Cryptic susceptibility to penicillin/β-lactamase inhibitor combinations in emerging multidrug-resistant, hospital-adapted Staphylococcus epidermidis lineages

Abstract

Global spread of multidrug-resistant, hospital-adapted Staphylococcus epidermidis lineages underscores the need for new therapeutic strategies. Here we show that many S. epidermidis isolates belonging to these lineages display cryptic susceptibility to penicillin/β-lactamase inhibitor combinations under in vitro conditions, despite carrying the methicillin resistance gene mecA. Using a mouse thigh model of S. epidermidis infection, we demonstrate that single-dose treatment with amoxicillin/clavulanic acid significantly reduces methicillin-resistant S. epidermidis loads without leading to detectable resistance development. On the other hand, we also show that methicillin-resistant S. epidermidis is capable of developing increased resistance to amoxicillin/clavulanic acid during long-term in vitro exposure to these drugs. These findings suggest that penicillin/β-lactamase inhibitor combinations could be a promising therapeutic candidate for treatment of a high proportion of methicillin-resistant S. epidermidis infections, although the in vivo risk of resistance development needs to be further addressed before they can be incorporated into clinical trials.

Fig. 1. Penicillin G susceptibility in the presence of 15 μg ml⁻¹ clavulanic acid.

a Broth microdilution determination of the minimum inhibitory concentration of penicillin G alone in a subset of 43 methicillin-resistant Staphylococcus epidermidis isolates, including 20 S2/S3 isolates and 23 R2 isolates. b Effect of clavulanic acid on susceptibility to penicillin. The Clinical and Laboratory Standards Institute breakpoint for penicillin G (≤0.125 μg ml⁻¹) is shown as grey vertical lines (a, b). PEN penicillin G, CLA clavulanic acid, MIC minimum inhibitory concentration. Source data are provided as a Source Data file.

Fig. 2. Amoxicillin susceptibility in the presence of 15 μg ml⁻¹ clavulanic acid.

a Broth microdilution determination of the minimum inhibitory concentration of amoxicillin alone in the 138 methicillin-resistant Staphylococcus epidermidis isolates, including 61 S2/S3 isolates and 77 R2 isolates. b Effect of clavulanic acid on susceptibility to amoxicillin. The empirical epidemiological cut-off for amoxicillin in the presence of 15 μg ml⁻¹ clavulanic acid (4.0 μg ml⁻¹) is shown as grey vertical lines (a, b). c Effect of amoxicillin alone or in combination with clavulanic acid against DEN09 in a mouse infection model. Mice were inoculated intramuscularly with around 8.5 × 10⁷ colony-forming units and treated subcutaneously with a single dose of the indicated drugs. Dunnett’s multiple comparisons tests were used to compare the effect of amoxicillin alone or in combination with clavulanic acid against vehicle treatment. Data are mean ± s.d. with n = 6 mice per group. d Population analysis profiles (PAPs) of DEN09 isolates recovered from mice treated with a single dose of vehicle alone (n = 3) or 250 mg kg⁻¹ amoxicillin in combination with 50 mg kg⁻¹ clavulanic acid (n = 3). A two-tailed unpaired Student’s t test was used to compare areas under the curve. Data are mean ± s.d. with n = 3 technical replicates per mouse. Bacterial colonies that were subjected to PAP testing are indicated in blue and orange (c, d). AMX amoxicillin, CLA clavulanic acid, VAN vancomycin, MIC minimum inhibitory concentration, c.f.u. colony-forming unit, AUC area under the curve. Source data are provided as a Source Data file.

Fig. 3. Correlation between the minimum inhibitory concentrations of amoxicillin/clavulanic acid and other β-lactam antibiotics.

Comparison of the minimum inhibitory concentrations (MICs) of amoxicillin in the presence of 15 μg ml⁻¹ clavulanic acid and the MICs of amoxicillin alone (a), clavulanic acid alone (b), penicillin G alone (c), and cefoxitin alone (d). Two-tailed Spearman correlation tests were used to compare MICs of different antibiotics. AMX amoxicillin, CLA clavulanic acid, PEN penicillin G, FOX cefoxitin, r Spearman correlation coefficient. Source data are provided as a Source Data file.

