Enhanced IFN-γ, but not IL-2, response to Mycobacterium tuberculosis antigens in HIV/latent TB co-infected patients on long-term HAART

Tuberculosis (TB) is the one of the leading deadly diseases worldwide with over 90% of deaths occurring in developing countries [1]. TB is caused by direct exposure to Mycobacterium tuberculosis (M. tuberculosis), and/or reactivation of latent TB infection. Once infected with M. tuberculosis, only about 5–10% of people directly develop active TB while 90–95% remain latently infected [2, 3]. In 2014, approximately 1.7 billion people were latently infected with M. tuberculosis globally, low-and middle-income countries accounting for around 80% of the prevalence [4]. Immunocompetent individuals control the infection by containing the mycobacteria in an inactive or latent state. Both the innate and adaptive arms of the immune system are involved in a collaborative way to control infection with M. tuberculosis and subsequent disease. Various T cells produce potent cytokines and the interaction of these cells with infected macrophages are crucial for anti-mycobacterial protective responses [2, 3, 5‐7]. People with latent TB infection have only 5–10% lifetime risk of reactivation [8]. However, following acquisition of HIV infection, the risk of reactivation of latent TB infection to active TB increases to 5–10% each year [3, 9]. This high rate of active TB development might be directly related to HIV-derived weakened host cell-mediated immunity in general, and impaired M. tuberculosis-specific immune responses in particular. M. tuberculosis-specific production of interferon-(IFN)-γ, interleukin (IL)-2 and Tumor necrosis factor (TNF)-α by T cells contribute substantially to elicit effective immunity to prevent reactivation of latent TB infection [7, 10‐12].

Highly active anti-retroviral therapy (HAART) reduces the incidence rate of TB in people living with HIV. A meta-analysis of various studies has shown that HAART is associated with a 67% reduction (range: 61–73%) in TB incidence [13‐16]. Reduction of TB incidence rate after initiation of HAART could likely be due to continuous CD4⁺ T cell recovery [14, 17] accompanied by restoration of functional anti-mycobacterial immunity. This is supported by studies which demonstrated initiation of HAART leads to rapid functional recovery of mycobacteria-specific immune responses, with increasing lymphocyte proliferation and secretion of IFN-γ by peripheral blood mononuclear cells (PBMCs) stimulated ex vivo with mycobacterial antigens [18‐20]. However, most investigators have evaluated responses within 1 year of HAART initiation. Immune reconstitution occurs in two or more phases over time. CD4⁺ T cells recovery observed in the first year of HAART is rapid and prominent, and primarily a result of an immediate redistribution of memory CD4⁺ T-cells from the lymphoid tissues into the blood streams [21‐23]. After a year of therapy, this phase is followed by a slow recovery of CD4⁺ T-cells and is mainly due to de novo production of new naïve T lymphocytes and reduced apoptosis [22, 24, 25]. Thus, the former process could explain enhanced immune reactivity to M. tuberculosis observed in the previous studies within the first year of HAART. In contrast, the functional immune response to M. tuberculosis in HIV/latent TB co-infected patients after prolonged HAART therapy has not been well studied. As a result, questions still remain regarding the extent and nature of the anti-mycobacterial immune reconstitution in the long-term of HAART. We therefore aimed at investigating the durability of HAART-driven anti-mycobacterial immune responses with the hypothesis that long-term HAART would still augment protective immune responses against M. tuberculosis in HIV/latent TB co-infected patients. In this study we observed an increased, but only partly, M. tuberculosis-specific functional immune responses in HIV/latent TB co-infected patients who received HAART for more than a year as compared to HAART-naïve patients.

