Autosomal Dominant Intellectual Developmental Disorders due to TRIO, NAA15, ASH1L, CAMK2B and TAOK1

Class article: single-gene autosomal dominant intellectual disability, with a uniform subsection for each gene

Literature and regulatory information updated: 28 October 2026.

0. Reference Summary Table

The entities brought together share the same clinical and genetic model: a neurodevelopmental disorder with intellectual disability (ID) or global developmental delay (GDD), caused by a heterozygous variant in a single gene, almost always de novo, without a biochemical marker and diagnosed by exome or genome sequencing. The first table sets out the identifiers; the second summarises the clinical and management data. OMIM designates the Online Mendelian Inheritance in Man catalogue; ORPHA, the Orphanet code; MRD, the OMIM series “intellectual developmental disorder, autosomal dominant”; ERN-ITHACA, the European Reference Network for rare malformation syndromes and rare neurodevelopmental disorders and ID. The phenotype and gene numbers, titles, inheritance and cytobands were confirmed in the six OMIM phenotype entries (last edited between November 2021 and May 2024) and match NCBI MedGen [1–7].

1. Introduction

1.1. Single-gene autosomal dominant intellectual disability as a class

The term “intellectual disability” replaced “mental retardation”, abandoned in the United States by federal law; in the United Kingdom “learning disability” and “learning difficulties” are used, and the International Classification of Diseases, 11th revision (ICD-11), adopted “intellectual developmental disorders”, from which derives the name “intellectual developmental disorder” used in the title of this article [19]. The traditional definition required an intelligence quotient (IQ) below 70 with onset in childhood and functional deficits; the fifth edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-5, 2013) stopped grading ID by IQ and began defining severity by adaptive functioning in the conceptual, social and practical domains, which determines the level of support needed, and the American Association on Intellectual and Developmental Disabilities (AAIDD) requires significant limitations of intellectual functioning and adaptive behaviour originating before 18 years of age [19]. The general approach to ID and global developmental delay, including epidemiology, the investigation algorithm and empirical recurrence risk, is not developed here; only what is needed to frame the five genes is summarised.

The introduction of trio exome sequencing (proband and parents) changed the understanding of sporadic severe ID in children of non-consanguineous parents: in series of severe non-specific ID, autosomal or X-linked de novo variants account for up to 60% of cases, and the British Deciphering Developmental Disorders (DDD) study identified 94 genes enriched for damaging de novo variants at genome-wide significance (P <7×10⁻⁷) [19]. The genes in this article belong to that generation of genes discovered through enrichment of de novo variants in large cohorts, data sharing between centres (GeneMatcher, DECIPHER) and functional confirmation in cells, mouse, Xenopus or Drosophila. In all of them the clinical picture is of low specificity, severity is variable and diagnosis depends on molecular testing.

1.2. Current classifications

Three systems classify these entities in non-matching ways. OMIM assigns each gene a phenotype number within the autosomal dominant ID series (MRD44 and MRD63 for TRIO, MRD50 for NAA15, MRD52 for ASH1L, MRD54 for CAMK2B) or a descriptive title (#619575 for TAOK1), and adds modifiers when the phenotype justifies it, such as “with microcephaly”, “with macrocephaly” or “with behavioral abnormalities” [7, 20]. In the current entries, the main title is “intellectual developmental disorder, autosomal dominant” and the old name “mental retardation, autosomal dominant” remains as an alternative title; MRD44 has the modifier “with microcephaly”, MRD63 “with macrocephaly” and MRD50 “with behavioral abnormalities”, whereas MRD52 and MRD54 have no modifier; the TAOK1 entity was left outside the MRD numbering, with the acronym DDIB and a title that allows developmental delay without ID; and all six entries indicate autosomal dominant inheritance with a known molecular basis (mapping key 3) [1–6]. Orphanet classifies rare ID by the presence or absence of other manifestations: “autosomal dominant non-syndromic ID” (ORPHA:178469), an aetiological subtype of “rare non-syndromic ID” (ORPHA:101685), groups 42 genes, including ASH1L, CAMK2B, TAOK1 and the paralogue CAMK2A; NAA15 is in “non-specific syndromic ID” (ORPHA:528084), a group of more than 100 genes in which ID is associated with other anomalies without amounting to a recognisable syndrome; and the TRIO loss-of-function form has its own entity among the malformation syndromes (ORPHA:476126), whereas the gain-of-function form has no associated entity [21–23]. The consistency of these boundaries is limited: entities classified as non-syndromic, such as those of TAOK1 and CAMK2B, have macrocephaly, dysmorphism, visual impairment or musculoskeletal anomalies, and ClinGen, the North American gene–disease curation consortium, classified the relationship between ASH1L and a “complex syndromic neurodevelopmental disorder” as definitive [14]. In clinical practice, the name most used in the recent literature is the gene–phenotype dyad, for example “NAA15-related neurodevelopmental disorder” or “TRIO-NDD” [8, 24].

1.3. Historical milestones by gene

TRIO. The gene was associated with ID in 2016 by two independent groups. Ba and colleagues, in Nijmegen, found a de novo intragenic deletion in a boy with ID and, by targeted sequencing in more than 2300 individuals with ID, three additional truncating variants, all in index cases with mild to borderline ID and behavioural abnormalities [25]. Pengelly and colleagues, in Wessex (United Kingdom), described a family with a frameshift variant transmitted from father to daughter and three children from the DDD study with de novo missense variants, and were the first to link missense variants in the guanine nucleotide exchange factor (GEF) domain to ID with microcephaly [26]. That family had already been published in 2008 by Mercer and colleagues, before a molecular diagnosis existed, as possible Stoll syndrome, owing to the combination of ventricular extrasystoles, short stature, craniofacial dysmorphism and digital hypoplasia, which makes the TRIO loss-of-function disorder one of the entities in this class with a clinical description predating identification of the gene [1, 26]. In 2020, Barbosa and colleagues brought together 24 individuals from 22 families and showed that variants in the seventh spectrin repeat hyperactivate the RAC1 GTPase and are associated with macrocephaly and more severe ID, whereas those in the first GEF domain hypoactivate it and are associated with microcephaly; this observation founded the division between gain and loss of function that OMIM enshrined in MRD63 and MRD44 [11, 27]. The GeneReviews entry was published in 2017 and revised in 2023 [8]; in April 2026, Mehta and colleagues added congenital ventriculomegaly and hydrocephalus to the allelic spectrum [9].

NAA15. The gene emerged among the risk genes for ASD and developmental delay in a 2017 targeted sequencing study, which sequenced 208 candidate genes using molecular inversion probes in more than 11,730 individuals with ASD, ID or GDD from 15 centres and identified 13 carriers of NAA15 variants [3, 28]; before that, the gene had already been noted with two loss-of-function variants in 362 cases of severe congenital heart disease [12]. The clinical delineation is owed to Cheng and colleagues (2018), in a “genotype-first” study of 38 individuals from 33 families and 25 truncating variants, which included Stessman's patients and some individuals identified earlier in large congenital heart disease and congenital diaphragmatic hernia cohorts [3, 12]. In 2019, the same group described 7 individuals with NAA15 missense variants, with biochemical effects in different directions on NatA activity and stability [3, 29]. G. J. Lyon's group later extended the phenotype through parent interviews and proposed the name “NAA15-related neurodevelopmental syndrome” [24], and in 2025 published the first natural history study up to adolescence [30]. In 2021, Ritter and colleagues described paediatric HCM [13].

ASH1L. The first de novo variants were identified in 2012 in a series of severe ID [14], that of de Ligt and colleagues, who studied 100 patients with severe ID by exome sequencing; there followed a boy from a Chinese cohort of more than 1045 ASD trios (Wang and colleagues, 2016) and five patients from large ASD and developmental delay studies (Krumm and colleagues, 2015; Stessman and colleagues, 2017), in whom the excess of truncating and missense variants reached statistical significance [4, 28]; Okamoto and colleagues described in 2017 a syndrome of multiple congenital anomalies and ID [31], and Shen and colleagues (2019) characterised an “emerging disorder” due to loss of function [32]. In 2025, Di Muro and colleagues showed that the smallest region of overlap of the 1q22 microdeletions is reduced to ASH1L, which unites MRD52 and the 1q22 microdeletion into a single disorder [33].

CAMK2B. Küry and colleagues (2017) brought together 24 unrelated individuals with 19 de novo variants in CAMK2A or CAMK2B, from six European countries and the United States, and showed that variants that increase and those that decrease kinase autophosphorylation both disrupt neuronal migration [16]. Akita and colleagues (2018) confirmed the association in a Japanese cohort and described progressive cerebellar atrophy [15].

TAOK1. Dulovic-Mahlow and colleagues (2019) compared the frequency of de novo variants in 2030 trios with neurodevelopmental disorders and 2755 trios without these disorders and brought together 8 individuals [34]; van Woerden and colleagues (2021) extended the series to 23 individuals, including four 17q11.2 microdeletions proximal to the NF1 region [35]; Hunter and colleagues (2022) described the first sibship with a variant inherited from a mildly affected mother [36]; and Elkhateeb and colleagues (2025) brought together 50 individuals and, in the same paper, delineated the disorder of the paralogue TAOK2 [18].

1.4. Definition of the entities in scope

The TRIO-related disorder is defined by a heterozygous pathogenic or likely pathogenic TRIO variant in an individual with developmental delay or ID; it is divided into a loss-of-function form (MRD44: missense variants in the first GEF domain, truncating variants along the gene and intragenic deletions) and a gain-of-function form (MRD63: missense variants in the seventh spectrin repeat) [8]. TRIO-associated congenital ventriculomegaly and hydrocephalus constitute a recent allelic phenotype, still without its own OMIM number [9]. The NAA15-related disorder results mainly from haploinsufficiency due to truncating variants and, in a fraction, from missense variants [12, 30]. The ASH1L-related disorder includes intragenic loss-of-function variants and 1q22 deletions encompassing the gene [33]. The CAMK2B-related disorder results from variants that alter kinase regulation in opposite directions or from loss of function [16, 37]. The TAOK1-related disorder results from haploinsufficiency and from missense variants in the kinase domain, some with a dominant-negative effect [18, 35]. The other genes in Orphanet group 178469 are outside the scope, except as differentials.

2. Molecular Genetics and Pathophysiology

2.1. Principles common to the class

The five genes encode proteins of different pathways — Rho GTPase signalling and the actin cytoskeleton (TRIO), co-translational acetylation of nascent proteins (NAA15), histone methylation and maintenance of transcriptional programmes (ASH1L), calcium-dependent synaptic plasticity (CAMK2B) and MAP kinase signalling and microtubule stability (TAOK1) — which converge on cortical development, neuronal migration, dendritic arborisation and synapse formation. The inheritance pattern is autosomal dominant and most cases result from de novo variants, but three mechanisms coexist and have different consequences for variant interpretation. The first is haploinsufficiency, in which loss of one allele is enough to cause disease: truncating variants subject to nonsense-mediated messenger RNA decay (NMD) and whole-gene deletions are pathogenic in themselves, and ClinGen assigns a haploinsufficiency score of 3 (sufficient evidence) to TRIO, NAA15, ASH1L and TAOK1 [10]. The second is gain of function, in which a missense variant increases protein activity, as in the TRIO spectrin variants and in several CAMK2B missense variants [11, 16]. The third is the dominant-negative effect, described in missense variants in the TAOK1 kinase domain [35]. For CAMK2B, ClinGen assigns haploinsufficiency 1 (“little evidence”), so a deletion that includes the gene should not be reported as a cause of MRD54 on the basis of the gene–disease association alone; this caution applies to the large 7p deletions described in Pallister-Hall and Greig syndromes [10].

2.2. TRIO

TRIO lies at 5p15.2, has 57 exons spread over 366.5 kb and generates six isoforms by alternative splicing, five of which contain the first GEF domain, with nervous system-specific isoforms [26]. The protein is a Dbl-family guanine nucleotide exchange factor with an N-terminal CRAL-Trio/Sec14 domain, several spectrin-like repeats, two Dbl-homology–Pleckstrin-homology (DH-PH) modules, two SH3 domains, an immunoglobulin-like domain and a C-terminal serine/threonine kinase domain [8, 26]. The first DH-PH module (GEFD1) activates RAC1 and RhoG; the second (GEFD2) activates RhoA. Through RAC1 and actin remodelling, TRIO regulates cytokinesis, cell migration, axon guidance, neurite outgrowth, lamellipodia formation and endocytosis of AMPA receptors at excitatory synapses [8, 10]. Complete Trio knockout in the mouse is lethal in the embryonic period, with a small brain and disorganised hippocampus, and deletion restricted to the hippocampus and cortex impairs learning [26]. In the rat hippocampus, Trio expression is concentrated in the early postnatal period and falls after day 11; reducing expression by RNA interference in cultured neurons increases the number of primary dendrites and branches, and in hippocampal slices increases AMPA receptor-mediated transmission (without changing NMDA receptor-mediated transmission) through reduced endocytosis of these receptors, indicating that TRIO brakes the early maturation of excitatory synapses, a phenomenon also described in ASD models [1, 25].

The pathophysiology is bidirectional. In HEK293T cells, phosphorylation of PAK1 (a marker of RAC1 activity) decreases with all GEFD1 missense variants studied, except p.Pro1461Thr (no effect; protein numbering of NP_009049.2, MANE transcript NM_007118.4, valid for all TRIO variants in this paragraph) and p.Pro1461Leu (slight increase), and with the truncation p.Gln1489Ter; it increases markedly with all variants in the seventh spectrin repeat (p.Thr1075Ile, p.Arg1078Trp, p.Arg1078Gly, p.Arg1078Gln and p.Asn1080Ile), which also increase neurite outgrowth and lamellipodia in neuroblastoma cells; and a p.Arg1078Trp double mutant with an inactive GEFD1 no longer activates RAC1, which shows that the hyperactivation depends on GEFD1 [11]. The level of RAC1 activation correlates with head circumference (HC), in parallel with variants in RAC1, PAK1 and PAK3, whose opposing effects on HC are known [11]. The assays diverge for p.Asn1080Ile, which did not alter RAC1 in a pull-down assay and hyperactivated it in the phospho-PAK1 assay [11, 26]. The divergence extends to classification: the OMIM MRD44 entry describes the p.Asn1080Ile carrier from the Pengelly series (without microcephaly and with seizures) among the MRD44 patients, whereas the MRD63 entry includes the same variant among the gain-of-function variants in the seventh spectrin repeat [1, 2]. In practice, location in spectrin 7 and the phospho-PAK1 assay bring this variant closer to the MRD63 group [11]. In Xenopus tropicalis, truncation in GEFD1 reproduces microcephaly and forebrain deformation, and truncation in GEFD2 does not [11]. The interaction with MgcRacGAP in cytokinesis suggests a defect in the progenitor cell cycle as the basis of the microcephaly [26]. Later studies added three findings: a cluster of ASD-associated de novo variants in GEF1 generates hypo- or hyperfunctional forms of the protein, with fewer synaptic AMPA receptors or more glutamatergic synaptogenesis [38]; variants in the β3-β4 loop of the PH2 domain release GEF2 autoinhibition and activate RhoA [39]; and mice heterozygous for three distinct human variants have different behavioural, synaptic and brain size phenotypes, including a compensatory increase in RAC1 activity in a hypofunctional GEF1 variant [40].

