The human heart is the first functional organ to form during embryonic development, beginning to beat around three weeks after fertilization. It starts as a simple tube assembled from scattered progenitor cells and, through a tightly orchestrated series of loops, folds, and molecular signals, transforms into a four-chambered pump complete with valves, coronary vessels, and an electrical wiring system. That transformation involves at least a dozen distinct cell populations talking to each other through chemical and mechanical cues, and when any step goes wrong, the result is congenital heart disease, the most common category of birth defect worldwide.1PubMed Central. Decoding congenital heart disease: a multi-omic framework for cardiac lineage and regulatory dysfunction What follows is the story of how a heart gets built, told roughly in the order development happens.
Where the Heart’s Cells Come From
The cells that eventually become the heart originate in a region of the embryo called the mesoderm, a middle tissue layer that forms during the second and third weeks of development. These progenitor cells are organized into two overlapping populations known as the first and second heart fields. The first heart field contributes mainly to the left ventricle and parts of the atria, while the second heart field adds cells to the right ventricle, the outflow tract, and parts of both atria. Early fate-mapping work in chick embryos showed that these progenitors form a continuous arc across the front of the embryo rather than discrete, separated patches, and that much of their displacement toward the midline is driven by overall tissue motion rather than individual cell migration.2PubMed Central. Dynamic Positional Fate Map of the Primary Heart-Forming Region
Getting progenitor cells to commit to a cardiac fate requires a cocktail of transcription factors. Among the most studied are Nkx2-5, Tbx5, Gata4, and Myocardin. When researchers introduced these factors into non-cardiac mouse cells, they found that the combination of Tbx5, Gata4, and Myocardin together was the most effective at switching on heart-muscle genes while keeping non-cardiac genes quiet.3PLoS One. Cardiac gene activation analysis in mammalian non-myoblasic cells by Nkx2-5, Tbx5, Gata4 and Myocd No single factor was enough on its own. Tbx5, for instance, activated Wnt-related signaling pathways while actually suppressing some cardiac programs, and Myocardin did roughly the opposite. The two-factor pair of Tbx5 plus Myocardin could induce some cardiac markers, but the three-factor combination produced the cleanest cardiac signature. This layered logic, where individual signals counterbalance each other and only the right combination unlocks the cardiac program, is a recurring theme throughout heart development.
Forming and Looping the Heart Tube
Once progenitor cells converge at the embryo’s midline, they fuse into a straight tube. This tube has a simple inside-outside structure: an inner layer of endocardial cells (the future lining of the heart chambers) surrounded by an outer layer of myocardial cells (the future muscle). Between them sits a jelly-like matrix called cardiac jelly. At this point the tube is symmetric, and the next step is one of the most visually dramatic events in all of embryology: the tube bends to the right and coils into a loop, establishing the left-right asymmetry that eventually places the ventricles below and the atria above.
The rightward bend of this loop is remarkably consistent, and researchers have been working out what drives its directionality. Studies in chick embryos showed that the heart muscle cells themselves have an intrinsic chirality, a built-in “handedness” at the cellular level. These cells exhibit a dominant clockwise rotation in culture, a rightward bias in their internal alignment, and an asymmetric distribution of structural proteins like N-cadherin even before the loop starts to form. Experimentally reversing this cellular handedness through a signaling pathway called protein kinase C reversed the direction of looping.4PubMed Central. Intrinsic cellular chirality regulates left-right symmetry breaking during cardiac looping
Alongside this intrinsic cell chirality, a molecular signal called Nodal plays an important coordinating role. In the mouse, Nodal is transiently active in precursors at both poles of the heart tube before looping begins. But here is the surprise: when researchers knocked out Nodal in those precursor cells, asymmetric bending still started on its own. Nodal was not required to initiate the loop. Instead, it acts as a bias signal, amplifying and coordinating opposed asymmetries at the two ends of the tube so the final shape is a consistent, robust helix rather than a randomly oriented kink.5Developmental Cell. Nodal Determines the Laterality and Shape of the Asymmetric Heart Loop through the Amplification of Opposite Pole Asymmetries In other words, the heart has a built-in random generator of asymmetry, and Nodal’s job is to make sure that randomness consistently tips the same way.
