Association of Genetic Variations in Energy Metabolism Genes with Noise-Induced Hearing Loss in a Chinese Population: A Case–Control Study

Abstract

Objective:

This study explored the relationship between susceptibility to single-nucleotide polymorphisms (SNP) and noise-induced hearing loss (NIHL) in a population exposed to occupational noise.

Methods:

Workers exposed to noise in a steel enterprise in Henan Province were included in the study. Workers with a hearing threshold of ≥40 dB (A) for binaural high-frequency (3000, 4000, and 6000 Hz) in the pure tone audiometry were included in the case group (393 workers in total). Individuals whose hearing threshold for any frequency (500, 1000, and 2000 Hz) was ≤25 dB (A) and an average hearing threshold of <35 dB (A) for high frequencies were included in the control group (731 individuals in total). A SNPscan multiple SNP typing kit was used for SNP typing, and PLINK software was used in analyzing the correlation between each gene locus and NIHL susceptibility. Cumulative noise exposure (CNE) was stratified as CNE < 97 dB (A) · year and CNE ≥ 97 dB (A) · year.

Results:

Regarding rs11204100, compared with subjects with the TT genotype, subjects with the TC+CC genotype were less susceptible to NIHL (odds ratio [OR] [95% CI] = 0.712 [0.554, 0.913], P = 0.009). After CNE stratification, subjects with the TC+CC genotype were less susceptible to NIHL than those with the TT genotype in the CNE ≥97 dB (A) · year group (OR [95% CI] = 0.614 [0.433, 0.871], P = 0.007). As for the rs10503675, subjects with the AG+GG genotype were less susceptible to NIHL than subjects with the AA genotype (OR [95% CI] = 0.797 [0.541, 0.925], P = 0.011) in the general population. Haplotype results showed that CGT (rs11204100-rs10503675-rs17412009) is associated with lowered susceptibility to NIHL.

Conclusion:

The ATP6V1B2 gene plays an important role in the risk of NIHL, and the C allele of rs11204100 and G allele of rs10503675 are associated with lowered susceptibility to NIHL.

KEY MESSAGES

• (1)Energy metabolism pathways play an important role in the process of hearing impairment
• (2)The core of energy metabolism is ATP, which plays an important role in the conversion of mechanical energy into electrical energy in hair cells.
• (3)The ATP6V1B2 gene is correlated with noise-induced hearing loss.

INTRODUCTION

Noise-induced hearing loss (NIHL) is a complex disease influenced by environmental and genetic factors.[1] Energy metabolism pathways play an important role in the process of hearing impairment.[2] Adenosine triphosphate (ATP) produced by mitochondria in outer hair cells supports the cells and blood vessels of the cochlea. ATP provides a considerable amount of energy for the physiological process of sound transmission in the cochlea,[3] and produces a large number of free radicals, such as reactive oxygen species (ROS).[4] Under normal physiological conditions, redundant ROS can be removed by the scavenging system of activated oxygen in the cochlea. However, continuous noise stimulation causes an energy metabolism overload of mitochondria in hearing cells, leading to an increase in ATP requirement for hair cells and an increase in oxygen consumption in hair and supporting cells. Therefore, energy metabolism genes related to the ATP pathway may be closely related to NIHL susceptibility. The core of energy metabolism is ATP, which plays an important role in the conversion of mechanical energy into electrical energy. ATP enzymes in the body mainly include Na⁺/K⁺-ATPase, Ca²⁺-ATPase, and H⁺-ATPase. These enzymes are responsible for converting electrochemical energy into mechanical energy and finally into chemical energy, which is stored in ATP terminals.[5] Our research focused on energy metabolism genes, including ATP1A2, ATP1A3, ATP1B1, ATP2B2, ATP6V0A4, ATP6V1B1, ATP6V1B2, and DNAH8. In humans, ATP1A3,[6,7] ATP6V0A4,[8] ATP6V1B1,[9,10] and ATP2B2[11,12,13] are associated with sensorineural hearing loss.

This study explored the relationship between gene polymorphisms and NIHL and identified relevant biomarkers potentially associated with the occurrence of NIHL. The aim was to provide a basis for the early diagnosis and prevention of subsequent NIHL.

MATERIALS AND METHODS

Objects and methods

From January 1, 2006 to December 31, 2015, workers in a steel plant in Henan Province were selected and included in a cohort study. The workers in this study were selected from this cohort and divided into case and control groups according to inclusion and exclusion criteria. The inclusion criteria were exposure to occupational noise (>80 dB[A]) and cumulative time of noise exposure of ≥3 years. The exclusion criteria were as follows: (1) have served in the air force or artillery; history of head trauma, explosive deafness or familial deafness; history of infectious diseases (mumps, measles, or rubella); and treatment with ototoxic drugs (aminoglycosides); (2) otitis media, Meniere syndrome, conductive hearing loss, exaggerated hearing loss, false hearing loss, sudden hearing loss, toxic hearing loss, infectious hearing loss, hearing loss due to tumors, and autoimmune diseases affecting hearing; (3) the pure tone audiogram is horizontal or nearly horizontal; and (4) more severe hearing loss in the speech spectrum as compared with hearing loss in high frequencies.