For laboratory purposes, gain-of-function missense variants have been described between p.Leu1071 and p.Glu1159 and GEFD1 variants between p.Arg1292 and p.Gln1591 (NM_007118.4, NP_009049.2) [8]. The rule “GEFD1 → microcephaly” has exceptions: p.Pro1461Leu, in two individuals, was not associated with microcephaly [11], and a Japanese boy with the de novo missense variant NM_007118.4:c.4104T>A p.(Asp1368Glu) in GEF1 had macrocephaly (+2.5 standard deviations, SD), severe ID, growth hormone deficiency, a thin corpus callosum and pituitary hypoplasia [41]. A missense variant in the second SH3 domain, c.7738A>T p.(Ile2580Phe), was associated with macrocephaly in an infant, with the first detailed HC growth curve in this disorder [27]. Very proximal truncating variants, such as p.Gln768Ter, probably undergo NMD; those in the last exon may escape it [8, 11].

2.3. NAA15

NAA15 (4q31.1) encodes the obligatory auxiliary subunit of the NatA complex, which anchors the catalytic subunit NAA10 to the ribosome and determines its substrate specificity; the complex exists as a trimer or tetramer with HYPK and NAA50 [13]. NatA acetylates the N-terminal α-amino group of about 40–50% of the human proteome, one of the most frequent co-translational modifications in eukaryotes; with NAA15, NAA10 localises to the ribosome and acetylates N-termini with serine, alanine, threonine, valine or glycine, and without NAA15 it acetylates acidic residues in the cytoplasm [30]. The main mechanism is haploinsufficiency: the 25 variants in the founding series were truncating (13 nonsense, 2 canonical splicing and 10 frameshift) in 12 of the 20 exons; in lymphoblastoid lines and induced pluripotent stem cells (iPSC) the mutant transcript was almost absent, owing to NMD; two truncating variants did not rescue the temperature-sensitive growth phenotype of NatA-deficient yeast; and deletions that include NAA15 and neighbouring genes give a similar and milder picture [12]. NAA15 missense variants have diverse biochemical effects on NatA, such as destabilisation of the complex (p.Lys450Glu; numbering of NP_476516.1, MANE transcript NM_057175.5, valid for the others), reduced thermostability (p.Lys338Asn), location in ribosome-binding regions (p.Arg27Gly, p.Ala585Thr) or in the C-terminus that stabilises the heterodimer and binds HYPK (p.Ala719Thr) [24]. In the mouse, loss of Naa15 increases the number of neurons and causes repetitive and anxious behaviours [42]. Duplication of NAA15 has been associated with short stature with intact endocrine function, and in zebrafish the orthologues negatively regulate myogenesis [30]. The mechanism is shared with NAA10.

2.4. ASH1L

ASH1L (1q22) encodes a trithorax-group histone lysine methyltransferase that trimethylates lysine 36 of histone H3 (H3K36me3), a mark of actively transcribed gene bodies, associated with transcribed regions including the HOX genes; it has AT-hook, SET, PHD finger and bromodomain domains [10, 31]. The deficiency impairs maintenance of transcriptional programmes during neurodevelopment [10]. Unlike other chromatin modifiers, there is no established DNA methylation episignature for ASH1L, so this test cannot be used to classify variants [10]. With a pLI of 1 in gnomAD, haploinsufficiency is the accepted mechanism; some series suggested that missense variants are associated with a more severe phenotype and a higher probability of ASD [14, 43], but a boy with the de novo missense variant NM_018489.3:c.4043A>G p.(Lys1348Arg), classified as of uncertain significance, had a mild picture of ASD and ADHD [44]. Animal models show autistic-like behaviour with haploinsufficiency or conditional brain deletion of Ash1l, epigenetic aberrations and seizures with deficiency in the prefrontal cortex [10]. A 2025 preprint, not peer reviewed, describes seizures, microcephaly and less complex dendrites in Ash1l mice, hyperexcitability of hippocampal CA1 neurons in females and sex-specific transcriptomic signatures, which the authors relate to a predominance of epilepsy in females and of ASD in males among patients [45].

2.5. CAMK2B

Ca²⁺/calmodulin-dependent kinase II (CaMKII) is a serine/threonine kinase that accounts for 2–6% of postsynaptic density protein; it forms homo- or heteromeric dodecamers of the α (CAMK2A), β (CAMK2B), γ (CAMK2G) and δ (CAMK2D) isoforms, with a predominance of α and β in the brain (about 9 α to 3 β in the forebrain, the reverse ratio in the cerebellum) [15]. At rest, the regulatory segment binds to the groove of the kinase domain; the Ca²⁺/calmodulin complex releases it, and autophosphorylation of threonine 287 (Thr286 in the α isoform) by neighbouring subunits maintains autonomous, calcium-independent activity, essential for long-term potentiation and learning [15]. CaMKIIβ also has a structural function independent of kinase activity, bundling actin filaments to form dendritic spines [10].

Pathogenic variants disrupt regulation, not the amount of kinase. In the Küry series (protein numbering identical in NP_742076.1, from the paper's transcript NM_172079.2, and in NP_001211.3, from MANE NM_001220.5), p.Glu110Lys, p.Pro139Leu and p.Glu237Lys increased Thr287 phosphorylation and severely impaired neuronal migration in utero (gain of function), whereas p.Lys301Glu, in the Ca²⁺/calmodulin-binding domain, almost abolished phosphorylation (loss of function); increased and decreased activity disrupt migration equally, indicating a dosage-sensitive relationship [16]. Because overexpression and reduction of normal CaMKIIβ also disrupt migration, and mutant subunits are incorporated into dodecamers with normal subunits, the OMIM synopsis interprets the missense variants in the series as probably dominant-negative in action; p.Glu110Lys and p.Pro139Leu also reduced the amount of mutant protein in cells, unlike the other missense variants [5, 16]. The data on p.Pro213Leu diverge: in Neuro-2a cells, Akita and colleagues observed increased Thr287 autophosphorylation [15]; in 2026, knock-in mice with this variant showed reduced cortical CaMKIIβ expression and Thr287 phosphorylation, aberrant low gamma oscillations (20–50 Hz) at rest and greater susceptibility to pentylenetetrazol-induced seizures, a phenotype reproduced by the knockout, which classifies it in vivo as loss of function [46]. The truncating variant p.(Arg29Ter), described de novo by Küry and inherited from a mother without manifestations by Heiman, suggests that loss of function of one allele can also cause the disease, a question Küry had left open [16, 37]. Akita proposed, on the basis of few patients, that CaMKIIβ dysfunction is associated with cerebellar symptoms (episodic ataxia, cerebellar atrophy) and CaMKIIα dysfunction with seizures, a distinction that the Küry series did not confirm as regards seizure frequency [15].

2.6. TAOK1

TAOK1 (17q11.2) belongs to the TAO kinase family, with TAOK2 (16p11.2) and TAOK3; all three have a conserved N-terminal kinase domain, a C-terminal regulatory domain and coiled-coil motifs [18]. TAOK1 activates the p38 pathway by phosphorylating MAP2K3 and MAP2K6, participates in the G2/M checkpoint after DNA damage, regulates microtubule stability through MARK2 and is required for neuronal maturation and cortical development [10]. The protein is organised into a kinase domain (amino acids 34–295), a substrate-binding domain (296–431), a spacer and a tail with coiled-coil [35]. In the mouse, both Taok1 knockdown by in utero electroporation and overexpression of the normal protein impair neuronal migration (70% versus 95% of cells in the cortical plate) and reduce neurite arborisation, indicating that the expression level must be regulated [35]. Truncating variants undergo NMD in fibroblasts (reduced RNA recoverable with cycloheximide), and knockdown of Tao1 in Drosophila causes developmental delay, neuromuscular junction abnormalities and smaller mitochondria in motor axons [34]. Among the missense variants (numbering of NP_065842.1, MANE transcript NM_020791.4), p.Arg150Ile and p.Leu548Pro behave as loss of function, p.Leu167Arg and probably p.Leu315Phe as dominant-negative, and p.Met231Val had no demonstrable effect [35]; p.(Ser111Phe), p.(Leu167Arg), p.(Ala219Val) and p.(Arg269Gln) reduce catalytic activity in vitro [18]. The variant p.(Arg865Ter), in exon 20, escapes NMD, and no pathogenic variants have been described in the C-terminal tail (residues 901–1001), which is dispensable [18]. The Taok1+/− mouse has macrocephaly, autistic-like behaviours and changes in the dorsal raphe nucleus, reversed by reintroduction of the normal protein but not of the kinase-dead protein [18, 47].

2.7. Variant spectrum and normalisation

The variants in the table were normalised in VariantValidator to the MANE Select transcript and to the GRCh38 and GRCh37 genomes; the historical nomenclature of the papers is indicated where it differs [48]. Two CAMK2B variants from the Küry series (p.Glu237Lys and p.Lys301Glu) are kept on the paper's transcript (NM_172079.2), because projection to MANE was not obtained; the protein numbering of isoforms NP_742076.1 and NP_001211.3 coincides at the positions checked.

The mutational spectrum differs between genes. In TRIO, sequencing detects about 95% of known pathogenic variants and deletion and duplication analysis about 5% [8]; the splicing variant NM_007118.4:c.4860-2A>G activates an acceptor 51 bp upstream of exon 33 and generates p.(Arg1620SerfsTer10) [11]. In NAA15, besides the three recurrent variants in the table, NM_057175.5:c.228_232del p.(Asp76GlufsTer20), published as c.228_232delCTTGA, occurred in two families [12], and in a Chinese series the intronic variant NM_057175.5:c.692-5A>G reduced messenger RNA [20]. In ASH1L there are truncating variants along the gene, intragenic deletions (exons 2–12 in a sibship) and 1q22 microdeletions, and partial duplications may be misinterpreted: optical genome mapping (OGM) showed, in a case that array and exome had described as a single event, two tandem duplications, one of them intragenic, and RNA sequencing showed no reduction of ASH1L messenger RNA [33, 50]. In the largest TAOK1 cohort, the 37 variants were 17 missense, 9 nonsense, 5 frameshift, 4 splicing or intronic and 4 multiexon deletions; 11 of the 17 missense variants were in the kinase domain, and the variant NM_020791.4:c.1909-7G>A included 5 bp in the RNA, with frameshift and NMD [18]. No founder variants have been described for any of the five genes in the sources consulted.

2.8. Genotype-phenotype correlations, penetrance and allelic disorders

In TRIO, the correlation is the most structured in the class: spectrin missense variants are associated with more severe ID, macrocephaly and a higher risk of seizures and scoliosis, and those at residue Arg1078 with the most severe phenotype; GEFD1 missense variants and truncating variants are associated with less severe ID and microcephaly [8]. Truncating variants have more variable and often milder phenotypes than GEFD1 missense variants, which is consistent with GEFD1 missense variants being de novo in 18 of 18 cases with known inheritance and truncating variants in 60% (10/17); seizures occur in 30–35% of carriers of truncating variants and in about 10% of carriers of GEFD1 missense variants, and 2-3 toe syndactyly, broad proximal interphalangeal (PIP) joints and 5th finger clinodactyly have only been described in the loss-of-function forms [8]. Penetrance appears complete: in 7 individuals from 5 families with an inherited variant, the parent was similarly or mildly affected [8].

In NAA15 no difference in adaptive functioning between truncating and missense variants has been shown [30]. From the incidence of developmental delay in the general population (5.12%) and the excess of likely disruptive variants in cases, Stessman and colleagues estimated the penetrance of these NAA15 variants for developmental delay at 35.3%; this is a statistical calculation on selected cohorts, not an observation in families, but it is consistent with transmission by parents with mild ID or without ID and with the three mothers with mild ID in the 2017 series [3, 12, 28]. For missense variants, biochemical effects in opposite directions (increase or decrease in enzyme activity, stabilisation or destabilisation of the NatA complex) have been proposed as an explanation for the variability, and two missense variants inherited from unaffected or mildly affected parents were present in gnomAD, which weakens their classification as pathogenic [3, 29]. In CAMK2B, ID is more severe when the variant affects the autoregulatory domain (severe or mild to moderate in 6 of 6, versus 10 of 17 in the kinase domain), and carriers of CAMK2B variants have more hypotonia (90% versus 50%) and more digestive problems than those with CAMK2A variants [16]; transmission of p.(Arg29Ter) by a mother without manifestations demonstrates incomplete penetrance [37]. In TAOK1, carriers of missense variants, compared with those with loss-of-function variants, had no seizures or neonatal growth delay, had more genital anomalies (33% versus 6%) and more overweight (43% versus 17%) and less macrocephaly (57% versus 89%) and hypotonia (40% versus 67%); inheritance from parents with mild ID or learning difficulties indicates reduced penetrance and variable expressivity [18, 36]. In ASH1L, the denominators are too small for reliable correlations [14].

As regards allelic disorders, the 2023 GeneReviews stated that there was no other phenotype associated with germline TRIO variants [8]; Mehta and colleagues found, in a cohort of 2697 trios with congenital ventriculomegaly or hydrocephalus studied by exome sequencing, five probands with de novo TRIO variants in addition to the index case, with significant enrichment (adjusted p 6.12×10⁻⁵), including p.(Arg1078Trp) and two substitutions (numbering of NP_009049.2, MANE NM_007118.4) of residue Glu1299 (p.Glu1299Lys and p.Glu1299Gly) [9]. Because p.Arg1078Trp is a spectrin gain-of-function variant and p.Glu1299Lys a GEFD1 variant associated with microcephaly in the Barbosa series [11], ventriculomegaly is not limited to one of the two mechanisms. For NAA15, paediatric HCM, dystonia and adult-onset parkinsonism have broadened the spectrum [13, 51–53]. For CAMK2B, the association with tetralogy of Fallot rests on one case with p.(Lys299Glu) (NP_001211.3, MANE NM_001220.5) and on a de novo variant described in a tetralogy cohort [54]. No source describes established genetic modifiers; dual diagnoses are frequent (14% in the TAOK1 cohort) and should be sought in exome analysis [18].

