Increased levels of regulatory T cells (T_regs) in human immunodeficiency virus-infected patients after 5 years of highly active anti-retroviral therapy may be due to increased thymic production of naive T_regs

Abstract

This study determines levels of regulatory T cells (T_regs), naive T_regs, immune activation and cytokine patterns in 15 adult human immunodeficiency virus (HIV)-infected patients receiving prolonged highly active anti-retroviral therapy (HAART) who have known thymic output, and explores if naive T_regs may represent recent thymic emigrant T_regs. HIV-infected patients treated with HAART with a median of 1 and 5 years were compared with healthy controls. Percentages of T_regs (CD3⁺CD4⁺CD25⁺CD127^low), naive T_regs (CD3⁺CD4⁺CD25⁺CD45RA⁺) and activation markers (CD38⁺human leucocyte antigen D-related) were determined by flow cytometry. Forkhead box P3 mRNA expression and T cell receptor excision circles (T_REC) content in CD4⁺ cells were determined by polymerase chain reaction and cytokines analysed with Luminex technology. Levels of T_regs were significantly higher in HIV-infected patients compared with controls, both after 1 and 5 years of HAART (P < 0·001), despite fully suppressed HIV-RNA and normalization of both CD4 counts, immune activation and cytokine patterns. Furthermore, levels of naive T_regs were elevated significantly in HIV-infected patients (P < 0·001) and were associated with thymic output measured as the T_REC frequency in CD4⁺ cells (P = 0·038). In summary, T_reg levels in HIV-infected patients are elevated even after 5 years of HAART. Increased thymic production of naive T_regs may contribute to higher T_reg levels in HIV-infection.

Introduction

Recently, a subpopulation of CD4⁺ cells named naturally regulatory T cells (T_regs) has gained attention, as they are believed to play a central role in regulation of immune responses in chronic viral infections such as human immunodeficiency virus (HIV) [1–3]. HIV is characterized by a progressive loss of CD4⁺ cells leading to immune deficiency. Chronic immune activation is an important cause of CD4⁺ cell depletion in HIV infection and a strong prognostic marker for disease progression [4,5]. The function of T_regs is down-regulation of the immune response, and as such down-regulation of the chronic immune activation. Thus, the number and function of T_regs in HIV infection may be crucial for disease progression [1].

In HIV infection most studies demonstrate increased levels of T_regs compared with healthy controls [1,3,6–8]. T_reg levels are not normalized after a short course of highly active anti-retroviral therapy (HAART) [7,9]. Whether or not T_reg levels in HIV-infected patients remain high even after long-term HAART is as yet not clear. Neither has the precise mechanism leading to higher T_reg levels in HIV infection been examined.

Peripheral T_regs have been defined primarily as belonging to the memory T cell population, and maintenance of the T_reg pool is dependent on peripheral proliferation [10]. However, evidence of T_regs with naive CD4⁺CD25⁺CD45RA⁺ phenotype and immunosuppressive properties within the peripheral T_reg pool exists [11–13]. Whether or not naive T_regs represent recent thymic T_regs has not been clarified. In adults naive T_regs decline with age, as do thymic output and other naive T cells [11,12,14,15]. Understanding the importance of naive T_regs is only just beginning. Naive T_regs seem to represent a subpopulation of T_regs with unique self-generating capacities, and also seem to be more resistant to apoptosis [16]. Furthermore, naive T_reg levels may be critical for the suppressive function of the entire T_reg pool [17].

The objective of the present study was to determine levels of T_regs during long-term HAART. Levels of T_regs, percentages of activated cells and cytokine patterns were determined in adult HIV-infected patients after 1 and 5 years of HAART. Furthermore, the study determined levels of naive T_regs in HIV-infected patients and investigated the association between naive T_regs and thymic output.

