Cerebral blood flow exhibits autoregulation over

Transcription

1 102 Pressure-Induced Myogenic Activation of Cat Cerebral Arteries Is Dependent on Intact Endothelium David R. Harder These studies were designed to determine the role of cerebral vascular endothelium in the "myogenic" depolarization and contraction observed in isolated cat middle cerebral arteries exposed to high transmural pressures. With intact endothelial cells we observed, on elevation of transmural pressure in cannulated isolated arteries, significant membrane depolarization, action potential generation, and reduction in internal diameter. After perfusion of the same vessels with collagenase and elastase for short periods of time to disrupt the endothelial layer, all previous responses to elevation of transmural pressure were no longer seen. Even though enzyme perfusion had no effect on membrane potential at "control" levels of transmural pressure, it abolished the pressure-dependent depolarization, action potential generation, and constriction. Furthermore, the contractile response to agonist stimulation was maintained after endothelial disruption via enzymes, showing that this method of endothelial disruption did not appreciably damage muscle cells. The data document a dependence of an intact endothelium in mediating the activation of isolated cat cerebral arteries in response to a changing transmural pressure. Thus, it is possible that the endothelial cell may serve as a transducer in the autoregulatory response to pressure. (Circulation Research 1987;60: ) Cerebral blood flow exhibits autoregulation over a wide range of arterial blood pressures. We have recently shown that when isolated cerebral arteries are cannulated and placed in an appropriate muscle myograph a pressure-dependent reduction in internal diameter occurs, which is mediated by muscle cell membrane depolarization and action potential generation. 1>2 This action is not mediated by adventitial nerves in that it occurs in the presence of neural blockade. 1 It was the purpose of this study to examine the role of the endothelium in pressure-induced activation of isolated cat cerebral arteries. With an intact, undisturbed endothelium, isolated middle cerebral arteries exhibited membrane depolarization, action potential generation, and reduction in diameter as observed previously. However, when the endothelium is disrupted via perfusion of collagenase and elastase this same pressure-dependent arterial muscle cell activation is no longer observed. This lack of pressure-mediated tone is not due to damage of the muscle cells by the enzyme since the response to chemical stimulation is maintained. Thus, these data suggest that the endothelial cell may serve as the transducer mediating changes in transmural pressure to activation of cerebral arterial muscle. From the Departments of Neurology and Physiology, Medical College of Wisconsin, Milwaukee, Wis., and Veterans Administration Medical Center, Milwaukee, Wis. Supported by NIH grants and 31871, and the Veterans Administration. Dr. Harder is an Established Investigator of the American Heart Association and a Research Career Scientist of the Veterans Administration. Address for reprints: David R. Harder, PhD, VA Medical Center, Research/151, 5000 W. National Avenue, Milwaukee, WI Received May 21, 1986; accepted September 22, Materials and Methods Adult mongrel cats ( kg of either sex) were anesthetized with ketamine hydrochloride and sodium pentpbarbital (30 mg/kg), decapitated, and brains were removed. The left middle cerebral artery was dissected free of arachnoid and placed in cold (4 C) physiological salt solution (PSS) consisting of (in mm) Na + 141, Cl- 125, Ca , K + 4.7, Mg , H 2 PO 4 " 1.7, HCO 3 ~ 22.5, glucose 11, and N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid (HEPES) 5. HEPES was used as a partial buffer to control ph more accurately. One end of a 6-8 mm segment of middle cerebral artery was threaded onto a 50-/u.m diameter plastic cannula and tied in place with a 22-fx.m suture. The opposite end was likewise cannulated. All side branches were tied off with the above silk suture material. The cannulated artery was placed inside a muscle myograph in which one cannula was fixed in plastic jaws while the opposite end was connected to a micrometer so the vessel could be maintained at its in vivo length. The inflow cannula was connected to a pressure reservoir filled with PSS. Aligned between the pressure reservoir and the arterial preparation was a pressure transducer to monitor applied transmural pressure. The system could either be flow-through by leaving the outflow cannula patent or closed by clamping off the outflow cannula. The entire preparation was suffused with PSS maintained at 37 C via a water jacket. The PSS was aerated with 95% air and 5% CO 2, yielding a Pco 2 of torr and ph of Solution gases and ph were monitored via periodic sampling measured with a Radiometer gas anaiyzer. The PSS in the reservoir that perfused the vessel was aerated and maintained identical to the suffusion solutions. Pressure measurements were made when the

