|Home | About | Journals | Submit | Contact Us | Français|
HIV+ elite controllers are a unique group of rare individuals who maintain undetectable viral loads in the absence of antiretroviral therapy. We studied immune responses in these subjects to inform vaccine development, with the goal of identifying the immune correlates of protection from HIV.
We compared markers of cellular activation, HIV-specific immune responses, and regulatory T (Treg) cell frequencies in 4 groups of subjects: HIV-negative healthy controls, elite controllers (HIV RNA level <75 copies/ml), individuals on highly active antiretroviral therapy (HAART), and subjects with HIV RNA level >10,000 copies/ml (non-controllers).
Elite controllers possessed significantly lower levels of activated HIV-specific CD8+ T cells and of recently divided HIV-specific CD4+ T cells than non-controllers, while these differences were not seen in the respective CMV-specific T cell populations. Elite controllers also mounted a stronger and broader cytokine and chemokine response following HIV-specific stimulation than individuals on HAART and non-controllers. Finally, we found that HAART suppressed subjects had elevated Treg cell frequencies, while elite controllers and non-controllers maintained normal percentages of Treg cells.
Elite controllers maintain high levels of HIV-specific immune responses with low levels of HIV-specific T cell activation, and do not have elevated Treg cell levels. Based on these data an ideal HIV vaccine would induce strong HIV-specific immune responses while minimizing HIV-specific T cell activation.
HIV infection is characterized by chronic immune activation, uncontrolled viral replication and a loss of CD4+ T cells, resulting in progression to AIDS. A population of rare individuals (approximately 1% of the HIV infected population) has been shown to maintain clinically undetectable plasma viral loads (<75 copies/ml) in the absence of therapy (“elite controllers”) [1, 2]. Elite controllers have decreased T cell activation [3, 4] and stronger HIV-specific CD4+ T cell IFNγ and IL-2 responses targeted towards the Gag protein than individuals who do not control HIV replication .
It is well documented that there is an increase in the level of CD4+ and CD8+ T cells expressing “activation” markers such as CD38, HLA-DR and/or Ki67 expression in HIV infection [3, 6–9]. For reasons that have not yet been fully defined, a higher frequency of these cells is associated with more rapid disease progression, independent of viral load and CD4+ T cell counts [3, 10–12]. Recent work by our group found that despite having undetectable viral loads, elite controllers had higher levels of immune activation than HIV seronegatives and HAART suppressed individuals [2, 3]. The mechanism accounting for this higher than expected level of immune activation is not known, but may be due to higher levels of HIV production, higher levels of microbial translocation, and/or lower levels of immunoregulatory cells.
Since immune activation appears to play an adverse role in HIV infection, cells with the ability to decrease inflammation may be beneficial. Regulatory T (Treg) cells are a unique population of CD4+ T cells that have the ability to suppress the activation and proliferation of T cells [13–16]. This population of CD4+ T cells has traditionally been identified by the expression of CD25 and the transcription factor FoxP3, and more recently in combination with other markers including CD127 and CD152 [13, 17–23]. Studies of the role of Treg cells in HIV infection show conflicting results. Some studies have shown an increase in the number of Treg cells in HIV infection [24–26], while others have shown a decrease [27–32]. Higher levels of Treg function could prevent some of the harmful effects of immune activation on disease progression, while lower levels could allow a strong and durable antigen-specific T cell response.
Finally, cytokine and chemokine profiles may differ between elite controllers and non-controllers. However, there is considerable variability in the levels of cytokines produced between infected individuals [33–38]. Studies of the earliest phase of HIV infection in plasma donors acquiring HIV revealed a striking pro-inflammatory cytokine cascade beginning within days of the first appearance of viremia [39, 40]. While there have been numerous studies of cytokine profiles in HIV infection, broad analysis of the cytokine responses in elite controllers has not been reported.
Despite strong and growing interest in the elite controllers as a model for understanding the optimal HIV-specific host response, no study to our knowledge has attempted to fully characterize the nature of the T cell response in these individuals. In our current study we assessed the activation and proliferation status of antigen-specific CD4+ and CD8+ T cells, the cytokine and chemokine response to HIV and CMV stimulation, and the number of Treg cells in four groups of subjects. Our sample cohort included HIV seronegative individuals and three classes of HIV infected individuals: elite controllers, HAART suppressed individuals and non-controllers. One sample from each individual was assessed for all parameters at the same time point, with the goal of identifying immunological factors that drive the maintenance of robust T cell responses and control of viral replication in elite controllers in the absence of therapy.
