Transmission of the Potential Pathogen Atypical Enteropathogenic Escherichia coli by Fecal Microbiota Transplant.

Here we report evidence of transfer of atypical enteropathogenic Escherichia coli (aEPEC) using fecal microbiota transplant (FMT) to a patient treated for recurrent Clostridium difficile infection (CDI). FMT is an effective treatment for recurrent CDI.1Cammarota G. et al.Gut. 2019; 68: 2111-2121Crossref PubMed Scopus (233) Google Scholar Extensive screening ensures donors are devoid of numerous medical and significant health measures including but not limited to gastrointestinal symptoms and infections.2Kassam Z. et al.N Engl J Med. 2019; 381: 2070-2072Crossref PubMed Scopus (81) Google Scholar However, a US Food and Drug Administration (FDA) alert announced adverse patient outcomes after FMT due to the development of EPEC and Shiga toxin-producing E coli infections.3U.S. FDAhttps://www.fda.gov/vaccines-blood-biologics/safety-availability-biologics/safety-alert-regarding-use-fecal-microbiota-transplantation-and-risk-serious-adverse-events-likelyGoogle Scholar Direct transfer from donor to recipient has been confirmed for Shiga toxin-producing E coli but not EPEC.4Zellmer C, et al. 2021;72:e876–e880.Google Scholar Before 2020, there was no specific recommendation for EPEC screening of donor stool.1Cammarota G. et al.Gut. 2019; 68: 2111-2121Crossref PubMed Scopus (233) Google Scholar Immediately after the FDA alerts, more rigorous donor screening methods for infectious agents, including EPEC, were instated.5U.S. FDAhttps://www.fda.gov/vaccines-blood-biologics/safety-availability-biologics/information-pertaining-additional-safety-protections-regarding-use-fecal-microbiota-transplantation-0Google Scholar Although EPEC has long been considered a pathogen primarily affecting infants in developing nations, its increased detection and association with disease in children and adults in the United States6Carlino M.J. Kralicek S.E. et al.Gut Microbes. 2020; 12: 1-21Crossref PubMed Scopus (6) Google Scholar,7Kralicek S.E. Sitaraman L.M. et al.Gastroenterology. 2022; 163: 1321-1333Abstract Full Text Full Text PDF PubMed Scopus (1) Google Scholar warrants careful consideration when detected in donor stool. EPEC has extensive genetic diversity and can be detected in asymptomatic individuals, leading to questions regarding strain pathogenicity. Therefore, we investigated if EPEC was transferred from an FMT donor to a recipient treated for recurrent CDI as well as the in vitro pathogenic potential of isolated strains. In collaboration with the FMT stool bank OpenBiome and the American Gastroenterological Association Fecal Microbiota Transplantation National Registry, we identified 1 donor–recipient pair for which EPEC-positive donor material and corresponding recipient stool were available (see Supplementary Methods). The donor was a 41-year-old woman who provided normal consistency stool (Bristol stool score 4) originally deemed EPEC-negative by stool culture and met all previous release criteria.2Kassam Z. et al.N Engl J Med. 2019; 381: 2070-2072Crossref PubMed Scopus (81) Google Scholar However, FDA alerts triggered further retrospective screening using the BioFire Gastrointestinal Panel assay (Biomerieux Clinical Diagnostics) that revealed donor samples to be EPEC-positive; this was confirmed by quantitative polymerase chain reaction. The recipient was a 78-year-old man with a history of chronic kidney disease who had experienced 4 episodes of CDI when referred for FMT. He underwent a single FMT delivered by colonoscopy using frozen banked material. At week 1 after FMT follow-up, the patient reported ongoing diarrhea. Repeat tests including glutamate dehydrogenase and immunoassay toxin were positive, indicating CDI; the patient was restarted on an extended course of vancomycin, and the CDI resolved. No gastrointestinal symptoms were reported at 30 days after FMT, and cure was reported to the FMT registry at 60 days. We determined retrospectively that stool collected 1 week after FMT was EPEC-positive by quantitative polymerase chain reaction. EPEC loads of donor and recipient stool were determined to be low (below 0.05% total stool bacteria) and similar between donor and recipient (Figure 1A). Low EPEC stool loads are characteristic of asymptomatic individuals.7Kralicek S.E. Sitaraman L.M. et al.Gastroenterology. 