The case for better models is straightforward. Mouse and rat hearts, the traditional workhorses of cardiovascular research, differ from human hearts in fundamental ways. They beat far faster, burn energy at a higher basal rate, and rely predominantly on a different myosin heavy chain isoform than the adult human ventricle, which is enriched in the beta form. These species differences help explain why promising preclinical results so often collapse in human trials, producing either unexpected cardiotoxicity or disappointing therapeutic benefit. Two-dimensional cultures, meanwhile, are cheap and fast but strip away the three-dimensional architecture, multicellular crosstalk and maturation cues that define living heart tissue. Cardiac MPS emerged to fill this void, and the review traces their remarkable evolution from the first beating embryoid bodies generated from stem cells in 1993, through engineered heart tissue in 2004, to the landmark 2021 creation of scaffold-free, self-organizing cardiac organoids called cardioids that recapitulate key features of early human heart development.

The review organizes cardiac MPS into two construction paradigms. The first is developmentally guided self-organization, in which human pluripotent stem cells are coaxed through timed bursts and pauses of WNT signaling, together with factors such as BMP4, activin A and FGF2, to form chamber-like organoids containing cardiomyocytes, fibroblasts, endothelial cells, neural crest cells and more. Elaborate versions add vascular-promoting cues like VEGF, foregut patterning, or even blood-generating signals to produce organoids with vascular compartments and hematopoietic progenitors. The second paradigm is engineering-driven assembly, which imposes external control through biomaterial scaffolds, 3D bioprinting and microfluidic heart-on-a-chip platforms. Self-organizing organoids offer unmatched developmental fidelity, making them ideal for studying embryonic cardiogenesis and congenital defects, while engineered tissues deliver superior reproducibility, spatial control and compatibility with quantitative drug screening. The review stresses that neither approach is universally superior; the choice depends entirely on the scientific question.

Biomaterials sit at the heart of the engineering approach, and the review provides an unusually detailed tour of the options. Naturally derived materials such as collagen, fibrin, hyaluronic acid, gelatin and decellularized extracellular matrix offer rich biological signals but suffer from batch variability and weak mechanics. Synthetic polymers like PEG, PCL and PLGA provide precise control and manufacturability but must be functionalized with cell-adhesive motifs. Hybrid systems blend the two philosophies. Electroconductive hydrogels laced with carbon nanotubes or conductive polymers can boost electrical coupling and connexin 43 expression, improving contraction synchrony, though dosing must be tuned carefully to avoid cytotoxicity. Scaffold stiffness matters enormously: native myocardium sits near 10 to 30 kilopascals, while scarred, fibrotic tissue stiffens to 35 to 70 kilopascals or more, and abnormal stiffness can actively drive disease-like cell states through epigenetic remodeling rather than merely accompanying them. Topographic cues from electrospun and micropatterned scaffolds guide cardiomyocyte alignment and restore the anisotropic architecture of real heart muscle.

Bioprinting adds another layer of control, depositing cell-laden bioinks layer by layer to fabricate vascular-like cardiac constructs with defined geometry. The review distinguishes carefully between biomaterial scaffolds, which can be processed under cell-free conditions, and true bioinks, which must keep living cells viable during extrusion or photopolymerization. Extrusion printing demands shear-thinning materials that recover quickly after deposition, while light-based printing requires finely tuned cross-linking kinetics. The technology’s strengths include spatial patterning of endothelialized structures, reproducible geometry, multicellular integration and controllable anisotropy. But a persistent trade-off between printing resolution and cell survival remains unresolved, and constructing a full hierarchical vascular network, from large vessels down to capillaries, still eludes the field. Microfluidics complements these efforts by providing perfusion, controlled shear stress, real-time biosensing and, in multi-organ chip formats, the ability to couple heart tissue with liver or kidney modules to capture systemic drug responses.

