Filamin C is Essential for mammalian myocardial integrity

FLNC, encoding filamin C, is one of the most mutated genes in dilated and hypertrophic cardiomyopathy. However, the precise role of filamin C in mammalian heart remains unclear. In this study, we demonstrated Flnc global (FlncgKO) and cardiomyocyte-specific knockout (FlnccKO) mice died in utero from severely ruptured ventricular myocardium, indicating filamin C is required to maintain the structural integrity of myocardium in the mammalian heart. Contrary to the common belief that filamin C acts as an integrin inactivator, we observed attenuated activation of β1 integrin specifically in the myocardium of FlncgKO mice. Although deleting β1 integrin from cardiomyocytes did not recapitulate the heart rupture phenotype in Flnc knockout mice, deleting both β1 integrin and filamin C from cardiomyocytes resulted in much more severe heart ruptures than deleting filamin C alone. Our results demonstrated that filamin C works in concert with β1 integrin to maintain the structural integrity of myocardium during mammalian heart development.


Author summary
The precise role of filamin C in mammalian heart development had not been determined, in part due to the lack of cardiac phenotypes in previously described Flnc knockout mice, which still had truncated filamin C expressed in the heart. In this study, we analyzed a true Flnc knockout mouse line, in which filamin C protein was completely ablated. Flnc knockout mice developed massive ruptures in their myocardium but not in the endocardium, suggesting filamin C is essential for the structural integrity of myocardium. On the other hand, we did not find overt abnormalities of sarcomeric structure in cardiomyocytes of Flnc knockout mice, indicating that filamin C is likely not required for sarcomeric assembly as previously observed in FLNC null iPSC-CMs. Moreover, contrary to the dogma that filamins are integrin inactivators, we found that filamin C plays an unexpected role in integrin activation and works in concert with β1 integrin to ensure the structural integrity of the myocardium. a1111111111 a1111111111 a1111111111 a1111111111 a1111111111

Introduction
Cardiomyopathy is one of the leading causes of morbidity and mortality around the world [1]. Genetic causes, including detrimental deletions, insertions, nonsense or missense mutations identified in nearly 100 genes, account for diverse forms of hypertrophic, dilated, restrictive, and arrhythmogenic cardiomyopathy [2,3]. One of the most mutated genes is FLNC (encoding filamin C), which has 77 variants identified among dilated cardiomyopathy (DCM) and 57 variants in hypertrophic cardiomyopathy (HCM) patients, many of which are pathogenic [4].
Filamins (FLNA, FLNB, FLNC) are large actin-binding and -crosslinking dimeric proteins, with each subunit ranging from 240 to 280 kDa [5]. Filamin C (FLNC) is predominantly expressed in striated muscle tissues [6], and is localized to the Z-disc [7], intercalated disc (ICD) [8], and costamere [6]. Filamin C contains an N-terminal actin-binding domain (ABD) and 24 C-terminal immunoglobulin (Ig)-like domains [5], which are responsible for protein dimerization and interacting with myotilin and FATZ-1 at Z-discs [9,10]. C-terminal Ig-like domains also interact with β1 integrin [11] and sarcoglycans [6] at the costamere, a structural and functional component that bridges and strengthens the connection of the Z-discs to the sarcolemma [12]. Thus, filamin C is proposed to serve as a link between myofibrils and sarcolemma [7,13,14]. In vitro studies have demonstrated filamins inactivate integrin by competing with talin for binding to the cytoplasmic domain of the integrin β subunit [15]. However, the functional consequences of loss of filamins, especially filamin C, on integrin activation and its potential role in filamin C-related cardiomyopathy has not been explored in vivo.
Several studies sought to elucidate the function of filamin C in heart. A nonsense mutation identified in the teleost fish medaka causes myocardial rupture in heart ventricles, suggesting that filamin C is involved in maintenance of structural integrity of cardiac muscle [16]. Ablating filamin C in human induced pluripotent stem cell-derived cardiomyocytes (hiPSC-CMs) led to sarcomere disarray [17]. Surprisingly, mice with homozygous deletion of the last 8 exons of Flnc did not show overt cardiac phenotypes [18]. However, these Flnc knockout mice still expressed a truncated form of filamin C protein in the heart [18]. Thus, a bona fide Flnc knockout mouse model is required to study the precise role of filamin C in mammalian heart.
To this end, we generated a floxed Flnc mouse line [19] and analyzed Flnc global knockout (Flnc gKO ) and Flnc cardiomyocyte-specific knockout (Flnc cKO ) mice. Both Flnc gKO and Flnc cKO mice died before embryonic day (E) 11.5 from severely ruptured ventricular myocardium, indicating filamin C is required to maintain the structural integrity of myocardium in mammalian heart. By immunofluorescence analyses, we found downregulation of key extracellular matrix (ECM) proteins which might partially explain the heart rupture phenotype. Surprisingly, we did not observe obvious sarcomere disarray in cardiomyocytes of Flnc gKO mice, suggesting that filamin C is not required for sarcomere assembly in vivo. Interestingly, we observed attenuated activation of β1 integrin specifically in myocardium of Flnc gKO mice. However, deleting β1 integrin from cardiomyocytes did not recapitulate the heart rupture phenotype in Flnc knockout mice, whereas deleting both β1 integrin and filamin C from cardiomyocytes resulted in larger heart ruptures. Our results demonstrated that filamin C works in concert with β1 integrin to maintain the structural integrity of myocardium during mammalian heart development.

