MSACL 2023 Abstract

Screening a Clinical Cohort for Urinary Tract Infections Directly from Specimen via Lipidomics Linda K Nartey (1), Richard Smith (2), Mert Pekcan (3), Hyojik Yang (2), Michael X Chen (4), Robert K Ernst (2), David R Goodlett (1) University of Victoria (1), University of Maryland-Baltimore (2), Ankara University (3), University of British Columbia (4) Linda Nartey, MSc (Presenter) University of Victoria >> POSTER (PDF)

Presenter Bio: I am currently a PhD student in Microbiology at University of Victoria, working under Professor David R Goodlett. My research is focused on investigating microbial lipidomes and proteomes directly from clinical specimen. I’m particularly interested in using mass spectrometry-based phenotypic profiling of essential species-specific lipids to identify microbes directly from clinical specimen without the need for culture.
I hold a B.Sc. in Botany from the University of Ghana where I researched on fungal and bacterial contaminants in food and how they can be controlled. After my first degree, I worked at the Noguchi Memorial Institute for Medical Research where I had hands-on experience in diagnostic problems involved in determining the cause, control and drug discovery for Buruli ulcer as well as Schistosomiasis. I also hold MSc. in Biology from Zhejiang Sci-Tech University where I researched on fungal biocontrol agents.
I can be contacted at: lnartey@uvic.ca

INTRODUCTION:
Early detection of pathogens is essential for management of urinary tract infections (UTIs). Current tests require 2-3 days for culture and antimicrobial susceptibility tests for disease diagnosis. Protein analysis techniques have become the standard for clinical microbial identification, replacing phenotypic characterization and microscopic methods. These techniques are reliable, but require culture (24-48 hours) and are often labor-intensive, which increases the cost and burden of diagnostic laboratory support. A faster workflow that maintains accuracy of protein-based methods while reducing costs and time to identification is essential.

OBJECTIVES:
To use mass spectrometry (MS)-based phenotypic profiling of species-specific membrane lipids to accurately identify microbes directly from clinically specimen without culture.

METHODS:
Urine samples were collected from the Victoria general hospital and analyzed using our novel fast lipid analysis technique (FLAT). The FLAT method allows direct lipid extraction on a MALDI plate for pathogen identification by MALDI-TOF MS (Sorensen et al. Sci Repo 2020). Briefly, 1 µL of urine was spotted on an MFX µFocus MALDI plate. Acidified and incubated in a humidified chamber at 110°C for 30 min. The plate was rinsed with distilled water and then 1 µL of Norharmane (10mg/mL) was added to each spot. After drying, samples were analyzed using a Bruker Microflex in negative ion, linear mode with automated laser operation. Results were compared against the hospital’s standard protein biotyper identification.

RESULTS:
302 urine samples were collected for this study. We identified clinically significant pathogens such as E. coli, P. aeruginosa, Proteus sp, Klebsiella sp and even as polymicrobial directly from 1 µL of urine. Overall, FLAT data produced a sensitivity of 94% and specificity of 99% with positive and negative predictive value of 93 and 99%, respectively for Gram-negatives. Two patients were found to harbor a mobile colistin resistance (mcr) gene as noted by chemical modifications to their lipid A by phosphoenthanolamine. These specimens were confirmed positive for mcr-1 by polymerase chain reaction. Additionally, ~ 40% of the negative samples showed an ion at 1446 m/z, which is reported to be the signature ion for P. aeruginosa lipid A. However, these samples containing this 1446 ion by FLAT failed to produce viable colonies when isolated on a culture plate. Tandem MS results of the 1446 ion confirmed it was a most likely a host cardiolipin (Kim et al. Journal of Lipid Research, 2011). Furthermore, FLAT detected an ion at m/z 1230 present in all urine samples with blood. Tandem MS results of this ion compared against a plasma standard analyzed by FLAT confirmed it was a heme dimer.

CONCLUSIONS:
UTIs are common infections that produce a substantial workload for a clinical microbiology laboratory. The ability to identify pathogens without need for culture allows for faster pathogen identification, reduced time to appropriate antimicrobial therapy, and improved patient outcomes. When compared to the hospital’s standard protein-based MALDI-TOF MS assay, FLAT produced accurate microbial identifications within one hour of receipt in the analytical laboratory. In addition, FLAT identified pathogens present in both mono- and poly-microbial infected UTI samples directly from specimen without culture. FLAT is an affordable, one-hour rapid test that has the potential to rule out suspected UTIs in the general population greatly reducing healthcare costs by circumventing the need for culturing negative samples. Finally, FLAT can be used to improve antimicrobial stewardship by detecting 1) UTIs negative for Gram negative pathogens circumventing the need for culture required for protein-based identification and 2) select antibiotic resistance markers, which is a major advantage of lipidomic-based microbial identification over standard protein-based methods currently used in the clinical laboratory. Our study is ongoing to optimize FLAT detection for Gram positives that have a worse limit of detection than Gram negatives.