Basic Three-Dimensional (3D) Intestinal Model System with an Immune Component

Summary

Here we describe constructing a basic three-dimensional (3D) intestinal cell line model system and a paraffin embedding protocol for light microscopic evaluation of fixed intestinal equivalents. Staining of selected proteins permits the analysis of multiple visual parameters from a single experiment for potential use in preclinical drug screening studies.

Abstract

There has been an increase in the use of in vivo and in vitro intestinal models to study the pathophysiology of inflammatory intestinal diseases, for the pharmacological screening of potentially beneficial substances, and for toxicity studies on potentially harmful food components. Of relevance, there is a current demand for the development of cell-based in vitro models to substitute animal models. Here, a protocol for a basic, “healthy tissue” three-dimensional (3D) intestinal equivalent model using cell lines is presented with the dual benefit of providing both experimental simplicity (standardized and easily repeatable system) and physiological complexity (Caco-2 enterocytes with a supporting immune component of U937 monocytes and L929 fibroblasts). The protocol also includes paraffin embedding for light microscopic evaluation of fixed intestinal equivalents, thereby providing the advantage of analyzing multiple visual parameters from a single experiment. Hematoxylin and eosin (H&E) stained sections showing the Caco-2 columnar cells forming a tight and regular monolayer in control treatments are used to verify the efficacy of the model as an experimental system. Using gluten as a pro-inflammatory food component, parameters analyzed from sections include reduced monolayer thickness, as well as disruption and detachment from the underlying matrix (H&E), decreased tight junction protein expression as shown from occludin staining (quantifiable statistically), and immune-activation of migrating U937 cells as evidenced from the cluster of differentiation 14 (CD14) staining and CD11b-related differentiation into macrophages. As shown by using lipopolysaccharide to simulate intestinal inflammation, additional parameters that can be measured are increased mucus staining and cytokine expression (such as midkine) that can be extracted from the medium prior to fixation. The basic three-dimensional (3D) intestinal mucosa model and fixed sections can be recommended for inflammatory status and barrier integrity studies with the possibility of analyzing multiple visual quantifiable parameters.

Introduction

The intestinal epithelial barrier, a one-cell-thick internal lining containing different types of epithelial cells, constitutes the first physical defensive barrier or interface between the outside and the internal milieu of the body^1,2. Columnar-type enterocytes constitute the most abundant type of epithelial cells. These are responsible for maintaining epithelial barrier integrity through interactions between several barrier components, including tight junctions (TJs), playing a significant role in barrier tightening^1,3. The TJ structure consists of intracellular plaque proteins, such as zonula occludens (ZO) and cingulin, cooperating with transmembrane proteins, including occludins, claudins, and junctional adhesion molecules (JAMs) that form zipper-like structure tightly linking the neighboring cells^3,4. The transmembrane proteins regulate the passive paracellular diffusion of small compounds and exclude toxic large molecules.

Potentially toxic food compounds and food contaminants stimulate inflammatory cytokine production that disrupts the epithelial permeability, activating immune cells and causing chronic intestinal tissue inflammation^5,6,7. In contrast, various anti-oxidant and anti-inflammatory phytochemicals have been reported to reduce inflammatory cytokine expression and enhance intestinal TJ barrier integrity through the restoration of TJ protein expression and assembly^4,6,8. Hence, the regulation of epithelial barrier integrity by both beneficial and harmful compounds has seen an increase in the use of both in vivo and in vitro models aimed at mimicking the intestinal barrier for pharmaceutical screening and toxicity studies. This is particularly relevant given the increasing interest in understanding the pathophysiology of intestinal bowel diseases (IBD), necrotizing enterocolitis, and cancer, which can be simulated in experimental models^8,9,10.

There has been a demand for the development of cell-based in vitro models in order to achieve the objective of the “3Rs” in animal testing. These include replacement alternatives to the use of animals, reduction in the number of animals used, and refinement in adopting methods that alleviate distress^11,12,13. Moreover, the underlying molecular, cellular, and physiological mechanisms between human and murine models (rodents being the most widely used species) are distinctive, leading to controversy regarding the efficacy of the murine models as predictors in human responses^12,13. Numerous advantages of in vitro human cell-line models include target-restricted experimentation, direct observation, and continuous analysis¹³.

Single-cell-type monolayers in two-dimensional (2D) cultures have served as powerful models. However, these cannot precisely reproduce the physiological complexity of human tissues^8,13,14. As a result, 3D culture systems are being developed with ever-increasing improvements to recapitulate the physiological complexity of both healthy and diseased intestinal tissues as next-generation risk assessment toolboxes^13,14. These models include 3D Transwell scaffolds with diverse cell lines, organoid cultures, and microfluidic devices (intestine-on-chip) using both cell lines and organoids (derived from both healthy and diseased tissues)^8,13,14. 

