Effective treatment of NR2F1 -related epilepsy with perampanel

Background NR2F1 mutations are associated with Bosch-Boonstra-Schaaf optic atrophy syndrome (BBSOAS). Although ~ 46.7% of BBSOAS patients present with epilepsy, which is always drug-resistant and associated with higher rates of behavioral and cognitive problems, the treatment and outcomes of NR2F1 -related epilepsy have rarely been described. Here, we present new cases of BBSOAS-related epilepsy and summarize all previously reported cases to explore the effective treatment for this type of epilepsy. Methods We identified six new Chinese cases of BBSOAS with epilepsy. Five different de novo heterozygous NR2F1 mutations were identified in these cases, including two novel mutations c.365G > T, p.Cys122Phe and c.449G > T, p.Gly150Val. By combining the six cases and 14 previously reported cases, we analyzed the characteristics and treat-ment outcomes of NR2F1 -related epilepsy. Results Twelve of the 20 patients (60%) had infantile epileptic spasms, while the other patients had generalized tonic/tonic-clonic, focal, myoclonic, absence, or unclassified seizures. Several anti-seizure medications, steroids, and a ketogenic diet were administered in these cases. However, seizures were controlled in only 50% of previously reported cases, while all of the six new cases became seizure-free after perampanel as an add-on treatment. The average time from the addition of perampanel to seizure control was 7.33 ± 4.59 months (range, 1–12 months). The median time to seizure freedom was 14 months (1–32 months, > 19 months in 3 cases). The average dosage of per-ampanel needed for epilepsy control was 0.22 ± 0.17 mg/kg per day. Conclusions In this paper, we comprehensively summarized the clinical characteristics, treatments and outcomes of NR2F1 -related epilepsy for the first time. Perampanel exhibits dramatic efficacy for NR2F1 -related epilepsy. This will help optimize the treatment of this type of epilepsy and provide clues for its pathogenic mechanisms. The two novel mutations expand the genotype spectrum of this disease.

Epilepsy caused by NR2F1 mutations was first reported in 2015, which expanded the clinical manifestations of BBSOAS [2].To date, various types of epilepsy have been identified in patients with BBSOAS, including infantile epileptic spasms syndrome (IESS)/West syndrome [2][3][4][5][6][7][8].Unfortunately, these NR2F1-related seizures are resistant to multiple anti-seizure medications (ASMs) and there were incomplete epileptic data at the time of reporting [5][6][7][8].The presence of early-onset epileptic encephalopathy in infancy (primarily with IESS) is of particular importance as it is always associated with poor neurodevelopmental outcomes.Previously, ASMs for NR2F1related epilepsy were selected according to the epilepsy syndrome and seizure type owing to the lack of specific and effective drugs.
In this study, we summarized the characteristics of 20 cases of NR2F1-related epilepsy with detailed epilepsy records (including our six new cases and 14 previously reported cases) to analyze the epileptic symptoms and treatment outcomes.In all of our six cases, perampanel effectively controlled the NR2F1-related epilepsy.The NR2F1 pathogenic variants in our six cases are also provided, including two novel mutations, which expand the clinical phenotypes and genotypes of NR2F1-related epilepsy in BBSOAS.

Participants
Six Chinese patients with epilepsy carrying NR2F1 mutations were diagnosed with BBSOAS in our epilepsy clinic from 2018.Complete clinical data of epilepsy were available from all of the patients, including epilepsy characteristics, electroencephalogram (EEG) data, brain MRI scans, treatment procedure and prognosis.Another 14 cases of NR2F1-related epilepsy reported in literature were included in this study with detailed epileptic data, including epilepsy characteristics, EEG data, brain MRI scans, and sequencing information (Fig. 1).

Ethics approval and consent to participate
This study was approved by the Clinical Research Ethics Committee of Peking University First Hospital.Written informed consent was obtained from the participants and their parents.

Evaluation of ASM efficacy
Outcomes of ASM treatment were classified as seizurefree (seizures controlled completely, i.e., absence of hypsarrhythmia on EEG), ASM-responsive (seizures decreased by ≥ 50%), and no effect (seizures decreased by < 50%).In this retrospective study, the medication efficacy was assessed at 12 months after ASM treatment.

