Research report: Hypertrophic Cardiomyopathy 12 (CMH12, CSRP3-related HCM)

Target: MONDO:0012804 · Mendelian · prepared 2026-10-13

How to read the evidence. Quotes marked [cached] are exact text from abstracts already in this worktree's references_cache/, so they can be used as snippets as they stand. References marked [not cached] were found by web search. They need just fetch-reference before anything from them is used as evidence, and I have not quoted them verbatim. Ontology CURIEs marked ✔ were checked against the committed cache/*/terms.csv. CURIEs without ✔ are leads and must be looked up before they are bound.

1. Disease information

Overview. CMH12 is familial hypertrophic cardiomyopathy caused by variants in CSRP3 (11p15.1). CSRP3 encodes muscle LIM protein (MLP), a small protein with two LIM domains. MLP sits at the Z-disc and costamere and also moves into the nucleus. Unlike most HCM genes, CSRP3 is non-sarcomeric. Its role is mechanosensing and signalling, not generating force.

"MLP, in contrast to the proteins encoded by the other recognised HCM disease genes, is non-sarcomeric, and has important signalling functions in cardiomyocytes." — PMID:30048712 [cached]

Identifiers | System | ID | |---|---| | MONDO | MONDO:0012804 hypertrophic cardiomyopathy 12 (already in the draft entry) | | OMIM (phenotype) | 612124 (CMH12; listed as AD) | | OMIM (gene) | 600824 CSRP3 | | HGNC | hgnc:2472 CSRP3 (taken from the ClinGen rows below; not yet checked in cache/hgnc) | | MedGen | 393755 | | LOVD disease | 03091 | | Orphanet | No CMH12-specific entry was found. The umbrella entry is "Familial isolated hypertrophic cardiomyopathy" (ORPHA:217569, not verified*; check with just fetch-reference ORPHA:…) | | ICD-10 / ICD-11 | Only the general HCM codes apply (I42.1 obstructive / I42.2 other HCM; ICD-11 BC43.1). Neither has been checked against the ICD10CM cache |

Synonyms: CMH12; cardiomyopathy, familial hypertrophic, 12; CSRP3-related hypertrophic cardiomyopathy; MLP-related HCM. Allelic disorder: dilated cardiomyopathy 1M (CMD1M, MONDO:0011840).

Data provenance: All of the data are disease-level aggregates (family cosegregation studies, case–control cohorts, and the ClinGen and OMIM curations). No EHR-derived data were found.

2. Etiology

Primary cause: Germline CSRP3 variants. Three kinds are reported: - Heterozygous missense variants in the LIM1 domain. Examples are C58G, L44P, S54R/E55G, K42Q and C150Y. - Homozygous or biallelic variants, including truncating variants that effectively knock the gene out. Examples are p.Arg122Ter and other truncations. - The W4R variant (p.Trp4Arg). It is common (about 1% in Europeans) and now looks like a modifier rather than a cause of disease on its own.

"We found 3 different missense mutations in 3 unrelated patients with familial HCM but detected no mutation in the DCM group or the controls. All mutations predicted an amino acid exchange at highly conserved residues in the functionally important LIM1 domain" — PMID:12642359 [cached]

Gene-disease validity: the classification moved between curations, and the mode of inheritance is disputed - ClinGen 2019 (Ingles et al., Circ Genom Precis Med) put CSRP3 in the Moderate group, together with TNNC1 and JPH2 [not cached; ClinGen HCM gene curation]. - The current ClinGen record is Definitive, with MOI recorded as SD (semidominant), from the Hereditary Cardiovascular Disease GCEP, dated 2023-08-09:

