Excessive conversion and impaired thymic output contribute to disturbed regulatory T-cell homeostasis in AIDS patients with low CD4 cell counts

Abstract

Objective: 

T regulatory (Treg) cells are a heterogeneous population that consists of CD4⁺CD25^lowCD45RA⁺ [naive Treg (nTreg) cells] and CD4⁺CD25^highCD45RA^– [activated Treg (aTreg) cells] subsets. We investigated the effects of HIV infection and HAART on distinct Treg subsets.

Methods: 

HIV-infected adult patients naive to HAART (n = 57), patients with acute HIV infection (n = 13), and healthy controls (n = 92) were recruited for a cross-sectional study. Patients receiving HAART were followed up in a longitudinal study.

Results: 

Compared with healthy controls, we observed a reduced proportion of nTreg cells and an elevated frequency of aTreg cells in peripheral blood from HAART-naïve patients. Moreover, nTreg cells showed a decreased CD31⁺ frequency, whereas aTreg cells showed an increased CD31⁺ frequency, indicating impaired thymic output and excessive conversion from nTreg to aTreg cells. nTreg and aTreg cells both displayed higher levels of Ki67⁺, reflecting hyperproliferation. The longitudinal study showed that HAART successfully recovered nTreg but not aTreg cell frequency. Higher baseline naïve CD4⁺ T-cell percentages were associated with faster reconstitution of nTreg cell frequency as well as CD4⁺ T-cell count.

Conclusion: 

Our data suggest that the disturbed homeostasis of Treg cells in HIV-infected patients is probably caused by excessive conversion from nTreg to aTreg cells, and impaired thymic output of nTreg cells. nTreg cells can be recovered by HAART, which was associated with baseline naive CD4⁺ T-cell percentages, indicating that reconstitution of nTreg cells may benefit from earlier antiretroviral treatment.

In Brief

Supplemental Digital Content is available in the text

Introduction

It has been shown that CD4⁺ T-cell depletion and clinical progression to AIDS are attributed to many factors, including HIV-1 infection, immune activation, and reduced thymic function ^[1–6]. In the past decade, regulatory T (Treg) cells have attracted a great amount of interest. Due to their potent inhibitory function on the activation and proliferation of CD4⁺ and CD8⁺ effector T cells ^[7–16], Treg cells appear to play opposing roles in the HIV disease progression. On the one hand, expansion of Treg cells in HIV infection might dampen effective antiviral immune responses, which may lead to a high viral load and make the patients more susceptible to other pathogens ^[9,13–15]. On the other hand, Treg cells are beneficial during AIDS progression by controlling excessive immune activation and limiting subsequent T-cell depletion ^[10,16].

In HIV-1-infected individuals, functional studies have demonstrated that Treg cells isolated from both peripheral blood and lymphoid tissues preserve the suppressive capacity ^[9,13]. Meanwhile, most phenotypic studies have shown a fall in the absolute number but a rise in the frequency of circulating Treg cells in chronic, treatment-naive patients when compared with healthy controls ^{[7,13,14,17–20]}. Moreover, despite immunological recovery and viral suppression after successful antiviral therapy, it has been found that the frequency of circulating Treg cells does not recover and even continues an increasing trend ^[19,20]. However, the results about Treg cells are still conflicting, due to a disparity in the markers used to identify Treg cells, as well as the heterogeneity in the HIV populations studied ^{[8,10,11,21–25]}.

Many studies have demonstrated that forkhead box P3 (FoxP3)⁺ Treg cells are heterogeneous and can be further subdivided into distinct subpopulations ^[26–30]. Human naive Treg (nTreg) cells with a CD4⁺FoxP3^lowCD45RA⁺ phenotype and upregulated CD62L and chemokine CXC receptor (CXCR)4, activated Treg (aTreg) cells with a CD4⁺FoxP3^highCD45RA^– phenotype with upregulated cytotoxic T-lymphocyte antigen (CTLA)-4, glucocorticoid-induced TNFR-related protein (GITR), CD95, HLA-DR and chemokine CC receptor (CCR)5, are both suppressive in vitro, whereas CD4⁺FoxP3^lowCD45RA^– T cells lack suppressive activity but perform a proinflammatory function ^[26]. The homeostasis of Treg cells is tightly regulated by nTreg production and aTreg depletion ^[26]. Similar to the total naive CD4⁺ T-cell fraction, nTreg cells derive from the thymus ^[27,31]. After stimulation, nTreg cells upregulate FoxP3 expression, proliferate, and differentiate to aTreg cells to expand the total Treg cell population. Subsequently, they exhibit suppression after their proliferation and conversion to aTreg cells, which die after proliferation and exertion of suppression ^[26].

