Highly dampened HIV-specific cytolytic effector T cell responses define viremic non-progression

Abstract

This study reports on HIV-specific T cell responses in HIV-1 infected Viremic Non-Progressors (VNPs), a rare group of people living with HIV that exhibit asymptomatic infection over several years accompanied by stable CD4+ T cell counts in spite of ongoing viral replication. We attempted to identify key virus-specific functional attributes that could underlie the apparently paradoxical virus-host equilibrium observed in VNPs. Our results revealed modulation of HIV-specific CD4+ and CD8+ effector T cell responses in VNPs towards a dominant non-cytolytic profile with concomitantly diminished degranulation (CD107a+) ability. Further, the HIV specific CD8+ effector T cell response was primarily enriched for MIP-1β producing cells. As expected, concordant with better viral suppression, VCs exhibit a robust cytolytic T cell response. Interestingly, PuPs shared features common to both these responses but did not exhibit a CD4+ central memory IFN-γ producing Gag-specific response that was shared by both non-progressor (VC and VNP) groups, suggesting CD4 helper response is critical for non-progression. Our study also revealed that cytolytic response in VNPs is primarily limited to polyfunctional cells while both monofunctional and polyfunctional cells significantly contribute to cytolytic responses in VCs. To further understand mechanisms underlying the unique HIV-specific effector T cell response described here in VNPs we also evaluated and demonstrated a possible role for altered gut homing in these individuals. Our findings inform immunotherapeutic interventions to achieve functional cures in the context of ART resistance and serious non AIDS events.

1. Introduction

In the absence antiretroviral therapy (ART), human immunodeficiency virus (HIV) infection typically causes rapid immune impairment through progressive CD4+ T cell depletion and uncontrolled viral replication, eventually leading to acquired immunodeficiency syndrome (AIDS) (Boasso et al., 2009; Okoye and Picker, 2013). ART-naïve HIV disease progression can be variable, categorised into rapid progression, typical progression and long term non-progression (Pantaleo and Fauci, 1996). In light of the continuing setbacks in prophylactic strategies for prevention of HIV acquisition (Corey et al., 2021; “HIV Vaccine Candidate Does Not Sufficiently Protect Women Against HIV Infection | National Institutes of Health (NIH),” n.d.; Rerks-Ngarm et al., 2009) as well as growing ART resistance (Hamers et al., 2018), studying mechanisms of non-progression remain important for elucidation of putative immune based interventions to achieve ‘functional cure’ and manage serious non AIDS events (Davenport et al., 2019; Prabhu et al., 2019). Long term non-progressors (LTNPs) like Elite Controllers (EC; <50 RNA copies/ml) and Viremic Controllers (VC; ≤ 2000 RNA copies/ml) efficiently control viremia while maintaining high CD4+ T cell count (Gaardbo et al., 2012; Okulicz et al., 2009). On the other hand, Viremic Non-Progressors (VNP) are a distinct group of individuals who continuously maintain near-normal but stable CD4+ T cell count for several years (≥7 years) despite persistently high viral replication (Choudhary et al., 2007; Gaardbo et al., 2013; Klatt et al., 2014; Singh et al., 2020). Also, HIV-1 infection in VNPs resembles non-pathogenic simian immunodeficiency virus (SIV) infection of natural hosts such as Sooty mangabey (SM) and African green monkey (AGM) (Chahroudi et al., 2012; Pandrea et al., 2008b). We recently reported that VNPs, despite ongoing viral replication, have intact CD4+ T cell homeostasis supported by sustained IL-7 mediated thymic repopulation of CD4+ T cells as well as preserved CD4+ central memory compartment (Singh et al., 2020). Preservation of this compartment in VNPs also indicates operational ‘CD4 help’, critical for development and maintenance of anti-viral effector CD8+ T cell responses and which is progressively depleted in typical progression (Kalams and Walker, 1998; Letvin et al., 2006; Nakanishi et al., 2009; Okoye et al., 2007; Shedlock and Shen, 2003; Sun and Bevan, 2003).

Notwithstanding intact homeostatic mechanisms to resist CD4+ T cell depletion, a cellular immune response capable of effectively suppressing viral replication seems unlikely in VNPs. However, experimental in vivo depletion of either CD4+ or CD8+ T cells in SIV-infection has been shown to abrogate decline of post peak and chronic viremia, thereby emphasizing a role for anti-viral T cell responses in resisting disease progression (Chowdhury et al., 2015; Ortiz et al., 2011). Furthermore, another in vivo CD8+ T cell depletion study in non-pathogenic SIV infection of AGM revealed importance of anti-viral CD8+ T cell responses in partial control of post-acute viral replication (Cardozo et al., 2018; Gaufin et al., 2010). Thus, a similar role for HIV-specific T cell responses within VNPs is likely. We hypothesized that a unique modulated cell-mediated antiviral response, providing limited virological control may also exist to resist disease progression in VNPs. To this end, we compared HIV-specific responses in VNPs to that of putative progressors (PuPs) - recently infected ART-naïve progressors with a history of rapid CD4+ T cell depletion, high viral replication and CD4+ T cell count of ≥ 500 cells/mm³ at the time of recruitment. In addition, the study also included a VC group where cell-mediated immunity is known to play a dominant role in control of viral replication (Freel et al., 2011; Jones et al., 2021).

