Research: Immune Phenotype and functionality of Mtb-specific T-cells in HIV/TB co-infected patients on antiretroviral treatment

SATVI researchers Professor Tom Scriba, Associate Professor Elisa Nemes, Cheleka Mpande and Tim Reid have co-authored a journal article titled "Immune Phenotype and Functionality of Mtb-Specific T-Cells in HIV/TB Co-Infected Patients on Antiretroviral Treatment" appearing in the Pathogens journal.

Abstract: 

The performance of host blood-based biomarkers for tuberculosis (TB) in HIV-infected patients on antiretroviral therapy (ART) has not been fully assessed. We evaluated the immune phenotype and functionality of antigen-specific T-cell responses in HIV positive (+) participants with TB (n = 12) compared to HIV negative (-) participants with either TB (n = 9) or latent TB infection (LTBI) (n = 9). We show that the cytokine profile of Mtb-specific CD4+ T-cells in participants with TB, regardless of HIV status, was predominantly single IFN-γ or dual IFN-γ/ TNFα. Whilst ESAT-6/CFP-10 responding T-cells were predominantly of an effector memory (CD27-CD45RA-CCR7-) profile, HIV-specific T-cells were mainly of a central (CD27+CD45RA-CCR7+) and transitional memory (CD27+CD45RA+/-CCR7-) phenotype on both CD4+ and CD8+ T-cells. Using receiving operating characteristic (ROC) curve analysis, co-expression of CD38 and HLA-DR on ESAT-6/CFP-10 responding total cytokine-producing CD4+ T-cells had a high sensitivity for discriminating HIV+TB (100%, 95% CI 70-100) and HIV-TB (100%, 95% CI 70-100) from latent TB with high specificity (100%, 95% CI 68-100 for HIV-TB) at a cut-off value of 5% and 13%, respectively. TB treatment reduced the proportion of Mtb-specific total cytokine+CD38+HLA-DR+ CD4+ T-cells only in HIV-TB (p = 0.001). Our results suggest that co-expression of CD38 and HLA-DR on Mtb-specific CD4+ T-cells could serve as a TB diagnosis tool regardless of HIV status.

Our results show that the activation and memory phenotype of Mtb-specific CD4+ T-cells can be used to discriminate between active TB and LTBI as well as to monitor declines in mycobacterial burden after two months of ATT. These findings extend the results previously reported in ART naïve participants [12,13,20] to individuals on ART.

Conflicting reports on the assessment of the functional profile in HIV+TB have been published, reflecting the challenges of relying on cytokine profiles as biomarkers of TB. While some studies have shown that in active TB there is a depletion of single IFN-γ expression [21,22,23] accompanied by a dominant TNFα profile [24], others have reported that a polyfunctional profile is associated with HIV+TB cases [25]. We observed dominant single IFN-γ and dual functional IFN-γ+TNFα+ CD4+ T-cells in response to ESAT-6/CFP-10 and Mtb-Lysate stimulation consistent with reports from South African cohorts [12,13]. Whether these inconsistencies are due to differences in demographics of the study cohorts, study design, or flow cytometry technical differences is not very clear. More useful than the polyfunctional profile was the PFS that increased in response to ATT in HIV−TB, although there was no difference in scores between HIV+TB and HIV−TB. Therefore, using functional profiles as assessed by COMPASS may be a better measure of response to treatment. Stimulation with Mtb-Lysate also revealed a dominant single TNFα response that had not been evident on ESAT-6/CFP-10 CD4+ T-cells. In addition, all participants responded to Mtb-Lysate, which suggests that this antigen formulation may offer advantages for detecting responses in participants who are unresponsive to ESAT-6/CFP-10 [13]. This is because Mtb-Lysate detects mycobacterial responses induced by Mtb, BCG vaccination, and/or exposure to environmental mycobacteria, [13] which could be a drawback on the specificity of the responses obtained.

HIV is a major risk factor for TB, with deficiencies in Mtb-specific CD4+ T-cells possibly contributing to the risk of active TB prior to declines in peripheral blood CD4+ T-cells [26]. While ART-induced immune reconstitution is well described, the impact of ART on co-pathogen specific CD4+ T-cell responses is poorly understood [17,27]. For example, although ART induced declines in bulk CD4+ and CD8+ T-cells co-expressing CD38 and HLA-DR after a year of ART, the activation was still higher than on T-cells from HIV-uninfected individuals [18]. Since the activation level of T-cells is one of the factors influencing the extent of immune restoration [18], the use of blood-based markers may be an effective tool to assess the risk of active TB prior to ART initiation. In our cohort of participants who had been on ART for a median of one year, we found that the immune activation in HIV+TB, measured as a proportion of CD38+HLA-DR+totalcytokine+ ESAT-6/CFP-10-responding CD4+ T-cells, was comparable to HIV−TB but higher than LTBI at the time of TB diagnosis. In addition, significant declines in activation at two months post-ATT were observed only in HIV−TB. This could indicate that we needed a longer follow-up period and a larger sample size to observe a decline in response to TB treatment in the HIV+TB group.

