2.1 Advancements and limitations of brain organoids

The human brain comprises a diverse array of cell types and presents complex challenges, including intricate neural circuits, vascular circulation, and immune evasion.³⁷ Animal models and 2D cell cultures often fail to fully address these issues due to ethical concerns and limitations in model complexity.³⁸ However, brain organoids replicate several crucial aspects of early human brain development, encompassing molecular, cellular, structural, and functional dimensions.²⁶ In 2013, Lancaster et al.²⁴ pioneered the use of iPSC technology to construct brain organoids that exhibited key characteristics of human cortical development, including a distinct progenitor zone rich in radial glial stem cells. In 2018, Mansour et al.³⁹ successfully transplanted human brain organoids into mouse brains, enabling vascularization and functional integration. In 2020, Pellegrini et al.⁴⁰ advanced this work by creating human choroid plexus organoids with selective barriers and cerebrospinal fluid-like secretion in independent compartments. Most recently, in 2023, Schafer et al.⁴¹ constructed vascularized brain organoids containing microglia, effectively modeling the ongoing neuroinflammatory processes observed in the brains of autistic children.

Brain organoids have successfully recapitulated central nervous system viral infections and genetic brain diseases. For example, there is evidence that Zika virus infection causes a reduction in brain organoid size and a loss of surface folds.⁴² Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection in human brain organoids leads to neuron-neuron and neuron-glial cell fusion, resulting in cell death and synaptic loss.^{43, 44} In 2019, Velasco et al.⁴⁵ utilized four distinct protocols to produce 3D brain organoids, revealing that early brain region patterning initiates cell specialization and maturation; extrinsic factors also play crucial roles.⁴⁵ Additionally, serum-treated brain organoids recapitulate Alzheimer's disease-like pathology, inducing levels of β-secretase 1 and glycogen synthase kinase 3 α/β that lead to increased levels of Aβ and p-Tau.⁴⁶

Despite the extensive insights provided by brain organoids, future developments should address several key considerations. First, brain organoids offer a unique opportunity for scientific exploration of consciousness, underscoring the importance of further ethical research to assess potential risks associated with human brain organoids and human–animal chimeric organoids.^{47, 48} Second, although Trujillo et al.⁴⁹ found that cortical organoids can generate continuous electrical activity signals, questions remain concerning whether neurons in brain organoids can establish proper synaptic connections to form mature cortical circuits. Third, iPSC-derived brain organoids require further maturation, particularly through the inclusion of essential components (e.g., vascular endothelium, microglia, and the blood–brain barrier) that are critical for the creation of a fully functional brain model.⁵⁰

2.2 Advancements and limitations of retinal organoids

The retina is a critical structure in the eye responsible for processing photoreceptive information. The differentiation of retinal organoids mirrors the formation of optic vesicles and can include most retinal neuronal cell types, such as cone cells, ganglion cells, bipolar cells, horizontal cells, and amacrine cells.⁵¹ In 2011, Eiraku et al.⁵² reported the dynamic and autonomous formation of optic cups (retinal primordia) from 3D cultures of mouse embryonic stem cell aggregates. Nakano et al.⁵³ later discovered that optic cup structures could self-organize in hESC cultures; the neural retina formed a thick layer that spontaneously curved into an apical structure. In 2014, Zhong et al.⁵⁴ achieved advanced maturation of photoreceptors in hiPSC-derived retinal organoids, demonstrating the formation of outer segment disks and light sensitivity. In 2019, Achberger et al.⁵⁵ combined organoid and organ-on-chip technologies to create a sophisticated human multilayered tissue retinal platform. This model provided vessel-like perfusion; for the first time, it captured the interactions between in vitro matured photoreceptors and retinal pigment epithelium.

Retinal organoids are a valuable tool for studies of eye development and visual function, as well as the acquisition of retinal tissue. However, they continue to exhibit deficiencies that require improvement. For example, although mature retinal organoids develop photoreceptor outer segment-like structures containing opsins, they do not exhibit proper disk stacking and orientation.⁵⁶ Additionally, retinal organoids lack vasculature and immune-related cells; they are unable to replicate the foveal structure needed to study diseases such as macular degeneration.^{57, 58}

2.3 Advancements and limitations of kidney organoids

Kidney development is believed to originate from the interaction between two embryonic cell populations: ureteric bud and metanephric mesenchyme.⁵⁹ In 2014, Taguchi et al.⁶⁰ and Takasato et al.^{60, 61} successfully differentiated human iPSCs and hESCs into metanephric mesenchyme organoids and organoids containing nephron structures. Combes et al.⁶² utilized single-cell sequencing to compare kidney organoids with human fetal kidneys, revealing a high fidelity in nephron, stromal, and endothelial cell types. Relative to immortalized human podocyte lines, podocytes in kidney organoids demonstrated correct basolateral polarity and gene expression profiles, offering a more accurate representation of in vivo conditions.^{63, 64} Considering that the kidney is the primary organ responsible for urine concentration, drug metabolism, and toxin filtration, human kidney organoids have been used to simulate autophagy processes in tacrolimus-damaged renal cells, thereby revealing the mechanisms of drug-induced nephrotoxicity.⁵¹ Concerning pharmacological screening platforms that use organoids, Lawlor et al.⁶⁵ in 2021 employed 3D bioprinting technology to precisely manipulate the biophysical properties of kidney organoids and assess the relative toxicities of drugs such as aminoglycosides.

