POLD3 haploinsufficiency is linked to non-syndromic sensorineural adult-onset progressive hearing and balance impairments

Abstract

Hearing impairment (HI) is a significant health concern globally, influenced by genetic and environmental factors. We had identified a homozygous pathogenic variant in POLD3 in a Lebanese patient with an autosomal congenital recessive syndromic hearing loss (MIM#620869). This variant was found at heterozygous state in the parents, who developed progressive hearing impairment around age 40. We conducted a thorough clinical and genetic assessment of sixteen family members, including physical exams, audiometry and vestibular function evaluations. Additionally, gene expression analysis of the Pold3 gene was performed in mice using RNAscope. Twelve individuals were heterozygous for the variant in POLD3, of whom eight showed bilateral adult-onset HI, typically starting around ages 40–50, and two older patients displaying unilateral vestibular weakness. Additionally, two carriers of the variant developed cancer at an early age. RNAscope confirmed Pold3 expression in auditory and vestibular neurons. Exome sequencing analysis excluded the presence of pathogenic variants in any known hearing impairment or cancer predisposition genes. We present herein, for the first time, evidence of a heterozygous pathogenic POLD3 variant associated with a novel form of autosomal dominant progressive adult-onset hearing and vestibular impairments. We also highlight the necessity for further exploration of the role of POLD3 in cancer predisposition.

Introduction

Hearing impairment (HI) refers to an abnormal hearing function leading to partial or complete inability to hear sounds. It is a significant global health issue affecting people of all ages with estimated 2.5 billion individuals projected to be affected by 2050, according to the WHO (https://www.who.int/; February 2024).

HI can be caused by genetic, and/or environmental factors including infections during pregnancy, premature birth and low birth weight, ear-malformations, exposure to harmful substances, teratogens, ototoxic medications, as well as noise exposures [1, 2]. Its prevalence increases with age, becoming the predominant sensory deficit among the elderly; often starting at an earlier age but remaining imperceptible. Indeed, in some cases, hearing can start to decline in adulthood, as early as 30 s or 40 s, but is typically noticed in the 50 s to early 60 s [3]. Studies examining heritability in twins and families have underscored a significant role for genetic predisposition for adult-onset HI, revealing heritability rates of up to 50% [4–7]. Recent family-based studies, combined with genome-wide association studies (GWAS), in cases with adult-onset HI have resulted in the identification of several candidate HI genes [8–11]. Most of these are inherited in an autosomal dominant (AD) pattern. Interestingly, many examples of variants involved in AD adult-onset HI are associated, when present at a homozygous state, with autosomal recessive (AR) non-syndromic or syndromic early-onset deafness forms [12] [https://hereditaryhearingloss.org/]. For instance, CDH23 variants cause a broad range of phenotypes of non-syndromic hearing loss (OMIM #601386, DFNB12); from AR congenital syndromic (Usher syndrome) and non-syndromic profound deafness to autosomal dominant (AD) late-onset progressive HI [13]. Similarly, HI associated with dizziness/vertigo due to mild vestibular dysfunction has been reported in elderly patients harboring heterozygous variants in WSF1 gene associated with either AR syndromic hearing loss (OMIM #222300), or AD adult-onset HI (DFNA6 and DFNA38; OMIM #600965) [14, 15] In addition, mutations in WFS1 have been reported, in a Finnish family with age-related HI (ARHI) [16]. This observation is not restricted to HI. Indeed, as in several conditions, carriers often described as ‘unaffected’, may actually show milder symptoms for the recessive disease they are carrying [17, 18].

A homozygous pathogenic POLD3 variant, was recently reported by our team in a Lebanese patient presenting with an AR syndromic form of immunodeficiency (immunodeficiency 122, MIM#620869) including congenital hearing loss [19]. Interestingly, both his parents and their siblings, who are heterozygous carriers for the same variant, were found to exhibit mild to severe progressive HI starting between 40 and 50 years of age.

Here we report for the first time, the involvement of a heterozygous pathogenic variant in POLD3 in a new form of AD progressive adult-onset hearing and vestibular impairments, in a large Lebanese family spanning three generations. We also highlight the necessity for further exploration of the role of POLD3 in cancer predisposition.

