中文导读
类器官技术为癌症与遗传疾病研究提供了接近人体的模型。本文系统介绍疾病建模类器官的构建思路、应用场景与研究进展。
以下为英文全文（English full text）
Abstract
Organoid technology has emerged as one of the most transformative advances in biomedical research over the past decade. These three-dimensional (3D) self-organizing tissue structures, derived from adult stem cells or pluripotent stem cells, recapitulate the architecture, cellular heterogeneity, and functional properties of their tissue of origin with remarkable fidelity. This article provides a comprehensive overview of disease modeling using organoids, tracing the historical evolution of the field, comparing organoid models with conventional 2D cell cultures and animal models, and reviewing applications across cancer, genetic disorders, infectious diseases, and neurodegenerative diseases. We also discuss the powerful synergy between organoid technology and CRISPR-Cas9 gene editing, which has opened new avenues for mechanistic studies and personalized therapeutic strategies.
1. Historical and Current Landscape of Disease Modeling
The practice of modeling human diseases in vitro has evolved dramatically over the past century. Early approaches relied on immortalized 2D cell lines grown on flat plastic surfaces, which offered experimental tractability but failed to replicate the complex 3D architecture, cell-cell interactions, and tissue-specific microenvironments of human organs. Animal models, particularly genetically engineered mice, provided valuable in vivo context but are limited by species-specific differences in physiology, immune responses, and drug metabolism. These limitations have contributed to the high attrition rates in clinical drug development, with many promising preclinical candidates failing in human trials.
The advent of organoid technology began with seminal work by Sato et al. (2009), who demonstrated that single Lgr5-positive intestinal stem cells could self-organize into 3D epithelial structures resembling intestinal crypts when cultured in a matrix containing basement membrane components and a defined growth factor cocktail [1]. This breakthrough established the foundational principle that organ-specific stem cells retain intrinsic self-organization programs when provided with appropriate niche signals. Over the subsequent decade, organoid protocols were developed for virtually every major organ system, including the colon, liver, pancreas, lung, breast, brain, and kidney.
Today, organoid technology sits at the intersection of stem cell biology, tissue engineering, and precision medicine. Living biobanks of patient-derived organoids (PDOs) have been established for numerous cancer types, and clinical trials are underway to evaluate the utility of organoid-based drug sensitivity testing in guiding treatment decisions for individual patients [2].
2. Advantages of Organoids in Disease Modeling
2.1 Versus 2D Cell Lines
Traditional 2D cell cultures have been the workhorse of biomedical research for decades, but they suffer from fundamental limitations that organoids address. Two-dimensional cultures lack the spatial gradients of nutrients, oxygen, and signaling molecules that characterize native tissues. Cells grown on rigid plastic substrates adopt flattened morphologies and altered gene expression profiles compared to their in vivo counterparts. Most critically, 2D cultures fail to capture the cellular heterogeneity and 3D architecture of tumors, which are increasingly recognized as critical determinants of therapeutic response.
Organoids, by contrast, self-organize into structures that mimic the polarized epithelial architecture of native tissues. They exhibit apical-basal polarity, form lumen-like structures, and establish cell-cell interactions that more faithfully represent in vivo biology. Studies have shown that gene expression profiles of organoids more closely match those of primary tissues than 2D cultures, and organoid drug responses are significantly more predictive of clinical outcomes [3].
2.2 Versus Animal Models
Animal models have been indispensable for studying disease pathophysiology and evaluating drug candidates. However, fundamental species differences limit their translatability. For example, mouse liver cytochrome P450 enzymes differ substantially from their human counterparts in substrate specificity and expression patterns. Mouse immune systems, while similar in broad architecture, differ in the specifics of cytokine signaling, Toll-like receptor repertoires, and immune cell subsets. These differences contribute to the well-documented problem of preclinical-to-clinical translation failure.
