Excessive diet consumption induces obesity, type 2 diabetes mellitus and the metabolic syndrome, risk factors associated with non-alcoholic fatty liver disease (NAFLD), which includes fatty liver diseases in individuals without significant alcohol consumption, ranging from fatty liver to steatohepatitis and cirrhosis(Reference Chalasani, Younossi and Lavine1). The global prevalence of NAFLD in patients with type 2 diabetes mellitus is 55·5 %(Reference Younossi, Golabi and de Avila2), while 29 and 25 % of the entire population in Japan(Reference Eguchi, Hyogo and Ono3) and the world(Reference Younossi, Koenig and Abdelatif4), respectively, are currently believed to have NAFLD. NAFLD can progress to cirrhosis and is the most rapidly growing cause of hepatocellular carcinoma(Reference Younossi, Stepanova and Ong5). Thus, in order to prevent the other liver diseases, there is an urgent need to find strategies that reduce the risk of NAFLD.
In rat experiments, excessive energy consumption in the form of a high-fat diet enhances the secretion of 12α-hydroxylated (12αOH) bile acid (BA)(Reference Yoshitsugu, Kikuchi and Iwaya6). In addition, 12αOH BA concentrations in the liver, portal blood and faeces are associated with hepatic lipid accumulation(Reference Hori, Abe and Lee7). Such an association has also been reported between 12αOH BA and fatty liver index in a clinical study(Reference Haeusler, Astiarraga and Camastra8). BA synthesised from cholesterol (Chol) are secreted into the duodenum. Cholic acid (CA; 12αOH BA) enhances the absorption of dietary Chol, but chenodeoxycholic acid (non-12αOH BA) does not(Reference Uchida, Nomura and Takeuchi9), which suggests that 12αOH BA contribute to the absorption of hydrophobic nutrients. In a previous study carried out by our group on rats(Reference Lee, Shimizu and Hagio10), dietary supplementation of the diet with 0·5 g/kg CA increased the proportion of 12αOH BA in the portal plasma, liver, caecal contents and faeces. Also, the dietary CA supplementation in rats led to various abnormalities, such as the development of hepatic steatosis without obesity, hepatic injury and dysbiosis in the gut microbiota. These observations suggest that an increase in the levels of 12αOH BA in the body associates hepatic steatosis with related diseases.
In addition to dietary energetic restriction and exercise, oligosaccharides also may serve as an effective way to prevent hepatic steatosis. Raffinose (Raf), an indigestible oligosaccharide, is a trisaccharide composed of galactose, glucose and fructose (β-D-fructofuranosyl-O-α-D-galactopyranosyl-(1,6)-α-D-glucopyranoside). Dietary Raf increases the proportion of Bifidobacterium, while reducing that of Clostridium in the microbiota of rats(Reference Dinoto, Suksomcheep and Ishizuka11), which may contribute to reducing the conversion of BA to toxic secondary BA(Reference Hayakawa and Hattori12). An increase in Bifidobacterium upon oral Raf administration has also been confirmed in humans(Reference Dinoto, Suksomcheep and Ishizuka11). Some indigestible oligosaccharides increase BA excretion(Reference Garcia-Diez, Garcia-Mediavilla and Bayon13,Reference Hoving, Katiraei and Heijink14) , and a diet that includes an unpurified Raf-containing supplement lowers hepatic BA concentration as well as hepatic TAG concentration in rats fed a high-fat diet(Reference Dai, Feng and Liu15). However, the mechanism in which dietary Raf reduces hepatic TAG concentration remains unknown. In our previous study, we observed association between hepatic TAG concentration and 12αOH BA levels in enterohepatic circulation(Reference Hori, Abe and Lee7). Those observations suggest that dietary Raf reduces hepatic TAG via modulation of BA metabolism in enterohepatic circulation.
We investigated whether Raf ingestion alleviates hepatic lipid accumulation by modulating BA metabolism in a novel steatosis model with dietary CA supplementation.
Methods
Animal and diets
The animal experiments were approved by the Institutional Animal Care and Use Committee of the National Corporation Hokkaido University (approval number: 17-0119), and all the animals were maintained according to the Hokkaido University Manual for Implementing Animal Experimentation. Male Wistar rats (n 24; 3 weeks old; Japan SLC Inc.) were individually housed in a controlled environment at 22 (sem 2)°C and 55 (sem 5) % humidity. The light period was from 08.00 to 20.00 hours.
