What Can an Organ-on-a-Chip Teach Us About Human Lung Pathophysiology?

Abstract

The intertwined relationship between structure and function has been key to understanding human organ physiology and disease pathogenesis. An organ-on-a-chip (organ chip) is a bioengineered microfluidic cell culture device lined by living cells and tissues that recapitulates organ-level functions in vitro. This is accomplished by recreating organ-specific tissue-tissue interfaces and microenvironmental biochemical and mechanical cues while providing dynamic perfusion through endothelium-lined vascular channels. In this review, we discuss how this emerging technology has contributed to the understanding of human lung structure-function relationships at the cell, tissue, and organ levels.

Introduction

Effective characterization of biological structure-function relations requires availability of experimental models that incorporate a sufficient amount of structural information while still allowing functional data to be interpreted with reasonable simplicity (1). This represents a major challenge during physiological investigation of the human lung, which is an organ that consists of >50 heterogeneous types of cells (2) that are arranged into specific architectural units with the primary function of gas exchange between the air and blood. The complexity is heightened further by the fact that the lung is a highly dynamic organ such that its constituent tissues and cells experience a variety of mechanical forces, including cyclic deformation (strain), hydrostatic pressure, and shear stress due to flow of both air and blood (3). Moreover, the influence of these mechanical cues can be significantly impacted by the relative stiffness or compliance of the extracellular matrix (ECM) that underlies the lung epithelium and endothelium and surrounds cells within the intervening connective tissue (4).

The rise of molecular biology and pursuit of therapeutic interventions for human diseases have elevated cellular physiology and the investigation of molecular mechanisms to the forefront of physiology research. This trend is underpinned by the development of human two-dimensional (2-D) culture models that enable high-throughput cell biological analyses and have resulted in some drug discovery success. For example, the use of patient-derived human bronchial epithelial cells contributed to the development of cystic fibrosis (CF) treatments (5). However, the limitations of conventional in vitro culture technology in modeling complex responses, disease phenotypes, and pharmacological behaviors necessitate additional testing in animal models (6), especially genetically engineered mice models that preserve in vivo structural and functional complexity while enabling investigation of the physiological impact of particular gene defects. This research paradigm has been a core pillar of basic biomedical research and preclinical drug discovery for the last 40 years (6). However, the extremely low success rate of preclinical-to-clinical translation has highlighted the genetic, structural, and functional differences among species. For example, results from mouse models fail to predict human responses to inflammatory diseases (7), viral vector-mediated gene therapy (8), and side effects of RNA vaccines against COVID-19 (9). The human lung also differs substantially from its mouse counterpart in terms of structure, constituent cell types, and patterns of gene expression (2, 10, 11). Thus, it is imperative to develop models of the human lung that balance simplicity with the need for relevant organ-level structures and functions, which are required to generate adequate functional information that is directly relevant to lung physiology.

Human organoids and organ chips are two technologies that have emerged over the last decade to address this need. Although they employ distinct design principles, both aim to replicate critical structural and functional features of an organ in vitro with three-dimensional (3-D) cell culture. Organoids are generated by culturing stem cell-derived or tissue progenitor cells within an ECM gel or synthetic hydrogels, where they differentiate and compartmentalize into closed spheroids that exhibit tissue structures with forms that often resemble those seen in vivo. For example, lung alveolar organoids and airway organoids lined by a differentiated lung epithelium have been generated and applied to study lung regeneration, cell fate decisions, and disease processes in vitro (12, 13). Organ chips, on the other hand, build upon the idea that development of differentiated features at the cell and tissue levels alone is not sufficient to fully replicate organ-level responses and that reconstitution of organ-level structures in a physiologically relevant microenvironment containing organ-specific mechanical cues is required as well, as recent evidence suggests that extrinsic microenvironmental controllers complement cell self-organization to facilitate spatial and temporal organoid patterning (14). By using an engineering approach to reproduce cellular scaffolds, multicellular structures, tissue-tissue interfaces, physicochemical microenvironments, and circulatory perfusion, organ chips can achieve tissue- and organ-level complexity, maturity, and functionality that are not attainable with standard 2-D or 3-D culture techniques, including organoids (15, 16).

