|Home | About | Journals | Submit | Contact Us | Français|
We used a murine model of catheter-associated biofilm formation to determine whether the mutation of the staphylococcal accessory regulator (sarA) has an impact on the susceptibility of established Staphylococcus aureus biofilms to treatment with daptomycin in vivo. The experiments were done with two clinical isolates, one of which (UAMS-1) was obtained from the bone of a patient suffering from osteomyelitis, while the other (UAMS-1625) is an isolate of the USA300 clonal lineage of community-acquired methicillin (meticillin)-resistant S. aureus. UAMS-1625 had a reduced capacity to form a biofilm in vivo compared to that of UAMS-1 (P = 0.0015), but in both cases the mutation of sarA limited biofilm formation compared to that of the corresponding parent strain (P ≤ 0.001). The mutation of sarA did not affect the daptomycin MIC for either strain, but it did result in increased susceptibility in vivo in the context of an established biofilm. Specifically, daptomycin treatment resulted in the clearance of detectable bacteria from <10% of the catheters colonized with the parent strains, while treatment with an equivalent daptomycin concentration resulted in the clearance of 46.4% of the catheters colonized with the UAMS-1 sarA mutant and 69.1% of the catheters colonized with the UAMS-1625 sarA mutant. In the absence of daptomycin treatment, mice with catheters colonized with the UAMS-1625 parent strain also developed skin lesions in the region adjacent to the implanted catheter. No such lesions were observed in any other experimental group, including untreated mice containing catheters colonized with the UAMS-1625 sarA mutant.
A primary concern with all bacterial pathogens is the continued emergence of antibiotic-resistant strains. This is perhaps most apparent in Staphylococcus aureus, as evidenced by the current predominance of methicillin (meticillin)-resistant strains even among community-acquired isolates (27). However, many S. aureus infections are recalcitrant to antimicrobial therapy even in the absence of acquired resistance (14). One factor that contributes to this recalcitrance is the formation of a bacterial biofilm on host tissues and indwelling medical devices (20, 33). This biofilm not only limits the efficacy of antimicrobial therapy but also provides some level of intrinsic resistance to host defenses (12, 32). For this reason, the effective treatment of biofilm-associated S. aureus infections often requires surgical intervention to remove infected tissues and/or indwelling devices (6, 19).
The growth of bacteria in a biofilm is a lifestyle option rather than a necessity, and in this context it is unlikely that any intervention strategy targeting biofilm formation per se would be therapeutically effective in and of itself. Nevertheless, such strategies have the potential to limit the intrinsic resistance associated with the biofilm mode of growth and thereby enhance the efficacy of conventional antimicrobial therapy. The studies described in this report were aimed at testing this hypothesis. We placed a specific emphasis on daptomycin based on reports concluding that it retains activity against slowly or even nongrowing bacteria (21) and has greater activity than other antibiotics in the context of an established biofilm (18, 25, 42).
Biofilm formation in S. aureus is a complex process that is influenced by numerous genes. However, in many cases, the impact of individual genes is strain dependent (10, 23). For this reason, we chose to focus our efforts on the staphylococcal accessory regulator (sarA), the mutation of which has been shown to limit biofilm formation both in vivo and in vitro not only in S. aureus but also in S. epidermidis (1, 11, 17, 28, 34, 35, 37, 39). To avoid any bias associated with strain choice, we included two genetically distinct S. aureus clinical isolates and their isogenic sarA mutants. One of these (UAMS-1) is a well-characterized, biofilm-positive osteomyelitis isolate (2, 8, 30), while the other (UAMS-1625) is a methicillin-resistant S. aureus (MRSA) isolate of the USA300 clonal lineage that was isolated from a patient with a fatal brain abscess (29). The experiments employed three different biofilm models, with a primary emphasis in the context of daptomycin susceptibility on an in vivo, murine model of catheter-associated biofilm formation (8, 26). The results confirmed that the mutation of sarA limited biofilm formation in both strains and that this limitation was correlated with increased daptomycin susceptibility in vivo in the specific context of an established biofilm.
