The application of explants, crypts, and organoids as models in intestinal barrier research

ABSTRACT

In vitro models are of great importance in advancing our understanding of human diseases, especially complex disorders with unknown etiologies like inflammatory bowel diseases (IBD). One of the key IBD features is the increased intestinal permeability. The disruption of the intestinal barrier can occur due to a destructive inflammatory response involving intestinal cell death. Alternatively, proteins that form tight junctions (TJ) fail to form function complexes and promote epithelial barrier disruption. The mechanisms behind this process are not fully understood. Thus, in vitro models that facilitate studying the intestinal barrier and its molecular components are of particular importance in the context of IBD. There are in vitro and ex vivo models that can be used to recapitulate some aspects of IBD. Among these are intestinal explants, crypts, and epithelial 3D-organoids. Here we describe some practical limitations of isolated crypts, gut tissue explants, and intestinal organoids as models in epithelial barrier biology, and TJ in particular. Our findings demonstrate that only 3D intestinal organoids formed from single cells are suitable to study barrier permeability in vitro, as primary crypt-derived organoids do not retain epithelial integrity due to cell death. Importantly, 3D organoids raised in culture conditions may fail to recapitulate inflammatory and barrier phenotypes of the source mouse model. To study the features of the inflamed epithelium, ex vivo intestinal explants and crypts were employed. We show here that isolated crypts do not preserve native TJ structure in a long-term experimental setting and tend to disintegrate in the unsupported culture environment. However, intestinal explants were stable in culture conditions for about 24 hours and demonstrated their applicability for short-term living tissue imaging and fluorescence recovery after photobleaching (FRAP). Thus, a combination of 3D organoids and intestinal explants provides a more accurate experimental platform to understand the intestinal epithelial barrier.

Introduction

Increased intestinal permeability is one of the key components of the inflammatory bowel diseases (IBD) and other gastrointestinal conditions. IBD encompasses chronic relapsing disorders of the gastrointestinal system including Crohn’s disease and ulcerative colitis. Patients diagnosed with IBD often demonstrate higher intestinal permeability than healthy subjects.¹ It is believed that epithelial barrier opening serves as a protective mechanism enabling immune cells to combat invading pathogens. However, dysregulation of the intestinal permeability can lead to excessive loss of water and ions, along with the uncontrolled transport of metabolites pathogen-associated molecular patterns throughout the body, potentially causing toxicity and exaggerating the immune response.²

The mechanistic background of the increased intestinal permeability is not fully understood, but it is closely associated with alterations in tight junctions (TJ). These apical junction complexes connect adjacent epithelial cells, thereby maintaining tissue integrity. They are anchored to the cortical actin filaments underlying plasma membranes and therefore connect to intracellular skeleton.^1,2 Research has consistently demonstrated that TJ functionality is dependent on actin dynamics and cytoskeleton stability.^3–6 To unravel the mechanisms underlying actin and TJ dysfunction in the context of the intestinal barrier, the researchers are in need of relevant models that accurately reconstitute the complexity of the native epithelial tissue. Given that in vivo research implies anatomical, technical, and ethical challenges, alternative approaches are being developed.

Cell cultures are broadly used for evaluating molecular mechanisms of the intestinal permeability regulation.^7,8 However, it is worth mentioning that most in vitro cell models cannot replicate the complex architecture of the native tissue and have their own limitations. Many widely used cultured cell lines do not reflect the diversity of cell types in the intestinal epithelium, for example, Caco-2 cell line lacks goblet-like cells.^9,10 Colonic cancer cell lines typically contain multiple karyotypic mutations that could influence junctional integrity.¹¹ The characteristics of commonly used Caco-2 cells can vary depending on culture conditions and passage number.¹² HT29 cells may not express all the necessary hydrolases and exhibit distinct ion transport properties.¹² An example of a cell line close to the natural intestinal cells is a normal non-transformed non-immortalized human intestinal epithelial crypt (HIEC) cell model. HIEC cells are considered to be undifferentiated crypt-like progenitor cells that keep the potential for specialization into stem cells by switching the activation statuses of the WNT/BMP pathways.¹³ However, the HEIC cells do not express the tight junction protein Occludin, which limits their use in barrier research.¹⁴ Primary human intestinal epithelial cells (HInEpCs) and induced pluripotent stem cell (iPSC)-derived intestinal cells represent better examples of cell lines that can be used for barrier research, but their cell type diversity requires further assessment.¹⁵ Finally, organoid-derived cultured cell monolayers established by immortalization appear to be a good model for barrier research. They retain the cell-type diversity of the natural tissue and form functional adhesion complexes.¹⁶

