Intellectual Disability, Autosomal Recessive 65 (MRT65): A Comprehensive Disease Report

Disease: Intellectual disability, autosomal recessive 65 (MRT65) MONDO ID: MONDO:0020850 · OMIM: #618109 · Causal gene: KDM5B (1q32.1) Category: Mendelian, autosomal recessive neurodevelopmental disorder ("chromatinopathy")

Summary

Intellectual disability, autosomal recessive 65 (MRT65) is an ultra-rare, autosomal-recessive neurodevelopmental disorder caused by biallelic (homozygous or compound-heterozygous) loss-of-function variants in KDM5B, a gene on chromosome 1q32.1 encoding a histone H3K4 (tri-/di-/mono-methyl) demethylase. Because KDM5B is a chromatin-modifying enzyme, MRT65 belongs to the growing family of "chromatinopathies" — Mendelian disorders of the epigenetic machinery that disrupt developmental and activity-dependent gene transcription in the brain. The disorder was first delineated by Faundes and colleagues in 2018, who described "a recessive histone lysine-methylation defect caused by homozygous or compound heterozygous KDM5B variants and resulting in a recognizable syndrome with developmental delay, facial dysmorphism, and camptodactyly" (PMID: 29276005).

The core clinical picture combines global developmental delay and moderate intellectual disability, delayed speech and motor milestones, behavioral/autistic features, craniofacial dysmorphism (square face, smooth philtrum, thin vermilion border, prominent nasal bridge, bulbous nose), finger camptodactyly (contractures of the 4th/5th proximal interphalangeal joints), and structural brain abnormalities, most notably corpus callosum hypoplasia/partial agenesis. Onset is neonatal to infantile. Importantly, the same gene shows a striking dosage relationship: KDM5B is loss-of-function–tolerant in the general population (gnomAD pLI ≈ 0, LOEUF 0.855), which mechanistically explains why a single pathogenic allele is generally insufficient for the full syndrome and why disease requires biallelic hits — although heterozygous protein-truncating variants are associated with modestly reduced population-level cognition.

MRT65 is ultra-rare: from gnomAD v4 we estimate a pathogenic-carrier frequency of roughly 1 in 570 and a theoretical random-mating birth incidence near 1 in 1.3 million (higher in consanguineous populations). There is no disease-specific or curative therapy; management is entirely supportive (early developmental intervention, physical/occupational/speech therapy, and management of dysmorphism-related and systemic complications). Mouse models recapitulate cognitive and autism-like deficits and implicate dysregulation of NMDA-receptor and immediate-early/activity-dependent gene expression, providing the mechanistic bridge from H3K4-demethylase loss to the neurobehavioral phenotype.

Key Findings

Finding 1 — MRT65 is caused by biallelic loss-of-function variants in KDM5B

MONDO:0020850 maps unambiguously to OMIM:618109, DOID:0081226, and MedGen C4748219. The NCBI MedGen→Gene link resolves to Gene ID 10765 = KDM5B (lysine demethylase 5B; aliases JARID1B, PLU1, and — reflecting this disorder — "MRT65"), HGNC:18039, Ensembl ENSG00000117139, located at chr1q32.1 (GRCh38 chr1:202,724,495–202,808,487). The disorder was first described by Faundes et al. 2018 (Am J Hum Genet), who reported "a recessive histone lysine-methylation defect caused by homozygous or compound heterozygous KDM5B variants and resulting in a recognizable syndrome with developmental delay, facial dysmorphism, and camptodactyly" (PMID: 29276005). A subsequent genotype/phenotype study explicitly confirmed that "bi-allelic disruptive variants (nonsense, frameshift, and splicing variants) in KDM5B have been identified as causative for autosomal recessive intellectual developmental disorder type 65" (PMID: 39202393).

