Neurodevelopmental Disorders Due to Specific Genes (GRIN1, KDM6B, TBR1, SCAF4, SPTBN1, RERE, CLCN3)

Seven monogenic neurodevelopmental disorders, almost always due to de novo variants, together with the classes of the GRINopathies (GRIN2B, GRIN2D) and the endolysosomal CLC exchangers (CLCN4).

Literature and regulatory information updated: 18 October 2026.

0. Initial Reference Table

Seven entities are brought together here that share the same mode of diagnosis, by broad sequencing in a child with developmental delay, and the same management logic, symptomatic and multidisciplinary, with the exception of the targeted therapy under study for the GRIN class. Each gene has a uniform structure (identification, inheritance, mechanism, phenotype, frequency, distinguishing signs, management and specific therapy). The genes GRIN2B and GRIN2D and the gene CLCN4 are treated as members of the classes of GRIN1 and CLCN3; GRIN2A-related disorder, developmental and epileptic encephalopathies (DEE) in general and the general approach to intellectual disability (ID) are not developed.

0.1. Identification of the entities

0.2. Essential clinical data

1. Introduction

1.1. Scope and rationale of the grouping

The seven disorders in this article were delineated through the same methodological route, the "genotype-first" approach: exome or genome sequencing in large series of children with developmental delay, ID or ASD identified recurrent de novo variants in a gene, and only afterwards did the international gathering of patients, through platforms such as GeneMatcher, the Deciphering Developmental Disorders (DDD) project or the Undiagnosed Diseases Network (UDN), define the phenotype [13, 18, 27]. None of the entities has a facial gestalt that allows reliable clinical recognition, and the authors of the largest cohorts conclude that the diagnosis is molecular [13, 18]. They are grouped here for that reason, with separate subsections whenever genetics, phenotype or management differ. Three genes belong to families with other clinically relevant members: GRIN1 to the family of N-methyl-D-aspartate (NMDAR)-type glutamate receptors, whose diseases are termed "GRINopathies" [68]; CLCN3 to the family of endolysosomal chloride/proton exchangers, which includes CLCN4 [46, 64]; and SPTBN1 to the cerebral spectrinopathies [60, 69]. Following the class approach, the members GRIN2B, GRIN2D and CLCN4 are described here in proportionate depth; GRIN2A, associated with epilepsies with aphasia and with DEE, appears only in the class tables and in the targeted therapy common to the family.

1.2. Historical milestones by gene

For GRIN1, Hamdan and colleagues reported in 2011 an excess of deleterious de novo variants in genes of the glutamatergic system in non-syndromic ID [70]. Lemke and colleagues delineated in 2016 the phenotypic spectrum in 14 new and 9 published individuals, with profound delay, severe ID without speech, hypotonia, hyperkinetic movement disorder, oculogyric crises, cortical blindness, cerebral atrophy and epilepsy, and described the first homozygous variants in consanguineous families, proposing a distinct "NMDA receptor genetic encephalopathy" [71]. Fry and colleagues associated in 2018 de novo missense variants in the S2 domain and the M3 helix with extensive bilateral polymicrogyria, in an exome study of 57 trios with polymicrogyria from the United Kingdom supplemented by cohorts from France, the USA and a European brain malformation network [72]. The GeneReviews by Platzer and Lemke, published in 2019 and revised in 2021 only to incorporate data on parental mosaicism, brought together 72 individuals [1].

For KDM6B, Stolerman and colleagues described in 2019 twelve unrelated patients with de novo variants identified by exome sequencing, with developmental delay, ID, facial dysmorphism and other findings [73]. OMIM then created entry 618505 with the name "neurodevelopmental disorder with coarse facies and mild distal skeletal abnormalities"; Rots and colleagues assembled in 2023 a cohort of 85 individuals, showed that coarse facies (4%) and distal skeletal anomalies (9%) are rare and proposed the designation "KDM6B-related NDD"; the OMIM entry is now called "Stolerman neurodevelopmental syndrome" [12, 13].

For TBR1, the gene was proposed as a candidate in a boy with a de novo 5.3 Mb deletion at 2q24.2 (Magri and colleagues, 2011, cited by Nambot) and later emerged in the large exome studies of ASD; Deriziotis and colleagues showed in 2014 that the de novo variants of sporadic ASD disrupt subcellular localisation, interaction with co-regulators and transcriptional repression, unlike the missense variants inherited from unaffected parents [18, 74]. Nambot and colleagues described in 2020, through the French AnDDI-Rares network and international collaboration, 25 new individuals and reviewed 13 from the literature, and concluded that TBR1 is more correctly an ID gene with frequent autistic traits than an "autism gene" [18, 19].

For SCAF4, Fliedner and colleagues (Erlangen) started in 2020 from a de novo splice variant in a patient with mild ID and seizures and gathered, through GeneMatcher, 11 individuals, nine of whom had likely pathogenic variants [75]. OMIM assigned number 620511 to Fliedner–Zweier syndrome, and Schmid and colleagues added in 2025 fifty individuals from 41 families, with centres in Europe, the USA, Pakistan and Brazil [22, 23]. The second published report of SCAF4 came from the University of São Paulo [76].

For SPTBN1, Cousin and colleagues published in 2021 the defining cohort of 29 individuals from 28 families, assembled by exome or genome sequencing in diagnostic practice, GeneMatcher, UDN and DDD/DECIPHER, and, almost simultaneously, Rosenfeld and colleagues described seven further individuals [27, 69].

For RERE, Fregeau and colleagues identified in 2016 de novo variants in ten individuals with ID, delay and/or ASD and showed that the spectrum overlaps that of 31 individuals with isolated 1p36 deletions including the gene [77]. Jordan and colleagues brought together nine individuals in 2018 and established genotype–phenotype correlations, the basis of the 2019 GeneReviews with 19 patients [31, 78]. Curtis and colleagues published in 2026 a cohort of 54 individuals with new correlations and management recommendations [62].

For CLCN3, Duncan and colleagues described in 2021, through MatchMaker Exchange, eleven individuals: nine with eight heterozygous missense variants and two siblings homozygous for a frameshift variant [64]. Nakashima and colleagues added in 2023 two boys with de novo variants at residue Gly327, and Polovitskaya and colleagues extended the cohort in 2026 to 26 individuals and identified the gain-of-function mechanism through the TMEM9 subunit [63, 79]. For the class member CLCN4, the families MRX15 (Raynaud, 1996) and MRX49 (Claes, 1997) preceded identification of the gene by Hu and colleagues in 2016 [46].

1.3. Nomenclature and reclassifications

Several designations have changed and it is useful to know them in order to read older reports. OMIM entry 614254 for GRIN1 still appears in databases as "mental retardation, autosomal dominant 8" (MRD8), a name replaced by NDHMSD [2, 3]. In 2022 OMIM separated from the recessive presentations with severe ID and involuntary movements (NDHMSR, 617820) DEE101 (619814), reserved for biallelic nulls with lethal neonatal encephalopathy [9, 11]. Entry 618505 for KDM6B changed from a morphological description to Stolerman's eponym [12]. For TBR1, GeneGuidance also associates OMIM number 606053 with the Orphanet 2q24 microdeletion (ORPHA:1617); OMIM confirms that 606053 is IDDAS due to a heterozygous variant in TBR1, and Orphadata lists TBR1 only as a candidate gene tested in the microdeletion, which is another entity, with dopa-responsive dystonia and parkinsonism [17, 21]. The recessive form of CLCN3 (NEDSBA) retains the "?" prefix in OMIM, because it rests on a single family [37, 38]. Finally, the OMIM designation for CLCN4 is Raynaud–Claes syndrome, with X-linked dominant inheritance, and Orphanet retains only an obsolete entry [47].

1.4. Definition of the entities within scope

GRIN1-related disorder is defined as developmental delay or ID caused by a pathogenic or likely pathogenic de novo heterozygous missense variant, or by biallelic missense or truncating variants, in GRIN1 [1]. The others are defined in the same way by gene and mode of inheritance: a heterozygous loss-of-function variant or a missense variant in the functional domains of KDM6B [13]; a heterozygous variant, mostly de novo, in TBR1, distinguishing it from the larger 2q24.2 deletions that include other genes [18]; a heterozygous loss-of-function variant in SCAF4 [23]; a heterozygous variant in SPTBN1 [27]; a heterozygous pathogenic or likely pathogenic variant in RERE [31]; and, for CLCN3, a de novo heterozygous missense variant (NEDHYBA) or biallelic loss of function (NEDSBA) [64].

2. Molecular Genetics and Pathophysiology

2.1. GRIN class: the NMDA receptor and GRIN1

Glutamate is the main excitatory neurotransmitter of the central nervous system (CNS), present at about 90% of excitatory synapses, and acts on metabotropic receptors (mGluR) and on ionotropic receptors of the NMDA, AMPA and kainate types; NMDARs are critical for neuronal survival, circuit formation and synaptic plasticity [68]. Each NMDAR is a heterotetramer of two GluN1 subunits, which bind glycine or D-serine, and two GluN2 subunits (encoded by GRIN2A to GRIN2D), which bind glutamate; opening of the cation pore requires binding of both agonists and membrane depolarisation, because Mg²⁺ blocks the channel in a voltage-dependent manner, and allows Ca²⁺ influx [1, 3]. GluN1 is the obligatory subunit of all NMDARs and is expressed ubiquitously from the embryo to adulthood, with eight isoforms generated by alternative splicing; the clinically relevant transcript has 20 exons and encodes a protein of 938 amino acids with an amino-terminal domain (ATD), two segments of the agonist-binding domain (S1 and S2), four transmembrane segments (M1 to M4), a calmodulin-binding domain and a C-terminal domain [1]. The GluN2B and GluN2D subunits are expressed mainly in the prenatal period, whereas GluN2A and GluN2C increase after birth and GluN2B becomes progressively restricted to the forebrain, which helps to explain the early severity of GRIN2B and GRIN2D variants [39]. The genes GRIN2C, GRIN3A and GRIN3B had no associated Mendelian disease at the time of the GeneReviews [43].

In the dominant form of GRIN1 only de novo missense variants have been described, clustered in or near the S2 domain and the M1 to M4 segments; no de novo truncating variant has been associated with disease [1, 71]. The reason lies in the recessive form: parents heterozygous for null variants are healthy, so GRIN1 haploinsufficiency causes no neurological phenotype, and the dominant variants act through a dominant-negative effect or gain of function [11, 71]. Functionally, some missense variants reduce receptor function and others increase it [1, 80]: Lemke and colleagues described in Xenopus oocytes mainly loss of function of variable severity with a dominant-negative effect, and Fry and colleagues showed that three variants associated with polymicrogyria increase agonist potency and one reduces proton inhibition, that is, gain of function [71, 72]. In the recessive form, three families have homozygous missense variants in the ATD and one family has the homozygous nonsense variant NM_007327.4:c.1666C>T p.(Gln556Ter), with a non-functional channel [1, 11, 81]; a Pakistani infant of consanguineous parents with a homozygous splice variant was the second case of DEE101 described in OMIM [11].

Functional classification of GRIN missense variants as gain of function (GoF) or loss of function (LoF) has become the therapeutic axis of the class. Myers and colleagues proposed in 2023 an analytical framework that integrates six in vitro assays (sensitivity to glutamate and glycine, sensitivity to the endogenous modulators Mg²⁺, protons and zinc, membrane trafficking, time course of the response and open probability) with confidence thresholds, to estimate the net effect on receptor-mediated charge transfer [82]. In GRIN2A, missense variants in the transmembrane domains and linkers produce gain of function and severe infantile-onset DEE, and those in the ATD and the agonist-binding domain (LBD) produce loss of function and speech disorders and/or seizures with mild or no ID [83, 84]. In GRIN2B, null alleles (nonsense, frameshift, exonic or whole-gene deletions, translocations and inversions) are associated with mild or moderate ID and missense variants with severe ID (Fisher's exact test, p = 0.0079) [39, 85]; in a review of 105 patients, transmembrane domain variants were predominantly GoF, GoF variants were associated with more epilepsy and microcephaly, and LoF missense variants were more severe than truncating LoF variants [86]. In GRIN2D all causal variants are de novo missense, with no null variants or large deletions associated, and with a loss- or gain-of-function effect; the regions controlling channel gating (pre-M1 helix, SYTANLAAF motif of M3 and M4/pre-M4) are the same ones whose variants cause disease in GRIN1, GRIN2A and GRIN2B [43].

The most robust genotype–phenotype correlation in GRIN1 is that of malformations of cortical development (MCD): the de novo variants of patients with MCD are located in the S2 and M3 domains, with the recurrent NM_007327.4:c.1975C>T p.(Arg659Trp), in the DRPEER motif of S2, and NM_007327.4:c.2365G>A p.(Asp789Asn) [81]; these regions are rarely mutated in the population and in patients without polymicrogyria [1, 72]. A study of 25 patients with NMDAR-linked cortical malformations showed clinical and imaging overlap between GRIN1 and GRIN2B, with diffuse dysgyria and dysmorphic basal ganglia and hippocampi, and the neuropathology of two GRIN1 fetuses revealed impaired proliferation and radial and tangential migration; homozygosity in GRIN1 produced microcephaly with a simplified gyral pattern [87]. The family with the homozygous nonsense variant had three children with fatal DEE, who died between 5 days and 5 months of age, a course not observed in the five people with homozygous ATD missense variants or in any de novo heterozygote [1]. Penetrance of the dominant variants of GRIN1, GRIN2B and GRIN2D is considered to be 100%, and there are no known allelic disorders of GRIN1 [1, 39, 43].

The glutamate receptor class also includes AMPA, kainate and metabotropic genes. The following table brings together the members with disease described in the sources, with the feature that distinguishes them.

2.2. KDM6B: histone H3K27 demethylase

KDM6B, also called JMJD3, encodes a JmjC-domain histone demethylase that removes di- and trimethylation of lysine 27 of histone H3 (H3K27me2/me3), a repressive chromatin mark; its action derepresses developmental genes, including HOX genes, is involved in macrophage differentiation in inflammation and regulates, through a role independent of demethylase activity, T-box factor-dependent transcription via the SWI/SNF complex [14]. The cohort of Rots and colleagues brought together 71 distinct variants in 85 individuals: 73 predicted protein-truncating variants (PTV; nonsense, frameshift and canonical splice) and 12 protein-altering variants (PAV; missense and in-frame indels) [13]. PTVs are distributed throughout the gene, including the last and penultimate exons, and three of these, which escape nonsense-mediated mRNA decay (NMD), were considered likely pathogenic because they eliminate the zinc-binding domain; 18 of 21 PAVs cluster at the C-terminus, in the JmjC domain (amino acids 1157–1485) or the zinc-containing domain (1563–1620) and the stabilising segments (1490–1558 and 1623–1635), with corrected p of 2.0 × 10⁻⁶ [13]. In a dual gain-of-function assay in Drosophila, eight of nine JmjC PAVs showed loss of function, and in the zinc domain the effect was variable: p.(Cys1575Ser) with complete loss, NM_001348716.2:c.4696C>A p.(Arg1566Ser) and c.4708G>C p.(Glu1570Gln) with reduced function and c.4750A>C p.(Asn1584His) without loss (c. numbering confirmed in ClinVar [81]); eleven PAVs were classified as pathogenic or likely pathogenic and ten remained variants of uncertain significance (VUS) [13]. The established mechanism is haploinsufficiency (ClinGen haploinsufficiency score 3 and triplosensitivity score 0) [14]. The Drosophila orthologue (Utx) is required for short- and long-term memory and regulates activity and sleep [13]. KDM6B is enriched for de novo variants in about 31,000 individuals with neurodevelopmental disorders and 11,000 with ASD, and rare de novo variants of the gene were associated with schizophrenia in another study; Rots and colleagues liken this trajectory (ID in childhood, psychosis in adolescence) to that of SETD1A, KMT2C, SRCAP and EHMT1 [13]. A benign control variant, NM_001348716.2:c.1244C>A p.(Pro415Gln) [4], has a frequency of 3.62 × 10⁻⁵ in gnomAD, a reminder that not every rare missense variant is pathogenic [13].

2.3. TBR1: cortical T-box transcription factor

TBR1 encodes a brain-specific T-box transcriptional repressor, expressed in glutamatergic pyramidal neurons, with maximal expression in the preplate and layer 6, from which corticothalamic axons arise; it regulates neuronal migration, laminar and areal identity and axonal projection, and represses FEZF2, restricting the origin of corticospinal axons to layer 5 [18, 19]. It is part of the PAX6–TBR2–NEUROD–TBR1 cascade, forms with CASK a nuclear complex that induces RELN, activates AUTS2, directly controls GRIN2B in mature neurons and homodimerises and interacts with FOXP1, FOXP2 and BCL11A [18]. De novo truncating variants disrupt transcriptional repression, subcellular localisation, homodimerisation and interaction with these partners, and de novo missense variants have milder effects but with an impact on function and dosage; pathogenic variants disrupt the TBR1–FOXP2 interaction, which relates to the language deficit [18, 74]. In the homozygous null mouse there is abnormal differentiation of early cortical neurons and severe disruption of lamination, with a thin cortex and reeler-type "mirror-image" lamination; in the heterozygote there are abnormal amygdalar axonal projections and deficits in social behaviour, cognitive flexibility and associative memory [18]. The gene is intolerant of loss of function (pLI = 1; 1 variant observed versus 24.5 expected in gnomAD) and of missense variation (z = 3.64), and the ClinGen haploinsufficiency score is 3 [18, 19].

The mutational spectrum of the 38 individuals reviewed by Nambot includes 29 different single-nucleotide variants (10 frameshift, 6 nonsense, 1 in-frame deletion and 10 missense) and deletions [18]. The relevant domains are the DNA-binding T-box (amino acids 206–393), the T-box-associated and dimerisation domain (418–680), the CASK-interacting region (342–681) and the FOXP2-interacting region (213–393); 20 variants lay in the T-box and 13, all frameshift, in the associated domain [18]. The recurrent last-exon variant, in eight individuals, seven of them with severe delay or ID, corresponds on the MANE transcript to NM_006593.4:c.1588_1594dup p.(Thr532ArgfsTer144), recorded in ClinVar as pathogenic or likely pathogenic [18, 90]. Three different missense variants affect residue p.Trp271 (two p.(Trp271Arg), with severe ID, autistic traits and seizures; p.(Trp271Ser), with moderate ID without autism or seizures; p.(Trp271Cys), with ID and autism), which illustrates variable expressivity at the same codon [18]. The deletions include a 122 kb deletion containing only TBR1 and a de novo deletion at 2q24.2 of up to 549 kb with TANK, PSMD14 and TBR1; a mosaic deletion of 0.422 Mb (about 61% of cells) was described in an 18-year-old man with mild ID and short stature [18]. Biallelic variants have never been described in living humans [18].

