T cell functionality in HIV‐1, HIV‐2 and dually infected individuals: correlates of disease progression and immune restoration

Summary

The role of suppressive anti‐retroviral therapy (ART) in eliciting restoration of dysregulated immune function remains unclear in HIV‐1 infection. Also, due to tailoring of therapeutic regimens towards HIV‐1, this possible impairment of therapy may be even more pronounced in HIV‐2 and dual (HIV‐D) infection. Thus, we evaluated the impact of ART on immune restoration by assessing T cell functions, including HIV specific responses in HIV‐1‐, HIV‐2‐ and HIV‐D‐infected individuals. Both ART‐treated and naive infected subjects showed persistently altered frequency of CD4⁺ T cell subsets [regulatory T cells (T_reg), naive/central memory, effector memory], increased immune activation, cytoxicity and decreased frequency of natural killer T (NKT)‐ like cells and T helper type 17 (Th17)/T_reg ratio with elevated microbial translocation. Further, HIV‐specific responses were dominated by gag‐specific CD4⁺ T cells in virologically suppressed HIV‐D individuals, suggesting retention of T cell memory for both viruses. Increased antigen‐specific responses, including dual‐functional interleukin (IL)‐2/interferon (IFN)‐γ CD4⁺ T cells, were detected in therapy receiving HIV‐2‐infected individuals indicative of a greater and more functionally diverse T cell memory repertoire. We delineated immune signatures specific to therapy‐naive single HIV infection, as well as a unique signature associated with HIV‐2 disease progression and immune restoration. Circulating T_reg frequency, T cell activation and microbial translocation levels correlated with disease progression and immune restoration among all types of HIV infection. Also, memory responses negatively correlated, irrespective of type of infection, in ART receiving infected individuals, with CD4 rebound and decreased pan T cell activation. Our data highlight the need for adjunct immunomodulatory therapeutic strategies to achieve optimal immune restoration in HIV infection.

Introduction

The HIV epidemic in India is relatively unique, in that infection with both HIV‐1 and HIV‐2 (i.e. dual infection; HIV‐D), along with single HIV‐1 and HIV‐2 infection, are clearly present 1. However, limited data exist on the impact of dual infection on T cell dysregulation, immune activation and ex‐vivo antigen‐specific responses 2. Pathogenesis associated with HIV‐D has been hypothesized to follow an exacerbated course due to possible hyperimmune activation, resulting from additive or synergistic interactions of both viruses. Alternatively, a disease course either similar to or attenuated compared to that of HIV‐1 single infection may also be occurring due to the interplay of individual viral replication kinetics 2, 3. HIV‐2 infection is considered to be less pathogenic, as it is associated with a significantly longer asymptomatic stage, lower plasma viral load, slower decline in CD4⁺ T cell count and lower mortality rate attributable to AIDS compared to HIV‐1 4, 5. Also, a well‐preserved and polyfunctional HIV‐specific memory CD4⁺ T cell response has been shown to be a hallmark of HIV‐2 infection 6, 7.

According to the National AIDS Control Organization (NACO)‐2013 anti‐retroviral therapy (ART) guidelines for HIV‐infected adults and adolescents, the clinical goal of ART is to provide maximal and durable suppression of viral load, which is expected to lead to an increase in the CD4 count and an accompanying complete/partial restoration of pathogen‐specific immune function 8. It is important to note, however, that most of the anti‐viral drugs and regimens have been designed and optimized for HIV‐1, and cannot be assumed to provide optimal viral suppression for HIV‐2 and HIV‐D infection 9. Due to the presence of natural polymorphisms, HIV‐2 is intrinsically resistant to the fusion inhibitors and the non‐nucleoside reverse transcriptase inhibitor (NNRTI)‐based regimens, which are standard therapy for HIV‐1 in India 9, 10. Furthermore, data with respect to effect of ART on immune function restoration is divergent and has reported only on HIV‐1 infection 11, 12. Also, limited studies on HIV‐1‐specific responses in ART‐receiving infected individuals show conflicting results 13. Some studies have reported continuous virological suppression leading to a time‐dependent reduction in the frequencies of HIV‐specific CD4⁺ T and CD8⁺ T cells 14, 15, 16, while others have reported detection of sustained HIV‐specific responses in ART‐receiving HIV‐1‐infected individuals 17, 18.

Thus, in this study we sought to investigate impact of ART on immune restoration by assessing systemic T cell functions and HIV‐specific T cell responses, particularly in HIV‐2 and HIV‐D, together with HIV‐1‐infected individuals from the same clinical setting. We assessed important cellular parameters, such as level of immune activation – a marker of disease progression and frequency of cytotoxic T cells (CTL) expressing the effector molecule granzyme B (GrzB), that are reported to confer host cell immunity against viral pathogens 19, 20, 21. CD4⁺ T cell subsets such as regulatory T cells (T_regs) (CD25^highCD127^low), effector memory (CD127^–CD25^–) and naive/central memory (CD127⁺CD25^low/−), defined on the basis of expression of two important homeostatic markers, interleukin (IL)‐2 receptor (CD25) and IL‐7α receptor (CD127) proposed as targets for immune‐based interventions, were also studied 22, 23, 24, 25, 26. In addition, the frequency of NK T cells, a unique immunoregulatory T cell population, shown to be involved in host immunity against HIV‐1 associated opportunistic infections, was examined in this study 27.

T helper type 1 (Th1) and Th17 cells, recognized as prominent proinflammatory cell types, are linked functionally and are vital for host defense against pathogens 28, 29. T_regs are developmentally linked with Th17 cells 30, and their interplay with the aforementioned subsets is thought to be critical in HIV disease progression. Thus, the impact of HIV infection as well as ART on the interplay of Th1/Th17/T_reg subsets was examined in this study.

Microbial translocation has been shown to play a key role in driving persistent immune activation driving disease progression in chronic HIV infection 31, 32. Also, Th17/T_reg balance predominantly maintains gut homeostasis 30. Thus, both these parameters and impact of ART on them were examined concurrently in single‐ as well as dually infected individuals.

Methods

Study subjects

A total of 126 HIV‐infected individuals and 16 HIV‐seronegative individuals were recruited from the ART Centers, Grant Medical College and Sir J. J. Group of Hospitals, Mumbai with approval from the NIRRH Institutional Clinical Ethics Committee (Project no.: 225/2012), after providing informed consent for participation in the study. The HIV‐infected individuals were stratified into three major groups: 61 HIV‐1‐, 50 HIV‐2‐ and 15 HIV‐D‐infected individuals. The HIV‐1 and HIV‐2 groups were further stratified by the presence and absence of ART into the defined subgroups: ART‐naive and ART‐receiving groups, as outlined in Table 1. Total nucleic acid from blood was isolated using the MagNa pure Compact Nucleic Acid Automated System (Roche Diagnostic, Mannheim, Germany) and the plasma viral load of HIV‐1‐infected individuals was estimated using Cobas TaqMan real‐time polymerase chain reaction (PCR) (Roche Molecular Systems, Piscataway, NJ, USA), with a limit of detection of 34 RNA copies/ml. Due to the non‐availability of a standard, reproducible HIV‐2 viral load estimation assay and thus the inability to compare viral loads from HIV‐1 infected individuals, we were unable to estimate the HIV‐2 viral load in HIV‐2‐ as well as HIV‐D‐infected individuals.

Table 1.

