- Email: [email protected]
Structural changes of gut microbiota in a rat non-alcoholic fatty liver disease model treated with a Chinese herbal formula
Systematic and Applied Microbiology 36 (2013) 188–196
Contents lists available at SciVerse ScienceDirect
Systematic and Applied Microbiology journal homepage: www.elsevier.de/syapm
Structural changes of gut microbiota in a rat non-alcoholic fatty liver disease model treated with a Chinese herbal formula Xiaochen Yin a , Jinghua Peng b , Liping Zhao a,c , Yunpeng Yu a , Xu Zhang a , Ping Liu d , Qin Feng b , Yiyang Hu b,d,∗∗ , Xiaoyan Pang a,∗ a
State Key Laboratory of Microbial Metabolism, School of Life Sciences & Biotechnology, Shanghai Jiao Tong University, Shanghai 200240, China Institute of Liver Diseases, Shuguang Hospital, Shanghai University of Traditional Chinese Medicine, Shanghai 201203, China c Ministry of Education Key Laboratory of Systems Biomedicine, Shanghai Center for Systems Biomedicine, Shanghai 200240, China d E-Institute of TCM Internal Medicine, Shanghai Municipal Education Commission, Shanghai 201203, China b
a r t i c l e
i n f o
Article history: Received 15 August 2012 Received in revised form 6 December 2012 Accepted 6 December 2012 Keywords: Chinese herbal formula Gut microbiota Non-alcoholic fatty liver disease (NAFLD) Pyrosequencing Polymerase chain reaction-denaturing gradient gel electrophoresis (PCR-DGGE)
a b s t r a c t Accumulating evidence indicates that disruption of the gut microbiota by a high-fat diet (HFD) may play a pivotal role in the progression of metabolic disorders such as non-alcoholic fatty liver disease (NAFLD). In this study, the structural changes of gut microbiota were analyzed in an HFD-induced NAFLD rat model during treatment with an ancient Chinese herbal formula (CHF) used in clinical practice – Qushi Huayu Fang. CHF treatment significantly reduced body weight, alleviated hepatic steatosis, and decreased the content of triglycerides and free fatty acids in the livers of the rats. Gut microbiota of treated and control rats were profiled with polymerase chain reaction-denaturing gradient gel electrophoresis and bar-coded pyrosequencing of the V3 region of 16S rRNA genes. Both analyses indicated that the CHFtreated group harbored significantly different gut microbiota from that of model rats. Partial least squares discriminant analysis and taxonomy-based analysis were further employed to identify key phylotypes responding to HFD and CHF treatment. Most notably, the genera Escherichia/Shigella, containing opportunistic pathogens, were significantly enriched in HFD-fed rats compared to controls fed normal chow (P < 0.05) but they decreased to control levels after CHF treatment. Collinsella, a genus with short chain fatty acid producers, was significantly elevated in CHF-treated rats compared to HFD-fed rats (P < 0.05). The results revealed that the bacterial profiles of HFD-induced rats could be modulated by the CHF. Elucidation of these differences in microbiota composition provided a basis for further understanding the pharmacological mechanism of the CHF. © 2013 Elsevier GmbH. All rights reserved.
Introduction Non-alcoholic fatty liver disease (NAFLD), a manifestation of the metabolic syndrome in the liver, is closely associated with insulin resistance, obesity, and diabetes mellitus [27]. Although it is a major threat to public health, the pathogenesis of NAFLD is not yet well understood, nor are there effective therapies [28]. Emerging evidence in the last two decades indicates a possible role of gut microbiota in the etiology of NAFLD [1]. Small intestinal bacterial overgrowth is commonly found in NAFLD patients and results in increased intestinal permeability, endogenous ethanol production, and choline deficiency [3,11,44]. In
∗ Corresponding author at: State Key Laboratory of Microbial Metabolism, School of Life Sciences & Biotechnology, Shanghai Jiao Tong University, Shanghai 200240, China. Tel.: +86 21 34204878; fax: +86 21 34204878. ∗∗ Corresponding author. Tel.: +86 21 20256160; fax: +86 21 20256521. E-mail addresses: [email protected] (Y. Hu), [email protected] (X. Pang). 0723-2020/$ – see front matter © 2013 Elsevier GmbH. All rights reserved. http://dx.doi.org/10.1016/j.syapm.2012.12.009
addition, the gut microbiota promotes absorption of monosaccharides and short chain fatty acids by fermentation, and thus increases de novo hepatic lipogenesis and enhances fat storage by regulating lipoprotein lipase activity [12]. NAFLD is typically associated with low-grade, chronic inflammation [13], and a major cell wall component of Gram-negative bacteria, lipopolysaccharide (LPS), is known to be potent in promoting inflammation [5]. When binding to CD14 and toll-like receptor 4 on the surface of immune cells, a low concentration of LPS from gut microbiota can trigger a series of inflammatory processes [31,45], which in turn contributes to the pathogenesis of NAFLD. Consequently, the gut microbiota has become a novel target for potential therapies [14], with probiotics, prebiotics, and antibiotics being tested for treatment or prevention of NAFLD [4,19,22]. Traditional medicine has long been used in China, and most of the formulas are orally administered. Recent research has revealed that many ingredients in these herbs could only be absorbed and they exert their biological effects with the help of the host’s gut microbiota [8]. Furthermore, intriguing but limited data also
X. Yin et al. / Systematic and Applied Microbiology 36 (2013) 188–196
189
Table 1 Herbs used in the Chinese herbal formula. Chinese name
Pharmaceutical name
Family name
Production place
Processing method
Yin chen
Artemisia capillaries Thunb. (above-ground parts, dried) Polygonum cuspidatum Sieb. et Zucc (rhizome, root, dried) Curuma longa L. (rhizome, dried) Hypericum japonicum (whole plant, dried) Gardenia jasminoides Ellis (fruit, dried)
Compositae
Anhui Province, China
Ethanol extraction
Polygonaceae
Jiangsu Province, China
Ethanol extraction
Zingiberaceae Clusiaceae Rubiaceae
Sichuang Province, China Jiangxi Province, China Fujian Province, China
Ethanol extraction Mixed at an equal mass ratio in distilled water, then concentrated to a relative density of 1.08–1.12 (80 ◦ C), followed by ethanol precipitation
Hu zhang Jiang huang Tian ji huang Zhi zi
show that the efficacy of some traditional herbs may be related to modulation of gut microbiota [48]. One formula, Qushi Huayu Fang, has a long history of use in clinical practice to alleviate NAFLD. However, to date, no systematic and credible explanations have been proposed for its mechanism. Whether Qushi Huayu Fang can modulate the host’s gut ecosystem while alleviating NAFLD symptoms and whether gut bacteria are a potential target for the herbal formula are still unanswered questions. Nowadays, advanced technologies have facilitated the exploration of gut microbiota, especially with the development of high-throughput sequencing technology. Researchers are able to look in detail at the structure of gut microbiota, and study how external factors, such as age and drug use among others, influence gut microbiota. In this study, gut microbial composition during Qushi Huayu Fang treatment in a high-fat diet (HFD)-induced NAFLD rat model was monitored by polymerase chain reactiondenaturing gradient gel electrophoresis (PCR-DGGE) and bar-coded pyrosequencing. Using a microbiome-wide association strategy [34,41,47], the modulation of this Chinese herbal formula (CHF) in the gut microbiome was validated, and key phylotypes closely related to NAFLD development and CHF treatment were expected to be identified. Materials and methods Preparation of the CHF The compound was prepared from five herbs, as listed in Table 1. Herbs were obtained from qualified suppliers and on the basis of standards specified in the Chinese Pharmacopoeia (1995 edition) and Chinese Materia Medica in Shanghai. Extracts of the herbs (Table 1) were mixed at a mass ratio of 13:7:7:7:7 (13 parts Artemisia capillaries Thunb., and 7 parts of each of the other four herbs). High-performance liquid chromatography was used for quality control by monitoring three active components in the formula: polydatin (retention time: 10 min; concentration: 1.779 mg g−1 ), emodin (19 min; 5.394 mg g−1 ), and geniposide (35 min; 2.850 mg g−1 ). The final solution was stored at 4 ◦ C at two concentrations, 0.93 and 0.47 g mL−1 , as high and low dosages, respectively. The formula has been issued a patent by the State Intellectual Property Office of PR China under the ID: ZL200610009140.0. Animals and experimental protocol Ethics statement All procedures of the experiments in this study were approved by the Animal Ethics Committee of the Shanghai University of Traditional Chinese Medicine (Approval No. SCXK 2003-0003). The care and use of animals were carried out under the Guidelines for Animal Experiment of the Shanghai University of Traditional Chinese Medicine (Shanghai, China), and all efforts were made to minimize the number of animals and their suffering [49].
Twenty-six male Sprague–Dawley rats were purchased from the Shanghai Laboratory Animal Co. Ltd., China, and raised in a specific pathogen-free barrier system in the Laboratory Animal Center of the Shanghai Traditional Chinese Medicine University. After acclimatization, rats were divided into two groups: one group (n = 21) was fed an HFD (w/w, 87.5% normal chow, 10% lard, 2% cholesterol, and 0.5% sodium cholate), and a control group (n = 5) was fed normal chow (NC). After 6 weeks on the diet, HFD-fed rats were randomly divided into three subgroups (each n = 7). Two of the subgroups maintained the HFD diet in conjunction with high (H) and low (L) dosage (0.93 g/100 g body weight and 0.47 g/100 g body weight, respectively) CHF (Qushi Huayu Fang) through intragastric administration. The third subgroup remained on the HFD without additional treatment (HFD). Food intake in the groups was controlled to the same level. Control volumes (equivalent to those of the CHF compound given to experimental groups) containing water were administrated to NC- and HFD-fed rats. Body weight was measured every 14 days. At the end of week 10 (i.e. the CHF treatment lasted for 4 weeks), all animals from the four groups were sacrificed and their livers were removed and stored at −70 ◦ C for histological and lipid content analysis, including oil Red O and hematoxylin–eosin (H&E) staining, TG and FFA analysis (see Supplementary Information). Liver index was calculated as the ratio of liver-to-body weight. Before sacrifice, fecal samples were collected from 23 animals, including 6 from HFD, 5 NC, 5 CHF (H), and 7 CHF (L). Unfortunately, the feces collected from the other three animals were insufficient, so they were omitted from the gut microbiota analysis. All stool samples were stored at −70 ◦ C.
PCR-DGGE and multivariate analysis of the 16S rRNA gene V3 region DNA was extracted from each fecal sample by the bead-beating method [50]. Bacterial universal primers P2 (5 ATTACCGCGGCTGCTGG-3 ) and P3 (5 -CGCCCGCCGCGCGCGGCGGGCGGGGCGGGGGCACGGGGGGCCTACGGGAGGCAGCAG-3 ) were used to amplify the 16S rRNA gene V3 region from each DNA sample by hot-start touchdown PCR [30]. PCR amplification reactions, DGGE analysis, and phylogenetic identification of major bands were performed as described previously [46] and the DGGE denaturing gradient was 30–54% (7 M urea and 40% deionized formamide were considered to be the 100% denaturant). The sequences retrieved from major DGGE bands have been deposited in GenBank (http://www.ncbi.nlm.nih.gov/Genbank/) with accession numbers JF729214–JF729235. The gel images were converted into digital data using Quantity One 4.4.0 (Bio-Rad, Hercules, CA, USA). Principal component analysis (PCA) was employed to compare the gut microbiota composition between treatment groups in the MATLAB 7.11.0 (R2010b) environment (The MathWorks, Inc., Natick, MA, USA).
