A novel antibacterial compound decreases MRSA biofilm formation without the use of antibiotics in a murine model

Despite significant advancements in material science, surgical site infection (SSI) rates remain high and prevention is key. This study aimed to demonstrate the in vivo safety and antibacterial efficacy of titanium implants treated with a novel broad‐spectrum biocidal compound (DBG21) against methicillin‐resistant Staphylococcus aureus (MRSA). Titanium (Ti) discs were covalently bound with DBG21. Untreated Ti discs were used as controls. All discs were implanted either untreated for 44 control mice or DBG21‐treated for 44 treated mice. After implantation, 1 × 107 colony forming units (CFU) of MRSA were injected into the operating site. Mice were killed at 7 and 14 days to determine the number of adherent bacteria (biofilm) on implants and in the peri‐implant surrounding tissues. Systemic and local toxicity were assessed. At both 7 and 14 days, DBG21‐treated implants yielded a significant decrease in MRSA biofilm (3.6 median log10 CFU [99.97%] reduction [p < 0.001] and 1.9 median log10 CFU [98.7%] reduction [p = 0.037], respectively) and peri‐implant surrounding tissues (2.7 median log10 CFU/g [99.8%] reduction [p < 0.001] and 5.6 median log10 CFU/g [99.9997%] reduction [p < 0.001], respectively). There were no significant differences between control and treated mice in terms of systemic and local toxicity. DBG‐21 demonstrated a significant decrease in the number of biofilm bacteria without associated toxicity in a small animal implant model of SSI. Preventing biofilm formation has been recognized as a key element of preventing implant‐related infections.

Bacterial infections, especially when associated with biofilm, represent one of the most serious complications following the implantation of medical devices. 1 Surgical-site infections (SSIs) and periprosthetic joint injections (PJIs) are particularly devastating for orthopaedic patients as antibiotics have difficulty reaching dormant bacteria in low-nutrient microenvironments such as implant surfaces and bone. 2 Biofilm is an exopolysaccharidic matrix comprised of bacteria with reduced antibiotic sensitivity and poor mechanical accessibility.
Biofilm formation plays a central role in the failure of conservative treatments (debridement, antibiotics, irrigation, and implant retention) for implant-related infections. 3Preventing biofilm formation has been recognized as a key element of SSI and implant-related infection prevention.Yet, in most medical disciplines, nothing new has been implemented in clinical practice that effectively reduces biofilm formation at the surface of implants.Bacterial biofilm is strongly associated with failure of infection control, infection recurrence, and the development of chronic infections. 4e emergence of antibiotic-resistant bacterial strains has pushed for the development of nonselective antibacterial coatings and next-generation innovative solutions. 5Those attempts have been unsuccessful at translating from the bench to the bedside to date.
8][9][10][11][12][13] Nanostructured surfaces aimed at preventing bacterial adhesion are not bactericidal and therefore, display a low efficacy in preventing the emergence of biofilm in vivo. 14Ideally, a permanent surface modification that would be both antibacterial and biocompatible would reduce the incidence of implant-related infection and make implant removals unnecessary.Surface modification of implants is superior to coatings in terms of stability and durability, since coatings tend to scale off with time, while surface modification relies on covalent bonding of compounds on the target substrate. 15This is the rationale behind the clinical need for permanent antimicrobial protection of surfaces which transient eluting coatings cannot fully address.
An ideal antimicrobial surface protection should be able to support the following claims: prevention of implant-related infection, long-lasting protection of implant surfaces from late onset bacterial hematogenous spread, indirect decrease of surrounding tissue bacterial load by drastic biofilm inhibition, excellent local and systemic biocompatibility profile, stability (no release of potentially toxic compounds), full sterilizability, scalability, cost-effectiveness, and no alteration of bone ingrowth.Therefore, a novel ready-to-use antimicrobial compound (DBG21) graftable on titanium-alloy implants (Ti-6Al-4V) was developed with the goal to form a permanently modified surface that would inhibit the growth of biofilm.DBG21 is a bifunctional polymer that has both the ability to covalently bind to surfaces and to eliminate bacteria.To proceed with a preclinical assessment, the objectives of this study were (1) to assess safety (local and systemic toxicity study of treated surfaces versus nontreated surfaces implanted subcutaneously in mice in the absence of infection); (2) to assess the antibacterial efficacy of treated implant surfaces versus nontreated implant surfaces using a previously validated MRSA subcutaneous infection murine model; (3)   to provide pathology data on the peri-implant tissue response in the setting of the same infection murine model.

