← All decks
ENT
Genetics in Otology and Neurotology
Built from Haynes-Glasscock-Shambaugh Surgery of the Ear, Seventh Edition
What’s inside
7 sections · 118 slides
Overview
- Genes and proteins that make hearing possible Dominant, recessive, X-linked, mitochondrial Common non-syndromic genes: GJB2, SLC26A4 Syndromic hearing loss and its red flags Genetic testing and result interpretation NF2-related schwannomatosis and paraganglioma
Foundations of human hearing genetics
- Altered gene expression in sound-detecting cells Over 200 genes linked to hearing loss Number rises month by month Both syndromic and non-syndromic forms Hearing loss is genetically heterogeneous
- Current standard test for hearing loss Panels of about 150 to 200 known genes Covers syndromic and non-syndromic genes Yield 40 to 60% in congenital or childhood onset Streamlines query of causative variants
- Pathogenic or likely pathogenic: cause disease Benign or likely benign: do not cause disease Variant of unknown significance: neither Classification drives clinical action Reclassification over time is expected
- 'Mutation' now used less and less often Term lacks specificity for clinicians Carries a negative connotation for patients Use 'pathogenic variant' instead Or 'likely pathogenic variant'
- Hair cells must survive and make energy Plus specialised roles unique to the cochlea Sound detection, ion transport, structure Loss of one protein can kill the cell Surrounding cells carry the same dependence
- Stereocilia and mechanotransduction
- Mechanotransduction sequence
- Haploid set of 23 chromosomes from each parent Diploid complement of 46 in every nucleus 44 autosomes plus two sex chromosomes Two alleles of nearly every gene per cell Exception: X and Y genes in a male
- Central dogma of molecular biology
- Insertions and deletions of nucleotides Splicing abnormalities of the mRNA Nonsense variant: an early stop codon Missense variant: amino acid substitution Copy number variant: large coding deletions Any class may be benign or pathogenic
- One normal copy is enough for the cell Both alleles must carry a pathogenic variant Term for this is bi-allelic Carriers have one normal copy and are unaffected One variant from each parent
- One pathogenic variant harms the cell Loss of one functional copy is enough Some proteins tolerate no reduction Same gene may act dominantly or recessively Depends on the type of variant present
- Dominant versus recessive at the cell
Hearing genetics: the big picture
- Causes of deafness and hard of hearing
- Over 60% of deaf children: identifiable genetic cause 70% of affected individuals: non-syndromic 30% of affected individuals: syndromic Over 400 forms of syndromic hearing loss All present first with hearing loss
- Named syndromes and their frequency
- Syndromic cause is hard to detect a priori Retinal or renal signs may lag years behind Testing may be the only early warning Identifies risk before other symptoms begin Changes habilitation planning in childhood
- Most congenital hearing loss genes are recessive One pathogenic variant from each parent Typical phenotype: congenital bilateral SNHL Highest prevalence in profound congenital loss Few exceptions to this phenotype
- Curated non-syndromic hearing loss genes
- ClinGen expert panels curate gene validity Hereditary Hearing Loss Homepage lists loci Lists are dynamic, not fixed MYO1A and TSPEAR now disputed genes Panels should exclude disputed genes
- Seen previously in other affected individuals Low prevalence in the general population Measurable impact in laboratory assays Impact on gene function in animal models Evidence is cumulative, not single-source
- ClinVar: general clinical variant archive ClinGen: curated gene and variant evidence Deafness Variation Database: ear-specific All list benign to pathogenic tiers Hearing-specific interpretation guidelines exist
- Autosomal dominant hearing loss
- Most dominant hearing loss is high frequency WFS1-related loss is typically low frequency Finding may vary with the specific variant Audiogram shape can point to the gene Exception worth remembering in clinic
- Gene lies on the X chromosome Males have a single X: fully penetrant Typically more severe and earlier onset No known hearing loss genes on the Y Accounts for 1-2% of non-syndromic cases
- One X inactivated per cell to balance dosage Named Lyonisation after Mary Lyon Females can still be affected Onset usually later, hearing loss milder Skewed inactivation worsens the phenotype
- Mitochondrial inheritance
- First reported in the early 1990s Maternally inherited familial deafness Linked to aminoglycoside antibiotic exposure Not driven by peak or trough levels Can follow a single exposure within days
- MT-RNR1 encodes the 12S ribosomal RNA m.1555A>G and m.1494C>T best described Both confer aminoglycoside sensitivity Loss is bilateral profound SNHL within days MT-TS1 m.7445>C/T/G: no aminoglycoside link
