- Email: [email protected]
Insight into alteration of gut microbiota in Clostridium difficile infection and asymptomatic C. difficile colonization
Anaerobe 34 (2015) 1e7
Contents lists available at ScienceDirect
Anaerobe journal homepage: www.elsevier.com/locate/anaerobe
Molecular Biology, Genetics and Biotechnology
Insight into alteration of gut microbiota in Clostridium difficile infection and asymptomatic C. difficile colonization Lihua Zhang, Danfeng Dong, Cen Jiang, Zhen Li, Xuefeng Wang, Yibing Peng* Department of Laboratory Medicine, Ruijin Hospital, Shanghai Jiaotong University School of Medicine, 197 Ruijin No.2 Road, Shanghai 200025, China
a r t i c l e i n f o
a b s t r a c t
Article history: Received 6 October 2014 Received in revised form 23 March 2015 Accepted 25 March 2015 Available online 26 March 2015
Clostridium difficile is well recognized as the common pathogen of nosocomial diarrhea, meanwhile, asymptomatic colonization with C. difficile in part of the population has also drawn public attention. Although gut microbiota is known to play an important role in the pathogenesis of C. difficile infection (CDI), whether there is any alteration of gut microbial composition in asymptomatic C. difficile carriers hasn't been clearly described. The purpose of this study was to explore the differences in gut microbiome among CDI patients, asymptomatic C. difficile carriers and healthy individuals. We performed fecal microbiota analysis on the samples of eight CDI patients, eight asymptomatic C. difficile carriers and nine healthy subjects using 16S rRNA gene pyrosequencing. CDI patients and asymptomatic carriers showed reduced microbial richness and diversity compared with healthy subjects, accompanied with a paucity of phylum Bacteroidetes and Firmicutes as well as an overabundance of Proteobacteria. Some normally commensal bacteria, especially butyrate producers, were significantly depleted in CDI patients and asymptomatic carriers. Furthermore, the differences observed in microbial community structure between CDI patients and asymptomatic carriers suggested that the gut microbiota may be a potential factor of disease state for CDI. Our study demonstrates the characterization and diversity of gut microbiota in CDI and asymptomatic C. difficile colonization, which will provide new ideas for surveillance of the disease state and development of microbiota-targeted agents for CDI prevention and treatment. © 2015 Elsevier Ltd. All rights reserved.
Keywords: Clostridium difficile infection Asymptomatic carrier Gut microbiota
1. Introduction Clostridium difficile, the common causative agent of antibioticassociated diarrhea and pseudomembraneous colitis, is known to be associated with a large part of nosocomial infections. Over the last decades, the incidence of C. difficile infection (CDI) has been increasing in hospitalized patients, with an estimated mortality rate ranging from 1% to 2.5% [1]. Many factors, such as increased antibiotic prescription, prolonged hospitalization and changes in diet, may have contributed to the wide spread of CDI. The hypervirulent strain ribotype 027, which caused outbreaks for many times in North America and Europe [2], adds much to the economic burden on healthcare settings and patients. In some Asian countries, the rising incidence of CDI has also aroused clinical concerns [3,4]. In spite of C. difficile-associated diseases, asymptomatic carriage of C. difficile in children and adults is also well recognized [5,6].
* Corresponding author. E-mail address: [email protected] (Y. Peng). http://dx.doi.org/10.1016/j.anaerobe.2015.03.008 1075-9964/© 2015 Elsevier Ltd. All rights reserved.
Conventional antibiotic therapy for CDI includes administration of metronidazole and vancomycin. However, there are still about 20% of the patients who will experience a recurrent episode of CDI after initial antibiotic treatment [7]. Thus, clinicians often resort to fecal microbiota transplantation (FMT) as an alternative approach to treat recurrent or refractory CDI if antibiotic regimens come to a failure [8]. CDI is believed to result from the disruption of gut microbiota, and FMT aims to restore the normal gut microbiota of the patient instead of direct eradication of the pathogen by certain antibiotics. Since first report of FMT in 1958, many studies have focused on the dynamic changes of the intestinal microbiome in response to FMT and the efficacy of FMT on CDI patients [9e13]. It has been discovered that patients with recurrent C. difficile-associated diarrhea are characterized by decreased diversity of gut microbiota and alteration of microbial composition [14]. However, whether people asymptomatically colonized with C. difficile have any notable difference from healthy individuals in their gut microbiota, or whether the intestinal bacterial communities of asymptomatic C. difficile carriers are just similar to those of CDI patients has yet to be clarified.
2
L. Zhang et al. / Anaerobe 34 (2015) 1e7
In this study, we analyzed the fecal samples from CDI patients, asymptomatic C. difficile carriers and healthy subjects by highthroughput 16S rRNA gene pyrosequencing so as to uncover the differences in intestinal microbial communities among the three groups. We hypothesized that changes in the intestinal microbiome may contribute to the pathogenesis of CDI and its disease state. 2. Materials and methods 2.1. Study design and specimen collection Our study involved twenty-five participants who were divided into 3 groups: patients with CDI (group 1, n ¼ 8), asymptomatic C. difficile carriers (group 2, n ¼ 8) and healthy individuals (group 3, n ¼ 9). CDI was defined as the presence of diarrhea along with positive results for toxigenic C. difficile in the patients' diarrheal fecal samples. A passage of more than 3 times of unformed stools was considered as diarrhea. For CDI patients, the presence of C. difficile was confirmed by stool culture and C. difficile toxin A & B was also tested (see below). The asymptomatic carriers had a positive stool culture for C. difficile but showed no diarrhea or other gastrointestinal disorders. The CDI patients and asymptomatic C. difficile carriers were selected from hospitalized patients. Healthy volunteers who underwent a stool test negative for C. difficile had been screened to ensure that they didn't have any gastrointestinal complaints, underlying chronic diseases or surgery and didn't receive any medications (within the preceding 3 months). The participants selected of the three groups were age-matched (group 1: mean 58.88 ± 22.24 years; group 2: mean 60.50 ± 20.79 years; group 3: mean 60.67 ± 16.02 years, P ¼ 0.98). Basic information of participants (including co-morbidities and antibiotic use) was displayed in Table S1. The fecal samples of each participate were collected in sterile containers and sent to the laboratory within 24 h. All samples were kept frozen at 80 C before DNA extraction.
sequencing primers and the letters in bold denote the universal 16S rRNA primers. The 8-bp sequence tag within the forward primer denoted by 8 Ns is used to barcode each sample. The fusion primer sequences were listed in Table S2 in the supplemental material. Amplification of the 16S rRNA V3eV4 region sequences was performed by PTC-100 thermal controller (MJ Research) using ExTaq DNA polymerase (Takara) with 0.25 mM forward and reverse primers and 5 ng of template DNA in a total reaction volume of 20 mL. Thermocycling involved 5 min of denaturation at 98 C, followed by 20 cycles of denaturation (98 C for 10 s), annealing (56 C for 30 s) and elongation (72 C for 30 s), with a final extension step at 72 C for 7 min. PCR products were confirmed by 2% agarose gel electrophoresis and purified by QIAquick Gel Extraction Kit (QIAGEN), after which quantification was done by the NanoDrop spectrophotometer (Thermo Scientific, USA). Equimolar amounts (50 ng) of the PCR amplicons of each sample were mixed together in a single tube and were purified using the AMPure XP Beads (Agencourt) to remove primers and reaction buffer. The amplicon mixtures were sequenced by 454 GS FLX þ Sequencer (Roche) at Chinese National Human Genome Center (CHGC) in Shanghai using standard procedures recommended from the Center. 2.5. Sequence processing and OTU classification Sequence reads were binned by samples with the samplespecific barcode sequences. After trimming the barcode and primer sequences, the raw sequence reads were filtered by QIIME pipeline (version 1.8.0) to optimize the quality of data with previously described criteria [15]. Operational taxonomic units (OTUs) were defined using a sequence similarity threshold of 97%. Taxonomic classification of each processed 16S rRNA gene sequence was performed at a genus level using the Ribosomal Database Project (RDP) Naïve Bayesian Classifier. All sequence data in this study were deposited in the NCBI Sequence Read Archive (http://www.ncbi. nlm.nih.gov/sra/; study accession number SRP051411).