Clinical characteristic of participants

	HIV‐1		HIV‐2		HIV‐D ART‐receiving	Seronegative
	ART‐naive	ART‐ receiving	ART‐naive	ART‐ receiving
Number of participants (female/male ratio)	31 (0·34)	30 (0·36)	18 (0·38)	32 (0·33)	15 (0·36)	16 (0.45)
Age † (years)	46·00 (28–59)	42·00 (31–56)	50·50 (34–58)	45·00 (31–58)	47·00 (30–51)	43.50 (35–56)
CD4 cell count † (cells/μl)	463·5 (196·2–795·1)	472·3 (132·4–1196)	560·8 (367·7–1229)	393·2 (179·2–1005)	494·8 (203·5–1044)	765.3 (523.5–1401)
CD8 cell count † (cells/μl)	1291 (250·3–3503)	946·9 (225·7–2323)	616·2 (461·8–2413)	748·7 (232·5–1912)	1217 (260·1–1619)	533.5 (106.9–1407)
CD4/CD8 ratio	0·3485	0·5077	0·8449	0·6417	0·5901	1.361
Viremia † (RNA copies/ml)	8200 (1120–196732)	TND	n.a.	n.a.	TND ‡	–
Duration between diagnosis and sampling † (years)	2 (1–4)	4 (1–8)	2 (1–4)	4 (1–7)	3 (1–6)	–
Duration between ART initiation and sampling (years)	–	3 (1–4)	–	2·5 (1–5)	3 (1–4)	–

^†Data are expressed as the median with range; TND = target not detectable/viral load < 34 copies/ml;

^‡HIV‐1 viral load in HIV‐D‐infected individuals; n.a. = data not available; ART = anti‐retroviral therapy.

Flow cytometry staining and analysis

For immunophenotypical staining, peripheral blood collected in an ethylenediamine tetraacetic acid (EDTA) vacutainer was stained with appropriate fluorochrome‐conjugated surface antibodies, including anti‐CD3 (clone: SK7), anti‐CD4 (clone: RPA‐T4), anti‐CD8 (clone: SK1), anti‐CD25 (clone: M‐A251), anti‐CD127 (clone: HIL‐7R‐M21), anti‐human leukocyte antigen D‐related (HLA‐DR) (clone: L243), anti‐CD38 (clone: HIT2) and anti‐CD45RA (clone: HI100), anti‐granzyme (clone: GB11), purchased from either BD Biosciences (San Jose, CA, USA) or Biolegend (San Diego, CA, USA). Intracellular staining for GrzB was performed according to the manufacturer’s instructions (BD Cytofix/Cytoperm™ Plus, cat. no.: 554715) after surface staining with specific surface marker antibodies. The samples were processed on the same day of sampling for ex‐vivo staining and Intracellular cytokine staining assay (ICCS) assay for GrzB detection. Flow cytometric acquisition and analysis were performed on at least 50 000 acquired events (gated on lymphocytes) on a BD ACCURI C6 flow cytometer (BD Biosciences). Data analysis was performed using FlowJo (Tree Star Inc., Ashland, OR, USA).

HIV‐1 and HIV‐2 gag and env immunodominant peptide set

Four different sets of immunodominant peptides were utilized in these experiments: (i) four immunodominant peptides, 15–20 amino acids in length derived from HIV‐1 env (PNNNTRKSIRIGPGQTFYA; QSNLLRAIEAQQHMLQLTVW; LSIVNRVRQGYSPLSFQTLT; RDRSIRLVSGFLALAWDDLR) corresponded to the sequence of isolate 96ZM651.8 (Accession no: AF286224) described in 33; (ii) three immunodominant peptides 15–20 amino acids in length derived from HIV‐2 env (PGNKTVTPITLMSGYKFHS; QQQLLDVVKRQQEMLRLTVW; GYRSPWPIAYIHFLIHLLIR) corresponded to the sequence isolate of CRIK147 (Accession no.: ABC39624.1); (iii) two immunodominant peptides 15–20 amino acids in length derived from HIV‐1 gag (FKTLRAEQATQEVKNWMTDT; MFTALSEGATPQDLNTMLNT) corresponded to the sequence of isolate 96ZM651.8 (Accession no.: AF286224) described in 34; and (iv) three immunodominant peptides 15–20 amino acids in length derived from HIV‐2 gag (CTPYDINQMLNCVGD; MFRPQNPVPVGNIYRRWIQI; RFYKSLRAEQTDPAVKNWMT) corresponded to the sequence isolate of HIV‐2 ROD isolate (Accession no.: M15390) previously described in 34. The peptides were synthesized from ABclonal Biotechnology (Woburn, MA, USA).

Ex‐vivo stimulation of cells

Peripheral blood mononuclear cells (PBMC) were isolated using lymphocyte separating media (Himedia) density centrifugation; 1 × 10⁶ PBMC were incubated with each peptide set (final concentration, 2 µg/ml/peptide) together with co‐stimulatory antibodies (anti‐CD28 and anti‐CD49d; 1 µg/ml each; BD Biosciences). HIV‐1 or HIV‐2 peptide sets were utilized according to the status of the individual and PBMCs from HIV‐D‐infected individuals were stimulated separately with both HIV‐1 and HIV‐2 peptide sets. In every experiment a negative control containing only anti‐CD28 and anti‐CD49d was used to measure antigen‐independent stimulation and a positive control, phorbol 12‐myristate 13‐acetate (PMA) (100 ng/ml) and ionomycin (1 mg/ml; Sigma‐Aldrich, St Louis, MO, USA) were used to ensure that cells were responsive during each experiment. The cultures were incubated for 1 h at 37°C in a 5% CO₂ incubator, followed by an additional 5 h in the presence of the secretion inhibitor monensin (Golgistop, 0·7 µl/ml; BD Biosciences). To estimate the frequency of Th1 and Th17 cells, the intracellular interferon (IFN)‐γ and IL‐17 production was accessed from PBMC stimulated with negative and positive controls.

Intracellular cytokine staining assay

Following incubation, cells were washed using wash buffer [phosphate‐buffered saline (PBS containing 1% serum bovine albumin)] and stained with pretitered CD3, CD4/CD8 antibodies. Cells were then washed and fix/permeabilized with Cytofix/Cytoperm according to the manufacturer’s instructions (BD PharMingen, San jose, CA, USA). Following permeabilization, cells were washed twice with Cytofix/Cytoperm wash buffer and stained intracellularly with pretitered antibodies specific for IL‐17 (clone: SCPL1362), IFN‐γ (clone: 4S.B3), IL‐2 (clone: MQ1‐17H12). Following staining, cells were washed and resuspended in wash buffer and acquired on an ACCURI C6 flow cytometer (BD Biosciences). At least 100 000 events/lymphocyte gates were collected per sample for analysis. FlowJo software was used to perform the analysis.

Markers of microbial translocation

Plasma levels of sCD14 (µg/ml) was quantified by using a human sCD14 enzyme‐linked immunosorbent assay (ELISA) kit (R&D systems, Minneapolis, MN, USA). Plasma (10 µl) was diluted 400‐fold and assayed as per the manufacturer’s recommendations. Plasma lipopolysaccharide (LPS) levels (pg/ml) were measured by using a LPS ELISA kit (MyBioSource; cat. no.: MBS702450), according to the manufacturer’s recommendations.

