Mechanical Loading of Ex Vivo Bovine Trabecular Bone in 3D-Printed Bioreactors

Introduction In 2021, 18.3% of the world’s population was diagnosed with osteoporosis [1]. Bone biomechanical behaviour studies have shown that better understanding of bone tissue mechanics are key in the development of more effective strategies to prevent bone fracture, regenerate, and repair osteoporotic bone defects [2]. Although polycarbonate (PC) bioreactors have been successfully used to study tissue mechanics in bone cores ex vivo [3], they are difficult and expensive to fabricate, and the height-diameter ratio for bone cores is below the standard for mechanical compression tests [4]. A bioreactor system was developed using polyjet 3D printing and MED610™, which has been suggested as suitable for long-term culture [5]. This system had been validated in a previous study [6], however, leakage and other experiment limitations were observed. Thus, the objective of this study was to improve and continue the validation of the new bioreactor system in a long -term ex vivo bovine trabecular bone core experiment.

Methods Twenty-two viable trabecular bone cores (10 mm height, 10 mm diameter) retrieved from a 3-year-old bovine sternum were individually cultured in custom -made 3D-printable MED610™ bioreactor chambers at 37°C and 5% CO ₂, with constant culture medium perfusion (6.6 mL/hr) , for 21 -days (Figure 1 a). Bone cores in bioreactors were ranked and sorted evenly based on their apparent elastic modulus ( 𝐸𝑎𝑝𝑝) into control and loading groups (n = 11/group). The load group was stimulated mechanically (Figure 1b) five times per week with a “trapezoid” waveform that mimicked a physiological high -impact stimulus, with a ∆3000 µƐ displacement and a strain rate of 3000 µƐ/s for 120 cycles. Bone core 𝐸𝑎𝑝𝑝 was assessed on days 0 and 21 in random order at room temperature with an initial 10 N pre-load, followed by a quasi-static -3000 µƐ strain at -50 µƐ/s. Data were analyzed in MATLAB 2020b (MathWorks) and 𝐸𝑎𝑝𝑝 was calculated assuming linear elastic behavior with the last 50% of the linear force-displacement curve . A two -way ANOVA and a Tukey post-hoc tests (α = 0.05) were used to compare statistical differences between groups on days 0 and 21. A two-tailed t-test was used to compare between groups for percent differences in 𝐸𝑎𝑝𝑝 after the 21 -days (%∆𝐸𝑎𝑝𝑝). Shapiro-Wilk tests confirmed normality across all groups. All statistical analysis were performed in SPSS® Statistics 23 (IBM) and GraphPad Prism® 6.1 (Dotmatics).

Results Bone cores tested on day 0 had a mean and a standard deviation 𝐸𝑎𝑝𝑝 of 97.7 ± 30.3 MPa, with no significant difference between groups after sorting. After 21-days, the 𝐸𝑎𝑝𝑝 decreased in all bone cores by 22.4% (± 20.6% SD) on average (Figure 1c). No significant differences in neither 𝐸𝑎𝑝𝑝 nor %∆𝐸𝑎𝑝𝑝 were observed between timepoints nor groups. Though not statistically different, there was a low % difference on day 0 between the 𝐸𝑎𝑝𝑝 means of the load and control groups (0.1%), which increased to 7.3% after 21-days. Thus, the load group’s 𝐸𝑎𝑝𝑝 decreased less, on average, than the control group (19.8% ± 19.6% decrease versus 25.0% ± 22.3%).

Discussion The 21-days experiment was successful: no system leakage nor infections were observed. And although data analyses results do not follow the same trends in 𝐸𝑎𝑝𝑝 changes as other ex vivo trabecular bone core studies using bioreactors [3,6], it is still relevant that the %∆𝐸𝑎𝑝𝑝 decrease was higher in the control group . It is thought, then, that the mechanical stimulation might have helped in the overall bone health of the load group, to some extent . Stress-relaxation, culture medium, and histology analysis are being done for a deeper understanding of the unexpected bone behavior observed. More studies are required to address the small sample size limitations, as well as other possible sources of error, to confirm if this 3D-printed bioreactor system can be used (or not) for long-term ex vivo organ cultures.