Filamin C Cardiomyopathy Variants Cause Protein and Lysosome Accumulation
Dominant heterozygous variants in filamin C (FLNC) cause diverse cardiomyopathies, although the underlying molecular mechanisms remain poorly understood.
We aimed to define the molecular mechanisms by which FLNC variants altered human cardiomyocyte gene and protein expression, sarcomere structure, and contractile performance.
Methods and Results:
Using CRISPR/Cas9, we introduced FLNC variants into human induced pluripotent stem cell–derived cardiomyocytes (hiPSC-CMs). We compared isogenic hiPSC-CMs with normal (wild-type), ablated expression (FLNC−/−), or haploinsufficiency (FLNC+/−) that causes dilated cardiomyopathy. We also studied a heterozygous in-frame deletion (FLNC+/Δ7aa) which did not affect FLNC expression but caused aggregate formation, similar to FLNC variants associated with hypertrophic cardiomyopathy. FLNC−/− hiPSC-CMs demonstrated profound sarcomere misassembly and reduced contractility. Although sarcomere formation and function were unaffected in FLNC+/− and FLNC+/Δ7aa hiPSC-CMs, these heterozygous variants caused increases in lysosome content, enhancement of autophagic flux, and accumulation of FLNC-binding partners and Z-disc proteins.
FLNC expression is required for sarcomere organization and physiological function. Variants that produce misfolded FLNC proteins cause the accumulation of FLNC and FLNC-binding partners which leads to increased lysosome expression and activation of autophagic pathways. Surprisingly, similar pathways were activated in FLNC haploinsufficient hiPSC-CMs, likely initiated by the loss of stoichiometric FLNC protein interactions and impaired turnover of proteins at the Z-disc. These results indicate that both FLNC haploinsufficient variants and variants that produce misfolded FLNC protein cause disease by similar proteotoxic mechanisms and indicate the therapeutic potential for augmenting protein degradative pathways to treat a wide range of FLNC-related cardiomyopathies.
Meet the First Author, see p 701
Cardiomyopathy, characterized by pathological cardiac enlargement and contractile dysfunction, is a leading cause of heart failure. Among patients with nonischemic cardiomyopathy, a cause is identified in over 20%, including damaging variants (loss of function, in-frame deletions/insertions, and deleterious missense) in over 40 genes with diverse functions in cardiomyocyte biology.1,2 These variants cause diverse forms of cardiac remodeling which are clinically subclassified into hypertrophic cardiomyopathy (HCM), dilated cardiomyopathy (DCM), and restrictive cardiomyopathies, among others. A critical next step to improve the understanding of human heart failure is to define shared and distinct mechanisms by which these variants cause disease.
Variants in FLNC, encoding filamin C, have emerged as a more common cause of cardiomyopathy.3–6FLNC truncating variants are found in ≈3% to 4% of patients with DCM which presents in early-to-mid adulthood and is associated with a high rate of ventricular arrhythmias and sudden cardiac death.6–9FLNC missense variants are more commonly associated with HCM and can cause pathological FLNC protein aggregation.9 However, there is considerable overlap between the cardiomyopathy phenotypes caused by distinct classes of FLNC variants, with missense, splice site, and in-frame indels reported in DCM, HCM, and restrictive cardiomyopathy.3,9,10
Filamins (FLNA, FLNB, FLNC) are dimeric proteins, consisting of 270 to 290 kDa subunits, that crosslink filamentous (F−) actin.11,12 FLNs contain an N-terminal ABD (actin-binding domain) followed by 24 immunoglobulin-like repeats, the last of which mediates protein dimerization.13,14 FLNC (Figure 1) is uniquely and highly expressed in cardiac and skeletal muscle15,16 and has an immunoglobulin-like domain insertion that enables sarcomere binding at Z-discs, where it crosslinks actin to anchor thin filaments.16 FLNC is less abundantly found at costameres (specialized focal adhesion complexes that connect sarcomeres to the sarcolemma) and at intercalated discs.12,17,18 At these locations and at Z-discs, the C-terminal immunoglobulin domains (immunoglobulin 16–24) of FLNC interact with a large number of proteins with diverse structural and signaling roles, including the dystrophin-glycoprotein-complex, SGCG (γ-sarcoglycan),12 ITGB1 (β1-integrin tail),17 SORBS1 (costamere-associated protein),19 FBLIM1 (migfilin),20 PGM5 (aciculin),21 XIRPs (xin proteins; XIRP1, XIRP2),21 SYNPO2 (synaptopodin 2),22 MYOT (myotilin),16 MYOZs (myozenins; MYOZ1, MYOZ2),23 TTN (titin),24 and HSP (heat shock proteins; HSPB1,18 HSPB725).
Homozygous ablation of Flnc in mice, flies, and fish produces abnormal Z-disc morphology, compromised sarcomere integrity, and striated muscle dysfunction.24,26–28 However, heterozygous FLNC variants in these model systems do not show overt striated muscle phenotypes, and thus their utility in understanding human cardiomyopathy pathogenesis has remained limited. A related disease, myofibrillar myopathy (MFM), is also caused by truncating FLNC variants and shows large protein aggregates composed of FLNC and its binding partners in skeletal muscle.29,30 Although some patients with MFM also have DCM the converse is not true—the vast majority of DCM patients with pathogenic FLNC variants lack skeletal muscle involvement. Moreover, large protein aggregates are not reported in cardiac biopsies from FLNC-DCM patients.7,10 As such, the relatedness of pathogenic mechanisms involved in FLNC myofibrillar myopathy and FLNC cardiomyopathy remains unclear.
Here, we explored the mechanism of FLNC cardiomyopathy by employing CRISPR/Cas9 methodologies to produce and study a series of damaging FLNC variants in human induced pluripotent stem cell–derived cardiomyocytes (hiPSC-CMs). Unexpectedly, our studies reveal that homozygous and heterozygous FLNC variants have largely distinct consequences. From analyses of transcriptomic, proteomic, and live-cell imaging, we define a critical role for FLNC expression in sarcomere assembly and demonstrate proteopathic effects that ensue from heterozygous FLNC variants.
Culture and Maintenance of hiPSCs
PGP1 hiPSCs (derived from PGP1 donor from the Personal Genome Project) with an endogenous green florescent protein (GFP) tag on the sarcomere gene TTN31 were grown in Matrigel-coated 6-well tissue culture dishes, maintained in stem cell maintenance media (mTESR1, STEMCELL Technologies), and passaged every 2 to 3 days (passage range, 55–70) in the presence of 5 μM ROCK (Rho kinase) inhibitor Y-27632 (125410; R&D Systems).
CRISPR/Cas9-Mediated Gene Editing of hiPSCs
The Benchling design tool was used for guideRNA and homology-directed repair arm design. Adherent hiPSCs were dissociated into single cells from the tissue culture plate using accutase, and nucleofected with 2 μg of Cas9 plasmid (pSpCas9[BB]-2A-Puro [PX459] V2.0, no. 62988; Addgene Plasmid), 4 μg of guideRNA pCR-BluntII-TOPO plasmid, and 4 μg of homology-directed repair arm using a Lonza nucleofector, as previously described.32 Edited colonies were subcloned and resequenced by MiSeq to confirm editing and population clonality.
Differentiation of hiPSC-CMs
hiPSCs seeded in Matrigel-coated 6-well culture plates were grown to 90% confluency for hiPSC-CM differentiation.33 hiPSC-CMs were enriched by glucose deprivation at differentiation day 11 to 17. Unless indicated, all hiPSC-CMs were studied on or after day 30 to maximize cardiomyocyte maturity (eg, MYH7/MYH6 [myosin heavy chain] proteins=0.97±0.02).34
Replating hiPSC-CMs for Imaging
hiPSC-CMs were dissociated and replated onto Matrigel-coated glass-bottom imaging-optimized 12-well plates (Mat-Tek) as described.35
Contractility Analysis of hiPSC-CMs
Replated hiPSC-CMs were imaged as described35 using a ×100 objective of a fluorescence microscope (Keyence BZ X-710) at an acquisition rate of 30 frames per second; sarcomere lengths and percent of sarcomere contraction were quantified using SarcTrack software.36
Immunofluorescent Analyses of iPSC-CMs
Replated hiPSC-CMs were fixed, permeabilized, and stained with an F-actin molecular probe, FLNC (1:100, abcam Rb mAb, ab180941), TFEB (1:500, Cell Signaling Technology [CST], Rb mAb 37785), or NucBlue before imaging with a ×100 oil objective. For all immunofluorescent analyses, secondary antibody-only controls (abcam Rb [rabbit] monoclonalantibody ab150113, ab150116, ab150077, ab150080) were used to distinguish genuine target staining from background. Sarcomere content, anisotropy, and sarcomere persistence were quantified using custom MatLab algorithms (Extended Methods in the Data Supplement).
