Advertisement
Research Article

Microbiota Dynamics in Patients Treated with Fecal Microbiota Transplantation for Recurrent Clostridium difficile Infection

  • Yang Song,

    Affiliation: Institute for Genome Sciences, University of Maryland School of Medicine, Baltimore, Maryland, United States of America

    X
  • Shashank Garg,

    Affiliation: Division of Gastroenterology, Sinai Hospital of Baltimore, Baltimore, Maryland, United States of America

    X
  • Mohit Girotra,

    Affiliation: Division of Gastroenterology, Sinai Hospital of Baltimore, Baltimore, Maryland, United States of America

    X
  • Cynthia Maddox,

    Affiliation: Institute for Genome Sciences, University of Maryland School of Medicine, Baltimore, Maryland, United States of America

    X
  • Erik C. von Rosenvinge,

    Affiliation: Division of Gastroenterology and Hepatology, University of Maryland School of Medicine, Baltimore, Maryland, United States of America

    X
  • Anand Dutta,

    Affiliation: Division of Gastroenterology, Sinai Hospital of Baltimore, Baltimore, Maryland, United States of America

    X
  • Sudhir Dutta,

    Affiliations: Division of Gastroenterology, Sinai Hospital of Baltimore, Baltimore, Maryland, United States of America, Department of Medicine, University of Maryland School of Medicine, Baltimore, Maryland, United States of America

    X
  • W. Florian Fricke mail

    wffricke@som.umaryland.edu

    Affiliation: Institute for Genome Sciences, University of Maryland School of Medicine, Baltimore, Maryland, United States of America

    X
  • Published: November 26, 2013
  • DOI: 10.1371/journal.pone.0081330

Abstract

Clostridium difficile causes antibiotic-associated diarrhea and pseudomembraneous colitis and is responsible for a large and increasing fraction of hospital-acquired infections. Fecal microbiota transplantation (FMT) is an alternate treatment option for recurrent C. difficile infection (RCDI) refractory to antibiotic therapy. It has recently been discussed favorably in the clinical and scientific communities and is receiving increasing public attention. However, short- and long-term health consequences of FMT remain a concern, as the effects of the transplanted microbiota on the patient remain unknown. To shed light on microbial events associated with RCDI and treatment by FMT, we performed fecal microbiota analysis by 16S rRNA gene amplicon pyrosequencing of 14 pairs of healthy donors and RCDI patients treated successfully by FMT. Post-FMT patient and healthy donor samples collected up to one year after FMT were studied longitudinally, including one post-FMT patient with antibiotic-associated relapse three months after FMT. This analysis allowed us not only to confirm prior reports that RCDI is associated with reduced diversity and compositional changes in the fecal microbiota, but also to characterize previously undocumented post-FMT microbiota dynamics. Members of the Streptococcaceae, Enterococcaceae, or Enterobacteriaceae were significantly increased and putative butyrate producers, such as Lachnospiraceae and Ruminococcaceae were significantly reduced in samples from RCDI patients before FMT as compared to post-FMT patient and healthy donor samples. RCDI patient samples showed more case-specific variations than post-FMT patient and healthy donor samples. However, none of the bacterial groups were invariably associated with RCDI or successful treatment by FMT. Overall microbiota compositions in post-FMT patients, specifically abundances of the above-mentioned Firmicutes, continued to change for at least 16 weeks after FMT, suggesting that full microbiota recovery from RCDI may take much longer than expected based on the disappearance of diarrheal symptoms immediately after FMT.

Introduction

Clostridium difficile, the pathogen associated with the majority of infective antibiotic-associated diarrhea and causative agent of pseudomembraneous colitis [1], is responsible for a large fraction of nosocomial, or hospital-acquired, disease [2]. Today, in parts of the U.S., the incidence of infections with C. difficile is higher than that of methicillin-resistant Staphylococcus aureus [3]. C. difficile infection (CDI) is believed to result from gastrointestinal dysbiosis, i.e., the disruption of the resident microbiota, often caused by antibiotic treatment, which enables C. difficile to establish an infection. C. difficile can be acquired via fecal-oral transmission of spores that survive atmospheric oxygen and gastric acid exposure and germinate in the large intestine. However, carriage of C. difficile is not always associated with disease, as asymptomatic C. difficile colonization is well recognized [4], especially in newborns and infants of <1 year age [5].

Besides treatment with almost any antibiotic [6][14], other factors associated with increased risk for C. difficile infection include old age, recent hospitalization, tube feeding, use of gastric acid-suppressing drugs and underlying chronic disease, including inflammatory bowel disease [15][19]. Recent evidence suggests that excessive inflammatory responses in the human host enhance the severity of CDI [20].

Standard treatment for C. difficile infection consists of metronidazole or vancomycin administration and, more recently, fidaxomicin. However, the rate of recurrent C. difficile infection (RCDI) after initial therapy is about 20% [21] and even higher after subsequent antibiotic courses and recurrences [8], [22]. Consequently, despite current therapeutic options, RCDI treatment has become increasingly challenging and the incidence of RCDI has been rising during the past decade resulting in increased healthcare cost and significant morbidity [23].

Fecal microbiota transplantation (FMT), which aims to restore a normal, functional intestinal microbiota from a healthy donor in the RCDI patient, has recently received increasing attention in clinical and research communities [24][27] and has also become a popular subject of discussion in other media. First documented in the fourth century in China and in 1958 in the U.S., FMT was shown in a recent systematic review of 317 patients in 27 separate studies to have an overall success rate of 92% [28]. The exact mechanism of action responsible for the success of FMT to treat RCDI remains unknown and there is no clinically validated set of parameters to define a suitable donor or ideal donor microbiota, although attempts in this direction have been made [29]. Short- and long-term effects of FMT on the recipient microbiota remain a concern, especially in light of the growing body of literature that implicates the gastrointestinal microbiota in a large number of diseases [30]. For the same reason, there is significant clinical interest in therapeutic options to target the microbiota to treat microbiota-associated health problems besides RCDI. As a result, attempts to treat IBD [31][33], metabolic syndrome [34] and other diseases [35], [36] by FMT have been made.

Clinical concerns and the increasing number of FMT procedures performed by U.S. physicians recently led the U.S. Food and Drug Administration (FDA) to release new guidelines that define FMT as a biologic therapy that requires physicians to obtain an investigational new drug (IND) application [37]. Shortly after this guideline was a released, however, the FDA announced a decision to exercise enforcement discretion in order to allow physicians to perform FMT in patients with RCDI not responsive to standard therapy. The urgency for further research into the short- and long-term effects of FMT is highlighted by the fact that the public awareness of FMT as a treatment option for RCDI has increased to a degree where do-it-yourself protocols have become available over the Internet and the procedure is being performed without medical surveillance.

In this study, we applied 16S rRNA amplicon pyrosequencing to analyze fecal samples from RCDI patients and their corresponding donors before and after FMT. For the first time, we included longitudinal simultaneous sampling of both post-FMT patients and healthy donors for up to one year after FMT. This unique sample set allowed us to describe previously undocumented microbiota dynamics in post-FMT patients after resolution of CDI. In addition, inclusion of a patient, who was initially treated successfully by FMT but experienced relapse after new antibiotic treatment, provided us with the unique opportunity to distinguish microbiota changes seen in a previously asymptomatic patients after relapse of CDI from those apparent in RCDI patients with long-term disease and multiple courses of anti-C. difficile antibiotic treatment.

