Unraveling the pathogenesis of ARX polyalanine tract variants using a clinical and molecular interfacing approach

Further characterisation of ARX-related disorders in females due to inherited or de novo variants

The Aristaless-related homeobox (ARX) gene is located on the X chromosome and encodes a transcription factor that is essential for brain development. While the clinical spectrum of ARX-related disorders is well described in males, from X linked lissencephaly with abnormal genitalia syndrome to syndromic and non-syndromic intellectual disability (ID), its phenotypic delineation in females is incomplete. Carrier females in ARX families are usually asymptomatic, but ID has been reported in some of them, as well as in others with de novo variants. In this study, we collected the clinical and molecular data of 10 unpublished female patients with de novo ARX pathogenic variants and reviewed the data of 63 females from the literature with either de novo variants (n=10), inherited variants (n=33) or variants of unknown inheritance (n=20). Altogether, the clinical spectrum of females with heterozygous pathogenic ARX variants is broad: 42.5% are asymptomatic, 16.4% have isolated agenesis of the corpus callosum (ACC) or mild symptoms (learning disabilities, autism spectrum disorder, drug-responsive epilepsy) without ID, whereas 41% present with a severe phenotype (ie, ID or developmental and epileptic encephalopathy (DEE)). The ID/DEE phenotype was significantly more prevalent in females carrying de novo variants (75%, n=15/20) versus in those carrying inherited variants (27.3%, n=9/33). ACC was observed in 66.7% (n=24/36) of females who underwent a brain MRI. By refining the clinical spectrum of females carrying ARX pathogenic variants, we show that ID is a frequent sign in females with this X linked condition.

Genome-wide prediction of disease variant effects with a deep protein language model

Predicting the effects of coding variants is a major challenge. While recent deep-learning models have improved variant effect prediction accuracy, they cannot analyze all coding variants due to dependency on close homologs or software limitations. Here we developed a workflow using ESM1b, a 650-million-parameter protein language model, to predict all ~450 million possible missense variant effects in the human genome, and made all predictions available on a web portal. ESM1b outperformed existing methods in classifying ~150,000 ClinVar/HGMD missense variants as pathogenic or benign and predicting measurements across 28 deep mutational scan datasets. We further annotated ~2 million variants as damaging only in specific protein isoforms, demonstrating the importance of considering all isoforms when predicting variant effects. Our approach also generalizes to more complex coding variants such as in-frame indels and stop-gains. Together, these results establish protein language models as an effective, accurate and general approach to predicting variant effects.

Scalable detection of technically challenging variants through modified next-generation sequencing

BACKGROUND

Some clinically important genetic variants are not easily evaluated with next-generation sequencing (NGS) methods due to technical challenges arising from high- similarity copies (e.g., PMS2, SMN1/SMN2, GBA1, HBA1/HBA2, CYP21A2), repetitive short sequences (e.g., ARX polyalanine repeats, FMR1 AGG interruptions in CGG repeats, CFTR poly-T/TG repeats), and other complexities (e.g., MSH2 Boland inversions).

METHODS

We customized our NGS processes to detect the technically challenging variants mentioned above with adaptations including target enrichment and bioinformatic masking of similar sequences. Adaptations were validated with samples of known genotypes.

RESULTS

Our adaptations provided high-sensitivity and high-specificity detection for most of the variants and provided a high-sensitivity primary assay to be followed with orthogonal disambiguation for the others. The sensitivity of the NGS adaptations was 100% for all of the technically challenging variants. Specificity was 100% for those in PMS2, GBA1, SMN1/SMN2, and HBA1/HBA2, and for the MSH2 Boland inversion; 97.8%-100% for CYP21A2 variants; and 85.7% for ARX polyalanine repeats.

CONCLUSIONS

NGS assays can detect technically challenging variants when chemistries and bioinformatics are jointly refined. The adaptations described support a scalable, cost-effective path to identifying all clinically relevant variants within a single sample.

A novel ARX loss of function variant in female monozygotic twins is associated with chorea

Pathogenic variants in ARX lead to a variety of phenotypes with intellectual disability being a uniform feature. Other features can include severe epilepsy, spasticity, movement disorders, agenesis of the corpus callosum, lissencephaly, hydranencephaly and ambiguous genitalia in males. We present the first report of monozygotic female twins with a de novo ARX pathogenic variant (c.1406_1415del; p. Ala469Aspfs*20), predicted to result in a truncated ARX protein missing the important regulatory Aristaless domain. The twins presented with profound developmental delay and seizures, consistent with the known genotype-phenotype correlation. Twin 2's features were significantly more severe. She also developed chorea; the first time this movement disorder has been seen in an ARX variant other than an expansion of the first polyalanine tract. Differential X-chromosome inactivation was the most likely explanation for the differing severities but could not be conclusively proven.

