A biohybrid system to reconstruct epithelial morphomechanics in vitro

TL;DR

A new biohybrid system called BIOPRECS mimics how epithelial tissues compact in response to muscle contraction, enabling detailed study of tissue mechanics and morphogenesis.

Contribution

The novel contribution is the development of BIOPRECS, a programmable in vitro system that replicates multi-axial epithelial compaction and organ-specific architectures.

Findings

BIOPRECS enables programmable multi-axial epithelial compaction mirroring in vivo processes.

The system allows examination of tissue-scale folding and subcellular deformation responses.

It reproduces organ-specific epithelial architectures similar to those in the lung, intestine, and stomach.

Abstract

Epithelial tissue compaction driven by smooth muscle contraction is central to epithelial morphogenesis and mechanotransduction. However, in vitro systems that recapitulate this mechanical process – especially those capable of mimicking the anisotropic and multi-axial nature of smooth muscle contraction – remain limited. Here, we present the biohybrid programmable epithelial tissue compaction system (BIOPRECS), consisting of an overlying contraction-responsive epithelial tissue (ET) and underlying thermo-responsive contractile hydrogel (CH). By coupling isotropic hydrogel contraction with defined epithelial geometric anisotropy, BIOPRECS enables programmable multi-axial epithelial compaction that mirrors in vivo processes. Its in situ culture-compatible setup allows straightforward examination of both tissue-scale folding and subcellular deformation, including cytoplasmic and nuclear responses. Furthermore, by modulating tissue geometry and boundary conditions, BIOPRECS reproduces organ-specific epithelial architectures reminiscent of those observed in the lung, intestine, and stomach. This system has a significant potential to become a powerful tool for investigating how mechanical compaction regulates epithelial morphogenesis and mechanotransduction.

Image 1

Figures

Biohybrid programmable epithelial tissue compaction system (BIOPRECS) replicating smooth muscle-induced epithelial tissue compaction. (A) Schematic of anisotropic smooth muscle layers underlying epithelial tissues *in vivo*, generating deformation from tissue-scale to subcellular-scale. (B) Schematic of BIOPRECS composed of an engineered epithelial tissue and a thermo-responsive contractile hydrogel. The pNIPAAm hydrogel contracts and transmits mechanical forces to the overlying epithelium in a geometry-dependent, anisotropic manner. (C) Schematic and photograph of the pNIPAAm hydrogel undergo

Multi-scale deformation and multi-axial compaction of epithelial tissue in BIOPRECS(A and B) Representative fluorescence images showing tissue-scale deformation of epithelial wrinkling, cytoplasmic compression, and nuclear elongation of epithelial cells in the BIOPRECS.(C) 3D reconstructed immunofluorescence images of epithelial tissue under programmed multi-axial compaction of BIOPRECS. Scale bar represents 100 μm.Fig. 2

Geometrical design criterion of BIOPRECS for anisotropic compaction of the ET. (A) Schematic of BIOPRECS consisting of a polygonal ET on a circular CH. (B) Time-dependent contractile strain of the CH at 37 °C with corresponding images. (C) Particle image velocimetry (PIV) confirming isotropic contraction of the CH. (D) Schematic illustrating the definition of the contractile strain ratio (γ) between ET and CH. (E) Theoretical model with 3D surface plot of *γ* as a function of normalized initial dimensions with empirical fitting. (F) Design criterion for anisotropic compaction; yellow star indi

Multi-scale characterization of anisotropic compacted ET within BIOPRECS.(A and B) 3D and cross-sectional fluorescence images of ET stained with phalloidin (cyan) and DAPI (white), illustrating tissue-scale wrinkling and morphological changes upon contraction.(C) Quantification of individual nuclear positions along the z-axis in the xz-plane.(D) Statistical comparison of the standard deviation of nuclear z-positions with and without contraction of ETs.(E and F) Quantification of cytoplasmic and nuclear aspect ratios in the xy-plane.(G) Immunofluorescence images of ETs with and without contract

Programmable multi-axial compaction in BIOPRECS and organ-specific epithelial morphologies. (A-C) Schematics of geometrically designed BIOPRECS and representative fluorescence images of compacted ETs stained with DAPI (white) and phalloidin (cyan). (D) Quantitative orientation distributions of wrinkle and folding patterns in rectangular (top), cross-shaped (middle), and tripod-shaped (bottom) ETs. Shaded areas indicate distributions, and solid lines represent mean values. Blue, yellow, and red arrows indicate programmed compaction directions. Black dashed lines mark orientations aligned with p

Full text

Introduction

1

Tonic contraction of the smooth muscle layers under epithelial tissues, which is a long-lasting and stable form of contractile activity, induces epithelial compaction and exerts sustained mechanical stress on adjacent epithelial layer. This process plays a crucial role in shaping epithelial microarchitecture and regulating tissue function [[1], [2], [3], [4]]. Such biomechanical compaction, which is predominantly anisotropic and occasionally multi-axial, is fundamental to morphogenetic epithelial folding and the establishment of tissue-specific patterns during development [[5], [6], [7], [8]]. For example, in the lung, circumferential smooth muscle contraction mechanically guides the repetitive wrinkle pattern of the airway epithelium. Also, in developing intestine, bi-axial contraction of the two smooth muscle layers generates compressive force in orthogonal directions, driving zigzag epithelial folding and the formation of villi that define the stem cell niche and enhance nutrient absorption [[9], [10], [11], [12], [13], [14]].

