Sexual Dimorphism in the Response to Broad-spectrum Antibiotics During T Cell-mediated Colitis

Abstract

Background: Broad-spectrum antibiotics [Abx], including combination therapy with ciprofloxacin and metronidazole, are often prescribed during the treatment of inflammatory bowel disease [IBD] to alleviate symptoms, but with varying success. In this pilot study, we studied the effects of Abx on the course of experimental colitis, with a particular focus on sex as a determinant of the microbial and inflammatory responses.

Methods: The effects of Abx were tested on colonic inflammation and microbiome in male and female Rag-/- mice, using adoptive transfer of naïve T cells to induce colitis in a short-term [2-week] and long-term [9-week] study.

Results: We observed disparities between the sexes in both the response to adoptive T cell transfer and the effects of Abx. At baseline without Abx, female mice displayed a trend toward a more severe colitis than males. In both the short- and the long-term experiments, gut microbiota of some female mice exposed to Abx showed weak, delayed, or negligible shifts. Caecum weight was significantly lower in Abx-treated females. Abx exposure favoured a quick and persistent rise in Enterococcaceae exclusively in females. Males had higher relative abundance of Lactobacillaceae following Abx exposure relative to females. Abx-treated females trended toward higher colitis scores than Abx-treated males, and towards higher levels of IL-17A, NOS2, and IL-22.

Conclusions: Although preliminary, our results suggest a differential response to both inflammation and Abx between male and female mice, The findings may be relevant to current practice and also as the basis for further studies on the differential gender effects during long-term antibiotic exposure in IBD.

Figure 1.

Males and females respond to differently to antibiotics and inflammation. Male and female Rag2^-/- mice were placed on metronidazole and ciprofloxacin [Abx] in drinking water, or alternatively maintained on autoclaved water. They were subsequently injected with naïve CD45⁺CD45RB^hi T cells and maintained for either 2 or 9 weeks in separate cohorts. Histological inflammation scores were assessed for proximal and distal colon separately and added to the result as total colonic score. [A] Summary from the 2-week and 9-week studies, showing changes to the colonic inflammation over time. Two-way ANOVA with a Tukey’s multiple comparisons test was used for analysis. [B-C] Caecum weight was calculated as a percentage of total body weight to avoid gross differences in size between the sexes. One-way ANOVA with Tukey’s multiple comparisons test was used for analysis: *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001 for post hoc tests.

Figure 2.

Long-term exposure to broad-spectrum antibiotics tends to promote higher mucosal cytokine expression in females. Following nine weeks of adoptive transfer and antibiotic exposure, proximal and distal colonic segments were resected, snap-frozen, and RNA-extracted for downstream analysis by RT-qPCR, analysed by the 2^-∆∆CT method against a GAPDH housekeeping gene [A]. IL-22 mRNA expression in the proximal colonic mucosa. [B] IL-17A expression in the proximal and distal colonic mucosa. [C] NOS2 mRNA expression in the distal colonic mucosa. For each treatment group [H₂O or Abx] males served as controls for data normalisation. A ShapiroWilke normality test was used to determine the normality of each group. Normally distributed groups were analysed by Student’s t test, and non-parametric groups were analysed by a Mann-Whitney U test. Only trends were observed without reaching statistical significance of p ≤0.05. RT-qPCR, real-time quantitative polymerase chain reaction.

Figure 3.

Long-term antibiotic administration significantly alters NHE3 mRNA expression in females but not males. RT-qPCR showing relative NHE3 expression in the distal colon of male and female mice given either autoclaved water or ciprofloxacin and metronidazol [Abx]. RNA was collected from snap-frozen distal colonic segments collected from the 9-week cohort. The 2^-∆∆CT method was used for analysis against a GAPDH housekeeping gene. Antibiotic groups were normalised to the groups given autoclaved water in each sex group; *p < 0.05 MannWhitney U test [n = 6]. RT-qPCR, real-time quantitative polymerase chain reaction.

Figure. 4.

The gut microbiota of females on antibiotics shows dissimilarity from males on antibiotics. Non-metric multidimensional scaling [NMDS] based on Bray-Curtis dissimilarities, as analysed in R from faecal 16S rRNA amplicon sequences from Illumina MiSeq. Each colour-coded dot represents the gut microbiota of a single mouse at a given time point; all time points are shown from the 9-week cohort.

Figure 5.

Microbiota of female mice responds to Abx treatment in delayed, inconsistent, and at times marginal fashion. Unweighted UniFrac distance matrix analysis; principal coordinate analysis [PCoA] plots were generated by Qiiime with explicit time axis [weeks] for both the 2- [A] and the 9- week study [B]. Antibiotic administration is given on the first day of collection, Week/Day 0.

Figure 6.

Alpha diversity of the gut microbiota is differently affected in antibiotic-treated males and females. Species richness [number of operational taxonomic units, or OTUs] calculated in R from 16S faecal microbiota sequenced on Illumina MiSeq. [A] All timepoints from the 9-week cohort are shown. The Kruskal-Wallace chi square test was used; p < 2.2. [B] Data categorised by group and time of treatment.

Figure 7.

Compositional analysis indicates sex-specific changes in gut microbiota in response to broad-spectrum antibiotics. A family-level taxonomic analysis [SILVA database] of males and females given either autoclaved water [A] or metronidazole and ciprofloxacin [Abx] [B]. Each bar represents the mean of each group [n = 6], and each colour represents a bacterial family. The most significantly affected families, Lactobacillaceae [C] and Enterococcaceae [D], are highlighted [mean abundance; n = 6] over time.

Figure 8.

Phylogenetic Investigation of Communities by Reconstruction of Unobserved States [PICRUSt] prediction of the metagenome functional content based on 16S amplicon profiling in male vs female mice treated with antibiotics during the progression of T-cell mediated colitis. Data from final collection at 9-weeks following T cell transfer were used for analysis. All presented Kyoto Encylopedia of Genes and Genomes [KEGG] orthology groups showed statistical differences between male and female mice.