Patterning of hiPSC-CMs and Quantification of Sarcomere Striation
hiPSC-CMs were replated onto 2000 μm2 fibronectin-coated rectangular polydimethylsiloxane micropatterns.37 At 2 and 8 days after patterning, hiPSC-CMs were fixed, permeabilized, and stained with NucBlue and paxillin (610569, 1:100 dilution; BD Biosciences), and imaged using a ×100 oil objective.
Live Imaging of Lysosomes in hiPSC-CMs
hiPSC-CMs were stained using 75 nM LysoTracker probe (LysoTracker Red DND-99, Thermo Fisher Scientific L7528) in RPMI B27 plus insulin media. Videos were acquired in an incubation chamber (37 °C, 5% CO2) using a widefield fluorescence microscope (Keyence BZ X-710; ×100 objective, 30 frames/s). Lysosomes were also imaged using a ×100 objective of a spinning disk confocal microscope. Image data were processed using Fiji software.
Flow Cytometry of eGFP-TTN Expression in hiPSC-CMs
hiPSC-CMs were dissociated, strained through a 100 µm cell strainer, and eGFP signal was quantified with a FACSCanto flow cytometer. hiPSC-CMs without the eGFP-TTN tag were analyzed to set appropriate gates for the identification of GFP-TTN hiPSC-CMs. The resultant GFP (FITC-A) intensity frequency distributions were plotted using FlowJo software.
RNA Sequencing of hiPSC-CMs
Cells were lysed in Trizol Reagent (Life Technologies) for RNA extraction and stored at −80 °C until RNA was extracted. All RNA quality (RIN) and quantity were assessed on the TapeStation 2200 (Agilent) and all samples had an RNA integrity number (RIN) of >8. Library preparation was conducted using the Nextera library preparation method (Nextera XT DNA Library Preparation Kit, Illumina) and sequenced (Illumina NextSeq500 platform). All data were combined into a single fastq file. Samples typically had 30M to 50M reads each. Seventy-five base pair-end reads were aligned to the human reference genome hg38 using Spliced Transcripts Alignment to a Reference. Raw reads were normalized to the total number of reads per kilobase of transcript per million.
Protein Isolation From Cell Lysates
hiPSC-CMs were washed and lysed with cold RIPA buffer containing protease and phosphatase inhibitors (Sigma-Aldrich 11873580001, Sigma-Aldrich 4906837001). To ensure lysis of nuclei, the lysate was passed 20× through a 21G needle, then shaken gently on ice for 15 minutes, and centrifuged (14 000g for 15 minutes). The supernatant was transferred to a fresh tube and protein concentration was quantified using the bicinchoninic assay kit (Pierce 23227).
For each sample, 12.5 μg (or 20 μg for soluble/insoluble experiments) of protein lysate was loaded and resolved under denaturing conditions on a NuPAGE 3% to 8% Tris-Acetate Protein Gel (or 4%–20% Tris-Glycine Gel for soluble/insoluble and autophagic flux experiments). Proteins were wet transferred to a polyvinylidene fluoride (PVDF) membrane overnight at 40 mA at 4 °C or 70 V for 3 hours. Membrane was blocked for 1 hour at room temperature in 5% skim milk in TBST buffer, before addition of primary and secondary antibodies (Expanded Methods), and then imaged on an LI-COR Biosciences Odyssey CLx (infrared) or ChemiDoc BioRad imaging system.
Tandem-Mass-Tag Proteomics of hiPSC-CMs
Equal amounts of protein (≈100 μg) from each sample were digested into peptides for tandem-mass-tag quantitative proteome profiling as described.38 Peptides were tagged with unique, sample-specific isobaric chemical labels (tandem-mass-tags) for multiplexing, and data were collected on an Orbitrap Fusion mass spectrometer in line with a Proxeon NanolC-1200 UHPLC. Database searching and reporter-ion quantification was performed using an in-house SEQUEST-based pipeline. To define relative abundance measurements, each protein was scaled such that the summed signal-to-noise for that protein across all channels was 100. Intensity-based absolute quantification values were used39 to estimate absolute protein abundance using mass spectra processed using MaxQuant (Version 184.108.40.206). Otherwise, mass spectra were processed using a SEQUEST-based pipeline.38 Protein quantification values were exported for further analysis in R (histograms, heatmaps, scatterplots) and bar graphs (GraphPad Prism).
Soluble/Insoluble Protein Fractionation of hiPSC-CMs
hiPSC-CMs protein lysates were fractionated into soluble and insoluble fractions using previously described methods.40 Cells were washed with PBS and then lysed in ice-cold CellLytic M buffer (Sigma) containing protease inhibitors (Roche) for 30 minutes. The cells were centrifuged at 12 000g for 15 minutes and supernatants were collected, generating the soluble fraction. The pellets were dissolved in DNAse I (Roche), homogenized, and gently sonicated for 4 minutes, generating the insoluble fraction. Insoluble fractions were then used for Western blotting or mass spectrometry.
Autophagic Flux Analyses
hiPSC-CMs were washed 1X with PBS and cultured in either RPMI B27 plus insulin (fed) or DMEM minus glucose (starved) for 4 hours. Cells were treated with and without 400nM of bafilomycin A1 for 4 hours. Western blotting was used to determine the expression of LC3b-I, LC3b-II, (microtubule associated protein 1 light chain 3 beta) and TUBB (β-tubulin), and densitometric quantitation was performed using the ChemiDoc BioRad imaging system. The absolute difference in the LC3-IIb/TUBB ratio between bafilomycin-treated and untreated cells was used to compute autophagic flux.
Statistical analyses were performed using nonparametric tests (Wilcoxon rank-sum tests), or Student t tests (2-sided, assuming similar variances) when the dataset was large and normally distributed. P<0.05 was considered statistically significant. For transcriptome and proteome-wide analyses, Benjamini-Hochberg corrections were applied for multiple comparisons with an α=0.05 after restricting the analyses by fold-change, as specified in the main text. To assess the effects of genotype on levels of particular sets of proteins (including endoplasmic reticulum [ER] proteins, autophagy proteins, FLNC interactors, and lysosome protein levels), linear mixed models were fitted with protein concentrations as the outcome variable and the interaction of genotype and marker protein as the fixed effect. To allow for correlated measurements between experiments, the replicate number was used as a random variable, and χ2 values and P values are reported. For differential expression analyses, fold-change cutoffs were determined by varying the fold-change criteria and selecting a threshold at the inflection point of the resultant graph, such that the fold-change cutoff was neither too permissive nor too restrictive. DAVID functional annotation clustering was used to conduct enrichment analyses and the topmost significant functional categories were selected for display.41
For all figures that include representative images, the individual sample that best represented the group mean was selected.
CRISPR/Cas9 Genome Engineering of FLNC Mutant hiPSC Lines
FLNC variants were introduced via CRISPR/Cas9 into the hiPSC line PGP1, which contains an endogenous GFP tag on the N-terminus of the sarcomere protein titin (eGFP-TTN). We targeted exon 30, a region that encodes the 15th immunoglobulin domain (Figure 2A), and produced isogenic hiPSCs with wild-type (WT) or mutated FLNC sequences. To model human DCM variants, we introduced a heterozygous truncating variant (FLNC+/−, Table). For comparison, we also produced cells that carry this variant on both alleles (FLNC−/−, Table).
|cell line||Coding sequence Δ||predicted protein sequence Δ||FLNC RNA (RPKM), (n)||FLNC protein (%WT), (n)|
|WT||WT||WT||283 (3)||100 (3)|
|FLNC−/−||c.5021_5022insT||p.Gly1674X||20* (3)||6±0.5 (4)|
|FLNC+/−||c.5021_5022insT||p.Gly1674X||93* (2)||47±5 (3)|
|FLNC+/Δ7aa||c.5002_5022del21||p.Val1668_Gly1674del||324 (2)||90±21 (2)|
|FLNCG1674S/G1674S||c.5020G>A||p.Gly1674Ser||249 (1)||111±10 (3)|
Spurious CRISPR mutagenesis also resulted in a heterozygous in-frame deletion of amino acids 1668-1674 (FLNC+/Δ7aa, Table). As FLNC in-frame insertion/deletions occur in patients with restrictive, hypertrophic, and dilated cardiomyopathies,3,10 we included FLNC+/Δ7aa in our studies, recognizing that although not identified in patients, this variant had utility in understanding the fundamental biological mechanisms of mutant alleles that do not cause FLNC haploinsufficiency. We also produced and studied FLNCG1674S/G1674S cells. At the time, this study was initiated FLNC G1674S was a variant of unknown significance but has since been reclassified as likely benign given its high frequency (1.35×10-4) in the gnomAD (Genome Aggregation Database) database.42 As such, FLNCG1674S/G1674S hiPSC-CMs, which are isogenic to parental PGP1-derived CMs, serve as a control cell line.