2.2. Detection of C. difficile 2.6. Biodiversity and microbial community analyses The stool test for C. difficile toxin A & B was performed on a VIDAS automatic analyzer (Biomerieux, Marcy-l'Etoile, France) using enzyme-linked fluorescence assay (ELFA). Culturing for C. difficile vegetative cells was done by inoculating alcoholpretreated feces on cycloserine cefoxitin fructose agar (CCFA) medium (Oxoid Ltd., Basingstoke, UK) and incubating at 35 C for 48 h in anaerobic condition. Typical colonies were identified by latex agglutination test (C. difficile Agglutination Test Kit; Oxoid, UK). The existence of C. difficile in the fecal samples was also checked by amplification of gluD and tcdB gene using fecal genomic DNA template. 2.3. DNA extraction Fecal genomic DNA was extracted from 180 to 220 mg of each stool specimen with TIANamp Stool DNA Kit (Tiangen biotech, Beijing, CHN) by following the manufacturer's instructions. 2.4. 16S rRNA gene pyrosequencing The fusion primers consisted of the 454 FLX sequencing primers, a unique barcode sequence, and universal primers for amplification of the V3eV4 regions of 16S rRNA genes. The primers were as follows: forward-50 -CCATCTCATCCCTGCGTGTCTCCGACGACTNNNNNNNNTACGGRAGGCAGCAG-30 and reverse-50 CCTATCCCCTGTGTGCCTTGGCAGTCTCAGAGGGTATCTAATCCT-30 . The underlined sequences in the primer pairs stand for 454 FLX
The rarefaction curves based on OTU counts and ecological parameters of each fecal specimen (including abundance-based coverage estimator [ACE] richness index, Chao richness index, Shannon diversity index and Simpson diversity index) were calculated using mothur (version 1.32.1) [16]. The difference of these indices among the three sample groups was analyzed by an oneway ANOVA test, followed by a NewmaneKeuls test for multiple comparisons between any two groups (GraphPad Prism version 5.0). Principal coordinate analysis (PCoA) of unweighted UniFrac distance metric was conducted by QIIME to evaluate the variability in OTUs among CDI patients, asymptomatic carriers and healthy individuals. The violin plot showing the UniFrac distances within the groups and Venn diagram examining the community overlaps were all made by R (version 3.1.2). As for the analysis of intestinal microbiota composition, the relative abundances of all identified microbial taxa were compared between every two groups by Metastats 1.0 based on pooled data of each group. P-value < 0.05 represents significant difference. 3. Results 3.1. Sequence reads and ecological diversity Stool samples from twenty-five participants were processed, from which we obtained an average of 8048 pyrosequencing reads per sample with an average length of 416bp. The mean numbers of
L. Zhang et al. / Anaerobe 34 (2015) 1e7
OTUs in CDI patients (group 1), asymptomatic C. difficile carriers (group 2) and healthy controls (group 3) were 169, 203 and 300, respectively (P ¼ 0.014). As shown by the rarefaction curves (Fig. 1A), the fecal microbial diversity from CDI patients, asymptomatic carriers to healthy controls was in increasing order. When all specimens were pooled by group, we discovered significant difference in community richness and diversity among CDI patients, asymptomatic C. difficile carriers and healthy subjects (Fig. 1B, P < 0.05). Compared with the healthy controls, microbial richness decreased significantly in the samples of CDI patients and asymptomatic carriers, as demonstrated by ACE and Chao indices (P < 0.05). Besides, microbial diversity was also reduced in CDI patients and asymptomatic carriers compared to the healthy samples, based on the comparison of mean Shannon diversity/Simpson diversity index calculations (group 1 vs. group 3, P < 0.001; group 2 vs. group 3, P < 0.01). Notably, CDI patients and asymptomatic carriers did not vary significantly in their microbial richness and diversity. 3.2. Analysis of microbial community structure PCoA plot of unweighted UniFrac distances shown in Fig. 2A revealed that CDI patients, asymptomatic C. difficile carriers and healthy individuals were probably different subject clusters due to the composition of their microbial communities. Maximum parsimony analysis indicated significant differences in microbial community structure between every two groups (P < 0.001), which can account for the discrepancies in the presence or proportions of certain genera in separate groups. Moreover, the interindividual variation within the samples of CDI patients and asymptomatic carriers turned out to be greater than that within the healthy samples (Fig. 2B). A comparison of OTU structures across the gut microbial communities are shown by Venn diagram (Fig. 2C). The total number of OTUs in the three groups was 2905, from which the greatest number of OTUs (376 OTUs) was shared between the samples of asymptomatic C. difficile carriers and healthy controls, followed by the 349 OTUs shared between the CDI patients and asymptomatic carriers. The lowest number of shared OTUs (268 OTUs) was that between the samples of CDI patients and healthy controls. In addition to that, 1067 OTUs were only identified in the healthy fecal samples, but never in the samples of CDI patients and asymptomatic C. difficile carriers, which suggests that lack of certain commensal bacteria in the healthy intestine may be associated with C. difficile infection or colonization. 3.3. Phylum- and genus-level differences in microbial composition among groups Examination of the differences in the bacterial community composition at the phylum level revealed two trends. Generally, healthy subjects demonstrated a predominance of Bacteroidetes (48.2%) and Firmicutes (41.5%), whereas the gut microbiota of the CDI patients consisted of a greater proportion of Proteobacteria (33.6%) as well as a smaller proportion of Bacteroidetes (23.8%) and Firmicutes (27.3%) than that of the asymptomatic C. difficile carriers and healthy controls (Fig. 3A). Furthermore, the higher percentages of Fusobacteria in the CDI patients and asymptomatic carriers were also documented. Detailed data were shown in Table S3. At the genus-level (Fig. 3B), the relative abundance of Clostridium XI (which is in accord with C. difficile) in the gut microbiota of CDI patients, asymptomatic C. difficile carriers and healthy subjects were 6.5%, 1.6% and 0.0%, respectively. Within Bacteroidetes, we noticed significant decreases (P < 0.05) in the abundance of Alistipes, Bacteroides, and Prevotella, along with a significant
3
increase (P < 0.05) of Parabacteroides in the gut microbiota of CDI patients and asymptomatic carriers compared with that of healthy controls. By contrast, for Proteobacteria, significant increases in the abundance of Escherichia/Shigella, Klebsiella, Haemophilus, Pseudomonas, and Bilophila were found in CDI patients and/or C. difficile carriers (P < 0.05), but the percentages of Parasutterella and Gemmiger reduced in these two groups. As to Firmicutes, genera such as Clostridium XlVa, Clostridium sensu strict, Enterococcus, Veillonella, and Lactobacillus accounted for significantly higher proportions (P < 0.05) in the samples of CDI patients and asymptomatic carriers, whilst significantly lower proportions (P < 0.05) of Phascolarctobacterium, Roseburia, Megamonas, Ruminococcus, Faecalibacterium, and Coprococcus were also observed in these two groups than that in healthy subjects. It is noteworthy that the relative abundance of Bacteroides increased progressively from the CDI patients (10.8%), asymptomatic carriers (24.0%) to healthy controls (31.0%), and conversely, these three groups (from CDI patients, asymptomatic carriers to healthy controls) were in descending order (23.9%, 8.7%, 3.6%) in regards to the proportion of Escherichia/Shigella (P < 0.05). Besides, CDI patients showed a relative paucity of Bifidobacterium in the gut microbiota compared with asymptomatic carriers and healthy subjects. Fig. 3B only shows some of the genera holding relatively high percentages. Specific data of all the genera analyzed were displayed in Table S4. 