Bile Acid Is a Host Factor That Regulates the Composition of the Cecal Microbiota in Rats

Transcription

1 GASTROENTEROLOGY 2011;141: Bile Acid Is a Host Factor That Regulates the Composition of the Cecal Microbiota in Rats K. B. M. SAIFUL ISLAM,* SATORU FUKIYA,* MASAHITO HAGIO, NOBUYUKI FUJII, SATOSHI ISHIZUKA, TADASUKE OOKA, YOSHITOSHI OGURA,,, TETSUYA HAYASHI,,, and ATSUSHI YOKOTA* *Laboratory of Microbial Physiology and Laboratory of Nutritional Biochemistry, Research Faculty of Agriculture, Hokkaido University, Sapporo, Japan; Division of Microbiology, Department of Infectious Diseases, Faculty of Medicine, and Division of Bioenvironmental Science, Frontier Science Research Centre, University of Miyazaki, Miyazaki, Japan BACKGROUND & AIMS: Alterations in the gastrointestinal microbiota have been associated with metabolic diseases. However, little is known about host factors that induce changes in gastrointestinal bacterial populations. We investigated the role of bile acids in this process because of their strong antimicrobial activities, specifically the effects of cholic acid administration on the composition of the gut microbiota in a rat model. METHODS: Rats were fed diets supplemented with different concentrations of cholic acid for 10 days. We used 16S ribosomal RNA gene clone library sequencing and fluorescence in situ hybridization to characterize the composition of the cecal microbiota of the different diet groups. Bile acids in feces, organic acids in cecal contents, and some blood parameters were also analyzed. RESULTS: Administration of cholic acid induced phylum-level alterations in the composition of the gut microbiota; Firmicutes predominated at the expense of Bacteroidetes. Cholic acid feeding simplified the composition of the microbiota, with outgrowth of several bacteria in the classes Clostridia and Erysipelotrichi. Externally administered cholic acid was efficiently transformed into deoxycholic acid by a bacterial 7 -dehydroxylation reaction. Serum levels of adiponectin decreased significantly in rats given the cholic acid diet. CONCLUSIONS: Cholic acid regulates the composition of gut microbiota in rats, inducing similar changes to those induced by high-fat diets. These findings improve our understanding of the relationship between metabolic diseases and the composition of the gastrointestinal microbiota. Keywords: Large Intestine; Commensal Bacteria; Microbiota Analysis; Host-Microbe Interaction. Recently, the idea that gut microbiota influence host health has become popular, and it has been argued that an imbalanced bacterial population associated with a high-fat diet is a trigger for the development of metabolic diseases such as obesity, 1 5 diabetes, 6,7 and hypercholesterolemia 8 in animal models. In human studies, altered gut microbiota has been reported in subjects diagnosed with type 2 diabetes mellitus 9 and in patients with inflammatory bowel disease. 10,11 Currently, much emphasis is put on assessing changes in the gut microbiota associated with high-fat diets and/or diseases and on clarifying how such reshaped gut microbiota mediate the development of metabolic diseases. However, important questions about why and how high-fat diets and/or diseases induce changes in the bacterial population remain to be elucidated. 12 This seems to be due to a relative lack of knowledge about the effect of host factors on gut microbiota formation in vivo. In this context, we hypothesized that bile acids might be good candidates that have not yet been well characterized. Bile acids are sterol compounds that are synthesized from cholesterol in the liver and secreted into the duodenum as the main component of bile. Because of their amphipathic properties, bile acids emulsify lipids, thereby aiding the absorption of liposoluble dietary nutrients. A high-fat diet enhances bile secretion to facilitate lipid digestion. 13 During transit to the large intestine, bile acids undergo modifications to the steroid nucleus by some members of the gut microbiota, yielding secondary bile acids. 14 The most typical secondary bile acid is deoxycholic acid (DCA; Supplementary Table 1), which arises from cholic acid (CA; Supplementary Table 1), the most abundant bile acid in biliary bile in humans. The conversion to secondary bile acids is mediated by a 7 -dehydroxylation reaction catalyzed by some species of Clostridium; nearly 100% of CA is converted into DCA in the large intestine. Another important feature of bile acids in this study is their strong antimicrobial activity. 15 Previously we have shown, with in vitro experiments using lactic acid bacteria and bifidobacteria, that the primary mechanism underlying the antimicrobial action of bile acids is membrane damage. 15 Physiologic concentrations of CA and DCA in the intestine disturb the integrity of membranes as a result of their detergent effect, leading to the leakage of ions and cellular components and eventually to cell death. DCA is one of the most potent antimicrobial bile acids, with 10 times the bactericidal activity of CA. 15 Thus, bile acids seem to exert strong selective pressure on the Abbreviations used in this paper: CA, cholic acid; DAPI, 4=,6-diamidino-2-phenylindole; DCA, deoxycholic acid; FISH, fluorescence in situ hybridization; H-CA, diet supplemented with 5 mmol/kg cholic acid; M-CA, diet supplemented with 1.25 mmol/kg cholic acid; OTU, operational taxonomic unit; RDP, Ribosomal Database Project; rrna, ribosomal RNA; SCFA, short-chain fatty acid; TCA, taurocholic acid by the AGA Institute /$36.00 doi: /j.gastro