Genotypes were confirmed with next-generation sequencing (Figure I in the Data Supplement). All lines differentiated into beating hiPSC-CMs, and principal component analyses of RNA sequencing profiles (Data Set I and Figure IIA in the Data Supplement) showed comparable degrees of maturity amongst all lines after 30 days of differentiation, with robust expression of prototypic cardiomyocyte genes, including PLN, SERCA2A, and TTN (Table I in the Data Supplement).
Using tandem-mass-tag quantitative mass spectrometry-based proteomics, we studied replicates of WT and FLNC mutant hiPSC-CMs, where each replicate represents total cellular protein extracted from an independent differentiation. In an 8-plex experiment, we quantified the relative expression levels of 7123 proteins from WT (n=4) and FLNC−/− (n=4) hiPSC-CMs (Data Set II in the Data Supplement), and in an 11-plex experiment, we quantified 6863 proteins from WT (n=3), FLNC+/− (n=3), FLNC+/Δ7aa (n=2), and FLNCG1674S/G1674S (n=3) hiPSC-CMs (Data Set III in the Data Supplement). Hierarchical clustering analyses of protein expression data from the 8-plex and 11-plex tandem-mass-tag experiments showed appropriate clustering by genotype (Figure IIB and IIC in the Data Supplement).
Mass spectrometry of WT hiPSC-CMs identified over 100 FLNC peptides, which was significantly greater than the average number (≈8) of mapped peptides per protein, hinting that FLNC was abundantly expressed in hiPSC-CMs (Data Set IV in the Data Supplement). Intensity-based absolute quantification43 estimated that FLNC was more abundant than 97.8% of proteins identified in hiPSC-CMs (Data Set IV in the Data Supplement).
Analysis of FLNC expression in WT and mutant hiPSC-CMs (Table) showed that FLNC−/− and FLNC+/− hiPSC-CMs had markedly reduced levels of FLNC transcript (7% and 33% of WT levels), likely reflecting mRNA nonsense-mediated decay. Mass spectrometry confirmed that FLNC protein levels were similarly reduced to 6% and 47% of WT, respectively (Table). By contrast, FLNC+/Δ7aa and FLNCG1674S/G1674S hiPSC-CMs had FLNC transcript and protein levels comparable to WT cells (Table). Western blotting using N-terminal and C-terminal FLNC antibodies confirmed these findings (Figure 2B).
FLNC Is Necessary for Actin and Thin Filament Gene Expression and Sarcomere Assembly
Proteome-wide comparisons of WT and FLNC−/− hiPSC-CMs (7123 total proteins) revealed 136 proteins that were differentially expressed (DE; Data Set II in the Data Supplement). Analyses of these DE proteins showed significant enrichment of actin-binding and sarcomere proteins (Table II in the Data Supplement). Parallel analyses of RNA sequencing data indicated that these decreased protein levels were largely attributable to reduced transcription, as there was a strong correlation between DE protein and transcript expression in FLNC−/− hiPSC-CMs (r=0.91, Figure 3A). Included among these DE genes were the FLNC-binding partner synaptopodin (SYNPO), thin filament-associated sarcomere components (tropomyosin (TPM2), troponin I (TNNI3), and troponin T (TNNT1)), as well as Z-disc components (synaptopodin 2-like [SYNPO2L], ankyrin repeat domain 1 [ANKRD1], PDZ and LIM domain 3 [PDLIM3], leomodin 2 [LMOD2], cysteine and glycine-rich protein 3 [CSRP3]; Figure 3A). By contrast, proteins of the thick filament and M-band (TTN, MYBPC3, [myosin binding protein C] nebulette, MYOM1, MYOM2) were only minimally (>75% of WT level) or not significantly affected, although the ratios of MYH6/MYH7 and MYL7/MYL2 (myosin light chain) transcripts were >5-fold increased, reflecting transcriptional isoform switches that are frequently observed in states of cardiomyocyte stress44 (Data Set I and II in the Data Supplement).
Immunofluorescent analyses of actin and sarcomere structure in hiPSC-CMs (Figure 3B) showed that FLNC−/− hiPSC-CMs had significantly reduced sarcomere content (Figure 3C; mass spectrometry and flow cytometry showed comparable expression of eGFP-TTN in WT and FLNC−/− hiPSC-CMs (Figure IIIA and IIIB in the Data Supplement), thereby excluding an artifactual effect of the tagged TTN protein). FLNC−/− hiPSC-CMs also had disordered actin cytoskeletons (Figure 3B and 3D). Examination of individual replated cells also showed aberrant actin crosslinking in FLNC−/− hiPSC-CMs demonstrated by the marked size variation and breakage of actin bundles (Figure IIIC in the Data Supplement).
In our prior work, we demonstrated that sarcomere assembly and organization can often be promoted by culturing hiPSC-CMs on surfaces with micropatterned fibronectin matrices, which support the alignment of costameres with the ECM (extracellular matrix).37 However, even when plated on micropatterned surfaces, FLNC−/− hiPSC-CMs had reduced sarcomere content (Figure 3E through 3G, Figure IIID in the Data Supplement). Although these surfaces improved sarcomere persistence, a metric of myofibril continuity, in both WT and FLNC−/− hiPSC-CMs, myofibrils in FLNC−/− hiPSC-CMs still remained less persistent compared with WT at both 2 and 8 days postreplating (Figure 3E through 3G). These studies indicate that even in the presence of an organized ECM, FLNC−/− hiPSC-CMs are unable to properly assemble and organize sarcomeres.
Despite severe deficits in sarcomere assembly, residual sarcomeres in FLNC−/− hiPSC-CMs were still contractile (Movies I and II in the Data Supplement). However, these sarcomeres were significantly less contractile relative to WT, as assessed by high-throughput analysis of thousands of sarcomeres36 (Figure 4, Data Set V in the Data Supplement). Many FLNC−/− sarcomeres exhibited a conspicuous bending motion during cardiomyocyte relaxation (Movie III in the Data Supplement), a phase of the cardiac cycle when tropomyosin undergoes azimuthal movements to reposition it close to actin where electrostatic interactions maintain tropomyosin in a blocked state.45 We suspect that inadequate actin crosslinking in FLNC−/− hiPSC-CMs disrupted these movements and resulted in relaxation-dependent bending.
Taken together, these studies of FLNC−/− hiPSC-CMs demonstrate multiple essential roles of FLNC in sarcomere biology including expression of a network of actin and thin filament genes, actin crosslinking at Z-discs, and maintenance of sarcomere structure during relaxation.
Distinct Abnormalities in hiPSC-CMs With Damaging Heterozygous FLNC Variants
Having identified impaired sarcomere structure, contractility, and relaxation in FLNC−/− hiPSC-CMs, we expected similar defects in the heterozygous mutant lines. However, neither FLNC+/−, FLNC+/Δ7aa nor FLNCG1674S/G1674S had significant changes in sarcomere structure or contractile performance in vitro (Figure 4, Movies IV through VI in the Data Supplement). Furthermore, unlike FLNC−/− hiPSC-CMs, the expression levels of most thin filament (TNNC1, TNNT2, TPM2) and thick filament (TTN, MYH6, MYH7, MYBPC3) proteins in these three lines were comparable to WT (Data Sets II and III in the Data Supplement).
