Kynurenine Reduces Memory CD4 T-Cell Survival by Interfering with Interleukin-2 Signaling Early during HIV-1 Infection

ABSTRACT

Early HIV-1 infection is characterized by enhanced tryptophan catabolism, which contributes to immune suppression and disease progression. However, the mechanism by which kynurenine, a tryptophan-related metabolite, induces immune suppression remains poorly understood. Herein, we show that the increased production of kynurenine correlates with defective interleukin-2 (IL-2) signaling in memory CD4 T cells from HIV-infected subjects. Defective IL-2 signaling in these subjects, which drives reduced protection from Fas-mediated apoptosis, was also associated with memory CD4 T-cell loss. Treatment of memory CD4 T cells with the concentration of kynurenine found in plasma inhibited IL-2 signaling through the production of reactive oxygen species. We further show that IL-2 signaling in memory CD4 T cells is improved by the antioxidant N-acetylcysteine. Early initiation of antiretroviral therapy restored the IL-2 response in memory CD4 T cells by reducing reactive oxygen species and kynurenine production. The study findings provide a kynurenine-dependent mechanism through IL-2 signaling for reduced CD4 T-cell survival, which can be reversed by early treatment initiation in HIV-1 infection.

IMPORTANCE The persistence of functional memory CD4 T cells represents the basis for long-lasting immune protection in individuals after exposure to HIV-1. Unfortunately, primary HIV-1 infection results in the massive loss of these cells within weeks of infection, which is mainly driven by inflammation and massive infection by the virus. These new findings show that the enhanced production of kynurenine, a metabolite related to tryptophan catabolism, also impairs memory CD4 T-cell survival and interferes with IL-2 signaling early during HIV-1 infection.

INTRODUCTION

Progressive depletion of CD4 T cells by pyroptosis and activation-related apoptosis is the hallmark of HIV-1 infection (1–4). The loss of memory CD4 T cells and, in particular, central memory CD4 T cells (T_CM cells), which are critical to ensure long-lasting immune protection, occurs from the onset of infection and independently predicts disease progression (5–8). Memory T-cell survival depends on signals provided by the gamma-chain-receptor cytokines, such as interleukin-2 (IL-2) and IL-7 (9–12). Previous data showed an impaired response to these cytokines in T cells during HIV-1 infection (13–16). However, our knowledge of the molecular mechanisms responsible for these immune defects is still incomplete.

CD4 T-cell depletion is also linked with persistent inflammation, which takes place in the gut and lymphoid tissues early after infection and in the bloodstream later after infection (17–19). This inflammation is clinically relevant, since it has not been fully abrogated by long-term antiretroviral therapy (ART) (20–23). Persistent inflammation during HIV-1 infection is associated with enhanced metabolic activity, which drives a Warburg-like effect on lipid, glucose, and amino acid pathways (22, 24–27). The catabolism of an essential amino acid like tryptophan into kynurenine (Kyn), which is mediated by the indoleamine 2,3-dioxygenase 1 (IDO-1) expressed in dendritic cells (DCs) and monocytes, is increased in HIV-1-infected subjects (22, 28–30). The enhanced tryptophan catabolism detected early during HIV-1 infection correlates with proinflammatory molecules like IL-6, soluble CD40 (sCD40), and gamma interferon (IFN-γ)-inducible protein 10 (IP-10) (22, 25, 31, 32) and is associated with CD4 T-cell decay (33–35). Many studies showed that the excessive tryptophan metabolism occurring during primary HIV-1 infection is associated with increased immune dysfunction. This state is characterized by the decrease of tryptophan availability and the increased production of the immunosuppressive compound Kyn. Enhanced tryptophan/kynurenine catabolism leads to the loss of Th₁₇ and Th₂₂ cells, which play key protective roles in the gut for microbial defense (36), the generation of regulatory T cells (25, 29, 32), and the inhibition of T-cell proliferation (37, 38).

In the present study, we investigated the impact of Kyn on memory CD4 T-cell survival in the early and chronic stages of HIV-1 infection. Our results identified a critical role of Kyn in the loss of memory CD4 T cells during HIV-1 infection by interfering with IL-2 signaling through the production of reactive oxygen species (ROS).

MATERIALS AND METHODS

Ethics statement.

Control subjects and HIV-1-infected subjects were participants in the Montreal, Quebec, Canada, primary HIV infection study that received approval from the McGill University Health Centre Ethical Review Board (ethics reference number SL-00.069). All subjects provided informed and written consent for participation in the study.

Study population.

