Hypoxia and Multilineage Communication in 3D Organoids for Human Disease Modeling

Organoids, three-dimensional (3D) multicellular microtissues derived from stem cells or tissues, have revolutionized the study of human biology and disease [1]. These self-organizing structures closely mimic the architecture and function of human organs, providing a more physiologically relevant model than traditional two-dimensional (2D) cell cultures or animal models [2].

Animal models afford robust platforms for investigating complex inter-organ interactions, particularly in systemic diseases, such as allergies and other non-infectious conditions, that involve multi-etiological factors [3]. However, the utilization of animal models is limited by substantial interspecies variations in gene expression and environmental influences, which restrict their ability to emulate human pathophysiology [4]. Notably, the regulation of gene expression, immune responses, and metabolic processes often diverges markedly between human and animal tissues [5,6]. To address these gaps, there is a pressing need for human-derived in vitro models, such as organoids or engineered tissues, that better replicate physiological and disease processes [7,8]. Organoids replicate vital processes of embryonic development and perform several organ-specific functions [9].

The advent of human pluripotent stem cells, particularly human embryonic stem cells (hESCs) and human-induced pluripotent stem cells (hiPSCs), has revolutionized research on human organ development and disease mechanisms [10,11]. A key advantage of hiPSCs lies in their ability to retain the genetic profile of the donor, making them a powerful in vitro model for examining critical cell–cell interactions and inter-organ communication under disease conditions [12,13]. Moreover, hiPSCs can be directed to differentiate into specific cell types, enabling the study of organ-level interactions using 2D co-culture systems [14]. The ongoing progress in biotechnology and tissue engineering has led to the development of advanced 3D organoid systems, including multilineage organoids, assembloids, and organoids-on-a-chip, that more accurately replicate the architecture and function of human tissues [15,16]. These innovations hold significant promise for generating complex, functional organoids from hiPSCs, enabling the investigation of inter-organ communication during development and disease [17]. In addition, cutting-edge omics tools, such as single-cell and spatial transcriptomics, allow for more detailed analysis of these diverse and intricate cellular systems [18,19].

In recent years, a diverse array of these tissue models has been developed, including organoids that represent the digestive organs [20], kidney [21], lung [22], brain [23], and liver [24].

Organoid models are beneficial for investigating the impacts of genetic differences or environmental factors on organ development, offering insights into the pathophysiological mechanisms of human diseases [7]. They are increasingly used in various research areas, including drug discovery, disease modeling, and personalized medicine [25].

A critical component of organoid development and function is multilineage communication, which refers to the dynamic interactions among various differentiated cell types within the organoid microenvironment [26]. These interactions orchestrate tissue patterning, maintain homeostasis, and drive functional maturation, closely mirroring the complex intercellular signaling found in native organs. The assimilation of multiple cell lineages, including epithelial, stromal, endothelial, and immune cells, enhances the physiological relevance of organoids, particularly for modeling disease processes that involve dynamic crosstalk between diverse cell types such as fibrosis, cancer, and inflammatory disorders [27,28]. The 2D differentiation methods, which begin with pluripotent stem cells, often fail to replicate the complex cellular interactions necessary for organ formation. In contrast, 3D organoid systems can generate more intricate, organ-mimicking tissue architectures [2]. However, the influence of diverse cell-type interactions (heterotypic interactions) on lineage specification remains insufficiently understood.

Physiological "normoxia" varies among embryonic and adult cell types, typically ranging from 2–9% O₂ (14.4–64.8 mm Hg), compared with 21% O₂ in the external atmosphere [29,30]. Specific tissues, such as the thymus, kidney medulla, and bone marrow, can function at ≤1% O₂ (7.2 mm Hg) owing to unique vascular architectures. In adult stem cell niches, which are typically located away from direct vascular supply, oxygen levels can drop to around 1% (7.2 mm Hg) [31,32,33]. One critical aspect of organoid development and function is the role of hypoxia, a condition of low oxygen levels. In organoids, hypoxia can influence cell differentiation, metabolism, and signaling pathways, affecting their overall development and function [34]. Hypoxia is a common feature of many diseases, including cancer, where it can promote tumor growth, invasion, and metastasis [35]. In obesity, adipocytes can grow to 150–200 μm in diameter, surpassing the ~100 μm oxygen diffusion range from nearby vessels, resulting in cellular hypoxia [36,37]. Low oxygen levels, or hypoxia, are not only characteristic of pathological states, such as ischemia, inflammation, and solid tumors [38], but are also present under physiological conditions, including during embryonic development and within certain adult tissues. These low-oxygen environments, often referred to as hypoxic niches, are crucial for guiding cellular differentiation [39,40]. Notably, stem cells typically inhabit these specialized anatomical regions, which play a vital role in regulating their function during development, as well as in tissue repair and maintenance.

Another critical factor in organoid biology is multilineage communication, the intricate interplay between different cell types within the organoid [41,42]. This communication is essential for maintaining tissue homeostasis, regulating cell differentiation, and coordinating responses to external stimuli. In disease models, the disruption of multilineage communication can contribute to disease progression and impact treatment responses.

The link between oxygen availability and mammalian embryogenesis was first recognized in the 1970s, when Morriss and New showed that neural fold formation in ex utero mouse embryos required low oxygen conditions to proceed normally [43]. Subsequent research has revealed the underlying molecular pathways involved in oxygen-mediated developmental regulation, particularly following the identification and characterization of hypoxia-inducible factors (HIFs), dimeric transcription factors that orchestrate cellular responses to reduced oxygen availability [44,45].

Studies using genetic models across various species have further emphasized the critical role of HIFs in development, as embryos lacking key HIF components frequently exhibit multiple developmental abnormalities [46,47]. These findings underscore the role of oxygen as a central regulator of embryogenesis. In this review, we examine how oxygen tension and HIFs influence both embryonic development and stem cell function, while also addressing alternative, HIF-independent pathways, such as the mTOR signaling axis, that contribute to hypoxia adaptation and embryonic growth.

