The inflammatory phase in wound healing is considered to be a preparatory process for the formation of new tissue. A monocyte-derived cytokine, tumor necrosis factor-alpha (TNF-alpha), is a highly conserved molecule known to play a major role in the pathogenesis of gram-negative shock. Besides this, previous experimental studies show that TNF-alpha may have either a beneficial or detrimental role in wound healing. The purpose of the present study was to examine the effects of TNF-alpha on developing granulation tissue in rats as well as on rat and human granulation tissue cells in culture. Subcutaneously implanted cylindrical hollow sponges were used for studying the effects of locally applied TNF-alpha on granulation tissue in rats. These implants were treated either on the day of implantation or for the first 4 or 7 days after implantation with a solution containing various amounts of TNF-alpha while the control implants were treated correspondingly with the carrier solution only. The analyses of the granulation tissue were carried out 4, 7, 14 and 21 days after implantation. In the histological specimen these sponges were cut into small pieces and stained with Weigert van Gieson to visualize collagen. The amount of granulation tissue grown into the sponge was calculated from the cross section of every sponge. For the cell culture studies fibroblasts were released from human and rat granulation tissue which was cut into small pieces and digested by collagenase and DNase in Hank's balanced salt solution. The cells were exposed to 1, 10, or 100 ng/ml of TNF-alpha and the rate of collagen synthesis was measured as synthesis of protein-bound 3H-hydroxyproline. The number of cells in the culture dishes was counted with Bürger's hemocytometer after detaching the cells with trypsin treatment. As interleukin-1 (IL-1) and TNF-alpha overlap in many of their functions, the effects of lipopolysaccharide (LPS), human interleukin 1 beta (IL-1) and prostaglandin E2 (PGE2) on experimental granulation tissue in rats as well as on rat granulation tissue cells in culture were studied with the same method. After a single application of TNF-alpha into the sponge, no essential differences between the groups were detected. However, after daily applications of TNF-alpha for 4 days, an inhibitory effect on tissue repair was observed after 4 and 7 days. Collagen formation, indicated by the hydroxyproline content of the sponge, was significantly lower in the group treated with TNF-alpha than in the controls. This effect was not observed after 14 or 21 days. These findings were confirmed in the histological samples. In the cultures of rat granulation tissue fibroblasts TNF-alpha decreased 3H-hydroxyproline production to about 75% of that in the controls and it had also a decreasing effect on pro alpha 1(I) and pro alpha 1(III) collagen mRNA levels maximally by 67% and 77% of the control level, respectively. In the cultures of human granulation tissue fibroblasts a similar inhibiting effect on the production of collagen was seen. TNF-alpha decreased the production of 3H-hydroxyproline to 56% of the control value with a dose of 100 ng/ml. Similarly, IL-1 beta decreased hydroxyproline content of granulation tissue seven days postoperatively and PGE2 decreased nonsignificantly the amounts of hydroxyproline but the steady-state levels of pro alpha 1(I) and pro alpha 1(III) collagen chain mRNAs were slightly elevated. In the IL-1 beta-treated fibroblast cultures collagen production decreased by 15% compared with that of the controls. PGE2 decreased collagen production by 34% of that in the controls. This effect could be abolished with indomethacin. Indomethacin alone stimulated collagen production by 40%. In vivo IL-1 decreases the formation of normal granulation tissue. This effect may be partly due to IL-1 stimulated secretion of PGE2.