In Vitro Modeling of Interorgan Crosstalk: Multi-Organ-on-a-Chip for Studying Cardiovascular-Kidney-Metabolic Syndrome

Abstract

Cardiovascular-kidney-metabolic syndrome is a progressive disorder driven by perturbed interorgan crosstalk among adipose, liver, kidney, and heart, leading to multiorgan dysfunction. Capturing the complexity of human cardiovascular-kidney-metabolic syndrome pathophysiology using conventional models has been challenging. Multi-organ-on-a-chip platforms offer a versatile means to study underlying interorgan signaling at different stages of cardiovascular-kidney-metabolic syndrome and bolster clinical translation.

The interaction between metabolic risk factors, chronic kidney disease (CKD), and cardiovascular disease (CVD) prompted the American Heart Association to establish a cardiovascular-kidney-metabolic (CKM) syndrome staging construct aimed at promoting cardio-kidney-metabolic health.¹ CKM syndrome is a progressive and systemic disorder characterized by aberrant interorgan communication networks driving several feedback loops leading to multiorgan dysfunction (Figure 1).² It is divided into stages 0 through 4, with stage 0 being individuals without CKM risk factors (ie, normal body mass index, waist circumference, fasting blood glucose, lipid profiles), and no evidence of CKD or CVD. Stage 1 is defined as individuals with excess or dysfunctional adiposity (ie, prediabetes) without other metabolic risk factors or CKD. Stage 2 is defined by metabolic risk factors (ie, hypertriglyceridemia, hypertension, metabolic syndrome [MetS], diabetes) or CKD. Stage 3 is defined by subclinical atherosclerotic CVD or heart failure [HF]) with excess or dysfunctional adiposity, metabolic risk factors, or CKD. Lastly, stage 4 is defined by clinical CVD, excess or dysfunctional adiposity, other CKM risk factors, or CKD. This stage can be further divided into stages 4a and 4b, which refer to those with and without kidney failure.¹ Notably, around 90% of adults in the United States meet the stage 1 CKM syndrome criteria or higher.³

Figure 1. Cardiovascular-kidney-metabolic (CKM) syndrome staging and pathophysiology.

CKM syndrome stages 0 through 4; clinical definition and prevalence (left). Interorgan crosstalk between the heart, kidney, and metabolic organs propagate CKM syndrome pathology through the exchange of key signaling molecules (depicted outside of the central blood vessel). Further research will help elucidate novel mediators of interorgan communication in CKM syndrome (within the depicted vessel) to guide the development of new therapies.^1,3 ASCVD indicates atherosclerotic cardiovascular disease; CKD, chronic kidney disease; CVD, cardiovascular disease; EVs, extracellular vesicles; FA, fatty acid; IL-6, interluekin-6; MetS, metabolic syndrome; RAAS, renin-angiotensin-aldosterone system; and TNFα, tumor necrosis factor alpha. Figure created with BioRender.com.

The strong linkage between cardiac, renal, and metabolic diseases (ie, CKM syndrome) suggests a shared pathological basis.⁴ However, major research efforts are needed to elucidate the interorgan communication networks responsible for disease progression, with the goal of developing targeted therapeutics and alleviating the economic and individual burden of CKM syndrome.^1,5 Many successful CVD therapeutics, including those targeting the renin-angiotensin-aldosterone system or adrenergic system, modulate key interorgan signaling pathways, and highlight the importance of further elucidating these networks. Understanding complex systemic multiorgan diseases will require refinement to sophisticated human tissue-based multi-organ-on-a-chips (multi-OoCs) to complement existing approaches. In this review, we summarize our understanding of CKM syndrome, current modeling platforms, examine the basis of in vitro modeling of interorgan crosstalk, and explore how multi-OoCs can drive progress in the field.

ELUCIDATING THE INTERORGAN COMMUNICATION IN CKM SYNDROME

Pathophysiology of CKM Syndrome

Excess and dysfunctional adipose tissue is a defining feature of CKM syndrome, where the activation of inflammatory signaling pathways, including TNFα (tumor necrosis factor alpha) and IL (interleukin)-6, lead to insulin resistance, oxidative stress, lipolysis, free fatty acid (FA) release, and the infiltration of proinflammatory immune cells.⁶ Because of its critical role in regulating insulin sensitivity and action as an endocrine organ, adipose tissue dysfunction is considered a major contributor to metabolic disease.⁷ Altered secretion of adipose-derived signaling molecules, such as adipokines, cytokines, lipokines, and free FAs, drives inflammation and dysfunction in various tissues, contributing to the development of other metabolic risk factors (ie, MetS, hypertriglyceridemia, diabetes, hypertension), as well as increasing the risk of CKD and CVD (Figure 1).^8,9

Sustained adipose tissue dysfunction and accompanying systemic inflammatory signaling contribute to the progression of CKM syndrome.² Stage 2 CKM syndrome is the most prevalent among the 4 stages and represents a critical time point for intervention before the development of CVD.^3,10 MetS is a state of multiorgan dysfunction involving the dysregulation of tissues such as adipose, liver, and vasculature.⁴ In this context, dysfunctional adipose tissue mediates altered lipid fluxes and leads to the accumulation of lipids in the liver and insulin resistance.¹¹ Excess hepatic lipids can alter the release of bioactive molecules, such as hepatokines, that exert various effects on tissues throughout the body to influence global metabolism.¹² In addition, endothelial dysfunction is strongly associated with metabolic, cardiovascular, and kidney disorders, directly contributing to impaired organ function and signaling.^13,14

The relationship between the heart and kidney is apparent, with cardiovascular mortality accounting for 40% to 50% of deaths in patients with stage 4 or 5 CKD and a rapid decline in kidney function after the onset of advanced HF.^15,16 The causes of HF with reduced ejection fraction are relatively well defined, with around half of cases being secondary to ischemic heart disease.¹⁷ Notably, the more straightforward cause of HF with reduced ejection fraction has enabled several effective treatment strategies. In contrast, HF with preserved ejection fraction (HFpEF) is increasingly common and is strongly associated with an obesity-related phenotype in CKM syndrome.¹⁸ However, HFpEF is highly heterogeneous, presenting with a spectrum of subtypes driven by comorbid conditions such as obesity, diabetes, hypertension, aging, and CKD. Unlike HF with reduced ejection fraction, no single cause has been identified, making the development of effective therapies challenging.¹⁹

CHALLENGES OF MODELING CKM SYNDROME

Animal Models of CKM Syndrome

Although the CKM syndrome staging construct is novel, researchers have long investigated the cardiovascular and renal effects of obesity and metabolic dysfunction using animal models such as mice, rats, rabbits, pigs, dogs, and nonhuman primates.^20–23 Although widely adopted mouse models, including leptin receptor null mice and diet-induced models, display many hallmarks of MetS, species-specific metabolic differences in lipid transport, glucose disposal, fat distribution, and adipose biology between mice and humans limit direct translatability.^24–27 In addition, MetS-related cardiac and renal phenotypes often differ between mice and humans. For example, although MetS is a major risk factor for the development of coronary atherosclerosis in humans, most mouse models of MetS do not develop overt coronary atherosclerosis or thrombosis. As a result, ischemic heart disease is typically modeled using coronary ligation or ischemia-reperfusion protocols, which do not fully capture human disease progression.²⁸ Similarly, many mouse models of MetS fail to develop advanced hypertension or renal pathology.^24,29