Statistical analysis

Differences between groups were analyzed by paired and unpaired t‐tests with Welch’s correction (95% confidence interval). Correlations between two variables were evaluated by non‐parametric Spearman’s Pearson correlation was used to plot graphs. Statistical analysis was performed using GraphPad Prism version 5.00 (GraphPad Software, San Diego, CA, USA). P‐values < 0·05 were considered significant.

Results

Study participant characteristics

Clinical characteristics of all study groups are shown in Table 1. In total, 31 ART‐naive HIV‐1, 30 ART‐receiving HIV‐1, 18 ART‐naive HIV‐2, 32 ART‐receiving HIV‐2, 15 ART‐receiving HIV‐D and 16 HIV‐seronegative individuals were enrolled into this study. ART regimens were based on NACO guidelines 8 as follows: HIV‐1, zidovudine (AZT) + lamivudine (3TC) + nevirapine (NVP); HIV‐2, zidovudine (AZT) + lamivudine (3TC) + lopinavir/ritonavir (LPV); HIV‐D, zidovudine (AZT) + lamivudine (3TC) + lopinavir/ritonavir (LPV). In ART‐receiving individuals, therapy had been initiated for at least 1 year with no significant difference in duration. Male/female ratio and age were equally distributed across groups. The absolute CD4 count and CD4/CD8 ratio of all HIV groups was significantly lower compared to the seronegative group. The absolute CD4 count as well as the CD4/CD8 ratio was significantly higher in ART‐naive HIV‐2 compared to ART‐naive HIV‐1. The absolute CD8 count in all HIV groups was higher compared to the seronegative group. The ART‐naive HIV‐1 group had a significantly higher absolute CD8 count than ART‐naive HIV‐2 and ART‐receiving HIV‐1. To evaluate the efficacy of ART in terms of CD4 rebound, we compared the absolute CD4 counts at ART start‐time and CD4 count at the sampling time in all therapy‐receiving individuals. A significant increase in the CD4 count was observed in all ART‐receiving groups. Also, the fold change observed in the CD4 count was similar for all ART‐receiving groups (Supporting information, Fig. S1). These data also suggest that ART‐receiving HIV‐2 infected individuals were responding to ART, although the CD4 count was lower at the time of sampling in these individuals compared to ART‐receiving HIV‐1 and ART‐naive HIV‐2‐infected individuals.

Increased systemic level of immune activation and T cell cytotoxicity across all HIV‐infected groups irrespective of viremia and ART status

The frequency of activated CD4⁺ T cells (CD4⁺HLA‐DR⁺CD38⁺), activated CD8⁺ T cells (CD4⁺HLA‐DR⁺CD38⁺), CD4⁺ cytotoxic T cells (CD4⁺Grz‐B⁺) and CD8⁺cytotoxic T cells (CD8⁺Grz‐B⁺) were significantly higher in HIV‐1, HIV‐2 and HIV‐D groups compared to the seronegative group (Fig. 1). Moreover, the frequencies of activated CD4⁺ T cells and activated CD8⁺ T cells were decreased, although not significantly (P = 0·0836 and P = 0·0553, respectively), in ART‐receiving compared to ART‐naive HIV‐1 subjects. The frequency of activated CD4⁺ and CD8⁺ T cells were significantly lower in the ART‐naive HIV‐2 group compared to the ART‐naive HIV‐1 group (Fig. 1a,b). The observed differences could possibly be explained in part by the higher absolute CD4 count (Table1) as well as the lower viral load, as has been reported previously for HIV‐2 infection 35, 36. Additionally, no significant differences in the frequency of cytotoxic T cells were observed between the ART‐naive and ‐receiving groups of infected individuals (Fig. 1c,d).

Figure 1.

Immune activation and cytotoxicity profiles in HIV‐1, HIV‐2 and HIV‐D infection. Comparison of expression of immune activation markers [CD38 and human leukocyte antigen D‐related (HLA‐DR)] on CD4⁺ T (a) and CD8⁺ T cells (b) as well as granzyme B (GrzB)‐producing CD4⁺T (c) and CD8⁺ T cells (d) among HIV‐1‐, HIV‐2‐, HIV‐D‐infected and seronegative individuals. Gating strategy of activated cells and GrzB⁺ cells are outlined in Supporting information, Fig. S2. Statistical significance was evaluated by unpaired t‐test; *P < 0·05; **P < 0·01; ***P < 0·001.

When frequencies of activated CD4⁺ T cells and activated CD8⁺ T cells were correlated with absolute CD4 count, a significant negative correlation was observed in all HIV‐infected groups. However, only CD8⁺ cytotoxic T cells showed a significant negative association with absolute CD4 count in ART‐receiving HIV‐1‐ and HIV‐2‐infected groups (Table 2).

Table 2.

Association between absolute CD4 count and T cell functions of HIV disease progression

	Treg	CD4+ naive/ central memory subset	CD4+ effector memory subset	CD4+ HLA‐DR+ CD‐38+	CD8+ HLA‐DR+ CD38+	CD4+ GrzB+	CD8+ GrzB+	sCD14	LPS	Th17 cells	Th17/Treg
HIV‐1 ART‐naive	–0·377, 0·0434	0·5362, 0·0027	–0·4062, 0·0288	–0·5660, 0·0011	–0·4764, 0·0078	0·0243, 0·9101	0·1626, 0·4477	–0·8179, 0·0002	–0·828, 0·0001	0·4531, 0·0176	0·3870, 0·0419
HIV‐1 on‐ART	–0·364, 0·0475	0·3004, 0·1068	–0·1218, 0·5214	–0·5004, 0·0049	–0·4928, 0·0057	–0·144, 0·5330	–0·539, 0·0117	–0·8163, 0·0004	–0·748, 0·0021	–0·290, 0·1597	–0·0713, 0·7079
HIV‐2 ART‐naive	–0·822, < 0·0001	0·6945, 0·0014	–0·4798, 0·0439	–0·7028, 0·0011	–0·5555, 0·0167	0·0833, 0·7505	0·0797, 0·7611	–0·7399, 0·0016	–0·864, < 0·0001	0·570, 0·0212	0·6069, 0·0098
HIV‐2 on‐ART	–0·360, 0·0428	0·3711, 0·0365	–0·5617, 0·0008	–0·4029, 0·0246	–0·4428, 0·0126	0·0451, 0·8094	–0·365, 0·0433	–0·7399, 0·0016	–0·828, 0·0001	–0·261, 0·1624	–0·3185, 0·0922
HIV‐D on‐ART	–0·571, 0·0261	0·5964, 0·0189	–0·6404, 0·0101	–0·5714, 0·0261	–0·5357, 0·0396	0·1182, 0·7345	–0·109, 0·7545	–0·6756, 0·0057	–0·792, 0·0004	–0·252, 0·3644	–0·2359, 0·3973

Associations were tested for significance with Spearman’s rank correlation. P‐values of 0·05 were considered significant and are shown in bold type. Value (r, P) for infected cohorts. LPS = lipopolysaccharide; ART = anti‐retroviral therapy; HLA‐DR = human leukocyte antigen D‐related; GrzB = granzyme B; Th17 = T helper type 17.