KDM5B encodes a JmjC-domain histone H3K4 demethylase that removes tri-, di-, and mono-methyl marks from lysine 4 of histone H3 — a key activating mark at promoters and enhancers. Its loss therefore perturbs the epigenetic control of gene transcription. Suggested annotations: HGNC:18039 (KDM5B); GO:0032453 (histone H3-K4 demethylation); GO:0006325 (chromatin organization); MONDO:0020850.

Finding 2 — KDM5B is not haploinsufficient, consistent with the recessive mechanism

Population constraint data explain the recessive inheritance. gnomAD constraint for KDM5B (ENSG00000117139) shows pLI ≈ 0 (7.9×10⁻³⁶), an observed/expected LoF ratio of 0.74 (90% CI 0.65–0.86; LOEUF 0.855), and missense Z = 0.53. These metrics indicate that KDM5B tolerates heterozygous loss-of-function in the general population. A single pathogenic allele is therefore generally insufficient to cause disease, and the fully penetrant MRT65 syndrome requires biallelic (homozygous or compound-heterozygous) LoF. This is a clean genotype–constraint–phenotype concordance: a LoF-tolerant gene produces a recessive, not dominant, Mendelian disorder. Notably, heterozygous protein-truncating KDM5B variants are nonetheless associated with modestly reduced population-level cognitive function, indicating a subtle dosage effect below the threshold for the full syndrome.

Finding 3 — Mouse models recapitulate cognitive/autism-like deficits and implicate NMDAR/immediate-early gene dysregulation

Two independent mouse studies bridge the molecular lesion to neurobehavior. Kdm5b^ΔARID/ΔARID mice (lacking demethylase activity) show hyperactivity and hippocampus-dependent long-term memory deficits, with downregulated baseline and hyperactivated post-learning immediate-early/activity-dependent gene expression (Pérez-Sisqués et al., J Neurosci 2024, PMID: 38575342). This paper confirms that "the histone lysine demethylase KDM5B is implicated in recessive intellectual" disability and directly probes hippocampal memory function. A companion study reports that KDM5B-deficient mice display autism-like phenotypes with increased NMDAR2D (Grin3b/GRIN2D) expression (Pérez-Sisqués et al., Sci Adv 2026, 42160407). Together these provide a mechanistic link from H3K4me3 demethylase loss → impaired transcriptional control of synaptic/learning ("plasticity") genes → cognitive and autism-like deficits.

Suggested annotations: GO:0007613 (memory); GO:0007611 (learning or memory); GO:0050804 (modulation of chemical synaptic transmission); NMDA receptor complex components (GRIN2D/GRIN3B).

Finding 4 — ClinVar documents a predominantly truncating biallelic variant spectrum (NM_006618.5)

ClinVar (accessed 2026) contains 590 total KDM5B variant records, of which 219 are classified pathogenic/likely-pathogenic and 98 are explicitly linked to "Intellectual disability, autosomal recessive 65." The pathogenic spectrum is dominated by protein-truncating variants. Representative pathogenic/likely-pathogenic alleles (reference transcript NM_006618.5) include:

Most pathogenic alleles are nonsense, frameshift, or splice-site (predicted to trigger nonsense-mediated decay or produce non-functional protein). Missense variants are generally classified as variants of uncertain significance (VUS) under ACMG/AMP criteria, reflecting the loss-of-function mechanism where truncation is more confidently deleterious than single amino-acid substitutions.

Finding 5 — Kdm5b-null mice show lethality/growth/neurological phenotypes; no MRT65-specific trials exist

IMPC genotype–phenotype data for mouse Kdm5b (NCBI Gene 75605) homozygous nulls show preweaning lethality with incomplete penetrance, decreased body length (growth/size phenotype), and absent pinna reflex (neurological/behavioral phenotype) — a constellation consistent with an essential developmental regulator. A ClinicalTrials.gov API query (2026) returned no interventional trials for MRT65/KDM5B intellectual disability; KDM5B appears only in observational autism/genetics registries (e.g., Simons SPARK, NCT01238250). Small-molecule KDM5B inhibitors exist but only in oncology contexts (e.g., 42324589) and are mechanistically inappropriate for a loss-of-function disorder — inhibiting an already-lost enzyme cannot restore function. Consequently, MRT65 management remains entirely supportive.