2. Materials and methods

2.1. Study participants

HIV-1 infected participants for this study were recruited from ART Centre at Grant Medical College & Sir J. J. Group of Hospitals, Mumbai. Study participants were selected from a previously described cohort based on availability of stored samples (Singh et al., 2020). Three groups of HIV-1 infected, ART naïve participants were recruited based on the following criteria: (i)Viremic Non-Progressors (VNP; n = 12) who had ≥ 7 years of asymptomatic HIV-1 infection with CD4+ T cell counts of ≥ 500 cells/mm³ and viral load (VL) of greater than 10,000 RNA copies/ ml; (ii)Viremic controllers (VC; n = 7) who had ≥ 7 years of asymptomatic HIV-1 infection with CD4+ T cell counts of ≥ 500 cells/mm³ and viral load (VL) of ≤ 2,000 RNA copies/ml; and (iii) Putative progressors (PuP; n = 11) who had recent (6 months – 3 years) HIV-1 infection but were recruited with CD4+ T cell counts of ≥ 500 cells/mm³, before dramatic CD4+ T cell depletion in absence of ART, to avoid the effects of immunological impairment. PuPs also had ongoing viral replication with detectable viral load (VL) of greater than 2,000 RNA copies/ml. Absolute CD4+ T cell count and viral load were measured as previously described (Singh et al., 2020). Clinical characteristics of study participants are summarised in Table 1. Signed informed consent was obtained from study participants as per recommendations of ICMR-NIRRH Institutional Ethics Committee for clinical research. ICMR-NIRRH Institutional Ethics Committee Review Board approved the study protocols (Project no. 225/2012).

Table 1. Clinical characteristics of the study participants.

	Viremic Controllers [VCs, n = 07]	Viremic Non-Progressors [VNPs, n = 12]	Putative Progressors [PuPs, n = 11]
Agea (Years), Range	42 (29–60)	39 (30–49)	35 (21–59)
Gender	Female = 03 Male = 04	Female = 04 Male = 08	Female = 06 Male = 05
CD4+ T cell count a (cells/mm3), Range	900 (501–1469)	680 (501–910)	553 (514–908)
Viral Load a,b (log10 copies/ml)	2.95 (1.73–3.04)	4.73 (4.01–5.35)	4.71 (3.58–5.98)
Duration of infectiona (Years), Range	10 (8–24)	10 (7–16)	1 (0.5 – 03)
Antiretroviral therapy (ART) status	Naïve	Naïve	Naïve
HLA-B*27/B*57 status c	HLA-B*27 (+ve = 0/06) & HLA-B*57 (+ve = 1/07)	HLA-B*27 (+ve = 1/09) & HLA-B*57 (+ve = 0/11)	HLA-B*27 (+ve = 02/10) & HLA-B*57 (+ve = 1/11)d

^aData are expressed as the median (range).

^bViral load was estimated at the time of sampling.

^cHLA-B*27 and HLA-B*57 allele status of 5 and 1 of the HIV-1 infected participants respectively, was not available due to lack of SSP-PCR amplification.

^dOnly one participant was positive for both HLA-B*27 and HLA-B*57 allele.

2.2. Antibodies

The following fluorescently labelled monoclonal antibodies (mAb) were used to define distinct maturation phenotypes and analyse 5 different T cell functions: anti-CD3-APC-Cy7 (Clone: SK7), anti-CD4-BV480 (Clone: RPA-T4), anti-CD8-BV605 (Clone: SK1), anti-CD45RA-PerCPCy5.5 (Clone: HI100), anti-CCR7-PECF594 (Clone: 150503), anti-IFN-γ-AF488 (Clone: B27), anti-IL-2-APC (Clone: MQ1-17H12), anti-MIP-1β-PE (Clone: D21-1351), anti-TNF-α-PECy7(Clone: MAb11) and anti-CD107a-BV786 (Clone: H4A3).

2.3. Ex-vivo PBMC stimulation and Intracellular cytokine staining

Thawed cryopreserved PBMCs were resuspended in complete medium (RPMI-1640 medium supplemented with 10% FBS and 1% peni-cillin/streptomycin) and rested overnight at 37 °C in a humidified atmosphere at 5% CO₂. Rested PBMCs were stimulated in presence of a pool of 15 mer peptides overlapping by 11 amino acids spanning HIV-1 (subtype C) consensus Gag (121 peptides) and Env (212 peptides) region. The peptides sets were obtained through NIH HIV Reagent Program, Division of AIDS, NIAID, NIH and used at final concentration of 2 μg/ml/peptide. An unstimulated negative control and a positive control stimulated with 50 ng/ml PMA and 1 ng/ml ionomycin was also set up for each sample. All samples were incubated at 37 ^°C in a humidified atmosphere at 5% CO₂ in presence of co-stimulatory antibodies (CD28/ CD49d, 1 μg/ml; BD Biosciences) and anti-CD107a mAb conjugated with BV786 to detect degranulation (Betts et al., 2003). Protein transport Inhibitors, monensin (Golgi stop, 0.7 μl/ml) and brefeldin A (Golgi plug, 1 μl/ml) were added after 1 h of incubation, mixed briefly and incubated for an additional 11 h at 37 °C in a humidified atmosphere at 5% CO₂.

Following stimulation, PBMCs were washed with PBS and labelled using LIVE/DEAD™ Fixable Violet Dead Cell Stain Kit (Invitrogen, Germany) for 30 min in dark to determine viability. After labelling with amine-reactive violet viability dye (ViViD), cells were washed with stain buffer (PBS containing 2% FBS) and stained with mAb specific for surface antigen for 20 min at 37 °C. Cells were then washed and fixed/ permeabilized using Cytofix/Cytoperm kit as per manufacturer’s instructions (BD Biosciences). Following permeabilization, cells were washed with BD perm/wash buffer and stained with mAb specific for cytokines and chemokines for 30 min at room temperature. Next, cells were washed with stain buffer and acquired immediately on flow cytometry.