Dysregulation in the frequencies of homeostatically significant CD4⁺ T cell subsets

The relative proportion of CD4⁺ T cell subsets (T_reg, effector memory and naive/central memory subsets) based on the expression of homeostatic markers CD25 (IL‐2Rα) and CD127 (IL‐7R) were examined in ART‐naive and ‐receiving HIV‐1, HIV‐2 and HIV‐D subjects. We observed a significant increase in the frequency of T_regs (CD25^highCD127^low) and effector memory (CD127^–CD25^–) subsets and a decline in the fraction of naive/central memory (CD127⁺CD25^low/−) T cell subsets in all ART‐naive and ‐receiving groups compared to seronegative controls. Also, the frequency of these CD4⁺ T cell subsets was found to be similar in ART‐naive and ‐receiving groups. These data confirm our previous findings, and extend these to the HIV‐D group 37. Also, T_reg (CD25^highCD127^low) frequency was lower, although not significantly (P = 0·0599), in ART‐receiving, compared to ART‐naive, HIV‐1‐infected individuals (Fig. 2).

Figure 2.

Identification of CD4⁺ T cell subsets based on the expression of CD127 [interleukin (IL)‐7R] and CD25 (IL‐2Rα). Comparison of frequencies of CD4⁺ T cell subsets in HIV‐1‐, HIV‐2‐ and HIV‐D‐infected individuals and seronegative individuals. Gating strategy is outlined in Supporting information, Fig. S2. Statistical significance was evaluated by unpaired t‐test; *P < 0·05; **P < 0·01; ***P < 0·001.

T_reg (CD25^highCD127^low) frequency was consistently and significantly negatively associated with absolute CD4 count in all HIV‐infected groups, suggesting its utility as a correlate of immune restoration in pan‐HIV infection that could be evaluated along with the CD4 count. Effector memory (CD127^–CD25^–) cells were found to be inversely correlated and the naive/memory (CD127⁺CD25^low/−) T cell subset was found to be positively correlated with absolute CD4 count in ART‐naive HIV‐1, ART‐naive HIV‐2, ART‐receiving HIV‐2 and ART‐receiving HIV‐D groups (Table 2). Also, activated CD4⁺ T cells and activated CD8⁺ T showed a significant positive correlation with frequency of T_regs in the the ART‐naive HIV‐1, ART‐naive HIV‐2 groups and ART‐receiving HIV‐D group. Only activated CD4⁺ T cells showed a significant positive correlation with T_regs in ART‐receiving HIV‐1 and ART‐receiving HIV‐2 groups (Table 3).

Table 3.

Association data of T cell functions with each other across HIV groups

		Treg	CD4+ HLA‐DR+CD38+	CD8+ HLA‐DR+CD38+	Th17 cells	Th17/Treg	sCD14
HIV‐1 ART‐naive	Treg	–	0·3896, 0·0367	0·3765, 0·0441	–0·4605, 0·0179	0·1327, 0·5181	–
	CD4+HLA‐DR+CD38+	–	–	–	–0·4630, 0·0150	–0·3782, 0·0358	–
	CD8+HLA‐DR+CD38+	–	–	–	–0·4874, 0·0099	–0·5992, 0·0012	–
	sCD14	0·7418, 0·0037	0·7407, 0·0024	0·6205, 0·0179	–0·7626, 0·0015	–0·6154, 0·0332	–
	LPS	0·7582, 0·0027	0·6308, 0·0156	0·5545, 0·0396	–0·6835, 0·0070	–0·6154, 0·0332	0·6750, 0·0058
HIV‐1 on‐ART	Treg	–	0·4008, 0·0282	0·2880, 0·1227	–0·06618, 0·7533	–0·2370, 0·2540	–
	CD4+HLA‐DR+CD38+	–	–	–	–0·4115, 0·0410	0·2192, 0·2924	–
	CD8+HLA‐DR+CD38+	–	–	–	0·1885, 0·3670	0·07846, 0·7093	–
	sCD14	0·5303, 0·0511	0·6271, 0·0164	0·6953, 0·0058	–0·6190, 0·0241	–0·5990, 0·0396	–
	LPS	0·5579, 0·0382	0·5413, 0·0456	0·5501, 0·0416	–0·6743, 0·0229	–0·5990, 0·0396	0·7137, 0·0042
HIV‐2 ART‐naive	Treg	–	0·6078, 0·0075	0·5008, 0·0343	–0·7025, 0·0024	–0·4911 0·0534	–
	CD4+HLA‐DR+CD38+	–	–	–	–0·5994, 0·0141	–0·5478, 0·0280	–
	CD8+HLA‐DR+CD38+	–	–	–	–0·5497, 0·0274	–0·5109, 0·0431	–
	sCD14	0·6506, 0·0086	0·5076, 0·0534	0·5939, 0·0196	–0·5458, 0·0435	–0·5745, 0·0316	–
	LPS	0·5321, 0·0412	0·6857, 0·0048	0·3038, 0·2709	–0·6806, 0·0074	–0·7563, 0·0017	0·6810, 0·0052
HIV‐2 on‐ART	Treg	–	0·5311, 0·0021	–0·0779, 0·6768	–0·3961, 0·0334	0·1762, 0·3606	–
	CD4+HLA‐DR+CD38+	–	–	–	0·04736, 0·8108	0·02967, 0·8809	–
	CD8+HLA‐DR+CD38+	–	–	–	–0·4535, 0·0175	–0·09732, 0·6223	–
	sCD14	0·6184, 0·0140	0·5780, 0·0304	0·6571, 0·0107	–0·5429, 0·0449	–0·6149, 0·0193	–
	LPS	0·6988, 0·0037	0·5604, 0·0371	–0·0241, 0·9346	–0·5769, 0·0390	–0·6285, 0·0161	0·5357, 0·0396
HIV‐D on‐ART	Treg	–	0·7750, 0·0007	0·7250, 0·0022	–0·2093, 0·4541	–0·4828, 0·0683	–
	CD4+HLA‐DR+CD38+	–	–	–	–0·5581, 0·0306	–0·2322, 0·4049	–
	CD8+HLA‐DR+CD38+	–	–	–	–0·4973, 0·0593	–0·3456, 0·2070	–
	sCD14	0·6273, 0·0123	0·5952, 0·0192	0·4290, 0·1106	–0·6750, 0·0058	–0·6827, 0·0050	–
	LPS	0·7357, 0·0018	0·5214, 0·0462	0·4964, 0·0598	–0·6494, 0·0088	–0·5961, 0·0190	0·6077, 0·0163

Value (r, P) for infected cohort. LPS = lipopolysaccharide; ART = anti‐retroviral therapy; HLA‐DR = human leukocyte antigen D‐related; GrzB = granzyme B; – = already analyzed in the same table; Th17 = T helper type 17. P‐values of 0·05 were considered significant and are shown in bold type.

NK T‐like cell depletion also occurs during HIV‐2 and HIV‐D infection

We analyzed the total NK T cell frequency from whole blood, identified as a double‐positive population expressing CD3 and CD16 (CD3⁺CD16⁺NK T‐like cells), which broadly corresponds to the CD3⁺CD1d⁺ NK T subset described to be an important immunoregulatory subset implicated in HIV‐1 pathogenesis 27. NK T‐like cells were delineated as CD4⁺ and CD8⁺ cells as well as functionally, in terms of cytotoxicity, through estimation of GrzB expression.

An apparent reduction in the frequency total peripheral NK T‐like cells was observed in all HIV‐infected participants compared to seronegative individuals. No significant difference was observed between the ART‐naive and ‐receiving HIV‐1‐ and HIV‐2‐infected individuals (Fig. 3a). A marked depletion in the frequency of CD4⁺NK T‐like cells and elevated frequency of CD8⁺NK T‐like cells was observed in all HIV‐infected groups (Fig. 3b,c).

