Effects of Consuming Xylitol on Gut Microbiota and Lipid Metabolism in Mice

The sugar alcohol xylitol inhibits the growth of some bacterial species including Streptococcus mutans. It is used as a food additive to prevent caries. We previously showed that 1.5–4.0 g/kg body weight/day xylitol as part of a high-fat diet (HFD) improved lipid metabolism in rats. However, the effects of lower daily doses of dietary xylitol on gut microbiota and lipid metabolism are unclear. We examined the effect of 40 and 200 mg/kg body weight/day xylitol intake on gut microbiota and lipid metabolism in mice. Bacterial compositions were characterized by denaturing gradient gel electrophoresis and targeted real-time PCR. Luminal metabolites were determined by capillary electrophoresis electrospray ionization time-of-flight mass spectrometry. Plasma lipid parameters and glucose tolerance were examined. Dietary supplementation with low- or medium-dose xylitol (40 or 194 mg/kg body weight/day, respectively) significantly altered the fecal microbiota composition in mice. Relative to mice not fed xylitol, the addition of medium-dose xylitol to a regular and HFD in experimental mice reduced the abundance of fecal Bacteroidetes phylum and the genus Barnesiella, whereas the abundance of Firmicutes phylum and the genus Prevotella was increased in mice fed an HFD with medium-dose dietary xylitol. Body composition, hepatic and serum lipid parameters, oral glucose tolerance, and luminal metabolites were unaffected by xylitol consumption. In mice, 40 and 194 mg/kg body weight/day xylitol in the diet induced gradual changes in gut microbiota but not in lipid metabolism.


Introduction
Gut microbiota form many bioactive metabolites from dietary components which can regulate host metabolism [1][2][3][4][5]. For example, an improvement in glucose metabolism induced by dietary fiber is associated with the increased abundance of Prevotella [2]. Similarly, some food derivatives and food additives can affect host metabolism after interactions with gut microbiota [1,5].
Xylitol has a caries preventative effect via its capacity to inhibit the growth of Streptococcus mutans [6]. Dietary xylitol, metabolized into D-xylulose-5-phosphate, activates the carbohydrate response element binding protein (ChREBP) [7]. We previously reported that dietary xylitol combined with a high-fat diet (HFD) induced hepatic lipogenic gene expression via ChREBP mRNA expression [8]. In this report, we revealed that xylitol can improve HFD-induced hypertriglyceridemia and hypercholesterolemia with cecum enlargement in mice. In another report, the administration of a 2.5-10% xylitol solution reduced serum cholesterol and low density lipoprotein-cholesterol in diabetic mice [9]. Moreover, mice supplemented with 5% xylitol and 0.05%

PCR-DGGE Analysis
Denaturing gradient gel electrophoresis (DGGE) was performed as previously described [24] using the DCode TM Universal Mutation Detection System instrument and model 475 gradient former according to the manufacturer's instructions (Bio-Rad Labs, Hercules, CA, USA). The V2-V3 region of the 16S rRNA genes (positions 339-539 in the Escherichia coli gene) of bacteria in gut samples was amplified with the primers HDA1-GC and HDA2. PCR reaction mixtures and the amplification program were the same as described previously [24]. The denaturing gradient was formed with two Nutrients 2017, 9, 756 4 of 12 8% acrylamide (acrylamide-bis 37.5:1) with denaturing gradients ranging from 20-80% for analysis of the amplified 16S rRNA fragments. The 100% denaturant solution contained 40% (v/v) deionized formamide and 7 M urea. PCR product (40 µL) was mixed with 40 µL dye before loading. Gels were run in 0.5× Tris/Acetate/EDTA buffer at 60 • C for 5.2 h at 180 V, 210 mA, stained with Gel Star (Lonza Japan, Tokyo, Japan) for 30 min, and analyzed by ChemiDoc MP (Bio-Rad, Hercules, CA, USA). Image Lab software, version 5.0 (Bio-Rad) was used for the identification of bands and normalization of band patterns from DGGE gels.

