Introduction
Presbyacusis, also known as age-related hearing loss, gradually occurs with age. The 2012/2013 Canadian Health Measures Survey reported on the marked age-dependent difference in the prevalence of audiometric confirmed hearing loss; this increased from 10 per cent of the population aged 50 years and younger, to 65 per cent at ages 70–79 years.Reference Feder, Michaud, Ramage-Morin, McNamee and Beauregard1 The global population of people aged 60 years and over was 962 million in 2017, equating to 13 per cent of the total population, and this number is projected to be 1.4 billion in 2030 and 2.1 billion in 2050.2
Despite the increasing proportion of the population affected by age-related hearing loss, the underlying pathogenetic mechanism remains poorly understood. Based on post-mortem pathological analysis, Schuknecht and Gacek classified age-related hearing loss into six types: sensory (loss of greater than 10 mm of cochlear sensory hair cells); neuronal (loss of greater than 50 per cent of spiral ganglion neurons); strial (atrophy of more than 30 per cent of the stria vascularis); cochlear conductive (where the pathological changes do not meet any of the previous criteria, but there is a characteristic audiometric hearing loss pattern); mixed (a combination of the first three types); and unclassified.Reference Schuknecht and Gacek3 Age-related hearing loss is believed to be a consequence of a combination of genetic, cumulative environmental exposures, and age-related pathophysiological change.Reference Gates and Mills4 Hypertension, diabetes, noise exposure and ototoxicity are amongst the proposed risk factors for its development.Reference Huang and Tang5 No cure has been identified so far.
Sudden sensorineural hearing loss (SNHL) is another acquired SNHL syndrome, of unknown aetiology in 85–90 per cent of the cases. It is defined as a hearing loss of greater than or equal to 30 dB, spanning at least three contiguous audiometric frequencies, which develops over 72 hours.Reference Stachler, Chandrasekhar, Archer, Rosenfeld, Schwartz and Barrs6 The incidence rate reported in large twenty-first century national studies varies between 27 and 60 per 100 000 population in the USA and Japan respectively.Reference Alexander and Harris7,Reference Nakashima, Sato, Gyo, Hato, Yoshida and Shimono8 The US study identified a slight male predominance and the Japanese study reported a 3:1 female predominance. Alexander and Harris’ study findings supported earlier reports that the incidence rate increased with age.Reference Alexander and Harris7 In this respect, sudden SNHL mirrors the age-dependant prevalence pattern of age-related hearing loss.
The recovery rate varies, and is proposed to depend on several factors, including: the initial degree of hearing loss, the time to treatment, age, the presence of vertigo and the pattern of hearing loss.Reference Suzuki, Tabata, Koizumi, Hohchi, Takeuchi and Kitamura9 The overall spontaneous clinically important hearing recovery rate can be as high as 73 per cent.Reference Mattox and Simmons10 The recovery rate with current treatment protocols is similar, reported as 78 per cent in a recent study (Lee et al., unpublished data).
Histopathological changes associated with sudden SNHL are similar to those of age-related hearing loss, including the loss of sensory hair cells and cochlear neurons, and atrophy of the stria vascularis.Reference Merchant, Adams and Nadol11 The pathogenesis of sudden SNHL, like age-related hearing loss, is unknown.
MicroRNAs are small (approximately 21–23 nucleotides long), non-coding, single-stranded RNAs. They are found in animals, plants and some viruses that repress messenger RNA transcription by binding to complementary sequences in the 3′ -untranslated region (3′-UTR) of messenger RNA. MicroRNAs are highly evolutionarily conserved, and have been found to participate in the regulation of gene expression of nearly all cellular processes (e.g. cell proliferation, differentiation, migration and apoptosis). Currently, more than 2500 microRNAs have been discovered in humans. Altered microRNA expression has been identified in ageing, and pathological states such as cancer, and cardiovascular and neurological diseases.
Recent studies have determined that microRNAs play a fundamental role in inner-ear development.Reference Rudnicki and Avraham12 There is increasing evidence that microRNA regulation of gene expression in the post-developmental inner ear contributes to the development of acquired hearing loss.Reference Pang, Xiong, Yang, Ou, Xu and Huang13
Identifying changes in microRNA expression in acquired SNHL may offer new and exciting insights into disease pathogenesis. Therefore, a review of studies investigating the role of microRNAs in age-related hearing loss and sudden SNHL was undertaken.
