Otology & Neurotology 35:358Y365 Ó 2014, Otology & Neurotology, Inc.
Genetic Variants of CDH23 Associated With Noise-Induced Hearing Loss *Tomasz Jarema Kowalski, †Malgorzata Pawelczyk, †Elzbieta Rajkowska, †Adam Dudarewicz, and †Mariola Sliwinska-Kowalska *Department of Orthopaedic and Musculoskeletal Trauma, Central Medical Clinic of Ministry of the Interior, Warsaw; and ÞDepartment of Audiology and Phoniatrics, Nofer Institute of Occupational Medicine, Lodz, Poland
Objectives: Noise-induced hearing loss (NIHL) is a complex disease resulting from the interaction between external and intrinsic/genetic factors. Based on mice studies, one of the most interesting candidate gene for NIHL susceptibility is CDH23encoding cadherin 23, a component of the stereocilia tip links. The aim of this study was to analyze selected CDH23 single nucleotide polymorphisms (SNPs) and to evaluate their interaction with environmental and individual factors in respect to susceptibility for NIHL in humans. Methods: A study group consisted of 314 worst-hearing and 313 best-hearing subjects exposed to occupational noise, selected out of 3,860 workers database. Five SNPs in CDH23 were genotyped using real-time PCR. Subsequently, the main effect of genotype and its interaction with selected environmental and individual factors were evaluated. Results: The significant results within the main effect of genotype were obtained for the SNP rs3752752, localized in exon 21. The effect was observed in particular in the subgroup of
young subjects and in those exposed to impulse noise; CC genotype was more frequent among susceptible subjects, whereas genotype CT appeared more often among resistant to noise subjects. The effect of this polymorphism was not modified by none of environmental/individual factors except for blood pressure; however, the latter one should be further investigated. Smoking was shown as an independent factor determining NIHL development. Conclusion: The results of this study confirm that CDH23 genetic variant may modify the susceptibility to NIHL development in humans, as it was earlier proven in mice. Because the differences between the 2 study groups were not necessarily related to susceptibility to noise but they also were prone to agerelated cochlear changes, these results should be interpreted with caution until replication in another population. Key Words: Association studyVCDH23VNoise-induced hearing lossVSNP. Otol Neurotol 35:358Y365, 2014.
Noise-induced hearing loss (NIHL) is a world-wide leading occupational health risk in industrialized countries and the second most common form of sensorineural hearing impairment, after presbyacusis. According to European Communities estimates (2004), approximately
20% of European workers are exposed, half or more of their working time, to noise so loud that they have to raise the voice to talk to other people. It has been estimated that in total as much as 500 million individuals might be at risk of developing NIHL (1). It is believed that NIHL is a complex disease resulting from the interaction between environmental and intrinsic factors. Over the last years, we have gathered much information concerning the genetic factors of NIHL, but this quest is definitely far from finished. To date, a number of association studies on candidate genes have been performed in humans. Rabinowitz et al. (2) demonstrated the protective effect for the GSTM1 genotype status on outer hair cell function in noise-exposed individuals, and Fortunato et al. (3) suggested a predisposing role of PON1 and SOD2; however, the results of both studies should be treated with caution as they were obtained in populations of a limited size (58 and 94 individuals, respectively). A much higher power to detect a
Address correspondence and reprint requests to Mariola SliwinskaKowalska, Department of Audiology and Phoniatrics, Nofer Institute of Occupational Medicine, 8 Sw. Teresy Street, 91-348 Lodz, Poland; E-mail: [email protected] Conflict of interests: This article has not been submitted for publication nor has it been published in whole or in part elsewhere. The study has not been supported by any pharmaceutical company. All authors listed on the title page have directly participated in the planning, execution, or analysis of this study and are not affiliated with any organization with financial interest in the subject matter discussed in the manuscript. The study was supported by the 6th European Framework Project under the Marie Curie Host Fellowship for the Transfer of Knowledge ‘‘NOISEHEAR’’ (Contract No. MTKD-CT-2004-003137) and by the State Scientific Committee for Research (Grant No. PB0911/P05/2004/26).
358
Copyright © 2014 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
CDH23 AND NOISE-INDUCED HEARING LOSS causative variant was achieved by Carlsson et al. (4), who, in the study of 215 Swedish individuals, excluded the influence of GJB2 35delG mutation on individual susceptibility to noise. In 2004, our department at the Nofer Institute of Occupational Medicine, Lodz, Poland, started an epidemiologic genetic study in cooperation with the Department of Medical Genetics at the University of Antwerp, Belgium. In a large cohort of Polish noiseexposed subjects (702 individuals), we definitely excluded GJB2 35delG as NIHL susceptibility variant (5), thereby confirming the results of Carlsson et al. (4). A thorough analysis of hundreds of SNPs in over 60 candidate genes performed at the University of Antwerp in the Polish and concurrently in the Swedish population has revealed that genetic variability in CAT (6); some potassium-recycling genes (7,8); PCDH15, MYO7, and HSP70 genes (9,10) might contribute to the development of NIHL. The effect of the last one was earlier shown also in the Chinese population (11). These studies, however, do not deplete the list of all NIHL candidate genes that should be analyzed. Based on animal studies, one of the most interesting genes to be investigated in humans is CDH23 encoding cadherin 23, a component of the stereocilia tip links, which are thought to gate mechanotransduction channel in hair cells (12). Cdh23 is the first and, so far, the only gene linked with predisposition to noise-induced hearing loss in waltzer mice (13). The genetic analyses revealed that homozygosity at Cdh23753A is a primary determinant of age-related hearing loss (AHL) as well (14). Genetic basis for AHL and NIHL seems to be correlated because in many studies, Ahl (locus for CDH23) was reported to render mice more susceptible to noise than strains that do not carry the susceptibility allele (15,16). Mutations in Cdh23 in mice cause stereocilia disorganization and lead to deafness and vestibular disorders. In humans, mutations in CDH23 lead to both nonsyndromic and syndromic hearing loss (e.g., Usher syndrome 1DYdeafness accompanied by retinitis pigmentosa and vestibular dysfunction). Despite a strong evidence that Cdh23 gene polymorphisms may play a role in susceptibility to noise in mice, to our best knowledge, to date, only one study by Yang et al. (17) described the positive association of CDH23 polymorphisms with NIHL in the population of Chinese workers exposed to noise. Unfortunately, not much information can be retained from this study as it was published in Chinese. More recently, PCDH15, a member of the cadherin superfamily encoding a protein that mediates calcium-dependent cell-cell adhesion has been indicated as NIHL susceptibility gene in the Polish and Swedish population (10). In this study, 63 polymorphisms in CDH23 were also examined, and the results were negative. However, the selection of polymorphisms was based mainly on different SNP databases and did not reflect the real SNP frequencies in the population under study. In 2008 within the framework of collaboration with National Institutes of Health (NIH), the sequence of CDH23 gene was determined in the Polish population, and the selection of its most frequent polymorphisms was
359
possible (18). Based on these results, and taking into account the positive associations with NIHL found by Konings et al. (10) for PCDH15, a member of cadherin superfamily genes, we found as well-grounded the extension of CDH23 association analysis for SNPs identified in the Polish population. Additionally, because it has been suggested that the genetic predisposition to NIHL is determined not only by a single gene but rather by an interaction of genes and environmental factors, here we present a thorough analysis of CDH23 polymorphisms along with their interaction with other environmental and individual factors known to determine NIHL development.
MATERIALS AND METHODS Subjects A study group consisted of noise-exposed male subjects. A detailed description of the methods used to select subjects has been published previously (8,19). Briefly, the study group was drawn from a database of more than 3,860 entries from noiseexposed workers. Information on individual features of these subjects was gathered by a questionnaire, which requested information on a general health state, individual factors predisposing to NIHL, including hereditary ones, and exposures to noise and ototoxic chemicals at workplace, among others. All subjects with a history of hearing loss in family, middle ear diseases, conductive hearing loss, meningitis, and the differences in hearing thresholds exceeding 40 dB between the right and the left ear were excluded from the study. For each subject, a Z-score that corrects for age, sex, tenure of exposure, and noise exposure level, was generated based on the ISO 1999 (ISO 1999VInternational Organization of Standardization, 1990). Subsequently, based on the mean hearing thresholds (HTL) for the left ear at 4 and 6 kHz, which are the frequencies most sensitive to noise trauma, we selected 20% most resistant (with better HTLs than predicted in the ISO model) and 20% most sensitive (with worse HTLs than predicted in the ISO model) to noise subjects. Such selection was made separately for workers of every single industry, that is, miners, high power station workers, dockyard workers, and paint and lacquer factory workers. The maximum Z-scores in the resistant subgroup were z = j0.39 at 4 kHz and z = j0.43 at 6 kHz. The minimum Zscores in the susceptible subgroup were z = 0.20 at 4 kHz and z = 0.12 at 6 kHz, thus no overlapping cases were noted. For the purpose of this study, 627 subjectsV314 susceptible and 313 resistant to noiseVhave been selected. The selected study groups have been further analyzed with Mann-Whitney U test in respect of age, tenure (years of work in noise conditions), noise exposure level, height, weight, and body mass index (BMI). Significant differences have been noticed for height and noise exposure level, and the latter one has been therefore included in statistical analysis as a confounding factor. The detailed description of the study groups is presented in Table 1. Because the mechanism and the degree of hearing loss differ between the exposure to continuous steady-state and impulse noise, the statistical analysis was subsequently performed in the subgroups exposed to the impulse noise. As for the entire group, here we also observed significant between-group difference for noise exposure level and further statistical analysis accounted for this feature. The detailed characteristics are presented in Table 2. Otology & Neurotology, Vol. 35, No. 2, 2014
Copyright © 2014 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
360
T. J. KOWALSKI ET AL. TABLE 1.
Feature
Characteristics of subjects selected to genetic analysis
Susceptible subjects (n = 314)
Resistant subjects (n = 313)
pa
40.4 (T7.9) 17.1 (T7.1) 174.0 (T7.05) 80.3 (T12.7) 26,5 (T3,4) 84.1 (T5.6)
41.3 (T6.5) 17.8 (T5.0) 175.2 (T8.9) 80.7 (T12.5) 26.1 (T3.4) 86.5 (T4.3)
0.212 0.138 0.003 0.818 0.116 0.0001
Age (yr) Tenure (yr) Height (cm) Weight (kg) Body mass index (BMI) Noise exposure level [dB(A)] a
Statistical analysis performed with Mann-Whitney U test.