Determination of Bacterial Strain by Sequence Analysis
Specific bands from DGGE gels were excised for DNA extraction, mashed, and incubated overnight in a diffusion buffer (0.5 M ammonium acetate, 1 mM EDTA, 0.1% SDS, 15 mM magnesium acetate). DNA was purified by the standard ethanol precipitation method. The V2-V3 region of the 16S rRNA genes were amplified by PCR, and purified DNA was used as a template. PCR products were cloned into the pCR2.1-TOPO vector (Invitrogen, Carlsbad, CA, USA), sequenced, and the bacterial genus was identified by BLAST.

Plasma and Hepatic Lipid Concentrations
Hepatic lipids were extracted and measured as previously described [25]. Plasma and liver triglycerides (TG) and total cholesterol concentration were measured by using Triglyceride-E and Cholesterol-E tests (Wako Pure Chemical Industries, Osaka, Japan), respectively.

RNA Preparation and Quantitative Reverse Transcriptase PCR
Extraction of total RNA, cDNA synthesis, and real-time PCR analysis were performed as described previously [25]. The relative abundance of each target transcript was calculated by normalization to the amount of amplified product from constitutively expressed β-actin mRNA (see primer sequences in Table 1).

Metabolome Analysis of Cecum Luminal Content by Capillary Electrophoresis Electrospray Ionization Time-of-Flight Mass Spectrometry
The cecum luminal content was immediately frozen in liquid nitrogen and stored at −80 • C until metabolite extraction. Sample tissues were weighed and completely homogenized in 0.5 mL ice-cold methanol containing 50 µM methionine sulfone and camphor-10-sulfonic acid as internal standards. The homogenates were mixed with 0.5 mL chloroform and 0.2 mL ice-cold Milli-Q water. After centrifugation at 2300× g for 5 min, the supernatant was centrifugally filtrated through 5-kDa cut-off filters (Millipore, Bedford, MA, USA) at 9100× g for 4-5 h to remove proteins. The filtrate was centrifugally concentrated in a vacuum evaporator, dissolved with Milli-Q water, and analyzed by capillary electrophoresis electrospray ionization time-of-flight mass spectrometry (CE-TOFMS).
CE-TOFMS analysis was performed using an Agilent CE system combined with a TOFMS (Agilent Technologies, Palo Alto, CA, USA) as reported by previously [24,26,27]. Each metabolite was identified based on a reference which containing internal standards including 110 metabolites (H3304-1002, Human Metabolome Technology (HMT), Inc., Tsuruoka, Japan) to m/z and migration time, and quantified by peak area.

Statistical Analyses
Data are expressed as means ± standard errors of the mean (SEM). A significant difference between groups was assessed via an unpaired two-tailed t-test in experiment 2 and FMT experiment. For comparisons among more than three groups, we employed analysis of variance (ANOVA) or the Kruskal-Wallis test in experiment 1. When a significant difference was found by ANOVA or the Kruskal-Wallis test, post hoc analyses were performed using the Tukey-Kramer protected least significant difference test. Concentration-dependent effects were identified via linear regression analysis. Spearman's rank correlation coefficient was used to calculate correlation coefficients between selected variables. Differences were considered significant at p < 0.05. Statistical analyses were performed using Mass Profiler Professional and Excel-Toukei 2006 (SSRI, Tokyo, Japan).

Results
To elucidate the effect of consuming low-dose xylitol on gut microbiota and lipid metabolism, the mean xylitol dosage administered to mice after weaning was 40 ± 5 mg/kg body weight/day (CD-LX), 194 ± 24 mg/kg body weight/day (CD-MX), and 194 ± 25 mg/kg body weight/day (HFD-MX) ( Figure 1A,B). During the treatment periods, body weight, relative epididymal fat weight per body weight, relative liver weight per body weight, and relative cecum weight per body weight were not different between the xylitol-fed groups and the control group of mice in experiment 1 and 2 ( Figure 1C,D, and Table 2). The relative amount of total fecal bacteria to fecal DNA displayed a trend towards an increase in the feces of CD-MX mice and was significantly increased in the feces of HFD-MX mice when compared with control mice (Figure 2A,B). In contrast, Bacteroides, a phylum of bacteria, was reduced in both MX mice fed a CD or HFD (Figure 2A,B). In addition, the combination of an HFD and ingestion of a medium-dose xylitol solution showed that an increased amount of Firmicutes phylum, the Prevotella genus, and the relative ratio of Firmicutes/Bacteroides and Prevotella/Bacteroides than those of HFD fed control mice ( Figure 2B,C). To explore in detail the microbiome bacterial composition, we carried out DGGE analysis. We identified five genera, which included two species of Clostridium and a Faecalibaculum genus which were increased in the MX mice and one from both the Clostridium and Barnesiella genera which were reduced in the MX mice; different analysis bands were significantly different ( Figure 2D