The 3D “healthy tissue” intestinal equivalent protocol presented in the present study was based on striking a balance between physiological complexity and experimental simplicity¹³. The model is representative of a 3D Transwell scaffold, comprised of three cell lines (enterocytes [the gold-standard colon adenocarcinoma Caco-2 line] with a supporting immune component [U937 monocytes and L929 fibroblasts]), constituting a standardized and easily repeatable system applicable for the preliminary screening of dietary molecules of interest on intestinal epithelial barrier integrity and immune response. The protocol includes paraffin embedding for light microscopic evaluation of epithelial barrier integrity using fixed intestinal equivalents. The advantage of the present approach is that numerous sections of the embedded tissues can be made to stain for multiple parameters from a single experiment.

Protocol

1. Preparation of the basic 3D reconstructed intestinal mucosa model

NOTE: The entire procedure must be carried out in a sterile laminar flow hood. All steps in the procedure involving the use of the cell incubator signify that cultures are incubated at 37 °C in a humidified atmosphere containing 5% CO₂ (unless stated otherwise in the protocol)_.

1. Prior preparation of the cell lines used in the intestinal model system
   1. Seed L929 mouse fibroblast cells at a concentration of 5 x 10⁵ in 5 mL of Dulbecco's Modified Eagle Medium (DMEM) containing 2 mM L-glutamine, 10% Fetal Bovine Serum (FBS), and 1% Penicillin-Streptomycin (Pen-Strep) in a F25 flask and culture it in an incubator 4 days before the construction of the intestinal models. After 48 h, remove the medium using a pipette. Then add fresh medium (5 mL), and incubate the cells further for 48 h.
      NOTE: The cells must be 80% confluent before being used.
   2. Seed Caco-2 cells (concentration of 2 x 10⁶) in 10 mL of DMEM medium (with 10% FBS and 1% Pen-Strep) in a F75 cell flask and culture it in an incubator 4 days before the construction of the intestinal mucosa model. After 48 h, change the medium as described in step 1.1.1 and further incubate for 48 h to a confluency of 80%.
      NOTE: The cells must be in an active proliferating phase: not too sparse nor too confluent. A 50-60% confluence is recommended. The cells must not be sown the day before making the model because this could slow down the proliferative capacity of the cells, which would then not take root perfectly in the reconstructed 3D model.
   3. Add U937 cells that grow in suspension (concentration of 1 x 10⁶) to 10 mL of Roswell Park Memorial Institute (RPMI) medium (containing 2 mM L-glutamine, 1% sodium pyruvate, 10% FBS, and 1% Pen-Strep) in a F75 cell flask 2 days prior to the model set-up, and place in an incubator for 48 h.
2. Preparation of the Transwell co-culture plate inserts
   1. Select a 24-well plate containing inserts with 0.4 µm filters.
      NOTE: The 0.4 µm filter is a standard choice in drug transport studies. The 3 µm and 8 µm filter sizes are not recommended to prevent any possible losses of the collagen-embedded cells.
   2. Using a pipette, hydrate the Transwell filters (henceforth referred to as membrane inserts) with 500 µL of Hanks' Balanced Salt Solution (HBSS) below and above the filter insert.
   3. Close the multi-well plate and place it in an incubator for a minimum of 2 h.
      NOTE: The membrane inserts can remain in the HBSS for up to 24 h. This operation can be performed the day before constructing the model. It is important that the membrane inserts are completely hydrated and that the filter does not dry out, as this could make it harder for the sample to adhere properly.
   4. Remove the plates from the incubator after 2 h (or 24 h). Carefully aspirate the HBSS from above and below the membrane inserts using a pipette and leave it to dry for 10 min.
3. Preparation of the cell-free collagen lamina propria of the basic 3D intestinal model system (DAY 1)
   1. Prepare a cell-free collagen solution in a 50 mL sterile tube on ice containing the following components in DMEM: 10% Fetal Bovine Serum (FBS), 200 mM L-glutamine, 1% sodium bicarbonate and 1.35 mg/mL Type 1 rat tail collagen.
      NOTE: All these components must be kept on ice and added to the DMEM with cooled pipette tips. The Type 1 rat tail collagen must be added last as this polymerizes with increasing temperature and pH. The amount of cell-free collagen solution prepared will depend on the number of intestinal equivalents required.
   2. Add 250 µL of the solution to each plate insert (above the membrane filter) for the number of intestinal equivalents selected and place the lid over the multi-well plate. Allow the collagen solution to polymerize at room temperature (RT) under the laminar flow hood.
      NOTE: Apart from the transition to a more solid phase, polymerization is also evident from a change in color from yellow to pink. Polymerization is usually complete within 10-15 min.
4. Preparation, cell counting, and addition of fibroblast (L929) and monocyte (U937) cells to the model system (DAY 1)
   1. Remove the L929 cell culture (step 1.1.1) from the incubator. Using a vacuum pump, aspirate the medium, replace it with 5 mL of sterile phosphate buffer saline (PBS; without Ca and Mg), and rinse the cells.
   2. Aspirate the PBS using a vacuum pump. Add 2 mL of a pre-prepared trypsin-EDTA solution (0.05% trypsin and 0.02% EDTA in PBS) and place in an incubator for 3-5 min.
   3. Use an inverted microscope (for example, an Eclipse Ts2, Nikon) to establish whether the cells are detaching from the adhesion surface. If this is occurring, immediately add 2 mL of DMEM (containing 10% FBS) to block the trypsin reaction and rinse the cells.
   4. Transfer the cell solution to a sterile 15 mL tube and centrifuge at 645 x g for 5 min. Using a vacuum pump, aspirate the supernatant.
