Truncating Titin and Lamin A/C Variants in Anthracycline-Induced Cardiomyopathy

Abstract

Background

Anthracycline chemotherapy-related cardiomyopathy (CCM) is a serious adverse event that can occur several years after completion of therapy. Demographic and clinical risk factors have failed to predict which patients will experience CCM. Genetic variants may account for a significant proportion of interindividual variation and CCM.

Objectives

This study aimed to identify genetic variants in the known idiopathic cardiomyopathy genes that predispose patients to CCM.

Methods

We developed a cardiotoxicity registry. Patients were (and continue to be) enrolled and consented for chart review and DNA sequencing. We sequenced whole exomes of the first 136 patients (anthracycline, n = 55; anti-HER2 [no anthracycline], n = 71; other chemotherapy, n = 10), primarily focusing on titin (TTN) truncating variants, known to be present in ∼25% of patients with primary dilated cardiomyopathy, and previously reported in 7.5% of patients with CCM, followed by exploration of rare nonsynonymous variants in 62 established genes for idiopathic cardiomyopathy.

Results

Eighteen of 55 patients treated with anthracycline experienced CCM. TTN truncating variants were identified in 2 of 18 (11%) CCM patients and absent in 37 patients who did not experience CCM. We identified a pathogenic variant in lamin A/C (p.Arg190Gln) in 1/18 (5.5%) patients and the same rare (p.Glu1127Gly) variant in ryanodine receptor 2 occurred in 2/18 (11%) of patients. We observed enrichment of rare missense variants in patients with anthracycline CCM compared to anti-HER2 therapy without anthracycline CCM (P < 0.00001).

Conclusions

Three of 18 anthracycline CCM patients carried likely pathogenic variants in the most common causative genes for idiopathic dilated cardiomyopathy, TTN, and lamin A/C. Rare nonsynonymous variants in ryanodine receptor 2 and other idiopathic cardiomyopathy genes warrant further investigation.

Central Illustration

The anthracycline, doxorubicin, is one of the most widely used and effective chemotherapy agents in oncology. However, an important side effect of this group of drugs is cumulative, dose-related, cardiomyopathy, which may result in persistent and potentially irreversible myocardial dysfunction, reflecting long-term structural and functional cardiac damage.^1,2 This limits the lifetime dose a patient can receive, even if the cancer recurs. Twenty-six percent of patients experience dox-related congestive heart failure (HF) at a cumulative dose of 550 mg/m².¹ Even at low “safe” doses (<300 mg/m²), 26% of adults develop a significant decline in left ventricular ejection fraction (LVEF) at 6 months.³

Current risk models of demographic and lifestyle factors do not effectively predict which patients will progress to HF, suggesting a genetic component for interindividual variation in outcome and severity.⁴ Identification of pathogenic rare variants in cardiomyopathy is challenging in both the field of chemotherapy-related cardiomyopathy (CCM)⁴ and idiopathic forms of dilated and hypertrophic cardiomyopathy (DCM, HCM, respectively).⁵ Even for familial autosomal dominant forms of DCM, genetic architecture encompasses ∼62 known genes and can be complex.6, 7, 8 This is strongly illustrated by the 2 most common causative DCM genes, titin (TTN) and lamin A/C (LMNA) which account for ∼25 and 7.5% of familial and 18 and 3.6% of sporadic DCM cases, respectively.^9,10 The relatively high proportions of pathogenic TTN truncating variants and LMNA point mutations in sporadic DCM are suggestive of multigene causation⁶ and/or the need for a second environmental hit, particularly for TTN, where truncating variants are observed in ∼10% of individuals with alcoholic cardiomyopathy,¹¹ and 15% of women with peripartum cardiomyopathy,^12,13 compared to ∼1% of the general population,¹⁴ leading to the hypothesis that rare variants in the established idiopathic DCM genes play a role in the development of CCM.

In the current literature, patients with CCM who carry truncating variants in the 364 exon gene, TTN, are reported to experience more HF and atrial fibrillation and impaired myocardial recovery than those without, a finding which was also consistent with a prior case study of 2 patients with breast cancer who developed severe HF within months after chemotherapy.¹⁵ Specifics of other known cardiomyopathy genes as a cause of CCM are less conclusive, but a recent case-control study of 46 patients with a history of CCM suggested an increased burden of variants across 59 established idiopathic cardiomyopathy genes.¹⁶ Therefore, further investigation (Central Illustration) of TTN frameshift variants and further exploration of this larger group of established idiopathic DCM is warranted in the setting of CCM.

Central Illustration.