2.4. Flow cytometry analysis

Samples were acquired on BD FACS Aria Fusion flow cytometer. Compensation was set using anti-mouse Ig, κ/negative control compensation beads (BD Biosciences). Gates were drawn with the help of fluorescence minus one (FMO) control. Approximately, 100,000 to 400,000 viable T cells were acquired for each samples and data was analysed using Flowjo software, version 10.6.2 (BD Biosciences). Gating strategy is shown in Fig. 1. To begin with, forward scatter area (FSC-A) versus height (FSC-H) scatter plot was used to exclude doublets. Next, Viable T cells were gated based on the expression of CD3 and exclusion of ViViD dye. Subsequently, based on expression of CD4 and CD8, CD4+ and CD8+ T cells, respectively, were identified. To analyse robust antigen-specific functional attributes, most abundant CD4+ and CD8+ T cell subsets were identified based on expression of CD45RA and CCR7. In CD4+ T cell compartment, CD4+ central memory (CD4+ CM; CD45RA-CCR7+), CD4+ effector memory (CD4+ EM; CD45RA-CCR7-) and CD8+ T cell compartment, CD8+ effector memory (CD8+ EM; CD45RA-CCR7-), CD8+ terminally differentiated effectors (CD8+ TD; CD45RA+ CCR7-), respectively, were identified for functional analysis (Fig. 1A). Subsequently, gates were created for each of the 5 functions (CD107a, IFN-γ, IL-2, TNF-α and MIP-1β) for all the identified CD4+ and CD8+ memory T cell subsets (Fig. 1B). Next, Boolean combination gating was performed using Flowjo software to create all possible combinations (2ⁿ) of the 5 functions, generating frequency data for each of 32 different response patterns (Fig. 1C). Non-specific background (detected in unstimulated negative control) was subtracted from each of the 32 response patterns obtained after antigenic stimulation, individually. A threshold of 0.02% response after background-subtraction with a minimum of 10 events was considered for further analysis (Cossarizza et al., 2012) and background-subtracted response values below the threshold were set to zero. Polyfunctionality index (PI) was calculated by employing Funky cell software (Boyd et al., 2015; Larsen et al., 2012).

Fig. 1. Evaluation of HIV-specific T cell functionality.

(A) Representative gating strategy for identification of viable CD4+ and CD8+ T cell subsets (in red boxes) for functionality evaluation. (B) Representative plots displaying background response in absence of HIV-antigen (top row) and functional profile in presence of HIV-antigen (Bottom row) individually for each function. (C) Data representing 31 different response patterns (except all negative) obtained through Boolean combination gating of individual function gates. Stacked bars (color coded for VC, VNP and PuP groups) represent mean CD8+ EM Gag-specific response for distinct functional profiles (x-axis). Refer “Material and Methods” for detailed methodology. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

2.5. Immunophenotyping

Immunophenotyping of thawed PBMCs were performed using following fluorescently labelled monoclonal antibodies: anti-CD3-APC-Cy7 (Clone: SK7), anti-CD4-BV480 (Clone: RPA-T4), anti-CD45RA-PerCPCy5.5 (Clone: HI100), anti-CCR7-PECF594 (Clone: 150503), anti-CD28-PECY7 (Clone: CD28.2), anti-CCR5-APC (anti-CCR5-APC) (Clone: 2D7) and anti-integrin-β7 (Clone: FIB504). PBMCs were washed with PBS and labelled using violet viability dye (ViViD) for 30 min in dark to determine viability. After labelling, cells were washed with stain buffer. Next, PBMCs were incubated with fluorescently labelled monoclonal antibodies for 20 min at room temperature. Cells were washed and resuspended in staining buffer. Approximately, 50,000 to 400,000 viable T cells were acquired for each samples and data was analysed using Flowjo software, version 10.6.2 (BD Biosciences). Anti-mouse Ig, κ/Negative Control compensation beads (BD Biosciences) were used to set compensation parameters. Fluorescence minus one (FMO) control was used to identify and gate cells.

2.6. Statistical analysis

Statistical analysis was performed on GraphPad Prism software, version 8 (San Diego, California, USA). GraphPad Prism software was also used for calling definite outliers with the help of ROUT (Q = 0.1) method with least stringency (Motulsky and Brown, 2006). The outliers identified by ROUT method also corresponded to a minimum of 2 SDs above mean response. Comparisons among study groups were made using nonparametric one-way analysis of variance (ANOVA, Kruskal-Wallis test) followed by post hoc Dunn’s test. Mann-Whitney nonparametric test was performed for comparison between study groups. Bivariate correlations were determined by Spearman’s rank correlation test. For all statistical calculations, p < 0.05 was considered significant.

3. Results

3.1. Evaluation of Gag and Env-specific T cell functionality across disease progression

Intracellular staining and multi-parametric flow cytometry was used to evaluate HIV specific responses from 12 Viremic Non-Progressors (VNPs), 7 viremic controllers (VCs) and 11 putative progressors (PuPs) (Table 1) as shown in Fig. 1. Gag and Env-specific T cell responses were measured in functionally distinct non-naïve memory subsets (CD45RA and CCR7) of both CD4+ and CD8+ T cells (Fig. 1A) (Mahnke et al., 2013; Sallusto et al., 1999). CD4+ central memory (CD4+ CM), CD4+ effector memory (CD4+ EM), CD8+ effector memory (CD8+ EM) and CD8+ terminally differentiated effectors (CD8+ TD), the most abundant memory phenotypes, were examined for concurrent CD107a mobilization (degranulation) and cytokine (IFN-γ, IL-2, TNF-α), chemokine (MIP-1β) production (Fig. 1B) (Betts et al., 2003; Makedonas and Betts, 2006). Representative analysis of CD8+ EM Gag-specific responses using Boolean gating yielding 31 unique combinations (except all negative) of the measured 5 functions is shown in Fig. 1C.

To begin with, aggregating frequencies of all Gag or Env-specific cells positive for at least one function (Fig. 2A) revealed that the total Env-specific responses were significantly higher compared to Gag-specific responses for both CD4+ CM and CD4+ EM subsets across all study groups (p = 0.0001 and p = 0.001, respectively). Both Gag and Env-specific responses in VNPs were aligned with those of VCs across CD4+ T cell subsets in spite of significantly divergent virological suppression (Fig. 2A). Notably, CD4+ EM Gag-specific responses detected in recently infected PuPs with comparable CD4+ T cell counts and high viremia were significantly higher compared to both VNPs and VCs (p = 0.0357 and p = 0.0256, respectively) (Fig. 2A). Although, there was no difference in total Env-specific responses among study groups in both these subsets (Fig. 2A), a significant positive correlation of CD4+ EM Env-specific response was observed with absolute CD4+ T cell count (p = 0.0187) (Fig. 2B). Surprisingly, we did not observe any statistically significant differences between total Gag and Env- specific response in CD8+ EM and CD8+ TD subsets across VC, VNP and PuP groups (Fig. 2C).