The S. aureus osteomyelitis isolate UAMS-1 and its isogenic sarA strain (UAMS-929) were described elsewhere (3). The USA300 isolate used in these experiments (designated here UAMS-1625) was isolated from a patient that died from a brain abscess (29). Because this isolate is resistant to kanamycin but sensitive to tetracycline (data not shown), the UAMS-1625 sarA mutant (UAMS-1653) was generated by 11-mediated transduction from UAMS-950 (3) rather than UAMS-929. Transductants were screened by PCR using primers that flank the insertion site of the resistance genes within the sarA open reading frame (3). The identity of the parent strains and their isogenic sarA mutants was confirmed by the pulsed-field gel electrophoresis of SmaI-digested genomic DNA (31). The complementation of the sarA mutations in UAMS-929 and UAMS-1653 was done as previously described (1, 3).
The production of SarA was assessed by the Western blotting of whole-cell lysates using rabbit polyclonal anti-SarA serum as previously described (4). To assess hemolytic activity and estimate the role of individual hemolysins, each strain was grown on rabbit or sheep blood agar using the laboratory strain RN4220 as a reporter as described by Traber et al. (36). Specifically, because RN4220 produces only beta toxin, and because beta toxin and alpha toxin are functionally antagonistic while beta toxin and delta toxin are functionally synergistic (36), perpendicular streaks of the test strain and RN4220 allows an estimation of the activity of individual S. aureus hemolysins. This is particularly true in this case, because both UAMS-1 and UAMS-1625 are hlb lysogens (data not shown) and consequently do not produce beta toxin. To directly assess the impact of sarA on the production of alpha toxin, supernatants from overnight (16 h) cultures grown in tryptic soy broth (TSB) were filter sterilized and resolved by sodium dodecyl sulfate-polyacrylamide gel electrophoresis prior to Western blot analysis using a commercially available rabbit, anti-staphylococcal alpha toxin antiserum (Sigma Chemical Co., St. Louis, MO) and a horseradish peroxidase-conjugated goat anti-rabbit secondary antibody (Jackson ImmunoResearch, West Grove, PA). Nonspecific binding was blocked using human immunoglobulin G prepared in 0.5% nonfat dry milk (Sigma Chemical Co.). The MIC of daptomycin was determined by Etest as described by the manufacturer (AB Biodisk, Solna, Sweden), except that tryptic soy agar (TSA) was used as the test medium.
Biofilm formation initially was assessed using a microtiter plate assay as previously described (1, 8). Daptomycin susceptibility in the context of an established biofilm was assessed using a catheter-associated model of biofilm formation (42). Briefly, 1-cm segments of fluorinated ethylene propylene (FEP) catheters (14 gauge; Introcan Safety, B. Braun, Bethlehem, PA) first were coated with human plasma (1) and then placed in the wells of a 12-well microtiter plate containing 2 ml of TSB supplemented with glucose and sodium chloride (biofilm medium) as previously described (1, 38, 42). Each well then was inoculated with the test strain at an optical density at 600 nm of 0.05. After overnight incubation at 37°C, catheters were removed, rinsed in phosphate-buffered saline (PBS) to remove nonadherent bacteria, and then placed into fresh biofilm medium supplemented with 2.5 mM CaCl2 with or without the indicated amounts of daptomycin.
The daptomycin concentrations used in these experiments were 5.0, 10.0, and 20.0 μg/ml, which correspond to 5, 10, and 20 times the concentration defined by the Clinical and Laboratory Standards Institute (CLSI) as the breakpoint MIC for a daptomycin-sensitive strain of S. aureus (≤1.0 μg/ml). Catheters (n = 3) were recovered at daily intervals for 3 days, with the spent medium being replaced in its entirety with fresh medium each day. After recovery at each time point, catheters were rinsed in PBS to remove nonadherent bacteria, placed into a test tube containing 5 ml of sterile PBS, and sonicated to remove adherent bacteria (26). Aliquots (100 μl) of appropriately diluted samples then were plated on TSA and incubated overnight at 37°C. The total number of bacteria recovered from each catheter then was calculated based on the number of colonies obtained and the corresponding dilution factor.