Materials and Methods

Study cohort and sample collection

The Institutional Review Board of Sinai Hospital Baltimore approved the study under protocol number #1826 and all subjects provided their written informed consent to participate in the study. FMT was performed at Sinai Hospital of Baltimore, Baltimore, MD by infusion of a fecal solution prepared by a predefined protocol (Dutta et al., submitted) based on Aas et al. [38]. Potential donors were thoroughly clinically evaluated based on history, physical examination and serological screening for HIV, syphilis, hepatitis A, B and C and Helicobacter pylori infection. Fecal specimens of patients and donors were tested 3–5 days before FMT for the presence of pathogenic bacteria (salmonella, shigella, yersinia), parasites (entamoeba, giardia, worms), and C. difficile. Patients were admitted to the hospital the day before and bowel prep administered the night before FMT. Patients were also administered a proton pump inhibitor (omeprazole, 20 mg) on the evening and morning before the procedure. Donor fecal samples (25–30 g) were mixed with 250 ml of sterile saline buffer, mixed into slurry and filtered once with surgical gauze for large particles and twice with a coffee filter. The volume of the filtrate was increased to 450 ml with sterile saline buffer and divided into 5 aliquots of 90 ml. For FMT, two aliquots (180 ml) were endoscopically delivered by spray catheter into the jejunum. The remaining three aliquots were instilled by colonoscopy into the right colon (180 ml) and transverse and upper descending colon (90 ml).

The clinical aspects of this study, including a comprehensive description and discussion of the FMT-treated patient population and individual case metadata, are provided in a separate publication (Dutta et al., submitted). Fecal samples were collected from 14 patient-donor pairs and used for this study (Fig. 1; Table 1). All patients had at least three recurrences of C. difficile infection and were treated with at least three courses of antibiotics. Fecal samples were collected before and after FMT from patients and, at corresponding time points, from their respective donors, which included family members (spouses and children) and friends (Fig. 1).

thumbnail

Figure 1. Overview of analyzed patient and donor samples.

RCDI patient samples are marked in red, post-FMT patient samples in blue and donor samples in green. *Patient #6a experienced antibiotic-induced relapse of C. difficile infection and was treated successfully with a second round of FMT as patient #6b. In the NCBI short read archive, samples referred to as #6b are designated as #7 samples.

doi:10.1371/journal.pone.0081330.g001
thumbnail

Table 1. RCDI patient study population.

doi:10.1371/journal.pone.0081330.t001

Sample collection and nucleic acid isolation

All fecal samples were self-collected by patients and donors without bowel preps, stored in the freezer and within 24 hours brought to Sinai Hospital, after which they were stored at –80°C. Patients stopped antibiotic use 5 days before the FMT procedure; RCDI patient samples were taken 1–2 days prior to FMT. For processing, samples were thawed at 4°C and in aliquots of 0.15 g per tube re-suspended in 1 ml of 1 × phosphate-buffered saline. Cell lysis was initiated with two enzymatic incubations, first using 5 µl of lysozyme (10 mg ml−1; Amresco, Solon, OH, USA), 13 µl of mutanolysin (11.7 U µl−1; Sigma-Aldrich) and 3 µl of lysostaphin (4.5 U µl−1; Sigma-Aldrich) for an incubation of 30 min at 37°C and, second, using 10 µl Proteinase K (20 mg ml−1; Research Products International, Mt Prospect, IL, USA), 50 µl 10% SDS and 2 µl RNase (10 mg ml−1) for an incubation of 45 min at 56°C. After the enzyme treatments, cells were disrupted by bead beating in tubes with Lysing Matrix B (0.1 mm silica spheres, MP Biomedicals, Solon, OH, USA), at 6 m s−1 for 40 s at room temperature in a FastPrep-24 (MP Biomedicals). The resulting crude lysate was processed using the ZR Fecal DNA mini-prep kit (Zymo, Irvine, CA, USA) according to the manufacturer’s recommendation. The samples were eluted with 100 µl of ultra pure water into separate tubes. DNA concentrations in the samples were measured using the Quant-iT PicoGreen dsDNA assay kit (Molecular Probes, Invitrogen, Carlsbad, CA, USA).

Amplification and sequencing

In brief, hypervariable regions V1–V3 of the bacterial 16S rRNA gene were amplified with primers 27F and 534R as described previously [39]. DNA amplification of 16S rRNA genes was performed using AccuPrime Taq DNA polymerase High Fidelity (Invitrogen) and 50 ng of template DNA in a total reaction volume of 25 µl, following the AccuPrime product protocol. Reactions were run in a PTC-100 thermal controller (MJ Research, Waltham, MA, USA) using the following protocol: 3 min of denaturation at 94°C, followed by 30 cycles of 30 s at 94°C (denaturation), 30 s at 52°C (annealing) and 45 ss at 68°C (elongation), with a final extension at 68°C for 5 min.

Equimolar amounts (50 ng) of the PCR amplicons were mixed in a single tube. Amplification primers and reaction buffer were removed using the AMPure Kit (Beckman Coulter, Brea, CA, USA) and purified amplicon mixtures sequenced at the Institute for Genome Sciences, University of Maryland, using 454 primer A and protocols recommended by the manufacturer (Roche, Branford, CT, USA). Raw sequences were deposited in the Short Read Archive Database (http://www.ncbi.nlm.nih.gov/sra; project number SRP016902). In the NCBI short read archive, samples referred to as #6a are designated as #6 samples and samples referred to as #6b as #7 samples.

Sequence processing and analysis

16S rRNA sequence reads were processed with QIIME [40] and CloVR [41], using the automated CloVR-16S pipeline as described in the corresponding standard operating procedure [42]. Briefly, using the QIIME split_libraries.py tool sequences were binned based on sample-specific barcodes, trimmed by removal of barcode and primer sequences and filtered for quality, using the default parameters, except for "—barcode-type "variable_length". Chimeric sequences were removed with UCHIME [43] using MicrobiomeUtilities (http://microbiomeutil.sourceforge.net/) and the rRNA16S.gold.fasta reference database. Reads were clustered into operational taxonomic units (OTUs) using a similarity threshold of 95%. On average, OTUs were classified using the RDP Naive Bayesian Classifier [44] with a score filtering threshold of 0.5. Rarefaction curves were calculated based on OTU counts using the rarefaction.single routine of the Mothur package [45]. Hierarchical clustering, boxplots, and statistical calculations (Wilcoxon rank sum tests, Jensen-Shannon divergence etc.) were performed in R. Differentially abundant OTUs were determined with Metastats [46]. Phylogenetic trees were created with FastTree2 [47] using trimmed alignments generated with NAST. Dot plots to evaluate phylogenetic distances and Jensen-Shannon divergence between sample pairs and changes in relative abundance of specific taxonomic families over time were generated with Prism5 (version 6 for Mac, GraphPad Software, San Diego CA, USA).