- Mitochondrial syndromes with deafness
- Type 2 diabetes and hypothyroidism Renal dysfunction Early onset stroke Myopathy and neuropathy Cardiomyopathy
- Up to 3.5% of non-syndromic loss in one cohort 2434 probands, North American repository Samples collected between 1997 and 2009 Predates widespread targeted panel testing Other causes in these patients not excluded
- DFN prefix denotes a deafness locus DFNA: autosomal dominant DFNB: autosomal recessive DFNX: X-linked Numbered by order of historical discovery DFNB1 is GJB2 and GJB6 on chromosome 13
Common non-syndromic genes
- Most common hearing loss genes worldwide Up to 25% of deaf or hard of hearing in N America Up to 50% in Europe and the Mediterranean Similar figures in East Asia High carrier frequency in these populations
- c.35delG: Europe and North America Carried by 4% of hearing individuals there c.235delC: up to 2% of East Asian heritage c.167delT: up to 3% of Ashkenazi heritage p.V37I: up to 10% of individuals in East Asia
- Nucleotide deletions cause a frameshift Frameshift creates a premature stop codon Result is loss of the connexin 26 protein High carrier frequency drives disease prevalence Explains why GJB2 is seen so often
- Large deletions common in Europe and Middle East GJB6 sits adjacent to GJB2 on chromosome 13 Deletion abolishes GJB2 and GJB6 expression Acts in trans with a GJB2 variant Detected by deletion analysis, not sequencing
- Connexins in the cochlea
- Two premature stop codons: severe to profound Two missense variants: mild to moderate One missense plus one truncating: mild to severe GJB2 frameshift with GJB6 deletion: profound Variability still observed within genotypes
- Once believed to be non-progressive Two longitudinal studies show mild progression Non-syndromic: no other health problems Temporal bone anomalies are rare Follow-up audiometry still warranted
- A small number of individuals affected Bilateral SNHL from mild to profound Usually syndromic with hyperkeratotic skin Vohwinkel, Bart-Pumphrey, KID syndrome Palmoplantar keratoderma, Clouston syndrome Mechanism is gain of function, not loss
- Chloride, bicarbonate and iodine transporter Expressed in cochlear supporting cells Also expressed in thyroid and kidney Recessive variants cause hearing loss with EVA Same gene underlies Pendred syndrome
- Second-most common identified genetic cause Bi-allelic variants in up to 40% with EVA Applies to patients imaged for hearing loss Unilateral EVA with unilateral loss also linked Temporal bone imaging guides the request
- Severity ranges from mild to profound Progression is typical, sometimes rapid Range varies even within one family Little genotype-phenotype correlation Detected at birth or shortly thereafter
- Penetrance of thyroid disease is variable Some relatives develop goitre, others do not Traditionally expected after the second decade Thyroid insufficiency can occur at birth Infants may present with congenital hypothyroidism
- Tip links composed of PCDH15 and CDH23 Link runs shortest to tallest stereocilium Upper tip link density: MYO7A Plus harmonin (USH1C) and sans (USH1G) Apical surface packed with actin fibrils
- Tip link genes and their two phenotypes
- MYO7A: recessive DFNB2 and dominant DFNA11 CDH23: recessive DFNB12 PCDH15: recessive DFNB23 USH1C: recessive DFNB18A All also expressed in the retina
- Ambiguity remains, but correlations exist CDH23 missense with residual function: DFNB12 CDH23 absent protein: Usher syndrome MYO7A residual function: hearing loss only PCDH15 missense retaining function: DFNB23
- Full channel composition still unknown Two decades of research, incomplete answer TMC1, LHFPL5 and TMIE form part of it TMC1: recessive or dominant SNHL LHFPL5 and TMIE: recessive congenital SNHL Phenotype is progressive bilateral SNHL
Syndromic hearing loss
- Profound congenital bilateral SNHL Balance dysfunction from infancy Motor delay: not walking before 16 months Retinitis pigmentosa develops later Genes: MYO7A, CDH23, PCDH15, USH1C, USH1G
- Retinitis pigmentosa in Usher syndrome
- Mild to moderate congenital SNHL Hearing loss does not progress Vision loss begins in the second decade Presents as peripheral field difficulty Genes: USH2A, ADGRV1 and WHRN
- Most variable of the three types Later onset hearing loss Balance involvement may or may not occur Retinitis pigmentosa follows Clarin (CLR) was the only known gene Small molecule therapy heading to trial
- Profound congenital bilateral SNHL Prolonged corrected QT interval on ECG Risk of life-threatening tachyarrhythmia Autosomal recessive: KCNQ1 or KCNE1 May need drug and device management
- QT interval within one family
