To evaluate the accuracy of in vivo estimates of protein synthesis and breakdown, measurements of plasma and tissue leucine kinetics were made in rat tumor tissues at different conditions of growth by use of constant intravenous infusion of [14C]leucine. These measurements were made in Yoshida sarcoma tumors on days 10 and 13 after implantation, with and without tumor necrosis factor (TNF) infusion and on day 10 in Walker-256 carcinosarcoma. Expressed as micromoles of leucine per gram tissue, tumor protein breakdown increased (P less than 0.01) from 0.32 +/- 0.02 to 0.52 +/- 0.09 (SE) mumol/h, with progress of the Yoshida sarcoma tumor between days 10 and 13 after implantation. Similarly, TNF increased tumor proteolysis on day 10 (0.43 +/- 0.03 mumol.h-1.g-1, P less than 0.05 vs. day 10 control) but not on day 13 after implantation of the Yoshida tumor. Estimates of growth derived from the difference between protein synthesis and breakdown rates were not statistically different from those based on actual tumor volume changes in both tumor models. However, estimates of "whole body" protein metabolism (plasma leucine flux) were not affected either by tumor aging or by treatment with TNF. This study shows that in vivo estimates of tissue protein metabolism based on our [14C]leucine constant infusion model closely reflect the growth characteristic of that tissue. A cytotoxic perfusion-independent effect for intravenous TNF on growing tumor tissue is demonstrable as increased protein breakdown. Furthermore, the commonly used concept of whole body protein metabolism, derived solely from tracer dilution in plasma, is an oversimplification.