ESAT6-Induced IFNγ and CXCL9 Can Differentiate Severity of Tuberculosis

Background Protective responses against Mycobacterium tuberculosis are dependent on appropriate T cell and macrophage activation. Mycobacterial antigen six kDa early secreted antigenic target (ESAT6) and culture filtrate protein 10 (CFP10) can detect M. tuberculosis specific IFNγ responses. However, most studies have been performed in non-endemic regions and to study pulmonary tuberculosis (PTB). We have studied ESAT6 and CFP10 induced cytokine and chemokines responses in PTB and extrapulmonary (EPul) TB. Methodology IFNγ, IL10, CXCL9 and CCL2 responses were determined using an ex vivo whole blood assay system in PTB (n = 30) and EPulTB patients with limited (LNTB, n = 24) or severe (SevTB, n = 22) disease, and in healthy endemic controls (ECs). Responses to bacterial LPS were also determined. Principal Findings ESAT6- and CFP10-induced IFNγ was comparable between ECs and TB patients. Both ESAT6- and CFP10-induced IFNγ secretion was greater in LNTB than PTB. ESAT6-induced CXCL9 was greater in EPulTB as compared with PTB, with an increase in SevTB as compared with LNTB. CFP10-induced CCL2 was higher in PTB than LNTB patients. LPS-stimulated CXCL9 was greatest in SevTB and LPS-induced CCL2 was increased in PTB as compared with LNTB patients. A positive correlation between ESAT6-induced IFNγ and CXCL9 was present in all TB patients, but IFNγ and CCL2 was only correlated in LNTB. ESAT-induced CCL2 and CXCL9 were significantly associated in LNTB while correlation in response to LPS was only present in SevTB. Conclusions ESAT6 induced IFNγ and CXCL9 can differentiate between limited and severe TB infections.