Methods

Patients

We herein describe a large three-generation Lebanese family originating from North Lebanon (Fig. 1), referred to our clinic for clinical and genetic evaluation of hearing impairment. A branch of this family has been previously evaluated and reported by our team [19].

Fig. 1. Pedigree of the investigated Lebanese family.

N/+: Heterozygous carrier of the p.Ile10Thr variant in POLD3. N/N: Homozygous normal. POLD3 genotypes were added for available patients.

Physical examination

We conducted a comprehensive physical examination on all individuals of this family (Fig. 1, Table 1). The examination encompassed assessment of vital signs, including temperature, pulse, blood pressure, and respiratory rate. A thorough evaluation of general appearance was performed, followed by detailed examinations of the head and neck, cardiovascular system, respiratory system, abdomen, musculoskeletal system, and neurological function. Skin examination for abnormalities or lesions was also conducted. Additionally, we assessed the cranial nerve function, motor skills, reflexes, and sensory perception.

Table 1.

List of the participants in this study with their age at physical examination.

Participants	Age at examination (years)	POLD3 genotype	Audiology testing
III-1	–	Not available	Not available
III-2	54	N/N	Normal hearing
III-3	–	Not available	Not available
III-4	49	N/+	Hearing impairment
III-5	46	N/N	Normal hearing
III-6	45	N/N	Normal hearing
III-7	44	N/+	Hearing impairment
III-8	–	Not available	Not available
III-9	42	N/+	Not available
III-10	36	N/+	Not available
III-11	–	Not available	Not available
III-12	30	N/+	Normal hearing
III-13	37	N/+	Hearing impairment
III-14	40	N/+	Hearing impairment
III-15	–	Not available	Not available
III-17	–	Not available	Not available
III-18	–	Not available	Not available
III-19	50	N/+	Hearing impairment
III-20	–	Not available	Not available
III-21	–	Not available	Not available
III-22	42	N/+	Hearing impairment
III-23	40	N/N	Normal hearing
III-24	37	N/+	Normal hearing
IV-2	14	N/N	Normal hearing
IV-4	9	N/+	Normal hearing
IV-6	6	+/+	Congenital hearing loss

This table includes the genotype of all available participants (N/+: Heterozygous carrier of the p.Ile10Thr variant in POLD3. N/N: Homozygous normal for the POLD3 variant. +/+: Homozygous mutated for the p.Ile10Thr variant in POLD3) as well as the results of hearing testing.

Hearing evaluation

Audiometry tests

A comprehensive audiometric assessment was completed on all available members (Table 1), to determine the type of hearing impairment. These tests include: (1) Tympanometry that was performed to evaluate the middle ear function using a probe tone of 226 Hz and the results were tabulated based on the Jerger classification of tympanograms; (2) Acoustic reflex test to assess the integrity of the auditory system including the neural pathway and the response of the stapedius muscle in the presence of an acoustically loud sound; (3) Pure Tone Audiometry, both air (frequencies ranging from 250 to 8000 Hz) and bone conduction (frequencies ranging from 500 to 4000 Hz) performed for identification of the quietest sound the patient can hear or hearing threshold levels; (4) Speech discrimination test to assess how well the word recognition is at different levels of loudness and in quiet and over noise; in addition to (5) Otoacoustic emission (OAE) testing performed to assess the status of outer hair cells in the cochlea. OAEs are measured using stimulus level of 55 to 65 dB SPL intensity, a frequency range between 500 Hz and 10000 Hz, and relative to the noise floor.

Vestibular function assessment

Balance or vestibular function was assessed in individuals III.4, III.14 and III.19, using the following tests:

• Videonystagmography (VNG): This test included spontaneous nystagmus test, gaze test, optokinetic nystagmus test, positional and positioning nystagmus testing, fixation suppression, and saccade tests. The assessment of the function of both horizontal semicircular canals was performed using the caloric irrigation test.

• Caloric testing: Air was used as the stimulant during the test with warm irrigation performed first at a temperature of 50 °C. The air volume used was 8 liters with 60 s duration of airflow and a minimum period of 7 min between irrigations.

