The importance of a collagen-binding adhesin in the pathogenesis of septic arthritis has been examined by comparing the virulence of two sets of Staphylococcus aureus mutants in an animal model. Collagen adhesin-negative mutant PH100 was constructed by replacing the chromosomal collagen adhesin gene (cna) in a clinical strain, Phillips, with an inactivated copy of the gene. Collagen adhesin-positive mutant S. aureus CYL574 was generated by introducing the cna gene into CYL316, a strain that normally lacks the cna gene. Biochemical, immunological, and functional analyses of the generated mutants and their respective parent strains showed that binding of 1251-labeled collagen, expression of an immunoreactive collagen adhesin, and bacterial adherence to cartilage were directly correlated with the presence of a functional cna gene. Greater than 70% of the mice injected with the Cna+ strains developed clinical signs of arthritis, whereas less than 27% of the animals injected with Cna-strains showed symptoms of disease. Furthermore, mice injected with the Cna+ strain Phillips had remarkably elevated levels of immunoglobulin Gl and interleukin-6 compared with mice injected with the Cnamutant PH100. Taken together, these results demonstrate that collagen adhesin plays an important role in the pathogenesis of septic arthritis induced by S. aureus.

Fifty-four Staphylococcus epidermidis strains responsible for infections of joint prostheses and 23 strains isolated from skin flora were studied for markers of virulence, to discriminate invasive strains from normal flora. They were screened for binding to polystyrene and matrix proteins and for the presence of staphylococcal genes involved in adhesion. The ica operon involved in biofilm formation was the only marker discriminating between these 2 categories of strains.

We have studied the role of the accessory gene regulator (agr) of Staphylococcus aureus as a virulence determinant in the pathogenesis of septic arthritis. At least 15 genes coding for potential virulence factors in Staphylococcus aureus are regulated by a putative multicomponent signal transduction system encoded by the agr/hld locus. agr and hid mutants show a decreased synthesis of extracellular toxins and enzymes, such as alpha-, beta-, and delta-hemolysin, leucocidin, lipase, hyaluronate lyase, and proteases, and at the same time an increased synthesis of coagulase and protein A as compared with the wild-type counterpart. We have used a recently described murine model of S. aureus-induced arthritis to study the virulence of S. aureus 8325-4 and two agrihld mutants derived from it. Sixty percent of the mice in,jected with the wild-type strain developed arthritis, whereas agrA and hid mutants displayed joint involvement in only 10 and 30%o, respectively. In addition, 40% of the mice inoculated with the wild-type strain displayed an erosive arthropathy; such changes were not detectable at all in mice inoculated with the agrA mutant. Serum levels of interleukin-6, a potent B-cell differentiation factor, were significantly higher (P < 0.001) in the mice inoculated with the wild-type strain than in those inoculated with the agrA mutant counterpart. Overall, our results suggest that the agr system of S. aureus is an important virulence determinant in the induction and progression of septic arthritis in mice.

In contrast, binding was not inhibited by a bacterial lysate from an osteopontin A gtll lysogen, nor by N-linked oligosaccharide isolated from bone sialoprotein or by proteoglycan from rat chondrosarcoma containing clustered 0-linked oligosaccharides of the same structure as those of bone sialoprotein. These results indicate that the major staphylococcal-binding site resides in the bone sialoprotein core protein and not in the carbohydrate side chains. No inhibition of bone sialoprotein binding could be detected for whole human serum or purified plasma proteins such as fibronectin, fibrinogen and IgG. Likewise, staphylococcal protein A or rat collagen type I did not inhibit the binding of bone sialoprotein. The latter results indicate that the binding site for bone sialoprotein on staphylococcal cells was not any of the hitherto described staphylococcal cell-surface proteins. Binding data indicated an average of 1000 bone-sialoprotein-binding sites/bacterial cell.