2 1774 ISLAM ET AL GASTROENTEROLOGY Vol. 141, No. 5 gut microbiota. Accordingly, only microbial populations able to tolerate physiologic concentrations of bile acids can survive in the gut. This seems especially evident when administering a high-fat diet causing increased intestinal bile acid flow. Accordingly, it has recently been shown in a mouse model that a high-fat diet, or the so-called Western diet, shifts the balance of 2 major phyla of gut microbiota, in many cases increasing Firmicutes and decreasing Bacteroidetes. 1 4 Therefore, to examine the possibility that the gut microbiota population is controlled by bile acids in vivo, rats were fed a basal diet or a CA-supplemented diet and their cecal microbiota were analyzed by 16S ribosomal RNA (rrna) gene clone library sequencing. The results were further confirmed by fluorescence in situ hybridization (FISH). The composition of fecal bile acids and cecal organic acids was analyzed, as were some host responses. The overall results reveal remarkable effects of CA on both the gut microbiota population and their metabolism. In particular, the changes in the gut microbiota were similar to those observed in mouse models fed high-fat diets. 1 4 These findings highlight a hitherto unknown role of bile acids in the control of the gut microbiota population, thereby providing new insight into the understanding of the metabolic diseases and their pathophysiology in relation to gut microbiota composition. Materials and Methods Design of Animal Experiments Male WKAH/HkmSlc rats (3 weeks old; Japan SLC Inc, Hamamatsu, Japan) were housed individually in a controlled environment 16 and had free access to food and water for the entire study period. The rats were acclimated on a control diet for 4 days and then divided into 3 groups. Each group was fed one of 3 diets for 10 days: a control diet (control group), a diet supplemented with 1.25 mmol/kg CA (M-CA group), or a diet supplemented with 5 mmol/kg CA (H-CA group) (see Supplementary Table 2 for each diet composition). Body weight gain and food intake were measured daily. Fecal samples were collected for 1 day at the end of the experimental period and used for bile acid analysis. The rats were killed by exsanguination under sodium pentobarbital anesthesia (50 mg/kg body wt), and the cecal contents were immediately collected for analyses of ph, organic acids, and microbiota population. The cecal bacterial population was analyzed by sequencing 16S rrna gene clone libraries, FISH, and 4=,6-diamidino-2-phenylindole (DAPI) staining. Cecum and colon tissues were collected for preparation of histologic specimens. Adipose tissues for weight measurement and blood samples for the measurements of nonesterified fatty acids and adiponectin were also collected (see Supplementary Materials and Methods for detailed sample preparations for each analysis and the methods for analyzing bile acids, organic acids, ph, nonesterified fatty acids, and adiponectin). Bacterial Population Analysis by Sequencing 16S rrna Gene Clone Libraries Total DNA was purified from the cecal contents using mechanical cell disruption with glass beads and phenol-chloroform extraction. Bacterial 16S rrna gene fragments were amplified by polymerase chain reaction with the primers B34F 17 and 1391R, 18 ligated into the pgem-t Easy Vector System I (Promega Corp, Madison, WI), and cloned into Escherichia coli DH5. Clone libraries were generated for individual rats (6 rats per diet group, 48 clones per rat), and the nucleotide sequences were determined by Takara Bio Inc (Otsu, Shiga, Japan); analysis and sequence classification were conducted using the myrdp pipeline tool from the Ribosomal Database Project (RDP). 19,20 Sequences ( 97% identity) were assembled into operational taxonomic units (OTUs). BLASTN program for similarity analysis and MEGA4 program 21 for multiple sequence alignment and phylogenetic analyses were used. Shannon index of diversity was also calculated for each diet group (see Supplementary Materials and Methods for additional details). Enumeration of Microbiota in Cecal Contents by FISH and DAPI The sample fixation procedures for FISH and DAPI staining were conducted as described previously. 