In vitro biomimetic models for respiratory diseases: progress in lung organoids and lung-on-a-chip

Abstract

Respiratory diseases pose a growing global public health challenge due to their rising incidence and mortality. Conventional cellular and animal models fall short in replicating the complex three-dimensional (3D) microenvironments of the lung and are limited by species differences, restricting their clinical relevance. Emerging technologies, such as lung organoids (LOs) and lung-on-a-chip (LOC), combine stem cell differentiation, microfluidic systems, and biomechanical cues to create advanced in vitro models that address these limitations. LOs leverage the innate ability of human pluripotent stem cells (hPSCs) to self-organize, faithfully mimicking lung development, cellular diversity, and the native tissue environment. In comparison, LOC platforms employ flexible biomaterials to recreate the respiratory microenvironment dynamically, enabling a detailed study of normal physiology and disease states. This review highlights recent advancements in LOs- and LOC-based models across five key respiratory conditions: infections, chronic obstructive pulmonary disease (COPD), asthma, pulmonary fibrosis (PF), and lung cancer. Furthermore, it discusses their capacity to deepen understanding of disease mechanisms, accelerate drug discovery, enhance pharmacological evaluation, and serve as comprehensive tools for respiratory research.

Introduction

The respiratory system facilitates gas exchange and maintains the body’s internal balance with the external environment. Beyond serving as a multi-layered defense system through ciliary movement, mucus production, and coordinated immune responses, it also supports vital physiological functions, including acid–base regulation, metabolic control, and the transformation of bioactive molecules, all of which are essential for preserving overall systemic homeostasis [1, 2]. Amid the swift rise in population aging, worsening air pollution, and changing smoking trends characterized by higher rates among younger people and women, chronic respiratory diseases and lung cancer, which are key contributors to the four leading global non-communicable diseases, have increasingly become a major global health concern [3,4,5,6,7,8]. At the same time, the COVID-19 pandemic has highlighted the persistent danger of respiratory infectious diseases, which are anticipated to continue posing significant challenges to global health security for the foreseeable future [9, 10]. These challenges emphasize the critical importance of advancing our knowledge of respiratory disease mechanisms, innovating new treatment strategies, and speeding up the development of effective pharmacological therapies.

Developing in vitro models that accurately mimic in vivo physiological conditions is essential for progress in disease research. While animal models can partially reproduce human lung characteristics, their usefulness is limited by species-specific differences, prolonged study durations, and significant expenses [11, 12]. Traditional two-dimensional (2D) cell cultures, which lack a 3D microenvironment and dynamic cellular interactions, cannot reproduce the intricate tissue architecture and functional complexity of the human lung [13, 14]. A significant technological breakthrough has been using microengineering techniques to create 3D cell culture systems. The landmark creation of the first 3D intestinal organoids by Clevers et al. [15] in 2009 set a crucial precedent, paving the way for further progress in organoid research. Lung organoids, generated from pluripotent stem cells (PSCs) or adult stem cells (ASCs), demonstrate self-organization and spatial differentiation, forming 3D microtissues that faithfully mimic the spatiotemporal development of lung tissue in vitro [16]. Compared to conventional models, LOs provide cellular diversity, a 3D structure, and air–liquid interface (ALI) conditions that more accurately replicate the in vivo lung environment while avoiding the species-related limitations seen in animal models [12, 17,18,19,20].

Alongside organoid-based models, LOC technology employs microfluidics and bioengineering techniques to mimic the structural and functional dynamics of the human lung at the microscale in vitro [21,22,23]. LOC platforms combine programmable microfluidic channels, flexible biomaterials, and multicellular co-cultures to reconstruct the alveolar-capillary interface and replicate respiratory motion through cyclic mechanical stretching [24, 25]. The immune system plays a central role in respiratory diseases, especially in the immune responses triggered by infections and lung injury. Incorporating immune cells into co-culture systems with LOs or LOC platforms allows for a more accurate simulation of the complex interactions between pathogens, immune responses, and pulmonary damage. These advances are transforming respiratory disease modeling, driving forward personalized medicine, and developing targeted therapies.

This review emphasizes the unique strengths of LOs and LOC in respiratory disease research, outlines their structural characteristics and recent technological progress, and explores how integrating these approaches could address existing challenges to improve physiological accuracy and translational potential (Fig. 1).

