Invasin-functionalized PIC hydrogels enable long-term 3D culture of epithelial organoids

Significance

This study introduces a biomaterial by incorporating the integrin-binding domain of Yersinia Invasin protein into polyisocyanide (PIC) hydrogels, creating a multivalent three-dimensional (3D) Invasin environment for long-term culture of tissue stem cell (TSC)-derived epithelial organoids. Unlike Matrigel/Basement Membrane Extract (BME), this system is fully defined, animal-free, and cost-effective. Additionally, PIC hydrogels offer significant advantages due to their reversible thermoresponsive properties, making the gels easy to handle and adaptable for organoid culture. The integration of two-dimensional (2D) and 3D Invasin-based cultures provides a robust Matrigel/BME alternative, enhancing organoid accessibility and enabling automation, validation, and standardization of research in academia and industry.

Abstract

Tissue stem cell (TSC)-derived epithelial organoids are typically cultured in Matrigel [T. Sato et al., Nature 459, 262–265 (2009)], an extracellular matrix-like hydrogel produced from Engelbreth–Holm–Swarm sarcoma cells. This tumor is grown in the mouse abdomen [R. W. Orkin et al., J. Exp. Med. 145, 204–220 (1977)]. Previously, we demonstrated that the Yersinia membrane protein Invasin, coated on transwells, replaces Matrigel by activating β1-integrins, allowing long-term expansion of primary epithelial cells as 2D organoid sheets [J. J. A. P. M. Wijnakker et al., Proc. Natl. Acad. Sci. U.S.A. 122, e2420595121 (2025)]. Here, we functionalize a synthetic polyisocyanide (PIC) hydrogel with the integrin-activating domain of Invasin (INV). PIC hydrogels are soluble at 4 °C and form a gel at 37 °C [P. H. J. Kouwer et al., Nature 493, 651–655 (2013)]. When INV is covalently linked to PIC, the resulting hydrogel supports multipassage 3D growth of human intestinal and airway organoids. Self-renewal, polarization, and differentiation are maintained. The 3D swelling assay for cystic fibrosis drug testing (S. F. Boj et al., J. Vis. Exp. (2017), 10.3791/55159] was validated using PIC-INV. With PIC-INV hydrogels, we establish a fully defined and animal-free system for 3D TSC-derived organoid culture.

Matrigel, or Basement Membrane Extract (BME), is a protein extract that is used as an in vitro substitute for the basement membrane of epithelia. Basement membrane components bind and activate integrin complexes and thus play a crucial role in establishing apical–basal polarity of epithelial cells and in preventing their anoikis. Correct polarity ensures the proper distribution of growth factor receptors -typically to the basolateral side-, allowing cells to receive external signals necessary for survival, growth and differentiation (1–3). Matrigel/BME is produced by purifying extracellular matrix proteins secreted by Engelbreth–Holm–Swarm (EHS) sarcoma cells grown in the mouse abdomen (4). Early on, it was found that Matrigel contains approximately 10% collagenous material, primarily collagen IV, a key component of the basement membrane (5). Unlike cartilage-and in resemblance to the basement membrane-, Matrigel lacks proteoglycans and collagen fibrils. Later studies revealed that Matrigel consists of ~70% of embryonic laminin-1, also known as laminin-111 (6). While being liquid at 4 °C, Matrigel’s components can form a network at 37 °C, creating a hydrogel that provides a supportive microenvironment for cells (7). Mina Bissell championed the importance of the basement membrane proteins, particularly laminin-1, in understanding tissue architecture (8, 9). Furthermore, Bissell identified integrins as key mediators of this process, enabling cells to sense and respond to their microenvironment (10, 11).

Building on Bissell’s extracellular matrix (ECM) work and having identified the key growth factors of the intestinal stem cell niches, we established tissue stem cell (TSC)-derived organoids, miniature 3D versions of the gut epithelium (12, 13). These organoids can be cultured long-term and can generate all cell types of the intestinal epithelium. Since then, protocols have been developed that can obtain similar organoids for virtually any epithelial cell type (12, 14). Intestinal epithelial organoids require a niche-mimicking cocktail that consists of the following four components: Wnt pathway agonists (Wnt; nowadays replaced by Wnt surrogate), and R-spondin (15), epidermal growth factor (EGF), the BMP inhibitor Noggin, and Matrigel (12). As mentioned above, Matrigel exerts its positive effects on epithelial cell survival and polarity by binding and activating ECM receptors of the integrin class. Standard TSC-derived organoid protocols are based on a cocktail of fully defined, synthetic growth factors and do not require a source of serum. Matrigel/BME has remained the only nonsynthetic component of this culturing platform. It is produced in animals, is costly, and shows batch-to-batch variability while its composition remains poorly defined (16).

