The human heart pumps between about 5 and 20 litres of blood through the body every minute1. To receive enough oxygen to fulfil this tremendous task, heart-muscle cells need their own blood supply. This is provided by specialized blood vessels, including coronary arteries. Defects in these arteries can lead to coronary heart disease and even heart attack2,3. Understanding how coronary arteries form during embryonic development is therefore of great interest, because such knowledge might help in developing strategies to prevent or treat coronary heart disease. In a paper in Nature, Su et al.4 provide a detailed picture of the sequence of events that leads to coronary artery development.
The cells that generate coronary arteries originate from various regions of the embryo, including a sac-like structure called the sinus venosus that adjoins the embryonic heart5,6. From these sites, the cells invade the heart’s muscle-cell layer. Here, they form an immature blood-vessel network called a plexus that is subsequently remodelled into functional arteries and veins.
Su and colleagues set out to investigate how cells from the sinus venosus develop into coronary arteries, using single-cell RNA sequencing (scRNA-seq) — a technique that enables precise identification of the genes being expressed in each cell of a tissue7. Gene-expression patterns change during tissue differentiation, for example as sinus venosus cells mature into coronary arteries. Comparison of the gene-expression patterns for individual cells of a given type can therefore reveal the cells’ relationships to one another.
The authors extracted single endothelial cells, which make up the inner lining of blood vessels, from the hearts of mouse embryos at a developmental time point just before coronary artery formation. They reasoned that, at this embryonic stage, they would obtain cells at the various stages leading to coronary artery maturation, including sinus venosus and plexus cells. They then used bioinformatics to investigate the lineage relationships between these cells.
It has been thought that the remodelling of the plexus into arteries and veins starts only after the plexus has connected to the aorta (the main heart artery), and therefore after the onset of blood flow5. But, unexpectedly, Su et al. found that several cells from their embryos, in which the plexus had not yet received blood, had a gene-expression profile associated with mature arteries. They called these cells pre-artery cells.
The authors used a genetic strategy to indelibly label the pre-artery cells with a marker protein, such that these cells and the lineages they give rise to could be tracked during embryonic development. This lineage tracing revealed that, although most pre-artery cells did go on to form coronary arteries, some were incorporated into capillaries, which connect coronary arteries with veins. Thus, it seems that, although certain endothelial cells are genetically predisposed to form arteries, they also have a degree of developmental plasticity (Fig. 1).
Next, Su and colleagues performed a detailed analysis of the gene-expression patterns of cells on the developmental spectrum from sinus venosus to pre-artery cells. Changes in gene expression towards more arterial-like profiles occurred only gradually along most of the spectrum. However, there was a sharp change as cells crossed a threshold to adopt a pre-artery state. The researchers showed that the greatest difference in expression in pre-artery cells compared with other cells in their analysis occurred in genes implicated in regulating the cell cycle. Furthermore, in mouse embryos, pre-artery cells proliferated less than did cells in the plexus. Thus, limiting cell divisions might be a prerequisite for coronary artery maturation.
Indeed, the authors found that overexpression of the transcription factor COUP-TF2 in mice inhibited pre-artery formation by upregulating cell-cycle genes. COUP-TF2 was previously thought to limit the growth of arteries by suppressing the Notch signalling pathway8. But Su et al. showed that activation of Notch signalling could not prevent the defects caused by COUP-TF2 overexpression in mouse embryos. By contrast, pharmacological inhibition of the cell cycle increased artery formation in an ex vivo experiment. Thus, COUP-TF2 has functions in artery development that are independent of Notch signalling. Together, Su and colleagues’ work provides exciting insights into coronary artery formation. It will be interesting to discover whether the findings apply to artery development in other settings.
It will also be valuable to delineate the signalling pathways that lead certain cells to adopt the gene-expression profile of pre-artery cells. The Notch signalling pathway, acting independently of COUP-TF2, is a prime candidate. Studies in several developmental settings9–11 have suggested that new blood-vessel sprouts initially emanate from veins and capillaries and only subsequently form arteries, with each activity inhibiting the other. Inhibiting Notch signalling can lead to excessive blood-vessel sprouting, while at the same time preventing artery formation12,13, and a similar effect is seen during coronary artery development14. It will therefore be interesting to investigate how Notch signalling affects the gene-expression profiles that lead to the formation of pre-artery cells.
Although it is becoming increasingly clear that artery differentiation is intimately linked to cell-cycle state, the underpinnings of this relationship need further investigation. A report last year showed that cell-cycle inhibition is important for proper arterial gene expression in cells in vitro15. In addition, the signalling pathways involving Notch and vascular endothelial growth factor, which are indispensable for the establishment of new blood-vessel networks, are both implicated in influencing endothelial-cell proliferation16. However, it is not known how cells interpret signalling inputs from these pathways to balance the demand for proliferation of cells in the blood-vessel network with the need to establish new arteries.
One must also bear in mind that cell-lineage trajectories obtained from scRNA-seq might not reflect true developmental relationships. For instance, cells that have similar gene-expression patterns are not necessarily derived from the same precursor population. New techniques that unite cell-lineage tracing with scRNA-seq17 will help to bridge this gap, and will surely provide further insights into coronary artery development.
Nature 559, 335-336 (2018)