How to cite this article: Saison J, Demaret J, Venet F, Chidiac C, Malcus C, Poitevin-Later F, Tardy J, Ferry T, Monneret G. CD4+CD25+CD127− assessment as a surrogate phenotype for FOXP3+ regulatory T cells in HIV-1 infected viremic and aviremic subjects. Cytometry Part B 2012; 84B: 50–54.
Although likely pivotal, the role of regulatory T cells (Tregs) in HIV pathogenesis remains elusive. This can be partly explained by analytical issues regarding their phenotypic identification in clinical studies. Instead of intracellular FOXP3 staining, CD4+CD25+CD127− phenotype has been proposed as an alternative to identify Tregs in clinical samples. However, its use remains controversial in viremic patients. Therefore, the objective of the present study was to assess the correlation between frequencies of CD4+CD25+CD127− and CD4+CD25+FOXP3+ lymphocytes in viremic and matched aviremic HIV-infected patients.
Peripheral blood was collected from HIV-1 infected patients. Eleven viremic patients (Viral Load > 40 copies/mL) were matched (age, sex, CD4+ cell number) with 8 aviremic patients under highly active antiretroviral therapy (HAART). Fresh whole blood was immediately stained to analyze by flow cytometry the correlation between CD4+CD25+CD127− and the reference phenotype CD4+CD25+FOXP3+ lymphocytes in the same tube (four color staining CD4/CD25/CD127/FOXP3 for concomitant analysis of cell surface and intracellular markers).
In both groups, no significant differences were observed when comparing CD4+CD25+CD127− and CD4+CD25+FOXP3+ cell frequencies. In line, a strong correlation between CD4+CD25+CD127− and CD4+CD25+FOXP3+ lymphocyte percentages was observed in the whole patient population (r: 0.948, P < 0.001) or each group separately: aviremic (r: 0.968, P < 0.001), viremic (r: 0.9, P < 0.001). Finally, we found that most CD4+FOXP3+ cells were indeed CD25+CD127−, both in viremic and aviremic groups (88.5% and 90.9%, respectively).
HIV infection severely impacts immune response homeostasis, leading to development of chronic immune activation and gradual immune cell depletion. Regulatory T cells (Tregs) are known to modulate the host response to viral infections (1, 2), but their role in HIV pathogenesis is still controversial. They are either supposed to suppress protective anti-HIV cell mediated immunity (3–5) or to control immune activation (6, 7), which plays a major role in disease progression. More recently, Moreno et al. also reported that Tregs may control viral replication during the early phase of infection (8). Several reasons may explain these discrepancies. On the one hand, the heterogeneity of patients' population has to be taken into account (level of viremia, type of progressors, treatments duration …). On the other hand, analytical issues may have contributed to discordances: markers used for Tregs phenotyping, staining of whole blood or purified cells, sample conservation (i.e., cryopreserved or not) (9, 10).
Today the most widely accepted phenotype for Tregs is the coexpression of CD4, CD25 (α-chain of the IL-2 receptor), and of FOXP3 (11). FOXP3 is the key transcriptional factor for Tregs. However, its intracellular staining is hardly usable on a daily clinical routine basis in large series of samples (several steps of incubation/washing, long incubation times, use of isotype control). As a consequence, some authors took advantage of an alternatively described phenotype: the coexpression of CD4 and CD25 associated with the absence or the low expression of CD127 (α-chain of IL-7 receptor) (i.e., CD4+CD25+CD127−) (12). A good correlation between these two phenotypes was reported in healthy subjects (13) and in septic patients (14). However, a debate remains in HIV infected patients. Indeed, this correlation was found in aviremic patients but not in viremic ones (15). In contrast, other studies, based on cryopreserved or purified cells, found an acceptable correlation in viremic patients (16, 17).
Considering the potential clinical interest of monitoring Tregs during HIV infection (18–20), and the need to further investigate their role in the different steps of disease progression, (especially during the phase of viremia), the objective of the present study was to assess in fresh samples the correlation between frequencies of CD4+CD25+CD127− and CD4+CD25+FOXP3+ lymphocytes in viremic and matched-aviremic HIV infected patients.
