Local bone destruction and osteoporosis in rheumatoid arthritis (RA) is mediated by the interaction of the receptor activator of nuclear factor kappa B (RANK) with its ligand (RANKL) and osteoprotegerin (OPG). Effector cells – osteoclasts – are generated upon the contact between RANK-positive (RANK⁺) pre-osteoclasts with RANKL expressed on the surface of osteoblasts or inflammatory cells [1]; a role of the soluble form of RANKL (sRANKL) in synovial fluid (SF) or serum in the osteoclast maturation has also been discussed. We have previously reported higher local production of OPG and sRANKL in RA active joints in comparison with their serum concentration, lower local and systemic concentrations of OPG in RA patients in comparison with those suffering from osteoarthritis (OA) and no difference in SF-sRANKL levels between RA and OA groups [2].

Surface expression of RANKL in T lymphocytes and monocytes was evaluated by flow cytometry using goat biotinylated polyclonal anti-RANKL and anti-RANK antisera visualized by streptavidin-phycoerythrin and FITC-conjugated anti-CD3, anti-CD4 and anti-CD14 monoclonal antibodies. Six paired RA samples of SF and peripheral blood were analysed. The results were compared with data from peripheral blood of 18 RA donors, 14 OA donors and 14 healthy donors, respectively.

In SF of RA patients, higher numbers (17%) of CD3⁺RANKL⁺ mononuclear cells (synovial fluid mononuclear cells) were found than in peripheral blood mononuclear cell (PBMC) preparations (5%, P = 0.07). In contrast, relatively more RANKL⁺CD14⁺ monocytes (79%) exist in peripheral blood than in SF (20%) in the same individuals. When PBMCs in different groups were analysed, significantly higher numbers of CD3⁺ (8%) and CD4⁺ (8.9%) PBMCs bearing RANKL on the surface were found in the RA group than in controls (4.3% and 4.7%, P = 0.03 and 0.02). In addition, RA patients had significantly higher proportions of CD14⁺RANKL⁺ PBMC (45.6%) than controls (30.2%, P = 0.013) or OA patients (29.7%, P = 0.046). On the other hand, RANK expression on CD14⁺ PBMC appeared diminished in RA patients (17.3%) when compared with controls (35.5%, P = 0.013).

Our findings suggest that the induction of bone erosions may depend rather on surface-bound RANKL than on its soluble form. This hypothesis is supported by our previous findings (see rationale) as well as by differences in RANKL expression on SF and peripheral blood CD3⁺ cells and the increased numbers of RANKL⁺ lymphocytes in the peripheral blood CD3⁺ and CD4⁺ compartments in RA patients. The lower ratio of CD14⁺RANK⁺ monocytes in peripheral blood of RA patients could reflect selective recruitment of RANK⁺ cells into the sites with elevated RANKL expression. In such a process, a soluble form of RANKL would probably play an important role [3].