2.4. SCAF4: antitermination of RNA polymerase II transcription

SCAF4 encodes a serine- and arginine-rich (SR) factor that associates with the phosphorylated C-terminal domain (CTD) of the largest subunit of RNA polymerase II and acts as an antiterminator: it binds nascent RNA upstream of early polyadenylation sites, suppresses premature cleavage and polyadenylation and prevents the production of truncated proteins, a function it shares with SCAF8 [24, 75]. The gene has 20 exons and encodes 1147 amino acids with an N-terminal CTD-interacting domain (CID), an SR domain, an RNA recognition motif (RRM) and a proline- and glutamine-rich domain [75]. The nine likely pathogenic variants of the original description (two splice and seven truncating) lay in the N-terminal two thirds, and NMD degradation was demonstrated by RT-PCR for p.(Ser434Ter) (published designation, NM_020706.2); of the 14 truncating variants in gnomAD, ten were in the last exon, where they escape NMD and are probably benign [75]. In the 2025 series, the 39 variants included an intragenic deletion of exons 1 to 12, 25 truncating variants distributed throughout the gene, 8 splice and 5 missense (in the CID p.(Leu11Pro), p.(Ser13Leu) and p.(Gln118Pro), published designations on NM_020706.2, and in the RRM NM_020706.2:c.1621C>G p.(Pro541Ala) [81]); the missense variants and two other variants remained classified as VUS, and the missense phenotype was indistinguishable from the truncating one [23]. The gene is intolerant of loss of function (pLI = 1; observed/expected LoF ratio 0.19 and LOEUF 0.29 in gnomAD v4) and weakly constrained for missense variation (z = 1.48) [23]. RNA sequencing of blood from four patients revealed 9038 differentially expressed genes and differential splicing of 2942 genes, with shorter transcripts in more than 70%, reflecting a defect of mRNA processing; knockdown of the orthologue CG4266 in Drosophila caused locomotor, learning and short-term memory deficits and an increase in active zones at the neuromuscular junction [75]. These RNA data are research findings and do not constitute a validated clinical assay [24]. In 19 individuals of the 2025 series there were additional variants in other genes (for example, 9q33 deletion or TLK2), which the authors interpret as a possible oligogenic contribution [23].

2.5. SPTBN1 and the cerebral spectrinopathies

SPTBN1 encodes the ubiquitous βII-spectrin, which forms with αII-spectrin the membrane-associated cytoskeletal meshwork; it binds actin and calmodulin in a Ca²⁺-dependent manner, anchors transmembrane proteins and positions organelles [28]. The protein has two calponin homology domains (CH1 and CH2), which bind filamentous actin (F-actin), a series of spectrin repeats (SR), of which SR1 is required for dimerisation with αII-spectrin, and a C-terminal pleckstrin homology (PH) domain, with ankyrin-binding and tetramerisation sites [27]. In Cousin's cohort, the 28 unique variants were 22 missense, 3 nonsense and 3 canonical splice, with recurrent codons Gly205, Thr268, Arg411 and Arg1003; about half lay in the CH2 domain [27]. In HEK293T cells, CH domain variants (for example, NM_003128.3:c.614G>A p.(Gly205Asp) and NM_003128.3:c.749T>G p.(Leu250Arg) [81]) formed cytosolic aggregates that sequester F-actin and αII-spectrin, suggesting a dominant-negative effect, and several Thr268 and SR variants increased membrane protrusions, an alteration of cytoskeletal dynamics compatible with gain of function or a dominant-negative effect; the variants also disrupted axonal growth, dendritic arborisation and axonal transport of organelles [27]. The haploinsufficient mouse has hypoplasia of the corpus callosum and social deficits, and neural deletion is lethal by day 40, with hyperactivity and spontaneous seizures; the mechanism is thus mixed, with haploinsufficiency and dominant-negative or gain-of-function effects of CH domain missense variants [27]. Genetic constraint is high (missense Z = 4.54; pLI = 1; LOEUF = 0.09) [27]. ClinVar held 216 pathogenic and 47 likely pathogenic variants against 547 VUS, a number that reflects testing volume and not disease frequency [28].

In the spectrin family, SPTA1 and SPTB cause red cell disorders (elliptocytosis, spherocytosis and pyropoikilocytosis), SPTAN1 causes de novo epileptic encephalopathy with cerebral atrophy and progressive microcephaly, SPTBN2 causes spinocerebellar ataxia type 5 (SCA5, dominant) and a recessive ataxia (SCAR14), and SPTBN4 causes, in the biallelic state, congenital hypotonia, areflexia, neuropathy and profound delay [69]. A review of 91 patients in the β-spectrin family associated epilepsy with SPTBN5 (83.3%) and SPTBN1 (45.5%), ataxia with SPTBN2 (82.4%) and hypotonia with SPTBN4 (91.7%) [60].

2.6. RERE: nuclear receptor co-regulator in the proximal 1p36 region

RERE encodes a protein of the atrophin family, a nuclear receptor co-regulator in complexes that positively and negatively modulate transcription; it associates with histone deacetylase complexes, positively regulates retinoic acid signalling and, when overexpressed, recruits BAX to nuclear PML domains and activates caspase-3 [31, 77, 91]. The protein has 1566 amino acids (172 kDa) and BAH-MTA, ELM2, SANT, ZnF_GATA and atrophin-1 domains, the latter with two regions of arginine–glutamate dipeptide repeats and a 21-amino-acid histidine-rich region (HRR) (p.His1425 to p.Pro1445), fully conserved down to Xenopus and zebrafish [31, 78]. In the combined series of 19 patients, six (31%) had loss-of-function variants (partial deletions, nonsense, frameshift) and 12 (63%) point variants in the atrophin-1 domain, nine of them in the HRR [78]. Loss-of-function variants act through haploinsufficiency, including a 317 kb intragenic deletion of exons 1 to 10 detected by microarray; missense variants may produce a non-functional protein or have a dominant-negative effect [31]. Structural defects of the eye, heart and kidney and sensorineural hearing loss were significantly more frequent with atrophin-1 point variants than with loss-of-function variants (P = 0.0004) [78]. The recurrent two-amino-acid duplication NM_001042681.2:c.4313_4318dup p.(Leu1438_His1439dup), with identical c. numbering on NM_012102.3, was found in two individuals meeting criteria for CHARGE syndrome without a CHD7 variant [4, 78]. The Curtis cohort (54 individuals) separated five subgroups and confirmed that loss-of-function variants are associated with fewer multisystem anomalies and are more often inherited from a mildly symptomatic or asymptomatic parent, whereas missense variants and in-frame HRR duplications have more multisystem phenotypes and are generally de novo; HRR missense variants stabilise the structure, which points to gain of function or a dominant-negative effect [62]. Of the 54 individuals, 30 (from 25 families) were previously unpublished and were gathered through GeneMatcher, DECIPHER and personal contacts, and the cohort included pathogenic, likely pathogenic and uncertain significance variants [62]. Among the loss-of-function variants is a synonymous variant, NM_012102.4:c.1740G>A p.(Ser580=) (NC_000001.11:g.8364056C>T; identical numbering on MANE Select NM_001042681.2), which affects the last nucleotide of the exon and which the authors showed to cause loss of RERE function; it was inherited from the mother and present in an affected sibling [4, 62]. In an allelic series of mice built with a null allele (om) and a hypomorphic allele (eyes3, p.Val193Ala), animals with two deficient alleles had microphthalmia, high perinatal mortality, postnatal growth deficit, cerebral hypoplasia with fewer NeuN-positive neurons in the hippocampus, hearing loss, cardiovascular malformations (aortic arch anomalies, double-outlet right ventricle, transposition of the great arteries, perimembranous ventricular septal defect), spontaneous cardiac fibrosis and renal agenesis; ventricular myocardial fibrosis appeared in 2 of 3 mice aged 16–20 months and in 1 of 4 aged 3–5 months, and in no control, whereas Rere+/om heterozygotes did not differ from controls in prepulse inhibition and conditioned fear tests [92]. Subsequent models from the same group showed delay of the principal cerebellar fissures and of Purkinje cell migration and maturation, with abnormal foliation of the vermis and cerebellar hemispheres [93]; fewer mesenchymal cells in the atrioventricular endocardial cushions, due to reduced epithelial–mesenchymal transition, with a genetic interaction between Rere and Gata4 and lower GATA4 expression, a mechanism proposed for atrial and ventricular septal defects [94]; cleft palate in most deficient embryos on a C57BL/6 background, with delayed elevation of the palatal shelves when the gene is deleted in cranial neural crest cells [95]; and progressive loss of retinal cells, with apoptosis in the ganglion cell layer from embryonic day 17.5, leading to atrophy of the retina and optic nerve [96]. The Curtis cohort summarises these models as the biological basis of multisystem involvement [62]. The mouse with total loss of Rere dies at around embryonic day 9.5 with an open neural tube defect, cardiac failure and fusion of the telencephalic and optic vesicles, and the deficient zebrafish has microphthalmia and auditory changes [77]. RERE lies in the proximal critical region of the 1p36 deletion (chr1:8,395,179–11,362,893, hg19), and its haploinsufficiency probably explains many features of proximal 1p36 deletions [31, 77, 97].

2.7. CLCN3 and the class of endolysosomal CLC exchangers

CLCN3 encodes ClC-3, a strongly outwardly rectifying electrogenic chloride/proton exchanger located in early and late endosomes, lysosomes and the Golgi apparatus, where it contributes to endolysosomal ion composition and acidification, including that of GABAergic synaptic vesicles [36]. In the family of vesicular CLCs, ClC-3, ClC-4 and ClC-6 are in late endosomes, ClC-5 in early endosomes (CLCN5, Dent disease) and ClC-7 with Ostm1 in lysosomes (CLCN7, osteopetrosis and storage disease); a recurrent gain-of-function variant in CLCN6 causes severe global delay, hypotonia, respiratory failure and neurodegeneration [64]. ClC-4 forms heterodimers with ClC-3 and depends on this heterodimerisation for endosomal trafficking [46, 98]. The Clcn3−/− mouse has neurodegeneration with loss of the hippocampus months after birth and early retinal degeneration [64].

CLCN3 variants have two mechanisms. In the recessive form, the variant c.336_339del p.(Lys112AsnfsTer6), which truncates the protein before the first transmembrane domain, abolishes function [4, 64]. In the dominant form, missense function is not uniform: in Duncan's work, p.(Ile607Thr) and p.(Thr570Ile) increased currents at negative cytoplasmic voltages and lost inhibition by acidic luminal pH, and NM_001829.4:c.1238C>T p.(Ala413Val) and NM_001829.4:c.2315T>C p.(Val772Ala) did not differ from wild type [81] [64]. Polovitskaya and colleagues explained this heterogeneity in 2026: the proteins TMEM9 (T9A) and TMEM9B are obligatory β-subunits of ClC-3, -4 and -5 and tonically inhibit transport with a C-terminal domain that blocks the cytoplasmic opening of the pore, stabilised by the phosphoinositide PI(3,5)P₂ [63]. Of the 20 missense variants studied, 12 altered exchanger properties in two classes (residues near the TMEM9 binding site, and variants that widen the voltage range of conduction), both with reduced inhibition by TMEM9, that is, toxic gain of function, with endolysosomal vacuolisation; more than 65% of the missense variants were hyperactive in the presence of T9A [63]. The two variants of residue Gly327 (5th transmembrane segment) were associated with small, malrotated hippocampi and amygdalae, as in homozygotes with loss of function, which led Nakashima to propose a dominant-negative effect; terminal 4q deletions including 4q33 have few brain anomalies, and a maternally derived partial deletion of CLCN3 in DECIPHER has no clinical data, so the gene is probably not sensitive to haploinsufficiency [79]. Constraint is high (missense z = 4.37; pLI = 1; LOEUF = 0.22), and the CADD score (22.3–28.7) did not correlate with severity [64].

In the class member CLCN4, all the missense variants tested in GeneReviews were loss-of-function, and haploinsufficiency (exonic deletions, frameshift, nonsense) produces a relatively milder phenotype than missense variants [46]. Palmer and colleagues studied 59 variants in oocytes in 2023: 15 (25%) caused loss of function by shifting voltage-dependent activation to more positive potentials and 9 caused toxic gain of function by altering gating, with inward transport at negative voltages [66]. Some variants reduce heterodimerisation with ClC-3, such as p.(Gly480Arg) and p.(Gly342Arg) (published designations on NM_001830.4) [99, 100], and studies in 2025 provided the first evidence of a dominant-negative effect of CLCN4 variants in ClC-3/ClC-4 heterodimers [98]. The genes CLCN2 (leukoencephalopathy) and CLCN7 (osteopetrosis) fall outside the scope [46].

A nomenclature note avoids a search error: in some book indexes, "CLCN3" appears through an optical character recognition error in place of CLN3, the gene for juvenile neuronal ceroid lipofuscinosis (Batten disease), which has no relationship to CLCN3.

2.8. Normalisation of the variants cited

The variants named in this article are written on the MANE Select transcript (GRCh38), confirmed in VariantValidator or ClinVar; for the variants in the table, the c. numbering coincides between the transcript used by the KDM6B (NM_001080424.2) and CLCN3 (NM_173872.3) cohorts and the MANE transcript. Variants cited only at protein level, without an unambiguous c. allele in ClinVar, keep the published designation [4, 81].

3. Clinical Spectrum and Manifestations by System

3.1. Common syndromic description

The seven entities share a core: developmental delay with onset in infancy or early childhood, ID of variable severity, disproportionate language delay in several of them, behavioural changes (ASD, ADHD, anxiety, self-injury) and, in a variable proportion, epilepsy, tone abnormalities and anomalies on brain magnetic resonance imaging (MRI). They differ in average severity, which ranges from the predominantly mild ID of SCAF4 and KDM6B to the severe or profound ID of GRIN1 and CLCN3, and in extraneurological signs, frequent in RERE, SCAF4 and CLCN3 and scarce in TBR1 and SPTBN1. All the following data come from series of patients assembled by genetic diagnosis, with ascertainment bias towards severe presentations and without population studies; there are almost no adult data, and the frequencies describe patients followed with usual care, not the natural history without intervention.

3.2. GRIN1: dominant form (NDHMSD)

The GeneReviews phenotype is based on 62 of the 72 published individuals with sufficient information, 54 de novo heterozygotes and 8 homozygotes, with data on 11 heterozygotes from DECIPHER [1]. All have developmental delay or ID, profound in 17%, severe in 71%, moderate in 7% and mild in 5%, and 48% have no active speech [1]. Epilepsy affects 65%, with onset between birth and 11 years (median 22.5 months); where the type was specified, 13% had epileptic spasms, 68% generalised seizures and 20% focal seizures, and the range includes tonic, tonic–clonic, atonic and myoclonic seizures, eyelid myoclonia, absences, gelastic seizures and status epilepticus; of the 27 patients with therapeutic follow-up, 17 had refractory epilepsy and 10 were controlled with usual antiseizure medicines [1]. The motor system is affected in a characteristic way: hypotonia in 66%, spasticity in 40% and movement disorder in 48% (dystonia 13%, dyskinesia 11%, choreiform movements 15%) [1]. Cortical visual impairment (34%) and oculogyric crises (11%) complete the neuro-ophthalmological picture that Lemke and colleagues identified as discriminating: severe ID with cortical visual impairment and oculomotor and movement disorders [1, 71]. Behaviourally, there are signs of ASD in 22%, stereotypies in 32%, self-injury in 7% and sleep disturbance in 15% [1]. Feeding difficulties (31%) result from severe hypotonia, gastro-oesophageal reflux disease (GORD) and oropharyngeal dysphagia, and a subgroup needs gastrostomy; there is growth restriction or short stature in 11%, microcephaly in 27% and scoliosis in 11%, with no specific facial dysmorphism [1]. The absence of described psychosis in GRIN1-NDD was the subject of a commentary in 2025 [102].

Neuroimaging divides patients into two groups. In a subgroup of 11 individuals there is extensive diffuse bilateral polymicrogyria, frontal, perisylvian, parietal and temporal, with relative occipital sparing, enlarged extra-axial spaces, enlarged lateral ventricles, reduced white matter volume, thin corpus callosum and abnormal hippocampi, an appearance similar to the dysgyria of the tubulinopathies or of GRIN2B [1, 72]. Patients with polymicrogyria share severe or profound delay, postnatal microcephaly, cortical visual impairment and refractory epilepsy, and at least four had abnormal movements or gaze deviations [72]. Without malformation, 23% have generalised volume loss, and two individuals had non-specific white matter hyperintensities [1]. Regression is documented in one individual from 3.5 years of age, with loss of speech and social interaction, drooling and loss of sphincter control [1, 103]. Life expectancy is unknown and adults are probably underdiagnosed [1].

3.3. GRIN1: recessive forms (NDHMSR and DEE101)

NDHMSR is characterised by severe psychomotor delay, severe ID and involuntary movements (stereotypies, spasticity, dystonia), generally without independent walking and with absent or poor speech [9]. The spectrum includes siblings with a homozygous missense variant, ID and autistic traits without epilepsy [104]. DEE101 is much more severe: in the family with the homozygous nonsense variant p.(Gln556Ter), three siblings of consanguineous parents had neonatal epileptic encephalopathy with intractable seizures from the first day and died between 5 days and 5 months; the infant with a homozygous splice variant had seizures on day 2, myoclonic encephalopathy, axial hypotonia, contractures, bradycardia with asystole and episodic complete heart block at 35 days, nasogastric tube (NGT) feeding and no developmental acquisitions at 12 months [11]. The heterozygous parents in both families had no disease [11].

3.4. GRIN class members: GRIN2B and GRIN2D

In Platzer's GRIN2B series (61 individuals with clinical evaluation, ID specified in 54 and MRI in 47), all had delay or ID, severe or profound in 61% (33/54), moderate in 24% (13/54) and mild in 15% (8/54); there was regression in 7% (4/61), hypotonia in 56% (34/61), five with tube feeding, and spasticity in 23% (14/61), all with severe ID [39]. Epilepsy (51%, 31/61) begins between birth and 9 years, is refractory in about half of those treated and includes generalised seizures (58%), focal seizures (48%) and spasms (35%); most patients with spasms have hypsarrhythmia and meet criteria for West syndrome [39]. There is ASD or autistic traits in 26%, microcephaly in 18%, movement disorder in 10% and cortical visual impairment in 8%; MCD (13%, 6/47) is a diffuse polymicrogyria-like cortical dysplasia with hypoplastic corpus callosum, enlarged and dysplastic basal ganglia, hippocampal dysplasia, enlarged tectum and absent septum pellucidum, similar to the tubulinopathies [39]. In a review of 105 patients, ID or delay was present in 100%, hypotonia in 87.27% and epilepsy in 53.08% [86]. GRIN2B also features among the genes of Rett-like presentations [89].

In GRIN2D (22 published patients), ID was severe or profound in 59% (13/22) and 36% (8/22) had no speech [43]. Seizures affect 86% (19/22), with mean onset at 11 months (1 day to 41 months; median 4 months), varied initial types (spasms and generalised seizures in five each, focal in four, absences in three, febrile in one) and multiple types in 11/19; no antiseizure regimen was particularly effective, and five patients, all with epilepsy, had regression with an encephalopathic course [43]. There is hypotonia in 54%, athetoid or myoclonic movements, ballism, dyskinesia or chorea in 36%, cortical visual impairment in 32%, spasticity in 18% and ataxic gait in 3 of 22; MRI (17/22) was normal in nine and showed atrophy in seven, without MCD, unlike GRIN1 and GRIN2B [43]. No adults have been described [43]. In 2026 a series of GRIN2D DEE with polymorphic seizures including spasms was published [105].

3.5. KDM6B (Stolerman syndrome)

The twelve original patients, aged 2 to 17 years, walked at 2–3 years, had verbal apraxia or dyspraxia, intelligence quotients (IQ) between 80 and 97 in several and ASD in four, and had a prominent nasal root, coarse features and musculoskeletal findings such as broad hands and fingers, hypermobility, 5th finger clinodactyly and toe syndactyly [12, 14]. The cohort of 85 individuals of Rots and colleagues changed this picture. Developmental delay was present in all but two and becomes less evident after 5 years of age; ID is mainly mild, severe in only two individuals (one with a dual diagnosis in HNRNPU), and about 90% have ID, ASD or both [13]. Psychosis occurred in 20% of individuals aged 12 years or more, a figure probably underestimated because only 20 of the 85 had reached that age [13]. Movement disorders (gait abnormality, dystonia-like presentations, spasticity, hypertonia with toe-walking) were in several cases the reason for genetic testing, resolved in two and in one required botulinum toxin; severe hypotonia prompted muscle biopsy or a myopathy panel in several, and severe neonatal feeding difficulties led to NGT feeding or admission to a neonatal intensive care unit (NICU) [13]. Nobody had short stature, and seven of ten with high birth weight retained overgrowth [13]. Dysmorphism is mild and variable, without a gestalt: anteverted nares with a depressed nasal root, deep-set eyes with narrow downslanting palpebral fissures, prominent forehead, sometimes a flat face, synophrys or overfolded helices [13].