Organ chips have been utilized to model numerous human tissues, organs, and organ systems, including lung, heart, gastrointestinal (GI) tract, liver, kidney, and brain (15, 17). More general discussions on the use of organ chips in basic biological research, disease modeling, drug development, and personalized medicine can be found in recent reviews (15–17). Here, we focus on the lung organ chip, the first described multitissue organ chip (18) and most well-characterized system. We begin by summarizing lung chip designs with varying degrees of structural complexity for mimicking the lung alveolus or the conducting airway. Then, we describe new functional insights into human lung pathophysiology at the cell, tissue, organ, and organism levels that these models have enabled. Finally, we discuss challenges and future opportunities so that the full potential of organ chips in advancing lung biology and pulmonary medicine can be achieved.

Human Lung Chips

Although the concepts of tissue engineering and “human-body-on-a-chip” have been around since the late 1990s (16), the first microfluidic culture device that recapitulated organ-level structures (FIGURE 1A) and functions of a major functional unit of a human organ—the distal lung gas exchange unit or lung alveolus—was described in Science in 2010 (18). This lung alveolus chip is composed of optically clear, elastomeric polydimethylsiloxane (PDMS) material that contains two parallel hollow channels separated by a porous ECM-coated membrane lined with human lung alveolar epithelial cells on one side and with human vascular endothelium on the other, recreating the native gas-exchanging alveolar-capillary interface (FIGURE 1B). Culture medium is flowed through the endothelium-lined channel to mimic vascular perfusion, and air is introduced into the epithelial channel to mimic the air-liquid interface (ALI). In addition, cyclic suction is applied to hollow side chambers to induce the flexible tissue-tissue interfaces that extend and retract to the same degree and at the same rate as the alveolus experiences because of lung breathing motions. All these elements combined result in a microfluidic microphysiological system that enables unprecedented structural and functional recapitulation of the human distal lung. This reductionist approach to reconstructing the structure and function of human organs created a conceptual framework for the organ chip field. Since then, various organ chip models of the lung and other organs have been designed, fabricated, and used for preclinical disease modeling and drug testing (15). Some have even greater cellular and structural complexity to achieve a higher level of in vivo mimicry, whereas others sacrifice structural complexity for throughput (21).

FIGURE 1.

Human alveolus chips designsA: a human alveolus structure showing major cell types, including alveolar type I cells (ATI), alveolar type II cells (ATII), microvascular endothelial cells, stromal cells, and alveolar macrophage (AM). B: a well-established 2-channel, mechanically actuable alveolus chip. Adapted from Ref. 15 with permission from Nature Reviews Genetics. C: a simple 1-channel alveolus chip with cyclic airflow, simulating cyclical stretching while breathing. Additionally, the breathing process displays cellular mechanical images. AC, alveolar cell; Fb, fibroblast; NP, nanoparticle; PDMS, polydimethylsiloxane. Adapted from Ref. 19 with permission from Advanced Materials. D: a 3-channel Open-Top alveolus chip contains an alveolar epithelium, an endothelium lining the lower spiral-fluidic channel, and a stroma compartment separating these 2 channels. Modified from Ref. 33 with permission from Biomaterials. E: example of a complexed alveolus chip model aimed to mimic 3-dimensional (3-D) architecture of the human alveolus. Reproduced from Ref. 20 with permission from Proceedings of the National Academy of Sciences USA.

Human Lung Alveolus Chips

Initial versions of the two-channel alveolus chip model used an established human alveolar epithelial cell line (A549 or NCI-H441) and human umbilical vein endothelial cells (18, 22). The model has since been improved by replacing cells with more relevant primary cells from commercial vendors (23–27), human cadaver lungs (24), or organoid cultures (28). These alveolar chip models have been used to study pulmonary nanotoxicology (18, 29), bacterial infection (18), drug toxicity-induced pulmonary edema (22), orthotopic lung cancer growth (25), intravascular thrombosis (23), and infection by influenza (24, 30) and SARS-CoV-2 (26, 27, 30, 31) viruses.

Single-channel microfluidic devices that only contain alveolar epithelial cells have been developed as well. One model used an elastic membrane consisting of poly(ε-)caprolactone and gelatin that has similar stiffness (Young’s modulus) as the basement membrane of the lung alveolus, to mimic the microenvironment of human alveolar epithelium (32). Despite lacking endothelium, this microfluidic model can be useful for studying stretch-induced functional changes of alveolar epithelium. Another single-channel alveolus chip enables real-time monitoring of cellular mechanics during breathing movements by integrating photonic nanoparticles in the underlying membrane on which the alveolar epithelial cells were seeded (19) (FIGURE 1C). This model allows the investigation of dynamic relationships between cellular deformations and disease phenotypes, such as idiopathic pulmonary fibrosis (IPF) (19).