The rats had free access to food and water for the entire study period. After acclimation with a control diet based on AIN-93G(Reference Reeves, Nielsen and Fahey16) for 2 weeks, the rats were divided into four dietary groups (n 6 each): control, CA supplementation (CA; 0·5 g CA/kg diet), Raf supplementation (Raf; 30 g Raf/kg diet) and a combined supplementation with CA and Raf (CA + Raf; 0·5 g CA/kg diet and 30 g Raf/kg diet) to have comparable average body weights based on measurements on the day when the test diet was initiated, as shown in Table 1. In this study, the investigators and outcome assessors were not blinded to the intervention. We confirmed that deoxycholic acid (DCA) is the major form of 12αOH BA in human faeces(Reference Hashimoto, Matsui and Ishizuka17). The range of DCA concentration is from 0·07 to 0·73 mmol/l in human faecal water on a high-fat diet(Reference Bernstein, Bernstein and Payne18), and it is possible to damage bacterial membrane in the DCA levels(Reference Kurdi, Kawanishi and Mizutani19). In our previous observation(Reference Islam, Fukiya and Hagio20), DCA concentration was calculated from 0·07 to 0·98 µmol/g caecal contents in rats fed diet supplemented with CA at 0·5 g/kg diet. In the present study, we used this CA supplementation level to mimic BA environment in those observations. A standard supplementation level of dietary fibre is 50 g/kg diet according to AIN-93 formulation(Reference Reeves, Nielsen and Fahey16). We observed increase in the proportion of bifidobacteria accompanied by reduction of pH in caecal contents and an increase in caecal tissue weight in rats fed a diet supplemented with Raf at 30 g/kg and crystalline cellulose at 50 g/kg diet (80 g fibre source/kg diet in total)(Reference Dinoto, Suksomcheep and Ishizuka11,Reference Ishizuka, Iwama and Dinoto21) . The supplementation levels of indigestible saccharides correspond daily dietary fibre intake in humans (0·5–8·0 g/MJ diet)(Reference Janzi, Dias and Martinsson22).
Table 1. Diet compositions
* NZMP Acid Casein (Fonterra Co-Operative Group Limited).
† TK-16 (Matsutani Chemical Industry Co. Ltd).
‡ Nippon Beet Sugar Manufacturing Co. Ltd.
§ J-Oil Mills, Inc.
|| Crystalline cellulose (Ceolus PH-102, Asahi Kasei Chemicals Corp.).
¶ AIN-93G Mineral mixture.
** AIN-93 Vitamin mixture.
†† Wako Pure Chemical Industries, Ltd.
Body weights and food intake were measured every 2 d, and the entire experimental period was 4 weeks long. Faecal samples were collected from days 27 to 28 and stored at −30°C for analysis of BA composition and energy content. At the end of the experimental period, portal and aortic blood were collected under anaesthesia via an intraperitoneal injection of sodium pentobarbital (50 mg/kg body weight, Somnopentyl, Kyoritsuseiyaku Corporation) and added to heparin (final concentration at 50 U/ml) and aprotinin (final concentration at 500 KIU/ml). After the rats were euthanised by exsanguination, the caecum was collected and the weights of the whole caecum and the caecal contents were measured. The caecal contents were diluted four times with deionised water and homogenised with a Teflon homogeniser. The pH levels of the homogenates were measured using a semiconducting electrode (Argus ISFET pH sensor; Sentron). Supernatants of the caecal contents were kept in liquid N2 and stored at −80°C for organic acid analysis. Ileal contents and mucosa were collected and stored at −30°C and −80°C for BA analysis and real-time quantitative PCR, respectively. Liver and epididymal adipose tissues were collected and weighed. Blood plasma was separated using centrifugation at 2000 g for 10 min at 4°C. The plasma and liver samples were kept in liquid N2 and stored at −80°C until analysis.
Biochemical analyses
Lipid extraction from liver and faecal samples was performed as previously described(Reference Hori, Hara and Ishizuka23). The activities of plasma aspartate aminotransferase and alanine aminotransferase were measured using a transaminase CII-test Wako kit (Wako Pure Chemical Industries). Chol and TAG levels were measured using a Cholesterol E-test kit and Triglyceride E-test kit (both from Wako), respectively. Hepatic NEFA concentration was measured using a NEFA C-test kit (Wako).
