VEGF-C and aortic cardiomyocytes guide coronary artery stem development

Coronary arteries (CAs) stem from the aorta at 2 highly stereotyped locations, deviations from which can cause myocardial ischemia and death. CA stems form during embryogenesis when peritruncal blood vessels encircle the cardiac outflow tract and invade the aorta, but the underlying patterning mechanisms are poorly understood. Here, using murine models, we demonstrated that VEGF-C–deficient hearts have severely hypoplastic peritruncal vessels, resulting in delayed and abnormally positioned CA stems. We observed that VEGF-C is widely expressed in the outflow tract, while cardiomyocytes develop specifically within the aorta at stem sites where they surround maturing CAs in both mouse and human hearts. Mice heterozygous for islet 1 (Isl1) exhibited decreased aortic cardiomyocytes and abnormally low CA stems. In hearts with outflow tract rotation defects, misplaced stems were associated with shifted aortic cardiomyocytes, and myocardium induced ectopic connections with the pulmonary artery in culture. These data support a model in which CA stem development first requires VEGF-C to stimulate vessel growth around the outflow tract. Then, aortic cardiomyocytes facilitate interactions between peritruncal vessels and the aorta. Derangement of either step can lead to mispatterned CA stems. Studying this niche for cardiomyocyte development, and its relationship with CAs, has the potential to identify methods for stimulating vascular regrowth as a treatment for cardiovascular disease.

The vascular beds of each individual organ are shaped to optimize blood flow through the different tissues of the body. In the heart, 2 coronary artery (CA) stems are located at specific sites on the aorta, one at the right sinus and one at the left sinus of the aortic valve (Figure 1A). The stem sites are positioned to efficiently deliver oxygenated blood to the ventricular myocardium. Congenital anomalies causing stems to arise at abnormal sites can cause myocardial ischemia and sometimes lead to sudden death (1, 2). Most dramatic is when CAs stem from the pulmonary artery, which carries deoxygenated blood. Usually fatal if left untreated, this malformation (anomalous left CA from the pulmonary artery [ALCAPA]) requires surgery soon after birth and often entails lifelong complications (3, 4). Despite the clinical relevance of proper stem formation, the mechanisms that dictate this process during embryonic development are unknown.

Figure 1

CA stem development in mice. (A) Schematics showing top and right lateral views of the heart, including CA stem locations (black arrows). (B) Confocal images of the right lateral side of hearts from the indicated embryonic days labeled for VE-cadherin. Coronary vessels (cv) and ASVs (arrowheads) grow toward the future right CA stem site (red asterisk) on the aorta (ao, dotted lines). (C) Tissue section through the aorta showing the presence of ASVs (red, arrowheads) directly beneath the epicardium (WT1⁺, green) and a vessel-free region around the aorta. The right panel is a high-magnification view of the boxed region. (D) Luminal endothelial cells lining the aorta, but not the pulmonary artery (pa), form sprouts (arrow). Inset is a high-magnification view of the boxed region. (E) Apelin-nlacZ expression (green, arrows) in VE-cadherin⁺ (red) aortic endothelium and adjacent sprouts. (F–I) Image projections of the CA stem site. (F) Small endothelial extensions from the aorta (arrow). (G) Initial interactions between coronary vessels and the aorta consist of thin, single-celled connections (arrowhead). (H) Multiple, lumenized connections that receive blood flow (arrowheads) develop at the future stem site and begin to acquire smooth muscle cell (SMC) coverage. (I) The mature pattern is a single, larger bore vessel stemming from the right coronary sinus (arrowheads) directly distal to the valves. (J) CA stem development. epi, epicardium; EC, endothelial cell; L atrium, left atrium; lca, left CA; ot, outflow tract; ra, right atrium; R atrium, right atrium; r ven, right ventricle; RCA, right CA. Scale bars: 100 μm (B–D); 50 μm (F–I); 25 μm (E, right panel in C, and inset in D).

Much of what we have learned about CA stem formation is from studies performed in chick embryos (5–8). For over a century, observations made using general histological stains suggested that CA stems arise by endothelial budding off the aorta (9, 10). A breakthrough in our understanding came from careful histological analyses and transplantation studies in chicks showing that, instead of outward budding, blood vessels that surround the cardiac outflow tract — termed peritruncal vessels — grow inward and fuse with the aorta (6, 7, 11). Recent lineage tracing has suggested a similar invasion process in the mammalian heart (12). These studies led to the current model in which peritruncal vessels first invade and anastomose with the aorta at multiple sites around its base, before remodeling into 2 larger-bore vessels at the right and left stem sites (Figure 1A and ref. 5). In chicks, this process requires PDGF, FGF, and VEGF-B, since inhibiting these molecules during coronary development decreases the number of stems formed (5, 13, 14). In mice, the remodeling of CA stems requires regulation of BMP signaling by BMP-binding endothelial regulator (BMPER) (15). However, the specific mechanisms patterning their spatial arrangement on the aorta are unknown.

In all species studied, including humans, correct positioning of CA stems depends on proper morphogenesis of the outflow tract. Inhibition of cardiac neural crest (16) and second heart field development (17) results in outflow tract septation and rotation defects that are associated with abnormally placed CA stems on the aorta. Interestingly, CAs do not connect with the pulmonary artery under these conditions, even in the presence of dramatic morphological defects. Similar phenotypes have been observed in Cx43-, Tbx1-, and perlecan-null mice (18–20) as well as in retinoic acid–treated embryos (21). Coronary stems are also abnormal in human hearts with rotation defects called transposition of the great arteries (TGA) (22). Thus, the mechanisms patterning CAs on the aorta are altered when outflow tract rotation is disturbed, but the factors restricting them from the pulmonary artery remain intact.

Our study provides insight into the factors around the outflow tract and aorta that facilitate CA stem formation at the proper location. VEGF-C is highly expressed around the outflow tract and is required for normally patterned stems because it attracts robust vessel growth near the presumptive stem sites on the aorta. Although necessary for the proper establishment of peritruncal vessels, VEGF-C expression throughout the aorta and pulmonary artery suggests that it does not specify the exact location of the stem sites. Instead, we found that cardiomyocytes, a highly angiogenic cell type (23, 24), are present specifically in the wall of the aorta, but not in the pulmonary artery, and that they surround developing stems in both mouse and human hearts. Heterozygosity for islet 1 (Isl1) decreases aortic cardiomyocytes and leads to delayed, abnormally targeted stems. Misplaced stems remain tightly associated with aortic cardiomyocytes on hearts with outflow tract rotation defects, and cardiomyocytes induce ectopic endothelial connections with the pulmonary artery in vitro. Our data suggest that VEGF-C first induces vascular expansion around the outflow tract, followed by cardiomyocyte-assisted vessel growth specifically at the CA stem sites. Inhibition of either mechanism results in mistargeted CA stems. Understanding this process will begin to shed light on the molecular defects that cause congenitally misplaced CA stems and could suggest methods of generating new CAs following injury or disease.