3. Clinical Spectrum and Manifestations by System

3.1. Common syndromic description

The picture common to the five disorders is global developmental delay recognised in the first or second year of life, with disproportionate delay in expressive language, ID of variable severity (from borderline functioning to severe ID), behavioural abnormalities (attention deficit and hyperactivity, autistic features or ASD, obsessive behaviours, aggression, stereotypies), hypotonia, feeding difficulties in infancy and, in variable proportions, epileptic seizures, mild facial dysmorphism without a diagnostic gestalt and anomalies of other organs. These comorbidities are those of ID in general: in a systematic review cited in the Oxford Desk Reference, 22% of people with ID had epilepsy, 20% cerebral palsy, 17% anxiety disorder, 12% oppositional defiant disorder and 10% autistic disorder [19]. What distinguishes each gene is the combination of features, above all HC (microcephaly in the TRIO loss-of-function forms; macrocephaly in the TRIO gain-of-function forms and in TAOK1), cardiac findings (NAA15), visual impairment, ataxia and cerebellar atrophy (CAMK2B), and neonatal hypotonia with joint hypermobility (TAOK1). The available data come from series gathered by ascertainment (patients with an exome requested for developmental delay, case sharing between centres, family registries), with denominators that vary from row to row, assessed at different ages and with almost no follow-up of adults; there are no studies of untreated natural history in the strict sense, because all patients described received early intervention and educational support.

3.2. TRIO: loss-of-function and gain-of-function forms

GeneReviews brought together 56 individuals with sufficient clinical information, published between 2016 and 2021 or communicated by the authors, 20 with gain-of-function variants and 36 with loss-of-function variants; the denominators in the table correspond to the individuals assessed for each feature [8].

Source: GeneReviews TRIO-NDD, table 2 [8].

Development and cognition. In the gain-of-function forms, only a third (6/17) sat unsupported at 14 months and 75% (13/17) at 36 months; one child sat only after 6 years of age [55]; seven children walked, with onset between 20 months and 7 years, and ten were not walking at the last assessment, the oldest aged 6 to 10 years; seven of 16 were verbal, with first words between 12–15 months and 5.5 years [8]. ID is probably universal, but no formal IQ assessments have been published in these forms; in the Barbosa series, ID was moderate in 22% and severe in 78% of the spectrin group, and motor milestones were later than in the GEFD1 group (walking, p = 0.035) [8, 11]. In the loss-of-function forms, 80% of those who sat did so by 14 months, 20 of 24 walked (80% by 22 months) and 80% of the 20 verbal individuals said their first words by 48 months; four were non-verbal, one of them aged 16 years with a multiexon in-frame intragenic deletion encompassing GEFD1, with refractory seizures [8, 56]. Only four individuals had a formal IQ: two with borderline functioning (IQ 81 and 78) and two with mild ID (IQ 62 and 68) [8].

Behaviour. In the gain-of-function forms, attention deficit occurred in 56% (9/16), stereotypies and obsessive-compulsive behaviours in 47%, aggression in 29% and ASD or autistic features in 27% (4/15) [8]. In the loss-of-function forms, 94% had behavioural abnormalities: attention deficit in 79% (22/28), obsessive-compulsive behaviour in 45%, aggression in 44% (two with self-injury), ASD or autistic features in 40% and stereotypies in about 25% [8]. Insensitivity or increased tolerance to pain was described in 3 of 18 individuals in the Barbosa series [11]. Sleep disturbances are probably under-reported [8].

Growth, feeding and the digestive system. Early feeding difficulties (weak suck, swallowing, vomiting, gastro-oesophageal reflux) are frequent in both forms; about 30% of carriers of gain-of-function variants and 22% of those with loss-of-function variants needed nutritional support, usually by tube or gastrostomy [8]. At the last assessment, weight was ≤ −2 SD in 67% (12/18) of the gain-of-function forms and in 30% (8/27) of the loss-of-function forms, and short stature (≤ −2 SD) affected 31% and 14%, respectively; constipation was more frequent with GEFD1 missense variants (63%) than with truncating variants (36%) [8].

Head and central nervous system. Absolute macrocephaly affects 61% of the gain-of-function forms, with HC ≥ +3 SD in four individuals and a maximum of +4.7 SD; in the Barbosa series, mean HC was +2.6 SD in the spectrin group and −3.82 SD in the GEFD1 group, where microcephaly was universal [8, 11]. In the loss-of-function forms, microcephaly is severe (< −3 SD) in at least 15 individuals, three with mean HC of −6 SD, and two had, conversely, macrocephaly [8, 57]. Seizures in the gain-of-function forms include nocturnal seizures, absences and myoclonus of the extremities and eyelids, with one onset described at 30 years of age; in the loss-of-function forms they include febrile seizures, myoclonic seizures and refractory absences with eyelid myoclonia [8]. Brain magnetic resonance imaging (MRI) showed, in five carriers of gain-of-function variants with abnormal MRI, dilated ventricles (4), enlarged cerebrospinal fluid (CSF) spaces (2), hydrocephalus (2) and corpus callosum abnormalities (3), and in six carriers of loss-of-function variants partial agenesis or hypoplasia of the corpus callosum (3), ventricular abnormalities (2), progressive leukoencephalopathy, lissencephaly and an interhemispheric cyst (one each) [8]. Tremor, dystonia and ataxia were described in two individuals each in the gain-of-function forms [8].

Face, hands, teeth and skeleton. Quantitative facial phenotyping showed significant clustering only in the gain-of-function forms (high or prominent forehead, frontal bossing, arched eyebrows, low-set ears, wide mouth), suggesting a recognisable gestalt; in the loss-of-function forms there was no clustering, although upslanting palpebral fissures, a tubular or bulbous nose and micro- or retrognathia are frequent [8, 11]. One individual with p.Asn1080Ile had an Angelman-like facial appearance [26]. Dental anomalies affect about 60% of the loss-of-function forms: delayed eruption in 30%, crowding in 30%, hypodontia or oligodontia, macrodontia of the incisors [8]. Scoliosis affects 44% of the gain-of-function forms, sometimes early-onset and requiring surgery, and 15% of the loss-of-function forms [8].

Heart and other organs. Structural cardiac anomalies are infrequent: bicuspid aortic valve with aortic regurgitation and atrioventricular septal defect with interrupted aortic arch in the gain-of-function forms; atrial septal defect (2) and tetralogy of Fallot (1) in the loss-of-function forms; there was also arrhythmia in two carriers of truncating variants, in one of them attributable to a coexisting pathogenic variant in KCNJ2 [8, 26]. Recurrent infections (3), neurogenic bladder, enuresis and urinary incontinence, optic atrophy, strabismus, cleft lip and aplasia cutis have also been described [8, 57].

Families and founding series. The family described by Mercer in 2008 and reassessed by Pengelly in 2016 shows the range of expressivity within a single pedigree: two brothers with mild learning difficulties, special education and microcephaly (−3 and −5 SD), the mother with normal intelligence but school difficulties, an affected maternal cousin and six more probably affected relatives over four generations; the straight, pointed nose, micrognathia, mild brachydactyly with tapering fingers and hypoplastic distal phalanges, prominent interphalangeal joints and hypodontia (3 of 4) recurred, and one of the brothers and the mother had multiple ventricular extrasystoles without syncope [1, 26]. The daughter of one of the brothers, index case of the reassessment, had mild GDD with speech delay, microcephaly of −5 SD, right radial aplasia with a rudimentary thumb and absent first metacarpal, and cardiac septal defects without clinical consequence [1]. In the Ba series (boys aged 7 and 10 years, a 35-year-old woman without family history and a 20-year-old man with an affected father), HC was between −1 and −2.5 SD, ID was borderline to mild, and three had recurrent infections, the feature that gave its name to the Orphanet entity; hyperacusis, swallowing difficulties, facial asymmetry, kyphosis and pectus excavatum were also described, and the 7-year-old boy, with an intragenic deletion, had been identified in a copy number variant screen of 5531 patients [1, 23, 25]. In the nine MRD63 patients in the Barbosa series (aged 3 to 19 years and one aged 40 years), walking usually began between 3 and 5 years, three of the youngest were not walking, ID was severe in seven and moderate in two, all but two had relative macrocephaly (+2 to +4.7 SD), and six had a substitution of residue Arg1078; the variant was de novo in eight, with unknown origin in the ninth [2, 11].

Allelic phenotype of ventriculomegaly and hydrocephalus. In Mehta's index case, a child with syndromic congenital hydrocephalus required CSF shunting and had de novo p.(Arg1078Trp); all six probands in the cohort had ventriculomegaly, often with GDD and other structural brain anomalies, and the literature review found six more individuals with de novo TRIO variants and ventriculomegaly or hydrocephalus [9]. Congenital hydrocephalus affects about 0.5–1 per 1000 live births, according to the authors [9]. A cohort of 25 unpublished individuals, reviewed with 32 published ones, added brain malformations (corpus callosum, ventriculomegaly), psychological and dental problems and a facial gestalt dependent on the domain and the functional effect [58].

3.3. NAA15

Cheng's founding series (38 individuals from 33 families, of various origins, studied by exome, genome or targeted sequencing) is the main source of frequencies, with different denominators for each feature [12].

Source: Cheng and colleagues, 2018, table 1 [12].

Development and cognition. ID ranges from mild to severe; some individuals do not speak and use sign language [12]. In a boy with HCM, the full-scale IQ was 94, with specific difficulties in attention, working memory and fine motor skills, which shows that not all carriers have ID [13]. In the Lyon cohort (66 individuals with NAA15, including 19 interviewed directly, recruited through family self-referral), ID was in general much less severe than in the NAA10 disorder, and overall adaptive functioning measured by the Vineland-3 was significantly higher [24]. In Makwana's longitudinal study (27 individuals aged 1 to 24.5 years, 18 male, recruited in the United States by the same group), the Vineland-3 adaptive behaviour composite (ABC), with a population mean of 100 and SD of 15, averaged 68.5 (SD 22.1); girls performed better in motor and fine motor skills (13.3 ± 3.5 versus 7.8 ± 3.2; p = 0.001) [30]. The relationship between age and functioning is analysed in section 11.

Behaviour and neurology. Self-injurious or aggressive behaviours, impulsivity, short attention span and autistic features are frequent [24]. Hypotonia was described in 14% in the founding series [12]. Childhood-onset dystonia has been described in a few individuals, including two cases of isolated gait dystonia published in 2026, and a man with ID and speech delay developed dystonia-parkinsonism in adulthood [30, 51–53]. An adolescent in the longitudinal cohort developed focal seizures and catatonia [30]. Brain MRI re-read blindly by two neuroradiologists was normal in 6 of 8 probands with NAA15 (versus 3 of 18 with NAA10), with a mean of 2.8 anatomical anomalies (SD 2.3) versus 5.7 (SD 3.0; p = 0.02) and no changes between scans in three patients with more than one MRI [59].

Heart. The four individuals with heart disease in the founding series had multifocal ectopic atrial tachycardia with hypertension, a repaired ventricular septal defect (VSD), a marfanoid habitus with an aortic root at the upper limit of normal, and heterotaxy syndrome with dextrocardia, tricuspid atresia, double outlet right ventricle and transposition of the great arteries [12]. The authors acknowledge that, with a prevalence of 21%, severe heart disease might depend on a second locus, and in a cohort of 2871 probands with congenital heart disease they found no new NAA15 variants [12]. Ritter and colleagues described two boys with HCM: one presenting at 2 months (tachypnoea, tachycardia, asymmetric septal hypertrophy of 9 mm, Z +8, maximum gradient 52 mmHg) with de novo p.(Glu337ArgfsTer5); another with a murmur at 6 years and septal hypertrophy that progressed from 10 mm (Z +5) to 14 mm (Z +10), without obstruction, with de novo p.(Arg27Gly) [13]. ClinGen considers the evidence for the association with congenital heart disease to be emerging [10].

Eyes, growth and other features. Strabismus, amblyopia, astigmatism and cortical visual impairment occur in a subset; a 19-year-old woman with a frameshift variant had unilateral microphthalmia [24]. In a study of 19 patients with NAA15 and 67 with NAA10, the prevalence of myopia, hypermetropia, strabismus, nystagmus, astigmatism and cortical visual impairment did not differ between the two genes [60]. Some individuals have short stature and low weight; one had corneal calcifications and persistent hypercalcaemia without hyperparathyroidism, of uncertain relationship to the variant [24]. In a São Paulo series of 39 children with a clinical diagnosis or suspicion of Silver-Russell syndrome, one of the alternative causes found was a NAA15 variant, which includes growth restriction in the clinical spectrum [61]. A Chinese series of four families (7 patients aged 5 to 35 years) described ID, language delay and learning difficulties with several affected members per family [20].

Early series and missense variants. In Stessman's 13 patients, ascertained in large cohorts and with scant clinical data, there was GDD or ID in 10 of 11, speech delay in 5 of 6, formal ASD in 5 of 8 (a 15-year-old boy with Asperger syndrome) and motor delay in 2 of 4, with selective mutism or aggression in some; none had seizures and neuroimaging, when performed, was normal [3, 28]. In Cheng's 7 patients with missense variants (2019), aged 6 to 16 years, the spectrum ranged from an 11-year-old girl with severe ID, verbal IQ of 36, ASD and seizures to two children with initial delay and subsequently normal cognition, one with Asperger syndrome and the other with mild dysmorphism and skeletal anomalies; four had seizures (febrile, afebrile or absences) or an abnormal electroencephalogram (EEG), and there was no common gestalt [3, 29].

3.4. ASH1L

Up to 2024 only 11 individuals with heterozygous loss of function had been described with detailed phenotyping; Cordova and colleagues added three boys from the Brain Gene Registry (a registry of 13 US ID and developmental research centres) and aggregated the frequencies, with small denominators [14].

Source: Cordova and colleagues, 2024 [14].