| Animals
A total of 88 BALB/c mice (11-week-old, 22-24 g) were used (Charles River, France) for the entire study (Table 1).These animals were housed in a protected area at the small animal facility of the University of Burgundy, Dijon, France (Biosafety level 2 facility) and were fedad libitum

| Titanium-alloy implants
Forty-four Ti-6Al-4V discs, 6 mm Ø, 0.5 mm thick, were sonicated in pure ethanol for 10 min to remove impurities, air-dried, activated, dip-coated in an ethanolic polymer solution (proprietary DBG21 highdensity quaternary ammonium polymer, DeBogy Molecular Inc.) and heated at 130°C for 3 h to produce covalently bound DBG21-treated discs.All the unbound polymer was removed by sonicating DBG21treated discs in pure ethanol for 30 min.Discs were then air-dried.
Forty-four control Ti-6Al-4V discs underwent the same activation process without being treated with the polymer solution.All discs were individually stored in double peel packs.Following packaging, all discs underwent subsequent sterilization by 25 kGy irradiation using a Cobalt-60 gamma irradiator (VPT Rad).

| Tolerance study
The experimental model was performed as previously described. 16iefly, mice were anesthetized by an intraperitoneal (IP) injection of

| Efficacy study
The surgical insertion of implants was performed as described above.
Immediately after disc implantation, an inoculum of 100 µL (1 × 10 7 CFU/mouse of ATCC 43300 MRSA strain) of the bacterial culture was injected onto the implant.Mice received either an untreated implant or a DBG21-treated implant.The day of infection was referred to as D0.At Day 7 (D7), 36 mice (18 control mice, 18 treated mice) were killed by cervical dislocation performed under anesthesia.The implant was collected and then used for bacterial enumeration.At D14, the same procedure was repeated with the 30 remaining mice (15 control mice, 15 treated mice).Histopathological analysis was performed in three additional mice of each group since histopathological analysis and microbiological enumeration cannot be performed in the same mice.

| Bacterial load on implants
Each implant was individually washed under aseptic conditions in an Eppendorf tube (three successive washing steps with 300, 400, and 500 µL of sterile saline).After the last wash, the implant was suspended into 1 mL of sterile saline, placed into an ultrasonic bath (AdvantageLab) for 3 min at room temperature before being vigorously vortexed to detach all adherent bacteria from the implant.Several successive dilutions of this suspension (undiluted, 10 −2 , 10 −4 ) were then cultured onto Chapman agar plates for 24-48 h at 37°C.If required, dilutions were repeated in case of unconvincing or inadequate results, the stability of the suspension at 4°C for 48 h having been previously verified.The limit of detection was determined as being 1 log 10 CFU on implants.

| Bacterial load in surrounding tissues
Adjacent tissue (mostly subcutaneous fat and occasional muscle fibers) of each implant were dissected, weighed, resuspended into 1 mL of saline solution and homogenized using a bead beating grinder and lysis system (FastPrep-24 5G, MP Biomedical; 1 cycle of 30 s at 4 m/s with 1 ceramic beads).Crushed tissues were serially diluted down to 10 −6 and 10 µL of each dilution was plated by spotting-and-tilt-spreading (SATS) approach on Chapman agar plates.The limit of detection was determined as being 1 log 10 CFU/g of surrounding tissues.

| Pathology analysis
During the necropsy at D7 and D14 postimplantation for infected mice and D11 for noninfected mice, the tissue surrounding the implant was excised from animals and then kept in a histological cassette to avoid distortion of the sample and fixed in 10%   Of note, 10 out of 14 (71.4%)mice with DBG21-treated discs had no bacterial growth in the surrounding tissues while all of the control mice had bacterial contamination in the surrounding tissues.