The resistance to end-organ damage is often overcome using multihit strategies that combine CKM syndrome-related stressors to enhance disease phenotypes. Multihit approaches commonly pair diet-induced obesity (eg, high-fat or high-fat/high-sugar diets) with additional factors such as aging, high salt intake, deoxycorticosterone, angiotensin II, nitric oxide synthase inhibitors, or unilateral nephrectomy to mimic comorbidities such as hypertension.³⁰ Although these strategies are used to model HFpEF, several widely used rodent models do not recapitulate the severe diastolic dysfunction observed in patients. Moreover, 3 widely accepted HFpEF models (eg, pig, rat, and mouse) show divergent myofibril mechanics,³¹ highlighting the limitations of relying on a single model.

In Vitro Models of CKM Syndrome

Complementary in vitro models for CKM syndrome-related pathologies have largely relied on conventional static culture in a 2-dimensional manner, where organ-specific cell types of interest (eg, cardiomyocytes, endothelial cells [ECs], hepatocytes, adipocytes) are monocultured or cocultured in the presence of CKM syndrome-related factors. However, in recognizing the importance of 3-dimensional (3D) microenvironmental interactions in maintaining cellular structure and function, various microphysiological systems including organoids, spheroids, assembloids, engineered tissues, and organ-on-a-chips (OoCs) have emerged to more accurately capture the dynamic 3D interactions occurring within organ systems and improve disease modeling (Figure 2).^32–36

Figure 2. Conceptualization and assembly of single-organ models.

Devices are constructed using a variety of cell sources ranging from primary cells and cell lines to human induced pluripotent stem cell–derived organ progenitors, which are combined in optimized ratios within hydrogels or engineered matrices to form functional tissues. Cells can be matured within the device using mechanical stimulation (eg, through pulsatile flow or cyclic stretch) or electrical pacing. Integrated analytical components, such as microelectrodes, electrochemical sensors, optical sensors, and strain/impedance sensors are used to provide real-time readouts of tissue performance. Platform biofabrication is achieved using diverse techniques such as microfluidics, micropatterning, injection molding, soft lithography, and 3-dimensional (3D) bioprinting. Compartmentalization of cells is accomplished using permeable membranes, with surface topography and nanofabrication techniques used to guide cell alignment. Figure created with BioRender.com.

Although animal models and conventional cell culture methods have been important in the study of CKM syndrome, species-specific differences, variability in disease pathology, and the importance of microenvironmental interactions highlight the challenges of recapitulating human pathophysiology. Moreover, these shortcomings are reflected by the failure rates of clinical trials, which are as high as 80%.³⁷ In that way, there is a growing need for complementary approaches to elucidate the mechanisms of CKM syndrome, including human tissue-based OoC models, which have the unique capacity to deeply interrogate multiorgan interaction.

VERSATILITY OF OoC TECHNOLOGY

Although organoids and other 3D self-organizing structures capture the cellular heterogeneity and cytoarchitectural attributes of the corresponding organ, these model systems do not incorporate analytical components to elucidate tissue functional properties. It is also difficult to uniformly apply controlled biophysical stimuli to organoids. In contrast, OoC platforms are designed to reproduce 3D tissue-like architecture, allowing for the application of defined electromechanical stimuli and simultaneous collection of physiological readouts in the same platform. Operating on a scale similar to organoids, OoC devices generally consist of cells and their associated matrices in a microfabricated platform (Table 1).

Table 1.

Basic Components of OoC Devices

Model components	Heart-on-a-chip models	Kidney-on-a-chip models	Liver-on-a-chip models	Adipose-on-a-chip models
Cell sources	hiPSC-derived CMs38 hESC-derived CMs39 Human cardiac fibroblasts40 Endothelial cells (eg, HUVECs, cardiac ECs, hiPSC-ECs)41 Cardiac resident macrophages42 Epicardial cells43	Glomerular endothelial cells44 Proximal tubule epithelial cells45,46 Vascular ECs47 Podocytes44 Mesangial cells48 HK-2,49 MDCK50 cell lines hiPSC-derived renal cells51,52	Primary hepatocytes53 Liver sinusoidal ECs54 Kupffer cells54 Hepatic stellate cells54 HepG2,55 HepaRG56 cell lines hiPSC-derived hepatic cells57	Primary mature adipocytes58 Tissue-resident immune cells59,60 Stromovascular cells61 Microvascular endothelial cells61 Adipocyte-derived stem cells62
Biochemical stimulation	Small molecule-based modulation of signaling pathways (Wnt/p-catenin, Hippo, GSK-3P)63 Fatty acid-based maturation media64 Growth factor supplementation (IGF-1, NRG1)65 O2 gradient66 Cardiovascular medications67	Retinoic acid68 Activin A69 Small molecule-based modulation of signaling pathways (Wnt, FGF, BMP, TGF-P)69 Extracellular vesicles70 Nephrotoxic compounds71	Small molecule-based modulation of signaling pathways (FGF, BMP, HGF, Oncostatin M)72 O2 gradient54 Glucose, glucagon Palmitic acid and oleic acid56 Hepatotoxic drugs, ethanol53	Proinflammatory cytokines (TNF-a, IL-6) Anti-inflammatory cytokines (IL-10 and IL-4)60 Insulin, glucose73 Dexamethasone, PDE inhibitors, indomethacin74,75 Isoproterenol to stimulate lipolysis58 PBMC-derived monocytes73
Electromechanical conditioning	External pacing76 Intermittent or continuous perfusion77 Pneumatic actuators for cyclic strain78 Suspension microvalves for unidirectional flow79 Flow constrictors for elevated afterload79 Cell patterning with topographical cues80	Bidirectional, unidirectional, or countercurrentflowthrough microfluidic channels81 Vacuum pumps to apply cyclic stretch and simulate pulsatile flow51	Micropatterned surfaces82 Sinusoid-mimetic spatial zonation54,55 Diffusion microchannels55 Pumpand gravity-driven flow55 Flow-guided perfusable angiogenesis54	Flow through perfusable media channels58 Micropores, micropillars, or semi-permeable membrane to separate convective and diffuse transport58,74
Analytical components	Microelectrodes/micropillars83 Elastic cantilevers84 Calcium imaging38 Field effect transistors85 Impedance or strain sensors84,86	Porous membranes/hydrogels may be ECM-coated87,88 Oxygen, temperature, pH sensors89 TEER sensors90 Pressure and flow sensors89 Biosensors for glucose, lactate, glutamine91 Perfused dye for permeability assessment92	Biosensors for albumin, urea, lactate, GST-α93 Fluorescent-tagged bile acid analogs94 CYP450 enzymatic probe94 substrates56,95 Oxygen sensors96	Fluorophore-tagged fatty acid transport58 Dil-coupled LDL uptake58 Biosensors for glucose, insulin73 Localized-surface plasmon resonance-based biosensors for multiplexed cytokine profiling60