Bile acid analysis
BA extraction and HPLC separation were performed as previously described(Reference Hagio, Matsumoto and Fukushima24–Reference Lee, Hori and Kohmoto26). Individual BA concentration was measured using nordeoxycholic acid (23-nor-5β-cholanic acid-3α,12α-diol) as an internal standard.
The 12αOH BA measured in this experiment were as follows: CA (5β-cholanic acid-3α,7α,12α-triol), DCA (5β-cholanic acid-3α,12α-diol), ursocholic acid (5β-cholanic acid-3α,7β,12α-triol), taurocholic acid (5β-cholanic acid-3α,7α,12α-triol-N-(2-sulfoethyl)-amide), taurodeoxycholic acid (5α-cholanic acid-3α,12α-diol-N-(2-sulfoethyl)-amide), glycocholic acid (5β-cholanic acid-3α,7α,12α-triol-N-(carboxymethyl)-amide), glycodeoxycholic acid (5β-cholanic acid-3α,12α-diol-N-(carboxymethyl)-amide), 7-oxo-deoxycholic acid (5β-cholanic acid-3α,12α-diol-7-one), 12-oxo-lithocholic acid (12oLCA; 5β-cholanic acid-3α-ol-12-one) and 5β-cholanic acid-12α-ol-3-one (3o12α).
The non-12αOH BA measured in this experiment were as follows: α-muricholic acid (5β-cholanic acid-3α,6β,7α-triol), β-muricholic acid (5β-cholanic acid-3α,6β,7β-triol), ω-muricholic acid (5β-cholanic acid-3α,6α,7β-triol), chenodeoxycholic acid (5β-cholanic acid-3α,7α-diol), hyocholic acid (5β-cholanic acid-3α,6α,7α-triol), hyodeoxycholic acid (5β-cholanic acid-3α,6α-diol), ursodeoxycholic acid (5β-cholanic acid-3α,7β-diol), lithocholic acid (LCA; 5β-cholanic acid-3α-ol), tauro-α-muricholic acid (5β-cholanic acid-3α,6β,7α-triol-N-(2-sulfoethyl)-amide), tauro-β-muricholic acid (5β-cholanic acid-3α,6β,7β-triol-N-(2-sulfoethyl)-amide), tauro-ω-muricholic acid (5β-cholanic acid-3α,6α,7β-triol-N-(2-sulfoethyl)-amide), taurochenodeoxycholic acid (5β-cholanic acid-3α,7α-diol-N-(2-sulfoethyl)-amide), taurohyodeoxycholic acid (5β-cholanic acid-3α,6α-diol-N-(2-sulfoethyl)-amide), taurolithocholic acid (5β-cholanic acid-3α-ol-N-(2-sulfoethyl)-amide), tauroursodeoxycholic acid (5β-cholanic acid-3α,7β-diol-N-(2-sulfoethyl)-amide), glycochenodeoxycholic acid (5β-cholanic acid-3α,7α-diol-N-(carboxymethyl)-amide), glycohyodeoxycholic acid (5β-cholanic acid-3α,6α-diol-N-(carboxymethyl)-amide), glycoursodeoxycholic acid (5β-cholanic acid-3α,7β-diol-N-(carboxymethyl)-amide), glycolithocholic acid (5α-cholanic acid-3α-ol-N-(carboxymethyl)-amide) and 7-oxo-lithocholic acid (5β-cholanic acid-3α-ol-7-one).
All BA, except for ursocholic acid, were purchased from Steraloids Inc. Ursocholic acid was obtained from Toronto Research Chemicals.
