A review of the clinical spectrum of BRAT1 disorders and case of developmental and epileptic encephalopathy surviving into adulthood

Highlights • We report a patient with a BRAT1 disorder who survived into adulthood.• Infantile-onset developmental and epileptic encephalopathy rather than an ataxic presentation was the presenting feature.• Compound heterozygous BRAT1 variants may be associated with later onset of seizures and longer survival reflecting a spectrum of clinical disease.


Introduction
Sporadic developmental and epileptic encephalopathies (DEE) are most commonly attributable to monoallelic variants arising de novo but can be transmitted with recessive and x-linked variants. Pathogenic biallelic variants in BRAT1 are associated with a spectrum of disease including Lethal Neonatal Rigidity and Multifocal Seizure Syndrome (RMFSL, OMIM: 614498), Epilepsy of Infancy with Migrating Focal Seizures (EIMFS), and Neurodevelopmental Disorder with Cerebellar Atrophy and with or without Seizures (NEDCAS, OMIM: 618056) [1].
RMFSL is a severe disorder characterized by microcephaly, axial and appendicular rigidity, early-onset (possibly in utero) multifocal seizures, and episodes of bradycardia and apnea typically resulting in death during infancy [2][3][4]. A RMFSL-like phenotype has been described, with patients sharing some of the clinical features of RMFSL, but with a longer survival and later onset of epilepsy [5][6][7][8]. Bialleic variants in BRAT1 can also cause EIMFS, an early-onset encephalopathic syndrome characterized by seizures that appear to 'migrate' between cerebral hemispheres and significant cognitive impairment and developmental delay [9,10]. Individuals with EIMFS caused by BRAT1 variants may have an earlier onset and greater severity than those with EIMFS due to other causes [10]. RMFSL and EIMFS have genetic and phenotypic overlap and may represent similar disorders or different perspectives of the same disorder [11]. In contrast to these two, NEDCAS is a milder phenotype where patients may survive past infancy and exhibit intellectual disability, and ataxia with cerebellar degeneration with or without seizures [8,12]. The current understanding of the spectrum of BRAT1 disorders reported in the literature is summarized in Table 1 which demonstrates the clinical, radiological, and electrophysiological heterogeneity seen in disorders caused by BRAT1 variants.
We describe the history of a man with BRAT1 variants and an RMFSL-like phenotype with microcephaly, limb hypertonia, developmental delay, and multifocal epilepsy. Significantly, he survived until the age of twenty years, substantially longer than any previously reported case.