Integrin receptors consist of an α- and a β-chain, with distinct ligand recognition determined by specific α/β combinations (17). Currently, 24 integrin receptors have been identified, categorized into different classes based on their ligand recognition (18). Among these, integrin β1 is one of the most versatile β-chain integrins, forming a distinct family by pairing with the largest variety of α-subunits to create functional receptors. In epithelial cells, these α/β1 integrin complexes interact with specific basement membrane proteins. For example, α5β1 and αVβ1 recognize the well-known RGD amino acid sequence in fibronectin, whereas α6β1 and α3β1 belong to the laminin-binding class, interacting with trimeric laminin proteins (composed of α, β, and γ chains) (18, 19). Through interactions between integrin β1 receptors and basement membrane proteins such as laminin, cells regulate actin filaments, thereby controlling polarity and growth (20). Our recent finding shows that the integrin-binding domain of the Yersinia protein Invasin provides survival and polarity signals for epithelial cells, comparable to those of basement membrane proteins (21). The truncated Invasin protein uniquely overrides the ligand-recognition specificity of integrin classes, enabling interactions with integrins across different families. In other words, Invasin—used as a single protein—mimics the integrin ligands fibronectin and the various laminin isoforms (including laminin-1) by interacting with integrins α3β1, α5β1, α6β1, α7β1, as well as αVβ1 (22).

We hypothesized that presenting this universal ECM-replacing protein in a multimerized form in 3D would provide a valuable tool for organoid cultures, complementing the existing 2D organoid sheet cultures. To achieve this, the current study utilizes polyisocyanide (PIC) polymers, a synthetic polymer with a helical backbone and several side chains (23). The helical structure induces rigidity, resembling collagen fibers, while the side chains confer solubility (24). Upon heating, water molecules are squeezed out from the solubilized polymer, promoting interactions between PIC polymers which drive network formation (Fig. 1A), causing gelation. Uniquely, this process is fully reversible—upon cooling, the hydrogel redissolves—allowing convenient retrieval of embedded organoids. Like Matrigel/BME, the PIC hydrogel is designed to remain soluble at 4 °C and form a stable gel at 37 °C. Here, we covalently functionalize PIC polymers with Invasin proteins, generating a multimeric Invasin-rich microenvironment for epithelial cells. We then address whether this fibrous 3D Invasin network can support the formation and maintenance of various epithelial organoids.

Fig. 1.

Functionalization of PIC hydrogel with Invasin. (A) Schematic of PIC hydrogel formation and covalent linkage of Invasin. 1) PIC polymer. 2) Solubilized (H₂O) PIC forms bundles. 3) PIC sidechains modified with azide groups, Invasin (green) labeled with DBCO groups on free lysines (K) residues. 4) Spontaneous SPAAC reaction covalently links Invasin to PIC. 5) PIC bundles decorated with Invasin proteins (green), 6) PIC-INV hydrogel formation upon temperature increase. (B) Homogenous distribution of Invasin-Alexa647 covalently linked to the PIC hydrogel at varying concentrations. (Scale bar, 200 μm.) (C) Brightfield images of ileum intestinal organoids (line: N39) cultured in PIC hydrogels covalently linked with different concentrations of Invasin. Images show organoids at day 0 and day 1. Arrows indicate representative healthy cystic organoids. (Scale bar, 2,000 μm.)

Results

To covalently link the Invasin protein to the PIC polymer, we first labeled Invasin with dibenzyl cyclooctyne (DBCO) groups via primary lysines (K). The sidechains of the PIC polymer were modified with chemically reactive azide groups (0.21 mM). By mixing the azide-functionalized soluble PIC polymer with the DBCO-modified Invasin protein, spontaneous strain-promoted azide-alkyne cycloaddition (SPAAC) reaction occurred, covalently linking Invasin to the PIC bundles. Upon heating, the modified PIC polymer formed a PIC-Invasin hydrogel (PIC-INV) (Fig. 1A). Using an Invasin-Alexa647 bioconjugate, we confirmed the homogeneous distribution of covalently linked Invasin molecules within PIC hydrogels (Fig. 1B). To determine the Invasin concentration required for organoid formation, we titrated Invasin protein within PIC hydrogels and analyzed the formation of cystic organoids from ileum (N39) organoids. At a high concentration (35 μM), the physical properties of the PIC hydrogels were affected, as premature gelation occurred even on ice. Conversely, at a low Invasin concentration (10 μM), stimulation was insufficient to induce organoid formation, likely due to an inadequate number of ligands to drive integrin clustering, required for downstream intracellular signaling (25, 26). Notably, unbound (soluble) Invasin was nonfunctional within the hydrogel environment and did not promote organoid growth (SI Appendix, Fig. S1). This indicated that only covalently linked Invasin supports organoid growth in PIC hydrogels. We identified 20 to 25 μM Invasin as the optimal concentration range, as cystic intestinal organoids were formed already after 1 d of culture. Additionally, at this concentration, we found that a PIC polymer length of 2 K was optimal for organoid growth (SI Appendix, Fig. S2). We continued experiments with PIC-INV hydrogels, created under these optimized conditions.

