Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Potential Role for HIV-Specific CD38/HLA-DR+ CD8+ T Cells in Viral Suppression and Cytotoxicity in HIV Controllers

  • Stéphane Hua,

    Affiliations INSERM U1012, Le Kremlin-Bicêtre, France, Université Paris-Sud 11, Le Kremlin-Bicêtre, France

  • Camille Lécuroux,

    Affiliations INSERM U1012, Le Kremlin-Bicêtre, France, Université Paris-Sud 11, Le Kremlin-Bicêtre, France

  • Asier Sáez-Cirión,

    Affiliation Institut Pasteur, Unité de Régulation des Infections Rétrovirales, Paris, France

  • Gianfranco Pancino,

    Affiliation Institut Pasteur, Unité de Régulation des Infections Rétrovirales, Paris, France

  • Isabelle Girault,

    Affiliations INSERM U1012, Le Kremlin-Bicêtre, France, Université Paris-Sud 11, Le Kremlin-Bicêtre, France

  • Pierre Versmisse,

    Affiliation Institut Pasteur, Unité de Régulation des Infections Rétrovirales, Paris, France

  • Faroudy Boufassa,

    Affiliations Université Paris-Sud 11, Le Kremlin-Bicêtre, France, INSERM U1018, Le Kremlin-Bicêtre, France, Assistance Publique-Hôpitaux de Paris (AP-HP), Hôpital Bicêtre, Département d'épidémiologie, Le Kremlin-Bicêtre, France

  • Olivier Taulera,

    Affiliation AP-HP, Hôpital Saint-Louis, Paris, France

  • Martine Sinet,

    Affiliations INSERM U1012, Le Kremlin-Bicêtre, France, Université Paris-Sud 11, Le Kremlin-Bicêtre, France

  • Olivier Lambotte,

    Affiliations INSERM U1012, Le Kremlin-Bicêtre, France, Université Paris-Sud 11, Le Kremlin-Bicêtre, France, AP-HP, Hôpital Bicêtre, Service de Médecine Interne, Le Kremlin-Bicêtre, France

  • Alain Venet

    alain.venet@u-psud.fr

    Affiliations INSERM U1012, Le Kremlin-Bicêtre, France, Université Paris-Sud 11, Le Kremlin-Bicêtre, France

Potential Role for HIV-Specific CD38/HLA-DR+ CD8+ T Cells in Viral Suppression and Cytotoxicity in HIV Controllers

  • Stéphane Hua, 
  • Camille Lécuroux, 
  • Asier Sáez-Cirión, 
  • Gianfranco Pancino, 
  • Isabelle Girault, 
  • Pierre Versmisse, 
  • Faroudy Boufassa, 
  • Olivier Taulera, 
  • Martine Sinet, 
  • Olivier Lambotte
PLOS
x

Abstract

Background

HIV controllers (HIC) are rare HIV-1-infected patients who exhibit spontaneous viral control. HIC have high frequency of CD38/HLA-DR+ HIV-specific CD8+ T cells. Here we examined the role of this subset in HIC status.

Materials and Methods

We compared CD38/HLA-DR+ CD8+ T cells with the classical CD38+/HLA-DR+ activated phenotype in terms of 1) their activation status, reflected by CD69, CD25, CD71, CD40 and Ki67 expression, 2) functional parameters: Bcl-2 expression, proliferative capacity, and IFN-γ and IL-2 production, and 3) cytotoxic activity. We also investigated how this particular profile is generated.

Results

Compared to CD38+/HLA-DR+ cells, CD38/HLA-DR+ cells exhibited lower expression of several activation markers, better survival capacity (Bcl-2 MFI, 367 [134–462] vs 638 [307–747], P = 0.001), higher frequency of polyfunctional cells (15% [7%–33%] vs 21% [16%–43%], P = 0.0003), greater proliferative capacity (0-fold [0–2] vs 3-fold [2][11], P = 0.007), and higher cytotoxicity in vitro (7% [3%–11%] vs 13% [6%–22%], P = 0.02). The CD38/HLA-DR+ profile was preferentially generated in response to low viral antigen concentrations.

Conclusions

These data highlight the role of CD38/HLA-DR+ HIV-specific CD8+ T cell cytotoxicity in HIC status and provide insights into the mechanism by which they are generated. Induction of this protective CD8+ subset may be important for vaccine strategies.

Introduction

A very small number of HIV-1-infected individuals spontaneously control viral replication. These so-called HIV controllers (HIC) usually have relatively high CD4+ T cell counts and remain AIDS-free for several decades [1], [2]. Several factors have been implicated in HIC status, including viral defects, cellular factors, and the innate immune system [3], but T cells, especially HIV-specific CD8+ T cells, are considered to play a major role. Overrepresentation of certain HLA class I alleles (especially HLA-B*57 and HLA-B*27) is associated with strong HIV-specific CD8+ T cell responses [4][6]. Studies of simian disease models also show that expression of the MHC class I alleles Mamu-B*08 and Mamu-B*17 correlates with viral control [7], [8]. Several studies have linked viral control to HIV-specific CD8+ T cell polyfunctionality [9], proliferation to HIV antigens, and lytic granule content [10], [11].

We and others have shown that HIC usually possess HIV-specific CD8+ T cells capable of suppressing HIV replication ex vivo [12], [13]. Although HIV-specific CD8+ T cells usually express similar levels of two activation markers – CD38 and HLA-DR – which correlate with viral load in non-controllers [14], they exhibit a peculiar activation phenotype in HIC, with low CD38 expression and high HLA-DR expression [13], [15]. Immune activation can have opposite effects on CD8+ T cell functional capacities. Indeed, a degree of immune activation is necessary for efficient effector responses (resting CD8+ T cells usually lack effector functions but can become cytotoxic when activated [16]), whereas overactivation may lead to exhaustion [17] and functional inefficiency [18]. The CD38/HLA-DR+ phenotype might reflect a low level of activation and could thereby contribute to HIC status [13], [15], [19].

Here we studied CD38/HLA-DR+ HIV-specific CD8+ T cells in a large population of HIC, focusing on functional activation status, memory/effector functions, and the mechanism underlying HLA-DR expression in the absence of CD38 expression.

Materials and Methods

Ethics statement

All the subjects provided their written informed consent to participation. The study was approved by the local investigational review board (Comité de Protection des Personnes Ile de France VII, Paris, France) and performed in accordance with the tenets of the Declaration of Helsinki.