Peripheral blood mononuclear cells (PBMCs) and plasma samples were collected from subjects in whom viral acquisition was estimated to have occurred less than 90 days earlier (primary HIV-1-infected [PHI] subjects), untreated ART-naive patients who had been infected with HIV-1 for a mean 4 years (chronically HIV-1-infected [CHI] subjects), and age-matched HIV-uninfected control subjects (HIV^free subjects) (n = 6 subjects in each group). All HIV-1-infected subjects included in the overall study were homogeneously selected, and clinical and virological data were similar for all HIV-1-infected subjects. Table 1 summarizes the clinical features of the 12 PHI and CHI subjects, including the viral load (VL) and CD4/CD8 ratio.

TABLE 1.

Clinical features of all HIV-1-infected subjects

Characteristic	Result for study HIV-1-infected population
	PHI patients	CHI patients
Time of infection before the study (days)
Mean ± SD	61.3 ± 18.4	647.1 ± 380.7
Range	31–90	167–1,278
Age (yr)
Mean ± SD	35.7 ± 8.1	37.2 ± 7.5
Range	22–48	21–47
No. (%) of patients of male sex	18 (90)	12 (100)
CD4 count (no. of cells/μl)
Mean ± SD	510 ± 238	489 ± 229
Range	240–1,120	260–1,128
CD8 count (no. of cells/μl)
Mean ± SD	1,094 ± 728	829 ± 319
Range	410–3,450	410–1,270
CD4/CD8 ratio
Mean ± SD	0.61 ± 0.41	0.66 ± 0.32
Range	0.13–1.40	0.32–1.30
VL (log10 no. of copies/ml)
Mean ± SD	5.11 ± 0.96	4.88 ± 0.65
Range	3.71–7.48	4.10–6.34

Reagents and antibodies.

RPMI 1640 medium, fetal bovine serum (FBS), antibiotics, and phosphate-buffered saline (PBS) were obtained from Wisent Inc. (Saint-Jean-Baptiste, QC, Canada). Anti-Fas CH11 antibody was from MBL International Corporation (Woburn, MA, USA). Recombinant cytokines (IL-2 and IL-7), hydrogen peroxide (H₂O₂) solution, phorbol myristate acetate, and ionomycin were provided from Sigma-Aldrich (Oakville, ON, Canada). We purchased all antibodies and reagents for flow cytometry from BD Biosciences (San Jose, CA, USA), except for the antibody to CD45RA-ECD, which was from Beckman Coulter (Montreal, QC, Canada). 7-Aminoactinomycin D (7-AAD) came from Invitrogen (Burlington, ON, Canada).

Quantitation of plasma markers.

The plasma levels of proinflammatory IL-6, soluble CD14 (sCD14), lipopolysaccharide (LPS)-binding protein (LBP), IP-10, and the tryptophan-related metabolite Kyn were measured by enzyme-linked immunosorbent assays (ELISAs) according to the manufacturers' instructions. Commercially available ELISA kits were purchased from R&D Systems (Minneapolis, MN, USA) for IL-6, IP-10, and sCD14; from Hycult Biotech for LBP; and Antibodies Online for Kyn.

Purification of CD4 T cells.

Total CD4 T cells were purified using an untouched CD4 isolation kit (EasySep human CD4 T cell enrichment kit; Stem Cell Technologies, Vancouver, BC, Canada), allowing more than 96% purification without any cell stimulation or apoptosis.

STAT5 phosphorylation assay.

Cells from PHI, CHI, and HIV^free subjects were stimulated with IL-2 (25 IU/ml) or IL-7 (0.3 ng/ml) for 15 min or were unstimulated. STAT5 phosphorylation was then measured by the BD Biosciences Phosflow protocol using anti-phospho-STAT (pSTAT5)-Alexa Fluor 647-specific antibody as previously described (39). The following multiparameter antibody cocktail was used: anti-CD3-phycoerythrin (PE), anti-CD4-V450, anti-CD45RA-ECD, anti-CD27-allophycocyanin (APC) H7, and anti-CCR7-PE Cy7 antibodies. The viability marker 7-AAD was always used to exclude dead cells from analysis. We collected approximately 20,000 gated events on a BD LSRII Fortessa flow cytometer (BD) and analyzed the data using DIVA and FlowJo software. Results shown are the expression levels of pSTAT5 within gated 7-AAD-negative CD3⁺ CD4⁺ CD45RA⁻ memory CD4 T cells.

Fas-mediated apoptosis.

Similarly to the procedure described in our previous study (8), we first cultured cells from PHI, CHI, and HIV^free subjects with 1.25 μg/ml of anti-Fas CH11 antibody in the presence or absence of 25 IU/ml IL-2 for 24 h. We then detected apoptotic cells on gated CD3⁺ CD4⁺ CD45RA⁻ memory CD4 T cells with annexin V-V450 labeling (BD Biosciences) using the following surface antibody cocktail: anti-CD3-PE, anti-CD4-APC H7, and anti-CD45RA-ECD antibodies. The results shown represent the percentage of cells protected from Fas-mediated apoptosis when cells were treated with IL-2. The percent inhibition of apoptosis by IL-2 was determined using the following formula: [(percentage of cells with Fas-mediated apoptosis for cells not treated with IL-2 − percentage of cells with Fas-mediated apoptosis for cells treated with IL-2)/percentage of cells with Fas-mediated apoptosis for cells not treated with IL-2] × 100.

