Risk of sudden cardiac death in EXOSC5-related disease

Abstract

The RNA exosome is a multi-subunit complex involved in the processing, degradation, and regulated turnover of RNA. Several subunits are linked to Mendelian disorders, including pontocerebellar hypoplasia (EXOSC3, MIM #614678; EXOSC8, MIM #616081: and EXOSC9, MIM #618065) and short stature, hearing loss, retinitis pigmentosa, and distinctive facies (EXOSC2, MIM #617763). More recently, EXOSC5 (MIM *606492) was found to underlie an autosomal recessive neurodevelopmental disorder characterized by developmental delay, hypotonia, cerebellar abnormalities, and dysmorphic facies. An unusual feature of EXOSC5-related disease is the occurrence of complete heart block requiring a pacemaker in a subset of affected individuals. Here, we provide a detailed clinical and molecular characterization of two siblings with microcephaly, developmental delay, cerebellar volume loss, hypomyelination, with cardiac conduction and rhythm abnormalities including sinus node dysfunction, intraventricular conduction delay, atrioventricular block and ventricular tachycardia (VT) due to compound heterozygous variants in EXOSC5: 1) NM_020158.4:c.341C>T (p.Thr114Ile; pathogenic, previously reported) and 2) NM_020158.4:c.302C>A (p.Thr101Lys; novel variant). Review of the literature revealed an additional family with biallelic EXOSC5 variants and cardiac conduction abnormalities. These clinical and molecular data provide compelling evidence that cardiac conduction abnormalities and arrhythmias are part of the EXOSC5-related disease spectrum and argue for proactive screening due to potential risk of sudden cardiac death.

Introduction

The exosome is a multi-subunit complex which plays an essential role in RNA regulation via mRNA turnover, rRNA processing, degradation of unstable RNAs, and nonsense-mediated decay (Kilchert et al., 2016). The exosome complex is encoded by the EXOSC family of genes and consists of a barrel-shaped core (EXOSC4, EXOSC5, EXOSC6, EXOSC7, EXOSC8 and EXOSC9) with a three-subunit cap (EXOSC1, EXOSC2, and EXOSC3). RNA molecules are delivered to the complex by exosome specificity factors where they are unwound by helicases and threaded through the exosome’s central channel in a 3’ to 5’ direction. Exonucleases (e.g. EXOSC10, DIS3, DIS3L) involved in processing RNA are bound to the catalytically inactive core and cap subunits of the exosome (Kilchert et al., 2016).

Pathogenic variants in the exosome complex result in Mendelian disorders. Biallelic variants in EXOSC3, EXOSC8, and EXOSC9 (MIM# 614678, 616081, and 618065, respectively) cause cerebellar volume loss, motor neuron degeneration, and variable extra-central nervous system (CNS) manifestations including short stature, facial dysmorphology, ocular abnormalities, sensorineural hearing loss, and skeletal anomalies (Fasken et al., 2020). Additionally, biallelic variants in EXOSC2 are associated with retinitis pigmentosa, progressive hearing loss, premature aging, short stature, intellectual disability, and facial dysmorphology (MIM# 617763, Fasken et al., 2020). More recently, EXOSC5 was identified as a cause of developmental delay, cerebellar abnormalities, short stature, and hypotonia (MIM *606492, Slavotinek et al., 2020). A unique feature of EXOSC5-related disease is the occurrence of bradycardia or complete heart block requiring a pacemaker in four of the eight previously reported affected individuals (Slavotinek et al., 2020; Beheshtian et al., 2019). Conduction abnormalities and arrhythmias were not previously reported in EXOSC-related disease; it is unclear if these are a common trait that should be proactively screened and followed in affected individuals.

Here we report two siblings with developmental delay, microcephaly, cerebellar volume loss, hypomyelination, cardiac conduction abnormalities and ventricular tachycardia due to compound heterozygosity for a known pathogenic EXOSC5 variant c.341C>T (p.Thr114Ile) and novel variant c.302C>A (p.Thr101Lys) (Slavotinek et al., 2020). The presence of sinus node dysfunction, intraventricular conduction delay, atrioventricular (AV) block, and ventricular tachycardia confirms and expands the arrhythmias observed in EXOSC5-related disease.

