1 INTRODUCTION

Neuromuscular disease and oculoauditory anomalies, or NMOAS (MIM #618733), is a condition with a variable clinical presentation associated with heterozygous pathogenic genetic variants in DHX16. Paine et al. (2019) first reported a cohort of four individuals with de novo missense variants in DHX16; one died shortly after birth and the other three presented with various combinations of developmental delay, neuromuscular disease, seizures, ocular nerve and/or retinal abnormalities, and sensorineural hearing loss (SNHL). Heterozygous pathogenic DHX16 missense variants have since been reported as de novo in an individual with congenital myopathy (Park et al., 2022) and with unknown inheritance in an individual noted to have infantile spasms, hypotonia, severe bilateral SNHL, and retinal and optic nerve abnormalities (Archana et al., 2022).

DHX16 encodes DEAH-box helicase 16 (DHX16), a nearly ubiquitously expressed ATP-dependent RNA helicase that participates in spliceosome complex B which executes pre-mRNA splicing. Mutant DHX16 has been shown in vitro to cause defective spliceosome activity that leads to retention of un-spliced pre-mRNA transcripts from numerous intron-containing genes in the cell nuclei; this appears to be caused by the formation of a catalytically inactive spliceosome complex by mutant DHX16 and its interaction with G-patch partner GPKOW, exerting a dominant negative effect (Bohnsack et al., 2021; Gencheva, Kato, et al., 2010; Gencheva, Lin, et al., 2010; Zang et al., 2014). The resulting outcome is nuclear accumulation of aberrant transcripts, in which the abundance of aberrant mRNA transcripts is detrimental to proper cellular function; this is likely in addition to a harmful reduction in expression of genes whose un-spliced mRNA transcripts remain in the nucleus and are thus not translated into functional proteins. Pathogenic variants in other members of the DExD/H-box RNA helicase superfamily of genes are known to cause moderately similar phenotypes that include abnormal central nervous system (CNS) development with or without extra-CNS anomalies (Paine et al., 2019; Snijders Blok et al., 2015; Wang et al., 2018).

In this report, we present a patient with features overlapping NMOAS, in addition to previously unreported mitochondrial deficiency and primary ovarian insufficiency (POI), who was identified to have a novel de novo missense variant in DHX16.

2 CASE REPORT

Our patient was the product of a twin gestation conceived by in vitro fertilization, though the second fetus did not survive beyond 17 weeks gestation. She was delivered at 36 weeks gestation after development of oligohydramnios, weighing 5 pounds and 7 ounces. She failed the newborn hearing screening and was determined to have profound bilateral sensorineural hearing loss (SNHL). Following clinical evaluation, she received genetic testing for Usher syndrome and other common forms of genetic hearing loss which were non-diagnostic. She had bilateral hearing aids followed by placement of bilateral cochlear implants with surgeries performed at 13 and 22 months of age.

Symptoms of vision impairment were noted early in life, starting with nyctalopia. She underwent an electroretinogram at the age of 2 years which demonstrated markedly attenuated amplitudes consistent with severe cone- and rod-mediated disease. She has been followed for routine examination, and her ocular phenotype is most consistent with retinitis pigmentosa. She continues to have good central vision (20/25 bilaterally) with severely depressed peripheral and night vision.

Hypotonia and motor delays were present at 12–18 months of age; on neurological assessment she was noted to have muscle weakness and hyporeflexia but was able to pull to stand and cruise short distances. Our patient never walked independently and demonstrated regression of motor abilities after 18 months of age. Nerve conduction studies demonstrated length-dependent peripheral neuropathy characterized primarily by axonal loss, with more marked motor changes than sensory. Genetic testing for Charcot–Marie-Tooth disease was uninformative.

On neuropsychiatric examination, our patient was noted to have slight deficits in her verbal skills and working memory, consistent with an attention deficit hyperactivity disorder (ADHD) inattentive type presentation. There are no other notable differences in her cognitive function. At age four, our patient underwent adrenarche, prompting an endocrinology work-up which detected elevated luteinizing hormone (LH) and follicle-stimulating hormone (FSH) levels. A pelvic ultrasound revealed apparently normal uterine and ovarian anatomy. Around age eight or nine, the patient began thelarche, but development arrested by age 12 at Tanner stages II to III. By the time of evaluation for primary amenorrhea at age 15, she had not experienced a menstrual cycle nor instances of vaginal spotting. This workup also found our patient to still have elevated LH and FSH (53.04 and 62.5 mLU/mL, respectively) levels, with associated low estradiol (7 pg/mL) and anti-Mullerian hormone (0.01 ng/mL) levels; she was given a diagnosis of primary ovarian insufficiency (POI) and began treatment with estradiol and medroxyprogesterone. Chromosome analysis confirmed a 46,XX karyotype, FMR1 analysis demonstrated biallelic CGG repeat sizes within the normal range (29/33), and a gene panel test for Perrault syndrome was non-diagnostic.