Blood samples were obtained from HIV infected individuals enrolled in the University of California, San Francisco (UCSF) SCOPE cohort and from 20 healthy HIV seronegative individuals. All subjects underwent informed consent under protocols approved by the UCSF Committee on Human Research. HIV infected subjects were divided into three groups: (1) Elite controllers: at least two viral loads <75 copies/ml spanning ≥6 months (median of 10 undetectable viral loads/subject). A viral load >1000 copies was recorded in 3 subjects over 4.5 years prior to study sample acquisition. Three controllers had ART exposure 7, 10, and 14 years prior to study. (2) HAART suppressed: antiretroviral treated with at least two viral loads <75 copies/ml spanning ≥12 months. (3) Non-controllers: untreated with an index viral load >10,000 copies/ml (Table 1). Since advanced immunodeficiency can cause T cell dysfunction, we excluded from our analysis any subject with a CD4+ T cell count <300 cells/µl.
Cryopreserved PBMC were rapidly thawed into RPMI supplemented with 10% heat inactivated human AB serum (Sigma-Aldrich), 10 mM Hepes and 50 IU/ml penicillin/streptomycin (UCSF Cell Culture Facility). One million cells/well were stimulated in 96 well “U” bottom plates (Falcon, BD Labware) for one hour at 37°C, 5% CO2, with either 5µg/ml of an HIV-1 p55 or CMV pp65 peptide pool (123 and 138 peptides, respectively, 15 amino acids long with 11 amino acid overlap, NIH AIDS Research & Reference Reagent Program), 200ng/ml of staphylococcal enterotoxin B (SEB; Sigma-Aldrich) as a positive control, or were left unstimulated as a negative control. Brefeldin A (Sigma-Aldrich) and GolgiStop (BD Biosciences) were added together at final concentrations of 10µg/ml and 8µM, respectively, prior to overnight incubation at 37°C, 5% CO2. Duplicate cultures without inhibitors were set up when cell numbers allowed for analysis of cytokines/chemokines. Supernatants were harvested between 18–24 hours and frozen for batch analysis.
Unless otherwise noted, all reagents were obtained from BD Pharmingen or BD Biosciences. Activation was assessed with the following panel: CD3-Pacific Blue, CD4-Alexa-fluor 700, CD8-APC-H7, CD45RA-PE-Cy5.5 (Caltag Laboratories), CD27-APC, HLA-DR-PE-Cy5, CD38-PE, IFNγ-PE-Cy7, Ki67-FITC and aqua amine-reactive dye (Invitrogen-Molecular Probes). The Treg panel consisted of: CD3-Pacific Blue, CD4-Alexa-fluor 700, CD25-PE-Cy7, CD127-PE, CD152-APC, FoxP3-Alexa-fluor 488 and aqua amine-reactive dye. Gating on CD4+CD25+CD127−CD152+ T cells has been shown to define Treg cells [18, 19, 23, 41]. Our results were consistent with these prior studies showing that this population was also FoxP3+ (data not shown). Activation panel intracellular staining was accomplished using Cytofix and Cytoperm reagents (BD Biosciences) and Treg intracellular staining was performed using a Foxp3 staining buffer set (eBioscience) according to the respective manufacturer’s instructions. Compensation controls included CompBeads or cells stained with equivalent quantities of test antibody. Fluorescence minus one (FMO) and isotype stained controls were used to set gating. Gates for IFNγ were set using unstimulated negative control samples.
The Treg panel was acquired on an LSRII flow cytometer and the activation panel on a FACS Aria. Instrument set-up was standardized to reduce batch-to-batch variation. Pre-optimized target channel voltages were set using mid-range FL1 rainbow fluorescent particles (BD Biosciences). Single stained compensation tubes were checked to ensure each stain was the brightest in its own channel. A median of 150,000 viable CD3+ events was collected for each panel. Data was analyzed using FlowJo 8.7.3 software (TreeStar).