Results and Discussion

Patient population, sample set and sequence data

For this longitudinal study, fecal samples were collected from 14 pairs of RCDI patients, treated successfully by FMT, and their respective donors (Fig. 1). In addition to the 14 donor samples used for FMT, 11 samples from pre-FMT RCDI patients and 17 samples from eight post-FMT patient samples, as well as 14 samples from eight healthy donors collected after FMT were analyzed, collected between one week and one year after the procedure, (total number of samples: 56). This allowed us to perform the first characterization of long-term microbiota changes in patients after FMT. All treated RCDI patients experienced resolution of diarrheal symptoms within 2–3 days after FMT (Table 1), in accordance with previous reports [27]. Of the post-FMT samples collected from asymptomatic patients, 14 were paired with donor samples collected at the same time points to serve as a control for intra-individual, longitudinal variations not associated with RCDI. RCDI patient #6a was successfully treated by FMT but experienced recurrence of C. difficile infection one month later, after being treated for a urinary tract infection with ciprofloxacin. Subsequent oral vancomycin and intravenous immunoglobulin therapy did not resolve the problem. The patient #6a was treated successfully for a second time by FMT, three months after the first FMT (designated as case #6b). Selected characteristics of all cases for which samples were analyzed are summarized in Table 1. Additional clinical aspects of this study have been described in a separate publication [48] FMT donors for this study were chosen by the RCDI patients and included genetically unrelated individuals living in the same household (8x spouses), as well as genetically related (2x children) or unrelated (3x friends) individuals living in households separate from those of the RCDI patients (Table 1). On average, 3,315 sequence reads were obtained per sample using the Roche/454 GS FLX Titanium platform (average sequence length: 527 bp). A list of read numbers and identified operational taxonomic units (OTUs) for each of the samples is part of the supplement (Table S1).

Reduced microbiota diversity in RCDI patients increases after FMT

Reduced microbiota diversity associated with C. difficile infection is reported in humans [49]-[51] and mice [52], [53]. This finding was confirmed in our study with multiple post-FMT samples collected up to one year after the procedure. Compared to healthy donors the fecal microbiota diversity of RCDI patients was reduced, as shown by rarefaction analysis of OTU counts (Fig. 2). Microbiota diversity increased significantly in post-FMT patient samples, as demonstrated by Shannon diversity index calculations (p<0.01, Wilcoxon rank sum test) between RCDI (mean 1.68± 0.75) and post-FMT (mean 3.37± 0.46) patient samples (Fig. 3). Microbial richness was also increased in post-FMT compared to RCDI patient samples, based on the comparison of mean ACE indices (46%; p < 0.001). Interestingly, no significant difference in microbial diversity or richness was noted between post-FMT patient and donor samples as determined by Shannon and ACE indices. Shannon diversity increased in all 17 post-FMT patients as soon as one week after FMT and remained stable and comparable among different patients for up to one year afterwards (Fig. S1).

thumbnail

Figure 2. Microbiota rarefaction curves showing fecal microbiota diversity in RCDI (red) and post-FMT (blue) patient and donor (green) samples.

Each curve shows the average number of OTUs found in a given number of sampled sequences. OTUs can be treated as equivalent to taxonomic species in the sequence space. RCDI samples are marked from patient #6a (*), who experienced antibiotic-induced relapse and was treated by FMT again as patient #6b (**).

doi:10.1371/journal.pone.0081330.g002
thumbnail

Figure 3. Microbiota diversity (Shannon) and richness (ACE) of RCDI and post-FMT patient and donor samples.

(A) Shannon index; (B) ACE index. Significant differences are shown (*, p<0.01; **, p<0.001) as measured by Wilcoxon rank sum test. RCDI samples from patient #6a (+), who experienced antibiotic-induced relapse and was treated by FMT again as patient #6b (++) are marked.

doi:10.1371/journal.pone.0081330.g003

Compared to the RCDI sample collected before the first FMT treatment (#6a_P0), microbial diversity in the RCDI sample from the same patient collected three months later after RCDI relapse (#6b_P0) showed a 2-fold increase based on the Shannon index but was still low compared to healthy donor samples (Fig. 3). These results suggest that FMT restores the reduced microbiota diversity associated with RCDI. Furthermore, diversity increases immediately after FMT and remains stable over time.

FMT shifts fecal microbiota towards healthy donor composition

To gain further insights into the effects of FMT on the patient microbiota, shared OTUs between RCDI patients, post-FMT patients and healthy donor samples were determined (Fig. S2). Using a threshold of at least five supporting reads across all 38 samples for OTUs to be considered in the comparison, a total of 1,321 OTUs were identified of which 876 (65%) were only identified in post-FMT patient and healthy donor samples but never in RCDI patient samples. This finding could be interpreted to indicate that post-FMT patients acquired donor OTUs as a consequence of FMT. However, the applied analysis has a detection limit of approximately 0.03% and does not allow for the distinction of different bacterial strains from the same OTU. It is therefore impossible to distinguish between OTUs that might have been present in RCDI patients below the detection limit and those that were acquired from the donors.

Microbiota compositions were analyzed based on phylogenetic distance calculations between samples using the unweighted, i.e., comparing OTU presences/absences, and weighted, i.e., including quantitative information about detected OTUs, UniFrac metric (Fig. 4). Principal coordinate analyses (PCoA) of the unweighted UniFrac comparison showed that most of the compositional variation among samples is accounted for by post-FMT patient and healthy donor samples (Fig. 4A). In contrast, when OTU abundance is also taken into consideration (weighted UniFrac analysis) most of the variation within the entire sample set is observed among RCDI patient samples (Fig. 4B), suggesting that relative abundances of major microbiota members can vary substantially not only between RCDI patient and healthy donor samples but also among different RCDI patient samples.

thumbnail

Figure 4. Unscaled principal coordinate analysis (PCoA) plots showing unweighted (A) and weighted (B) UniFrac analysis of RCDI (red) and post-FMT (blue) patient and healthy donor (green) samples.

RCDI patient samples are circled in red. RCDI samples from patient #6a (*), who experienced antibiotic-induced relapse and was treated by FMT again as patient #6b (**) are marked in dark red. Sample names indicate case numbers, patient or donor source and time point of collection ("0" time point refers to pre-FMT sampling time points; other time points are abbreviated as weeks [w], months [m] and year [y]).

doi:10.1371/journal.pone.0081330.g004

In most cases, FMT resulted in the adoption of a fecal microbiota composition in post-FMT samples that was similar to that of healthy donors. This is apparent in the clustering of post-FMT patient and healthy donor samples in unweighted UniFrac analysis (Fig. 4A). However, several patients appeared to at least temporarily return to pre-FMT fecal microbiota composition states (e.g., Patient #8 at 5 months and Patient #14 at 3 weeks after FMT), although all treated patients were reported to be symptom-free within 2–3 days after FMT. The adoption of a fecal microbiota composition in post-FMT patient samples similar to that of healthy donors was also supported by comparing mean phylogenetic UniFrac distances. These were significantly larger between RCDI and post-FMT patient samples than between post-FMT and donor samples both in unweighted (p<0.05) and weighted (p<0.01) UniFrac analysis.