Fig. 4. Phylogenetic tree of 138 methicillin-resistant Staphylococcus epidermidis BPH0662 clone, ST2-mixed, ST5, and ST23 isolates from Australia, Denmark, and Germany.

The maximum-likelihood phylogeny was built from a core-genome single-nucleotide polymorphism (SNP) alignment (3102 SNPs) after putative recombination sites were removed. The tree was rooted at the midpoint. Branch support values above 90% are indicated by filled circles at the nodes. Isolates were deemed phenotypically resistant or susceptible to amoxicillin in the presence of 15 μg ml⁻¹ clavulanic acid if they had minimum inhibitory concentrations above or at/below the empirical epidemiological cut-off (4.0 μg ml⁻¹), respectively. The scale bar represents the number of nucleotide substitutions per variable site. MLST multilocus sequence type, AMX amoxicillin, CLA clavulanic acid.

Fig. 5. Resistance development.

Serial passaging of BPH0719 (a) and DEN09 (b) in 15 μg ml⁻¹ clavulanic acid with increasing concentrations of amoxicillin and of vancomycin and rifampicin alone over 30 days. Representatives of up to six independent serial passaging experiments (E1–E6) are shown. Minimum inhibitory concentrations and growth rates of BPH0719 (c) and DEN09 (d) and their descendants during serial passaging in 15 μg ml⁻¹ clavulanic acid with increasing concentrations of amoxicillin and of vancomycin and rifampicin alone over 30 days. The descendants are named after the parent strain (e.g., BPH0719), the experiment (e.g., E1), and the day of isolation (e.g., D01). AMX amoxicillin, CLA clavulanic acid, PEN penicillin G, FOX cefoxitin, VAN vancomycin, RIF rifampicin, C_max the highest drug concentrations that allowed growth after 24 h of incubation at 37 °C. Source data are provided as a Source Data file.

Fig. 6. Mutational trajectories.

Mutations in serially passaged descendants of BPH0719 (a) and DEN09 (b) associated with increased resistance to amoxicillin in the presence of 15 μg ml⁻¹ clavulanic acid. Descendants are named after the parent strain (e.g., BPH0719), the experiment (e.g., E1), and the day of isolation (e.g., D01). Amoxicillin/clavulanic acid minimum inhibitory concentrations and growth rates are indicated. Boxes show open reading frames (ORFs) and intergenic regions (IRs) containing single-nucleotide polymorphisms. ORFs and IRs that were altered in more than one of the evolving populations are indicated in coloured boxes. ORFs were named through RPS-BLAST searches against the NCBI Conserved Domain Database (version 3.20), setting the E-value cut-off at 1 × 10^-25. A single ORF did not produce hits and was therefore named ORF-1. Amino acid changes are included in parentheses. Stop codons are represented by asterisks. Three independent serial passaging experiments (E1–E3) are shown. Source data are provided as a Source Data file.

Fig. 7. Assessment of biofilm formation and anti-biofilm activity of amoxicillin in the presence of clavulanic acid.

Assessment of biofilm formation by the 138 methicillin-resistant Staphylococcus epidermidis (MRSE) isolates (a) and in a subset of 71 MRSE isolates that were phenotypically susceptible to amoxicillin/clavulanic (b). The ability to form biofilms (strong, moderate, weak, and none) was categorised using a previously described scoring system. Dose-response effect of amoxicillin in the presence of 15 μg ml⁻¹ clavulanic acid against BPH0719 allowed to form biofilms for 6-h (c) and 24-h (d). The y axes show the amounts of biomass as percentages of the results for the matching control (c, d). Data are mean ± s.d. with n = 3 technical replicates. AMX amoxicillin, CLA clavulanic acid. Source data are provided as a Source Data file.