Likewise, intestinal 3D organoids also resemble native epithelium representing major natural cell types with functional TJ complexes and appear a promising alternative to study barrier permeability and programmed cell death.¹¹ While phenotypes arising from defined genetic mutations can be successfully reconstituted in tissue organoids, complex inflammatory phenotype from the IBD patients is not always retained in culture conditions.^17–19 These data agree with transcriptomic studies, showing that culturing intestinal organoids in vitro leads to the loss of specialized gene expression profiles emphasizing the role of niche and systemic factors in epithelial physiology.¹⁶

Thus, for a more in-depth investigation of intestinal barrier dysfunctions arising from the complex inflammatory environment, ex vivo organ culture models like intestinal explants can be employed. An explant is a piece of a tissue or an organ that has been removed from a living system and maintained under culturing conditions.²⁰ Explants have been actively used in a number of applications both in animal models and in human samples.^21–23 Explant models retain the intestinal mucosa’s characteristics, enabling controlled experimental conditions and reduced animal use. A number of complex devices as well as methodological variants to culture explant tissues have been offered like mechanically separated mucosa or isolated crypts.^24,25 Isolated crypts seeded on a specialized matrix proved to be a good model to study energy metabolism and mitochondrial respiration in vitro.^26–28 Yet, these models come with technical challenges, such as the need for costly equipment and reagents, as well as the elaborate techniques, which can pose difficulties for the beginner researchers.

Ex vivo explants could serve a perfect model in terms of the intestinal barrier dysfunction that originate within complex inflammatory environment. However, current evidence on the stability of F-actin and TJ structures during explant culture is limited. In this brief report, we present possible experimental applications for colonic organoids and explants, focusing on F-actin, TJ structure, and barrier integrity. We also tested isolated colonic crypts as an alternative to full-size explants in a short culture experiment. We found that isolated colonic crypts could not sustain viability and tissue structure in a short-term setting. Our data suggest that long-term culture of the colonic explants causes the loss of regular F-actin structure and results in delocalization of claudins. At the same time, short-term experiments on live-cell actin dynamics can be successfully performed with tissue explants. To further develop the use of intestinal explants, we developed an imaging setting to immobilize intestinal tissue under an inverted microscope, enabling a from-the-lumen perspective of the epithelial surface. The explant experiments were complemented with a functional permeability test in 3D organoid culture. Our data suggest that careful evaluation of organoids based on size, developmental stage, and living cell count is needed in order to provide reproducible results. According to our findings, organoids derived from Muc2^-/- transgenic mouse model of intestinal inflammation do not retain cellular phenotypes when cultured in vitro. Thus, we suggest that combining different in vitro and ex vivo models provides a better insight into molecular and cellular mechanisms behind IBD.

Materials and methods

Animals

The experiments were performed using C57BL/6J mice (our in-house sub-colony of C57BL/6J mice) and Muc2^−/− mouse strain. Muc2^−/− mice were obtained by the rederivation of previously generated Muc2^tm1Avel/Muc2^tm1Avel mice on the C57BL/6 genetic background in SPF CD1 female mice and backcrossing to C57BL/6. Mutant mice and their wild-type littermates (Muc2^+/+ mice) were obtained by crossing Muc2^+/- females with Muc2^+/- males.

Mice were housed in groups of the same-sex siblings in open cages (3W, Russia) with birch sawdust as litter and paper cups as shelter. The housing conditions were: 12 h/12 h light/dark photoperiod [12:00 (OFF) : 00:00 (ON)] with 22–24°C temperature, 30–60% humidity; food (BioPro, Russia) and water were provided ad libitum. For all experiments, mice were humanely euthanized via cervical dislocation.

This study was conducted in accordance with the recommendations of the Animal Research Guidelines of the Ethics Committee on Animal and Human Research of the Institute of Molecular and Cellular Biology (Novosibirsk, Russia). All the protocols were approved by the Ethics Committee on Animal and Human Research of the Institute of Molecular and Cellular Biology (Novosibirsk, Russia), protocol №02/21 from 04 August 2021.