| Weight and clinical scores
The day following the surgical procedure, mice with treated implants lost 10% of their body weight but regained a weight level at Day 2 comparable to that of animals with untreated control implants.
During the 11 days following the subcutaneous implantation, mice gained a similar amount of weight in both the control group and the DBG21-treated group. 19There was no observed difference in weight between treated and control animals (p > 0.05) (Figure 2A).Clinical scores 17 were not statistically different between DBG21-treated and control mice (p > 0.05) (Figure 2B).All surgical wounds healed uneventfully.

| Toxicity tests
Eleven days after subcutaneous implantation, no significant biochemical alterations were recorded, regardless of the group (control or treated).No statistically significant differences between groups were observed (p > 0.05) (Figure 2C).

| DISCUSSION
Data from the American Joint Replacement Registry showed that infection is the number one cause of early and late failure for both hip and knee replacement. 20Despite best efforts of prevention, the number of primary hip and knee replacement leading to infection is still on the rise and affects patient quality of life, risk of infection recurrence, amputation, and death.In this in vivo study, we demonstrated that a novel covalently grafted antimicrobial compound significantly decreased MRSA growth both on a titanium disc and in the surrounding soft tissues.Moreover, our finding also indicates the absence of local and systemic toxicity.High bacterial median log 10 reductions on implants were achieved at both 7 and 14 days postoperatively despite the use of a high inoculum, the absence of antibiotics, and the use of a virulent strain with known capability to adhere to biomaterials (MRSA).Furthermore, the bacterial reductions observed in the soft tissues improved with time between Day 7 and Day 14. DBG21-treated implants did not induce any acute or subacute systemic toxicity in mice (ISO 10993-11:2017 17 ).The histopathological analysis revealed no differences in local toxicity between the treated and control groups.Taken together, these findings strongly support an excellent biocompatibility profile of DBG21-treated titanium implants.
In this investigation, given that the compound was designed to be noneluting, one may hypothesize that the strong biofilm inhibition Comparison of the bacterial reductions on titanium implants between DBG21 and data from previously published peer-reviewed studies (Table 2) showed that the current study had one of the longest (14 days) time periods of observation after MRSA inoculation.The bacterial reductions reported on implants and in the surrounding tissues in the present study outperformed the scientific literature on comparable subcutaneous infection rodent models with titanium implants.
Indeed, most published studies do not exceed 2.5-2.75 log 10 bacterial reductions both in the surrounding tissues and on implants for bacterial strains with less virulence (Staphylococcus epidermidis), lower inoculum (6 log 10 and below), and shorter time points (under 7 days).Also, most of the processes described in the literature are exceedingly complex, requiring the use of toxic reagents and/or solvents, and do not meet standards of scalability in the orthopaedic industry.
Interestingly, the bacterial log reductions observed in our study in the surrounding tissue that improved between Days 7 and 14 were not found in the literature pertaining to covalently bound antimicrobials. 9,10Typically, authors reported no reduction of bacterial growth in surrounding tissues for covalently bound antimicrobials.This could be due to their lower bactericidal effect on contact and thus lower biofilm inhibition since biofilm and tissue   Indeed, antibiotics are known to be synergistic with antimicrobial surface technologies. 6Second, we acknowledge that this small animal trial does not provide data regarding bone ingrowth in contact with the implant.A large animal trial is planned to provide data on efficacy in a bone-relevant model and on osseointegration, which is clinically paramount.This trial was designed as an initial proof of concept for antimicrobial surface modification of medical devices, not solely for orthopedic applications, as titanium-alloy is a gold-standard metal for numerous other medical applications.
Small animal models also allow the collection of more complete safety and efficacy data with the use of larger sample sizes.Third, we acknowledge the low sample size in the safety study.We found no histological differences between treated and control in the absence of infection.Moreover, our findings are consistent with Schaer et al. 23