Fig. 1

Using Lung Organoids and Lung-on-a-Chip Platforms to Explore the Pathogenesis and Therapeutic Approaches for Respiratory Diseases, including Infections, asthma, COPD, pulmonary fibrosis, and lung cancer

In vitro pulmonary physiological models: bridging the gap from 2 to 3D and from static to dynamic paradigms

The emergence and evolution of lung organoids technology

Lung development begins in the ventral foregut endoderm of the primitive gut tube, where the initial lung buds form. Multipotent epithelial progenitor cells, known as “apical progenitor cells,” are located at the tips of these buds and are surrounded by mesenchymal tissue. As lung growth progresses, epithelial cells repeatedly branch through branching morphogenesis, creating a complex tree-like structure typical of the mature lung. During this process, apical progenitor cells maintain their ability to self-renew and proliferate while differentiating into all major lung epithelial cell types. Early in development, branching morphogenesis produces a tubular network of bronchi and bronchioles that supports air conduction. At later stages, apical progenitor cells at the distal ends of the airways mature into alveolar epithelial cells [16].

Induced pluripotent stem cells (iPSCs), reprogrammed from human somatic cells, have emerged as a vital source for generating LOs due to their pluripotent differentiation potential [26]. The standard differentiation process typically proceeds through three key stages: definitive endoderm, foregut endoderm, and NKX2-1⁺ lung progenitor cells (Fig. 2). Researchers have progressively refined the culture protocols for producing LOs from iPSCs by precisely modulating critical signaling pathways (Fig. 3).

Fig. 2

Schematic overview of lung organoids generation from induced pluripotent stem cells (iPSCs) and primary lung tissues. iPSCs are sequentially differentiated through three key stages: definitive endoderm, foregut endoderm, and NKX2-1⁺ lung progenitor cells. Activation of the WNT signaling pathway (via CHIR99021) and inhibition of TGF-β signaling (via SB431542) in high-density sorted NKX2-1⁺ cells promotes the formation of alveolar-like organoids containing type I and type II alveolar epithelial cells (SFTPB⁺, SFTPC⁺). Stimulation with FGF10 further supports the development of lung organoids enriched with type I and II alveolar epithelial cells (SFTPB⁺, SFTPC⁺), goblet cells (SCGB1A1⁺), basal cells (TP63⁺), ciliated cells (FOXJ1⁺), and mesenchymal components (ECAD⁺). Further induction using CHIR99021, FGF7, FGF10, BMP4, and retinoic acid leads to branching lung organoids containing mesenchymal cells (ZEB1⁺), airway-like cells (SOX2⁺), and diverse cell populations. Furthermore, airway organoids can be generated by embedding lung progenitor cells, fibroblasts, and lung microvascular endothelial cells from adult lung tissue in a 3D Matrigel matrix within Transwell or culture dishes, followed by growth factor stimulation resulting in pseudostratified ciliated epithelium and other airway cell types. Fetal lung organoids are derived by isolating stem cells from fetal amniotic fluid (indicated by the dashed arrow)

Fig. 3

Key Milestones in LOs Development from 2015 to 2025. Abbreviations: LOs, lung organoids; iPSCs, induced pluripotent stem cells; ASCs, adult stem cells; LFOs, fetal lung organoids; AOs, airway organoids

In 2015, Dye et al. [27] introduced a landmark protocol for differentiating iPSCs into human lung organoids and lung bud organoids. This method employed Activin A to induce endoderm formation, combined with dual inhibition of transforming growth factor-β (TGF-β) and bone morphogenetic protein (BMP) pathways, alongside synergistic activation of WNT, Sonic hedgehog (Shh) and fibroblast growth factor 4 (FGF4) signaling. The resulting LOs demonstrated characteristic pulmonary epithelial-mesenchymal features. Based on this foundation, Chen et al. [28] developed an improved protocol in 2017 to generate more mature LOs featuring branched airway structures and early alveolar morphologies. Later, in 2019, Miller and Dye’s group [16] further refined the methodology to produce more precise and developmentally advanced LOs. Han et al. [29] proposed an efficient induction strategy based on the work of Chen et al. [28], which employed only six cytokines and reduced the differentiation timeline to 41 days to address the lengthy culture duration. To better recapitulate human lung architecture, researchers have developed specialized LOs representing alveolar structures and bronchial architectures [30, 31]. However, key physiological features, such as functional cilia, which are critical for modeling mucociliary clearance, remain incompletely described, and further integration with complex LOC systems is still underway.