- 707 Californian children screened, no cases Same study found 3 with long QT of other cause Higher in Scandinavia and the Middle East Founder variants and consanguinity explain this Long QTc commoner in deaf than general population Screening ECG for deaf children still advised
- Proteinuria and haematuria Corneal defects Late childhood onset progressive SNHL Defective basement membranes in all three organs Collagen 4 subunits: COL4A5, COL4A3, COL4A4
- X-linked COL4A5 form is up to 80% of cases Autosomal forms from COL4A3 and COL4A4 Autosomal cases found more often with NGS Kidney disease presents before hearing loss X-linked form most severe in males Faster progression to end stage kidney disease
- Congenital unilateral or bilateral SNHL Severe to profound and non-progressive Pigment abnormalities of hair, skin, eyes White forelock or lightly pigmented skin Heterochromia iridis
- The four types of Waardenburg syndrome
- Penetrance of hearing loss is incomplete Up to 60% affected within type I families Hearing loss usually severe to profound Enlarged vestibular aqueduct may be seen Modiolar and internal auditory canal hypoplasia
- PAX3 dominant: types I and III Recessive PAX3 variants also cause type III MITF dominant: type II Type II has higher hearing loss penetrance EDNRB, EDN3, SNAI2 and SOX10: type IV
- SOX10 variants cause Waardenburg type IV May cause eighth nerve hypoplasia or absence Resolve this question by MRI Direct implication for cochlear implantation Image before counselling on implant outcome
- Most common cause of X-linked hearing loss Mixed hearing loss with stapes fixation Locus designated DFNX2 Imaging: bilateral dilated internal auditory canal Imaging occasionally reported as normal Males pre-lingual; females milder and later
- Surgery may provoke endolymphatic gusher Acute loss of hearing can follow Anticipation permits alternative strategies Plan the approach before opening the footplate Consider genetics before stapes surgery
- Affects the eighth nerve, cochlea works May appear non-syndromic Also seen in Friedreich ataxia And in Charcot-Marie-Tooth disease Four genes: OTOF, PJVK, DIAPH3 and GJB2
- Multiple centres reported success in 2024 Target was OTOF-related auditory neuropathy First gene therapy success in this field Makes the genetic diagnosis actionable Reinforces the case for early testing
Testing, interpretation and counselling
- Until 15 years ago: one gene at a time Genomic DNA from peripheral blood leukocytes In practice, GJB2 sequencing plus GJB6 deletion Diagnostic in only 25% in North America Gene discovery outpaced single-gene testing
- First panel came from the University of Iowa Sequence 80 to 200 known hearing loss genes Faster diagnosis demonstrated by many groups Flags risk of progression and systemic disease Limits further investigation and reduces cost
- Three kinds of genetic test result
- Recessive: two confirmed pathogenic variants Each variant inherited from a different parent Dominant: one pathogenic variant is enough Phenotype informs the interpretation Large families may need segregation testing De novo change possible with hearing parents
- No variants found on that panel Not the same as 'hearing loss is not genetic' Causative gene may not be on the panel Panels sequence coding regions mainly Regulatory regions poorly represented Consider genome sequencing if syndrome suspected
- One pathogenic variant in a recessive gene Carrier or a second undetected variant? Or variants of unknown significance only Evidence insufficient either way Additional testing and follow-up indicated
- Each variant must sit on a different allele Both on one copy is not a recessive condition Sequencing cannot resolve phase directly Exception: variants within tens of base pairs Laboratories request parental samples
- Segregation analysis means setting phase One available parent may be sufficient Both variants from one parent rules them out Example: two MYO15A variants, both paternal Mother carried neither: cause excluded
- Commonly reported, often several at once Also abbreviated VUS or VOUS Cannot be called pathogenic or benign Test other affected relatives if available Segregation with affected only raises suspicion
- gnomAD gives frequency in unaffected people Common variants are unlikely to be pathogenic But some deafness variants are common GJB2 variants reach high population frequency ClinGen sets hearing-specific parameters
- SIFT and PolyPhen predict protein impact Mutation Taster and REVEL also used SpliceAI predicts mRNA splicing effects Programs may give conflicting results Conservation and gene tolerance are adjuncts
- Protein structural modelling Functional modelling in cell systems Functional modelling in animal models Used when clinical evidence is insufficient Can reclassify a variant of unknown significance
- Counselling is integral, not optional Pre-test and post-test sessions ideally Explain the three possible result types State that negative does not exclude genetics Set expectations before samples are drawn