Introduction
Protective immunity against Mycobacterium tuberculosis is dependent on the interplay between activated T cells, macrophages and other leucocytes, proinflammatory cytokines such as, interferon gamma (IFNc), tumor necrosis factor-alpha (TNFa) and interleukin (IL)-12, and down-modulatory cytokines IL-10 and transforming growth factor beta (TGF)-b [1,2]. IL10 produced by macrophages is important in regulating the Th1 cytokine balance and down regulates proinflammatory responses [3]. TNFa regulated granuloma formation relies on effective recruitment of leucocytes, dependent on chemokines such as the CC chemokines; CCL2, CCL3, CCL4, CCL5 and CXC chemokines; CXCL8, CXCL9 and CXCL10 [4][5][6][7]. Coordinate expression of these chemokines with their receptors is essential for granuloma formation. Deficiency of the CCL2 receptor, CCR2, has been shown to result in progressive infection by M. tuberculosis in mice [8]. CXCL9 stimulates T lymphocytes [9] and has previously been shown to be a sensitive predictive marker for antigen-specific IFNc production and IFNc secreting cells [10].
Rapid diagnosis of tuberculosis (TB) infections is essential for early detection and control of disease. Six kDa early secreted antigenic target (ESAT6) and culture filtrate protein 10 (CFP10) are both encoded by the region of difference 1 (RD1) which is present in M. tuberculosis and M. bovis, but absent from M. bovis BCG and most environmental mycobacteria [11,12]. ESAT6 is a immunodominant T cell-stimulatory antigen and is recognized by specific IFNc-secreting T cells present in greater numbers in patients with active disease as compared with those who are uninfected [13,14]. Commercial tests employing IFNc release assays (IGRA) such as, ESAT6 Quantiferon TB-2G [15,16] and T-SPOT TB are increasingly available as diagnostic and predictive tests for vaccination. Therefore, ESAT6 and CFP10 induced IFNc responses have been shown to be useful in discriminating infected individuals from healthy controls [17][18][19]. However, there are controversial reports from endemic regions. In TB endemic regions, IFNc can be modulated by natural exposure to M. tuberculosis or non-tuberculous mycobacteria as well as BCG vaccination [20,21]. Recent studies have shown that the sensitivity of IGRA may be related to the ethnic origin and age of the individual tested [22]. The sensitivity of ESAT6 and CFP10 induced Mycobacterium-specific T cell responses is greatest in a BCG unvaccinated population in a non-endemic region and most studies have been performed in areas of low tuberculosis (TB) transmission [23], with less data available from high transmission TB endemic regions [24]. A recent study has shown that the utility of using additional biomarkers such as CCL2 and CXCL10 together with IFNc to assess mycobacterial antigen induced responses in identifying patients with M. tuberculosis infection [25]. However, most research has been performed on patients with pulmonary tuberculosis (PTB). Although PTB is the predominant form of tuberculosis, extrapulmonary disease (EPulTB) involving; lymph nodes, skeletal system, abdomen, meningeal, miliary or disseminated forms of the disease remain common [26,27]. EPulTB presents an additional burden due to the problems associated with diagnosis and treatment of infections in extra-pulmonary sites. It is therefore essential to assess IFNc induced responses to mycobacterial antigens in endemic areas and in patients with PTB and EPulTB. This work has been carried out in Pakistan which is ranked 8 th amongst high TB burden countries and has an incidence of 181/100,000 cases/year, where 44% pulmonary and 15% extrapulmonary samples are listed under new case detection for DOTS [28].
The ex vivo whole blood assay model is useful in assessing host immune responses in tuberculosis patients [29] [30], despite its limitations in identifying the cellular sources of secreted products. We have investigated mycobacterium specific responses using antigens ESAT6 and CFP10 and also response to bacterial lipopolysaccharide (LPS) in whole blood cultures of patients with PTB (n = 30) and EPulTB (n = 46) with differing disease severity. Cytokines (IFNc and IL10) and chemokines (CXCL9 and CCL2) were measured in each donor. Our results indicate that ESAT6induced IFNc and CXCL9 responses do not discriminate between active TB patients and healthy endemic controls (ECs), but that they do differentiate severity of infections. Table 1 illustrates hematological characteristics of patients in the study. EPulTB patients were divided according to WHO ranking of severity into less severe TB (LNTB) or severe TB (SevTB) [28], see methods and Table 2. The study groups were comparable with respect to age and gender. As expected [27], erythrocyte sedimentation rate (ESR) was found to be raised in TB patients as compared with ECs; PTB, LNTB and SevTB (p,0.01, 0.01, 0.01, respectively). Total leucocyte (TLC) and monocyte counts were comparable between patients and controls. However, in PTB and SevTB, lymphocyte counts were decreased (p = 0.002, p,0.01, respectively), while neutrophil counts were increased (p = 0.02, p = 0.008, respectively).

Characteristics of study group
Cytokine and chemokine secretion in unstimulated whole blood cultures of TB patients and endemic controls IFNc, IL10 and CXCL9 and CCL2 secretion from unstimulated cells was measured in healthy endemic controls (ECs, n = 17) and in TB patients (n = 76, Table 3). IFNc, IL10 and CCL2 secretion from unstimulated whole blood cells of ECs and TB patients were comparable. However, CXCL9 secretion from unstimulated whole blood cells of ECs was greater than in TB patients (p = 0.031).
Mycobacterial antigens ESAT6-induced IFNc, IL10, CCL2 and CXCL9 in TB patients ESAT6-induced cytokine and chemokines responses were determined in whole blood cells of 76 TB patients. Upon stimulation by ESAT6, the trend of IFNc secretion at 5 day post-stimulation was greater in TB patients as compared with ECs but did not achieve statistical significance (Table 3). ESAT6induced IL10 secretion was also increased in TB patients but was not statistically significant as compared with ECs (p = 0.07).
CXCL9 has been shown to be induced by ESAT6 in patients with pulmonary TB [17]. We found ESAT6-induced CXCL9 to be significantly greater in TB patients as compared with ECs (p = 0.003). While, no difference was observed between ESAT6induced CCL2 in TB patients and ECs.