Next, we investigated whether PIC-INV hydrogels could sustain human organoid growth over multiple days. Different organoid types (colon, ileum, airway) were cultured in the following hydrogels: PIC-INV, unmodified PIC, PIC functionalized with RGD peptides, and BME (Fig. 2A). Seminal integrin studies have identified RGD as the minimal functional fibronectin motif (27). The RGD tripeptide sequence has been used extensively since, due to its ease of synthesis and compatibility with coupling to synthetic hydrogels. As shown in Fig. 2A, PIC-RGD hydrogels failed to support the growth and maintenance of cystic intestinal or lung organoids after 9 d of culture, as did unmodified PIC hydrogels. Similar results were observed before with human liver organoids (28). In contrast, PIC-INV hydrogels enabled the formation of cystic intestinal organoids that appeared comparable to those grown in Matrigel/BME. Notably, airway organoids started from single cells at low seeding densities (1,000 cells/μL hydrogel) were able to form 3D organoids in PIC-INV, which did not occur in Matrigel/BME. Indeed, we have never been able to expand single cells derived from airway, at low cell densities, in Matrigel/BME. The proliferation results were quantified after 9 d of culture using an ATP-sensitive luminescent assay (Fig. 2A).

Fig. 2.

Growth, polarization, and characterization of human epithelial organoids in PIC-INV hydrogels. (A) Representative images of airway (LU30) (start 10,000 cells/well), colon (P26N), and ileum (N39) organoids after 9 d in different PIC hydrogels. Growth was quantified using an ATP-based luminescent assay (CellTiter-Glo) in relative light units (RLU). (B) Characterization of the different epithelial organoids in PIC-INV hydrogel. Ileum (N39) organoids were stained for the basal marker integrin β1, apical marker F-actin, and proliferation marker Ki67. Colon organoids were stained for the apical marker ZO-1 (zonula occludens-1). Airway organoids were characterized using the basal cell marker keratin 5 (KRT5) and F-actin staining. (Scale bar, 50 μm.)

To characterize human ileal and colonic organoids cultured in PIC-INV hydrogels, immunostaining for Ki67 demonstrated the proliferative activity of the 3D cultures. Correct apical–basal polarity was visualized using the polarity markers ZO-1, F-actin (apical markers), and integrin β1 (basal marker) (Fig. 2B). Human airway epithelial organoids were characterized using immunostaining for the basal cell marker keratin 5 and F-actin (Fig. 2B).

Further growth characterization in PIC-INV hydrogel was performed using ileal intestinal organoids endogenously labeled with the fluorescent ubiquitin cell cycle indicator (FUCCI) (21, 29). Flow cytometry analysis (Fig. 3A) confirmed comparable cell cycle states between intestinal epithelial cells cultured in PIC-INV and those cultured in Matrigel/BME hydrogel. Importantly, this reporter organoid line expanded with consistent cell cycle profiles in three additional, independently generated PIC-INV hydrogel batches (SI Appendix, Fig. S3).

Fig. 3.

Epithelial organoids maintain growth and differentiation capacity in PIC-INV hydrogels. (A) Cell cycle analysis by flow cytometry of human ileal intestinal organoids genetically modified with the FUCCI construct, cultured in PIC-INV and BME hydrogels. (B) Confocal images of Lgr5-GFP-tagged mouse small intestinal epithelial organoids cultured in PIC-INV and BME hydrogels. Arrows indicate the triangular Lgr5-positive stem cells. (Scale bar, 50 μm.) (C) Long-term culture of epithelial organoids from different species in PIC-INV hydrogels. Dots indicate passages. Passage ratios: 1:2 for N39 and P26N, 1:4 for mSI, and 1:3 for LU30. (D) Differentiation of epithelial organoids in 2D on Invasin-coated transwells after multiple passages in PIC-INV hydrogels. Endogenously tagged ileum epithelial cells were differentiated into goblet cells (MUC2), enteroendocrine cells (EEC, marked by CHGA), and Paneth cells (DEFA5). Airway epithelial cells cultured on Invasin were differentiated into ciliated cells (acetylated tubulin) and club cells (CC10), while retaining the basal cells (KRT5).

Next, murine intestinal organoids that carry a genetically engineered Lgr5-GFP reporter (30), were cultured in PIC-INV hydrogels to visualize the localization of the Lgr5-positive stem cells within these organoids. We observed that Lgr5-positive stem cells were located in crypt-like regions, interspersed between granule-containing Paneth cells (Fig. 3B) in the PIC-INV-cultured organoids. This was identical to the Matrigel/BME-cultured organoids (Fig. 3B).

Epithelial organoids from both human and mouse tissue could be expanded over multiple passages in PIC-INV hydrogels (Fig. 3C). To assess whether epithelial cells maintained the capacity to differentiate into functional cell types after several PIC-INV passages, the epithelial cells were cultured in a 2D format on Invasin-coated surfaces as reported previously (21). For validation of their differentiation capacity, intestinal and airway epithelial organoids were dissociated to single cells at passage 3 and passage 5, respectively. These single cells were seeded onto Invasin-coated transwell inserts. Once confluency was reached, differentiation was induced using an air–liquid interface (ALI) approach. After 2 wk of differentiation, organ-specific functional epithelial cell types were observed, including Paneth, goblet, and enteroendocrine cells for the intestinal epithelium, as well as ciliated and club cells in the airway epithelium (Fig. 3D). SI Appendix, Fig. S4 illustrates the same differentiation capacity when these organoids were grown in 3D in PIC-INV hydrogels. These findings demonstrate that the characterized 3D epithelial organoids, expanded over multiple passages in PIC-INV hydrogels, retained their ability to differentiate into mature functional epithelial cell types (SI Appendix, Fig. S5). From these experiments, we deduced that—at least in the experimental settings used—the epithelial cells behaved equally under 2D and 3D conditions. It will be of interest in future studies to determine which—if any—functional characteristics of epithelial cells rely specifically on a 3D shape of the structure in which they reside.