To assess for other potentially pathogenic effects of these variants, we compared the entire proteomes (6863 proteins) of FLNC+/−, FLNC+/Δ7aa, FLNCG1674S/G1674S and WT hiPSC-CMs. These proteome-wide analyses showed that FLNCG1674S/G1674S was indistinguishable from WT (Figure 5A), confirming G1674S as a benign variant and allowing its use as an additional independent control line. By contrast, FLNC+/− and FLNC+/Δ7aa hiPSC-CMs had 212 and 568 DE proteins relative to WT (Data Set III in the Data Supplement); >50% of these DE proteins were shared by both FLNC+/− and FLNC+/Δ7aa hiPSC-CMs, suggesting overlapping pathogenic effects (Figure 5A). The identity and expression patterns of DE proteins in FLNC+/− and FLNC+/Δ7aa hiPSC-CMs were largely distinct from those of FLNC−/− hiPSC-CMs (Figure IVA in the Data Supplement), further implying that FLNC heterozygous alleles evoked different consequences than those caused by complete absence of FLNC.
Among the 350 DE proteins shared between the FLNC+/− and FLNC+/Δ7aa hiPSC-CMs, there was significant enrichment for lysosomal proteins (Table III in the Data Supplement, Figure 5B). Levels of lysosomal membrane proteins, as well as proteins that mediate lysosomal acidification, and lysosomal enzymes including proteases, glycosidases, sulfatases, phospholipases, and ceramidases were increased (Figure 5B). Neither FLNC−/− nor FLNCG1674S/G1674S hiPSC-CMs showed large changes in lysosomal protein expression (Figure IVB in the Data Supplement).
Additional enrichment was observed for proteins containing signal peptides, disulfide bonds, glycoproteins, and extracellular proteins in both FLNC+/− and FLNC+/Δ7aa hiPSC-CMs (Table III in the Data Supplement, Figure 5C). As these terms reflect features of proteins that are destined for membrane incorporation or secretion,46 we hypothesized that the increased abundance of these proteins within the cell could reflect ER stress and dysfunctional protein trafficking. Consistent with this hypothesis, FLNC+/− and FLNC+/Δ7aa hiPSC-CMs had increased protein levels of ER stress markers including PERK (PRK-like ER kinase), CHOP (C/EBP homologous protein 10), PDI (prolyl 4-hydroxylase subunit beta), BiP/GRP78 (endoplasmic reticulum chaperone), and ERp72 (protein disulfide isomerase family A member 4), which infers activation of the PERK and ATF6 arms of the unfolded protein response (Figure 5D).47 Involvement of the IRE-1 (inositol-requiring enzyme) arm of the unfolded protein response appeared minimal, as splicing of the XBP1 mRNA48 was not observed in either the WT or FLNC heterozygous cell lines (Figure IVC in the Data Supplement).
The levels of most lysosomal proteins, trafficked proteins, and ER stress markers were even higher in FLNC+/Δ7aa than in FLNC+/− hiPSC-CMs (Figure 5B through 5D), inferring that the Δ7aa allele caused a similar, but more severe, cellular phenotype than that of the haploinsufficient allele.
Increased Lysosome Content in FLNC+/− and FLNC+/Δ7aa hiPSC-CMs
We considered whether increased lysosomal protein levels in FLNC+/− and FLNC+/Δ7aa hiPSC-CMs reflected changes in organelle number or size. Using LysoTracker, a red fluorescent membrane-permeable dye that accumulates in acidic organelles, we directly visualized lysosomes in WT, FLNC+/−, and FLNC+/Δ7aa hiPSC-CMs (Figure 5E). Live-cell imaging showed active movement of Lysotracker-stained structures along linear paths in all cell lines (Movies VII through IX in the Data Supplement), indicating that these were appropriately acidic and dynamically motile along microtubules.49 Quantification of LysoTracker staining demonstrated that LysoTracker structures were significantly larger in FLNC+/− hiPSC-CMs and FLNC+/Δ7aa (Figure 5F), as well as more numerous in FLNC+/Δ7aa relative to WT hiPSC-CMs (Figure 5G). Although the RNA and protein levels of TFEB (transcription factor EB), a master regulator of lysosomal biogenesis,50 were not significantly changed in FLNC+/− or FLNC+/Δ7aa hiPSC-CMs (Figure VA in the Data Supplement), the proportion of nuclear to cytoplasmic TFEB was increased in FLNC+/Δ7aa hiPSC-CMs relative to WT (Figure VB and VC in the Data Supplement), indicating TFEB activation.51,52 Additionally, genes containing a Coordinated Lysosomal Expression and Regulation motif through which TFEB activates transcription50 had modest but significantly increased transcription in FLNC+/− and FLNC+/Δ7aa hiPSC-CMs (Figure VD, Data Set I in the Data Supplement), suggesting that increased TFEB activity may have partially contributed to the increased expression of lysosomal genes, particularly in FLNC+/Δ7aa hiPSC-CMs, although additional mechanisms are likely involved.
Increased Autophagic Flux and Altered Z-Disc Protein Handling in FLNC+/− and FLNC+/Δ7aa hiPSC-CMs
The increase in lysosome expression and abundance in FLNC+/− and FLNC+/Δ7aa hiPSC-CMs suggested that autophagic pathways could be activated in these mutants. Consistent with this hypothesis, the levels of several autophagic proteins, including those involved in early phagophore formation (BECN1 [beclin-1], ATG13 [autophagy-related gene]), autophagosome formation (ATG2B, MAP1LC3A [microtubule associated protein 1 light chain 3 alpha], ATG7, ATG12, ATG5, ATG16L1), and regulators of autophagy including PRKACA and PRKACB53–55 were found in lower abundance in FLNC+/− and FLNC+/Δ7aa compared with WT and FLNCG1674S/G1674S hiPSC-CMs (Figure 6A, Data Set III in the Data Supplement). We hypothesized that the depletion of these proteins reflected their consumption due to increased autophagic flux.
To measure autophagic flux, we assessed LC3b phosphatidylethanolamine conjugation in hiPSC-CMs cultured in both full and starvation (glucose-depleted) media (Figure 6B and 6C). Autophagic flux increased by genotype from WT to FLNC+/− to FLNC+/Δ7aa hiPSC-CMs. This increase in flux by genotype was observed under both fed and starved conditions (Figure 6C). As autophagic flux was not significantly further upregulated by starvation in WT hiPSC-CMs, it is likely that 4 hours of glucose depletion was an insufficient stressor to result in a further increase in flux; nevertheless, the effect of genotype on flux is significant regardless of condition.
We hypothesized that the increase in lysosome content and autophagic flux in FLNC+/− and FLNC+/Δ7aa hiPSC-CMs was triggered by protein aggregation of FLNC or other FLNC-binding partners. To test this hypothesis, total cellular protein from WT and FLNC mutant hiPSC-CMs was fractionated into soluble and insoluble fractions, and the insoluble fractions were subjected to analysis by mass spectrometry (Data Set VI in the Data Supplement). Analysis of FLNC-binding partners and associated Z-disc proteins showed that these proteins were enriched in the insoluble fractions of FLNC+/Δ7aa hiPSC-CMs relative to WT (Figure 6D). The proteins with the greatest enrichment in FLNC+/Δ7aa hiPSC-CMs included FLNC, striated muscle actin (ACTA1), PDLIM3 (PDZ and LIM Domain 3), DES (desmin), FILIP1 (FLNA interacting protein 1), and migfilin/FBLIM1 (FLN-binding LIM protein 1). The enrichment of these proteins in the insoluble fraction indicated that the FLNCΔ7aa protein formed aggregates of FLNC and other FLNC-binding partners, similar to those that form with FLNC HCM missense variants.9 Western blotting further confirmed the enrichment of FLNC and DES in the insoluble fractions of FLNC+/Δ7aa hiPSC-CMs (Figure 6E). In contrast, FLNC+/− hiPSC-CMs did not show significant enrichment of FLNC or its binding partners in the insoluble fraction relative to WT (Figure 6D and 6E; although FLNC+/− also carry only 50% of the amount of total FLNC protein).