Study participants

We collected samples from 120 HIV-infected individuals, comprising 80 patients enrolled in the French ANRS CO21 CODEX cohort of HIC (inclusion criteria: HIV infection at least 5 years previously; 5 latest plasma HIV RNA values <400 copies/mL; no HAART), 21 untreated, chronically viremic patients, and 19 HAART-treated aviremic individuals (HIV RNA <50 copies/mL) enrolled in the French ANRS CO6 PRIMO cohort. Clinical and biological characteristics of the study participants are shown in Table 1. Peripheral blood samples from seven HIV-uninfected healthy blood donors (HD) were obtained from Etablissement Français du Sang (Paris, France).

Cell preparation

Peripheral blood mononuclear cells (PBMC) were isolated from EDTA-anticoagulated blood by Ficoll density gradient centrifugation and stored in liquid nitrogen. Human leukocyte antigen typing used the complement-dependent microlymphocytotoxic technique (Ingen).

Flow cytometry of HIV-specific CD8+ T cells

Specific CD8+ T cells were characterized by staining, first for 15 minutes at room temperature with APC-labeled peptide-HLA class I multimers (Proimmune and Immudex for HLA-B*57 peptides) derived from the HIV proteins Gag, Nef, Pol, and Env and from the Epstein-Barr virus (EBV) proteins BMLF-1 and BZLF-1, and then with labeled antibodies for 15 minutes at 4°C. Anti-CD38-PerCP-Cy5.5 and anti-HLA-DR-PE-Cy7 were used to identify and sort the different activation subsets. The following antibodies were used to characterized HIV-specific CD8+ T cells: FITC–coupled anti-CD71, -CD40, -Bcl-2 and -Granzyme B; PE-coupled anti-Ki67, -CD69 and -CD25; V450-coupled anti-CD8 (BD Biosciences); VioGreen-coupled anti-CD3 (Miltenyi Biotec); and PE-coupled anti–perforin (clone D48, Diaclone). Dead cells were excluded by using the Live Dead Fixable Near-IR Dead Cell Stain kit (Molecular Probes, Invitrogen). For intracellular staining, cells were incubated with the appropriate antibodies for 30 minutes at 4°C after incubation with FACS permeabilizing solution (BD Biosciences). Samples were acquired on an LSR Fortessa cell analyzer (BD Biosciences) and analyzed with FACS DIVA software (BD Biosciences). An example of the gating strategy is shown in the Figure S1 and an example of the expression of the activation markers and cytokine production by the different subsets is shown in Figure S2. We assessed an average of two tetramers per patient and acquired at least 80 events per subset studied.

CD8+ T cells were sorted with an ARIA cell sorter (BD Biosciences) into the following activation subsets: CD38/HLA-DR+, CD38+/HLA-DR, and CD38+/HLA-DR+, using anti-CD56-PE to exclude NK cells, anti-CD8-V450 and anti-CD4-APC-H7 antibodies (BD Biosciences) according to the manufacturer's guidelines. After sorting, cells were resuspended at a density of 1×106/mL and cultured for five days before used in the cytotoxicity assay described below. Purity was routinely >95% for each subset (data not shown). Too few cells expressed the CD38+/HLA-DR phenotype to obtain interpretable data.

Intracellular cytokine production

PBMC were stimulated for 15 hours in medium containing the relevant optimal HIV peptide (2 µM). After 1 hour of stimulation, cytokine secretion was blocked by adding brefeldin A (10 µg/mL, Sigma-Aldrich Chemie). After further incubation, samples were stained as described above. Anti-IFN-γ-APC and anti-IL-2-PE antibodies (BD Biosciences) were used to detect intracellular cytokine production. A negative control (medium) and a positive control (staphylococcal enterotoxin B, SEB) were included in each experiment. The polyfunctionality was assessed by measuring the dual production of IFN-γ and IL-2. The results were expressed as frequency of IFN-γ+/IL-2+ producing cells among IFN-γ+ or IL-2+ producing cells.

In vitro HIV suppression assay

The capacity of CD8+ T cells to suppress HIV-1 infection of autologous CD4+ T cells ex vivo was assessed as described in details elsewhere [20]. Briefly, CD4+ T cells were activated with phytohemagglutinin (PHA) and IL-2 for three days, then infected with HIV-1 BaL and cultured alone or with unstimulated autologous CD8+ T cells (1∶1 ratio). P24 antigen was measured in culture supernatants with an ELISA method (Zeptometrix) as a measure of viral replication. The capacity of CD8+ T cells to suppress HIV infection was expressed as the log decrease in p24 production when superinfected CD4+ T cells were cultured in the presence of CD8+ T cells.

Cytotoxicity assay

The Grantoxilux Plus! cytotoxicity assay was used according to the manufacturer's instructions (OncoImmunin, Inc). It is based on Granzyme-B-mediated intracellular cleavage of a fluorogenic substrate in live target cells. Briefly, PBMC were used as target cells, CD8+ T cells as effectors, and CD8-depleted PBMC as feeders. Target cells were stimulated with PHA and IL-2 for five days. Effector cells were cocultured with feeders at a ratio of 1∶1 and with HIV peptide (2 µM) for five days. On day 5, target cells were primed for 1 h with the cognate or an irrelevant peptide, labeled with TFL4 (OncoImmunin, Inc.), and then cocultured with CD8+ T cells at an Effector:Target ratio of 50∶1. The cells were then incubated with Granzyme B substrate (OncoImmunin Inc.) for 1 h and analyzed immediately by flow cytometry. The results are expressed as the difference between the frequency of lysed target cells incubated with the cognate peptide and the frequency of lysed target cells incubated with the irrelevant peptide. An example of the cytotoxic assay is shown in Figure S3.

Healthy donor cell culture

CD8+ T cells were isolated by using a human anti-CD56 antibody (Miltenyi Biotec) to deplete NK cells and a human anti-CD8 antibody (Miltenyi Biotec). Purity was greater than 95%. Isolated CD8+ T cells were cultured in vitro (2×106 cells/mL) with medium alone, with the optimal concentrations of peptides, or with anti-CD3/CD28 antibodies as control (Miltenyi Biotec). IFN-α (PBL InterferonSource) was used at 500 IU/mL.