Since the transfer of organoid cells from 3D to 2D culture was successful, we next investigated whether the reverse was also feasible. Primary airway epithelial cells (line AO#1 from ref. 21) that had been expanded for 15 passages as 2D Invasin organoid sheets, were subsequently embedded in PIC-INV hydrogels. These cells successfully formed self-organizing organoids containing distinct airway epithelial cell types, including basal cells (KRT5), club cells (CC10), and goblet cells (MUC5AC) (Fig. 4A). Furthermore, when single airway organoid cells were replated in 3D PIC-INV, new organoids emerged, and this culture remained stable over multiple passages (Fig. 4C).

Fig. 4.

Establishment and functional validation of organoids derived from primary epithelial cells in PIC-INV hydrogels. (A) Confocal images of primary airway epithelial cells, initially established on Invasin and cultured for 15 passages in 2D (line AO#1 from ref. 21), then transferred to PIC-INV hydrogels. The cells formed airway organoids containing airway-specific cell types, including basal cells (KRT5), club cells (CC10), and goblet cells (MUC5AC). Nuclei are stained with DAPI. (Scale bar, 50 μm.) (B) Brightfield images showing multiple passages (P0 to P2) of airway cultures derived from freshly isolated tissue biopsies, grown in PIC-INV and BME hydrogels. Arrows indicate emerging organoids. Scale bar, 2,000 μm (overview images), and 400 μm (zoom-in images). (C) Long-term culture of primary airway epithelial organoids in PIC-INV hydrogel. Dots represent passages; passage ratio was 1:3. (D) Brightfield images of ileal organoids (N39) in PIC-INV hydrogels at T = 0 and T = 150 min during a forskolin-induced swelling assay. Organoids were segmented and tracked over time under control and forskolin-treated condition in the two hydrogels (PIC-INV and BME). Swelling was normalized to the size of organoid at T = 0; a 1.5-fold threshold (dashed line) was used to define swelling. (Scale bar, 0.25 mm.)

To further assess the robustness of this system, we aimed to establish organoid cultures directly from freshly isolated airway biopsies. The gradual emergence of airway epithelial organoids over several passages is shown in Fig. 4B. When replated as single cells, the organoids could be propagated over multiple passages in PIC-INV, maintaining long-term expansion potential (Fig. 4C). These cells were also capable of differentiating after 4 passages to functional airway epithelial cells using ALI procedures on 2D Invasin and BME, as shown in SI Appendix, Fig. S6. PIC-INV-derived airway organoids exhibited morphological characteristics indistinguishable from the airway cultures in BME.

While the 2D format dramatically simplifies the morphological and functional analysis of organoid cells, certain assays require a 3D culture format. A key example is the so called “swelling assay,” which is used to evaluate the function of the cystic fibrosis transmembrane conductance regulator (CFTR) protein in organoids derived from patients with CFTR mutations. Forskolin is added to the culture medium to stimulate intracellular cAMP production, which activates and opens CFTR channels in healthy organoids. This in turn leads to rapid influx of fluid into the organoid lumen, observed as organoid swelling over 1 to 2 h. In organoids with CFTR-null or nonfunctional CFTR, this swelling response is absent, allowing for the assessment of CFTR-targeting drugs to restore CFTR-channel function and consequently of the swelling capacity (31). Intestinal epithelial organoids were cultured in PIC-INV, and after one passage, forskolin was added to assess whether swelling could be induced. Within 2 h of forskolin incubation, all organoids show swelling (Fig. 4D), while none of the organoids in the control condition reached the swelling-threshold. The result was comparable to the observation made with organoids grown in Matrigel/BME cultures (Fig. 4D).

Discussion

Here, we show that coupling of the integrin-activating domain of the recombinant Yersinia protein Invasin to a PIC hydrogel results in a biologically active, functionalized hydrogel that shows the same temperature sensitivity as Matrigel/BME. The resulting PIC-INV hydrogel appears to be at least equipotent to Matrigel/BME in its support of epithelial organoid expansion. While standard TSC-derived organoid protocols are based on a cocktail of fully defined growth factors, Matrigel/BME has remained the only nonsynthetic component of this culturing platform. It is a costly animal product, while its composition remains poorly defined (16). We previously identified the Yersinia protein Invasin as a fully defined, synthetic replacement of Matrigel/BME for 2D “organoid sheet” cultures. The current study shows that coupling of Invasin to a PIC hydrogel can also replace Matrigel/BME in 3D.

We observe that PIC-INV hydrogels support organoid growth comparably to Matrigel/BME, whereas PIC-RGD hydrogels did not. This finding contrasts with previous results observed with poly(ethylene glycol) (PEG)-RGD hydrogels (32). These hydrogels decorated with RGD were predominantly effective for mouse organoid culture for a limited number of passages (three). For human epithelial organoid cultures, collagen-derived peptides, synergy sites with RGD sequences or ECM protein binders were added to enhance the performance of these hydrogels (33–35). However, these studies often lack data to demonstrate the feasibility of long-term, stable expansion of the different organoid types, which in our hands has been the single most critical parameter in the development of a new organoid protocol.