The spectrum extends from average cognitive functioning with ASD and adaptive deficit to severe pictures. At the severe end, Okamoto described a multiple congenital anomaly syndrome with severe ID, growth delay, microcephaly, facial dysmorphism, delayed myelination and skeletal anomalies [31]: the Japanese boy, son of unrelated parents, had hypotonia and GDD from the first year, feeding difficulties and chronic constipation; at 4 years he could not sit, crawl or speak and had self-injurious behaviours, with microcephaly of −3.9 SD, nystagmus, strabismus and visual impairment, sleep disturbance, unilateral sensorineural deafness, congenital fusion of C2 with C3 and lumbar scoliosis [4, 31]. The first two cases also had severe pictures: the boy in the de Ligt series, with severe ID, absent speech, hypotonia, cryptorchidism, pectus carinatum and dysmorphism (deep-set eyes, synophrys, small asymmetric ears, beaked nose), and the Chinese boy from Wang's ASD cohort, with IQ below 30, ASD, hyperactivity, anxiety, obsessive behaviour, mild regression, macrocephaly, sleep disturbance and digestive complaints; in Stessman's three patients, aged 6 to 17 years, ID was mild to moderate, and two of three had speech delay, ASD and seizures [4]. These descriptions use OMIM protein notation, without a stated transcript: A724S (de Ligt), V2080I (Wang) and A2786P (Okamoto) (historical notation; the reference residue matches at positions p.Ala724Ser, p.Val2080Ile and p.Ala2786Pro of NP_060959.2, MANE transcript NM_018489.3), all de novo missense variants without functional studies [4]; Shen described multiple congenital anomalies, learning difficulties, ADHD, sleep apnoea and scoliosis [32]. At the other end, a boy from the Brain Gene Registry with p.(Ser1635CysfsTer18) (NP_060959.2, MANE NM_018489.3) had average cognition (Shipley scale 101), ASD, ADHD, obsessive behaviours, dysmetria and ataxic gait, and another, with p.(Gln1637Ter), had borderline functioning, ASD, mild bilateral sensorineural hearing loss and epilepsy with absences, focal motor and myoclonic seizures resistant to clonazepam [14]. The genital and urinary malformations described include penile torsion and cryptorchidism; the musculoskeletal ones, pectus carinatum or excavatum, scoliosis and cervical vertebral fusion; the ophthalmological ones, hypermetropia, astigmatism, strabismus and nystagmus [14]. Epilepsy may be the dominant problem: twin girls with mild ID and seizures had a nonsense variant in exon 3 [43], and a 7-year-old boy with ID and myoclonic seizures had an epileptic encephalopathy with spike-wave activation in sleep that evolved into photosensitive epilepsy [62]. There are also pictures of isolated mild ID [63] and non-syndromic ID in three siblings with an intragenic deletion [33]. Most published probands are male, without differences in presentation by sex in the clinical series [14]. Progressive neurodegeneration has not been described [10].

3.5. CAMK2B

The Küry series (24 individuals, 10 with CAMK2B variants and 14 with CAMK2A variants, from the United States, France, the United Kingdom, the Netherlands, Germany and Norway) is the source of the frequencies for the group; some rows distinguish the two genes [16].

Source: Küry and colleagues, 2017, table 2 [16].

Considering only the 10 carriers of CAMK2B variants in the Küry series (aged 14 months to 14 years, of Northern European or US ancestry), seven had severe ID, six did not walk and six did not speak; four had seizures and four an abnormal EEG; half had reduced overall growth, but only one had frank microcephaly (−3.4 SD); seven had visual anomalies and eight relevant digestive difficulties, mainly feeding, reflux and constipation; and one or two had ectodermal dysplasia, food allergies or respiratory problems [5, 16]. The seven variants were four missense (p.Pro139Leu in four patients), two splicing and one nonsense [5]. The phenotype is only mildly dysmorphic (hypotelorism, downslanting palpebral fissures, epicanthus), with variable ID and particularly affected language, and diagnosis without molecular testing is practically impossible [16]. The recurrent variant p.(Pro139Leu) accounted for 40% of published cases in 2020; the oldest known patient with this variant, at 22 years, had severe ID without language, hypotonia, microcephaly, epilepsy, motor stereotypies, optic atrophy, scoliosis and progressive cerebellar atrophy [17]. In Akita's Japanese cohort, a 7-year-old boy with p.(Pro213Leu), who walked at 20 months and had his first seizures at 23 months, had mild ID with few words [5], seizures during sleep partially controlled with clonazepam, episodic ataxia and acquired microcephaly (HC −2.3 SD at 5 years) with normal MRI at 4 years; another, aged 8 years, with p.(Arg284Ser), had profound ID without head control or words, myoclonus, ocular convergence spasm and oculogyric crises without epileptic seizures, with HC that went from +1.1 SD at birth to −2.4 SD at 2 years and progressive cerebellar atrophy from 2 years of age [15]. A 2-year-old girl with p.(Lys299Glu) in the autoregulatory segment had growth restriction, microcephaly, GDD, dysmorphism and tetralogy of Fallot [54]. In Heiman's family, the 3-year-old proband had focal seizures from 11 months, severe GDD and expressive language delay; her 8-year-old sister, with the same variant, had speech delay, visual impairment and gross motor problems, and the carrier mother had no manifestations [37]. The case of periodic neuropsychiatric episodes with p.(Glu110Lys) is described in section 10 [49].

3.6. TAOK1

Two cohorts provide frequencies: that of van Woerden, with 20 individuals added to the 8 of Dulovic-Mahlow (28), and that of Elkhateeb, with 50 individuals (31 male) from 40 families, aged 2 to 52 years, gathered through DDD, a North American commercial laboratory, DECIPHER, GeneMatcher and MyGene2 [18, 35].

Sources: van Woerden and colleagues, 2021, table 2; Elkhateeb and colleagues, 2025 [18, 35].

The difference in the frequency of macrocephaly between the two cohorts (38% and 78%) reflects different denominators and recruitment criteria; Elkhateeb's abstract gives 83%, whereas the count in the results is 32 of 41 [18]. Global delay may be very mild, and the reason for referral may be behaviour or dysmorphism [35]. Motor milestones catch up, but speech difficulties persist in the 23 individuals assessed between 3 and 45 years of age; 18 attended mainstream education with support and 9 special education; there was regression in one individual who also had a variant of uncertain significance in CDKL5 [18]. Macrocephaly was primary at birth in 4 of 11; abnormal neuroimaging included ventriculomegaly, congenital hydrocephalus, cortical atrophy, white matter and corpus callosum abnormalities, cerebellar atrophy or hypoplasia and focal cortical dysplasia [18]; one patient had multisutural craniosynostosis, possibly related to prenatal ventriculomegaly and macrocephaly [35]. Childhood-onset tremor, sometimes progressive, has been proposed as a manifestation that should prompt consideration of TAOK1 [64, 65]. The 2025 cohort described for the first time male genital anomalies (hypospadias, cryptorchidism) and transient or recurrent neonatal hypoglycaemia, as well as prematurity in 11 individuals, hearing loss and food selectivity into adolescence [18]. Automated facial analysis found no recognisable gestalt (area under the curve 0.757; p = 0.124), although a broad or high forehead, thin upper vermilion, hypertelorism, slanting palpebral fissures, low-set ears, high palate, micrognathia, bulbous nose and 5th finger clinodactyly are frequent [18, 35]. In 2022, Hunter and colleagues highlighted the behavioural and gastrointestinal abnormalities, joint laxity and hypotonia [36]. Of the four children in that series, all had behavioural abnormalities (autistic features in three and anxious behaviours in three), all had an increased body mass index and three had macrocephaly; the mother of the two siblings, initially considered unaffected, had macrocephaly, mild learning difficulties, ankle laxity in childhood and stress-triggered tremor, which both children also had [6, 36]. In the van Woerden series, behavioural abnormalities ranged from ADHD to social anxiety, with one case of self-amputation [6]. Blaschek's case, a 16-year-old girl with special educational support in primary school, had a postural and action tremor with onset in childhood, worsening over the previous two years, worse at the end of the day and in the right hand, which impaired writing and eating, with mild saccadic abnormalities and mild dysmetria on serial finger movements and a de novo frameshift variant [6, 64].

3.7. Comparative table of the class

4. Clinical Recognition and Red Flags

4.1. When to suspect

None of the five disorders has a pathognomonic sign, and clinical diagnosis alone is not possible; what is recognised is the indication for broad genetic testing. This exists in any child with global developmental delay or ID without an evident acquired cause, especially if expressive language delay is disproportionate, if there are autism spectrum or hyperactivity behavioural abnormalities, if HC is outside the limits (below −2 SD or above +2 SD) or if anomalies of other organs coexist [8, 19]. In infants, hypotonia with feeding difficulties and poor weight gain is the most frequent presentation of the TRIO, TAOK1 and CAMK2B forms [8, 16, 18]. At preschool and school age, speech delay, attention deficit and ASD become the reasons for referral. In adulthood, the diagnosis arises mainly when investigating a parent of a diagnosed child, because variable expressivity allows parents with mild learning difficulties to transmit the variant [8, 12, 36].

GeneReviews proposes considering the TRIO disorder in the presence of any combination of early feeding difficulties with poor weight gain, motor and speech delay, intellectual functioning between borderline (IQ 70–85) and severe ID, neurobehavioural abnormalities (stereotypies, obsessive-compulsive behaviour, ASD or autistic features, ADHD, aggression, sleep disturbance), abnormal HC, minor finger anomalies (short or tapering fingers, broad PIP joints, 5th finger clinodactyly, 2-3 toe syndactyly), dental anomalies (crowding, delayed or failed eruption) and scoliosis or kyphosis [8]. By analogy, the clues that guide exome interpretation towards the other genes are congenital heart disease or HCM with developmental delay (NAA15) [13], the combination of ASD, ADHD, sleep disturbance and genital malformations (ASH1L) [14], marked hypotonia with visual impairment, episodic ataxia or acquired microcephaly (CAMK2B) [15, 16], and macrocephaly with neonatal hypotonia, joint hypermobility or tremor (TAOK1) [18, 64]. There is no newborn screening for any of these disorders.

4.2. Warning and emergency signs

5. Diagnosis: Criteria, Testing and Algorithms

5.1. Diagnostic criteria

There are no formal clinical criteria for any of the five disorders. The diagnosis is established in a proband with compatible findings and a heterozygous pathogenic or likely pathogenic variant in the gene, classified according to the American College of Medical Genetics and Genomics and Association for Molecular Pathology (ACMG/AMP) standards, in which “pathogenic” and “likely pathogenic” are synonymous for clinical decision-making; a variant of uncertain significance (VUS) neither establishes nor excludes the diagnosis [8, 66]. For ASH1L, the 1q22 deletion that includes the gene also establishes the diagnosis [10, 33]. For CAMK2B, a whole-gene deletion is not enough, on its own, to assign MRD54, given the haploinsufficiency score of 1 [10].

5.2. Genetic investigation guidelines

There is no specific guideline for any of the five genes; GeneReviews records the absence of clinical practice guidelines for the TRIO disorder and the literature search found no consensus statements for the other four [8]. The general guidelines for ID and global developmental delay apply. The ACMG evidence-based guideline (2021), using GRADE methodology, strongly recommends that exome or genome sequencing be considered as a first- or second-tier test in paediatric patients with one or more congenital anomalies with onset before 1 year, or with GDD or ID with onset before 18 years, because it has a higher diagnostic yield than conventional tests and may be more cost-effective when requested early [67]. The meta-analysis and consensus by Srivastava (2019), with 30 studies, estimated an exome yield of 36% overall, 31% in isolated neurodevelopmental disorders and 53% when associated with other conditions, versus 15–20% for chromosomal microarray (CMA), and recommended the exome as a first-tier test in unexplained neurodevelopmental disorders [68]. The 2025 American Academy of Pediatrics (AAP) clinical report replaces the 2014 report and proposes phenotype-driven testing when there is a clinical hypothesis and, in its absence, an agnostic approach chosen by yield and practical considerations [69]; the 2014 report recommended CMA as first tier, fragile X testing and the exome as an emerging option, and remains of historical interest [70]. The 2013 ACMG guideline on the aetiological evaluation of ASD predates the 2021 guideline [71]. The position of the Oxford Desk Reference (2017), which placed the exome after CMA, has been superseded by the 2019–2021 guidelines [19].

5.3. Practical algorithm

The proposed pathway, consistent with the guidelines above, begins with a complete clinical assessment (three-generation pedigree with consanguinity, pregnancy and delivery, milestones and regression, HC, growth, dysmorphism, neurological and skin examination) [19]. If the phenotype suggests a recognisable syndrome, a targeted test is requested. Otherwise, an exome or genome is requested, preferably as a trio, with copy number variant analysis; when the exome does not include reliable CNV analysis, CMA is added [8, 67, 68]. Single-gene testing for TRIO is rarely useful and is not recommended, and some ID panels may not include the most recent genes, which favours the exome [8]. FMR1 testing remains indicated in undiagnosed boys and girls, because the exome does not detect repeat expansions [19]. If the exome is negative, periodic reanalysis of the data and the genome are options: in the patient in section 10, the CAMK2B variant was identified on reanalysis three years after the initial exome [49]. Once the variant has been identified, characterisation is completed with the investigations in section 7.

5.4. Sensitivity and limits of testing

The published detection percentages refer to the proportion of known pathogenic variants detectable by each method and not to clinical sensitivity: for TRIO, sequencing detects about 95% and targeted deletion and duplication analysis about 5% [8]. For the other genes there are no formal estimates. The relevant limits are as follows. Intragenic and multiexon deletions, frequent in TAOK1 (4 of 37 variants), described in ASH1L (exons 2–12) and in TRIO, may escape an exome without CNV analysis and require CMA, validated exome CNV analysis or a genome [8, 18, 33]. Partial duplications may be misinterpreted: in ASH1L, array and exome described as a single event two tandem duplications that only optical mapping resolved, and the absence of messenger RNA reduction on RNA sequencing contradicted the prediction of frameshift with NMD [50]. Intronic variants near the acceptor (TAOK1 c.1909-7G>A, NAA15 c.692-5A>G, TRIO c.4860−2A>G) were only classified with RNA studies [11, 18, 20]. Parental mosaicism, somatic or germline, may not be detected in blood DNA: in the Di Muro sibship, the parents were not carriers in blood and three siblings had the deletion [33]. Undescribed missense variants often require functional studies; the Erasmus MC natural history study of the CAMK2 disorders excludes VUS without functional analysis [72]. A negative exome does not exclude these disorders. There is no validated methylation episignature for ASH1L [10].