• Video Head Impulse Test (vHIT): This test was employed to assess the function of all semicircular canals.

• Cervical Vestibular Evoked Myogenic Potentials (cVEMPs): cVEMPs were performed in individuals III.4, III.14 and III.19, to assess otolithic function. Air-conducted sound using Tone Burst (TB) at frequencies 500 Hz and 1000 Hz activate the macula hair cells and their afferent neurons resulting in myogenic evoked potentials detected on the sternocleidomastoid muscle.

Genetic analysis

Isolation of genomic DNA

Written informed consent was obtained from all participants and from the legally authorized representatives of minor individuals (parental consent) to participate in this study and its publication. EDTA blood samples from all members of the family were collected for genetic studies. DNA was extracted from leucocytes by standard salt-precipitation methods.

Exome sequencing (ES)

ES was carried out in three individuals (III.13, III.14 and III.19). Briefly, exome was captured and enriched using Agilent SureSelect Human All Exon kit version 5.0 and after adding unique barcodes for each sample, samples were then multiplexed and subjected to sequencing on an Illumina HiSeq 2500 PE100–125. Reads files (FASTQ) were generated from the sequencing platform via the manufacturer’s proprietary software. Reads were aligned to the hg19/GRCh37 reference genome using the Burrows-Wheeler Aligner (BWA) package version 0.7.11. Variant calling was subsequently performed using the Genome Analysis Tool Kit (GATK) version 3.3. Variants were called using high stringency settings and annotated with VarAFT software 1.61 containing information from dbSNP147 and the Genome Aggregation database (gnomAD, http://gnomad.broadinstitute.org). Filtering of the variants was initially performed according to the frequency of the variant in the gnomAD database v2.1.1 (GRCh37) ( <0.01% and <50 heterozygous carriers or <5 homo-/hemizygous carriers), and in our in-house database ( <1 homozygous occurrence) that includes 940 Lebanese individuals. All remaining variants were assessed by several prediction tools (e.g., SIFT, PolyPhen-2, MutationTaster) and according to the guidelines established by the ACMG (American College of Medical Genetics and Genomics) for variant classification [20]. Among the retained variants, (1) those predicted to be benign/likely benign were excluded; (2) those predicted as pathogenic/likely pathogenic by at least one of the prediction tools were evaluated one by one based on the function of the gene and its involvement in Human diseases. As per the ACMG guidelines, secondary findings were reported to the patients during genetic counseling sessions. Following this step, (3) all remaining variants classified as variants of unknown significance in the coding regions and in the –20/ + 20 intronic boundaries were individually assessed as above, based on the function of the gene and its involvement in Human diseases.

Sanger sequencing

Sanger sequencing was performed in all available members (Table 1). Genomic sequence of POLD3 (NM_006591.3) was obtained from UCSC Genomic Browser (Feb. 2009 (GRCh37/hg19)). Primers used for PCR amplification were designed using Primer3 software (http://frodo.wi.mit.edu) to amplify the exon 1 of the POLD3 gene including the p.Ile10Thr (c.29T > C) detected by ES in the patients. PCR reactions were performed using Taq DNA polymerase (Invitrogen Life Technologies, Carlsbad, CA, USA). PCR fragments were run on 1% agarose gel. The fragments were purified using « SIGMA-ALDRICH TM» kit and then sequenced using the Big Dye_Terminator v1.1 Cycle Sequencing Kit (Applied Biosystems, Foster City, CA, USA). Sequence reaction was purified on Sephadex G50 (Amersham Pharmacia Biotech, Foster City, CA), and then loaded into an ABI3500 system after the addition of Hidi formamide. Electropherograms were analyzed using Sequence Analysis Software version 5.2 (Applied Biosystems) and then aligned with the reference sequences using ChromasPro v1.7.6.1 (Technelysium, Queensland, Australia).

