A total of 72 RG-2 transplanted gliomas were studied in 58 rats at three time points (1, 30, 240 min) after intravenous injection of [125I]radioiodinated serum albumin ([125I]RISA). The animals were divided into two groups: a control group that received no treatment and a second group that was treated with five doses of 1.5 mg/kg of dexamethasone over 2.5 days. Local tissue concentrations of [125I]RISA were measured with quantitative autoradiography based on morphological features of the tumors and used to calculate the tissue distribution space. Two models were used to analyze the data. A two compartment model yielded estimates of local blood-to-tissue influx constants (K1), lower limit extracellular volumes (Ve), and plasma vascular volumes (Vp) in different tumor regions. Treatment with dexamethasone consistently reduced the RISA distribution space in the RG-2 tumors; the reduction in Ve was statistically significant in almost all tumor regions: whole tumor Ve (mean +/- SE) was reduced from 0.14 +/- 0.02 ml g-1 in control animals to 0.08 +/- 0.01 ml g-1 in dexamethasone treated animals. K1 and Vp were also decreased in all tumor regions after treatment with dexamethasone (whole tumor K1 decreased from 2.36 +/- 0.89 to 0.83 +/- 0.29 microliter g-1 min-1 and Vp decreased slightly from 0.016 +/- 0.013 to 0.010 +/- 0.005 ml g-1 after dexamethasone treatment), but these changes were not statistically significant. A comparison of the tumor influx constants in control animals and the aqueous diffusion constants of two different size molecules (RISA and aminoisobutyric acid) suggests that the "pores" across RG-2 capillaries are large and may not restrict the free diffusion of RISA (estimated minimum pore diameter greater than 36 nm) and that the total pore area is approximately 6.2 X 10(-5) cm2 g-1 in RG-2 tumor tissue. The second model, which allows for diffusion and solvent drag of RISA across tumor capillaries and through the tissue, was used to analyze the distribution profiles of RISA in peripheral tumor and adjacent brain. This analysis was consistent with a small bulk flow of plasma-derived edema fluid (capillary filtration rate approximately equal to 0.8 microliter g-1 min-1) and a larger component of free diffusion of RISA (K approximately equal to 2 microliter g-1 min-1) through pores in the tumor vessels of control animals. Dexamethasone treatment markedly reduced or eliminated the filtration of plasma-derived fluid across tumor capillaries and the movement of RISA through the extracellular space by solvent drag.(ABSTRACT TRUNCATED AT 400 WORDS)