Immunity to tuberculosis is dependent on type 1 responses (interferon (IFN)γ, interleukin (IL)12, tumour necrosis factor (TNF)α) but these do not provide a complete explanation for the regulation of the immune response seen. Additional effector or regulatory mechanisms need to be sought.1 Recently, a new T cell subset, IL-17 producing CD4+ Th cells (Th17), has been described.2 In the mouse at least, IL-17 secreting cells may be distinct from conventional Th1, Th2 and regulatory T cells.2 These IL-17-secreting T cells drive secretion of TNFα, IL1β and IL6, as well as chemokines CXCL1, 2 and 8, and enhance inflammation and influx of neutrophils.2 IL17 has tentatively been implicated in mouse models of immunity to mycobacteria.3 ,4 However, there are no data from the human lung or from patients with tuberculosis (TB).

To ascertain whether Th17 related cytokines are expressed in tuberculous lungs and to evaluate the anatomical (blood vs lung) and cellular (CD4 vs CD8) compartmentalisation of these cytokines, we measured expression of genes encoding IL4, IFNγ, IL17 and IL23 (a cytokine that increases expression of IL17 by T cells) in bronchoalveolar lavage (BAL) cells (unfractionated) of 14 HIV negative culture positive TB patients and 14 matched controls with no laboratory evidence of latent TB infection (determined by antigen specific IFNγ responses; T SPOT TB), and correlated these findings with clinical parameters. Blood was obtained from all participants, and BAL from all TB patients but only from six control subjects. Bronchoscopy, isolation of cells and qRT-PCR (normalised to a validated reference after quality control of RNA), were performed as previously described.1 Blood derived peripheral mononuclear cells were separated into CD3, CD4 and CD8 fractions.

We showed previously using the same samples that IFNγ and IL4 levels were significantly elevated in BAL fluid of patients compared with controls.1 Here we show that this is also true for IL23 mRNA (fig 1A). In contrast, patients had reduced IL23 mRNA levels in blood, suggesting that in TB there is sequestration of IL23 expressing cells in the lung (fig 1B). However, IL17 mRNA levels were similar in patients and controls, both in blood and BAL fluid (fig 1B). In blood, IL17 was expressed predominantly in CD8 rather than CD4 cells of TB patients (fig 1C). Neither IL17 nor IL23 mRNA levels correlated radiological disease extent or presence of cavitation on chest radiography (p = 0.09).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1 Gene expression profiles of interleukin (IL)17 and IL23 in whole blood and bronchoalveolar lavage (BAL) cells of patients with tuberculosis (TB) compared with controls. IL23 mRNA levels were significantly higher in TB patients in BAL cells (A) but not in peripheral blood (B); IL17 mRNA levels were not significantly different in blood or BAL cells (C) but were prominently expressed in CD8+ T cells (D). RNA was qualitatively and quantitatively assessed using a bioanalyser and mRNA values were normalised to 1 million copies of HuPO, a validated reference gene.

In the mouse, IL17 secreting CD4+ cells have been considered to be a distinct helper T cell lineage.2 Nevertheless, in mice infected with M tuberculosis, most of the IL17 was expressed by γ/δ T cells.4 In humans, the situation is even more confusing. IL17 can be expressed in both naïve and memory T cells after non-specific stimulation, in CD8+ cells and in cells simultaneously expressing IFNγ or transforming growth factor β, suggesting Th1-like and regulatory properties, respectively.5 There might not be a distinct Th17 lineage. Thus although we did not observe altered expression of IL17 in TB, it will be essential to study in detail the effect of counter-regulatory cytokines and the phenotypes of the IL17 expressing cells, since this might be altered even though the overall levels of IL17 mRNA were not. Furthermore, an apparently “normal” IL17 expression profile does not preclude its importance in the early stages of infection. Answering these questions will require a new study because the stored clinical material is now exhausted. A limitation of this preliminary study is the measurement of mRNA and not protein levels. However, measurement of biomarker protein levels, due to massive dilution and lack of a reliable normalisation strategy, is problematic in BAL fluid.1

In conclusion, although similarly expressed in patients and controls, IL17 may be a potentially important cytokine in the immunopathogenesis of TB. Increased IL23 expression in TB lung is likely to be inducing IL17 expression by some cell types that now need to be identified.

Acknowledgements

This work was funded in part by the British Lung Foundation. KD is supported by a SA NRF Research Chair Award and a MRC Career Development Award.