Immune Correlates of Natural HIV Elite Control and Simultaneous HCV Clearance—Supercontrollers

HIV-elite controllers are a minority group of HIV-infected patients with the ability to maintain undetectable HIV viremia for long time periods without antiretroviral treatment. A small group of HIV-controllers are also able to spontaneously clear the hepatitis C virus (HCV) whom we can refer to as “supercontrollers.” There are no studies that explore immune correlates looking for the mechanisms implicated in this extraordinary phenomenon. Herein, we have analyzed HCV- and HIV-specific T-cell responses, as well as T, dendritic and NK cell phenotypes. The higher HCV-specific CD4 T-cell polyfunctionality, together with a low activation and exhaustion T-cell phenotype was found in supercontrollers. In addition, the frequency of CD8 CD161high T-cells was related with HIV- and HCV-specific T-cells polyfunctionality. Interesting features regarding NK and plasmacytoid dendritic cells (pDCs) were found. The study of the supercontroller's immune response, subjects that spontaneously controls both chronic viral infections, could provide further insights into virus-specific responses needed to develop immunotherapeutic strategies in the setting of HIV cure or HCV vaccination.


INTRODUCTION
HIV-controllers are a rare group of HIV-infected patients who can maintain HIV viremia at low levels in the absence of antiretroviral treatment (1,2). Although several efforts are being made to elucidate mechanisms responsible of this phenomenon, multiple factors seem to be involved (3,4). Interestingly, the extraordinary capacity of these subjects allows them to show differential characteristic regarding hepatitis C virus (HCV) coinfection. We have previously communicated that Caucasian HIV-controllers exhibited lower HCV viral loads (VL) than non-controllers, and a different HCV genotype distribution (5). Furthermore, Sajadi et al. reported a high rate of spontaneous clearance of HCV among HIV-controllers (6). We referred as supercontrollers those patients that simultaneously, and spontaneously, control HIV and clear HCV. The existence of these patients suggests the idea of some common or additive mechanisms concur in the control of both viruses.
Virus-specific T-cell response has been extensively studied and a higher virus-specific T-cell response and polyfunctionality has been unequivocally related with HIV and HCV control (7,8). Besides, innate immunity has been also related with natural viral control. Higher NK frequency and anti-viral activity has been found in HIV-controllers (9, 10) and a high differentiated NK phenotype (defined by CD57 and HLA-C-binding killer cell immunoglobulin-like receptor (KIR) expression) has been reported in HCV spontaneous clearers (11,12). Also, myeloid dendritic cells (mDCs) and plasmacytoid dendritic cells (pDCs) frequency, phenotype and antigen-presenting capacity have been related with HIV control (13)(14)(15) and HCV clearance (16,17). However, there are no studies focused on patients which spontaneously control both viruses. For this purpose, we have selected the extraordinary group of supercontrollers to perform deep immunophenotyping of T-cells, NK and dendritic cells and to analyze HCV-and HIV-specific T-cell responses in order to find shared mechanisms of control, and to know whether supercontrollers have immunological characteristics that enable them to control more than one virus infection.

Samples and Patients
Peripheral blood samples from 50 HIV and HCV coinfected patients were collected from three Spanish Hospitals: Virgen del Rocio University Hospital (Seville), San Cecilio University Hospital (Granada) and Joan XXIII University Hospital (Tarragona). Of those, eight were HCV spontaneous clearers and HIV-elite controllers (supercontrollers, SC), 14 non-spontaneously HCV clearers HIV-elite controllers (nSC), 13 HCV spontaneous clearers non-HIV-elite controllers (SnC), and 15 non-spontaneously HCV clearers non-HIV-elite controllers (nSnC). HIV-elite controllers were defined as HIV-infected patients with VL<50 HIV-RNA copies/mL in absence of antiretroviral treatment for at least 1 year. Non-HIV-elite controllers were patients with confirmed VL > 2 × 10 3 HIV-RNA copies/mL in absence of antiretroviral treatment. HCV spontaneous clearers were defined as HCV infected patients (anti-HCV antibody positive) with <10 HCV-RNA copies/ml in absence of anti-HCV treatment. Non-spontaneously HCV clearers or HCV chronic patients were those with confirmed detectable HCV VL, without anti-HCV treatment. This study was carried out in accordance with the recommendations of "Comité de Etica de la Investigación de Centro Hospital Universitario Virgen del Rocío de Sevilla" with written informed consent from all subjects. All subjects gave written informed consent in accordance with the Declaration of Helsinki. The protocol was approved by the Comité de Etica de la Investigación de Centro Hospital Universitario Virgen del Rocío de Sevilla (2012PI/240).

