Microfluidic intestinal-chips as next-generation modeling strategies: Multifaceted emerging applications in food research, functional foods and personalized nutrition

Abstract

Food digestion and absorption are essential functions of the human gut and the cornerstone of food research. This article has reviewed the core modeling strategies of intestinal-chips, and their application in the field of food science research, including precise evaluation of nutritional efficacy, food safety risk prediction, innovative development of functional foods and construction of personalized nutritional intervention strategies. Although microfluidic intestinal-chips have addressed some limitations inherent to traditional models, they still face technical challenges in fully replicating the intestinal microenvironment, as well as neural innervation, vascular perfusion and systemic immune responses. Currently, intestinal-chips are accelerating their integration with advanced technologies such as biomaterials, stem cell science, gene editing, 3-D bioprinting and artificial intelligence, to continuously enhance the physiological fidelity, support the optimization of food processing techniques and accelerate the development of new functional foods. These advancements of intestinal-chips won't only substantially promote the industrialization of next-generation food innovations such as cell-cultured meat and synthetic biology foods, but also drive nutritional science research from group recommendations to individualized decision-making. This technology thus provides an indispensable in vitro experimental platform and transformation acceleration engine for achieving "precision nutrition targeting the gut".

Graphical abstract

The application advantages of intestinal-chips in food science. The integration of intestinal-chips with advanced technologies can significantly enhance the in vitro intestinal physiological simulation, drive the optimization of food processing techniques and nutritional fortification, accelerate the healthy food design and new products development, and promote the in-depth innovation in precise nutrition.

Highlights

• •Microfluidic intestinal-chip strategies and food research applications were reviewed.
• •Integration of intestinal-chips with cutting-edge technologies were analyzed.
• •AI-powered intestinal-chips enable rational food design and precision nutrition.
• •Technical bottlenecks, challenges, and potential applications were discussed.

1. Introduction

The gut, as a complex and important organ, plays a pivotal role in food nutrients metabolism. It serves not only as the primary site for digestion and nutrient absorption, but also supports overall health via immune regulation and dynamic interactions with gut microbiota (Xiang et al., 2023; Hevia et al., 2023; Wheeler et al., 2024; Garcia-Gutierrez and Cotter, 2022). These complex interactions make the study of intestine an important subject in the fields of food science and nutrition.

The traditional in vitro models have obvious limitations in simulating the complex physiological environment of human intestine. As shown in Fig. 1, although static cell models such as Transwell systems are commonly employed in intestinal-related research, their two-dimensional (2-D) monolayer culture approach oversimplifies the intricate physiological architecture of the intestine. These models also fail to replicate the three-dimensional (3-D) villus structure of the epithelium, let alone emulate dynamic mechanical stimuli such as intestinal peristalsis, or the inter-organ crosstalk between the intestine and distal organs (Li et al., 2022). Furthermore, static models cannot establish the characteristic aerobic-anaerobic stratified microenvironment of the gut, which impedes the investigation of core physiological processes like probiotic adhesion and microbial fermentation, and prevents comprehensive elucidation of the action mechanisms of food functional components (Garcia-Gutierrez and Cotter, 2022). Although in vivo animal experiments can partially mimic intestinal physiology, yet the obtained findings have limited applicability to humans due to interspecies differences in lifespan, gut anatomy and physiological traits. For instance, the murine intestine lacks circular folds, resulting in a chyme transit time only one-fifth of that observed in humans. Additionally, the mouse gut microbiota is predominantly dominated by Firmicutes (accounting for over 80%), whereas the human gut microbiota harbors a relatively high abundance of Bacteroidetes (40-60%) (Wheeler et al., 2024). Meanwhile, animal experiments are further constrained by high costs, prolonged experimental cycles, exclusive endpoint detection and ethical concerns, which render these traditional intestinal models inadequate to meet the rapid screening requirements of the food industry.

Fig. 1.

A comparative analysis of the advantages and limitations of biological research models in food science.

In recent years, microfluidic intestinal-chips have emerged as powerful alternatives for intestinal research. The intestinal-chip is a micrometer-scale biomimetic model engineered based on microfluidic technology, which enables precise manipulation of fluids at the nanoliter-to-microliter scale, and supports co-culture and regulation of intestinal systems under controllable dynamic flow conditions. By integrating principles from cell biology, engineering and materials science, this model aims to recapitulate key intestinal physiological processes in vitro, including digestion and absorption, barrier maintenance, and gut-microbiota interactions. The core components of a microfluidic chip include the cell manipulation unit (reagent injection system and cell culture chambers) and the fluid control unit (multi-channel peristaltic pumps), which enable accurate distribution and transport of culture medium across all functional chambers (Xiang et al., 2023; Li et al., 2022). Through rational design and structural optimization, microfluidic intestinal-chip systems achieve realistic simulation of the human intestine in four key aspects: (1) By incorporating micrometer-scale villus arrays and a multi-chamber structures separated by porous membranes, the model reproduces the spatial heterogeneity of intestinal lumen and vascular barrier, thus mimicking the complex 3-D villus architecture of gut epithelium; (2) Through periodic mechanical stretching of membranes, fluid shear stress and periodic deformation are introduced to simulate intestinal peristalsis and precisely regulate intestinal epithelial cell differentiation; (3) By modulating microfluidic structural design and material permeability, a physiological oxygen concentration gradient is established to recapitulate nutrients transport and metabolism; (4) Utilizing advanced co-culture techniques to integrate intestinal epithelial cells, goblet cells, immune cells, gut microbiota, and even enteric neurons, the model constructs a sophisticated biological interface to simulate the neuromodulatory mechanisms in intestinal immunity, and dissect the interactions between the gut, symbiotic microorganisms, and distal tissues (Xiang et al., 2023; Thomas et al., 2023). Overall, the core characteristics of microfluidic intestinal-chips lie in their high integration, high-throughput capacity, and real-time in-situ monitoring capabilities. These features enable rapid acquisition of highly reliable data related to intestinal physiology and function in a precisely controlled experimental environment, thus offering broad application prospects in dietary nutrition profiling, health risk factor assessment, and bioactive food component screening.

Although existing reviews have summarized the application scenarios and technical value of intestinal-chip in food and nutrition research, most of them conduct isolated discussions on single topics, thus the systematic analysis remains largely underexplored. Several recent reviews have preliminarily discussed emerging directions such as the integration of intestinal-chip with artificial intelligence (AI) and its implementation in personalized nutrition, but these studies have not fully discussed the actual transformative value of this technology for the modern food industry. And the systematic evaluation of implementation barriers, translational bottlenecks and regulatory adaptation challenges encountered during its large-scale deployment in the food industry is also lacking. The core innovation of this review lies in integrating these scattered advances into a unified analytical framework, achieved by (1) systematically delineating the representative applications of intestinal-chips in the food industry, (2) critically examining key challenges hindering their industrialization, and (3) underscoring their distinctive value in personalized and precision nutrition.