2 Harder Transducer for Myogenic Tone in Cerebral Arteries system was closed so that it was not necessary to account for the flow resistance dk he vessel. Internal diameter was monitored witji pr video system composed of a camera (RCA), TV monitor (Sony), and VCR (Panasonic), and measured throughout the experiment with a Colorado Video, Inc., measuring system (Instruments for Physiology, Model 907, San Diego, Calif.). Arterial dimensions were obtained 5 minutes after a step increase in luminal pressure of 20 mm Hg beginning at 20 mm Hg and ending at 160 mm Hg. Magnification on the video screen was 150x. Changes in diameter could be measured with an accuracy of ± 2 (im, which was the relative thickness of the electronically produced lines generated by the Colorado Video, Inc., measuring system. Intracellular electrical activity was recorded with 103 glass microelectrodes using previously described techniques.1 Glass microelectrodes were filled with 3 M KC1, had tip resistances of Mft and tip potentials less than 3 mv. Criteria for successful impalements of arterial muscle cells have been described in detail elsewhere.1 Measurements of electrical events and internal diameter were made when the system was closed so that specific values of transmural pressure were accurately known. The endothelium was disrupted by perfusing a PSS containing 1 /ug/ml collagenase and 0.5 /tg/ml elastase through the arterial segment for 3 minutes. After this 3-minute period, cold (20 C) PSS was perfused through the vessel at 100 mm Hg to inactivate and wash out enzymes. Efficacy of endothelial disruption was assessed via electron microscopy. Figure 1 is an FIGURE 1. Electron micrograph of a cat middle cerebral artery section (5820 X magnification) before (panel A) and after (panel B) perfusion of collagenase and elastase showing disruption of the endothelial surface layer upon perfusion. Note the tight endothelial-endothelial cell junction (E).

3 104 Circulation Research Vol 60, No 1, January ENDOTHELIUM INTACT INTERNAL DIAMETER 400 (urn) B ENDOTHEUUM REMOVED FIGURE 2. Panel A: Change in internal diameter in response to changing transmural pressure in isolated cannulated cat middle cerebral arteries. Due to the variability of initial vessel dimensions at equilibrating (20 mm Hg) transmural pressure, the response in each of the 8 separate arteries studied is given. Note the tendency of vessels 2-7 to decrease in internal diameter beyond mm Hg transmural pressure. Vessels 1 and 8 maintained diameter from mm Hg before becoming smaller beyond 120 mm Hg. The endothelium in these cerebral arteries is intact. Panel B: A graph depicting the same vessels as in Panel A (vessel numbers are the same in both so all data are paired). Note that after perfusion of cerebral arteries with collagenase and elastase to disrupt the endothelium, all 8 vessels increased in diameter as a function of transmural pressure TRANSMURAL PRESSURE ( mm Hg) electron micrograph (5,280 x) from a vessel perfused with enzymes showing disruption of the endothelial layer (Figure IB), compared to a nonperfused vessel showing intact endothelium with tight endothelial-endothelial cell junctions (Figure 1A). Details for electron microscopy are given in a separate communication. 3 To test the viability of the muscle cells within the artery after endothelial disruption with enzymes the response to 30 mm KC1 and serotonin (3 X 10 ~ 7 M) before and after perfusion were compared. In separate experiments, we could never observe a dilatory response to acetylcholine in serotonin contracted arteries when endothelium was disrupted in such a fashion with enzyme perfusion (observed in 5 of 5 vessels). All data in this manuscript are paired, i.e., the same vessel is compared before and after endothelial disruption. Data were generated from cerebral arteries of 8 individual animals. Results Effect of Endothelial Disruption on Pressure-Mediated Maintenance and Reduction of Internal Diameter Internal diameter was measured as a function of transmural pressure in 8 individual