Hyperosmotic Agents and Antibiotics Affect Dissolved Oxygen and pH Concentration Gradients in Staphylococcus aureus Biofilms

ABSTRACT

Biofilms on wound surfaces are treated topically with hyperosmotic agents, such as medical-grade honey and cadexomer iodine; in some cases, these treatments are combined with antibiotics. Tissue repair requires oxygen, and a low pH is conducive to oxygen release from red blood cells and epithelialization. We investigated the variation of dissolved oxygen concentration and pH with biofilm depth and the variation in oxygen consumption rates when biofilms are challenged with medical-grade honey or cadexomer iodine combined with vancomycin or ciprofloxacin. Dissolved oxygen and pH depth profiles in Staphylococcus aureus biofilms were measured using microelectrodes. The presence of cadexomer iodine with vancomycin or ciprofloxacin on the surface of the biofilm permitted a measurable concentration of oxygen at greater biofilm depths (101.6 ± 27.3 μm, P = 0.02; and 155.5 ± 27.9 μm, P = 0.016, respectively) than in untreated controls (30.1 μm). Decreases in pH of ∼0.6 and ∼0.4 units were observed in biofilms challenged with medical-grade honey alone and combined with ciprofloxacin, respectively (P < 0.001 and 0.01, respectively); the number of bacteria recovered from biofilms was significantly reduced (1.26 log) by treatment with cadexomer iodine and ciprofloxacin (P = 0.002) compared to the untreated control. Combining cadexomer iodine and ciprofloxacin improved dissolved oxygen concentration and penetration depth into the biofilm, while medical-grade honey was associated with a lower pH; not all treatments established a bactericidal effect in the time frame used in the experiments.

IMPORTANCE Reports about using hyperosmotic agents and antibiotics against wound biofilms focus mostly on killing bacteria, but the results of these treatments should additionally be considered in the context of how they affect physiologically important parameters, such as oxygen concentration and pH. We confirmed that the combination of a hyperosmotic agent and an antibiotic results in greater dissolved oxygen and reduced pH within an S. aureus biofilm.

INTRODUCTION

Biofilms form on tissue wounds (1) and on the surfaces of implanted medical devices (2, 3). In both locations, biofilms can cause persistent infections. Biofilms protect bacteria from opsonization, antibiotic exposure, and phagocytosis (4–6). Staphylococcus aureus, a common human pathogen, is often associated with biofilm infections (7–9) in chronic wounds (8, 10, 11). Such infections are problematic to treat because of the slow penetration of antimicrobials and antibiotics through the biofilm matrix (10, 12–14) and the formation of persister cells (15). In vitro experiments have demonstrated that the bacterial gene regulation and resistance to antibiotics inside a biofilm differ from cell responses in planktonic culture (2, 5, 16–18). Furthermore, biofilm formation causes oxygen depletion (19), an increase in pH near the tissue surface, and delays in wound healing (20, 21). Availability of dissolved oxygen (DO) and reduced pH are two important factors that promote angiogenesis and increase macrophage infiltration and fibroblast activity (21–23). The pH of healthy skin ranges from 4.2 to 5.6 (24). Therefore, faster wound healing is likely to occur when wounds are treated in a manner that improves antibiotic delivery while maintaining oxygen availability and reducing pH in the wound bed.

Some researchers report that antibiotics do not effectively penetrate biofilms (25, 26), while others report rapid penetration (2, 27). Consistent with this discrepancy is the observation that specific antibiotics are required at higher dosages to treat biofilms (28–30). Additionally, the balance of water activity plays a critical role in a wide range of cellular and antibiotic functions and in mass transport in biofilms (31–33). A shift to a hyperosmotic environment challenges biofilm homeostasis, which in turn threatens the integrity of bacterial cellular membranes (34–36), improves antibiotic effectiveness (37), and exhausts the buffering capacity of bacterial cytoplasm, which may also result in a reduction in pH (38, 39). Thus, combining a high concentration of antibiotics with a hyperosmotic agent offers an alternative treatment strategy for biofilm communities.