CFP-10 induced cytokine and chemokine responses
We also compared CFP10 induced whole blood cell responses in ECs and TB patients. CFP10-induced IFNc was increased in TB as compared with ECs but was not significantly different (Table 3). CFP10-induced IL10 (p = 0.002) secretion and also CXCL9 (p = 0.001) secretion was significantly raised in TB patients. However, there was no difference in CFP10-induced CCL2 between patients and controls. Although the trend of responses to ESAT6 and CFP10 were comparable, the magnitude of cytokine responses to CFP10 was approximately 2 fold greater than that induced by ESAT6.

LPS-induced cytokine and chemokine secretion in TB patients
To investigate whether the differences in cytokine secretion observed were specific to mycobacterial antigens, we also studied cytokine and chemokine responses to bacterial lipopolysaccharide (LPS), a strong activator of macrophages. LPS-induced IL10 was not different between TB patients and ECs, Table 3. However, LPSinduced CXCL9 was significantly greater (p = 0.008), while CCL2 was reduced (p,0.01) in TB patients as compared with ECs.

Differential ESAT6-induced IFNc in pulmonary and extrapulmonary TB
We further investigated IFNc responses to mycobacterial antigens in TB patients with pulmonary and extrapulmonary disease of differing severity. Pulmonary TB patients included in    the study comprised those with minimal (n = 6) and moderate (n = 24) disease. However, we found antigen stimulated IFNc secretion in minimal and moderate PTB patients to be comparable between groups (data not shown) and have therefore performed further analysis on the combined minimal/ moderate PTB group. Responses in PTB patients were compared with those of EPul-TB; limited (LNTB) or severe (SevTB). As illustrated in Fig. 1A, ESAT6-induced IFNc in LNTB (median, 253 pg/ml) was significantly greater than in PTB (median, 3 pg/ml, p = 0.011). IFNc responses in LNTB were significantly greater as compared with SevTB (median, 1 pg/ml, p = 0.029). ESAT6-induced IL10 was negligible in all TB groups (PTB, LNTB, SevTB; 0, 3, 3 pg/ ml respectively, data not shown).

ESAT6-induced responses
The balance between proinflammatory IFNc and downmodulatory IL10 has been shown to be critical in determining outcome of mycobacterial infections [31], and the IFNc (5d) /IL10 (2d) ratio induced in response to mycobacterial antigens has been shown to be to be an indicator of disease severity in tuberculosis [32]. We found the ESAT6-induced IFNc/IL10 ratio to be significantly greater in LNTB (median, 3.5) as compared with the SevTB group (median, 1; p = 0.011, data not shown), consistent with these earlier reports.
Coincident with results observed in response to ESAT6, CFP10induced IFNc /IL10 ratio was also greater in LNTB patients (median, 35) as compared with those with SevTB (median, 0.1; p = 0.001, data not shown).
No differences were observed in CFP10-induced CXCL9 in PTB (51 pg/ml) and EPulTB groups, although CFP10 induced a greater trend of CXCL9 secretion in LNTB and SevTB (543, 192 pg/ml, respectively, data not shown). As observed in the case of IFNc, CFP10-induced CXCL9 was 2 fold greater in magnitude as compared with that in response to ESAT6.

LPS-induced CXCL9 and CCL2 is greater in severe as compared with limited TB
When LPS-induced CXCL9 was measured in PTB, LNTB and SevTB groups it was observed that chemokine levels were significantly greater in SevTB as compared with PTB and LNTB (p = 0.001, 0.004, respectively), Fig. 3A. In contrast, LPS-induced CCL2 was greater in PTB as compared with LNTB (p = 0.01) and also in SevTB patients as compared with LNTB (p = 0.003), Fig. 3B.