Building the Chambers
After looping, the heart tube balloons outward at specific regions that will become the future atria and ventricles. This ballooning, or “chamber formation,” involves rapid cardiomyocyte proliferation and a process called trabeculation, in which the inner wall of the ventricles develops muscular ridges and grooves that increase surface area and help the embryonic heart pump before the coronary circulation is established. Trabeculation also sets up the structural framework for the future conduction system and papillary muscles.
A signaling pathway called Notch is critical for this step. Notch1 activity is highest in the endocardial cells lining the regions where trabeculae will sprout. When Notch1 or its downstream partner RBPJk was knocked out in mouse embryos, trabeculation failed. Ventricular ridges did not form, and key downstream signals like EphrinB2, Neuregulin-1 (NRG1), and BMP10 were diminished. What made this finding particularly interesting was that Notch appeared to control two processes independently: cardiomyocyte proliferation, which could be rescued by adding BMP10, and cardiomyocyte differentiation, which could be rescued by adding NRG1.6PubMed Central. Notch signaling is essential for ventricular chamber development These two processes, growing more cells and maturing them into the right type, must be exquisitely balanced. Tilt the ratio either way and you get a wall that is too thin (leading to cardiac weakness) or too thick (obstructing blood flow).
Beyond molecular signals, mechanical forces actively shape the developing chambers. Blood flow through the looping heart tube creates shear stress on the endocardial lining and pressure loads on the myocardium. These forces feed back into gene expression programs, effectively making the developing heart a sensor of its own function.7PubMed Central. Mechanical regulation of cardiac development A heart that pumps weakly during development does not just underperform; it remodels differently, because the mechanical cues it generates are altered.
Dividing the Outflow Tract
The early heart tube has a single outflow vessel that must eventually split into the aorta and pulmonary artery, which route blood to the body and lungs, respectively. This division depends on a population of cells that does not originate in the heart at all: cardiac neural crest cells. These cells arise from the dorsal neural tube, a structure along the embryo’s back that also gives rise to parts of the face, skull, and peripheral nerves. They then migrate long distances into the developing heart.8PubMed Central. Cardiac Neural Crest
Once they arrive, neural crest cells condense along the inner surface of the single outflow tube and physically force it to split into two parallel tubes.9PubMed Central. Dullard-mediated Smad1/5/8 inhibition controls mouse cardiac neural crest cells condensation and outflow tract septation When this process goes wrong, the result is persistent truncus arteriosus, a serious congenital defect in which the aorta and pulmonary artery remain fused as a single vessel. Experimental work in mice showed that disrupting a chromatin-regulating gene called Ankrd11 specifically in neural crest cells produced exactly this defect, along with ventricle dilation and weakened contraction.10PubMed Central. The chromatin regulator Ankrd11 controls cardiac neural crest cell-mediated outflow tract remodeling and heart function The vulnerability here is clear: because neural crest cells have to travel from a distant origin and then execute a precise mechanical task, anything that slows their migration, reduces their number, or disrupts their condensation behavior can leave the outflow tract unseparated.
How Heart Valves Form
Valves ensure that blood flows in one direction through the heart’s chambers, and their formation is a multi-step process that begins surprisingly early. Around four weeks of human gestation, specialized endocardial cells in the atrioventricular canal and outflow tract region undergo a dramatic identity change called endocardial-to-mesenchymal transformation, or EndMT. These cells delaminate from the endocardial lining and invade the cardiac jelly, creating swellings called endocardial cushions that serve as the primitive valve precursors. The transcription factor NFATC1 plays a central role in regulating this transformation, with its expression detected in valve endocardial cells as early as four weeks and persisting through later stages of leaflet formation.11PubMed Central. Endocardial-to-mesenchymal transformation and mesenchymal cell colonization at the onset of human cardiac valve development
Once the cushions are populated with mesenchymal cells, the next phase is remodeling. The cushions elongate and thin out into recognizable leaflets. The extracellular matrix, the scaffolding material that gives each leaflet its physical properties, changes composition over time. Early on, the cushions are rich in glycosaminoglycans, gel-like molecules that allow flexibility. As development proceeds, stress-resistant fibrous proteins like collagen accumulate.12PubMed Central. Expression and Deposition of Fibrous Extracellular Matrix Proteins in Cardiac Valves during Chick Development In the atrioventricular valves specifically, this matrix patterning is not completed until about a week after birth in mice, with the atrial side of the leaflet retaining glycosaminoglycans and versican while the ventricular side develops densely packed collagen fibers. Nodular thickenings form at the leaflet closure points during the late embryonic and neonatal period and then remodel further.13PubMed. Atrioventricular valve development during late embryonic and postnatal stages involves condensation and extracellular matrix remodeling This means valve development is not finished at birth. The leaflets continue to mature and refine their mechanical properties during infancy.
