Experimental myocardial infarction triggers canonical Wnt signaling and endothelial-to-mesenchymal transition

Abstract

Despite available therapies, myocardial infarction (MI) remains a leading cause of death worldwide. Better understanding of the molecular and cellular mechanisms that regulate cardiac repair should help to improve the clinical outcome of MI patients. Using the reporter mouse line TOPGAL, we show that canonical (β-catenin-dependent) Wnt signaling is induced 4 days after experimental MI in subepicardial endothelial cells and perivascular smooth muscle actin (SMA)-positive (SMA(+)) cells. At 1 week after ischemic injury, a large number of canonical-Wnt-positive cells accumulated in the infarct area during granulation tissue formation. Coincidently with canonical Wnt activation, endothelial-to-mesenchymal transition (EndMT) was also triggered after MI. Using cell lineage tracing, we show that a significant portion of the canonical-Wnt-marked SMA(+) mesenchymal cells is derived from endothelial cells. Canonical Wnt signaling induces mesenchymal characteristics in cultured endothelial cells, suggesting a direct role in EndMT. In conclusion, our study demonstrates that canonical Wnt activation and EndMT are molecular and cellular responses to MI and that canonical Wnt signaling activity is a characteristic property of EndMT-derived mesenchymal cells that take part in cardiac tissue repair after MI. These findings could lead to new strategies to improve the course of cardiac repair by temporal and cell-type-specific manipulation of canonical Wnt signaling.

Fig. 1.

Induction of Wnt and Dkk family members after MI. (A) RT-PCR analysis for the expression of Wnt pathway mediators was performed on mRNA isolated from the hearts of C57BL/6J mice 5 days (5d) after being subjected to either permanent LAD occlusion or sham surgery. Results from three representative independent control and MI mice are shown (from six samples). (B) Quantification by real-time quantitative RT-PCR. Bars represent mean ± s.e.m. using Student’s t-test: *P<0.05; ***P<0.001; NS, not significant. Control sham, n=6; MI samples, n=6.

Fig. 2.

Canonical Wnt signaling activity in the normal adult mouse heart. Whole-mount and histological analyses of canonical Wnt activity in the adult heart using the TOPGAL mouse line. (A) X-gal-stained whole-mount heart shows canonical Wnt activity in the great vessels at the base of the heart. Tissue sections of this area show canonical Wnt activity in the media and intima of aortic (center left image), pulmonary (center right), and coronary arteries and subepicardial microvasculature (right). (B) Whole-mount X-gal staining of partially dissected heart to show cells with canonical Wnt activity in cardiac valves (left). Histological sections (two right-most panels) reveal cells positive for canonical Wnt activity in the connective tissue of outflow tract (aortic) and atrioventricular (mitral) valves. (C) Anti-CD31 antibody stain (brown) on an X-gal-stained section from TOPGAL mouse aorta demonstrates canonical Wnt activity in vascular endothelial cells (arrow). (D) Anti-SMA antibody binding (brown) on an X-gal-stained aorta section indicates canonical Wnt activity (blue) in perivascular SMA⁺ cells. The three right panels are IF images of tissue sections stained with antibodies recognizing SMA in vascular smooth muscle cells (green) and β-galactosidase (red). Merged images (far right) show colocalization of the two antigens in cells of the intima and media areas of the aorta. DAPI was used for counterstaining of cellular nuclei (blue).

Fig. 3.

Activation of the canonical Wnt signaling pathway after experimental MI. MI was induced in TOPGAL mice by permanent LAD ligation. Sham-operated animals underwent the same surgical procedure without blood vessel ligation. (A) Schematic outline of the experimental time line and corresponding stages of the ischemic injury response. (B) Hearts were isolated at 24 hours, 4 days, 7 days and 3 weeks post-MI, and stained with X-gal to assess canonical Wnt activity. Upper panels: front views with visible sutures; middle panels: rear views; lower panels: higher-magnification images of interest areas and infarct sites. In sham-operated control hearts, β-galactosidase activity staining is indistinguishable from normal (i.e. non-operated) hearts at all time points after surgery, with canonical Wnt activity around the great vessels. The displayed sham examples are from 7 days after surgery. Similarly, there are no visible changes in β-galactosidase staining 24 hours after LAD occlusion, compared with non-operated heart. By contrast, starting at day-4 post-MI, numerous X-gal-positive cells (arrows) appear throughout the heart around blood vessels. By 7-days post-MI, X-gal-positive cells with canonical Wnt activity are present around the infarct and peri-infarct areas (arrows). β-galactosidase (canonical Wnt pathway) activity is undetectable in the injury area 3 weeks after experimental MI. (C) Real-time quantitative RT-PCR analysis using RNA isolated from mouse hearts at the indicated time points after LAD occlusion. RNA samples from sham-operated animals isolated at equivalent time points after surgery served as controls. Sham samples had comparable expression levels at all stages after injury. In MI samples, peak expression levels of putative target genes of canonical Wnt signaling coincided with the time of highest β-galactosidase activity, at day-7 post-MI, except Myc, which was induced earlier (at 24 hours). *P<0.05; **P<0.01; ***P<0.001; NS: not significant. Sham, n=6; 24-hours post-MI, n=3; 7-days post-MI, n=6; 3-weeks post-MI, n=3.

