Mycobacterial antigen-induced T helper type 1 (Th1) and Th2 reactivity of peripheral blood mononuclear cells from diabetic and non-diabetic tuberculosis patients and Mycobacterium bovis bacilli Calmette–Guérin (BCG)-vaccinated healthy subjects

Abstract

Patients with diabetes mellitus are more susceptible to tuberculosis (TB), and the clinical conditions of diabetic TB patients deteriorate faster than non-diabetic TB patients, but the immunological basis for this phenomenon is not understood clearly. Given the role of cell-mediated immunity (CMI) in providing protection against TB, we investigated whether CMI responses in diabetic TB patients are compromised. Peripheral blood mononuclear cells (PBMC) obtained from diabetic TB patients, non-diabetic TB patients and Mycobacterium bovis bacilli Calmette–Guérin (BCG)-vaccinated healthy subjects were cultured in the presence of complex mycobacterial antigens and pools of M. tuberculosis regions of difference (RD)1, RD4, RD6 and RD10 peptides. The PBMC were assessed for antigen-induced cell proliferation and secretion of T helper 1 (Th1) [interferon (IFN)-γ, interleukin (IL)-2, tumour necrosis factor (TNF)-β], and Th2 (IL-4, IL-5, IL-10) cytokines as CMI parameters. All the complex mycobacterial antigens and RD1_pool stimulated strong proliferation of PBMC of all groups, except moderate responses to RD1_pool in healthy subjects. In response to complex mycobacterial antigens, both IFN-γ and TNF-β were secreted by PBMC of all groups whereas diabetic TB patients secreted IL-10 with concentrations higher than the other two groups. Furthermore, in response to RD peptides, IFN-γ and IL-10 were secreted by PBMC of diabetic TB patients only. The analyses of data in relation to relative cytokine concentrations showed that diabetic TB patients had lower Th1 : Th2 cytokines ratios, and a higher Th2 bias. The results demonstrate a shift towards Th2 bias in diabetic TB patients which may explain, at least in part, a faster deterioration in their clinical conditions.

Introduction

Tuberculosis (TB) and diabetes are among the most important diseases of global concern. TB ranks among the top 10 causes of world mortality [1]. It is estimated that approximately one-third of the world's population is infected with Mycobacterium tuberculosis; 8 million people develop active disease, leading to about 2 million deaths each year [1]. With respect to diabetes, the statistics released by the World Health Organization and the International Diabetes Federation shows that the number of diabetics in the world is expected to increase from 194 million in 2003 to 330 million in 2030, with 75% diabetic patients living in developing countries [2].

In addition to being two independent major global health problems of infectious and non-infectious nature, there is a strong association between TB and diabetes [3]. Patients with diabetes are more susceptible to bacterial infections, including pulmonary TB, and both types 1 and 2 diabetes are linked to TB [3]. The prevalence of TB is two to five times higher in diabetic patients than in non-diabetic patients and the disease is more aggressive in poorly controlled diabetes [4,5,6]. In diabetic patients, pulmonary TB may progress rapidly and thus a high index of suspicion is required [7]. Before the discovery of insulin a diagnosis of diabetes was a death sentence within 5 years, and the usual cause of that death was TB [3]. Although the frequent association of diabetes mellitus and TB has been observed for more than 2000 years, the reasons for this association are largely unknown [3].

It is now well established that cellular immune responses mediated by T cells and macrophages play the major role in defence against TB [8]. The few studies that investigated the cellular immunity in diabetic TB patients have shown that, compared to non-diabetic TB patients, diabetic TB patients had depressed cellular immunity as evidenced by a decrease in the number of T cells and their ability to proliferate in vitro in response to antigenic stimuli using a T cell transformation test as indicator [9]. These immunological parameters became normal following successful therapies [9]. In addition, macrophage functions including bacterial phagocytic capacity, resistance to tubercle bacilli and T cell function in delayed type hypersensitivity, were also depressed in mice with experimentally induced diabetes [10]. These studies suggest that the functions of both T cells and macrophages are compromised in diabetes.

Resistance to mycobacterial infections is mediated largely by T helper type 1 (Th1) cells and their cytokines, whereas Th2 cells and their cytokines correlate with disease susceptibility and pathology in TB [11–14]. In particular, the Th1 cytokine interferon (IFN)-γ is considered a principal mediator of protective immunity against TB [11,13]. Individuals with mutations in IFN-γ receptors or IFN-γ intracellular pathway components have experienced fatal disseminated mycobacterial infections during childhood [15], and inclusion of aerosolized IFN-γ in the treatment regimen of multi-drug-resistant pulmonary TB patients cured the disease [16]. In addition, the secretion of IFN-γ in response to M. tuberculosis-specific region of difference (RD)1 antigens has been shown to have diagnostic significance [8,17–19]. Moreover, Th1 cells also produce tumour necrosis factor (TNF)-β (lymphotoxin-alpha), a cytokine that is essential for the control of pulmonary TB [20]. TNF-β has a critical role in the activation of T cells and macrophages and local organization of the granulomatous response. On the other hand, the Th2 responses, characterized by the secretion of interleukin (IL)-4, IL-5 and IL-10, are associated with the lack of protection in TB [12,13]. In particular, IL-10 is associated with reduced resistance and chronic progressive TB [14]. Furthermore, IL-10 deactivates macrophages and down-regulates the secretion of Th1 cytokines [14], and some human populations exhibit increased expression of IL-10, which correlates with ineffective bacilli Calmette–Guérin (BCG) vaccination [21]. In addition, in vitro secretion of IL-4 in response to M. tuberculosis by CD8⁺ and γ/δ T cells of health care workers was associated with the development of progressive TB within 2–4 years [22].