Filamin C is essential for mammalian heart development
By in situ hybridization, we demonstrated that Flnc was specifically expressed in heart and somites from embryonic day (E) 9.5 to 11.5 (S1A Fig). The expression pattern of Flnc is consistent with the observation that heart and skeletal muscle are most affected in patients with mutations in FLNC [4,20]. However, previously described Flnc knockout mice only had defects in skeletal muscles but not in heart, probably owing to the hypomorphic nature of the mutant Flnc allele in that study [18]. Thus, a bona fide Flnc knockout mouse model is required to fully understand roles of filamin C in heart. To this end, we generated global Flnc knockout mice (Flnc -/or Flnc gKO ) by crossing Flnc fl/fl [19] mice with Sox2 Cre mice [21] (Fig 1A). Western blot and immunofluorescence analyses confirmed that filamin C protein was completely absent in Flnc gKO mice (Figs 1B and S1B).

Filamin C maintains the integrity of the myocardial wall
Upon close examination of E10.5 Flnc gKO mouse hearts, we observed ruptures in the ventricular wall (Fig 1D). To determine the exact location of the ruptures, we intercrossed Flnc +/mice with Rosa26 tdTomato /Xmlc2 Cre to label cardiomyocytes with tdTomato fluorescence [25,26]. Results showed that Flnc gKO mice had multiple ruptures in myocardium but not in endocardium (Fig 2A). Interestingly, the location of rupture sites varied between individual Flnc gKO mice (Fig 2A).
To investigate whether the heart rupturing was caused by cardiomyocyte hypoplasia resulting from decreased cardiomyocyte proliferation and/or increased cardiomyocyte apoptosis, we measured cardiomyocyte proliferation and apoptosis rates in E8.5 to E10.5 Flnc gKO hearts and littermate controls. Although cardiomyocyte proliferation was markedly reduced and cardiomyocyte apoptosis was increased in E10.5 Flnc gKO hearts compared with controls, both parameters were indistinguishable between Flnc gKO and controls from E8.5 to E9.5 (S2B- S2C Fig).
Because the heart rupture phenotype was already evident in E9.5 Flnc gKO hearts, these findings indicated that heart rupturing was not caused by cardiomyocyte hypoplasia.
As filamin C is thought to play a role in sarcomere assembly in iPSC-CMs in vitro [17], we examined the overall sarcomere structure in E9.5 Flnc gKO cardiomyocytes by IF using antibodies against α-actinin (Z line) or myomesin (M line). However, we did not find any obvious sarcomere disarray in Flnc gKO hearts (S2D Fig), indicating filamin C is dispensable for sarcomere assembly in vivo.

Wound healing and blood coagulation pathways were activated in Flnc gKO
To assess transcriptomic changes in Flnc gKO mice, we extracted RNA from E9.5 Flnc gKO and littermate control hearts and performed RNA sequencing (RNA-seq). Using false discovery rate (FDR) < 0.05, we identified 901 significantly upregulated and 315 significantly downregulated differentially expressed genes (DEGs) in Flnc gKO hearts ( Fig 3A-3B and S1 Table). Among the most downregulated DEGs was Flnc, indicating our RNA-seq faithfully reflected gene expression changes between Flnc gKO and controls ( Fig 3B). Gene ontology analysis revealed downregulated DEGs enriched in molecular pathways related to cardiac chamber morphogenesis and function (Fig 3C), which might contribute to the cardiac phenotypes observed in Flnc gKO mice.
On the other hand, we found that genes involved in blood coagulation, including Pdgfb [27], Ppbp [28] and Gp5 [29] (Fig 3B) were dramatically upregulated in Flnc gKO hearts. Gene ontology analysis demonstrated upregulated DEGs mostly enriched in blood coagulation and wound healing processes (Fig 3C), in agreement with the formation of thrombi at rupture sites in Flnc gKO hearts (Fig 2B-2D). In addition, we found compensatory upregulation of Flna and Flnb in Flnc gKO hearts (Fig 3B).