190
X. Yin et al. / Systematic and Applied Microbiology 36 (2013) 188–196
Bar-coded pyrosequencing and multivariate analysis of the 16S rRNA gene V3 region Forward 5 -NNNNNNNNATTACCGCGGCTGCT-3 and reverse primers were used to amplify the 16S rRNA gene V3 region. The NNNNNNNN was the sample-unique 8-base barcode used for sorting of PCR amplicons. PCR reactions, pyrosequencing, and sequence quality controls were performed as described previously with minor modification [46]. After cluster database at high identity with tolerance (CD-HIT) clustering (100% identity), sequences were aligned using nearest alignment space termination (NAST) [10,21]. Qualified V3 region sequences were imported to the ribosomal database project (RDP) classifier for taxonomical assignment at a 50% confidence level [7]. The sequences were also imported into the ARB software environment [24] in order to calculate the distance matrix based on the neighbor-joining algorithm for online UniFrac analysis [18]. A distance matrix was also used to define the operational taxonomic unit (OTU) in DOUTR [38]. Rarefaction analysis and the Shannon diversity index were calculated as described previously [46]. The unique sequences from pyrosequencing have been deposited in the GenBank Sequence Read Archive database (http://www.ncbi.nlm.nih.gov/Traces/sra) with accession number SRA054076. Partial least squares discriminant analysis (PLS-DA) was used to test whether these groups could be separated based on the OTU data. The correct classification rate of the PLS-DA model was performed with leave-one-out cross-validation [33]. Martens’ uncertainty test [43] followed by the Mann–Whitney test (P < 0.05) were used to select important OTUs for classification. All multivariate statistical analyses were performed in the MATLAB 7.11.0 (R2010b) environment (The MathWorks, Inc.).
vs. 601.22 ± 74.73 mol g−1 protein, CHF (H) vs. HFD; P < 0.05, 394.79 ± 14.28 mol g−1 vs. 601.22 ± 74.73 mol g−1 protein, CHF (L) vs. HFD; Fig. 1F).
5 -NNNNNNNNCCTACGGGAGGCAGCAG-3
Results Weight reduction and alleviation of NAFLD in rats treated with the CHF At the end of a 10-week intervention period, NC group rats weighed significantly less than the HFD group (P < 0.01; 484.20 ± 20.76 g vs. 540.29 ± 14.02 g, respectively; Fig. 1A). Rats in the CHF high dosage treatment (CHF (H)) and CHF low dosage treatment (CHF (L)) groups also weighed less than those in the HFD group (P < 0.01, 439.86 ± 7.22 g vs. 540.29 ± 14.02 g, CHF (H) vs. HFD; P < 0.01, 438.71 ± 13.94 g vs. 540.29 ± 14.02 g, CHF (L) vs. HFD) and even less than NC rats. Liver histological analysis showed extensive micro/macrovesicular steatosis in the hepatocytes of the HFD group (Fig. 1B and C). In this group, large lipid droplets filled the cytoplasm of hepatocytes, displacing the nucleus to the periphery. The two groups of CHF-treated rats had significantly less fat deposition in hepatocytes compared to their HFD counterparts. CHF high-dosage treatment exerted a better effect than low-dosage. The liver index showed that all HFD-fed animals, including the HFD group and the two CHF groups, had higher liver indices than the NC group, while the significantly reduced index was observed in the CHF (H) group compared to the HFD group (P < 0.05; 0.02 ± 6.03 E−4 vs. 0.04 ± 5.27 E−4, CHF (H) vs. HFD; Fig. 1D). For liver triglycerides (TG), a significant decrease was also observed in all CHF-treated rats compared to HFD rats (P < 0.01; 22.98 ± 1.30 mg g−1 vs. 40.12 ± 1.64 mg g−1 , CHF (H) vs. HFD; Fig. 1E), and the effect was greater in CHF (H)-treated than CHF (L)-treated rats. Liver free fatty acids (FFA) analysis had a similar trend to the TG. The FFA content of the two CHF groups was also significantly reduced (P < 0.01, 294.27 ± 20.47 mol g−1
Structural changes in gut microbiota revealed by PCR-DGGE fingerprints According to the 16S rDNA V3 region PCR-DGGE patterns, the overall composition after the CHF (H) treatment showed a significant difference from that in rats fed NC or HFD (Fig. 2A). The PCA of the fingerprints showed that rats fed NC, HFD and CHF (H) fell into three clusters, with the first principal component (PC1) accounting for 32.8% of the variation (Fig. 2B). In HFD-fed rats, the intensities of band 1 (b1), b2, b3 and b4 were significantly reduced. By sequence analysis, these bands were related to Prevotella copri strain CB7 (98% homology), Prevotella sp. DJF RP53 (98%), Prevotella copri strain CB7 (99%) and Bacteroides plebeius (96%), respectively (Fig. 2 and Supplementary Table S1). Rats in the HFD group showed different dominant bands, including b5, b6 and b7, which were closest to Coprococcus catus GD/7 (100%), Ruminococcus gauvreauii strain CCRI-16110 (100%), and Escherichia coli strain CAIM 1647 (100%), respectively. In CHF (H)-treated rats, there was a significant difference in the composition of gut microbiota compared to other groups. The intensities of b5, b6 and b7 were significantly reduced in some CHF (H)-treated rats. Notably, b9 (Collinsella aerofaciens strain KCTC5916, 97% homology) was promoted in the CHF (H) group, but b8 (Clostridium sordellii strain 2TT1, 100%), which was apparent in NC and HFD rats, was not detected in this group. DGGE analysis was also performed in order to test for a difference between the effects of CHF (H) and CHF (L) on gut microbiota. However, there were no significant differences in the fingerprints of the two dosages (Fig. 2C), as confirmed by PCA (Fig. 2D), since the two groups could not be divided into different clusters.
Structural changes of gut microbiota revealed by bar-coded pyrosequencing The high-throughput technology of bar-coded pyrosequencing was also introduced in order to monitor the structure shift of gut microbiota for all four groups of rats. In total, 41,375 high quality sequences were obtained with an average of 1500 per sample (Supplementary Fig. S1A). Using 97% identity as the cutoff, 1240 OTUs were delineated (Supplementary Fig. S1B). The rarefaction curves for all animals did not reach a plateau at the sequencing depth used (Supplementary Fig. S1C). However, the Shannon diversity indices all reached stable values, indicating that bacterial diversity in these communities was mostly covered (Supplementary Fig. S1D). Unweighted UniFrac analysis was used to discriminate the microbiota composition of the different groups based on the evolutionary distance. All samples fell into three clusters with CHF (H) and CHF (L) in one cluster (Fig. 3A and B). CHF (H) and CHF (L) were then combined into a single CHF treatment group in the subsequent analysis. Taxonomy-based analysis at the phylum level showed that there were much less Bacteroidetes in HFD- vs. NC-fed rats (P < 0.05; median 7.09% vs. 38.80%, respectively; Fig. 4A) and more Firmicutes (P < 0.05; 83.83% vs. 56.19%; Fig. 4B), whereas Proteobacteria did not change significantly (P = 0.11; 4.10% vs. 6.02%, NC vs. HFD; Fig. 4C). After CHF treatment, the relative abundance of these three phyla did not change significantly but Actinobacteria increased significantly (P < 0.05; 3.86% vs. 1.25%, CHF vs. HFD), and was even higher than in NC-fed rats (P < 0.05; 3.86% vs. 0.47%, CHF vs. NC; Fig. 4D).
X. Yin et al. / Systematic and Applied Microbiology 36 (2013) 188–196
191
Fig. 1. Alleviation of non-alcoholic fatty liver disease by the Chinese herbal formula (CHF). (A) Growth curves based on average body weight of the four groups measured every 14 days. (B) Oil Red O staining of the rat hepatocytes (magnification, 200×). (C) Hematoxylin–eosin (H&E) staining of hepatocytes (magnification, 400×). (D) Liver index (the ratio of liver to body weight) at the end of week 10. (E) Liver triglycerides at the end of week 10. (F) Liver free fatty acids at the end of week 10. Data are expressed as mean values ± SE. **P < 0.01 and *P < 0.05 vs. HFD; # P < 0.05 vs. CHF (H) by one-way analysis of variance (ANOVA) followed by least-significant difference post hoc multiple comparison. Note: Part of the phenotypic data from an animal in this study has been published previously [20]. NC: normal chow-fed group; HFD: high-fat diet-fed group; CHF (H): Chinese herbal formula high dosage treatment group; CHF (L): Chinese herbal formula low dosage treatment group.
Taxonomy-based comparison at the genus level further showed that, in healthy NC-fed rats, Prevotella and Bacteroides from phylum Bacteroidetes, and the seven genera of Lactobacillus, Oscillibacter, Blautia, Coprococcus, Faecalibacterium, Sporacetigenium, and Allobaculum from phylum Firmicutes were predominant (each genus had a relative abundance >1%). While in HFD-fed rats, Prevotella, Bacteroides, and Allobaculum significantly decreased, whereas Escherichia/Shigella, Coprococcus, Blautia, and Roseburia increased. CHF treatment significantly reduced Escherichia/Shigella (P < 0.01; median 0.06% vs. 1.45%, CHF vs. HFD) and also significantly diminished Sporacetigenium (P < 0.05; 0.39% vs. 2.42%, CHF vs. HFD), while Collinsella was significantly enhanced (P < 0.05; 3.15% vs. 0.55%, CHF vs. HFD) after CHF treatment (Fig. 5 and Supplementary Table S2).
Key OTUs responding to HFD and CHF treatment identified by partial least squares discriminant analysis (PLS-DA) PLS-DA models were established between the NC vs. HFD groups and HFD vs. CHF groups. Score plots based on the first two components showed that different groups were well separated (Supplementary Fig. S2A and S2B). In leave-one-out cross-validation, the PLS-DA model between NC vs. HFD groups with three components yielded the high correct classification rate of nearly 100%. The Martens’ uncertainty and Mann–Whitney tests (P < 0.05) further identified 41 OTUs as key variables for classification in this model. The PLS-DA model between HFD vs. CHF with five components yielded a correct classification rate of 78%. Seven phylotypes were also identified using the same protocol. In all, 45 OTUs responding to HFD and CHF treatments were identified (Fig. 6 and Supplementary Table S3). In HFD-fed rats, OTUs belonging to Porphyromonadaceae (5 OTUs), Prevotella (4 OTUs), Bacteroides (2 OTUs), Parabacteroides (1 OTU), Bacteroidales (1 OTU), Allobaculum (1 OTU), Veillonellaceae (1 OTU), Parasutterella
(1 OTU), and Desulfovibrio (1 OTU) were significantly reduced, whereas Escherichia/Shigella (1 OTU), Erysipelotrichaceae (1 OTU) and Peptococcaceae (1 OTU) were significantly increased. Although one OTU belonging to Collinsella was significantly increased in the HFD group compared to NC, taxonomy-based analysis showed that there was no significant difference for Collinsella (including 204 unique sequences) between these two groups. CHF treatment significantly reduced the abundance of Barnesiella (1 OTU), Peptococcaceae (1 OTU), Erysipelotrichaceae (1 OTU), Clostridiales (1 OTU), Lactobacillus (1 OTU), Coprococcus (1 OTU) and Ruminococcaceae (2 OTUs). Notably, OTU 19 belonging to Escherichia/Shigella was also significantly reduced after CHF treatment (P < 0.05; median 0.06% vs. 1.39%, CHF vs. HFD). Another OTU (OTU 11), belonging to Lactobacillus, was significantly increased after CHF treatment. However, OTUs belonging to the same family/genus, such as in Ruminococcaceae (9 OTUs), Lactobacillus (4 OTUs), Barnesiella (4 OTUs), Lachnospiraceae (3 OTUs), Clostridiales (2 OTUs) and Coprococcus (2 OTUs), responded differently to HFD and CHF treatment (Supplementary Table S3). For example, of the nine OTUs belonging to Ruminococcaceae, four decreased in the HFD group but increased in the CHF group. Another two OTUs in Ruminococcaceae increased in the HFD group but decreased in the CHF group, and the remaining three OTUs were reduced in both the HFD and CHF groups vs. the NC group.
Discussion Due to the close anatomical relationship and metabolic interactions between the liver and intestinal tract, imbalances in gut microbiota may influence diverse physiological processes [1,11]. Li et al. [22] reported that treatment with a probiotic (VSL#3) or anti-TNF-␣ antibodies alleviated NAFLD in ob/ob mice, leading to an improvement in liver steatosis, as well as reduced levels
192
X. Yin et al. / Systematic and Applied Microbiology 36 (2013) 188–196
Fig. 2. Comparison of gut microbiota by denaturing gradient gel electrophoresis (DGGE) fingerprinting. (A) DGGE fingerprints of the V3 region of 16S rRNA genes from fecal bacterial communities of NC, HFD and CHF (H). (B) Principal component analysis (PCA) of DGGE fingerprints in (A). (C) DGGE profiles of HFD, CHF (H), and CHF (L) groups. (D) PCA analysis of DGGE fingerprints in (C). NC: normal chow-fed group; HFD: high-fat diet-fed group; CHF (H): Chinese herbal formula high dosage treatment group; CHF (L): Chinese herbal formula low dosage treatment group. PC: principal component. Numbers at the top of each lane correspond to the animal IDs in the experiment.
