|Home | About | Journals | Submit | Contact Us | Français|
The acid-fast bacillus Mycobacterium tuberculosis (MTB) is often the first manifestation of AIDS in HIV infected patients. This study was conducted to better understand the mechanism underlying MTB-specific pathogenicity early in HIV infection.
MTB-specific T Helper 1 (TH1) cells were studied in HIV negative (n=114) and chronically HIV infected (n=68) Tanzanian subjects usinEarly Secreted Antigenic Target 6 (ESAT6) protein or Tuberculin (PPD) by Interferon gamma (IFNγ) ELISPOT and intracellular cytokine staining. In a longitudinal study the effect of acute HIV infection on MTB-specific TH1 cells was determined by polychromatic flow cytometric analysis in 5 subjects with latent MTB infection, who became infected with HIV.
In Tuberculosis (TB) asymptomatic subjects, chronic HIV infection was associated with a decreased frequency of responders with detectable MTB-specific TH1 cells (p-value = 0.0003) that was not observed in subjects with active TB. Acute HIV infection induced a rapid depletion of MTB-specific TH1 cells in 4 subjects who remained TB asymptomatic, but were stable in subjects who remained HIV negative (p<0.01).
Together these data suggest a mechanism of rapid MTB-specific TH1 cell depletion that may contribute to the early onset of TB in latently MTB infected individuals who become HIV infected.
HIV infection is characterized by the progressive depletion of CD4 T cells eventually leading to the Acquired Immunodeficiency Syndrome (AIDS) as defined by the onset of opportunistic infections. The mechanism of pathogenesis underlying HIV associated immune damage may differ between the acute and chronic phase of infection. It is widely accepted that ongoing viral replication and virus-induced cell death are responsible for the massive depletion of memory CD4 T cells from mucosal sites during acute SIV and HIV infection [1-5]. Yet, despite this rapid depletion of memory CD4 T cells early during HIV infection, most opportunistic infections typically cause complications only after extended time periods of HIV disease progression. Few pathogens cause disease early after HIV infection. One such pathogen of high clinical relevance is the acid-fast bacillus Mycobacterium tuberculosis (MTB). During the first year of HIV/MTB co-infection the risk of developing active Tuberculosis (TB) increases dramatically [6, 7]. TB disease occurs in HIV infected persons at all CD4 T lymphocyte counts  and especially in the developing world pulmonary TB frequently is the first manifestation of AIDS , suggesting that MTB-specific pathology in HIV infected individuals may differ from that observed for most other opportunistic infections.
MTB commonly causes latent infections of the lung that are tightly controlled by the MTB-specific cellular immune response in healthy individuals, but result in disease during periods of immunosuppression. Interferon gamma (IFNγ) secreting CD4 T Helper 1 cells (TH1 cells) can activate MTB infected macrophages and contribute to the containment of the intraphagosomal pathogen [8, 9] and hence are important in the control of MTB infection. Optimal activation of infected macrophages also may require involvement of costimulatory cell surface proteins  and therefore is dependent on cellular contact with MTB-specific lymphocytes. Interestingly, the same interactions also contribute to maximal HIV-1 replication inside alveolar macrophages  and may lead to efficient transmission of HIV from MTB/HIV coinfected macrophages to MTB-specific CD4 T cells. However, despite the fact that MTB-specific immunity apparently is greatly suppressed in individuals coinfected with both pathogens, the affects of HIV infection on MTB-specific CD4 T cells are not well understood. MTB-specific CD4 T cell responses are present in HIV-infected subjects with pulmonary TB [11, 12], but at the same time HIV infection is associated with a reduced Delayed Type Hypersensitivity reaction after the Tuberculin skin test (TST), suggesting that HIV does affect Mycobacteria-specific TH1 cell responses in vivo.