While significant efforts have been made in engineering in vitro epithelial tissue models incorporating mechanical stimuli such as microfluidic shear flow or cyclic stretching, systems reproducing epithelial tissue compaction remain underdeveloped despite its recognized importance in tissue morphogenesis and function [[15], [16], [17], [18], [19]]. Existing approaches typically allow only uniaxial compaction, failing to recapitulate the anisotropic and multi-axial contractile dynamics generated by multiple smooth muscle layers in vivo. Furthermore, many of these systems rely on bulky and expensive electromechanical components [5,16,[20], [21], [22], [23], [24]], which hinder their integration with standard cell culture workflows and make it difficult to maintain long-term in situ culture. These constraints limit the ability to monitor dynamic cellular or tissue-level responses to sustained compaction over time.

To overcome these challenges, we developed a biohybrid programmable epithelial tissue compaction system (BIOPRECS) that replicates in vivo smooth muscle-induced epithelial compaction using a geometrically engineered epithelial tissue (ET) and a thermo-responsive poly(N-isopropylacrylamide) (pNIPAAm) hydrogel. Acting as an artificial smooth muscle, the pNIPAAm hydrogel provides controllable and anisotropic compaction of the ETs composed of both cells and extracellular matrix (ECM) hydrogel in a simple, low-cost, and in situ culture-compatible method. This setup enables sustained mechanical strain that drives tissue-scale morphogenesis while simultaneously inducing subcellular responses in epithelial cells. Notably, by programming the directionality of compaction through defined ET geometries and boundary conditions, BIOPRECS represents the first in vitro system capable of reproducing tissue-specific epithelial architectures driven by multi-axial mechanical compaction, closely mimicking tissue-specific morphogenetic patterns observed in organs such as the lung, small intestine, and stomach.

Results and discussion

2

Biohybrid programmable epithelial tissue compaction system (BIOPRECS) recapitulating multi-axial epithelial compaction

2.1

The epithelial tissue, in hollow organs such as the small intestine and stomach, is supported by underlying smooth muscle layers that exhibit distinct anisotropic contractile directionality (Fig. 1A) [[25], [26], [27]]. The contractile forces from the smooth muscle layers are transmitted to epithelium through the intervening connective tissue composed of ECM. Such mechanical cues drive tissue-scale morphological deformation in epithelium such as wrinkling and folding and extend to the cellular level, inducing cytoplasmic compression, nuclear polarization, and ultimately modulation of mechanotransducive signaling pathways that regulate cellular function [[28], [29], [30]].Fig. 1. Biohybrid programmable epithelial tissue compaction system (BIOPRECS) replicating smooth muscle-induced epithelial tissue compaction. (A) Schematic of anisotropic smooth muscle layers underlying epithelial tissues in vivo, generating deformation from tissue-scale to subcellular-scale. (B) Schematic of BIOPRECS composed of an engineered epithelial tissue and a thermo-responsive contractile hydrogel. The pNIPAAm hydrogel contracts and transmits mechanical forces to the overlying epithelium in a geometry-dependent, anisotropic manner. (C) Schematic and photograph of the pNIPAAm hydrogel undergoing volumetric contraction above its lower critical solution temperature (LCST). Scale bar represents 100 μm.Fig. 1

To emulate this physiological mechanics on in vitro, we developed BIOPRECS (Fig. 1B), which integrates two major components: an overlying contraction-responsive engineered epithelial tissue (ET) and an underlying thermo-responsive contractile hydrogel (CH). The ET mimics the bilayer architecture of native tissues, consisting of an epithelial monolayer seeded on top of an ECM hydrogel. The CH is made up of pNIPAAm hydrogel, which undergoes isotropic volumetric contraction at physiological temperature of 37 °C above its lower critical solution temperature (LCST) (Fig. 1C) [[31], [32], [33]]. Although the contraction of the pNIPAAm hydrogel is intrinsically isotropic, anisotropic deformation of epithelial tissue can be achieved by imposing geometrical constraints. Specifically, by patterning the epithelial tissue into a high-aspect ratio rectangular geometry aligned with the desired compaction direction, isotropic contraction is converted into preferential transmission along the long axis, resulting in anisotropic compaction of the epithelial tissue.