There were no significant differences between truncating and non-truncating variants after Bonferroni correction; clefting (p = 0.03 uncorrected) was the only nominal difference [13].

3.6. TBR1 (IDDAS)

Nambot's 25 new individuals (12 female and 13 male, 2 to 29 years, healthy non-consanguineous parents) all had delay or ID, moderate in 10 and severe in 15 [18]. Five walked unaided only after 24 months, four only with support and two did not walk; seven had no speech and most of the others had severe language delay [18]. There were autistic traits in 19/25 (formal ASD by DSM-IV or ADOS/ADI criteria in five), with stereotypies, rituals and echolalia, and behavioural changes in 22/25 (attention deficit, self- and other-directed aggression) [18]. Hypotonia and fine motor delay were present in 15/21 each, abnormal movements (dystonia, chorea, tremor) in 7/21 and seizures in five, with onset between 6 months and 28 years, including a refractory generalised epilepsy sensitive to geometric patterns and light [18]. Dysmorphism is heterogeneous (high or broad forehead and long philtrum in 6/18), and there are skeletal anomalies in 13/24 (laxity, pes planus, scoliosis), intrauterine growth restriction (IUGR) in 7/23, constipation in 7/25 and feeding difficulties in 4/25, some with Nissen fundoplication [18].

ID severity did not differ between truncating and missense variants (p = 0.99), nor for the recurrent last-exon variant [18]. MRI was described as abnormal in 10/22 by the referring centres, but specialist review of seven scans found anomalies in all of them, including three initially considered normal: thin or absent anterior commissure in 7/7, dysplastic hippocampi in three and cortical dysplasia with severe gyriform anomalies in two, as well as subtle cortical thickening, a relatively small brain and reduced gyral complexity [18]. These findings are subtle and not diagnostic in isolation, but the authors propose them as supporting signs for variant interpretation [18, 19]. Cases without ID exist but are the exception [19].

3.7. SCAF4 (Fliedner–Zweier syndrome)

In the original description (eight analysable individuals), all had delay or ID, mainly mild, with speech more affected than motor function: seven walked between 12 and 18 months, but only two said their first words before 15 months and four after 2 years [75]. Half had seizures (myoclonic in two, intractable in one, myoclonic–astatic epilepsy at 4 years in another), five had behavioural changes and there were renal anomalies (unilateral agenesis, multicystic kidneys, hydronephrosis), cardiac anomalies (ventricular septal defect, VSD, bicuspid aortic valve, hypoplastic aortic arch, dilated cardiomyopathy) and skeletal anomalies (sacral segmentation anomaly, brachydactyly, kyphoscoliosis, hallux valgus, 2–3 toe syndactyly, pectus excavatum); one individual had tracheo-oesophageal fistula, imperforate anus, tethered cord and hypothyroidism [75]. Schmid's series added 50 individuals and showed that delay or ID is absent in few (three tested because of epilepsy or hypotonia and behavioural changes), with cognition ranging from normal IQ to severe ID, mostly borderline to mild [23].

Seizures in Schmid's individuals began between 9 months and 13 years, focal or generalised, with an electroencephalogram (EEG) showing polymorphic waves and multifocal spikes; MRI (27 patients) showed non-specific changes in eight, such as periventricular leukomalacia, prominent ventricles or Chiari malformation [23]. The Chinese series describe epilepsy with infrequent seizures and a good response to antiseizure medicines, with a median age at onset of 4 years [106], and focal epilepsy without recurrence after short-term therapy, with skeletal anomalies and one case of compound heterozygosity, atypical inheritance for a dominant disease [107]. A review of 68 published cases recorded delay or ID in 59, language disorder in 46, distinctive facies in 38, epilepsy in 35 (mainly focal, with a good response), skeletal anomalies in 31, cardiovascular in 8 and renal in 8 [108]. Dysmorphism is mild and non-specific; epicanthus, deep or flat nasal root, bulbous nasal tip and deep philtrum recur in some, and a broad nasal root in others [23, 75]. The seven adults of the 2025 series with a mild presentation in childhood had mild learning difficulties, employment and a family, without late or progressive problems, although the data are insufficient to define the natural history [23]. The Brazilian report describes a 16-year-old boy with ID, epilepsy, behavioural changes, speech impairment, dysmorphism, heart disease and obesity [76].

3.8. SPTBN1 (DDISBA)

In Cousin's cohort (17 male, 12 female, 6 months to 26 years), all but one had early-onset developmental delay, 21/29 had ID and three regressed [27]. Speech and language impairment gives the entity its name and is prominent, with variable hearing loss [28]. There were seizures in 9/29 (four with frontal lobe or generalised epilepsy), autistic traits or ASD in six and other behavioural changes in 15: ADHD in 12, emotional lability in 8, aggressiveness or self-injury in 7, anxiety in 3 and obsessive behaviour in 3; sleep disturbances (seven) sometimes accompanied the seizures [27]. MRI was abnormal in seven, with thin corpus callosum, ventriculomegaly, delayed myelination, diffuse parenchymal atrophy and mild cerebellar atrophy [27]. There are tone abnormalities, abnormal movements, variable dysmorphism (hypertelorism in some), micro- or macrocephaly and short stature in some; a 13-year-old individual had mild delay and dystonia as the main symptom, and another a predominantly hepatic phenotype [27]. The same variant NM_003128.3:c.3007C>T p.(Arg1003Trp), with divergent classifications in ClinVar [81], produced, in three carriers, seizures in only one and abnormal MRI in only another [27]. In Rosenfeld's series, all seven had delay, the older ones with mainly mild to moderate ID, abnormal behaviour in 6/7 (ADHD in 4), epilepsy in 2/7, high narrow palate in the two missense carriers, mild dilatation of the ascending aorta (Z 2.7) in a 25-year-old woman and mild hemihypertrophy in a boy [69]. The 2026 review estimated seizures in 45.5% of carriers of SPTBN1 variants, and the five new patients with a de novo variant all had seizures [60]. The heterogeneity extends to a Korean infant with infantile epileptic spasms syndrome refractory to vigabatrin and corticosteroids, microcephaly and hypoplasia of the corpus callosum [109], to children with complex or primary ADHD [110] and to a child with recent-onset refractory seizures mimicking autoimmune encephalitis [111]. No cancer predisposition, progressive neurodegeneration or systematic extraneurological involvement has been established [28].

3.9. RERE (NEDBEH)

In the 19 GeneReviews individuals, delay or ID is present in most, from mild to profound, and some reach early milestones at the normal time; hypotonia and feeding difficulties are frequent and may result from brainstem or cranial nerve anomalies, and behaviour includes ADHD, self-injury and ASD [31]. Structural ocular anomalies (6/19, about 30%) include iris and chorioretinal colobomas, optic nerve atrophy or hypoplasia, microphthalmia and Peters anomaly, without blindness; there is also myopia, anisometropia, astigmatism, strabismus and ptosis [31]. Sensorineural hearing loss (4/19, about 20%) is mild, unilateral or bilateral, and progressed in one case to severe; congenital heart disease (8/19, about 40%) is mainly VSD; epilepsy affects 2/19 and neuroimaging is abnormal in 13/19 (about 70%), with reduced white matter, abnormal or thin corpus callosum and ventriculomegaly [31]. In the genitourinary system there is vesicoureteric reflux (VUR), cryptorchidism and hypospadias; there is hip dysplasia in 2/19 and scoliosis without vertebral anomalies in 3/19, the latter probably underestimated because of young age [31]. Choanal atresia was seen only in carriers of the duplication p.(Leu1438_His1439dup), which produces a CHARGE-like presentation with coloboma, heart disease, growth deficit and genitourinary anomalies; one of these patients had persistent pulmonary hypertension with heart failure [31, 78]. In the 2026 cohort, delay or ID and ASD were frequent in all subgroups, but multisystem anomalies were concentrated in HRR missense variants and duplications [62]. Recent reports have added Chiari type I malformation [112] and enuresis with scoliosis [113]. In a 6-year-old Chinese boy with autistic behaviour, impaired verbal communication, ID, growth delay and irritability in social interactions, the de novo loss-of-function variant NM_012102.4:c.3732del p.(Tyr1245ThrfsTer12) (NC_000001.11:g.8358808del; identical numbering on MANE Select NM_001042681.2), in exon 21, was reproduced in a CRISPR/Cas9-edited cell line, which showed downregulation of the sonic hedgehog pathway and of autism-associated genes (CNTNAP2, STX1A, FARP2, GPC1) and changes in HDAC1 and HDAC2 [4, 114]. Orofacial clefts are rare: cleft lip was described in one of the ten individuals of the original series and cleft palate in another patient with NEDBEH [77, 95]. Two individuals were alive in their early twenties, and survival into adulthood is possible [31].

3.10. CLCN3 (NEDHYBA and NEDSBA)

In Duncan's series (11 individuals of Turkish, European, Ashkenazi Jewish, Uruguayan and mixed ethnicity), all had global delay or ID, from mild to profound, with gross and fine motor and language delay, and five did not speak [64]. MRI was abnormal in 9/11: total or partial agenesis of the corpus callosum in 6/9, disorganised cerebellar folia in 4/9, delayed myelination, reduced white matter, pontine hypoplasia and dysmorphic dentate nuclei in 3/9 each, and increased frontal gyration in the two carriers of p.(Thr570Ile) [64]. There were seizures in 6/11, all with brain anomalies, visual impairment in 9/11, hypotonia in 7/11, mood or behavioural changes in 6/11 (severe anxiety, self-injury, explosive behaviour), autism in four, feeding difficulties in 9/11, growth delay in 4/11 and dysmorphism in 10/11 (microcephaly, prominent forehead, hypertelorism, downslanting palpebral fissures, full cheeks, micrognathia) [64]. The carrier of p.(Ile607Thr) had a complex brain malformation (hypoplastic pons, aqueductal stenosis, small vermis), arthrogryposis multiplex congenita, hip dislocation, head circumference of +4.5 standard deviations (SD) and died at 23 days [64]. The two homozygous siblings had the most severe form: no speech, no walking, neonatal focal seizures and multifocal tonic and myoclonic seizures from 3 months, absent visual fixation, nystagmus and "salt-and-pepper" fundus pigmentation, reduced white matter, thin corpus callosum and small, malrotated hippocampi; one died at 14 months [64]. The two boys with Gly327 variants had polymicrogyria (insular in one, diffuse in the other), complete agenesis of the corpus callosum, hypoplastic pons and coloboma or optic nerve hypoplasia; one of them, at 17 years, did not speak and had never walked [79].

The Polovitskaya cohort (16 individuals with functionally altered variants or biallelic loss of function; 7 female and 9 male; median 9 years, range 23 days to 25 years) defines the current spectrum [63]. Initial concern arose at a mean of 6.7 months (neonatal to 18 months), with neonatal findings in 63%: hypotonia in 8/13, feeding difficulties in 8/16 and growth delay in 4/16 [63]. All those assessable had delay or ID (13/13), ID was present in 11/12 (mild to severe after 6 years) and four of 12 were non-verbal; they sat between 6 and 18 months and walked unaided between 1.3 and 6 years, but at least two were not walking after 3 years, and there was ataxia in 3/10 and dystonia in 3/11 [63]. Behavioural or psychiatric changes affected 10/14 (autistic traits 8/13, self-injury 6/13, anxiety 6/11, ADHD 3/10), sleep disturbances 7/12 (severe apnoea in two, narcolepsy in one) and regression 4/12, all with autistic traits [63]. Epilepsy (7/16, 44%) began between the neonatal period and 6.5 years, with spasms, hypermotor focal seizures, myoclonic seizures and absences; two of seven were seizure-free and one was drug-resistant [63]. The eye and vision were affected in 11/13 (optic and retinal dystrophy, nystagmus, pigment loss, strabismus, anisometropia) and the gastrointestinal tract in 10/16; one patient died at 15 years, and there was also cryptorchidism, microdontia, gingival hypertrophy, scoliosis, acquired contractures, hypothyroidism and partial growth hormone deficiency [63].

3.11. CLC class member: CLCN4

The GeneReviews table compares hemizygous males, females with a de novo variant and females with an inherited variant [46].

Epilepsy in males (61%) was drug-resistant in 12 of 22, with DEE, but the frequency may reflect recruitment in cohorts with severe epilepsy; there was one seizure-related death at 2 years 9 months and one family with three males who died in adolescence, and the risk of sudden unexpected death in epilepsy (SUDEP) remains to be defined [46]. Late spasticity may progress, ataxic gait occurs in 5/26 and, in older males, there is depression, obsessive–compulsive disorder, bipolar disorder and anxiety with improvement on treatment; the face lengthens subtly and the chin becomes square with age [46]. Palmer's series, with 122 published individuals, confirmed almost constant ID in males, mainly moderate to severe or profound, and a broad spectrum in females, half of the heterozygous carriers apparently without disease; ADHD in 59% of males and 46.7% of de novo females, ASD in 54.5% and 40%, anxiety in 27.2% and 53%, epilepsy in 59% and 20%, abnormal neuroimaging in 58.8% and 61.5%, and late neurological signs (tremor, ataxia, hyperkinesia, stooped gait, progressive spasticity) in 31.8% and 40% [66]. The four females with the recurrent de novo gain-of-function variant NM_001830.4:c.1664C>T p.(Ala555Val) [81] had severe growth and feeding restriction, two with gastrostomy, without growth hormone deficiency; progressive microcephaly affected 80.8% of de novo females versus 20% of males [66]. Speech and language impairment is almost universal (12/13 in a dedicated series), with eight minimally verbal participants using augmentative and alternative communication (AAC) [115]. In 13 Chinese patients, seizures affected 8/13, refractory in 62.5%, and delay preceded the seizures [99].

3.12. Comparison between genes

4. Clinical Recognition and Warning Signs

4.1. When to suspect

None of these disorders can be recognised from the phenotype with sufficient confidence to dispense with genetic testing, and the KDM6B and TBR1 cohorts explicitly recommend a genotype-first strategy, with broad ID panels or exome/genome sequencing [13, 18]. The clinician's role is to recognise the child who should be tested and to guide interpretation of the result. General suspicion arises from global developmental delay or ID without an acquired cause, especially when accompanied by disproportionate language delay, ASD, early-onset epilepsy, movement disorder or congenital anomalies; in that context, the American College of Medical Genetics and Genomics (ACMG) guideline recommends exome or genome sequencing as a first- or second-tier test [49]. For GRIN1-NDD, GeneReviews considers suggestive the association of delay or ID with, in infancy or childhood, epilepsy, abnormal tone, dystonic, dyskinetic or choreiform movement disorder, ASD, microcephaly or cortical visual impairment, and diffuse bilateral polymicrogyria on MRI in a subgroup [1]. For RERE, the diagnosis is suggested by delay or ID and/or ASD with any of: ADHD or self-injury, infantile hypotonia, feeding difficulties, structural eye defects or visual problems, sensorineural hearing loss, septal heart disease, epilepsy, genitourinary anomalies or choanal atresia [31].

4.2. Clues by gene

4.3. Warning and emergency signs

Emergency signs are concentrated in the most severe forms. In DEE101, intractable seizures begin on the first day of life, and the infant with the splice variant had bradycardia with asystole and episodic complete heart block [11]. In de novo GRIN1 with polymicrogyria, seizures may begin on the first day with focal tonic seizures accompanied by apnoea, bradycardia and desaturation [117]. In the carrier of CLCN3 p.(Ile607Thr), the brain malformation with aqueductal stenosis and arthrogryposis led to death at 23 days [64]. In the CLCN4 class, seizure-related deaths in childhood and adolescence have been described [46]. Developmental regression, described in GRIN1, GRIN2B, GRIN2D, SPTBN1 and CLCN3, requires a search for unrecognised epilepsy and reassessment of the EEG [1, 27, 43, 63]. In SPTBN1, a presentation with recent-onset refractory seizures mimicking autoimmune encephalitis shows that a genetic diagnosis should be considered even in an acute setting [111]. In KDM6B, the appearance of psychotic symptoms from adolescence is a warning sign for psychiatric assessment [13]. In RERE, persistent pulmonary hypertension with heart failure in a carrier of the CHARGE-like duplication is a reminder of the need for early echocardiography [78].

4.4. Prenatal findings

None of these entities has newborn screening. The prenatal findings described are non-specific: in CLCN3, one fetus had increased nuchal translucency (NT) and a hypoplastic nasal bone at 13 weeks and suspected hypogenesis of the corpus callosum at 25 weeks [79]; in CLCN4, anomalies of the corpus callosum allow prenatal detection, and prenatal diagnosis in fetuses with congenital brain anomalies has been published [46, 118]; in TBR1, 30% of the individuals reviewed had IUGR [18]; in KDM6B, 17% had a birth weight above 2 SD [13].

5. Diagnosis: Criteria, Tests and Algorithms

5.1. Criteria

There are no formal clinical criteria for any of the entities, and the diagnosis is established by identification of a pathogenic or likely pathogenic variant in the gene, with a compatible mode of inheritance [1, 31]. For GRIN1, this means a de novo heterozygous missense variant or biallelic missense or truncating variants; a VUS neither confirms nor excludes the diagnosis [1]. For RERE, a heterozygous pathogenic or likely pathogenic variant [31]. For CLCN3, distinguishing between the dominant and recessive forms requires parental testing [64].

5.2. Testing algorithm

The guidelines converge on broad sequencing as the first line. The ACMG recommends, with a strong recommendation by the GRADE method, considering exome or genome sequencing as a first- or second-tier test in paediatric patients with congenital anomalies with onset before 1 year or with developmental delay or ID with onset before 18 years, because of the higher diagnostic yield and possible cost-effectiveness when requested early [49]. The National Society of Genetic Counselors (NSGC) strongly recommends genetic testing in all individuals with unexplained epilepsy, without an age limit, with exome, genome and/or a multigene panel of more than 25 genes as first line followed by microarray, conditionally preferring exome or genome to a panel, and requires that tests be chosen, ordered and interpreted by a qualified professional with genetic counselling before and after testing [119]. The Genetics Commission of the International League Against Epilepsy (ILAE) recommends exome or genome sequencing with copy-number variant (CNV) analysis as first line in most cases, indicates testing in severe childhood-onset epilepsies, especially DEE, in epilepsies with ID, autism or other comorbidities and in MCD, and proposes periodic genetic reassessment of undiagnosed patients, for example after two years [120]. The classic order in ID, microarray first, is retained in the RERE GeneReviews, which recommends starting with chromosomal microarray analysis (CMA) and, if not diagnostic, moving on to an ID panel or exome, considering single-gene testing rarely useful [31]. For GRIN2B, GeneReviews recommends a panel with deletion and duplication analysis, because 4 of 86 alterations are deletions or duplications and CMA covers 12p13.1 [39]. Trio testing (patient and parents) is particularly useful because de novo status is a strong pathogenicity criterion in all these genes and because SCAF4, KDM6B and RERE have described parental transmission [13, 23, 62].

