Auditory neuropathy (AN), which is a clinical disorder featuring auditory processing dysfunction, is characterized by normal outer hair cell (OHC) function in the cochlea and abnormal neural function at the level of the inner hair cells (IHCs), the cochlear nerve, or their junction. Patients with AN commonly have normal otoacoustic emissions (OAEs) and cochlear microphonics (CMs) but absent or severely abnormal auditory brainstem responses (ABRs). Pure-tone audiometry can vary widely from normal hearing thresholds to profound hearing loss (HL). Acoustic reflexes (ARs) are often absent, and speech recognition is worse than would be predicted from the pure-tone thresholds, particularly in ambient noise (Starr et al., 1996). Although individuals with AN retain OHC function, they usually have impaired neural coding of sound stimuli as a result of lesions involving the auditory nerve itself, the IHCs and/or their synapses or auditory cortex abnormalities (Kaga et al., 1996). Accordingly, the terms “auditory synaptopathy” and “auditory neuropathy” are applied to this disease when it is due to synaptic and neural deficits, respectively (Moser and Starr, 2016).

Among patients with sensorineural HL, AN accounts for 0.5–15% of cases (Madden et al., 2002; Sanyelbhaa Talaat et al., 2009; Gohari et al., 2019), with a prevalence of approximately 0.23% in at-risk children (Rance et al., 1999); however, the diverse etiologies of AN are only beginning to be characterized. Genetic factors and the effects of a wide range of other etiologies (anoxia, certain infectious diseases, and hyperbilirubinemia) are estimated to account for 42 and 10% of all cases of AN, respectively, and the remaining 48% of cases lack a defined etiology (Starr et al., 2000, 2008). All forms of AN may be present in isolation (non-syndromic AN) or with multisystem involvement, including peripheral and/or optic neuropathies as well as various central nervous system (CNS) disorders (syndromic AN) (Santarelli, 2010).

Due to the distinct clinical and pathological features of non-syndromic AN (NSAN), its diagnosis mainly relies on audiological measures and electrophysiological tests along with ancillary methods, such as OAEs, acoustic immittance, temporal bone high-resolution computed tomography (HRCT) and magnetic resonance imaging (MRI). Pure-tone audiometry enables the degree, type, and configuration of HL to be identified. In addition to revealing impaired speech perception that is out of proportion to HL, electrophysiological tests [including ABRs and electrocochleography (ECochG)] and OAEs can assist in identifying the sites of the lesions along the auditory pathway since dysfunction of the auditory nerve, IHCs and/or IHC ribbon synapses [absent/abnormal ARs, ABRs, and action potential (AP) on ECochG] is accompanied by preserved OHC measurements (OAEs and/or CMs) (Santarelli, 2010). Furthermore, HRCT and MRI approaches have become effective methods for identifying possible structural or inflammatory abnormalities of the auditory nerve and auditory pathway that cause HL (Sharma et al., 2018). Therefore, when diagnosing AN, both imaging modalities are powerful methods for ruling out any organic lesions.