Interestingly, the RCDI sample from the patient (#6a/b), who relapsed after unrelated antibiotic treatment, showed a microbiota composition that was similar to that of other post-FMT and healthy donor samples, especially in the weighted UniFrac analysis (Fig. 4). This second RCDI episode lasted only two months and included treatment with a single antibiotic (vancomycin) compared to 4.5–72 months duration and at least three different antibiotic treatments in other RCDI patients, It is therefore possible that several of the phenotypes observed in other RCDI samples are reflective of long-term disease and multiple antibiotic treatment courses. The data presented here suggest that RCDI is associated with the presence or absence of specific fecal microbiota members (i.e., co-clustering of all RCDI samples in unweighted UniFrac analysis, including #6b_P0), rather than significant changes in the relative abundance of major microbiome components (i.e., separate clustering of different RCDI samples and of #6b_P0 with healthy donor samples in weighted UniFrac analysis), which could represent a consequence of long-term disease.

FMT affects predominantly Firmicutes and Proteobacteria

The identification of specific microbiota members associated with RCDI and successful FMT treatment bears the potential to identify new diagnostic markers to predict susceptibility to C. difficile infection or infection relapse in at-risk populations. In addition, this knowledge may provide the insights required to assemble culture-based "probiotic" bacterial mixtures as substitutes for transplantation of fecal samples, as has recently been demonstrated in humans [54] and the mouse model [55]. Towards this goal, the relative abundances of all identified microbial taxa were compared between RCDI and post-FMT patient and healthy donor sample groups using Metastats [46]. Among these three groups, bacteria from only three taxonomic orders, belonging to two phyla, showed significant changes, i.e., Clostridiales and Lactobacillales (both from phylum Firmicutes) and Enterobacteriales (phylum Proteobacteria) (Fig. 5). Clostridiales, which include the species C. difficile, were present at only 12.8% in RCDI patient samples and significantly increased in post-FMT samples (55%) but still remained lower compared to healthy donor samples (70%) (p<0.001, unpaired t-test with unequal variance). Lactobacillales, which were present at high abundance in RCDI patient samples (mean: 58%), were significantly decreased in post-FMT patient (22%) and healthy donor (5%) samples. However, abundance of Lactobacillales remained higher in post-FMT patient compared to donor samples (p<0.01). Enterobacteriales, present at 6.5% in RCDI patient samples, were less than 1% in post-FMT patient and donor samples (p<0.001).

thumbnail

Figure 5. Microbiota changes between RCDI and post-FMT patient and healthy donor sample groups at the taxonomic order level.

Significant differences between sample groups as calculated with the Metastats tool are marked with asterisks (p<0.01).

doi:10.1371/journal.pone.0081330.g005

Three taxonomic families within the order Clostridiales (phylum: Firmicutes) significantly increased in relative abundance between RCDI and post-FMT patient samples (p<0.01), Lachnospiraceae, Peptostreptococcaceae, and Ruminococcaceae (Fig. 6). Most prominently, an uncharacterized genus within the Lachnospiraceae family (Lachnospiraceae Incertae Sedis) increased from on average 3% in RCDI patient samples to 30% in post-FMT patient samples and was 39% in healthy donor samples (p<0.01). The dominant OTU within this genus (99% identical to GenBank Acc.-No.: EF399262) was identified in all 28 donor samples (27 samples with >4 reads), 15 out of 17 post-FMT patient samples (14 samples with >4 reads), and 8 out of 11 RCDI patient samples (#6b was the only sample with >4 reads). C. difficile is a member of the Peptostreptococcaceae [56], which increased in patients after FMT. Moreover, an unknown genus within this family accounts for >2% of the fecal microbiota in healthy donor samples (Fig. 6), demonstrating that taxonomically close relatives of C. difficile exert non-pathogenic or even beneficial functions in the healthy intestinal microbiota.

thumbnail

Figure 6. Microbiota changes between RCDI and post-FMT patient and healthy donor sample groups at the taxonomic family and genus levels.

Significant differences between sample groups as calculated with the Metastats tool are marked with asterisks (p<0.01). Note that standard deviations are smaller for genera that increased in post-FMT relative to RCDI patient samples (e.g., Lachnospiraceae Incertae Sedis) compared to those that decreased (e.g. Streptococcus), which reflects differences in the relative abundances of major microbiota members among RCDI patient samples.

doi:10.1371/journal.pone.0081330.g006

Within the orders Lactobacillales (phylum: Firmicutes) and Enterobacteriales (phylum: Proteobacteria), the genera Enterococcus and Klebsiella, which were present on average at 18% and 4% in RCDI patient samples, respectively, were significantly reduced to less than 0.1% in post-FMT patient samples (p<0.01). Members of the Streptococcaceae (phylum: Firmicutes), the dominant taxonomic family in RCDI patient samples (mean: 30.1%), were reduced on average by more than 10% after FMT, although this change was not statistically significant due to large variations between RCDI patients. With the exception of the genus Streptococcus, none of these families or genera showed significant differences in relative abundance between post-FMT patient and healthy donor samples (p<0.05). Streptococcus was the only genus with a significant difference in relative abundance between both RCDI patient and donor samples and between post-FMT patient and donor samples. As post-FMT patients appear to show increased susceptibility to C. difficile infection compared to healthy donors, if additional antibiotic medication to treat unrelated infections becomes necessary [27], the increased abundance of the Streptococcus genus in this population could play a role for this susceptibility. However, not all RCDI samples contained high counts of Streptococcus sequences (range: 0.1% to 82.4%). In general, different RCDI samples showed more variation in the abundance of microbiota members that were increased relative to healthy donors (e.g., Enterococcaceae and Streptococcaceae) than of microbiota members that were reduced (see error bars in Fig. 6). This may suggest that the second group provides a better target for the identification of diagnostic markers for RCDI (e.g., among the Lachnospiraceae, Peptostreptococcaceae, and Ruminococcaceae).

In contrast to all other cases, the fecal RCDI microbiota from patient #6b, who experienced antibiotic-induced relapse of C. difficile infection, contained large fractions of Lachnospiraceae (11% compared to no detection before the first FMT and on average 1% in other RCDI samples) and Akkermansia (60% compared to on average 0.1% in other RCDI samples and 1.8% in healthy donor samples) (Fig. S3). This atypical composition could be responsible for the clustering of this sample with healthy donor and post-FMT patient samples in the weighted UniFrac analysis (Fig. 4B). It is therefore possible that the reductions in Lachnospiraceae characteristic of the other RCDI samples, rather than being a cause of disease susceptibility, represent an effect of disease duration and number of antibiotic treatment regimens exceeding those that patient #6b experienced after recurrence. Interestingly, Akkermansia spp. have recently received special attention in human microbiome research because of their ability to colonize the intestinal mucosa and to utilize mucus as a sole carbon and nitrogen source [57], [58]. While A. municiphila has been proposed as a marker of a healthy intestine, due to its production of short chain fatty acids and its negative correlation with inflammatory bowel diseases, appendicitis and obesity (reviewed here:[58]), its high abundance in the fecal sample of patient #6b might also be an indicator of high concentrations of mucus in the stool, which could be the result of acute diarrhea.