Figure 3.

Distribution of natural killer (NK) T‐like cell frequency and cytotoxicity in HIV pathogenesis. Comparison of frequency of NK T‐like cells (a), CD4⁺NK T‐like cells (b) and CD8⁺NK T‐like cells (c); comparison of frequency of granzyme B (GrzB)‐producing NK T‐like cells (d), CD4⁺NK T‐like cells (e) and CD8⁺NK T‐like cells (f) in HIV‐1‐, HIV‐2‐ and HIV‐D‐infected individuals and seronegative individuals. Gating strategy is outlined in Supporting information, Fig. S3. Statistical significance was evaluated by unpaired t‐test; *P < 0·05; **P < 0.01; ***P < 0·001.

Additionally, the GrzB level was significantly increased in total NK T‐like cells among all HIV‐infected participants compared to seronegative individuals. Interestingly, both at the total as well as subset (CD4⁺ and CD8⁺) levels, NK T‐like cells showed an increased level of GrzB in the ART‐receiving HIV‐1 group compared to the ART‐receiving HIV‐2 group (Fig. 3d–f) despite a similar expression seen in ART‐naive conditions. Also, in agreement with earlier findings with respect to frequency of this subset in HIV‐1 chronic infection 38, 39, NK T‐like cell frequency and cytotoxicity showed no correlation with absolute CD4 count and level of activation (data not shown) in HIV‐2 and HIV‐D individuals.

Persistent loss of Th17/T_reg balance in HIV‐1, HIV‐2 and HIV‐D groups

A significant reduction in the frequency of Th17 cells (total IL‐17‐producing CD4⁺ T cells) and Th17/T_reg ratio were seen in all HIV‐infected groups, irrespective of ART‐status compared to seronegative individuals. A concomitant significant increase in the frequencies of total IFN‐γ‐producing CD4⁺ T (Th1) and CD8⁺ T cells was observed in HIV‐infected groups compared to seronegative subjects (Fig. 4a–c,g).

Figure 4.

Interplay of T helper type 1 (Th1)/Th17/regulatory T cells (T_reg) among HIV‐1, HIV‐2 and HIV‐D groups. Comparison of frequency of Th17 cells (a), Th1 cells (b), total interferon (IFN)‐γ‐producing CD8⁺ T cells (c), interleukin (IL)‐17⁺single‐positive (SP) CD4⁺ T cells (d), IFN‐γ⁺SP CD4⁺ T cells (e), IFN‐γ⁺IL‐17⁺CD4⁺ T cells (f) and Th17/T_reg ratio (g) in HIV‐1, HIV‐2, HIV‐D and seronegative individuals. Gating strategy is outlined in Supporting information, Fig. S4. Statistical significance was evaluated by unpaired t‐test; *P < 0·05; **P < 0·01; ***P < 0·001.

Further, we quantified the frequency of CD4⁺ T cells producing IFN‐γ only [single‐positive (SP)], IL‐17 only or both IFN‐γ and IL‐17. Interestingly, increased frequency of IFN‐γ‐producing cells (Th1) in all HIV‐infected groups was found to be contributed by IFN‐γ⁺IL‐17⁺CD4⁺ T cells and not by IFN‐γ⁺SP CD4⁺T cells (Fig. 4d–f).

Th17 cell frequency and Th17/T_reg ratio exhibited a significant positive correlation with absolute CD4 count (Table 2) and a negative correlation with level of immune activation (Table 3) in ART‐naive HIV‐1‐ and HIV‐2‐infected individuals. Further, only Th17 cells but not the Th17/T_reg ratio showed an inverse correlation with level of immune activation in ART‐treated HIV‐infected groups. A significant negative correlation was also observed between Th17 and T_reg frequencies in ART‐naive HIV‐1, ART‐naive HIV‐2 and ART‐treated HIV‐2‐infected individuals (Table 3). In summary, disruption of Th17/T_reg homeostasis was apparently not restored upon ART treatment or lower in therapy‐naive conditions within HIV‐1‐, HIV‐2‐ and HIV‐D‐infected individuals.

Increased microbial translocation associated with dysfunctional T cell immunity

LPS, a component of gram negative bacterial cell walls, is considered a major marker of microbial translocation. sCD14, which is shed by innate immune cells such as monocytes, binds to LPS, thereby serving as a marker of LPS‐induced monocyte or macrophage activation 31, 32. Thus, LPS and sCD14 levels were measured in plasma to determine the degree of microbial translocation as well as monocyte activation, respectively, in all study groups. LPS and sCD14 levels remained elevated in HIV‐1‐, HIV‐2‐ and HIV‐D‐infected individuals irrespective of ART status compared to healthy individuals (Fig. 5).

Figure 5.

Level of microbial translocation markers across the study cohort. Comparison of lipopolysaccharide (LPS) (a) and sCD14 (b) level in HIV‐1, HIV‐2, HIV‐D and seronegative subjects. Statistical significance was evaluated by unpaired t‐test; *P < 0·05; **P < 0·01; ***P < 0·001.

We found a significant positive correlation between LPS and sCD14 in all HIV‐infected groups. Also, LPS and sCD14 showed a significant negative correlation with absolute CD4 count in all HIV‐infected groups (Table 2). In addition, a significant positive correlation of LPS and sCD14 levels with T_reg frequency and level of immune activation (activated CD4 T cells) was observed in all HIV‐infected individuals. Further, we found a significant negative correlation of LPS and sCD14 with Th17 cells and Th17/T_reg ratio in all HIV‐infected groups (Table 3). Interestingly, HIV‐2 infection in therapy‐naive conditions resulted in lower immune activation as described above but similar levels of microbial translocation and monocyte activation, which is in agreement with previously reported observations that only examined LPS in the absence of activation data 40. Also, HIV‐D individuals did not seem to have significantly different pathology compared to single infections.

Magnitude and persistence HIV‐specific responses following extended ART and association with disease progression

We aimed to investigate whether, following prolonged suppression of viremia through ART, HIV‐specific responses were detectable in ART‐receiving infected individuals and whether these responses were associated with systemic dysfunctional cellular immunity. Using ex‐vivo stimulation with immunodominant peptide pools specific for gag and env proteins of HIV‐1 and HIV‐2, frequencies of virus‐specific CD4⁺ T and CD8⁺ T cells producing IFN‐γ only (IFN‐γ‐SP), IL‐2 only (IL‐2‐SP) or both IFN‐γ and IL‐2 were estimated by ICCS assay.

Overall, virus‐specific responses were mainly detected in memory CD4⁺ T cells and higher for gag than env (Fig. 6a–d). Thus, further analysis was carried out on CD4⁺ T cell responses. Interestingly, in all ART‐receiving HIV groups, gag‐specific CD4⁺T cell responses (Fig. 6a) were mainly contributed by IL‐2⁺IFN‐γ⁺ and IFN‐γ‐SP cells in contrast to env‐specific CD4⁺ T cell responses (Fig. 6b), which were mainly IFN‐γ‐SP cells. Further, total gag and env‐specific CD4⁺ T cell responses as well as CD4⁺ T cells producing both IL‐2⁺IFN‐γ⁺ were significantly higher in HIV‐2‐infected individuals. HIV‐D‐infected individuals retained similar total HIV‐specific CD4⁺ T cell memory responses when stimulated with either gag or env immunodominant peptide sets of HIV‐1 and HIV‐2 separately (Fig. 6e–h).

