In The Gambia, Mycobacterium tuberculosis (Mtb) and Mycobacterium africanum (Maf) are major causes of tuberculosis (TB). Maf is more likely to cause TB in immune suppressed individuals, implying differences in virulence. Despite this, few studies have assessed the underlying immunity to the two pathogens in human. In this study, we analyzed T-cell responses from 19 Maf- and 29 Mtb-infected HIV-negative patients before and after TB chemotherapy following overnight stimulation of whole blood with TB-specific antigens. Before treatment, percentages of early secreted antigenic target-6(ESAT-6)/culture filtrate protein-10(CFP-10) and purified protein derivative-specific single-TNF-α-producing CD4+ and CD8+ T cells were significantly higher while single-IL-2-producing T cells were significantly lower in Maf- compared with Mtb-infected patients. Purified protein derivative-specific polyfunctional CD4+ T cells frequencies were significantly higher before than after treatment, but there was no difference between the groups at both time points. Furthermore, the proportion of CD3+CD11b+ T cells was similar in both groups pretreatment, but was significantly lower with higher TNF-α, IL-2, and IFN-γ production in Mtb- compared with that of Maf-infected patients posttreatment. Our data provide evidence of differences in T-cell responses to two mycobacterial strains with differing virulence, providing some insight into TB pathogenesis with different Mtb strains that could be prospectively explored as biomarkers for TB protection or susceptibility.
Tuberculosis (TB) remains a major threat to public health worldwide; with 1.4 million deaths in 2011 and an estimated one-thirds of the global population infected by the causative pathogen, Mycobacterium tuberculosis complex (MTBC) . In West Africa, TB is caused by two distinct strains Mycobacterium africanum (Maf) and Mycobacterium tuberculosis (Mtb), with Maf causing up to half of all TB . In The Gambia, TB patients infected with Maf West African type 2 are more likely to be older, HIV coinfected, and/or severely malnourished . In addition, their exposed household contacts have a slower rate of progression to active disease compared with their Mtb-exposed counterparts , suggesting that Maf might be the less virulent of the two prevalent strains [5, 6]. The heterogeneity in chest x-rays, bacillary load in sputum and clinical symptoms, likely reflects the variation in host immune response to the invading pathogen, and/or the differences between the lineages within the MTBC. For instance, Maf has slower growth in culture medium compared with Mtb [6, 7], which may influence progression to active disease .
The recently published genome of Maf reveals a high content of pseudogenes and the presence of a unique sequence — the region of difference 900 “RD900” that has been deleted evolutionarily from Mtb and M. bovis strains . Many studies have concentrated on the pathogen strain differences in various geographical locations, with only a few analyzing host-related factors (reviewed in ). This includes a decreased IFN-γ response to ESAT-6 in Maf-infected patients and exposed household contacts . Untreated Maf-infected patients also tend to have greater production of IL-10 and TNF-α compared with Mtb-infected patients in whole blood culture supernatant following stimulation with purified protein derivative (PPD) and TB10.4 . However, there have been no studies analyzing responses to TB antigens using multiparameter flow cytometry. Furthermore, despite differences in virulence, little analysis has been done on posttreatment recovery of TB patients infected with the distinct strains.
In this study, we used polychromatic flow cytometry to determine differences in T-cell responses to Mtb-specific antigens and whole mycobacteria between 19 Maf- and 29 Mtb-infected patients before and after treatment. We identified differences in host immune responses to mycobacteria-specific antigens between TB patients infected with Mtb and Maf. In particular, there was significantly higher proportion of single-TNF-α and lower proportion of single-IL-2-producing CD4+ and CD8+ T cells in Maf- compared with Mtb-infected patients pretreatment, and persistently high proportion of CD3+CD11b+ T cells (activated T cells) in Maf-infected patients after treatment. These differences have implications for TB disease protection or susceptibility and will form the basis of subsequent biomarker investigations.
Forty-eight HIV-negative TB patients were included in this study. Following spoligotyping, 19 were classified with Maf West African type 2 infection and 29 with Mtb Europe, America, and Africa lineage-infection (Supporting Information Table 1). Both groups had similar clinical severity of disease at baseline as shown by the body mass index, skinfold thickness, and chest X-ray score (Table 1). There was no significant difference in age, gender, and ethnicity between Maf- and Mtb-infected patients (Table 1). The hemoglobin concentration, as well as the red blood cell, lymphocyte, and monocyte counts all significantly increased while granulocyte, platelet, and total white blood cell counts significantly decreased after treatment; with no differences in any parameters seen between the two groups (data not shown). Following flow cytometry analysis, three patients had poor quality data at baseline; therefore, data from 18 Maf- and 27 Mtb-infected patients were examined before treatment. Thirty-three patients were successfully followed up after treatment (14 Maf- and 19 Mtb-infected); all were sputum acid-fast bacilli (AFB) smear-negative after 6 months of treatment.