In relation to diabetes mellitus, impaired host resistance against M. tuberculosis infection in diabetic mice is associated with reduced production of Th1-related cytokines IL-12 and IFN-γ and impaired development of Th1 cells [23]. In humans, conflicting reports have appeared, which suggest no difference [24], and lower [4] or higher production [25] of IFN-γ by diabetic than non-diabetic TB patients. However, to our knowledge, a systematic analysis of cell-mediated immunity (CMI) responses in diabetic TB patients, and in particular the profile of Th1 and Th2 cytokines in response to M. tuberculosis-specific RD antigens, is lacking in the published literature.

In this study, to explore further the immunological mechanism of reduced resistance to TB in diabetic patients, we have studied CMI parameters that included in vitro cell proliferation and secretion of a battery of Th1 (IFN-γ, IL-2, TNF-β), Th2 (IL-4, IL-5, IL-10) cytokines by peripheral blood mononuclear cells (PBMC) obtained from diabetic and non-diabetic pulmonary TB patients and M. bovis BCG-vaccinated healthy subjects. PBMC were cultured in the absence and presence of complex mycobacterial antigens (whole-cell M. bovis BCG and whole cells, culture filtrate and cell walls of M. tuberculosis), and pools of synthetic peptides corresponding to M. tuberculosis-specific open reading frames (ORFs) of RD1, RD4, RD6 and RD10 that are deleted in all vaccine strains of M. bovis BCG [26,27]. The CMI responses were determined by evaluating antigen-induced cell proliferation and the secretion of Th1 (IFN-γ, IL-2, TNF-β), Th2 (IL-4, IL-5, IL-10) cytokines and the responses were compared among the three groups of donors.

Materials and methods

Donor groups

Heparinized venous blood was collected from newly diagnosed and culture-confirmed non-diabetic pulmonary TB patients (n = 18) and those suffering from pulmonary TB and type II diabetes (n = 11) attending the Chest Diseases Hospital, Kuwait. In addition, buffy coats were obtained from M. bovis BCG-vaccinated and purified protein derivative (PPD)-positive healthy subjects (n = 20) donating blood at the Central Blood Bank, Kuwait. All the patients and healthy subjects were residing in Kuwait at the time of the study. The patients were diagnosed on the basis of clinical and radiological features and positive cultures for M. tuberculosis. TB patients were diagnosed for having diabetes mellitus based on two blood glucose measurements in accordance with World Health Organization (WHO) criteria. The diabetic patients had uncontrolled diabetes for a period of 1–10 years. In addition, quality of glycaemic control was assessed by measuring haemoglobin A1c (HbA1c) levels in the peripheral blood. Demographic data of the three study groups are detailed in Table 1. The groups of healthy donors, non-diabetic TB patients and diabetic TB patients were all serologically negative for HIV infection. Informed consent was obtained from all the subjects and the study was approved by the Ethical Committee of the Faculty of Medicine, Kuwait University, Kuwait.

Table 1.

Demographic and clinical data of non-diabetic tuberculosis (TB) patients, diabetic TB patients and bacille Calmette–Guérin (BCG)-vaccinated healthy subjects.

Parameter	Diabetic TB patients	Non-diabetic TB patients	BCG-vaccinated healthy subjects
Ethnicity [no. (%) of patients]
Kuwaiti	0 (0·0)	0 (0·0)	2 (10)
Indian subcontinent	6 (54·5)	8 (44·4)	7 (35)
Asian	5 (45·5)	10 (55·6)	11 (55)
Age range (years)	20–57	28–57	20–47
Sex (male : female)	10:1	17:1	18:2

Complex antigens and synthetic peptides of M. tuberculosis

The complex mycobacterial antigens used in this study were irradiated whole cell M. tuberculosis H37Rv and M. bovis BCG [28], and M. tuberculosis culture filtrate (MT-CF) and cell walls (MT-CW) (provided by J. T. Belisle, Fort Collins, CO, USA and produced under NIH contract HHSN266200400091C/ADB contract NO-AI40092, ‘Tuberculosis Vaccine Testing and Research Materials Contract’). A total of 607 peptides (25-mers overlapping with neighbouring peptides by 10 aa) spanning the sequence of putative proteins encoded by genes in RD1, RD4, RD6 and RD10 genomic regions were designed based on the amino acid sequence deduced from the nucleotide sequence of the respective genes (Table 2). These peptides were synthesized commercially by Thermo Hybaid GmBH (Ulm, Germany) using fluonerylmethoxycarbonyl chemistry, as described previously [29]. The stock concentrations (5 mg/ml) of the peptides were prepared in normal saline (0·9%) by vigorous pipetting, and the working concentrations were prepared by further dilution in tissue culture medium RPMI-1640, as described previously [30].