Despite lacking enrichment of FLNC and FLNC-binding partners in the insoluble protein fraction, further analyses of total cellular lysate from FLNC+/− hiPSC-CMs showed significant accumulation of FLNC-binding partners and Z-disc proteins relative to WT (Figure VI, Data Set III in the Data Supplement); this accumulation was even greater in FLNC+/Δ7aa hiPSC-CMs compared to FLNC+/− hiPSC-CMs. Sarcomere proteins that were not part of the Z-disc including those of the thin (eg, TNNC1, TPM2) and thick filament (eg, MYH6, MYH7, MYBPC3) did not show these effects (Figure VI in the Data Supplement). These data suggest that Z-disc protein turnover is modestly affected in FLNC+/− hiPSC-CMs, and more severely so in FLNC+/Δ7aa hiPSC-CMs which form aggregates of FLNC and FLNC-binding proteins.
We demonstrate that FLNC is critical for sarcomere assembly and define the effects of damaging FLNC variants on cardiomyocyte biology (Figure 7). Complete loss of FLNC expression (FLNC−/−) disrupted the coordinated expression of actin and thin filament genes, which reduced sarcomere content and impaired sarcomere function. FLNC haploinsufficiency (FLNC+/−), a cause of DCM, did not disrupt sarcomere assembly and performance but rather impaired cardiomyocyte proteostasis, as reflected by the presence of ER stress, increased lysosome expression, and enhanced autophagic flux. Even more profound abnormalities in proteostasis occurred with an in-frame deletion mutant encoding a misfolded FLNC protein (FLNC+/Δ7aa), which formed aggregates akin to those that accrue with damaging FLNC HCM missense variants.9
The severe consequences of FLNC ablation on sarcomere structure in hiPSC-CMs are consistent with observations in other organisms and partially explained by the known role of FLNC in crosslinking actin at Z-discs.26,27 Surprisingly, FLNC−/− hiPSC-CMs also failed to transcribe (and translate) an entire network of sarcomere and actin-associated genes, including SYNPO, TPM2, TNNI3, TNNT1, SYNPO2L, ANKRD1, PDLIM3, LMOD2, and CSRP3. The disruption of RNA transcription of these genes indicates that FLNC has underappreciated roles in the coordinated regulation of the thin filament gene network. Ultimately, loss of FLNC-mediated actin crosslinking and deficient thin filament gene expression caused severe impairment of both contractility and relaxation in FLNC−/− hiPSC-CMs. These effects explain the embryonic lethality of homozygous null alleles in mice26,28 and the absence of homozygous null alleles in humans.42
In contrast to FLNC−/− hiPSC-CMs, the sarcomere content, organization, and function of FLNC+/− hiPSC-CMs was indistinguishable from WT. Although experimental in vitro conditions cannot fully recapitulate in vivo mechanics and hemodynamics, the substantial differences between FLNC−/− and FLNC+/− hiPSC-CMs implied that haploinsufficiency evoked a different pathogenetic mechanism. The FLNC+/Δ7aa mutant, which encoded a misfolded FLNC protein and shared >50% of DE proteins with FLNC+/− hiSPC-CMs, caused a similar but more profound cellular phenotype and was ultimately quite useful in elucidating abnormal proteostasis as a key disease mechanism for both heterozygous variants.
Proteomic analyses of insoluble protein fractions revealed that FLNC+/Δ7aa hiPSC-CMs formed prominent aggregates of FLNC and its binding partners, akin to those that occur with FLNC missense alleles that also cause FLNC misfolding.9 Given its considerable abundance (top 3% of cardiomyocyte proteins), we suspect that aggregation of FLN imposes a profound burden on protein clearance mechanisms, which may evoke hypertrophic remodeling similar to other proteopathies such as cardiac amyloidosis.56 It is possible that the variable sequestration of particular FLNC-binding partners caused by distinct FLNC variants may contribute to the diversity of clinical phenotypes caused by FLNC variants.10
Proteome-wide comparisons of the FLNC+/Δ7aa and FLNC+/− lines showed that FLNC+/− had a similar, although less severe phenotype than the FLNC+/Δ7aa hiPSC-CMs. Like FLNC+/Δ7aa hiPSC-CMs, FLNC+/− hiPSC-CMs also showed evidence of ER stress as well as increased lysosome content and autophagy, implying that the FLNC+/− variant also activated protein handling pathways. We suspect that the activation of protein handling pathways in the FLNC+/− and FLNC+/Δ7aa lines occurred by similar mechanisms induced by inefficient protein turnover at the Z-disc. The greater severity of phenotype in FLNC+/Δ7aa is likely secondary to the formation of aggregates of FLNC and a wide range of FLNC-binding partners in this mutant, which places an even greater burden on protein clearance pathways. However, FLNC+/− hiPSC-CMs produce half-normal levels of FLNC protein, and thus the accumulation of FLNC and its binding partners is likely more modest, resulting in an intermediate phenotype. Although our investigations suggest that both FLNC haploinsufficient and aggregate forming variants ultimately both cause proteotoxicity by similar mechanisms, one limitation of this study is that FLNC+/Δ7aa is not a clinical variant. Future studies will investigate whether other FLNC clinical missense and indel variants also produce proteotoxicity, as proposed by our model (Figure 7).
Our investigations add to a growing body of work that recognizes the importance of lysosomes in the turnover of Z-disc proteins. The central role of lysosomes in this process is highlighted by the profound cardiomyopathy that occurs with damaging variants in X-linked LAMP2 which deplete the encoded lysosomal-associated membrane protein-2 and cause multisystem disease in affected males.57 Electrostimulation of soleus muscles isolated from Lamp2-null mice results in Z-disc disintegration—another indication that lysosome-mediated degradation is essential at the Z-disc.58 The importance of lysosomes in Z-disc protein turnover may explain their accumulation with damaging heterozygous FLNC variants and also why lysosome content was not increased with homozygous null alleles, in which Z-disc and thin filament gene expression is inhibited. The downregulation of thin filament gene transcription in FLNC−/− hiPSC-CMs suggests that cardiomyocytes may have intrinsic mechanisms to sense Z-disc stoichiometry and negatively regulate their transcription and assembly in the absence of key components such as FLNC, perhaps so as to prevent pathological protein aggregation. Further study of the mechanisms by which cardiomyocytes determine the stoichiometry of Z-disc protein complexes and then dispose of excess or damaged proteins in lysosomes is likely to provide further insights into the fundamental processes of sarcomere assembly and repair.
Impaired proteostasis is an emerging mechanism of cardiomyopathy that also occurs with pathogenic variants in other Z-disc genes including MYOT and DES, both of which, like FLNC, cause DCM and MFM.59 Furthermore, loss of function variants in Z-disc chaperones also cause DCM and MFM. These include the DES chaperone, αβ-cystallin (CRYAB),59,60 and the BAG domain–containing cochaperone (BAG3), which concentrates at Z-discs and cooperates with the autophagy adaptor protein, p62, to dispose of ubiquitinated FLNC.58FLNC+/− hiPSC-CMs had depleted levels of two additional Z-disc chaperones—HSPB1 and HSPB7—both of which interact with FLNC and promote its proper localization and folding.18,25 Loss of HSPB7 leads to FLNC aggregation, cardiomyocyte dysfunction, and myopathy in zebrafish, mice, and cultured human cardiomyocytes.25,61,62 Replenishment of these critical Z-disc chaperone proteins, and further augmentation of protein handling pathways that are inadequate in FLNC+/− cardiomyocytes, may be broadly therapeutic for patients with Z-disc-related cardiac and skeletal myopathies. Additional interventions to increase autophagic flux through pharmacological modulation and lifestyle modifications, such as endurance exercise and caloric restriction, may also have clinical value.63,64 hiPSC-CMs carrying FLNC variants provide an ideal model system to test the efficacy of these interventions.
In conclusion, damaging FLNC variants produce proteopathic cardiomyopathies and reinforce the critical roles for the lysosome-dependent turnover of excess, misfolded, or damaged proteins in cardiomyocytes. The growing recognition of pathological protein accumulation that occurs in many genetic and unsolved causes of DCM, HCM, and MFM indicates that an improved understanding of associated mechanisms and development of appropriate interventions has the potential to benefit many patients with these proteopathic myopathies.
migfilin/filamin binding LIM protein 1
filamin A interacting protein 1
human induced pluripotent stem cell
human induced pluripotent stem cell–derived cardiomyocytes
heat shock proteins
PDZ and LIM domain 3
We thank the Harvard Medical School (HMS) Micron Core for assistance with cardiomyocyte imaging and the National Science Foundation (NSF) Engineering Research Center on Cellular Metamaterials (EEC-1647837) for expertise in patterning human induced pluripotent stem cell–derived cardiomyocytes (hiPSC-CMs). Graphics including graphical abstract were created with BioRender.com.