Results
To elucidate the effect of consuming low-dose xylitol on gut microbiota and lipid metabolism, the mean xylitol dosage administered to mice after weaning was 40 ± 5 mg/kg body weight/day (CD-LX), 194 ± 24 mg/kg body weight/day (CD-MX), and 194 ± 25 mg/kg body weight/day (HFD-MX) ( Figure 1A,B). During the treatment periods, body weight, relative epididymal fat weight per body weight, relative liver weight per body weight, and relative cecum weight per body weight were not different between the xylitol-fed groups and the control group of mice in experiment 1 and 2 ( Figure 1C,D, and Table 2). The relative amount of total fecal bacteria to fecal DNA displayed a trend towards an increase in the feces of CD-MX mice and was significantly increased in the feces of HFD-MX mice when compared with control mice (Figure 2A,B). In contrast, Bacteroides, a phylum of bacteria, was reduced in both MX mice fed a CD or HFD (Figure 2A,B). In addition, the combination of an HFD and ingestion of a medium-dose xylitol solution showed that an increased amount of Firmicutes phylum, the Prevotella genus, and the relative ratio of Firmicutes/Bacteroides and Prevotella/Bacteroides than those of HFD fed control mice ( Figure 2B,C). To explore in detail the microbiome bacterial composition, we carried out DGGE analysis. We identified five genera, which included two species of Clostridium and a Faecalibaculum genus which were increased in the MX mice and one from both the Clostridium and Barnesiella genera which were reduced in the MX mice; different analysis bands were significantly different ( Figure 2D       Our study and others report that a high dose of xylitol improved hyperlipidemia in mice fed an HFD and in diabetic mice [8][9][10]. To reveal the effect of a low dose of xylitol on lipid metabolism, we investigated cholesterol and triglyceride concentrations in the liver and serum, parameters which were not different among the three groups of mice maintained on the control diet ( Table 2). In contrast, an HFD induced hypertriglyceridemia and hypercholesterolemia in the liver, but xylitol supplementation did not ameliorate dyslipidemia ( Table 2). We also found that hepatic ChREBP and the expression of its target genes were increased in HFD-MX mice compared with control mice ( Figure 3A) as was reported in a previous study [8]. In addition, we investigated glucose tolerance in mice fed an HFD because two reports have shown an abundance of several different species of the genus Prevotera that are linked with glucose intolerance or insulin resistance in humans and mice [2,28]. We could not detect any changes in glucose tolerance, as well as the expression of inflammation-related genes, in mice fed the HFD with or without dietary xylitol supplementation ( Figure 3A,B).
Nutrients 2017, 9, 756 7 of 12 < 0.05 between CD and CD-MX. b: p < 0.05 between CD-LX and CD-MX. c: p < 0.05 between HFD and HFD-MX. Table 2. Body weight, organ weight, and plasma parameters of mice fed the control diet or the high-fat diet with or without xylitol. Data represent the mean ± SEM (n = 5-6).