Materials and methods
A narrative review of English-language articles published between January 2000 and December 2017 was undertaken. Medline, PubMed, Web of Science and Embase databases were searched, using the Medical Subject Headings (MeSH) term ‘microRNA’, combined with the MeSH terms ‘hearing loss, sensorineural’, ‘hearing loss, sudden’, ‘presbycusis’, ‘organ of Corti’, ‘labyrinth supporting cells’, ‘hair cell, auditory’, ‘ear, inner’, ‘deafness’ and the non-MeSH term ‘progressive hearing loss’.
Inclusion criteria were primary studies investigating the role of microRNAs in the pathogenesis of age-related hearing loss or sudden SNHL, in animal and/or human subjects aged over 18 years. Studies were excluded if they: did not use the American Academy of Otolaryngology – Head and Neck Surgery sudden SNHL diagnostic criteria;Reference Stachler, Chandrasekhar, Archer, Rosenfeld, Schwartz and Barrs6 did not record the authors; or included mostly patients with conditions such as diabetes mellitus or acoustic trauma, which are known hearing loss aetiological factors. Similarly, studies reporting on non-mammals or microRNA biogenesis components only were excluded.
The results of the search were tabulated in the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (‘PRISMA’) guidelines recommended format.Reference Liberati, Altman, Tetzlaff, Mulrow, Gotzsche and Ioannidis14 The quality of evidence for individual studies was graded independently by two authors (HHRC and PW) as poor, fair or good, using the National Heart, Lung, and Blood Institute of the National Institutes of Health Study Quality Assessment Tools (Table 1).15 The grade recorded was that agreed by both evaluators. When there was a discrepancy between the two evaluators that was not resolved through discussion into a consensual grade, then the senior author (DAN) assessed the study and allocated the study grade.
NIH = National Institutes of Health
Results and discussion
A total of 207 records were identified with the MeSH terms search after removing duplicates. Eighty-six records were excluded for the following reasons: review articles (n = 51), conference abstracts (n = 22), letters (n = 3), editorial (n = 1), survey (n = 1), report (n = 1), non-English language (n = 6) and anonymous authorship (n = 1). On further screening, of the remaining 121 primary article abstracts, 112 failed to meet the study inclusion criteria. Figure 1 shows the Preferred Reporting Items for Systematic Reviews and Meta-Analyses flow chart.
Evidence of microRNA modulation was reported in eight age-related hearing loss studies and in one sudden SNHL study; these formed the basis of this review. Two were animal studies,Reference Zhang, Liu, McGee, Walsh, Soukup and He16,Reference Zhang, Liu, Soukup and He17 four included findings from animal cell culture studies,Reference Huang, Zheng, Ou, Xiong, Yang and Zhang18–Reference Xiong, Pang, Yang, Dai, Liu and Ou21 one comprised findings from animal and human subjects,Reference Pang, Xiong, Yang, Ou, Xu and Huang13 and two concerned findings from human subjects alone.Reference Li, Peng, Huang, Li, Wang and Wang22,Reference Sekine, Matsumura, Takizawa, Kimura, Saito and Shiiba23
The quality of evidence of the studies, based on the National Institutes of Health quality assessment criteria, was judged as good in six, fair in one and poor in two studies (Table 2).
NIH = National Institutes of Health; HL = hearing loss; ABR = auditory brainstem response; mRNA = messenger RNA; HEI-OC1 = House Ear Institute-organ of Corti 1; siRNA = small interfering RNA; miRNA = microRNA; SNHL = sensorineural hearing loss
MicroRNAs in programmed cell death
Programmed cell death, which includes apoptosis, autophagy, and programmed necrosis or necroptosis, is defined as regulated cell death executed by an intracellular programme.Reference Su, Yang, Xu, Chen and Yu24 Accumulated evidence has shown a close interaction between apoptosis, autophagy and necroptosis, and the molecular mechanism underlying the crosstalk is regulated by microRNAs. One microRNA may target more than one component of a programmed death pathway and even more than one death pathway.