      NOTE: Care is needed not to disturb the pellet. The effect of DMEM was tested on the L929 and U937 cells and was not shown to have adverse effects on growth.
   5. Add 1 mL of DMEM to the pellet and suspend the cells.
      NOTE: The cells must be homogeneously suspended in the solution.
   6. Remove 20 µL of the cells in DMEM and add 20 µL of Trypan blue solution. Remove 20 µL of the mix and evaluate cell density microscopically using a cell counting chamber.
   7. Remove the U937 cell culture (step 1.1.3) from the incubator. Centrifuge the cell solution at 645 x g for 5 min. Using a vacuum pump, aspirate the supernatant to avoid disturbing the pellet.
   8. Suspend the cells in 1 mL of RPMI. As with the L929 cells, make sure the cells are homogeneously suspended in the solution.
   9. Similarly, establish the cell density as reported in step 1.4.6.
   10. Prepare a collagen solution as described in step 1.3.1.
       NOTE: The amount of solution to be made must take into consideration 450 µL for each insert (or model).
   11. Prepare a solution of DMEM to contain a count of 50,000 L929 cells and 15,000 U937 cells, respectively, in a volume of 50 µL for each intestinal equivalent model to be constructed.
       NOTE: The number of cells is a critical factor. Too many of both cell types would result in a lamina propria excessively full of cells that would not be adequately organized. An inferior number of fibroblasts (which generate collagen) would result in a less compact lamina propria, whereas too few monocytes would impede the immune response to stimuli. Provided the correct count of cells is present within each 50 µL aliquot - a total volume of 600 µL can be prepared for the 12 filter inserts present in each 24 well-plate.
   12. Add each 50 µL aliquot containing the cells to 450 µL of collagen solution in step 1.4.10. Mix thoroughly.
   13. Overlay the pre-coated cell-free collagen lamina propria with 500 µL of the collagen-containing cell solution for each model.
       NOTE: It is important to rapidly add the volumes to each insert, and as such, it is advisable to limit the number of reconstructed intestinal mucosa models to 12 or fewer at a time.
   14. Close the plate and place it in an incubator for 2 h to allow the solution to set.
5. Preparation, cell counting, and addition of epithelial Caco-2 cells to the intestinal model (DAY 1)
   1. Remove the Caco-2 cell culture (step 1.1.2) from the incubator. Repeat the procedures from step 1.4.1 using 10 mL of PBS for preliminary rinsing. Then add 5 mL of a pre-prepared trypsin-EDTA solution (0.05% trypsin and 0.02% EDTA in Ca- and Mg-free PBS) and place in an incubator for 5-8 min.
   2. Repeat steps 1.4.3 (but using 5 mL of DMEM to block the trypsin reaction) through to 1.4.6 to count the cells using the cell counting chamber.
   3. Prepare a solution of DMEM to contain a count of 150,000 Caco-2 cells in 50 µL.
      NOTE: The number of cells used is a critical point. Exceeding 150,000 cells could result in a compact disorganized epithelial layer, whereas with too little (less than 100,000), the cells struggle to grow and do not adequately cover the basement membrane creating a discontinuous intestinal epithelial layer. Make sure the cells are effectively suspended. The tip of a 200 µL pipette can be used to ensure homogenous distribution. Given that 12 models can be constructed at any given time, a 600 µL cell solution can be prepared.
   4. After the 2 h required in step 1.4.14, add 50 µL of Caco-2 cells suspended in DMEM to the middle of each basement membrane. Close the lid of the multi-well plate.
   5. Incubate under the sterile laminar flow hood for 10 min. Then transfer to the incubator for 30 min.
   6. Prepare a DMEM solution containing 10% FBS and 1% Pen/Strep.
      NOTE: Prepare a sufficient solution to use a 1 mL volume per model.
   7. Add 500 µL of the medium above the reconstructed model and 500 µL below the filter.
      NOTE: Care is required when adding the medium above the filter to avoid detaching or stressing the cells.
   8. Close the multi-well plate system and place it in the incubator.
6. Model preparation/formation and use (DAY 2 to DAY 6)
   1. DAY 2: Carefully remove the solution from both above the reconstructed model and below the filter using a pipette. Replace with fresh 500 µL of fresh DMEM (10% FBS and 1% Pen/Strep) above and below the filter, respectively.
   2. DAY 3: Repeat as above in step 1.6.1.
   3. DAY 4: Repeat as above in step 1.6.1.
      NOTE: This intestinal model is a static cellular model; therefore, to favor the release of molecules and the growth and stimulation of cells, it is very important that the medium is changed every day.
   4. DAY 5: At this point, the model is fully formed/developed. Use these models for further studies.
      NOTE: The best time to use the model is at 5 days. Although the cells can be maintained in an incubator for longer periods, the more time that passes, the greater the probability that the epithelial cells may grow in an uncontrolled manner, resulting in an unorganized and compact layer that is difficult to use.
   5. At 5 days, incubate the models with either toxic (gluten or lipopolysaccharide [LPS]) or beneficial components (polyphenols) of interest. Add these to the upper portion of the reconstructed model suspended in the DMEM medium.
      NOTE: The suitable concentration of each component of interest must be calculated and suspended in a DMEM medium. Non-treated controls containing only DMEM medium must be set up for comparison to the experimental models.
   6. Incubate the control and experimental models for 24 h in an incubator.
   7. DAY 6: Remove the medium above and below the filter with a pipette.
      ​NOTE: The medium can be stored for subsequent enzyme-linked immunosorbent assay (ELISA)-type tests to measure the release of inflammatory cytokines. For this purpose, the medium must be added to sterile vials and stored at -20 ° C for further analysis.