It has been demonstrated that HIV infection can trigger immune activation and is associated with impaired thymic output ^[32,33]. In the present study, we revealed a reduced proportion of nTreg cells with an elevated proportion of aTreg cells in peripheral blood from HAART-naive patients. The disturbed homeostasis was associated with excessive Treg cell proliferation and conversion resulting from immune activation, and impaired thymic output of nTreg cells. More importantly, our longitudinal study showed that HAART successfully recovered nTreg cell frequency, whereas aTreg cell percentages remained at high levels.

Methods

Study participants

The study was approved by the Committee of Ethics at Beijing Ditan Hospital, Capital Medical University, Beijing, China. All human blood samples were collected with informed consent. Fifty-seven chronically HIV-infected, HAART-naive patients, 13 individuals with acute HIV infection, and 92 healthy controls were enrolled for a cross-sectional study from 2008 to 2011. Chronic-HIV-infected patients (88% men, 50/57) had a mean age of 37 years (range 23–67 years) with a mean CD4⁺ T-cell count of 157 ± 89 cells/μl (16–330 cells/μl) and a range of plasma viral loads (2.25–6.12 log₁₀copies/ml). Patients with acute HIV infection (all men) had a mean age of 34 years (range 21–61 years) with a mean CD4⁺ T-cell count of 508 ± 321 cells/μl (210–1122 cells/μl) and a range of plasma viral loads (4.41–6.87 copies/ml). Healthy controls (68% men, 63/92) had a mean age of 41 years (range 26–59 years). For a longitudinal study, the 57 drug-naive, chronically HIV-infected patients received HAART, which consisted of two nucleotide reverse transcriptase inhibitors (NRTIs) and a nonnucleoside reverse transcriptase inhibitor (NNRTI), and were further followed up every 12 weeks for 2 years.

Flow cytometric cell phenotypic analysis

All antibodies were purchased from BD Biosciences (San Jose, California, USA). Three-color and four-color flow cytometry were performed on fresh whole blood samples to identify Treg cell phenotype. Treg cells were gated on the basis of forward scatter/side scatter characteristics and CD25/FoxP3/CD45RA expression patterns as described previously ^[26–30]. After staining of surface markers, cells were fixed and permeabilized using the fluorescence-activated cell sorting lysing/permeabilizing buffers (BD Biosciences), followed by intracellular staining with FoxP3 and Ki-67 antibodies. Data were acquired on a CellQuest system (BD Biosciences) and analyzed using Flowjo software (Treestar, Ashland, Oregon, USA).

Plasma HIV-1 viral load and CD4⁺ T-cell count

Plasma HIV-1-RNA copy number was measured using the Standard Amplicor HIV Monitor assay, version 1.5 (Roche Diagnostics, Indianapolis, Indiana, USA). CD4⁺ T-cell count was determined in EDTA-treated whole blood and was performed using a standard flow cytometry technique with a Trucount tube in routine laboratories (BD Biosciences). The absolute numbers of lymphocytes stained for CD45/CD3/CD4 and CD45/CD3/CD8 were analyzed with MultiSET software (BD Biosciences). HIV-1-RNA levels and CD4⁺ T-cell counts were determined in a single laboratory using standard methodology that were included in the national QA programs twice a year.

Statistical analysis

All statistical analyses were performed using SPSS 16.0 software (SPSS Inc., Chicago, Illinois, USA). Data are expressed as mean ± SD. Flow cytometric data of Treg cell subsets from different stages of infection were analyzed using the one-way t-test. P values were derived from the one-way t-test to determine differences among several groups. The correlation between CD4⁺ T-cell count, plasma HIV-1 viral load, and the proportions of Treg cell subsets was analyzed with the Pearson correlation. For all comparisons, P less than 0.05 was considered statistically significant.