When designing heart-on-a-chip, kidney-on-a-chip, liver-on-a-chip, and adipose-on-a-chip organ modules, one must consider cell sources, biochemical stimulation, electromechanical conditioning, and incorporation of analytical components. BMP indicates bone morphogenic protein; CM, cardiomyocte; CYP450, cytochrome P450; EC, endothelial cell; ECM, extracellular matrix; FGF, fibroblast growth factor; GSK-3β, glycogen synthase kinase 3 beta; GST-α, glutathione-transferase alpha; hESC, human embryonic stem cell; hiPSC, human induced pluripotent stem cell; HGF, hepatocyte growth factor; HK-2, human kidney-2; HUVEC, human umbilical vein EC; IGF-1, insulin-like growth factor 1; IL, interleukin; LDL, low-density lipoprotein; MDCK, Madin-Darby canine kidney; NRG1, neuregulin 1; OoC, organ-on-a-chip; PBMC, peripheral blood mononuclear cell; PDE, phosphodiesterase; TEER, transepithelial/transendothelial electrical resistance; TGF-β, transforming growth factor beta; and TNF, tumor necrosis factor.

To be considered an OoC, the system must possess 3 components (1) 3D structure and spatial arrangement, (2) multiple cell types reflective of the physiological diversity within an organ, and (3) incorporation of relevant biomechanical forces and readouts.⁹⁷ Designs with open configurations facilitate ease of cell seeding, tissue manipulation, and fluid sampling via automated equipment. In contrast, sealed systems are pump-driven and superior in controlling flow dynamics and molecule delivery.

Heart-on-a-Chip

Heart-on-a-chip systems commonly produce a cylindrical structure mimicking trabeculae anchored at 2 opposing ends on 2 posts or wires to align cells and drive tissue maturation. The deflection of the post or the wire enables the measurement of contractile properties.^38,76,98 Cylindrical cardiac structures can also be engineered and confined in a microfluidic channel.^{64,67,99–103} Systems such as microfluidic miniPUMP recapitulate 3D ventricular architecture and incorporate elevated afterload and on-chip valves for unidirectional flow to replicate exposure to physiological conditions and measurement of ejection fraction.⁷⁹ Because adult human cardiomyocytes are terminally differentiated, these platforms use human induced pluripotent stem cell–derived cardiomyocytes (hiPSC-CMs) with supporting cells in an exogenous ECM (extracellular matrix) to enable tissue compaction and CM alignment. Strategies such as patterned cell seeding, dynamic loading, electrical stimulation, and metabolic maturation via culture media can be used to mature cardiomyocytes in these platforms.^64,76,104 Cardiac OoC platforms can be applied to the study of ischemic, structural, and congenital heart disease, as well as inheritable arrhythmias, cardiomyopathies, and HF.^{41,67,84,105,106}

Liver-on-a-Chip

Biomimetic multilayered liver-on-a-chip have also recently been fabricated to replicate the 3D microstructure of hepatic lobules.^54,55,107 A network of hexagonal tissue-culture chambers and universal feeding ports enabled flow circulation between layers to mimic the function of the central vein in liver lobules, directing the formation of self-assembled perfusable hepatic sinusoids from seeded hepatocytes under micropillar guidance.⁵⁴ These devices also incorporate perfusion inlets and outlets to replicate the portal veins and hepatic arteries, along with a regulating chip to ensure physiologically accurate oxygen concentrations.⁵⁵ Platform fidelity can be evaluated via functionality assays that examine biomarkers such as albumin excretion and urea synthesis.¹⁰⁷

Kidney-on-a-Chip

Kidney-on-a-chip models similarly hold promise in advancing our understanding of renal pathology, discovering therapeutics for CKDs, and predicting drug nephrotoxicity.¹⁰⁸ Although OoC models have been developed that attempt to independently replicate the physiology of the glomerulus,^44,51 proximal tubule,^45,46 and distal tubule,¹⁰⁹ the integration of these components towards the development of a true kidney-on-a-chip platform has yet to be achieved. Moreover, the structure and function of different parts of the nephron, along with the countercurrent exchange process, rely on the coordination of various cell types (like glomerular ECs and podocytes) within specific, organized environments, and on adjusting fluid flow to balance electrochemical and osmotic pressures¹¹⁰

Adipose-on-a-Chip

Compartmentalized microfluidic systems that enable the coculture of primary human adipocytes with immune cells and vascular cells have also been developed to generate adipose-on-a-chip systems for the study of inflammation in obesity and insulin resistance.^58–60 Porous barriers enable paracrine signaling via diffusion, with nutrient and oxygen delivery through perfused channels. These platforms support real-time FA uptake monitoring and multiplexed cytokine analysis via biosensors.

Controlled Microenvironments and Integrated Readouts

To advance the physiological relevance of multi-OoCs, these platforms must incorporate relevant cell populations, often derived from pluripotent stem cell sources, with precisely controlled electromechanical conditions and biochemical stimuli to form functional biomimetic microenvironments (Table 1). For example, heart-on-a-chip models use FA-based maturation media and intensity training regimens of electrical pacing to drive accelerated cardiomyocyte maturation, with subsequent quantification of the tissue’s contractile function, calcium handling, and response to β-adrenergic agonists. Kidney-on-a-chip systems incorporate multiple cell types (proximal tubule epithelial cells, glomerular ECs, podocytes) with countercurrent flow through microfluidic channels with porous membranes, along with TEER (transepithelial/ transendothelial electrical resistance) sensors and perfusion of dyes such as dextran for permeability assessment. Structurally complex liver-on-a-chip models use micropatterned surfaces and indentations that mimic sinusoids, with the introduction of hepatotoxic compounds and tuning of media oxygenation to replicate in vivo differences between portal venous and hepatic arterial blood. The performance of these platforms can be analyzed using fluorescent-tagged bile acid derivatives, assays for albumin and urea synthesis, and CYP450 (cytochrome P450) enzyme probe substrates. Finally, adipose-on-a-chip platforms incorporate tissue-resident immune cells and stromovascular cells along with insulin, glucose, and pro/anti-inflammatory cytokines to create a biomimetic microenvironment, with fluorophore-tagged FAs to evaluate lipogenesis/lipolysis and incorporated biosensors for cytokine profiling.

Biophysical forces must also be controlled. Shear stress in the endothelium-lined microchannels connecting organ modules should be in the physiological range of 0.1 to 5 Pa. Mean intraluminal pressures should range from 70 to 100 mm Hg to mimic those observed in healthy arterial vasculature and replicate physiological cardiac afterload, with incorporation of both cyclic circumferential (5%–10%) and axial (0%–20%) tensile strain. Hepatocytes similarly experience shear stress due to blood flow, with a reported value of 0.1 to 0.5 dynes/cm² at the sinusoid. Glomerular ECs, podocytes, and proximal tubule cells are all highly sensitive to deviations in shear stress, with the normal physiological set point ranging from 1 to 2 dyn/cm². Precise flow control is achieved via pumps, pressure controllers, and optimized channel designs, with microvalves and resistors enabling compartmentalization and mimicking vascular resistance.