According to the Chinese diagnosis of occupational noise-induced deafness (GBZ 49-2014),[14] the case group was defined as an average binaural hearing loss reflecting a frequency ≥40 dB (A). The control group with a hearing threshold of ≤25 dB (A) and high-frequency average hearing threshold of <35 dB (A) at 500, 1000, and 2000 Hz frequency bands were selected. The subjects in the case and control groups were essentially doing the same type of work and had the same level of noise exposure. This study included 393 participants in the case group and 731 participants in the control group.

General physical examination and questionnaire survey

The face-to-face survey method was used for investigative inquiry. A questionnaire was designed to collect the following information: demographic characteristics (including gender and age), living habits (smoking and drinking), medical history (including family history of deafness and poisoning history), drug use history (aminoglycoside antibiotics and other drugs), occupational history, and noise exposure. Basic information about ear health included ear pain and trauma and the use of protective measures (including the use of ear plugs and other protective devices during working time). Face-to-face interviews were conducted by trained investigators, and all the physical examinations were performed by trained physicians. Parameters, such as weight, height, and systolic and diastolic blood pressure levels, were measured using a standard protocol, and blood samples were obtained.

Pure tone listening test and field measurement of noise

Pure tone hearing test

The procedures were performed by highly trained and experienced occupational health-monitoring doctors and based on the diagnosis of occupational noise-induced deafness (GBZ 49-2014).[14] In a sound insulation room with background noise <25 dB (A), the acoustics threshold was tested with an As216 audiometer (Interacoustics International Hearing Equipment and Instrument Company in Denmark) at 500, 1000, 2000, 3000, 4000, and 6000 Hz. Before the examination, the subjects were required to avoid an occupational noise for at least 12 hours, and the pure tone hearing test results were adjusted for age and sex according to “Acoustics Statistical distribution of hearing thresholds as a function of age” (GB/T 7582–2004).[15] Ear morphology and otoscopy were required, including bilateral auricle malformation, abnormal stenosis of the external ear canal, tympanic membrane perforation, adhesions, and calcification. On the basis of the “’Occupational Health Standard of the People’s Republic of China: Measurement of Physical Agents in Workplace − Part 8: Noise” (GBZ/T 189.8-2007),[16] the measuring instrument was calibrated according to the calibration requirements before measurements, and each subject’s 8h equivalent continuous A-weighted sound pressure level (A) was measured with a Noisepro multifunctional individual noise meter (Quest, USA; equivalent continuous A-weighted sound pressure level, L_Aeq and 8h). The cumulative noise exposure (CNE) of each operator was calculated on the basis of the sound pressure level to which workers were exposed obtained from the field occupational health survey and the occupational history of noise operation obtained from the questionnaire combined with the CNE of each worker recorded in each workplace over the years. The calculation formula is

where, n is the total number of different jobs exposed to noise operations, L_Aeq and 8h are the equivalent continuous A sound level, and T is the length of service of each segment.

Quality control

(1) The personnel who conducted the measurements had been trained and passed the examination. (2) The instruments and equipment used in the measurement were within the verification cycle. (3) A qualified sound calibrator was used in calibrating the instrument before and after measurement. When the deviation was greater than 0.5 dB, the measurement result was discarded, the measuring instrument was replaced, and measurement was performed again. (4) Information collectors were trained and qualified, collected information was checked on the same day, and missing items were timely communicated and supplemented. (5) Questionnaire input was double input, and consistency test was carried out. (6) Pure tone hearing threshold was measured by professional medical examiners with relevant work experience. Measurements were performed 24 hours after the subjects were pulled out from the noisy environment. (7) After the examination, the test results were corrected for age and gender.

Single-nucleotide polymorphism site selection and genotyping

Single-nucleotide polymorphisms (SNP) selection principles are based on the sites studied in previous studies and in promotor 2K⁺ all gene regions in accordance with the tagSNP+ functional region, and routine minor allele frequency of >0.1 and r² value of >0.8 were used as conditions for selecting research sites. Genotyping: 2 mL of peripheral blood was collected from each subject and transferred to an ethylenediaminetetraacetic acid (EDTA) tube, and 2 mL of genomic DNA extraction kit (Shanghai Lifefeng Biotech Co., Ltd., Shanghai, China) was used in extracting peripheral blood DNA. The concentration and purity of genomic DNA were detected using a Nano Photometer P360 (Shanghai Boyibio Biotech Co., Ltd., Shanghai, China) and stored in a −80°C refrigerator for future use. In this study, a SNPscan multiple SNP typing kit (Shanghai Tianhao Biotechnology Co., LTD., Shanghai, China) was used for SNP typing, and an ABI3730XL DNA analyzer was used for sequencing. GeneMapper 4.1 software (Applied Biosystems, Foster City, CA, USA) was used in analyzing and obtaining the genotypes of each SNP site.