arteries. At trans- mural pressures beyond 60 mm Hg the internal diameter was maintained or became smaller (Figure 2A). In actual numbers, 2 preparations (1 and 8) maintained a constant diameter in the face of an increasing pressure load between 60 and 120 mm Hg and exhibited a reduction in diameter only at pressures beyond 120 mm Hg. In the 6 other preparations depicted in Figure 2A, internal diameter gradually became less at each pressure step beyond mm Hg, which is consistent with autoregulatory ability, i.e., flow resistance would increase at those higher transmural pressures. Each point depicted in Figure 2 is the value of a single diameter measurement taken 5 minutes after a 20 mm Hg step increase in transmural pressure. Conversely, disruption of the endothelium caused all 8 arterial preparations to dilate passively as transmural pressure was elevated (Figure 2B). The numbers in Figures 2A and 2B correspond to the same vessel before and after endothelial disruption. The inability of the vessels to respond to increased transmural pressure with reduction in diameter did not appear to be a consequence of muscle cell damage during the enzyme perfusion since there was no significant difference in the degree of vasoconstriction to high K + or serotonin before or after perfusion (in response to 30 mm KC1,

4 Harder Transducer for Myogenic Tone in Cerebral Arteries ±4% [SEM] of control diameter before vs. 24 ± 6% after enzyme perfusion; in response to 3 x 10" 7 serotonin, 28 ± 5% vs. 32 ± 8%). The integrity of the arterial muscle cell after endothelium disruption is further verified by the normal level of membrane potential, -61 ± 2.4 mv (SEM) after enzyme perfusion (see below). This increase in radial dimensions in the face of an increasing pressure load is most likely completely passive since it occurs in the absence of any changes in intracellular membrane potential (E ra ) and is similar to that occurring when Ca 2+ influx is blocked by verapamil, e.g., a mean diameter increase of 34% between mm Hg in the absence of endothelium vs. 38% in a separate set of experiments in the presence of verapamil (D.R. Harder and J.A. Madden, unpublished observations). Inhibition of Pressure-Dependent Arterial Muscle Cell Depolarization on Endothelial Disruption E m was measured in 45 cell impalements from 8 arteries at various transmural pressures. As seen in Figure 3, E m decreases as a function of transmural pressure from a control value of -63 ±1.8 mv (SEM) at 20 mm Hg to as low as - 22 mv at pressures beyond 140 mm Hg. Disruption of the cerebral arterial endothelium with collagenase and elastase prevented this "pressure-dependent membrane depolarization" (Figure 3B). Note, again, in Figure 3B that enzyme perfusion in and of itself had no significant effect on E m (-61 ± 2.4 mv at 20 mm Hg) at low transmural pressures. Comparison of Change in Diameter as a Function of Change in E m The diameter and E m data look at the pressure response in individual preparations with and without endothelium. Because of the variation in size of individual animals, there was a variation in size of any given middle cerebral artery; the scale is large to accommodate such variation, and the differences in diameter appear small. In an effort to compare more directly the diameter response to increasing transmural pressure with concomitant changes in E m, the diameter data were normalized as percent change from control and plotted against the change in E m. Resultant comparisons are depicted in Figure 4, which shows that in these 8 cerebral arteries with normal intact endothelium there is a significant positive relation between an increasing transmural pressure and reduction in internal diameter. The maximum reduction in internal diameter occurs at 160 mm Hg and is approximately 20%. There is also a significant correlation between the observed reduction in diameter and level of membrane depolarization. The slope predicts a 1% reduction in internal radial dimension for a 1.9 mv reduction in E m. The meaning of such a relation will require further study regarding cause and effect but does suggest that a membrane electrical mechanism may be responsible for the myogenic response to pressure. Any significant correlation between pressure and di MEMBRANE POTENTIAL ^ 1 ENDOTHELIUM INTACT * J * * * H 1 ^ B 60' -SO MEMBRANE POTENTIAL (mv) ENDOTHELIUM REMOVED TRANSMURAL PRESSURE (mm Hg) FIGURE 3. Panel A: Graph depicting membrane potential (E m ) measured with glass microelectrodes in 45 cells from 8 different arteries measured as a function of transmural pressure