Traditional LOs protocols typically rely on Matrigel, a basement membrane matrix derived from Engelbreth-Holm-Swarm mouse sarcomas [32], and serum-based media, contributing to variability and limiting reproducibility. To overcome these challenges, several innovative approaches have been developed: Loebel et al. [33] introduced tunable synthetic hydrogels as Matrigel substitutes; Budeus et al. [34] established a matrix-free culture method; and Leko et al. [30] proposed a serum-free induction protocol, which significantly reduced batch-to-batch variability and improved reproducibility and overall success rates.

Beyond iPSCs, adult stem cells have also gained prominence as a valuable source for lung organoids generation, owing to their tissue specificity and more controllable differentiation potential [35, 36]. In 2017, Q et al. [37] successfully established airway organoids (AOs) using a 3D co-culture system composed of primary adult bronchial epithelial cells, lung fibroblasts, and microvascular endothelial cells. Salahudeen et al. [38] developed a feeder-free, chemically defined culture system to improve model stability and ensure greater experimental reproducibility. Gerli et al. [39] demonstrated the feasibility of deriving LOs from fetal epithelial stem/progenitor cells isolated from human amniotic fluid in fetal lung development and congenital disease modeling. Moreover, they generated lung fetal organoids (LFOs) from amniotic fluid collected from fetuses diagnosed with congenital diaphragmatic hernia, successfully reproducing key molecular and pathological features of the condition. These advances provide valuable tools for investigating human lung development and contribute to studying congenital respiratory diseases.

Lung organoids should recapitulate lung tissue’s cellular composition and tissue structure characteristics and systematically integrate their biomechanical properties to simulate the biomechanical microenvironment accurately in vivo developmental context. During development, lung epithelial cells are continuously exposed to various mechanical signals, including shear stress, tensile stress, and matrix stiffness. These play a crucial role in lung branching morphogenesis and alveolar specification [40, 41]. For example, in the developing mouse lung, removing the mechanosensitive transcriptional activator Yes-associated protein (YAP) has been shown to affect epithelial cell proliferation and interfere with branching morphogenesis. During the lung bud formation stage, altering external tension can induce the formation of lung buds or cause ectopic branching [42]; in the alveolar formation stage, cyclical stretching activates the YAP/TAZ pathway, inducing the formation of alveolar type I cells (AT1), while shear stress determines whether progenitor cells differentiate into the flat AT1 or maintain a cuboidal alveolar type II cells (AT2) state [43].

Based on previous advances, recent studies have focused on incorporating biomechanical cues into lung organoids culture systems. Gjorevski et al. [44] pioneered tunable stiffness hydrogels to investigate how matrix rigidity influences stem cell fate. Expanding on this approach, Liu et al. [45] developed a tunable gelatin methacrylate hydrogel system tailored for lung organoids culture, enabling the exploration of how exogenous matrix stiffness modulates lung development. Integrating microfluidic chip platforms that mimic biomechanical forces, such as cyclic stretch from fetal breathing motions and shear stress generated by amniotic fluid flow, has enhanced LOs structural and functional maturation [46]. These findings underscore the essential role of biomechanical signals in lung morphogenesis and highlight their importance in advancing organoid functionality. In the future, developing multi-dimensional culture platforms that combine precise regulation of developmental signaling pathways with the recreation of biomechanical environments will be critical for promoting LOs maturation and physiological relevance.

Advances in biotechnology drive the development of lung-on-a-chip

LOC technology has continuously been refined since the landmark development of a biomimetic multi-chamber LOC replicating the alveolar-capillary interface by Hub et al. [24] in 2010 (Fig. 4A). The central design of the model incorporated a dual-channel microfluidic system divided by a porous polydimethylsiloxane (PDMS) membrane fabricated using soft lithography-based microfabrication techniques. Human alveolar and pulmonary microvascular endothelial cells were cultured on opposite membrane surfaces. The upper channel was exposed to air to establish an ALI culture, while the lower channel was continuously perfused with a blood-mimicking fluid. Cyclic mechanical stretching was applied to simulate the biomechanical forces of respiratory motion. Subsequent studies have explored alternative membrane materials, including parchment paper [47], poly(lactic-co-glycolic acid) (PLGA) electrospun nanofibre membranes [48, 49], and extracellular matrices(ECM) [50]. Guenat et al. [51] developed a second-generation LOC platform incorporating collagen–elastin composite biomembranes, which more accurately recapitulated the biomechanical properties of the alveolar network and basement membrane (Fig. 4B).