Sources of Funding
This work was supported in part by grants from the National Institutes of Health (NIH) National Heart Lung and Blood Institute (F30HL147389 to R. Agarwal), the NIH National Heart Lung and Blood Institute, and Leducq Foundation (HL084553 and HL080494 to J.G. Seidman and C.E. Seidman), the Howard Hughes Medical Institute (to C.E. Seidman), NIH National Institute of General Medical Sciences (R01 GM132129 to J.A. Paulo and GM97645 to S.P. Gygi), the British Heart Foundation Centre for Research Excellence award (RE/18/3/34214 to C.N. Toepfer), the National Science Foundation Graduate Research Fellowships Program (DGE-1840990 to J.K. Ewoldt), and the Sir Henry Wellcome Fellowship (206466/Z/17/Z to C.N. Toepfer).
Expanded Materials and Methods
Data Supplement Figures I–VI
Data Supplement Movies I–IX
Data Sets I–VI
Disclosures C.E. Seidman and J.G. Seidman are founders and own shares in Myokardia Inc., a startup company that is developing therapeutics that target the sarcomere. C.N. Toepfer is a consultant for Myokardia Inc. The other authors report no conflicts.
Schultheiss HP, Fairweather D, Caforio ALP, Escher F, Hershberger RE, Lipshultz SE, Liu PP, Matsumori A, Mazzanti A, McMurray J,. Dilated cardiomyopathy.Nat Rev Dis Primers. 2019; 5:32. doi: 10.1038/s41572-019-0084-1CrossrefMedlineGoogle Scholar
Cho KW, Lee J, Kim Y. Genetic variations leading to familial dilated cardiomyopathy.Mol Cells. 2016; 39:722–727. doi: 10.14348/molcells.2016.0061MedlineGoogle Scholar
Ader F, De Groote P, Réant P, Rooryck-Thambo C, Dupin-Deguine D, Rambaud C, Khraiche D, Perret C, Pruny JF, Mathieu-Dramard M,. FLNC pathogenic variants in patients with cardiomyopathies: prevalence and genotype-phenotype correlations.Clin Genet. 2019; 96:317–329. doi: 10.1111/cge.13594CrossrefMedlineGoogle Scholar
Begay RL, Tharp CA, Martin A, Graw SL, Sinagra G, Miani D,. HHS public access. 2017; 1:344–359. doi: 10.1016/j.jacbts.2016.05.004.FLNCGoogle Scholar
Janin A, N’Guyen K, Habib G, Dauphin C, Chanavat V, Bouvagnet P, Eschalier R, Streichenberger N, Chevalier P, Millat G. Truncating mutations on myofibrillar myopathies causing genes as prevalent molecular explanations on patients with dilated cardiomyopathy.Clin Genet. 2017; 92:616–623. doi: 10.1111/cge.13043CrossrefMedlineGoogle Scholar
Ortiz-Genga MF, Cuenca S, Dal Ferro M, Zorio E, Salgado-Aranda R, Climent V, Padrón-Barthe L, Duro-Aguado I, Jiménez-Jáimez J, Hidalgo-Olivares VM,. Truncating FLNC mutations are associated with high-risk dilated and arrhythmogenic cardiomyopathies.J Am Coll Cardiol. 2016; 68:2440–2451. doi: 10.1016/j.jacc.2016.09.927CrossrefMedlineGoogle Scholar
Begay RL, Graw SL, Sinagra G, Asimaki A, Rowland TJ, Slavov DB, Gowan K, Jones KL, Brun F, Merlo M,. Filamin C truncation mutations are associated with arrhythmogenic dilated cardiomyopathy and changes in the cell–cell adhesion structures.JACC: Clin Electrophysiol. 2018; 4:504–514. doi: 10.1016/j.jacep.2017.12.003CrossrefMedlineGoogle Scholar
Peters S, Kumar S, Elliott P, Kalman JM, Fatkin D. Arrhythmic genotypes in familial dilated cardiomyopathy: implications for genetic testing and clinical management.Heart Lung Circ. 2019; 28:31–38. doi: 10.1016/j.hlc.2018.09.010CrossrefMedlineGoogle Scholar
Valdés-Mas R, Gutiérrez-Fernández A, Gómez J, Coto E, Astudillo A, Puente DA, Reguero JR, Álvarez V, Morís C, León D,. Mutations in filamin C cause a new form of familial hypertrophic cardiomyopathy.Nature Commun. 2014; 5:1–9. doi: 10.1038/ncomms6326CrossrefGoogle Scholar
Verdonschot JAJ, Vanhoutte EK, Claes GRF, Helderman-van den Enden ATJM, Hoeijmakers JGJ, Hellebrekers DMEI, de Haan A, Christiaans I, Lekanne Deprez RH, Boen HM,. A mutation update for the FLNC gene in myopathies and cardiomyopathies.Hum Mutat. 2020; 41:1091–1111. doi: 10.1002/humu.24004CrossrefMedlineGoogle Scholar
Hartwig JH, Stossel TP. Structure of macrophage actin-binding protein molecules in solution and interacting with actin filaments.J Mol Biol. 1981; 145:563–581. doi: 10.1016/0022-2836(81)90545-3CrossrefMedlineGoogle Scholar
Thompson TG, Chan YM, Hack AA, Brosius M, Rajala M, Lidov HG, McNally EM, Watkins S, Kunkel LM. Filamin 2 (FLN2): a muscle-specific sarcoglycan interacting protein.J Cell Biol. 2000; 148:115–126. doi: 10.1083/jcb.148.1.115CrossrefMedlineGoogle Scholar
van der Flier A, Sonnenberg A. Structural and functional aspects of filamins.Biochim Biophys Acta. 2001; 1538:99–117. doi: 10.1016/s0167-4889(01)00072-6CrossrefMedlineGoogle Scholar
Pudas R, Kiema TR, Butler PJ, Stewart M, Ylänne J. Structural basis for vertebrate filamin dimerization.Structure. 2005; 13:111–119. doi: 10.1016/j.str.2004.10.014CrossrefMedlineGoogle Scholar
Razinia Z, Mäkelä T, Ylänne J, Calderwood DA. Filamins in mechanosensing and signaling.Annu Rev Biophys. 2012; 41:227–246. doi: 10.1146/annurev-biophys-050511-102252CrossrefMedlineGoogle Scholar
van der Ven PF, Wiesner S, Salmikangas P, Auerbach D, Himmel M, Kempa S, Hayess K, Pacholsky D, Taivainen A, Schröder R,. Indications for a novel muscular dystrophy pathway. gamma-filamin, the muscle-specific filamin isoform, interacts with myotilin.J Cell Biol. 2000; 151:235–248. doi: 10.1083/jcb.151.2.235CrossrefMedlineGoogle Scholar
Kiema T, Lad Y, Jiang P, Oxley CL, Baldassarre M, Wegener KL, Campbell ID, Ylänne J, Calderwood DA. The molecular basis of filamin binding to integrins and competition with talin.Mol Cell. 2006; 21:337–347. doi: 10.1016/j.molcel.2006.01.011CrossrefMedlineGoogle Scholar
Collier MP, Alderson TR, de Villiers CP, Nicholls D, Gastall HY, Allison TM, Degiacomi MT, Jiang H, Mlynek G, Fürst DO,. HspB1 phosphorylation regulates its intramolecular dynamics and mechanosensitive molecular chaperone interaction with filamin C.Sci Adv. 2019; 5:eaav8421. doi: 10.1126/sciadv.aav8421CrossrefMedlineGoogle Scholar
Zhang M, Liu J, Cheng A, Deyoung SM, Saltiel AR. Identification of CAP as a costameric protein that interacts with filamin C.Mol Biol Cell. 2007; 18:4731–4740. doi: 10.1091/mbc.e07-06-0628CrossrefMedlineGoogle Scholar
Tu Y, Wu S, Shi X, Chen K, Wu C. Migfilin and Mig-2 link focal adhesions to filamin and the actin cytoskeleton and function in cell shape modulation.Cell. 2003; 113:37–47. doi: 10.1016/s0092-8674(03)00163-6CrossrefMedlineGoogle Scholar
Molt S, Bührdel JB, Yakovlev S, Schein P, Orfanos Z, Kirfel G, Winter L, Wiche G, van der Ven PF, Rottbauer W,. Aciculin interacts with filamin C and Xin and is essential for myofibril assembly, remodeling and maintenance.J Cell Sci. 2014; 127:3578–3592. doi: 10.1242/jcs.152157MedlineGoogle Scholar