Table 2.

Synthetic peptides corresponding to open reading frames (ORFs) predicted in the regions of difference (RDs) present in M. tuberculosis H37Rv but absent in all strains of Mycobacterium bovis bacille Calmette–Guérin (BCG).

Region deleted*	ORF designation†	No. of ORFs	No. of synthetic peptides
RD1	ORF2-ORF11, ORF14, ORF15	12	220
RD4	Rv0221-Rv0223c	3	80
RD6	Rv1506c-Rv1516c	11	236
RD10	Rv1255c-Rv1257c	3	71
Total		29	607

^*RD designations are according to Behr et al.[27].

^†ORFs designations for RD1 are according to Amoudy et al.[26], whereas for other RDs, the ORF designations are according to Behr et al.[27].

Isolation of PBMC and antigenic stimulation

PBMC were isolated from blood of TB patients and healthy subjects by flotation on Lymphoprep gradients using standard procedures [28,29]. The cells were finally suspended in complete tissue culture medium [RPMI-1640 + 10% human AB serum + penicillin (100 U/ml) + streptomycin (100 µg/ml) + gentamycin (40 µg/ml) + fungizone (2·5 µg/ml)] and counted in a Coulter Counter (Coulter Electronics Ltd, Luton, UK).

Antigenic stimulation of PBMC was performed according to standard procedures [28,31]. In brief, 2 × 10⁵ PBMC suspended in 50 µl were seeded into the wells of 96-well tissue culture plates (Nunc, Roskilde, Denmark) and stimulated with complex mycobacterial antigens and peptide pools of RD1, RD4, RD6 and RD10, as described previously [31,32]. The cells in the control wells were cultured in the absence of exogenously added mycobacterial antigens. The final volume of the culture wells was adjusted to 200 µl. The plates were incubated at 37°C in a humidified atmosphere containing 5% CO₂ and 95% air. On day 6, culture supernatants (100 µl) were collected from each well and frozen at −70°C until used to determine cytokine concentrations. The remaining cultures were pulsed with 1 µci [³H]-thymidine (Amersham Life Science, Amersham, UK) and harvested (Skatron Instruments As, Oslo, Norway) according to standard procedures [28,29].

Interpretation of proliferation results

The radioactivity incorporated was obtained as counts per minute (cpm). The average cpm was calculated from triplicate cultures stimulated with each antigen or peptide pool, as well as from triplicate wells of negative control cultures lacking antigen. The cell proliferation results were presented as stimulation index (SI), which is defined as follows: Sl = cpm in antigen or peptide-stimulated cultures/cpm in cultures lacking antigen or peptide. A patient was considered responder to a given antigen if the PBMC yielded SI ≥ 3 [28]. Based on the median SI and percentages of responders in antigen-induced proliferation assays, the antigens were considered strong (median SI > 10 with > 70% responders), moderate (median SI > 3·0 but < 10 with 50–70% responders) or weak (median SI < 3·0 with < 50% responders) stimulators of PBMC.

Immunoassays for the quantification of cytokines

Supernatants (0·1 ml) from each well of 96-well plates were collected from the cultures of PBMC before [³H]-thymidine pulse. The supernatants were kept frozen at −20°C until assayed for cytokine concentration using FlowCytomix kits (Bender Medsystems Inc., Vienna, Austria), which allows for the simultaneous quantification of several analytes in one sample using a low sample volume (25 µl), as described previously [31]. The cytokines tested included Th1 (IFN-γ, IL-2, TNF-β) and Th2 (IL-4, IL-5, IL-10) cytokines. The test samples were analysed by flow cytometry using Coulter EPICS FC500 (Beckman Coulter Inc., Fullerton, CA, USA). For each analysis, up to 10 000 events were acquired. The mean concentration of each cytokine was expressed as pg/ml. By using the above kits, the minimum detectable concentrations of IL-2, IL-4, IL-5, IL-10, TNF-β and IFN-γ were 8·9 pg/ml, 6·4 pg/ml, 5·3 pg/ml, 6·9 pg/ml, 3·2 pg/ml and 7·0 pg/ml, respectively. In response to antigenic stimuli, experimental/control (E/C) values above 3 were considered a positive response [11]. E/C is the cytokine concentration in the supernatants of PBMC cultured in the presence of M. tuberculosis antigens/cytokine concentration in the supernatants of PBMC cultured in the absence of M. tuberculosis antigens. The antigen-induced positive responses were considered weak, moderate and strong when E/C-values were > 3–5, > 5–10 and > 10, respectively [31].

Statistical analysis

Statistical analyses were performed using SPSS 16·0 for Windows software package. The results of cytokine secretion were analysed statistically for significant differences between cultures of patients with TB and diabetes, patients with TB alone and healthy subjects using the non-parametric Mann–Whitney U-test. P-values of < 0·05 were considered significant.

