Common Bean Suppresses Hepatic Ceramide Metabolism in a Mouse Model of Metabolic Dysfunction-Associated Steatotic Liver Disease

Abstract

Background/Objectives: The incidence of metabolic dysfunction-associated steatotic liver disease (MASLD), a condition linked to the ongoing obesity pandemic, is rapidly increasing worldwide. In turn, its multifactorial etiology is consistently associated with low dietary quality. Changing dietary macronutrient and phytochemical quality via incorporating cooked common bean into an obesogenic diet formulation has measurable health benefits on the occurrence of both obesity and hepatic steatosis in C57BL/6 mice. Methods: A cohort of C57BL/6 mice were randomized into experimental diets containing multiple dietary concentrations of common bean. The primary endpoint of this study was comparing metabolomic analyses from liver and plasma of different treatment groups. Additionally, RNA sequencing and protein expression analysis via nanocapillary immunoelectrophoresis were used to elucidate signaling mediators involved. Results: Herein, global metabolomic profiling of liver and plasma identified sphingolipids as a lipid subcategory on which bean consumption exerted significant effects. Of note, C16 and C18 ceramides were significantly decreased in bean-fed animals. Hepatic RNAseq data revealed patterns of transcript expression of genes involved in sphingolipid metabolism that were consistent with metabolite profiles. Conclusions: Bean incorporation into an otherwise obesogenic diet induces effects on synthesis, biotransformation, and degradation of sphingolipids that inhibit the accumulation of ceramide species that exert pathological activity. These effects are consistent with a mechanistic role for altered sphingolipid metabolism in explaining how bean inhibits the development of MASLD.

Keywords: ceramides; common bean; lipid metabolism; metabolic dysfunction-associated steatotic liver disease; metabolomics; serine palmitoyl transferase; sphingolipids.

Figure 1

Bean prevents the accumulation of lipid droplets in the liver of male animals in a dose-dependent manner. (a) Oil Red O- and hematoxylin-stained liver sections across different diets: percentage indicates the amount of total dietary protein derived from common bean. Extracellular and intracellular lipid is stained red and hepatic nuclei are stained blue. (b) Box plots of the bean dose effect on the hepatic lipid. Values represent the amount of lipid in mg normalized to g of dry liver weight across the diet groups. Groups indicate total dietary protein percent sourced by beans. Kruskal–Wallis testing showed significant differences in the diet effect (χ² = 36.09, p-value = 7.167 × 10⁻⁸) with the large effect size (η² = 0.435). Pairwise comparisons between the diet groups were conducted using the post-hoc Dunn test: *** q-value < 0.001; **** q-value < 0.0001 (adapted from [18]).

Figure 2

PCoAs of metabolomic profiles in liver (a) and plasma (b). Ordination was performed with the Bray–Curtis dissimilarity index, and 95% confidence intervals are shown as ovals.

Figure 3

Box plots of metabolites involved in de novo ceramide synthesis and resulting ceramides in liver. Differences indicated where * q-value < 0.05; ** q-value < 0.01; *** q-value < 0.001.

Figure 3

Box plots of metabolites involved in de novo ceramide synthesis and resulting ceramides in liver. Differences indicated where * q-value < 0.05; ** q-value < 0.01; *** q-value < 0.001.

Figure 4

Box plots of sphingosine and HCER metabolites involved in salvage pathway in liver. Differences indicated where * q-value < 0.05; ** q-value < 0.01; *** q-value < 0.001; **** q-value < 0.0001.

Figure 5

Box plots of selected sphingomyelin metabolites from the sphingomyelin hydrolysis pathway in liver. Differences indicated where * q-value < 0.05; ** q-value < 0.01.

Figure 6

PCoA of the selection of genes within the Ceramide Biosynthesis canonical pathway from IPA (de novo ceramide synthesis). Ordination was performed with the Bray–Curtis dissimilarity index, and 95% confidence intervals are shown as ovals. PERMANOVA testing indicated R² = 0.42, pseudo-F-value =12.876, and p-value = 0.001.

Figure 7

De novo ceramide biosynthesis pathway with an overlay of transcriptomic and metabolomic data. Blue indicates observed reduction, red—increase, gray—statistically insignificant result, and white—lack of observation in bean compared to control.

Figure 8

Overlook of ceramide metabolism and bean-induced effects. Ceramide metabolism includes degradation into sphingosine and biotransformation into cermide-1-phosphate, sphingomyelin, and hexosylceramide. Metabolon metabolites and RNAseq gene transcription data are overlaid: blue shows decrease while red shows increase in bean compared with the control.