Engineered decellularized matrix hydrogels and methods for formation of functional intestinal crypt-villus epithelium

Abstract

A method for making an intestinal tissue model includes depositing a decellularized small intestine submucosa-norbornene (dSIS-NB) hydrogel within a sacrificial hydrogel intestine mold having negative crypts-villi structures formed thereon, and dissolving the sacrificial hydrogel intestine mold to thereby generate intestinal tissue model having positive crypts-villi structures formed thereon.

Description

PRF-71031-03ENGINEERED DECELLULARIZED MATRIX HYDROGELS AND METHODS FOR FORMATION OF FUNCTIONAL INTESTINAL CRYPT- VILLUS EPITHELIUMCROSS-REFERENCE TO RELATED APPLICATIONS

[0001] The present non-provisional patent application is related to and claims the priority benefit of U.S. Provisional Patent Application Serial No. 63 / 778,683, filed March 27, 2025, and which is related to and claims the priority benefit of U.S. Provisional Patent Application Serial No.63 / 904,178, filed October 23, 2025, the contents of each of which are hereby incorporated by reference in its entirety into the present disclosure.GOVERNMENT SUPPORT CLAUSE

[0002] This invention was made with government support under DK127436 awarded by the NIH. The government has certain rights in the invention.TECHNICAL FIELD

[0003] The present disclosure generally relates to methods and materials for generating artificially constructed structures which mimic intestinal epithelium useful for drug testing and disease modeling.BACKGROUND

[0004] The intestinal epithelium is a specialized layer of epithelial cells that lines the gastrointestinal tract in mammals. It plays a critical role in digestion, absorption, and barrier function, protecting the underlying tissues from pathogens and toxins. The epithelium is characterized by its high regenerative capacity, which is essential for maintaining intestinal homeostasis due to the constant exposure to mechanical stress and microbial challenge.PRF-71031-03

[0005] Human intestinal organoids (hTO) have emerged as a promising technology for mimicking the structure and phenotype of the small intestine. Derived from intestinal stem cells (ISC) or isolated intestinal crypts, intestinal organoids exhibit cellular diversity and selfrenewing properties. For example, Lee et al. differentiated ISCs into intestinal organoids containing crypt- villus-like domains that resembled the spatial architecture of native intestinal epithelium. Jelinsky et al. generated monolayers of patient-derived human colonic or ileum organoids capable of forming a functional epithelial barrier with sustained transepithelial / transendothelial electrical resistance (TEER). Nonetheless, most organoid models are limited by their inability to control crypt- villus architecture. In this regard, Anseth group developed engineered photo-responsive hydrogels to generate intestinal organoids with controlled crypt morphogenesis. Unfortunately, organoids generated within three-dimensional (3D) matrices often adopted an inverse polarity, with the apical surface facing inward and inaccessible to drug transport and absorption. While methods have been developed to generate intestinal organoid models with normal epithelial polarity (i.e., apical-out), the basolateral surfaces of the epithelium are still inaccessible for sampling transcytosed molecules. Therefore, there is an unmet need for a better culture system that addresses the organoid limitations for assay performance yet preserves the physiologically relevant 3D structure.SUMMARY

[0006] A method for making an intestinal tissue model is disclosed. The method includes depositing a decellularized small intestine submucosa-norbornene (dSIS-NB) hydrogel within a sacrificial hydrogel intestine mold having negative crypts-villi structures formed thereon, and dissolving the sacrificial hydrogel intestine mold to thereby generate intestinal tissue model having positive crypts-villi structures formed thereon.

[0007] In the above method, the dSIS-NB hydrogel is formed by a photochemical reaction after mixing a) decellularized small intestine submucosa (dSIS) matrix having one or more norbornenes (dSIS-NB), with b) 4-arm polyethylene glycol) (PEG) thiol (PEG4SH) as a crosslinker, and c) lithium phenyl-2,4,6-trimethylbenzoylphosphinate (LAP) as a photo-initiator.

[0008] In the above method, the sacrificial hydrogel intestine mold is a PEG-norbornene-tyramide (PEGNB-T) hydrogel network.PRE-71031-03

[0009] In the above method, the PEGNB-T hydrogel network is formed by a photochemical reaction after mixing a) PEGNB-T, with b) 4-arm PEG thiol (PEG4SH) as a cross-linker, and c) LAP as a photo-initiator.

[0010] In the above method, the sacrificial hydrogel intestine mold is dissolved by a water-based solvent.

[0011] In the above method, PEGNB-T is synthesized by:a) reacting 4-arm PEG-OH structure with carbic anhydride in the presence of 4- Dimethylaminopyridine (DMAP) as a catalyst, to form PEGNB -carboxylate (PEGNBCA) structures; andb) reacting the PEGNBCA with tyramine in the presence of 1,3-Diisopropylcarbodiimide (DIC), 1 -Hydroxybenzotriazole (HOBt), N,N-Diisopropylethylamine (DIPEA), and DMAP as catalysts to form PEGNB-T.

[0012] In the above method, the photochemical reaction is based on a concentration of LAP between about 0.5 mM to about 34 mM.

[0013] In the above method, the photochemical reaction is based on light having a wavelength of between about 365 nm to about 600 nm.

[0014] In the above method, the photochemical reaction is based on having an intensity of between about 2 to about 20 mW / cm2.

[0015] In the above method, the sacrificial hydrogel intestine mold is made by a digital light processing (DLP) three dimensional (3D) printer.PRF-71031-03

[0016] A sacrificial hydrogel intestine mold is also disclosed. The sacrificial hydrogel intestine mold includes a mold made from a hydrogel, having formed thereon negative crypts-villi structures. The hydrogel dissolves when placed in contact with a water-based solvent.

[0017] In the above artificial hydrogel intestine mold, the sacrificial hydrogel intestine mold is a polyethylene glycol)-norbornene-tyramide (PEGNB-T) hydrogel network.

[0018] In the above artificial hydrogel intestine mold, the sacrificial hydrogel intestine can be dissolved by a water-based solvent.

[0019] In the above artificial hydrogel intestine mold, the PEGNB-T hydrogel network is formed by a photochemical reaction after mixing a) PEGNB-T, with b) 4-arm PEG thiol (PEG4SH) cross-linkers, c) lithium phenyl-2,4,6-trimethylbenzoylphosphinate (LAP) as a photo-initiator, and d) tartrazine as a photo-absorber.

[0020] In the above artificial hydrogel intestine mold, PEGNB-T is synthesized by:a) reacting 4-arm PEG-OH structure with carbic anhydride in presence of 4- Dimethylaminopyridine (DMAP) as a catalyst, to form PEGNBCA structures; and b) reacting the PEGNBCA with tyramine in the presence of 1,3-Diisopropylcarbodiimide (DIC), 1-Hydroxybenzotriazole (HOBt), N,N-Diisopropylethylamine (DIPEA), and DMAP as catalysts to form PEGNB-T.

[0021] In the above artificial hydrogel intestine mold, the photochemical reaction is based on a concentration of LAP between about 0.5 mM to about 34 mM.

[0022] In the above artificial hydrogel intestine mold, the photochemical reaction is based on light having a wavelength of between about 365 nm to about 600 nm.

[0023] In the above artificial hydrogel intestine mold, the photochemical reaction is based on having an intensity of between about 12 to about 20 mW / cm2.

[0024] In the above artificial hydrogel intestine mold, the photochemical reaction is based on having a tartrazine concentration of between about 0.5 to about 2 mM.

[0025] In the above artificial hydrogel intestine mold, the sacrificial hydrogel intestine mold is made by a digital light processing (DLP) 3D printer.BRIEF DESCRIPTION OF FIGURESPRF-71031-03

[0026] FIG. la is a schematic of chemical synthesis of decellularized small intestine submucosa-norbornene (dSIS-NB) using triethylamine (TEA) as a base catalyst by reacting dSIS with carbic anhydride with a degree of NB substitution determined to be about 10%.

[0027] FIG. lb is a schematic of a chemical synthesis which provides a reaction scheme of thiolnorbornene photo-click reaction used for dSIS-NB hydrogel crosslinking.

[0028] FIG. 1c is a graph of shear moduli (G') in Pa vs. % by weight of PEG4SH representing a graph of hydrogel stiffness which was adjusted by tuning crosslinker (i.e., 4-arm PEG-thiol or PEG4SH) concentration (i.e., 1.2 wt%, 1.0 wt%, 0.8 wt%, or 0.6 wt%) at fixed dSIS-NB (1.2 wt%) and LAP contents (7mM).

[0029] FIG. Id provides representative immuno staining images of F-actin (white) and paxillin (red) staining showing differences in protrusions and focal adhesion 1-day post cell seeding on flat dSIS-NB hydrogel (top) or Matrigel (bottom). Scale bars represent 100 pm.

[0030] FIGs. le-lg provide graphs of (FIG. le) Caco-2, (FIG. If) HT29-MTX, and (FIG. 1g) Caco-2 / HT29-MTX vs. days (where 9:1, total cell density of IxlO5cells cm-2was observed) referring to live cell tracking of intestinal cell growth kinetics on thick (about 1.15 mm) and flat dSIS-NB hydrogels or Matrigel.

[0031] FIG. Ih provides representative immuno staining images of F-actin, EpCAM, E-Cad, Mucin-2, and ZO-1 in Caco-2 / HT29-MTX co-culture on flat dSIS-NB hydrogels (top) or Matrigel (bottom) after 21 days of culture, where scale bars represent 200 pm.

[0032] FIG. li provides a schematic for transepithelial / transendothelial electrical resistance (TEER) measurements across intestinal epithelium and TEER values of Caco-2 and HT29-MTX monocultures and cocultures on flat dSIS-NB hydrogel or Matrigel, measured on day 21, TEER values are presented as mean ± SEM (n = 2 for Matrigel or 4 for dSIS-NB hydrogels).

[0033] FIG. Ij provides a bar graph showing proteomic profiling of extracellular matrix proteins identified in dSIS and dSIS-NB, showing relative abundance of collagen types I- VI and fibrillins (FBN1, FBN2).

[0034] FIG. Ik provides snap-shot images of Caco-2 cells cultured on soft (about IkPa) and medium stiffness (about 2 kPa) dSIS-NB hydrogel.

[0035] FIG. 11 provides snap-shot images of Caco-2 cells cultured on flat dSIS-NB hydrogel (about 2 kPa) at various seeding densities, providing the optimal seeding density at IxlO5cells cm-2, where the cells form a uniform, well-organized monolayer, at lower seeding densities, thePRF-71031-03confluence of the monolayer was delayed, leading to gaps in the monolayer and a less uniform structure.

[0036] FIGs. Im, In, and lo provide snap-shot images showing biocompatibility assessment based on intestinal cell behaviors on flat dSIS-NB hydrogels and Matrigel. These figures show representative time-lapse image sequences that compare the behavior of intestinal cell lines cultured on flat dSIS-NB hydrogel and Matrigel over time. The three conditions include: FIG. Im for Caco-2 monoculture, FIG. In for HT29-MTX monoculture, FIG. lo for Caco-2 / HT29-MTX co-culture.

[0037] FIG. Ip provides a series of graphs showing gene expression results comparing dSIS-NB hydrogel and Matrigel.

[0038] FIGs. Iq. Ir, and Is provide image that show effect of hydrogel biochemical composition and stiffness on Caco-2 / HT29-MTX cell behavior. FIGs. Iq and Ir show GelNB-PEG4SH hydrogels were used as a control for culturing Caco-2 / HT29-MTX cells. GelNB at 5 wt% was crosslinked with PEG4SH at 0.9 wt% for a soft hydrogel with G’ about 1 kPa or with 1.3 wt% PEG4SH for a medium stiffness hydrogel with G’ of about 2 kPa. Caco-2 / HT29-MTX cell adhesion and growth on dSIS-NB hydrogels with similar stiffness of 1 kPa to 2 kPa were shown for comparison. FIG. Is shows Transwell used as a control for conventional models), suggesting both matrix stiffness and compositions are important in promoting the proliferation of intestinal cells.

[0039] FIG. It provides images showing assessment of morphology and cell-to-cell interaction in Caco-2 monoculture on flat dSIS-NB hydrogels or Matrigel.

[0040] FIG. 2a provides a schematic of a workflow of the inverse molding to generate dSIS-NB hydrogels with a positive crypt / villus topography. This process utilizes rapid and autonomous dissolution of sacrificial hydrogels in a trans-well insert, allowing for the precise formation of crypt / villus structures in the dSIS-NB hydrogel.

[0041] FIG. 2b provides an image of a design of the negative crypt / villus mold by Tinker CAD for DLP printing, according to one embodiment.

[0042] FIG. 2c provides an In situ photorheometry of PEGNB-T (4 wt%) crosslinked with PEG4SH (3 wt%).

[0043] FIG. 2d provides graphs showing a representative DLP-printed sacrificial PEGNB-T hydrogel with negative crypt / villus structures.PRF-71031-03

[0044] FIG. 2e provides 3D reconstruction of a confocal z-stack view and cross-section view of the printed PEGNB-T hydrogel with negative crypt / villus structure. A trace of rhodamine-PEG-thiol (Rh-PEG-SH) was added after the DLP-printing of sacrificial PEGNB-T hydrogel for visualization.

[0045] FIG. 2f provides an image of dSIS-NB crypt / villus hydrogel being placed in a trans-well insert following the degradation of the sacrificial PEGNB-T hydrogel mold.

[0046] FIG. 2g provides an image of a 3D reconstruction of a confocal z-stack view and crosssections of dSIS-NB hydrogels with positive crypt / villus structures. A trace of Rh-PEG-SH was added to the dSIS-NB hydrogel for visualization.