3D organoid culture

Crypt isolation and organoid culture were performed similarly with^29,30 and according to StemCell’s Intesticult Organoid Growth Medium directions for use. Colon specimens were excised from mice, rinsed using 23 G catheter with cold (+4°С) Dulbecco’s Phosphate Buffered Saline (DPBS) (D5652, Sigma, USA) containing 1% antibiotic-antimycotic solution (ABM) (A5955, Sigma, USA), then cut longitudinally and washed in cold DPBS+ABM 3 times. The tissue was then sliced into small pieces on ice and digested for 30 minutes at 37°C in DMEM/F12 (D8900, Sigma, USA) medium containing 2 mg/ml collagenase I (Gibco, USA), 10% FBS, and 1% ABM. The crypt suspension was filtered through a 70 µm strainer, washed twice with DMEM/F12 supplemented with 10% FBS and 1% ABM, and pelleted by centrifugation at 270 g, 4°C. The crypts were resuspended in Organoid-Star Matrigengel (0827555, ABW, USA). The drops of 25 µl were seeded into 48-well culture plates. Plates were incubated for 5 minutes at 37°C until the matrix polymerized. Culturing was performed in IntestiCult Organoid Growth Medium (06000, Stemcell Technologies, Canada) with 1% ABM at 37°C and 5% CO₂. The medium was changed every three days. After 7–10 days of culture, organoids were passaged. For passaging, IntestiCult was removed, Matrigengel drops were resuspended in 0.25% trypsin-EDTA (T4049, Sigma, USA) and incubated for 5 minutes at 37°C, 5% CO₂. Organoids were washed with DMEM/F12 + 10% FBS + 1% ABM, centrifuged at 200 g for 5 min. The supernatant was removed, organoids were resuspended in Matrigengel and seeded as described earlier. Further procedures (fixation or permeability assay) were conducted after 3 days post passaging.

Crypts culture on flat matrix

Seeding of colonic crypts on flat matrix performed as described in,^26–28 Colon specimens were excised from mice, rinsed using 23 G catheter with cold (+4°С) Dulbecco’s Phosphate Buffered Saline (DPBS) (D5652, Sigma, USA) containing 1% antibiotic-antimycotic solution (ABM) (A5955, Sigma, USA), then cut longitudinally and washed in cold DPBS+ABM 3 times. The tissue was then sliced into small pieces on ice and digested for 30 minutes at 37°C in DMEM/F12 (D8900, Sigma, USA) medium containing 2 mg/ml collagenase I (Gibco, USA), 10% FBS, and 1% ABM. The crypt suspension was filtered through a 70 µm strainer, washed twice with DMEM/F12 supplemented with 10% FBS and 1% ABM, and pelleted by centrifugation at 270 g at 4°C. The crypts were resuspended in complete Intesticult Organoid Growth Medium (06000, Stemcell Technologies, Canada) with 1% ABM. About 2000–3000 crypts in 500 µl Intesticult were seeded in 24-well plate pre-coated with Organoid-Star Matrigengel (0827555, ABW, USA) in DPBS (1:50) and incubated for 1, 2 and 3 hours at 37°C, 5% CO₂. Following incubation, the medium with crypts was removed, 50 µl of suspension was taken to assess viability and the remaining crypts were pelleted by centrifugation at 270 g, 4°C and fixed in 4% paraformaldehyde for immunohistochemical analysis.

Permeability assay

Permeability testing followed the protocol outlined in.³¹ In brief, all wells were rinsed with DMEM/F12 containing 10% FBS and 1% ABM to collect cultured organoids from the plate. After centrifugation at 300 g at 4°C, the supernatant was removed, and the organoids were resuspended in a 2 mg/ml FITC-Dextran 4kDa (FD4-100 MG, Sigma) + 10 µg/ml Hoechst 33342 (G1127, ServiceBio) + 10 µg/ml 7-Aminoactinomycin D (A1310, Invitrogen) in the culture medium. For experimental groups, 10 µM U-73122 (ab120998, Abcam, USA) or 5 mm phosphatidylinositol (# P6636, Sigma, USA) were added. After 1 hour of incubation, organoids were transferred to the 8-well µ-slide (SPL Lifesciences, Korea) and observed under a confocal microscope LSM 710 (Zeiss) with the ZEN 2012 software. The µ-slide was placed on a temperature-controlled microscope stage (37 ℃) with CO₂ supplied. The microscopy was conducted at the Joint Access Center for Microscopy of Biological Objects with the Siberian Branch of the Russian Academy of Sciences.

Fluorescence intensity was quantified using ImageJ software. Permeability was evaluated as FD4 fluorescence intensity in the lumen of organoid divided on the mean of three background regions around the organoid (Figure 1(b)).

Figure 1.