RNAscope and imaging

RNA expression analysis was performed using RNAscope in situ hybridization and according to the RNAscope Multiplex Fluorescent Reagent Kit v2 Assay (ACD-Bio, Document number: 323100-USM; Catalogue number: 323270). Inner ear tissue (cochleae and vestibule) from P6 mice were dissected in ice-cold PBS and placed in 4% PFA at 4 °C for 24 h with gentle shaking. After fixation, samples were washed 3x with ice-cold PBS and dehydrated overnight with 10%, 20% and 30% sucrose at 4 °C, for 24 h, then embedded in OCT. Fourteen-micrometer inner ear sections were cut using a microtome. The manufacturer designed probes used for labelling: RNAscope® Probe - Mm-Pold3-C1 (ACD-Bio, Catalogue number: 1294121-C1), and RNAscope® Probe - Mm-Cacna1a-C2 (ACD-Bio, Catalogue number: 493141-C2). The fluorophores used for imaging: TSA Vivid™ Fluorophore Kit 520 (Tocris Bioscience a Bio-Techne Brand, Catalogue number: 7523) and TSA Vivid™ Fluorophore Kit 570 (Tocris Bioscience a Bio-Techne Brand, Catalog number: 7526). The samples were mounted in the VECTASHIELD® Antifade Mounting Medium (Vector Laboratories, catalogue number: H-1000-10) on the SuperFrost Plus™ Adhesion slides (Epredia, catalogue number: 10149870). ZEISS LSM 980 Airyscan 2 microscope was used for imaging (ZEISS, Germany). Z-stack images were captured at 0.150 µm steps. The images were then processed using the Fiji ImageJ software [21].

Results

Physical examination

All available affected and non-affected individuals underwent physical examination that was unremarkable for all participants (Table 1). The age of the participating subjects ranged from 30 to 77 years (Fig. 1, Table 1). Past medical history for most of the participants was unremarkable except for the following patients: (1) the patient III.18 who died from brain cancer at the age of 30 years, (2) patient III.13 who was diagnosed with a left muscle tissue neck rhabdomyosarcoma at age 16, treated with chemotherapy and radiation therapy, (3) and the proband IV.6, a 6-year-old male child of first-cousin consanguineous parents, who presented with recessive congenital syndromic severe deafness that was detected in him and in three of his older siblings (IV.1, IV.3 and IV.5). Clinical data of this family branch is detailed in a paper published by our team [19]. According to the family, Individuals I.1, I.2, II.1, II.2, and II.4 lived to old age without any significant health complications. Only II.1 had a history of hypertension and hypothyroidism, while II.2 died from a heart attack.

Eight subjects reported progressive mild to severe HI (indicated in grey in Fig. 1). The self-reported onset of HI ranged from around the age of 40 to 50 years of age. Patient III.13 exhibits a more severe HI, primarily attributable to the cancer she developed and the subsequent treatment. Two of the oldest patients exhibited mild vestibular dysfunction, but no other sensory defects nor dysmorphic features were detected.

Hearing evaluation

Audiometric tests revealed that affected members of this three-generation Lebanese family exhibit adult-onset, non-syndromic, progressive, bilateral sensorineural hearing impairment. Hearing impairment starts in the fifth decades between 40 and 50 years of age as a mild loss and progresses to severe in older ages. These tests (Fig. 2) revealed an intra and inter-individual variability but, in most cases, the high frequencies are affected first (Supplementary Fig. 1).

Fig. 2. Representative audiograms of three patients (III.19, III.4, and III.14) demonstrating intra- and inter-subject variability in hearing level.

A Patient III.19, age 50, exhibits moderate to severe sensorineural hearing impairment (SNHI) in the right ear and mild downsloping to moderate SNHI at high frequencies in the left ear. B Patient III.4, age 49, has bilateral mild downsloping to severe SNHI at mid- to high frequencies. C Patient III.14, age 40, has mild downsloping to moderate SNHI in the right ear and mild downsloping to severe SNHI in the left ear both affecting the high frequencies.

We performed a comprehensive assessment of the auditory function on three members (III.4, III.14 and III.19) representing two of the oldest (III.19 and III.4, 50 and 49 years old respectively) and one of the youngest (III.14, 40 years old) affected family members, in order to better understand the relationship between age and the progression of HI and speech recognition. Our studies included tympanometry, acoustic reflexes assessment, pure tone audiometry, speech audiometry tests, and cochlear response using otoacoustic emissions (OAE) test.