Truncating Titin and Lamin A/C Variants in Anthracycline-Induced Cardiomyopathy

We developed an in-house cardiotoxicity registry/biorepository for genetic screening of cancer patients at risk of chemotherapy-related cardiomyopathy (CCM). This report focuses on 62 genes previously associated with idiopathic forms of cardiomyopathy. Exome sequence data of 55 participants treated with the chemotherapy drug doxorubicin identified 2 participants with truncating variants in the gene encoding titin (TTN), and both of these participants required cardiac transplants. A variant that altered the amino acid arginine to glutamine at amino acid position 190 (Arg190Gln) in the gene encoding lamin A/C (LMNA) was identified in one of 18 study patients who experienced CCM deemed to be caused by doxorubicin. A very rare variant (frequency ∼0.03% of Europeans in the general population) ryanodine receptor 2 gene (RYR2) p.Glu1127Gly was identified in 2 of 18 (11%) of study participants who experienced CCM following doxorubicin treatment

Methods

Patient population

Patients at Mayo Clinic Jacksonville, Mayo Clinic Arizona, University of Florida, and Moffit Cancer Center were enrolled to a cardiotoxicity registry in protocols approved by the Mayo Clinic Institutional Review Board (IRB #19-002566 and #IRB 22-001501), in which patients consented for clinical chart review and DNA sequencing. Eligibility to the registry and sequencing project (began in 2019) includes any patient receiving or who had previously been treated with any potentially cardiotoxic cancer therapy. Patients are enrolled either prospectively prior to therapy initiation and followed for 2 years to determine whether or not they develop CCM or, as retrospectively when first identified with chemotherapy-related cardiotoxicities. Patients with prior cardiovascular disease such as coronary artery disease were included. Clinical data were collected at baseline conditions (prior to chemotherapy) including pre-existing comorbidities of current or past smoking, history of alcohol use, prior drug use, coronary artery disease, hypertension, hyperlipidemia, diabetes, renal failure, prior history of myocardial infarction, prior history and of stroke or transient ischemic attack, previous percutaneous coronary intervention, previous peripheral bypass, previous carotid stent, previous carotid endarterectomy, previous vascular claudication, known cardiac arrhythmia including atrial fibrillation or flutter, renal failure, peripheral vascular disease, automatic implantable cardioverter defibrillator, and any cardiac medications prior to cancer therapy. Oncological disease characteristics and treatment regimens were collected including: cumulative anthracycline dose, exposure to anti-HER2 therapy, exposure to other chemotherapies, radiation to the chest, dose of radiation, maximum dose of radiation to the heart, start and end dates of each cancer treatment. The cardiotoxicity data collected included echocardiography and cardiac magnetic resonance imaging (MRI) parameters, time to cardiotoxicity, and duration of follow-up. Cardiac outcomes collected included cardiac side effects (recorded as chest pain, shortness of breath, weight gain due to fluid overload, leg swelling, syncope), left ventricular systolic dysfunction, HF, left ventricular assist device, myocardial infarction, arrhythmia, atrial fibrillation, myocarditis, critical limb ischemia, cancer treatment interruption, duration of interruption, cancer treatment rechallenge, cardiac transplantation, and cardiac death. Initial chart review and data entry were performed by a trained data coordinator. These data were then exported from our in-house Scientific Database Management System. Oncology data were reviewed and checked by the team oncologist (P.A.). Data fields from patients with any type of cardiotoxicity were then exported and sent to the dedicated cardio-oncology subspecialty program (A.M. and J.C.R.) for validation. The Scientific Database Management System database was then updated with any corrections. The data fields collected were designed to be used to determine the cardiac outcome by any definition for the purpose of collaboration and data harmonization. For the purpose of this report, cardiomyopathy and HF were defined according to the American college of Cardiology (ACC) and American Heart Association (AHA) Guideline for the Management of Heart Failure.¹⁷ Cardiomyopathy was defined as an LVEF <50% and ACC/AHA Stage B. HF was defined as an ACC/AHA Stage C or D. The diagnosis of cancer-related cardiac toxicity was made as a diagnosis of exclusion based off extensive work-up ruling out other causes for cardiomyopathy including ischemic and inflammatory. These work-ups included ischemic evaluation and coronary angiography (computed tomography or invasive) and cardiac MRI.

At the time of the current data cutoff (May 2025), a total of 185 patients have been enrolled to the registry, and exome sequencing is completed for 136 patients. Here we focus on the exome sequences of 55 patients who were treated with anthracycline, primarily on TTN truncating variants, then expanding to rare nonsynonymous variants in 61 additional established idiopathic cardiomyopathy genes (Supplemental Table 1). Frequencies of putative pathogenic variants were also assessed in 71 patients treated with anti-HER2 therapy (no anthracycline) and 10 patients treated with other types of chemotherapy.

