategy is added as Supplemental Digital Content,http://links.lww.com/QAI/A520).

Monoclonal antibodies used to determine lymphocyte subsets were isotype control immunoglobulin G1 (IgG1)/IgG2a phycoerythrin (PE), IgG1 peridinin chlorophyll proteins-cyanine (PerCP-Cy5.5), IgG1/IgM fluorescein isothiocyanate (FITC), IgG1/IgG2b allophycocyanin (APC), IgG1 PE-Cy7, IgG1 APC-H7, Foxp3-PE, CD8⁻PerCP-Cy5.5, CD25-PerCP-Cy5.5, CD3⁻FITC, CD127-FITC, HLA-DR-APC, CD45RA-APC, IL-10-APC, CD69-PECy7, CD38-PE-Cy7, and CD4-APC-H7, all purchased from BD. Acquisition was performed using a 6-color FACS Canto, and data were processed using FACS Diva software (BD). For each sample minimum 50,000 cells were acquired.

Interleukin 10

To determine production of IL-10, 0.4 mL peripheral blood was cultured in 1.6 mL RPMI 1640 and phytohemagglutinin. The culture was incubated at 37° C for 24 hours after which the supernatant was collected and stored at −80° C until use. Cultures were performed in 85 participants (INR = 13, IR = 30, RES = 25, and controls = 17). IL-10 was measured in the supernatants by the Fluorokine Human MultiAnalyte Profiling Base Kit assay (R&D Systems, Minneapolis, MN) according to manufacturer's instructions and analyzed on Luminex 100 platform (Luminex Corp, Austin, TX). Samples were measured in duplicates.

Tonsil Biopsies

Tonsil biopsies were obtained from participants with preserved tonsils willing to donate a biopsy (n = 36). Clinical characteristics of participants with biopsy are presented in Table 1. After applying local anesthesia, a biopsy was obtained using Weils forceps and knife from the easiest accessible tonsil. Biopsies were performed by an experienced otorhinolaryngologist (K.S.). Biopsies were fixed in neutral buffered formalin (4%) overnight, followed by paraffin embedding. Serial 2 μm sections were cut and processed for staining and immunolabeling. Affinity-purified monoclonal mouse anti-human Foxp3 236A/E7was purchased from eBioscience (San Diego, CA). Heat-induced epitope retrieval was performed by Dako PT-link (Glostrup, Denmark). Thereafter, sections were incubated for 20 minutes with a dilution 1:40 of Foxp3 at room temperature, followed by staining using Dako EnVision Flex High pH Link detection system (EnVision FLEX + Mouse, EnVision FLEX HRP, EnVision FLEX DAB + Chromogen, EnVision Substrate Buffer) for Dako Autostainer Link Instruments. The sections were counterstained using EnVision Flex Hematoxylin and mounted by Pertex mounting media (Leica Microsystems, Wetzlar, Germany).

Scoring of Foxp3-positive cells was done by counting the number of positive cells in 3 interfollicular hot spot areas at ×40 magnification and calculating the average (Fig. 1F). Areas of high Foxp3 density were selected as hotspots. Immunolabeling was always nuclear and strong in intensity.²⁹ The pathologist was blinded to clinical characteristics. Biopsies from 3 participants did not include 3 representative areas of lymphoid tissue and were excluded from further analysis.

Statistical Analyses

Age and data on intracellular IL-10 production are given as median and interquartile range due to nonparametric distributions otherwise results are given as mean and 95% confidence intervals. Differences between groups were analyzed using analysis of variance/Kruskal–Wallis and unpaired t test/Mann–Whitney U test. Log transformation was applied where appropriate. As CD4 T-cell count and nadir differed between groups, a general univariate linear model was made and adjustment for CD4 nadir and age was performed. Comparisons between healthy controls and HIV-infected patients were only adjusted for age. All P values presented are adjusted except intracellular IL-10 where nonparametric analyses were used. Correlations were calculated by Pearson correlation. Comparison of CD4 T-cell count at inclusion and after 1 year was made with a paired t test. To determine the influence of Tregs on the change in CD4 T-cell count, linear regression was used. Two-tailed P < 0.05 were considered significant. Statistical analyses were performed using the Statistical Package for Social Sciences (SPSS version 11.5.0; SPSS Inc, Chicago, IL) and GraphPad Prism 6.0 (GraphPad Software, La Jolla, CA).

RESULTS

Tregs in Peripheral Blood From INR Are Extremely Prevalent and Have an Activated Phenotype

The percentage Tregs of CD4⁺ cells in INR, IR, RES, and controls were 10.2% (7.9–12.6), 7.2% (6.5–7.9), 5.4% (4.9–5.9) and 5.1% (4.5–5.6), respectively (Fig. 2A). INR had higher percentages of Tregs compared with IR, RES and controls (P = 0.0030, <0.0001, and <0.0001, respectively). Likewise, IR had higher percentages of Tregs compared with RES and controls (P < 0.001 and P < 0.0001), whereas RES were comparable with controls (P = 0.3190).