Apoptosis is a process by which the body rids itself of up to 70 billion superfluous cells a day without triggering an inflammatory process. It occurs normally during development and ageing, and as a homeostatic mechanism to maintain cell populations in tissues. Inappropriate apoptosis is a factor in many human conditions, including neurodegenerative diseases, ischaemic damage, autoimmune disorders and many types of cancer. Many microRNAs have been reported as regulating the apoptotic signalling pathways, such as executioner caspases, and key regulators transcription factor p53 and death-associated protein kinase (DAPK).Reference Su, Yang, Xu, Chen and Yu24
Pro-apoptotic microRNA upregulation in presbyacusis
Pro-apoptotic microRNAs, such as miR-29a/b/c, miR-34a/b/c, let-7a/b/c/e/f/g/i, miR-141, miR-146, miR-203 and miR-429, work through different direct and indirect pathways to induce apoptosis.Reference Zhang, Liu, McGee, Walsh, Soukup and He16 Upregulation of both pro-apoptotic miR-29a/b/c and miR-34a/b/c families in the organ of Corti likely contribute to hair cell apoptosis and tissue degeneration in age-related hearing loss. MiR-29 regulates genes upstream of p53 pathways and miR-34a regulates genes downstream of p53 pathways, suggesting that microRNAs play a role in p53-mediated apoptosis in the organ of Corti.Reference Tadros, D'Souza, Zhu and Frisina25 The protein p53 is a key transcription factor which regulates many genes that initiate anti-proliferative responses, such as cell-cycle arrest, DNA repair, apoptosis and cellular senescence. At an organism level, p53 activity has been implicated in tissue degeneration and ageing.
In age-related hearing loss animal models, signalling pathways miR-34a/B-cell lymphoma2 (Bcl-2)Reference Huang, Zheng, Ou, Xiong, Yang and Zhang18 and miR-34a/sirtuin 1 (SIRT1)/p53Reference Xiong, Pang, Yang, Dai, Liu and Ou21 were demonstrated to be associated with cochlear hair cell apoptosis. Genes Bcl-2 and SIRT1 are direct targets of miR-34a.Reference Huang, Zheng, Ou, Xiong, Yang and Zhang18,Reference Xiong, Pang, Yang, Dai, Liu and Ou21 The Bcl-2 family contains both anti-apoptotic and pro-apoptotic members, and their interactions determine the likelihood of the cell undergoing apoptosis.Reference Gross, McDonnell and Korsmeyer26 Anti-apoptotic Bcl-2 family proteins (Bcl-2, Bcl-2l1, Bcl-2l2, Mcl-1 and Bcl-2a1) are thought to exert their effects by stabilising the mitochondrial membrane potential and preventing the release of cytochrome C and apoptosis-inducing factors. The binding of miR-34a to the 3’-untranslated region of the Bcl-2 gene halts its transcription, thus cancelling the anti-apoptotic effect of Bcl-2. In aged mice, higher expression of Bcl-2 and Bcl-211 genes was demonstrated to exert a protective effect during cochlear ageing.Reference Tadros, D'Souza, Zhu and Frisina25
Similarly, miR34a overexpression suppressed SIRT1 gene messenger RNA and induced cochlear hair cell apoptosis via a p53-mediated pathway.Reference Xiong, Pang, Yang, Dai, Liu and Ou21 The SIRT1 protein is a nicotinamide adenine dinucleotide dependent deacetylase that regulates apoptosis in response to oxidative and genotoxic stress by deacetylating its substrates, including p53.Reference Yamakuchi and Lowenstein27 Thus, miR-34a blockage of SIRT1 transcription induces cochlear hair cell apoptosis via a p53-mediated pathway.
The E2F3 gene (E2 transcription factor 3), which regulates cell-cycle progression, might be another direct target of miR-34a.Reference Welch, Chen and Stallings28 Pang and colleagues’ (2016) study demonstrated a simultaneous increase of miR-34a expression in the auditory pathways and plasma in aged mice.Reference Pang, Xiong, Yang, Ou, Xu and Huang13 The messenger RNA levels of genes SIRT1, Bcl-2 and E2F3 were correspondingly decreased within the auditory pathway and plasma with ageing. However, simultaneous increases of miR-34a expression in non-hearing organs such as the heart and liver suggest that miR-34a exerts a non-tissue-specific ageing effect in mice.