Real-time quantitative PCR
mRNA expression was measured in the liver and ileal mucosa using real-time quantitative PCR(Reference Hori, Hara and Ishizuka23). Total RNA was extracted using an RNeasy® Mini Kit (Qiagen) according to the manufacturer’s instructions, and the RNA concentration was measured using a NanoDrop™ Lite Spectrophotometer (Thermo Scientific). Complementary DNA was synthesised from 1 μg of RNA using ReverTraAce® quantitative PCR RT master mix with gDNA remover (Toyobo Co. Ltd), according to the manufacturer’s instructions. Quantitative PCR was performed on an Mx3000P™ Real-Time PCR system (Stratagene) using the SYBR® Green method or TaqMan® method, as described previously(Reference Hori, Hara and Ishizuka23). The SYBR® Green method was performed with specific primer pairs for ribosomal protein lateral stalk subunit P0 (Rplp0, forward: 5′-GGCAAGAACACCATGATGCG-3′; reverse: 5′-GTGATGCCCAAAGCTTGGAA-3′, 5 μm), apical Na-dependent bile acid transporter (Asbt, SLC10A2; forward: 5′-GGTGACATGGACCTCAGTGTT-3′; reverse: 5′-GTAGGGGATCACAATCGTTCCT-3′, 10 μm), ATP-binding cassette subfamily C member 2 (Abcc2, MRP2; forward: 5′-GCTGGTTGGAAACTTGGTCG-3′; reverse: 5′-CAGGGCGCCTCATTCTCTAC-3′, 10 μm), organic solute transporter subunit α (Osta, SLC51A; forward: 5′-CCCTCATACTTACCAGGAAGAAGCTAC-3′; reverse: 5′-CCATCAGGAATGAGAAACAGGC-3′, 10 μm) and organic solute transporter subunit β (Ostb, SLC51B; forward: 5′-TAT TCCATCCTGGTTCTGGCAGT-3′; reverse: 5′-CGTTGTCTTGTGGCTGCTTCTT-3′, 10 μm). Relative expression levels were calculated for each sample after normalisation to Rplp0 as a reference gene using the standard curve method. The TaqMan® Gene Expression Assays (Thermo Fisher Scientific) used in this study were as follows: Rn03302271_gH for RPLP0, Rn01495769_m1 for sterol regulatory element binding protein-1 (Srebp1), Rn01463550_m1 for fatty acid synthase (Fasn), Rn04181355_m1 for cell death-inducing DNA fragmentation factor-α-like effector A (Cidea) and Rn00580702_m1 for carnitine O-palmitoyltransferase 1 (Cpt1). Relative expression levels of these target mRNAs were calculated for each sample, as shown in the SYBR® Green method.
Organic acid analysis
Organic acids in the caecal contents were measured using HPLC (Shimadzu) with crotonic acid (Wako Pure Chemical Industries) as an internal standard, according to the method of Hoshi(Reference Hoshi, Sakata and Mikuni27). The caecal contents were homogenised and neutralised with sodium hydroxide to prevent the extraction of SCFA. Fat-soluble substances in the supernatant were removed using chloroform, and the aqueous phase was passed through a membrane filter (cellulose acetate, 0·20 μm pore size; DISMIC-13cp; Toyo Roshi Kaisha). The samples were analysed using an HPLC system (Shimadzu). The equipment of the HPLC system consisted of LC-20AD pump, SIL-20AC auto-sampler and CDD-6A electro-conductivity detector. The mobile phase was 5 mmol/l p-toluenesulfonic acid, while the detection solution was 5 mmol/l p-toluenesulfonic acid containing 100 μmol/l EDTA and 20 mmol/l bis-tris.
Bomb calorimeter analysis
The faecal energy content was determined by combustion of approximately 0·3 g of dry faeces in a CA-4PJ automatic adiabatic bomb calorimeter (Shimadzu).
Statistical analyses
To determine an adequate sample size to identify significant differences in hepatic TAG concentration, a power analysis based on the experimental design was performed using G*Power (version 3.1.9.7)(Reference Faul, Erdfelder and Lang28). Based on an α probability of 0·05 and a power of 0·80, the effect size was estimated using the variance by special effects and error variances in an experiment (n 10 per group)(Reference Shimoda, Hori and Maegawa29) for hepatic TAG concentration. The required sample size per group was determined to be 4–10 per group. All data have been presented as mean values with their standard errors. All the data were evaluated for normality and homogeneity of variance using the Shapiro–Wilk and Brown–Forsythe tests, respectively. The data were log-transformed to improve the homogeneity of the variance before ANOVA if necessary. Two-way ANOVA was used to evaluate significant interaction between the two factors, CA and Raf. Tukey–Kramer test was applied as a post-hoc test if a significant interaction was observed (CA × Raf). Multiple regression analysis was performed to identify predictors in parameters with significant differences in main effects of two-way ANOVA. Statistical analysis was performed using JMP version 14.0 (SAS Institute Inc.).