Finally, integrated pressure and flow sensors allow real-time feedback and adaptive regulation. Multiorgan systems can, therefore, incorporate the cyclic strain and pulsatile pressure essential for myocardial maturation with the low shear conditions required for sinusoid mimicry and the segment-specific shear stresses present along the nephron. Moreover, they can alter these biomechanical cues to replicate pathologies, such as hypertension, atherosclerosis, and microvascular disease observed in CKM syndrome.

BASIS OF MULTI-OoC MODELING

Isolated OoCs are constrained due to the absence of a broader biological context involving communication between different organ systems. Conceptually, multi-OoCs aim to address this limitation by coupling multiple OoCs (Figure 3). Two major strategies for linking multiple OoCs (multi-OoC) are the modular and the combined chip approaches.¹¹¹ In the modular approach, each tissue chip is assembled individually and later linked via vascular-mimicking channels. This approach allows for the incorporation of organ-specific vasculature or tissue maturation.^111,112 For example, hiPSC-CMs incorporated into engineered heart tissues can be electromechanically matured via pacing before linking them into multi-OOC.¹¹¹

Figure 3. Approaches to modeling cardiovascular-kidney-metabolic (CKM) syndrome using a multi-organ-on-a-chip (multi-OoC).

Multi-organ systems incorporating cardiac, kidney, liver, pancreas, and adipose organ modules connected via endothelium-lined microchannels can be used for systems-level modeling of complex interorgan signaling and soluble drug responses in CKM syndrome. Two major strategies for linking multiple multi-OoCs are the combined and the modular chip approaches—in the former, organ modules are developed in parallel in a common ‘blood mimetic’ medium, offering advantages with regard to ease of assembly. For the modular approach, each tissue chip is developed individually and later linked via microchannels with recirculating flow, allowing for incorporation of organ-specific vasculature and tissue features. Multiple organ modules can be obtained via differentiation of patient-specific human induced pluripotent stem cells (hiPSCs). Endothelium-lined vascular channels are fabricated using a porous membrane that allows for exchange of signaling factors, drugs, and their metabolites; more sophisticated models incorporate continuous or period injection of immune cells. Pumps, mechanical actuators, and electrical field stimulation can be used for systemic or modular exposure to physiological stimuli, with tissue-specific readouts quantified using various sensors. Evolving responses can be evaluated using functional assays, immunohistochemical analysis, transcriptomics/proteomics, and cytokine profiling. ECM indicates extracellular matrix. Figure created with BioRender.com.

In contrast, the combined multi-OoC devices are smaller and typically have higher throughput, use less media, and may offer advantages regarding ease of assembly.¹¹¹ This approach is better suited for drug screening due to its simpler design and less complex tissue composition. For example, an organ may be represented by only a few key cell types, making assembly and experimentation more streamlined. However, the fixed design of the chip may limit the ability to address specific questions.

MULTI-OoCs FOR ELUCIDATING INTERORGAN COMMUNICATION IN CHRONIC DISEASE

Establishing an in vitro model of CKM syndrome will be crucial for the elucidation of pathological mechanisms and for testing potential therapeutics. In this section, we will highlight recent studies where multi-OoCs are being developed to study disease pathophysiology, focusing on tissues relevant to CKM syndrome, such as the heart, kidney, liver, adipose, and immune system. We then discuss the prospects of combining models to study the complex interactions between multiple tissues in chronic conditions such as CKM syndrome.

Modeling Immune Interactions

Chronic low-grade inflammation is strongly associated with cardiovascular, kidney, and metabolic disorders, where it is hypothesized to contribute to disease pathophysiology.¹¹³ One approach is to mimic the function of circulatory immune cells by directly incorporating them into a circulating medium. For example, human Tohoku Hospital Pediatrics-1 cells were introduced into the recirculating medium of a gravity-driven multi-OoC device containing hiPSC-CMs, primary hepatocytes, and primary skeletal muscle. Using a cantilever deflection and patterned custom microelectrode array, this system was able to emulate a targeted immune response to amiodarone-induced heart tissue damage, which was characterized by infiltration of Tohoku Hospital Pediatrics-1 cells and impaired contractile force, reduced conduction velocity, and beat frequency in the cardiomyocytes without affecting the liver (CYP activity) or skeletal muscle (contractile force). In contrast, treatment with lipopolysaccharide and IFNγ (interferon gamma) triggered a broad inflammatory response with damage manifesting in all organ compartments.¹¹⁴

Another study incorporated human CD (cluster of differentiation) 14⁺ monocytes into the vascular compartment of their multi-OoC containing engineered and matured human liver, cardiac, bone, and skin tissue niches derived from hiPSCs and primary stromal cells. Notably, each tissue niche was maintained in its optimized environment, linked via a recirculating vascular flow and selectively permeable endothelial barrier. The monocyte’s locality and phenotypes were profiled after 4 weeks of culture and in response to cryoinjury inflicted upon the heart tissue.¹¹² After 4 weeks of culture, the monocytes retained a CD14⁺/CD16⁻ status, whereas the monocytes in the platform lacking an endothelial layer transitioned to a more proinflammatory state. Similarly, after cryoinjury, the monocytes gained a proinflammatory phenotype and homed in on the cardiac chamber (site of injury), suggesting that this system could also model inflammatory responses to injury.

Another approach is to generate a bone marrow niche chip, containing the precursors required to produce immune cells and allowing them to egress into the circulating vascular compartment upon stimulation.¹¹⁵ One study aimed to assess the effect of acute and protracted exposure to cosmic radiation on cardiac, liver, and immune tissues within a multi-OoC containing a selectively permeable endothelial barrier separating each tissue module. Engineered cardiac tissues and liver spheroids were generated from hiPSC-CMs or hiPSC-derived hepatocytes and primary human fibroblasts. Engineered bone marrow was created using decellularized trabecular bone block-derived scaffolds seeded with iPSC-derived mesenchymal stem cells, which were differentiated into osteoblasts. After osteogenic maturation, human umbilical vein ECs, iPSC-derived mesenchymal stem cells, and CD43+ cord blood-derived hematopoietic/stem cell progenitors were introduced into the bone tissue in a fibrin hydrogel.¹¹⁶ After irradiation, the heart module displayed altered troponin release and contractile phenotypes, whereas the liver displayed altered albumin and urea secretion. Further analysis revealed altered cytokine levels in the vascular compartment medium and significant population shifts in the bone marrow compartment characterized by myeloid skewing. Notably, the bone marrow niche model could recapitulate immune cell mobilization into circulation.