Human studies on the relationship between circulating miR-34a expression level and auditory function are inconclusive. MiR-34 expression level was significantly upregulated in age-related hearing loss patients, compared to normal hearing age-matched and younger controls.Reference Pang, Xiong, Yang, Ou, Xu and Huang13 However, the messenger RNA levels of miR-34a target genes SIRT1, Bcl-2 and E2F3 were similar in age-related hearing loss patients and controls, which questions the validity of a miR-34a/Bcl-2 or miR-34a/SIRT1/p53 mediated pathway in the pathogenesis of human age-related hearing loss. In addition, the similar circulating miR-34a expression levels in healthy human controls of different age groups indicates that ageing in humans is not associated with a non-tissue-specific increase in miR-34a level as demonstrated in animal models.Reference Pang, Xiong, Yang, Ou, Xu and Huang13
MiR-29b overexpression was shown to induce mitochondrial dysfunction and hair cell loss via downregulation of the SIRT1/PGC-1α (peroxisome proliferator-activated receptor-gamma coactivator 1α) in aged mice.Reference Xue, Wei, Zha, Qiu, Chen and Qiao20 SIRT1 is a direct target of miR-29b. Oxidative stress is believed to cause age-related hearing loss by apoptosis of auditory system cells.Reference Someya, Yu, Hallows, Xu, Vann and Leeuwenburgh29 SIRT1 regulates intracellular oxidative stress by the deacetylation of its substrates including PGC-1α, a transcriptional co-regulator that binds to numerous transcription factors to promote mitochondrial biogenesis and oxidative metabolism.Reference Tan, Tang, Zhang, Cheng, Cai and Chen30 Previous studies, as cited by Xue et al.,Reference Xue, Wei, Zha, Qiu, Chen and Qiao20 have shown that the upregulation of miR-29 microRNAs promotes p53-dependent apoptosis by suppressing several key regulators of cell survival, such as genes p85α and CDC42 (cell division cycle 42). There is evidence from animal studies that miR-29 contributes to age-related hearing loss, but supporting human studies are lacking.
Anti-apoptotic microRNA downregulation in presbyacusis
With ageing, more microRNA expression levels are downregulated than upregulated.Reference Zhang, Liu, McGee, Walsh, Soukup and He16,Reference Sekine, Matsumura, Takizawa, Kimura, Saito and Shiiba23 Amongst them are miR-181 and miR-183 family members, which are known mediators of cell proliferation and differentiation pathways. The miR-181 family consists of four members: miR-181a, miR-181b, miR-181c and miR-181d. All of these, with the exception of miR-181c, are associated with organ of Corti degeneration in aged mice.Reference Zhang, Liu, McGee, Walsh, Soukup and He16 Upregulation of miR-181a correlated with inner-ear sensory epithelium regeneration in chickens, by suppressing protein p27, a cell-cycle inhibitor and differentiator in the auditory sensory epithelium.Reference Frucht, Santos-Sacchi and Navaratnam31
The miR-183 family is a triad consisting of miR-96, miR-182 and miR-183; these originate from a common primary transcript that begins to be expressed as early as embryonic day 9.5 within the otic vesicle.Reference Sacheli, Nguyen, Borgs, Vandenbosch, Bodson and Lefebvre32 Its unique expression pattern within the inner ear correlates with inner-ear development and differentiation.Reference Sacheli, Nguyen, Borgs, Vandenbosch, Bodson and Lefebvre32
MiR-182 is expressed during the differentiation of inner-ear stem/progenitor cells into a hair-cell-like fate. Its function may be associated with its putative target protein Tbx1 (T-box transcription factor 1), a transcription factor that has been implicated in inner-ear development and hair cell fate.Reference Wang, Zhang, Du and Jiang33 MiR-182 overexpression induced ectopic hair cells in developing zebra fish, while knockdown of the miR-183 family led to reduced numbers of sensory hair cells and defects in the semi-circular canals.Reference Li, Kloosterman and Fekete34