FIGURE 2: Distribution of percentage Tregs, absolute numbers, and subpopulations in INR (n = 14), IR (n = 33), RES (n = 30), and HIV-negative controls (n = 34). A, Percentages CD4+CD25+CD127lowFOX3+ Tregs of CD4+ cells. B, Absolute numbers of Tregs. C, Percentages activated Tregs (CD4+CD25+CD127lowFoxp3highCD45RA−) of Tregs. D, Percentages resting Tregs (CD4+CD25+CD127lowFoxp3lowCD45RA+) of Tregs. E, Percentages nonsuppressive Tregs (CD4+CD25+CD127lowFoxp3lowCD45RA−) of Tregs. Bars indicate mean, whiskers 95% confidence intervals of the mean.

In contrast, absolute numbers of Tregs were lower in both INR and IR compared with controls (P = 0.001 and P < 0.0001, respectively; Figure 2B).

The percentage of Tregs, which were activated in INR, IR, RES and controls, was 64.4% (57.8–71.0), 50.3% (44.8–55.8), 34.6% (30.3–38.8), and 28.6% (25.2–32.0), respectively (Fig. 2C). The percentage activated Tregs in INR was higher compared with IR, RES, and controls (P = 0.053, <0.0001, and <0.0001). Likewise, IR had higher percentage activated Tregs compared with RES and controls (P = 0.035 and P < 0.0001), and RES had higher percentage activated Tregs compared with controls (P = 0.037). Regarding resting Tregs, the opposite was found with the lowest percentage resting Tregs in INR compared with IR, RES, and controls (3.9% [2.3–5.5], 9.7% [7.2–12.1], 18.6% [14.6–22.6], and 23.4% [19.4–27.4], respectively; Fig. 2D). Similar findings were done for the percentage of nonsuppressive Tregs (Fig. 2E).

Poor Immunological Reconstitution Is Accompanied by High Immune Activation

INR had higher percentages recently activated CD4⁺ and CD8⁺ cells compared with controls, whereas IR and RES were closer to controls, (Figs 3A, B). Also, INR displayed the highest percentages of chronically activated CD4⁺ and CD8⁺ cells, and all HIV-infected individuals displayed elevated percentages compared with controls (Figs. 3C, D).

FIGURE 3: Percentages recently (CD69+) and chronically activated (HLA-DR+CD38+) CD4 T cells and CD8 T cells (A, B, C, and D, respectively) in INR (n = 14), IR (n = 33), RES (n = 30), and HIV-negative controls (n = 34).

Tregs Production of IL-10 Is Partly Preserved in INR but Reduced in RES

Concentration of IL-10 in stimulated whole-blood cultures in INR, IR, RES, and controls were 87 pg/mL (51–123), 293 pg/mL (210–375), 272 pg/mL (222–323), and 531 pg/mL (438–625), respectively. INR had lower production of IL-10 compared with IR, RES, and controls (P = 0.0021, <0.0001and <0.0001, respectively). Furthermore, IR displayed lower values compared with controls (P = 0.0005), and RES had lower production of IL-10 compared with controls (P < 0.0001).

To determine if low IL-10 concentration in INR was because of reduced production of IL-10 by Tregs, percentage Tregs producing IL-10 was measured. Percentage Tregs producing IL-10 was higher in INR compared with RES although reduced compared with controls (64.9% [55.4–81.2], 26.9% [16.4–37.9], and 82.8% [79.9–89.0], P = 0.0022 and P = 0.026, respectively; Fig. 4A).

FIGURE 4: A, Percentages Tregs producing intracellular IL-10 after stimulation with phytohemagglutinin in INR (n = 6), RES (n = 6), and controls (n = 6), line indicates median. B, Number of Foxp3+ cells/interfollicular hotspot in lymphoid tissue from tonsils in INR (n = 5), IR (n = 11), RES (n = 9), and HIV-negative controls (n = 8), line indicates mean. C–F, Linear regression was performed between percentage increase in CD4 T-cell count and percentage Tregs of CD4+ cells/number of Foxp3+ cells in lymphoid tissue in HIV-infected individuals with CD4 T-cell count <500 cells per microliter (C, D) and CD4 T-cell count >500 cells per microliter (E, F). Percentages Tregs producing intracellular IL-10 after stimulation with phytohemagglutinin in INR (dots, n = 6), RES (squares, n = 6), and controls (triangles, n = 6), line indicates median.

