Preventing hypocontractility-induced fibroblast expansion alleviates dilated cardiomyopathy

Abstract

Cardiomyocyte hypocontractility underlies inherited dilated cardiomyopathy (DCM). Yet, whether fibroblasts modify DCM phenotypes remains unclear despite their regulation of fibrosis, which strongly predicts disease severity. Expression of a hypocontractility-linked sarcomeric variant in mice triggered cardiac fibroblast expansion from the de novo formation of hyperproliferative mechanosensitized fibroblast states, which occurred prior to eccentric myocyte remodeling. Initially, this fibroblast response reorganized fibrillar collagen and stiffened the myocardium, albeit without depositing fibrotic tissue. These adaptations coincided with heightened matrix-integrin receptor interactions and diastolic tension sensation at focal adhesions within fibroblasts. Targeted p38 deletion arrested these cardiac fibroblast responses in DCM mice, which prevented cardiomyocyte remodeling and improved contractility. p38-mediated fibroblast responses were essential regulators of DCM severity, marking a potential cellular target for therapeutic intervention.

Fig. 1.. Cardiomyocyte hypocontractility stiffens and aligns the myocardium prior to eccentric hypertrophic remodeling and fibrosis.

(A and B) Quantification of (A) isolated cardiomyocyte length and (B) length:width ratio in relaxed conditions from the described genotypes [CON n = 399/402 (2/4 month), I61Q n = 357/399 (2/4 month)]. (C and D) Quantification of (C) heart weight:body weight ratio by gravimetrics and (D) left ventricular diastolic chamber diameter by echocardiography at 2 (I61Q n = 12, CON n = 12) and 4 (I61Q n = 16, CON n = 15) months of age. (E) Quantification of diastolic sarcomere length in intact cardiomyocytes from I61Q and CON hearts [CON n = 399/402 (2/4 month), I61Q n = 357/399 (2/4 month)]. (F) Quantification of EDPVR by invasive hemodynamics [CON n = 5/9 (2/4 month), I61Q n = 5/8 (2/4 month)]. (G) Representative developed pressure traces from stepwise inflation of a balloon inside a blebbistatin-treated intact (solid line) and decellularized (dotted line) heart. (H and I) Pressure-volume curves of (H) intact and (I) decellularized mouse hearts at 2 months of age (n = 7 both genotypes). (J to L) Quantification of (J) titin isoforms N2B and N2BA, (K) N2B S267 phosphorylation site, and (L) serine PEVK region S267 phosphorylation by Coomassie staining and western blot (n = 3 both genotypes). The N2B isoform is expressed relative to the total titin (N2B + N2BA), and phosphospecific antibodies were normalized to the titin stain on the Coomassie gel. (M) Heatmap of differentially expressed matrisome proteins color coded by z-score identified by MS in 4-month-old decellularized ECM (CON n = 3, I61Q n = 4). (N and O) Representative images (N) and quantification of fibrosis (O) in cardiac sections stained with picrosirius red-fast green [PSR/FG; scale bars, 1 mm (left and middle) and 50 μm (right)]. (P) Representative two-photon maximum-intensity projection images of SHG [scale bar, 100 μm (left)] and (Q) masking of the collagen fibers (scale bar, 20 μm) in decellularized hearts in which Z-stacks were taken starting 10 μm below the epicardial surface. Fibrillar collagen (magenta) and ECM autofluorescence (green). (R and S) CurveAlign quantification of collagen fiber length (R) and alignment (S) from 2-month-old I61Q (n = 10) and CON (n = 11) hearts. (T) Representative TEM images of collagen fibrils in 2-month-old I61Q and CON hearts. (U and V) Quantification of fibril diameter (U) and density (V) from TEM images (n = 20 ROIs/mouse, 3 mice per genotype). (W) Quantification of decorin abundance, identified by MS. [(A), (B), and (E)] Four biological replicates per genotype were used for isolated myocyte experiments. Data represent mean ± SEM. ns, not significant; **P < 0.01; ***P < 0.005; ****P < 0.001 by two-way ANOVA with Holm-Sidak’s multiple comparisons test [(A) to (D), (F), and (O)] or two-tailed unpaired t test [(E), (J) to (L), (R), (S), (V), and (W)]. Single-letter abbreviations for the amino acid residues referenced throughout this paper are as follows: I, Ile; Q, Gln; S, Ser; P, Pro; E, Glu; V, Val; K, Lysine.

Fig. 2.. Hyperproliferative fibroblast states drive fibrosis-independent tissue stiffening in I61Q hearts.

