|Home | About | Journals | Submit | Contact Us | Français|
The generation of polyfunctional CD8+ T cells, in response to vaccination or natural infection, has been associated with improved protective immunity. However, it is unclear whether the maintenance of polyfunctionality is related to particular cellular phenotypic characteristics. To determine whether the cytokine expression profile is linked to the memory differentiation stage, we analyzed the degree of polyfunctionality of HIV-specific CD8+ T cells within different memory subpopulations in 20 ART-naïve HIV-1 infected individuals at approximately 34 weeks post infection. These profiles were compared with CMV-specific CD8+ T cell responses in HIV-uninfected controls and in individuals chronically infected with HIV. Our results showed that the polyfunctional abilities of HIV-specific CD8+ T cells differed according to their memory phenotype. Early-differentiated cells (CD45RO+CD27+) exhibited a higher proportion of cells positive for three or four functions (p<0.001), and a lower proportion of mono-functional cells (p<0.001) compared to terminally-differentiated (CD45RO−CD27−) HIV-specific CD8+ T cells. The majority of terminally-differentiated HIV-specific CD8+ T cells were mono-functional (median 69% [IQR: 57–83]), producing predominantly CD107a or MIP1β. Moreover, proportions of HIV-specific mono-functional CD8+ T cells positively associated with proportions of terminally-differentiated HIV-specific CD8+ T cells (p=0.019, r=0.54). In contrast, CMV-specific CD8+ T cell polyfunctional capacities were similar across all memory subpopulations, with terminally- and early-differentiated cells endowed with comparable polyfunctionality. Overall, these data show that the polyfunctional abilities of HIV-specific CD8+ T cells are influenced by the stage of memory differentiation, which is not the case for CMV-specific responses.
An increasing body of evidence suggests that HIV-1 replication can be partially controlled by T cell immune responses (1–3). However, progression to AIDS occurs in almost all untreated individuals, reflecting the inability of the immune system to mount effective, sustained responses. It has been clearly demonstrated that the overall magnitude of IFNγ-producing HIV-specific CD8+ T cells does not associate with viral control or the establishment of the viral set point (4–8). However, Betts and others have reported that long term non-progressors (LTNPs) were characterized by having higher frequencies of polyfunctional HIV-specific CD8+ T cells compared to non-controllers (9–12), introducing the concept that the quality, rather than the quantity, of antigen-specific T cell responses may dictate T-cell antiviral capacity and play a role in controlling viral replication.
In addition to the range of functional abilities that antigen-specific T cells exhibit, there is enormous phenotypic heterogeneity within the memory T cell compartment. Tremendous effort has been deployed to decipher the phenotypic characteristics of HIV-specific T cells that correlate with viral control. HIV-specific CD8+ T cells possess a distinct memory differentiation profile when compared to other virus-specific CD8+ T cells such as EBV, CMV and HCV (13), exhibiting mainly a transitional memory phenotype (i.e. CD45RO+CD27+CCR7−) (14–16). Moreover, the distribution of HIV-specific CD8+ memory T cells in early infection influences the subsequent viral set point, where proportions of terminally-differentiated HIV-specific CD8+ T cells (defined as CD45RA−CCR7− or CD45RO+CCR7−CD27−) associated with high viral set points (15, 17, 18). This would suggests that excessive maturation of HIV-specific memory CD8+ T cells to the late stage of differentiation could impact on the ability to control virus.
Distinct memory T cell subsets have different proliferation, survival and homing capacities (19–21), so it is therefore plausible that cell maturation could impact on the functional abilities of antigen-specific CD8+ T cells. Indeed, different CD8+ T cell memory subsets have variable capacities to produce cytokines such as IFNγ or IL-2 in response to stimulation with PMA/ionomycin (22), establishing that distinct memory subpopulations have differing inherent functional abilities. However, the relationship between the differentiation stage of a T cell and its polyfunctional profile appears to be more complex, since antigen-specific CD8+ T cells exhibit unique memory maturation profiles depending on their viral specificity (13, 23). Thus, it is likely that antigen load, antigen persistence, quality of co-stimulation as well as the cytokine milieu are all contributing factors controlling both cell differentiation and the polyfunctional profile of antigen-specific CD8+ T cells (24–27). However, it is still unclear whether the stage of memory maturation defines the polyfunctional capacity of CD8+T cell after antigen re-stimulation. In this study, we sought to examine the link between memory phenotype and the polyfunctional ability of antigen-specific CD8+ T cells, by assessing the degree of polyfunctionality of HIV-specific CD8+ T cells within distinct memory subsets in 20 ART-naïve HIV-1 infected individuals and comparing them to CMV-specific CD8+ T cell responses.