Exome sequencing and variant analysis

DNA extractions were performed at the Mayo Clinic Biospecimens and Processing core facility. Exome sequencing was performed through the Mayo Clinic NGS core. Briefly, paired-end libraries were prepared with 400 ng of genomic DNA using the SureSelect XT Low Input Reagent Kit (Agilent). The concentration and size distribution of the completed libraries were determined using an Agilent Bioanalyzer DNA 1000 chip or Advance Fragment Analyzer and Qubit fluorometry (Invitrogen). Adaptor-ligated DNA was amplified with the SureSelect Post-Capture forward and specific index reverse primers. Whole exon capture was carried out using 750 ng of the prepped library following the protocol for Agilent’s SureSelect Human All Exon v5 + UTRs 75 MB kit. The purified capture products were then amplified and final library quality assessment completed using Qubit (Invitrogen) and an Agilent Bioanalyzer DNA 1000 chip. Libraries were sequenced at an average coverage of approximately 120x following Illumina's standard protocol for the Illumina NovaSeq 6,000 and S4 flow cell. The flow cells were sequenced as 150 X 2 paired end reads using the NovaSeq S4 v1.5 sequencing kit and NovaSeq Control Software v1.8.0. Base-calling was performed using Illumina’s RTA version 3.4.4.

Utilizing Mayo Clinic’s Genome GPS pipeline v 5.0.6 for variant calling, reads were aligned to the reference genome (GRCh38) using Burrows Wheeler Aligner-Maximum Exact Matches (v 0.7.10) and variant calling and joint genotyping was performed with Genome Analysis ToolKit (GATK) v3.6, while implementing the Best Practices Workflow.¹⁸ Following evaluation of sequencing and alignment quality metrics, comprehensive quality control procedures for WES data were implemented to evaluate samples from coverage, alignment quality, transition to transversion ratio, contamination, call rate, sex check, relatedness, and population substructure. Bi-allelic variants passing GATK Variant Quality Score Recalibration filter, having a genotyping rate ≥98% and minor allele frequency (MAF) ≥0.1% and Hardy-Weinberg P value >5e-08 were retained. Variants were annotated to various annotation sources using ANNOVAR. Combined annotation-dependent depletion (CADD) scores (v1.7)¹⁹ were utilized to evaluate pathogenicity of coding and noncoding variants. TTN exon numbering and variants were annotated as per.²⁰ Established cardiomyopathy genes are listed in central illustration and Supplemental Table 1. We filtered variants across 62 established primary cardiomyopathy and channelopathy genes, retaining nonsynonymous variants with minor allele frequencies of ≤0.5% in GnomAD European non-Finnish, African, and East Asian ancestries,²¹ and with a CADD.Phred score ≥20.^19,22

Statistical analysis

Comparisons between CCM/HF and non-CCM/HF groups in Table 1 were assessed by unpaired t-test for continuous data and by chi-square test for categorical data. Fisher exact test was used for categorical data with cell counts <5. Analyses were performed in GraphPad Prism version 10.0.0. P values <0.05 were considered statistically significant. P values were not corrected for multiple testing and considered exploratory.

Table 1.

Clinical Characteristics of Patients Treated With Anthracycline, Grouped by Cardiac Outcome

	Anthracycline-Related CCM/HF (n = 18)	No Anthracycline-Related CCM/HF (n = 37)	P Value
Baseline
Age at cancer diagnosis, y	Range 20-66, mean 46	Range 31-70, mean 52	0.087
Female, n (%)	16 (89)	37 (100)	0.103
Race, n (%)
White	13 (72)	28 (76)	0.863e
Black	4 (22)	5 (13)
Asian	1 (6)	4 (11)
Hispanic	0 (0)	3 (8)
Comorbidities: n (%)
Current or past smokers	5 (28)	10 (27)	0.953
History of alcohol use	10 (56)	24 (65)	0.505
Coronary artery disease	3 (17)	3 (8)	0.381
Hypertension	12 (67)	13 (35)	0.028
Hyperlipidemia	8 (44)	14 (38)	0.639
Diabetes	6 (33)	9 (24)	0.482
Oncological disease, n (%)
Breast cancer	14 (78)	37 (100)	0.009
Hematological malignancy	4 (22)	0 (0)
Cumulative anthracycline dose, mg/m2, range (mean)	133-310 (222)a	152-252 (232)c	0.459
Exposure to anti-HER2, n (%)	7 (39)	0 (0)	0.0002
Radiation to chest, n (%)	13 (72)	28 (76)	0.783
LVEF prior to chemotherapy, range (mean)	42-65 (58)b	55-72 (65)d	0.002
At cardiotoxicity diagnosis
Lowest LVEF, range (mean)	8-51 (27)	55-69 (61)	<0.0001
Time to cardiotoxicity, y, median (IQR)	3.5 (1-10.75)	NA	NA
NYHA functional class III-IV, n (%)	8 (44)	NA	NA
Follow-up
Duration of follow-up, y, median (IQR)	14.5 (5.25-20)	4 (2-4)	<0.0001
Final LVEF, %, range (mean)	23-63 (47)	55-71 (63)	<0.0001
Cardiac transplantation, n (%)	3 (17%)	NA	NA
Cardiac death, n (%)	1 (6%)	NA	NA