Organoids derived from human tissues bypass these species-specific limitations. Human tumor organoids preserve the genetic landscape, histopathological features, and drug response patterns of the patient's original tumor. Furthermore, organoid cultures can be established from normal human tissues, enabling direct comparisons between healthy and diseased states within the same genetic background. The ethical and logistical advantages of organoid models—faster establishment, lower cost, and reduced animal use—have made them increasingly attractive for both basic and translational research [4].
3. Cancer Organoid Modeling: PDOs in Tumor Research
Patient-derived tumor organoids (PDOs) represent the most extensively developed application of organoid technology in disease modeling. PDOs can be established from surgical resections, needle biopsies, and even liquid biopsies across a wide range of solid tumor types, including colorectal, lung, breast, pancreatic, prostate, liver, and gastric cancers.
The establishment process involves enzymatic digestion of tumor tissue into small epithelial fragments or single-cell suspensions, which are then embedded in basement membrane matrix (such as Matrigel) and overlaid with a tissue-specific growth medium. For epithelial cancers, the medium typically contains Wnt-conditioned medium or recombinant Wnt3a, R-spondin 1, Noggin, epidermal growth factor (EGF), and additional tissue-specific factors [5].
A landmark study by van de Wetering et al. (2015) established a living biobank of 22 colorectal cancer (CRC) organoid lines, demonstrating that PDOs closely recapitulate the histopathology, mutational landscape, and transcriptome of the parental tumors [6]. Subsequent studies have extended this approach to numerous other cancer types. Critically, PDOs have been shown to predict patient drug responses with high accuracy. In a prospective study of metastatic gastrointestinal cancers, PDO drug sensitivity profiles achieved 100% sensitivity and 93% specificity in predicting clinical responses [7].
The ability to cryopreserve and revive PDOs enables the creation of permanent living biobanks that can be shared across research institutions and used for large-scale drug screening campaigns. These biobanks capture the inter-patient heterogeneity that is a hallmark of human cancer, providing platforms for studying rare mutations and resistance mechanisms that would be difficult to model with conventional cell lines [8].
4. Genetic Disease Modeling
Genetic diseases affecting epithelial organs have been particularly amenable to organoid modeling. The ability to derive organoids from patient induced pluripotent stem cells (iPSCs) or adult stem cells carrying disease-causing mutations enables researchers to study disease mechanisms in human-relevant tissue contexts.
4.1 Cystic Fibrosis (CF)
Cystic fibrosis is caused by mutations in the CFTR gene, which encodes a chloride channel essential for epithelial ion transport. Intestinal organoids derived from CF patients have emerged as a powerful model system. Dekkers et al. (2013) demonstrated that CF intestinal organoids show a characteristic absence of swelling in response to forskolin-induced cAMP elevation, due to the lack of functional CFTR-mediated chloride secretion [9]. This organoid-based readout has been developed into a diagnostic and drug screening platform, with the CFTR potentiator ivacaftor restoring forskolin-induced swelling in organoids from patients with responsive genotypes. The assay is now used clinically to predict patient responses to CFTR modulators [10].
4.2 Polycystic Kidney Disease (PKD)
Polycystic kidney disease, both autosomal dominant (ADPKD) and autosomal recessive (ARPKD) forms, has been successfully modeled using kidney organoids derived from human pluripotent stem cells. Freedman et al. (2015) used CRISPR-Cas9 to introduce biallelic PKD1 and PKD2 mutations in hPSCs, and the resulting kidney organoids developed cystic structures that recapitulated key features of human PKD [11]. More recent studies have demonstrated that flow-induced mechanical stress acts as a novel cystogenic mechanism in PKHD1-mutant organoids cultured on microfluidic chips, highlighting the importance of integrating organoid technology with organ-on-chip platforms [12].
5. Infectious Disease Modeling
Organoids have proven invaluable for studying host-pathogen interactions in human tissues that are difficult to access in vivo. The SARS-CoV-2 pandemic dramatically accelerated the use of organoid models for infectious disease research. Human airway organoids, established from nasal or bronchial epithelial cells, were rapidly deployed to study SARS-CoV-2 tropism and pathogenesis. These models revealed that the virus preferentially infects ciliated and secretory cells in the proximal airway, and that interferon responses play a critical role in restricting viral replication [13].