Materials and methods

Human immunodeficiency virus-infected patients and healthy blood donors

A total of 25 male HIV-infected patients were included. Data on CD4 counts and thymic output have been presented previously [18,19]. Of the 25 patients frozen blood samples were available in 15 patients, and the present study demonstrates data on these 15 patients only. Patients were enrolled after 6–18 months (median 1 year) of HAART, all were naive to any anti-retroviral treatment prior to HAART. In nine of the 15 patients additional frozen blood samples were available after 54–66 months (median 5 years) of HAART. Clinical characteristics of the 15 patients at time of inclusion and at follow-up are presented in Table 1. For comparison 20 male HIV-negative age-matched controls were recruited from hospital personnel. A total of 10 controls were recruited for measurements of T_regs and naive T_regs, and an additional 10 controls were recruited for measurements of activation markers and cytokine profiles. The study was approved by the local ethics committee (KFE 01-181/00 and 11-142/04) and informed consent was obtained from all participants. Ethylenediamine tetraacetic acid blood was used to obtain a full blood count and for flow cytometry. Plasma was used to determine cytokine levels. Plasma HIV-RNA levels were measured with a polymerase chain reaction (PCR) quantitative kit (Amplicore Monitor HIV-1 PCR; Roche, Branchburg, NJ, USA) according to the manufacturer's instructions. Detection threshold was 20 copies/ml. Heparin blood was used to obtain peripheral blood mononuclear cells (PBMC) by means of density gradient centrifugation. Freshly isolated PBMCs were used for enrichment of CD4⁺ cells, determination of T cell receptor (TCR) excision circles (T_RECs) and quantification of forkhead box P3 (FoxP3) mRNA expression by PCR. Remaining PBMCs were frozen. Thawed PBMC samples were used for flow cytometric analysis of T_regs and naive T_regs.

Table 1.

Clinical characteristics of 15 human immunodeficiency virus (HIV)-infected patients at initiation of highly active anti-retroviral therapy (HAART), at inclusion in the study (1 year of HAART) and at follow-up (5 years of HAART).

Patient ID	CD4 count at initiation of HAART, cells/µl	Viral load at initiation of HAART, copies/ml	Age at inclusion, years	CD4 count at inclusion, cells/µl	Viral load at inclusion, copies/ml	HIV-related illness	HAART at inclusion	Months of HAART treatment at inclusion	CD4 count at follow-up, cells/µl	Viral load at follow-up, copies/ml	Months of treatment at follow-up
1	17	123 000	38	91	20	Herpes zoster, HIV-wasting	3TC, AZT, IND, RTV	9	n.f.d.	n.f.d.	n.f.d.
2	212	Missing	32	623	20		3TC, AZT, RTV, SQV	12	1023	<20	60
3	238	16 900	45	324	20		3TC, AZT, NEV	12	n.f.d.	n.f.d.	n.f.d.
4	24	78 500	45	116	20	Kaposi's sarcoma	3TC, AZT, NEV, NFV	17	247	<20	65
5	42	90 300	47	160	275	AIDS-dementia	3TC, D4T, RTV, SQV	15	502	20	63
6	234	>150 000	35	414	20	HIV-encephalitis	3TC, AZT, RTV, SQV	16	n.f.d.	n.f.d.	n.f.d.
7	264	48 400	28	409	20	Herpes zoster	3TC, D4T, NEV, NFV	18	489	<20	66
8	96	>150 000	37	410	20		3TC, AZT, NFV	16	1374	<20	64
9	136	>150 000	41	366	20		DDI, D4T, EFV	9	707	<20	57
10	66	>150 000	38	322	44		3TC, AZT, RTV, SQV	18	737	<20	66
11	288	>150 000	47	506	20		3TC, AZT, NEV, NFV	15	n.f.d.	n.f.d.	n.f.d.
12	289	21 000	30	515	263		DDI, D4T, ABV	18	n.f.d.	n.f.d.	n.f.d.
13	230	8 870	29	150	20		3TC, AZT, NEV	11	293	<20	59
14	5	>150 000	33	95	20	Kaposi's sarcoma, PCP	3TC, AZT, EFV	6	457	<20	54
15	119	77 600	45	472	47	Kaposi's sarcoma	3TC, AZT, ABV	9	n.f.d.	n.f.d.	n.f.d.
Mean		N.D.	37·9	324	57			13·4	617	N.D.	61·4
SEM		N.D.	1·7	45	22			1	112	N.D.	1·3

3TC, lamivudine; ABV, abacavir; AZT, zidovudine; D4T, stavudine; DDI, didanosine; EFV, efavirenz; IND, indinavir; n.d., not done; n.f.d., no follow-up data; NEV, nevirapine; NFV, nelfinavir; RTV, ritonavir; SQV, saquinavir.