P values < 0.05 are highlighted in bold.

CCM = chemotherapy-related cardiomyopathy; HF = heart failure; LVEF = left ventricular ejection fraction.

^aUnknown for 10/18 patients.

^bBaseline LVEF missing for 10/18 patients.

^cUnknown for 3/37 patients.

^dUnknown for 1/37 patients.

^eComparison for White vs non-White; NA, not applicable.

Rare nonsynonymous, potentially deleterious variants were determined as those with an MAF ≤0.5% in 3 different GnomAD populations (European non-Finnish, African, and East Asian) and CADD.Phred score ≥20. Exploratory comparisons of the total number of rare, potentially deleterious nonsynonymous variants in 62 idiopathic cardiomyopathy genes between CCM patients treated with anthracycline and CCM patients treated with anti-HER2 regimens that did not include anthracycline were assessed by chi-square test.

Results

Patient populations

Whole exome sequencing was completed for a total of 136 patients in the registry. Fifty-five patients received anthracycline. Eighteen of 55 (33%) patients experienced CCM or HF diagnosis that was attributed to cancer treatment. Clinical characteristics of patients treated with anthracycline, with and without CCM, are detailed in Table 1. Fifty-one of 55 (93%) patients were treated for breast cancer and 4/55 patients (7%) were treated for hematological malignancy. Patient population consisted of 41/55 (75%) White, 9/55 (16%) Black, 5/55 (9%) Asian, and 3/55 (5%) Hispanic. Mean dose of anthracycline was similar between patients with (222 mg/m²) and without (232 mg/m²) CCM. Overall, the median follow-up time was 4 years (IQR: 2-11.5). In patients with and without CCM, the median follow-up time was 14.5 years (IQR: 5.25-20) and 4 years (IQR: 2-4), respectively. Median time to cardiotoxicity was 3.5 years (IQR: 1-10.75).

None of the patients in the group without CCM received anti-HER2 therapy. Seven of 18 patients (39%) with CCM or HF were treated with both anthracycline and anti-HER2 therapy, whereas none of the patients in the group without CCM received anti-HER2 therapy (trastuzumab, pertuzumab, T-DM1), and this difference was statistically significant (P = 0.0002). Prior to chemotherapy, we observed a higher proportion of patients with hypertension (P = 0.028), and lower LVEF (P = 0.002) in the CCM group compared to the non-CCM group. Seventy-one patients received anti-HER2 therapy and no anthracycline, of which 7/71 (10%) experienced CCM. Ten patients received other treatments than anthracycline and anti-HER2, none experiencing CCM.

Known likely pathogenic and pathogenic variants in idiopathic DCM genes occurred in patients with anthracycline-related cardiomyopathy

TTN truncating variants were identified in 2 of 18 (11%) patients who experienced CCM and/or HF following treatment with doxorubicin. TTN truncating variants were absent in 37 patients who were treated with doxorubicin, who did not experience CCM or HF. A likely pathogenic variant in LMNA (LMNA p.Arg190Gln) was identified in one of 18 (5.5%) patients who experienced CCM or HF following treatment with doxorubicin. We also observed the same rare variant in ryanodine receptor 2 (RYR2) (p.Arg1084Lys) in 2 of 18 (11%) CCM patients, as was KCNQ1, p.Lys393Asn. Neither of these variants is previously reported as likely pathogenic in idiopathic DCM. Case history and variant details are described for each individual below and further detailed in Table 2.

Table 2.