To this aim, we performed a comprehensive search of relevant literatures from the PubMed, Web of Science, and Scopus databases science 2021. The search strategy combined keywords related to microfluidic intestinal-chip (or "gut-on-a-chip"), gut microbiota (or "intestinal microbiota"), dietary nutrients, nutrition-related adverse outcomes (e.g., "malnutrition", "obesity", "undernutrition", or "metabolic syndrome"), as well as "artificial intelligence", "food industry innovation", "personalized nutrition" and "precision nutrition". Only original experimental research, mechanism-focused investigations and high-quality review articles published in English are included. All of the retrieved records were first subjected to deduplication, followed by a two-stage screening workflow. The first stage is a preliminary screening based on titles and abstracts to evaluate their relevance to the core scope of this review, and then entered the second stage for a detailed full-text reassessment. This two-step procedure ensures that the finally included references meet requirements in terms of thematic relevance, research dimension coverage and academic quality, and can provide sufficient and reliable literature supports for the core conclusions of this review.

2. Core technologies and modeling strategies of intestinal-chip

Since its inception, the intestinal-chip has undergone remarkable advances, evolving from early conceptual proof-of-concept models into sophisticated, functionally integrated systems. Initially, intestinal-chips were designed as a single intestinal epithelial model to recapitulate the fundamental structural and functional traits of intestinal epithelial cells. Subsequently, to more accurately mimic the complexity of the human intestinal microenvironment, intestinal-chip gradually shifted toward multi-cell co-culture models, which gave rise to the innovative "intestinal organoid-on-a-chip" (Qi et al., 2023). In recent years, with the deepening research on the crosstalk between the intestine and distal organs, including the gut-liver axis, gut-adipose axis, and gut-skin axis, intestinal-chips have further evolved toward multi-organ-on-a-chip tandem systems and eventually the "human-on-a-chip" platforms (Kim and Sung, 2024). As illustrated in Fig. 2, by integrating microscale channels, dynamic fluid control mechanisms, and advanced multi-cell co-culture techniques, current intestinal-chips can effectively simulate the physiological architecture and key functions of the human intestine (Zhang et al., 2024; Tian et al., 2023).

Fig. 2.

Core technologies and modeling strategies for building the intestinal-chips. A: 3-D microchannel fabrication and surface modification of substrate material utilized in intestinal-chips; B: Cells/microbiota co-culture system on intestinal-chips; C: Intestinal-chips integrate multidisciplinary technologies, including materials science, microfluidics and artificial intelligence (AI), thereby achieving a more accurate in vitro simulation of human intestinal physiological functions.

2.1. Core technologies and engineering principles

One of the primary advantages of intestinal-chip lies in its remarkable enhancement of absorptive surface area by mimicking 3-D villus structures. To achieve this, researchers have employed 3-D bioprinting or soft lithography techniques to fabricate micropillar arrays resembling human intestinal villi (natural height of 500-1200 μm) (Thomas et al., 2023; Shin and Kim, 2022). For instance, Altay et al. (2022) fabricated hydrogel scaffolds using photolithography with poly (ethylene glycol) diacrylate, and obtained villus structures with a height of 350 μm and a density of 16 villi per mm, which approximates the physiological height range of native human intestinal villi. Xi et al. (2022) successfully replicated both the crypt-villus microarchitectures via a two-step soft-lithography molding process, which supported long-term expansion of intestinal organoids on matrigel-coated surfaces and enabled the formation of a confluent, polarized monolayer with morphology highly similar to intestinal epithelium in vivo. Song et al. (2025) introduced a coaxial bioprinting approach to construct intestinal-chips featuring 3-D tubular structures, on which the cultured intestinal epithelial cells formed well-structured villi exceeding 100 μm in height, effectively mimicking the crypt-villus unit and its key physiological functions.

The selection and modification of substrate materials constitute another critical determinant to guarantee the physiological relevance and functionality of intestinal-chips. Material selection must balance biocompatibility, mechanical and optical properties, manufacturing convenience, sterilization processes and costs. At present, polydimethylsiloxane (PDMS) is the most widely adopted material due to its excellent gas permeability and favorable biocompatibility. However, PDMS has certain limitations such as its inherent hydrophobicity, which hinders cell adhesion, and its elastic modulus (1-3 MPa) is far higher than native intestinal tissue ($0.1‐1kPa$). In contrast, hydrogels (such as collagen and matrigel) offer superior biomimetic potential. Among them, collagen is a naturally extracellular matrix (ECM) component, it promotes cell adhesion, proliferation and differentiation. Matrigel is rich in growth factors and cell-adhesive molecules, which enhances cellular interactions and signal transduction. Besides, substrate stiffness of the chip exerts a direct regulatory effect on fate of intestinal epithelial cell, namely the rigid substrates (e.g., PDMS) promote monolayer epithelial formation, whereas softer substrates are more conducive to organoid self-assembly (Agarwal et al., 2021; Das et al., 2024). Thermoresponsive hydrogels, which undergo reversible phase transitions near physiological temperature (32°C), enable dynamic, spatiotemporally modulation of ECM stiffness and subsequently the differentiation fate of epithelial cells.

Surface functionalization is also important for enhancing the biocompatibility and functionality of intestinal-chips (Fig. 2). Coating chip surfaces with ECM components such as collagen, fibronectin or laminin can significantly enhance cell adhesion and differentiation. For instance, Wang et al. (2025a) coated the vertical walls of hydrogel chips with cross-linked type I collagen gel, and introduced DAPT (γ-secretase inhibitor) and prostaglandin E2 into the culture medium. By integrating with air-liquid interface culture technique, these treatments successfully increased the mucus layer thickness from approximately 5 μm to 50 μm. Researchers also find the supplementation of hydrogels with laminin-111 is essential for the formation and expansion of epithelial organoids (Fang et al., 2023).

2.2. Simulation of dynamic mechanical properties

In addition to villus structures, another prominent feature of intestinal-chips is their capacity to accurately recapitulate peristalsis, which is the fundamental mechanism for food mixing and nutrient absorption. Through the synergistic action of flexible porous membranes and micro-pump systems, the in vitro simulation of intestinal peristalsis can be achieved. In detail, the culture medium is continuously perfused at a desired flow rate into the cell culture channels, while periodic vacuum suction is applied to induce stretch deformation of porous membranes, thereby mimicking peristaltic movements of intestinal tracts (Fig. 2). This physiological mechanical stimulus can effectively promote the differentiation of intestinal epithelial cells and induce the formation of villus-like projections with typical morphological characteristics (Li et al., 2022; Thomas et al., 2023). Additionally, mechanical stimuli, as a biomechanical signal, can regulate the polarization and collective migration of intestinal epithelial cells (thereby affecting intestinal barrier function), and also exert considerable impacts on gut microbiota (Constantinou and Bastounis, 2023).