For HIV-specific CD8+ T cell responses, a subset of 20 individuals, from the Centre for AIDS Programme of Research in South Africa (CAPRISA) HIV-1 acute infection cohort were analyzed. This cohort, located in Durban, South Africa, has been described previously (28, 29). The time post infection was estimated either by a prospective RNA positive/antibody negative measurement or taken as the midpoint between the last antibody negative test and first antibody positive enzyme-linked immunosorbent assay test. Study participants were followed for 12 months, and follow-up is ongoing. Data from samples collected at approximately 8 months post infection (median 34 weeks, ranging from 22 to 42), are reported herein. All studied subject were antiretroviral therapy (ART) naïve. The University of KwaZulu-Natal, University of Witwatersrand and University of Cape Town Research Ethics Committees approved the study, and all participants provided written informed consent for participation in the study.
Since no data were collected on CMV-specific responses in the CAPRISA cohort, we were obliged to use a different cohort to compare HIV and CMV-specific CD8+ T cell responses. Hence for CMV-specific CD8+ T cell responses, 12 individuals from another cohort (Canadian/African Prevention Trial, CAPT) recruited from the Perinatal HIV Research Unit, Johannesburg, South Africa, were analyzed. Six individuals were HIV-uninfected and the remaining six were chronically infected with HIV-1 and were ART-untreated. The University of Witwatersrand Research Ethics Committee approved the study, and all participants provided written informed consent for participation in the study.
Plasma HIV-1 RNA levels were quantified using the COBAS AMPLICOR™ HIV-1 monitor test version 1.5 (Roche Diagnostics). Absolute blood CD4+ and CD8+ T cell counts were measured using a FACSCalibur flow cytometer and expressed as cells/mm3. For the CAPRISA cohort (n=20), the median plasma viral load, at the time point studied, was 26,550 HIV RNA copies/ml (ranging from 400 to 425,000) and the median CD4+ count was 484 cells/mm3 (ranging from 342 to 1411). For the HIV-infected subjects (n=6) from the CAPT cohort, the median viral load was 35,625 HIV RNA copies/ml (ranging from 440 to 281,000) and the median CD4 count was 542 cells/mm3 (ranging from 391 to 713). Of note, there were no statistical differences in the absolute CD4 count and viral load between the two HIV-infected groups.
A set of 432 synthetic overlapping peptides spanning the entire expressed HIV-1 clade C proteome corresponding to gene products from the HIV-1 consensus C (Gag), isolate Du151 (Pol and Nef) and isolate Du179 (gp160 Env) were synthesized using 9-Fluorenylmethoxy carbonyl chemistry and standard based solid phase techniques (Natural and Medical Sciences Institute, University of Tubingen, Tubingen, Germany). The non-consensus synthesized peptides were based on sequences from isolates used for manufacture of a clade C vaccine (30). The CMV peptide pool, consisting of a set of 138 peptides (15mers overlapping by 11 amino acids) corresponding to HCMV pp65, was obtained from the NIH AIDS Research and Reference Reagent Program. Pooled peptides were used at a final concentration of 1 μg/ml.
In three individuals, HIV-specific CD8+ T cell polyfunctional profiles were also measured in response to autologous 9mer single peptides. For CAP210, p24-Gag164-172 (Gag-YL9: 164YVDRFFKTL172) and Vif79-88 (Vif-WI9: 79WHLGHGVSI88) peptides were used. For CAP228, Env56-64 (Env-DV9: 56DAKAYEREV64) peptide was used, and for CAP239, Nef82-90 (Nef-KF9: 82KAAVDLSFF90) was used. Each peptide was used at a final concentration of 2μg/ml. Viral sequencing was carried out as previously described (31).