Detection of intracellular ROS.

ROS measurement was assessed by flow cytometry using a CM-H₂DCFDA (chloromethyl 2′,7′-dichlorodihydrofluorescein diacetate acetyl ester) probe (Life Technologies, Invitrogen) according to the manufacturer's protocol. Briefly, cells were incubated for 45 min with 10 μM CM-H₂DCFDA in a cell incubator (37°C), then washed three times in 1× PBS, and, finally, subjected to flow cytometry analyses. Upon oxidation by ROS, the nonfluorescent probe CM-H₂DCFDA is converted to the highly fluorescent 2′,7′-dichlorofluorescein (DCF), that is detectable at 525 nm (40).

Data analysis.

All statistical analyses comparing the groups of subjects (HIV^free, PHI, and CHI subjects) were performed using the nonparametric Mann-Whitney U test with the assumption of independent samples. This test, which uses the rank of the data rather than their raw values to calculate statistical significance, is an alternative to the t test when the assumption of normality cannot be tested in the case of a moderate sample size. A sample size of 6 subjects per group is sufficient to achieve a significant statistical power based on the observed changes. In contrast, we used the paired t test for analyzing the results of in vitro assays using noninfected memory CD4 T cells. Spearman's correlation test was used to identify the association among study clinical and immunological variables. We considered P values of less than 0.05 to be significant.

RESULTS

Loss of memory CD4 T cells during HIV-1 infection involves defective IL-2 signaling.

Early HIV-1 infection is characterized by the loss of CD4 T cells, including CD4 T cells from the memory compartment (5–7). To newly identify the underlying mechanisms associated with the depletion of memory CD4 T cells, we first measured their frequencies in primary HIV-1-infected (PHI), chronically HIV-1-infected (CHI), and HIV^free subjects. Table 1 summarizes the clinical features of the two groups of viremic subjects. The frequencies of viable CD3⁺ CD4⁺ CD45RA⁻ memory CD4 T cells were significantly lower in PHI and CHI subjects than HIV^free donors (7.4% ± 4.3%, 9.7% ± 3.5%, and 14.7% ± 4.7%, respectively; P < 0.05; n = 6) (Fig. 1A). Since IL-2, IL-7, and IL-15 play critical roles in the long-term survival of memory T cells (9–12), we next investigated the response to these cytokines in memory CD4 T cells from all groups by measuring the phosphorylation of STAT5. Ex vivo, uninduced and IL-7- and IL-15-induced phospho-STAT5 (pSTAT5) expression levels in memory CD4 T cells were similar in all groups. In contrast, we found that IL-2-induced pSTAT5 levels were significantly reduced in PHI and CHI subjects compared with those in HIV^free donors (mean fluorescence intensities [MFI], 4,223.8 ± 2,060.1, 4,794 ± 2,178.9, and 9,033.7 ± 1,303.8, respectively; P < 0.01; n = 6) (Fig. 1B). Similarly, the frequency of pSTAT5-positive memory CD4 T cells upon IL-2 stimulation was lower in PHI and CHI subjects than HIV^free subjects (36.4% ± 16.5%, 44.9% ± 15.8%, and 61.1% ± 4.3%, respectively, for PHI, CHI, and HIV^free subjects) (data not shown). We found no significant difference in the levels of IL-2-induced pSTAT5 in naive CD45RA⁺ cells among the subjects (Fig. 1B). The reduced response to IL-2 in memory CD4 T cells during HIV-1 infection concerned key memory subsets, such as the CD27⁺ CCR7⁺ central memory CD4 T cells (T_CM cells) and CD27⁺ CCR7⁻ transitional memory CD4 T cells (T_TM cells) (Fig. 1C) (41). The defective IL-2 response in memory CD4 T cells from HIV-1-infected subjects was not related to the decreased expression of the surface IL-2 receptor (CD25), since we found similar percentages of CD25-positive cells, including CD8 T cells and memory and naive CD4 T cells, in all subjects (Fig. 1D).

FIG 1.