The fecal microbiota continues to change in asymptomatic post-FMT patients

Asymptomatic post-FMT patients appear to be at higher risk for recurrence of C. difficile infection compared to patients without a history of RCDI, if additional antibiotic medication to treat unrelated infections becomes necessary [27]. Whether specific microbiota features, such as the increased abundance of Streptococcus in post-FMT patient compared to healthy donor samples, are responsible for this susceptibility is unknown, but the susceptibility of post-FMT patients to RCDI may decrease over time and little is known about the long-term dynamics of FMT-induced microbiota changes. In order to characterize microbiota changes after FMT over time, fecal samples from post-FMT patients, all of which were asymptomatic with respect to RCDI, were compared longitudinally. Microbiota diversity in post-FMT patient samples did not change significantly over time, as measured by comparing the Shannon diversity index (Fig. S1). To study changes in microbiota composition over time, weighted and unweighted UniFrac distances and the Jensen-Shannon divergence were calculated between (i) RCDI and post-FMT patient sample pairs, (ii) donor and post-FMT patient samples pairs and, as a control for temporal variations in healthy individuals, between (iii) sample pairs collected from the same donor before and after FMT (Fig. 7). For the comparison of post-FMT and RCDI patient samples, both unweighted UniFrac and Jensen-Shannon distance metrics displayed a significant linear change over time when plotted on a logarithmic scale. However, comparison of post-FMT patient and donor samples or of donor samples collected before and after FMT did not. That this correlation is only apparent if temporal changes are plotted on a logarithmic scale shows that the most significant changes happen immediately after FMT and that the microbiota continues to evolve over time albeit at a decreasing rate.

thumbnail

Figure 7. Post-FMT microbiota changes.

Unweighted (A) and weighted (B) UniFrac distances and Jensen-Shannon divergence (C) metrics were calculated between post-FMT and RCDI patient sample pairs (red), post-FMT patient and donor sample pairs (green) and between donor sample pairs collected over time (blue) and plotted on logarithmic scales. R2 values and p-values to establish whether the slope of the curve was significantly different from zero are shown with asterisks indicating significance (p<0.05, F-test). The 20-week data point of patient #8 was classified as outlier and not included in the analyses, based on the Bonferroni-adjusted outlier test, and is shown with parentheses. One-year time points (patient and donor #1) were also classified as outliers and omitted from the analysis and plot. A plot showing all data points including those omitted is part of the supplement (Fig. S4).

doi:10.1371/journal.pone.0081330.g007

Individual taxonomic families showed similar trends in post-FMT patients over time, if compared case-by-case, i.e. increases in Lachnospiraceae and Ruminococcaceae and decreases in Streptococcaceae (Fig. 8). However, in contrast to changes in relative abundance between the pre- and post-FMT patient microbiota (Fig. 6), changes in post-FMT patients over time were not significant for the three studied Firmicutes families. This suggests that, while changes in the abundance of Lachnospiraceae and/or Streptococcaceae might play important roles for RCDI or successful recovery after FMT in some patients, general post-FMT microbiota dynamics across the entire patient population are better described using metrics that take account of the microbiota as a whole, i.e., UniFrac distances and Jensen-Shannon divergence.

thumbnail

Figure 8. Post-FMT changes in selected microbiota members by case (genus level).

(A) Lachnospira Incertae Sedis; (B) Ruminococcus; (C) Streptococcus. Genus-specific changes in relative abundance over time were not significant (p>0.05)when samples were grouped by time periods (1 week, 2–4 weeks, 6–8 weeks, 12–20 weeks) and groups compared with a non-parametric statistical test (Wilcoxon rank sum test).

doi:10.1371/journal.pone.0081330.g008

‘Keystone’ species are not identified in RCDI or FMT

The concept of keystone species has been used to describe the disproportionate importance of a single or a few organisms for the structure or function of an entire environment [59], [60], e.g. in the oral cavity where colonization with the commensal bacterium Porphyromonas gingivalis even at low abundance can play a major role for microbiota changes associated with periodontitis [61]. In the context of RCDI and FMT, keystone bacteria could be crucial for the identification of diagnostic markers to predict susceptibility to C. difficile infection and as substitutes for fecal samples of largely unknown composition to be used in transplantation. That RCDI can principally be treated by transplantation of in vitro-assembled microbial communities instead of fecal material was shown recently in humans [54] and mice [52], although little justification was provided for the selection of specific bacterial species or strains. While, based on our findings and previous data, members of the Lachnospiraceae family, for example, might present themselves as keystone candidates [50], [62], [63], at least one case was found in our cohort where RCDI was associated with relatively high counts of Lachnospiraceae (i.e., #6b). In another case (#9), Lachnospiraceae did only increase temporarily six weeks after FMT but dropped to pre-FMT levels 12 weeks after FMT. Khoruts et al. found a relatively large proportion (>5%) of Lachnospiraceae Inc. Sed. in an RCDI sample before FMT treatment [25]. Interestingly, the dominant representative of the genus Lachnospiraceae Inc. Sed. associated with successful FMT treatment, which was identified in the Canadian study by Shahinas et al. [50], is different from the one identified here (Shahinas: 97% identical to GenBank Acc.-No. JX230866, compared to this study: 99% identical to EF399262). This difference could either result from variations in the applied pyrosequencing protocols (e.g., Shahinas et al. used primers specific for hypervariable regions V5–V6 instead of primers specific for V1–V3 used here) or indicate that different species or strains of the genus Lachnospiraceae Inc. Sed. circulate in U.S. and Canadian human populations. In any case, it seems as if neither RCDI nor FMT are associated with the presence or absence of a single specific microbiota fraction.

Instead of bacterial keystone taxa, specific microbial microbiota genes or transcripts could be associated with health and disease and, thus, serve as "keystone functions" with potential as diagnostic markers. A redundancy and similarity of functional microbiota compositions between individuals despite significant taxonomic variation has previously been demonstrated for the healthy human microbiota [64]. These functions could be predominantly but not exclusively associated with certain members of the fecal microbiota, which would then still show statistical correlations with health and disease states. Short-chain fatty acid (SCFA) production plays an important role in the regulation of intestinal inflammatory processes [65] and intestinal barrier maintenance [66][68] and has been discussed in the context of RCDI, as C. difficile infection in the mouse model was shown to alter SCFA profiles [52]. Consequently, the reduction of Lachnospiraceae and Ruminococcaceae has been interpreted as a depletion in butyrate-producing bacteria [51]. Shotgun sequencing of total metagenomic DNA and/or metatranscriptomic RNA isolates will be needed to confirm the lack of butyrate production in the fecal RCDI microbiota or to associated other "keystone functions" with RCDI and FMT.

Concomitant effects of antibiotics and diarrhea

Previous RCDI microbiota studies have had difficulty determining the chain of events leading to disease as well as the relationship between observed microbiota phenotypes and disease. C. difficile infection is typically initiated by antibiotic treatment and phenotypically characterized by severe diarrhea. Both events by themselves have a massive impact on the fecal microbiota independent of the disease caused by the C. difficile infection [69], [70]. It is therefore difficult to distinguish between microbiota changes that play a causative role in RCDI and those that simply co-occur. The data presented here also include an RCDI patient with successful FMT and subsequent relapse of CDI after antibiotic treatment, whose fecal microbiota showed characteristics described for healthy individuals as opposed to RCDI patients (e.g. relatively high Lachnospiraceae abundance). This single patient may therefore suggest that multiple rounds of antibiotic treatment and/or long-term duration of the disease are needed to induce some of the microbiota changes previously reported to be associated with CDI. In order to determine the exact time line of events, prospective studies are needed starting before antibiotic treatment and following patients during the onset and course of CDI.