5.5. Interpretation and report content

Some interpretation rules are specific to the class. TRIO truncating variants exist in population controls, especially in the last exon, which makes it necessary to confirm the prediction of NMD before applying the loss-of-function criterion [8]. TRIO missense variants should be located relative to spectrin 7 and GEFD1, because location predicts the phenotype [8, 11]. In TAOK1, truncating variants in the C-terminal tail (residues 901–1001) have no demonstrated pathogenicity [18]. In NAA15, in April 2022 there were 64 missense or in-frame variants in ClinVar, many classified as VUS [24]; three were reclassified as likely pathogenic in the longitudinal cohort [30]. In CAMK2B, the low haploinsufficiency score and the existence of variants with opposite effects make functional studies particularly useful [10, 16]. Dual diagnoses are frequent (14% in the TAOK1 cohort, including 17q11.2 deletions encompassing NF1, SLC6A1 variants and congenital adrenal hyperplasia) and the report should list all relevant variants [18]. The report includes the gene, the versioned transcript (MANE Select: NM_007118.4, NM_057175.5, NM_018489.3, NM_001220.5, NM_020791.4), the HGVS description at DNA and protein level, zygosity, origin (de novo, inherited or unknown), the ACMG/AMP class with the criteria applied and, for CNVs, the genomic coordinates and the genes included [48, 66]. The GeneReviews authors offer to review TRIO VUS [8].

5.6. Non-molecular investigations

There is no biochemical biomarker, nor any non-molecular test that confirms any of these disorders [10]. Extensive laboratory panels (digestive, metabolic, hormonal) were not useful in the NAA10/NAA15 group [24]. Brain MRI is indicated in the presence of micro- or macrocephaly, neurological signs, seizures or regression [19]; it is useful for documenting ventriculomegaly (TRIO, TAOK1) and cerebellar atrophy (CAMK2B), but a normal MRI does not exclude later cerebellar atrophy [15]. EEG is performed when seizures are suspected and should be considered in all carriers of TRIO loss-of-function or truncating variants [8]. Echocardiography and ECG are indicated in NAA15 carriers and in the presence of cardiac signs in the others [8, 13].

5.7. Prenatal and preimplantation diagnosis

Prenatal diagnosis and preimplantation genetic testing (PGT) are only possible when the familial variant has been identified [8, 19]. Prenatal ultrasound may show ventriculomegaly or macrocephaly (TRIO, TAOK1) or heart disease (NAA15) [9, 12, 35], but these findings are non-specific; no source describes prenatal diagnosis of these disorders by fetal exome with yield data. Reproductive options are covered in section 8.

6. Differential Diagnosis

The differential of each of these disorders is, in practice, the full set of monogenic and chromosomal causes of ID without distinctive signs, including the OMIM series of autosomal dominant, autosomal recessive and X-linked, syndromic and non-syndromic ID [8]; discrimination is therefore made by broad genetic testing and not by successive targeted tests. The table brings together the differentials with a direct relationship (same pathway, paralogues, CNVs containing the genes, allelic phenotypes) and other relevant entities.

NAA10-related disorder. This is the closest differential of NAA15, because the two proteins form the same complex and the disorders share ID, speech delay, behavioural abnormalities and ophthalmological and cardiac findings [24, 60]. Automated facial analysis (GestaltMatcher, with 78 individuals with NAA10 and 33 with NAA15) showed a mean distance below the threshold for distinction, that is, an overlapping gestalt; on classification by Human Phenotype Ontology (HPO) terms, the two disorders cluster separately with overlap [24]. The NAA10 disorder is X-linked, generally more severe, with lower adaptive functioning and more MRI anomalies [24, 30, 59]. In an international cohort of 30 individuals with missense variants in NAA10 or NAA15, no consistent facial pattern was identified, although thick eyebrows and a wide philtrum were frequent [29].

TAOK2 and the 16p11.2 region. TAOK2 is the paralogue of TAOK1 located in the proximal 16p11.2 region. In the same publication that delineated the largest TAOK1 cohort, 10 individuals with TAOK2 variants all had a neurodevelopmental abnormality, macrocephaly and ASD in 75%, overweight or obesity in 70% and tall stature in four [18]. Overweight and tall stature point towards TAOK2, and macrocephaly with neonatal hypotonia and hypermobility points towards TAOK1; the 16p11.2 deletion and duplication enter the differential.

CaMKII paralogues. CAMK2A variants cause an overlapping picture, with ID and behavioural abnormalities; in the Küry series, CAMK2A carriers had less hypotonia (7/14) and fewer digestive problems (1/12) than CAMK2B carriers [16]. Akita suggested that CAMK2A is more associated with seizures and CAMK2B with cerebellar symptoms, on the basis of few patients [15]. The CAMK2G and CAMK2D forms are part of the same family of “CAMK2 synaptopathies” followed in the Erasmus MC natural history study [72].

CNVs containing the genes. For ASH1L, the 1q22 microdeletion and the intragenic variant are the same disorder [33]. For TAOK1, 17q11.2 deletions proximal to the usual NF1 region (807 bp to 2 Mb) reproduce the phenotype, and deletions extending to NF1 combine the two disorders; the microcephaly described in a 1.69 Mb deletion does not appear to be explained by loss of TAOK1 [18, 35]. For TRIO, the 5p deletions of cri-du-chat syndrome include the gene, and the microcephaly of these deletions is attributed in part, by positional inference, to TRIO haploinsufficiency. For CAMK2B, inclusion of the gene in a deletion is not enough to explain the phenotype, because of the low haploinsufficiency score [10].

Allelic phenotypes expressed in other areas. Congenital hydrocephalus with a TRIO variant enters the differential of genetic hydrocephalus [9]; paediatric HCM with developmental delay should prompt consideration of NAA15, according to the authors who described it [13]; childhood-onset tremor should prompt consideration of TAOK1 [64, 65]. ADHD and ASD are frequent comorbidities of the class.

7. Treatment and Risk Management Strategies

7.1. General principles

There is no specific, disease-modifying or approved treatment for any of the five disorders, and no registered interventional clinical trial; management is directed at the manifestations, multidisciplinary and coordinated by the attending physician, with the developmental paediatrician, the paediatric neurologist and the geneticist [8, 10, 74]. A 2023 review of CAMK2B mechanisms expressly states that there is no targeted therapy [75]. The GeneReviews entry for the TRIO disorder is the only source with structured recommendations for assessment, treatment and surveillance; because the clinical problems of the other four genes are of the same type, these recommendations serve as a model for the class, with gene-specific additions taken from the published series. This extrapolation is an editorial decision and not a recommendation of a scientific society.

7.2. Assessment after diagnosis

7.3. Non-pharmacological interventions

Development and education. GeneReviews describes the North American model (early intervention from 0 to 3 years with occupational therapy, physiotherapy, speech and feeding therapy; a preschool programme with an individualised education plan from 3 to 5 years; an education plan reviewed annually at school age and a transition plan in adolescence), whose equivalents in Portugal and Brazil are the national early intervention and inclusive education systems [8]. Physiotherapy improves mobility and reduces late orthopaedic complications such as scoliosis; occupational therapy addresses fine motor skills; oromotor dysfunction is assessed at every visit [8]. Augmentative and alternative communication (AAC) should be assessed by a speech and language therapist, because it does not delay speech and supports language development; this indication is particularly relevant in CAMK2B, in which many patients do not speak after 5 years of age [8, 10, 16].

Behaviour. ASD interventions, including individual applied behaviour analysis (ABA) with a certified analyst, and the behavioural management strategies of the developmental paediatrician are the first line; severe aggressive or destructive behaviour warrants referral to child psychiatry [8]. In ASD, the AAP recommends standardised screening at 18 and 24 months, attention to comorbidities (sleep, feeding, gastrointestinal problems, obesity, seizures, ADHD, anxiety, elopement) and evidence-based behavioural interventions, with shared decision-making with the family [76].

Feeding and growth. Feeding therapy, a food diary with calorie counting, thickening or cooling of food and, in severe or persistent difficulties, a nasogastric tube or gastrostomy are the measures described [8]. A radiological swallowing study is indicated in the presence of choking, poor weight gain, frequent respiratory infections or food refusal [8]. In the combined NAA10/NAA15 experience of the Lyon group, calorie supplementation and feeding therapy did not normalise weight, two children with growth hormone deficiency who were treated continued to have growth delay, and gastrostomy was effective for weight gain and made care easier; these data come mainly from patients with NAA10 [24].

Orthopaedics, teeth and other areas. Scoliosis and kyphosis are followed by orthopaedics, crowding and malocclusion by a dentist or orthodontist, recurrent infections by immunology, and heart disease by paediatric cardiology [8]. Constipation and gastro-oesophageal reflux receive standard treatment [8].

7.4. Symptomatic pharmacological treatment

The drugs used treat comorbidities (irritability and aggression associated with ASD, ADHD, insomnia, epileptic seizures) and not the genetic disorder; no source gives doses for these disorders in particular, nor is there a dosing guideline for them. The doses in the table are those of the official label for the comorbidity (Food and Drug Administration, FDA, in the United States; European Medicines Agency, EMA, in the European Union) and their application to the disorders in this article is an extrapolation, declared as such. Names follow INFARMED; brand names are given according to the jurisdiction of the label consulted. INFARMED summaries of product characteristics (SmPC) and the package inserts of the Brazilian Health Regulatory Agency (ANVISA) were not consulted.

Three observations limit the application of this table. The harmonised European indication for risperidone in children, set by EMA referral in 2008, is short-term symptomatic treatment (up to 6 weeks) of persistent aggression in conduct disorder, from 5 years of age, in children and adolescents with below-average intellectual functioning or intellectual disability, and not the irritability associated with autistic disorder of the US indication; the CHMP excluded autistic disorder, so use for irritability in ASD in Europe is an extrapolation from the US label [77, 81]. The European dosage is 0.25 mg/day below 50 kg (optimum 0.5 mg/day; maximum 0.75 mg/day) and 0.5 mg/day from 50 kg (optimum 1 mg/day; maximum 1.5 mg/day), with increases at most every other day [81]. The methylphenidate label advises against use in cardiomyopathy and severe structural heart disease, so a carrier of an NAA15 variant should have a cardiac assessment before starting stimulants [13, 79]. Aripiprazole should be used with caution in patients with a history of seizures, which are frequent throughout the class [78].

Epileptic seizures. Treatment is standard, chosen by the neurologist according to seizure type and electroclinical syndrome, without a preferred drug for any of these genes [8, 10]. Individual reports describe seizures partially controlled with clonazepam in CAMK2B [15], clonazepam-resistant epilepsy in ASH1L [14] and, in a 3-year-old child with CAMK2B, a marked reduction in seizure frequency with levetiracetam 350 mg twice daily, topiramate 30 mg twice daily and vitamin B6 50 mg daily, after previous treatment failures [37]. These doses are those of the report and do not constitute a recommendation.

7.5. Treatments described in case reports

The following treatments have been published in isolated patients and have the level of evidence of a case report; they are not routine recommendations. In an adolescent girl with CAMK2B p.(Glu110Lys), a gain-of-function variant, with periodic neuropsychiatric episodes, an empirical combination, adjusted over 2 years, of verapamil, lithium, lamotrigine, valproic acid, magnesium oxide and riboflavin reduced the severity of the episodes; the verapamil dose was limited by bradycardia and Wenckebach-type atrioventricular block, and prochlorperazine and paroxetine worsened agitation, the latter also being stopped because of prolongation of the corrected QT interval [49]. The doses are not given in the report. The authors propose, as a hypothesis, choosing drugs by their predicted effect on CaMKII activity (reducing calcium signalling in gain of function), without validation [49]. In NAA15, an infant with obstructive HCM remained stable on propranolol monotherapy and a child with multifocal ectopic atrial tachycardia was treated with verapamil, without doses in the sources [12, 13]; management of HCM follows the general guidelines. In the adolescent girl with postural and action tremor associated with TAOK1, propranolol produced a good symptomatic response; the dose and duration are not given in the synopsis consulted, and use in TAOK1-associated tremor is off-label, so dosage follows the rules for tremor of other causes and the national label of the drug, with the usual precautions for beta-blockers [6, 64]. A participant in the NAA15 longitudinal cohort with a favourable adaptive course was receiving hyperbaric oxygen therapy sessions, a relationship the authors consider speculative [30]. A 2026 letter describes psychosis after hormonal therapy in the NAA15 disorder; the content was not consulted, but the title warrants caution with hormonal treatments in these patients [82].

7.6. Regulatory status

The dates of initial authorisation and of the extensions of these drugs are not given in the label sections consulted.

7.7. Clinical trials and studies

A ClinicalTrials.gov search in October 2026 for the five genes found no interventional trials; there are two observational studies of direct relevance [72, 74]. For the symptomatic drugs, the trials supporting the indications were conducted in children with autistic disorder in general, without genetic stratification.

The limits of these data are clear. The antipsychotic trials were conducted in populations with autistic disorder without a genetic diagnosis, and it cannot be assumed that response or tolerability are the same in patients with these monogenic disorders, who have more seizures, hypotonia and feeding problems. The trials are not compared with one another. The absence of registered interventional trials does not prove that no studies are in preparation. No negative, suspended or withdrawn trials were found for these genes.

7.8. Developments in the last 12 months and research

Other, older preclinical data point to hypothetical pathways: PAK1 inhibition reversed behavioural defects in mice, suggesting a pathway for the TRIO forms with RAC1 hyperactivation [11]; chemogenetic inhibition of the prefrontal cortex improved social deficits and absence-like seizures in an Ash1l mouse [10]; and reintroduction of normal Taok1 in adult mice attenuated autistic-like behaviours [47]. None of these approaches has been tested in humans, and they should not be presented to families as treatment.

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

There is no validated list of contraindicated drugs in these disorders. The available signals come from single cases or labels: severe worsening with prochlorperazine and paroxetine in a patient with gain-of-function CAMK2B [49]; caution with stimulants in NAA15-associated HCM [13, 79]; caution with antipsychotics in patients with seizures [78]; and caution with hormonal treatments in NAA15 [82]. The sources consulted do not describe pregnancies in women with these disorders, nor data on lactation; planning a pregnancy in a carrier follows the general rules for ID and for the drugs in use, and reproductive counselling is covered in section 8. Nor do the sources specifically address the perioperative period; the risks foreseeable from the phenotype are aspiration in patients with dysphagia, scoliosis surgery and, in NAA15, heart disease, which should be known before anaesthesia. Transition to adult care should be planned in adolescence, with an educational and social transition plan [8], and with attention to the psychiatric disorders described in adolescents and adults with TAOK1 and to the adaptive decline described in girls with NAA15 [18, 30].

