Spatial analysis of the osteoarthritis microenvironment: techniques, insights, and applications

**Background:** Osteoarthritis (OA) is the most prevalent joint disorder, affecting an estimated 528 million people worldwide. It is characterized by molecular malfunction and anatomic degeneration of the osteochondral unit, including cartilage degradation, bone remodeling, osteophyte production, and joint inflammation. Current diagnosis relies on clinical symptoms and X-ray, which provides limited information for nonmineralized tissues like cartilage. The complexity and patient diversity of OA pathogenesis hinder the discovery of robust early biomarkers and effective disease-modifying treatments. Deep spatial phenotyping—using advanced imaging to quantify morphological, functional, and biochemical changes across tissue matrices—offers a way to overcome these challenges by preserving spatial architecture and enabling molecular-level analysis.

**Methods:** This narrative review surveys state-of-the-art spatial phenotyping techniques applicable to OA research. Techniques are categorized into: (1) vibrational spectroscopic imaging (near-infrared [NIR], mid-infrared/Fourier transform infrared [MIR/FTIR], nano-FTIR, Raman spectroscopy); (2) spatial multiomics imaging (spatial transcriptomics, spatial proteomics via mass spectrometry imaging [MSI] including MALDI-TOF and DESI, targeted multiplexed methods like digital spatial profiler [DSP], imaging mass cytometry [IMC], multiplexed ion beam imaging [MIBI]); (3) spatial lipidomics and metabolomics (MALDI-TOF, TOF-SIMS); (4) functional enzyme imaging (functional MSI [fMSI]); (5) elemental phenotyping imaging (LA-ICP-MS, SR-μXRF, TOF-SIMS); and (6) mechanical phenotyping (microindentation, nanoindentation, atomic force microscopy [AFM]). The review also discusses sample preparation challenges, AI/deep learning for data analysis, and existing databases.

**Key Results:** - **Vibrational spectroscopy:** NIR imaging (resolution ~100 μm) can detect proteoglycan (PG) loss via C-H, N-H, O-H bond changes; studies show NIR arthroscopy correlates with Knee Injury and Osteoarthritis Outcome Score and can classify healthy vs. OA cartilage with high accuracy using machine learning. MIR/FTIR (resolution ~10 μm) enables simultaneous compositional and morphologic assessment; Mao et al. achieved 95.7% training accuracy and 94.3% cross-validation for OA classification using PCA-FDA. Raman spectroscopy (resolution 1 μm–250 nm) detects biomarkers like PG, amide I/III, and phosphate species; Gupta et al. found increased mineralization in early OA but decreased in advanced OA. - **Spatial multiomics:** Spatial transcriptomics (e.g., 10X Visium, resolution ~200 nm) has identified distinct zonal chondrocyte populations in mouse embryo cartilage. Spatial proteomics via MALDI-MSI revealed fibronectin upregulation in OA synovium and cartilage, and altered N-glycan distribution in deep OA cartilage. Targeted methods like IMC and MIBI can multiplex up to 50 proteins at 1 μm resolution. - **Lipidomics/metabolomics:** TOF-SIMS and MALDI-MSI showed altered lipid profiles (e.g., increased phosphatidylcholines, decreased lysophosphatidylcholine) in OA synovium. MALDI-MSI identified ~40 N-glycan structures in cartilage and subchondral bone, with high-mannose N-glycans differentially distributed. - **Functional enzyme imaging:** fMSI demonstrated elevated phospholipase A2 (PLA2) activity in OA cartilage, with PLA2G2A identified as the predominant enzyme. - **Elemental imaging:** SR-μXRF (resolution ~1 μm) revealed distinct elemental patterns between normal and OA tissue. TOF-SIMS detected elevated calcium and phosphate colocalization in early OA cartilage, associated with MMP-3 and MMP-13 synthesis. - **Mechanical phenotyping:** AFM detected changes in cartilage stiffness at grades 1–2 OA (Outerbridge scale), while microindentation showed decreased cartilage rigidity in OA patients.

**Clinical Implications:** Deep spatial phenotyping techniques can transform OA understanding and management by: (1) elucidating disease mechanisms through spatial molecular signatures; (2) improving risk assessment and patient stratification based on molecular profiles (e.g., lipid/protein differences in OA with type 2 diabetes); (3) enhancing prognosis and diagnosis via integration with arthroscopy (NIR, FTIR, Raman) and correlation with blood/urine biomarkers; (4) enabling precision treatment by tracking drug penetration (e.g., triamcinolone acetonide in cartilage via MSI) and guiding targeted therapy. Challenges include sample preparation for bone/cartilage, high cost, need for subcellular resolution, and lack of dedicated OA databases. Future directions include 3D imaging, spatiotemporal analysis, and nondestructive sequential multiomics to support development of disease-modifying osteoarthritis drugs (DMOADs).