Blood Collection From all subjects selected for genetic analysis, blood samples were collected, and DNA was isolated with QIAamp DNA Blood Mini Kit (QIAGEN) according to the manufacturer’s guidelines. Briefly, 200 Kl of blood was mixed with 20 Kl QIAGEN Protease and 200 Kl Buffer AL. The samples were then incubated at 56-C for 10 minutes. Next, 200 Kl of ethanol was added to the sample and mixed again by pulse-vortexing for 15 seconds. The mixture was subsequently applied to the QIAamp Spin Column, centrifuged at 8,000 rpm for 1 minute and placed in a clean collection tube. Next, 500 Kl of Buffer AW1 was added, and samples were centrifuged at 8,000 rpm for 1 minute and placed in a clean collection tube. The mixture was subsequently washed with 500 Kl of Buffer AW2 and centrifuged at 14,000 rpm for 3 minutes. The QIAamp Spin Column was thenplaced in a clean tube; 200 Kl of Buffer AE was added, which was followed by incubation at room temperature for 1 minute and centrifugation at 8,000 rpm for 1 minute. DNA samples were stored at j20-C.
SNP Selection and Genotyping SNP selection was based on the results of our previous analysis in CDH23 gene (NG_008835.1), where coding exons were sequenced and a total of 35 polymorphisms were identified (18). Of 35 SNPs identified by sequencing in 10 Polish workers, the choice of polymorphisms for the further analysis was performed from 19 variants localized in coding regions of the gene. We have selected 5 the most common variants representing different location within extracellular protein ectodomains, that is, location within domain ‘‘2’’ for rs3802720, ‘‘7’’ for rs3752751, ‘‘8’’ for rs3752752, and ‘‘19’’ for rs11592462 and rs10466026. We included 2 SNPs nonsynonymous and 3 synonymous. The detailed characteristics of selected SNP are presented in Table 3. The SNPs were genotyped with real-time PCR technique (iCycler iQ; Bio-Rad). The TaqMan probes and all necessary reagents were designed and supplied by Applied Biosystems (Foster City, CA, U.S.A.). The designed assay ID numbers for each SNP under study are provided in Table 3. All reactions
TABLE 2.
Data Quality Control Several data quality control steps were performed to eliminate the samples that did not meet quality standards. As individuals with an outlying genetic background, when compared with the rest of the population, decrease the power of a genetic association study, they were removed after running the CHECKHET program (20). Identity-by-state (IBS) values between pairs of individuals were calculated with GRR program to highlight unknown family relationships among subjects and to detect sample duplications (21). All individual SNPs were tested for Hardy-Weinberg equilibrium with a W2 test for goodness-of-fit using genotype transposer (22). To experimentally rule out the possibility of population stratification, we performed a genomic control analysis using the methods described by Devlin et al. (23). In brief, 72 random markers across the genome were genotyped. These markers were independent of each other. Each of the 72 markers was tested for association with the phenotype, and the W2 test statistic was calculated. The lambda inflation factor of the p values was obtained by calculating the median of these 72 W2 statistics and dividing it by 0.456, which is the median W2 in the absence of population stratification. The data quality control checks did not indicate any sample to be removed from the study.
Characteristics of subjects selected to genetic analysis, exposed to impulse noise
Feature Age (yr) Tenure (yr) Height (cm) Weight (kg) Body Mass Index (BMI) Noise exposure level [dB(A)] a
were performed according to the manufacturer’s guidelines. Briefly, 9 Kl of DNA was mixed with 10 Kl TaqMan PCR Mastermix No AmpErase UNG (2) and 1 Kl Assay Mix (20). The real-time PCR reaction consisted of the following steps: initial denaturation of 10 minutes at 95-C and 50 cycles of 15-second denaturation at 92-C and 1-minute annealing at 60-C. The results were analyzed with iCycler iQ real-time PCR Detection System (Bio-Rad). To limit the costs of genetic analysis, SNP genotyping was performed in 2 steps. In the first step, 60% to 85% samples (depending on SNP) were genotyped. After an initial statistical analysis, one genetic variant, namely, rs3752752, was indicated as possibly associated with hearing impairment, and subsequently, it was genotyped in the whole study group (627 subjects).
Susceptible subjects (n = 137)
Resistant subjects (n = 132)
pa
39.9 (T8.1) 17.1 (T6.9) 175.0 (T7.0) 81.4 (T13.6) 26.6 (T3.4) 84.8 (T3.4)
41.5 (T5.9) 18.1 (T4.0) 176.0 (T6.5) 81.9 (T11.9) 26.5 (T3.8) 85.8 (T2.7)
0.082 0.190 0.147 0.540 0.441 0.003
Statistical analysis performed with Mann-Whitney U test.
Otology & Neurotology, Vol. 35, No. 2, 2014
Copyright © 2014 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
CDH23 AND NOISE-INDUCED HEARING LOSS TABLE 3. rs number (SNP)
361
Characteristics of single nucleotide polymorphisms selected for the purpose of genetic analysis Exon
HapMap-CEU genotype frequencies
Amino acid change
Assay ID (applied biosystems)
rs3802720 (366C9T)
5
None
C_25592461_10
rs3752752 (2388C9T)
21
None
C_27477092_10
rs3752751 (2316C9T)
21
None
C_25806240_10
rs11592462 (5996G9C)
45
Ser1999Thr
C_11214268_10
rs10466026 (6130G9A)
46
C/C = 0.522 C/T = 0.363 T/T = 0.115 C/C = 0.469 C/T = 0.451 T/T = 0.080 C/C = 0.469 C/T = 0.451 T/T = 0.080 C/G = 0.550 C/C = 0.333 G/G = 0.117 G/G = 0.500 A/G = 0.402 A/A = 0.098
Glu2044Lys
C_2154820_10
Statistical Analysis All statistical analysis were performed with SPSS 12.0 software (SPSS Inc., Chicago, IL, U.S.A.; www.spss.com) and consisted of the following steps: 1. Analysis of the frequency of factors that might influence the development of NIHL in the susceptible and resistant to noise groups (W2 test). Here, the following factors were taken into account: noise exposure, organic solvents exposure, vibrations at workplace, impulse noise, the use of hearing protectors, smoking, pigmentation (eyes and hair color), alcohol consumption, head trauma, treatment with ototoxic drugs, acoustic trauma, blood pressure, hearing loss in the family, and tinnitus. 2. Evaluation of the genotype main effect on the disease status (logistic regression test), corrected for exposure and age. Similar analysis was performed in the subgroup of subjects exposed to impulse noise. All calculations were repeated in 3 noise exposure subgroups (G85 dB(A), 85Y92 db(A) and 992 dB(A)) and 3 age subgroups (G35, 35Y50, and 950 yr). 3. Analysis of correlation between genotype and auditory threshold level for the given frequencies, taking into account noise exposure level and age as confounding factors (ANCOVA). Here, SNP was regarded as independent variable, and an observed hearing threshold level was a dependent variable. 4. Analysis of interaction between genotype and environmental or individual factors that might influence the shift of hearing threshold level (logistic regression, Wald test). These were as follows: organic solvents, smoking, height, pigmentation, blood pressure, and impulse noise. Noise exposure level, organic solvents exposure index, and height were treated as quantitative variables. The discreet variables included genotype, smoking (subjects smoking at present and/or within the last 10 yr), pigmentation (darkVeyes brown/black, hairVbrown/black), blood pressure (systolicVbelow or above 120 mmHg, diastolicVbelow or above 90 mmHg; at least 1 parameter above the norm was decisive for the classification into the group with elevated blood pressure). For all tests applied in this study, p values at the 5% level were regarded as significant.
RESULTS Analysis of Individual and Environmental Factors in the Susceptible and Resistant to Noise Groups Analysis of the frequency of factors that might influence the development of NIHL in the susceptible and
resistant to noise groups revealed significant differences only for tinnitus, which was reported more frequently among subjects susceptible to noise. No differences in frequencies were observed for the remaining analyzed factors. The results are presented in Table 4. Evaluation of the Genotype Effect on the Disease Status To analyze whether selected polymorphisms in CDH23 gene influence the susceptibility to NIHL, differences in genotype frequencies were statistically compared between resistant and susceptible subjects. Here, the significant results were obtained for one SNP rs3752752, located in exon 21. The main effect of genotype was observed in the whole groups of workers as well as in the subgroup exposed to impulse noise. Because part of the study group was exposed to noise and organic solvents, we have performed additional analysis in the subgroup exposed only to noise and a subgroup exposed to noise and organic solvents. Significant results were obtained for rs3752752 only in the first subgroup under study (p = 0.021, data not shown), which implies this genetic variant might be associated with individual susceptibility to noise and not organic solvents-induced hearing loss. We also observed no significant effect of genotype in none of the 3 analyzed noise exposure subgroups ( Tables 5 and 6). Because the development of hearing loss may be associated not only with noise exposure but also age, we have compared the main effect of genotype and interaction of genotype with noise of rs3752752 variant in 3 age subgroups: younger than 35 years, 35 to 50, and older than 50 years. Here, significant results were obtained only for the main effect of genotype in young subjects (G35 yr old). The CC genotype was more frequent among susceptible subjects, whereas genotype CT appeared more often among resistant to noise subjects (Fig. 1). Analysis of Interaction Between Genotype and Environmental Individual Factors The interaction between genetic and environmental/ medical factors was analyzed for the SNP with significant Otology & Neurotology, Vol. 35, No. 2, 2014
Copyright © 2014 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
362 TABLE 4.
T. J. KOWALSKI ET AL. Prevalence (%) of selected environmental and individual factors that might influence hearing threshold level after exposure to noise in the study group
Environmenta/medical factor
Susceptible subjects (n = 314)
Resistant subjects (n = 313)
p
95.8 10.2 43.6 83.4 15.0 19.5 6.1 32.3 1.3 17.8 18.2 41.2 7.4 54.1 9.5 12.0 11.5 12.1 2.6 28.6 6.7
96.5 8.9 42.2 80.5 14.7 15.1 7.1 33.8 1.7 16.6 16.6 35.8 7.7 47.3 12.4 9.7 8.6 12.5 3.8 12.5 5.1
0.557 0.684 0.747 0.406 1.000 0.169 0.632 0.734 1.000 0.751 0.674 0.186 1.000 0.094 0.123 0.438 0.261 0.780 0.496 G0.001 0.498
Sex (men) Organic solvents exposure (yes) Impulse noise exposure (yes) Hearing protectors (yes) Vibrations nowadays (yes) Noise exposure in the past (yes) Organic solvents in the past (yes) Noise in the army (yes) Organic solvents in the army (yes) Noise in the leisure time (yes) Organic solvents in the leisure time (yes) Hair color (Brown/Black) Eyes color (Brown/Black) Smoking (yes)a Alcohol (yes)b Hypertension (yes) Head trauma in the past (yes) Ototoxic drugs (yes) Acoustic trauma (yes) Tinnitus (yes) Presbyacusis in family (yes) a b
Smoking nowadays and/or in the past (the last 10 yr). Frequently, at least once a week; 46 records missing.