Statistical analysis

Excel software was used for data entry, and SPSS 20.0 software (SPSS Inc., Chicago, IL, USA) was used for statistical analysis. Age, CNE, and average high-frequency binaural hearing loss were not consistent with normal distribution after the normality test and were represented by M (P25, P75). Wilcoxon’s rank sum test was used in comparing differences between the case and control groups. Classification variables, such as inter-group differences in hypertension, smoking, drinking, physical exercise, and population composition were statistically described by the case number and component ratio (%). Inter-group differences were compared with the χ² test, in which CNE was stratified, that is, CNE < 97 dB (A) year and CNE ≥ 97 dB (A) year. PLINK 1.9 software (Shaun Purcell, Harvard, USA, http://pngu.mgh.harvard.edu/purcell/plink/) was used in analyzing the correlation between each gene locus and NIHL susceptibility, and the Hardy–Weinberg equilibrium (HWE) was considered statistically significant at P-value of >0.05. The OR (95% CI) was considered statistically significant at P-value of <0.05. The linkage disequilibrium test (LD) command was used in establishing haplotype blocks and constructing the haplotype to determine the possible haplotype. PLINK software was used.

RESULTS

Basic characteristics

The segment and general attributes of the subjects are depicted in Table 1. No significant differences in age, sex, CNE, hypertension, smoking, and drinking were found between the groups (P > 0.05). However, a notable difference in the level of hearing loss was found as follows: mean of 49.50 dB (case group) versus 18.17 dB (control group; P < 0.001). Additionally, a significant difference in exercise habits was observed (P < 0.05; Table 1).

Table 1.

Basic characteristics of the case and control subjects

Characteristics	Case (n = 393)	Control (n = 731)	Z*/χ2#	P-value
Age/years, M (P25, P75)	42.167 (34.833, 47.750)	41.500 (32.417, 46.333)	−1.600*	0.110
Working year/years, M (P25, P75)	21.420 (9.250,27.750)	21.080 (9.080, 27.080)	−1.066*	0.286
CNE/dB (A) · year, M (P25, P75)	97.443 (95.101, 101.538)	97.252 (94.900, 101.076)	−1.181*	0.237
HTL/dB (A), M (P25, P75)	49.500 (43.833, 56.083)	18.167 (12.833, 24.500)	−27.283*	<0.001
Gender, n (%)
Male	376 (95.670)	698 (95.490)	0.021#	0.884
Female	17 (4.330)	33 (4.510)
Smoking, n (%)
Yes	253 (64.380)	429 (58.690)	3.468#	0.063
No	140 (35.620)	302 (41.310)
Drinking, n (%)
Yes	270 (68.700)	492 (67.310)	0.299#	0.633
No	123 (31.300)	239 (32.690)
Exercise, n (%)
Yes	162 (41.220)	350 (47.880)	4.569#	0.033
No	231 (58.780)	381 (52.120)
Hypertension, n (%)
Yes	161 (40.970)	303 (41.450)	0.025#	0.875
No	232 (59.030)	428 (58.550)

Notes: CNE = cumulative noise exposure; HTL= hearing threshold level; * = Z-value, # = χ²-value.

The basic information and genotypes of the SNPs of 34 energy metabolism genes in the case and control groups are shown in Table 2. rs2072462 and rs1026435 did not conform to HWE (P < 0.05) and were excluded from further analysis [Table 2].

Table 2.

Basic information of the 34 SNPs

Genes	SNP	A1	A2	MAF	A1A1/A1A2/A2A2		P (HWE)
					Case	Control
ATP1A2	rs2070703	T	C	0.152	6/104/283	22/182/527	0.644
ATP1B1	rs1200130	T	C	0.119	6/71/316	12/161/558	0.569
ATP1B1	rs1200135	A	G	0.143	9/88/296	15/185/531	0.807
ATP1B1	rs1358714	G	A	0.143	6/87/300	17/188/526	1.000
ATP1B1	rs1200137	T	C	0.118	4/77/312	8/165/558	0.389
ATP1B1	rs3766031	T	C	0.158	15/94/284	20/190/521	0.115
ATP6V1B1	rs11681642	C	T	0.390	71/171/151	105/353/273	0.531
ATP6V1B1	rs2266918	C	T	0.481	84/204/105	176/358/197	0.115
ATP6V1B1	rs2072462	C	T	0.369	16/256/121	28/486/217	<0.001
ATP2B2	rs35678	C	T	0.420	54/207/132	132/365/234	0.142
ATP2B2	rs2289273	A	G	0.269	17/159/217	59/294/378	0.449
ATP2B2	rs751122	T	C	0.343	54/178/161	82/321/328	0.643
DNAH8	rs1738254	A	G	0.420	83/180/130	123/352/256	0.358
DNAH8	rs1678664	C	T	0.267	38/139/216	47/290/394	0.445
DNAH8	rs3823430	G	A	0.379	63/189/141	90/357/284	0.311
DNAH8	rs1678674	A	G	0.258	38/135/220	46/277/408	0.161
DNAH8	rs874808	G	A	0.298	23/163/207	64/332/335	0.087
DNAH8	rs1678690	C	G	0.221	13/129/251	32/277/422	0.100
ATP6VA04	rs1026435	G	A	0.184	15/111/267	33/206/492	0.047
DNAH8	rs9357283	A	G	0.358	58/182/153	89/329/313	0.698
DNAH8	rs1678729	C	T	0.283	43/144/206	52/302/377	0.463
DNAH8	rs2061907	C	T	0.314	27/170/196	71/339/321	0.096
DNAH8	rs6458080	C	T	0.150	6/91/296	18/199/514	0.816
DNAH8	rs4714192	T	C	0.381	64/190/139	96/346/289	0.752
DNAH8	rs4452640	G	A	0.247	31/136/226	39/280/412	0.873
DNAH8	rs1537232	T	C	0.303	40/164/189	59/319/353	0.621
DNAH8	rs3737094	G	A	0.274	36/155/202	51/286/394	0.653
ATP6V0A4	rs3807154	A	G	0.170	11/108/274	28/196/507	0.169
ATP6V1B2	rs10091503	A	G	0.343	38/173/182	90/341/300	0.643
ATP6V1B2	rs11778205	G	A	0.259	28/133/232	56/282/393	0.188
ATP6V1B2	rs11204100	C	T	0.377	50/169/174	112/355/264	0.800
ATP6V1B2	rs10503675	G	A	0.177	13/96/284	19/238/474	0.608
ATP6V1B2	rs17412009	T	C	0.109	3/66/324	9/155/567	0.761
ATP1A3	rs919390	G	C	0.177	8/122/263	19/221/491	0.123