The Region of Difference 1 antigens (RD1), Early Secreted Antigenic Target 6 (ESAT6) and Culture Filtrate Protein 10 (CFP10) are more specific for MTB than Tuberculin (PPD) and have more recently been used to identify MTB-specific cellular immune responses [12-16]. The antigens ESAT6 and CFP10, especially in combination with PPD, are therefore suitable to study MTB-specific CD4 T cell responses and the effect of HIV on these responses in greater detail.
To investigate the mechanisms underlying the early onset of pulmonary tuberculosis often observed after HIV infection, we have first studied the effect of chronic HIV infection on MTB-specific TH1 cell responses in TB asymptomatic subjects and subjects with active pulmonary TB. We then further dissected events early after HIV infection by following up 5 women from a commercial sex workers cohort in Tanzania, who were latently infected with MTB before they became infected with HIV. We analyzed the MTB-specific CD4 T cell response during and up to 12 months after HIV seroconversion and finally determined clinical outcome in all 5 subjects of the MTB/HIV coinfection by a final interview that took place 3 years after the last study follow up.
For the analysis of MTB-specific TH1 cell responses volunteers were recruited from three different ongoing studies that have been described elsewhere [17-19]. This sub-study received a separate ethical clearance at the local and national IRB and in case additional specimens were taken a new consent form was signed with the volunteer. For the cross-sectional analysis of MTB-specific TH1 cell responses in fourteen acid-fast bacilli (AFB) smear positive patients with clinical symptoms of pulmonary TB (productive cough >4 weeks, night sweats, weight loss, loss of appetite) were recruited from an ongoing TB diagnostic trial . During the study, all participants received health care for acute medical problems. Patients with diagnosed TB were referred for TB treatment according to WHO standards. During the course of this study all individuals were anti-retroviral naïve. TB was diagnosed using 3 independent sputum samples for AFB-staining and clinical diagnosis. Latent infection with MTB was defined by cellular responses to RD1 antigens in the absence of clinical symptoms suggestive of TB. HIV-1 status was determined using two diagnostic HIV enzyme-linked immunoassay (ELISA) tests (Enzygnost® Anti HIV1/2 Plus, Dade Behring, Liederbach, Germany; and Determine™ HIV 1/2, Abbott, Wiesbaden, Germany). CD4 counts were determined from fresh whole blood using the BD-Multitest kit (Becton Dickinson).
ESAT6 and CFP10 (Lionex, Braunschweig, Germany) and Purified Protein Derivative for in vitro use (PPD, Tuberculin, Statens Serum Institute, Copenhagen, Denmark) were used at a final concentration of 10μg/ml. Peptides overlapping by 11 were designed for ESAT6 (AF420491.1) and CFP10 (AAC83445) using PeptGen Peptide generator from the HIV Molecular Immunology Database.
Freshly isolated Peripheral Blood Mononuclear Cells (PBMC) were screened for responses specific for recombinant ESAT6 and PPD by stimulation of 2.5 × 105 PBMC/well in duplicates overnight. The assay was performed as previously described . Responses with at least 20 Spot-forming cells (SFC)/106 PBMC and at least 3-fold the negative control were scored as positive. Discordant results, when only one of the duplet wells was scored as positive, were excluded from analysis.
Following antibodies were used: CD3-Cy7APC, IFNγ–FITC, CCR5-Cy7PE from BD Biosciences; CD27-Cy5PE and CD45RO–Texas red–PE (TRPE) from Beckman Coulter; CD4-PECy5.5 from Caltag. The following antibodies were conjugated inhouse according to standard protocols (http://drmr.com/abcon/index.html): CD8–quantum dot (QD) 655 and TNFα–Alexa680.
Cell stimulation and staining were performed using a modification of the method described previously . After 6h stimulation, PBMC were washed once with PBS and stained with Vivid (Molecular probes, ) and anti-CCR5 for 10 minutes at room temperature (RT) in the dark. After washing, surface proteins were stained for 20 min. The cells were then washed again and permeabilized using the cytofix/cytoperm kit (BD Biosciences). After intracellular staining cells were washed and fixed with 1% paraformaldehyde. Cells were analyzed with a modified flow cytometer (LSRII; BD Immunocytometry Systems). Electronic compensation was conducted with antibody capture beads (BD Biosciences) stained separately with individual antibodies used in the test samples. Data analysis was performed using FlowJo version 8.2 (TreeStar).