This review article further explores the interplay between hypoxia and multilineage communication within 3D organoids and disease models. We discussed how hypoxia modulates organoid development and functionality, the role of multilineage interactions in maintaining tissue homeostasis and modeling disease, and how the convergence of these two factors shapes organoid biology, as presented in Figure 1.

Organoid technology represents a transformative advancement in biomedical research, enabling the formation of miniature, 3D organ-like structures from stem cells or tissue fragments in vitro [187]. These "mini-organs" mimic the architecture, cellular heterogeneity, and functional characteristics of their in vivo counterparts, thereby bridging the gap between traditional 2D cell cultures and animal models [188].

Organoids can be derived from various sources, including embryonic stem cells, induced pluripotent stem cells, adult stem cells, cancer cells, primary tissues, xenografts, and even mature cells [189,190,191]. Cancer-derived organoids, or tumoroids, are generated by dissociating tumor tissue into single cells or small clusters and embedding them in an extracellular matrix (e.g., Matrigel or defined hydrogel) under conditions that preserve tumor heterogeneity [192,193,194]. They replicate the architecture, histopathology, genetic and mutational profiles, and therapeutic responses of the original tumor. ECM is essential, as it provides structural support and biochemical signals that direct cellular self-organization, proliferation, and differentiation. Growth factors added to the culture media, such as EGF for epithelial organoids and fibroblast growth factor 10 (FGF10) for hepatic and gastric tumoroids, further guide organ-specific development [188]. The generation of hiPSC-derived organoids typically follows a stepwise differentiation process, guiding hPSCs into lineage-specific progenitors using stage-appropriate morphogens and growth factors that mimic developmental signaling [195]. These protocols aim to replicate human organogenesis, resulting in organoids that often resemble immature or fetal tissue rather than fully developed adult tissue, as seen in patient-derived organoids [196,197]. A defective differentiation strategy can lead to the presence of unintended cell types [198,199,200]. To address this, researchers benchmark hiPSC-derived organoids against fetal tissue, using it as an in vivo reference to refine developmental accuracy and cellular composition [42,198,201].

hPSC-derived organoids represent a rapidly advancing area of research, with broad applications in disease modeling and drug discovery. Owing to their ability to differentiate into virtually any cell type, hPSCs enable the generation of diverse organoid systems. Vandana et al. reviewed a wide range of these models, encompassing endoderm-derived organoids (lung, liver, intestine, pancreas), mesoderm-derived organoids (heart, kidney, skeletal muscle, blood vessels), and ectoderm-derived organoids (neural, retinal) [202]. This enables access to cell types that were previously challenging to isolate and culture, including the generation of complex structures such as brain and cardiac organoids [203]. Patient-derived cells can be reprogrammed into hPSC-derived organoids using specialized differentiation media and growth factors. These organoids serve as versatile platforms for drug screening, immune response studies [202,203] and investigations into genetic disorders and infectious diseases [204].

Organoids offer significant advantages. They recapitulate organ-specific microanatomy and cell interactions that are absent in monolayer cultures, thus providing high physiological relevance [205]. Furthermore, organoids retain the genetic, morphological, and functional characteristics of their tissue of origin, which makes them powerful models for studying development, disease mechanisms, and personalized therapies [206]. Organoids are highly versatile, having been established for a wide array of tissues, including the intestine, liver, brain, kidney, and tumors, and can be propagated in the long term or cryopreserved [7]. Genetic manipulation tools, such as CRISPR/Cas9, have been effectively used in organoid systems, enhancing disease modeling by enabling precise gene editing and simulation of disease progression [189,207].

Applications of organoids bridge diverse research domains, including developmental biology, regenerative medicine, toxicology, infectious diseases, and high-throughput drug screening [208,209]. Organoids also present notable advantages, including compatibility with comprehensive omics analyses, a small size that enables high-resolution 3D and even 4D imaging, and the capacity to generate large datasets suitable for training artificial intelligence models [210]. On the other hand, the use of organoids in high-throughput screening is hindered by some limitations. First, although their small size aids in high-quality images and precise machine learning assessment, their small size interferes with their ability to be scaled for in vivo or clinical transplant [211]. Second, organoids are static, which contradicts the dynamic nature of humans, as organoids lack blood vessels, immune cells, and microbes; their absence affects cell-to-cell interaction and scalability [212]. Moreover, in complex organs, like the brain, organoids might lack heterogeneity and reproducibility [213]. Third, the variation in organoid size, morphology, and maturation across batches disrupts assay consistency and hampers automation, while manual handling requirements and dependence on extracellular matrices, such as Matrigel, pose additional obstacles to scalable, fully automated workflows. Finally, high-content imaging generates complex datasets requiring advanced computational analysis; slow-growing or rare organoid types further restrict throughput, making rapid, large-scale screening challenging [214].

Patient-derived tumor organoids are increasingly used to investigate cancer heterogeneity and predict drug response, while kidney organoids serve as models for nephrotoxicity and congenital disorders [215]. Integration with organoid-on-a-chip platforms and microfluidics enables researchers to simulate inter-organ interactions and improve physiological relevance, thus moving closer to complex disease modeling and personalized therapeutics [216,217].

Recent innovations include the use of biomaterial scaffolds and microfluidic systems to mimic the physical and biochemical environment of native tissues, as well as the incorporation of single-cell sequencing and artificial intelligence for deeper phenotypic analysis and predictive modeling [217,218,219]. Despite these advances, organoid research still faces challenges in scalability, reproducibility, vascularization, and clinical standardization areas that continue to be intensively explored. Integrating organoid technology with tissue engineering has led to the creation of organoids-on-a-chip using microfluidic platforms [220]. These systems enable the construction of advanced, tissue-integrated biomimetic organoids that closely replicate human organ function and disease progression in vitro [221]. These humanoid organoid platforms facilitate both macroscopic and microscopic investigations of intercellular signaling and communication across tissues, paving the way for multi-organ models that faithfully replicate complex inter-organ interactions. These systems are proving valuable for studies in organogenesis, drug toxicity testing, and therapeutic efficacy evaluation [222].