CA stem formation in the mammalian heart. A stepwise, high-resolution characterization of CA stem development has not been performed in the mammalian heart. Therefore, we used whole-mount confocal microscopy to image blood vessels in mouse hearts isolated from E10.5 to birth. We found that peritruncal vessels surrounding the aortas and pulmonary arteries first appeared at E11.5 and arose from 2 different locations. First, coronary vessels migrated toward the base of the outflow tract as they grew over and into the heart (Figure 1B). Second, a previously undescribed plexus was detected on the aorta, which began as a single vessel that circled its lateral side before expanding at later developmental stages (Figure 1B and Supplemental Figure 1; supplemental material available online with this article; doi:10.1172/JCI77483DS1). These vessels developed directly beneath the aortic epicardium (Figure 1C) and were therefore termed aortic subepicardial vessels (ASVs). Similar vessels were seen occasionally in the analogous region of the pulmonary artery but with much less frequency and at lower densities (data not shown). ASVs appeared to be blood vessels, since they contained erythrocytes (Supplemental Figure 2A) and connected to aortic endothelium (see below). However, a subset of early ASV endothelial cells was positive for the lymphatic marker PROX1 but negative for the mature lymphatic marker LYVE-1 (Supplemental Figure 2, B and C). LYVE-1⁺ vessels appeared on the aorta much later at E16.5 (Supplemental Figure 2D). Both coronary vessels and ASVs expanded as development progressed (Figure 1B), but only coronary vessels persisted while ASVs mostly disappeared by E17.5 (Supplemental Figure 1). Thus, peritruncal vessels surrounding the outflow tract are composed of developing coronary vessels and ASVs.

Prior to any anastomosis event between the aorta and peritruncal vessels, the aortic endothelium often extended sprouts (Figure 1, D and E, and Supplemental Figure 3A), which never appeared on the pulmonary artery (Figure 1D and Supplemental Figure 3A). Sprouts were frequently 1 cell extending into the medial layer surrounding the aorta (Figure 1, D and E), but those consisting of 2 to 5 cells were also observed (Figure 1F). Extending cells and adjacent aortic endothelial cells expressed apelin, a marker for sprouting endothelium (Figure 1E). These sprouts were independent of coronary vessel or ASV development, since they still formed when peritruncal vessels were inhibited by an early injection of the VEGFR antagonist axitinib (Supplemental Figure 3B). The presence of aortic sprouts is in contrast to that in avian experiments, which report only the ingrowth of peritruncal vessels, with no budding of the aorta. This could be due to species differences or our use of high-resolution confocal microscopy, which can detect single-cell processes deep within tissue. It does not appear that the sprouts robustly contributed to the stem, since they were very small (1–5 cells) and did not grow extensively in the absence of peritruncal vessels (Supplemental Figure 3B). Taken together, these observations suggest that the environment surrounding the aorta is angiogenic, while the analogous region on the pulmonary artery is not, but that ingrowth of coronary vessels is the primary mechanism of stem formation.

Connections between the aorta and peritruncal vessels emerged between E12.5 and E13.5 and formed at locations that include both future CA stem sites (Figure 1G) and various points higher on the artery (Supplemental Figure 3A and Supplemental Figure 4). Connections never formed on the pulmonary artery (Supplemental Figure 3A). Initial fusions consisted of lumenless, single-cell thick bridges (Figure 1G). If these connections were positioned at the future stem site adjacent to the valve, they expanded in the next developmental stage. Expansion involved the formation of multiple, clustered connections with increased diameters that began to receive blood from the aorta (Figure 1H and Supplemental Figure 5A). The stem site was not located at the exact same spot in each heart but occurred within a range at or above where the valve leaflets converge (Supplemental Figure 6, A and B). At E14.5, vascular smooth muscle development was initiated around the CA stems (Supplemental Figure 5B), and, by E16.5, the mature stem morphology was seen (Figure 1I). In summary, the process of CA stem formation in the mammalian heart includes anastomoses of peritruncal vessels with a sprouting aortic endothelium, followed by the remodeling and maturation of only those connections at the stem sites (Figure 1J and Supplemental Figure 4).

Coronary vessels originate from at least 3 progenitor populations, the sinus venosus (25, 26), endocardium (25, 27), and proepicardium (28), the latter of which may first transition through the former cell types (28). We next performed Cre recombinase–based lineage tracing with Cre lines expressed in the sinus venosus and endocardium to assess their contribution to CA stems. Mice containing either the Rosa^mTmG or Rosa^tdTomato Cre reporter alleles were crossed with Apj-CreER or Nfatc1-CreER transgenic mice to lineage label the sinus venosus and endocardium, respectively. When analyzed at later embryonic stages, stems contained lineage-labeled cells from both Apj-CreER and Nfatc1-CreER traces, showing that both populations contributed to the CA stem endothelium (Figure 2). Lineage-labeled cells from both Cre lines were regularly found in both the right and left CA stems. Thus, both progenitor cell types contribute to right and left CA stem formation, a reasonable outcome given the close apposition of the stem region to growing sinus venosus–derived coronary vessels and the endocardium.

Figure 2

The sinus venosus and endocardium both contribute to CA stems. Tissue sections of hearts immunostained with endothelial markers (VE-cadherin or CD31) and a nuclear dye (DAPI). Lineage-traced cells (arrowheads) are present in CA stems from (A) Apj-CreER, Rosa^mTmG and (B) Nfatc1-CreER, Rosa^tdTomato embryos. Right panels are high-magnification views of the boxed regions. Scale bars: 100 μm.

VEGF-C is required for peritruncal vessel development. We next sought to identify the molecules involved in CA stem formation. Previous studies identified a role for VEGF-B in avian stem development (13), although, in mice, it appears to have a more prominent role in regulating cardiac vascularization and metabolism during injury rather than in development (29–32). To investigate whether other VEGF family members could have a role during mammalian coronary development, embryos were treated with the pan-VEGFR inhibitor axitinib in utero. This treatment could completely inhibit coronary vessel growth, identifying the VEGF pathway as a potential player during stem genesis (data not shown). Investigating the expression of different VEGF family members using lacZ knockin alleles revealed that VEGF-C was expressed in a pattern highly suggestive of a role in stem development. Vegfc-lacZ expression was detected within the walls of the outflow tract vessels and around their base (Figure 3, A–C). VEGF-C receptors were expressed in nearby cardiac endothelial cells. Peritruncal vessels (coronary vessels and ASVs) and the endocardium were positive for VEGFR2 and VEGFR3 (Figure 3D), while antibodies recognizing VEGFR2 also labeled aortic endothelium (Figure 3D). VEGF-A was also expressed in the heart, but it was largely expressed in cardiomyocytes throughout the ventricle (33) and not robustly expressed in the aortic wall or near its endothelium where stems normally form (Supplemental Figure 7). Thus, VEGF-C expression overlaps with the outflow tract and stem-forming regions (Figure 3C) and could signal to peritruncal and endocardial cells through VEGFR2 and VEGFR3.