Fig. 3. The overall gut microbiota structural change revealed by 454 pyrosequencing. (A) UniFrac analysis (unweighted) based on the pyrosequencing operational taxonomic units (OTUs) (97% similarity level) data. Each point represents the mean of PC scores in a group at the end of week 10, and error bars represent the standard error. (B) Clustering of gut microbiota based on distances between the groups, calculated by multivariate analysis of variance tests of the first 10 PCs of UniFrac (unweighted) distance. **P < 0.01. NC: normal chow-fed group; HFD: high-fat diet-fed group; CHF (H): Chinese herbal formula high dosage treatment group; CHF (L): Chinese herbal formula low dosage treatment group. PC: principal coordinate.
X. Yin et al. / Systematic and Applied Microbiology 36 (2013) 188–196
193
Fig. 4. Taxonomy-based analysis at the phylum level. In each of the panels, the phylum analyzed is indicated by the label on the y-axis. *P < 0.05 by the Mann–Whitney test. The boundary of the box closest to zero indicates the 25th percentile, a line within the box marks the median, and the boundary of the box farthest from zero indicates the 75th percentile. Whiskers (error bars) above and below the box indicate the 90th and 10th percentiles. NC: normal chow-fed group; HFD: high-fat diet-fed group; CHF: Chinese herbal formula treatment group.
of hepatic total fatty acid and serum alanine aminotransferase. Further clinical trials showed VSL#3 exerted beneficial functional effects on chronic liver diseases in humans [23]. Oligofructose, a widely studied prebiotic, has been reported to protect rats against liver TG accumulation induced by fructose but was not able to prevent fructose-induced hypertriglyceridemia [19]. Antibiotic treatment of fructose-fed mice has been shown to markedly reduce hepatic lipid accumulation, liver index, and portal endotoxin levels [4]. However, the mechanisms underlying the beneficial effects of probiotics and prebiotics in NAFLD are not well characterized. Wang et al. [42] used a culture-based technology to analyze changes of gut microbiota in rats fed a cholesterol-enriched diet supplemented with Lactobacillus plantarum MA2 for 5 weeks. This
probiotic strain was found to significantly lower total hepatic cholesterol and TG, as well as serum total cholesterol, low-density lipoprotein cholesterol and TG. There were increases in lactic acid bacteria and bifidobacteria, while E. coli was unchanged. However, the use of enrichment media to culture these three groups of bacteria has limitations, because most of the gut microbes are unknown and resistant to traditional culture methods [2]. In our study, comparable beneficial effects of one CHF (Qushi Huayu Fang) against NAFLD were observed. Liver histology in CHFtreated rats showed significant improvement with less hepatocyte steatosis and inflammation. Hepatic TG and FFA contents were also significantly reduced after CHF treatment. DGGE fingerprints and high-throughput bar-coded pyrosequencing based on the 16S rRNA
Fig. 5. Taxonomy-based analysis at the genus level. In each of the panels, the genus analyzed is indicated by the label on the y-axis. **P < 0.01 and *P < 0.05 by the Mann–Whitney test. The boundary of the box closest to zero indicates the 25th percentile, a line within the box marks the median, and the boundary of the box farthest from zero indicates the 75th percentile. Whiskers (error bars) above and below the box indicate the 90th and 10th percentiles, respectively. NC: normal chow-fed group; HFD: high-fat diet-fed group; CHF: Chinese herbal formula treatment group.
194
X. Yin et al. / Systematic and Applied Microbiology 36 (2013) 188–196
Fig. 6. Abundance distribution of the 45 phylotypes identified as key classification variables. The OTUs were based on the pyrosequencing (97% similarity level) data. Partial least squares discriminant analysis (PLS-DA) was used followed by Martens’ uncertainty test and the Mann–Whitney test (P < 0.05) in order to identify the key OTUs. To show the distribution of the OTUs with lower abundance, the colored squares of each column have been scaled to indicate the relative ratios of the OTU between the animals. The phylogenetic relationships of these OTUs are displayed to the left of the heatmap, with animals arranged from left to right according to the increase in liver triglyceride content. NC: normal chow-fed group; HFD: high-fat diet-fed group; CHF (H): Chinese herbal formula high dosage treatment group; CHF (L): Chinese herbal formula low dosage treatment group. The numbers along the x-axis are animal IDs in each of the groups.
gene V3 region were used to detect wider overall changes in microbiota. Previous studies have shown that some CHFs are capable of modulating gut microbiota. Kato et al. [17] observed that Juzentaihoto could decrease unculturable bacteria and increase Lactobacillus johnsoni in mice, which in turn modulated heat shock protein gene expression. The main components of some herbs, such as the polysaccharides of cassiae seeds, have been shown to promote growth of Lactobacillus and Bifidobacterium while inhibiting some opportunistic pathogens in piglets [9]. However, none of these studies provide a complete profiling of the gut microbiota. Through DGGE and UniFrac analysis based on pyrosequencing data, we were able to observe a significant structural difference between gut microbiota in CHF-treated rats and their NAFLD counterparts. Multivariable analysis, such as PLS-DA, enabled some key phylotypes to be pinpointed that may be relevant to the development of NAFLD and the action of the CHF studied. In the present study, Escherichia/Shigella were most abundant in HFD-fed rats and significantly reduced in CHF-treated rats. This is a group of Gram-negative bacteria that contain LPS in their cell walls. It is known that a large number of Gram-negative bacteria in the gut may impair the gut barrier, thus releasing LPS into the blood, and triggering low-grade chronic inflammation [5]. Escherichia/Shigella are reported to be particularly efficient at translocation [39] and are closely associated with small intestinal bacterial overgrowth. They may also contribute to the high concentration of serum LPS and TNF-␣ expression usually found in NAFLD patients [35]. In this study, reduced levels of Escherichia/Shigella after CHF treatment may suggest a possible reduced antigen load and improved inflammatory status of the host. However, the causal relationship between the microbiome and the disease still needs further confirmation. Among the key variables identified, Collinsella increased more than 6-fold in CHF-treated animals, which correlated with significantly alleviated NAFLD compared to that in HFD-fed rats. However, there was no difference in Collinsella between NC- and HFD-fed rats. Promotion of Collinsella specifically by CHF treatment may contribute to the different position of CHF group in UniFrac plot from
NC and HFD group. Bacteria in this genus have been reported to ferment carbohydrates to hydrogen, ethanol, formate, and lactate [15]. Tannock et al. [40] reported that consumption of oligosaccharidecontaining biscuits elevated the number of C. aerofaciens in human feces in 11 out of 15 subjects. Thus, the ability of Collinsella to produce short-chain fatty acids (SCFA) may contribute to the efficacy of this formula, since SCFA have been reported to stimulate epithelial cell proliferation and play an important role in improving gut barrier integrity [36,37]. High levels of C. aerofaciens were also closely associated with a lower risk for colon cancer in a human cohort including 18 polyp patients, 15 Japanese-Hawaiians, 17 North American Caucasians, 22 rural native Japanese and 16 rural Africans [29]. Less Collinsella has been detected in patients with irritable bowel syndrome compared to healthy controls [16]. A negative correlation between Collinsella and body mass index was further observed in irritable bowel syndrome patients [26]. Lower levels of Collinsella have also been observed in elderly people taking non-steroidal anti-inflammatory drugs vs. young people, or in elderly people taking no drugs [25], which indicates an underlying relationship between Collinsella and inflammation. The significant increase of Collinsella in our CHF-treated rats may therefore suggest some other unknown protective mechanism that warrants further investigation. The major effects on the balance of gut microbiota by CHF treatment could be reduction of certain opportunistic pathogens, such as Escherichia/Shigella, accompanied by an increase of potentially beneficial bacteria like Collinsella. Such changes may account for alleviation of NAFLD by CHF directly or indirectly. Further work will be addressed to validate and examine this interesting hypothesis. When analyzing bar-coded pyrosequencing data, PLS-DA was a robust method for identifying key phylotypic differences [46]. In this study, such differences were confirmed by DGGE fingerprints. Important genera, such as Prevotella spp., Bacteroides spp., E. coli, and Collinsella, were identified by both methods. However, due to different sensitivities of the methods, some low abundance bacteria were only identified by PLS-DA (e.g. OTU 259). Interestingly, phylotypes identified by PLS-DA, even though they belonged to the same taxonomic level, responded differently to HFD and CHF
X. Yin et al. / Systematic and Applied Microbiology 36 (2013) 188–196
treatment. This indicated that the response of gut microbiota to diet and drugs may be strain-dependent. The same phenomenon was also observed in key phylotypes in other studies [41,46], suggesting that strain- or gene-specific analysis is necessary for a more precise evaluation. Traditional Chinese medicine is frequently used as an alternative medicine and shows encouraging efficacy for some diseases [6,32]. However, it is still difficult to decipher its mode of action due to the complex ingredients and their interactions. Our study validates the effect of one Chinese herbal medicine in treating NAFLD, and further shows that this formula can modulate the gut microbiota structure of NAFLD rats by an increase in opportunistic pathogens and a decline in butyrate-producing bacteria, which may directly or indirectly affect the course of the disease. This work also showed that a microbiome-wide association study, correlating the molecular profiling data of gut microbiota with physiological/clinical parameters of the host, may provide a novel method for studying the mode of action of other CHFs. Acknowledgements This work was supported by Key Projects 30730005, 30800155 and 81001575 of the National Natural Science Foundation of China (NSFC), Project 2007DFC30450 of the International Science and Technology Cooperation Program in China, and the National Science and Technology Major Project of China 2009ZX10004-601. Financial support was also received from the Innovative Research Team of the Universities, the Shanghai Municipal Education Commission and the Key Laboratory of Liver and Kidney Diseases (Shanghai University of Traditional Chinese Medicine), Ministry of Education, Shanghai, China. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.syapm. 2012.12.009. References [1] Abu-Shanab, A., Quigley, E.M. (2010) The role of the gut microbiota in nonalcoholic fatty liver disease. Nat. Rev. Gastroenterol. Hepatol. 7, 691–701. [2] Amann, R.I., Ludwig, W., Schleifer, K.H. (1995) Phylogenetic identification and in situ detection of individual microbial cells without cultivation. Microbiol. Rev. 59, 143–169. [3] Baker, S.S., Baker, R.D., Liu, W., Nowak, N.J., Zhu, L. (2010) Role of alcohol metabolism in non-alcoholic steatohepatitis. PLoS ONE 5, e9570. [4] Bergheim, I., Weber, S., Vos, M., Kramer, S., Volynets, V., Kaserouni, S., McClain, C.J., Bischoff, S.C. (2008) Antibiotics protect against fructose-induced hepatic lipid accumulation in mice: role of endotoxin. J. Hepatol. 48, 983–992. [5] Cani, P.D., Amar, J., Iglesias, M.A., Poggi, M., Knauf, C., Bastelica, D., Neyrinck, A.M., Fava, F., Tuohy, K.M., Chabo, C., Waget, A., Delmee, E., Cousin, B., Sulpice, T., Chamontin, B., Ferrieres, J., Tanti, J.F., Gibson, G.R., Casteilla, L., Delzenne, N.M., Alessi, M.C., Burcelin, R. (2007) Metabolic endotoxemia initiates obesity and insulin resistance. Diabetes 56, 1761–1772. [6] Chan, K. (1995) Progress in traditional Chinese medicine. Trends Pharmacol. Sci. 16, 182–187. [7] Cole, J.R., Wang, Q., Cardenas, E., Fish, J., Chai, B., Farris, R.J., Kulam-SyedMohideen, A.S., McGarrell, D.M., Marsh, T., Garrity, G.M., Tiedje, J.M. (2009) The Ribosomal Database Project: improved alignments and new tools for rRNA analysis. Nucleic Acids Res. 37, D141–D145. [8] Crow, J.M. (2011) Microbiome: that healthy gut feeling. Nature 480, S88–S89. [9] Deng, Z.Y., Zhang, J.W., Li, J., Fan, Y.W., Cao, S.W., Huang, R.L., Yin, Y.L., Zhong, H.Y., Li, T.J. (2007) Effect of polysaccharides of cassiae seeds on the intestinal microflora of piglets. Asia Pac. J. Clin. Nutr. 16 (Suppl. 1), 143–147. [10] DeSantis, T.Z., Jr., Hugenholtz, P., Keller, K., Brodie, E.L., Larsen, N., Piceno, Y.M., Phan, R., Andersen, G.L. (2006) NAST: a multiple sequence alignment server for comparative analysis of 16S rRNA genes. Nucleic Acids Res. 34, W394–W399. [11] Dumas, M.E., Barton, R.H., Toye, A., Cloarec, O., Blancher, C., Rothwell, A., Fearnside, J., Tatoud, R., Blanc, V., Lindon, J.C., Mitchell, S.C., Holmes, E., McCarthy, M.I., Scott, J., Gauguier, D., Nicholson, J.K. (2006) Metabolic profiling reveals a contribution of gut microbiota to fatty liver phenotype in insulin-resistant mice. Proc. Natl. Acad. Sci. U.S.A. 103, 12511–12516.