[0047] FIG. 2h provides 1H NMR spectra of PEG-OH, PEGNBca, and PEGNB-T.

[0048] FIG. 2i provides a graph showing fidelity of DLP-printed PEGNB-T mold and dSIS-NB villus / crypt topography, specifically, showing fidelity of DLP-printed PEGNB-T hydrogels with negative villus / crypt topography.

[0049] FIG. 2j provides a graph showing fidelity of dSIS-NB villus / crypt topography was performed on five samples (n=5), focusing on key geometrical parameters such as base diameter and height of the villi.

[0050] FIG. 2k provides a graph showing hydrolytic degradation of DLP-printed PEGNB-T hydrogels.

[0051] FIG. 3a provides a schematic of cell seeding and monolayer formation on the inverse-molded dSIS-NB hydrogel with positive crypt / villus structure.

[0052] FIG. 3b provides brightfield images of Caco-2 / HT29-MTX cells distributed on the scaffold surface after seeding (i.e., day 0) and day 3.

[0053] FIG. 3c provides a heatmap showing the average local cell coverage on 2D flat and 3D crypt / villus dSIS-NB hydrogels.

[0054] FIG. 3d provides a 3D confocal front-to-back view of an intestinal epithelium (stained with F-actin and nuclei formed over the dSIS-NB hydrogels with crypt / villus topography on day 3 of culture. Zoom-in images show the cell coverage on the top, middle, and bottom sections of a villus.

[0055] FIG. 3e provides images which show representative immuno staining images of E-cadherin and F-actin of day 3 to assess cell-cell interaction and cytoskeleton structure, respectively.PRF-71031-03

[0056] FIG. 3f provides immunostaining images of Ki67 and vinculin on day 3 to assess cell proliferation and focal adhesion complex formation, respectively.

[0057] FIG. 3g provides images showing formation of Caco-2 / HT29-MTX cell monolayer on the crypt-villus dSIS-NB hydrogels for the first three days.

[0058] FIG. 3h and FIG. 3i provide brightfield images of selective areas on dSIS-NB hydrogels with crypt / villus structure (see FIG. 3h) or flat surface (see FIG. 3i) after 3 days of Caco-2 / HT29-MTX cell culture. The crypt / villus topography (shown in FIG. 3h) promoted more organized monolayer formation, while the flat surface (shown in FIG. 3i) showed an uneven spreading.

[0059] FIG. 3j and FIG. 3k provide confocal z-stack images of immuno staining (Ki67-green; Vinculin-red, and Nuclei-blue) of Caco-2 / HT29-MTX cells cultured on dSIS-NB hydrogels with crypt / villus structure (shown in FIG. 3j) and flat surface (shown in FIG. 3k).

[0060] FIGs. 4a and 4b provide representative confocal images of immuno staining (Villin-red; F-actin-while; and Nuclei-blue) of Caco-2 / HT29-MTX monolayers cultured on crypt-villus dSIS-NB hydrogels, flat dSIS-NB hydrogels, and conventional Transwell membrane at day 7 (FIG. 4a) and day 21 (FIG. 4b).

[0061] FIG. 4c provides higher magnification of the dashed box regions provided in FIG. 4b.

[0062] FIG. 4d provides mRNA expression of TJP-1, the gene encoding tight junction protein ZO1.

[0063] FIG. 4e provides immunofluorescence staining of ZO1 and nuclei.

[0064] FIGs. 4f, 4g, 4h, and 4i provide images that show immunofluorescence staining for 3D crypt / villus structure after 21 days of culture.

[0065] FIG. 5a shows a schematic of crypt / villus barrier integrity and permeability assays.

[0066] FIG. 5b provides TEER measurements of the intestinal epithelial barrier on conventional trans-well, 2D flat, and 3D crypt / villus models. Data are presented as mean ± SEM (n = 4).

[0067] FIG. 5c provides confocal images of 4.4k-TD (i.e., tetramethyl rhodamine isothiocyanate glucan-dextran dye) and 500k-FD (i.e., fluorescein isothiocyanate-dextran dye) permeation in dSIS-NB hydrogels with flat, 3D crypt / villus structure, and cell free-hydrogel.

[0068] FIG. 5d provides a graph that shows quantification of fluorescence intensity for 4.4k-TD permeation from the apical to basolateral side of 2D flat and 3D crypt / villus dSIS-NB models. Normalized RFU values were determined by the RFU(tn) / RFU(to) ratio (n = 3).PRF-71031-03

[0069] FIG. 5e provides a graph that shows calculated apparent permeability coefficients (Papp) from 4.4k-TD permeation assays.

[0070] FIG. 6a provides brightfield images of the cell-laden crypt / villus layers under toxic stress conditions with different concentrations of staurosporine (after 24 hours of exposure).

[0071] FIG. 6b provides epithelial barrier disruption with different staurosporine concentration evaluated by TEER values over time for 24 hr. N

[0072] FIG. 6c provides confocal images present dextran leakage of 3D crypt / villus model after 24 hours of staurosporine exposure using 4.4k-TD and 500k-FD.

[0073] FIG. 6d provides timeline of the toxicity and transport assay. Staurosporine was added at the start of the experiment, while dextran was added at the time indicated post-staurosporine addition (i.e., 0, 9. 24 hr).

[0074] FIG. 6e and FIG. 6f provide effect of Staurosporine (5pM) on permeability of (FIG. 6e) 4.4k-TD and (FIG. 6f) 500k-FD in dSIS-NB hydrogel with crypt / villus structure.

[0075] FIG. 6g and FIG. 6h provide graphs showing effect of Staurosporine (5 pM) on the permeability of (FIG. 6g) 4.4 kDa TRITC-dextran and (FIG. 6h) 500 kDa FITC-dextran.

[0076] FIG. 7a provides a schematic of molecular transport in celiac disease and healthy crypt / villus epithelium.

[0077] FIG. 7b provides an image of design of a negative pizza mold by Tinker CAD and molded dSIS-NB ‘pizza’ hydrogel.

[0078] FIG. 7c provides brightfield images of the intestinal cells culture on the ‘pizza’ hydrogels.

[0079] FIG. 7d provides an image which shows a representative 3D overall view for dextran transport across the flat and crypt / villus epithelium on the pizza model.

[0080] FIG. 7e provide images that show representative cross-section view (x-z) of the pizza hydrogel model with dextran transport through the intestinal epithelial cell layer into the hydrogel.

[0081] FIG. 7f provides a graph that shows quantification of 4.4k-TD and 500k-FD transport along the distance from flat to transition and to crypt / villus zones.DETAILED DESCRIPTIONPRF-71031-03

[0082] For the purposes of promoting an understanding of the principles of the present disclosure, reference will now be made to the embodiments illustrated in the drawings, and specific language will be used to describe the same. It will nevertheless be understood that no limitation of the scope of this disclosure is thereby intended.

[0083] In the present disclosure, the term “about” can allow for a degree of variability in a value or range, for example, within 15%, within 10%, within 5%, or within 1% of a stated value or of a stated limit of a range.

[0084] In the present disclosure, the term “substantially” can allow for a degree of variability in a value or range, for example, within 85%, within 90%, within 95%, or within 99% of a stated value or of a stated limit of a range.

[0085] A novel method for making an intestinal tissue model utilizing a sacrificial hydrogel intestine mold made by made by a digital light processing (DLP) printing or other three-dimensional (3D) printing technologies is presented herein.

[0086] In the present disclosure, we report a novel two-step biofabrication approach to generate a biomimetic crypts / villi structure. In the first step, a sacrificial hydrogel mold (SHM) is formulated by 3D DLP printing of photocrosslinkable and dissolvable polyethylene glycol)-norbornene-tyramine (PEGNB-T) hydrogels. Next, the DLP-printed SHM is directly used to mold cell-instructive dSIS-norbornene (dSIS-NB) hydrogels without any surface coating process. Both dissolvable SHM hydrogels and cell-instructive dSIS-NB hydrogels were crosslinked by rapid thiol-norbornene photo-click reaction. After these two steps, the construct was inverted and incubated in trans- well inserts to allow for rapid and autonomous dissolution of the PEGNB-T-based SHM, leaving the cell-instructive dSIS-NB hydrogel with crypt / villus topography. The biological relevance of the dSIS-NB crypt / villi hydrogels was evaluated via intestinal epithelial cell growth and polarity, macromolecular permeability, and responses to toxic agents. Finally, a proof-of-concept intestinal disease model was generated using the same biofabrication design principle but with both intact crypt / villus and impaired / flat topographies on the same substrate, permitting side-by-side observation and comparison of dysfunction and healthy intestinal tissues.

[0087] Advances in 3D printing technologies have propelled the development of scaffold-based intestinal epithelium models. 3D-printed scaffolds can be tailored to mimic the native cryptvillus structure and biomechanical properties of intestinal tissue. Additionally, 3D-printed scaffolds maintain an open epithelial surface, making them suitable for biological assays,PRF-71031-03transepithelial / transendothelial electrical resistance (TEER) measurements, and drug absorption. A significant challenge of 3D-printed crypt-villus structures is the poor printability of most bioactive matrices. On the other hand, materials suitable for 3D printing are suboptimal for intestinal cell culture. Engineered matrices, such as gelatin methacryloyl (GelMA) and polyethylene glycol) diacrylate (PEGDA), could be co-polymerized into bioactive hydrogels to support intestinal epithelial cell growth. However, these materials lacked the structural complexity required to mimic the native crypt- villus arrangement. The crypt- villus structures can be fabricated using digital light projection stereolithography (DLP-SLA). However, Caco-2 cells cultured on this scaffold took 21 days to reach confluence and exhibited abnormal TEER values.

[0088] Intestinal epithelial cells provided with physiologically relevant bioactive cues, stiffness, and 3D crypt-villus structure significantly improve their barrier function. Towards this end, we adopted a decellularized small intestine submucosa-norbornene (dSIS-NB) hydrogels for their inherent bioactivity and their rapid and tunable thiol-norbornene crosslinking. While dSIS-NB can be 3D printed via extrusion and digital light processing (DLP) bioprinting, the dSIS-NB precursor solutions were too viscus for generating the intricate 3D crypt-villus structures using 3D printing. To address this issue, we generated dSIS-NB hydrogels with the critical crypt-villus structures through an inverse molding process. First, PEG-norbornene-tyramine (PEGNB-T), a photocrosslinkable sacrificial bioink, was 3D printed with the inverse crypt-villus structures by a DLP bioprinter. dSIS-NB and PEG-tetra-thiol (PEG4SH) crosslinker solution was then cast over the PEGNB-T hydrogel mold, and photocrosslinked into thiol-norbornene hydrogels with the correct crypt-villus structures. Finally, the entire construct (i.e., dSIS-NB + PEGNB-T hydrogel) was inverted and incubated in cell culture medium to allow for the rapid dissolution of the sacrificial PEGNB-T hydrogels. These inverse-molded dSIS-NB hydrogels with crypt-villus structures were placed into cell culture inserts to allow intestinal epithelium formation. How this 3D environment affects other aspects of intestinal epithelial cell function, such as barrier formation, permeability, and toxicity responses, were explored. To fully realize its potential in mimicking the intestinal epithelium in vitro, we generated a Celiac disease mimicking “pizza” like dSIS-NB hydrogel surface, which combined both crypt-villus structures and impaired / flat regions. We evaluated the impact of intestinal surface topography on cell proliferation rate and barrier function.PRF-71031-03

[0089] The ability to model the intestinal epithelium, including crypt structures, would be greatly useful for research and study of cell populations and formation of intestinal lining, as well as for drug discovery and study, and the examination of disease processes and morphologies.

[0090] Manufactured trans-well inserts have been widely adopted for studying intestinal cell functions due to their simplicity and accessibility. Intestinal epithelial cells are typically grown on a flat and rigid polycarbonate trans-well membrane coated with a thin layer of extracellular matrix (ECM), such as Matrigel and collagen. While drug transport studies can be readily performed using the trans-well inserts, this approach fails to mimic the intricate 3D crypt-villus architecture and mechanical properties of the native small intestine. These limitations hinder natural cell-matrix interactions essential for maintaining intestinal functions. Consequently, this model does not present physiological relevance and often leads to underestimated paracellular absorption and abnormally high transepithelial electrical resistance (TEER).

[0091] The advances in 3D printing technologies have propelled the development of scaffoldbased intestinal epithelium models. 3D-printed scaffolds can be tailored to mimic the native crypt-villus structure and biomechanical properties of intestinal tissue. Additionally, 3D-printed scaffolds maintain an open epithelial surface, making them suitable for biological assays, TEER measurements, and drug absorption. A significant challenge of 3D-printed crypt- villus structures is the poor printability of most bioactive matrices. On the other hand, materials suitable for 3D printing are suboptimal for intestinal cell culture. Engineered matrices, such as gelatin methacryloyl (GelMA) and polyethylene glycol) diacrylate (PEGDA), could be co-polymerized into bioactive hydrogels to support intestinal epithelial cell growth. However, these materials lacked the structural complexity required to mimic the native crypt-villus arrangement.

[0092] The present disclosure provides materials and methods for constructing structures which will provide intestinal epithelial cells with physiologically relevant bioactive cues, stiffness, and 3D crypt-villus structures, all of which are important for the formation of coherent epithelium with selective barrier functions. The crypt-villus structures can be fabricated using digital light projection stereolithography (DLP-SLA). However, Caco-2 cells cultured on this scaffold took 21 days to reach confluence and exhibited abnormal TEER values.