Foxp3⁺ Cells in Lymphoid Tissue Are Depleted in HIV-Infected Individuals With CD4 T-Cell Counts <500 Cells per Microliter

In tonsil biopsies, number of Foxp3⁺ cells/interfollicular hotspot in INR, IR, RES and controls were 63 (21–106), 69 (29–109), 97 (54–141), and 147 (100–193), respectively (Fig. 4B). Numbers of Foxp3⁺ cells in lymphoid tissue in INR and IR were significantly lower than in controls (P = 0.013 and P = 0.022), whereas RES were comparable with controls (P = 0.11). Differences were not found between groups of HIV-infected individuals.

Tregs in Blood and Foxp3⁺ Cells in Lymphoid Tissue Predict Immunological Reconstitution

To assess the impact of Tregs on immunological reconstitution CD4 T-cell counts in HIV-infected individuals were determined at inclusion in the study and after 1 year of follow-up, (Fig 1A). Seven patients had transferred out, and 1 patient was excluded due to treatment interruption. INR, IR, and RES had a mean CD4⁺ cell increase of 33 cells per microliter (P = 0.0428), 34 cells per microliter (P = 0.0168), and 19 cells per microliter (P = 0.1622), respectively.

In both INR and IR, a high percentage of Tregs in peripheral blood and high numbers of Foxp3⁺ cells in lymphoid tissue predicted increase in CD4 T-cell count after 1 year (r² = 0.1712; P = 0.0072 and r² = 0.2444; P = 0.0723; Figs. 4C, D). In contrast, in RES, a negative association was found between Tregs in peripheral blood and increase in CD4 T-cell count (r² = 0.1753; P = 0.0418), whereas no association was found between numbers of Foxp3⁺ cells in lymphoid tissue and increase in CD4 T-cell count (r² = 0.0539; P = 0.5802) (Figs. 4E, F).

DISCUSSION

This study was designed to investigate the role of Tregs in immunological reconstitution in HIV-infected individuals. In INR, the percentage of Tregs in peripheral blood was high, and most Tregs were activated. Furthermore, in INR, low production of IL-10 was found in stimulated whole-blood cultures, whereas near-normal percentages of Tregs producing IL-10 as measured by intracellular staining were found. In contrast, RES presented with percentages of Tregs, Treg subpopulations, and IL-10 production in whole blood comparable with healthy controls, whereas intracellular IL-10 production was reduced. Foxp3⁺ cells in lymphoid tissue from patients with CD4 T-cell counts <500 cells per microliter were depleted, whereas the number in RES was comparable with healthy controls. Importantly, the percentage of Tregs in peripheral blood and number of Foxp3⁺ cells in lymphoid tissue predicted future increase in CD4 T-cell count in INR and IR. These findings suggest the distribution of Tregs to be altered in HIV-infected INR and IR compared with RES and healthy controls, whereas the function seems partly preserved. As those patients with CD4 T-cell counts <500 cells per microliter, high percentage of Tregs in peripheral blood and high numbers of Foxp3⁺ cells in lymphoid tissue achieved higher CD4 T-cell increase after 1 year, Tregs may play a beneficial role in immunological reconstitution at low CD4 T-cell counts.

The definition of immunological nonresponse suffers from lack of consensus impeding comparison of findings. Immunological nonresponse is often defined as CD4 T-cell counts <200 cells per microliter while the treatment duration needed is variable.¹⁰ Furthermore, some study groups define nonresponders by the CD4⁺ cell increase in percentages.^30–33 However, there seems to be agreement that an optimal immunological response to cART includes a CD4 T-cell count >500 cells per microliter, mainly because HIV-infected patients with this level of immunological reconstitution have morbidity and mortality approaching that of HIV-negative individuals.^24,34 Consequently, patients with CD4 T-cell counts <500 cells per microliter are suboptimal responders. Suboptimal responders constitute a heterogeneous population because nonresponders are included within this group. The group of suboptimal responders with intermediate response (CD4 T-cell counts between 200 and 500 cells/μL) are poorly described, and they may have a morbidity and mortality distinct from both nonresponders and optimal immunological responders.^24,34,35 In this study, all participants had been on cART for a minimum of 2 years at the time of inclusion. We defined 3 groups INR, IR, and RES on the basis of current CD4 T-cell counts. Because low CD4 nadir and hepatitis C virus coinfection are associated with nonresponse, all patients had to have a low nadir <250 cell per microliter and to be hepatitis C virus negative.