Defective IL-2 signaling in memory CD4 T cells is associated with cell loss during HIV-1 infection. (A) Ex vivo frequency of memory CD4 T cells among PBMCs from PHI, CHI, and HIV^free subjects. (B) PBMCs from all groups were stimulated or not with cytokines for 15 min and then collected to assess pSTAT5 levels. (Inset) Representative dot plots for IL-2-induced pSTAT5 expression. (C) IL-2-induced pSTAT5 levels in viable CD3⁺ CD4⁺ CD45RA⁻ CD27⁺ CCR7⁻ T_CM cells, CD27⁺ CCR7⁻ T_TM cells, and CD27⁻ CCR7⁻ T_EM cells for all subject groups. (D) Percentage of CD25-positive cells. In panels A to D, data are for 6 subjects in each group. (E) Correlation between the level of IL-2-induced pSTAT5 and the frequency (i) (n = 18) or the absolute number (ii) (n = 12) of memory CD4 T cells. *, 0.05 > P > 0.01; **, 0.01 > P > 0.001.

We further determined whether the loss of memory CD4 T cells during HIV-1 infection is associated with defective IL-2 signaling. We confirmed correlations between the IL-2-induced pSTAT5 level and (i) the memory CD4 T-cell frequency determined in all study subjects or (ii) the absolute number of memory CD4 T cells by using cell counts, which were available only for HIV-1-infected subjects (P = 0.0127 [n = 18] and 0.0366 [n = 12], respectively) (Fig. 1E).

Finally, we assessed the protection from Fas-mediated apoptosis in memory CD4 T cells from PHI, CHI, and HIV^free subjects upon IL-2 stimulation. To this end, cells from all groups were first treated with anti-Fas CH11 antibody for 24 h in the presence or absence of IL-2 and then collected to assess apoptosis levels by annexin V staining. Memory CD4 T cells from PHI and CHI subjects displayed a higher sensitivity to Fas-mediated apoptosis and concomitantly showed increased expression of Fas receptor (CD95) compared to the response of memory CD4 T cells from HIV^free donors (Fig. 2A and B). IL-2 stimulation was effective at reducing Fas-mediated apoptosis in memory CD4 T cells in all subject groups (Fig. 2A). However, the levels of IL-2-associated protection were significantly lower in memory CD4 T cells from PHI and CHI subjects than in those from HIV^free donors (35.1% ± 12.8%, 44.8% ± 22.4%, and 80.2% ± 15.9%, respectively; P < 0.01; n = 6) (Fig. 2C). Interestingly, we found a significant correlation between the protection from Fas-mediated apoptosis with IL-2 treatment and the frequency of memory CD4 T cells for all study subjects (P = 0.0083; n = 18) (Fig. 2D).

FIG 2.

Memory CD4 T cells from HIV-1-infected subjects displayed reduced protection from apoptosis upon IL-2 stimulation. (A) Fas-mediated apoptosis of memory CD4 T cells for PHI, CHI, and HIV^free subjects in the presence or absence of IL-2 treatment. This was calculated using the following formula: percent apoptosis of cells treated with anti-Fas antibody − percent apoptosis of anti-Fas-untreated cells. (B) Ex vivo percentage of CD95-positive memory CD4 T cells. (C) Levels of apoptosis protection when cells were treated with IL-2. In panels A to C, data are for 6 subjects in each group. (D) Correlation between the level of apoptosis protection with IL-2 and the ex vivo frequency of memory CD4 T cells from all subjects (n = 18). *, 0.05 > P > 0.01; **, 0.01 > P > 0.001.

Overall, these data demonstrate that memory CD4 T cells, which are depleted during HIV-1 infection, showed a reduced capacity to resist Fas-mediated apoptosis in the presence of IL-2.

Defective IL-2 signaling in memory CD4 T cells is associated with the Kyn level during HIV-1 infection.

As early HIV-1 infection is characterized by persistent inflammation and enhanced tryptophan catabolism (22, 24–27, 42), we first evaluated the levels of several proinflammatory molecules and Kyn in plasma from the PHI, CHI, and HIV^free subjects. Consistent with earlier reports (22, 25, 43), we found higher levels of IL-6, sCD14, IP-10, and Kyn in PHI and CHI subjects than HIV^free donors (Fig. 3A to D). However, we did not observe any significant differences for LPS-binding protein (LBP) (Fig. 3E). More importantly, we found significant correlations between IL-2-induced pSTAT5 levels in memory CD4 T cells from all subjects and the CD4/CD8 ratio or the IL-6, sCD14, IP-10, or Kyn level (P = 0.0159, 0.023, 0.0267, 0.0149, and 0.0028, respectively; n = 18) (Fig. 3F).

FIG 3.

Reduced IL-2 signaling correlates with inflammation and Kyn levels during HIV-1 infection. (A to E) Levels of proinflammatory soluble factors and Kyn in plasma samples from all subject groups (n = 6). (F) Correlations between IL-2-induced pSTAT5 levels and the CD4/CD8 ratio (i), the viral load (VL) (ii), the concentrations of proinflammatory molecules (iii to v), or the Kyn concentration (vi) for all subjects (n = 18). Tryp, tryptophan. *, 0.05 > P > 0.01; **, 0.01 > P > 0.001; ns, not significant.