Renal organoid research must address several key points as it progresses toward the creation of a fully functional in vitro kidney model.⁶⁶ For instance, iPSC-derived kidney organoids face challenges in fully replicating the characteristics of nongenetic kidney diseases due to the removal of epigenetic marks during reprogramming.⁶⁷ Additionally, although iPSC-derived kidney organoids contain nephron structures, they exhibit limited functionality, growth, and lifespan, likely due to the absence of blood supply, immune cells, and neural cells.⁶⁸ Furthermore, the complexity of culturing iPSC-derived kidney organoids and tubules poses challenges in maintaining experimental reproducibility. The excessive use of growth factors can also influence immune cell differentiation. For example, A8301 affects T helper cell polarization by altering transforming growth factor-β signaling,⁶⁹ and the inclusion of interleukin-2 in T cell cultures impacts podocyte injury and apoptosis.⁷⁰

2.4 Advancements and limitations of cardiac organoids

The heart is the first functional organ to develop during human embryogenesis, consisting of cardiomyocytes, cardiac fibroblasts, and endothelial cells.⁷¹ The unique 3D morphology of cardiac organoids has significantly advanced the field by enabling studies of cavity formation, wall thickness, ejection fraction during contraction, and endothelial tubulogenesis.⁷² Compared with organoids from other organs, cardiac organoids initially exhibited slower development. In 2017, Giacomelli et al.⁷³ demonstrated that hiPSCs could codifferentiate into 3D cardiac microtissues composed of cardiomyocytes and endothelial cells. In 2021, Hofbauer et al.⁷⁴ utilized hiPSC technology to create the first self-organizing cardiac organoids by regulating the mesodermal WNT–bone morphogenetic protein signaling axis, resulting in chamber-like cardioids with cavities that replicate cardiac lineage structures. In the same year, Lewis-Israeli et al.⁷⁵ developed complex luminal structures with multiple cardiac cell types that could reproduce heart field formation and atrioventricular features, reflecting the metabolic disruptions associated with congenital heart defects. Drakhlis et al.³¹ reviewed early heart and foregut development, contributing to the investigation of cardiac genetic defects in vitro. Additionally, cardiac organoids have shown promise in transplantation studies. Between 2018 and 2021, several studies demonstrated the transplantation of hESC- and hiPSC-derived cardiac organoids into mouse brain and abdominal cavities.⁷⁶ Overall, cardiac organoids have rapidly progressed in developmental biology, disease modeling, and transplantation research, gaining substantial attention.

Despite these advancements, cardiac organoids require extensive improvements. For instance, iPSC-derived organoids continue to exhibit embryonic characteristics.⁷⁷ Currently, they can replicate cell types such as ventricular cardiomyocytes, endothelial cells, and pericytes, but they lack the ability to model the cardiac conduction system that is necessary for studies of arrhythmogenic cardiomyopathy. Additionally, they do not yet incorporate the vascular structures needed to simulate transient ischemic events or the immune cell coculture systems required in analyses of autoimmune interactions.^{39, 41, 78} Further research is essential to enhance the complexity of cardiac organoids by applying principles of cardiac development to achieve greater maturation and functional simulation in vitro.

2.5 Advancements and limitations of lung organoids

The primary function of the lungs is gas exchange, involving various cell types such as basal cells, club cells, goblet cells, and alveolar epithelial cells (AECs). AEC type I (AEC1) cells are primarily responsible for gas exchange, whereas AEC type II (AEC2) cells secrete surfactants.⁷⁹ Specific regions from proximal airways to distal alveoli contain distinct stem and progenitor cell populations. In 2013, Barkauskas et al.⁸⁰ revealed that AEC2 cells play a key role in alveolar maintenance and repair by forming self-renewing “alveolospheres.” Subsequent studies have shown that the SMAD and NOTCH signaling pathways are crucial for airway stem cell proliferation and differentiation, regulating basal cell behavior.^{81, 82} In 2019, Sachs et al.⁸³ developed human airway organoids that included basal cells, functional multiciliated cells, and mucus-producing secretory cells. Lung organoids have become powerful tools for simulating lung physiology and disease. Lung organoids derived from human lung tumors recapitulate the morphological, histological, and genetic characteristics of the primary tumors.^84-86 Additionally, in 2017, Chen et al.⁸⁷ reported that the introduction of Hermansky-Pudlak syndrome 1 mutations into lung organoids could replicate key features of fibrotic lung disease in vitro. The coronavirus disease 2019 (COVID-19) pandemic accelerated the use of lung organoids in infectious disease research. These organoids have been utilized to study the entry mechanisms of the SARS-CoV-2 virus into human lungs, the pathways of viral transmission, and the mechanisms of viral shedding, revealing that AEC2 cells and the transmembrane protease serine 2 are critical for infection.^88-90 Organoid models have also been used to study other pathogens, including explorations of the differences during infection and replication of human adenovirus types 3 and 55, assessments of potential antiviral drugs,⁹¹ observations of the replication adaptability of human respiratory syncytial virus and its induced innate cytokine responses in lung organoids,⁹² and analyses of avian influenza A virus tropism in human airway organoids, along with changes in cytokine and chemokine profiles.⁹³