The Epicardium and Coronary Vessels
The outer covering of the heart, the epicardium, has a surprisingly important developmental role that goes far beyond being a simple wrapping. It originates from a transient structure called the proepicardium, a cluster of cells that protrudes from near the venous pole of the embryonic heart into the pericardial cavity.14PubMed. Embryonic development of the proepicardium and coronary vessels These cells migrate over the surface of the heart tube and form the epicardial layer. But they do not stay on the surface. A subset of epicardial cells undergoes another round of epithelial-to-mesenchymal transformation and dives into the heart wall, where they differentiate into the smooth muscle cells and fibroblasts of the coronary arteries.15PubMed. Smooth muscle cells and fibroblasts of the coronary arteries derive from epithelial-mesenchymal transformation of the epicardium
This was originally demonstrated using quail-chick chimera experiments, in which proepicardial tissue from quail embryos was transplanted into chick embryo hearts. The quail cells could be tracked by their distinctive nuclear markers, and researchers confirmed that they populated the walls of the developing coronary vessels, contributing both smooth muscle and fibroblasts as well as the endothelium.16PubMed. Epicardium-derived cells contribute a novel population to the myocardial wall and the atrioventricular cushions The epicardium is now recognized as a progenitor reservoir for the heart’s vascular supply, and understanding how to reactivate epicardial programs in adult tissue is an active area of regenerative medicine research.
Dividing the Atria and Ventricles
The looped, ballooned heart still needs internal walls, or septa, to separate oxygenated blood from deoxygenated blood. Atrial septation is a two-step curtain process. A first partition (the septum primum) grows down from the roof of the atrium toward the endocardial cushions. Before it fully closes the gap, a second opening (the foramen secundum) perforates its upper portion. Then a second, thicker partition (the septum secundum) grows alongside the first, creating a flap-valve arrangement. The result is the foramen ovale, a one-way passage that allows blood to bypass the lungs in the fetus, since the lungs are not yet breathing air. At birth, when the lungs inflate and left atrial pressure rises, the flap is pushed shut. In about a quarter of adults, this closure is incomplete, leaving a patent foramen ovale, a small tunnel-like gap between the two atria that is anatomically distinct from a true atrial septal defect, which involves an actual hole within the septum itself.17PubMed Central. Anatomy of the atrial septum and interatrial communications
Ventricular septation proceeds from a different direction: a muscular ridge grows upward from the floor of the ventricles while the endocardial cushions and outflow tract ridges grow downward. The final closure of the interventricular septum requires contributions from multiple tissue sources converging at the same point, making it another vulnerable step. A small membranous gap where these components meet is the most common location for ventricular septal defects.
What Maternal Metabolism Can Disrupt
Because heart development depends on precise gene activation at specific times, environmental factors that alter the chromatin landscape of progenitor cells can have outsized effects. Maternal diabetes is one of the strongest known environmental risk factors for congenital heart defects. Research using both cell-culture and mouse models of high glucose exposure found widespread changes in chromatin accessibility, with roughly 49,000 regions of the genome showing altered openness compared to normal-glucose controls. One specific casualty was the Nos3 gene, which encodes an enzyme needed for nitric oxide production. High glucose caused the chromatin around Nos3 to close down, reducing its expression. At the same time, reactive oxygen species production increased, creating a double hit on the same signaling pathway.18The Journal of Clinical Investigation. Epigenetic mechanisms underlying maternal diabetes-associated risk of congenital heart disease Nitric oxide is involved in endocardial cushion formation and outflow tract remodeling, among other steps, so reducing it at the wrong moment can derail multiple aspects of heart development simultaneously.