Upon thermal activation at 37 °C, the BIOPRECS induces distinct morphological changes within the ET, most notably the formation of surface wrinkles with characteristic wavelengths on the order of tens of micrometers (Fig. 2A). These tissue-scale deformations are accompanied by cytoplasmic distortion and nuclear reshaping, confirming effective transmission of mechanical cues to the cellular and subcellular levels (Fig. 2B).Fig. 2. Multi-scale deformation and multi-axial compaction of epithelial tissue in BIOPRECS(A and B) Representative fluorescence images showing tissue-scale deformation of epithelial wrinkling, cytoplasmic compression, and nuclear elongation of epithelial cells in the BIOPRECS.(C) 3D reconstructed immunofluorescence images of epithelial tissue under programmed multi-axial compaction of BIOPRECS. Scale bar represents 100 μm.Fig. 2

Furthermore, by tailoring the geometrical design of ET, BIOPRECS generates organotypic, multi-axial compaction fields. This capability allows the recapitulation of microarchitectures on ET that closely resemble those observed in the lung, small intestine, and stomach (Fig. 2C). Collectively, these results demonstrate the versatility and physiological relevance of BIOPRECS as a powerful platform for studying epithelial morphodynamics and mechanotransduction.

Geometric system design for inducing anisotropic compaction of epithelial tissue

2.2

To achieve anisotropic compaction of the ET on the thermo-responsive CH within BIOPRECS, we established a fundamental geometric design framework for both components. The primary goal was to impose defined directional orientation in the ET while ensuring predictable and quantifiable contraction of the CH. The ET was fabricated as a rectangular prism to define orthogonal axes, thereby enabling programmed anisotropic deformation. In contrast, the CH was designed as a circular disk, taking advantage of the intrinsically isotropic contractile behavior of pNIPAAm hydrogel and allowing uniform contraction analysis across all directions. This geometric framework provides the structural basis for achieving programmable anisotropic epithelial compaction on isotropic contraction of the CH (Fig. 3A).Fig. 3. Geometrical design criterion of BIOPRECS for anisotropic compaction of the ET. (A) Schematic of BIOPRECS consisting of a polygonal ET on a circular CH. (B) Time-dependent contractile strain of the CH at 37 °C with corresponding images. (C) Particle image velocimetry (PIV) confirming isotropic contraction of the CH. (D) Schematic illustrating the definition of the contractile strain ratio (γ) between ET and CH. (E) Theoretical model with 3D surface plot of γ as a function of normalized initial dimensions with empirical fitting. (F) Design criterion for anisotropic compaction; yellow star indicates the configuration selected for subsequent experiments ( $[eqn]$  = 0.5). (G) Top- and side-view images before and after contraction, demonstrating directional deformation under the designed geometry. (H) Quantitative comparison of contractile strain in CH and ET along x- and y-axes over time. (I) Anisotropy index of ET during CH contraction. Scale bar represents 10 mm. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)Fig. 3

The CH was fabricated from pNIPAAm hydrogel, which undergoes isotropic volumetric contraction upon thermal activation above its LCST (∼32 °C) (Fig. S1) [34,35]. Considering that tissue compaction strain in vivo can reach up to ∼ 0.5, BIOPRECS was designed to generate strains within this range. To achieve sufficient compressive strain, the pNIPAAm hydrogel was prepared at a very high concentration (∼90% w/v) [36]. Fabrication parameters were optimized using a Taguchi design-of-experiments approach, systematically varying crosslinker concentration, dry CH thickness, and UV polymerization time. The optimal conditions of 2 wt% crosslinker concentration, 1 mm of dry CH thickness, and 2 min of UV exposure yielded stable and homogeneous contraction without hydrogel blebbing or irregular deformation [37,38] (Fig. S2).

Characterization revealed that, upon incubation at 37 °C, the optimized CH exhibited predictable contraction kinetics, plateauing within ∼15 min and reaching a contractile strain of ∼0.7 (Fig. 3B). Particle image velocimetry (PIV) further confirmed radially symmetric displacement vectors during contraction (Fig. 3C), consistent with uniform planar and homogeneous CH contraction. Notably, the strain rate followed a log-normal profile resembling smooth muscle contraction kinetics (Fig. S3), underscoring the biomimetic relevance of the CH actuation [[39], [40], [41]].

A key design requirement of BIOPRECS is efficient transmission of CH contraction into ET compaction. This requires sufficient interfacial adhesion between the two components. We observed that adequate interfacial adhesion formed after ∼10 min of dwelling time of ET on the CH, likely mediated by physical entanglement and interlocking between fibrous pNIPAAm polymer chains in the CH and collagenous fibers in the ET matrix (Fig. S4) [[42], [43], [44]].

To quantify ET deformation in response to CH contractile, we defined a contractile strain ratio (γ) between ET and CH and developed a geometric model linking γ to interfacial dimensions (Fig. 3D). Assuming quasi-static loading and negligible lateral confinement in the ECM matrix of the ET, the system was modeled as a compressible neo-Hookean solid [[45], [46], [47], [48]]. For assuming quasi-static loading, stress relaxation time comparison between epithelial tissue and actuation is important [49]. Epithelial monolayers subjected to controlled mechanical strain exhibit a biphasic stress relaxation response, with a dominant exponential relaxation characterized by a time constant on the order of 10-20 s, governed primarily by actomyosin remodeling [50]. Moreover, the ECM hydrogel exhibited rapid stress relaxation, reaching a steady-state value within a second (τ < 1s) (Fig. S13). In addition, the operating conditions of our system impose specific constraints that significantly limit the relevance of temperature-dependent viscoelastic effects for the epithelial tissue investigated in this study (Fig. S12). Although the ECM hydrogel adjacent to the thermo-responsive contractile hydrogel (CH) may transiently experience a temperature change when the CH is transferred from room temperature (∼20 °C) to the incubator, we experimentally evaluated the mechanical response of the ECM hydrogel across this temperature range. As shown in Fig. S11, no statistically significant differences in mechanical properties were observed, indicating that the ECM hydrogel exhibits temperature-insensitive mechanical behavior within the relevant temperature window. This yielded the following expressions:

$$[eqn]$$

and

$$[eqn]$$

where $[eqn]$ and $[eqn]$ denote initial ET dimensions in the x- and y-axes, respectively, and $[eqn]$ is the initial CH diameter. This dimensionless model enables predictive control of ET deformation by programming interfacial geometry (Fig. 3E).

For effective anisotropic compaction of ET in the x-axis, $[eqn]$ should approach 1 (indicating efficient strain transfer), while $[eqn]$ should remain small where we defined $[eqn]$ as a criterion for minimal transverse strain transfer, consistent with anisotropic compaction regimes reported in previous studies [[51], [52], [53]]. Based on this criterion, we identified interfacial geometries that supported anisotropic compaction and selected ET dimensions of $[eqn]$  = 20 mm and $[eqn]$  = 4 mm (yellow star, Fig. 3F).

As anticipated, BIOPRECS with this geometry exhibited pronounced anisotropic ET compaction along the x-axis in response to isotropic CH contraction, while the y-dimension remained ∼4 mm (its original value). This directional deformation – characterized by substantial shortening in the x-direction and minimal change in the y-direction – was evident in both top- and side-view images of the BIOPRECS (Fig. 3G).

Quantitative analysis showed ∼50% contractile strain in the x-axis and only ∼10% strain in the y-axis, compared to ∼70% isotropic contraction of the CH after 20 min at 37 °C (Fig. 3H). The anisotropy index ( $[eqn]$ ) increased markedly after thermal activation and stabilized above 4, indicating strong directional bias (Fig. 3I). Model predictions closely matched experimental measurements across different ET geometries (Fig. S5). In summary, geometric optimization of BIOPRECS enables anisotropic compaction of epithelial tissues despite isotropic hydrogel actuation. The established design criterion offers a quantitative framework for programming both the magnitude and directionality of ET compaction with tunable anisotropy.

Anisotropic contraction-driven morphological and proteomic alterations in ET

2.3

To investigate the biological impacts of anisotropic ET compaction in BIOPRECS, we examined structural changes at tissue, cellular, and nuclear levels. Prior to contraction, the ET exhibited a flat morphology with a uniform epithelial monolayer (Fig. S6). Following CH contraction, the geometrically predesigned ET compacted anisotropically along the x-axis, producing tissue-scale deformation characterized by repetitive, periodic wrinkle patterns in the xz-plane, accompanied by increased cellular crowding and elevated cell density (Fig. S7).

High-magnification imaging of F-actin and nuclei revealed multi-scale deformations, spanning tissue-, cellular- and nuclear levels [[54], [55], [56]]. Epithelial cells exhibited distinct morphological changes, with both cytoplasmic and nuclei elongation, oriented predominantly perpendicular to the x-axis of compaction (Fig. 4A and B). Nuclear positional mapping along the z-axis revealed periodic fluctuations consistent with wrinkle morphology (Fig. 4C), with a significant increase in positional variance (Fig. 4D). The wrinkle wavelength was ∼100 μm, in line with prior reports of epithelial wrinkling under ∼0.5 compressive strain [5]. Furthermore, quantitative analysis confirmed cytoplasmic and nuclear elongation in the xy-plane (Fig. 4E and F, Figure S8), with complementary deformation and reorientation in the xz-plane, further supported the influence of tissue-scale compaction on cellular- and nuclei-level morphology (Fig. S9).Fig. 4. Multi-scale characterization of anisotropic compacted ET within BIOPRECS.(A and B) 3D and cross-sectional fluorescence images of ET stained with phalloidin (cyan) and DAPI (white), illustrating tissue-scale wrinkling and morphological changes upon contraction.(C) Quantification of individual nuclear positions along the z-axis in the xz-plane.(D) Statistical comparison of the standard deviation of nuclear z-positions with and without contraction of ETs.(E and F) Quantification of cytoplasmic and nuclear aspect ratios in the xy-plane.(G) Immunofluorescence images of ETs with and without contraction, in the xy-plane, stained with ZO-1 (magenta), Lamin A (red), phalloidin (cyan), and DAPI (white).(H and I) Normalized fluorescence intensity of ZO-1 and Lamin A as a function of culture duration and contraction state. Scale bar represents 100 μm. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)Fig. 4

We next examined whether anisotropic compaction alters expression of functional proteins. Immunostaining for zonula occludens-1 (ZO-1), a tight junction protein, and Lamin A, a key regulator of nuclear envelope stability [[57], [58], [59]], revealed significant upregulation under compaction (w/Contr.) (Fig. 4G). Notably, these elevations occurred rapidly, as expression levels were already distinct from those of the control group (w/o Contr.) by culture day 1. ZO-1 expression was elevated as early as day 1 and remained localized at junctional sites up to day 7, contrasting with reported effects of mechanical stretching, which often disrupts junctions and suppresses ZO-1. Lamin A expression increased progressively in both compacted and control groups, but compacted ET displayed markedly stronger and more distinct nuclear staining by day 7 (Fig. 4H and I).

