Recapitulation of developmental mechanisms to revascularize the ischemic heart

Figures

• 

  Figure 1. The endogenous neovascular response after myocardial infarction. Immunostained transverse heart sections reveal considerable expansion of the capillary network by day 7 after infarction (B; compared with sham, A). Note the robust upregulation of Emcn, particularly in the endocardium, and the de novo capillary network that forms within the expanded epicardium (dotted line indicates epicardial-myocardial boundary; solid line indicates infarcted myocardium; representative of n = 15 day 7 MI hearts and n = 4 day 7 sham hearts). Scale bar: 200 μm (A and B). By qRT-PCR, upregulation of endothelial genes, exemplified by Pecam1, precedes induction of smooth muscle genes, Acta2 and Sm22α (C) (mean ± SEM; n = 4 hearts per time point). Two-tailed Kruskal-Wallis nonparametric test with Dunn’s post-hoc test for multiple comparisons; *P ≤ 0.05, **P ≤ 0.01.

• 

  Figure 2. Lineage-tracing neovascularization in the infarcted heart. PdgfbCreERT2; R26R-EYFP mice were injected with 2 doses of 2 mg tamoxifen to prelabel coronary endothelial cells, 2 weeks prior to myocardial infarction (A), and assessed before/after infarction by immunostaining. In noninfarcted mice, capillary (B and C), venous (D and E), and arterial (F and G) endothelial cells were efficiently labeled with EYFP (EYFP+ cells indicated with arrows); only the endocardium remained 100% unlabeled with the reporter (arrowheads in C). In the border zone after infarction (day 7), 73.6% ± 3.4% of capillary endothelial cells were EYFP+ (H–M; box in M enlarged in N and O; arrows indicate EYFP+ cells), thus preexisting or derived via angiogenesis from existing endothelium. However, 26.4% ± 3.4% of infarct/border zone endothelial cells were EYFP– (H–J and P), including long capillaries extending into the infarct (K–O; arrowheads indicate EYFP– ECs). EYFP– cells were Pdgfb– prior to injury and thus newly derived, either via vasculogenesis from a nonendothelial source or from the endocardium. MI, myocardial infarction; I/B, infarct/border zone; myo, myocardium; epi, epicardium; Tam, tamoxifen. Scale…

• 

  Figure 3. Remodeling of the endocardium after myocardial infarction. Immunostaining revealed increased trabeculation of the endocardial surface following MI: a representative sham-operated heart (A), compared with MI hearts after 2 (B) and 7 (C) days. Whereas, in the uninfarcted heart, large coronary arteries (red arrows) and veins (green arrows) were restricted to the epicardial side of the ventricle (D), new vessels appeared on the endocardial side of the infarct (E and F), coincident with compaction of trabeculae. Altered endocardial morphology was detected as early as 24 hours, with formation of cavities and finger-like protrusions (G–I). Coalescence of endocardial cells and trabecular compaction between days 4 and 7 (J–M; box in K enlarged in L) coincided with appearance of subendocardial vessels (arrowheads, M). LV, left ventricle; RV, right ventricle; epi, epicardium; endo, endocardium. Scale bars: 1 mm (A–F); 100 μm (G, K, and J); 200 μm (H, I, and M). The boxed area in K is magnified 2-fold in L. Representative of 24 hours: n = 5 (2 sham); 2 days: n = 6 (3 sham); 4 days: n = 5 (2 sham); 7 days: n = 15 (4 sham); 14 days: n = 5 (2 sham).

• 

  Figure 4. Endocardial remodeling contributes new subendocardial vessels. Detection of αSMA+ cells by immunostaining suggested either recruitment of smooth muscle precursors or endothelial-to-mesenchymal transition (A–D, arrowheads in B), consistent with expression of Snai1, in endocardial cells (A). Serial sections, approximately 50 μm apart, revealed progressive closure of subendocardial vessels: in (E), the lumen on the endocardial surface was incompletely enclosed (arrowhead) and only the enclosed side had αSMA support, whereas, in a more apical section (F), the equivalent lumen was fully enclosed (arrowhead) and surrounded by αSMA+ mural cells. Later detection of SM-MHC+ cells, from 7–14 days (arrowheads, G), confirmed a mature smooth muscle phenotype (G). By day 14, some subendocardial vessels were fully enclosed by SM-MHC+ cells (red arrowheads, H; enlarged in I), whereas larger vessels remained devoid of VSMC support (green arrowheads, H; enlarged in J). Scale bars: 100 μm (A–D); 500 μm (E, F, and H); 50 μm (G and J); 10 μm (I). An increase in small/medium SM-MHC+ and SM-MHC– vessels was observed by day 7 and an increase in large SM-MHC+ vessels was observed by day 14.…

