Bradykinin (BK) is an important endogenous mediator of microvascular flow modulation. Since the structure of the microcirculation is very different in tumor tissues than in normal tissues, bradykinin may elicit different responses in tumors. This study was designed to test the hypothesis that local administration of bradykinin increases blood flow preferentially in normal tissue relative to adjacent tumor tissue, resulting in a 'vascular steal' phenomenon. Microvessel diameters (D), velocities (V(c)), length densities, shear rates, and intermittent flow frequencies were measured every 10 min before, during, and after 40 min exposure to BK in rats with dorsal flap window chambers 9 days after tumor implantation. Separate studies were made of normal vessels outside the tumor margin, the hypervascular tumor periphery, and the tumor center. Bradykinin was administered with a suffusion medium flowing over the tissue at 1-2 ml/min with a BK concentration of 1.6 x 107 M. Administration of BK created five distinct changes in normal and tumor vessel function that varied over time, but coincidentally reached a maximum effect after 20 min exposure to BK. In normal vessels, increased V(c) and D led to increased flow, which reached a peak 20 min after onset of suffusion with BK. In contrast, in centrally located tumor vessels, decreased D and V(c) were observed in most vessels during the initial 10-20 min of suffusion. In addition, there was a significant increase in intermittent flow frequency in tumor central vessels, which peaked after 20 min of suffusion with BK. These five separate observations that coincided at 20 min of suffusion are consistent with a 'vascular steal' phenomenon. The increase in normal microvessel D and V(c) at 20 min suggests that BK causes vasodilation in arterioles. The coincident decrease in tumor microvessel D and V(c) suggests that tumor feeding vessels are less able to respond to BK by vasodilating. The concomitant increase in intermittent flow frequency in tumor vessels suggests that a reduction in pressure drop occurred after 20 min exposure to BK, which is also consistent with 'vascular steal.' Since BK is also known to increase vascular permeability, it is possible that increases in interstitial fluid pressure brought on by exposure to BK contributed to the observed reduction in tumor blood flow. In normal vessels, reduced D and V(c), relative to peak values, were noted after 40 min suffusion with BK. Adherence of leukocytes to the vessel walls was prominent and microthrombi were also observed during this period. No evidence of such adhesion was seen in tumor vessels, although microthrombi were observed. The mechanism for an increase in flow in some tumor vessels at the time that flow in normal vessels was declining could be related to loss of vascular steal, accentuated by adhesion of leukocytes to the walls of normal microvessels. Heterogeneity was observed in these effects between and within individual tumors. Possible reasons for the heterogeneity are discussed. The temporal variability in the effects of BK on blood flow in the tumor and normal tissue microvasculature suggests, for this model, that tumor oxygenation could be influenced by local release of BK, but that the magnitude would be highly dependent on time after the release of BK and would be variable from one tumor to the next.