Despite these advancements, lung organoids have inherent limitations in fully replicating the cell maturity, interactions, and multicellular complexity observed in vivo. In particular, although lung organoids are primarily derived from AEC2 cells, they often lack the conditions necessary for differentiating into the AEC1 cells required for gas exchange.⁹⁴

2.6 Advancements and limitations of liver organoids

The liver is a crucial hub for numerous physiological processes, including bile metabolism, vitamin metabolism, hormone regulation, blood volume control, immune system support, and endocrine regulation of growth signaling pathways.^{95, 96} The classic method for amplifying liver organoids was established by Huch et al.⁹⁷ in 2013; they demonstrated that Lgr5⁺ cells from injured mouse livers could produce functional liver and bile duct organoids in vitro. In 2015, Huch et al.⁹⁸ achieved long-term expansion of bipotent progenitor cells from adult bile ducts in human livers, which could differentiate into functional hepatocytes. They also found that organoids derived from patients with α1-antitrypsin deficiency and Alagille syndrome accurately reflected in vivo pathology and genetic profiles. In 2023, Zhang et al.⁹⁹ developed 3D liver bud organoid tissues, reconstructing the structural elements of functional and vascularized organs from liver organoid tissues. Liver organoids have significantly advanced our understanding of various liver diseases; they are valuable for disease mechanism research and drug screening.¹⁰⁰ In 2021, Hendriks et al.¹⁰¹ demonstrated that clustered regularly interspaced short palindromic repeats (CRISPR)-associated protein 9 (Cas9) gene editing could be performed in organoids, facilitating studies of liver development and congenital liver diseases. In the same year, Guan et al.¹⁰² developed a human multilineage liver organoid model exhibiting abnormal bile ducts and fibrosis characteristics. In 2023, Zhang et al.¹⁰³ used iPSCs to create three distinct liver organoids for the prediction of drug-induced liver injury based on albumin production, cytochrome P450 expression, and alanine aminotransferase/aspartate aminotransferase release.

Despite rapid advancements in liver organoid research, future improvements may focus on the development of more complex coculture models that include hepatocytes, immune cells, and endothelial cells, as well as more practical and efficient microfluidic devices.¹⁰⁴ Another critical area for development is the exploration of transplantation potential in liver organoids. Although Sampaziotis et al.¹⁸ provided evidence in 2021 that bile duct cell organoids could repair bile ducts after transplantation into human livers, additional evidence is needed to ensure the safety of such procedures for actual patient transplants.

2.7 Advancements and limitations of intestinal organoids

Intestinal epithelial cells are essential for the absorption and metabolism of drugs, nutrients, and water. Conventional 2D cell lines often fail to accurately replicate the complex characteristics of the gut, such as the distinct genetic profiles of epithelial cells in various regions (e.g., ileum and colon), appropriate pH levels, water absorption, and metabolic products. These limitations hinder their effectiveness in mimicking the true physiological environment within the intestine. In 2009, Clevers et al.³ successfully created small intestinal organoids in vitro using mouse Lgr5 intestinal stem cells for the first time. In 2021, Qu et al.¹⁰⁵ developed a proliferative intestinal organoid culture system exhibiting complex crypt-villus structures and significantly enhanced regeneration features related to tissue damage. Intestinal organoids retain the histopathological architecture of their tissue of origin. Using organoids, Lindemans et al.¹⁰⁶ found that interleukin-22 promotes epithelial regeneration mediated by intestinal stem cells. Through single-cell RNA sequencing of cancer organoids and colorectal cancer (CRC) patient biopsies, Dmitrieva-Posocco et al.¹⁰³ revealed that the ketone body β-hydroxybutyrate acts through the Hcar2-Hopx axis, inhibiting cell viability. Fujii et al.¹⁰⁷ constructed 55 CRC organoids to analyze gene mutation profiles and cellular heterogeneity, showing that the dependency on microenvironmental factors in CRC organoids is highly variable. Organoids not dependent on WNT3A/R-spondin1 often carried mutations in the WNT signaling pathway, including adenomatous polyposis coli, catenin beta 1, and transcription factor 7 like 2.