DO from the atmosphere is an important source of oxygen during wound healing (40, 41), because tissue oxygenation is limited to capillaries near the skin surface (42). Oxygen is needed for collagen formation, angiogenesis, and epithelialization (42). When a biofilm is present on the wound surface, it competes for DO and depletes it before the oxygen can diffuse into the wound bed (43–45). Furthermore, chronic wounds have an alkaline environment, which retards wound healing and affects oxygen availability (46, 47). Therefore, it is crucial to find effective strategies to increase the DO concentration at the bottom of the biofilm and create a more acidic environment. Hyperosmotic concentrations of agents, such as medical-grade honey (48, 49) and cadexomer iodine (50, 51), have been considered for wound treatment. Recently, it was shown that hyperosmotic concentrations of maltodextrin increased the efficiency of vancomycin against S. aureus biofilms by reducing biofilm coverage, thickness, and diffusion distance, as well as by improving the DO concentration within the biofilm (37). To the best of our knowledge, the variation in DO concentration and pH in S. aureus biofilms that are treated with medical-grade honey or cadexomer iodine combined with vancomycin or ciprofloxacin has not been investigated.

The goal of this research was to investigate DO penetration and pH change in S. aureus biofilms challenged with medical-grade honey or cadexomer iodine in combination with vancomycin or ciprofloxacin. Ciprofloxacin (52, 53) and vancomycin (2, 27) are commonly used to treat bacterial infections. Ciprofloxacin is a fluoroquinolone that interferes with DNA replication through direct interaction with topoisomerases and DNA gyrase proteins (54). Vancomycin is a glycopeptide antibiotic that works by binding to the d-alanyl–d-alanine portion of the peptide side chain of the peptidoglycan strand that encompasses the cell wall, leading to the inhibition of cell wall synthesis (54). Biofilms were grown on membranes resting on agar surfaces where biofilm communities received nutrients from growth medium agar and oxygen from the air. This allowed growth of biofilms in a manner relevant to infections growing on wound surfaces. We treated S. aureus biofilms with (i) medical-grade honey alone, (ii) cadexomer iodine alone, (iii) sucrose alone, (iv) vancomycin alone, (v) ciprofloxacin alone, (vi) medical-grade honey with vancomycin or ciprofloxacin, (vii) cadexomer iodine with vancomycin or ciprofloxacin, and (viii) sucrose with vancomycin or ciprofloxacin, with the goal of identifying the conditions that allow the highest oxygen delivery. DO and pH depth profiles were measured using microelectrodes to quantify the oxygen consumption rate and the oxygen penetration depth in the biofilms and to investigate how pH changes inside the biofilms. Finally, we quantified the number of recovered cells after each treatment.

RESULTS AND DISCUSSION

The MIC for planktonic S. aureus culture was 0.74 μg/ml (0.5 μM) for vancomycin and 0.77 μg/ml (2 μM) for ciprofloxacin (see Fig. SI1 in the supplemental material). These values are consistent with the values reported by other researchers (55, 56). Because biofilms protect bacterial communities (5, 57), we used a higher concentration (4 mM) of ciprofloxacin and vancomycin to increase the probability of killing the bacteria.

Effects of hyperosmotic agents and antibiotics on dissolved oxygen penetration.

DO depth profiles from four locations in an untreated S. aureus biofilm showed almost identical trends (Fig. 1). This indicates that the profiles were reproducible when they were measured at different locations. Almost 99% of oxygen was consumed within the top 50 ± 5 μm of the biofilm.

FIG 1.

Dissolved oxygen (DO) depth profiles of untreated S. aureus biofilm measured at randomly selected different locations (100 μm apart from each other) in the biofilm.

Biofilms treated with vancomycin and ciprofloxacin showed similar DO depth profiles (Fig. 2A), indicating that they have a similar effect on the DO availability in biofilms. DO was no longer detectable below a depth of 55.2 ± 5 μm from the biofilm surface in the control, whereas DO was detectable to a depth of 60.2 ± 5 μm in biofilms treated with vancomycin or ciprofloxacin. When the biofilm was treated sequentially with 90% medical-grade honey and 4 mM vancomycin or ciprofloxacin (Fig. 2B), DO penetrated further into the biofilm, although the effect was not significant compared to the untreated control (P = 0.18 and 1.0, respectively) or compared to the individual treatments of vancomycin and ciprofloxacin (P > 0.05).

FIG 2.