5.3. Yield and limitations of testing

These percentages describe the proportion of published variants detectable by each method and are not equivalent to the clinical sensitivity of the test in a patient with the phenotype [1]. There are five practical limitations. First, panels vary in content, and some ID panels include neither GRIN1 nor RERE [1, 31]. Second, detection of intragenic deletions depends on CNV analysis, and larger contiguous gene deletions (1p36 deletion, 2q24 deletion) may escape targeted methods and require CMA [18, 31]. Third, low-level mosaicism may not be detected in blood: in SPTBN1, half-siblings inherited the variant from a mother with 1.8% of reads, and one proband had 13.3% [27]; in SCAF4, a patient with a mild phenotype had 8% [23]; in GRIN2B, one patient had about 50% in blood [39]. Fourth, last-exon truncating variants escape NMD and require specific interpretation: they are probably benign in SCAF4 when present in gnomAD [75], but three PTVs in the last or penultimate exon of KDM6B were considered likely pathogenic because they eliminate the zinc domain [13], and the recurrent TBR1 variant lies in the last exon [18]. Fifth, the reference transcript matters: the MANE Select for KDM6B (NM_001348716.2) and RERE (NM_001042681.2) differs from that used in the cohorts and in GeneReviews, although the verified variants retain the same numbering [4]. A negative test does not exclude the diagnosis.

5.4. Interpretation of missense variants and functional testing

Interpretation of missense variants is the main clinical problem with these genes. In May 2022, 73% (111/153) of CLCN4 missense variants in ClinVar were VUS [66]; in SPTBN1, VUS (547) exceeded pathogenic and likely pathogenic variants combined (263) [28]; in an AlphaMissense assessment of 135 SPTBN1 variants, most were classified in ClinVar as of uncertain significance [110]. De novo segregation, location in domains with clustering of pathogenic variants (S2 and M1–M4 of GluN1; JmjC and zinc domain of KDM6B; CH2 of βII-spectrin; HRR of RERE) and functional data help with reclassification [1, 13, 27, 78]. In the GRIN class, functional classification is mandatory before any targeted therapy: the Myers framework integrates six in vitro assays [82], the Broad Institute GRIN-Portal (version 1.1, updated September 2026) and the GRIN Database bring together variants and classifications [56, 121], and the Leipzig authors of the GeneReviews make themselves available to assess VUS in NMDAR genes and to initiate studies at the Center for Functional Evaluation of Rare Variants of Emory University (Atlanta) [43]. For CLCN3, CLCN4 and SPTBN1, electrophysiology in Xenopus oocytes and studies in HEK293 cells are research tools and are not available as routine clinical tests [27, 64, 66]. Blood RNA sequencing in SCAF4 is likewise a research observation [24].

5.5. Ancillary investigations

5.6. Content of the report

The report should state the gene, the transcript with version (preferably MANE Select), HGVS c. and p. nomenclature, the genomic coordinate, zygosity, parental origin when studied, the ACMG/AMP class with the criteria applied and, for GRIN missense variants, the functional classification when available [4, 82]. The NSGC and the ILAE add the need for pre-test counselling on the indication, the choice of test, the possibility of VUS and non-medical implications, and show with an illustrative case the interpretation errors that arise from a request with insufficient phenotypic information [119, 120].

5.7. Prenatal and preimplantation diagnosis

Once the familial variant has been identified, prenatal diagnosis and preimplantation genetic testing (PGT) are possible in the dominant and recessive forms; in de novo forms they may be considered because of the residual risk of germline mosaicism [1, 31]. In the absence of a known variant, ultrasound findings (agenesis of the corpus callosum, increased NT) are non-specific, and prenatal diagnosis of CLCN4 in fetuses with congenital brain anomalies has been published [79, 118]. Prenatal diagnosis was performed in a 23-year-old pregnant woman with de novo Fliedner–Zweier syndrome, whose variant was found in the fetus by amniocentesis [59].

6. Differential Diagnosis

6.1. Principles

The differential diagnosis of these entities operates on two levels. Clinically, each overlaps with the hundreds of causes of ID (more than 180 genes in the OMIM phenotypic series of autosomal dominant, autosomal recessive and X-linked ID), of early-onset epileptic encephalopathy (more than 50 genes) and of polymicrogyria (about 50 genes), so the differential is resolved, in practice, by exome or genome sequencing [1, 39]. Molecularly, it is important to distinguish a point variant from a contiguous gene deletion (RERE and 1p36 deletion; TBR1 and 2q24 microdeletion), the dominant from the recessive forms of the same gene and the members of each class from one another, because prognosis, recurrence risk and, in the GRIN class, targeted therapy depend on this distinction.

6.2. Differential diagnosis table

6.3. Main differentials

For RERE, overlap with CHARGE syndrome is the most frequent pitfall, and the two-amino-acid duplication in the HRR produces a presentation that meets CHARGE criteria: Jordan and colleagues recommend testing RERE in CHARGE patients without a CHD7 variant [78]. Abnormal semicircular canals and tracheo-oesophageal fistula point to CHARGE [31], although a tracheo-oesophageal fistula has been described in one individual with SCAF4 [75]. The 1p36 deletion is distinguished by the dysmorphism and the fontanelle, and CMA identifies it and allows the extent to be assessed when the deletion includes RERE [31, 97].

For TBR1, the differential with the 2q24 microdeletion is mainly one of nomenclature and extent: Nambot's cohort included deletions affecting only TBR1 or TBR1 with TANK and PSMD14 and excluded deletions with other OMIM genes, which are distinct entities; Orphanet entry 1617 refers to a deletion with dopa-responsive dystonia–parkinsonism, not to IDDAS [17, 18, 21].

For GRIN1, the combination of cortical visual impairment, oculogyric crises and movement disorder distinguishes it from most DEEs, but oculogyric crises have their own metabolic differential: neurotransmitter and PNPO deficiencies, GLUT1DS, Wilson disease and pantothenate kinase-associated neurodegeneration (PKAN), among others, have biochemical markers and some have specific treatment, so metabolic assessment should not be abandoned because a GRINopathy is suspected [116]. The diffuse polymicrogyria with relative occipital sparing of GRIN1 and GRIN2B resembles that of the tubulinopathies, and correlation with imaging is useful when reanalysing exome data [1, 39, 87]. Anti-NMDAR encephalitis shares the glutamatergic pathophysiology and hyperkinetic movement disorder, but has an acute onset and immunological markers [122].

Within the classes, the differential is molecular and has consequences. In the GRIN class, the dominant form of GRIN1 must be distinguished from the recessive forms, because parents heterozygous for nulls are healthy and the recurrence risk changes from about 1% to 25% [1, 11]. In the CLC class, the X linkage of CLCN4 explains affected boys with mothers of normal cognition, whereas CLCN3 is autosomal and usually de novo [46, 64]. In the spectrin family, ataxia with cerebellar atrophy points to SPTBN2, hypotonia with areflexia to SPTBN4 and epilepsy with normal MRI to SPTBN5, whereas SPTBN1 combines delay, speech impairment and ADHD [60, 69]. Optical character recognition in some book indexes confuses CLCN3 with CLN3, the gene for juvenile neuronal ceroid lipofuscinosis, and the reader should always check the gene symbol.

7. Treatments and Risk Management Strategies

7.1. General principles

None of the entities has an authorised treatment, and no intervention has been studied in a randomised trial in GRIN1-NDD, KDM6B, TBR1, SCAF4, SPTBN1, RERE or CLCN3; the GeneReviews management recommendations are expert opinion, and ClinGen has no actionability assertion for these genes [3, 14, 19, 24, 28, 31, 36]. Management is multidisciplinary, guided by the problems present in each patient and by the risks specific to each gene, and follows the general principles of ID and epilepsy. The only area with targeted therapy under clinical investigation is the GRIN class, where the choice depends on the functional direction of the variant.

7.2. Evaluations following diagnosis

7.3. Epilepsy

Epilepsy is treated with antiseizure medicines chosen according to seizure type and syndrome, prescribed by an experienced neurologist; the GRIN1 GeneReviews stresses that many antiseizure medicines may be effective, but none has shown specific efficacy in this disease, and the same is stated for GRIN2B, GRIN2D and RERE [1, 31, 39, 43]. Educating carers about recognising and treating seizures is part of treatment [1]; first aid aims to protect the patient from injury during the seizure and, in repeated or prolonged seizures, to prevent status epilepticus, and the surveys reviewed by O'Hara show a lack of knowledge of these measures among parents, teachers and the public [123]. Patient sources record responses to specific medicines that illustrate, without establishing, choices: in CLCN3, myoclonic seizures at 4 years controlled with clobazam and oxcarbazepine and seizures at 6 months well controlled with levetiracetam [64]; in the boys with Gly327 variants, focal seizures controlled with levetiracetam and spasms refractory to intramuscular corticotrophin (ACTH), valproate, zonisamide and topiramate, with a partial response to clobazam, discontinued because of hypersecretion and respiratory difficulty [79]; in CLCN4, some response to lamotrigine in 5 of 25 patients with epilepsy, without clear benefit from acetazolamide, and a report of the ketogenic diet [46, 66, 124]; in SCAF4, a good response to antiseizure medicines in most [106, 108]; in SPTBN1, persistent spasms despite vigabatrin and corticosteroids [109]. Patients with GRIN disease with gain of function and refractory spasms are the group in whom memantine is discussed (section 7.8).

As there are no doses specific to these diseases, the doses of the antiseizure medicines mentioned in the series are those of the official labels for the authorised epilepsy indications; the extrapolation to these disorders is stated and the choice rests with the neurologist. The table summarises the doses in the US Food and Drug Administration (FDA) label.

In the European Union, levetiracetam (Keppra®, centralised authorisation of 29 September 2000) has the same brand name in Portugal and differs from the FDA label in three respects: monotherapy is authorised only from 16 years, in partial seizures of newly diagnosed epilepsy; paediatric adjunctive therapy from 1 month is 7 to 21 mg/kg twice daily before 6 months and 10 to 30 mg/kg twice daily from 6 months up to 50 kg, with oral solution preferred before 6 years; and myoclonic and primary generalised tonic–clonic seizures are adjunctive indications from 12 years; in adults and adolescents weighing 50 kg or more, treatment starts at 500 mg twice daily (or 250 mg twice daily) up to a maximum of 1500 mg twice daily [129]. Lamotrigine has a national authorisation; in INFARMED's summary of product characteristics (SmPC) for Lamictal®, the paediatric indications from 2 to 12 years are adjunctive therapy of partial and generalised seizures, including Lennox–Gastaut syndrome, and monotherapy of typical absences, with monotherapy of other seizures only from 13 years [130]. The paediatric adjunctive regimen coincides with that of the FDA at initiation (0.15 → 0.3 mg/kg/day with valproate; 0.3 → 0.6 mg/kg/day without valproate or inducers; 0.6 → 1.2 mg/kg/day with inducers), but maintenance without valproate or inducers is 1 to 10 mg/kg/day, with a maximum of 200 mg/day, versus 4.5–7.5 mg/kg/day and a maximum of 300 mg/day in the FDA label; with inducers, 5 to 15 mg/kg/day up to 400 mg/day in both texts [126, 130]. Vigabatrin (Sabril®) is indicated in Portugal as monotherapy for infantile spasms (West syndrome), with a starting dose of 50 mg/kg/day, which may be escalated over one week, and doses up to 150 mg/kg/day used with good tolerability, which coincides with the FDA label; in resistant partial epilepsy, the Portuguese SmPC indicates 40 mg/kg/day at initiation and weight-based maintenance (0.5–1 g/day between 10 and 15 kg, 1–1.5 g/day between 15 and 30 kg, 1.5–3 g/day between 30 and 50 kg, 2–3 g/day above 50 kg), and clinically significant improvement in spasms is generally seen within 2 to 4 weeks [131]. Clobazam is marketed in Portugal as Castilium®, indicated for anxiety and as adjunctive treatment of epilepsy in patients not stabilised on monotherapy; the SmPC states that, for children aged 3 to 15 years, a daily dose of 5 to 10 mg is often sufficient, starting at a low dose and increasing gradually, without the weight-based regimen of the Lennox–Gastaut syndrome indication of the FDA label [128, 132]. The sources consulted do not include labels from Brazil's National Health Surveillance Agency (ANVISA) for these medicines.

The doses of ACTH, felbamate and vigabatrin used in Xu's case (section 10) are those of the report and do not replace institutional protocols [101].

7.4. Tone, movement disorders and rehabilitation

Hypotonia, spasticity and movement disorder are managed by orthopaedics, physical and rehabilitation medicine, physiotherapy and occupational therapy, with stretching to prevent contractures and falls, positioning and mobility devices [1]. For hypertonia and dystonia, the GRIN1 GeneReviews suggests considering, with specialists, baclofen, tizanidine, botulinum toxin, antiparkinsonian medicines or orthopaedic procedures [1]; in KDM6B, one individual needed botulinum toxin for spasticity [13]. GeneReviews gives no doses; those of the FDA labels follow as a reference, with the caveat that the oral safety and efficacy of baclofen and tizanidine are defined in the label for adults and that paediatric use is an extrapolation of the spasticity indication.

In Portugal, the SmPCs differ from the FDA label in respects with practical consequences. Baclofen (Lioresal®) has a paediatric indication below 18 years in spasticity of cerebral origin, including cerebral palsy, starting at about 0.3 mg/kg/day in 2 to 4 divided doses, with weekly increases, usual maintenance of 0.75 to 2 mg/kg/day and a maximum of 40 mg/day below 8 years and 60 mg/day above 8 years; the tablets are not suitable below 33 kg, and discontinuation should be gradual over 1 to 2 weeks; in adults, treatment starts at 15 mg/day, with increases of 15 mg/day every 3 days, up to 30–80 mg/day [136]. Tizanidine (Sirdalud®) is not recommended below 18 years because of limited experience; in adults, the starting dose for spasticity should not exceed 6 mg/day in 3 divided doses, with an optimal dose of 12 to 24 mg/day and a maximum of 36 mg/day, and it is contraindicated in severe hepatic impairment [137]. Botulinum toxin (Botox®) has in Portugal a paediatric indication only for focal spasticity of the ankle and foot in ambulant patients with cerebral palsy from 2 years of age, as an adjunct to rehabilitation, with 4 to 8 U/kg divided among the affected muscles and a maximum of 8 U/kg or 300 U per session; paediatric upper limb use, authorised by the FDA, and use outside cerebral palsy, as in the diseases in this article, are off-label in the EU [135, 138]. The sources consulted do not include ANVISA labels.

Dysphagia is assessed clinically or by videofluoroscopy when there is choking, poor weight gain, frequent respiratory infections or food refusal; thickening of foods, feeding and oromotor therapy and, in severe cases, NGT feeding or gastrostomy are the measures described [1, 31]. In CLCN4, the frequency of reflux (36.4% of males) and constipation (36.4%) leads the authors to recommend the involvement of neurogastroenterology [66].

7.5. Development, communication and behaviour

Developmental intervention follows the model of the RERE GeneReviews: early intervention from 0 to 3 years, preschool programmes from 3 to 5 years and an individualised education programme from 5 to 21 years, with physiotherapy, occupational therapy and speech therapy [31]. Augmentative and alternative communication should be introduced early for expressive deficits and does not hinder speech acquisition [1]; in CLCN4, eight of 13 minimally verbal participants used low- and high-tech AAC, and the authors recommend individualised and early speech therapy [115]. For TBR1, SCAF4 and SPTBN1, in which language impairment is prominent, speech therapy and audiological assessment are part of initial management [19, 24, 28]. For autistic traits, applied behaviour analysis (ABA) is used, ADHD is treated with medication when necessary, and severe aggression is managed by child psychiatry [1, 31]. In CLCN4, the frequency of psychiatric complications warrants surveillance and referral to a psychiatrist experienced in neurodevelopmental disorders [66]. In KDM6B, psychosis in 20% of the adolescents assessed requires psychiatric assessment at the first symptom; Rots and colleagues point out that in Kleefstra syndrome psychosis has therapeutic implications, but there is no specific treatment recommendation for KDM6B [13].

7.6. Congenital anomalies and sensory organs

The anomalies of RERE are treated by the corresponding specialty in the standard way: correction of refractive errors and strabismus, hearing aids for sensorineural hearing loss, treatment of heart disease, GORD, genitourinary anomalies and VUR, and orthopaedics for scoliosis and hip dysplasia [31]. In the subgroup with in-frame HRR duplications, in which choanal atresia has been described [31, 78], management is that of the malformation, for example according to the 2023 Slovenian national recommendations, which bring together 34 recommendations derived from a systematic review (eight systematic reviews or meta-analyses and four randomised trials), a survey and consensus meetings, and which were not developed for RERE [139]. In the newborn with respiratory distress, nasal obstruction is confirmed with a saline test followed by a 6–8 Fr suction catheter, and the diagnosis with computed tomography (CT) of the face, nose, paranasal sinuses and skull base with 2–5 mm slices; transfontanellar ultrasound excludes central nervous system involvement, and in bilateral atresia a clinical genetics consultation, cardiological assessment before surgery, audiovestibular assessment and swallowing assessment are recommended [139]. Bilateral atresia is operated on between the 10th and 13th day of life, even in preterm infants, preferably by transnasal endoscopic choanoplasty, with removal within seven days of any stent placed; unilateral atresia is operated on after 6 months; follow-up lasts at least two years or until the end of growth, and surgical revision is indicated in the presence of symptoms, poor nasal patency or reduction of the choanal lumen of more than 50% [139]. Cleft palate in one patient and cleft palate in most deficient murine embryos make examination of the palate pertinent at the initial assessment [95]. Spontaneous myocardial fibrosis in adult Rere-deficient mice and the location of the gene in one of the two critical regions for cardiomyopathy in the 1p36 deletion (chr1:8,395,179–11,362,893, hg19) support the hypothesis that the gene contributes to cardiomyopathy [92, 140], but the sources consulted do not describe cardiomyopathy in patients with RERE point variants, nor data that would establish cardiac surveillance beyond the initial echocardiogram and follow-up of structural heart disease [31]. For the renal, cardiac and skeletal anomalies of SCAF4, orthopaedic assessment and assessment of structural anomalies are part of initial management [24], and Chen and colleagues recommend monitoring for multisystem involvement and epilepsy [108]. The cortical visual impairment of GRIN1 has no specific treatment, and early intervention with vision therapy is the indicated measure [1]. In a boy with SPTBN1 and mild hemihypertrophy, the team maintained tumour surveillance, with no tumours identified; this was an individual decision linked to the hemihypertrophy and not a recommendation for the entity [69].

7.7. Targeted therapy in the GRIN class: the functional principle

Targeted therapy of the GRINopathies rests on a simple rule with safety consequences: gain-of-function variants are candidates for NMDAR blockers or negative modulators, and loss-of-function variants and null alleles are candidates for increasing receptor activation with the co-agonist D-serine, given as L-serine [83, 120]. The ILAE summarises this in a table of precision therapies: memantine, dextromethorphan or ketamine for GoF variants of GRIN1, GRIN2A, GRIN2B and GRIN2D, serine for LoF [120]. Reversing this rule is dangerous: a child with a GoF missense variant mistakenly treated with L-serine had immediate and temporary behavioural deterioration, and the GRIN2A GeneReviews recommends avoiding agonists and other activators in activating variants and using NMDAR blockers (memantine, dextromethorphan, ketamine) with caution in inhibitory and null variants [83, 141]. Correct classification of pathogenicity and functional direction, according to the Myers framework, is therefore a prerequisite for any treatment [50, 82]. The evidence is, to date, observational for memantine and L-serine: a 2025 review of precision therapies in the genetic epilepsies points out that only everolimus has class I evidence and that most of the precision therapies proposed rest on uncontrolled reports [142]. These therapies are not recommended as routine for GRIN1 in any guideline, and the 2021 GRIN1 GeneReviews does not include them [1, 3].