• 

  Figure 5. Marker profiling supports an endocardial origin of de novo subendocardial vessels. By immunostaining, VEGFR2 was expressed in capillaries (A) and veins weakly (B and C) but not arteries (arrowhead, B) or endocardium (A and D). Subendocardial vessels, formed after myocardial infarction (MI), did not express VEGFR2 (D, enlarged in E–G; I–K). Medium-power view of VEGFR2+ capillaries (H). Emcn (A and D–J) and endoglin/CD105 (L) were highly upregulated in endocardium and expressed in subendocardial vessels. Cx40-EGFP was expressed in preexisting coronary arteries (shown in uninjured heart, M), but not in de novo subendocardial vessels (N), consistent with a nonarterial origin. Cx40-EGFP is expressed in Purkinje fibers (white arrowheads, N and O), and these were displaced around forming subendocardial vessels in MI (yellow arrowheads, O) but not uninjured, hearts (N). In the PdgfbCreERT2; R26R-EYFP reporter line, the endocardium (arrowheads, Q) and all endothelial cells of subendocardial vessels formed after MI were entirely negative for EYFP (P–R), confirming their derivation from a cell type that did not express Pdgfb/EYFP prior to MI. Representative of n = 15 WT MI hearts (n = 4…

• 

  Figure 6. The coronary sinus contributes angiogenic sprouts after myocardial infarction, recapitulating the developmental role of the sinus venosus. Immunostaining on transverse sections of day 7 hearts after MI reveals overt sprouting from the coronary sinus vein (A, enlarged in B–D; sprouting capillaries indicated by arrowheads). Apelin+ cells, a marker of sinus venosus in the embryo, comigrate with the PECAM1+ cells (E, enlarged in F–H), a smaller proportion of which also coexpress αSMA (G and H). A gradient of apelin+ cells was seen, diminishing in number from the atrioventricular sulcus (I, enlarged in J–L), tracking along the epicardium. In cultures 2 days after MI (M), atrioventricular explants produced PECAM1+ vascular sprouts by day 7, following outgrowth of epicardial cells at 3–4 days (N, enlarged in O). In contrast, ventricular explants resulted in outgrowth of epicardial cells but no vascular sprouts, only isolated PECAM1+ cells (P). Representative immunostaining of n = 15 day 7 MI hearts (4 sham); n = 4 day 2 MI explant cultures (from each heart, 1 right AVS, 1 left AVS, and 4 ventricular explants were plated in separate wells). LV, left ventricle; LA, left atrium; epi,…

• 

  Figure 7. Growth of a vascular network after myocardial infarction within the reactivated epicardium. Shown by immunostaining, the single-cell layer epicardium of the noninjured mouse heart (A) expands rapidly in response to injury, from day 2 after MI (B), continuing to day 14 (C–F). A network of capillaries arises within the activated epicardium between day 4 and day 7 (PECAM1+ endothelial cells in green; αSMA+ vascular smooth muscle cells in red (C–E). Epicardial capillaries remodel to form arterioles (F and G), acquiring smooth muscle support (D–H) and connecting with the underlying coronary vasculature (arrowhead in G). Maximal epicardial activity was observed in the proximity of the infarct (H), with only modest thickening over remote regions of the LV (I). Representative of 24 hours: n = 5 (2 sham); 2 days: n = 6 (3 sham); 4 days: n = 5 (2 sham); 7 days: n = 15 (4 sham); 14 days: n = 5 (2 sham). qRT-PCR confirmed the reexpression of fetal epicardial genes, Wt1, Tbx18, Tcf21, and Aldh1a2 (J, fold change relative to day 2 MI; n = 4 separate animals per time point; mean ± SEM; 2-tailed Kruskal-Wallis nonparametric test with Dunn’s post-hoc test for multiple comparisons; *P ≤ 0.05, ***P…

• 

  Figure 8. The epicardial lineage contributes pericytes, but not endothelial or smooth muscle cells, after myocardial infarction. Wt1 lineage tracing (Wt1CreERT2/+; R26REYFP) with immunostaining revealed no contribution to PECAM1+ endothelial cells after infarction (A) but revealed extensive contribution of NG2+ pericytes by day 7 (B and C), some of which incorporated into infarct/border zone vessels by day 14 (D, enlarged in E). No contribution to αSMA+/SM-MHC+ smooth muscle cells was detected at any stage, shown at day 14, when nonlabeled smooth muscle was abundant in the epicardial vasculature (F, equivalent regions in I–N). EYFP-labeled epicardial cells (GFP+) migrate from the epicardium toward the infarct (F). Arterioles within the expanded epicardium express the venous marker Emcn, yet they are supported by SM-MHC+ cells (G, enlarged in H). Epicardial lineage cells (GFP+) do not contribute endothelial or smooth muscle cells (SM-MHC+; I, enlarged in J; same arterioles shown in G and H; αSMA+; K, enlarged in L; M, enlarged in N), although they closely associate with newly formed vessels (J, L, and N). Representative of n = 8 hearts at day 7 and n = 6 hearts at day 14. epi, epicardium;…

