Result
Design and operation of the generic organ-on-a-chip platform
As shown in Fig. 1A , each chip consists of 32 chip units, with each chip unit consisting of three horizontally arranged wells. The holes at both ends are connected by the passageway of the lower channel, where culture the medium was perfused with a rocker. The middle hole is designated for cell culture and is separated from the lower layer by a transparent PET porous membrane. The specifications of the porous membrane can be chosen according to specific requirements. This design not only reduces the consumption of culture medium and other materials but also facilitates pipetting operations and characterizations with enhanced convenience.
To validate the usability of the chip, we constructed a simple gut-on-a-chip model. In the specific experiments, Caco-2 and HT-29 cells were seeded in a 9:1 ratio, totaling about 50,000 cells, in the central well of the chip unit to form a compact layer of intestinal epithelial cells on the membrane (Fig. 1B) . Then the chip was placed on a rocker at a speed of 10 rpm and an angle of ± 8° to induce basal fluid flow. To further investigate the fluid dynamics within the chip, computational simulations were performed using COMSOL software. The results indicate that the fluid shear stress (~0.4 dyn/cm2) at the center of the membrane was relatively lower compared to the surrounding regions (0.8~2 dyn/cm2) (Fig. 1C) . Additionally, the fluid shear stress beneath the membrane remained stable at around 0.4 dyn/cm2(Fig. S3) . These observed shear force values align well with physiological data.[26]
It should be noted that although we focus on seeding cells on the upper side of the membrane in this work, the chip design (Fig. 1B) allows for easy cell seeding on the opposite side of the membrane by simply seeding cells in the bottom channel and inverting the chip. This unique feature enables the co-culture of multiple cell types, thus highlighting the scalability of our system and its potential for diverse applications. For instance, a multicellular model of inflammatory bowel disease can be established on this platform by seeding intestinal epithelial cells in the central well, inoculating intestinal endothelial cells on the opposite side of the membrane, and introducing immune cells into the bottom channel. Additionally, the versatility of the system extends further with the ability to control the cultivation conditions. By adjusting the angle and speed of the rocker, different shear forces can be easily applied to achieve distinct cultivation conditions for various experimental needs.
Characterization of the intestinal epithelium on theorgan-on-a-chip platform
Previous studies have elucidated the crucial role of basal-side fluid flow in promoting the morphogenesis of Caco-2 cells.[27, 28] This effect is attributed to the accumulation of Wnt signaling pathway inhibitors, such as DKK-1, beneath the membrane under static conditions, leading to hindered cell proliferation and impaired morphogenesis of the Caco-2 cell layer. The introduce of fluid flow can effectively eliminate the accumulation of these inhibitory molecules. In our research, a rocker with adjustable angle and speed was employed to induce reciprocating fluid flow. This dynamic flow not only dilutes the inhibitory molecules but also facilitates the occurrence of intestinal-like morphogenesis that closely mimics in vivo conditions.
In our chip system, a prominent distinction was observed between the perfused and static conditions. Under perfusion conditions, the intestinal cell layer cultured on the chip developed 3D villous structures, while cells without perfusion predominantly remained in a single-layer state. After three days of culture with fluid, noticeable disparities were observed under an optical microscope (Fig. 2A) . The intestinal cells under flow conditions formed convex and stacked structures, appearing as dark arc curves or circular structures with well-defined boundaries. Similar to transwell or petri dish culture, intestinal cells can be cultured on the chip for up to 30 days, but after the sixth day, there is no significant change in the morphology of the cells under the optical microscope (Fig. S4) . Consequently, cells after six days of fluid introduction were taken for characterization.
Live/dead staining images revealed that most of the cells exhibited normal growth after six days of perfused culture on the chip(Fig. 2D) . Additionally, a CCK-8 assay was performed to assess cell viability under static and perfused conditions, demonstrating that the viability of cells cultured on the chip was comparable to or even higher than that of cells cultured in a static state (Fig. 2E) . Furthermore, differences in cell layer height and microvillus morphology were observed through H&E staining (Fig. 2B) and SEM(Fig. 2C) , respectively. These results highlight the successful establishment of an intestinal-like morphogenesis on our chip system under perfusion conditions.
Improved functionality of intestinal epithelium under flow condition
To assess the integrity of the cell layer, we evaluated the permeability of the intestinal barrier by measuring the fluorescence intensity of FITC-dextran with varying molecular weights that permeated the cell layer within a specific time frame (Fig. 2F) . Notably, regardless of molecular weight, the Papp of dextran in the perfused chip was significantly lower than that observed in the static culture. In the 1-hour experiment, the Papp of 4 kDa-dextran was as low as 5 x 10-7cm/s, indicating a substantial enhancement in the integrity of the intestinal barrier in the perfusion environment.
To further assess the integrity of the intestinal barrier, we measured the TEER using a commercially available instrument, EVOM3, which was facilitated by our chip design (Fig. S3) . Interestingly, there was minimal disparity between the TEER values of the perfused and static cultures (Fig. 2G) , deviating from previous studies. Both groups displayed an upward trend in TEER over the six-day culture period, eventually reaching an approximate value of 100 Ω·cm2. While this value may seem relatively low compared to other reports, it is essential to note that co-culturing Caco-2 and HT-29 cells typically leads to lower TEER values.[29, 30] Consequently, the observed TEER value aligns more closely with physiological conditions in humans.
Immunofluorescence staining of ZO-1 (Fig. 3C) revealed the presence of well-formed tight junctions and brush border formation in the perfusion environment, indicating a more robust barrier compared to static culture. Moreover, the detection of Villin (Fig. 3F) in apical cells provided additional evidence for the presence of intestinal villi in the perfusion environment, which was further supported by SEM images showing well-defined microvilli (Fig. 2C) . Furthermore, by analyzing the distribution of nuclei in images at different levels, we could clearly visualize the hollow structure of intestinal villi(Fig. 3A) . The expression of MUC-2 was also detected. As shown in Fig. 3D , MUC-2 was evenly distributed on the intestinal villus, and superposition with the bright field image provided a clear view of the villi’s shape. Side-view confocal 3D images allowed us to calculate the height of the villi, and the results indicated that the average villi height in static culture was about 33 μm (Fig. 3E) , while the average villi height on the chip was close to 90 μm. Additionally, the expression of Ki67 (Fig. 3G) indicated active proliferation of intestinal cells, suggesting the potential for further growth into higher intestinal villi.