Pulse Crop Effects on Gut Microbial Populations, Intestinal Function, and Adiposity in a Mouse Model of Diet-Induced Obesity

Abstract

The dietary fiber gap that is present in many countries co-exists with a low intake of grain legumes (pulses) that have 2-3 times more dietary fiber than cereal grains that are commonly recommended to increase fiber intake. Given the relationships among dietary fiber, gut health and chronic disease risk, a study was undertaken in a preclinical mouse model for obesity to examine how commonly consumed pulses, i.e., chickpea, common bean, dry pea and lentil, would impact gut microbes, intestinal function, and adiposity. Pulses were fed to C57BL/6 mice at similar levels of protein and fiber. Bacterial count in the cecum was elevated 3-fold by pulse consumption. At the phylum level, a 2.2- to 5-fold increase in Bacteriodetes relative to Firmicutes was observed. For Akkermansia muciniphila, a health-beneficial bacterium, differential effects were detected among pulses ranging from no effect to a 49-fold increase. Significant differences among pulses in biomarkers of intestinal function were not observed. Pulses reduced accumulation of lipid in adipose tissue with a greater reduction in the subcutaneous versus visceral depots. Metabolomics analysis indicated that 108 metabolites were highly different among pulse types, and several compounds are hypothesized to influence the microbiome. These results support recent recommendations to increase consumption of pulse-based foods for improved health, although all pulses were not equal in their effects.

Keywords: Akkermansia muciniphila; Bacteriodetes; Firmicutes; adiposity; chickpea; common bean; dry pea; intestinal function; lentil; pulses.

Figure 1

Effect of feeding pulse diets on body weight. HF Control n = 7, LF Control n = 7, Bean n = 6, Chickpea n = 7, Dry Pea n = 8, Lentil n = 8; HF: high fat; LF: low fat. Values are means ± SEM. Differences among groups at each timepoint were not statistically significant.

Figure 2

Effect of feeding pulse diets on cecal bacteria populations. (A) 16S Ct for 20 ng genomic DNA, mean ± SEM; ANOVA:LF Control p < 0.0001, Bean p < 0.0001, Chickpea p < 0.0001, Dry Pea p < 0.0001, Lentil p < 0.0001; HF Control vs. all pulses p < 0.0001; LF Control vs. all pulses p = 0.0443 (B) Akkermansia muciniphilia ∆Ct, ANOVA: LF Control p = 0.0071, Bean p = 0.0001, Chickpea p = 0.9300, Dry Pea p = 0.2131, Lentil p < 0.0001; HF Control vs. all pulses p = 0.1342; LF Control vs. all pulses p = 0.0006 (C) Firmicutes ∆Ct, ANOVA: LF Control p = 0.7762, Bean p = 0.0370, Chickpea p = 0.1450, Dry Pea p = 0.8696, Lentil p < 0.0052; HF Control vs. all pulses p = 0.0937; LF Control vs. all pulses p = 0.1497 (D) Bacteroidetes ∆Ct, ANOVA: LF Control p = 0.0003, Bean p < 0.0001, Chickpea p = 0.0001, Dry Pea p < 0.0001, Lentil p < 0.0001; HF Control vs. all pulses p = < 0.0001; LF Control vs. all pulses p = 0.2189; FC relative to HF control diet; ^* different relative to HF control diet (p ≤ 0.05); ^# different relative to LF control diet (p ≤ 0.05); FC calculated relative to HF Control; HF Control n = 7, LF Control n = 7, Bean n = 6, Chickpea n = 7, Dry Pea n = 8, Lentil n = 8; FC: fold change; HF: high fat, LF: low fat.

Figure 3

Farnesoid X Receptor (FXR) expression. (A) 18S Ct, ANOVA: LF Control p = 0.0268, Bean p = 0.9537, Chickpea p = 0.3215, Dry Pea p = 0.5767, Lentil p = 0.2114; (B) FXR ∆Ct, ANOVA: LF Control p = 0.6239, Bean p = 0.1225, Chickpea p = 0.2717, Dry Pea p = 0.1044, Lentil p = 0.0787; ^* different relative to HF Control (p ≤ 0.05); ^# different relative to LF Control (p ≤ 0.05). Fold change was calculated relative to HF Control; ANOVA: HF Control vs. all pulses p = 0.05, LF Control vs. all pulses p = 0.01, HF Control vs. LF Control p = 0.26; HF Control n = 7, LF Control n = 7, Bean n = 6, Chickpea n = 7, Dry Pea n = 8, Lentil n = 8; HF: high fat; LF: low fat.

Figure 4

Metabolite variation among pulse types. Principal component analysis (PCA, top) and orthogonal projection to latent structures discriminant analysis (OPLS-DA, bottom) was performed on 17 varieties spanning four pulse types. The data matrix consisted of 2433 metabolites for n = 2 replicates per variety. Figures are scores plots for select components of the PCA and OPLS-DA.