Collectively, these results indicate that HIV-1 disease progression, inflammation, and, in particular, the level of Kyn production were associated with reduced IL-2 signaling in memory CD4 T cells from infected subjects.

Treatment of CD4 T cells with Kyn inhibits IL-2 signaling through ROS.

From the plasma markers studied (Fig. 3), our observations indicated that the enhanced production of Kyn was the best correlate of reduced levels of IL-2-induced pSTAT5 expression during HIV-1 infection. To determine whether Kyn directly interferes with IL-2 signaling in memory CD4 T cells, purified CD4 T cells from HIV^free donors were treated in vitro with different concentrations of recombinant Kyn (0 to 50 μM) before being stimulated with IL-2. Our data showed that Kyn treatment led to a dose-dependent inhibition of IL-2-induced pSTAT5 levels in memory CD4 T cells, starting with a 5 μM concentration (which reduced the pSTAT5 levels 23% ± 8.9% compared to the level in untreated cells) (Fig. 4A). Of note, no apoptosis of memory CD4 T cells was induced after 48 h following Kyn treatments (with 2.5 to 50 μM concentrations) (data not shown). Since oxidative stress has been reported to be responsible for the apoptosis of natural killer cells mediated by concentrations of Kyn (44), we hypothesized that defective IL-2 signaling in memory CD4 T cells treated with Kyn could implicate ROS. To test this hypothesis, we first investigated whether 5 μM Kyn treatment resulted in ROS production by memory CD4 T cells. Our results confirmed that memory CD4 T cells cultured with 5 μM Kyn for 24 h displayed higher levels of expression of intracellular ROS than untreated cells (MFI, 1,798.4 ± 376.9 and 1,277.2 ± 214.5, respectively; P = 0.0385; n = 5) (Fig. 4B). We also preincubated CD4 T cells from uninfected subjects with increasing concentrations of H₂O₂ (0.1 to 100 μM), a powerful oxidant, prior to the stimulation with IL-2. Treatment with H₂O₂ (starting with a 1 μM concentration) led to both higher levels of ROS expression by memory CD4 T cells (Fig. 4B) and a decrease in the level of IL-2-induced pSTAT5 (Fig. 4C), confirming the contribution of the oxidative stress pathway.

FIG 4.

Kyn treatment inhibits the protection from Fas-mediated apoptosis provided by IL-2 and involves ROS production. (A) (Bottom) IL-2-induced pSTAT5 level on memory CD4 T cells in response to various Kyn concentrations. The percent inhibition of IL-2 signaling is indicated for each Kyn concentration. (Top) Representative histograms. (B) ROS expression by memory CD4 T cells in the presence or absence of 5 μM Kyn. We also stimulated CD4 T cells with increasing concentrations of H₂O₂. (C) CD4 T cells were preincubated with H₂O₂ for 1 h, stimulated with IL-2 for 15 min, and then collected to determine pSTAT5 levels in memory cells. (D) After 24 h of Kyn and/or NAC treatment, CD4 T cells were stimulated with IL-2 and pSTAT5 levels in memory cells were assessed. (E) Protection from Fas-mediated apoptosis by IL-2 in memory CD4 T cells in the presence or absence of Kyn and/or NAC. In panels A to E, data are for 5 subjects in each group. *, 0.05 > P > 0.01; **, 0.01 > P > 0.001; ***, P < 0.001.

To further demonstrate that Kyn-mediated defective IL-2 signaling in memory CD4 T cells involved ROS production, cells treated with Kyn were stimulated or not with the antioxidant N-acetylcysteine (NAC). Briefly, CD4 T cells were cultured with 5 μM Kyn for 24 h in the presence or absence of 5 mM NAC and then stimulated with IL-2 for 15 min. Cells were collected to determine pSTAT5 levels in the memory CD4 T-cell subset. Our results showed that the induction of pSTAT5 mediated by IL-2 was not affected by NAC treatment alone (Fig. 4D). We confirmed that treatment of cells with 5 μM Kyn for 24 h led to a significant reduction of IL-2-induced pSTAT5 levels. Defective IL-2 signaling in memory CD4 T cells treated with Kyn was partially ROS dependent, since the addition of NAC led to the significant but incomplete restoration of pSTAT5 levels (Fig. 4D). Kyn treatment of CD4 T cells also resulted in reduced protection of memory cells from Fas-mediated apoptosis by IL-2 (P = 0.0206 compared to the results for Kyn-untreated cells) (Fig. 4E). Similarly, the addition of NAC when cells were treated with Kyn led to significant improvements of memory CD4 T-cell survival (Fig. 4E).

Altogether, these data indicate that the oxidative stress generated by elevated levels of Kyn production during HIV-1 infection significantly interferes with IL-2 signaling in memory CD4 T cells.