The PIC-RGD hydrogel used in the current study has been successfully used for long-term (more than 10 passages) culture of mouse mammary gland organoids, while it should be noted that the morphology of the organoid cultures in RGD-containing hydrogels has been different from those cultured in laminin-rich Matrigel (36). In addition, various human cancer cell types—which typically are less dependent on niche conditions than their wildtype counterparts—failed to form lumens when cultured in PIC-RGD hydrogels but were able to grow as dense spheres (37). These observations are consistent with findings from Schneeberger et al, who also failed to culture human ductal liver organoids in PIC-RGD hydrogels (28), in contrast to a PEG-RGD hydrogel study (32). Cyclic and linear RGD have distinct affinities for the integrin α5β1 receptor (38), the affinity of the ligand can affect the stimulation and growth of organoids in 3D. For example, Invasin has a dissociation constant (Kd) of 5 nM, whereas fibronectin has a much higher Kd of 800 nM toward the integrin α5β1 receptor (39). Affinities of RGD peptides fall within a range of 100 to 300 nM dissociation constants (38). Laminin molecules, on the other hand, have higher affinity for their integrin receptors, with a Kd around 5 nM, similar to that of Invasin (40). While direct comparison of these studies is challenging, they provide insights into the stability of integrin–ligand interactions, which may explain why PIC-RGD hydrogels are less effective. Improved ligand binding can be achieved by increasing the expression of the corresponding integrin on epithelial cells (41), such as α5β1 in the case of RGD-hydrogel grown epithelial organoids (42). However, several studies have shown that α5β1 integrin expression is low in various organoid types (21, 33, 34). As a result, additional elements are often incorporated into hydrogels to support long-term culture of multiple human organoids, with full-length laminin molecules proving to be most efficient. This has been demonstrated in studies using fibrinogen hydrogels (43), PEG hydrogels (44, 45), and PIC hydrogels (28). These findings suggest that laminin, rather than RGD-containing proteins, may provide the necessary stimuli for prolonged culture of multiple types of human organoids in defined synthetic hydrogels. Our study proposes that laminin (often isolated from the EHS-tumor) molecules can be replaced by the easy-to-produce Invasin protein, which, in addition to its high affinity for RGD-recognizing integrins, also exhibits a strong affinity for laminin-recognizing integrins such as α6β1 and α3β1. This likely explains why the PIC-Invasin hydrogel behaves similarly to laminin-rich Matrigel.

Future optimization of PIC-INV hydrogels will improve culture conditions for specific organoid types. For example, further optimization of Invasin, such as enhancing its specificity or universality for the various integrin complexes, in combination with fine-tuning hydrogel properties (e.g., stiffness or Invasin:DBCO ratio) could allow for the generation of more specific organoid types and cell populations within the organoids. Here, we observed that a 2 K PIC length is optimal for human intestinal organoids. Upon gelation, the polymer chains assemble into bundles that form networks, giving rise to a broad distribution of pores ranging from tens of nanometer to microns, depending on polymer length and concentration (46). This could be further optimized for each organoid type, as the pore size contributes to the inherent deformability of PIC gels, allowing physical remodeling that enables organoids to grow freely. The stiffness of the PIC gels is comparable to Matrigel (46) and could be adjusted for different organoid types, reflecting the unique mechanical requirements of each organ.

While many aspects of the hydrogel characteristics and applications remain to be explored, this study highlights the potential of using recombinant Invasin for 3D organoid cultures. We believe that integrating both 2D and 3D culture systems will increase the accessibility of organoid technology and allow for automation, validation, and standardization of organoid-based approaches in academia and industry.

Materials and Methods

Organoid Cultures.

BME-established 3D or Invasin-derived 2D organoids were obtained as described below and maintained in culture media specified in Table 1.

Table 1.

Table presenting media composition of different organoid types

Reagent	Source	Conc.	Colon Ileum	Airway	Mouse intestine
F12 advanced	Gibco	NA	+	+	+
Hepes	Gibco	10 nM	+	+	+
Glutamax	Gibco	1:100	+	+	+
N-acetylcysteine (Nac)	Sigma-Aldrich	1,25 mM	+	+	+
Nicotinamide (Nic)	Sigma-Aldrich	10 mM	+	+ (5 mM)
B27	Sigma-Aldrich	1:50	+	+ (B27+vitA.)	+
p38 inhibitor SB202190	Sigma-Aldrich	10 µM	+	+ (1 µM)
A83-01 inhibitor TGF-β type I receptor	Tocris	0,5 µM	+	+
PGE2	Tocris	1 µM	+
EGF	Peprotech	50 ng/mL	+		+
FGF10	Peprotech	0,1 µg/mL		+
FGF7/KGF	Peprotech	100 ng/mL		+ (25 ng/mL)
NGS Wnt Surrogate-Fc	#: N001 IpA	0,5 nM	+		+
Rspondin3- Fc	#: R001 IpA	3%	+	+ (1%)	+
Noggin-Fc	#: N002 IpA	2%	+	+ (1%)	+
Penecillin/Streptomycin	Gibco	100 U/mL	+	+	+
Primocin	Invivogen	50 µg/mL		+

Intestinal organoids: Mouse and human intestinal organoids were embedded in BME domes (R&D system) and cultured under standard conditions. Organoids were mechanically passaged every 5 to 7 d. For passaging, BME domes were disrupted in cold DMEM/F12 (Gibco) using a P1000 pipette tip, and the contents were transferred to a 15 mL tube. Organoids were centrifuged at 500 × g for 5 min at 4 °C. After centrifugation, the supernatant containing DMEM/F12 and residual BME was aspirated. New 500 μL DMEM/F12 was added, the organoid pellet was then fragmented mechanically using P1000 and P200 pipette tips.