195
[12] Gordon, J.I., Backhed, F., Ding, H., Wang, T., Hooper, L.V., Koh, G.Y., Nagy, A., Semenkovich, C.F. (2004) The gut microbiota as an environmental factor that regulates fat storage. Proc. Natl. Acad. Sci. U.S.A. 101, 15718–15723. [13] Haukeland, J.W., Damas, J.K., Konopski, Z., Loberg, E.M., Haaland, T., Goverud, I., Torjesen, P.A., Birkeland, K., Bjoro, K., Aukrust, P. (2006) Systemic inflammation in nonalcoholic fatty liver disease is characterized by elevated levels of CCl2. J. Hepatol. 44, 1167–1174. [14] Jia, W., Li, H.K., Zhao, L.P., Nicholson, J.K. (2008) Gut microbiota: a potential new territory for drug targeting. Nat. Rev. Drug Discov. 7, 123–129. [15] Kageyama, A., Benno, Y., Nakase, T. (1999) Phylogenetic and phenotypic evidence for the transfer of Eubacterium aerofaciens to the genus Collinsella as Collinsella aerofaciens gen. nov., comb. nov. Int. J. Syst. Bacteriol. 49, 557–565. [16] Kassinen, A., Krogius-Kurikka, L., Makivuokko, H., Rinttila, T., Paulin, L., Corander, J., Malinen, E., Apajalahti, J., Palva, A. (2007) The fecal microbiota of irritable bowel syndrome patients differs significantly from that of healthy subjects. Gastroenterology 133, 24–33. [17] Kato, M., Ishige, A., Anjiki, N., Yamamoto, M., Irie, Y., Taniyama, M., Kibe, R., Oka, J., Benno, Y., Watanabe, K. (2007) Effect of herbal medicine Juzentaihoto on hepatic and intestinal heat shock gene expression requires intestinal microflora in mouse. World J. Gastroenterol. 13, 2289–2297. [18] Knight, R., Lozupone, C., Hamady, M. (2006) UniFrac – an online tool for comparing microbial community diversity in a phylogenetic context. BMC Bioinform. 7, 371. [19] Kok, N., Roberfroid, M., Delzenne, N. (1996) Dietary oligofructose modifies the impact of fructose on hepatic triacylglycerol metabolism. Metabolism. 45, 1547–1550. [20] Li, H.S., Feng, Q., Hu, Y.Y., Chen, S.D., Li, X.M., Peng, J.H., Xu, L.L., Liu, C. (2009) Effect of Qushi Huayu Decoction on high-fat diet induced hepatic lipid deposition in rats. Chin. J. Integr. Tradit. West. Med. 29, 1092–1095. [21] Li, W., Godzik, A. (2006) Cd-hit: a fast program for clustering and comparing large sets of protein or nucleotide sequences. Bioinformatics 22, 1658–1659. [22] Li, Z., Yang, S., Lin, H., Huang, J., Watkins, P.A., Moser, A.B., Desimone, C., Song, X.Y., Diehl, A.M. (2003) Probiotics and antibodies to TNF inhibit inflammatory activity and improve nonalcoholic fatty liver disease. Hepatology 37, 343–350. [23] Loguercio, C., Federico, A., Tuccillo, C., Terracciano, F., D’Auria, M.V., De Simone, C., Del Vecchio Blanco, C. (2005) Beneficial effects of a probiotic VSL#3 on parameters of liver dysfunction in chronic liver diseases. J. Clin. Gastroenterol. 39, 540–543. [24] Ludwig, W., Strunk, O., Westram, R., Richter, L., Meier, H., Yadhukumar, Buchner, A., Lai, T., Steppi, S., Jobb, G., Forster, W., Brettske, I., Gerber, S., Ginhart, A.W., Gross, O., Grumann, S., Hermann, S., Jost, R., Konig, A., Liss, T., Lussmann, R., May, M., Nonhoff, B., Reichel, B., Strehlow, R., Stamatakis, A., Stuckmann, N., Vilbig, A., Lenke, M., Ludwig, T., Bode, A., Schleifer, K.H. (2004) ARB: a software environment for sequence data. Nucleic Acids Res. 32, 1363–1371. [25] Makivuokko, H., Tiihonen, K., Tynkkynen, S., Paulin, L., Rautonen, N. (2010) The effect of age and non-steroidal anti-inflammatory drugs on human intestinal microbiota composition. Br. J. Nutr. 103, 227–234. [26] Malinen, E., Krogius-Kurikka, L., Lyra, A., Nikkila, J., Jaaskelainen, A., Rinttila, T., Vilpponen-Salmela, T., von Wright, A.J., Palva, A. (2010) Association of symptoms with gastrointestinal microbiota in irritable bowel syndrome. World J. Gastroenterol. 16, 4532–4540. [27] Marchesini, G., Bugianesi, E., Forlani, G., Cerrelli, F., Lenzi, M., Manini, R., Natale, S., Vanni, E., Villanova, N., Melchionda, N., Rizzetto, M. (2003) Nonalcoholic fatty liver, steatohepatitis, and the metabolic syndrome. Hepatology 37, 917–923. [28] Marchesini, G., Moscatiello, S., Agostini, F., Villanova, N., Festi, D. (2011) Treatment of non-alcoholic fatty liver disease with focus on emerging drugs. Expert. Opin. Emerg. Drugs 16, 121–136. [29] Moore, W., Moore, L. (1995) Intestinal floras of populations that have a high risk of colon cancer. Appl. Environ. Microbiol. 61, 3202–3207. [30] Muyzer, G., de Waal, E.C., Uitterlinden, A.G. (1993) Profiling of complex microbial populations by denaturing gradient gel electrophoresis analysis of polymerase chain reaction-amplified genes coding for 16S rRNA. Appl. Environ. Microbiol. 59, 695–700. [31] Neal, M.D., Leaphart, C., Levy, R., Prince, J., Billiar, T.R., Watkins, S., Li, J., Cetin, S., Ford, H., Schreiber, A., Hackam, D.J. (2006) Enterocyte TLR4 mediates phagocytosis and translocation of bacteria across the intestinal barrier. J. Immunol. 176, 3070–3079. [32] Normile, D. (2003) The new face of traditional Chinese medicine. Science 299, 188–190. [33] Osten, D.W. (1988) Selection of optimal regression models via cross-validation. J. Chemom. 2, 39–48. [34] Qin, J., Li, Y., Cai, Z., Li, S., Zhu, J., Zhang, F., Liang, S., Zhang, W., Guan, Y., Shen, D. (2012) A metagenome-wide association study of gut microbiota in type 2 diabetes. Nature 490, 55–60. [35] Ruiz, A.G., Casafont, F., Crespo, J., Cayon, A., Mayorga, M., Estebanez, A., Fernadez-Escalante, J.C., Pons-Romero, F. (2007) Lipopolysaccharide-binding protein plasma levels and liver TNF-alpha gene expression in obese patients: evidence for the potential role of endotoxin in the pathogenesis of nonalcoholic steatohepatitis. Obes. Surg. 17, 1374–1380. [36] Sakata, T. (1987) Stimulatory effect of short-chain fatty acids on epithelial cell proliferation in the rat intestine: a possible explanation for trophic effects of fermentable fibre, gut microbes and luminal trophic factors. Br. J. Nutr. 58, 95–103. [37] Scheppach, W., Bartram, P., Richter, A., Richter, F., Liepold, H., Dusel, G., Hofstetter, G., Rüthlein, J., Kasper, H. (1992) Effect of short-chain fatty acids on the human colonic mucosa in vitro. J. Parenter. Enteral. Nutr. 16, 43–48.
196
X. Yin et al. / Systematic and Applied Microbiology 36 (2013) 188–196
[38] Schloss, P.D., Handelsman, J. (2005) Introducing DOTUR, a computer program for defining operational taxonomic units and estimating species richness. Appl. Environ. Microbiol. 71, 1501–1506. [39] Steffen, E.K., Berg, R.D., Deitch, E.A. (1988) Comparison of translocation rates of various indigenous bacteria from the gastrointestinal tract to the mesenteric lymph node. J. Infect. Dis. 157, 1032–1038. [40] Tannock, G.W., Munro, K., Bibiloni, R., Simon, M.A., Hargreaves, P., Gopal, P., Harmsen, H., Welling, G. (2004) Impact of consumption of oligosaccharidecontaining biscuits on the fecal microbiota of humans. Appl. Environ. Microbiol. 70, 2129–2136. [41] Wang, T., Cai, G., Qiu, Y., Fei, N., Zhang, M., Pang, X., Jia, W., Cai, S., Zhao, L. (2011) Structural segregation of gut microbiota between colorectal cancer patients and healthy volunteers. ISME J. 6, 320–329. [42] Wang, Y., Xu, N., Xi, A., Ahmed, Z., Zhang, B., Bai, X. (2009) Effects of Lactobacillus plantarum MA2 isolated from Tibet kefir on lipid metabolism and intestinal microflora of rats fed on high-cholesterol diet. Appl. Microbiol. Biotechnol. 84, 341–347. [43] Westad, F., Martens, H. (2000) Variable selection in near infrared spectroscopy based on significance testing in partial least squares regression. J. Near Infrared Spectrosc. 8, 117–124. [44] Wigg, A.J., Roberts-Thomson, I.C., Dymock, R.B., McCarthy, P.J., Grose, R.H., Cummins, A.G. (2001) The role of small intestinal bacterial overgrowth, intestinal permeability, endotoxaemia, and tumour necrosis factor alpha in the pathogenesis of non-alcoholic steatohepatitis. Gut 48, 206–211.
[45] Wright, S.D., Ramos, R.A., Tobias, P.S., Ulevitch, R.J., Mathison, J.C. (1990) CD14, a receptor for complexes of lipopolysaccharide (LPS) and LPS binding protein. Science 249, 1431–1433. [46] Zhang, C.H., Zhang, M.H., Wang, S.Y., Han, R.J., Cao, Y.F., Hua, W.Y., Mao, Y.J., Zhang, X.J., Pang, X.Y., Wei, C.C., Zhao, G.P., Chen, Y., Zhao, L.P. (2010) Interactions between gut microbiota, host genetics and diet relevant to development of metabolic syndromes in mice. ISME J. 4, 312–313. [47] Zhang, X., Zhao, Y., Zhang, M., Pang, X., Xu, J., Kang, C., Li, M., Zhang, C., Zhang, Z., Zhang, Y., Li, X., Ning, G., Zhao, L. (2012) Structural changes of gut microbiota during berberine-mediated prevention of obesity and insulin resistance in high-fat diet-fed rats. PLoS ONE 7, e42529. [48] Zhang, X., Zhao, Y.F., Hu, Y.Y., Liu, P., Zhao, L.P. (2011) Whole-body systems biology based on functional metagenomics – new opportunity for the modernization of Traditional Chinese Medicine. Modern. Tradit. Chin. Med. Mater. Med./World Sci. Technol. 13, 202–212. [49] Zimmermann, M. (1983) Ethical guidelines for investigations of experimental pain in conscious animals. Pain 16, 109. [50] Zoetendal, E.G., von Wright, A., Vilpponen-Salmela, T., Ben-Amor, K., Akkermans, A.D., de Vos, W.M. (2002) Mucosa-associated bacteria in the human gastrointestinal tract are uniformly distributed along the colon and differ from the community recovered from feces. Appl. Environ. Microbiol. 68, 3401–3407.
