Top-Down Genomic Surveillance Approach To Investigate the Genomic Epidemiology and Antibiotic Resistance Patterns of Enterococcus faecium Detected in Cancer Patients in Arkansas

Abstract

Control of hospital-associated Enterococcus faecium infection is a strenuous task due to the difficulty of identifying transmission routes and the persistence of this nosocomial pathogen despite the implementation of infection control measures that have been successful with other important nosocomial pathogens. This study provides a comprehensive analysis of over 100 E. faecium isolates collected from 66 cancer patients at the University of Arkansas for Medical Sciences (UAMS) between June 2018 and May 2019. In the top-down approach used in this study, we employed, in addition to the 106 E. faecium UAMS isolates, a filtered set of 2,167 E. faecium strains from the GenBank database to assess the current population structure of E. faecium species and, consequently, to identify the lineages associated with our clinical isolates. We then evaluated the antibiotic resistance and virulence profiles of hospital-associated strains from the species pool, focusing on antibiotics of last resort, to establish an updated classification of high-risk and multidrug-resistant nosocomial clones. Further investigation of the clinical isolates collected from UAMS patients using whole-genome sequencing analytical methodologies (core genome multilocus sequence typing [cgMLST], core single nucleotide polymorphism [coreSNP] analysis, and phylogenomics), with the addition of patient epidemiological data, revealed a polyclonal outbreak of three sequence types occurring simultaneously in different patient wards. The integration of genomic and epidemiological data collected from the patients increased our understanding of the relationships and transmission dynamics of the E. faecium isolates. Our study provides new insights into genomic surveillance of E. faecium to assist in monitoring and further limiting the spread of multidrug-resistant E. faecium. IMPORTANCE Enterococcus faecium is a member of the gastrointestinal microbiota. Although its virulence is low in healthy, immunocompetent individuals, E. faecium has become the third leading cause of health care-associated infections in the United States. This study provides a comprehensive analysis of over 100 E. faecium isolates collected from cancer patients at the University of Arkansas for Medical Sciences (UAMS). We employed a top-down analytical approach (from population genomics to molecular biology) to classify our clinical isolates into their genetic lineages and thoroughly evaluate their antibiotic resistance and virulence profiles. The addition of patient epidemiological data to the whole-genome sequencing analytical methodologies performed in the study allowed us to increase our understanding of the relationships and transmission dynamics of the E. faecium isolates. This study provides new insights into genomic surveillance of E. faecium to help monitor and further limit the spread of multidrug-resistant E. faecium.

Keywords: Enterococcus faecium; daptomycin; daptomycin resistance; genomic epidemiology; genomic surveillance; population structure; vancomycin; vancomycin resistance.

FIG 1

Pairwise genomic distance-based tree. Pairwise genomic distances were obtained using FastANI. The tree shows a clear division of the species into clades A1, A2, and B. The first ring indicates whether the isolate was environmental or clinical. The second ring represents the isolation source. The presence of the vancomycin resistance determinants VanA, VanB, and VanD is represented in rings 3 to 5. The outer ring represents the presence of mutations in liaSR genes, related to development of the DAP-nonsusceptible phenotype. The location on the tree of the 106 E. faecium isolates collected for this study is indicated with dashed red boxes.

FIG 2

(A) Pairwise genomic distance-based tree using ANI measurements of 1,384 clade A1 E. faecium strains. The outer rings of the tree represent the resistome of each strain. The innermost ring represents sequence type affiliation. The location on the tree of the 106 E. faecium isolates collected for this study is highlighted with dashed red boxes. (B) Stacked bar plot of ARDs identified in strains from clade A1 with MLST distribution. A total of 48 nonredundant ARDs and 75 sequence types were annotated for strains in clade A1.

FIG 3

Pairwise coreSNP heat map of 106 E. faecium isolates from 66 cancer patients. The heat map was colored using pairwise coreSNP distance. Warmer colors in the heat map represent clusters of E. faecium clones. Metadata, including patient number, sequence type, isolation source, presence of vanA cluster, and presence of daptomycin nonsusceptibility related mutations, are annotated. Data for patients with only one isolate are shown in gray.

FIG 4

Minimum spanning tree of 106 E. faecium isolates from cancer patients. Numbers in nodes represent the cgMLST classification. Node size represents the number of strains for each cgMLST. Node color represents the sequence type. Different combinations of presence/absence of the vanA cluster and DAP nonsusceptibility-related mutations are shown in the outer ring of each node in blue, purple, and red. Percentages in parentheses are the relative abundances of strains in that category. A gray background delineates isolates that were detected as part of a clonal cluster with a coreSNP threshold of 15.

FIG 5

Maximum-likelihood phylogenetic tree of 105 clade A1 and A2 E. faecium isolates. The phylogeny of the 105 isolates was obtained using the alignment of 1,266 core genome alleles and IQ-TREE software with a maximum-likelihood approach. Isolates belonging to each of the four clonal clusters are marked in the inner ring of the tree. The presence of the vanA cluster of genes conferring resistance to vancomycin and mutations in liaSR genes related to the daptomycin-nonsusceptible phenotype are represented in rings 2 and 3, respectively.