Figure 3

VEGF-C is required for peritruncal vessel growth and CA stem formation and patterning. (A) Tissue sections through Vegfc–lacZ hearts showing expression (blue) in the aorta and pulmonary artery. (B) Boxed region in A showing expression throughout the vessel wall (brackets) and base of the outflow tract. (C) Schematic of VEGF-C expression (blue). (D) VEGFR2 is expressed in all cardiac endothelial cell types while VEGFR3 is expressed in coronary and ASV endothelium and in the endocardium (endo). Right panels are high-magnification views of the boxed regions and show expression in coronary vessels that share cell-cell junctions with endocardial cells (dotted lines). (E) Confocal images showing the absence of peritruncal vessels (VE-cadherin⁺, blue) around the outflow tract in VEGF-C–deficient hearts at E12.5 and E14.5. Cardiomyocytes are shown in red (cTnT). (F) Heterozygous (het) hearts have hypoplastic (hypo) peritruncal vessels. (G and H) CA stems (arrowheads) are present in wild-type but absent (arrows) or abnormal (arrowheads) in knockout (ko) hearts. The trough of each aortic valve sinus is traced with a dotted line for reference in comparison of stem location. (G) In many of the mutant hearts, sprouts connected to the endocardium (open arrowhead) appear to be extending up toward the aorta, even when other peritruncal vessels are absent. (I) Distribution of CA stem phenotypes in VEGF-C knockout embryos. n values are shown on the right. (J) Schematics of representative CA stems in VEGF-C–deficient outflow tracts. ao lum, aortic lumen; val, valve; ven, ventricle. Scale bars: 100 μm (A, B, left panel in D, and E–H); 50 μm (right panel in D).

Loss-of-function experiments were next performed to investigate the role of VEGF-C in CA stem formation. Depletion of VEGF-C had a dramatic effect. Although heart morphogenesis was grossly normal and other vessel beds were not affected (26), ASVs were completely missing and peritruncal coronary vessels were severely reduced in VEGF-C knockout animals (Figure 3E). Heterozygous animals exhibited an intermediate phenotype (Figure 3F). Many VEGF-C–deficient hearts did not have CA stems (Figure 3, G–I). Interestingly, in these knockout hearts, sprouts connected to the endocardium appeared to extend up toward the aorta (Figure 3G). When mutant animals did have stems, they were abnormally low on the aorta, connecting to the trough of the valve and pointing downward instead of to the side of the aorta above the aortic valvular sinus (Figure 3, H–J). This low positioning was closer to the deeper coronary vasculature that forms in knockout hearts. Data are shown for the right lateral side (right CA), but in most cases the left CA exhibited the same phenotype (Supplemental Figure 8A). These data indicate that VEGF-C is required for normal CA stem development due to its role in stimulating the first step of the process: the establishment of peritruncal vessels. In the absence of this growth factor, anomalous compensatory stems arise, possibly through interactions with endocardial sprouts.

Cardiomyocytes specifically populate the aortic wall where they are associated with developing CA stems. The knockout analysis and expression data described above show that VEGF-C is necessary for proper CA stem formation but that it does not pattern the precise connection site on the aorta. More specifically, VEGF-C is a powerful inducer of peritruncal vessel growth, but its expression extends into areas in which endothelial cells are not found, including the pulmonary artery and the vessel-free zone that surrounds the outflow vessels (Figure 1C and Figure 3, A–C). These observations suggest that other factors collaborate with VEGF-C–induced vessel growth to specify stem sites on the aorta. To identify candidate mechanisms, we looked for differences between the aorta and pulmonary artery by examining the spatial distribution of different cell types present at the relevant developmental stages. Our analysis included neural crest cells, smooth muscle cells, neurons, macrophages, epicardial cells, and cardiomyocytes. Of these, cardiomyocytes were the only cell type displaying an aorta-specific localization, suggesting a possible role in CA stem patterning.

Whole-mount imaging of E13.5 hearts showed that cardiomyocytes were abundant within the wall of the aorta but not in the pulmonary artery, in which they formed a sharp border at the level of the valve (Figure 4A). When this parameter was measured in tissue sections, cardiomyocytes were found to be significantly more distal from the valve in the aorta than in the pulmonary artery (Figure 4B). Measuring cardiomyocytes around the circumference of the aorta showed that they are significantly more concentrated at the left and right coronary sinuses when compared with the noncoronary sinus (Figure 4, C–E). The presence of aortic cardiomyocytes preceded stem formation, with aortic cardiomyocytes first observed at E11.5 (Figure 4E). At this time, they were in direct contact with luminal endothelial sprouts from the aorta (Figure 4F). In contrast, cardiomyocytes were mostly separated from the pulmonary artery endothelium where sprouting was never seen (Figure 4G). Sectioning through the entire outflow tract and using serial sections to quantify the total amount of contact between cardiomyocytes and luminal endothelium revealed that there are significantly more direct interactions with the aorta compared with the pulmonary artery (Figure 4H). Thus, cardiomyocytes are present at the correct location and appropriate developmental stage to pattern CA stem formation.

Figure 4

Aorta-specific cardiomyocyte development precedes CA stem formation. (A) Confocal image of the right lateral side of an E13.5 heart showing that cardiomyocytes (red, arrowheads) develop on the aorta distal to the valves but are absent in the analogous region of the pulmonary artery. The right coronary sinus (rcs) and noncoronary sinus (ncs) regions are indicated by square and curly brackets, respectively. (B) Quantification of cardiomyocytes within the aorta and pulmonary artery. (C) Quantification of cardiomyocytes among the different regions of the aorta. (D) Right and left lateral views of the aorta showing concentration of cardiomyocytes specifically around the right and left coronary sinuses (lcs) at the maturing stem sites (arrowheads). (E) Cardiomyocytes (arrowheads) are present in the aortic wall around the right and left coronary sinuses at E11.5 before stem formation. Aortic and pulmonary valve sinuses are schematized with dotted lines. (F) Cardiomyocytes directly contact the aortic endothelium (arrowhead) where aortic sprouts (sp) form — (G) but not the pulmonary artery (arrow), which never sprouts — as quantified in H. High-magnification views of the boxed regions are shown. (I) Developing CA stems (arrowhead) are closely associated with aortic cardiomyocytes. (J) Aortic cardiomyocytes are no longer present in adult hearts. (K) Developing human CA stems (arrowheads, dotted lines) are associated with aortic cardiomyocytes. OFT, outflow tract. All data represent mean ± SD. Each dot represents a value obtained from one sample. ****P < 0.0001; ***P = 0.0001 to 0.001; **P = 0.001 to 0.01. Scale bars: 100 μm (A, D, E, left panel in F, right panel in G, J, and K); 25 μm (right panel in F and left panel in G); 50 μm (I).

We next observed the relationship between aortic cardiomyocytes and developing CA stems. Both the right and left stems were always closely associated with aortic cardiomyocytes in embryonic hearts (Figure 4I and Supplemental Figure 9, A and B). However, these cardiomyocytes disappeared postnatally, and stems were surrounded instead by smooth muscle (Figure 4J). A similar association between aortic cardiomyocytes and CA stems was seen during human embryogenesis (Figure 4K). This was particularly notable, since human CAs, unlike rodent CAs, are located on top of ventricular myocardium. We observed that human CAs connect with the aorta through a layer of aortic cardiomyocytes before they localize to the heart surface (Figure 4K). Thus, cardiomyocytes could play an important role in guiding CA stem development in mouse and human hearts.

Understanding how aortic cardiomyocytes arise on the aorta could be useful in devising loss-of-function experiments that test their role in stem formation. There are at least 2 routes by which cardiomyocytes could localize to the aorta: by migrating up from the ventricle or by directly differentiating from progenitor cells present in the aortic wall. The latter would be analogous to the mechanism of early heart growth, in which there are cardiomyocyte progenitors at the outflow and inflow poles that continuously differentiate and add to the heart tube (34, 35). To investigate these possibilities, we performed time-lapse microscopy on isolated hearts with fluorescently labeled cardiomyocytes (Myh6-Cre, Rosa^mTmG). While labeled cells could be seen proliferating, they never appeared to migrate (Figure 5). Instead, we frequently observed the initiation and gradual increase in cardiomyocyte-specific GFP expression on the aorta distal to the main ventricle, suggesting de novo differentiation (n = 7 hearts). This was in contrast to control experiments in which fluorescently labeled endothelial cells (VE-cadherin-CreER, Rosa^tdTomato) migrated rapidly while continuously extending and retracting filopodia (data not shown). These data suggest that aortic cardiomyocytes differentiate in situ and that inhibiting progenitor populations could decrease their presence.

