Department of Internal Medicine, Seoul National University Bundang Hospital, Seongnam, South Korea.
Department of Internal Medicine and Liver Research Institute, Seoul National University College of Medicine, Seoul, South Korea.
Tumor Microenvironment Global Core Research Center, College of Pharmacy, Seoul National University, Seoul, South Korea.
Cancer Research Institute, Seoul National University, Seoul, South Korea.
Abstract
The frequency of azoxymethane/dextran sulfate sodium (AOM/DSS)-induced carcinogenesis in male mice is higher than that in female mice. Previous studies have reported that 17β-estradiol inhibits tumorigenesis in males by modulating nuclear factor-erythroid 2-related factor 2 (Nrf2). This study aimed to investigate the changes in mouse gut microbiome composition based on sex, AOM/DSS-induced colorectal cancer (CRC), and Nrf2 genotype. The gut microbiome composition was determined by 16S rRNA gene sequencing fecal samples obtained at week 16 post-AOM administration. In terms of sex differences, our results showed that the wild-type (WT) male control mice had higher alpha diversity (i.e. Chao1, Shannon, and Simpson) than the WT female control mice. The linear discriminant analysis effect size (LEfSe) results revealed that the abundances of Akkermansia muciniphila and Lactobacillus murinus were higher in WT male control mice than in WT female controls. In terms of colon tumorigenesis, the alpha diversity of the male CRC group was lower than that of the male controls in both WT and Nrf2 KO, but did not show such changes in females. Furthermore, the abundance of A. muciniphila was higher in male CRC groups than in male controls in both WT and Nrf2 KO. The abundance of Bacteroides vulgatus was higher in WT CRC groups than in WT controls in both males and females. However, the abundance of L. murinus was lower in WT female CRC and Nrf2 KO male CRC groups than in its controls. The abundance of A. muciniphila was not altered by Nrf2 KO. In contrast, the abundances of L. murinus and B. vulgatus were changed differently by Nrf2 KO depending on sex and CRC. Interestingly, L. murinus showed negative correlation with tumor numbers in the whole colon. In addition, B. vulgatus showed positive correlation with inflammatory markers (i.e. myeloperoxidase and IL-1β levels), tumor numbers, and high-grade adenoma, especially, developed mucosal and submucosal invasive adenocarcinoma at the distal part of the colon. In conclusion, Nrf2 differentially alters the gut microbiota composition depending on sex and CRC induction.
Figure 1Experimental scheme for evaluating the differential effects of Nrf2 KO on the gut microbial composition depending on sex and CRC induction. (A) Experimental design. Male and female WT and Nrf2 KO mice were treated with AOM/DSS to induce colitis-associated CRC. The mice were injected with AOM (10 mg/kg bodyweight) on Day 0. At week 1 post-AOM injection, DSS (2.5% w/v) was supplemented in the drinking water for one week. The fecal samples were collected and the mice were sacrificed at week 16 post-AOM injection (aged 24 weeks). (B) Data analysis scheme. Examination of sex-specific changes in the gut microbiome composition by comparing the gut microbiota composition between males and females in the following groups: WT control, WT AOM/DSS-treated, Nrf2 KO control, and Nrf2 KO AOM/DSS-treated groups. Based on the results of this analysis, the effect of Nrf2 KO on the gut microbiome composition was examined. Furthermore, the CRC-specific differences in the gut microbiome composition were analyzed by comparing the gut microbiota composition between control and AOM/DSS-treated animals in the following groups: male WT, female WT, male Nrf2 KO, and female Nrf2 KO groups. Based on this analysis, the effect of Nrf2 KO on the CRC-specific changes in the gut microbiome composition was examined. CRC, colorectal