Adipokines and the risk of active TB: a nested case-control study

SUMMARY

BACKGROUND:

Adipokines are emerging mediators of immune response, and may affect susceptibility to active TB.

OBJECTIVE:

To examine the associations between adipokines and the risk of active TB.

METHODS:

In a case-control study nested within a prospective cohort of middle-aged and older adults in Singapore, 280 incident active TB cases who donated blood for research before diagnosis were matched with 280 controls. Serum levels of adiponectin, resistin, leptin and ghrelin were measured. Multivariable logistic regression models were used to compute the adjusted odds ratios (ORs) and 95% confidence intervals (CIs) for the associations between adipokines and the risk of active TB.

RESULTS:

Higher levels of leptin and resistin were associated with reduced risk of TB in a dose-dependent manner. Compared to those in the lowest quartile of leptin levels, those in the highest quartile had an OR of 0.46 (95%CI 0.26–0.82; P for trend = 0.009). Similarly, compared to those in the lowest quartile of resistin levels, those in the highest quartile had an OR of 0.46 (95%CI 0.24–0.90; P for trend = 0.03). Adiponectin and ghrelin levels were not associated with TB risk.

CONCLUSION:

Increased serum levels of leptin and resistin may be associated with reduced susceptibility to active TB infection.

TB is the leading cause of global mortality from an infectious disease. Approximately 1.7 billion people are infected with Mycobacterium tuberculosis worldwide and are at risk of developing active disease. The WHO estimates that around 10 million people fall ill with TB each year.¹

The immune mechanisms that control progression from infection to active disease are still unclear. Although the adipose tissue is traditionally believed to be inert and used only for energy storage, increasing evidence suggests that it plays a role in the regulation of immune and inflammatory mediators. Adipokines such as adiponectin, resistin and leptin are among these mediators.^2–7 The role of adipokines in metabolic health has been extensively studied,⁸ but emerging evidence also indicates they modulate immune responses to infections.^9,10 Ghrelin, the counterpart of leptin in regulating appetite and energy expenditure, also possesses anti-inflammatory properties that could potentially counteract the effects of leptin.^11,12

Several retrospective case-control studies have examined the associations of ghrelin and adipokines with risk of active TB. However, findings on the association between ghrelin and TB risk have been inconsistent.^13–15 Although some studies found higher plasma adiponectin levels in TB patients than controls, these differences were not statistically significant.^16,17 Resistin levels were significantly elevated in TB patients compared to controls, especially in those with severe disease.^13,18 On the other hand, TB patients often had lower leptin levels than those without active TB.^14,16,17

Although these case-control studies suggest a link between these biomarkers and active TB, reverse causality could not be ruled out, as biomarker levels may already have been altered by disease progression and/or treatment since these were measured after clinical diagnosis of disease. Furthermore, TB patients often suffer from wasting, which can in turn induce changes in the levels of these biomarkers.^19,20

A case-control study nested in a prospective cohort would overcome such shortcomings. Using resources of the Singapore Chinese Health Study, we aimed to examine the prospective association between adiponectin, resistin, leptin and ghrelin, and the subsequent development of active TB using a nested case-control study design. As serum samples for biomarkers measurement were collected from participants several years prior to their clinical diagnosis of TB, this study could correctly establish the temporal relationship between these adipokines and active TB disease. Participants of this cohort went through periods when TB was highly prevalent in the country several decades ago. Hence, those who had acquired latent TB infection in those early years would be at risk of disease reactivation in advanced age.²¹

METHODS

Study population

The Singapore Chinese Health Study is a prospective population-based cohort comprising 63 257 Chinese adults aged 45–74 years from 1993 to 1998.²² All participants were citizens or permanent residents of Singapore. Study participants were restricted to the two major dialect groups of Chinese in Singapore—Hokkiens who originated from Fujian Province, and Cantonese who came from Guangdong Province in Southern China. An in-person interview was conducted using a structured questionnaire during recruitment. Information collected included demographics, height and weight, diet, smoking and alcohol drinking habits. Self-reported history of physician-diagnosed medical conditions, such as diabetes (DM) and age at diagnosis were also collected.