Figure 5

Aortic cardiomyocytes do not migrate up from the ventricle but differentiate in situ during explant culture. Images taken from time-lapse movies of E12.5 Myh6-Cre, Rosa^mTmG heart cultures. Cardiomyocyte migration is not observed (arrows) at the junction of the ventricle and aorta. Instead, there is the appearance of a new cardiomyocyte (arrowheads), which seems to have differentiated in situ (images are shown at higher magnification in bottom panels). Scale bars: 100 μm (top panels); 20 μm (bottom panels).

Isl1 heterozygous hearts exhibit decreased aortic cardiomyocytes and abnormally low CA stems. Deletion of the transcription factor Isl1 causes embryonic lethality and cardiac developmental defects (36). Without Isl1, cardiomyocyte progenitors in the second heart field fail to expand and contribute to heart growth, resulting in the lack of later developing structures, such as the right ventricle and outflow tract. We crossed the Mef2c-AHF-Cre mouse (37) to a Rosa^mTmG Cre reporter to lineage trace the second heart field and found that aortic cardiomyocytes descend from this cell population (Figure 6A). Since aortic cardiomyocytes arise from the second heart field lineage and appear to develop de novo from progenitor cells within the outflow tract, we tested whether deletion of one Isl1 allele would affect their cell numbers in the aorta. Observing cardiomyocytes using whole-mount confocal micrographs of wild-type and Isl1 heterozygous hearts revealed a reduction of this cell type in the aortic wall at both the right and left CA stem regions (Figure 6, B and C, and Supplemental Figure 8B). Measuring the presence of cardiomyocytes distal to the valves via optical sections showed a significant decrease in the cardiomyocyte distribution in mutant embryos (Figure 6D). We did not find any gross defects in heart morphogenesis, as outflow tract rotation was normal (Figure 6B) and myocardial thickness and total heart size were similar in wild-type littermates (Supplemental Figure 10). The amount of smooth muscle around the aorta near the stem site was not seriously affected (Figure 6E), and there were no obvious abnormalities in other cell types such as neurons (data not shown). Isl1 mutation did not produce a general vessel growth defect, since coronary sprouting over the ventricles in heterozygous mice was normal (Figure 6, B and F). Although we cannot rule out differences in cell types or functions not analyzed, Isl1 heterozygous mice appear to have a relatively well-isolated defect in aortic cardiomyocytes, making them useful in testing the role of these cells in CA stem formation.

Figure 6

ISL1 heterozygosity decreases aortic cardiomyocytes and disrupts normal stem development and positioning. (A) Second heart field lineage tracing (Mef2c-AHF-Cre, Rosa^mTmG) labels aortic cardiomyocytes (CMs, arrows). (B) Confocal images of wild-type and Isl1 heterozygous embryos showing decreased numbers of cardiomyocytes within the aorta (dotted line). The pulmonary artery is also outlined (solid lines). Cardiomyocytes (cTnT⁺) are shown in red; peritruncal vessels (VE-cadherin⁺) are shown in blue. The bottom panel shows only the cTnT channel where red lines indicate the stem position. (C) Boxed regions in A showing only the cTnT channel. (D) Quantification of the length distal to the valve occupied by aortic cardiomyocytes shows a significant decrease in Isl1 mutant hearts. (E and F) Other developmental parameters, (E) smooth muscle cell development and (F) coronary vessel growth, are unchanged in mutant hearts. (D–F) Data represent mean ± SD. Each dot represents a value obtained from one sample. (G) Isl1 heterozygous hearts have abnormally low stem positioning with respect to the valves (lines) in comparison to wild-type hearts. (H) Distribution of stem phenotypes observed in Isl1 mutant and wild-type hearts. n values are shown on the right. (I) The height of CA stems does not significantly rise above the level of aortic cardiomyocytes in both wild-type and mutant hearts, as evidenced by the paucity of points in the pink area (y > x) of the graph. These data were measured at E14.5. A linear regression for the data set (dotted line, y = 0.5999x – 36.39) indicates a positive correlation. ****P < 0.0001; NS ≥ 0.05. Scale bars: 100 μm.

CA stems in Isl1 mice were dramatically different than those in wild-type mice. Stems attached much lower on the aorta, often localizing to the trough of the valve (Figure 6G and Supplemental Figure 6, A and B). In addition, initial connections between peritruncal vessels and the aorta were frequently delayed (Figure 6H). This phenotype was seen in both the right and the left CA stems (Figure 6 and Supplemental Figure 8C). Stem sites were quantified by drawing a line at the top of the valves where the leaflets converge and measuring the distance from this line to where the stem connects to the aorta (Figure 6G and Supplemental Figure 6, A and B). Positive values indicate the stem orifice was above (distal to) the valve. Negative values indicate the stem orifice was below the valve. A value of 0 was assigned to a stem that connects at the exact level of the valve. Plotting this location against the region occupied by cardiomyocytes revealed a positive correlation between the 2 parameters (Figure 6I). This analysis highlighted that stems did not attach above the region where cardiomyocytes were found (i.e., no data points were significantly above the line y = x), even in Isl1 mutants with dramatically lowered aortic cardiomyocytes (Figure 6I). These data support a model in which cardiomyocytes within the aortic wall guide the formation of CA stems at their proper location on the lateral sides of the aorta just distal to the valves.

Misplaced CA stems associate with aortic cardiomyocytes in hearts with outflow tract rotation defects. During heart development, the outflow tract undergoes a rotation event as it is transformed from a single vessel into the aorta and pulmonary artery. This process positions the aorta between the atria and pulmonary artery (Figure 7A). TGA is a congenital heart defect that is a consequence of insufficient rotation, causing the aorta and pulmonary artery to be reversed or side by side (Figure 7B). This malformation is also associated with variations in CA stem placement (Figure 7B). We used Pax3-null mice (38), which have defective neural crest cell migration that frequently leads to TGA, to examine whether malrotation affects aortic cardiomyocyte growth and whether these cells still associate with CA stems under abnormal conditions.

Figure 7

Misplaced CA stems are correlated with aortic cardiomyocytes in hearts from Pax3-null embryos. (A and B) The side-by-side positioning of the aorta and pulmonary artery in hearts from (B) Pax3-null embryos compared with (A) wild-type embryos. Blue lines indicate the locations of CA stems. (C–H) Confocal images of hearts immunofluorescently labeled for VE-cadherin (endothelium) and cTnT (cardiomyocytes). CA stems (arrowheads) associate with aortic cardiomyocytes (brackets) in wild-type and in Pax3 knockouts. CA stems are outlined in solid lines in D–F. Outflow tracts are outlined with dotted lines in C–F. Images in G and H are maximum projections of optical sections through the aortic lumen captured from the right lateral side of the heart. Scale bars: 100 μm.