Laboratory Measurements
Absolute counts of CD4+ and CD8+ T-cells were determined in fresh whole blood by using an Epic XL-MCL flow cytometer (Beckman-Coulter, Brea, CA) according to the manufacturer's instructions. The plasma HIV-1 RNA concentration was measured by using quantitative PCR (Cobas Ampliprep/Cobas TaqMan HIV-1 test; Roche Molecular Systems, Basel, Switzerland) according to the manufacturer's protocol. The detection limit for this assay was 50 HIV-1 RNA copies/mL. Hepatitis C virus (HCV) RNA was determined using an available PCR procedure kit (Cobas Amplicor; Roche Diagnostics, Barcelona, Spain) with a detection limit of 10 HCV-RNA copies/ml. Plasma samples were tested for anti-HCV antibodies using HCV-ELISA (Siemens Healthcare Diagnosis, Deerfield, IL, USA). HCV genotype was determined using a reversehybridization assay (InnoLIPA HCV II; Innogenetics, Barcelona, Spain).

Cell Stimulation
Peripheral blood mononuclear cells (PBMCs) were thawed, washed and rested for 2h in DNase I (Roche Diagnostics, Indianapolis, IN)-containing R-10 complete medium. 1.5 × 10 6 PBMCs including 10 µg/mL of brefeldin A (Sigma Chemical Co, St. Louis MO), and 0.7 µg/mL of monensin (BD Biosciences), were stimulated in vitro for 6 h with 2 µg/ml of each peptide pool of HCV NS4A, HCV NS4B, HCV NS3, and HCV Core (BEI Resources Repository, Manassas, VA, USA). HCV peptides were based on HCV 1a H77 sequence. In addition, 1.5 × 10 6 PBMCs were stimulated with 2 µg/ml of an overlapped HIV (Gag)-specific peptide pool (NIH AIDS Reagent Program [https://www.aidsreagent.org/index.cfm]). 1.5 × 10 6 unstimulated cells and cells stimulated with staphylococcal enterotoxin B (SEB) as a positive control were included in each experiment. The stimulation was performed in the presence of titrated amounts of anti-CD107a-BV605 (clone H4A3; BD Biosciences, USA) monoclonal antibody as previously described (18). T-cell specific response was defined as the frequency of cells with detectable intracellular cytokine production, after background subtraction of the unstimulated condition, after stimulation with HCV NS3, NS4A, NS4B, and Core peptide and HIV Gag peptides. For this analysis 1 × 10 6 events were acquired and a median of 4.72 × 10 5 live T-cells were gated.

Dendritic Cells Immunophenotyping
When samples were available, PBMCs were stained with zombie UV dye (Biolegend, USA) and surface stained with Lineage

Statistical Analysis
Differences between unpaired samples were analyzed by Mann-Whitney U-tests. Correlations between variables were assessed using the Spearman rank test. All differences with a P < 0.05 were considered statistically significant. Statistical analyses were performed by using Statistical Package for the Social Sciences software (SPSS 22.0; SPSS Inc., Chicago, IL). Graphs were generated with Prism, version 5.0 (GraphPad Software, Inc.) and R Statistical Software (Foundation for Statistical Computing, Vienna, Austria) (19). Polyfunctionality was defined as the percentage of lymphocytes producing multiple cytokines. Polyfunctionality pie charts were constructed using Pestle version 1.6.2 and Spice version 5.2 (provided by M. Roederer, NIH, Bethesda, MD) and was quantified with the polyfunctionality index algorithm (20) employing the 0.1.2 beta version of the FunkyCells Boolean Dataminer software provided by Martin Larson (INSERM U1135, Paris, France). Differences between unpaired distributions in pie charts were analyzed by Permutation test, Spice version 5.2. In this test a P < 0.05 was considered statistically significant.

