Introduction

Bone and joint infection (BJI) is often characterized by severe inflammation and progressive bone destruction. Osteocytes are the most numerous and long-lived bone cell type, and therefore represent a potentially important long-term reservoir of bacterial infection. Staphylococcus aureus is known to establish stable intracellular osteocytic infections, however, little is known about the less virulent yet second most prevalent BJI pathogen, S. epidermidis, associated with late-diagnosed, chronic BJI. Thus, this study sought to establish an in vitro model to study the infection characteristics of S. epidermidis in human osteocyte-like cells.

In chronically infected fracture non-unions, treatment requires extensive debridement to remove necrotic and infected bone, often resulting in large defects requiring elaborate and prolonged bone reconstruction. One approach includes the induced membrane technique (IMT), although the differences in outcome between infected and non-infectious aetiologies remain unclear. Here we present a new rabbit humerus model for IMT secondary to infection, and, furthermore, we compare bone healing in rabbits with a chronically infected non-union compared to non-infected equivalents.

A 5 mm defect was created in the humerus and filled with a polymethylmethacrylate (PMMA) spacer or left empty (n=6 per group). After 3 weeks, the PMMA spacer was replaced with a beta-tricalcium phosphate (chronOs, Synthes) scaffold, which was placed within the induced membrane and observed for a further 10 weeks. The same protocol was followed for the infected group, except that four week prior to treatment, the wound was inoculated with Staphylococcus aureus (4×10⁶ CFU/animal) and the PMMA spacer was loaded with gentamicin, and systemic therapy was applied for 4 weeks prior to chronOs application.

All the animals from the infected group were culture positive during the first revision surgery (mean 3×10⁵ CFU/animal, n= 12), while at the second revision, after antibiotic therapy, all the animals were culture negative. The differences in bone healing between the non-infected and infected groups were evaluated by radiography and histology. The initially infected animals showed impaired bone healing at euthanasia, and some remnants of bacteria in histology. The non-infected animals reached bone bridging in both empty and chronOs conditions.

We developed a preclinical in vivo model to investigate how bacterial infection influence bone healing in large defects with the future aim to explore new treatment concepts of infected non-union.

Fracture related infection remains a major challenge in musculoskeletal trauma surgery. Despite best practice, treatment strategies suffer from high failure rates due to antibiotic resistance and tolerance. Bacteriophages represent a promising alternative as they retain activity against such bacteria. However, optimal phage administration protocols remain unknown, although injectable hydrogels, loaded with phage and conventional antibiotics, may support conventional therapy.

In this study we tested the activity of meropenem, and two newly isolated bacteriophages (ϕ9 and ϕ3) embedded within alginate-chitosan microbeads and a hydrogel. Antibiotic and phage stability and activity were monitored in vitro, over a period of 10 days. In vivo, the same material was tested in treatment of a 5-day old Pseudomonas aeruginosa infection of a tibial plate osteotomy in mice. Treatment involved debridement and 5 days of systemic antibiotic therapy plus: i- saline, ii-phages in saline, iii-phages and antibiotics loaded into a hydrogel (n=7 mice/group). To assess the efficacy of the treatments, the infection load was monitored during revision surgery with debridement of the infected tissue after 5,10 and 13 days (euthanasia) by CFU and PFU quantification.

In vitro testing confirmed that the stability of meropenem and activity of ϕ9 and ϕ3, was not affected within the alginate beads or hydrogel over 10 days. The in vivo study showed that all mice receiving phages and antibiotics loaded into a hydrogel survived the infection with a reduction of the bacterial load in the soft tissue. Active phages could be recovered from the infected site at euthanasia (10⁴ PFU/g).

The hydrogel loaded with bacteriophages and meropenem showed a positive result in locally reducing the infection load indicating a synergistic effect of the selected antimicrobials. Overall, our new strategy shows encouraging results for improving the treatment of antibiotic-resistant biofilm infections that are related to medical implants.