The estimated purity of peptides was >80% as measured by high performance liquid chromatography and mass spectrometry. Individual peptides were diluted in dimethyl sulfoxide (DMSO, Sigma-Aldrich) and prepared as previously described (6). All prepared individual peptides or peptide pools were stored at −80°C prior to use.
Peripheral blood mononuclear cells (PBMC) were isolated by standard Ficoll-Hypaque density gradient centrifugation (Amersham Pharmacia) and cryopreserved in 90% heat inactivated fetal bovine serum (FBS, Invitrogen) with 10% DMSO in liquid nitrogen until needed. Frozen PBMC were thawed and rested in RPMI 1640 (Invitrogen) containing 10% heat-inactivated FBS and 50 IU of Gentamycin (Invitrogen) at 37°C and 5% CO2 for 18 hours prior to use in intracellular cytokine staining assays.
Flow cytometric detection of phenotypic and functional markers was performed as described in (32). The following antibodies and fluorophores were used: CD3-APCcy7, CD45RO-PE TexasRed, CD27-PEcy5, IFNγ-FITC and TNFα-PEcy7 (all Beckman Coulter), CD4-PEcy5.5 (Caltag), CD8-QD705 (Invitrogen), IL-2-APC and CD107a-Alexa680 (ebioscience), MIP1β-PE (R&D), CD14-Pacific Blue and CD19-Pacific Blue (both Biolegend), and the violet amine reactive dye “Vivid” (Molecular Probes). All antibodies were pre-titered to optimal concentrations. Briefly, PBMC were stimulated with anti-CD28 and anti-CD49d (1 μg/ml) and one to four peptide pools (Gag, Pol, Nef and Env), autologous peptides or the CMV peptide pool for 6 hours in the presence of brefeldin-A (10 μg/ml, Sigma), monensin (0.7 μg/ml, BD Biosciences) and CD107a antibodies. Cells were first stained with Vivid for 10 min and washed once with PBS. Cells were then surface-stained with CD4, CD8, CD45RO, CD27, CD14 and CD19 antibodies. Cells were fixed and permeabilized using Cytofix/Cytoperm buffer (BD Biosciences) and stained intracellularly with CD3, MIP1β, TNFα and IFNγ and IL-2. After washing, cells were resuspended in 1% paraformaldeyde (Electron Microscopy Solutions). Approximately 500 000 events were acquired per stimulation on an LSRII flow cytometer (BD Biosciences) and analysis was performed using FlowJo (v9.4.3, Treestar). Dead cells (ViVid+), monocytes (CD14+) and B cells (CD19+) were excluded from the analysis. Cells were gated on singlets, live CD3+, small lymphocytes and CD8+ cells. Naïve CD8+ T cells (CD45RO−CD27+) were excluded to identify total CD8+ antigen-experienced T cells (hereafter referred to as ‘memory’ cells). Different memory subsets were identified using CD45RO and CD27 expression, and HIV− or CMV-specific CD8+ T cells were identified based on CD107a, MIP1β, TNFα and/or IFNγ expression. We do not report on IL-2 production as it was absent or below our cutoff for a positive response (see below) for HIV and CMV-specific CD8+ T cells. Memory subsets and cytokine producing CD8+ T cells are expressed as a percentage of total CD8+ memory cells. The gating strategy is provided in Supplementary Figure S1. A positive cytokine response was defined as at least twice the background (no antigen, only co-stimulatory antibodies), greater than 0.05% of total memory T cells and at least 40 events. The latter criterion was introduced so as to minimize the possibility of error due to a low number of events when further subdividing these cells into the four memory subsets. Polyfunctionality of antigen-specific cells was analyzed using a boolean gating strategy and represented visually using Pestle (v1.6.2) and Spice (v5.1) software (provided by the NIH) (33).
Statistical analysis and graphical presentation were performed using InStat and Prism software (v5, GraphPad). Data are expressed as median values with interquartile ranges and analyzed by the use of nonparametric statistics. Statistical analysis of significance was calculated using either Mann-Whitney or Kruskal-Wallis ANOVA using Dunn’s test for multiple comparisons. All tests were two-tailed and a value of p<0.05 was considered statistically significant. The relationship between memory populations was analyzed using Spearman rank correlations.