Biofilm formation and daptomycin susceptibility were assessed in vivo using a murine model of catheter-associated biofilm formation (2). Specifically, a 1-cm FEP catheter segment was placed into a subcutaneous pocket generated in each flank of NIH Swiss mice. Both to avoid an immune response to human proteins and because the catheter presumably would be coated with the corresponding murine proteins once it was implanted, catheters in these experiments were not precoated with human plasma. After the implantation of all catheters (approximately 1 h), 105 CFU of the test strain in a total volume of 100 μl of PBS was introduced directly into the lumen of each catheter. After 24 h, mice were randomly divided into experimental groups (n = 15). Because each mouse had two catheters implanted, and because preliminary experiments confirmed the absence of cross-contamination between catheters in opposite flanks of the same mouse (data not shown), each catheter subsequently was treated as an independent data point (n = 30).
In the untreated groups, 100 μl of sterile PBS was injected in the lumen of each catheter at daily intervals. In the treated groups, 100 μl of sterile PBS containing the indicated amounts of daptomycin was injected into the lumen, also at daily intervals. Based on the results of our in vitro experiments demonstrating that biofilms formed by UAMS-1625 were more susceptible than those formed by UAMS-1 (see below), in vivo experiments with UAMS-1625 and its sarA mutant (UAMS-1653) were done with a daptomycin concentration corresponding to 10× the CLSI-defined breakpoint MIC, while those with UAMS-1 and its sarA mutant (UAMS-929) were done with a concentration corresponding to 20× the MIC. In all cases, treatment was continued daily for 7 days, at which point catheter segments were harvested, rinsed in sterile PBS to remove nonadherent bacteria, and sonicated in a total volume of 5.0 ml of PBS to remove adherent bacteria (26). The total number of viable bacteria colonizing each catheter then was determined by plating appropriately diluted samples on TSA as described above.
Bacterial count data were analyzed using both parametric and nonparametric methods. Specifically, Wilcoxon rank-sum tests were used to make comparisons between untreated samples, while Fisher's exact tests were used to compare clearance percentages among the daptomycin-treatment groups. We also performed analysis of variance (ANOVA) on the logarithmically transformed bacterial count data to evaluate the effect of the sarA mutation in the context of daptomycin exposure. Using these ANOVA models, we were able to assess not only the effect of the sarA mutation but also any interaction between the effect of the mutation and daptomycin exposure. The significance of the ANOVA test statistics were calculated using permutation tests. Pair-wise testing was performed using t tests on the logarithmically transformed data, and P values were calculated using permutation tests. All statistical analyses were performed using R (version 2.7; The Foundation for Statistical Computing), with P values ≤0.05 considered significant.
The mutation of sarA and the identity of each mutant with respect to its corresponding parent strain was confirmed by the PCR analysis of the sarA gene itself, Western blot analysis of cell-free lysates using an anti-SarA antibody, and pulsed-field gel electrophoresis (data not shown). The mutation of sarA in both UAMS-1 and UAMS-1625 resulted in a reduced capacity to form a biofilm as assessed using a microtiter plate assay, and this effect was reversed in both strains by the introduction of a plasmid-borne copy of the functional sarA gene (Fig. (Fig.1).1). The mutation of sarA had no impact on the daptomycin MIC for either strain as determined by E-strip susceptibility testing (data not shown). The MIC for UAMS-1625 and its sarA mutant was lower than that for UAMS-1 and its sarA mutant (0.38 and 0.50 μg/ml, respectively), but in both cases the MIC was below the CLSI-defined breakpoint for a daptomycin-sensitive strain of S. aureus (≤1.0 μg/ml).
Because the microtiter plate assay does not lend itself to the assessment of antibiotic susceptibility, we addressed this issue in the context of daptomycin using a catheter-associated model of biofilm formation (42). When the ability of each parent strain to form a biofilm was assessed in vitro, UAMS-1 and UAMS-1625 were found to colonize catheters initially to a comparable degree (1.8 × 108 and 5.47 × 107 CFU per catheter, respectively), but the number of bacteria colonizing each catheter increased over time with UAMS-1 but remained essentially unchanged with UAMS-1625 (Fig. (Fig.2).2). By day 1, the difference had become statistically significant (P = 0.037). However, even on day 3 the difference was less than 10-fold (3.20 × 108 and 8.90 × 107 CFU/catheter, respectively). Nevertheless, this difference appeared to be relevant with respect to daptomycin susceptibility, in that the lowest counts observed with catheter-associated biofilms formed by UAMS-1 and exposed to daptomycin were between 101 and 102 CFU/catheter, and achieving this degree of clearance required exposure to ≥10.0 μg/ml (10× MIC) for 3 days (Fig. (Fig.2,2, left). Moreover, none of the catheters colonized with UAMS-1 were cleared even after continuous exposure to 20 μg/ml of daptomycin for 3 days. In contrast, similar counts were observed with UAMS-1625 biofilms after exposure to daptomycin at a concentration as low as 5.0 μg/ml, and all catheters were cleared by exposure to 20 μg/ml for 3 days (Fig. (Fig.2,2, right).
