To assess the temporal relationship between periodontal tissue destruction and the activity of collagenase, exudate from inflamed periodontal tissues was collected and latent and active collagenase activities were measured by a functional assay in a longitudinal cohort study. Comparisons were made between human subjects with either: 1) inflammation with a previous history of progressive loss of connective tissue and bone support (n = 14); 2) inflammation and previous history of bone loss but now clinically stable (n = 27); or 3) inflammation and no loss of bone support (n = 17). Experiments using specific enzyme inhibitors, blocking antibodies and SDS-PAGE fluorograph to identify the pattern of collagen substrate degradation demonstrated that the collagenase activity was derived from neutrophils and not from bacteria or other host cells. Active collagenase activity pooled from 6 sites per subject was respectively 5 and 6-fold higher in the group with progressive loss of connective tissue compared to the groups with either inflamed tissues alone or with inflammation and previous bone loss. In contrast, latent collagenase was increased up to 2 fold higher in the group with inflammation but no bone loss compared to the group with progressive lesions. Moreover, the ratio of active to total collagenase activity was 50% higher in the group with progressive lesions. Although in all subjects successive measurements of site-specific active collagenase 1 month apart demonstrated wide variation (r < 0.50), only in sites with progressive periodontal destruction was there significant increase of active collagenase with time (1.28 x 10(-4) collagenase units per day). There were also sharp elevations in active enzyme level at the time of detection of loss of connective tissue attachment in specific sites of 8 subjects. At the time of detection of connective tissue attachment loss, there was an overall 40% increase of pooled active collagenase activity in all subjects with progressive loss of connective tissue compared to pre-breakdown sampling times. These data provide strong in vivo evidence for a direct role of active neutrophil collagenase in the pathological destruction of periodontal connective tissue.