Table 1. Demographic and baseline clinical characteristics of M. africanum and M. tuberculosis-infected patients
T-cell cytokines production in Maf and Mtb-infected patients differed pre- and posttreatment
The frequencies of CD4+ (Fig. 1) and CD8+ (data not shown) T cells producing IFN-γ, TNF-α, and/or IL-2 in both groups differed depending on the antigen used (Fig. 1). A distinct cluster of CD4+ T cells producing two cytokines following stimulation with PPD was consistently observed before treatment and significantly reduced after treatment independent of the strain of infection (Fig. 1B and C). The proportions of CD4+IFNγ+ and CD4+TNF-α+ significantly increased in both groups posttreatment following phytohaemagglutinin (PHA) stimulation, but were not significantly different between the groups at both time points (Fig. 2A). In response to early secreted antigenic target-6 kDa (ESAT-6)/culture filtrate protein-10 kDa (CFP-10) peptides pool (EC), the proportions of CD4+IFN-γ+, CD4+IL-2+, and CD8+IL-2+ T cells were significantly higher in Mtb- than Maf-infected patients before treatment (p = 0.025, p = 0.018, and p = 0.0005, respectively; Fig. 2C and D), but with no differences seen posttreatment. In response to PPD and live Mtb (LMtb) and Maf (LMaf), there were no significant differences in CD4+ and CD8+ T cells cytokine responses between the two groups before and after treatment (Supporting Information Fig. 1 A–D and Fig. 2G andH). However, stimulation with LMaf induced significantly higher proportions of CD4+ (IFN-γ+ and IL-2+) and CD8+ (all three cytokines) T cells in Mtb-infected patients posttreatment compared to pretreatment (Fig. 2G and 2H). Similarly, the proportion of CD8+IFN-γ+ T cells significantly increased posttreatment in Maf-infected patients following stimulation with LMaf (Fig. 2H). Heat-killed Maf (KMaf) induced a significantly higher proportion of CD8+IL-2+ T cells in Mtb- compared with Maf-infected patients before treatment (p = 0.04; Fig. 2F). Following treatment, the proportion of CD4+IFN-γ+ T cells was significantly increased in Mtb-infected patients in response to KMaf (Fig. 2E). Whereas, stimulation with heat-killed Mtb (KMtb) induced no significant responses between the two groups and no significant changes between the two time points (Supporting Information Fig. 1E and F).
Quality of T-cell responses differs between Mtb and Maf-infected patients pretreatment
Polyfunctionality of the CD4+ and CD8+ T cells was assessed using SPICE software . Prior to treatment, the proportion of single TNF-α-producing CD8+ T cells was significantly higher in Maf than Mtb-infected patients in response to EC and heat-killed Maf stimulation (p = 0.035 and p < 0.0001, respectively; Fig. 3B and D) and a similar result was observed with single TNF-α+CD4+ T cells in response to PPD (p = 0.026, Fig. 3C). However, single TNF-α+CD4+ T cell subsets in response to EC and PPD and single TNF-α+CD8+ in response to heat-killed Maf significantly decreased in Maf-infected patients posttreatment (Fig. 3B, C, and D). Single IL-2+CD8+ T cells were significantly higher in Mtb- than Maf-infected patients before treatment in response to EC, PPD, KMaf, and KMtb stimulation (p = 0.002, p = 0.009, p = 0.002, and p = 0.035, respectively; Fig. 3B, C, and D and Supporting Information Fig. 2C). The proportion of single IFN-γ+ (CD4+ and CD8+) T cells were similar between the groups and significantly increased in both groups posttreatment in response to KMaf stimulation (Fig. 3D). Single IFN-γ+CD4+ T cells significantly increased posttreatment in both groups following stimulation with PPD (Fig. 3C) and in Mtb-infected patients only following stimulation with EC (Fig. 3B). Similarly, single IFN-γ+CD8+ T cells significantly increased posttreatment only in Mtb-infected patients in response to LMtb and KMtb stimulation (Supporting Information Fig. 2B and C). Stimulation with PHA showed no differences in single cytokine producing T cell subsets between Maf and Mtb-infected groups.
