Let me fetch the primary abstracts for exact quotes and PMIDs.

I'm pulling the BAG5/CMD2F primary papers into the reference cache so the report can quote their exact wording. Next I'll read them.

Reading the cached Gupta 2016 paper and confirming the BAG5–Hsp70 structural paper; the report comes after that.

Cardiomyopathy, Dilated, 2F (CMD2F; BAG5-related autosomal recessive dilated cardiomyopathy): research report

Prepared 2026-10-08 for the dismech entry kb/disorders/Cardiomyopathy_Dilated_2F.yaml. Quotes marked "verbatim" are copied from cached abstracts or full texts in references_cache/. Ontology terms marked ✔ were checked against the local cache/<prefix>/terms.csv label caches. Those caches confirm that a CURIE exists and has that label. They do not confirm membership in a dynamic enum.

Summary

CMD2F is a rare, early-onset, severe dilated cardiomyopathy. It is caused by biallelic loss-of-function (truncating) variants in BAG5 (hgnc:941 ✔). BAG5 encodes a five-BAG-domain co-chaperone that acts as a nucleotide exchange factor for HSC70/Hsp70. Patients present in the teens or twenties with severe systolic dysfunction and refractory ventricular arrhythmias. Many progress to an LVAD or heart transplantation.

Two mechanisms have been proposed, both from mouse work: - Junctional membrane complex and calcium-handling failure under catecholamine stress (Hakui 2022). - An impaired cardiomyocyte ER-stress response leading to apoptosis (Wongong 2024; Gupta 2016).

AAV9-BAG5 gene replacement rescued the knock-in mouse phenotype. About 12 cases from roughly 7 families have been published, so the evidence base is small.

1. Disease information

All information comes from aggregated case reports and family studies. There is no EHR- or registry-level data.

Grouping context: CMD2F belongs to the recessive "CMD2" series (2A TNNI3, 2B GATAD1, 2C PPCS, 2D RPL3L, …), and to the BAG family of cardiomyopathy genes alongside BAG3 (autosomal dominant CMD1HH). The 2026 ClinGen DCM reassessment lists BAG5 among nine newly assessed genes classified as high evidence (PMID:42708185):

"Nine newly assessed genes were classified as high evidence: BAG5, FLII, LMOD2, MYLK3, MYZAP, NRAP, PPA2, PPP1R13L, and RPL3L. Twelve genes (11 newly appraised) were rated as high evidence with an autosomal recessive (AR) MOI." (verbatim)

2. Etiology

Primary cause: biallelic germline truncating BAG5 variants (PMID:35044787):

"we identified that homozygous truncating mutations in the gene encoding Bcl-2–associated athanogene (BAG) co-chaperone 5 (BAG5) caused inherited DCM in five patients among four unrelated families with complete penetrance." (verbatim)

Genetic risk factors - Biallelic truncation is the cause: homozygous, or compound heterozygous as in Inoue 2023 (PMID:37873655). - Heterozygous truncation may predispose to tachycardia-induced cardiomyopathy, but this has not been established. PMID:35044787 reports:

"We also identified heterozygous truncating mutations in three patients with tachycardia-induced cardiomyopathy, a reversible DCM subtype associated with abnormal calcium homeostasis." (verbatim)

This is a separate, susceptibility-level claim. Do not fold it into the CMD2F inheritance block. - Heterozygous carriers were clinically unaffected in Hakui 2022. In Wongong 2024, two of three female heterozygous relatives had mild or borderline ECG abnormalities (PMID:38796549):

"suggesting a potential predisposition to arrhythmia in female carriers of the BAG5 variant." (verbatim)

Environmental and stress modifiers. These are inferred from mouse models only: - Catecholamine (β-adrenergic) stimulation unmasked dilatation, arrhythmia and death in Bag5 knock-in mice (PMID:35044787). - ER stress (tunicamycin) was required to reveal systolic dysfunction in a second knock-in line (PMID:38796549):

"To develop DCM in Bag5 point mutant KI mice, ER stress is required." (verbatim; full-text Discussion)

The equivalent human triggers have not been identified. Tachyarrhythmia, adrenergic surges and intercurrent illness are plausible candidates but unproven.

Protective factors and gene-environment interactions: none reported in humans. The mouse data suggest a gene × adrenergic/ER-stress interaction, which is a hypothesis.