Organoids in barrier research. (a) Permeability assay for organoids. 7-AAD stains permeable cells causing leakage in organoids. (b) Basal-out organoid shows both epithelial integrity and paracellular permeability for FITC-dextran 4 kDa (FD4) (arrowheads). Blue and pink ROIs are shown as an example for relative FD4 fluorescence quantification using the formula. (c) Organoid permeability assay. (d) Inhibitor of phospholipase C U-73122 increases permeability of organoids, while exposure to phosphatidylinositol (PI) does not affect permeability (**p < 0.01, Student’s test). Bars for A-C −50 µm. (e) Localization of Claudin 7 (green) and F-actin (red) are disrupted in the Muc2^-/- colon (left), while Muc2^-/- derived organoids (right) do not retain these features. (f) Fluorescence intensity quantification of Claudin 7 in the colon and colonic organoids of Muc2^+/+ and Muc2^-/- mice (***p < 0.001, n = 3, Student’s t-test). Bars for E-F −10 µm. (g) A scheme representing the algorithm for Cldn7 quantification in colon or colonoids. More detailed information can be found in the methods section.

Explants culture

Explant cultivation was performed similarly with.³² Mice were humanely euthanized via cervical dislocation, and descending colon samples were obtained. The colonic contents were immediately washed out using 23 G catheter with cold (+4°С) Dulbecco’s Phosphate Buffered Saline (DPBS) (D5652, Sigma, USA) containing 1% antibiotic-antimycotic solution (ABM) (A5955, Sigma, USA), then cut longitudinally and sliced into 0.5 × 0.5 cm fragments. Then colon fragments were washed three times with cold DPBS to remove dead cells and debris. The prepared explants were incubated on the surface of complete IntestiCult Organoid Growth Medium (06000, Stemcell Technologies, Canada) with 1% ABM at 37°C in a 5% CO₂ incubator with gentle horizontal shaking at 90 rpm up to 30 h. For keeping tissue on the air-liquid surface mini cell strainers with 100 µm nylon membrane were inserted in 24-well plate. After incubation, explant was cut in half for viability assessment and for immunohistochemistry assay.

The “Tube-gut sandwich” explants were made as follows: A 1.5–2 cm colonic fragment was rinsed in cold DPBS with 1% ABM, then opened longitudinally. The tissue was placed between two frames made out of a 0.5 ml microcentrifuge tube (1110–00, SSI Bio, USA) (Figure 4(a)). The inner frame was made out of the cap, and the outer frame was made out of the upside of the tube (2 mm in height). The inner frame was covered with colonic tissue, the mucosa facing up. The tissue was then clamped from above with the outer frame, as if the tube was locked. These explants were incubated in DMEM/F12 culture medium with 10% FBS and 1% ABM at 37°C in a 5% CO₂ incubator with gentle horizontal shaking at 90 rpm.

Figure 4.

The “tube-gut sandwich” for live imaging. (a) The scheme of the “tube-gut sandwich” assemble. Alive tissue imaging must be performed for 1 hour, approximately. (b) Live confocal imaging of the intestinal epithelium in the “tube-gut sandwich”. Blue – hoechst 33258, red – SiR-actin. (c) Immunohistochemistry of the intestinal epithelium in the “tube-gut sandwich”. Blue – DAPI, red – F-actin, green - Claudin 7. Bar −50 μm.

Assessment of actin dynamics in living explants

Colon explants were placed in DMEM/F12 (D8900, Sigma, USA) with 10% FBS, 1% ABM, and CellMask™ Green Actin Tracking Stain (SC001, Invitrogen, USA) diluted according to the manufacturer instructions. Samples were stained for 30 minutes at 37°C, 5% CO₂, in a 48-well culture plate at 90 rpm with gentle horizontal shaking. 150 nM Jasplakinolide (sc-202191A, Santa Cruz Biotechnology, USA) was added to the experimental group and the equivalent volume of DMSO was added to the control group. The explants were observed under a confocal microscope LSM 710 (Zeiss, Germany) with the ZEN 2012 software. For live-cell imaging, tissue was placed in the Attofluor™ Cell Chamber (Invitrogen, USA).