4. Discussion It is well known that human gut is inhabited by trillions of microbes, which play important roles in nutrient metabolism, energy supply and immunologic functions. Alteration of normal intestinal microbiota can result from various factors and has close relevance to certain disease states, e.g. inflammatory bowel disease (IBD) [17], obesity [18] et al. Here the CDI cases and asymptomatic C. difficile carriers in this study were from hospitalized patients, whose intestinal microbiota may be influenced by factors such as antibiotic use or comorbidities. We observed the differences in gut microbial composition among CDI patients, asymptomatic C. difficile carriers and healthy individuals, which indicated that CDI status and C. difficile colonization possibly associated with the changes in gut microbiota. In recent years, several researches have focused on intestinal microbial communities in C. difficile-associated diseases [10,14,19e23]. Shahinas et al. [21] explored the microbiota dynamics in patients with CDI before and after FMT, and found those pre-transplant or failed transplant specimens demonstrated a decrease in microbial diversity and richness, in accordance with what we discovered in CDI patients of this study. Interestingly, the species richness and microbial diversity of asymptomatic C. difficile carriers seemed to be similar to that of CDI patients, however, the microbial community structure of the two groups were significantly different. Since the perturbations of gut microbiota play an important part in the etiology of CDI, we can infer that the microbial community structure of asymptomatic C. difficile carriers hasn't reached that of the CDI disease state, and colonization of C. difficile may not directly lead to C. difficile-associated diseases. Previous studies [21,24] have revealed that a successful FMT results in the eradication of Proteobacteria and the restoration of Bacteroidetes and Firmicutes, which implied that some species in Bacteroidetes or Firmicutes may be vital to outcompete or inhibit the growth of C. difficile in a way. Antharam et al. [25] reported a depletion of Ruminococcaceae, Lachnospiraceae, and other butyrateproducing bacteria in the patients with nosocomial diarrhea including CDI. Similarly, we noted a relative paucity of butyrateproducing Coprococcus and Roseburia (from Lachnospiraceae family) as well as Faecalibacterium and Ruminococcus (from
4
L. Zhang et al. / Anaerobe 34 (2015) 1e7
Fig. 1. (A) Rarefaction curves for samples of CDI patients (red), asymptomatic C. difficile carriers (blue) and healthy controls (green). Fecal microbial diversity among the groups is plotted against the number of V3eV4 region sequence reads for grouped samples. Each curve shows the average number of OTUs in samples of the respective group. (B) Indices of microbial richness and diversity for all samples (shown in group). The mean value and standard error of each group are denoted by a thin line in the middle of the plot and two thick lines above and below that line. CDI: CDI patients (group 1); AS: asymptomatic C. difficile carriers (group 2); CT: healthy controls (group 3). *P < 0.05; **P < 0.01; ***P < 0.001. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
L. Zhang et al. / Anaerobe 34 (2015) 1e7
5
Fig. 2. (A) Principal coordinate analysis (PCoA) of bacterial communities using unweighted UniFrac distances of 16S rRNA gene sequences. The horizontal axis and vertical axis represent 19.80% and 10.36% of the intersample variation, respectively. (B) Violin plot of unweighted UniFrac distances within the groups. The 3 spots of each “violin” represent upper quartile, median and lower quartile of the group. The mean distance value within healthy controls was significantly lower than that within CDI patients and asymptomatic carriers (t-test, P < 0.001). (C) Overlapping OTUs across the groups displayed by Venn diagram. OTUs were designated by a threshold of 0.03 genetic distance. The comparison of microbial communities between groups was made by pooled samples. CDI: CDI patients; AS: asymptomatic C. difficile carriers; CT: healthy controls.
Ruminococcaceae family) in CDI patients and C. difficile carriers. Anaerobic intestinal bacteria produce a variety of fermentation substrates that maintain the homeostasis of intestinal epithelial cells, including acetate, butyrate, lactate, and other short-chain fatty acids [26]. It is well documented that butyric acid can decrease intestinal permeability and enhance colonic defense barriers by increasing mucin production and antimicrobial peptide levels, thus prevents host from infection [25,27e29]. Butyrateproducing bacteria have been implicated as important members of the gut microbiota. Butyric acid serves as a preferred energy source of colonocytes [25], and has been observed to have antiinflammatory effects [29,30] and to downregulate bacterial virulence [28]. Therefore, depletion of these bacteria in CDI patients could potentially lead to epithelial dysfunction and higher osmotic load in the intestinal lumen, ultimately causing C. difficile-associated diarrhea. For asymptomatic carriers, lack of butyrate producers suggests that their gut microbiota would be susceptible to C. difficile infection, but we speculate the occurrence of disease may depend on the intestinal microenvironment as well as the host
immune state. In addition to butyrate-producing bacteria, Buffie et al. [31] recently discovered a bile acid 7a-dehydroxylating intestinal bacterium, Clostridium scindens, which was associated with resistance to CDI. Besides, Bacteroides, which appeared to be reduced in CDI patients and C. difficile carriers, are mainly work on the digestion of carbohydrates in the intestine, resulting in the production of essential substrates for the homeostasis of colonocytes [26]. Other genera with higher percents in CDI and C. difficile carrier groups, such as Enterococcus and Lactobacillus, might have some relations to the disease state, but the specific role of them is still undecided yet. Furthermore, due to the small sample size, it was difficult to define exact genus-level changes between the CDI group and C. difficile carrier group. Exposure to antibiotics is confirmed to be a major risk factor for CDI, because antibiotic-associated disturbance of the indigenous gut microbiota gives rise to shifts in carbohydrate, amino acid and bile acid metabolism [32], which may provide a niche for C. difficile to colonize and grow. In the present study, all of the CDI patients and 75% of the asymptomatic C. difficile carriers had received
6
L. Zhang et al. / Anaerobe 34 (2015) 1e7
antibiotic treatment (mainly cephalosporins or carbapenems) during the hospitalization before fecal sample collection (Table S1). Broad-spectrum antibiotic treatment may result in their decreased abundance of Bacteroides, thus converting the intestinal microenvironment to one that favors the germination and growth of C. difficile. On the other hand, the healthy individuals in the control group were chosen from the community (out of the hospital) with no exposure to antibiotics. In fact, the study has not separated the factor “living environment” that can lead to some inherent difference in gut microbial communities. In conclusion, CDI patients and asymptomatic C. difficile carriers demonstrated a significant decrease in gut microbial richness and diversity compared to healthy subjects. Intestinal dysbiosis, especially the depletion of certain butyrate-producing bacteria, may contribute to CDI occurrence or susceptibility to this disease. We considered that asymptomatic C. difficile colonization was associated with the intestinal microbial communities apart from the extrinsic factors. Despite the small sample number and other nonCDI factors affecting the gut microbiota, this study serves as a first step in addressing the role of gut microbiota not only as a risk factor for CDI but also as a potential factor of disease state, which will give enlightenment for disease surveillance and exploiting specific probiotics for prevention and treatment of CDI. Conflict of interest The authors have no conflict of interest to declare. Acknowledgments This study was supported by the National Natural Science Foundation of China (Grant No. 81371873), the Key Basic Research Project of Science and Technology Commission of Shanghai (Grant No. 12JC1406100) and the Research Project of Science and Technology Commission of Shanghai (Grant No. 15YF1407300). Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.anaerobe.2015.03.008. References
Fig. 3. (A) Relative abundance of phyla in pooled samples of different groups. Phyla were shown as the percentage of total OTUs. (B) Relative percentages of the most abundant bacterial genera of the three analyzed groups. CDI: CDI patients; AS: asymptomatic C. difficile carriers; CT: healthy controls.