Results

Circulating naive T regulatory and activated T regulatory cell frequencies in HAART-naive, HIV-infected individuals

In line with most of the previous studies ^{[13,14,17,18]}, we found that compared with healthy controls, chronic HIV-infected and treatment-naive patients displayed higher percentages of circulating CD4⁺ CD25⁺ cells (19.43 ± 2.0%, n = 57 versus 7.38 ± 0.37%, n = 92, P < 0.001), with a concomitant decline in absolute number (30 ± 4 versus 52 ± 3 cells/μl, P < 0.001) due to lower absolute CD4⁺ T-cell count.

We further defined T-cell subpopulations in the light of the expression patterns of CD45RA and CD25 ^[26]. The results based upon CD25 were in accordance with those based upon FoxP3 (Fig. 1a). In comparison with nTreg cells, aTreg cells expressed higher levels of CTLA-4, inducible T-cell co-stimulator (ICOS), CD95, HLA-DR, and CCR5, but lower levels of CD31, CD62L, CD38, and CXCR4 (Fig. 1b). Despite a normal nTreg cell frequency, a slightly (about two-fold) but statistically significant increase of aTreg cell frequency was observed in acute HIV-infected individuals (Fig. 2a and b). In HAART-naive, chronic HIV-infected patients, the increase of aTreg cell percentage was more dramatic (about six-fold; Fig. 2a and b), whereas nTreg cell percentages (Fig. 2a and b) as well as absolute number (Supplementary Fig. 1a, http://links.lww.com/QAD/A302) were dramatically reduced.

Fig. 1: Delineation of CD4+ T cells into subsets by cell surface and intracellular molecules and phenotypic characterization of T regulatory subsets.(a) Fractions (Frs) I–V of CD4+ T lymphocytes in a representative healthy control and a representative patient with chronic HIV infection. CD4+ naive T regulatory (nTreg) and activated T regulatory (aTreg) cell subsets were defined as CD25/FoxP3lowCD45RA+ (Fr I) and CD25/FoxP3highCD45RA– (Fr II). Cytokine-secreting nonsuppressive T cells, memory/effector T cells, and naive CD4+ T cells were defined as CD25/FoxP3lowCD45RA– (Fr III), CD25/FoxP3–CD45RA– (Fr IV), and CD25/FoxP3–CD45RA+ (Fr V). Numbers indicate percentages of nTreg or aTreg cells. The results based upon CD25 and CD45RA in both healthy controls and HIV-infected patients were in accordance with those based upon FoxP3 and CD45RA. (b) Phenotype of nTreg and aTreg cell subsets in patients with chronic HIV infection.

Fig. 2: Disturbed homeostasis of T regulatory cells is correlated with disease progression during HIV infection.The frequency of T regulatory (Treg) cell subsets was evaluated in peripheral whole blood samples from healthy controls (n = 92), acute HIV-infected individuals (n = 13), and HAART-naive chronic HIV-infected patients (n = 57). (a) Representative dot plots of CD25 and CD45RA expression on CD4+ T cells in peripheral blood from acutely and chronically infected HAART-naive patients. Numbers indicate percentages of naive T regulatory (nTreg) or activated T regulatory (aTreg) cells. (b) The frequency of Treg cell subsets in peripheral blood from healthy controls, acutely and chronically infected HAART-naive individuals. (c) Frequency of Treg cell subsets in peripheral blood from healthy controls and patients with naive CD4 cell percentages more than 15 (n = 27) and 15 or less (n = 30). Data represent mean ± SD. P < 0.05 indicates significant difference.

