Integrative circulatory engineering for regenerative therapy: Dynamic perfusion design and metabolic maturation in cardiac tissue engineering Authors: Jun Homma, Hidekazu Sekine Published: 2026 ABSTRACT AbstractThe adult myocardium couples extreme metabolic demand with a finely tuned coronary microvasculature. Continuous, high-work output is sustained by dense capillary networks, short diffusion distances and dominant reliance on mitochondrial oxidative phosphorylation. By contrast, current cardiac tissue engineering platforms, including scaffold-free myocardial constructs, have achieved robust contraction and basic drug responsiveness but remain constrained by fetal-like cardiomyocyte metabolism and incomplete vascular integration. As pluripotent stem cell–derived cardiomyocytes are driven toward more adult-like phenotypes using metabolic maturation strategies—such as substrate switching from glycolysis to fatty acid oxidation—oxygen consumption rises sharply, exposing the limitations of diffusion-limited culture and underspecified vascular design.This Review proposes integrative circulatory engineering as a framework in which metabolic maturation, vascular architecture and perfusion are co-designed rather than optimized in isolation. Scaffold-free myocardial tissues are highlighted as a particularly suitable platform, enabling close cell–cell contact, self-organized microvascular networks and dynamic remodeling of extracellular matrix. We examine how cell type composition, paracrine crosstalk, matrix mechanics and spatial patterning can be orchestrated to align metabolic demand with vascular supply.Perfusion bioreactors are treated as active components of the engineered circulation, providing controlled oxygen delivery, shear conditioning of endothelial networks, dynamic substrate provision and tunable mechanical loading. The concept is extended to in vitro circulatory units that couple myocardial modules to vascular beds and, in advanced implementations, to other metabolic organs. Finally, translational scenarios, disease modelling opportunities, quality control and computational design are discussed. Together, these elements outline a path toward myocardial tissues that approximate the structural, metabolic and functional complexity required for regenerative therapy and high-fidelity disease modeling.Keywords: Cardiac tissue engineering, Scaffold-free myocardium, Metabolic maturation, Vascular design, Perfusion bioreactor, Integrative circulatory engineering Abbreviations three-dimensional adenosine diphosphate adenosine triphosphate extracellular matrix fatty acid oxidation human pluripotent stem cell induced pluripotent stem cell nitric oxide optical coherence tomography poly(N-isopropylacrylamide) vascular endothelial growth factor. Integrating metabolic maturation and vascular design The adult human heart is simultaneously one of the most metabolically demanding and most vascularly constrained organs. At rest, cardiomyocytes rely predominantly on mitochondrial oxidative phosphorylation, with a major contribution from fatty acid β-oxidation, and must maintain continuous force generation over billions of beats [1]. This demand is supported by a dense, highly organized coronary microvasculature, with capillaries spaced within tens of micrometres of every myocyte and finely regulated flow, oxygen delivery and substrate supply [2,3]. Any attempt to engineer functional myocardial tissue for regenerative therapy must therefore confront both metabolic specialization and vascular integration from the outset [3]. Traditionally, cardiac tissue engineering has treated these challenges as largely separate—cell survival and contractile function on one side, vascularization and perfusion on the other. This separation was tolerable in thin, diffusion-supported tissues containing relatively glycolytic cells. As constructs thicken and pluripotent stem cell–derived cardiomyocytes are driven toward more adult-like phenotypes, it becomes increasingly important to consider oxygen demand, substrate utilization, and mitochondrial remodeling together with capillary-scale transport distances, endothelial phenotype, and perfusion dynamics [2,[4], [5], [6]]. This realization motivates a shift toward a more integrative framework in which metabolic maturation and vascular design are coordinated from the earliest stages of construct development. Limitations of current myocardial tissue constructs Over the past two decades, a diverse array of constructs has been developed, including hydrogel-based patches, decellularized matrix repopulation, microfabricated scaffolds, organoids and scaffold-free cell aggregates or sheets [7,8]. Many systems achieve synchronous contraction, reproducible calcium transients and basic drug responses, and some have entered preclinical evaluation as epicardial patches or injectable matrices. Yet several recurrent limitations distinguish them from native adult myocardium. First, cardiomyocytes commonly retain a fetal-like phenotype, with disorganized sarcomeres, small cell size, immature electrophysiology and a metabolic profile dominated by glycolysis and lactate utilization [9]. Immature fetal or PSC-derived cardiomyocytes are relatively small and rounded or spindle-like, whereas mature adult ventricular cardiomyocytes are rod-shaped and typically reach approximately 10–25 μm in diameter and about 100 μm in length [10,11]. As cardiomyocytes enlarge during fetal and postnatal maturation, diffusion distances increase accordingly [10]. In parallel, cardiac vascularization progresses from an initially diffusion-supported myocardium to an expanding and remodeling coronary vascular plexus that supports the growing compact wall [12]. Thus, from fetal to adult stages, the coronary vasculature expands substantially to support the growing myocardium, although cardiomyocyte hypertrophy can outpace increases in capillary density during postnatal growth [13,14]. Structural and electrophysiological aspects improve with prolonged culture, electrical pacing, mechanical load, hormonal cues, and also with appropriately designed two-dimensional culture environments, including extracellular matrix composition, physiological substrate stiffness, and cell shape or geometry [4,[15], [16], [17], [18], [19], [20]]; however, metabolic remodeling often lags behind. Second, vascularization remains suboptimal in most thick constructs. Diffusion alone limits viable thickness to a few hundred micrometres; beyond this, hypoxia and necrosis occur [6,21,22], as schematically illustrated for stacked cardiac cell sheets that exceed four layers (Fig. 1). Co-culture with endothelial cells, pre-vascularized scaffolds, microfluidic channels and inosculation with host vessels improve perfusion but rarely recapitulate the capillary density, hierarchy and flow distribution of native myocardium. Poorly perfused regions can compromise both survival and function. Diffusion-limited oxygen supply and necrotic core formation in stacked cardiac cell sheets. Schematic illustration of oxygen and nutrient diffusion in multilayered cardiac cell sheets. When more than four sheets are stacked, only the surface layers form an oxygen-rich zone, whereas deeper layers become hypoxic, leading to the development of a central necrotic region. Third, metabolic and vascular