7.8. Memantine in gain-of-function variants

Memantine is an NMDAR channel blocker, authorised in the USA (Namenda®, NDA 021487, approved on 16 October 2003) and in the EU (Ebixa® on 15 May 2002 and Axura® on 17 May 2002, centralised procedure, plus generics) only for moderate to severe Alzheimer's disease in adults; its use in GRIN diseases is off-label in all jurisdictions [143, 144]. In the FDA label, the adult dose starts at 5 mg once daily and increases in 5 mg steps, at intervals of at least one week, up to 20 mg/day (10 mg twice daily), with a target of 5 mg twice daily in severe renal impairment (creatinine clearance 5–29 mL/min) and increased levels in conditions that alkalinise the urine [143]. The European SmPC of Ebixa® gives, in adults, 5 mg/day in the first week and increases of 5 mg per week up to 20 mg once daily, 10 mg/day in moderate renal impairment (increasing to 20 mg/day if well tolerated) and severe renal impairment, and does not recommend use in severe hepatic impairment; the practical difference from the FDA label is the once-daily dosing [144]. Neither label authorises paediatric use: the FDA label records that memantine showed no efficacy in two 12-week controlled studies in 578 children aged 6 to 12 years with ASD, with extended-release capsules of 3, 6, 9 or 15 mg/day for weights <20 kg, 20–39 kg, 40–59 kg and ≥60 kg, started at 3 mg/day and titrated until week 6, and with a risk profile similar to that in adults; the European SmPC has no paediatric data [143, 144]. The most frequent adverse effects in adults are dizziness, headache, confusion and constipation [143].

In GRIN diseases, the doses used in the series were calculated by weight. In Xu's case, memantine was started at 14 months at 0.2 mg/kg/day, increased to 0.4 mg/kg/day over two weeks and was at 0.5 mg/kg/day at the last visit [101]. In Pierson's GRIN2A patient, titration over weeks reached about 0.5 mg/kg/day without adverse effects, and the authors state that paediatric safety guidance limited higher doses [145]. In Karnstedt's series, the standard dose was 0.5 mg/kg/day, increased in a few cases up to 2 or 4 mg/kg/day, with a duration of days to 22 months [50]. These doses were not established by dose-finding studies.

In Karnstedt's series, response depended on the distance of the variant from the memantine binding site: among GoF variants more than 9 Å away, seven of nine had a positive effect; less than 9 Å away, only three of ten, because nearby variants alter drug potency [50]. Among the GRIN1 patients in this series, the M3 helix variant p.(Met641Ile) responded in one patient and not in another, and the S2 variant p.(Met706Val) responded in a 17-year-old patient and not in another aged 22, which shows that the same variant does not guarantee the same response [50]. Two GRIN2B patients with NM_000834.5:c.1922T>C p.(Ile641Thr) and p.(Met818Leu) (published designation; ClinVar has two c. alleles with this protein change) achieved seizure control with memantine in a Chinese series [86]. NMDAR antagonists are limited by off-target behavioural and cardiovascular effects [142]. Intrathecal injection of magnesium sulphate for a Mg²⁺-insensitive GRIN1 GoF variant was described in a short, non-randomised, unblinded N-of-1 trial, with a suggestion of benefit on neurophysiological monitoring; it is an invasive, investigational procedure [147].

7.9. L-serine in loss-of-function variants

L-serine is a non-essential amino acid given orally that increases D-serine, the NMDAR co-agonist at the GluN1 subunit [83, 148]. In a 5-year-old child with severe Rett-like encephalopathy and the loss-of-function GRIN2B variant p.(Pro553Thr) (published designation, NM_000834.5), D-serine restored function in vitro, and dietary supplementation with L-serine increased plasma and brain D-serine and was associated with motor, cognitive and communication improvement at 11 and 17 months [148]. Krey's retrospective series brought together ten individuals with GRIN2A or GRIN2B diseases, each in an N-of-1 trial: in the nine with LoF missense variants or null alleles there was improvement in behaviour in 8/9 (89%), in development in 4/9 (44%) and in the EEG or seizure frequency in 4/9 (44%), without adverse effects [141].

The phase 2A trial of Juliá-Palacios (NCT04646447), open-label, single-arm and non-randomised, treated for 52 weeks 24 Spanish participants aged 2 to 18 years with GRIN LoF variants (GRIN2B 12, GRIN1 6, GRIN2A 5; 23 completed) [51]. The protocol dose was 250 mg/kg/day in the first two weeks and 500 mg/kg/day from week 3 to week 52, orally, divided into three daily doses, in participants aged 2 to 18 years, with no absolute maximum dose defined in the source, using L-serine powder manufactured by Nutricia and described by the FDA as a substance generally recognised as safe (GRAS) as a food additive [149]. There was improvement in the GMFM-88 gross motor scale (P = 0.002) and in PedsQL quality of life (P = 0.00068) regardless of severity, in several Vineland subdomains, mainly in the mild group, and in the Bayley-III cognitive scale in the severe group (+21.6 points; P = 0.016); the EEG normalised in five and seizure frequency in one child [51]. One patient discontinued because of irritability and insomnia [51]. Wang's review adds that another child had a first seizure on L-serine [142]. The Dutch protocol of N-of-1 trials in GRIN2B LoF used 500 mg/kg/day in four equal doses and mentions paediatric use of 400–600 mg/kg/day for decades in 3-phosphoglycerate dehydrogenase deficiency without reported adverse effects [150]. The first randomised, double-blind, placebo-controlled N-of-1 trials, in four GRIN2B LoF children, showed significant improvement in one performance measure at 6 weeks (p = 0.0373) and clinical improvement in two of the four patients, without features predicting response [151]. The trial registries and summary consulted do not state a daily maximum in grams.

L-serine has no authorisation as a medicine for GRIN diseases from the FDA or the European Medicines Agency (EMA), nor a European orphan designation for this indication; in the trials it is used as a supplement or as a food for special medical purposes [149, 152].

7.10. Radiprodil and other modulators

Radiprodil is a selective negative allosteric modulator of the GluN2B subunit. In a study of three infants with infantile spasms syndrome, one became spasm-free and two improved partially; after discontinuation, one remained spasm-free and two worsened and required the ketogenic diet and other antiseizure medicines [83, 153]. UCB's phase 2 trial in drug-resistant infantile spasms (NCT02829827) ended with three participants for feasibility reasons [154]. GRIN Therapeutics then developed the drug for GRIN diseases with gain of function: the Honeycomb study (NCT05818943, open-label phase 1b, 6 months to 12 years) uses an oral suspension of 0.25 mg/mL or 2.5 mg/mL twice daily, orally or by tube, individually titrated to the maximum tolerated dose, and excludes felbamate, memantine, perampanel and ACTH or prednisolone [155]. Topline results were released in a sponsor press release of 9 September 2024 and presented at the ILAE European Epilepsy Congress: in 15 patients with GoF variants in GRIN1, GRIN2A or GRIN2B, with escalation from 0.05 mg/kg and 8 weeks of maintenance, the median reduction in countable motor seizures was 86%, 71% had a reduction greater than 50% and 43% greater than 90%, one became seizure-free, and the three serious adverse events (obstructive bronchitis, viral pneumonia, adenovirus infection) were not considered related [156]. These are press release data, not peer-reviewed. The phase 3 Beeline trial (NCT07224581), randomised 1:1, double-blind and placebo-controlled, with an open-label extension, plans to enrol 100 participants aged 1 month to 18 years with GoF variants in GRIN1, GRIN2A, GRIN2B or GRIN2D, with the primary outcome being countable motor seizures at 12 weeks (epilepsy cohort) and safety at 24 weeks (cohort with neurodevelopmental symptoms), and began on 5 January 2026 at centres in Belgium, France, Germany, Italy, the Netherlands, Poland, Slovenia, Spain, the United Kingdom and the USA [52, 157].

7.11. Comparative regulatory status

7.12. Clinical trials

These studies do not allow indirect comparison: they differ in population (LoF or GoF, genes, age), design (open-label, N-of-1, retrospective) and outcomes (developmental scales, countable seizures, global impression). Percentage improvements in open-label series without a control group overestimate the effect, because development progresses with age and seizures fluctuate; the 86% reduction in Honeycomb is a median of 15 patients in a press release. The negative or interrupted trials (NCT02829827 and NCT06965881) and the ineffectiveness of memantine in 7 of 15 cases should be read alongside the positive ones [142]. There are no trials registered in ClinicalTrials.gov for SPTBN1, CLCN3, SCAF4, KDM6B or RERE; for CLCN4 and KDM6B there are only observational studies such as Simons Searchlight (NCT01238250) and the epilepsy precision medicine study NCT05450822 [162, 163]. A negative search does not prove that no studies exist.

7.13. Investigational therapies in other genes

For TBR1, the pilot trial at the CHU de Dijon (NCT06776848) is testing lithium carbonate for 24 months in patients aged 6 years or more, with the dose increased up to the target serum lithium level and blood monitoring every 4 days until it is reached, exclusion of patients with renal or hepatic impairment, decompensated thyroid disease, long QT or Brugada syndrome, heart failure or unstabilised epilepsy, and mandatory highly effective contraception [58]. The rationale is preclinical: lithium chloride restored dendritic spine maturation and synaptogenesis in adult Tbr1 mutant mice, and zinc supplementation rescued learning in the Tbr1+/− mouse [58]. Outside the trial, the dosing reference is the lithium label for bipolar disorder from 7 years of age: 300 mg three times daily above 30 kg or 300 mg twice daily between 20 and 30 kg, titration to serum concentrations of 0.8–1.0 mEq/L for maintenance, with a lithium level on day 3 (12 hours after the last dose) and prior assessment of renal and thyroid function and electrolytes; toxicity occurs close to therapeutic concentrations (boxed warning) [159]. In Portugal, the SmPC of Priadel® (modified-release lithium carbonate) does not recommend use in children, measures the lithium level 4 to 7 days after initiation (12 or 24 hours after the previous dose), weekly until stable and then at intervals of no more than 2–3 months, with a target of 0.5–1.0 mmol/L for prophylaxis and not exceeding 1.5 mmol/L; it also includes control of aggressive and self-mutilating behaviour among its indications [160]. This use in TBR1 is experimental. For CLCN3, pharmacological inhibition of ClC-3 could benefit hyperactive variants, but no specific inhibitors exist [63].

7.14. Developments in the last 6–12 months

Between October 2025 and October 2026, developments are classified by level of evidence. Peer-reviewed publications: the retrospective series of 34 patients treated with memantine (Epilepsia, April 2026) [50]; the description of the TMEM9 gain-of-function mechanism in CLCN3 (EMBO Molecular Medicine, 2026) [63]; the cohort of 54 individuals with RERE variants and variant-specific management recommendations (Genetics in Medicine, 2026), whose full text was not accessible [62]; the review of the β-spectrinopathies with 91 patients [60]; the report of SPTBN1 mimicking autoimmune encephalitis [111]; the first randomised N-of-1 trials of L-serine in GRIN2B [151]; a letter on L-serine in GRIA3 and GRIN2B [164]; and the review of 69 patients with SCAF4 on the occasion of a prenatal diagnosis [59]. Trial registrations: the start of Beeline (radiprodil phase 3) in January 2026 [52, 157], of TAP-GRIN in August 2025 [152] and of the lithium trial in TBR1 in September 2025 [58]. Regulatory opinion: the European orphan designation of radiprodil (May 2025) [158]. No marketing authorisation has been granted for any of these diseases. ERN EpiCARE published in 2026 the European research priorities in epilepsy, including genetics, targeted therapies and innovative trial designs [165].

7.15. Agents to avoid, pregnancy, perioperative care and transition

In GRIN gain-of-function variants, NMDAR agonists and other activators, including L-serine, should be avoided, and in loss-of-function or null variants NMDAR blockers (memantine, dextromethorphan, ketamine) should be used with caution [83]. This rule should be known to anaesthesiology, because ketamine and dextromethorphan are on that list; the sources consulted do not describe specific anaesthetic recommendations for these diseases. Levetiracetam is associated with behavioural changes and psychotic symptoms, which weighs in patients with ASD, ADHD or psychosis (KDM6B, CLCN4), and vigabatrin requires visual field monitoring, which is difficult to assess in patients with cortical visual impairment or retinal dystrophy (GRIN1, CLCN3) [125, 127]. Pregnancy has been described in women with inherited SCAF4, KDM6B and RERE variants, and in a pregnant woman with Fliedner–Zweier syndrome the variant was diagnosed in the fetus [13, 59, 166]; for memantine, the label states that there are no adequate human data and that there were developmental effects in rats at doses higher than human doses [143]. Transition to adult care should be planned from 12 years of age, in its financial, vocational and medical aspects [31, 39].

7.16. Guidelines and divergences

There is no specific clinical guideline for any of the seven entities, nor a published management consensus for the GRINopathies [3, 19]. The applicable guidelines are generic and converge on diagnosis: ACMG 2021 (exome or genome as first- or second-tier testing in ID and congenital anomalies) [49], NSGC 2023 (genetic testing in all unexplained epilepsies, with sequencing before microarray) [119] and ILAE 2022 (exome or genome with CNV analysis as first line, periodic reanalysis, precision therapies by functional direction) [120]. The divergence lies in the order of testing: the 2019 RERE GeneReviews still puts microarray first, whereas the NSGC and the ILAE give priority to sequencing with CNV analysis [31, 119, 120]. The ILAE document describes current practice and does not grade the strength of its recommendations [120]. In therapy, the ILAE includes memantine, dextromethorphan and ketamine for GoF and serine for LoF in a table of precision therapies, whereas the GRIN1 GeneReviews does not recommend targeted treatment and the 2025 review classifies the evidence as limited to uncontrolled reports [1, 120, 142].

8. Reproductive Options and Family Counselling

8.1. Recurrence risk by scenario

8.2. Counselling by gene

In the de novo forms, all GRIN1 probands with tested parents had a de novo variant; germline mosaicism is theoretically possible, and in a study of mosaicism in apparently asymptomatic parents of children with DEE the frequency of parental mosaicism, somatic and inferred germline, was 10% [1, 168]. The recurrence risk for siblings is estimated at about 1% [167]. Confirmation of biological parentage and awareness that leucocyte DNA may not detect somatic or germline mosaicism are part of counselling [43], and tests sensitive to low-level mosaicism, such as allele-specific PCR or deep next-generation sequencing (NGS), may be considered [39]. The patients with GRIN1, GRIN2B and GRIN2D described have not reproduced, many because they have not reached reproductive age [1, 39].

Three genes require particularly careful parental testing, because the variant may come from a parent with a mild phenotype. In KDM6B, 64/85 variants were de novo and 9/85 inherited (five maternal, four paternal, all truncating) from parents with mild delay, learning difficulties or ASD or without clinical disease, which reflects variable expressivity and incomplete penetrance [13]. In SCAF4, 23 variants were de novo and 13 inherited, and six of eight transmitting parents had learning difficulties or mild behavioural changes; one variant was inherited from a healthy mother in the original series [23, 75]. In RERE, loss-of-function variants are more often inherited from a mildly symptomatic or asymptomatic parent, and the first transmission described was from a mother with ADHD, myopia and mild speech delay to a son with delay, autism and hypermetropia [62, 166]. In the family with the synonymous variant c.1740G>A, the mother transmitted the variant to two affected children; sequencing of maternal blood showed no mosaicism, but the authors do not exclude maternal somatic mosaicism [62]. The probabilistic models of Campbell and colleagues show that recurrence risk depends on parental origin and mosaicism: in the female germline, whose clonal expansion ends early in development, the proportion of mutant cells varies more, so an observed maternal transmission sharply increases the expected risk in another pregnancy, and somatic mosaicism raises the risk in both parents [169]. Transmission of loss-of-function variants by mildly affected or unaffected parents fits with the incomplete penetrance of haploinsufficient ID genes: Ropers and Wienker found 56 loss-of-function variants in ASXL1 in ExAC (60,706 exomes from individuals selected to exclude severe childhood disease), four of them already described as pathogenic, and similar variants in other ID genes [170], and penetrance and expressivity estimated in clinical series tend to be overestimated compared with population cohorts [171]. In TBR1, missense variants inherited from unaffected parents identified in autism studies did not disrupt protein function, unlike de novo variants, which makes inheritance from a healthy parent an argument against the pathogenicity of a missense variant in this gene [18, 74].

In the recessive forms, the parents are obligate, asymptomatic carriers, the proband's siblings have a 25% risk of being affected, and the parents' siblings have a 50% risk of being carriers; carrier detection requires knowledge of the familial variants [1]. In AR CLCN3, the parents heterozygous for the loss-of-function variant were healthy [64].

In CLCN4, about 16% of affected males have a de novo variant, and five of eight females with manifestations were de novo; heterozygous females with inherited variants mostly have normal cognition or mild learning difficulties (22/25), but three had ID and one DEE, and the X-inactivation pattern does not predict severity [46, 66]. Somatic mosaicism has been described in a female proband and in a severely affected male [46]. Of the 64 females published up to 2023, 44 were identified after the diagnosis of a relative [66].

8.3. Prenatal and preimplantation diagnosis and minors

Once the familial variant has been identified, prenatal diagnosis and PGT are possible for all modes of inheritance, and in de novo forms they may be offered because of the residual risk of germline mosaicism [1, 31]. The example of the 23-year-old pregnant woman with de novo SCAF4, whose variant was found in the fetus by amniocentesis at 19 weeks, shows the scenario of transmission by an affected mother, with a 50% risk [59]. Testing of asymptomatic sisters or daughters who are minors in CLCN4 raises the question of testing minors without immediate medical benefit, and counselling before and after any test, by a qualified professional, is recommended by the NSGC and the ILAE [119, 120]. A VUS neither confirms nor excludes the diagnosis and does not serve as a basis for prenatal diagnosis until it is reclassified [1].

9. Special Considerations: Population Diversity and Epidemiology

9.1. Frequencies and what they measure

There are no population prevalence studies for any of the entities, and all the available figures measure something else. GRIN1-NDD has unknown prevalence, with fewer than 100 patients published at the date of the GeneReviews [1]. For GRIN2B, the proportion of about 0.2% refers to tested individuals with neurodevelopmental disorder and/or childhood-onset epilepsy, not to the population [39]. For GRIN2D, an empirical incidence algorithm predicted 175 newborns with a de novo variant in the USA in 2018, and the contribution to neurodevelopmental disorders was estimated at 0.006% from two de novo variants in the largest study of de novo variants [43]. For GRIN2A, 7 of 3038 individuals screened by panel (0.23%) had a pathogenic variant [83]. For TBR1, recurrent disruptive variants in a small number of genes, TBR1 included, may together explain about 1% of cases of sporadic autism; this is a series proportion covering several genes, not a prevalence [19, 74]. For RERE, Orphanet assigns the class <1/1,000,000 [33], whereas the 1p36 deletion, of which RERE is one of the critical genes, has an estimated prevalence of 1/5000 [97]. For KDM6B, SCAF4, SPTBN1 and CLCN3, the only figures are cohort sizes (85, about 69, 29 and 26 individuals) [13, 27, 59, 63]. The sex imbalance in KDM6B (73% male) and SCAF4 (45 males to 23 females in the review of 68 cases) may reflect ascertainment bias for ASD and ID and is unexplained [13, 108].

