Multicenter clinical and functional evidence reclassifies a recurrent noncanonical filamin C splice-altering variant

Abstract

Background: Truncating variants in filamin C (FLNC) can cause arrhythmogenic cardiomyopathy (ACM) through haploinsufficiency. Noncanonical splice-altering variants may contribute to this phenotype.

Objective: The purpose of this study was to investigate the clinical and functional consequences of a recurrent FLNC intronic variant of uncertain significance (VUS), c.970-4A>G.

Methods: Clinical data in 9 variant heterozygotes from 4 kindreds were obtained from 5 tertiary health care centers. We used in silico predictors and functional studies with peripheral blood and patient-specific induced pluripotent stem cell-derived cardiomyocytes (iPSC-CMs). Isolated RNA was studied by reverse transcription polymerase chain reaction. iPSC-CMs were further characterized at baseline and after nonsense-mediated decay (NMD) inhibition, using quantitative polymerase chain reaction (qPCR), RNA-sequencing, and cellular electrophysiology. American College of Medical Genetics and Genomics (ACMG) criteria were used to adjudicate variant pathogenicity.

Results: Variant heterozygotes displayed a spectrum of disease phenotypes, spanning from mild ventricular dysfunction with palpitations to severe ventricular arrhythmias requiring device shocks or progressive cardiomyopathy requiring heart transplantation. Consistent with in silico predictors, the c.970-4A>G FLNC variant activated a cryptic splice acceptor site, introducing a 3-bp insertion containing a premature termination codon. NMD inhibition upregulated aberrantly spliced transcripts by qPCR and RNA-sequencing. Patch clamp studies revealed irregular spontaneous action potentials, increased action potential duration, and increased sodium late current in proband-derived iPSC-CMs. These findings fulfilled multiple ACMG criteria for pathogenicity.

Conclusion: Clinical, in silico, and functional evidence support the prediction that the intronic c.970-4A>G VUS disrupts splicing and drives ACM, enabling reclassification from VUS to pathogenic.

Keywords: Arrhythmia; Arrhythmogenic cardiomyopathy; Functional genetics; Nonsense-mediated decay; Splicing.

Figure 1.. Family pedigrees.

A) 4-generation pedigree of variant heterozygotes cared for by USA centers. All siblings of proband generation are heterozygous for the variant and have a history of ACM. The mother died at 42 of sudden death following the diagnosis of an unknown arrhythmia, and the maternal grandfather suddenly died at the age of 60 one year after a negative ischemic evaluation. B) 3-generation pedigree of a young proband followed at an Australian center. The index patient had early onset DCM requiring heart transplant. Her mother and sister were both negative for the variant. Family history was otherwise remarkable for both paternal and maternal grandfathers with diagnosed DCM. C) 3-generation pedigree from Netherlands. The proband presented later in life with severe DCM. The mother and maternal grandmother both died suddenly. A maternal aunt and child were variant heterozygotes in apparently normal health, who were unable to be contacted for additional phenotyping. D) 2-generation pedigree of additional family from the Netherlands. The proband presented in mid-adulthood with DCM. Both parents had a history of cardiomyopathy, but further were precluded by estrangement from his family. + indicates genetic testing and heterozygous status for c.970–4A>G. – indicates wildtype status. • indicates obligate heterozygote. + or – absent when genetic testing was not completed. Arrow indicates proband within each family. Dash line across relationships indicates estrangement.

Figure 2.. Exercise stress test of USA proband.

A) Proband ECG (Family 1) at rest prior to exercise. Late-coupled PVCs (3 morphologies). B) Proband ECG at Stage 3 of stress test with an 8-beat episode of relatively monomorphic NSVT.

Figure 3.. FLNC c.970–4A>G Variant Characteristics.

A) SpliceAI scores for the FLNC c.970–4A>G. Strongly predictive of the introduction of a competing splice donor site immediately adjacent to the variant. B) Schematic of the functional consequences for variant heterozygotes. Competing splice site introduces a 3-bp premature termination codon into the aberrantly spliced exon. C) Sanger trace of cDNA analysis of peripheral blood obtained from the proband of Family C (The Netherlands I). Trace shows introduction of the predicted premature termination codon in one transcript.

Figure 4.. FLNC c.970–4A>G iPSC-CM transcriptional investigations.

A) iPSC-CM RNA was reverse transcribed into cDNA and amplified by PCR. Gel shows amplified DNA from a healthy control (Pop. Con.), USA proband, and a negative control (Neg. Con.) without RNA input to RT-PCR. B) Sanger sequencing result of the gel extracted band from USA proband iPSC-CM RT-PCR product. Sequence shows only normal splicing of exon 5 to exon 6. C) Chemical structure of cycloheximide, a natural product inhibitor of nonsense-mediated decay. D) Sanger sequencing of RT-PCR products from patient and control lines treated with DMSO or 100 μM cycloheximide (CHX). E) TA cloning of gel extracted bands from 100 μM CHX treatment in panel E. Identification of aberrantly spliced transcript degraded in absence of NMD inhibition. F) Relative fold change of transcript abundance as determined by qPCR. Two-tailed t-test. G) Sashimi plot visualization of RNA-seq data in patient cell line. Low level indel incorporation is observed after vehicle treatment, which is markedly upregulated after inhibition of NMD.

Figure 5.. Spontaneous action potentials of FLNC c.970–4A>G proband and healthy population control derived iPSC-CMs.

A) Example spontaneous action potentials measured in control iPSC-CMs. B) Example spontaneous action potentials in FLNC proband iPSC-CMs.