654. Core Antibiotic-Induced Transcriptional Signatures Reflect Susceptibility to All Members of an Antibiotic Class

BACKGROUND: Current growth-based antibiotic susceptibility testing (AST) is too slow to guide key clinical decisions. We previously demonstrated a combined Genotypic and Phenotypic AST assay using RNA detection (GoPhAST-R) that can provide AST in < 4h directly from blood culture, 24-36h faster than standard growth-based methods. GoPhAST-R quantifies specific mRNA expression signatures using the multiplexed hybridization RNA-detection platform, NanoString. After brief antibiotic exposure, susceptible cells become stressed, eliciting transcriptional changes that distinguish them from unharmed resistant cells. Here, we assess the generalizability of transcriptional signatures of susceptibility within an antibiotic class. METHODS: For Escherichia coli and Klebsiella pneumoniae, we assessed the ability of the top ten antibiotic-responsive genes previously identified for ciprofloxacin, gentamicin, and meropenem to predict susceptibility to two other fluoroquinolones (FQ), two other aminoglycosides (AG), and six other beta-lactams (BL), respectively, across 6-8 clinical isolates for each drug for a total of 184 pathogen-drug pairs. After standardized antibiotic exposure (60m for FQs and AGs, 120m for BLs, each at its CLSI breakpoint MIC), samples were mechanically lysed and used as input for NanoString assays as previously described (Bhattacharyya, Nat Med 2019). RESULTS: In both species, the top ten genes identified for AST of ciprofloxacin, gentamicin, and meropenem showed similar normalized fold-induction upon treatment with three FQs, three AGs, and seven BLs, respectively, allowing robust distinction of susceptible and resistant isolates (Fig 1). Figure 1 [Image: see text] CONCLUSION: We show that a shared set of genes optimized for AST of one antibiotic can predict susceptibility to all members of that drug class, consistent with a conserved core transcriptional response related to mechanism of action. We demonstrate this phenomenon for two common pathogens with propensity for multidrug resistance treated with multiple members of three major antibiotic classes in common clinical use. This unified set of genes for susceptibility prediction would streamline GoPhAST-R implementation, in turn facilitating efficient deployment of antibiotics and reducing unnecessary broad-spectrum use. DISCLOSURES: All Authors: No reported disclosures