"CSRP3 | HGNC:2472 | hypertrophic cardiomyopathy | MONDO:0005045 | SD | Definitive" — CGGV:assertion_012b9b15-798a-46e2-8164-61a829ceeda1-2023-08-09T160000.000Z [cached] - Curation caveat: this ClinGen assertion is for the general HCM term MONDO:0005045, not MONDO:0012804. Under the CLAUDE.md rule ("The assertion must be for this entry's disease"), it is copied onto CMH12's gene_disease_validity only if CMH12 maps to that term with exactMatch. Otherwise it is an other_disease lead. - For the allelic DCM phenotype: "CSRP3 | HGNC:2472 | dilated cardiomyopathy 1M | MONDO:0011840 | AD | Limited" — CGGV:assertion_beaafc90-…-2026-03-04 [cached]. - Recessive inheritance was argued early on: "we suggest that the association of CSRP3 as a validated HCM-causing gene require additional studies and those CSRP3 variants could result in HCM with an autosomal recessive inheritance rather than with an autosomal dominant transmission as usually reported on HCM (OMIM 612124)." — PMID:30012424 [cached] - Population-scale association (2026): "Several definitive genes which are much less commonly affected in HCM (CSRP3, MYL3, ACTC1, TPM1, FHOD3, MYL2, and TNNC1) did not pass our Bonferroni corrected-significance threshold, but all had positively associated effect sizes with HCM." — PMID:41893039 [cached]

Risk factors - Genetic: Biallelic state, which consanguinity makes more likely (see section 9). Co-inheritance of a sarcomeric variant, which W4R makes worse (see Modifiers). - Age: Penetrance depends strongly on age (C150Y; PMID:33035702). - Environmental: No CSRP3-specific environmental risk factor has been reported. General HCM modifiers such as hypertension and obesity are plausible but unstudied for CMH12.

Protective factors: None reported. Mechanistically, lowering sarcomeric crossbridge load protects iPSC tissue (PMID:35384713, see section 6), but this is a treatment effect and not a protective factor in a population.

Gene–environment interactions: No human data. MLP's subcellular location responds to load and contractility (PMID:19376126). This suggests mechanical load is a plausible G×E axis, but that is an inference.

3. Phenotypes

All of the HP CURIEs below are ✔ in cache/hp/terms.csv.

Severity: Varies with dose. The phenotype is mild, late-onset and low-risk in heterozygotes, and more severe in homozygotes:

"Ten of the index cases were heterozygous and one was homozygous. Homozygous had a more severe phenotype." — PMID:33035702 [cached]

Quality of life: No CMH12-specific QoL data were found. General HCM data (e.g. KCCQ in the mavacamten trials) apply only by inference.

4. Genetic / molecular information

• Gene: CSRP3 (hgnc:2472; OMIM *600824; NCBI Gene 8048), 11p15.1, 194-aa protein with two LIM domains (LIM1 and LIM2), a nuclear localization signal, and a glycine-rich region.
• Reference sequences: NM_003476.5 / NP_003467.1 (used in PMID:34558151).

Key variants | Variant | Type | Zygosity | Status / notes | Source | |---|---|---|---|---| | p.Cys58Gly (C58G) | missense, LIM1 zinc finger | het (familial) | The only variant rated likely pathogenic in the 2018 ACMG meta-analysis. Abolishes zinc coordination | PMID:12642359, 15205937, 30012424, 30048712 [cached] | | p.Leu44Pro, p.Ser54Arg/Glu55Gly | missense, LIM1 | het | Cause protein depletion in cells | PMID:30048712 [cached] | | p.Cys150Tyr | missense, LIM2 | het (10) / hom (1) | 11/6456 HCM probands vs 0 in other cardiomyopathy and control groups (p<0.0001). Late-onset, low risk | PMID:33035702 [cached] | | p.Arg122Ter (c.364C>T) | nonsense | hom (Polish), het (Chinese family, cosegregating) | Homozygous null; the heterozygous Chinese family conflicts with recessive-only models | PMID:31919335, 34558151 [cached] | | Homozygous truncating (2 probands) | nonsense/frameshift | hom | First homozygous CSRP3 knockouts in HCM | PMID:30012424 [cached] | | p.Trp4Arg (W4R; rs45550635, look up) | missense, N-terminal | het, ~1% of Europeans | Modifier, not causal on its own | PMID:35384713 [cached]; Geier 2008 PMID:18505755 [not cached] |