limitations reinforce each other. Metabolically immature, glycolytic cardiomyocytes can survive in poorly vascularized constructs because their demands are low. Conversely, pushing cells toward adult-like oxidative metabolism without adequate vascular support exacerbates hypoxia, cell loss and arrhythmia risk [5,6,23]. Many platforms therefore settle into a semi-fetal equilibrium compatible with limited vascular supply but insufficient for adult-like performance. Hydrogel-based engineered heart tissues (EHTs) offer controllable geometry and tunable mechanics but often rely on bulk diffusion unless explicitly perfused; bioprinting enables architectural predefinition but can be limited by cell viability during fabrication and by the maturation gap of printed constructs; decellularized matrices provide native ECM cues but face challenges in recellularization uniformity and perfusion access. These trade-offs motivate scaffold-free strategies combined with designed perfusion as a complementary route toward jointly achieving thickness, maturation and vascular integration. Metabolic maturation as a design driver Metabolic state is now recognized as an actively programmable trait rather than a passive readout of maturation. Manipulation of substrate availability, nutrient composition, endocrine cues and culture duration can drive pluripotent stem cell–derived cardiomyocytes from glycolytic, glucose- and lactate-preferring phenotypes toward increased mitochondrial content, oxidative phosphorylation and fatty acid β-oxidation, as demonstrated by metabolic selection and maturation strategies developed by Tohyama and colleagues, who exploited distinct metabolic dependencies of pluripotent stem cells and cardiomyocytes for lactate-based purification and subsequent oxidative maturation [9,24,25]. In these Tohyama-style protocols, lactate-based selection, glucose restriction, supplementation with long-chain fatty acids, and combinations with thyroid hormone, glucocorticoids, and insulin have shown that metabolic maturation can be accelerated and partially standardized. Similar improvements in maturation have also been reported across independent platforms using defined metabolic media and/or multimodal maturation cues (e.g., electromechanical training and advanced culture systems), supporting the generality of this metabolic-design perspective [4,26,27]. These approaches yield cardiomyocytes with improved contractile force, action potential morphology and calcium handling, and transcriptomic and proteomic profiles closer to adult myocardium. They also increase oxygen consumption and alter redox homeostasis, making mitochondrial-rich, fatty acid oxidation (FAO)–competent cardiomyocytes more sensitive to oxygen and substrate deprivation. Metabolic maturation therefore tightens the constraints on oxygen delivery and waste removal within engineered tissues and transforms a cellular optimization problem into a tissue- and system-level design constraint. Vascular design in metabolically demanding tissues Vascular design encompasses the creation of perfusable networks that are structurally, functionally and adaptively compatible with myocardial demands. This includes capillary-level structures with appropriate spacing and connectivity, feeding arterioles and draining venules, suitable hydraulic resistances and physiologically relevant shear stress patterns [[28], [29], [30], [31]]. Simple parallel channels or sparse capillary-like structures may suffice in thin or immature tissues, but as thickness and metabolic demand rise, diffusion distances shrink and perfused vessel density must approach adult values. A useful framework for considering endothelial maturation is that endothelial cells do not converge on a single generic “mature” state, but instead acquire organ-specific phenotypes [32,33]. Across organs, capillary endothelial cells range from tight barrier-forming phenotypes in the brain to highly fenestrated filtration-specialized phenotypes in the renal glomerulus [32,33]. In the heart, by contrast, the microvascular endothelium is predominantly continuous and non-fenestrated, and is specialized for rapid but regulated exchange in a tissue with extreme oxidative demand and continuous mechanical activity [32,34]. Accordingly, cardiac endothelial maturation should be understood not as maximization of impermeability or filtration, but as acquisition of a phenotype that combines junctional integrity, shear-responsive homeostasis, and efficient support of adjacent cardiomyocytes [32,34]. Cardiac endothelial cells actively regulate nitric oxide availability, angiocrine and paracrine signaling, and transendothelial transport of metabolites and nutrients, thereby influencing cardiomyocyte survival, contractile performance, and metabolic state [34,35]. For engineered myocardial tissues, this distinction is important: endothelial maturation should not be judged solely by the appearance of vessel-like networks or perfusable lumens, but also by whether endothelial cells acquire cardiac-supportive functional traits such as stable barrier properties, physiological flow responsiveness, and endothelial–cardiomyocyte metabolic coupling [32,34]. From this perspective, perfusion conditioning is not only a transport strategy but also an instructive cue for establishing a more coronary-like endothelial phenotype [32,34]. Within this organ-specific framework, coronary endothelial phenotype becomes a key determinant of vascular performance. In coronary microvessels, the endothelium regulates vascular tone, barrier function and substrate transport, and it dynamically responds to shear stress via nitric oxide and other mediators [3]. Endothelial metabolism is largely glycolytic, supporting proliferation and barrier integrity while conserving oxygen for cardiomyocytes. In engineered tissues, endothelial cells must be conditioned with appropriate shear forces and microenvironmental cues to adopt similarly specialized roles. Unstable, leaky or inflammatory vessels may provide inadequate and heterogeneous perfusion despite appearing structurally intact. Thus, vascular design cannot be reduced to “more vessels are better.” Instead, vascular topology, perfusion dynamics, endothelial phenotype and cardiomyocyte metabolism must be orchestrated to recreate the functional coupling of native myocardium. Scaffold-free myocardial tissues as a platform for integration Scaffold-free myocardial tissue engineering—via cell-sheet stacking, self-assembled spheroids and bio-printed aggregates—uses cells and their own extracellular matrix as primary structural elements [7,8,36,37]. This approach can facilitate closer cell–cell and cell–matrix interactions, improve electrical coupling, support spontaneous alignment and allow natural remodeling of microarchitecture over time. In typical cell-sheet platforms, temperature-responsive culture surfaces enable enzyme-free detachment of confluent monolayers when the temperature is lowered, thereby preserving native cell–cell junctions and endogenously deposited extracellular matrix (Fig. 2). This preservation is not merely technical but foundational for scaffold-free myocardial engineering: intact intercellular coupling and matrix continuity enable hierarchical stacking without artificial scaffolds, maintain electrophysiological coherence, and