Functional avidity

Functional avidity was measured in an IFN-γ ELISpot assay using single-epitope peptides corresponding to optimal HIV-CTL epitopes (National Institutes of Health HIV Molecular Immunology Database: http://www.hiv.lanl.gov/content/immunology/tables/optimal_ctl_ summary.html) according to the subjects' HLA types. Functional avidity of CD8+ T cell responses was assessed by performing limiting peptide dilutions ranging from 10−5 to 10−11 M in in vitro assays as described elsewhere [21]. It was defined as the peptide concentration required to achieve 50% of the maximal response (EC50%) and was expressed as log EC50%.

Statistical methods

Data were analyzed with Prism software (GraphPad Software Inc.). Groups were compared using a nonparametric Mann-Whitney or Wilcoxon paired test for continuous data or a chi-squared test for categorical data. Correlations were evaluated using Spearman's rank correlation coefficient. The Spearman r correlation and (in the Figures) the Pearson correlation curve are indicated for significant correlations. The threshold for statistical significance was set to P<0.05.

Results

Study population

Baseline data for the 80 HIC, 21 viremic patients and 19 HAART-treated patients are shown in Table 1. HIC differed significantly from the other two groups of patients in terms of the sex ratio, age, CD4+ and CD8+ T cell counts, viral RNA loads, and time since diagnosis.

Higher frequencies of CD38/HLA-DR+ HIV-specific CD8+ T cells in HIV controllers

We first compared CD38 and HLA-DR expression ex vivo on bulk CD8+ T cells in the healthy donors (HD), viremic patients, HAART-treated patients, and HIV controllers (HIC) (Figure 1A–B) and on HIV-specific CD8+ T cells in the three groups of HIV-infected patients (Figure 1C–D). As previously reported by our group [15], median [1st–3rd interquartile range] CD38 expression was lower in HD, HAART-treated patients and HIC (11% [6%–14%], 12% [5%–14%], 9% [5%–14%] respectively) than in viremic patients (20% [14%–39%]) (P = 0.0001 for each comparison, Fig. 1A). HLA-DR expression was higher in viremic patients and HIC (29% [23%–38%] and 27% [19%–32%], respectively) than in HD and HAART-treated patients (14% [11%–21%] and 11% [7%–19%]) (P = 0.0002 and P<0.0001 for viremic patients versus HD and HAART, respectively; P = 0.006 and P = 0.0008 for HIC versus HD and HAART, respectively; Fig. 1A). Consequently, the CD38+/HLA-DR+ phenotype was more frequent in viremic patients (14% [8%–19%]) than in HD, HAART-treated patients and HIC (2% [1%–3%], 3% [1%–4%] and 4% [3%–8%], respectively) (P<0.0001 for all three comparisons, Fig. 1B). CD38/HLA-DR+ expression was more frequent in HIC than in HAART-treated patients (21% [15%–28%] vs 9% [5%–12%], P = 0.0003, Fig. 1B).

thumbnail
Figure 1. CD38 and HLA-DR expression on bulk and HIV-specific CD8+ T cells.

(A) Proportions of bulk CD8+ T cells from HD (n = 16, light gray bars), HAART-treated patients (n = 19, mid gray bars), viremic patients (n = 21, dark gray bars) and HIC (n = 79, black bars) expressing CD38 and HLA-DR. (C) Proportions of HIV-specific CD8+ T cells from HAART-treated patients (n = 13, mid gray bars), viremic patients (n = 39, dark gray bars) and HIC (n = 80, black bars) expressing CD38 and HLA-DR. Pie charts representing CD38/HLA-DR (white), CD38/HLA-DR+ (black), CD38+/HLA-DR (light gray) and CD38+/HLA-DR+ (dark gray) cells among bulk (B) and HIV-specific CD8+ T cells (D). Statistical differences shown in the pie charts are based on the difference in the frequency of the CD38/HLA-DR+ subset between the different groups. * P<0.05, ** P<0.01, *** P<0.001.

https://doi.org/10.1371/journal.pone.0101920.g001

HIV-specific CD8+ T cell frequency was higher in HIC (2705 SFC/106 PBMC [1135–5675]) and in viremic patients (2255 SFC/106 PBMC [1334–6933]) than in HAART-treated patients (1138 SFC/106 PBMC [179–2733]) (P = 0.009 and P = 0.03 between HIC and HAART-treated patients and between viremic patients and HAART-treated patients respectively). HIV-specific CD8+ T cells with high CD38 expression were also less frequent in HAART-treated patients and HIC (10% [6%–16%] and 13% [11%–39%], respectively) than in viremic patients (28% [15%–48%]) (P<0.0001 for both comparisons, Fig. 1C).

HLA-DR expression in HAART-treated patients (19% [10%–24%]) was far lower than in HIC (47% [32%–61%], P<0.0001), and in viremic patients (50% [32%–58%], Fig. 1C). The frequency of CD38/HLA-DR+ was thus higher in HIC (33% [20%–46%]) than in both HAART-treated and viremic patients (13% [6%–19%] and 20% [15%–47%], respectively; P<0.0001 for both comparisons, Figure 1D). Moreover, there is a correlation between the frequency of CD38/HLA-DR+ and the frequency of HIV-specific CD8+ T cells (P = 0.002, r = 0.22) which suggested that CD38/HLA-DR+ may contribute mainly to the high frequency of HIV-specific CD8+ T cells

CD38/HLA-DR+ HIV-specific CD8+ T cells are more weakly activated than their CD38+/HLA-DR+ counterparts

Although expression ex vivo of the activation markers CD69 (activation marker), CD40 (costimulatory receptor), CD71 (transferrin receptor protein 1), CD25 (IL-2Rα) and Ki67 (proliferative marker) was rather low on HIV-specific CD8+ T cells in HIC (3% [2%–6%], 4% [2%–8%], 4% [3%–7%], 1% [1%–2%] and 1% [0%–9%], respectively, Fig. 2A), classical CD38+/HLA-DR+ activated cells exhibited higher expression of CD69 (9% [4%–15%]), CD40 (14% [7%–20%]), CD71 (13% [7%–16%]), CD25 (3% [1%–8%]) and Ki67 (25% [12%–49%]), than the other subsets. Strikingly, expression of these markers on the CD38/HLA-DR+ subset was very low and similar to that observed on resting CD38/HLA-DR cells (median 0–2% for both subsets and all activation markers). Thus, the CD38/HLA-DR+ subset is more weakly activated than the CD38+/HLA-DR+ subset.

thumbnail
Figure 2. Activation phenotype of CD38/HLA-DR, CD38/HLA-DR+ and CD38+/HLA-DR+ subsets from HIC.

