Jejunum-derived NF-κB reporter organoids as 3D models for the study of TNF-alpha-induced inflammation
Size, morphology and replicative capacity differ between organoids derived from small intestine and colon
Intestinal organoids were established from isolated crypts obtained from the small and large intestine regions of male and female NF-κB-RE-luc transgenic mice. The same procedure and the biopsy size (3 cm pieces of tissue) were used for obtaining duodenum, jejunum, ileum, and colon-derived crypts. Nevertheless, the efficiency of the isolating procedure varied between the intestinal regions, being the ileum the segment with the lowest number of crypts recovered (data not shown).
After 24 h of culture, crypts closed themselves, leading to a cyst or a spherical cellular structure (spheroid) (Fig. 1a, Day 1). By day 4 of culture, spheroids grew in size and complexity, leading to the characteristic 3D architecture of cyst with a central lumen flanked by a simple polarized epithelium and the basal side of the cells oriented toward the outside. Organoids derived from the small intestine showed multiple new crypt-like structures or buds emerging from the center to the surrounding matrix, whereas organoids derived from the colon showed a symmetric spherical shape (Fig. 1a, Day 4).
To better characterize the different types of organoids, morphological and growth parameters were analyzed from single-cell dissociated organoid cultures (Fig. 1b–e). The number of organoids generated from 1000 single epithelial cells seeded, was evaluated after 4 days in culture. This parameter ranged between 11.33 ± 9.15 to 7.35 ± 4.20 organoids, with no statistically differences among the different intestinal organoids (P = NS) (Fig. 1b). Regarding growth parameters, organoids developed from single cells took more days to grow in size, compared with the culture isolated from crypts. By day 8, all organoids had increased their area several times compared with day 4 (Fig. 1c) and their aspect resembled those of organoids derived from intestinal crypts at day 4 of culture (data not shown). There was a significant difference between jejunum and colon-derived organoids with respect to the numbers of buds developed in each organoid. Jejunum-derived cultures presented a higher percentage of organoids with more than four buds (55.02 ± 5.76%, for jejunum; 26.25 ± 7.86% for colon, *P < 0.05), whereas colon-derived organoids showed a higher percentage of structures without buds (29.12 ± 3.82% for jejunum; 49.73 ± 2.82% for colon, #P < 0.05) (Fig. 1d). Circularity was in accordance with these results; small intestinal derived-organoids significantly decreased their circularity at day 8 of culture compared with day 4 (P < 0.05), whereas colon-derived organoids maintained the highest value for this parameter (day 8, 0.72 ± 0.19 circularity factor) throughout the evaluated period (Fig. 1e).
Histological analysis of organoids derived from jejunum and colon revealed multiple well-differentiated and preserved ductal structures, organized from the outside by a delicate basement membrane (Fig. 2). In colonoids a remnants of a strongly eosinophilic, amorphous matrix with a diffusely vacuolated appearance is observed. In both organoids the epithelial ducts have at least one-cell continuous layer of columnar-type epithelial cells, supported at their basal pole by a subtle eosinophilic extracellular matrix, with round to ovoid strongly basophilic nuclei. In those cases where clusters of ductal structures were observed, the epithelium were constituted by a epithelial cell multilayer (from 2 to 5 cells thickness), with similar histologic appearance to the previous description, always maintaining the ductal organization. Frequently, on the luminal border of these ducts, other less abundant rounded cells were observed, interspersed among the epithelial cell layers. These cells were identified by their clear eosinophilic foamy cytoplasm, with the presence of larger vacuoles, reminiscent of goblet cells. In the lumen of the ducts, it was also frequent to observe detachments of necro-apoptotic cellular masses, being rounded in appearance, with cariolisis, and with abundant and foamy cytoplasm (ghost cells).
Characterization of the epithelial cell population was performed by quantitative analysis of mRNA expression from organoids and intestinal tissue samples. Five specific genes, representative for the main intestinal epithelial cell types were evaluated, Lgr5 (intestinal stem cell), Lysozyme (Paneth cell), Villin (enterocytes), Chromogranin A (enteroendocrine cells), and Mucin 2 (Goblet cells). As shown in Fig. 3a, the expression pattern of Villin, Chromogranin A and Lysozyme were similar between tissue and organoids from the same intestinal region. Lysozyme could not be detected in samples from tissues and organoids derived from the colon, since Paneth cells do not populate this intestinal region. Mucin 2 expression was reduced in colon-derived organoids, compared with the same tissue (P < 0.05). Of note, except for ileum, Lgr5 levels were higher in organoids compared with the corresponding tissue (P < 0.05). The presence of Paneth cells in small intestine-derived organoids were corroborated at the protein level by immunodetection of Lysozyme, whereas actively proliferating cells were found in both organoids, as indicated by the presence of Ki67 positive cells (Fig. 3b).