| CONCLUSION
Overall, these results showed that the MRSA biofilm was reduced by up to 99.97% in mice that received DBG21-treated titanium-alloy implants compared to untreated control implants without causing measurable toxicity.This level of protection provided by a noneluting surface treatment despite a high bacterial inoculum of a virulent strain is novel and has the potential to prevent implant-related infections in humans.
according to the current recommendations of the European Institute of Health.No fasting was required for this study.Before each experiment, animals were housed for 1 or 2 weeks at the animal facility.During this period and for the duration of the study, qualified staff members checked on animals twice a day and assessed their well-being.The animal facility was authorized by the French authorities (Agreement N°C 21 464 04 EA).Animal housing and experimental procedures were conducted according to the French and European Regulations and NRC Guide for the Care and Use of Laboratory Animals.All procedures using animals were submitted to the Animal Care and Use Committee C2EA accredited by the French authorities (APAFIS #33499-2021101914348582).

a
mix of ketamine (50 mg/kg) (Virbac) and xylazine (10 mg/kg) (Bayer HealthCare).The flank on the right side was shaved and then disinfected by three consecutive applications of betadine/sterile water.A cutaneous incision of 0.5 cm was made under sterile conditions and an untreated implant (control) or a treated implant (DBG21-treated) was subcutaneously inserted and placed at about 2 cm from the incision site.Five mice received an untreated implant (control) and five mice received a treated implant.The incision was sutured and disinfected once a day for 3 days after surgery.Based on the ISO 10993-11:2017 standard, 17 systemic acute and subacute toxicity was evaluated in mice receiving DBG21-treated implants compared to mice receiving untreated implants (controls) in the absence of infection.These mice were monitored daily over an 11day period (weight, daily clinical score).Clinical scoring (ISO 10993-11:2017 standard 17 ) included: movement, body posture, fur quality, degree of eye opening, body weight.At Day 11, blood samples were collected through intracardiac puncture on all animals previously anesthetized via a mix of ketamine and xylazine.Euthanasia was performed immediately after the intracardiac puncture by cervical dislocation.A blood biochemical analysis included urea, creatinine, protein, albumin, liver function tests (alkaline phosphatase, alanine aminotransferase, glutamate dehydrogenase, total bilirubin), electrolytes (Na + , K + , Cl − ) and glucose.Blood was collected into purple-top potassium EDTA tubes and stored at 4°C until shipment.The relevant tubes were sent to Cerbavet (Massy) for analysis and results were available within 24 h.
formaldehyde.All samples (22 subcutaneous murine tissue specimens with titanium implants and apical orientation sutures) were then sent for paraffin-embedding and further histological analysis (Haematoxylin/Eosin/Saffron staining) to Atlantic Bone Screen (Saint-Herblain).The samples (subcutaneous tissue with the titanium implant) were processed at Atlantic Bone Screen.The samples were stored at room temperature in a dedicated location until the start of the experiments.The titanium discs were removed, and the tissue samples were embedded in paraffin and stained with Haematoxylin/ Eosin/Saffron.For each block, sections of 3-4 µm were made and placed on Superfrost slides.The slides were dried under a fume hood overnight at room temperature before being used for HES staining.The quality of the histological sections present on each slide was individually assessed before any processing.Similarly, the quality of each staining was individually checked at the end of the procedure.A veterinary pathologist further performed the histological analysis of the produced microscopic slides (graduate of the European College of Veterinary Pathologists).The veterinary pathologist separately documented, illustrated, and commented on any notable events.2.6.4 | Microscopic examinationAll samples (corresponding to a total of 22 sections, 10 from mice used in the tolerance study and 12 from mice in the antibacterial efficacy study) were observed by a veterinarian pathologist in a blinded fashion.All significant events were listed, recorded, and documented.Studied parameters were inflammation, fibrosis, vascularization (neo-angiogenesis), and necrosis.18

2. 7 |
Statistical analysisMicrobiology statistical analyses were performed with GraphPad Prism software using Mann-Whitney tests.Histological statistical analyses were performed using Kolmogorov-Smirnov tests.The results were expressed as the median ± interquartile range (IQR).IQR is the difference between the 75th and 25th percentiles of the data.p-values were calculated and specified as *p < 0.05; **p < 0.005; ***p < 0.001; ****p < 0.0001.