22 For FISH analysis, fixed cells were hybridized with Cy3-labeled oligonucleotide probes Eub (5=-GCTGCCTCCCGTAGGAGT-3=), Erec (5=-GCTTCTTAGTCARGTACCG-3=), and Bac (5=-[A/G]GCTGCCTTCGCAATCGG-3=) as previously described, 25 with the modification that a hybridization solution of ph 8.0 and 2 washes after hybridization were used. Details of imaging and counting are described in Supplementary Materials and Methods. Statistical Analysis and Ethical Considerations The feeding trial was conducted in 5 different experiments at 5 different times to verify the reproducibility of the results. The number of rats per experiment and the analytical parameters corresponding to each experiment are listed in Supplementary Table 3. Results are expressed as the mean SEM. Statistical analyses were performed among different diet groups using a Tukey Kramer test. This study was approved by the Animal Use Committee at Hokkaido University. The animals were maintained according to the Guidelines for the Care and Use of Laboratory Animals, Hokkaido University. Results Metabolic Parameters in Rats Fed CA During the study period, all the animals were healthy and there was no significant difference in food intake, final body weight, or body weight gain among the groups (Supplementary Table 4). Table 1 shows the tissue and serum parameters at the end of the experimental period. Significantly reduced epididymal adipose tissue weight was measured in the H-CA group. Lower serum adiponectin level was observed in the CA-fed groups. Reproducibility of these results was verified in another experimental cohort (Supplementary Table 5). Furthermore, histologic observations indicated that there was no apparent inflammation in the cecal and colonic mucosa (Supplementary Figure 1). Profiles of Bile Acids in Feces The bile acid profile in the intestine is determined by a combination of bile acid biosynthesis/modification

3 November 2011 BILE ACID REGULATES CECAL MICROBIOTA 1775 Table 1. Effect of CA Supplementation to the Diet on Tissue Weights and Tissue Parameters Tissue weights and serum parameters at day 10 Control group M-CA group H-CA group Tissue weights (g/100g body weight) Mesenteric adipose tissue Epididymal adipose tissue a a,b b Retroperitoneal adipose tissue Liver Cecum b a a Cecal contents Serum parameters Nonesterified fatty acids (mmol/l) Adiponectin ( g/ml) a b b NOTE. Data from the second and the third cohorts are used for tissue weights, and data from the fourth cohort are used for serum parameters. Values are expressed as mean SEM. Values not sharing the same lowercase letter in a row are significantly different among the diet groups (n 6, P.05; Tukey Kramer test). in the liver and bioconversion by intestinal bacteria. The bile acid profiles in the large intestine of rodents are dominated by muricholic acids, which have one hydroxyl group each at carbons 3, 6, and 7 of the steroid nucleus. 26 The remaining components are generally similar to those found in humans. As shown in Figure 1A, increased intestinal bile acid flow as a result of the M-CA and H-CA diets dramatically increased the fecal concentrations of both the highly bactericidal DCA and the administered CA, whereas the biosynthesis and metabolism of muricholic acids were apparently unaffected (see Supplementary Tables 1 and 6 for the chemical structures and concentrations of all the detected bile acids, respectively). This clearly indicates that CA in the diet is extensively transformed into DCA by bacterial 7 -dehydroxylation in the rat cecum 14 ; a similar effect occurs in the human colon. 14 In addition, 7-oxo-deoxycholic acid, another metabolite formed from CA by the 7 -hydroxysteroid dehydrogenases of some intestinal bacteria, 27 was significantly increased in the H-CA group. However, the antimicrobial activity of 7-oxo-deoxycholic acid is far lower than that of DCA (unpublished data, September 2011). The total bile acid concentration for the M-CA and H-CA groups increased 6-fold and 20-fold, respectively, compared with the control group. The cecal bile acid concentration can be roughly estimated from a dry fecal sample, assuming that the water content of the biological sample is 75%. Thus, the average cecal DCA concentrations in the control and M-CA groups were calculated to be 0.07 mmol/l and 0.98 mmol/l, respectively; these values are close to those reported for the DCA content of fecal water from humans on a normal ( mmol/l) or high-fat diet ( mmol/l), respectively. 28 By contrast, the DCA concentration in the H-CA group was 2.55 mmol/l, which is far higher than the physiologic range in the human colon. Also, it is important to note that Figure 1. Effect of CA feeding on bile acid and organic acid compositions in rats of different diet groups. (A) Major bile acids found in feces. For the control and H-CA groups, n 12; for the M-CA group, n 9. (B) Major organic acids found in the cecal contents. For the control and H-CA groups, n 6; for the M-CA group, n 9. Total SCFAs are the sum of acetate, propionate, and n-butyrate. Total organic acids are the sum of all the organic acids detected. The mean SEM was plotted. Different lowercase letters indicate significant differences in each acid (P.05; Tukey Kramer test).