Rare Nonsynonymous Variants Identified in Patients With Anthracycline CCM

Patient ID and Outcome	Gene, Variant	Hg38 Position	rsID	CADD PHRED Score (vs1.7)	gnomAD Frequency (EUR, AFR, EAS)	Previously Reported as (Likely) Pathogenic in ClinVar
CRT-001 Cardiac transplant	TTN c.89900-89903del p.Asn29967MetfsTer27	Chr2: 178552996	rs869312081	54.0	0.00, 0.00, 0.00	y: 223,324
	LAMA4 c.C2603G p.Pro868Arg	Chr6: 112142183	rs368543835	23.1	0.00004, 0.00, 0.00	n
	RYR2 p.Arg1084Lys	Chr1: 237566603	rs193922624	21.2	0.0002. 0.00, 0.00	n
CRT-079 Cardiac transplant	TTN c.89846-89847del p.Thr29949SerfsTer11	Chr2: 178553052	None	57.0	Not present	n
	RBM20 c.C1994G p.Ser665Cys	Chr10: 110812391	rs750494502	22.1	0.00003, 0.00, 0.00	n
	CAV3 c.C233T P.Thr78Met	Chr3: 8745644	rs72546668	33.0	0.0027, 0.004, 0.00	n
	KCNQ1 c.G1179T: p.Lys393Asn∗	Chr11: 2587620	rs12720457	21.5	0.0004, 0.001, 0.00	n
CRT-003 Cardiomyopathy, lowest LVEF 15, final LVEF 43	LMNA c.G569A p.Arg190Gln	Chr1: 156,134,458	rs267607571	32.0	0.000003, 0.00, 0.00	y: 66,910
	LAMA4 c.G4505A p.Arg1502His	Chr6: 112120443	rs781976795	21.0	0.00002, 0.00, 0.00007	n
	KCNE2 c.T170C p.Ile57Thr	Chr21: 34370648	rs74315448	25.9	0.0001, 0.0001, 0.00	n
	FLNC c.G1657A p.G553S	Chr7: 128840655	rs201572079	29.8	0.0001, 0.0000, 0.00003	n
CRT-090 Cardiac transplant	RYR2 c.A3380G p.Glu1127Gly	Chr1: 237566732	rs200525962	24.3	0.0003, 0.0002, 0.006	n
CRT-094 Cardiomyopathy, lowest LVEF 40	RYR2 c.A3380G p.Glu1127Gly	Chr1: 237566732	rs200525962	24.3	0.0003, 0.0002, 0.006	n

Variant details and outcomes are given for known likely pathogenic or known pathogenic variants, including additional rare nonsynonymous variants found in the same patient. Rare variants in RYR2 and KCNQ1 that were not previously identified as (likely) pathogenic in idiopathic cardiomyopathy are reported where they occurred in more than one patient with CCM.

EUR = European (non-Finnish); EAS = East Asian; AFR = African in GnomAD; TTN = titin; LMNA = lamin A/C; RYR2 = ryanodine receptor 2; other abbreviation as in Table 1.

^∗variant also identified in patient CRT-086 (treated with anthracycline, outcome cardiomyopathy).

TTN truncating variant: CRT-001 is a 26-year-old female treated for non-Hodgkin lymphoma with 4 cycles of cyclophosphamide, doxorubicin, vincristine, and prednisone (CHOP) followed by preconditioning regimen for transplant including carmustine, etoposide, and cyclophosphamide. The cumulative anthracycline dose was 200 mg/m². She then received consolidation radiation therapy to the anterior mediastinum. A multigated-acquisition (MUGA) scan performed after transplant, prior to radiation therapy, revealed an ejection fraction of 53%.

Three years later, she presented to the emergency department with 3 days of progressive dyspnea and persistent cough which led to a diagnosis of acute systolic HF. Subsequent evaluations supported nonischemic DCM with an estimated LVEF of 20%. HF was managed with carvedilol, lisinopril, spironolactone, digoxin, and furosemide. Left ventricular function did not improve on maximum tolerated guideline-directed medical therapy; therefore, a biventricular implantable cardiac defibrillator was placed for primary prevention. She progressed from an NYHA functional class I to class III over the course of 3 years and finally received orthotopic heart transplant at the age of 46 years.

CRT-001 carried a TTN truncating variant (rs869312081) deletion in exon 336 resulting in a frameshift, p.Asn29967MetfsTer27. This variant maps within the TTN A-band region where the majority of truncating pathogenic variants associated with DCM have been reported⁹ and has been previously identified as a likely pathogenic variant in primary DCM.^9,23 It is extremely rare in the general population, occurring only 3 times in a population of 1,179,770 alleles (589,885 individuals) of European ancestry and is not present in >37,000 individuals of African ancestry and not present in >22,000 individuals of East Asian ancestry in the GnomAD database and was previously determined as pathogenic by the Ambry Variant Classification Scheme (2023).

Following our exome variant filtering criteria, CRT-001 also carried missense variants in the LAMA4 and RYR2 genes, also known to be causative of familial forms of cardiomyopathy. LAMA4 p.Pro868Arg (rs368543835) and RYR2 p.Arg1084Lys (rs193922624) are also very rare in the general population data in GnomAD. Multiple reports for these variants in ClinVar for conditions of primary DCM and catecholaminergic polymorphic ventricular tachycardia 1 include classifications of likely benign and uncertain significance.