More complicated models have been created by incorporating additional cell types of the distal lung, such as stromal cells. To accomplish this, the most frequently used design incorporates a middle layer seeded with ECM gel alone or with lung fibroblasts (33) (FIGURE 1D). An open-top chip also has been engineered that allows analysis of stromal-epithelial interactions and cellular cross talk between the three chip compartments. Similar 3-D alveolar barrier models have also been fabricated by inkjet printing alveolar epithelial cells, lung fibroblasts, ECM, and lung microvascular endothelial cells (34), although vascular perfusion and cyclic mechanical strain are absent in this static model. Another strategy for increasing complexity focused on modeling not only the tissue interface but also the 3-D architectures of the alveolar air sac. A biological, stretchable, and biodegradable membrane made of collagen and elastin was cast on a thin gold mesh to emulate an array of tiny alveoli with in vivo-like dimensions (35). Similarly, a 3-D porous hydrogel made of low-stiffness gelatin methacryloyl (GelMA) has been used to construct an inverse opal structure, which is bonded to a compartmentalized PDMS chip device that provides the medium supply, ALI, and the cyclic mechanical movements (20) (FIGURE 1E). Although these more complex chips provide more accurate 3-D structural representations of the in vivo lung alveolus than the two-channel models, it remains to be demonstrated whether they can provide additional novel biological insights.

Human Lung Airway Chips

Human lung airways are divided into two compartments, the larger conducting airways (bronchi) and the smaller respiratory airways (bronchioles), and organ chips have been designed to model both compartments (FIGURE 2A). The two-channel microfluidic design used in the alveolus chip was adapted to create a human lung small airway chip that supports full differentiation of a columnar, pseudostratified, mucociliary bronchiolar epithelium (39). This model was successfully used to model disease processes associated with the small airway, including asthma (39, 40), chronic obstructive pulmonary disease (COPD) (39), and the effects of exposure to cigarette smoke (41). More recently, the chip was modified to contain a membrane with different chemistry (PDMS vs. polyethylene terephthalate) and material properties (flexible vs. rigid) and larger pores (7 mm vs. 0.4 mm) to culture human bronchial airway basal cells and to model the human large bronchial airway (30) (FIGURE 2B). This model has proven to be powerful for studying viral-host interactions (24, 30), modeling disease processes related to CF (42), identifying candidate antiviral therapeutics (30), testing drug efficacy of novel biologics including RNA therapeutics (43), and assessing influenza viral evolution in responses to antiviral drug treatment in vitro (30, 37). Various human diseased airway chips also have been generated by populating chips with primary bronchial or bronchiolar epithelial cells obtained from patients with lung diseases such as COPD (39) or CF (42). These chips have been shown to recapitulate disease-specific phenotypes, for example, selective cytokine hypersecretion, increased neutrophil recruitment, and clinical exacerbation by exposure to viral and bacterial infections (39, 42).

FIGURE 2.

Human airway chips designsA: human airway consists of many distinct cell types that form a pseudostratified layer. Adapted from Ref. 36 with permission from Nature Reviews Molecular Cell Biology. B: a well-established 2-channel human bronchial/airway chip, reproduced from Ref. 37 with permission from Microbiology Spectrum. C: a simple and higher-throughput plastic organ chip that includes 2 parallel channels separated by a rigid microporous membrane. Adapted from Ref. 15 with permission from Nature Reviews Genetics. D: a complexed bronchiole model with air in the center, epithelial cell-lined lumen, and media in the 2 side endothelial cell-lined lumens. ALI, air-liquid interface; PDMS, polydimethylsiloxane. Reproduced from Ref. 38 with permission from Nature Communications.

A simplified one-channel microfluidic device composed of proximal airway epithelial progenitor cells derived from human induced pluripotent stem cells (iPSCs) enabled the investigation of fluid shear stress during ciliated cell development and disease modeling for primary ciliary dyskinesia (44). The use of gravity-assisted media exchange instead of perfusion pumps also allowed for an increase in throughput. For example, a human small airway chip employing this method allowed high-throughput studies of direct gas/aerosol exposures to inhaled toxicants (45). A streamlined plastic 384-well format organ chip design also was described recently and used to create a human bronchial airway cultured under an ALI (46) (FIGURE 2C). Although this chip lacked endothelial cells, the model enabled high-throughput infection and drug screening against respiratory viral infections (46).