Results

Antigen-induced proliferation of PBMC in response to complex mycobacterial antigens and RD peptides

All complex mycobacterial antigens were strong stimulators of PBMC proliferation, as indicated by high median SI in diabetic TB patients, non-diabetic TB patients and healthy subjects (median SI = 23–45, 13·5–76·5 and 15·6–29, respectively) and high percentage of responders (82–100%, 88–100% and 76–100%, respectively) (Table 3). In response to RD peptides, the proliferation of PBMC from diabetic TB and non-diabetic TB patients was strong in response to RD1_pool (median SI of 13 and 16·5, respectively, and 100% responders) whereas, in healthy subjects, RD1_pool induced weak responses (median SI of 4·5 and 57% responders) (Table 3). All other RD peptides (RD4_pool, RD6_pool and RD10_pool) induced weak proliferation responses in all of the tested groups (Table 3). The statistical analysis showed that none of the complex mycobacterial antigens induced proliferation responses that were significantly different in TB patients than in healthy subjects (P > 0·05), whereas RD1_pool induced significantly higher responses in both the diabetic and non-diabetic TB patients than in healthy subjects (P < 0·05). Furthermore, RD1-induced proliferation responses were significantly higher than the responses induced by RD4_pool, RD6_pool and RD10_pool in both the diabetic and non-diabetic TB patients (P < 0·05), but not the healthy subjects (P > 0·05) (Table 3).

Table 3.

Antigen-induced proliferation of peripheral blood mononuclear cells of diabetic tuberculosis (TB) patients, non-diabetic TB patients and healthy subjects in response to complex mycobacterial antigens and peptide pools of regions of difference (RD1), RD4, RD6 and RD10.

	Diabetic TB patients		Non-diabeticTB patients		Healthy subjects
Antigen	SI*	P/T†(%)	SI	P/T (%)	SI	P/T (%)
M. bovis BCG	45 (< 1–280)	10/11 (91)	32 (< 1–672)	15/17 (88)	23 (< 1–64)	17/20 (85)
M. tuberculosis	23 (< 1–346)	9/11 (82)	13·5 (1–464)	17/18 (94)	15·6 (4–91)	20/20 (100)
MT-CF	32 (8–281)	7/7 (100)	76·5 (10–505)	13/13 (100)	29 (3–328)	7/7 (100)
MT-CW	38 (3–342)	11/11 (100)	39 (4–3233)	18/18 (100)	20 (< 1–179)§	13/17 (76)
RD1pool	13 (6–78)	7/7 (100)	16·5 (7–250)	13/13 (100)	4·5 (< 1–29)‡,§	4/7 (57)
RD4pool	1·7 (1–8·5)	2/7 (28·5)	2·3 (1–8)	4/12 (33)	2 (< 1–27)	1/7 (14)
RD6pool	1 (< 1–7)	1/7 (14)	1·2 (< 1–8)	2/12 (17)	1 (< 1–2)	0/7 (0)
RD10pool	1 (< 1–3)	1/7 (14)	1 (< 1–5)	3/12 (25)	1 (< 1–7)	1/7 (14)

^*Stimulation index (SI) is presented as median (range) and is significant if ≥ 3 (shown in bold type).

^†P/T: positive responders per total tested (% = percentage of positive responders).

^‡P < 0·05 when diabetic TB compared with healthy subjects;

^§P < 0·05 when non-diabetic TB compared with healthy subjects. M. bovis, Mycobacterium bovis; BCG, bacille Calmette–Guérin; MT-CF, Mycobacterium tuberculosis culture filtrate; MT-CW, M. tuberculosis cell walls.

Cytokine secretion by PBMC in the absence of mycobacterial antigens

Only PBMC of diabetic TB patients demonstrated spontaneous secretion of detectable levels of all the tested Th1, except for IL-2 (median concentrations: IFN-γ = 48, TNF-β = 35 and IL-2 < 8·9 pg/ml), and Th2 (median concentrations: IL-4 = 82, IL-5 = 95 and IL-10 = 29 pg/ml) cytokines (data not shown). However, compared to the other two groups, PBMC of diabetic TB patients produced significantly higher concentrations of IL-4 alone (P < 0·05) (data not shown).

Th1/Th2 cytokine secretion by PBMC stimulated with complex mycobacterial antigens and RD peptides

PBMC from all groups of subjects showed strong (E/C > 10) secretion of Th1 cytokine IFN-γ in response to all the complex mycobacterial antigens with comparable concentrations in diabetic and non-diabetic TB patients (P > 0·05) (Table 4). In response to RD peptides, only RD1_pool induced strong IFN-γ secretion by all the tested groups but the diabetic TB patients secreted significantly more IFN-γ than non-diabetic TB patients and healthy subjects (P < 0·05) (Table 4). However, the peptide pools of the other RDs induced significantly higher IFN-γ by PBMC of diabetic TB patients than non-diabetic TB patients and healthy subjects (P < 0·05) (Table 4). Statistical analysis of the data showed that higher levels of IFN-γ were induced by the RD1_pool compared to the peptide pools of other RDs in both diabetic and non-diabetic TB patients (P < 0·05) (Table 4).

Table 4.