[1] A.N. Ananthakrishnan, Clostridium difficile infection: epidemiology, risk factors and management, Nat. Rev. Gastroenterol. Hepatol. 8 (2011) 17e26. [2] J.R. O'Connor, S. Johnson, D.N. Gerding, Clostridium difficile infection caused by the epidemic BI/NAP1/027 strain, Gastroenterology 136 (2009) 1913e1924. [3] Y.S. Kim, D.S. Han, Y.H. Kim, W.H. Kim, J.S. Kim, H.S. Kim, et al., Incidence and clinical features of Clostridium difficile infection in Korea: a nationwide study, Epidemiol. Infect. 141 (2013) 189e194. [4] P.L. Lim, T.M.S. Barkham, L.M. Ling, F. Dimatatac, T. Alfred, B. Ang, Increasing incidence of Clostridium difficile-associated disease, Singapore, Emerg. Infect. Dis. 14 (2008) 1487e1489. [5] A.L. Galdys, J.S. Nelson, K.A. Shutt, J.L. Schlackman, D.L. Pakstis, A.W. Pasculle, et al., Prevalence and duration of asymptomatic Clostridium difficile carriage among healthy subjects in Pittsburgh, Pennsylvania, J. Clin. Microbiol. 52 (2014) 2406e2409. [6] M. Furuichi, E. Imajo, Y. Sato, S. Tanno, M. Kawada, S. Sato, Characteristics of Clostridium difficile colonization in Japanese children, J. Infect. Chemother. 20 (2014) 307e311. [7] C.P. Kelly, J.T. LaMont, Clostridium difficileemore difficult than ever, N. Engl. J. Med. 359 (2008) 1932e1940. [8] S. Di Bella, C. Drapeau, E. Garcia-Almodovar, N. Petrosillo, Fecal microbiota transplantation: the state of the art, Infect. Dis. Rep. 5 (2013) e13. [9] M.J. Hamilton, A.R. Weingarden, T. Unno, A. Khoruts, M.J. Sadowsky, Highthroughput DNA sequence analysis reveals stable engraftment of gut microbiota following transplantation of previously frozen fecal bacteria, Gut Microbes 4 (2013) 125e135. [10] F. Broecker, M. Kube, J. Klumpp, M. Schuppler, L. Biedermann, J. Hecht, et al., Analysis of the intestinal microbiome of a recovered Clostridium difficile patient after fecal transplantation, Digestion 88 (2013) 243e251.
L. Zhang et al. / Anaerobe 34 (2015) 1e7 [11] Z. Kassam, C.H. Lee, Y. Yuan, R.H. Hunt, Fecal microbiota transplantation for Clostridium difficile infection: systematic review and meta-analysis, Am. J. Gastroenterol. 108 (2013) 500e508. [12] M.J. Koenigsknecht, V.B. Young, Faecal microbiota transplantation for the treatment of recurrent Clostridium difficile infection: current promise and future needs, Curr. Opin. Gastroenterol. 29 (2013) 628e632. [13] R. Pathak, H.A. Enuh, A. Patel, P. Wickremesinghe, Treatment of relapsing Clostridium difficile infection using fecal microbiota transplantation, Clin. Exp. Gastroenterol. 7 (2013) 1e6. [14] J.Y. Chang, D.A. Antonopoulos, A. Kalra, A. Tonelli, W.T. Khalife, T.M. Schmidt, et al., Decreased diversity of the fecal microbiome in recurrent Clostridium difficile-associated diarrhea, J. Infect. Dis. 197 (2008) 435e438. [15] R. Flores, J. Shi, M.H. Gail, P. Gajer, J. Ravel, J.J. Goedert, Assessment of the human fecal microbiota: II. Reproducibility and associations of 16S rRNA pyrosequences, Eur. J. Clin. Invest. 42 (2012) 855e863. [16] P.D. Schloss, S.L. Westcott, T. Ryabin, J.R. Hall, M. Hartmann, E.B. Hollister, et al., Introducing mothur: open-source, platform-independent, communitysupported software for describing and comparing microbial communities, Appl. Environ. Microbiol. 75 (2009) 7537e7541. [17] J.R. Allegretti, M.J. Hamilton, Restoring the gut microbiome for the treatment of inflammatory bowel diseases, World J. Gastroenterol. 20 (2014) 3468e3474. [18] M. Million, J.C. Lagier, D. Yahav, M. Paul, Gut bacterial microbiota and obesity, Clin. Microbiol. Infect. 19 (2013) 305e313. [19] T.A. Rubin, C.E. Gessert, J. Aas, J.S. Bakken, Fecal microbiome transplantation for recurrent Clostridium difficile infection: report on a case series, Anaerobe 19 (2013) 22e26. [20] A.R. Manges, A. Labbe, V.G. Loo, J.K. Atherton, M.A. Behr, L. Masson, et al., Comparative metagenomic study of alterations to the intestinal microbiota and risk of nosocomial Clostridium difficile-associated disease, J. Infect. Dis. 202 (2010) 1877e1884. [21] D. Shahinas, M. Silverman, T. Sittler, C. Chiu, P. Kim, E. Allen-Vercoe, et al., Toward an understanding of changes in diversity associated with fecal microbiome transplantation based on 16S rRNA gene deep sequencing, MBio
7
3 (2012). [22] Y. Song, S. Garg, M. Girotra, C. Maddox, E.C. von Rosenvinge, A. Dutta, et al., Microbiota dynamics in patients treated with fecal microbiota transplantation for recurrent Clostridium difficile infection, PLoS One 8 (2013) e81330. [23] J. Skraban, S. Dzeroski, B. Zenko, D. Mongus, S. Gangl, M. Rupnik, Gut microbiota patterns associated with colonization of different Clostridium difficile ribotypes, PLoS One 8 (2013) e58005. [24] A.M. Seekatz, J. Aas, C.E. Gessert, T.A. Rubin, D.M. Saman, J.S. Bakken, et al., Recovery of the gut microbiome following fecal microbiota transplantation, MBio 5 (2014) e00893e14. [25] V.C. Antharam, E.C. Li, A. Ishmael, A. Sharma, V. Mai, K.H. Rand, et al., Intestinal dysbiosis and depletion of butyrogenic bacteria in Clostridium difficile infection and nosocomial diarrhea, J. Clin. Microbiol. 51 (2013) 2884e2892. [26] S. Bibbo, L.R. Lopetuso, G. Ianiro, T. Di Rienzo, A. Gasbarrini, G. Cammarota, Role of microbiota and innate immunity in recurrent infection, J. Immunol. Res. 2014 (2014) 462740. [27] J.M. Wong, R. de Souza, C.W. Kendall, A. Emam, D.J. Jenkins, Colonic health: fermentation and short chain fatty acids, J. Clin. Gastroenterol. 40 (2006) 235e243. [28] P. Guilloteau, L. Martin, V. Eeckhaut, R. Ducatelle, R. Zabielski, F. Van Immerseel, From the gut to the peripheral tissues: the multiple effects of butyrate, Nutr. Res. Rev. 23 (2010) 366e384. [29] S.I. Cook, J.H. Sellin, Review article: short chain fatty acids in health and disease, Aliment. Pharmacol. Ther. 12 (1998) 499e507. [30] H.M. Hamer, D. Jonkers, K. Venema, S. Vanhoutvin, F.J. Troost, R.J. Brummer, Review article: the role of butyrate on colonic function, Aliment. Pharmacol. Ther. 27 (2008) 104e119. [31] C.G. Buffie, V. Bucci, R.R. Stein, P.T. McKenney, L. Ling, A. Gobourne, et al., Precision microbiome reconstitution restores bile acid mediated resistance to Clostridium difficile, Nature 517 (7533) (2015) 205e208, http://dx.doi.org/ 10.1038/nature13828. [32] C.M. Theriot, V.B. Young, Microbial and metabolic interactions between the gastrointestinal tract and Clostridium difficile infection, Gut Microbes 5 (2014) 86e95.