create a permissive microenvironment for microvascular self-organization and inosculation. In this sense, temperature-responsive harvesting directly supports the integrative co-design of tissue architecture, metabolic maturation, and vascular development. Temperature-responsive culture surfaces for enzyme-free harvesting of cell sheets. (A) Schematic of cell behavior on a poly (N-isopropylacrylamide) (PIPAAm)–grafted surface. At temperatures above 32 °C, the surface is relatively hydrophobic and cells show active spreading and firm adhesion, whereas at temperatures below 32 °C the surface becomes hydrophilic and cell adhesion is weakened. (B) Comparison between conventional trypsin treatment and temperature-mediated detachment. On standard culture dishes, confluent cell monolayers are dissociated into single cells by trypsinization. On PIPAAm-grafted dishes, lowering the temperature induces a hydrophobic-to-hydrophilic transition, allowing the confluent cell monolayer to detach as an intact cell sheet together with its deposited extracellular matrix. Compared with other vascularized cardiac tissue platforms, cell-sheet assemblies represent a distinct design option. Microfluidic systems such as the vascularized hiPSC-derived cardiac microtissue-on-chip reported by Ulgu et al. [38] enable precise control of perfusion and support flow-dependent anastomosis between internal and external microvascular networks, making them particularly powerful for studying endothelial–cardiomyocyte crosstalk, barrier function, and drug or inflammatory responses. Likewise, the vascularized cardiac spheroid-on-chip platform reported by Di Cio et al. [39] provides a useful framework for perfusion-based therapeutic toxicity testing in beating microtissues. By contrast, cell-sheet platforms are especially attractive when the design goal is to build dense laminar myocardial tissues while preserving endogenous extracellular matrix and cell–cell coupling and to scale these tissues through hierarchical stacking or integration with perfusable vascular beds. Thus, these methods may be viewed as complementary rather than competing approaches: microfluidic vascularized microtissues and spheroids are highly effective for mechanistic analysis and screening, whereas cell-sheet-based scaffold-free tissues are particularly well suited to constructing thicker myocardial modules for regenerative and circulatory-unit applications. For vascularization, scaffold-free constructs permit endothelial and mural cells to self-organize into networks that can anastomose with host vessels or engineered channels. The absence of dense synthetic scaffolds may lower barriers to inosculation and remodeling and allow microvessels to align with stress fields and metabolic gradients. Layering or modular assembly of pre-vascularized sheets or spheroids can produce thicker, yet perfusable, slabs. Metabolically, scaffold-free constructs respond dynamically to perfusion and substrate changes. As cardiomyocytes mature and mitochondrial content increases, local oxygen consumption shapes gradients that influence angiogenic signaling and vascular remodeling. This reciprocity between metabolism and vascular growth is harder to capture in rigidly patterned scaffolds but can emerge more readily in systems where cells can reorganize. However, this plasticity also poses challenges for reproducibility, making perfusion bioreactors, mechanical loading regimens and well-timed maturation protocols essential to steer self-organization toward desired states. Concept of integrative circulatory engineering These considerations motivate a broader conceptual framework, integrative circulatory engineering, which treats engineered myocardium as a component of a designed circulation in which perfusion, oxygenation, substrate delivery and waste removal are explicitly specified and actively controlled [6,21]. Within this circulation, metabolic maturation of cardiomyocytes and vascular network development are co-designed over time, with dynamic perfusion used to synchronize changes in demand and supply. The central design question becomes: What combination of cardiomyocyte metabolic state, vascular architecture and perfusion dynamics yields a self-sustaining, adaptable myocardial unit capable of integrating with host circulation? Scope of the review Within this conceptual landscape, Chapters 2-4 examine how metabolic maturation strategies and vascular design principles can be combined to advance scaffold-free myocardial tissue engineering. Chapter 2 focuses on cellular and structural design inputs (cell composition, paracrine crosstalk, ECM architecture/mechanics, and spatial patterning) that set boundary conditions for metabolic demand–supply balance. Chapter 3 focuses on perfusion bioreactors as implementation levers to control oxygenation, shear conditioning, substrate delivery, and sensing/feedback. Chapter 4 then synthesizes these elements into integration principles that coordinate metabolic maturation with vascular architecture and perfusion in time and space. We move from cellular metabolism and oxygen dynamics through microvascular architecture and perfusion bioreactor design to in vitro circulatory units and translational deployment. Perfusion is emphasized as a dynamic design variable rather than a static culture condition, and metabolic and vascular factors are treated as interdependent. Integrative circulatory engineering, grounded in scaffold-free platforms and informed by physiology and systems-level thinking, provides a unifying framework for this endeavor. Cellular and structural foundations of Metabolic–Vascular integration This chapter defines the cellular and structural design inputs that govern the coupling between metabolic demand and vascular supply in engineered myocardium. We focus on tissue-building variables—cell-type composition, paracrine interactions, ECM architecture/mechanics, and spatial patterning—that determine diffusion distances, angiogenic capacity, and the microenvironmental context in which perfusion strategies (Chapter 3) and maturation–vascular integration principles (Chapter 4) operate [28,40]. Although these boundary conditions apply across scaffold-free tissues, hydrogel-based EHTs, bioprinted constructs and organ-on-chip systems, their practical impact differs depending on how perfusion access and microvascular organization are implemented in each platform. Together, these inputs set the boundary conditions for demand–supply matching in engineered myocardium. Cellular composition as a coupled design variable The metabolic and vascular phenotype of engineered myocardium reflects the ensemble of cell types, not cardiomyocytes alone. Native heart tissue contains endothelial cells, smooth muscle cells, pericytes, fibroblasts, resident immune cells and specialized conduction cells [41]. Reconstituting at least a subset of this diversity is crucial for capturing interdependent maturation of metabolism and vasculature. Cardiomyocyte source and differentiation protocol influence baseline metabolic state, mitochondrial content and stress responsiveness [4,27]. Some protocols yield cells that rapidly upregulate oxidative phosphorylation; others maintain glycolytic dominance for extended periods. These starting conditions dictate how aggressively metabolic maturation can be