Flow cytometry

Determination of activated lymphocyte subsets was performed on full blood, as described previously [19]. Percentages of CD4⁺ and CD8⁺ cells co-expressing CD69⁺, CD25⁺ or CD38⁺human leucocyte antigen D-related (HLA-DR⁺) were determined. Samples were analysed using a fluorescence activated cell sorter (FACScan) (Becton Dickinson (BD), San Jose, CA, USA) equipped with a 488 nm argon-ion laser. Data were processed using cellquest software (BD) and analysed using an isotype control strategy.

Determination of T_regs and naive T_regs was conducted on freshly thawed PBMC samples. T_regs were determined as CD3⁺CD4⁺CD25⁺CD127^low cells and naive T_regs as CD3⁺CD4⁺CD45RA⁺CD25⁺ cells (Fig. 1a, b). Approximately 100 000 cells were incubated with 10 µl of fluorescence-conjugated monoclonal antibodies at room temperature for 15 min. Cells were washed with phosphate-buffered saline (PBS) and resuspended in PBS supplemented with 10% CellFIX. Samples were analysed using a FACSCalibur (BD) equipped with a 488 nm argon-ion laser and a 635 nm red diode laser. Data were processed using cellquestpro software (BD) and analysed using a flourescence-minus-one strategy. The fluorescence of 50 000 cells was measured.

Fig. 1.

Representative flow cytometric panels showing the gating strategy on regulatory T cells (T_regs) and naive T_regs. (a) The expression of CD25 and CD127 gated on CD3⁺CD4⁺ lymphocytes. The R4 region includes CD25⁺CD127^low T_regs. (b) Naive T_regs were defined as CD3⁺CD4⁺ lymphocytes co-expressing CD25 and CD45RA. This is equivalent to region R5.

Monoclonal antibodies used were isotype control γ1-fluorescein isothiocyanate (FITC)/γ1-phycoerythrin (PE)/γ1-peridinin chlorophyll (PerCP)/γ1-allophycocyanin (APC), CD3-APC, CD4-PerCP, CD8-PerCP, CD25-FITC, CD38-PE CD45RA-PE, CD69-PE, CD127-PE and HLA-DR-FITC, all purchased from BD.

Enrichment of CD4⁺ cells

CD4⁺ cells were enriched from freshly collected PBMC using a magnetic affinity cell separator (Miltenyi Biotec, Bergisch Gladbach, Germany), as described previously [19]. The purity of sorted populations was determined by flow cytometry and was always >90%. Enriched CD4⁺ cells were used for PCR.

FoxP3 by reverse transcription–PCR

Measurement of Foxp3 mRNA and CD4 mRNA in separated CD4⁺ cells was performed by a reverse-transcribed real-time quantitative PCR (Taqman) assay. Preparation of mRNA was performed using an automated procedure (KingFisher), based on magnetic beads coated with oligo-dT. The eluted mRNA was reverse-transcribed using random hexamer priming and the Taqman assay was performed using the following intron-spanning primer probe sets: FoxP3 forward primer minus; GCTACGATGCAGCAGGAGC, FoxP3 reverse primer minus; CTGGGAAAATGGGACTGACC; FoxP3 Taqman probe: 6-FAM-AGGCTTCATCTGTGGCATCATCCGAC-TAMRA; CD4 forward prime minus; TCTTTCCCTGAGTGGCTGCT, CD4 reverse primer minus; GTCCCTTTTAGGCACTTGCTTCT; and CD4 Taqman probe: 6-FAM-TGCTGCAACTGGCGCTCCTCC-TAMRA [20]. Threshold cycle during the exponential phase of amplification was determined by real-time monitoring of fluorescence emission after cleavage of the sequence-specific probes by nuclease activity of the Taq polymerase. Equal amplification efficiency in the FoxP3 and CD4 assay was assessed by titration yielding essentially parallel tendency lines. As FoxP3 is expressed mainly by CD4⁺ cells, the FoxP3 signal was expressed relative to the CD4 signal [20] by the delta Ct method [21] and converted into relative units (RU). All measurements were performed in duplicate.

Determination of T_RECs in purified CD4⁺ cells

DNA was extracted from CD4⁺ cells using a salting-out procedure [22], and quantification of signal-joint T_RECs was performed by real-time quantitative PCR with the 5′-nuclease (TaqMan) assay, as described previously [19]. Samples were analysed in triplicate that did not vary by more than 10%, and the results were averaged.

