Dietary short-chain fatty acid intake improves the hepatic metabolic condition via FFAR3

Fermented foods represent a significant portion of human diets with several beneficial effects. Foods produced by bacterial fermentation are enriched in short-chain fatty acids (SCFAs), which are functional products of dietary fibers via gut microbial fermentation. In addition to energy sources, SCFAs also act as signaling molecules via G-protein coupled receptors such as FFAR2 and FFAR3. Hence, dietary SCFAs in fermented foods may have a direct influence on metabolic functions. However, the detailed mechanism by dietary SCFAs remains unclear. Here, we show that dietary SCFAs protected against high-fat diet-induced obesity in mice in parallel with increased plasma SCFAs without changing cecal SCFA or gut microbial composition. Dietary SCFAs suppressed hepatic weight and lipid synthesis. These effects were abolished in FFAR3-deficient mice but not FFAR2-deficient. Thus, SCFAs supplementation improved hepatic metabolic functions via FFAR3 without influencing intestinal environment. These findings could help to promote the development of functional foods using SCFAs.

small intestine 25 . However, the detailed mechanisms underlying the difference in the metabolic benefits conferred by the direct intake of dietary SCFAs (such as through fermented foods) and those provided by gut microbes producing SCFAs from dietary fibers in the colon remain unclear.
To clarify these effects, in this study, we investigated the SCFAs-mediated systemic effects via circulating plasma upon dietary SCFA intake in a mouse model of high-fat diet (HFD)-induced obesity. We confirmed that dietary SCFA intake such as fermented foods mainly causes improved hepatic metabolic conditions via FFAR3 in HFD-induced obese mice.