Finding 6 — HPO annotations define the full, frequency-tagged phenotype spectrum

The HPO disease annotation for OMIM:618109 (gene KDM5B/NCBIGene:10765; autosomal recessive; source 29276005) comprises 37 terms, with frequencies expressed as n/3 biallelic individuals from the original cohort. The phenotype spectrum:

Onset: neonatal in 2/3 and infantile in 1/3.

Finding 7 — Carrier frequency ~1 in 570; theoretical birth incidence ~1 in 1.3 million

Using gnomAD v4 (8,525 KDM5B variants), high-confidence (LOFTEE HC) rare (allele frequency <0.1%) predicted loss-of-function alleles across 411 sites sum to a pathogenic allele frequency q ≈ 8.8×10⁻⁴. This gives:

• Carrier frequency 2q ≈ 0.0018 (≈ 1 in 567)
• Theoretical random-mating homozygous/compound-heterozygous birth incidence q² ≈ 7.8×10⁻⁷ (≈ 1 in 1.3 million)

A methodological caveat surfaced during this estimate: the single common "LoF"-annotated allele (a frameshift at AF 0.185) is LOFTEE low-confidence (LC) — i.e., not a true LoF. Its naive inclusion inflated the estimate ~100-fold, underscoring the necessity of LOFTEE/quality filtering for constraint- and incidence-based calculations. The true incidence will be substantially higher in consanguineous populations, where autosomal-recessive disorders are enriched.

Mechanistic Model / Interpretation

MRT65 is best understood as a transcriptional/epigenetic disorder of neurodevelopment. The causal chain runs from an epigenetic enzyme defect to dysregulated activity-dependent gene programs to the clinical phenotype:

1.  Biallelic LoF variants in KDM5B (nonsense / frameshift / splice)
        │  leads to
        ▼
2.  Nonsense-mediated decay / truncated non-functional protein
        │  results in
        ▼
3.  Loss of H3K4me3/me2/me1 demethylase activity (JmjC + ARID domains)
        │  leads to
        ▼
4.  Aberrant H3K4-methylation landscape at promoters/enhancers
        │  results in
        ▼
5.  Dysregulated developmental + activity-dependent transcription
        │  (immediate-early genes down at baseline, hyperactivated post-learning;
        │   NMDAR subunit GRIN2D/GRIN3B up — shown in mouse)   [model organism]
        ├─────────────► Brain: impaired synaptic plasticity / hippocampal memory
        │                        └► intellectual disability, developmental delay,
        │                           autistic/behavioral features, gait ataxia
        ├─────────────► Neuroanatomy: corpus callosum hypoplasia / partial agenesis
        ├─────────────► Craniofacial morphogenesis → dysmorphism
        └─────────────► Limb/digit development → camptodactyly (4th/5th PIP contractures)

Upstream vs downstream. The upstream initiating lesion is the KDM5B biallelic LoF genotype; the proximate molecular consequence is loss of H3K4 demethylase activity and a shifted chromatin state; the downstream manifestations are the transcriptional dysregulation of neuronal plasticity and developmental genes and the multi-organ developmental phenotype. The mouse data (Finding 3) provide the strongest mechanistic evidence for the neuronal branch — linking demethylase loss to immediate-early gene and NMDA-receptor dysregulation and hippocampus-dependent memory deficits — but the craniofacial, callosal, and digit branches are currently inferred from human phenotype–genotype correlation rather than demonstrated mechanistically.

Cell types and processes. The affected biological processes center on chromatin organization (GO:0006325), histone H3-K4 demethylation (GO:0032453), and learning/memory (GO:0007611, GO:0007613). Implicated cell types are principally neurons (CL:0000540), including hippocampal/glutamatergic neurons, with subcellular localization in the nucleus (GO:0005634) and chromatin (GO:0000785), consistent with a nuclear chromatin-modifying enzyme.

