Polymorphisms of the BCL2 gene associated with susceptibility to tuberculosis

ABSTRACT

Although tuberculosis (TB) is a serious public health concern, we still don’t understand why only 10% of people infected will develop the disease. Apoptosis plays a role in the interaction of Mycobacterium tuberculosis (Mtb) with the human host and it may be modified by subtle alterations in the B-cell lymphoma 2 (BCL2) gene, an anti-apoptotic regulatory element. Therefore, we investigated whether there is an association between BCL2 polymorphisms and susceptibility to TB by analyzing 130 TB cases, 108 subjects with latent TB infection (LTBI), and 163 healthy controls (HC). Logistic regression was used to calculate odds ratios (ORs) and 95% confidential intervals (95% CIs) for possible associations between single nucleotide polymorphisms (SNPs) in BCL2 and the risk of tuberculosis. We found that the G allele of rs80030866 (OR=0.62, 95%CI:0.42-0.91, P=0.015), and also the G allele of rs9955190 (OR=0.58, 95%CI:0.38-0.88, P=0.011) were less frequent in the TB group compared with the LTBI group. In addition, individuals with rs2551402 CC genotype were more likely to have LTBI than those with AA genotype (OR=2.166, 95%CI:1.046-4.484, P=0.037). Our study suggests that BCL2 gene polymorphisms may be correlated with susceptibility to both TB and LTBI.

Tuberculosis; Susceptibility; Latent tuberculosis infection; BCL2; Polymorphism

INTRODUCTION

Tuberculosis (TB) is a transmittable infectious disease caused by Mycobacterium tuberculosis (Mtb) that primarily targets the lungs. Globally, an estimate of 9.9 million people have TB and 1.3 million TB deaths among HIV-negative people were reported in 2020¹, confirming that it is still a leading public health problem. As a consequence of disruptions in TB control caused by the COVID-19 pandemic, deaths due to TB in high-burden settings may increase by up to 20% over the next 5 years².

Individuals who are in close contact with active TB cases have a higher risk of developing latent TB infection (LTBI). People with LTBI show a persistent immune response to Mtb antigens, but have no evidence to clinically manifest TB disease³. However, in a small percentage of individuals with LTBI, the immune system will fail to control the infection at some point in their lives and they will develop active TB, so they represent a potential reservoir for new TB cases⁴. It is estimated that about one-fourth of the world’s population is infected with Mtb⁵ but only 5%-10% of infected individuals will develop TB disease⁶. The immunological mechanisms that either restrict the infection or allow it to progress to active disease remain poorly understood, but host genetic factors have been suspected to play an important role⁷.

Studies have linked TB disease susceptibility to polymorphisms in several candidate target genes⁸, but many of these associations have been difficult to ascertain. Genome-wide association (GWAS) studies have also been used to find loci associated with susceptibility to TB, but surprisingly, GWAS did not find any association with the loci identified in the candidate gene studies⁹. Therefore, the specific genetic elements that influence the risk of TB infection and disease are still largely unknown. Nevertheless, if TB risk-associated genes could be identified, they would help to clarify the pathogenesis of this disease and enable personalized treatment based upon an individual’s risk of infection and progression to disease.

Previous studies have shown that apoptosis of macrophages¹⁰ and T cells¹¹ plays a vital role in host defense against Mtb and TB pathogenesis. The B-cell lymphoma 2 (BCL2) gene functions as an antiapoptotic regulatory element and its down-regulation correlate with an increased risk of TB disease progression¹². Therefore, we decided to investigate the possible role of BCL2 in TB genetic susceptibility through a case-control study of the association of BCL2 gene polymorphisms with TB/LTBI risk in the Chinese Han population.

MATERIALS AND METHODS

Study participants

The study enrolled 130 pulmonary TB patients and 279 close contacts of individuals with sputum smear or culture-positive TB. Eight out of 279 contacts were ruled out in the quality control of genotyping, leaving 271 cases for the final analysis. All participants were recruited from the Chinese Han population visiting the Shenzhen Nanshan Center for Chronic Disease Control (22° N 113° E) from May 2017 to July 2018 and all were vaccinated with Bacillus Calmette-Guerin (BCG) in their infancy. TB patients were newly diagnosed by clinical specialists according to microbiological criteria (positive sputum smear or cultures), clinical manifestations, and radiology (chest X-rays or computed tomography scans). To detect LTBI, both interferon-gamma release assays (IGRA) and the Mantoux tuberculin skin test (TST) are recommended by WHO¹³,¹⁴. In our study, we used IGRA to separate the contacts into those with LTBI and the uninfected healthy controls (HC). LTBI were IGRA positive, while HC had negative IGRA tests. LTBI and HC individuals had no TB-related symptoms and negative sputum for acid-fast bacilli by microscopy. Participants with cancer, concomitant chronic obstructive pulmonary disease, HIV infection, hepatitis B virus (HBV) infection, HCV infection, or immune-mediated disorders were excluded.

Extraction of genomic DNA

Blood samples were collected from all participants in EDTA tubes, followed by genomic DNA extraction using QIAamp^® DNA blood mini kit (QIAGEN Inc., Hilden, Germany), by the manufacturer’s instructions. The extracted DNA was diluted to a working concentration of 50 ng/μL for further genotyping.