[0093] Suitable materials for the compositions and methods of the present invention include dSIS-NB (US Provisional Patent Application 63 / 554603) and PEGNB-T (US Patent publicationPRF-71031-0320240052097 Al ) contents of each of which are incorporated by reference in its entirety into the present disclosure.

[0094] The present disclosure provides for a method for making a hydrogel model of intestinal lining with an inverse-molded dSIS-NB crypt-villus hydrogel comprising inverse molding dSIS-NB over a sacrificial poly(ethylene glycol)-norbomene-tyramine (PEGNB-T) hydrogel 3D-printed by a digital light processing (DLP) bioprinter, where the 3D-printed sacrificial hydrogel was generated for molding intestinal crypt / villus structures on a highly bioactive hydrogel to form a model surface for growing epithelial cells.

[0095] The present disclosure also provides for a method of modeling a Celiac disease using a hydrogel molded structure mimicking dSIS-NB hydrogel surface, which combined both cryptvillus structures and impaired / flat regions

[0096] The present disclosure provides for the method of mimicking celiac disease as described where the method provides that the epithelial cells are grown onto a structure as shown in Figures discussed below, to mimic celiac disease structure.

[0097] Thus, the present disclosure provides for modeling in vivo intestinal epithelial structures that were both intact and showing disease states, where the method provides for hydrogels that are generated with the critical crypt- villus structures through an inverse molding process, such as:a. PEG-norbornene-tyramine (PEGNB-T), a photocrosslinkable sacrificial bioink, was 3D printed with the inverse crypt-villus structures by a DLP bioprinter;b. dSIS-NB and PEG-tetra-thiol (PEG4SH) crosslinker solution was then cast over the PEGNB-T hydrogel mold, and photocrosslinked into thiol-norbornene hydrogels with the correct crypt-villus structures;c. the entire construct of dSIS-NB + PEGNB-T hydrogel was inverted and incubated in cell culture medium to allow for the rapid dissolution of the sacrificial PEGNB-T hydrogels. d. the inverse-molded dSIS-NB hydrogels with crypt-villus structures were placed into cell culture inserts to allow intestinal epithelium formation.

[0098] It is further provided for in the present disclosure, completed hydrogel structures generated by the methods described above, which provide models for intact and disease state intestinal epithelial linings.PRF-71031-03

[0099] It is well-established that Engineered matrices that mimic intestinal epithelium are highly desirable for testing drug adsorption and modeling intestinal diseases. However, generating a functional small intestine model is challenging owing to the complex geometry of the cryptvillus structure. In this disclosure, we demonstrate a unique biofabrication strategy to address this challenge. Specifically, a 3D-printed sacrificial hydrogel was generated for molding intestinal crypt / villus structures on a highly bioactive hydrogel. Hydrogels with stiffness suitable for intestinal cell growth were prepared by dSIS-NB via thiol-norbornene photoclick reaction. The crypt- villus structures were generated by inverse molding of dSIS-NB over a sacrificial polyethylene glycol)-norbornene-tyramine (PEGNB-T) hydrogel 3D-printed by a digital light processing (DLP) bioprinter. The combination of bioactive dSIS substrates with proper stiffness and crypt- villus structure was proven to be essential for rapid formation of functional intestinal epithelium, which was achieved in three days, an achievement that has not been reported in the literature. Through macromolecular transport and transepithelial electrical resistance (TEER) measurements, the new inverse-molded dSIS-NB crypt-villus hydrogels further demonstrated the selective barrier functions under normal culture and toxic compound treatment. Finally, the utility of this hydrogel model for modeling Celiac disease was demonstrated via generating a hybrid crypt-villus / flat substrate on the same hydrogel matrix.

[0100] Generating a functional intestine model is imperative for drug testing and disease modeling, but it is challenging owing to the intestinal crypt / villus structure. The present disclosure provides an inverse molding biofabrication technique to address this challenge. To generate the desired crypts / villi structures, a sacrificial hydrogel mold (SHM) with negative crypt / villus features was first fabricated using digital light processing (DLP) 3D printing of poly(ethylene glycol)-norbornene-tyramine (PEGNB-T) thiol-norbornene hydrogels, although other 3D printing techniques are within the ambit of the present disclosure. Next, decellularized small intestine submucosa-norbornene (dSIS-NB) solution was cast and photopolymerized over the SHM, also via efficient thiol-norbornene photoclick reaction using PEG-tetra-thiol as the crosslinker. The hydrogel construct was placed in a buffer solution to induce autonomous and rapid dissolution of the SHM, generating dSIS-NB hydrogels with the positive crypts / villi structure. Intestinal epithelial cells (Caco-2 and MTX-HT29) seeded on the dSIS-NB crypts / villi matrices formed a confluent monolayer within 3 days and displayed correct intestinal polarity. Through transepithelial electrical resistance (TEER) measurements and macromolecularPRF-71031-03transport studies, the new inverse-molded dSTS-NB crypt / villus model further demonstrated the selective and drug-responsive barrier functions. Finally, the unique biofabrication technique was leveraged to generate an intestinal disease model carrying regions of both normal crypts-villi and flattened epithelium. Only the crypts-villi regions permitted selective molecular transport.

[0101] Referring to FIG. la, a schematic of chemical synthesis of dSIS-NB is provided using triethylamine (TEA) as a base catalyst by reacting dSIS with carbic anhydride with a degree of NB substitution determined to be about 10%. Proteomics analysis results show that both dSIS and dSIS-NB exhibited similar protein makeup, with more than 60% of type I collagen (COL1A1 and COL1A2), about 20% of type III collagen (COL3A1), and about 15% of fibrillins (FBN1 and FBN2) (see FIG. Ij, which provides a bar graph showing proteomic profiling of extracellular matrix proteins identified in dSIS and dSIS-NB, showing relative abundance of collagen types I-VI and fibrillins (FBN1, FBN2)). These structural proteins establish a fibrillar matrix that is distinctly different from commonly used matrices such as Matrigel, which is rich in glycoproteins (e.g., laminin) and collagen IV. The immobilization of NB did not alter protein composition in dSIS, but afforded thiol-norbornene photocrosslinking (see FIG. lb, which provides a reaction scheme of thiol-norbornene photo-click reaction used for dSIS-NB hydrogel crosslinking) into hydrogels with physiologically relevant moduli (G’ about 1.0 to about 2.1 kPa, see FIG. 1c which provides a graph of hydrogel stiffness which was adjusted by tuning crosslinker (i.e., 4-arm PEG-thiol or PEG4SH) concentration (i.e.. 1.2 wt%, 1.0 wt%, 0.8 wt%, or 0.6 wt%) at fixed dSIS-NB (1.2 wt%) and LAP contents (7mM)). Light at wavelength: 365 nm, intensity: 5 mW cm-2for 2 minutes was used. Matrigel was used as a control. Data provided represents mean ± standard error of the mean (SEM); n = 3, * and **** representp < 0.05 and 0.0001, respectively, by ordinary one-way ANOVA with Dunnert’s multiple comparisons test, with a single pooled variance. Chemically crosslinked dSIS-NB hydrogels supported early Caco-2 cell adhesion on dSIS-NB hydrogels, as revealed by prominent stress fibers and lamellipodia / filopodia-like protrusions (see FIG. Id which provides representative immuno staining images of F-actin (white) and paxillin (red) staining showing differences in protrusions and focal adhesion 1-day post cell seeding on dSIS-NB hydrogel (top) or Matrigel (bottom). Scale bars represent 100 pm. In contrast, Caco-2 cells seeded on Matrigel presented an aggregated morphology with few stress fibers. Interestingly, dSIS-NB hydrogels with shear moduli (G’) of about 1 kPa supported initial cell attachment and proliferation, but the gelsPRF-71031-03degraded after 10 days of culture (see FIGs. Ik and 11 which provide images showing effect of hydrogel stiffness and Caco-2 seeding density on the formation of cell monolayers on dSIS-NB hydrogel). Specifically, FIG. Ik provides snap-shot images of Caco-2 cells cultured on soft (about IkPa) and medium stiffness (about 2 kPa) dSIS-NB hydrogel. The (X) image refers to a situation where the hydrogel was degraded and unable to support cells form a complete monolayer. Furthermore, FIG. 11 provides snap-shot images of Caco-2 cells cultured on dSIS-NB hydrogel (about 2 kPa) at various seeding densities, providing the optimal seeding density at IxlO5cells cm'2, where the cells form a uniform, well-organized monolayer, at lower seeding densities, the confluence of the monolayer was delayed, leading to gaps in the monolayer and a less uniform structure. Conversely, at higher seeding densities, the cells accumulated unevenly, leading to overcrowding that did not achieve a uniform monolayer. Scale bars represent 200 pm. At the same dSIS-NB concentration (1.2 wt%), increasing crosslinker PEG4SH content led to stiffer dSIS-NB hydrogels (G’ about 2 kPa) that supported long-term stability (over three weeks) for cell attachment and proliferation (FIG. Ik). This result aligns with the existing literature on the effects of substrate stiffness on cytoskeleton and focal adhesion dynamics. The optimal cell seeding density was evaluated, with IxlO5cells cm-2determined to be ideal for producing a uniform epithelial monolayer (see FIG. 11).

[0102] After identifying dSIS-NB gel stiffness suitable for long-term intestinal cell culture, cell spreading was tracked over three weeks using live cell imaging (see FIGs. Im, In, and lo which provide snap-shot images showing biocompatibility assessment based on intestinal cell behaviors on dSIS-NB hydrogels and Matrigel). These figures show representative time-lapse image sequences that compare the behavior of intestinal cell lines cultured on dSIS-NB hydrogel and Matrigel over time. The three conditions include: FIG. Im for Caco-2 monoculture, FIG. In for HT29-MTX monoculture, FIG. lo for Caco-2 / HT29-MTX co-culture. Scale bar represents 200 pm. The images demonstrated that dSIS-NB hydrogel supports the proliferation and organization of intestinal epithelial cells, both monocultures and co-cultures, more effectively than Matrigel. A cellular network covering more than 50% of the dSIS-NB hydrogel surface was observed by day 7, and full coverage occurred within 13 days (see FIG. Im). In contrast, Caco-2 cell aggregates formed on Matrigel covered only about 30% of the surface after 21 days (see FIG. Im and FIGs. le-lg which provide graphs of le) Caco-2, If) HT29-MTX, and 1g) Caco-2 / HT29-MTX vs. days (where 9:1, total cell density of IxlO5cells cm-2was observed) referring to livePRF-71031-03cell tracking of intestinal cell growth on thick (about 1.15 mm) and flat dSIS-NB hydrogels or Matrigel). The maturation of Caco-2 cells after 21 days of culture was verified by higher expression of key intestinal epithelial genes, including VIL-1, MDR1, CCND1.ALPI, and SLC15A (see FIG. Ip which provides a series of graphs showing maturation of Caco-2 cells cultured on dSIS-NB hydrogel and Matrigel). mRNA expression levels of intestinal epithelial markers in Caco-2 cells cultured on dSIS-NB hydrogels and Matrigel are provided including VIL1: Villin, MDR1: multidrug resistance, CCND1: cell cycle marker, cyclin DI, ALPI: intestinal alkaline phosphatase, and SLC15A: transporter gene. Data are presented as mean ± SEM; (n = 3, *p < 0.05, * *p < 0.01, ***p < 0.001, andmp < 0.0001 by Two-way ANOVA Multiple Comparisons Tukey’s Post Hoc Test). The expression of these markers significantly increased on day 21 in Caco-2 cells cultured on both dSIS-NB hydrogel and Matrigel, reflecting enhanced epithelial differentiation and functional maturation. VIL-1 and CCND1 were particularly upregulated in cells grown on dSIS-NB hydrogels, highlighting their role in promoting functional monolayer formation and cell proliferation, respectively.

[0103] dSIS-NB hydrogels also supported faster spreading of HT29-MTX cells, a mucussecreting goblet-like cell line. However, unlike Caco-2 cells that formed flattened morphology on dSIS-NB hydrogels, HT29-MTX cells formed multicellular clusters with less spreading, reaching about 88% confluence after 21 days (see FIG. In and FIG. If). Mixing Caco-2 and HT29-MTX cells yielded a similar growth pattern to Caco-2 monoculture, achieving complete coverage of the dSIS-NB hydrogel surface within 13 days (see FIG. Io and FIG. 1g). In contrast, co-culture on Matrigel never reached monolayer coverage. As Matrigel was substantially softer than dSIS-NB hydrogels, we performed another control experiment using gelatin-norbornene (GelNB) hydrogels with similar stiffness to that of dSIS-NB hydrogels (i.e., G’ about 1 or 2 kPa). Five cell imaging of Caco-2 / HT29-MTX cells on GelNB hydrogels revealed limited cell expansion over 4 days on GelNB gels, regardless of gel stiffness (see FIGs. Iq, Ir. and Is which provide image that show effect of hydrogel biochemical composition and stiffness on Caco-2 / HT29-MTX cell behavior. FIGs. Iq and Ir show GelNB-PEG4SH hydrogels were used as a control for culturing Caco-2 / HT29-MTX cells. GelNB at 5 wt% was crosslinked with PEG4SH at 0.9 wt% for a soft hydrogel with G’ at about 1 kPa or with 1.3 wt% PEG4SH for a medium stiffness hydrogel with G’ about 2 kPa. Caco-2 / HT29-MTX cell adhesion and growth on dSIS-NB hydrogels with similar stiffness of 1 kPa to 2 kPa were shown for comparison. FIG. Is showsPRF-71031-03Transwell used as a control for conventional models), suggesting both matrix stiffness and compositions are important in promoting the proliferation of intestinal cells. Additionally, Caco-2 / HT29-MTX cells seeded on Transwell membrane achieved a confluent cell monolayer within four days of culture (see FIG. Is).