4 1776 ISLAM ET AL GASTROENTEROLOGY Vol. 141, No. 5 the estimated cecal DCA concentrations in CA-fed groups are high enough to severely inhibit the in vitro growth of typical intestinal bacteria ( 1 mmol/l DCA), 15,29 suggesting that the DCA levels in the M-CA and H-CA groups exert great environmental stress on the cecal microbiota. Cecal Organic Acid Profiles Bile acid stress was expected to affect the fermentative metabolism of the gut microbiota. As shown in Figure 1B (see Supplementary Table 7 for the concentrations of all detected organic acids), the concentrations of short-chain fatty acids (SCFAs; ie, acetate, propionate, and butyrate) decreased as the CA concentration in the diet increased. The decreases became significant in the H-CA group, in which the total SCFA concentration and the total organic acid concentration both decreased to approximately 57% of the control group concentrations. Among the SCFAs, acetate and butyrate showed marked and significant reductions. In response to these changes, the ph of the cecal contents became significantly higher in the H-CA group (Supplementary Table 4). CA Feeding Alters the Gut Microbiota Population Microbial communities of the cecal contents were next investigated by DAPI staining, sequencing of 16S rrna gene clone libraries, and FISH analysis. As observed under an epifluorescence microscope after DAPI staining, the control group showed a variety of morphologies, from cocci to long rods (Figure 2A), whereas the CA-fed groups showed much less variety and primarily displayed cocci and short rods (Figure 2B and C). Total bacterial counts decreased with increasing dietary CA concentration (Figure 2), probably due to the antimicrobial activity of DCA. A significant reduction in total bacterial counts, to 51% that of the control group, was observed in the H-CA group. These results suggest that the bacterial populations in both the M-CA and H-CA groups are altered. Indeed, sequencing 16S rrna gene clone libraries revealed striking alterations in the bacterial populations of both the M-CA and H-CA groups (of a total 854 clones sequenced, 825 sequences could be aligned successfully on the RDP database: control group, n 266; M-CA group, n 274; H-CA group, n 285) (Figure 3; see Supplementary Tables 8 and 9 for further information). Our results show that the bacterial population in control rat ceca was dominated by Firmicutes (54.1%) and Bacteroidetes (30.7%), together with minority populations such as Proteobacteria (6.1%) and Actinobacteria (2.6%) (Figure 3A). This distribution is similar to those previously found in humans 30 and mice. 18 In contrast, in the CA-fed groups, Firmicutes expanded significantly, to approximately 95% of the total clones, at the expense of Bacteroidetes and Actinobacteria (Figure 3A). The increase in Firmicutes resulted from expansion of the Clostridia class in particular, as well as the Erysipelotrichi class (Figure 3B). Within these 2 classes, the genus Blautia and the genus Allobaculum dominated, respectively. Clones from Blautia and Allobaculum increased to account for approximately 60% and 15% of the total clones, respectively, in the M-CA and H-CA groups (Figure 3C). In Proteobacteria, the class Gammaproteobacteria expanded with increasing CA concentration, and the remaining classes nearly disappeared in the M-CA and H-CA groups (Figure 3B). Because biases are inherent in all molecular methods used for microbial ecology studies, we also applied FISH analysis, which is based on molecular principles different from 16S rrna gene clone library sequencing, to confirm the observed alterations in the bacterial populations. For this purpose, 2 group-specific probes were used to detect phylum-level alterations together with an Eub probe for the domain bacteria. To detect Firmicutes, an Erec probe targeting the Ruminococcus-Eubacterium- Clostridium cluster was used. The Erec482 probe can interact with 88% of the species in the genus Blautia, a major population in the M-CA and H-CA groups (Figure 3C), although the coverage of this probe for Firmicutes is only 13%. To detect Bacteroidetes, a Bac probe was used; it can interact with 52% of the class Bacteroidia, to which Figure 2. Effect of CA on bacterial morphology in rat cecal contents and total cell count as revealed by DAPI staining. (A) Control group, (B) M-CA group, and (C) H-CA group. Values for the total cell count per gram wet cecal content are given as the mean SEM. The values labeled with different lowercase letters are significantly different (n 6 for the control and H-CA groups, and n 9 for the M-CA group, P.05; Tukey Kramer test).

5 November 2011 BILE ACID REGULATES CECAL MICROBIOTA 1777 Figure 3. Composition of cecal microbiota of rats from different diet groups as revealed by sequencing of 16S rrna gene clone libraries. Population analyses for each diet group included (A) phyla, (B) classes, and (C) major genera found. Mean percentage (%) of the total population is shown (n 6 for each diet group). Classification is based on a 95% confidence threshold using a total of 825 sequence reads (control group 266 reads, M-CA group 274 reads, and H-CA group 285 reads). Unclassified means bacteria not able to be classified by the RDP pipeline. See Supplementary Tables 8 and 9 for detailed distribution of bacteria by different diet groups. all the detected Bacteroidetes clones belonged (Figure 3B). The coverage of this probe for the phylum Bacteroidetes is 32%, which is the highest possible value at present for a single probe. The results shown in Figure 4 indicate that although Bac719-positive cells remained detectable in the M-CA and H-CA groups, these populations were significantly smaller than those in the control group and decreased