To investigate the relationship between the polyfunctional repertoire of HIV-specific CD8+ T cells and their maturation profiles, we selected positive responders to Gag, Env, Nef and Pol peptide pools amongst HIV-1 subtype C infected individuals. Twenty individuals were tested for Gag responses, 16 for Env and Nef responses and 8 for Pol responses. Clinical characteristics of study participants are reported in the Methods section. PBMCs, collected at a median of 34 weeks post infection were stimulated with the relevant peptide pools and labeled with a panel of antibodies to assess four different functions and four distinct memory subsets. The functions measured were the cytokines IFNγ and TNFα, the chemokine MIP1β, and CD107a, used as a surrogate marker of degranulation. The four memory subsets were early-differentiated cells (ED: CD45RO+CD27+), late-differentiated cells (LD: CD45RO+CD27−), intermediate cells (Inter: CD45RO−CD27dim) and terminally-differentiated cells (TD: CD45RO−CD27−). The average magnitude of total HIV-specific CD8+ T cells in responding individuals was 1.8% (ranging from 0.15% to 11.2% of total memory CD8+ T cells, Figure 1A). No statistically significant differences were observed between the frequencies of Gag, Env, Nef or Pol responses. Of note, the frequency of HIV-specific CD8+ T cells strongly associated with the absolute number of HIV-specific CD8+ T cells (calculated based on absolute CD8+ count, r=0.84, p<0.0001, data not shown). Further assessment of the combination of functions exhibited by HIV-specific CD8+ T cell responses revealed that approximately 40% of the total responses consisted of cells positive for three or four functions, with no significant differences in the polyfunctional abilities of cells between the different HIV peptide pools tested (Figure 1B). Regardless of the specificity of the CD8+ T cell responses against the HIV peptide pools tested, the most prevalent population observed consisted of cells simultaneously producing IFNγ, MIP1β, and degranulating (IFNγ+MIP1β+CD107a+; median 25% of total response [IQR: 11–34]). HIV-specific CD8+ T cells endowed with two functions were mostly distributed amongst the CD107a+MIP1β+, IFNγ+MIP1β+ and CD107a+IFNγ+ cells (median 11% [IQR: 7–17], 8% [IQR: 4–9] and 6% [IQR: 1.6–7] of total response, respectively). Mono-functional HIV-specific CD8+ T cells were predominantly CD107a+ or MIP1β+ (median 18% [IQR: 10–29] and 14% [IQR: 8–23] of total response, respectively). Of note, the proportion of cells producing only IFNγ was marginal, representing less than 3% of the total measured response, and TNFα was detected only in cells endowed with 4 functions (Figure 1B). These profiles are consistent with previous reports on the functional characteristics of HIV-specific CD8+ T cells (9, 34–36).
We next assessed the memory maturation profiles of HIV-specific CD8+ T cells based on the expression of CD45RO and CD27. A representative dot plot of the memory profile of HIV-specific CD8+ T cells is presented in Figure 1C. Total HIV-specific CD8+ T cells exhibited predominantly an early-differentiated (ED) memory phenotype (CD45RO+CD27+); however, the frequencies of ED HIV-specific CD8+ T cells were variable amongst the individuals analyzed (median 57%, ranging from 7% to 88%, Figure 1D). The percentage of ED HIV-specific CD8+ T cells was negatively correlated with viral load (p<0.0001, r=−0.6, data not shown), suggesting that these variations likely reflect the differences in viral replication levels in these individuals. This is in accordance with data previously reported from the same cohort (15). Of note, the memory maturation profiles did not differ significantly between Gag, Env, Nef and Pol-specific responses (Supplementary Figure S2A).
In summary, our results show that Gag-, Env-, Nef- and Pol-responsive CD8+ T cells were quantitatively and qualitatively similar in our cohort, and HIV-s pecific CD8+ T cells exhibited mainly an ED memory phenotype and moderate polyfunctionality.