Airway organoids: Human airway organoids were passaged approximately every 10 to 12 d via enzymatic dissociation. BME domes were disrupted in cold DMEM/F12, and organoids were pelleted at 500 × g for 5 min at 4 °C. The pellet was resuspended in 1 mL of prewarmed TrypLE Express (Gibco) and incubated at 37 °C. Dissociation was monitored by brightfield microscopy. Upon visible epithelial cluster formation, the suspension was pipetted using a P1000 tip to aid dissociation. The cell suspension was then centrifuged again (500 × g, 5 min, 4 °C), and the pellet was resuspended in cold BME.

For all organoids cultured, approximately 35 μL of the BME-cell suspensions was plated as domes in suspension culture plates (Greiner) and incubated at 37 °C. After 10 to 15 min, once the BME domes solidified, the appropriate culture medium was added.

The origin and derivation of organoid lines used in this study are as follows:

Human colon (P26N) organoids were previously described (47).

Human Ileum (N39) organoids were previously described (48).

Human airway (LU30) organoids were previously described (49).

Mouse small intestinal Lgr5-GFP organoids were previously described (30).

Human airway 2D-organoid sheets were previously described (21).

Bacterial Invasin Production, Purification, and DBCO Modification.

Invasin proteins were produced and purified as previously described (21), with modifications to enhance yield. Escherichia coli shuffle T7 K12 cells (New England Biolabs) were cultured in Terrific Broth (TB) medium (VWR) supplemented with 1% glucose to suppress premature protein expression. Upon reaching optimal growth, protein expression was induced by 0.4 mM IPTG (isopropyl β-D-1-thiogalactopyranoside, Sigma), without glucose, followed by overnight incubation at 16 °C. Purified Invasin proteins were chemically modified with DBCO groups by targeting the free lysine residues using an EZ-link DBCO protein labeling kit (ThermoFisher), following the manufacturer’s instructions. Prior to conjugation, Invasin was concentrated to approximately 5 mg/mL using Amicon Ultra-15 mL centrifugal filters (Amicon, UFC8010). Protein concentration was determined via Nanodrop A280 values, and the protein characteristics (extinction coefficient and Abs0.1%) were evaluated with ProtParamTool. A 20 mM excess of DBCO crosslinker was added relative to the protein concentration to obtain efficient coupling. The final Invasin concentration was optimized to 20 µM in the PIC hydrogel to prevent excessive dilution of the PIC hydrogel.

Preparation and Functionalization of PIC Hydrogels.

Azide-functionalized (0.21 mM Azide) or RGD-modified polyisocyanopeptide (PIC) polymers were obtained from the Paul Kouwer laboratory (46). A 10 mg/mL stock solution was prepared by dissolving 10 mg PIC in 1 mL of cold PBS, followed by overnight incubation at 4 °C. The solution was thoroughly mixed by pipetting to ensure a homogeneous solution. The 1 K, 2 K, and 5 K lengths of PIC polymers were measured based on viscosity-averaged weights M_V using Mark–Houwink equation with MH constants determined for other polyisocyanides. Using this protocol, the M_v values were determined as follows 251, 401, and 521 kg/mol for PIC 1 K, 2 K, and 5 K, respectively. The length, e.g., 2 K, stands for a polymerization with a monomer:initiator ratio of 2,000:1. To prepare functional PIC-INV hydrogels, solubilized PIC-azide solution was incubated overnight at 4 °C with DBCO-modified Invasin protein, achieving a final Invasin concentration of 20 µM in the hydrogel. The overnight incubation facilitated a slow and efficient SPAAC reaction, ensuring maximal covalent attachment of Invasin molecules to the PIC hydrogel. The PIC-INV was prepared at a concentration at 5 mg/mL and stored at 4 °C for up to 1 wk or at −20 °C for longer storage. The initial concentration of PIC 5 mg/mL accounted for dilution caused by residual organoid medium or DMEM/F12 in the 15 mL Falcon tube during organoid culture and passaging. As a result, the final PIC working concentration for organoid growth was consistently around 2 mg/mL.

Organoid Culture in PIC Hydrogels.

