Molecular Genetics of Nonsyndromic Auditory Neuropathy

In recent years, the identification of genes underlying mendelian hearing impairment has greatly accelerated our understanding of the molecular processes of hearing. In general, acquired causes and genetic causes account for an equal proportion of childhood hearing loss, with 40% recessive, 10% dominant, and 3% X-linked or mitochondrial hearing loss comprising the genetic portion (Fraser, 1964; Marazita et al., 1993). There are 400 described genetic syndromes associated with hearing loss (Gorlin et al., 1967), such that families with similar clinical features accompanying hearing loss can be grouped for study. However, the vast majority of sensorineural hearing loss (SNHL) is nonsyndromic, or unassociated with other clinical features (Bergstrom et al., 1971). Recessive nonsyndromic hearing impairment (NSHI) (designated by DFNB loci) is almost always congenital, profound, and bilateral, while dominant NSHI (designated by DFNA loci) is usually delayed onset and progressive.

The inability to use phenotypic features to differentiate recessive deafness of different genetic etiologies has been overcome by studying isolated and/or consanguineous populations. However, most families from the United States that are powerful enough for linkage analysis have dominant deafness. Dominant loci have been mapped and cloned in single large families or by using mapping data and phenotypic features to combine genetically homogeneous families. Despite the obstacles of genetic heterogeneity, the high prevalence of acquired or environmental deafness, reduced penetrance, and variable expressivity even within families, at least 40 deafness genes have been identified to date (Hereditary Hearing Loss Homepage).

Both positional cloning and positional candidate approaches have been successful in identifying deafness genes belonging to diverse families such as atypical myosins, collagens, and cadherins (Resendes et al., 2001). Depending on the mutation, the phenotype may be dominant or recessive, nonsyndromic or syndromic. The vast majority of dominant traits are actually semidominant, since homozygosity usually results in a more severe phenotype than heterozygosity for a mutated allele (Nussbaum et al., 2001). While true dominance is exceedingly rare among all genetic traits, the notable exception is Huntington's disease, which is caused by abnormal expansion of triplet repeats (Wexler et al., 1987).

Mutations in MYO7A may segregate as dominant (DFNA11) (Liu et al., 1997b) or recessive NSHI (DFNB2) (Liu et al., 1997a; Weil et al., 1997) or as Usher syndrome (USH1B) (Weil et al., 1995). Mutations in Protocadherin 15 (PDCH15) underlie USH1F and DFNB23 (Ahmed et al., 2003), and the mouse homolog Pcdh15 underlies the Ames waltzer mutant (av) (Alagramam et al., 2001). Mutations in CDH23 (Cadherin 23) underlie both USH1D and DFNB12 (Bork et al., 2001), while murine Cdh23 underlies age-related hearing loss (ahl), modifier of deaf waddler (mdfw), and waltzer (w) (Di Palma et al., 2001; Wilson et al., 2001; Noben-Trauth et al., 2003). Recent reports implicate the CHD23 protein in mechanoelectrical transduction occurring in tip links as well as the organization of stereociliary bundles (Siemens et al., 2004; Sollner et al., 2004).

In 1997, mutations in the GJB2 gene were identified and found to be a common cause of nonsyndromic recessive deafness (Kelsell et al., 1997; Estivill et al., 1998). This discovery revolutionized genetic testing for deafness because of three fortuitous circumstances: 1) GJB2 is a small gene; 2) the prevalence of mutations in the deaf population approaches 50%; 3) specific alleles occur frequently in certain ethnic groups such that sequencing of the entire gene is necessary only in a minority of cases. Thus, it is practical in terms of cost and effort to screen GJB2 even without significant a priori evidence for a genetic etiology. Heterozygosity for a GJB2 deafness allele may result in deafness when inherited with a deletion in GJB6 (del Castillo et al., 2002). Computerized tomography (CT) of the temporal bone may identify morphological abnormalities such as Mondini dysplasia and enlarged vestibular aqueduct (EVA) known to be associated with mutations in SLC26A4 (Prasad et al., 2000). Thus, for nonsyndromic recessive loss, genetic advances have been readily translated into clinical evaluation and management.

In contrast, information to guide the clinical algorithm for dominant nonsyndromic hearing loss is limited. GJB2 screening will identify the rare dominant allele (Cx26 Homepage) or pseudominant inheritance of GJB2 deafness. However, for the majority of dominant families lacking GJB2 mutations or inner ear morphological abnormalities, there is no one gene most commonly found in dominant deafness. Screening is not practical without a priori evidence of linkage, especially for large genes with numerous exons. For many loci there is no predominant mutation, and often, there are only 1 or 2 families linked. However, in contrast to recessive deafness, audiologic characteristics can be used in dominant hearing loss to define genetically homogeneous types. Dominant hearing loss may present in a variety of phenotypes, affecting selected frequencies (low, mid or high) or all frequencies. In addition, the rate of progression varies from very slow (with years of stable hearing) to rapid development of profound loss.

We have demonstrated that low frequency sensorineural hearing loss (LFSNHL) segregating as a dominant trait is almost always due to mutations in WFS1 at the DFNA6 locus (see Preliminary Data). The only other LFSNHL locus, DFNA1, was mapped in a family with a distinct form of LFSNHL caused by mutations in DIAPH1 (Lynch et al., 1997), a human homolog of Drosophila diaphanous. The molecular genetic study of auditory neuropathy (AN) presents another opportunity to focus on a specific audiologic phenotype.