Linnemann A, van der Ven PF, Vakeel P, Albinus B, Simonis D, Bendas G, Schenk JA, Micheel B, Kley RA, Fürst DO. The sarcomeric Z-disc component myopodin is a multiadapter protein that interacts with filamin and alpha-actinin.Eur J Cell Biol. 2010; 89:681–692. doi: 10.1016/j.ejcb.2010.04.004CrossrefMedlineGoogle Scholar
Takada F, Woude DLV, Tong H, Thompson TG, Watkins SC, Kunkel LM, Beggs AH. Myozenin: an alpha-actinin and gamma-filamin-binding protein of skeletal muscle Z lines.Proc Natl Acad Sci USA. 2001; 98:1595–1600. doi: 10.1073/pnas.041609698MedlineGoogle Scholar
González-Morales N, Holenka TK, Schöck F. Filamin actin-binding and titin-binding fulfill distinct functions in Z-disc cohesion.PLoS Genet. 2017; 13:e1006880. doi: 10.1371/journal.pgen.1006880CrossrefMedlineGoogle Scholar
Juo LY, Liao WC, Shih YL, Yang BY, Liu AB, Yan YT. HSPB7 interacts with dimerized FLNC and its absence results in progressive myopathy in skeletal muscles.J Cell Sci. 2016; 129:1661–1670. doi: 10.1242/jcs.179887MedlineGoogle Scholar
Dalkilic I, Schienda J, Thompson TG, Kunkel LM. Loss of FilaminC (FLNc) results in severe defects in myogenesis and myotube structure.Mol Cell Biol. 2006; 26:6522–6534. doi: 10.1128/MCB.00243-06CrossrefMedlineGoogle Scholar
Fujita M, Mitsuhashi H, Isogai S, Nakata T, Kawakami A, Nonaka I, Noguchi S, Hayashi YK, Nishino I, Kudo A. Filamin C plays an essential role in the maintenance of the structural integrity of cardiac and skeletal muscles, revealed by the medaka mutant zacro.Dev Biol. 2012; 361:79–89. doi: 10.1016/j.ydbio.2011.10.008CrossrefMedlineGoogle Scholar
Yangzhao Z, Ze’e C, Lunfeng Z, Mason Z, Changming T, Xinmin Z, Sylvia ME, Xi F, Wei F, Ju C. Loss of filamin C is catastrophic for heart function.Circulation. 2020; 141:869–871. doi: 10.1161/CIRCULATIONAHA.119.044061LinkGoogle Scholar
Vorgerd M, van der Ven PF, Bruchertseifer V, Löwe T, Kley RA, Schröder R, Lochmüller H, Himmel M, Koehler K, Fürst DO,. A mutation in the dimerization domain of filamin c causes a novel type of autosomal dominant myofibrillar myopathy.Am J Hum Genet. 2005; 77:297–304. doi: 10.1086/431959CrossrefMedlineGoogle Scholar
Kley RA, Serdaroglu-Oflazer P, Leber Y, Odgerel Z, van der Ven PF, Olivé M, Ferrer I, Onipe A, Mihaylov M, Bilbao JM,. Pathophysiology of protein aggregation and extended phenotyping in filaminopathy.Brain. 2012; 135:2642–2660. doi: 10.1093/brain/aws200CrossrefMedlineGoogle Scholar
Sharma A, Toepfer CN, Ward T, Wasson L, Agarwal R, Conner DA, Hu JH, Seidman CE. CRISPR/Cas9-mediated fluorescent tagging of endogenous proteins in human pluripotent stem cells.Curr Protoc Hum Genet. 2018; 96:220.127.116.11.20. doi: 10.1002/cphg.52CrossrefGoogle Scholar
Yang L, Yang JL, Byrne S, Pan J, Church GM. CRISPR/Cas9-directed genome editing of cultured cells.Curr Protoc Mol Biol. 2014; 2014:31.1.1–31.1.17. doi: 10.1002/0471142727.mb3101s107Google Scholar
Lian X, Zhang J, Azarin SM, Zhu K, Hazeltine LB, Bao X, Hsiao C, Kamp TJ, Palecek SP. Directed cardiomyocyte differentiation from human pluripotent stem cells by modulating Wnt/β-catenin signaling under fully defined conditions.Nat Protoc. 2013; 8:162–175. doi: 10.1038/nprot.2012.150CrossrefMedlineGoogle Scholar
Toepfer CN, Garfinkel AC, Venturini G, Wakimoto H, Repetti G, Alamo L, Sharma A, Agarwal R, Ewoldt JF, Cloonan P,. Myosin sequestration regulates sarcomere function, cardiomyocyte energetics, and metabolism, informing the pathogenesis of hypertrophic cardiomyopathy.Circulation. 2020; 141:828–842. doi: 10.1161/CIRCULATIONAHA.119.042339LinkGoogle Scholar
Sharma A, Toepfer CN, Schmid M, Garfinkel AC, Seidman CE. Differentiation and contractile analysis of GFP-sarcomere reporter hiPSC-cardiomyocytes.Curr Protoc Hum Genet. 2018; 96:21.12.1–21.12.12. doi: 10.1002/cphg.53CrossrefGoogle Scholar
Toepfer CN, Sharma A, Cicconet M, Garfinkel AC, Mücke M, Neyazi M, Willcox JAL, Agarwal R, Schmid M, Rao J,. SarcTrack an adaptable software tool for efficient large-scale analysis of sarcomere function in hiPSC-cardiomyocytes.Circ Res. 2019; 124:1172–1183. doi: 10.1161/CIRCRESAHA.118.314505LinkGoogle Scholar
Chopra A, Kutys ML, Zhang K, Polacheck WJ, Sheng CC, Luu RJ, Eyckmans J, Hinson JT, Seidman JG, Seidman CE,. Force generation via β-cardiac myosin, titin, and α-actinin drives cardiac sarcomere assembly from cell-matrix adhesions.Dev Cell. 2018; 44:87–96.e5. doi: 10.1016/j.devcel.2017.12.012CrossrefMedlineGoogle Scholar
Navarrete-Perea J, Yu Q, Gygi SP, Paulo JA. Streamlined Tandem Mass Tag (SL-TMT) protocol: an efficient strategy for quantitative (phospho)proteome profiling using tandem mass tag-synchronous precursor selection-MS3.J Proteome Res. 2018; 17:2226–2236. doi: 10.1021/acs.jproteome.8b00217CrossrefMedlineGoogle Scholar
Schwanhäusser B, Busse D, Li N, Dittmar G, Schuchhardt J, Wolf J, Chen W, Selbach M. Global quantification of mammalian gene expression control.Nature. 2011; 473:337–342. doi: 10.1038/nature10098CrossrefMedlineGoogle Scholar
Xu N, Gulick J, Osinska H, Yu Y, McLendon PM, Shay-Winkler K, Robbins J, Yutzey KE. Ube2v1 positively regulates protein aggregation by modulating ubiquitin proteasome system performance partially through K63 ubiquitination.Circ Res. 2020; 126:907–922. doi: 10.1161/CIRCRESAHA.119.316444LinkGoogle Scholar
Huang da W, Sherman BT, Lempicki RA. Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources.Nat Protoc. 2009; 4:44–57. doi: 10.1038/nprot.2008.211CrossrefMedlineGoogle Scholar
Karczewski KJ, Francioli LC, Tiao G, Cummings BB, Alföldi J, Wang Q, Collins RL, Laricchia KM, Ganna A, Birnbaum DP,; Genome Aggregation Database Consortium. The mutational constraint spectrum quantified from variation in 141,456 humans.Nature. 2020; 581:434–443. doi: 10.1038/s41586-020-2308-7CrossrefMedlineGoogle Scholar
Arike L, Valgepea K, Peil L, Nahku R, Adamberg K, Vilu R. Comparison and applications of label-free absolute proteome quantification methods on Escherichia coli.J Proteomics. 2012; 75:5437–5448. doi: 10.1016/j.jprot.2012.06.020CrossrefMedlineGoogle Scholar
Harvey PA, Leinwand LA. Cellular mechanisms of cardiomyopathy.JCB: Rev. 2011; 194:355–365. doi: 10.1083/jcb.201101100CrossrefMedlineGoogle Scholar
Kiani FA, Lehman W, Fischer S, Rynkiewicz MJ. Spontaneous transitions of actin-bound tropomyosin toward blocked and closed states.J Gen Physiol. 2018; 151:4–8. doi: 10.1085/jgp.201812188CrossrefMedlineGoogle Scholar