NF-κB reporter ileum- and jejunum-derived organoids are most sensitive to TNF-α stimulus
The activation of the NF-κB transcription factor was evaluated by stimulating transgenic and wild-type organoids derived from the small intestine with different concentrations of Tumor Necrosis Factor-alpha (TNF-α) as the pro-inflammatory stimulus. After 24 h of incubation, luciferase activity was measured. As shown in Fig. 4a, luciferase activity was only detected in NF-κB-RE-Luc organoids, which showed a concentration-dependent response to the inflammatory stimulus. The maximum NF-κB activation levels were attained at 100 ng/mL TNF-α, with more than 20 fold change increase compared with the control value (un-stimulated organoids) (P < 0.05), confirming the ability of the organoids to report the activation of NF-κB signaling pathway.
To better characterize the model, the responsiveness of organoids derived from duodenum, jejunum, ileum and colon were also evaluated with TNF-α. The sensitivity of the reporter response varied between the intestinal regions and within the same region, among different cultures (see supplementary material, Table S1 and Figs. S1 and S2). Ileum- and jejunum-derived organoids were the most sensitive to the stimulus, showing a significant increase in the reporter signal with concentrations equal or higher than 1 and 2.5 ng/mL of TNF-α, respectively. On the opposite side, it was not possible to detect any reporter signal from colonic organoids. Organoids obtained from duodenum produced a detectable luciferase reporter response only at higher concentrations of TNF-α (at 10 and 100 ng/mL) (Fig. 4b).
Organoids obtained from a responsive (jejunum) and unresponsive (colon) intestinal region were further studied for the presence of TNF-α Receptor 1 and TNF-α Receptor 2 (TNF-R1 and TNF-R2). Both intestinal organoids were able to express mRNA from TNF-R1 and TNF-R2 (Fig. 4c), which suggests that they could bind the TNF-α and respond to this stimulus.
Regarding the stability of the reporter system over time, the NF-κB reporter activity of jejunum-derived organoids was evaluated with different concentrations of TNF-α at different passages. As indicated in Table 1, the concentration of TNF-α that gives half-maximal response (EC50 values) were similar among all the passages, indicating a stable reporter activity at least during 16 passages. Beyond this fact, we defined passage 12 as the maximum passage number for other assays.
Direct measurement of NF-κB pathway functionality was evaluated by comparing TNF-α-induced translocation of NF-κB in jejunum- and colon-derived organoids. Organoids were stimulated with 50 ng/mL of TNF-α for 3 h and the nuclear translocation of NF-κB p65 subunit was analyzed by confocal microscopy (Fig. 5). Images show that in TNF-α-stimulated organoids, NF-κB p65 signal is more homogeneously distributed between the nucleus and the cytoplasm, while in unstimulated controls, p65 signal predominate in the cytoplasm, being more evident the presence of non labeled-nuclei (Fig. 5a, jejunum and colon). These findings suggest that TNF-α induced the translocation of the p65 subunit to the nucleus, while in non-stimulated cells, NF-κB was mainly detected in the cytoplasm. Nuclear translocation was quantified by calculating the ratio between the signal intensity of the nucleus and the cytoplasm (N/C ratio). In both jejunum and colon-derived organoids, the N/C ratio was increased (P < 0.05) in the TNF-α-stimulated organoids compared to non-stimulated conditions, indicating nuclear translocation of NF-κB transcription factor, a key step in the performance of the reporter assay (Fig. 5b,c).
These results confirmed the presence of a functional NF-κB signaling pathway in both jejunum- and colon-derived organoids. Thus the lack of reporter activity in colon-derived organoids would not be associated with an alteration of this signaling pathway.
In addition to TNF-α, the intestinal epithelium is frequently exposed to other inflammatory stimuli. We next evaluated the ability of organoids to respond to three well known intestinal inflammatory compounds: Lipopolysaccharide (LPS), a heat-inactivated Salmonella enterica extract22 and interleukin 1 beta (IL-1β). Neither of them induced a detectable reporter response in the NF-κB-RE-Luc organoids, regardless of the intestinal region analyzed (Fig. 6).
Natural and synthetic compounds modulate TNF-α- induced NF-κB activation in jejunum-derived organoids
Based on our results, TNF-α induced a potent luciferase response at 10 ng/mL in reporter organoids derived from jejunum. Therefore to validate the model, we selected this system and tested synthetic and natural compounds known to interfere with the NF-κB signaling pathway.
Dexamethasone (Dex) and Bay11-7082 (Bay) significantly reduced TNF-α-induced NF-κB activation, as indicated by the decrease in the luciferase activity (NF-κB activation as fold change with respect to the unstimulated control: TNF-α = 5.92 ± 1.12; Dex = 3.11 ± 0.89; Bay = 4.05 ± 0.27) (Fig. 7a). The natural peptide Vioprolide A (VioA) as well as Lactobacillus plantarum and L. reuteri conditioned media (CM), all were able to significantly reduce the TNF-α-induced NF-κB activation as indicated by the reduction in the reporter response, without altering the basal value, when compared with the unstimulated control (NF-κB activation in fold change respect to the unstimulated control: L. reuteri CM = 2.57 ± 0.71; L. plantarum CM = 2.07 ± 0.27; VioA = 3.45 ± 0.36) (Fig. 7b).