The proportion of CD4+ and CD8+ IFN-γ+TNFα+ T cells both significantly increased in Maf-infected patients posttreatment, and CD8+ IFN-γ+TNFα+ was significantly higher in Maf- than Mtb-infected groups posttreatment in response to PHA stimulation (p = 0.045 and p = 0.032, Fig. 3A). However, in response to PPD, CD8+IFN-γ+TNFα+ T cells were higher in Maf than Mtb-infected groups before treatment (p = 0.036, Fig. 3C). Following PHA stimulation, CD4+ and CD8+ IL-2+TNF-α+ T cells both significantly increased in Mtb-infected patients posttreatment and were higher in Mtb- than Maf-infected group posttreatment (p = 0.048 and p = 0.007, respectively, Fig. 3A). In addition, CD8+IL-2+TNF-α+ T cells frequency was significantly higher in Mtb- than Maf-infected groups pretreatment in response to EC (p = 0.032, Fig. 3B).
The proportion of T cells producing all three cytokines was similar in Mtb and Maf-infected groups regardless of stimulation (Fig. 3 and Supporting Information Fig. 2). However, both groups had significantly reduced polyfunctional CD4+ T cell subsets following treatment in response to PPD (p = 0.005 and p = 0.008, respectively, Fig. 3C). Whereas, polyfunctional CD8+ T cells significantly increased in Mtb-infected patients in response to PPD, LMaf, and LMtb stimulation (p = 0.01, Fig. 3C; p = 0.007 and p = 0.022, Supporting Information Fig. 2A and B).
The overall functional profile of CD8+ T cells was significantly different between the Mtb and Maf-infected groups before treatment in response to EC, PPD, and KMaf (pie graphs; Fig. 3B–D). In addition, CD8+ T cells functional profile was different pre- and posttreatment in both groups following stimulation with KMaf but only in Maf-infected group in response to PPD. Likewise, CD4+ T cells functional profile was different pre- and posttreatment in both groups following stimulation with PPD but only in Maf group in response to EC (pie graphs; Fig 3B and C).
CD3+CD11b+ T cells are higher in Maf than Mtb-infected patients after treatment
The complement receptor 3 (CR3) or CD11b/CD18 is a surface marker commonly expressed on DCs, macrophage monocytes, and NK cells populations, it is also known as a homing receptor to sites of inflammation [13-15]. Recent studies have shown CD11b expression by highly activated T cells that are mainly cytotoxic [16, 17]. Following gating on the lymphocyte population, activated T cells were defined as CD3+CD11b+ cells and the proportions of IFN-γ, IL-2, and TNF-α positive subsets within this population were determined. CD3+CD11b+ cells produced a higher proportion of cytokines compared to CD3+CD11b− cells (Fig. 4A) and were predominantly CD8+ T cells as expected (not shown).
Similar levels of CD3+CD11b+ T cells were seen in both groups prior to treatment, but lower proportion was seen after treatment following stimulation with PPD in Mtb-infected patients compared to pretreatment (p = 0.0045 Fig. 4B). Consequently, the proportion of CD3+CD11b+ T cells was significantly lower in Mtb- compared to Maf-infected patients posttreatment (p = 0.0061, Fig. 4B), similar findings were observed with all other stimuli (Supporting Information Table 2).
We further dissected the differential proportions of CD3+CD11b+ T cells between the groups posttreatment by investigating whether Maf-infected patients also had higher frequency of T cells coexpressing CD11b and CD56 (Fig. 4C). Although, CD11b+CD56+ T cells represented a smaller proportion of CD3+CD11b+ T cells in both groups their proportion was significantly higher in Maf than Mtb-infected patients posttreatment in response to all stimulant (Fig. 4D). As expected, CD11b+CD56+ T cells were mainly made of CD8+ than CD4+ T cells in both Maf and Mtb-infected patients (p = 0.0006 and p = 0.0002, respectively; Fig. 4E).
CD3+CD11b+ T cells produce less cytokines in Maf than Mtb-infected patients posttreatment
CD3+CD11b+ T cells produced cytokines in response to all stimulants tested with no significant differences between Maf and Mtb-infected patients before treatment. However, after treatment, the proportion of CD3+CD11b+TNF-α+ T cells was significantly higher in Mtb- than Maf-infected patients in response to PPD, LMtb LMaf, and KMtb stimulation (p = 0.013, 0.006, 0.008, and 0.012, respectively; Fig. 5C, 5D, E, and F). Similar result was seen for CD3+CD11b+IFN-γ+ T cells in response to LMtb and KMtb stimulation (p = 0.0065 and p = 0.013, respectively; Fig. 5D and F). Likewise, CD3+CD11b+IL-2+ was significantly higher in Mtb- than Maf-infected patients posttreatment in response to PPD and KMtb stimulation (p = 0.028 and p = 0.025, respectively; Fig. 5C and 5F).