Why recessive. Finding 2 ties the genetics together: because KDM5B is LoF-tolerant (pLI ≈ 0, LOEUF 0.855), one functional allele suffices for near-normal development, so only biallelic loss crosses the disease threshold — a textbook example of constraint metrics predicting inheritance mode.

Section-by-Section Detail

1. Disease Information

MRT65 is an ultra-rare autosomal-recessive neurodevelopmental chromatinopathy. Identifiers: MONDO:0020850; OMIM #618109; DOID:0081226; MedGen C4748219. Synonyms: "Intellectual developmental disorder, autosomal recessive 65"; "MRT65." Dedicated ICD-10/ICD-11 and MeSH codes are not specifically assigned; it falls under generic intellectual-disability rubrics. Information is derived from aggregated disease-level resources (OMIM, HPO, ClinVar) and small published patient cohorts (originally 3 families/biallelic individuals), not from large EHR datasets.

2. Etiology

Primary cause: genetic — biallelic loss-of-function variants in KDM5B. Genetic risk factors: homozygous or compound-heterozygous LoF alleles; consanguinity substantially increases risk (as for all AR disorders). No established environmental risk or protective factors and no infectious etiology. Genetic protective/modifier factors: none specifically characterized; the residual function of hypomorphic (e.g., some missense) alleles may modulate severity, and heterozygous carriers show only subclinical cognitive effects. Gene–environment interactions: none documented.

3. Phenotypes

See Finding 6 table. Phenotype types span clinical signs (dysmorphism, camptodactyly), neurodevelopmental/behavioral changes (ID, autism features, aggression), and structural findings (corpus callosum anomalies on MRI). Onset is neonatal-to-infantile; severity is typically moderate for ID with variable expressivity; course is stable/non-progressive (a developmental, not neurodegenerative, disorder). Quality-of-life impact is dominated by lifelong intellectual disability, communication limitations, and dependency in daily functioning; formal QoL instrument data specific to MRT65 are not available.

4. Genetic / Molecular Information

Causal gene: KDM5B (HGNC:18039; gene MIM 605393; Ensembl ENSG00000117139). Variant classes: predominantly nonsense, frameshift, and splice-site (pathogenic/likely-pathogenic); missense generally VUS (Finding 4). Reference transcript: NM_006618.5. Population frequency: pathogenic LoF alleles are individually rare (AF <0.1%). Origin: germline (no somatic disease role). Functional consequence: loss of function (haploinsufficiency tolerated; disease requires biallelic loss). Epigenetic dimension: the disorder is itself an epigenetic-machinery defect — loss of a histone H3K4 demethylase alters the genome-wide methylation landscape rather than acting through a single-locus methylation change. Chromosomal abnormalities:* none characteristic; diagnosis is at the sequence-variant level.

5. Environmental Information

No environmental, lifestyle, or infectious contributors are implicated. MRT65 is a monogenic Mendelian disorder.

6. Mechanism / Pathophysiology

See the causal-chain diagram above and Findings 1–3. Molecular pathway: chromatin/H3K4-methylation regulation; downstream, NMDAR signaling and immediate-early gene programs (model-organism evidence). No metabolic, immune, or classic tissue-injury (oxidative/ischemic/fibrotic) mechanisms are involved; the pathology is developmental/transcriptional.

7. Anatomical Structures Affected

Primary organ/system: central nervous system (brain, UBERON:0000955; nervous system, UBERON:0001016), with corpus callosum (UBERON:0002336) hypoplasia/partial agenesis. Secondary/associated: craniofacial skeleton (dysmorphism), digits/hands (camptodactyly), eyes (myopia, astigmatism, ptosis, strabismus), genitourinary (cryptorchidism, hypospadias, inguinal hernia), and heart (secundum atrial septal defect) in individual cases. Cell type: neurons (CL:0000540). Subcellular: nucleus/chromatin (GO:0005634 / GO:0000785). Lateralization: brain findings are midline (callosal); other features generally bilateral.