cancer; WT, wild-type; KO, knockout; Nrf2, nuclear factor-erythroid 2-related factor 2; AOM, azoxymethane; DSS, dextran sodium sulfate; ♂, male; ♀, female.Figure 2UniFrac-based PCoA of 16S rRNA sequence from 64 fecal samples. Samples derived from control (A) and AOM/DSS-treated groups (B) with “Sex and KO” criteria, as well as those derived from male (C) and female mice (D) with “AOM/DSS and KO” criteria, were clustered using the Generalized UniFrac method at the species level. PERMANOVA was performed to examine the dissimilarity of bacterial population structures in (A–D). Below each PCoA plot, the PERMANOVA statistics between the two groups are shown. Significance for p- and q-values is indicated by asterisks: *, < 0.05; **, < 0.01; n.s. not significance. The clustering of each group is marked with a different color: WT_M_CON, filled blue ellipse; WT_F_CON, filled red ellipse; WT_M_AOM/DSS, filled green ellipse; KO_M_CON, blanked blue ellipse; KO_F_CON, blanked red ellipse; KO_M_AOM/DSS, blanked green ellipse; KO_F_AOM/DSS, blanked purple ellipse. PCoA, principal coordinate analysis; WT, wild-type; KO, Nrf2 knockout; AOM, azoxymethane; DSS, dextran sodium sulfate; M, male; F, female; PERMANOVA, permutational multivariate analysis of variance.Figure 3Analysis of species richness and diversity of the gut microbiome. (A–D) Observed OTU count, (E–H) species richness (Chao1 index), and (I–L) alpha diversity (Shannon index) of the microbial community in the groups with “Sex and KO” criteria (A, B, E, F, I, J), as well as with AOM/DSS and KO criteria (C, D, G, H, K, L). Data are expressed as the mean ± SEM. Whiskers show the minimum and maximum values. The p-values obtained from the Kruskal–Wallis test is shown in the figure; *p < 0.05 for comparison between two groups (Mann–Whitney U-test). OTU, operational taxonomic unit; WT, wild-type; Nrf2 KO, Nrf2 knockout; Con, control; AOM, azoxymethane; DSS, dextran sodium sulfate; SEM, standard error of mean.Figure 4Gut microbiota composition at the phylum level. (A–D) Taxonomic composition of fecal samples obtained from the groups with “Sex and KO” criteria (A, B), as well as with “AOM/DSS and KO” criteria (C, D). The abundance indicates the percentage of each phylum in total microorganisms. Stacked bars indicate mean ± SEM. Mann–Whitney U-test was used to analyze the difference between two independent groups in (A–D). ap < 0.05, males vs. females; bp < 0.05, controls vs. AOM/DSS-treated groups; cp < 0.05, WT controls vs. Nrf2 KO controls; dp < 0.05, WT AOM/DSS-treated vs. Nrf2 KO AOM/DSS-treated. (E–H)Firmicutes/Bacteroidetes ratio calculated by dividing the abundance of Firmicutes with that of Bacteroidetes in the groups with “Sex and KO” criteria, as well as with “AOM/DSS and KO” criteria. Data are expressed as the mean ± SEM. Whiskers show the minimum and maximum values. The p-values calculated from the Kruskal–Wallis test is shown in the figure. WT, wild-type; Nrf2 KO, Nrf2 knockout; Con, control; AOM, azoxymethane; DSS, dextran sodium sulfate; SEM, standard error of mean.Figure 5Sex-specific alterations in the abundance of the gut microbiota composition examined using LEfSe. Bar plots of the LEfSe results, which were obtained based on the following criteria: 1) the alpha value for the factorial Kruskal–Wallis H test between assigned taxa when compared with that of the groups < 0.05; 2) the alpha value for the pairwise Wilcoxon test among the taxonomic compositions < 0.05; 3) the threshold of the logarithmic LDA score for discriminative features < 2.0; 4) a multi-class analysis set as all-against-all; the LEfSe plot was further simplified; 5) overlapping bacteria selection; 6) identification and classification of the bacterial characteristics based on previous reports as “commensal bacteria,” “opportunistic pathogens,” and “not characterized;” and 7) removal of non-overlapping