2.3. Intestinal cells and cells/microbiota co-culture systems

During the designing of intestinal-chip, cell composition is a critical factor that influences both the physiological relevance and the validity of experimental models (Fig. 2). Guided by specific experimental objectives, different studies selectively incorporate various intestinal cell types into the chip systems (Kim and Sung, 2024; Constantinou and Bastounis, 2023). For instance, Walsh et al. (2024) developed a co-culture model consisting of Caco-2 intestinal epithelial cells and THP-1 macrophages, to recapitulate intestinal inflammation and assess whether synthetic human milk oligosaccharides exert immunomodulatory effects comparable to their natural counterparts. Their findings support the development of cost-effective, functionally equivalent breast milk substitutes for infant formula. Chen et al. (2021) co-cultured Caco-2 cells with vascular endothelial EA.hy926 cells, and demonstrated that dietary phenolic compounds derived from legumes can be efficiently metabolized and absorbed by intestinal epithelial cells, and subsequently exerted antioxidant and anti-inflammatory effects against oxidative stress-induced damages on vascular endothelial cells, thereby contributing to the amelioration of vascular-related inflammatory responses and cardiovascular diseases.

As the understanding of the gut microbiota's impacts on health continues to deepen, the significance of microbial co-culture in intestinal-chip research becomes evident. Given that intestinal microbiota primarily thrive in the anaerobic environment, while sufficient oxygen supply is essential for the survival and proliferation of intestinal epithelial cells, precise reconstruction of the intricate oxygen concentration gradient in the intestine poses a significant challenge for developing intestinal-chip models (Xiang et al., 2023). This can be addressed by integrating gas regulation modules and designing special culture media to simultaneously establish the hypoxic and normoxic zones, enabling long-term co-culture of human intestinal epithelial cells with stable aerobic or anaerobic microbiota. The innovative intestinal-chip design provides a valuable tool for exploring the effects of gut microbiota abundance fluctuations (such as those triggered by antibiotic exposure or hormone stimulation) on the physiological functions of the intestinal epithelium (Xiang et al., 2023). Furthermore, these systems have also advanced the research and development of probiotics and microbiota-related nutritional supplements.

2.4. Simulation of multi-tissue interfaces

The innovative multi-channel intestinal-chip can integrate multiple intestine-related cell types to establish highly biomimetic multi-tissue interaction interfaces. This is achieved through a controllable fluid environment, physiologically relevant mechanical stimuli, as well as separate parenchymal and vascular chambers (Fig. 2). This system exhibits higher physiological fidelity than conventional in vitro models, and can more accurately predict the overall in vitro effects after the intake of dietary bioactive components and nutraceuticals (Garcia-Gutierrez and Cotter, 2022; Thomas et al., 2023). For example, Lee and Sung (2022) developed the "gut-skin axis" multi-organ chip, which enables independent culture and differentiation of intestinal and skin cells, and realizes cross-tissue mass transport through microfluidic channels. Subsequent experiments confirmed that impaired intestinal barrier function can significantly amplify the negative effects of dietary fatty acids on skin cells, manifested as reduced cell viability, up-regulated pro-inflammatory factor secretion, and elevated expression of inflammatory markers. This platform can be applied to the toxicity assessment of food allergens such as peanut and gluten, and provides a novel tool for dissecting the regulatory mechanisms of the "gut-skin axis". Ramadan et al. (2022) constructed an immunot-integrated intestinal-chip, which sets up three independent fluid channels for medium perfusion, immune cell culture, and cytokines capture and quantification, respectively. In lipopolysaccharide-induced intestinal inflammation, this research verified that both curcumin and docosahexaenoic acid, two representative dietary supplements, can significantly inhibit the secretion of pro-inflammatory factors, including TNF-α, IL-6 and IL-1β, confirming the clear application potential of such multi-chamber chips in high-throughput screening of immunomodulatory food components. Besides these, the integrated-gut-liver-on-a-chip platform, by co-culturing Caco-2 and HepG2 cells, helped clarify the metabolic pathways and potential toxicity of dietary free fatty acids (Yang et al., 2023), while the gut-brain organoid model was applied to explore how food components affect cognitive function and emotion modulation (Kim et al., 2024).

3. Technical innovation of intestinal-chip

3.1. Neuronalized and vascularized intestinal-chip

Although intestinal-chips can replicate certain traits of the intestinal epithelium, they typically lack critical components such as nerve cells and vasculature that are vital for gut physiology. The integration of advanced 3-D bioprinting with cell patterning on microfluidic chip platforms allows for the construction of innervated and vascularized intestinal-chips, which can more accurately reconstruct the complex intestinal microenvironment (Yin et al., 2024; Zheng et al., 2024). For instance, Li et al. (2024a) introduced a novel photocurable biomaterial for the fabrication of an intestinal-chip model with intraluminal microvilli, crypt-villus-like topology and functional tight junctions. In their study, human umbilical vein endothelial cells encapsulated in the model could autonomously form well-aligned microvascular networks under dynamic perfusion, effectively mimicking the dense and orderly capillary beds in human intestine.

However, the co-culture of the enteric nervous system and microvessels in intestinal-chips still faces substantial challenges. The primary limitation is the difficulty in fully replicating the intricate microenvironment necessary for neuron development and functional maintenance. Additionally, only a limited subset of neural cells has been focused on currently, resulting in an insufficient representation of the diversity and complexity of neurons and glial cells present in the enteric nervous system (Wu et al., 2023). Overcoming these challenges will deepen the understanding of intestinal physiological functions, but also provide mechanistic evidence for clarifying the pathological basis of neurogastrointestinal dysfunction and developing precise nutritional intervention strategies targeting the "gut-brain-vessel" axis (Zheng et al., 2024).