ROS inhibits IL-2 signaling in memory CD4 T cells during HIV-1 infection.

We next investigated the involvement of oxidative stress in the reduced response to IL-2 in memory CD4 T cells from HIV-1-infected subjects. We first measured ex vivo the expression of intracellular ROS in memory CD4 T cells from PHI, CHI, and HIV^free subjects. We showed that the levels of ROS expression measured in memory CD4 T cells from PHI and CHI subjects were significantly increased compared to those measured in memory CD4 T cells from HIV^free subjects (MFI, 2,225.2 ± 568.9, 2,344.5 ± 469, and 1,267.2 ± 237.7, respectively; P = 0.0022; n = 6) (Fig. 5A). All memory cell subsets, including T_CM cells, T_TM cells, and CD27⁻ CCR7⁻ effector memory CD4 T cells (T_EM cells), from HIV-1-infected subjects displayed increased levels of ROS expression compared to the memory cell subsets from HIV^free donors (Fig. 5B). Correlations between the levels of ROS expression detected in memory CD4 T cells and the plasma Kyn level (Fig. 5C) or the level of IL-2-induced pSTAT5 expression (Fig. 5D) were observed for all subjects (P = 0.0022 and P < 0.0001, respectively; n = 18).

FIG 5.

Reduced IL-2 signaling in memory CD4 T cells during HIV-1 infection involves ROS production. (A) (Bottom) ROS levels in ex vivo memory CD4 T cells from PHI, CHI, and HIV^free subjects (n = 6 subjects in each group). (Top) Representative histograms, including histograms for H₂O₂ as a positive control. (B) ROS level in gated T_CM, T_TM, and T_EM cell subsets for all subjects (n = 6 subjects in each group). (C and D) Correlation between ROS and plasma Kyn level (C) or IL-2-induced pSTAT5 levels in memory CD4 T cells (D) (n = 18). (E) PBMCs from PHI, CHI, and HIV^free subjects were first incubated or not incubated with NAC for 24 h and stimulated with IL-2 for 15 min, and then pSTAT5 levels in memory CD4 T cells were assessed (n = 6). *, 0.05 > P > 0.01; **, 0.01 > P > 0.001.

Finally, PBMCs from PHI, CHI, and HIV^free subjects were cultured for 24 h with NAC, stimulated with IL-2 for 15 min, and then collected to assess the level of IL-2-induced pSTAT5 in memory CD4 T cells. Pretreatment of memory CD4 T cells from HIV-1-infected subjects with NAC resulted in a significant improvement in the level of pSTAT5 expression (P = 0.0079 and 0.0317 for PHI and CHI subjects, respectively; n = 6), but the levels did not reach those for HIV^free subjects (Fig. 5E).

Together, these results showed that ROS are involved with the inability of memory CD4 T cells to optimally respond to IL-2 during HIV-1 infection.

ART improves IL-2 signaling in memory CD4 T cells by reducing ROS and Kyn levels.

In addition to improving the life expectancy of treated patients by controlling HIV-1 replication, ART inhibits the immunosuppressive tryptophan catabolism (22, 34, 45, 46). To test whether ART could restore the ability of memory CD4 T cells from PHI subjects to respond to IL-2, we collected plasma and PBMCs from HIV-1-infected subjects before treatment (pre-ART) and after 1 year of effective treatment (on ART). Figure 6A summarizes the clinical features of HIV-1-infected subjects pre-ART and on ART. First, we showed by analyzing plasma concentrations that ART was effective in reducing Kyn production in HIV-1-treated subjects (2.0 ± 0.5 μM and 1.3 ± 0.4 μM, respectively, for the pre- and on-ART conditions; P = 0.0317; n = 5) (Fig. 6B). Although the Kyn concentration appeared to be lower in plasma from HIV^free donors (0.61 μM ± 0.03), only a trend between Kyn levels from subjects on ART and uninfected subjects was noted (P = 0.1905; n = 5). We next determined the levels of ROS expression by memory CD4 T cells from HIV^free donors and subjects pre-ART and on ART. Whereas memory CD4 T cells from subjects pre-ART displayed significantly increased ROS levels compared to those of memory CD4 T cells from subjects on ART and HIV^free subjects (MFI, 2,128 ± 981.9, 933.8 ± 156.2, and 993.4 ± 147.1, respectively; P = 0.0079; n = 5), we found a complete suppression of oxidative stress in ART recipients (Fig. 6C). PBMCs from HIV^free subjects and subjects pre-ART and on ART were also stimulated for 15 min with IL-2 to determine pSTAT5 levels in memory CD4 T cells. Memory CD4 T cells from subjects pre-ART displayed lower pSTAT5 levels than memory CD4 T cells from subjects on ART and HIV^free subjects (P = 0.0079 for both comparisons; n = 5) (Fig. 6D). Once again, we found no difference between pSTAT5 levels in memory CD4 T cells from ART recipients and those from uninfected donors (P = 0.5476; n = 5). Importantly, we observed correlations between the plasma concentrations of Kyn and (i) ROS or (ii) IL-2-induced pSTAT5 levels with all study subjects (P = 0.002 and 0.014, respectively; n = 15) (Fig. 6E).

