Transcription Factor 21 Regulates Cardiac Myofibroblast Formation and Fibrosis

Abstract

Background: TCF21 (transcription factor 21) is a bHLH (basic helix-loop-helix) protein required for the developmental specification of cardiac fibroblasts (CFs) from epicardial progenitor cells that surround the embryonic heart. In the adult heart, TCF21 is expressed in tissue-resident fibroblasts and is downregulated in response to injury or stimuli leading to myofibroblast differentiation. These findings led to the hypothesis that TCF21 regulates fibroblast differentiation in the adult mammalian heart to affect fibrosis.

Methods: Tamoxifen-inducible Cre genetic mouse models were used to permit either Tcf21 gene deletion or its enforced expression in adult CFs. Histological and echocardiographic analyses were used, as well as transcriptomic analysis to determine the consequences of TCF21 gain-of-function and loss-of-function in vivo. Genomic Tcf21 occupancy was identified by chromatin immunoprecipitation and sequencing in CFs. Myocardial infarction and AngII (angiotensin II)/phenylephrine served as models of cardiac fibrosis.

Results: Acute and long-term deletion of Tcf21 in CFs of the adult mouse heart does not alter fibroblast numbers, myofibroblast differentiation, or fibrosis. Fibroblast-specific Tcf21 gene-deleted mice demonstrate no significant alterations in cardiac function or scar formation in response to cardiac injury compared with control mice. In contrast, enforced expression of TCF21 in CFs inhibits myofibroblast differentiation and significantly reduces cardiac fibrosis and hypertrophy in response to 1 week of Ang II/phenylephrine infusion. Mechanistically, sustained TCF21 expression prevents the induction of genes associated with fibrosis and ECM (extracellular matrix) organization.

Conclusions: TCF21 expression is not required to maintain the cell state of CFs in the adult heart. However, preventing the normal downregulation of TCF21 expression with injury reduces myofibroblast formation, cardiac fibrosis, and the acute cardiac hypertrophic response following 1 week of Ang II/phenylephrine stimulation.

Keywords: cardiomyopathies; extracellular matrix; fibroblasts; fibrosis; myofibroblasts.

Figure 1.. Genetic loss of Tcf21 in adult cardiac fibroblasts does not promote myofibroblast differentiation.

(A) Quantitative real-time polymerase chain reaction (qRT-PCR) analysis of Tcf21 mRNA and (B) immunoblot analysis of TCF21 and smooth muscle α-actin (αSMA) protein in cultured adult murine cardiac fibroblasts from Tcf21^fl/fl mice infected with adenovirus-encoding beta-galactosidase (Ad-βgal) or Cre recombinase (Ad-Cre). Fifty micrograms of protein were loaded and GAPDH was used as a loading/normalization control. (C) qRT-PCR analysis of Acta2 and Tagln mRNA from the indicated samples in cultured adult murine cardiac fibroblasts from the indicated genotypes of mice with the 2 adenoviral infections in culture. (D) Schematic of the genetically modified mouse alleles used and temporal experimental outline. This cross also contained the Rosa26-eGFP conditional reporter allele to show fluorescence within recombined fibroblasts. Tamoxifen (tam) was administered to 8–12-week-old mice for 5 consecutive days by intraperitoneal injection (i.p) and harvested 14 days later. (E) qRT-PCR analysis of Tcf21 mRNA expression from GFP+ sorted cardiac fibroblasts from Tcf21^fl/MCM mice versus Tcf21^MCM controls. (F and G) Immunofluorescence microscopy analysis of heart histological sections for the myofibroblast markers (F) αSMA and (G) transgelin (SM22) as denoted by red fluorescence. GFP (green) shows the presence of the recombined fibroblasts due to the Rosa26-eGFP reporter allele and nuclei are shown in blue with DAPI staining. The arrows show areas of cardiac vessels as controls that are positive for smooth muscle gene expression with αSMA and SM22. Scale bars are 50 μm. (H) Volcano plot of differentially expressed mRNAs as detected by bulk transcriptomic sequencing from fluorescence-activated cell sorting of GFP+ cardiac fibroblasts (CF) from Tcf21^fl/MCM mice versus Tcf21^MCM controls. (I) Normalized Tcf21 transcript counts from these samples in “H”. (J) Heatmap of Z-score normalized read counts of manually selected fibrosis-associated genes in cardiac fibroblasts of Tcf21^fl/MCM versus Tcf21^MCM control mice. (K) Schematic of the genetically modified mouse alleles used and experimental outline. Tam was administered to 8–12-week-old mice for 5 consecutive days by intraperitoneal injection (i.p). followed by 1 week of angiotensin II/phenylephrine (AngII/PE) stimulation prior to analysis. (L) Heart weight to body weight ratio (HW/BW). (M) Representative whole heart histological images stained with Sirius red for collagen deposition from the 2 groups of mice shown. Scale bars are 1 mm. (N) Quantification of percentage area of tissue fibrosis as shown in “M”. All data shown are mean +/− standard error of the mean. Samples were analyzed by a Kolmogorov-Smirnov test (exact p values are shown; E, N) or a Scheirer–Ray–Hare test with a Dunn’s post-hoc test and Benjamini Hochberg method for multiple test correction (exact adjusted p values are shown; L). For cell culture experiments, each n value is an independent pool of cardiac fibroblasts obtained from one male and one female mouse heart. All other data points in graphs represents biological replicates.

