Properties of tissue from around cemented joint implants with erosive and/or linear osteolysis.

Levels of bone remodeling agents were measured in conditioned media from cultures of periprosthetic pseudosynovial membranes and related to the radiographic features of the failed joint implants. Radiographs of both cemented hip (n = 28) and cemented knee (n = 11) implants were examined and the pattern of radiolucency was classified as erosive linear, or mixed. Similar levels of interleukin-1-beta (IL-1 beta), interleukin-6, tumor necrosis factor alpha (TNF-alpha), transforming growth factor beta-1, and prostaglandin E2 (PGE2) were found in pseudosynovial membrane conditioned media from all 3 groups of hips and the knee group (all linear). Significant correlations were evident only between PGE2 and TNF-alpha levels in the linear hip group and PGE2 and IL-1 beta levels in the knee group. A close correlation was found between IL-1 beta and TNF-alpha in both linear and erosive hips. It is suggested that coregulation of these bone remodeling agents differs with the radiographic appearance of the failed joint implants. As all the implants were cemented and the results contrast with those of others obtained with pseudosynovial membrane conditioned media from cementless implants, it is considered that cement critically influences the process of implant failure.

Osteolytic properties of the synovial-like tissue from aseptically failed joint prostheses.

Relationships were found between the bone-resorbing ability of conditioned media (CMs) from culture of peri-prosthetic tissues and their levels of bone-remodelling agents. Bone-resorbing activity was measured by 45Ca release from pre-labelled mouse calvaria and 23 of 40 CMs exhibited bone-resorbing activity. Cytokine and prostanoid levels in the CMs were measured by immunoassay, and the levels of interleukin (IL)-1 beta, IL-6, tumour necrosis factor alpha (TNF alpha) and prostaglandin E2 (PGE2) correlated with each other, except for the latter two. Significantly higher levels of IL-6 were present in those CMs with bone-resorbing activity than in those without, and a similar pattern was observed for PGE2 and IL-1 beta. However, some CMs with high levels of IL-1 beta, IL-6, TNF alpha and PGE2 failed to induce resorption, whereas a few CMs with low levels of these agents induced resorption. Moreover, neither dialysis of CMs nor addition of neutralizing antisera to IL-1 alpha and IL-1 beta to CMs, either alone or in combination, reduced the bone-resorbing activity of the CMs. It is considered that these agents may act synergistically to mediate osteolysis around failed joint implants, but that other unidentified bone-resorbing agent(s) must be involved.

Analysis of cell types and mediator production from tissues around loosening joint implants.

Quantitative immunophenotypic analysis of cell types present in peri-prosthetic tissue [pseudosynovial membrane (PSM)] from aseptically loose joint implants revealed considerable heterogeneity between tissues from different individuals. The monocyte/macrophage was the commonest leucocyte type; however, its proportion varied widely. T cells normally accounted for approximately 5% of cells, but in a few cases formed > 20% of cells. In all cases, there was a high ratio of CD4 to CD8 cells. PSM leucocytes were activated in most PSMs as judged by surface expression of CD23, CD25 and CD71. Analysis of the proportions of cell types in PSM, OA synovium and RA synovium revealed similarities between the different tissue types. The levels of IL-1, IL-6 and prostaglandin E produced by the PSM were correlated, but only IL-1 and IL-6 levels correlated with markers of the monocyte/macrophage lineage. This result suggests that prostaglandin E is produced in vivo by many PSM cell types.

Radiological scoring of osteoarthritis progression in STR/ORT mice.

The progressive changes observed in the knee and ankle joints of male STR/ORT mice, which spontaneously develop osteoarthritis (OA), were quantified by a radiological scoring system. The knee scores for males increased with age whereas those for females plateaued from 5 months. Comparison of scores for the knee and ankle joints showed that the male knee scores increased directly with age but were not significantly different from female scores until 7 months whereas the male ankle scores increased dramatically at 5-6 months and plateaued thereafter. In addition, correlations between patellar and calcaneal displacement showed that they rarely occurred together in the same limb, suggesting that knee and ankle OA are almost independent events. Although an association between high knee scores and patellar displacement was noted, even at 11 months of age not all mice with OA had displaced patellae, demonstrating that patellar displacement cannot be a primary event. Remarkably, summing the scores of all calcified structures in the ankle and knee joints revealed that the scores became significantly greater for males than for females by 3 months of age. It is considered that calcification is an earlier event than cartilage erosion and that OA in male STR mice may result from calcification placing abnormal stresses on joints.

Differential effects of immunisation with mycobacterial 65 kD heat shock protein on two models of autoimmunity.

The effects of preimmunisation with the 65 kD mycobacterial heat shock protein (hsp65) on 2 murine models of autoimmunity were compared. Experimental autoimmune haemolytic anaemia (AIHA) can be provoked in mice by repeated injection with rat red blood cells (RBC). In this model, preimmunisation with hsp65 10 days before induction of disease resulted in a partial, but significant, reduction in RBC-bound autoantibody levels measured by Coombs' test. However, preimmunisation with human IgG (hIgG) was associated with a similar suppressive effect. Administration of neither hsp65 nor hIgG affected the direct or indirect anti-rat agglutinin titres of mice subsequently injected with rat RBC. Injection of hsp65 or hIgG prior to induction of AIHA elicited the production of IgG antibodies against the respective immunogen, as judged by enzyme-linked immunosorbent assays. In contrast to the results in experimental AIHA, pristane-induced arthritis (PIA) was effectively prevented by preimmunisation with hsp65, but not with hIgG. It is considered that, whilst hsp65 injection may slightly reduce subsequent anti-RBC autoantibody production in AIHA by antigenic competition, such a mechanism cannot account for the substantial protection against PIA afforded by hsp65 preimmunisation. We suggest that the high, sustained production of anti-hsp65 antibodies observed in mice given hsp65 and pristane may play a role in specifically suppressing arthritogenic immune responses in PIA.

Rheumatoid factor has increased reactivity with IgG from synovial fluids of patients with rheumatoid arthritis and osteoarthritis.

Synovial fluid IgG may be altered in rheumatoid arthritis (RA) and promote the formation of immune complexes with rheumatoid factor. To investigate this possibility, monomeric IgG was prepared from synovial fluids from a range of arthritides for use as the antigen in a rheumatoid factor microplate radioimmunoassay. In comparisons with normal serum IgG antigen, increased rheumatoid factor binding was shown to IgG antigens prepared from synovial fluids from patients with RA and osteoarthritis (OA). Increased binding was also shown to RA sera IgG, but not to OA sera IgG. This increased binding was not due to increased IgG antigen binding to the plate or to IgG rheumatoid factor in the antigen preparations. It was considered that cause was a structural alteration of the IgG as a result of inflammation within the rheumatoid and OA joint.

Identification of an erythrocyte autoantigen using monoclonal autoantibodies induced by immunization of mice with rat erythrocyte.

An erythrocyte autoantigen has been identified by means of monoclonal autoantibodies raised by immunizing mice with rat red blood cells (RBC). The autoantibodies reacted with intact rat and mouse RBC as judged by a cellular radioimmunoassay, and with a 52K band on western blots of rat and mouse RBC. They did not react with intact sheep RBC or blotted sheep erythrocyte membranes. Although anti-rat erythrocyte antibodies did react with bands in the molecular weight region of 50-55 K (on blots of rat erythrocyte membranes), these bands were susceptible to neuraminidase digestion, thus distinguishing them from the 52 K band recognized by monoclonal autoantibodies. The implications of the above results for the known autoantibody specificity of suppression is discussed, and it is suggested that they favour the existence of autoantigen-specific suppressor cells.