Results
To confirm whether Raf ingestion improves CA-induced steatosis and hepatic injury, we analysed parameters involved in hepatic steatosis. As shown in our previous study(Reference Lee, Shimizu and Hagio10), CA feeding induced hepatic lipid accumulation (Fig. 1(a)) and increased plasma Chol, aspartate aminotransferase and alanine aminotransferase levels (Fig. 1(b) and (c)). In contrast, a reduction in plasma TAG was observed in the 4-week experiment (Fig. 1(b)). Notably, Raf ingestion normalised hepatic TAG and FFA concentrations in the CA-supplementation group (Fig. 1(a)). On the other hand, Raf diet did not affect liver Chol or plasma parameters (TAG, Chol and transaminase activities) (Fig. 1(a)–(c)). There were no significant differences in terms of total food intake, final body weights and epididymal adipose tissue weights between the groups (Table 2).
Fig. 1. Liver lipid and plasma parameters of rats fed control, cholic acid (CA), raffinose (Raf) or CA + Raf diet for 4 weeks. Liver lipids have been shown as (a) liver TAG, NEFA and cholesterol (Chol) concentrations. Plasma levels of (b) TAG, Chol and (c) transaminase activities (aspartate aminotransferase (AST) and alanine aminotransferase (ALT)). Parameters of rats fed control and CA diets have been shown using blank and filled symbols, respectively. Parameters of rats fed diets with and without Raf supplementation have been shown in black and grey, respectively. The values have been expressed as mean values with their standard errors (n 5–6). P-values of two-way ANOVA for CA and Raf have been shown in the inset tables. Different superscript letters represent significant differences among the groups (P < 0·05), obtained using Tukey–Kramer’s test.
Table 2. Growth parameters and tissue weights in the rats fed the experimental diets for 4 weeks (Mean values with their standard errors)
Thirty molecular species of BA in the liver, ileal contents, blood plasma and faeces were analysed using LC-MS at the end of the experimental period (Fig. 2). CA supplementation increased the concentrations of taurocholic acid and taurodeoxycholic acid in the liver. Two-way ANOVA revealed an interaction between CA and Raf in determination of the hepatic taurocholic acid concentration (Fig. 2(a)). The same was however not observed in the ileal contents (Fig. 2(b)). In case of portal and aortic plasma, alterations in CA concentrations were similar to those observed in the liver (Fig. 2(c) and (d)). CA supplementation increased the faecal DCA concentration, but a subtle influence of Raf diet was found on the composition of 12oLCA and 3o12α (Fig. 2(e)). There was an increase in faecal wet weight per day upon Raf supplementation (Table 2).
Fig. 2. Bile acid (BA) composition in samples from rats fed control, cholic acid (CA), raffinose (Raf) or CA + Raf diet for 4 weeks. Concentration of each BA molecule in (a) liver, (b) ileal contents, (c) portal plasma, (d) aortic plasma and (e) faeces. Concentrations of 12α-hydroxylated (12αOH) BA in (f) liver, ileal contents, portal plasma, aortic plasma and faeces. (g) Pearson’s correlation between the concentrations of portal 12αOH BA and liver TAG. Parameters of rats fed control and CA diets have been shown using blank and filled symbols, respectively. Parameters of rats fed diets with and without Raf supplementation have been shown in black and grey, respectively. The values have been expressed as mean values with their standard errors (n 5–6). P-values of two-way ANOVA for CA and Raf have been shown in the inset tables. Different superscript letters represent significant differences among the groups (P < 0·05), obtained using Tukey–Kramer’s test.
We also determined 12αOH BA in the liver, ileal contents, plasma and faeces (Fig. 2(f)). Ingestion of Raf diet partially reduced the 12αOH BA concentration in the livers of CA-fed rats. On the other hand, a high level of 12αOH BA was observed in the ileal contents of the CA + Raf-fed rats. The compositions of 12αOH BA in the portal and aortic plasmas were similar to those in the liver. In case of faecal BA, the increase in 12αOH BA upon CA supplementation was not altered upon further Raf supplementation. Excreted amount of 12αOH BA in the faeces was tended to be increased as follows: 3·05 (sem 0·50) µmol/d in control, 12·64 (sem 1·25) µmol/d in CA, 3·43 (sem 0·30) µmol/d in Raf and 15·14 (sem 2·22) µmol/d in CA + Raf (two-way ANOVA P-values were < 0·0001 for CA, NS for Raf and NS for CA × Raf; NS). It is possible that an enhancement of 12αOH BA excretion in the urine although we have not measured in the present study. We consider that Raf supplementation enhanced faecal 12αOH BA excretion. We deduced a relationship between 12αOH BA and hepatic lipid accumulation and examined the association between 12αOH BA and liver TAG using Pearson’s correlation analysis. The portal plasma 12αOH BA concentration correlated with the liver TAG concentration (Fig. 2(g); R 2 = 0·52, P < 0·001). Similar correlations of liver TAG concentration were found with the hepatic 12αOH BA concentration (R 2 = 0·49, P < 0·001) or aortic 12αOH BA concentration (R 2 = 0·47, P < 0·001).