Luminex

Cytokine analysis [interleukin (IL)-2, IL-4, IL-5, IL-10, interferon (IFN)-gamma] was performed with Luminex technology using an in-house assay as described previously [23], with some alterations. In brief, 50 µl plasma or standard diluted 1 : 10 in assay buffer (PBS containing 0·5% Tween 20 and 1% bovine serum albumin) were incubated for 1·5 h with a 50 µl suspension of capture-antibody-conjugated beads, 1500 beads per analysis per sample diluted in assay buffer. The beads were washed twice with 200 µl washing buffer (PBS containing 0·5% Tween) per well and incubated for 1·5 h with a mixture (50 µl) of biotinylated detection antibodies, each diluted 1:600 in assay buffer. Fifty µl of streptavidin–PE 20 µg/ml (Molecular Probes, Leiden, the Netherlands) were added to the wells and the incubation continued for additional 30 min. The beads were washed twice and resuspended in 100 µl washing buffer. After 15 min of shaking, samples were analysed on Luminex 100^(tm) (Luminex, Austin, TX, USA) according to the manufacturer's instructions. The results are given in median fluorescence intensities.

Statistical analysis

Data are given as means (±standard error of the mean) or medians (25%; 75% quartiles). Comparisons between HIV-infected patients and controls were made using Student's t-test or Wilcoxon's two-sample test. Assumptions for using Student's t-test were evaluated using the Shapiro–Wilks test for normality and visual inspection of residual plots for evaluating equal variances. Some of the outcome variables were log-transformed prior to the analysis because of skewed distribution. Correlations between efficacy measures were calculated using Spearman's correlation coefficient. A linear regression model was used to evaluate the association between thymic output and naive T_regs. A 5% significance level was used. The analyses were performed using Statistical Analysis Systems (version 9.1.3).

Results

Elevated levels of T_regs in HIV-infected patients after 1 and 5 years of HAART

The T_reg levels in 12 HIV-infected patients after 1 year of HAART were determined as percentages (%) of CD4⁺ cells co-expressing CD3⁺CD4⁺CD25⁺CD127^low as well as absolute numbers and for comparison in seven controls. In HIV-infected patients median % T_regs was 6·9% (6·3; 9·5%) versus 2·0% (1·4; 2·7%) of CD4⁺ cells in controls, (P < 0·001). Median absolute number of CD4⁺CD25⁺CD127^low cells in HIV-infected patients tended to be higher than in controls [23 cells/µl (11; 37 cells/µl) versus 10 cells/µl (7; 16 cells/µl), P = 0·090] (Fig. 2a, b). Higher absolute numbers of T_regs in patients were found despite significantly higher CD4 counts in controls [740 cells/µl (±59 cells/µl) versus 324 cells/µl (±45 cells/µl) in HIV-infected patients, P < 0·001] (Fig. 2c). No association was found between % T_regs and CD4 counts or months of treatment.

Fig. 2.

Regulatory T cells (T_regs) and CD4 counts in human immunodeficiency virus (HIV)-infected patients after 1 and 5 years of highly active anti-retroviral therapy (HAART). Percentages (a) and absolute numbers (b) of T_regs measured as CD3⁺CD4⁺CD25⁺CD127^low of CD4⁺ cells and CD4 counts (c) in HIV-infected patients after 1 year of HAART, after 5 years of HAART and for comparison in healthy controls. (a) Percentages of T_regs in HIV-infected patients were significantly different from controls both after 1 and 5 years of HAART (P < 0·001). (b, c) Absolute numbers of T_regs tended to be higher in HIV-infected patients after 1 year of HAART (P = 0·090) despite significantly higher CD4 counts in controls (P < 0·001). After 5 years of HAART CD4 counts in patients were normalized compared with control (P = 0·36) and absolute numbers of T_regs were significantly higher than controls (P = 0·011). Data in (b) and (c) are given as medians (25%; 75% quartiles). Differences between patients and controls were calculated using Student's t-test.

Furthermore, FoxP3 mRNA levels were measured in 13 HIV-infected patients and 10 controls. The median FoxP3 mRNA RU in HIV-infected patients was 0·31 (0·19; 0·35) and in controls 0·11 (0·09; 0·14, P = 0·038). No association was found between FoxP3 levels and CD4 counts or months of treatment.