BME-grown epithelial cells derived from mouse small intestine, human colon, human ileum, or human airway were treated with Dispase (5 U/mL) (ThermoFisher) to release the epithelial cells from the BME matrix. Dispase was added to the culture medium, and BME domes were mechanically disrupted before incubation at 37 °C for 30 min. The degradation of the BME was monitored using brightfield microscopy. After treatment, organoids were washed twice with cold DMEM/F12 by centrifugation (500 x g, 5 min, 4 °C). Intestinal organoids were mechanically fragmented, while airway organoids were enzymatically dissociated into single cells using TrypLE, as described previously. The cells were then resuspended in cold PIC-INV (PIC, or PIC-RGD) solutions, and hydrogel domes (10 to 20 μL) were formed in suspension culture plates (Greiner). After incubation at 37 °C for 10 to 15 min, the polymerized PIC hydrogels were overlaid with the appropriate culture medium. For passaging, organoid media were aspirated, and cold DMEM/F12 was added to dissolve the PIC hydrogel, which liquifies at low temperatures. After 5 to 10 min of incubation, with intermittent pipette mixing, organoids were collected in a 15 mL tube, centrifuged (500 g, 5 min, 4 °C), and processed as described earlier: mechanically fragmented (intestinal organoids) or enzymatically dissociated (airway organoids). The dissociated cells were then replated in fresh PIC-INV hydrogels for continued culture.

Primary Tissue Culture in PIC Hydrogel.

Healthy airway cells were derived from biopsies obtained after surgery at the Diakonessen Hospital Utrecht. Informed consent was obtained from the patient, and this study was approved by the ethics committee of the University Medical Center Utrecht. The tissue was minced using scissors and razor blades and transferred to a 50 mL tube. Tissue was further digested using 0.5 mg/mL collagenase (Sigma Aldrich) in DMEM/F12 and incubated at 37 °C on a shaker for 25 min. DNAse (50 μg/mL, Sigma Aldrich) was then added and samples were incubated on ice for 5 min. The digested tissue was filtered through a 500 μm filter (pluriSelect) using a syringe plunger, rinsed with 10 mL cold DMEM/F12, and then filtered through a 100 μm filter (Greiner bio-one). The cell suspension was centrifuged (500 x g, 5 min, 4 °C), and the pellet was resuspended in 1 mL red blood cell (RBC) (Roche) lysis buffer, incubated at room temperature for 5 min, and then quenched with 15 mL DMEM/F12. After centrifugation, the single cells were plated in PIC-INV or BME hydrogel domes. The domes were covered with airway medium containing 10 μM Y-27 (Rho kinase inhibitor, Abmole) for the first 24 h.

Quantification of PIC Cultures.

After 9 d of culture, human ileum, colon, and airway organoid cells grown in different PIC hydrogels and BME were imaged using an EVOS microscope. On the same day, organoid viability was quantified using an ATP-sensitive luminescent assay (CellTiter-Glo, Promega). Organoid media were aspirated, and domes were overlaid with CellTiter-Glo solution (1:1 diluted in DMEM/F12) and incubated at room temperature for 10 min. After incubation, the CellTiter-Glo solution was transferred to a white 96-well plate (Greiner, 655074), and ATP-driven luminescence was measured using a Tristar multimode plate reader (Berthold) to assess cell viability.

Microscopy and Immunofluorescent Staining.

Epithelial organoids (colon, ileum, mouse intestine, airway) grown in PIC-INV hydrogels were fixed with prewarmed (37 °C) 4% formaldehyde (pH 7.4) for 30 min. Unlike organoids grown in BME, those grown in PIC hydrogels were not removed from their hydrogel before fixation, as there was no risk of mouse-derived contaminants interfering with primary or secondary mouse antibodies in PIC hydrogels. Following fixation, the organoids embedded in the 3D hydrogel domes were washed with PBS and permeabilized with 1% Triton X-100 (VWR) in PBS for 30 min at room temperature. The fixed and permeabilized organoids were then incubated overnight at 4 °C with primary antibodies (details below). After incubation, the hydrogel domes were extensively washed with PBS to remove unbound antibodies. To detect primary antibody binding, species-specific fluorophore-conjugated secondary antibodies (1:1,000) were applied for 1 h at room temperature. After secondary antibody incubation, the organoids were thoroughly washed (4×) in PBS. During the first 10-min wash step, organoids were incubated with DAPI (ThermoFisher), 1:1,000 in PBS, to stain nuclei. Fluorescently labeled organoids were imaged using an SP8 confocal microscope (Leica). Image processing and analysis were performed using ImageJ or Imaris software.

Phalloidin 647 (Sigma, 65906)	1:10,000
Somatostatin (Proteintech, 17512-1-AP)	1:250
Acetylated tubulin (Santa Cruz, 6-11B-1, sc-23950)	1:250
Uteroglobin/CC10 (R&D, MAB4218)	1:250
Ki67 (Millipore, AB9260)	1:250
KRT5 (AF138, PRB 160P-100; 905501)	1:1,000
MUC5AC (Invitrogen, MA5-12178, 45M1)	1:250
ZO-1 (Invitrogen, PA5-19090)	1:250
AIIB2 (kind gift of A.Sonnenberg) (1 mg/mL)	1:250

Differentiation in 2D Invasin Transwell Systems.