The dissolved oxygen (DO) depth profiles inside the S. aureus colony biofilms. (A) Ciprofloxacin (C; 4 mM) and vancomycin (V; 4 mM) were applied to the top of colony biofilm and incubated for 8 h separately. (B) Medical-grade honey (MGH; 90%) was applied for 20 h, and then solutions with and without 4 mM C and 4 mM V were applied for 8 h sequentially. (C) Cadexomer iodine (CI; 90%) was applied for 20 h, and then solutions with and without 4 mM C and 4 mM V were applied for 8 h sequentially. The data are means, and the error bars represent the standard deviations of the means of the results from at least triplicate measurements.

DO concentration within biofilms was increased when they were treated with 90% cadexomer iodine and antibiotics (Fig. 2C); at a 60 ± 5 μm depth from the biofilm surface, the DO concentrations were 0.07 ± 0.03 mg/liter, 0.31 ± 0.1 mg/liter, 3.06 ± 1.5 mg/liter, and 3.4 ± 0.80 mg/liter for the control, 90% cadexomer iodine only, cadexomer iodine with vancomycin, and cadexomer iodine with ciprofloxacin, respectively. At a depth of 60 ± 5 μm, the DO concentration was significantly greater when the biofilm was treated with cadexomer iodine and vancomycin or ciprofloxacin (P = 0.02 and 0.016, respectively) with respect to the untreated control. In addition, the DO concentration at a depth of 60 ± 5 μm was significantly greater in biofilms treated with cadexomer iodine and ciprofloxacin than in the biofilm treated with ciprofloxacin alone (P = 0.002). Furthermore, the combined treatments were associated with detectable DO at greater depths within the biofilms. To illustrate, a DO concentration of ∼1 mg/liter was evident at depths of 30.1 μm, 101.6 ± 27.3 μm, and 155.5 ± 27.9 μm for the control, cadexomer iodine with vancomycin, and cadexomer iodine with ciprofloxacin, respectively. The depth of detectable DO increased when the biofilm was challenged with cadexomer iodine and vancomycin or ciprofloxacin (P = 0.01 and 0.001, respectively, compared to the untreated control). Moreover, the depth of detectable DO with cadexomer iodine and ciprofloxacin treatment was significantly greater than when vancomycin was used in conjunction with cadexomer iodine (P = 0.044 between cadexomer iodine with vancomycin and cadexomer iodine with ciprofloxacin treatment). Similarly, DO concentration increased when biofilm was subjected to combined treatments of sucrose and vancomycin or ciprofloxacin (data not shown); at a depth 60.2 ± 5 μm, the DO concentrations for sucrose alone, sucrose with vancomycin, and sucrose with ciprofloxacin were 0.43 ± 0.16 mg/liter, 0.95 ± 0.60 mg/liter, and 1.43 ± 0.57 mg/liter, respectively.

Treatment with cadexomer iodine resulted in the presence of DO at a greater depth than in biofilm treated with medical-grade honey, even when combined with antibiotics (Fig. 2B and C). In addition, the results indicate that the application of a hyperosmotic agent with an antibiotic improves the amount of DO available deep within the biofilm. This is further supported by the same results being obtained when the biofilm was treated with a combination of sucrose and antibiotics. In a surface wound environment, the presence of DO aids collagen formation, angiogenesis, and epithelialization of human tissues for improved wound healing (42).

Effects of hyperosmotic agents and antibiotics on biofilm oxygen consumption.

A change in the slope of a DO profile relative to an untreated control indicates a difference in the rate of oxygen consumption throughout the biofilm (Fig. 3A). When the biofilm was treated with 4 mM vancomycin or 4 mM ciprofloxacin alone, the oxygen consumption rate decreased by 11.8% ± 3.69% or 15.0% ± 4.96%, respectively (P > 0.05); treatment with 90% medical-grade honey alone or 90% cadexomer iodine alone also failed to induce a significant change (P = 0.99 in both cases) compared to the untreated control. Importantly, however, the rate of oxygen consumption decreased for biofilms that were challenged with medical-grade honey and vancomycin (P = 0.006). The greatest reduction was observed when biofilms were treated with cadexomer iodine and vancomycin or ciprofloxacin. Sequentially treating biofilms with cadexomer iodine and then either vancomycin or ciprofloxacin reduced the oxygen consumption rate by 65.0% ± 3.86% (P < 0.001) and 69.1% ± 5.96% (P < 0.001), respectively. These findings are consistent with the increasing DO concentration after these treatments (reported above and in Fig. 2C). A reduced rate of oxygen consumption is consistent with cell death or with a transition to anaerobic respiration.