Intestinal organoids have proven to be valuable tools for studying microbe–epithelium interactions. Since 2015, pathogens such as Salmonella Typhimurium, Shigella flexneri, and astroviruses have been shown to infect intestinal organoids.^108-110 These findings have significantly advanced our understanding of the lifecycle of viruses, bacteria, and eukaryotic parasites in the gut, including their processes of adhesion, invasion, infection, and replication. These studies have also provided insights into the cell-specific tropism of these pathogens for intestinal epithelial cells, goblet cells, enteroendocrine cells, and Paneth cells. In 2019, it was found that Cryptosporidium parvum could infect human intestinal and airway organoids, allowing new oocysts to be isolated for downstream analysis.¹¹¹ Additionally, in 2023, researchers demonstrated that SARS-CoV-2 could infect intestinal organoids; its infection efficiency decreases when the farnesoid X receptor was blocked.¹¹²

Although intestinal organoid development has made considerable advances, several critical issues must be addressed. First, the 3D spherical structure and extracellular matrix gel used in these models limit drug and microbial access to the lumen, thereby affecting accurate assessment of intestinal permeability and drug absorption.¹¹³ Puschhof et al.¹¹⁴ suggested that bacterial microinjection could be a feasible solution to overcome this barrier. Second, the intestinal epithelium's sensitivity to pH, shear forces from fluid flow, and oxygen gradients poses challenges.¹¹⁵ Although microfluidic technology has been developed to address these factors, the complexity of this technology indicates that its true clinical translation remains a distant goal. Finally, the creation of multicellular organoid systems that replicate the full range of intestinal functions remains a significant challenge.

2.8 Challenges and innovations in organoid engineering

The current challenges facing organoids can be summarized as follows. First, most organoids lack vascularization and neural system integration. Second, the determination of cell fate and tissue organization, driven by biochemical factors and the extracellular environment, is impacted by variability in the quality control of Matrigel and cytokines, affecting the standardization and reproducibility of organoids. Finally, the achievement of high-throughput, homogeneous, and standardized production, along with automated operations and intelligent monitoring, assessment, and control of organoids, requires the development of more advanced algorithms and methods.

At present, the engineering of organoids, which leverages bioengineering techniques, biomaterials, and AI technology, is collectively advancing the creation of predictive models, accurate disease simulations, and personalized medical strategies.¹³ Methods for engineering organoids can be broadly categorized into four areas: engineered cells, engineered microenvironments, engineered detection methods, and engineered cell–cell interactions (Figure 2).

Engineered cells encompass techniques such as CRISPR–Cas9, air–liquid interface, microbial injection, and viral injection. CRISPR/Cas9 has been widely applied in various organoids to study organ embryonic development,¹¹⁶ tumor suppression,¹¹⁷ and oncogenic transformation.¹¹⁸ The air–liquid interface method more effectively simulates the physiological environment of organoids. For example, brain organoids cultured at the air–liquid interface show improved neuron survival and axon growth.¹¹⁹ Similarly, air–liquid interface culture allows endometrial organoids to mimic the anatomical structure, cellular composition, hormone-induced menstrual cycle changes, gene expression profiles, and dynamic ciliogenesis of the endometrium.¹²⁰ Finally, microinjection is used to study epithelial-microbe interactions in organoid research. For instance, the exposure of human intestinal organoids to genotoxic pks⁺ Escherichia coli can induce CRC mutation signatures.¹²¹

Organoid cultures primarily rely on extracellular matrices derived from animals or tumors. However, the complex and undefined nature of extracellular matrices sourced from murine tumor stroma limits their use in large-scale drug screening and regenerative medicine applications.¹²² To address this challenge, Rizwan et al.¹²³ designed a well-defined viscoelastic hyaluronic acid hydrogel for organoid culture. They discovered that by mimicking the stress relaxation rate of liver tissue, they could induce the growth of cholangiocyte organoids. Additionally, bio-scaffolds offer another approach to the modification of physiological environments within organoids. For example, whereas intestinal organoids cultured in Matrigel typically form randomly developed closed cystic structures, scaffolding can guide the formation of “mini-guts” that more closely resemble physiological states.¹²⁴ Brain organoids also benefit from improved support structures; Ritzau-Reid et al.¹²⁵ developed a microfibril scaffold to guide cavity formation in stem cells and introduced a high-throughput method for the generation, culture, and analysis of engineered brain organoids. Furthermore, the addition of exogenous substances can simulate organoid environments or test developmental processes. For instance, the introduction of oleic acid and free fatty acids to liver organoids can mimic the formation of metabolic-associated fatty liver disease (MAFLD), whereas the addition of troglitazone can simulate drug-induced cholestasis.¹²⁶ The effects of 4-aminopyridine on iPSC-derived brain organoids can also be evaluated in this context.¹²⁷

Despite considerable success in the cultivation of physiologically relevant organoids, the integration of AI, microfluidics, and imaging technologies has accelerated rapid screening, cost-effective extraction of multiscale image features, streamlined multiomics data analysis, and precise preclinical assessment and application. Microfluidic chips have been instrumental in the simulation of high-fidelity disease models and the provision of diagnostic and therapeutic solutions. For example, Abdulla and Quintard^{128, 129} developed a 3D microfluidic brain organoid platform featuring dynamic fluid perturbations and integrated functional vascularized organoid chips, thereby enhancing organoid growth, maturation, and function. Additionally, AI and imaging technologies have undergone rapid advancement. For instance, Kong et al.¹² utilized machine learning to predict drug responses in colorectal and bladder organoids. Renner et al.¹³⁰ combined automated organoid workflows with AI-based analysis, enabling large-scale drug screening for Parkinson's disease and other neurodegenerative diseases. Furthermore, Singaraju et al.¹³¹ developed Organalysis, multifunctional image preprocessing and analysis software for cardiac organoid research, capable of calculating features such as organoid size and fluorescence intensity.