To investigate if T_reg levels normalize after prolonged HAART, we measured CD3⁺CD4⁺CD25⁺CD127^low cells in nine patients after 5 years of HAART. Both median percentage of T_regs and absolute number of CD4⁺CD25⁺CD127^low cells at 5 years of HAART were significantly different from controls, 7·3% of CD4⁺ cells (5·2; 10·8%, P < 0·001) and 53 cells/µl (29; 58 cells/µl, P = 0·0011) (Fig. 2a, b). CD4 counts at 5 years of HAART had a mean of 617 cells/µl (±112 cells/µl) and were not different from controls (P = 0·36) (Fig. 2c).

Elevated levels of naive T_regs in HIV-infected patients after 1 year of HAART

Percentages and absolute numbers of naive T_regs were measured in 12 HIV-infected patients and 10 controls. The median percentage of naive T_regs in HIV-infected patients was significantly higher than in controls [2·6% (1·6; 3·1%) versus 0·7% (0·4; 0·8%) of CD4⁺ cells, P < 0·001] (Fig. 3a).

Fig. 3.

Naive regulatory T cells (T_regs) and thymic output in human immunodeficiency virus (HIV)-infected patients. Percentages of naive T_regs in HIV-infected patients after 1 year of highly active anti-retroviral therapy (HAART) compared with controls (a) and the association between thymic output and percentages (b) and absolute numbers (c) of naive T_regs. (a) Percentages of naive T_regs were higher in patients than controls (P < 0·001). (b) In a linear regression model a trend that percentages of naive T_regs depend on T cell receptor excision circles (T_RECs) frequency was found (P = 0·077). (c) Taking into account CD4 counts, the multiple regression showed absolute numbers of naive T_regs to depend significantly on both CD4 count (P < 0·001) and T_REC frequency (P = 0·036). The difference between patients and controls in (a) was calculated using Student's t-test.

The median absolute number of naive T_regs in patients was 8 cells/µl (4; 15 cells/µl) compared with controls, 5 cells/µl (3; 6 cells/µl, P = 0·18). No association was found between percentage of naive T_regs and CD4 counts.

Higher levels of naive T_regs in HIV-infected patients depend on thymic output

To examine if naive T_regs may represent recent thymic emigrant T_regs, the T_REC frequency in CD4⁺ cells was determined (published previously in [19]). In a regression model a trend that proportions of naive T_regs depend on T_REC frequency was found (P = 0·077) (Fig. 3b). Analysing absolute numbers and taking into account CD4 counts, the multiple regression showed absolute numbers of naive T_regs to depend significantly on both CD4 count (P < 0·001) and T_REC frequency (P = 0·036) (Fig. 3c). In both regression models, however, one observation was very influent on the result.

Immune activation

Immune activation was measured in all 15 HIV-infected patients after 1 year of HAART and in 10 controls (Table 2). Because T_regs inhibit activated T cells, the association between T_reg levels and immune activation was investigated. A negative correlation was found between FoxP3 RU and percentage of CD4⁺CD38⁺HLA-DR⁺ cells (r = −0·57; P = 0·041) (Fig. 4), but no association was found between percentage of T_regs and CD4⁺CD38⁺HLA-DR⁺ cells or between percentage of T_regs or FoxP3 levels and percentage of CD8⁺CD38⁺HLA-DR⁺ cells. Duration of treatment (months) was associated negatively with percentage of CD8⁺CD38⁺HLA-DR⁺ cells (r = −0·65; P = 0·008) and tended to be associated with percentage of CD4⁺CD38⁺HLA-DR⁺ cells (r = −0·49; P = 0·063).

Table 2.

Activation markers and cytokine levels in 15 human immunodeficiency virus (HIV)-infected patients after 1 year of highly active anti-retroviral therapy and in 10 controls.