"We demonstrate that the common muscle LIM protein-W4R variant is an important modifier, exacerbating the phenotypic expression of HCM, but alone may not be a disease-causing mutation." — PMID:35384713 [cached]

Functional consequence: The pathogenic missense variants cause loss of function through protein destabilization and proteasomal depletion, which amounts to haploinsufficiency or hypomorphism. They do not act as gain of function (functional_impact_category: LOSS_OF_FUNCTION or PARTIAL_LOSS_OF_FUNCTION):

"Intriguingly, the abundance of MLP decreased in KI/KI mice by 80% and in KI/+ mice by 50%. Protein depletion was also observed in cellular studies for two further HCM-causing CSRP3 mutations (L44P and S54R/E55G). We show that MLP depletion is caused by proteasome action." — PMID:30048712 [cached]

Yield and frequency: About 3% of HCM probands carry a CSRP3 variant (12/389 in the Mayo cohort; Bos 2006, PMID:16352453 [not cached]). This share should be read as an upper bound, because pathogenicity is unresolved for many of those variants. Geier 2008 reported a large linkage family plus 4 more families with segregating missense variants (PMID:18505755 [not cached]).

Modifiers: W4R (above). Sarcomeric co-variants such as MYH7 interact with MLP depletion (PMID:35384713).

Epigenetic changes / chromosomal abnormalities: None reported. CNVs involving 11p15.1 are not known as a CMH12 mechanism.

Somatic origin: Not applicable (germline).

5. Environmental information

There are no known environmental, lifestyle or infectious causes. Mechanical load and contractility regulate where MLP sits in the cell (PMID:19376126), which makes exercise and pressure load plausible modulators. This is an inference, not an observation in patients. Following CLAUDE.md, an environmental entry should either be left out or carry a "Left deliberately uncited." waiver.

6. Mechanism / pathophysiology

Ordered causal chain

1. Germline CSRP3 variant (LIM1/LIM2 missense, or biallelic truncation) leads to misfolding of a zinc-finger LIM domain:

   "this single point mutation prevents zinc coordination and proper folding of the second zinc-finger in the first LIM domain, which consequently renders the protein less stable and more susceptible to proteolysis." — PMID:15205937 [cached; IN_VITRO]

2. Unstable MLP is degraded by the proteasome, which results in MLP depletion (about 50% in heterozygous and about 80% in homozygous knock-in mice; PMID:30048712 [cached; MODEL_ORGANISM]). Truncating biallelic alleles reach the same end state, a null, directly.
3. Branch 2b (proteotoxic): Mutant MLP C58G binds BAG3, which triggers a proteotoxic, heat-shock-protein stress response (PMID:30048712).
4. Loss of MLP binding partners at the Z-disc: binding to α-actinin, N-RAP and telethonin (T-cap) drops:

   "Protein-binding studies indicate that mutations in the CRP3 gene lead to a decreased binding activity of MLP to alpha-actinin." — PMID:12642359 [cached] W4R specifically disrupts the interaction with telethonin and its localization (Knöll 2002, Cell 111:943, PMID:12507422 [not cached]).

5. The Z-disc stretch-sensing complex fails, which leads to a mechanosensing defect, disinhibition of calcineurin–NFAT signalling, and hypertrophic gene expression. This step was shown in human iPSC engineered heart tissue, where MLP depletion was secondary to MYH7 load. Applying it to primary CSRP3 depletion is a mechanistic inference:

   "Subsequent reduction in the sarcomeric muscle LIM protein level caused disinhibition of calcineurin-nuclear factor of activated T-cells signaling, which promoted cardiac hypertrophy." — PMID:35384713 [cached; IN_VITRO]

6. Disrupted nucleocytoplasmic shuttling of MLP impairs myofilament remodelling and stress gene expression (sarcomere disarray, BNP). This was demonstrated in rat neonatal myocytes:

   "cyclic strain of myocytes after prior NLS treatment to remove nuclear MLP resulted in disarrayed sarcomeres" — PMID:19376126 [cached; IN_VITRO]

