Intestinal Organoids in Colitis Research: Focusing on Variability and Cryopreservation

Abstract

In recent years, stem cell-derived organoids have become a cell culture standard that is widely used for studying various scientific issues that were previously investigated through animal experiments and using common tumor cell lines. After their initial hype, concerns regarding their standardization have been raised. Here, we aim to provide some insights into our experience in standardizing murine colonic epithelial organoids, which we use as a replacement method for research on inflammatory bowel disease. Considering good scientific practice, we examined various factors that might challenge the design and outcome of experiments using these organoids. First, to analyze the impact of antibiotics/antimycotics, we performed kinetic experiments using ZellShield® and measured the gene expression levels of the tight junction markers Ocln, Zo-1, and Cldn4, the proliferation marker Ki67, and the proinflammatory cytokine Tnfα. Because we found no differences between cultivations with and without ZellShield®, we then performed infection experiments using the probiotic Escherichia coli Nissle 1917 as an already established model setup to analyze the impact of technical, interexperimental, and biologic replicates. We demonstrate that interexperimental differences pose the greatest challenge for reproducibility and explain our strategies for addressing these differences. Additionally, we conducted infection experiments using freshly isolated and cryopreserved/thawed organoids and found that cryopreservation influenced the experimental outcome during early passages. Formerly cryopreserved colonoids exhibited a premature appearance and a higher proinflammatory response to bacterial stimulation. Therefore, we recommend analyzing the growth characteristics and reliability of cryopreserved organoids before to their use in experiments together with conducting several independent experiments under standardized conditions. Taken together, our findings demonstrate that organoid culture, if standardized, constitutes a good tool for reducing the need for animal experiments and might further improve our understanding of, for example, the role of epithelial cells in inflammatory bowel disease development.

Figure 1

ZellShield® does not affect organoid kinetics. Comparison of the gene expression levels in colonoids over the course of 12 hours after the administration of fresh medium with (+ ZS) and without (Ø ZS) ZellShield® (each: n = 5). The graphs show the means with SDs. ^∗P < 0.05, ^∗∗P < 0.01, ^∗∗∗P < 0.001, and ^∗∗∗∗P < 0.0001. (a) Relative Ocln expression, ANOVA (F (9, 41) = 0.9771; P = 0.4728). (b) Relative Zo-1 expression, ANOVA (F (9, 42) = 1.088; P = 0.3915). (c) Relative Cldn4 expression, ANOVA (F (9, 42) = 0.9054; P = 0.5295). (d) Relative Tnfα expression, ANOVA (F (9, 42) = 0.4397; P = 0.9055). (e) Relative Ki67 expression, ANOVA (F (9, 41) = 1.946; P = 0.0719).

Figure 2

Experimental data are predominantly affected by interexperimental differences. The technical and experimental reproducibility and the biological variability in colonoids were compared using the EcN infection model. ^∗P < 0.05, ^∗∗P < 0.01, ^∗∗∗P < 0.001, and ^∗∗∗∗P < 0.0001. (a) Technical replicates. The graphs plot individual values of five representative organoid lines with the technical replicate mean and SD values; the outliers, as identified using the Grubbs outlier test, are shown as black dots. (b) Experimental and biological replicates. The graphs plot five independent experiments with the technical replicate means from three different organoid lines (biological replicates) per experiment plus the overall mean and SD per experiment; the outliers, as identified using the Grubbs outlier test, are shown as red triangles; ANOVA: Ocln (F (9, 20) = 2.609, P = 0.0355), Zo-1 (F (9, 20) = 1.538, P = 0.2019), Cldn4 (F (9, 20) = 4.227, P = 0.0035), Tnfα (F (9, 20) = 29.69, P < 0.0001), and Ki67 (F (9, 20) = 4.150, P = 0.0039). (c) Pooled data. The graphs plot the technical replicate means of all biological replicates (n = 15) from all independent experiments plus the overall mean and 95% CI.

Figure 3

Characterization of freshly isolated and cryopreserved colonoids during growth and after EcN infection. (a) Representative light phase-contrast images. (b) Immunofluorescent staining for epithelial cell adhesion molecule (CD326), KI67, and tight junction protein 1 (ZO-1). (c) The gene expression levels in freshly isolated colonoids (n = 17) and cryopreserved and thawed organoids (n = 12) of the same passage were compared before and after infection with EcN. The graphs plot the pooled technical replicate means of biological replicates plus the overall mean and 95% CI. ^∗P < 0.05, ^∗∗P < 0.01, ^∗∗∗P < 0.001, and ^∗∗∗∗P < 0.0001. Relative Slc5a1 expression (ANOVA: F (3, 54) = 1.179; P = 0.3264). Relative Chga1 expression (ANOVA: F (3, 53) = 1.104; P = 0.3555). Relative Muc2 expression (ANOVA: F (3, 54) = 2.932; P = 0.0416). Relative Ki67 expression (ANOVA: F (3, 54) = 3.435; P = 0.0231).

Figure 4

Fresh and cryopreserved organoids exhibit a similar gene expression profile in response to EcN injection. Heat map displaying significantly (adjusted P value < 0.01) differentially expressed genes detected between EcN and control samples in fresh colonoids. Besides the normalized and log transformed expression levels of the samples originating fresh organoids, the heat map includes in an analogous manner the samples of thawed colonoids.

Figure 5

Gene set enrichment analysis. Significantly (adjusted P value < 0.01) upregulated genes in EcN compared to control samples originating from fresh and cryopreserved colonoids, respectively, were examined for their similarity to gene sets of the KEGG 2019 Mouse library via Enrichr gene set enrichment analysis web server. (a) Top enriched terms in the KEGG 2019 Mouse library, with P values. Asterisk symbolizes the term has an adjusted P value < 0.05. (b) Each hexagon represents one gene set from the KEGG 2019 Mouse library. The brighter the blue color is, the more similar and, therefore, significant the specific gene set. Hexagons that are grouped together represent similar gene sets.

Figure 6

Cryopreservation affects gene expression levels in colonoids after EcN infection. The gene expression levels in freshly isolated colonoids (n = 17) and cryopreserved and thawed organoids (n = 12) of the same passage after infection with EcN were compared. The graphs plot the pooled technical replicate means of biological replicates plus the overall mean and 95% CI. ^∗P < 0.05, ^∗∗P < 0.01, ^∗∗∗P < 0.001, and ^∗∗∗∗P < 0.0001. (a) Relative Ocln expression (ANOVA: F (3, 53) = 0.4720; P = 0.7031). (b) Relative Zo-1 expression (ANOVA: F (3, 54) = 1.010; P = 0.3953). (c) Relative Cldn4 expression (ANOVA: F (3, 53) = 4.962; P = 0.0041). (d) Relative Tnfα expression (ANOVA: F (3, 54) = 91.30; P < 0.0001).