Variant analysis
Five milliliters of peripheral venous blood were collected from patients and their parents.Genomic DNA was extracted for trio-based whole exome sequencing (WES).The pathogenicity of the variants was predicted by more than two bioinformatics programs, such as SIFT, PolyPhen-2, CADD, Protein Variation Effect Analyzer (PROVEAN), and MutationTaster, based on the American College of Medical Genetics and Genomics guidelines in 2015.Variants were evaluated against the general population using the ExAC, gnomAD, and dbSNP databases (Chigene company, Beijing, China).Sanger sequencing was performed for variant validation, and segregation analyses were performed following standard protocols.

Statistical analysis
Continuous variables are reported as the mean ± standard deviation.Categorical variables are presented as frequencies and percentages.All analyses were conducted using the SPSS 19.0 software.

Demographics
The mean age of the six cases was 46.5 ± 29.83 months (range, 20-104 months), and the male-to-female ratio (See figure on next page.)Fig. 1 EEG recordings before and after administration of perampanel to Patients 1, 2, 3 and 5. a: In patient 1, before perampanel administration, bursts of high-amplitude spikes, multiple spikes and spike-slow waves, and multiple clinical attacks were detected (with a spasm of the whole body trunk) (age 5 months).At 15 months after perampandel administration, low-amplitude fast waves, spikes and spike-slow waves in were detected in the right middle and posterior temporal regions during sleep (age 2 years and 5 months).b: In patient 2, before perampanel administration, EEG recording detected hypsarrhythmia and bilateral discharges primarily in the posterior region (4 months); at 6 months after perampanel administration, no hypsarrhythmia, spike-slow waves, or sharp-slow waves were detected in the bilateral posterior regions (10 months).c: In patient 3, before perampanel administration, EEG recording detected spike-slow/sharp-slow waves, polyspike-slow/slow waves mainly in the posterior and the midline areas.Isolated and series of spasms as well as mild tonic attacks were detected (age 7 years and 6 months).At 4 months after perampanel administration, EEG recorded spike-slow/sharp-slow, slow, sharp, and spike waves mainly in the posterior and the midline areas (age 8 years).d: In patient 5, before perampanel administration, EEG recorded multifocal slow, spike and sharp waves, and one attack of focal origin (age 8 months).At 10 months after perampanel administration, 6-7 Hz waves with medium-to-high amplitudes and 15-20 Hz fast waves with a large number of continuous discharges were detected in the bilateral occipitotemporal regions (age 2 years and 7 months) Fig. 1 (See legend on previous page.) was 5:1.The detailed clinical information is summarized in Table 1.