FIG 6.

ART restores proper IL-2 signaling in memory CD4 T cells by reducing ROS and Kyn levels. (A) Clinical features of 5 HIV-1-infected patients in each subject group at the pre- and on-ART longitudinal time points. d, number of days; undetectable#, the VL was <40 copies/ml. (B to D) Measurement of Kyn (B), ROS (C), and IL-2-induced pSTAT5 (D) levels for all subject groups. In panels A to D, data are for 5 subjects in each group. (E) Correlations between Kyn and ROS (i) or IL-2-induced pSTAT5 (ii) levels for subjects pre-ART, subjects on ART, and HIV-1^free subjects (n = 15). *, 0.05 > P > 0.01; **, 0.01 > P > 0.001.

Collectively, these results provide evidence that 1 year of ART restored the response to IL-2 and reduced the levels of both Kyn production and oxidative stress in memory CD4 T cells from treated HIV-1-infected patients (Fig. 7).

FIG 7.

Interplay between inflammation, increased kynurenine levels, and reduced IL-2 signaling in memory CD4 T cells during HIV-1 infection.

DISCUSSION

Herein, we revealed a newly discovered metabolism-related mechanism responsible for inhibition of the IL-2 response and the subsequent loss of memory CD4 T cells during HIV-1 infection. Although we found no significant difference in the IL-7 and IL-15 responses in memory CD4 T cells from all subject groups, our results also showed that following cytokine stimulation, CD27⁻ CCR7⁻ T_EM cells from HIV^free donors displayed lower pSTAT5 levels than those from PHI and CHI subjects, as previously reported (14) (P = 0.055 and 0.026, respectively, for IL-7 and P = 0.0087 and 0.0043, respectively, for IL-15).

Collectively, the ex vivo and in vitro data demonstrated that the enhanced production of Kyn was one of the molecular mechanisms associated with memory CD4 T-cell loss and interference with IL-2 signaling (Fig. 7). Tryptophan catabolism into Kyn is driven by the IDO-1 enzyme expressed by monocyte/macrophages and DCs. Several soluble factors associated with immune activation, including IFN-γ, tumor necrosis factor alpha, IL-1β, sCD40, IL-32, and Toll-like receptor ligands, are known to induce IDO-1 activity in these cells (42). Strain-dependent HIV-1 infection and Tat expression also lead to increased IDO-1 activity in innate immune cells (47). Interestingly, the in vitro assays using Kyn treatment indicated that a low 5 μM concentration was sufficient to reduce the ability of noninfected memory CD4 T cells to respond to IL-2 (Fig. 4A). Similar concentrations were previously reported in plasma from viremic subjects (22, 29).

The fact that NAC partially improved memory CD4 T-cell survival from Fas-mediated apoptosis demonstrates that ROS production is one of the mechanisms contributing to defective cell survival during HIV-1 infection. Similarly, previous observations showed that oxidative stress is responsible for the induction of apoptosis in several cell types, including natural killer cells, enterocytes, podocytes, and astrocytes, during HIV-1 infection (44, 48–50). Recent studies, such as those conducted by Kalinowska et al. (14) and Trautmann et al. (16), also showed that decreased IL-7-induced pSTAT5 levels in CD8 T cells during HIV-1 infection were associated with oxidative stress.

Kyn-induced defective IL-2 signaling in memory CD4 T cells during HIV-1 infection not only was mediated by elevated oxidative stress but also involved other ROS-independent mechanisms. Although in the present study the results of the in vitro assays using Kyn treatment implicated oxidative stress, the addition of NAC only partially restored IL-2 signaling and cell survival compared to the levels of IL-2 signaling and survival found for cells cultured in the absence of Kyn (Fig. 4D and E). These observations were further supported by similar results with memory CD4 T cells from PHI and CHI subjects and those from HIV^free donors (Fig. 5E). Finally, our data also showed that 1 μM H₂O₂ treatment resulted in similar levels of ROS production following IL-2 treatment and Kyn treatment (Fig. 4B) but led to lower levels of reduction of pSTAT5 levels following IL-2 treatment than Kyn treatment (Fig. 4C and D). Collectively, such results indicate underlying mechanisms which are Kyn dependent but ROS independent. A previous study demonstrated that the activation of the aryl hydrocarbon receptor, which is triggered by several agonists, including Kyn (51), interferes with the activation of STAT5 by IL-2 through direct binding with the STAT protein (52). Therefore, we cannot exclude the possibility of the involvement of the aryl hydrocarbon receptor in reducing IL-2 signaling in memory CD4 T cells as an additional mechanism which does not necessarily involve oxidative stress.