A1 = mutant type, A2 = wild type, HWE = Hardy–Weinberg equilibrium, MAF = mutant type frequency, SNP = single-nucleotide polymorphisms.

Association between energy metabolism genes, single-nucleotide polymorphisms, and noise-induced hearing loss

The frequencies of the rs874808 G allele were 0.266 and 0.315 in the case and control groups, respectively (OR [95% CI] = 0.789 [0.650, 0.957], P = 0.0160), indicating a significant difference in rs874808 G allele between the groups and low risk of NIHL. The frequencies of the rs2061907C allele were 0.285 and 0.329 in both groups (OR [95% CI] = 0.813 [0.673, 0.982], P = 0.032), indicating a significant difference in the rs2061907C allele between the groups. The C allele indicated a lowered risk of NIHL. The rs11204100C allele frequencies of the case and control groups were 0.342 and 0.396, respectively (OR [95% CI] = 0.794 [0.662, 0.951], P = 0.012), indicating a statistically significant difference in the rs11204100C allele between the groups and the lowered risk of NIHL. As for the rs10503675 G allele, the frequencies were 0.155 and 0.189 in the case and control groups, respectively (OR [95% CI] = 0.790 [0.625, 0.997], P = 0.047), indicating a statistically significant difference of G allele between cases and controls and the lowered risk of NIHL [Table 3].

Table 3.

Effects of alleles on susceptibility to NIHL

SNP	A1	F-A	F-U	P	OR (95% CI)	SNP	A1	F-A	F-U	P	OR (95% CI)
rs2070703	T	0.147	0.155	0.660	0.947 (0.743, 1.207)	rs1678690	C	0.197	0.233	0.494	0.808 (0.652, 1.000)
rs1200130	T	1.056	1.265	0.144	0.815 (0.619, 1.073)	rs9357283	A	0.379	0.347	0.127	1.150 (0.961, 1.377)
rs1200135	A	0.135	0.147	0.431	0.904 (0.703, 1.162)	rs1678729	C	0.293	0.278	0.454	1.076 (0.888, 1.303)
rs1358714	G	0.126	0.152	0.094	0.805 (0.624, 1.038)	rs2061907	C	0.285	0.329	0.032	0.813 (0.673, 0.982)
rs1200137	T	0.108	0.124	0.273	0.858 (0.653, 1.128)	rs6458080	C	0.131	0.161	0.060	0.787 (0.613, 1.011)
rs3766031	T	0.158	0.157	0.978	1.000 (0.791, 1.273)	rs4714192	T	0.405	0.368	0.088	1.167 (0.977, 1.394)
rs11681642	C	0.398	0.385	0.543	1.060 (0.885, 1.262)	rs4452640	G	0.252	0.245	0.712	1.038 (0.850, 1.269)
rs2266918	C	0.473	0.486	0.576	0.952 (0.800, 1.132)	rs1537232	T	0.310	0.299	0.571	1.056 (0.875, 1.274)
rs35678	C	0.401	0.430	0.177	0.886 (0.743, 1.056)	rs3737094	G	0.289	0.265	0.235	1.124 (0.927, 1.363)
rs2289273	A	0.246	0.282	0.065	0.830 (0.680, 1.011)	rs3807154	A	0.165	0.172	0.675	0.952 (0.755, 1.200)
rs751122	T	0.364	0.332	0.126	1.152 (0.961, 1.382)	rs10091503	A	0.317	0.356	0.059	0.838 (0.696, 1.007)
rs1738254	A	0.440	0.409	0.153	1.136 (0.953, 1.354)	rs11778205	G	0.241	0.270	0.134	0.858 (0.702, 1.048)
rs1678664	C	0.274	0.263	0.578	1.057 (0.870, 1.285)	rs11204100	C	0.342	0.396	0.012	0.794 (0.662, 0.951)
rs3823430	G	0.401	0.367	0.119	1.152 (0.964, 1.376)	rs10503675	G	0.155	0.189	0.047	0.790 (0.625, 0.997)
rs1678674	A	0.268	0.252	0.407	1.087 (0.893, 1.324)	rs17412009	T	0.092	0.118	0.052	0.751 (0.564, 1.004)
rs874808	G	0.266	0.315	0.016	0.789 (0.650, 0.957)	rs919390	G	0.176	0.177	0.925	0.989 (0.788, 1.242)