Supernatants were assayed using the high-sensitivity LincoPlex kit (Millipore) for IL-1β, IL-2, IL-4, IL-5, IL-6, IL-7, IL-8, IL-10, IL-12 p70, IL-13, IFNγ, granulocyte-macrophage colony stimulating factor (GM-CSF) and tumor necrosis factor-α (TNFα), and the standard-sensitivity Milliplex Map kit (Millipore) for IFNα2, IL-15, IL-17, inducible protein -10 (IP-10), monocyte chemotactic protein-1 (MCP-1), MCP-3, macrophage-derived chemokine (MDC), macrophage inflammatory protein-1α (MIP-1α), MIP-1β, soluble CD40 ligand (sCD40L) and TNFβ following the manufacturer’s protocols. Samples were acquired on a Labscan 100 analyzer (Luminex) using Bio-Plex manager 4.1 software (Bio-Rad).
Non-parametric tests were used because the data were not normally distributed. The Kruskal-Wallis test followed by Dunn’s test for multiple comparisons was used to assess differences in the distribution of values for HIV seronegative individuals and infected groups. Spearman’s rank test was used to determine the correlation between Treg number and age using Prism 5 software (Graphpad). To examine the association between HAART therapy and the frequency of Treg cells, multivariate linear regression models were generated that included therapy (HAART therapy vs. untreated), age, and CD4+ T cell count as independent variables using Stata 10.1 SE software (StataCorp LP). Heat maps of the cytokine data were generated using Aabel 3 software (Gigawiz). Results were considered statistically significant if p<0.05.
Subjects from four groups were studied, HIV negatives, HIV controllers, HAART suppressed and HIV non-controllers, including 18 HIV controllers and 20 of each of the other groups. Most subjects were men, and the median age for the respective groups was 43, 48, 53 and 44 years (Table 1). The median CD4+ T cell counts for the controllers, HAART-suppressed and non-controllers were 764, 725 and 588 cells/µl (CD4+ T cell counts were not available for the HIV seronegatives).
For unclear reasons, at least a subset of elite controllers is able to maintain vigorous HIV-specific T cell responses for extended periods, with some patients now known to be able to maintain control for up to 30 years. It is known that elite controllers have high frequencies of HIV-specific CD4+ and CD8+ T cells and intermediate levels of general T cell activation when compared to HIV uninfected or HAART suppressed subjects (low activation) and non-controllers (high activation) [3, 5], and we confirmed those findings in the current study (data not shown). However, we hypothesized that HIV-specific cells might be relatively less activated in elite controllers compared to non-controller subjects. We defined activated HIV-specific cells as those co-expressing CD38 and HLA-DR (gated on IFNγ+ cells following HIV p55 stimulation, Fig. 1a). We found significantly lower levels of activated CD8+ but not CD4+ HIV-specific T cells among elite controllers and HAART suppressed individuals compared to non-controllers (Fig. 1b,c). Elite controllers also showed the lowest levels of recently divided HIV-specific CD4+ T cells (expressing Ki67), and their numbers were significantly lower than non-controllers and HAART suppressed subjects (Fig. 1d). There was no significant difference in the turnover of HIV-specific CD8+ T cells among the HIV+ groups (Fig. 1e).
The lower levels of HIV-specific CD8+ T cell activation in elite controllers could have been due to decreased antigen exposure in subjects with lower viral load, as recently suggested [42–44]. To test this we measured activation of CMV-specific T cells and found no significant differences compared to non-controllers (Fig. 2a,b). These results imply that the increased activation seen in HIV-specific CD8+ T cells from non-controllers is due to greater antigenic stimulation compared to elite controllers. It has recently been suggested that CD4+ T cell proliferation (measured by BrdU staining) is determined by both CD4+ T cell depletion and HIV viral burden . One would expect that response to CD4+ T cell depletion would occur equally in HIV-and CMV-specific CD4+ T cells, while HIV viral burden would induce more proliferation in HIV-specific CD4+ T cells in an antigen-specific fashion or equally affect HIV- and CMV-specific T cells through direct mechanisms such as infection of activated cells or induction of bystander apoptosis. Consistent with viral burden rather than CD4+ T cell depletion driving proliferation, frequencies of recently divided CMV-specific T cells were lower compared to HIV-specific T cells among non-controllers (p<0.001 and p<0.05 for CD4+ and CD8+ T cells, respectively). Additionally, the CMV-specific T cells showed no increase in recently divided cells in non-controllers compared to elite controllers (Fig. 2c,d), illustrating no global increase in proliferation of antigen-specific cells in non-controllers compared to elite controllers.