Disturbed homeostasis of T regulatory cell subsets is correlated with disease progression

To investigate the relationship between the proportions of different Treg cell subsets and HIV replication or CD4⁺ T-cell count, we performed correlation analysis. However, we only found a weak correlation between nTreg cell frequency and CD4⁺ T-cell count (r = 0.340, P = 0.010). aTreg cell frequency was inversely correlated with CD4⁺CD45RA⁺CD25^– naive T-cell frequency (r = −0.466, P < 0.001), and nTreg cell frequency was strongly correlated with CD4⁺CD45RA⁺CD25^– naive T-cell frequency (r = 0.674, P < 0.001). The proportion of naive CD4⁺ T cells at baseline in AIDS patients has been reported to reliably predict immune reconstitution ^[34]; therefore, we stratified chronic HIV-infected patients (n = 57) based on the median naive CD4⁺ T-cell frequency of this cohort (∼15% of total CD4⁺ T cells). Compared to the group with naive CD4⁺ T-cell percentages more than 15, those with naive CD4⁺ T-cell percentages less than 15 showed a significantly lower frequency of nTreg cells and a higher percentage of aTreg cells (Fig. 2c). These data suggest that a disturbed homeostasis of Treg cell subsets is correlated with disease progression in HIV-infected and HAART-naive patients.

Hyperproliferation of T regulatory cells, excessive conversion from naive T regulatory to activated T regulatory cells, and impaired thymic output of naive T regulatory cells in HIV-infected patients

To investigate the mechanisms that regulate homeostasis of Treg cell subsets during HIV infection, we assessed expression of Ki-67 ^[35], a nuclear protein that is expressed at a higher level in proliferating cells. Compared with healthy controls, both acute and chronic HIV-infected patients displayed higher percentages of circulating Ki-67⁺ cells in the nTreg and aTreg subsets (Fig. 3a), which indicates that HIV infection promotes proliferation of Treg cells.

Fig. 3: Analysis of Ki-67 and CD31 expression indicates hyperproliferation of T regulatory cells, excessive conversion from naive T regulatory to activated T regulatory cells, and impaired thymic output of naive T regulatory cells.The frequency of Ki-67+ and CD31+ Treg cells was analyzed in peripheral blood from healthy controls (n = 30), acute HIV-infected individuals (n = 13), and HAART-naive chronic HIV-infected patients (n = 38). (a) Frequency of Ki-67+ Treg cells in peripheral blood from healthy controls, acute HIV-infected individuals, chronic HIV-infected HAART-naive patients with CD4 cell percentage more than 15 (n = 24), and chronic HIV-infected HAART-naive patients with CD4 cell percentage 15 or less (n = 14). (b) Frequency of CD31+ Treg cells in peripheral blood from healthy controls, acute HIV-infected individuals, chronic HIV-infected HAART-naive patients with CD4 cell percentage more than 15, and chronic HIV-infected HAART-naive patients with CD4 cell percentage 15 or less. Data represent mean ± SD. P < 0.05 indicates significant difference.

Next, we evaluated the expression of CD31 (platelet endothelial cell adhesion molecule-1, PECAM-1), a marker that is expressed on the new emigrants from the thymus ^[36]. We found an increase in the CD31⁺ fraction in aTreg cells from patients with acute and chronic HIV infection, which suggests a rapid transition of aTreg cells from newly emigrated nTreg cells from the thymus (Fig. 3b). However, a decrease in the CD31⁺ fraction in nTreg cells was noted only in the group with naive CD4⁺ T-cell percentages less than 15 (Fig. 3b), which suggests that, different from hyperproliferation and rapid transition of Treg cells emerging at an early stage, impairment of thymus function is associated with later progression of HIV infection.

Distinct response of naive T regulatory and activated T regulatory cell subsets to antiviral therapy

We analyzed the dynamic changes of Treg cell subsets in patients receiving HAART in a longitudinal study. Optimal suppression of HIV replication, restoration of CD4⁺ T-cell count (Fig. 4a), increased frequency of CD4⁺CD28⁺ T cells (Supplementary Fig. 2a, http://links.lww.com/QAD/A302) as well as decreased frequency of CD8⁺CD38⁺⁺, CD8⁺CD38⁺, and CD8⁺CD38⁺HLA-DR⁺ T cells (Supplementary Figs. 2b, 2c and 2d, http://links.lww.com/QAD/A302) were achieved in this cohort. In line with the results from above parameters, the proportions of nTreg cells started to rise after 24 weeks of treatment and reached a plateau after 48 weeks (Fig. 4b), with an increase in the absolute number of nTreg cells (Supplementary Fig. 3a, http://links.lww.com/QAD/A302). However, the proportions of aTreg cells remained at a higher level compared with healthy controls over a period of 96 weeks of antiviral treatment (Fig. 4b).