Overall, these reports demonstrate the feasibility of modeling the inflammatory responses within multi-OoC models, focusing mainly on innate immune cell responses, which are critically involved in disease pathology. However, the immune system is composed of innate and adaptive components, the latter of which also have important roles in disease. Modeling adaptive immune responses would require additional lymphoid tissue niches to investigate key processes such as T-cell activation and maturation. Further research is needed to generate a truly immunocompetent OoC model and capture the complexities of the immune system.¹¹⁷

Cardiorenal Axis

The reciprocal relationship between the cardiac and renal systems is apparent, with hemodynamic factors playing a major role. However, the nonhemodynamic interaction between the heart and kidney in CKM syndrome remains poorly defined. Although there are models for different structures of the human kidney, multi-OoCs containing renal and other tissues are scarce.¹⁰⁸ One study linked multiple organ chips generated using primary human cells (heart, lung, intestine, liver, kidney, skin, blood-brain-barrier, and brain) by vascular perfusion, where an endothelial barrier facilitated the separation of vascular and parenchymal compartments. Notably, the kidney tissue maintained its functionality (eg, albumin reabsorption) over extended culture, which is important for modeling chronic diseases such as CKM syndrome.¹¹⁸ Another study aimed to develop a model to decipher different modes of interorgan communication between the heart and kidney by generating a microfluidic system containing iPSC-derived cardiac microtissues and kidney organoids that maintained viability and function for 72 hours.¹¹⁹ Interestingly, they also reported a lower contractile amplitude in their cardiac microtissues when cocultured with the kidney organoids on their microfluidic chip, which may indicate crosstalk between the tissues.

Metabolic Tissue Interactions

Metabolic dysregulation is a central component of CKM syndrome, which is heavily intertwined with systemic inflammation and organ dysfunction. Metabolic homeostasis is maintained through constant communication between organs such as the liver, adipose, and pancreas. To investigate these aspects, several multi-OoC approaches have been taken to model interorgan crosstalk in metabolic disorders such as metabolic dysfunction-associated steatotic liver disease, diabetes, and obesity.¹²⁰

Zandi Shafagh et al¹²¹ investigated the liver-pancreas axis during prediabetic hyperglycemia by generating a multi-OoC containing primary human liver spheroids and islets. Their system recapitulated tissue-typical organ functions such as insulin secretion and hepatic insulin responses after a glucose challenge. In addition, the investigators used bulk RNA sequencing to reveal counter-regulatory gene expression signatures in the pancreas and liver tissues after high glucose treatment.

Tao et al¹²² developed a multi-OoC containing hiPSC-derived liver and pancreatic islet organoids to model the liver-pancreatic islet axis. They found that the system can be cultured for up to 30 days with synergistic effects of the coculture on liver and islet organoid functions (ie, albumin and insulin production) while maintaining their relevant genetic profiles (ie, RNA sequencing). Importantly, the investigators demonstrated the glucose-responsiveness of the system and recapitulated a type 2 diabetes-like phenotype characterized by altered metabolism, which was restored by metformin.

Similarly, Aleman et al¹²³ coupled a vascularized liver acinus and pancreatic islet model generated from human Tohoku Hospital Pediatrics-1 cells, primary hepatocytes, liver sinusoidal ECs, hepatic stellate cells, and islets to elucidate the impact of lipid-induced hepatic insulin resistance on islet dysfunction. By simulating early MetS culture conditions, they found that islet dysfunction was dependent on the presence of the liver acinus module, which displayed many of the hallmarks of metabolic dysfunction-associated steatotic liver disease.

To study the influence of adipose inflammation on hepatic metabolism and insulin resistance, Qi et al¹²⁴ developed a multi-OoC comprised of hiPSC-derived adipocytes and hepatocytes. When cultured with the adipose module, the hepatic module demonstrated enhanced functionality (ie, albumin and urea production) compared with isolated culture conditions. In addition, the investigators demonstrated that their system could recapitulate hepatic responses to glucagon (eg, gluconeogenesis) and insulin (eg, glucose uptake and metabolic gene expression). They also demonstrated that the introduction of hiPSC-derived proinflammatory macrophages promoted adipose tissue inflammation and subsequent hepatic lipid accumulation and insulin resistance. At the same time, a glucagon-like peptide-1 receptor agonist improved hepatocyte function indirectly by acting on the adipose module.

Slaughter et al¹²⁵ also studied the adipose-liver axis by generating a multi-OoC containing primary human hepatocytes and adipocytes, which were cultured in a recirculating serum-free medium over 14 days. Their adipose-liver multi-OoC recapitulated several aspects of nonalcoholic fatty liver disease, such as insulin resistance biomarkers, altered adipokines, and inflammation-induced steatosis. Importantly, they demonstrated that TNFα-induced steatosis was dependent on the presence of the adipose module, suggesting an indirect mechanism involving the crosstalk between adipose and liver, potentially through altered adipokine profiles and elevated lipolysis.

UNIQUE REQUIREMENTS FOR MULTI-OoCs IN CKM SYNDROME MODELING

Modeling CKM Syndrome Using Multi-OoCs

Multi-OoC systems offer a promising in vitro platform for dissecting the complex interorgan signaling mechanisms in CKM syndrome and assessing potential therapeutic strategies. To our knowledge, no multi-OoC model has been developed to investigate CKM syndrome specifically. However, recent advances in multi-OoC-based modeling of specific CKM syndrome features suggest that such a model could be feasible.

Modular multi-OoC platforms offer advantages over fixed tissue-chip designs by allowing greater experimental control and flexibility in combining relevant tissues. One recent advancement uses a vascular endothelial barrier to separate individual parenchymal and vascular compartments.^112,118 This new design eliminates the need for a common parenchymal media while enabling tissues to maintain organotypic functions and accounting for the critical role of the endothelium in mediating the exchange of signaling factors and immune cells between the blood and tissues.^112,118 In that way, modular platforms containing endothelial barriers are particularly suited for modeling interorgan crosstalk in vitro. Furthermore, endothelial dysfunction, which is common in metabolic, cardiovascular, and kidney diseases, represents a crucial interaction with model.

Stage-Specific Multi-OoC Approach to Modeling CKM Syndrome

Stage 1 CKM syndrome is defined by the presence of excess/dysfunctional adipose tissue, an early feature of metabolic disease. The consequences of adipose tissue dysfunction on distal organs remain underexplored but will be crucial for understanding the early pathogenesis of CKM syndrome. A stage 1 multi-OoC could be established by inducing adipose tissue dysfunction through the introduction of proinflammatory mediators, such as IL-6 and TNFα, or immune cells such as M1-polarized macrophages.¹²⁴ In addition, the number of adipose tissue modules could be increased to account for increased adipose volume in obesity. This approach could be useful in determining the paracrine-mediated effects of dysfunctional adipose tissue on cardiac, renal, and hepatic tissues.

