The quantitative and functional relation between insulin‐like growth factor‐I (IGF) and IGF‐binding proteins during human osteoarthritis

A previous hypothesis stated that during osteoarthritis (OA) increased insulin‐like growth factor (IGF) binding proteins (IGFBPs) sequester IGFs and limit their access to the cell. The objective of this article was to test this by: (1) quantifying IGF and IGFBP‐3 as well as their ratios in human OA cartilages, and (2) measuring the metabolic responses of diseased cartilage to IGF‐I and its IGFBP‐insensitive analogs. Knee or hip OA cartilages were staged for OA by histology. Cartilage slices were either extracted for assays of IGF proteins, or maintained intact as organ cultures. Proteoglycan (PG) metabolism ± IGFs was measured by use of the ³⁵S‐sulfate precursor. IGFBP‐3 (ng/mg protein) was weakly correlated with OA score by regression analysis (R² = 0.122; p = 0.040; n = 35). IGF‐I (ng/mg protein) was constant across all OA groups (ANOVA; p = .428, n = 18) and the IGF‐I/IGFBP‐3 ratios were > 1 in most samples. All OA cartilages responded to hrIGF‐I by increasing PG synthesis [average 2.29‐fold ± 0.55 (±SD) at saturation, n = 12] irrespective of OA score. The des (1–3) IGF‐I analog (which lacks the three N‐terminal amino acids) had similar maximal effects (average 2.23‐fold stimulation ± 0.71, n = 10), but it was more effective in two out of three samples at suboptimal doses. The effect of hrIGF‐I, des (1–3) IGF‐I, or the B‐chain analog on degradation was minimal. In summary, catabolism was insensitive to IGF‐I, and this was probably not due to IGFBPs. By contrast, IGF‐I exerted a robust stimulation of anabolism at sufficiently high doses, even though IGFBPs could tone down the ligand effect at low doses.