(A) Proportions of HIV-specific CD8+ T cells expressing CD38, HLA-DR, CD69, CD25, CD71, CD40 and Ki67. (B–F) Proportions of CD38/HLA-DR, CD38/HLA-DR+ and CD38+/HLA-DR+ HIV-specific CD8+ T cells expressing CD69 (B), CD25 (C), CD71 (D), CD40 (E) and Ki67 (F) (n = 8). * P<0.05, ** P<0.01, *** P<0.001.

https://doi.org/10.1371/journal.pone.0101920.g002

The CD38/HLA-DR+ subset contains a higher frequency of polyfunctional cells and has higher proliferative and cytotoxic capacities than the CD38+/HLA-DR+ subset

The survival capacity was assessed by measuring the difference of the median fluorescence intensity of Bcl-2 expression of one subset and the isotype. Bcl-2 expression ex vivo, was higher for CD38/HLA-DR+ cells than for CD38+/HLA-DR+ cells (638 [307–747] vs 367 [134–462], P = 0.001, respectively, Fig. 3A).

thumbnail
Figure 3. Qualitative features of CD38/HLA-DR+ and CD38+/HLA-DR+ HIV-specific CD8+ T cell subsets in HIC.

(A) Bcl-2 expression on CD38/HLA-DR+ and CD38+/HLA-DR+ HIV-specific CD8+ T cells (n = 11). (B) Proportion of CD38/HLA-DR+ and CD38+/HLA-DR+ HIV-specific CD8+ T cells producing both IFN-γ and IL-2 among HIV-specific CD8+ T cells producing IFN-γ or IL-2 (n = 35). (C) Fold increase in CD38/HLA-DR+ and CD38+/HLA-DR+ HIV-specific CD8+ T cell numbers after 5 days of culture with HIV peptides (2 µM) (n = 7). (D–E) Proportion of CD38/HLA-DR+ and CD38+/HLA-DR+ HIV-specific CD8+ T cells producing perforin (D) and granzyme B (E) (n = 35). (F) Graphs representing percentage cytotoxicity (measured as granzyme-B-mediated intracellular cleavage of a fluorogenic substrate) of CD38/HLA-DR+ and CD38+/HLA-DR+ HIV-specific CD8+ T cells (n = 11). * P<0.05, ** P<0.01, *** P<0.001.

https://doi.org/10.1371/journal.pone.0101920.g003

We then analyzed dual production in vitro of IFN-γ and IL-2 as a marker of polyfunctionality. As shown in Figure 3B the CD38/HLA-DR+ subset contained more IFN-γ+IL-2+-producing cells than the CD38+/HLA-DR+ subset (21% [16%–43%] vs 15% [7%–33%], P = 0.0003, Fig. 3B). We analyzed the proliferation capacity in vitro by assessing the number of HIV-specific CD8+ T cells at the beginning and the end of the culture and calculated the fold increase of HIV-specific CD8+ T cells. After sorting and culture with cognate peptides, proliferation of CD38/HLA-DR+ cells increased 3.2-fold [1.2–7.8], whereas CD38+/HLA-DR+ cells did not proliferate significantly (0.4-fold [0.0–1.7]) (P = 0.007, Fig. 3C).

Although ex vivo perforin expression did not differ between the two subsets (52% [21%–67%] and 52% [26%–67%], respectively, P = 0.26, Fig. 3D), CD38/HLA-DR+ cells exhibited lower granzyme B ex vivo expression than CD38+/HLA-DR+ cells (29% [8%–54%] and 36% [12%–64%], respectively, P = 0.01, Fig. 3E). However, ex vivo after sorting cells, cytotoxicity was very low and similar with the two subsets (0% [0%–8%] vs 4% [4%–16%], respectively, P = 0.2; data not shown). After sorting and coculture with CD8-depleted PBMC for 5 days, CD38/HLA-DR+ cells showed increased cytotoxicity (13% [7%–23%]), contrary to the CD38+/HLA-DR+ subset (7% [3%–11%]) (P = 0.001, Fig. 3F). CD38/HLA-DR+ CD8+ T cells thus responded significantly more strongly to cognate antigen than did their CD38+/HLA-DR+ counterparts.

Ex vivo HIV-suppressive activity was routinely evaluated with freshly isolated whole-blood T cells which often include several HIV specificities while cytotoxicity was evaluated only on immunodominant peptides. Of note, HIV-specific CD8+ T cell cytotoxicity correlated with HIV-suppressive activity, expressed as the log decline in p24 antigen (P = 0.004; r = 0.37, data not shown). Interestingly, only the frequency of the CD38/HLA-DR+ HIV-specific CD8+ T cells correlated strongly with the log p24 decrease (P<0.0001; r = 0.32, data not shown). No such correlation was found with CD38+/HLA-DR+ cells (P = 0.5).

The conditions of stimulation determine CD38 and HLA-DR expression by specific CD8+ T cells

Finally, we examined the mechanisms by which resting CD38/HLA-DR cells become CD38/HLA-DR+ cells in the context of viral infection. For this purpose, we used an in vitro resting and fully unactivated model in which EBV-specific CD8+ T cells which are mostly CD38/HLA-DR as shown in Figure 4 are stimulated by EBV peptides in various conditions.

thumbnail
Figure 4. Influence of stimulatory conditions on the activation phenotype of specific CD8+ T cells.