p value within the main effect of genotype, that is, for rs3752752. Significant associations were observed for a single SNP under study and smoking. By contrast, height, pigmentation, and impulse noise yielded insignificant values. Additionally, none of the interactions of rs3752752 polymorphism and environmental factors was significant except for blood pressure. The results are presented in Table 7.
DISCUSSION The general aim of this study was to investigate whether genetic variability in CDH23 may underlie an increased susceptibility to the development of NIHL in the Polish population. Because hearing impairment may be caused not only by exposure to noise but also by other environmental and individual factors, we have assessed whether these factors alone or in combination with the genotype variant may increase the risk of NIHL development. We selected study subjects out of the 2 extremes of the phenotypic spectrum (resistant and susceptible), based on the ISO 1999 model, which corrects for age, sex, years of exposure, and noise exposure level. In our opinion, such selection might significantly increase the power to detect TABLE 5.
Evaluation of the genotype main effect on the disease status in the whole group and subgroups exposed to impulse noise
SNP (no. of genotyped samples) rs3752751 (n rs3752752 (n rs3802720 (n rs10466026 (n rs11592462 (n
the causative variants associated with NIHL over the previous studies where the selection was simply based on the degree of hearing loss. However, this method seems to have also one serious problem that makes interpretation difficult. Although in this study, both groups were relatively young (average about 40), they were already old enough that there should be substantial variation in thresholds even in the absence of noise exposure. In addition, the average noise exposures were rather modest (85 dBA), and the amount of noise-induced threshold shift would therefore be small compared with the amount of age-related threshold shift. Thus, the 2 study subgroups cannot be considered to differ only in susceptibility to noise. Those with the best hearing might be also highly resistant to aging changes, as well as other factors influencing hearing (ototoxic substances, diet) or their combination. Those with the poorest hearing may have the opposite set of traits. On the other hand, subjects susceptible to the development of NIHL might be susceptible also to the aging changes because similar pathogenic processes are involved in the development of both NIHL and presbyacusis. This has been also proven at the gene levelVseveral SNPs have been shown to have overlapping functions in both these complex disorders (5). Moreover, the genetic analyses in mice revealed that homozygosity at Cdh23753A is a primary determinant of
Main effect of genotype ( p value)
OR (95% CI)
Impulse noise ( p value)
0.191 0.037 0.499 0.525 0.184
0.81 (0.591Y1.112) 0.767 (0.598Y0.984) 1.104 (0.828Y1.471) 0.889 (0.618Y1.279) 1.194 (0.918Y1.553)
0.099 0.02 0.557 0.142 0.782
= 341) = 627) = 468) = 374) = 410)
Otology & Neurotology, Vol. 35, No. 2, 2014
Copyright © 2014 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
OR (95% CI) 0.684 0.632 1.139 0.655 1.059
(0.434Y1.079) (0.429Y0.933) (0.735Y1.766) (0.371Y1.156) (0.705Y1.589)
CDH23 AND NOISE-INDUCED HEARING LOSS TABLE 6. SNP rs3752751 rs3752752 rs3802720 rs10466026 rs11592462
363
The effect of genotypes on the disease status in 3 noise exposure subgroups
OR(95% CI)/p value
OR(95% CI)/p value
OR(95% CI)/p value
G85 dB
85Y92 dB
992 dB
0.761 (0.43Y1.346) /0.344 0.782 (0.497Y1.23) /0.285 0.939 (0.57Y1.546) /0.802 0.544 (0.279Y1.057) /0.07 1.124 (0.672Y1.88) /0.654
0.864 (0.547Y1.364) /0.528 0.737 (0.52Y1.042) /0.083 1.16 (0.749Y1.797) /0.505 1.069 (0.621Y1.839) /0.809 1.088 (0.758Y1.561) /0.645
1.504 (0.534Y4.234) /0.428 1.373 (0.583Y3.232) /0.459 1.323 (0.494Y3.54) /0.569 1.292 (0.402Y4.153) /0.659 2.345 (0.912Y6.026) /0.069
both age-related and noise-induced hearing loss (14). Hearing protectors could have less contribution to the variability of hearing thresholds because their use in a particular enterprise was very uniform. The initial analysis of our study groups revealed that resistant individuals were exposed to higher noise levels than susceptible subjects. To overcome a bias related to a so called ‘‘healthy workers syndrome,’’ we have included noise exposure level as a confounding factor in all statistical analysis. Another factor that differentiated susceptible subjects from resistant ones was height, whereby resistant individuals were slightly higher than susceptible ones. Despite certain literature data indicating that low height might be a risk factor in the development of hearing loss (24), our further analysis within the logistic regression framework did not confirm this hypothesis in the group of workers under study. Another feature that differentiated both subgroups (susceptible and resistant) was significantly more frequent prevalence of tinnitus among noise-sensitive individuals. This constitutes the replication of previously described significant correlation between tinnitus and NIHL (25). The genetic component of complex diseases, such as NIHL, has been attributed to a ‘‘large number of smalleffect common variants across the entire allele frequency spectrum or a large number of large-effect rare variants.’’ The genome-wide association studies (GWASs) are neither powered nor designed to detect variation under any of these models on a consistent basis, so there is insufficient empirical data to resolve this debate. Although the discovery of a rare variant near a common variant might
be particularly informative, rare variants are challenging to find since their low frequency renders current cohorts underpowered to detect all but the strongest effects, and lack of correlation to other markers often prevents them from being picked up by standard genotyping marker panels (26). Based on our earlier sequencing analysis of CDH23 done specifically for the Polish population (18), we have therefore focused on most frequent genetic variants, and for the current study, we selected 5 of 35 identified SNPs. Only one of these SNPs (rs3802720) was in common with previous study of Konings et al. (10). The selection in this study was based on the localization of the variant in respect of the promoter region, its frequency, significance for the amino acid change, and varied distribution in previously analyzed subjects. The selected SNPs were localized in exons 5, 21, 45, and 46, two of them leading to amino acid change. In many association studies, the selection of polymorphisms is based mainly on their localization and SNPs in coding sequences are preferred over those found in introns. However, different complex diseases, including NIHL, have varied background in different populations and it is hard to unambiguously determine which SNP selection strategy is the best. Also, the extent to which SNPs leading to amino acid change contribute to disease predisposition represents one of unanswered questions in human genetics. TABLE 7.
Interaction between rs3752752 polymorphism and environmental/medical factors
Interaction
p
rs3752752*smoking
rs3752752*height
rs3752752*complexion
rs3752752*blood pressure
rs3752752*impulse noise FIG. 1. Analysis of genotype frequencies of rs3752752 polymorphism across 3 age subgroups.
SNP SNP*noise smoking SNP*smoking SNP SNP*noise height SNP*height SNP SNP*noise complexion SNP*complexion SNP SNP*noise blood pressure SNP*blood pressure SNP SNP*noise impulse noise SNP*impulse noise
0.014 0.893 0.027 0.513 0.017 0.884 0.051 0.614 0.014 0.922 0.131 0.977 0.016 0.827 0.173 0.016 0.014 0.893 0.791 0.233