Analyzing Pax3 mutants that exhibit TGA (outflow tract septation without proper rotation) revealed that neither normal neural crest migration nor complete rotation was required for aorta-specific cardiomyocyte growth (Figure 7, C–H). Although the aorta and pulmonary artery were side by side in mutant hearts, cardiomyocytes were specifically present in large numbers on the aorta (Figure 7, D and F). Mutant hearts also had coronary vessels that exited the aorta ventrally (from the front) and dorsally (from the back) instead of laterally, from the right and left sides (Figure 7, A and B). However, the misplaced CA stems were always associated with a concentration of cardiomyocytes growing within the aortic wall (Figure 7, D, F, and H) (n = 12 hearts). These data provide further evidence of a developmental relationship between cardiomyocytes and CA stems.

Cardiomyocytes can induce ectopic connections with the pulmonary artery in vitro. We next performed gain-of-function experiments to investigate whether cardiomyocytes could induce vascular connections where they normally do not occur. Outflow tract explants were cultured adjacent to ventricular myocardium or control tissues. After 2 to 3 days, the explants were immunostained to determine whether ectopic vascular connections occurred specifically in the presence of cardiomyocytes. When cultured alone, explanted pulmonary arteries maintained their general shape, with a nonsprouting luminal endothelial layer surrounded by a vessel-free region (Figure 8A). Endothelial cells from lung explants did not invade the vessel-free region, even though they readily migrated into the myocardium at the base of the artery explant (Figure 8, B and D). In contrast, culturing myocardium adjacent to outflow tracts frequently resulted in connections between ventricular coronary vessels and pulmonary artery luminal endothelium (Figure 8, C and D). Thus, myocardium can mediate ectopic connections between the pulmonary artery and the coronary vessels in vitro, supporting the hypothesis that aortic cardiomyocytes facilitate this process during stem development.

Figure 8

Ectopic vessel connections with the pulmonary artery lumen can form in the presence of cardiomyocytes. (A–C) Confocal images of pulmonary artery outflow tract explants cultured alone or adjacent to other tissues. Endothelial cells are labeled in blue (VE-cadherin⁺), and cardiomyocytes are labeled in red (cTnT⁺). Schematics summarizing experimental setup and results are shown. (A) Pulmonary artery explants retain a luminal endothelial layer that does not sprout into the vessel-free zone (dotted line). (B) Endothelial cells within lung tissue do not connect with the pulmonary artery lumen but do migrate into the myocardial (myo) region at its base (arrow), which contains cardiomyocytes. (C) When cardiomyocytes from ventricular myocardium (ven myo) are placed alongside artery explants, ventricular coronary vessels frequently form connections (arrowhead) with the pulmonary artery endothelium. (D) Percentage of explant samples that did or did not form connections with the pulmonary artery lumen when cultured beside lung or ventricular myocardium. Scale bars: 100 μm (left and center panels in A–C); 25 μm (right panels in A–C).

In this study, we describe 2 morphological steps that are required for CA stem patterning — (a) the establishment of the peritruncal vessels that surround the aorta and (b) anastomosis of these vessels specifically with the aorta — and provide insight into the molecular and cellular mechanisms guiding both processes. Our data show that VEGF-C mediates step one. It is highly expressed throughout the outflow tract, and VEGF-C–deficient mice have absent ASVs and hypoplastic peritruncal coronary vessels. Step one is required for CA stem patterning, as evidenced by the delayed, misplaced stems that form in VEGF-C mutant hearts when the peritruncal vessels are reduced or missing. VEGF-C is expressed throughout the outflow tract, including regions in which endothelial cells do not grow, suggesting that it alone is not sufficient for stimulating connections with the aorta at defined locations. We found that cardiomyocytes are localized in a pattern suggestive of a role in step two. They are concentrated along the wall of the aorta in which stems form but are mostly absent from the pulmonary artery. Isl1 heterozygosity decreases aortic cardiomyocytes and results in delayed, abnormally placed stems. Misplaced stems on hearts with outflow tract rotation defects formed within regionally shifted aortic cardiomyocytes, and cardiomyocytes induced ectopic connections with the pulmonary artery in culture. These data support a step-wise model in which peritruncal vessels are attracted to the outflow tract from nearby vessel beds (the sinus venosus, endocardium, and surface of the aorta) in response to VEGF-C, after which cardiomyocytes facilitate the positioning of their fusion sites on the aorta (Figure 9). Failure to complete either step results in abnormal CA stem patterning.

Figure 9

Proposed model for the role of VEGF-C and aortic cardiomyocytes in CA stem formation. (A) VEGF-C stimulates vascular growth near the outflow tract, while a vessel-free zone directly surrounds the aorta and pulmonary artery. (B) Peritruncal vessels develop around the outflow tract but do not invade the vessel-free zone. Wild-type Isl1 expression levels in the embryo allow cardiomyocytes to differentiate specifically in the aortic wall where they support vessel growth and facilitate connections between peritruncal vessels and lumen endothelium. (C) The result is a correctly positioned CA stem (ca) on the aorta.

CA anomalies in humans can range from nonstandard stem origin on the aorta to defects in which the left coronary stem branches from the pulmonary artery (ALCAPA). Anomalous origins on the aorta have been estimated to occur in approximately 1% of the population (1). These are frequently asymptomatic but can cause myocardial ischemia and sudden cardiac death in a significant number of cases (2, 39). Anomalous CA stems can attach to the aorta in a number of different configurations, which may explain the spectrum of symptoms. The most clinically concerning anomalies for sudden cardiac death occur when the artery stem attaches to the opposite aortic sinus, which causes the vessel to travel between the aorta and pulmonary artery, e.g., the left CA arising from the right sinus. The direct cause of ventricular ischemia in these cases is not known but is thought to arise when the aorta and pulmonary artery apply pressure to the intervening CA, obstructing the vessel lumen. Understanding how to handle CA anomalies is particularly important for athletes and military recruits preparing to undergo intense training. CA anomalies are estimated to be responsible for 17% of sudden cardiac deaths in young competitive athletes (40). An American Armed Forces Institute of Pathology study found that 33% of cardiac-related deaths in new recruits undergoing training were associated with anomalous left CAs arising from the right coronary sinus (41). Specialists in the field are focused on developing imaging techniques for accurate screening and devising prophylactic strategies following diagnosis (42).

The condition in which coronary arteries connect to the pulmonary artery (ALCAPA) is a much more serious and life-threatening anomaly than instances of nonstandard stem placement on the aorta. ALCAPA has been calculated to occur in 1 in 300,000 births, and approximately 90% of patients with this condition do not survive past infancy without surgical repair (43, 44). Those who do survive develop extensive, disorganized compensatory vasculature around the heart and are at increased risk for severe coronary complications in adulthood. Interestingly, unlike many congenital cardiac malformations, ALCAPA is usually an isolated defect observed in the absence of other abnormal structural changes, though an underlying genetic cause for this condition has not yet been identified. Our data suggest that Vegfc or Isl1 mutations could be possible candidates. A better understanding of the developmental and potential genetic basis of this syndrome could lead to less invasive, alternative interventions for ALCAPA.