9.2. Geography, ancestry and founder populations

The cohorts are mainly North American and European. The KDM6B cohort brought together patients from Radboudumc, the UK DDD, the 100,000 Genomes Project and North American autism consortia [13]; the TBR1 cohort originated in the French AnDDI-Rares network [18]; the SPTBN1 cohort combined US laboratories, the UDN and the DDD [27]; the memantine series included patients from the USA, Germany, the Netherlands, Italy and the United Kingdom [50]; the L-serine trial took place in Spain [51]. Outside these regions there are series from China (SCAF4, SPTBN1, GRIN2B, CLCN4, with NM_001830.4:c.823G>A p.(Val275Met) in four unrelated Chinese patients as a possible regional hotspot), Japan (CLCN3), Korea (SPTBN1) and Colombia (RERE) [60, 79, 86, 99, 106, 107, 109, 112]. Duncan's CLCN3 series included patients of Turkish, Ashkenazi Jewish, Uruguayan and mixed ancestry, and Polovitskaya's included centres in Georgia, Pakistan and several European countries [63, 64]. No founder variants have been described in any of these genes, and in CLCN4 the GeneReviews authors explicitly state the absence of founder variants [46]. The recessive forms arose in consanguineous families (GRIN1 DEE101; CLCN3 NEDSBA of Turkish origin; Pakistani infant with a GRIN1 splice variant) [11, 64].

From Brazil there is the SCAF4 report from the University of São Paulo, the second published on the gene, and the participation of Brazilian centres in the 2025 series [23, 76]; the Brazilian book of genetic syndromes lists GRIN2B among the genes of Rett-like presentations [89]. From Portugal, the sources consulted record only the study of an SPTBN1 patient at the CGC Genetics laboratory, with Portuguese co-authors, without the patient's nationality being stated [69]. The lack of national data in the sources consulted does not demonstrate their absence from the literature.

9.3. Special populations

Adults are under-represented: the oldest described are 29 years in TBR1, 26 in SPTBN1 and 25 in CLCN3, seven adults are followed in SCAF4, two individuals with RERE were alive in their early twenties and there are no adults with GRIN2D [18, 23, 27, 31, 43, 63]. Adults with a mild phenotype (transmitting parents in KDM6B, SCAF4, RERE and CLCN4 females) are generally diagnosed after their child, which has implications for reproductive health and psychiatric surveillance [13, 23, 66]. In pregnancy, there are only isolated descriptions [59].

9.4. European networks and patient associations

The sources consulted identified no specific association for KDM6B, TBR1, SCAF4, RERE or CLCN3; for these families, the European networks, Simons Searchlight and EURORDIS are the available points of contact.

10. Illustrative Clinical Case

Xu and colleagues, from Emory University, described in 2021 a boy born at term to healthy non-consanguineous parents, with a birth weight of 6 pounds 15 ounces and Apgar scores of 8 and 9, without neonatal problems [101]. In the first and second weeks of life the parents noticed "startles", with sudden extension of the arms lasting 1 to 2 seconds, at least ten times a day, which lasted until 10 weeks; at 12 weeks brief episodes of head turning to the right with rapid lateral eye movements, arm extension and leg flexion appeared, which became countless [101]. At 4 months, prolonged video-EEG monitoring showed infantile spasms and tonic seizures, with bursts of high-amplitude discharges alternating with periods of suppression of 4 to 5 seconds, compatible with early infantile epileptic encephalopathy; MRI at 3 months was normal [101]. ACTH (150 U/m²) reduced the frequency of spasms by 50%, which returned to 30–50 per day during the 4-week taper; felbamate (10 mg/kg/day, rapidly titrated up to 100 mg/kg/day), clobazam (described as 2.5 mg at night in one passage of the text and 2.5 mg/kg/day in another) and vigabatrin (175 mg/kg/day) followed, without control of the spasms or the tonic seizures, and from 10 months tonic–clonic seizures appeared two to three times a day [101]. At 12 months he was encephalopathic, without eye contact or visual tracking, with axial hypotonia and poor head control, appendicular hypertonia and spasticity and diffuse hyperreflexia; at 24 months he had intermittent chorea of the upper limbs [101].

Trio exome sequencing performed at 6 months identified the de novo heterozygous variant NM_007327.4:c.1923G>A p.(Met641Ile) in exon 13 of GRIN1, in the M3 transmembrane helix, at a residue conserved in vertebrates, absent from gnomAD and classified as likely pathogenic by ACMG criteria [4, 101]. In Xenopus oocytes and HEK cells, receptors with GluN1-M641I showed increased glycine potency, reduced Mg²⁺ block and changes in open probability and membrane trafficking, that is, gain of function; all the FDA-approved channel blockers tested (memantine, ketamine, dextromethorphan and its metabolite dextrorphan, amantadine) were 3.4 to 20 times more potent at the mutant receptors [101]. The same variant had been described in another patient with epileptic encephalopathy with onset at 2 months, cerebellar atrophy and epilepsy refractory to valproate, carbamazepine and phenobarbital [101].

After the risks and benefits had been discussed, memantine was added at 14 months, off-label, at 0.2 mg/kg/day, increased to 0.4 mg/kg/day over two weeks [101]. Three weeks later, spasms and tonic seizures had fallen from 20–30 per day to a few per day, according to the parental diary; after 3 months, spasms occurred once every 2–3 weeks, the EEG showed only mild slowing without interictal discharges, and the child was more alert and calm, with sleep improved with clonidine [101]. At 17 months, missing three doses of memantine (two days, in the legend of Figure 6) was accompanied by a brief recurrence of 5–7 spasms per day and tonic–clonic seizures, which ceased two days after resuming the drug; six months after restarting, seizures remained controlled (10–20 per week), and at the last visit he was taking memantine 0.5 mg/kg/day, vigabatrin 90 mg/kg/day and felbamate 70 mg/kg/day, with no adverse effects reported by the parents [101]. Figure 1 of the article shows the normal MRI at 3 months and the ictal EEG at 9 months, with decrement and superimposed fast activity during a typical spasm (https://pmc.ncbi.nlm.nih.gov/articles/PMC8283169/figure/acn351406-fig-0001/), and Figure 6 shows the course of seizure frequency on memantine and the EEGs with hypsarrhythmia at 4 months, multifocal discharges at 9 months and only mild slowing at 18 months (https://pmc.ncbi.nlm.nih.gov/articles/PMC8283169/figure/acn351406-fig-0006/); the legends of both were consulted [101].

The case illustrates three points: the presentation of GRIN1 with infantile spasms and normal MRI, without the polymicrogyria that characterises the S2/M3 subgroup; the value of functional characterisation of the variant before choosing a blocker; and the single-patient design, with relapse on interruption and improvement on resumption, as a form of individual evidence. The management is that of a case report: memantine has no authorised indication for GRIN diseases, GeneReviews does not recommend it and, in Karnstedt's series, the same variant did not respond in another patient [1, 50, 143].

11. Prognosis and Long-term Follow-up

11.1. Prognosis

Prognosis differs between genes and within each gene. In GRIN1, ID is severe or profound in 88%, epilepsy is refractory in 17 of 27 patients with therapeutic follow-up, and life expectancy is unknown; DEE101 is lethal between 5 days and 5 months in the cases described [1, 11]. In GRIN2D, regression with an encephalopathic course occurred in five of 22, and no adults have been described [43]. In KDM6B, developmental problems become less evident after 5 years and ID is mainly mild, but 20% of the adolescents assessed had psychosis [13]. In SCAF4, adults with a mild presentation in childhood retain mild learning difficulties, with employment and a family [23]. In TBR1, 61% have severe ID and severity does not depend on variant type [18]. In RERE, survival into adulthood is possible and prognosis depends mainly on the associated malformations [31, 62]. In CLCN3, the severe gain-of-function and recessive forms include deaths at 23 days, 14 months and 15 years, and regression occurred in one third of those assessed [63, 64]. In CLCN4, there have been seizure-related deaths in childhood and adolescence, neurological signs may progress in adulthood (31.8% of males) and epilepsy may begin at 13 years [46, 66]. In SPTBN1, three of 29 regressed and the adult course has not been described [27].

11.2. Surveillance

The GeneReviews set numerical intervals only for some assessments; for the others they state "as clinically indicated" or "at each visit", and the table reproduces that wording without inventing intervals.

12. Summary and Final Considerations

The neurodevelopmental disorders related to GRIN1, KDM6B, TBR1, SCAF4, SPTBN1, RERE and CLCN3 have been defined over the last fifteen years by sequencing of large series of children with developmental delay, and share the absence of a recognisable gestalt, an almost always de novo origin and an exclusively molecular diagnosis, by exome or genome sequencing, preferably as a trio and with CNV analysis, as recommended by the ACMG, the NSGC and the ILAE. They differ in severity and associated signs: GRIN1 and CLCN3 produce the most severe forms, with DEE, brain malformations and visual impairment; TBR1 combines severe ID with autistic traits and a thin anterior commissure; KDM6B and SCAF4 produce mainly mild ID, with psychosis in adolescence in the former and renal, cardiac and skeletal malformations in the latter; SPTBN1 stands out for speech impairment and ADHD; and RERE for ocular, cardiac and auditory anomalies and overlap with CHARGE syndrome and the 1p36 deletion. In KDM6B, SCAF4 and RERE, transmission from parents with a mild phenotype makes it necessary to test the parents before giving a recurrence risk of 1%.

Management is multidisciplinary and symptomatic for all genes, with antiseizure medicines chosen according to seizure type, rehabilitation, early augmentative communication, ophthalmological and audiological surveillance and attention to mental health. The GRIN class is the only one with targeted therapy under investigation, and the rule that guides it is the functional direction of the variant: memantine, off-label, and radiprodil, in a phase 3 trial since January 2026, for gain-of-function variants; L-serine for loss-of-function variants and nulls, with the documented risk of worsening if the rule is reversed. The available evidence is observational or from open-label and N-of-1 trials, and the ongoing controlled trials (Beeline, TAP-GRIN) and the lithium trial in TBR1 will define the place of these therapies. For CLCN3, the gain-of-function mechanism via TMEM9 opens a line of research without an available drug.