MATERIALS AND METHODS
Study Population and Blood Samples
HIV-1 infected patients from the Hospices Civils de Lyon HIV cohort, located in Croix-Rousse Hospital (Lyon, France) were consecutively included in the study between February 15th and 29th, 2012. These patients were concomitantly matched (sex, age, CD4+ cell number) with aviremic patients under highly active antiretroviral therapy (HAART). Analyses were performed on available residual blood after routine analysis (i.e., CD4+ lymphocyte count) at the Cellular Immunology Laboratory (Hôp. E. Herriot, Lyon, France).
Expression of cell surface and intracellular markers was assessed on fresh whole EDTA-anticoagulated peripheral blood. Flow cytometry protocol has been previously described (14). Briefly, lymphocytes were stained with monoclonal antibodies according to manufacturer's recommendation: ECD-labeled anti-CD4, PE-labeled anti-CD127, PC5-labeled anti-CD25 (Beckman Coulter, Hialeah, FL), FITC-labeled anti-FOXP3 (eBioscience, San Diego, CA). Appropriate isotypic control (Rat IgG 2a-FITC, eBioscience) was used to evaluate nonspecific staining and set up positive threshold for FOXP3. A four color staining (CD4/CD25/CD127/FOXP3) for concomitant analysis of cell surface and intracellular markers was performed. Cell surface staining was performed first. Then, after fixing and permeabilizing, samples were stained for intracellular FOXP3. Red blood cells were lysed using Versalyse reagent (Beckman-Coulter, Miami, FL). Data were acquired on a NAVIOS flow cytometer and analyzed using the NAVIOS software (Beckman Coulter). Results are expressed as percentages of CD4+CD25+CD127− or CD4+CD25+FOXP3+ cells out of the total CD4+ lymphocytes.
Viral Load (VL) Quantification
After extraction of 0.6 mL of plasma, reverse transcription, PCR amplification, and detection/quantification reactions were performed using Abbott RealTime HIV-1 assay and m2000 system platform (Abbott Molecular, Des Plaines, IL). The detection limit was 40 copies/mL.
Normality of the variables was assessed using the Kolmogorov-Smirnov test. Exact Fisher test for categorical variables, and T-test for continuous variables were performed. Correlations between CD4+CD25+CD127− and CD4+CD25+FOXP3+ cell frequencies were studied using the Pearson's correlation coefficient test. The statistical analyses were performed using SPSS software (version 16.0, SPSS, Chicago, IL). A significance was assumed for P < 0.05.
Population at Baseline
Nineteen HIV-1 infected patients were included in the study. None of them were coinfected with hepatitis B or hepatitis C viruses. Of them, 11 viremic patients [men n = 8 (73%), median age = 33 years; Inter Quartile Range (IQR): 22-61, median CD4+ cell count = 363 cells/mm3, IQR: 115–482] were enrolled. In parallel, 8 patients receiving HAART [men n = 6 (75%), median age = 40 years, IQR: 24–59, median CD4+ cell count = 357 cells/mm3, IQR: 84–1196] were recruited and classified in the aviremic group. There were no statistical differences at baseline between the two groups either for sex (P = 0.91), age (P = 0.70), or absolute CD4+ cell counts (P = 0.33). Aviremic patients had no detectable viremia (VL < 40copies/mL), whereas median VL in the viremic group was 27304 copies/mL (IQR 3,500–138,336) (P < 0.001).
Gating strategy and illustrative examples of CD4-CD25-CD127 and CD4-CD25-FOXP3 stainings are shown in Figure 1. Respective phenotype frequencies, percentages of CD4+CD25+CD127− lymphocytes expressing FOXP3, and percentages of CD4+CD25+FOXP3+ cells not expressing CD127 are reported in Table 1, as individual values. Percentages of Tregs identified by the two phenotypes, i.e. CD4+CD25+CD127− and CD4+CD25+FOXP3+ cells were not statistically different in the viremic group (P = 0.74), in the aviremic group (P = 0.36), or in the whole population (P = 0.45). In line, we found a strong correlation between frequencies of CD4+CD25+CD127− and CD4+CD25+FOXP3+ cells, when considering the whole study population (r: 0.948, P < 0.001) or each group separately: aviremic (r: 0.968, P < 0.001), viremic (r: 0.94 P < 0.001). Another important observation was that most CD4+FOXP3+ cells were indeed found to be CD25+CD127−. Similar results were obtained for the viremic group and the aviremic group (88.5% and 90.9%, respectively, Table 1).
