The keystone gut species Christensenella minuta boosts gut microbial biomass and voluntary physical activity in mice

Abstract

The gut bacteria of the family Christensenellaceae are consistently associated with metabolic health, but their role in promoting host health is not fully understood. Here, we explored the effect of Christensenella minuta amendment on voluntary physical activity and the gut microbiome. We inoculated male and female germ-free mice with an obese human donor microbiota together with live or heat-killed C. minuta for 28 days and measured physical activity in respirometry cages. Compared to heat-killed, the live-C. minuta treatment resulted in reduced feed efficiency and higher levels of physical activity, with significantly greater distance traveled for males and higher levels of small movements and resting metabolic rate in females. Sex-specific effects of C. minuta treatment may be in part attributable to different housing conditions for males and females. Amendment with live C. minuta boosted gut microbial biomass in both sexes, immobilizing dietary carbon in the microbiome, and mice with high levels of C. minuta lose more energy in stool. Live C. minuta also reduced within and between-host gut microbial diversity. Overall, our results showed that C. minuta acts as a keystone species: despite low relative abundance, it has a large impact on its ecosystem, from the microbiome to host energy homeostasis.IMPORTANCEThe composition of the human gut microbiome is associated with human health. Within the human gut microbiome, the relative abundance of the bacterial family Christensenellaceae has been shown to correlate with metabolic health and a lean body type. The mechanisms underpinning this effect remain unclear. Here, we show that live C. minuta influences host physical activity and metabolic energy expenditure, accompanied by changes in murine metabolism and the gut microbial community in a sex-dependent manner in comparison to heat-killed C. minuta. Importantly, live C. minuta boosts the biomass of the microbiome in the gut, and a higher level of C. minuta is associated with greater loss of energy in stool. These observations indicate that modulation of activity levels and changes to the microbiome are ways in which the Christensenellaceae can influence host energy homeostasis and health.

Keywords: Christensenella; energy; microbiome.

Fig 1

Effects of live or heat-killed C. minuta amendment to fecal transplants to GF mice on C. minuta, murine adiposity, feed efficiency, and fecal energy loss 4 weeks post-inoculation. (A) Overview of the experimental set-up. Inocula consisted of a slurry derived from a fecal sample from an obese human donor amended with living (Live-CM) or heat-killed C. minuta (Killed-CM). Each experiment included 16 mice of one sex, with 8 mice per treatment group; each experiment was replicated six times. During the first 25 days, mice were housed in groups of four (females) or singly (males) at 22°C. After 25 days, mice were transferred to a sterilized behavioral phenotyping respirometry cage system (Promethion Sable Line, Nevada, USA) and housed singly at a temperature of 26°C. The first 24 h were considered as acclimatization and excluded from the analysis. (B) Quantification by qPCR of C. minuta in murine cecal contents, normalized by wet weight of cecal contents used for DNA extraction collected on day 28 post-inoculation. (C) Percent change in adiposity from day 0 to day 25 post-inoculation. (D and E) Average daily food intake (FI) from day 25 to day 28 post-inoculation: (D) FI raw values correlated with body weight and (E) FI residuals adjusted for sex, batch, and weight. (F) Feed efficiency (FE) over the duration of the experiment calculated using weight gain from day 0 to day 28 post-inoculation and daily food intake. (G) Residuals of daily energy excreted via feces (fecal energy) measured via bomb calorimetry adjusted for sex, batch, and weight. Asterisks indicate statistical significance of the linear mixed model correcting for sex, batch, and (D, E, and G) mouse weight. *P < 0.1; **P < 0.01; and ***P < 0.001. Adj. = adjusted and GE = genome equivalents.

Fig 2

Higher physical activity and metabolic energy expenditure in mice with live C. minuta. (A) Number of beam breaks by treatment group. (B) Distance traveled by sex and treatment. (C) Beam breaks plotted against normalized and standardized C. minuta. (D) Distance traveled plotted against normalized and standardized C. minuta, by sex. (E) Feed efficiency plotted against beam breaks. (F) Feed efficiency plotted against distance traveled, by sex. (G) Average hourly energy expenditure plotted against body weight for males and females. (H) Energy expenditure by treatment and sex. (I) Resting metabolic rate during the dark and light cycles. “Adj”: data were adjusted for effects of cabinet, batch, and sex (A–F), and body weight, batch, and sex (H and I). Asterisks indicate levels of statistical significance: *P < 0.1; **P < 0.01; and ***P < 0.001. Adj. = adjusted; GE = genome equivalents; and norm. = normalized.

Fig 3

C. minuta effects on gut microbiome biomass and diversity. (A) Quantification of microbial biomass (genome equivalents) in ceca via qPCR with universal 16S rRNA primers determined by the microbial standards and normalized by wet weight of cecal contents used for the DNA extraction and plotted by treatment group. (B) Microbial biomass (normalized plus standardized) plotted against C. minuta (normalized plus standardized). (C and D) α-diversity of phylogenetically profiled metagenomic cecal sequences by treatments. (E) Intra-group weighted and (F) unweighted UniFrac distances of the treatments. (G) Differential abundance analysis of microbial species using MaAsLin2 (29) and multiple hypothesis correction with the Benjamini-Hochberg method. Coefficients indicate associations with treatments (red: negative values = higher in Killed-CM and blue: positive values = higher in Live-CM). Asterisks indicate statistical significance of (A, C, and D) the linear mixed model correcting for sex and batch or (E and F) a permuted Wilcoxon rank sum test. (B) R² and P values of Spearman’s correlation analyses. *P < 0.1; **P < 0.01; and ***P < 0.001. Adj. = adjusted; F = females; GE = genome equivalents; Norm. = normalized; and M = males.

Fig 4

Changes of metabolism in relation to activity and energy expenditure. (A) For each sex, the sum of measured SCFA concentrations in murine cecal contents on day 28 post-inoculation was determined via gas chromatography-mass spectrometry. (B) Butyrate concentrations per sex and treatment. (C) Total SCFAs (normalized and standardized) plotted against adjusted distance traveled. (D) Butyrate (normalized and standardized) plotted against adjusted distance traveled. (E) Serum corticosterone levels at day 28 post-inoculation measured via liquid chromatography-mass spectrometry (LC-MS) per sex and treatment group. (F) Predicted metabolic pathways significantly associated with female energy expenditure including animals from both treatment groups, determined by analyzing the entire LC-MS peak spectrum of murine serum samples with XCMS (v.2.7.2) (31), MaAsLin2, including multiple hypothesis adjustment (29), and MetaboAnalyst5.0 (www.metaboanalyst.ca) (32). Pathways with a combined P value (GSEA and mummichog) <0.05 were considered significant. Asterisks indicate statistical significance of the linear mixed model correcting for sex and batch. (B and D) Marginal R² and P values of linear mixed model analyses are stated in the figure. **P < 0.01. Adj. = adjusted; NES = normalized enrichment score; Norm. = normalized; and SCFAs = short-chain fatty acids.