Histological and immunohistochemical analyses were performed on skeletal muscle biopsies. These studies demonstrated areas of denervation and reinnervation with a mild increase in internal nuclei and scattered necrotic or regenerating fibers. Muscle fibers showed a decrease in cytochrome c oxidase activity, though complete cytochrome c oxidase negativity was not seen. Type I fibers were more abundant than type II and some fascicles consisted almost entirely of type I fibers. These findings were considered indicative of a myopathic process with associated denervation. Electron transport chain enzyme analyses performed on the muscle tissue demonstrated reduced citrate synthase and complex IV activity, with complex II activity preserved. Mitochondrial DNA copy number results (39% of the mean value of age) were consistent with mitochondrial depletion syndrome; mitochondrial DNA sequencing and deletion analyses were negative. A 14-gene mitochondrial DNA depletion panel was negative. Her creatine kinase was significantly elevated at approximately 1000 U/L. Our patient was subsequently prescribed vitamin C, coenzyme Q10, and carnitine for mitochondrial supplementation. Trio-based exome sequencing (ES) with copy number analysis and mitochondrial genome analysis conducted at a reference laboratory were also non-diagnostic.

Trio-based genome sequencing (GS) with copy number analysis was subsequently performed as part of an ongoing research protocol at Ann & Robert H. Lurie Children's Hospital of Chicago (IRB #2021-4158). This analysis revealed the de novo heterozygous likely pathogenic missense variant in DHX16 NM_003587.4:c.2033A > G (p.Glu678Gly) which was clinically confirmed by the laboratory using Sanger sequencing. The reported parental relationships were also confirmed using short tandem repeat (STR) analyses. No CNVs of potential clinical significance were identified by GS.

3 DISCUSSION

We report a case of possible NMOAS with additional features of POI and mitochondrial deficiency in a patient found to have a confirmed de novo heterozygous likely pathogenic variant in DHX16. This variant is absent from gnomAD (Lek et al., 2016) and is strongly predicted to be deleterious (REVEL = 0.931) (Ioannidis et al., 2016; Pejaver et al., 2022). ACMG-AMP criteria applied: PS2, PP3, and PM2 (Richards et al., 2015).

Our patient's variant represents the sixth unique missense, and the fifth known de novo DHX16 variant reported in association with disease (Table 1). The gnomAD missense constraint (Z-score = 3.08) and loss-of-function intolerance (pLI = 0.00) metrics (Karczewski et al., 2020) for DHX16 suggest that it is missense variation, rather than protein truncation or loss of allelic expression, that is likely to be the main mechanism that causes disease. These constraint metrics and the current understanding of the mutational spectrum are consistent with the hypothesized dominant negative mechanism of disease for DHX16, in which pathogenic missense variants lead to a non-functional protein that still interacts with GPKOW but forms a catalytically inactive spliceosome, resulting in nuclear accumulation of many un-spliced pre-mRNA transcripts (Gencheva, Kato, et al., 2010; Gencheva, Lin, et al., 2010; Paine et al., 2019; Zang et al., 2014).

We also performed 3D protein structural modeling and in silico pathogenicity predictions using Missense3D (Ittisoponpisan et al., 2019) (PDB code 7DVQ) and REVEL (Ioannidis et al., 2016), respectively, for seven variants: these include the five different variants previously reported in affected individuals (Archana et al., 2022; Paine et al., 2019; Park et al., 2022), the helicase domain variant used in the first studies demonstrating mutant DHX16's role in aberrant splicing activity (Gencheva, Kato, et al., 2010; Gencheva, Lin, et al., 2010; Zang et al., 2014), and that found in our patient (Table 2). Four of the seven variants, including our patient's, were predicted to induce one or more significant changes to DHX16 structure, such as replacing buried glycine residues, altering cavity sizes or dimensions, and breaking buried hydrogen bonds or salt bridges (Table 2). The high REVEL scores (≥0.890) for three of those are concordant with the predicted impacts to the encoded protein's structure; note that because p.Gly724Asn is only possible via a multinucleotide deletion–insertion, there is no REVEL score available for this variant. Interestingly, three variants previously reported in patients (p.Arg482His, p.Phe582Ile, p.Gln697His; Archana et al., 2022; Paine et al., 2019) were not predicted to induce structural changes to DHX16, and the latter two variants' REVEL scores (0.128 and 0.314, respectively) similarly supported minimal or no predicted functional impacts to the protein. It is therefore possible that some variants in DHX16 may impact spliceosome functioning in a manner to which current in silico prediction methodologies are ineffective at capturing, while the high REVEL scores and predicted structural impacts of the other four variants, including our patient's, clearly support their role in NMOAS pathogenesis.