(A) Representative dot plots of HLA-DR and CD38 expression in unstimulated conditions (upper graph), after IFN-α stimulation (middle graph) or peptide stimulation (2 µM, lower graph) among specific (dark dots) and non-specific (gray dots) CD8+ T cells from healthy donors after a four-day culture period. CD38 and HLA-DR expression on bulk (C left panel) and specific CD8+ T cells (B and C right panel). (B) Flow cytometry histograms showing representative results for cells from one individual in unstimulated conditions (dark lines), after IFN-α stimulation (light gray histograms) or peptide stimulation (dark gray histograms). (C) Surface expression of CD38 (white bars) and HLA-DR (gray bars) on bulk and specific CD8+ T cells (n = 4).

https://doi.org/10.1371/journal.pone.0101920.g004

We postulated that the CD38+/HLA-DR+ profile might result from a dual activation pathway, in which antigenic peptide stimulation leads to HLA-DR expression while indirect activation by IFN-α leads to CD38 expression [22], [23]. As expected, IFN-α stimulation led to CD38 expression on bulk CD8+ T cells (14% [7%–17%]) and, to a lesser degree, on EBV-specific CD8+ T cells (9% [4%–28%]), whereas neither bulk nor EBV-specific CD8+ T cells expressed HLA-DR (0% [0%–1%] and 0% [0%–1%] respectively; Figure 4, middle panel), confirming the non-specific nature of IFN-α-induced activation. However, stimulation with EBV peptides led to strong expression of both CD38 and HLA-DR by EBV-specific cells (69% [53%–84%] vs 72% [61%–83%], respectively; Figure 4 lower panel), while no change was observed in the total CD8+ T cell population. We suspected that the increase in CD38 expression might be related to IFN-α production by non T cells in this model, but neither blockade with an anti-IFN-αR antibody nor the use of highly purified CD8+ T cells significantly modified our results.

Interestingly, the antigen concentration correlated positively with the degree of activation and also affected the profile of activated cells. At low antigen levels, most activated cells expressed HLA-DR but not CD38. Indeed, the frequency of CD38/HLA-DR+ cells among activated cells was higher at low antigen levels (89% [44%–98%] and 2% [1%–11%] at 0.2 nM and 2000 nM, respectively; P = 0.0006, Fig. 5A), whereas CD38+/HLA-DR+ cells exhibited the opposite behavior (4% [0%–14%] vs 91% [78%–97%] at 0.2 nM and 2000 nM, respectively; P = 0.001, Fig. 5B).

thumbnail
Figure 5. Stimulation with a low antigen concentration induces the CD38/HLA-DR+ phenotype on specific CD8+ T cells.

(A) Graphs representing the frequency of (A) CD38/HLA-DR+ cells (dark histograms) and (B) CD38+/HLA-DR+ cells (dark gray histograms) among activated EBV-specific healthy donor CD8+ T cells (i.e. those expressing CD38 and/or HLA-DR) after a four-day culture period (n = 8).

https://doi.org/10.1371/journal.pone.0101920.g005

The CD38/HLA-DR+ phenotype by specific CD8+ T cells in HIC is conditioned by low viral loads and high antigen sensitivity

HIV-specific CD8+ T cell sensitivity to cognate antigen, measured with an ELISpot assay, showed an overall positive correlation between the frequency of ex vivo CD38/HLA-DR+ in these cells and the antigen sensitivity of HIV-specific CD8+ T cells (r = 0.34, P = 0.01, Figure 6A), whereas no such correlation was seen with CD38/HLA-DR cells (P = 0.39) or CD38+/HLA-DR+ cells (P = 0.05, Figure 6B).

thumbnail
Figure 6. High antigen sensitivity is associated with a high frequency of CD38/HLA-DR+ cells in HIC.

The antigen sensitivity of HIV-specific CD8+ T cells was measured in ELISpot assays with serial limiting dilutions of antigenic peptides (from 10−5 to 10−11 M) and was expressed as the log molar concentration of peptide yielding 50% of the maximum response (EC50%). Correlations between the proportion of CD38/HLA-DR+ (A) or CD38+/HLA-DR+ (B) and the antigen sensitivity of HIV-specific CD8+ T cells from HIC. Correlations were evaluated using the Spearman rank correlation coefficient. The Spearman r correlation and the Pearson correlation curve are indicated for significant correlations (n = 47).

https://doi.org/10.1371/journal.pone.0101920.g006

These results suggest strongly that the CD38/HLA-DR+ subset is induced in the context of low viral loads and expressed on highly avid cells.

Discussion

HIV-specific CD8+ T cells in HIV controllers were initially shown to exhibit (i) efficient proliferation and cytotoxicity [10], [11], (ii) strong virus suppression ex vivo [13], and (iii) a particular activation phenotype: CD38/HLA-DR+ [13], [15]. Little is known about this CD8+ T cell subset and its possible role in HIC. The activation status of this CD38/HLA-DR+ subset in HIC is puzzling, as HLA-DR expression suggests an activated state, while the lack of CD38 expression suggests the opposite. None of the other activation markers studied here were expressed by these cells, including those expressed early after activation, such as CD69 and CD25, and those expressed later, such as CD71 and CD40 [24], [25]. This exclusive HLA-DR expression suggests a low state of activation, and is in keeping with absent expression of Ki-67, a marker of cell cycling. A previous study showed coexpression of HLA-DR and Ki-67 by these cells after immunotherapy, but the authors did not evaluate CD38 expression [26].

This low level of activation is compatible with efficient survival evaluated here in terms of Bcl-2 expression level, and also with good proliferative potential. Previous studies showed an association between HLA-DR expression and proliferation [26], [27]. And indeed, these CD38/HLA-DR+ cells expanded 3-fold in culture, whereas their CD38+/HLA-DR+ counterparts did not expand significantly. Of note, this expansion of CD38/HLA-DR+ cells was associated with differentiation into effector cells, as their HIV-specific cytotoxicity was higher than that of CD38+/HLA-DR+ cells, even after adjusting for the HIV-specific-effector-to-target ratio. These data confirm those reported by Blankson et al., who showed that sorted CD38/HLA-DR+ CD8+ T cells from HIC, when placed in short-term culture with autologous unstimulated CD4+ T cells, were the first cells to suppress viral replication [28]. Additionally, although CD38/HLA-DR+ and CD38+/HLA-DR+ showed similar ex vivo cytotoxic capacity, the very low frequency of CD38+/HLA-DR+ cells, and their poor proliferative capacity shown here and elsewhere [29], make it unlikely that this population has a major role in suppressing viral replication. These data suggest that in vivo CD38/HLA-DR+ cells which are less susceptible than CD38+/HLA-DR+ cells to exhaustion, have a greater capacity to control viral replication than CD38+/HLA-DR+ cells. Although we and others have observed that HIV-specific CD8+ T cells from some HIC patients lack HIV-suppressive activity ex vivo, they are capable of differentiating when exposed to cognate antigens, acquiring the ability to lyse infected cells and to suppress HIV replication [30].

