Long chain polyunsaturated fatty acids alter membrane-bound RANK-L expression and osteoprotegerin secretion by MC3T3-E1 osteoblast-like cells

Abstract

Inflammation triggers an increase in osteoclast (bone resorbing cell) number and activity. Osteoclastogenesis is largely controlled by a triad of proteins consisting of a receptor (RANK), a ligand (RANK-L) and a decoy receptor (osteoprotegerin, OPG). Whilst RANK is expressed by osteoclasts, RANK-L and OPG are expressed by osteoblasts. The long chain polyunsaturated fatty acid (LCPUFA) arachidonic acid (AA, 20:4n-6) and its metabolite prostaglandin E2 (PGE2), are pro-inflammatory and PGE2 is a potent stimulator of RANKL expression. Various LCPUFAs such as eicosapentaenoic acid (EPA, 20:5n-3), docosahexaenoic acid (DHA, 22:6n-3) and gamma-linolenic acid (GLA, 18:3n-6) have anti-inflammatory activity. We aimed to determine if AA itself can stimulate RANKL expression and whether EPA, DHA and GLA inhibit RANKL expression in osteoblasts. MC3T3-E1/4 osteoblast-like cells were cultured under standard conditions with each of the LCPUFAs (5 μg/ml) for 48 h. Membrane-bound RANKL expression was measured by flow cytometry and OPG secretion measured by ELISA. In a second experiment, RANKL expression in MC3T3-E1/4 cells was stimulated by PGE2 treatment and the effect of EPA, DHA and GLA on membrane-bound RANKL expression and OPG secretion determined. The percentage of RANKL-positive cells was higher (p < 0.05) than controls following treatment with AA or GLA but not after co-treatment with the cyclooxygenase inhibitor, indomethacin. DHA and EPA had no effect on membrane-bound RANKL expression under standard cell culture conditions. Secretion of OPG was lower (p < 0.05) in AA-treated cells but not significantly different from controls in GLA, EPA or DHA treated cells. Treatment with prostaglandin E2 (PGE2) resulted in an increase (p < 0.05) in the percentage of RANK-L positive cells and a decrease (p < 0.05) in mean OPG secretion. The percentage of RANKL positive cells was significantly lower following co-treatment with PGE2 and either DHA or EPA compared to treatment with PGE2 alone. Mean OPG secretion remained lower than controls in cells treated with PGE2 regardless of co-treatment with EPA or DHA. Results from this study suggest COX products of GLA and AA induce membrane-bound RANKL expression in MC3T3-E1/4 cells. EPA and DHA have no effect on membrane-bound RANKL expression in cells cultured under standard conditions however both EPA and DHA inhibit the PGE2-induced increase in RANKL expression in MC3T3-E1/4 cells.

Introduction

Pro-inflammatory signalling has a fundamental role in the initiation of bone remodelling, triggering both osteoclastogenesis and osteoblastogenesis. Osteoclasts are large, multi-nucleated cells of the monocyte/macrophage lineage which resorb fatigued bone. Osteoblasts are mononuclear cells of mesenchymal origin which synthesise new bone tissue and replace that which osteoclasts remove. Under normal conditions the processes of bone resorption and formation are coupled and there is no net change in overall bone mass. However conditions leading to elevated levels of inflammatory mediators result in an imbalance between osteoclast and osteoblast formation and activity and consequently, an overall loss of bone mass [1].

Osteoclastogenesis is largely controlled by a triad of proteins consisting of two receptors, receptor activator of nuclear factor κ B (RANK) and osteoprotegerin (OPG), and a ligand, receptor activator of nuclear factor κ B ligand (RANK-L) [2], [3]. RANK is a membrane-bound receptor present on osteoclast precursors [4]. RANK-L and OPG are expressed by osteoblasts [5], [6] as well as various other cell types including activated T-cells [7] and fibroblasts [8]. OPG is a soluble protein whereas RANK-L is largely expressed as a membrane-bound protein although small amounts of intact RANK-L are also secreted by osteoblasts [9]. Binding of RANK to RANK-L stimulates osteoclastogenesis and promotes mature osteoclast survival [10]. However binding of RANK-L to OPG leads to a rapid decline in osteoclast number due to both prevention of osteoclastogenesis and increased apoptosis of mature osteoclasts [10], [11], and a decrease in bone resorbing activity of existing osteoclasts [12]. Inflammatory mediators are involved in the regulation of RANK-L and OPG expression [13]. The level of RANK-L expression relative to OPG is a major controlling factor governing osteoclastogenesis [14].

Prostaglandin E2 (PGE2), a pro-inflammatory lipid mediator derived from the n-6 long chain polyunsaturated fatty acid (LCPUFA) arachidonic acid (AA, 20:4n-6) by the activity of cyclooxygenase (COX), promotes osteoclastogenesis by stimulating RANK-L expression and inhibiting OPG secretion by osteoblasts [15]. Similarly, treatment of osteoblasts with AA stimulates secretion of soluble RANK-L and inhibits secretion of OPG, and both effects are largely blocked by COX inhibition [16].

The n-3 family of LCPUFAs, particularly eicosapentaenoic acid (EPA, 20:5n-3) and docosahexaenoic acid (DHA, 22:6n-3) and another member of the n-6 LCPUFA family, gamma-linolenic acid (GLA, 18:3n-6), are generally regarded as anti-inflammatory lipids [17], [18], [19]. Dietary consumption of oils rich in EPA, DHA and/or GLA has been linked with anti-resorptive effects [20] or bone-protective effects [21], [22] in animal and human studies. The potential effects of these three LCPUFAs on membrane-bound RANK-L expression in osteoblasts are unknown.

The MC3T3-E1/4 cell line is a pre-osteoblast cell line derived from murine (Mus muscularis) calvarial cells. Under appropriate growth conditions, MC3T3-E1/4 cells differentiate into mature osteoblasts and are capable of synthesising mineralised extracellular matrix [23]. MC3T3-E1/4 cells have also been shown to increase TRAP+ cell number, (indicative of osteoclast cell number) when co-cultured with bone marrow cells indicating they are capable of supporting osteoclastogenesis [24], [25]. They have also been shown to induce osteoclast activity in vitro [26].

We hypothesised that treatment of MC3T3-E1/4 cells with the anti-inflammatory LCPUFAs, GLA, EPA and DHA, would reduce membrane-bound RANKL expression whereas treatment with the pro-inflammatory LCPUFA would increase membrane-bound RANKL expression. To this end, the effect of AA, GLA, EPA and DHA on membrane-bound RANKL expression in cells under standard culture conditions was investigated. Secondly, the effect of GLA, EPA and DHA on membrane-bound RANKL expression in cells in which RANKL expression had been stimulated by PGE2 was assessed.