Ultrasmall Surface-Charge-Modified Tantalum Oxide Nanoparticles for the Assessment of Articular Cartilage Using Contrast-Enhanced Computed Tomography.

抄録

The size and surface properties of nanoparticles (NPs) define their interactions with biological tissues and impact their effectiveness for targeted imaging and therapeutic applications. Here, we investigate ultrasmall tantalum oxide nanoparticles (Ta₂O₅-NPs) as computed tomography (CT) contrast agents and the effects of NP surface charge on diffusion in ex vivo human cartilage samples, as well as after intra-articular administration in intact rabbit and equine joints. Controlled hydrolysis of tantalum(V) ethoxide, followed by coating with positively charged trimethylammonium and nonionic poly(ethylene glycol) (PEG) ligands at different ratios, affords the approximately 3 nm positively charged (Ta₂O₅-cNPs) and neutral (Ta₂O₅-nNPs) tantalum oxide nanoparticles. Ta₂O₅-cNPs readily diffuse into human cartilage as measured by microcomputed tomography, and the Ta₂O₅-cNPs partition positively correlates with key indicators of early-stage osteoarthritis to include the proteoglycan content, equilibrium Young's modulus, and porosity (stress-relaxation time constant) while inversely with viscosity (phase shift). In contrast, the neutral Ta₂O₅-nNPs reside at the cartilage surface, triple the attenuation difference at the cartilage-fluid boundary, accumulate within microscopic surface injuries, and enable clear detection of surface injuries and lesions. When extended to intact ex vivo animal models, Ta₂O₅-cNPs diffuse into rabbit knee cartilage over 24 h, while Ta₂O₅-nNPs delineate fissures and partial erosions in equine carpal cartilage. In conclusion, customizing the Ta₂O₅-NPs surface charge allows, or prevents, their diffusion into cartilage, enabling distinct CT imaging applications: diffusible Ta₂O₅-cNPs assess the cartilage structure and function, while nondiffusible Ta₂O₅-nNPs enhance cartilage segmentation and detect lesions in both human in vitro and animal ex vivo models.