Tympanometry performed using a probe tone of 226 Hz revealed normal middle ear pressure and compliance in both ears in all three patients.

Acoustic reflexes were present within normal thresholds at 2000 Hz but were absent at 500 and 1000 Hz in both ears; and contralateral acoustic reflexes were absent at all tested frequencies except at 2000 Hz in the right ear.

For patient III.19, the audiogram revealed a moderate to severe sensorineural hearing loss at low frequencies and a moderate loss at high frequencies on the right side. The left ear showed a mild downsloping to moderate sensorineural hearing loss starting at 250 Hz (Fig. 2A). The speech reception threshold was consistent with the pure tone average bilaterally. The word recognition score was poor on the right and excellent on the left, which aligns with the patient’s hearing results (data not shown).

Interestingly, OAEs findings were consistent with the audiometric testing in all three patients (Fig. 3). For instance, in the 50-year-old patient III.19, OAEs were absent at all tested frequencies in the right ear while they were present but reduced at 1000, 1500, 2000, and 3000 Hz, and absent at the remaining frequencies in the left ear (Fig. 3A), similar to the pattern of hearing loss observed on the audiogram (Fig. 2A). Similarly, the audiogram of the 49-year-old patient III.4 revealed severe sensorineural hearing loss involving the mid- and high frequencies bilaterally (Fig. 2B) reflected by abnormal OAE measurements at all frequencies (Fig. 3B). Audiometry in the 10 years younger patient III.14 revealed, as expected, a milder phenotype compared to older patients, with a faster decline in hearing at the high frequencies in the left ear as compared to the right ear (Fig. 2C). The OAEs correlated with the audiometric findings as well (Fig. 3C).

Fig. 3. Representative Distortion Product Otoacoustic Emission (DPOAE) measurements of three patients demonstrating outer hair cell dysfunction in subjects with hearing loss.

A Patient III.19 has present but abnormal DPOAEs in the right ear across all tested frequencies (red line). On the left side, DPOAEs are present and normal at 1–3 KHz while absent at 4–10 KHz. B Patient III.4 has absent DPOPAEs bilaterally. C Patient III.14 has present and normal DPOAEs at 1–1.5 KHz, present and abnormal DPOAEs at 2–5 KHz and absent DPOAEs at 5–10 KHz in the right ear. DPOAEs in the left ear were present and normal at 1.5 KHz, present and abnormal at 2–4 KHz and at 8–9 KHz, and absent at 1 KHz, at 5–7 KHz, and 10 KHz.

Balance evaluation

Owing to the fact that age-related decline in cochlear function generally precedes the decline of vestibular function [22], we have selected the oldest patients and the youngest patient in the family to be screened for vestibular dysfunction, and compared their results. Vestibular Function Tests (VFTs) in patients III.4 (49 years old), III.14 (40 years old) and III.19 (50 years old) revealed normal oculomotor testing and normal vHIT in all three patients. Patients III.4 and III.14 showed normal caloric responses however the caloric responses in the oldest patient III.19 (age 50) revealed a 30% unilateral weakness to the right, involving the lateral semicircular canal at low frequency movements and/or its vestibular ganglion afferent neurons pathway (data not shown). Interestingly, the cVEMPs tests revealed an abnormal result in patient III.19 (age 50) suggested by the delay in the latency of the peak (P1) waves in both ears; right ear latency (R: 22.67 ms) and left ear latency (L: 21.67) tested at 1000 Hz of frequency and 110 dB of stimulus intensity (Fig. 4A). At 500 Hz this patient had no waves detected in the right ear and delayed latency in the left ear (L: 22 ms). At 1000 Hz, the N1-P1 or P1 amplitude was reduced in the right ear (0.383 μv) but normal in the left ear (0.874 μv). Patient III.4 (age 49) showed undetected cVEMP waves at 500 Hz of frequency and 105 dB of stimulus intensity in the right (Fig. 4B). Waves were detected in the left ear with normal latency (16.67 ms) and reduced amplitude (0.303 μv) at 500 Hz (Fig. 4B). This patient was not tested at the 1000 Hz frequency. Patient III.14 (age 40) had normal cVEMPs at both frequencies (1000 and 500 Hz) (Fig. 4C). All values were estimated as normal or abnormal based on the normal values indicated in the parameters (Fig. 4) and previously published [23, 24].