2. Paraffin embedding of the reconstructed intestinal mucosa models

NOTE: The entire procedure must be carried out under a chemical fume hood. Each step and respective time allocation must be strictly adhered to. For this reason, it is important to have all reagents prepared ahead of time.

1. Set the paraffin machine to 58 °C so that it is ready for use.
2. Transfer the membrane inserts to clean wells using pliers in a sterile 24-well plate.
3. Add 500 µL of 37% buffered formalin in PBS above the filter and 1 mL under the filter. Close the lid and leave it under the chemical fume hood for 2 h at RT.
   NOTE: As an alternative, 4% buffered formalin can be used at RT for 1 h.
4. Remove the formalin and add HBSS solution both above and below the filter. Then remove the HBSS.
5. Detach the lamina propria from the membrane insert.
   NOTE: The cells of the intestinal mucosa usually detach very easily from the membrane insert as they are not attached to the latter. Additionally, the two sample compartments (apical and basal) remain attached, retaining the 3D structure. If, for some reason, they are not readily detached, it will be important to use a sterile disposable scalpel blade to remove the intestinal mucosa from the membrane insert. The membrane insert can be cut, thus detaching it from the plastic. Be careful never to touch the sample with the scalpel. The motive for removing the membrane insert from the collagen-embedded cells is that this filter may become detached in the subsequent fixation or paraffin embedding phases, rendering comparisons between intestinal mucosa models disproportional. Moreover, the membrane inserts within the embedded section have a different consistency which can interfere with sectioning.
6. Place the 100 mL beakers under the chemical fume hood, each correctly labeled to include one reconstructed intestinal mucosa model system under investigation. Add 25 mL of 35% ETOH to each beaker and then add the reconstructed intestinal mucosa. Incubate for 10 min.
7. Replace the 35% ethanol with 25 mL of 50% ethanol and incubate for 10 min.
8. Replace the 50% ethanol with 25 mL of 70% ethanol and incubate for 10 min.
9. Replace the 70% ethanol with 25 mL of 80% ethanol and incubate for 10 min.
10. Replace the 80% ethanol with 25 mL of 95% ethanol and incubate for 10 min.
11. Replace the 95% ethanol with 25 mL of 100% ethanol and incubate for 10 min.
12. Replace the 100% ethanol with another 25 mL change of 100% ethanol and incubate for 10 min.
13. Place 100 mL beakers under the fume hood, each correctly labeled to include one model under investigation. Add 50 mL of xylene or a histological clearing agent for 10-20 min.
    NOTE: Histo-Clear (histological clearing agent) is recommended as the agent enhances the clarity and vibrancy of acidophilic stains. The clearing time can vary from one reconstructed intestinal mucosa sample to another and must be checked for transparency in appearance every 2-3 min.
14. As soon as the samples are transparent, place them in a metal tissue cassette holder which is then immersed in liquid paraffin inside the heated machine for 45 min.
15. Change the paraffin and leave the samples inside the heated machine for a further 45 min.
16. Remove the tissue cassette holder and place on ice to cool. When the cooled sample blocks detach from the metal holder, these can be further cooled at room temperature.
17. Store the blocks at RT.
    NOTE: If the objective is to prepare sections immediately, then the blocks can be placed at 4 ° C so that they are well-cooled when used.
18. Cut 4 µm sections using a microtome.
19. Place the cut sections on slides and dry them in an oven at 37 ° C for 24 h.
20. The slides are ready for use. Store them at RT until performing either H&E histological staining or immune-histochemical reaction staining (antigen-antibody type).
    1. For immune-histochemical staining, select the antibodies of interest (either for the study of TJ proteins such as occludin or for monocyte activation, migration, and differentiation) and detect expression (via staining) using commercial kits.
    2. Similarly, measure mucus using the Alcian blue and periodic acid-Schiff (PAS) staining kit.