Hypoxia, a condition characterized by reduced oxygen availability, plays a pivotal role in embryonic development and organogenesis. In the context of organoid cultures, hypoxic conditions have been shown to affect multilineage communication, thereby affecting organoid maturation and functionality [41].

Generally, oxygen tension in most tissues ranges between 2% and 9%, with an average of approximately 3% [29]. As a result, numerous studies have focused on how hypoxia influences various aspects of stem cell biology, including their proliferation, differentiation capabilities, and gene expression patterns [282,283,284].

During embryogenesis, physiological hypoxia represents a critical regulator of cellular differentiation and tissue patterning. Similarly, in organoid cultures, hypoxia has been observed to modulate interactions among various cell lineages, including epithelial, mesenchymal, and endothelial cells. These interactions are essential for the development of complex tissue structures and the establishment of functional vasculature within organoids [285,286].

Recently, several studies have demonstrated the role of hypoxia in promoting vascularization within organoids. For instance, co-culturing mesodermal progenitor cells with neural spheroids under hypoxic conditions led to the formation of vascularized organoids, demonstrating the importance of oxygen gradients in facilitating endothelial and mesenchymal cell interactions [287].

Furthermore, hypoxia has been shown to influence the differentiation pathways of various cell types within organoids. A research team demonstrated how oxygen tension shapes the differentiation of human iPSC-derived liver buds (hiPSC-LBs), with a focus on its modulation of TGFB-mediated intercellular signaling during hepatogenesis [288]. This study highlighted the role of controlled hypoxia in influencing early liver development and enhancing organoid maturation in vitro. Mild hypoxia (10% O₂) yielded the most significant and most numerous hiPSC-LBs, with enhanced albumin and vitronectin secretion, CYP3A4 activity, and urea production, indicating improved hepatoblast differentiation compared to excess hypoxia or ambient conditions [288]. This oxygen level also preserved stromal cells (HUVECs and MSCs) and suppressed TGFB1, TGFB3, and INHBA expression, fostering a pro-hepatocytic environment. TGFB1 and TGFB3 were predominantly produced by mesenchymal and endothelial cells in both fetal liver and hiPSC-LBs, identifying stromal cells as a primary TGFB source. Excess hypoxia upregulated TGFB family genes and cholangiocyte markers while reducing albumin secretion [288]. In sum, this report concluded that controlled oxygen delivery, particularly mild hypoxia, promotes stable hepatocyte differentiation in hiPSC-derived liver buds by modulating TGFB signaling from mesenchymal and endothelial cells. Both excessive TGFB activity under strong hypoxia and excessive suppression hinder hepatocyte maturation. Replicating physiological oxygen conditions alongside multicellular co-culture may enhance liver organoid maturation in vitro, supporting applications in regenerative medicine and drug screening.

A study examined hiPSC-LBs to compare 3D multilineage organoid interactions with conventional 2D differentiation [42]. In 2D, iPSCs were sequentially differentiated into hepatocyte-like cells (day 0–21), serving as a monoculture baseline. In contrast, 3D LBs were generated by co-culturing iPSC-derived hepatic endoderm with mesenchymal (MC) and endothelial cells (ED), which self-organized into tissue within 24 h.

Hepatic cells in LBs followed distinct lineage trajectories, showing intermediate maturation with epithelial migration signatures (e.g., PROX1, ONECUT2) similar to fetal hepatoblasts [42]. Transcriptomic analyses revealed that LB cells, ECs, and MCs closely resembled fetal liver lineages and engaged in extensive crosstalk, with MCs showing the strongest interactions. Pathway analyses identified TNF, FGF, JAK/STAT, NF-κB, HIF, and VEGF as key drivers; notably, VEGFR2 inhibition impaired endothelial sprouting and hepatoblast differentiation in 3D but not 2D systems.

LBs displayed hypoxic stress (HIF-1α upregulation) in vitro, which was rapidly resolved after transplantation into mice, as vascularization was established within 48 h, promoting hepatic maturation. Hypoxia activates multiple signaling pathways (TNF, FGF, JAK/STAT, NF-κB, HIF, VEGF) and induces extracellular matrix production along with remodeling enzymes (MMP2/4/19) [42]. Together, these changes shape the organoid microenvironment, promoting multilineage communication that is critical for hepatocyte maturation.

The interplay between hypoxia and multilineage communication is also evident in disease modeling. In cerebral organoids, chronic hypoxia was found to remodel the TME, supporting the expansion of glioma stem cells and highlighting the role of oxygen availability in tumor progression [289].

These findings underscore the significance of hypoxia as a modulator of multilineage interactions in organoid systems. By carefully controlling oxygen levels, researchers can enhance the physiological relevance of organoid models, thereby improving their utility in studying human development and disease.

Hypoxia is a crucial factor that impacts the survival and successful engraftment of islets following islet transplantation [290]. Islets, especially β-cells, have limited capacity to cope with low oxygen levels and oxidative stress due to their naturally low antioxidant enzyme expression [291]. Although the body initiates natural responses to promote neovascularization, the formation of new blood vessels typically takes around 7 to 14 days, during which a significant proportion of islets, particularly the larger ones, fail to survive [292]. The low oxygen availability at common implantation sites, such as the liver, further exacerbates this challenge. Encouragingly, several strategies aimed at promoting vascularization and enhancing islet viability, even within challenging sites, such as the subcutaneous space, have yielded promising results, demonstrating strong potential for clinical translation [293,294].