significantly as the dietary CA concentration increased. The population of Bac719-positive cells decreased from 31.1% of the total population in the control group to 5.6% in the H-CA group. We consider the discrepancies in our data to derive from the biases intrinsic to the different molecular methods for analyzing microbial ecology. In sharp contrast, the Erec482-positive population was significantly increased in the CA-fed groups as compared with the control group. Overall, the results obtained by FISH agree well with those obtained using the clone library method, and thus we conclude that CA in the diet increases the relative abundance of Firmicutes and decreases that of Bacteroidetes in the rat cecal microbiota. Additionally, a decreasing trend was observed in the Eub338-positive fraction of DAPI-positive cells in the CAfed groups, suggesting that these populations lose viability due to the antimicrobial activity of DCA. Diversity Analysis of the Cecal Microbiota Analysis of OTUs (using 97% nucleotide identity as a cutoff for OTUs) identified 117 OTUs and revealed that specific bacteria at the genus and species levels were highly concentrated in the M-CA and H-CA groups (Figure 5) (see Supplementary Table 10 for OTU information and Supplementary Figure 2 for the phylogenetic tree). Thus, the diversity of the OTUs in the M-CA and H-CA groups, expressed as Shannon indices (2.936 and 3.487, respectively), was lower than that of the control group (5.839). OTU1 (Blautia species), which is characterized as a close relative of Blautia schinkii (96% identity), increased dramatically from 2% of the total clones in the control group to 44% and 26% in the M-CA and H-CA groups, respectively. OTU3 (Blautia glucerasea) and OTU4 (Blautia producta) showed similar alterations. The latter appeared to be especially concentrated in the H-CA group, representing 20% of the total clones. Phylogenetically, OTU1 and OTU3 are closely related. The genus Blautia is a recently proposed phylogenetic group in the Clostridia class that incorporates several species of Clostridium and Ruminococcus within the clostridial rrna cluster XIVa. 31 Their cell morphologies are coccoid or ovoid, in agreement with our

6 1778 ISLAM ET AL GASTROENTEROLOGY Vol. 141, No. 5 (Clostridium bartlettii) were also detected only in the CA-fed groups, and the latter was detected only in the H-CA group, representing 8% of the total clones. OTU9 (Eubacterium cylindroides), closely related to OTU6, was also detected more frequently in the CA-fed groups than in the control group. In contrast, OTUs 12, 18, and 20 in the class Clostridia were detected only in the control group, demonstrating the opposite response to the introduction of CA. Among the Proteobacteria, the Gammaproteobacteria that bloomed in the H-CA group consisted solely of OTU11 (E coli), amounting to 5% of the total clones in the H-CA group. Figure 4. Analysis of microbiota in rat cecal contents by FISH using the Eub338, Erec482, and Bac719 probes. The mean SEM was plotted for each group. Different lowercase letters represent significant differences (n 6 for each diet group, P.05; Tukey Kramer test). microscopic observations (Figure 2B and C). The blooming of these OTUs accounted for the significant increase of the Clostridia class in the CA-fed groups. Another notable OTU was OTU2 (Allobaculum species), in the class Erysipelotrichi, because it was detected only in the CA-fed groups and expanded dramatically up to 15% and 11% of the total clones in the M-CA and H-CA groups, respectively. However, the phylogenetically closely related OTU5 was detected less frequently with increasing CA concentration, suggesting a difference in bile acid resistance between the 2 OTUs. The genus Allobaculum has been recently proposed as a member of the clostridial rrna cluster XVI. 32 OTU6 (Clostridium innocuum) and OTU7 Discussion Using simple experiments in a rat model, this study clearly shows the hitherto unexplored role of bile acid as a host factor that controls the gut microbiota population in vivo. This has been an unanswered question since early in the last century when the antimicrobial activity of bile acid was first described, 33 although the bactericidal action of bile acid on typical intestinal bacteria has often been examined in vitro. 15,29,34 Our previous work has shown the primary mechanism underlying the bactericidal action of bile acid to be membrane damage. 15 The bactericidal activity of a bile acid molecule corresponds to its hydrophobicity, which increases its affinity for the phospholipid bilayer of the bacterial cell membrane. 35 Bile acids are synthesized as conjugated bile acids; taurine or glycine is amide bonded to the carboxyl group of the free bile acid (Supplementary Table 1). These conjugated bile acids are then hydrolyzed to yield free bile acids in the small and large intestines by the bacterial bile salt hydrolase reaction. This reaction increases the pka of the molecules to the neutral range. Thus, free bile acids become weak acids, giving them strong bactericidal activities in the neutral physiologic ph range in which a substantial proportion of the bile acid molecules are in a nonionized form. Subsequently, the bacterially facilitated Figure 5. OTU analysis of rat cecal microbiota populations from different diet groups. OTUs are indicated by numbers. OTUs that comprise at least 4 sequences are displayed. Total sequence reads used for OTU analysis was 845 (control group, 276 reads; M-CA group, 279 reads; H-CA group, 290 reads). O-Fir, the other Firmicutes; O-Bac, the other Bacteroidetes; O-Pb, the other Proteobacteria; O-Ab, the other Actinobacteria. OTUs consisting of 1 to 3 sequences are combined into O-Fir, O-Bac, O-Pb, and O-Ab.