3.2. Integration with AI

Currently, the deep integration of personalized intestinal-chip systems and AI has become a core strategy to overcome technical bottlenecks in precision nutrition and drive the upgrading of the food industry toward customization. The core of this system is to integrate nutrition databases, gut microbial multi-omics profiles, host genomics and other digital data via multiple machine learning (ML) algorithms, then construct a quantifiable prediction framework for individual dietary responses, and output personalized nutritional intervention. The intestinal-chip recapitulates individual-specific intestinal physiology, microbiota composition and metabolic response, thereby functionally validation all AI-derived predictions. This bridges computational modeling directly to real intestinal physiology, avoiding false positive biases inherent in purely data-driven approaches. Ultimately, it was speculated to form a full-chain workflow of "AI multi-omics prediction - intestinal-chip functional validation - personalized nutritional intervention output". This paradigm shifts the "diet-gut microbiota-host response-health efficacy" network from correlation observation to systematic mechanistic modeling, and overcomes the long-standing industry pain of "one-size-fits-all" in functional food development. Thus, it is expected to play a critical role in the formulation of new-generation dietary guidelines, functional foods development and precision nutrition intervention schemes (Greathouse and Choudhury, 2025a). The core technical modules of this system can be specifically decomposed as followings:

• (1)Multi-paradigm ML algorithm system‌: Due to the complexity of dietary components, substantial inter-individual variability and the dynamic nature of gut microbial communities, different ML algorithms can complement each other in analyzing complex cell-cell interactions and dynamic physiological responses to stimuli. For example, random forest can capture the non-linear associations between different dietary components and gut microbiota composition, while recurrent neural networks are suitable for time-series modeling of the dynamic changes of microbial communities under long-term dietary interventions (He et al., 2026). Unsupervised clustering algorithms can identify the differences among probiotic strains, multivariate adaptive regression splines fit the non-linear dose-effect relationships in metabolic pathways, and the combination of random forest feature selection and deep neural networks enables the in-depth association mining of multi-source heterogeneous data (He et al., 2026). Besides, data augmentation, federated learning, incremental learning and transfer learning enhance the robustness and adaptability of the models (Patel et al., 2024); Explainable AI methods such as SHapley Additive exPlanations and Grad-CAM can visualize the decision-making logic of black-box models, which can significantly enhance the model reliability in nutritional function analysis and industrial implementation (Arrighi et al., 2026).
• (2)Digital twin and predictive modeling methods‌: Digital twin (DT) refers to a dynamic virtual mirror of physical entities constructed using data from sensors, the Internet of Things, virtual reality and AI. By coupling computational models with DT, it can accurately reproduce the traits, behaviors and operation processes of physical objects, and support simulation deduction and intelligent decision optimization. DT is rapidly penetrating into food and nutritional science, forming two core applications: (a) Full-link digital management of the food industry. Relying on the real-time mapping and simulation capabilities of DT, it supports the full life-cycle management of food production lines, full-chain food safety traceability and customized functional food research. (b) Personalized precision nutrition. Centered on the "digital gut twin" framework proposed by (Greathouse and Choudhury, 2025b), it fuses multi-source heterogeneous data and integrates multiple algorithms (including knowledge graph reasoning, large language model based association analysis, Bayesian network probabilistic inference and reinforcement learning), to model the non-linear "diet - gut microbiota metabolism - host response" cascade. Its aim is to realize personalized prediction of individual dietary intervention responses and provide a highly credible pathway for microbiota-targeted precision nutritional interventions.
• (3)Image analysis methods‌: Relying on the deep integration of high spatiotemporal resolution microscopic imaging and AI semantic segmentation, the system enables continuous, non-invasive and real-time tracking of dynamic physiological processes in intestinal-chip. The system can automatically and accurately identify, segment and quantify core physiological indicators, while simultaneously correlating the gene expression and metabolite data within matched time window. Thus it can capture the full temporal cascade from dietary bioactive compound exposure to physiological effects. This method completely replaces traditional endpoint detection mode, providing high spatiotemporal-resolution quantitative support for the mechanism analysis of dietary bioactive components.
• (4)Automated chip performance optimization‌: Driven by DT and global parameter sensitivity analysis, multi-objective iterative optimization is carried out for core physicochemical microenvironment parameters such as pH gradient distribution in microchannels, transmembrane flux of nutrients, physiological gradient of oxygen partial pressure and rheological properties of the mucus layer. The design of chip topology, selection of substrate material and control of perfusion fluid are also auto-optimized, thus the research cycle and trial-and-error costs were greatly shortened when compared with the traditional manual trial-and-error mode. Meanwhile, biomimetic degree of intestinal-chip and the stability of experimental results were further improved (Filippello et al., 2022).

At present, there are several representative research cases in this field. For example, Li et al. (2026a) coupled ML with finite element analysis to construct a parameterized COMSOL Multiphysics model of the intestinal-chip, establishing quantitative links among channel geometry, mass transfer coefficient, cellular oxygen consumption and the internal oxygen distribution. The resulting oxygen concentration prediction model can directly support chip design, and enables precise microenvironment regulation for studying intestinal barrier function, immune response and host-microbe interactions. Wu et al. (2024) reported the development of an intelligent intestinal-chip system integrated with ML algorithms for high-throughput screening of probiotic strains. By mimicking the gut microenvironment, it can comprehensively assess interactions between probiotic strains, host intestinal epithelial cells and immune cells. Using unsupervised ML algorithms, the system detects subtle therapeutic variations among strains and ranks them by efficacy scores. Thereby, it facilitated the cost-effective and function-driven screening of probiotics. Wang et al. (2026) proposed the AI-driven platform Gi-MAPS, which combined quantitative microbiome analysis with ecological modeling to simulate the gut microbial community responses to dietary interventions. Based on infant 16S rRNA datasets and the Q-net DT platform, Sizemore et al. (2024) captured the temporal dynamics of microbial communities, to simulate the potential effects of dietary interventions on infant gut microbial communities and neurodevelopment.

Nevertheless, this technical system still has clear implementation limitations. First, multi-dimensional heterogeneity including dietary nutrient intake, microbial composition, host sensitivity and environmental exposure continuously restricts the generalization of the model. The limited size of training data further biases performance across specific populations. Second, existing AI models still have limited capability to capture the dynamic temporal traits of microbiota-host interactions in the gut microenvironment, and it is difficult to fully replicate long-term intervention effects in vivo. Third, with the continuous expansion in parameter dimension of DT models, some black-box deep learning models lack sufficient interpretability, which may undermine mechanistic analysis of food bioactives and impede subsequent traceability.

Overall, the current AI-integrated intestinal-chip in the food and nutrition field is still in the stage of retrospective analysis, offline simulation and decision-support modeling, rather than strict causal relationships. Subsequent efforts are required to finally realize an operable DT workflow through pathways including long-term longitudinal data collection, cross-laboratory standardized data stream construction, dynamic model updating and iterative calibration, as well as external validation across populations. Combining with explainable AI tools will further strengthen the credibility for industrial applications. These advances will jointly shift nutrition research from the stage of empirical induction toward mechanistic elucidation and quantitative prediction, thereby paving the way for a paradigm shift from "population-based recommendations" to "individualized dietary prescriptions" in future nutrition practices.

4. Application of intestinal-chip in food science

Microfluidic intestinal-chips enable real-time monitoring of intestinal barrier function, transporter activity, and dynamic changes of immune markers, offering a novel quantitative research platform for assessing nutrients' digestibility, bioaccessibility, safety, allergenicity and their health impacts (Fig. 3). This section systematically reviews representative research advances in this field by comparing the technical features and application scenarios of intestinal-chip models with different cell sources and structural designs, and summarizes their respective advantages and limitations. The detailed information is presented in Table 1.