Fig. 4

Reproduced with permission from Refs. [24, 51,52,53]. Copyright 2010 AAAS; Copyright 2021 Springer Nature; Copyright 2018 Royal Society of Chemistry; Copyright 2018 IOP Publishing

A First-generation lung-on-a-chip developed by the A. Hub team. B Second-generation lung-on-a-chip incorporating a collagen-elastin biomembrane. C Lung-on-a-chip employing a thermoplastic gel-based biomembrane. D Advanced lung-on-a-chip constructed using a decellularized ECM combined with 3D printing technology.

Humayun et al. [52] developed a microfluidic airway-on-a-chip (AOC) platform using thermoplastic materials (Fig. 4C). Fabricated via micro-milling and solvent bonding techniques, this design replaces conventional PDMS with a suspended hydrogel layer. It consists of an epithelial cell layer, a suspended hydrogel layer, and a smooth muscle cell layer. The extractable hydrogel layer enables dynamic epithelial–smooth muscle cell interaction investigation. Park et al. [53] employed 3D printing to construct an in vitro AOC replicating the functional interface between the airway epithelium and vascular network (Fig. 4D). To replicate airway microenvironments with functional capillary networks, decellularized ECM from various tissues and organs was used as a bioink to fabricate cell-laden constructs. This approach allows precise modulation of microenvironmental conditions and seamless integration with vascularized platforms, resulting in an AOC that more accurately reflects the structural and functional properties of native lung tissue.

Advancements in bioengineering and biomaterials are continually improving the physiological accuracy of LOC, therefore strengthening their applicability in modeling respiratory diseases and facilitating drug screening efforts [54].

Construction and application of disease-specific in vitro models for respiratory diseases

We will provide a detailed description of the applications of lung organoids and lung-on-a-chip in the five most common respiratory diseases and explore how they contribute to the advancement of disease research and treatment. Table 1 summarizes the development history, main applications, and advantages and disadvantages of various disease models, offering important supplementary information for the subsequent discussion.

Table 1 Representative LOs and LOC for Common Respiratory Diseases: Timeline, Cell Composition, Applications, and Evaluation

Respiratory infections

The LOC technologies provide distinct advantages for modeling pathogen invasion and investigating host responses. These biomimetic platforms offer high fidelity in replicating interactions between pathogens and host cells and associated immune responses and disease pathogenesis. While LOs and LOCs have been employed across various areas of biomedical research, their value in studying respiratory infections gained widespread attention during the COVID-19 pandemic. Given the limited susceptibility of conventional animal models to SARS-CoV-2 and the ethical and logistical challenges posed by large animal studies, these advanced in vitro systems have emerged as vital alternatives for respiratory infection research.

Infection with SARS-CoV-2, the causative agent of COVID-19, compromises the pulmonary epithelial barrier, triggers excessive immune activation and apoptosis, and can ultimately lead to respiratory failure and multi-organ damage [55, 56]. Hashimoto et al. [57] developed an AOC model to understand its pathogenic mechanisms better and facilitate therapeutic development. Their findings revealed that the virus disrupts endothelial barrier integrity by downregulating the tight junction protein Claudin-5 (CLDN5) and impairing cadherin-mediated adhesion junctions. This suggests that restoring CLDN5 expression could offer a potential therapeutic strategy. In a separate study, Han et al. [58] employed LOs derived from hPSCs to model SARS-CoV-2 infection and replicate key pulmonary pathological features observed in COVID-19 patients. Single-cell RNA sequencing confirmed that AT2 cells express critical viral entry factors, including ACE2, TMPRSS2, and FURIN. To investigate host immune responses to SARS-CoV-2, Chen et al. [59] established a co-culture system combining macrophages differentiated from hPSCs with LOs. Their results demonstrated that M1 and M2 macrophages suppress viral replication. Only M1 macrophages significantly increase proinflammatory cytokine expression (e.g., IL-6 and IL-18), inhibit lung cell proliferation, and promote apoptosis. Importantly, a combination therapy involving ACE2-blocking antibodies and enhanced M2 macrophage activity nearly eliminated viral presence and protected lung cells from injury, offering promising new directions for immune-targeted interventions.

In addition to investigating the pathogenesis of SARS-CoV-2, LOs and LOC have become valuable platforms for drug discovery and screening. Capitalizing on the high-throughput capabilities of LOs, Han et al. [58] identified three promising antiviral candidates: imatinib, mycophenolic acid, and quinacrine hydrochloride which significantly suppressed SARS-CoV-2 infection. Similarly, Si et al. [60] employed an airway chip system to evaluate antiviral efficacy and successfully pinpointed amodiaquine, toremifene, and clomifene as potential therapeutic agents.

