Osteoarthritis-associated fibrosis is related with both elevated pyridinoline cross-link formation and lysyl hydroxylase 2b expression

Abstract

Purpose: Fibrosis is a major contributor to joint stiffness in osteoarthritis (OA) and associated with joint pain. Lysyl hydroxylase 2b (LH2b) increases pyridinoline cross-links during collagen synthesis, resulting in harder to degrade collagen. These pyridinoline cross-links are elevated in different fibrotic diseases. In previous studies in murine knee joints, we found a relationship between LH2b and TGF-β-induced irreversible synovial fibrosis. In this study, we examined whether the amount of pyridinoline cross-links per triple helix collagen was elevated in the synovium of murine knee joints during collagenase-induced OA-related fibrosis. In addition, to learn more about the mechanisms by which LH2b is induced and how to prevent this induction, we investigated whether the induction of LH2b was Smad2P or Smad3P depended. Methods: We induced OA by intra articular injection of bacterial collagenase into the right knee joint of C57Bl/6 mice (collagenase model). Mice were sacrificed at day 7, 21, 28 and 42 after collagenase injection and the mRNA was isolated from the synovium for Q-PCR analysis. Paraffin sections of the murine joints were stained immunohistochemically for LH2 to determine the LH2 expression. The amount of pyridinoline cross-links per triple helix in synovium was determined biopsies with HPCL. All animal experiments were approved by the local animal ethics committee. Human fibroblasts (hSF) were isolated from synovial tissue of knee joints of OA patients undergoing arthroplasty. The hSF were stimulated with TGF-β with and without the Smad3P inhibitor SIS3 or the ALK5 kinase inhibitor SB-505124 (SB-5). RNA was isolated and the gene expression for LH2b and collagen type 1 (COL1A1) were analyzed with Q-PCR. Results: LH2b mRNA expression in synovium of murine knee joints was significantly upregulated in the OA-affected joints compared to the healthy joints on al measured days. Histological sections of murine knee joints with collagenase-induced OA showed a mild increase in the thickness of the synovial membrane at day 7 whereas a large increase was seen on later days. Day 7 showed a strong increase in LH2 staining, at later time points there was still a clear increase but less intence than at day 7 (Figure 1). There was a significant increased in the amount of pyridinoline cross-links per triple helix after day 7 compared to control knee joints. TGF-β upregulated both LH2b and Col1A1 gene expression in hSF, however when SIS3 the Smad3P inhibitor was added Col1A1 was strongly (+/- 4 Ct cycles) down-regulated. In contrast to Col1A1, LH2b was still induced by TGF-β in the presence of SIS3. SB-5 blocked both TGF-β induced LH2b and Col1A1. Conclusions: We have published that TGF-β is crucial in irreversible synovial fibrosis in experimental OA. Now we show strong LH2b expression in synovium of murine knee joints with collagenase-induced OA and that pyridinoline cross-links were more than doubled per triple helix in the synovium of murine knee joints with collagenase-induced OA. Most likely, TGF-ß that is elevated during OA is the driving force of enhanced LH2b expression, which drives pyridinoline cross link formation. We showed previously that TGF-β induced LH2b relies exclusively on TGF-β ALK5 (Smad2/3) and not ALK1 (Smad1/5/8) signaling. Blocking both Smad2 and Smad3 signaling did prevent TGF-β induced LH2b, whereas LH2b was still induced by TGF-β when alone Smad3P was inhibited. This suggests that LH2b is mediated through ALK5/Smad2P. We propose that LH2b is responsible for the persistence of fibrosis during OA. Blocking LH2b, or the Smad2p route, in OA may therefore prevent the formation of persistent fibrosis.