Results
Dietary SCFA intake increases plasma SCFA levels. We firstly investigated changes in incorporated SCFAs level following SCFAs feeding in a mouse model of HFD-induced obesity. In this experiment, 7-weekold mice were fed HFDs containing 5% SCFAs (acetate, propionate, and butyrate, respectively), 5% cellulose or normal HFD as control for 4 weeks (Supplementary Table S1). Feeding of diets supplemented with each SCFA significantly increased the corresponding SCFA levels compared with those of the cellulose-supplemented and control HFD-fed mice ( Supplementary Fig. S1A), whereas there were no changes in the cecal SCFAs levels ( Supplementary Fig. S1B). Moreover, 16S rRNA gene amplicon sequencing confirmed that the SCFA diets did not alter the relative abundance of the major phyla constituting the gut microbiota ( Supplementary Fig. S1C) or the gut microbiota composition, as indicated by principal coordinates analysis with reference to taxonomic datasets ( Supplementary Fig. S1D).
Dietary SCFA intake exerts metabolic benefits. We next investigated changes in metabolic parameters following SCFAs feeding. HFD feeding to adult mice for 4 weeks can sufficiently induce obesity and related metabolic disturbances as compared to that in normal chow-fed mice ( Fig. 1A-F) [41][42][43] . Body weights in the SCFAs-fed mice were significantly lower than those in the cellulose-fed and control mice during growth (Fig. 1A). The white adipose tissue (WAT) mass and liver weights (Fig. 1B), blood glucose levels (Fig. 1C), and plasma insulin levels ( Fig. 1D) were also significantly lower in the SCFA groups than those in the cellulose-fed and control mice at 11 weeks of age. Furthermore, HFD-induced insulin resistance and impaired glucose tolerance, as determined by the glucose tolerance test (GTT) and insulin tolerance test (ITT), respectively, were significantly attenuated in SCFA-fed mice as compared to those in cellulose-fed and control mice (Fig. 1E,F). However, the plasma levels of the gut hormones glucagon like peptide-1 (GLP-1) and peptide YY (PYY) were similar among all groups, although plasma GLP-1 tended to be slightly higher in the SCFA-fed groups (Fig. 1G,H). Food intake was also similar among all groups ( Supplementary Fig. S2).
Dietary SCFA intake improves hepatic metabolic conditions. We further examined changes in lipid metabolism. Although plasma triglycerides (TGs), non-esterified fatty acids (NEFAs), and total cholesterol levels were comparable between each SCFAs-fed group and the cellulose-fed or control group ( Fig. 2A), hepatic TG levels in each group of SCFA-fed mice were significantly lower than those in cellulose-fed and control mice (Fig. 2B). We also investigated the expression profiles of hepatic genes related to energy metabolism. The expression levels of Fas and Chrebp-which are related to fatty acid synthesis-decreased, while that of Ppara-as a key regulator of lipid metabolism with acetate and propionate feeding-was increased in the SCFAs-fed mice compared to those of cellulose-fed and control mice (Fig. 2C). However, these changes in genes related to energy metabolism were not observed in the WAT or muscle ( Supplementary Figs S3 and 4). Thus, dietary SCFA intake suppressed the HFD-induced accumulation of hepatic TGs via changing hepatic lipid metabolism but not the WAT and muscle metabolism.
FFAR3 deficiency abolishes the metabolic benefits of dietary SCFA intake. SCFAs exhibit various physiological functions related to energy regulation 13 . Therefore, we next investigated the roles of the SCFA receptors FFAR2 and FFAR3 in the observed metabolic improvement by increased plasma SCFAs from dietary SCFA intake using SCFA receptor-deficient mice. Before HFD feeding, initial body and tissue weights, and blood glucose, as well as hepatic lipid metabolic related genes such as Fas, Chrebp, and Ppara were similar among three groups ( Supplementary Fig. S5). Although food intake was comparable among the three groups ( Supplementary Fig. S2), the suppression of HFD-induced weight gain by SCFAs supplementation was attenuated in Ffar3 −/− mice, especially with butyrate feeding, whereas the effects in Ffar2 −/− mice were comparable with those in wild-type mice (Fig. 3A). Moreover, although the HFD-induced WAT mass increase was suppressed by the SCFAs in both Ffar3 −/− and Ffar2 −/− mice, similar to the wild-type mice, the suppression of HFD-induced liver weight gain by SCFAs feeding was abolished in the Ffar3 −/− mice but not in the Ffar2 −/− mice (Fig. 3B). The changes in blood glucose by SCFAs feeding were also completely abolished in Ffar3 −/− mice and were attenuated in Ffar2 −/− mice (Fig. 3C). Although basal plasma insulin levels in HFD-fed mice were lower in Ffar3 −/− mice and higher in Ffar2 −/− mice, the SCFAs suppressed these changes in both Ffar2 −/− and Ffar3 −/− mice, similar to the effects observed in wild-type mice (Fig. 3D). These results indicated that the functions of FFAR3 in the liver are partially related to the metabolic benefits from dietary SCFA intake.
www.nature.com/scientificreports www.nature.com/scientificreports/ Finally, we investigated whether SCFAs influence hepatic lipid metabolism under direct stimulation rather than as a secondary long-term effect. Intraperitoneal administration of propionate, as the most potent agonist for FFAR3, transiently increases plasma propionate levels 37 . Similar to the effects of long-term dietary SCFAs feeding, www.nature.com/scientificreports www.nature.com/scientificreports/ were almost completely abolished in Ffar3 −/− mice (Fig. 5B). These results confirmed that dietary SCFA intake improves hepatic metabolic conditions via plasma SCFA-stimulated FFAR3.