with intact endothelium. Each point is the value of a single impalement at that pressure. There is a linear relation between reduction ofe m.and elevation of transmural pressure with a correlation coefficient of The E m at 20 mm Hg averaged 63 ±1.8 mv (SE). As can be seen, the E m fell as low as 22 mvat pressures beyond 140 mm Hg. Panel B: Membrane potential (E m ) as a function of transmural pressure after a 3-minute perfusion with collagenase and elastase at 100 mm Hg perfusion pressure. Each point is an impalement at that pressure. Note that enzyme perfusion did not significantly change E m at 20 mm Hg, demonstrating maintenance of cell integrity, and that after enzymatic disruption of endothelium, there is no longer a relation between E m and transmural pressure as observed in A. ameter, and E m and diameter, is abolished upon disruption of the endothelium (Figures 2B and 3B). Inhibition of Action Potential Generation in Response to Elevated Transmural Pressure on Endothelial Disruption In 6 of 8 vessels studied, action potentials could be recorded when transmural pressure exceeded 60-70

5 106 mm Hg. These action potentials were recorded only at the base of side-branching arteries and were not observed for more than mm on either side of a tied off branch. Seen in Figure 5A (top) is a record of action potentials recorded approximately 80 jinn from that point where the cell in panel A (bottom) was impaled; note how it appears to be propagated in that there is no prominent prepotential such as that depicted in Figure 5A (bottom), which was recorded as close to the point of bifurcation as possible. Even though this is an interesting finding, the point most relevant to this manuscript is that after disruption of the endothelium with enzymes, no action potentials could be recorded from the very same areas of the artery at the same transmural pressure (Figure 5A vs. Figure 5B). However, in 8 preparations action potentials in areas visibly void of branching arteries were not recorded. Note also, that the membrane is hyperpolarized compared to the control after endothelial disruption. It should be noted that in those cells that did not generate action potentials, elevation in transmural pressure induced depolarization only, and Figure 3 represents data from all cells. In spiking cells, the level of E m was taken as the most hyperpolarized state INTERNAL DIAMETER TRANSMURAL PRESSURE (mm Hg) % CHANGE IN 90 INTERNAL DIAMETER 80 B ^_ 40 ^ ^ 5 (mmhg) SO N. 100 «^^ R = 0.98 ^^S^ CHANGE IN MEMBRANE POTENTIAL (mv) 120 ^ \ _ FIGURE 4. Representation of normalized data from Figure 2 A. Data is normalized as percent of diameter from control (i.e., 100% is control) at each transmural pressure. Regression analysis depicts a positive relation between pressure and reduction of diameter (negative slope), with a correlation coefficient ofo. 91. In Panel B, the normalized diameter data is plotted as a function of change in E m at each transmural pressure, the slope of which indicates a 1.9 mv change in E m for each 1% change in internal diameter and a highly significant correlation coefficient of The solid unsymboled lines in A and B represent the calculated slope obtained from regression analysis Circulation Research Vol 60, No 1, January 1987 TRANSMURAL A,Intact Endothelium 1 i i i i i n PRESSURE =120 mmhg Tsec A,Intact Endothelium r V E mv m, 4sec Ssec '"Ise e 4lei sec -60 FIGURE 5. Actual chart record depicting spontaneous action potentials fpanel A, top and bottom) recorded at branch points in isolated arteries exposed to 120 mm Hg transmural pressure. Panel B (top and bottom) represents cell impalements in the exact area after perfusion ofcollagenase and elastase to disrupt the endothelium. Apart from the inhibition of action potential generation after endothelial disruption, note also that the E is more polarized. This figure depicts action potentials recorded at a point of bifurcation (A, bottom) between the main artery and a branch demonstrating prominent pacemaker activity. The action potentials in the top of panel A were recorded 80 yjn away from that point impaled in the bottom record of the same panel. The absence of prominent pacemaker prepotentials suggests that the action potentials may be propagated; no action potentials were recorded greater than 2 mm away from the point impaled in Panel A (bottom). All impalements were from the same vessel. Discussion Much has been written about the influence of substances released from the endothelial cell layer on the muscle cells within the same arteries. There is little direct evidence identifying the nature and/or composition of those endothelial derived substances; however, the action of these agents appears most often to be dilatory. 