Matériaux biomimétiques de comblement pour des applications osseuses : influence du dopage de céramiques à base d'hydroxyapatite co-dopées par des carbonates et des silicates sur leur dégradation et leur potentiel de régénération osseuse

Calcium phosphate-based bioceramics, such as hydroxyapatite (HA) or β-tricalcium phosphate (β-TCP), are widely used in cranio-maxillofacial and orthopedic surgery owing to their chemical similarity to bone mineral, ensuring their excellent biocompatibility and well- established osteoconductivity. Doping HA with bioactive ions is a promising strategy to modulate ceramic properties and improve performance. Carbonate, the most abundant ionic substitute in bone mineral, enhances HA resorption when incorporated into the apatite lattice. Silicate substitution has also attracted interest, as its bioavailable form, Si(OH)4 was shown to stimulate osteogenic differentiation and mineralization in vitro. However, silicate- substituted HAs (SiHAs) has frequently shown limited degradability and solubility comparable to HA. On this basis, co-doping HA with carbonate and silicate (CSiHA) could couple the advantages of both dopants, enhancing ceramic degradation while supporting bone regeneration, yet its in vivo behaviour remains largely unexplored. The studies in this thesis investigated the impact of silicate and/or carbonate substitutions in HA on ceramic degradation and bone regeneration, focusing on the mechanisms that drive degradation. We directly compared C0.8Si0.4HA with monosubstituted (C0.8HA, Si0.4HA) and undoped HA controls, using β-TCP as a soluble ceramic reference. To isolate composition effects, all ceramics were synthesized and shaped by identical routes and sintered to yield equivalent microstructural features; substitutions were confirmed by XRD/FTIR/elemental analyses. We then assessed the influence of ionic substitution on bone regeneration and degradation, distinguishing passive dissolution from cell-mediated resorption using complementary in vivo models. In the subcutaneous model, all apatites and β-TCP exhibited negligible volume loss over 32 weeks, indicating minimal physicochemical dissolution in vivo. In osteogenic environments, osteoclast-mediated degradation increased over 8 weeks and varied by composition, following the trend observed under acidic in vitro conditions (pH 5.5): HA ≲ SiHA ≲ CHA < CSiHA ≈ β-TCP. Thus, in vivo, CSiHA was resorbed to an extent comparable to β-TCP and elicited a distinct osteoclast gene-expression profile. its behaviour is consistent with grain-boundary-first dissolution, likely reflecting greater lattice destabilization than in the other apatites. Nevertheless, these resorption differences did not translate into increased bone formation in either the ectopic or the orthotopic (condyle) rodent models. The second study complemented the in vivo analyses by examining, under standardized conditions, the differentiation and resorptive capacity of human CD14⁺ monocyte–derived osteoclasts in vitro. Unexpectedly, the in vitro osteoclastic responses did not match the in vivo trends: CSiHA suppressed osteoclastogenesis, whereas CHA markedly promoted it, resulting in greater and more consistent resorption. Although silicate ion release from CSiHA was confirmed in vitro, it did not translate into enhanced bone formation in vivo. In a third study, we therefore evaluated the effects of orthosilicic acid (Si(OH)₄) across a concentration range of 0.03–3 mM on human mesenchymal stem cells. Only the highest dose (3 mM) modestly increased viability, proliferation, and matrix mineralization under osteogenic conditions. In summary, CSiHA underwent osteoclast-mediated resorption in vivo to an extent comparable to β-TCP and greater than control apatites. This enhanced resorption, however, was not reproduced in vitro with human osteoclasts. Furthermore, silicate release during degradation did not improve bone formation, consistent with in vitro findings showing only minimal effects of Si ions on hMSC mineralization.