Gut Dysbiosis and Its Associations with Gut Microbiota-Derived Metabolites in Dogs with Myxomatous Mitral Valve Disease

Abstract

Gut dysbiosis and gut microbiota-derived metabolites, including bile acid (BA), short-chain fatty acid, and trimethylamine N-oxide (TMAO), are associated with cardiovascular disease. Canine myxomatous mitral valve disease (MMVD) is a model for human MMVD. The aim of the study is to evaluate gut microbial dysbiosis and its relationship with gut-produced metabolites in dogs with MMVD. Fecal samples from 92 privately owned dogs, including 17 healthy, 23 and 27 asymptomatic MMVD dogs without (stage B1) and with (stage B2) secondary cardiac enlargement, respectively, and 25 MMVD dogs with history of congestive heart failure (stage C or D), were analyzed by 16S rRNA sequencing. Alpha and beta diversities were different between healthy and MMVD dogs (adjusted P < 0.05). The average dysbiosis indexes were -1.48, -0.6, 0.01, and 1.47 for healthy, B1, B2, and C/D dogs, respectively (P = 0.07). Dysbiosis index was negatively correlated with Clostridium hiranonis (P < 0.0001, r = -0.79). Escherichia coli, capable of trimethylamine production in the gut, had an increased abundance (adjusted P < 0.05) and may be responsible for the increased circulating TMAO levels in stage B2 and C/D MMVD dogs. Primary and secondary BAs showed opposite associations with C. hiranonis, a key BA converter (P < 0.0001 for both, r = -0.94 and 0.95, respectively). Secondary BAs appeared to promote the growth of Fusobacterium and Faecalibacterium but inhibit that of E. coli Multivariate analysis revealed significant but weak associations between gut microbiota and several circulating metabolites, including short-chain acylcarnitines and TMAO.IMPORTANCE Our study expands the current "gut hypothesis" to include gut dysbiosis at the preclinical stage, prior to the onset of heart failure. Gut dysbiosis index increases in proportion to the severity of myxomatous mitral valve disease (MMVD) and is inversely associated with Clostridium hiranonis, a key bile acid (BA) converter in the gut. Secondary BAs appear to promote the growth of beneficial bacteria but inhibit that of harmful ones. An intricate interplay between gut microbiota, gut microbiota-produced metabolites, and MMVD pathophysiological progression is implicated.

Keywords: Clostridium hiranonis; bile acid; canine; congestive heart failure; dysbiosis; metabolite; microbial metabolite; microbiome; microbiota; mitral valve disease; trimethylamine N-oxide.

FIG 1

Faith’s phylogenetic diversity (PD) index (A) and number of unique bacterial species (B). P values were from Tukey’s post hoc tests following ANOVA. (C) Bootstrap experiments. Only subsamples with no difference in age are shown. Distributions of P values on Faith’s PD index, body weight (BW), age, body condition score (BCS), and sex from a bootstrap study (left) and histogram of P values on Faith’s PD index (right). P values are expressed as −log₁₀ P. (D) Bar plots of the five predominant phyla. *, P < 0.05; **, P < 0.01.

FIG 2

(A) Principal-coordinate analysis on the Bray-Curtis distance of the four groups. The first two principal coordinates (PCs), PC1 and PC2, are shown. The x and y axes indicate data variances captured by PC1 and PC2, respectively. (B) Box plots of PC1 by groups. ANOVA and Tukey’s tests found differences on PC1. No difference was found on PC2. Adjusted P value: *, P < 0.05.

FIG 3

Significant operational taxonomical units (OTUs). Bacteria that shared at least 97% sequence similarity in their 16S marker genes were considered to be the same OTU. The lowest taxonomical rank in the OTU lineage is shown, with its OTU identifier (ID) inside the parentheses. These OTUs represented five genera, Megamonas, Blautia, Turicibacter, Bacteroides, and Oscillospira (A to E), six species, E. dolichum, F. prausnitzii, B. pullicaecorum, B. producta, E. coli, and E. uniformis (F to K), and three families, Ruminococcaceae, Bacteroidaceae, and Erysipelotrichaceae (L to O). N and O were two different OTUs. The horizontal lines indicated means. The percentage of nonzero samples is indicated below each group. Nonparametric Kruskal-Wallis test was performed on each OTU. Significant OTUs were subject to post hoc Dunn’s tests with corrections for false-discovery rate (FDR). ⁺, FDR < 0.1; *, FDR < 0.05; **, FDR < 0.01; ***, FDR < 0.001.

FIG 4

PCR-based fecal dysbiosis index using a panel of gut bacteria. Dysbiosis index (A), total bacteria (B) and seven bacterial genera and species, Faecalibacterium, Turicibacter, Streptococcus, E. coli, Blautia, Fusobacterium, and C. hiranonis (C to I) in dogs with MMVD. Bacterial abundances were measured using quantitative PCR. The horizontal lines denoted medians. Adjusted P values: *, P < 0.05, **, P < 0.01.

FIG 5

Alpha and beta diversity analyses on DI and C. hiranonis using the 16S sequencing data. Pearson’s correlations between Faith’s PD index and DI (A) and C. hiranonis (B). Principal-coordinate analysis using the Bray-Curtis distances on DI (C) and C. hiranonis (D). Samples were colored based on DI, cyan (L, DI ≤ 0), gray (M, 0 < DI < 2), or orange (H, DI ≥ 2) (C), and on C. hiranonis abundance, cyan (L, log₁₀ DNA < 4.5) or orange (H, log₁₀ DNA ≥ 4.5) (D). The first two principal coordinates, PC1 and PC2, are displayed with the percentages of data variation denoted in the x and y axes, respectively. (A and B) Fitted linear regression lines with 95% confidence intervals are included. Clustering of samples along PC1 was evidenced in panels C (P = 1.9e⁻¹⁴) and D P = 8.5e⁻¹⁴. DI, dysbiosis index.

FIG 6

Pearson’s correlations between fecal bile acids (BAs) and dysbiosis index. CA, cholic acid; CDCA, chenodeoxycholic acid; DCA, deoxycholic acid; LCA, lithocholic acid; UDCA, ursodeoxycholic acid; 1° BA/2° BA, the ratio of primary to secondary BAs. The percentage of each BA was calculated as the ratio of the BA to the sum of primary and secondary BAs. Only unconjugated BAs were considered. Correlation coefficients (r) were indicated on the top right corner. P < 1e⁻⁴ in all cases.

FIG 7

Pearson’s correlations between fecal bile acids (BAs) and gut microbes. CA, cholic acid; LCA, lithocholic acid; DCA, deoxycholic acid; CDCA, chenodeoxycholic acid; UDCA, ursodeoxycholic acid; 1° BA/2° BA, the ratio of primary to secondary BAs; r, correlation coefficient. (A to L) The percentage of each BA was calculated as the ratio of BA to the sum of primary and secondary BAs (y axis). Bacterial abundance was expressed as log₁₀ DNA abundance (x axis). Only unconjugated BAs were considered. Samples were colored by dysbiosis index (DI): green for DI ≤ 0, red for DI ≥ 2, gray for others. P < 1e⁻⁵ in all cases.