Stool substitute transplant therapy for the eradication of Clostridium difficile infection: 'RePOOPulating' the gut

Abstract

Background: Fecal bacteriotherapy ('stool transplant') can be effective in treating recurrent Clostridium difficile infection, but concerns of donor infection transmission and patient acceptance limit its use. Here we describe the use of a stool substitute preparation, made from purified intestinal bacterial cultures derived from a single healthy donor, to treat recurrent C. difficile infection that had failed repeated standard antibiotics. Thirty-three isolates were recovered from a healthy donor stool sample. Two patients who had failed at least three courses of metronidazole or vancomycin underwent colonoscopy and the mixture was infused throughout the right and mid colon. Pre-treatment and post-treatment stool samples were analyzed by 16 S rRNA gene sequencing using the Ion Torrent platform.

Results: Both patients were infected with the hyper virulent C. difficile strain, ribotype 078. Following stool substitute treatment, each patient reverted to their normal bowel pattern within 2 to 3 days and remained symptom-free at 6 months. The analysis demonstrated that rRNA sequences found in the stool substitute were rare in the pre-treatment stool samples but constituted over 25% of the sequences up to 6 months after treatment.

Conclusion: This proof-of-principle study demonstrates that a stool substitute mixture comprising a multi-species community of bacteria is capable of curing antibiotic-resistant C. difficile colitis. This benefit correlates with major changes in stool microbial profile and these changes reflect isolates from the synthetic mixture.

Clinical trial registration number: CinicalTrials.gov NCT01372943.

Figures

Figure 1

Clinical timeline of events for Patients 1 and 2. Sequence of events for the first two patients enrolled in the study. (A) Patient 1 had Clostridium difficile initially occurring after a pre-operative course of cefazolin for elective total knee arthroplasty. (B) Patient 2 had C. difficile initially occurring after a course of cefazolin for cellulitis. Both patients had multiple courses of antibiotic treatment for the C. difficile infection with both vancomycin and metronidazole prior to enrollment, as indicated. In addition, Patient 1 received the probiotic Saccharomyces boulardii. Prior to treatment with the stool substitute preparation RePOOPulate (RP), stool collection on each patient was carried out at 2 days pre treatment (PT), day 2 post treatment (D2), week 2 post treatment (W2), week 4 post treatment (W4), and 6 months post treatment (6 M). Toxin assays for C. difficile were also performed (purple boxes), with results as shown. Incidental antibiotic use post treatment is indicated. AMX, amoxicillin; CFZ, cefazolin; CIP, ciprofloxacin; CLI, clindamycin; CRO, ceftriaxone; LEX, cephalexin; MTZ, metronidazole; NIT, nitrofurantoin; SXT, trimethoprim-sulfamethoxazole; VAN, vancomycin

Figure 2

Distance tree of weighted UniFrac distances between samples for Patient 1 amplified and sequenced independently. Distance tree calculated by the unweighted pair group method with arithmetic mean. Branch tips are colored by sample: red, pre-treatment; blue, RePOOPulate formulation. Post-treatment samples are colored green (D2), cyan (W2), and purple (W4). Tip label fields are separated by an underscore character and the fields are: Ion Torrent run ID, person and time of amplification, sample identifier, barcode sequence

Figure 3

Principle component coordinates of patient time points and most abundant sequences clustered at family level. Weighted UniFrac principle coordinates were generated by QIIME for each patient independently. These time points are denoted PT for pre treatment, RP for the RePOOPulate formula, and as the day (D), week (W) or month (M) time point post treatment. The weighted mean abundance of family-level taxonomic groups is indicated by the size and position of the open circles. For example, for Patient 1 Bacteriodaceae are abundant in the day 2 and week 2 post-treatment samples, less abundant in the week 4 post-treatment sample, and are rare in all other samples. Only the 10 most abundant groupings of organisms are shown, and these differ between the two patients, although the Lachnospiraceae family is abundant in both

Figure 4

Barplot of abundance at the family level. Operational taxonomic units (OTUs) that comprised more than 0.5% of the OTUs in any sample were grouped into the appropriate family and plotted. These plots show how the actual composition of each sample changes over time. Note that the two patients had very different initial microbiota compositions. The compositional differences were maintained at all time points, suggesting that environmental or genetic factors were important in shaping community structure

Figure 5

Weighted abundance overlap at the identical sequence unit and 97%-clustered operational taxonomic unit levels. Proportion of sequence counts that correspond exactly to those in the RePOOPulate (RP) formulation and found in each patient sample as a function of time post treatment. Red, RP formulation; dark blue, samples from Patient 1; cyan, samples from Patient 2. There is an initial increase in reads identical to the RP reads immediately after treatment, and a steady decline in proportion for each patient with time since treatment. Both patients had similar RP-identical reads at 6 months post treatment, even though their microbiota profiles were different