TTN truncating variant: CRT-0079 is a 60-year-old female treated for right-sided ductal breast cancer with mastectomy and adjuvant chemotherapy including 4 cycles of dose-dense doxorubicin and cyclophosphamide (240 mg/m² of doxorubicin) followed by 4 cycles of paclitaxel. Transthoracic echocardiogram completed prior to chemotherapy noted normal LVEF.

Two years following completion of chemotherapy, she presented to the emergency department for 1 week of dyspnea on exertion, paroxysmal nocturnal dyspnea, and orthopnea. She was diagnosed with acute systolic HF with an estimated LVEF of 10% to 15%. Subsequent testing established a diagnosis of nonischemic cardiomyopathy. She was referred to advance HF clinic for consideration of cardiac transplantation. Two months later, she was hospitalized for cardiogenic shock. She was deemed ineligible for transplant due to diagnosis of bladder cancer that was incidentally discovered during transplant work-up. With HF optimal medical therapy, LVEF improved to 52% 1 year later.

CRT-079 carried a 2 nucleotide (TG) deletion resulting in a frameshift, p.Thr29949SerfsTer11, also in exon 336 in the TTN A-band region. This variant has not been previously reported in the literature and is not present in GnomAD at this time. CRT-079 also presented with rare missense variants in RBM20, p.Ser665Cys (rs750494502) and CAV3, T78M (rs72546668) and KCNQ1, p.Lys393Asn. These variants have not previously been reported in ClinVar, but we note that KCNQ1 p.Lys393Asn, very rare in GnomAD populations, was also identified in one other patient in this study, also treated with anthracycline who experienced cardiomyopathy without HF (at the time of current data freeze).

LMNA variant: CRT-003 is a 41-year-old female treated for left-sided ductal breast cancer with lumpectomy, adjuvant chemotherapy, and radiation therapy (5040 cGy). Her chemotherapy regimen included 4 cycles of dose-dense doxorubicin and cyclophosphamide (240 mg/m² of doxorubicin) followed by 4 cycles of paclitaxel.

Fourteen years later she noticed shortness of breath with minimal exertion. Transthoracic echocardiogram performed estimated LVEF of 15% with subsequent testing being consistent with nonischemic cardiomyopathy. She never progressed beyond an NYHA functional class II or required hospitalization for HF. With guideline-directed medical therapy including sacubitril-valsartan, metoprolol, spironolactone, and dapagliflozin, her LVEF improved to 44% 1 year later.

CRT-003 carried a very rare missense variant in LMNA, p.Arg190Gln (rs267607571). After TTN, the most frequent pathogenic variants in familial DCM are those in LMNA (6%).^7,10 p.Arg190.Gln is previously reported as likely pathogenic for DCM and arrhythmogenic right ventricular cardiomyopathy (ClinVar ID: 66910).^10,24 CRT-003 also carried very rare missense variants in LAMA4 p.Arg1502His and KCNE2 p.Ile57Thr which have been previously reported in ClinVar as likely benign or of uncertain significance for conditions of DCM, Long QT, and atrial fibrillation.

RYR2 rare missense variants: CRT-090 and CRT-094 both carried the same RYR2 missense variant, p.Glu1127Gly (rs200525962) (2/18, 11% patients with CCM). This variant is of low frequency in GnomAD, MAF, 0.00075 including homozygous genotypes in 14 of from a total of 107,000 individuals. ClinVar reports include benign, likely benign, and unknown significance for conditions of sudden cardiac death, cardiomyopathy, HCM, arrhythmogenic right ventricular dysplasia 2, catecholaminergic polymorphic ventricular tachycardia 1. We note here that CRT-090 and 094 were negative for other rare coding variants in known cardiomyopathy and channelopathy genes, and a different rare missense variant in this gene (p.Arg1084Lys (rs193922624) was identified one patient with a TTN truncating variant (CRT-001). Clinical descriptions of CRT-090 and CRT-094 are given below.

CRT-090 is a 55-year-old female treated for diffuse large B-cell lymphoma with 6 cycles of rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisone (R-CHOP) plus lenalidomide. Her pretreatment LVEF was estimated at 51%. She complained of dyspnea after cycle 6 which led to a diagnosis of acute systolic HF with an estimated LVEF of 27%. Subsequent testing was consistent with dilated cardiomyopathy. HF was managed with metoprolol, sacubitril-valsartan, and spironolactone. LVEF did not improve on maximum-tolerated guideline-directed medical therapy; therefore, an implantable cardiac defibrillator was placed for primary prevention. A cardiac ablation was later performed for a ventricular ectopy burden of 23%. Paroxysmal atrial fibrillation was also noted and managed with amiodarone. At the age of 64 years, she was hospitalized for cardiogenic shock and underwent an orthotopic heart transplant the same year.