In summary, geometrically designed BIOPRECS recapitulated anisotropic epithelial compaction, inducing hierarchical deformation from tissue-level wrinkling to cytoplasmic and nuclear elongation. Importantly, beyond morphological alterations, the BIOPRECS also elicited functional proteomic responses: reinforcement of ZO-1 at junctions, supporting epithelial barrier integrity, and upregulation of Lamin A, promoting nuclear stiffening and mechanoprotection. Together, these findings highlight compaction as a potent regulator of epithelial morphodynamics and mechanotransduction, influencing both cytoskeletal-junctional organization and nuclear mechanobiology.

Programmable multi-axial anisotropic compaction of ET recapitulating organ-specific morphomechanics

2.4

In vivo, epithelial tissues are subjected to combinatorial contractile forces generated by multi-layered smooth muscles with distinct anisotropically oriented contractile directions [25,26]. These multi-axial contractions are fundamental to organ-specific morphomechanics, yet reconstructing them in vitro has remained technically challenging. Building on our design principle of inducing anisotropic compaction, we systematically varied the geometric configuration of the ET to program multi-axial compaction profiles. The core concept was to guide contractile forces from the isotropically contracting CH so that compaction is preferentially directed along defined axes. While a single elongated rectangular ET produced uniaxial contraction, assembling multiple rectangular units into cross-shaped or tripod-shaped configurations enabled simultaneous bi-axial or tri-axial compaction at the central region.

As shown in Fig. 4, distinct ET geometries produced characteristic morphological outcomes. Rectangular ETs generated wrinkle patterns oriented predominantly perpendicular (90°) to the x-axis (marked with a blue arrow at 0°) of compaction, consistent with uniaxial compression. Cross-shaped ETs, designed for bi-axial compaction, produced folding patterns aligned with both principal compaction directions (marked with blue and yellow arrows at 0° and 90°) and their verctorial summation near 45°. Tripod-shaped ETs, designed for tri-axial compaction (marked with blue, yellow, and red arrow at 30°, 90°, and 150°), displayed more complex wrinkle orientations, with distribution peaks at around 30°, 90°, and 150°, reflecting higher-order mechanical interactions between multiple anisotropic compaction vectors and the emergent morphogenetic patterning of the epithelial surface. Orientation distribution analysis confirmed that wrinkle and folding orientations were directly governed by the imposed compaction geometry (Fig. 5A–C).Fig. 5. Programmable multi-axial compaction in BIOPRECS and organ-specific epithelial morphologies. (A-C) Schematics of geometrically designed BIOPRECS and representative fluorescence images of compacted ETs stained with DAPI (white) and phalloidin (cyan). (D) Quantitative orientation distributions of wrinkle and folding patterns in rectangular (top), cross-shaped (middle), and tripod-shaped (bottom) ETs. Shaded areas indicate distributions, and solid lines represent mean values. Blue, yellow, and red arrows indicate programmed compaction directions. Black dashed lines mark orientations aligned with primary compaction axes, while gray dashed lines mark orientations corresponding to vector summation of compaction directions (n = 3, mean ± s.d.). (E) Organ-mimetic epithelial morphologies generated from epithelial cells derived from the lung airway (uni-axial), small intestine (bi-axial), and stomach (tri-axial). Scale bar represents 100 μm. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)Fig. 5

To demonstrate the biomechanical relevance, we applied this strategy using epithelial cells derived from the lung airway, small intestine, and stomach. By tailoring BIOPRECS to replicate the characteristic smooth muscle contraction orientations of each organ, i.e., uni-axial (0°) for lung airway, bi-axial (0°, 90°) for intestine, and tri-axial (0°, 45°, 90°) for stomach, we successfully reproduced organotypic epithelial architectures (Fig. 5D). The emergent wrinkling and folding patterns reminiscent of native epithelial tissue architecture, underscoring the role of mechanical compaction in shaping epithelial organization.

Together, these results establish BIOPRECS as a versatile and physiologically relevant platform for programming multi-axial compaction, enabling both fundamental dissection of epithelial morphogenesis and the development of advanced in vitro models for mechanobiology.

Contractile forces are pivotal mechanical regulators of epithelial tissue architecture, coordinating macroscopic organization with intracellular dynamics. In vivo, epithelial tissues continuously adapt to contractile inputs from one or more smooth muscle layers, undergoing tissue-scale deformations coupled to cytoplasmic and nuclear remodeling. These biomechanical processes underpin tissue development and physiology, driving processes such as morphogenesis, cell polarization, and region-specific cell fate decisions through mechanotransduction. With growing recognition of their importance, there is increasing demand for in vitro systems that can reconstruct such phenomena using physiologically relevant components, rational structural design, in situ culture compatibility, and experimental accessibility (Fig. S10).