Elite control of HIV infection has been shown to correlate with the presences of cells able to secrete IFNγ, IL-2, TNFα, and MIP-1β. We aimed to extend these observations by focusing on other factors that might be preferentially expressed in T cells from elite controllers. Stimulation with p55 induced pro-inflammatory cytokines and chemokines, including IL-2, IFNγ, TNFα, IP-10, GM-CSF, MCP-3 and MIP-1β. PBMC from elite controllers produced significantly more IFNγ, GM-CSF, IP-10, MCP-3, TNFα and IL-2 than the HIV seronegative individuals after p55 stimulation (Fig. 3a). In contrast, the HAART suppressed group only produced significantly more IFNγ and IP-10 upon p55 stimulation compared to the HIV seronegative individuals. The HAART suppressed group also produced more IL-2 and IL-13 than the non-controllers (Fig. 3a). The non-controllers produced the lowest levels of HIV-induced cytokines, only secreting IFNγ and IP-10 (Fig. 3a). The non-controllers had the lowest mean-fold increase in IFNγ compared to elite controllers and HAART suppressed individuals (mean-fold change of 31.69 vs. 424.23 for elite controllers and 246.50 for HAART suppressed). The non-controllers also secreted less TNFα, GM-CSF and MCP-3 than the elite controllers (Fig. 3a). Collectively, these data demonstrate the elite controllers can produce a wide breadth of cytokines and chemokines after HIV stimulation, including antiviral factors such as MIP-1β.
Stimulation of PBMC with the CMV pp65 peptide pool induced cytokines and chemokines to the same degree in all groups, with the exception of non-controllers who produced significantly less MDC than HIV seronegative individuals and less GM-CSF than HAART suppressed individuals (Fig. 3b). Therefore the alterations in the cytokine and chemokine environment were HIV-specific and not a result of general immune activation resulting from a chronic viral infection.
The decreased levels of activation and proliferation seen in HIV-specific T cells from elite controllers may have been due to higher magnitude Treg cell levels in the elite controllers. We defined Treg cells as CD4+CD25+CD127−CD152+ (Fig. 4a). The frequency and absolute number of these cells in the elite controllers was comparable to that in HIV seronegatives and non-controllers, while the HAART suppressed patients showed an elevated percentage of Treg cells compared to HIV seronegatives and elevated absolute Treg cell numbers compared to non-controllers (Fig. 4b,c). We found no significant correlation between subject age and absolute Treg cell count among HIV+ samples (Fig. 4d). However, when all samples (both HIV+ and HIV−) were analyzed, a positive correlation between the frequency of Treg cells and age was found (data not shown, p<0.05). The HAART suppressed group, which had the highest frequency and absolute number of Treg cells, also had the oldest median age of 53 years (range 35 to 80 years), significantly older than the median age of the HIV seronegative individuals and the HIV+ non-controllers (Table 1, p<0.05 for both comparisons). Multivariable linear regression analysis showed that after adjusting for age and CD4 count, HAART was significantly associated with a 0.60% increase in the frequency of Treg cells on average (p=0.03). Age (p=0.63) and CD4 count (p=0.60) were not significantly associated with an increase in the frequency of Treg cells.
We found that elite controllers maintained lower levels of activated HIV-specific CD8+ and recently divided HIV-specific CD4+ T cells compared to non-controllers, and this was not explained by increased proportions of Treg cells in elite controllers. Decreased antigen-specific T cell activation was limited to the HIV-specific subsets of cells and was not found in the respective CMV-specific populations. Additionally, the elite controllers possessed the strongest and broadest HIV-specific immune responses, with seven cytokines and chemokines induced by HIV stimulation (IL-2, IFNγ, TNFα, GM-CSF, IP-10, MCP-3 and MIP-1β). In summary, elite controllers maintained the strongest HIV-specific immune responses, with markers of inflammation on HIV-specific cells significantly lower than those in subjects unable to control viral replication.
