A novel approach for analyzing foodborne zoonotic E. coli

By Bhavana KunkalikarReviewed by Lily Ramsey, LLMMar 29 2023

In a recent study published in the One Health journal, researchers in the United States (US) identified extraintestinal Escherichia coli infections using source-related mobile genetic elements.

Study: Using source-associated mobile genetic elements to identify zoonotic extraintestinal E. coli infections. Image Credit: Surrender/Shutterstock.com

Background

While E. coli urinary tract infections (UTIs) seldom lead to major invasive diseases, its prevalence makes it the leading reason for E. coli-related sepsis mortalities reported worldwide.

However, a fundamental comprehension still needs to be improved related to the origination and transmission dynamics associated with causative E. coli strains, especially zoonotic types that originate from food animals and are acquired via food intake. Closing these knowledge gaps is essential to curb and prevent these prevalent infections.

About the study

In the present study, researchers compared E. coli genomes derived from retail meat items with human clinical samples to recognize source-associated mobile genetic elements (MGEs).

In Flagstaff, Arizona, retail meat specimens and clinical isolates were obtained between 1 January 2012 and 31 December 2012. Across this 12-month timespan, all brands of turkey, raw chicken, and pork samples were collected every two weeks from Flagstaff's nine largest grocery stores.

Simultaneously, all human clinical E. coli isolates were collected from urinary and blood specimens at the Flagstaff Medical Center, the principal clinical laboratory covering Flagstaff and the surrounding area. An enrichment technique obtained a single E. coli isolate from each meat item. Using disk diffusion, the antimicrobial susceptibility of the meat isolates was evaluated. Using standard procedures, E. coli isolates from urine and blood were obtained in the clinical laboratory.

For patients with E. coli-positive urine and blood cultures, symptom keywords, primary and secondary diagnoses, and clinical laboratory findings were extracted from their medical records whenever available. The team also created deoxyribonucleic acid (DNA) sequence libraries from 1,188 human clinical blood and urine and 1,923 meat isolate samples, including turkey, chicken, and pork.

Results

During the 12-month study period, there was a discernible population diversity among the retail meat and human clinical E. coli isolate samples obtained contemporaneously at the study site. Some sequence types (STs) involved only animal or human E. coli isolates, whereas others included isolates belonging to both sample types. In particular, 443 STs were identified among the 3,120 E. coli isolates, of which 247 STs included only meat, 120 STs contained only human, and 76 STs contained both human and meat isolates.

To study potential zoonotic transmissions of E. coli, the team constructed rooted phylogenies associated with 56 individual STs comprising four or more E. coli isolates from human and meat isolates. Even among isolates derived from the same sample type and ST, the derived phylogenies demonstrated high genetic variation among the E. coli populations and some close clonal connections.

This indicated that core-genome phylogenetic assessment by itself could not be utilized to detect recent zoonotic transmission incidences easily. The team also noted 366 source-associated accessory genes, which were non-redundant.

Most of these accessory genes belonged to multigene clusters, of which 17 had MGE characteristics. Eleven plasmids, one integrative and conjugative element, one integrative and mobilizable element, and four prophages were among the 17 source-associated MGEs. Each MGE had between two and 42 source-related accessory genes that consistently co-occurred across strains. Six of the 17 source-related MGEs were connected with humans, whereas the remaining 11 were associated with meat.

The team utilized a two-class Bayesian latent class model (BLCM) to detect human E. coli isolates derived from food animals via contaminated meat. The criterion for evaluating the model was congruence between its anticipated origin and the true sample type.

The curated isolate collection was randomly divided into two datasets, one involving two-thirds of the information while the second contained the remaining one-third. With the isolates from the validation set, 96.3% of the model's projected source matched the known sample type. The BLCM origin predictions for the entire set of isolates revealed a bimodal distribution consistent with the sample type.

Conclusion

The study findings highlighted the creation of new tools for analyzing E. coli and quantifying the fraction of human E. coli infections elicited by foodborne zoonotic E. coli strains. This was achieved by detecting 17 source-associated MGEs to estimate the origination of E. coli isolates. Strains of E. coli may lose or acquire host-adaptive MGEs during host migrations.

The approach described here can be adopted in different settings to recognize the highest-risk foodborne zoonotic strains, assess their sources, and inform the development of new initiatives to reduce the heavy impact of extraintestinal E. coli infections on public health.