(A) Representative images of (left) 2-month-old CON and I61Q cardiac sections stained for αSMA and PDGFRα and (right) Postn lineage-traced cell density in 8-month-old cardiac sections. (B) Quantification of immunofluorescent imaging for the percentage of fibroblasts (PDGFRα⁺) expressing αSMA (2 months old: n = 14 CON, n = 7 I61Q; 4 months old: n = 10 CON, n = 9 I61Q). (C) Quantification of Postn⁺ cell density in I61Q (n = 5) and CON (n = 8) hearts at 8 months of age. (D) UMAP dimensionality reduction plot for all sequenced cardiac fibroblast nuclei from 2-month-old I61Q and CON hearts. Each color represents distinct fibroblast clusters (states) based on differential gene expression. (E) Proportion analysis showing genotype-dependent changes in the percentage of the cardiac fibroblast population residing in each of the defined cell clusters (states). (F) Dot plot of the top five expressed genes that define each fibroblast cluster. (G) Kegg pathways and gene ontology biological processes (GO:BP) identified by differentially regulated genes in fibroblast clusters 1 and 3 that are specific to the I61Q genotype. (H) Heatmap showing expression levels of cell cycle genes. (I to L) Representative images (scale bar, 20 μm) (I) and quantification of immunofluorescent staining for phospho-histone H3 (pHH3) and PDGFRα (J); pHH3⁺, PDGFRα⁺ cells per total nuclei (Hoechst) (K); and pHH3⁺, PDGFRα⁺ cells as a percentage of the total number of PDGFRα⁺ cells in I61Q and CON cardiac sections (L). (M) Quantification of pHH3⁺ nonfibroblasts in I61Q [n = 8 (p14), 9 (1 month), 6 (2 months)] and CON [n = 8 (p14), 15 (1 month), 6 (2 months)] cardiac sections. (N and O) Representative images (scale bar, 5 mm) (N) and quantification (O) of the compaction of free-floating collagen gels seeded with cardiac fibroblasts (CFs) isolated from I61Q and CON hearts ± cell cycle inhibitor (CDKi, dinaciclib, 5 μM); n = 5 per genotype. (P) Quantification of cardiac fibroblast proliferation by 5-ethynyl-2′-deoxyuridine (EdU) incorporation ± cell cycle inhibitor (CDKi, dinaciclib, 5 μM). (Q) Representative images of fibrin tissues seeded with cardiac fibroblasts mounted between posts (scale bar, 1 mm). (R and S) Quantification of length (R) and passive force production (S) by tissues seeded with cardiac fibroblasts from 2-month-old I61Q cTnC transgenic or CON mice (n = 11 per genotype). (T) Quantification of cellular and ECM alignment in fibrin tissues seeded with cardiac fibroblasts derived from I61Q cTnC transgenic and CON hearts by wheat germ staining (n = 11 CON, n = 8 I61Q). (U) Average twitch forces generated by EHTs 2 weeks after cardiomyocytes were adenovirally transduced with either control (AdWT, n = 8) or I61Q mutant cTnC (AdI61Q, n = 6). (V to X) Quantification of twitch force (V), passive tension (W), and tissue alignment (X) by wheat germ staining in EHTs. Data represent mean ± SEM. ns, not significant; *P < 0.05; **P < 0.01; ***P < 0.005; ****P < 0.001 by two-way ANOVA with Holm-Sidak’s multiple comparisons test [(B), (J) to (M), (O), and (P)] or two-tailed unpaired t test [(C), (R) to (T), and (V) to (X)]. All scale bars are 50 μm unless otherwise noted.

Fig. 3.. Diastolic mechanosensations at fibroblast focal adhesions are accentuated by ECM-receptor interactions that trigger proliferation.

(A) Ligand-receptor analysis of myocyte-fibroblast interactions predicted from snRNA-seq of 2-month-old CON and I61Q hearts where the ligand is expressed in myocytes and receptor is expressed in fibroblasts. (B) Proliferation by EdU assay of CON and I61Q fibroblasts seeded within PEG gels decorated with ECM peptides from CON and I61Q hearts. (C) (Top) Representative images of cardiac fibroblasts seeded onto the ECM screening array stained for EdU to mark proliferating cells (scale bar, 100 μm) and (bottom) quantification of EdU⁺ fibroblast counts on ECM-coated microspots. (D) Schematic of cocultures with cardiomyocyte monolayers adenovirally transduced with either CON or I61Q cTnC that were sparsely overlaid with cardiac fibroblasts genetically encoded with a FRET tension sensor in the focal adhesion protein vinculin. (E to G) Quantification of the average FRET efficiency (correlates with tension) at fibroblast focal adhesions in (E) relaxed (+blebbistatin), (F) diastolic [Ca²⁺], and (G) submaximal activating conditions (n = 25 fibroblasts per group). (H) Representative images (top scale bar, 100 μm; bottom scale bar, 30 μm) of CON and I61Q cardiac fibroblasts stained for vinculin and filamentous actin (F-actin) and quantified for focal adhesion (I) area and (J) eccentricity (n = 681 CON and 622 I61Q fibroblasts from three different mice per genotype). (K and L) Quantification of (K) the percentage of proliferating cardiac fibroblast as measured by Ki67 positivity (n = 15 unloaded-WT, 15 preloaded-WT, 15 unloaded-I61Q, and 15 preloaded-I61Q) and (L) collagen alignment (n = 18 unloaded-WT, 18 preloaded-WT, 15 unloaded-I61Q, and preloaded-I61Q) in unloaded or chronically preloaded EHTs generated with cardiomyocytes adenovirally transduced with WT or I61Q cTnC. *P < 0.05; **P < 0.01; ***P < 0.005 by either two-way or one-way ANOVA with Holm-Sidak’s multiple comparisons test [(B), (C), (K), and (L)], or two-tailed unpaired t test [(E) to (G), (I), and (J)].