Conclusion

In accordance with previous reports, we found a reduction in microbiota diversity and richness in fecal samples from RCDI patients compared to healthy donors, which was restored after FMT. Similarly, our results confirm previous findings that FMT changes the RCDI fecal microbiota to become more similar to that of healthy donors. We extend current knowledge by demonstrating that there are different varieties of dysbiosis in RCDI patient samples, that FMT predominantly affects Firmicutes and Proteobacteria, and that the fecal microbiota continues to change in post-FMT patients. We did not identify a ‘keystone’ species in RCDI or FMT, but our findings suggest that butyrate producing bacteria may be important. We believe that additional longitudinal studies, ideally beginning before initial infection and including metagenomic and metatranscriptomic analyses, will lead to improved outcomes in C. difficile infection.

Supporting Information

Figure S1.

Fecal microbiota diversity in patient and donor samples depending on collection time points. The Shannon index of all samples is plotted over time, split into donor (A, blue) and patient (B, red) samples.

doi:10.1371/journal.pone.0081330.s001

(PDF)

Figure S2.

Venn diagram showing shared OTUs between RCDI and post-FMT patient and donor samples. Only OTUs represented by at least 5 reads across all 56 samples are shown.

doi:10.1371/journal.pone.0081330.s002

(PDF)

Figure S3.

Microbiota changes between RCDI samples collected from the same patient before the first FMT (#6a) and, after antibiotic-induced relapse, before the second FMT (#6b). Relative abundances of all taxonomic genera (>1%) are shown.

doi:10.1371/journal.pone.0081330.s003

(PDF)

Figure S4.

Post-FMT microbiota changes. Unweighted (A) and weighted (B) UniFrac distances and Jensen-Shannon divergence (C) metrics were calculated between post-FMT and RCDI patient sample pairs (red), post-FMT patient and donor sample pairs (green) and between donor sample pairs collected over time (blue). This figures shows that both patient and donor samples from case #1 collected one year aft FMT show an unusual small divergence (Unweighted/weighted UniFrac distances and Jensen-Shannon divergence) from the donor sample collected before FMT.

doi:10.1371/journal.pone.0081330.s004

(PDF)

Table S1.

Numbers of reads and identified operational taxonomic units (OTUs) by sample.

doi:10.1371/journal.pone.0081330.s005

(XLSX)

Acknowledgments

We are grateful for generous support from the Weinberg Foundation, the Friedman and Friedman Group and Eric Cowan.

Author Contributions

Conceived and designed the experiments: YS SD WFF. Performed the experiments: YS SG MG CM AD SD. Analyzed the data: YS WFF. Contributed reagents/materials/analysis tools: SD WFF. Wrote the paper: YS ECvR SD WFF.