Interferon (IFN)-γ secretion by peripheral blood mononuclear cells (PBMC) of diabetic tuberculosis (TB) patients, non-diabetic TB patients and healthy subjects in response to complex mycobacterial antigens and peptide pools of regions of difference (RD1), RD4, RD6 and RD10.

	Concentration of IFN-γ (pg/ml) in PBMC culture supernatants of
	Diabetic TB patients		Non-diabetic TB patients		Healthy subjects
Antigen	Median (range)	E/C*	Median (range)	E/C	Median (range)	E/C
No antigen	48·5 (≤ 7–434)		≤ 7 (≤ 7–434)		12 (≤ 7–334)
M. bovis BCG	2248 (273–10408)	46	2157 (38–10507)	308	787 (≤ 7–18279)‡	65·5
M. tuberculosis	1997 (253–7439)	41	1479 (174–14188)	211	972 (29–10914)	81
MT-CF	2275 (1143–19322)	47	2692 (836–13421)	384	3990 (202–15945)	332
MT-CW	2474 (1672–6467)	51	2277 (343–14785)	325	1868 (170–9112)	156
RD1pool	1789 (251–11194)	37	405 (≤ 7–10507·3)†	58	164 (≤ 7–15870)‡	13·6
RD4pool	368 (23–790)	7·5	5·5 (≤ 7–562)†	0·7	12·6 (≤ 7–2270)‡	1
RD6pool	233 (64–1276)	4·7	7·6 (≤ 7–1254)†	1	61 (≤ 7–3699)‡	5
RD10pool	23 (≤ 7–386)	0·4	≤ 7 (≤ 7–160)†	1	14·5 (≤ 7–229)‡	1·1

^*E/C: cytokine concentration in antigen stimulated cultures/cytokine concentration in cultures lacking antigen, and is considered positive if ≥ 3 (shown in bold type).

^†P < 0·05 when diabetic TB compared with non-diabetic TB patients;

^‡P < 0·05 when diabetic TB compared with healthy subjects. BCG, bacille Calmette–Guérin; MT-CF, Mycobacterium tuberculosis culture filtrate; MT-CW, M. tuberculosis cell walls.

The results of quantification of other Th1 cytokines in the culture supernatants showed that none of the antigens induced secretion of IL-2 in any group of subjects tested (data not shown), whereas all the complex mycobacterial antigens induced the secretion of TNF-β (E/C ≥ 3) by PBMC of all the groups, except for whole cells of M. bovis BCG and M. tuberculosis in the diabetic TB patients (Table 5). The median concentrations of TNF-β in response to complex mycobacterial antigens in the tested groups of subjects were comparable (P > 0·05) (Table 5), except for M. bovis BCG, which induced significantly less TNF-β in PBMC cultures of healthy subjects compared to the diabetic patients (median concentration n = 21 pg/ml and 90 pg/ml, respectively). However, the non-diabetic TB patients and healthy subjects showed higher values of E/C (9·4–20 and 4·6–21·5, respectively) than the diabetic TB patients (E/C = 1·5–6·4). Among the RD peptides, only RD1_pool induced weak secretion of TNF-β by PBMC of diabetic and non-diabetic TB patients (E/C = 3·4, 4·4, respectively) but, as expected, none of the RD peptide pools induced positive responses in BCG-vaccinated healthy subjects (E/C < 3) (Table 5).

Table 5.

Tumour necrosis factor (TNF)-β secretion by peripheral blood mononuclear cells (PBMC) of diabetic patients, non-diabetic tuberculosis (TB) patients and healthy subjects in response to complex mycobacterial antigens and peptide pools of regions of difference (RD1), RD4, RD6 and RD10.

	Concentration of TNF-β (pg/ml) in PBMC culture supernatants of
	Diabetic TB patients		Non-diabetic TB patients		Healthy subjects
Antigen	Median (range)	E/C*	Median (range)	E/C	Median (range)	E/C
No antigen	35 (≤ 3·2–70 283)		≤ 3·2 (≤ 3·2–153)		≤ 3·2 (≤ 3·2–147)
M. bovis BCG	90 (≤ 3·2–138 445)	2·5	30 (≤ 3·2–221)	9·4	21 (≤ 3·2–450)	6·6
M. tuberculosis	54 (≤ 3·2–68 537)	1·5	33 (4–238)	10·4	15 (≤ 3·2–343)	4·6
MT-CF	225 (≤ 3·2–212 980)	6·4	64 (8–647)	20	69 (≤ 3·2–151)	21·5
MT-CW	140 (≤ 3·2–124 181)	4	59 (≤ 3·2–283)	18·4	31 (≤ 3·2–398)	9·7
RD1pool	120 (≤ 3·2–141 116)	3·4	14 (≤ 3·2–225)	4·4	≤ 3·2 (≤ 3·2–398)§	1
RD4pool	12 (≤ 3·2–47 277)	0·3	≤ 3·2 (≤ 3·2–38)	1	≤ 3·2 (≤ 3·2–41)‡	1
RD6pool	37 (≤ 3·2–55 723)	1	≤ 3·2 (≤ 3·2)†	1	≤ 3·2 (≤ 3·2–15)	1
RD10pool	≤ 3·2 (≤ 3·2–43 632)	0·1	≤ 3·2 (≤ 3·2–17)	1	≤ 3·2 (≤ 3·2–5)	1

^*E/C: cytokine concentration in antigen stimulated cultures/cytokine concentration in cultures lacking antigen, and is considered positive if ≥ 3 (shown in bold type).

