Gene expression-based classifiers identify Staphylococcus aureus infection in mice and humans.

Area of Science:

• Infectious Diseases
• Immunology
• Computational Biology

Background:

• Staphylococcus aureus infections present diagnostic challenges, leading to treatment delays and antibiotic misuse.
• Host inflammatory responses are increasingly recognized as potential diagnostic markers for infections.
• Current diagnostic methods for S. aureus bloodstream infections (BSI) can be slow and lack specificity.

Purpose of the Study:

• To investigate if host inflammatory responses to Staphylococcus aureus differ quantitatively from those to Escherichia coli infection.
• To develop and validate gene-expression-based classifiers for diagnosing S. aureus infection using host response data.
• To explore conserved and disparate host response pathways between murine and human S. aureus infections.

Main Methods:

• Bayesian sparse factor modeling and penalized binary regression were employed to identify gene-expression signatures.
• Classifiers were trained on peripheral blood gene expression data from murine models of S. aureus and E. coli infection.
• Human subjects with S. aureus BSI, E. coli BSI, and healthy controls were included for classifier development and validation.

Main Results:

• A murine-derived classifier accurately distinguished S. aureus from healthy controls and E. coli-infected mice (AUC > 0.97).
• A human-derived classifier differentiated S. aureus BSI from healthy controls (AUC 0.99) and E. coli BSI (AUC 0.84).
• The murine classifier successfully predicted human S. aureus BSI (AUC 0.84), and both classifiers showed high performance in independent human cohorts (AUCs 0.95 and 0.92).

Conclusions:

• Host gene-expression patterns provide a robust basis for differentiating S. aureus infections from other conditions, including E. coli infections.
• The study identified conserved host response pathways between mice and humans, validating the use of murine models for studying human S. aureus infections.
• The developed classifiers represent a promising new avenue for rapid and accurate diagnostics of S. aureus bloodstream infections, potentially improving patient outcomes and guiding antibiotic stewardship.