cFBs are geometrically interspersed between cardiomyocytes and, in normal circumstances, lead to extracellular matrix (ECM) homeostasis in the cardiac tissue. Immediately after IRI, activated fibroblasts differentiate into myofibroblasts, triggering an initially protective form of cardiac fibrosis aimed at preserving structural integrity. Fibrotic EMC remodeling and inflammatory activation can lead to long-term cardiac complications due to excessive scarring (Hinz et al., 2012). Myofibroblast differentiation is driven by the expression of contractile genes, such as ACTA2, which encode for smooth-muscle α-actin (α-SMA) (Tallquist and Molkentin, 2017). Myofibroblasts can worsen local inflammation by activating the NLRP3 inflammasome response (Sandanger et al., 2013). Once differentiated, myofibroblasts increase the ECM, deposition factors like collagen -I, -III, -IV, -V, and -VI, glycoproteins such as fibronectin and tenascin-C, and proteoglycans, all responsible for the healing and scarring processes, as reviewed in detail elsewhere (Tallquist and Molkentin, 2017; Kanisicak et al., 2016). The scarring process is necessary for maintaining the structural integrity and functionality of the heart. However, excessive fibrosis can result in adverse remodeling, increased stiffness, reduced contractility, and, ultimately, heart failure (Zhang et al., 2024).

ECs are more than just a protective barrier controlling the exchange of nutrients and gasses to the cardiac muscle. In a stress and inflammatory setting, ECs regulate the adhesion and transmigration of immune cells, contributing to the inflammatory response within the cardiac tissue. During IRI, ECs play an important role in the aggravation of and protection against IRI. ECs impair vasodilation by reducing NO production, increasing vascular permeability inflammation that could contribute to microvascular obstruction, one of the hallmarks of IRI. Additionally, in the context of IRI, ECs may drive tissue remodeling and hypertrophy after ischemia, as reviewed by Yang et al. (2015) and Singhal et al. (2010). EC-derived extracellular vesicles (EC-EVs) play a crucial role in the alleviation of IRI (Yadid et al., 2020; Li et al., 2023), illustrated by a ∼50% increase in cardiomyocyte viability after an IRI stimulus when EC-EVs were added (Yadid et al., 2020). Mechanistically, Liu et al. discovered that exosomes, a small subtype of extracellular vesicles, derived from human-induced pluripotent stem cell-derived endothelial cells (hiPSC-ECs) restored the expression and activity of reticulum Ca²⁺ ATPase 2a (SERCA-2a) and ryanodine receptor (RYR2) in cardiomyocytes. Consequently, intracellular Ca²⁺ transient and cardiomyocyte contractions were enhanced after MI (Li et al., 2023).

When developing an appropriate research model to study the pathophysiology of IRI, it is crucial to consider both beneficial and deleterious local interactions. Incorporating these interactions will enable the creation of an in vitro model that more accurately reflects the human in vivo environment.

Engineered heart tissues (EHTs) are 3D structures typically created by mixing hiPSC-CMs with a hydrogel into a casting mold, providing a highly reproducible tissue structure. CMs in EHTs align along the force lines of the cell or culture they are placed on, such as rings (Katare et al., 2010), sheets—as used by Yadid et al. (2020), measuring 3.2 mm × 4.2 mm—or elastomeric pillars (Ahmed et al., 2020; Arslan et al., 2022; Mastikhina et al., 2020). Mastikhina et al. (2020) included fibroblasts in their cardiac EHT and investigated the fibrotic effects after ischemia using 2.5*10^5 cells placed between two polydimethylsiloxane (PDMS) pillars. The CMs created a coherent beating syncytium that allows for the detailed measurement of contractile function, including force generation and electrophysiological properties. EHTs are generally easy to produce and can be manufactured on a relatively large scale; however, due to the lack of flow, they are prone to the development of necrotic inner tissue due to oxygen deprivation (Stein et al., 2021).