8. Temporal Development

Onset: congenital/neonatal (2/3) to infantile (1/3). Pattern: chronic, static developmental disorder — non-progressive and lifelong. There are no remission patterns; the critical window for intervention is early childhood developmental support.

9. Inheritance and Population

Inheritance: autosomal recessive. Penetrance: high/complete for biallelic LoF (based on limited cohorts). Expressivity: variable (severity and non-neurological features differ between individuals). Carrier frequency: ≈ 1 in 570; theoretical birth incidence: ≈ 1 in 1.3 million under random mating (Finding 7), higher with consanguinity. No anticipation (not a repeat-expansion disorder). Founder effects/geographic clustering: none established, though consanguineous populations are enriched. Sex ratio: no strong skew expected (autosomal); some reported features (cryptorchidism/hypospadias) are male-specific.

10. Diagnostics

Genetic testing is diagnostic. Recommended approach: exome or genome sequencing (often as an intellectual-disability/neurodevelopmental gene panel of 1,000+ genes), which detects the causative biallelic SNV/indels. Long-read genome sequencing can add value for phasing biallelic variants in autosomal-recessive genes and resolving structural/complex variants (PMID: 41514368). Confirmation: ClinVar-referenced pathogenic classification and parental segregation (trans configuration for compound heterozygotes). Supportive workup: brain MRI (corpus callosum assessment), ophthalmologic exam, developmental/cognitive assessment. No specific biochemical biomarker exists. Differential diagnosis: other chromatinopathies and ID syndromes — notably KDM4B-related disorders (autosomal dominant IDD from heterozygous KDM4B LoF, and a reported biallelic KDM4B case) which share developmental delay, brain anomalies, and digital findings (PMID: 37526414); metabolic ID mimics should be excluded when biochemical clues are present.

11. Outcome / Prognosis

MRT65 is a non-lethal, non-progressive developmental disorder in humans (contrast with murine preweaning lethality, Finding 5). Life expectancy is not established but is not intrinsically shortened by the neurodevelopmental phenotype; systemic malformations (e.g., cardiac septal defect) may modify individual prognosis. Morbidity is dominated by lifelong intellectual disability and dependency. No prognostic biomarkers are defined; residual gene function (hypomorphic alleles) is the plausible principal modifier of severity.

12. Treatment

No disease-specific or curative therapy exists (Finding 5). Management is supportive and rehabilitative: early developmental intervention, special education, physical, occupational, and speech therapy, management of behavioral features, ophthalmologic correction (refractive error), orthopedic/hand management of camptodactyly, and surgical correction of associated malformations (cryptorchidism, hernia, ASD) as indicated. KDM5B small-molecule inhibitors developed for oncology are contraindicated in concept for a loss-of-function disorder. No pharmacogenomic or gene/RNA-based therapies are available. Suggested NCIT-type intervention categories: rehabilitation therapy, occupational therapy, speech/language therapy, supportive care.

13. Prevention

No primary prevention beyond genetic counseling and reproductive options for at-risk (e.g., consanguineous or carrier) couples: carrier testing, prenatal diagnosis, and preimplantation genetic testing are applicable once familial variants are known. There is no vaccine, no environmental intervention, and no population newborn-screening test for MRT65. Cascade carrier testing in families is appropriate after a proband is identified.

14. Other Species / Natural Disease

Orthologue: mouse Kdm5b (NCBI Gene 75605). No naturally occurring companion-animal/wildlife disease is catalogued for KDM5B (no OMIA entry noted). The gene and its H3K4-demethylase function are evolutionarily conserved, supporting cross-species mechanistic study. No zoonotic dimension (non-infectious disorder).