Contents lists available at ScienceDirect
Anaerobe journal homepage: www.elsevier.com/locate/anaerobe
Molecular Biology, Genetics and Biotechnology
Insight into alteration of gut microbiota in Clostridium difficile infection and asymptomatic C. difficile colonization Lihua Zhang, Danfeng Dong, Cen Jiang, Zhen Li, Xuefeng Wang, Yibing Peng* Department of Laboratory Medicine, Ruijin Hospital, Shanghai Jiaotong University School of Medicine, 197 Ruijin No.2 Road, Shanghai 200025, China
a r t i c l e i n f o
a b s t r a c t
Article history: Received 6 October 2014 Received in revised form 23 March 2015 Accepted 25 March 2015 Available online 26 March 2015
Clostridium difficile is well recognized as the common pathogen of nosocomial diarrhea, meanwhile, asymptomatic colonization with C. difficile in part of the population has also drawn public attention. Although gut microbiota is known to play an important role in the pathogenesis of C. difficile infection (CDI), whether there is any alteration of gut microbial composition in asymptomatic C. difficile carriers hasn't been clearly described. The purpose of this study was to explore the differences in gut microbiome among CDI patients, asymptomatic C. difficile carriers and healthy individuals. We performed fecal microbiota analysis on the samples of eight CDI patients, eight asymptomatic C. difficile carriers and nine healthy subjects using 16S rRNA gene pyrosequencing. CDI patients and asymptomatic carriers showed reduced microbial richness and diversity compared with healthy subjects, accompanied with a paucity of phylum Bacteroidetes and Firmicutes as well as an overabundance of Proteobacteria. Some normally commensal bacteria, especially butyrate producers, were significantly depleted in CDI patients and asymptomatic carriers. Furthermore, the differences observed in microbial community structure between CDI patients and asymptomatic carriers suggested that the gut microbiota may be a potential factor of disease state for CDI. Our study demonstrates the characterization and diversity of gut microbiota in CDI and asymptomatic C. difficile colonization, which will provide new ideas for surveillance of the disease state and development of microbiota-targeted agents for CDI prevention and treatment. © 2015 Elsevier Ltd. All rights reserved.
Keywords: Clostridium difficile infection Asymptomatic carrier Gut microbiota
1. Introduction Clostridium difficile, the common causative agent of antibioticassociated diarrhea and pseudomembraneous colitis, is known to be associated with a large part of nosocomial infections. Over the last decades, the incidence of C. difficile infection (CDI) has been increasing in hospitalized patients, with an estimated mortality rate ranging from 1% to 2.5% [1]. Many factors, such as increased antibiotic prescription, prolonged hospitalization and changes in diet, may have contributed to the wide spread of CDI. The hypervirulent strain ribotype 027, which caused outbreaks for many times in North America and Europe [2], adds much to the economic burden on healthcare settings and patients. In some Asian countries, the rising incidence of CDI has also aroused clinical concerns [3,4]. In spite of C. difficile-associated diseases, asymptomatic carriage of C. difficile in children and adults is also well recognized [5,6].
* Corresponding author. E-mail address: [email protected] (Y. Peng). http://dx.doi.org/10.1016/j.anaerobe.2015.03.008 1075-9964/© 2015 Elsevier Ltd. All rights reserved.
Conventional antibiotic therapy for CDI includes administration of metronidazole and vancomycin. However, there are still about 20% of the patients who will experience a recurrent episode of CDI after initial antibiotic treatment [7]. Thus, clinicians often resort to fecal microbiota transplantation (FMT) as an alternative approach to treat recurrent or refractory CDI if antibiotic regimens come to a failure [8]. CDI is believed to result from the disruption of gut microbiota, and FMT aims to restore the normal gut microbiota of the patient instead of direct eradication of the pathogen by certain antibiotics. Since first report of FMT in 1958, many studies have focused on the dynamic changes of the intestinal microbiome in response to FMT and the efficacy of FMT on CDI patients [9e13]. It has been discovered that patients with recurrent C. difficile-associated diarrhea are characterized by decreased diversity of gut microbiota and alteration of microbial composition [14]. However, whether people asymptomatically colonized with C. difficile have any notable difference from healthy individuals in their gut microbiota, or whether the intestinal bacterial communities of asymptomatic C. difficile carriers are just similar to those of CDI patients has yet to be clarified.
2
L. Zhang et al. / Anaerobe 34 (2015) 1e7
In this study, we analyzed the fecal samples from CDI patients, asymptomatic C. difficile carriers and healthy subjects by highthroughput 16S rRNA gene pyrosequencing so as to uncover the differences in intestinal microbial communities among the three groups. We hypothesized that changes in the intestinal microbiome may contribute to the pathogenesis of CDI and its disease state. 2. Materials and methods 2.1. Study design and specimen collection Our study involved twenty-five participants who were divided into 3 groups: patients with CDI (group 1, n ¼ 8), asymptomatic C. difficile carriers (group 2, n ¼ 8) and healthy individuals (group 3, n ¼ 9). CDI was defined as the presence of diarrhea along with positive results for toxigenic C. difficile in the patients' diarrheal fecal samples. A passage of more than 3 times of unformed stools was considered as diarrhea. For CDI patients, the presence of C. difficile was confirmed by stool culture and C. difficile toxin A & B was also tested (see below). The asymptomatic carriers had a positive stool culture for C. difficile but showed no diarrhea or other gastrointestinal disorders. The CDI patients and asymptomatic C. difficile carriers were selected from hospitalized patients. Healthy volunteers who underwent a stool test negative for C. difficile had been screened to ensure that they didn't have any gastrointestinal complaints, underlying chronic diseases or surgery and didn't receive any medications (within the preceding 3 months). The participants selected of the three groups were age-matched (group 1: mean 58.88 ± 22.24 years; group 2: mean 60.50 ± 20.79 years; group 3: mean 60.67 ± 16.02 years, P ¼ 0.98). Basic information of participants (including co-morbidities and antibiotic use) was displayed in Table S1. The fecal samples of each participate were collected in sterile containers and sent to the laboratory within 24 h. All samples were kept frozen at 80 C before DNA extraction.
sequencing primers and the letters in bold denote the universal 16S rRNA primers. The 8-bp sequence tag within the forward primer denoted by 8 Ns is used to barcode each sample. The fusion primer sequences were listed in Table S2 in the supplemental material. Amplification of the 16S rRNA V3eV4 region sequences was performed by PTC-100 thermal controller (MJ Research) using ExTaq DNA polymerase (Takara) with 0.25 mM forward and reverse primers and 5 ng of template DNA in a total reaction volume of 20 mL. Thermocycling involved 5 min of denaturation at 98 C, followed by 20 cycles of denaturation (98 C for 10 s), annealing (56 C for 30 s) and elongation (72 C for 30 s), with a final extension step at 72 C for 7 min. PCR products were confirmed by 2% agarose gel electrophoresis and purified by QIAquick Gel Extraction Kit (QIAGEN), after which quantification was done by the NanoDrop spectrophotometer (Thermo Scientific, USA). Equimolar amounts (50 ng) of the PCR amplicons of each sample were mixed together in a single tube and were purified using the AMPure XP Beads (Agencourt) to remove primers and reaction buffer. The amplicon mixtures were sequenced by 454 GS FLX þ Sequencer (Roche) at Chinese National Human Genome Center (CHGC) in Shanghai using standard procedures recommended from the Center. 2.5. Sequence processing and OTU classification Sequence reads were binned by samples with the samplespecific barcode sequences. After trimming the barcode and primer sequences, the raw sequence reads were filtered by QIIME pipeline (version 1.8.0) to optimize the quality of data with previously described criteria [15]. Operational taxonomic units (OTUs) were defined using a sequence similarity threshold of 97%. Taxonomic classification of each processed 16S rRNA gene sequence was performed at a genus level using the Ribosomal Database Project (RDP) Naïve Bayesian Classifier. All sequence data in this study were deposited in the NCBI Sequence Read Archive (http://www.ncbi. nlm.nih.gov/sra/; study accession number SRP051411).