FIG 3.

The relative rates of oxygen consumption of S. aureus biofilm (A) and planktonic culture (B) after various treatments. Data are normalized with the control to show percentage change from the control. The data are means, and the error bars represent the standard deviations of the means of the results from at least triplicate measurements. Stars indicate statistically significant differences of treated samples from the untreated controls (⋆, P < 0.05; ⋆⋆, P < 0.001). V, vancomycin; C, ciprofloxacin; MGH, medical-grade honey; CI, cadexomer iodine; S, sucrose.

To provide insight into the mechanism behind these observations, we measured the oxygen consumption rates of planktonic cultures during exponential growth. The rates of oxygen consumption of planktonic cultures treated with vancomycin or ciprofloxacin alone were not significantly different (P = 1 for both) from that of the untreated control culture (Fig. 3B). It has been reported in the literature that treatment with bactericidal concentrations of antibiotics is associated with accelerated cell respiration (58). This respiratory acceleration leads to a state of “overflow metabolism” and formation of reactive species that contribute to the lethality of the antibiotics (59). Given the bactericidal nature of vancomycin and ciprofloxacin, this is a plausible explanation for the apparent lack of effect on oxygen consumption relative to the untreated control culture. In contrast, planktonic cultures treated with medical-grade honey alone or cadexomer iodine alone showed a significant decrease in the rate of oxygen consumption (P = 0.006 and <0.001, respectively) compared to the untreated planktonic culture, consistent with reduced cell respiration. The effect of hyperosmotic agents alone was augmented in the planktonic cultures; this is possibly explained by the slow release of the antimicrobial iodine from the cadexomer polymer and the hyperosmotic effect from medical-grade honey and sucrose that resulted in slow growth. Moreover, the planktonic cultures lack the restricted diffusion present in biofilms due to the extracellular polymeric substance (EPS).

A combination of medical-grade honey or cadexomer iodine with vancomycin or ciprofloxacin reduced the rate of oxygen consumption (Fig. 3A), a result that is consistent with higher concentrations of DO at greater depths in the biofilms. The increased DO penetration is consistent with either cell death or quiescence. Besides the antibacterial activity of the two antibiotics, medical-grade honey (60, 61) induces cell death through the production of reactive oxygen species (ROS), while cadexomer iodine rapidly penetrates cells, where it damages proteins, nucleotides, and fatty acids, culminating in cell death (62–64). Cell death is consistent with higher DO concentrations deeper into the biofilm because diffusing oxygen is no longer used by bacteria. This is important because it is generally accepted that the correction of wound hypoxia is needed to support healing (41). For planktonic culture, the rate was also significantly decreased when hyperosmotic agents were combined with antibiotics (Fig. 3B) compared to the untreated planktonic culture. Here, we have shown that treatment with a hyperosmotic agent and an antibiotic together improves DO concentration at deeper levels within a biofilm.

Effects of hyperosmotic agents and antibiotics on biofilm pH.

The pH depth ranged from ∼7.3 to ∼7.0 from the liquid-biofilm surface to within the control biofilm (Fig. 4). Similarly, the pH of biofilm treated with 4 mM vancomycin or ciprofloxacin ranged from ∼7.2 to ∼7.0 (Fig. 4A). The experimentally measured pH levels of the vancomycin and ciprofloxacin solutions before they were added to the biofilms were ∼3.7 and ∼4.8, respectively. However, it is evident that the biofilms treated with vancomycin alone or ciprofloxacin alone showed a neutral pH, similar to the control. Biofilms are able to sustain a neutral or an alkaline state, which is a suitable environment for biofilm growth and maintenance, in the presence of an antibiotic (22, 65, 66). S. aureus has an optimal pH for growth of ∼7 (67, 68).

FIG 4.