2022; 163: 1321-1333Abstract Full Text Full Text PDF PubMed Scopus (1) Google Scholar EPEC isolates were purified from donor and recipient stool, genomes analyzed, and determined to be bfpA-negative indicating aEPEC (Supplementary Methods). Simultaneous colonization by unique isolates within the donor was detected (OB56, OB126, and OB127) as indicated by the various phylogroups, multilocus sequence type, and serotypes present (Figure 1B). Donor-OB127 and recipient isolate, OB-REC, originate from the same phylogroup (E), multilocus sequence type (2569), and serotype (H49:O–) (Figure 1B). Sequence coverage and small nucleotide polymorphism (SNP) analysis identified 99.9% coverage and only 16 SNPs between OB-REC and donor-OB127, in contrast to 89% coverage and 48,336 SNPs detected between OB-REC and donor-OB126, which also falls within the E phylogroup (Figure 1B). The corresponding SNP neighbor-joining tree indicated a close relatedness of donor-OB127 and OB-REC isolates (Figure 1C). These data suggest transfer of aEPEC from donor to recipient after FMT. To explore the pathogenic potential of these aEPEC isolates, genetic and phenotypic virulence properties known to be involved in typical EPEC (tEPEC) pathogenesis were examined. Proteomic analysis of genes within the locus of enterocyte effacement (LEE) and non-LEE pathogenicity islands characteristic of tEPEC strain E2348/69 revealed the same protein homology pattern between OB-REC and donor-OB127 but not OB126 or OB56 (Figure 1E). These data further support transfer of aEPEC from donor to recipient. High homology within the LEE to tEPEC-E2348/69 (Figure 1D) indicates the pathogenic potential of these aEPEC isolates. However, divergence within non-LEE genetic elements could cause variation in pathogenicity. The initial step in EPEC pathogenicity is attachment to host cells. Defined bacterial attachment patterns are associated with diarrheal disease.6Carlino M.J. Kralicek S.E. et al.Gut Microbes. 2020; 12: 1-21Crossref PubMed Scopus (6) Google Scholar,8Mare A.D. et al.Gastroenterol Insights. 2021; 12: 28-40Crossref Google Scholar Therefore, we assessed attachment levels and pattern of isolates on SKCO-15 colonic-like intestinal epithelial cells. Attachment levels of donor isolates are similar to the control strain tEPEC E2348/69 mutant lacking bfpA (ΔbfpA) (Figure 1E). A modest increase in attachment of OB-REC was observed, however, the level fell below the threshold of significance for aEPEC isolates from symptomatic individuals, which often have 2- to 5-fold higher attachment than ΔbfpA6Carlino M.J. Kralicek S.E. et al.Gut Microbes. 2020; 12: 1-21Crossref PubMed Scopus (6) Google Scholar (Figure 1E). All isolates attach in an undefined manner (Figure 1F), suggesting that “healthy” donor isolates may have reduced adherence properties. A hallmark of tEPEC pathogenicity is formation of actin-rich clusters under attached bacteria termed pedestals. Pedestal formation was detected for OB56 but not for OB-REC, donor-OB126, or donor-OB127 under these in vitro conditions (Figure 1G). In conclusion, BioFire Gastrointestinal Panel assay–EPEC-positive carriers can simultaneously house unique aEPEC strains. Our data suggest the transfer of at least 1 strain from donor to recipient after FMT for recurrent CDI treatment. Because our screening approach did not encompass detection of every EPEC strain within the stools and recipient stools were not assessed for EPEC before or after FMT treatment, it is possible that other aEPEC strains were present in donor or recipient stool. Although EPEC load was relatively low in stool and the examined isolates have diminished attachment properties, high protein homology within the LEE pathogenicity island and the ability of OB56 to form pedestals indicate pathogenic potential. EPEC and C difficile co-infections are prevalent.6Carlino M.J. Kralicek S.E. et al.Gut Microbes. 