Discussion
Dietary fiber and its gut microbial metabolite SCFAs are well known to exert metabolic benefits to the host, and the mechanisms have been extensively examined. However, the underlying mechanism of the metabolic benefits provided by dietary SCFA intake, such as with fermented food enriched in SCFAs, are less well understood. Here, we demonstrate that dietary SCFA intake increases plasma SCFA levels to active FFAR3 and improves hepatic metabolic conditions without changing the intestinal environment.
Indeed, dietary SCFA intake increased plasma SCFAs levels but not cecal SCFAs levels without changing the gut microbial compositions in HFD-fed mice. Consequently, dietary SCFA intake dramatically suppressed the HFD-induced liver and WAT weight gains, without influencing gut hormones. This result indicates that direct intake of dietary SCFAs such as consumption of fermented foods also exhibits metabolic benefits. Although these effects appear to be similar to those of dietary fiber, the mechanism may differ to some degree. For example, dietary SCFA intake did not sufficiently change plasma gut hormone levels, although plasma GLP-1 tended to be slightly higher in the SCFA-fed groups, because almost all SCFAs are absorbed in the small intestine and they are not readily transfered to the colon 25 , whereas dietary fiber intake produces SCFAs via fermentation by gut www.nature.com/scientificreports www.nature.com/scientificreports/ microbiota mainly in the colon. L-cells producing GLP-1 and PYY are localize mainly in the distal ileum and colon. Therefore, direct SCFA intake might mainly exert systemic effects via the plasma SCFAs rather than via distal intestinal SCFAs. Further investigation is needed to test this hypothesis and clarify the difference between the effects of direct SCFA intake and gut microbiota-produced SCFAs.
We found that dietary SCFA intake suppressed the HFD-induced liver weight gain and hepatic TGs accumulation along with a change in hepatic lipid metabolism-related genes, and these effects were abolished by FFAR3 deficiency but not FFAR2 deficiency. Similarly, a previous study showed that dietary SCFA intake suppressed the synthesis of hepatic fatty acids 44 . Hence, we concluded that dietary SCFA intake improves hepatic metabolic conditions via FFAR3. However, since FFAR3 is barely expressed in the liver 36 , it is more likely that plasma SCFAs indirectly influence hepatic lipid metabolic-related genes such as Fas and Ppara in the liver via other FFAR3-expressing tissues. Given the high expression of FFAR3 in the peripheral nerves 36 , FFAR3-mediated neural activity might explain the observed improvement of hepatic metabolic conditions by SCFAs. The FFAR3-FFAR2 heteromer and species difference influence different intracellular signaling pathways 45,46 and therefore, it might also exert other physiological effects. To further clarify this FFAR3-mediated molecular mechanism, further experiments are needed with tissue-specific FFAR3-deficient mice.
SCFAs are known to influence insulin actions via their receptors 47,48 . Indeed, the plasma insulin levels in HFD-fed control and cellulose-supplemented mice were already drastically lower at baseline in Ffar3 −/− mice compared with those of wild-type mice. Hence, this difference might partially explain why the protection of dietary SCFA intake against HFD-induced obesity and hyperglycaemia was abolished in the Ffar3 −/− mice. However, acute SCFAs intraperitoneal administration in Ffar3 −/− mice also abolished the changes of SCFAs-induced hepatic lipid metabolism-related genes observed in wild-type mice. Accordingly, improvement of hepatic metabolic conditions by SCFAs appears to be due, at least in part, to the direct effects via FFAR3. Moreover, the protective effect of dietary SCFA intake against HFD-induced WAT weight gain was not abolished in both Ffar3 −/− and Ffar2 −/− mice. This indicates that other SCFA receptors beside FFAR2 and FFAR3, SCFA-mediated bioactivities, or the confounding effects of interactions between FFAR2 and FFAR3 might be related to the SCFA-mediated suppression of adipose fat accumulation. To further clarify this SCFA and receptor-mediated molecular mechanism, further experiments are also needed with FFAR2/FFAR3-double-deficient mice. www.nature.com/scientificreports www.nature.com/scientificreports/ used to determine the relative expression levels with the mRNA levels of the housekeeping 18S gene as reference. The primer sequences are shown in Supplementary Table S2. Statistical analysis. All values are presented as the mean ± SEM. The statistical significance of differences between groups was determined using a two-tailed unpaired Student's t-test (two groups) or two-tailed one-way analysis of variance, followed by Dunnett's post hoc test (≥three groups). To estimate the required sample size for each experiment, a priori power analysis was performed using the G*Power software ver3.1 (Franz Faul, Universiät Kiel, Germany; http://www.gpower.hhu.de/). Sample size used in the present study will achieve 95% actual power to detect an effect of each experiment assuming a one-way ANOVA with a 0.05 significance level. Except for the power analysis, all experimental data analyses were performed using GraphPad Prism 7.0 (Graphpad Software, San Diego, CA, USA).

Data availability
The raw 16S rRNA sequence data have been deposited into the DNA Data Bank of Japan (DDBJ) database under accession no. DRA008619 [https://ddbj.nig.ac.jp/DRASearch/submission?acc = DRA008619].