Data analyses were carried out using GraphPad Prism software. Tests used for statistical analysis are mentioned in the figure legends.
A total of 182 subjects were tested in a cross-sectional study with an IFNγ ELISPOT assay for recognition of recombinant ESAT6 protein. Table 1 summarizes CD4 counts, detected MTB-specific CD4 T cells and the results for AFB staining for HIV negative and TB asymptomatic subjects (HIV-/TB-), HIV positive and TB asymptomatic subjects (HIV+/TB-), HIV negative subjects with pulmonary TB (HIV-/TB+) and HIV positive subjects with pulmonary TB (HIV+/TB+).
HIV infection was associated with a significant decrease in the frequency of detectable IFNγ-ELISPOT responses to ESAT6 (p-value = 0.0003, Fisher’s Exact test) in patients with no clinical signs of TB. While 47% (50 of 106, 5 subjects were excluded due to discordant results between duplet wells) of the HIV-/TB-subjects responded to ESAT6, only 19% (10 of 54, 3 subjects were excluded due to discordant results between duplet wells) of HIV+/TB-subjects had detectable responses. Intracellular cytokine staining (ICS) demonstrated that detectable IFNγ responses were CD4 T cell mediated (figure 1). When the overall median magnitude of all 106 HIV-subjects (18 SFC/106PBMC) was compared with the 54 HIV+subjects (2 SFC/106PBMC), HIV infection was associated with a nine-fold reduction of MTB-specific CD4 T cells (p-value < 0.01, figure 2A). PPD-specific responses followed an identical trend (p<0.0001, figure 2C). Contrary to our expectation, there was no linear correlation between MTB-specific CD4 T cells and the total CD4 T cell count (figure 2B, D). Together these results demonstrate that at least half of the studied population from Mbeya region was probably latently infected with MTB and that HIV infection is greatly decreasing the MTB-specific TH1 cell response in subjects with “latent” MTB infection.
In order to study differences between latently and actively infected individuals we measured in a second step the ESAT6-specific TH1 cells in HIV- and HIV+ individuals with AFB smear positive TB. In line with previous reports  10 of 11 (91%) HIV+/TB+patients (Figure 2A) had detectable ESAT6 responses with a median of 118 SFC/106 PBMC. In agreement with the previous observation, detection of MTB-specific CD4 T cell responses did not correspond to the total CD4 count. 3 of 3 (100%) HIV-/TB+ patients responded to ESAT6. Together, our results suggest that MTB-specific TH1 cells present during latent infection are depleted after HIV infection and that only after MTB reactivation or de novo infection is associated with expansion of these cells, independent of the total CD4 T cell count. We therefore hypothesize that detection of MTB-specific TH1 cells in HIV positive individuals is a marker for exposure of the immune system with new MTB antigen and therefore indicates at least transient reactivation with MTB.
Most newly transmitted HIV viruses infect cells that express surface viral receptors CD4 and CCR5 . To study whether MTB-specific TH1 cells are potentially susceptible to viral infection during primary HIV infection, we next analyzed CCR5 expression in 9 HIV- subjects with latent MTB infection and compared these (gated in figure 3A) to different CD4 T cell subsets as defined by their expression of the T cell memory markers CD27 and CD45RO (figure 3C). Naïve CD4 T cells (CD27+CD45RO-) that do not express CCR5 (blue line in figure 3B) were used for the determination of cut offs. As expected, a higher fraction of CD27-CD45RO+ CD4 T cells (median 50.7%) expressed CCR5 as compared with CD27+CD45RO+(median 18.7%). A median of 58.8% MTB-specific TH1 cells (range: 29.4% - 69.3%) expressed CCR5, constituting a 2-fold increase compared to total memory CD4 T cells (median 27.9%). Also the level of CCR5 cell surface expression was significantly increased (p<0.01). Together these data demonstrate that a large fraction of MTB-specific TH1 cells from subjects with latent MTB infection express the cellular receptors required for HIV entry of most newly transmitted viruses.