The BIOPRECS introduced in this study represents a remarkably simple yet powerful system to reproduce epithelial tissue compaction in vitro and investigate its biological consequences. Comprising only three basic elements of pNIPAAm hydrogel, ECM hydrogel, and epithelial cells, the system is functionally optimized for controlled induction of compaction and subsequent mechanobiological analyses. The thermo-responsive pNIPAAm hydrogel, which isotropically contracts at physiological temperature (37 °C), provides a robust and biocompatible source of contractile force. Through a strategically designed interface, isotropic hydrogel contraction is converted into anisotropic epithelial tissue compaction without introducing additional heterogeneity in material composition or fabrication processes, enabling programmable control over uni-, bi-, and tri-axial deformation. This low-cost, minimal design allows direct visualization of outcomes across scales, from tissue wrinkling to cytoplasmic and nuclear elongation, while supporting in situ culture for downstream functional studies.

During hydrogel contraction, limited water outflow may occur due to ECM volumetric shrinkage. However, under quasi-static conditions, fluid release is expected to occur slowly and primarily through laterally unconstrained regions, making fluid-flow–induced stresses on the epithelial monolayer minimal. Consistently, Live/Dead assays showed no detectable change in cell viability during compaction (Fig. S10), supporting those epithelial responses are governed predominantly by solid-phase mechanical compaction rather than fluid-mediated effects.

Our findings demonstrate that mechanical compaction strengthens epithelial junctions, as evidenced by enhanced ZO-1 upregulation, and simultaneously promotes nuclear mechanoadaptation through Lamin An upregulation. These results highlight that compressive stress elicits not only morphological remodeling but also molecular and genetic responses, positioning BIOPRECS as a model for investigating mechanotransduction in epithelial tissues.

A particular notable feature is the system's geometric programmability. By varying tissue configurations, we demonstrated that multi-axial compaction that has been previously difficult to achieve can be readily reproduced. The design principles presented here provide a framework for engineering epithelial tissue subjected to bi-, tri-, or higher-order axial contractions, directly reflecting the morphomechanics of organs such as lung airway, intestine, and stomach. Beyond these examples, programmable compaction may be extended to reconstruct complex morphologies in skin, craniofacial tissues, and other mechanically patterned epithelia. This versatility expands the utility of BIOPRECS from fundamental mechanobiology to regenerative medicine and tissue engineering. To further advance the system, future studies should incorporate systematic rheological characterization of hydrogels as a function of polymer concentration and composition. In addition, given that CH operates through thermo-responsive actuation, temperature-dependent rheological analysis will be essential to quantitatively link CH contraction behavior with the resulting mechanical loading conditions imposed on the ET.

In summary, BIOPRECS provides a cost-effective, experimentally accessible, and biologically relevant platform for studying epithelial compaction. Its ability to bridge morphological and proteomic responses to compressive stress establishes it as a versatile tool for advancing research in tissue morphomechanics and mechanotransduction.

Conclusion

3

In this study, we introduced the biohybrid programmable epithelial tissue compaction system (BIOPRECS), a minimal yet powerful system for in vitro reconstruction of epithelial tissue compaction. By integrating strategic geometric design of epithelial tissue (ET) with a thermo-responsive pNIPAAm hydrogel contractile layer, BIOPRECS translates isotropic hydrogel contraction into programmable anisotropic and multi-axial compaction of ET. This simple and cost-effective system faithfully recapitulates epithelial morphomechanics, driving tissue-scale wrinkling, cytoplasmic compression, and nuclear deformation, while also inducing functional adaptations such as enhanced ZO-1 and Lamin A expression.

Beyond uniaxial modes, BIOPRECS enables biaxial and triaxial compaction, reproducing organ-specific epithelial architectures of the lung airway, intestine, and stomach. As the first broadly applicable in vitro system capable of systematically programming epithelial morphomechanics under multi-axial cues, it provides critical insights into how compressive stresses regulate epithelial morphogenesis and mechanotransduction across molecular, cellular, and tissue scales.

Collectively, BIOPRECS offers a physiologically relevant and experimentally accessible tool for dissecting the biomechanical regulation of epithelial morphology and function. Its modular design and organotypic reconstruction capabilities position it as a versatile platform with applications ranging from fundamental mechanobiology to regenerative medicine, disease modeling, and tissue engineering.