Clinical report

The proband (P1) is a 10-year-old female of Mexican descent born to unrelated parents from a small town (estimated population 3,000-4,000) (Fig. 1a,b). She has a 20-year-old affected brother (BAB14925) and three unaffected siblings. Common features between the two siblings include developmental delay, microcephaly, cerebellar volume loss, hypomyelination, myopia, astigmatism, esotropia, and tapering fingers. Distinct features of the proband include dysmetria, hypotonia, and mild retrognathia (Fig. 1b). At 10 years of age, her growth parameters and Z-scores using Centers for Disease Control data were length:135 cm (Z=−0.45) and weight: 44.9 kg (Z=1.4). Her fronto-occipital circumference (FOC) was 48.0 cm (Z=−2.77). Due to the patient’s age, FOC Z-scores were calculated using Nellhaus data (Weaver et al., 1980). She has mild intellectual disability (ID). The older brother (P2) is more severely affected; he is nonverbal and unable to walk. He has profound ID. Other distinctive features of the older sibling include optic atrophy, short stature, dysmorphic facies, camptodactyly, clindodactyly, scoliosis, spasticity, orolingual dyskinesia, and refractory myoclonic epilepsy (Fig. 1c).

Figure 1. Biallelic variants in EXOSC5 cause dysmorphology and abnormal brain findings.

(a) Family pedigree, genetic variants, and segregation analysis of biallelic EXOSC5 variants.

(b) Pictures of proband P1 showing 1. strabismus, 2. mild retrognathia and microcephaly, 3. absence of scoliosis, 4 and 5. grossly normal appearance of hands with tapering of the fingers, 6 and 7. normal feet.

(c) Pictures of P2 showing several dysmorphic features including 1 and 2. microcephaly, sloping forehead, proportionally large ears, high nasal bridge, prominent nose with deviated septum, large mouth with thick lips, large tongue, misalignment of teeth, prognathism, 3. scoliosis, 4 and 5. severe multidigit camptodactyly of bilateral hands with knuckle pads on proximal and distal interphalangeal joints, 6 and 7. outward deviated toes with nodules of metatarsophalangeal joints.

(d) and (e) Brain MRI of proband P1 at age 2 and 4 years, respectively, showing microcephaly, cerebellar atrophy, and hypomyelination. 1. Sagittal T1, 2. Coronal T2, 3. Axial T2, and 4. Axial T2/FLAIR.

(f) Brain MRI of P2 at age 10 years, showing microcephaly, cerebellar atrophy, and white matter signal abnormalities. 1. sagittal T1, 2. coronal T2, 3. axial T2, and 4. axial T2/FLAIR.

(g) Brain MRI of P2 at age 20 years, showing microcephaly, progressive cerebellar and cerebral atrophy, thinning of the corpus callosum, and white matter signal abnormalities. 1. Sagittal T1, 2. Coronal T2, 3. Axial T2, and 4. Axial T2/FLAIR.

To determine an etiology for the microcephaly and developmental delay, brain magnetic resonance imaging (MRI) was performed for the proband at ages 2 and 4 years old (Fig. 1d, e) and the older brother at 10 and 20 years of age (Fig. 1f, g). Both affected siblings have cerebellar atrophy and the white matter was incompletely myelinated for age. Repeat imaging of the proband after 2 years demonstrated interval progressive volume loss of the cerebellum and minimal progression in myelination consistent with hypomyelination and potentially neurodegeneration (Fig. 1e). Repeat imaging of the older brother showed severe thinning of the corpus callosum and interval progressive cerebellar and cerebral volume loss consistent with neurodegeneration (Fig. 1g).

The proband was referred to cardiology due to ventricular bigeminy noted during anesthesia induction for surgical repair of esotropia. ECG showed a nonspecific intraventricular conduction delay, left axis deviation, junctional rhythm, and high-grade AV block (Fig. 2a). Twenty-four hour Holter monitoring demonstrated: 1) junctional rhythm with relative bradycardia for age (heart rate ranged 37-98bpm, average 51bpm), intermittent sinus with first degree block and variable AV conduction concerning for second degree, Type I and Type II AV block, and 3) monomorphic ventricular tachycardia ranging from 120-150 bpm (Fig. 2b). Evaluation of the older brother showed more advanced conduction disease with trifasicular block on ECG and similar high grade AV block and ventricular tachycardia on Holter monitoring (Fig. 2c, d). Echocardiogram was normal in both siblings (Fig. 2e). Risk of sudden cardiac death from heart block and ventricular tachycardia and management options were discussed and dual chamber implantable cardioverter defibrillators (ICD) were implanted in both siblings. During ICD implantation and while under anesthesia, both siblings demonstrated intermittent complete heart block and profound sinus node dysfunction.

