Comparison of antibacterial and antibiofilm activity of bioactive glass compounds S53P4 and 45S5

Bone loss and deformation due to damage caused by injury or recurrent invasive infections presents a major clinical obstacle. While bone substitute biomaterials promote osseous tissue regeneration, their application in sites complicated by microbial infections such as osteomyelitis, is limited. Bioactive glass biomaterials (Bioglass) have been shown to have efficient mechanisms of repairing the integrity of bone, while inhibiting growth of a range of bacterial strains. There are several commercially available bioactive glass compounds, each with a unique chemical composition. One compound in particular, S53P4, has demonstrated antimicrobial effects in previous studies but the antimicrobial activity of the parent compound 45S5 has not been investigated. To assess whether antimicrobial activity is common among bioglass compounds, 45S5-the parent compound, was evaluated in comparison to S53P4 for antibacterial and antibiofilm effects against multiple strains of aerobic and anaerobic bacteria associated with various types of osteomyelitis. Experiments of antimicrobial effects in liquid cultures demonstrated that both compounds were antimicrobial against various microbial genera including S. gordonii, V. parvula, P. aeruginosa and MRSA; particles of the smallest size (32–125 µm) invariably showed the most robust antimicrobial capabilities. When employed against biofilms ecological biofilms grown on hydroxyapatite, 45S5 particles produced a stronger reduction in biofilm mass compared to S53P4 particles when considering small particle ranges. We found that 45S5 seems to be as effective as S53P4 and possibly even more capable of limiting bacterial infections. The efficacy of bioactive glass was not limited to inhibition of planktonic growth, as it also extended to bacterial biofilms. The increased antibacterial activity of 45S5 compared to S53P4 is true for a variety of size ranges.


Introduction
Globally, the need for bone and joint surgeries is constantly increasing, and bone substitute biomaterials are being developed to meet this demand as well as to optimize repair and regeneration of bone tissues [1]. While biomedical biomaterials research has mainly focused on tissue regenerative effects of bone substitute biomaterials, combating infections is equally as important as regenerating tissue. This need for both the ability to regenerate tissue and combat infection has shifted development focus towards multifunctional biomaterials with antibacterial properties. Bioglasses are bioactive materials which promote bone regeneration. Some bioglass formulations have been shown to have antibacterial properties, making them particularly suitable for use in compromised bone regeneration surgical procedures which Open Access *Correspondence: kotsakis@uthsca.edu have high infection risk, such as oral osseous and osteomyelitis defects [1]. The efficacy of S53P4 bioactive glass, a variant of 45S5, in inhibiting bacterial growth has been extensively documented both in vitro and in vivo [2][3][4]. Prior work has shown that S53P4 inhibits the growth of methicillin-resistant Staphylococcus aureus (MRSA), Klebsiella pneumoniae, and Acinetobacter baumannii by causing deformation of the cells, and hole formation in the cell membranes [5,6]. In addition, bioglass 45S5 has been reported to exhibit an antibacterial effect against S. aureus and Staphylococcus epidermidis [7]. However, information is scarce on its antimicrobial effects due to a variety of size ranges. If equivalence can be determined, then the indications for 45S5 could be broadened to additionally include the indications for S53P4 related to clinical deployment in infected sites in vivo, as well as set the foundation for chemical optimization of antibacterial bioactive glass formulations.
Osteomyelitis itself is a costly medical issue; when antibiotic-resistant microbes are involved, it can lead to life-threatening complications. An array of bacteria has been implicated in both bone surgery complications and osteomyelitis resulting in substantial morbidity and medical costs. Usually, osteomyelitis is managed with IV antibiotic courses that may be empirical or adjusted based on microbial identification and requires several weeks of hospitalization [8]. One challenge is that the characteristics of the pathogens vary widely across clinical cases with an array of aerobic and anaerobic bacteria capable of destructive bone inflammation, thus limiting the efficacy of systemic antibiotics [9]. Furthermore, several pathogenic microorganisms that cause osteomyelitis are resistant to antibiotics. For instance, Veillonella parvula, which is a causative agent in vertebral osteomyelitis, was within the spectrum of Penicillin G and cephalosporins in the 1980s when its pathogenic role was established [10]. However, antibiotic targeting of Veillonella has led to considerable resistance in strains implicated in osteomyelitis and septicemia, with a recorded resistance to penicillin exceeding 30 μg/ml [11]. These findings highlight the need for developing local antimicrobial therapies for bone infections that are resistant to systemic antibiotic treatment.
In addition to preventing long-term resistance, it is essential to prevent systemic mortality and morbidity due to direct bacterial infection in bone lesions. In some cases, multidrug resistant bacteria implicated in osteomyelitis can lead to life-threatening complications. For instance, Pseudomonas aeruginosa is multidrug resistant, thus leading to mortality in hospitalized patients [12]. Similarly, MRSA is a notorious pathogen that causes severe infections in the bone. MRSA is more difficult to eliminate than most S. aureus strains due to its extraordinary capability of resistance to commonly used antimicrobials [13]. Additionally, Streptococcus gordonii, which is implicated in spondylodiskitis, can ultimately result in bacteremia and infective endocarditis [14,15]. Therefore, the overall goal of this work is to identify whether the antimicrobial effects of multifunctional bioactive glass bone substitutes are present in all commonly employed compositions, and to determine efficacy against clinically relevant bacteria and their biofilms. The primary objective of this study is to compare anti-bacterial and anti-biofilm properties between Bioglass 45S5 and Bioglass S53P4 compositions.