More complicated models have been created by incorporating other relevant cell types in the conducting airways, such as fibroblasts and smooth muscle cells (SMCs). A common design approach for multilayer microfluidic airway chips is to use a three-channel configuration: a top channel for airway epithelial cells, a middle channel for interstitial fibroblasts, and a bottom channel for microvascular endothelial cells (47). A similar three-compartment design was used to incorporate SMCs by replacing the middle channel with a thin hydrogel layer for ECM and a bottom medium reservoir for SMC culture (48). Another model comprises three hollow channels within a 3-D collagen matrix containing pulmonary fibroblasts; the central channel is lined with primary human bronchial epithelial cells cultured under ALI, whereas the two flanking channels are lined with primary human lung microvascular cells to form the vascular compartment (38) (FIGURE 2D).

A more complicated airway chip disease model was created to study idiopathic pulmonary fibrosis (IPF), in which healthy or IPF primary fibroblasts were cultured in a fibrin-collagen hydrogel filling one channel in parallel to an airway epithelial-endothelial barrier. This model revealed communication between the interstitium and the airway epithelium as well as fibrotic changes in club cells and ciliated cells in the airway epithelium (49). In addition to a single endothelial channel, 3-D printing using a decellularized ECM bioink has been used to encapsulate and print endothelial cells and fibroblasts to form a vascular network in an airway chip (50). Physical forces such as bronchial compressive stress can be introduced to the airway chip as well. Another bronchial airway chip model reconstitutes 3-D cocultures of primary human airway SMCs and bronchial epithelial cells that are fully differentiated under an ALI (51). In addition, physiologically relevant compressive stress was applied to the pseudostratified epithelium by regulating air pressure, and the resulting changes in mechanical properties of SMCs were quantified in real time with optical magnetic twisting cytometry (51). This model can be used to study the interplay of mechanical and biochemical signals in bronchospasm. Finally, by linking a nasal, a bronchus, and an alveolar chip through the air channel, a multicompartment human airway-on-chip platform was fabricated to emulate the airborne route of respiratory pathogens and to investigate regional lung cross talk for viral infection pathways (52).

Human Organ Chips Inform Lung Physiology at the Cell Level

Among the >50 cell types found in the human lung, only a small subset has been well studied (2), and most of these are epithelial cells. Located at the interface between the environment and the organism, epithelial cells serve many important functions including barrier protection, fluid balance, clearance of particulate, initiation of immune responses, production of mucus, and regeneration after injury. Although much of our understanding of these cells has come from mouse studies, recent work using human organ chips has revealed some unique insights.

The two major epithelial lineages within the alveoli are alveolar type 1 (AT1) and AT2 epithelial cells. The thin flat AT1 cells that cover >95% of the lung surface area are responsible for gas exchange, whereas the less frequent cuboidal AT2 cells produce pulmonary surfactant to preserve alveolar surface tension as well as acting as progenitors for AT1 cells. Several studies from human alveolus chips have found that cyclic mechanical strain (which mimics physiological breathing motions) increases surfactant production, decreases cell proliferation, and enhances ECM deposition (18, 19, 24, 53). Cyclic strain also increases nanoparticle absorption across the alveolar-capillary barrier (18), enhances interleukin-2 (IL-2)-induced pulmonary edema via a signaling pathway involving the transient receptor potential vanilloid-type 4 (TRPV4) mechanosensitive ion channel (22), and suppresses lung cancer cell growth and invasion (25).

More recently, physical forces associated with cyclic breathing motions also were found to stimulate lung innate immune responses to viral infection in alveolar epithelial cells and microvascular endothelial cells (24). RNA sequencing (RNA-seq) analysis revealed that TRPV4 and the receptor for advanced glycation end products (RAGE), an inflammatory mediator and marker for ATI cells, play central roles in this response (24). Modulating these pathways may offer great therapeutic potential against viral pneumonia, ventilator-induced lung injury, and other inflammatory respiratory diseases. Moreover, although physiological strain was protective, application of hyperphysiological strains on-chip revealed that they can induce cell apoptosis, activate inflammatory responses, and impair barrier integrity (24, 32) as experienced by patients on mechanical ventilators or with abnormal lung compliance.