[0104] Images from live cell tracking over 21 days suggest that stiffer dSIS-NB hydrogels (G’ about 2 kPa) support long-term intestinal cell culture (see FIGs. Im-lo). On dSIS-NB hydrogels, both Caco-2 monoculture (see FIG. It which provides images showing assessment of morphology and cell-to-cell interaction in Caco-2 monoculture on dSIS-NB hydrogels or Matrigel). Immuno staining of key markers involved in cell morphology and cell-to-cell interactions in Caco-2 monoculture cultured on dSIS-NB hydrogels or Matrigel. F-actin (actin cytoskeleton), EpCAM (epithelial cell adhesion protein), E-Cad (E-cadherin, a cell-cell adhesion protein), Mucus-2 (mucin protein), and ZO-1 (tight junction protein). Mucus-2 was not detected in Caco-2 monocultures) and Caco-2 / HT29-MTX coculture (see FIG. Ih which provides representative immuno staining images of F-actin, EpCAM, E-Cad, Mucin-2, and ZO-1 in Caco-2 / HT29-MTX co-culture on dSIS-NB hydrogels (top) or Matrigel (bottom) after 21 days of culture, where scale bars represents 200 pm) formed a thin flat monolayer (20 - 25 pm), with strong EpCAM and uniform E-Cad expression. In contrast, cells formed large (600 - 700 pm in diameter) and thicker (120 - 130 pm in thickness) clusters on Matrigel (see FIG. Ih and FIG. It). Additionally, more uniform ZO-1 expressions were found in Caco-2 monoculture on dSIS-NB hydrogels but not in cell aggregates formed on Matrigel, while co-culture weakened the expression of ZO-1 (see FIG. Ih), suggesting a weaker tight junction than in Caco-2 monoculture. The addition of HT29-MTX cells also led to Mucin-2 secretion on both matrices, but the accumulation of Mucin-2 was more pronounced and wide spread on dSIS-NB hydrogels than on Matrigel (see FIG. Ih). TEER measurements confirmed the intestinal barrier integrity in Caco-2 monoculture and Caco-2 / HT29-MTX co-culture (see FIG. li which provides TEER values of Caco-2 and HT29-MTX monocultures and cocultures on dSIS-NB hydrogel or Matrigel, measured on day 21, TEER values are presented as mean ± SEM (n = 2 for Matrigel or 4 for dSIS-NB hydrogels)). Low TEER values were detected from cells cultured on Matrigel regardless of cell compositions and HT29-MTX monoculture on dSIS-NB hydrogels, reflecting their lack of monolayer coverage (see FIGs. le-lg, and FIGs. Im-lo). The incorporation of HT29-MTX cells also caused a slight reduction in TEER value (624 Q-cm2) as compared toPRF-71031-03Caco-2 monoculture (746 fi-cm2, see FTG. li). This result is consistent with the disrupted ZO-1 expression patterns in Caco-2 / HT29-MTX coculture (see FIG. Ih). The notable differences in cell behaviors could be attributed to hydrogel stiffness and composition. The dSIS-NB hydrogel, with its adjustable mechanical properties (G’ about 2 kPa), provided the necessary mechanical cues for cell attachment and spreading. This was in stark contrast to the performance of Matrigel, where Caco-2 cells exhibited only limited growth and multicellular clustering, likely due to the low stiffness of Matrigel (G' < 100 Pa). Scaffold stiffness is widely recognized as an important factor in regulating cell behavior, particularly for epithelial cells. Previous studies have shown that soft substrates can limit epithelial cell spreading and promote cluster formation rather than the formation of a continuous monolayer. Zhang et al. used a tunable matrix composed of Matrigel and synthetic oligo(ethylene glycol)-grafted polyisocyanides to show that cell morphology shifted from round hollow cysts to 2D monolayers at higher matrix stiffness. Our proteomics results also show that collagen I, collagen III, and Fibrillin made up more than 95% of dSIS and dSIS-NB (see Ij), whereas Matrigel was rich in laminin and collagen IV. The differences in biochemical compositions may also contribute to the different cell morphologies observed between dSIS-NB hydrogels and Matrigel.

[0105] While dSIS-NB hydrogels supported the formation of monolayer intestinal epithelium within 2 weeks (FIGs. la- li), the flat surface did not mimic the topography of intestinal tissue in vivo. One option to generate crypt / villus topography is using DLP bioprinting. Although dSIS-NB can be used as a bioink for 3D printing, its high viscosity prevents its use in DLP printing to produce the fine crypt-villus structure with a dimension between 50 to 200 pm. Therefore, an inverse molding technique was utilized where the bioactive dSIS-NB hydrogel was photocrosslinked over a sacrificial hydrogel mold (SHM) with negative crypt / villus structures printed by a DLP bioprinter (see FIG. 2a which provides a schematic of a workflow of the inverse molding to generate dSIS-NB hydrogels with a positive crypt / villus topography. This process utilizes rapid and autonomous dissolution of sacrificial hydrogels in a trans-well insert, allowing for the precise formation of crypt / villus structures in the dSIS-NB hydrogel). The negative crypt / villus mold was designed using Tinker CAD, with larger wells for molding villi (500x300x100 pm, for the depth, upper, and lower diameter, respectively) and small pillars for molding crypts (200x100x150 pm for the height, upper, and lower diameter, respectively) (see FIG. 2b which provides an image of a design of the negative crypt / villus mold by Tinker CADPRF-71031-03for DLP printing, according to one embodiment). The dSIS-NB precursor solution was cured over SHM, which undergoes rapid dissolution upon contact with PBS at 37 °C (see FIG. 2c which provides an In situ photorheometry of PEGNB-T (4 wt%) crosslinked with PEG4SH (3 wt%) and 10 mM LAP under 405 nm light at 34 mW cm'2. The gelation was modulated by the addition of photoabsorber tartrazine at concentrations of 0 mM, 0.75 mM, and 1.5 mM, which corresponded to gel points of about 2 s, about 5 s. and about 12 s, respectively, indicating tunable photopolymerization speed). PEGNB-T was synthesized via reacting PEGNBCA with tyramine using standard carbodiimide chemistry (see FIG. 2h which provides 1H NMR spectra of PEG-OH, PEGNBCA, and PEGNB-T. The comparison of the proton nuclear magnetic resonance (1H NMR) spectra of three different polymers: PEG-OH, PEGNBCA, and PEGNB-T are provided. Peak a represents PEG alkane protons. Peak b represents norbornene alkane protons, Peaks c represents norbornene alkene protons, and peaks d represents tyramine aromatic protons. PEG-OH shows the characteristic peaks of the polyethylene glycol backbone, PEGNBCA shows the NMR spectrum of PEG functionalized with carbic anhydride clearly displays additional peaks for corresponding to the strained alkene (c, about 6.4 ppm) and the alkane protons (b, about 1.4), PEGNB-T shows the spectrum of PEGNB-T which exhibits the typical PEG backbone peak a (about 4 ppm), norbornene peaks b (about 1.4 ppm) and c (about 6.4 ppm), as well as additional peaks for aromatic ring protons, and (d) in the range of 6.6-7.0 ppm, indicating successful conjugation of tyramine(T) to PEGNB). Similar to other norbornene-modified macromers, the crosslinking of PEGNB-T was rapid under 405 nm light, even in the presence of photoabsorber tartrazine (see FIG. 2d which provides graphs showing a representative DLP-printed sacrificial PEGNB-T hydrogel with negative crypt / villus structures), which was added to reduce out-of-focus polymerization and improve the printing fidelity. The PEGNB-T SHM was printed (see FIG. 2e which provides 3D reconstruction of a confocal z-stack view and cross-section view of the printed PEGNB-T hydrogel with negative crypt / villus structure. A trace of rhodamine-PEG-thiol (Rh-PEG-SH) was added after the DLP-printing of sacrificial PEGNB-T hydrogel for visualization, and see FIG. 2f which provides an image of dSIS-NB crypt / villus hydrogel being placed in a trans-well insert following the degradation of the sacrificial PEGNB-T hydrogel mold) with high fidelity (90% for the negative crypts and nearly 100% for the negative villi) (see FIG. 2i which provides a graph showing fidelity of DLP-printed PEGNB-T mold and dSIS-NB villus / crypt topography, specifically, showing fidelity of DLP-printed PEGNB-T hydrogels withPRF-71031-03negative villus / crypt topography. Quantitative analysis was performed on seven samples (n=7), focusing on key geometrical parameters such as upper diameter of the wells and the tip diameter of the pillars. DLP printing achieved high fidelity in replicating the intricate topography, reflecting the precision of the photopolymerization process).

[0106] Following DLP-printing of SHM, dSIS-NB / PEG4SH / LAP solution was pipetted into the mold and crosslinked by thiol-norbornene photo-click reaction. The construct was inverted and placed in a trans-well insert (see FIG. 2g which provides an image of a 3D reconstruction of a confocal z-stack view and cross-sections of dSIS-NB hydrogels with positive crypt / villus structures. A trace of Rh-PEG-SH was added to the dSIS-NB hydrogel for visualization) to allow for the rapid and autonomous dissolution of PEGNB-T SHM. Positive crypts / villi topography on dSIS-NB hydrogel (8 mm in diameter) was revealed after the dissolution of SHM, which occurred in 2 hours without any manual peeling or addition of any degradation-facilitating reagent (see FIG. 2k which provides a graph showing hydrolytic degradation of DLP-printed PEGNB-T hydrogels. Tartrazine concentrations tested are 0 mM (control, red diamonds), 0.75 mM (blue triangles), and 1.5 mM (green squares). The G’ of the hydrogels is measured upon fabrication, after 60 minutes and 120 minutes (n=3) to assess the impact of tartrazine on shortterm hydrolytic degradation. Tartrazine did not affect significantly to the final hydrolytic degradation rate of PEGNB-T hydrogels after 120 min). The average villus height and crypt depth were 318 pm and 88 pm, and the average diameters of the villus base and tip were 228 pm and 99 pm, respectively (see FIG. 2h). Molded dSIS-NB crypts / villi hydrogel achieved high fidelity and consistent uniformity (p > 0.05) (see FIG. 2j which provides a graph showing fidelity of dSIS-NB villus / crypt topography was performed on five samples (n=5), focusing on key geometrical parameters such as base diameter and height of the villi. Molded dSIS-NB achieved high fidelity and consistent uniformity (p > 0.05, t-test) across the samples. The results confirmed the reproducibility of the sacrificial molding approach to fabricating the intricate crypt-villus topography on dSIS-NB hydrogel with high fidelity).

[0107] Compared with prior studies using gelatin methacryloyl (GelMA) and PEG-diacrylate, our biofabrication method presents unique advantages, including the use of highly bioactive dSIS matrices and high-fidelity DLP-printing of autonomous PEGNB-T SHM. Elomaa et al. used DLP printing to generated villi structures from dSIS-methacrylate (dSIS-MA). Unfortunately, the high viscosity of dSIS-MA resulted in low fidelity (about 58%), high swelling (30 to 40%) post-PRF-71031-03printing, and a lack of crypts. Rudolph et al. employed a similar inverse-molding technique to produce silk-based scaffolds featuring crypt / villus structures. However, the multi-step process involved (1) DLP-printing of a polymer resin, (2) casting of PDMS negative mold, (3) coating and curing of a silk film on the PDMS mold, (4) curing of silk-spongy scaffold, (5) manual peeling of the scaffold, and (6) coating of collagen to provide cell adhesion sites. In contrast, our PEGNB-T SHM was DLP-printed in a single step, followed by rapid gelation of dSIS-NB gels over the SHM without the need for any coating. The removal of SHM occurred autonomously through rapid hydrolysis and without user intervention. This simple process generated bioactive dSIS-NB hydrogels with physiologically relevant stiffness (G’ about 2 kPa) and intricate crypt / villus topography.