Our data also indicate that the loss of memory CD4 T cells and interference with IL-2 signaling during HIV-1 infection depend on other Kyn-independent mechanisms. In agreement with findings reported in the literature, we demonstrated that the diminished STAT5 phosphorylation found during the early phase of HIV-1 infection correlated not only with enhanced ROS and Kyn levels but also with enhanced levels of several inflammation markers, such as IL-6, sCD14, and IP-10 (Fig. 3A to C) (22, 25, 29). Therefore, we cannot exclude the possibility that other mechanisms, which could involve the release of intermediate molecules such as IFN type I and other proinflammatory markers by monocyte/macrophages and DCs, are at play in our observations. This is particularly true considering that previous reports have indicated that the IFN-α produced by plasmacytoid DCs interferes with the ability of CD4 T cells to respond to IL-2 and survive (53, 54). Novel discoveries on the interplay between inflammation and viral persistence shed new light on the key role of the inflammasome, especially for CD4 T-cell loss, during HIV-1 infection (1, 24, 55). The inflammasome is a key signaling platform that detects ROS, pathogen-associated molecular patterns, and metabolic perturbations (56–58). Inflammasome activation accounts for the majority of CD4 T-cell death by pyroptosis in HIV-1-infected lymphoid tissues (1, 2, 59). In fact, the results of studies with macaques indicate that inflammasome activation represents one of the earliest events which follow mucosal simian immunodeficiency virus infection (60). Interestingly, Sagulenko et al. recently demonstrated that the inflammasome activates both apoptotic and pyroptotic cell death, whose balance depended on the amounts of stimuli (61). Therefore, since the results implicate both inflammation and ROS in the loss of memory CD4 T cells in HIV-1-infected patients, inflammasome activation could represent an additional mechanism at play in our study.

Gut damage and the subsequent microbial translocation from tissue to the periphery have also been recognized to be major contributors to HIV-1-related inflammation (21, 62). Results from a study by Gaardbo et al. showed that microbe-related effects in the gut are associated with CD4 T-cell loss during HIV-1 infection (35). The dysbiosis that occurs in the gut of HIV-1-infected subjects persisting on ART describes the imbalance of bacterial diversity that is illustrated by increased proinflammatory pathogenic strains (63, 64). Importantly, the enhanced tryptophan metabolism related to the composition of the gut microbiota has been linked to the disease progression contributed by enhanced Kyn production in the gut (65). Conversely, the administration of probiotic bacteria in HIV-1-infected subjects transiently increases the CD4 T-cell count and reduces gut inflammation (66, 67). Therefore, because of the accumulation of Kyn, the gut microbial imbalance could interfere with memory CD4 T-cell survival.

Finally, our data reinforce the notion that the early initiation of ART could limit metabolic and CD4 T-cell immune dysfunction. Despite relatively low HIV-1-specific CD8 T-cell responses, recent observations demonstrate a beneficial effect of early initiation of ART in preserving mucosal CD4 T cells, possibly limiting the release of microbial products associated with immune activation (68–70). It is important to note that the complete reestablishment of IDO-1 activity occurs only when ART is initiated early, in contrast to the partial normalization that occurs when ART is initiated during the chronic phase of infection (34, 45, 46). Accordingly, we also found a drastic suppression of the Kyn-dependent inhibition of the IL-2 response in memory CD4 T cells after 1 year of ART when ART was initiated during the first months of HIV-1 infection (initiation at 124.6 ± 106.4 days postinfection) (Fig. 6). However, in ART recipients, Kyn levels were not fully normalized, while the oxidative stress and IL-2-induced pSTAT5 levels were similar to those seen in HIV^free subjects (Fig. 6B to D). These data strongly suggest that the reduced memory CD4 T-cell survival occurring during HIV-1 infection depends on multiple parameters, including Kyn-independent mechanisms.

In conclusion, our study provides another piece of evidence indicating that the enhanced metabolic activity occurring at the onset of primary infection plays a key role in interfering with the host molecular signaling involved in the survival of memory CD4 T cells. This information should be considered in the development of therapeutic strategies.

Funding Statement

This work was supported in part by the Fonds de la Recherche Québec-Santé (FRQ-S): Réseau SIDA/Maladies infectieuses and Thérapie cellulaire, the Canadian Institutes of Health Research (CIHR) (grants MOP 103230 and CTN 257), and the Canadian HIV Cure Enterprise (CanCURE) (Team Grant HIG-133050). CanCURE is a partnership of the CIHR, the Canadian Foundation for AIDS Research (CANFAR), and the International AIDS Society (IAS). The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.