Temperature-sensitive auditory neuropathy (TSAN) is a very rare form of NSAN with distinct phenotypes. The most striking feature of TSAN is that patients present with transient HL after a rise in the core body temperature. This disorder was first described by Gorga et al. (1995) in a child with recurrent but completely reversible HL during febrile conditions. Clinical examinations showed that poor speech understanding emerged earlier than pure-tone HL and resolved more slowly than the loss of auditory function. Tests of cochlear function, including OAEs and ECochG, were normal; however, all the tests related to central connections within the brainstem, such as ARs, ABRs, and mechanically evoked trigeminofacial reflexes, were abnormal regardless of whether auditory function was normal or impaired at the time. Moreover, the middle-latency responses and auditory late potentials appeared normal. The authors predicted that the site of the lesion was the brainstem, rather than the peripheral nervous system or higher levels in the auditory pathway. Three years later, Starr et al. (1998) identified three children—two siblings and an unrelated child—who carried similar phenotypes, as confirmed by tests performed in both febrile and afebrile states. Two of the three children may have had a summating potential (SP), the existence of which was not examined in the first reported patient. Furthermore, all three patients had mild HL either at all frequencies or specifically at low frequencies when afebrile. In addition, one patient showed abnormal median sensory nerve conduction. The authors proposed that the patients’ transient deafness resulted from a demyelinating disorder of the auditory nerve and concluded that the reported cases represented an atypical form of AN, which was termed TSAN. Examples of both sporadic and familial TSAN have been described (Cianfrone et al., 2006; Varga et al., 2006; Madanoglu and Derinsu, 2007; Marlin et al., 2010; Dimitrijevic et al., 2011; Matsunaga et al., 2012; Wynne et al., 2013; Kaga, 2016; Zhang et al., 2016). Although fever-associated HL is reversible and OAEs are always normal, all reported cases have two abnormal auditory findings in common: ABRs and ARs are abnormal during both febrile and afebrile episodes. Further research revealed that variants of the OTOF gene are the main cause of this disease (Varga et al., 2006; Romanos et al., 2009; Marlin et al., 2010; Wang et al., 2010; Matsunaga et al., 2012; Zhang et al., 2016).

The pathogenic mechanism of TSAN is only beginning to be understood. To date, only one study performed by Strenzke et al. (2016) successfully built an Otof^I515T/I515T mutant mouse model, which exhibited a similar severity of HL to that in humans. Further research revealed that impaired exocytosis at hair cell ribbon synapses due to abnormal synaptic vesicles caused by Ile515Thr-otoferlin was the main pathogenic mechanism of TSAN. Notably, the mutant mice failed to show more severe hearing loss at high body temperatures, unlike the patients. The authors concluded that the lack of a human RXR motif (20 amino acids), which is thermally sensitive in mice, was the cause of inconsistent phenotypes between mutant mice and TSAN patients. However, TSAN patients are rare, the relationship between the severity of phenotypes and variants of the OTOF gene is still unclear, and clinical tests have never detected any demonstrable structural lesion or underlying mechanism to explain this uncommon type of HL.

Herein, we describe a TSAN-afflicted family with a distinctive pattern of phenotypic and genetic features. Four siblings, including a pair of identical twins, were examined in our clinic, and diagnosed with TSAN. Using next-generation sequencing (NGS) technology, we identified two compound heterozygous variants of the OTOF gene that are associated with TSAN. These data reveal genotype–phenotype correlations in TSAN, providing an improved understanding of the diagnostic features and natural history of this disease.