Figure 2.. Loss of Tcf21 does not affect cardiac function or fibrosis after myocardial infarction (MI)

(A) Schematic of experimental outline for single cell RNA sequencing (scRNA-seq) analysis from isolated cardiac interstitial cells from uninjured hearts and 3 days post MI (3dpMI) from Tcf21^fl/MCM mice and Tcf21^MCM controls. (B) Unifold manifold projection (UMAP) of all single cells that passed quality control filtering from hearts of the groups discussed in “A”. (C) DittoPlot of the cardiac cell populations in the groups of mice shown, controls were Tcf21^MCM mice. (D) UMAP of re-clustered uninjured GFP+ (>2 counts) fibroblasts from hearts of the 2 groups of mice shown. (E) Schematic of the genetically modified mouse alleles used and temporal experimental outline. Tamoxifen (tam) was administered to 8–12-week-old mice for 5 consecutive days by intraperitoneal injection (i.p) followed by 1 week of tam chow. One week later, mice were subjected to either sham or MI surgery before harvesting 14 days post MI (14dpMI) for analysis of (F) heart weight to body weight (HW/BW) ratio and (G) echocardiographic analysis of left ventricular (LV) wall thickness in diastole, (H) ejection fraction (EF) (%) and (I) LV internal chamber diameter in diastole (LVID). (J) Representative whole heart histological images stained with Sirius red for collagen deposition from the 2 groups of mice shown. Scale bar is 1 mm. (K) Quantification of percentage area of tissue fibrosis in Tcf21^fl/MCM mice versus Tcf21^MCM controls as shown in “J”. All data shown are mean +/− standard error of the mean. Samples were analyzed by a Scheirer–Ray–Hare test with a Dunn’s post-hoc test and Benjamini Hochberg method for multiple test correction (exact adjusted p values are shown; F-I, K). All data points in graphs represents biological replicates. FBs; fibroblasts. SCs; Schwann cells. MΦ; macrophages. NK; natural killer. DCs; dendritic cells. EC; endothelial cells. SMCs; smooth muscle cells.

Figure 3.. Generation of tamoxifen-inducible fibroblast specific mice with enforced TCF21 expression

(A) Schematic of the CRISPR modified mouse Col1a1 genetic locus with inverted loxP sites and the human TCF21 cDNA in the reverse direction, which upon Cre-mediated recombination gives proper orientation for gene expression. (B) Quantitative real time polymerase chain reaction (qRT-PCR) mRNA analysis for TCF21 expression in cultured adult murine cardiac fibroblasts infected with adenovirus-encoding beta-galactosidase (Ad-βgal) or Cre recombinase (Ad-Cre). Data were normalized to GAPDH mRNA. (C) Experimental schematic of crossed mice with the engineered Col1a1-TCF21 allele, the Rosa26-eGFP conditional reporter allele, and the Pdgfra-CreERT2 allele with the temporal experimental outline at the bottom. Tamoxifen (tam) was administered to 8–12-week-old mice for 1 or 3 consecutive days by intraperitoneal injection (i.p) and harvested 7 days later. (D) PCR analysis of Cre presence and DNA inversion of the TCF21 cassette within the modified Col1a1 locus upon Cre-mediated recombination in hearts from tamoxifen-treated mice as shown in “C”. Pdgfra-CreERT2 mice were used as controls. (E) Immunofluorescent images from heart histological sections of the 2 indicated groups of mice for TCF21 (red), recombined fibroblasts (GFP+; green) and nuclei (blue; DAPI staining). The left 2 panels show all 3 channels while the right 2 panels show only TCF21 antibody reactivity (red). The scale bars are 20 μm. (F) Experimental scheme. Tam was administered by i.p. injection on 3 consecutive days to 8–12-week-old Pdgfra-CreERT2:Col1a1-TCF21 mice and either Pdgfra-CreERT2 or Col1a1-TCF21 control mice and analyzed 3 weeks later. (G) qRT-PCR analysis for TCF21 cDNA expression from GFP+ cardiac fibroblasts isolated by fluorescence-activated cell sorting from hearts of the 2 groups of mice shown. (H) Heart weight to body weight (HW/BW) ratio in the 2 indicated groups of mice. (I) Echocardiographic analysis of left ventricular (LV) wall thickness in diastole, (J) fractional shortening (FS) % and (K) LV internal diameter in diastole in the 2 indicated groups of mice. All data shown are mean +/− standard error of the mean. Samples were analyzed by a Kolmogorov-Smirnov test (exact p values are shown; B, H-K). For cell culture experiments, each n value shown represents fibroblasts from an independent animal. All data points in graphs represent biological replicates.

