Advances in stem cell biology and engineering have led to the development of organs-on-chips (OOCs) and organoids, which are
in vitro models designed to emulate the structures and functions of human tissues in a more physiologically relevant manner than standard two-dimensional (2D) cultures do. After decades of technological progress, these models are now officially recognized tools in drug discovery and safety testing, with their roles evolving from bridging the gap between animal and human studies to validated approaches in reducing animal tests [
1]. Despite this great progress, only partial physiological structures and simplified functions can be recapitulated with current
in vitro tissues, compelling us to seek more improvements in OOC and organoid construction methods.
The organization of multicellular structures provides the foundation for tissue functionality. OOCs and organoids adopt distinct strategies for multicellular structure construction. Most OOCs feature a microfluidic chip with multiple chambers separated by a porous polymeric membrane (such as polydimethylsiloxane), inherited from the prototype lung-on-a-chip [
2]. The chambers are designed to accommodate different types of cells in specific areas, aligning with the heterogeneous cell distribution of native tissue. The inserted membrane facilitates substance transport while maintaining the space partition between chambers. Therefore, OOC is highly favorable for modeling tissue-tissue barriers, and multiple preclinical trials have verified that OOCs can enhance drug discovery and disease modeling. However, some considerable issues persist. The materials involved in the molding of the microfluidic chip and porous membrane significantly adsorb hydrophobic small molecules, although substantial efforts have been made to ameliorate this process [
3], [
4], [
5]. Such adsorption could compromise the reliability of drug tests, undermining the predictive value of OOCs as drug screening models. Nonspecific absorption could also alter the chemical constituents of the cellular microenvironment, causing unpredictable influences on cell behaviors and fates. Another major issue is the overly simplistic design of the tissue barrier, which relies on a porous membrane, in comparison to the multiple complex features of native physiological interfaces [
6], [
7], [
8]. Taking the alveolus as an example, such a blood-air barrier has a curvature radius at the scale of hundreds of micrometers for maximizing the gas exchange efficiency. The modulus of the tissue barrier is maintained on the order of a few kilopascals with a thickness of less than one micrometer, allowing efficient contraction and expansion of the alveolus. These features are not easily recapitulated in current OOC technology. Such challenges also exist in the construction of perfusable three-dimensional (3D) capillary vessels, which are indispensable for the healthy culture of large-scale tissue.
Organoid technology involves constructing
in vitro tissues in a different manner. Organoids originate from the self-assembly of stem cells, progenitor cells, and/or differentiated cells through cell-cell and cell-extracellular matrix (ECM) interactions, where Matrigel or Geltrex is often used to embed and support these cells [
9], [
10], [
11]. The generated organoids contain multiple cell lineages and sophisticated structures resembling the
in vivo organs, and some of them are reported to display mature phenotypes. As organoids are generated through cell self-assembly without external engineering manipulation, the fabricated organoids inevitably exhibit stochasticity [
12], [
13], [
14], [
15]. For instance, multipotent progenitor cells may undergo differentiation along unanticipated directions because of a lack of robustness. The ECM substitute Matrigel, derived from mouse sarcoma tissue, shows high batch-to-batch variations in its physical and chemical properties, such as rigidity changes [
16] and ligand density [
17], and fails to provide consistent guidance for cell differentiation and growth. The spatial structures organized by various cell populations might differ from our expectations, as any slight disturbances could impact the self-assembly behaviors. With respect to multiple unpredictable variables, ensuring consistency between organoids and physiological tissues in terms of both structure and function is difficult, making standardization and reproducibility challenging. As organoids are dependent on dish culture, different cell populations within them are maintained under homogenized culture conditions, resulting in a failure to provide heterogeneous and appropriate microenvironments for different types of cells [
18]. These issues are likewise faced by the assembly of different organoids called assembloids, which are highly significant models for studying organoids interactions.
For both OOCs and organoids, the limitations in mimicking the cellular architecture of native tissue and the associated microenvironment are largely attributable to the lack of effective compartmentalized strategies for tissue construction. The compartmentalization concept is extensively employed in cell biology [
19], [
20], [
21] and is often used to describe the microdomains formed by organelles within cells. These microdomains are spatially distinct yet selectively communicate with each other, allowing thousands of intracellular microreactions to occur simultaneously and in a cascade, guaranteeing that various metabolic reactions proceed in the right sequence. Similarly, cells surrounded by a phospholipid bilayer membrane are also regarded as compartmentalization, where the intracellular and extracellular environments are separated but intercommunicable to maintain normal cell morphology and function. Considering the independence and interaction of hierarchical multicellular architectures and associated microenvironments in physiological tissues, the principle of compartmentalization can also be extended to the tissue level [
22], [
23], [
24], [
25] (
Fig. 1). In native tissues and organs, different types of cells, along with their own microenvironments, occupy specific compartmentalized regions, thereby allowing specialized functions. Cells of the same type perform distinct functions when they reside in different positions and microenvironments. Taking intestinal tissue as an example, the villus domain consists of absorptive cells, goblet cells, enteroendocrine cells, and tuft cells; the crypt domain contains stem cells, Paneth cells, and other progenitor cells. The cells in the villi are responsible for the absorption of nutrients and water and the secretion of mucus and hormones, maintaining barrier integrity. The undifferentiated cells at the bottom of the crypt self-renew, differentiate, and migrate upward to the villus for the renewal and repair of the intestinal epithelium. These cells are confined to specific regions and have different functions but communicate with each other. Such structural and functional compartmentalization across multiple hierarchical levels is also applicable to other native tissues and organs, such as the skin, hepatic lobule, and vascular vessels.
