Polyfunctional T cells in human tuberculosis


  • Katalin A. Wilkinson,

    1. MRC National Institute for Medical Research, Mill Hill, London, UK
    2. Clinical Infectious Diseases Research Initiative, Institute of Infectious Diseases and Molecular Medicine, University of Cape Town, South Africa
    Search for more papers by this author
  • Robert J. Wilkinson

    Corresponding author
    1. MRC National Institute for Medical Research, Mill Hill, London, UK
    2. Clinical Infectious Diseases Research Initiative, Institute of Infectious Diseases and Molecular Medicine, University of Cape Town, South Africa
    3. Division of Medicine, Imperial College London, UK
    • Room 3.03.05 Clinical Infectious Diseases Research Initiative, Wolfson Pavilion, Institute of Infectious Diseases and Molecular Medicine, Faculty of Health Sciences, University of Cape Town, Anzio Road, Observatory 7925, South Africa Fax: +27-21-406-6796
    Search for more papers by this author


Studies of chronic viral infections have highlighted the phenotypic and functional heterogeneity of antigen-specific T cells and suggested that polyfunctional T cells that secrete multiple cytokines and are able to proliferate on encounter with antigen are more likely than single cytokine secretors to represent correlates of protective immunity. These findings have prompted the evaluation of such T–cell responses in chronic bacterial infections, such as tuberculosis (TB). A number of studies in humans suggested that polyfunctional T cells may indeed be involved in mediating protection in TB; however, studies that question these findings are also emerging, including a study published in this issue of the European Journal of Immunology. These differing findings highlight the difficulties of studying human immunity to TB and the need for polyfunctional T cells to be evaluated in longitudinal studies as opposed to case-control analyses.

Based on surface phenotype and function, T cells can be divided into naïve (which give rise to memory cells following antigen encounter), central memory (which home to the lymph nodes following immune control of the pathogen and need co-stimulation to expand upon rechallenge) and effector memory/terminally differentiated effector subsets (which home to disease sites and are able to perform immediate effector functions without needing co-stimulation) 1, 2. The simultaneous measurement of a large number of surface and intracellular markers using polychromatic flow cytometry has recently enabled a multitude of T-cell subsets to be distinguished; however, the sequence of development of these different T-cell subsets remains undefined in humans 3.

Different pathogens pose different challenges to the immune system and variations in antigen exposure, persistence, load, localisation and pathways of presentation, cytokine environments and co-stimulation mechanisms/molecules may all affect T-cell differentiation 4, 5. Studies of chronic viral infections such as cytomegalovirus, Epstein–Barr Virus, and others have highlighted the phenotypic and functional heterogeneity of the antigen-specific CD4+ and CD8+ T-cell responses, and provided evidence that they are dependent on, and modulated by, antigen concentration 6. Overall it has been shown that polyfunctional T cells that secrete multiple cytokines and are able to proliferate are more likely than single cytokine secretors to represent correlates of protective antiviral immunity in chronic infections (when antigen load is low), while single IFN-γ-secreting CD4+ and CD8+ T cells are characteristic of acute infections (when antigen load is high) 7. If chronic infection ensues after failure of complete immune control, the balance of responding T cells tends to shift towards the single IFN-γ-secreting phenotype. This process is particularly skewed in the case of HIV-1 infection, as the HIV-1-specific CD4+ and CD8+ T-cell response is overwhelmingly dominated by a single IFN-γ-secreting effector response during both the primary and chronic phases of infection 7. Antiretroviral treatment (ART) during chronic HIV-1 infection results in a shift to a memory response composed of a polyfunctional phenotype that has also been observed in individuals who spontaneously control HIV-1 replication 8, 9.

These advances in understanding what may constitute protective anti-viral immune responses in chronic viral infections have prompted the evaluation of polyfunctional T-cell responses in chronic bacterial infections, such as tuberculosis (TB). In humans, vaccination with a novel tuberculosis vaccine based on modified Vaccinia virus Ankara expressing Ag85A (MVA85A) induces T cells with complex phenotypic and polyfunctional cytokine-secreting profiles, including a less differentiated group that co-expresses IFN-γ, TNF and IL-2, and with the ability to proliferate. Such cells are considered more likely than single cytokine secretors to be associated with protection 10, 11. Another novel TB vaccine, based on a recombinant replication-deficient adenovirus (Ad35)-expressing TB antigens, also induces a robust polyfunctional CD4+ T-cell response in adults 12. Similar studies of BCG vaccination in 10-wk-old infants showed the emergence of a diverse set of T cells expressing various combinations of IFN-γ, IL-2, and TNF 13.

Polyfunctional Mycobacterium tuberculosis (MTB)-specific T cells secreting combinations of IFN-γ/TNF or IFN-γ/IL-2 have also been detected in people dually infected with MTB and HIV. A negative correlation was found between HIV-viral load and the proportion of MTB-specific IL-2-secreting (IFN-γ+/IL-2+ and IFN-γ/IL-2+) CD4+ T cells and a positive correlation was found with IFN-γ single positive CD4+ T cells and HIV-1 viral load 14. These data suggest that MTB-specific T cells secreting IL-2 (with or without IFN-γ) may be proportionally diminished when HIV-1 viral load is high, corresponding with greater susceptibility to develop active TB. Therefore MTB-specific CD4+ T cells secreting IL-2 alone or in combination with other cytokines could be potential correlates of MTB protection. Treatment of MTB in HIV-uninfected TB patients leads to a shift from predominantly IFN-γ single positive CD4+ T cells to predominantly IFN-γ+/IL-2+ and newly detectable IL-2 single positive T cells 15. Together with data derived from mouse models of Leishmania and TB 16, 17, the above studies suggest that polyfunctional T cells may be associated with protection against intracellular pathogens in general, not just viruses; however, it is important to point out that none of the above studies provided direct evidence that these polyfunctional T cells are associated with protection, for example by comparing patients who do or do not have active TB disease.