Liu et al. developed a bioelectronic heart-on-a-chip platform to study the impact of acute hypoxia on cardiac function [4]. This system integrates microfluidics with extracellular or intracellular bioelectronics, enabling rapid oxygen modulation to simulate coronary occlusion and reperfusion by switching between normoxic (21% O₂), hypoxic (1–4% O₂), and recovery (21% O₂) conditions. Using HL-1 mouse atrial cells, the model effectively reproduced normoxia, hypoxia, and early recovery phases. Hypoxia (1% O₂ for 2.5 h) induced strong nuclear HIF-1α expression, which normalized after 90 min of reoxygenation. Acute hypoxia (1% O₂ for 5 h) initially caused tachycardia (beating rate from 3.2 to 4.2 Hz), followed by arrhythmia with prolonged firing intervals (4.9 ± 1.5 s), mirroring responses in primary cardiomyocytes. This platform offers real-time, multiplexed readouts, making it valuable for studying hypoxia, ischemia, and reperfusion injuries, as well as for drug testing [4]. The authors propose future enhancements, including the incorporation of ischemia-related factors (e.g., acidosis, hyperkalemia), expanded multiplexing, stable long-term intracellular recordings, and adaptation for 3D cardiac tissues.

Hypoxia serves as a major driver of numerous cancer-related processes, including enhanced aerobic glucose metabolism [295,296], activation of angiogenesis via factors like VEGF and vascular endothelial growth factor receptor (VEGFR) [297,298], the initiation of EMT pathways [299,300,301], and the engagement of immune cells that support tumor growth such as myeloid cells [302], regulatory T cells [303], and tumor-associated macrophages [304]. Although hypoxia plays a crucial role in tumor progression, conventional 2D cell cultures fail to replicate the hypoxic conditions of the TME accurately [305]. In contrast, 3D models, including spheroids, organoids, 3D scaffolds, and microfluidic systems, have been shown to better mimic this key characteristic of the TME in numerous studies [306,307,308].

Zhi et al. developed a cortical organoid-on-a-chip model that mimics the physiological hypoxia of the human fetal brain to study neural differentiation and assess the effects of Tanshinone IIA (Tan IIA) [220]. The platform integrates a microfluidic chip with a micropillar array, enabling controlled formation of cortical organoids from hiPSCs. Featuring four independent chambers, the chip supports efficient drug screening. Embryoid bodies seeded onto the chips differentiated into millimeter-sized cortical organoids under specific culture conditions.

To simulate in vivo hypoxia, organoids were cultured in 5% O₂ for 30 or 50 days. The model successfully replicates the hypoxic microenvironment of early brain development, enhancing neurogenesis, synaptogenesis, and neuronal maturation compared to normoxic conditions [220]. Tanshinone IIA (Tan IIA), a traditional Chinese herbal medicine, further promoted neural progenitor differentiation, neuronal activity, and functional network formation. This organoid-on-a-chip system offers a powerful tool for studying human cortical development under hypoxia and for evaluating neurotherapeutics.

Hypoxia in the core regions of larger organoids, caused by the lack of functional vasculature, often leads to nutrient deprivation, necrosis, and diminished functionality [309]. To fully realize the therapeutic potential of these 3D organoid systems, it is essential to engineer sophisticated vascular networks within their structure [310].

In organoids, the development and branching of blood vessels are finely orchestrated by spatial gradients of pro-angiogenic factors, a complexity that is challenging to replicate within organoid systems [310]. Hypoxic conditions stimulate the formation of a VEGF gradient, where cells in low-oxygen regions produce higher levels of VEGF compared to well-oxygenated areas. Endothelial cells can sense these VEGF signals through VEGFR2 and initiate a Notch-dependent lateral inhibition mechanism. This signaling cascade designates specific cells with elevated VEGFR2 expression as tip cells to guide the sprouting process along the VEGF gradient, while adjacent stalk cells, characterized by increased VEGFR1 expression, contribute to vessel elongation [310].

A recent study developed the vascular network quality index (VNQI), a standardized and quantitative metric to evaluate the physiological relevance of vascularized microphysiological systems (vMPS) based on oxygen transport [311]. This study addresses the limitations of current user-dependent, non-standardized, and morphology-based assessments that poorly reflect oxygenation function. The researchers generated a dataset of 500 diverse vMPS samples, varying parameters such as endothelial cell source and density, fibroblast density, growth factor levels, and ECM stiffness [311]. Vascularized islet-chips with higher VNQI scores preserved and enhanced insulin secretion under hypoxic conditions, unlike poorly vascularized controls.

A chained neural network effectively derived the VNQI, which correlated strongly with actual oxygen transport. This metric offers a robust, standardized tool for evaluating vascularization in vMPS, bridging morphology and function through the application of machine learning.

Low-oxygen conditions have been demonstrated to contribute to the development of fetal organs and the placenta. However, it has also been found that hypoxic culture conditions, by themselves, are insufficient to promote progenitor cell expansion [288]. A recent study tackled the challenge of limited size and functionality in hiPSC-derived liver organoids, which has restricted their potential in regenerative medicine [312]. In this study, researchers explored progenitor expansion during mouse liver development to identify key factors that could enhance hiPSC-liver organoid growth. They focused on the interaction between the placenta and fetal liver, visualizing placenta-derived blood perfusion by injecting AlexaFluor647-conjugated lectin and anti-CD31 antibody through the umbilical vein. Using confocal microscopy, the researchers generated 3D images and quantified liver vasculature at embryonic stages E9.5, E10.5, and E11.5.

