Urokinase and tissue-type plasminogen activators (u--PA and t--PA) are serine proteases that convert plasminogen into plasmin, which degrades matrix proteins and activates metalloproteinases. The PAs are balanced by specific inhibitors (PAI--1 and PAI--2). Local production of t--PA and PAI--2 was recently demonstrated in human gingival tissues. The aim now was to investigate the production and localization of t--PA and PAI--2 in gingival tissues from dogs in three well-defined periodontal conditions; clinically healthy gingiva, chronic gingivitis and an initial stage of ligature-induced loss of attachment. At the start of the experiment the gingiva showed clear signs of inflammation. Clinically healthy gingiva were obtained after 21 days period of intense oral hygiene. Attachment loss was induced by placing rubber ligatures around the neck of some teeth. Biopsies were taken from areas representing the different conditions and prepared for in situ hybridization and immunohistochemistry. In clinically healthy gingiva both t--PA mRNA and antigen were expressed in a thin outer layer of the sulcular and junctional epithelia. No t--PA signals or staining were seen in connective tissue. Both mRNA signaling and immunostaining for t--PA were stronger in chronic gingivitis. In areas with loss of attachment, t--PA mRNA as well as antigen were found in the sulcular and junctional epithelia to a similar degree as in gingivitis. Occasionally the connective tissue was involved, especially in connection with vessels. PAI--2 mRNA was seen in a thin outer layer of the sulcular and junctional epithelia in clinically healthy gingiva, but no signals were seen in connective tissue. PAI--2 antigen was found primarily in the outer layer of the sulcular and junctional epithelia. Some cells in the connective tissue were stained. In gingivitis, PAI--2 signals were mainly found in the same locations, but more intense and extending towards the connective tissue. Immunostaining was seen in the outer half of the sulcular and junctional epithelia as well as in the upper part of the connective tissue, close to the sulcular epithelium. In sites with loss of attachment, PAI--2 mRNA was found throughout the sulcular and junctional epithelia, as was the antigen, which stained intensely. No PAI--2 mRNA was seen in connective tissue; the antigen was found scattered, especially near vessels. This study shows that the expression of both t--PA and PAI--2 increases with experimental gingival inflammation in the dog, and furthermore, the two techniques demonstrate a strong correlation between the topographical distribution of the site of protein synthesis and the tissue location of the antigens for both t--PA and PAI--2. The distribution correlates well with previous findings in humans.