Feige MJ, Hendershot LM. Disulfide bonds in ER protein folding and homeostasis.Curr Opin Cell Biol. 2011; 23:167–175. doi: 10.1016/j.ceb.2010.10.012CrossrefMedlineGoogle Scholar
Gardner BM, Pincus D, Gotthardt K, Gallagher CM, Walter P. Endoplasmic reticulum stress sensing in the unfolded protein response.Cold Spring Harb Perspect Biol. 2013; 5:a013169. doi: 10.1101/cshperspect.a013169CrossrefMedlineGoogle Scholar
Uemura A, Oku M, Mori K, Yoshida H. Unconventional splicing of XBP1 mRNA occurs in the cytoplasm during the mammalian unfolded protein response.J Cell Sci. 2009; 122:2877–2886. doi: 10.1242/jcs.040584CrossrefMedlineGoogle Scholar
Guardia CM, Farías GG, Jia R, Pu J, Bonifacino JS. BORC functions upstream of kinesins 1 and 3 to coordinate regional movement of lysosomes along different microtubule tracks.Cell Rep. 2016; 17:1950–1961. doi: 10.1016/j.celrep.2016.10.062CrossrefMedlineGoogle Scholar
Sardiello M, Palmieri M, di Ronza A, Medina DL, Valenza M, Gennarino VA, Di Malta C, Donaudy F, Embrione V, Polishchuk RS,. A gene network regulating lysosomal biogenesis and function.Science. 2009; 325:473–477. doi: 10.1126/science.1174447CrossrefMedlineGoogle Scholar
Settembre C, Zoncu R, Medina DL, Vetrini F, Erdin S, Erdin S, Huynh T, Ferron M, Karsenty G, Vellard MC,. A lysosome-to-nucleus signalling mechanism senses and regulates the lysosome via mTOR and TFEB.EMBO J. 2012; 31:1095–1108. doi: 10.1038/emboj.2012.32CrossrefMedlineGoogle Scholar
Medina DL, Di Paola S, Peluso I, Armani A, De Stefani D, Venditti R, Montefusco S, Scotto-Rosato A, Prezioso C, Forrester A,. Lysosomal calcium signalling regulates autophagy through calcineurin and TFEB.Nat Cell Biol. 2015; 17:288–299. doi: 10.1038/ncb3114CrossrefMedlineGoogle Scholar
Dikic I, Elazar Z. Mechanism and medical implications of mammalian autophagy.Nat Rev Mol Cell Biol. 2018; 19:349–364. doi: 10.1038/s41580-018-0003-4CrossrefMedlineGoogle Scholar
Pohl C, Dikic I. Cellular quality control by the ubiquitin-proteasome system and autophagy.Science. 2019; 366:818–822. doi: 10.1126/science.aax3769CrossrefMedlineGoogle Scholar
Stephan JS, Yeh Y, Ramachandran V, Deminoff SJ, Herman PK. The Tor and PKA signaling pathways independently target the Atg1 / Atg13 protein kinase complex to control autophagy.Proc Natl Acad Sci USA. 2009; 106:17049–17054. doi: 10.1073/pnas.0903316106CrossrefMedlineGoogle Scholar
Bistola V, Parissis J, Foukarakis E, Valsamaki PN, Anastasakis A, Koutsis G, Efthimiadis G, Kastritis E. Practical recommendations for the diagnosis and management of transthyretin cardiac amyloidosis.Heart Fail Rev. 2021; 26:861–879. doi: 10.1007/s10741-020-10062-wCrossrefMedlineGoogle Scholar
Samad F, Jain R, Jan MF, Sulemanjee NZ, Menaria P, Kalvin L, Bush M, Jahangir A, Khandheria BK, Tajik AJ. Malignant cardiac phenotypic expression of Danon disease (LAMP2 cardiomyopathy).Int J Cardiol. 2017; 245:201–206. doi: 10.1016/j.ijcard.2017.06.031CrossrefMedlineGoogle Scholar
Arndt V, Dick N, Tawo R, Dreiseidler M, Wenzel D, Hesse M, Fürst DO, Saftig P, Saint R, Fleischmann BK,. Chaperone-assisted selective autophagy is essential for muscle maintenance.Curr Biol. 2010; 20:143–148. doi: 10.1016/j.cub.2009.11.022CrossrefMedlineGoogle Scholar
McLendon PM, Robbins J. Desmin-related cardiomyopathy: an unfolding story.Am J Physiol Heart Circ Physiol. 2019; 3:1220–1228. doi: 10.1152/ajpheart.00601.2011Google Scholar
Tannous P, Zhu H, Johnstone JL, Shelton JM, Rajasekaran NS, Benjamin IJ, Nguyen L, Gerard RD, Levine B, Rothermel BA,. Autophagy is an adaptive response in desmin-related cardiomyopathy.Proc Natl Acad Sci USA. 2008; 105:9745–9750. doi: 10.1073/pnas.0706802105CrossrefMedlineGoogle Scholar
Wu T, Mu Y, Bogomolovas J, Fang X, Veevers J, Nowak RB, Pappas CT, Gregorio CC, Evans SM, Fowler VM,. HSPB7 is indispensable for heart development by modulating actin filament assembly.Proc Natl Acad Sci USA. 2017; 114:11956. doi: 10.1073/pnas.1713763114CrossrefGoogle Scholar
Mercer EJ, Lin YF, Cohen-Gould L, Evans T. Hspb7 is a cardioprotective chaperone facilitating sarcomeric proteostasis.Dev Biol. 2018; 435:41–55. doi: 10.1016/j.ydbio.2018.01.005CrossrefMedlineGoogle Scholar
Gámez A, Yuste-Checa P, Brasil S, Briso-Montiano Á, Desviat LR, Ugarte M, Pérez-Cerdá C, Pérez B. Protein misfolding diseases: prospects of pharmacological treatment.Clin Genet. 2018; 93:450–458. doi: 10.1111/cge.13088CrossrefMedlineGoogle Scholar
Galluzzi L, Bravo-San Pedro JM, Levine B, Green DR, Kroemer G. Pharmacological modulation of autophagy: therapeutic potential and persisting obstacles.Nat Rev Drug Discov. 2017; 16:487–511. doi: 10.1038/nrd.2017.22CrossrefMedlineGoogle Scholar
Novelty and Significance
What Is Known?
Dominant heterozygous variants in filamin C (FLNC) can cause diverse cardiomyopathies.
FLNC is a muscle-specific member of the class of actin crosslinking filamin proteins that is expressed primarily at sarcomere Z-discs.
FLNC homozygous variants that cause loss of FLNC protein expression result in early lethality.
FLNC heterozygous variants cause diverse forms of human cardiomyopathy with missense variants most often associated with hypertrophy and truncating variants most often associated with dilation.
What New Information Does This Article Contribute?
FLNC homozygous variants profoundly impair sarcomere assembly and thin filament gene expression in human induced pluripotent stem cell–derived cardiomyocytes.
Diverse FLNC heterozygous variants including those that cause haploinsufficiency and those that cause FLNC misfolding impair the turnover of Z-disc proteins, activate autophagy, and increase lysosome content in human induced pluripotent stem cell–derived cardiomyocytes.
Dominant heterozygous variants in FLNC cause diverse presentations of cardiomyopathy. Complete ablation of FLNC expression (FLNC−/−) in human induced pluripotent stem cell–derived cardiomyocytes causes profound impairment of sarcomere assembly and thin filament gene expression. Human induced pluripotent stem cell–derived cardiomyocytes bearing FLNC heterozygous variants assemble sarcomeres but increase lysosome content and activate protein handling pathways due to impaired turnover of Z-disc proteins (FLNC+/−, FLNC+/Δ7aa) and protein aggregation of FLNC and its binding partners (FLNC+/Δ7aa). Therapies to facilitate the appropriate degradation and disposal of accumulating Z-disc proteins may help antagonize cardiomyopathy progression caused by FLNC variants.