pursued without exceeding oxygen delivery capacity. Endothelial cells derived from pluripotent stem cells, organ-specific progenitors or primary isolates differ in angiogenic potential, barrier behaviour and metabolic profile [29,42]. Coronary-like endothelial cells may better form stable, shear-resistant microvessels and support cardiomyocyte function, but may require specific cues to maintain identity [33,43]. Mural cells modulate vessel stability and reactivity, while fibroblasts regulate extracellular matrix deposition, mechanics and paracrine signaling [44,45]. The relative proportions and maturation states of these populations should therefore be treated as active design parameters, as they strongly influence diffusion pathways, angiogenic capacity and metabolic cross-feeding. Paracrine and metabolic crosstalk Metabolic maturation and vascular development are tightly linked through paracrine communication [[46], [47], [48]]. Cardiomyocytes experiencing increased workload or transient hypoxia upregulate pro-angiogenic factors such as VEGF and angiopoietins, which stimulate endothelial sprouting and network remodeling [46,47,49]. As perfusion improves, oxygen and substrate delivery feed back onto cardiomyocyte metabolism, enabling further mitochondrial biogenesis and shifts toward FAO [47,48,50]. Non-myocytes participate in this dialogue. Cardiac fibroblasts secrete extracellular matrix proteins and cytokines that influence both angiogenesis and metabolic gene expression. Depending on activation state, fibroblasts can support metabolic maturation by promoting appropriate matrix stiffness and alignment, or drive maladaptive remodeling that stiffens tissue and limits diffusion [47,48]. Endothelial cells produce angiocrine factors that modulate cardiomyocyte growth, survival and metabolism, and regulate trans-endothelial transport of fatty acids and other substrates via transporters and lipases [47,48,50]. Culture timelines—when endothelial and mural cells are introduced, how long pro-angiogenic cues are maintained—critically affect how this crosstalk shapes metabolic–vascular integration [47,48]. Matrix architecture and mechanics Even in scaffold-free systems, the extracellular matrix (ECM) produced by cardiac and stromal cells is a major determinant of metabolic maturation and vascular patterning. Composition, alignment and crosslinking of collagen, fibronectin and laminin modulate cell shape, traction forces and mechanotransductive signaling, and recent work by Tani and colleagues has shown that heart-derived collagen within engineered heart tissue can enhance sarcomeric organization and mitochondrial maturation, underscoring the importance of matrix origin [17,18]. Cardiomyocytes cultured on substrates with physiological stiffness and anisotropy exhibit improved sarcomere organization, conduction and mitochondrial function compared with those on excessively soft or stiff surfaces [19,20]. In three-dimensional tissues, constructs that self-organize into aligned laminar structures with appropriate tension support more adult-like energetics than isotropic, mechanically disordered aggregates [15,20]. For microvessel formation, ECM density and crosslinking must permit endothelial sprouting and lumen formation while providing sufficient mechanical support for stable perfusion [51]. Cyclic stretch and contraction further couple metabolism and vascular development. As scaffold-free constructs beat spontaneously or under pacing, regional strain fields guide vessel orientation and redistribute metabolic demand. Culture conditions that promote synchronized, physiologic contraction—without overstretch—provide another lever to shape both mitochondrial maturation and capillary patterning [15]. Spatial patterning for controlled metabolic landscapes Spatial arrangement of cardiomyocytes, endothelial cells and stromal populations determines how oxygen and metabolites are consumed and distributed. Rather than attempting to eliminate heterogeneity, integrative design can exploit intentional patterning [40]. Strategies include stacking cardiomyocyte-rich sheets with interposed vascularized stromal layers to create repeating laminar units in which perfusion pathways are regularly positioned relative to zones of high metabolic demand [37,52,53]. Compared with purely geometric patterning in printed or scaffolded constructs, layer-by-layer assembly allows the spatial relationship between demand (cardiomyocyte-rich laminae) and supply (vascularized interlayers) to be tuned while preserving anisotropic conduction and contraction; stacked-sheet approaches and related modular fusion or vascularized microtissue implementations provide examples of this design logic [36,53]. Spheroids or microtissues with predefined compositions can be fused into larger constructs, with some aggregates serving as “perfusion hubs” that preferentially support microvascular branching [36,53]. Gradients in cell composition or maturation state can further align oxidative maturation with local perfusion capacity—for example, placing more glycolytic, immature cardiomyocytes in regions where perfusion is initially limited and more oxidative, mature cells near perfusion inlets, then advancing maturation distally as networks refine and perfusion improves [36]. Quantitative assessment and design workflows Quantitative metrics are needed to evaluate metabolic–vascular alignment. At the metabolic level, regional oxygen tension, mitochondrial membrane potential, ATP/ADP ratios and substrate utilization rates can be mapped using oxygen-sensitive probes, NADH/FAD autofluorescence and genetically encoded sensors [[53], [54], [55]]. Coupled with functional readouts such as calcium transients and strain, these maps reveal how metabolic state correlates with workload. For the vascular compartment, imaging of network topology, vessel diameter, perfusion fraction and shear surrogates using confocal or light-sheet microscopy, optical coherence tomography angiography or micro-CT provides structural context [56]. Integrating these datasets reveals regions with high metabolic demand but sparse perfusion, or well-vascularized areas with persistently glycolytic cardiomyocytes, highlighting misalignments that call for adjustment in cell composition, maturation or perfusion strategy. Design workflows can proceed from top–down, specifying a target phenotype (for example, a slab with near-adult oxidative capacity and defined capillary spacing) and inferring constraints on cell composition and perfusion, or bottom–up, mapping intrinsic self-organization of available populations and then steering it through modest interventions. In practice, both approaches will be combined, with scaffold-free tissues offering a particularly fertile ground for this interplay. Perfusion bioreactors and control of the Metabolic–Vascular milieu Metabolic maturation and vascular architecture realize their full potential only when embedded in an appropriately designed perfusion environment. Static culture can support thin tissues and relatively immature, glycolytic cardiomyocytes, but it becomes insufficient as constructs thicken and oxidative metabolism increases. Perfusion bioreactors therefore serve not simply as culture accessories, but as active design tools that define convective transport, endothelial conditioning, and