Fig. 2. Total HIV-specific T cell responses.

(A) Frequency of total Gag and Env-specific response for both CD4+ CM and CD4+ EM subsets. (B) Correlation between frequency of total Env-specific CD4+ EM response and absolute CD4+ T cell count. (C) Frequency of total Gag and Env-specific response for both CD8+ EM and CD8+ TD subsets. Data is represented as Box and whiskers (Min. to Max. and show all points) plots and are color coded for VC, VNP and PuP groups. Comparisons were performed using nonparametric one –way ANOVA followed by post hoc Dunn’s test. Bivariate correlations were determined by Spearman’s rank correlation test. p-value < 0.05 (*p < 0.05; **p < 0.01; ***p < 0.001) were considered statistically significant.

3.2. Predominantly non-cytolytic HIV-specific T cell responses characterise viremic non-progression

We next attempted to delineate the individual contribution of CD4+ and CD8+ T cell recall responses to HIV antigens and their possible association with disparate disease progression. In addition to evaluating magnitude of these responses within T cell subsets (Figs. S1 and S2) we also determined relative contribution (proportion) of CD107a, IFN-γ, IL-2, TNF-α and MIP-1β response to that of total Gag and Env-specific response (Figs. 3 and 4). VCs demonstrated higher proportion of CD4+ CM Gag-specific IFN-γ response compared to both VNPs and PuPs (p = 0.0390 and p = 0.0117, respectively) (Fig. 3A). This signature was shared, albeit in a diminished capacity, by VNPs who, like VCs as reported by us previously (Singh et al., 2020), preserve their CD4+ CM pool and probably appear to better preserve the ability to secrete IFN-γ compared to recently infected PuPs (Fig. 3A and S1). In contrast, PuPs had higher proportion of CD4+ EM Gag-specific IFN-γ response compared to both VCs and VNPs (p = 0.0377 and p = 0.0885, respectively) (Fig. 3B). We observed no clear differences in overall functionality within CD4+ CM Env-specific response except for the fact that CD107a and TNF-α were the dominant responses observed in all study groups (Fig. 3C). Intriguingly and in contrast, the Env-specific response within the CD4+ EM subset revealed a distinct functional profile for VNPs (Fig. 3D). Here, a significantly lower proportion of CD4+ EM Env-specific cells with ability to degranulate (CD107a mobilization; cytolytic potential) were observed in VNPs compared to both VCs and PuPs (p = 0.0064 and p = 0.0785, respectively) (Fig. 3D). Similarly, lower proportion of Gag-specific CD4+ EM cytolytic responses was also observed in VNPs compared to VCs (p = 0.0356) (Fig. 3B). Interestingly, our data also revealed trends indicative of an increased proportion of CD4+ EM Env-specific MIP-1β and TNF-α production in VNPs (Fig. 3D). Further-more, a lower proportion of CD4+ EM Env-specific IFN-γ response was observed in VNPs compared to PuPs (p = 0.0176) (Fig. 3D). Although clearly detectable, HIV-specific production of IL-2 did not seem to exhibit any unique distribution across groups (Fig. 3A-D). Next, to evaluate the implications of the functional profiles of HIV-specific responses observed in our groups for non-progression we correlated these profiles with viral load (VL) and absolute CD4+ T cell count (Fig. 3E). Proportion of CD4+ CM Gag-specific IFN-γ response and proportion of both CD4+ EM Gag and Env-specific cells with ability to degranulate (CD107a+) correlated negatively with VL (p = 0.0285, p = 0.0493 and p = 0.0390, respectively) (Fig. 3E and S3A-C). Interestingly, although overall CD4+ EM Env-specific response positively correlated with CD4+ T cell count (Fig. 2B), neither CD107a (predominant in VC) nor MIP-1β and TNF-α (predominant in VNP) correlated with CD4+ T cell count (Fig. 3E).

Fig. 3. Modulation of HIV-specific CD4+ T cell responses.

(A, B) Proportion of Gag-specific response observed for each of the studied functions (CD107a, IFN-γ, IL-2, TNF-α and MIP-1β) to that of total Gag-specific CD4+ CM and CD4+ EM cells, respectively. (C, D) Proportion of Env-specific response observed for each of the studied functions (CD107a, IFN-γ, IL-2, TNF-α and MIP-1β) to that of total Env-specific CD4+ CM and CD4+ EM cells, respectively. Data is represented as Box and whiskers (Min. to Max. and show all points) plots and are color coded for VC, VNP and PuP groups. Comparisons were performed using nonparametric one -way ANOVA followed by post hoc Dunn’s test. (E) Proportion of Gag and Env-specific CD4+ CM and CD4+ EM responses, respectively, observed for each of the studied functions (CD107a, IFN-γ, IL-2, TNF-α and MIP-1β as per color key) were correlated with viral load and absolute CD4+ T cell counts and depicted in the correlation matrix. Bivariate correlations were determined by Spearman’s rank correlation test. p-value < 0.05 (*p < 0.05; **p < 0.01; ***p < 0.001) were considered statistically significant.

Fig. 4. Modulation of HIV-specific CD8+ T cell responses.

(A, B) Proportion of Gag-specific response observed for each of the studied functions (CD107a, IFN-γ, IL-2, TNF-α and MIP-1β) to that of total Gag-specific CD8+ EM and CD8+ TD cells, respectively. (C, D) Proportion of Env-specific response observed for each of the studied functions (CD107a, IFN-γ, IL-2, TNF-α and MIP-1β) to that of total Env-specific CD8+ EM and CD8+ TD cells, respectively. Data is represented as Box and whiskers (Min. to Max. and show all points) plots and are color coded for VC, VNP and PuP groups. Comparisons were performed using nonparametric one -way ANOVA followed by post hoc Dunn’s test. (E) Proportion of Gag and Env-specific CD8+ EM and CD8+ TD responses, respectively, observed for each of the studied functions (CD107a, IFN-γ, IL-2, TNF-α and MIP-1β as per color key) were correlated with viral load and absolute CD4+ T cell counts and depicted in the correlation matrix. Bivariate correlations were determined by Spearman’s rank correlation test. p-value < 0.05 (*p < 0.05; **p < 0.01; ***p < 0.001) were considered statistically significant.