Otology & Neurotology, Vol. 35, No. 2, 2014
Copyright © 2014 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
364
T. J. KOWALSKI ET AL.
Statistically significant results within the main effect of genotype were obtained for one SNP localized in exon 21Yrs3752752. The significance level was 0.037 in the entire group and 0.02 in impulse noise-exposed subgroup. The statistical method used in this study strictly followed the protocol of all our previous association studies conducted in the collaboration with the Department of Medical Genetics at the University of Antwerp (6Y10). Such approach allowed to make data comparable. Indeed, a lack of positive association signal for rs3802720 variant in CDH23 gene in this study constitutes a replication of previous negative finding (10). A Bonferroni correction was not used because its application in gene polymorphism association studies is still a matter of a critical debate among genetic epidemiologists, as it can overcorrect true positive associations. Nevertheless, we cannot rule out that the positive finding regarding one CDH23 SNP of 5 tested may be spurious and due to chance. To avoid false-positive results, it has been agreed that the replication of an association in multiple independent sample sets should be considered more important than obtaining very small p values (27). In the study of Yang et al. (17), published in Chinese, the authors suggested other SNPs (rs1227049 and rs3802711) in CDH23 that play an important role in the development of NIHL in Chinese population. The difference in indicated genetic variants may be caused by ethnic differences between analyzed populations such as allele frequencies or LD patterns of associated regions or by different strategies of study groups’ selection for the purpose of genetic analysis. The observed differences in associated genetic variants in the 2 populations do not, in our opinion, disprove CDH23 as gene for NIHL. Moreover, it has been suggested that different SNPs but within the same gene may be regarded as replication in independent populations (27). It should be underlined that the rs3752752 variant was in particular significant in the subgroup of subjects exposed to impulse noise ( p = 0.02). Taking into account the expression pattern and the role of CDH23 in the inner ear, this finding is particularly important as impulse noise was found to lead to mechanical damage of hair cell stereocilia. As it has been earlier suggested, tip link abnormalities related to Cdh23ahl variant in mice may cause an increased probability in the opening of the mechanotransduction channels, which leads to the leakage of potassium cations and a frequent state of depolarization. This depolarization in turn opens calcium channels in the basolateral membrane causing excessive entry of Ca2+ that might lead to hair cell death (28). It has been also hypothesized that mutations in CDH23 may alter the structure of the encoded protein in such a way that the stereocilia bundles do not attain maximum strength and are easily damaged by even lower noise levels, thereby rendering the subject more susceptible to hearing loss when compared with the rest of population (29). Further analysis across three age subgroups (G35, 35Y50 and 950 yr old) of subjects susceptible and resistant
to noise revealed that rs3752752 SNP displayed its effect in young (G35 yr old) individuals and that the CC genotype was more frequent among susceptible subjects, whereas genotype CT appeared more often among resistant-tonoise subjects. Identification of genetic variants displaying protective effects on the auditory organ is particularly important in young individuals, which are shortly exposed to noise as they were described as mostly vulnerable to the development of noise-induced hearing loss (30). The subsequent analysis between medium auditory hearing threshold levels (HTLs) and rs3752752 CC and CT genotypes confirmed that HTLs indeed were significantly worse in carriers of CC genotype than in heterozygous subjects in both (left and right ear) at 4 and 6 kHz only, that is, the 2 frequencies that are most easily affected by NIHL (data not shown). NIHL is a complex disease; thus, the research regarding combined effects of genetic, environmental and medical/individual factors is of great importance. Most of already performed association studies focused only on identification of causal genes; however, this seems to be insufficient. Recently Carlsson et al. (31) have shown that the effect of smoking on susceptibility to NIHL is dependent on the presence of the GSTM1 deletion, suggesting a substantial interaction of genes and environmental factors in NIHL development. In this study, we show that smoking is independent factor for developing hearing loss, which is in agreement with several previous publications (32Y34). Our data show also that the effect of rs3752752 polymorphism is not modified by none of environmental/individual factors except for blood pressure. However, because ‘‘elevated blood pressure’’ was stated on one single measurement and not on the medical diagnosis of hypertension; this issue should be further investigated. To conclude, this study shows that genetic variation in CDH23 may be an independent major factor playing a role in determining individual susceptibility to NIHL, although a substantial variation in thresholds because of aging should be also taken into account. The effect seems to be particularly important in young individuals and in those exposed to impulse noise and can be related to the loss of mechanical stiffness of stereocilia. Taking into account shortcoming of the statistics, further analyses on CDH23 as NIHL susceptibility gene in carefully selected independent populations are required to confirm these findings. REFERENCES 1. Alberti P. Noise-induced hearing lossVa global problem. In: Prasher D, Luxon L, eds. Advances in noise research, Vol. 1, Protection against noise. London, UK: I. Whurr Publisher Ltd, 1998:7Y15. 2. Rabinowitz PM, Pierce Wise J, Hur Mobo B, et al. Antioxidant status and hearing function in noise-exposed workers. Hear Res 2002;173:164Y71. 3. Fortunato G, Marciano E, Zarrilli F, et al. Paraoxonase and superoxide dismutase gene polymorphisms and noise-induced hearing loss. Clin Chem 2004;50:2012Y8. 4. Carlsson PI, Borg E, Grip L, Dahl N, Bondeson ML. Variability in noise susceptibility in a Swedish population: the role of 35delG
Otology & Neurotology, Vol. 35, No. 2, 2014
Copyright © 2014 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
CDH23 AND NOISE-INDUCED HEARING LOSS
5. 6. 7. 8.
9. 10. 11. 12. 13. 14. 15. 16. 17. 18.
19.
mutation in the Connexin26 (GJB2) gene. Audiol Med 2004; 2:123Y30. Van Eyken E, Van Laer L, Fransen E, et al. The contribution of GJB2 (Connexin 26) 35delG to age related hearing impairment and noise-induced hearing loss. Otol Neurotol 2007;28:970Y5. Konings A, Van Laer L, Pawelczyk M, et al. Association between variations in CAT and noise-induced hearing loss in two independent noise-exposed populations. Hum Mol Genet 2007;16:1872Y83. Van Laer L, Carlsson PI, Ottschytsch N, et al. The contribution of genes involved in potassium recycling in the inner ear to noiseinduced hearing loss. Hum Mutat 2006;27:786Y95. Pawelczyk M, van Laer L, Fransen E, et al. Analysis of gene polymorphisms associated with K+ ion circulation in the inner ear of patients susceptible and resistant to noise-induced hearing loss. Ann Hum Genet 2009;73:411Y21. Konings A, Van Laer L, Michel S, et al. Variations in HSP70 genes associated with noise-induced hearing loss in two independent populations. Eur J Hum Genet 2008;17:329Y35. Konings A, van Laer L, Wiktorek-Smagur A, et al. Candidate gene association study for noise-induced hearing loss in two independent noise-exposed populations. Ann Hum Genet 2009;73:215Y24. Yang M, Tan H, Yang Q, et al. Association of hsp70 polymorphisms with risk of noise-induced hearing loss in Chinese automobile workers. Cell Stress Chaperones 2006;11:233Y9. Kazmierczak P, Sakaguchi H, Tokita J, et al. Cadherin 23 and protocadherin 15 interact to form tip-link filaments in sensory hair cells. Nature 2007;449:87Y91. Holme RH, Steel KP. Progressive hearing loss and increased susceptibility to noise-induced hearing loss in mice carrying a Cdh23 but not a Myo7a mutation. JARO 2004;5:66Y79. Noben-Trauth K, Zheng QY, Johnson KR. Association of cadherin 23 with polygenic inheritance and genetic modification of sensorineural hearing loss. Nat Genet 2003;35:21Y3. Erway LC, Shiau YW, Davis RR, Krieg EF. Genetics of age-related hearing loss in mice. III. Susceptibility of inbred and F1 hybrid strains to noise-induced hearing loss. Hear Res 1996;93:181Y7. Davis RR, Newlander JK, Ling X, Cortopassi GA, Krieg EF, Erway LC. Genetic basis for susceptibility to noise-induced hearing loss in mice. Hear Res 2001;155:82Y90. Yang M, Tan H, Zheng JR, et al. Association of cadherin CDH23 gene polymorphisms with noise induced hearing loss in Chinese workers. Wei Sheng Yan Jiu 2006;35:19Y22. Sliwinska-Kowalska M, Noben-Trauth K, Pawelczyk M, Kowalski TJ. Single Nucleotide Polymorphisms in the cadherin 23 (CDH23) gene in Polish workers exposed to industrial noise. Am J Hum Biol 2008;20:481Y3. Sliwinska-Kowalska M, Dudarewicz A, Kotylo P, ZamyslowskaSzmytke E, Pawlaczyk-Luszczynska M, Gajda-Szadkowska A. Individual susceptibility to noise-induced hearing loss: choosing an
365
optimal method of retrospective classification of workers into noisesusceptible and noise-resistant groups. IJOMEH 2006;19:235Y45. 20. Curtis D, North BV, Gurling HMD, Blaveri E, Sham PC. A quick and simple method for detecting subjects with abnormal genetic background in case-control samples. Am Hum Genet 2002;66: 235Y44. 21. Abecasis GR, Cherny SS, Cookson WOC, Cardon LR. GRR: graphical representation of relationship errors. Bioinformatics 2001; 17:742Y3. 22. Cox DG, Canzian F. Genotype transposer: automated genotype manipulation for linkage disequilibrium analysis. Bioinformatics 2001;17:738Y9. 23. Devlin B, Roeder K, Wasserman L. Genomic control, a new approach to genetic-based association studies. Theor Popul Biol 2001;60:155Y66. 24. Burr H, Lund SP, Sperling BB, Kristensen TS, Poulsen OM. Smoking and height as risk factors for prevalence and 5-year incidence of hearing loss. A questionnaire-based follow-up study of employees in Denmark aged 18Y59 years exposed and unexposed to noise. Int J Audiol 2005;44:531Y9. 25. Rosenhall U, Karlsson AK. Tinnitus in old age. Scand Audiol 1991;20:165Y71. 26. Raychaudhuri S. Mapping rare and common causal alleles for complex human diseases. Cell 2011;147:57Y69. 27. Neale BM, Sham PC. The future association studies: gene-based analysis and replication. Am J Hum Genet 2004;75:353Y36228. 28. Johnson KR, Yu H, Ding D, Jiang H, Gagnon LH, Salvi RJ. Separate and combined effects of SOD1 and Cdh23 mutations on agerelated hearing loss and cochlear pathology in C57BL/6J mice. Hear Res 2010;268:85Y92. 29. Wang Y, Hirose K, Liberman MC. Dynamics of noise-induced cellular injury and repair in the mouse cochlea. J Assoc Res Otolaryngol 2002;3:248Y68. 30 Corso JF. Age correction factor in noise-induced hearing loss: a quantitative model. Audiology 1980;19:221Y32. 31. Carlsson PI, Fransen E, Stenberg E, Bondeson ML. The influence of genetic factors, smoking and cardiovascular diseases on human noise susceptibility. Audiol Med 2007;5:82Y91. 32. Sliwinska-Kowalska M, Zamyslowska-Szmytke E, Szymczak W, et al. Effects of coexposure to noise and mixture of organic solvents on hearing in dockyard workers. J Occup Environ Med 2004; 46:30Y8. 33. Nomura K, Nakao M, Morimoto T. Effect of smoking on hearing loss: quality assessment and meta-analysis. Prev Med 2005;40: 138Y44. 34. Uchida Y, Nakashimat T, Ando F, Niino N, Shimokata H. Is there a relevant effect of noise and smoking on hearing? Int J Audiol 2005;44:86Y91.
