Myocardial TGFβ2 Is Required for Atrioventricular Cushion Remodeling and Myocardial Development

Abstract

Among the three transforming growth factor beta (TGFβ) ligands, TGFβ2 is essential for heart development and is produced by multiple cell types, including myocardium. Heterozygous mutations in TGFB2 in patients of connective tissue disorders result in congenital heart defects and adult valve malformations, including mitral valve prolapse (MVP) with or without regurgitation. Tgfb2 germline knockout fetuses exhibit multiple cardiac defects but the role of myocardial-TGFβ2 in heart development is yet to be elucidated. Here, myocardial Tgfb2 conditional knockout (CKO) embryos were generated by crossing Tgfb2^flox mice with Tgfb2^+/-; cTntCre mice. Tgfb2^flox/- embryos were normal, viable. Cell fate mapping was done using dual-fluorescent mT/mG^+/- mice. Cre-mediated Tgfb2 deletion was assessed by genomic PCR. RNAscope in situ hybridization was used to detect the loss of myocardial Tgfb2 expression. Histological, morphometric, immunohistochemical, and in situ hybridization analyses of CKOs and littermate controls at different stages of heart development (E12.5-E18.5) were used to determine the role of myocardium-derived TGFβ2 in atrioventricular (AV) cushion remodeling and myocardial development. CKOs exhibit a thin ventricular myocardium, AV cushion remodeling defects and developed incomplete AV septation defects. The loss of myocardial Tgfb2 resulted in impaired cushion maturation and dysregulated cell death. Phosphorylated SMAD2, a surrogate for TGFβ signaling, was "paradoxically" increased in both AV cushion mesenchyme and ventricular myocardium in the CKOs. Our results indicate that TGFβ2 produced by cardiomyocytes acting as cells autonomously on myocardium and via paracrine signaling on AV cushions are required for heart development.

Keywords: AVSD; SMAD2; TGFβ2; atrioventricular cushion; mitral valve; myocardium.

Figure 1

cTntCre efficiently deletes Tgfb2 in the myocardium. (A) Genomic PCR with DNA extracted from the myocardium shows a near complete deletion of the Tgfb2 floxed allele in the presence of Cre recombinase. All embryos containing the cTntCre transgene (CKO) contain the deleted 174bp band while their non-transgenic or wildtype (WT) littermates have the intact ~1 kb product amplification. DNA from tail biopsy (ctrl1) does not show the deleted band, suggesting Cre expression is specific to the myocardium (ctrl1). There is no amplification with wild-type DNA (ctrl2) as this PCR only detects floxed and deleted bands for Tgfb2. No DNA template control also indicated (ctrl3). (B–D) The expression of Tgfb2 mRNA is significantly reduced in the ventricular myocardium of E12.5–13.5 cTntCre; Tgfb2 conditional knock out (CKO) embryos compared to littermate controls (p = 0.0374, multiple t-test; n = 3). scale bar = 10 μm. (E–J) Cardiomyocytes were lineage traced using mT/mG dual fluorescent reporter where GFP expression acts as a surrogate for Cre activity. Some conditional knockouts (CKOs) show a greater myocardialization of AV cushion, marked by an increased number of GFP positive cells (F), compared to littermate controls (E). CKOs also have a range of cardiac malformations including dysmorphic AV cushion (F,G) and ventricular septal defect (VSD; F,G; white arrows) which may continue to persist over development (J) or be sealed (I). scale bar = 200 μm.

Figure 2

Conditional deletion of Tgfb2 in early cardiomyocytes disrupts myocardial development, AV cushion remodeling, and AV septation. (A–D) Hematoxylin and eosin (H&E) (A,B) and cardiac muscle actin (C,D) staining show thinning of right ventricular myocardium in cTntCre; Tgfb2^−/flox (CKO) embryos compared to littermate controls (E13–E13.5). (C′,D′) are magnified views of the regions boxed in (C,D), respectively. (E–G) H&E (E14.5) images show well developed interventricular septa in the control (E) but CKOs show atrioventricular septal defects, perimembranous ventricular septal defect, and abnormal ventricular myocardium (F,G). Additionally, AV cushions are dysmorphic and abnormally remodeled, compared to littermate/age-matched wild type embryos. (H,I) Although the surface area of AV cushions (H) shows a decreasing trend in CKOs, the data are not statistically significant (p > 0.05, Student’s t-test). There is significant thinning of the right ventricular myocardium (p = 0.011, two-tailed Student’s t test with Welch’s correction; n = 4) (I). Scale bars, 200 µm (A–D,E–G); 100 µm (C′,D′).

Figure 3

Cardiomyocyte-specific loss of Tgfb2 results in defective cushion remodeling and myocardial cell proliferation. (A–C) Periostin immunohistochemistry. Myocardial Tgfb2 CKOs have significantly reduced periostin expression (Mann–Whitney test, p < 0.05) in their endocardial cushions, compared to littermate controls (D–F). Phospho-histoneH3 immunohistochemistry and the quantification of myocardial cell proliferation between the two groups at E12.5–14.5 (Mann–Whitney, Exact p = 0.0571 (median: ctrl 3.0, cko 2.08)). Scale bars, 20 μm (A,B); 200 µm (D,E).

Figure 4

SMAD2 activation is increased in myocardial Tgfb2 CKO hearts. (A–F) phospho-SMAD2 immunohistochemistry shows that CKOs have higher pSMAD2 (i.e., surrogate marker of TGFβ signaling) in both AV cushions (A–C) and myocardium (D–F), compared to littermate wild type controls at E13.5–14.5 (AV cushions: unpaired t-test with Welch’s correction, two-tailed, * p = 0.0172 (mean: ctrl 37.2, cko 65.61); Myocardium: unpaired t-test, two-tailed, * p = 0.029 (mean: ctrl 37.5, cko 63.7). Asterisks indicate statistically significant values (C,D). Scale bars, 20 μm.