This approach was further applied to assess the multidimensional functional profiles of HIV-specific CD8+ T cell response. As shown in Fig. 4A and 4B, we found no statistically significant difference in proportion of Gag-specific CD8+ EM and CD8+ TD cells expressing CD107a, IFN-γ, IL-2, TNF-α and MIP-1 β among VC, VNP and PuP groups. However, evaluation of Env-specific responses revealed that, VNPs tend to have decreased proportion of degranulating (CD107a+) Env-specific CD8+ EM and CD8+ TD cells compared to VCs and PuPs (p = 0.0161, p = 0.0804 and p = 0.0084, p = 0.1025, respectively) (Fig. 4C and 4D). Notably, a similar trend of reduced cytolytic potential (degranulation) was also demonstrated by Gag-specific CD8+ TD cells of VNPs compared to VCs (p = 0.0591) (Fig. 4B). Taken together, these trends align with the profile observed for HIV-specific cells in the CD4+ EM subset (Fig. 3B and 3D). Furthe-rmore, subsequent correlation analysis revealed that Gag and Env-specific CD8+ EM and CD8+ TD cells with ability to degranulate (CD107a+) associated negatively with viral load (p = 0.0166, p = 0.0793 and p = 0.0262, p = 0.0260, respectively) (Fig. 4E and S4A-D). Conversely, we observed increased proportion of Env-specific CD8+ EM and CD8+ TD MIP-1β response in VNPs compared to VCs (p = 0.0928 and p = 0.0370, respectively), which positively correlated with viremia (p = 0.0236 and p = 0.0709, respectively) (Fig. 4E and S4E-F). We also observed a significant positive correlation between absolute CD4+ T cell count and CD8+ TD Gag and Env-specific TNF-α response, respectively (p = 0.0246 and p = 0.0366, respectively) (Fig. 4E and S4G-H). Thus, overall, our results indicated modulation of HIV-specific T cell response towards a primarily non-cytolytic response in VNPs.

3.3. VNPs, VCs and PuPs exhibit dominant Gag-specific polyfunctional T cell responses

Previous studies have reported on the importance of polyfunctional HIV-specific CD4+ and CD8+ T cells in controlling viremia (Betts et al., 2006; Kannanganat et al., 2007). Here, we assessed HIV-specific polyfunctional responses in both CD4+ and CD8+ T cell subsets by comparing proportion of aggregate HIV-specific T cells displaying more than one function. In addition, we also calculated polyfunctionality Index (PI) using Funky cell software which employs a novel algorithm to quantify the degree and variation of polyfunctionality (Boyd et al., 2015; Larsen et al., 2012). We observed that polyfunctional CD4+ CM Gag and Env-specific responses were predominantly limited to 2-functions while those for CD4+ EM cells demonstrated more than 2-function responses (Fig. 5A) across all study groups. Calculation of PI revealed that Gagspecific CD4+ CM and CD4+ EM polyfunctionality was observed to be significantly higher compared to Env-specific polyfunctionality (p = 0.0322 and p = 0.0031, respectively) for all the study groups (Fig. 5A and 5B). These results, in contrast to the results observed in Fig. 2A, with respect to total HIV-specific response, demonstrated that Env-specific CD4+ CM and CD4+ EM responses are predominantly monofunctional. Polyfunctional profile of HIV-specific response in the VNP group was similar to that of VCs and PuPs except that VCs displayed a significantly higher CD4+ CM Gag-specific polyfunctionality compared to VNPs (p = 0.0079) (Fig. 5B). Furthermore, CD4+ CM Gag-specific polyfunctionality correlated negatively with VL (p = 0.0070) (Fig. 5C), suggesting a possible role for polyfunctional Gag-specific CD4+ CM responses in viral suppression. Interestingly, our data showed that recently infected PuPs seemed capable of generating effective Gag-specific CD4+ EM polyfunctionality (Fig. 5A and 5B). On evaluation of HIV-specific polyfunctionality in CD8+ EM and CD8+ TD subsets, both these subsets showed up to 4-function Gag and Env-specific responses across study groups (Fig. 5D). In addition, comparison of PI demonstrated a strong trend towards a dominant Gag-specific polyfunctional response for both CD8+ EM and CD8+ TD subsets consistently across study groups (p = 0.0799 and p = 0.0574, respectively) (Fig. 5E).

Fig. 5. HIV-specific polyfunctional T cell responses.

(A) Concentric donut pie charts show composition of Gag and Env-specific response for both CD4+ CM and CD4+ EM subsets, respectively for VC, VNP and PuP groups (Inside to outside); each pie slice represents the mean proportion of single function (blue) or multiple concurrent functions (as per color code). (B) Poly-functionality index of Gag and Env-specific CD4+ CM and CD4+ EM subsets. (C) Correlation between Polyfunctionality index of Gag-specific CD4+ CM response and viral load. (D) Concentric donut pie charts show composition of Gag and Env-specific response for both CD8+ EM and CD8+ TD subsets, respectively for VC, VNP and PuP groups (Inside to outside); each pie slice represents the mean proportion of single function (blue) or multiple concurrent functions (as per color code). (E) Polyfunctionality index of Gag and Env-specific CD8+ EM and CD8+ TD subsets. Data is represented as Box and whiskers (Min. to Max. and show all points) plots and are color coded for VC, VNP and PuP groups. Comparisons were performed using nonparametric one -way ANOVA followed by post hoc Dunn’s test. Bivariate correlations were determined by Spearman’s rank correlation test. p-value < 0.05 (*p < 0.05; **p < 0.01; ***p < 0.001) were considered statistically significant. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