CD (n = 5) CD-LX (n = 5) CD-MX (n = 5) HFD (n = 5) HFD-MX (n = 6)
Our study and others report that a high dose of xylitol improved hyperlipidemia in mice fed an HFD and in diabetic mice [8][9][10]. To reveal the effect of a low dose of xylitol on lipid metabolism, we investigated cholesterol and triglyceride concentrations in the liver and serum, parameters which were not different among the three groups of mice maintained on the control diet ( Table 2). In contrast, an HFD induced hypertriglyceridemia and hypercholesterolemia in the liver, but xylitol supplementation did not ameliorate dyslipidemia ( Table 2). We also found that hepatic ChREBP and the expression of its target genes were increased in HFD-MX mice compared with control mice ( Figure 3A) as was reported in a previous study [8]. In addition, we investigated glucose tolerance in mice fed an HFD because two reports have shown an abundance of several different species of the genus Prevotera that are linked with glucose intolerance or insulin resistance in humans and mice [2,28]. We could not detect any changes in glucose tolerance, as well as the expression of inflammation-related genes, in mice fed the HFD with or without dietary xylitol supplementation ( Figure 3A,B). To further investigate the effects of xylitol intake on luminal metabolites, we conducted a CE-MS analysis. We identified 94 metabolites from a metabolite list provided by HMT. From the PCA plot, we were unable to distinguish any metabolite patterns among the groups of mice in experiment 1 fed the AIN93G diet with or without supplemental dietary xylitol in their drinking water ( Figure 3C,D). Only dihydroxyacetone phosphate concentration was different between CD and CD-MX groups. These results suggest that the changes in luminal content microbiota in xylitol supplemented groups had little, if any, effect on overall metabolism. water ( Figure 3C,D). Only dihydroxyacetone phosphate concentration was different between CD and CD-MX groups. These results suggest that the changes in luminal content microbiota in xylitol supplemented groups had little, if any, effect on overall metabolism.
Finally, we attempted to detect microbiota-dependent effects of xylitol feeding in mice fed an HFD via FMT. One day after the final transplantation, the microbiota was clearly different between the mice that were recipients of feces transplanted from mice fed an HFD (HFD-FMT) and fed an HFD with medium-dose xylitol (HFD-MX-FMT) ( Figure 4A,B). These perceptible differences between the two groups disappeared 18 day after the transplantation ( Figure 4A,B). No changes in luminal metabolites, body weight, and relative tissue weight between HFD-FMT and HFD-MX-FMT mice were detected ( Figure 4C, Table 3). These results indicate that changes in the fecal microbiota of mice fed xylitol are transient and likely continuous xylitol supplementation is necessary to sustain the changes observed. Interestingly, serum cholesterol in the HFD-MX-FMT mice was slightly, but significantly, higher than that of the HFD-FMT mice (Table 3). This suggests that changes in the composition of microbiota induced by dietary xylitol increase serum cholesterol. Finally, we attempted to detect microbiota-dependent effects of xylitol feeding in mice fed an HFD via FMT. One day after the final transplantation, the microbiota was clearly different between the mice that were recipients of feces transplanted from mice fed an HFD (HFD-FMT) and fed an HFD with medium-dose xylitol (HFD-MX-FMT) ( Figure 4A,B). These perceptible differences between the two groups disappeared 18 day after the transplantation ( Figure 4A,B). No changes in luminal metabolites, body weight, and relative tissue weight between HFD-FMT and HFD-MX-FMT mice were detected ( Figure 4C, Table 3). These results indicate that changes in the fecal microbiota of mice fed xylitol are transient and likely continuous xylitol supplementation is necessary to sustain the changes observed. Interestingly, serum cholesterol in the HFD-MX-FMT mice was slightly, but significantly, higher than that of the HFD-FMT mice (Table 3). This suggests that changes in the composition of microbiota induced by dietary xylitol increase serum cholesterol. Table 3. Body weight, organ weight, and plasma parameters of mice fed the high-fat diet following fecal transplantation from mice fed a high-fat diet with or without xylitol.

