Imipenem-induced viscoelastic reinforcement drives physical tolerance in Pseudomonas aeruginosa biofilms.

Sub-inhibitory antibiotic exposure can alter biofilm architecture, but the effects on matrix mechanics remain poorly understood. Here, imipenem at half the minimum inhibitory concentration (MIC) is shown to induce viscoelastic reinforcement and physical tolerance in Pseudomonas aeruginosa (P. aeruginosa) biofilms. Ten imipenem-resistant isolates from ready-to-eat foods were cultured with and without sub-inhibitory imipenem. Treated biofilms transitioned from smooth, flat colonies to rugose, wrinkled, hyper-hydrophobic architectures, with contact angle increasing from 37.5° to 90.2°. Capillary rheometry revealed a 3.4-fold increase in apparent viscosity (185.6 vs. 54.2 mPa s) and pronounced shear-thinning behavior. Creep-recovery tests showed lower creep compliance (0.72 vs. 1.65 Pa-1) and higher elastic recovery (48.3% vs. 17.5%), indicating a stiffer, more cross-linked matrix. Biofilm inhibitory concentrations (BIC) exceeded 128 μg/mL in all 10 imipenem-treated rugose phenotypes (range >128 μg/mL for each isolate), despite planktonic MICs of only 8-32 μg/mL. Using the Stokes-Einstein equation, the 3.4-fold viscosity increase was calculated to proportionally reduce the imipenem diffusion coefficient from 1.5 × 10-11 to 4.2 × 10-12 m2/s (71% reduction), directly limiting antibiotic flux into deep biofilm layers. Imipenem-treated biofilms required 3.6-fold higher shear stress for initial detachment and 2.7-fold higher peeling force (33.7 vs. 12.4 N/m), with failure mode shifting from adhesive to cohesive. Notably, imipenem MIC did not correlate with viscosity parameters (r = 0.24, p > 0.05), demonstrating that mechanical reinforcement operates independently of genetic resistance. These in vitro findings highlight the potential significance of matrix mechanics in biofilm persistence and provide a mechanistic explanation for reduced antimicrobial efficacy in biofilm-associated contexts. Conventional susceptibility testing fails to capture this adaptive response.