Case report
A 20-year-old man with non-consanguineous parents was born by caesarean section at 35 weeks due to severe intrauterine growth restriction. His mother and father did not have any history of epilepsy, neurological disorders, or any other medical problems.  His birth weight was 1215 g and his head circumference was 27.2 cm, both below the third centile. He spent four weeks on the neonatal unit due to low birth weight and was noted to have abdominal distention, axial hypotonia, limb hypertonia and episodes of both bradycardia and apnea. At the age of one year, he still exhibited axial hypotonia and limb hypertonia in addition to horizontal and upbeat nystagmus, a sluggish pupillary reflex to light, and inability fix his gaze on objects. He had persistent constipation requiring treatment with regular laxatives and enemas. A viral serology screen did not demonstrate evidence of congenital viral infection with CMV, parvovirus B19, Rubella, HSV, or congenital toxoplasmosis.
He had severe global developmental delay without regression. He did not smile until the age of 16 months, made babbling noises at the age of 21 months, and rolled at 22 months. He was never able to walk and required a wheelchair to mobilize. He had a wide nasal bridge, bilateral epicanthic folds, and dysmorphic ears.
At 8 months of age, he presented with an afebrile transient loss of consciousness lasting 5 min. He was reported to have been unresponsive, grinding his teeth with his eyes rolled backwards, and was grey and floppy without limb movements.
By the age of 3 years, his parents had witnessed at least six unresponsive, cyanotic, and hypotonic episodes with twitching of his limbs lasting for a few minutes each which were presumed to be bilateral tonic-clonic seizures of unknown onset. This was in addition to multiple daily brief episodes of apparent loss of awareness and occasional myoclonic jerks. At this point, epilepsy was diagnosed, and he commenced sodium valproate 200 mg twice a day. After two-years without tonic-clonic seizures, these relapsed, increasing in frequency to clusters of 3-4 seizures occurring approximately once a week. We have not been able to obtain medical records or a collateral history covering the period from childhood to until he represented at the age of 15 years, when he commenced topiramate but continued to have seizures weekly. This was thought to be contributed to, in part, by nonconcordance with anti-seizure medication when he stopped eating and drinking. A gastrostomy tube was subsequently inserted to support nutrition and medication administration. Despite this, seizure frequency worsened over the following five years, and at the age of 20, levetiracetam was added with an improvement in seizure frequency. He died aged 20 years due to aspiration pneumonia complicating unexplained gastroparesis with vomiting.
Blood tests in early childhood including full blood count, urea and electrolytes, and liver and thyroid profiles were normal, as were serum lactate and pyruvate. While the original data from electroencephalography (EEG) study at four years of age is no longer available, it was reported as showing symmetric background alpha, beta, and theta rhythms with occasional increased amplitude delta activity, which was notched and sharp. There were also frequent focal sharp wave and spike discharges, both simultaneously and independently over both hemispheres, consistent with multifocal epilepsy. The report does not mention capture of the events characterized by loss of awareness, that we classify as unknown onset behavior arrest seizures on purely clinical grounds. Similarly, the original magnetic resonance images acquired the same year are no longer available for review but were reported as showing cerebellar hypoplasia with a prominent vermis. EEG during his terminal illness in the intensive care unit without sedation was consistent with severe encephalopathy, demonstrating widespread low amplitude theta and delta rhythms without reactivity to auditory stimuli.
Genetic testing yielded a normal karyotype with no evidence of fragile X syndrome. Whole exome sequencing was preformed during his terminal admission at 20 years of age, and reported as an in silico panel of selected genes (gene list provided as a supplement) by GEMINI (Cambridge University Hospitals NHS Foundation  [7,13,14]. We interpret them according to Amercan College of Medical Genetics and Genomics criteria as pathogenic and likely pathogenic respectively [15]. The lack of parental DNA and the distance between variants has meant that it has not been possible to confirm that these variants are in trans (compound heterozygous).

