Nonsteroidal Anti-Inflammatory Drug-Induced Leaky Gut Modeled Using Polarized Monolayers of Primary Human Intestinal Epithelial Cells

Abstract

The intestinal epithelium provides a critical barrier that separates the gut microbiota from host tissues. Nonsteroidal anti-inflammatory drugs (NSAIDs) are efficacious analgesics and antipyretics and are among the most frequently used drugs worldwide. In addition to gastric damage, NSAIDs are toxic to the intestinal epithelium, causing erosions, perforations, and longitudinal ulcers in the gut. Here, we use a unique in vitro human primary small intestinal cell monolayer system to pinpoint the intestinal consequences of NSAID treatment. We found that physiologically relevant doses of the NSAID diclofenac (DCF) are cytotoxic because they uncouple mitochondrial oxidative phosphorylation and generate reactive oxygen species. We also find that DCF induces intestinal barrier permeability, facilitating the translocation of compounds from the luminal to the basolateral side of the intestinal epithelium. The results we outline here establish the utility of this novel platform, representative of the human small intestinal epithelium, to understand NSAID toxicity, which can be applied to study multiple aspects of gut barrier function including defense against infectious pathogens and host-microbiota interactions.

Keywords: NSAIDs; bacterial translocation; leaky gut; mitochondria; small intestine; superoxide.

Figure 1. Overview of the two-dimensional human small intestine (HSI) monolayer utilized

(A) Schematic of the Transwell apparatus containing a collagen scaffold, upon which human small intestinal monolayers are cultured. (B) Longitudinal cross section of monolayers stained with F-actin and Integrin β4 to label the luminal and basolateral regions, respectively. (C) En face view of monolayers stained with F-actin and E-cadherin reveals characteristic flagstone staining pattern. Scale bar, 25 μm. (D) HSI monolayers consist primarily of proliferating, EdU+ cells. Colors used: EdU, red; DNA, blue.

Figure 2. Diclofenac exerts cytotoxicity by reducing mitochondrial membranepotential and inducing O₂^•−

(A) DCF-treated cells have reduced Mitotracker CMH₂-XRos staining compared to vehicle-treated control. Images were acquired at 20x and are representative images from three independent experiments. (B) Quantification of the mean fluorescence intensity of (A) demonstrates that DCF significantly reduces CMH₂-XRos staining. (C) Compared to untreated control, 24h treatment with DCF exerts significant dose-dependent cytotoxicity, as measured by CellTox Green endpoint assay. ** p<0.01 by one-way ANOVA with Dunnett’s test for multiple comparisons to the control. (D) MitoSOX staining reveals high levels of O₂^•− generated in response to DCF, within 24h. Cells are counterstained in both (A) and (D) with Hoechst to label nuclei, and Mitotracker Green to label mitochondria. Scale bar, 25 μm.

Figure 3. DCF reduces the proliferation of stem/progenitor cells in HSI monolayers

(A) DCF-treated HSI monolayers have a smaller proportion of EdU+ nuclei compared to controls. Scale bar = 200 μm. Following 24h treatment, while (B) cellular density (number of cells per square millimeter) is unchanged, DCF decreases (C) the total number of nuclei in HSI monolayers. (D) DCF reduces EdU fluorescence intensity of HSI, and (E) reduces the histological score, which is described in Methods. ** p<0.01, ***p< 0.001 by one-way ANOVA with Dunnett’s test for multiple comparisons to the control.

Figure 4. DCF breaches barrier integrity and causes permeability in HSI monolayers

(A) Sequence of the experimental workflow detailing points at which TEER was measured. (B) 1000 and 2000 μM of DCF significantly reduces ΔTEER (%) within 24h. Mean ΔTEER (%) of DCF-treated samples were compared to mean of vehicle-treated cells using one-way ANOVA with Dunnett test for multiple comparisons. *p<0.05. (C) DCF increases the P_app of LY through HSI within 24h, with significant differences observed for 1000 and 2000 μM DCF. Means of treated cells were compared to vehicle-cells using one-way ANOVA with Dunnet test for multiple comparisons. *p<0.05, ***p<0.001. (D) 24h treatment with LPS along did not significantly reduces ΔTEER (%) (p=0.266) but DCF significantly reduces ΔTEER%, and this reduction is also observed when LPS is also added to cells, indicating an overall loss of barrier integrity. Means of experimental groups were compared to each other using two-way ANOVA with Sidak’s correction for multiple comparisons. **p<0.01, ***p<0.001. (E) Immunofluorescence of F-actin and E-cadherin demonstrates a loss of flagstone pattern staining with DCF treatment, which is also observed with LPS co-treatment. Scale bar, 25 μm.