SNP selection and genotyping

The SNPs in the BCL2 gene were selected from the 1000 Genomes Project based on tag SNPs with a minor allele frequency ≥ 0.05 and r² > 0.8, using Haploview software (version 4.2, Broad Institute of MIT and Harvard, Cambridge, MA, USA). These tag SNPs were annotated for their location and predictive functions¹⁵. Study subjects were genotyped using the Illumina HumanOmniZhongHua-8 BeadChip. The analysis was repeated on 5% of the samples to evaluate the reproducibility and quality of genotyping. To exclude possible genotyping errors, we generated genotype cluster plots from the selected SNPs and visually inspected them using GenomeStudio software (version 2.0, Illumina, San Diego, CA, USA).

Ethics approval

Signed informed consent was obtained from all participants and ethics approval was granted by the Ethics Review Committee of the Shenzhen Nanshan Center for Chronic Disease Control (ll20170018).

Statistical analyses

T-test and χ² test were used to compare age and gender distributions between the patients and controls. Deviation from Hardy–Weinberg equilibrium (HWE) was assessed using the SNPassoc package of R (version 4.0.3, R Project for Statistical Computing, Vienna, Austria). The differences in allele frequencies between the three groups of subjects were performed with the χ² test. Logistic regression analysis, under genotypes and genetic models (dominant, recessive, and additive model)¹⁶, adjusting for age and gender, was used to assess possible associations between BCL2 SNPs and TB susceptibility, and to calculate 95% confidence intervals (95% CIs), odds ratios (ORs), and P values. Subgroup analysis was performed by age, gender, sputum smear status, pulmonary cavities, and course of treatment, obtained from the clinical records. Haploview (version 4.2) was used to assess linkage disequilibrium (LD) of the SNP sites and haplotype analysis was performed using the SHEsis online software platform. Power calculations were performed with PASS software (version 11.0, NCSS, Kaysville, Utah, USA). P values were from two-tailed tests and statistical significance was set at P < 0.05. P values were then adjusted by Bonferroni correction for the ten polymorphisms examined so that a P-value < 0.005 (0.05/10) was considered statistically significant. The genotyping data were processed using GenomeStudio 2.0 and PLINK (version 1.90) for Windows, and statistical analyses were performed using SPSS Statistics (version 22.0, IBM, Chicago, USA).

RESULTS

Characteristics of the enrolled subjects

Of the total 409 subjects enrolled in the study, 401(98.04%) were successfully genotyped, including 130 TB cases, 108 LTBI individuals, and 163 HCs. The characteristics of the three groups are shown in Table 1. There was no significant difference in the gender distribution, but the TB patients were significantly younger than the LTBI and HC cohorts (P<0.05).

Table 1
Demographic and clinical characteristics of the study groups.

	TB (n=130)	LTBI (n=108)	HC (n=163)	TB vs. LTBI P-value	TB vs. HC P-value	LTBI vs. HC P-value
Age, mean ±S.D.	31.18±10.85	40.09±12.42	34.15±12.57	<0.001	0.034	<0.001
Gender, n (%)				0.286	0.415	0.729
Male	74(56.9)	54(50.0)	85(52.1)
Female	56(43.1)	54(50.0)	78(47.9)
Sputum smear status, n (%)
Smear-positive	53(40.8)
Smear-negative (Culture-positive)	77(59.2)
Pulmonary cavity, n (%)
Present	43(33.1)
Absent	87(66.9)
Course of treatment, n (%)
> 6 months	49(38.3)
≤ 6 months	79(61.7)

Association of BCL2 SNPs and TB / LTBI susceptibility

The reproducibility of the genotyping was 100% according to the duplicate genotyping results. The distribution of the genotypes of 10 BCL2 SNPs among 163 HCs fully met the Hardy Weinberg equilibrium (P>0.05). We found that five SNPs were significantly associated with TB or LTBI (rs80030866, rs3744933, rs9955190, rs1801018 and rs2551402, all P<0.05) (Table 2). However, the associations did not reach the statistical significance of 0.005 stipulated by the Bonferroni correction for 10 separate comparisons.

Table 2
The positive results of genetic association analysis in three groups.

