CDH2 mutation affecting N-cadherin function causes attention-deficit hyperactivity disorder in humans and mice

Attention-deficit hyperactivity disorder (ADHD) is one of the most common childhood-onset neuropsychiatric conditions, characterized by a persistent pattern of inattention, impulsivity, and hyperactivity, with complications often continuing into adulthood. Affected individuals have difficulties in higher-level executive functions, mediated by late-developing frontal-striatal-parietal and frontal-cerebellar neuronal networks, including inhibition, working memory, sustained attention, and temporal processing. Although its etiology is not well defined, ADHD has substantial heritability, and rare monogenic variants may play an essential role in its pathogenesis. The authors describe three siblings from a consanguineous kindred presenting with severe ADHD from early childhood. Through linkage analysis, whole-exome sequencing and biochemical studies, they identified a homozygous missense mutation in CDH2 affecting proteolysis and maturation of N-cadherin, a protein critical for synaptogenesis and neurite outgrowth, with known roles in regulating dopaminergic progenitors in ventral midbrain and prefrontal cortex. Knock-in mice bearing the human mutation exhibited hyperactivity and sensorimotor integration deficits, implicating CDH2-related pathways in ADHD pathophysiology.

Cell adhesion molecules, especially cadherins, are key mediators of neuronal development, synapse formation and plasticity. Prior genome-wide association studies have linked several cadherin genes (e.g., CDH9/10, CDH5, CDH11, CDH8, CDH23, CDH12/18, CDH13, CDH18, CDH28, CDH7) to psychiatric disorders including autism spectrum disorder, schizophrenia, depression and bipolar disorder. CDH13 has been associated with ADHD in GWAS without causal or functional proof. N-cadherin (CDH2) is essential for brain morphogenesis and synaptogenesis; its prodomain cleavage in the Golgi is critical for maturation, and aberrant cleavage impairs adhesive properties and synapse formation. N-cadherin also regulates presynaptic function and vesicle clustering via trans-synaptic mechanisms and interacts with synaptic organizers (neuroligin1, LRRTM2, Cadm1). Loss of N-cadherin has been associated with reduced dopaminergic neurons via Wnt/β-catenin signaling, a pathway implicated in dopamine neurogenesis and synaptic assembly. Despite these roles, cadherins had not been causally implicated in ADHD pathophysiology prior to this study.

Human studies: A consanguineous Bedouin family with three siblings diagnosed with severe ADHD was clinically characterized using interviews, observations, standardized questionnaires (CELF-5, Conners Parent Rating), and cognitive testing (WISC-III), excluding other neuropsychiatric comorbidities. Linkage analysis in nine family members identified a single homozygous locus on chromosome 18 (~11 Mb). Whole-exome sequencing (WES) of an affected individual was filtered against public databases and an in-house control set, focusing on autosomal recessive segregation. Sanger sequencing validated the candidate variant and segregation; 400 ethnically matched controls were screened. Protein analyses: In silico modeling (SWISS-MODEL) of N-cadherin and peptide docking to furin (HPEPDOCK) assessed the impact of p.H150Y on prodomain cleavage. Biochemical cleavage assays used synthesized FITC- and biotin-tagged 22-aa peptides (WT and mutant) encompassing the furin recognition motif; peptides were incubated with recombinant furin, reactions stopped at multiple time points, and products quantified by LC-MS to determine cleavage efficacy. Mouse model: CRISPR/Cas9 was used to generate two founder knock-in mouse lines harboring the Cdh2 p.H150Y mutation; founders were bred to establish homozygous mutant and WT lines. Behavioral testing: Male KI and WT mice underwent a battery assessing motor activity and anxiety (open-field test, elevated plus maze, rotarod), cognition (Y-maze spontaneous alternation), sensorimotor gating (acoustic startle reflex and pre-pulse inhibition), and social behavior (resident-intruder, three-chamber sociability and social novelty). Methylphenidate intervention: WT and KI mice received intraperitoneal methylphenidate (10 mg/kg) or vehicle 30 min prior to OFT and ASR; outcomes were analyzed as fold changes versus vehicle. Neuronal and synaptic assays: Primary dense hippocampal neuron cultures (DIV 8 and 14) from WT and KI littermates were immunostained for presynaptic markers (Synaptobrevin2/Syb2; vGlut1). Synaptic puncta intensity and distribution were quantified, including Full Width at Half Maximum (FWHM) analysis. FM1-43 dye loading/unloading quantified the readily releasable pool (RRP) and recycling pool (RcP). Synaptic vesicle cycling was measured using synaptopHluorin (sypHy), including alkaline trapping with bafilomycin A and NH4Cl unquenching to determine RcP fraction and kinetics of exocytosis/endocytosis. Electrophysiology: In acute hippocampal slices, whole-cell recordings of CA1 pyramidal neurons quantified miniature EPSCs (mEPSCs) in TTX and bicuculline; extracellular field recordings measured short-term synaptic facilitation (P2/P1 ratio) in the Schaffer collateral-CA1 pathway at 5–20 Hz. Presynaptic calcium dynamics were assessed using SypI-GCaMP6f during field stimulation. Molecular assays: Micro-dissected ventral midbrain (vMB) and prefrontal cortex (PFC) tissues underwent RT-qPCR for tyrosine hydroxylase (TH) and dopamine transporter (DAT), normalized to 18S rRNA. Tyrosine hydroxylase immunofluorescence quantified TH-positive neuron density in vMB sections. Dopamine concentrations in PFC and vMB were measured by high-sensitivity ELISA. Transcriptomics: RNA-seq on PFC and vMB (WT vs KI) identified differentially expressed genes (DEGs) using DESeq2 (adjusted p ≤ 0.05, |FC| ≥ 1.3), followed by Gene Ontology enrichment using DAVID. Statistics: Data are mean ± SEM with appropriate two-sided Student’s t-tests or nonparametric tests; mixed-effects models for repeated measures and TH quantification; significance threshold p < 0.05 unless noted.