The researchers found that placenta-derived perfusion began at E10.5, primarily targeting the ventral lobes, which subsequently grew 9.1-fold by E11.5 compared to 3.7-fold for dorsal lobes [312]. Notably, despite active perfusion, the liver remained hypoxic at E10.5 and became oxygenated only by E11.5, when hepatoblasts transitioned to oxidative metabolism, highlighting the role of oxygen in growth post E11.5. In line with these findings, IL1α treatment under hypoxia significantly promoted the expansion of hiPSC-liver organoids and hepatoblasts (ALB⁺CK19⁺). The study effectively recreated the transient placenta–liver interaction by sequentially supplying placenta-derived IL1α under hypoxia, followed by oxygenation [312]. This approach, which mimics the physiological oxygen profile and IL-1α signaling, was identified as critical for optimal progenitor expansion and organoid development.

Kakni et al. developed an innovative in vitro model, hypoxia-tolerant apical-out human small intestinal organoids, to study host–microbiome interactions under physiologically relevant conditions [313]. This model addresses the limitations of existing gastrointestinal systems that fail to replicate the low-oxygen environment essential for the dominant anaerobic gut microbiota. Human intestinal organoids were efficiently differentiated under hypoxia (5% O₂), maintaining expected cellular composition and gene expression profiles for proliferation, differentiation, and mesenchymal markers, comparable to normoxic and Matrigel-embedded organoids. The organoids actively responded to hypoxia, with elevated HIF-1α protein expression and significant upregulation of HIF-1α target genes linked to mucosal barrier function (e.g., MUC2, MUC3A, TFF3, KRT20, SLC2A1). The hypoxic apical-out organoids exhibited intact epithelial barrier integrity and polarized nutrient absorption, effectively excluding 4 kDa FITC-dextran and absorbing fluorescent fatty acids from the medium [313]. This study presents the first human small intestinal organoid model with reversed polarity, capable of sustained culture and function under hypoxic conditions (5% O₂).

Kip et al. developed and characterized human small intestinal organoids (hSIOs) as a multicellular 3D model to investigate intestinal ischemia–reperfusion (IR) injury [314]. Unlike conventional 2D cell lines, hSIOs more accurately mimic the cellular complexity and architecture of the intestinal epithelium. The study differentiated hSIOs into crypt-like (CL) organoids with proliferative features and villus-like (VL) organoids enriched in differentiated enterocytes and goblet cells, emulating the distinct susceptibilities of crypts and villi to IR injury in vivo [314]. VL organoids showed pronounced cellular stress responses, with increased expression of oxidative stress markers (e.g., CHCHD2, ERO1A), pro-apoptotic proteins (e.g., DIABLO, LCN2, BCAP31), and complex I subunits linked to elevated ROS production. Enriched processes were selectively enriched in VL organoids, including the upregulation of basement membrane components, which is consistent with the early damage observed at villus tips during IR injury. Kim et al. developed 3D neural organoids derived from adult fibroblast–reprogrammed neural stem cells, which structurally mimic the human cerebral cortex and enable the modeling of hypoxic brain injury [315]. Following induced hypoxia and reoxygenation, the organoids showed restored neuronal proliferation but impaired maturation, highlighting their potential for neurodegenerative disease research and tailored drug evaluation.

Physiological hypoxia has been shown to enhance the growth and functional maturation of human intestinal epithelial organoids. Under oxygen levels of 2% compared to 20%, these cultures exhibited altered gene expression profiles associated with differentiation, metabolism, mucus barrier formation, and immune regulation [34]. Oxygen concentration significantly influences hPSC hepatic differentiation outcomes and should be regarded as an essential translational variable in scalable liver organoid systems for therapeutic and pharmacological use. Table 2 summarizes representative studies utilizing hypoxia-modulated organoid-on-chip platforms across various organ systems. These examples demonstrate how controlled low-oxygen environments have been integrated into advanced 3D microphysiological systems, including bioelectronic, microfluidic, and vascularized chips, to mimic in vivo physiology and investigate organ-specific responses. By applying defined hypoxic conditions, these models have provided new insights into cardiac function, neurodevelopment, and endocrine regulation, underscoring the importance of oxygen tension in shaping tissue maturation, function, and disease phenotypes.

Taken together, controlled hypoxia emerges as a double-edged regulator in organoid systems. On the one hand, insufficient oxygen can induce cellular stress, apoptosis, and impaired maturation, limiting organoid viability. On the other hand, carefully tuned physiological hypoxia enhances multilineage communication, stimulates vascularization, and drives tissue-specific differentiation, thereby increasing the physiological relevance of organoid models. By harnessing hypoxia-responsive signaling pathways and integrating oxygen-engineering strategies, researchers can overcome inherent limitations and generate advanced platforms for studying development, disease mechanisms, and therapeutic applications. Figure 6 illustrates this concept, highlighting both the detrimental consequences of excessive hypoxia and the beneficial outcomes of controlled oxygen regulation, underscoring its pivotal role in engineering functional and clinically relevant organoids.

Current research is now being directed towards the use of 3D models as an application for in vitro modeling, to enhance cellular interaction, aiming for interactions similar to those observed in vivo models [316,317]. The 3D models include spheroids, organoids, or hydrogels [318]. To obtain an optimum 3D model, optimization of both the 3D matrix and the oxygen concentration is crucial [318]. Although well established in 2D models, PO₂ in 3D systems remains not fully understood, due to the abnormalities and variety in cellular structure thickness that impede gas passage and generate oxygen concentration gradient, as reported [319].

Egger et al. presented a high-throughput cell-encapsulated hydrogel system made of polyether ketone that helped to establish a hypoxic/anoxic environment [318]. Standard hydrogel crosslinking methods, such as GelMA, are known to consume oxygen during the process [319,320]. This challenge can be addressed by using oxygen-permeable biomaterials, which enable the controlled and sustained release of oxygen to enhance tissue repair efficacy [321]. Oxygen release can be achieved either through oxygen generation, such as by employing photosynthetic algae [321,322,323], or by using peroxides [321,324]. To ensure sustained oxygen delivery and prevent cytotoxicity from rapid peroxide decomposition, the encapsulation of peroxide is required [324]. Among these peroxides is the micro-sized calcium peroxide, which was loaded on polylactic acid and embedded into the silk fibroin and oxidized pectin hydrogel, where its effect lasted for 14 days in bone models [325].