Early 17β-estradiol treatment reduces seizures but not abnormal behaviour in mice with expanded polyalanine tracts in the Aristaless related homeobox gene (ARX)

Children with severe intellectual disability have an increased prevalence of refractory seizures. Steroid treatment may improve seizure outcomes, but the mechanism remains unknown. Here we demonstrate that short term, daily delivery of an exogenous steroid 17β-estradiol (40 ng/g) in early postnatal life significantly reduced the number and severity of seizures, but did not improve behavioural deficits, in mice modelling mutations in the Aristaless-related homeobox gene (ARX), expanding the first (PA1) or second (PA2) polyalanine tract. Frequency of observed seizures on handling (n = 14/treatment/genotype) were significantly reduced in PA1 (32% reduction) and more modestly reduced in PA2 mice (14% reduction) with steroid treatment compared to vehicle. Spontaneous seizures were assessed (n = 7/treatment/genotype) at 7 weeks of age coinciding with a peak of seizure activity in untreated mice. PA1 mice treated with steroids no longer present with the most severe category of prolonged myoclonic seizures. Treated PA2 mice had an earlier onset of seizures coupled with a subsequent reduction in seizures later in postnatal life, with a complete absence of any seizures during the analysis at 7 weeks of age. Despite the reduction in seizures, 17β-estradiol treated mice showed no improvement in behavioural or cognitive outcomes in adulthood. For the first time we show that these deficits due to mutations in Arx are already present before seizure onset and do not worsen with seizures. ARX is a transcription factor and Arx PA mutant mice have deregulated transcriptome profiles in the developing embryonic brain. At postnatal day 10, treatment completion, RNAseq identified 129 genes significantly deregulated (Log2FC > ± 0.5, P-value<0.05) in the frontal cortex of mutant compared to wild-type mice. This list reflects genes deregulated in disease and was particularly enriched for known genes in neurodevelopmental disorders and those involved in signalling and developmental pathways. 17β-estradiol treatment of mutant mice significantly deregulated 295 genes, with only 23 deregulated genes overlapping between vehicle and steroid treated mutant mice. We conclude that 17β-estradiol treatment recruits processes and pathways to reduce the frequency and severity of seizures in the Arx PA mutant mice but does not precisely correct the deregulated transcriptome nor improve mortality or behavioural and cognitive deficits.

Aristaless-Related Homeobox (ARX): Epilepsy Phenotypes beyond Lissencephaly and Brain Malformations

The Aristaless-related homeobox (ARX) transcription factor is involved in the development of GABAergic and cholinergic neurons in the forebrain. ARX mutations have been associated with a wide spectrum of neurodevelopmental disorders in humans and are responsible for both malformation (in particular lissencephaly) and nonmalformation complex phenotypes. The epilepsy phenotypes related to ARX mutations are West syndrome and X-linked infantile spasms, X-linked myoclonic epilepsy with spasticity and intellectual development and Ohtahara and early infantile epileptic encephalopathy syndrome, which are related in most of the cases to intellectual disability and are often drug resistant. In this article, we shortly reviewed current knowledge of the function of ARX with a particular attention on its consequences in the development of epilepsy during early childhood.

A new mouse model of ARX dup24 recapitulates the patients' behavioral and fine motor alterations