The pH depth profiles inside S. aureus biofilm treated with and without 4 mM ciprofloxacin (C) or vancomycin (V) (A), with 90% medical-grade honey (MGH) and 4 mM C or 4 mM V sequentially (B), and with 90% cadexomer iodine (CI) and 4 mM C or 4 mM V sequentially (C). The initial pH values of the vancomycin and ciprofloxacin solutions were ∼3.7 and ∼4.8, respectively. One replicate is shown here to represent all three replicates.

At a depth of 150 ± 10 μm, the pH of the biofilm decreased significantly, to 6.2 ± 0.1 (P < 0.001), when 90% medical-grade honey alone was introduced (Fig. 4B) compared to the untreated control. Similarly, a significant decrease in pH was observed for biofilm treated with medical-grade honey and ciprofloxacin, with a P value of 0.01 (Fig. 4B). With respect to the untreated control, cadexomer iodine with or without antibiotic did not show a significant effect (P > 0.05) on pH change (Fig. 4C). Thus, the colony biofilms treated with medical-grade honey alone and sequentially with ciprofloxacin showed low pH; decreases in pH of ∼0.6 and ∼0.4 units, respectively, were observed compared to the control at a depth of ∼150 μm. A key factor in the efficacy of honey is its low pH, because this can be inhibitory to bacterial growth (69, 70) and likely induces acid stress in the biofilm. In addition, enzymatic production of hydrogen peroxide (an ROS) occurs naturally in honey (60, 61). With a higher DO concentration facilitated by low pH, ROS generation is presumably increased. Other mechanistic effects of medical-grade honey include low water activity and high sugar content, leading to osmotic stress (71). Our findings for medical-grade honey are consistent with the reported pH of honey in the literature (70).

A low pH on the wound bed promotes epithelialization and oxygen delivery from the blood capillaries (46, 72–74). Ono et al. (24) reported a decrease in pH along with the promotion of epithelialization in noninfected burn wounds, whereas an increase in pH was observed in wounds as infections progressed inside them. Thus, it is crucial to stimulate an acidic environment on the wound bed to promote healing. Additionally, antibiotics can fail to penetrate microbial biofilms (25); the ability of an antibiotic to penetrate a biofilm depends on the rate at which it is deactivated. Agents that experience reactive neutralization in the biofilm are prone to penetration failure (75). Moreover, antibiotic efficacy is affected by pH (76); ciprofloxacin was found to decrease its activity in a low-pH environment (77, 78). Ciprofloxacin is a piperazine-containing fluoroquinolone which is negatively charged at physiological pH (79). This is found to be its most active form. However, when the pH is low, the charge becomes positive. This positive charge may affect the penetration of ciprofloxacin into the bacteria and decrease their activity. Based on Fig. SI2, the MIC of ciprofloxacin increased by 3-fold when the pH of the medium was 6 rather than 7. Moreover, the MIC increased by 15-fold when the pH of the medium was 5 rather than 7. The MIC results (Fig. SI2) confirm the negative effect of low pH on the activity of ciprofloxacin.

Cell viability after treatments with hyperosmotic agents and antibiotics.

The CFU counts of control and treated biofilms were determined under conditions identical to those used in the DO and pH depth profiling. Figure 5 shows the viable cells (in CFU per milliliter) inside the S. aureus colony biofilm under the various treatment conditions. Vancomycin alone had little effect (P > 0.05) on recovered bacterial counts (0.5-log reduction; Fig. 5), whereas ciprofloxacin alone reduced CFU counts by 0.8 log. A reduction of only 0.4 log was observed in biofilms treated with medical-grade honey, which is not significantly different (P > 0.05) from the untreated control. When both hyperosmotic agents and antibiotics were administered, log reductions of 0.6 to 1.2 were observed. Among the treatments, cadexomer iodine with ciprofloxacin yielded the highest significant reduction (1.26-log reduction; P < 0.001) compared to the untreated control. The controlled release of iodine and improved penetration of ciprofloxacin possibly contributed to the significant killing with the combined treatment. It is important to note that the molecular mass of vancomycin (vancomycin hydrochloride hydrate) is 1,485.71 g/mol and that of ciprofloxacin (ciprofloxacin hydrochloride monohydrate) is 385.82 g/mol. Among all the active compounds, the smaller molecules have higher activities for specific cells (80). In addition, antibiotics are effective when cells are metabolically active and use oxygen as an electron acceptor; aminoglycosides are oxygen transport dependent, while beta-lactams and vancomycin inhibit actively dividing cells (81–83). Furthermore, the availability of oxygen deeper within the biofilm can stimulate dormant cells to respire aerobically, leading to an active metabolism that is favorable for antibiotic killing.