Acquisition settings for FRAP timelapses were adjusted with considerations described in.³³ Time-lapse images were taken with a 63×/1.4 plan-apo objective with oil immersion. The microscope was controlled by ZEN software, and the settings were as follows: format 512 × 512, speed 12, unidirectional with one line averaging, and pinhole adjusted to 1 airy unit. Circular ROI with radius of 1 µm was drawn on the brush border of the enterocyte. The frame of 40,81 × 40,81 µm including the ROI, was imaged 3 followed by photobleaching that was achieved with one scan (total laser power = 100%) of the ROI. Imaging of the area was resumed immediately after the photobleaching in a total of 37 cycles with 10-seconds interval. 16 timelapses per group were analyzed. Measurements were performed with ImageJ software in three regions: the bleached ROI, the reference region (non-bleached brush border), and the background region (Figure 5(b)). Fluorescence recovery was quantified using the following formula: (ROI_bleached – ROI_background)/(ROI_reference – ROI_background) at each time point.

Figure 5.

Fluorescence recovery after photobleaching in colon explants. (a) Timelapse of F-actin recovery with and without jasplakinolide (JP). (b) FRAP curves. Data is presented as mean ± standard error of the mean, n = 16. An analysis of variance yielded significant effect of the group, F(65, 1089) = 11.9, p < 0.001. A post hoc Tukey test showed that at 10 s and further p-value is < 0.01 between groups. Timelapse is an example, where ROIs for quantification are marked with circles: white – bleached ROI, magenta – reference ROI, blue – background ROI. Bars −5 μm.

Immunohistochemistry

Samples from the descending colon, cultured explants, or organoid suspensions were fixed in 4% paraformaldehyde. The explants and colons then were kept in 15% sucrose for 12 h and in 30% sucrose for another 12 h. Sucrose-protected tissue was embedded in Tissue-Tek® O.C.T. Compound (4583, Sakura, USA) and quick-frozen in liquid nitrogen. Sections of 40 μm in thickness were prepared using a CM1850 UV cryostat (Leica Biosystems, Germany). Tissue sections and organoids were washed, blocked, and permeabilized with 0.5% BSA, 0.3% Triton X-100 in PBS for 2 h at 4°С. Primary antibodies to Claudin-7 (# 37–4800, Invitrogen, USA) were incubated overnight at 4°C in PBST + 0.5% BSA then washed three times with PBS. Secondary antibodies (A-11001, Invitrogen, USA) with 1 µg/ml DAPI (HiMedia, USA) and Alexa Fluor™ 568 Phalloidin (A12380, Invitrogen, USA) were incubated for 2 h at room temperature in PBST + 0.5% BSA. Samples were stored in a 1:1 PBS + glycerol solution on glass slides at 4°C. Imaging was performed with an LSM 710 Axio Observer confocal microscope (Zeiss, Germany) in Z-stack mode using same settings for all shots, and fluorescence intensity was quantitatively analyzed from 11 merged stacks (2.412 µm) using ImageJ software. Each measurement was calculated as mean intensity in the line area of enterocyte lateral membrane, and black background intensity was subtracted (Figure 1(g)). Lateral membranes were initially traced at F-actin channel, then corresponding area of Claudin 7 channel was measured. Plots were created with the PlotsOfData.³⁴

Explants and crypts viability

Viability of crypts and explants was assessed by staining with 1 µg/ml DAPI (HiMedia, USA) for 5 minutes and observing under Zeiss Axio Imager M2 microscope (Zeiss, Germany). For all photos the following exposure settings were set: 200 ms for DAPI channel and 1.2 ms for bright field channel. Explants killed by incubation in EtOH for 1 hour were taken as reference of dead tissue.

Statistics and data analysis

Intensity quantification was carried out using ImageJ software. The background intensity was subtracted from every measurement. For Claudin 7 distribution, at least 30 independent measurements were made for each of three animals per group. Claudin 7 distribution data and organoid permeability assay were analyzed using Student’s t-test. For FRAP data, a two-way repeated-measures ANOVA with post-hoc Tukey test were applied.

Results & discussion

Among in vitro and ex vivo models, organoids provide one of the most convenient objects for studying the properties of epithelial tissue and selecting personalized treatments for a number of diseases,^35–38 Organoids serve as an ideal model for screening the effects of drugs on the epithelial barrier using a 4 kDa FITC-dextran (FD4) permeability assay. In vivo permeability tests require over 20 animals per assay.^26,28 Organoid technology significantly reduces animal use. A mere 2–3 animals can provide sufficient material to populate over a 100 organoid wells, with each well serving as an individual technical replica.