^†P < 0·05 when diabetic TB compared with non-diabetic TB patients;

^‡P < 0·05 when diabetic TB compared with healthy subjects;

^§P < 0·05 when non-diabetic TB compared with healthy subjects. BCG, bacille Calmette–Guérin; MT-CF, Mycobacterium tuberculosis culture filtrate; MT-CW, M. tuberculosis cell walls.

With respect to Th2 cytokines, none of the complex mycobacterial antigens and RD peptide pools induced the secretion of IL-4 and IL-5 by PBMC of all tested groups (E/C < 3, data not shown), whereas IL-10 was secreted in response to all complex mycobacterial antigens in the three groups of subjects, except MT-CF which did not induce IL-10 secretion in healthy subjects (E/C = 1·4) (Table 6). Statistical analysis of the data showed that in response to M. bovis BCG, PBMC of diabetic TB patients secreted significantly higher concentrations of IL-10 than the non-diabetic TB patients and healthy subjects (P < 0·05) (Table 6). In addition, the diabetic TB patients secreted significantly higher concentrations of IL-10 than the healthy subjects in response to whole cells of M. tuberculosis and MT-CW (P < 0·05) (Table 5). In response to RD peptide pools, RD1_pool, RD4_pool and RD10_pool induced secretion of IL-10 by PBMC of diabetic patients only (E/C = 7, 4, 5·6, respectively), and the levels of IL-10 were significantly higher in all RD peptide pools in diabetic TB patients than in non-diabetic TB patients and healthy subjects (P < 0·05) (Table 6).

Table 6.

Interleukin (IL)-10 secretion by peripheral blood mononuclear cells (PBMC) of diabetic patients, non-diabetic tuberculosis (TB) patients and healthy subjects in response to complex mycobacterial antigens and peptide pools of regions of difference (RD1), RD4, RD6 and RD10.

	Concentration of IL-10 (pg/ml) in PBMC culture supernatants of
	Diabetic TB patients		Non-diabetic TB patients		Healthy subjects
Antigen	Median (range)	E/C*	Median (range)	E/C	Median (range)	E/C
No antigen	29 (≤ 6·9–837)		≤ 6·9 (≤ 6·9–666)		≤ 6·9 (≤ 6·9–297)
M. bovis BCG	601 (162–1740)	20·4	169 (≤ 6·9–2427)†	24·4	38 (≤ 6·9–1109)‡	5·5
M. tuberculosis	974 (250–6536)	33	221 (≤ 6·9–4934)	32	330 (≤ 6·9–2186)‡	47·8
MT-CF	157 (≤ 6·9–1063)	5·3	81 (≤ 6·9–430)	12	10 (≤ 6·9–273)	1·4
MT-CW	747 (54–10 949)	25·3	104 (≤ 6·9–3390)	15	25 (≤ 6·9–2389)‡	3·6
RD1pool	207 (13–519)	7	≤ 6·9 (≤ 6·9–329)†	1	≤ 6·9 (≤ 6·9–198)‡	1
RD4pool	120 (14–352)	4	≤ 6·9 (≤ 6·9–20)†	1	≤ 6·9 (≤ 6·9–166)‡	1
RD6pool	67 (≤ 6·9–1222)	2·2	≤ 6·9 (≤ 6·9–21)†	1	≤ 6·9 (≤ 6·9–62)‡	1
RD10pool	167 (≤ 6·9–561)	5·6	≤ 6·9 (≤ 6·9–14)†	1	≤ 6·9 (≤ 6·9–116)‡	1

^*E/C: cytokine concentration in antigen stimulated cultures/cytokine concentration in cultures lacking antigen, and is considered positive if ≥ 3 (shown in bold type).

^†P < 0·05 when diabetic TB compared with non-diabetic TB patients;

^‡P < 0·05 when diabetic TB compared with healthy subjects. BCG, bacille Calmette–Guérin; MT-CF, Mycobacterium tuberculosis culture filtrate; MT-CW, M. tuberculosis cell walls.

Th1 : Th2 cytokine ratios

To determine Th1 versus Th2 cytokine bias in response to various mycobacterial antigens in the tested groups of subjects, Th1 : Th2 cytokine ratios were calculated and compared between groups (Table 7). The results showed that IFN-γ : IL-4 ratios to all complex mycobacterial antigens and RD1_pool were lower in diabetic TB patients compared to non-diabetic TB patient and healthy subjects (Table 7). A similar observation was made regarding IFN-γ : IL-5 and IFN-γ : IL-10 ratios (Table 7). Furthermore, and in general, ratios of TNF-β : IL-4, TNF-β : IL-5 and TNF-β : IL-10 to various antigens were also lower in the diabetic TB patients compared to the other two groups (Table 7). A similar analysis for RD4_pool, RD6_pool and RD10_pool was not possible because they were weak stimulators of Th1 and Th2 cytokines, as reported above. The overall analysis showed that Th1 : Th2 cytokine ratios were consistently lower in diabetic TB patients than non-diabetic TB patients and healthy subjects (Table 7), thus suggesting a Th2 cytokine bias in diabetic TB patients.