Epilepsy
To gain a better understanding of NR2F1-related epilepsy, we included only BBSOAS patients with seizures in our epilepsy clinic.The six patients had a mean epilepsy onset age of 3.08 ± 1.96 months (range, 0.5-6 months).The seizure types mainly included focal seizures, infantile epileptic spasms (IES), myoclonic seizures, atonic attacks, and generalized tonic seizures (GTS) (Table 2).IES, which is difficult to control, was observed in all six patients, with onset before 6 months of age.
We reviewed all previously published BBSOAS cases to collect detailed clinical information (Additional file 1: Fig. S1; Table 3).One hundred and twelve BBSOAS patients with NR2F1 mutations had been reported previously, of whom 45 BBSOAS patients were documented to have seizure attacks.However, only 14 individuals had available information on seizure type, seizure onset age, brain MRI scans, and EEG recording.The outcomes of epilepsy treatment were reported in eight of the 14 cases, and 4 of the 8 patients reported effective treatment with ASMs.Together with the six patients in our epilepsy clinic, a total of 20 BBSOAS patients were included for review (Table 3).The seizure types included IES, GTS, generalized tonic-clonic seizure (GTCS), focal seizures, myoclonus, absence, atonic, and some undetermined seizure patterns.Thirteen patients (65%) had IES, and the mean age of IES onset was 3.50 ± 1.54 months (range, 0.5-9 months).
Since IES was the primary seizure type present in all of the six new cases, adrenocorticotropic hormone (ACTH), vigabatrin, vitamin B6, and topiramate (TPM) were first considered for treatment, followed by valproate (VPA), clonazepam (CZP), levetiracetam (LEV), and lamotrigine (LTG).A ketogenic diet was administered.The seizures were uncontrolled until the addition of perampanel (Fig. 1, Additional file 1: Fig. S4).After the addition of perampanel, all six (100%) patients became seizure-free.The average time from the addition of perampanel to seizure control was 7.33 ± 4.59 months (range, 1-12 months).The average time from perampanel addition to the appearance of effect was 4.5 ± 2.66 weeks.The median time to the seizure freedom status was 14 months (1-32 months, 3 cases > 19 months).The median followup time was 26 months (range: 11-32 months).The average perampanel dosage for epilepsy control was 0.22 ± 0.17 mg/kg per day.The DD/ID symptoms were improved in all six patients after seizure control.
According to the literature, ASMs used for NR2F1epilepsy included VPA, oxcarbazepine, pyridoxine, nitrazepam, steroids, CZP, phenobarbital, phenytoin, TPM, LEV, LTG, and clobazam.Only four of eight patients (50%) with recorded ASM treatment became seizure-free.The effective ASMs in these four patients were VPA and ACTH, and the seizure types were GTCS, myoclonic seizures, absence seizures, and IES.Moreover, seven cases of IES have been reported previously, but only two had detailed treatment information.In one patient, the IES was controlled by ACTH application, whereas for the other patient, the seizures were uncontrolled and the detailed treatment data were not provided.
Vision impairment was identified in all the patients by professional ophthalmologists.Three patients displayed optic atrophy (OA) or a pale/small optic disc.Four patients showed a prolonged latency and/or decreased amplitudes of visual-evoked potential (Table 1).
All of the six patients manifested moderate-to-severe DD/ID.Congenital developmental delay deteriorated after seizures, especially infantile spasms, and was improved after seizure control.Speech was affected profoundly as patients 3 and 5 could only make a sound resembling "mama" or "baba", and the other patients could just make babbling sound.The DD/ID symptoms of the 6 patients were all improved after seizure control.Patient 1 could sit and stand independently after seizure control.Patients 2 and 3 could not turn body over by themselves, but right after seizure control, they could sit independently and their cognitive ability was greatly improved.Patients 4 and 6 could walk with help, crawl independently and understand some simple commands after seizure control.
ASD is a major clinical comorbidity.Due to the profound DD/ID and young age, formal ASD evaluation was not applied, but ASD symptoms were observed in three of the six patients.Hypotonia was observed in all patients.Oromotor dysfunction, high pain tolerance, and common facial features such as prominent ears, epicanthal folds, a tall forehead, a thin upper lip, and downturned mouth corners were also observed in the six patients.

Brain MRI
Brain MRI revealed a thin corpus callosum in two of our patients and non-specific slightly delayed myelination in one patient (Additional 1: Fig. S2).The other three cases had normal MRI finding.

Table 1 Detail clinical data of the six BBSOAS patients in our clinic
NA not available, ID intellectual disability, IES infantile epileptic spasms, AS atonic seizure, GTS generalized tonic seizure, DM delayed myelination, DD developmental delay, ASD autism spectrum disorder, P/SOD pale/small optic disc, VEP visual evoked potentials, OA optic atrophy

NR2F1-related genotyping
Five de novo heterozygous NR2F1 mutations in the DNA-binding domain (DBD) were identified in our cases by trio-whole exome sequencing, including four missense mutations (at highly evolutionarily conserved residues across species) and one deletion mutation; two novel mutations were detected, including c.365G > T, p.Cys122Phe in one patient (Patient 1) and c.449G > T, p.Gly150Val in two patients (Patients 4 and 5) (Table 4, Fig. 2, Additional file 1: Fig. S3).Combined with the cases previously reported, there were 13 NR2F1-related patients with IESS (65%, 13/20).Twelve patients with IESS had mutations located in the DBD (nine missense mutations and three deletions), whereas the remaining patient had an 8.4-Mbp deletion in 5q14.3-15[87,086298-95538640-699] (Table 1).The three patients found with novel mutations all had IES, severe DD, visual problems, and swallowing disorders.Patient 6 also had a clear tendency towards selfharming and was extremely emaciated to need a gastric tube feeding.