The trachea and proximal airways are lined by a pseudostratified epithelium that contains multiple cell lineages, including multiciliated cells, secretory cells, goblet cells, and basal stem cells, as well as many other less abundant cell types, such as neuroendocrine cells (54). The two-channel human small airway chip enables the generation of cilia with the same structure, length, beating frequency, and transport velocity as their in vivo counterparts on the apical surface of the polarized epithelium (39). In addition, fluid shear stress applied through the vascular channel promotes synchronized cilium beating and coordinated mucociliary movement (39). A recent study using a monoculture of iPSC-derived airway epithelial cells exhibited similar behavior even though cells were cultured under a submerged condition, which was previously regarded as not suitable for inducing cilium differentiation (20). It is possible that fluid shear stress can compensate for ALI through an unidentified mechanotransduction mechanism. Investigating how different types of physical stresses alter differentiation, gene expression, and responses to external stimuli in lung epithelial and endothelial cells on-chip will provide critical insights into lung mechanobiology.

Human Organ Chips Inform Lung Physiology at the Tissue Level

The respiratory system in its totality is composed of multiple tissue systems, including the epithelial, endothelial, mesenchymal, and immune components. So far, the greatest understanding from studies of lung organ chips has been obtained for the epithelial tissues. Lung epithelium serves as the first-line defense against environmental pathogens by creating a relatively tight permeability barrier. Human lung organ chips are well suited to study the formation, damage, and repair of this critical tissue interface. Barrier integrity can be measured by quantifying permeability of fluorescent tracers of different sizes or using integrated electrical sensors to measure transepithelial electrical resistance (TEER) on-chip (before these assays, medium needs to be introduced to the air channel in chips under ALI culture conditions) (55, 56). Tissue barrier permeabilities measured within two-channel lung chips are similar to those measured in vivo (18, 41). Compromised barrier function also can be detected on-chip when exposed to a number of insults that are known to contribute to acute lung injury, such as increased levels of IL-2 (22) and TNF-α (23), as well as exposure to bacterial lipopolysaccharide endotoxin (LPS) (23) or infection by influenza virus (24, 30). With easy access to both the air and blood sides of the lung barrier on-chip, it is possible to determine whether any of these insults lead to direct (pulmonary) or indirect (extrapulmonary) lung injury in vitro. For example, LPS causes significant barrier disruption only when dosed in the air channel of the alveolus chip, supporting the hypothesis of direct epithelial injury (23). Interestingly, LPS also induced vascular thrombosis in this model via its effect on the epithelium, as it had no effect on coagulation when infused directly within the endothelium-lined vascular channel.

Both gas-exchanging alveoli and conducting airways are capable of regeneration after injury. However, our understanding of this process remains limited, with much of our knowledge derived from studies of the mouse trachea, which most closely resembles the structure of human proximal airway (57). Human lung organ chips are beginning to produce valuable insights into this process. Similar to in vivo (58, 59), alveolar epithelial cells and endothelial cells are relatively quiescent on-chip but proliferate rapidly upon influenza-induced lung injury (24). Future work is needed to demonstrate whether the airway has similar regenerative capacity on-chip after viral infection.

Human organ chip experiments also have demonstrated airway cell plasticity. Goblet cells constitute ∼10–20% of the total cell population in the human small airway epithelium, but the percentage increased significantly after exposure of the epithelium to interleukin-13 (IL-13) (39). This goblet cell hyperplasia mimics the clinical presentation in the airway mucosa of individuals with asthma (39). The application of recently developed lineage tracing methodologies, such as cellular barcoding (60), to the organ chip field is likely to shed fresh light on cell fate determination during lung development and regeneration.

Human Organ Chips Inform Lung Physiology at the Organ Level

The ability to coculture multiple tissues and cells in an organ-relevant structural context represents the greatest advantage of human organ chips. This allows exploration of cellular cross talk that is vital for the development, injury response, and regeneration of the respiratory system. The process involved in forming the tissues in a lung chip in vitro are also reminiscent of lung developmental processes that occur in vivo. For example, time course RNA-seq analysis suggests significant epithelial-endothelial signaling cross talk that may contribute to distal lung epithelial maturation and vascularization under these conditions (24). Future work is required to examine whether any of the signaling axes that were previously reported in mice (e.g., BMP4-NFATc1-thrombospondin-1, MMP14-EGFR; Ref. 61) are also utilized during tissue differentiation in human lung chips. Another microphysiological model of the airway has revealed cellular cross talk between airway epithelial cells and smooth muscle cells (51). A mechanical stimulus mimicking a bronchospastic challenge was shown to trigger a marked contraction and delayed relaxation of airway smooth muscle. This was mediated by the discordant expression of cyclooxygenase genes in epithelial cells and regulated by the mechanosensor and transcriptional coactivator Yes-associated protein.