Rapid depletion of memory CD4 T cells is a hallmark of acute SIV and HIV infection [1-5]. Despite this significant decrease of memory CD4 T cells, Candida albicans-, CMV-and tetanus toxoid-specific CD4 T cells usually can still be detected early in HIV infection [24, 25]. Especially the lack of correlation of MTB-specific TH1 cells with the total CD4 cell count led us to hypothesize that MTB-specific CD4 T cells are depleted early during HIV infection. In order to clarify this, we studied the dynamics of RD1- and PPD-specific CD4 T cell responses in 5 latently MTB infected subjects who became HIV infected during the HISIS study .
Before HIV infection, TH1 cell responses targeting either one of the RD1 peptide sets (ESAT6 or CFP10) and PPD were detected in all 5 subjects (figure 4). As expected, during the last HIV seronegative follow up RD1-specific responses (range 0.09 – 0.21% of memory CD4 T cells, background subtracted) and PPD-specific responses (range 0.1 to 0.8% of memory CD4 T cells) were of relatively low magnitude. Although the RD1 peptide sets were specifically designed to study MTB-specific CD8 T cell responses, none were detected in these or any other HIV-subjects with latent MTB infection, which were tested throughout this study. In 4 of 5 subjects with latent infection, MTB-specific responses were rapidly depleted within the first year after HIV seroconversion. In contrast, MTB-specific TH1 cell responses did not fluctuate much in latently MTB infected subjects, who remained HIV seronegative (p<0.01 data not shown). Importantly, none of these four subjects had been diagnosed with active TB within 4 years after HIV infection.
In the 5th subject H19, who had the most dramatic drop in CD4 counts (58% decrease from 810/μl to 343/μl) and a high viral load of continuously above 100.000 vRNA copies/ml, MTB-specific TH1 responses increased after HIV infection and especially PPD-specific CD4 T cells expanded dramatically 1 year after HIV seroconversion (figure 4). In addition, a strong CD8 T cell response targeting CFP10 emerged after HIV infection reaching 0.57% of CD8 T cells. Subject H19 was diagnosed with, and treated for, active TB within 15 months after HIV seroconversion.
The present study was primarily designed to dissect the underlying mechanism associated with the dramatic risk increase to develop pulmonary TB early after HIV infection in subjects latently infected with MTB. Three questions were addressed in particular (1) Does chronic HIV infection affect MTB-specific cellular Immunity in latently TB-infected individuals? (2) Do these cells express the HIV coreceptor CCR5? (3) How does acute HIV infection impact on the MTB-specific cellular immune responses?
The high prevalence of ESAT6-specific responses observed in HIV negative TB asymptomatic subjects suggests that at least half of the studied population from Mbeya region is latently infected with MTB. However, the absence of these cells in many HIV+subjects does not exclude latent MTB infection as demonstrated by the detection of MTB-specific TH1 cell responses before HIV infection and their disappearance thereafter. Together these data suggest that typically co-infected individuals in TB endemic regions are already latently infected with MTB before HIV infection. Although not sufficient, this may contribute to the early manifestation of TB associated with HIV infection.
Chronic HIV infection was associated with a reduced frequency of responders to MTB-antigen, indicating that cellular immunity to MTB was reduced by chronic HIV infection. The rapid depletion of MTB-specific responses early after HIV infection and their presence in HIV+/TB+ subjects suggests that these cells (re)expand after ongoing or transient reactivation of mycobacterial growth or de novo exposure. This interpretation is further supported by two recent studies [26, 27].