MiR-96 overexpression demonstrated the same effect as miR-182 in developing zebra fish.Reference Li, Kloosterman and Fekete34 Mutation in the miR-96 seed region (2–7 nucleotides in length), which confers binding specificity,Reference Lewis, Burge and Bartel35 results in autosomal dominant, progressive hearing loss in humansReference Mencia, Modamio-Hoybjor, Redshaw, Morin, Mayo-Merino and Olavarrieta36 and in mice.Reference Lewis, Quint, Glazier, Fuchs, De Angelis and Langford37 In addition, a mutation located in the non-seed region of the miR-96 identified in non-syndromic hearing loss in humans was shown to interfere with miR-96 pre-microRNA secondary structure.Reference Solda, Robusto, Primignani, Castorina, Benzoni and Cesarani38 MiR-96 mutation in mice alters the function and gene expression profile of the organ of Corti.Reference Lewis, Quint, Glazier, Fuchs, De Angelis and Langford37 Its downstream effect on Oncomodulin, Pitpnm1, Slc26a5 (prestin), Ptprq and Gfi1 genes results in impaired hair cell function and hair cell degeneration, and causes deafness.Reference Solda, Robusto, Primignani, Castorina, Benzoni and Cesarani38 Interestingly, Zhang and colleagues’ (2013) microarray analysis detected miR-96 downregulation only in C57BL/6J mice, and not in CBA/J mice, suggesting that the different genetic backgrounds of these two mouse strains may result in differences in the microRNAs involved in the pathogenesis of age-related hearing loss.Reference Zhang, Liu, McGee, Walsh, Soukup and He16
Anti-autophagy microRNA upregulation in presbyacusis
In a recent study, miR-34a was shown to induce cochlear hair cell death via the suppression of autophagy.Reference Pang, Xiong, Lin, Lai, Yang and Liu19 Autophagy is a regulated intracellular programme that seeks to clear accumulated intracellular toxins; if unsuccessful, it results in cell death. It is triggered by growth signal deficiency, nutrient deprivation, genotoxic stress, hypoxic stress, endoplasmic reticulum stress, and/or reactive oxygen species accumulation.
In Pang and colleagues’ (2017) study, miR-34a overexpression was associated with a reduction in autophagy in aged mice and HEI-OC1 (House Ear Institute-organ of Corti 1) cells.Reference Pang, Xiong, Lin, Lai, Yang and Liu19 In HEI-OC1 cells, miR-34a overexpression suppressed protein ATG9A (autophagy-related protein 9A) level and impaired autophagy via p62 elevation.Reference Tai, Wang, Gong, Han, Zhou and Wang39 ATG9A is one target gene of miR-34a, and its corresponding protein AT9A is necessary for optimal autophagy.Reference Huang, Sun, Huang, Ye, Pan and Zhong40
Further research is required to demonstrate the role of miR-34a mediated autophagy in human age-related hearing loss.
MicroRNAs in stria vascularis
A number of differentially expressed microRNAs have been reported in the ageing lateral wall of the cochlear duct or stria vascularis.Reference Zhang, Liu, Soukup and He17 MiR-203, a pro-apoptotic microRNA, was upregulated in the lateral wall of aged mice. Genotoxic stress agents (e.g. camptothecin) were demonstrated to induce miR-203 expression by p53 acetylation.Reference Chang, Davis-Dusenbery, Kashima, Jiang, Marathe and Sessa41 MiR-203 overexpression downregulates anti-apoptotic genes Bclw or Bcl-2l2 (a member of the Bcl-2 family), and promotes cell death in a p53-dependent manner. Other upregulated microRNAs (miR-762 and miR-1224) and downregulated microRNAs (miR-107, miR-145, miR-342 and miR-455) are also suggested to play a crucial role in the function and maintenance of the stria vascularis.
Potential microRNA–messenger RNA networks in age-related hearing loss
Zhang et al. (2014) proposed networks of microRNA–messenger RNA interactions in age-related hearing loss based on studies of ageing mice.Reference Zhang, Liu, Soukup and He17 Differentially expressed microRNAs and their predicted target messenger RNAs associated with apoptosis were used. The networks were complex, with examples of downregulated and upregulated microRNAs targeting upregulated and downregulated pro- and anti-apoptotic genes. For example, Tnfsf10, a critical pro-apoptotic gene was predicted to be upregulated by the downregulation of anti-apoptotic miR-181, amongst others.