During the first follow-up phase (1999–2004), 52 322 participants were successfully re-contacted via telephone interviews to update information on selected lifestyle habits (e.g., smoking and alcohol drinking) and medical history, including DM. They were also invited to donate biospecimens, and 28 330 consenting participants donated blood samples. Cases and controls for this study were identified from this subpopulation who donated blood samples.

The study was approved by the Institutional Review Board at the National University of Singapore, Singapore. All enrolled participants gave informed consent for interviews and blood sampling.

Cases and controls selection

History of active TB among cohort participants was obtained using record linkage with the Singapore National TB Notification Registry up to 31 December 2014. In Singapore, it is mandatory by law to notify all suspected and confirmed TB cases to the registry within 72 hours of starting TB treatment and/or confirmation based on culture-positive results. Diagnosis of TB cases in Singapore are confirmed using chest radiograph, followed by sputum smear and culture testing. All data on culture-positive TB patients are comprehensively captured in this registry via electronic linkage with the two mycobacterial laboratories in Singapore. Approximately 88% of all the incident active TB cases in this cohort were pulmonary TB cases.²³

A total of 1249 individuals were found to have developed active TB after recruitment into the study. Of these, 280 donated blood for research before their TB diagnosis and were thus selected as study cases. Compared to the remaining 969 who had active TB but who either did not give blood for research (n = 830) or who gave blood for research after TB diagnosis (n = 139), the cases included in this study were significantly younger (64.2 years vs. 67.4 years), heavier (body mass index [BMI] 22.3 kg/m² vs. 21.7 kg/m²) and had a higher level of education (30.4% vs. 19.5% with secondary school or higher education). In all other respects, TB cases in this study and those who were not included were not significantly different by sex, smoking status, history of DM or frequency of alcohol consumption.

Each TB case was matched to a control with available samples and who remained free of TB until 31 December 2014. The matching criteria included age at blood collection (± 2 years), sex, dialect group (Hokkien/Cantonese) and date of blood sample collection (± 6 months).

Laboratory procedures

Leptin and resistin levels in serum samples were measured using Luminex^® xMAP^® technology (ProcartaPlex^™ Multiplex Panel, Thermo Fisher Scientific, Waltham, MA, USA). Serum samples were incubated overnight with fluorescent-coded magnetic beads, each conjugated to capture antibodies against the analyte of interest. The plates were washed, and biotinylated detection antibodies were incubated with the complex for 1 h. Streptavidin-phycoerythrin was then added and incubated for 30 min. The plates were washed again and resuspended in sheath fluid for subsequent measurement using Luminex^™ FLEX-MAP 3D^®. Data were acquired using xPONENT^® 4.0 (Luminex) software and analysed using Bio-Plex Manager^™ Software 6.1.1 (Bio-Rad, Hercules, CA, USA). Concentrations of samples were calculated using a 4PL or 5PL algorithm based on the median fluorescence intensity (MFI) generated by each standard.

Adiponectin (Abcam, Cambridge, UK) and ghrelin (eBioscience, Thermo Fisher Scientific, Waltham, MA, USA) were measured using enzyme-linked immunosorbent assay. Serum samples and detection antibodies were incubated in target-specific, pre-coated 96-well plates. After incubation, wells were washed to remove unbound material. Finally, a chromogen substrate was added and catalysed for streptavidin-peroxidase enzymatic reaction. The reaction was terminated and absorbance measured immediately on the EnVision^® 2104 multimode microplate reader (Perkin Elmer, Waltham, MA, USA). The detection range was 4.13–16 900 pg/ml for resistin, 14.65–60 000 pg/ml for leptin, 16–1000 pg/ml for ghrelin and 0.781–50 ng/ml for adiponectin.