To determine whether the polyfunctional properties exhibited by antigen-specific CD8+ T cells were dictated by their differentiation stage, we next compared the functional profiles of HIV-specific CD8+ T cells within different memory subsets. Since the functions and phenotypes of CD8+ T cells specific for Gag, Env, Nef and Pol were similar (Supplementary Figure S2B), we combined all HIV-specific responses for these analyses, and here report on 43 peptide pool responses in 20 individuals. Briefly, total HIV-specific CD8+ T cells were gated on distinct memory subsets (ED, Inter, LD and ED), and we then measured the distribution of HIV-specific CD8+ T cells expressing 4, 3, 2 or 1 function in each memory subpopulation. Figure 2A shows representative dot plots of memory profiles of HIV-specific T cells positive for 4, 3, 2 or 1 function, from two individuals with a high and a low magnitude Gag-specific CD8+ T cell response. Depending on their memory profiles, HIV-specific CD8+ T cells had distinct polyfunctional abilities, with decreasing polyfunctionality coinciding with an increase in differentiation from ED to TD memory subsets (Figure 2B). HIV-specific CD8+ T cells with a TD phenotype were primarily mono-functional (median 69% [IQR: 57–83]), whilst ED CD8+ T cells had a significantly lower proportion of mono-functional cells (median 27% [IQR: 16–38], p<0.001) and a significantly higher proportion of cells positive for 4 and 3 functions (p<0.001) compared to TD CD8+ T cells. Inter and LD HIV-specific CD8+ T cells showed an intermediate polyfunctional profile consisting of: 1) significantly more mono-functional cells compared to the ED subset (median 51% [IQR: 22–61], p<0.01 and 43% [IQR: 30–54], p<0.01, respectively), and 2) significantly more cells positive for 4 and 3 functions compared to the TD subset (Figure 2B). It is worth noting that although the TD subset had the highest proportion of mono-functional cells (Figure 2B), they contributed only approximately 20% to the absolute number of HIV-specific mono-functional CD8+ T cells (data not shown), since the TD subset itself represents only a median of 14% of all HIV-specific CD8+ T cells (as shown in Figure 1D). Consequently, even though there was a lower proportion of mono-functional CD8+ HIV-specific cells within the ED subset compared to the TD subset, in absolute numbers there were more circulating mono-functional ED cells than mono-functional TD cells, since ED cells make up almost 60% of total HIV-specific CD8+ T cells.
To determine whether differences in HIV-specific CD8+ T cell polyfunctional profiles observed within each memory subset could also be identified using single peptides, and were not an artifact of heterogeneous specificities in the pools, we assessed the degree of polyfunctionality in each memory subpopulation in response to stimulation with optimal autologous 9mer peptides. Figure 2C shows the polyfunctional profile of Gag-VL9-specific CD8+ T cells within the different memory subsets in one individual (CAP210). We found that increasing cell differentiation from ED to TD subsets resulted in a decreased proportion of epitope-specific cells exhibiting 3 and 4 functions, and an elevated proportion of mono-functional epitope-specific cells. Four autologous peptide responses (Gag-VL9, Vif-WI9, Env-DV9 and Nef-KF9) were measured in three individuals (CAP210, CAP228 and CAP239). As demonstrated with peptide pools, epitope-specific CD8+ responses displayed greater polyfunctionality in the ED subset when compared to the TD subset (p<0.05; Figure 2D).
To assess the specific cytokines that made up the different polyfunctional subsets, we next compared the proportions of the eight most prevalent combinations of functions detected within each HIV-specific CD8+ T cell memory subset (Figure 3A). Both CD107a+ and MIP1β+ cells accounted for the increased proportion of mono-functional cells in the TD memory subset; the proportion of CD107a+ cells were almost three-fold higher in ED compared to TD subsets, from a median of 13% (IQR: 6–17) to 33% (IQR: 16–45; p<0.0001), whilst the proportion of CD8+ T cells producing MIP1β alone was a median of 13% (IQR: 5–15) in ED HIV-specific cells compared to 34% (IQR: 19–43) in TD HIV-specific cells (p<0.0001; Figure 3A). Interestingly, no significant change in the proportion of cells expressing IFNγ only was observed between the different memory subpopulations. Figure 3B summarizes the mean distribution of detectable cytokine combinations within each HIV-specific CD8+ T cell memory subset, illustrating the progressive increase in the proportion of CD107a+ and MIP1β+ mono-functional cells and the decrease in the proportion of cells endowed with 3 functions (CD107a+MIP1β+IFNγ+) and 4 functions in HIV-specific cells from early to terminal differentiation. Importantly, the degree of polyfunctionality of HIV-specific CD8+ T cells in each memory subset was similar between individuals with low (VL<3,000 copies/ml) or high viral load (VL>100,000 copies/ml; Supplementary Figure S3A), indicating that the distribution of polyfunctional cells within each memory subset was independent of viraemia.