SNPs	Genotypes*/Alleles	TB	LTBI	HC	TB vs. LTBI		TB vs. HC		LTBI vs. HC
		N (%)	N (%)	N (%)	P	OR (95%CI)	P	OR (95%CI)	P	OR (95%CI)
rs80030866	AA	65 (50.4)	42 (38.9)	70 (43.0)	Rf
A>G	GA	55 (42.6)	48 (44.4)	75 (46.0)	0.531	0.837 (0.512-1.369)	0.479	0.837 (0.512-1.369)	0.805	1.07 (0.623-1.838)
	GG	9 (7.0)	18 (16.7)	18 (11.0)	0.077	0.575 (0.238-1.388)	0.218	0.575 (0.238-1.388)	0.460	1.344 (0.613-2.949)
	A	185 (71.7)	132 (61.1)	215 (66.0)
	G	73 (28.3)	84 (38.9)	111 (34.0)	0.015	0.62 (0.42-0.91)	0.137	0.76 (0.54-1.09)	0.250	1.23 (0.86-1.76)
rs3744933	CC	97 (74.6)	92 (85.2)	139 (85.3)	Rf
C>A	AC	32 (24.6)	15 (13.9)	23 (14.1)	0.063	1.979 (0.965-4.058)	0.02	2.014 (1.105-3.672)	0.865	1.064 (0.519-2.184)
	AA	1 (0.8)	1 (0.9)	1 (0.6)	——	——	——	——	——	——
	C	226 (86.9)	199 (92.1)	301 (92.3)
	A	34 (13.1)	17 (7.9)	25 (7.7)	0.067	1.76 (0.95-3.25)	0.03	1.81(1.05-3.12)	0.931	1.03( 0.54-1.95)
rs9955190	AA	81 (62.3)	57 (52.8)	89 (54.6)	Rf
A>G	GA	47 (36.2)	38 (35.2)	65 (39.9)	0.463	1.254 (0.685-2.295)	0.513	0.85 (0.521-1.385)	0.467	0.819 (0.478-1.403)
	GG	2 (4.5)	13 (12.0)	9 (5.5)	0.028	0.165 (0.033-0.821)	0.098	0.265 (0.055-1.278)	0.279	1.684 (0.656-4.324)
	A	209 (80.4)	152 (70.4)	243 (74.5)
	G	51 (19.6)	64 (29.6)	83 (25.5)	0.011	0.58 (0.38-0.88)	0.094	0.71 (0.48-1.06)	0.285	1.23 (0.84-1.81)
rs1801018	AA	107 (82.3)	85 (78.7)	141 (86.5)	Rf
A>G	GA	22 (16.9)	19 (17.6)	22 (13.5)	0.786	1.108 (0.529-2.317)	0.368	1.346 (0.704-2.573)	0.358	1.382 (0.693-2.754)
	GG	1 (0.8)	4 (3.7)	0 (0.0)	——	——	——	——	——	——
	A	236 (90.8)	189 (87.5)	304 (93.3)
	G	24 (9.2)	27 (12.5)	22 (6.7)	0.201	0.68 (0.38-1.23)	0.336	1.35 (0.73-2.48)	0.022	1.97( 1.09-3.57)
rs2551402	AA	45 (34.6)	34 (31.5)	65 (39.9)	Rf
A>C	CA	63 (48.5)	50 (46.3)	75 (46.0)	0.733	1.114 (0.598-2.079)	0.454	1.215 (0.729-2.025)	0.533	1.196( 0.681-2.103)
	CC	22 (16.9)	24 (22.2)	23 (14.1)	0.422	0.725 (0.331-1.59)	0.430	1.328 (0.656-2.686)	0.037	2.166 (1.046-4.484)
	A	153 (58.8)	118 (54.6)	205 (62.9)
	C	107 (41.2)	98 (45.4)	121 (37.1)	0.355	0.84 (0.58-1.21)	0.319	1.18 (0.85-1.65)	0.055	1.41 (0.99-2.00)

• — = all those frequencies <0.03 will be ignored in analysis; ^*Adjusted for age and gender.

For rs80030866 (A>G), the frequency of the G allele was lower in the TB group (OR=0.62, 95%CI:0.42-0.91, P=0.015) compared with the LTBI group (Table 2).

For rs3744933 (C>A), the frequency of the A allele was higher in the TB group (OR=1.81, 95%CI:1.05-3.12, P=0.031) when compared to the HC group, and the heterozygous AC genotype was more frequent in the TB group than the CC genotype (OR=2.014, 95%CI:1.105-3.672, P=0.022) (Table 2). Compared to the HC group, the association of TB with the rs3744933 A allele remained significant using either dominant (OR=1.986, 95%CI: 1.099-3.587, P=0.023) or additive (OR=0.498, 95%CI: 0.273-0.907, P=0.023) genetic models (Table 3).

Table 3
The positive results of the association between BCL2 genotypes and LTBI/TB based on genetic model analysis.

SNPs	Models		Groups, N (%)			TB vs. LTBI		TB vs. HC		LTBI vs. HC
			TB	LTBI	HC	P*	OR (95%CI)*	P*	OR (95%CI)*	P*	OR (95%CI)*
rs3744933 C>A	Dominant	AA+AC	33 (25.4)	16 (14.8)	24 (14.7)	0.078	1.880 (0.932-3.793)	0.023	1.986 (1.099-3.587)	0.795	1.097 (0.544-2.214)
		CC	97 (74.6)	92 (85.2)	139 (85.3)	Rf
	Recessive	AA	1 (0.8)	1 (0.9)	1 (0.6)	——	——	——	——	——	——
		AC+CC	129 (99.2)	107 (99.1)	162 (99.4)
	Additive	AA+CC	98 (75.4)	93 (86.1)	140 (85.9)	0.060	0.502 (0.245-1.029)	0.023	0.498 (0.273-0.907)	0.878	0.945 (0.461-1.939)
		AC	32 (24.6)	15 (13.9)	23 (14.1)	Rf
rs2551402 A>C	Dominant	CC+CA	85 (65.4)	74 (68.5)	98 (60.1)	0.954	0.983 (0.550-1.759)	0.379	1.242 (0.767-2.012)	0.199	1.412 (0.834-2.388)
		AA	45 (34.6)	34 (31.5)	65 (39.9)	Rf
	Recessive	CC	22 (16.9)	24 (22.2)	23 (14.1)	0.281	0.681 (0.339-1.369)	0.595	1.191 (0.626-2.266)	0.044	1.960 (1.018-3.772)
		CA+AA	108 (83.1)	84 (77.8)	140 (85.9)	Rf
	Additive	CC+AA	67 (51.5)	58 (53.7)	88 (54.0)	0.427	0.799 (0.459-1.390)	0.638	0.894 (0.561-1.425)	0.770	1.078 (0.652-1.784)
		CA	63 (48.5)	50 (46.3)	75 (46.0)	Rf