In the mammalian inner ear, hair cells (HCs) transform mechanical vibrations into electrical signals (Wang et al., 2017; He et al., 2019; Liu Y. et al., 2019; Qi et al., 2019, 2020; Jiang et al., 2020; Zhou et al., 2020), and spiral ganglion neurons (SGNs) function as electric signal-transduction cells (Sun et al., 2016; Liu et al., 2018, 2021; Guo et al., 2019, 2020, 2021; Zhao et al., 2019). While supporting cells function as potential resources to regenerate HCs after damage (Lu et al., 2017; Zhang et al., 2017; Tan et al., 2019; Zhang S. et al., 2019, 2020; Zhang Y. et al., 2020; Chen et al., 2021). Sensorineural hearing loss has many etiologies, varying from genetic defects (Zhu et al., 2018; Fang et al., 2019; Yang et al., 2019; Qian et al., 2020; Cheng et al., 2021; Fu et al., 2021a; Lv et al., 2021; Zhang et al., 2021) to environmental factors, including aging (Zhang et al., 2017; Fu et al., 2018; Li H. et al., 2018; He et al., 2020, 2021), noise and ototoxic drugs (Liu et al., 2016; He et al., 2017; Li A. et al., 2018; Zhang Y. et al., 2019; Gao et al., 2020; Zhong et al., 2020; Fu et al., 2021b). TSAN is a rare phenotype with high clinical heterogeneity, manifesting as transient bilateral HL due to an increase in the core body temperature. The transient, conditional nature of this HL makes diagnosis challenging. In the present study, the diagnosis of AN was based on normal CM responses and abnormal ABRs during febrile episodes, although the tympanometry and AR measurement results were abnormal in one of four cases at the beginning of febrile episodes. However, further audiometric evaluations showed that the patients’ behavioral thresholds, ABRs, and ECochG signals recovered to normal during afebrile episodes. Unlike other patients with TSAN or other forms of AN, the ECochG pattern in our patients during afebrile episodes was reversible, suggesting that auditory function almost completely recovered, although ARs continued to be absent. Therefore, certain exceptions to the usual features of AN often need to be considered when making a diagnosis, as 40% of affected individuals have abnormal rather than absent ABRs, ARs are sometimes present at elevated thresholds rather than being absent, and OAEs are absent in approximately 30% of AN patients upon retesting (Starr et al., 2001, 2008). In the present study, the proband’s abnormal DPOAEs with a type C tympanogram and absent ARs during the febrile state were very likely attributable to dysfunction of the middle ear rather than OHC and not related to TSAN, although no abnormality was detected in physical or neuroimaging examinations. More notably, however, negative middle ear pressure should result in a retracted TM, which may or may not be observable upon physical examination, depending on the experience of the person conducting the examination, the amount of negative pressure, and the use of a standard otoscope versus microscopic examination of the TM. Furthermore, imaging studies will not show a retracted TM as an isolated finding. In contrast, his twin brother (II-3) showed consistent normal DPOAEs and a type A tympanogram, regardless of the body temperature, although they shared the same genotype. However, the changes in CMs and ABRs during each febrile event in the twins were the same, indicating that both tests are more reliable than others when diagnosing AN, especially when DPOAEs and tympanograms differ among patients in the same family. Consequently, present CMs with absent ABRs are recommended to diagnose AN (Mittal et al., 2012). Therefore, patients in this study were diagnosed with AN when febrile despite the recovery of their auditory function after fever.

Although several genes are involved in AN, only mutations of the OTOF gene are responsible for TSAN (Del Castillo and Del Castillo, 2012). To date, five loci (including four identified genes) and six genes are known to be associated with NSAN: DFNB9 (OTOF gene), DFNB59 (PJVK gene), and GJB2 for autosomal recessive AN; AUNA1 (DIAPH3 gene), AUNA2 (undefined gene), SLC17A8, PCDH9, DIAPH1, and TMEM43 for autosomal dominant AN; DFNX5 (AIFM1 gene) for X-linked recessive AN; and mitochondrial 12S rRNA (T1095C) (Varga et al., 2003; Cheng et al., 2005; Wang et al., 2005; Delmaghani et al., 2006; Ruel et al., 2008; Santarelli et al., 2008; Grati et al., 2009; Schoen et al., 2010; Zong et al., 2015; Lang-Roth et al., 2017; Wu et al., 2020; Jang et al., 2021). However, variants of the OTOF gene have been shown to be major contributors to NSAN (Varga et al., 2006; Rodríguez-Ballesteros et al., 2008; Romanos et al., 2009; Chiu et al., 2010; Wang et al., 2010). Moreover, DNA variations in OTOF differ among regions and ethnic populations (Rodríguez-Ballesteros et al., 2008; Del Castillo and Del Castillo, 2012; Matsunaga et al., 2012; Almontashiri et al., 2018). Two families with familial TSAN have been shown to carry OTOF variants. The first family was reported by Starr et al. in 1998 (Starr et al., 1998). Varga et al. (2006) showed that in this family, the two siblings with TSAN carried a heterozygous c.1544T > C (p.Ile515Thr) variant in the OTOF gene (Varga et al., 2006). This variant was inherited from their unaffected father and was absent in their mother and maternal siblings, and it is known to contribute to profound prelingual HL in humans and TSAN or NSAN in mouse models (Mirghomizadeh et al., 2002; Strenzke et al., 2016; Michalski et al., 2017). Further study revealed that the other mutant allele containing c.3346C > T (p.Arg1116Ter) in the OTOF gene (Strenzke et al., 2016). The second familial TSAN cluster was found in a consanguineous family with three members carrying a homozygous p.Glu1804del variant (Marlin et al., 2010). Furthermore, several sporadic cases of TSAN have been reported to be caused by variants in the OTOF gene (Romanos et al., 2009; Matsunaga et al., 2012; Zhang et al., 2016). To date, a total of nine variants of the OTOF gene have been identified as associated with TSAN (Table 1 and Figure 6A). In this study, we reported a third cluster of familial TSAN, which includes three cases with two variants of the OTOF gene (which encodes otoferlin protein), c.5098G > C (p.Glu1700Gln) and c.4882C > A (p.Pro1628Thr). The reasons why both variants were considered part of the cause of TSAN in this family are as follows: (1) both are missense variants affecting residues that are completely preserved among different species; (2) p.Glu1700Gln has been shown to be pathogenic (Chiu et al., 2010; Chen et al., 2018; Qiu et al., 2019; Wu et al., 2019); (3) p.Pro1628Thr, which is reported here for the first time, was absent in 200 normal controls and has a very low allele frequency in a public database (8.67 × 10^–6 in ExAC); and (4) both variants strongly co-segregated with the TSAN phenotype.