Extracellular matrix proteins are downregulated in the myocardium of Flnc gKO
As cell-cell junctions of cardiomyocytes are critical for structural integrity of the heart [30], we examined expression and localization of key cell-cell junction proteins, including cadherins and desmoplakin, in E9.5 Flnc gKO hearts (Fig 4A-4B). However, we found their expression and localization were comparable between Flnc gKO and controls, indicating the heart rupture phenotype was not caused by diminished expression or mislocalization of cell-cell junction proteins.
Because filamin C is localized to costameres and interacts with the dystrophin-associated glycoprotein complex (DGC) [6], we then examined expression and localization of the DGC proteins β-dystroglycan and γ-sarcoglycan and found they were not reduced or mislocalized in E9.5 Flnc gKO hearts (Fig 4C-4D), indicating the heart rupture phenotype in Flnc gKO hearts was not caused by dysregulation of the DGC complex.
We previously reported extracellular matrix (ECM) disorganization caused heart rupture phenotypes in kindlin-2 cardiomyocyte-specific knockout mice [31]. To determine whether filamin C is required for the proper organization of ECM, we performed IF of ECM proteins collagen I and laminin on E9.5 Flnc gKO hearts. Although the expression of collagen I and laminin were upregulated in endocardium, they were markedly downregulated in myocardium of Flnc gKO hearts (Fig 4E), which may partially account for the myocardial rupture phenotype in Flnc gKO mice.

Both filamin C and β1 integrin are required to maintain the structural integrity of myocardium
Filamins are known integrin inactivators and abnormal activation of β1 integrin can lead to impaired cell proliferation, differentiation and migration [15]. To determine whether β1 integrin, a dominant integrin β isoform in cardiomyocytes [31], was ectopically activated in Flnc gKO mice, we performed IF using an antibody (9EG7) [32] against the activated ligand-bound conformation of β1 integrin and an antibody against total β1 integrin. Surprisingly, we found that activated β1 integrin was reduced in the myocardium of Flnc gKO hearts without changes in the total β1 integrin expression (Fig 5A). In contrast, both activated and total β1 integrin were upregulated in regions of endocardium proximal to rupture sites of Flnc gKO hearts (Fig 5A).
To assess whether decreased activation of β1 integrin in cardiomyocytes contributed to the heart rupture phenotype observed in Flnc gKO mice, we generated cardiomyocyte-specific Itgb1 (encoding β1 integrin) knockout mice using Xmlc2 Cre (Itgb1 cKO ) and compared them with cardiomyocyte-specific Flnc knockout mice (Flnc cKO ) (Fig 5B). Consistent with observations in Flnc gKO mice, Flnc cKO mice had obvious ruptures in their myocardium at E10.5 (Fig 5C). Although we previously reported that Itgb1 cKO develop heart ruptures at E14.5 [31]. Itgb1 cKO hearts did not show rupture at E10.5 (Fig 5C). These findings indicated that decreased β1 integrin activation in Flnc gKO mice may only partially account for the myocardial rupture phenotype.

PLOS GENETICS
Filamin C is essential for mammalian myocardial integrity To further explore the genetic interaction of Flnc and Itgb1, and its contribution to myocardial wall integrity, we generated cardiomyocyte-specific Flnc/Itgb1 double knockout mice (Flnc/Itgb1 dcKO ) (Fig 5B). Strikingly, Flnc/Itgb1 dcKO mice had much more severe heart rupturing which led to myocardial disintegration (Fig 5C), compared with Flnc cKO mice. Consequently, the endocardium of Flnc/Itgb1 dcKO mice became inflated, presumably due to the lack of mechanical support from the myocardium (Fig 5C). Flnc/Itgb1 dcKO mice also had larger pericardial effusions and more pronounced overall growth retardation (Fig 5C). Taken together, our findings suggest that filamin C works in concert with β1 integrin to maintain the structural integrity of myocardium during mammalian heart development.