    3. Quantify the stained TJ proteins of interest, such as occludin, by calculating the percentage of positive pixels on micrographs taken from the microscope.
    4. Analyze pictures of the cells using image analysis software (e.g., ImageJ2 software [Wayne Rasband, version 2.9.0/1.53t]).
       1. To perform the analysis of the pixels, process the digital images to 300 pixels/inch and convert them to 8 bits. Then, process the binary images by Color Deconvolution plugin to analyze the staining of the protein of interest, in this case, the permanent red staining of occludin.
       2. Save the selected picture as a tiff and subject it to a "clean-up" procedure to eliminate artifacts with a graphics editor (e.g., Adobe PhotoshopCC [version 20.0.4]).
       3. Thereafter, measure all fields of interest with the application Analyze Particle of ImageJ2, and report the data as the number of pixels.
          NOTE: Each experiment should be performed in triplicate with three internal replicate fields analyzed for each replicate. For quantifying mucus, the same principle of measuring the pixels can be applied to the images taken of the purple-magenta-stained neutral mucins and bright blue-stained acidic mucus, respectively.

Representative Results

The first important aspect is to determine the acceptability of the basic 3D intestinal equivalent mucosa for experimental purposes. This is performed with the most widely used stain in histology and histopathology laboratories, namely hematoxylin (stains nuclear material deep blue-purple color) and eosin (stains cytoplasmic material varying shades of pink). The H&E staining is first performed on an untreated control, which is cultured under the same conditions and timeframe as the experimental treatments. From visualizing the pattern, shape, and structure of cells, the success of this intestinal cell model system is determined by the Caco-2 cells forming a tight and regular monolayer above the extracellular matrix (ECM)-rich lamina propria after 5 days (Figure 1A). Examples of model systems that would be deemed unacceptable for further analyses include those showing excessive growth with disorganized epithelial layers, as shown after 10 days (Figure 1B). An additional example is an epithelial layer that is too thick and disorganized (Figure 1C), caused either by the seeding of excess Caco-2 cells or an excessive incubation. Malformation (Figure 1D), attributable either to seeding cells that were not in an active phase of proliferation, or problems in the preparation of the lamina propria or during fixation with paraffin (cells became detached /ruined), is also an example of an unacceptable reconstructed intestinal mucosa. Similarly unacceptable is the lack of epithelial layer formation (Figure 1E), indicative of the seeding of too few Caco-2 cells or problems in the preparation of the lamina propria. Using acceptable 3D intestinal equivalents, from which multiple sections can be generated, various parameters can then be investigated from a single experiment.

Of the various parameters that can be analyzed, barrier integrity in response to pro-inflammatory inducers can be visualized, and TJ proteins stained and quantified and compared to untreated control model systems. Here we demonstrate the pro-inflammatory effect of 40 µg/mL total wheat protein from a durum wheat source with a high gluten content⁷, on both cellular integrity as well as TJ occludin protein content (Figure 2A). Using H&E histological staining, the control model system shows the Caco-2 columnar cells forming a tight and regular monolayer. In contrast, the effect of the gluten-containing protein sample was shown to cause disruption of the monolayer, with more nonspecific eosin staining (Figure 2A). Moreover, the thickness of the epithelial layer can be measured and is shown to be significantly reduced after exposure to pro-inflammatory gluten (Figure 2A,B). Using an antibody to occludin (conjugated to red chromogen staining), occludin protein could be visualized and quantified. Significantly higher occludin protein content was evident in the control group compared to the intestinal mucosa exposed to gluten protein for 24 h (Figure 2A,C).