7. Cardiomyocyte hypertrophy and myocyte disarray, plus compensatory N-RAP upregulation (proposed in PMID:15205937), lead to asymmetric septal LV hypertrophy.
8. Fetal gene programme, profibrotic signalling and interstitial fibrosis result in diastolic dysfunction, and with biallelic or severe loss, systolic dysfunction with progression towards DCM:

   "Transcriptome analysis by RNA-seq identified activation of pro-fibrotic signalling, induction of the fetal gene programme and activation of markers of hypertrophic signalling in these hearts." — PMID:30048712 [cached]

9. Clinical expression (dyspnea, arrhythmia, rarely SCD) depends on gene dose and age. Heterozygous carriers have a late-onset, low-risk course (PMID:33035702).

Upstream vs downstream: Steps 1–3 are molecular (MOLECULAR scale). Steps 4–5 are cellular signalling (CELLULAR). Steps 6–7 are tissue and organ (TISSUE). Step 8 is organism-level (ORGANISM).

Suggested terms (✔ = in cache) - GO BP: response to muscle stretch GO:0035994 ✔; calcineurin-NFAT signaling cascade GO:0033173 ✔ (modifier INCREASED); cardiac muscle hypertrophy GO:0003300 ✔; sarcomere organization GO:0045214 ✔; cardiac myofibril assembly GO:0055003 ✔; proteasome-mediated ubiquitin-dependent protein catabolic process GO:0043161 ✔ (MLP depletion) - GO CC: Z disc GO:0030018 ✔; costamere GO:0043173 ✔; nucleus GO:0005634 ✔ - CL: cardiac muscle cell CL:0000746 ✔; fibroblast of cardiac tissue CL:0002548 ✔ (fibrosis node) - Module conformance candidates: run just list-modules cardiomyopathy and just list-modules fibro. cardiomyopathy_maladaptive_remodeling and fibrotic_response are named in CLAUDE.md; check them before conforming.

Molecular profiling: The only omics dataset found is the bulk RNA-seq of C58G knock-in mouse hearts (PMID:30048712; locate the GEO accession using just discover-datasets). No human single-cell, proteomic or metabolomic data specific to CSRP3 were found.

7. Anatomical structures

• Organ: heart UBERON:0000948 ✔; cardiovascular system.
• Site: interventricular septum UBERON:0002094 ✔ (asymmetric). The left ventricle term was not found in the cache; look it up.
• Tissue: cardiac muscle tissue UBERON:0001133 ✔.
• Cells: cardiomyocytes (CL:0000746 ✔); cardiac fibroblasts (secondary).
• Subcellular: Z-disc, costamere, nucleus and nucleolus (PMID:19376126: "The NLS of MLP also localizes to the nucleolus.").
• Secondary organs: None. MLP is also expressed in skeletal muscle, but no skeletal myopathy is documented in CMH12.

8. Temporal development

• Onset: Usually adulthood (mean 55 ± 13 y for C150Y). Biallelic cases can present in adolescence or young adulthood (PMID:30012424 MeSH lists adolescent and adult; check the full text for exact ages).
• Pattern: Insidious. Hypertrophy develops progressively with age-dependent penetrance.
• Course: Chronic and lifelong. Heterozygotes appear to have a stable, low-event course (no events over a mean 11 y follow-up). Homozygous and knockout states are more severe and can progress to systolic dysfunction.
• Remission: None.
• Critical period: Not defined. Serial screening of carriers into later life is warranted because of late penetrance (an inference from PMID:33035702).