Improved Polyfunctionality of HCV-and HIV-Specific CD4 T-Cells in Supercontrollers
We analyzed the T-cell specific response defined as the frequency of cells (after background subtraction of the unstimulated condition) with detectable intracellular cytokine production after stimulation with HCV NS3, NS4A, NS4B, and Core peptide and HIV Gag peptides. CD4 HCV and HIV specific T-cell gating strategies are shown in Figure 1A. T-cells from SC exhibited the highest HCV-specific CD4 T-cell polyfunctionality, in terms of simultaneously production of IFNγ, IL2, and TNFα ( Figure 1B, red portion of pie chart). Although the overall distribution of cytokine production was not statistically significant, when we analyzed the proportion of CD4 T-cells expressing IFNγ, IL2, and TNFα (3 functions) at the same time in response to HCV, the highest levels were shown in SC group ( Figure 1C). Interestingly, this cell subset was higher in SC than SnC, the group of patients that spontaneously clear the HCV. When the HIV-specific CD4 T-cell response was analyzed, a higher polyfunctionality was present in HIV-controllers, independently of HCV clearance (SC and nSC groups) ( Figure 1D, red portions). Also, the frequency of HIV specific CD4 T-cells producing the combination IFNγ+IL2+TNFα-was higher in the HIV-controller groups (SC and nSC) respect non-HIV-controller groups (SnC and nSnC) ( Figure 1E). The CD4 and CD8 T-cell polyfunctionality index (pINDEX) in response to HCV positively correlated with the CD8 T-cell pINDEX in response to HIV (r = 0.364, p = 0.019; and r = 0.441, p = 0.004, respectively) ( Figure 1F), reinforcing the idea of the better T-cell response to HCV corresponded with a better T-cell response to HIV. When CD4 T-cell cytotoxicity was analyzed, assessed by CD107a and Granzyme B production in any combination with other cytokine, no differences were observed in response to HCV and HIV stimulation (data not shown). Interestingly, controllers (SC and nSC) exhibited higher CD107a+GranzymeB-IFNg-IL2-TNFa-CD4 HIV-specific T-cells than non-controllers (Supplementary Figure 1). HCVand HIV-specific CD8 T-cell response was also analyzed (Supplementary Figure 2). Gating strategies are summarized in Supplementary Figure 2A. No significant differences in HCV-and HIV-specific CD8 T-cell response among groups  Frontiers in Immunology | www.frontiersin.org were observed (Supplementary Figure 2B). We observed differences in some cytokine combination, such as higher proportion of IFNγ+IL2-TNFα+ CD8 HCV-specific Tcells of SC compared with SnC and nSnC (p = 0.025 and 0.049, respectively) (Supplementary Figure 2C). Interestingly, we observed a trend to higher CD8 HIV-specific T-cells polyfunctionality (Supplementary Figure 2D) and higher proportion of IFNγ+IL2-TNFα+ and IFNγ+IL2+TNFα+ CD8 Figure 2E) in nSC group than SnC and nSnC.

Low Levels of T-Cell Exhaustion in HIV-Controllers Independently of HCV Clearance
To determine whether HCV-specific CD4 T-cell polyfunctionality was associated to lower T-cell exhaustion we quantified the expression of the exhaustion markers Lag3, PD1, TIGIT, and Tim3 in CD4 and CD8 memory T-cell subsets (gating strategy is shown in Supplementary Figure 3). The "multiple exhausted phenotype" (simultaneous expressions of three or more of the analyzed exhaustion markers) was represented by pie charts. SC exhibited low multiple exhausted phenotypes in both CD4 and CD8 T-cells, in all memory subpopulations. These levels were similar to the other HIV-controller group, nSC (Figures 2A,B). This trend was observed when specific combinations of exhaustion markers were analyzed in each group (Figure 2C). Regarding CD8 T-cell subsets similar patterns to CD4 T-cell subsets were shown when exhaustion marker combinations were analyzed ( Figure 2D).

T-Cell Activation and CD161 high CD8 T-Cell Correlated With HCV and HIV T-Cell Response
T-cell activation was assessed by HLA-DR expression both in CD4 and CD8 T-cell subpopulations. SC presented lower CD4 HLA-DR+ T-cell levels ( Figure 3A) and CD8 HLA-DR T-cells ( Figure 3B) in every subpopulation studied and, similarly to Tcell exhaustion, at the same levels than the other HIV controller group (nSC). Interestingly, this activated phenotype of CD8 and CD4 central and effector memory T-cells showed a common pattern of negative correlations with CD4 and CD8 HCVand HIV-specific T-cell responses ( Figure 3C). In addition, the chemokine receptor CXCR3, involved in T-cell trafficking to inflamed tissue (21) was also analyzed in CD4 and CD8 Tcells. We did not found differences in CXCR3 expression in any memory subset, neither in CD4 nor CD8 T-cell (data not shown). Unexpectedly, we observed lower levels of CXCR3+ naïve CD4 and CD8 T-cells in HIV-controllers (SC and nSC) compared with non-HIV controllers (SnC and nSnC) (Figures 4A,B).
The expression of the C-type lectin CD161 was also analyzed in CD4 and CD8 T-cells. No significant differences in CD161 expression of CD4 T-cells were found among groups (data not shown). Higher frequency of CD161 high CD8 T-cell was presented in HIV-controllers, SC and nSC, compared with non-HIV controllers, SnC and nSnC ( Figure 5A). Notably, we observed higher frequency of CD161 high CD8 effector memory T-cells in SnC than nSnC. This peculiar CD8 T-cell phenotype from central and effector memory subsets strongly correlated with HCV-and HIV-specific CD4 and CD8 T-cell responses ( Figure 5B).