13. References

1. Platzer K, Lemke JR. GRIN1-Related Neurodevelopmental Disorder. 2019 Jun 20 [Updated 2021 Apr 1]. In: Adam MP, Bick S, Mirzaa GM, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. Available from: https://www.ncbi.nlm.nih.gov/books/NBK542807/. PMID: 31219694.
2. Online Mendelian Inheritance in Man, OMIM®. MIM 614254: NEURODEVELOPMENTAL DISORDER WITH OR WITHOUT HYPERKINETIC MOVEMENTS AND SEIZURES, AUTOSOMAL DOMINANT; NDHMSD. Baltimore (MD): Johns Hopkins University. Available from: https://omim.org/entry/614254.
3. GeneGuidance. GRIN1 — gene and conditions NDHMSD/NDHMSR. Version of 2026-09-26.
4. Freeman PJ, Wagstaff JF, Fokkema IFAC, Cutting GR, Rehm HL, Davies AC, et al. Standardizing variant naming in literature with VariantValidator to increase diagnostic rates. Nat Genet. 2024;56(11):2284-2286. doi:10.1038/s41588-024-01938-w. PMID: 39358598. Available from: https://variantvalidator.org/.
5. Online Mendelian Inheritance in Man, OMIM®. MIM 138249: GLUTAMATE RECEPTOR, IONOTROPIC, N-METHYL-D-ASPARTATE, SUBUNIT 1; GRIN1. Baltimore (MD): Johns Hopkins University. Available from: https://omim.org/entry/138249.
6. Orphanet. Autosomal dominant non-syndromic intellectual disability. ORPHA:178469. https://www.orpha.net/en/disease/detail/178469
7. Orphanet. Early infantile developmental and epileptic encephalopathy. ORPHA:1934. https://www.orpha.net/en/disease/detail/1934
8. Orphanet. Bilateral generalized polymicrogyria. ORPHA:208447. https://www.orpha.net/en/disease/detail/208447
9. Online Mendelian Inheritance in Man, OMIM®. MIM 617820: NEURODEVELOPMENTAL DISORDER WITH OR WITHOUT HYPERKINETIC MOVEMENTS AND SEIZURES, AUTOSOMAL RECESSIVE; NDHMSR. Baltimore (MD): Johns Hopkins University. Available from: https://omim.org/entry/617820.
10. Orphanet. Autosomal recessive non-syndromic intellectual disability. ORPHA:88616. https://www.orpha.net/en/disease/detail/88616
11. Online Mendelian Inheritance in Man, OMIM®. MIM 619814: DEVELOPMENTAL AND EPILEPTIC ENCEPHALOPATHY 101; DEE101. Baltimore (MD): Johns Hopkins University. Available from: https://omim.org/entry/619814.
12. Online Mendelian Inheritance in Man, OMIM®. MIM 618505: STOLERMAN NEURODEVELOPMENTAL SYNDROME; NEDSST. Baltimore (MD): Johns Hopkins University. Available from: https://omim.org/entry/618505.
13. Rots D, Jakub TE, Keung C, Jackson A, Banka S, Pfundt R, et al. The clinical and molecular spectrum of the KDM6B-related neurodevelopmental disorder. Am J Hum Genet. 2023;110(6):963-978. doi:10.1016/j.ajhg.2023.04.008. PMID: 37196654. PMC10257005.
14. GeneGuidance. KDM6B — gene and Stolerman syndrome. Version of 2026-09-26.
15. Online Mendelian Inheritance in Man, OMIM®. MIM 611577: LYSINE DEMETHYLASE 6B; KDM6B. Baltimore (MD): Johns Hopkins University. Available from: https://omim.org/entry/611577.
16. Orphanet. Non-specific syndromic intellectual disability. ORPHA:528084. https://www.orpha.net/en/disease/detail/528084
17. Online Mendelian Inheritance in Man, OMIM®. MIM 606053: INTELLECTUAL DEVELOPMENTAL DISORDER WITH AUTISM AND SPEECH DELAY; IDDAS. Baltimore (MD): Johns Hopkins University. Available from: https://omim.org/entry/606053.
18. Nambot S, Faivre L, Mirzaa G, Thevenon J, Bruel AL, Mosca-Boidron AL, et al. De novo TBR1 variants cause a neurocognitive phenotype with ID and autistic traits: report of 25 new individuals and review of the literature. Eur J Hum Genet. 2020;28(6):770-782. doi:10.1038/s41431-020-0571-6. PMID: 32005960. PMC7253452.
19. GeneGuidance. TBR1 — gene and IDDAS (OMIM 606053). Version of 2026-09-26.
20. Online Mendelian Inheritance in Man, OMIM®. MIM 604616: T-BOX, BRAIN, 1; TBR1. Baltimore (MD): Johns Hopkins University. Available from: https://omim.org/entry/604616.
21. Orphanet. Developmental delay-language impairment-dopa responsive dystonia-parkinsonism syndrome due to 2q24 microdeletion. ORPHA:1617. https://www.orpha.net/en/disease/detail/1617
22. Online Mendelian Inheritance in Man, OMIM®. MIM 620511: FLIEDNER-ZWEIER SYNDROME; FZS. Baltimore (MD): Johns Hopkins University. Available from: https://omim.org/entry/620511.
23. Schmid CM, Gregor A, Ruiz A, Manso Bazús C, Herman I, Ammouri F, et al. Further delineation of the SCAF4-associated neurodevelopmental disorder. Eur J Hum Genet. 2025;33(5):588-594. doi:10.1038/s41431-024-01760-2. PMID: 39668183. PMC12048650.
24. GeneGuidance. SCAF4 — gene and Fliedner-Zweier syndrome. Version of 2026-09-26.
25. Online Mendelian Inheritance in Man, OMIM®. MIM 616023: SR-RELATED C-TERMINAL DOMAIN-ASSOCIATED FACTOR 4; SCAF4. Baltimore (MD): Johns Hopkins University. Available from: https://omim.org/entry/616023.
26. Online Mendelian Inheritance in Man, OMIM®. MIM 619475: DEVELOPMENTAL DELAY, IMPAIRED SPEECH, AND BEHAVIORAL ABNORMALITIES; DDISBA. Baltimore (MD): Johns Hopkins University. Available from: https://omim.org/entry/619475.
27. Cousin MA, Creighton BA, Breau KA, Spillmann RC, Torti E, Dontu S, et al. Pathogenic SPTBN1 variants cause an autosomal dominant neurodevelopmental syndrome. Nat Genet. 2021;53(7):1006-1021. doi:10.1038/s41588-021-00886-z. PMID: 34211179. PMC8273149.
28. GeneGuidance. SPTBN1 — gene and DDISBA (OMIM 619475). Version of 2026-09-26.
29. Online Mendelian Inheritance in Man, OMIM®. MIM 182790: SPECTRIN, BETA, NONERYTHROCYTIC, 1; SPTBN1. Baltimore (MD): Johns Hopkins University. Available from: https://omim.org/entry/182790.
30. Online Mendelian Inheritance in Man, OMIM®. MIM 616975: NEURODEVELOPMENTAL DISORDER WITH OR WITHOUT ANOMALIES OF THE BRAIN, EYE, OR HEART; NEDBEH. Baltimore (MD): Johns Hopkins University. Available from: https://omim.org/entry/616975.
31. Scott DA, Sherr EH. RERE-Related Disorders. 2019 Mar 21. In: Adam MP, Bick S, Mirzaa GM, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. Available from: https://www.ncbi.nlm.nih.gov/books/NBK538938/. PMID: 30896913.
32. Online Mendelian Inheritance in Man, OMIM®. MIM 605226: ARGININE-GLUTAMIC ACID DIPEPTIDE REPEATS; RERE. Baltimore (MD): Johns Hopkins University. Available from: https://omim.org/entry/605226.
33. Orphanet. RERE-related neurodevelopmental syndrome. ORPHA:494344. Available from: https://www.orpha.net/en/disease/detail/494344.
34. Orphanet. 1p36 deletion syndrome. ORPHA:1606. https://www.orpha.net/en/disease/detail/1606
35. Online Mendelian Inheritance in Man, OMIM®. MIM 619512: NEURODEVELOPMENTAL DISORDER WITH HYPOTONIA AND BRAIN ABNORMALITIES; NEDHYBA. Baltimore (MD): Johns Hopkins University. Available from: https://omim.org/entry/619512.
36. GeneGuidance. CLCN3 — gene, NEDHYBA (OMIM 619512) and NEDSBA (OMIM 619517). Version of 2026-09-26.
37. Online Mendelian Inheritance in Man, OMIM®. MIM 600580: CHLORIDE CHANNEL 3; CLCN3. Baltimore (MD): Johns Hopkins University. Available from: https://omim.org/entry/600580.
38. Online Mendelian Inheritance in Man, OMIM®. MIM 619517: NEURODEVELOPMENTAL DISORDER WITH SEIZURES AND BRAIN ABNORMALITIES; NEDSBA. Baltimore (MD): Johns Hopkins University. Available from: https://omim.org/entry/619517.
39. Platzer K, Lemke JR. GRIN2B-Related Neurodevelopmental Disorder. 2018 May 31 [Updated 2021 Mar 25]. In: Adam MP, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. Available from: https://www.ncbi.nlm.nih.gov/books/NBK501979/
40. Online Mendelian Inheritance in Man, OMIM®. MIM 613970: INTELLECTUAL DEVELOPMENTAL DISORDER, AUTOSOMAL DOMINANT 6, WITH OR WITHOUT SEIZURES; MRD6. Baltimore (MD): Johns Hopkins University. Available from: https://omim.org/entry/613970.
41. Online Mendelian Inheritance in Man, OMIM®. MIM 616139: DEVELOPMENTAL AND EPILEPTIC ENCEPHALOPATHY 27; DEE27. Baltimore (MD): Johns Hopkins University. Available from: https://omim.org/entry/616139.
42. Online Mendelian Inheritance in Man, OMIM®. MIM 138252: GLUTAMATE RECEPTOR, IONOTROPIC, N-METHYL-D-ASPARTATE, SUBUNIT 2B; GRIN2B. Baltimore (MD): Johns Hopkins University. Available from: https://omim.org/entry/138252.
43. Platzer K, Krey I, Lemke JR. GRIN2D-Related Developmental and Epileptic Encephalopathy. 2022 Jul 28. In: Adam MP, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. Available from: https://www.ncbi.nlm.nih.gov/books/NBK582335/
44. Online Mendelian Inheritance in Man, OMIM®. MIM 617162: DEVELOPMENTAL AND EPILEPTIC ENCEPHALOPATHY 46; DEE46. Baltimore (MD): Johns Hopkins University. Available from: https://omim.org/entry/617162.
45. Online Mendelian Inheritance in Man, OMIM®. MIM 602717: GLUTAMATE RECEPTOR, IONOTROPIC, N-METHYL-D-ASPARTATE, SUBUNIT 2D; GRIN2D. Baltimore (MD): Johns Hopkins University. Available from: https://omim.org/entry/602717.
46. Palmer EE, Nguyen MH, Forwood C, Kalscheuer V. CLCN4-Related Neurodevelopmental Disorder. 2021 Dec 16. In: Adam MP, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. Available from: https://www.ncbi.nlm.nih.gov/books/NBK575836/
47. Online Mendelian Inheritance in Man, OMIM®. MIM 300114: RAYNAUD-CLAES SYNDROME; MRXSRC. Baltimore (MD): Johns Hopkins University. Available from: https://omim.org/entry/300114.
48. Online Mendelian Inheritance in Man, OMIM®. MIM 302910: CHLORIDE CHANNEL 4; CLCN4. Baltimore (MD): Johns Hopkins University. Available from: https://omim.org/entry/302910.
49. Manickam K, McClain MR, Demmer LA, Biswas S, Kearney HM, Malinowski J, et al. Exome and genome sequencing for pediatric patients with congenital anomalies or intellectual disability: an evidence-based clinical guideline of the American College of Medical Genetics and Genomics (ACMG). Genet Med. 2021;23(11):2029-2037. doi:10.1038/s41436-021-01242-6. PMID: 34211152.
50. Karnstedt M, Perszyk RE, Myers SJ, McDaniels E, Somorai M, Borggraefe I, et al. Memantine treatment in individuals with GRIN gain-of-function variants is associated with improvements in behavior, development, and seizure frequency. Epilepsia. 2026;67(4):1961-1974. doi:10.1002/epi.70090. PMID: 41489401. PMC13075606.
51. Juliá-Palacios N, Olivella M, Sigatullina Bondarenko M, Ibáñez-Micó S, Muñoz-Cabello B, Alonso-Luengo O, et al. L-serine treatment in patients with GRIN-related encephalopathy: a phase 2A, non-randomized study. Brain. 2024;147(5):1653-1666. doi:10.1093/brain/awae041. PMID: 38380699.
52. GRIN Therapeutics. A Multinational, Multicenter Study With an Open-Label Phase 1b and a Randomized, Double-Blind, Placebo-Controlled Phase 3 Followed by an Open-Label Extension to Assess the Efficacy, Safety, Tolerability, and Pharmacokinetics of Radiprodil in Participants With GRIN-Related Neurodevelopmental Disorder (Beeline). ClinicalTrials.gov ID: NCT07224581. Available from: https://clinicaltrials.gov/study/NCT07224581.
53. ERN EpiCARE — European Reference Network for Rare and Complex Epilepsies. Available from: https://epi-care.eu/.
54. ERN ITHACA — European Reference Network on Rare Congenital Malformations and Rare Intellectual Disability. Available from: https://ern-ithaca.eu/.
55. CureGRIN Foundation. Available from: https://curegrin.org/.
56. GRIN-Portal (Broad Institute). Available from: https://grin-portal.broadinstitute.org/.
57. Simons Searchlight. What we study. Available from: https://www.simonssearchlight.org/research/what-we-study/.
58. Centre Hospitalier Universitaire Dijon. A Pilot, Multicentre, Controlled, Open-label Study Evaluating 24 Months of Lithium Carbonate Treatment in Patients With TBR1-related Neurocognitive Disorder. ClinicalTrials.gov ID: NCT06776848. Available from: https://clinicaltrials.gov/study/NCT06776848.
59. Yan L, Chen C, Zhang Y, Cao J, Han C, Li H. [Clinical features and genetic etiology analysis in a patient with Fliedner-Zweier syndrome caused by a de novo SCAF4 variant]. Zhonghua Yi Xue Yi Chuan Xue Za Zhi. 2025;42(12):1453-1458. doi:10.3760/cma.j.cn511374-20250415-00227. PMID: 41811042.
60. Luo J, Wang T, Yan H, Zhang Y, Wang J, Ouyang S, et al. Clinical characterization of SPTBN1, SPTBN2, and SPTBN5 variants: A case series and systematic review. Seizure. 2026;137:81-90. doi:10.1016/j.seizure.2026.02.018. PMID: 41819009.
61. Beta Spectrin Foundation (SPTBN1). Available from: https://www.sptbn1.org/.
62. Curtis D, Zhao X, Owen NM, Azamian MS, Lalani SR, Keller K, et al. New genotype-phenotype correlations and management recommendations for individuals with RERE variants. Genet Med. 2026;28(6):102580. PMID: 41988794.
63. Polovitskaya MM, Tkemaladze T, Jensen L, Kar S, Planells-Cases R, Agolini E, et al. Abnormal ClC-3/TMEM9-mediated endosomal ion transport in CLCN3-associated neurodevelopmental disease. EMBO Mol Med. 2026;18(9):3474-3494. doi:10.1038/s44321-026-00486-6. PMID: 42486908. PMC13562645.
64. Duncan AR, Polovitskaya MM, Gaitán-Peñas H, Bertelli S, VanNoy GE, Grant PE, et al. Unique variants in CLCN3, encoding an endosomal anion/proton exchanger, underlie a spectrum of neurodevelopmental disorders. Am J Hum Genet. 2021;108(8):1450-1465. doi:10.1016/j.ajhg.2021.06.003. PMID: 34186028. PMC8387284.
65. GRIN2B Foundation. Available from: https://www.grin2b.com/.
66. Palmer EE, Pusch M, Picollo A, Forwood C, Nguyen MH, Suckow V, et al. Functional and clinical studies reveal pathophysiological complexity of CLCN4-related neurodevelopmental condition. Mol Psychiatry. 2023;28(2):668-697. doi:10.1038/s41380-022-01852-9. PMID: 36385166. PMC9908558.
67. Cure CLCN4. Available from: https://www.cureclcn4.org/.
68. García-Cazorla Á, Artuch R, Pearl PL. Disorders of Neurotransmission. In: Saudubray JM, Baumgartner MR, García-Cazorla Á, Walter J, editors. Inborn Metabolic Diseases: Diagnosis and Treatment. 7th ed. Berlin: Springer; 2022. Chapter 30, section 30.2 (30.2.1 Glutamate Receptor Mutations; Table 30.2). doi:10.1007/978-3-662-63123-2_30.
69. Rosenfeld JA, Xiao R, Bekheirnia MR, Kanani F, Parker MJ, Koenig MK, et al. Heterozygous variants in SPTBN1 cause intellectual disability and autism. Am J Med Genet A. 2021;185(7):2037-2045. doi:10.1002/ajmg.a.62201. PMID: 33847457. PMC11182376.
70. Hamdan FF, Gauthier J, Araki Y, Lin DT, Yoshizawa Y, Higashi K, et al. Excess of de novo deleterious mutations in genes associated with glutamatergic systems in nonsyndromic intellectual disability. Am J Hum Genet. 2011;88(3):306-316. doi:10.1016/j.ajhg.2011.02.001. PMID: 21376300. PMC3059427.
71. Lemke JR, Geider K, Helbig KL, Heyne HO, Schütz H, Hentschel J, et al. Delineating the GRIN1 phenotypic spectrum: A distinct genetic NMDA receptor encephalopathy. Neurology. 2016;86(23):2171-2178. doi:10.1212/wnl.0000000000002740. PMID: 27164704. PMC4898312.
72. Fry AE, Fawcett KA, Zelnik N, Yuan H, Thompson BAN, Shemer-Meiri L, et al. De novo mutations in GRIN1 cause extensive bilateral polymicrogyria. Brain. 2018;141(3):698-712. doi:10.1093/brain/awx358. PMID: 29365063. PMC5837214.
73. Stolerman ES, Francisco E, Stallworth JL, Jones JR, Monaghan KG, Keller-Ramey J, et al. Genetic variants in the KDM6B gene are associated with neurodevelopmental delays and dysmorphic features. Am J Med Genet A. 2019;179(7):1276-1286. doi:10.1002/ajmg.a.61173. PMID: 31124279.
74. Deriziotis P, O'Roak BJ, Graham SA, Estruch SB, Dimitropoulou D, Bernier RA, et al. De novo TBR1 mutations in sporadic autism disrupt protein functions. Nat Commun. 2014;5:4954. doi:10.1038/ncomms5954. PMID: 25232744. PMC4212638.
75. Fliedner A, Kirchner P, Wiesener A, van de Beek I, Waisfisz Q, van Haelst M, et al. Variants in SCAF4 Cause a Neurodevelopmental Disorder and Are Associated with Impaired mRNA Processing. Am J Hum Genet. 2020;107(3):544-554. doi:10.1016/j.ajhg.2020.06.019. PMID: 32730804. PMC7477272.
76. Carvalho LML, Pinto CF, de Oliveira Scliar M, Otto PA, Krepischi ACV, Rosenberg C. SCAF4-related syndromic intellectual disability. Am J Med Genet A. 2023;191(2):570-574. doi:10.1002/ajmg.a.63032. PMID: 36333968.
77. Fregeau B, Kim BJ, Hernández-García A, Jordan VK, Cho MT, Schnur RE, et al. De Novo Mutations of RERE Cause a Genetic Syndrome with Features that Overlap Those Associated with Proximal 1p36 Deletions. Am J Hum Genet. 2016;98(5):963-970. doi:10.1016/j.ajhg.2016.03.002. PMID: 27087320. PMC4863473.
78. Jordan VK, Fregeau B, Ge X, Giordano J, Wapner RJ, Balci TB, et al. Genotype-phenotype correlations in individuals with pathogenic RERE variants. Hum Mutat. 2018;39(5):666-675. doi:10.1002/humu.23400. PMID: 29330883. PMC5903952.
79. Nakashima M, Argilli E, Nakano S, Sherr EH, Kato M, Saitsu H. De novo CLCN3 variants affecting Gly327 cause severe neurodevelopmental syndrome with brain structural abnormalities. J Hum Genet. 2023;68(4):291-298. doi:10.1038/s10038-022-01106-0. PMID: 36536096. PMC12327483.
80. XiangWei W, Jiang Y, Yuan H. De Novo Mutations and Rare Variants Occurring in NMDA Receptors. Curr Opin Physiol. 2018;2:27-35. doi:10.1016/j.cophys.2017.12.013. PMID: 29756080. PMC5945193.
81. National Center for Biotechnology Information. ClinVar. Bethesda (MD): National Library of Medicine. Available from: https://www.ncbi.nlm.nih.gov/clinvar/.
82. Myers SJ, Yuan H, Perszyk RE, Zhang J, Kim S, Nocilla KA, et al. Classification of missense variants in the N-methyl-d-aspartate receptor GRIN gene family as gain- or loss-of-function. Hum Mol Genet. 2023;32(19):2857-2871. doi:10.1093/hmg/ddad104. PMID: 37369021. PMC10508039.
83. Strehlow V, Myers KA, Morgan AT, Scheffer IE, Lemke JR. GRIN2A-Related Disorders. 2016 Sep 29 [Updated 2024 Jul 4]. In: Adam MP, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. Available from: https://www.ncbi.nlm.nih.gov/books/NBK385627/
84. Strehlow V, Heyne HO, Vlaskamp DRM, Marwick KFM, Rudolf G, de Bellescize J, et al. GRIN2A-related disorders: genotype and functional consequence predict phenotype. Brain. 2019;142(1):80-92. doi:10.1093/brain/awy304. PMID: 30544257. PMC6308310.
85. Platzer K, Yuan H, Schütz H, Winschel A, Chen W, Hu C, et al. GRIN2B encephalopathy: novel findings on phenotype, variant clustering, functional consequences and treatment aspects. J Med Genet. 2017;54(7):460-470. doi:10.1136/jmedgenet-2016-104509. PMID: 28377535. PMC5656050.
86. Xie C, Kessi M, Liu F, He F, Peng J. GRIN2B-related neurodevelopmental disorders: genotype-phenotype correlations and therapeutic implications. Orphanet J Rare Dis. 2025;20(1):541. doi:10.1186/s13023-025-04055-x. PMID: 41146259. PMC12560415.
87. Brock S, Laquerriere A, Marguet F, Myers SJ, Hongjie Y, Baralle D, et al. Overlapping cortical malformations in patients with pathogenic variants in GRIN1 and GRIN2B. J Med Genet. 2023;60(2):183-192. doi:10.1136/jmedgenet-2021-107971. PMID: 35393335. PMC10642159.
88. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 412–413 (Epilepsy in infants and children).
89. Ribeiro EM, coord. Síndromes Genéticas: Reconhecer para Tratar. Fortaleza: EdUnichristus; 2021. p. 118–119 (Síndrome de Rett).
90. National Center for Biotechnology Information. ClinVar; VCV000224144 — NM_006593.4(TBR1):c.1588_1594dup (p.Thr532fs). Available from: https://www.ncbi.nlm.nih.gov/clinvar/variation/224144/.
91. GeneGuidance. RERE — gene and NEDBEH (OMIM 616975). Version of 2026-09-26.
92. Kim BJ, Zaveri HP, Shchelochkov OA, Yu Z, Hernández-García A, Seymour ML, et al. An allelic series of mice reveals a role for RERE in the development of multiple organs affected in chromosome 1p36 deletions. PLoS One. 2013;8(2):e57460. doi:10.1371/journal.pone.0057460. PMID: 23451234. PMC3581587.
93. Kim BJ, Scott DA. Mouse model reveals the role of RERE in cerebellar foliation and the migration and maturation of Purkinje cells. PLoS One. 2014;9(1):e87518. doi:10.1371/journal.pone.0087518. PMID: 24466353. PMC3900724.
94. Kim BJ, Zaveri HP, Jordan VK, Hernandez-Garcia A, Jacob DJ, Zamora DL, et al. RERE deficiency leads to decreased expression of GATA4 and the development of ventricular septal defects. Dis Model Mech. 2018;11(9):dmm031534. doi:10.1242/dmm.031534. PMID: 30061196. PMC6176990.
95. Kim BJ, Zaveri HP, Kundert PN, Jordan VK, Scott TM, Carmichael J, et al. RERE deficiency contributes to the development of orofacial clefts in humans and mice. Hum Mol Genet. 2021;30(7):595-602. doi:10.1093/hmg/ddab084. PMID: 33772547.