Discussion
NR2F1 was first linked to OA with intellectual disability in 2014.This syndrome was named BBSOAS in a subsequent study.Although NR2F1-related epilepsy was first reported in 2015, most studies on NR2F1 mutationrelated BBSOAS have focused on visual impairment, ID/ DD and psychobehavioral disorders.Less attention has been paid to related epilepsy.
Our results showed that the NR2F1-related epilepsies were mostly IESS (65%), which generally occurs within six months of life and is difficult to control.This type of epilepsy is always accompanied by visual impairment and ID/DD.Therefore, for patients with triad presentation including epilepsy, ID/DD, and vision defects, NR2F1 sequencing should be conducted in a timely manner for early precision treatment.
We administered perampanel to patients 1, 2, 3, 4, and 5 who presented with seizures resistant to multiple treatments, and this drug showed dramatic efficacy in reducing the seizures.Based on this experience, the following patient 6 was treated with perampanel as an initial ASM when he was diagnosed with NR2F1-related epilepsy (DD, visual impairment, infantile spasm and WES presenting NR2F1 de novo variant).This patient reached seizure-freedom with perampanel treatment alone within 12 months, and the seizure-free status remained 9 months later.In addition, we tried to withdraw ASMs except perampanel in patients 1, 2, 3, 4 and 5 after reaching the seizure-free status.The seizures relapsed in two patients during TPM withdrawal; therefore, we maintained TPM with perampanel in patients 1, 2, 4 and 5 while withdrawal of other ASMs continued.No recurrence or worsening of seizures occurred in these five cases.This indicates that perampanel combined with TPM may be an effective combinational regimen for NR2F1-related epilepsy.
We reviewed recent studies on the efficacy of ACTH, oral steroids or vigabatrin, which are the first-line treatments for IESS.In 2022, the National Infantile Spasms Consortium conducted a large-scale retrospective study on pediatric spasms in 23 medical centers in the United States.ACTH, oral steroids or vigabatrin was applied in 205, 99 or 91 children with IESS, respectively.Ultimately, the highest clinical remission rate was found at 30 days after treatment, which was, however, only 48% [14].The latest randomized controlled trial on the standard treatment of IESS reported that, at 30 days after two-week treatment, 9/12 (75%) of the IESS children were controlled with ACTH, 1/9 (11%) with vigabatrin and 5/13 (38%) with combined therapy [15].In 2018, 66 newly diagnosed patients with infantile spasms underwent sequential treatment and long-term follow-up.After 7 months of follow-up, a total of 48 patients (72.7%) were controlled [16].However, there was no systematic research on the treatment of BBSOAS-related IESS epilepsy.In this retrospective study, we set 12 months as the time point for assessing medication efficacy, which is longer than the time frame in traditional studies of IESS.However, the pathology of BBSOAS-relate IESS is distinct from the general IESS, and vigabatrin and hormones can cause more severe long-term side effects than perampanel.Thus, considering the high seizure control rate in our patients (100%) compared with traditional treatments for corresponding epilepsy syndrome (e.g., ACTH and vigabatrin for IESS), we suggest that perampanel could be considered as the first-choice ASM for NR2F1-related epilepsy.Large prospective studies on the treatment of the BBSOAS-related IESS are needed in the future.
The pharmacological mechanism underlying the effectiveness of perampanel may be related to the pathogenicity of NR2F1 mutations.NR2F1 has been verified to participate in many aspects of early nervous system development [17][18][19][20][21][22][23][24].A review summarizing all the functional effects of reported NR2F1 mutations proposed that NR2F1 mutations identified in patients with BBSOAS mainly result in haploinsufficiency of the Nr2f1 protein [3,4,20].Moreover, NR2F1 variants in the DBD domain of the protein are thought to cause a greater loss of the transcriptional regulation function, possibly due to the dominant negative effect [20].In 2020, Del Pino et al. identified that the loss of Nr2f1 protein in cortical progenitors resulted in increased intrinsic excitability in a mouse model [22].Bertacchi  et al. found that the brains of NR2F1-knockout mice display increased cortical plate thickness [8].Conditional inactivation of Nr2f1 in mouse interneuron precursors results in decreased numbers of interneurons and aberrant migration [21].Moreover, Teratani-Ota et al. reported that the majority of neuron-like cells generated from embryonic stem cells by Nr2f1 induction are GABA-positive and express other GABAergic neuronal markers [23].These findings indicate that the normal expression of Nr2f1 is important for maintaining normal neurogenesis of GABAergic neurons and optimal excitement of cortical neurons.Thus, mutations of NR2F1 may affect neurogenesis and the functions of GABAergic neurons, leading to an excessive increase in cortical excitability, and finally epilepsy.However, the first "patient-specific" NR2F1-R112K mutant mice generated by Zhang et al. [24] only recapitulated ASDlike traits while no epilepsy phenotype was observed.In summary, previous studies have indicated that the NR2F1-related epilepsy is primarily attributed to haploinsufficiency of the Nr2f1 protein caused by NR2F1 variants, as Nr2f1 plays a role in maintaining the normal number and migration of interneurons, as well as the normal excitability of excitatory neurons in brain development.Consequently, decreased functionality of the Nr2f1 protein may lead to abnormal excitability of excitatory neurons as well as decreased number and aberrant migration of interneurons.Perampanel selectively inhibits glutamate receptors [25,26].TPM is thought to exert antiepileptic effects through enhancement of GABAergic activity and inhibition of kainate/AMPA-type glutamate receptors [27].The effectiveness of perampanel and TPM in our cases further supports the hypothesis that NR2F1-related epilepsy is due to the increased excitability and number of pyramidal neurons, as well as decreased number and aberrant migration of interneurons.To elucidate the mechanism of NR2F1-related epilepsy, experimental models with patient-specific NR2F1 mutations, such as knock-in mouse models and brain organoids originating from patient-induced pluripotent stem cells, are needed.
A limitation of this study was that it was an observational and retrospective study without a randomized control group.We only recruited BBSOAS patients with NR2F1-related epilepsy; therefore, the number of patients recruited was relatively small.Moreover, all the six patients in our clinic suffered from IES.In future studies, patient data from multiple hospitals are needed to reduce the sampling error and focus on more types of NR2F1related intractable epilepsy.Second, we did not conduct any experiments to verify our hypothesis that the effectiveness of perampanel on NR2F1-related epilepsy is due to the enhancement of pyramidal neurons and weakening of interneurons.Experimental models carrying patientspecific NR2F1 mutations may be needed to fully explore the underlying mechanisms of NR2F1-related epilepsy.