Device-associated infection remains a serious clinical problem in orthopaedic and trauma surgery. The emergence of resistant organisms such as methicillin resistant Staphylococcus aureus (MRSA) has further exacerbated this problem by limiting the range of treatment options. Currently, systemic antibiotic therapy is the cornerstone of treatment, alongside surgical resection of infected tissues and implant removal. The potential for antibiotic loaded biomaterials to support the prevention and treatment of infection is significant, although the currently available options are limited in number and often re-purposed from other applications e.g. antibiotic loading of bone cement. The first part of the talk will cover the basic concepts involved in antibiotic treatment, with an emphasis on the ideal antibiotic release kinetics from biomaterials, and how bacterial biofilms and antibiotic resistance influence antimicrobial efficacy. The next generation of biomaterials for antibiotic delivery should be specifically designed with this knowledge in mind. Regulatory approval of antimicrobial combination devices, however, is an evolving process as regulatory bodies seek more robust and clinically relevant efficacy data. Approval will require preclinical efficacy using standardized animal models that recapitulate the key features of the clinical disease. The second part of this talk will cover best practice in this important stage of development.

This longitudinal microCT study revealed the osteolytic response to a Staphylococcus epidermidis-infected implant in vivoand also demonstrates how antibiotics and/or a low bone mass state influence the morphological changes in bone and the course of the infection.

Colonisation of orthopaedic implants with Staphylococcus aureusor S. epidermidisis a major clinical concern, since infection-induced osteolysis can drastically impair implant fixation or integration within bone. High fracture incidence in post-menopausal osteoporosis patients means that this patient group are at risk of implant infection. The low bone mass in these patients may exacerbate infection-induced osteolysis, or alter antibiotic efficacy. Therefore, the aims of this study were to examine the bone changes resulting from a S. epidermidisimplant infection in vivousing microCT imaging, and to determine if a low bone mass stateinfluences the course of the infection and the efficacy of antibiotic therapy. An in vivomodel system using microCT scanning [1], involving the implantation of either a sterile or a S. epidermidis-colonised PEEK screw into the proximal tibia of 24 week-old female Wistar rats, was used to investigate the morphological changes in bone following infection over a 28 day period. In addition, the efficacy of a combination antibiotic therapy (rifampin and cefazolin: administered twice daily from days 7–21 post-screw implantation) for affecting osteolysis was also assessed. A subgroup of animals was subjected to ovariectomy (OVX) at 12 weeks of age, allowing for a 12 week period for bone loss prior to screw implantation at 24 weeks. Bone resorption and formation rates, bone-implant contact and peri-implant bone volume in the proximity of the screw were assessed by microCT scanning at days 0, 3, 6, 9, 14, 20 and 28 days post-surgery. Following euthanasia at day 28, the implanted screw, bone and soft tissues were subjected to quantitative bacteriology as a measure of the efficacy of the antibiotic regimen. In non-OVX animals S. epidermidisinfection induced marked osteolysis, which peaked between 9 and 14 days post-screw implantation. Peak bone resorption was detected at day 6, before recovering to baseline levels at day 14. Infection also resulted in extensive deposition of mineralised tissue, initially within the periosteal region (day 9–14), then subsequently in the osteolytic region at day 20–28. Quantitative bacteriology indicated all non-OVX animals remained infected. Rifampin and cefazolin successfully cleared the infection in 5/6 non-OVX animals group although there was no difference observed in CT-derived bone parameters. OVX resulted in extensive loss of trabecular bone but this did not alter the temporal pattern of infection-induced osteolysis, or mineralised tissue deposition, which was similar to that observed in the non-OVX animals. Similarly, there was no difference in bacterial counts between non-OVX and OVX animals (39,005 colony-forming units (CFU) [range: 3,675–156,800] vs 37,665 CFU [range 3,250–84,000], respectively). Interestingly, antibiotic treatment was less effective in the OVX animals (3/5 remained infected), suggesting that antibiotics have reduced efficacy in OVX animals. This study demonstrates S. epidermidis-induced osteolysis displays a similar temporal pattern in both normal and low bone mass states, with comparable bacterial loads present within the localised infection site.

Summary

An in vivo model of implant infection was developed to assess immune response. Titanium and PEEK implants were tested in the presence of an osteotomy and Staphylococcus aureus contamination. Immune response differed yet the outcome of contamination did not.

Summary

Staphylococcus aureus isolates from Fracture fixation device related infections contained fewer isolates that form a strong biofilm in comparison with isolates from Prosthetic joint infections. Both orthopaedic implant related infection groups possessed fnbB and sdrE more frequently than the non-implant related infection groups.

Summary Statement

In vivo microCT allows monitoring of subtle bone structure changes around infected implants in a rat model.