Elite controllers had the ability to mount a strong and broad antiviral cytokine response when stimulated with a p55 peptide pool, producing IL-2, IFNγ, TNFα, IP-10, GM-CSF, MIP-1β and MCP-3. Whilst the non-controllers also induced some of the same cytokines as elite controllers (IL-2, IFNγ, TNFα, GM-CSF, IP-10 and MCP-3), they produced significantly less of each than elite controllers. The lower quantities of antiviral cytokines may reflect weak antiviral T cell responses and be associated with exhaustion of adaptive immune responses. Antiretroviral therapy was clearly associated with a reduced ability to secrete a broad cytokine and chemokine response, with these individuals only secreting significant amounts of IL-2, IFNγ and IP-10. Interestingly, HAART suppression was associated with the production of IL-13, a cytokine involved in B cell growth and differentiation that can inhibit macrophage inflammatory cytokine production. The reduction of inflammatory cytokines could also be responsible for the reduction in T cell activation and proliferation seen during HAART treatment. Our multiplex cytokine testing data support previous reports of viral control being associated with polyfunctional T cell responses [42, 45, 46], and the cytokines identified above further clarify which specific responses are associated with control of viral replication during chronic HIV infection.
The increase in the number of Treg cells following HAART that we observed is consistent with some previous studies. Weiss et al. found an expanded number of Treg cells in HIV+ individuals receiving HAART, with a Treg cell phenotype similar to that of normal donors and cancer patients . Lim et al. also observed an increase in the number of Treg cells identified by an increase in FoxP3 mRNA expression in individuals who suppressed viremia with HAART . Kolte et al. found that both absolute Treg cell numbers and the percentage of Treg cells were increased after one and five years of receiving HAART and were associated with an increase in the thymic output of naïve Treg cells . Two other studies showed no effect of HAART on Treg cell numbers despite suppression of viral replication and immunological recovery [49, 50]. The precise mechanism of Treg cell expansion during HAART remains unknown and requires further investigation. An increase in the peripheral Treg cell pool by proliferation, increased survival of Treg cells or an increase in the thymic generation of Treg cells all could be responsible [41, 51–53]. As we saw no correlation between the number of Treg cells and HIV-specific or CMV-specific T cell responses (data not shown), it would appear that Treg cells do not strongly interfere with HIV-specific immune responses, raising the possibility of inducing these cells to ameliorate the effects of immune activation in the setting of high viral loads during chronic HIV infection.
Whilst our data mostly agree with those of Chase et al. , we did see a difference in which HIV infected group had the highest number of Treg cells. Elite controllers in the Chase et al. study had the highest number of Treg cells, whereas we saw the highest number of Treg cells in our HAART suppressed group. One possible explanation for this is confounding by age, since older individuals have higher Treg cell numbers [55–57]. In both our study and the Chase study the groups with the highest number of Treg cells were also the oldest. In the Chase study elite controllers were the oldest (median age = 54 years), while their HAART suppressed group was the youngest (median age = 46 years). In contrast, our HAART suppressed group was the oldest (median age = 53 years) and the elite controllers were younger (median age = 48 years). Multivariate analysis of our data, which controlled for confounding by age, showed that the increase in Treg cells was due to the therapy and not age. Whether this would be the case in the Chase et al. study was not addressed .
In conclusion, lower levels of HIV-specific T cell activation and in vivo proliferation combined with stronger, broader HIV-specific cytokine responses likely play a role in the control of HIV infection by elite controllers. However, elite controllers do not completely clear the virus  and may eventually lose their elite status and progress towards the development of AIDS [1, 3]. A therapeutic vaccine or immune modulation that could reduce immune activation, potentially by the induction of Treg cells, and generate a more appropriate balance of immune responses (such as those seen in elite controllers) may allow non-controllers to decrease HIV replication and delay the progression to AIDS.
This work was supported in part by grants from the NIAID Center for HIV/AIDS Vaccine Immunology (CHAVI) AI-067854, the UCSF/Gladstone Center for AIDS Research (P30 AI27763, NIAID (AI069994, AI44595), the UCSF Clinical and Translational Science Institute (UL1 RR024131-01) and American Foundation for AIDS Research (106710-40-RGRL). The following reagents were obtained through the AIDS Research and Reference Reagent Program, Division of AIDS, NAID, NIH: HIV-1 consensus B Gag peptides (#8117), HCMV pp65 peptide pool (#11549). The UCSF AIDS Specimen Bank provided cryopreserved PBMC from the SCOPE cohort. We thank P.W. Hunt and J.M. McCune for helpful discussions.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
R.E.O. planned the studies, performed experiments, analyzed data, and wrote the manuscript. J.W.H. and D.F.H. performed experiments, M.C.L. planned the studies, H.H.B. performed statistical analyses, J.N.M. provided clinical samples, M.R.K. maintained the database, S.G.D. planned the studies and provided clinical samples, and P.J.N. planned the studies and wrote the manuscript.