Overall, these data suggest that in HIV infection the polyfunctional profile of HIV-specific cells is related to their differentiation stage, where early-differentiated cells exhibit a greater degree of polyfunctionality, and terminally-differentiated cells are more mono-functional.
To investigate whether our observations linking the degree of polyfunctionality with the memory maturation profile in HIV infection holds true for another pathogen, we measured the polyfunctional attributes of distinct CMV-specific CD8+ T cell subsets in 12 participants, six of whom were HIV-uninfected and six chronically infected with HIV. All HIV-infected participants were ART-naïve. CMV-specific CD8+ T cells were highly polyfunctional, with approximately 75% (IQR: 63–85) of the total response consisting of cells positive for 4 functions (i.e. CD107a, MIP1β, IFNγ and TNFα; Figure 4A). No significant differences were detected between HIV-infected and uninfected individuals. Interestingly, the prevalent cytokine combinations observed in CMV-specific CD8+ T cells were distinct from HIV-specific responses, with CMV-specific cells positive for 3 functions consisting mostly of MIP1β+IFNγ+TNFα+ cells. CMV-specific cells endowed with 2 functions were present at low levels or were undetectable, and mono-functional CMV-specific cells expressing TNFα only were detectable (Figure 4A), a subset that was absent for HIV-specific responses. Representative examples of CMV-specific CD8+ T cells and their memory subset distribution are depicted in Figure 4B.
When we compared the cell memory distribution for each antigen, we found that the proportion of TD cells was significantly higher in CMV-specific CD8+ T cells when compared to HIV-specific CD8+ T cells (median 31% [IQR: 24–40] and 14% [IQR: 9–20], respectively, p=0.003), and this was true regardless of HIV infection status (Figure 4C). Thus, CMV-specific responses exhibited a high degree of polyfunctionality and a differentiated phenotype, as previously described (13, 16, 34).
We next examined the distribution of polyfunctionality between different CMV-specific CD8+ T cell memory subsets, as we had done for HIV. Interestingly, and in contrast to HIV-specific CD8+ T cells, the polyfunctional capacity of CMV-specific CD8+ T cells was similar across all memory subsets (Figure 5), suggesting that for CMV-specific CD8+ T cell responses, polyfunctional profiles are neither linked to specific memory subsets nor enriched in particular memory subsets. Of note, the degree of polyfunctionality of CMV-specific CD8+ T cells within each subset was comparable in HIV-infected and uninfected individuals (Supplementary Figure S3B). The difference between HIV− and CMV-specific CD8+ T cells was further emphasized when we compared the polyfunctional properties of each memory subset. Figure 6A shows that ED HIV-specific CD8+ T cells exhibited a significantly higher proportion of cells positive for 2 or 3 functions and a lower proportion of 4-function cells (p<0.0001) compared to ED CMV-specific CD8+ T cells. Moreover, TD HIV-specific CD8+ T cells exhibited a significantly higher proportion of cells positive for 1 or 2 functions (p<0.0001 and p=0.0004, respectively) and a lower proportion of cells positive for 3 or 4 functions (p=0.01 and p<0.0001, respectively) compared to CMV-specific responses. Furthermore, there was a significant positive association between the proportion of HIV-specific CD8+ T cells endowed with one function and the proportion of TD HIV-specific CD8+ T cells (p=0.019, r=0.546), whilst the proportion of HIV-specific CD8+ T cells positive for 3 functions was inversely correlated with the proportion of TD HIV-specific CD8+ T cells (p=0.01, r=−0.587; Figure 6B, left panel). No such associations were observed for CMV-specific CD8+ T cells (Figure 6B, right panel). Overall, these results show that for HIV-specific CD8+ T cells, polyfunctional properties were linked to differentiation levels, in contrast to CMV-specific CD8+ T cells, where polyfunctional capacities were similar across all memory subsets.
