Aminoglycoside antibiotics are powerful bactericidal therapeutics that are often used in the treatment of critical Gram-negative systemic infections. The emergence and global spread of antibiotic resistance, however, has compromised the clinical utility of aminoglycosides to an extent similar to that found for all other antibiotic-drug classes. Apramycin, a drug candidate currently in clinical development, was suggested as a next-generation aminoglycoside antibiotic with minimal cross-resistance to all other standard-of-care aminoglycosides. Here, we analyzed 591,140 pathogen genomes deposited in the NCBI National Database of Antibiotic Resistant Organisms (NDARO) for annotations of apramycin-resistance genes, and compared them to the genotypic prevalence of carbapenem resistance and 16S-rRNA methyltransferase (RMTase) genes. The 3-N-acetyltransferase gene aac(3)-IV was found to be the only apramycin-resistance gene of clinical relevance, at an average prevalence of 0.7%, which was four-fold lower than that of RMTase genes. In the important subpopulation of carbapenemase-positive isolates, aac(3)-IV was nine-fold less prevalent than RMTase genes. The phenotypic profiling of selected clinical isolates and recombinant strains expressing the aac(3)-IV gene confirmed resistance to not only apramycin, but also gentamicin, tobramycin, and paromomycin. Probing the structure–activity relationship of such substrate promiscuity by site-directed mutagenesis of the aminoglycoside-binding pocket in the acetyltransferase AAC(3)-IV revealed the molecular contacts to His124, Glu185, and Asp187 to be equally critical in binding to apramycin and gentamicin, whereas Asp67 was found to be a discriminating contact. Our findings suggest that aminoglycoside cross-resistance to apramycin in clinical isolates is limited to the substrate promiscuity of a single gene, rendering apramycin best-in-class for the coverage of carbapenem- and aminoglycoside-resistant bacterial infections.

The aminoglycoside apramycin has been proposed as a drug candidate for the treatment of critical Gram-negative systemic infections. However, the potential of apramycin in the treatment of drug-resistant bloodstream infections (BSIs) has not yet been assessed. The resistance gene annotations of 40 888 blood-culture isolates were analysed. In vitro profiling of apramycin comprised cell-free translation assays, broth microdilution, and frequency of resistance determination. The efficacy of apramycin was studied in a mouse peritonitis model for a total of nine Escherichia coli and Klebsiella pneumoniae isolates. Genotypic aminoglycoside resistance was identified in 87.8% of all 6973 carbapenem-resistant Enterobacterales blood-culture isolates, colistin resistance was shown in 46.4% and apramycin in 2.1%. Apramycin activity against methylated ribosomes was > 100-fold higher than that for other aminoglycosides. Frequencies of resistance were < 10^-9 at 8 × minimum inhibitory concentration (MIC). Tentative epidemiological cut-offs (TECOFFs) were determined as 8 µg/mL for E. coli and 4 µg/mL for K. pneumoniae. A single dose of 5 to 13 mg/kg resulted in a 1-log colony-forming unit (CFU) reduction in the blood and peritoneum. Two doses of 80 mg/kg resulted in an exposure that resembles the AUC observed for a single 30 mg/kg dose in humans and led to complete eradication of carbapenem- and aminoglycoside-resistant bacteraemia. Encouraging coverage and potent in vivo efficacy against a selection of highly drug-resistant Enterobacterales isolates in the mouse peritonitis model warrants the conduct of clinical studies to validate apramycin as a drug candidate for the prophylaxis and treatment of BSI.

Background: Whole genome sequencing of clinical bacterial isolates holds promise in predicting their susceptibility to antibiotic therapy, based on a detailed understanding of the phenotypic manifestation of genotypic variation. The aac(6′) aminoglycoside acetyltransferase gene family is the most abundant aminoglycoside resistance determinant encountered in clinical practice. A variety of AAC(6′) isozymes have been described, suggesting a phenotypic distinction between subtype I, conferring resistance to amikacin (AMK), and subtype II, conferring resistance to gentamicin (GEN) instead. However, the epidemiology and thus clinical relevance of the various and diverse isozymes and their phenotypic distinction demand systematic and contemporary re-assessment to reliably predict bacterial susceptibility to aminoglycoside antibiotics. Methods: We analyzed the resistance gene annotations of 657,603 clinical bacterial isolates to assess the prevalence and diversity of aac(6′) genes. Seventeen unique aac(6′) amino acid sequences were cloned and expressed under defined promoter control in otherwise isogenic E. coli cells for phenotypic analysis with twenty distinct aminoglycoside antibiotics. A panel of clinical isolates was analyzed for the genotype–phenotype correlation of aac(6′). Results: An aac(6′) resistance gene annotation was found in 139,236 (21.2%) of the clinical isolates analyzed. AMK resistance-conferring aac(6′)-I genes dominated in Enterobacterales (28.5%). In Pseudomonas aeruginosa and Acinetobacter baumannii, a gene conferring the aac(6′)-II phenotype but annotated as aac(6′)-Ib₄ was the most prevalent. None of the aac(6′) genes were annotated as subtype III, but gene aac(6′)-Ii identified in Gram-positive isolates displayed a subtype III phenotype. Genes that were annotated as aac(6′)-Ib₁₁ in Enterobacterales conferred resistance to both AMK and GEN, which we propose constitutes a novel subtype IV when applying established nomenclature. A phenotypic assessment facilitated structural re-assessment of the substrate promiscuity of AAC(6′) enzymes. Conclusions: Our study provides the most comprehensive analysis of clinically relevant aac(6′) gene sequence variations to date, providing new insights into a differentiated substrate promiscuity across the genotypic spectrum of this gene family, thus translating into a critical contribution towards the development of amino acid sequence-based in silico antimicrobial susceptibility testing (AST).