Lewis et al. (2016) also explored the potential networks controlled by miR-96 in miR-96 mutant mice (diminuendo mice).Reference Lewis, Buniello, Hilton, Zhu, Zhang and Evans42 Three methods of network analysis were used: a manually created regulatory network, protein–protein interactome analysis and gene set enrichment analysis. Myc, Gfi1 and Fos genes were suggested as important targets of miR-96 mediated regulatory networks involved in hearing and deafness. Fos was consistently identified by more than one network analytical method. It regulates genes Slc26a5 (prestin) and Ocm (oncomodulin), both of which are associated with hearing and deafness.Reference Tong, Hornak, Maison, Ohlemiller, Liberman and Simmons43 Gfi1 controls Fos via Myc.Reference Lewis, Buniello, Hilton, Zhu, Zhang and Evans42 However, more work is required in this area.
MicroRNAs in sudden sensorineural hearing loss pathogenesis
Li et al. predicted that the microRNAs hsa-miR-34a/548n/15a/143/23a/210/18b regulated target genes which may have a critical role in sudden SNHL, based on a small sample of nine sudden SNHL patients and three controls.Reference Li, Peng, Huang, Li, Wang and Wang22 Experimental studies not specifically focused on sudden SNHL corroborate a potential role for miR-34a,Reference Pang, Xiong, Yang, Ou, Xu and Huang13 miR-210,Reference Tra, Frisina and D'Souza44 miR18B,Reference Niceta, Stellacci, Gripp, Zampino, Kousi and Anselmi45 miR-23AReference Ohlemiller, Rosen and Gagnon46 and miR15a-5pReference Hedlund, Tangvoranuntakul, Takematsu, Long, Housley and Kozutsumi47 in sudden SNHL. Further studies of larger patient samples are required to validate a role for microRNAs in human sudden SNHL.
Two human-subject studies investigated microRNA biogenesis associated components in sudden SNHL, though neither sought direct evidence of altered microRNA levels.Reference Han, Kim, Shin, Cho and Nam48,Reference Kim, Lee and Nam49 The microRNA biogenesis pathway proteins DiGeorge syndrome critical region gene 8 (DGCR8) and argonaute 2 (AGO2) directly influence the biosynthesis of all microRNAs.Reference Han, Kim, Shin, Cho and Nam48 AGO2 is a component of the RNA-induced silencing complex (RISC) that silences microRNA activity.Reference Kobayashi and Tomari50 DGCR8 is a unit of the microprocessor complex, which mediates microRNA biogenesis via the generation of a precursor microRNA (pre-microRNA).Reference Lee, Ahn, Han, Choi, Kim and Yim51
Han and colleagues’ study identified elevated AGO2 gene expression in sudden SNHL patients’ peripheral blood samples, and it was positively correlated with DGCR8 gene expression in both sudden SNHL patients and healthy controls.Reference Han, Kim, Shin, Cho and Nam48
Kim and colleagues’ study on circulating messenger RNA levels of RNAse III endonucleases Dicer and Drosha identified dysregulation of Dicer, another RISC component that was reduced in sudden SNHL patients compared to controls.Reference Kim, Lee and Nam49 Drosha expression was not altered, suggesting that the processing of microRNAs at the nuclear level was not affected. Dicer is essential for the processing of mature microRNAs from their pre-microRNA form.Reference Bernstein, Kim, Carmell, Murchison, Alcorn and Li52 Normal development of the functioning inner ear is strongly dependent on normal microRNA maturation, and interrupting the microRNA maturation process results in profound inner-ear malformation.Reference Friedman and Avraham53
The expression of individual microRNA biogenesis related components has been found to be changed in various human diseases.Reference Sand, Skrygan, Georgas, Arenz, Gambichler and Sand54 Therefore, the impact of alterations in AGO2 and Dicer expressions on the differential expression of microRNAs in sudden SNHL patients is worthy of further study.
Conclusion
MicroRNA regulation of gene expression plays a role in the development, differentiation, proliferation, autophagy and apoptosis of cells. In age-related hearing loss, microRNAs in apoptotic pathways, predominantly miR-34 and miR-29 families, are significantly upregulated, while regenerative and developmental pathway microRNAs, namely miR-181 and miR-183 families, are downregulated. MicroRNA autophagy-mediated effects may also contribute to age-related hearing loss. The observed changes in microRNA expression levels are consistent with the sensory hair cell loss and elevated hearing thresholds associated with age-related hearing loss. There is limited evidence that microRNAs have an aetiological role in sudden SNHL.
Competing interests
None declared