7.10. Guidelines and divergences

There is no specific guideline or consensus for any of the five genes, and the GeneReviews entry for the TRIO disorder states the absence of clinical practice guidelines [8]. The applicable general guidelines are those for the genetic diagnosis of ID (section 5.2), the AAP guideline for ASD [76] and, for NAA15-associated HCM, the 2024 AHA/ACC guideline [85]. The relevant divergence between sources is regulatory and not clinical: the paediatric indication for risperidone differs between the USA and the EU [77, 81]. Among primary sources there is a divergence on the need for cardiac surveillance in NAA15: Ritter and colleagues propose lifelong cardiac surveillance, whereas the ClinGen curation considers the evidence emerging and no source issues a formal screening recommendation [10, 13].

8. Reproductive Options and Family Counselling

8.1. Inheritance and recurrence risks

All five disorders have autosomal dominant inheritance. Most cases result from de novo variants, but the proportion of inherited variants, almost always from a parent with a mild phenotype or without manifestations, varies between genes and is important for counselling. In the TRIO disorder, about 15% of the 56 individuals reviewed inherited the variant from an affected parent; missense variants of both mechanisms are typically de novo, and truncating variants may be de novo or inherited [8]. In the NAA15 founding series, inheritance was familial in 3 of 33 families, with a mildly affected parent and affected siblings [12]; the statistically estimated penetrance of 35.3% for likely disruptive variants means that a carrier parent may have no recognised developmental delay [3, 28], and a Chinese series described four families with several affected members [20]. In TAOK1, 11 of the 50 individuals in the largest cohort inherited the variant from an affected parent (15 de novo, 24 of unknown origin), and a sibship with a variant transmitted by a mildly affected mother has been described [18, 36]. In CAMK2B, a truncating variant was transmitted by a mother without manifestations, of normal intelligence, to two daughters with different expressivity [37]. In ASH1L, three siblings inherited an intragenic deletion from parents who were not carriers in blood, through gonadal or gonosomal mosaicism [33].

8.2. Parental testing and cascade testing

Testing the parents for the proband's variant is recommended in all cases, even when they are clinically healthy, because transmission by parents without manifestations or with mild manifestations has been demonstrated in CAMK2B, TAOK1, NAA15 and TRIO [8, 12, 36, 37]. As well as defining the recurrence risk, the result contributes to variant classification, because a de novo variant with confirmed parentage is a strong criterion of pathogenicity [66]. If the variant is absent from the blood of both parents, analysis of another tissue or of the variant in sperm is not routine and does not exclude exclusively germline mosaicism [8]. When a parent is a carrier, siblings can be tested to guide developmental support and, in adults, reproductive planning. Other relatives are only at risk if one of the parents is a carrier [8].

8.3. Prenatal diagnosis and preimplantation genetic testing

Once the familial variant has been identified, prenatal diagnosis by chorionic villus sampling or amniocentesis and PGT are possible; views on the use of prenatal diagnosis vary among professionals and families and the decision is personal [8]. The ideal time to discuss the risk and these options is before pregnancy [8]. In de novo cases, prenatal diagnosis may be offered for reassurance, given the residual risk due to mosaicism. Variable expressivity makes it impossible to predict severity in a carrier fetus, especially for genes with incomplete penetrance, which should be explained before testing.

8.4. Content of counselling

Counselling covers an explanation of the mechanism and the gene in accessible language; the usually de novo nature and the absence of parental responsibility; the recurrence risk according to the result of parental testing, including the residual risk due to mosaicism; clinical variability, even within the family; the 50% risk to the offspring of affected people themselves, which should be addressed in adolescence and at transition, adapted to their capacity for understanding; and referral to patient organisations and research registries [8, 33, 37]. Testing of asymptomatic minors with a carrier parent is justified when there is clinical benefit, namely early developmental intervention or cardiac screening in families with NAA15. In families without a molecular diagnosis, the empirical recurrence risks for ID apply [19].

9. Special Considerations: Population Diversity and Epidemiology

9.1. What the figures measure

There are no population prevalence or incidence estimates for any of the five disorders, and Orphanet has no prevalence record for the corresponding entities [10, 21]. The published figures are of a different nature and should not be converted into prevalences. ID as a whole affects 1–2.5% of the population (IQ < 70) and 0.3–0.5% in its severe form (IQ < 50), with a male predominance of 1.3:1 [19]; monogenic disorders revealed by exome sequencing account for about a quarter to a third of ID with an aetiological diagnosis [19]. Within this universe, each gene represents a small fraction, measured as yield in selected cohorts.

The denominators of the clinical series (24 individuals with TRIO in the Barbosa series, 38 with NAA15, about 16 with ASH1L, 10 with CAMK2B in the Küry series, 50 with TAOK1) reflect the number of cases gathered and not the frequency of the disease; underdiagnosis, incomplete penetrance and recruitment through families seeking registries make them poorly representative of mild pictures [8, 18, 24].

9.2. Geographical distribution of the series

The TRIO series come mainly from Europe: Wessex and the DDD study in the United Kingdom and an international cohort that brought together data from DDD, the 100,000 Genomes Project, a North American genomics institute and clinical practice [11, 26], with isolated cases in Japan and China [27, 41]. The NAA15 series come mainly from the United States and from international collaborations of the Lyon group [12, 24, 30], with a large Chinese ASD cohort (64 distinct variants in individuals with neurodevelopmental disorders, in the cohort and in the literature review) and four Chinese families [20, 42]. The ASH1L cases come from the United States, Japan, China, Italy, France, Uzbekistan, Russia and Saudi Arabia [14, 31, 33, 43, 44, 50, 62, 73]. The CAMK2B founding series included patients from the United States, France, the United Kingdom, the Netherlands, Germany and Norway, with later Japanese and Italian cases [15–17, 54]. The largest TAOK1 cohort was assembled from the British DDD, a North American laboratory and international platforms [18]. No founder variants or differences in frequency between populations have been described; the absence of series in many regions reflects access to the exome and not the real distribution.

9.3. Portugal and Brazil

The sources consulted contain no data from Portugal on these genes. In Brazil, a series of 39 children with a clinical diagnosis or suspicion of Silver-Russell syndrome studied at a tertiary centre in São Paulo included a NAA15 variant among the alternative diagnoses [61]. The lack of data in the sources consulted does not demonstrate the absence of diagnosed patients in these countries.

9.4. Special populations

Girls with NAA15 had better motor functioning than boys, but adaptive decline with age in the longitudinal cohort, in contrast with boys [30]. In ASH1L, published probands are mostly male, and a preprint suggests a predominance of epilepsy in females and of ASD in males [14, 45]. Adults are under-represented in all series: the TAOK1 cohort includes individuals up to 52 years of age, with psychiatric disorders in six [18], and the oldest patient described with CAMK2B p.(Pro139Leu) was 22 years old [17]. Parents diagnosed after their child are a particular population, often with mild learning difficulties and without previous follow-up [8, 36].

9.5. Reference networks and patient organisations

In Europe, the relevant reference network is ERN-ITHACA, which brings together centres specialised in rare malformation syndromes and rare neurodevelopmental disorders and ID [87]. The organisations and registries identified are Team TRIO (www.teamtrio.org), a support network for families and professionals [8]; Simons Searchlight, an international registry and family community that includes ASH1L, NAA15, TAOK1 and TRIO [74]; the CAMK2 Therapeutics Network (camk2.org) and the Erasmus MC ENCORE specialist clinic for the CAMK2 disorders [72, 88]; the ASH1L Syndrome & Gene portal (ash1l.org) [89]; the Brain Gene Registry of the North American ID and developmental research centres [14]; and the Human Disease Gene website series, which aggregates data from families with NAA15, TRIO and other genes [8, 24]. For ID in general, the AAIDD in the United States and Mencap in England and Wales are the reference organisations cited [8, 19].

10. Illustrative Clinical Case

In 2022 Dwyer and colleagues described a Caucasian girl, then aged 15 years, without dysmorphism, with a gain-of-function CAMK2B variant [49]. In the neonatal period she lacked the oral coordination to breastfeed, which persisted until 6 months, and she cried inconsolably every day. She developed moderate to severe psychomotor delay, with hypotonia and apraxia: she sat at 12 months, crawled at 13 and walked at 17, with intensive occupational therapy; she was continent at 4 years; with speech therapy, she formed two- to three-word sentences at 3–4 years and told short narratives at 6, with near-normal comprehension. At 15 years she read with difficulty, because of attention deficit and visual apraxia, added single-digit numbers, showed empathy and enjoyed interacting, but still had difficulty sustaining a conversation and reacted to frustration with self-injury and aggression. Brain MRI, karyotype, CMA and metabolic studies had been normal.

At 7 years she had a one-week episode of encephalopathy, with sudden onset and offset and no trigger: confusion, agitation, inconsolable crying, regression of skills, mutism except for stereotyped phrases, and refusal of food and fluids. Over the following months the episodes recurred, lasting 7–10 days, and came to occur every 21–28 days for 8 months, requiring hospital admissions for sedation, nutrition and hydration and bladder catheterisation for urinary retention, with constipation requiring disimpaction. In an emergency department, prochlorperazine given for suspected headache caused severe disorientation and agitation with stereotyped movements. Investigation of altered mental status was negative: repeat MRI, EEG, anti-NMDA receptor and antistreptococcal antibodies in CSF, metabolic studies and porphyria screening. During one crisis, the patient complained of severe headache and had unilateral ptosis with ipsilateral scleral injection; with a working hypothesis of cluster headache she started verapamil and lamotrigine, with a reduction in symptoms at each increase in verapamil, and then lithium, by analogy with Kleine-Levin syndrome, with immediate improvement of the disorientation; she no longer needed admission and began attending school. Reduction of verapamil, required because of bradycardia and intermittent Wenckebach block, worsened the episodes; paroxetine was associated with almost daily confusion, agitation and aggression and was stopped because of prolongation of the corrected QT, with resolution of the aggression between episodes [49].

Trio exome sequencing, performed at 7 years, was initially negative; reanalysis at 10 years identified de novo NM_001220.5:c.328G>A p.(Glu110Lys) in CAMK2B, a variant that increases Thr287 phosphorylation and reduces protein expression in HEK293T cells, and the patient was included in the Küry series [16, 49]. After 2 years of empirical adjustments, the combination of verapamil, lithium, lamotrigine, valproic acid, magnesium oxide and riboflavin controlled the symptoms; 8 years after the episodes were recognised, they persist every 21–28 days, lasting 7–10 days, but with much lower severity, in the form of difficulty calming down, self-injury, impulsivity, poor concentration and transient loss of language and complex motor coordination [49]. The paper presents the course on a timeline (figure 1 of the paper), which was not consulted as an image.

The case illustrates three points. First, reanalysis of a negative exome may yield the diagnosis years later, as new genes are described. Second, the CAMK2 disorders may have episodic manifestations, here neuropsychiatric and in other patients in the form of episodic ataxia [15], which must be distinguished from autoimmune encephalitis, non-convulsive status epilepticus and metabolic diseases. Third, the therapy used was empirical and based on a single patient; the authors propose selecting drugs by their predicted effect on CaMKII activity, an untested hypothesis, and the regimen does not constitute a recommendation nor correspond to any guideline.

11. Prognosis and Long-Term Follow-Up

11.1. Prognosis

Prognosis is determined by the severity of the ID, by epilepsy and by organ comorbidities, and there are no life expectancy data for any of the five disorders. None is, in the published series, a typical neurodegenerative disease, but there are three partial exceptions: progressive cerebellar atrophy and acquired microcephaly in CAMK2B [15, 17], progressive leukoencephalopathy in a carrier of a TRIO variant [8] and regression in a patient with TAOK1 who also had a CDKL5 variant [18].

In the TRIO disorder, outcome data are scattered: individuals with gain-of-function variants were not walking at 6–10 years, one had a first seizure at 30 years, and individuals with loss of function were non-verbal at 14 and 16 years [8, 11]. In NAA15, the longitudinal study of 27 individuals found a negative correlation between age and overall adaptive functioning (r = −0.38; p = 0.01), with decline in daily living, socialisation and the motor domain and an increase in internalising behaviours with age (r = 0.49); the decline was evident in girls (r = −0.51; p = 0.04), whereas in boys, excluding a patient with focal seizures and catatonia with onset in adolescence, the correlation was positive (r = 0.38; p = 0.04) [30]. These results come from 17 patients assessed twice and two assessed three times, with a mean interval of 1.45 years, and contradict the slight improvement described in a series of four children; the authors call for more longitudinal data [30]. NAA15-associated HCM has a variable onset (2 months and 6 years in the cases described) and a course yet to be defined [13]. In TAOK1, motor milestones catch up, but speech difficulties persist into adulthood, and anxiety, depression and bipolar disorder have been described in adults [18]. In ASH1L there are no outcome data beyond isolated cases, and in CAMK2B the oldest patient described was 22 years old [14, 17]. All data come from patients with early intervention and educational support, without untreated comparison groups.

11.2. Surveillance

The table adopts the recommendations of the GeneReviews entry for the TRIO disorder as a common basis and adds the gene-specific elements described in the sources. The sources do not set numerical intervals: surveillance is carried out “at each visit”, with a frequency defined by the care team according to age and active problems [8].

12. Summary and Concluding Remarks

The intellectual developmental disorders caused by heterozygous variants in TRIO, NAA15, ASH1L, CAMK2B and TAOK1 form a class with a common model: global developmental delay recognised in the first years, ID of variable severity that may be absent, disproportionate language delay, autism spectrum and hyperactivity behavioural abnormalities, hypotonia and feeding difficulties in infancy, seizures in some patients and no biochemical marker. Most variants arise de novo, but all five can be transmitted by parents with a mild phenotype or without manifestations, and in ASH1L recurrence due to gonadal mosaicism has been documented; parental testing is therefore indispensable for counselling.

Each gene adds to the common core elements that guide exome interpretation and surveillance. In TRIO, head circumference distinguishes the variants in the first GEF domain and truncating variants, with loss of function, microcephaly and milder ID, from the variants in the seventh spectrin repeat, with gain of function, macrocephaly, more severe ID, severe feeding difficulties and scoliosis; congenital ventriculomegaly and hydrocephalus were added to the spectrum in 2026. In NAA15, the disorder is generally milder than that of NAA10, but it includes congenital heart disease, paediatric HCM, dystonia and, in girls, adaptive decline with age. In ASH1L, the intragenic variant and the 1q22 microdeletion are the same disorder, with ASD, ADHD, sleep disturbance, genital malformations and sometimes dominant epilepsy. In CAMK2B, variants with opposite effects on autophosphorylation produce ID with severely impaired speech, hypotonia, visual impairment, episodic ataxia and, in some patients, progressive cerebellar atrophy and periodic neuropsychiatric episodes. In TAOK1, macrocephaly with neonatal hypotonia and hypermobility accompanies a cognitive deficit that may be minimal, with psychiatric disorders in adulthood.