Relationship between CXCL9, CCL2 and IFNc secretion in tuberculosis patients
We next determined the association between cytokine and chemokine responses induced to mycobacterial antigen ESAT6 and to LPS in different TB patient groups. A Spearman's rank correlation analysis showed ESAT6-induced secretion of IFNc and CXCL9 to be positively associated in all TB groups (Table 4). A significant association between ESAT6-induced IFNc and CCL2 (R = 0.471, p = 0.027) and between CCL2 and CXCL9 (R = 0.57, p = 0.006) was observed only in the LNTB group, data not shown. No association was found between LPS-induced IL10 and either CXCL9 or CCL2. However, LPS-induced CXCL9 and CCL2 were found to be positively associated in patients with SevTB but not in PTB or LNTB groups (Table 4).

Discussion
The relationship between cytokines, IFNc and IL10 in regulating the outcome of M. tuberculosis infections is well studied. We have recently shown that circulating levels of IFNc and CXCL9 differ in patients with pulmonary and extrapulmonary disease [33]. We show here that the relationship between mycobacterial antigen induced IFNc and CXCL9 may play a role in determining disease severity in tuberculosis.
Our study illustrates that in a TB endemic region, ESAT6induced IFNc responses may not be as effective in distinguishing patients with active disease from healthy individuals in endemic regions. This may be attributed to immune modulation of IFNc responses to specific M. tuberculosis antigens as has been shown in a recent study of household contacts of TB patients [31]. Also, a high rate of transmission in endemic regions may result in modulation of IFNc and IL10 responses due to increased exposure to environmental mycobacteria [21,34].  Amongst healthy controls we studied asymptomatic individuals who were both TST2 and TST+. However, we did not observe any difference between cytokine or chemokine induced responses from whole blood cells of either group either after measurement from unstimulated cells or from ESAT6 or CFP10 stimulated cells (data not shown), corresponding with previous reports [35]. However, when both EC groups; TST2 and TST+ donors were compared with TB patients, a significant differences in ESAT6and CFP10-induced CXCL9 and CCL2 secretion was found in TB patients as compared with TST2 individuals but not in the TST+ group (data not shown). This may be attributed to latent infection in the TST+ controls which would raise specific cytokine responses to mycobacterial antigens. Previous reports have shown Th1-like effector memory cells to be increased in bronchoalveolar cells from TST+ individuals as compared with TST2 individuals, with coordinate increase in IFNc and IFNc regulated chemokines [36]. Therefore, in our study we have compared responses of TB patients with Ecs who are TST2 in order to have the least exposed group as a control.
In all cases, antigen-induced IFNc secretion increased in magnitude from 2 to 5 days post-stimulation (data not shown). This may be attributed to the differing cellular sources of IFNc whereby at the earlier time point there would be a greater contribution from effector T cells, NK cells, cd T cells and activated macrophages, while at the latter time point the IFNc may primarily be released from mycobacterium specific T memory cells.
We found ESAT6-induced IFNc response to be decreased in patients with severe disseminated TB, comprising of those with spinal, meningeal, abdominal, miliary involvement, as compared with localised disease. Tuberculous lymphadenitis is the most common presentation of extrapulmonary TB. It is often a selflimiting disease and is less severe as compared with other forms. Increased IFNc responses in less severe TB infections correlates with previous work which has shown the frequency of antigen specific CD4 T cells to be greater in latently infected individuals and in those with minimal disease with a low bacterial burden as compared with patients with a higher bacterial load [37]. Therefore, patients with limited extrapulmonary disease may be more likely to have an increased number of IFNc secreting T cells as compared with those with progressive disease [37]. Previous reports haves shown that in pulmonary TB, there is a decrease in IFNc responses related to increased severity of disease [38,39].