Faecal lipids and faecal energy content were analysed. A significant increase in TAG was found upon Raf supplementation, while a reduction in faecal FFA was observed upon CA supplementation (Fig. 3(a)). Faecal energy excretion correlated with the faecal weight (P < 0·001, R 2 = 0·91). There was no difference in faecal Chol excretion upon the dietary interventions (Fig. 3(a)). Notably, Raf ingestion increased faecal energy excretion (Fig. 3(a)). We measured the levels of caecal organic acids to find an association between hepatic steatosis and caecal fermentation (Fig. 3(b)). Raf supplementation increased the concentrations of acetate and succinate (Fig. 3(b)), accompanied by a reduction in the caecal pH (Table 2). There was a significant increase in the weights of the whole caecum, caecal contents and faeces upon Raf ingestion (Table 2). We conducted multiple linear regression analysis to determine predictors of liver TAG and found that the concentrations of aortic 12αOH BA and caecal acetic acid could serve as predictors of hepatic steatosis (Fig. 3(c), inset table).
Fig. 3. Faecal lipid excretion, energy content, caecal parameters and multiple regression analysis of rats fed control, cholic acid (CA), raffinose (Raf) or CA + Raf diets for 4 weeks. (a) Faecal excretions of TAG, NEFA, cholesterol (Chol) and energy content per d were calculated from faecal weight collected for 24 h at the end of the experiment. (b) Organic acid concentrations. (c) Predictors of liver TAG concentration. Parameters in rats fed control and CA diets have been shown using blank and filled symbols, respectively. Parameters in rats fed diets with and without Raf supplementation have been shown in black and grey, respectively. The values have been expressed as mean values with their standard errors (n 5–6). P-values of two-way ANOVA for CA and Raf have been shown in the inset table.
No significant differences were observed in the hepatic mRNA expression levels of Srebp1, Fasn, CideA and Cpt1 (data not shown). In case of ileal mRNA expression of BA transporters responsible for enterohepatic BA circulation (Fig. 3), CA supplementation resulted in an increase in Abcc2 expression. In addition, an interaction between CA and Raf was observed in Asbt expression, although no significant difference was observed in a post hoc test. There were no significant differences in the expression levels of Osta and Ostb in the ileal mucosa (Fig. 3).
Discussion
In this study, we investigated whether dietary Raf suppresses hepatic lipid accumulation induced by the CA supplementation to diet(Reference Lee, Shimizu and Hagio10) based on a report that shows Raf modulates BA metabolism in mice fed a high-fat/western-style diet(Reference Dai, Feng and Liu15). We successfully detected the CA-induced hepatic TAG accumulation in week 4 and also observed dietary Raf normalised hepatic TAG accumulation in the CA-fed rats accompanied by reduced 12αOH BA concentrations in liver and portal plasma.
In our CA-feeding hepatic steatosis model(Reference Lee, Shimizu and Hagio10), the nutrient composition of the CA-supplemented diet was almost the same as that of the control diet, other than the addition of CA at 0·5 g/kg diet. In addition, there was no alteration in food consumption and body weight. Nevertheless, the CA diet induces hepatic steatosis without obesity and liver inflammation(Reference Lee, Shimizu and Hagio10), which may represent an early stage of fatty liver. Such simple fatty liver without obesity is also observed in humans(Reference Kojima, Watanabe and Numata30). In order to find a way to prevent NAFLD, it is important to identify what happens at the onset of steatosis that can be easily modified with a simple dietary intervention. Based on this novel fatty liver model, we investigated whether the use of oligosaccharides that affect BA metabolism leads to the prevention of hepatic steatosis. The present study showed that dietary raffinose reduced hepatic lipid accumulation via suppression of enterohepatic 12αOH BA circulation.