Fig. 4. Cervical Vestibular Evoked Myogenic Potential (cVEMP) in three patients (III.19, III.4 and III.14).

Increased cVEMP latency is noted at 1000 Hz in both ears of patient A (III.19) and reduced P1 amplitude in the right ear. At 500 Hz no waves were recorded in the right ear, increased latency and reduced P1 amplitude were observed in the left ear. Increased Interaural Asymmetry Ratio (IAR) greater than 36% (39%). In Patient B (III.4) cVEMP at 500 Hz showed absence of waves from the right ear, increased latency and reduced P1 amplitude in the left ear. Augmented IAR (100%). Patient C (III.14) showed good waveform morphology in both ears at both 1000 and 500 Hz. The IAR were in the normal ranges (22% at 1000 Hz) and (17% at 500 Hz). Note that these frequencies were tested at 105 and 110 dB of stimulus intensities and normal values are indicated in the parameters table (Lat Latency, Amp Amplitude, P1 Marking of the first positive peak, N1 Marking of the first negative peak, NA Data not available, IAR interaural asymmetry ratio).

The above presented data suggests a defect in the vestibular system involving the lateral semicircular canals in the oldest patient and the otolithic organs and/or vestibular ganglion neurons (VGN) of both older patients tested.

Genetic studies

Assuming an autosomal dominant mode of disease inheritance in the family, ES was performed in patients III.13, III.14 and III.19. Around 98.5% of the exons were captured and fully covered at a minimum of 20x sequencing depth. Complete exomes were analyzed, encompassing all variants shared by the three patients who underwent ES. Analysis did not reveal any pathogenic variant involved in the patients’ progressive HI, nor in genes predisposing them to higher risk of cancer or malignancy. To avoid missing any pathogenic variant involved in the disease of the patients, ES and coverage data were then reassessed for a panel of 420 genes implicated in syndromic and non-syndromic HI (Supplementary Table 1) and a panel of 105 genes involved in cancer predisposition (Supplementary Table 2). Accordingly, all panel genes were well covered (99% at a minimum of 20x) and no pathogenic point variations in genes involved in similar diseases were identified in any of these patients. Furthermore, analysis of the genes involved in HI did not show any variant of uncertain significance shared between the affected individuals. Sanger sequencing revealed that the variant p.Ile10Thr (c.29T > C) in POLD3 gene (NM_006591.3) is heterozygous in the following patients: II.3, III.4, III.7, III.9, III.10, III.12, III.13, III.14, III.19, III.22, III.24 and IV.4, and homozygous in patient IV.6.

Pold3 gene expression

To validate the expression of Pold3 in both cochlear and vestibular tissues we have used RNA-scope technology consistent of an in situ hybridization approach with low background signal, that allows us to visualize and/or co-localize two or more probes [25]. We designed and tested probes against murine Pold3 (green) and Cacna1a (red) genes. We have used Cacna1a probe in parallel to Pold3 as a marker for subset of spiral ganglion neurons (SGN) as we have previously shown [26]. However, Pold3 expression, is not restricted to only Cacna1a positive cells.

Our results suggest expression of Pold3 in both SGN and vestibular ganglion neurons (VGN) in addition to its presence in the surrounding supporting cells (Fig. 5).

Fig. 5. RNA-scope of Cacna1a and Pold3.

A Upper panel is a representative image from RNAscope in situ hybridizations of Cacn1a1 (red) and Pold3 (green) at postnatal day 6 from cryopreserved of mouse cochlear/partial vestibular tissue. A Lower panel is a representative image of negative control probes provided by the ACD RNAscope company. Similar observation of the negative control probe in the VGN area (not shown). A White boundaries indicate spiral ganglion neuron (SGN) and vestibular ganglion neuron (VGN) cells and surrounding cells that might be neurons or supporting cells. B White dotted lines outlining single-neuron cells. Green arrows indicate cells expressing Pold3, and red arrows indicate cells expressing Cacna1a. DAPI was used as a counterstain to detect nuclei.