dynamic control of the tissue milieu. In scaffold-free myocardial engineering, perfusion design is the principal means by which metabolic trajectories and vascular remodeling are coordinated in space and time. Fig. 3 provides a comparative overview of representative perfusion strategies used across engineered cardiac tissues, including interstitial perfusion, luminal perfusion, vascular network perfusion, and surface-fed perfusion. This chapter builds on that comparative framework to discuss how perfusion geometry, flow regime, oxygen and substrate delivery, and sensing-enabled control together shape the metabolic–vascular microenvironment and determine whether engineered cardiac tissues can progress from diffusion-limited survival toward robust oxidative function. Representative perfusion strategies in engineered cardiac tissues. Schematic comparison of commonly used perfusion configurations that define oxygen and nutrient transport environments in engineered cardiac tissues. (A) Interstitial perfusion, in which culture medium is convectively delivered through tissue matrices or porous structures surrounding the cells. (B) Luminal perfusion, in which flow is directed through endothelialized channels that establish defined vascular lumens and provide shear conditioning to endothelial cells. (C) Vascular network perfusion, in which flow is routed through self-organized or engineered microvascular networks (including implementations in microfluidic/organ-on-chip platforms) to support exchange across dense tissues and to condition endothelium under defined hemodynamics. (D) Surface-fed perfusion, in which medium flows along the tissue surface and transport into the construct occurs primarily through diffusion from the surrounding medium. These configurations represent complementary engineering strategies for establishing transport pathways in engineered cardiac tissues and are shown here to illustrate how perfusion geometry, vascular interfaces, and flow paths influence oxygen delivery and metabolic demand. From diffusion-limited culture to representative perfusion configurations As outlined in Chapter 1, static culture is fundamentally constrained by diffusion, and these limitations become increasingly restrictive as cardiomyocytes mature metabolically and oxygen demand rises. Perfusion bioreactors address this limitation by introducing convective transport through or around the construct. As summarized in Fig. 3, representative configurations include interstitial perfusion through tissue matrices or porous supports, luminal perfusion through endothelialized channels, perfusion through vascularized networks, and surface-fed culture in which transport remains largely diffusion-dominant despite medium flow along the tissue boundary. These configurations create distinct oxygen and substrate landscapes, expose endothelial cells to different shear environments, and impose different architectural constraints on tissue design. Accordingly, perfusion geometry should be selected with the anticipated mature metabolic state in mind, rather than the initial low-demand state alone. Flow regime, shear stress and endothelial phenotype Beyond transport itself, perfusion defines the mechanical microenvironment experienced by endothelial cells. Shear stress is a central regulator of endothelial phenotype and metabolism, and therefore a key design variable in vascularized cardiac tissues. In native coronary vessels, laminar and pulsatile shear promote an anti-inflammatory, nitric oxide-producing, barrier-competent phenotype that remains metabolically supportive of adjacent cardiomyocytes. Conversely, disturbed or chronically low shear favours inflammatory activation, barrier disruption, and maladaptive endothelial states. These considerations apply differently across the representative perfusion strategies shown in Fig. 3. For example, luminal and vascular network perfusion permit more direct control of endothelial shear exposure, whereas interstitial and surface-fed approaches influence endothelial state more indirectly through local transport fields and boundary conditions. In engineered tissues, physiologically appropriate shear should stabilize endothelial monolayers, promote junctional integrity, and maintain an endothelial state compatible with support of oxidative cardiomyocytes. Oxygenation and substrate delivery in metabolically advancing tissues As cardiomyocytes transition toward more oxidative phenotypes, oxygen delivery and substrate availability become increasingly coupled. Oxygen supply depends not only on inlet oxygen tension and flow rate, but also on the available exchange surface and on local tissue consumption. In parallel, perfusion systems define the chemical environment by controlling the delivery of glucose, lactate, fatty acids, ketone bodies, amino acids, and endocrine factors. The representative configurations in Fig. 3 differ fundamentally in how oxygen and substrates are delivered: interstitial perfusion enhances convective exchange through porous surroundings, luminal perfusion couples delivery to defined endothelialized channels, vascular network perfusion distributes flow through branched microvascular pathways, and surface-fed perfusion relies primarily on diffusion from the flowing external medium. From a design perspective, oxygenation and substrate composition should therefore be staged together with perfusion architecture: diffusion-dominant conditions may suffice for early survival and network formation, whereas progression toward fatty acid oxidation and adult-like workload requires increased oxygen delivery and more precisely tuned media composition. Thus, oxygenation strategy is not separate from metabolic maturation, but one of its principal enabling conditions. Controlled, transient hypoxia can be used as an angiogenic or metabolic cue, but must be carefully bounded to avoid irreversible injury. Hemoglobin-based oxygen carriers or perfluorocarbon emulsions can extend viable thickness but introduce additional variables such as viscosity and potential endothelial interactions that must be managed [57]. Control of workload and synchronized maturation cues Perfusion bioreactors do not merely deliver medium; they establish the broader physicochemical context in which cardiomyocytes mature. Mechanical load, whether imposed by intraluminal pressure, external stretch, pacing, or afterload elements, directly alters ATP demand and oxygen consumption, and perfusion must scale accordingly. Likewise, staged changes in substrate composition and endocrine cues can push constructs along different metabolic trajectories. These considerations reinforce the need to treat perfusion as a dynamic variable rather than a fixed setting. In practice, maturation is best supported when oxygen delivery, substrate composition, and workload are increased in a coordinated manner, thereby minimizing mismatches between rising demand and available vascular support. Across the representative strategies shown in Fig. 3, the central design challenge is not simply to maintain flow, but to match perfusion mode and control logic to the evolving metabolic and vascular state of the tissue. Sensing-enabled feedback control Because maturation and vascular remodeling are non-linear and heterogeneous processes, fixed perfusion settings are