Statistical analysis

As biomarker levels were significantly different between men and women in our study, participants were divided into sex-specific categories of quartiles based on the distribution of individual biomarkers separately for men and women among the controls. Conditional logistic regression models were used to estimate adjusted odds ratios (ORs) with 95% confidence intervals (CIs) of active TB associated with biomarker levels. Any case-control set with undetectable measurement of a specific biomarker in either the case or the control, likely due to problematic laboratory measurement, was excluded from matched analysis. The multivariable model included age at blood collection, level of education, smoking status, alcohol intake and BMI and history of DM. All statistical analysis was conducted using SAS v9.4 (SAS Institute, Cary, NC, USA). All reported P values were two-sided; P < 0.05 was considered statistically significant.

RESULTS

Among the 280 cases of active TB, the mean duration from blood collection to TB diagnosis was 6.5 years (standard deviation [SD] 3.6). The mean age of all participants was 64.1 years (SD 7.3) at the time of blood sample collection; 76.1% of TB cases were male. As expected, TB cases had significantly lower BMI and higher prevalence of smoking (Table 1). The correlation among the four biomarkers measured was weak; Spearman correlation coefficient ranged from – 0.11 to 0.19 for pairwise correlations among these biomarkers.

Table 1.

Characteristics of active TB cases and controls, Singapore Chinese Health Study, Singapore

Characteristics	TB cases* (n = 280) n (%)	Controls* (n = 280) n (%)	P value†
Age at blood collection, years, mean ± SD	64.2 ± 7.3	64.1 ± 7.3	0.85
Men	213 (76.1)	213 (76.1)	—
Dialect group			—
Cantonese	138 (49.3)	138 (49.3)
Hokkien	142 (50.7)	142 (50.7)
Body mass index, kg/m2, mean ± SD	22.3 ± 3.6	23.1 ± 3.1	0.006
Level of education			0.14
No formal education	48 (17.1)	40 (14.3)
Primary school	147 (52.5)	133 (47.5)
Secondary school or above	85 (30.4)	107 (38.2)
History of diabetes	58 (20.7)	49 (17.5)	0.33
Smoking status			<0.001
Never smokers	106 (37.9)	147 (52.5)
Former smokers	64 (22.9)	67 (23.9)
Current smokers	110 (39.3)	66 (23.6)
Alcohol intake			0.25
Non-drinkers	229 (81.8)	223 (79.6)
Monthly or weekly	39 (13.9)	50 (17.9)
Daily	12 (4.3)	7 (2.5)

^*TB cases and controls were matched according to age at blood collection (± 2 years), sex, dialect and date of blood collection (± 6 months).

^†Based on χ² test for categorical variables and Student’s t-test for continuous variables.

SD = standard deviation.

The associations between biomarkers and risk of active TB are given in Table 2. In the multivariable model, the highest quartile of leptin levels was associated with an OR of 0.46 (95%CI 0.26–0.82; P_trend = 0.009), while the highest quartile of resistin levels was associated with an OR of 0.46 (95%CI 0.24–0.90; P_trend = 0.03) when compared to the lowest quartile of the biomarker, respectively. Conversely, adiponectin and ghrelin levels were not significantly associated with risk of active TB.

Table 2.

Risk of active TB according to quartiles of biomarkers, Singapore Chinese Health Study, Singapore^*

Biomarker	Median level	Cases/controls	OR (95%CI)
Leptin, pg/ml
Quartile 1	926	114/64	1.00
Quartile 2	2435	53/65	0.53 (0.31–0.91)
Quartile 3	3651	40/66	0.36 (0.20–0.64)
Quartile 4	6603	54/66	0.46 (0.26–0.82)
Ptrend			0.009
Resistin, pg/ml
Quartile 1	2848	77/55	1.00
Quartile 2	4222	57/55	0.63 (0.34–1.15)
Quartile 3	6057	49/60	0.44 (0.23–0.85)
Quartile 4	9696	48/61	0.46 (0.24–0.90)
Ptrend			0.03
Ghrelin, ng/ml
Quartile 1	1.83	67/69	1.00
Quartile 2	2.97	60/70	0.83 (0.49–1.41)
Quartile 3	4.05	76/69	1.13 (0.67–1.88)
Quartile 4	4.80	75/70	0.90 (0.53–1.53)
Ptrend			0.95
Adiponectin, ng/ml
Quartile 1	764	55/59	1.00
Quartile 2	1325	61/64	1.05 (0.58–1.92)
Quartile 3	1862	55/63	0.96 (0.48–1.90)
Quartile 4	3078	82/67	1.43 (0.72–2.83)
Ptrend			0.20

^*Model was adjusted for age at blood collection (years) and level of education (no formal, primary school, secondary and above), smoking status (never, former, current), alcohol intake (none, monthly/weekly, daily), body mass index (kg/m²) and diabetes

OR = odds ratio; CI = confidence interval.