3 | RESULTS 3 . 1 |
Efficacy study 3.1.1| Evaluation of the efficacy of DBG21-treated versus untreated titanium implants in a MRSA biofilm murine model At Day 7 postinoculation, comparable bacterial loads were obtained in control groups (untreated) on implants and surrounding tissues.The median level ± IQR of bacterial colonization remained stable over the 14 day-period of infection in the tissues (7.18 ± 1.75 Log 10 CFU/g at D7 and 6.55 ± 1.99 Log 10 CFU/g at D14).A slight decrease in the bacterial load was observed on untreated control implants (6.51 ± 0.90 Log 10 CFU at D7 and 5.84 ± 1.68 Log 10 CFU at D14); thus, the bacterial colonization was overall quite stable.At D7 postinoculation, a significant decrease (p < 0.0001) in bacterial load was observed in animals that received the DBG21treated implants, (−2.69 Log 10 CFU/g in tissues and −3.57Log 10 CFU on implants).Of note, 3 mice out of 18 (16.6%)had no bacterial growth on their tissues and implants.At D14 postinoculation, this bacterial load decrease was still confirmed, with a more pronounced effect in the surrounding tissues compared to the implant (−5.55 Log 10 CFU/g in tissues and −1.93 Log 10 CFU on implants) (Figure1).

BOULOUSSA ET AL. | 205 3 . 2 . 3 |
Figure 3A,B.A side-by-side photographic comparison between a control and a treated mouse is shown in Figure 3C.After 11 days of subcutaneous implantation, the microscopic analysis of the HESstained slides showed no significant cytological alterations, no increased fibrosis or inflammation, and no necrosis or neo-angiogenesis in the treated group versus controls (p > 0.05).

3. 3 |study 3 . 3 . 1 |
Pathology Histopathological impact of DBG21-treated implants versus untreated titanium-alloy implants on surrounding tissues in an MRSA biofilm model The use of treated implants trended toward decreased inflammation, fibrosis, vascularization, and necrosis rates at D7 and D14 postimplantation, as seen in Figure 4 (p > 0.05).These results indicate that in the presence of an infection, DBG21-treated implants did not generate suppurative and necrotic inflammation over time, unlike control implants with the same bacterial inoculum.The analysis over the 2-week study time period of either control or DBG21-treated implant effects did not have any significant differences between the two time points.As observed in Figure 4, the presence of inflammation, fibrosis, vascularization, and necrosis on D7 persisted until D14 for control implants.

F I G U R E 1
Antibacterial efficacy of DBG21-treated titanium implants versus controls against methicillin-resistant Staphylococcus aureus (MRSA) (ATCC 43300) biofilm in a mouse model of implant-associated infection after 7 and 14 days of infection in surrounding tissues (A) and on implants (B).F I G U R E 2 (A) Body weight measurement (expressed as % body weight, mean ± SD) in DBG21-treated and untreated mice up to 11 days postimplantation.(B) Clinical scores (expressed as % body weight, mean ± SD) in DBG21-treated and untreated mice up to 11 days postimplantation.(C) Biochemical assessment (median is presented) from animals receiving untreated or DBG21-treated implants at D11 postimplantation.(A: Urea, B: Creatinine, C: Total Protein, D: Serum albumin, E: Alkaline Phosphatase, F: Transaminases S.G.P.T, G: Glutamate dehydrogenase, H: Sodium, I: Potassium, J: Chlore, K: Glycaemia).allowed the immune response to build over time and clear bacteria in the surrounding soft tissues.These findings have to be interpreted in light of the context: a stringent model using a high MRSA inoculum (1 × 10 7 CFU) with direct injection into the operative site after skin closure in the absence of perioperative antibiotics.This surpasses almost all scenarios of surgical as it was demonstrated based on experimental animal data that approximately 10 2 CFU of S. aureus are sufficient to establish infection if inoculated at the time of a hip hemiarthroplasty in a rabbit model.21Moreover, despite the smaller sample size of the histopathological efficacy substudy, there was demonstrated decreased inflammation, fibrosis, vascularization, and necrosis around the treated versus control discs.In the presence of infection, all these positive findings (bacterial counts and histopathological analysis) showed that DBG21-treated discs were able to mitigate infection, or even bring the infection under control in some of the mice, despite a high MRSA bacterial load and without the use of antibiotics.