Contents lists available at SciVerse ScienceDirect
Systematic and Applied Microbiology journal homepage: www.elsevier.de/syapm
Structural changes of gut microbiota in a rat non-alcoholic fatty liver disease model treated with a Chinese herbal formula Xiaochen Yin a , Jinghua Peng b , Liping Zhao a,c , Yunpeng Yu a , Xu Zhang a , Ping Liu d , Qin Feng b , Yiyang Hu b,d,∗∗ , Xiaoyan Pang a,∗ a
State Key Laboratory of Microbial Metabolism, School of Life Sciences & Biotechnology, Shanghai Jiao Tong University, Shanghai 200240, China Institute of Liver Diseases, Shuguang Hospital, Shanghai University of Traditional Chinese Medicine, Shanghai 201203, China c Ministry of Education Key Laboratory of Systems Biomedicine, Shanghai Center for Systems Biomedicine, Shanghai 200240, China d E-Institute of TCM Internal Medicine, Shanghai Municipal Education Commission, Shanghai 201203, China b
a r t i c l e
i n f o
Article history: Received 15 August 2012 Received in revised form 6 December 2012 Accepted 6 December 2012 Keywords: Chinese herbal formula Gut microbiota Non-alcoholic fatty liver disease (NAFLD) Pyrosequencing Polymerase chain reaction-denaturing gradient gel electrophoresis (PCR-DGGE)
a b s t r a c t Accumulating evidence indicates that disruption of the gut microbiota by a high-fat diet (HFD) may play a pivotal role in the progression of metabolic disorders such as non-alcoholic fatty liver disease (NAFLD). In this study, the structural changes of gut microbiota were analyzed in an HFD-induced NAFLD rat model during treatment with an ancient Chinese herbal formula (CHF) used in clinical practice – Qushi Huayu Fang. CHF treatment significantly reduced body weight, alleviated hepatic steatosis, and decreased the content of triglycerides and free fatty acids in the livers of the rats. Gut microbiota of treated and control rats were profiled with polymerase chain reaction-denaturing gradient gel electrophoresis and bar-coded pyrosequencing of the V3 region of 16S rRNA genes. Both analyses indicated that the CHFtreated group harbored significantly different gut microbiota from that of model rats. Partial least squares discriminant analysis and taxonomy-based analysis were further employed to identify key phylotypes responding to HFD and CHF treatment. Most notably, the genera Escherichia/Shigella, containing opportunistic pathogens, were significantly enriched in HFD-fed rats compared to controls fed normal chow (P < 0.05) but they decreased to control levels after CHF treatment. Collinsella, a genus with short chain fatty acid producers, was significantly elevated in CHF-treated rats compared to HFD-fed rats (P < 0.05). The results revealed that the bacterial profiles of HFD-induced rats could be modulated by the CHF. Elucidation of these differences in microbiota composition provided a basis for further understanding the pharmacological mechanism of the CHF. © 2013 Elsevier GmbH. All rights reserved.
Introduction Non-alcoholic fatty liver disease (NAFLD), a manifestation of the metabolic syndrome in the liver, is closely associated with insulin resistance, obesity, and diabetes mellitus [27]. Although it is a major threat to public health, the pathogenesis of NAFLD is not yet well understood, nor are there effective therapies [28]. Emerging evidence in the last two decades indicates a possible role of gut microbiota in the etiology of NAFLD [1]. Small intestinal bacterial overgrowth is commonly found in NAFLD patients and results in increased intestinal permeability, endogenous ethanol production, and choline deficiency [3,11,44]. In
∗ Corresponding author at: State Key Laboratory of Microbial Metabolism, School of Life Sciences & Biotechnology, Shanghai Jiao Tong University, Shanghai 200240, China. Tel.: +86 21 34204878; fax: +86 21 34204878. ∗∗ Corresponding author. Tel.: +86 21 20256160; fax: +86 21 20256521. E-mail addresses: [email protected] (Y. Hu), [email protected] (X. Pang). 0723-2020/$ – see front matter © 2013 Elsevier GmbH. All rights reserved. http://dx.doi.org/10.1016/j.syapm.2012.12.009
addition, the gut microbiota promotes absorption of monosaccharides and short chain fatty acids by fermentation, and thus increases de novo hepatic lipogenesis and enhances fat storage by regulating lipoprotein lipase activity [12]. NAFLD is typically associated with low-grade, chronic inflammation [13], and a major cell wall component of Gram-negative bacteria, lipopolysaccharide (LPS), is known to be potent in promoting inflammation [5]. When binding to CD14 and toll-like receptor 4 on the surface of immune cells, a low concentration of LPS from gut microbiota can trigger a series of inflammatory processes [31,45], which in turn contributes to the pathogenesis of NAFLD. Consequently, the gut microbiota has become a novel target for potential therapies [14], with probiotics, prebiotics, and antibiotics being tested for treatment or prevention of NAFLD [4,19,22]. Traditional medicine has long been used in China, and most of the formulas are orally administered. Recent research has revealed that many ingredients in these herbs could only be absorbed and they exert their biological effects with the help of the host’s gut microbiota [8]. Furthermore, intriguing but limited data also
X. Yin et al. / Systematic and Applied Microbiology 36 (2013) 188–196
189
Table 1 Herbs used in the Chinese herbal formula. Chinese name
Pharmaceutical name
Family name
Production place
Processing method
Yin chen
Artemisia capillaries Thunb. (above-ground parts, dried) Polygonum cuspidatum Sieb. et Zucc (rhizome, root, dried) Curuma longa L. (rhizome, dried) Hypericum japonicum (whole plant, dried) Gardenia jasminoides Ellis (fruit, dried)
Compositae
Anhui Province, China
Ethanol extraction
Polygonaceae
Jiangsu Province, China
Ethanol extraction
Zingiberaceae Clusiaceae Rubiaceae
Sichuang Province, China Jiangxi Province, China Fujian Province, China
Ethanol extraction Mixed at an equal mass ratio in distilled water, then concentrated to a relative density of 1.08–1.12 (80 ◦ C), followed by ethanol precipitation
Hu zhang Jiang huang Tian ji huang Zhi zi
show that the efficacy of some traditional herbs may be related to modulation of gut microbiota [48]. One formula, Qushi Huayu Fang, has a long history of use in clinical practice to alleviate NAFLD. However, to date, no systematic and credible explanations have been proposed for its mechanism. Whether Qushi Huayu Fang can modulate the host’s gut ecosystem while alleviating NAFLD symptoms and whether gut bacteria are a potential target for the herbal formula are still unanswered questions. Nowadays, advanced technologies have facilitated the exploration of gut microbiota, especially with the development of high-throughput sequencing technology. Researchers are able to look in detail at the structure of gut microbiota, and study how external factors, such as age and drug use among others, influence gut microbiota. In this study, gut microbial composition during Qushi Huayu Fang treatment in a high-fat diet (HFD)-induced NAFLD rat model was monitored by polymerase chain reactiondenaturing gradient gel electrophoresis (PCR-DGGE) and bar-coded pyrosequencing. Using a microbiome-wide association strategy [34,41,47], the modulation of this Chinese herbal formula (CHF) in the gut microbiome was validated, and key phylotypes closely related to NAFLD development and CHF treatment were expected to be identified. Materials and methods Preparation of the CHF The compound was prepared from five herbs, as listed in Table 1. Herbs were obtained from qualified suppliers and on the basis of standards specified in the Chinese Pharmacopoeia (1995 edition) and Chinese Materia Medica in Shanghai. Extracts of the herbs (Table 1) were mixed at a mass ratio of 13:7:7:7:7 (13 parts Artemisia capillaries Thunb., and 7 parts of each of the other four herbs). High-performance liquid chromatography was used for quality control by monitoring three active components in the formula: polydatin (retention time: 10 min; concentration: 1.779 mg g−1 ), emodin (19 min; 5.394 mg g−1 ), and geniposide (35 min; 2.850 mg g−1 ). The final solution was stored at 4 ◦ C at two concentrations, 0.93 and 0.47 g mL−1 , as high and low dosages, respectively. The formula has been issued a patent by the State Intellectual Property Office of PR China under the ID: ZL200610009140.0. Animals and experimental protocol Ethics statement All procedures of the experiments in this study were approved by the Animal Ethics Committee of the Shanghai University of Traditional Chinese Medicine (Approval No. SCXK 2003-0003). The care and use of animals were carried out under the Guidelines for Animal Experiment of the Shanghai University of Traditional Chinese Medicine (Shanghai, China), and all efforts were made to minimize the number of animals and their suffering [49].
Twenty-six male Sprague–Dawley rats were purchased from the Shanghai Laboratory Animal Co. Ltd., China, and raised in a specific pathogen-free barrier system in the Laboratory Animal Center of the Shanghai Traditional Chinese Medicine University. After acclimatization, rats were divided into two groups: one group (n = 21) was fed an HFD (w/w, 87.5% normal chow, 10% lard, 2% cholesterol, and 0.5% sodium cholate), and a control group (n = 5) was fed normal chow (NC). After 6 weeks on the diet, HFD-fed rats were randomly divided into three subgroups (each n = 7). Two of the subgroups maintained the HFD diet in conjunction with high (H) and low (L) dosage (0.93 g/100 g body weight and 0.47 g/100 g body weight, respectively) CHF (Qushi Huayu Fang) through intragastric administration. The third subgroup remained on the HFD without additional treatment (HFD). Food intake in the groups was controlled to the same level. Control volumes (equivalent to those of the CHF compound given to experimental groups) containing water were administrated to NC- and HFD-fed rats. Body weight was measured every 14 days. At the end of week 10 (i.e. the CHF treatment lasted for 4 weeks), all animals from the four groups were sacrificed and their livers were removed and stored at −70 ◦ C for histological and lipid content analysis, including oil Red O and hematoxylin–eosin (H&E) staining, TG and FFA analysis (see Supplementary Information). Liver index was calculated as the ratio of liver-to-body weight. Before sacrifice, fecal samples were collected from 23 animals, including 6 from HFD, 5 NC, 5 CHF (H), and 7 CHF (L). Unfortunately, the feces collected from the other three animals were insufficient, so they were omitted from the gut microbiota analysis. All stool samples were stored at −70 ◦ C.
PCR-DGGE and multivariate analysis of the 16S rRNA gene V3 region DNA was extracted from each fecal sample by the bead-beating method [50]. Bacterial universal primers P2 (5 ATTACCGCGGCTGCTGG-3 ) and P3 (5 -CGCCCGCCGCGCGCGGCGGGCGGGGCGGGGGCACGGGGGGCCTACGGGAGGCAGCAG-3 ) were used to amplify the 16S rRNA gene V3 region from each DNA sample by hot-start touchdown PCR [30]. PCR amplification reactions, DGGE analysis, and phylogenetic identification of major bands were performed as described previously [46] and the DGGE denaturing gradient was 30–54% (7 M urea and 40% deionized formamide were considered to be the 100% denaturant). The sequences retrieved from major DGGE bands have been deposited in GenBank (http://www.ncbi.nlm.nih.gov/Genbank/) with accession numbers JF729214–JF729235. The gel images were converted into digital data using Quantity One 4.4.0 (Bio-Rad, Hercules, CA, USA). Principal component analysis (PCA) was employed to compare the gut microbiota composition between treatment groups in the MATLAB 7.11.0 (R2010b) environment (The MathWorks, Inc., Natick, MA, USA).