The hallmark of AN is preservation of outer hair cell (OHC) function with absent or abnormal auditory brainstem responses (ABRs) (Starr et al., 1996). Some authors prefer the term "auditory dys-synchrony" since synchronous potentials are necessary for a measurable electrophysiological response (Berlin et al., 2003). The cochlear microphonic (CM) response precedes wave I of the ABR and is generated by OHC and inner hair cells (IHC), but it is chiefly an OHC potential when recorded by surface electrodes in humans. The discovery that OHCs generate otoacoustic emissions (OAEs) in response to an acoustic stimulus has greatly facilitated screening for AN, such that the traditional concept of SNHL defined on pure tone audiometry can be further refined. Absent OAEs confirm a "sensory" loss or OHC disorder, as opposed to "neural" disorders affecting IHCs, the auditory nerve, and/or the synapses between the IHC and auditory nerve. While "auditory neuropathy" may seem a misnomer for hearing loss caused by a selective IHC disorder, at this time there is no specific clinical test to assess IHC function.

Despite the extensive body of literature describing clinical findings associated with the myriad of NSHI loci and genes, little is known about OHC function in subjects with NSHI. Most studies mapping and cloning NSHI genes have characterized the phenotype by pure tone audiometry measuring thresholds by air and bone conduction. Studies conducted in foreign populations often must rely on portable battery-powered equipment, especially for regions without electricity. Without a sound attenuated booth, background noise makes it difficult to record an OAE response above the noise floor, particularly at low frequencies. Even in the United States, where ABRs and OAEs are readily available at clinical centers, it is more convenient for subjects to undergo testing in their homes or in the field.

If subjects with NSHI are not tested early in life by OAEs and/or CMs and ABR, hearing loss initially presenting as AN may not be recognized as such. Approximately 20-30% of AN subjects will ultimately demonstrate loss of OHC function to develop hearing loss that is both sensory and neural (Starr et al., 2001). AN is increasingly identified in newborns undergoing hearing screening by ABR and OAEs, but acquired causes of AN (anoxia, prematurity, and hyperbilirubinemia) (Stein, 1996) are more prevalent in this population as well. It is estimated that 10% of infants with absent ABR will in fact have OAEs (Berlin et al., 2003).

AN may have virtually any type of pattern on pure tone audiometry (Hood, 1998; Sininger and Starr, 2001). Thresholds may be normal or elevated, ranging from mild to profound levels. Patterns include a flat loss across all frequencies, down-sloping (high frequency) loss, or quite commonly, up-sloping (low frequency) loss (Starr et al., 1996; Doyle et al., 1998; Starr et al., 2000). A study of children with AN found that thresholds in the mild to severe range were more commonly seen with up-sloping loss, while profound hearing loss was associated with flat or corner audiograms (Rance et al., 1999).

AN may accompany peripheral neuropathy in a variety of autosomal dominant syndromes including Freidreich's ataxia (Satya-Murti et al., 1980) and Charcot-Marie-Tooth disease, which has been attributed to mutations in several genes including MPZ, PMP22, GJB1, and EGR2 (Lupski et al., 1991; Bergoffen et al., 1993; Warner et al., 1998; Starr et al., 2003). Mutations in NDRG1 underlie the autosomal recessive hereditary motor and sensory neuropathy-Lom (Kalaydjieva et al., 2000). Nonsyndromic AN (unassociated with peripheral neuropathy) is most commonly described as a recessive trait (Madden et al., 2002; Varga et al., 2003; Wang et al., 2003). Mutations in the otoferlin (OTOF) gene, the cause of DFNB9 recessive deafness, are implicated in recessive nonsyndromic AN as well. A study of recessive AN in 4 sibships of 2-3 affected individuals described linkage to the DFNB9 locus (Varga et al., 2003); of the four families in the study, one had homozygous mutations; 2 had heterozygous mutations, and 1 had no detectable OTOF mutation. Examination of transient evoked OAEs in subjects with known homozygous OTOF mutations revealed responses that were unilateral in 4/21 and bilateral in 6/21; 2 additional subjects with AN were found heterozygous for OTOF mutations (Rodriguez-Ballesteros et al., 2003). In addition, X-linked recessive (Wang et al., 2003) and autosomal dominant inheritance (Bonfils et al., 1991) of nonsyndromic AN have been described, but no genetic studies have been reported in these families.

In order to develop a practical clinical algorithm incorporating genetic testing for dominant hearing loss, we believe it is imperative to continue to map and clone new deafness genes. While functional studies of known deafness genes are clearly important, it may be very challenging to explain function for novel genes such as WFS1, which encodes a protein completely unrelated to any other known. Family-based linkage analysis and positional candidate/cloning approaches are powerful methods to unequivocally identify deafness genes rather than attempting association studies in potentially heterogeneous populations. Genetic testing can then be offered to families that are too small for linkage analysis or even to individual patients.

Despite the relatively weak genetic evidence, many clinical centers have already begun incorporating OTOF screening for patients with AN (Rodriguez-Ballesteros et al., 2003). Like LFSNHL and WFS1 mutations (Lesperance et al., 2003; Smith and Huygen, 2003), the specific phenotype of preservation of OAEs with apparent SNHL on pure tone audiometry can define a more genetically homogeneous population. Thus, identifying a gene responsible for AN is likely to have immediate clinical impact, as centers currently offering genetic testing can easily add AUNA1 into their protocol for AN patients. An increasing awareness of AN as a hereditary trait will highlight the importance of using clinical testing to identify the site of the deafness-causing lesion at the OHC, IHC, synapses, or auditory nerve to improve our understanding of the pathophysiology of deafness. The long-term goal of this work is to develop a strategy for the clinical evaluation of dominant genetic hearing loss based on translation of genetic research advances.