• — = all those frequencies <0.03 will be ignored in analysis; ^*Adjusted for age and gender.

For rs9955190 (A>G), the frequency of the G allele was lower in the TB group when compared to the LTBI group (OR=0.58, 95%CI:0.38-0.88, P=0.011), and the TB group had significantly fewer GG genotypes than AA genotypes (OR=0.165, 95%CI:0.033-0.821, P=0.028) (Table 2).

For rs1801018 (A>G), the frequency of the G allele was higher in the LTBI group when compared to the HC group (OR=1.97, 95%CI:1.09-3.57, P=0.022) (Table 2).

For rs2551402 (A>C), the CC genotype was more common than the AA genotype in the LTBI group when compared to the HC group (OR=2.166, 95%CI:1.046-4.484, P=0.037) (Table 2), and significant in a recessive model (OR=1.960, 95%CI: 1.018-3.772, P=0.044) (Table 3).

Power analysis

We calculated the power of the study using ORs of 2.0, 3.0, and 4.0 to determine whether the study sample size was adequate for the detection of associations¹⁷. The results indicated that 8 of the analyzed SNPs had a power approaching 80% to find an association with an OR ≥ 2 (Table 4). However, for the rs3744933 and rs1801018 loci, the statistical power was insufficient, and larger sample sizes would be needed to explore possible associations with tuberculosis.

Table 4
Power of the study with different odds ratios (OR) in an allelic model.

SNPs	MAF	Power in TB vs. LTBI			Power in TB vs. HC			Power in LTBI vs. HC
		OR=2	OR=3	OR=4	OR=2	OR=3	OR=4	OR=2	OR=3	OR=4
rs1564483	0.3998	0.7531	0.987	0.9995	0.8345	0.9964	0.9999	0.7936	0.9927	0.9998
rs956572	0.4643	0.7499	0.9841	0.9991	0.8304	0.9954	0.9999	0.7883	0.9909	0.9997
rs12454712	0.4583	0.7508	0.9845	0.9992	0.8313	0.9955	0.9999	0.7894	0.9912	0.9997
rs80030866	0.3383	0.7414	0.9872	0.9996	0.8257	0.9964	1.000	0.7848	0.9927	0.9999
rs3744933	0.1349	0.5201	0.9252	0.9942	0.6213	0.9656	0.9986	0.583	0.9484	0.9967
rs9955190	0.2669	0.7045	0.9832	0.9995	0.7947	0.9948	0.9999	0.7531	0.99	0.9998
rs12458289	0.2907	0.7202	0.9852	0.9996	0.8081	0.9956	1.000	0.7668	0.9913	0.9998
rs949037	0.3065	0.7286	0.9861	0.9996	0.8151	0.996	1.000	0.774	0.992	0.9998
rs1801018	0.0923	0.4058	0.8376	0.9746	0.5014	0.9089	0.991	0.4708	0.8807	0.9839
rs2551402	0.4177	0.7537	0.9865	0.9994	0.8347	0.9962	0.9999	0.7935	0.9924	0.9998

• SNP = single nucleotide polymorphism; MAF = minor allele frequency.

Subgroup analyses

Subgroup analyses were performed on rs80030866 and rs9955190 based on age, gender, sputum smear status, pulmonary cavities, and course of treatment. We found significant differences in the allelic frequencies of rs80030866 (OR=0.545, 95%CI:0.318-0.933, P=0.026) and rs9955190 (OR=0.530, 95%CI:0.299-0.942, P=0.029) between TB and LTBI in males (Table 5). The rs80030866*GA+GG was less prevalent than the rs80030866*AA in patients requiring treatment for more than 6 months compared to patients treated for the standard 6 months or less (OR=0.461, 95%CI: 0.221-0.959, P=0.037) (Table 5).

Table 5
Association statistics of BCL2 rs80030866 and rs9955190.