Bacterial strains and culture conditions
Representative Gram positive (Gram +) and Gram negative (Gram-) bacterial species / strains that have been found to be pathogenic in various forms of bone infections or osteomyelitis were used in this study; listed in Table 1. S. gordonii DL1 was grown in Brain Heart Infusion (BHI) broth (Oxoid ™ , Thermo Scientific ™ , USA), and V. parvula PK1910 was cultured in BHI supplemented with 0.6% sodium lactate (BHIL). These bacteria were grown anaerobically at 37 °C. P. aeruginosaPAO1 was cultured in Luria-broth (LB) (BD Difco ™ , USA), and MRSA ATCC BAA-2313 ™ was cultured in Nutrient Broth (NB) (BD Difco ™ , USA). These two bacteria were routinely grown aerobically at 37 °C. Culture of all strains was performed under BSL2 conditions and was approved by the Institutional Biosafety Committee. Because bacteria form multispecies biofilms on the bone that confer resistance to antimicrobials as compared to that of planktonic bacteria, for greater clinical relevance, multi-species bacterial biofilms were also employed. A previously characterized ex vivo clinical sample isolated from a peri-implant osteolytic lesion was cultured on porous hydroxyapatite discs to better simulate anti-biofilm efficacy under translational clinical conditions of bacterial bone infection. Isolation and culture conditions of the ex vivo ecological biofilm, which includes over thirty distinct bacterial taxa, have been previously described [16,17].

Preparation of bioglass particles
Bioglass 45S5 and S53P4 were used in this study and were separated by particle size to determine particle sizedependent bacterial inhibitory effects as well as to mimic marketed product configurations for orthopedic bone graft applications. Bulk bioglass frit was placed in grinding jars with burundum grinding media and placed on a jar mill (U. S. Stoneware, East Palestine, OH) to pulverize the bioglass. The bioglass particulate was then sieved using stainless steel sieves on a mechanical shaker to separate it into the various size ranges (1 mm and 2 mm sieves for 1-2 mm particles; 500 µm and 710 µm sieves for 500-710 µm particles; 125 µm and 500 µm sieves for 90-710 µm particles; and 32 µm and 125 µm sieves for 32-125 µm particles.) Aliquots of particulate were packaged in heat-sealed pouches and sterilized using gamma irradiation (25 kGy).