Mapping Heart Development Cell by Cell
Much of what we know about cardiac embryology came from decades of painstaking work in animal models, from quail-chick chimeras to genetically modified mice. In recent years, single-cell RNA sequencing has opened a new window by allowing researchers to read the gene-expression profile of thousands of individual cells from developing hearts. One study profiled nearly 5,000 cells from human embryonic hearts spanning 5 to 25 weeks of gestation and identified four major cell clusters, including cardiomyocytes, fibroblast-like cells, endothelial cells, and valvar cells, each with distinct gene signatures that changed over developmental time.19Cell Reports. Cardiac Embryology: From Progenitor Cells to Valves
Other single-cell work has gone beyond gene expression to map chromatin accessibility, identifying eight major differentiation trajectories in human fetal hearts, each defined by a unique combination of active transcription factors.20PubMed Central. Integrative single-cell analysis of cardiogenesis identifies developmental trajectories and non-coding mutations in congenital heart disease This kind of data is valuable not only for understanding normal development but also for pinpointing where congenital heart defects originate, since mutations in the non-coding regulatory regions that control when and where genes turn on may be just as consequential as mutations in the genes themselves. Separate profiling across multiple gestational timepoints identified at least eleven major cell types in the developing human heart, including not just the expected cardiomyocytes and endothelial cells but also macrophages, pericytes, and neural cells, underscoring how many non-muscle populations are essential to building a working heart.21PubMed. Single-cell RNA sequencing reveals the gene expression profile and cellular communication in human fetal heart development
Growing Hearts in a Dish
One of the more striking developments in the field is the ability to grow miniature heart-like structures, called organoids, from human stem cells. These self-assembling organoids recapitulate many features of the real thing. One protocol produced organoids with myocardial tissue, endocardial-lined internal chambers, a network of endothelial cells, cardiac fibroblasts, and an outer epicardial layer, mirroring the layered architecture of an actual embryonic heart. Imaging revealed four to six interconnected chambers, and electron microscopy showed sarcomeres, mitochondria, gap junctions, and structures resembling T-tubules, all hallmarks of real cardiac muscle.22Nature Communications. Self-assembling human heart organoids for the modeling of modeling of cardiac development and congenital heart disease
A more recent approach went further, generating organoids that exhibit anterior-posterior patterning with distinct atrial and ventricular-like chamber regions, proepicardial organ formation, and responsiveness to retinoic acid gradients, mimicking the spatial organization seen in a post-heart-tube-stage embryo.23Nature Communications. A patterned human primitive heart organoid model generated by pluripotent stem cell self-organization These organoids are not replacement hearts, and they lack the full vascular supply and nervous connections of a real organ. But they are proving useful as platforms to study how congenital defects arise in human tissue rather than mouse tissue, and to test whether candidate drug therapies can rescue developmental errors in a controlled setting.
An Evolutionary Perspective on Heart Complexity
The four-chambered heart of mammals and birds is often treated as a pinnacle of circulatory design, but comparative work across vertebrates suggests it is better understood as an elaboration of an ancient and remarkably conserved plan. Core transcription factors like Tbx5 and Tbx20 drive heart formation from jawless lampreys to humans, while related factors like Tbx2 and Tbx3 suppress chamber identity at the boundaries between heart regions in essentially all vertebrates studied. Even the electrical behavior of the heart shows deep conservation: electrocardiograms from fish, amphibians, and mammals share similar waveform features despite the fish heart having only two chambers and operating at far lower heart rates. The implication is that the mammalian four-chambered heart did not appear from scratch but was assembled by modifying and partitioning structures that already existed in simpler form in cold-blooded ancestors. Septation, trabeculation, and the division of the outflow tract are developmental add-ons layered onto a tube-based architecture that has been around for hundreds of millions of years.