References

  1. 1. Bartlett JG (2002) Clinical practice. Antibiotic-associated diarrhea. N Engl J Med 346: 334–339. doi: 10.1056/nejmcp011603
  2. 2. Jobe BA, Grasley A, Deveney KE, Deveney CW, Sheppard BC (1995) Clostridium difficile colitis: an increasing hospital-acquired illness. Am J Surg 169: 480–483. doi: 10.1016/s0002-9610(99)80199-8
  3. 3. Miller BA, Chen LF, Sexton DJ, Anderson DJ (2011) Comparison of the burdens of hospital-onset, healthcare facility-associated Clostridium difficile Infection and of healthcare-associated infection due to methicillin-resistant Staphylococcus aureus in community hospitals. Infect Control Hosp Epidemiol 32: 387–390. doi: 10.1086/659156
  4. 4. Rivera EV, Woods S (2003) Prevalence of asymptomatic Clostridium difficile colonization in a nursing home population: a cross-sectional study. J Gend Specif Med 6: 27–30.
  5. 5. Jangi S, Lamont JT (2010) Asymptomatic colonization by Clostridium difficile in infants: implications for disease in later life. J Pediatr Gastroenterol Nutr 51: 2–7. doi: 10.1097/mpg.0b013e3181d29767
  6. 6. McFarland LV, Elmer GW, Surawicz CM (2002) Breaking the cycle: treatment strategies for 163 cases of recurrent Clostridium difficile disease. Am J Gastroenterol 97: 1769–1775.
  7. 7. Pepin J, Routhier S, Gagnon S, Brazeau I (2006) Management and outcomes of a first recurrence of Clostridium difficile-associated disease in Quebec, Canada. Clin Infect Dis 42: 758–764. doi: 10.1086/501126
  8. 8. Surawicz CM, McFarland LV, Greenberg RN, Rubin M, Fekety R, et al. (2000) The search for a better treatment for recurrent Clostridium difficile disease: use of high-dose vancomycin combined with Saccharomyces boulardii. Clin Infect Dis 31: 1012–1017. doi: 10.1086/318130
  9. 9. Musher DM, Logan N, Mehendiratta V, Melgarejo NA, Garud S, et al. (2007) Clostridium difficile colitis that fails conventional metronidazole therapy: response to nitazoxanide. J Antimicrob Chemother 59: 705–710. doi: 10.1093/jac/dkl553
  10. 10. Buggy BP, Fekety R, Silva J Jr (1987) Therapy of relapsing Clostridium difficile-associated diarrhea and colitis with the combination of vancomycin and rifampin. J Clin Gastroenterol 9: 155–159. doi: 10.1097/00004836-198704000-00009
  11. 11. Berman AL (2007) Efficacy of rifaximin and vancomycin combination therapy in a patient with refractory Clostridium difficile-associated diarrhea. J Clin Gastroenterol 41: 932–933. doi: 10.1097/01.mcg.0000225685.37465.e7
  12. 12. Johnson S, Schriever C, Galang M, Kelly CP, Gerding DN (2007) Interruption of recurrent Clostridium difficile-associated diarrhea episodes by serial therapy with vancomycin and rifaximin. Clin Infect Dis 44: 846–848. doi: 10.1086/511870
  13. 13. Garey KW, Jiang ZD, Bellard A, Dupont HL (2009) Rifaximin in treatment of recurrent Clostridium difficile-associated diarrhea: an uncontrolled pilot study. J Clin Gastroenterol 43: 91–93. doi: 10.1097/mcg.0b013e31814a4e97
  14. 14. Johnson S, Schriever C, Patel U, Patel T, Hecht DW, et al. (2009) Rifaximin Redux: treatment of recurrent Clostridium difficile infections with rifaximin immediately post-vancomycin treatment. Anaerobe 15: 290–291.
  15. 15. Navaneethan U, Venkatesh PG, Shen B (2010) Clostridium difficile infection and inflammatory bowel disease: understanding the evolving relationship. World J Gastroenterol 16: 4892–4904. doi: 10.3748/wjg.v16.i39.4892
  16. 16. Forster AJ, Taljaard M, Oake N, Wilson K, Roth V, et al. (2012) The effect of hospital-acquired infection with Clostridium difficile on length of stay in hospital. CMAJ 184: 37–42. doi: 10.1503/cmaj.110543
  17. 17. Daneman N, Stukel TA, Ma X, Vermeulen M, Guttmann A (2012) Reduction in Clostridium difficile infection rates after mandatory hospital public reporting: findings from a longitudinal cohort study in Canada. PLoS Med 9: e1001268. doi: 10.1371/journal.pmed.1001268
  18. 18. O'Keefe SJ (2010) Tube feeding, the microbiota, and Clostridium difficile infection. World J Gastroenterol 16: 139–142. doi: 10.3748/wjg.v16.i2.139
  19. 19. Pohl JF (2012) Clostridium difficile infection and proton pump inhibitors. Curr Opin Pediatr 24: 627–631. doi: 10.1097/mop.0b013e328355a3e1
  20. 20. Sartor RB (2008) Microbial influences in inflammatory bowel diseases. Gastroenterology 134: 577–594. doi: 10.1053/j.gastro.2007.11.059
  21. 21. Kelly CP, LaMont JT (2008) Clostridium difficile—more difficult than ever. N Engl J Med 359: 1932–1940. doi: 10.1056/nejmra0707500
  22. 22. Maroo S, Lamont JT (2006) Recurrent clostridium difficile. Gastroenterology 130: 1311–1316. doi: 10.1053/j.gastro.2006.02.044
  23. 23. McDonald LC, Killgore GE, Thompson A, Owens RC Jr, Kazakova SV, et al. (2005) An epidemic, toxin gene-variant strain of Clostridium difficile. N Engl J Med 353: 2433–2441. doi: 10.1056/nejmoa051590
  24. 24. Lund-Tonnesen S, Berstad A, Schreiner A, Midtvedt T (1998) [Clostridium difficile-associated diarrhea treated with homologous feces]. Tidsskr Nor Laegeforen 118: 1027–1030.
  25. 25. Khoruts A, Dicksved J, Jansson JK, Sadowsky MJ (2010) Changes in the composition of the human fecal microbiome after bacteriotherapy for recurrent Clostridium difficile-associated diarrhea. J Clin Gastroenterol 44: 354–360. doi: 10.1097/mcg.0b013e3181c87e02
  26. 26. Wettstein A, Borody TJ, Leis S, Chongan J, Torres M, et al.. (2007) Faecal bacteriotherapy – an effective treatment for relapsing symptomatic Clostridium difficile infection. 15th Gut United European Gastroenterology Week (France) 39(Suppl 1): A303 Vienna: United European Gastroenterology Federation.
  27. 27. Brandt LJ, Aroniadis OC, Mellow M, Kanatzar A, Kelly C, et al. (2012) Long-term follow-up of colonoscopic fecal microbiota transplant for recurrent Clostridium difficile infection. Am J Gastroenterol 107: 1079–1087.
  28. 28. Gough E, Shaikh H, Manges AR (2011) Systematic review of intestinal microbiota transplantation (fecal bacteriotherapy) for recurrent Clostridium difficile infection. Clin Infect Dis 53: 994–1002. doi: 10.1093/cid/cir632
  29. 29. Bakken JS, Borody T, Brandt LJ, Brill JV, Demarco DC, et al. (2011) Treating Clostridium difficile infection with fecal microbiota transplantation. Clin Gastroenterol Hepatol 9: 1044–1049. doi: 10.1016/j.cgh.2011.08.014
  30. 30. Vrieze A, de Groot PF, Kootte RS, Knaapen M, van Nood E, et al. (2013) Fecal transplant: a safe and sustainable clinical therapy for restoring intestinal microbial balance in human disease? Best Pract Res Clin Gastroenterol 27: 127–137. doi: 10.1016/j.bpg.2013.03.003
  31. 31. Anderson JL, Edney RJ, Whelan K (2012) Systematic review: faecal microbiota transplantation in the management of inflammatory bowel disease. Aliment Pharmacol Ther 36: 503–516. doi: 10.1111/j.1365-2036.2012.05220.x
  32. 32. Musa S, Thomson S, Cowan M, Rahman T (2010) Clostridium difficile infection and inflammatory bowel disease. Scand J Gastroenterol 45: 261–272. doi: 10.3109/00365520903497098
  33. 33. Kump PK, Grochenig HP, Lackner S, Trajanoski S, Reicht G, et al. (2013) Alteration of Intestinal Dysbiosis by Fecal Microbiota Transplantation Does not Induce Remission in Patients with Chronic Active Ulcerative Colitis. Inflamm Bowel Dis 19: 2155–2165. doi: 10.1097/mib.0b013e31829ea325
  34. 34. Vrieze A, Van Nood E, Holleman F, Salojarvi J, Kootte RS, et al.. (2012) Transfer of intestinal microbiota from lean donors increases insulin sensitivity in individuals with metabolic syndrome. Gastroenterology 143: 913–916 e917.
  35. 35. Borody TJ, Paramsothy S, Agrawal G (2013) Fecal microbiota transplantation: indications, methods, evidence, and future directions. Curr Gastroenterol Rep 15: 337. doi: 10.1007/s11894-013-0337-1