Experimental section

4

Fabrication of BIOPRECS

4.1

For fabrication of the contractile hydrogel (CH) of BIOPRECS, pNIPAAm hydrogel precursor solution was prepared by dissolving N-isopropylacrylamide (NIPAAm, 97%; Sigma-Aldrich, USA) in deionized (DI) water at a weight ratio of 9:1. For crosslinking, 2 wt% N, N′-methylenebisacrylamide (MBAAm; Sigma-Aldrich, USA) was added to the precursor solution and mixed thoroughly. Additionally, 0.05 wt% 2-hydroxy-1-[4-(2-hydroxyethoxy) phenyl]-2-methyl-1-propanone (Irgacure 2959; BASF, Germany) was incorporated as a photoinitiator to facilitate UV-induced polymerization. The solution was sonicated (UCP10; Lab Companion, Republic of Korea) for 60 min at high intensity to ensure homogeneous mixing. The hydrogel was molded into the desired shape using a polydimethylsiloxane (PDMS) mold. The PDMS molds were fabricated by mixing PDMS base and curing agent (Sylgard 184; Dow Corning, USA) at a 10:1 wt ratio, followed by degassing and curing at 80 °C for 4 h. The prepared pNIPAAm precursor solution was then poured into the PDMS molds and exposed to UV light (365 nm) for polymerization. Following polymerization, the hydrogels were demolded and laser-cut into circular shapes using a laser cutting machine (IS350, INNOSTA, Republic of Korea). To activate the hydrogels for biological use, samples were immersed in DI water and cell culture medium (Advanced DMEM; Gibco, USA) for 10 h at room temperature (RT) to induce swelling.

Fabrication of epithelial tissue (ET) followed protocols from previous work [5]. Briefly, to construct the epithelium-ECM hydrogel bilayer, PDMS molds were prepared to define the ECM hydrogel substrate. PDMS base and curing agent were mixed at a 10:1 wt ratio, poured into custom-designed molds, and cured at 80 °C for 4 h. The ECM hydrogel solution was prepared by mixing rat-tail type I collagen (Corning, USA), 1 M sodium hydroxide (NaOH; Sigma-Aldrich, USA), and 10 × Dulbecco's modified Eagle's medium (DMEM; Gibco, USA) at a volume ratio of 1:0.025:0.1. To achieve a final collagen concentration of 3 mg mL^−1^, additional 1 × DMEM was added. The ECM solution was poured into PDMS molds and allowed to gel at 37 °C in 5% CO_2_ incubator for 1 h.

Three human epithelial cell lines of small intestine (Caco-2; Korean Cell Line Bank, Republic of Korea), lung (A549; Korean Cell Line Bank, Republic of Korea), stomach (AGS; Korean Cell Line Bank, Republic of Korea) were cultured in their respective media: Advanced Dulbecco's Modified Eagle's Medium (Advanced DMEM; Gibco, USA) for Caco-2, Roswell Park Memorial Institute 1640 (RPMI 1640; Gibco, USA) for A549 and AGS. All media were supplemented with 10% fetal bovine serum (FBS; Gibco, USA) and 1% amphotericin B (Sigma-Aldrich, USA). Cells were detached using TrypLE (Thermo Fisher Scientific, USA) and seeded onto ECM hydrogel at 1 × 10^6^ cells mL^−1^. After 2 - 3 days of culture, stable epithelial monolayers formed on top of the ECM hydrogel. The ET constructs were then placed onto the central region of the CH.

Optimization of pNIPAAm fabrication parameter using the taguchi method

4.2

Fabrication parameters for pNIPAAm hydrogels were optimized using the Taguchi method to minimize surface roughness during contraction. Three factors were considered: crosslinker concentration (2, 6, 10%), dry CH thickness (0.5, 1.0, 1.5 mm), and UV exposure time (120, 240, 360 s). Surface roughness was measured for each condition in triplicates. Signal-to-noise ratio (SNR) was calculated for all conditions, and mean SNR values were analyzed to identify a condition minimizing roughness and ensuring stable hydrogel structure. Analysis of variance (ANOVA) was performed to evaluate the contribution of each factor to variation in roughness. All analyses, including SNR calculations, factor level evaluations, and ANOVA, were conducted using MATLAB. Optimal fabrication conditions were determined based on the parameter set yielding the lowest SNR and most stable surface morphology.

Thermo-responsive ET compaction in BIOPRECS

4.3

To induce thermo-responsive compaction of ET on CH, pre-fabricated ET constructs of defined geometry were carefully positioned at the central region of swollen CH in culture dishes. After 10 min dwelling to ensure stable adhesion at the ET-CH interface, 30 mL of cell-type-specific culture medium was gently added to fully submerge both ET and CH. For Caco-2 cells, Advanced Dulbecco's Modified Eagle's Medium (Advanced DMEM; Gibco, USA) was used. Constructs were incubated at 37 °C for 2 h to trigger thermally induced contraction of the pNIPAAm. After contraction, compacted ET-CH constructs were carefully transferred to fresh culture dishes containing 10 mL of medium and maintained under standard condition (37 °C, 5% CO_2_) for subsequent experiments.

Mechanical characterization of ECM hydrogel

4.4

To evaluate the mechanical properties of the ECM hydrogel, stress relaxation and rheological measurements were performed. Stress relaxation behavior was assessed using cylindrical collagen hydrogel samples (diameter: 10 mm, height: 4 mm), fabricated following the same protocol as that used for all ECM hydrogels. All measurements were conducted at physiological temperature (37 °C) to match the experimental conditions used during tissue compaction. Uniaxial compression tests were performed using a universal testing machine (INSTRON 6800, Instron, USA), subjected to step compressive strains of 10%, 20%, and 30%, which were subsequently held constant while the temporal evolution of the compressive force was recorded. To assess the temperature dependence of the viscoelastic properties of the ECM hydrogel, oscillatory shear rheology was performed using cylindrical collagen samples (diameter: 10 mm, height: 3 mm), prepared using the same fabrication protocol. Rheological measurements were conducted using a rotational rheometer (Discovery HR-2, TA Instruments, USA) equipped with a temperature-controlled stage. Frequency sweep tests were carried out over an angular frequency range of 0.1 to 100 rad s^−1^ under a constant strain amplitude of 2%, which was confirmed to be within the linear viscoelastic regime for all tested conditions.