9. Inheritance and population

• Inheritance: OMIM lists AD and ClinGen lists semidominant. The literature supports gene-dose effects: heterozygous variants have low penetrance, and homozygous variants give a more severe phenotype.
• Suggested inheritance blocks: Autosomal dominant (HP:0000006) and Autosomal recessive (HP:0000007). Neither is in the cache list I checked, so look both up. Name semidominance in description. Look up whether HPO has a "Semidominant inheritance" term (possibly HP:0032113, unverified).
• Penetrance: Incomplete and age-dependent (PMID:33035702).
• Expressivity: Variable; modified by W4R and sarcomeric co-variants.
• Consanguinity: Biallelic variants are enriched:

  "variants in the minor HCM genes MYL2, MYL3, and CSRP3 more likely to present in homozygosity than the major genes, suggesting these variants are less penetrant in heterozygosity." — PMID:37431535 [cached; Egyptian cohort n=514]

• Founder effects: Not established. C150Y was found in a Spanish/European multicentre cohort, and R122Ter recurs in Polish and Chinese families.
• Prevalence: HCM overall affects about 1/500 ("estimated prevalence of 1/500", PMID:30012424 [cached]). CSRP3 accounts for a small minority of genotyped HCM: 11/6456 probands (about 0.17%) for C150Y alone, and about 3% of the Mayo cohort carried any CSRP3 variant. No CMH12-specific prevalence exists. Record it as measure_type: UNKNOWN / prevalence_class: RARE rather than deriving a number.
• Sex ratio and geography: Not reported for CMH12 specifically.
• Germline mosaicism / anticipation: Not reported.

10. Diagnostics

• Imaging: Echocardiography and cardiac MRI showing LV wall ≥15 mm (≥13 mm in relatives) without another explanation, which is the general ESC/AHA HCM criterion. MRI LGE can assess fibrosis.
• ECG: LVH and repolarization changes; Holter monitoring for AF and NSVT.
• Biomarkers: NT-proBNP and BNP (non-specific). BNP expression depends on MLP in the myocyte models (PMID:19376126).
• Histopathology: Myocyte hypertrophy, disarray and fibrosis (the proband biopsy in PMID:34558151).
• Genetic testing: A multigene HCM panel or exome. CSRP3 is on the ClinGen-curated HCM panel and is typically included (PMID:30012424: "CSRP3 is usually included on HCM gene panels used for molecular diagnosis by next-generation sequencing"). Points for interpreting results:
• Rule out W4R as the sole explanation.
• Look for a second allele or a sarcomeric co-variant.
• Use ancestry-matched controls (PMID:37431535).
• CMA, karyotype, FISH, mtDNA and repeat expansion testing are not indicated.
• Differential diagnosis: Sarcomeric HCM (MYBPC3 and MYH7, which account for most genotype-positive cases). Phenocopies: Fabry disease, PRKAG2, Danon disease (LAMP2), TTR amyloidosis, RASopathies, mitochondrial disease. Also athlete's heart and hypertensive LVH. TRIM63 is another recessive non-sarcomeric HCM gene.
• Screening: Cascade genetic testing and serial imaging of first-degree relatives. This matters particularly because of late penetrance.

11. Outcome / prognosis

• Heterozygous C150Y: favourable. "The p.(Cys150Tyr) variant in CSRP3 causes late-onset and low risk form of hype[rtrophic cardiomyopathy]" (PMID:33035702; the cached abstract is cut off there, so quote only up to the truncation).
• Homozygous / null: more severe, with risk of systolic dysfunction (human case reports plus the KI/KI mouse). There are too few cases to estimate survival.
• No CMH12-specific survival, mortality or QoL statistics exist. Prognostic factors follow general HCM SCD risk models (ESC HCM Risk-SCD), which have not been validated for CSRP3.

12. Treatment

No CSRP3-specific therapy exists. Management follows the general HCM guidelines (2024 AHA/ACC; 2023 ESC).

• Gene therapy, ASO, siRNA: None for CSRP3. A loss-of-function mechanism points towards replacement or stabilization strategies, but this is speculative. PMID:35384713 frames "mechanism-based treatments for HCM that stabilize the Z-disc MLP-mechanosensory complex."
• Trials: No CSRP3-specific trials were found on ClinicalTrials.gov. General HCM trials (EXPLORER-HCM NCT03470545, SEQUOIA-HCM) included unselected genotypes. Verify the NCT IDs with just fetch-reference before citing them.
• Pharmacogenomics: None specific.