[0108] The formation of intestinal cell monolayer on dSIS-NB hydrogels with crypt / villus topographies was evaluated using intestinal epithelial cells Caco-2 and mucus- secreting HT29-MTX cells (see FIG. 3a which provides a schematic of cell seeding and monolayer formation on the inverse-molded dSIS-NB hydrogel with positive crypt / villus structure). After seeding, cells clustered predominantly within the crypts and surrounding the bases of the villi, and very few cells were found on the tips of villi (day 0 of FIG. 3b which provides brightfield images of Caco-2 / HT29-MTX cells distributed on the scaffold surface after seeding (i.e., day 0) and day 3. Cell clustering in a crypt on day 0 and a full coverage of a villus on day 3 were highlighted). Over time, the cells proliferated to cover the remaining area and eventually formed a complete epithelial monolayer within 3 days (see day 3 of FIG. 3b). This was significantly faster than the flat 2D culture counterpart, which took 13 days to reach complete surface coverage (see FIGs. le-lg). While cell coverage on both flat dSIS-NB gel surface and crypt / villus structure was about 12 - 13% of the gel surfaces following initial cell seeding (day 0 of FIG. 3c which provides a heatmap showing the average local cell coverage on 2D flat and 3D crypt / villus dSIS-NB hydrogels. Coverage was assessed from 4 random regions of each hydrogel, with 3 independent replicates per condition. Data are presented as mean ± SEM. Cell coverage reached nearly 100% on dSIS-NB gel with crypt / villus topography on day 3, outperforming flat hydrogels, n.s.: Nonsignificant; ** and *** represent p<0.01 and 0.001, respectively, by multiple paired t tests at the same time point), cells seeded on 2D flat surfaces clustered randomly, with a large portion of the area void of cells, prolonging the time it took to reach full coverage. After one day, cells on the flat surface had just begun spreading in a few regions, reaching about 18% confluence, whilePRF-71031-03cells on crypt / villus scaffolds had already started spreading at the base of villi, achieving about 38% confluence (see FIG. 3c, day 1). Monolayer formation over the entire crypt / villus structure was verified through immunofluorescence staining of F-actin (see FIG. 3d which provides a 3D confocal front-to-back view of an intestinal epithelium (stained with F-actin and nuclei formed over the dSIS-NB hydrogels with crypt / villus topography on day 3 of culture. Zoom-in images show the cell coverage on the top, middle, and bottom sections of a villus), (see FIG. 3g which provides images showing formation of Caco-2 / HT29-MTX cell monolayer on the crypt-villus dSIS-NB hydrogels for the first three days. Cell morphology and organization assessed using F-actin staining and DAPI counterstaining for nuclei. Scale bars represents 200 pm. This early morphogenesis suggests that the crypt-villus architecture of dSIS-NB hydrogel promotes rapid and organized epithelial layer formation). On the first day, cells efficiently clustered together in the crypt while cells on the base of villi began to spread. By the second day, the crypts and villus base were almost completely covered while no cells were found on the villus projections.Impressively, a complete cell monolayer covering the entire crypt / villus structure was observed by the third day. In contrast, cells on the flat surface still exhibited high variation in local cell coverage, with some areas full of cells while others showed very low coverage (See FIG. 3c), (see FIG. 3h and FIG. 3i which provide rightfield images of selective areas on dSIS-NB hydrogels with crypt / villus structure (see FIG. 3h) or flat surface (see FIG. 3i) after 3 days of Caco-2 / HT29-MTX cell culture. The crypt / villus topography (shown in FIG. 3h) promoted more organized monolayer formation, while the flat surface (shown in FIG. 3i) showed an uneven spreading). Additional immunostaining of E-cadherin (see FIG. 3e which provides images which show representative immunostaining images of E-cadherin and F-actin of day 3 to assess cellcell interaction and cytoskeleton structure, respectively. Nuclei were counter-stained with DAPI. Strong E-cadherin expression was observed on the villi), as well as Ki67 and vinculin (see FIG.3f which provides immunostaining images of Ki67 and vinculin on day 3 to assess cell proliferation and focal adhesion complex formation, respectively. Ki67-positive cells were evenly distributed along crypt / villus axis, suggesting active cell division, while vinculin, marking focal adhesion sites, was concentrated in the crypts, reflecting stronger cell-ECM interactions in these areas. Nuclei were counter-stained with DAPI), (see FIG. 3j and FIG. 3k which provide confocal z-stack images of Caco-2 / HT29-MTX cells cultured on dSIS-NB hydrogels with crypt / villus structure (shown in FIG. 3j) and flat surface (shown in FIG. 3k). Immuno staining forPRF-71031-03Ki67, a proliferation marker, and Vinculin, a focal adhesion protein, was performed to assess cell proliferation and adhesion on day 3. Cells cultured on the crypt / villus structure (FIG. 3j) exhibited evenly Ki67 expression along with well-defined Vinculin localization. In contrast, cells on the flat surface (FIG. 3k) showed uneven Ki67 expression and less defined Vinculin staining. Scale bars represent 200 pm) revealed that the cells were highly proliferative and formed strong focal adhesion complexes throughout the 3D crypt / villus surface than on 2D flat surface (see FIGs. 3h-3k). In a previous work, Castano et al. used a photomask to generate villi-like pillars on much stiffer PEGDA hydrogels (>10 kPa). While Caco-2 monolayer was achieved in about 50 hours, the PEGDA hydrogel did not contain crypts, and a secondary protein conjugation step was needed to allow cell attachment on the otherwise inert PEGDA gel surface. In contrast, our dSIS-NB hydrogels were inherently cell adhesive and the inverse-molding process generated both crypts and villi on gels with physiological relevant stiffness (about 2 kPa). Evenly distributed crypts may act as cell "niches" that promote early cell aggregation and efficient cell-to-cell interaction, leading to faster spreading. All together, these results suggest that the topographical features of the crypt / villus structure not only promote more uniform cell seeding but also enhance initial cell adhesion and even spreading, resulting in more efficient coverage compared to flat surfaces.

[0109] To further assess epithelial polarization, we performed immunofluorescence staining and imaging of Villin, a well-established marker of brush border and microvilli formation (see FIGs.4a-4c, which provide images and a graph that show process of evaluation of epithelial polarization and tight junction on different scaffolds, specifically, FIGs. 4a and 4b provide representative confocal images of Caco-2 / HT29-MTX monolayers cultured on crypt-villus dSIS-NB hydrogels, flat dSIS-NB hydrogels, and conventional Transwell membrane at day 7 (FIG. 4a) and day 21 (FIG. 4b). Cells were stained for Villin to visualize the brush border and microvilli formation. F-actin was stained to assess cytoskeletal organization. Nuclei were counterstained with DAPE FIG. 4c provides higher magnification of the dashed box regions provided in FIG. 4b. The orientation of the monolayers shows the apical side facing up and the basal side facing down. FIG. 4d provides mRNA expression of TJP-1, the gene encoding tight junction protein ZO1. Data are presented as mean ± SEM; (n = 3,*p < 0.05, * *p < 0.01, ***p < 0.001, and *** *p < 0.0001 by Two-way ANOVA Multiple Comparisons Tukey’s Post Hoc Test). FIG. 4e provides immunofluorescence staining of ZO1, aPRF-71031-03tight junction protein). Caco-2 / HT29-MTX monolayers were cultured on crypt- villus dSTS-NB hydrogels, flat dSIS-NB hydrogels, and conventional Transwell membranes. On day 7, cells on crypt-villus hydrogels and Transwell membranes displayed a polarized phenotype, with Villin detected on the epical surface and absent in the basolateral side. Similarly, F-actin was enriched at the apical membrane, while also outlining intercellular junctions along the apical-basolateral border. In contrast, cells grown on flat dSIS-NB hydrogels on day 7 showed less distinct polarization, as F-actin was distributed more evenly between apical and basolateral surfaces (see FIG. 4a and FIG. 4c). This could be attributed to the fact that cells on crypts / villi surface and Transwell membrane reached confluence in 3 to 4 days (see FIGs. 3a-3f), but it took 13 days on flat dSIS-NB hydrogel (FIGs. la- Id). By day 21, all three conditions exhibited well-polarized epithelial monolayers, with Villin and F-actin signals concentrated at the apical surface (see FIG.4b and FIG. 4c). These results indicate that the crypt-villus topography accelerates polarization and brush border maturation compared to flat hydrogel surfaces, supporting its role in promoting more physiologically relevant epithelial organization.

[0110] The extent of tight junction formation was further evaluated at both mRNA (T.IP-J) and protein (ZO-1) levels. Quantitative RT-PCR revealed that TJP-1 expression was consistently higher in cells cultured on Transwell than on dSIS-NB hydrogel (see FIG. 4d). Immuno staining of ZO-1 showed more intense and continuous tight junctions along each cell border in Transwells, whereas cells on flat and crypt- villus dSIS-NB hydrogels exhibited weaker and less continuous staining at both time points (see FIG. 4e). These results indicate that Transwell cultures with excessively high expression of tight junction proteins may hinder molecular transport across the intestinal epithelium. Although TJP-1 transcript levels were comparable between flat and crypt- villus dSIS-NB hydrogels, ZO-1 protein distribution revealed important structural differences. The lower and more spatially heterogeneous ZO-1 signal on crypt- villus hydrogel may reflect localized regulation of junctional complexes in response to curvature and microtopography. Together, these results highlight that scaffold geometry not only influences the degree of tight junction formation but also their spatial organization.

[0111] Taken all together, these results highlight the distinct roles of scaffold topography and composition in guiding epithelial polarization. The crypt-villus hydrogel accelerated polarization and brush border formation, as shown by earlier apical localization of Villin and F-actin compared to flat hydrogels. However, when examining tight junction formation, both hydrogel-PRF-71031-03based models showed lower TJP-l / ZO-1 expression compared to Transwells, which are known to generate overly tight barriers not representative of the human intestine. This suggests that the dSIS-NB hydrogel provides a more physiologically relevant environment for epithelial organization and barrier function, balancing polarization and tight junction formation in a way that better mimics the in vivo intestinal epithelium. The combination of early formation of polarized epithelium from the crypt- villus structure and more moderate tight junction expression resulting in a model that aligns more closely with physiological conditions.

[0112] To assess the long-term epithelium viability and stability in our 3D dSIS-NB crypt / villus model, the structure was stained with Ep-CAM, E-cadherin, and F-actin after 21 days of in vitro culture (see FIGs. 4f-4i which provide images that show immunofluorescence staining for 3D crypt / villus structure after 21 days of culture. FIG. 4f and FIG. 4g provide confocal z-stack images of E-cad, EpCAM, F-actin, and nuclei staining of the Caco-2 / HT29-MTX cell monolayer. The staining revealed prominent expressions of E-cadherin and EpCAM (FIG. 4f), both of which are indicative of strong cell-cell adhesion, which was maintained throughout the culture period, suggesting a well-formed epithelial layer. The F-actin (FIG. 4g) staining highlights organized cytoskeletal networks, further supporting the well-structured monolayer. Cross-sectional images of a single layer (crypts, villus base, and villi) and a 3D view obtained by merging all markers. The cross-sectional images (FIG. 4h) demonstrate the preservation of the crypt-villus architecture with well-defined crypts and villus structures, even after 21 days of culture. The 3D view (FIG. 4i) reinforces the integrity and stability of the crypt / villus morphology throughout the culture period. Scale bars represent 200 pm). The staining results revealed that all crypts and villi remained intact after 21 days, with no visible defects in the monolayer. Longitudinal monitoring of Caco-2 / HT29-MTX co-culture revealed that the rigid trans-well inserts yielded a rapid rise of TEER values, increasing significantly from 287 Q-cm2on day 4 to 1,284 Q-cm2on day 22 (see FIG. 5b, described below). Meanwhile, cells cultured on flat dSIS-NB hydrogels showed minimal TEER increases (under 50 Q-cm2) for the first 10 days, followed by a sharp increase afterward (76 Q-cm2on day 10 to 304 Q-cm2on day 13) when the cells reached full surface confluence (FIG. 1g). TEER values were higher for 3D crypt / villus gel than on the flat counterpart during the first week, with values rising from 6 Q-cm2on day 1 to 50 Q-cm2by day 4, reflecting the rapid epithelialization over the crypt / villus architecture (FIGs. 3a-3f). Note that although the cells reached a complete monolayer after 3 days of culture on thePRF-71031-03crypt / villus model, they might not be fully differentiated and polarized. The TEER value continued to increase over time and reached a more stable value on day 7, which is consistent with the polarized phenotype of cell culture on crypt / villus structure (FIGs. 4a-4e). After one week, the 3D crypt / villus model presented a limited increase, exhibiting values within a range of 90 to 150 Q- cm2from day 7 to day 22. These values were comparable to those obtained from in vivo intestinal tissues (40 - 100 Q-cm2). As the total surface of the crypt / villus hydrogel was bigger than flat surface, we accounted for the total villi and crypt numbers and the surface area in each sample. In total, 147 villi and 145 crypts were formed in a single scaffold. The effective surface area of crypts / villi hydrogel is 0.74 cm2, only 23% higher than the flat hydrogel or Transwell membrane (0.60 cm2). When normalized to the increased surface area, the TEER value of 3D crypt / villus model corresponded to values from 112 to 185 Q-cm2over day 7 to day 22. These normalized values were still much closer to the physiological range in native intestinal tissue than in the flat hydrogel and Transwell models.

[0113] Reference is made to FIGs. 5a-5e which provide schematics, images and graphs that show functional characteristics of 2D flat and 3D crypt / villus models. In particular FIG. 5a shows a schematic of crypt / villus barrier integrity and permeability assays. FIG. 5b provides TEER measurements of the intestinal epithelial banner on conventional trans-well, 2D flat, and 3D crypt / villus models. Data are presented as mean ± SEM (n = 4). The dashed line indicates a physiological TEER value (in vivo) obtained from Srinivasan et al. FIG. 5c, FIG. 5d, and FIG. 5e show permeability studies using dextran permeation from the apical to basolateral side for 2D flat and 3D crypt / villus models. In particular, FIG. 5c provides confocal images of 4.4k-TD and 500k-FD permeation in dSIS-NB hydrogels with flat. 3D crypt / villus structure, and cell free-hydrogel. FIG. 5d provides a graph that shows quantification of fluorescence intensity for 4.4k-TD permeation from the apical to basolateral side of 2D flat and 3D crypt / villus dSIS-NB models. Normalized RFU values were determined by the RFU(tn) / RFU(to) ratio (n = 3). FIG. 5e provides a graph that shows calculated apparent permeability coefficients (Papp) from 4.4k-TD permeation assays. Data are presented as mean ± SEM (n = 3, **p < 0.01 by a two-tailed Student’s Ftest). The 3D crypt / villus model exhibited a significantly higher Pappvalue compared to the 2D flat model, indicating enhanced permeability function in the 3D model.