2020; 12: 1-21Crossref PubMed Scopus (6) Google Scholar However, the role of interspecies interactions during CDI9Abbas A. Zackular J.P. Curr Opin Microbiol. 2020; 53: 19-25Crossref PubMed Scopus (27) Google Scholar are unknown, and the contribution of aEPEC to the clinical course reported in this case remains unclear, warranting further investigation into aEPEC pathogenic mechanisms, effects of aEPEC on CDI, and differences between aEPEC detected in symptomatic vs asymptomatic individuals. This report highlights the importance of screening stool donors for EPEC and avoiding the use of EPEC-positive stool for FMT until more refined distinguishing factors of aEPEC pathogenicity are determined. The results published here are, in part, based on data obtained from the American Gastroenterological Association Fecal Microbiota Transplantation National Registry, funded by the National Institute of Allergy and Infectious Diseases at the National Institutes of Health (award no. R24 AI118629). The authors thank the American Gastroenterological Association Fecal Microbiota Transplantation registry team for coordinating the acquisition of recipient stool from Dr Allegretti. We thank OpenBiome, Inc for providing donor stool material and the Genomics Facility at Loyola University Chicago for their efficient delivery of genomic sequencing results. Sarah E. Kralicek, MS (Conceptualization: Equal; Data curation: Equal; Formal analysis: Equal; Investigation: Lead; Methodology: Lead; Visualization: Lead; Writing – original draft: Lead; Writing – review & editing: Equal). Celeste Jenkins, BS (Investigation: Supporting; Writing – review & editing: Supporting). Jessica R. Allegretti, MD, MPH (Resources: Supporting; Writing – original draft: Supporting; Writing – review & editing: Supporting). James D. Lewis, MD, MSCE (Investigation: Supporting; Writing – review & editing: Supporting). Majdi Osman, MD (Resources: Supporting; Writing – review & editing: Supporting). Gail A. Hecht, MD (Data curation: Equal; Formal analysis: Equal; Funding acquisition: Lead; Resources: Lead; Supervision: Lead; Writing – review & editing: Equal). Donor and recipient stools were acquired under Loyola University Chicago institutional review board exempt status for retrospective analysis. The FDA alerts which reported infections with EPEC and Shiga toxin-producing E coli after FMT prompted the stool bank OpenBiome to screen all available donor material using the Biofire Gastrointestinal Panel assay. Through the aid of the American Gastroenterological Association Fecal Microbiota Transplantation National Registry, BioFire Gastrointestinal Panel–EPEC-positive donor material was cross-referenced to those individuals who had received the corresponding FMT. Donor material from the FDA alert was no longer available. One donor–recipient pair was identified for which donor and recipient stool were available. Two donor stool were acquired and corresponded to the sample used to create the FMT material and a stool sample donated 2 days before. Donor stools were received as either raw stool or glycerol-preserved stool (0.1 g of stool resuspended in a glycerol–saline solution) and stored at –80oC. Recipient stool was collected 1 week after FMT and raw stool frozen and stored at –80oC. Relative EPEC load was performed as described.1Kralicek S.E. Sitaraman L.M. et al.Gastroenterology. 