The potential use of TNF as a therapeutic agent requires knowledge of the rate and mechanisms of its plasma and tissue clearance and its cellular localization. We have explored these issues using a male Sprague-Dawley rat model and a specific radioimmunoassay for human TNF that does not cross-react with rodent TNF. Unlabeled and 125I-labeled TNF were administered by intracardiac injection. A biexponential clearance was observed with t1/2 (beta-phase, metabolic) of 280 min for TCA-precipitable 125I-labeled TNF and t1/2 (beta-phase) of 30 min for unlabeled TNF. The Sephadex G-75 profile for 125I-labeled TNF obtained 2 min and 4 h after intracardiac injection did not differ significantly except for an increase in the amount of free 125I at 4 h compared to 2 min. Scatchard plot analysis suggests that 125I labeling alters significantly the binding of TNF to its receptors on FRTL-5 cells. 125I-labeled TNF uptake by 18 different tissues was corrected for the 125-labeled TNF in the vascular space by simultaneous injection with 131I- or 99mTc-labeled albumin. The 125I-labeled TNF tissue/plasma ratio was greater than 1.0 in adrenal, pituitary, lung, spleen, liver, and kidney. Most of these tissues have previously been shown to have TNF receptors and to experience the greatest endothelial damage in response to TNF administration.