Stage 2 CKM syndrome is defined by the presence of excess/dysfunctional adipose tissue in combination with additional metabolic risk factors such as hypertriglyceridemia, hypertension, MetS, diabetes, or CKD. This stage encompasses the transition from isolated adipose tissue dysfunction to overt MetS, an important driver of CKM syndrome progression. A stage 2 multi-OoC could be constructed by inducing adipose tissue dysfunction (as described in stage 1) in tandem with hepatic dysfunction while incorporating additional metabolic stressors such as elevated or oscillating glucose, altered insulin concentrations to model diabetes, free FA to model dyslipidemia and lipotoxicity; adrenergic agonists to simulate sympathetic activation; and renin-angiotensin-aldosterone system activators to mimic hypertension and fibrotic signals (Figure 4). This approach could be useful for studying the progression of CKM syndrome from stage 2 to stage 3, including the onset of renal dysfunction and CKD. Established CKD could further be simulated by introducing uremic toxins, such as uric acid and other metabolic byproducts.

Figure 4. Establishing a multi-organ-on-a-chip (multi-OoC) model of stage 2 cardiovascular-kidney-metabolic (CKM) syndrome.

(1) Individual OoCs for heart, kidney, liver, and adipose are assembled. (2) Adipose and liver dysfunction are induced by pathological inflammation while cardiac and renal tissues are cultured in normal medium. (3) All tissues are linked into a multi-OoC and further stimulated with MetS factors (3.1) or normal medium (3.2). (4) Each tissue module can be analyzed using on-chip or off-chip approaches to decipher the impact of metabolic tissue dysfunction on the multi-OoC system. Omics analysis will reveal key signaling pathways involved. EV indicates extracellular vesicles; FFA, free fatty acid; MetS, metabolic syndrome; and RAAS, renin-angiotension-aldosterone system. Figure created with BioRender.com.

Stage 3 CKM syndrome is defined by the presence of subclinical atherosclerotic CVD or HF along with excess/dysfunctional adipose tissue, other metabolic risk factors, or CKD. This stage represents the transition from metabolic stress to early organ dysfunction. A stage 3 multi-OoC could be generated by prolonging the duration of stage 2 culture conditions to simulate chronic metabolic stress. Alternatively, higher concentrations of MetS factors (as described in stage 2) could be used to accelerate tissue dysfunction.

Stage 4 CKM syndrome is defined by clinical CVD with excess/dysfunctional adiposity, and other metabolic risk factors or CKD, with stages 4a and 4b representing cases with or without kidney failure respectively. This stage marks the transition from subclinical to clinical disease, often involving MI, HF, or significant renal impairments. A stage 4 multi-OoC could be established by extending stage 3 culture conditions along with heightened MetS-related stress. Alternatively, cardiac-targeted cryoinjury or hypoxia/reoxygenation could simulate MI and model HF with reduced ejection fraction. In contrast, HFpEF could be modeled through prolonging stage 3 culture conditions in combination with additional inflammatory factors, such as IL-1β and IFNγ.¹²⁶

Analyzing a CKM Syndrome Multi-OoC

Understanding the interactions between organ systems in a CKM syndrome multi-OoC will require a combination of real-time and end point assessments (Figure 4). Although real-time assessments offer insights into the OoC function without major perturbations to the multi-OoC, they are often technologically challenging and require advanced microfabrication techniques and sophisticated OoC designs. Real-time measurements can be achieved by incorporating biosensors and microscopy windows that provide longitudinal function assessment. In contrast, end point assessments are typically performed after the tissue module or medium is extracted from the multi-OoC and can include similar methodologies used for single-OoCs.

Considering the wide range of cardiac phenotypes observed in CKM syndrome, methods to assess cardiac tissue function should be tailored to specific research interests and the type of cardiac tissue model being used. Heart failure is prevalent among CKM syndrome patients, making the assessment of contractile and relaxation properties an essential readout. 3D cardiac tissue constructs such as organoids and engineered heart tissues can be assessed using force transducers, strain sensors, or video microscopy techniques in conjunction with elastic cantilever micropillars.^127,128 Electrical abnormalities or arrhythmic phenotypes or can be assessed using EP methods, such as multielectrode arrays, field effect transistors, patch clamps, or calcium imaging techniques.^{85,126,129–131} One key advantage of incorporating a multielectrode array into an OoC is that it can provide real-time, longitudinal monitoring of cardiomyocyte electrical activity.^132,133 Alternatively, key end point assessments also include structural analysis for cardiomyocyte hypertrophy and fibrosis.

Impedance spectroscopy is another electrophysiology method frequently used to provide real-time assessment of barrier-forming cells such as the endothelium and epithelium. Altered endothelial barrier properties are a hallmark of endothelial dysfunction. Therefore, the incorporation of impedance electrodes will provide a robust readout of endothelial status over time.^133,134 Similarly, fluorescently labeled dextran can be introduced into the circulatory compartment, where extravasation into the tissue compartment can be quantified via fluorescence microscopy to assess the endothelial barrier integrity further.¹³⁵ Nitric oxide bioavailability and reactive oxygen species are critical indicators of endothelial function and can be measured using various fluorescent probes or electron paramagnetic resonance.^136–138

Immune cells play a major role in CKM syndrome by propagating inflammatory signaling and disrupting tissue function. The bone marrow niche module and the circulating immune cells can be profiled by flow cytometry to quantify the dynamic population changes.¹¹⁶ In addition, the extravasation of immune cells into each tissue compartment can be quantified using fluorescently labeled immune cells or via immunofluorescence. Immune cell responses (eg, homing in on injured tissue) could also be profiled in response to relevant stimuli.¹¹²

CKD is characterized by a progressive decline in renal function and depending on the anatomic region of the kidney being modeled, various functional outputs can be assessed. For example, the glomerular filtration barrier function can be evaluated using fluorescently labeled albumin, while renal tubular epithelial barrier integrity can be monitored over time using impedance spectroscopy or TEER sensors alongside ECM-coated porous membranes. Passive diffusion assays with fluorescently labeled molecules such as inulin or dextran can also be used to assess permeability, with pressure and flow sensors added if the platform incorporates perfusion. Similarly, the functional transport of substrates such as glucose, lactate, and glutamine can be assessed using biosensors.¹³⁹ Additional measures, such as albumin or glucose reabsorption, and end point analyses, including the quantification of apoptosis and fibrosis, can further provide insights into kidney function.^108,118,140

Adipose tissue dysfunction is characterized by enhanced inflammation, increased lipolysis, impaired insulin sensitivity, hypertrophy, fibrosis, and cell death. Dedicated tests can be used to assess each of these parameters. For instance, fluorescently labeled FA can be used to quantify the kinetics of lipid uptake. Lipolysis can be assessed by monitoring the release of fluorescent FA or glycerol into the medium. In addition, insulin and glucose challenges can be performed using standard protocols. Lastly, structural components can be assessed by standard staining procedures for lipid droplets, cell size, fibrosis, or immunofluorescence.^61,124 Lastly, biosensors can be incorporated into these organ modules for multiplexed cytokine profiling after exposure to immune cells or pro/anti-inflammatory stimuli.