Figure 2. Cardiac conduction and rhythm abnormalities in patients with biallelic EXOSC5 variants.

(a) Electrocardiogram (ECG) of proband P1 showing junctional bradycardia with intermittent sinus capture beats, first degree AV block, and nonspecific intraventricular conduction delay.

(b) Holter monitor/Event recorder findings of P1 showing 1. Sinus rhythm with Mobitz Type I Wenckebach, 2. High grade AV block, 3 and 4. Junctional rhythm and ventricular tachycardia.

(c) ECG of P2 showing sinus rhythm with first and second degree (Mobitz type I) AV block, complete right bundle branch block and left axis deviation (trifasicular block).

(d) Holter monitor findings of P2 showing 1. 2:1 AV block, 2. 1st degree block, Mobitz Type I AV block, 3. Junctional rhythm with pause, 4. Ventricular tachycardia.

Molecular analysis

The family provided written informed consent including permission for publication of photographs under Institutional Review Board (IRB)-approved research protocol (H-29697). Proband exome sequencing (ES) was performed at Baylor Genetics (Houston, TX). The resulting FASTQ file was transferred to the Baylor-Hopkins Center for Mendelian Genomics for reanalysis through an in-house pipeline as previously described (Pehlivan et al., 2019). Two variants in EXOSC5 were identified in the proband. The first variant [Chr19:41897789:G>A (hg19); NM_020158.4; c.341C>T, p.Thr114Ile] is a known pathogenic variant (Beheshtian et al., 2019; Slavotinek et al., 2020). It is ultra-rare (defined as a MAF<1/10,000; 17 heterozygotes in gnomAD v2.1.1) and well-conserved (Fig. 3a) (Hansen et al., 2019). The second variant [Chr19:41897828:G>T (hg19); NM_020158.4; c.302C>A, p.Thr101Lys] is ultra-rare (1 heterozygote in our database of >12,000 exomes, 8 heterozygotes in gnomAD v2.1.1) and well-conserved (Fig. 3a). It is predicted damaging by PolyPhen, MutationTaster, Sorting Intolerant From Tolerant (SIFT) and Likelihood Ratio Test (LRT) and has a Combined Annotation Dependent Depletion (CADD) score (v1.6) of 27.5 (Adzhubei et al, 2010; Chun et al., 2009; Schwarz et al., 2014; Sim et al., 2012; Rentzsch et al., 2021). EXOSC5 occurs on chromosome 19 and consists of six exons (Fig. 3b). Both variants are found in the third exon. Segregation analysis and variant validation was performed by Sanger sequencing which, in accordance with Mendelian expectations, confirmed the compound heterozygous state of the two affected individuals in trans with inheritance of each variant from one parent (Fig. 1a). As anticipated by the hypothesized model tested, all unaffected siblings were either homozygous wild-type or heterozygotes.

Figure 3. Amino acid residue conservation and location of EXOSC5 variants.

(a) Both EXOSC5 variants present in P1 and P2 are fully conserved across species. Conservation data obtained from Vertebrate Multiz Alignment & Conservation (46 Species) in UCSC Genome Browser (http://genome.ucsc.edu).

(b) Structure of EXOSC5 mRNA and location of EXOSC5 variants present in P1 and P2 (red) and previously reported variants (black). The EXOSC5 variant observed in all patients with cardiac conduction abnormalities and/or arrhythmias, p.Thr114Ile, is indicated with a heart symbol.