VEGF receptor signaling is an important regulator of coronary vessel development, but a site-specific role of individual family members has not been delineated for stem development. Cardiomyocytes within the main ventricle produce VEGF-A (33, 45), and coronary vascular development is decreased when VEGF-A is inhibited (13, 46, 47). Specific deletion of this factor from cardiac cells primarily affects the microvasculature (48) and coronary vessels within the myocardium (27). VEGF-A expression is downstream of a signaling cascade in which FGFs induce sonic hedgehog to regulate coronary development (49, 50). VEGF-A also has an important role in myocardial and valve development. With VEGF-A inhibition or deletion, the myocardial layer is thin, containing fewer cardiomyocytes (13, 27, 48). The developing valves require a complex temporal sequence of VEGF signaling, involving first VEGFR1 and then VEGFR2, to carry out the process of cushion formation and subsequent maturation (51–53). Thus, VEGF-A is multifunctional during heart development, possibly complicating interpretations of its role in coronary vascularization, which is thought to be dependent on hypoxic signals triggered within the expanding myocardium (54).

Our data indicate that VEGF-C has a more specific function during cardiac development, even within the coronary vasculature. Myocardial thickness and valve structure are not perturbed in knockout hearts, which is consistent with the more restricted expression pattern of VEGF-C in the outflow tract vessels and epicardium (26). In another study, we showed that VEGF-C deletion specifically affects sinus venosus–derived coronary development but not coronary vascularization from other sources on the ventral side of the heart and septum (likely from the endocardium, refs. 25, 27, or proepicardium, ref. 28) (26). This specific effect is unsurprising, given that sinus venosus sprouting is initiated directly beneath the epicardium (which expresses VEGF-C), while ventral sprouting is first seen deeper within the myocardium (which expresses VEGF-A) (25–27). VEGF-C expression in the outflow tract is required for the first step in CA stem development. Few peritruncal vessels develop in the absence of VEGF-C, precluding their connections with the aorta. In knockout hearts, endothelial projections that are connected to endocardium can be seen sprouting toward the outflow tract. We hypothesize that these vessels give rise to the delayed, abnormally low stems that were seen in some of the mutants. Indeed, Nfatc1-CreER lineage tracing experiments show that endocardial-derived coronary vessels contribute to CA stems. The formation of these lowered stems could involve endocardial activation in response to VEGF-A expressed by ventricular cardiomyocytes. In summary, VEGF-C is important for CA stem patterning, because, in the absence of peritruncal vessels, stems form in abnormal locations.

Our results identified cardiomyocytes as a cellular feature specific to the aorta. Cardiomyocytes are known to foster vascular growth by secreting angiogenic molecules (23, 24, 55), supporting the hypothesis that they facilitate the establishment of stems at the correct location. However, preliminary experiments deleting Vegfa using various cardiac Cre lines did not suggest a role for this growth factor. Future experiments will focus on identifying the specific cardiomyocyte-derived molecule mediating stem positioning. An additional possibility is that these cells are antagonizing an antiangiogenic factor present in the walls of the outflow tract. Indeed, large arteries are frequently surrounded by an avascular zone, as is the case with the developing aorta and pulmonary artery. In our experiments, endothelial cells did not migrate into the region surrounding outflow tract explants (unless cardiomyocytes were present), suggesting the existence of antiangiogenic factors. Precedent for such a mechanism is seen with the extracellular matrix protein chondromodulin-1, which restricts vessel growth in the cardiac valves during development and in the adult (56). Similarly, the vessel-free zone around the outflow arteries expresses numerous extracellular matrix molecules, which, in other systems, have been shown to liberate antiangiogenic peptides following proteolytic processing (57).

The mechanisms by which cardiomyocytes specifically populate the aortic wall are actively under investigation. Aortic cardiomyocyte development is disrupted by heterozygosity for the transcription factor Isl1. This transcription factor marks the second heart field, which contributes to heart growth, by adding cells onto the early heart tube to produce the outflow tract, right ventricle, and atria (58). Functionally, Isl1 supports cardiomyocyte progenitor proliferation, causing Isl1-null animals to lack the majority of second heart field–derived cardiac structures (36). In fact, we found that aortic cardiomyocytes derive from the second heart field. These observations suggest that aortic cardiomyocytes may develop via mechanisms analogous to those of second heart field maturation but in an aorta-specific manner. There may be an aorta-specific delay in the differentiation of second heart field progenitors in response to molecules specifically expressed in this outflow vessel. There are genetic distinctions between the aorta and pulmonary artery. The base of the pulmonary artery specifically expresses a Myf5 promoter fragment (96-16 transgene) (59) and Sema3c (18), while the corresponding region on the aorta expresses an Fgf10 fragment (T55 transgene) (18). Ultimately, insight into the embryonic mechanisms that trigger CA stem development on the aorta could be commandeered to establish new vessels in the adult as treatments for congenital cardiac defects and coronary heart disease.

View Supplemental data

We thank Sylvia Evans (Isl1), Lucile Miquerol (Vegfa-lacZ), Annette Damert (Vegfa-lacZ), Andras Nagy (Vegfa-lacZ), Brian Black (Mef2c-AHF-Cre), and Louisa Iruela-Arispe (VE-cadherin-CreER) for mouse strains. H.I. Chen was partially supported by a summer research grant administered by the Stanford University Vice Provost for Undergraduate Education. B. Zhou is supported by the Ministry of Science and Technology (2013CB945302 and 2012CB945102) and the National Natural Science Foundation of China (91339104, 31271552, 31222038). J.C. Wu is supported by the NIH (R01 HL093172) and the American Heart Association Established Investigator Award. K. Red-Horse is supported by the NIH (4R00HL10579303) and the Searle Foundation.

Address correspondence to: Kristy Red-Horse, Gilbert Hall, 371 Serra Mall, Stanford, California 94305, USA. Phone: 650.724.3135; E-mail: kredhors@stanford.edu.

Conflict of interest: Kari Alitalo is a consultant for Herantis Pharma Ltd. Joseph C. Wu is a consultant for the publically traded company Ivivi.

Reference information: J Clin Invest. 2014;124(11):4899–4914. doi:10.1172/JCI77483.