Discussion
This case illustrates some important clinical lessons applicable to the genetic investigation of epilepsy in adults. First, recessive forms of DEE may be identified with panel testing in adults, even without consanguinity. However, compared to pediatric patients, adults may be less likely to have parental DNA available to confirm the suspected genetic mechanism through segregation of variants; this is also a challenge to the interpretation of monoallelic potentially de novo variants. In the very near future, advances in genetic sequencing technologies will help with the interpretation of variants in recessive disease genes where segregation cannot be demonstrated. In this case, the variants were 5 kilobases apart, raising the possibility that a single long-read may span both loci. Second, among the increasing number of young adults transitioning to adult services with DEE, some have disorders that were previously not recognized in adulthood due to the biases of genetic discovery and initial phenotypic studies, which focus on pediatric cohorts, often comprising children with particularly severe phenotypes. Consequently, some genetic causes may thus far have been associated only with the more severe extreme of their latent clinical spectrum. Finally, for adult patients, the threshold for testing should take account of the clinical heterogeneity of each gene, and the possible incompleteness of historical records of clinical phenomenology, development, EEG, and MRI. These may have been destroyed or incomplete, or not followed the patient from hospital to hospital. When assessing an adult with early onset epilepsy, the historical records and previous investigations are likely to predate the discovery of the particular cause of their epilepsy. Accordingly, the absence of the corresponding classic gene-phenotype association in these medical records should not be assumed to mean that the phenotype was never present -it may not have been actively sought.
The BRAT1 gene (7p22.3) encodes the BRCA1-associated ATM activator 1 protein. This protein is involved in the response to DNA damage and apoptosis and in the regulation of mitochondrial homeostasis [10]. Few detailed descriptions of people with disorders attributable to BRAT1 variants have been published, but these have revealed heterogeneous phenotypic constellations with a varying degree of clinical severity. If seizures are present, they are typically but not exclusively seen in early infancy. The median age of seizure onset in the reported cases of BRAT1-related disease is the first day of life (Table 1).
At the most severe end of the clinical spectrum of disease is RMFSL, typified by severe drug-resistant seizures beginning in utero or within the first days of life, microcephaly, rigidity, swallowing difficulties, and poor psychomotor development. Infants with RMFSL experience apneic and bradycardic episodes culminating in cardiorespiratory arrest and death, typically within the first year of life [5].
A less severe clinical course is characterized by features similar to RMFSL but with survival past infancy, later onset epilepsy and continued development [5][6][7][8]. Our case matches this phenotype, with evidence of microcephaly, limb hypertonia, severe global developmental impairment and drug resistant multifocal epilepsy, and represents the longest reported survival in such an individual with this phenotype to date: the previous oldest patient with this phenotype was reported to still be alive at the age of six years and could be as old as 11 now [7]. The mildest reported constellation of BRAT1-related disease is characterized by intellectual disability, ataxia, and cerebellar atrophy with or without microcephaly and epilepsy [8,12].
It appears that the severity of disease is linked to zygosity of the BRAT1 variants with most of the reported homozygous cases developing seizures immediately after birth or suspected to have had seizures in utero. The longest reported survival for a homozygous patient is 12 months [16]. The reported cases with compound heterozygous BRAT1 variants, however, show a relatively delayed onset of seizures, typically within the first few months of life but in a few cases, after one year of age. These patients generally survive longer, with several previously reported cases of survival into childhood and the present case showing that survival into adulthood is possible. However, two sets of siblings with compound heterozygous variants have shown significant phenotypic variability, indicating that genetic expressivity is variable and other genetic or non-genetic factors contribute [6,17].
Overall, it appears that BRAT1-related epilepsies can include both focal-onset and generalized-onset seizures. Beyond the individual we present here with unknown onset for seizures associated with behavioral arrest, only one other has been reported to have seizures compatible with absences (case 3 in Srivastava, 2016 [7]). This person had late onset epilepsy given the of context of BRAT1-related disorder (3 years) and survival to at least 4 years. This may reflect the difficulty of detecting non-motor seizures in infancy or the increased propensity for absences to occur as the brain develops towards mid-childhood (onset of absences before 4 years of age is rare across the epilepsies and should prompt consideration of genetic testing). In either case, the recognition of non-motor seizures in a person with BRAT1-related epilepsy may be a marker of survival beyond infancy, and thus a milder form of these diseases.
A person with identical BRAT1 variants has previously been described by Mundy et al., who were able to confirm compound heterozygosity and emphasized survival beyond infancy [7]. Similarly, to the present case, they describe microcephaly, mild dysmorphic features, abnormal tone, and developmental delay in addition to drug-resistant seizures. However, the individual that they reported had developed seizures at a younger age (3 months vs 8 months). The three BRAT1 cases reported in Stödberg et al all had shared the c.294dupA, p.(Leu99Thrfs*92) genotype with our case; all three were reported to have pachygyria -not reported in the individual we report, for whom we were not able to review the original MRI or repeat the study [18]. The other variant we report here, c.1925C>A, p.(Ala642Glu), has been described in compound heterozygosity in someone with a neurodevelopmental syndrome with progressive neurodegeneration that fits neither the syndromes of EIMFS nor RMFLS, and who had focal impaired awareness seizures [7]. It is possible that the events with apparent loss of awareness of the person we report were focal impaired awareness seizures.
The prevalence of epilepsy due to BRAT1 variants in adults is currently unknown, although none was detected among 150 adult patients with neurodevelopmental disorders with epilepsy (using a gene panel that included BRAT1) [19]. The national screening study of children with seizures in Scotland did not include BRAT1 on its panel [20].

Conclusion
Biallelic pathogenic BRAT1 variants cause a spectrum of clinical disease with and without seizures. With wider testing and a higher index of suspicion, we may see cases with longer survival in the context of a later age of seizure onset and compound heterozygosity. This case expands the clinical spectrum further to include people reaching adulthood with multifocal epilepsy, dysmorphic features, and developmental delay.

Declaration of Competing Interest
The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: R.H.T. reports Honoraria from Arvelle, Bial, Eisai, GW Pharma, Sanofi, UCB Pharma, UNEEG and Zogenix. E.A. reports Honoraria from Arvelle. R.F., M.E., C.J.M., and D.L.-S. have nothing to declare.