Heart-on-a-chip models are based on a microfluidic system that enables dynamic culture, perfusion, and the addition of a vascular channel (Arslan et al., 2022; Paz-Artigas et al., 2023). Heart-on-chips utilize a range of biocompatible natural and synthetic materials, including hydrogels, (3D-printed) synthetic polymers such as PDMS, or extracellular matrix, to provide structural support and guide cell alignment and organization. These materials allow for the precise spatial arrangement of cell culture chambers, fluidic channels, and measuring devices, such as electrodes (Liu et al., 2020; Paloschi et al., 2021). Nevertheless, heart-on-a-chip systems require specialized equipment, like microfluidic pumps, and can be equipped with biosensors, like microelectrode arrays or intracellular electrodes, for measuring extracellular field potentials and action potentials (Liu et al., 2020) or mechanical biosensors, like force transducers or cantilevers, to measure contractile forces (Yadid et al., 2020). The chip to be cultured on, the microfluidic pump system, and potential sensors add technical complexity and costs compared to 2D and scaffold-free 3D systems.

Spheroids belong to the scaffold-free models, together with organoids. Spheroids are simple micro-size 3D aggregates often generated by hanging drop or ultra-low-attachment plate methods (Paz-Artigas et al., 2023; Cho et al., 2021). They are generated from hiPSCs alone or in a mixture with ECs, FBs, or even smooth muscle cells (SMCs). These cells can exhibit synchronized contractions, which may be perceived as beating heart tissue. Regarding their compact nature, they are at an increased risk of O₂-mediated cell death; this feature can be used to recreate differences in oxygenation, which can be translated to gradients of IRI damage (Richards et al., 2020).

Organoids are more complex, self-organizing 3D structures that mimic the structural and functional properties of the heart without the need for external support to maintain mechanical integrity (Kaushik et al., 2018). Organoids are made using cells that grow in a 3D structure that resembles the organ in structure and function; this development is similar to the mesodermal development of the human heart (Song et al., 2024; Hofbauer et al., 2021a). These effects can be enhanced by the addition of BMP, VEGF, FGF, and TGF-beta-containing medium, creating organoids consisting of further differentiated cells into different lineages. Adding these factors led to further development of other cell types of the heart, reflected by a significant increase in endothelial VE-cadherin expression, increasing from 4% to 15%. Additionally, CD90 positivity, a fibroblast marker, increased from 6% to 27% when compared to CaO of a similar age and size (Song et al., 2024). Organoids are more complex structures than spheroids and more closely resemble the human heart (Gunti et al., 2021). However, they are more challenging to generate and may exhibit greater structural and functional heterogeneity.

Although some previous IRI-on-chip models were developed utilizing hiPSC-CMs (Paz-Artigas et al., 2023), Yadid et al. (2020) advanced the field by creating a multicellular model consisting of an EHT assembled on a flexible chip mimicking ventricular-like muscle. Although indirect, this model was designed to investigate IRI on a microfluidic platform (Yadid et al., 2020). A flexible cantilever chip was used to study the effects of exosomes on hiPSC-CM survival and function. The CM cantilever chip consists of micro-patterned films to facilitate hiPSC-CM alignment, creating an organized CM sheet more similar to a naïve myocardium than to a 2D monolayer. Hypoxia was induced by exposing the cells to a 1% O₂ environment for 3 h, combined with glucose-depleted and slightly acidified media (pH 6.4) to simulate the ischemic event. This was followed by 1.5 h under normal culture conditions to mimic the reperfusion phase. Exosome production was approximately 6.5 times higher in hypoxic EC-EXs than in normoxic EC-EXs. Additionally, the heart cantilever receiving EC-EXs 3 h prior to ischemia demonstrated improved cell survival and enhanced twitch stress recovery compared to non-EX-treated controls (Yadid et al., 2020). Although the uptake was equal, the positive results could not be replicated when neonatal rat cardiomyocytes were stimulated using human EC-EXs. These results could indicate a species-specific effect of human EC-EXs.