rarely optimal throughout culture. Modern perfusion systems increasingly incorporate measurements of oxygen tension, pH, glucose, lactate, pressure, and flow, together with optical or electrophysiological readouts that report tissue function. These data provide the basis for feedback control, enabling adjustment of flow rate, pressure, substrate composition, and mechanical stimulation in response to the evolving state of the construct. Although Fig. 3 summarizes structural categories of perfusion strategy rather than control architecture, each of these configurations can in principle be combined with sensing-enabled regulation. At the simplest level, outlet oxygen or pressure can be maintained within predefined limits; more advanced strategies use integrated monitoring to anticipate changes in metabolic demand and adjust the culture environment pre-emptively. The long-term objective is a self-tuning circulatory environment that supports engineered cardiac tissues through the transition from fetal-like glycolytic metabolism to stable oxidative performance. Integrating metabolic maturation with vascular design This chapter consolidates solution-oriented integration principles that link metabolic maturation to vascular architecture and perfusion. We organize the discussion around time (when to advance oxidative maturation relative to vascular capacity), space (how to engineer oxygen/substrate landscapes), and feedback (how sensing-informed adjustments in perfusion and media maintain alignment between demand and supply). By framing maturation and vascularization as a coupled design problem, this chapter builds on the design inputs in Chapter 2 and the implementation levers in Chapter 3. Importantly, integrative circulatory engineering is intended as an organizing framework rather than a single fabrication method. The same coupling principles can be implemented using scaffold-free tissues, hydrogel-based EHTs or microfluidic perfusion platforms, provided that maturation, vascular architecture and perfusion are co-designed and linked by measurable feedback. Key elements include. •Temporal co-design: increases in oxidative capacity occur in parallel with, or shortly after, establishment of sufficient microvascular density and perfusion capacity. Temporal co-design: increases in oxidative capacity occur in parallel with, or shortly after, establishment of sufficient microvascular density and perfusion capacity. •Spatial control: oxygen and substrate gradients are engineered, not merely tolerated, to pattern cardiomyocyte phenotypes and vascular remodeling. Spatial control: oxygen and substrate gradients are engineered, not merely tolerated, to pattern cardiomyocyte phenotypes and vascular remodeling. •Endothelial–cardiomyocyte coupling: endothelial cells are treated as active metabolic regulators and paracrine partners, with shear and substrate exposure tuned to promote stable, supportive phenotypes. Endothelial–cardiomyocyte coupling: endothelial cells are treated as active metabolic regulators and paracrine partners, with shear and substrate exposure tuned to promote stable, supportive phenotypes. •Feedback and adaptation: real-time measurements of oxygen, metabolites and functional performance inform adjustments in perfusion, mechanical load and media composition. Feedback and adaptation: real-time measurements of oxygen, metabolites and functional performance inform adjustments in perfusion, mechanical load and media composition. Metabolic maturation as a driver of perfusion requirements Protocols that promote metabolic maturation of hPSC-derived cardiomyocytes typically combine substrate manipulation, provision of long-chain fatty acids and endocrine cues [4,58]. These reprogram metabolism from fetal-like glycolytic states toward oxidative phosphorylation and FAO, accompanied by increased mitochondrial biogenesis, sarcomeric organization and electrophysiological maturity. Oxygen consumption and redox burden rise accordingly. In immature constructs, modest oxygen demand can be met by surface diffusion or sparse vascular channels. As oxidative capacity increases, oxygen consumption approaches adult myocardium, requiring microvascular spacing on the order of tens of micrometres and continuous convective supply [2,5]. Strategies that accelerate metabolic maturation must therefore be co-designed with vascular architecture; otherwise, the most advanced regions may be selectively lost to hypoxia. Oxygen landscapes during maturation Oxygen distribution in thick sheets or aggregates is governed by reaction–diffusion–convection processes, with consumption strongly dependent on developmental stage. Early in differentiation, low mitochondrial content yields shallow gradients; upregulation of FAO and oxidative phosphorylation steepens gradients and shrinks diffusion distances [4,5,27]. Classical models indicate that even modest increases in consumption substantially reduce allowable diffusion distances. For scaffold-free constructs produced by cell-sheet stacking or condensation, this means that the number of layers supportable under static conditions is tightly constrained once adult-like metabolism is targeted. Vascular design must anticipate future oxygen demand. Dynamic perfusion can reshape oxygen landscapes over time, with flow rates, medium oxygen tension and oxygen carriers adjusted to maintain physiological profiles during maturation. Mild, controlled gradients may be used as patterning cues, but require precise coordination with network geometry and metabolic stage. Substrate switching and spatial nutrient control Metabolic maturation also depends on temporal and spatial delivery of energy substrates. Fetal and early postnatal cardiomyocytes favour glucose and lactate; adult cells rely primarily on FAO but retain metabolic flexibility [59]. Tohyama-inspired protocols exploit this trajectory via glucose-deprived, lactate-rich media, followed by FAO-promoting formulations [9,24,25]. In parallel, we have optimized fatty acid formulations for engineered myocardial tissues, demonstrating that tailored lipid composition can increase the contractility of cell sheet–based constructs and thereby couple metabolic programming to improved pump performance [60]. In perfused constructs, substrate manipulations acquire a spatial dimension. Gradients inevitably arise along flow paths and across thickness. If deliberately designed, inlet regions may be enriched in fatty acids and oxygen to promote highly oxidative cardiomyocytes, while downstream regions experience slightly lower oxygen or different substrate ratios, supporting heterogeneity reminiscent of in vivo zonation. Achieving this safely requires careful control of flow, channel architecture and medium formulation to avoid lipotoxicity or energy failure in less well-perfused areas. Endothelial support and cardio–vascular metabolic coupling Endothelial cells actively shape cardiac metabolism. In adult heart, endothelial metabolism regulates barrier function, NO bioavailability, angiogenic responsiveness and substrate traffic [[61], [62], [63]]. Endothelial cells consume relatively little oxygen yet influence the microenvironment through glycolysis, lactate production and regulation of FA transport. In engineered tissues, metabolically competent endothelial networks can both stabilize