Fig. 4: Distinct response of naive T regulatory and activated T regulatory cell subsets to antiviral therapy.Profiles of plasma HIV-1 RNA level, CD4+ T-cell count (a); proportions of naive T regulatory (nTreg) and activated T regulatory (aTreg) cell subsets (b); CD31+ (c) and Ki-67+ (d) fractions in nTreg and aTreg cells in HIV-infected patients receiving HAART in the longitudinal study (n = 57). Dashed lines indicate the value of healthy controls. An asterisk indicates significant difference between HIV-infected patients and healthy controls.

Meanwhile, earlier than the recovery of nTreg cells, CD31⁺ fractions among nTreg cells were restored to the normal levels after 12 weeks of treatment (Fig. 4c). By contrast, the frequency of CD31⁺ fractions in aTreg cells started to decrease after 24 weeks of treatment (Fig. 4c). These data indicate that elevated turnover from nTreg to aTreg cells was tempered by HAART and the recovery of nTreg cells depended on the restoration of thymic function in these patients. However, despite a significant decrease in the Ki-67⁺ fraction in both nTreg and aTreg cells after treatment (Fig. 4d), the patients still exhibited a higher Ki-67⁺ fraction in Treg cells than in healthy controls, which is indicative of an imperfect effect of antiviral therapy on proliferation of Treg cells.

Recovery of naive T regulatory cells was correlated with naive CD4⁺ T-cell level before initiation of therapy

Although the average proportion of nTreg cells among CD4⁺ T cells reached a plateau after 48 weeks of treatment, approximately 50% (27/57) of the patients in the cohort failed to reach the level of the healthy controls at this time point. Thus, we analyzed the kinetic changes in nTreg cells after HAART in patients stratified on the basis of their naive CD4⁺ T-cell percentages before initiation of therapy. A significant difference in the kinetic recovery of nTreg cells was observed between the groups with lower (≤15%) and higher (>15%) naive CD4⁺ T-cell percentages at baseline. Compared to the group with lower naive CD4⁺ T-cell percentages, the group with higher naive CD4⁺ T-cell percentages achieved significantly faster reconstitution of nTreg cell frequency (Fig. 5c), CD4⁺ T-cell count (Fig. 5a), the percentage of recent thymic immigrants (Fig. 5b), and absolute number of CD31⁺ nTreg cells (Fig. 5e), as well as a significantly lower frequency of aTreg cells (Fig. 5d) and a marked decrease in Ki67⁺ nTreg cell percentages (Fig. 5f).

Fig. 5: Recovery of naive T regulatory cells is correlated with naive CD4+ T-cell level before initiation of therapy.Differences in immunological response to antiviral therapy among patients with naive T-cell percentage 15 or less (green line, n = 30) and more than 15 (red line, n = 27) at baseline in the longitudinal study, including CD4+ T-cell count (a); percentages of recent thymic immigrants (b); naive T regulatory (nTreg) cell frequency (c); activated T regulatory (aTreg) cell frequency (d); absolute number of CD31+ nTreg cells (e); and Ki-67+ nTreg cell frequency (f). Dashed lines indicate the value of healthy controls. An asterisk indicates significant difference between two groups.

Discussion

Previous studies have given conflicting results on the roles of Treg cells in disease progression in either treatment-naive HIV-infected individuals or antiviral-treated patients. Except for the heterogeneity in the HIV populations studied, the most likely explanation for this discrepancy is that human circulating CD4⁺ Treg cells are considered as a homogeneous population by the markers of FoxP3, CD25, and/or CD127 ^{[8,10,11,21–25,37]}. Recently, investigations have proved the heterogeneity of Treg cells, which could be affected by many factors at the stages of intrathymic generation of nTregs cells and extrathymic activation/conversion from nTreg to aTreg cells ^[26–30]. In this study, we reported a notable characteristic of reduced nTreg frequency and elevated aTreg cell proportion in HIV-infected HAART-naive patients, as well as distinct responses of Treg cell subpopulations to antiviral therapy.