Antimicrobial effects of bioglass particles
All particles were weighed and mixed with different bacteria in corresponding broth. Based on previously published studies suggesting that concentrations up to 800 mg/ml of bioglass are required for complete growth inhibition of some bacterial species/strains [5], strainspecific pilot experiments were conducted to determine inhibitory concentrations with doses ranging from 50 mg/ml to 800 mg/ml. Overnight cultures of S. gordonii, V. parvula, P. aeruginosa and MRSA were centrifuged and resuspended with fresh corresponding media to an OD 600 of 1.0. Next, resuspended bacterial cultures were diluted 1:1,000 into the media with pre-weighted bioglass particles. Bacterial cultures without particles were used as positive controls, and cultures with particles and without bacteria were used as sterility controls. The bacteria were incubated at 37 °C in aerobic (P. aeruginosa and MRSA) or anaerobic conditions (S. gordonii and V. parvula) as indicated for 24, 48 h or 72 h. Before measuring bacterial growth, bacteria were vortexed and suspended, and tubes were left standing until particles precipitated. Optical densities were measured at 600 nm, which directly corresponds to bacterial cell numbers in broth, to determine bacterial growth as compared to controls.

Antibiofilm effects of bioglass
The efficacy of bioglass against established biofilms was assessed using the crystal violet staining method. For biofilm assays, MRSA and V. parvula were employed based on their ability to form single-species biofilms. Bacterial cells in log-phase were centrifuged and resuspended with fresh corresponding media to an OD 600 of 1.0 for standardization. For biofilm formation, bacteria were inoculated at a 1:1,000 dilution in BHI broth supplemented with 0.6% sodium lactate (BHIL; V. parvula), or TSB broth with 0.5% Yeast Extract and 0.5% glucose (TSBYEG; MRSA). Then, 200 µl aliquots were added into sterile 96-well plates and incubated anaerobically (V. parvula) or aerobically (MRSA) at 37 °C for 24 h. Subsequently, to determine antibiofilm effects, Bioglass 45S5 and S53P4 with different particle sizes were weighed and aliquoted into the 96-well plates with the 24 h established biofilms at final concentrations of 100-200 mg/ml. After an additional 24 h incubation, the plates were washed with PBS three times to remove non-adhered cells, biofilm, and bioglass particles. Washed plates were then stained with crystal violet, dissolved with 30% acetic acid, and biomass measured at OD 562 nm. Each assay was performed in triplicate wells and repeated three times. Biofilms stained with crystal violet without any stimulations served as controls.

Ex vivo ecological biofilm formation and scanning electron microscopy (SEM)
To further study the antibiofilm effects of bioglass on multispecies biofilms, the multi-species ex vivo ecological biofilms described above were grown on porous HA-discs and SEM was used to assess the biofilm after treatment of bioglass 45S5/32-125 µm and S53P4/32-125 µm. Briefly, the working culture was prepared by inoculating frozen ex vivo biofilm stocks in TSB broth with 0.5% Yeast Extract and 0.5% sucrose (TSBYES) and grown anaerobically at 37 °C for 18 h. The overnight culture was diluted 1:10 into fresh TSBYES media and 1 mL bacterial culture was aliquoted into 24-well plate containing HA discs for ecological biofilm formation. To facilitate bacterial biofilm attachment, the sintered HA discs were modified with a round carbide burr at 2,000 rpm followed by acid etching with 37% orthophosphoric acid for 45 s, thus resulting in a microrough porous surface. The HA discs were sonicated in 70% ethanol prior to the experiment, and then placed into a 24-well plate and seeded with bacterial inoculums. The samples were cultured in an anaerobic jar at 37 °C for 24 h for the formation of ecological biofilms on HA-discs. After incubation, bioglass 45S5/32-125 µm and S53P4/32-125 µm were added into 24-well plate at the final concentration of 400 mg/mL, respectively. HA-disc without bioglass was used as control. The plate was cultured anaerobically at 37 °C for 24 h for bioglass treatment. After 24 h incubation, the HA-discs were washed with PBS times to remove non-adhered cells and biofilm and bioglass particles. The HA-discs were then fixed in a phosphate buffered 4% formaldehyde and 1% glutaraldehyde solution for 2 h, triple-washed in PBS for 3 min and post fixed in 1% osmium tetroxide for 1 h. After rinsing in Zetterquist's buffer for 2 min, the HA-discs were then dehydrated and triple-washed with 70% and 95% ethanol for 15 min, followed by two 20 min rinses in 100% ethanol. The specimens were then treated in Hexamethyldisilazane (HMDS) for 5 min and air-dried in a desiccator. The samples were mounted on aluminum stubs, sputter-coated with gold palladium, and then examined with scanning electron microscopy.