Table 7.

T helper type 1 (Th1) : Th2 type cytokine ratios in culture supernatants of peripheral blood mononuclear cells (PBMC) obtained from diabetic and non-diabetic tuberculosis (TB) patients and Mycobacterium bovis bacille Calmette–Guérin (BCG)-vaccinated healthy subjects in the presence of complex mycobacterial antigens and region of difference (RD)1_pool.

	Cytokine ratios in culture supernatants of
Antigen Th1 : Th2 cytokine	Diabetic TB patients	Non-diabetic TB patients	Healthy subjects
M. bovis BCG
IFN-γ : IL-4	31	337	123
IFN-γ : IL-5	16·7	240	148
IFN-γ : IL-10	3·7	13	21
TNF-β : IL-4	1·3	5	3·3
TNF-β : IL-5	0·7	3·3	4
TNF-β : IL-10	0·2	0·2	0·6
M. tuberculosis
IFN-γ : IL-4	27	231	152
IFN-γ : IL-5	10·5	134	183†
IFN-γ : IL-10	2	6·6	3
TNF-β : IL-4	0·7	5	2·3
TNF-β : IL-5	0·3	3	2·8
TNF-β : IL-10	0·1	0·2	0·1
MT-CF
IFN-γ : IL-4	103	420*	623†
IFN-γ : IL-5	30	448*	752†
IFN-γ : IL-10	14·5	33	399†
TNF-β : IL-4	10	10	11
TNF-β : IL-5	3	11	13
TNF-β : IL-10	1·4	0·8	7
MT-CW
IFN-γ : IL-4	28	355	291
IFN-γ : IL-5	12	189*	352†
IFN-γ : IL-10	3·3	22	74
TNF-β : IL-4	1·6	9	5
TNF-β : IL-5	0·7	5	6
TNF-β : IL-10	0·2	0·6	1·2
RD1pool
IFN-γ : IL-4	17	63	25·6
IFN-γ : IL-5	15·3	58	31
IFN-γ : IL-10	8·6	58*	23‡
TNF-β : IL-4	1·1	2·2	0·5
TNF-β : IL-5	1	2	0·6
TNF-β : IL-10	0·6	2	0·5

^*P < 0·05 when diabetic TB compared with non-diabetic TB patients;

^†P < 0·05 when diabetic TB compared with healthy subjects.

^‡P < 0·05 when non-diabetic TB compared with healthy subjects. IFN, interferon; IL, interleukin; TNF, tumour necrosis factor; MT-CF, Mycobacterium tuberculosis culture filtrate; MT-CW, M. tuberculosis cell walls.

Discussion

To understand the immunological mechanisms involved in increased pathogenesis of TB in diabetic patients, we have analysed CMI responses using PBMC from pulmonary TB patients with and without type II diabetes and M. bovis BCG-vaccinated healthy subjects. The assays used have been shown previously to correlate with protective immunity (antigen-induced cell proliferation and secretion of Th1 cytokines IL-2, IFN-γ and TNF-β) and susceptibility/pathogenesis in TB (secretion of Th2 cytokines IL-4, IL-5 and IL-10) [11–14]. A number of complex mycobacterial antigen preparations and peptides corresponding to M. tuberculosis-specific regions RD1, RD4, RD6 and RD10 were used, because they may induce immune responses that vary in type and magnitude based on antigenic differences [33]. To study antigen-induced cellular responses the cultures were harvested on day 6, because previous studies have shown that this time period is optimal for mycobacterial antigen-induced cellular proliferation and cytokine responses, e.g. IFN-γ, of human PBMC [31,34].

The results of antigen-induced cell proliferation studies showed that all the complex mycobacterial antigen preparations were strong stimulators of PBMC of diabetic and non-diabetic TB patients and M. bovis BCG-vaccinated healthy subjects (Table 3). These findings are in agreement with our recent study using various preparations of complex antigens of M. tuberculosis and defined secreted antigens of cross-reactive nature, i.e. MPT59 and MPT64 [33]. However, among RD peptides, only RD1_pool induced strong responses in diabetic and non-diabetic TB patients and moderate responses in M. bovis BCG-vaccinated healthy subjects (Table 3). These results suggest that diabetic and non-diabetic TB patients respond in a similar way to both cross-reactive and specific antigens of M. tuberculosis, and only RD1 contains immunodominant antigens of M. tuberculosis. By using one of the antigens constituting the RD1 peptide pool, i.e. early secreted antigenic target 6 (ESAT6), we have shown previously that ESAT6 induces equally strong responses in diabetic and non-diabetic TB patients but weak responses in M. bovis BCG-vaccinated healthy subjects [33], due most probably to latent TB [11,28,31,36,37]. The moderate responses observed to RD1_pool in this study in M. bovis BCG-vaccinated healthy subjects could also have been due to the presence of Pro-Pro-Glu 68 (PPE68), because the sequence of the immunodominant peptide of PPE68 is also present in PPE-family proteins of M. bovis BCG [32,35]. To study the profile of Th1 and Th2 cytokines, we first determined their spontaneous ex vivo secretion by PBMC, which implies prior in vivo stimulation, and thus represents cell-specific activity in the context of current infection [38,39]. The results showed that PBMC of only diabetic TB patients spontaneously secreted detectable concentrations of Th1 (IFN-γ and TNF-β) and Th2 cytokines (IL-4, IL-5 and IL-10), but the concentration of IL-4 alone was significantly higher in diabetic TB patients compared to non-diabetic TB patients and healthy subjects. The elevated IL-4 secretion may contribute to increased pathogenesis in diabetic TB patients because IL-4 impairs anti-microbial activity of infected cells, increases availability of iron to intracellular M. tuberculosis and drives pulmonary fibrosis, etc. [22].