3.4. Contribution of HIV-specific polyfunctional T cell responses to progression specific signatures

In depth polyfunctional analysis was undertaken for responses found in more than 25% of the study participants (responders) in each group to generate polyfunctional response profiles corresponding to all 31 combinations represented in Fig. 1C. Due to the general lack of antigen specific IL-2 production by polyfunctional cells, we did not detect responders with 5-function HIV-specific CD8+ EM and CD8+ TD responses in our study groups (Fig. 6A-D). Notably, we observed less diverse polyfunctional profiles for Gag and Env-specific CD8+ TD cells (Fig. 6C and 6D). Polyfunctional cytolytic cells (CD107a+) with ability to secrete IFN-γ, TNF-α and MIP-1β (4-function) were observed and occurred at similar proportions among VC, VNP and PuP groups for Gag and Env in both CD8+ EM and CD8+ TD subsets (Fig. 6A-D). VNPs however exhibited elevated Gag and Env-specific 3-function (CD107a+ IFN-γ+ MIP-1β+) and 2-function (CD107a+ MIP-1β+) CD8+ EM responses compared to VCs (Fig. 6A and 6B). A similar pattern was shown by Gag but not Env-specific CD8+ TD cells, where VNPs lacked 2-function response (CD107a+ MIP-1β+) (Fig. 6C and 6D). Furthermore, 2-function Gag-specific CD8+ EM cells (CD107a+ IFN-γ+) lacking MIP-1β were found to be preferentially elevated in VCs compared to both VNPs and PuPs (Fig. 6A). Similarly, a unique 2-function (CD107a+ TNF-α+) Env-specific CD8+ EM response but lacking MIP-1β was found to be significantly elevated in VCs compared to both VNPs and PuPs (Fig. 6B). We were also surprised to note that although the total cytolytic (CD107a+) response for VNPs was significantly lower compared to other groups (Fig. 4) this response when examined in the context of polyfunctionality remained largely comparable to the other groups.

Fig. 6. Polyfunctional response profiles of CD8+ T cell subsets.

(A) Polyfunctional response profile (x-axis) of Gag-specific CD8+ EM cells. (B) Polyfunctional response profile (x-axis) of Env-specific CD8+ EM cells. (C) Polyfunctional response profile (x-axis) of Gag-specific CD8+ TD cells. (D) Polyfunctional response profile (x-axis) of Env-specific CD8+ TD cells. Data is represented as Box and whiskers (Min. to Max. & show all points) plots and are color coded for VC, VNP and PuP groups. Comparisons were performed using nonparametric one –way ANOVA followed by post hoc Dunn’s test. p-value < 0.05 (*p < 0.05; **p < 0.01; ***p < 0.001) were considered statistically significant.

We observed a more diverse Env-specific CD4+ CM polyfunctional response profile compared to that of Gag-specific polyfunctional response profile (Figs. S5A and S5B). Gag-specific CD4+ EM polyfunctionality was observed to be up to 4-function response while Env-specific CD4+ EM polyfunctionality was limited to 2-function response (Figs. S5C and S5D). Overall, combinations of responses for the CD4+ T cell subset seemed to involve HIV-specific TNF-α production. Interestingly however, the unique signature of Gag specific IFN-γ production (total response) by CD4+ CM observed in Fig. 3A was not constituted by polyfunctional cells for VNP and PuP groups but rather by monofunctional production of this cytokine.

3.5. Cytolytic activity is primarily limited to polyfunctional cells in Viremic Non-Progressors

Our data so far revealed a distinct modulation of HIV-specific responses in VNPs, primarily with respect to CD107a mobilization (degranulation) and MIP-1β secretion in all effector subsets (CD4+ EM, CD8+ EM and CD8+ TD) (Figs. 3 and 4). This observation prompted us to delineate the contribution of both these functional signatures: degranulation (Cytolytic response) and MIP-1β (Non-cytolytic response) production towards polyfunctionality across study groups. First, we compared proportion of Gag and Env-specific monofunctional and polyfunctional cells with cytolytic potential (CD107a mobilisation) in CD4+ EM, CD8+ EM and CD8+ TD subsets across VC, VNP and PuP groups. As shown in the heat map (Fig. 7A), we observed that VNPs in general have substantially reduced proportion of monofunctional cells with cytolytic potential (Range:10%-81%; Median: 45%) in comparison to that of polyfunctional cells (Range:10%-95%; Median: 89%). Also, VNPs have significantly reduced proportion of cytolytic monofunctional cells (Range:12%-55%; Median: 38%) compared to that observed for VCs (Range:39%-81%; Median: 56%). However, proportion of cytolytic polyfunctional cells was observed to be comparable across all the study groups, particularly CD8+ effector T cell subsets (Fig. 7A). Next, we compared proportion of Gag and Env-specific monofunctional and polyfunctional CD4+ EM, CD8+ EM and CD8+ TD cells secreting MIP-1β, across VC, VNP and PuP groups (Fig. 7B). Here, except for Gag-specific CD4+ EM response, VNPs demonstrated increased proportion of Gag and Env-specific MIP-1β secreting monofunctional cells (Range:20%-37%; Median: 33%) compared to both VC (Range:08%-28%; Median: 11%) and PuP (Range:12%-35%; Median: 18%) groups. VNPs, together with PuPs, had higher proportion of Gag and Env-specific MIP-1β secreting polyfunctional cells (Range:63%-94%; Median: 87% and Range:81%-95%; Median: 94%, respectively) in the studied subset, except for CD4+ EM subset, compared to VC (Range:46%-87%; Median: 62%) group. The boxed regions in Fig. 7 most clearly represent this dichotomy between monofunctional and polyfunctional response.

Fig. 7. Contribution of CD107a and MIP-1β towards polyfunctionality.