In the context of CKM syndrome, hepatic dysfunction is characterized by impaired insulin sensitivity, increased lipid accumulation, inflammation, and eventually fibrosis. Insulin response, gluconeogenesis, and lipid accumulation will be key functional readouts for the hepatic module.¹⁴¹ The contributions of adipose tissue dysfunction on hepatic lipid accumulation can also be studied by loading the adipose tissue module with fluorescent FA and tracing their release and accumulation in the hepatic module.¹²⁴ In addition, assays to detect hepatic cytochrome p450 enzymatic activity, urea, albumin, and lactate production may also be relevant functional readouts.^118,142

Lastly, in-depth molecular profiling of each tissue module via transcriptomics, epigenomics, proteomics, lipidomics, and metabolomics will offer valuable insights into each tissue’s state (Figure 4). These data can be correlated to the functional readouts and integrated with medium analysis to yield a comprehensive snapshot of the multi-OoC state. Moreover, the molecular profiles can be compared with published primary human tissue datasets to further validate the CKM syndrome model.

Analyzing the Mediators of Interorgan Crosstalk in a CKM Syndrome Multi-OoC

To identify specific molecular targets of interorgan communication in CKM syndrome, the medium can be profiled for secreted proteins, EVs, and metabolites (Figure 4), which can be integrated with the omics data from each tissue to elucidate causal signaling networks. Other approaches to identifying specific mediators of interorgan crosstalk include various molecular labeling strategies. For example, Ronaldson-Bouchard et al used a green fluorescent protein labeled CD63 reporter to track cardiac-derived exosomal distribution within their multi-OoC.¹¹² In addition, proximity labeling strategies such as biotin identification (BioID) and turbo-powered proximity labeling (TurboID) could be used to trace the cellular sources of secreted factors when paired with proteomic analysis.¹⁴³

LIMITATIONS AND FUTURE DIRECTIONS

Whereas multiorgan systems have the potential to drive significant advancements, several challenges remain (Figure 5). The preparation of organ-specific cells and tissues, the need for expertise in microfabrication, and the associated costs and labor requirements present substantial hurdles. The inclusion of organ-specific stromal support and immune cells (eg, resident macrophages^42,144) is needed to capture emergent cellular behaviors. High-density vascularization required to grow miniature high-density tissue volumes and connect the compartments appropriately is missing. This advancement is required to solve the scaling issues¹⁴⁵ so that metabolic rates, media residence times, masses, and surface areas all fall into place perfectly in multi-organ systems. It is still unclear whether organ-specific ECs are essential for this process. Several studies report efficacy with ETV2 inducible ECs.¹⁴⁶

Figure 5. Emerging challenges in multi-organ-on-a-chip (multi-OoC) development.

Substantial hurdles include optimization of organ modules via incorporation of cellular heterogeneity and tissue vascularization, microfabricating appropriately scaled designs that recreate complex, physiologically relevant stimulation patterns, achieving platform integration and high-throughput screening using automation, and ensuring clinical relevance through benchmarking and in silico models. 3D indicates 3-dimensional; and PDMS, polydimethylsiloxane. Figure created with BioRender.com.

The integration of perfusable vascular networks into individual organ modules is critical for achieving functional maturation and adequate tissue size.¹⁴⁷ This can be achieved by either mixing ECs and supporting stromal cells with the primary organ cell type to self-assemble into prevascularized tissues, or by embedding tissues in a preestablished capillary bed fabricated within a hydrogel or polymeric material. Functional anastomoses form by the expansion and outgrowth of organ-derived vasculature into the surrounding matrix or by angiogenic sprouts from the surrounding microvascular network penetrating the tissue to connect with organ-derived vessels. These vascularized modules can also be incorporated into preestablished polymer conduits or microfluidic channels lined with ECs to achieve hierarchical branching and perfusable vascular networks. Although many models continue to rely on primary ECs or human umbilical vein ECs for vascularization, this introduces patient source variation and is associated with a tendency to form unstable capillary networks in vitro. An emerging alternative is hiPSC-derived ECs, which can provide a limitless supply of patient- and organ-specific vascular cells for multi-OoCs.

Future OoC fabrication will rely heavily on developing biocompatible, optically transparent materials and bio-inks that overcome the nonspecific small molecule absorption and high gas permeability classically associated with polydimethylsiloxane,¹⁴⁸ which challenge the control of physiological features and accuracy of drug screening. We must also develop platforms with standardized chip interfaces that accommodate the incorporation of multiple interchangeable modular organ subunits in a plug-and-play format to facilitate large-scale multi-OoC fabrication and device commercialization.

Physical stimuli are important for maturation, but they are often applied in a rhythmic, metronomic manner, unlike the chaotic and fractal patterns found in physiological systems. Moreover, critical pressure variations across biological barriers, essential for normal body function (eg, formation of glomerular slits), are frequently lacking. Future efforts should prioritize improving our ability to extract data from models—through scalable analytics—alongside developing more advanced models. Currently, functional readouts in most OoC devices often rely on semi-manual methods, where image or sample analysis occurs off-chip requiring advancements in automation.

Advances must be made in platform integration and scaling to capture tissue physiological responses over periods longer than a few days or weeks, particularly if we are to model therapeutic interventions in chronic disease states. This requires the incorporation of automated liquid handling technologies with high-throughput multidimensional screening via live cell imaging and real-time biochemical assays. Incorporating computer vision and deep learning approaches can streamline data collection and analysis, enabling parallel readouts and the identification of emerging trends. Scalable and automated production of multi-OoCs is essential to move beyond traditional screening methods using a limited number of well plates, which require transitioning from standard soft lithography and polydimethylsiloxane to plastics processing techniques such as hot embossing or injection molding.¹⁴⁸

As the throughput of multiorgan systems increases, managing flow becomes increasingly challenging. Microfluidic setups with open-well configurations offer greater scalability potential, as they eliminate the need for additional tubing and pump connections. However, the unphysiologically high media-to-cell ratio in such systems may dilute signaling molecules.

The clinical translation of multi-OoC predictions relies heavily on developing measurable standards that not only quantify the physiological features of healthy organs but also their deviations in disease phenotypes. Such measurements should be accessible in public databases to facilitate benchmarking with in vivo data and accelerate multi-OoC acceptance by regulatory agencies. One excellent example includes the IQ (International Consortium for Innovation and Quality) in Pharmaceutical Development and its affiliations like the IQ Consortium Drug Induced Liver Injury Initiative. The initial stages will focus heavily on benchmarking parameters that are associated with individual cell types (Table 2) but will eventually need to expand to encompass organ-level parameters as our models become more sophisticated. Importantly, these benchmarking parameters should be applied to both iPSC and primary cell–derived models.

Table 2.