190
X. Yin et al. / Systematic and Applied Microbiology 36 (2013) 188–196
Bar-coded pyrosequencing and multivariate analysis of the 16S rRNA gene V3 region Forward 5 -NNNNNNNNATTACCGCGGCTGCT-3 and reverse primers were used to amplify the 16S rRNA gene V3 region. The NNNNNNNN was the sample-unique 8-base barcode used for sorting of PCR amplicons. PCR reactions, pyrosequencing, and sequence quality controls were performed as described previously with minor modification [46]. After cluster database at high identity with tolerance (CD-HIT) clustering (100% identity), sequences were aligned using nearest alignment space termination (NAST) [10,21]. Qualified V3 region sequences were imported to the ribosomal database project (RDP) classifier for taxonomical assignment at a 50% confidence level [7]. The sequences were also imported into the ARB software environment [24] in order to calculate the distance matrix based on the neighbor-joining algorithm for online UniFrac analysis [18]. A distance matrix was also used to define the operational taxonomic unit (OTU) in DOUTR [38]. Rarefaction analysis and the Shannon diversity index were calculated as described previously [46]. The unique sequences from pyrosequencing have been deposited in the GenBank Sequence Read Archive database (http://www.ncbi.nlm.nih.gov/Traces/sra) with accession number SRA054076. Partial least squares discriminant analysis (PLS-DA) was used to test whether these groups could be separated based on the OTU data. The correct classification rate of the PLS-DA model was performed with leave-one-out cross-validation [33]. Martens’ uncertainty test [43] followed by the Mann–Whitney test (P < 0.05) were used to select important OTUs for classification. All multivariate statistical analyses were performed in the MATLAB 7.11.0 (R2010b) environment (The MathWorks, Inc.).
vs. 601.22 ± 74.73 mol g−1 protein, CHF (H) vs. HFD; P < 0.05, 394.79 ± 14.28 mol g−1 vs. 601.22 ± 74.73 mol g−1 protein, CHF (L) vs. HFD; Fig. 1F).
5 -NNNNNNNNCCTACGGGAGGCAGCAG-3
Results Weight reduction and alleviation of NAFLD in rats treated with the CHF At the end of a 10-week intervention period, NC group rats weighed significantly less than the HFD group (P < 0.01; 484.20 ± 20.76 g vs. 540.29 ± 14.02 g, respectively; Fig. 1A). Rats in the CHF high dosage treatment (CHF (H)) and CHF low dosage treatment (CHF (L)) groups also weighed less than those in the HFD group (P < 0.01, 439.86 ± 7.22 g vs. 540.29 ± 14.02 g, CHF (H) vs. HFD; P < 0.01, 438.71 ± 13.94 g vs. 540.29 ± 14.02 g, CHF (L) vs. HFD) and even less than NC rats. Liver histological analysis showed extensive micro/macrovesicular steatosis in the hepatocytes of the HFD group (Fig. 1B and C). In this group, large lipid droplets filled the cytoplasm of hepatocytes, displacing the nucleus to the periphery. The two groups of CHF-treated rats had significantly less fat deposition in hepatocytes compared to their HFD counterparts. CHF high-dosage treatment exerted a better effect than low-dosage. The liver index showed that all HFD-fed animals, including the HFD group and the two CHF groups, had higher liver indices than the NC group, while the significantly reduced index was observed in the CHF (H) group compared to the HFD group (P < 0.05; 0.02 ± 6.03 E−4 vs. 0.04 ± 5.27 E−4, CHF (H) vs. HFD; Fig. 1D). For liver triglycerides (TG), a significant decrease was also observed in all CHF-treated rats compared to HFD rats (P < 0.01; 22.98 ± 1.30 mg g−1 vs. 40.12 ± 1.64 mg g−1 , CHF (H) vs. HFD; Fig. 1E), and the effect was greater in CHF (H)-treated than CHF (L)-treated rats. Liver free fatty acids (FFA) analysis had a similar trend to the TG. The FFA content of the two CHF groups was also significantly reduced (P < 0.01, 294.27 ± 20.47 mol g−1
Structural changes in gut microbiota revealed by PCR-DGGE fingerprints According to the 16S rDNA V3 region PCR-DGGE patterns, the overall composition after the CHF (H) treatment showed a significant difference from that in rats fed NC or HFD (Fig. 2A). The PCA of the fingerprints showed that rats fed NC, HFD and CHF (H) fell into three clusters, with the first principal component (PC1) accounting for 32.8% of the variation (Fig. 2B). In HFD-fed rats, the intensities of band 1 (b1), b2, b3 and b4 were significantly reduced. By sequence analysis, these bands were related to Prevotella copri strain CB7 (98% homology), Prevotella sp. DJF RP53 (98%), Prevotella copri strain CB7 (99%) and Bacteroides plebeius (96%), respectively (Fig. 2 and Supplementary Table S1). Rats in the HFD group showed different dominant bands, including b5, b6 and b7, which were closest to Coprococcus catus GD/7 (100%), Ruminococcus gauvreauii strain CCRI-16110 (100%), and Escherichia coli strain CAIM 1647 (100%), respectively. In CHF (H)-treated rats, there was a significant difference in the composition of gut microbiota compared to other groups. The intensities of b5, b6 and b7 were significantly reduced in some CHF (H)-treated rats. Notably, b9 (Collinsella aerofaciens strain KCTC5916, 97% homology) was promoted in the CHF (H) group, but b8 (Clostridium sordellii strain 2TT1, 100%), which was apparent in NC and HFD rats, was not detected in this group. DGGE analysis was also performed in order to test for a difference between the effects of CHF (H) and CHF (L) on gut microbiota. However, there were no significant differences in the fingerprints of the two dosages (Fig. 2C), as confirmed by PCA (Fig. 2D), since the two groups could not be divided into different clusters.
Structural changes of gut microbiota revealed by bar-coded pyrosequencing The high-throughput technology of bar-coded pyrosequencing was also introduced in order to monitor the structure shift of gut microbiota for all four groups of rats. In total, 41,375 high quality sequences were obtained with an average of 1500 per sample (Supplementary Fig. S1A). Using 97% identity as the cutoff, 1240 OTUs were delineated (Supplementary Fig. S1B). The rarefaction curves for all animals did not reach a plateau at the sequencing depth used (Supplementary Fig. S1C). However, the Shannon diversity indices all reached stable values, indicating that bacterial diversity in these communities was mostly covered (Supplementary Fig. S1D). Unweighted UniFrac analysis was used to discriminate the microbiota composition of the different groups based on the evolutionary distance. All samples fell into three clusters with CHF (H) and CHF (L) in one cluster (Fig. 3A and B). CHF (H) and CHF (L) were then combined into a single CHF treatment group in the subsequent analysis. Taxonomy-based analysis at the phylum level showed that there were much less Bacteroidetes in HFD- vs. NC-fed rats (P < 0.05; median 7.09% vs. 38.80%, respectively; Fig. 4A) and more Firmicutes (P < 0.05; 83.83% vs. 56.19%; Fig. 4B), whereas Proteobacteria did not change significantly (P = 0.11; 4.10% vs. 6.02%, NC vs. HFD; Fig. 4C). After CHF treatment, the relative abundance of these three phyla did not change significantly but Actinobacteria increased significantly (P < 0.05; 3.86% vs. 1.25%, CHF vs. HFD), and was even higher than in NC-fed rats (P < 0.05; 3.86% vs. 0.47%, CHF vs. NC; Fig. 4D).
X. Yin et al. / Systematic and Applied Microbiology 36 (2013) 188–196
191
Fig. 1. Alleviation of non-alcoholic fatty liver disease by the Chinese herbal formula (CHF). (A) Growth curves based on average body weight of the four groups measured every 14 days. (B) Oil Red O staining of the rat hepatocytes (magnification, 200×). (C) Hematoxylin–eosin (H&E) staining of hepatocytes (magnification, 400×). (D) Liver index (the ratio of liver to body weight) at the end of week 10. (E) Liver triglycerides at the end of week 10. (F) Liver free fatty acids at the end of week 10. Data are expressed as mean values ± SE. **P < 0.01 and *P < 0.05 vs. HFD; # P < 0.05 vs. CHF (H) by one-way analysis of variance (ANOVA) followed by least-significant difference post hoc multiple comparison. Note: Part of the phenotypic data from an animal in this study has been published previously [20]. NC: normal chow-fed group; HFD: high-fat diet-fed group; CHF (H): Chinese herbal formula high dosage treatment group; CHF (L): Chinese herbal formula low dosage treatment group.
Taxonomy-based comparison at the genus level further showed that, in healthy NC-fed rats, Prevotella and Bacteroides from phylum Bacteroidetes, and the seven genera of Lactobacillus, Oscillibacter, Blautia, Coprococcus, Faecalibacterium, Sporacetigenium, and Allobaculum from phylum Firmicutes were predominant (each genus had a relative abundance >1%). While in HFD-fed rats, Prevotella, Bacteroides, and Allobaculum significantly decreased, whereas Escherichia/Shigella, Coprococcus, Blautia, and Roseburia increased. CHF treatment significantly reduced Escherichia/Shigella (P < 0.01; median 0.06% vs. 1.45%, CHF vs. HFD) and also significantly diminished Sporacetigenium (P < 0.05; 0.39% vs. 2.42%, CHF vs. HFD), while Collinsella was significantly enhanced (P < 0.05; 3.15% vs. 0.55%, CHF vs. HFD) after CHF treatment (Fig. 5 and Supplementary Table S2).
Key OTUs responding to HFD and CHF treatment identified by partial least squares discriminant analysis (PLS-DA) PLS-DA models were established between the NC vs. HFD groups and HFD vs. CHF groups. Score plots based on the first two components showed that different groups were well separated (Supplementary Fig. S2A and S2B). In leave-one-out cross-validation, the PLS-DA model between NC vs. HFD groups with three components yielded the high correct classification rate of nearly 100%. The Martens’ uncertainty and Mann–Whitney tests (P < 0.05) further identified 41 OTUs as key variables for classification in this model. The PLS-DA model between HFD vs. CHF with five components yielded a correct classification rate of 78%. Seven phylotypes were also identified using the same protocol. In all, 45 OTUs responding to HFD and CHF treatments were identified (Fig. 6 and Supplementary Table S3). In HFD-fed rats, OTUs belonging to Porphyromonadaceae (5 OTUs), Prevotella (4 OTUs), Bacteroides (2 OTUs), Parabacteroides (1 OTU), Bacteroidales (1 OTU), Allobaculum (1 OTU), Veillonellaceae (1 OTU), Parasutterella
(1 OTU), and Desulfovibrio (1 OTU) were significantly reduced, whereas Escherichia/Shigella (1 OTU), Erysipelotrichaceae (1 OTU) and Peptococcaceae (1 OTU) were significantly increased. Although one OTU belonging to Collinsella was significantly increased in the HFD group compared to NC, taxonomy-based analysis showed that there was no significant difference for Collinsella (including 204 unique sequences) between these two groups. CHF treatment significantly reduced the abundance of Barnesiella (1 OTU), Peptococcaceae (1 OTU), Erysipelotrichaceae (1 OTU), Clostridiales (1 OTU), Lactobacillus (1 OTU), Coprococcus (1 OTU) and Ruminococcaceae (2 OTUs). Notably, OTU 19 belonging to Escherichia/Shigella was also significantly reduced after CHF treatment (P < 0.05; median 0.06% vs. 1.39%, CHF vs. HFD). Another OTU (OTU 11), belonging to Lactobacillus, was significantly increased after CHF treatment. However, OTUs belonging to the same family/genus, such as in Ruminococcaceae (9 OTUs), Lactobacillus (4 OTUs), Barnesiella (4 OTUs), Lachnospiraceae (3 OTUs), Clostridiales (2 OTUs) and Coprococcus (2 OTUs), responded differently to HFD and CHF treatment (Supplementary Table S3). For example, of the nine OTUs belonging to Ruminococcaceae, four decreased in the HFD group but increased in the CHF group. Another two OTUs in Ruminococcaceae increased in the HFD group but decreased in the CHF group, and the remaining three OTUs were reduced in both the HFD and CHF groups vs. the NC group.
Discussion Due to the close anatomical relationship and metabolic interactions between the liver and intestinal tract, imbalances in gut microbiota may influence diverse physiological processes [1,11]. Li et al. [22] reported that treatment with a probiotic (VSL#3) or anti-TNF-␣ antibodies alleviated NAFLD in ob/ob mice, leading to an improvement in liver steatosis, as well as reduced levels
192
X. Yin et al. / Systematic and Applied Microbiology 36 (2013) 188–196
Fig. 2. Comparison of gut microbiota by denaturing gradient gel electrophoresis (DGGE) fingerprinting. (A) DGGE fingerprints of the V3 region of 16S rRNA genes from fecal bacterial communities of NC, HFD and CHF (H). (B) Principal component analysis (PCA) of DGGE fingerprints in (A). (C) DGGE profiles of HFD, CHF (H), and CHF (L) groups. (D) PCA analysis of DGGE fingerprints in (C). NC: normal chow-fed group; HFD: high-fat diet-fed group; CHF (H): Chinese herbal formula high dosage treatment group; CHF (L): Chinese herbal formula low dosage treatment group. PC: principal component. Numbers at the top of each lane correspond to the animal IDs in the experiment.
