Fecal Microbial Transplants Reduce Antibiotic-resistant Genes in Patients With Recurrent Clostridium difficile Infection

Abstract

Background: Recurrent Clostridium difficile infection (RCDI) is associated with repeated antibiotic treatment and the enhanced growth of antibiotic-resistant microbes. This study tested the hypothesis that patients with RCDI would harbor large numbers of antibiotic-resistant microbes and that fecal microbiota transplantation (FMT) would reduce the number of antibiotic-resistant genes.

Methods: In a single center study, patients with RCDI (n = 20) received FMT from universal donors via colonoscopy. Stool samples were collected from donors (n = 3) and patients prior to and following FMT. DNA was extracted and shotgun metagenomics performed. Results as well as assembled libraries from a healthy cohort (n = 87) obtained from the Human Microbiome Project were aligned against the NCBI bacterial taxonomy database and the Comprehensive Antibiotic Resistance Database. Results were corroborated through a DNA microarray containing 354 antibiotic resistance (ABR) genes.

Results: RCDI patients had a greater number and diversity of ABR genes compared with donors and healthy controls. Beta-lactam, multidrug efflux pumps, fluoroquinolone, and antibiotic inactivation ABR genes were increased in RCDI patients, although donors primarily had tetracycline resistance. RCDI patients were dominated by Proteobacteria with Escherichia coli and Klebsiella most prevalent. FMT resulted in a resolution of symptoms that correlated directly with a decreased number and diversity of ABR genes and increased Bacteroidetes and Firmicutes with reduced Proteobacteria. ABR gene profiles were maintained in recipients for up to a year following FMT.

Conclusions: RCDI patients have increased numbers of antibiotic-resistant organisms. FMT is effective in the eradication of pathogenic antibiotic-resistant organisms and elimination of ABR genes.

Keywords: C. difficile; antibiotic resistance; antibiotics; colitis; intestinal microbiome.

Figure 1.

Microbial profile of stool samples from recurrent Clostridium difficile infection patients and donors. Microbial composition in stool samples from single successful fecal microbiota transplant (FMT) recipients shown at the phyla (A) and genus (C) levels before FMT (BFT; n = 11), donor samples (Donors; n = 11), and post FMT (PFT; n = 11). Microbial composition in stool samples from repeated FMT cohort shown at the phyla (B) and genus (D) level before FMT (BFT; n = 9), first donor samples (1st Donors; n = 9), patients following initial failed FMT (PFT 1; n = 9), second donor samples (2nd Donors; n = 9), and patients following successful second FMT (PFT2; n = 6). Data are presented as the fraction of total reads in each sample at the phyla level and 14 most prevalent genera.

Figure 2.

Principal coordinates analysis (PCoA) of taxonomic data from recurrent Clostridium difficile infection (RCDI) patients and donors. Biplot vectors show the 4 genera with the strongest magnitude that contributed to sample dissimilarity. A, Before FMT (BFT), RCDI patients were defined by higher proportions of Klebsiella and Escherichia, whereas donors were predominated by Ruminococcus and Bacteroides. Post FMT (PFT), all but one RCDI patient clustered together with donor samples. B, RCDI patients who failed the initial FMT were also defined by Klebsiella, Enterobacteria, and Escherichia BFT, and did not resemble the donors until after the second FMT. PCoA was performed using Bray-Curtis dissimilarity to quantify the compositional dissimilarity between samples. A, Green circles: BFT samples; Black triangles: PFT samples; Red squares: donors samples. B, Repeated FMT cohort: Blue circles: BFT samples; Red squares: 1st donor samples; Black triangles: post 1st FMT recipient samples; Red diamonds: 2nd Donors samples; Brown circle: post 2nd FMT recipient samples. Donor cohort: n = 31; Single FMT Cohort: n = 11; Repeated FMT Cohort n = 9. Abbreviation: FMT, fecal microbiota transplant.

Figure 3.

Number of antibiotic-resistance (ABR) genes in recurrent Clostridium difficile infection (RCDI) and donor stool samples pre and post fecal microbiota transplantation (FMT). Metagenomic data was aligned using Bowtie2 to the Comprehensive Antibiotic Resistance Database (http://arpcard.mcmaster.ca) to detect ABR genes. A, RCDI patients at baseline (before FMT [BFT]) had increased numbers of ABR genes compared with the donors and healthy controls (HC). B, Following successful FMT, the number of detected ABR genes decreased in the RCDI patients and this was maintained over time. All patients had a post-FMT (PFT) sample taken between 1–4 weeks following FMT (Group 1: n = 11); 9 patients had a sample taken between 4 and 8 weeks following FMT (Group 2: n = 9); 7 patients had a sample taken between 8 and 22 weeks following FMT and 2 patients had a 1 year PFT sample (Combined in Group 3). C, In the patients which did not respond to the initial FMT, there was no decrease seen in the number of ABR genes following the first FMT (PFT1). In this cohort, a decrease in the number of ABR genes was seen following the second FMT and this was maintained over time. All patients had a post-FMT (PFMT1) taken between 1–3 weeks following FMT2 (Group 1: n = 9); 4 patients had a sample taken between 3 and 12 weeks following FMT2 and 2 patients had a 28 week follow-up (Combined in Group 2: n = 6). Healthy cohort: n = 87; Donor cohort: n = 29; Single FMT Cohort: n = 11; Repeated FMT Cohort n = 9. Whiskers denote mean with the standard deviation. *P-value <.05; **P-value <.005; **P-value <.0001.

Figure 4.

Relative abundance of antibiotic resistance (ABR) gene types assigned to antibiotic class resistance in recurrent Clostridium difficile infection (RCDI) patients and donors prior to fecal microbiota transplantation (FMT). Identified ABR genes were classified according to Comprehensive Antibiotic Resistance Database (http://arpcard.mcmaster.ca). Genes that confer resistance to multiple antibiotics were included in the analysis. A, The RCDI patients had a greater diversity of ABR gene classes compared with donors. B, Venn diagram showing the shared and unique ABR genes in the RCDI patients and donors. C, Venn diagram showing the shared and unique ABR gene classes in the RCDI patients and donors. Green: single FMT cohort; blue: repeated FMT cohort; red: donors. Donor cohort: n = 29; Single FMT Cohort: n = 11; Repeat FMT Cohort n = 9.

Figure 5.

Pearson similarity index of donors over time and recurrent Clostridium difficile infection (RCDI) patients to donors. Samples were analyzed using a DNA microarray. A, Donors retained a high degree of similarity to their own samples over their donation time period. Donor 1 = 18 months; Donor 2 = 12 months; Donor 3 = 7 months (B) Percent similarity of the single fecal microbiota transplantation (FMT) recipient samples at those time points (1 = 1–4 weeks; 2 = 8–52 weeks post FMT [PFT]) to their respective donor. Following FMT, the similarity of the RCDI patients increased to more closely resemble the donors; ****P-value <.0001 calculated using paired t-test; ***P-value =.0005 calculated using Mann–Whitney test. C, Percent similarity of the repeated FMT recipient samples at those time points to the respective donor of the most recent FMT (PFT1: 1–2 weeks; PFT2: 3–28 weeks); *P-value <.05; **P-value <.005 both calculated using Mann–Whitney test. Whiskers denote mean with the standard deviation. Abbreviation: BFT, before FMT.