Fluorescence microscopy analysis
To validate the antibiofilm effect of bioglass, the ex vivo multi-species biofilms were grown on HA-discs as described above, and the mature biofilm was treated by 45S5 particles 32-125 µm, which were determined to have the best antibiofilm effects in the previous experiments and analyzed by fluorescence microscopy. Biofilms were prepared exactly as described above by inoculating HA-discs in 24-well plates.

Statistical analysis
To examine the differences between different bioglass compositions and particle sizes in treatment of the different bacterial strains, we performed statistical analyses via repeated-measures two-way ANOVA at different time points. Post-hoc testing using the False Discovery Rate (FDR) was used to determine differences in pairwise comparisons for bacterial inhibition related to particle type and concentration. A standard two-way ANOVA with post-hoc testing was used to examine variation in V. parvula and MRSA biofilm inhibition due to the particle type and particle size used. P-values were assessed as significant using FDR-adjusted alpha levels.

Antibacterial effects of bioglass particles on planktonic bacteria
To evaluate the antibacterial effects of bioglass particles, the bacterial cultures were treated with various size ranges of 45S5 and S53P4 bioglass. Bacterial growth was quantitated by measuring an optical density at 600 nm (OD 600 nm). Veillonella are strictly anaerobic Gramnegative cocci and play crucial roles for oral biofilm formation and the V. parvula species are opportunistic pathogens [18][19][20][21][22]. Being that the V. parvula is causative to certain forms of osteomyelitis and is known to be resistant to multiple antibiomicrobials [10,23], it was first assessed to determine composition-and particle-related antimicrobial effects of bioglass. Because Veillonella viability in liquid broth diminishes after 48 h, we measured OD 600 nm after 24 h and 48 h incubation. As shown in Fig. 1, when 50 mg/ml (A&B) and 100 mg/ml (C&D) bioglass particles were incubated with V. parvula PK1910, the growth of this bacterium was completely inhibited by 45S5 and S53P4 with the size of 32-125 µm at 24 h time point (p < 0.001). However, the population partially recovered after 48 h of incubation, suggesting that V. parvula was not eradicated by either bioglass at the two concentration levels. In addition, after 48 h incubation, the 32-125 µm bioglass particles reduced bacterial growth as compared to the positive control (p < 0.001), while bacterial viability in the 90-710 µm particle range was comparable to controls (p > 0.05). Differential antimicrobial effects were noted when cultured for 48 h with 500-700 µm and 1-2 mm for both the 50 mg/ml and 100 mg/ ml concentrations (Fig. 1B, D). Bacterial viability was fully recovered in these particle ranges in S53P4 groups, while 45S5 showed significantly greater inhibitory effects on the growth of V. parvula (p < 0.001) (Fig. 1). Because previous studies have suggested that bioglass antimicrobial effects are pH-mediated [7], we validated previous investigations by measuring the pH after 24 h culture and found that the maximal effect noted with the 32-125 µm range was consistent with the more alkaline conditions they produce as compared to larger particle groups (data not shown). Due to its resistance to commonly used antibiotics, MRSA is notoriously more difficult to treat than most Gram-positive pathogens. Thus, we validated the findings of particle range effects on MRSA. In this study, we assessed the inhibitory effects of all sizes of bioglass particles on MRSA at a concentration of 200 mg/ml. Compared to non-treated bacterial cultures, all particles reduced MRSA growth at 24 h, and, notably, both 45S5 and S53P4 of 32-125 µm completely inhibit MRSA growth ( Fig. 2A). After 48 h incubation, we did not observe any growth increase for the S53P4 bioglass group compared to 24 h cultures, however, MRSA growth was recovered in the group of 45S5/90-710 µm and 500-700 µm. Results were similar at 24 h, with no growth of MRSA measured for 45S5 and S53P4 particles of smallest size (Fig. 2B). Finally, we observed no difference for 45S5/1-2 mm and S53P4/90-710 µm, 500-700 µm and 1-2 mm after 72 h incubation. MRSA growth was recovered at levels similar to controls in the group of 45S5/90-710 µm and 500-700 µm. Interestingly, MRSA growth was slightly recovered in particle S53P4/32-125 µm, suggesting that MRSA could not be killed by this particle at 200 mg/ml concentration (Fig. 2C). The fact that MRSA was unable to grow in 45S5/32-125 µm suggests that this bioglass particle is the most effective at inhibiting MRSA.
To further assess antibacterial effects of bioglass particles, S. gordonii and P. aeruginosa were employed in this study. The results of Veillonella and MRSA have demonstrated that both bioglass particles of the smallest size (32-125 µm) showed the most robust antimicrobial capability, so our studies for S. gordonii and P. aeruginosa focused on 45S5/32-125 µm and S53P4/32-125 µm. As shown in Fig. 3A, when 50 mg/ml bioglass particles were incubated with S. gordonii DL1, the growth of this strain was completely inhibited by 45S5 and S53P4 of the smallest size for 24 h and 48 h time points. Additionally, compared to bacterial control, particles 45S5 and S53P4 of the largest size (1-2 mm) can partially reduce DL1 growth after 24 h and 48 h incubation. As expected, when treated by 100 mg/ml bioglass particles, DL1 cannot grow in broth with 45S5/32-125 µm and S53P4/32-125 µm. Interestingly, 45S5/1-2 mm completely inhibited DL1 growth, and S53P4/1-2 mm showed slight inhibitory effect (Fig. 3B). As P. aeruginosa PAO1 is a clinical isolate, it may be more resistant to bioglass challenge. As expected, low concentrations of particles (100 mg/ml and 200 mg/ml) have no effect on PAO1 growth (data not shown). Thus, 500 mg/ml particles were used in this study (Fig. 3C). After 24 h incubation, 45S5/32-125 µm and S53P4/32-125 µm appear to impair PAO1 growth, and 45S5/500-700 and S53P4/500-700 µm showed slight inhibitory effect. Unexpectedly, PAO1 growth was further reduced in the mixture with 45S5 particles at 48 h, and we could not detect PAO1 growth in 45S5/32-125 µm, implying this bacterium might lyse after long-term incubation with 45S5 particles. Contrarily, PAO1 continued to grow in the presence of S53P4, and no difference was observed between bacterial control and S53P4/500-700 µm at 24 h and 48 h time points.