FIG 5.

The effect of sequential treatment with hyperosmotic agents and antibiotics on CFU of S. aureus biofilms. The treatments are 4 mM vancomycin (V), 4 mM ciprofloxacin (C), 90% medical-grade honey (MGH), 90% cadexomer iodine (CI), and 90% sucrose (S). The data are means of the results from at least three biological replicates. The error bars represent standard deviations of the means of the results from triplicate measurements. The stars indicate statistically significant differences from the untreated control biofilm (⋆⋆, P < 0.001; ⋆, P = 0.002 for 4 mM C and P = 0.004 for 90% CI; n = 3).

Despite the successful reduction in the number of viable bacteria, we generally failed to achieve the >3-log reduction (99.9% decrease in cell number) that is generally considered an effective bactericidal effect (84). However, clinical treatments with compounds, such as cadexomer iodine and medical-grade honey, are usually administered for weeks, and the applications are repeatedly replaced with fresh dressings within the treatment period (85–88). Consequently, while the effect may appear limited, this does not account for the repeated and lengthy exposure that would occur in clinical cases. In the case of planktonic cultures, higher log reductions are observed overall for all treatments compared to biofilms: sucrose with a ciprofloxacin-treated culture had a 3-log reduction, and medical-grade honey alone reduced CFU about 2.29 log (Fig. SI3). This is expected, since planktonic cultures do not have the protection of extracellular polymeric substance (EPS) that biofilms have.

In summary, combining cadexomer iodine with vancomycin or ciprofloxacin significantly increased DO concentration and penetration depth within the biofilm compared to the untreated biofilm, with no significant bactericidal effect. In addition, the oxygen consumption rate was significantly reduced when the biofilm was treated sequentially with medical-grade honey and vancomycin, or with cadexomer iodine with vancomycin or ciprofloxacin. The addition of hyperosmotic agents and antibiotics caused a pH decrease, especially in biofilms treated with medical-grade honey only and sequentially with ciprofloxacin. The poor activity of ciprofloxacin in an acidic environment (90% medical-grade honey) could explain its minimal effect on DO penetration compared to that of vancomycin. Last, cell viability was reduced significantly for biofilms challenged with cadexomer iodine combined with ciprofloxacin in comparison to the untreated biofilm; the reduction was, however, not significant as a bactericidal effect. The resulting increased DO penetration depth and acidic environment in the biofilm, due to sequential treatment with hyperosmotic agents and antibiotics, are key points to consider when evaluating treatments for wound infection. This comprehensive study details an antibiotic delivery approach that both serves the purpose of effective biofilm elimination from wound surfaces and allows for a wound environment suitable for faster wound healing.

MATERIALS AND METHODS

Bacterial strains and media.

Pure cultures of Staphylococcus aureus (ATCC BAA-1747) were used in this study. Bacteria were streaked and grown on Trypticase soy agar (TSA) (catalog no. DF0369-17-6; Fisher Scientific, Palatine, IL, USA) at 37°C for 18 h. A single colony from the TSA plates was transferred into a 50-ml culture tube containing approximately 5 ml of Trypticase soy broth (TSB) (catalog no. DF0370-17-3; Fisher Scientific). The culture in TSB was incubated on a shaker (55 to 60 rpm) at 37°C and grown overnight.

MIC measurements.

Before selecting the doses for the antibiotics to treat S. aureus biofilms, we estimated the MICs of these two antibiotics against an overnight culture of S. aureus. The procedure defined by the Clinical and Laboratory Standards Institute (CLSI) was used, with minor changes (89). Briefly, 11.89 μg/ml and 3.09 μg/ml concentrations of vancomycin and ciprofloxacin, respectively, were diluted at 1:2 using TSB to a concentration close to zero. Finally, optical density (OD) at 595 nm was recorded at regular intervals using a BioAssay reader (PerkinElmer HTS 7000 Plus).