Most patients with NSAN caused by OTOF variants present with severe prelingual or profound deafness; however, the phenotypes of TSAN patients reported in this and previous studies did not follow this pattern. In previous studies, most TSAN patients had variable HL (normal to mild) when afebrile, and their hearing thresholds dropped to a greater degree of HL, with associated communication difficulties only during febrile episodes. In addition, patients with AN caused by OTOF variants tended to have more frequently affected and more severely affected DPOAEs than those with no OTOF variants (Kitao et al., 2019). Otoferlin, which is encoded by the OTOF gene, is a six-C2 domain (C2A–F) protein that also has a transmembrane domain (TMD), and its function is auditory neurotransmission at the hair cell ribbon synapse in the inner ear (Roux et al., 2006; Vogl et al., 2016; Michalski et al., 2017). C2 domains have been implicated in Ca²⁺ binding, and they play a crucial role in the function of otoferlin, which may contribute to the location specificity of most pathogenic variants of the OTOF gene (Del Castillo and Del Castillo, 2012). Notably, at least one of the alleles observed in these cases carries a missense variant affecting a C2 domain of otoferlin (Figure 6A). Moreover, subjects with two C2 domain-affecting variants presented with more severe phenotypes than did those with only one (Table 1). Relative to subjects in earlier studies, the three subjects in this study showed mild phenotypes, with the improvement in pure-tone thresholds from mild HL during the febrile condition to normal levels during the afebrile periods and the recovery of auditory function (except for ARs) during afebrile periods. Neither of the variants identified in this study is located in a C2 domain (both are between domains C2E and C2F); in addition, the p.Glu1700Gln variant has been shown to cause progressive mild to moderate prelingual HL, and OAEs in patients with AN carrying homozygous or compound heterozygous p.Glu1700Gln may be consistently present, consistently absent, or even present in the first test and absent later (Chiu et al., 2010; Wu et al., 2018). However, in the present study, the causes of the proband’s absent DPOAEs only when febrile are still unknown, although he has the same genetic background as his twin brother. Remarkably, we did not identify modifiers using WES technology, suggesting that the variable phenotypes associated with these genotypes may result from environmental and/or genetic factors that have not yet been identified. In addition, the temperature-sensitive phenotype of HL may also be a direct consequence of defects in normal otoferlin-mediated synaptic vesicle trafficking in the inner ear. This issue may be addressed using temperature-sensitive variants that have been reported in several human diseases, including albinism and cystic fibrosis (King et al., 1991; Sharma et al., 2001; Wang et al., 2008), and decreased activity of otoferlin due to structural changes with rising temperatures may have an additive effect on the phenotype, as verified in previous reports (Strenzke et al., 2016).