Otology & Neurotology, Vol. 35, No. 2, 2014
Copyright © 2014 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
Genetic Variants of CDH23 Associated With Noise-Induced Hearing Loss *Tomasz Jarema Kowalski, †Malgorzata Pawelczyk, †Elzbieta Rajkowska, †Adam Dudarewicz, and †Mariola Sliwinska-Kowalska *Department of Orthopaedic and Musculoskeletal Trauma, Central Medical Clinic of Ministry of the Interior, Warsaw; and ÞDepartment of Audiology and Phoniatrics, Nofer Institute of Occupational Medicine, Lodz, Poland
Objectives: Noise-induced hearing loss (NIHL) is a complex disease resulting from the interaction between external and intrinsic/genetic factors. Based on mice studies, one of the most interesting candidate gene for NIHL susceptibility is CDH23encoding cadherin 23, a component of the stereocilia tip links. The aim of this study was to analyze selected CDH23 single nucleotide polymorphisms (SNPs) and to evaluate their interaction with environmental and individual factors in respect to susceptibility for NIHL in humans. Methods: A study group consisted of 314 worst-hearing and 313 best-hearing subjects exposed to occupational noise, selected out of 3,860 workers database. Five SNPs in CDH23 were genotyped using real-time PCR. Subsequently, the main effect of genotype and its interaction with selected environmental and individual factors were evaluated. Results: The significant results within the main effect of genotype were obtained for the SNP rs3752752, localized in exon 21. The effect was observed in particular in the subgroup of
young subjects and in those exposed to impulse noise; CC genotype was more frequent among susceptible subjects, whereas genotype CT appeared more often among resistant to noise subjects. The effect of this polymorphism was not modified by none of environmental/individual factors except for blood pressure; however, the latter one should be further investigated. Smoking was shown as an independent factor determining NIHL development. Conclusion: The results of this study confirm that CDH23 genetic variant may modify the susceptibility to NIHL development in humans, as it was earlier proven in mice. Because the differences between the 2 study groups were not necessarily related to susceptibility to noise but they also were prone to agerelated cochlear changes, these results should be interpreted with caution until replication in another population. Key Words: Association studyVCDH23VNoise-induced hearing lossVSNP. Otol Neurotol 35:358Y365, 2014.
Noise-induced hearing loss (NIHL) is a world-wide leading occupational health risk in industrialized countries and the second most common form of sensorineural hearing impairment, after presbyacusis. According to European Communities estimates (2004), approximately
20% of European workers are exposed, half or more of their working time, to noise so loud that they have to raise the voice to talk to other people. It has been estimated that in total as much as 500 million individuals might be at risk of developing NIHL (1). It is believed that NIHL is a complex disease resulting from the interaction between environmental and intrinsic factors. Over the last years, we have gathered much information concerning the genetic factors of NIHL, but this quest is definitely far from finished. To date, a number of association studies on candidate genes have been performed in humans. Rabinowitz et al. (2) demonstrated the protective effect for the GSTM1 genotype status on outer hair cell function in noise-exposed individuals, and Fortunato et al. (3) suggested a predisposing role of PON1 and SOD2; however, the results of both studies should be treated with caution as they were obtained in populations of a limited size (58 and 94 individuals, respectively). A much higher power to detect a
Address correspondence and reprint requests to Mariola SliwinskaKowalska, Department of Audiology and Phoniatrics, Nofer Institute of Occupational Medicine, 8 Sw. Teresy Street, 91-348 Lodz, Poland; E-mail: [email protected] Conflict of interests: This article has not been submitted for publication nor has it been published in whole or in part elsewhere. The study has not been supported by any pharmaceutical company. All authors listed on the title page have directly participated in the planning, execution, or analysis of this study and are not affiliated with any organization with financial interest in the subject matter discussed in the manuscript. The study was supported by the 6th European Framework Project under the Marie Curie Host Fellowship for the Transfer of Knowledge ‘‘NOISEHEAR’’ (Contract No. MTKD-CT-2004-003137) and by the State Scientific Committee for Research (Grant No. PB0911/P05/2004/26).
358
Copyright © 2014 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
CDH23 AND NOISE-INDUCED HEARING LOSS causative variant was achieved by Carlsson et al. (4), who, in the study of 215 Swedish individuals, excluded the influence of GJB2 35delG mutation on individual susceptibility to noise. In 2004, our department at the Nofer Institute of Occupational Medicine, Lodz, Poland, started an epidemiologic genetic study in cooperation with the Department of Medical Genetics at the University of Antwerp, Belgium. In a large cohort of Polish noiseexposed subjects (702 individuals), we definitely excluded GJB2 35delG as NIHL susceptibility variant (5), thereby confirming the results of Carlsson et al. (4). A thorough analysis of hundreds of SNPs in over 60 candidate genes performed at the University of Antwerp in the Polish and concurrently in the Swedish population has revealed that genetic variability in CAT (6); some potassium-recycling genes (7,8); PCDH15, MYO7, and HSP70 genes (9,10) might contribute to the development of NIHL. The effect of the last one was earlier shown also in the Chinese population (11). These studies, however, do not deplete the list of all NIHL candidate genes that should be analyzed. Based on animal studies, one of the most interesting genes to be investigated in humans is CDH23 encoding cadherin 23, a component of the stereocilia tip links, which are thought to gate mechanotransduction channel in hair cells (12). Cdh23 is the first and, so far, the only gene linked with predisposition to noise-induced hearing loss in waltzer mice (13). The genetic analyses revealed that homozygosity at Cdh23753A is a primary determinant of age-related hearing loss (AHL) as well (14). Genetic basis for AHL and NIHL seems to be correlated because in many studies, Ahl (locus for CDH23) was reported to render mice more susceptible to noise than strains that do not carry the susceptibility allele (15,16). Mutations in Cdh23 in mice cause stereocilia disorganization and lead to deafness and vestibular disorders. In humans, mutations in CDH23 lead to both nonsyndromic and syndromic hearing loss (e.g., Usher syndrome 1DYdeafness accompanied by retinitis pigmentosa and vestibular dysfunction). Despite a strong evidence that Cdh23 gene polymorphisms may play a role in susceptibility to noise in mice, to our best knowledge, to date, only one study by Yang et al. (17) described the positive association of CDH23 polymorphisms with NIHL in the population of Chinese workers exposed to noise. Unfortunately, not much information can be retained from this study as it was published in Chinese. More recently, PCDH15, a member of the cadherin superfamily encoding a protein that mediates calcium-dependent cell-cell adhesion has been indicated as NIHL susceptibility gene in the Polish and Swedish population (10). In this study, 63 polymorphisms in CDH23 were also examined, and the results were negative. However, the selection of polymorphisms was based mainly on different SNP databases and did not reflect the real SNP frequencies in the population under study. In 2008 within the framework of collaboration with National Institutes of Health (NIH), the sequence of CDH23 gene was determined in the Polish population, and the selection of its most frequent polymorphisms was
359
possible (18). Based on these results, and taking into account the positive associations with NIHL found by Konings et al. (10) for PCDH15, a member of cadherin superfamily genes, we found as well-grounded the extension of CDH23 association analysis for SNPs identified in the Polish population. Additionally, because it has been suggested that the genetic predisposition to NIHL is determined not only by a single gene but rather by an interaction of genes and environmental factors, here we present a thorough analysis of CDH23 polymorphisms along with their interaction with other environmental and individual factors known to determine NIHL development.
MATERIALS AND METHODS Subjects A study group consisted of noise-exposed male subjects. A detailed description of the methods used to select subjects has been published previously (8,19). Briefly, the study group was drawn from a database of more than 3,860 entries from noiseexposed workers. Information on individual features of these subjects was gathered by a questionnaire, which requested information on a general health state, individual factors predisposing to NIHL, including hereditary ones, and exposures to noise and ototoxic chemicals at workplace, among others. All subjects with a history of hearing loss in family, middle ear diseases, conductive hearing loss, meningitis, and the differences in hearing thresholds exceeding 40 dB between the right and the left ear were excluded from the study. For each subject, a Z-score that corrects for age, sex, tenure of exposure, and noise exposure level, was generated based on the ISO 1999 (ISO 1999VInternational Organization of Standardization, 1990). Subsequently, based on the mean hearing thresholds (HTL) for the left ear at 4 and 6 kHz, which are the frequencies most sensitive to noise trauma, we selected 20% most resistant (with better HTLs than predicted in the ISO model) and 20% most sensitive (with worse HTLs than predicted in the ISO model) to noise subjects. Such selection was made separately for workers of every single industry, that is, miners, high power station workers, dockyard workers, and paint and lacquer factory workers. The maximum Z-scores in the resistant subgroup were z = j0.39 at 4 kHz and z = j0.43 at 6 kHz. The minimum Zscores in the susceptible subgroup were z = 0.20 at 4 kHz and z = 0.12 at 6 kHz, thus no overlapping cases were noted. For the purpose of this study, 627 subjectsV314 susceptible and 313 resistant to noiseVhave been selected. The selected study groups have been further analyzed with Mann-Whitney U test in respect of age, tenure (years of work in noise conditions), noise exposure level, height, weight, and body mass index (BMI). Significant differences have been noticed for height and noise exposure level, and the latter one has been therefore included in statistical analysis as a confounding factor. The detailed description of the study groups is presented in Table 1. Because the mechanism and the degree of hearing loss differ between the exposure to continuous steady-state and impulse noise, the statistical analysis was subsequently performed in the subgroups exposed to the impulse noise. As for the entire group, here we also observed significant between-group difference for noise exposure level and further statistical analysis accounted for this feature. The detailed characteristics are presented in Table 2. Otology & Neurotology, Vol. 35, No. 2, 2014
Copyright © 2014 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
360
T. J. KOWALSKI ET AL. TABLE 1.
Feature
Characteristics of subjects selected to genetic analysis
Susceptible subjects (n = 314)
Resistant subjects (n = 313)
pa
40.4 (T7.9) 17.1 (T7.1) 174.0 (T7.05) 80.3 (T12.7) 26,5 (T3,4) 84.1 (T5.6)
41.3 (T6.5) 17.8 (T5.0) 175.2 (T8.9) 80.7 (T12.5) 26.1 (T3.4) 86.5 (T4.3)
0.212 0.138 0.003 0.818 0.116 0.0001
Age (yr) Tenure (yr) Height (cm) Weight (kg) Body mass index (BMI) Noise exposure level [dB(A)] a
Statistical analysis performed with Mann-Whitney U test.
