Cytokine response to selected MTB antigens in Ghanaian TB patients, before and at 2 weeks of anti-TB therapy is characterized by high expression of IFN-γ and Granzyme B and inter- individual variation

Despite the availability since the 1980′s of an inexpensive, efficacious, and moderately well-tolerated therapy, that can cure 90% of cases, tuberculosis (TB) continues to be a major worldwide health problem, with 8.6 million new cases and 1.3 million deaths in 2012 [1]. One of the factors mitigating against a decline in tuberculosis cases is the long duration of treatment (6 months) which makes adherence difficult and also puts pressure on health care systems in developing countries [2]. Clinical, bacteriological and radiological improvement in TB patients is achieved within 2 months upon effective chemotherapy with a multi-drug regimen [3], making “month 2 sputum culture status” the only acceptable biomarker for TB treatment response [4]. However, waiting for 2 months to reach a decision on the treatment outcome could be damaging to a patient as it unduly delays the need for drug sensitivity testing or changes in treatment regimes. During this 2-month period, primary multi-drug resistant organisms will remain untreated and drug-resistant mycobacteria may have time to develop resistance to additional drugs [5].

Additionally, use of classical microbiological methods like “sputum culture status” as a biomarker for TB treatment response has limited usefulness in children and extra-pulmonary cases, where appropriate quality sputum samples are hard to get. Given the shortcomings associated with using sputum as a sample for monitoring TB treatment response, blood has been proposed as a better alternative. Blood is easier to obtain and assessment of immunological parameters in blood can be done within days and could be well adapted for field use. Measurement of Mtb-specific antigen induced responses in peripheral blood has been used often in recent times, however, most studies using this approach have only focused on IFN-γ as a sole marker of response. It has been proposed that sets of markers rather than a single marker may increase the predictive power [2]. Multiple cytokines assessed at week two, a much earlier time point during treatment than the current stipulated “month two”, could be useful in detecting early cellular response to TB treatment. There has been a long held belief that patients with drug-susceptible TB are non-infectious after two weeks of therapy [6, 7] although recent microbiological and epidemiological evidence has challenged this dogma [8, 7, 9‐11]. However, the nature of the Mtb-specific immunological response during this period of TB treatment has not been adequately investigated. It has been demonstrated that Acid fast bacilli (AFB) counts fall by about 20-fold in the first 2 days and by a further 200-fold in the next 12 days to reduce the counts of an initially smear-positive patient to about 10³ per ml at 2 weeks of short course chemotherapy [12]. These levels are below the estimates of 10^3.5 to 10⁴ per ml which are the limits indicating a change from smear-positive to smear-negative, culture-positive in untreated patients [7]. If and whether this is characterized by an improvement in cellular response can potentially be exploited as a possible tool to predict early treatment response. To address this knowledge gap, we identified, in addition to IFN-γ, 5 key cytokines; Granzyme B (grzB), TNF-α, IL-17, IL-10 and sIL-2R-α most of which are important to early control and containment of tuberculosis infection.

Using the Luminex platform, we measured the levels of the 6 selected host markers in the PBMC culture supernatant of TB patients after 6 days long term stimulation with ESAT-6/CFP-10 and 3 DosR antigens (Rv1733c, Rv2029c, Rv2628) known to be immunogenic in African populations [13] before and after 2 weeks of treatment. A profile consisting of high levels (pg/ml) of IFN-γ followed by Granzyme B and TNF-α and low levels of IL-17, IL-10 and sIL-2R-α in PBMC before treatment and at 2 weeks of treatment was observed. We show that there is general improvement in cellular response at 2 weeks of therapy but there is wide inter-individual variation, suggesting that factors other than anti TB therapy account for cellular response observed. These ought to be investigated to further explore the potential to use cytokine response patterns for monitoring treatment response.

Cytokine secretion profile in response to mycobacterial antigens has been extensively studied in identifying immunological differences between active TB and latent infection or in TB patients before and after treatment. The first two weeks of TB treatment are known for the rapid reduction of metabolically active mycobacteria, yet studies on changes in the host immunological response during this period are scanty. Such changes could be exploited in the design of putative biomarkers for early treatment response. Determining the cytokine secretion profile of Mtb-specific antigens in different settings would also expand our knowledge of the performance of potential immunogenic M. tuberculosis antigens which could be worthy candidates for designing new diagnostics or vaccines [17]. In this study, the four tested MTB antigens were immunogenic in our cohort as evidenced by the highly significant differences (P < 0.01 to P < 0.0001) seen between stimulated and un-stimulated cultures.