Derivation of contractile strain ratio

4.5

To model ET compaction driven by CH contraction, we assumed the mechanical characteristics of collagen hydrogel and contraction properties of pNIPAAm: (i) quasi-static contraction of CH allows fluid-solid equilibrium, (ii) negligible lateral confinement of collagen permits approximation of Poisson's ratio as zero (ν ≈ 0), and (iii) nonlinear collagen mechanics follows a compressible neo-Hookean model. The force exerted by CH on ET was expressed as proportional to contact area and average shear stress. From the generalized neo-Hookean framework, the strain ratio between ET and CH was derived and experimentally calibrated (Fig. S5):

$$[eqn]$$

where $[eqn]$ and $[eqn]$ denote initial ET dimensions in x- and y-axis, respectively, and $[eqn]$ denotes the initial CH diameter. This dimensionless scaling highlights that ET strain under CH contraction is governed predominantly by ET geometry, providing a design framework for programming anisotropic compaction profiles in vitro.

Imaging and image analysis

4.6

To visualize ECM hydrogels, Rhodamine 123 (Sigma-Aldrich, USA) was incorporated into the hydrogel solution prior to gelation. For immunofluorescent staining, cells were fixed with 4% paraformaldehyde (Chembio, USA) for 10 min at RT and permeabilized using 0.1% (v/v) Triton X-100 (Sigma-Aldrich, USA) for 20 min. Non-specific binding was blocked with 1% (w/v) bovine serum albumin (BSA) in phosphate-buffered saline (PBS) for 1 h at RT. Nuclei and F-actin were stained using 4′,6-diamidino-2-phenylindole (DAPI; Thermo Fisher Scientific, USA) and Alexa Fluor 647 phalloidin (F-actin; A22286, Invitrogen, USA) at dilutions of 1:200 and 1:500 (v/v), respectively, in PBS for 20 min at RT.

For Caco-2 differentiation under compaction, Lamin A (MA1-06101, Invitrogen, USA) and Zonula occludens-1 (ZO-1; 40-2200, Invitrogen, USA) were immunostained. After fixation and permeabilization, cells were incubated with primary antibodies (1:100 in BSA, 1 h), followed by three PBS washes. After washing, secondary antibodies of Alexa Fluor 547 goat anti-mouse (A11004, Invitrogen, USA) and Alexa Fluor 488 goat anti-rabbit (A11008, Invitrogen, USA) were applied at 1:150 in BSA for 1 h at RT.

Cell viability was assessed using a LIVE/DEAD assay (Molecular Probes, USA) with Calcein AM (green) and Eethidium homodimer-1 (red) at 1:150 and 1:200, respectively, in PBS for 20 min at 37 °C. Immunofluorescence images were captured using a confocal microscope (Olympus, Japan) and analyzed with Imaris Viewer (Ver.10.0.0, Oxford Instruments, UK) and ImageJ (NIH, USA). Structural features of the epithelial bilayers were further characterized using Adobe Illustrator (Adobe, USA).

CH contractile motion analysis

4.7

Strain distribution and magnitude of pNIPAAm hydrogel contraction were analyzed using particle image velocimetry (PIV) using an open-source software, OpenPIV-MATLAB (Mathworks, USA). This technique quantitatively assessed the strain field by tracking the displacement of features between sequential images. Images were divided into 32 × 32 pixels interrogation windows, and cross-correlation was applied to track displacement between sequential images. Local velocity vectors were used to generate strain maps illustrating the magnitude and orientation of hydrogel contraction.

Post-compaction ET structural analysis

4.8

Wrinkle and folding patterns of compacted ETs were analyzed in MATLAB using an edge-detection-based approach followed by Hough transform to identify pattern orientations. Extracted pattern orientations were quantified, and their distributions were visualized as histogram plots and polar vector charts. The MATLAB script used for this analysis computed the normal vectors of the detected wrinkle or folding structures and classified their orientations within 0° - 180°. To ensure statistical robustness, we processed four images per condition and calculated the mean and standard deviation of orientation distributions. Data were normalized for cross-condition comparisons.

Statistical analysis

4.9

Statistical significance was determined using Student's t-test for pairwise comparisons and one-way ANOVA with Tukey's multiple comparison test for multiple groups.

CRediT authorship contribution statement

Jangwon Yoon: Conceptualization, Formal analysis, Methodology, Writing – original draft. Jaeseung Youn: Formal analysis, Investigation, Methodology, Writing – original draft, Writing – review & editing. Dong Sung Kim: Funding acquisition, Supervision, Writing – review & editing.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.