vasculature and tune cardiomyocyte maturation. Shear stress modulates endothelial phenotype and metabolism, promoting NO production, anti-inflammatory signaling and, in some contexts, enhanced FAO [61,[64], [65], [66]]. These changes influence cardiomyocytes via paracrine factors and by gating substrate delivery. Flow regimes in engineered networks should therefore be chosen to approximate physiological coronary shear, fostering adult-like endothelial phenotypes and beneficial metabolic coupling. Coupled remodeling and design principles In vivo, metabolic demand and vascular supply are coupled through angiogenesis, vasomotor control and structural remodeling. Hypoxia and metabolic stress trigger VEGF and other signals; shear drives arteriogenesis and pruning [[67], [68], [69]]. In engineered tissues, analogous feedbacks can emerge only if perfusion conditions allow adaptation. Fixed, uniform flows and static media hinder co-evolution. Perfusion systems endowed with sensing and control can deliberately harness these feedbacks. Rising oxygen extraction or lactate production can trigger increases in flow or targeted angiogenic stimuli; signs of hypoxia can prompt redistribution of perfusion or adjustment of substrate supply. Scaffold-free constructs, unconstrained by rigid scaffolds, are well suited to such adaptive schemes. Integrating these considerations suggests several design principles. Metabolic state should be treated as a programmable variable with explicit trajectories co-designed with vascular structure and perfusion. Oxygen and substrates should be delivered not only to avoid necrosis but also to provide physiologically relevant gradients. Endothelial and mural phenotypes must be tuned by shear and substrate exposure to support stable, supportive networks. Bioreactors should implement multi-parametric sensing and closed-loop adaptation to match rising metabolic demand with vascular capacity. Quantitative benchmarks—oxygen consumption, FAO/glucose oxidation ratio, mitochondrial content, capillary spacing, perfused fraction, shear profiles and endothelial NO production—are needed to operationalize and compare designs [56,70]. In vitro circulatory units, translational integration and design optimization The convergence of metabolic maturation and vascular design reaches its fullest expression when myocardial constructs are embedded in higher-order circulatory environments. At this scale, scaffold-free myocardial modules, perfusable vascular beds, controlled hemodynamics and programmable metabolic inputs are combined into in vitro circulatory units that approximate key organizational features of the heart within the body. These systems enable advanced maturation, provide a bridge to translational deployment and expose a complex design space that benefits from computational and control-theoretic tools. From tissue modules to in vitro circulatory units Conventional cardiac tissue engineering typically yields single functional modules—a contractile patch, a sheet stack or a vascularized spheroid—cultured in static medium or under simple perfusion. In contrast, an in vitro circulatory unit is conceived as a closed or semi-closed loop in which one or more myocardial modules are integrated with engineered vascular networks, compliance elements and pumps that generate physiologically inspired pressure–flow relationships [71,72]. A conceptual example is shown in Fig. 4A, where cardiac organoid-based modules are integrated with a perfusable vascular bed, and in Fig. 4B, which illustrates a hierarchical framework linking cell sheet–based myocardial tissues to progressively more complex in vitro circulatory units and multi-organ circulatory platforms. Conceptual framework for integrative circulatory engineering. Schematic illustration of how tissue modules, vascular interfaces, and perfusion-based control may be integrated to support metabolic maturation and functional coupling in engineered cardiac tissues. This figure is presented as a conceptual synthesis of the future design direction proposed in this review, rather than as a representative image of any single group's experimental system. Within this architecture, perfusate composition, oxygenation and flow waveforms are explicit design variables. Early phases can employ low afterload, moderate flow and substrates compatible with relatively glycolytic cells; later phases gradually increase hemodynamic load, oxygen content and fatty acid availability in concert with vascular maturation. Endothelial networks are conditioned by pulsatile shear, steering them toward anti-inflammatory, NO-producing phenotypes compatible with coronary behaviour [71,73]. Myocardial modules and vascular compartments form a coupled system in which changes in metabolic state alter oxygen extraction and perfusate composition, which in turn affect vascular tone and microvascular flow, feeding back onto myocardial workload. Hemodynamic loading as a regulator of metabolic maturation The adult heart operates under substantial and varying hemodynamic load. Preload, afterload and heart rate modulate mechanical work and, consequently, energy demand and substrate utilization [74,75]. Engineered myocardium can exploit this relationship by treating hemodynamic load as a controllable input for maturation rather than merely a stressor. In vitro circulatory units allow imposition of defined pressure–volume conditions through combinations of pumps, resistive elements and compliance chambers. By regulating these components, target stroke volume, end-diastolic pressure and systolic load can be prescribed. Stepwise escalation of afterload and pacing frequency, synchronized with introduction of fatty acid–rich, oxygenated perfusate, can drive cardiomyocytes toward adult-like excitation–contraction–metabolism coupling. Vascular capacity must keep pace: as work and oxygen consumption rise, capillary density, perfused fraction and shear profiles must remain within physiological ranges. Load ramping thus becomes both a maturation tool and a functional stress test revealing whether metabolic and vascular integration are adequate [74,76]. Multi-organ coupling and systemic metabolic context In vivo, myocardial metabolism is tightly interwoven with the function of liver, adipose tissue, skeletal muscle and kidney. Circulating levels of glucose, fatty acids, lipoproteins, ketone bodies and amino acids reflect whole-body metabolic status and strongly influence cardiac substrate selection [77,78]. To capture this context, circulatory units can be extended to multi-organ platforms linking myocardial modules with liver-like and adipose-like tissues. Recent work on perfusable endocrine tissues with anastomosable vasculature provides a template for such organ-specific modules within multi-organ circulatory systems [79]. In such systems, a shared circulation delivers substrates and hormones to all modules, while organ-specific microcirculatory beds and local control elements tailor pressure and flow. A liver module may regulate glucose output and lipid processing; an adipose module may modulate fatty acid release in response to hormonal cues. Myocardial modules then experience systemic substrate patterns akin to those in health or disease. Metabolic maturation thus becomes a multi-organ design problem, and disease states—such as insulin