Biofilm age
Several biofilm models have been developed, each with many experimental parameters that can be adjusted. This flexibility inevitably makes it difficult to compare results obtained with varying conditions chosen in different studies. In the current study we confirmed previous findings that mature biofilm have reduced antibiotics susceptibility compared with young biofilms (22–25). However, no definition of young and mature biofilm has been universally adopted. In the case of P. aeruginosa, some considered 4 h biofilm as young and 24 h as mature (26), while others considered 24 h as young and 12 days biofilm as mature (24). Similarly, 6 h S. aureus biofilm was considered as young and 24 h as mature (27), whereas some considered 7 days old biofilm as mature (28). The inconsistency in the different biofilm studies underlines the need for a form of consensus definition and a simple way to measure maturity. The textbook version of biofilm formation involves bacterial initial attachment to a solid surface, the formation of microcolonies on the surface, and finally differentiation of microcolonies into exopolysaccharide-encased, mature biofilms. However, studies often assume the maturity of the biofilm without looking into the structure of the biofilms or even CFUs of biofilm. In the case of MIC testing, a crucial parameter is inoculum size which is set to be 5 x 105 CFU mL-1. It is because MIC values can increase concurrently with increasing number of CFUs (29).
The current study treated 24 h biofilms as young and 72 h as mature. The CFU per biofilm shown in Figure S1 and S2 indicated continuous growth in cell number after 24 h for up to 1-log. In batch culture, bacterial growth curve defines the different stages of planktonic culture growth. Similarly, the biofilm formation curves can be established for each strain and growth condition. It was shown previously (30) that using CBD the number of E. coli ATCC 25922 and P. aeruginosa ATCC 27853 continuously increased over 24 h while the growth of S. aureus stagnated after 7 h under the same condition. As the growth phase of the biofilm influences antimicrobial susceptibility, it is therefore important to construct the biofilm growth curve for each strain under the chosen conditions.
Growth media
Biofilm eradication was found different with the two media (Figure 1 and Figure 2). Different composition of media is reported to change the activity of antibiotics (31–33). The Ca2+ and Mg2+ ions in CaMHB media are required for a correct antimicrobial susceptibility testing because those ions reflect the divalent cation concentration in human blood (34–37). Neither of the two tested media, TSB and CaMHB, resembles in vivo conditions. However, use of CaMHB makes it possible to compare with MIC results, while TSB has been frequently used in other publications (Tables 1 and 2). Other media such as brain-heart infusion broth (38), TSB supplemented with glucose (39), LB (40), and chemically defined media such as basal medium 2 and M9 minimal media (41) have also been used in studies. The choice of media is known to affect biofilm formation (42, 43), but a standardized medium to assess the activity of antibiofilm agents has not been established. It is difficult to standardize because the in vivo environment of biofilm infections varies depending on the location of the infection, hence the optimal medium should be developed for each infection, for example, medium supplemented with mucin for studying cystic fibrosis lung infection (44), or saliva containing medium for studying oral biofilms (45), or human urine for urinary tract infections (46, 47). Besides nutrient source, in vivo conditions are far more complex with presence of immune systems and varying oxygen level etc., the antibiotics concentration needed for biofilm eradication will most likely be different from in vitro results. For comparison across different studies, a simple and widely available culture medium is suitable, but for estimation of in vivo biofilm killing host factors in form of, for example, serum, plasma, or blood should be included in testing medium.
Antibiotics exposure time
Vancomycin displayed a time-dependent eradication of S. aureus biofilms (Table 2) which has been demonstrated in other studies (28, 48, 49). Post et al. have shown continuous reduction of viable S. aureus biofilm cells over 28 days (28). This indicates that further killing could be possible by prolonging the antibiotic exposure time in the current study and complete eradication could be achieved at lower vancomycin concentration.
In contrast to vancomycin, tobramycin exhibits concentration-dependent activity (50–53). The current study indicated that tobramycin displayed concentration-dependent activity for 24 h PA14 biofilms. However, increased killing of 72 h biofilms were observed with prolonged exposure. Castaneda et al. found increased biofilm antimicrobial susceptibility with increasing antimicrobial exposure time including tobramycin against P. aeruginosa biofilms (54), whereas Walters et al. only found little reduction in P. aeruginosa biofilm cell count with longer tobramycin treatment (55). Futures studies are needed to investigate the time-dependency of tobramycin antibiofilm effect.
Regardless of the antimicrobials being time-dependent or concentration-dependent on planktonic bacteria, it may be different on biofilm cells because of the presence of biofilm matrix. Exposure time may play an important role in determination of killing effect, because the biofilm matrix may slow down antimicrobial penetration (56). Therefore, a killing curve is much more informative than a definitive MBEC value determined at a fixed time point.
OSTEOmycinTM
Since the antibiotic concentration needed for biofilm eradication is far above the parenterally administrated levels, local delivery of antibiotics may achieve concentrations high enough for biofilm killing. In this study, OSTEOmycinTM showed a strong biofilm eradication efficacy and completely removed biofilm in all tested conditions except three 72 h S. aureus biofilms. OSTEOmycinTM is a product developed based on Winkler et al. 2000 (20). According to the study, 1 g human cancellous bone impregnated with vancomycin released around 20000 mg L-1 vancomycin in 3 mL of 5% human albumin solution after one day and decreased to around 100 mg L-1 after seven days. Accordingly, it implies that approximately 16800 mg L-1 of vancomycin after one day and 84 mg L-1 after seven days were released with the applied amount in this study. When impregnated with tobramycin, it released more than 10000 mg L-1 tobramycin after one day and decreased to around 30 mg L-1 after seven days (20), suggesting 6600 mg L-1 of vancomycin after one day and 19.8 mg L-1 were released after seven days with the applied amount in this study. These concentrations are much higher than the MBEC values found in Figure 1 and 2, which likely explains the high efficacy. This indicates that prolonged antibiotics treatment may not be necessary when sufficiently high concentration of antibiotics is administered in the beginning of treatment. OSTEOmycin was also shown to be efficient for local treatment of osteomyelitis in the clinic although recurrence may still occur in complex cases within an unknown period of time (19). The limitation of this study is that OSTEOmycinTM was not tested in a medium resembling the nutrient composition in the bone under in vivo like conditions, and more clinically used or candidate antibiofilm products for osteomyelitis could have been evaluated, such as antibiotics impregnated cement or hydrochlorous acid. Ideally, the in vitro effect of these antibiofilm products could be compared to clinical outcome to validate the assay. Assays developed in such as way could be used to guide the dose of antibiotics for clinical application.
In this study, the used conditions (nutrient rich media, pH, atmospheric oxygen level, shear, biofilm growth in static system, mono species biofilm etc.) were not specific for a distinct biofilm infection and more suitable for initial testing of antibiofilm product. It was by no means meant as a standardization or guideline for clinical application. The purpose was to raise the awareness that biofilm eradication depends on many factors, including the ones mentioned here, but also pH, oxygen level, temperature, shear, and complicated by polymicrobial community interactions and the presence of human factors such as the human immune systems. For specific biofilm infections, we think it is necessary to develop assays with in vivo like environment and validate obtained results by comparing with clinical outcome.