Since de novo lipogenesis is an important event in CA-induced steatosis, we analysed mRNA expression of genes closely associated with de novo lipogenesis. However, no significant alteration was detected in the expression levels of Srebp1 and Fasn, even in the livers of CA-fed rats. From another point of view, the involvement of dietary lipids is also of interest in steatosis. To ascertain a balance of dietary fat, we measured the faecal excretions of energy and lipids such as TAG, FFA and Chol (Fig. 3). There was no difference in faecal energy excretion between the control and CA diet groups, which suggests that the retention of dietary fat is almost comparable between the groups. On the other hand, enhanced faecal energy excretion upon supplementation with dietary Raf is another possible explanation for the partial amelioration of dyslipidaemia by Raf. However, in the present study, this effect may not contribute to hepatic lipid normalisation by dietary Raf in the CA-fed condition because alteration of Raf-induced energy excretion was not consistent with hepatic TAG accumulation and simply correlated with faecal weight. Both de novo lipogenesis and dietary fat were not clearly associated with changes in hepatic TAG concentration in this study, which is probably due to the short term of the experiments, which makes it difficult to detect some subtle effects sharply. In other words, 12αOH BA metabolism is a sensitive parameter that reflects hepatic TAG concentration.
Modulation of enterohepatic 12αOH BA circulation is presumably a key event in Raf-mediated prevention of CA-induced hepatic steatosis. Raf feeding significantly reduced the portal CA concentration, while increasing the taurocholic acid concentration in ileal contents, especially under the CA-fed condition, which suggests suppression of enterohepatic 12αOH BA circulation. In addition, 12αOH BA in the portal blood was associated with liver TAG concentration. Enterohepatic BA circulation is mediated by ileal BA transporters(Reference Dawson, Haywood and Craddock31). ASBT is the major transporter in ileal epithelial cells that incorporates BA from the luminal contents, while ABCC2 is another transporter that secretes BA from the enterocyte to the lumen(Reference Dawson and Karpen32). Incorporated BA are secreted from the enterocytes to the portal blood via organic solute transporter subunit α and organic solute transporter subunit β (Reference Dawson and Karpen32). Increased ABCC2 gene expression may contribute to BA excretion to the lumen, which contributes to the suppression of enterohepatic 12αOH BA circulation. Such reduction of enterohepatic 12αOH BA circulation is not necessarily observed in other types of dietary oligosaccharides. Dietary difructose anhydride III does not lower the portal 12αOH BA concentration in CA-fed rats under similar experimental conditions(Reference Lee, Shimizu and Hagio10). In the literature, Chol-enriched diet triggers down-regulation of ASBT expression, a reduction in ileal BA uptake and an increase in the faecal excretion of BA in mice(Reference Thomas, Landrier and Gaillard33). Another study showed that a decrease in luminal BA induces up-regulation of ASBT expression and promotes uptake of ileal BA, which results in reduction of BA excretion in ampicillin-administered mice(Reference Miyata, Yamakawa and Hamatsu34). These studies suggest that ASBT expression depends on the luminal BA environment in order to maintain BA homoeostasis in the body. In this experiment, a significant interaction was observed between CA and Raf supplementation in case of Asbt expression. These observations suggest that suppression of enterohepatic 12αOH BA circulation plays a preventive role in the development of hepatic TAG accumulation.
Raf feeding increased the levels of acetic acid (Fig. 3) and lowered the caecal pH (Table 2). In addition, acetic acid and aortic 12αOH BA were predictors of liver TAG concentration (Fig. 3). Oral administration with acetate has been shown to improve hepatic steatosis in Otsuka Long-Evans Tokushima Fatty rats, accompanied by an increase in the AMP:ATP ratio, which contribute to the promotion of AMP-activated protein kinase and reduction in fatty acid synthesis(Reference Yamashita, Fujisawa and Ito35). Although such an effect was expected in the present study, a clear influence on gene expression of the hepatic de novo lipogenesis-related genes was not identified (Fig. 4).