Our data corresponds with previous work which has shown that differences in the severity of extrapulmonary TB can be defined by the IFNc: IL10 ratio to mycobacterial antigens, with a direct relationship between disseminated disease or disease localized to peripheral sites without lung involvement [32,40].
CXCL9 is IFNc induced and has been shown to be an early marker of IFNc activation and of T cells secreting IFNc [10]. CXCL9 is a chemotactic ligand produced by macrophages and other antigen presenting cells at the site of inflammation [9]. We found ESAT6-induced CXCL9 to be increased in TB patients, corresponding with previous reports [17]. This corresponds with reports by Gonzalez-Juarrero et al. who show upregulated expression of IFNc regulated chemokines transcripts in response to M. tuberculosis coordinate with inflammation in the mouse model [41]. The increase in CXCL9 observed in severe extrapulmonary disease corresponds with literature which have shown chemokines responses to be increased in severe TB such as in tuberculous meningitis [42,43]. CFP10-induced activation of CXCL9 was greater than that observed for ESAT6 just as observed in the case of IFNc. Although the trend of responses to ESAT6 and CFP10 was comparable, ESAT6 appeared to elicit more significant difference in responses between limited and severe TB. This may be due to differential recognition of cell surface receptors for the recombinant antigens on T cells which may determine the magnitude of IFNc response and subsequently impact macrophage activation.
ESAT6 stimulated secretion of induced-CCL2 was not different between TB groups, further indicating the utility of this antigen in eliciting T cell specific responses.
LPS is a strong stimulant of monocyte responses, and both CXCL9 and CCL2 are produced by monocytes. Raised LPSinduced CXCL9 in severe extrapulmonary TB may be indicative of the increased inflammatory processes ongoing in progressive, disseminated disease and an absence of regulation. CCL2 has previously been shown to be increased in PTB patients as compared with Epul-TB patients and our results with LPS support this. In addition, we observe that as in case with CXCL9, CCL2 levels are raised in severe TB as compared with limited disease in patients with extrapulmonary TB.
Our correlation analysis between the cytokines and chemokines produced by whole blood cells of TB patients indicates differential regulation of cytokines in response to mycobacterial antigen stimulation in patients with differing disease severity. A positive association was found between ESAT6 induced IFNc and CXCL9 in all TB patients indicating that this balance may play a predictive role in the Th1-like cytokine balance required for protection against M. tuberculosis. However, as association between the macrophage activating CCL2 and the T cell activation IFNc and CXCL9 was only found in LNTB patients. When LPSinduced responses were investigated, there appeared to be no association between CCL2 and CXCL9 in limited disease (PTB and LNTB), but a correlation between these two chemokines was present in severe extrapulmonary TB. It is possible that in severe extrapulmonary disease where there is an overall reduction in IFNc possibly due to T cell anergy, CXCL9 secretion may lead to increased macrophage recruitment and inflammation, the lack of association between the chemokines and T cell regulation may result in progressive disease related pathology in the host. Overall, this work proposes differential chemokine regulation in limited and severe tuberculosis infections which may explain subsequent pathophysiological changes in the host. Further investigations into the mechanisms responsible for the control of macrophage activation in different disease states may further explain this phenomenon.