To our experience, organoids derived from large crypts or cultured long enough to form crypt-like extensions are often very fragile, fracture while transferred from matrix, and exhibit elevated permeability. In this regard, “young” organoids derived from 1–3 cells prove to be more convenient. They require more time (about a week) to form luminal 3D-structures and bear fewer shedding cells. According to our data and published protocols, the optimal age of organoids for the permeability assay appears to be 7–10 days post-seeding.^39,40 A perfect organoid must be 50–200 µm in diameter, spherical in shape with no significant outgrowths, and free of dead, membrane-permeable cells in the epithelial monolayer. The amount of such organoids can be greatly increased by breaking and passaging large branching colonoids, which close into perfect spheres 3 days after seeding. Dead cells within organoids can be identified using 7-aminoactinomycin D (7-AAD), which is typically excluded from living cells (Figure 1(a)). While propidium iodide is commonly used for this purpose,^41–44 our study suggests that 7-AAD can serve as an alternative and is also suitable for fixed organoids.⁴⁵ EDTA-treated organoids are commonly used as a negative control in permeability assay.³¹ Some researchers prefer to analyze only those organoids that exhibit dye uptake following EDTA treatment, which is also a reliable approach.³⁹

Crypts typically form organoids with apical cell membranes facing the lumen with rare exceptions. There are protocols that allow remodeling organoids into the apical-out conformation.^31,46 Notably, FITC-dextran enters paracellular space but not the lumen, due to the barrier formed by the tight junctions (TJs) at the apical region of the lateral membranes (Figure 1b). In case of necessity to have access to both apical and basal sides of the epithelium 2D organoid monolayers can come in handy.⁴⁷ Permeability assay can be conducted on monolayers as assessing of FD4 translocation from apical to basolateral compartment.⁴⁸ Transepithelial electrical resistance may provide information of epithelial integrity, but it is most reflective for ion permeability than for large molecules diffusion.^49,50 To our experience, the basal-out organoids are also suitable for permeability assays. As shown, inhibitor of phospholipase C (U-73122) induces intestinal barrier disruption and FITC-dextran entering the luminal space, while phosphatidylinositol (PI) has no significant effect (Figure 1(c-d)). To summarize, establishing reliable criteria to select organoids for permeability assays allows excluding false-positive results originating from degenerated organoids. We propose the following formal selection criteria: 7-day-old spherical organoids derived from single stem cells 50–200 µm in diameter, composed entirely of living cells (7-AAD-negative). Imaging conditions are also key to successful permeability assay: using a thermostatic CO₂ chamber for microscopy significantly improves the outcome since organoids left on the standard imaging platform for too long lose epithelial barrier integrity.^31,39 In case a stage with CO₂-incubation is not available for the microscope, immediate imaging upon sample preparation can be considered.

There is evidence demonstrating the phenotypic consistency of organoids obtained both from animal models and from IBD and colon cancer patients,^51–54 However, this data is controversial, and some researchers demonstrate opposite findings,^17–19 In this study, we derived organoids from Muc2-knockout mice, a well-characterized chronic IBD model, to determine whether they retain an inflammatory phenotype similar to the in vivo model.²⁶ We have previously shown that F-actin structure and Claudin 3 localization are perturbed in the descending colon of Muc2-knockout mice.²⁶ Our latest findings show that Claudin 7 localization in the intestinal epithelium is also disrupted in this mouse strain (Figure 1(e-f)). Here we show that under cell culture conditions, Muc2-derived organoids do not retain the characteristic delocalization of Claudin 7 from the lateral membranes found in native tissue of the mutant mice (Figure 1(e-f)). It appears that the local inflammatory environment is crucial for epithelial cell structure in the Muc2-knockout model. Similar observations have been made for IBD-derived 3D-epithelial cultures.⁵⁵ Organoids from IBD patients have been shown to lose the inflammatory phenotype in culture over time in the absence of inflammatory stimuli.^17,19 A number of studies indicate that they are morphologically identical to controls and have similar distribution of TJ proteins.^56,57 The viability of organoids also strongly depends on the severity of the disease in patient tissues.⁵⁸ Consequently, organoids may not always be suitable for barrier research upon IBD.