3. Phenotypes

• Severity: severe and progressive. Wongong 2024 describes "consistently severe manifestations of the disease."
• Quality of life: no formal measures have been published. The burden is that of NYHA III–IV heart failure in adolescents and young adults, including an LVAD and/or transplant, an ICD and recurrent shocks.
• Extracardiac features: none reported. Note that Bag5-null male mice are infertile (IMPC), but human fertility has not been studied.

4. Genetic and molecular information

Gene: BAG5 (BCL2-associated athanogene 5), hgnc:941 ✔, chromosome 14q32. RefSeq used by Wongong: NM_001015049.2.

Reported pathogenic alleles. Variant list per Wongong 2024 discussion (PMID:38796549, verbatim: "c.589C > T (p.Arg197Ter), c.1168C > T (p.Arg390Ter) and c.18dupA (p.His7ThrTer5)"):

Caveats on this table - Wongong describes the Hakui variants as "de novo homozygous". That wording is almost certainly inaccurate, because homozygosity implies inheritance. Do not copy it. - The Inoue 2023 variant assignments come from Wongong's summary. The Inoue abstract could not be retrieved (PMID_37873655.md is cached as content_type: unavailable).

Other gene-level facts - ACMG/ClinGen: ClinGen Gene-Disease Validity, BAG5–dilated cardiomyopathy (MONDO:0005021), AR, Moderate, DCM GCEP, 2024-06-14 (cached CGGV:assertion_97fe83b7-…-2024-06-14T160000.000Z). Quotable row: BAG5 | HGNC:941 | dilated cardiomyopathy | MONDO:0005021 | AR | Moderate. - The assertion is for the broader DCM term, not MONDO:0030680. Under the repository's Gene-Disease Validity Is Copied, Never Assigned rule, check whether it counts as "this entry's disease" before recording it in gene_disease_validity. - GenCC also lists a Moderate AR submission. PanelApp Australia rates BAG5 Green. - Variant class / functional consequence: all reported alleles are truncating. The proposed mechanism is loss of function. Hakui showed reduced protein; Wongong found Bag5 protein "completely absent" in knock-in hearts. Inoue reported reduced BAG5 and SERCA2 immunostaining in patient myocardium, per Wongong's summary. Suggested slot value: functional_impact_category: LOSS_OF_FUNCTION. - Origin: germline. - Allele frequency: not systematically reported. Wongong filtered at <1% in gnomAD and 1000 Genomes. Check gnomAD for p.Arg197Ter and p.Arg390Ter, because recurrence in Japanese families raises the possibility of a founder allele. This is unverified. - Modifier genes, epigenetics, chromosomal abnormalities: none reported. - Lesurf 2022 (npj Genomic Medicine 7:18, a whole-genome sequencing study of early-onset cardiomyopathy) is listed by GenCC as a BAG5 evidence source. Its BAG5-specific finding was not verified.

5. Environmental information

There is no environmental, lifestyle or infectious cause. The only "exposures" with data are experimental stressors in mice: isoproterenol/catecholamine and tunicamycin. These belong in animal_models or experimental_models, not in environmental:.

