Abstract
Aim
There is growing evidence that bacteria encountered in periprosthetic joint infections (PJI) form surface-attached biofilms on prostheses, as well as biofilm aggregates embedded in synovial fluid and tissues. However, models allowing the investigation of these biofilms and the assessment of their antimicrobial susceptibility in physiologically relevant conditions are currently lacking. To address this, we developed a synthetic synovial fluid (SSF) model and we validated this model in terms of growth, aggregate formation and antimicrobial susceptibility testing, using multiple PJI isolates.
Methods
17 PJI isolates were included, belonging to Staphylococcus aureus, coagulase negative staphylococci, Cutibacterium acnes, Pseudomonas aeruginosa, enterococci, streptococci, Candida species and Enterobacterales. Growth and aggregate formation in SSF, under microaerophilic or anaerobic conditions, were evaluated using light microscopy. The biofilm preventing concentration (BPC) and minimum biofilm inhibitory concentration (MBIC) of relevant antibiotics (doxycyclin, rifampicin and oxacillin) were determined for the staphylococcal strains (n=8). To this end, a high throughput approach was developed, using a fluorescent viability resazurin staining. BPC and MBIC values were compared to the minimum inhibitory concentration (MIC) obtained with conventional methods.
Results
The SSF model allowed all isolates to grow well under microaerophilic or anaerobic conditions. When cultured in SSF, all isolates formed biofilm aggregates, varying in size and shape along different species. A susceptibility testing method based on measuring resazurin-derived fluorescence was successfully developed, allowing high throughput determination of the BPC and the MBIC in SSF. For all staphylococci cultured in SSF a reduced susceptibility to the tested antibiotics was observed when compared to susceptibility data obtained in general medium. For rifampicin and doxycyclin the BPC was consistently higher than the MIC (two- to fourfold dilution difference for rifampicin and four- to sixfold dilution difference for doxycyclin). For oxacillin the MIC equaled the BPC for two isolates, while for the other isolates the BPC was higher than the MIC (two- to fourfold dilution difference). Expectedly, the MBIC was higher than the BPC and differences with the MIC were even more pronounced for all antibiotics tested (differences of six- to fourteenfold dilutions were observed).
Conclusion
Our data indicate that the in vitro SSF model could provide more insight in how PJI-related pathogens form biofilms in physiologically relevant conditions. The BPC and MBIC were consistently and substantially higher than MIC. This model could be a valuable addition to evaluate the antimicrobial susceptibility in biofilms in a PJI context.
Sources of funding: FWO-Vlaanderen (grant G066523N).
