Resensitizing carbapenem- and colistin-resistant bacteria to antibiotics using auranofin

Carbapenem-resistant Gram-negative bacteria, frequently driven by plasmid-encoded metallo-β-lactamases such as NDM-1, undermine nearly all β-lactam therapies and often co-harbor multiple resistance determinants. Colistin has been reintroduced as a last-resort agent, but its efficacy is compromised by plasmid-mediated MCR enzymes (for example MCR-1) that modify lipid A via phosphoethanolamine transfer, reducing colistin binding. Both NDM-1 (a B1 class MBL) and MCR-1 rely on Zn(II) ions at their catalytic centers. Prior work showed that metal-based approaches (e.g., bismuth compounds) can inactivate MBLs by displacing Zn(II). This study tests the hypothesis that auranofin, a gold(I)-based drug, can act as a dual inhibitor of MBLs and MCRs by displacing their essential Zn(II), thereby restoring bacterial susceptibility to carbapenems and colistin and suppressing resistance development.

The spread of NDM-1 since 2008 has been rapid, with blaNDM plasmids often co-carrying multiple resistance determinants, exacerbating multidrug resistance. Colistin’s membrane-disruptive action is neutralized by MCR-1-mediated phosphoethanolamine addition to lipid A, a mechanism disseminated globally. Previous strategies to inhibit MBLs include β-lactamase inhibitors and metal chelators or metal-based agents (e.g., colloidal bismuth subcitrate) that displace Zn(II). Attempts to resensitize MCR-1-positive bacteria have included combination therapies (e.g., clarithromycin) and substrate mimics (e.g., ethanolamine). However, given the divergent active sites and mechanisms of MBLs and MCRs, no clinically available inhibitor has simultaneously targeted both. This work positions auranofin as a potential dual inhibitor addressing both resistance mechanisms.

- Enzyme production and purification: Overexpressed and purified native Zn2+-bound NDM-1. Produced full-length MCR-1 and a soluble periplasmic domain (MCR-1-S) for biochemical and structural studies. - Use of AUR analog: For biophysical assays, used chloro(triethylphosphine)gold(I), Au(PEt3)Cl, to mimic the intracellular active species of auranofin. - NDM-1 inhibition assays: Assessed hydrolysis of nitrocefin in the presence of varying Au(PEt3)Cl concentrations to determine IC50 and inhibition kinetics. Conducted Lineweaver–Burk analyses to estimate effects on Vmax and Km. - Metal displacement and binding: Quantified metal content by ICP-MS. Used PAR assay to monitor Zn(II) release from proteins upon Au(PEt3)Cl exposure. Performed equilibrium dialysis to determine Au binding stoichiometry and effects on Zn(II) content in MCR-1-S. - Cellular engagement: Employed cellular thermal shift assays (CETSA) in NDM-1- and MCR-1-expressing E. coli to detect thermal stability shifts upon AUR treatment, indicating target engagement. - Structural biology: Obtained crystals of Au-bound NDM-1 (PDB: 6HLE) and of Zn- and Au-bound MCR-1-S (Zn-bound PDB: 6L14; Au-bound PDB: 6L16) via co-crystallization or soaking. Solved structures and identified Au occupancy and coordination at active sites. - MCR-1 activity assay: Measured cleavage/transfer activity on fluorescent PEA substrates (e.g., NBD-labeled substrates) using TLC-based readouts, comparing Zn-MCR-1 vs AUR/Au(PEt3)Cl-treated enzyme. - Antimicrobial synergy testing: Performed broth microdilution checkerboard assays to determine MICs and FIC indices for combinations of auranofin with meropenem (MER) against MBL-positive strains and with colistin (COL) against MCR-positive strains, including panels expressing diverse MCR variants/homologs. - Time–kill studies: Monitored CFU over 24 h for monotherapies and combinations against representative NDM-1- or MCR-1-positive strains. - Resistance development: Serial passage under sub-inhibitory concentrations of MER or COL with/without AUR; monitored changes in MIC/MBC and frequency of resistant mutants. - In vivo efficacy: Murine peritonitis models using MCR-1-positive K. pneumoniae and E. coli co-expressing MCR-1 and NDM-5. Single-dose intraperitoneal treatments of vehicle, AUR, COL, or combination; measured bacterial loads in organs and survival over 5 days. Doses were chosen based on in vitro synergy and pilot tolerability; COL dosed at approximately human-equivalent ranges; AUR at subclinical levels.