Innate Immune Cells Involvement in HIV Control and HCV Spontaneous Clearance
We then ought to analyze the association of innate immunity in the supercontroller phenotype. The gating strategies for different dendritic cell subsets are summarized in Figure 6A. As expected, higher frequency of pDCs were found in HIV-controllers (SC and nSC) compared with non-controllers (SnC and nSnC), independently of HCV spontaneous clearance (Figure 6B). No differences in mDCs were found among groups (data not shown). Interestingly, there was a trend to show higher frequency of pDCs expressing the lymph node homing marker CCR7 in SC ( Figure 6C). Furthermore, positive correlations were found between frequency of pDCs and HCV-specific CD4 T-cell pINDEX (r = 0.802, p < 0.001) ( Figure 6D) and frequency of IFNγ+IL2+TNFα+ CD4 HCV-specific T-cells (the combination shown in Figure 1C) (r = 0.571, p = 0.013) (Figure 6E). Also, positive correlations were found between pDCs levels and HCVspecific CD8 T-cell pINDEX (r = 0.585, p = 0.011) ( Figure 6F) and frequency of IFNγ+IL2-TNFα+ HIV-specific CD8 T-cells (r = 0.515, p = 0.029) (Figure 6G).
Regarding NK cells, we also observed higher levels of NK CD56 dim cells among HIV-controller group, SC and nSC  ( Figure 7A). No differences of NK CD56 high cells were found among groups (Figure 7B). In line with the results for CD4 and CD8 T-cell activation, HIV controller groups, SC and nSC, presented lower levels of NK CD56 dim cells expressing HLA-DR (Figure 7C), likewise some exhaustion markers such as TIGIT ( Figure 7D) and Lag3 (Figure 7E), and CXCR6