Activation markers	Patients, %	Controls, %	P
CD4+CD25+ cells	4·4 (3·4; 10)	4·8 (3·1; 6·6)	0·96
CD4+CD69+ cells	4·8 (3·2; 10)	4·2 (4·2; 8·5)	0·98
CD4+CD38+HLA-DR+ cells	3·4 (2·4; 9·7)	3·1 (2·7; 5·4)	0·31
CD8+CD25+ cells	0·4 (0·02; 1·1)	1·3 (0·5; 2·1)	0·96
CD8+CD69+ cells	3·4 (2·3; 5·7)	5·7 (3·3; 6·8)	0·24
CD8+CD38+HLA-DR+ cells	5·5 (3; 6·4)	3·1 (1·9; 4)	0·21
Cytokine	Patients, MFI	Controls, MFI	P
IFN-γ	116 (92; 192)	148 (95; 181)	0·64
IL-2	160 (113; 223)	226 (131; 259)	0·31
IL-4	208 (163; 246)	245 (201; 302)	0·18
IL-5	214 (176; 277)	298 (184; 351)	0·71
IL-10	163 (139; 243)	224 (126; 244)	0·68

All activation markers are given as medians (25%; 75% quartiles). Cytokine levels are given as medians (25%; 75% quartiles) of median fluorescence intensity (MFI). IFN, interferon; IL, interleukin; HLA-DR, human leucocyte antigen D-related.

Fig. 4.

Forkhead box P3 relative unit (RU) levels in human immunodeficiency virus (HIV)-positive patients after 1 year of highly active antiretroviral therapy were negatively associated with the percentage of activated CD4⁺ cells (r = −0·57; P = 0·041). Spearman's correlation coefficient was used.

Cytokine profiles in HIV-infected patients and healthy controls

To investigate cytokine profiles in HIV-infected patients after 1 year of HAART plasma levels of the T helper type 1 (Th1) cytokines IL-2 and IFN-γ and the Th2 cytokines IL-4, IL-5 and IL-10 were measured in 15 patients and in 10 controls (Table 2). Th1 and Th2 cytokine patterns may be controlled by T_regs, which themselves act in a cytokine-dependent manner. We found no association between T_reg levels measured by flow cytometry or FoxP3 expression and the cytokines IFN-γ, IL-2, IL-4, IL-5 or IL-10.

Discussion

The purpose of the present study was to determine levels of T_regs in adult HIV-infected patients receiving prolonged HAART. We demonstrate elevated levels of T_regs despite 5 years of HAART, suppressed viral loads and normalized CD4 counts, immune activation and cytokine patterns. Thus, the reason for higher T_reg levels is not known and does not seem to be driven by immune activation or viraemia. Furthermore, significantly higher levels of naive T_regs in HIV-infected patients compared with healthy controls were found. Levels of naive T_regs were associated with thymic output, suggesting higher T_reg levels in HIV-infection explained in part by increased thymic production of naive T_regs.

In recent years T_regs have become acknowledged as important players in maintaining immunological self-tolerance [10,24]. T_regs share many characteristics with activated CD4⁺ cells and both cell types express the IL-2 receptor CD25⁺[25]. T_regs, however, express CD25 with higher intensity and can hereby be distinguished from activated cells [26]. Recently, lower levels of the α-chain of the IL-7 receptor CD127 have been shown to distinguish T_regs more accurately [7,27]. Another method for quantifying T_reg content is measuring the expression of the transcription factor FoxP3 responsible for the regulatory capabilities of the cells [28–31].

In HIV infection immune activation is a strong prognostic marker of disease progression [4]. Immune activation set point during early HIV infection has been shown to predict CD4 decline [5]. T_regs may therefore play a central role in HIV disease pathogenesis [1–3]. Most studies demonstrate increased levels of T_regs in blood and lymphoid tissue of HIV-infected patients compared with healthy controls [1,3,6–8,32]. However, the suppressive capacity of T_regs isolated from HIV-infected patients does not seem altered [1,3,29], even if it has been suggested that the function of T_regs in HIV-infected patients with more advanced disease is impaired [1]. Absolute numbers of T_regs decline with CD4 depletion in HIV infection, as T_regs are themselves infected by HIV. Thus, proportions of T_regs reflect the kinetics in HIV infection more clearly, and T_regs seem to decline independently from other CD4⁺ cells [29]. Higher proportions of T_regs have been found both in treatment-naive patients and patients treated with HAART. Interestingly, two recent studies of initiation of HAART with a median follow-up period of 24 and 36 weeks, respectively, showed no effect on T_reg level despite immunological recovery and viral suppression [7,9]. Our data are in line with these findings. We show elevated levels of T_regs measured both as percentages of CD3⁺CD4⁺CD25⁺CD127^low cells and FoxP3 mRNA RU in HIV-infected patients after 1 year of HAART. In the previously published studies, immune activation remained significantly higher in patients compared with controls. Interestingly, in this study, immune activation was normalized in patients. We did, however, find a negative association between FoxP3 RU and percentage of CD4⁺CD38⁺HLA-DR⁺ cells, but not between percentage of CD3⁺CD4⁺CD25⁺CD127^low cells and activated CD4⁺ or CD8⁺ cells. We found that proportions of T_regs after 5 years of HAART were unchanged despite normalization of CD4 counts. Thus, the expansion of T_regs in HIV infection may be irreversible and may not reflect viral load, CD4 depletion or level of immune activation.