Fig. 3. The overall gut microbiota structural change revealed by 454 pyrosequencing. (A) UniFrac analysis (unweighted) based on the pyrosequencing operational taxonomic units (OTUs) (97% similarity level) data. Each point represents the mean of PC scores in a group at the end of week 10, and error bars represent the standard error. (B) Clustering of gut microbiota based on distances between the groups, calculated by multivariate analysis of variance tests of the first 10 PCs of UniFrac (unweighted) distance. **P < 0.01. NC: normal chow-fed group; HFD: high-fat diet-fed group; CHF (H): Chinese herbal formula high dosage treatment group; CHF (L): Chinese herbal formula low dosage treatment group. PC: principal coordinate.
X. Yin et al. / Systematic and Applied Microbiology 36 (2013) 188–196
193
Fig. 4. Taxonomy-based analysis at the phylum level. In each of the panels, the phylum analyzed is indicated by the label on the y-axis. *P < 0.05 by the Mann–Whitney test. The boundary of the box closest to zero indicates the 25th percentile, a line within the box marks the median, and the boundary of the box farthest from zero indicates the 75th percentile. Whiskers (error bars) above and below the box indicate the 90th and 10th percentiles. NC: normal chow-fed group; HFD: high-fat diet-fed group; CHF: Chinese herbal formula treatment group.
of hepatic total fatty acid and serum alanine aminotransferase. Further clinical trials showed VSL#3 exerted beneficial functional effects on chronic liver diseases in humans [23]. Oligofructose, a widely studied prebiotic, has been reported to protect rats against liver TG accumulation induced by fructose but was not able to prevent fructose-induced hypertriglyceridemia [19]. Antibiotic treatment of fructose-fed mice has been shown to markedly reduce hepatic lipid accumulation, liver index, and portal endotoxin levels [4]. However, the mechanisms underlying the beneficial effects of probiotics and prebiotics in NAFLD are not well characterized. Wang et al. [42] used a culture-based technology to analyze changes of gut microbiota in rats fed a cholesterol-enriched diet supplemented with Lactobacillus plantarum MA2 for 5 weeks. This
probiotic strain was found to significantly lower total hepatic cholesterol and TG, as well as serum total cholesterol, low-density lipoprotein cholesterol and TG. There were increases in lactic acid bacteria and bifidobacteria, while E. coli was unchanged. However, the use of enrichment media to culture these three groups of bacteria has limitations, because most of the gut microbes are unknown and resistant to traditional culture methods [2]. In our study, comparable beneficial effects of one CHF (Qushi Huayu Fang) against NAFLD were observed. Liver histology in CHFtreated rats showed significant improvement with less hepatocyte steatosis and inflammation. Hepatic TG and FFA contents were also significantly reduced after CHF treatment. DGGE fingerprints and high-throughput bar-coded pyrosequencing based on the 16S rRNA
Fig. 5. Taxonomy-based analysis at the genus level. In each of the panels, the genus analyzed is indicated by the label on the y-axis. **P < 0.01 and *P < 0.05 by the Mann–Whitney test. The boundary of the box closest to zero indicates the 25th percentile, a line within the box marks the median, and the boundary of the box farthest from zero indicates the 75th percentile. Whiskers (error bars) above and below the box indicate the 90th and 10th percentiles, respectively. NC: normal chow-fed group; HFD: high-fat diet-fed group; CHF: Chinese herbal formula treatment group.
194
X. Yin et al. / Systematic and Applied Microbiology 36 (2013) 188–196
Fig. 6. Abundance distribution of the 45 phylotypes identified as key classification variables. The OTUs were based on the pyrosequencing (97% similarity level) data. Partial least squares discriminant analysis (PLS-DA) was used followed by Martens’ uncertainty test and the Mann–Whitney test (P < 0.05) in order to identify the key OTUs. To show the distribution of the OTUs with lower abundance, the colored squares of each column have been scaled to indicate the relative ratios of the OTU between the animals. The phylogenetic relationships of these OTUs are displayed to the left of the heatmap, with animals arranged from left to right according to the increase in liver triglyceride content. NC: normal chow-fed group; HFD: high-fat diet-fed group; CHF (H): Chinese herbal formula high dosage treatment group; CHF (L): Chinese herbal formula low dosage treatment group. The numbers along the x-axis are animal IDs in each of the groups.
gene V3 region were used to detect wider overall changes in microbiota. Previous studies have shown that some CHFs are capable of modulating gut microbiota. Kato et al. [17] observed that Juzentaihoto could decrease unculturable bacteria and increase Lactobacillus johnsoni in mice, which in turn modulated heat shock protein gene expression. The main components of some herbs, such as the polysaccharides of cassiae seeds, have been shown to promote growth of Lactobacillus and Bifidobacterium while inhibiting some opportunistic pathogens in piglets [9]. However, none of these studies provide a complete profiling of the gut microbiota. Through DGGE and UniFrac analysis based on pyrosequencing data, we were able to observe a significant structural difference between gut microbiota in CHF-treated rats and their NAFLD counterparts. Multivariable analysis, such as PLS-DA, enabled some key phylotypes to be pinpointed that may be relevant to the development of NAFLD and the action of the CHF studied. In the present study, Escherichia/Shigella were most abundant in HFD-fed rats and significantly reduced in CHF-treated rats. This is a group of Gram-negative bacteria that contain LPS in their cell walls. It is known that a large number of Gram-negative bacteria in the gut may impair the gut barrier, thus releasing LPS into the blood, and triggering low-grade chronic inflammation [5]. Escherichia/Shigella are reported to be particularly efficient at translocation [39] and are closely associated with small intestinal bacterial overgrowth. They may also contribute to the high concentration of serum LPS and TNF-␣ expression usually found in NAFLD patients [35]. In this study, reduced levels of Escherichia/Shigella after CHF treatment may suggest a possible reduced antigen load and improved inflammatory status of the host. However, the causal relationship between the microbiome and the disease still needs further confirmation. Among the key variables identified, Collinsella increased more than 6-fold in CHF-treated animals, which correlated with significantly alleviated NAFLD compared to that in HFD-fed rats. However, there was no difference in Collinsella between NC- and HFD-fed rats. Promotion of Collinsella specifically by CHF treatment may contribute to the different position of CHF group in UniFrac plot from
NC and HFD group. Bacteria in this genus have been reported to ferment carbohydrates to hydrogen, ethanol, formate, and lactate [15]. Tannock et al. [40] reported that consumption of oligosaccharidecontaining biscuits elevated the number of C. aerofaciens in human feces in 11 out of 15 subjects. Thus, the ability of Collinsella to produce short-chain fatty acids (SCFA) may contribute to the efficacy of this formula, since SCFA have been reported to stimulate epithelial cell proliferation and play an important role in improving gut barrier integrity [36,37]. High levels of C. aerofaciens were also closely associated with a lower risk for colon cancer in a human cohort including 18 polyp patients, 15 Japanese-Hawaiians, 17 North American Caucasians, 22 rural native Japanese and 16 rural Africans [29]. Less Collinsella has been detected in patients with irritable bowel syndrome compared to healthy controls [16]. A negative correlation between Collinsella and body mass index was further observed in irritable bowel syndrome patients [26]. Lower levels of Collinsella have also been observed in elderly people taking non-steroidal anti-inflammatory drugs vs. young people, or in elderly people taking no drugs [25], which indicates an underlying relationship between Collinsella and inflammation. The significant increase of Collinsella in our CHF-treated rats may therefore suggest some other unknown protective mechanism that warrants further investigation. The major effects on the balance of gut microbiota by CHF treatment could be reduction of certain opportunistic pathogens, such as Escherichia/Shigella, accompanied by an increase of potentially beneficial bacteria like Collinsella. Such changes may account for alleviation of NAFLD by CHF directly or indirectly. Further work will be addressed to validate and examine this interesting hypothesis. When analyzing bar-coded pyrosequencing data, PLS-DA was a robust method for identifying key phylotypic differences [46]. In this study, such differences were confirmed by DGGE fingerprints. Important genera, such as Prevotella spp., Bacteroides spp., E. coli, and Collinsella, were identified by both methods. However, due to different sensitivities of the methods, some low abundance bacteria were only identified by PLS-DA (e.g. OTU 259). Interestingly, phylotypes identified by PLS-DA, even though they belonged to the same taxonomic level, responded differently to HFD and CHF
X. Yin et al. / Systematic and Applied Microbiology 36 (2013) 188–196
treatment. This indicated that the response of gut microbiota to diet and drugs may be strain-dependent. The same phenomenon was also observed in key phylotypes in other studies [41,46], suggesting that strain- or gene-specific analysis is necessary for a more precise evaluation. Traditional Chinese medicine is frequently used as an alternative medicine and shows encouraging efficacy for some diseases [6,32]. However, it is still difficult to decipher its mode of action due to the complex ingredients and their interactions. Our study validates the effect of one Chinese herbal medicine in treating NAFLD, and further shows that this formula can modulate the gut microbiota structure of NAFLD rats by an increase in opportunistic pathogens and a decline in butyrate-producing bacteria, which may directly or indirectly affect the course of the disease. This work also showed that a microbiome-wide association study, correlating the molecular profiling data of gut microbiota with physiological/clinical parameters of the host, may provide a novel method for studying the mode of action of other CHFs. Acknowledgements This work was supported by Key Projects 30730005, 30800155 and 81001575 of the National Natural Science Foundation of China (NSFC), Project 2007DFC30450 of the International Science and Technology Cooperation Program in China, and the National Science and Technology Major Project of China 2009ZX10004-601. Financial support was also received from the Innovative Research Team of the Universities, the Shanghai Municipal Education Commission and the Key Laboratory of Liver and Kidney Diseases (Shanghai University of Traditional Chinese Medicine), Ministry of Education, Shanghai, China. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.syapm. 2012.12.009. References [1] Abu-Shanab, A., Quigley, E.M. (2010) The role of the gut microbiota in nonalcoholic fatty liver disease. Nat. Rev. Gastroenterol. Hepatol. 7, 691–701. [2] Amann, R.I., Ludwig, W., Schleifer, K.H. (1995) Phylogenetic identification and in situ detection of individual microbial cells without cultivation. Microbiol. Rev. 59, 143–169. [3] Baker, S.S., Baker, R.D., Liu, W., Nowak, N.J., Zhu, L. (2010) Role of alcohol metabolism in non-alcoholic steatohepatitis. PLoS ONE 5, e9570. [4] Bergheim, I., Weber, S., Vos, M., Kramer, S., Volynets, V., Kaserouni, S., McClain, C.J., Bischoff, S.C. (2008) Antibiotics protect against fructose-induced hepatic lipid accumulation in mice: role of endotoxin. J. Hepatol. 48, 983–992. [5] Cani, P.D., Amar, J., Iglesias, M.A., Poggi, M., Knauf, C., Bastelica, D., Neyrinck, A.M., Fava, F., Tuohy, K.M., Chabo, C., Waget, A., Delmee, E., Cousin, B., Sulpice, T., Chamontin, B., Ferrieres, J., Tanti, J.F., Gibson, G.R., Casteilla, L., Delzenne, N.M., Alessi, M.C., Burcelin, R. (2007) Metabolic endotoxemia initiates obesity and insulin resistance. Diabetes 56, 1761–1772. [6] Chan, K. (1995) Progress in traditional Chinese medicine. Trends Pharmacol. Sci. 16, 182–187. [7] Cole, J.R., Wang, Q., Cardenas, E., Fish, J., Chai, B., Farris, R.J., Kulam-SyedMohideen, A.S., McGarrell, D.M., Marsh, T., Garrity, G.M., Tiedje, J.M. (2009) The Ribosomal Database Project: improved alignments and new tools for rRNA analysis. Nucleic Acids Res. 37, D141–D145. [8] Crow, J.M. (2011) Microbiome: that healthy gut feeling. Nature 480, S88–S89. [9] Deng, Z.Y., Zhang, J.W., Li, J., Fan, Y.W., Cao, S.W., Huang, R.L., Yin, Y.L., Zhong, H.Y., Li, T.J. (2007) Effect of polysaccharides of cassiae seeds on the intestinal microflora of piglets. Asia Pac. J. Clin. Nutr. 16 (Suppl. 1), 143–147. [10] DeSantis, T.Z., Jr., Hugenholtz, P., Keller, K., Brodie, E.L., Larsen, N., Piceno, Y.M., Phan, R., Andersen, G.L. (2006) NAST: a multiple sequence alignment server for comparative analysis of 16S rRNA genes. Nucleic Acids Res. 34, W394–W399. [11] Dumas, M.E., Barton, R.H., Toye, A., Cloarec, O., Blancher, C., Rothwell, A., Fearnside, J., Tatoud, R., Blanc, V., Lindon, J.C., Mitchell, S.C., Holmes, E., McCarthy, M.I., Scott, J., Gauguier, D., Nicholson, J.K. (2006) Metabolic profiling reveals a contribution of gut microbiota to fatty liver phenotype in insulin-resistant mice. Proc. Natl. Acad. Sci. U.S.A. 103, 12511–12516.