Figure 6.

HIV‐specific immune responses in anti‐retroviral therapy (ART)‐receiving HIV‐1, HIV‐2 and HIV‐D subjects. Graphical presentation of gag‐specific CD4⁺ T cell responses (a), env‐specific CD4⁺ T cell responses (b), gag‐specific CD8⁺ T cell responses (c) and env‐specific CD8⁺ T cell responses (d) in ART‐receiving HIV‐1, HIV‐2 and HIV‐D subjects. Comparison of total HIV‐specific CD4⁺ T cells responses against gag (e) and env (f) as well as HIV‐specific CD4⁺ T cells producing interleukin (IL‐2)/interferon (IFN)‐γ against gag (g) and env (h) in ART‐receiving HIV‐1, HIV‐2 and HIV‐D subjects. Gating strategy is outlined in Supporting information, Fig. S5. Statistical significance was evaluated by unpaired t‐test; *P < 0·05; **P < 0·01; *** P < 0·001.

Spearman’s regression analyses, utilizing CD4 counts as a clinical disease progression marker, showed inverse correlation of gag‐specific responses, particularly CD4⁺ T cells producing IFN‐γ together with both IL‐2 and IFN‐γ, in all ART‐receiving HIV groups. However, env‐specific CD4⁺ T cell responses (total IL‐2 and IL‐2‐SP for HIV‐2; IL‐2⁺IFN‐γ⁺ for HIV‐D) were negatively correlated with CD4 count in ART‐receiving HIV‐2 and HIV‐D‐infected individuals (Table 4).

Table 4.

Association between absolute CD4 count and HIV‐specific CD4⁺ T cell responses in ART‐receiving HIV groups

	gag					env
	Total IFN‐γ+	Total IL‐2+	IL2+ IFN‐γ+	IL‐2+ SP	IFN‐γ+ SP	Total IFN‐γ +	Total IL‐2+	IL2+ IFN‐γ+	IL‐2+ SP	IFN‐γ+ SP
HIV‐1 on‐ART	–0·7939, 0·0061	–0·5030, 0·1383	–0·8909, 0·0005	0·1885, 0·6021	0·5515, 0·0984	–0·2848, 0·4250	–0·2000, 0·6059	–0·3833, 0·3085	–0·3939, 0·2600	–0·5152, 0·1276
HIV‐2 on‐ART	–0·7818, 0·0045	–0·8091, 0·0026	–0·8091, 0·0026	–0·4727, 0·1420	–0·6364, 0·0353	–0·4182, 0·2006	–0·6545, 0·0289	–0·5636, 0·0710	–0·6545, 0·0289	–0·3545, 0·2847
HIV‐D on‐ART	1; –0·5152, 0·1276	1; –0·0909, 0·8028	1; –0·5152, 0·1276	1; 0·1152, 0·7514	1; –0·0787, 0·8287	1; –0·4681, 0·1725	1; –0·2848, 0·4250	1; –0·7212, 0·0186	1; 0·07879, 0·8287	1; –0·2000, 0·5796
	2; –0·6485, 0·0425	2; –0·5758, 0·0816	2; –0·745, 0·0133	2; –0·2432, 0·4984	2; –0·05455, 0·8810	2; 0·04863, 0·8939	2; –0·1636, 0·6515	2; –0·5833, 0·0992	2; –0·0668, 0·8544	2; –0·3818, 0·2763

1 and 2, HIV‐specific response against gag or env immunodominant peptide sets of HIV‐1 and HIV‐2, respectively. Value (r, P) for infected cohort. P‐values of 0·05 were considered significant and are shown in bold type. IFN = interferon; IL = interleukin; SP = single‐positive; ART = anti‐retroviral therapy.

Next, HIV‐specific CD4⁺T cell responses were associated with level of immune activation, a functional immunological marker. Most associations observed were for gag‐specific responses (total IFN‐γ, total IL‐2 and IL‐2⁺IFN‐γ⁺), which included positive correlation with level of immune activation (activated CD4⁺ T cells) in all ART‐receiving HIV groups. However, env‐specific CD4⁺ T cell responses (total IFN‐γ and total IL‐2) only in ART‐receiving HIV‐2‐infected individuals exhibited a positive correlation with level of immune activation (Table 5). Thus, gag‐specific CD4⁺ T cell responses were observed to be reliable indicators of disease progression and immune dysfunction in the absence of viremia. Also, HIV‐2‐infected individuals showed relatively higher antigen‐specific responses that correlated more frequently with systemic markers of disease progression.

Table 5.

Association between HIV‐specific CD4⁺T cell responses and level of immune activation in ART‐receiving HIV groups

		gag					env
		Total IFN‐γ+	Total IL‐2+	IL2+ IFN‐γ+	IL‐2+ SP	IFN‐ γ+ SP	Total IFN‐γ+	Total IL‐2+	IL2+ IFN‐γ+	IL‐2+ SP	IFN‐γ+ SP
HIV‐1 on‐ART	CD4+ HLA‐DR+CD38+	0·757, 0·0111	0·660, 0·0376	0·806, 0·0049	0·006079, 0·9867	–0·2727, 0·4458	0·3333, 0·3466	0·3500, 0·3558	0·500, 0·1705	0·224, 0·5334	0·381, 0·2763
	CD8+ HLA‐DR+CD38+	0·6242, 0·0537	0·5030, 0·1383	0·697, 0·0251	0·006079, 0·9867	–0·2727, 0·4458	0·224, 0·5334	0·300, 0·4328	0·600, 0·0876	–0·05455, 0·8810	0·200, 0·5796
HIV‐2 on‐ART	CD4+ HLA‐DR+CD38+	0·6848, 0·0289	0·6970, 0·0251	0·6970, 0·0251	0·1879, 0·6032	0·357, 0·3104	0·7818, 0·0075	0·7091, 0·0217	0·454, 0·1869	0·248, 0·4888	0·272, 0·4458
	CD8+ HLA‐DR+CD38+	–0·2848, 0·4250	–0·3818, 0·2763	–0·3818, 0·2763	0·006061, 0·9867	–0·2848, 0·4250	0·6364, 0·0479	0·2242,0·5334	0·139, 0·7009	0·248, 0·4888	0·272, 0·4458
HIV‐D on‐ART	CD4+ HLA‐DR+CD38+	1; 0·6485,0·0425	1; 0·442, 0·2004	1; 0·709, 0·0217	1; –0·151, 0·6761	1; 0·04242,0·9074	1; 0·492, 0·1482	1; 0·430, 0·2145	1; 0·297, 0·4047	1; 0·03030, 0·9338	1; 0·539, 0·1076
		2; 0·7455, 0·0133	2; 0·757, 0·0111	2; 0·854, 0·0016	2; 0·0668, 0·8544	2; –0·103, 0·7770	2; 0·243, 0·4984	2; 0·406, 0·2443	2; 0·533, 0·1392	2; 0·0911, 0·8022	2; 0·527, 0·1173
	CD8+ HLA‐DR+CD38+	1; 0·127, 0·7261	1; 0·260, 0·4671	1; 0·224, 0·5334	1;0·0060, 0·9867	1; 0·575, 0·0816	1; –0·109, 0·7635	1; 0·333, 0·3466	1; 0·260, 0·4671	1; 0·139, 0·7009	1; 0·333, 0·3466
		2; 0·224, 0·5334	2; 0·563, 0·0897	2; 0·357, 0·3104	2; 0·0060, 0·9867	2; 0·175, 0·6272	2; –0·145, 0·6876	2; 0·163, 0·6515	2; 0·316, 0·4064	2; 0·0729, 0·8413	2; 0·309, 0·3848

1 and 2, HIV‐specific response against gag or env immunodominant peptide sets of HIV‐1 and HIV‐2 respectively. Value (r, P) for infected cohort. IFN = interferon; IL = interleukin; SP = single‐positive; ART = anti‐retroviral therapy; HLA‐DR = human leukocyte antigen, D‐related.