The intestinal equivalents comprised of an immune component (represented by U937 monocytes) are also ideal in investigating the effects of pro-inflammatory inducers on monocyte activation, migration, and differentiation into macrophages. Once again, using 40 µg/mL total wheat protein extracted from a wheat source with a high gluten content, pro-inflammatory activation of U937 monocytes in the lamina propria was evident from fast red coupled CD14 staining (Figure 3A). Activation is detected from the increased expression of this red-stained membrane glycoprotein on monocytes and macrophages. Increased CD14 protein was also detected with fluorescein isothiocyanate (FITC), which stains green under immunofluorescence. Migration was evidenced from the presence of activated U937 cells close to the Caco-2 monolayer. Where the monolayer had ruptured, the U937 cells were shown to have entered within the Caco-2 cell monolayer (Figure 3A). There was no evidence of CD14-stained monocyte migration in the control group (Figure 3A). The differentiation of monocytes into macrophages is identifiable from the marker CD11b. No CD11b staining was evident in the control group with either red chromogen or FITC (Figure 3B). Instead, in the tissues exposed to gluten, CD11b macrophages were observed both in the ECM-rich lamina propria and within the ruptured Caco-2 monolayer (Figure 3B).

Using LPS to simulate intestinal inflammation, here we show that it is also possible to use this model system to investigate acidic mucin and mucopolysaccharides produced by the Caco-2 cells in the untreated control and the LPS treatment, respectively (Figure 4A). Using Alcian blue and PAS staining, acidic mucus is stained bright blue, and neutral mucins are stained purple-magenta, respectively. The extent of mucus staining can similarly be quantified and is shown to be significantly induced under 1 ng/mL LPS challenge (Figure 4B). Moreover, prior to fixation, the medium can be removed to investigate pro-inflammatory cytokine production. Although numerous excretory cytokines can be measured, in this particular instance, midkine (MDK) was selected. MDK is an endogenous inflammatory marker induced by the transcription factor, nuclear factor kappa-light-chain-enhancer of activated B cell (NF-κB) pathway, and is also associated with inflammatory diseases of intestinal cells¹⁵. Under the LPS challenge, MDK is significantly induced in comparison to the control (Figure 5).

Figure 1: Untreated control 3D intestinal equivalents. Hematoxylin and eosin (H&E) staining of untreated control 3D intestinal equivalents (Caco-2 /U937/L929 co-cultures) for 24 h as a verification of the efficacy of the basic 3D intestinal mucosa model for experimental purposes. (A) An acceptable experimental model is compared to unacceptable models, showing excessive (B) epithelial growth and disorganized layers after 10 days, (C) excessive epithelial thickness and disorganization, (D) malformation, and (E) no epithelial cell layer. The scale bar in each image equals 50 µm. Please click here to view a larger version of this figure.

Figure 2: Treated 3D intestinal equivalents. (A) Hematoxylin and eosin (H&E) staining, and occludin expression of 3D intestinal equivalents (Caco-2 /U937/L929 co-cultures) exposed for 24 h to 40 µg/mL total protein from a high gluten containing modern wheat mix compared to the control (CTRL). The scale bar in each image equals 50 µm. (B) Quantification of Caco-2 cell thickness and (C) occludin-red chromogen staining. Significant differences were determined by one-way variance (ANOVA) and represented as *** 0.0001< P < 0.001. The black dots represent the number of replicates. This figure has been adapted with permission from Truzzi et al.⁷. Please click here to view a larger version of this figure.

Figure 3: CD14 and CD11B staining. CD14 staining of U937 monocytes and CD11B U937 differentiated macrophages in 3D intestinal equivalents (Caco-2/U937/L929 co-cultures) exposed to no added protein (control [CTRL]) or 40 µg/mL total protein from a high gluten containing modern wheat mix after a period of 24 h. For CD14 and CD11b staining, fast red was used as a chromogen and nuclei were counterstained with eosin. The scale bar in each image equals 20 µm. CD14 and CD11b staining with fluorescein isothiocyanate (green) was also shown under immunofluorescence. The nuclei were counterstained with 4',6-diamidino-2-phenylindole (DAPI). The scale bar in each image equals 50 µm. Arrows indicate the presence of monocytes (CD14 positive cells) and macrophages (CD11b positive cells). This figure has been adapted with permission from Truzzi et al.⁷. Please click here to view a larger version of this figure.