Conclusions
In summary, in this paper we summarize the clinical characteristics of NR2F1-related epilepsy, including treatments and outcomes.Infantile epileptic spasms was the most common seizure type of NR2F1-related epilepsy.Although NR2F1-related epilepsy was previously resistant to multiple anti-seizure medications and steroids, in our patients, perampanel exhibited dramatic effects on NR2F1-related epilepsy.This finding will help optimize the treatment of this type of epilepsy and provide insights into the pathogenesis of this epilepsy.We also reported two novel mutations, c.365G > T, p.Cys122Phe and c.449G > T, p.Gly150Val, which expand the genotype spectrum of this disease.

Fig. 2
Fig. 2 Distribution of the pathogenic variants in the Nr2f1 protein revealed in our cases.Two novel mutations are marked in red, including p. Cys122Phe in patient 2 and p.Gly150Val in patients 5 and 6.Abbreviations: DBD: DNA binding domain; LBD: ligand-binding domain

Table 2
Seizure information of the six patients with BBSOAS in our clinic

Table 2 (continued) Patient Age at last follow-up Age at SZ onset Seizure type Brain MRI EEG ASMs used ASMs in use at last follow-up
SZ seizure, IES infantile epileptic spasms, AS atonic seizure, GTS generalized tonic seizure, DM delayed myelination, ASMs anti-seizure medications, TPM topiramate, VPA valproate, CZP clonazepam, LEV levetiracetam, LTG lamotrigine, ACTH adrenocorticotropic hormone, KD ketogenic diet

Table 3
Seizure information of the 20 BBSOAS patients Patient (age/sex)