1. Introduction
Childhood hearing impairment (HI) is associated with impaired language acquisition, learning, speech development and affects 34 million children worldwide (World Health Organization). Approximately 1/1000 children are born with hearing loss, of which approximately 80% is genetic [
1]. HI can be part of a syndrome with the presence of other medical anomalies, or it can be nonsyndromic. Currently, 120 nonsyndromic HI genes have been identified, with 59% having an autosomal recessive (AR), 37% an autosomal dominant (AD), and 5% an X-linked mode of inheritance (Hereditary hearing loss homepage). However, many genes remain to be identified due to the complexity of the hearing system and due to the understudy of some ancestries [
2].
Nonsyndromic HI has no association with additional features or abnormalities. However, it can be associated with abnormalities of the middle ear and/or inner ear [
1]. A large number of these abnormalities are mild, but bilateral cochlear aplasia, i.e., bilateral absence of the cochlea, is an ultra-rare and severe developmental abnormality of the inner ear. Approximately 0.3% of children with congenital sensorineural HI are estimated to have bilateral cochlear aplasia [
3]. However, this estimate is predominately based on children who were candidates for a cochlear implant, and they usually present with severe-to-profound HI [
3,
4,
5].
We previously identified de novo loss-of-function variants in
GREB1L in two individuals with profound nonsyndromic HI with inner ear and cochleovestibular nerve (or 8th cranial) malformations (
Table 1) [
5,
6]. Affected individuals had either absent cochleae bilaterally [p.(Glu1410fs)] or an absent cochlea on the right and incomplete partition type I on the left [p.(Arg328*)]. Both individuals also displayed abnormalities of their vestibules and absent 8th cranial nerves [
6]. In addition,
greb1l−/− zebrafish exhibit a loss of and/or abnormal sensory epithelia innervation, including a loss of the anterior cristae nerve and an abnormal innervation pathway from the occipital lateral line neuromast. These findings in humans and model organisms confirm the importance of
GREB1L in sensory innervation [
6]. Furthermore, Greb1l is widely expressed during craniofacial development, including the otic vesicle [
6,
7], and
Greb1l−/− mice are embryological lethal and demonstrate severe abnormalities, including craniofacial and renal abnormalities [
8].
Greb1l+/− mice show an abnormal embryo size, growth retardation [
9] and mild abnormalities to their kidneys and ureters [
8].
In addition, de novo or autosomal dominantly inherited variants (often with reduced penetrance) have previously been implicated in individuals with renal, bladder and genital malformations [
8,
13,
14]. Renal hypoplasia/aplasia 3 (RHDA3) is a severe developmental disorder characterized by abnormal kidney development and is caused by heterozygous
GREB1L variants. Although the phenotype can be highly variable, the disorder falls within the most severe end of the spectrum of congenital anomalies of the kidney and urinary tract. In many of these cases, children were aborted or stillborn due to the severity of the malformations, such as bilateral renal aplasia [
8,
13,
14].
In this article, we have, for the first time, identified a family with congenital profound HI that segregates a missense variant in GREB1L with an AD mode of inheritance and also report on an additional case with bilateral cochlear and cochlear nerve aplasia with a GREB1L variant.
4. Discussion
HI in children is both genetically and phenotypically heterogeneous. Identification of novel genes implicated in congenital HI is important to understand normal hearing and ear development, for patient management and intervention and for the development of novel therapeutic strategies.
We identified two families with congenital profound nonsyndromic sensorineural HI that segregate missense variants [p.(Asn283Ser) and p.(Thr116Ile)] in
GREB1L (
Figure 1). GREB1L is a premigratory neural crest (NC) regulatory molecule implicated in the embryonic development of many tissues [
40]. The cranial NC is important in the development of the peripheral nervous system and non-neural tissues, including craniofacial connective and skeletal tissues [
41]. In addition, it also gives rise to the stria vascularis of the inner ear and the glia cells of the cochleovestibular nerve and inner ear ganglion [
42].
greb1l has also been implicated in Hoxb1 and Shh
a signaling in zebrafish [
14], important pathways in the inner ear and cranial nerve development [
43,
44,
45].
Previous reports on disease-related
GREB1L variants showed that a variable phenotype is present, including within families segregating the same variant (e.g., left vs. right ear) [
36,
37]. In addition, a high level of reduced penetrance has been reported, including in family 2 of this study. There is no evidence that variants cluster within specific domains of the protein (
Table 2;
Figure 2). This finding is similar to what was observed for
EYA1, an NC regulatory molecule which is involved branchio-oto-renal (BOR)/branchio-otic (BO) syndrome etiology [
46].