Figure 4: Hematoxylin and eosin (H&E) staining, and mucus expression from Alcian Blue/Periodic Acid-Schiff Stain (PAS) staining. (A) Caco-2 cells in basic 3D intestinal equivalents (Caco-2 /U937/L929 co-cultures) after 24 h in the untreated equivalent and reconstructed intestinal mucosa model systems exposed to 1 ng/mL lipopolysaccharide. The scale bar in each image equals 50 µm. (B) Quantification of mucus expression was determined by one-way variance (ANOVA) with significance reported as **** P < 0.0001. The black dots represent the number of replicates. This figure has been adapted with permission from Truzzi et al.¹⁶. Please click here to view a larger version of this figure.

Figure 5: Midkine expression. Midkine expression in the medium of 3D intestinal equivalents (Caco-2 /U937/L929 co-cultures) after 24 h in the untreated control (CTRL) and intestinal mucosa model systems exposed to 1ng/mL lipopolysaccharide. Quantification of midkine expression was determined by one-way variance (ANOVA) with significance reported as **** P < 0.0001. This figure has been adapted with permission from Truzzi et al.¹⁶. Please click here to view a larger version of this figure.

Figure 6: Schematic diagram representing the construction and experimental use of intestinal equivalents, as well as the fixation and microscopic analyses of parameters from tissue sections. The schematic diagram sums up the construction of the intestinal equivalents and possible experimental treatments that can be administered to these models. The generation of multiple tissue sections from a single experiment permits the analyses of numerous structural and immune aspects of the intestinal mucosa, as illustrated in Figure 1, Figure 2, Figure 3, and Figure 4. Please click here to view a larger version of this figure.

Discussion

The basic reconstructed intestinal mucosa model system presented here (Figure 6) combines physiological complexity (more physiologically relevant 3D cell cultures containing a Caco-2 monolayer with an ECM-rich lamina propria support containing fibroblasts and monocytes) with experimental simplicity (using commercial human cell lines to produce standardized and easily repeatable system)¹³. As such, this model system is deemed a suitable alternative to murine models aimed at evaluating the effect of potential medicinal drugs/food components on intestinal barrier integrity. In this regard, previous articles substantiate the efficacy of the present cell culture model in testing both potentially beneficial compounds (plus dose dependency) as well as harmful food components^7,16,17. Moreover, LPS or alternative pro-inflammatory reagents (2,4,6-trinitrobenzene sulfonic acid or dextran sulfate sodium) can be added to the present intestinal mucosa model to simulate IBD symptoms. In so doing, it is potentially possible to study the etiology of intestinal diseases from a structural aspect as well as from the expression of cytokines, such as pro-inflammatory MDK, which is a marker of IBD¹⁵. This basic model system also has eventual drug screening potential for the broader population. Although the model system is physiologically more complex than 2D models^13,14, the limitations in reproducing complex human physiology include the absence of both a microfluidic system (fluidic perfusion of a continuous supply of nutrients and removal of waste products as well as mechanical stimuli)^8,14 and the microbiome (evaluating host-microbiome interactions in intestinal diseases)^8,14.

The advantage of using the present basic 3D intestinal mucosa model in combination with paraffin embedding for light microscopic evaluation of epithelial barrier integrity is that several tissue sections can made from a single experiment, thereby permitting the analysis of numerous parameters. Transepithelial electrical resistance (TEER)³ is a commonly used method to assess intestinal barrier integrity and has been used in conjunction with macromolecular permeability studies^3,18, as well as widely reported studies on Western blot evaluation of TJ protein expression.  Western blot evaluation provides quantitative insight into overall changes in protein expression, and TEER serves as a quantitative measure of barrier integrity. Instead, microscopic evaluation provides a valuable tool for the qualitative visualization and assessment of local protein changes and interactions¹⁸.

Microscopic evaluation of a model with an immune component permits the visualization of immune responses, such as the differentiation of monocytes into macrophages and the migration of activated monocytes in the ECM, as demonstrated here by CD14 and CD11b staining. Identification can be made via immunohistochemistry and/or immunofluorescence staining on individual tissue sections, as was shown for CD14 and CD11b in Figure 3. Immunohistochemistry permits the structural visualization of the reconstructed intestinal mucosa tissue as a whole. Moreover, it is possible to identify both the nuclei and the cytoplasm in cells. In contrast, confocal immunofluorescence does not allow the tissue to be seen as a whole (only the markers of interest and DAPI-stained nuclei). However, it provides a very clean image where the background color is eliminated. The advantage of this is that it permits greater precision in evidencing the positivity or negativity of the markers and as such, the calculation of any differences in color intensity (for example, a marker that is more expressed in one cell than in another). Potentially the extent of migration of the CD14 monocytes from within the lamina propria to the Caco-2 monolayer could also be estimated from multiple replicate images taken at various time points. Nonetheless, evidence of inflammation was evident from the presence of both CD14 and CD11b stained cells.