SNP	Group	Genotype, n (%)		P	OR (95%CI)	Allele, n (%)		P	OR (95%CI)
		AA	GA+GG			A	G
rs80030866 A>G	Males TB	41 (55.4)	33 (44.6)	0.152	0.597 (0.294-1.212)	111 (75.0)	37 (25.0)	0.026	0.545 (0.318-0.933)
	Males LTBI	23 (42.6)	31 (57.4)			67 (62.0)	41 (38.0)
	Females TB	24 (43.6)	31 (56.4)	0.367	0.701 (0.324-1.517)	74 (67.3)	36 (32.7)	0.276	0.735 (0.423-1.280)
	Females LTBI	19 (35.2)	35 (64.8)			65 (0.60)	43 (0.40)
	≤ 31-year TB	45 (51.7)	42 (48.3)	0.062	0.444 (0.188-1.053)	124 (71.3)	50 (28.7)	0.221	0.684 (0.371-1.260)
	≤ 31-year LTBI	10 (32.3)	21 (67.7)			39 (62.9)	23 (37.1)
	>32-year TB	20 (47.6)	22 (52.4)	0.524	0.782 (0.367-1.667)	61 (72.6)	23 (27.4)	0.059	0.575 (0.322-1.025)
	>32-year LTBI	32 (41.6)	45 (58.4)			93 (60.4)	61 (39.6)
	Smear-positive	23 (43.4)	30 (56.6)	0.185	1.611 (0.795-3.267)	70 (66.0)	36 (34.0)	0.091	1.598 (0.925-2.761)
	Smear-negative	42 (55.3)	34 (44.7)			115 (75.7)	37 (24.3)
	Pulmonary Cavity (Yes)	23 (53.5)	20 (46.5)	0.618	0.830 (0.399-1.728)	63 (73.3)	23 (26.7)	0.696	0.891 (0.499-1.591)
	Pulmonary Cavity (No)	42 (48.8)	44 (51.2)			122 (70.9)	50 (29.1)
	Course of Treatment (>6 months)	30 (61.2)	19 (38.8)	0.037	0.461 (0.221-0.959)	74 (75.5)	24 (24.5)	0.211	0.695 (0.392-1.231)
	Course of Treatment (≤6 months)	32 (42.1)	44 (57.9)			105 (68.2)	49 (31.8)
rs9955190 A>G	Males TB	46 (62.2)	28 (37.8)	0.337	0.706 (0.346-1.439)	119 (80.4)	29 (19.6)	0.029	0.530 (0.299-0.942)
	Males LTBI	29 (53.7)	25 (46.3)			74 (68.5)	34 (31.5)
	Females TB	35 (62.5)	21 (37.5)	0.259	0.646 (0.302-1.382)	90 (0.80)	22 (0.20)	0.156	0.636 (0.339-1.191)
	Females LTBI	28 (51.9)	26 (48.1)			78 (72.2)	30 (27.8)
	≤ 31-year TB	58 (66.7)	29 (33.3)	0.828	0.909 (0.385-2.149)	143 (82.2)	31 (17.8)	0.788	0.903 (0.431-1.893)
	≤ 31-year LTBI	20 (64.5)	11 (35.5)			50 (80.6)	12 (19.4)
	>32-year TB	23 (53.5)	20 (46.5)	0.568	0.804 (0.381-1.699)	66 (76.7)	20 (23.3)	0.088	0.594 (0.326-1.085)
	>32-year LTBI	37 (48.1)	40 (51.9)			102 (66.2)	52 (33.8)
	Smear-positive	32 (60.4)	21 (39.6)	0.706	1.148 (0.559-2.360)	84 (79.2)	22 (20.8)	0.701	1.129 (0.608-2.097)
	Smear-negative	49 (63.6)	28 (36.4)			125 (81.2)	29 (18.8)
	Pulmonary Cavity (Yes)	27 (62.8)	16 (37.2)	0.936	0.970 (0.456-2.063)	70 (81.4)	16 (18.6)	0.773	0.908 (0.470-1.752)
	Pulmonary Cavity (No)	54 (62.1)	33 (37.9)			139 (79.9)	35 (20.1)
	Course of Treatment (>6 months)	32 (65.3)	17 (34.7)	0.441	0.747 (0.356-1.570)	80 (81.6)	18 (18.4)	0.589	0.839 (0.442-1.590)
	Course of Treatment (≤6 months)	45 (58.4)	32 (41.6)			123 (78.8)	33 (21.2)

Association between BCL2 haplotypes and TB / LTBI susceptibility

Two BCL2 haplotypes were constructed with Haploview (version 4.2), and relatively strong LDs were observed between rs80030866 and rs9955190 (pairwise D’=0.88, r²=0.51) and between rs12458289 and rs2551402 (pairwise D’=1.00, r²=0.56) (Figure 1). The rs80030866-rs9955190 AA haplotype was more frequent in the TB group than in the LTBI group (P=0.018193), while the GG haplotype was less frequent in the TB group than in either the LTBI group (P=0.003391) or the HC group (P=0.034859) (Table 6). No significant associations with TB were observed for the rs12458289-rs2551402 haplotypes.

Figure 1
Plot of linkage disequilibrium (LD) by Haploview (version 4.2). Pair-wise LD between all pairs of SNPs in BCL2 was evaluated by D’ and r2 statistics. The D’ and r2 values (%) are presented in the squares.

Table 6
Haplotype analysis of four SNPs of BCL2.