Diagnosis rests on first- or second-tier exome or genome sequencing, according to the general ID guidelines, with CNV analysis, RNA studies for splicing variants and functional studies for undescribed missense variants. There is no specific therapy; treatment is supportive, with early intervention, AAC, behavioural interventions, feeding management, standard seizure treatment and use of psychotropic drugs according to the labels for the comorbidities. Natural history data come from small series recruited by ascertainment, without systematic follow-up of adults, and therapeutic research is still preclinical; the international registries and the prospective study of the CAMK2 disorders are, at present, the main sources of new knowledge.

13. References

1. OMIM® Online Mendelian Inheritance in Man. #617061 INTELLECTUAL DEVELOPMENTAL DISORDER, AUTOSOMAL DOMINANT 44, WITH MICROCEPHALY; MRD44. Johns Hopkins University; last edited 25/02/2022. https://omim.org/entry/617061. Accessed 28 October 2026.
2. OMIM® Online Mendelian Inheritance in Man. #618825 INTELLECTUAL DEVELOPMENTAL DISORDER, AUTOSOMAL DOMINANT 63, WITH MACROCEPHALY; MRD63. Johns Hopkins University; last edited 21/03/2023. https://omim.org/entry/618825. Accessed 28 October 2026.
3. OMIM® Online Mendelian Inheritance in Man. #617787 INTELLECTUAL DEVELOPMENTAL DISORDER, AUTOSOMAL DOMINANT 50, WITH BEHAVIORAL ABNORMALITIES; MRD50. Johns Hopkins University; last edited 10/11/2021. https://omim.org/entry/617787. Accessed 28 October 2026.
4. OMIM® Online Mendelian Inheritance in Man. #617796 INTELLECTUAL DEVELOPMENTAL DISORDER, AUTOSOMAL DOMINANT 52; MRD52. Johns Hopkins University; last edited 28/04/2022. https://omim.org/entry/617796. Accessed 28 October 2026.
5. OMIM® Online Mendelian Inheritance in Man. #617799 INTELLECTUAL DEVELOPMENTAL DISORDER, AUTOSOMAL DOMINANT 54; MRD54. Johns Hopkins University; last edited 15/05/2024. https://omim.org/entry/617799. Accessed 28 October 2026.
6. OMIM® Online Mendelian Inheritance in Man. #619575 DEVELOPMENTAL DELAY WITH OR WITHOUT INTELLECTUAL IMPAIRMENT OR BEHAVIORAL ABNORMALITIES; DDIB. Johns Hopkins University; last edited 22/05/2023. https://omim.org/entry/619575. Accessed 28 October 2026.
7. National Center for Biotechnology Information. MedGen [Internet]. Bethesda (MD): National Library of Medicine [cited 28 Oct 2026]. Concepts C5394205, C4310740, CN381290, C4540470, C4540478, C4540484, C5562004. Available from: https://www.ncbi.nlm.nih.gov/medgen/
8. Varvagiannis K, Vissers LELM, Baralle D, de Vries BBA, Gazdagh G. TRIO-related neurodevelopmental disorder. 2017 Aug 10 [updated 2023 Mar 23]. In: Adam MP, Feldman J, Mirzaa GM, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. Available from: https://www.ncbi.nlm.nih.gov/books/NBK447257/
9. Mehta NH, Dennis E, Allington G, Mekbib KY, Hale AT, Davalan WC, et al. De Novo TRIO Missense Variants Disrupt Ras-GEF Domains and Cause Congenital Ventriculomegaly and Hydrocephalus. Hum Mutat. 2026;2026:8870037. doi:10.1155/humu/8870037 PMID: 42051466. PMC13115915.
10. GeneGuidance [curation database]. Records for the genes TRIO, NAA15, ASH1L, CAMK2B and TAOK1, with ClinGen, OMIM, UniProt and Orphanet data. Extract of 26 Sep 2026 [cited 28 Oct 2026].
11. Barbosa S, Greville-Heygate S, Bonnet M, Godwin A, Fagotto-Kaufmann C, Kajava AV, et al. Opposite Modulation of RAC1 by Mutations in TRIO Is Associated with Distinct, Domain-Specific Neurodevelopmental Disorders. Am J Hum Genet. 2020;106(3):338-355. doi:10.1016/j.ajhg.2020.01.018 PMID: 32109419. PMC7058823.
12. Cheng H, Dharmadhikari AV, Varland S, Ma N, Domingo D, Kleyner R, et al. Truncating Variants in NAA15 Are Associated with Variable Levels of Intellectual Disability, Autism Spectrum Disorder, and Congenital Anomalies. Am J Hum Genet. 2018;102(5):985-994. doi:10.1016/j.ajhg.2018.03.004 PMID: 29656860. PMC5986698.
13. Ritter A, Berger JH, Deardorff M, Izumi K, Lin KY, Medne L, et al. Variants in NAA15 cause pediatric hypertrophic cardiomyopathy. Am J Med Genet A. 2021;185(1):228-233. doi:10.1002/ajmg.a.61928 PMID: 33103328. PMC8007079.
14. Cordova I, Blesson A, Savatt JM, Sveden A, Mahida S, Hazlett H, et al. Expansion of the Genotypic and Phenotypic Spectrum of ASH1L-Related Syndromic Neurodevelopmental Disorder. Genes (Basel). 2024;15(4):423. doi:10.3390/genes15040423 PMID: 38674358. PMC11049257.
15. Akita T, Aoto K, Kato M, Shiina M, Mutoh H, Nakashima M, et al. De novo variants in CAMK2A and CAMK2B cause neurodevelopmental disorders. Ann Clin Transl Neurol. 2018;5(3):280-296. doi:10.1002/acn3.528 PMID: 29560374. PMC5846454.
16. Küry S, van Woerden GM, Besnard T, Proietti Onori M, Latypova X, Towne MC, et al. De Novo Mutations in Protein Kinase Genes CAMK2A and CAMK2B Cause Intellectual Disability. Am J Hum Genet. 2017;101(5):768-788. doi:10.1016/j.ajhg.2017.10.003 PMID: 29100089. PMC5673671.
17. Rizzi S, Spagnoli C, Salerno GG, Frattini D, Caraffi SG, Trimarchi G, et al. Severe intellectual disability, absence of language, epilepsy, microcephaly and progressive cerebellar atrophy related to the recurrent de novo variant p.(P139L) of the CAMK2B gene: A case report and brief review. Am J Med Genet A. 2020;182(11):2675-2679. doi:10.1002/ajmg.a.61803 PMID: 32875707.
18. Elkhateeb N, Crookes R, Spiller M, Pavinato L, Palermo F, Brusco A, et al. Expanding the phenotype and genotype spectrum of TAOK1 neurodevelopmental disorder and delineating TAOK2 neurodevelopmental disorder. Genet Med. 2025;27(3):101348. doi:10.1016/j.gim.2024.101348 PMID: 39737487. PMC13479925.
19. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. Intellectual disability; p. 188-93.
20. Xuan X, Tang J, Zhao X. [Clinical and genetic analysis of four Chinese pedigrees affected with NAA15-related intellectual developmental disorder]. Zhonghua Yi Xue Yi Chuan Xue Za Zhi. 2026;43(6):408-414. doi:10.3760/cma.j.cn511374-20250711-00421 PMID: 42091191.
21. Orphanet. Autosomal dominant non-syndromic intellectual disability. ORPHA:178469 [Internet]. Paris: INSERM [cited 28 Oct 2026]. Available from: https://www.orpha.net/en/disease/detail/178469
22. Orphanet. Non-specific syndromic intellectual disability. ORPHA:528084 [Internet]. Paris: INSERM [cited 28 Oct 2026]. Available from: https://www.orpha.net/en/disease/detail/528084
23. Orphanet. Micrognathia-recurrent infections-behavioral abnormalities-mild intellectual disability syndrome. ORPHA:476126 [Internet]. Paris: INSERM [cited 28 Oct 2026]. Available from: https://www.orpha.net/en/disease/detail/476126
24. Lyon GJ, Vedaie M, Beisheim T, Park A, Marchi E, Gottlieb L, et al. Expanding the phenotypic spectrum of NAA10-related neurodevelopmental syndrome and NAA15-related neurodevelopmental syndrome. Eur J Hum Genet. 2023;31(7):824-833. doi:10.1038/s41431-023-01368-y PMID: 37130971. PMC10325952.
25. Ba W, Yan Y, Reijnders MR, Schuurs-Hoeijmakers JH, Feenstra I, Bongers EM, et al. TRIO loss of function is associated with mild intellectual disability and affects dendritic branching and synapse function. Hum Mol Genet. 2016;25(5):892-902. doi:10.1093/hmg/ddv618 PMID: 26721934. PMC4754042.
26. Pengelly RJ, Greville-Heygate S, Schmidt S, Seaby EG, Jabalameli MR, Mehta SG, et al. Mutations specific to the Rac-GEF domain of TRIO cause intellectual disability and microcephaly. J Med Genet. 2016;53(11):735-742. doi:10.1136/jmedgenet-2016-103942 PMID: 27418539. PMC5264232.
27. Zhang SY, Jin JY, Zeng L, Niu RC, Wang X. Identification and growth features of developmental delay with macrocephaly caused by a novel TRIO variant affecting the second SH3 domain. BMC Med Genomics. 2026;19(1):152. doi:10.1186/s12920-026-02422-6 PMID: 42421094. PMC13632391.
28. Stessman HA, Xiong B, Coe BP, Wang T, Hoekzema K, Fenckova M, et al. Targeted sequencing identifies 91 neurodevelopmental-disorder risk genes with autism and developmental-disability biases. Nat Genet. 2017;49(4):515-526. doi:10.1038/ng.3792. PMID: 28191889.
29. Cheng H, Gottlieb L, Marchi E, Kleyner R, Bhardwaj P, Rope AF, et al. Phenotypic and biochemical analysis of an international cohort of individuals with variants in NAA10 and NAA15. Hum Mol Genet. 2019;28(17):2900-2919. doi:10.1093/hmg/ddz111. PMID: 31127942.
30. Makwana R, Christ C, Patel R, Marchi E, Harpell R, Lyon GJ. Natural History of NAA15 -Related Neurodevelopmental Disorder Through Adolescence. Am J Med Genet A. 2025;197(6):e64009. doi:10.1002/ajmg.a.64009 PMID: 39991982. PMC12052496.
31. Okamoto N, Miya F, Tsunoda T, Kato M, Saitoh S, Yamasaki M, et al. Novel MCA/ID syndrome with ASH1L mutation. Am J Med Genet A. 2017;173(6):1644-1648. doi:10.1002/ajmg.a.38193 PMID: 28394464.
32. Shen W, Krautscheid P, Rutz AM, Bayrak-Toydemir P, Dugan SL. De novo loss-of-function variants of ASH1L are associated with an emergent neurodevelopmental disorder. Eur J Med Genet. 2019;62(1):55-60. doi:10.1016/j.ejmg.2018.05.003 PMID: 29753921.
33. Di Muro E, Petracca A, Castori M, Palumbo O. Gonadal Mosaicism for an ASH1L Intragenic Deletion Makes a Bridge Between MRD52 and 1q22 Microdeletion. Am J Med Genet A. 2025;197(4):e63960. doi:10.1002/ajmg.a.63960 PMID: 39655631.
34. Dulovic-Mahlow M, Trinh J, Kandaswamy KK, Braathen GJ, Di Donato N, Rahikkala E, et al. De Novo Variants in TAOK1 Cause Neurodevelopmental Disorders. Am J Hum Genet. 2019;105(1):213-220. doi:10.1016/j.ajhg.2019.05.005 PMID: 31230721. PMC6612514.
35. van Woerden GM, Bos M, de Konink C, Distel B, Avagliano Trezza R, Shur NE, et al. TAOK1 is associated with neurodevelopmental disorder and essential for neuronal maturation and cortical development. Hum Mutat. 2021;42(4):445-459. doi:10.1002/humu.24176 PMID: 33565190. PMC8248425.
36. Hunter JM, Massingham LJ, Manickam K, Bartholomew D, Williamson RK, Schwab JL, et al. Inherited and de novo variants extend the etiology of TAOK1-associated neurodevelopmental disorder. Cold Spring Harb Mol Case Stud. 2022;8(2):a006180. doi:10.1101/mcs.a006180 PMID: 35091509. PMC8958914.
37. Heiman P, Drewes S, Ghaloul-Gonzalez L. A familial case of CAMK2B mutation with variable expressivity. SAGE Open Med Case Rep. 2021;9:2050313X21990982. doi:10.1177/2050313X21990982. PMID: 33796307.
38. Sadybekov A, Tian C, Arnesano C, Katritch V, Herring BE. An autism spectrum disorder-related de novo mutation hotspot discovered in the GEF1 domain of Trio. Nat Commun. 2017;8(1):601. doi:10.1038/s41467-017-00472-0 PMID: 28928363. PMC5605661.
39. Carrizales MG, Boulton AD, Koleske AJ. Genetic variants linked to neurodevelopmental disorders within the β3-β4 loop of the TRIO PH2 domain release autoinhibition of GEF2 activity. J Biol Chem. 2025;301(8):110429. doi:10.1016/j.jbc.2025.110429 PMID: 40581118. PMC12309612.
40. Ishchenko Y, Jeng AT, Feng S, Nottoli T, Manriquez-Rodriguez C, Nguyen KK, et al. Heterozygosity for neurodevelopmental disorder-associated TRIO variants yields distinct deficits in behavior, neuronal development, and synaptic transmission in mice. Elife. 2025;13:RP103620. doi:10.7554/eLife.103620 PMID: 40488445. PMC12148328.
41. Hiraide T, Hayashi T, Yamoto K, Masunaga Y, Asahina M, Ogata T, et al. An atypical case of macrocephaly and severe intellectual disability associated with a missense variant in the guanine nucleotide exchange factor-1 domain of TRIO. Brain Dev. 2025;47(5):104405. doi:10.1016/j.braindev.2025.104405 PMID: 40712355.
42. He M, Du B, Chen G, Lyu Y, Guo H, Jia X, et al. Naa15 Haploinsufficiency and De Novo Missense Variants Associate With Neurodevelopmental Disorders and Interfere With Neurogenesis and Neuron Development. Autism Res. 2025;18(5):954-965. doi:10.1002/aur.3308 PMID: 39825710.