We also found that UAMS-1625 had a reduced capacity to form a biofilm compared to that of UAMS-1 in our in vivo experiments (P = 0.0015) (Fig. (Fig.3).3). However, the negative impact of mutating sarA on biofilm formation was evident in vivo in both UAMS-1 (P = 0.001) (Fig. (Fig.4A)4A) and UAMS-1625 (P < 0.001) (Fig. (Fig.4B),4B), and in both cases this was correlated with increased susceptibility to daptomycin. The latter was reflected both in the average counts for each experimental group (Fig. (Fig.4)4) and in the fact that only 2 of 30 catheters (6.7%) colonized with UAMS-1 and treated with 20× daptomycin and only 1 of 16 (6.3%) colonized with UAMS-1625 and treated with 10× daptomycin were cleared of bacteria (Fig. (Fig.5).5). In contrast, 13 of 28 (46.4%) colonized with UAMS-929 and 20 of 29 (69.0%) colonized with UAMS-1653 were cleared by treatment with daptomycin at the same concentrations as those used with their respective parent strains (Fig. (Fig.5).5). Statistical analysis confirmed a significant difference between the parent strains and their isogenic sarA mutants with respect to the percentage of catheters cleared by exposure to equivalent concentrations of daptomycin (P < 0.001).
Finally, 18 of the 30 catheters (60%) colonized with UAMS-1625 and left untreated were lost by the end of the 8-day infection period. This was correlated with the formation of skin lesions in the area immediately adjacent to the implantation site (Fig. (Fig.6).6). We did not observe the development of such lesions or this degree of catheter loss in any of the other experimental groups, including those infected with UAMS-1625 but treated with daptomycin (data not shown) or those in the untreated control group infected with the UAMS-1625 sarA mutant (Fig. (Fig.66).
One possible explanation for this is the impact of sarA on hemolytic activity. Specifically, UAMS-1625 was highly hemolytic on both rabbit and sheep blood agar (Fig. (Fig.7).7). UAMS-1653 also was hemolytic, but its activity was reduced compared to that of the UAMS-1625 parent strain, particularly on sheep blood agar. On rabbit blood, hemolytic activity in both UAMS-1625 and UAMS-1653 was inhibited in the region immediately adjacent to the RN4220 reporter strain (Fig. (Fig.7).7). Because RN4220 produces only beta toxin, and because beta and alpha toxins are functionally antagonistic (36), this suggests that the mutation of sarA resulted in the reduced production of alpha toxin in UAMS-1653. This was confirmed by Western blot analysis demonstrating that supernatants from UAMS-1653 contain reduced amounts of alpha toxin compared to those of UAMS-1625 (Fig. (Fig.7).7). The complementation of the sarA mutation in UAMS-1653 with a plasmid-borne copy of sarA restored alpha toxin production to wild-type levels, as assessed by Western blotting (Fig. (Fig.7,7, right).
In contrast to the antagonism between alpha and beta toxins, S. aureus delta and beta toxins are synergistic (36), and UAMS-1625 was found to produce an enhanced zone of hemolysis on sheep blood agar in the region immediately adjacent to the RN4220 cross-streak (Fig. (Fig.7).7). This zone was reduced in UAMS-1653, which suggests that the mutation of sarA in UAMS-1625 also resulted in the reduced production of delta toxin. This overall pattern of hemolytic activity also was restored by the introduction of a functional sarA gene into UAMS-1653 (Fig. (Fig.7,7, left). To the extent that both delta toxin and alpha toxin are produced under the regulatory control of agr, these results suggest that the failure of UAMS-1653 to produce skin lesions in the region adjacent to the implanted catheters are a function of the reduced production of alpha toxin, possibly as a result of the impact of sarA on the expression of agr.