195
[12] Gordon, J.I., Backhed, F., Ding, H., Wang, T., Hooper, L.V., Koh, G.Y., Nagy, A., Semenkovich, C.F. (2004) The gut microbiota as an environmental factor that regulates fat storage. Proc. Natl. Acad. Sci. U.S.A. 101, 15718–15723. [13] Haukeland, J.W., Damas, J.K., Konopski, Z., Loberg, E.M., Haaland, T., Goverud, I., Torjesen, P.A., Birkeland, K., Bjoro, K., Aukrust, P. (2006) Systemic inflammation in nonalcoholic fatty liver disease is characterized by elevated levels of CCl2. J. Hepatol. 44, 1167–1174. [14] Jia, W., Li, H.K., Zhao, L.P., Nicholson, J.K. (2008) Gut microbiota: a potential new territory for drug targeting. Nat. Rev. Drug Discov. 7, 123–129. [15] Kageyama, A., Benno, Y., Nakase, T. (1999) Phylogenetic and phenotypic evidence for the transfer of Eubacterium aerofaciens to the genus Collinsella as Collinsella aerofaciens gen. nov., comb. nov. Int. J. Syst. Bacteriol. 49, 557–565. [16] Kassinen, A., Krogius-Kurikka, L., Makivuokko, H., Rinttila, T., Paulin, L., Corander, J., Malinen, E., Apajalahti, J., Palva, A. (2007) The fecal microbiota of irritable bowel syndrome patients differs significantly from that of healthy subjects. Gastroenterology 133, 24–33. [17] Kato, M., Ishige, A., Anjiki, N., Yamamoto, M., Irie, Y., Taniyama, M., Kibe, R., Oka, J., Benno, Y., Watanabe, K. (2007) Effect of herbal medicine Juzentaihoto on hepatic and intestinal heat shock gene expression requires intestinal microflora in mouse. World J. Gastroenterol. 13, 2289–2297. [18] Knight, R., Lozupone, C., Hamady, M. (2006) UniFrac – an online tool for comparing microbial community diversity in a phylogenetic context. BMC Bioinform. 7, 371. [19] Kok, N., Roberfroid, M., Delzenne, N. (1996) Dietary oligofructose modifies the impact of fructose on hepatic triacylglycerol metabolism. Metabolism. 45, 1547–1550. [20] Li, H.S., Feng, Q., Hu, Y.Y., Chen, S.D., Li, X.M., Peng, J.H., Xu, L.L., Liu, C. (2009) Effect of Qushi Huayu Decoction on high-fat diet induced hepatic lipid deposition in rats. Chin. J. Integr. Tradit. West. Med. 29, 1092–1095. [21] Li, W., Godzik, A. (2006) Cd-hit: a fast program for clustering and comparing large sets of protein or nucleotide sequences. Bioinformatics 22, 1658–1659. [22] Li, Z., Yang, S., Lin, H., Huang, J., Watkins, P.A., Moser, A.B., Desimone, C., Song, X.Y., Diehl, A.M. (2003) Probiotics and antibodies to TNF inhibit inflammatory activity and improve nonalcoholic fatty liver disease. Hepatology 37, 343–350. [23] Loguercio, C., Federico, A., Tuccillo, C., Terracciano, F., D’Auria, M.V., De Simone, C., Del Vecchio Blanco, C. (2005) Beneficial effects of a probiotic VSL#3 on parameters of liver dysfunction in chronic liver diseases. J. Clin. Gastroenterol. 39, 540–543. [24] Ludwig, W., Strunk, O., Westram, R., Richter, L., Meier, H., Yadhukumar, Buchner, A., Lai, T., Steppi, S., Jobb, G., Forster, W., Brettske, I., Gerber, S., Ginhart, A.W., Gross, O., Grumann, S., Hermann, S., Jost, R., Konig, A., Liss, T., Lussmann, R., May, M., Nonhoff, B., Reichel, B., Strehlow, R., Stamatakis, A., Stuckmann, N., Vilbig, A., Lenke, M., Ludwig, T., Bode, A., Schleifer, K.H. (2004) ARB: a software environment for sequence data. Nucleic Acids Res. 32, 1363–1371. [25] Makivuokko, H., Tiihonen, K., Tynkkynen, S., Paulin, L., Rautonen, N. (2010) The effect of age and non-steroidal anti-inflammatory drugs on human intestinal microbiota composition. Br. J. Nutr. 103, 227–234. [26] Malinen, E., Krogius-Kurikka, L., Lyra, A., Nikkila, J., Jaaskelainen, A., Rinttila, T., Vilpponen-Salmela, T., von Wright, A.J., Palva, A. (2010) Association of symptoms with gastrointestinal microbiota in irritable bowel syndrome. World J. Gastroenterol. 16, 4532–4540. [27] Marchesini, G., Bugianesi, E., Forlani, G., Cerrelli, F., Lenzi, M., Manini, R., Natale, S., Vanni, E., Villanova, N., Melchionda, N., Rizzetto, M. (2003) Nonalcoholic fatty liver, steatohepatitis, and the metabolic syndrome. Hepatology 37, 917–923. [28] Marchesini, G., Moscatiello, S., Agostini, F., Villanova, N., Festi, D. (2011) Treatment of non-alcoholic fatty liver disease with focus on emerging drugs. Expert. Opin. Emerg. Drugs 16, 121–136. [29] Moore, W., Moore, L. (1995) Intestinal floras of populations that have a high risk of colon cancer. Appl. Environ. Microbiol. 61, 3202–3207. [30] Muyzer, G., de Waal, E.C., Uitterlinden, A.G. (1993) Profiling of complex microbial populations by denaturing gradient gel electrophoresis analysis of polymerase chain reaction-amplified genes coding for 16S rRNA. Appl. Environ. Microbiol. 59, 695–700. [31] Neal, M.D., Leaphart, C., Levy, R., Prince, J., Billiar, T.R., Watkins, S., Li, J., Cetin, S., Ford, H., Schreiber, A., Hackam, D.J. (2006) Enterocyte TLR4 mediates phagocytosis and translocation of bacteria across the intestinal barrier. J. Immunol. 176, 3070–3079. [32] Normile, D. (2003) The new face of traditional Chinese medicine. Science 299, 188–190. [33] Osten, D.W. (1988) Selection of optimal regression models via cross-validation. J. Chemom. 2, 39–48. [34] Qin, J., Li, Y., Cai, Z., Li, S., Zhu, J., Zhang, F., Liang, S., Zhang, W., Guan, Y., Shen, D. (2012) A metagenome-wide association study of gut microbiota in type 2 diabetes. Nature 490, 55–60. [35] Ruiz, A.G., Casafont, F., Crespo, J., Cayon, A., Mayorga, M., Estebanez, A., Fernadez-Escalante, J.C., Pons-Romero, F. (2007) Lipopolysaccharide-binding protein plasma levels and liver TNF-alpha gene expression in obese patients: evidence for the potential role of endotoxin in the pathogenesis of nonalcoholic steatohepatitis. Obes. Surg. 17, 1374–1380. [36] Sakata, T. (1987) Stimulatory effect of short-chain fatty acids on epithelial cell proliferation in the rat intestine: a possible explanation for trophic effects of fermentable fibre, gut microbes and luminal trophic factors. Br. J. Nutr. 58, 95–103. [37] Scheppach, W., Bartram, P., Richter, A., Richter, F., Liepold, H., Dusel, G., Hofstetter, G., Rüthlein, J., Kasper, H. (1992) Effect of short-chain fatty acids on the human colonic mucosa in vitro. J. Parenter. Enteral. Nutr. 16, 43–48.
196
X. Yin et al. / Systematic and Applied Microbiology 36 (2013) 188–196
[38] Schloss, P.D., Handelsman, J. (2005) Introducing DOTUR, a computer program for defining operational taxonomic units and estimating species richness. Appl. Environ. Microbiol. 71, 1501–1506. [39] Steffen, E.K., Berg, R.D., Deitch, E.A. (1988) Comparison of translocation rates of various indigenous bacteria from the gastrointestinal tract to the mesenteric lymph node. J. Infect. Dis. 157, 1032–1038. [40] Tannock, G.W., Munro, K., Bibiloni, R., Simon, M.A., Hargreaves, P., Gopal, P., Harmsen, H., Welling, G. (2004) Impact of consumption of oligosaccharidecontaining biscuits on the fecal microbiota of humans. Appl. Environ. Microbiol. 70, 2129–2136. [41] Wang, T., Cai, G., Qiu, Y., Fei, N., Zhang, M., Pang, X., Jia, W., Cai, S., Zhao, L. (2011) Structural segregation of gut microbiota between colorectal cancer patients and healthy volunteers. ISME J. 6, 320–329. [42] Wang, Y., Xu, N., Xi, A., Ahmed, Z., Zhang, B., Bai, X. (2009) Effects of Lactobacillus plantarum MA2 isolated from Tibet kefir on lipid metabolism and intestinal microflora of rats fed on high-cholesterol diet. Appl. Microbiol. Biotechnol. 84, 341–347. [43] Westad, F., Martens, H. (2000) Variable selection in near infrared spectroscopy based on significance testing in partial least squares regression. J. Near Infrared Spectrosc. 8, 117–124. [44] Wigg, A.J., Roberts-Thomson, I.C., Dymock, R.B., McCarthy, P.J., Grose, R.H., Cummins, A.G. (2001) The role of small intestinal bacterial overgrowth, intestinal permeability, endotoxaemia, and tumour necrosis factor alpha in the pathogenesis of non-alcoholic steatohepatitis. Gut 48, 206–211.
[45] Wright, S.D., Ramos, R.A., Tobias, P.S., Ulevitch, R.J., Mathison, J.C. (1990) CD14, a receptor for complexes of lipopolysaccharide (LPS) and LPS binding protein. Science 249, 1431–1433. [46] Zhang, C.H., Zhang, M.H., Wang, S.Y., Han, R.J., Cao, Y.F., Hua, W.Y., Mao, Y.J., Zhang, X.J., Pang, X.Y., Wei, C.C., Zhao, G.P., Chen, Y., Zhao, L.P. (2010) Interactions between gut microbiota, host genetics and diet relevant to development of metabolic syndromes in mice. ISME J. 4, 312–313. [47] Zhang, X., Zhao, Y., Zhang, M., Pang, X., Xu, J., Kang, C., Li, M., Zhang, C., Zhang, Z., Zhang, Y., Li, X., Ning, G., Zhao, L. (2012) Structural changes of gut microbiota during berberine-mediated prevention of obesity and insulin resistance in high-fat diet-fed rats. PLoS ONE 7, e42529. [48] Zhang, X., Zhao, Y.F., Hu, Y.Y., Liu, P., Zhao, L.P. (2011) Whole-body systems biology based on functional metagenomics – new opportunity for the modernization of Traditional Chinese Medicine. Modern. Tradit. Chin. Med. Mater. Med./World Sci. Technol. 13, 202–212. [49] Zimmermann, M. (1983) Ethical guidelines for investigations of experimental pain in conscious animals. Pain 16, 109. [50] Zoetendal, E.G., von Wright, A., Vilpponen-Salmela, T., Ben-Amor, K., Akkermans, A.D., de Vos, W.M. (2002) Mucosa-associated bacteria in the human gastrointestinal tract are uniformly distributed along the colon and differ from the community recovered from feces. Appl. Environ. Microbiol. 68, 3401–3407.