Discussion and Conclusions
In this study, we showed that the administration of xylitol at 40 and 194 mg/kg body weight/day significantly altered gut microbiota in mice. In particular, we noted the relative abundance of the Bacteroidetes phylum was reduced in mice in the CD-MX and HFD-MX groups, indicating that xylitol suppressed the growth of some bacterium, including the genus Barnesiella in mice fed either CD or HFD. In contrast, the relative abundance of Firmicutes phylum and the genus Prevotella were increased in the HFD-MX group. Contrary to the significant alteration of microbiota, body composition, lipid parameters, and luminal metabolites were not different between groups, regardless of xylitol consumption.
The improvement of glucose tolerance observed with increased dietary fiber intake is linked with a higher abundance of the genus Prevotella [2]. In contrast, the abundance of Prevotella copri was positively associated with microbial branched-chain amino acid (BCAA) biosynthesis in the gut and insulin resistance with a soy protein diet which contained a low level of BCAAs [28]. Our present study showed an increased abundance of Prevotella and an increase in the Prevotella/Bacteroidetes ratio, but no differences were observed in glucose tolerance or luminal BCAA concentrations between the HFD and HFD-MX groups. Because the mice were fed a diet containing casein, which has as the protein source a high BCAA content, we were unable to detect any changes in the luminal BCAA concentrations. These results suggest that changes in the bacterial composition and the supply of dietary components modulates host metabolism in a coordinated manner.
An amount of dietary indigestible fiber and gut microbiota which digest fiber regulates cecum weight [29,30]. In the experiment 1, we used AIN93G as a control diet which contains more fiber (5%) than the HFD which used in experiment 2 (2.9%). In the present study, xylitol feeding did not affect cecum weight, therefore a difference in the amount of dietary fiber might affect cecum weight.
In our study, daily supplemental dietary xylitol of 194 but not 40 mg/kg body weight induced significant changes in microbiota for the genera Barnesiella, which was reduced, and Feacalibaculum, which was increased. Barnesiella and Feacalibaculum have been detected in human or mice microbiota [31,32]. Barnesiella species have been negatively correlated with the colonization of vancomycin-resistant Enterococcus faecium in mice intestines [33] and the relative abundance of the bacterial genera Faecalibacterium was significantly decreased in children at risk of asthma [34]. In contrast, dietary xylitol suppressed lipopolysaccharide-induced inflammatory responses in male broiler chickens [35] and has been shown to ameliorate human respiratory syncytial virus infections in mice [36]. Collectively, changes in the fecal microbiota of animals fed xylitol might affect immune responses or colonization of some bacterial species.
Recently, Geidenstam et al. reported that baseline levels of serum xylitol showed an inverse association with a ≥10% weight loss in obese subjects fed low-calorie diet [37]. Firmicutes phylum accelerates degradation of food component to supply energy to host, it is, therefore, known as obesity-related bacterial phylum [38]. In our study, the total bacteria/DNA and the relative abundance of Firmicutes phylum were increased in the HFD-MX group. Geidenstam and colleague did not examine gut microbiota in their study, human metabolism of xylitol and potential involvement of the gut microbiota could help us to understand the effect of xylitol feeding on human lipid metabolism.
Xylitol metabolized into xylulose-5-phosphate (X-5-P) is synthesized via the pentose phosphate pathway [39] and activates ChREBP through protein phosphatase 2A [40]; this results in its binding to a specific DNA sequence which induces lipogenesis-related genes which increase lipogenesis from carbohydrates [41]. Daily dietary xylitol at exposure levels ranging between 1.5-4.0 g/kg body weight in combination with a HFD showed a trend towards increased expression of hepatic ChREBP mRNA and a reduction in hepatic triglycerides and total cholesterol as reported in a previous study [8]. These findings suggest xylitol has other functions unrelated to the ChREBP pathway. In the present study, we found that an HFD supplemented with 0.2 g/kg body weight/d of dietary xylitol also induced the increased expression of hepatic ChREBP mRNA and but had a tendency to increase hepatic triglycerides and total cholesterol. The differences between the studies may arise from differences in model species, xylitol dose, and diet which used to characterize the effect of xylitol on lipid metabolism. Because plasma triglyceride level was not increased by HFD feeding in this study, another study that uses another diet (e.g., high-fat high-sucrose diet) which strongly induces hypertriglyceridemia will help to understand the effect of xylitol to alter plasma TG levels. Hepatic total cholesterol in HFD-MX-FMT mice was slightly but significantly higher than that of HFD-FMT mice. Taken together, these changes in the gut microbiota induced by dietary xylitol may potentiate the accumulation of cholesterol and upregulation of hepatic ChREBP.
In conclusion, we found that 40 and 194 mg/kg body weight/day of dietary xylitol in mice induced gradual changes in gut microbiota, but did not ameliorate HFD-induced dyslipidemia.