Our results support a scenario of early MTB-specific TH1 cell depletion that could be caused by direct HIV infection. Active HIV viral replication indeed is a potent suppressor of MTB-specific TH1 cell responses as demonstrated by their dramatic expansion observed after initiation of antiretroviral therapy [28, 29]. High levels of cell surface expression of the viral co-receptor CCR5 should contribute to direct HIV infection. PPD antigen alone can trigger productive HIV-1 infection of CD4 T cells from HIV-1-infected co-cultured DCs in vitro  and it is sufficient to induce HIV replication and cell death in PBMC from HIV infected TST positive subjects . Alveolar macrophages are easily infected with HIV in vitro and are also infected in vivo during the acute and final phase of Simian Immunodeficiency Virus infection . In addition, anti-MTB-cellular immunity maximizes HIV replication inside alveolar macrophages , which could enhance “MTB-specific” HIV transmission to MTB-specific TH1 cells. A state of complete metabolic quiescence during latent MTB infection may undermine this specific mode of transmission. However, it has been shown that “rather than caused by metabolic quiescence, MTB latency is the result of a continuous cross talk between the host immune system and the persisting pathogen, as indicated by active sites of cell proliferaton and follicle like structures within well vascularized parts of the lungs from latently MTB infected subjects” . The high viral load found in bronchalveolar fluid (BAL) from sites of active TB disease, but not from unaffected sites [34, 35] further supports the hypothesis that the proinflammatory microenviroment caused by local MTB infection enhances HIV transmission to MTB-specific CD4 T cells in vivo.
The importance of TH1 cytokines in mediating protection from TB disease is well documented [36, 37]. Rapid depletion of MTB-specific TH1 cells early during HIV infection could therefore be key to the tremendous risk increase to develop TB. In contrast, CMV-specific CD4 T cells persist until late stages of HIV-infection  and typically CMV associated pathology does not occur until very late stages of AIDS, further supporting our hypothesis that rapid depletion of MTB-specific CD4 T cells is important in HIV/MTB-specific pathology. However, despite the depletion of this important T cell subset, only a minority of coinfected subjects develops TB soon after HIV infection. Specific CD8 T cells, non-conventional T cell subsets or alternative sources of IFNγ and TNFα, such as NK cells, may compensate some of the effector functions of the MTB-specific TH1 cells [8, 38]. In addition, TH17 cells that do not produce IFNγ or TNFα have been shown to participate in MTB-specific immunity  Such responses may still be sufficient in most cases to control MTB after depletion of MTB-specific TH1 cells. Alternatively, host genetic polymorphisms affecting anti-MTB immunity could account for differences in susceptibility to disease. Finally, differences in pathogen load or virulence may also contribute to different TB disease outcomes in HIV/MTB coinfected subjects. Interestingly, during early HIV infection the pulmonary and non-disseminated form of tuberculosis predominates, whereas in AIDS patients with very low CD4 counts, disseminated and extra pulmonary disease is frequently observed, indicating that even in such subjects there is residual anti-MTB-immunity left after acute HIV infection.
In conclusion, our results demonstrate that acute HIV infection is associated with the rapid loss of MTB-specific TH1 cells in the peripheral blood. This contrasts the gradual decline of the total CD4 count during the chronic phase and may be caused by direct HIV infection of these cells. Together these data suggest a mechanism of rapid MTB-specific TH1 cell depletion that may contribute to the early onset of TB often observed in latently infected individuals who become HIV infected.
This work was supported by the European Commission, DG XII, INCO-DC, (grant ICACT-2002-10048). There was no conflict of interests for any author.
We would like to thank all of the HISIS participants and the excellent staff at the Mbeya Medical Research Programme that conducted the HISIS study, especially Weston Assisya, Frowin Nichombe, Clemence Konkamkula and Vera Kleinfeldt. Furthermore we would especially like to thank Natanya Sandler, Philip Scheinberg and Jason Brenchley, (Vaccine Research Center, NIH) for many exiting discussions during the writing of the manuscript.