Although the precise synaptic mechanism of the action of otoferlin at hair cell ribbon synapses remains unclear, the impairment of otoferlin in temperature-sensitive synaptic neurotransmitter release has been suggested to contribute to the phenotype of TSAN (Kaga, 2016; Kindt and Sheets, 2018). Normal auditory function depends on faithful information transfer, which requires otoferlin-dependent IHC exocytosis to be indefatigable, highly efficient, and accurately synchronized (Johnson, 1980; Griesinger et al., 2005; Roux et al., 2006; Kitcher et al., 2021). Reliable and temporally precise cochlear potentials are characterized by fast rise times, short onsets, and short peak latencies (Liu W. et al., 2019; Rutherford et al., 2021). Based on the findings from the present and other studies (Table 1), the severity of the phenotype associated with TSAN due to dysfunctional neurotransmitter release appears to reflect variants that alter the C2 domains of otoferlin. Therefore, the patients in previous studies had more severe HL in the febrile and afebrile states than the patients in the present study. We suggest that the possible pathogenesis and phenotype-genotype relations of TSAN caused by the dysfunction of otoferlin-dependent exocytosis in the IHC ribbon synapses are shown in Figures 6B,C. According to this hypothesis, three different clinical phenotypes may be interpreted. First, in TSAN patients with two C2 domain-affecting variants (CAVs) who became febrile, synaptic transmitter release was transiently blocked in almost all IHC–SGN synapses, especially the more seriously affected remaining synapses. The recorded cochlear current waves were low and prolonged or, in some cases, were hardly detectable. Second, when febrile, TSAN patients with only one CAV had better hearing thresholds than those with two CAVs. Large numbers of IHC–SGN synapses temporarily failed to release vesicles, with the remaining synapses being more heavily affected than those during afebrile states. Moreover, the initiation and propagation of spikes along the SGNs tended to vary in time and intensity. The measurements revealed the occurrence of cochlear potentials of variable size and shape in this type of TSAN patient under febrile conditions. Third, as in TSAN patients with one or two CAVs, mutant otoferlin in TSAN patients with no CAV exhibited impaired neurotransmitter release at IHC–SGN synapses. However, in a febrile state, the impairment caused by mutant otoferlin was less severe in TSAN patients with no CAV than in patients with one or two CAVs. As a result, synaptic transmission was inhibited for a very short period in the febrile state and rapidly increased after the body temperature decreased. The cochlear potentials, i.e., ABRs and ECochG signals, improved from no response to almost normal after the body temperature dropped. Importantly, the small magnitude of the effect on synaptic transmission may be the main reason why fever-related auditory dysfunction is almost completely reversible (with continued absent ARs) in TSAN patients with no CAV. Although the real molecular mechanisms are only beginning to be understood, otoferlin-induced defective synaptic transmission at the IHC ribbon synapse is one of the most important underlying causes of TSAN.

In this study, we described the detailed courses of disease in four male siblings from a non-consanguineous Chinese TSAN family who complained of communication difficulties when febrile, and we explored the genetic etiology of their condition using NGS technology. We investigated the detailed clinical progression of autosomal recessive TSAN in four siblings with two OTOF variants that did not affect any C2 domain of otoferlin, and we also identified a unique characterization of normal ABRs in afebrile episodes that was different from previously reported cases. Moreover, the presence of CMs with absent (or markedly abnormal) ABRs is a reliable diagnostic criterion for AN. These observations enrich the current knowledge of the phenotypes and genotypes of TSAN and may lay a foundation for further studies of its pathogenesis.

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found below: the ClinVar repository (https://www.ncbi.nlm.nih.gov/clinvar/), accession numbers: SCV001787151 and SCV001787152.

The studies involving human participants were reviewed and approved by Committees of Medical Ethics at the Nanfang Hospital, the Chinese PLA General Hospital, and the Lanzhou University Second Hospital. Written informed consent to participate in this study was provided by the participants’ legal guardian/next of kin. Written informed consent was obtained from the individual(s), and minor(s)’ legal guardian/next of kin, for the publication of any potentially identifiable images or data included in this article.

Y-MZ collected and analyzed data and drafted the main manuscript. Y-FL and Y-LL collected and processed the data. QL and X-PL supervised and conceptualized the study. XG and PD designed and conceptualized the study, interpreted the data, and revised and finalized the manuscript. All authors read and approved the final manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.