Gastric organoids have been used to model Helicobacter pylori infection, the primary cause of gastric ulcers and gastric cancer. In co-culture experiments, H. pylori was shown to adhere to gastric epithelial cells within organoids, induce inflammatory responses, and promote epithelial-to-mesenchymal transition—key steps in the pathogenesis of gastric malignancy [14]. Intestinal organoids have also been employed to study enteric pathogens including Clostridioides difficile, Salmonella, and enteroviruses, providing insights into mechanisms of bacterial adhesion, toxin effects, and host defense [15].
6. Neurodegenerative Disease Modeling
Brain organoids, often referred to as "cerebral organoids" or "mini-brains," have emerged as powerful models for studying neurodevelopmental and neurodegenerative disorders. Derived from human pluripotent stem cells, these structures self-organize into regions resembling the cerebral cortex, hippocampus, and other brain areas, containing neuronal and glial cell types in a 3D arrangement [16].
Alzheimer's disease has been modeled using brain organoids carrying familial mutations in APP, PSEN1, or PSEN2. These models recapitulate amyloid-beta aggregation and tau hyperphosphorylation, enabling mechanistic studies of disease initiation and progression. Parkinson's disease organoids have been generated from patient iPSCs carrying mutations in LRRK2, SNCA, or PARK2, and exhibit alpha-synuclein aggregation and dopaminergic neuron dysfunction [17].
While brain organoids currently lack the vascularization, microglial populations, and long-range axonal connections of the intact brain, advances in co-culture systems and microfluidic platforms are rapidly addressing these limitations. The integration of microglia, derived from hematopoietic progenitors, into brain organoids has enabled the study of neuroinflammation in Alzheimer's disease models [18].
7. Organoids and CRISPR Gene Editing for Disease Modeling
The combination of organoid technology with CRISPR-Cas9 genome editing has created unprecedented opportunities for disease modeling. CRISPR enables the introduction of specific disease-causing mutations into wild-type organoids or the correction of mutations in patient-derived organoids, establishing isogenic pairs that differ only at the edited locus. This approach controls for genetic background effects and provides definitive evidence for genotype-phenotype relationships.
In cancer research, CRISPR has been used to engineer oncogenic mutations in normal organoids, sequentially introducing mutations in APC, TP53, KRAS, and SMAD4 to model the progression of colorectal cancer from adenoma to carcinoma [19]. Large-scale CRISPR screens in gastric organoids have identified gene-drug interactions and synthetic lethal targets that are being pursued for therapeutic development [20].
For genetic diseases, CRISPR correction of patient-derived iPSCs followed by organoid differentiation has validated disease mechanisms and demonstrated proof-of-concept for gene therapy approaches. In PKD, CRISPR-mediated correction of PKD1 mutations in patient iPSCs restored normal organoid morphology, providing compelling evidence that the cystic phenotype was directly caused by the mutation [21].
Base editing and prime editing technologies are further expanding the toolkit, enabling the introduction of precise point mutations without double-strand breaks. These advances are enabling the generation of more accurate disease models, particularly for disorders caused by single-nucleotide variants [22].
Conclusion
Organoid technology has fundamentally transformed disease modeling, offering human-relevant 3D tissue systems that bridge the gap between traditional cell culture and animal models. From cancer PDOs that predict patient drug responses to genetic disease models that enable mechanistic studies and therapeutic screening, organoids are accelerating the pace of biomedical discovery. The integration of CRISPR gene editing, organ-on-chip platforms, and advanced imaging techniques is further enhancing the power and physiological relevance of these models. As organoid-based clinical trials advance and biobanks expand, we anticipate that these technologies will play an increasingly central role in precision medicine and drug development.