- Human genetics: An ~11 Mb homozygous locus on chromosome 18 was shared by affected siblings. WES identified a single homozygous variant in CDH2 (NM_001792.4: c.355C>T; p.H150Y) within this locus, segregating recessively; absent from gnomAD; only 15 CDH2 LoF variants reported, none homozygous. One carrier found among 400 matched controls; the affected residue is highly conserved. - Protein modeling and biochemistry: The p.H150Y mutation lies near the furin protease consensus cleavage site (RXK/R-R) of the N-cadherin prodomain. Docking predicted impaired anchoring of the mutant peptide in the furin catalytic pocket. LC-MS cleavage assays showed substantially decreased proteolysis of the mutant peptide versus WT; after 180 min, cleavage was >20% weaker in mutant (t = 180 min p = 0.052). - Mouse behavior (founder line 1, 10-week males, n = 9/group): In open-field, mutants showed increased distance traveled (p = 0.04), velocity (p = 0.04), and mobility time (p = 0.03) with no change in center time or center/border crossings; no rotarod differences. Y-maze alternation showed a trend to lower performance (p = 0.06). ASR startle amplitude was higher in mutants (p = 0.03); similar pattern in PPI without change in PPI fraction. No differences in EPM; social tests were mixed (shorter resident-intruder interaction time, p = 0.03; no differences in three-chamber sociability/novelty). - Mouse behavior (founder line 2, 12-week males, n = 15/group): Replicated hyperactivity with larger effects: distance (p = 2E-9), velocity (p = 2E-9), mobility and cumulative movement duration (p = 1E-7 and 1E-8), increased center/border crossings (p = 1E-4); no difference in center time. - Methylphenidate (14-week males, n = 39): MPH (10 mg/kg i.p.) increased locomotion in both genotypes, but fold-changes were greater in mutants for velocity and distance (p = 0.0032) and mobility continuous (p = 0.0174); mutants showed a smaller fold-increase in center time (p = 0.034). In ASR, MPH-treated mutants had significantly lower startle amplitude than WT (p = 0.012). - Synaptic structure and vesicle pools: In hippocampal cultures, presynaptic SV cluster intensity was reduced in mutants: Syb2 at DIV14 (p = 0.007) and DIV8 (p = 0.01), vGlut1 at DIV14 (p = 3E-4) with no difference at DIV8; Syb2 puncta were narrower (FWHM p = 0.002), indicating smaller, more compact clusters. FM1-43 destaining showed a smaller RRP fraction in mutants (p = 0.017) with no change during prolonged destaining (RcP unaffected). - Synaptic function: SypHy imaging revealed reduced stimulus-evoked SV recycling in mutants (peak ΔF/F0 p = 0.007). Alkaline trapping showed similar RcP fractions (~50%, ns p = 0.78). No differences in exocytosis or endocytosis kinetics. mEPSC recordings showed reduced frequency (inter-event intervals increased, p = 0.0085) with unchanged amplitudes (ns p = 0.99), indicating presynaptic deficits. Field recordings demonstrated reduced frequency facilitation in mutants at 10 Hz (p = 0.032) and 20 Hz (p = 0.02); presynaptic calcium imaging with SypI-GCaMP6f showed no differences in basal fluorescence or peak responses (ns). - Dopaminergic markers and dopamine levels: TH mRNA was reduced in PFC (p = 0.03) with a trend in vMB (ns p = 0.09); DAT mRNA unchanged (ns). TH immunofluorescence in vMB showed fewer TH-positive neurons in mutants (one-tailed, p = 0.0435). ELISA showed decreased dopamine concentration in PFC (p = 0.013), with no significant change in vMB (ns p = 0.27). - Transcriptomics: RNA-seq identified 181 DEGs in PFC (99 down, 82 up) and 604 in vMB (383 down, 221 up). CDH2 transcripts were down in PFC (p = 2.7E-04, FC −1.49). Downregulated genes included CTNND1 (p120-catenin, FC −1.31) and IQGAP1 (FC −1.39); DDC (DOPA decarboxylase) was upregulated (FC +1.63), suggesting compensatory dopamine synthesis. GO terms were enriched for synaptic processes, neuronal projection, cell-cell adhesion, axon guidance, and glutamatergic/GABAergic synapses; multiple ADHD-linked genes were downregulated.

