A P. aeruginosa strain, which eventually caused a fatal epidemic cluster, was able to contaminate and likely replicate within a triclosan-soap dispenser. Triclosan is an anionic, lipophilic compound that is very poorly soluble in water . However, previous researchers have reported that various solubilizers, such as the surfactant sodium lauryl sulphate, increase the solubility of triclosan from 80- to 6,000-fold. Micellar solubilization and the formation of either salts or complexes are postulated as possible mechanisms for the increase in the solubility of triclosan . In this study, we decided to perform the antimicrobial susceptibility tests and the time kill assays using the commercial soap product A, in which triclosan (5,000 mg/L) is kept in solution by 10% sodium lauryl-myristyl ether sulfate. Assuming that this surfactant, or any other additive of product A, may have increased the effectiveness of triclosan in killing P. aeruginosa, this would paradoxically imply that the actual P. aeruginosa resistance to triclosan alone would be higher than that observed in our experimental system.
Determination of the bactericidal activity for triclosan confirmed that the epidemic strain could have contaminated the soap dispenser most probably because of its remarkable tolerance to triclosan. In fact this epidemic strain was originally isolated at relatively high concentration (5 × 104 CFU/mL) from the dispenser containing the triclosan-based commercial soap (product A) for healthcare workers' hand washing . These findings provide additional biological evidence to confirm our previous hypothesis, based on molecular identity between patients' and soap P. aeruginosa isolates , that the contaminated soap dispenser may have acted as the primary source of infection. We can also suppose that the activity of triclosan in the commercial formulation was reduced by partial inactivation of triclosan by micellar entrapping on inclusion which may have reduced the amount of free triclosan available in solution.
Conversely, the chlorhexidine-based disinfectant, which was employed in rotation of 3 months with triclosan in the oncohematology unit , had strong and fast bactericidal effect at concentrations far below those recommended by the manufacturer for practical use.
Another interesting result was that the level of resistance to some antibiotics, namely tetracycline, fluoroquinolones, aminoglycosides, carbenicillin and chloramphenicol, was increased by previous adaptation of P. aeruginosa to triclosan. As reported in other studies [10, 16], triclosan may induce the expression of multidrug efflux systems belonging to RND family in P. aeruginosa. These systems are relatively nonspecific in substrate recognition and enables bacteria to pump out numerous chemically unrelated substances, such as antibiotics and biocides. Here, we report that the pattern of increasing resistance to antibiotics was likely to be driven by activation of efflux pumps belonging to the RND family; in fact, the RND efflux pump inhibitor PAβN and the protonophore CCCP restored the baseline antibiotic susceptibility in the triclosan-adapted strain. Thus, while the link between triclosan usage and development of clinically significant clinical resistance to antibiotics remains controversial [9, 10], here we provide evidence that previous exposure to triclosan elevates the level of antibiotic resistance. Although the differences in MICs obtained for the different antibiotics are small (2-fold), they provide a clear evidence of the overall trend towards increasing antibiotic resistance associated with adaptation to tricosan. Increased resistance was also observed for those antibiotics to which P. aeruginosa L2 cells showed elevated baseline resistance. As noted for carbenicillin, adaptation to triclosan converted the susceptible phenotype (MIC ≤ 128 mg/L) into intermediate resistant (MIC = 256 mg/L).
To the best of our knowledge, this is the first time that such a high tolerance to triclosan is documented in a clinical P. aeruginosa isolate, since previous studies on both uncharacterized and type (PAO1) P. aeruginosa strains reported actual MICs of ~ 1,000 mg/L .
Some issues limit the generalizability our findings. Firstly, our data come from a single epidemic event, and therefore they may not apply in the same way to seemingly similar episodes in other institutions or to general clinical practices as biological features of the outbreak strain (e.g. virulence, infectivity, and/or baseline resistance patterns) and/or local situation (e.g. infection control policies, mode of transmission, and/or risk factors) may be unique. Secondly, we could not quantify the actual triclosan concentration in the triclosan-based commercial soap formulation - as it was in the soap dispenser at the time of the outbreak - since the oncohematology unit was closed due to legal issue, and the only soap sample obtained was insufficient for chemical analysis. Therefore we cannot rule out that accidental dilution (below the commercial concentration) of triclosan soap due to refilling procedure might have played a preeminent role in facilitating the growth of P. aeruginosa in the dispenser and the eventual spreading of the infection. Finally, we only performed phenotypic assays, and we cannot provide a genetic basis for the remarkably high resistance to triclosan of the P. aeruginosa outbreak strain. Intrinsic outer membrane impermeability, resistance of the FabV enoyl-acyl carrier protein reductase, expression of an additional triclosan-specific pump (TriABC-OpmH) or other unknown mechanisms could have acted synergistically with efflux in determining the unprecedented triclosan resistance observed our isolates.
Despite these limitations, this study provide biological evidence to confirm that an unprecedented environmental source, which was identified as a triclosan soap dispenser, actually acted as a continuous source of infection during the outbreak. In addition, we provide hints that exposure to sub-lethal concentration of triclosan may contribute to antibiotic resistance in P. aeruginosa clinical isolates.