15. Model Organisms

The principal model is the mouse. Informative genetic models: (1) Kdm5b^ΔARID/ΔARID (demethylase-dead) mice recapitulate hyperactivity, hippocampal memory deficits, and activity-dependent/immediate-early gene dysregulation (PMID: 38575342); (2) KDM5B-deficient mice with autism-like phenotypes and increased NMDAR (GRIN2D/GRIN3B) expression (Sci Adv 2026, 42160407); and (3) IMPC Kdm5b-null mice with preweaning lethality (incomplete penetrance), reduced body length, and absent pinna reflex (Finding 5). Phenotype recapitulation: strong for the cognitive/behavioral axis. Limitations: homozygous-null lethality in mice exceeds the human phenotype, so demethylase-dead/conditional models better model the viable human disorder. Resources: MGI, IMPC.

Evidence Base

Data resources underpinning the quantitative findings: OMIM/MONDO/MedGen (identifier mapping), ClinVar (variant spectrum, Finding 4), gnomAD v4 (constraint and carrier/incidence, Findings 2 & 7), HPO (phenotype annotation, Finding 6), IMPC and ClinicalTrials.gov API (mouse phenotypes and trial landscape, Finding 5).

Limitations and Knowledge Gaps

1. Small human cohort. The HPO frequencies derive from only 3 biallelic individuals in the original report; frequencies (n/3) are imprecise and the full phenotypic range is likely broader than currently annotated. The 2024 genotype/phenotype study (39202393) expands this, but MRT65 remains defined by tens, not hundreds, of patients.
2. Mechanistic branches inferred, not proven. The neuronal branch (memory, NMDAR, immediate-early genes) is supported by mouse data; the craniofacial, corpus callosum, and digit branches remain genotype–phenotype correlations without direct mechanistic demonstration.
3. Genotype–phenotype correlation. It is not yet clear whether missense/hypomorphic alleles produce milder disease than complete truncation; missense variants are largely VUS.
4. Incidence estimate is theoretical. The 1-in-1.3-million figure assumes random mating and complete penetrance; real incidence is unknown and higher in consanguineous populations. No formal prevalence study exists.
5. Model organism discordance. Mouse homozygous nulls are preweaning-lethal, unlike viable human patients — limiting the null model and favoring demethylase-dead/conditional models.
6. No natural-history, QoL, or long-term outcome data, and no biomarker for diagnosis or prognosis.

Proposed Follow-up Experiments / Actions

1. Aggregate a larger patient registry (GeneMatcher/MatchmakerExchange) to refine phenotype frequencies, expressivity, and genotype–phenotype correlations (truncating vs missense).
2. Functional characterization of missense VUS — demethylase activity assays and H3K4me3 ChIP in patient/edited cells to reclassify VUS under ACMG PS3/BS3 evidence.
3. Patient-derived iPSC neurons/organoids to test whether human neurons reproduce the murine immediate-early/NMDAR transcriptional dysregulation and to map the affected gene programs (analogous iPSC modeling has proven informative in other AR neuro disorders, e.g., PMID: 39617394).
4. Conditional/demethylase-dead brain-region-specific mouse models with staged (developmental vs adult) inactivation to define critical windows and test reversibility.
5. Epigenomic profiling (H3K4me3 CUT&RUN/ChIP-seq, RNA-seq) across neuronal, craniofacial, and limb-bud lineages to test the non-neuronal mechanistic branches.
6. Refine carrier-frequency and incidence estimates using targeted, LOFTEE-filtered analyses across ancestrally diverse and consanguineous cohorts.
7. Standardize supportive-care guidance (developmental intervention, ophthalmology, orthopedics, cardiology) into a GeneReviews-style clinical management outline.

Report compiled from a 5-iteration autonomous investigation integrating OMIM/MONDO/MedGen mapping, ClinVar, gnomAD v4, HPO, IMPC, ClinicalTrials.gov, and primary literature. Evidence types are labeled human clinical, model organism, or computational throughout.