Prior studies have shown that the activation profile of Mtb-specific CD4+ T-cells defined by CD38, HLA-DR, and Ki67 in combination [11,12,13] or alone [14] can distinguish latent from active TB with high sensitivity. Our results of a high sensitivity for the co-expression of CD38 and HLA-DR on Mtb-specific CD4+ T-cells as reflected by the ROC curve analysis agree with reports from Riou et al. [13] and Wilkinson et al. [12], in our analysis, the co-expression of CD38, HLA-DR, and Ki67 had poor sensitivity for TB diagnosis and the proportion of cells expressing this phenotype did not change after two months of TB treatment. The expression of Ki67 has also been shown to be a less robust marker for differentiating active disease from LTBI in a study of HIV infected participants who were not on ART [12].

The chronic antigen exposure that occurs in active TB disease drives the maturation of Mtb-specific CD4+ T-cells toward a late differentiated phenotype [28]. CD27 loss has been shown to differentiate active TB from LTBI [9,10,13,28] and associates with clinical disease severity and lung tissue damage in active TB [29]. Consistent with these reports, Mtb-specific CD4+ T-cells in response to ESAT-6/CFP-10 stimulation from active TB cases, regardless of HIV status, had a dominant effector memory phenotype compared to the central memory phenotype on LTBI. In contrast, Gag stimulated CD8+ T-cells had a dominant transitional memory phenotype that has been associated with HIV virologic control in participants not on ART and reduced immune activation [30].

Recently, a T-cell-activation marker-tuberculosis (TAM-TB) assay that measures the CD27 phenotype of IFN-γ producing cells was shown to have good performance characteristics in children [10]. We extended the analysis of the CD27 MFI ratio (measured as CD27 MFI on CD4+ T-cells divided by CD27 MFI on IFN-γ+CD4+ T-cells), suggested by Portevin et al. [10], to total cytokine+CD4+ T-cells, as recently reported [9]. We added new information regarding the CD27 MFI ratio in HIV+TB participants and showed that the ratio was highest on ESAT-6/CFP-10 responding total cytokine+ CD4+ T-cells in active TB participants regardless of HIV status. Previously, CD27 was analyzed as a % or MFI on IFN-γ+CD4+ T-cells of HIV+TB who were not on ART also showed good sensitivity in distinguishing active disease from infection [13]. In our analysis, TB treatment also caused declines in the CD27 MFI ratio in the HIV−TB participants. An increase in CD27 expression on Mtb-specific T-cells has also been shown to correlate with the conversion of sputum culture and lung repair measured as the reduction in lung tissue destruction on radiological examination at two months of ATT [29].

DiscussIon: Our results should be interpreted in light of some limitations to the analyses. The sample size of our study groups was very small and was limited by studying participants with paired samples to detect early changes during ATT. This further reduced the sample sizes and may have possibly masked some differences and limited our ability to detect changes in certain phenotypes. Larger cohorts would be required to extend our findings. Whilst biomarkers for early treatment monitoring are important for identifying patients that are likely to respond well to shorter treatment regimens [31], the lack of an end of treatment timepoint and the fact that all patients had culture negative results at two months, limits our ability to comment on whether changes in the biomarkers are associated with favorable responses to ATT. Nonetheless, our results show that although TB treatment generally lasts six months, there are a number of patients that will culture convert by the end of the intensive phase of ATT. In addition, the inclusion of Gag stimulation allowed us to observe that these changes in T-cell activation appeared to be specific to the reduction of Mtb load during ATT because Gag-specific cells remained unaffected. However, the functional impact of the divergence in the memory phenotypes observed between Mtb and HIV-specific T-cells could not be determined due to the nature of the study design. Furthermore, we did not include HIV+ participants without TB as a control, which limits our ability to comment on the specificity of such responses.

Overall, our results show that CD38 and HLA-DR co-expression on Mtb-specific T-cells has good performance characteristics for active TB regardless of HIV status and has potential for early treatment monitoring in HIV+TB individuals on ART. The CD27 MFI ratio performed well on ESAT-6/CFP-10 responding total cytokine+ CD4+ T-cells in HIV−TB and needs further validation in large, well-powered cohorts to define its potential in monitoring early treatment responses in HIV+TB.

Keywords: CD38; HLA-DR; immune activation; treatment response.