Irradiation-Sterilized Human Bone Allograft Toughened by Glyco-Oxidation Crosslinking

Introduction: Bone embrittlement due to γ-irradiation sterilization (25-35kGy range) is a concern for tissue banks (product quality) and for orthopaedic surgeons (clinical effectiveness and outcomes). This is especially true in cases of large structural allografts (e.g. massive intercalary defect reconstruction) because only approximately 20% of such a graft will be invaded and remodelled by host bone cells (5y; host-graft junction) [1]. The rest of the graft remains dead and yet must tolerate damage caused by millions of cycles of loading per year for many years. Fracture is a recognized mode of graft failure [1] and irradiation may increase this rate of fracture [2]. Irradiation-sterilized bone is thought to become brittle and less damage tolerant because the collagen becomes fractured [3, 4] and collagen network connectivity is lost [4]. Such irradiated bone acts more like a brittle ceramic with greatly reduced toughness [4, 5], fracture toughness [4, 6, 7], fatigue resistance [8] and altered microdamage formation [5]. Various alternatives include avoiding irradiation altogether, switching to a form of chemical sterilization or irradiating in the presence of radio-protective agents (e.g. free radical scavengers) [3]. Our novel approach consists of pre-treating the bone with a glyco-oxidation crosslinking (GOC) agent prior to irradiation to allow effective sterilization and producing a new toughening phase made of in situ modified collagen. Based on our previous work demonstrating that irradiation degrades collagen connectivity and that tougher irradiation-sterilized bovine cortical bone was produced when collagen network connectivity was restored [4], we hypothesized that glyco-oxidation crosslinking (GOC) pretreatment prior to irradiation would also toughen γ-irradiation-sterilized human cortical bone. This would provide a proof-ofprinciple that various endogenous crosslinking agents might be applied to toughen irradiation-sterilized bone allograft. Methods: This study used cortical bone from the diaphysis of five left human femurs. The donors were all male aged between 59 and 67 years. The bones were sourced through our collaborator Mount Sinai Allograft Technologies after ethics approval from the hospital review board. From each diaphysial site (proximal, mid-shaft, distal), six matched beams (termed a set) were cut with a metallurgical saw and hand polished to a 1-um surface finish. The beams measured 2x4x50 mm (±10% max.). 15 sets were tested. One beam from each set was randomly assigned to one of six groups: non-irradiated controls (N; kept frozen until testing), irradiated controls (I), and four GOC pre-treatment groups. The I and GOC groups were incubated in PBS at 60°C for 24 hours. The GOC agent was ribose and the buffers contained increasing concentrations of ribose (GOC1 = 0.06M, GOC2 = 0.3M, GOC3 = 0.6M, and GOC4 = 1.2M). The I and GOC groups were then packed on dry ice and irradiated at 34kGy (±10%). All specimens were then thawed and rehydrated in PBS for four hours before mechanical testing. The beams were tested to failure in three-point bending following ASTM D790 as closely as possible using an Instron ElectroPuls E1000 testing machine. Flexural modulus (E), yield stress and strain (σy and εy), ultimate stress (US), failure strain (εf) and work-of-fracture (Wfx) were determined from the test data. Statistical Analyses: Repeated measures ANOVA with Holms-Sidak tests post-hoc were used to test between group differences at the 95% confidence level (p<0.05). Results: γ-irradiation sterilization at 34kGY reduced the ultimate strength by 8% (p > 0.05), failure strain by 19% (p < 0.0001), and work-to-fracture by 29% (p < 0.0001) on average but did not detectably affect the flexural modulus (p=0.43) or yield point (p=0.22). GOC pre-treatment resulted in notable improvements of the affected measures in a concentration dependent manner. See Table 1 and Figure 1. In the GOC4 group, US reached N levels (106%; p > 0.05 N vs. GOC4) and GOC4 recovered 60% of the strain at failure and 76% of the Wfx lost in I (both cases: p < 0.001 GOC4 vs. I; p > 0.05 GOC4 vs N). The effects of I and the GOC treatments were not detectably dependent upon diaphysial harvest site (p > 0.05).