bacteria only at “not characterized” microbiome. In the LEfSe plot, all commensal bacteria and opportunistic pathogens and only overlapped “not characterized” bacteria are included. Bar plots at the species level with significant differences in abundance based on LEfSe. The color bars and lines show the LDA scores of species that enriched in indicated condition; (A) blue bar (WT male controls), (A) red bar (WT female controls), (B) blue line (Nrf2 KO male controls), (B) red line (Nrf2 KO female controls), (C) green bar (WT male AOM/DSS-induced CRC), (C) purple bar (WT female AOM/DSS-induced CRC), (D) green line (Nrf2 KO male AOM/DSS-induced CRC), and (D) purple line (Nrf2 KO female AOM/DSS-induced CRC). Each color on the species name indicates the characteristics of each genus: yellow for commensal bacteria, orange for opportunistic pathogens, and green for not characterized bacteria. The p-values were determined using the non-parametric factorial Kruskal–Wallis sum-rank test. (E) The ratio of commensal bacteria to opportunistic pathogens based on the LEfSe results in each comparison group. LDA, linear discriminant analysis; LEfSe, LDA effect size; CRC, colorectal cancer; WT, wild-type; KO, Nrf2 knockout; Con, control; AOM, azoxymethane; DSS, dextran sodium sulfate; M, male; F, female; N, number; NC, not calculated.Figure 6CRC-specific alterations in the abundance of gut microbiota examined using LEfSe. Bar plots of the LEfSe results, which were generated based on the criteria mentioned in the legend of Figure 5. Bar plots at the species level with significant differences in abundance based on LEfSe. The color bars and lines show the LDA scores of species that were enriched in the indicated condition; (A) blue bar (WT male controls), (A) green bar (WT male AOM/DSS-induced CRC), (B) blue line (Nrf2 KO male controls), (B) green line (Nrf2 KO male AOM/DSS-induced CRC), (C) red bar (WT female controls), (C) purple bar (WT female AOM/DSS-induced CRC), (D) red line (Nrf2 KO female controls), and (D) purple line (Nrf2 KO female AOM/DSS-induced CRC). Each color on the species name indicates the characteristics of each genus: yellow for commensal bacteria, orange for opportunistic pathogens, and green for not characterized bacteria. The p-values were calculated using the non-parametric factorial Kruskal–Wallis sum-rank test. (E) The ratio of commensal bacteria to opportunistic pathogens based on the LEfSe results in each comparison group. CRC, colorectal cancer; LDA, linear discriminant analysis; LEfSe, LDA effect size; WT, wild-type; KO, Nrf2 knockout; Con, control; AOM, azoxymethane; DSS, dextran sodium sulfate; M, male; F, female; N, number; NC, not calculated.Figure 7Nrf2 KO-specific alterations in the abundance of gut microbiota examined using LEfSe. Bar plots of the LEfSe results, which were generated based on the criteria mentioned in the legend of Figure 5. Bar plots at the species level with significant differences in abundance based on the LEfSe results. The color bars and lines show the LDA scores of species that are enriched in the indicated condition; (A) blue bar (WT male controls), (A) blue line (Nrf2 KO male controls), (B) red bar (WT female controls), (B) red line (Nrf2 KO female controls), (C) green bar (WT male AOM/DSS-induced CRC), (C) green line (Nrf2 KO male AOM/DSS-induced CRC), (D) purple bar (WT female AOM/DSS-induced CRC), and (D) purple line (Nrf2 KO female AOM/DSS-induced CRC). Each color on the species name indicates the characteristics of each genus: yellow for commensal bacteria, orange for opportunistic pathogens, and green for not characterized bacteria. The p-values were determined using the non-parametric factorial Kruskal–Wallis sum-rank test. (E) The ratio of commensal bacteria to opportunistic pathogens based on the LEfSe results in each comparison group. CRC, colorectal cancer; LDA, linear