The effect of EVs in IRI was exploited further by Ellis et al. (2022). Their heart-on-a-chip model comprising hiPSC-CMs and hiPSC-ECs was used to investigate the potential role of extracellular vesicle-derived miRNAs in the context of IRI. During ischemia, about 3.5 times more EC-EXs were secreted than during reperfusion or non-ischemic controls; these results were validated in clinical plasma samples. During reperfusion, an anion exchange membrane (AEM)-based miRNA sensor was used to investigate the presence of miRNA in chip effluents compared to clinical controls. miRNAs miR-208 and miR-499 were identified in the model effluent, and they were also present in clinical IRI samples, implicating the potential of this IRI-on-a-chip model for biomarker discovery (Ellis et al., 2022).

In mice, the administration of EC-EVs intranasally every day the first 3 days and twice a week up to 3 weeks after IRI has been demonstrated to improve cardiac function in mice after a 28-day period (Wang et al., 2023). The impact of EVs on reperfusion injury depends on their source. EVs obtained not only from ECs but also isolated from the ischemic heart were administered via an intracardiac injection into a second heart just before reperfusion exacerbated IRI in mice. Further analysis revealed increased M1 polarization of macrophages, as well as increased local cytokine expression (Ge et al., 2021). The development of technologies such as IRI-on-a-chip (Yadid et al., 2020; Ellis et al., 2022) and EHT fibrosis models (Mastikhina et al., 2020) could potentially enable a more detailed investigation of the content, secretion, and effects of these EVs in human models. These advancements offer an opportunity to more comprehensively explore potential therapeutic targets or biomarkers associated with EVs.

Veldhuizen et al. (2020) developed a microdevice with a heart-on-a-chip configuration and surface topographical patterning to facilitate two-dimensional cell alignment. This microfluidic chip was constructed from PDMS and filled with hiPSC-CMs and cFBs at a ratio of 4:1 in an ECM comprising a collagen/fibronectin (Veldhuizen et al., 2020) or collagen/matrigel (Veldhuizen et al., 2022) mixture. These cells formed aligned tissues around the embedded microposts. Culturing on the PDMS chip led to the increased maturation of hiPSC-CMs compared to 2D control, as evidenced by the upregulation of genes involved in calcium uptake and release, including HCN1, KCNQ1, CAV1.2, CAV3.1, PLN, and RYR2 (Veldhuizen et al., 2020). Placing this model in a hypoxic setting (1% O₂ for 24 h), followed by reperfusion for either 1 or 24 h, revealed no change in cell death during hypoxia. However, a significant increase in cell death was observed following oxygen reperfusion at both the 1-h and 24-h time points, indicating the successful induction of reperfusion-related damage (Veldhuizen et al., 2022). Furthermore, 24 h of hypoxia led to a near 2-fold increase in lactate secretion, which is comparable to human physiological levels during IRI, where ischemia only caused minor fluctuations in the inter-beat interval, while reperfusion led to a notable increase in inter-beat variability. Additionally, reperfusion also led to increased expression of α-SMA (Veldhuizen et al., 2022). This de novo expression of α-SMA is a hallmark of myofibroblast activation, allowing for the formation of stress fibers and the production of extracellular matrix, which induces fibrosis and scarring (Hinz et al., 2012). The transcriptomic analysis revealed a notable increase in glycolysis and other metabolic pathways, accompanied by a reduction in oxidative phosphorylation (OXPHOS). Collectively, similar pathways and matching functional readouts compared to human ischemia serve as validation for IRI-on-a-chip models and allow for further developments of these 3D-IRI-on-a-chip models (Veldhuizen et al., 2022).

Although both scaffold-based and scaffold-free models allow for the co-culture of cardiomyocytes with different cell types like EVs and cFBs, an acute limitation of scaffold-free models, both organoids and spheroids, is the lack of a functional vascular network to facilitate nutrient exchange and waste removal (Homan et al., 2019), especially in modeling the reperfusion phase, where fast nutrient/oxygen exchange and waste removal are warranted. In static circumstances, only spheroids with a diameter of less than 150 μm have been used to study cardiac ischemia (Richards et al., 2017) as in this size, passive oxygen diffusion is still possible. Small organoids or spheroids are usually less complex and exhibit greater variability in their properties and reduced contractile function than their larger (600 μm) counterparts (Hoang et al., 2021).