We also explored the conditions in which the CD38/HLA-DR+ phenotype is generated by using an in vitro model with EBV peptides and cells from healthy donors. We needed a non-activated condition to evaluate the mechanism of CD38 and HLA-DR expression. EBV infection is a model of infection with effective memory resting cells while HIV infection leads to persistent activation and inflammation even in HIC. Therefore, we believed that the EBV-model is a more appropriate model to address the question of CD38 and HLA-DR induction after activation. We confirmed that IFN-α induces CD38 expression but not HLA-DR expression [22], [23] and that this activation is not virus-specific. Conversely, EBV peptide stimulation led to activation of EBV-specific CD8+ T cells expressing both CD38 and HLA-DR. It has been shown that strong TCR stimulation by anti-CD3/CD28 leads to cd38 transcription in purified CD8+ T cells [23], possibly explaining how CD38 is expressed following peptide stimulation even in IFN-αR-antagonizing conditions. However, we found that low EBV peptide concentrations elicited the CD38/HLA-DR+ phenotype on the vast majority of activated cells, in keeping with the low viral load and lack of IFN-α production in experimental models of viral control [31]. Although plasmacytoid dendritic cells from HIC are able to produce IFN-α efficiently after HIV antigen stimulation in vitro [32], [33], HIC do not produce detectable IFN-α in vivo, as shown by tissue studies [34], which can explain the absence of IFN-α pathway mediated CD38 expression. The discrepancy between the low frequency of CD38/HLA-DR+ cells among virus-specific CD8+ T cells in our in vitro model and the high frequency observed in vivo in HIC may be due to the contrast between brief activation in vitro and chronic activation by low HIV viral loads in HIC. The preferential activation of the CD38/HLA-DR+ subset by low antigen concentrations is particularly interesting, as memory cells activated by low antigen concentrations show little senescence, have high proliferation rates, are able to persist for long periods [35], and maintain their TCR on the membrane surface, enabling a persistent cytotoxic response [36]. Furthermore, the peptide concentrations that yielded the CD38/HLA-DR+ phenotype in our in vitro EBV model are similar to the estimated concentrations of HIV antigens present in vivo [37], [38]. In addition, the high antigen sensitivity observed in HIC may reinforce preferential induction of the CD38/HLA-DR+ profile on the most sensitive cells, which are endowed with the best functional profile [21], [38][40]. Indeed, antigen sensitivity has been shown to correlate negatively with HIV cellular viral load [40]. However, high antigen sensitivity is not the only feature that can explain HIC status. Indeed, we observed no increase in the frequency of CD38/HLA-DR+ cells in patients positive for HLA-B*27 or HLA-B*57 ([21] and data not shown) whereas antigen sensitivity is particularly high in this context [21], [39]. We moreover observed no difference in the frequency of this subset when comparing different HIV peptide specificities (data not shown). In addition, we have previously reported a lack of difference in perforin expression and inhibition of viral replication between HIC with and without protective HLA alleles [21], [41].

T cell activation is necessary for effector functions such as cytotoxicity and suppression of viral replication, but the persistent immune activation associated with systemic inflammation is known to play a key role in HIV disease progression [42], [43]. This is also the case in HIC, who display low but higher frequencies of classical activated CD38+/HLA-DR+ CD8+ T cells than do healthy donors (HD) and HAART-treated patients [43]. This excessive immune activation may result from the persistence of extremely low levels of HIV replication [44] and/or enhanced microbial translocation from the gut [45]. HIC exhibit a degree of disease progression and sometimes experience a decline in CD4+ T cells or even lose their ability to control HIV [46], possibly as a result of persistent immune activation [46], [47]. Alternatively, CD4+ T cell activation might permit persistent low-level infection of these cells, in turn maintaining efficient stimulation of the HIV-specific immune response [21], [43] and thereby helping to control viral replication through inducing CD38/HLA-DR+ profile.

Conclusions

The paradoxical activation profile of some HIV-specific CD8+ T cells, with HLA-DR but not CD38 expression and weak expression of all other activation markers, might confer the capacity to differentiate into effective cytotoxic cells after moderate activation [28] while avoiding the deleterious effects of excessive immune activation. It would be a favorable condition as compared to HAART-treated patients who display a very low viral load which leads to an absence of immune activation and effective response and as compared to viremic patients who display high viral load which leads to excessive activation and exhaustion. A high frequency of CD38/HLA-DR+ cells among total CD8+ T cells has been linked to slow HIV disease progression and high CD4+ T cell counts [48]. We suspect that the expansion of this particular activated subset will lead to effective HIV-specific CD8+ T cells capable of rapid viral suppression. Induction of this protective cell subset by activation with low concentrations of HIV antigens, together with limited lymphocyte activation, might have implications for HIV vaccine strategies.

Supporting Information

Figure S1.

Example of gating strategy of subsets of HIV-specific CD8+ T cells in HIV controllers. Cells were gated (A) on live cells using live/dead stain cells kit, (B) on lymphocytes, (C) on CD3+/CD8+ expressing cells, (D) on HIV-specific CD8+ T cells using tetramer staining, (E) on different subset of CD38 and HLA-DR expressing cells using isotypes as control.

https://doi.org/10.1371/journal.pone.0101920.s001

(TIF)

Figure S2.

Dot plots of activation marker expression and cytokine production on different subsets of HIV-specific CD8+ T cells in HIV controllers. Representative dot plots of CD25 and CD40 (A), CD69 and CD71 (B), Ki67 (C) expression and IL-2 and/or IFN-γ secretion (D) on CD38/HLA-DR (left panel), CD38/HLA-DR+ (middle panel) and CD38+/HLA-DR+ (right panel) subsets

https://doi.org/10.1371/journal.pone.0101920.s002

(TIF)

Figure S3.

Dot plots of the cytotoxic assay of CD38/HLA-DR+ and CD38+/HLA-DR+ HIV-specific CD8+ T cells. Cytotoxic capacity was assessed by measuring the frequency of positive target cells for granzyme B substrate. The results were expressed as the difference between the frequency using effector cells co-cultured with target cells incubated with the relevant peptide (B and D) and the frequency in the negative control using effector cells co-cultured with target cells incubated with irrelevant peptide (A and C).

https://doi.org/10.1371/journal.pone.0101920.s003

(TIF)

Acknowledgments

The authors thank all the clinicians and the patients from all of the participating centers of the French HIV Controllers cohort (Agence Nationale de Recherche sur le SIDA, CO21-CODEX Cohort) and of the French PRIMO cohort (ANRS CO6-PRIMO Cohort). We also thank Christine Rouzioux for the data on viral RNA, Sandie Gérard for technical support, David Young who edited the English text and Christine Bourgeois and Nicolas Noel for helpful discussions.