2.2. Detection of C. difficile 2.6. Biodiversity and microbial community analyses The stool test for C. difficile toxin A & B was performed on a VIDAS automatic analyzer (Biomerieux, Marcy-l'Etoile, France) using enzyme-linked fluorescence assay (ELFA). Culturing for C. difficile vegetative cells was done by inoculating alcoholpretreated feces on cycloserine cefoxitin fructose agar (CCFA) medium (Oxoid Ltd., Basingstoke, UK) and incubating at 35 C for 48 h in anaerobic condition. Typical colonies were identified by latex agglutination test (C. difficile Agglutination Test Kit; Oxoid, UK). The existence of C. difficile in the fecal samples was also checked by amplification of gluD and tcdB gene using fecal genomic DNA template. 2.3. DNA extraction Fecal genomic DNA was extracted from 180 to 220 mg of each stool specimen with TIANamp Stool DNA Kit (Tiangen biotech, Beijing, CHN) by following the manufacturer's instructions. 2.4. 16S rRNA gene pyrosequencing The fusion primers consisted of the 454 FLX sequencing primers, a unique barcode sequence, and universal primers for amplification of the V3eV4 regions of 16S rRNA genes. The primers were as follows: forward-50 -CCATCTCATCCCTGCGTGTCTCCGACGACTNNNNNNNNTACGGRAGGCAGCAG-30 and reverse-50 CCTATCCCCTGTGTGCCTTGGCAGTCTCAGAGGGTATCTAATCCT-30 . The underlined sequences in the primer pairs stand for 454 FLX
The rarefaction curves based on OTU counts and ecological parameters of each fecal specimen (including abundance-based coverage estimator [ACE] richness index, Chao richness index, Shannon diversity index and Simpson diversity index) were calculated using mothur (version 1.32.1) [16]. The difference of these indices among the three sample groups was analyzed by an oneway ANOVA test, followed by a NewmaneKeuls test for multiple comparisons between any two groups (GraphPad Prism version 5.0). Principal coordinate analysis (PCoA) of unweighted UniFrac distance metric was conducted by QIIME to evaluate the variability in OTUs among CDI patients, asymptomatic carriers and healthy individuals. The violin plot showing the UniFrac distances within the groups and Venn diagram examining the community overlaps were all made by R (version 3.1.2). As for the analysis of intestinal microbiota composition, the relative abundances of all identified microbial taxa were compared between every two groups by Metastats 1.0 based on pooled data of each group. P-value < 0.05 represents significant difference. 3. Results 3.1. Sequence reads and ecological diversity Stool samples from twenty-five participants were processed, from which we obtained an average of 8048 pyrosequencing reads per sample with an average length of 416bp. The mean numbers of
L. Zhang et al. / Anaerobe 34 (2015) 1e7
OTUs in CDI patients (group 1), asymptomatic C. difficile carriers (group 2) and healthy controls (group 3) were 169, 203 and 300, respectively (P ¼ 0.014). As shown by the rarefaction curves (Fig. 1A), the fecal microbial diversity from CDI patients, asymptomatic carriers to healthy controls was in increasing order. When all specimens were pooled by group, we discovered significant difference in community richness and diversity among CDI patients, asymptomatic C. difficile carriers and healthy subjects (Fig. 1B, P < 0.05). Compared with the healthy controls, microbial richness decreased significantly in the samples of CDI patients and asymptomatic carriers, as demonstrated by ACE and Chao indices (P < 0.05). Besides, microbial diversity was also reduced in CDI patients and asymptomatic carriers compared to the healthy samples, based on the comparison of mean Shannon diversity/Simpson diversity index calculations (group 1 vs. group 3, P < 0.001; group 2 vs. group 3, P < 0.01). Notably, CDI patients and asymptomatic carriers did not vary significantly in their microbial richness and diversity. 3.2. Analysis of microbial community structure PCoA plot of unweighted UniFrac distances shown in Fig. 2A revealed that CDI patients, asymptomatic C. difficile carriers and healthy individuals were probably different subject clusters due to the composition of their microbial communities. Maximum parsimony analysis indicated significant differences in microbial community structure between every two groups (P < 0.001), which can account for the discrepancies in the presence or proportions of certain genera in separate groups. Moreover, the interindividual variation within the samples of CDI patients and asymptomatic carriers turned out to be greater than that within the healthy samples (Fig. 2B). A comparison of OTU structures across the gut microbial communities are shown by Venn diagram (Fig. 2C). The total number of OTUs in the three groups was 2905, from which the greatest number of OTUs (376 OTUs) was shared between the samples of asymptomatic C. difficile carriers and healthy controls, followed by the 349 OTUs shared between the CDI patients and asymptomatic carriers. The lowest number of shared OTUs (268 OTUs) was that between the samples of CDI patients and healthy controls. In addition to that, 1067 OTUs were only identified in the healthy fecal samples, but never in the samples of CDI patients and asymptomatic C. difficile carriers, which suggests that lack of certain commensal bacteria in the healthy intestine may be associated with C. difficile infection or colonization. 3.3. Phylum- and genus-level differences in microbial composition among groups Examination of the differences in the bacterial community composition at the phylum level revealed two trends. Generally, healthy subjects demonstrated a predominance of Bacteroidetes (48.2%) and Firmicutes (41.5%), whereas the gut microbiota of the CDI patients consisted of a greater proportion of Proteobacteria (33.6%) as well as a smaller proportion of Bacteroidetes (23.8%) and Firmicutes (27.3%) than that of the asymptomatic C. difficile carriers and healthy controls (Fig. 3A). Furthermore, the higher percentages of Fusobacteria in the CDI patients and asymptomatic carriers were also documented. Detailed data were shown in Table S3. At the genus-level (Fig. 3B), the relative abundance of Clostridium XI (which is in accord with C. difficile) in the gut microbiota of CDI patients, asymptomatic C. difficile carriers and healthy subjects were 6.5%, 1.6% and 0.0%, respectively. Within Bacteroidetes, we noticed significant decreases (P < 0.05) in the abundance of Alistipes, Bacteroides, and Prevotella, along with a significant
3
increase (P < 0.05) of Parabacteroides in the gut microbiota of CDI patients and asymptomatic carriers compared with that of healthy controls. By contrast, for Proteobacteria, significant increases in the abundance of Escherichia/Shigella, Klebsiella, Haemophilus, Pseudomonas, and Bilophila were found in CDI patients and/or C. difficile carriers (P < 0.05), but the percentages of Parasutterella and Gemmiger reduced in these two groups. As to Firmicutes, genera such as Clostridium XlVa, Clostridium sensu strict, Enterococcus, Veillonella, and Lactobacillus accounted for significantly higher proportions (P < 0.05) in the samples of CDI patients and asymptomatic carriers, whilst significantly lower proportions (P < 0.05) of Phascolarctobacterium, Roseburia, Megamonas, Ruminococcus, Faecalibacterium, and Coprococcus were also observed in these two groups than that in healthy subjects. It is noteworthy that the relative abundance of Bacteroides increased progressively from the CDI patients (10.8%), asymptomatic carriers (24.0%) to healthy controls (31.0%), and conversely, these three groups (from CDI patients, asymptomatic carriers to healthy controls) were in descending order (23.9%, 8.7%, 3.6%) in regards to the proportion of Escherichia/Shigella (P < 0.05). Besides, CDI patients showed a relative paucity of Bifidobacterium in the gut microbiota compared with asymptomatic carriers and healthy subjects. Fig. 3B only shows some of the genera holding relatively high percentages. Specific data of all the genera analyzed were displayed in Table S4. 