Against this background an interesting study by Caccamo et al. is published in this issue of European Journal of Immunology18, reporting two separate studies carried out in Italy and The Netherlands. Following short-term restimulation with recombinant antigens, PBMC from 20 Italian patients with untreated active TB were shown to have a higher frequency of polyfunctional CD4+ T cells simultaneously expressing IFN-γ, IL-2, and TNF than 18 subjects with latent TB infection (LTBI). Elegant analysis in the paper shows that during antitubercular therapy the frequency of polyfunctional CD4+ cells declines, and is accompanied by an increase in the frequency of IFN-γ+/IL-2+ and IFN-γ single positive antigen-specific cells. In The Netherlands, a smaller number of patients with previously treated TB were studied (four patients) but were nevertheless shown to have detectable polyfunctional CD4+ T cells after restimulation of PBMC for 6 days with peptide antigen (long-term stimulation), indicative of previous polyfunctional T cell responses, whereas such responses were infrequent amongst eight subjects with LTBI. The authors conclude that polyfunctional CD4+ T cells may be a useful biomarker of active TB and interpret their data as indicating that polyfunctional responses are not associated with protection against TB.

The findings and interpretation are similar to a report from the Gambia 19, also showing that HIV-uninfected TB patients had significantly higher levels of IFN-γ/TNF/IL-2 CD4+ T cells compared with healthy household contacts; however, the Gambian study 19 also found higher numbers of CD4+ and CD8+ T cells in active TB patients that were TNF single positive, and a significantly increased proportion of IL-2 single positive CD4+ and CD8+ T cells in PBMC from the household contact group in response to both PPD and ESAT-6/CFP-10 stimulation. In the Caccamo et al. study 18 TNF single positive T cells were not analysed, while the IL-2 single positive response was not different between the TB and LTBI groups. These differences may have arisen because of marked differences in the phenotype of the control (LTBI) group in these studies; household contacts in a relatively high incidence environment are assumed as recently exposed and infected, whereas the Italian and Dutch subjects are not described as having recent contact with TB patients.

This difference and the different interpretation the authors of 18 place on their findings compared with studies published hitherto encourages reflection on the problems faced when trying to study human immunity to TB. Does the presence of a particular immune response in blood during disease and its disappearance during therapy indicate that the response is non-protective? The authors of 18 note the essentiality of IFN-γ to protect against TB, yet also acknowledge it is often found to be elevated in the blood cells of TB patients and even more so at disease sites 20. Only one human study has so far assayed polyfunctional CD4+ T cells in the lungs of persons exposed to TB, finding them to be detectable in HIV-1-uninfected persons but essentially absent from the lungs of highly MTB-susceptible HIV-1-infected persons 21. It is recognised that the risk of TB is greatest early after infection so when we phenotype LTBI should household contacts be regarded as protected or at risk? Conversely, can TB infection truly be dichotomised into binary latent and active forms? Do all those with evidence of immune sensitisation (the only way to infer LTBI) harbour viable bacilli or can they be eliminated 22? If so, are such persons resistant or susceptible to reinfection?

What is then the best study design to learn about immunity to TB in humans? The most informative section of the work by Caccamo et al. 18 is the repeat testing after antitubercular therapy. Longitudinal ‘natural history’ studies such as those performed in the Gambia where contacts are followed until the development of disease or otherwise are also potentially powerful; however, such studies are expensive and not without ethical tension as preventive antitubercular therapy is routine and effective in many parts of the world and, in the case of HIV-1 co-infection, the benefits of antiretroviral therapy in preventing TB are also now well-recognised. Both antitubercular and antiretroviral therapy are associated with changes in the antitubercular immune response. In fact, studying the effect of ART on MTB-specific T-cell responses in persons co-infected with both HIV and TB may be a model to identify correlates of protection as very susceptible people become less susceptible and thus form their own controls. We have recently described the ‘returning’ MTB-specific T-cell responses in HIV-1-infected MTB-sensitised South African adults over 12 months from the time they started ART, and found that while ART is associated with an absolute increase in effector function (often evaluated via IFN-γ secretion alone), the response declined proportionally, and the strongest correlate of increased ART-mediated immunity was the central memory response 23. Subsequent adoption of the same study design in the Gambia showed complementary data, as 12 months of ART restored polyfunctional CD4+ T-cell responses in MTB-infected patients, with a significant increase as early as 3 months in response to PPD and 6 months in response to ESAT-6/CFP-10 24. The results were primarily due to significant increases in IL-2-expressing CD4+ T cells, in conjunction with either IFN-γ or TNF, while cells expressing IFN-γ alone significantly declined proportionally over time 24.

Another opportunity to learn from intervention is presented by trials of novel TB vaccines, several of which are now at a stage of advanced clinical evaluation. What is clear and necessary from all the studies discussed in the Commentary, including that by Caccamo et al. 18, is that we need to continue to bring cutting edge advances in flow cytometry to bear on critical global health problems in the hope that advances both in understanding but also in prevention of TB may occur.


The authors are funded by the Wellcome Trust (084323, 088316), Medical Research Council and European Union (Sante/2006/105-061).

Conflict of interest: The authors declare no financial or commercial conflict of interest.