LOs and LOC systems have also been instrumental in studying other respiratory pathogens, such as influenza viruses, which are known for their high mutation rates and transmission potential. LOC technology has enabled detailed observation of viral infection dynamics and antigenic variation. It has been shown that influenza viruses can develop drug-resistant mutations under selective pressure from antivirals like amantadine and oseltamivir. In comparison, treatment with the host protease inhibitor nafamostat did not induce resistance [61]. Moreover, integrating lung organoids with organ chip platforms has advanced research on multi-pathogen interactions. In a co-infection model involving the influenza virus and Staphylococcus aureus, Stefanie et al. [62] demonstrated that co-infection provokes a heightened inflammatory response, compromises endothelial integrity, and significantly disrupts lung barrier function, effects not observed with single-pathogen infections. These findings suggest that co-infections may intensify disease severity and offer new insights into the complex mechanisms underlying polymicrobial respiratory infection.

In summary, LOs and LOC address many limitations of traditional respiratory infection models. They are robust platforms for dissecting pathogen mechanisms and host immune responses, opening new avenues for drug screening, personalized therapy development, and investigating complex infection dynamics. These models enable detailed observation of crucial biological processes, such as airway epithelial remodeling, cell migration, immune cell recruitment, cytoskeletal rearrangement, apical protrusion formation, and the development of syncytium-like structures, thus driving significant progress in both fundamental and translational respiratory infection research [63].

Asthma and chronic obstructive pulmonary disease

Although asthma and COPD arise from distinct pathogenic processes, they share critical features, such as dysfunction of the small airway epithelium, abnormal immune cell recruitment, and persistent inflammation [64]. These pathological changes commonly involve epithelial barrier disruption, impaired ciliary clearance, and continuous secretion of inflammatory cytokines, contributing to disease relapse and exacerbations. However, traditional animal models and simple single-cell cultures are limited in their ability to dynamically replicate these complex pathological hallmarks and accurately mimic the human airway microenvironment, therefore constraining mechanistic insights and the development of novel treatments.

To address this challenge, LOC technologies have emerged as vital platforms for investigating asthma and COPD by accurately replicating the complex physiological structures and functional states. Benam et al. [65] developed a small airway-on-a-chip (SAOC) incorporating differentiated ciliated epithelium and vascular endothelium. Treatment with IL-13 successfully modeled asthmatic airway inflammation, mucus overproduction, and ciliary dysfunction, while stimulation with lipopolysaccharide (LPS) and polyinosinic: polycytidylic acid (Poly I: C) induced chronic inflammatory responses characteristic of COPD. Expanding on this work, Nawroth et al. [66] designed a novel chip model to simulate virus-induced asthma exacerbation. Using IL-13-pretreated differentiated airway epithelium to recreate an inflammatory milieu combined with human rhinovirus 16 (HRV16) infection, they reproduced viral cellular pathology and immune responses, including temporal cytokine fluctuations (e.g., IL-6, IFN-λ1, CXCL10) and neutrophil trans-epithelial migration. This model demonstrated that the CXCR2 antagonist MK-7123 effectively reduced neutrophil infiltration under viral-inflammatory conditions, supporting its potential in immunomodulatory drug screening. Furthermore, the findings underscored how IL-13 disrupts coordinated host immune responses to viral infection, highlighting the pivotal role of Th2-type inflammation in asthma exacerbation.

As one of the leading causes of chronic disease mortality worldwide, COPD has attracted significant focus in studies of its pathological mechanisms [67, 68]. Chen et al. [69] developed a 3D human small airway model showing COPD features by optimizing an ALI culture system. This model incorporates key functional cell types, including basal, goblet, and ciliated cells, more accurately reflecting COPD’s structural and cellular composition. Their study revealed significantly abnormal nuclear localization of the autophagy-related protein LC3B in COPD-derived human small airway epithelial cells (HSAECs). It suggests that nuclear autophagy may regulate disease-associated cellular differentiation. Further investigation showed that chloroquine selectively preserves cytoplasmic LC3B activity, promotes ciliated cell differentiation, and enhances ciliary beating, whereas ivermectin predominantly stimulates goblet cell differentiation. Furthermore, co-expression of LC3B and ACE2 was detected in ciliated cells, indicating a possible interplay between these molecules in viral susceptibility and COPD pathology. These findings underscore the use of LOC for recapitulating the human respiratory microenvironment, enabling dynamic modeling of disease processes, and providing vital platforms for exploring therapeutic targets.