Table 1. Percentages of Lymphocytes in Regard with Different Treg Phenotypes According to the Virologic Status of HIV-1 Infected Patients
% of CD25+CD127- (among CD4+)
% of CD25+FOXP3+ (among CD4+)
% of FOXP3+ (among CD4+CD25+CD127−
% of CD127- (among CD4+CD25+FOXP3+)
IQR, inter quartile range; P values: viremic vs. aviremic groups (student T test).
In this observational study we compared two Tregs phenotypes in an HIV-1-infected cohort. In both viremic and aviremic HIV-infected subjects, our purpose was to investigate whether CD4+CD25+CD127− phenotype identifies FOXP3+ Tregs in these particular clinical contexts.
Indeed, we observed that most CD4+CD25+CD127− cells expressed FOXP3 (Figs. 1a and 1b), either in the viremic (median: 88.50%, IQR: 84.10–92.90), or in the aviremic group (median: 90.9%, IQR: 80.80–97.70, Table 1). Conversely, using a backgating strategy (Figs. 1c and 1d), we highlighted that most CD4+CD25+FOXP3+ cells were negative for CD127 expression (Table 1). Our results are consistent with two other studies (16, 17), in which a close correlation (r: 0.81, P < 0.001, and r: 0.96 P < 0.0001, respectively) between the two phenotypes was observed in viremic patients. Nevertheless, these studies did not include aviremic patients. Overall, these results are important because some authors previously used the CD4+CD25+CD127− phenotype to characterize Tregs (without assessing the FOXP3+ intracellular staining). Our results confirmed that their conclusions on roles of Tregs in the pathophysiology of HIV infection are relevant for both the viremic and the aviremic subjects (21, 22).
However, the present report contrasts with Del Pozo and colleagues' results (15). They observed a lack of correlation between the two phenotypes in viremic subjects. This discrepancy could be attributed to different study protocols (both analytical and regarding inclusion criteria). First, we used fresh whole blood samples rather than isolated peripheral blood mononuclear cells. Second, we included patients with lower absolute CD4+ value, and not only those presenting with CD4+ count higher than 350 cells/mm3. Third, in our study, there was no statistical difference in age between the two groups while the aviremic group was older than the viremic one in Del Pozo's study (median age 43 and 31 years respectively, P = 0.001). This may explain the higher CD4+CD25+FOXP3+ frequency observed in the aviremic group (4.89 versus 3.90, P = 0.039).
Our study has some limitations. First, we did not specifically analyzed Tregs phenotypes and frequencies in long-term nonprogressors or in case of early HIV infection. We cannot rule out that some specificity may relate to these conditions. Second, the number of patients included is relatively modest. However, the very strong correlations between CD4+CD25+ FOXP3+ and CD4+CD25+CD127− percentages obtained in such small cohort tend to support our conclusion. Third, it has to be mentioned that a small percentage of FOXP3+ cells was not included in the CD4+CD25+CD127− cell population (around 10%, no significant difference between groups, Fig. 1d). These cells belong either to CD4+CD25+CD127+ or to CD4+CD25-CD127− populations for which published data remain very scarce. These cells deserve further investigation as their potential role in HIV pathophysiology has not yet been elucidated. In that sense, since FOXP3 is to date the universal marker for Tregs cells, development of a more practical technique to stain intracellular FOXP3 remains highly desirable to monitor those cells in clinical research dealing with large numbers of patients. An important step would be to shorten the intracellular FOXP3 staining in the future.
We confirm here that CD4+CD25+CD127− phenotype is a good surrogate phenotype for FOXP3+ regulatory T cells identification in both viremic and aviremic HIV-1-infected subjects. Awaiting for some robust and rapid intracellular FOXP3 staining, we suggest that the CD4+CD25+CD127− phenotype remains a reliable and easy tool (whole blood, no wash, less than 30 min to be completed, isotype control not necessary) for Tregs measurement in clinical research on HIV-infected patients.
The authors thank Anne Portier, Caroline Guignant, and Celine Griffon-Renoux for technical assistance.