Antibiofilm effects of bioglass particles on bacteria
A biofilm is a group of microbial cells that are enclosed in a complex extracellular matrix. Microbes in biofilms are protected by the biofilm matrix, which confers resistance to various antimicrobial agents [24,25]. To assess antibiofilm effects of bioglass particles, established biofilms of MRSA and V. parvula were treated with all size particles of 45S5 and S53P4. For the MRSA biofilm assay, the same concentration (200 mg/ml) used in antibacterial assays was used in this test. Interestingly, both 45S5 and S53P4 particles at 200 mg/ml completely impaired MRSA established biofilm (data not shown). Thus, we reduced particle concentrations to 100 mg/ml. As shown in Fig. 4A, after treatment with 100 mg/ml particles for 6 h, most bioglass particles except for S53P4/1-2 mm strongly reduced MRSA biofilm, and no obvious difference was observed among these bioglass particles. S53P4/1-2 mm slightly reduced MRSA biofilm compared to control.
To assess the antibiofilm activity of bioglass particles on Veillonella biofilm, 100 mg/ml bioglass particles were used to treat V. parvula PK1910 mature biofilm. As shown in Fig. 4B, all 45S5 particles can impair PK1910 biofilm, and 45S5/32-125 µm showed greater effect on antibiofilm compared to other size 45S5 particles. Similar to 45S5 group, all size S53P4 particles reduced PK1910 biofilm: S53P4/32-125 µm showed the strongest effect, and other particles slightly reduced mature PK1910 biofilm (Fig. 4B,  C).