Similarly, organoid studies have identified complex and unique structures in the human distal airways, specifically the respiratory bronchioles and alveolar gas exchange regions absent in murine models. This area harbors a specialized secretory cell population known as respiratory airway secretory (RAS) cells, which serve as unidirectional progenitors differentiating into AT2 cells and play crucial roles in alveolar maintenance and repair. In COPD patients, RAS cell transcriptional profiles are disrupted alongside AT2 cell dysfunction, with this dysregulation closely linked to chronic smoking exposure [70]. LOs thus offer powerful insights into human lung-specific cell–cell interactions, transcriptional regulation, and disease mechanisms, opening new avenues for therapeutic development in chronic airway diseases like COPD [71].

In summary, LOC and LOs offer experimental platforms that more accurately replicate human physiology for investigating the mechanisms and therapies of asthma and COPD. These sophisticated models capture essential pathological characteristics and faithfully recreate the human respiratory microenvironment, establishing a strong foundation for mechanistic studies and advancing personalized treatment approaches for chronic airway diseases.

Pulmonary fibrosis

Pulmonary fibrosis is a chronic, fatal disease characterized by epithelial-mesenchymal transition (EMT), abnormal fibroblast proliferation, excessive ECM deposition, and stiffening of lung tissue, ultimately resulting in loss of lung function [72]. Due to limited insights into its pathogenesis, effective prevention and treatment options remain scarce [73]. To develop LOs and LOC that faithfully replicate the key features of PF, researchers commonly employ inducers such as bleomycin, paraquat, and TGF-β1 [74,75,76,77,78]. Xia et al. [75] created a LOC incorporating macrophages and pulmonary epithelial-mesenchymal tissues further to explore the disease’s pathological mechanisms through immune co-culture. Transcriptomic analysis of this model revealed that activation of the PI3K-AKT signaling pathway drives the progression from inflammation to fibrosis. Moreover, the PI3K inhibitor LY294002 suppresses fibrosis, reduces activated M2 macrophage accumulation, and intensifies inflammation severity. These findings offer a theoretical foundation for investigating macrophage roles in PF and developing macrophage-targeted therapies. To evaluate the pharmacodynamics of anti-fibrotic drugs, Hong et al. [76] constructed alveolar organoids containing functional macrophages derived from human pluripotent stem cells (Mac-AOs). Agents like pirfenidone, Nursol, and NP-011 effectively reduced TGF-β1-induced collagen buildup and fibrotic lesions in Mac-AOs by inhibiting extracellular signal-regulated kinase (ERK) signaling. Similarly, Suezawa et al. [77] used a comparable model to demonstrate that ALK5 inhibitors and integrin αVβ6 blockers strongly suppress fibrosis.

Cystic fibrosis (CF), a genetic disorder caused by mutations in the CFTR gene and classified as a subtype of PF, is characterized by impaired airway ciliary clearance, heightened inflammation, and persistent infection, ultimately resulting in respiratory failure. Plebani et al. [79] developed a CF airway chip by co-culturing primary bronchial epithelial cells from CF patients with pulmonary microvascular endothelial cells. This model faithfully reproduces key CF pathological features, including increased mucus buildup, decreased ciliary density, and elevated IL-8 secretion. It also promotes neutrophil adhesion to the endothelium and migration into the airway while creating a favorable environment for Pseudomonas aeruginosa proliferation, further intensifying inflammation and neutrophil recruitment. This platform provides a valuable preclinical tool for studying CF pathophysiology, evaluating therapeutics, and advancing personalized medicine approaches for CF-associated fibrosis.

In summary, PF models developed by exposing LOs and LOC to specific inducers successfully replicate multicellular interactions and the lung tissue microenvironment, accurately reflecting critical signaling pathway alterations and immune responses involved in fibrosis. As these models become more complex and controllable, they are poised to play an increasingly important role in advancing the diagnosis and treatment of PF.

Lung cancer

Lung cancer remains the leading cause of cancer-related mortality worldwide. According to a 2024 report by the International Agency for Research on Cancer (IARC), lung cancer was responsible for approximately 1.8 million deaths in 2022, representing 18.7% of all cancer fatalities [80]. Non-small cell lung cancer (NSCLC) accounts for about 85% of these cases [81]. Although targeted therapies and immunotherapies have made considerable progress in recent years, challenges such as acquired drug resistance and tumor heterogeneity continue to hinder their long-term effectiveness. These challenges underscore an urgent clinical need for more physiologically relevant in vitro models to unravel resistance mechanisms, assess drug responses, and facilitate the discovery of novel treatment options.