2022; 163: 1321-1333Abstract Full Text Full Text PDF PubMed Scopus (1) Google Scholar In brief, genomic DNA was purified from frozen stool using the MagMAX Microbiome Ultra Nucleic Acid Isolation kit (Applied Biosystems, Foster City, CA). Quantitative polymerase chain reaction (PCR) was performed using eaeA-specific and 16S-rDNA universal primers with TBgreen Premix EXTaqII reagent (Takara Bio USA, San Jose, CA). PCR conditions were as described.2Carlino M.J. Kralicek S.E. et al.Gut Microbes. 2020; 12: 1-21Crossref PubMed Scopus (6) Google Scholar Relative EPEC load was determined by EPEC load divided by total bacteria load. EPEC isolates were purified as previously described.2Carlino M.J. Kralicek S.E. et al.Gut Microbes. 2020; 12: 1-21Crossref PubMed Scopus (6) Google Scholar In brief, frozen stools were plated on MacConkey II agar and single colonies screened using colony PCR for eaeA. eaeA-positive colonies were again streaked and screened using colony PCR to obtain double-purified eae-positive colonies that were then stored as Luria-Bertani-glycerol stocks at –80oC. Isolates were screened for bfpA as previously described2Carlino M.J. Kralicek S.E. et al.Gut Microbes. 2020; 12: 1-21Crossref PubMed Scopus (6) Google Scholar to determine typical or atypical designation. Donor and recipient EPEC isolate genomic DNA was extracted from overnight Luria-Bertani cultures using DNeasy Blood and Tissue kit (69504; Qiagen, Germantown, MD) and then submitted to the Loyola Genomic Facility for library preparation and shotgun sequencing as previously described.2Carlino M.J. Kralicek S.E. et al.Gut Microbes. 2020; 12: 1-21Crossref PubMed Scopus (6) Google Scholar Paired sequences were uploaded and assembled on the Enterobase pipeline.3Zhou Z. et al.Genome Res. 2020; 30: 138-152Crossref PubMed Scopus (378) Google Scholar In silico serotyping (H-antigen and O-antigen), Achtman multilocus sequence typing, SNP analysis, and SNP neighbor-joining tree analysis were performed using the Enterobase database3Zhou Z. et al.Genome Res. 2020; 30: 138-152Crossref PubMed Scopus (378) Google Scholar and visualized with GrapeTree.4Zhou Z. et al.Genome Res. 2018; 28: 1395-1404Crossref PubMed Scopus (376) Google Scholar Percentage of coverage analysis was performed according to Faith et al5Faith J.J. et al.Science. 2013; 3411237439Crossref Scopus (1383) Google Scholar; coverage score = (Xaln + Yaln) ÷ (X + Y), where X and Y are the lengths of genome X and Y, respectively, and Xaln and Yaln are the number of aligned bases of genome X and Y, respectively.5Faith J.J. et al.Science. 2013; 3411237439Crossref Scopus (1383) Google Scholar Protein homology comparisons of donor and recipient isolates to tEPEC strain E2348/69 were performed using the Proteome comparison tool available on the BV-BRC pipeline6Wattam A.R. et al.Methods Mol Biol. 2018; 1704: 79-101Crossref PubMed Scopus (63) Google Scholar and heatmaps generated in GraphPad Prism 9 (GraphPad Software). Attachment assays, Giemsa staining, and immunofluorescence microscopy were performed as previously described.2Carlino M.J. Kralicek S.E. et al.Gut Microbes. 2020; 12: 1-21Crossref PubMed Scopus (6) Google Scholar In brief, cells were fixed with 4% paraformaldehyde (20 minutes) and permeabilized with 0.1% Triton X-100 (15 minutes). Antibodies used were BODIPY-558/568 Phalloidin (B3475; Invitrogen, Waltham, MA) for F-actin and anti-E coli-FITC (1003; Virostat, Westbrook, ME) with the anti-rabbit AlexFluor-488 secondary antibody (A11034; Invitrogen). Nuclei were stained with Hoechst 33342 (H3570; Invitrogen). One-way analysis of variance with Tukey post-hoc tests and figure creation were performed with GraphPad Prism 9 and Illustrator 2021 (Adobe Inc, San Jose, CA). Data are presented as mean ± standard error of mean. Statistical significance was defined as P < .05, with “ns” indicating nonsignificant.