7 November 2011 BILE ACID REGULATES CECAL MICROBIOTA dehydroxylation reaction in the large intestine eliminates the functional hydroxyl group at C-7, further increasing the hydrophobicity of the molecules. Accordingly, the bactericidal activity of bile acid molecules generally increases as the molecules travel from the duodenum down to the distal colon. 35 Among these bile acids in the large intestine, DCA is extremely toxic, such that inclusion of 1 mmol/l DCA in growth medium severely inhibits the growth of many intestinal bacteria, including Clostridium perfringens, Bacteroides fragilis, lactobacilli, and bifidobacteria. 15,29 We calculated that CA feeding increased DCA concentrations to mmol/l in the rat cecum. Therefore, it is reasonable to infer that DCA applied a strong selective pressure that altered the composition of the cecal microbiota in the CA-fed groups. CA feeding increased the relative abundance of Firmicutes over Bacteroidetes in rat ceca (Figures 3 and 4). However, it is not clear why Firmicutes had a survival advantage over Bacteroidetes under these experimental conditions. From the OTU analysis, a reasonable explanation seems to be selection for specific strains of Firmicutes that have efficient systems for tolerating bile acid (Supplementary Discussion 1) and their subsequent establishment in niches in the rat cecum. Thus, bacteria belong to OTUs 1 to 4, 6, 7, and 9, which expanded in the M-CA and H-CA groups (Figure 5), must be more resistant to bile acid than those belong to the other OTUs. In accord with this interpretation, there are reports that C bartlettii (OTU7) 36 and E cylindroides (OTU9) 37 can tolerate a high concentration of bile. Furthermore, Enterobacteriaceae bacteria (class Gammaproteobacteria) are known to be highly tolerant of bile acid. 34 This is consistent with the preferential detection of OTU11 (E coli) in the H-CA group. To substantiate the role played by DCA in the control of the gut microbiota population, we have isolated bacteria from cecal contents and compared their sensitivities with DCA in in vitro growth experiments (see Supplementary Materials and Methods). Although not definitive, the DCA median inhibitory concentration values (DCA concentrations giving 50% growth inhibition) for bacteria belonging to a limited number of OTUs supported the hypothesis that the antimicrobial effect of bile acids induces a population shift at the phylum level, with significantly higher median inhibitory concentrations in Firmicutes (OTU6 [C innocuum] and OTU16 [Blautia coccoides]) ( mmol/l) than in Bacteroidetes (OTU21 [Bacteroides vulgatus] and OTU53 [Bacteroides sartorii]) ( mmol/l) and no growth inhibition detected in Gammaproteobacteria (OTU11) (Supplementary Results, Supplementary Figure 3, and Supplementary Table 11). It is interesting to note that the observed shift in the relative abundance of Firmicutes and Bacteroidetes in the cecal microbiota triggered by CA feeding resembles the phylum-level population shifts recently reported as high-fat diet induced alterations in a mouse model. 1 4 Because dietary fat increases bile secretion, 13 it can be speculated that these alterations in the bacterial population were promoted by increased DCA levels in the cecum resulting from increased fat intake. Moreover, OTU analysis revealed several phylogenetically overlapping bacteria that expanded under both conditions. For example, the genus Allobaculum has been reported to bloom in hamsters fed a diet supplemented with grain sorghum lipid extract. 8 Organisms closely related to C innocuum and Eubacterium dolichum (the latter was not detected in our study, but these 2 species are phylogenetically close relatives) have been reported to increase in fecal samples from mice fed a high-fat Western diet. 3 It has also been reported that the relative abundance of the order Enterobacteriales (Gammaproteobacteria) significantly increases in rats fed a high-fat diet. 5 High-fat diet induced decreases in bacterial population diversity (in mouse 1 and hamster 8 ) and in total cell density (in rat 5 ) have also been reported. Although we cannot yet draw any conclusions, these observations all suggest that bile acid is a mediator of high-fat diet induced gut microbiota alterations. High-fat diets not only induce Firmicutes-enriched microbiota but also induce obesity. 1 3 It has been proposed that increased digestive efficiency displayed by Firmicutesenriched microbiota enables enhanced SCFA production (extra caloric supply), which results in obesity. 