Blood Collection From all subjects selected for genetic analysis, blood samples were collected, and DNA was isolated with QIAamp DNA Blood Mini Kit (QIAGEN) according to the manufacturer’s guidelines. Briefly, 200 Kl of blood was mixed with 20 Kl QIAGEN Protease and 200 Kl Buffer AL. The samples were then incubated at 56-C for 10 minutes. Next, 200 Kl of ethanol was added to the sample and mixed again by pulse-vortexing for 15 seconds. The mixture was subsequently applied to the QIAamp Spin Column, centrifuged at 8,000 rpm for 1 minute and placed in a clean collection tube. Next, 500 Kl of Buffer AW1 was added, and samples were centrifuged at 8,000 rpm for 1 minute and placed in a clean collection tube. The mixture was subsequently washed with 500 Kl of Buffer AW2 and centrifuged at 14,000 rpm for 3 minutes. The QIAamp Spin Column was thenplaced in a clean tube; 200 Kl of Buffer AE was added, which was followed by incubation at room temperature for 1 minute and centrifugation at 8,000 rpm for 1 minute. DNA samples were stored at j20-C.
SNP Selection and Genotyping SNP selection was based on the results of our previous analysis in CDH23 gene (NG_008835.1), where coding exons were sequenced and a total of 35 polymorphisms were identified (18). Of 35 SNPs identified by sequencing in 10 Polish workers, the choice of polymorphisms for the further analysis was performed from 19 variants localized in coding regions of the gene. We have selected 5 the most common variants representing different location within extracellular protein ectodomains, that is, location within domain ‘‘2’’ for rs3802720, ‘‘7’’ for rs3752751, ‘‘8’’ for rs3752752, and ‘‘19’’ for rs11592462 and rs10466026. We included 2 SNPs nonsynonymous and 3 synonymous. The detailed characteristics of selected SNP are presented in Table 3. The SNPs were genotyped with real-time PCR technique (iCycler iQ; Bio-Rad). The TaqMan probes and all necessary reagents were designed and supplied by Applied Biosystems (Foster City, CA, U.S.A.). The designed assay ID numbers for each SNP under study are provided in Table 3. All reactions
TABLE 2.
Data Quality Control Several data quality control steps were performed to eliminate the samples that did not meet quality standards. As individuals with an outlying genetic background, when compared with the rest of the population, decrease the power of a genetic association study, they were removed after running the CHECKHET program (20). Identity-by-state (IBS) values between pairs of individuals were calculated with GRR program to highlight unknown family relationships among subjects and to detect sample duplications (21). All individual SNPs were tested for Hardy-Weinberg equilibrium with a W2 test for goodness-of-fit using genotype transposer (22). To experimentally rule out the possibility of population stratification, we performed a genomic control analysis using the methods described by Devlin et al. (23). In brief, 72 random markers across the genome were genotyped. These markers were independent of each other. Each of the 72 markers was tested for association with the phenotype, and the W2 test statistic was calculated. The lambda inflation factor of the p values was obtained by calculating the median of these 72 W2 statistics and dividing it by 0.456, which is the median W2 in the absence of population stratification. The data quality control checks did not indicate any sample to be removed from the study.
Characteristics of subjects selected to genetic analysis, exposed to impulse noise
Feature Age (yr) Tenure (yr) Height (cm) Weight (kg) Body Mass Index (BMI) Noise exposure level [dB(A)] a
were performed according to the manufacturer’s guidelines. Briefly, 9 Kl of DNA was mixed with 10 Kl TaqMan PCR Mastermix No AmpErase UNG (2) and 1 Kl Assay Mix (20). The real-time PCR reaction consisted of the following steps: initial denaturation of 10 minutes at 95-C and 50 cycles of 15-second denaturation at 92-C and 1-minute annealing at 60-C. The results were analyzed with iCycler iQ real-time PCR Detection System (Bio-Rad). To limit the costs of genetic analysis, SNP genotyping was performed in 2 steps. In the first step, 60% to 85% samples (depending on SNP) were genotyped. After an initial statistical analysis, one genetic variant, namely, rs3752752, was indicated as possibly associated with hearing impairment, and subsequently, it was genotyped in the whole study group (627 subjects).
Susceptible subjects (n = 137)
Resistant subjects (n = 132)
pa
39.9 (T8.1) 17.1 (T6.9) 175.0 (T7.0) 81.4 (T13.6) 26.6 (T3.4) 84.8 (T3.4)
41.5 (T5.9) 18.1 (T4.0) 176.0 (T6.5) 81.9 (T11.9) 26.5 (T3.8) 85.8 (T2.7)
0.082 0.190 0.147 0.540 0.441 0.003
Statistical analysis performed with Mann-Whitney U test.
Otology & Neurotology, Vol. 35, No. 2, 2014
Copyright © 2014 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
CDH23 AND NOISE-INDUCED HEARING LOSS TABLE 3. rs number (SNP)
361
Characteristics of single nucleotide polymorphisms selected for the purpose of genetic analysis Exon
HapMap-CEU genotype frequencies
Amino acid change
Assay ID (applied biosystems)
rs3802720 (366C9T)
5
None
C_25592461_10
rs3752752 (2388C9T)
21
None
C_27477092_10
rs3752751 (2316C9T)
21
None
C_25806240_10
rs11592462 (5996G9C)
45
Ser1999Thr
C_11214268_10
rs10466026 (6130G9A)
46
C/C = 0.522 C/T = 0.363 T/T = 0.115 C/C = 0.469 C/T = 0.451 T/T = 0.080 C/C = 0.469 C/T = 0.451 T/T = 0.080 C/G = 0.550 C/C = 0.333 G/G = 0.117 G/G = 0.500 A/G = 0.402 A/A = 0.098
Glu2044Lys
C_2154820_10
Statistical Analysis All statistical analysis were performed with SPSS 12.0 software (SPSS Inc., Chicago, IL, U.S.A.; www.spss.com) and consisted of the following steps: 1. Analysis of the frequency of factors that might influence the development of NIHL in the susceptible and resistant to noise groups (W2 test). Here, the following factors were taken into account: noise exposure, organic solvents exposure, vibrations at workplace, impulse noise, the use of hearing protectors, smoking, pigmentation (eyes and hair color), alcohol consumption, head trauma, treatment with ototoxic drugs, acoustic trauma, blood pressure, hearing loss in the family, and tinnitus. 2. Evaluation of the genotype main effect on the disease status (logistic regression test), corrected for exposure and age. Similar analysis was performed in the subgroup of subjects exposed to impulse noise. All calculations were repeated in 3 noise exposure subgroups (G85 dB(A), 85Y92 db(A) and 992 dB(A)) and 3 age subgroups (G35, 35Y50, and 950 yr). 3. Analysis of correlation between genotype and auditory threshold level for the given frequencies, taking into account noise exposure level and age as confounding factors (ANCOVA). Here, SNP was regarded as independent variable, and an observed hearing threshold level was a dependent variable. 4. Analysis of interaction between genotype and environmental or individual factors that might influence the shift of hearing threshold level (logistic regression, Wald test). These were as follows: organic solvents, smoking, height, pigmentation, blood pressure, and impulse noise. Noise exposure level, organic solvents exposure index, and height were treated as quantitative variables. The discreet variables included genotype, smoking (subjects smoking at present and/or within the last 10 yr), pigmentation (darkVeyes brown/black, hairVbrown/black), blood pressure (systolicVbelow or above 120 mmHg, diastolicVbelow or above 90 mmHg; at least 1 parameter above the norm was decisive for the classification into the group with elevated blood pressure). For all tests applied in this study, p values at the 5% level were regarded as significant.
RESULTS Analysis of Individual and Environmental Factors in the Susceptible and Resistant to Noise Groups Analysis of the frequency of factors that might influence the development of NIHL in the susceptible and
resistant to noise groups revealed significant differences only for tinnitus, which was reported more frequently among subjects susceptible to noise. No differences in frequencies were observed for the remaining analyzed factors. The results are presented in Table 4. Evaluation of the Genotype Effect on the Disease Status To analyze whether selected polymorphisms in CDH23 gene influence the susceptibility to NIHL, differences in genotype frequencies were statistically compared between resistant and susceptible subjects. Here, the significant results were obtained for one SNP rs3752752, located in exon 21. The main effect of genotype was observed in the whole groups of workers as well as in the subgroup exposed to impulse noise. Because part of the study group was exposed to noise and organic solvents, we have performed additional analysis in the subgroup exposed only to noise and a subgroup exposed to noise and organic solvents. Significant results were obtained for rs3752752 only in the first subgroup under study (p = 0.021, data not shown), which implies this genetic variant might be associated with individual susceptibility to noise and not organic solvents-induced hearing loss. We also observed no significant effect of genotype in none of the 3 analyzed noise exposure subgroups ( Tables 5 and 6). Because the development of hearing loss may be associated not only with noise exposure but also age, we have compared the main effect of genotype and interaction of genotype with noise of rs3752752 variant in 3 age subgroups: younger than 35 years, 35 to 50, and older than 50 years. Here, significant results were obtained only for the main effect of genotype in young subjects (G35 yr old). The CC genotype was more frequent among susceptible subjects, whereas genotype CT appeared more often among resistant to noise subjects (Fig. 1). Analysis of Interaction Between Genotype and Environmental Individual Factors The interaction between genetic and environmental/ medical factors was analyzed for the SNP with significant Otology & Neurotology, Vol. 35, No. 2, 2014
Copyright © 2014 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
362 TABLE 4.
T. J. KOWALSKI ET AL. Prevalence (%) of selected environmental and individual factors that might influence hearing threshold level after exposure to noise in the study group
Environmenta/medical factor
Susceptible subjects (n = 314)
Resistant subjects (n = 313)
p
95.8 10.2 43.6 83.4 15.0 19.5 6.1 32.3 1.3 17.8 18.2 41.2 7.4 54.1 9.5 12.0 11.5 12.1 2.6 28.6 6.7
96.5 8.9 42.2 80.5 14.7 15.1 7.1 33.8 1.7 16.6 16.6 35.8 7.7 47.3 12.4 9.7 8.6 12.5 3.8 12.5 5.1
0.557 0.684 0.747 0.406 1.000 0.169 0.632 0.734 1.000 0.751 0.674 0.186 1.000 0.094 0.123 0.438 0.261 0.780 0.496 G0.001 0.498
Sex (men) Organic solvents exposure (yes) Impulse noise exposure (yes) Hearing protectors (yes) Vibrations nowadays (yes) Noise exposure in the past (yes) Organic solvents in the past (yes) Noise in the army (yes) Organic solvents in the army (yes) Noise in the leisure time (yes) Organic solvents in the leisure time (yes) Hair color (Brown/Black) Eyes color (Brown/Black) Smoking (yes)a Alcohol (yes)b Hypertension (yes) Head trauma in the past (yes) Ototoxic drugs (yes) Acoustic trauma (yes) Tinnitus (yes) Presbyacusis in family (yes) a b
Smoking nowadays and/or in the past (the last 10 yr). Frequently, at least once a week; 46 records missing.