96. Kim BJ, Scott DA. RERE deficiency causes retinal and optic nerve atrophy through degeneration of retinal cells. Dev Dyn. 2021;250(10):1398-1409. doi:10.1002/dvdy.330. PMID: 33742727.
97. Jones KL, Jones MC, del Campo M. Smith's Recognizable Patterns of Human Malformation. 8th ed. Philadelphia: Elsevier; 2022. p. 76–78 (1p36 Deletion Syndrome).
98. Tettey-Matey A, Picollo A, Sbrana F, Coppola MA, Rubino E, Giusto A, et al. Dominant Action of CLCN4 Neurodevelopmental Disease Variants in Heteromeric Endosomal ClC-3/ClC-4 Transporters. Cells. 2025;14(24):1973. doi:10.3390/cells14241973. PMID: 41439993. PMC12731782.
99. He H, Li X, Guzman GA, Bungert-Plümke S, Franzen A, Lin X, et al. Expanding the genetic and phenotypic relevance of CLCN4 variants in neurodevelopmental condition: 13 new patients. J Neurol. 2024;271(8):4933-4948. doi:10.1007/s00415-024-12383-4. PMID: 38758281.
100. Sahly AN, Sierra-Marquez J, Bungert-Plümke S, Franzen A, Mougharbel L, Berrahmoune S, et al. Genotype-phenotype correlation in CLCN4-related developmental and epileptic encephalopathy. Hum Genet. 2024;143(5):667-681. doi:10.1007/s00439-024-02668-z. PMID: 38578438.
101. Xu Y, Song R, Chen W, Strong K, Shrey D, Gedela S, et al. Recurrent seizure-related GRIN1 variant: Molecular mechanism and targeted therapy. Ann Clin Transl Neurol. 2021;8(7):1480-1494. doi:10.1002/acn3.51406. PMID: 34227748. PMC8283169.
102. Colijn MA, Stowe RM. The curious absence of psychosis in GRIN1-related neurodevelopmental disorder. Eur Arch Psychiatry Clin Neurosci. 2025;275(8):2567-2568. doi:10.1007/s00406-024-01796-x. PMID: 38483682.
103. Papa FT, Mancardi MM, Frullanti E, Fallerini C, Della Chiara V, Zalba-Jadraque L, et al. Personalized therapy in a GRIN1 mutated girl with intellectual disability and epilepsy. Clin Dysmorphol. 2018;27(1):18-20. doi:10.1097/mcd.0000000000000205. PMID: 29194067.
104. Rossi M, Chatron N, Labalme A, Ville D, Carneiro M, Edery P, et al. Novel homozygous missense variant of GRIN1 in two sibs with intellectual disability and autistic features without epilepsy. Eur J Hum Genet. 2017;25(3):376-380. doi:10.1038/ejhg.2016.163. PMID: 28051072. PMC5315503.
105. Obregón Gómez LR, Juanes M, Touzon MS, Romero A, Maenza L, Borgani K, et al. GRIN2D-Related Developmental and Epileptic Encephalopathy Associated With Polymorphic Seizures Including Epileptic Spasms. Int J Dev Neurosci. 2026;86(1):e70108. doi:10.1002/jdn.70108. PMID: 41673952.
106. Hu Y, Zhang B, Chen L, He J, Yang L, Chen X. SCAF4 variants are associated with epilepsy with neurodevelopmental disorders. Seizure. 2024;116:113-118. doi:10.1016/j.seizure.2023.10.008. PMID: 37891035.
107. Lin H, Chen YH. SCAF4 variants associated with focal epilepsy accompanied by multisystem disorders. Seizure. 2024;116:65-73. doi:10.1016/j.seizure.2023.06.018. PMID: 37394306.
108. Chen Z, Zhao J, Fan X, Xuan X, Zhao X. A novel nonsense mutation in SCAF4 associated with fliedner-zweier syndrome: a case report and review of the literature. Front Genet. 2025;16:1487352. doi:10.3389/fgene.2025.1487352. PMID: 40290495. PMC12021868.
109. Jang HN, Ryu J, Kim SS, Moon JH. A Case of Infantile Epileptic Spasms Syndrome with the SPTBN1 Mutation and Review of βII-Spectrin Variants. Genes (Basel). 2025;16(8):904. doi:10.3390/genes16080904. PMID: 40869952. PMC12385928.
110. O'Connell M, Harstad E, Aites J, Hayes K, Arnett AB, Scotellaro J, et al. Diverse clinical presentation of SPTBN1 variants: Complex versus primary attention-deficit/hyperactivity disorder. Am J Med Genet A. 2025;197(1):e63851. doi:10.1002/ajmg.a.63851. PMID: 39162370.
111. Ramani PK, Gonzalez-Garcia A, Appleton T, Beemarajan E, Arya K. Novel SPTBN1 Nonsense Variant Presenting as Apparent Autoimmune Encephalitis: A Diagnostic Lesson in New-Onset Refractory Seizures. Pediatr Neurol. 2026;185:75-76. doi:10.1016/j.pediatrneurol.2026.09.009. PMID: 42837906.
112. Pachajoa H, Gutierrez-Obando AM, Leal AF. Expanding the Clinical Spectrum of RERE-Related Disorders: A Case Report of Neurodevelopmental Disorder with Brain Malformations Including Chiari Type I. Mol Syndromol. 2025. doi:10.1159/000550194. PMID: 41669386. PMC12885554.
113. Cai T, Li C, Chen L, Zeng H, Lu C, Yu Y. A de novo variant of RERE was identified in a patient with neurodevelopmental disorder, enuresis and scoliosis. Sci Rep. 2025;15(1):32179. doi:10.1038/s41598-025-17628-4. PMID: 40890291. PMC12402444.
114. Li Q, Li W, Hu K, Wang Y, Li Y, Xu J. A de novo variant in RERE causes autistic behavior by disrupting related genes and signaling pathway. Clin Genet. 2024;105(3):273-282. doi:10.1111/cge.14461. PMID: 38018232.
115. Garrett A, Kalscheuer VM, Figueroa RR, Palmer EE, Morgan AT. CLCN4 -Related Neurodevelopmental Condition: Characterization of Speech and Language Abilities. Am J Med Genet A. 2025;197(12):e64192. doi:10.1002/ajmg.a.64192. PMID: 40762462.
116. Saudubray JM, García-Cazorla Á. Clinical Approach to Inborn Errors of Metabolism in Paediatrics. In: Saudubray JM, Baumgartner MR, García-Cazorla Á, Walter J, editors. Inborn Metabolic Diseases: Diagnosis and Treatment. 7th ed. Berlin: Springer; 2022. Chapter 1, sections 1.3.4 and 1.5 (Tables 1.28–1.29). doi:10.1007/978-3-662-63123-2_1.
117. Eiler I, Reecher HM, Carlton K, Cabacungan E, Cohen S, Adams S, et al. GRIN1-related epilepsy in a neonate with response to memantine and vigabatrin. Ann Child Neurol Soc. 2024;2(4):299-302. doi:10.1002/cns3.20088. PMID: 42564024. PMC13359329.
118. Lam Z, Wall E, Ryan G, Barber R, Kilby MD, Williams DK. Prenatal diagnosis of CLCN4-related neurodevelopmental disorder in fetuses with congenital brain anomalies. Prenat Diagn. 2023;43(9):1247-1250. doi:10.1002/pd.6404. PMID: 37409888.
119. Smith L, Malinowski J, Ceulemans S, Peck K, Walton N, Sheidley BR, et al. Genetic testing and counseling for the unexplained epilepsies: An evidence-based practice guideline of the National Society of Genetic Counselors. J Genet Couns. 2023;32(2):266-280. doi:10.1002/jgc4.1646. PMID: 36281494.
120. Krey I, Platzer K, Esterhuizen A, Berkovic SF, Helbig I, Hildebrand MS, et al. Current practice in diagnostic genetic testing of the epilepsies. Epileptic Disord. 2022;24(5):765-786. doi:10.1684/epd.2022.1448. PMID: 35830287. PMC10752379.
121. GRIN Database. Available from: https://www.grin-database.de/.
122. Orphanet. Anti-NMDA receptor encephalitis. ORPHA:217253. https://www.orpha.net/en/disease/detail/217253
123. O'Hara KA. First aid for seizures: the importance of education and appropriate response. J Child Neurol. 2007;22(5 Suppl):30S-37S. doi:10.1177/0883073807303066. PMID: 17690085.
124. Sager G, Yukselmis U, Güzel O, Turkyilmaz A, Akcay M. Experience with the Ketogenic Diet in a Boy with CLCN4 Related Neurodevelopmental Disorder. Balkan J Med Genet. 2023;26(2):77-82. doi:10.2478/bjmg-2023-0013. PMID: 38482266. PMC10932601.
125. UCB, Inc. KEPPRA® (levetiracetam) tablets and oral solution — Prescribing Information. NDA 021035. openFDA set_id 3ca9df05-a506-4ec8-a4fe-320f1219ab21 (version 2026-01-28). Available from: https://dailymed.nlm.nih.gov/dailymed/lookup.cfm?setid=3ca9df05-a506-4ec8-a4fe-320f1219ab21.
126. GlaxoSmithKline. LAMICTAL® (lamotrigine) tablets, chewable dispersible tablets and orally disintegrating tablets — Prescribing Information. NDA 020241. openFDA set_id d7e3572d-56fe-4727-2bb4-013ccca22678 (version 2025-10-10). Available from: https://dailymed.nlm.nih.gov/dailymed/lookup.cfm?setid=d7e3572d-56fe-4727-2bb4-013ccca22678.
127. Lundbeck. SABRIL® (vigabatrin) tablets and powder for oral solution — Prescribing Information. NDA 020427. openFDA set_id a5d389d2-d0e1-4395-a2a2-b552808e7f98 (version 2021-10-20). Available from: https://dailymed.nlm.nih.gov/dailymed/lookup.cfm?setid=a5d389d2-d0e1-4395-a2a2-b552808e7f98.
128. Lundbeck. ONFI® (clobazam) tablets and oral suspension — Prescribing Information. NDA 202067 and 203993. openFDA set_id de03bd69-2dca-459c-93b4-541fd3e9571c (version 2024-03-12). Available from: https://dailymed.nlm.nih.gov/dailymed/lookup.cfm?setid=de03bd69-2dca-459c-93b4-541fd3e9571c.
129. European Medicines Agency. Keppra (levetiracetam) — EPAR, product information (summary of product characteristics, Portuguese version). EU/1/00/146; first authorisation 29/09/2000. Available from: https://www.ema.europa.eu/pt/documents/product-information/keppra-epar-product-information_pt.pdf.
130. INFARMED. Lamictal (lamotrigina) 25 mg — Resumo das Características do Medicamento. Infomed, med_id 4876. Available from: https://extranet.infarmed.pt/INFOMED-fo/download-ficheiro.xhtml?med_id=4876&tipo_doc=rcm.
131. INFARMED. Sabril 500 mg comprimidos revestidos por película (vigabatrina) — Resumo das Características do Medicamento. Infomed. Available from: https://extranet.infarmed.pt/INFOMED-fo/.
132. INFARMED. Castilium 10 mg comprimidos (clobazam) — Resumo das Características do Medicamento (approved 10/01/2024). Infomed, med_id 1494. Available from: https://extranet.infarmed.pt/INFOMED-fo/download-ficheiro.xhtml?med_id=1494&tipo_doc=rcm.
133. OZOBAX® DS (baclofen) oral solution 10 mg/5 mL — Prescribing Information. NDA 208193. openFDA set_id d8ddcec5-debe-42b4-a803-e65917cc6a54 (version 2026-03-13). Available from: https://dailymed.nlm.nih.gov/dailymed/lookup.cfm?setid=d8ddcec5-debe-42b4-a803-e65917cc6a54.
134. ZANAFLEX® (tizanidine hydrochloride) capsules and tablets — Prescribing Information. NDA 020397 and 021447. openFDA set_id 043d9e51-bfa2-4add-9058-5ece332f7e99 (version 2025-05-05). Available from: https://dailymed.nlm.nih.gov/dailymed/lookup.cfm?setid=043d9e51-bfa2-4add-9058-5ece332f7e99.
135. Allergan/AbbVie. BOTOX® (onabotulinumtoxinA) for injection — Prescribing Information. BLA 103000. openFDA set_id 33d066a9-34ff-4a1a-b38b-d10983df3300 (version 2023-11-18). Available from: https://dailymed.nlm.nih.gov/dailymed/lookup.cfm?setid=33d066a9-34ff-4a1a-b38b-d10983df3300.
136. INFARMED. Lioresal 10 mg e 25 mg comprimidos (baclofeno) — Resumo das Características do Medicamento. Infomed, med_id 5074. Available from: https://extranet.infarmed.pt/INFOMED-fo/download-ficheiro.xhtml?med_id=5074&tipo_doc=rcm.
137. INFARMED. Sirdalud 2 mg comprimidos e Sirdalud MR 6 mg cápsulas de libertação modificada (tizanidina) — Resumo das Características do Medicamento. Infomed. Available from: https://extranet.infarmed.pt/INFOMED-fo/.
138. INFARMED. BOTOX 100 Unidades Allergan pó para solução injetável (toxina botulínica A) — Resumo das Características do Medicamento. Infomed, med_id 29406. Available from: https://extranet.infarmed.pt/INFOMED-fo/download-ficheiro.xhtml?med_id=29406&tipo_doc=rcm.
139. Urbančič J, Vozel D, Battelino S, Boršoš I, Bregant L, Glavan M, et al. Management of Choanal Atresia: National Recommendations with a Comprehensive Literature Review. Children (Basel). 2023;10(1):91. doi:10.3390/children10010091. PMID: 36670642. PMC9856561.
140. Zaveri HP, Beck TF, Hernández-García A, Shelly KE, Montgomery T, van Haeringen A, et al. Identification of critical regions and candidate genes for cardiovascular malformations and cardiomyopathy associated with deletions of chromosome 1p36. PLoS One. 2014;9(1):e85600. doi:10.1371/journal.pone.0085600. PMID: 24454898. PMC3893250.
141. Krey I, von Spiczak S, Johannesen KM, Hikel C, Kurlemann G, Muhle H, et al. L-Serine Treatment is Associated with Improvements in Behavior, EEG, and Seizure Frequency in Individuals with GRIN-Related Disorders Due to Null Variants. Neurotherapeutics. 2022;19(1):334-341. doi:10.1007/s13311-021-01173-9. PMID: 34997442. PMC9130352.
142. Wang S, Perucca E, Berkovic SF, Perucca P. Precision therapies for genetic epilepsies in 2025: Promises and pitfalls. Epilepsia Open. 2025. doi:10.1002/epi4.70065. PMID: 40411479. PMC13394099.
143. Allergan/AbbVie. NAMENDA® (memantine hydrochloride) tablets and oral solution — Prescribing Information. NDA 021487 (initial approval 16/10/2003). openFDA set_id 72973b46-f307-4a9e-8fb1-ad2cbccfd736 (version 2018-11-01). Available from: https://www.accessdata.fda.gov/scripts/cder/daf/index.cfm?event=overview.process&ApplNo=021487.
144. European Medicines Agency. Ebixa (memantine) — EPAR, product information (summary of product characteristics, Portuguese version; first authorisation 15/05/2002); Axura (memantine), authorisation of 17/05/2002. Available from: https://www.ema.europa.eu/pt/documents/product-information/ebixa-epar-product-information_pt.pdf.
145. Pierson TM, Yuan H, Marsh ED, Fuentes-Fajardo K, Adams DR, Markello T, et al. GRIN2A mutation and early-onset epileptic encephalopathy: personalized therapy with memantine. Ann Clin Transl Neurol. 2014;1(3):190-198. doi:10.1002/acn3.39. PMID: 24839611. PMC4019449.
146. Li D, Yuan H, Ortiz-Gonzalez XR, Marsh ED, Tian L, McCormick EM, et al. GRIN2D Recurrent De Novo Dominant Mutation Causes a Severe Epileptic Encephalopathy Treatable with NMDA Receptor Channel Blockers. Am J Hum Genet. 2016;99(4):802-816. doi:10.1016/j.ajhg.2016.07.013. PMID: 27616483. PMC5065652.
147. Lewis SA, Shetty S, Gamble S, Heim J, Zhao N, Stitt G, et al. Intrathecal magnesium delivery for Mg++-insensitive NMDA receptor activity due to GRIN1 mutation. Orphanet J Rare Dis. 2023;18(1):225. doi:10.1186/s13023-023-02756-9. PMID: 37537625. PMC10398931.
148. Soto D, Olivella M, Grau C, Armstrong J, Alcon C, Gasull X, et al. L-Serine dietary supplementation is associated with clinical improvement of loss-of-function GRIN2B-related pediatric encephalopathy. Sci Signal. 2019;12(586):eaaw0936. doi:10.1126/scisignal.aaw0936. PMID: 31213567.
149. Fundació Sant Joan de Déu. Tolerability and Efficacy of L-Serine in Patients With GRIN-related Encephalopathy. ClinicalTrials.gov ID: NCT04646447. Available from: https://clinicaltrials.gov/study/NCT04646447.
150. den Hollander B, Rothuizen-Lindenschot M, Geertjens L, Vaz FM, Brands MM, Le HL, et al. Effectiveness of L-serine supplementation in children with a GRIN2B loss-of-function mutation: Rationale and protocol for single patient (n-of-1) multiple cross-over trials. Contemp Clin Trials Commun. 2023;36:101233. doi:10.1016/j.conctc.2023.101233. PMID: 38144875. PMC10746402.
151. den Hollander B, Rothuizen-Lindenschot M, Le HL, Ramautar JR, Müller AR, Geertjens L, et al. Potential benefits of l-serine in children with GRIN2B loss-of-function variants: Randomized n-of-1 trials. Mol Genet Metab. 2025;146(4):109268. doi:10.1016/j.ymgme.2025.109268. PMID: 41265180.
152. Meyer Children's Hospital IRCCS. L-serine Supplementation in Patients With GRIN-related Neurodevelopmental Disorders: Multicentre Protocol for an Aggregated Series of Randomised, Placebo-controlled N-of-1 Trials (TAP-GRIN). ClinicalTrials.gov ID: NCT07377032. Available from: https://clinicaltrials.gov/study/NCT07377032.
153. Auvin S, Dozières-Puyravel B, Avbersek A, Sciberras D, Collier J, Leclercq K, et al. Radiprodil, a NR2B negative allosteric modulator, from bench to bedside in infantile spasm syndrome. Ann Clin Transl Neurol. 2020;7(3):343-352. doi:10.1002/acn3.50998. PMID: 32106360. PMC7085998.
154. UCB Biopharma. An Open-label Adaptive Study for the Assessment of Safety, Tolerability, Pharmacokinetics, and Efficacy of Multiple Doses of Radiprodil in Subjects With Drug-resistant Infantile Spasms. ClinicalTrials.gov ID: NCT02829827. Available from: https://clinicaltrials.gov/study/NCT02829827.
155. GRIN Therapeutics. A Multicenter Study to Assess the Safety, Tolerability, Pharmacokinetics, and Effect on Seizures and Behavioral Symptoms of Multiple Individually Titrated Doses of Radiprodil in Children with GRIN-related Disorder (Honeycomb). ClinicalTrials.gov ID: NCT05818943. Available from: https://clinicaltrials.gov/study/NCT05818943.
156. GRIN Therapeutics. GRIN Therapeutics announces positive topline data from Honeycomb trial of radiprodil in GRIN-related neurodevelopmental disorder [press release, PR Newswire]. New York; 2024 Sep 9. Available from: https://www.biospace.com/press-releases/grin-therapeutics-announces-positive-topline-data-from-honeycomb-trial-of-radiprodil-in-grin-related-neurodevelopmental-disorder.
157. Phase 3 Beeline trial begins evaluating radiprodil in GRIN-related neurodevelopmental disorder. Contemporary Pediatrics [Internet]. 2026 Jan 9. Available from: https://www.contemporarypediatrics.com/view/phase-3-beeline-trial-begins-evaluating-radiprodil-in-grin-related-neurodevelopmental-disorder.
158. European Medicines Agency. EU/3/25/3061 — Orphan designation for the treatment of GRIN-related neurodevelopmental disorder (radiprodil). Designated 22/05/2025; published 25/06/2025. Available from: https://www.ema.europa.eu/en/medicines/human/orphan-designations.
159. Lithium carbonate capsules — Prescribing Information. ANDA 090702. openFDA set_id 01c4facd-ed79-4078-ba33-2044de372d0f (version 2026-04-27). Available from: https://dailymed.nlm.nih.gov/dailymed/lookup.cfm?setid=01c4facd-ed79-4078-ba33-2044de372d0f.
160. INFARMED. Priadel 400 mg comprimidos de libertação modificada (carbonato de lítio) — Resumo das Características do Medicamento (approved 15/11/2022). Infomed. Available from: https://extranet.infarmed.pt/INFOMED-fo/.
161. GRIN Therapeutics. A Phase 1, Open-label, 2-part Study to Evaluate the Pharmacokinetics and Safety Effects Following Coadministration of a CYP3A4 Inducer (Carbamazepine) or Inhibitor (Itraconazole), With Radiprodil in Healthy Adult Participants. ClinicalTrials.gov ID: NCT06965881. Available from: https://clinicaltrials.gov/study/NCT06965881.
162. Simons Searchlight. Online Study of People Who Have Genetic Changes and Features of Autism: Simons Searchlight. ClinicalTrials.gov ID: NCT01238250. Available from: https://clinicaltrials.gov/study/NCT01238250.
163. Knudsen GM (responsible party). The BrainDrugs-Epilepsy Study: A Prospective Open-label Cohort Precision Medicine Study in Epilepsy. ClinicalTrials.gov ID: NCT05450822. Available from: https://clinicaltrials.gov/study/NCT05450822.
164. Kim JS, Stone A, Cross J, Grinspan ZM. L-Serine for GRIA3 and GRIN2B-related neurodevelopmental disorders. Brain. 2026;149(8):e62-e64. doi:10.1093/brain/awag081. PMID: 41742885.
165. Tchaicha S, Auvin S, Beniczky S, Brunklaus A, Lagae L, Perucca E, et al. Areas of research priorities in epilepsy: A position paper of the European Reference Network for Rare and Complex Epilepsies, EpiCARE. Epilepsia Open. 2026;11(1):291-321. doi:10.1002/epi4.70210. PMID: 41489013. PMC12903829.
166. Niehaus AD, Kim J, Manning MA. Phenotypic variability in RERE-related disorders and the first report of an inherited variant. Am J Med Genet A. 2022;188(11):3358-3363. doi:10.1002/ajmg.a.62952. PMID: 36053530.
167. Rahbari R, Wuster A, Lindsay SJ, Hardwick RJ, Alexandrov LB, Turki SA, et al. Timing, rates and spectra of human germline mutation. Nat Genet. 2016;48(2):126-133. doi:10.1038/ng.3469. PMID: 26656846. PMC4731925.
168. Myers CT, Hollingsworth G, Muir AM, Schneider AL, Thuesmunn Z, Knupp A, et al. Parental Mosaicism in "De Novo" Epileptic Encephalopathies. N Engl J Med. 2018;378(17):1646-1648. doi:10.1056/nejmc1714579. PMID: 29694806. PMC5966016.
169. Campbell IM, Stewart JR, James RA, Lupski JR, Stankiewicz P, Olofsson P, et al. Parent of origin, mosaicism, and recurrence risk: probabilistic modeling explains the broken symmetry of transmission genetics. Am J Hum Genet. 2014;95(4):345-359. doi:10.1016/j.ajhg.2014.08.010. PMID: 25242496. PMC4185125.
170. Ropers HH, Wienker T. Penetrance of pathogenic mutations in haploinsufficient genes for intellectual disability and related disorders. Eur J Med Genet. 2015;58(12):715-718. doi:10.1016/j.ejmg.2015.10.007. PMID: 26506440.
171. Kingdom R, Wright CF. Incomplete Penetrance and Variable Expressivity: From Clinical Studies to Population Cohorts. Front Genet. 2022;13:920390. doi:10.3389/fgene.2022.920390. PMID: 35983412. PMC9380816.
172. EURORDIS — Rare Diseases Europe. Available from: https://www.eurordis.org/.