Ethics statement
This study was conducted according to the principles expressed in the Declaration of Helsinki. The study was approved by the Institutional Review Board of Aga Khan University hospitals. All patients provided written informed consent for the collection of samples and subsequent analysis.

Subject selection
Seventy six patients were recruited from the Aga Khan University Hospital and Medical College (AKUH) and Masoomeen Hospital, Karachi. All study subjects were examined and evaluated by infectious diseases consultants. Patients were newly diagnosed or had taken less than 7 days of anti-tuberculous therapy (ATT). Patients with significant co-morbid conditions including diabetes mellitus, chronic renal failure, chronic liver disease, patients on high dose corticosteroid therapy were excluded to assure relatively unmodulated immunological parameters. All patients were tested and found to be HIV negative.
Patients with pulmonary TB (PTB, n = 30) were diagnosed by clinical examination, chest X-ray, sputum acid fast bacillus (AFB) Zeihl Neelson staining, AFB culture and/ or clinical response to treatment (along with fever, cough and weight loss). All PTB patients had findings consistent with active TB as evaluated by one of the consulting physicians. Fifteen patients were found to have a positive AFB smear microscopy. Nine PTB patients had an AFB negative smear, but showed a favourable clinical response to ATT. Patients were diagnosed as having minimal, moderate or advanced pulmonary tuberculosis using a modified classification of the National Tuberculosis Association of the USA based on extent of lung tissue involvement [44,45]. Of the PTB patients 6 had minimal, while 24 had moderate disease.
Diagnosis of TB lymphadenitis, spinal, abdominal and genitourinary TB was based on histopathological findings of chronic granulomatous inflammation with caseous necrosis, AFB staining and culture and supportive radiological evidence on CT or MRI. Diagnosis of meningeal TB was based on CSF biochemical findings, supported by AFB culture and findings on contrastenhanced CT and/or MRI. Pleural TB was diagnosed on the basis of pleural fluid biochemical findings, AFB culture, histopathological findings on pleural biopsy and supportive radiological evidence on X-rays and/or contrast-enhanced CT scan Patients with extrapulmonary TB (Epul-TB) were divided into two groups according to the WHO ranking of diseases severity based on bacillary load, extent of disease and anatomical site [28]. Epul-TB patients were subsequently classified into less severe (LNTB) or severe (SevTB) disease. Of the 24 patients with LNTB, 23 had lymphadenitis and 1 had genitourinary TB. All LNTB patients were confirmed on histological findings consistent with tuberculosis. Twenty two patients had SevTB and their diagnostic criteria are provided in Table 2.
BCG-vaccinated asymptomatic healthy volunteers who were staff at AKU with no known exposure to TB were used as endemic controls (Ecs). All volunteers had a normal chest X-Ray. Tuberculin skin testing (TST) was assessed by administering five tuberculin units on the volar surface of the right arm subcutane-ously, and read by a single reader at 48 h. An induration of $10 mm was used as a cutoff for positive responses and TST2 (n = 17) Ecs were included in the study.

Reagents
Recombinant antigens ESAT6 and CFP10 were provided through the TB Vaccine Testing and Research Materials Contract, NIH, NIAID NO1-A1-40091, awarded to Colorado State University, USA.

Whole blood assay
Venous blood was diluted 1:10 in RPMI-1640 medium and 200 ml set up per well in a 96 well tissue culture plate as described previously [29]. Whole blood cells were stimulated with recombinant antigens ESAT-6 (5 mg/ml) and CFP-10 (5 mg/ml) and cultured for up to 5 days. All samples were set up in replicates. Supernatants were collected at 2 and 5 day post-stimulation for cytokine and chemokine measurements. Samples were centrifuged to collect any cellular debris, aliquoted and stored at -70 C until tested. Cytokines were measured at both 2 and 5 days poststimulation. We have presented IFNc responses from whole blood cells at 5 day of culture to focus on T cell responses [21,29]. IL10 increased only marginally from 2 to 5 day post-stimulation and there was no significant increase between the two times studied (data not shown). Therefore, we have further analyzed IL10 responses at 2 day post-stimulation. CCL2 and CXCL9 were optimally measured at 2 days post-stimulation.

ELISA for IFNc, IL-10 and CXCL9
IFNc and IL-10 was detected in cellular supernatants by using standards and ELISA reagents obtained from Endogen (Rockford, IL, USA). Cytokines were measured using a sandwich ELISA technique according to the manufacturer's instructions and as reported previously [30]. Recombinant human cytokine was used to obtain a dose response curve with a range of detection from 3.9-1000 pg/ml. All experimental samples were tested in duplicate.
CXCL9 and CCL2 standards and monoclonal antibody pairs for capture and detection were obtained from R&D Systems (Abingdon, UK). All measurements were carried out according to the manufacturer's recommendations and as described previously [30]. Recombinant human cytokine was used to obtain a dose response curve with a range of detection from 6.25-500 pg/ml for CXCL9 and 6.25-1000 pg/ml for CCL2.

Statistical analysis
Data is represented by box blots illustrating 25 th and 75 th quartiles with the median shown as a horizontal bar. Nonparametric statistical analysis was performed using Kruskal-Wallis and Mann-Whitney U tests as required. P values#0.05 were considered to be statistically significant. Spearman's Rank Correlation was also performed, all using the Statistical Package for Social Sciences software (SPSS).