Alternatively, we cultured isolated colonic crypts for 3 hours to evaluate the stability of their cytoarchitecture. The absence of cell-matrix interactions typically induces programmed cell death in enterocytes,^59,60 so we used a solubilized basement matrix membrane commonly used in organoid culture and used previously to evaluate mitochondrial respiration in isolated crypts,^26–28,61 Short-term culture resulted in a loss of cell polarization defined by the rearrangement of cytoskeleton and barrier structures (Figure 2(a)) the brush border disappeared, F-actin was redistributed to the cortical cytoskeleton, Claudin 7 scarcely retained its peripheral localization. Cell viability assay demonstrated that numerous cells within crypts did not survive the procedure (Figure 2(b)). Earlier studies involving culture of human crypts from large intestine for 24–72 hours reported similar findings, with epithelial cells displaying peripheral F-actin distribution, but no clear brush border was found.²⁴ TEM analysis of the colonocyte monolayer showed remnants of microvilli, intact cytoplasmic organelles, and junctional complexes.²⁴ However, recent studies suggest that crypts or isolated epithelial cells should not be cultured for more than 1 hour to preserve cytomorphology,^62–64 Crypts cultured for longer periods tend to lose their original barrier properties. Therefore, isolated crypts are not advisable for studying F-actin and TJs’ functionality.

Figure 2.

Cultured colonic crypts. (a) Loss of enterocytes polarization in isolated crypts cultured for 1-3 hours on a basement matrix membrane. Red – F-actin, green - Claudin 7, blue – nuclei, bar −50 µm. (b) Representative photo of cultured crypts, dead cells stained with DAPI. Bar −100 µm.

There are numerous ex vivo culture protocols for intestinal explants enabling researchers to maintain tissue viability ranging from 12–24 hours to 5–14 days²⁰. The most sophisticated approach utilizes specialized bioreactor to sustain explants’ viability for at least 28 days.⁶⁵ We utilized the simplest culturing conditions that have been successfully used previously.^32,66–70 We collected colon tissue from an 8–12-week-old wild-type mice and cultured these explants for the time points of 0, 1, 3, 6, 18, 24, and 30. After that, the explants were fixed and stained for F-actin and Claudin 7 as a marker of TJs (Figure 3(a)). At all time points, epithelial cells appeared to have a morphologically regular shape and only rare dead cells were observed as compared to the ethanol treated cells considered positive for cell death (Figure 3(b)). At the time points 1, 3, 6, and 18 hours the structure of F-actin and TJs was similar to that found in living tissue (time point 0): actin and Claudin 7 remained their peripheral and membrane-bound localization. At the 24-hour time point most tissue retained regular staining for F-actin ad Claudin 7, but some areas of structural changes were already seen, while by 30-hour time-point the loss of regular actin and TJ staining was evident (Figure 3(b)). This finding suggests that cultured intestinal explants are not suitable for barrier studies in long-term experiments, while could be considered as an option in a short-term setting. Previous research supports these findings: after a 24-hours long culture, biochemical studies revealed a progressive loss of brush border markers such as alkaline phosphatase, alpha-glucosidase, and leucil-beta-naphthylamide from the intestinal tissue.⁷¹ After 48 hours, the explant’s morphology deteriorated rapidly, with microvilli shortening and blunting and enterocyte damage.^71,72 Transmission electron microscopy (TEM) data indicate morphological changes after 24 hours, such as the relocation of enterocyte nuclei from basal to apical cell compartment and shortened microvilli.^71,73 There is also evidence of explants maintaining their morphology for several weeks with a slight change in histological appearance,⁶⁷ but these data are not supported by recent studies.

Figure 3.

Cultured colonic explants. (a) Cultured colonic explants exhibit destruction of actin filaments and TJ after 30 hours. Red – F-actin, green - Claudin 7, blue – nuclei, bar −50 µm. (b) Representative photo of cultured explants, dead cells stained with DAPI. Bar −100 µm.

Many researchers claim that explants obtained from embryos are more tolerant to culturing conditions.⁶⁷ The smaller size of fetal tissues, their resistance to anoxia, and the fact that they are in the process of morphogenesis compared to the homeostatic mature state of adult make embryonic intestines more suitable to cell culture. Although dissecting embryonic tissues is technically more challenging, they seem to offer better prospects for maintaining accurate cytomorphology and barrier function for long-term.^74,75