[0114] Permeability studies were conducted using fluorescently labeled dextran: 4.4 kDa TRITC-dextran (4.4k-TD) and 500 kDa FITC-dextran (500k-FD). Thick dSIS-NB hydrogelsPRF-71031-03(thickness about 1.15 mm) with flat and crypt / villus topographies were used for visualizing dextran distribution across the intestinal epithelial layer by confocal microscopy (see FIG. 5c). On flat hydrogels with cell monolayer coverage, most of the dextran, regardless of molecular size (4.4k-TD or 500k-FD), was retained at the apical surface of the intestinal cell layer, with very weak fluorescence signal observed in the hydrogel layer (see FIG. 5c, flat surface + cells). However, cell-laden dSIS-NB crypt / villus hydrogel provided selective barrier function for 500k-FD but allowed the gradual passage of 4.4k-TD, as demonstrated by higher red fluorescence signals in the hydrogel layer (see FIG. 5c, Crypt / villus + cells). As expected, cell-free 3D crypt / villus dSIS-NB hydrogel did not provide barrier function, as both 4.4k-TD and 500k-FD could be observed in the hydrogel layer (see FIG. 5c, Crypt / villus cell-free). These results indicate that intestinal epithelium formed on both flat and crypt / villus hydrogels restricted the passage of larger molecules (i.e., 500k-FD), consistent with the restrictive nature of tight junctions for high molecular weight molecules. Interestingly, small molecules (i.e., 4.4k-TD) were gradually transported across the crypt / villus epithelium, a phenomenon akin to the selective permeability observed in native intestinal tissue.

[0115] Next, transport of 4.4k-TD from the apical to the basolateral side was quantified in both models. Consistent with the confocal images, 4.4k-TD showed a progressive increase in transport across the 3D crypt / villus model over time, while transport in the flat model remained minimal (see FIG. 5d). Furthermore, the apparent permeability coefficient Papp) for 3D crypt / villus model was nearly five-fold higher than that in the flat counterpart (2.3 x 10‘6cm s’1vs 0.5 x 10’6cm s’1, respectively, see FIG. 5e). After normalization for the increased effective surface area, the Pappof the crypt / villus model remained higher than that of flat hydrogel (1.9 x 10’6cm s’1vs 0.5 x 10’6cm s’1), further highlighting the impact of tissue architecture on absorption function.

[0116] In addition to increasing surface area, the crypt / villus topography also alters epithelial behavior in other ways. The expression patterns of ZO-1 in cells over crypts / villi hydrogel were discontinuous and uneven, suggesting weakened intracellular junctions. In contrast, ZO-1 patterns in flat dSIS-NB gels were more organized but still uneven. Finally, ZO-1 expressions in cells on transwell membrane were highly organized, suggesting very tight cell-cell junctions. These results explain the higher Pappand lower TEER values of the crypt / villus model compared to conventional models. The combination of an increased surface area and looser barrierPRF-71031-03facilitates nutrient and drug absorption while maintaining selective permeability. Noted that, although our 3D crypt / villus model achieved full epithelial confluence within 3 days (see FIGs.3a-3f 3), we performed functional permeability assays on day 21 to ensure a fair comparison with the flat hydrogel control, which required up to 13 days to reach full confluence. Moreover, testing on day 21 also demonstrated the long-term stability and sustained barrier integrity of the epithelial monolayer maintained on the dSIS-NB crypt / villus scaffold.

[0117] Reference is made to FIGs. 6a-6f which provide images and graphs that show drug-induced loss of crypt / villus barrier integrity. In particular, FIG. 6a provides brightfield images of the cell-laden crypt / villus layers under toxic stress conditions with different concentrations of staurosporine (after 24 hours of exposure). FIG. 6b provides epithelial barrier disruption with different staurosporine concentration evaluated by TEER values over time for 24 hr. Normalized TEER values were determined by the TEER(tn) / TEER(to) ratio (n = 3). FIG. 6c provides confocal images present dextran leakage of 3D crypt / villus model after 24 hours of staurosporine exposure using 4.4k-TD and 500k-FD. FIG. 6d provides timeline of the toxicity and transport assay. Staurosporine was added at the start of the experiment, while dextran was added at the time indicated post-staurosporine addition (i.e., 0, 9, 24 hr). FIG. 6e and FIG. 6f provide effect of Staurosporine (5pM) on permeability of (FIG. 6e) 4.4k-TD and (FIG. 6f) 500k-FD in dSIS-NB hydrogel with crypt / villus structure. DMSO was used as a control to establish the baselines of molecular transport in the absence of drug treatment.

[0118] Having demonstrated the successful generation of a. functional 3D crypt / villus intestine model, we explored its potential to model barrier dysfunction, such as inflammatory bowel disease, indigestion, irritable bowel syndrome, and Celiac disease. The intestinal epithelium formed on the dSIS-NB crypt / villus hydrogel was challenged with staurosporine, a broadspectrum kinase inhibitor widely used to induce cellular apoptosis. A clear dose-dependent disruption of the intestinal epithelium was observed after 24 hours of staurosporine treatment (see FIG. 6a). Staurosporine-induced loss of intestinal epithelium integrity was verified via rapid and dose-dependent decreases of TEER values (see FIG. 6b). High concentrations (5 pM and 50 pM) of staurosporine treatment also clearly increased the transport of 500k-FD (see FIG. 6c), a phenomenon not observed under drug-free conditions (see FIG. 5c), indicating a severe breakdown of tight junction integrity. Molecular transport studies were conducted using 4.4k-TD and 500k-FD added to the crypt / villus model at different time points post-staurosporine exposurePRF-71031-03(i.e., 0 hr, 9 hr, 24 hr. see FTG. 6d). When 4.4k-TD and 500k-FD were added at the same time as staurosporine treatment (i.e., Dextran @ Ohr), no drastic changes were observed in molecular transport compared with the DMSO control, suggesting that the rapid reduction of TEER values (see FIG. 6b) preceded the changes in molecular transport. When 4.4k-TD and 500k-FD were added 9 hr or 24 hr post-staurosporine treatment (i.e., Dextran @ 9 hr and 24 hr), significant increases in both 4.4k-TD and 500k-FD transport across the crypt / villus intestinal layer were detected (see FIG. 6e and FIG. 6f).

[0119] Reference is made to FIG. 6g and FIG. 6h which provide graphs showing effect of Staurosporine (5 pM) on the permeability of (FIG. 6g) 4.4 kDa TRITC-dextran and (FIG. 6h) 500 kDa FITC-dextran. Staurosporine (5 pM), a known inducer of apoptosis, was added at the start of the experiment to assess its effect on the barrier integrity of the Caco-2 / HT29-MTX monolayer cultured on dSIS-NB hydrogels with crypt / villus structures. Permeability was measured by adding (FIG. 6g) 4.4k-TD and (FIG. 6h) 500k-FD at indicated time points (i.e. 0 hr, 9 hr, and 12 hr) after staurosporine treatment. The results show a significant increase in the permeability of both dextrans, indicating a loss of barrier function. The delayed and limited passage of 500 kDa dextran suggests that tight junctions, although compromised by staurosporine, still partially restricted the movement of larger molecules. Data are presented as mean + SEM; (n = 3, *p < 0.05, * *p < 0.01, ***p < 0.001, and ****p < 0.0001 by repeated measures one-way ANOVA).

[0120] A clear increase in Pappwas noted for both dextran sizes over time after staurosporine exposure (see FIG. 6g and FIG. 6h), indicating that the drug disrupted the epithelial barrier, causing higher permeability. Our data indicated that the TEER measurement is more sensitive to early epithelial barrier damage than the fluorescent leakage assay. The immediate decline in TEER values upon exposure to staurosporine highlights the sensitivity of TEER as an early indicator of tight junction dysfunction. In the context of drug testing for gastrointestinal toxicity, TEER can serve as a rapid, non-invasive measurement to identify compounds that compromise barrier function. In contrast, the fluorescent leakage assay provided valuable insights into molecular permeability, though its sensitivity lagged behind TEER response. No significant fluorescent leakage was observed in the first two hours, suggesting that while tight junctions were compromised, the barrier was still able to restrict molecular diffusion. However, as barrier damage progressed, increased leakage of small molecular weight probes was observed after threePRF-71031-03hours, with clear permeability across the epithelial layer by the 9-hour and 24-hour time points. The delayed response of the fluorescent leakage assay suggests that molecular permeability assays may miss early signs of barrier disruption, particularly when the damage primarily affects tight junctions without immediately compromising the entire epithelial layer. However, at later stages, when the barrier becomes more permeable, fluorescent leakage assays provide a complementary approach to TEER by identifying changes in molecular permeability that TEER cannot detect. The sensitivity of TEER to early-stage barrier disruption and the molecular insights provided by fluorescent leakage assays make this approach highly effective for evaluating drug-induced epithelial damage. This finding provides valuable insights into the utility of different assays for assessing drug effects on epithelial integrity, where TEER can serve as a rapid, non-invasive measurement to identify compounds that compromise barrier function, but the results must be confirmed with actual molecular transport studies.

[0121] Reference is now made to FIGs. 7a-7f which provide a schematic, images, and a graph showing a pizza-like dSIS-NB hydrogel to model intestinal disease. In particular, FIG. 7a provides a schematic of molecular transport in celiac disease and healthy crypt / villus epithelium. FIG. 7b provides an image of design of a negative pizza mold by Tinker CAD and molded dSIS-NB ‘pizza’ hydrogel. FIG. 7c provides brightfield images of the intestinal cells culture on the ‘pizza’ hydrogels. The first two images indicate similar degree of cell spreading on both flat and crypt / villus structures after seeding (day 0). The last image shows the difference in cell coverage on day 3 between flat and crypt / villus topography. White dashed lines mark the cell boundary while the thin yellow dashed lines mark the flat and crypt / villus topography boundaries. FIG. 7d, FIG. 7e, and FIG. 7f provides an image and graphs that show permeability assay using the dSIS-NB ‘pizza’ hydrogel with flat, transition, and crypt / villus topography. Specifically, FIG. 7d provides an image which shows a representative 3D overall view for dextran transport across the flat and crypt / villus epithelium on the pizza model. FIG. 7e provide images that show representative cross-section view (x-z) of the pizza hydrogel model with dextran transport through the intestinal epithelial cell layer into the hydrogel. FIG. 7f provides a graph that shows quantification of 4.4k-TD and 500k-FD transport along the distance from flat to transition and to crypt / villus zones. The y axes represent the signal intensity and the x axes represent the distance corresponding with FIG. 7e. To ensure an accurate comparison of fluorescence intensity, the entire hydrogel channel was selected.PRF-71031-03

[0122] Celiac disease (CD) is a chronic autoimmune disorder of the small intestine triggered by the ingestion of gluten in genetically predisposed individuals. Intestinal villous atrophy, crypt hyperplasia, and a flat intestinal surface was notable in CD, causing severely hindered nutrient absorption (see FIG. 7a). In vitro CD models are limited in their capacity to mimic the complexity of the disease, particularly the structural changes from healthy villi to flattened mucosa. This work presents a proof-of-concept ‘pizza’ -like dSIS-NB hydrogels with features of both villus-rich architectures of the healthy small intestine and the progressively flattened surface of celiac disease (see FIG. 7b), ensuring consistent experimental conditions (e.g., cell density) and allowing side-by-side comparisons of molecular transport. The ‘pizza’ hydrogel was fabricated using DLP printing of sacrificial PEGNB-T hydrogel with both flat and negative crypt / villus topographies (see FIG. 7b). After the crosslinking of dSIS-NB hydrogels and the dissolution of PEGNB-T sacrificial hydrogels, Caco-2 / HT29-MTX cells were seeded on the ‘pizza’ dSIS-NB hydrogels. Similar to that observed in separate hydrogel cultures (see FIGs. la-li and see FIGs. 3a-3f), on the ‘pizza’ hydrogel, regions with intact crypt / villus structures were fully covered by Caco-2 / HT29-MTX cells as early as day 3, whereas the flat regions exhibited delayed cell spreading (see FIG. 7c). A molecular transport study using 4.4k-TD and 500k-FD across the ‘pizza’ hydrogel further revealed the clearer differences in permeability patterns across gel with different topographies (see FIGs. 7d-7f). The ‘pizza’ hydrogel surface was separated into flat, transition, and crypt / villus zones, representing celiac disease, transition, and healthy intestinal epithelium, respectively. Similar to the transport study results shown in FIGs.5a-5e, the flat zone restricted the transport of both 4.4k-TD and 500k-FD, whereas the crypt / villus zone permitted the passage of 4.4k-TD and a limited amount of 500k-FD. This observation was consistent with the notion that the healthy crypt / villus architecture promotes efficient mass transfer, while a flattened epithelium hampers normal nutrient absorption, a signature characteristic of celiac disease in vivo.

[0123] As for the benefits of the ‘pizza-like’ hydrogel, we believe that (1) it will allow one to observe cell proliferation in different zones without having to fabricate different scaffolds and ensures consistency in the same experimental environment, (2) it can eliminate sample-to-sample variability while allowing for simultaneous, side-by-side comparisons of molecular transport, and (3) it can capture the ‘patchy’ appearance of Celiac disease to reflect the spatial variation of villus atrophy. The dual-region scaffold that mimics healthy (with crypt- villi) and diseased (flat)PRE-71031-03intestinal tissue in a single construct offers a simple yet effective within-sample comparison. Our engineered dSIS-NB crypt / villus hydrogels could also be leveraged to study immune-mediated damage to the crypt / villus architecture. To our knowledge, engineered matrices that can recapitulate the spatiotemporal variation of intestinal diseases have not been reported in the literature.

