Lack of Small Intestinal Dysbiosis Following Long-Term Selective Inhibition of Cyclooxygenase-2 by Rofecoxib in the Rat

Abstract

Intestinal dysbiosis is linked to numerous gastrointestinal disorders, including inflammatory bowel diseases. It is a question of debate if coxibs, selective inhibitors of cyclooxygenase (COX)-2, cause dysbiosis. Therefore, in the present study, we aimed to determine the effect of long-term (four weeks) selective inhibition of COX-2 on the small intestinal microbiota in the rat. In order to avoid mucosal damage due to topical effects and inflammation-driven microbial alterations, rofecoxib, a nonacidic compound, was used. The direct inhibitory effect of rofecoxib on the growth of bacteria was ruled out in vitro. The mucosa-sparing effect of rofecoxib was confirmed by macroscopic and histological analysis, as well as by measuring the intestinal levels of cytokines and tight junction proteins. Deep sequencing of bacterial 16S rRNA revealed that chronic rofecoxib treatment had no significant influence on the composition and diversity of jejunal microbiota. In conclusion, this is the first demonstration that long-term selective inhibition of COX-2 by rofecoxib does not cause small intestinal dysbiosis in rats. Moreover, inhibition of COX-2 activity is not likely to be responsible per se for microbial alterations caused by some coxibs, but other drug-specific properties may contribute to it.

Keywords: cyclooxygenase-2; enteropathy; inflammatory bowel diseases; intestinal dysbiosis; microbiota; rofecoxib.

Figure 1

The effect of rofecoxib (ROF, 1 mg/kg, n = 6; 5 mg/kg, n = 8; 10 mg/kg, n = 5) on the levels of PGE2 in the gastric mucosa (A) and pouch exudate (B) in the carrageenan-airpouch model. The effect of 5 mg/kg rofecoxib was also assayed 24 h after the final gavage (n = 8). The results are expressed as the mean ± SEM percent of the control PGE2 levels measured in vehicle-treated rats. ***p < 0.001 compared to control (one-way ANOVA, Holm-Sidak post hoc test).

Figure 2

The effects of four-week vehicle (VEH, 1% methylcellulose) and rofecoxib (ROF, 5 mg/kg) treatment on the body weight (A) and gastrointestinal mucosa. (B): Macroscopic scores of gastric mucosa; (C): Representative photos of the gastric mucosa and histological micrographs (haematoxylin-eosin staining); (D): Macroscopic scores of small intestinal mucosa; (E): Length of small intestines; (F): Representative photos of the jejunal mucosa and histological micrographs (haematoxylin-eosin staining), scale bar: 200 µM. There are no signs of any macroscopic or histological tissue damage. (A): Results are expressed as the mean ± SEM. Panels (B), (D), and (E): Circles represent the data of each rat, bars indicate the mean ± SEM. For statistical analysis two-way repeated measures ANOVA followed by Holm-Sidak post hoc test (A), Mann-Whitney U test (B,D), and Student’s t test (E) were used, n = 8/group.

Figure 3

The effect of four-week vehicle (VEH, 1% methylcellulose) and rofecoxib (ROF, 5 mg/kg) treatment on the tissue protein levels of TNF-α (A, n = 8/group), IL-1β (B, n = 7–8/group), IL-10 (C, n = 7–8/group), occludin (D, n = 6/group), and claudin-1 (E, n = 6/group) in the distal jejunum of rats. Circles represent the data of each rat, bars indicate the mean ± SEM. For statistical analysis, Student’s t test was used. Panel F: Representative Western blots for occludin and claudin-1 proteins in the distal jejunum of vehicle- and rofecoxib-treated rats.

Figure 4

The relative abundance of bacterial families in jejunal samples of rats treated with vehicle (VEH, 1% methylcellulose) and rofecoxib (ROF, 5 mg/kg) for four weeks, determined by deep sequencing of 16S rRNA. Each vertical bar represents the sequencing data for one rat. Unclassified families and families with an abundance less than 0.1% are summarized as “Other”: Relative abundances of the bacterial families were compared between vehicle- and rofecoxib-treated groups by Wald test with Benjamini-Hochberg correction, which did not show any significant difference.

Figure 5

The effects of four-week vehicle (VEH, 1% methylcellulose) and rofecoxib (ROF, 5 mg/kg) treatment on the jejunal microbiota. (A): Relative abundances of the most abundant bacterial families in the jejunum. Box and whisker plots indicate the medians, first and third quartiles, and the minimum and maximum values. Panels (B) and (C): Bacterial richness (observed operational taxonomic units) and diversity estimated by the Shannon index, circles represent the data of each rat, bars indicate the mean ± SEM. (D): Principal component analysis (PCA) plot comparing the microbiota composition of vehicle- and rofecoxib-treated rats. The percentage of variation explained by the principal components (PC1 and PC2) is indicated on the axes. There was no clustering between rats treated with vehicle versus rofecoxib. For statistical analysis, Wald test with Benjamini-Hochberg correction (A), Mann-Whitney U test (B,C), and Hotelling’s T-square test (D) were used, n = 8/group.