Richards et al. (2020) used the absence of a vascular network as a means of gradually introducing ischemia. In their study, 300-μm cardiac organoids were subjected to a 10-day, 10% oxygen + adrenergic stimulation via norepinephrine. The quantification of oxygen diffusion in these microtissues revealed a reduction to 1% oxygen at the core of the organoid. Not only did the low-oxygen treatment result in significant cell death, visualized by TUNEL staining, in accordance with metabolic responses in IRI, but a reduction in non-mitochondrial respiration and an increase in glycolysis during ischemia were also observed. A meta-analysis of transcriptomic changes in these human cardiac infarct organoids revealed significant similarities with the transcriptomes of acute post-infarct tissues from animal models and human cardiac samples affected by ischemic cardiomyopathy. Gene Ontology terms between control and infarcted organoids revealed changes in pathways indicative of altered calcium handling, such as ion transport, calcium signaling, and arrhythmogenic right ventricular cardiomyopathy. A reduction in well-studied calcium handling components, including ATP2A2, RYR2, CACNA1C, and SLC8A1, was observed, and the peak calcium ion concentration, crucial for CM contraction, was reduced in the interior of the infarcted organoid compared to the edge or control CMs. In addition to altered calcium handling, bulk RNA sequencing analysis revealed changes in pathways related to fibrosis, with genetic alterations mirroring those observed in the infarcted mouse heart. Functionally, there was a notable increase in myofibroblast-like structures (α-SMA + fibroblasts) and a significant increase in tissue stiffness, following infarction (Richards et al., 2020). To conclude, this model provides a valuable opportunity to investigate the mechanisms of fibrosis and calcium fluxes, and it may also serve as a model for exploring druggable targets in ischemia. However, a significant limitation of this model is that due to the distance to the center of the microtissue, even under normoxic conditions, oxygen levels do not exceed 6%, limiting effective reperfusion and potentially affecting the study of reperfusion injury.

As an indirect co-culture containing both iPSC-CMs and human umbilical vein endothelial cells (HUVECs), Sebastião et al. (2020) developed cardiac spheroids, measuring approximately 260 μm, which were cultured for 18 days. The cardiac hiPSC-CM spheroids were subjected to a 5-h-long ischemic protocol, comprising both nutrient and oxygen deprivation with low pH and increased lactic acid supplementation, resulting in sarcomere disturbances and a reduction in viability in the spheroid core. The angiogenic potential of HUVECs was enhanced when stimulated with conditioned media from IRI spheroids, as opposed to hiPSC-CM control media. The present study reveals the indirect, angiogenic effect of the IRI spheroid secretome on the vasculature.

In intestinal organoids, IRI has successfully been modeled to recapitulate properties of in vivo IRI responses while reaching a larger size between 200 μm and 400 μm (Kip et al., 2021). One of the defining characteristics of intestinal organoids is their formation of a lumen-enclosed structure, which allows for more effective oxygen differentiation. Although not yet deployed in IRI, the development of cardioids, a specific type of organoid that self-organizes into chamber-like structures (Hofbauer et al., 2021b), could offer the potential to work with scaffold-free models while maintaining proper perfusion.

Recent advantages in the synergic integration of the flow dynamics of the heart-on-a-chip model, together with the self-organizing capacity of cardiac organoids, revealed great advancements over static models and could facilitate the use of larger, more stable and more complex organoids. In a kidney organoid model with fluidic culturing, transcriptomic analysis revealed 229 signaling pathways not identified in the static model (Hiratsuka et al., 2024). Min et al. (2024) revealed the impact of flow EHTs comprising hiPSC-CMs, ECs, and CFs within a 3D heart extracellular matrix hydrogel using a microfluidic chip. Placing these large 1-mm EHTs in a chip system, providing them with nutrient flow, led to increased oxygen concentrations within the cardiac tissues and reduced expression of cleaved caspase-3 in the organoid, compared to no-flow or flowing the organoids in a plate. Functionally, the application of flow on a microfluidic chip resulted in an increase in sarcomere length, the contraction amplitude, both indicators of hiPSC-CM maturation, and a heart rate of 76.59 ± 15.7 beats per minute (BPM) within the frequency range of a healthy human heart. RNA sequencing revealed significant differences between static and plate-flowed EHTs, with an observed increase in genes related to heart development, extracellular matrix organization, and angiogenesis. The relative mRNA expression of CM (MYH7 and TNNT2), EC (VWF), and cFB markers (COL1A1 and PDGFRA) increased in the chip-flow model compared to the no- or plate flow models, revealing increased functional differentiation of different cell types (Min et al., 2024). Altogether, the combination of EHTs with chip-based flow enables the functional maturation of diverse cell types within organoids while simultaneously facilitating an increase in size up to 1 mm.