43. Liu H, Liu DT, Lan S, Yang Y, Huang J, Huang J, et al. ASH1L mutation caused seizures and intellectual disability in twin sisters. J Clin Neurosci. 2021;91:69-74. doi:10.1016/j.jocn.2021.06.038 PMID: 34373061.
44. Pulatov O, Nguyen W, Vega DA, Barros R. A novel de novo missense variant in ASH1L associated with mild autism spectrum disorder and an uneven cognitive profile: a case report. J Med Case Rep. 2025;19(1):616. doi:10.1186/s13256-025-05675-4 PMID: 41291913. PMC12649092.
45. Papendorp CM, Nolan E, Higashimori H, Jhanji M, Goicouria L, Pieplow C, et al. Mutations in ASH1L cause a neurodevelopmental disorder with sex differences in epilepsy and autism [preprint]. bioRxiv. 2025 (v2 9 Apr 2025). doi:10.1101/2025.02.21.639570.
46. Mutoh H, Aoto K, Fukuda A, Saitsu H. CaMKIIβ insufficiency disrupts cortical networks, producing aberrant low-gamma oscillations and seizure susceptibility. Epilepsia. 2026;67(7):3766-3778. doi:10.1002/epi.70250 PMID: 41984532. PMC13361058.
47. Wang J, Li W, Li Z, Xue Z, Zhang Y, Yuan Y, et al. Taok1 haploinsufficiency leads to autistic-like behaviors in mice via the dorsal raphe nucleus. Cell Rep. 2023;42(9):113078. doi:10.1016/j.celrep.2023.113078 PMID: 37656623.
48. Freeman PJ, Hart RK, Gretton LJ, Brookes AJ, Dalgleish R. VariantValidator: accurate validation, mapping, and formatting of sequence variation descriptions. Hum Mutat. 2018;39(1):61-68. doi:10.1002/humu.23348. PMID: 28967166.
49. Dwyer BK, Veenma DCM, Chang K, Schulman H, Van Woerden GM. Case Report: Developmental Delay and Acute Neuropsychiatric Episodes Associated With a de novo Mutation in the CAMK2B Gene (c.328G>A p.Glu110Lys). Front Pharmacol. 2022;13:794008. doi:10.3389/fphar.2022.794008 PMID: 35620293. PMC9127182.
50. Blavier G, Lecoquierre F, Guerrot AM, Hélas GJ, Rondeau S, Boland A, et al. Revealing the impact of partial gene duplications in ASH1L: integration of optical genome mapping and RNA sequencing. Mol Cytogenet. 2025;19(1):2. doi:10.1186/s13039-025-00740-5 PMID: 41327308. PMC12777024.
51. Yubero D, Martorell L, Nunes T, Lyon GJ, Ortigoza-Escobar JD. Neurodevelopmental Gene-Related Dystonia: A Pediatric Case with NAA15 Variant. Mov Disord. 2022;37(11):2320-2321. doi:10.1002/mds.29241 PMID: 36221186. PMC10266335.
52. Straka I, Švantnerová J, Minár M, Stanková S, Zech M. Neurodevelopmental Gene-Related Dystonia-Parkinsonism with Onset in Adults: A Case with NAA15 Variant. Mov Disord. 2022;37(9):1955-1957. doi:10.1002/mds.29125 PMID: 35730864.
53. Yang K, Montomoli M, Rong J, Tam A, Ferrer Socorro M, Miller C, et al. Expanding the Motor Phenotype and Molecular Spectrum of NAA15-Related Disorder: Two Cases of Isolated Childhood-Onset Gait Dystonia. Mov Disord. 2026. doi:10.1002/mds.70447 PMID: 42549786.
54. Horii Y, Kuroda Y, Saito Y, Enomoto Y, Naruto T, Kurosawa K. A CAMK2B variant associated with tetralogy of Fallot, developmental delay, and growth retardation. Eur J Med Genet. 2023;66(10):104845. doi:10.1016/j.ejmg.2023.104845 PMID: 37734707.
55. Kloth K, Graul-Neumann L, Hermann K, Johannsen J, Bierhals T, Kortüm F. More evidence on TRIO missense mutations in the spectrin repeat domain causing severe developmental delay and recognizable facial dysmorphism with macrocephaly. Neurogenetics. 2021;22(3):221-224. doi:10.1007/s10048-021-00648-3. PMID: 34013494.
56. Kolbjer S, Martin DA, Pettersson M, Dahlin M, Anderlid BM. Lissencephaly in an epilepsy cohort: Molecular, radiological and clinical aspects. Eur J Paediatr Neurol. 2021;30:71-81. doi:10.1016/j.ejpn.2020.12.011. PMID: 33453472.
57. Schultz-Rogers L, Muthusamy K, Pinto E Vairo F, Klee EW, Lanpher B. Novel loss-of-function variants in TRIO are associated with neurodevelopmental disorder: case report. BMC Med Genet. 2020;21(1):219. doi:10.1186/s12881-020-01159-y. PMID: 33167890.
58. Gazdagh G, Hunt D, Gonzalez AMC, Rodriguez MP, Chaudhry A, Madruga M, et al. Extending the phenotypes associated with TRIO gene variants in a cohort of 25 patients and review of the literature. Am J Med Genet A. 2023;191(7):1722-1740. doi:10.1002/ajmg.a.63194 PMID: 36987741.
59. Patel R, Makwana R, Christ C, Marchi E, Miyake CY, Goncalves FG, et al. Neuroanatomical features of NAA10 and NAA15-related neurodevelopmental syndromes. J Neuroradiol. 2025;52(4):101339. doi:10.1016/j.neurad.2025.101339. PMID: 40204117.
60. Patel R, Park AY, Marchi E, Gropman AL, Whitehead MT, Lyon GJ. Ophthalmic manifestations of NAA10-related and NAA15-related neurodevelopmental syndromes: Analysis of cortical visual impairment and refractive errors. Am J Med Genet A. 2024;194(12):e63821. doi:10.1002/ajmg.a.63821 PMID: 39012200.
61. Braga BL, da Cunha Scalco R, Homma TK, Freire BL, Cellin LP, Canton APM, et al. Rare Causes and Differential Diagnosis in Patients With Silver-Russell Syndrome. Clin Genet. 2025;107(4):441-445. doi:10.1111/cge.14659. PMID: 39586716.
62. Khachatrian RG. [Evolution of DEE-SWAS into photosensitive epilepsy in a patient with an ASH1L gene mutation]. Zh Nevrol Psikhiatr Im S S Korsakova. 2025;125(10. Vyp. 2):114-119. doi:10.17116/jnevro2025125102114 PMID: 41283842.
63. Liao B, Xie W, He S. Novel heterozygous ASH1L nonsense variant involved in mild intellectual disability. Front Neurol. 2025;16:1524532. doi:10.3389/fneur.2025.1524532 PMID: 39902220. PMC11788156.
64. Blaschek A, Sitzberger A, Brugger M, Graf E, Berutti R, Zech M, et al. TAOK1-related neurodevelopmental disorder: A new differential diagnosis for childhood-onset tremor? Parkinsonism Relat Disord. 2023;109:105320. doi:10.1016/j.parkreldis.2023.105320 PMID: 36841152.
65. Krygier M, Limanówka M, Pietruszka M, Chylińska M, Mazurkiewicz-Bełdzińska M, Zech M. TAOK1-related neurodevelopmental disorder: A new differential diagnosis for childhood-onset tremor! Parkinsonism Relat Disord. 2025;133:107323. doi:10.1016/j.parkreldis.2025.107323 PMID: 39956038.
66. Richards S, Aziz N, Bale S, Bick D, Das S, Gastier-Foster J, et al. Standards and guidelines for the interpretation of sequence variants: a joint consensus recommendation of the American College of Medical Genetics and Genomics and the Association for Molecular Pathology. Genet Med. 2015;17(5):405-24. doi:10.1038/gim.2015.30. PMID: 25741868.
67. 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.
68. Srivastava S, Love-Nichols JA, Dies KA, Ledbetter DH, Martin CL, Chung WK, et al. Meta-analysis and multidisciplinary consensus statement: exome sequencing is a first-tier clinical diagnostic test for individuals with neurodevelopmental disorders. Genet Med. 2019;21(11):2413-2421. doi:10.1038/s41436-019-0554-6 PMID: 31182824. PMC6831729.
69. Rodan LH, Stoler J, Chen E, Geleske T, Council on Genetics. Genetic Evaluation of the Child With Intellectual Disability or Global Developmental Delay: Clinical Report. Pediatrics. 2025;156(1):e2025072219. doi:10.1542/peds.2025-072219 PMID: 40545261.
70. Moeschler JB, Shevell M, Committee on Genetics. Comprehensive evaluation of the child with intellectual disability or global developmental delays. Pediatrics. 2014;134(3):e903-18. doi:10.1542/peds.2014-1839 PMID: 25157020. PMC9923626.
71. Schaefer GB, Mendelsohn NJ, Professional Practice and Guidelines Committee. Clinical genetics evaluation in identifying the etiology of autism spectrum disorders: 2013 guideline revisions. Genet Med. 2013;15(5):399-407. doi:10.1038/gim.2013.32 PMID: 23519317.
72. Erasmus Medical Center. CAMK2-related Synapthopathies Natural History Study [Internet]. ClinicalTrials.gov identifier: NCT07372833. Registered 28 Jan 2026 [cited 28 Oct 2026]. Available from: https://clinicaltrials.gov/study/NCT07372833
73. Al-Aamri M, Alshaqaq M, Al-Abdi SY. A Saudi Girl With Co-occurring CHD1 (Pilarowski-Bjornsson Syndrome) and ASH1L Gene Variants. Cureus. 2023;15(12):e49905. doi:10.7759/cureus.49905. PMID: 38174187.
74. Simons Searchlight. Online Study of People Who Have Genetic Changes and Features of Autism: Simons Searchlight [Internet]. ClinicalTrials.gov identifier: NCT01238250. Updated 23 Jul 2026 [cited 28 Oct 2026]. Available from: https://clinicaltrials.gov/study/NCT01238250
75. Borghi R, Trivisano M, Specchio N, Tartaglia M, Compagnucci C. Understanding the pathogenetic mechanisms underlying altered neuronal function associated with CAMK2B mutations. Neurosci Biobehav Rev. 2023;152:105299. doi:10.1016/j.neubiorev.2023.105299. PMID: 37391113.
76. Hyman SL, Levy SE, Myers SM, COUNCIL ON CHILDREN WITH DISABILITIES, SECTION ON DEVELOPMENTAL AND BEHAVIORAL PEDIATRICS. Identification, Evaluation, and Management of Children With Autism Spectrum Disorder. Pediatrics. 2020;145(1):e20193447. doi:10.1542/peds.2019-3447 PMID: 31843864.
77. Risperidone tablets — prescribing information (generic ANDA201003; reference label text Risperdal®) [Internet]. DailyMed set id 03d3b52d-3a7d-4595-84c9-332da6ca9171, version 25 Jun 2026 [cited 28 Oct 2026]. Available from: https://dailymed.nlm.nih.gov/dailymed/lookup.cfm?setid=03d3b52d-3a7d-4595-84c9-332da6ca9171
78. Otsuka America Pharmaceutical. ABILIFY (aripiprazole) tablets — prescribing information [Internet]. NDA021436. DailyMed set id c040bd1d-45b7-49f2-93ea-aed7220b30ac, version 29 Jan 2025 [cited 28 Oct 2026]. Available from: https://dailymed.nlm.nih.gov/dailymed/lookup.cfm?setid=c040bd1d-45b7-49f2-93ea-aed7220b30ac
79. Sandoz Inc. RITALIN (methylphenidate hydrochloride) tablets — prescribing information [Internet]. NDA010187. DailyMed set id d6fb2750-cdab-4749-ba0d-7534840a5892, version 5 Feb 2025 [cited 28 Oct 2026]. Available from: https://dailymed.nlm.nih.gov/dailymed/lookup.cfm?setid=d6fb2750-cdab-4749-ba0d-7534840a5892
80. European Medicines Agency. Slenyto (melatonin) prolonged-release tablets — EPAR product information [Internet]. Amsterdam: EMA [cited 28 Oct 2026]. Available from: https://www.ema.europa.eu/en/documents/product-information/slenyto-epar-product-information_en.pdf
81. European Medicines Agency. Risperdal — Article 30 referral: Annex I–IV (harmonised summary of product characteristics, sections 4.1 and 4.2); CHMP opinion of 24 Jul 2008 and European Commission decision of 7 Oct 2008. Available from: https://www.ema.europa.eu/en/medicines/human/referrals/risperdal
82. Christ C, Makwana R, Appah FSK, Cheng H, Moreno C, Marchi E, et al. Letter: Psychosis Following Hormonal Therapy in NAA15-Related Neurodevelopmental Disorder. J Child Adolesc Psychopharmacol. 2026;36(7):454-456. doi:10.1177/10445463261462385 PMID: 42316871.
83. Schultz DC, Parr LC, Sweet H, Lamb S, Engers JL, Napier NC, et al. Discovery of VU6083859, a TAOK1 Selective Inhibitor, and VU6080195, a pan-TAOK Activator. ACS Chem Neurosci. 2026;17(3):624-634. doi:10.1021/acschemneuro.5c00906 PMID: 41530942. PMC12879741.
84. Liu Y, Sun L, Zhang H, Ma Y, Li M, Li J, et al. NAA10 and NAA15 mutations in cardiovascular disorders: Recent advances and mechanistic insights. Ageing Res Rev. 2026;122:103364. doi:10.1016/j.arr.2026.103364 PMID: 42749089.
85. Ommen SR, Ho CY, Asif IM, Balaji S, Burke MA, Day SM, et al. 2024 AHA/ACC/AMSSM/HRS/PACES/SCMR Guideline for the Management of Hypertrophic Cardiomyopathy: A Report of the American Heart Association/American College of Cardiology Joint Committee on Clinical Practice Guidelines. Circulation. 2024;149(23):e1239-e1311. doi:10.1161/CIR.0000000000001250 PMID: 38718139.
86. 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.
87. ERN-ITHACA. European Reference Network on Rare Congenital Malformations and Rare Intellectual Disability [Internet] [cited 28 Oct 2026]. Available from: https://ern-ithaca.eu/
88. CAMK2 Therapeutics Network [Internet] [cited 28 Oct 2026]. Available from: https://camk2.org/
89. ASH1L Syndrome & Gene: Family, Clinical & Research Hub [Internet] [cited 28 Oct 2026]. Available from: https://ash1l.org/