UAMS-1 was nonhemolytic on sheep blood agar and minimally hemolytic on rabbit blood (Fig. (Fig.7).7). However, hemolytic activity on rabbit blood was increased in UAMS-929 compared to that of the UAMS-1 parent strain. No alpha toxin was detected in Western blots of supernatants from UAMS-1 or its sarA mutant (Fig. (Fig.7),7), which is consistent with the observation that UAMS-1 and other closely related strains (e.g., MRSA252) have a nonsense mutation in hla that precludes the production of functional alpha toxin (8). Unlike UAMS-1625, we did not observe an RN4220-associated zone of enhanced hemolysis on sheep blood agar with either UAMS-1 or its sarA mutant (Fig. (Fig.7).7). These results demonstrate that, unlike UAMS-1625, the mutation of sarA results in increased hemolytic activity in UAMS-1. However, the absence of an enhanced zone of hemolysis adjacent to the RN4220 cross-streak on sheep blood agar, together with the observation that UAMS-1 is an hlb lysogen (3) and has a nonsense mutation in hla, suggests that the increased hemolytic activity observed in the UAMS-1 sarA mutant is a function of something other than alpha, beta, or delta toxin.
Biofilm-associated S. aureus infections are particularly problematic, because the biofilm compromises antimicrobial therapy even in the absence of acquired resistance (6, 20). This is evidenced by the fact that resolving such infections often requires the removal of infected devices and/or surgical intervention to debride infected tissue (6, 19). The intrinsic resistance of S. aureus within an established biofilm also is evident in the observation that, in our in vitro experiments, which effectively eliminated issues related to antibiotic delivery, none of the catheters colonized with UAMS-1 were cleared of detectable bacteria, even when exposed for as long as 3 days to daptomycin concentrations 20-fold higher than the CLSI-defined breakpoint MIC for a sensitive strain of S. aureus (≤1.0 μg/ml). The clearance of catheters colonized with UAMS-1625 was observed in vitro, but only after 3 days of exposure to a daptomycin concentration 10-fold higher than the CSLI-defined breakpoint.
Based on such results, we believe that the development of adjunct therapies capable of enhancing the efficacy of conventional antimicrobial agents in the specific context of an established S. aureus biofilm is a viable clinical objective. This belief was the motivation behind these experiments. Specifically, we used sarA mutants as a model to assess whether an effective inhibitor of biofilm formation in S. aureus would enhance the efficacy of conventional antimicrobial therapy to the point that surgical intervention would become an option rather than an imperative. This choice was based on the numerous independent studies demonstrating that the mutation of sarA limits, but does not abolish, biofilm formation not only in S. aureus but also in S. epidermidis (1, 11, 17, 28, 34, 35, 37, 39, 42). Our decision to include both UAMS-1 and UAMS-1625 was based on the fact that these strains are genetically and phenotypically distinct from each other but nevertheless are representative of prominent clonal lineages of S. aureus clinical isolates (13, 15). The observation that the mutation of sarA limited biofilm formation in both strains despite this genotypic and phenotypic diversity provides further support for our focus on sarA as a potential therapeutic target.
UAMS-1625 was found to have a reduced capacity to form a biofilm compared to that of UAMS-1 both in vitro and in vivo. This difference was modest, with average counts from catheters colonized by the two strains differing by less than 10-fold. However, it appeared to be relevant in the context of antimicrobial susceptibility, in that the efficacy of daptomycin treatment was greater with UAMS-1625 than with UAMS-1. In our in vitro experiments, which allowed for sampling at multiple time points, this was evident in terms of both the time of exposure and the daptomycin concentration. Specifically, with UAMS-1625, we observed a reduction of greater than 4 logs after 1 day of exposure to as little as 5× daptomycin, and the clearance of all catheters was observed after 3 days of exposure to 20× daptomycin. Neither of these was true of biofilms formed by UAMS-1.