Fig. 3.

Applications of intestinal-chips in food science research. These application scenarios cover several critical areas, including nutritional research, food toxicology and safety assessment, probiotics development, and metabolic disease and personalized nutrition.

Table 1.

Systematic comparison of intestinal-chip models applied in food science research.

Research Topics	Core Characteristics of the Intestinal-chip	Key Research Findings and Technical Value	Limitations	References
Applications in dietary nutrition research Flavor and sensory modulation	Caco-2/HT29 co-cultured multi-channel intestinal-chip incorporating human taste receptors	Taste receptors modulate satiety signaling and nutrient absorption processes.	Lack of host–microbiota interactions	(Li et al., 2024b), (Wan, 2024)
Nutritional evaluation of formula foods	Intestinal organoid chip recapitulating key physiological features of infant gut	Developmental pathways driven by Lactococcus lactis and lactoferrin, to support precise design of infant formula	Limited infant-derived organoid availability	Wang et al. (2024a)
Applications in food toxicology and safety assessment Long-term dietary lipid toxicity	Intestinal-liver tandem microfluidic chip or intestinal organoid chip	Identify in vivo metabolic pathways of dietary lipids and their potential hepatotoxic mechanisms	High operational complexity in multi-organ chip	(Yang et al., 2023), (Filippello et al., 2022)
Heavy metal pollutants	Colon organoid chip, with physiologically relevant mucus	Cadmium bioavailability assessment has enhanced physiological relevance	Unable to represent distant organ accumulation toxicity	Li et al. (2025)
Microplastics	Colon organoid chip, recapitulating the physiologically relevant hypoxic microenvironment	Quantitative assessment of intestinal cytotoxicity and inflammation-inducing potential across microplastics with different size	Lack of evaluation the effects mediated by mucus layer and commensal microbiota	Park et al. (2023)
Pathogens	3D Intestinal–Microbiota Symbiosis Chip	Dynamic tracking of the complete pathogenic cascade initiated by bacterial pathogens	Challenges in maintaining anaerobic conditions and high technical requirements for imaging systems	Lee et al. (2024)
Mycotoxin	Caco-2 tubular dynamic intestinal-chip	Allows parallel analysis of multiple toxins, offering high screening throughput	Lower physiological relevance	(Morelli et al., 2024), (Pöschl et al., 2023)
Application in risk assessment of food additives Food colorants	Intestinal organoid chip	Comparative intestinal toxicity of synthetic food colorants vs. natural ones	Lack of toxicity analysis related to gut microbiota	(Kong et al., 2021), (Tan et al., 2024)
Food emulsifiers	Dual-chamber anaerobic perfusion chip for co-culture of colonic organoids and anaerobic microbiota	Identification of gut microbiota–mediated intestinal mucus barrier disruption induced by food emulsifiers	Hard to sustain anaerobic microenvironment over long-term	Ogulur et al. (2023)
Nanoparticle additives	Intestinal organoid chip	Quantitative detection of nanoparticle trans-barrier transport process	Insufficient data for long-term exposure	(Malaisé et al., 2024), (Qiao et al., 2025)
Application in food allergenicity assessment High-throughput screening of food allergens	3D-printed electrochemical integrated intestinal organoid chip	Non-invasive electrochemical detection with high throughput	Lack of allergic mechanism analysis	Brooks et al. (2023)
Quantitative assessment of food allergens	Three-layer chip co-culture of intestinal epithelium and immune cells	Dynamic tracking of allergen-induced temporal cascade reactions	High cost for isolation of immune cells	Zhang et al. (2025)
Applications in nutrition-related metabolic disease research Dietary interventions for specific populations	Microbiota-epithelial co-cultured intestinal- chip	Reproduce how bioactives reprogram the gut microbiota, generate immunomodulatory metabolites and regulate intestinal barrier homeostasis	Challenges in maintaining microbial diversity	Wang et al. (2024b)

4.1. Applications of intestinal-chip in nutrition research

Focusing on core application scenarios including the digestion and absorption of nutrients, flavor regulation, health benefits verification and industrial implementation, different intestinal-chip models exhibit differentiated technical advantages. For instance, the intestinal-chip integrated with taste receptors can dissect the mechanisms of nutrient absorption and related physiological responses at the molecular level, enabling rational flavor optimization of healthy foods with reduced sugar and salt (Li et al., 2024b; Wan, 2024). In food formula optimization, Feihe Dairy has developed a human intestinal organoid/multi-omics combined platform, to evaluate the nutritional values of three types of stage 1 infant formula. This platform captured subtle development-related regulatory pathways driven by Lactococcus lactis and lactoferrin that cannot be detected by traditional approaches. Therefore, it can support the precise and physiology-informed design of formula nutrients (Wang et al., 2024a).

4.2. Applications of intestinal-chip in food toxicology and safety assessment

Microfluidic intestinal-chips with varied designs have also been employed to investigate the toxic effects and mechanisms of food contaminants (such as heavy metals, pesticide residues, pathogens, mycotoxins and microplastics) on intestinal barrier function and epithelial homeostasis (Fig. 4). For example, Li et al. (2025) developed a single-channel 3-D intestinal-chip, to assess toxic effects of heavy metals in foods. Yang et al. (2023) constructed a microfluidic gut-liver-on-a-chip platform to elucidate the metabolic pathways and potential toxicity of dietary free fatty acids mediated by the gut-liver axis. Coupled with high spatiotemporal resolution imaging technology, this system can track the cascade damage process after contaminant exposure in real time, making it suitable for assessing the effects of chronic dietary exposure (Filippello et al., 2022; Park et al., 2023; Lee et al., 2024). And the tubular multi-parameter, high-throughput intestinal-chip was also constructed to evaluate the responses of intestinal epithelial cells to mycotoxins present in cereals (Morelli et al., 2024; Pöschl et al., 2023).

Fig. 4.