Another study by Tehrani et al. fabricated a chitosan/β-glycerol phosphate thermosensitive hydrogel enclosing polylactic acid as a capsule to enclose hydrogen peroxide, along with polyvinylpyrrolidone; it was coated with a layer of decellularized human amniotic membrane for wound healing that sustained oxygen release for a week [326]. Nevertheless, the use of peroxides still has the drawbacks of the optimizations required to maintain oxygen release, as well as bio-incompatibility and toxicity that can result from the fabricated hydrogels [321]. In addition to oxygen generation, oxygen release can be accomplished via the oxygen-carrying system that involves the loading of oxygen on hydrogels and their release to the desired site by the integration of hemoglobin [327,328] or perfluorocarbons [329,330,331] into the hydrogels.

The 3D organoid model provides a physiologically relevant platform for replicating HIF-1α–regulated metabolic processes in hepatocellular carcinoma (HCC). Unlike chemical induction with cobalt chloride (CoCl₂), this model naturally recreated hypoxic conditions and the associated shift to HIF-1α–dependent glycolysis [332].

During organoid development, hypoxia favored the selection of oxygen-tolerant cell populations, which may underline the malignant characteristics of pancreatic ductal adenocarcinoma (PDAC) [333]. This phenomenon provides a valuable platform for investigating disease mechanisms and identifying novel therapeutic strategies. Physiological hypoxia enhanced both the growth and functional maturation of human intestinal epithelial organoids. Specifically, lowering oxygen levels from 20% to 2% in colon-derived differentiated intestinal organoids altered gene expression profiles associated with differentiation, metabolism, mucus barrier formation, and immune regulation, thereby providing a valuable platform for disease modeling and personalized drug testing [34].

Colonoids cultured at 2% oxygen developed greater cell mass than those at 20%. Growth was similar until day six, after which differentiated colonoids showed enhanced expansion at 2% oxygen. In both oxygen conditions, colonoids proliferated and differentiated into goblet, absorptive, and enteroendocrine cells [34]. Undifferentiated colonoids expressed higher levels of the proliferation marker Ki67, while differentiated colonoids had elevated MUC2 (goblet) and CK20 (absorptive) expressions, with a trend toward more CGA-positive enteroendocrine cells. Protein-level expressions of Ki67, MUC2, CK20, and CGA did not differ significantly between oxygen levels, indicating that oxygen tension does not substantially alter intestinal epithelial cell composition [34]. Collectively, physioxia (2% oxygen) is preferable for colonoid research when closely replicating in vivo conditions. Then, human colonoids can be cultured from single cells in continuous low oxygen, often performing better than in 20% oxygen.

A recent study examined how oxygen concentration during the initial establishment of 3D pancreatic cancer organoids (PCOs) influences their properties, aiming to minimize selection bias from conventional normoxic culture and more accurately model the hypoxic tumor microenvironment in vivo [333]. The researchers utilized human pancreatic ductal adenocarcinoma (PDAC) surgical specimens to generate 3D organoids. The same tissue samples were digested into single cells and cultured from the outset under normoxia (20% O₂, producing NORMO-PCOs) or hypoxia (1% O₂, producing HYPO-PCOs) [333]. To evaluate phenotype stability, NORMO-PCOs were later switched to 1% O₂ and HYPO-PCOs to 20% O₂ for three passages before further analysis.

Both normoxic and hypoxic pancreatic cancer organoids were successfully established and maintained for over 20 passages. NORMO-PCOs displayed glandular morphology, while HYPO-PCOs formed solid structures. The source PDAC tissue contained both moderately and poorly differentiated regions; NORMO-PCOs reflected the well-differentiated areas with strong E-cadherin, absent vimentin, and high GATA6, whereas HYPO-PCOs mirrored the poorly differentiated regions with weak E-cadherin, high vimentin, and minimal GATA6 [333]. These phenotypic patterns were reproduced in xenografts, where NORMO-PCOs formed adenoductal structures and HYPO-PCOs produced undifferentiated regions, indicating oxygen-driven cell type selection. Transcriptomic profiling revealed distinct signatures: NORMO-PCOs were enriched for Moffitt classical signaling, and HYPO-PCOs for epithelial–mesenchymal transition (EMT) and Moffitt basal-like signaling [333]. HYPO-PCOs expressed higher VIM, S100A2, and TP63, and lower CDH1 and GATA6, with PurIST scoring confirming classical and basal-like subtype classifications for NORMO- and HYPO-PCOs, respectively.

An organoid system generated from human glioblastomas successfully replicated the hypoxic gradients and cancer stem cell heterogeneity seen in vivo, providing a platform to study both microenvironmental dynamics and cancer stem cell (CSC) biology simultaneously [306]. Physiological hypoxia enhanced vascular development in iPSC-derived kidney organoids. Culturing under 7% O₂ increased average vessel length, which correlated with upregulation of VEGFA-189 and VEGFA-121 and suppression of the anti-angiogenic isoform VEGF-A165b [334].

In prevascularized human cardiac organoids, stabilization of HIF-α represents an effective strategy to enhance endothelial cell expression and lumen development, providing a strong foundation for advancing prevascularization approaches in cardiac and other tissue transplantation therapies [335]. Furthermore, HIF-1α activation promoted vascular network formation in kidney organoids and conferred protection against cisplatin-induced injury under hypoxic conditions [336].