1 However, our data showed that CA-induced Firmicutes-enriched microbiota resulted in reduced digestive efficiency with respect to SCFA production (Figure 1B). The major potential cause of these disparate physiologic traits of apparently similar bacterial communities is likely reduced metabolic activities of gut microbiota due to the antimicrobial activity of DCA as represented by the decreased cell density in the CA-fed groups (Figure 2). Another possible explanation could be increased consumption of SCFAs by the host due to the enhanced energy expenditure induced by CA through promotion of intracellular thyroid hormone activation 38 (see next paragraph) (Supplementary Discussion 2). We also investigated the impact of CA administration on host parameters (Table 1). In the rats in the H-CA group, a significant weight reduction in epididymal adipose tissue was observed, and similar trends were also observed in the other 2 adipose tissues. Such a difference was not found between rats in the control group and the M-CA group. It has been reported in mice that CA ingestion (5 g [12 mmol]/kg added to a high-fat diet) promotes energy expenditure by promoting intracellular thyroid hormone activation; taurocholic acid (TCA) was found to be most effective. 38 In our experiment, a large amount of TCA was found in the feces of rats in the H-CA group but not in that of rats in the control group or M-CA group (Figure 1A). Thus, it is possible that promotion of energy expenditure occurred in the rats in the H-CA group but not rats in the M-CA group, which explains the significant reduction of adipose tissue weight in the rats in the H-CA group. Considering the detected amounts of fecal TCA, the intestinal bile acid flow of the rats in the M-CA group still appeared within a normal range, whereas that of the rats in the H-CA group seemed to exceed the physiologic

8 1780 ISLAM ET AL GASTROENTEROLOGY Vol. 141, No. 5 range. Interestingly, serum adiponectin concentration decreased significantly in rats in both the M-CA and H-CA groups relative to the rats in the control group. Adiponectin is a protein secreted from white adipose tissue and is involved in the suppression of obesity-related disorders such as inflammation, atherosclerosis, and type 2 diabetes mellitus. 39 However, in this study, no symptom of inflammation was observed in the large intestinal mucosa in any of the CA-fed rats (Supplementary Figure 1). This indicated that CA-induced gut microbiota change took place independently of the inflammation under the conditions examined. CA feeding for longer than 10 days might be required to induce apparent inflammation. As shown in the present study, bile acid is clearly a host factor that is important for understanding alterations in the gut microbiota. Moreover, bile acid appeared to affect host health by modulating the secretion of adiponectin. These findings surely open a new door to a better understanding of the relationship between metabolic syndrome and the gut microbiota population. Supplementary Material Note: To access the supplementary material accompanying this article, visit the online version of Gastroenterology at and at doi: /j.gastro References 1. Turnbaugh PJ, Bäckhed F, Fulton L, et al. Diet-induced obesity is linked to marked but reversible alterations in the mouse distal gut microbiome. Cell Host Microbe 2008;3: Hildebrandt MA, Hoffmann C, Sherrill-Mix SA, et al. High-fat diet determines the composition of the murine gut microbiome independently of obesity. Gastroenterology 2009;137: Turnbaugh PJ, Ridaura VK, Faith JJ, et al. The effect of diet on the human gut microbiome: a metagenomic analysis in humanized gnotobiotic mice. Sci Transl Med 2009;1:6ra14. doi: / scitranslmed Murphy EF, Cotter PD, Healy S, et al. Composition and energy harvesting capacity of the gut microbiota: relationship to diet, obesity and time in mouse models. Gut 2010;59: de La Serre CB, Ellis CL, Lee J, et al. Propensity to high-fat diet-induced obesity in rats is associated with changes in the gut microbiota and gut inflammation. Am J Physiol Gastrointest Liver Physiol 2010;299:G440 G Cani PD, Neyrinck AM, Fava F, et al. Selective increases of bifidobacteria in gut microflora improve high-fat-diet-induced diabetes in mice through a mechanism associated with endotoxaemia. Diabetologia 2007;50: Cani PD, Delzenne NM. The role of the gut microbiota in energy metabolism and metabolic disease. Curr Pharm Des 2009;15: Martínez I, Wallace G, Zhang C, et al. Diet-induced metabolic improvements in a hamster model of hypercholesterolemia are strongly linked to alterations of the gut microbiota. Appl Environ Microbiol 2009;75: Larsen N, Vogensen FK, van den Berg FWJ, et al. Gut microbiota in human adults with type 2 diabetes differs from non-diabetic adults. PLoS ONE 2010;5:e Manichanh C, Rigottier-Gois L, Bonnaud E, et al. Reduced diversity of faecal microbiota in Crohn s disease revealed by a metagenomic approach. Gut 2006;55: Frank DN, St. Amand AL, Feldman RA, et al. Molecular-phylogenetic characterization of microbial community imbalances in human inflammatory bowel diseases. Proc Natl Acad Sci U S A 2007;104: Bajzer M, Seeley RJ. Obesity and gut flora. Nature 2006;444: Reddy BS. Diet and excretion of bile acids. Cancer Res 1981;41: Ridlon JM, Kang DJ, Hylemon PB. Bile salt biotransformations by human intestinal bacteria. J Lipid Res 2006;47: Kurdi P, Kawanishi K, Mizutani K, et al. Mechanism of growth inhibition by free bile acids