Discussion

This investigation is one of the few to extensively examine clinical and immunological profiles of individuals who are infected with single HIV‐1, single HIV‐2 and dually infected with HIV‐1 and HIV‐2 (HIV‐D) from India, another major site for HIV‐2 and HIV‐D infection that remains largely under‐reported. Another novel aspect of this study is that we have examined the impact of ART on different T cell functions and HIV‐specific responses concurrently in all three virological groups (HIV‐1, HIV‐2 and HIV‐D).

As reported 12, 41, 42 for single HIV‐1 and HIV‐2 infection we found that immune activation remained high, at similar levels to single infection, and was associated with disease progression in ART‐receiving HIV‐D individuals. This implies similar mechanistic impairments of gut CD4 repopulation, due to residual viral replication and/or chronic immune activation that are discordant with peripheral rebound during the pathogenesis of dual infection 43, 44. Also, HIV‐2 and HIV‐D individuals may be expected to suffer similar consequences of this activation in terms of immune restoration, as have been reported for HIV‐1 45, 46. CD4⁺ T and not CD8⁺ T cell activation showed a significant positive association with circulating T_reg frequency in all study cohorts irrespective of viremia or ART status, highlighting the importance of monitoring the latter as a potential peripheral surrogate of gut immune restoration as well indicative of expanded suppressive responses to curb systemic activation.

Cytotoxic T cells confer host cell immunity against viral pathogens through release of perforin and granzyme (particularly GrzB) from cytolytic granules 19, 20, 21. While HIV‐specific defects in CTL activity have been reported to occur in ART receiving HIV‐1‐infected individuals 47, we wanted to evaluate if, as a consequence of other systemic T cell dysfunction such as elevated activation and T_reg frequency, increased cytotoxic potential was observed and influenced by virological suppression in all three infections. We observed elevated levels of GrzB in CD4⁺ and CD8⁺ T cell compartments in HIV‐1, HIV‐2 and HIV‐D groups compared to the seronegative group, irrespective of viremia.

The current study showed apparent alteration in homeostatic CD4⁺ T cell subsets, based on the expression of IL‐2 (CD25) and IL‐7α (CD127) receptors, and this dysregulation was found to be associated with disease progression irrespective of ART status in HIV‐1 and HIV‐2 groups. A similar alteration in these subsets was also observed in the HIV‐D group, highlighting a pan‐HIV‐induced alteration in cytokine networks with respect to IL‐2 and IL‐7 homeostasis. Both IL‐2 and IL‐7 are being investigated as immunomodulatory therapies in HIV‐infected individuals 22, 23, 24, 25.

CD3⁺CD16⁺NK T‐like cells are a group of T cells that have important immunoregulatory functions against cancer, autoimmune diseases and immunodeficiency disease, including HIV infection 27. We investigated whether circulating NK T‐like cell frequencies were perturbed and associated with HIV disease progression in therapy‐receiving and naive conditions. Contrary to earlier reports 48 that described a depletion in both CD4⁺ and CD8⁺ subsets of these cells following infection, we observed that reduction in frequency was contributed by significant depletion only in the CD4⁺NK T‐like cell subset, known to be susceptible to infection and to express high levels of CCR5 27. Our study also revealed persistent perturbation of NK T‐like cell subsets in HIV‐infected individuals receiving ART, similar to that reported in therapy‐naive conditions 49, that could contribute to opportunistic infections if disease progression worsens due, for instance, to acquisition of drug resistance mutations. Encouragingly, however, IL‐2 combined with ART has been reported to significantly expand NK T cells in HIV‐1‐infected individuals 50, and similar adjunct therapy may be considered for HIV‐2 and HIV‐D infections. It is important to note that, although we found differential alteration in NK T‐like cell subsets following HIV infection, the cytotoxicity associated with both CD4⁺ and CD8⁺ subsets was found be elevated in all HIV‐infected individuals irrespective of virological suppression, suggesting systemic and persistent dysregulation of these cells following infection.

With respect to the interplay of Th1/Th17/T_reg phenotypes, our data are in agreement with previous findings where increased production of IFN‐γ, a known inhibitor of Th17 cell differentiation 51 in all infected groups irrespective of therapy status, indicates the presence of a common pathogenic signature contributing to increased chronic immune activation, dysfunction and exhaustion. The inability to effectively repopulate Th17 cells suggests that strong, probably early and persistent pathogenic mechanisms affecting either the induction or differentiation of Th17 cells may be involved in all types of HIV infection. These data, together with associated elevated levels of activation and microbial translocation markers (LPS and sCD14), correlated as expected with disease progression, increased level of immune activation and T_reg frequency. Interestingly, the use of probiotics to reverse or blunt persistent HIV‐induced damage to gut immunity and integrity has been suggested as a possible supplementary modality for restoration of Th17 cells 52, but further consolidative studies have been lacking since then.

The observed functional immune impairment reported in this study and by others could be due to the delay and incomplete restoration CD4 count at gut‐associated lymphoid tissue (GALT), a primary and early site of HIV replication, even after prolonged ART 43, 44. Indeed, it can be argued that the time of ART initiation would also matter in the observed functional impairment because in our study cohort, individuals had started ART relatively late according to NACO 2013 guidelines (CD4 count < 350 cells/mm3) 8. These have now been revised to the Joint United Nations Programme on HIV and AIDS (UNAIDS) recommended ‘test and treat’ algorithm 53. Also, studies have shown that a rebound of some immune function is possible in individuals where ART has initiated during early infection 27, 41, 54.

Continuing low‐level viral replication from reservoirs has been demonstrated in the majority of ART‐receiving individuals 55. Furthermore, memory cells have been shown to persist for prolonged periods in the absence of specific antigens 56. Also, therapeutic vaccination while receiving ART has been shown to be a valuable additive treatment modality 57, 58. Our aim was thus to evaluate the presence, quality and quantity of virus‐specific responses retained after extended ART in all three types of HIV infections. While earlier studies have demonstrated a drastic decline for markers such as IFN‐γ to undetectable levels of recall HIV‐1‐specific cellular immune responses during successful and extended ART 13, we show here that CD4⁺, more than CD8⁺T cell responses, possibly including long‐term memory populations, persist in treated individuals, with gag‐specific CD4⁺ T cell responses dominating. Also, specific responses to both HIV‐1 and HIV‐2 were detected in ART‐receiving HIV‐D individuals, suggesting that T cell memory for both viruses is retained and could potentially play a role in the immune control of HIV‐D infection. Also, more total HIV‐specific responses as well as dual‐functional IL‐2/IFN‐γ HIV‐specific CD4⁺ T cell responses were detected in HIV‐2‐infected individuals indicative of their maintaining a greater and more functionally diverse repertoire of virus‐specific memory CD4⁺ T cell responses. Further, HIV‐specific responses along with immune activation were associated negatively with absolute CD4 count, suggesting that rebound of non‐virus‐specific CD4⁺ T cells in individuals with potentially smaller reservoirs of ongoing viral replication was occurring, resulting in dilution of circulating antigen‐specific T cells.