• 

  Figure 9. Thymosin β4 expression after myocardial infarction. qPCR shows that Tmsb4x, the gene encoding Tβ4, is endogenously upregulated in the heart within 24 hours of infarction (A; n = 3 separate hearts per time point; box-and-whisker plots show mean ± minimum/maximum; 2-tailed Kruskal-Wallis nonparametric test with Dunn’s post-hoc test for multiple comparisons; *P ≤ 0.05, **P ≤ 0.01. ). By immunofluorescence, Tβ4 levels were highest in the expanding epicardium and upregulated in capillaries (B). Tβ4 levels in endocardial cells peaked at day 4 (C and D). Staining on Tβ4KO sections confirms KO of Tβ4 and antibody specificity (E). Upregulation can be appreciated when compared with a sham heart (F). By day 7, Tβ4 levels were reduced in endocardium (G), but elevated in cardiomyocytes surrounding remodeling subendocardial vessels (arrowheads, H) and in cells underlying the endocardium (I), many of which coexpressed αSMA (J; indicated by arrowheads). Of note, fully formed vessels without associated delaminating cells, indicated by the asterisk in G, contained the lowest Tβ4 levels. Representative of n = 4 WT day 4 hearts; n = 8 WT day 7 hearts and n = 8 KO day 7 hearts. MI, myocardial…

• 

  Figure 10. Thymosin β4 is required for epicardial activation, epithelial-to-mesenchymal transition, and expansion of the epicardial neovascular network. Epicardial activation following myocardial infarction was diminished in Tβ4KO hearts (A and B). By immunostaining, large regions of WT epicardium contained WT-1+ cells by day 7 (A), compared with only small regions in Tβ4KO hearts (B). This was reflected in whole-heart qRT-PCR, with reduced expression of Wt1 (C), Tbx18 (D) and Aldh1a2 (E) by 2-way ANOVA (n = 3 separate hearts per time point, mean ± SEM indicated). Post-hoc tests with Bonferroni correction confirmed significant reduction of Wt1 at day 7 and Aldh1a2 at day 4. **P ≤ 0.01, ***P < 0.001. Defects in epicardial mobilization were confirmed in explant culture: WT explants outgrew within 7 days (F), whereas KO explants produced fewer epicardial cells, and these failed to spread and form an epithelial layer (G; n = 3 of 8 cultures); moreover, several failed to outgrow altogether (H; n = 5 of 8 cultures). Coincident with expansion, a smooth muscle–lined vascular network extended throughout the WT epicardium (I). Even selecting for regions of greatest epicardial expansion in Tβ4KO…

• 

  Figure 11. Thymosin β4 is required for trabecular compaction and endocardial contribution of de novo vessels. Compared with WT (A), accumulation of trabeculae (B) and abnormal lumen morphology (C) were evident, by immunofluorescence, in Tβ4KO hearts from day 7 and more apparent by day 14 (D–F), a stage when compaction is mostly complete in WT (D); yet large trabeculae and lumina persisted in Tβ4KO hearts (E and F), suggesting a failure to undergo compaction remodeling and coalescence of new vessels. These defects manifested as a significant reduction in the incidence of subendocardial vessels formed (representative sections and quantification, G) from n = 10 hearts per genotype; 1-way ANOVA with Bonferroni correction for multiple comparisons; *P ≤ 0.05, **P ≤ 0.01, ***P < 0.001. This was associated with a striking reduction in αSMA+ cells in close proximity to Emcn+ endocardial cells (H–K). Scale bars: 1 mm (A, B, D, and E); 100 μm (C, F, H, and J); 20 μm (I and K)

• 

  Figure 12. Neovascularization is diminished in the infarct border zone of thymosin β4 KO hearts. By immunofluorescence, vascular density within the border zone was significantly reduced in Tβ4KO hearts, compared with WT (A–D, quantified in E), and vessels appeared less mature (B, compared with A). Fewer vessels acquired smooth muscle support (D, compared with C, quantified in F). By qRT-PCR, endothelial cell markers were not significantly reduced, with the exception of endoglin (G), whereas smooth muscle markers, Acta2, Sm22a and Notch3, were all significantly reduced in Tβ4KO hearts (H). These differences were reflected in quantification of immunofluorescence signal intensity, showing no significant reduction of endothelial markers (I) but a significant reduction of smooth muscle markers (J). Sections are representative of n = 9 hearts per genotype and n = 9 quantified. BZ, border zone; endo, endocardium; epi, epicardium. Scale bars: 500 μm (A and B); 50 μm (C and D). Statistical analyses (E and F): Mann-Whitney test (2-tailed); (G–J) 2-way ANOVA with Bonferroni correction for multiple comparisons; scatter plots: each data point represents a separate animal with mean ± SEM; box-and-whisker…