(A) Heat-map depicting proportion of CD107a among aggregate monofunctional and polyfunctional Gag and Env-specific CD4+ EM, CD8+ EM and CD8+ TD response, respectively, across study groups (as per color key). (B) Heat-map depicting proportion of MIP-1β among aggregate monofunctional and polyfunctional Gag and Env-specific CD4+ EM, CD8+ EM and CD8+ TD response, respectively, across study groups. The boxed regions highlight the strongest disparity.

3.6. Altered gut homing of memory CD8 T cells in VNPs

To understand underlying mechanisms of non-progression other than dominant non-cytolytic antiviral response observed in VNPs, we performed immunophenotyping of thawed PBMCs from eight VNPs and eight PuPs to evaluate systemic distribution of CCR5 (HIV Co-receptor) and integrin-β7 (Gut homing marker) across CD4+ and CD8+ T cell subsets (Gorfu et al., 2009; Kasarpalkar et al., 2021; Reynes et al., 2000) (Fig. S6). We observed no significant difference of CCR5 expressing CD4+ T cell subsets between VNPs and PuPs (Fig. 8A). We also explored the possibility of altered gut homing of circulating T cells and intriguingly, we noticed a significantly lower expression of integrin-β7 on memory CD8+ T cell subsets (Fig. 8B and 8C) suggesting lower mobilization of these cells to the gut in VNPs.

Fig. 8. Distribution of CCR5 and Integrin-β7 across CD4+ and CD8+ T cell subsets.

(A) Frequency of CCR5 expressing CD4+ T cell subsets (Central memory, CM; Transitional memory, TM; Effector memory, EM; Terminally differentiated, TD1 and TD2) in VNPs and PuPs (as per color key). (B) Frequency of integrin-β7 expressing CD4+ T cell subsets (CM; TM; EM; TD1 and TD2) in VNPs and PuPs. (C) Frequency of integrin-β7 expressing CD8+ T cell subsets (CM; TM; EM; TD1 and TD2) in VNPs and PuPs. Comparison between groups was performed using nonparametric Mann-Whitney test. p-value < 0.05 (*p < 0.05; **p < 0.01; ***p < 0.001) were considered statistically significant.

4. Discussion

The role of antigen-specific T cell responses and their contribution towards progression free survival of VNPs in the face of robust viral replication remains poorly understood. We undertook a systematic investigation of HIV-specific T cell responses present in VNPs compared to two groups - (1) VCs, known to efficiently control viremia and (2) PuPs, unable to control viremia in the context of continually depleting CD4+ T cell count. VNPs resemble non-pathogenic SIV infection of natural hosts who maintain low level of immune activation compared to pathogenic SIV infection (Choudhary et al., 2007; Pandrea et al., 2008a; Rotger et al., 2011; Silvestri et al., 2003). Our previous work has shown that VNPs have similar levels of immune activation compared to recently infected putative progressors, most likely due to ongoing viral replication that in turn drives chronic immune activation (Singh et al., 2020). Nevertheless, maintenance of CD4+ regulatory T cells (Tregs) seemed to play an important role in mitigating effects of immune activation in VNPs, as has also been reported in their paediatric counterparts (Roider et al., 2019; Singh et al., 2020). Our study also revealed that VNPs maintained intact IL-7 mediated CD4+ T cell homeostasis and efficiently preserved central memory CD4+ T cells (Singh et al., 2020). In addition, limited HIV-1 infection of central memory and stem cell central memory CD4+ T cells has also been reported in VNPs (Klatt et al., 2014). Taken together, in the context of HIV-specific T cell immunity, these results suggest that VNPs have an intact ‘CD4 help’ component but may have unique, possibly impaired anti-viral responses compared to VCs.

To begin with, we showed that aggregate Gag and Env-specific responses detected in both CD4+ and CD8+ T cell subsets are not drastically different among VC, VNP and PuP groups and do not correlate with viremia. This finding, in line with earlier work, asserts that control over viral replication is not exclusively dependent on the total HIV-specific response (Betts et al., 2001). Also, these observations differ from our previously reported studies with virologically suppressed ART receiving individuals where Gag specific responses dominated (Salwe et al., 2019). Consequently, we studied the multidimensionality of total Gag and Env-specific response by comparing relative contribution of IFN-γ, IL-2, CD107a, TNF-α and MIP-1β to that of total HIV-specific responses and their concurrent detection (polyfunctionality). Our data revealed a unique reciprocal functional profile where highest CD4+ CM and CD4+ EM Gag-specific IFN-γ responses were observed in VC and PuP groups, respectively, with VNPs demonstrating intermediate levels of these responses. Our previous results demonstrated that both VCs and VNPs preserve CD4+ CM counts despite several years of viremic disease, while PuPs despite similar CD4+ T cell count at recruitment experienced CD4+ CM depletion soon after infection (Singh et al., 2020). Also, in this study, CD4+ CM Gag-specific IFN-γ responses strongly correlated negatively with viremia overall. Thus, our results suggest that VNPs, through CD4+ CM preservation, retain the ability to mount CD4+ CM Gag-specific IFN-γ responses that may contribute to limited control over viral replication. Additionally, we observed low magnitude but no significant difference in HIV-specific IL-2 secreting cells among study groups. Furthermore, these cells were predominantly monofunctional in studied CD4+ T cell subsets. Since viremic HIV infection has been associated with suppressed IL-2 mediated CD4+ T cell proliferative capacity, this observation is a likely outcome of persistent viremia in our study groups (McNeil et al., 2001; Palmer et al., 2002; Platten et al., 2016; Younes et al., 2003). Coupled with the observation, in this study CD4+ CM Gag-specific polyfunctional cells secreting IL-2 were found to be TNF-α positive but lacked IFN-γ secretion. Thus, our data suggests a mechanism shared by VCs and VNPs, where Gag-specific CD4+ memory cells producing IFN-γ with limited autocrine proliferative capacity (IL2-IFN-γ+) may be generated through proliferation and differentiation of CD4 + CM (IL2 + IFN-γ-) cells.