Overall, our study delineates immune signatures that are specific to therapy conditions (naive versus ART‐receiving), as illustrated by the Th17/T_reg ratio for all types of naive HIV infection. A unique signature associated with HIV‐2 disease progression and immune restoration and/or preservation was the frequency of CD4⁺ naive/central memory and effector memory subsets. Interestingly, this was observed in individuals who had a range of CD4 increases from as low as 1% to more than 300% (Supporting information, Table S1). Correlates of disease progression and immune restoration among all types of HIV infection include circulating T_reg frequency, T cell activation and microbial translocation levels. Also, we show that HIV‐specific memory responses can be followed in ART‐receiving HIV‐infected (including dually infected) individuals as correlates of CD4 rebound or a potential marker of viral reservoir (Table 6).

Table 6.

Correlates of disease progression and immune restoration

	Disease progression		Immune restoration
	ART‐naive HIV‐1	ART‐naive HIV‐2	ART‐receiving HIV‐1	ART‐receiving HIV‐2	ART‐receiving HIV‐D
Treg	+	+	+	+	+
CD4+ naive/central memory subset	+	+	–	+	+
CD4+ effector memory subset	+	+	–	+	+
T cell activation	+	+	+	+	+
Th17/Treg ratio	+	+	–	–	–
Microbial translocation	+	+	+	+	+
HIV‐specific responses	n.a.	n.a.	+	+	+

The table depicts the presence (+) or absence (–) of correlations between immune signatures and disease progression (therapy‐naive) or immune restoration (under ART) in various types of infection. n.a. = data not available.

One clear limitation of this study is the lack of ART‐naive HIV‐D data, as it would have been better to compare the ART‐treated group to ART‐naive group of HIV‐D subjects. Another limitation is the lack of the viral load data of HIV‐2 and HIV‐D which would have enabled us to know the viral load status of these individuals. Also, there is a need to consider the entire breadth of the cellular immune response to HIV, rather than limiting studies to immunodominant peptide pools.

Conclusion

To the best of our knowledge, this is the first report to focus on concurrent evaluation of homeostatic and pathogen‐specific T cell functions and the impact of ART in HIV‐1, HIV‐2 and HIV‐D subjects from India – a major site of the global pandemic. Our data do not support a purported disparate pathogenesis for HIV‐D infection, and revealed that HIV‐D infected individuals undergoing ART showed a similar degree of immune dysfunction compared to that observed in either single HIV infection. Although ART has led to successful viral load suppression and demonstrable CD4 count rebound, our data revealed the presence of persistent functional impairment of T cell functions even after prolonged therapy, highlighting the need for novel immunomodulatory therapeutic strategies for HIV‐1, HIV‐2 and HIV‐D infection. The demonstration of persistent HIV‐specific responses after prolonged viral suppression indicates continuing reservoir‐driven viral replication in these individuals, which may undermine the long‐term efficacy of anti‐viral therapy alone as a sustainable therapeutic modality against HIV eradication and raises the possibility of eventual acquisition of drug resistance. Our data also support the rationale for using therapeutic vaccine‐based strategies for improved immune control of viral replication under ART.

Disclosures

All authors declare that they have no conflicts of interest.

Author contributions

V. P.1 and S. S. designed the study and analyzed the data. S. S. and V. P.2 recruited the participants and performed the experiments. V. N. and P. P. assisted with participant recruitment and obtaining clinical history. V. P.1 and S. S. wrote the manuscript. All authors read and approved the final manuscript. Vainav Patel and Varsha Padwal are designated as V.P.1 and V.P.2 respectively.

Supporting information

Fig. S1. CD4⁺T cell rebound in ART‐receiving HIV‐1, HIV‐2 and HIV‐D infected group. Comparison of absolute CD4 counts at ART‐start time (represented as “0”) and CD4 count at the sampling time (represented as “1”) (a, b & c); Comparison of fold change fold change in the CD4 count (CD4 count at the time of sampling/CD4 count at the ART start) (d); Comparision of duration on therapy (e) in ART‐treated HIV‐1, HIV‐2 and HIV‐D infected individuals. Statistical significance was evaluated by paired t test (for a, b & c) and unpaired t test (for d); *P < 0·05; **P < 0·01; and ***P < 0·001

Fig. S2. Gating strategy for defining subsets of CD4⁺ T cells using CD127 and CD25: Cells were gated based on characteristic light scatter properties FSC against SSC, followed by gating on CD4⁺ T cells. Thereafter based on expression of CD127 and CD25, CD4⁺T cells (a) were further demarcated as naive/memory (CD127⁺CD25^low/‐), effector (CD127‐CD25‐) and T_regs (CD25^highCD127^low). Gating strategy for activation marker: Cells were gated based on characteristic light scatter properties FSC against SSC, followed by gating on CD4⁺ T cells and CD8⁺T cells. Thereafter based on expression of HLADR and CD38, CD4⁺T (b) and CD8⁺T (c) cells were further demarcated as HLADR⁺CD38⁺ population. The HLADR⁺CD38⁺ population was reported as the activated population. The FMO control was used for gating positive population of CD38. Gating strategy for granzyme‐B: The lymphocyte population was gated, followed by gating on CD4⁺ T cells and CD8⁺T cells. Thereafter based on expression of GrzB, CD4⁺T (d) and CD8⁺T (e) cells were further analyzed for granzyme‐B positivity (compared to FMO control) and this population was reported as cytotoxic T cells.

Fig. S3. Gating strategy for NKT‐like cells, its subsets and cytotoxicity of these subsets. Cells were gated based on characteristic light scatter properties FSC against SSC, followed by gating on CD3⁺ T cells. Based on expression of CD16, CD3⁺T cells (a) were further analyzed for CD16 positivity and these population is known as NKT‐like cells.Thereafter based on expression of CD4, NKT‐like cells were demarcated as CD4⁺ and CD4^‐ (referred as CD8⁺)further demarcated as CD4⁺NKT and CD8⁺NKT population (b). Further, based on expression of GrzB, CD4⁺NKT, CD8⁺NKT and NKT‐like cells were further analyzed for granzyme‐B positivity (c). The FMO control was used for gating positive population of CD16 and GrzB.

Fig. S4. Gating strategy for Th1 and Th17 cells. Cells were gated based on characteristic light scatter properties FSC against SSC, followed by gating on CD4⁺ T cells and CD8⁺T cells. Thereafter based on the expression of IL‐17 and IFN‐γ, CD4⁺T and CD8⁺T cells were further demarcated as IFN‐γ‐SP, IL‐17‐SP and both IFN‐γ/IL‐17 (compared to unstimulated) population

Fig. S5. Gating strategy for HIV‐specific responses. Cells were gated based on characteristic light scatter properties FSC against SSC, followed by gating on CD4⁺ T cells and CD8⁺T cells. Thereafter based on the expression of IL‐17 and IFN‐γ, CD4⁺T and CD8⁺T cells were further demarcated as IFN‐γ‐SP, IL‐2‐SP and both IFN‐γ/IL‐17 (compared to unstimulated) population

Table S1. Descriptive statistics for CD4 rebound (fold change) across HIV‐1, HIV‐2 and HIV‐D groups.