To determine the effect of pH on the activity of ciprofloxacin, MIC assays were performed on S. aureus planktonic culture for ciprofloxacin using various pH-adjusted batches of TSB (pHs of 5, 6, and 7 were tested). Similarly, the procedure defined by CLSI was used, with minor changes (89); 98.77 μg/ml ciprofloxacin was diluted at 1:2 using TSB at various pHs to a concentration close to zero. A bioassay reader was used, as described above.

Oxygen consumption rate assay for planktonic culture.

To assess the oxygen consumption of planktonic culture, a MitoXpress Xtra oxygen consumption assay (catalog no. MX-200; Luxcel Biosciences, Amsbio, Cambridge, MA, USA) was used. A 96-well black plate, TSB, and HS mineral oil (provided in the assay kit) were prewarmed separately to 37°C. Overnight cultures were diluted to an OD at 600 nm (OD₆₀₀) of 0.22 for the control and treatments with medical-grade honey alone, cadexomer iodine alone, vancomycin alone, and ciprofloxacin alone. Overnight cultures were diluted to an OD₆₀₀ of 0.25 for combined treatments with a hyperosmotic agent (2.9% medical-grade honey, 2.9% cadexomer iodine, or 2.9% sucrose) and an antibiotic (86.4 μM vancomycin or 86.4 μM ciprofloxacin). Aliquots (150 μl of OD₆₀₀ of 0.22 and 140 μl of OD₆₀₀ of 0.25) were dispensed into the 96-well plate. Ten microliters of autoclaved deionized water and individual treatments were added to the 150-μl cultures. For combined treatments, 10 μl of a hyperosmotic agent (medical-grade honey, cadexomer iodine, or sucrose) and 10 μl of an antibiotic (vancomycin or ciprofloxacin) were added to the 140-μl culture. A 10-μl aliquot of the reconstituted MitoXpress Xtra reagent (in autoclaved deionized [DI] water) was then added to each well. Finally, two drops of prewarmed HS mineral oil were added to each well to limit diffusion of oxygen from air to the sample. The concentrations of hyperosmotic agents and antibiotics used against planktonic culture were selected to produce equivalent micrograms of treatment per CFU per milliliter to which biofilms were exposed. The well plate was read for 5 h at a time-resolved fluorescence setting with 380 ± 20 nm excitation, 650 ± 15 nm emission, a 30-μs delay, and a 100-μs measurement time (Cytation 5 microplate reader; Bio-tek, Winooski, VT, USA). The oxygen consumption rate was calculated from the linear portion of the signal profile over time.

Biofilm preparation.

Colony biofilms were grown as described previously (57), with minor modifications. Briefly, overnight planktonic cultures (18-h incubation) of S. aureus were diluted to an OD₆₀₀ (1-cm path length) of 0.5 in TSB. A 5-μl aliquot of diluted culture was used to inoculate individual sterile black polycarbonate membrane filters (25-mm diameter, 0.2-μm pore size, catalog no. GTBP 025 00; Fisher Scientific, Palatine, IL, USA) resting on TSA plate culture medium (membranes were sterilized by UV exposure, 15 min per side, prior to inoculation). The plates were inverted and incubated at 37°C for 48 h. The membrane-supported biofilms were transferred to fresh TSA every 6 to 10 h.

Hyperosmotic agents and antibiotics.

Medical-grade honey (catalog no. UMF20+; Manuka Natural, Christchurch, Canterbury, New Zealand) and cadexomer iodine gel (catalog no. b12261; Active Forever, Peoria, AZ, USA) were used in this study. Vancomycin hydrochloride and ciprofloxacin hydrochloride were purchased from Sigma-Aldrich (catalog no. 861987 and PHR1044, respectively). All chemical compounds were used as received, without further purification, and were dissolved in autoclaved filtered (0.2-μm pore size) nanopure water.

Treatments.

We used the following treatments and treatment combinations: (i) medical-grade honey alone, (ii) cadexomer iodine alone, (iii) vancomycin alone, (iv) ciprofloxacin alone, (v) medical-grade honey and vancomycin, (vi) medical-grade honey and ciprofloxacin, (vii) cadexomer iodine and vancomycin, (viii) cadexomer iodine and ciprofloxacin, (ix) sucrose alone, (x) sucrose and vancomycin, and (xi) sucrose and ciprofloxacin.