Consistent with other studies, [13, 18, 19] cytokine levels in response to ESAT-6/CFP-10 fusion protein stimulation were generally high. ESAT-6 is a virulence factor secreted by actively replicating bacteria early during infection, so high responses in TB patients at the early stage of disease at are not unusual. However, cytokines were secreted in response to the latency associated antigens as well. The recognition of the latency-associated antigens by PBMC from active TB patients could reflect the fact that most TB patients undergo a latent infection prior to TB disease [20] or indicate the simultaneous involvement of latent and metabolically active bacteria in active disease. Thus an individual could harbour actively replicating bacteria and succumb to TB, while having latent foci that will lead to recognition of latency related antigens. The immunodominance of Rv1733 across populations was affirmed as in agreement with other studies, [13, 18], Rv1733 was the most recognized (of the 3 DosR antigens used) in our cohort in terms of positive cytokine responders and frequency of IFN-γ + T cell response.

The cytokine profile in response to all antigenic stimulation consisted of high levels of IFN-γ followed by grzB and TNF-α and lower levels of IL-17, sIL2Rα and IL-10 (Figure 1). It would be expected that ESAT-6/CFP-10 being virulent factors would induce both pro and anti inflammatory cytokines whilst the dormancy -related Rv1733, may only induce pro inflammatory responses and relatively little anti-inflammatory response. This actually confirms that the immune responses during the active stage of the disease (TB patients) is characterized by both pro and anti-inflammatory cytokines irrespective of the nature of the antigen. While the high IFN-γ levels were expected due to its critical role in the control of Mtb infection, [21‐23] the higher levels of Granzyme B in this study in response to all antigens was an interesting finding. In a previous study by Toosii et al. [24] IFN-y and grzB were the only Mtb effector molecules that were inducible in PBMC from Mtb-sensitized subjects. In Mtb infections, the combined action of perforin and the antibacterial agent granulysin, both of which are expressed in the granules of CTLs and NK cells, influences the outcome of infection [25]. Studies using mouse models indicate that the majority of Mtb-specific CD8 + T cells are limited to either cytotoxicity or the secretion of IFN-γ, [26]. High levels of grzB could indicate that the early stage of disease control is characterized by lysis of infected cells to quell the spread of infection and re-enforces the role of CD8+ T cells in the control of human tuberculosis infection. This was evident as we saw a higher frequency of CD8+ T cells than CD4+ at baseline and a decline in CD8+ T cells at week two and an increase in IFN-γ + CD4 T cells. Marginal increases in cytokine levels from baseline to week two as well as a significant increase in the frequencies of CD4+ T cells was observed, suggestive of an improvement in cellular response with therapy. However, only the increase in median Rv1733- specific Granzyme B secretion was of statistical significance. Although effect of time point and antigen on cytokine secretion could be observed most of the trends were non-significant. It could be because the sample size was too small to show any differences or cytokines assessed remain relatively unchanged or stable from baseline to week two of treatment. Improvement in cellular response with therapy has been reported in previous studies [27‐30].

Two reasons have been adduced for this cellular improvement, firstly an increase in the number of peripheral CD4+ T cells that produce IFN-γ, owing to the fact that CD4+ T cells responsive to a vast array of Mtb epitopes are sequestered or compartmentalized at the site of the disease, and appear in the peripheral blood after effective chemotherapy, which reverses the state of anergy seen in these patients [27, 30]. This is actually corroborated by our study as the median frequency of IFN-γ + CD4 T cells increased during week 2 from baseline levels. Secondly a shift in cytokine production by PBMC from cytokines that down regulates the activation of Th1 cells and their cytokines such as IL-10. The levels of these regulatory cytokines are high in active TB patients and decrease upon treatment with anti-TB drugs [31, 32].

In-spite of this general improvement in cytokine responses, inter-individual variation was observed with 3 distinct patterns of Increased, decreased or fluctuating levels of all cytokines. We could not detect an association between “age” and “sputum smear result at diagnosis” with cytokine response to antigenic stimulation. The variation could be best explained by host intrinsic factors beyond the scope of this study.

This study had some limitations, including the small sample size (n = 20), and limited number of analytes tested for (6 cytokines). Also, we could not take sputum samples at week 2 which would have permitted us to directly compare the changes in the cytokine profile to smear status (bacterial load) at week 2 and perhaps would have helped to better explain the individual variation in response. Also, epidemiological factors were not considered.

In conclusion, effective chemotherapy improved cellular responses of TB patients to Mtb stage specific antigens, as early as two weeks of anti-TB therapy, characterized by a general increase in secretion of various cytokines. However, using cytokine levels to predict early treatment response to anti-TB treatment will be difficult due to the wide inter-individual variation observed. While the amount of cytokines produced may not be an indication of its biological activities, the high quantities of Granzyme B warrants further investigation of its role in the control of TB infection and the possible utility as a marker for early TB treatment response.