Human diseases associated with the dysregulation of ubiquitously expressed genes that are active in regulating the expression of other genes, namely conditions such as the chromatinopathies (Fahrner & Bjornsson, 2014), are typically multisystemic and often exhibit considerable variability in phenotypic expression. Our patient has features consistent with those seen previously in individuals with DHX16 pathogenic variants, but it is notable that no single feature is present in all patients (Table 1). Though the understanding of DHX16-related disease is still nascent, it is likely that certain phenotypes will emerge as more common in patients with pathogenic DHX16 variants: six of seven patients were reported to have some form of muscle weakness with hypotonia and/or neuropathy; five of seven had one or more ocular anomalies including various retinal manifestations, with a range of severity and progression; and four of seven had severe to profound bilateral SNHL, with the one patient with none of these features deceasing before any of these phenotypes were presumably ascertainable. A history of epileptic activity was noted in three of the previously reported patients, and a myopathy was described in two individuals. More studies are needed to determine how DHX16's role in the spliceosome complex can give rise to the diverse phenotypic abnormalities observed in the different organ systems in individuals with DHX16 pathogenic variants.

Delineation of common phenotypic features associated with a genetic condition can assist clinicians in the diagnostic process in the future, but the presence of atypical phenotypes, such as POI, mitochondrial deficiency, flexion contractures, and others reported in a minority of patients, may be confounding. Indeed, with a multitude of conditions with overlapping presentations, deriving a diagnosis for this condition solely based on clinical features may be extremely challenging or even impossible. Genetic testing, either with gene panels including DHX16 or via trio-based ES or GS, may be necessary to achieve a molecular diagnosis for these patients, while simultaneously effectively ruling out possible differential diagnoses.

This report serves to strengthen the evidence of an association between pathogenic missense variants in DHX16 and NMOAS, as well as to potentially expand the phenotypic spectrum associated with variation in this gene. Given the variable expressivity common in chromatinopathies as well as mitochondrial and mitochondrial-like disorders (Fahrner & Bjornsson, 2014; Stenton & Prokisch, 2020), and the emerging evidence of DHX16's role in disease, DHX16 analysis should be considered during the diagnostic testing process for individuals with a combination of neuromuscular disease, SNHL, and ocular abnormalities, even if additional, potentially confounding phenotypes are also present. Furthermore, this gene has sufficient evidence supporting its disease association to be included in gene lists derived for phenotype-driven analysis of ES and GS datasets from patients presenting with the mentioned overlapping features. However, testing laboratories must assess the effectiveness of their informatics pipelines to gather genes with emerging evidence of disease associations for analysis; this finding was not reported during the clinical ES by a reference laboratory, evidently due to an unrecognized disease association that was published in the literature approximately 4 months prior to the date of our patient's test order and 7 months prior to it being reported. We expect that a clinical ES reanalysis would have detected and appropriately prioritized this variant with incorporation of the new DHX16 disease association. Time-lapsed reanalyses of ES data have been repeatedly shown to yield additional molecular diagnoses and this report adds yet more anecdotal evidence supporting the utility of this practice (Liu et al., 2019; Wenger et al., 2017).

Additional investigations can elucidate the precise role of this gene in the developmental period and throughout life. Identification and study of larger cohorts with DHX16 variants will facilitate investigations of the full spectrum of associated clinical manifestations and may eventually allow for examination of genotype–phenotype correlations.

4 CONCLUSION

We report a patient with axonal sensorimotor peripheral neuropathy, congenital profound bilateral SNHL, progressive retinopathy, mitochondrial dysfunction, and POI who was found to have a novel diagnostic de novo missense variant in DHX16 following previous comprehensive non-diagnostic genetic testing. This case represents a potential expansion of the clinical features associated with pathogenic missense variation in DHX16 and provides additional evidence that this gene should be routinely included in analyses for individuals with similar clinical presentations. Additionally, de novo missense variants in this gene have the most well-established disease associations in the current literature, which is further strengthened by our report. Taken together, this information supports the diagnostic utility of including DHX16 in genetic work-up of patients presenting with a unique constellation of muscular disease, hearing loss, and ocular anomalies.

AUTHOR CONTRIBUTIONS

Andy Drackley: Conceptualization, data curation, formal analysis, investigation, project administration, writing—original draft preparation, and writing—review & editing. Lenika De Simone: Conceptualization and writing—review & editing. Nancy Kuntz: Conceptualization and writing—review & editing. Safa Rahmani: Writing—review & editing. Alexander Ing: Writing—review & editing. Vamshi K. Rao: Writing—review & editing. Pamela Rathbun: Writing—review & editing. Kai Lee Yap: Conceptualization, formal analysis, project administration, resources, supervision, and writing—review & editing.

CONFLICT OF INTEREST STATEMENT

Dr. Rao has received grants from NS Pharma, RegenxBio, Alexion, and Sarepta and nonfinancial support from Ann and Robert H. Lurie Children's Hospital for conduct of clinical trials. Dr. Rao has also received personal fees from Biogen, Avexis/Novartis, Capricor, NSPharma, Regenxbio, Genentech-Roche, Scholar Rock, PTC Therapeutics, Sarepta Therapeutics, France Foundation, and MDA outside the submitted work.

PATIENT CONSENT STATEMENT

This patient and family have provided consent to their involvement in this report.