Based on the increased susceptibility of UAMS-1625 biofilms observed in vitro, we used different daptomycin concentrations for UAMS-1 and UAMS-1625 in our in vivo experiments. Although this complicated comparisons between the parent strains in the treatment groups, we note that there was no statistically significant difference between the counts observed with the parent strains after treatment with daptomycin despite our use of a lower concentration with UAMS-1625 (10×) than with UAMS-1 (20×). The concentrations used in these experiments were based on the CLSI-defined breakpoints rather than the actual MIC for the test strains, and the MIC for UAMS-1625 was lower (0.38 μg/ml) than that for UAMS-1 (0.50 μg/ml), but even after adjusting for this difference the concentrations employed with UAMS-1625 were lower (26.3×) than those employed with UAMS-1 (40.0×). Taken together, these results support the conclusion that the differences we observed between UAMS-1 and UAMS-1625, while modest, nevertheless were relevant in the context of antimicrobial therapy, at least as defined by susceptibility to daptomycin.
The reasons for such strain-dependent differences have not been clearly defined (5, 16, 24, 40, 41, 43). However, the more important point in the context of this report is that the mutation of sarA limited biofilm formation in both UAMS-1 and UAMS-1625 to an extent that could be correlated with increased susceptibility to daptomycin in vivo. While this was reflected in average counts from each experimental group, it was perhaps most evident in the context of the clearance of established biofilms. Specifically, treatment with daptomycin for 7 days resulted in the clearance of detectable bacteria from only 6.7 and 6.3% of the catheters colonized with the parent strains UAMS-1 and UAMS-1625, respectively. In contrast, treatment with an equivalent concentration of daptomycin (compared to that for the respective parent strains) resulted in the clearance of 46.4 and 69.1% of the catheters colonized with the isogenic sarA mutants. Such results provide further support for the hypothesis that inhibitors of sarA have the capacity to enhance the efficacy of conventional antimicrobial agents, perhaps to the point that at least some biofilm-associated infections could be resolved effectively without surgical intervention and/or the removal of the offending device. Having said that, it must be acknowledged that the results we report represent a best-case scenario in which sarA function is abolished prior to establishing the biofilm, and it remains unclear whether it is possible to achieve a similar therapeutic effect by inhibiting sarA function in an established biofilm. This depends on several factors, most notably the ability to achieve a sufficient level of inhibition throughout the organized structure of the biofilm. Experiments aimed at addressing this issue are ongoing. It also is unclear why biofilms formed by sarA mutants are more easily cleared; one possibility is that they simply contain fewer bacteria at the outset. However, in previous in vitro experiments, statistical analysis confirmed that biofilms formed by an UAMS-1 sarA mutant were in fact more susceptible to daptomycin, vancomycin, and linezolid, even after accounting for the reduced ability of the sarA mutant to form a biofilm (42). Although that was not the case in the in vivo experiments reported here, the results nevertheless support the conclusion that targeting sarA has therapeutic potential in the specific context of S. aureus biofilm-associated infections.
Finally, >50% of the mice in the untreated group infected with UAMS-1625 developed skin lesions at the site of implantation that, in the absence of daptomycin treatment, ultimately resulted in the loss of the catheter. This was not observed in any other experimental group, including untreated mice colonized with the UAMS-1653 sarA mutant. One explanation for this is that catheters colonized with the UAMS-1653 sarA mutant contained fewer bacteria and, thus, failed to reach some critical threshold, leading to the formation of skin lesions. The alternative explanation is that the mutation of sarA limited the production of one or more virulence factors that contribute to the formation of such lesions. In this respect it is interesting that the mutation of sarA in UAMS-1625 resulted in the reduced production of alpha toxin, which has been proposed as the primary virulence determinant of USA300 clonal isolates (7). We also demonstrated that the mutation of sarA in UAMS-1625 resulted in the reduced production of delta toxin, which suggests that the impact of sarA on the production of alpha toxin in UAMS-1625 is mediated, at least in part, by the altered expression of agr (9, 22). However, the more important point is that, irrespective of the mechanism involved, these results suggest that the inhibitors of sarA not only increase antimicrobial susceptibility in the context of an established biofilm but also limit the pathology of at least some S. aureus biofilm-associated infections, including those caused by community-acquired methicillin-resistant S. aureus isolates of the USA300 clonal lineage.
This research was supported by funding from Cubist Pharmaceuticals (M.S.S.), grants AI043356 and AI069087 from the National Institute of Allergy and Infectious Disease (M.S.S.), and a research fellowship awarded to E.C.W. from the Alpha Omega Alpha Medical Honor Society.
Published ahead of print on 3 August 2009.