The incidence rate of TB in HIV-infected patients who initiate HAART is significantly decreased compared to patients without therapy, but even on antiretroviral therapy, the risk for TB is higher than HIV uninfected individuals [14, 32]. Clarification of the quality and quantity of the incomplete immune recovery and developing means to improve clinical outcomes thus remain an important priority in the HIV field. HAART-driven immune restoration is thought to occur in two phases; the first and fast phase is due to immediate redistribution of memory CD4⁺ T cells from lymphoid tissues to peripheral blood, but does not involve significant net changes in CD4⁺ T cell number, followed by slow gradual immune recovery as a result of de novo production of naïve or memory T lymphocytes and reduced apoptosis [21‐25, 33, 34]. Most studies that have assessed restoration of M. tuberculosis-specific responses have typically involved cohort studies evaluating immune reconstitution before and after up to one year of therapy, which would be predicted to address changes primarily due to the fast redistribution phase. We carried out this cross-sectional study to enumerate TB-specific immune responses among patients without HAART and with a range of duration on therapy from 1.5 up to 9 years. We found that PPD and ESAT-6-specific IFN-γ responses were significantly higher in HIV-infected patients who received HAART for more than a year compared to the pre-HAART participants, consistent with the observations of cohort studies evaluating responses within the first 6 to 12 months of HAART [18, 35, 36]. Hence, this shows that anti-mycobacterial IFN-γ responses are restored after initiation of HAART and maintained for prolonged period of the therapy. In contrast to the IFN-γ response, however there was no significant difference between HAART-naive and-treated patients in M. tuberculosis-specific IL-2 production. These results contrast with studies done on patient cohorts with less than one year of HAART [37, 38]. Notably, we observed that M. tuberculosis-specific IL-2 responses were inversely correlated with duration of HAART, and not related to either CD4⁺ T cell count or CD4⁺ T cell gain after HAART. These findings raise the possibility that IL-2 responses may be restored only transiently during the initial redistribution phase, which would accommodate previous studies, but are unable to be maintained during long-term antiretroviral therapy in the secondary phase.

T cell subpopulations including CD4⁺ and CD8⁺ T cells are known to produce IFN-γ and/or IL-2. These T cells can be divided into distinct populations of effector cells, effector memory and central memory cells [39]. Effector T cells show immediate effector function of secreting IFN-γ whereas effector memory T cells produces both IFN-γ and IL-2 cytokines [40, 41], but in advanced HIV disease tend to produce relatively more IFN-γ [42, 43]. On the other hand, central memory T cells typically produce predominantly IL-2 [40, 41]. Presumably the differences we observed in IL-2 responses between the relatively early HAART and prolonged HAART reflect differences in the frequencies of these T cells subsets. However, since our studies were done on unfractionated cells, we could not enumerate the relative contribution of these subsets, nor the frequency of IFN-γ/IL-2 co-producing cells which may have impacted the observed results. Nonetheless, the fact that we did not observe correlations between the cytokine producing cells and CD4⁺ T cells counts, between IFN-γ and IL-2 responses in patients are findings consistent with the possibility that the IFN-γ and IL-2 responses measured here are being produced by different subsets of T cells. Thus, we would speculate that some subsets are preferentially induced or maintained at the expense of others in patients undergoing immune restoration after HAART. An example of how this might occur can be ascertained from known differential requirements of CD8⁺ and CD4⁺ T cell subsets for IL-7 and IL-15. In general, memory CD8⁺ T cells appear more dependent on IL-15 than IL-7, whereas memory CD4⁺ T cells may be more dependent on IL-7 [40, 41]. Consistent with these findings, CD4⁺ T cell counts are improved with IL-7 supplementation in humans on HAART [44] and in simian immunodeficiency virus (SIV)-infected and treated macaques, but not with IL-15 therapy in macaques [45, 46]. The findings of reduced IL-7 producing capacity of lymph node stromal cells associated with fibrosis in the para-cortical T cell zone [47], as well as improvements with anti-fibrotic therapy [48] contribute to the view that IL-7 may be particularly important in the maintenance of adequate memory CD4⁺ T cells or subsets thereof involved in IL-2 production. Deficiency of such cells could lead to suboptimal IL-2 production as we have observed here. Alternatively, failure to attain or maintain adequate IL-2 production may also reflect continued high rates of apoptosis of CD4⁺ T cells or their subsets [24, 25, 49]. Hence, further studies are required to investigate whether IL-7 and other factors are involved in restoration of IL-2 responses of M. tuberculosis-specific memory T cells in HIV infected patients receiving HAART.