Posttreatment differences between the groups were due to significant increase in cytokine production by CD3+CD11b+ T cells in Mtb- more than Maf-infected patients. For instance, CD3+CD11b+IFN-γ+ T cells was significantly increased posttreatment in Mtb-infected patients in response to all seven stimuli except PPD (Fig. 5C), while in Maf-infected patients increase was seen only in response to EC (Fig. 5B). CD3+CD11b+IL-2+ T cells were significantly increased in Mtb-infected patients posttreatment following stimulation with PHA, KMtb, and KMaf (p = 0.026, p = 0.011, and p = 0.012, respectively; Fig. 5A, F, and G). Similarly, CD3+CD11b+TNF-α+ T cells significantly increased in Mtb-infected patients following treatment in response to PHA and EC (p = 0.0006 and p = 0.015, respectively; Fig. 5A and 5B). In Maf-infected patients CD3+CD11b+TNF-α+ T cells significantly increased in response to PHA stimulation following treatment, while there was a significant decrease in response to LMtb stimulation (p = 0.00047 and p = 0.023, respectively; Fig. 5A and 5D).
This study provides the first detailed analysis of host immunological responses to two different strains of the Mycobacterium tuberculosis complex: M. africanum and M. tuberculosis affecting the same host population. The study design and the choice of antigens allowed the investigation of host immune response before and after treatment to the two common strains that cause TB disease in our setting. We observed differences between TB patients caused by Mtb and Maf in their T-cell responses that varied according to the mycobacterial antigens and time of anti-tuberculosis treatment (pre- or posttreatment).
Despite similar age, ethnic background, and degree of disease severity between the groups prior to treatment, stimulation with heat-killed mycobacteria as well as with ESAT-6/CFP-10 peptides pool and PPD, induced a significantly lower proportion of single-positive IL-2 and higher proportion of single-positive TNF-α CD4+ and CD8+ T cells in the Maf- compared with the Mtb-infected groups. Interestingly, this profile was mycobacterial-antigens-specific as stimulation with PHA (mitogen) showed no differences between the groups. A higher proportion of single-TNF-α producing CD4+ and CD8+ T cells in active TB patients compared with latent TB infected (LTBI) or treated patients is well documented, with high single-TNF-α production associated with active TB disease [18-25]. TNF-α controls MTBC infection progression to disease as evidenced by the increase incidence of active TB in patients treated with anti-TNF-α [26, 27]. However, its protective role appears to be dose dependent with increased levels promoting immunopathology [28, 29]. IL-2 also is important in the host by stimulating the proliferation and differentiation of T cells and NK cells for increased effectiveness in MTBC elimination. Decreased Mtb-specific CD4+ T cells producing IL-2 in HIV-infected patients, has been suggested to increase the risk of TB reactivation , implying that higher single-IL-2 producing T cells may be protective against TB [20, 21]. Although this functional profile of mycobacteria-specific CD4+ and CD8+ T cells observed in Maf-infected patients prior to treatment might suggest why they developed disease from a less virulent bacillus, the lack of a threshold for protection or susceptibility for these cytokines requires interpretation with caution. Nevertheless, further prospective works on this signature are required to conclude the role of these as biomarkers in TB pathogenesis.
We saw a similar proportion of polyfunctional CD4+ T cells producing all three cytokines, which significantly reduced posttreatment compared to pretreatment in both groups, as shown in previous studies from The Gambia and The Netherlands [21, 22]. This suggests that they might play a similar role in TB disease induced by Maf and Mtb and that their presence is associated with the bacteria load independent of the infecting strain type. The current paradigm in TB is that an increase in T-cell responses might be exploited by the mycobacteria as a survival strategy, by increasing lung immunopathology (cavity formation) and subsequently promoting transmission . Clearly, other T cell subsets (i.e. Th17, Th2, Treg ) will play a role in overall immunity to MTBC. We are currently assessing this using global cytokine production (Luminex technology) to determine the correlation with clinical recovery after treatment.