Despite abundant literature on both the memory differentiation profiles and polyfunctional capacities of HIV-specific CD8+ T cells at different stages of infection (9–11, 13–15), less is known about how these T cell characteristics are associated with each other. In this study, we examined the relationship between the memory phenotype of CD8+ T cells and the polyfunctional responses of these cells, as measured by IFNγ, TNFα, MIP1β and CD107a expression. We studied HIV-specific CD8+ T cell responses in 20 HIV-1 infected individuals at approximately 34 weeks post infection, and compared them to CMV-specific T responses. We detected a higher proportion of polyfunctional CD8+ T cells specific for CMV compared to HIV, which agrees with previous observations (34). HIV-specific responses were enriched for cells with an early-differentiated phenotype, whilst CMV-specific cells were mainly of a terminally-differentiated phenotype, consistent with earlier findings (14, 37). Our novel observation was a distinct pattern of polyfunctionality between HIV− and CMV-specific CD8+ T cell memory subsets. For HIV-specific CD8+ T cells, early-differentiated memory cells (CD45RO+CD27+) were enriched for a polyfunctional response, and there was a significant reduction in the number of functions in terminally-differentiated cells (CD45RO−CD27−). In contrast, CMV-specific CD8+ T cells exhibited a polyfunctional profile that was similar across early, late or terminally-differentiated memory subsets. Moreover, the CMV response was highly polyfunctional (80–90% of cells expressing 3 or 4 functions) in both HIV-infected and uninfected individuals, compared to HIV-specific CD8+ T cells (40–50%). This implies that there is not a global decrease in CD8+ T cell polyfunctionality in the background of HIV infection. Differences in antigen load or recurrence or CD4+ help may account for differential (poly)function between HIV-specific CD8+ T cells and cells of other specificities. Overall, this suggests that the hierarchical loss in the number of functions by an antigen-specific CD8+ T cell as memory differentiation proceeds is dependent on the infecting pathogen, and that memory phenotype and polyfunctional characteristics of CD8+ T cells can differ significantly.
Seminal work from Sallusto and colleagues (20) introduced the concept of antigen-specific cells being divided into memory subsets expressing different phenotypic markers with distinct homing and survival abilities, and also alluded to a functional distinction between different memory subsets. A number of factors have been shown to shape this functional and phenotypic heterogeneity of antigen-specific T cells. Co-stimulatory signals engaged during T cell priming, as well as cytokines such as IL-2 and IL-21, can modulate T cell functionality (38–41). Major determinants of CD8+ T cell polyfunctionality are antigen concentration (26, 42, 43) and TCR affinity (44). Work from murine models has proposed that the clonal expansion process can modulate the profile of secreted cytokines (45, 46). With regard to memory differentiation of CD8+ T cells, the degree of cell maturation is dependent on similar factors, namely antigen load, co-stimulation signals and the cytokine environment (reviewed in (47)). Hence, being regulated by similar factors, it can be speculated that cell differentiation and polyfunctional potential could be co-dependent phenomena, where the generation and maintenance of late-stage differentiated effector cells, endowed with a rapid response to pathogens and a high degree of polyfunction, would be favorable to ensure viral control. Indeed, in certain well-controlled infections or successful vaccinations (such as CMV, vaccinia or yellow fever vaccine), antigen-specific CD8+ T cells that are generated are highly differentiated and highly polyfunctional (23, 48). In contrast, during uncontrolled HIV infection, HIV-specific CD8+ T cells exhibit mainly an early-differentiated memory phenotype, often regarded as an immature stage (14, 37). However, the relationship between maturation and polyfunction appears to be more complex, as antigen-experienced cells exhibit heterogeneous memory and cytokine secretion profiles dependent upon different antigen specificities (reviewed in (21)). Our data are in agreement with the latter observation, where we show that differentiation towards late memory for HIV-specific CD8+ T cells is accompanied by a progressive loss of polyfunctional capacities, whilst CMV-specific CD8+ T cells retain their polyfunctional capacities regardless of memory differentiation.