Preservation of a CD4+ helper response in VNPs would also be expected to enhance the anti-viral cytotoxic T cell response (Johnson et al., 2015; Migueles et al., 2008, 2002; Patel et al., 2010; Rosenberg et al., 1997; Swain et al., 2012). Surprisingly, our results demonstrated that VNPs who exhibited high viral replication had significantly reduced proportions of HIV-specific cytolytic CD4+ EM cells, denoted by their degranulation ability (CD107a+), compared to both VCs and recently infected PuPs. Similarly, diminished CD8+ cytolytic potential was observed, predominantly for Env-specific CD8+ EM and CD8+ TD cells, in VNPs. Thus, our study highlights a unique dampened cytolytic HIV-specific response associated with viremic non-progression spanning both CD4+ and CD8+ T cell responses against Gag and Env that underlies the impaired virological suppression observed in these individuals. Indeed, an earlier study by Hersperger et al., focussing on Elite Controllers (ECs) has also reported reduced perforin production by HIV-specific CD8+ T cells in a smaller group of VNPs (n = 6) compared to ECs (Hersperger et al., 2010). The major constituent of the HIV-specific response in VNPs was represented by an increased proportion of MIP-1β producing cells and a concomitant lack of CD107a expression. This altered profile was also dominant in HIV-specific CD8+ EM, CD8+ TD cells as well as in Env-specific CD4+ EM cells. Previous reports have demonstrated that production of soluble factors like MIP-1β, a specific ligand of CCR5 (HIV co-receptor on CD4 T cells), along with other members of β-chemokines (MIP-1α and RANTES), suppress HIV-replication via. non-cytolytic mechanisms (Cocchi et al., 1995; Kaur et al., 2007; McBrien et al., 2018; Ondoa et al., 2002; Saunders et al., 2011). Our data thus seemed to support the presence of such mechanisms operating within VNPs that could contribute to partial control of viral replication.

Although many studies including our data strongly support non-cytolytic mechanisms of HIV/SIV-specific CD8+ and CD4+ T cell in suppressing viral replication (Klatt et al., 2010; Morvan et al., 2021; Seich al Basatena et al., 2013; Walker and Levy, 1989; Wiviott et al., 1990; Wong et al., 2010; Zanoni et al., 2020), the complex multifactorial mechanism of anti-viral non-cytolytic control remains poorly understood.

A significant feature of our study was the incorporation of the PuP group that had a history of relatively rapid CD4+ T cell depletion but comparable CD4+ T cell counts to VCs and VNPs at the time of recruitment. This afforded us an opportunity to evaluate and compare HIV-specific responses in typical progression at apparently similar levels of immune competence compared to non-progressor phenotypes. Interestingly, as opposed to either a dominant HIV-specific cytolytic (CD107a+) or non-cytolytic (MIP-1β) response in VCs and VNPs, respectively, PuPs clearly shared features common to both these responses. This suggests that typical disease progression occurred in these individuals in spite of a balanced and multi-dimensional HIV specific response. However, in contrast to VCs and VNPs, depletion of CD4+ central memory compartment in PuPs and associated impairment in mounting HIV-specific responses therein (IFN-γ production) may be the determinant for progression.

Of late, evidence has been accumulating to suggest that T cell exhaustion in chronic viral infection may occur as a result of altered differentiation by which T cells functionally adapt their effector capacity (Petrovas et al., 2006; Speiser et al., 2014) to mitigate immunopathology (Cornberg et al., 2013; Schönrich and Raftery, 2019; Zehn et al., 2016). Downregulation of PD-1, an immune checkpoint inhibitor, has also been associated with increased cytolytic CD8+ T cell response (Taylor et al., 2016). Additionally, we have previously reported that VNPs show moderate increase in PD-1 expression on CD8+ effectors compared to progressors (Singh et al., 2020). Taken together, these findings implicate a possible role for immune regulators like PD-1 expressed on HIV-specific effectors that may play a role in the emergence of a non-cytolytic response in VNPs. However, it is hard to ascertain whether a modulation of overall HIV-specific responses towards a dominant non-cytolytic signature in VNPs is established early during infection or is a result of extended viral-host coevolution. Interestingly, our results demonstrating significantly reduced frequency of gut-homing circulating memory CD8+ T cells in VNPs lead us to speculate on possibly lower cytotoxic T lymphocyte (CTL) associated gut immune pathology within VNPs (Panetti et al., 2021; Rouse and Sehrawat, 2010) and warrant further investigation.

Of note, in contrast to overall HIV-specific responses, VNPs retained polyfunctional profiles with cytolytic ability. Further analysis revealed that both monofunctional and polyfunctional cells significantly contribute to cytolytic responses in VCs, while in VNPs this response is primarily contributed by polyfunctional cells. Conversely, VNPs had increased proportion of both monofunctional and polyfunctional cells capable of secreting MIP-1 β. These findings highlight the preservation of HIV-specific polyfunctional cells with intact cytolytic potential, albeit a minority, despite several years of viremic infection. Although this study has been undertaken with a relatively limited sample size, we believe it provides significant novel information on modulation of HIV-specific T cell responses within VNPs in the context of relatively high viral replication, but stable CD4+ T cell counts. Additionally, predicting the importance of the quality and quantity of the HIV-specific response in influencing disease progression has been systematically evaluated using a mathematical model by Graw et al (Graw and Regoes, 2014). They report: ‘We find that the strength of the response is a good predictor of disease progression, while functional diversity has only a minor influence.’ Further, they suggest that an improved non-cytolytic effector function would lead to increased probability of CD4+ T cell preservation and improved control of viremia. Thus, non-progression in VNPs seems to fit into the model with presence of functional non-cytolytic response, stable CD4+ T cell count with accompanying retention of ‘CD4 help’ and limited but sustained viral control. Hence, further studies are necessary to understand the multifactorial mechanism of sustained non-cytolytic effector response in VNPs. If discovered, immunotherapeutic interventions can be employed to achieve drug-free remission, i.e., a functional cure, among HIV-infected individuals.

Abbreviations

VNP

Viremic Non-Progressors

VC

Viremic Controllers

PuP

Putative Progressors.

Data availability

The data generated for this study are available on request to the corresponding author.