Perhaps the use of human organ chips in elucidating tissue cross talk is best exemplified by recent work that used lung chips to study respiratory viral infections. When respiratory viruses, such as influenza virus or SARS-CoV-2, infect airway or alveolar epithelial cells in lung chips, significant inflammation and injury also can be observed in endothelial cells, even though these viruses do not infect endothelial cells in monoculture (24, 26, 27, 31). At least in the case of influenza virus infection, these effects on endothelium are likely due to paracrine signals from the infected epithelial cells, as viral RNA was not detected in the endothelium (24). Alternatively, in the case of SARS-CoV-2 infection, it may be due to mitochondrial dysfunction and activation of the cyclic GMP-AMP synthase (cGAS)-stimulator of interferon genes (STING) pathway (27). The activation of endothelial cells then initiates a series of responses, including the adherence of neutrophils, monocytes, B cells, and T cells to the endothelium and their subsequent transmigration into the air space, which can further amplify release of inflammatory cytokines (24, 30, 31). This vicious cycle can lead to hyperinflammation that causes irreversible tissue damage and loss of lung function (62). A number of immunomodulatory drugs have been tested and showed efficacy in disrupting this process on-chip, including IL-6 antibody tocilizumab (26) and the RAGE inhibitor azeliragon (24). The use of organ chip technology to identify targets that disrupt cellular communication should result in the identification of more effective therapies against human infectious diseases.

Human Organ Chips Inform Lung Physiology at the Organism Level

Compared with other 3-D tissue culture models, one of the greatest advantages of organ chips is their ability to study interorgan communications because they contain vascular channels that experience dynamic flow. This is achieved by fluidically linking the vascular channels of different organ chips to form the so-called human “body-on-chips” (15, 63, 64). The most frequently used application of such multiorgan chip systems is to mimic drug absorption, distribution, and clearance via metabolism or excretion. When combined with mass spectrometry and computational simulation, these body chip systems may offer a replacement to animal models to measure drug pharmacokinetics (PK) and pharmacodynamics (PD) (64–69). By linking the lung with other organs such as the liver, it is possible to evaluate drug delivery and PK/PD of aerosolized small molecular drugs or other therapeutic modalities (70, 71). For example, a multiorgan system composed of the lung, liver, and breast cancer that was created to mimic inhalation therapy of curcumin revealed that the air route is more effective and allows more frequent dosing than intravenous delivery in the same system (72).

Human Organ Chips Inform Lung Microbe-Host Interactions

The lung is susceptible to infection by a wide variety of microorganisms, including viruses, bacteria, and fungi. Human lung chips have been used to model pathogen-host interaction associated with these pathogens. Besides studying host responses and tissue cross talk during infection as discussed above, lung chips can also be used to study how the host shapes the viral evolution landscape. For example, serial passaging of influenza virus between multiple human airway chips (mimicking human-to-human transmission) revealed mutations that have been observed in human transmission of influenza viruses, as well as novel ones that have not been previously reported in human clinical isolates (37). Recent research also suggests that ferrets or mink (which are often used to model viral infections) are poor models for human lung, particularly when investigating spike RBD mutants, as mutations that adapt SARS-CoV-2 to mink or ferret do not increase fitness in the human airway (73). Because the human airway chip enables more faithful mimicry of host innate immune response and host selection pressure compared with traditional cell line models or animal models, it provides a better approach to examine and predict antigenic shift and antigenic drifts, which could inform future vaccine designs (37).

In addition to modeling viral infection, lung chips have also been used to model various bacterial infections, such as those caused by Pseudomonas aeruginosa (42) or Mycobacterium tuberculosis (Mtb) (74); a similar approach has been used to study effects of aerosolized LPS droplets that mimic bacterial infection (75). These studies have provided many insights into host defense. For example, a two-channel lung chip revealed that pulmonary surfactant plays an important role in host innate immunity during early Mtb infection (74). Recently, a human alveolus chip was used to model Aspergillus fumigatus infection, and it was shown that human macrophages partially inhibited the growth of the fungus and contributed to the release of proinflammatory cytokines, which was associated with an increased number of invasive hyphae (76). Organ chips can even be used to study multikingdom (bacterial-fungal-human) human microbiome and their interaction with the host, as demonstrated by a human bronchiole model containing airway, vascular, and ECM compartments (38).