The engineering of cell–cell interactions can effectively address the challenges associated with the construction of organoids of high complexity, size, and structure. Coculture systems provide a relatively straightforward method for the enhancement of organoid complexity and realism. Examples include organoid–fibroblast coculture systems,¹³² intestinal organoid–macrophage systems,¹³³ and liver cancer organoid–endothelial cell systems.¹³⁴ Bioprinting offers a rapid and high-throughput approach to the production of kidney organoids with consistent cell numbers and viability, enabling the assessment of the relative toxicities of drugs such as aminoglycosides.⁶⁵ A key advantage of 3D bioprinting is its ability to more accurately replicate the tumor microenvironment by allowing precise control over multiple biomaterials, cells, and extracellular matrices within predefined architectures.¹³⁵ Considering that drug metabolism and distribution often involve multiple organs, Nguyen et al.¹³⁶ developed a multiorgan chip model incorporating human kidney and liver organoids to study the therapeutic effects and distribution of extracellular vesicles.

3.1 Organoids for disease models

The development of organoids has provided a clearer perspective on human biology, offering significant advantages over conventional disease models (Table 1). Historically, animal models have been the primary tools for studies of disease mechanisms. However, these models face challenges, including genetic and structural differences between species, ethical concerns, long cultivation periods, and high financial costs.¹³⁷ Although cost effective and readily available, cell lines often display altered genetic backgrounds and limited physiological functionality. Additionally, cell line contamination, as encountered with HeLa cells,¹³⁸ LO2 cells,¹³⁹ and the MGc80-3 cell line,¹⁴⁰ has been a serious problem for biomedical research.

Patient-derived organoids retain their genetic complexity and functional differences, providing high-fidelity models for studies of organ development, tissue maintenance, and pathogenesis. As shown in Table 2, organoid models have significantly advanced research into various diseases, including genetic disorders, infectious diseases, metabolic conditions, and cancer.

3.2 Organoids for drug discovery

Organoids have played a pivotal role in the evaluation of new drugs, significantly enhancing the efficiency and accuracy of in vitro validation and testing.⁴¹ Compared with primary cells and cell lines, organoids demonstrate superior organ-specific functions and genetic characteristics. Furthermore, they are more cost effective and time efficient than animal models, such as rodents and rhesus monkeys.^{235, 236}

Patient-derived organoids effectively capture the heterogeneity of patient tumors and facilitate drug screening. For example, Mao et al.²³⁷ constructed CRC organoids to screen 335 drugs, identifying 34 with anti-CRC activity. Fatima and Sun^{238, 239} also developed colorectal and liver cancer organoids to evaluate the antitumor effects of albendazole and romidepsin, respectively. Shukla et al.²⁴⁰ assessed the efficacy of 5-fluorouracil and PRIMA-1Met in 3D bioprinted breast cancer models, which improved organoid reproducibility. Organoids are also valuable for the evaluation of drug toxicity in organs such as the liver, kidneys, and heart. Treatments for neurological diseases often rely on neurotherapeutics, but there are significant concerns about their potential adverse effects on reproductive health. Brain chimeroids, brain chimeras, and testicular and ovarian organoids have been used to assess drug toxicity in these contexts.^{241, 242} For instance, Shinozawa et al.¹⁵ reported that iPSC-derived liver organoids exhibited high predictive value (sensitivity: 88.7%, specificity: 88.9%) for drug-induced liver injury. Ding²⁴³ utilized kidney organoids to evaluate the nephrotoxicity of aspirin, penicillin G, and cisplatin. Organoids are also proving useful for toxicity evaluation in nanomedicine research. Li et al.²⁴⁴ constructed intestinal and hepatic organoid models to assess the cytotoxicity and drug-loaded effects of three typical metal-organic frameworks: ZIF-8, ZIF-67, and MIL-125. Additionally, considering the increasing awareness of microplastic hazards, Qin et al.²⁴⁵ used endometrial organoids to investigate microplastic pollution and its potential impact on reproductive health.