p value within the main effect of genotype, that is, for rs3752752. Significant associations were observed for a single SNP under study and smoking. By contrast, height, pigmentation, and impulse noise yielded insignificant values. Additionally, none of the interactions of rs3752752 polymorphism and environmental factors was significant except for blood pressure. The results are presented in Table 7.
DISCUSSION The general aim of this study was to investigate whether genetic variability in CDH23 may underlie an increased susceptibility to the development of NIHL in the Polish population. Because hearing impairment may be caused not only by exposure to noise but also by other environmental and individual factors, we have assessed whether these factors alone or in combination with the genotype variant may increase the risk of NIHL development. We selected study subjects out of the 2 extremes of the phenotypic spectrum (resistant and susceptible), based on the ISO 1999 model, which corrects for age, sex, years of exposure, and noise exposure level. In our opinion, such selection might significantly increase the power to detect TABLE 5.
Evaluation of the genotype main effect on the disease status in the whole group and subgroups exposed to impulse noise
SNP (no. of genotyped samples) rs3752751 (n rs3752752 (n rs3802720 (n rs10466026 (n rs11592462 (n
the causative variants associated with NIHL over the previous studies where the selection was simply based on the degree of hearing loss. However, this method seems to have also one serious problem that makes interpretation difficult. Although in this study, both groups were relatively young (average about 40), they were already old enough that there should be substantial variation in thresholds even in the absence of noise exposure. In addition, the average noise exposures were rather modest (85 dBA), and the amount of noise-induced threshold shift would therefore be small compared with the amount of age-related threshold shift. Thus, the 2 study subgroups cannot be considered to differ only in susceptibility to noise. Those with the best hearing might be also highly resistant to aging changes, as well as other factors influencing hearing (ototoxic substances, diet) or their combination. Those with the poorest hearing may have the opposite set of traits. On the other hand, subjects susceptible to the development of NIHL might be susceptible also to the aging changes because similar pathogenic processes are involved in the development of both NIHL and presbyacusis. This has been also proven at the gene levelVseveral SNPs have been shown to have overlapping functions in both these complex disorders (5). Moreover, the genetic analyses in mice revealed that homozygosity at Cdh23753A is a primary determinant of
Main effect of genotype ( p value)
OR (95% CI)
Impulse noise ( p value)
0.191 0.037 0.499 0.525 0.184
0.81 (0.591Y1.112) 0.767 (0.598Y0.984) 1.104 (0.828Y1.471) 0.889 (0.618Y1.279) 1.194 (0.918Y1.553)
0.099 0.02 0.557 0.142 0.782
= 341) = 627) = 468) = 374) = 410)
Otology & Neurotology, Vol. 35, No. 2, 2014
Copyright © 2014 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
OR (95% CI) 0.684 0.632 1.139 0.655 1.059
(0.434Y1.079) (0.429Y0.933) (0.735Y1.766) (0.371Y1.156) (0.705Y1.589)
CDH23 AND NOISE-INDUCED HEARING LOSS TABLE 6. SNP rs3752751 rs3752752 rs3802720 rs10466026 rs11592462
363
The effect of genotypes on the disease status in 3 noise exposure subgroups
OR(95% CI)/p value
OR(95% CI)/p value
OR(95% CI)/p value
G85 dB
85Y92 dB
992 dB
0.761 (0.43Y1.346) /0.344 0.782 (0.497Y1.23) /0.285 0.939 (0.57Y1.546) /0.802 0.544 (0.279Y1.057) /0.07 1.124 (0.672Y1.88) /0.654
0.864 (0.547Y1.364) /0.528 0.737 (0.52Y1.042) /0.083 1.16 (0.749Y1.797) /0.505 1.069 (0.621Y1.839) /0.809 1.088 (0.758Y1.561) /0.645
1.504 (0.534Y4.234) /0.428 1.373 (0.583Y3.232) /0.459 1.323 (0.494Y3.54) /0.569 1.292 (0.402Y4.153) /0.659 2.345 (0.912Y6.026) /0.069
both age-related and noise-induced hearing loss (14). Hearing protectors could have less contribution to the variability of hearing thresholds because their use in a particular enterprise was very uniform. The initial analysis of our study groups revealed that resistant individuals were exposed to higher noise levels than susceptible subjects. To overcome a bias related to a so called ‘‘healthy workers syndrome,’’ we have included noise exposure level as a confounding factor in all statistical analysis. Another factor that differentiated susceptible subjects from resistant ones was height, whereby resistant individuals were slightly higher than susceptible ones. Despite certain literature data indicating that low height might be a risk factor in the development of hearing loss (24), our further analysis within the logistic regression framework did not confirm this hypothesis in the group of workers under study. Another feature that differentiated both subgroups (susceptible and resistant) was significantly more frequent prevalence of tinnitus among noise-sensitive individuals. This constitutes the replication of previously described significant correlation between tinnitus and NIHL (25). The genetic component of complex diseases, such as NIHL, has been attributed to a ‘‘large number of smalleffect common variants across the entire allele frequency spectrum or a large number of large-effect rare variants.’’ The genome-wide association studies (GWASs) are neither powered nor designed to detect variation under any of these models on a consistent basis, so there is insufficient empirical data to resolve this debate. Although the discovery of a rare variant near a common variant might
be particularly informative, rare variants are challenging to find since their low frequency renders current cohorts underpowered to detect all but the strongest effects, and lack of correlation to other markers often prevents them from being picked up by standard genotyping marker panels (26). Based on our earlier sequencing analysis of CDH23 done specifically for the Polish population (18), we have therefore focused on most frequent genetic variants, and for the current study, we selected 5 of 35 identified SNPs. Only one of these SNPs (rs3802720) was in common with previous study of Konings et al. (10). The selection in this study was based on the localization of the variant in respect of the promoter region, its frequency, significance for the amino acid change, and varied distribution in previously analyzed subjects. The selected SNPs were localized in exons 5, 21, 45, and 46, two of them leading to amino acid change. In many association studies, the selection of polymorphisms is based mainly on their localization and SNPs in coding sequences are preferred over those found in introns. However, different complex diseases, including NIHL, have varied background in different populations and it is hard to unambiguously determine which SNP selection strategy is the best. Also, the extent to which SNPs leading to amino acid change contribute to disease predisposition represents one of unanswered questions in human genetics. TABLE 7.
Interaction between rs3752752 polymorphism and environmental/medical factors
Interaction
p
rs3752752*smoking
rs3752752*height
rs3752752*complexion
rs3752752*blood pressure
rs3752752*impulse noise FIG. 1. Analysis of genotype frequencies of rs3752752 polymorphism across 3 age subgroups.
SNP SNP*noise smoking SNP*smoking SNP SNP*noise height SNP*height SNP SNP*noise complexion SNP*complexion SNP SNP*noise blood pressure SNP*blood pressure SNP SNP*noise impulse noise SNP*impulse noise
0.014 0.893 0.027 0.513 0.017 0.884 0.051 0.614 0.014 0.922 0.131 0.977 0.016 0.827 0.173 0.016 0.014 0.893 0.791 0.233
