Kinetic analysis of cerebrovascular transport based on indicator diffusion technique.

The indicator diffusion method was used for studies of the blood-brain barrier in rats and [131I]iodoantipyrine (IAP) was used as a highly diffusable model test substance. Interlaminar (Taylor) diffusion and effects of red cell carriage were studied with 57Co-diethylene-triaminepentaacetic acid (DTPA), 125I-human serum albumin, and 46Sc-microspheres (15 microns diameter). Vascular shunting from the pterygopalatine artery (PPA) into the torcula sinus was observed in some animals, and ligation of the PPA was required to obtain reliable data. Dilution curve of the reference compound was corrected to compensate for any difference in interlaminar diffusion and red cell plus protein carriage of the test substance. Apparent extraction ratio of IAP was calculated for each torcula sinus sample and found to increase during the initial phase of the dilution curve, reach a peak of approximately 0.85, and fall during the latter portion of the curve. These results suggest a heterogeneity of intravascular transit times in the cerebral circulation and a rapid efflux of IAP from brain into venous blood. Because of the topography of cerebral capillaries, we developed a modification of the distributed model for intravascular transit and capillary exchange proposed by Goresky et al. (J. Clin. Invest. 52: 991-1009, 1973) and Rose and Goresky (Circ. Res. 34:541-554, 1976); this modification included a well-mixed tissue compartment, as suggested by Johnson and Wilson (Am. J. Physiol. 210: 1299-1303, 1966) and is named the tissue homogeneity model. The experimental data was analyzed by both the tissue homogeneity and Goresky models. The estimated mean extraction E (0.95 and 0.94) and the estimated permeability-surface area product of influx (PS)1 (3.1 and 2.8 ml.min-1.g-1) for IAP in a whole blood injectate were similar using the two different models. The efflux rate constant (k2) for IAP was consistently smaller when the tissue homogeneity model was used (0.13 +/- 0.02 s-1) vs. that obtained with the Goresky model (0.18 +/- 0.02 s-1). Model simulations also indicated that the efflux parameter k2 was most sensitive to the choice of kinetic models, but we could not discriminate between the two model analyses on the basis of the "quality of fit." Nevertheless, from anatomical considerations, we suggest that the tissue homogeneity model may be more appropriate fro brain.