The integration of bioengineering and AI technologies with organoid technology has significantly advanced drug screening and toxicity assessment. Microfluidic technology offers a more physiologically relevant environment for organoids, enabling real-time monitoring of factors such as oxygen tension, cytokine concentrations, and shear stress. For example, Kasendra et al.²⁴⁶ developed a duodenal small intestine chip that featured polarized cell structures, intestinal barrier function, and expression of key intestinal drug transporters. This model demonstrated improved cytochrome P450 3A4 (CYP3A4) expression and induction compared with Caco-2 cells.²⁴⁶ Zhang et al.²⁴⁷ designed an automated microfluidic chip-based system for continuous monitoring of organoid drug responses. Whereas single liver organoids can only reflect hepatotoxicity, multiorgan organoid models provide a more comprehensive evaluation of drug-specific organ toxicity and metabolism. In 2017, Skardal et al.²⁴⁸ constructed a tri-tissue organ-on-chip system, which included liver, heart, and lung tissues, thereby mimicking the interactive nature of the human body. In 2020, the team further advanced this concept by developing an integrated organoid chip system that incorporated liver, heart, lung, endothelium, brain, and testis organoids to study the metabolic toxicities of drugs such as 5-fluorouracil and ifosfamide.²⁴⁹ AI technologies play a crucial role in the analysis of organoid-based data, including microscopic images, transcriptomics, metabolomics, and proteomics, thereby facilitating high-throughput drug screening.²⁵⁰ In 2021, Renner et al.¹³⁰ combined automated organoid workflows with AI-based analysis to develop next-generation interdisciplinary high-throughput screening methods for Parkinson's disease and other conditions. In 2022, Matthews et al.²⁵¹ introduced a robust image analysis platform capable of automatic identification, labeling, and tracking of individual organoids in brightfield and phase-contrast microscopy experiments, enabling the calculation of dose effects on organoid circularity, solidity, and eccentricity. In 2023, Compte et al.²⁵² used pancreatic ductal adenocarcinoma organoids combined with AI-driven live-cell image analysis to match retrospective clinical patient responses and screen for chemotherapy drug sensitivity.

The high cost of drug development, estimated at approximately $1 billion per drug, presents significant challenges for pharmaceutical companies. These expenses contribute to elevated drug prices and can inhibit innovation, primarily due to extensive research and development efforts and the high risks associated with clinical trial failures.²⁵³ Major companies, including Johnson & Johnson, Roche, Takeda, and Merck, have already initiated investments in high-throughput drug screening using organoids and organ-on-chip technologies.^{254, 255} However, before organoids can be widely adopted in clinical settings, several critical issues must be addressed by clinicians, pharmacologists, bioengineering laboratories, and regulatory agencies. First, there is a pressing need for more mature and standardized organoid models. Researchers must reach a consensus regarding the specific media, initial cell types, induction methods, and development cycles required for different organoids. Second, the integration of materials science, bioengineering, and AI technologies is crucial to improve the scalability and automation of drug screening processes. This integration will enable more accurate and high-throughput simulations of drug metabolism across various organs, including the skin, intestine, blood vessels, liver, and kidneys. Finally, ethical and regulatory concerns surrounding the use of organoids must be thoroughly addressed. Although organoids pose relatively low ethical risks, they retain patients' genetic information, and the potential for advanced brain organoids to develop consciousness remains a significant concern.

3.3 Organoids for personalized medicine

Precision medicine aims to enhance disease characterization at the molecular and genomic levels, thereby improving drug screening. In recent years, extensive organoid biobanks have been established for various organs, including the brain,¹⁶⁴ stomach,^{227, 256} liver,^{257, 258} kidneys,²⁵⁹ intestines^{184, 260} pancreas,²⁶¹ breast,¹⁴³ ovaries,²⁶² cervix,²⁶³ and bladder.²⁶⁴ Organoids derived from tumors and genetic diseases can faithfully replicate the phenotypic and genomic characteristics of primary tumors. For example, Sachs et al.¹⁴³ constructed over 100 breast cancer organoids, discovering that the DNA copy number variations and sequence changes in the organoids were consistent with those features in the original tumor tissues. Similarly, Ji et al.²⁵⁷ developed 65 human liver cancer organoids, with proteomic analysis revealing that the organoids retained the molecular and phenotypic features of the parent cancer tissues. This study demonstrated that the organoid platform could capture both intra- and inter-patient heterogeneity through multiomics analyses, including histopathology, genomics, transcriptomics, and single-cell sequencing. Moreover, organoids are valuable tools for the investigation of rare genetic diseases and identification of personalized treatments. For instance, Yao et al.²⁶⁵ and Soroka et al.¹⁷⁴ created transcriptional atlases of primary sclerosing cholangitis and cholangiocyte organoids, showing that even in nontumorous conditions, organoids retained the specific pro-inflammatory gene signatures characteristic of primary sclerosing cholangitis. Additionally, Geurts et al.²⁶⁶ constructed a biobank of intestinal organoids from 664 cystic fibrosis (CF) patients and used CRISPR gene editing to correct CF-related gene mutations in these organoids.

Given the high-fidelity genetic profiles of organoids, drug sensitivity testing with these models can quickly identify the most effective treatments for individual patients. For example, Yuan et al.²⁶⁷ developed organoids from gallbladder cancer, normal gallbladder tissue, and benign gallbladder adenomas. They found that the dual PI3K/HDAC inhibitor CUDC-907 significantly inhibited the growth of various gallbladder cancer organoids while displaying minimal toxicity to normal gallbladder organoids. Similarly, Phan et al.²⁶⁸ tested 240 United States Food and Drug Administration (US FDA)-approved or clinically developed protein kinase inhibitors on patient-derived organoids from clear cell/high-grade serous tumors and platinum-resistant high-grade serous ovarian cancer; their findings could aid clinical decision-making. These examples underscore the potential for organoid-based drug sensitivity testing to offer personalized treatment options and advance precision medicine.