Sensitivity analysis was performed by examining the relation between biomarkers and TB risk separately for cases with longer follow-up (≥5 years) and cases with a shorter follow-up (<5 years) after blood collection; results of the two groups were similar (data not shown).

DISCUSSION

In this case-control study nested within a prospective population-based cohort, increased serum leptin and resistin levels were associated with lower risk of active TB. Our study provided novel evidence that those with higher serum leptin and resistin levels subsequently had lower susceptibility to developing active disease.

First discovered as a hormone that regulates energy homeostasis,²⁴ leptin has been shown to modulate immune responses by stimulating phagocytosis and production of pro-inflammatory cytokines interleukin (IL) 6 and tumour necrosis factor alpha (TNF-α).²⁵ In experimental research, mycobacterial infected leptin-deficient ob/ob mice had higher bacterial loads in the lungs and lower protective cytokine production than wild-type mice. The reduced cytokine response was restored with leptin replacement, demonstrating the role of leptin in immune response to pulmonary TB.²⁶ A meta-analysis of 12 case-control studies showed higher serum leptin levels in healthy controls than in TB patients, suggesting decreased leptin to be associated with active TB.²⁷ Our study provided novel evidence that those with lower serum leptin levels years before diagnosis of active TB subsequently had increased risk of developing active disease.

Ghrelin, which has opposing effects from leptin in energy regulation and inflammatory responses,^11,28 did not modulate TB risk in our study. There are inconsistent findings on the link between ghrelin levels and active TB,^13–15 and a case-control study that also examined the nutritional status of the TB patients found ghrelin levels to be significantly lower among malnourished than well-nourished TB patients, while there was no significant difference in ghrelin levels between TB patients and healthy controls.²⁹ These findings collectively suggest that while leptin and ghrelin are counterparts in many of their physiological functions, only leptin may have immunomodulatory effects on TB disease.

Human resistin is induced in response to various inflammatory stimuli such as lipopolysaccharide (LPS), TNF-α or IL-6, and resistin itself induces proinflammatory cytokines, suggesting a role for resistin in inflammation in humans.^30–32 Serum resistin was elevated in patients with severe TB or DM compared with mild TB cases and healthy controls.¹⁸ However, unlike our prospective study, these retrospective studies were unable to determine whether the increased resistin level in TB patients was a cause or sequel of active disease.

The strengths of our study include the prospective design, which allowed us to examine biomarker levels prior to active TB. Recall bias in the collection of other exposure data was also minimised. As TB cases in our cohort were identified via linkage with the nationwide registry, notification of which is mandatory, our study has highly valid definition of cases and controls. Our analysis was also able to adjust for many well-established TB risk factors. A limitation of the study is that history of TB exposure was not captured, and cases and controls were assumed to have equal risk of acquiring TB infection, whether before or after blood sampling. Whether the effect of adipokines was at the level of acquisition of TB infection or progression to active disease, or both, could not be ascertained. We did not collect data on HIV status, but with a very low prevalence (117.8 cases per million population) of this condition in Singapore,³³ any confounding effect from not adjusting for HIV status in this cohort is likely to have negligible impact on the associations observed. Finally, residual confounding may be present due to unknown or unmeasured factors.

In conclusion, we found increased serum leptin and resistin to be inversely associated with TB risk. Our study provides novel evidence on the prospective association between adipokines and TB risk, and strengthens evidence on the potential role of adipokines in TB reactivation. More studies are necessary to validate these findings and elucidate the underlying mechanisms.