A1 = mutant type, F-A = case frequency, F-U = control frequency, SNP = single-nucleotide polymorphisms. Bold signifies a P-value of <0.05.

After adjusting for CNE, hypertension, exercise, smoking, and drinking status [Table 4], subjects carrying the G allele of rs874808 were significantly associated with the lowered risk of NIHL (OR [95% CI] = 0.795 [0.651, 0.971], P = 0.025). Subjects carrying the C allele of rs2061907 were significantly associated with the lowered risk of NIHL (OR [95% CI] = 0.821 [0.674, 0.999], P = 0.485). As for rs11204100, subjects carrying the C allele were significantly associated with the lowered risk of NIHL (OR [95% CI] = 0.790 [0.659, 0.948], P = 0.011). Regarding rs10503675, subjects carrying the G allele were significantly associated with the lowered risk of NIHL (OR [95% CI] = 0.782 [0.617, 0.991], P = 0.042).

Table 4.

Associations of mutant allele with the risk of NIHL

SNP	A1	A2	P	OR (95% CI)	SNP	A1	A2	P	OR (95% CI)
rs2070703	T	C	0.733	0.959 (0.751, 1.223)	rs1678690	C	G	0.058	0.808 (0.648, 1.007)
rs1200130	T	C	0.119	0.803 (0.610, 1.058)	rs9357283	A	G	0.155	1.140 (0.952, 1.364)
rs1200135	A	G	0.343	0.885 (0.688, 1.139)	rs1678729	C	T	0.544	1.061 (0.877, 1.284)
rs1358714	G	A	0.062	0.783 (0.606, 1.013)	rs2061907	C	T	0.048	0.821 (0.674, 0.999)
rs1200137	T	C	0.241	0.846 (0.639, 1.119)	rs6458080	C	T	0.945	0.805 (0.625, 1.038)
rs3766031	T	C	1.000	0.998 (0.790, 1.262)	rs4714192	T	C	0.124	1.152 (0.962, 1.380)
rs11681642	C	T	0.584	1.051 (0.880, 1.254)	rs4452640	G	A	0.781	1.029 (0.842, 1.258)
rs2266918	C	T	0.650	0.960 (0.807, 1.144)	rs1537232	T	C	0.570	1.057 (0.873, 1.280)
rs35678	C	T	0.130	0.869 (0.724, 1.042)	rs3737094	G	A	0.221	1.128 (0.930, 1.369)
rs2289273	A	G	0.065	0.826 (0.674, 1.012)	rs3807154	A	G	0.631	0.946 (0.753, 1.188)
rs751122	T	C	0.116	1.157 (0.965, 1.387)	rs10091503	A	G	0.052	0.830 (0.688, 1.001)
rs1738254	A	G	0.162	1.132 (0.952, 1.347)	rs11778205	G	A	0.130	0.858 (0.704, 1.046)
rs1678664	C	T	0.632	1.049 (0.864, 1.273)	rs11204100	C	T	0.011	0.790 (0.659, 0.948)
rs3823430	G	A	0.136	1.149 (0.957, 1.378)	rs10503675	G	A	0.042	0.782 (0.617, 0.991)
rs1678674	A	G	0.471	1.074 (0.885, 1.304)	rs17412009	T	C	0.052	0.747 (0.557, 1.003)
rs874808	G	A	0.025	0.795 (0.651, 0.971)	rs919390	G	C	0.923	1.012 (0.800, 1.28)

A1 = mutant type, A2 = wild-type, SNP = single-nucleotide polymorphisms, bold signifies a P-value of <0.05.

Stratification analysis of energy metabolism genes by CNE

Sites associated with the risk of NIHL in the general population were examined [Table 5]. Regarding rs2289273, subjects with the AA genotype were less susceptible to NIHL compared with subjects with the GG+GA genotype (OR [95% CI] = 0.515 [0.296, 0.896], P = 0.019). As for rs1678674, subjects with the AA genotype were more susceptible to NIHL than subjects with the GG+GA genotype (OR [95% CI] = 1.594 [1.018, 2.496], P = 0.044). For rs874808 and rs10503675, subjects with the AG+GG genotype were less susceptible to NIHL than subjects with the AA genotype (OR [95% CI] = 0.760 [0.595, 0.972], P = 0.033; OR [95% CI] = 0.797 [0.541, 0.925], P = 0.011). For rs1678690, subjects with the GC+CC genotype were less susceptible to NIHL than subjects with the GG genotype (OR [95% CI] = 0.773 [0.600, 0.995], P = 0.048). For rs1678729, subjects with the CC genotype were more susceptible to NIHL than subjects with the TT+TC genotype (OR [95% CI] = 1.604 [1.050, 2.452], P = 0.032). For rs11204100, subjects with the TC+CC genotype were less susceptible to NIHL than subjects with the TT genotype (OR [95% CI] = 0.712 [0.554, 0.913], P = 0.009).