- Content tailored to the result obtained Recessive result: 1 in 4 risk each pregnancy Negative or ambiguous: arrange re-evaluation Interpretation may change over time Newer testing may become available
Hereditary tumour predisposition
- Genetically distinct from neurofibromatosis type 1 Both thought identical until the mid-1980s Separate clinical criteria set in the late 1980s Linked to chromosome 22 NF2 gene encodes the protein merlin
- Vestibular schwannoma on MRI
- Cranial and spinal meningioma Schwannomas of other cranial and spinal nerves Spinal cord ependymoma, peripheral nerve tumour Subcutaneous tumours and skin plaques Cortical wedge cataract, thickened optic nerves Retinal hamartoma and epiretinal membrane
- Prevalence about 1 in 25,000 in UK studies Onset from birth through old age Median age of onset around 60 years Penetrance approaches 100% Wide age range defies a single screening age
- Derived from the Manchester diagnostic criteria Route 1: bilateral vestibular schwannomas Route 2: identical NF2 variant in two tumours Those tumours must be anatomically distinct Meningiomas or ependymomas qualify Route 3: major and minor criteria combination
- Two major, or one major plus two minor Unilateral vestibular schwannoma First-degree relative, not a sibling, with NF2 Two or more meningiomas NF2 variant in unaffected tissue such as blood Allele fraction well under 50% means mosaic
- More than one of a type may be counted Two distinct schwannomas count as two minors Ependymoma, meningioma or schwannoma If major is unilateral VS, one must be dermal Juvenile subcapsular or cortical cataract Retinal hamartoma; epiretinal membrane under 40
- Sequence the NF2 gene Add deletion and duplication analysis Missense variants and copy number variants occur Consider ring 22q if developmental disability Relax Manchester criteria in the young Any schwannoma in a child warrants testing
- 155 isolated schwannoma patients aged under 25 30 of 155 had germline NF2 variants 7 had germline SMARCB1 variants 10 had pathogenic LZTR1 variants Vestibular schwannomas do occur in both Test all three genes if presenting under 30
- 25 to 50% of individuals are mosaic Variant arose after conception Present only in a subset of cells Blood studies are often negative Tumour shows two NF2 hits Test blood and tumour together
- Standard NGS depth of read is about 30x That depth misses low-level mosaicism Mosaic testing needs depth in the thousands Choose the laboratory accordingly A negative shallow test is not an answer
- Hereditary tumour predisposition condition Arises from neural crest paravertebral tissue Paraganglioma: skull base to pelvis Phaeochromocytoma: within the adrenal medulla Most benign; malignancy is genotype-specific
- Paraganglioma histology
- Skull base and neck tumours: parasympathetic They typically do not secrete catecholamines Thoracic, lumbar, sacral paraspinal: sympathetic These typically do secrete catecholamines Adrenal medullary tumours secrete as well
- SDHD was the first gene identified Then SDHA, SDHAF2, SDHB and SDHC Collectively known as SDHx Subunits of succinate dehydrogenase Over 40% carry a germline driver variant Highest heritability of any human tumour
- 2017 consensus from the NGS in PPGL group Published in Nature Reviews Endocrinology Lists 27 genes associated with the condition Panel adds FH, MAX, NF1, RET, TMEM127, VHL Also EPAS1, KIF1B, MET, MDH2 and others Nearly all inherited as autosomal dominant
- Reported for pathogenic variants in SDHD Paternal transmission: offspring affected Maternal transmission: carrier but unaffected Explains skipped generations in pedigrees This observation has since been questioned
- SDHB variants carry much higher malignancy risk Resect even small paragangliomas in SDHB carriers Rationale is metastatic potential Mosaicism has been reported in this syndrome If blood negative, test tumour-derived DNA
- Genetic testing informs medical management Applies to hearing loss and head and neck tumours Conditions once simple are highly heterogeneous We can sequence faster than we can interpret Establishing pathogenicity is the limiting step
- Clinical geneticists and genetic counsellors Otologists and neurotologists Shared testing and interpretation pathway Informs management and decision-making Re-evaluate ambiguous results over time
- 60% of childhood deafness has a genetic cause Panels diagnose 40-60% of congenital cases GJB2 and SLC26A4 top the non-syndromic list Every deaf child needs an ECG and a retina plan POU3F4 and SOX10 change surgical decisions Test blood and tumour when NF2 is suspected
- References
- References (continued)
- References (continued)
- References (continued)
- References (continued)
- References (continued)
- Haynes-Glasscock-Shambaugh Surgery of the Ear, Seventh Edition
- Join our WhatsApp groups