Fig. 4.

Canonical Wnt activity marks endothelial cells and myofibroblasts after MI. Histological analysis of canonical Wnt activity was performed on TOPGAL mouse hearts 7 days after LAD occlusion or sham surgery. (A) Tissue sections of X-gal-stained mouse hearts shows that, in sham cardiac tissue, β-galactosidase (canonical Wnt signaling) activity is limited to the endothelium and pericytes of coronary vessels and in microvascular endothelial cells (left panel). After MI, in addition to coronary vasculature and microvasculature, canonical Wnt signaling marks newly appearing subepicardial mesenchymal-like cells surrounding the infarct area and a subset of epicardial cells (middle two panels). By contrast, cells with active canonical Wnt signaling are not present in subendocardial regions (right panel). (B) Immunohistological analyses of sections from TOPGAL mouse cardiac tissue. IHC using the anti-CD31 antibody on X-gal-stained cardiac tissue shows canonical Wnt activity (blue) in endothelial cells (brown) in the peri-infarct area (white arrows). The three right panels display IF results obtained using the anti-β-galactosidase (green) and -CD31 (red) antibodies, and the corresponding merged image. Canonical Wnt activity marks endothelial cells (white arrows). Examples of endothelial cells without canonical Wnt activity (red arrows) and non-endothelial, canonical-Wnt-pathway-active cells (green arrows) are indicated. DAPI was used for counterstaining of cellular nuclei (blue). (C) IHC analysis using the anti-SMA antibody on sections of X-gal-stained cardiac tissue from TOPGAL mice shows canonical Wnt activity (blue) in SMA⁺ cells (brown) in the peri-infarct area (white arrows). The three right panels display IF results obtained using the anti-β-galactosidase (green) and -SMA (red) antibodies, and the corresponding merged image. Blue is nuclear DAPI counterstain. Canonical Wnt activity marks SMA⁺ cells (white arrows). Green arrows point to SMA^– cells with canonical Wnt pathway activity and red arrows to SMA⁺ cells lacking canonical Wnt activity. A portion of SMA^– cells with canonical Wnt pathway activity are macrophages (supplementary material Fig. S1). (D) Canonical-Wnt-activity-marked endothelial cells were quantified by counting the number of double X-gal/CD31-stained and single CD31-positive cells per visual field. Canonical-Wnt-active SMA⁺ cells were quantified by counting the number of SMA-expressing X-gal-positive cells as a percentage of the total cell population in a given visual field, or as a percentage of the total SMA⁺ cell population. *P<0.05; ***P<0.001. Control sham, n=5; MI samples, n=5.

Fig. 5.

EndMT is induced post-MI. (A) IF analysis with antibodies recognizing CD31 (red) and SMA (green) on sections of mouse heart isolated 7 days post-MI. Lower panels are higher magnifications of the boxed areas. The corresponding merged images on the right show cells coexpressing CD31 and SMA (arrowheads). DAPI was used for counterstaining of cellular nuclei (blue). (B) FACS analysis with anti-CD31 and -SMA antibodies using cells isolated from peri-infarct tissue 7 days after MI or sham surgery. There is a marked increase in the population of CD31 and SMA double-positive cells on day 7 post-MI (red circle). The complete set of FACS data is included in supplementary material Fig. S2. (C) Quantification of CD31 and SMA double-positive cells in FACS analysis samples. ***P<0.001. Control sham, n=3; 7 days post-MI, n=3. (D) Real-time quantitative RT-PCR analysis for the expression of epithelial-mesenchymal transition (EMT)- or EndMT-associated genes using RNA samples from TOPGAL mouse hearts isolated at defined stages after MI as indicated. Peak expression levels of putative EMT- and/or EndMT-associated genes take place during granulation tissue formation, at 7 days post-MI. ***P<0.001; NS: not significant. Control sham, n=6; 24 hours post-MI, n=3; 7 days, n=6; 3 weeks, n=3. (E) IF analysis on sections from TOPGAL mouse hearts using the anti-β-galactosidase (green) and -Snail (red) antibodies. DAPI was used for nuclear counterstaining (blue). Right panel is a higher magnification of the boxed area in the merged image. Merged images illustrate that canonical-Wnt-signaling-active cells also express the EMT-regulator protein Snail (orange; arrowheads).