6. Mechanism and pathophysiology

Causal chain

1. Biallelic truncating BAG5 variants lead to loss of BAG5 protein through nonsense-mediated decay or truncation. This is demonstrated in knock-in mouse hearts and inferred from reduced immunostaining in patient tissue.
2. Loss of BAG5 leads to loss of nucleotide-exchange activity on HSC70/Hsp70, so ADP release slows and chaperone-mediated folding is less efficient. The biochemistry is demonstrated in vitro (Arakawa 2010, PMID:20223214; PMID:35044787: "BAG5 acts as a nucleotide exchange factor for heat shock cognate 71 kDa protein (HSC70), promoting adenosine diphosphate release and activating HSC70-mediated protein folding"). That this is the operative cardiac lesion is inferred.
3. The pathway then branches:
4. 3a. JMC/calcium branch (Hakui 2022, mouse plus cardiomyocytes). Under catecholamine stimulation, BAG5 deficiency leads to reduced abundance of functional junctional membrane complex (JMC) proteins and disrupted JMC structure, which results in abnormal Ca²⁺ handling. PMID:35044787 (verbatim): "BAG5 localized to junctional membrane complexes (JMCs), critical microdomains for calcium handling. Bag5-mutant mouse cardiomyocytes exhibited decreased abundance of functional JMC proteins under catecholamine stimulation, disrupted JMC structure, and calcium handling abnormalities."
   • The specific JMC proteins named in the full text (e.g., JPH2, RyR2, LTCC) were not verified from the abstract.
   • Human support for this branch is reduced SERCA2 in patient myocardium (Inoue 2023, as reported secondhand by Wongong).
5. 3b. ER-stress branch (Wongong 2024; Gupta 2016). During ER stress, BAG5 normally stabilises GRP78 (HSPA5) and limits CHOP-mediated apoptosis (in vitro, rat cardiomyocytes). PMID:26729625 (verbatim): "Bag5 protein expression is significantly increased in the ER during ER stress and that this in turn modulates GRP78 protein stability and reduces ER stress." Without BAG5, ER stress leads to excess cardiomyocyte apoptosis (TUNEL in knock-in mice, n = 1 per genotype). PMID:38796549 (verbatim): "Their cardiac tissues exhibited a notable increase in apoptotic cells, despite non-distinctive changes in CHOP and GRP78 levels." The CHOP-independent result means the downstream effector is unresolved. The authors suggest ATF6 or XBP1, which is speculation.
6. Ca²⁺ mishandling leads to triggered and re-entrant ventricular arrhythmias (mouse arrhythmogenicity under catecholamine stress; refractory VT/VF in patients). It also contributes to contractile dysfunction.
7. Cardiomyocyte loss and contractile failure lead to LV dilatation and systolic dysfunction (DCM), then advanced heart failure (LVAD or transplant), and sudden death. The step from 3a/3b to DCM is shown in mice under stress. In humans it is inferred.

Which steps are upstream: steps 1–2 are upstream and certain. Branch 3a has the stronger evidence: a rescue experiment plus the arrhythmia phenotype. Branch 3b rests on small-n data. The two branches are not mutually exclusive.

Suggested annotations (✔ cache-verified)

Molecular profiling, single-cell, spatial and CRISPR data

None has been published for CMD2F. No GEO datasets were identified. The structural PDB entry 3A8Y covers the BAG5 BD5–Hsp70 NBD complex (Arakawa 2010).

7. Anatomical structures affected

• Organ: the heart. The left ventricle is primary (UBERON:0002084 ✔). Right-ventricular involvement is not well described.
• Tissue and cells: ventricular myocardium and cardiomyocytes (CL:0000746 ✔).
• Subcellular: junctional membrane complexes (sarcolemma/T-tubule–junctional SR dyads) and the ER/SR. BAG5 also has reported mitochondrial and cytosolic roles in non-cardiac cells, which may not be relevant here.
• Secondary organs: those affected by low-output heart failure.
• Laterality: not applicable.

8. Temporal development

• Onset: adolescence to young adulthood. The youngest reported death was at 12 years and the oldest diagnosis at 34. Onset is insidious, sometimes first noticed as an arrhythmia.
• Progression: rapid and progressive toward advanced heart failure. Wongong describes "consistently severe" disease that is less variable than BAG3 DCM. There are no natural-history cohorts, and no remission has been reported.
• Suggested progression: phases: preclinical (carrier or presymptomatic homozygote); symptomatic DCM with arrhythmia; advanced HF needing mechanical support or transplant.

9. Inheritance and population

• Inheritance: autosomal recessive (HP:0000007 ✔), with complete penetrance in biallelic individuals (PMID:35044787).
• Expressivity: severe but variable. In the Wongong sibship LVEF ranged from 36% to 46%, and one sibling had a normal LV size.
• Heterozygotes: possible minor ECG findings in women, and possible susceptibility to tachycardia-induced cardiomyopathy. Both are unconfirmed.
• Consanguinity: one Middle Eastern consanguineous family (PMID:38796549).
• Populations: Japanese (Hakui and Inoue, Osaka and Tokyo) and Middle Eastern (studied in Thailand).
• Prevalence: unknown, ultra-rare. About 12 published patients (GenCC counts 12 cases across 5 publications). Suggested encoding: measure_type: CASES_IN_LITERATURE.
• Sex ratio, anticipation, mosaicism: not reported, and anticipation is not expected for this mechanism.