In iPSC-derived kidney organoids, hypoxia promoted vascularization by extending vessel length, elevating VEGFA-189 and VEGFA-121, and downregulating VEGF-A165b. This underscored the role of hypoxic culture in improving organoid vascular networks [334]. Moreover, lung tumor organoids exhibited hypoxic conditions that modulated the expression of lung cancer markers, including thyroid transcription factor-1, cytokeratin 7, and the ΔNP63 variant, as well as the upregulation of stem cell-associated markers, resulting in an increased proportion of cancer stem cells. Furthermore, these exhibited unique migratory behavior upon reoxygenation, driven by epithelial-mesenchymal transition and the elevated expression of matrix metalloproteinases 7 and matrix metalloproteinases 9, indicating their enhanced invasive potential [337]. However, in breast cancer 3D models, both hypoxic and normoxic conditions produced cellular phenotypes closely resembling those grown on 3D-printed scaffolds, whereas cells cultured in 2D or within Matrigel displayed distinct characteristics independent of oxygen levels [338].

In summary, controlled hypoxia is a pivotal factor in replicating in vivo physiology within organoid systems. Advances in biomaterial-based oxygen delivery now enable precise regulation of oxygen levels, ensuring a balance between cell survival, maturation, and functional performance. As summarized in Figure 7, the engineering of hypoxic microenvironments provides a powerful strategy to bridge the gap between in vitro and in vivo systems, laying the foundation for more accurate models in regenerative medicine, disease research, and therapeutic discovery.

Multilineage organoids are offering a platform that helps to mimic organogenesis during embryonic development, opening doors for their use as disease models in studies [339]. Among the most studied multilineage organoids were liver organoids [340]. A study performed by Takebe et al. developed a robust protocol for generating interconnected hepatic, biliary, and pancreatic (HBP) structures from hPSCs by fusing anterior and posterior gut spheroids [341]. The resulting multi-organ model mimics in vivo organogenesis and offers a promising platform for studying development, disease, and regenerative therapies. Guan et al. engineered an autosomal recessive polycystic kidney disease (ARPKD) mutated multilineage liver organoid to study hepatic fibrosis, showing elevation in collagen production and offering a fibrotic pathology, offering a model that can be used for further fibrosis drug screening [232]. Multilineage live organoids have attracted increasing attention from researchers as powerful models for studying disease. Goulart et al. proved that multilineage liver organoids produce more albumin and metabolic enzymes, offering another model with which to study different pathways of liver development [342].

Camp et al. advanced multilineage differentiation by employing single-cell RNA sequencing before organoid formation, demonstrating that the resulting model exhibits a gene expression profile comparable to that of the developing fetal liver [42]. With a special interest in studying fetal organ development, Branco et al. created a multilineage pro-epicardium, which gives rise to the outermost layer of the heart [343]. In this study, the pro-epicardium organoids were co-cultured with cardiomyocytes, giving rise to a functional heart organoid to help further study the crosstalk between myocardium and epicardium in health and disease [343].

Voges et al. demonstrated that human cardiac organoids (hCOs), which mimic immature heart tissue, can fully recover function after injury without fibrosis or hypertrophy, indicating a regenerative capacity similar to that seen in neonatal hearts [344]. Using a cryoinjury model, they showed that the injured cardiac tissue could achieve complete functional recovery within 14 days, highlighting the intrinsic regenerative potential of the immature heart.

Researchers developed heart-forming organoids (HFOs) from human pluripotent stem cells using Matrigel embedding and WNT pathway modulation, resulting in complex structures that mimic early heart development [345]. These organoids not only model key cardiac architecture and tissue interactions but also enable the in vitro analysis of genetic defects, as shown by NKX2.5 knockout-induced malformations. Another studied organ was the lung, where Miller et al. recruited hiPSCs to form both human lung organoids resembling the bronchi and bud tip progenitor organoids [346]. Both types of respiratory tract organoids proved to be representative of human fetal-like tissue, offering a model with which to investigate cellular crosstalk and epithelial fate throughout lung development [346]. de Carvalho et al. presented a 3D collagen I culture system that enables multilineage maturation of hPSC-derived lung progenitors into both proximal and distal airway cells by modulating GSK3, NOTCH, and WNT signaling [347,348]. The findings reveal key mechanisms that drive lung cell differentiation and provide a platform for advancing human lung development models.

Peng et al. reported that the overexpression or stabilization of HIF-1α enhances vascularization in kidney organoids derived from human iPSCs and enhances protection against cisplatin-induced injury under hypoxic conditions [336]. By improving endothelial cell development and activating antioxidant pathways, HIF-1α supports both organoid maturation and their potential for disease modeling. These findings highlight HIF-1α as a promising target for enhancing the physiological relevance and therapeutic utility of kidney organoids.

Using hypoxia-induced differentiation, Lim et al. established kidney organoids with interconnected collecting ducts and nephrons, enhancing structural and functional maturity [349]. One of the diseases where hypoxia plays a primary role is cancer. Establishing hypoxic organoids enhances the replication of TME. Ziółkowska-Suchanek reviewed the influence of hypoxia on lung organoids, highlighting current model limitations, particularly inadequate surface diffusion [350]. Dauer et al. also reviewed the impact of hypoxia on cancer and concluded that, in pancreatic adenocarcinoma, hypoxia is more closely associated with chemoresistance than with proliferation or angiogenesis, underscoring its relevance in chemotherapy drug screening [351].

A research report introduced a scaffold-free 3D cartilage regeneration approach using cartilaginous organoid bio-assembly (COBA) [352]. The method consistently achieved efficient production, structural integration, and functional restoration. Mechanistic analysis revealed that cell–cell adhesion proteins drive aggregation and cytoskeletal reorganization during organoid formation, while hypoxia within dense aggregates activates HIF-1α–mediated glycolysis to sustain the chondrogenic phenotype and enhance extracellular matrix deposition. In vivo, separated cartilaginous organoids fused into continuous cartilage, enabling applications, such as minimally invasive nasal dorsum augmentation, in goats [352]. Although a proof of concept, these findings highlight COBA's potential for clinical translation in plastic and reconstructive surgery.