Author Contributions

Conceived and designed the experiments: AS-C GP OL MS AV. Performed the experiments: SH CL IG AS-C PV. Analyzed the data: SH CL MS AV. Contributed to the writing of the manuscript: SH CL AV. Helped to coordinate the French CODEX cohort: FB OT OL.

References

  1. 1. Deeks SG, Walker BD (2007) Human immunodeficiency virus controllers: mechanisms of durable virus control in the absence of antiretroviral therapy. Immunity 27: 406–416
  2. 2. Lambotte O, Boufassa F, Madec Y, Nguyen A, Goujard C, et al. (2005) HIV Controllers: A Homogeneous Group of HIV-1—Infected Patients with Spontaneous Control of Viral Replication. Clin Infect Dis 41: 1053–1056
  3. 3. Sáez-Cirión A, Pancino G (2013) HIV controllers: a genetically determined or inducible phenotype? Immunol Rev 254: 281–294
  4. 4. Carrington M, O'Brien SJ (2003) The influence of HLA genotype on AIDS. Annu Rev Med 54: 535–551
  5. 5. Carrington M, Walker BD (2012) Immunogenetics of spontaneous control of HIV. Annu Rev Med 63: 131–145
  6. 6. Goulder PJR, Walker BD (2012) HIV and HLA class I: an evolving relationship. Immunity 37: 426–440
  7. 7. Loffredo JT, Maxwell J, Qi Y, Glidden CE, Borchardt GJ, et al. (2007) Mamu-B*08-positive macaques control simian immunodeficiency virus replication. J Virol 81: 8827–8832
  8. 8. Yant LJ, Friedrich TC, Johnson RC, May GE, Maness NJ, et al. (2006) The High-Frequency Major Histocompatibility Complex Class I Allele Mamu-B*17 Is Associated with Control of Simian Immunodeficiency Virus SIVmac239 Replication. J Virol 80: 5074–5077
  9. 9. Betts MR, Nason MC, West SM, De Rosa SC, Migueles SA, et al. (2006) HIV nonprogressors preferentially maintain highly functional HIV-specific CD8+ T cells. Blood 107: 4781–4789
  10. 10. Migueles SA, Laborico AC, Shupert WL, Sabbaghian MS, Rabin R, et al. (2002) HIV-specific CD8+ T cell proliferation is coupled to perforin expression and is maintained in nonprogressors. Nat Immunol 3: 1061–1068
  11. 11. Migueles SA, Osborne CM, Royce C, Compton AA, Joshi RP, et al. (2008) Lytic granule loading of CD8+ T cells is required for HIV-infected cell elimination associated with immune control. Immunity 29: 1009–1021
  12. 12. Chen H, Piechocka-Trocha A, Miura T, Brockman MA, Julg BD, et al. (2009) Differential Neutralization of Human Immunodeficiency Virus (HIV) Replication in Autologous CD4 T Cells by HIV-Specific Cytotoxic T Lymphocytes. J Virol 83: 3138–3149
  13. 13. Sáez-Cirión A, Lacabaratz C, Lambotte O, Versmisse P, Urrutia A, et al. (2007) HIV controllers exhibit potent CD8 T cell capacity to suppress HIV infection ex vivo and peculiar cytotoxic T lymphocyte activation phenotype. Proc Natl Acad Sci U S A 104: 6776–6781
  14. 14. Doisne J-M, Urrutia A, Lacabaratz-Porret C, Goujard C, Meyer L, et al. (2004) CD8+ T cells specific for EBV, cytomegalovirus, and influenza virus are activated during primary HIV infection. J Immunol Baltim Md 1950 173: 2410–2418.
  15. 15. Sáez-Cirión A, Sinet M, Shin SY, Urrutia A, Versmisse P, et al. (2009) Heterogeneity in HIV suppression by CD8 T cells from HIV controllers: association with Gag-specific CD8 T cell responses. J Immunol Baltim Md 1950 182: 7828–7837
  16. 16. Kaech SM, Wherry EJ, Ahmed R (2002) Effector and memory T-cell differentiation: implications for vaccine development. Nat Rev Immunol 2: 251–262
  17. 17. Sachdeva M, Fischl MA, Pahwa R, Sachdeva N, Pahwa S (2010) Immune Exhaustion Occurs Concomitantly With Immune Activation and Decrease in Regulatory T Cells in Viremic Chronically HIV-1–Infected Patients: JAIDS. J Acquir Immune Defic Syndr 54: 447–454
  18. 18. Trautmann L, Janbazian L, Chomont N, Said EA, Gimmig S, et al. (2006) Upregulation of PD-1 expression on HIV-specific CD8+ T cells leads to reversible immune dysfunction. Nat Med 12: 1198–1202
  19. 19. Deeks SG, Kitchen CMR, Liu L, Guo H, Gascon R, et al. (2004) Immune activation set point during early HIV infection predicts subsequent CD4+ T-cell changes independent of viral load. Blood 104: 942–947
  20. 20. Sáez-Cirión A, Shin SY, Versmisse P, Barré-Sinoussi F, Pancino G (2010) Ex vivo T cell-based HIV suppression assay to evaluate HIV-specific CD8+ T-cell responses. Nat Protoc 5: 1033–1041
  21. 21. Lécuroux C, Sáez-Cirión A, Girault I, Versmisse P, Boufassa F, et al. (2014) Both HLA-B*57 and plasma HIV RNA levels contribute to the HIV-specific CD8+ T cell response in HIV controllers. J Virol 88: 176–187
  22. 22. Manion M, Rodriguez B, Medvik K, Hardy G, Harding CV, et al. (2012) Interferon-alpha administration enhances CD8+ T cell activation in HIV infection. PloS One 7: e30306
  23. 23. Hervas-Stubbs S, Riezu-Boj J-I, Gonzalez I, Mancheño U, Dubrot J, et al. (2010) Effects of IFN-α as a signal-3 cytokine on human naïve and antigen-experienced CD8(+) T cells. Eur J Immunol 40: 3389–3402
  24. 24. Bourgeois C, Rocha B, Tanchot C (2002) A Role for CD40 Expression on CD8+ T Cells in the Generation of CD8+ T Cell Memory. Science 297: 2060–2063
  25. 25. Caruso A, Licenziati S, Corulli M, Canaris A, De Francesco M, et al. (1997) Flow cytometric analysis of activation markers on stimulated T cells and their correlation with cell proliferation. Cytometry 27: 71–76