Table 5.

General population and stratified analysis of associated SNPs by CNE

SNP	Model	Genotype	Case [n (%)]	Control [n (%)]	OR (95% CI)	P-value
General population
rs2289273	REC	GG+GA	376 (95.674)	672 (91.929)	1
		AA	17 (4.326)	59 (8.071)	0.515 (0.296, 0.896)	0.019
rs1678674	REC	GG+GA	355 (0.903)	685 (93.707)	1
		AA	38 (9.669)	46 (6.293)	1.594 (1.018, 2.496)	0.044
rs874808	DOM	AA	207 (52.672)	335 (45.828)	1
		AG+GG	186 (47.328)	396 (54.172)	0.760 (0.595, 0.972)	0.033
rs10503675	DOM	AA	284 (38.851)	474 (64.842)	1
		AG+GG	109 (27.735)	257 (35.157)	0.797 (0.541, 0.925)	0.011
rs1678690	DOM	GG	251 (63.868)	422 (57.729)	1
		GC+CC	142 (36.132)	309 (42.271)	0.773 (0.600, 0.995)	0.048
rs1678729	REC	TT+TC	350 (89.059)	679 (92.886)
		CC	43 (10.941)	52 (7.113)	1.604 (1.050, 2.452)	0.032
rs11204100	DOM	TT	174 (44.275)	264 (36.115)	1
		TC+CC	219 (55.725)	467 (63.885)	0.712 (0.554, 0.913)	0.009
CNE < 97 dB (A) · year
rs1200135	DOM	GG	151 (80.319)	247 (70.774)	1
		GA+AA	37 (19.681)	102(29.226)	0.593 (0.387, 0.910)	0.017
rs1358714	DOM	AA	151 (80.319)	249 (71.347)	1
		AG+GG	37 (19.681)	100 (28.653)	0.610 (0.398, 0.936)	0.023
rs1200137	DOM	AA	157 (83.511)	261 (74.785)	1
		AG+GG	31 (16.489)	88 (25.215)	0.586 (0.372, 0.923)	0.022
rs2289273	DOM	AA	6 (3.191)	29 (8.309)	1
		GG+GA	182 (96.809)	320 (91.691)	0.364 (0.148,0.893)	0.027
CNE ≥ 97 dB (A) · year
rs10091503	DOM	GG	95 (46.341)	140 (36.649)	1
		GA+AA	110 (53.659)	242 (63.351)	0.670 (0.475, 0.945)	0.027
rs11778205	DOM	AA	128 (62.439)	193 (50.524)	1
		GA+GG	77 (37.561)	189 (49.476)	0.614 (0.434, 0.869)	0.007
rs11204100	DOM	TT	90 (43.903)	124 (32.461)	1
		TC+CC	115 (56.097)	258 (67.539)	0.614 (0.433, 0.871)	0.007

CNE = cumulative noise exposure, bold signifies P < 0.05. OR (95%CI) = odds ratio (95% confidence interval), DOM = dominant model (dominant model refers to the phenomenon that in genetics, one allele has a stronger or more dominant influence on the phenotype than other alleles), REC = recessive model (recessive model is a genetic phenomenon in the genome that can be expressed only when an individual carries two of the same recessive alleles), SNP = single-nucleotide polymorphisms.

At noise exposure levels of <97 dB (A) [Table 5], for rs1200135, subjects with the GA+AA genotype were less susceptible to NIHL than subjects with the GG genotype (OR [95% CI] = 0.593 [0.387, 0.910], P = 0.017); Similarly, for rs1358714, rs1200137, and rs2289273, subjects with the AG+GG genotype were less susceptible to NIHL than subjects with the AA genotype (OR [95% CI] = 0.610 [0.398, 0.936], P = 0.023; OR [95% CI] = 0.586 [0.372, 0.923], P = 0.022; OR [95% CI] = 0.364 [0.148, 0.893], P = 0.027).

Three SNPs had a significant P-value for high noise levels (CNE≥ 97 dB(A) · year). For rs10091503, subjects with the GA+AA genotype were less susceptible to NIHL than subjects with the GG genotype (OR [95% CI] = 0.670 [0.475, 0.945], P = 0.027). For rs11778205, subjects with the GA+GG genotype were less susceptible to NIHL than subjects with the AA genotype (OR [95% CI] = 0.614 [0.434, 0.869], P = 0.007). For rs11204100, subjects with the TC+CC genotype were less susceptible to NIHL than subjects with the TT genotype (OR [95% CI] = 0.614 [0.433, 0.871], P = 0.007). The complete data on the association between SNP and NIHL risk in the general population, the CNE < 97 dB(A) year group, and the CNE ≥ 97 dB(A) year group are provided in Supplementary Material Tables 1-3.