This study addresses whether a monogenic variant can causally contribute to ADHD and delineates a mechanistic pathway. Identification of a homozygous CDH2 p.H150Y mutation in a family with severe ADHD, coupled with functional evidence that the mutation disrupts N-cadherin prodomain cleavage by furin, supports a pathogenic effect on protein maturation and adhesion. Knock-in mice recapitulated hyperactivity and sensorimotor phenotypes modifiable by methylphenidate, aligning with clinical responsiveness patterns in ADHD. At the synaptic level, mutants exhibited smaller, more compact presynaptic vesicle clusters, a reduced readily releasable pool, attenuated evoked release, decreased spontaneous release frequency, and diminished short-term facilitation, consistent with weakened presynaptic function. Molecular changes included reduced TH expression and dopamine levels in the mesocortical pathway and transcriptomic alterations in adhesion and synaptic pathways (including p120-catenin and IQGAP1), suggesting that impaired N-cadherin-mediated adhesion and Wnt/β-catenin-related signaling converge to reduce dopaminergic tone. These findings link CDH2 dysfunction to the hypo-dopaminergic hypothesis of ADHD and highlight adhesion-dependent synaptic organization as a critical upstream determinant of neurotransmission relevant to ADHD pathophysiology.

A homozygous missense mutation in CDH2 (p.H150Y) causes a familial, non-syndromic ADHD phenotype. The mutation impairs N-cadherin maturation, leading to deficits in presynaptic vesicle clustering and release, altered short-term synaptic plasticity, reduced tyrosine hydroxylase expression, and decreased dopamine in the prefrontal cortex. CRISPR/Cas9 knock-in mice carrying the human mutation exhibit hyperactivity and altered responses to methylphenidate. Whole-transcriptome analyses reveal coordinated changes in adhesion and synaptic genes. Together, these results establish CDH2-related pathways as contributors to ADHD pathophysiology. Future work should dissect the precise molecular mechanisms linking N-cadherin processing to Wnt/β-catenin signaling, dopaminergic development and function, evaluate chronic therapeutic responses, include female cohorts, and expand genetic studies to assess the prevalence and spectrum of CDH2 variants in ADHD.

Human evidence derives from a single consanguineous family with three affected siblings, limiting generalizability. The peptide cleavage assay showed a trend-level difference at the longest time point (p = 0.052). Behavioral outcomes had some mixed or non-significant domains (e.g., Y-maze trend, social tests). Most behavioral assessments used male mice, limiting sex-specific insights. Acute methylphenidate administration may not reflect chronic therapeutic effects. Calcium imaging did not resolve nanodomain calcium dynamics that may influence facilitation. Mechanistic links to Wnt/β-catenin signaling and dopaminergic development require further validation.