PLOS GENETICS
Filamin C is essential for mammalian myocardial integrity

Discussion
In this study, we demonstrated that filamin C played an essential role in maintaining the structural integrity of myocardium, as Flnc gKO mice had severely ruptured ventricular myocardium but intact endocardium. Interestingly, CD31-positive thrombi and chest wall overgrowth were observed at the rupture sites, and β1 integrin and ECM proteins were upregulated in the endocardium. These phenomena are likely compensatory mechanisms to prevent complete heart rupturing. However, Flnc gKO mice did not survive past E11.5, indicating that filamin C is essential for heart development and embryonic survival. Mechanistically, although several cell junction and dystrophin-associated glycoprotein complex (DGC) proteins were unchanged, key extracellular matrix (ECM) proteins were downregulated in myocardium of Flnc gKO mice which may partially explain the heart rupture phenotype, reminiscent of our findings in kindlin-2 knockout mice [31]. Contrary to the belief that filamin C functions as an integrin inactivator, we observed attenuated activation of β1 integrin specifically in the myocardium of Flnc gKO mice. To further investigate whether downregulation of activated β1 integrin was key

PLOS GENETICS
Filamin C is essential for mammalian myocardial integrity to cardiac phenotypes in Flnc gKO mice, we generated β1 integrin cardiomyocyte-specific knockout mice (Itgb1 cKO ). However, Itgb1 cKO mice did not recapitulate the early heart rupture phenotype observed in Flnc knockout mice, whereas deleting β1 integrin and filamin C simultaneously from cardiomyocytes resulted in much more severe heart ruptures. Our results suggest filamin C works in concert with β1 integrin to maintain the structural integrity of myocardium during mammalian heart development.
FLNC is among the most mutated genes in dilated cardiomyopathy (DCM) and hypertrophic cardiomyopathy (HCM) patients [4], underscoring the essential role of filamin C in cardiac development and function. However, previously described Flnc knockout mice with homozygous deletion of the last 8 exons of Flnc only had defects in skeletal muscle but not in cardiac muscle [18]. Further studies revealed that these mice still expressed a truncated form of filamin C protein lacking the last four immunoglobulin (Ig)-like repeats and the hinge 2 region [18]. While the truncated filamin C protein was expressed at a lower level than wildtype, the reduction in Flnc mRNA levels was less pronounced in heart than in limb muscle [18], which may explain why there are phenotypes in skeletal muscles but not in heart of Flnc knockout mice. These observations also suggested that the N-terminal actin-binding domain and 19 Ig-like repeats (~82% of wild-type protein), even at much lower levels than that of wild-type FLNC proteins, are sufficient for filamin C to function normally in heart. Thus, the hypomorphic nature of the mutant Flnc allele renders it unsuitable for studying the function of filamin C in the heart. To address this problem, we generated Flnc knockout mice by deleting exon 9-13 of the Flnc gene, which introduced a premature termination codon (PTC) within exon 14 and subjected Flnc mRNA to nonsense-mediated mRNA decay (NMD). In line with this, Flnc mRNA levels were drastically downregulated in Flnc global knockout mice according to our RNA-seq data (Fig 3B and S1 Table). Although a small amount of N-terminal truncated protein (not recognizable by our filamin C antibodies that were raised against C-terminal regions of filamin C protein) could be generated, the truncated protein is unlikely to be functional as it only includes the N-terminal actin-binding domain and two Ig-like domains (486 amino acids,~17% of wild-type protein).
In a recent report, filamin C was ablated in in vitro cultured human induced pluripotent stem cell-derived cardiomyocytes (FLNC −/− hiPSC-CMs), which exhibited defects in sarcomere assembly and decreased thin filament gene expression, suggesting that filamin C plays a role in sarcomere assembly and thin filament gene expression [17]. To determine whether filamin C possesses similar functions in vivo, we examined overall sarcomere structure in E9.5 Flnc gKO cardiomyocytes by immunofluorescence analyses. However, we did not observe any sarcomere disarray as seen in FLNC −/− hiPSC-CMs. In addition, our RNA-seq analysis revealed very modest downregulation of thin filament genes including Lmod2 (Log 2 FC = -0.33), Tnni3 (Log 2 FC = -0.30) and Synpo2 (Log 2 FC = -0.31) (S1 Table), which is in stark contrast to the dramatic downregulation of thin filament genes in FLNC −/− hiPSC-CMs [17]. Our findings suggest that filamin C is dispensable in sarcomere assembly and has minimal impact on the expression of thin filament genes in vivo. It is worth noting that filamin A and filamin B were not upregulated in FLNC −/− hiPSC-CMs according to our examination of the transcriptomics and proteomics data from that study [17]. Thus, upregulation of filamin A and filamin B, or lack thereof, could explain why there are no defects of sarcomere assembly in cardiomyocytes of Flnc gKO mice but sarcomere disarray in FLNC −/− hiPSC-CMs. Future studies, i.e., ablating all three filamins from cardiomyocytes in vivo, might be necessary to elucidate roles of filamins in sarcomere assembly.
Filamin C interacts with β1 integrin [11] and sarcoglycans [6] at the costamere to serve as a link between myofibrils and sarcolemma. Our discovery of myocardial wall ruptures in Flnc knockout mice provided strong support for filamin C's essential structural role in myocardium integrity. On the other hand, filamins are well-known integrin inactivators that function by competing with talin for binding to the cytoplasmic domain of the integrin β subunit [15], and abnormal activation of β1 integrin can lead to impaired cell proliferation, differentiation and migration [15]. To determine whether β1 integrin is ectopically activated in filamin C-ablated cardiomyocytes which could account for the observed cardiac phenotypes, we examined expression and localization of activated and total β1 integrin by immunofluorescence. Surprisingly, while total β1 integrin expression and localization were unchanged, activated β1 integrin was reduced in myocardium of Flnc gKO hearts. To further investigate whether the attenuated activation of β1 integrin was key to cardiac phenotypes in Flnc gKO mice, we generated cardiomyocyte-specific Itgb1 knockout mice (Itgb1 cKO ) and compare them with cardiomyocyte-specific Flnc knockout mice (Flnc cKO ). However, ablating β1 integrin in cardiomyocytes did not cause myocardial rupture at E10.5, a stage when Flnc cKO mice already had severe rupturing in their myocardium. Considering that Itgb1 cKO indeed develop heart ruptures later at E14.5 [31], these findings indicated that attenuated β1 integrin activation alone may only partially account for myocardial ruptures in Flnc gKO mice. Another possibility is that some residual β1 integrin proteins may still remain in E10.5 Itgb1 cKO mice due to their remarkably long half-life [33], and these remaining β1 integrin proteins can be normally activated in the presence of filamin C. If this is the case, simultaneously ablating filamin C and β1 integrin should recapitulate the phenotypes of Flnc cKO mice. However, the Flnc/Itgb1 dcKO mice we generated had even more severe heart rupturing phenotypes than Flnc cKO mice, suggesting that filamin C maintains the integrity of myocardium through both integrin-dependent and integrin-independent pathways. Future studies are needed to delineate detailed molecular mechanisms by which filamin C facilitates β1 integrin activation in cardiomyocytes.