Examples of microfluidic intestinal-chips in food safety assessment. (A) An in vitro human NAFLD model constructed on an integrated-gut-liver-on-a-chip platform. Under free fatty acid induction, the microfluidic co-cultured Caco-2 gut-mimicking cells and HepG2 hepatocytes exhibited anti-apoptotic protective effects, whereas monolayer cultured cells showed induced apoptosis. These results demonstrate that this platform effectively mimics the pathological interactions mediated by the gut-liver axis, and holds promise as a reliable, animal-free tool for gaining deeper insights into the mechanisms underlying NAFLD development. Reproduced from Yang et al. (2023) (https://doi.org/10.1038/s42003-023-04710-8, open access). (B) Schematic representation of the 3-D intestinal model from Caco-2 using a microfluidic chip. The Cd bioavailability in rice measured by the 3-D intestinal-chip model was comparable to the mouse model, but significantly lower than the Caco-2 monolayer model. Reproduced from Li et al. (2025) (https://doi.org/10.26599/FSHW.2024.9250364, open access). (C) An representation of a dynamic 3-D gut-microbes interface, composed of cell culture chambers and collagen channels forming a miniature intestinal lumen, which effectively mimicking key structural and functional features of the human intestine. The intestinal-chip can secrete mucus and sustain microbial adhesion and growth over extended periods. Compatible with high-resolution imaging techniques, it enables observation of morphological changes in microvilli-like structures and the intestinal barrier, as well as the impact of microbial communities on the miniaturized intestine. The scale bars represent 20 μm. Reproduced from Lee et al. (2024) (https://doi.org/10.1002/advs.202302113, open access). (D) An representation of a high-throughput intestinal-chip for studying epithelial responses to enterotoxins, with chips embedded in a standard 384-well microtiter plate. Each chip is equipped with an optically transparent observation window, enabling high-resolution imaging and real-time live-cell dynamic monitoring. Exposure to enterotoxins leads to significant dose-dependent disruption of epithelial barrier integrity, characterized by a continuous decline in transepithelial electrical resistance values as toxin concentration increases. Significance was determined using the Wallis test followed by Dunn's post hoc test, where ****P < 0.0001; ***P < 0.001. Reproduced from Morelli et al. (2024) (https://doi.org/10.1038/s41598-024-56520-5, open access).

4.3. Applications of intestinal-chip in food additive analysis

Food additives including stabilizers, emulsifiers, food colorants and nano-antimicrobials are widely used in food processing and play roles in optimizing food flavor, improving processing efficiency and extending shelf life. However, the potential health risks under long-term dietary exposure lack precise evaluation (Song et al., 2023). Intestinal-chips enable long-term dynamic culture to faithfully model interactions among food additives, intestinal epithelium and gut microbiota, thus can reveal the sub-toxic damage mechanisms that undetectable by traditional methods. In relevant studies, human intestinal-chips have confirmed the significant intestinal toxicity differences between synthetic pigments and natural bioactive components (Kong et al., 2021; Tan et al., 2024). The colon organoid-microbiota co-culture chip is used to dissect the microbiota-mediated toxicity mechanisms of food additives (Ogulur et al., 2023), providing molecular-level mechanistic evidence for the risk management of emulsifiers. Dynamic exposure on intestinal-chip can clearly distinguish the health effects across different types of nanomaterials, providing critical in vitro evidence for safety evaluation of nano food additives (Malaisé et al., 2024; Qiao et al., 2025).

4.4. Applications of intestinal-chip in food allergen assessment

Food allergy is growing as a global public health concern-especially among children, which has an estimated global prevalence of approximately 10% (Florsheim et al., 2023). On the other hand, as novel protein-rich foods derived from plants, fungi, insects and microalgae, as well as artificial cultured meat, gradually enter the public diet, the potential risks of unknown new food allergens continue to rise. Thus, there is an urgent need to develop screening platforms to effectively assess the safety of these new proteins, particularly their sensitization risk (Dijk et al., 2023). The differently designed microfluidic intestinal-chips have their respective focuses in allergenicity assessment. For example, Brooks et al. (2023) reported a 3-D intestinal-chip integrated with electrochemical detection modules, which is suitable for high-throughput preliminary screening of food intolerance, and supports rapid identification of food allergies (Brooks et al., 2023). Intestinal-chip systems co-cultured with immune cells can fully recapitulate the entire in vivo sensitization process, thereby enabling quantitative and precise assessment of food protein allergenicity (Zuurveld et al., 2023; Janssen et al., 2024; Zhang et al., 2025).

4.5. Applications of intestinal-chip in nutrition-related metabolic disease research

Relying on the highly biomimetic and modular traits, microfluidic intestinal-chip can accurately model the intestinal physiology, genetic background and microbiota traits of different individuals, thus enhancing both the research depth and implementation feasibility of dietary interventions for metabolic diseases and personalized nutrition strategies (Fig. 3). For example, intestinal-chip co-cultured with microorganisms can fully recapitulate the complete action chain by which dietary bioactives (polyphenols, dietary fiber, carotenoids, etc.) regulate gut microbiota and subsequently target the modulation of intestinal barrier homeostasis (Wang et al., 2024b). And the modular interconnected multi-organ chip enables dissection of inter-individual differences in dietary response, revealing how immune background variations drive opposite health effects from the same nutrients.

5. Contribution of intestinal-chip to food industry

At present, the microfluidic intestinal-chip has exhibited irreplaceable value in core application scenarios including targeted screening of bioactive compounds, multi-dimensional safety assessment and systematic evaluation of novel protein-rich foods. Thus, it can advance both the optimization of food processing technologies and the precision design of functional foods (Fig. 5).

• (1)Accelerating the optimization of food processing technologies‌: Food processing methods (e.g., baking, boiling and frying) and environmental factors (e.g., pH levels and temperature) profoundly influence the composition, digestibility, absorption, metabolic conversion and health effects of foods. The in vitro microfluidic intestinal-chips provide a physiologically relevant and effective tool for evaluating how processing modulates nutrient properties and bioavailability. For example, Wang et al. (2023) developed an intestinal-chip to systematically analyze the transmembrane transport, uptake and metabolic transformation kinetics of carotenoid across the intestinal epithelial barrier. They found that the absorption of carotenoid is highly dependent on its ability to co-assemble with cholesterol, long-chain fatty acids and phospholipids to form mixed micelles in intestinal lumen, confirming that simply increasing dietary intake cannot effectively improve bioaccessibility of carotenoid. The core optimization direction should focus on adjusting the fatty acid-cholesterol ratio of co-ingested lipids and adding medium-chain triglycerides to enhance micelle stability. Meanwhile, they also verified that flavonoid glycosides (such as quercetin glycosides in apples and hesperidin in citrus fruits) promote micelle formation and subsequent absorption by upregulating the expression of CD36 and SR-BI receptors. These mechanistic findings provide precise targets for the development of carotenoid-fortified foods and designing of microencapsulated delivery systems for active components (Wang et al., 2023).

Fig. 5.

Technical bottlenecks and innovation strategies of intestinal-chip in food science. A: Technical bottlenecks include the stability of long-term cell cultures, the integration with multiple organs, and the absence of standardization and high-throughput detection capabilities. B: The integration with advanced 3-D bioprinting technology allows for the construction of vascularized and innervated intestinal-chips; C: The integration of intestinal-chips with artificial intelligence facilitates a clearer understanding of the intricate interactions among nutrition, host factors, and gut microbiota. This advancement is anticipated to foster the development of personalized dietary intervention strategies.