CRT-094 is an 18-year-old female treated for diffuse large B-cell lymphoma with 6 cycles of CHOP followed by 30 fractions of radiation therapy to a left lung mass. She was then treated for left-sided ductal breast cancer at the age of 48 years with docetaxel, carboplatin, trastuzumab, and pertuzumab (TCHP). Her pretreatment LVEF was estimated at 55% to 60%. After 2 cycles, her LVEF dropped to 40% to 45%. She did not develop clinical HF. TCHP was resumed and she completed a total of 6 cycles. Her post-treatment LVEF was estimated at 55% to 60%.

Rare variants in 62 known idiopathic cardiomyopathy genes

Finally, we assessed the total number of rare nonsynonymous variants across 62 idiopathic cardiomyopathy genes in patients groups by CCM and treatment status, after filtering for MAF ≤0.5% in 3 different GnomAD populations (European non-Finnish, African, and East Asian) and a CADD.Phred score ≥20 (Table 3). Within this group of genes and variants, we observed a significantly higher proportion of variants in CCM patients treated with anthracycline compared to CCM patients treated with anti-HER2 regimens that did not include anthracycline (P = 0.0001).

Table 3.

Total Number of Rare Nonsynonymous Variants in 62 Known Cardiomyopathy Genes Was Significantly Higher in CCM Patients Treated With Anthracycline Compared to CCM Patients Treated With anti-HER Regimens Without Anthracycline

	Total Variants in Group	Unique Variants in Group	Variants in >1 Patient in Group
Anthracycline CCM, n = 18	47	45	KCNQ1 c.G1179T: p.Lys393Asn rs12720457 present in 2 patients RYR2 c.A3380G p.Glu1127Gly rs200525962 present in 2 patients
Anthracycline no CCM, n = 37	70	69	OBSCN rs542888890 present in 2 patients
AntiHER2 therapy, no anthracycline, CCM, n = 7	15	15	None
AntiHER2 therapy, no anthracycline, no CCM, n = 64	136	134	TTN, rs72677232 present in 2 patients DES, rs57965306 present in 2 patients

Abbreviations as in Tables 1 and 2.

Nonsynonymous variants in 62 idiopathic cardiomyopathy genes were retained after filtering for minor allele frequency ≤0.5% in 3 different GnomAD populations (European non-Finnish, African, and East Asian) and a CADD.Phred score ≥20. Variants not present in GnomAD were retained. Within this group of genes and variants, we observed a significantly higher proportion of variants in CCM patients treated with anthracycline compared to CCM patients treated with anti-HER2 regimens that did not include anthracycline, P < 0.00001.

Discussion

Identification of pathogenic rare variants is a key challenge in the field of CCM.^4,5 The barriers in answering this question are largely explained by the difficulty in collecting both genetic and detailed clinical and follow-up data from patients receiving chemotherapy. Together, data presented in our study and 3 published studies14, 15, 16 provide converging evidence of TTN truncating variants as risk variants for CCM, particularly, anthracycline-related CCM. Additionally, we show evidence of a pathogenic LMNA variant in a patient with CCM, suggest the possibility of a role for a rare variant in RYR2 as risk modifiers for CCM and increased burden of rare nonsynonymous variants in 62 idiopathic cardiomyopathy genes in patients with anthracycline CCM compared to CCM patients treated with anti-HER2, nonanthracycline regimens.

The strongest evidence of genetic risk variants in established idiopathic cardiomyopathy genes is for truncating variants in TTN. We identified 2 TTN truncating variants in 18 patients (11%), specifically patients with anthracycline-related cardiomyopathy, slightly higher than that observed in a larger study of 213 patients with CCM, which reported 7.5%.¹⁴ Both variants mapped to the A-band, where the majority of truncating pathogenic variants associated with DCM have been reported,⁹ one of which has previously been identified as pathogenic in primary DCM.^9,23

There are some differences between our study and that of Garcia-Pavia et al such that our study, although smaller, offers additional and converging evidence for the role of TTN truncating variants in anthracycline-related CCM. Firstly, 90% of the 213 patients in the Garcia-Pavia study were treated with anthracyclines, whereas we report the frequency of TTN truncating variants in a smaller cohort in which 100% of patients were treated with anthracycline, which is perhaps the reason why our frequency was slightly higher. Secondly, we also examined genetic data from 37 patients treated with anthracyclines who did not present with CCM, and the Garcia-Pavia study examined The Cancer Genome Atlas participants and healthy volunteers as controls (for which cardiac outcome is unknown). Our finding of lack of observation of TTN truncating variants in patients treated with anthracycline who did not present with CCM is a strength, and our failure to observe TTN truncating variants in 71 patients (7 with CCM) treated with anti-HER2 therapy, without anthracyclines is further evidence of TTN truncating variants as modifiers specifically for anthracycline-related cardiomyopathy.