Figure 4.. Cardiac fibroblast-specific enforced expression of TCF21 reduces cardiac fibrosis.

(A) Schematic of the genetically modified mouse alleles used and temporal experimental outline. This cross also contained the Rosa26-eGFP conditional reporter allele to show labeling within recombined fibroblasts. Pdgfra-CreERT2 mice were used as controls. Tamoxifen (tam) was administered to 8–12-week-old mice for 3 consecutive days, followed by 1 week of angiotensin II/phenylephrine (Ang II/PE) stimulation prior to analysis. (B) Immunoblot analysis of TCF21 and smooth muscle ɑ-actin (ɑSMA) protein expression in the 2 indicated groups of mice with Ang II/PE treatment as shown in “A” in cardiac fibroblasts isolated by magnetic bead sorting with a mEF-SK4 antibody. Thirty micrograms of protein were loaded and GAPDH was used as a loading/normalization control. (C) Heart weight/body weight (HW/BW) ratio and (D) ventricular weight to body weight ratio (VW/BW) ratio in the indicated groups of mice. (E) Immunofluorescent histological images from hearts of the 2 indicated groups of mice with Ang II/PE stimulation stained for ɑSMA (white) and TCF21 (red) expression. Blue staining with DAPI shows nuclei while green staining is GFP expression in fibroblasts due to recombination of the Rosa26-eGFP reporter allele. Scale bars are 50 μm. The lower row of images is a higher magnification of the boxed area in the upper images. Scale bars are 10 μm. (F) Representative whole heart histological sections stained with Sirius red for collagen deposition from the 2 groups of mice shown with Ang II/PE stimulation. Scale bars are 1 mm. (G) Quantification of percentage area of tissue fibrosis as shown in “F”. (H) Immunofluorescent images from heart histological sections of Pdgfra-CreERT2 control versus Pdgfra-CreERT2:Col1a1-TCF21 mice for expression of periostin (POSTN, red) after 1 week of Ang II/PE. Nuclei are shown in blue with DAPI staining. Scale bars are 200 μm. All data shown are mean +/− standard error of the mean. Samples were analyzed by a Scheirer–Ray–Hare test with a Dunn’s post-hoc test and Benjamini Hochberg method for multiple test correction (exact adjusted p values are shown; C-D, G). All data points in graphs represent biological replicates.

Figure 5.. Cardiac fibroblast-specific enforced expression of TCF21 reduces expression of fibrosis genes.

(A) Volcano plot of differentially expressed genes detected by bulk transcriptomic sequencing of GFP+ cardiac fibroblasts (CFs) isolated by fluorescence-activated cell sorting (FACS) from hearts of Pdgfra-CreERT2:Col1a1-TCF21 versus Pdgfra-CreERT2 controls subject to 1 week of angiotensin II/phenylephrine (Ang II/PE) stimulation. (B) Dot plot of top gene ontology (GO) terms of biological processes with a false discovery rate (FDR) < 0.05 of all downregulated genes with a Log2 fold change (FC) > −1, FDR < 0.1. The gene ratio is the proportion of genes that are present within the GO category. (C) Heatmap of select z-score normalized downregulated genes associated with fibroblast differentiation and fibrosis function from hearts of the 2 groups of mice shown with 1 week of Ang II/PE stimulation. The legend above Panel C corresponds to the 3 and 2 samples shown by color above the heatmap. (D-G) Quantitative real time PCR (qRT-PCR) validation of select genes in GFP+ cardiac fibroblasts isolated by FACS of the 2 groups of mice after Ang II/PE. (H) Immunofluorescent histological images from the hearts of Pdgfra-CreERT2:Col1a1-TCF21 versus Pdgfra-CreERT2 control mice following 1 week of Ang II/PE stimulation, stained with antibody against cartilage oligomeric matrix protein (COMP). Nuclei are shown in blue with DAPI staining. The scale bar is 200 μm. All data shown are mean +/− standard error of the mean. Samples were analyzed by a Kolmogorov-Smirnov test (exact p values are shown; D-G). All data points in graphs represent biological replicates.

Figure 6.. TCF21 binds to genomic regions associated with cardiac fibroblast activation and fibrosis.

(A) HOMER analysis of DNA sequence binding motifs in TCF21 chromatin immunoprecipitation (ChIP)-seq peaks in cultured cardiac myofibroblasts with enforced TCF21 expression. Col1a1-TCF21 fibroblasts were infected with an adenovirus-encoding Cre recombinase (Ad-Cre) to induce TCF21 expression and expanded for 5 days. Transforming growth factor β (10 ng/mL) was added for the last 24 hours to induce myofibroblast differentiation. (B) Pie chart showing peak distribution of annotated peaks. (C-F) ChIP-seq peaks of select genes that are transcriptionally downregulated by TCF21. Each row represents a biological replicate (rep) that was derived from cardiac fibroblasts from 3 independent male and female mice.