Haplotype	TB (freq)	LTBI (freq)	HC (freq)	TB vs. LTBI		TB vs. HC		LTBI vs. HC
				P	OR (95% CI)	P	OR (95% CI)	P	OR (95% CI)
rs80030866-rs9955190
AA	178.08 (0.690)	128.71 (0.596)	210.54 (0.646)	0.018193	1.592 (1.081-2.344)	0.166053	1.286 (0.900-1.837)	0.242827	0.808 (0.565-1.156)
GA	29.92 (0.116)	23.29 (0.108)	32.46 (0.100)	0.744731	1.100 (0.619-1.956)	0.487492	1.205 (0.712-2.040)	0.751721	1.095 (0.624-1.923)
GG	43.08 (0.167)	60.71 (0.281)	78.54 (0.241)	0.003391	0.519 (0.333-0.808)	0.034859	0.641 (0.423-0.971)	0.288908	1.236 (0.835-1.828)
rs12458289-rs2551402
AC	75.00 (0.288)	65.00 (0.301)	84.00 (0.258)	0.766381	0.942 (0.634-1.399)	0.404881	1.168 (0.810-1.683)	0.269410	1.240 (0.846-1.818)
CA	153.00 (0.588)	118.00 (0.546)	205.00 (0.629)	0.354998	1.188 (0.825-1.710)	0.319292	0.844 (0.604-1.179)	0.055233	0.711 (0.501-1.008)
CC	32.00 (0.123)	33.00 (0.153)	37.00 (0.114)	0.347447	0.778 (0.461-1.314)	0.720852	1.096 (0.662-1.815)	0.181882	1.408 (0.850-2.333)

• Haplotypes with a frequency of <0.05 in the subjects were not evaluated.

DISCUSSION

In this study, we focused on the relationship between 10 SNPs in the BCL2 gene and two states of TB infection – latent infection (LTBI) and active disease (TB). We compared the three study groups – HC, LTBI, and TB – for the frequency of alleles in the 10 SNPs to try to identify genetic markers associated with either LTBI or TB. We found that both the G allele of rs80030866 and the G allele of rs9955190 were more common in the LTBI group when compared to those with active TB. In addition, LTBI individuals were more likely to have the rs2551402 CC genotype than the AA genotype. Although several of these associations had P values < 0.05, they did not reach the more stringent 0.005 level of statistical significance stipulated by the Bonferroni correction for comparing 10 different polymorphisms. In a study with a larger cohort or fewer polymorphisms tested, the associations would likely reach significance.

The BCL2 gene, located on human chromosome 18, consists of three exons and two introns. BCL2 interacts with other pro and anti-apoptotic BCL2 family proteins to regulate mitochondrial permeability and affect endogenous apoptosis, thereby altering the fate of cells. Studies using in vitro¹⁸ and mouse models¹⁹ have shown that the expression levels of BCL2 in macrophages containing Mtb may affect the intracellular survival of the bacilli, and a population-based study suggested that the BCL2 expression levels may predict the onset of active TB at a very early stage after infection²⁰. Individuals with LTBI, as defined by a positive TST (tuberculin skin test) or IGRA test without clinical TB, can reactivate and develop active TB disease, but neither test can distinguish between LTBI and active TB nor predict which LTBI individuals will progress to active TB²¹. It is possible that BCL2 expression levels change with the status of Mtb infection and thus may provide a way to distinguish between latent and active TB infection, or perhaps serve as an early marker to predict which LTBI individuals will develop active TB.

Lyu et al.²² showed that BCL2 variants may be associated with drug-induced liver injury associated with anti-tuberculous therapy. There is an interaction between BCL2 and glutathione, an important regulator of lung inflammation, and restoration of BCL2 expression leads to the replenishment of glutathione and a reduction in the levels of reactive oxygen species²³. This suggests that BCL2 could help limit the lung damage caused by TB and thus may warrant investigation as an adjunct therapy for the treatment of TB.

The rs80030866 and rs9955190 SNPs map to the intron region of the BCL2 gene, and studies have shown that some introns encode small RNA transcripts that regulate gene expression and function²⁴. HaploReg (version 4.1, Massachusetts Institute of Technology, Cambridge, MA, USA) indicates that both rs80030866 and rs9955190 are related to promoter and enhancer histone marks. Chromatin modification of histone marks is known to modify gene expression²⁵, so perhaps rs80030866 and rs9955190 regulate the expression of BCL2. The rs2551402 SNP is in strong linkage disequilibrium with rs1462129 (r²=0.99), both of which are also located in the intron region of the BCL2 gene. A large two-stage study reported that rs2551402 and rs1462129 play important roles in the regulation of apoptosis²⁶. While these four SNPs are all located in the intron region of the BCL2 gene, fine-mapping and functional studies are needed to identify the causal SNPs and elucidate the biological mechanisms underlying the associations we found with TB/LTBI risk.

The association of LTBI susceptibility with polymorphisms in several genes has been reported (e.g., in SP110²⁷), but the relation of polymorphisms in BCL2 to LTBI has not been previously studied. Although we found no associations of LTBI or TB with the BCL2 variants rs1564483, rs956572, and rs12454712, it has been suggested that these may be associated with cancer²⁸ and the aging process¹⁵, and therefore these polymorphisms deserve to be investigated further.