resistance, obesity or ketogenic adaptation—can be modelled by adjusting organ-specific functions and perfusate composition [72,80,81]. Translational scenarios, host integration and preconditioning Translating metabolically and vascularly integrated constructs into regenerative therapies requires careful consideration of implantation scenarios and host environments. Early applications are likely to involve relatively small, metabolically competent patches intended to support infarct border zones and prevent adverse remodeling, whereas later strategies may target wall-spanning constructs for ventricular reconstruction [82,83]. Each scenario imposes distinct demands on capillary density, perfusion routes, mechanical resilience and metabolic flexibility. In vitro circulatory units can serve as preconditioning platforms. Perfusion pressures and waveforms can be tuned to approximate those at the planned implantation site, conditioning endothelial and mural cells to expected hemodynamics. Substrate composition and oxygen tension can be adjusted to mirror the host's metabolic state—normoglycemic, diabetic, obese or otherwise—allowing cells to adapt before transplantation [73,75]. Pre-implantation stress tests, such as simulated ischemia–reperfusion, acute afterload surges or abrupt substrate shifts, can be used to evaluate graft resilience. Constructs that sustain function and structural integrity under these challenges are more likely to withstand perioperative insults. Host–graft metabolic coupling must also be considered. A highly FAO-competent graft implanted into hypoperfused or metabolically compromised myocardium may experience energy imbalance, whereas a relatively immature, glycolytic graft may be overwhelmed by a highly oxidative milieu. Tailoring in vitro maturation and terminal preconditioning to specific implantation scenarios may minimize these mismatches and promote gradual convergence toward the host phenotype. Disease modelling and precision medicine The same integrative principles can be harnessed to build high-fidelity disease models. Many cardiomyopathies feature combined disturbances in substrate use, mitochondrial function and microvascular integrity, as in diabetic cardiomyopathy, pressure-overload hypertrophy and mitochondrial diseases [72,84]. Scaffold-free myocardial modules derived from patient-specific iPSCs can be matured within circulatory units under conditions chosen to reproduce disease-relevant metabolic and hemodynamic states. Diabetic cardiomyopathy, for example, may be modelled by exposing constructs to hyperglycemic, lipid-rich perfusate, altered hormonal profiles, increased afterload and modest microvascular rarefaction. Phenotypes such as diastolic dysfunction, impaired metabolic flexibility and microvascular dysfunction can then be probed. Vascular design also enables modelling of microvascular pathologies; endothelial cells carrying disease-associated variants can line perfused networks, with downstream cardiomyocytes experiencing consequences of abnormal vasomotor responses or barrier properties. These platforms are naturally suited to precision medicine: patient-specific constructs can be exposed to panels of metabolic agents, vasodilators or combination regimens, and integrated readouts—contractile performance, arrhythmia propensity, metabolic fluxes and microvascular behaviour—can generate multi-dimensional response fingerprints. These efforts build on the broader landscape of engineered heart tissue and organoid models for disease modelling and drug discovery developed by Tohyama and co-workers. Computational design, optimization and standardization The design space of integrative circulatory engineering is high-dimensional, spanning cell composition, metabolic protocols, vascular topology, perfusion parameters and loading. Exploring this space empirically is inefficient. Multi-scale computational models that link cellular metabolism, tissue-level diffusion–perfusion and network-scale hemodynamics can guide design by predicting how specific parameter combinations affect oxygen distribution, substrate utilization and functional outputs [76,85]. Data from well-characterized constructs—oxygen consumption, metabolite fluxes, capillary density, flow maps and functional performance—are needed to calibrate and validate such models. Optimization frameworks, including evolutionary algorithms and machine learning–based surrogate models, can then search for designs that satisfy multiple objectives, such as maximizing contractile output while minimizing hypoxic fraction and arrhythmia risk. Candidate “operating points” emerging from in silico exploration can be realized experimentally and iteratively refined. In parallel, standardization will be essential. Reference protocols for perfusion-based maturation (for example, “early glycolytic,” “intermediate mixed,” “late FAO-dominant” stages) could specify typical ranges for flow, oxygenation and substrates. Reporting guidelines that include detailed perfusion parameters, capillary metrics and metabolic benchmarks would facilitate cross-study comparison and meta-analysis. Over time, a library of validated design templates may emerge, tailored to regenerative patches, pediatric grafts or disease-specific models. Outlook within an integrative circulatory framework Embedding scaffold-free myocardial constructs into in vitro circulatory units, and coupling them to multi-organ contexts and computational design tools, completes the conceptual arc of integrative circulatory engineering. Metabolic maturation is no longer a purely cell-intrinsic refinement, and vascularization is no longer a structural add-on; both become elements of a designed circulation dynamically tuned to support, test and prepare tissues for their intended roles. Future progress will depend on tighter integration of sensing, control and modelling [71,72]. Real-time imaging of mitochondrial function and microvascular flow, genetically encoded metabolic reporters and adaptive control algorithms could transform perfusion systems into self-optimizing environments that shepherd constructs through complex maturation trajectories. As these technologies mature, integrative circulatory engineering may provide not only more faithful cardiac models and more robust grafts, but also a framework for understanding how metabolism, vasculature and hemodynamics jointly define cardiac health and disease—and how they can be re-engineered for regenerative therapy. Declaration of generative AI and AI-assisted technologies in the manuscript preparation process During the preparation of this work, the authors used ChatGPT (OpenAI, San Francisco, CA, USA) in order to assist with editing and refining the English language of the manuscript (e.g., phrasing, grammar and clarity). After using this tool, the authors carefully reviewed and edited all content and take full responsibility for the content of the published article. Declaration of competing interest The authors declare that they have no conflicts of interest. Acknowledgements This work was supported by JSPS KAKENHI Grant Number JP24H00803. Footnotes This article is part of a special issue entitled: In Vitro Vascular Integration published in Regenerative Therapy. Peer review under responsibility of the Japanese Society for Regenerative Medicine. References