In conclusion, a meta-analysis of transcriptomic changes in human ischemic organoids revealed a strong resemblance to post-infarct murine hearts while still preserving key human-specific characteristics (Richards et al., 2020). The utility of organoids in IRI research is currently constrained by the absence of a vascular network, which is a crucial element in reperfusion studies. However, recent developments in organoid technology have led to the creation of organoids containing chamber-like structures, which have opened new avenues of research in the field of organoid-ischemia–reperfusion studies (Hofbauer et al., 2021b). Moreover, the integration of chip flow, intersecting scaffold-free self-organization and complexity with scaffold-based nutrient exchange (Min et al., 2024), presents promising opportunities for bridging the knowledge and treatment gap in IRI.

As the knowledge on maturation is rapidly evolving, mature hiPSC-CMs will come closer to but remain different from adult human CMs (Ahmed et al., 2020; Tu et al., 2018). Relative immaturity of hiPSC-CMs can lead to reduced translatability of IRI findings compared to more established animal models, potentially limiting their predictive accuracy for human outcomes.

Microtissues comprising hiPSC-CMs and cFBs demonstrated enhanced electrophysiology and contractility superior sarcomere structure and augmented mitochondrial respiration, compared to hiPSC-CMs alone (Giacomelli et al., 2020). In addition, the amplitude of intracellular calcium flux during the contraction–relaxation cycle of hiPSC-CMs increased when co-cultured with cFBs (King et al., 2022), indicating enhanced cardiomyocyte maturation. Both fibroblasts and the co-culture of hiPSC-CMs with ECs result in the increased maturation of hiPSC-CMs (Giacomelli et al., 2020; Abecasis et al., 2019; Gisone et al., 2022; Liu et al., 2021), resulting in more organized and longer sarcomeres than the single-culture controls (Gisone et al., 2022), a feature suggesting a more adult-like phenotype and often associated with improved contractile function (Skorska et al., 2022). Additionally, the co-culture resulted in the increased expression of proteins involved in the deposition of several extracellular matrix components such as collagens and fibronectin (Abecasis et al., 2019), increased the contractility and higher expression of the ventricular cardiomyocyte marker IRX4 (Liu et al., 2021), and reduced the beating frequency (King et al., 2022).

Both the integration of multiple cell types and functional heart-on-a-chip integration of electrical pacing capabilities (Schneider et al., 2022) or mechanical stretch/stress (Mozneb et al., 2024) can significantly increase hiPSC-CM maturation. A recently developed multi-cell heart-on-a-chip model using multi-cell-type hiPSC-CMs and HUVECs could incorporate both dynamic fluid flow (shear stress) and biomechanical cyclic stretch. The combination of maturation strategies led to an improvement in the functional and transcriptional maturity of hiPSC-CMs, an improvement in the alignment of hiPSC-ECs grown on a heart chip and the facilitation of the formation of a tube-like EC network (Mozneb et al., 2024). Although hiPSC-CMs have yet to reach the maturity or mitochondrial capacity of adult CMs (Wu et al., 2021), it becomes increasingly feasible to combine multiple maturation methods to optimize hiPSC-CMs for modeling IRI in this fast progressing field.