The aristaless-related homeobox (ARX) transcription factor is involved in the development of GABAergic and cholinergic neurons in the forebrain. ARX mutations have been associated with a wide spectrum of neurodevelopmental disorders in humans, among which the most frequent, a 24 bp duplication in the polyalanine tract 2 (c.428_451dup24), gives rise to intellectual disability, fine motor defects with or without epilepsy. To understand the functional consequences of this mutation, we generated a partially humanized mouse model carrying the c.428_451dup24 duplication (Arx^dup24/0) that we characterized at the behavior, neurological and molecular level. Arx^dup24/0 males presented with hyperactivity, enhanced stereotypies and altered contextual fear memory. In addition, Arx^dup24/0 males had fine motor defects with alteration of reaching and grasping abilities. Transcriptome analysis of Arx^dup24/0 forebrains at E15.5 showed a down-regulation of genes specific to interneurons and an up-regulation of genes normally not expressed in this cell type, suggesting abnormal interneuron development. Accordingly, interneuron migration was altered in the cortex and striatum between E15.5 and P0 with consequences in adults, illustrated by the defect in the inhibitory/excitatory balance in Arx^dup24/0 basolateral amygdala. Altogether, we showed that the c.428_451dup24 mutation disrupts Arx function with a direct consequence on interneuron development, leading to hyperactivity and defects in precise motor movement control and associative memory. Interestingly, we highlighted striking similarities between the mouse phenotype and a cohort of 33 male patients with ARX c.428_451dup24, suggesting that this new mutant mouse line is a good model for understanding the pathophysiology and evaluation of treatment.

Mutations of ARX and non-syndromic intellectual disability in Chinese population

Mutations of Aristaless-related homeobox (ARX) gene were looked as the third cause of non-syndromic intellectual disability (NSID), while the boundary between true disease-causing mutations and non-disease-causing variants within this gene remains elusive. To investigate the relationship between ARX mutations and NSID, a panel comprising six reported causal mutations of the ARX was detected in 369 sporadic NSID patients and 550 random participants in Chinese. Two mutations, c.428_451 dup and p.G286S, may be disease-causing mutations for NSID, while p.Q163R and p.P353L showed a great predictive value in female NSID diagnosis with significant associations (X² = 19.60, p = 9.54e-6 for p.Q163R; X² = 25.70, p = 4.00e-07 for p.P353L), carriers of these mutations had an increased risk of NSID of more than fourfold. Detection of this panel also predicted significant associations between genetic variants of the ARX gene and NSID (p = 3.73e-4). The present study emphasized the higher genetic burden of the ARX gene on NSID in the Chinese population, molecular analysis of this gene should be considered for patients presenting NSID of unknown etiology.

ARX polyalanine expansion mutations lead to migration impediment in the rostral cortex coupled with a developmental deficit of calbindin-positive cortical GABAergic interneurons

The Aristaless-related homeobox gene (ARX) is indispensable for interneuron development. Patients with ARX polyalanine expansion mutations of the first two tracts (namely PA1 and PA2) suffer from intellectual disability of varying severity, with seizures a frequent comorbidity. The impact of PA1 and PA2 mutations on the brain development is unknown, hindering the search for therapeutic interventions. Here, we characterized the disturbances to cortical interneuron development in mice modeling the two most common ARX polyalanine expansion mutations in human. We found a consistent ∼40-50% reduction of calbindin-positive interneurons, but not Stt+ or Cr+ interneurons, within the cortex of newborn hemizygous mice (p=0.024) for both mutant strains compared to wildtype (p=0.011). We demonstrate that this was a consequence of calbindin precursor cells being arrested or delayed at the ventral subpallium en route of tangential migration. Ex-vivo assay validated this migration deficit in PA1 cells (p=0.0002) suggesting that the defect is contributed by intrinsic loss of Arx function within migrating cells. Both humans and mice with PA1 mutations present with severe clinical features, including intellectual disability and infantile spasms. Our data further demonstrated the pathogenic mechanism was robustly shared between PA1 and PA2 mutations, as previously reported including Arx protein reduction and overlapping transcriptome profiles within the developing mouse brains. Data from our study demonstrated that cortical calbindin interneuron development and migration is negatively affected by ARX polyalanine expansion mutations. Understanding the cellular pathogenesis contributing to disease manifestation is necessary to screen efficacy of potential therapeutic interventions.

Embryonic forebrain transcriptome of mice with polyalanine expansion mutations in the ARX homeobox gene