Discussion

This study describes the identification of a novel form of non-syndromic sensorineural adult-onset hearing and vestibular impairments associated with a heterozygous pathogenic variant in POLD3. The variant p.Ile10Thr in POLD3 was first identified at a homozygous state in a patient (IV.6) affected with a recessive form of congenital syndromic deafness combined with immunodeficiency and neurodevelopmental delay [19]. A thorough investigation of all members of this family showed correlation between the presence of this variant at a heterozygous state and the occurrence of autosomal dominant (AD) progressive HI in all carriers aged above 40 years. In fact, among carriers of the variant, 67% experienced hearing impairment, while 33% (aged less than 40) had normal hearing. As for non-carriers, 100% had normal hearing. This association was statistically significant with a p-value of 0.03 ( <0.05) based on the IBM’s Statistical Package for the Social Sciences (SPSS). Further assessment of the role of this variant in the maintenance of other parts of the inner ear, namely the vestibular system, suggests age-related dysfunction of the vestibular systems including the semicircular canal (cristae) and the saccular system (macula) and/or the vestibular afferent ganglion neurons.

Our previous work had shown that p.Ile10Thr is a loss of function variant leading to complete absence of the POLD3 protein expression in the homozygous proband, and to haploinsufficiency in the carrier parents [19]. This study reports haploinsufficiency in POLD3 as being involved in non-syndromic progressive adult-onset hearing and vestibular impairments.

POLD3 is a component of the DNA polymerase δ (POLD) complex that is composed of the catalytic subunit POLD1 and three accessory subunits, POLD2, POLD3, and POLD4 [27]. It is a vital enzyme complex essential for high-fidelity chromosomal DNA replication [28]. Interestingly, hearing impairment has been reported in patients with mutations in POLD genes. Indeed, biallellic mutations in POLD1 variants were linked to either autosomal recessive non-syndromic sensorineural hearing loss [29], or syndromic forms of deafness [28, 30, 31]. This underscores the significance of POLD in maintaining the integrity of the hearing system. In addition to POLD subunits, the connection between DNA stability, replication, and repair, and progressive HI has been well documented in both humans and mice [32, 33].

The association between adult-onset hearing and vestibular impairments observed in the patients presented in this study is not surprising as several genome wide association studies in ARHI have identified candidate genes that are also expressed in the mouse vestibular system, suggesting a potential connection between the genetic factors of ARHI and age-related vestibular impairment (ARVI) [34–37]. For instance, older patients with AD progressive HI due to WSF1 mutations experience dizziness attributed to mild vestibular dysfunction [14, 15]. In the current report, this link is further supported by the gene expression studies herein performed (Fig. 5), revealing the expression of Pold3 in the auditory/cochlear and vestibular neurons and their supporting cells. Our expression studies also showed that pold3 is expressed in the Greater Epithelial Ridge (GER) cells and other supporting cells (data not shown). Similarly, studies including single cell transcriptomic [38], SHIELD database (https://shield.hms.harvard.edu/) and gEAR portal (https://umgear.org/), revealed the expression of Pold3 in cochlear and vestibular cells including the GER Cells and the auditory and vestibular neurons.

While ARHI and ARVI are complex disorders influenced by gene-environment interactions, ARHI is well studied but much less is known about ARVI. Studies on twins, siblings, and GWAS on familial progressive ARHI have revealed a highly significant genetic predisposition [5, 39, 40]. Single gene variants may contribute to ARHI. Interestingly, the majority of the genes involved in dominantly-inherited non-syndromic deafness [https://hereditaryhearingloss.org/], show progressive hearing loss, many with adult-onset.

Progressive hearing loss phenotype typically precedes the age-related decline in the vestibular function, a clinical aspect that has been under-diagnosed and received limited research focus [22]. Indeed, approximately 20–30% of normal individuals over the age of 65 experience various forms of dizziness, and around 40–50% of those aged 85 and above complaining of vertigo [41, 42]. Therefore, the absence of vestibular dysfunction in young patients carrying POLD3 variants and exhibiting adult-onset HI at approximately 40 years old, is not surprising.