As advanced 3D biomimetic platforms, LOC and lung cancer organoids (LCOs) have demonstrated significant value in lung cancer drug discovery and the study of resistance mechanisms. LOC technology recreates the dynamic tumor microenvironment (TME) architecture using microfluidic systems, enabling the simulation of critical processes such as cell–cell interactions, signal transduction, and drug diffusion [82]. Tan et al. [83, 84] developed a three-cell co-culture LOC comprising the NSCLC cell line HCC827, human fetal lung fibroblasts (HFL-1), and human umbilical vein endothelial cells (HUVECs). Their findings revealed that IL-6 signaling drives HFL-1 differentiation into cancer-associated fibroblasts (CAFs), which promote EMT and induce resistance of HCC827 cells to the EGFR inhibitor osimertinib. Treatment with the IL-6 antibody tocilizumab partially reversed this resistance, highlighting IL-6 as a pivotal mediator of TME-driven drug resistance and suggesting that combining tocilizumab with EGFR-targeted therapies could offer novel strategies to overcome resistance in NSCLC. Yang et al. [49] further showed that insulin-like growth factor 1 (IGF-1) secreted by HFL-1 also contributes to resistance against the EGFR inhibitor gefitinib, emphasizing the broad role of fibroblast-derived paracrine factors in resistance development.

LCOs, by preserving the genetic background and tissue architecture of patient tumors, hold substantial promise for personalized drug screening. Since Clevers et al. [85] established the first lung cancer organoids in 2018, research has expanded its use in high-throughput drug testing, predicting drug responses, and screening combination therapies [86, 87]. Kim et al. [88] generated LCOs representing five histological subtypes of lung cancer as well as non-tumor bronchial mucosa, while Ebisudani et al. [89] created a biobank of 43 patient-derived LCOs from tissue, sputum, and blood samples, providing a valuable resource for personalized lung cancer research. Moreover, integrating tumor organoids with immune co-culture approaches offers novel means to evaluate immunotherapy sensitivity and resistance mechanisms. Voest et al. [90] developed an immune organoid platform enabling co-culture of patient-derived T cells with tumor organoids, effectively modeling tumor-immune interactions to assess tumor susceptibility to T cell-mediated cytotoxicity and immune evasion mechanisms. This platform does not require excised tumor tissues and can expand circulating tumor-reactive T cells from peripheral blood for functional assays. It also facilitates testing candidate small molecules or antibodies for combination therapies that enhance T-cell killing, providing a preclinical foundation to improve immunotherapy outcomes in NSCLC.

LOC and LCOs are complementary in elucidating resistance mechanisms and advancing drug development, accelerating the clinical translation of personalized lung cancer therapies.

Multi-organ-on-a-chip (MOC) systems have recently emerged as highly physiologically relevant platforms for studying metastatic lung cancer and drug screening. By linking lung organoids or chips with models of the brain, liver, bone, and other organs, these systems simulate metastatic pathways and inter-organ interactions involved in distant cancer spread. Metastatic lung cancer, a leading cause of lung cancer mortality, frequently affects multiple organs such as the brain, bone, and liver and is characterized by significant heterogeneity and drug resistance. Traditional models fail to systematically replicate cancer cell migration, organ colonization, and drug responses across different sites, limiting the detailed investigation of metastatic mechanisms and combination treatment strategies. MOC systems overcome these challenges by connecting lung and distant organ modules (e.g. brain, liver, and bone) via microfluidic platforms, enabling dynamic metastasis modeling and simultaneous drug response assessment at multiple metastatic locations. Xu et al. [91] developed a multi-organ microfluidic chip system featuring an upstream lung tissue chamber that models primary lung cancer and three downstream culture chambers representing brain, bone, and liver metastases. Constructed with three PDMS layers and two microporous membranes, this system supports the parallel culture of bronchial epithelium, lung cancer cells, endothelial, immune, and stromal cells, astrocytes, osteoblasts, and hepatocytes. This model successfully recapitulates the full metastatic cascade of lung cancer and promotes tumor spheroid formation, and findings validated through in vivo imaging in nude mouse models. It offers a powerful platform to dissect metastatic pathways and resistance mechanisms. Furthermore, MOC platforms are well-suited for high-throughput drug efficacy and toxicity evaluation. Zhu et al. [92] created a microphysiological system chip (MSCP) integrating intestinal, liver, heart, and lung cancer modules to mimic in vivo drug absorption, distribution, metabolism, and excretion processes. This system enables concurrent assessment of the efficacy and toxicity of frontline lung cancer drugs such as cisplatin, pemetrexed, docetaxel, and doxorubicin, significantly improving the efficiency and predictive accuracy of candidate drug screening.