During HIV infection, dysregulation of the cytokine network has been suggested with a shift towards type 2 cytokines associated with disease progression. In contrast, a balanced cytokine pattern may be associated with long-term non-progression [29,33,34]. Type 1 and type 2 cytokine patterns have been proposed to be controlled by T_regs, which themselves may depend upon cytokines [6,35]. We did not find any association between types 1 or 2 cytokines and T_reg levels; nor did we find different levels of cytokines in patients and controls, indicating that the patients investigated in this study are not so immunologically disturbed.

The precise mechanism underlying T_reg expansion in HIV infection is not yet known. The expansion of T_regs may be a result of increased peripheral proliferation in the T_reg pool, prolonged survival of existing T_regs, or may be due to increased generation of recent thymic emigrant T_regs. Traditionally, peripheral T_regs are considered as belonging to the memory T cell population and the majority of T_regs in adults do express the memory marker CD45RO [10,12]. However, a subpopulation of T_regs with the naive phenotype CD4⁺CD25⁺CD45RA⁺ is detectable within the peripheral T_reg pool and with suppressive capacity comparable with memory T_regs[11–13]. Naive T_regs may be de novo generated cells released recently from the thymus, supported by the demonstration of both longer telomeres and higher T_REC contents in naive T_regs, similar to telomere length and T_REC contents of other naive CD4⁺ cells [10,11]. T_RECs are generated as a by-product of the TCR rearrangement process in the thymus, and are enriched in newly generated cells [14]. However, the T_REC frequency is dependent not only upon thymopoiesis, but also on peripheral T cell proliferation and cell death and therefore must be interpreted with caution [36]. Interestingly, naive T_regs seem to represent a subpopulation of T_regs with special proliferative activity and unique self-generating capacities being more resistant to apoptosis [10,16]. In fact, naive T_reg levels may be critical for the suppressive function of the entire population of T_regs. In patients with relapsing–remitting multiple sclerosis, T_regs are functionally impaired and less potent suppressors of T cells, and this T_reg defect has been proposed to be explained by low numbers of naive T_regs[17]. Thus, naive T_regs from the thymus may be essential to maintain full suppressive capacity of the entire T_reg pool.

The thymus retains some ability to generate T cells in adults. Thus, in HIV-infected patients, higher T_REC frequencies correlate with higher absolute and naive CD4 counts and are found in patients with a better immunological response to anti-retroviral therapy [14,15,19,37–41]. Whether or not the adult thymus also plays a role in maintaining T_reg levels contributing with naive T_regs has not been clarified. We found increased levels of naive T_regs in HIV-infected patients compared with controls, and naive T_regs were associated with increased thymic output, suggesting that the adult thymus exports both naive CD4⁺ cells and naive T_regs to the periphery. Pregnancy supports the theory that the thymus is involved in regulating both central and peripheral tolerance even in adulthood. Here an immunosuppressing state develops in the maternal immune system in order to secure tolerance of fetal paternal antigens [42,43]. The thymus may be responsible for the development of this tolerance undergoing temporary changes in size and function and modulating both thymic export of naive T cells (central tolerance) and T_regs (peripheral tolerance) [42,44–46].

In conclusion, we found elevated levels of T_regs even after 5 years of HAART in HIV-infected patients with suppressed viral replication and normalization of both CD4 counts and immune activation. We present data suggesting that elevated levels of T_regs may be due to increased thymic production of naive T_regs. Thus, levels of naive T_regs were associated with thymic output measured as the T_REC frequency. However, it still remains unanswered whether T_regs in HIV-infection play a harmful or useful role: harmful if T_regs suppress HIV-specific effectors and thereby limit the body's immune response to HIV; useful if T_regs help to down-regulate the chronic immune activation characteristic of HIV infection. Further studies to enlighten this are warranted. Such studies may include studies on long-term non-progressors or on HIV-infected women during pregnancy.