This innovative approach aims to streamline the drug discovery process, reduce costs, and increase the success rate of new treatments by providing more accurate models of human physiology and disease. As of July 2024, a search for the keyword “organoids” on ClinicalTrials.gov identified a total of 86 registered clinical trials (Table 3). These trials are predominantly sponsored by institutions or individuals from China and the United States; most have been initiated within the past 5 years. The majority of the trials target cancers; colon cancer (15 trials), breast cancer (16 trials), ovarian cancer (five trials), and gastric cancer (eight trials) are the most common focus areas. These findings underscore the growing momentum in precision medicine research concerning patient-derived organoids. Organoids are emerging as powerful tools for disease gene screening and personalized drug testing, which will facilitate advancements in precision medicine.

3.4 Organoids for regenerative medicine

Organ transplantation remains the most definitive and effective treatment for diseases that result in irreversible organ failure. However, due to the shortage of donor organs and the adverse effects of immunosuppressive therapies, thousands of patients die each year while waiting for a transplant.²⁶⁹ The advent of organoid technology has introduced a powerful new tool to the field of regenerative medicine. Organoids have the potential to integrate, mature, vascularize, and develop specific targeted physiological functions in vivo, suggesting that in vitro cultivation of fully functional organs could one day become a reality.²⁷⁰

Significant progress has been made in the transplantation of human organoids derived from various organs, including the intestine, retina, kidney, liver, pancreas, brain, lung, and heart. Intestinal organoids have been particularly well studied. In 2014, Watson et al.²⁷¹ generated human intestinal organoids from hESCs or iPSCs and observed significant expansion and maturation of the epithelium and mesenchyme upon transplantation into mice; the organoids exhibited functional attributes, such as permeability and peptide uptake. Xenotransplanted human ileal organoids retained their regional specificity and formed new villus structures in the mouse colon, offering a potential treatment for patients with short bowel syndrome. In July 2022, a Japanese research team conducted the first autologous transplantation using intestinal organoids cultured from the healthy gut mucosal stem cells of ulcerative colitis patients.²³⁵ Retinal organoids have also made substantial progress toward clinical application. Since 2011, iPSC-derived retinal organoids transplanted into animal models of retinal diseases have shown promising results, including the development of mature and functional photoreceptors, integration with the host retina, and improvements in vision and light sensitivity.^{272, 273} In 2023, Hirami et al.⁶ transplanted allogeneic iPSC-derived retinal organoid sheets into two patients with advanced retinitis pigmentosa, observing increased retinal thickness at the transplant site without severe adverse events. Organoids derived from cholangiocytes and liver cells have also undergone rapid advancement. Takebe et al.²⁷⁴ used iPSC technology to create vascularized, functional liver organoids that, when transplanted into immunodeficient mice, quickly formed functional vasculature and exhibited drug metabolism activity. In 2021, Sampaziotis et al.¹⁸ demonstrated that human bile duct organoids maintained their plasticity and restored function in vivo when transplanted back into the human biliary tree, despite the loss of transcriptional diversity in vitro, as shown by single-cell RNA sequencing. In 2022, Willemse et al.²⁷⁵ enhanced the clinical translation potential of bile duct organoids by using extracellular matrix-derived hydrogels from decellularized human or pig livers for culture, instead of the commonly used tumor-derived basement membrane extract. Other organoids, such as those derived from the brain, pancreas, and lungs, also show regenerative potential. In 2018, Mansour et al.³⁹ found that brain organoid transplantation led to integration, gradual maturation, neuronal differentiation, synapse formation, and gliogenesis, resulting in the development of a functional vascular system within the grafts. In 2020, Yoshihara et al.²⁷⁶ reported that human islet-like organoids rapidly reestablished glucose homeostasis in diabetic immunodeficient mice. In 2019, Weiner et al.²⁷⁷ transplanted dissociated AEC2 organoids into influenza-infected mice, demonstrating that the organoids proliferated and differentiated into AEC2 cells, which were able to support lung regeneration.

Overall, these studies suggest that organoids could become a viable alternative to conventional organ transplantation, potentially alleviating future organ shortages. Table 4 presents a selection of representative papers on human organoid transplantation from the past two decades, highlighting the progress and potential of this emerging field.

Despite their promising potential in regenerative medicine, organoids face several challenges and limitations. First, organoid size and expansion are critical for their clinical application. Transplantable organoids must be as large and numerous as possible, but their proliferation rate is significantly lower than that of cancer cell lines. One feasible approach to address this challenge involves the use of rotating bioreactors. For instance, Qian et al.^{217, 278} developed a small rotating bioreactor, known as SpinΩ, to optimize brain organoid maturation while reducing costs. Second, organoid maturation presents a significant hurdle. The culture requirements for organoids differ substantially from those for conventional monolayer cell cultures. Most iPSC- and hESC-derived organoids tend to resemble fetal rather than adult cells, which makes it difficult to replicate complete vascularization and neural networks in vitro.^{279, 280} Although scientists have developed various bioreactors to meet the oxygenation, lineage progression, and 3D spatial organization needs of organoids, progress remains challenging.²⁸¹ Limitations such as high fluid shear and impeller instability can negatively affect shear-sensitive organoid cultures. Although technologies such as SpinΩ have significantly advanced the bioexpression of forebrain organoids, 3D printing technology remains exceedingly complex for widespread use.²⁸² Third, the need for robust preclinical validation models poses another challenge. The immune-deficient models essential for preclinical evaluation are costly, limiting their accessibility and use.^{283, 284} Finally, the assurance of transplantation safety is a central requirement. Before organoid transplantation can become a standard practice, concerns regarding potential tumorigenicity and the differentiation capacity of organoids must be addressed.⁶⁶ The establishment of protocols that comply with Good Manufacturing Practice guidelines to evaluate the long-term risks of in vitro-cultured organoids and their potential tumorigenicity post-transplantation offers a rational solution to these safety concerns.²⁸⁵