Regardless of mechanisms of immune memory restoration or homeostatic maintenance, it is likely that M. tuberculosis-specific T cell production of both IFN-γ and IL-2 is important for optimal prevention of reactivation of latent TB infection. The necessity for IFN-γ is aptly demonstrated by the increased risk of TB and other mycobacterial diseases in individuals with genetic defects in the production or action of IFN-γ [50], yet IFN-γ is clearly not sufficient because most TB patients readily produce IFN-γ. IL-2 promotes the expansion of the antigen specific T cells and likely play an indispensable role in control of the M. tuberculosis infection [11, 31, 51]. Consistent with the role of IL-2 in proliferation, we confirmed on a subset of the patients in this study that M. tuberculosis-specific IL-2 response correlated positively with CD4⁺ T cell proliferation ex vivo. The negative correlation of IL-2 production with duration of HAART suggests that proliferative capacity of M. tuberculosis-specific CD4⁺ T cells in HIV-infected patients could be diminished over time after antiretroviral therapy regardless of the gain in CD4⁺ T cell count. Hence, a failure to maintain adequate levels of M. tuberculosis-specific IL-2 secreting cells in patients on long-term HAART, predicts that such patients, despite circulating levels of IFN-γ producing cells, may be impaired in the development or augmentation of adaptive immune responses to either reactivated latent TB or new TB infections.

Despite the evidence of the importance of IL-2 for adequate immune responses, a recent prospective study in South Africa evaluating the ability of cytokine levels to predict risk of TB observed that TB-specific IL-2 detected in vitro was actually higher among patients who eventually developed disease [52]. This observation is not consistent with what we would have predicted although we did not compare cytokine production in the latent TB infection with active TB in the present study. These findings serve as a reminder that, like IFN-γ, levels of IL-2 and other cytokines, even if presumed necessary for protection, do not necessarily predict disease. It further underscores the importance of prospective studies to correlate immune responses with actual TB disease.

The cross-sectional approach we have utilized here has the advantage that patients on long-term HAART can be more conveniently recruited and evaluated, but the limitation is that, unlike prospective studies, no baseline values prior to HAART were tested. Thus, we cannot rule out the possibilities that prior to HAART, those patients on long-term HAART had lower levels of IL-2 than those of early HAART subjects, and that either IL-2 increased proportionally with IFN-γ in the HAART period similarly in both groups, or that IL-2 never changed in the long-term HAART group. In addition to baseline testing, an IL-2 assay of greater sensitivity than the currently available are required since the levels of most participants almost approached the limit of detection. In any case, it is clear that definitive testing of the hypothesis that IL-2 is only transiently increased during the fast redistribution phase of immune reconstitution, but not sustained during the long-term HAART, requires both a prospective study evaluating patients at baseline, and a much longer follow-up period than has previously been employed, and one which, like the aforementioned South African study evaluates disease outcomes.

In summary, this study has shown an increased anti-mycobacterial immune response in HIV/latent TB co-infected patients with long-term HAART as evaluated by IFN-γ response. However, the failure of M. tuberculosis-specific IL-2 responses to improve may indicate that patients on therapy remain only partly competent to mount and augment an effective M. tuberculosis-specific response. This finding is in keeping with the observation that patients on long-term HAART remain at risk, albeit reduced, for new opportunistic infections, and require careful monitoring and follow-up. These findings also call for more prospective studies which evaluate subsets of T cells in both pre-HAART and over a much longer time period of the therapy to more accurately reveal the immune potential and ability to protect against disease of patients both during the short-term and long-term phases of immune reconstitution.