  36. 36. Borody TJ, Khoruts A (2012) Fecal microbiota transplantation and emerging applications. Nat Rev Gastroenterol Hepatol 9: 88–96. doi: 10.1038/nrgastro.2011.244
  37. 37. U.S. Department of Health and Human Services FaDA, Center for Biologics Evaluation and Research (2013) Guidance for Industry: Enforcement Policy Regarding Investigational New Drug Requirements for Use of Fecal Microbiota for Transplantation to Treat Clostridium difficile Infection Not Responsive to Standard Therapies. 1–4.
  38. 38. Aas J, Gessert CE, Bakken JS (2003) Recurrent Clostridium difficile colitis: case series involving 18 patients treated with donor stool administered via a nasogastric tube. Clin Infect Dis 36: 580–585. doi: 10.1086/367657
  39. 39. Ravel J, Gajer P, Abdo Z, Schneider GM, Koenig SS, et al. (2011) Vaginal microbiome of reproductive-age women. Proc Natl Acad Sci U S A 108 Suppl 14680–4687. doi: 10.1073/pnas.1002611107
  40. 40. Caporaso JG, Kuczynski J, Stombaugh J, Bittinger K, Bushman FD, et al. (2010) QIIME allows analysis of high-throughput community sequencing data. Nat Methods 7: 335–336. doi: 10.1038/nmeth.f.303
  41. 41. Angiuoli SV, Matalka M, Gussman A, Galens K, Vangala M, et al. (2011) CloVR: a virtual machine for automated and portable sequence analysis from the desktop using cloud computing. BMC Bioinformatics 12: 356. doi: 10.1186/1471-2105-12-356
  42. 42. James White WFF, Cesar Arze, Malcolm Matalka, The CloVR Team, Owen White, et al. (2011) CloVR-16S: Phylogenetic microbial community composition analysis based on 16S ribosomal RNA amplicon sequencing – standard operating procedure, version1.0. Nature Precedings.
  43. 43. Edgar RC, Haas BJ, Clemente JC, Quince C, Knight R (2011) UCHIME improves sensitivity and speed of chimera detection. Bioinformatics 27: 2194–2200. doi: 10.1093/bioinformatics/btr381
  44. 44. Wang Q, Garrity GM, Tiedje JM, Cole JR (2007) Naive Bayesian classifier for rapid assignment of rRNA sequences into the new bacterial taxonomy. Appl Environ Microbiol 73: 5261–5267. doi: 10.1128/aem.00062-07
  45. 45. Schloss PD, Westcott SL, Ryabin T, Hall JR, Hartmann M, et al. (2009) Introducing mothur: open-source, platform-independent, community-supported software for describing and comparing microbial communities. Appl Environ Microbiol 75: 7537–7541. doi: 10.1128/aem.01541-09
  46. 46. White JR, Nagarajan N, Pop M (2009) Statistical methods for detecting differentially abundant features in clinical metagenomic samples. PLoS Comput Biol 5: e1000352. doi: 10.1371/journal.pcbi.1000352
  47. 47. Price MN, Dehal PS, Arkin AP (2010) FastTree 2—approximately maximum-likelihood trees for large alignments. PLoS One 5: e9490. doi: 10.1371/journal.pone.0009490
  48. 48. Dutta SK, Garg S, Dutta A, von Rosenvinge EC, Maddox C, et al.. (2013) Jejunal and Colonic Fecal Microbiota Transplantation for Recurrent Clostridium difficile Infection with genomic analysis. Gastroenterology ID: AJG-13-1181submitted.
  49. 49. Skraban J, Dzeroski S, Zenko B, Mongus D, Gangl S, et al. (2013) Gut microbiota patterns associated with colonization of different Clostridium difficile ribotypes. PLoS One 8: e58005. doi: 10.1371/journal.pone.0058005
  50. 50. Shahinas D, Silverman M, Sittler T, Chiu C, Kim P, et al.. (2012) Toward an understanding of changes in diversity associated with fecal microbiome transplantation based on 16S rRNA gene deep sequencing. MBio 3.
  51. 51. Antharam VC, Li E, Ishmael A, Sharma A, Mai V, et al.. (2013) Intestinal dysbiosis and depletion of butyrogenic bacteria in Clostridium difficile infection and nosocomial diarrhea. J Clin Microbiol.
  52. 52. Lawley TD, Clare S, Walker AW, Stares MD, Connor TR, et al. (2012) Targeted restoration of the intestinal microbiota with a simple, defined bacteriotherapy resolves relapsing Clostridium difficile disease in mice. PLoS Pathog 8: e1002995. doi: 10.1371/journal.ppat.1002995
  53. 53. Lawley TD, Clare S, Walker AW, Goulding D, Stabler RA, et al. (2009) Antibiotic treatment of clostridium difficile carrier mice triggers a supershedder state, spore-mediated transmission, and severe disease in immunocompromised hosts. Infect Immun 77: 3661–3669. doi: 10.1128/iai.00558-09
  54. 54. Petrof EO, Gloor GB, Vanner SJ, Weese SJ, Carter D, et al.. (2012) Stool substitute transplant therapy for the eradication of Clostridium difficile infection: ‘RePOOPulating’ the gut. Microbiome 1.
  55. 55. Lawley B, Tannock GW (2012) Nucleic acid-based methods to assess the composition and function of the bowel microbiota. Gastroenterol Clin North Am 41: 855–868. doi: 10.1016/j.gtc.2012.08.010
  56. 56. Ludwig W, Schleifer K-H, Whitman WB (2009) Revised road map to the phylum Firmicutes. In: Bergey's Manual of Systematic Bacteriology, ed., vol. 3 (The Firmicutes) (P. De Vos, G. Garrity, D. Jones, N.R. Krieg, W. Ludwig, F.A. Rainey, K.-H. Schleifer, and W.B. Whitman, eds.). Springer-Verlag, New York: 1–13.
  57. 57. Derrien M, Vaughan EE, Plugge CM, de Vos WM (2004) Akkermansia muciniphila gen. nov., sp. nov., a human intestinal mucin-degrading bacterium. Int J Syst Evol Microbiol 54: 1469–1476. doi: 10.1099/ijs.0.02873-0
  58. 58. Belzer C, de Vos WM (2012) Microbes inside– –from diversity to function: the case of Akkermansia. ISME J 6: 1449–1458. doi: 10.1038/ismej.2012.6
  59. 59. Power ME, Tilman D, Estes JA, Menge BA, Bond WJ, et al. (1996) Challenges in the quest for keystones. Bioscience 46: 609–620. doi: 10.2307/1312990
  60. 60. Ebenman B, Jonsson T (2005) Using community viability analysis to identify fragile systems and keystone species. Trends Ecol Evol 20: 568–575. doi: 10.1016/j.tree.2005.06.011
  61. 61. Hajishengallis G, Liang S, Payne MA, Hashim A, Jotwani R, et al. (2011) Low-abundance biofilm species orchestrates inflammatory periodontal disease through the commensal microbiota and complement. Cell Host Microbe 10: 497–506. doi: 10.1016/j.chom.2011.10.006
  62. 62. Hamilton MJ, Weingarden AR, Unno T, Khoruts A, Sadowsky MJ (2013) High-throughput DNA sequence analysis reveals stable engraftment of gut microbiota following transplantation of previously frozen fecal bacteria. Gut Microbes 4: 125–135. doi: 10.4161/gmic.23571
  63. 63. Reeves AE, Koenigsknecht MJ, Bergin IL, Young VB (2012) Suppression of Clostridium difficile in the gastrointestinal tracts of germfree mice inoculated with a murine isolate from the family Lachnospiraceae. Infect Immun 80: 3786–3794. doi: 10.1128/iai.00647-12
  64. 64. Human Microbiome Project C (2012) Structure, function and diversity of the healthy human microbiome. Nature 486: 207–214.
  65. 65. Maslowski KM, Vieira AT, Ng A, Kranich J, Sierro F, et al. (2009) Regulation of inflammatory responses by gut microbiota and chemoattractant receptor GPR43. Nature 461: 1282–1286. doi: 10.1038/nature08530
  66. 66. Roediger WE (1982) Utilization of nutrients by isolated epithelial cells of the rat colon. Gastroenterology 83: 424–429.
  67. 67. Koruda MJ, Rolandelli RH, Bliss DZ, Hastings J, Rombeau JL, et al. (1990) Parenteral nutrition supplemented with short-chain fatty acids: effect on the small-bowel mucosa in normal rats. Am J Clin Nutr 51: 685–689.
  68. 68. Peng L, He Z, Chen W, Holzman IR, Lin J (2007) Effects of butyrate on intestinal barrier function in a Caco-2 cell monolayer model of intestinal barrier. Pediatr Res 61: 37–41. doi: 10.1203/01.pdr.0000250014.92242.f3
  69. 69. Dethlefsen L, Relman DA (2011) Incomplete recovery and individualized responses of the human distal gut microbiota to repeated antibiotic perturbation. Proc Natl Acad Sci U S A 108 Suppl 14554–4561. doi: 10.1073/pnas.1000087107
  70. 70. Gorkiewicz G, Thallinger GG, Trajanoski S, Lackner S, Stocker G, et al. (2013) Alterations in the colonic microbiota in response to osmotic diarrhea. PLoS One 8: e55817. doi: 10.1371/journal.pone.0055817