Fig. 6.

Canonical-Wnt-signaling-marked SMA⁺ cells are of endothelial origin. End-SCL-lacZ double transgenic mice were treated every second day with tamoxifen for 10 days to induce Cre-ER^T recombinase and genetically mark endothelial cells and their progeny. At 7 days after the last tamoxifen dose, mice were subjected to LAD occlusion or sham surgery. (A) X-gal staining and histological analysis 7 days after coronary artery ligation or sham surgery. In sham-operated mice, induction of Cre recombinase exclusively labels endothelial cells in the subepicardial area (upper panel; arrows). By contrast, in day-7 post-MI tissue sections, β-galactosidase activity marks both endothelial and mesenchymal-like cells (arrows), suggesting that the latter are of endothelial origin (bottom panel). (B) IHC analysis using the anti-SMA antibody on sections of X-gal-stained cardiac tissue of tamoxifen-treated end-SCL-lacZ mice 7 days post-MI. Double β-galactosidase/SMA-positive cells are observed in the peri-infarct subepicardial region of MI samples, but not in sham controls, suggesting that SMA⁺ cells are of endothelial origin. The lower left panel depicts a higher magnification of the boxed area. Arrows indicate SMA/β-galactosidase double-positive cells. The graph is quantification of canonical Wnt activity and SMA double-positive cells as percentage of the total SMA⁺ cells in subepicardial tissue. ***P<0.001; n=14 visual fields. (C) IHC analysis using the anti-β-catenin antibody on sections of X-gal-stained cardiac tissue of tamoxifen-treated end-SCL-lacZ mice 7 days post-MI. Left panel: sham control shows β-catenin accumulation in the intercalated discs of cardiomyocytes (white arrowheads) and a subset of endothelial cells (white arrows). Middle and right panels: at 7 days post-MI, nuclear β-catenin is also present in mesenchymal cells (black arrows). Black arrowheads indicate β-catenin staining in the intercalated discs of cardiomyocytes. Right panel depicts a greater magnification of area boxed in the middle image.

Fig. 7.

Activation of canonical Wnt signaling causes mesenchymal transition in cultured endothelial cells. The effects of canonical Wnt signaling activation were investigated in adult BAECs and the mouse brain bEnd.3 endothelial cell line by treatment with the canonical Wnt signaling activator BIO. (A) BIO induces a mesenchymal phenotype in BAECs within 24 hours of treatment. (B) Real-time quantitative RT-PCR analysis using RNA samples from BIO-treated BAECs. BIO induces the expression of canonical Wnt signaling gene targets (Axin2, TCF7) and EndMT-associated genes (Slug, SMA, Col1A1), whereas it leads to the downregulation of the endothelial-specific gene CD31. Induction of canonical-Wnt-pathway target genes (Axin2, TCF7) and downregulation of CD31 occurs within 24 hours of BIO treatment, whereas induction of SMA and Col1A1 is not observed until after 3 days of exposure to BIO. (C) IF analysis of bEnd.3 cells treated for 48 hours with BIO (+) or its inactive analog MeBIO (–) and stained with antibodies recognizing CD31 (green) and SMA (red). BIO treatment leads to downregulation of CD31 and upregulation of SMA, consistent with the molecular data obtained in BAECs. (D) IF analysis of bEnd.3 cells treated for 1, 4 or 24 hours with BIO (+) or MeBIO (–) and stained with antibodies recognizing β-catenin (red). BIO treatment causes nuclear accumulation of β-catenin, a hallmark of canonical Wnt signaling activation. DAPI stain (blue) marks cellular nuclei. All BIO treatment time points (1, 4 or 24 hours) showed nuclear β-catenin (arrowheads). The images are from the 4-hour exposure. (E) Luciferase analysis using protein extracts isolated from bEnd.3 cells transiently transfected with Super TOPFlash and pRL-TK Renilla Luciferase Reporter for normalization. bEnd.3 cells were treated with BIO, MeBIO or vehicle (DMSO) for 6 hours. BIO addition increases luciferase activity, indicating induction of the canonical Wnt pathway.

Fig. 8.

Schematic model of cardiac repair cell production after MI. Canonical Wnt signaling marks endothelial cells undergoing mesenchymal transition (blue), yielding a migratory mesenchymal cell type that expresses both endothelial (CD31) and mesenchymal (SMA) markers at low levels. The mesenchymal intermediate cell type, which is also marked by canonical Wnt signaling, has dual potential to differentiate into endothelial cells and new blood vessels, or give rise to scar-tissue-producing myofibroblasts.