13. Prevention

• Primary: Not possible. Preimplantation or prenatal testing is an option in families with known biallelic alleles.
• Secondary: Cascade genotyping. Long-term echo surveillance of heterozygous carriers, given late-onset penetrance.
• Tertiary: SCD risk stratification, AF anticoagulation, and activity counselling per the HCM guidelines.
• Immunization / public health: Not applicable.

14. Other species / natural disease

• No naturally occurring CSRP3-related disease is recorded in OMIA (I did not search directly; verify). Feline HCM is mainly MYBPC3/ALMS1, not CSRP3.
• Orthologs: mouse Csrp3 (MGI). Look up the NCBI Gene ID; zebrafish csrp3.
• MLP's Z-disc mechanosensing role is conserved across vertebrates.

15. Model organisms

Resources: MGI (Csrp3 alleles), IMPC, ZFIN.

Key uncertainties for curators

1. Mode of inheritance. OMIM says AD, ClinGen says SD, and Janin 2018 and Lipari 2020 argue for AR. A heterozygous truncation cosegregates in a Chinese family (PMID:34558151), which conflicts with a purely recessive model. Encode multiple inheritance blocks, each with its own evidence.
2. ClinGen Definitive is recorded against general HCM (MONDO:0005045). Check mapping scope before copying it onto CMH12.
3. W4R should not be modelled as causal. Put it in genetic as relationship_type: MODIFIER.
4. Several canonical papers are not cached: Arber 1997, Knöll 2002, Bos 2006, Geier 2008, Ingles 2019 and Theis 2006. Fetch them before citing; the Arber PMID in particular was not confirmed.
5. No GeneReviews chapter specific to CMH12 is expected. Confirm with just check-genereviews; the general "Hypertrophic Cardiomyopathy Overview" chapter probably applies.

Sources

• Cached in this worktree: PMID:12642359, 15205937, 15781201 (letter; no abstract text), 19376126, 30012424, 30048712, 31919335, 33035702, 34558151, 35384713, 37431535, 41893039; ClinGen CGGV assertions 012b9b15 (HCM, Definitive) and beaafc90 (DCM 1M, Limited).
• Web: Arber 1997 (JCI reference list) · edoc.unibas.ch · Knöll 2002 abstract (Arizona) · Mount Sinai abstract page · WikiGenes 12507422 · LOVD CMH12 · MedGen 393755 · OMIM 600824 mirror · GenCC CSRP3 · PanelApp CSRP3 · Ingles 2019 (PMC6410971) · medRxiv ClinGen re-curation · Geier 2008 (ORA Oxford) · MDC edoc · Bos 2006 (PMC2756511) · Theis 2006 (dkNET) · OMIM 607487 mirror

Reference Validation

Checked with linkml-reference-validator 0.3.0.

All extracted references resolved successfully.

Term Validation

Checked with linkml-term-validator 0.4.5, through the ols: adapter.

Terms the report names something else

These identifiers resolve, so nothing about them looks wrong, and the ontology calls them something unrelated to what the report calls them. That usually means the identifier is not the one the sentence needs:

• MONDO:0005045 (3 mentions) - the report calls it "ClinGen Definitive is recorded against general HCM"; MONDO calls it hypertrophic cardiomyopathy

Terms whose name is worth a second look

The report's name for these is recognisably related to the term's own name without being one of them. A loose paraphrase reads the same way as a citation of the wrong sibling term - and so does a related synonym, which the ontology records precisely because it names something adjacent rather than the same thing - so these are listed rather than judged:

• HP:0000006 (1 mention) - the report calls it "Suggested inheritance blocks: Autosomal dominant"; HP calls it Autosomal dominant inheritance

Prefixes with no resolver

Terms carrying these prefixes were not checked either way, because no configured ontology covers them. An unrecognised prefix may name an ontology this run could not reach as easily as one that does not exist, so nothing here is evidence of fabrication: ORPHA.