in lactobacilli and bifidobacteria. J Bacteriol 2006;188: Ishizuka S, Iwama A, Dinoto A, et al. Synbiotic promotion of epithelial proliferation by orally ingested encapsulated Bifidobacterium breve and raffinose in the small intestine of rats. Mol Nutr Food Res 2009;53:S62 S Dewhirst FE, Chien CC, Paster BJ, et al. Phylogeny of the defined murine microbiota: altered Schaedler flora. Appl Environ Microbiol 1999;65: Ley RE, Bäckhed F, Turnbaugh P, et al. Obesity alters gut microbial ecology. Proc Natl Acad Sci U S A 2005;102: Cole JR, Wang Q, Cardenas E, et al. The Ribosomal Database Project: improved alignments and new tools for rrna analysis. Nucleic Acids Res 2009;37:D141 D Wang Q, Garrity GM, Tiedje JM, et al. Naïve Bayesian classifier for rapid assignment of rrna sequences into the new bacterial taxonomy. Appl Environ Microbiol 2007;73: Tamura K, Dudley J, Nei M, et al. MEGA4: Molecular Evolutionary Genetics Analysis (MEGA) software version 4.0. Mol Biol Evol 2007;24: Dinoto A, Suksomcheep A, Ishizuka S, et al. Modulation of rat cecal microbiota by administration of raffinose and encapsulated Bifidobacterium breve. Appl Environ Microbiol 2006;72: Amann RI, Binder BJ, Olson RJ, et al. Combination of 16S rrnatargeted oligonucleotide probes with flow cytometry for analyzing mixed microbial populations. Appl Environ Microbiol 1990;56: Franks AH, Harmsen HJM, Raangs GC, et al. Variations of bacterial populations in human feces measured by fluorescent in situ hybridization with group-specific 16S rrna-targeted oligonucleotide probes. Appl Environ Microbiol 1998;64: Takaishi H, Matsuki T, Nakazawa A, et al. Imbalance in intestinal microflora constitution could be involved in the pathogenesis of inflammatory bowel disease. Int J Med Microbiol 2008;298: Hofmann AF, Hagey LR. Bile acids: chemistry, pathochemistry, biology, pathobiology, and therapeutics. Cell Mol Life Sci 2008; 65: Fukiya S, Arata M, Kawashima H, et al. Conversion of cholic acid and chenodeoxycholic acid into their 7-oxo derivatives by Bacteroides intestinalis AM-1 isolated from human feces. FEMS Microbiol Lett 2009;293: Bernstein H, Bernstein C, Payne CM, et al. Bile acids as carcinogens in human gastrointestinal cancers. Mutat Res 2005;589: Floch MH, Binder HJ, Filburn B, et al. The effect of bile acids on intestinal microflora. Am J Clin Nutr 1972;25: Eckburg PB, Bik EM, Bernstein CN, et al. Diversity of the human intestinal microbial flora. Science 2005;308: Liu C, Finegold SM, Song Y, et al. Reclassification of Clostridium coccoides, Ruminococcus hansenii, Ruminococcus hydrogenotrophicus, Ruminococcus luti, Ruminococcus productus and Ruminococcus schinkii as Blautia coccoides gen.nov., comb. nov., Blautia hansenii comb. nov., Blautia hydrogenotrophica comb. nov., Blautia luti comb. nov., Blautia producta comb. nov., Blautia schinkii comb. nov.

9 November 2011 BILE ACID REGULATES CECAL MICROBIOTA 1781 and description of Blautia wexlerae sp. nov., isolated from human faeces. Int J Syst Evol Microbiol 2008;58: Greetham HL, Gibson GR, Giffard C, et al. Allobaculum stercoricanis gen. nov., sp. nov., isolated from canine feces. Anaerobe 2004;10: Stacey M, Webb M. Studies on the antibacterial properties of the bile acids and some compounds derived from cholanic acid. Proc R Soc Lond B 1947;134: Floch MH, Gershengoren W, Elliott S, et al. Bile acid inhibition of the intestinal microflora-a function for simple bile acids? Gastroenterology 1971;61: Margolles A, Yokota A. Bile acid stress in lactic acid bacteria and bifidobacteria. In: Sonomoto K, Yokota A, ed. Lactic acid bacteria and bifidobacteria: current progress in advanced research. Norfolk, VA: Caister Academic Press, 2011: Song YL, Liu CX, McTeague M, et al. Clostridium bartlettii sp. nov., isolated from human faeces. Anaerobe 2004;10: Wade WG. Genus I. Eubacterium Prévot 1938, 294 AL. In: De Vos P, Garrity GM, Jones D, et al, eds. Bergey s manual of systematic bacteriology, the Firmicutes. Volume 3. 2nd ed. Dordrecht, The Netherlands: Springer, 2009: Watanabe M, Houten SM, Mataki C, et al. Bile acids induce energy expenditure by promoting intracellular thyroid hormone activation. Nature 2006;439: Brochu-Gaudreau K, Rehfeldt C, Blouin R, et al. Adiponectin action from head to toe. Endocrine 2010;37: Received March 27, Accepted July 25, Reprint requests Address requests for reprints to: Atsushi Yokota, PhD, Laboratory of Microbial Physiology, Research Faculty of Agriculture, Hokkaido University, Sapporo, , Japan. yokota@chem. agr.hokudai.ac.jp; fax: (81) Acknowledgments Polymerase chain reaction derived 16S ribosomal RNA gene sequences have been deposited in the DDBJ/EMBL/GenBank under accession number AB AB620513, AB AB621132, AB AB621136; AB AB621139, AB AB621211, and AB AB K.B.M.S.I., S.F. and M.H. contributed equally to this work. Conflicts of interest The authors disclose no conflicts. Funding Supported in part by a Grant-in-Aid for Scientific Research (B) from the Japan Society for the Promotion of Science (no to A.Y.).