discriminant analysis; LEfSe, LDA effect size; WT, wild-type; KO, Nrf2 knockout; Con, control; AOM, azoxymethane; DSS, dextran sodium sulfate; M, male; F, female; N, number; NC, not calculated.Figure 8Negative correlation between Lactobacillus murinus and tumor numbers in the colon. (A–D) Abundance (%) of Lactobacillus murinus in the groups with “Sex and KO” criteria (A, B), as well as with “AOM/DSS and KO” criteria (C, D). Data are expressed as the mean ± SEM. Whiskers show the minimum and maximum values. The p-values calculated using the Kruskal–Wallis test are shown in the figure; *p < 0.05, comparison between two groups (Mann–Whitney U-test). CRC, colorectal cancer; WT, wild-type; Nrf2 KO, Nrf2 knockout; Con, control; AOM, azoxymethane; DSS, dextran sodium sulfate. (E) DAI score, (F) MPO concentration, (G) IL-1β concentration, (H) total tumor numbers, (I) grade of tumor arising from the proximal colon, and (J) grade of tumor arising from the distal colon. The abundance of Lactobacillus murinus was significantly correlated with total tumor numbers (H). p < 0.05 (Spearman’s correlation test) (E–J). Dotted line indicates the regression line. The correlations were analyzed using the Spearman’s rank correlation coefficient method (also known as Spearman’s rho) and linear regression. DAI, disease activity index; MPO, myeloperoxidase. Tumor grade: 0, no tumor; 1, low-grade adenoma; 2, high-grade adenoma; 3, mucosal invasive adenocarcinoma; 4, submucosal invasive adenocarcinoma.Figure 9Positive correlation between Bacteroides vulgatus and colitis-associated CRC indices. (A–D) Abundance (%) of Bacteroides vulgatus in the groups with “Sex and KO” criteria (A, B), as well as with “AOM/DSS and KO” criteria (C, D). Data are expressed as the mean ± SEM. Whiskers show the minimum and maximum values. The p-values calculated from the Kruskal–Wallis test are shown in the figure; *p < 0.05, comparison between two groups (Mann–Whitney U-test). CRC, colorectal cancer; WT, wild-type; KO, Nrf2 knockout; Con, control; AOM, azoxymethane; DSS, dextran sodium sulfate; SEM, standard error of mean. (E-J) The abundance of Bacteroides vulgatus was significantly correlated with tumor indices. (E) DAI score, (F) MPO concentration, (G) IL-1β concentration, (H) total tumor numbers, (I) grade of tumor arising from the proximal colon, and (J) grade of tumor arising from the distal colon. p < 0.05 obtained from Spearman’s correlation analysis (E–J). Dotted line indicates the regression line. The correlations were analyzed using Spearman’s rank correlation coefficient (also known as Spearman’s rho) and linear regression. DAI, disease activity index; MPO, myeloperoxidase. Tumor grade: 0, no tumor; 1, low-grade adenoma; 2, high-grade adenoma; 3, mucosal invasive adenocarcinoma; 4, submucosal invasive adenocarcinoma.Figure 10Illustrative summary. The findings of this study indicated that Nrf2 KO was associated with sex-specific and CRC-specific alterations in the gut microbiome composition in the mouse model. Nrf2 KO differentially altered the abundance of two bacterial species, which are L. murinus and B. vulgatus, depending on sex and CRC induction. Interestingly, the sex-specific difference in the abundance of L. murinus, which was modulated by Nrf2 KO, was negatively correlated with the colitis-associated CRC index. In contrast, the CRC-specific difference in the abundance of B. vulgatus, which was also regulated by Nrf2 KO, was positively correlated with colitis-associated CRC indices. WT, wild-type; KO, knockout; CRC, colorectal cancer; +, positive correlation; -, negative correlation.
그림 출처: PubMed Central 공개 논문 (PMC8261249)
Conflict of interest statement
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
그 밖의 정보
출판 형태
Journal Article, Research Support, Non-U.S. Gov't
저널
Frontiers in cellular and infection microbiology · eISSN 2235-2988