From the perspective of compartmentalization, compared with native tissues, OOCs and organoids currently remain poorly understood with respect to multicellular architectures and associated microenvironments, highlighting the need for strategies that reproduce compartmentalized architectures
in vitro more faithfully. Over the past few decades, numerous methods for constructing compartmentalized systems have been developed, some of which have been applied to engineer
in vitro models. For example, bioprinting [
26,
27], a bottom-up additive biomanufacturing approach, is important for architecting compartmentalized cells and their microenvironments. By precisely depositing bioinks composed of cells and matrix materials along the
X,
Y, and
Z axes, bioprinting allows for the direct construction of tissue-like structures with defined geometry. Through the alternating use of nozzles containing various bioinks, different types of cells along with specialized active components can be deposited point by point and layer by layer. In this manner, cells can be positioned into their favorable matrix at predesigned regions, resulting in the compartmentalization of cells and their microenvironment. In bioprinting, the bioink gelled hydrogel works both as a matrix for supporting cell growth and as a compartmentalized boundary for cell organization. Such a compartmentalized boundary mediated by a soft hydrogel creates a morphogen gradient resembling gradients found in native tissues [
28], guiding the external shape and internal hierarchical cellular architectures of
in vitro tissue. For example, intestinal organoid-derived cells were seeded into hydrogels with a predefined geometry similar to that of native crypts [
29,
30]. In the absence of external biochemical stimulus gradients, cells organize into crypt-villus-like structures with specialized cell types at specified differentiated regions and exhibit key physiological hallmarks of native intestinal tissues, which is rare in conventional intestinal organoid construction methods, in which the formation of crypt-villus-like structures is highly stochastic. Such a hydrogel-mediated soft barrier for compartmentalized tissue generation was also demonstrated in OOC construction, such as skin-on-a-chip [
31], where full-thickness skin tissue with multilayer structures was established with the assistance of fibrin hydrogel. In addition to crypt-like patterning, other hydrogel modules, such as microgels, spheres, micro/nanofibers, membranes, and defined blocks, can also be used to guide morphogenesis and differentiation [
32,
33], resulting in compartmentalized multicellular architectures and associated microenvironments.
Physical fields offer another route to spatial compartmentalization. Physical fields have been studied for years for manipulating the distribution, movement, trapping, and sorting of particles through interactions between physical fields and particles. More recently, increasing interest has focused on the application of various physical fields to guide cell assembly, including the acoustic [
34], magnetic [
35], and optical [
36] fields. Among them, acoustic field-mediated bioassembly has attracted considerable attention because of its noncontact, flexible, biosafe, and high-throughput features [
37,
38]. Under an acoustic field, the acoustic radiation force is generated as sound waves propagate through a fluid medium. Such a force would direct cells into specific regions of sound waves, such as pressure nodes or antinodes, with high cell density and spatial resolution. By modulating the range, frequency, intensity, and mode of the acoustic field, numerous cell assembly patterns can be generated simultaneously. Over time, acoustic bioassembly has evolved from generating 2D symmetric patterns of a single cell type to constructing 3D tissues with multiple cell types, nonsymmetrical shapes, and closely packed architectures [
38,
39] (
Fig. 2(a) [
40]).
These advantages have enabled the use of the acoustic field as a promising tool for biofabrication. Some studies have reported the use of an acoustic field during and after bioprinting [
37,
38] (
Fig. 2(b) [
41]). With the assistance of an acoustic field, the dispersed cells in extruded bioinks could be aggregated and aligned, by which the cell density is enhanced significantly and the feature size of bioprinting is reduced. Using this approach, some improved bioprinting tissues, such as muscle, vasculature, tendons, and ligaments, have been demonstrated to have good cell viability [
37]. In addition to bioprinting, an acoustic field can also be used in microfluidic chips, referred to as an acoustic chip [
42,
43]. Leveraging the particle manipulation capability of the acoustic field, cell movement, separation, clustering, and 2D and 3D patterning have been achieved within the channels, without the need for the physical or chemical gradients required for traditional microfluidic chips (
Fig. 2(c) [
44]), exemplifying the way of merging the acoustic field manipulation method into OOCs. In addition to OOCs, the acoustic field could also be applied in organoid technology. As shown in
Fig. 2(d) [
45], the acoustic field has been demonstrated in the controllable fusion of organoids into assembloids with specific architectures. Promisingly, the acoustic field can not only manipulate the spatial architecture of cells but also steer the maturation and programming of engineered tissues both directly and indirectly [
37]. Directly, the acoustic field can apply force to the cells, inducing the compression, stretching, and shear of the cells and thereafter the rearrangement of the cytoskeleton. Indirectly, the acoustic field can modify the properties of the ECM, such as stiffness, to control cell behaviors and trigger the release of active molecules loaded in the ECM to stimulate cells. The capability of cell spatial manipulation and the potential for regulating cellular behavior demonstrate that physical fields are promising for engineering tissues and improving current OOC and organoid platforms. Notably, some issues remain to be addressed for current acoustic field technology, including the difficulties associated with selectively manipulating one single cell among a large population and generating highly complex 3D patterns, and the temperature of the system should be carefully monitored considering the acoustic energy input [
43,
46].