Although sustained HIV replication appears to play a predominant role in driving CD8+ T cells towards a late stage of memory maturation (15, 17, 18), we found no difference in the memory-polyfunction association between those with high and low HIV viral loads, i.e. those controlling viral replication did not resemble the stable ‘polyfunctionality regardless of memory phenotype’ pattern of CMV-specific CD8+ T cells. It would be of interest to determine whether this is also the case for HIV-specific cells from individuals on long-term, successful ART, where antigen load is reduced.
Our starting point for this study was that polyfunctional T cells, capable of carrying out a range of functions simultaneously, exhibit superior protective immunity (9, 49). We can speculate that measuring multiple functions on a per cell basis may more closely reflect the ability of CD8+ T cells to impart antiviral effects and may be more relevant than focusing only on one function, such as IFNγ. However, the combination of specific functions rather the number of functions per se may offer a more refined and accurate assessment of protective immunity (50). Upregulation of perforin has recently been highlighted as an important indicator of cytotoxic potential for control of HIV (51, 52). This is also a consideration in the question of how maturation profiles of CD8+ T cells and functional abilities are linked, since specific functions such as perforin and IL-2 may indeed be associated with particular memory phenotypes, whilst other functions may not be as tightly regulated. Loss of CD28 expression appears to be coupled to loss of IL-2 production, whilst T-bet expression correlates with perforin upregulation (53). In our study, we did not measure perforin and nor could we detect IL-2 responses. Indeed, CMV-specific CD8+ T cells rarely produce IL-2 (53) and IL-2 production in the context of HIV infection is confined to long-term non-progressors (54, 55).
Excessive activation of the immune system can influence both the function and phenotype of CD8+ T cells. Several studies have described the upregulation of PD-1 and other inhibitory receptors such as CD160, 2B4 and LAG3 on CD8+ T cells in chronic viral infections, including HIV, which results in defective cytokine production and lack of polyfunctional responses. This can be partially reversed by blocking the interaction of these receptors with their ligands (56–58). Yamamoto and colleagues (37) recently described elevated levels and co-expression of several negative regulators on HIV-specific CD8+ T cells compared to CMV-specific cells. Co-expression of these inhibitory receptors correlated inversely with polyfunctionality, and PD-1 blockade restored cytokine production. It would thus be of interest to determine whether, in HIV infection, the memory maturation profile of HIV-specific CD8+ T cells coincides with their degree of functional inhibition.
In conclusion, our data show that the polyfunctional abilities of HIV-specific CD8+ T cells are influenced by the stage of memory differentiation, which is not the case for CMV-specific responses. This emphasizes that different pathogens generate CD8+ T cell responses with distinct polyfunctional-memory subset profiles; this may reflect the distinct life histories of pathogens and their interactions with the immune system. A better understanding of which memory-function combinations lead to superior and durable protective immunity against HIV is needed.
We thank the participants, clinical and laboratory staff at CAPRISA for the specimens. We thank Laurie Lamoreaux and Jennifer Fischer for technical help, and Steve Perfetto for LSRII training.
This work was funded by the National Institute of Allergy and Infectious Diseases, National Institutes of Health, U.S. Department of Health and Human Services Grant U19 A151794 (to S.A.K.) and R01 AI084387 (to W.A.B.); the Canada-Africa Prevention Trials Network (CAPTN); and the Bill and Melinda Gates Comprehensive T Cell Vaccine Immune Monitoring Consortium award. W.A.B. was supported by a Columbia University Fogarty AIDS International Training Research Fellowship, and is a Wellcome Trust Intermediate Fellow in Public Health and Tropical Medicine.
Disclosure: The authors have no conflict of interest
CONTRIBUTION:Conceived and designed the experiments: KM, RK, MR, SAK, MRA, CW, CMG and WAB. Performed the experiments: WAB. Analyzed the data: CR, WAB. Contributed reagents/materials/analysis tools: ML, NG, KM, SAK, GdB. Wrote the paper: CR, WAB.