1. Zeina AR, Blinder J, Sharif D, Rosenschein U, Barmeir E. Congenital coronary artery anomalies in adults: non-invasive assessment with multidetector CT. Br J Radiol. 2009;82(975):254–261.
   View this article via: PubMed CrossRef Google Scholar
2. Angelini P. Coronary artery anomalies: an entity in search of an identity. Circulation. 2007;115(10):1296–1305.
   View this article via: PubMed Google Scholar
3. Wesselhoeft H, Fawcett JS, Johnson AL. Anomalous origin of the left coronary artery from the pulmonary trunk. Its clinical spectrum, pathology, and pathophysiology, based on a review of 140 cases with seven further cases. Circulation. 1968;38(2):403–425.
   View this article via: PubMed CrossRef Google Scholar
4. Ginde S, Earing MG, Bartz PJ, Cava JR, Tweddell JS. Late complications after Takeuchi repair of anomalous left coronary artery from the pulmonary artery: case series and review of literature. Pediatr Cardiol. 2012;33(7):1115–1123.
   View this article via: PubMed CrossRef Google Scholar
5. Olivey HE, Svensson EC. Epicardial-myocardial signaling directing coronary vasculogenesis. Circ Res. 2010;106(5):818–832.
   View this article via: PubMed CrossRef Google Scholar
6. Bogers AJ, Gittenberger-De Groot AC, Poelmann RE, Péault BM, Huysmans HA. Development of the origin of the coronary arteries, a matter of ingrowth or outgrowth? Anat Embryol. 1989;180(5):437–441.
   View this article via: PubMed CrossRef Google Scholar
7. Waldo KL, Willner W, Kirby ML. Origin of the proximal coronary artery stems and a review of ventricular vascularization in the chick embryo. Am J Anat. 1990;188(2):109–120.
   View this article via: PubMed CrossRef Google Scholar
8. Ando K, Nakajima Y, Yamagishi T, Yamamoto S, Nakamura H. Development of proximal coronary arteries in quail embryonic heart: multiple capillaries penetrating the aortic sinus fuse to form main coronary trunk. Circ Res. 2004;94(3):346–352.
   View this article via: PubMed CrossRef Google Scholar
9. Riley PR, Smart N. Vascularizing the heart. Cardiovasc Res. 2011;91(2):260–268.
   View this article via: PubMed CrossRef Google Scholar
10. Hutchins GM, Kessler-Hanna A, Moore GW. Development of the coronary arteries in the embryonic human heart. Circulation. 1988;77(6):1250–1257.
    View this article via: PubMed CrossRef Google Scholar
11. Pérez-Pomares J-M, et al. Origin of coronary endothelial cells from epicardial mesothelium in avian embryos. Int J Dev Biol. 2002;46(8):1005–1013.
    View this article via: PubMed Google Scholar
12. Tian X, et al. Peritruncal coronary endothelial cells contribute to proximal coronary artery stems and their aortic orifices in the mouse heart. PLoS One. 2013;8(11):e80857.
    View this article via: PubMed CrossRef Google Scholar
13. Tomanek RJ, et al. VEGF family members regulate myocardial tubulogenesis and coronary artery formation in the embryo. Circ Res. 2006;98(7):947–953.
    View this article via: PubMed CrossRef Google Scholar
14. Tomanek RJ, Hansen HK, Christensen LP. Temporally expressed PDGF and FGF-2 regulate embryonic coronary artery formation and growth. Arterioscler Thromb Vasc Biol. 2008;28(7):1237–1243.
    View this article via: PubMed CrossRef Google Scholar
15. Dyer L, Wu Y, Moser M, Patterson C. BMPER-induced BMP signaling promotes coronary artery remodeling. Dev Biol. 2014;386(2):385–394.
    View this article via: PubMed CrossRef Google Scholar
16. Waldo KL, Kumiski DH, Kirby ML. Association of the cardiac neural crest with development of the coronary arteries in the chick embryo. Anat Rec. 1994;239(3):315–331.
    View this article via: PubMed CrossRef Google Scholar
17. Ward C, Stadt H, Hutson M, Kirby ML. Ablation of the secondary heart field leads to tetralogy of Fallot and pulmonary atresia. Dev Biol. 2005;284(1):72–83.
    View this article via: PubMed CrossRef Google Scholar
18. Théveniau-Ruissy M, et al. The del22q11.2 candidate gene Tbx1 controls regional outflow tract identity and coronary artery patterning. Circ Res. 2008;103(2):142–148.
    View this article via: PubMed CrossRef Google Scholar
19. Li Wei , et al. An essential role for connexin43 gap junctions in mouse coronary artery development. Development. 2002;129(8):2031–2042.
    View this article via: PubMed Google Scholar
20. González-Iriarte M, et al. Development of the coronary arteries in a murine model of transposition of great arteries. J Mol Cell Cardiol. 2003;35(7):795–802.
    View this article via: PubMed CrossRef Google Scholar
21. Ratajska A, ZBotorowicz R, BBazejczyk M, Wasiutyñski A. Coronary artery embryogenesis in cardiac defects induced by retinoic acid in mice. Birth Defects Res Part A Clin Mol Teratol. 2005;73(12):966–979.
    View this article via: PubMed CrossRef Google Scholar
22. Chiu IS, et al. Evolution of coronary artery pattern according to short-axis aortopulmonary rotation: a new categorization for complete transposition of the great arteries. J Am Coll Cardiol. 1995;26(1):250–258.
    View this article via: PubMed CrossRef Google Scholar
23. Zhou B, et al. Fog2 is critical for cardiac function and maintenance of coronary vasculature in the adult mouse heart. J Clin Invest. 2009;119(6):1462–1476.
    View this article via: JCI PubMed CrossRef Google Scholar
24. Heineke J, et al. Cardiomyocyte GATA4 functions as a stress-responsive regulator of angiogenesis in the murine heart. J Clin Invest. 2007;117(11):3198–3210.
    View this article via: JCI PubMed CrossRef Google Scholar
25. Red-Horse K, Ueno H, Weissman IL, Krasnow MA. Coronary arteries form by developmental reprogramming of venous cells. Nature. 2010;464(7288):549–553.
    View this article via: PubMed CrossRef Google Scholar
26. Chen HI, et al. The sinus venosus contributes to the coronary vasculature through VEGF-C stimulated angiogenesis. Development. In press.
27. Wu B, et al. Endocardial cells form the coronary arteries by angiogenesis through myocardial-endocardial VEGF signaling. Cell. 2012;151(5):1083–1096.
    View this article via: PubMed CrossRef Google Scholar
28. Katz TC, et al. Distinct compartments of the proepicardial organ give rise to coronary vascular endothelial cells. Dev Cell. 2012;22(3):639–650.
    View this article via: PubMed CrossRef Google Scholar
29. Kivelä R, et al. VEGF-B-induced vascular growth leads to metabolic reprogramming and ischemia resistance in the heart. EMBO Mol Med. 2014;6(3):307–321.
    View this article via: PubMed Google Scholar
30. Kärpanen T, et al. Overexpression of vascular endothelial growth factor-B in mouse heart alters cardiac lipid metabolism and induces myocardial hypertrophy. Circ Res. 2008;103(9):1018–1026.
    View this article via: PubMed CrossRef Google Scholar
31. Bellomo D, et al. Mice lacking the vascular endothelial growth factor-B gene (Vegfb) have smaller hearts, dysfunctional coronary vasculature, and impaired recovery from cardiac ischemia. Circ Res. 2000;86(2):E29–E35.
    View this article via: PubMed CrossRef Google Scholar
32. Li X, et al. Reevaluation of the role of VEGF-B suggests a restricted role in the revascularization of the ischemic myocardium. Arterioscler Thromb Vasc Biol. 2008;28(9):1614–1620.
    View this article via: PubMed CrossRef Google Scholar
33. Miquerol L, Gertsenstein M, Harpal K, Rossant J, Nagy A. Multiple developmental roles of VEGF suggested by a LacZ-tagged allele. Dev Biol. 1999;212(2):307–322.
    View this article via: PubMed CrossRef Google Scholar