One of the key findings of our study was the disturbed homeostasis of Treg cell subpopulations, which is characterized by increased aTreg frequency and ultimate exhaustion of nTreg cells. Previous data have shown that, following challenge by exogenous and endogenous antigens, nTreg cells are activated, then proliferate and eventually become aTreg cells, to expand the total Treg pool ^[26]. Thus, it is not surprising that we found that the increase in aTreg cell frequency started from the acute stage of HIV infection and was gradually aggravated with disease progression, which implies a rapid and sustained effect of HIV stimulation. aTreg cells are CD25/Foxp3^high; therefore, elevated aTreg frequency found in the present study is consistent with the results from most previous studies ^{[7,13,14,17–20]}, in which high levels of CD25 or Foxp3 are a hallmark of Treg cells. Moreover, continuous activation of Treg cells was demonstrated by an enhanced Ki67⁺ fraction in both nTreg and aTreg cells in HAART-naive patients. Meanwhile, enhanced conversion of Treg cells was further supported by an increased proportion of CD31⁺ aTreg cells and a decreased frequency of CD31⁺ nTreg cells.

Since the heterogeneity of Treg cells was demonstrated, several studies have reported an elevation in aTreg cell frequency in several common human diseases, including infectious diseases ^[38], type 1 diabetes ^[39], systemic lupus erythematosus ^[40], and graft-versus-host diseases ^[41]. Due to the higher proliferation characteristics and expression levels of CD95, aTreg cells rapidly undergo apoptosis after activation ^{[26,28,30,41]}. Thus, a continuous conversion from nTreg to aTreg cells is needed so as to maintain high levels of aTreg frequency. However, reduced nTreg cell frequency was not found in most of the studies, which indicates that the thymus can still compensate for activation-induced apoptosis of aTreg cells. Unlike most cases of infection, thymocytes are the target of HIV ^[42–44]. A deficit of recent thymic immigrants has been demonstrated to account for the reduction of naive T cells in HIV-infected individuals ^{[32,42,45–47]}. Here, we observed a strong correlation between nTreg and naive CD4⁺ T-cell frequencies. Moreover, when recent thymic output was assessed with CD31 as a direct marker, a decrease in recent thymic immigrants of nTreg cells was noticed in patients with chronic HIV infection and naive CD4⁺ T-cell percentages less than 15 (Fig. 3b). This suggests a failure of the thymus to generate sufficient nTreg cells at the late stage of disease progression. More importantly, we found that restoration of the CD31⁺ fraction in nTreg cells was much earlier than that of nTreg cells after HAART, which further proves the importance of the thymus in the maintenance of immune homeostasis. Nevertheless, we observed a declined proportion of CD31⁺ nTreg cells after 96 weeks of treatment (Fig. 4c). Considering a stable nTreg proportion at this point of time, homeostasis of nTreg cells could be maintained even at a decreased level of thymic output after treatment.

Furthermore, we found that the patients with higher CD4⁺ T-cell count at baseline achieved a better recovery of nTreg cells, but this difference was not statistically significant (data not shown). Previous studies have demonstrated that overall CD4⁺ T-cell reconstitution is dependent on the integrity of naive CD4⁺ T-cell population ^{[32,34,42,45–47]}. In a recent study, Schacker et al.^[34] identified that measurement of naive CD4⁺ T cells but not absolute CD4⁺ T-cell count predicted potential for immune reconstitution in HIV-infected patients. In accordance with this notion, we found that patients with higher baseline naive CD4⁺ T-cell frequency showed optimal CD4 cell reconstitution, and more importantly, faster reconstitution of nTreg cell frequency after HAART. Thus, thymic function before initiation of HAART might influence the restoration of nTreg cells. Of note, after nTreg cell frequency reached a plateau comparable to the levels of healthy controls after 48 weeks of treatment, CD4⁺ naive T-cell frequency kept increasing even after 96 weeks of treatment (Supplementary Fig. 4b, http://links.lww.com/QAD/A302). Whether earlier re-establishment of nTreg cell homeostasis may benefit long-term recovery of CD4 T cells should be further investigated.