Vestibular caloric responses performed in the older affected patient (III.19) showed a unilateral weakness to the right correlating with his more severe hearing phenotype in the same ear. This patient had an abnormal latency and amplitude of the cervical Vestibular Evoked Myogenic Potential (cVEMP) waves and augmented Interaural Asymmetry Ratio (IAR) (39%). However, the video head impulse test (vHIT) test was normal in this patient. This observation is not uncommon. Indeed, certain studies have indicated that vHIT outcomes tend to appear normal when the unilateral caloric weakness measures less than 40% [43]. It is also not uncommon to observe a normal caloric tests in patients with augmented cVEMP [23]. A similar but unilateral abnormal cVEMP waves were observed in the second older patient (III.4). An IAR value of 100% were estimated at 500 Hz for both older patients (III.4 and III.19). However, this patient exhibited normal caloric tests. Notably, vestibular tests conducted in the 10-year-younger patient (III.14) were all normal. This may suggest that the vestibular phenotype is progressive in this family and that the younger carriers family members are all at risk of developing vestibular and balance abnormalities as they age.

In parallel, two individuals (III.13 and III.18) in the family reported herein, heterozygous for the p.Ile10Thr, presented with different types of cancer at a young age. Exome Sequencing in both patients did not reveal the presence of any pathogenic variant in known genes that may predispose them to a higher risk of cancer. The potential link between haploinsufficiency in POLD3 and cancer predisposition is supported by previous studies on the role of POLD subunits and other DNA repair mechanisms in cancer. In fact, beyond DNA replication, POLD has been involved in DNA repair and recovery and genomic integrity and stability [27, 28, 44–46]. All POLD subunits have been shown to be overexpressed in human cancers [44, 45, 47]. Moreover, heterozygous germline mutations in the exonuclease proofreading domains of POLD1 are associated with colorectal cancer and cancers of the pancreas, prostate and small intestine [28, 48]. Therefore, reporting further cases with POLD3 mutations is essential to confirm a possible link between this gene and cancer.

Last but not least, given the worldwide rise in aging population, increasing awareness of age-related sensory alterations is imperative. HI not only contributes to social isolation and reduced confidence but also elevates the risk of dementia and falls among older individuals [49, 50]. Knowledge of these conditions and the identification of the causative genetic variants could lead to early diagnosis and potentially prevent fatalities in this age group.

Altogether, this study reports for the first time POLD3 haploinsufficiency as being responsible for adult-onset HI and VI, and as possibly involved in cancer predisposition. This suggests the emergence of a new category of HI-associated genes involved in DNA replication and repair. Reporting further cases with variants in POLD components is essential to better delineate the diseases spectrum and future mechanistic studies in mice models is crucial to further understand the mechanism of these diseases, which may allow the application of new therapeutic approaches.

Author contributions

EC, CM, MM and AM conceived, designed the study, performed data interpretation, and wrote the manuscript. SC an RK performed DNA experiments and NGS analysis. PZ and CG performed gene expression analysis. RS, RB performed clinical and physical examinations of the patients. MY, MB, AKAY, and FA performed the audiological evaluation. AH and JNE performed the literature review and helped in writing the manuscript. All authors have read and approved the manuscript.

Funding

This study was funded by the President Intramural Research Fund (PIRF) of the Lebanese American University (PIRF- I0004). Authors did not receive any personal funding. Medical Research Council: MR/X012077/1 (LMF). Medical Research Council: MR/S002510/1 (MMustapha). Imaging work was performed at the Wolfson Light Microscopy Facility, using the Zeiss LSM 980 Airyscan 2 confocal microscope (MRC grant MR/X012077/1).

Data availability

The datasets used and analyzed during the current study are available from the corresponding author upon a reasonable request.

Competing interests

The authors declare no competing interests.

Ethics approval and consent

Approval to conduct the study was obtained from the Institutional Review Board of the Lebanese American University, Beirut, Lebanon (IRB #:LAU.SOM.EC1.2020.R3.1/Feb/2024). All family members signed an informed consent for participation and sample collection.

Consent to publish

All family members signed an informed consent for data publication.

Supplementary information

The online version contains supplementary material available at 10.1038/s41431-024-01715-7.