34. Watanabe Y, Buckingham M. The formation of the embryonic mouse heart. Ann N Y Acad Sci. 2010;1188(1):15–24.
    View this article via: PubMed CrossRef Google Scholar
35. Laugwitz K-L, Moretti A, Caron L, Nakano A, Chien KR. Islet1 cardiovascular progenitors: a single source for heart lineages? Development. 2008;135(2):193–205.
    View this article via: PubMed Google Scholar
36. Cai C-L, et al. Isl1 identifies a cardiac progenitor population that proliferates prior to differentiation and contributes a majority of cells to the heart. Dev Cell. 2003;5(6):877–889.
    View this article via: PubMed CrossRef Google Scholar
37. Verzi MP, McCulley DJ, De Val S, Dodou E, Black BL. The right ventricle, outflow tract, and ventricular septum comprise a restricted expression domain within the secondary/anterior heart field. Dev Biol. 2005;287(1):134–145.
    View this article via: PubMed CrossRef Google Scholar
38. Engleka KA, et al. Insertion of Cre into the Pax3 locus creates a new allele of Splotch and identifies unexpected Pax3 derivatives. Dev Biol. 2005;280(2):396–406.
    View this article via: PubMed CrossRef Google Scholar
39. Yamanaka O, Hobbs RE. Coronary artery anomalies in 126,595 patients undergoing coronary arteriography. Cathet Cardiovasc Diagn. 1990;21(1):28–40.
    View this article via: PubMed CrossRef Google Scholar
40. Maron BJ, et al. Recommendations and considerations related to preparticipation screening for cardiovascular abnormalities in competitive athletes: 2007 update: a scientific statement from the American Heart Association Council on Nutrition, Physical Activity, and Metabolism: endorsed by the American College of Cardiology Foundation. Circulation. 2007;115(12):1643–1655.
    View this article via: PubMed CrossRef Google Scholar
41. Eckart RE, et al. Sudden death in young adults: a 25-year review of autopsies in military recruits. Ann Intern Med. 2004;141(11):829–834.
    View this article via: PubMed CrossRef Google Scholar
42. Angelini P. Novel imaging of coronary artery anomalies to assess their prevalence, the causes of clinical symptoms, and the risk of sudden cardiac death. Circ Cardiovasc Imaging. 2014;7(4):747–754.
    View this article via: PubMed CrossRef Google Scholar
43. Yau JM, Singh R, Halpern EJ, Fischman D. Anomalous origin of the left coronary artery from the pulmonary artery in adults: a comprehensive review of 151 adult cases and a new diagnosis in a 53-year-old woman. Clin Cardiol. 2011;34(4):204–210.
    View this article via: PubMed CrossRef Google Scholar
44. Frescura C, et al. Anomalous origin of coronary arteries and risk of sudden death: a study based on an autopsy population of congenital heart disease. Hum Pathol. 1998;29(7):689–695.
    View this article via: PubMed CrossRef Google Scholar
45. Tomanek RJ, Ratajska A, Kitten GT, Yue X, Sandra A. Vascular endothelial growth factor expression coincides with coronary vasculogenesis and angiogenesis. Dev Dyn. 1999;215(1):54–61.
    View this article via: PubMed Google Scholar
46. Gerber HP, et al. VEGF is required for growth and survival in neonatal mice. Development. 1999;126(6):1149–1159.
    View this article via: PubMed Google Scholar
47. Liu H, et al. Role of VEGF and tissue hypoxia in patterning of neural and vascular cells recruited to the embryonic heart. Dev Dyn. 2009;238(11):2760–2769.
    View this article via: PubMed CrossRef Google Scholar
48. Giordano FJ, et al. A cardiac myocyte vascular endothelial growth factor paracrine pathway is required to maintain cardiac function. Proc Natl Acad Sci U S A. 2001;98(10):5780–5785.
    View this article via: PubMed CrossRef Google Scholar
49. Lavine KJ, Long F, Choi K, Smith C, Ornitz DM. Hedgehog signaling to distinct cell types differentially regulates coronary artery and vein development. Development. 2008;135(18):3161–3171.
    View this article via: PubMed CrossRef Google Scholar
50. Lavine KJ, et al. Fibroblast growth factor signals regulate a wave of Hedgehog activation that is essential for coronary vascular development. Genes Dev. 2006;20(12):1651–1666.
    View this article via: PubMed CrossRef Google Scholar
51. Dor Y, et al. A novel role for VEGF in endocardial cushion formation and its potential contribution to congenital heart defects. Development. 2001;128(9):1531–1538.
    View this article via: PubMed Google Scholar
52. Chang C-P, et al. A field of myocardial-endocardial NFAT signaling underlies heart valve morphogenesis. Cell. 2004;118(5):649–663.
    View this article via: PubMed CrossRef Google Scholar
53. Stankunas K, Ma GK, Kuhnert FJ, Kuo CJ, Chang C-P. VEGF signaling has distinct spatiotemporal roles during heart valve development. Dev Biol. 2010;347(2):325–336.
    View this article via: PubMed CrossRef Google Scholar
54. Wikenheiser J, Doughman Y-Q, Fisher SA, Watanabe M. Differential levels of tissue hypoxia in the developing chicken heart. Dev Dyn. 2006;235(1):115–123.
    View this article via: PubMed CrossRef Google Scholar
55. Patten IS, et al. Cardiac angiogenic imbalance leads to peripartum cardiomyopathy. Nature. 2012;485(7398):333–338.
    View this article via: PubMed CrossRef Google Scholar
56. Yoshioka M, et al. Chondromodulin-I maintains cardiac valvular function by preventing angiogenesis. Nat Med. 2006;12(10):1151–1159.
    View this article via: PubMed CrossRef Google Scholar
57. Mundel TM, Kalluri R. Type IV collagen-derived angiogenesis inhibitors. Microvasc Res. 2007;74(2–3):85–89.
    View this article via: PubMed Google Scholar
58. Kelly RG. The second heart field. Curr Top Dev Biol. 2012;100:33–65.
    View this article via: PubMed Google Scholar
59. Bajolle F, et al. Rotation of the myocardial wall of the outflow tract is implicated in the normal positioning of the great arteries. Circ Res. 2006;98(3):421–428.
    View this article via: PubMed CrossRef Google Scholar
60. Karkkainen MJ, et al. Vascular endothelial growth factor C is required for sprouting of the first lymphatic vessels from embryonic veins. Nat Immunol. 2004;5(1):74–80.
    View this article via: PubMed CrossRef Google Scholar
61. Sun Y, et al. Islet 1 is expressed in distinct cardiovascular lineages, including pacemaker and coronary vascular cells. Dev Biol. 2007;304(1):286–296.
    View this article via: PubMed CrossRef Google Scholar
62. Monvoisin A, Alva JA, Hofmann JJ, Zovein AC, Lane TF, Iruela-Arispe ML. VE-cadherin-CreERT2 transgenic mouse: a model for inducible recombination in the endothelium. Dev Dyn. 2006;235(12):3413–3422.
    View this article via: PubMed CrossRef Google Scholar
63. Tian X, et al. Vessel formation. De novo formation of a distinct coronary vascular population in neonatal heart. Science. 2014;345(6192):90–94.
    View this article via: PubMed CrossRef Google Scholar
64. Sheikh AY, et al. In vivo genetic profiling and cellular localization of apelin reveals a hypoxia-sensitive, endothelial-centered pathway activated in ischemic heart failure. Am J Physiol Heart Circ Physiol. 2008;294(1):H88–H98.
    View this article via: PubMed Google Scholar
65. Sharan SK, Thomason LC, Kuznetsov SG, Court DL. Recombineering: a homologous recombination-based method of genetic engineering. Nat Protoc. 2009;4(2):206–223.
    View this article via: PubMed CrossRef Google Scholar
66. Warming S, Costantino N, Court DL, Jenkins NA, Copeland NG. Simple and highly efficient BAC recombineering using galK selection. Nucleic Acids Res. 2005;33(4):e36.
    View this article via: PubMed CrossRef Google Scholar