[0124] Generating a functional small intestine model is challenging owing to the complex topography of the crypt / villus structure. This work presents a novel and robust biofabrication platform to recapitulate the 3D crypt / villus architecture of the small intestine. Specifically, we leveraged the DLP-printed dissolvable PEGNB-T SHM to mold cell-responsive dSIS-NB hydrogel with the delicate crypt / villus structures. From early cell adhesion to complete monolayer formation, dSIS-NB hydrogels with crypts / villi topography outperform Matrigel. The only special instrument required for generating the crypts / villi topography is a DLP printer, which has become widely accessible in recent years. Both norbornene-functionalized macromers, PEGNB-T and dSIS-NB, could be synthesized in a standard fume hood, making this platform adaptable for molding matrices for in vitro modeling of epithelial diseases. Looking ahead, this platform opens new avenues for investigating the impact of various diseases on intestinal barrier physiology and function, such as inflammatory bowel disease (IBD), where the crypt / villus structure is often disrupted. Similarly, the model could be used to investigate pathogen invasion and host defense mechanisms in gastrointestinal infections by introducing relevant microbial or viral agents to the apical surface. Additionally, the model may be adapted to study early events in colorectal cancer progression by incorporating genetically modified or patient-derived cells into the crypt regions. Moreover, as dSIS-NB hydrogels are cytocompatible for in situ cell encapsulation, other stromal cells, including endothelial cells, fibroblasts, and immune cells, can be incorporated within the dSIS-NB hydrogels for modeling complex cellular dynamics. Another future direction is the adaptation of the current Transwell platform to enable the application of flow or peristaltic forces, thereby generating a microphysiological system that more closely mimics the in vivo intestinal microenvironment. In sum, the simplified and highly adaptable fabrication workflow holds significant promise for advancing gastrointestinal tissue engineering and drug discovery.

[0125] dSIS-NB was synthesized according to our established protocols. Briefly, fresh bovine small intestine was collected from local grocery stores and stored at -20 °C for no longer thanPRF-71031-03one month. One kilogram of intestine was washed carefully with tap water, then cut into 10 cm in length, and the mesenteric tissues were manually removed. The intestine segments were inverted and scrubbed to remove the mucosal epithelium and lamina propria. The segments were flipped back and the tunica serosa and tunica muscularis externa were removed. Prior to decellularization, the cleaned SIS tissues were rinsed 2 times with PBS, followed by stirring in 250 mL PBS containing 1% SDS and 2% penicillin-streptomycin for 4 days, with daily buffer change. Next, the tissues were treated with 1% triton-XlOO containing 0.05 mg mL-1gentamycin for 24 h before vigorous washing in autoclaved deionized water for 7 days (change every 8 to 12 h). The dSIS obtained was lyophilized for 72 h under -50 °C, 20 Pa, and stored at -80 °C for use within 12 months.

[0126] One hundred and fifty mg of desiccated dSIS was cut into small pieces using sterilized scissors and then digested in 75 mL of acidic solution (pH 2, adjusted by adding 0.01 N HC1) containing 150 units of pepsin per mg of dSIS at room temperature (RT, 20°C -22°C) for 7 days until fully dissolved. Norbornene conjugation was achieved by adding 135 mg of carbic anhydride (CA) to the acidic dSIS solution supplemented with about 255 pL of triethylamine (TEA) with magnetic stirring (650 rpm) at RT for 6 hr. The dSIS-NB solution was passed through a 40-pm cell strainer to remove debris, followed by dialysis (12 kDa molecular weight cut-off) against precooling autoclaved DI water at 4 °C for 3 days, with frequent changes of fresh autoclaved DI water every 12 hr. The dialyzed solution was freeze-dried for 3 days and the dried dSIS-NB was dissolved in sterile PBS using vigorous vortexing at 4 °C, stored at 4 °C, and used within 1 week. To ensure reproducibility, the gelation and mechanical properties of each batch were confirmed to fall within an acceptable range before use for further experiments.

[0127] Proteomic analysis was performed by Indiana Clinical and Translational Sciences Institute (CTSI) Center for Proteome Analysis. Briefly, 5 mg of lyophilized dSIS and dSIS-NB samples were resuspended in 40% acetonitrile at 2-4 pg / pL. 8 M urea (30 pL) was added to 10 pL of dSIS / dSIS-NB sample (about 50 pg). Disulfide bonds were reduced with 15 mM TCEP and alkylated with 50 mM chloroacetic acid (CAA). The samples were digested overnight with 1 pg trypsin / LysC - ( 1 hr at about 6 M urea, then diluted to 2 M urea with Tris and added 2 pL PNGase F, which was allowed to go overnight at 35 °C). The next morning, samples were quenched with formic acid and cleaned up with Waters Sep-Pak cartridge. Next, samples werePRF-71031-03resuspended in 40 pL of 0.1 % FA and analyzed by the Eclipse- Aurora column with FAIMS (CID). Analyzed in PEAKS 12 with custom modification of dSIS-NB.

[0128] PEGNB-T was synthesized by reacting tyramine with the carboxylic acid on PEGNB-carboxylate (PEGNBCA), which was synthesized as previously reported. Briefly, one gram (g) of 8-arm PEG-OH (0.4 mmol hydroxyl group), 0.328 g carbic anhydride (2 mmol, 5 equiv. to -OH), 49 mg 4-Dimethylaminopyridine (DMAP, Sigma-Aldrich) (0.4 mmol, 1 equiv. to -OH), and 6.64 mL of anhydrous tetrahydrofuran (THF, Sigma- Aldrich) were charged in a round bottom flask. The flask was placed in a 60°C oil bath with constant stirring for 8 hours. An additional 5 equivalents of CA and 1 equivalent DMAP were added to the flask and the reaction was continued for 24 hours. The product was precipitated using 20x cold diethyl ether twice, redissolving in minimal dichloromethane (DCM, Fisher Scientific) between precipitations, and dried in a vacuum overnight before dialyzing for 72 hours against ddHaO. After dialysis, the product, PEGNBCA, was lyophilized for 72 hours under 20 Pa at -50 °C.

[0129] Next, tyramine was conjugated to 1 g PEGNBCA (carboxylic group: 380 pmol) using the standard carbodiimide chemistry. Briefly, 148.8 pL of N,N’ -Diisopropylcarbodiimide (DIC, Chem-Impex) (950 pmol, 2.5 equiv.), 160 mg Hydroxybenzotriazole (HOBt, Oakwood Chemical) (950 pmol, 2.5 equiv.) were mixed in 10 mL amine-free anhydrous dimethylformamide (DMF, Fisher Scientific) and blanketed with nitrogen and stirred for 2 hours in the dark. 165 mg tyramine hydrochloride (Chem-Impex) (950 pmol, 2.5 equiv.), 165.4 pL N,N’ -diisopropylethylamine (DIPEA, TCI Chemicals) (950 pmol, 2.5 equiv.), and 23 mg DMAP (190 pmol, 0.5 equiv.) were dissolved in 1 mL amine-free DMF and added to the reaction. The mixture was allowed to react for 16 hours, followed by dialysis in methanol for one day and in ddH O for two more days, then lyophilized. The norbornene and tyramine substitution were characterized by 1H NMR spectra in D2O (Bruker 400 MHz).

[0130] dSIS-NB hydrogels were crosslinked by thiol-norbornene photo-click reaction. To prepare the hydrogel samples, 40 pL of the precursor solution (1.2 wt% dSIS-NB, various contents of PEG4SH, and 7 mM of photoinitiator LAP) was dispensed in between two glass slides separated with 0.8 mm Teflon® spacers and pre-treated with a water-repellent coating. The assembly was placed under 365 nm UV light at 5 mW cm-2for 2 minutes to yield dSIS-NB hydrogels with an approximate diameter of 8 mm and a thickness of 0.8 mm. Bulk hydrogel moduli were characterized by a modular rheometer (MCR 102, Anton Paar) operating in a strain-PRF-71031-03sweep mode with normal force (NF) control. The rheometrical testing was conducted at 25 °C, with a normal force of 0.25 N, shear strain from 1% to 5%, and a frequency of 1 Hz. The shear modulus (G’) in the linear viscoelastic region (LVE) was reported as the stiffness of the hydrogels.

[0131] Caco-2 (ECACC 86010202) and HT29-MTX (ECACC 12040401) cell lines were maintained in 100 mm tissue culture dish (229621, Celltreat) in high-glucose DMEM (SH30243.01, Cytiva) containing 10% fetal bovine serum (Gibco), 1 % v / v non-essential amino acid (Fisher Scientific), and 1 % v / v penicillin / streptomycin (Fisher Scientific) at 37 °C, 5% CO2 and 95% humidity. Cells were passaged at 80% confluence using 0.05% Trypsin EDTA (Fisher Scientific) and the culture media were refreshed every 2-3 days. Both Caco-2 and HT29-MTX cells at passages 5-12 (passage 1: the first subculture after thawing the acquired frozen stocks) were used for the experiments. Cells at passages 1-5 were maintained in cryogenic storage in liquid nitrogen.

[0132] Ninety microliters of dSIS-NB precursor (1.2 wt% dSIS-NB, 1.2 wt% PEG4SH, 7 mM LAP) or Matrigel (354230, Corning) were prepared in 24-well glass bottoms plates (229125, Celltreat) for time-lapse imaging, or in 24-well plates with culture inserts (PI8P01250, Millipore) for all other experiments. The hydrogels were polymerized under 365 nm light (5 mW cm-2) for 2 min when using dSIS-NB or incubated for 30 min at 37 °C when using Matrigel. Caco-2 and HT29-MTX cells were trypsinized, pelleted, and aliquoted. The cells were applied by seeding 1 x 105cells cm-2onto the surface of solidified hydrogels either mono or in co-culture with ratio of 9:1 (Caco-2:HT29-MTX) as recommended from previous study. Subsequently, the plate was incubated for at least 1 h to allow the cells to attach and maintained at 37 °C. 5% CO2. and 95% humidity. Medium was refreshed every 2-3 days and cultured over 21 days.

[0133] Time-lapse imaging was conducted using an automatic CELLCYTE X microscope (CELLINK AB) in a humidified, heated, C02-controlled chamber (VWR). Live cell imaging started 3 hours after cell seeding, using a 4X objective. Pictures of the same positions were captured every three hours over a recording duration from cell seeding to the confluence of cell monolayer. Four positions on each well were monitored and the experiment was repeated on three wells for each hydrogel condition. Images were analyzed to determine cell confluence and videos were compiled using the tools integrated into CellCyte Studio.PRF-71031-03

[0134] The overall process of creating an engineered human intestinal model is graphically described in FIGs. 2a-2g. We used a digital processing light (DLP) 3D printer (LumenX, CELLINK) to create PEGNB-T SHM with negative crypt / villus structures. The PEGNB-T SHMs were crosslinked by thiol-norbornene photo-click reaction at 405 nm. Bioink formulation and printing parameters (light intensity and exposure time) were obtained from our previous studies. PEGNB-T SHM were printed with 4 wt% PEGNB-T, 3 wt% PEG4SH, 1.5 mM tartrazine, and 10 mM LAP. 405 nm light (34 mW cm-2) was turned on for 9.5 seconds for each layer of printing. Printing fidelity was defined as the difference in the upper diameter of the wells and the tip diameter of the pillars between designed and printed structures:_. , ... / rr, . , \Desianed diameter -printed diameter] „ , Fidelity (%) = 1 -J- - - - -1x 100% > Equation 1Designed diameter

[0135] SHM was used to cast 40 pL of dSIS-NB precursor (1.2 wt% dSIS-NB, 1.2 wt% PEG4SH, 7 mM LAP), followed by exposure to 365 nm light at 5 mW cm-2for 1 min. The cast hydrogel assembly (i.e., SHM + dSIS-NB hydrogels) was placed upside down in a trans- well insert (PI8P01250, Millipore). The insert was precoated with 50 pL of the dSIS-NB precursor followed by another 365-nm light exposure for 1 min to secure the hydrogel assembly in the trans-well insert. After photocrosslinking, IX PBS was added to the trans-well insert, allowing hydrolytic degradation of the SHM at 37°C. After the autonomous removal of PEGNB-T SHM, the crypt / villus structure was rinsed twice with PBS, followed by Caco-2 and HT29-MTX cells seeding 1 x 105cells cm-2on the top of the crypt / villus structure. Cells were allowed to settle at 37 °C, 5% CO2 for 1 h before additional medium was added and incubated at 37 °C. 5% CO2.