Effects of bioglass particles on established biofilm of clinical sample on hydroxyapatite discs
To further assess the antibiofilm effects of bioglass, clinical samples were used in this study to form a complex multi-species biofilm on hydroxyapatite (HA)

Discussion
A series of independent and confirmatory experiments showed that regardless of other tested conditions, particle size was the key determinant of antibacterial activity; smaller particle diameters had greater effectiveness. Further, the role of the composition of the bioglass in antimicrobial efficacy varied among the tested strains, which responded differently to encounters with these antimicrobial particles. It has been reported that bioglass 45S5 and S53P4 have an antimicrobial effect against Streptococcus mutans [26,27], in this study, another important Streptococcus species, S. gordonii showed the most overall growth inhibition compared to other tested bacteria, suggesting streptococci are sensitive to bioglass antibacterial effect. S. gordonii effects also seemed to be more composition dependent, whereas V. parvula levels were similar for both compositions. In the case of S. gordonii, growth inhibition was dependent on the type of particle used across most sizes and timepoints favoring the 45S5 parent composition. Similarities between the two particle types holds true across tested particle sizes and timepoints. The timepoint measured makes a significant difference, with a much stronger inhibitory effect at 24 h and bacterial growth nearly reaching control levels after 48 h for larger particle sizes. This suggests that even though the bacteria are very susceptible to the bioglass at 24 h, coping mechanisms may develop that provide tolerance to particles so bacteria can recover. An alternative explanation is that diameter related effects of bioglass, such as dissolution rate, ionic exchange ratios and pH-modulation, which all vary across particle ranges may have key roles in antimicrobial effects. Zhang et al. have demonstrated that bioglass inhibited the proliferation of P. aeruginosa at a concentration of 100 mg/ml [28], however, we found a clinical wound isolate P. aeruginosastrain PAO1 was able to survive at 100 mg/ml and 200 mg/ml bioglass concentrations, and partially inhibited at a high bioglass concentration (500 mg/ml). This is likely due to the higher tolerance of antibiotics and antimicrobials in clinical isolates. For both 45S5 and S53P4, the smallest size particles showed the strongest antibacterial effects for all tested bacteria. It has been reported that the bioactive glass can release ions from surface to increase osmotic pressure and pH in the environment, thus making the environment antagonistic to microbial growth [29]. So, the smaller particles have a greater surface area per unit mass, and then have increased potential to change environment. Furthermore, the antibiofilm activity of bioglass S53P4 has been well studied. S53P4 showed the strong activity to reduce the biofilm produced by a broad range of microorganisms, such as S. aureus, P. aeruginosa, K. pneumoniae, A. baumannii and S. epidermidis [30][31][32].
In our study, 45S5 particles produced a stronger reduction in biofilm mass compared to S53P4 particles when considering small particle ranges, suggesting that this glass type has a greater potential for eradicating mature biofilm structures. Nonetheless, particle size remained the key determinant of effectiveness of these particles against biofilms as smaller particles reduced biofilms significantly more than larger particles across all experiments and consistent with planktonic assays. This is likely because smaller particles are closer in size to the bacteria, and thus are more apt to interact with and destroy bacterial biofilms. 32-125 µm particles of either type show much less remaining biofilm compared to each of the larger sizes, suggesting that this size of particle is most ideal for biofilm destruction. Interestingly, while the majority of previous published studies on antimicrobial effects of bioglass have focused on S53P4, the 32-125 µm 45S5 particles performed better at reducing biofilm biomass both against V. parvula and S. gordonii biofilms suggesting that the parent 45S5 may be more effective in eradicating infections caused by biofilms. In addition, the fact that the 32-125 µm 45S5 bioglass particles can kill and remove robust ex vivo clinical multi-species biofilms might help in deriving novel therapeutics for treatment and prevention of infectious diseases related to biofilm.

Conclusion
In conclusion, there is ample evidence that suggests 45S5 has a greater degree of antibacterial activity when compared to S53P4. Therefore, the indications for 45S5 could be broadened to include the indications of S53P4 for clinical deployment in infected sites in vivo. Additionally, formulations of bioglasses could be optimized to contain mostly smaller size particles (32-125 µm) for improved infection treatment purposes. However, more studies should be conducted to consider if other types of particles and specifications could make a particle even more ideal for treatment of these infections and if these inhibitory effects are also seen in a clinical setting.