10. Diagnostics

• Genetic testing: a DCM multigene panel, exome or genome sequencing. BAG5 is now on PanelApp Green lists, and Hakui 2022 recommends adding it to testing. Confirm biallelic status with parental segregation and Sanger sequencing. Truncating variants are well detected by exome sequencing.
• Imaging: echocardiography (LVEDD Z-score, LVEF) and cardiac MRI (hypokinesia). LGE patterns have not been reported.
• ECG and Holter: low limb-lead voltage, inferolateral T-wave inversion, poor R progression, and VT/VF.
• Biomarkers: BNP/NT-proBNP elevated (non-specific).
• Histology: reduced BAG5 and SERCA2 immunostaining in explanted myocardium (Inoue 2023). This is research-level, not diagnostic.
• Differential diagnosis:
• other recessive and early-onset DCMs: TNNI3-AR, GATAD1, RPL3L, NRAP, PPA2, PLEKHM2;
• BAG3 DCM (dominant, broader age range);
• arrhythmogenic cardiomyopathy (DSP, FLNC, PLN — overlapping arrhythmic DCM);
• myocarditis;
• tachycardia-induced cardiomyopathy (reversible after rate control);
• neuromuscular and metabolic DCMs (Duchenne/Becker, Barth, mitochondrial).
• Screening: cascade testing of siblings in affected families, followed by echo and ECG surveillance of biallelic relatives.

11. Outcome and prognosis

• Reported outcomes are poor. Every patient in the founding Japanese cohort needed an LVAD and at least one was transplanted (PanelApp summary of PMID:35044787). One Middle Eastern sibling died at 12 years awaiting transplant.
• Arrhythmic burden is high and refractory (PMID:36130910).
• No survival statistics exist.
• Prognostic factors are not established. Biallelic truncation appears uniformly severe.

12. Treatment

No disease-specific therapy is approved. Management follows standard heart-failure care for DCM.

• Clinical trials: no NCT or ICTRP trials specific to BAG5 were found.
• Pharmacogenomics: none.

13. Prevention

• Primary prevention: genetic counselling (NCIT:C15240), carrier testing in consanguineous kindreds, and prenatal or preimplantation diagnosis for families with known variants.
• Secondary prevention: cascade genotyping of siblings, then echo and ECG surveillance of biallelic relatives.
• Tertiary prevention: an ICD to prevent sudden death and early referral for advanced HF therapy.
• Avoiding catecholamine excess: plausible from mouse data, but not validated.

14. Other species and natural disease

No naturally occurring BAG5 cardiomyopathy has been reported in animals, and OMIA lists none. The mouse orthologue is Bag5 (MGI:1917619). BAG5 is conserved across vertebrates.

15. Model organisms

Key references

Gaps: - The Inoue 2023 and Hakui 2022 Circ J abstracts could not be retrieved. - The JMC protein names and per-patient clinical details from Hakui 2022 need the full text. - The OMIM numbers come from secondary sources. - There are no prevalence, natural-history or trial data.

Repository note: this research run added the following files to references_cache/, generated with just fetch-reference: - PMID_35044787.md, PMID_38796549.md, PMID_42708185.md, PMID_26729625.md — usable. - PMID_37873655.md, PMID_36130910.md — cached with no quotable text. - PMID_39712638.md — an unrelated sevoflurane paper picked up by the PubMed search. My attempt to delete it was not approved. Delete it before committing.

Sources: - Osaka University press release (Hakui/Asano 2022) - GenCC BAG5 submission - GenCC SGC-106981.2 - GenCC BAG5 gene page - Wongong et al. 2024, PMC11127938 - PanelApp Australia BAG5 - MGI OMIM:619747 - Disease Ontology DOID:0081162 - DrugMap DIS6YW38 - genebe BAG5 - NORD MONDO page - Gupta et al. 2016, PMC4909508 - RCSB PDB 3A8Y - IMPC Bag5 - Cyagen Bag5 KO - Lesurf 2022 (BSC record) - medRxiv Lesurf preprint

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:

• DOID:0081162 (5 mentions) - the report calls it "DOID"; DOID calls it dilated cardiomyopathy 2F

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:

• MONDO:0005021 (2 mentions) - the report calls it "ACMG/ClinGen: ClinGen Gene-Disease Validity, BAG5–dilated cardiomyopathy"; MONDO calls it dilated cardiomyopathy**, and lists "familial dilated cardiomyopathy" among its other names
• NCIT:C15240 (1 mention) - the report calls it "Primary prevention: genetic counselling"; NCIT calls it Genetic Counseling**

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: OMIM, MGI.