The present work also included studies to assess mycobacterial antigen-induced secretion of Th1 and Th2 cytokines by PBMC of various groups of subjects, which may be useful as early predictors of disease activity and contribute to the understanding of the disease pathogenesis [11–14]. The results showed that among the Th1 cytokines studied, large quantities of IFN-γ were secreted by PBMC of all groups tested in response to complex mycobacterial antigens, without statistical differences between the groups, except for significantly higher levels of IFN-γ produced by diabetic patients, compared to healthy subjects, in response to M. bovis BCG (Table 4). In response to RD peptides, i.e. RD1_pool, RD4_pool and RD6_pool, significantly higher concentrations of IFN-γ were secreted by PBMC of diabetic TB patients compared to non-diabetic TB patients and M. bovis BCG-vaccinated healthy subjects (P < 0·05). However, strong responses (E/C ≥ 10) were observed with RD1 only.

TNF-β was secreted by all groups in response to the complex mycobacterial antigens and/or RD1_pool (Table 5). However, in spite of positive antigen-induced proliferation responses to complex mycobacterial antigens and RD1_pool, IL-2 was not detected in any of the antigen-stimulated cultures (data not shown). This could have been due to utilization of IL-2 by T cells proliferating in response to antigenic stimuli [40]. Thus, the overall results suggest that the increased pathogenesis observed in diabetic patients may not be due to decreased mycobacterial antigen-induced secretion of Th1 cytokines in these patients.

Among the Th2 cytokines, mycobacterial antigen-induced secretions of IL-4 and IL-5 were not observed in any group of subjects included in this study (data not shown), thus supporting previous observations reported for non-diabetic TB patients and healthy subjects [11,31,41]. However, in response to the complex mycobacterial antigens and/or RD peptide pools, significantly higher concentrations of IL-10 were secreted by PBMC of diabetic TB patients than non-diabetic TB patients and healthy subjects (Table 6). Despite the abundant production of Th1 cytokines, secretion of IL-10 interferes with their protective function [42], and helps mycobacteria to survive intracellularly [43]. Therefore, our results suggest that, in spite of large quantities of IFN-γ secretion in response to mycobacterial antigens, the elevated secretion of IL-10 may contribute to increased pathogenesis in diabetic TB patients.

Because Th1 and Th2 cytokines regulate the secretion of each other, their net biological effect can be evaluated more clearly by comparing the cytokine ratios [44–46]. A higher Th1/Th2 ratio indicates a protective type 1 cytokine bias, while a lower ratio suggests a bias towards pathological type 2 cytokines in TB [44–50]. In this study, we found that the overall Th1/Th2 cytokine ratios (IFN-γ : IL-4, IFN-γ : IL-5 and IFN-γ : IL-10; and TNF-β : IL-4, TNF-β : IL-5 and TNF-β : IL-10) were lower in the diabetic TB patients than non-diabetic TB patients and healthy subjects (Table 7). The lower Th1 : Th2 ratios may also contribute to hampered resistance to M. tuberculosis infection under diabetic conditions. This study is restricted to selected Th1 and Th2 cytokines; however, it should be stressed that the failure to eradicate M. tuberculosis probably involves sophisticated immune evasion strategies of the pathogen as well as the immune regulatory mechanisms of the host [51]. In the context of regulatory mechanisms, several subsets of regulatory T cells (T_reg) have been identified such as IL-10 (Tr1) or TGF-β-secreting T_reg (Th3) and CD4⁺CD25⁺ forkhead box P3 (FoxP3)⁺ T_reg cells [22,51,52]. It has been shown that CD4⁺CD25⁺FoxP3⁺ T_reg cells capable of suppressing Th1 responses are present in high frequency in TB patients [52]. Furthermore, previous studies have shown a correlation of IL-4 antagonist IL-4δ2 with protective immunity in TB by demonstrating high levels of IL-4δ2 in healthy infected subjects compared to TB patients and increase in the levels of IL-4δ2 after chemotherapy of TB patients [53–55]. Therefore, in future, studies may be carried out to determine the role of T_reg cells and IL-4δ2 in understanding the immunological mechanisms responsible for the increased pathogenesis of TB in diabetic patients.

Disclosure

None of the authors has any conflict of interest to declare.