Fig. 4. Ileal gene expression in rats fed control, cholic acid (CA), raffinose (Raf) or CA + Raf diet for 4 weeks. Ileal mRNA expression levels of apical Na-dependent bile acid transporter (Asbt), ATP-binding cassette subfamily C member 2 (Abcc2), organic solute transporter subunit α (Osta) and organic solute transporter subunit β (Ostb). The data were normalised with ribosomal protein lateral stalk subunit P0 (Rplp0) mRNA expression. Parameters of rats fed control and CA diets have been shown using blank and filled symbols, respectively. Parameters of rats fed diets with and without Raf supplementation have been shown in black and grey, respectively. The values have been expressed as mean values with their standard errors (n 5–6). P-values of the two-way ANOVA for CA and Raf are shown in the inset table.
Acidification of the intestine by ingestion of difructose anhydride III alters the gut microbiota and suppresses secondary BA formation(Reference Minamida, Kaneko and Ohashi36). On the other hand, DCA production was not inhibited by dietary Raf in the CA-fed rats in the present study regardless of low pH environment in the caecal contents (Fig. 2(e)). Clostridium prefers a high pH of 6·5–7·5 in a continuous culture study(Reference Edwards, Duerden and Read37), and 7α-dehydroxylase displays optimal activity at pH 7–8, while getting inhibited at a pH < 6·5 in vitro (Reference Macdonald, Singh and Mahony38). Raf lowered the caecal pH, but not to below pH 6·5. Hence, it is possible that upon Raf supplementation, clostridium could survive and the 7α-dehydroxylase activity was maintained, which enabled the conversion of primary BA into secondary BA. Stable DCA production upon CA supplementation may increase plasma transaminases, regardless of Raf supplementation in the present study. In our previous study(Reference Lee, Shimizu and Hagio10), CA diet increased blood transaminase activities, which positively correlated with the DCA concentrations in the caecal contents in a long-term experiment. A study aimed to compare BA hepatotoxicity among CA, DCA and LCA showed that DCA supplementation associated with oxidative stress is the most hepatotoxic of the three(Reference Delzenne, Calderon and Taper39). In the present study, we expected that Raf supplementation would lead to altered gut microbiota and reduced secondary BA production, including DCA. However, the faecal DCA and liver taurodeoxycholic acid concentrations remained high under the CA-supplemented condition, regardless of Raf ingestion. Although Raf reduced 12αOH BA concentrations in the liver, the effect was partial and taurodeoxycholic acid remained at a high level upon CA supplementation. These observations suggest that DCA produced in the large intestine can be absorbed, which may be responsible for the increase in transaminase activities. Since a synbiotic combination in the diet reduces DCA production(Reference Lee, Yoshitsugu and Kikuchi40), synbiotics with Raf and appropriate bacteria can suppress secondary BA production. In the multiple hit hypothesis in liver disease(Reference Buzzetti, Pinzani and Tsochatzis41), an additional second hit with oxidative stress or endoplasmic reticulum stress is necessary for disease progression. Inhibition of DCA production may reduce these risks, thus preventing further progression of fatty liver.
There were several limitations associated with the present study. First, we do not confirm that an increase in 12αOH in enterohepatic circulation directly induces hepatic lipid accumulation in humans. Second, the effect of dietary raffinose in humans remains obscure because intestinal bacteria that assimilate raffinose are not exactly the same as those in rats. Third, we have not clarified the mechanism by which 12αOH induces hepatic TAG accumulation in the CA-fed rat model. Further studies are necessary to address these issues. Fourth, there might be underpowered in some parameters even within the required range in the power analysis because the determination of the sample size was calculated for hepatic TAG concentration.
In conclusion, we found that normalisation of elevated 12αOH BA enterohepatic circulation by dietary Raf is associated with the amelioration of hepatic TAG accumulation induced upon CA supplementation. The present study also suggests faecal 12αOH BA as a marker for hepatic lipid accumulation.
Acknowledgements
This study was supported in part by the Japan Society for the Promotion of Science KAKENHI (S. I., grant numbers 19H2900 and 16K14917), the Japan Science and Technology Agency Centre of Innovation (S. I., grant number JPMJCE1301) and Nippon Beet Sugar Manufacturing Co. Ltd..
The authors’ contributions are as follows: S. I., S. F. and A. Y. contributed to the study design and interpretation of the findings; K. M., H. K., K. U. and S. I. carried out the experiments and collected the data; K. M. and S. I. wrote the manuscript. All the authors have read and approved the final version of the manuscript.
The authors declare the following financial interest/relationship which may be considered as potential competing interests: S. I. and A. Y. received grants and were provided raffinose from Nippon Beet Sugar Manufacturing Co. Ltd. The company was not involved in this study.