One limitation of current 3D IRI models is the absence of the immune component as an injury-response mediator. Recently, Ze Lu et al. not only developed a heart-on-a-chip system that incorporates HUVECs, CFs, and hiPSC-CMs but also tested the addition of an immune fraction in the form of human peripheral blood mononuclear cells (PBMCs) within the vascular channel that initiates migration through the system under the appropriate conditions. This high-throughput setup allows for the straightforward collection of flow data and the measurement of cardiac redouts, such as beating force (Ze Lu et al., 2024). The introduction of SARS-CoV-2 to the 3D system did not result in significant alterations in the secretion of the cytokines IL-6, IL-8, and MCP-1, beating force, or contraction slope, while the addition of PBMCs caused significant alterations in these readouts. Although it has not been tested in an IRI setting, the incorporation of an immune component could potentially enhance physiological relevance of the 3D culturing system.

It is important to recognize the differences in in vitro protocols modeling IRI as the methods for inducing both ischemia and reperfusion can vary significantly between studies. This variation highlights the complexity of accurately modeling IRI in vitro and underscores the need for careful consideration when comparing findings across different experimental approaches. With oxygen modulation varying between 0.1% (Ellis et al., 2022) and 10% (Richards et al., 2020) and the time subjected to ischemia or reperfusion varying between 1 and 24 h, even at similar oxygen concentrations (Yadid et al., 2020; Veldhuizen et al., 2022), the majority of studies seek to achieve a total cell death rate between 20% and 50%.

Although oxygen deprivation is frequently used to induce ischemia, other physiological variables, like the lack of nutrients and the accumulation of cellular waste products, can also be introduced into the models to induce IRI. For instance, the accumulation of lactic acid during ischemic conditions (20 mM sodium lactate) and the corresponding reduction in pH 6.4 to 6.8 (Sebastião et al., 2020), instead of the physiological range of 7.2–7.4, combined with nutrient and oxygen deprivation, can be utilized to induce IRI in spheroids (Sebastião et al., 2020) and EHTs (Yadid et al., 2020; Chen and Vunjak-Novakovic, 2019). Hidalgo et al. (2018) compared oxygen reduction only to oxygen reduction together with physiological changes in pH and glucose availability during the ischemic episode in mature H9-NCX1+ CMs. A reduction in pH to 6.2, in conjunction with a 2-h period of 0% oxygen and 0 mM glucose, resulted in an approximate 60% increase in the death of CMs in vitro, measured by LDH concentration, compared to CMs only treated with 0% oxygen (Hidalgo et al., 2018).

To date, caution is still warranted when comparing and interpreting different IRI results as protocols are optimized per model as the field develops. Rapid developments in the field of hiPSC-CM maturation, as well as further development and standardization of IRI models, are expected to improve the consistency and reliability of IRI modeling in vitro.

Increased sensitivity and opportunities in functional electrophysiological readouts offer great opportunities in IRI development. Integrated live oxygen sensor integration (Schneider et al., 2022) and the integration of patterned intra- and extracellular electrodes on a heart-on-a-chip model (Liu et al., 2020) allow for the measurement of extracellular beating frequency, spatial waveform propagation, and precise action potential measurement, even in an IRI environment (Liu et al., 2020).

Progression of readout techniques has facilitated further progress in this field; Gao et al. pioneered using a multi-omics approach in an IRI setting (Gao et al., 2023). By integrating bulk and single-nucleus RNA sequencing with metabolomics profiling of reperfused rat hearts at various time points post-MI, it was discovered that early reperfusion reduced myocardial IRI by preserving fatty acid metabolism, a process regulated by PPARα. Functionally, pretreatment with the PPARα agonist fenofibrate upregulated genes associated with fatty acid oxidation and TCA pathways and revealed significantly upregulated PPARα expression, indicating that fenofibrate maintains energy metabolism post-infarct. These results were confirmed in animal experiments showing smaller infarct size and reduced fibrosis compared to non-fenofibrate-treated controls (Gao et al., 2023). Although yet to be used in organoid models, spatial transcriptomics combined with single-cell RNA-seq in embryonic hearts has enabled the generation of a three-dimensional cellular map, used to investigate cardiac organoid physiology and behavior in depth (Asp et al., 2019). Omics technology is becoming increasingly extensive and more readily available (Babu and Snyder, 2023) as cellular interactions and specific metabolic changes can be mapped in more detail, paving the way for future scientific advancements and therapeutic strategies.