Mass spectrometry-based proteomics identify novel serum osteoarthritis biomarkers

• Published in: Arthritis Research & Therapy
• Main Authors: Ginette Tardif, Frédéric Paré, Clarisse Gotti, Florence Roux-Dalvai, Arnaud Droit, Guangju Zhai, Guang Sun, Hassan Fahmi, Jean-Pierre Pelletier, Johanne Martel-Pelletier
• Format: Article in Journal/Newspaper
• Language: English
• Published: BMC 2022
• Subjects: Serum biomarkers; Proteomics; Osteoarthritis; Mass spectrometry; Diseases of the musculoskeletal system; RC925-935; Newfoundland
• Online Access: https://doi.org/10.1186/s13075-022-02801-1; https://doaj.org/article/cedea55d8a3f4c18b61c342dfbb76726

Abstract Background Osteoarthritis (OA) is a slowly developing and debilitating disease, and there are no validated specific biomarkers for its early detection. To improve therapeutic approaches, identification of specific molecules/biomarkers enabling early determination of this disease is needed. This study aimed at identifying, with the use of proteomics/mass spectrometry, novel OA-specific serum biomarkers. As obesity is a major risk factor for OA, we discriminated obesity-regulated proteins to target only OA-specific proteins as biomarkers. Methods Serum from the Osteoarthritis Initiative cohort was used and divided into 3 groups: controls (n=8), OA-obese (n=10) and OA-non-obese (n=10). Proteins were identified and quantified from the liquid chromatography–tandem mass spectrometry analyses using MaxQuant software. Statistical analysis used the Limma test followed by the Benjamini-Hochberg method. To compare the proteomic profiles, the multivariate unsupervised principal component analysis (PCA) followed by the pairwise comparison was used. To select the most predictive/discriminative features, the supervised linear classification model sparse partial least squares regression discriminant analysis (sPLS-DA) was employed. Validation of three differential proteins was performed with protein-specific assays using plasma from a cohort derived from the Newfoundland Osteoarthritis. Results In total, 509 proteins were identified, and 279 proteins were quantified. PCA-pairwise differential comparisons between the 3 groups revealed that 8 proteins were differentially regulated between the OA-obese and/or OA-non-obese with controls. Further experiments using the sPLS-DA revealed two components discriminating OA from controls (component 1, 9 proteins), and OA-obese from OA-non-obese (component 2, 23 proteins). Proteins from component 2 were considered related to obesity. In component 1, compared to controls, 7 proteins were significantly upregulated by both OA groups and 2 by the OA-obese. Among upregulated proteins ...