Studies in various countries and settings have shown that the rates of TB are significantly higher in males than in females²⁹, reflected by a male-to-female ratio for worldwide case notifications of 1.6. Host genetic factors may influence the overall outcome of Mtb infection and account for part of this disparity. Wu et al.¹⁷ showed that the association of STAT4 rs4853542 with TB was different in males than in females: the rs4853542*A allele may be more important in the development of male TB than female TB. Similarly, we found that BCL2 rs80030866*G and rs9955190*G were associated with a decreased risk of TB in male subjects but not in female subjects. However, neither BCL2 SNP was significantly associated with TB in males after the Bonferroni correction.

The standard duration of therapy for drug-sensitive pulmonary tuberculosis is 6 months³⁰, but treatment duration depends on the extent of the tuberculosis disease and the response to therapy. Poltavskaya et al.³¹ showed that the rs8341*TT genotype appears to have a protective effect against a more severe form of schizophrenia. Similarly, our result showed that rs80030866*GA+GG was less common in patients who received treatment courses longer than the usual 6 months. However, this association is difficult to explain, as we included only newly diagnosed patients with drug-susceptible TB and found no significant association of BCL2 variants with indications of TB disease severity such as sputum smear status or the presence of cavities. Saranathan et al.³² also suggested that there was no significant association between TIRAP (Toll/interleukin-1 receptor domain-containing adaptor protein) variants and smear grade or cavity status among TB patients. The haplotype analysis found that the rs80030866-rs9955190 haplotype was also associated with TB, with the AA haplotype again more frequent and the GG haplotype less frequent in the TB group when compared to the LTBI group.

The main weaknesses of our study were the limited sample size and the lack of a replication cohort to verify the associations that we have found. We recognize that this is a preliminary study and that further work is required to validate the associations we have identified and explore the underlying mechanisms.

CONCLUSION

In summary, our study suggests that BCL2 gene polymorphisms may be correlated with the susceptibility to LTBI and TB in the Chinese Han population, which could potentially help to elucidate the relationship of apoptosis with the development of TB. To the best of our knowledge, this is the first report that associates BCL2 polymorphisms with TB/LTBI susceptibility. These data suggest that the BCL2 SNP associations that we have identified could perhaps serve as biomarkers for discriminating between latent and active TB infection.