Ethics statement
All animal procedures were performed in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals and approved by the Institutional Animal Care and Use Committee of the University of California San Diego with approved protocol # S01049.

Western blots
Western blots were performed as previously described [26,35]. Briefly, embryonic mouse hearts were dissected and snap-frozen in liquid nitrogen. Total protein extracts were prepared by homogenization of hearts in RIPA buffer using a handheld pellet pestle (Sigma-Aldrich). Protein samples were separated on Bolt 4%-12% Bis-Tris gels (Life Technologies) and transferred to PVDF membrane (Bio-Rad). Membranes were then blocked and incubated with primary antibodies overnight at 4˚C. Membranes were then washed with TBST and incubated with HRP-conjugated secondary antibodies and visualized using enhanced chemiluminescence (ECL) reagent (Bio-Rad) and captured by Bio-Rad ChemiDoc Imaging System. Catalog numbers for antibodies used in western blots in this study: filamin C, NBP1-89300 (Novus); GAPDH, sc-32233 (Santa Cruz Biotechnology).

RNA Sequencing
RNA sequencing (RNA-seq) was performed as previously described [26]. E9.5 embryonic hearts or isolated ventricles were homogenized in TRIzol (Invitrogen) and total RNA was isolated according to the manufacturer's instructions. The concentration and quality of purified RNA was assessed by TapeStation (Agilent). cDNA libraries were prepared using an Illumina TruSeq stranded mRNA kit according to manufacturer's instructions. Libraries were sequenced with an Illumina NovaSeq 6000 sequencer. Sequencing reads were mapped to GEN-CODE mouse transcripts reference (release M22, GRCm38.p6) and transcription levels were quantified using salmon. Subsequently, differential expression analysis was carried out using DEseq2 (version: 1.22.2). Benjamini-Hochberg correction for multiple testing was applied to correct p-value of each gene as false discovery rate (FDR). FDR < 0.05 was used as a threshold for differentially-expressed genes (DEGs). Lists of downregulated DEGs and upregulated DEGs were separately examined for statistical enrichment of gene ontology (GO) terms and biological pathways in Toppgene (https://toppgene.cchmc.org). RNA-seq datasets were deposited in Gene Expression Omnibus (GEO) with the accession number GSE222542.

Statistical analysis
Data are presented as mean ± standard error of the mean (SEM). Statistical analysis was performed using GraphPad Prism 9 software, with Welch's t test used for comparisons among groups as indicated. P-values less than 0.05 were considered significant and reported as � p < 0.05, �� p < 0.01, ��� p < 0.001, ���� p < 0.0001.