This study introduces a 3D multicellular neurovascular unit (NVU) organoid that models hypoxia- and inflammation-induced blood–brain barrier (BBB) dysfunction, providing a physiologically relevant platform for neurological disease research, including ischemic stroke [353]. The approach addresses the limitations of conventional and animal models, which often fail to replicate the human BBB structural and functional complexity. In this study, the researchers generated a human cell–based 3D neurovascular organoid model composed of six cell types (brain microvascular endothelial cells, pericytes, astrocytes, microglia, oligodendrocytes, and neurons), using a staged assembly process with hanging-drop culture. The cell ratios were precisely controlled (e.g., 30% endothelial cells, 15% pericytes, 15% astrocytes) [353].

To mimic ischemic stroke, organoids were exposed to 0.1% O₂ for 24 h in a hypoxic chamber and compared with normoxic controls. BBB function was assessed by tracking FITC-labeled IgG, albumin, and dextran permeability, while cell viability was measured via calcein AM and ethidium homodimer-1 staining. Hypoxic organoids demonstrated greater permeability to large molecules (IgG, albumin) than normoxic controls, indicating BBB compromise [353]. These results concluded that the multicellular 3D neurovascular unit organoid reliably models BBB dysfunction under hypoxic and neuroinflammatory conditions, offering a robust in vitro platform for disease research and therapeutic testing.

Organoid research offers a more physiologically relevant in vitro representation of human biology, with features such as hypoxia and multilineage differentiation, which are shown to influence cellular behavior and immune responses. These attributes enhance the fidelity of disease models and their utility in drug screening. However, challenges remain, particularly in scalability and vascularization.

Although organoids hold great promise for therapeutic and clinical applications, several limitations currently restrict their broader utility. Their small size is advantageous for high-resolution imaging and AI-driven analysis. However, it hinders translation into transplantable or clinically relevant constructs. Organoids, particularly those modeling complex tissues, such as the brain, often exhibit heterogeneity and batch-to-batch variability in size, architecture, and maturation, which impairs reproducibility and large-scale automation. Dependence on manual handling and biological matrices, such as Matrigel, further complicates workflow standardization. Moreover, high-content imaging generates computationally intensive datasets; some organoid types are slow-growing or difficult to establish, limiting throughput. Consequently, issues related to scalability, vascularization, and standardization remain significant barriers to the widespread incorporation of organoids into high-throughput biomedical platforms.

Hypoxia-modulated organoids development affords a robust translational potential in personalized medicine, drug development, and regenerative therapy. By reconstructing tissue-specific microenvironments, including physiologically relevant low-oxygen conditions, they offer more accurate platforms for predicting therapeutic responses. In precision oncology, patient-derived tumor organoids cultured under hypoxic conditions enable the pre-clinical testing of individualized drug regimens to optimize efficacy and minimize toxicity.

For drug discovery, incorporating hypoxic conditions improves the identification of compounds effective against resistant cell populations and better emulates clinical metabolic and molecular tumor profiles. In regenerative medicine, low oxygen-conditioned organoids enhance stemness, differentiation, and repair capacities, enabling the modeling of hypoxia-induced injury responses (e.g., brain and intestinal damage) and the development of tissue-specific therapeutic strategies. Overall, oxygen modulation improves the physiological relevance of organoid systems, enhancing their value for clinical translation.

This review highlights two critical microenvironmental factors, hypoxia and multilineage communication, which impact organoid behavior and their utility in disease modeling. Hypoxia, primarily mediated by HIFs, is not merely a passive condition of 3D culture but a potent regulator of cell fate, metabolism, and gene expression. Deliberate modulation of hypoxic conditions enables more accurate modeling of pathological states such as ischemia, cancer, and fibrotic diseases. Concurrently, multilineage communication is driven by complex intercellular signaling; ECM remodeling enhances the structural and functional complexity of organoids, allowing for a more faithful recapitulation of in vivo multicellular interactions, including immune dynamics and tissue regeneration.

Recent advancements in the engineering of hypoxic niches and co-culture systems have substantially improved the physiological and immunological relevance of organoid models. These innovations have broadened the translational potential of organoids in applications such as high-throughput drug screening, regenerative medicine, and precision therapeutics. Nevertheless, key challenges persist, including the need for standardized protocols, improved vascularization, and the incorporation of systemic physiological cues to fully replicate in vivo conditions.

Incorporating the dynamic, spatially controlled hypoxic gradients alongside programmable ECMs will allow for more precise modeling of tissue- and disease-specific microenvironments. In parallel, the integration of multi-omics technologies with computational tools, such as single-cell multi-omics, spatial transcriptomics, and machine learning, is poised to deepen our understanding of lineage commitment, intercellular communication, and the molecular mechanisms underlying disease. Additionally, the development of perfusable, vascularized organoids that integrate immune components will significantly enhance their physiological relevance, particularly for modeling systemic diseases and investigating immuno–oncology interactions. In the realm of precision medicine, patient-derived organoids combined with CRISPR-based gene editing and high-throughput drug screening hold great promise for advancing personalized therapies and predictive diagnostics.

Integrating patient-derived cells into these systems also holds great promise for advancing precision medicine by improving model specificity and translational relevance. Notably, precise control of oxygen concentrations during hypoxia induction is critical, as varying oxygen levels can have diverse effects, depending on the context of the disease.

In this review, we underline the importance of integrating hypoxia and multilineage differentiation in the development of advanced organoid models that more accurately recapitulate organogenesis and pathological processes. Further research is essential to elucidate the role of hypoxia across various disease models, which will be instrumental in the design of more effective therapeutic strategies.

In conclusion, harnessing the synergistic interplay between hypoxia and multilineage communication offers significant promise for next-generation organoid platforms, positioning them as transformative tools in disease modeling and therapeutic discovery. Achieving this vision will require sustained interdisciplinary collaboration and the continuous refinement of organoid technologies.