PROJECT SUMMARY/ABSTRACT
Extended-spectrum beta-lactamase (ESBL)-producing Enterobacterales (ESBL-PE) have become widespread
in recent years, posing a serious threat to the treatment of common bacterial infections. Increased use of "last
resort" carbapenem antibiotics to treat ESBL-PE infections has fueled the emergence of carbapenem-resistant
Enterobacterales. Whole genome sequencing (WGS)-based prediction of antibiotic resistance phenotypes may
enable early initiation of appropriate treatment options for ESBL-PE infections, including selection of
carbapenem-sparing beta-lactam regimens. However, existing genomic prediction models have had relatively
poor accuracy for predicting resistance to beta-lactams in ESBL-PE. The genetic environment of ESBL genes,
including their location in the bacterial genome and association with mobile genetic elements (MGEs), is
unaccounted for in these models. These structural factors may contribute to beta-lactam resistance by providing
promotors modulating the expression of ESBL genes and enabling their duplication and mobilization within and
between Gram-negative bacteria. We hypothesize that including MGEs in genomic prediction models will
improve detection of beta-lactam resistance phenotypes in ESBL-PE. The objectives of this proposal are to 1)
determine how MGEs modulate beta-lactam minimum inhibitory concentrations (MICs) in ESBL-PE and 2)
develop and optimize multivariate regression models for predicting beta-lactam resistance from WGS data in
ESBL-PE. Our multidisciplinary research team has the necessary expertise in bacterial genomics, molecular
biology, infectious diseases epidemiology, and biostatistics to ensure successful completion of the proposed
studies. To characterize the genetic environment of ESBL genes, we will take advantage of the ability of
nanopore WGS to resolve structurally complex genomic regions including MGEs and plasmids. We will test our
hypothesis through the following three aims: 1) we will comprehensively genotype and phenotype a large
collection of clinical ESBL-PE isolates (n=450) to test the association between specific MGE genotypes and
beta-lactam MICs; 2) we will evaluate potential mechanisms whereby MGEs may affect beta-lactam resistance
phenotypes, focusing on the role that MGE-associated promotors play in modulating ESBL gene expression and
the impact of ESBL gene copy number on beta-lactam MICs; 3) we will use machine learning to build and
optimize multivariate regression models incorporating MGE genotypes and other genetic factors to predict beta-
lactam resistance phenotypes and validate our resulting models in a prospective collection of 200 ESBL-PE
isolated from urine cultures. The development of rapid diagnostic methods that predict antibiotic treatment
options to ESBL-PE should be a research priority. In addition to determining how MGEs contribute to ESBL-PE
resistance phenotypes, the proposed research will help enable integration of nanopore WGS into surveillance
and diagnostic approaches to detect beta-lactam resistance in ESBL-PE and facilitate the selection of
appropriate carbapenem-sparing regimens for ESBL-PE infections using genomic prediction models.