[0136] After 7, 12, and 21 days of culture on either dSIS-NB hydrogels or Matrigel, Caco-2 cells were collected for gene expression by reverse transcription-quantitative polymerase chain reaction (RT-qPCR). Total RNA was isolated from the collected cells using NucleoSpin® RNA kit (MACHEREY-NAGEL), followed by cDNA generation with PrimeScript RT reagent kit (Takara, Cat# RR037A). RT-qPCR was performed with TB Green Premix Ex TaqII kit (Takara, Cat# RR820L) for genes commonly selected for intestinal epithelial assessment (e.g., Villin VIL1), Mucus (MUC2), intestinal alkaline phosphatase (ALPI), proliferation marker (CCNDl), multidrug resistance transporter (MDR1), and transporter gene (SLC15A)). The primer sequences were listed in Table 1, provided below. To adjust for variations in the cDNA synthesis, gene expression data were normalized to housekeeping genes (GAPDH) and analyzed statistically toPRF-71031-03compare differences between time points and between hydrogel conditions. Fold changes were calculated by the 2-AACtmethod. Three independent replicates in each condition were performed. Table 1 - The primer sequences for qPCR used in present disclosure.Gene Forward primer sequence (5’-3’) Reverse primer sequence (5’-3’) GAPDH GCCTCCTGAAAAGAGAGTGGAAG GCCTCCTGAAAAGAGAGTGGAAG ALPI CCAGGACATCGCCACTCAGC CTCAGTGCGGTTCCACACATAC VIL1 TGCTATCTATGGTGTGGGAAGG TCCTGTAGTCTCTTGGTGTTGG MDR] GCCAAAGCCAAAATATCAGC TTCCAATGTGTTCGGCAT CCND1 CAATGACCCCGCACGATTTC CATGGAGGGCGGATTGGAA SLC15A TGTCCACCGCCATCTACCATA CCACGAGTCGGCGATAAGAG TJP-1 ACCAGTAAGTCGTCCTGATCC TCGGCCAAATCTTCTCACTCC

[0137] Cells cultured on 3D crypt / villus hydrogel models or 2D flat hydrogels were fixed with 4% w / v paraformaldehyde (Fisher Scientific) in PBS (Life Tech No. 20012068) for an hour at room temperature or 24 h at 4 °C, washed twice for 5 min with PBS, permeabilized with 0.3% Triton X-100 (Sigma # T8787) in PBS for 10 min, and blocked with 3% bovine serum albumin (BSA) (Sigma # A2153) in PBS with 0.05% v / v Tween-20 (Sigma Aldrich) for 1 h.Subsequently, cells were incubated with primary antibodies: Vinculin, Paxillin, Ki67, mucus-2 (MUC2), zonulaoccludens-1 (ZO-1), CD326 (EpCAM), and E-cadherin (E-Cad), at 4 °C overnight, followed by washing with PBS (3 times x 15 min). After incubating the corresponding secondary antibodies together with Alexa Fluor 647-conjugated phalloidin (1:100) (Cytoskeleton / F-actin) for 2 h at room temperature, cell nuclei were counterstained with DAPI (1:200) for 30 min. The antibodies used are detailed in Table 2, provided below. During staining, the entire device was covered with aluminum foil to reduce the fluorescence decrease due to photobleaching. Fluorescence images were taken with a bench-top confocal microscope (BC43, Oxford Instrument). All immuno staining s were repeated at least three times.Table 2 - List of antibodies used in the present disclosurePRF-71031-03Antibody Dilution Source / Isotype Supplier (Catalog #) Primary antibodiesE-Cadherin 1:200 Rabbit Cell Signaling (3195) EpCam (CD326) 1:100 Mouse Thermo Scientific (14-9326-82) Vinculin 1:50 Mouse Santa Cruz Biotech, (sc-25336) Paxillin 1:50 Rabbit Abeam (ab32084)ZO-1 1:100 Rabbit Thermo Scientific (61-7300) Mucin 2 / MUC2 1:100 Mouse Santa Cruz Biotech, (sc-515032) Ki67 1:200 Rabbit Cell Signaling (9129)Villin 1:200 Mouse Thermo Scientific (MA5-38658) Secondary antibodiesAnti-Rabbit Alexa Fluor 488 1:200 Donkey Invitrogen (A21206)Anti-Mouse Alexa Fluor 488 1:200 Goat BioLegend (405319)Anti-Rabbit Alexa Fluor 555 1:200 Goat Cell Signaling (4413S) Anti-Mouse Alexa Fluor 555 1:200 Goat Cell Signaling (4409S)

[0138] To assess the growth and integrity of epithelial barriers on flat or 3D crypt / villus models, TEER was measured every three days using an EVOM2 Epithelial Voltohmmeter with an STX03 electrode (World Precision Instruments). TEER measurements were taken over 22 days, including insert only (no hydrogel, no cells), insert + flat hydrogel (no cells), and insert + crypt / villus hydrogel (no cells) to set the initial blank values for the resistance of transwell membrane insert and hydrogel, which were subtracted from the TEER values of the experimental groups. Note that the hydrogel layer alone contributed minimally to the TEER values, while the resistance of membrane in culture medium was around 181 ± 80. Two independent replicates were conducted for the conditions without significant TEER increases, and four independent replicates were conducted for the conditions with significant TEER increases. Raw resistance data were converted into TEER values using the following equation:PRF-71031-03TEER ( -cm2) = (RM - Rc) x Membrane Area . Equation 2 where, RM represents the resistance value (Q) of the measured model with cells,Rc is the resistance value (Q) of the blank control group, andA is the membrane area of the Transwell insert (0.6 cm2, PI8P01250, Millipore).

[0139] The barrier integrity of intestinal epithelium on 2D flat and 3D crypt / villus topography was evaluated by assessing paracellular transport. One day before starting the experiment, cells were rinsed with PBS and incubated with fresh DMEM without phenol red. To assess the distribution of dextrans within the hydrogel, representative conditions were selected. A total of 400 pL fresh medium was added to the basolateral side, and 400 pL of the dextran mixture (4.4 kDa TRITC-dextran / 4.4k-TD (Sigma, T1037) and 500 kDa FITC-dextran / 500k-FD (Sigma, FD500S)) was applied to the apical side at 0.1 mg m "1in medium. After 60 minutes of incubation at 37°C, the solutions were removed, and a bench-top confocal microscopy was used to observe dextran distribution within the dSIS-NB hydrogel.

[0140] For permeability measurements, only 4.4k-TD was used. The medium was changed with 400 pL fresh medium in the basolateral side and 400 pL of 4.4k-TD (0.1 mg mL“l) on the apical side. Cell-free hydrogels with identical crypt / villus geometry were used as controls. Samples (100 pL) were collected from the basolateral side every 30 min for 3 hr into a black 96-well microplate (Greiner bio-one, 655076), and the same volume of fresh medium was replenished after each sampling into basolateral side. Fluorescence intensity was measured using a SpectraMax iD5 Microplate Reader (Molecular Devices, CA, USA) with excitation / emission at 530 / 590 nm. Fluorescence intensities were normalized to initial basolateral concentrations. The samples were correlated to a standard curve. The total amount (including the amount in 100 pl of took out samples) of compound determined at each time point were used to calculate the apparent permeability of the compound (from apical to basolateral side) using the following equation:dQPapp = / (A x C(}y . Equation 3 where, dQ / dt is the amount of compound (mg) passing through the intestinal barrier per unit time (s),A is the surface area (cm2), andPRF-71031-03Co is the initial concentration of the compound (mg mL'1). Experiments were performed with three replicates.

[0141] The disruption of barrier integrity under drug-induced conditions was evaluated by exposing models to increasing concentrations of staurosporine (i.e., 0, 0.5, 5, and 50 pM) in medium. Controls were treated with medium containing only 0.01% DMSO (Sigma, D8418). TEER measurements were taken prior to staurosporine exposure as a baseline and monitored hourly for the first 12 hours, and then again from 21 to 24 hours post-exposure. Between each interval, the models were placed back into the incubator. TEER values were normalized to the TEER measured before staurosporine exposure. The dextran distribution with 500k-TD and 4.4k-TD were conducted after 24 hours of exposure using the same method described for the absorption / barrier integrity assay.

[0142] The barrier disruption at 5 pM staurosporine was further analyzed by performing a leakage assay at 0, 9, and 24 hours of exposure. Four hundred microliters of 4.4k-TD or 500k-FD (0.1 mg mL-1) containing 5 pM staurosporine in medium was added separately to the apical side following the replacement of the basal medium with 400 pL of fresh medium. At each time point, 100 pL aliquots were sampled at the basolateral side and then replaced with 100 pl fresh medium. Fluorescence intensity with excitation / emission at 485 / 535 nm for FITC-dextran and 530 / 590 for TRITC-dextran. The normalized fluorescence intensity and Pappvalues were calculated as described above. These experiments were carried out across 3 replicates.

[0143] The process of creating the “Pizza” scaffold followed the same fabrication method used for molded dSIS-NB hydrogels with crypt / villus topography. The modification involved a CAD design that incorporated both flat surfaces and crypt / villus topography, as shown in FIG. 7b. Then, a mixture of Caco-2 and HT29-MTX cells were seeded onto pizza dSIS-NB scaffold to form a functional epithelium layer. The distribution of dextrans within the hydrogel was assessed using the previously described method. Confocal images were analyzed to quantify the fluorescent intensity of 4.4k-TD and 500k-FD transported along the distance from flat region through transition zone to crypt / villus zone using FlJI-ImageJ.

[0144] Statistical analyses were performed using GraphPad Prism 9.0 (GraphPad, La Jolla, USA). Relevant statistical methods were provided in the figure captions. All experiments were conducted with at least three independent repeats. Averaged values were expressed as mean ± SEM. Statistical significance was assumed for p < 0.05.PRF-71031-03

[0145] Those having ordinary skill in the art will recognize that numerous modifications can be made to the specific implementations described above. The implementations should not be limited to the particular limitations described. Other implementations may be possible.

Claims

PRh-71031-03Claims:

1. A method for making an intestinal tissue model, comprising:depositing a decellularized small intestine submucosa-norbornene (dSIS-NB) hydrogel within a sacrificial hydrogel intestine mold having negative crypts-villi structures formed thereon; anddissolving the sacrificial hydrogel intestine mold to thereby generate intestinal tissue model having positive crypts-villi structures formed thereon.

2. The method of claim 1, wherein the dSIS-NB hydrogel is formed by a photochemical reaction after mixing a) decellularized small intestine submucosa (dSIS) matrix having one or more norbornenes (dSIS-NB), with b) 4-arm polyethylene glycol) (PEG) thiol (PEG4SH) as a cross-linkers, and c) lithium phenyl-2,4,6-trimethylbenzoylphosphinate (LAP) as a photo-initiator.

3. The method of claim 1, wherein the sacrificial hydrogel intestine mold is a PEG- norbornene-tyramide (PEGNB-T) hydrogel network.

4. The method of claim 3, wherein the PEGNB-T hydrogel network is formed by a photochemical reaction after mixing a) PEGNB-T, with b) 4-arm PEG thiol (PEG4SH) as a cross-linker, and c) LAP as a photo-initiator.

5. The method of claim 3, wherein the sacrificial hydrogel intestine mold is dissolved by a water-based solvent.

6. The method of claim 5, wherein PEGNB-T is synthesized by:a) reacting 4-arm PEG-OH structure with carbic anhydride in presence of 4-Dimethylaminopyridine (DMAP) as a catalyst, to form PEGNB- carboxylate (PEGNBCA) structures; andb) reacting the PEGNBCA with tyramine in the presence of 1,3- Diisopropylcarbodiimide (DIC), 1 -Hydroxybenzotriazole (HOBt), N,N- Diisopropylethylamine (DIPEA), and DMAP as catalysts to form PEGNB-T.

7. The method of claim 2, wherein the photochemical reaction is based on a concentration of LAP between about 0.5 mM to about 34 mM.PRF-71031-038. The method of claim 2, wherein the photochemical reaction is based on light having a wavelength of between about 365 nm to about 600 nm.

9. The method of claim 2, wherein the photochemical reaction is based on having an intensity of between about 2 to about 20 mW / cm2.

10. The method of claim 1, wherein the sacrificial hydrogel intestine mold is made by a digital light processing (DLP) three dimensional (3D) printer.

11. A sacrificial hydrogel intestine mold, comprising:a mold made from a hydrogel, having formed thereon negative crypts- villi structures; wherein the hydrogel dissolves when placed in contact with a water-based solvent.

12. The sacrificial hydrogel intestine mold of claim 11, wherein the sacrificial hydrogel intestine mold is a PEG-norbornene-tyramide (PEGNB-T) hydrogel network.

13. The sacrificial hydrogel intestine mold of claim 12, wherein the sacrificial hydrogel intestine can be dissolved by a water-based solvent.

14. The sacrificial hydrogel intestine mold of claim 12, wherein the PEGNB-T hydrogel network is formed by a photochemical reaction after mixing a) PEGNB-T, with b) 4-arm PEG thiol (PEG4SH) as a cross-linker, c) lithium phenyl-2,4,6- trimethylbenzoylphosphinate (LAP) as a photo-initiator, and d) tartrazine as a photoabsorber.

15. The sacrificial hydrogel intestine mold of claim 14, wherein PEGNB-T is synthesized by:a) reacting 4-arm PEG-OH structures with carbic anhydride in presence of 4- Dimethylaminopyridine (DMAP) as a catalyst, to form PEGNB-carboxylate (PEGNBCA) structures; andb) reacting the PEGNBCA with tyramine in the presence of 1,3- Diisopropylcarbodiimide (DIC), 1 -Hydroxybenzotriazole (HOBt), N,N- Diisopropylethylamine (DIPEA), and DMAP as catalysts to form PEGNB-T.

16. The sacrificial hydrogel intestine mold of claim 14, wherein the photochemical reaction is based on a concentration of LAP between about 0.5 mM to about 34 mM.

17. The sacrificial hydrogel intestine mold of claim 14, wherein the photochemical reaction is based on light having a wavelength of between about 365 nm to about 600 nm.

18. The sacrificial hydrogel intestine mold of claim 14, wherein the photochemical reaction is based on having an intensity of between about 12 to about 20 mW / cm2.PRF-71031-0319. The sacrificial hydrogel intestine mold of claim 14, wherein the photochemical reaction is based on having a tartrazine concentration of between about 0.5 to about 2 mM.

20. The sacrificial hydrogel intestine mold hod of claim 11, wherein the sacrificial hydrogel intestine mold is made by a digital light processing (DLP) three dimensional (3D) printer.

Images