After 48 h, the biofilms were treated with 4 mM vancomycin (5,942.8 μg/ml) alone, 4 mM ciprofloxacin (1,471.2 μg/ml) alone, 90% medical-grade honey alone, or 90% cadexomer iodine alone. These compounds were added (15 μl for hyperosmotic agents and 10 μl for antibiotics) directly on top of the preformed biofilms. Treatment with an antibiotic or hyperosmotic agent alone was administered for 8 h or 20 h, respectively. Combined treatments were administered sequentially (20 h of treatment with a hyperosmotic agent and then 8 h of treatment with an antibiotic, for a total incubation of 28 h). Hyperosmotic agent and antibiotic concentrations were adjusted to maintain total concentrations of 90% and 4 mM, respectively. DO and pH depth profiles were measured inside the biofilms for each time frame of treatment. Control biofilms were measured after 48 h of growth.

Dissolved oxygen and pH microelectrodes.

DO and pH microelectrodes were constructed according to previously described protocols (90). The microelectrodes had tip diameters of <20 μm. DO microelectrodes were polarized to −800 mV_Ag/AgCl and calibrated using two-point calibration: in the air (oxygen saturation) and in a solution of 0.1 M sodium ascorbate and 0.1 M NaOH (without oxygen). The pH microelectrodes were constructed with liquid ion exchange (LIX; catalog no. 95297; Sigma-Aldrich, St. Louis, MO, USA) membrane tips (90). Custom-made external Ag/AgCl reference electrodes with an agar salt bridge at the tip were used for pH measurements. Then, pH microelectrodes were calibrated in buffer solutions (pH 4, 7, and 10; catalog no. 910104, 910107, and 910110, respectively; Cole-Palmer, Vernon Hills, IL, USA).

A Keithley 6517A electrometer/high-resistance meter was used to operate both DO microelectrodes and pH microelectrodes. The measurement setup is shown in Fig. 6. For pH measurements, 50 μl of liquid medium was added on top of the biofilm to provide an ionic flow between the pH and reference electrodes. The microelectrode (DO or pH) was initially placed on top of the biofilm. The surface of the biofilm sample was determined using a Zeiss Stemi 2000 stereomicroscope (Carl Zeiss Microscopy, Thornwood, NY, USA). After the microelectrode tip was located above the biofilm surface, the microelectrode was stepped down in 5-μm increments, except for the pH profiles measured for medical-grade honey and cadexomer iodine treatments. A 10-μm step size was used to measure pH profiles in biofilms treated with medical-grade honey and cadexomer iodine. A computer-controlled stepper motor (PI M-230.10S, part no. M23010SX; Physik Instrumente, Auburn, MA, USA) controlled with a custom-built microprofiling system was used to manipulate microelectrode movements. An analog-to-digital converter (ADC; USB-1608FS; Measurement Computing, Norton, MA, USA) was used to record the collected data.

FIG 6.

Colony biofilm and microelectrode experimental setup. The pH microelectrode setup differs by having an external reference electrode. AD, analog to digital.

The oxygen consumption rate was calculated using the DO depth profiles. At any given time, the oxygen consumption rate in a biofilm is equal to the oxygen flux through the biofilm-air interface (90, 91). Using Fick's first law of diffusion, the oxygen flux is calculated as the slope of the DO depth profile multiplied by the oxygen diffusion coefficient (0.178 cm²/s in air) (92).

Bacterial viability measurement.

CFU before and after treatments were quantified using methods described by Chen et al. (93). The results and errors represent average log₁₀-transformed values and standard deviations of the results from triplicate independent experiments.

Data analysis.

All experiments were performed at least in triplicate. Depth profiles, oxygen consumption rates, and CFU counts are presented and plotted as mean ± standard deviation, except for pH depth profiles. One replicate is shown to represent all three replicates. A one-way analysis of variance (ANOVA) with Tukey's post hoc test was performed to determine significant differences for all possible pairwise comparisons among treatments and control for oxygen consumption rates. A single-factor ANOVA with a post hoc pairwise comparison test and a Bonferroni correction were used to determine whether there was a significant difference between the DO concentrations and viabilities of control and treated biofilms. For statistically significant testing, we used a P value of ≤0.05. Calculations and statistical analyses were performed using SigmaPlot (versions 12.0 and 12.5).