Association of haplotypes with noise-induced hearing loss

Factors adjusted for logistic regression included CNE, hypertension, exercise, drinking, and smoking status. Haplotype blocks were estimated using the default procedure in Haploview, and only individuals with a non-missing phenotype were included in this analysis. Pairwise LD was only calculated for SNPs within 200 kb by default. Haplotype results showed that AGG (rs3823430-rs1678674-rs874808), CGT (rs11204100-rs10503675-rs17412009), and TAC (rs11204100-rs10503675-rs17412009) were significantly distributed in the case and control groups (P < 0.05). Subjects with the haplotypes AGG (rs3823430-rs1678674-rs874808) and CGT (rs11204100-rs10503675-rs17412009) were less susceptible to NIHL, and subjects with TAC (rs11204100-rs10503675-rs17412009) were more susceptible to NIHL [Table 6].

Table 6.

Analysis of haplotype

Haplotype blocks	Haplotype	P	OR	Haplotype blocks	Haplotype	P	OR
rs1200137-rs3766031	CT	0.989	0.998	rs1678690-rs9357283-rs1678729-rs2061907	GATT	0.141	1.140
rs1200137-rs3766031	TC	0.241	0.846	rs1678690-rs9357283-rs1678729-rs2061907	GGTT	0.743	0.930
rs1200137-rs3766031	CC	0.389	1.090	rs6458080-rs4714192	TT	0.124	1.150
rs1738254-rs1678664	AC	0.653	1.050	rs6458080-rs4714192	CC	0.095	0.805
rs1738254-rs1678664	AT	0.167	1.180	rs6458080-rs4714192	TC	0.758	0.973
rs1738254-rs1678664	GT	0.153	0.881	rs1537232- rs3737094	CG	0.244	1.120
rs3823430-rs1678674-rs874808	AGG	0.019	0.785	rs1537232- rs3737094	TA	0.611	1.050
rs3823430-rs1678674-rs874808	AAA	0.457	1.080	rs1537232- rs3737094	CA	0.108	0.862
rs3823430-rs1678674-rs874808	GGA	0.119	1.160	rs11204100-rs10503675-rs17412009	CGT	0.026	0.713
rs3823430-rs1678674-rs874808	AGA	0.883	0.974	rs11204100-rs10503675-rs17412009	CGC	0.518	0.893
rs1678690-rs9357283-rs1678729-rs2061907	CGTC	0.080	0.820	rs11204100-rs10503675-rs17412009	CAC	0.284	0.888
rs1678690-rs9357283-rs1678729-rs2061907	GGTC	0.566	0.913	rs11204100-rs10503675-rs17412009	TAC	0.019	1.240
rs1678690-rs9357283-rs1678729-rs2061907	GGCT	0.499	1.070	-	-	-	-

Haplotype blocks constitute a continuous region of the same chromosome that is in a state of linkage disequilibrium. Haplotype is a set of alleles inherited by an individual from a single parent. Bold signifies a P-value of <0.05

DISCUSSION

rs11204100 in the total population and CNE ≥97 dB (A) · year stratified analysis showed an association with the risk of NIHL, and rs874808 and rs10503675 sites suggested an association with the risk of NIHL in the CNE stratified analysis (CNE ≥ 97 dB [A] · year) results (P < 0.05). A set of alleles in a tissue is inherited entirely from either the father or from the mother and haplotypes are the sequences of these alleles, indicating where a homologous chromosome of each allele falls and containing more information than a genotype alone.[17] The effect of a single SNP of a gene may be weak at the onset of a disease, and haplotypes are better than a single SNP for analysis.[18] Haplotype results showed that AGG (rs3823430-rs1678674-rs874808), CGT (rs11204100-rs10503675-rs17412009), and TAC (rs11204100-rs10503675-rs17412009) were statistically significantly distributed in the case and control groups (P < 0.05). Subjects with the haplotypes AGG (rs3823430-rs1678674-rs874808) and CGT (rs11204100-rs10503675-rs17412009) were less susceptible to NIHL, whereas subjects with TAC (rs11204100-rs10503675-rs17412009) were more susceptible to NIHL. Thus, the C allele of rs11204100, the G allele of rs10503675, and the G allele of rs874808 were associated with lowered risk of NIHL.