A range of 20–40 μg total protein is used by the majority of laboratories for western blotting¹⁹. The quantification of numerous gap junction proteins (occludin, claudin, zonulin, and E-caderin) can be performed by immunohistochemical and/or fluorescent staining on individual tissue sections, respectively necessitating significantly less protein. Concurrently, from the same experiment, the integrity of the epithelial layer in response to compounds can also be visualized from the pattern, shape, and structure of cells with H&E staining.

The efficacy of a basic 3D reconstructed intestinal mucosa for use in staining mucus production, TJ proteins, and proteins produced from immune responses is dependent on overcoming critical steps, particularly in the construction of the model system and during tissue fixation.  Regarding model construction, using the correct confluence and seeding the correct number of Caco-2 cells (indicated by the notes in the protocol), respectively, is fundamental. Too many cells or an incorrect confluence would result in a thick and disorganized epithelial layer. Instead, too few cells would result in either a reduced or a lack of epithelial layer formation. Equally fundamental at seeding is the correct cell confluence and number of L929 and U937 cells in the ECM, as well as the correct handling of the collagen during preparation. Excess cells or rapid polymerization of the collagen would result in a compact and malformed ECM, in turn impacting on the structure of the epithelial cells above. Similarly, too few collagen-forming L929 cells would result in an incorrect consistency of the ECM, whereas too few U937 monocytes would not favor an immune response. Also important is that cells be used for experimentation after 5 days, as a prolonged incubation would result in excessive cell growth resulting in a compact epithelial layer. Moreover, given that the present design lacks a microfluidic system and is representative of a static intestinal cellular model, to favor the release of molecules and the growth and stimulation of cells, it is very important that the medium be changed every day. Regarding the fixation protocol, the clearing phase is critical and given that the time required for the tissues to become transparent in paraffin may vary, each intestinal mucosa system must be closely observed during this phase. Throughout the fixation procedure, care must be taken in handling these models in order to avoid damaging them.

In conclusion, provided strict adherence is paid to the critical points in the construction and fixation of the basic 3D intestinal equivalents, the models can provide numerous tissue sections for the evaluation of multiple parameters in pharmaceutical screening and toxicity studies.

Materials

Name	Company	Catalog Number	Comments
Alcian Blue/PAS kit	ScyTek Laboratories, Inc.	APS-1, APS-2	kit for immunohistochemical staining
Blue Trypan solution	Thermo Fisher	15250061	cell count analyses
Caco-2 colorectal adenocarcinoma cells	ATCC	ATCC HTB-37	cell line
Citro-Histo-Clear Limonene based	Histoline laboratories	R0050CITRO	reagent for paraffin embedding
Dulbecco's Modified Eagle Medium (DMEM)	GIBCO	11965092	cell colture reagent
Embedding Center	Histoline laboratories	TEC2900	instrument for paraffin embedding
Fetal Bovine Serum (FBS)	GIBCO	A5256701	cell colture reagent
Hanks' Balanced Salt Solution (HBSS)	GIBCO	12350039	cell colture reagent
Human Midkine ELISA kit	Cohesion Biosciences	CEK1270	kit ELISA
Inverted microscope	Eclipse Ts2, Nikon	MFA34100	microscope
L929 mouse fibroblasts	ATCC	ATCC ®-CCL1	cell line
LC3-II	Novus Biologicals	NB910-40752SuperNovusPack	antibody
L-Glutamine	GIBCO	A2916801	cell colture reagent
Occludin	Novus Biologicals	NBP1-87402	antibody
Paraffin Lab-O-Wax PLUS 56 °C–58 °C	Histoline laboratories	R0040 PLUS	instrument for paraffin embedding
Penicillin-Streptomycin	GIBCO	15070063	cell colture reagent
Penicillin-Streptomycin	GIBCO	15070063	cell colture reagent
rat tail collagen type I	GIBCO	A1048301	cell colture reagent
Roswell Park Memorial Institute Medium (RPMI)	GIBCO	21870076	cell colture reagent
Sodium pyruvate	GIBCO	11360070	cell colture reagent
Thermo Fisher Countess II FL Automated Cell Counter	Thermo Fisher	TF-CACC2FL	cell counting instrument
Transwell	Costar Corning	3413	plastic for cell colture
Trypsin	GIBCO	15090046	cell colture reagent
U937 a pro-monocytic, human myeloid leukemia cell line	ATCC	ATCC CRL-1593.2	cell line
UltraTek Alk-Phos Anti-Polyvalent (permanent red) Stain Kit	ScyTek Laboratories, Inc.	AMH080	kit for immunohistochemical staining