4.1 Future organoid research

Organoids have made substantial progress in disease simulation, drug development, patient precision medicine, and regenerative medicine. A promising future direction for organoid research lies in the development of multiorganoid systems that use microfluidics and biomaterials to integrate organ systems, thereby enabling highly detailed disease modeling. Perfusable microfluidic devices can ensure the efficient delivery of nutrients, gradient pressure, and oxygen to organoids, which are crucial for the maintenance of their viability and functionality. By manipulating biomaterial parameters such as cell-binding ligands, structural geometry, and mechanical properties, researchers can control organoid development and disease progression.^{312, 313} Because diseases, drugs, and microorganisms impact organ function and structure, as well as the organ's microenvironment, “multiorganoid chips” are poised to become a cutting-edge technology for the evaluation of new drug efficacy and toxicity.³¹⁴

Moreover, the integration of organoid technology, tissue engineering, and AI offers exciting possibilities for enhanced understanding of organoid engineering principles.¹³ Studies have begun to explore the use of AI in hydrogel design, optimization, and its applications in biomedicine.³¹⁵ For instance, Kowalczewski et al.³¹⁶ utilized unsupervised machine learning to cluster and refine cardiac organoids based on functional similarity, revealing unique features related to their geometric design. Additionally, Lefferts et al.³¹⁷ introduced OrgaSegment, a deep learning segmentation model based on MASK-RCNN, which quantifies organoid swelling. This model can differentiate among organoids with various CF transmembrane conductance regulator (CFTR) mutations and assess their responses to CFTR-modulating drugs. The integration of organoid technology with AI has enabled a more comprehensive analysis of organoid behaviors, intricate cellular interactions, and dynamic responses to various stimuli.³¹⁸ AI is revolutionizing our approach to in vitro modeling, pushing the boundaries of biomedical research.

4.2 Challenges and limitations of organoid research

Although the field of organoids holds broad prospects for various applications, it faces several significant challenges and limitations. First, organoid technology bridges the gap between cell lines and in vivo models; however, most organoids lack the surrounding stromal cells, immune cells, and vascular endothelial cells necessary for comprehensive modeling.³¹⁹ For example, liver organoids often lack hepatocyte zonation and essential components of metabolic fatty liver disease pathogenesis, such as vasculature, immune cells, and neural innervation.³²⁰ Second, the organoid industry is hindered by a lack of standardization, a situation exacerbated by the rapid progress in engineered organoids. Factors affecting reproducibility include batch-to-batch variation (e.g., patient tissue heterogeneity, as well as timing and method of iPSC induction), culture conditions (e.g., cytokine concentration, matrix gel concentration, and matrix gel composition), and cell composition and organoid structure.³²¹ The resolution of these issues will require collaborative efforts from biomedical scientists, clinicians, and regulatory agencies to standardize and homogenize organoid technology, which would facilitate the transition from scientific research to clinical practice and enable the production of larger-scale organoids for drug screening.²³⁵ Third, the cultivation of human organoids often relies on mouse-derived extracellular matrix substitutes, such as Matrigel and basement membrane extracts. However, issues such as the tumor origins of these materials, batch-to-batch variability, high costs, and safety concerns impact the reproducibility and practicality of experimental results. Additionally, the presence of unknown pathogens and potential immune responses make these substrates unsuitable for clinical transplantation.³²² Recent advances have introduced alternative materials, such as graphene oxide, biomimetic hydrogels, and self-assembling peptide hydrogels, which offer promising replacements for Matrigel and enhance the development of advanced organoid models.^323-325 Although organoids generally raise fewer ethical issues than other technologies, there are ongoing ethical deliberations concerning potentially controversial entities such as brain organoids and embryoid bodies.³²⁶ Much of this debate centers on the ethical status of such organoid models and the implications of their use.³²⁷

In summary, although the field of organoids is rapidly advancing, several challenges remain, including issues with standardization, matrix gel contamination, bioethics, and complexity. The resolution of these challenges through the construction of engineered organoids, the development of new materials, and the establishment of industry standards will be crucial for future progress in this field.

4.3 Conclusion

In summary, organoids have demonstrated immense potential in disease modeling, drug development, and precision medicine. They are capable of replicating and exhibiting complex biological processes, including multicellular and organ interactions, tumor immune microenvironments, and host-pathogen interactions. Moreover, the integration of organoids with advanced materials science, AI, and gene editing has reshaped our understanding of diseases and translational therapies, significantly advancing our efforts to address and conquer human diseases.