The AI-powered intestinal-chip can further realize intelligent iteration of processing technologies. For instance, by quantitatively analyzing the cell viability, morphological integrity and multi-omics metabolic profiles of intestinal-chip exposed to processed foods, AI-driven models can rapidly identify key parameters in thermal processing procedures such as baking, boiling and extrusion, thus effectively inhibiting the generation of cytotoxic by-products like acrylamide while maximizing the bioaccessibility of nutrients. This data will provide full-chain guidance for the process design of next-generation precision nutrition foods (Wang et al., 2025b).

• (2)Facilitating the targeted screening of bioactive compounds‌: The in vitro intestinal-chip provides a powerful platform for the screening of functional food bioactives, healthy benefits analysis and formula optimization, greatly reducing the trial-and-error costs. It enables systematic dissection of the entire process of absorption, distribution, metabolism and excretion of bioactive compounds, and precise mapping of underlying molecular pathways, signaling cascades and cellular responses. For instance, using the developed human intestinal-chip, Inoue et al. (2022) confirmed that the absorption of glucose, dipeptides and cholesterol strictly depends on the expression of transporters SGLT1/GLUT2, PEPT1 and NPC1L1, respectively. Meanwhile, they found that epigallocatechin gallate and naringenin (from green tea and citrus) significantly inhibit glucose absorption, while vegetable-derived flavonols and quercetin do not suppress cholesterol absorption by inhibiting NPC1L1. This finding corrected the false positive conclusions obtained from traditional 2-D culture systems, and greatly improved the efficiency of bioactive screening. Kim & Moon (Kim and Moon, 2026) utilized the intestinal-chip to screen and validate the function of bioactives in quinoa. They identified stigmasterol and β-sitosterol as the core functional factors that exert gastrointestinal protective effects. These researches provide direct mechanistic evidence for targeted screening of high-value cereal bioactives and formula optimization of functional foods.
• (3)Advancing food safety assessment‌: Intestinal-chips fully recapitulate the tripartite interactions of contaminants-intestinal epithelium-gut microbiota, and realize full-dimensional safety assessment from acute toxicity to long-term chronic effects. Relying on this platform, the effects of additives on intestinal barrier integrity, microbiota homeostasis and immune balance under different exposure durations and doses can be accurately tracked, and their potential long-term health risks such as inflammation induction and metabolic disruption can be clarified. In hypoallergenic food development, intestinal-chips monitor the entire process of trans-barrier transport, metabolism and immune activation of food components after desensitization modification, thus it can rapidly validate the efficacy of hypoallergenic formulas and greatly shortening the development timelines.
• (4)Systematic evaluation of novel protein-rich foods‌: For novel commercialized products such as cultivated meat and synthetic biology-derived foods, intestinal-chips can systematically evaluate the digestion kinetics, nutrient absorption efficiency and bioaccessibility of recombinant proteins, growth factors and other components introduced during manufacturing, as well as their potential allergenic risks. Intestinal-chips also accurately track the trans-barrier permeation process of protein components, and capture real-time allergic cascades such as immune cell activation and inflammatory factor release. Combined with AI algorithms, they support the systematic evaluation of nutritional attributes and allergenic risks, thus offering a solid safety foundation for commercialization of novel protein foods (Tan et al., 2024).
• (5)Promoting personalized nutrition‌: Due to the heterogeneity of gut microbiota composition and genetic background among different individuals, the absorption, metabolism and health effects of the same dietary component exhibit marked variability across individuals (Fig. 5). Relying on the modular co-culture capability of intestinal-chip, synthetic microbial community models can be customized for different individuals, to systematically dissect the diet-microbiota interactions. For example, the synthetic microbiota/intestinal-chip system proposed by Lee et al. (2023) can integrate strains with defined functions (based on individual gut microbiota testing results), to directly track the metabolic transformation process of different dietary components in the exclusive microbiota environment. Personalized intestinal-chips constructed from intestinal crypts or human induced pluripotent stem cells (hiPSCs) completely preserve the genetic and biological traits of the donor, overcoming the genetic homogeneity limitation of conventional cell models and enabling truly individualized physiological simulation (Wu et al., 2023; Eicher et al., 2022). In the relevant studies, the hiPSC-derived intestine-chip developed by Moerkens et al. (2024), which contains epithelial cells, stromal cells and neuronal cells, can autonomously form villus folds and lamina propria-like structures, and accurately recapitulate individual-specific mature intestinal phenotypes (Moerkens et al., 2025). These platforms enable direct assessment of inter-individual differences in absorption, metabolism and function of food components, reducing the trial-and-error costs in large-scale population interventions. For example, researchers have built personalized intestinal-chips tailored to individuals with specific phenotypes (such as lactose intolerance or specific metabolic traits), to directly screen dairy formulas in vitro. This approach facilitates rapidly distinguishing "responders" vs. "non-responders", ultimately supporting the design of precision dairy interventions grounded on genotypic evidence (Santos et al., 2024). Hasan et al. (2021) generated 3-D intestinal-chips from crypt stem cells of patients with morbid obesity, which completely retained the obesity-related metabolic traits of the patients such as abnormal expression of carbohydrate transporters and elevated glycolysis flux. This finding identifies targetable metabolic nodes for precision nutritional interventions in obesity subtypes.

At present, deep integration of intestinal-chip with transcriptomics, metabolomics, microbiomics and nutrigenomics remains limited. Deep integration of intestinal-chips with multi-omics technologies in the future is expected to provide unprecedented experimental controllability for unraveling the causal relationship chain between genetic variations, gut microenvironment, and host phenotypes, thereby offering a transformative platform to overcome key bottlenecks in functional food development—such as undefined targets, unclear mechanisms of action, and unpredictable individual responses (Cassotta et al., 2025). The emergence of intelligent and personalized intestinal-chip marks a shift in nutritional intervention from traditional passive remediation approaches to AI-driven proactive, predictive and regulatory frameworks. The "digital life twin" system promoted by BGI integrates intestinal-chip physiological simulation with multi-omics data, supporting precision nutrition assessment and customized gut microbiota regulation. It delivers a complete technical framework for dynamic monitoring of individual life status and real-time iteration of nutritional strategies (https://www.bgi.com/news/2026091011). Relying on the world's largest personalized nutrition randomized controlled trial system, the UK-based ZOE team deeply couples gut microbiota testing, continuous blood glucose monitoring, AI algorithms and intestinal-chip verification, has effectively optimized the gut microbiota composition and metabolic health status of subjects, and provided evidence-based personalized dietary solutions for over 250,000 members (Liu et al., 2026) (https://zoe.com/our-science). Viome, based on gut microbiota RNA sequencing, provides customized dietary advice and targeted nutritional supplements for over 500,000 users (https://www.viomepro.com/precision-health-pro). These advances have greatly improved the efficacy and safety of personalized nutrition, and hold promise for a paradigm shift in future nutrition practice-from "population-based recommendations" to "individualized prescriptions."