One of the possible applications of a short-term explant culture is microscopic analysis of living tissue, which is usually not possible in live animals. As F-actin is a dynamic molecule that impacts barrier integrity, live-cell imaging of F-actin in the native epithelial environment might be of interest. Commercial dyes such as SiR-actin (Spirochrome) and CellMask™ Green (Invitrogen) enable the tracking of polymerized actin in living cells for 24 hours or more, so we applied them to explants in a short-term experiment. Confocal microscopy in an inverted setting raised technical challenges such as tissue crumpling and poor adherence to the coverslip. To straighten the tissue, we devised a “tube-gut sandwich” by placing a piece of opened colonic tissue between two plastic frames made out of a 0.5-ml microcentrifuge tube (Figure 4(a)). This “tube-gut sandwich” can be cultured in 48-well plates under desirable conditions, stained, and observed under an inverted microscope for about 1 hour. With SiR-actin or CellMask™ Green stains, the brush border and lateral actin can be perfectly seen from the lumen (Figure 4(b)). If necessary, the “tube-gut sandwich” can be fixed with paraformaldehyde and stained after live-cell imaging (Figure 4(c)). The “tube-gut sandwich” allows for a from-the-lumen view at the intestinal surface. This perspective is ideal for imaging junction-specific proteins such as ZO-1. It’s important to note that even immobilized explants may exhibit slight muscle contractions, potentially causing artifacts during image capture. If these artifacts significantly impact the experiment, adding a spasmolytic agent into the medium may be beneficial.

We further used this technique to conduct a time-lapse FRAP assay to measure F-actin dynamics. In FRAP methodology, fluorescence recovery is measured in time lapses at a given location in a cell and the speed of recovery is considered as a protein turnover rate as described.³³ We traced actin dynamics using CellMask™ Green Actin Tracking Stain specific to F-actin. Thus, actin dynamics might be assessed by comparing actin turnover rates in different conditions.^76,77 FRAP was performed in the brush border region where F-actin signal is the strongest. We observed about 25% fluorescence recovery after 320 seconds post-bleaching in the intact intestinal explants (Figure 5(a-b)). Actin filaments show a relatively low level of fluorescence recovery due to slow turnover of cytoskeleton monomers in contrast to highly-mobile molecules.^78,79 Importantly, the dynamics of actin recovery in explants strongly resembles that found in published reports. For example, FRAP analysis of SiR-Jasp – stained human red blood cells demonstrate ∼25–30% recovery after 12 minutes post-bleaching.⁷⁷ Fluorescence recovery of GFP-Actin in stress fibers of chondrocytes is about 40% after 582 seconds.⁷⁸ FRAP assay conducted in C2C12 myocytes demonstrated ~30% recovery in 540 seconds.⁷⁶ Thus, in different FRAP assays, mobile actin comprises about a third of all actin, but its recovery rate varies depending on cells type, dye, conditions and equipment applied. To further valuate the validity of the FRAP procedure when applied to explants, we tested the effect of an F-actin polymerizing drug jasplakinolide on F-actin dynamics.⁸⁰ The addition of low-dose jasplakinolide significantly increased F-actin recovery: analysis of variance yielded significant effect of the group, F(65, 1089) = 11.9, p < 0.001, and a post hoc Tukey test showed that at 10 s and further on the samples differed significantly (p < 0.01 between groups, Figure 5(a-b)). Thus, F-actin retains physiological response to the drug within the explant environment.

In the conclusion it should be questioned if our considerations are applicable for human tissues. Numerous studies explored culturing human explants, crypts, and organoids obtained from patient colonoscopy samples.^{24,31,38,59,66,67,70} According to published data, human tissue is different from mouse models in at least following aspects: endoscopic biopsies consist of only mucosa with some sub-mucosa, human colonoscopy sample size is relatively small, culture conditions should be adjusted for human tissue.^67,81,82 While permeability assay can be conducted similarly for organoids of either species³¹ possible limitation is the need to perform living cell imaging immediately after biopsy collection to ensure cytoskeleton and TJs integrity. Another limitation is the small sample size, which is important to collect enough data for live-imaging. Otherwise the described above approaches and considerations apply to both animal models and human patients.

In this study, we highlighted and discussed several limitations of such models as isolated crypts, intestinal explants, and 3D-organoids for use in TJ biology and epithelial barrier integrity. According to our data, intestinal crypts and explants are poorly applicable in long-term experiments due to considerable tissue rearrangements and redistribution of TJs proteins under culturing conditions. Still, living intestinal explants can be used in short-term assessment of actin dynamics providing insight to mechanistic basis of TJs assembly. Evaluation of intestinal barrier function in various conditions can be performed in colonic organoids derived from small groups of stem epithelial cells. We hope that our technical tips would be helpful for those performing research in the field epithelial biology. Utilizing various methodologies will deepen our understanding of the intestinal barrier complexity and contribute to the development of novel therapeutic strategies aiming at intestinal health.

Funding Statement

This research was funded by the Russian Science Foundation, [grant number 20-74-10022-П].

Disclosure statement

No potential conflict of interest was reported by the author(s).