4 In some arterial beds and under certain conditions, constrictor substances that are released from vascular endothelium can also be identified. 5 " 7 A recent abstract also demonstrates a constrictor substance released from dog cerebral arteries upon stretch. 8 The data presented here are in precise agreement with those of Katusic et al. 8 An intact endothelium does indeed appear to be necessary in order to elicit the membrane depolarization, action potential generation, and pressure-dependent reduction of internal diameter in isolated cat cerebral arteries. The absence of pressure-mediated excitatory events does not appear to be due to damage of cerebral arterial muscle upon perfusion ofcollagenase and elastase in that 1) the response to chemical agonists is not altered by the treatment; and 2) a stable E m not different from control after

6 Harder Transducer for Myogenic Tone in Cerebral Arteries 107 enzyme perfusion is maintained, showing that the muscle cell membrane was not damaged. Furthermore, the dilatory action of transmural nerve stimulation is not affected by enzyme perfusion in cat middle cerebral arteries. 3 It appears that vascular endothelial cells, at least in cat and dog cerebral arteries, may serve a "transducer function" in that a mechanical force liberates a chemical mediator that can activate the adjacent arterial muscle cells. The increase in radial dimensions upon elevation of transmural pressure in the absence of endothelium is most likely a passive phenomenon, since there is no change in E m and increases in a linear fashion with pressure. A similar passive phenomenon is observed in cat cerebral arteries upon inhibition of Ca 2+ influx with verapamil. 9 The data in the present study suggest that an increase in transmural pressures creates a stress across the endothelial cell that activates the release of a chemical mediator. It is not possible at this point to determine the specificity of the type of mechanical stimulus, i.e., the sheer stress that would occur as transmural pressure increased, actual stretching of the cell surface, or a point source of pressure analogous to activation of a pacinian corpuscle. What is clear from this study is that the chemical mediator(s) act via electromechanical coupling initiating a change in ion conductance, which then activates the muscle cell. Indeed, this entire process can be inhibited by blocking Ca 2+ or increasing K + conductance 19 (unpublished observations). It is intriguing that action potentials are not universally recorded along the entire surface of a pressurized artery, but only at the bifurcation points of branching arteries. Obviously, this has important implications which require further independent study before any conclusions can be drawn regarding the function of these spontaneously active cells. It is of merit, however, to point out that an intact endothelium is required to initiate regenerative electrical activity in these cells. References 1. Harder DR: Pressure-dependent membrane depolarization in cat middle cerebral artery. Circ Res 1984;55: Harder DR, Lombard JH: Voltage-dependent mechanisms of receptor stimulation in cerebral arterial muscle, in Bevan JA, Godfraind T, Maxwell RA, Stoclet JC, Worcel M (eds): Vascular Neuroeffector Mechanisms. Amsterdam, The Netherlands, Elsevier, 1985; pp Harder DR, Madden JA: Electrical stimulation of the endothelial surface of pressurized cat middle cerebral arteries results in TTXTsensitive vasoconstriction. Circ Res 1987 (in press) 4. Furchgott RF: Role of endothelium in response to vascular smooth muscle. Circ Res 1983;53: Hickey KA, Rubanyi G, Paul RJ, Highsmith RF: Characterization of a coronary vasoconstriction produced by cultured endo- % thelial cells. Am J Physiol 1985;248:C550-C Gabor M, Vanhoutte PM: Hypoxia releases vasoconstrictor substances from the coronary endothelium (abstract). Circulation 1984;70: O'Brien RF, McMurtry IF: Endothelial cell supernates contract bovine pulmonary artery rings (abstract). Am Rev Respir Dis 1984;192: Katusic ZS, Shepherd JT, Vanhoutte PM: Endothelial-dependent contraction to stretch in canine basilar arteries (abstract). Fed Proc 1986;45: Lombard JH, Smeda J, Madden JA, Harder DR: Effect of reduced oxygen availability upon myogenic depolarization and contraction of cat middle cerebral artery. Circ Res 1986;58: KEY WORDS physiology cerebral arteries myogenic activation endothelium electro-