- Auranofin inhibits NDM-1: Au(PEt3)Cl inhibited NDM-1 nitrocefin hydrolysis with IC50 = 437.9 ± 29.1 nM. Kinetic analysis showed a marked decrease in apparent Vmax (from ~19.41 to ~3.35 μM·min−1) with little change in Km, consistent with effective inhibition. CETSA in NDM-expressing cells showed NDM-1 melting temperature shift from 37.0 °C (control) to 58.6 °C with AUR, supporting intracellular binding. - Metal displacement mechanism: PAR and ICP-MS demonstrated AUR-induced Zn(II) release from NDM-1 and MCR-1-S, with Au binding that Zn(II) could not reverse. Structural data (PDB 6HLE; 6L16) revealed Au occupying/coordinating at active sites, displacing Zn(II), consistent with irreversible catalytic inactivation. - Restoring carbapenem activity: Against an NDM-1-positive E. coli isolate (NDM-HK), AUR + meropenem showed strong synergy (FIC index 0.156) and reduced MER bactericidal concentration by 64-fold. Time–kill assays showed >5-log10 CFU reduction at 24 h with the combination versus no significant effect with either agent alone. Synergy extended to various MBL-producing Enterobacteriaceae with FIC indices ~0.133–0.375. - Slower resistance emergence to MER: Serial passage with MER alone led to ~8-fold increases in resistance; inclusion of AUR limited increases to ~2-fold. - Inhibiting MCR-1 catalysis: Full-length MCR-1 lost cleavage/transfer activity on fluorescent PEA substrates after treatment with AUR/Au(PEt3)Cl. MCR-1-S bound ~3 equivalents of Au with concurrent loss of ~2–2.2 equivalents of Zn. CETSA in MCR-1-expressing cells showed a decrease in melting temperature (~61.7 to 58.1 °C) upon AUR, indicating binding. - Restoring colistin activity: In MCR-1-positive E. coli, AUR resensitized to colistin with a 64-fold MIC reduction (from 8 to 0.125 μg mL−1); FIC index 0.125 (synergy). No synergy in MCR-1-negative strains (FIC ~0.351). Time–kill assays showed >10^7-fold CFU reductions at 24 h with AUR + COL vs monotherapies. AUR at fixed concentration (0.625 μg mL−1) reduced COL MICs by 8–16-fold across multiple MCR-1 variants and MCR homologs, indicating broad-spectrum effect. - Slower resistance to COL: Serial passage with COL alone increased resistance more markedly than with AUR + COL combination, which curtailed resistance escalation. - In vivo efficacy: In a sublethal K. pneumoniae (MCR-1+) peritonitis model, AUR + COL reduced bacterial loads in spleen and liver by >10-fold vs COL alone. In a lethal E. coli (MCR-1+, NDM-5+) model, combination therapy achieved 100% survival at 5 days, whereas vehicle or AUR alone led to complete mortality and COL monotherapy rescued only 2/6 mice.

This work demonstrates that auranofin functions as a dual inhibitor of two mechanistically and structurally distinct Zn(II)-dependent resistance enzymes, MBLs (e.g., NDM-1) and MCR-1. Biochemical assays, cellular engagement (CETSA), and crystallography support a mechanism in which Au(I) displaces essential Zn(II) from active sites, irreversibly disabling enzyme function. Functionally, AUR restores the activity of meropenem against MBL-positive strains and colistin against MCR-positive strains, producing robust synergistic killing and significantly suppressing resistance development during serial passage. In vivo, the AUR–COL combination reduced organ bacterial burdens and achieved full survival in a lethal co-resistant infection model, underscoring translational potential. The findings suggest a mechanistic synergy wherein inactivation of MCR-1 restores colistin’s membrane-disrupting action, which may also enhance uptake of hydrophobic AUR, reinforcing antibacterial activity. By simultaneously targeting MBL and MCR enzymes, AUR acts as an adjuvant that can potentially extend the clinical utility of last-resort antibiotics in the face of co-harbored resistance determinants.

Auranofin is identified as a potent dual inhibitor of metallo-β-lactamases and MCR-family enzymes by displacing catalytic Zn(II), thereby resensitizing carbapenem- and colistin-resistant Gram-negative bacteria to meropenem and colistin. It exhibits strong in vitro synergy across diverse resistant strains, slows resistance emergence, and demonstrates significant in vivo efficacy in murine infection models. These results support repurposing auranofin as an antibiotic adjuvant against co-resistant pathogens. Future work should evaluate pharmacokinetics/pharmacodynamics of AUR–antibiotic combinations, optimize dosing regimens, assess safety and toxicity in relevant models, investigate spectrum and off-target effects in complex microbiomes, and explore structural analogs with improved selectivity and potency.

- Some biochemical experiments used an AUR analog (Au(PEt3)Cl) to mimic the intracellular active species; translation of these results to clinical auranofin requires careful PK/PD validation. - The in vivo studies were limited to murine peritonitis models and a small number of strains; broader pathogen panels and infection models are needed. - Comprehensive toxicity and safety profiles for AUR in combination with colistin or meropenem at antibacterial dosing were not extensively characterized here. - The potential for off-target effects exists, as auranofin can interact with multiple cellular targets; long-term resistance evolution and fitness impacts were evaluated only over limited passages. - Reported FIC/MIC metrics in some textual sections were variable due to experimental conditions; standardized, multicenter validation would strengthen generalizability.