4. Discussion It is well known that human gut is inhabited by trillions of microbes, which play important roles in nutrient metabolism, energy supply and immunologic functions. Alteration of normal intestinal microbiota can result from various factors and has close relevance to certain disease states, e.g. inflammatory bowel disease (IBD) [17], obesity [18] et al. Here the CDI cases and asymptomatic C. difficile carriers in this study were from hospitalized patients, whose intestinal microbiota may be influenced by factors such as antibiotic use or comorbidities. We observed the differences in gut microbial composition among CDI patients, asymptomatic C. difficile carriers and healthy individuals, which indicated that CDI status and C. difficile colonization possibly associated with the changes in gut microbiota. In recent years, several researches have focused on intestinal microbial communities in C. difficile-associated diseases [10,14,19e23]. Shahinas et al. [21] explored the microbiota dynamics in patients with CDI before and after FMT, and found those pre-transplant or failed transplant specimens demonstrated a decrease in microbial diversity and richness, in accordance with what we discovered in CDI patients of this study. Interestingly, the species richness and microbial diversity of asymptomatic C. difficile carriers seemed to be similar to that of CDI patients, however, the microbial community structure of the two groups were significantly different. Since the perturbations of gut microbiota play an important part in the etiology of CDI, we can infer that the microbial community structure of asymptomatic C. difficile carriers hasn't reached that of the CDI disease state, and colonization of C. difficile may not directly lead to C. difficile-associated diseases. Previous studies [21,24] have revealed that a successful FMT results in the eradication of Proteobacteria and the restoration of Bacteroidetes and Firmicutes, which implied that some species in Bacteroidetes or Firmicutes may be vital to outcompete or inhibit the growth of C. difficile in a way. Antharam et al. [25] reported a depletion of Ruminococcaceae, Lachnospiraceae, and other butyrateproducing bacteria in the patients with nosocomial diarrhea including CDI. Similarly, we noted a relative paucity of butyrateproducing Coprococcus and Roseburia (from Lachnospiraceae family) as well as Faecalibacterium and Ruminococcus (from
4
L. Zhang et al. / Anaerobe 34 (2015) 1e7
Fig. 1. (A) Rarefaction curves for samples of CDI patients (red), asymptomatic C. difficile carriers (blue) and healthy controls (green). Fecal microbial diversity among the groups is plotted against the number of V3eV4 region sequence reads for grouped samples. Each curve shows the average number of OTUs in samples of the respective group. (B) Indices of microbial richness and diversity for all samples (shown in group). The mean value and standard error of each group are denoted by a thin line in the middle of the plot and two thick lines above and below that line. CDI: CDI patients (group 1); AS: asymptomatic C. difficile carriers (group 2); CT: healthy controls (group 3). *P < 0.05; **P < 0.01; ***P < 0.001. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
L. Zhang et al. / Anaerobe 34 (2015) 1e7
5
Fig. 2. (A) Principal coordinate analysis (PCoA) of bacterial communities using unweighted UniFrac distances of 16S rRNA gene sequences. The horizontal axis and vertical axis represent 19.80% and 10.36% of the intersample variation, respectively. (B) Violin plot of unweighted UniFrac distances within the groups. The 3 spots of each “violin” represent upper quartile, median and lower quartile of the group. The mean distance value within healthy controls was significantly lower than that within CDI patients and asymptomatic carriers (t-test, P < 0.001). (C) Overlapping OTUs across the groups displayed by Venn diagram. OTUs were designated by a threshold of 0.03 genetic distance. The comparison of microbial communities between groups was made by pooled samples. CDI: CDI patients; AS: asymptomatic C. difficile carriers; CT: healthy controls.
Ruminococcaceae family) in CDI patients and C. difficile carriers. Anaerobic intestinal bacteria produce a variety of fermentation substrates that maintain the homeostasis of intestinal epithelial cells, including acetate, butyrate, lactate, and other short-chain fatty acids [26]. It is well documented that butyric acid can decrease intestinal permeability and enhance colonic defense barriers by increasing mucin production and antimicrobial peptide levels, thus prevents host from infection [25,27e29]. Butyrateproducing bacteria have been implicated as important members of the gut microbiota. Butyric acid serves as a preferred energy source of colonocytes [25], and has been observed to have antiinflammatory effects [29,30] and to downregulate bacterial virulence [28]. Therefore, depletion of these bacteria in CDI patients could potentially lead to epithelial dysfunction and higher osmotic load in the intestinal lumen, ultimately causing C. difficile-associated diarrhea. For asymptomatic carriers, lack of butyrate producers suggests that their gut microbiota would be susceptible to C. difficile infection, but we speculate the occurrence of disease may depend on the intestinal microenvironment as well as the host
immune state. In addition to butyrate-producing bacteria, Buffie et al. [31] recently discovered a bile acid 7a-dehydroxylating intestinal bacterium, Clostridium scindens, which was associated with resistance to CDI. Besides, Bacteroides, which appeared to be reduced in CDI patients and C. difficile carriers, are mainly work on the digestion of carbohydrates in the intestine, resulting in the production of essential substrates for the homeostasis of colonocytes [26]. Other genera with higher percents in CDI and C. difficile carrier groups, such as Enterococcus and Lactobacillus, might have some relations to the disease state, but the specific role of them is still undecided yet. Furthermore, due to the small sample size, it was difficult to define exact genus-level changes between the CDI group and C. difficile carrier group. Exposure to antibiotics is confirmed to be a major risk factor for CDI, because antibiotic-associated disturbance of the indigenous gut microbiota gives rise to shifts in carbohydrate, amino acid and bile acid metabolism [32], which may provide a niche for C. difficile to colonize and grow. In the present study, all of the CDI patients and 75% of the asymptomatic C. difficile carriers had received
6
L. Zhang et al. / Anaerobe 34 (2015) 1e7
antibiotic treatment (mainly cephalosporins or carbapenems) during the hospitalization before fecal sample collection (Table S1). Broad-spectrum antibiotic treatment may result in their decreased abundance of Bacteroides, thus converting the intestinal microenvironment to one that favors the germination and growth of C. difficile. On the other hand, the healthy individuals in the control group were chosen from the community (out of the hospital) with no exposure to antibiotics. In fact, the study has not separated the factor “living environment” that can lead to some inherent difference in gut microbial communities. In conclusion, CDI patients and asymptomatic C. difficile carriers demonstrated a significant decrease in gut microbial richness and diversity compared to healthy subjects. Intestinal dysbiosis, especially the depletion of certain butyrate-producing bacteria, may contribute to CDI occurrence or susceptibility to this disease. We considered that asymptomatic C. difficile colonization was associated with the intestinal microbial communities apart from the extrinsic factors. Despite the small sample number and other nonCDI factors affecting the gut microbiota, this study serves as a first step in addressing the role of gut microbiota not only as a risk factor for CDI but also as a potential factor of disease state, which will give enlightenment for disease surveillance and exploiting specific probiotics for prevention and treatment of CDI. Conflict of interest The authors have no conflict of interest to declare. Acknowledgments This study was supported by the National Natural Science Foundation of China (Grant No. 81371873), the Key Basic Research Project of Science and Technology Commission of Shanghai (Grant No. 12JC1406100) and the Research Project of Science and Technology Commission of Shanghai (Grant No. 15YF1407300). Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.anaerobe.2015.03.008. References
Fig. 3. (A) Relative abundance of phyla in pooled samples of different groups. Phyla were shown as the percentage of total OTUs. (B) Relative percentages of the most abundant bacterial genera of the three analyzed groups. CDI: CDI patients; AS: asymptomatic C. difficile carriers; CT: healthy controls.