The higher response to Maf by the Mtb-infected patients posttreatment might indicate that they have a stronger immune profile allowing them to mount this robust inflammatory response to Maf. It could also be a sign of accelerated recovery from disease by the Mtb-infected patients following treatment or may suggest that Maf is the less virulent of the two pathogens and the group with a more competent immune profile will respond vigorously to the weaker pathogen [3, 4].
The loss of the difference between the two groups in T cells cytokine responses to ESAT-6/CFP-10 stimulation posttreatment was due to increased responses in Maf-infected patients, which might result from increased abundance of ESAT-6 that is an intracellular protein that is inefficiently secreted by live Maf . We speculate that the ESAT-6 is largely released following Maf destruction by anti-TB treatment, as per previous demonstration of higher ESAT-6 content in Maf than Mtb culture lysates .
One of the most important finding of this study is the higher proportion of activated T cells subsets (CD3+CD11b+) in Maf compared with Mtb-infected patients after treatment but was not seen before treatment. These were mostly cytotoxic CD8+ T cells expressing CD56 as previously described , with significantly lower production of cytokines (IL-2, IFN-γ, and TNF-α) in Maf- compared with Mtb-infected patients posttreatment. This result is in agreement with a previous study, which describes that activation induced increased CD11b expression by CD3+ T cells result in inhibition of proliferation and reduced IL-2 secretion . Therefore, CD11b+ being a marker of T-cell activation and trafficking [14, 16], might indicate an unrecovered immune profile in Maf-infected patients, or incomplete clearance of bacilli despite treatment completion and negative AFB in sputum. In support of this, the analysis of the clinical outcome in correlation with stimulated whole blood supernatant cytokines showed a significantly improved profile in Mtb- than Maf-infected patients following treatment (Tientcheu et al., unpublished observations). Although most reports on host immune responses to TB have focused on CD4+, CD8+ T cells are important for long-term responses [33, 34]. Interestingly, in this study, CD8+ T cells showed the greatest differential responses between the groups before treatment suggesting that they might be more sensitive to changes in the bacteria load with treatment than CD4+ T cells as previously shown [25, 34].
The main limitation of this study was our inability to assess at the point of infection before progressing to disease, which could help identify correlates of protection. Mixed-infection by both strains that might affect the immune responses was not addressed because of lack of appropriate technology. Whole genome sequencing of mycobacteria grown directly from sputum is a very promising approach [35, 36]. Adjustment for multiple testing analyses was not applied on the p-values, although some might become nonsignificant after adjusting, it is unlikely that this would affect the consistent trend of the result we presented.
Although difference before treatment cannot be related reliably to host or pathogen factors, those following treatment when bacilli are reduced or cleared are likely to indicate intrinsic host factors contributing to resistance or susceptibility to disease. Therefore, the proportion of CD3+CD11b+ T cells posttreatment might be a potential surrogate marker of response to TB treatment that can be investigated on larger sample size.
The lower single-IL-2 and higher single-TNF-α producing CD4+ and CD8+ T cells in patients infected with the less virulent Maf compared with Mtb-infected patients before treatment and the persistently high activated T cells in Maf-infected patients posttreatment provide evidence for TB susceptibility that could be further investigated for their role in TB pathogenesis.
Materials and methods
Study participants were HIV-negative patients with no history of previous TB disease consecutively enrolled following written informed consent. All patients underwent a routine chest x-ray, sputum smear examination, and hematological analyses. Tuberculosis was confirmed following sputum smear Auramine O and Ziehl-Neelsen (ZN) microscopy for AFB and culture in mycobacteria growth indicator tubes (Becton-Dickinson). All cases took a tuberculin skin test with PPD RT23 (Statens Serum Institut, Copenhagen, Demark). Induration ≥10 mm after 48 h was considered positive. Basic demographic and anthropometric data were recorded at enrolment, and all took the standard treatment for TB administered by the Gambian government TB clinics . TB treatment in The Gambia is based on the standard 6 months regimen using four drugs: Isoniazid, Rifampicin, Ethambutol, and Pyrazinamide for the first two months followed by 4 months with Isoniazid and Rifampicin irrespective of patient's infecting strain under the direct observed treatment short-course strategy . Heparinized blood samples were collected at baseline and 6 months of treatment for immunological investigations. Ethical approval was obtained from the Gambian government/MRC joint ethics committee and the London School of Hygiene and Tropical Medicine ethics committee.
Following liquid culture confirmation, bacterial isolates were obtained and genomic DNA extracted using the boiled lysate method . The genotype of infecting bacteria was determined using standard spoligotyping analysis and assessing the presence or absence of lineage defining large sequence polymorphism RD702 and TbD1 as previously described [11, 40].