Benchmarking Parameters for Heart-on-a-Chip, Liver-on-a-Chip, Kidney-on-a-Chip, and Adipose-on-a-Chip Organ Models

Model	Benchmarking parameters
Heart-on-a-chip	Elongated, rod-like cardiomyocytes149
	Expression of adult ventricular genes (MYH7, MYL2, and TNNI3)149
	Contractile stress 30–60 mN/mm3,150,151
	Positive force-frequency ratio, characteristic notch at the top of upstroke152
	RMP −90 mV, action potential amplitude 100–110 mV, action potential duration 230–300 ms152
	Depolarization velocity 250–300 V/s, conduction velocity 30–100 cm/s152
	Well-developed sarcoplasmic reticulum structure and function, gene expression of calcium-handling machinery149
	Cytosolic calcium concentration 1 μm during systole, 100 nmol/L during diastole153
	Reliance on fatty acid oxidation (>70%)154
Liver-on-a-chip	Albumin production >37 μg per day/1 million hepatocytes155
	Urea synthesis >56 μg per day/1 million hepatocytes155
	Expression of phase I and II CYP450 enzymes, hepatocyte uptake, and efflux transports155
	Polygonal, nonrounded, polarized hepatocytes155
	Presence of bile canaliculi (BSEP, MRP2, and AQP1) and bile acid synthesis155
	Presence of Kupffer cells (CD68) and stellate cells (desmin)155
Kidney-on-a-chip	Albumin reabsorption efficiency of at least 80%156
	Glucose reabsorption efficiency of at least 95%157
	Transepithelial electrical resistance 6–10 Ω·cm2 (proximal tubule),158 100–500 Ω·cm2 (renal cell lines)159
	Expression of podocyte markers (NPSH1, NPSH2, and WT1)160
	Expression of mesangial cell markers (GATA3, PDGFRA, and ITGA8)160
	Expression of glomerular endothelial cell markers (EHD3, EMCN, KDR, and CD31)160
	Brush border enzyme activity: Na+/K+-ATPase, alkaline phosphatase, γ-glutamyltransferase160
Adipose-on-a-chip	Adipocytes >55 μm in diameter161
	Unilocular lipid contents, expression of maturation markers (PPARG, FABP4, and adiponectin)162
	Evidence of insulin-stimulated glucose uptake162
	Evidence of lipid metabolism (free fatty acid uptake, glycerol release)162
	Incorporation of resident immune cells (CD68) and stromovascular cells (CD31, ACTA2)162

Parameters are reflective of values normally observed for mature cardiomyocytes, hepatocytes, proximal tubule epithelial cells/glomerular cells, and adipocytes. ACTA2 indicates alpha-smooth muscle actin; AQP1 indicates aquaporin 1; BSEP, bile salt export pump; CD, cluster of differentiation; CYP450, cytochrome P450; DES, desmin; EHD3, EH domain containing 3; EMCN, endomucin; FABP4, fatty acid-binding protein 4; GATA3, GATA-binding protein 3; ITGA8, integrin alpha 8; KDR, Kinase Insert Domain Receptor; MRP2, multidrug resistance-associated protein 2; MYH7, myosin heavy chain 7; MYL2, myosin light chain 2; NPSH1, nephrin; NPSH2, podocin; PDGFRA, platelet-derived growth factor receptor alpha; PPARG, peroxisome proliferator-activated receptor gamma; RMP, resting membrane potential; TNNI3, troponin I type 3; WT1, Wilms tumor-1; and WT, wild-type.

Although research involving hiPSCs is currently not at a stage where all benchmarking parameters can be established for mature cell populations, these values should act as guidelines for groups developing and modifying these platforms. Rather than hindering regulatory approval, these parameters could serve as 1 metric (among many others) for regulatory agencies when determining each device’s clinical relevance. Moreover, it would serve the multi-OoC community well to not only document benchmarking parameters for mature organs but also to assess cellular maturity for all hiPSC-derived cells and organoids used in organ modeling platforms. This would establish a much-needed range of acceptable parameters for each organ module, as has previously been done for hiPSC-CMs.¹⁵² In addition, the maturation of hiPSC-derived cell and organ models remains an active area of investigation and will be critical to fully realizing the potential of hiPSC technology in multi-OoC applications. Despite these limitations, hiPSC-based models offer a highly accessible, near-infinite supply of cells with patient-specific genetic backgrounds, paving the way for increasing regulatory acceptance in disease modeling and drug screening research.

We must also address cell line and batch-to-batch variability issues. Directed differentiation protocols have advanced significantly since iPSC generation in 2006.¹⁶³ On-chip differentiation in OoC systems adds complexity to the selection of culture media and matrices (eg, Matrigel), particularly in multi-OoCs. The use of defined media and small molecules, rather than growth factors, should become standard practice. In assembling body-on-a-chip systems with common media, secretion of factors by certain cell types that may harm or impair the function of others necessitates solutions for common culture media. In addition, understanding the impact of circadian rhythms on cell culture variability is required. Reporting the sex of cells and designing experiments with models of different sex characteristics are essential for capturing sex-specific drug responses and disease manifestations. This process can be expedited by deriving all cells in a model from an isogenic source.

With regards to CKM syndrome specifically, modeling the chronicity of this disease will be a significant impediment. CKM syndrome progresses over months to years, whereas multi-OoC platforms are typically optimized for short-to-intermediate culture durations on the order of days to weeks. The long-term modeling of progressive fibrosis, lipotoxicity, and organ failure in these systems remains technically challenging and may not fully recapitulate chronic disease states. Moreover, the intricate coordination of differing culture requirements, maturation timelines, mechanical cues, and biochemical stimuli, along with the integration of complex sensors, circulating immune cells, and endocrine signals, will make it difficult to maintain organotypic function under shared or modular perfusion conditions, capture systemic feedback loops, and perform continuous, noninvasive assessment of functional readouts in these systems.

CONCLUSIONS

Since its introduction in 2010, the OoC concept has rapidly advanced through interdisciplinary efforts in cell biology, bioengineering, and microfabrication.¹⁶⁴ These innovations have led to sophisticated human tissue-based in vitro models with the potential to transform biomedical research by reducing reliance on animal models while enhancing clinical translation. Although single-OoC systems have successfully recapitulated specific organotypic functions, enabling applications in drug screening and disease modeling, they cannot replicate the interorgan communication critical to understanding systemic physiology and pathology. To overcome this limitation, multi-OoC models have emerged as powerful tools for studying interorgan signaling mechanisms particularly relevant in complex disease states, such as CKM syndrome. These advanced platforms can provide insights into how interorgan communication promotes disease progression, deciphering poorly understood mechanisms and paving the way for identifying novel therapeutic targets. The rapid expansion of the multi-OoC field underscores its potential to transform drug development and disease modeling. However, continued research is needed to refine these systems and fully realize their potential for addressing the challenges of complex multiorgan disorders.

Nonstandard Abbreviations and Acronyms

3D

3-dimensional

CD

cluster of differentiation

CKD

chronic kidney disease

CKM

cardiovascular-kidney-metabolic

CVD

cardiovascular disease

ECM

extracellular matrix

EVs

extracellular vesicles

FA

fatty acids

HFpEF

heart failure with preserved ejection fraction

hiPSC

human induced pluripotent stem cell

hiPSC-CM

human induced pluripotent stem cell–derived cardiomyocytes

IFNγ

interferon gamma

IL

interleukin

IQ

International Consortium for Innovation and Quality

MetS

metabolic syndrome

Multi-OoC

multi-organ-on-a-chip

OoC

organ-on-a-chip

TNFα

tumor necrosis factor alpha