[1] A.N. Ananthakrishnan, Clostridium difficile infection: epidemiology, risk factors and management, Nat. Rev. Gastroenterol. Hepatol. 8 (2011) 17e26. [2] J.R. O'Connor, S. Johnson, D.N. Gerding, Clostridium difficile infection caused by the epidemic BI/NAP1/027 strain, Gastroenterology 136 (2009) 1913e1924. [3] Y.S. Kim, D.S. Han, Y.H. Kim, W.H. Kim, J.S. Kim, H.S. Kim, et al., Incidence and clinical features of Clostridium difficile infection in Korea: a nationwide study, Epidemiol. Infect. 141 (2013) 189e194. [4] P.L. Lim, T.M.S. Barkham, L.M. Ling, F. Dimatatac, T. Alfred, B. Ang, Increasing incidence of Clostridium difficile-associated disease, Singapore, Emerg. Infect. Dis. 14 (2008) 1487e1489. [5] A.L. Galdys, J.S. Nelson, K.A. Shutt, J.L. Schlackman, D.L. Pakstis, A.W. Pasculle, et al., Prevalence and duration of asymptomatic Clostridium difficile carriage among healthy subjects in Pittsburgh, Pennsylvania, J. Clin. Microbiol. 52 (2014) 2406e2409. [6] M. Furuichi, E. Imajo, Y. Sato, S. Tanno, M. Kawada, S. Sato, Characteristics of Clostridium difficile colonization in Japanese children, J. Infect. Chemother. 20 (2014) 307e311. [7] C.P. Kelly, J.T. LaMont, Clostridium difficileemore difficult than ever, N. Engl. J. Med. 359 (2008) 1932e1940. [8] S. Di Bella, C. Drapeau, E. Garcia-Almodovar, N. Petrosillo, Fecal microbiota transplantation: the state of the art, Infect. Dis. Rep. 5 (2013) e13. [9] M.J. Hamilton, A.R. Weingarden, T. Unno, A. Khoruts, M.J. Sadowsky, Highthroughput DNA sequence analysis reveals stable engraftment of gut microbiota following transplantation of previously frozen fecal bacteria, Gut Microbes 4 (2013) 125e135. [10] F. Broecker, M. Kube, J. Klumpp, M. Schuppler, L. Biedermann, J. Hecht, et al., Analysis of the intestinal microbiome of a recovered Clostridium difficile patient after fecal transplantation, Digestion 88 (2013) 243e251.
L. Zhang et al. / Anaerobe 34 (2015) 1e7 [11] Z. Kassam, C.H. Lee, Y. Yuan, R.H. Hunt, Fecal microbiota transplantation for Clostridium difficile infection: systematic review and meta-analysis, Am. J. Gastroenterol. 108 (2013) 500e508. [12] M.J. Koenigsknecht, V.B. Young, Faecal microbiota transplantation for the treatment of recurrent Clostridium difficile infection: current promise and future needs, Curr. Opin. Gastroenterol. 29 (2013) 628e632. [13] R. Pathak, H.A. Enuh, A. Patel, P. Wickremesinghe, Treatment of relapsing Clostridium difficile infection using fecal microbiota transplantation, Clin. Exp. Gastroenterol. 7 (2013) 1e6. [14] J.Y. Chang, D.A. Antonopoulos, A. Kalra, A. Tonelli, W.T. Khalife, T.M. Schmidt, et al., Decreased diversity of the fecal microbiome in recurrent Clostridium difficile-associated diarrhea, J. Infect. Dis. 197 (2008) 435e438. [15] R. Flores, J. Shi, M.H. Gail, P. Gajer, J. Ravel, J.J. Goedert, Assessment of the human fecal microbiota: II. Reproducibility and associations of 16S rRNA pyrosequences, Eur. J. Clin. Invest. 42 (2012) 855e863. [16] P.D. Schloss, S.L. Westcott, T. Ryabin, J.R. Hall, M. Hartmann, E.B. Hollister, et al., Introducing mothur: open-source, platform-independent, communitysupported software for describing and comparing microbial communities, Appl. Environ. Microbiol. 75 (2009) 7537e7541. [17] J.R. Allegretti, M.J. Hamilton, Restoring the gut microbiome for the treatment of inflammatory bowel diseases, World J. Gastroenterol. 20 (2014) 3468e3474. [18] M. Million, J.C. Lagier, D. Yahav, M. Paul, Gut bacterial microbiota and obesity, Clin. Microbiol. Infect. 19 (2013) 305e313. [19] T.A. Rubin, C.E. Gessert, J. Aas, J.S. Bakken, Fecal microbiome transplantation for recurrent Clostridium difficile infection: report on a case series, Anaerobe 19 (2013) 22e26. [20] A.R. Manges, A. Labbe, V.G. Loo, J.K. Atherton, M.A. Behr, L. Masson, et al., Comparative metagenomic study of alterations to the intestinal microbiota and risk of nosocomial Clostridium difficile-associated disease, J. Infect. Dis. 202 (2010) 1877e1884. [21] D. Shahinas, M. Silverman, T. Sittler, C. Chiu, P. Kim, E. Allen-Vercoe, et al., Toward an understanding of changes in diversity associated with fecal microbiome transplantation based on 16S rRNA gene deep sequencing, MBio
7
3 (2012). [22] Y. Song, S. Garg, M. Girotra, C. Maddox, E.C. von Rosenvinge, A. Dutta, et al., Microbiota dynamics in patients treated with fecal microbiota transplantation for recurrent Clostridium difficile infection, PLoS One 8 (2013) e81330. [23] J. Skraban, S. Dzeroski, B. Zenko, D. Mongus, S. Gangl, M. Rupnik, Gut microbiota patterns associated with colonization of different Clostridium difficile ribotypes, PLoS One 8 (2013) e58005. [24] A.M. Seekatz, J. Aas, C.E. Gessert, T.A. Rubin, D.M. Saman, J.S. Bakken, et al., Recovery of the gut microbiome following fecal microbiota transplantation, MBio 5 (2014) e00893e14. [25] V.C. Antharam, E.C. Li, A. Ishmael, A. Sharma, V. Mai, K.H. Rand, et al., Intestinal dysbiosis and depletion of butyrogenic bacteria in Clostridium difficile infection and nosocomial diarrhea, J. Clin. Microbiol. 51 (2013) 2884e2892. [26] S. Bibbo, L.R. Lopetuso, G. Ianiro, T. Di Rienzo, A. Gasbarrini, G. Cammarota, Role of microbiota and innate immunity in recurrent infection, J. Immunol. Res. 2014 (2014) 462740. [27] J.M. Wong, R. de Souza, C.W. Kendall, A. Emam, D.J. Jenkins, Colonic health: fermentation and short chain fatty acids, J. Clin. Gastroenterol. 40 (2006) 235e243. [28] P. Guilloteau, L. Martin, V. Eeckhaut, R. Ducatelle, R. Zabielski, F. Van Immerseel, From the gut to the peripheral tissues: the multiple effects of butyrate, Nutr. Res. Rev. 23 (2010) 366e384. [29] S.I. Cook, J.H. Sellin, Review article: short chain fatty acids in health and disease, Aliment. Pharmacol. Ther. 12 (1998) 499e507. [30] H.M. Hamer, D. Jonkers, K. Venema, S. Vanhoutvin, F.J. Troost, R.J. Brummer, Review article: the role of butyrate on colonic function, Aliment. Pharmacol. Ther. 27 (2008) 104e119. [31] C.G. Buffie, V. Bucci, R.R. Stein, P.T. McKenney, L. Ling, A. Gobourne, et al., Precision microbiome reconstitution restores bile acid mediated resistance to Clostridium difficile, Nature 517 (7533) (2015) 205e208, http://dx.doi.org/ 10.1038/nature13828. [32] C.M. Theriot, V.B. Young, Microbial and metabolic interactions between the gastrointestinal tract and Clostridium difficile infection, Gut Microbes 5 (2014) 86e95.