The influence of Tregs on HIV infection and vice versa is controversial. However, because of an overall depletion of CD4⁺ cells in HIV infection, decrease in absolute numbers of Tregs are found in this study along others, in particular in patients with very low CD4⁺ cell counts.^10,36 However, Tregs seem to be spared compared with other CD4⁺ cells resulting in increasing percentages in treated and untreated patients.^14–18,37 In this study, INR presented with very high percentages of Tregs, IR with elevated percentages, and RES with percentages comparable with controls. This is in line with 2 recent studies.^36,38 Interestingly, in lymphoid tissue reduced numbers of Foxp3⁺ cells were found in INR and IR compared with RES and controls. A recent study of INR with CD4 T-cell counts <300 cells per microliter found reduced percentages of Tregs in gut mucosa.³⁶ Previous studies performed in untreated patients have reported similar or increased absolute numbers and/or percentages of Tregs in gut mucosa or tonsil tissue in patients, compared with HIV-negative individuals.^19–21,39 These data suggest accumulation of Tregs in lymphoid tissue in untreated progressing disease. One study examined Tregs in gut mucosa in 13 HIV-infected patients on cART and reported similar absolute number and percentage in patients and controls.¹⁹ However, those 13 patients had very variable CD4 T-cell counts. As the percentage of Tregs seem to be related to the CD4 T-cell count,^38,40 this may explain the lack of consensus. Also, it remains unknown if lymphoid tissue in the tonsils and the gut is comparable. Ideally, paired samples from tonsils and gut should be examined and both percentages and absolute numbers should be included in future studies to clarify the importance of Tregs in lymphoid tissue.

Peripheral blood Tregs in INR were activated, whereas IR, and in particular RES, presented with higher percentages of resting and suppressive Tregs resembling healthy controls. The potential role of a skewed distribution of Treg subpopulations is unknown. Patients with active systemic lupus erythematosus have been reported to have reduced level of activated Tregs and increased level of resting Tregs,²² implying a sufficient level of activated Tregs to be of importance for normal immune function. In HIV infection, a recent study reported reduced percentage of activated Tregs in primary HIV infection,⁴¹ suggesting that HIV influences the percentage of activated Tregs. Also, our group recently showed that the percentage of activated Tregs is maintained in nonprogressors.²⁶ Moreover, INR displayed an altered production of IL-10 in whole blood consistent with an earlier study.³¹ To investigate if low IL-10 production was because of few Tregs producing IL-10, intracellular staining of IL-10 was performed. INR were found to have near-normal levels of Tregs producing IL-10, whereas RES had reduced percentage. This is in line with our findings of high levels of activated Tregs in INR, and resting Tregs in RES.

Differences in Treg subpopulation may partly explain the difficulty in answering the question of whether Tregs overall are beneficial or harmful in HIV infection. Thus, it is debated whether high levels of Tregs are beneficial as downregulators of harmful and unspecific immune activation, or conversely, that high levels of Tregs are harmful in downregulating HIV-specific beneficial immune responses. In this study, INR presented with high immune activation as shown multiple times.¹⁰ However, no associations between immune activation and Tregs were found (data not shown). Interestingly though, in patients with suboptimal immunological reconstitution (INR and IR), Tregs in both blood and lymphoid tissue were positively associated with the increase in CD4 T-cell count during 1 year of follow-up, whereas this was not found in RES. This suggests that Tregs play a beneficial role in patients with suboptimal immunological reconstitution. One could speculate whether or not low CD4 T-cell count leads to recruitment of Tregs from lymphoid tissue to peripheral blood and to activation of Tregs and production of IL-10 as a response to low CD4 cell counts and high immune activation.

This study was limited by a relatively small sample size albeit a high fraction of patients with paired samples from peripheral blood and lymphoid tissue biopsy material was studied. Lower CD4 nadir and higher age were found in INR, both of which are known predictors of poor immunological reconstitution. Finally, Foxp3 was used as the sole marker of Tregs in lymphoid tissue. Activated T cells may express Foxp3, and we cannot rule out that some Foxp3⁺ cells in the tonsils are activated T cells possibly resulting in overestimation of Tregs in lymphoid tissue. However, we have previously compared the use of Foxp3 alone with CD4⁺CD25⁺CD127lowFoxp3⁺ to quantify Tregs in peripheral blood, finding the frequency of Tregs to be overestimated by approximately one-third in both healthy controls and in HIV-infected patients suggesting the potential bias when comparing groups to be of limited impact.²⁶

In conclusion, patients with suboptimal immunological reconstitution had high percentages of Tregs and activated Tregs. In contrast, INR had low number of Tregs in lymphoid tissue. These quantitative alterations of homeostasis were accompanied by low IL-10 in whole blood but near-normal percentages of Tregs producing intracellular IL-10. Finally, a high percentage of Tregs in blood and high number of Foxp3⁺ cells in lymphoid tissue predicted increase in CD4 T-cell counts after 1 year. Altogether, these findings are suggestive of an altered homeostasis and function of Tregs in patients with suboptimal immunological reconstitution. Further prospective studies are warranted to clarify the relation between Tregs and immunological reconstitution.

ACKNOWLEDGMENTS

The authors gratefully acknowledge the participants who volunteered and made this study possible. Professor, DMSc Niels Obel and the Danish HIV Cohort study are acknowledged for identifying patients.