Nevertheless, unlike the recovery of nTreg cells, the frequency of aTreg cells remained at a higher level over a period of 96 weeks of antiviral treatment. Despite a decrease in Ki-67⁺ fraction in the nTreg cell subset after 96 weeks, the frequency of Ki-67⁺ aTreg cells in patients was still approximately five-fold higher than that in healthy controls after 96 weeks of antiviral treatment (14.40 ± 5.18 versus 3.11 ± 2.37%, P < 0.001, Fig. 4d). In addition, higher proportions of CD31⁺ aTreg cells and cytokine-secreting nonsuppressive T cells (fraction III) were also observed in patients after successful antiviral therapy (Fig. 4c and Supplementary Fig. 4a, http://links.lww.com/QAD/A302). Moreover, compared with that in healthy controls, CD8⁺CD38⁺HLADR⁺ T-lymphocyte frequency remained higher in HAART-treated patients (Supplementary Fig. 2d, http://links.lww.com/QAD/A302), despite CD8⁺CD38⁺⁺ T-lymphocyte frequency being restored (Supplementary Fig. 2c, http://links.lww.com/QAD/A302). The above results indicate the presence of residual immune activation, which might cause hyperproliferation of aTreg cells and excessive conversion from nTreg to aTreg cells, and therefore a higher frequency of aTreg cells. Given that HIV infection is characterized by significant abnormalities in the gastrointestinal tract mucosa and high levels of circulating lipopolysaccharide, microbial products in the plasma might cause an increase in T-cell activation ^[48,49]. Moreover, higher frequencies of Treg cells have been observed in lymphoid tissues of HIV-infected patients ^[50,51]. It has also been shown that existing viral replication in remote sites of the body such as the lymph nodes might contribute to residual immune activation even in patients with good virological suppression ^[42,46,52]. Thus, it is possible that aTreg cells in lymphoid tissues or sites in affected organs might emigrate into the peripheral blood and lead to a persistent higher frequency of circulating aTreg cells.

During the preparation of the study, Simonetta et al.^[53] reported unaltered nTreg cell frequency and a decreased aTreg cell proportion in HIV-infected patients based on a similar strategy. This discrepancy between the two studies could be explained by differences in antibodies used for flow cytometric analysis, samples analyzed (fresh or cryopreserved), and populations studied. First, Simonetta's study was based on the patients with higher CD4 T-cell counts (>358 cells/μl). In contrast, we recruited the patients with relatively lower CD4 T-cell counts (16–330 cells/μl). Second, for dynamic responses of nTreg and aTreg cells to antiviral therapy, a cross-sectional study with 18 patients was conducted in the study of Simonetta et al., whereas we used a cohort with 57 chronically infected individuals in a 2-year longitudinal study. Third, we performed flow cytometric analysis on fresh peripheral blood mononuclear cells to avoid a reduction of Treg frequencies caused by cryopreservation ^[54,55]. More importantly, we provided evidence that excessive conversion and impaired thymic output are involved in disturbed homeostasis of Treg cells.

Conclusion

In summary, we showed disturbed homeostasis of Treg cell subpopulations in HIV-infected patients and their distinct responses to antiviral therapy. Our findings add to the understanding of the biological basis of the clinical benefit related to early HAART, which is in line with the recently recommendations from the International Antiviral Society-USA panel that antiretroviral therapy (ART) is recommended for all HIV-infected adults regardless of CD4 cell count ^[56], as well as the reports from China's National Free ART Program ^[57–59]. As a regimen based on two NRTIs and a NNRTI was used in our study, how protease inhibitors affect Treg cell subpopulations remains unknown. Further investigation could offer a way for immunological surveillance and provide possible novel targets for AIDS therapy through Treg cell manipulation.

Acknowledgements

The present work was supported by National Key Technologies R&D Program for the 12th Five-year Plan (grant 2012ZX10004–904), National Natural Science Foundation of China (grant 81161120427) and 215 program (grant 2009–2–13).

H.Z. performed the experiments, analyzed the data, and wrote the article. Y.H., C.S., J.H., J.Z., and Y.L. performed the experiments and analyzed the data. G.G., N.H., D.Y., and F.Z. collected and analyzed clinical data. H.Z. and H.Z. designed the study, analyzed the data, and wrote the article.

The authors would like to thank Dr Yao Zhang for his assistance in the preparation of this article.

Conflicts of interest

The authors have no conflicts of interest.