In summary, LOC and LCOs are complementary in unraveling drug resistance mechanisms and driving novel drug development for lung cancer. LOC platforms excel at dynamically replicating the TME to uncover microenvironment-driven resistance, while LCOs emphasize tumor genetic heterogeneity and enable personalized drug efficacy evaluation. The advent of multi-organ chips and immune organoid systems further enhances the complexity and functionality of these models. Coupled with ongoing advances in drug delivery technologies, high-throughput screening, and immune co-culture techniques, these platforms are increasingly important tools to accelerate precision medicine and clinical translation in lung cancer therapy [93].

Conclusion and future perspectives

Despite the promising potential of LOs and LOC technologies in respiratory disease research and drug development, their clinical translation remains hindered by several key challenges. Advancing toward practical applications will require ongoing technological innovation and strong interdisciplinary collaboration. For example, the lack of essential cell types, such as pulmonary neuroendocrine cells (PNECs), in current models limits their ability to replicate critical physiological processes. These cells secrete hormones like serotonin, calcitonin, and bombesin, which are crucial for various functions, including platelet formation. Elevated serotonin levels, often seen in COVID-19 patients, have been linked to thrombosis and cardiovascular complications. Addressing this gap in cell composition will be crucial for advancing disease modeling and therapeutic development. Furthermore, enhancing the physiological relevance of organoid models requires the integration of vasculature, immune system components, and other critical features often missing in current systems. Achieving these advancements will require the incorporation of cutting-edge techniques, such as co-culturing iPSC-derived endothelial progenitor cells with immune cells, to create more functional and representative models [94,95,96,97]. Current validation methods for LOs rely largely on the expression of specific genes and proteins, which fall short of capturing the full functional complexity of human organs. Furthermore, LOs are characterized by increased cellular heterogeneity and complex self-organization processes. However, they frequently lack critical features such as a complete extracellular matrix, functional vasculature, and immune system component factors that limit their structural maturation, physiological relevance, and long-term viability. To overcome these limitations, co-culturing iPSC-derived endothelial progenitor cells with tissue-resident immune cells, such as alveolar macrophages, could enable the development of LOs with functional vasculature and immune responsiveness [98]. This strategy will support future efforts to model lung injury repair, immune responses to infection, and multicellular interactions within the TME. By incorporating hydrogels or microfluidic platforms, it is possible to create a multi-dimensional culture system replicating developmental signaling pathways and biomechanical cues, promoting LOs maturation and functional refinement.

The design of interfaces and hydrodynamic parameters such as shear stress and perfusion rate in MOC platforms remains unstandardized, resulting in limited comparability of experimental outcomes across different systems. To address this, it is crucial to develop standardized interface protocols and compile a comprehensive database of hydrodynamic parameters for various organ modules. Furthermore, integrating artificial intelligence algorithms can optimize the dynamic regulation of MOC systems, including the reconstruction of physiological oxygen gradients and pharmacokinetic predictions [99,100,101].

Integrating light-sheet fluorescence microscopy, organ-on-a-chip platforms, and machine learning can enable high-throughput 3D imaging and analysis of organoid development and drug responses. Moreover, the preservation conditions of LOs significantly affect their efficacy. Building on preliminary studies, developing cryopreservation techniques to extend the viability of LOs will support the large-scale establishment of biobanks [102,103,104,105].

One of the key challenges in current disease research is the limited ability to replicate inter-organ interactions within the human body accurately. To overcome this limitation, MOC platforms that integrate organoid and LOC technologies have been developed. These "patient-on-a-chip" systems enable the reconstruction of molecular and signaling networks across multiple organs, providing a novel framework for investigating cross-organ disease mechanisms and drug metabolism pathways [106, 107]. Moreover, the development of axis-on-a-chip models enables the exploration of inter-organ signaling pathways and their underlying mechanisms in complex conditions, such as metabolic disorders and neuroinflammation. These models offer innovative approaches to addressing critical regenerative and precision medicine challenges.