  26. 26. Speiser DE, Migliaccio M, Pittet MJ, Valmori D, Liénard D, et al. (2001) Human CD8(+) T cells expressing HLA-DR and CD28 show telomerase activity and are distinct from cytolytic effector T cells. Eur J Immunol 31: 459–466
  27. 27. Imamichi H, Lempicki RA, Adelsberger JW, Hasley RB, Rosenberg A, et al. (2012) The CD8+ HLA-DR+ T cells expanded in HIV-1 infection are qualitatively identical to those from healthy controls. Eur J Immunol 42: 2608–2620
  28. 28. Buckheit RW III, Salgado M, Silciano RF, Blankson JN (2012) Inhibitory potential of subpopulations of CD8+ T cells in HIV-1-infected elite suppressors. J Virol 86: 13679–13688
  29. 29. Sandoval-Montes C, Santos-Argumedo L (2005) CD38 is expressed selectively during the activation of a subset of mature T cells with reduced proliferation but improved potential to produce cytokines. J Leukoc Biol 77: 513–521
  30. 30. Ndhlovu ZM, Proudfoot J, Cesa K, Alvino DM, McMullen A, et al. (2012) Elite controllers with low to absent effector CD8+ T cell responses maintain highly functional, broadly directed central memory responses. J Virol 86: 6959–6969
  31. 31. Campillo-Gimenez L, Laforge M, Fay M, Brussel A, Cumont M-C, et al. (2010) Nonpathogenesis of simian immunodeficiency virus infection is associated with reduced inflammation and recruitment of plasmacytoid dendritic cells to lymph nodes, not to lack of an interferon type I response, during the acute phase. J Virol 84: 1838–1846
  32. 32. Barblu L, Machmach K, Gras C, Delfraissy J-F, Boufassa F, et al. (2012) Plasmacytoid dendritic cells (pDCs) from HIV controllers produce interferon-α and differentiate into functional killer pDCs under HIV activation. J Infect Dis 206: 790–801
  33. 33. Herbeuval J-P, Shearer GM (2007) HIV-1 immunopathogenesis: How good interferon turns bad. Clin Immunol 123: 121–128
  34. 34. Herbeuval J-P, Nilsson J, Boasso A, Hardy AW, Kruhlak MJ, et al. (2006) Differential expression of IFN-alpha and TRAIL/DR5 in lymphoid tissue of progressor versus nonprogressor HIV-1-infected patients. Proc Natl Acad Sci U S A 103: 7000–7005
  35. 35. West EE, Youngblood B, Tan WG, Jin H-T, Araki K, et al. (2011) Tight regulation of memory CD8(+) T cells limits their effectiveness during sustained high viral load. Immunity 35: 285–298
  36. 36. Betts MR, Price DA, Brenchley JM, Loré K, Guenaga FJ, et al. (2004) The functional profile of primary human antiviral CD8+ T cell effector activity is dictated by cognate peptide concentration. J Immunol Baltim Md 1950 172: 6407–6417.
  37. 37. Varela-Rohena A, Molloy PE, Dunn SM, Li Y, Suhoski MM, et al. (2008) Control of HIV-1 immune escape by CD8 T cells expressing enhanced T-cell receptor. Nat Med 14: 1390–1395
  38. 38. Almeida JR, Sauce D, Price DA, Papagno L, Shin SY, et al. (2009) Antigen sensitivity is a major determinant of CD8+ T-cell polyfunctionality and HIV-suppressive activity. Blood 113: 6351–6360
  39. 39. Berger CT, Frahm N, Price DA, Mothe B, Ghebremichael M, et al. (2011) High-functional-avidity cytotoxic T lymphocyte responses to HLA-B-restricted Gag-derived epitopes associated with relative HIV control. J Virol 85: 9334–9345
  40. 40. Almeida JR, Price DA, Papagno L, Arkoub ZA, Sauce D, et al. (2007) Superior control of HIV-1 replication by CD8+ T cells is reflected by their avidity, polyfunctionality, and clonal turnover. J Exp Med 204: 2473–2485
  41. 41. Hersperger AR, Pereyra F, Nason M, Demers K, Sheth P, et al. (2010) Perforin expression directly ex vivo by HIV-specific CD8 T-cells is a correlate of HIV elite control. PLoS Pathog 6: e1000917
  42. 42. Giorgi JV, Lyles RH, Matud JL, Yamashita TE, Mellors JW, et al. (2002) Predictive value of immunologic and virologic markers after long or short duration of HIV-1 infection. J Acquir Immune Defic Syndr 1999 29: 346–355.
  43. 43. Hunt PW, Brenchley J, Sinclair E, McCune JM, Roland M, et al. (2008) Relationship between T cell activation and CD4+ T cell count in HIV-seropositive individuals with undetectable plasma HIV RNA levels in the absence of therapy. J Infect Dis 197: 126–133
  44. 44. Pereyra F, Palmer S, Miura T, Block BL, Wiegand A, et al. (2009) Persistent Low-Level Viremia in HIV-1 Elite Controllers and Relationship to Immunologic Parameters. J Infect Dis 200: 984–990
  45. 45. Brenchley JM, Price DA, Schacker TW, Asher TE, Silvestri G, et al. (2006) Microbial translocation is a cause of systemic immune activation in chronic HIV infection. Nat Med 12: 1365–1371
  46. 46. Boufassa F, Saez-Cirion A, Lechenadec J, Zucman D, Avettand-Fenoel V, et al. (2011) CD4 dynamics over a 15 year-period among HIV controllers enrolled in the ANRS French observatory. PloS One 6: e18726
  47. 47. Okulicz JF, Marconi VC, Landrum ML, Wegner S, Weintrob A, et al. (2009) Clinical Outcomes of Elite Controllers, Viremic Controllers, and Long-Term Nonprogressors in the US Department of Defense HIV Natural History Study. J Infect Dis 200: 1714–1723
  48. 48. Giorgi JV, Ho HN, Hirji K, Chou CC, Hultin LE, et al. (1994) CD8+ lymphocyte activation at human immunodeficiency virus type 1 seroconversion: development of HLA-DR+ CD38− CD8+ cells is associated with subsequent stable CD4+ cell levels. The Multicenter AIDS Cohort Study Group. J Infect Dis 170: 775–781.