In addition to the solid materials leveraged for barrier construction in OOCs, nonsolid materials are also available in spatial compartmentalization by forming barriers driven by phase separation, such as liquid-liquid and gas-liquid interfaces, which have been studied for more than a century. A representative example of liquid-liquid interface-mediated compartmentalization is the droplet interface bilayer (DIB) proposed by Bayley’s group [
47], which was later employed to construct synthetic tissues (
Fig. 3(a) [
48]). They demonstrated that aqueous droplets, when introduced into an oil phase containing phospholipids, spontaneously formed lipid monolayers. Upon contact, adjacent droplets generated a lipid bilayer at the interface, creating stable droplet networks. These networks were later functionalized with membrane proteins to enable ion transport and signal transmission, mimicking the basic functions of muscle and nerve tissues [
48,
49]. This DIB-mediated network works as both a barrier and a bridge between the surrounding components, which could be further integrated into OOCs and organoids during the construction of diverse compartmentalized cells and their independent microenvironments. Currently, droplet microfluidics and 3D printing have been applied to generate DIB-mediated networks, through which thousands of networks with various architectures and formulations can be constructed in the short term [
50]. Limited by the current manufacturing technologies for handling ultrasoft materials, tissue such as the brain is difficult to reconstruct precisely, considering that different types of cells are accommodated to specific regions of a very soft ECM. For this purpose, this team utilized the droplets to compartmentalize neural stem cells and astrocytes (
Fig. 3(b) [
51]). By programming their positions, they recreated multiple stages of cortical development and significantly accelerated tissue maturation from months (as seen in organoids) to weeks [
51]. This study validated the feasibility of DIB in manipulating multicellular compartmentalization during
in vitro tissue construction and maturation, which could be further exploited in engineering the geometry of organoids and the assembly of diverse organoids into assembloids.
Interfaces are not only feasible for constraining the spatial distribution of cells but also amenable to cell adhesion, expansion, and directional differentiation. Keese and Giaever [
52] pioneered cell culture on liquid-liquid interfaces using oil and aqueous solutions. At the interface, a protein nanolayer spontaneously forms because of the denaturation and crosslinking of protein molecules, providing recognition domains for cell adhesion. Thus, the cells could spread and grow on the interfaces. By regulating the mechanical properties, such as shear modulus, the morphology of cell adhesion, the rate of cell proliferation, and the trend of cell differentiation can be controlled. For instance, Jia et al. [
53,
54] and Kong et al. [
55] reported that the ultrastructure of protein nanolayers formed on liquid-liquid interfaces was adaptive to cell traction forces, promoting mesenchymal stem cell differentiation toward neural cells (
Fig. 3(c) [
53]). Although numerous studies have confirmed the feasibility of using liquid interfaces in cell culture, to further apply them in 3D tissue culture, novel approaches are still needed to program these interfaces in 3D space. Studies on the mechanism of multi-interface formation and methods for inducing phase separation are still limited and require more attention and investigation. A comprehensive comparison table of the key features of different compartmentalization strategies is shown in
Table 1 [
14],[
15],[
37], [
38], [
39], [
50], [
51], [
56], [
57], [
58], [
59], [
14], [
60], [
61], [
62], [
63].
The established paradigm of OOCs and organoids has become deeply entrenched over the years of development. Although considerable efforts have been made, we must admit that the difference between OOCs, organoids, and native tissues is still substantial in both multicellular architectures and associated microenvironments. Rather than focusing solely on these established approaches, it is more productive to view OOCs and organoids as platforms for engineering in vitro 3D tissues with high physiological fidelity. Such a perspective opens the door to a wider range of feasible strategies for recapitulating the complex compartmentalized multicellular architectures and associated microenvironments being emulated. We highlight compartmentalization as a unifying and underexploited design principle underlying OOCs and organoids. Viewing compartmentalization not only as spatial partitioning but also as intercommunicable interactions between cells and their microenvironments, new conceptual and technical possibilities might be unlocked for further improvement of OOCs and organoids. Accordingly, emerging strategies, such as bioprinting, physical fields, and phase separation, have validated both the feasibility and the potential for engineering compartmentalized multicellular structures and their associated microenvironments. Integrating these compartmentalized strategies with existing OOC and organoid construction methods would provide a versatile toolkit for enhancing their structural fidelity, functional heterogeneity, and physiological relevance, thereby advancing the development of more faithful in vitro models.
PDF
(2594KB)