Interestingly, our study also identified a LMNA variant in one of 18 patients (5.5%) with anthracycline-related CCM, p.Arg190Gln. LMNA variants are present in 5.9% of idiopathic DCM¹⁰ and the same variant has previously been identified as pathogenic for familial DCM based on segregation with disease and additional in vitro functional validation.²⁵ In a larger published study,¹⁴ likely pathogenic LMNA variants were also identified in patients with CCM, but only TTN truncating variants were significantly enriched in CCM compared to controls, raising the question, would this patient have developed DCM regardless of anthracycline treatment? Our study and others are likely not statistically powered to answer this question unequivocally but are certainly of interest to a field which is increasingly moving toward predictive modeling in patient-derived-induced pluripotent stem cell-derived cardiomyocyte and cardiac organoid models.26, 27, 28 Given the cost and time invested in these models, genetic screening will be an important and necessary tool. Also of note, a recent study in human induced Pluripotent Stem Cell-derived cardiomyocytes with LMNA knockdown demonstrated reversal of a LMNA knockdown gene expression signature by pretreatment with the angiotensin receptor blocker, olmesartan.²⁹ In this model, olmesartan significantly elevated the expression of sarcomeric genes and rate and force of contraction and ameliorated arrhythmogenic potential, suggesting that perhaps, a strategy of genetic screening prior to anthracycline therapy could refine or improve patient-specific cardioprotective strategy.

A novelty of our study is the finding of a rare missense variant (p.Glu1127Gly) in RYR2 in 2 of 18 (11%) of patients with anthracycline CCM, the same frequency of TTN truncating variants, and a different RYR2 rare missense variant (p.Arg1084Lys) was also identified in a patient with a TTN truncating variant. Two of the 3 patients with rare RYR2 variants required cardiac transplantation. Although replication and further work is needed to determine the contribution of RYR2 missense variants to anthracycline CCM, in vitro and in vivo studies have demonstrated that doxorubicin is causative of calmodulin dissociation and calcium leakage from RYR2 and that pharmacological treatment with dantrolene improved cardiac function in a mouse model of doxorubicin-induced cardiotoxicity.³⁰

In summary, our findings are consistent with 3 published studies.14, 15, 16 Our observation of 2 TTN truncating variants in patients with CCM or HF and a pathogenic LMNA variant in a patient with CCM, attributed to anthracycline, strengthens current evidence of the implementation of genetic screening in this patient population, particularly for TTN truncating variants, and possibly for LMNA variants prior to chemotherapy. This finding alone yields a rate of return of 17% of patients with CCM having a pathogenic or likely pathogenic variant, although we acknowledge that due to small sample size and lack of statistical power, this finding may not be truly representative of the CCM population. Our study also provides very preliminary evidence for rare missense variants in RYR2 as potential risk variants of anthracycline-related cardiotoxicity, which can potentially be pharmacologically targeted and warrants further investigation of the broader group of known idiopathic cardiomyopathy genes.

Study Limitations

There are several caveats to our study design. Firstly, our sample size of 136 patients (55 treated with anthracycline, of which only 18 presenting with CCM) was small for meaningful genetic association studies or determination of multigenic effects and our findings are largely descriptive and may not be representative of the CCM population. Secondly, 39% of patients in the CCM group were treated with both anthracycline and trastuzumab, but 0% of control patients were treated with anthracycline and trastuzumab, which makes it difficult to determine whether the genetic contribution to CCM relate solely to anthracycline or anthracycline plus trastuzumab. Thirdly, patients presenting with HF in the anthracycline group presented many years beyond completion of cancer therapy, likely because we identified these patients retrospectively through our HF clinic and the follow-up time in our control group was much shorter, as these patients were enrolled to the study in a prospective approach. However, we believe our study provides valuable incremental data to move the field forward. Future directions are multidisciplinary evaluation of patients for CCM risk prediction incorporating genetic testing for DCM genes.

Funding support and author disclosures

This work is supported by grants from U.S. Department of Defense (W81XWH22-1-0288/289, PR210385) and National Heart, Lung, and Blood Institute (HL169268). The authors have reported that they have no relationships relevant to the contents of this paper to disclose.