We used PHA-L (Sigma-Aldrich, UK; 5 μg/mL) as a positive control and medium alone as a negative control. Mtb antigens were purified protein derivative (Mtb-PPD; Staten Serum Institute, Denmark; 10 μg/mL) and a pool of peptides (15 mers overlapping by 10 amino acids) from ESAT-6 (17 peptides) and culture filtrate protein 10 (CFP-10; 18 peptides) (ESAT-6/CFP-10) at 2.5 μg/mL final concentration for each peptide (ProImmune). Mtb H37Rv and Maf GM041182 were used both live (final multiplicity of infection of approximately 1:2 (bacteria: monocytes)) and heat-killed (6 × 105 cfu/mL). The heat-killed bacterial suspension was sixfold higher than the live bacteria suspension.
Two hundred microliters of fresh whole blood were incubated with the antigens/whole bacteria in 5 mL polypropylene tubes (Dako, Carpinteria, CA) at 37°C, 5% CO2. After 2 h of incubation, Brefeldin A (eBioscience, UK at 1X or 3.0 μg/mL) was then added and the samples incubated overnight. The next day, 20 μL of surface antibody cocktail (anti-CD3 Cascade Yellow (Dako), anti-CD4 peridinin chlorophyll (PerCP), anti-CD8 Pacific Blue anti-CD11b allophycocyanin-cyanine7 (allophycocyanin-Cy7), and anti-CD56 phycoerythrin-cyanine7 (PE-C7); all from BD Pharmingen, USA) was added. Cells were vortexed and incubated at room temperature (RT) for 30 min in the dark. Two milliliters of 1× FACS lysing buffer (BD Pharmingen, USA) was then added to each tube, mixed, and incubated at RT, for 9 min in the dark; followed by 5 min centrifugation at 500 × g. The supernatant was removed and cells permeabilized by adding 500 μL of 1× FACS Perm 2 solution (BD Pharmingen), vortexed and incubated at RT, for 20 min, in the dark. Cells were then centrifuged at 600 × g, the supernatant was carefully removed from each tube and 20 μL of intracellular cytokine-specific antibodies (anti-IFN-γ allophycocyanin), anti-TNF-α FITC, and anti-IL-2 phycoerythrin (PE); all from BD Pharmingen, USA) was added. The tubes were vortexed and incubated for 30 min, at RT, in the dark. The cells were then washed with 1 mL of FACS buffer centrifuged and the cells resuspended in 4% paraformaldehyde prior to acquisition.
Cells were acquired using a 3-laser, 9-color CyAn ADP (Beckman Coulter, USA) flow cytometer. At least 150 000 lymphocytes were acquired, and analyzed using FlowJo software (v9.3 TreeStar, USA). Compensation was performed using single-stained anti-mouse Ig, κ beads (BD Bioscience, USA). Doublets were excluded using forward scatter area versus forward scatter height gating (Fig. 1). Guided by the fluorescence minus one controls, appropriate gates were applied to derive the percentages of total CD4+, CD8+, CD3+CD11b+, and CD11b+CD56+ T cells, and those producing intracellular IFN-γ, IL-2, and TNF-α. Different combinations of cytokine positive CD4+ and CD8+ T cells were determined by Boolean gating and data analyzed using Pestle v1.7 and Spice v5.22 software downloaded from http://exon.niaid.nih.gov/spice .
Background frequency of cytokine producing T cells was subtracted and values ≤ 0.01 were adjusted to zero. Maf and Mtb-infected patients were compared using a two-tailed nonparametric Mann–Whitney U test at each time point and Wilcoxon matched-pairs signed-rank test to evaluate differences before and after treatment within each group. Categorical data were compared between the groups using a Fisher's exact test. All statistical analysis was performed with STATA12 (StataCorp, USA). The cytokine combinations were analysed as stated above and the pie graphs were assessed using the permutation test in-built in SPICE v5.22. Statistical significance was considered at p < 0.05. The graphs were drawn using GraphPad Prism software v5.0 (Software MacKiev, GraphPad, San Diego, CA) and IntaglioTM v3.2 (Purgatory Design).
We would like to thank all study participants, field workers for sample collection, MRC TB clinic staff, the TB Immunology and TB diagnostic laboratory staff and the Gambian National Tuberculosis and Leprosy Control programme for continuing collaboration. This work was funded by the MRC Unit, The Gambia as a PhD fellowship awarded to L.D.T.
Conflict of interest
The authors declare no financial or commercial conflict of interest.