Fig. 4.. Fibroblast-specific p38 deficiency corrects cardiac dilation and systolic dysfunction in I61Q cTnC transgenic mice.

(A) Schematic showing the generation of I61Q cTnC transgenic mice with tamoxifen-inducible fibroblast-specific p38 deletion and the experimental genotypes derived from the described breeding scheme. Here, I61Q cTnC and tTA transgenic animals were bred with a mouse line containing conditional p38α loss of function (p38^fl/fl) and a tamoxifen-inducible Cre recombinase knocked into the Tcf21 locus (Tcf21^iCre). (B) Experimental design schematic showing that the I61Q mutant cTnC was expressed just after birth (~p2), mice were allowed to develop normally for 1 month, and then tamoxifen was administered to induce fibroblast-specific p38 excision. Experimental endpoints were at 2 and 4 months of age. (C and D) UMAP dimensionality reduction plot (C) and proportion analysis (D) for all sequenced cardiac fibroblast nuclei from 2-month-old I61Q, p38-I61Q, and CON hearts. Each color represents distinct fibroblast clusters (states) based on differential gene expression. UMAPs for I61Q and CON were also shown in Fig. 2 for clarity, but all three groups were sequenced simultaneously as part of the same experiment. (E and F) Quantification of fibroblast proliferation (E) and fibroblast density (F) by immunofluorescent staining for pH3 and PDGFRα in 2-month-old myocardial sections (CON, n = 13; p38 KO, n = 13; I61Q, n = 12; p38 KO I61Q, n = 12). (G) Representative images of EGR1 (green) and PDGFRα (magenta) in myocardial sections. White arrows indicate EGR1⁺ nuclei (Hoechst stain; CON, n = 8; p38 KO, n = 8; I61Q, n = 7; p38 KO I61Q, n = 7; scale bar, 100 μm). (H) Quantification of the percentage of proliferating CON or I61Q cardiac fibroblasts ± dsRed (control) or human EGR1 retroviral vectors ± p38 inhibitor as determined by EdU positivity (EdU⁺). (I and J) Representative PSR/FG-stained myocardial sections (scale bars, 50 μm) (I) and quantification of collagen (red) (J) (CON, n = 7; p38 KO, n = 11; I61Q, n = 8; p38 KO I61Q, n = 8). (K) Quantification of collagen fiber alignment from decellularized CON (n = 11), p38 KO (n = 11), I61Q (n = 9), and p38 KO I61Q (n = 7) hearts. (L and M) Representative relationship between normalized tension and Ca²⁺ concentration (pCa) (L) and Ca²⁺ sensitivity of tension generation (pCa₅₀) (M) in membrane-permeabilized trabeculae of CON (n = 19), p38 KO (n = 15), I61Q (n = 10), and p38 KO I61Q (n = 6) mice. (N) Mean twitch forces from intact trabeculae of 4-month-old CON (n = 5), p38 KO (n = 5), I61Q (n = 5), and p38 KO I61Q (n = 4) mice. (O) Quantification of left ventricular ejection fraction measured by echocardiography from CON (n = 17), p38 KO (n = 20), I61Q (n = 12), and p38 KO I61Q (n = 23) mice. (P and Q) Representative pressure-volume loops (P) and quantification of stroke work (Q) from invasive hemodynamic measurements at 4 months (CON, n = 9; p38 KO, n = 9; I61Q, n = 7; p38 KO I61Q, n = 6). (R and S) Quantification of unloaded sarcomere shortening amplitude (CON, n = 85; p38 KO, n = 87; I61Q, n = 45; p38KO-I61Q, n = 85 cardiomyocytes) (R) and Ca²⁺ transient amplitude (CON, n = 75; p38 KO, n = 81; I61Q, n = 54; p38 KO I61Q, n = 66) (S) in isolated intact cardiomyocytes from the described genotypes. (T) Quantification of left ventricular (LV) diastolic diameter at 4 months of age by echocardiography [n was the same as in (O)]. (U and V) Quantification of isolated cardiomyocyte length (U) and area (V) from the described genotypes (CON, n = 250; p38 KO, n = 250; I61Q, n = 199; p38 KO I61Q, n = 200). Isolated myocyte experiments [(R), (S), (U), and (V)] were performed at 4 months of age with the following biological replicates per genotype: CON, n = 4; p38 KO I61Q, n = 5; I61Q, n = 4; and p38 KO I61Q, n = 5. Data are mean ± SEM. ns, not significant; *P < 0.05; **P < 0.01; ***P < 0.005; ****P < 0.001 by two-way ANOVA with Holm-Sidak’s multiple comparisons test.