Intestine ileum (N39) triple reporter (50), airway (LU30, primary tissue line AO#6), and primary airway tissue (AO#6) were passaged in PIC-INV hydrogels. After 3 to 5 passages, organoids were collected and enzymatically dissociated into single cells using prewarmed TrypLE (Gibco). The digestion was carried out at 37 °C for 10 min, and dissociation was monitored by brightfield microscopy. Dissociated cells were washed twice with cold DMEM/F12, and approximately 50,000 cells were plated in 12-well transwell systems (Greiner, 3 μm pore size) in appropriate culture medium. Transwells were precoated at 4 °C for 12 h with either 5 μg/mL Invasin or 2% BME. Once epithelial confluency was reached, differentiation was initiated using an air–liquid interface (ALI) protocol.

For intestine: A patterning medium [as described previously (50)] was added to the lower transwell compartment for 7 d, with media refreshed every 2 d, while the upper compartment was exposed to air. After 1 wk, the patterning medium was replaced with maturation medium (50), and the culture continued for an additional 3 to 5 d. Fully differentiated intestinal cell types were observed using a confocal SP8 (Leica) microscope.

For airway: Differentiation was initiated by adding PneumaCult medium (stem cell technology, #05001) to the lower compartment, with the upper compartment exposed to air. Media was changed every 2 to 3 d. After 10 to 14 d of differentiation, the presence of ciliated cells was observed using brightfield microscopy. Once ciliated cells appeared, the epithelial layer was fixed, and differentiated cells were imaged using a confocal SP8 (Leica) microscope and analyzed using Imaris software.

Ileum Differentiation in 3D PIC-INV Hydrogel.

Differentiation of 3D PIC-INV intestinal cultures was performed after 3 passages in PIC-INV hydrogels. Differentiation was initiated by switching the expansion medium to the patterning medium for 14 d. Afterward, the patterning medium was changed by the maturation medium for 7 d. The differentiation protocol and differentiation medium composition were described previously (50).

Immunohistochemistry Staining of Differentiated Intestinal Cells.

Differentiated (50) 2D-Invasin, PIC-INV, and BME-derived intestinal cultures were fixed in 4% paraformaldehyde for 30 min at room temperature, prior to medium removal. Following fixation, the samples were dehydrated, paraffin-embedded, and sectioned. Standard hematoxylin and eosin and periodic acid–Schiff stainings were performed. For immunohistochemistry, sections were stained using an antilysozyme antibody (DAKO, A0099, 1:4,000). Digital images were obtained using a VS200 slide scanner and analyzed with Olyvia software.

Ghrelin ELISA.

2D-Invasin cultures, 3D PIC-INV, and BME organoids were differentiated toward hormone-producing enteroendocrine cells. The presence of enteroendocrine cells in the cultures was verified using fluorescent markers included in the ileum (N39) organoid line. Two days after refreshing the differentiation medium, the culture medium was collected. The hormone ghrelin levels were measured using a human ghrelin ELISA kit (Abcam, GR3454028-1), following the manufacturer’s instructions.

Paneth Cell Extrusion Assay.

Ileum-derived intestinal organoid cultures under 2D Invasin conditions were differentiated toward Paneth cells as previously described (21). Paneth cells were identified based on DEFA5-dsRED expression. Following differentiation, cells were treated with 0.5 ng/mL interferon-gamma (IFNγ, PeproTech) for 2 d (51). After treatment, the number of Paneth cells was quantified by flow cytometry based on DsRED fluorescence.

Flow Cytometry Analysis.

The ileum FUCCI reporter organoids (from ref. 52) were cultured in PIC-INV hydrogels for 9 d. Then, the organoids were released from the hydrogel by incubating with ice-cold DMEM/F12. The liberated organoids were collected in a 15 mL tube and centrifuged (500 g, 5 min, 4 °C), and the resulting pellet was enzymatically dissociated into single cells using 1 mL TrypLE (Gibco) at 37 °C for 5 min, with intermittent gentle pipetting. Dissociation was monitored using brightfield microscopy. The dissociated cells were washed with DMEM/F12 to stop the enzymatic reaction and stained with DAPI to exclude dead cells. At least 5,000 DAPI-negative cells were analyzed using a LSRFortessa Cell Analyzer (BD Biosciences). To analyze batch-to-batch differences, we synthesized three independent PIC-INV and used three BME batches (Lot 1695623, Lot 1765326, Lot 1788484); we cultured the ileum FUCCI reporter in these hydrogels for 7 d. After 7 d, we quantified 8000 DAPI-negative single cells using LSRFortessa Cell Analyzer (BD Biosciences).

Swelling Assay.

Intestinal organoids were cultured in PIC-INV hydrogels. One day after passaging, organoids were exposed to 10 μM Forskolin, which was added to the intestinal expansion medium. To analyze swelling, the organoids were imaged every 30 min for a total of 150 min using the EVOS microscope. The control group remained in the expansion medium without Forskolin. Organoids were segmented and tracked over time using Fiji (ImageJ) software. The area of each organoid was measured at each time point, and swelling behavior was normalized to the initial organoid size at T = 0. These normalized swelling curves were plotted to determine the proportion of organoids that responded to Forskolin, and a defined swelling threshold (indicated in the graph) was defined.

Supplementary Material

Appendix 01 (PDF)

Data, Materials, and Software Availability

All study data are included in the article and/or SI Appendix.