Synovial intimal fibroblasts.

Cells of the normal synovial intima fall into two clearly separable types, macrophages and fibroblasts.' 2 In diseased tissue separation on traditional criteria such as ultrastructure may be less clear cut, but recent studies using techniques for identifying specific gene products related to cell lineage suggest that this distinction remains valid.3 ' The fibroblasts of synovial intima have been considered the likely source of synovial fluid hyaluronan for several decades, but direct evidence of specialisation of these cells has been difficult to obtain, because of the problems of separating intimal from other fibroblasts in vitro.5 6 Histochemical demonstration of hyaluronan in normal synovium suggests that the molecule is concentrated in the intima,7 but this could theoretically reflect diffusion into the intima from the synovial fluid. The first in situ evidence that intimal fibroblasts are specialised came from the work of Stevens and coworkers,8 showing that these cells, in contrast to most other fibroblasts, are associated with an unknown epitope recognised by the monoclonal antibody 67. Unfortunately, the molecule carrying this epitope has remained uncharacterised until now. More recently, Pitsillides et al,9"` using a quantitative cytochemical technique developed by Mehdizadeh et al,'2 were able to show that intimal fibroblasts do indeed differ from other fibroblasts in their potential for synthesis of hyaluronan. The activity of the enzyme uridine diphosphoglucose dehydrogenase (UDPGD) is four to nine times greater in synovial intimal fibroblasts than in any cells of the deeper layers ofthe tissue. Sheets ofintimal cells partially disaggregated from tissue and viewed as cytospin preparations show a clear distinction between fibroblastic cells, with high UDPGD activity, and macrophages with prominent cytoplasmic CD68." While UDPGD activity is not a direct measure of hyaluronan synthesising ability, there are reasons for believing that, in cells which show low levels of sulphate incorporation (intimal cells), it may be the most useful cytochemical marker of this specialised activity.'3 Knowledge of the specialisation of gene expression in intimal fibroblasts has subsequently expanded rapidly. These cells express vascular cell adhesion molecule-I (VCAM-1) even in normal tissue, at levels rarely achieved on endothelium (judged immunochemically) . They also show relatively prominent basal expression of the other adhesion molecules, CD44,15 and 1 integrins'6 (chiefly in combination with ox3, 5 and 6 chains). 17 As discussed by Revell in these Proceedings, they are probably largely responsible for the synthesis of a number of relatively specialised extracellular matrix components including types IV,`8 V and VI"9 20 collagen, laminin,'8 fibronectin,'8 chondroitin6-sulphate containing glycosaminoglycans2' and tenascin.22 These findings clearly establish that a functionally specialised type of fibroblast exists in synovial intima. The best term for describing these cells remains uncertain. The term synoviocyte may be useful, but unfortunately has been used in the past to include, not only intimal macrophages, but also unselected synovial cells studied in vitro. The ultrastructural distinction between type A and B cells' is less than ideal, partly because it is linked with the implication that the two cells are related in terms of ontogeny, and partly because ultrastructure is no longer considered a good gold standard for distinguishing macrophage and fibroblasts in the context of disease. The author currently favours the term synovial intimal fibroblast. Unfortunately, even this term is less than ideal because, in normal rabbit synovial intima, two fibroblast populations exist, one with high UDPGD activity and one expressing VCAM-1 (Kahn, in preparation). The more general term synovial fibroblast is perhaps best used for passageable cells, in tissue culture, which may be of both intimal and subintimal origin. Recent studies in our laboratory have focussed on the epitope recognised by monoclonal antibody (MAb) 67. The epitope is expressed almost exclusively on the intima in sections of both normal and diseased synovium. Stevens et al8 have demonstrated its specific association with intimal fibroblasts. However, studies of other tissues have shown the molecule to be expressed at a range of other highly specific sites: Bowman's capsule of the glomerulus and the juxtaglomerular apparatus (JGA) in the kidney, fetal synovial intima (being present even before joint cavity formation) fetal skin, cells within lymphoid follicular germinal centres, bone marrow stromal fibroblasts, elastin fibres in some tissues, placental chorionic villi, amniotic epithelium, and weakly in adult skin stratum granulosum. There is an interesting overlap in this distribution with that of VCAM-1, which is also present in adult (but not fetal) synovial intima, Bowman's capsule (but not the JGA), lymphoid germinal centres (but on a different cell population), and bone marrow stromal cells.23 It has proved possible to extract material from both synovium and amnion, giving a band on Western blots probed with MAb 67 corresponding to a molecular mass of about 55 kDa. This material is readily extracted Division of Rheumatology, University College London, United Kingdom J CW Edwards