6. Challenges and future perspectives

The deep integration of intelligent and personalized intestinal-chip has opened a new technical paradigm for food science and nutrition research. Rather than a simple upgrade of experimental tools, this system establishes a systemic framework of "AI prediction - chip validation - data feedback optimization", which allows high-fidelity reproduction of the dynamic physiological processes of the human intestine, flora interaction mechanisms and personalized metabolic characteristics on a microscale platform. This transformation is gradually driving the entire food industry to shift from experience-based evaluation to quantitative and precision-oriented assessment. However, the large-scale promotion of this technology in the food industry still faces multiple practical challenges:

• (1)Construction of regulatory validation systems. Currently, China's food nutrition and safety risk assessment still relies primarily on animal experiments, 2-D cell models and population sampling testing. The national standard “General Technical Requirements for Intestinal Chips”, led by the China National Center for Food Safety Risk Assessment, is under development. By defining core performance parameters, functional evaluation methods and data output specifications, this standard will set a unified threshold for this technology to enter the regulatory acceptance system. At the international level, the European Food Safety Authority is formulating a regulatory framework for the use of non-animal methods (NAMs, including human organ-on-a-chip technologies) in food safety and nutritional assessment. In April 2025, the U.S. FDA announced a 3-5 year transition toward NAMs and already accepted human-relevant toxicological data obtained from intestinal-chips as supplementary evidence in reviewing certain food additives and novel foods. The United Kingdom launched the "Replacing Animals in Science" strategy in 2025, under which the UK Centre for the Validation of Alternative Methods was established to coordinate cross-domain validation and foster collaboration between regulatory agencies and technology developers. Singapore's "Future-Ready Food Safety Centre" has launched an integrated system for food risk assessment and nutrition detection, which adopts organ-on-a-chip for toxicity testing. These initiatives not only validate the scientific maturity of intestinal-chips, but also highlight their emerging application value in food safety and human health (Li et al., 2026b).
• (2)Full-chain standardization of intestinal-chips. At present, unified standards have not yet been formed for the microfluidic intestinal-chip in terms of structural design, cell sources, model construction, analytical procedures and functional validation, which directly restricts the cross-platform/laboratory data comparability and experimental reproducibility (Carvalho et al., 2023). In 2024, the Chinese Society of Biotechnology released the first group standard for human intestinal-chips titled "Organ on a Chip: Gut" (Liu et al., 2024). This standard clarifies the scope, terminology, definitions, technical requirements and detection methods for the construction of human intestinal-chip models, laying the foundation for the standardization of operational procedures across the entire industry. Future efforts need to promote unified manufacturing protocols, standardized cell banks and AI-assisted data analysis frameworks. These advances will transition intestinal-chips from research tools to mass-producible products. Ultimately, they will gradually eliminate deviations among laboratories, completely resolve the long-standing reproducibility issues and significantly broaden its accessibility in global food research institutions and industries.
• (3)Scalability, fabrication costs and benefit optimization. In the process of transition from laboratory to industrial scenarios, the scalability of intestinal-chips remains a core bottleneck. Most existing intestinal-chip systems are developed for biomedical toxicology applications and lack pre-treatment modules needed for complex food matrices. Consequently, their direct deployment in food-related contexts often leads to poor adaptability. In addition, product consistency control during large-scale production, the maintenance of nutritional homeostasis in co-culture systems and high-precision modulation of environmental parameters will collectively increase operational complexity and raise implementation costs, rendering it difficult to be employed among small and medium-sized food industries.

Currently, the industry is adopting a modular plug-and-play architecture, which enables flexible integration of functional units such as intestine, liver and immune system through standardized fluid interfaces. Equipped with multi-channel sensors and cloud-based data acquisition modules, this system supports remote monitoring and cross-laboratory reproducibility. Meanwhile, AI-powered automatic perfusion and high-content imaging minimize deviations caused by manual operations. Advances in material science and high-precision bioprinting ensure batch-to-batch consistency. Furthermore, AI algorithms are used to assist cell interaction prediction and dynamic optimization of metabolic pathways, which can further balance performance, implementation costs and biomimetic fidelity. Ultimately, these advances will substantially lower the application threshold of intestinal-chips, enabling convenient use by non-specialized laboratories and industrial research teams (Zhou et al., 2025).

7. Conclusion

Food nutrition and safety are crucial for human health. As advanced 3-D in vitro systems, intestinal-chips have markedly improved the accuracy of food research by more faithfully recapitulating gut physiology in miniature. However, restricted by the complex anatomical structure of the human intestine, the dynamic interaction network of flora-immunity-metabolism, and the complexity of multi-organ system cascade responses, there is still a gap between the current technology and full replication of the in vivo microenvironment. At the industrial implementation level, key hurdles include poor comparability across laboratories, an immature regulatory validation system and the absence of full-chain standardization. Relying on multi-dimensional technological innovations such as AI-driven real-time microenvironment control, functional modification of chip interfaces and high-precision 3-D bioprinting, the platform will achieve an optimal balance among high biomimetic fidelity, scalable manufacturability and cost-effective implementation. The short- and medium-term research priorities should focus on the construction of standardized operating procedures, modular plug-and-play multi-organ integration, and deep AI/personalized intestinal-chip coupling across the full "prediction-validation-optimization" loop. The next-generation "4-D organoid-on-a-chip" system is expected to respond robustly to tissue signals and dynamically adapt to physiological rhythms. With the continuous development of microfluidic intestinal-chips and their integration with other disciplines, this technology is poised to drive innovation in the food industry and support the mechanism-guided precision design of the next generation functional foods.

Author contributions

Junli Feng: Funding acquisition, Investigation, Conceptualization, Investigation, Writing – original draft; Hejian Zhu: Data curation, Formal analysis; Yan Peng: Investigation, Visualization; Liqin Lin: Methodology, Resources; Jingzi Yao: Data curation; Xiuqin Fan: Resources; Jiayan Shen: Validation; Jing Xue: Supervision, Writing-review & editing; Zejun Wang: Conceptualization, Funding acquisition; Guofang Shen: Writing-review & editing; Qing Shen: Conceptualization, Resources, Writing-review & editing.

Disclosure statement

No potential conflict of interest was reported by the authors.

Data availability statement

Data will be made available on request.

Funding

This work was supported by the Zhejiang Provincial Medical and Health Science and Technology Program (2024XY031), Zhejiang Province Traditional Chinese Medicine Science and Technology Plan (2026ZL0702), Hangzhou Health Science and Technology Plan (B20261721),the Hangzhou Municipal Special Support for Biomedical and Health Industry Development Technology Program (2024WJC092), and the Jinhua Municipal Major Science and Technology Program (2026-3-078).

Declaration of competing interest

The authors declare no conflicts of interest.

Handling Editor: Dr. Yeonhwa Park