EYA1 is also characterized by a high level of phenotypic variation between patients, even within the same family, and the severity of the phenotype does not correlate with the type of variant nor with the domain involved. In BOR patients with
EYA1 variants, which presents with both ear and renal abnormalities, normal kidneys were often observed in family members with BOR while other family members had renal abnormalities [
46]. Many neurocristopathies typically show this variable phenotypic profile amongst patients, even within families or within the same individual (left vs. right) [
47], and multiple hypotheses have been suggested to explain this phenomenon, such as environmental factors and genetic modifiers [
6,
46,
48]. However, as the NC is a transient and migratory cell population during development, there are also complex micro-regulations that could disturb NC migration during development. Because of this, the path of NC migration that ends up affected due to
GREB1L dysfunction could perhaps be attributed to chance. An example of this can be found in knockout (Wv/Wv) mice. These mice have a defect in c-kit, a NC migration regulatory molecule involved in the migration and proliferation of melanocytes in the inner ear. Wv/Wv mice show uni- or bilateral inner ear issues with variable hearing levels, and this variability in inner-ear phenotype was found to be reflected by the number of melanocytes present and how far they migrated along each cochlea during development [
49].
The particular link between renal and ear abnormalities has previously been demonstrated [
47], including in neurocristopathies. Several neural crest regulatory molecules are known to cause ear/kidney syndromes with variable expression of both ear and kidney phenotypes (e.g.,
EYA1,
SIX1,
SIX5,
CHD7,
MASP1,
TBX1), involved in BOR/BO syndrome, CHARGE syndrome, 3MC syndrome and DiGeorge syndrome [
5,
47]. In addition to these, there are also several other disorders with a specific renal/ear link, such as Alport syndrome and Bartter syndrome [
50,
51]. Interestingly, when reviewing all variants reported in
GREB1L to date, we also demonstrate that 14% of
GREB1L variants (N = 7) have been associated with ear-related issues. It is to be noted however, that many of the previous reports (focused on renal malformations) included aborted/stillborn fetuses, in which hearing could not have been assessed. In addition, inner ear and cochlear nerve malformations cannot be assessed via prenatal ultrasound and if an autopsy was performed and would usually not be detected on routine autopsy. Therefore, the number of ear malformations associated with
GREB1L variants is likely under-reported. Last, this renal/ear link is also seen in Mayer-Rokitansky-Kuster-Hauser (MRKH) syndrome, characterized by abnormal development of the internal reproductive system in females, and is also caused by
GREB1L variants (
Table 2). Interestingly, HI is reported in 10–25% of individuals with MRKH syndrome [
37].
We also detected a maternal bias in the inheritance of
GREB1L variants (
Table 2). This maternal bias has previously observed and two mechanisms have been suggested: (1) imprinting [
8,
36] (2) or
GREB1L variants could affect male fertility resulting in a low rate of paternal inheritance [
8]. Genital issues, including uterus aplasia, are common and have been reported in many females (
Table 2), but the presence in males may be underestimated as the defect might not be a gross morphological abnormality that causes infertility.
De novo
GREB1L variants have been previously implicated in a phenotype which consists of profound HI and inner ear and cochleovestibular nerve malformations [
6]. The inner ear malformation seen in family 2 is remarkably similar to the patients previously reported with
de novo GREB1L variants (
Table 1) [
6], and includes cochlear aplasia, cochlear nerve aplasia and bilateral dysplastic vestibules and semicircular canals (
Figure 1), an ultra-rare phenotype. The finding of multiple independent cases with
GREB1L variants and this exact ultra-rare phenotype is significant [
6]. In addition,
greb1l−/− zebrafish (p.Gln408Ter) exhibit a loss of or abnormal sensory epithelia innervation [
6], supporting the importance of
GREB1L in the inner ear and nerve development.
Unfortunately, we were unable to perform temporal bone imaging in the affected members of family 1 since they are located in a remote village in Pakistan. The profound bilateral congenital HI phenotype observed for affected members of this family suggests that it may also be due to inner ear/cochleovestibular nerve malformations. Since sample collection for DNA extraction and genetic screening is easier to implement in areas with limited access to modern healthcare systems than temporal bone imaging, we believe including GREB1L in diagnostic screening for nonsyndromic HI is valuable.
In conclusion, we demonstrate that autosomal dominantly inherited variants in GREB1L are involved in profound sensorineural HI etiology and show that GREB1L behaves with a similar phenotypic variance compared to other neurocristopathies. In addition, we recommend including GREB1L in diagnostic screening panels for nonsyndromic HI.