Otology & Neurotology, Vol. 35, No. 2, 2014
Copyright © 2014 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
364
T. J. KOWALSKI ET AL.
Statistically significant results within the main effect of genotype were obtained for one SNP localized in exon 21Yrs3752752. The significance level was 0.037 in the entire group and 0.02 in impulse noise-exposed subgroup. The statistical method used in this study strictly followed the protocol of all our previous association studies conducted in the collaboration with the Department of Medical Genetics at the University of Antwerp (6Y10). Such approach allowed to make data comparable. Indeed, a lack of positive association signal for rs3802720 variant in CDH23 gene in this study constitutes a replication of previous negative finding (10). A Bonferroni correction was not used because its application in gene polymorphism association studies is still a matter of a critical debate among genetic epidemiologists, as it can overcorrect true positive associations. Nevertheless, we cannot rule out that the positive finding regarding one CDH23 SNP of 5 tested may be spurious and due to chance. To avoid false-positive results, it has been agreed that the replication of an association in multiple independent sample sets should be considered more important than obtaining very small p values (27). In the study of Yang et al. (17), published in Chinese, the authors suggested other SNPs (rs1227049 and rs3802711) in CDH23 that play an important role in the development of NIHL in Chinese population. The difference in indicated genetic variants may be caused by ethnic differences between analyzed populations such as allele frequencies or LD patterns of associated regions or by different strategies of study groups’ selection for the purpose of genetic analysis. The observed differences in associated genetic variants in the 2 populations do not, in our opinion, disprove CDH23 as gene for NIHL. Moreover, it has been suggested that different SNPs but within the same gene may be regarded as replication in independent populations (27). It should be underlined that the rs3752752 variant was in particular significant in the subgroup of subjects exposed to impulse noise ( p = 0.02). Taking into account the expression pattern and the role of CDH23 in the inner ear, this finding is particularly important as impulse noise was found to lead to mechanical damage of hair cell stereocilia. As it has been earlier suggested, tip link abnormalities related to Cdh23ahl variant in mice may cause an increased probability in the opening of the mechanotransduction channels, which leads to the leakage of potassium cations and a frequent state of depolarization. This depolarization in turn opens calcium channels in the basolateral membrane causing excessive entry of Ca2+ that might lead to hair cell death (28). It has been also hypothesized that mutations in CDH23 may alter the structure of the encoded protein in such a way that the stereocilia bundles do not attain maximum strength and are easily damaged by even lower noise levels, thereby rendering the subject more susceptible to hearing loss when compared with the rest of population (29). Further analysis across three age subgroups (G35, 35Y50 and 950 yr old) of subjects susceptible and resistant
to noise revealed that rs3752752 SNP displayed its effect in young (G35 yr old) individuals and that the CC genotype was more frequent among susceptible subjects, whereas genotype CT appeared more often among resistant-tonoise subjects. Identification of genetic variants displaying protective effects on the auditory organ is particularly important in young individuals, which are shortly exposed to noise as they were described as mostly vulnerable to the development of noise-induced hearing loss (30). The subsequent analysis between medium auditory hearing threshold levels (HTLs) and rs3752752 CC and CT genotypes confirmed that HTLs indeed were significantly worse in carriers of CC genotype than in heterozygous subjects in both (left and right ear) at 4 and 6 kHz only, that is, the 2 frequencies that are most easily affected by NIHL (data not shown). NIHL is a complex disease; thus, the research regarding combined effects of genetic, environmental and medical/individual factors is of great importance. Most of already performed association studies focused only on identification of causal genes; however, this seems to be insufficient. Recently Carlsson et al. (31) have shown that the effect of smoking on susceptibility to NIHL is dependent on the presence of the GSTM1 deletion, suggesting a substantial interaction of genes and environmental factors in NIHL development. In this study, we show that smoking is independent factor for developing hearing loss, which is in agreement with several previous publications (32Y34). Our data show also that the effect of rs3752752 polymorphism is not modified by none of environmental/individual factors except for blood pressure. However, because ‘‘elevated blood pressure’’ was stated on one single measurement and not on the medical diagnosis of hypertension; this issue should be further investigated. To conclude, this study shows that genetic variation in CDH23 may be an independent major factor playing a role in determining individual susceptibility to NIHL, although a substantial variation in thresholds because of aging should be also taken into account. The effect seems to be particularly important in young individuals and in those exposed to impulse noise and can be related to the loss of mechanical stiffness of stereocilia. Taking into account shortcoming of the statistics, further analyses on CDH23 as NIHL susceptibility gene in carefully selected independent populations are required to confirm these findings. REFERENCES 1. Alberti P. Noise-induced hearing lossVa global problem. In: Prasher D, Luxon L, eds. Advances in noise research, Vol. 1, Protection against noise. London, UK: I. Whurr Publisher Ltd, 1998:7Y15. 2. Rabinowitz PM, Pierce Wise J, Hur Mobo B, et al. Antioxidant status and hearing function in noise-exposed workers. Hear Res 2002;173:164Y71. 3. Fortunato G, Marciano E, Zarrilli F, et al. Paraoxonase and superoxide dismutase gene polymorphisms and noise-induced hearing loss. Clin Chem 2004;50:2012Y8. 4. Carlsson PI, Borg E, Grip L, Dahl N, Bondeson ML. Variability in noise susceptibility in a Swedish population: the role of 35delG
Otology & Neurotology, Vol. 35, No. 2, 2014
Copyright © 2014 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
CDH23 AND NOISE-INDUCED HEARING LOSS
5. 6. 7. 8.
9. 10. 11. 12. 13. 14. 15. 16. 17. 18.
19.
mutation in the Connexin26 (GJB2) gene. Audiol Med 2004; 2:123Y30. Van Eyken E, Van Laer L, Fransen E, et al. The contribution of GJB2 (Connexin 26) 35delG to age related hearing impairment and noise-induced hearing loss. Otol Neurotol 2007;28:970Y5. Konings A, Van Laer L, Pawelczyk M, et al. Association between variations in CAT and noise-induced hearing loss in two independent noise-exposed populations. Hum Mol Genet 2007;16:1872Y83. Van Laer L, Carlsson PI, Ottschytsch N, et al. The contribution of genes involved in potassium recycling in the inner ear to noiseinduced hearing loss. Hum Mutat 2006;27:786Y95. Pawelczyk M, van Laer L, Fransen E, et al. Analysis of gene polymorphisms associated with K+ ion circulation in the inner ear of patients susceptible and resistant to noise-induced hearing loss. Ann Hum Genet 2009;73:411Y21. Konings A, Van Laer L, Michel S, et al. Variations in HSP70 genes associated with noise-induced hearing loss in two independent populations. Eur J Hum Genet 2008;17:329Y35. Konings A, van Laer L, Wiktorek-Smagur A, et al. Candidate gene association study for noise-induced hearing loss in two independent noise-exposed populations. Ann Hum Genet 2009;73:215Y24. Yang M, Tan H, Yang Q, et al. Association of hsp70 polymorphisms with risk of noise-induced hearing loss in Chinese automobile workers. Cell Stress Chaperones 2006;11:233Y9. Kazmierczak P, Sakaguchi H, Tokita J, et al. Cadherin 23 and protocadherin 15 interact to form tip-link filaments in sensory hair cells. Nature 2007;449:87Y91. Holme RH, Steel KP. Progressive hearing loss and increased susceptibility to noise-induced hearing loss in mice carrying a Cdh23 but not a Myo7a mutation. JARO 2004;5:66Y79. Noben-Trauth K, Zheng QY, Johnson KR. Association of cadherin 23 with polygenic inheritance and genetic modification of sensorineural hearing loss. Nat Genet 2003;35:21Y3. Erway LC, Shiau YW, Davis RR, Krieg EF. Genetics of age-related hearing loss in mice. III. Susceptibility of inbred and F1 hybrid strains to noise-induced hearing loss. Hear Res 1996;93:181Y7. Davis RR, Newlander JK, Ling X, Cortopassi GA, Krieg EF, Erway LC. Genetic basis for susceptibility to noise-induced hearing loss in mice. Hear Res 2001;155:82Y90. Yang M, Tan H, Zheng JR, et al. Association of cadherin CDH23 gene polymorphisms with noise induced hearing loss in Chinese workers. Wei Sheng Yan Jiu 2006;35:19Y22. Sliwinska-Kowalska M, Noben-Trauth K, Pawelczyk M, Kowalski TJ. Single Nucleotide Polymorphisms in the cadherin 23 (CDH23) gene in Polish workers exposed to industrial noise. Am J Hum Biol 2008;20:481Y3. Sliwinska-Kowalska M, Dudarewicz A, Kotylo P, ZamyslowskaSzmytke E, Pawlaczyk-Luszczynska M, Gajda-Szadkowska A. Individual susceptibility to noise-induced hearing loss: choosing an
365
optimal method of retrospective classification of workers into noisesusceptible and noise-resistant groups. IJOMEH 2006;19:235Y45. 20. Curtis D, North BV, Gurling HMD, Blaveri E, Sham PC. A quick and simple method for detecting subjects with abnormal genetic background in case-control samples. Am Hum Genet 2002;66: 235Y44. 21. Abecasis GR, Cherny SS, Cookson WOC, Cardon LR. GRR: graphical representation of relationship errors. Bioinformatics 2001; 17:742Y3. 22. Cox DG, Canzian F. Genotype transposer: automated genotype manipulation for linkage disequilibrium analysis. Bioinformatics 2001;17:738Y9. 23. Devlin B, Roeder K, Wasserman L. Genomic control, a new approach to genetic-based association studies. Theor Popul Biol 2001;60:155Y66. 24. Burr H, Lund SP, Sperling BB, Kristensen TS, Poulsen OM. Smoking and height as risk factors for prevalence and 5-year incidence of hearing loss. A questionnaire-based follow-up study of employees in Denmark aged 18Y59 years exposed and unexposed to noise. Int J Audiol 2005;44:531Y9. 25. Rosenhall U, Karlsson AK. Tinnitus in old age. Scand Audiol 1991;20:165Y71. 26. Raychaudhuri S. Mapping rare and common causal alleles for complex human diseases. Cell 2011;147:57Y69. 27. Neale BM, Sham PC. The future association studies: gene-based analysis and replication. Am J Hum Genet 2004;75:353Y36228. 28. Johnson KR, Yu H, Ding D, Jiang H, Gagnon LH, Salvi RJ. Separate and combined effects of SOD1 and Cdh23 mutations on agerelated hearing loss and cochlear pathology in C57BL/6J mice. Hear Res 2010;268:85Y92. 29. Wang Y, Hirose K, Liberman MC. Dynamics of noise-induced cellular injury and repair in the mouse cochlea. J Assoc Res Otolaryngol 2002;3:248Y68. 30 Corso JF. Age correction factor in noise-induced hearing loss: a quantitative model. Audiology 1980;19:221Y32. 31. Carlsson PI, Fransen E, Stenberg E, Bondeson ML. The influence of genetic factors, smoking and cardiovascular diseases on human noise susceptibility. Audiol Med 2007;5:82Y91. 32. Sliwinska-Kowalska M, Zamyslowska-Szmytke E, Szymczak W, et al. Effects of coexposure to noise and mixture of organic solvents on hearing in dockyard workers. J Occup Environ Med 2004; 46:30Y8. 33. Nomura K, Nakao M, Morimoto T. Effect of smoking on hearing loss: quality assessment and meta-analysis. Prev Med 2005;40: 138Y44. 34. Uchida Y, Nakashimat T, Ando F, Niino N, Shimokata H. Is there a relevant effect of noise and smoking on hearing? Int J Audiol 2005;44:86Y91.
Otology & Neurotology, Vol. 35, No. 2, 2014
Copyright © 2014 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