Discussion

Within the initial EXOSC5 cohort (Slavotinek et al., 2020), one patient developed syncopal episodes at two years of age and was diagnosed with complete heart block requiring pacemaker placement. This individual is a compound heterozygote with a 1023 bp deletion involving exons 5-6 of EXOSC5 in addition to c.341C>T; p.Thr114Ile. EXOSC5 was independently reported as a candidate disease gene in a consanguineous Iranian family with three affected siblings homozygous for c.341C>T; p.Thre114Ile with developmental delay and adolescent onset right bundle branch block and bradycardia requiring pacemakers (Beheshtian et al., 2019). Therefore, six of the ten individuals with EXOSC5-related disease reported to date have some form of cardiac conduction abnormalities that have been clinically recognized (Suppl. Table 1). Curiously, these six individuals all share the p.Thr114Ile variant. The absence of recognition of cardiac disease in the remainder of the patients reported by Slavotinek et al. may reflect the young age of the cohort (three patients ≤14 months of age), as cardiac conduction abnormalities were not detected until early childhood or adolescence. Alternatively, this trait may exhibit reduced penetrance or be specific to the p.Thr114Ile variant. As the previously studied pathogenic EXOSC5 missense variants are hypomorphic alleles (Slatovinek et al., 2020), it is also possible the manifestation of cardiac disease may be a function of residual exosome complex activity resulting from the combination of patient’s alleles, perhaps with lower activity levels resulting in arrhythmias or conduction abnormalities. Better assays of exosome complex activity are needed to test the latter hypothesis. Regardless, these EXOSC5 variant alleles may place patients at risk of sudden cardiac death, warranting clinical evaluation, long-term follow-up and consideration of a pacemaker or defibrillator among affected individuals with evidence of conduction disease or arrhythmias. Individuals with EXOSC5-related disease may benefit from monitoring for cardiac conduction defects and ventricular arrhythmias. Selection pressure resulting from increased risk of sudden cardiac death may explain the observed young age of all identified patients. Considering the significant phenotypic overlap among known exosome disease genes, it may be prudent to perform cardiac evaluation in individuals with pathogenic variants in ‘exosome genes’ other than EXOSC5. Supporting this, AV block was described in a 21-year-old male with EXOSC3-related pontocerebellar hypoplasia type 1D, and sudden unexplained death was reported in a 10-year-old male with EXOSC9-related pontocerebellar hypoplasia type 1D (Le Duc et al., 2020; Sakamoto et al., 2021).

While the concurrence of neurodevelopmental disabilities and congenital heart defects is frequently seen in Mendelian disorders (Homsy et al., 2015), the combination of cardiac conduction abnormalities and arrhythmias, developmental delay, hypomyelination, and cerebellar abnormalities is unusual. On initial review of the literature, the only potential match was Zaki-Gleeson syndrome (MIM# 614407). Described in a single consanguineous family, key features include developmental delay, microcephaly, dysmorphic facies, camptodactyly, crowded toes, bradycardia, second-degree heart block, hypotonia, cerebellar hypoplasia, and white matter abnormalities (Zaki et al., 2011). Unfortunately, the locus responsible for Zaki-Gleeson syndrome remains unknown. A comparison between the two affected siblings described here, Zaki-Gleeson syndrome, and EXOSC5-related disease reveals considerable overlap (Suppl. Table 1).

While microcephaly was not seen in any patients with EXOSC5-related disease thus far, it is a feature of EXOSC3-related disease (Slavotinek et al., 2020). Therefore, the microcephaly seen in this family may represent phenotypic expansion of EXOSC5-related disease. The more severe phenotype of the older brother is notable and may reflect intrafamilial variability, disease progression, or a blended phenotype due to a dual molecular diagnosis (Posey et al., 2017). As clinical ES was not performed due to financial issues, this remains unresolved, but such intrafamilial variation might indeed reflect multilocus pathogenic variation (Karaca et al., 2018; Mitani et al., 2019).

Most reported human phenotypes associated with exosome genes involve cerebellar and/or motor neurons. As many cerebellar ataxia and motor neuron disease genes regulate RNA biology (e.g. SMN1, SETX, PUM1), it has long been hypothesized that these neurons are exquisitely sensitive to RNA processing defects (Conlon et al., 2017). However, the exosome complex is ubiquitously expressed, and its dysfunction must impact tissues outside the central nervous system. One example of extra-CNS manifestations is the defective class switch recombination and somatic hypermutation seen in B cell specific Exosc3 conditional knockout mice (Pefanis et al., 2014). Given all exosome subunits in yeast and all subunits examined in mice are required for viability (Kilchert et al., 2016; Dickinson et al., 2016; http://www.mousephenotype.org), biallelic loss-of-function (LoF) EXOSC variants are likely also incompatible with human life. The degree of residual exosome function may therefore determine the severity of the phenotype and the extent of extra-CNS involvement. Further patient identification, tissue-specific knockout of exosome genes, and functional study of patient EXOSC variants will help fully delineate the spectrum of exosome-related disease and its involvement in brain and cardiac developmental biology and physiology.