Perhaps the review’s most important contribution is its rigorous framework for assessing maturation, the property that ultimately determines whether a heart model is physiologically meaningful. Structural benchmarks include sarcomere length approaching the adult range of roughly 2.2 micrometers, organized Z-discs and A-bands, polarized connexin 43 at cell junctions, chamber-like cavity formation, and organized populations of fibroblasts, endothelial cells, macrophages and sympathetic neurons. Molecularly, the developmental shift from the TNNI1 to the TNNI3 troponin isoform is tracked quantitatively, though even the most mature organoids reported so far reach only about 18 percent TNNI3, corresponding roughly to a twenty-week human fetal heart, a sobering reminder that in vitro maturation still falls short of adulthood.

Functional maturation is assessed through an equally demanding battery of readouts. Electrophysiological maturity shows up as a more hyperpolarized resting membrane potential, adult-like action potentials and conduction velocities around 25 centimeters per second, with properly responsive beta-adrenergic signaling. Calcium handling reveals whether cardiomyocytes have built the T-tubule and dyad architecture needed for efficient excitation-contraction coupling, a feature most current systems still lack. Contractile performance is quantified by force generation; one progressively stretched engineered heart tissue achieved a specific twitch force of 11.28 millinewtons per square millimeter, a 5.1-fold improvement over controls. Metabolic maturation is measured through oxygen consumption rates and the shift from fetal glycolysis toward fatty acid beta-oxidation, which can be encouraged by substrate switching with fatty acids and galactose, thyroid hormone, or chronic electrical stimulation. In one conditioned tissue, mitochondria occupied 30 percent of the analyzed area, matching adult human myocardium. No single metric suffices, the review insists; maturity is a convergent, multidimensional phenotype.

The applications already demonstrated are striking. Multi-chamber cardioids have disentangled the regulatory relationship between the transcription factors HAND1 and NKX2-5 during early chamber formation. Organoids with microRNA-187 overexpression have reproduced congenital heart defect phenotypes that could be rescued by NIPBL supplementation. Gene-edited organoids carrying C10orf71 frameshift variants replicated the contractile defects of dilated cardiomyopathy, and the cardiac myosin activator omecamtiv mecarbil later restored function in a matching mouse model. Zonally oxygenated engineered tissues have recreated the infarct border zone with region-specific metabolic, fibrotic and calcium abnormalities that uniform hypoxia models miss entirely. Vascularized organoids showed that the blood pressure drug captopril protects cryoinjured tissue, while doxorubicin cardiotoxicity and even nanoplastic-induced injury have been modeled in organoid-on-chip formats. On the regenerative frontier, engineered heart muscle allografts have achieved durable remuscularization in rhesus macaques, though post-transplant arrhythmias highlight that electrical integration, not just survival, must be solved.

The review closes with a clear-eyed assessment of the remaining barriers: incomplete vascularization beyond the diffusion limit, fetal-like maturation, line-to-line variability, Matrigel batch effects, missing systemic physiology and the absence of standardized benchmarking. Its prescription is pragmatic rather than maximalist. Automated bioreactors, biosensors, AI-assisted closed-loop cultivation and standardized reference datasets could make cardiac MPS reproducible enough for regulated drug testing, while regenerative applications await perfusable vasculature and safe graft-host coupling. The message to the field is that progress should be judged not by how elaborate these tissues look, but by whether they become reliable, interpretable and genuinely predictive instruments of human cardiac medicine.

Subject of Research: Cardiac microphysiological systems for modeling human cardiac development, disease and drug responses

Article Title: Cardiac microphysiological systems in cardiovascular research: Construction paradigms, maturation trajectories, and translational frontiers

Article References: Xie, R., Zhao, H., Lei, Y., & Wang, X. (2026). Cardiac microphysiological systems in cardiovascular research: Construction paradigms, maturation trajectories, and translational frontiers. Bioengineering & Translational Medicine, Article e70167. https://doi.org/10.1002/btm2.70167

Image Credits: AI Generated

DOI: 10.1002/btm2.70167

Keywords: cardiac organoids, microphysiological systems, engineered heart tissue, cardiomyocyte maturation, 3D bioprinting, heart-on-a-chip, cardiotoxicity, disease modeling, vascularization, regenerative medicine, biomaterials, drug screening