F
I G U R E 3 (A) Representative picture of the implant cavity with different magnifications.Presence of an optically empty cavity (*) in the subcutaneous adipose-connective location, circumscribed by a light fibrous densification (arrow).Slight leukocytic densification in the loose connective tissue at the periphery of the polymorphic cavity, predominantly mononuclear (o).(B) Representative picture of the implant cavity with different magnifications.Absence of cavitary lesion in the subcutaneous connective tissue.Minimal leukocyte densification in loose, predominantly mononuclear connective tissue (o).(C) Representative pictures of the implant tolerance.Left panel: sample previously with neutral titanium implant.Right panel: sample previously with treated titanium implant.bacterial burden are not two separate entities but are interdependent mechanisms.As was shown over 60 years ago by Elek et al., 22 it is the presence of the foreign body that allows the infection to persist despite a low inoculum.Our investigation has some limitations.First, the antibacterial effect was more pronounced on treated implants at D7 than D14.This is most likely related to a limitation of the model to perform true bacterial enumeration on implants at Day 14.A capsule F I G U R E 4 Description of the effect of titanium implants + bacterial inoculum on the adjacent tissues at two timepoints (D7 and D14).(A-D) Results observed after 7 days of implantation.(E-H) Results after 14 days of implantation.(A, E) Inflammation rate.(B, F) Fibrosis rate.(C, G) Angiogenesis rate.(D, H) Necrosis rate.T A B L E 2 Comparative data for various antibacterial surface technologies.

|
209       formed on several discs at Day 14 in the presence of infection, which is a limitation of the animal model used.Because of this protocol, the capsule cannot be removed to count bacteria on the titanium surface.Therefore, the bacterial reduction described at Day 14 reflects what occurred either on the titanium surface or on a capsule, based on the amount of tissue response in the presence of infection.Interestingly, the soft tissues demonstrated no bacterial growth in most mice in the same timeframe.What occurs on the surface of the titanium after Day 14 remains an open question and most likely cannot be answered using this model.Indeed, bacteria are still detectable around peri-implant capsules in all mice at Day 14.The clinical impact is expected to be different as the standard treatment of implant-related infections involves the use of intravenous antibiotics, which were intentionally withheld from this animal trial to elucidate the specific role of DBG21.
whose team studied a different quaternary ammonium polymer (N,N-dodecyl,methyl-PEI coatings) on locking compression plates for tibial osteotomies in sheep.They found that their coating neither negatively impacted soft or calcified tissue physiology nor altered the healing response over at least a 3-month period.Schaer et al. also found less in vivo inflammation in the setting of infection on treated implants versus controls, which is in agreement with our study findings.Last, the present study was performed with MRSA only and further in-vivo studies are warranted to investigate the efficacy on other bacterial strains.Because quaternary ammonium polymers are already known to eliminate both gram-positive and gram-negative bacteria in vitro and in vivo, 24 similar findings could be possible in further DBG21 studies.Aside from cost, the treatment of implant-related and periprosthetic joint infections demands a minimum 6-week exposure to systemic antibiotics for the patient.These prolonged antibacterial treatments commonly cause profound microbiota disturbance producing gastrointestinal symptomatology and convey risk of both superimposed Clostridioides difficile colonic infection and drug-induced allergic reaction.Because treatment is often intravenous, catheter complications such as deep venous thrombosis and central line-associated bloodstream infection (CLA-BSI) are inherent risks.