REFERENCES

• ¹
  World Health Organization. Global tuberculosis report 2021. [cited 2022 Ago 12]. Available from: https://www.who.int/publications/i/item/9789240037021
  » https://www.who.int/publications/i/item/9789240037021
• ²
  Hogan AB, Jewell BL, Sherrard-Smith E, Vesga JF, Watson OJ, Whittaker C, et al. Potential impact of the COVID-19 pandemic on HIV, tuberculosis, and malaria in low-income and middle-income countries: a modelling study. Lancet Glob Heal. 2020;8:e1132-41.
• ³
  Lönnroth K, Migliori GB, Abubakar I, D’Ambrosio L, de Vries G, Diel R, et al. Towards tuberculosis elimination: an action framework for low-incidence countries. Eur Respir J. 2015;45:928-52.
• ⁴
  Carranza C, Pedraza-Sanchez S, de Oyarzabal-Mendez E, Torres M. Diagnosis for latent tuberculosis infection: new alternatives. Front Immunol. 2020;11:2006.
• ⁵
  Houben RM, Dodd PJ. The global burden of latent tuberculosis infection: a re-estimation using mathematical modelling. PLoS Med. 2016;13:e1002152.
• ⁶
  Vynnycky E, Fine PE. Lifetime risks, incubation period, and serial interval of tuberculosis. Am J Epidemiol. 2000;152:247-63.
• ⁷
  Takiff HE. Host genetics and susceptibility. In: Palomino JC, Leão SC, Ritacco V, editors. Tuberculosis 2007: from basic science to patient care. Brazil: Flying Publisher; 2007. p.207-62.
• ⁸
  Ghanavi J, Farnia P, Farnia P, Velayati AA. Human genetic background in susceptibility to tuberculosis. Int J Mycobacteriol. 2020;9:239-47.
• ⁹
  Grant AV, Sabri A, Abid A, Abderrahmani Rhorfi I, Benkirane M, Souhi H, et al. A genome-wide association study of pulmonary tuberculosis in Morocco. Hum Genet. 2016;135:299-307.
• ¹⁰
  Lam A, Prabhu R, Gross CM, Riesenberg LA, Singh V, Aggarwal S. Role of apoptosis and autophagy in tuberculosis. Am J Physiol Lung Cell Mol Physiol. 2017;313:L218-29.
• ¹¹
  Chernykh ER, Sakhno L V, Khonina MA, Tikhonova MA, Kozhevnikov VS, Nikonov SD, et al. T cell subsets undergoing apoptosis and anergy in patients with pulmonary tuberculosis. Probl Tuberk. 2002;7:43-8.
• ¹²
  Elliott TO, Owolabi O, Donkor S, Kampmann B, Hill PC, Ottenhoff TH, et al. Dysregulation of apoptosis is a risk factor for tuberculosis disease progression. J Infect Dis. 2015;212:1469-79.
• ¹³
  World Health Organization. Guidelines on the management of latent tuberculosis infection. [cited 2022 Ago 12]. Available from: https://www.who.int/publications/i/item/9789241548908
  » https://www.who.int/publications/i/item/9789241548908
• ¹⁴
  Andersen P, Munk ME, Pollock JM, Doherty TM. Specific immune-based diagnosis of tuberculosis. Lancet. 2000;356: 1099-104.
• ¹⁵
  Erdman VV, Nasibullin TR, Tuktarova IA, Mustafina OE. Association of polymorphic markers of CASP8, BCL2 and BAX genes with aging and longevity. Adv Gerontol. 2012;25:398-404.
• ¹⁶
  Zheng G, Zhang W, Xu J, Yuan A, Li Q, Gastwirth JL. Genetic risks and genetic model specification. J Theor Biol. 2016;403:68-74.
• ¹⁷
  Wu S, Wang M, Wang Y, Zhang M, He JQ. Polymorphisms in the STAT4 gene and tuberculosis susceptibility in a Chinese Han population. Microb Pathog. 2019;128:288-93.
• ¹⁸
  Klingler K, Tchou-Wong KM, Brändli O, Aston C, Kim R, Chi C, et al. Effects of mycobacteria on regulation of apoptosis in mononuclear phagocytes. Infect Immun. 1997;65:5272-8.
• ¹⁹
  Mogga SJ, Mustafa T, Sviland L, Nilsen R. Increased Bcl-2 and reduced Bax expression in infected macrophages in slowly progressive primary murine Mycobacterium tuberculosis infection. Scand J Immunol. 2002;56:383-91.
• ²⁰
  Sutherland JS, Hill PC, Adetifa IM, de Jong BC, Donkor S, Joosten SA, et al. Identification of probable early-onset biomarkers for tuberculosis disease progression. PLoS One. 2011;6:e25230.
• ²¹
  Campbell JR, Krot J, Elwood K, Cook V, Marra F. A systematic review on TST and IGRA tests used for diagnosis of LTBI in immigrants. Mol Diagnosis Ther. 2015;19:9-24.
• ²²
  Lyu M, Jiao L, Zhou J, Li H, Meng Z, Xie W, et al. Do genetic polymorphisms of B-cell CLL/lymphoma 2 confer susceptibility to anti-tuberculous therapy-associated drug-induced liver injury? Int J Infect Dis. 2020;91:223-31.
• ²³
  McCullough KD, Martindale JL, Klotz LO, Aw TY, Holbrook NJ. Gadd153 sensitizes cells to endoplasmic reticulum stress by down-regulating Bcl2 and perturbing the cellular redox state. Mol Cell Biol. 2001;21:1249-59.
• ²⁴
  Ding X, Qian J, Yang Y, Xu W, Liu D, Su B. Genetic variants of BCL2 gene predict clinical outcomes of non-small-cell lung cancer patients treated with platinum-based chemotherapy in a Chinese population. Am J Cancer Res. 2016;6:2310-22.
• ²⁵
  Liao X, Yap MK, Leung KH, Kao PY, Liu LQ, Yip SP. Genetic association study of KCNQ5 polymorphisms with high myopia. Biomed Res Int. 2017;2017:3024156.
• ²⁶
  Lin J, Lu C, Stewart DJ, Gu J, Huang M, Chang DW, et al. Systematic evaluation of apoptotic pathway gene polymorphisms and lung cancer risk. Carcinogenesis. 2012;33:1699-706.
• ²⁷
  Chang SY, Chen ML, Lee MR, Liang YC, Lu TP, Wang JY, et al. Sp110 polymorphisms are genetic markers for vulnerability to latent and active tuberculosis infection in Taiwan. Dis Markers. 2018;2018:4687380.
• ²⁸
  Xu P, Liu L, Wang J, Zhang K, Hong X, Deng Q, et al. Genetic variation in BCL2 3’-UTR was associated with lung cancer risk and prognosis in male Chinese population. PLoS One. 2013;8:e72197.
• ²⁹
  Hertz D, Schneider B. Sex differences in tuberculosis. Semin Immunopathol. 2019;41:225-37.
• ³⁰
  Bouchikh S, Stirnemann J, Prendki V, Porcher R, Kesthmand H, Morin AS, et al. Durée de traitement des tuberculoses extrapulmonaires : six mois ou plus ? Analyse de la base de données TB-INFO. Rev Med Interne. 2012;33:665-71.
• ³¹
  Poltavskaya EG, Fedorenko OY, Vyalova NM, Kornetova EG, Bokhan NA, Loonen AJ, et al. Genetic polymorphisms of PIP5K2A and course of schizophrenia. BMC Med Genet. 2020;21 Suppl 1:171.
• ³²
  Saranathan R, Sathyamurthi P, Thiruvengadam K, Murugesan S, Shivakumar S, Gomathi NS, et al. MAL adaptor (TIRAP) S180L polymorphism and severity of disease among tuberculosis patients. Infect Genet Evol. 2020;77:104093.

• FUNDING: This work was supported by the Sanming project of Medicine in Shenzhen (grant number SZSM201603029), the Natural Science Foundation of Guangdong Province (grant number 2018A030313123), and the Key Disciplines of Medicine in Nanshan District.

Publication Dates

• Publication in this collection
  30 Sept 2022
• Date of issue
  2022

History

• Received
  10 June 2022
• Accepted
  12 July 2022