DISCUSSION
HIV-elite controllers that spontaneously clear HCV (supercontrollers) are extremely difficult to investigate, as they are very uncommon. Considering that about 1% of patients are able to control HIV (22), and 10-20% of infected patients spontaneously clear the HCV (23), the proportion of patients that control both viruses is very low (we estimates <0.2%). Herein, we have described for the first time the immunological features of these abovementioned supercontrollers, consisting in high CD4 HCV-specific T-cell polyfunctionality together with low exhausted and activated phenotype of T-and innate immunity cells.
Virus specific CD4 T-cells have an important role in controlling viral infections, not only by helping the CD8 Tcell cytotoxic activity and protective antibodies production by B cells, but also exercising antiviral functions (24,25). We have found higher HCV-specific CD4 T-cell polyfunctionality in subjects that control HIV and clear HCV at the same time. That was in line with the only study of HCV T-cell response within HIV-controllers, in which the authors found a strong Th1 response against HCV (26). This study was performed with chronic HCV patients, not with HCV clearers, but an inverse correlation between Th1 frequency specific for HCV core protein and HCV VL was found (26). That emphasized the importance of CD4 T-cells in HCV control within HIVcontrollers.
Interestingly, a higher, broadly and long lasting HCV-specific CD4 T-cell response has been associated with spontaneous clearance of HCV compared with chronic HCV patients (27)(28)(29)(30)(31)(32). Also remarkably, these studies found broad CD4 T-cell responses to HCV non-structural protein NS3 and NS4 in HCV clearers, peptides that we have included in our study in the peptide pool for stimulations. Furthermore, Shulze et al. demonstrated that peptide recognized by CD4 T-cells from HCV spontaneous clearers were highly promiscuous, which can be restricted by multiple MHC-II molecules (32). Curiously, same phenomenon seems to occur with Gag peptide restriction in HIV-controllers, which same peptide can be restricted by up to five HLA-DR allomorphs to one Tcell receptor (33). Actually, several studies have associated some HLA-DRB1 alleles (34)(35)(36) and HLA-DQB1 (34,37,38) with HCV spontaneous clearance. Unfortunately, because of samples unavailability after flow cytometry experiments we were unable to test HLA-II typing. However, due to the small sample size this important immune correlate of virus control may be confirmed in larger cohorts of supercontrollers.
On the other hand, CD4 T-cell response were also related with HIV spontaneous control (39). It is known that HIV-controllers exhibit more vigorous, proliferative and polyfunctional CD4 HIV-specific T-cell responses than non-controller patients, even after the antiretroviral treatment onset (40)(41)(42)(43)(44). Furthermore, Potter et al. showed that CD4 T-cells from HIV-elite controllers were enriched of central memory, which expressed higher levels of CD127. This fact could contribute to a long-term antiviral memory activity in these patients (44), probably including HIV-and HCV-specific response. In our experiments, the CD4 HIV-specific response was similar in both HIV controller groups.
We did not observe significant differences in polyfunctionality of HCV-specific CD8 T-cell response among groups. It has been demonstrated that HCV-specific CD8 T-cell in chronic patients is difficult to find in peripheral blood, due to liver accumulation of these cells (45). Of note, although both SC and SnC spontaneously cleared the HCV, a higher IFNγ and TNFα HCV-specific CD8 T-cell production were found in patients that control HIV and clear HCV. However, we highlight the importance of the CD4 HCV-specific polyfunctionality found in SC. First, because a HCV-specific CD8 T-cell activity is known not to be enough to mount a protective response, as it has been demonstrate in the setting of HCV vaccines (46), probably because of the high scape mutation rate in this virus (47,48). And second, because the importance of developing adequate CD4 T-cell response over a CD8 T-cell has been demonstrated in an elegant experiment with the HCV chimpanzee model in which Grakoui et al. depleted CD4 T-cells from animals that previously cleared the HCV, and then they reinfected them with HCV thus, demonstrating that despite of having memory HCV-specific CD8 T-cells, it was not enough to control the virus scape mutations onset (47).
T-cell exhaustion and activation are hallmarks of the HIV infection (49,50). Interestingly, SC exhibited low levels of both measurements. Interestingly, CD4+ T-cell expressing PD1, Lag-3 and TIGIT has been strongly correlated with HIV reservoir size and HIV persistence (50). Maybe, this extraordinary group of controllers is enriched in patients with in addition to HCV spontaneous eradication, have minimal amount of integrated HIV. Unfortunately, we could not measure HIV reservoir because of lack of samples. Further analyses in this regard should be done. In this sense, Banga et al. argue that CD4 T-cell expressing CXCR3 represent the major blood compartment with HIV proviral population (51). No differences in CXCR3 expression in CD4 and CD8 T-cell memory subsets were found, but we observed lower frequency of naïve CD4 and CD8 T-cell expressing CXCR3 in HIVcontrollers. This molecule is involved in T-cell homing to inflamed peripheral tissue after maturation (21). We did not expect to find CXCR3 expression in naive T-cells but although the expression were low (<5% in HIV-controller groups), we observed differences with the non-controllers groups. We do not have a plausible explanation for this fact; maybe this is putative marker of lower inflammation/activation levels in HIVcontrollers.
Interestingly, we found higher frequency of CD161 high CD8 T-cells in HIV-controllers, with a trend to high levels in SC. The significance of CD161 expression in CD8 T-cell is controversial (52). It has been suggested that CD161 high CD8 T-cell population represents a memory stem cell subset (53), whereas other authors argue that these cells are IL-17 producing CD8 Tcells (54,55). Our results suggest that this CD8 T-cell subset could be involved somehow with the antiviral response (56), as they positively correlate with HIV-and HCV-specific T-cell responses.
Innate immunity also has a major role on viral control in both HIV and HCV (57,58). We have previously communicated that pDCs are related with HIV spontaneous control (13,14), and other authors showed higher responsiveness of pDCs from HCV spontaneous clearers (16). In this work we found higher pDCs frequency in HIV-controllers compared with non-controllers and interestingly, a trend to higher CCR7 expression of pDCs in SC. CCR7 is a homing molecule that enable pDC migrate to lymph nodes (59) and prime CD4 and CD8 T-cells (60). This is in agreement with the correlation found between pDCs and HIV-and HCV-specific T-cell levels. On the other hand, NK cells also play an important role in HIV control (9, 10) and HCV clearance (11,12). Similarly to T-cells, HIV-controllers showed lower activation and exhaustion of CD56 dim NK cells compare with non-controllers, the high cytotoxic NK subset (61).
In conclusion, CD4 T-cell response seems to be the most differential feature in SC. Furthermore, this extraordinary group of subjects exhibited low levels of T-cell exhaustion and activation. Other T-cell subsets, as CD161 high were interestingly related with HIV and HCV-specific T-cell responses, as well as pDCs. Further analyses with a large cohort of SC must be done to deeply analyze HIV-and HCV-specific T-cell phenotype and the correlates with the control of both viruses in relation with other players of adaptive and innate immunity. The study of these patients could serve as a model for the development of therapeutic strategies aimed to enhance antiviral responses.