The rs11204100 and G allele of rs10503675 belong to ATP6V1B2. ATP6V1B2 encodes the V1B2 subunit of V-ATPase. Vesicle-type proton pump (vesicle-type H⁺-ATPase and V-ATPase) is a special multi-submatrix pump that hydrolyses ATP and transports H⁺ against a concentration gradient.[19] The proton pump is responsible for the acidification of lysosomes and other membrane-binding compartments. Although the link between this gene and NIHL has not been confirmed, it has been linked to diseases associated with hearing loss. The ATP6V1B2 gene is highly expressed in the inner ear hair cells and spiral ganglion neurons of rats.[20] Qiu et al.[21] believed that ATP6V1B2 defects can cause the apoptosis of spiral ganglion neurons (SGNs) of the cochlea and affect hearing function; in mouse models with mutations in the ATP6V1B2 gene, lysosome dysfunction activates apoptosis, leading to the degeneration of SGNs and thereby causing hearing loss. The relationship between ATP6V1B2 and NIHL has not been examined. Here, we only find several case reports of ATP6V1B2 and deafness-related disorders, and the c.1516C>T mutation is associated with a genetic susceptibility to dominant deaf-nail dystrophy syndrome, which is an autosomal dominant disorder characterized by congenital sensorineural hearing loss accompanied by malnutrition or nail loss.[22,23] Zhao et al.[24] found that the disease caused by ATP6V1B2 is related to the damage of the hippocampus region; they investigated the pathological role of mutant ATP6V1B2 in the neurosensory system and generated a transgenic line of mice carrying c.1516C > T (p.Arg506∗) in Atp6v1b2, Atp6v1b2Arg506∗/Arg506∗(homozygous mutant); moreover, they identified that mutant mice display evident cognitive defects for which impairment in the hippocampal CA1 region might be the pathological basis. Dong et al.[25] demonstrated that the protein expression levels of matrix metalloproteinase (MMP)‑9 in the auditory cortex and hippocampus in 15‑month‑old mice were significantly higher than in the 3‑month‑old mice, indicating that MMP-9 is associated with age-related hearing loss. Although the relationship between ATP6V1B2 and occurrence of NIHL has not been explored, ATP6V1B2 gene mutation was found to cause hearing loss by affecting the hippocampus or directly damaging the cochlear nerve. Therefore, future studies on this gene should be increased.

In addition, the G allele of rs874808 in our results also suggests that it is related to NIHL, and rs874808 belongs to DNAH8. However, there have not been any previous studies on hearing loss and DNAH8. This result may be influenced by factors such as the study sample being sourced from a single factory, limitations in sample size, or racial restrictions. In reality, the effect strength of the gene itself may not be significant. Therefore, in order to further verify whether the gene locus is related to the occurrence of NIHL, as well as its possible mechanism and role, different industries can be collected. The biological information of different ethnic and occupational populations can also be validated using larger cohorts, or can be validated using animal experiments. Therefore, the mechanism of rs874808 and DNAH8 on NIHL is needed to be comprehensively analyzed and verified.

The main strength of this study is that the population survey strictly adhered to the sampling principles of the epidemiological survey. Second, all experiments (DNA extraction and genotyping) operators and participants were blinded to the disease state to draw relatively more objective conclusions.

This study also has some limitations that should be considered. Previous studies on the energy metabolism genes are insufficient to prove our results. Therefore, how ATP6V1B2 and its SNPs play a role in the occurrence and development of hearing loss and how the signaling pathway is governed should be further verified by animal experiments. In this study, the influences of high temperatures or other occupational factors were not excluded, and other occupational harmful factors and noise exert a synergistic effect. Furthermore, this study did not incorporate factors such as workers’ personal noise prevention awareness, the setting of factory protective facilities, and workers’ living habits into the research, all of which should be further explored. Once a hearing disease is diagnosed, there may be a conscious effort to protect hearing about the the hearing protection might overwhelm the genomic factors for reducing NIHL[26,27] as these factors are more or less affecting hearing.

CONCLUSION

The ATP6V1B2 gene plays an important role in the risk of NIHL, whereas the C allele of rs11204100 and G allele of 10503675 are associated with the lowered risk of NIHL. Genetic variations in the energy metabolism genes may play an important role in determining individual susceptibility to NIHL. These results provide new insight into the pathogenesis and early prevention of NIHL.

Conflicts of Interest

The authors have no conflicts of interest to declare.

Availability of data and materials

Data to support the findings of this study are available on reasonable request from the corresponding author.

Author contributions

Bing Wang: data acquisition, data analysis, literature search, manuscript preparation, manuscript editing;

Shanfa Yu: concepts, design, literature search, manuscript editing, manuscript review;

Jie Jiao: concepts, design, literature search, data analysis, manuscript editing, manuscript review;

Guizhen Gu: experimental studies, data acquisition, manuscript editing;

Guoshun Chen: experimental studies, data acquisition, manuscript editing;

Wenhui Zhou: experimental studies, data acquisition, manuscript editing;

Hui Wu: experimental studies, data acquisition, manuscript editing;

Yanhong Li: experimental studies, data acquisition, manuscript editing;

Huanling Zhang: experimental studies, data acquisition, manuscript editing.

Ethics Approval and Consent to Participate

Based on the Measures for the Ethical Review of Biomedical Research Involving Humans enacted by the People’s Republic of China, this study was approved by the Ethics Committee of the Third People’s Hospital of Henan Province (Henan Occupational Disease Hospital) (2013003). Prior to the survey, all subjects were informed of the purpose and content of the study and signed an informed consent form.

Funding Statement

This study was supported by the National Natural Science Foundation of China (81372940 and 81872574) and research Fund from National Science and Technology Infrastructure Program of the People’s Republic of China (2014BAI12B03).