The Aristaless-related homeobox (ARX) gene encodes a paired-type homeodomain transcription factor with critical roles in embryonic development. Mutations in ARX give rise to intellectual disability (ID), epilepsy and brain malformation syndromes. To capture the genetics and molecular disruptions that underpin the ARX-associated clinical phenotypes, we undertook a transcriptome wide RNASeq approach to analyse developing (12.5 dpc) telencephalon of mice modelling two recurrent polyalanine expansion mutations with different phenotypic severities in the ARX gene. Here we report 238 genes significantly deregulated (Log2FC  > +/-1.1, P-value <0.05) when both mutations are compared to wild-type (WT) animals. When each mutation is considered separately, a greater number of genes were deregulated in the severe PA1 mice (825) than in the PA2 animals (78). Analysing genes deregulated in either or both mutant strains, we identified 12% as implicated in ID, epilepsy and autism (99/858), with ∼5% of them as putative or known direct targets of ARX transcriptional regulation. We propose a core pathway of transcription regulators, including Hdac4, involved in chromatin condensation and transcriptional repression, and one of its targets, the transcription factor Twist1, as potential drivers of the ID and infantile spasms in patients with ARX polyalanine expansion mutations. We predict that the subsequent disturbance to this pathway is a consequence of ARX protein reduction with a broader and more significant level of disruption in the PA1 in comparison to the PA2 mice. Identifying early triggers of ARX-associated phenotypes contributes to our understanding of particular clusters/pathways underpinning comorbid phenotypes that are shared by many neurodevelopmental disorders.

An Emerging Female Phenotype with Loss-of-Function Mutations in the Aristaless-Related Homeodomain Transcription Factor ARX

The devastating clinical presentation of X-linked lissencephaly with abnormal genitalia (XLAG) is invariably caused by loss-of-function mutations in the Aristaless-related homeobox (ARX) gene. Mutations in this X-chromosome gene contribute to intellectual disability (ID) with co-morbidities including seizures and movement disorders such as dystonia in affected males. The detection of affected females with mutations in ARX is increasing. We present a family with multiple affected individuals, including two females. Two male siblings presenting with XLAG were deceased prior to full-term gestation or within the first few weeks of life. Of the two female siblings, one presented with behavioral disturbances, mild ID, a seizure disorder, and complete agenesis of the corpus callosum (ACC), similar to the mother's phenotype. A novel insertion mutation in Exon 2 of ARX was identified, c.982delCinsTTT predicted to cause a frameshift at p.(Q328Ffs^* 37). Our finding is consistent with loss-of-function mutations in ARX causing XLAG in hemizygous males and extends the findings of ID and seizures in heterozygous females. We review the reported phenotypes of females with mutations in ARX and highlight the importance of screening ARX in male and female patients with ID, seizures, and in particular with complete ACC.

Genetics and genotype-phenotype correlations in early onset epileptic encephalopathy with burst suppression

OBJECTIVE

We sought to identify genetic causes of early onset epileptic encephalopathies with burst suppression (Ohtahara syndrome and early myoclonic encephalopathy) and evaluate genotype-phenotype correlations.

METHODS

We enrolled 33 patients with a referral diagnosis of Ohtahara syndrome or early myoclonic encephalopathy without malformations of cortical development. We performed detailed phenotypic assessment including seizure presentation, electroencephalography, and magnetic resonance imaging. We confirmed burst suppression in 28 of 33 patients. Research-based exome sequencing was performed for patients without a previously identified molecular diagnosis from clinical evaluation or a research-based epilepsy gene panel.

RESULTS

In 17 of 28 (61%) patients with confirmed early burst suppression, we identified variants predicted to be pathogenic in KCNQ2 (n = 10), STXBP1 (n = 2), SCN2A (n = 2), PNPO (n = 1), PIGA (n = 1), and SEPSECS (n = 1). In 3 of 5 (60%) patients without confirmed early burst suppression, we identified variants predicted to be pathogenic in STXBP1 (n = 2) and SCN2A (n = 1). The patient with the homozygous PNPO variant had a low cerebrospinal fluid pyridoxal-5-phosphate level. Otherwise, no early laboratory or clinical features distinguished the cases associated with pathogenic variants in specific genes from each other or from those with no prior genetic cause identified.

INTERPRETATION

We characterize the genetic landscape of epileptic encephalopathy with burst suppression, without brain malformations, and demonstrate feasibility of genetic diagnosis with clinically available testing in >60% of our cohort, with KCNQ2 implicated in one-third. This electroclinical syndrome is associated with pathogenic variation in SEPSECS. Ann Neurol 2017;81:419-429.

A novel mutation in the OAR domain of the ARX gene

Mutations in ARX gene should be considered in patients with mental disability or/and epilepsy. It is an X-linked gene that has pleiotropic effects. Here, we report the case of a boy diagnosed with Ohtahara syndrome. We performed the molecular analysis of the gene and identified a new missense mutation.