Challenges and Future Directions

During the first decade of the organ chip field, much attention has been paid to microfluidic device design and engineering and demonstration of their capacity to emulate key tissue and organ functions. The advancements discussed above illustrate that human organ chips can be utilized to address a wide range of questions relevant to human lung physiology and pathophysiology at scales from single cells to body-level functions (FIGURE 3) that could not be addressed with conventional culture systems, static microphysiological systems, or even animal models. Moving forward, a few obstacles must be addressed before the full potential of organ chips in lung biology research can be realized. Apart from typical issues related to all organ chip models, such as procurement of high-quality human cells, long-term culture, alternate flexible elastomeric materials, online readouts, and benchmarking (15, 16), there are a few questions that are particularly pertinent to the lung.

FIGURE 3.

Organ chip models human lung physiology across different scales Human organ chips offer a versatile way of simulating human lung physiology at the cell, tissue, organ, organ system, organism, and microbe-host levels. ECM, extracellular matrix; PD, pharmacodynamics; PK, pharmacokinetics.

The first question is whether we can leverage organ chips to better define intercellular cross talk between the cells that comprise the multiple different tissues of human lung. Although many of the existing lung chip models still largely focus on the epithelium, other tissue types are equally important for maintaining lung structure and function. For example, endothelial cells are involved in the differentiation of alveolar epithelial cells, as well as in alveolar regeneration after acute injury (77–80). A growing body of evidence also indicates that stromal cells and the ECM play a critical role in the development and pathogenesis of interstitial lung diseases, such as pulmonary fibrosis (36, 81–83). However, none of the animal models that are commonly used in current preclinical research provides a faithful representation of pathophysiology of this fibrotic disease (84). The three-channel chip design that incorporates an ECM gel seeded with fibroblasts may represent a promising tool to study this process (33). Another component that is yet to be included on lung chip models is immune cells located in the lung parenchyma, such as alveolar macrophages and tissue-resident memory immune cells, both of which are critical in the host response to respiratory pathogens (85, 86). Finally, the potential to mimic anatomical regions unique to humans may provide a significant benefit in studying the physiology and disorders linked with those regions, such as the submucosal gland and the bronchiole.

Is it possible to model the heterogeneity of the lung with an organ chip? The lung is extremely heterogeneous in terms of its cell composition, topography, and mechanical forces acting on the cells. The variability of lung cell types was only recently mapped out with advances in single-cell analysis techniques. These studies have uncovered previously unidentified populations of cells, such as general capillary endothelial cells and aerocytes that are specialized for gas exchange and leukocyte trafficking inside the capillary endothelium (87). Although it is unrealistic and perhaps also unnecessary for organ chips to represent the full cellular variability of intact organs, characterizing cell heterogeneity on existing chip models or incorporating rare cell types, such as neuroendocrine cells (54) and ionocytes (88), in future models would enable more in-depth analysis of relevant biological questions (89). Similarly, engineering chip models that recapitulate 3-D tissue architecture may enhance their ability to mimic in vivo structural and mechanical anisotropy of lung tissues that are important for its function (20, 90).

Conclusions

The rapid growth of the organ chip field over the last decade has demonstrated that this technology is capable of not only recapitulating what we already know about human lung pathophysiology but also shedding light on previously unknown mechanisms. Many human organ chips have been designed and fabricated to model key functional units of the human respiratory system. However, to fully exploit the potential of lung chips in physiology research, it is critical that these tools be used to address biological questions on a wider scale. There are a few hurdles to this endeavor. Besides common considerations for the organ chip field, such as standardization, democratization, and regulatory challenges (15), model-builders must improve existing designs and develop new ones that can meet the ever-increasing needs of traditional lung biologists. On the other hand, lung biologists who are accustomed to using traditional models must be educated about advances in the organ chip field and be kept current on the rapidly expanding capabilities of these new tools in addressing research questions in lung biology and pulmonary medicine. As with organ chips, it is only through intimate interactions and cross talk within research communities that new knowledge about lung physiology can be obtained.