Background and Purpose Pial arterioles have diverse mechanisms for endothelium-dependent dilations. In mice, different mechanisms or endothelium-derived mediators exist for each of the following dilators: acetylcholine, bradykinin, and calcium ionophore A-23187. This study tests the response to each of these dilators during profound ischemia. The response to sodium nitroprusside, an endothelium-independent dilator, was also tested.
Methods In each mouse, ischemia was produced by bilateral carotid artery ligation that reduced cortical blood flow by approximately 90% as determined by laser-Doppler flowmetry. In separate studies of 10 mice each, dilations of pial arterioles to two doses of each dilator were compared before and after 10 minutes of occlusion, with the occlusion continuing during the second set of measurements. The dilator was applied in the suffusate bathing the pial surface exposed at a craniotomy site. Diameters were monitored by in vivo television microscopy and image splitting.
Results The dose-dependent dilations to acetylcholine, bradykinin, and calcium ionophore A-23187 were each profoundly depressed during ischemia. The response to sodium nitroprusside was not depressed. In all cases, the ischemia was accompanied by arteriolar narrowing of approximately 25% and by obvious slowing of blood flow observed by intravital microscopy. Superoxide dismutase plus catalase failed to prevent the depressed response to acetylcholine.
Conclusions Endothelium-dependent dilations, mediated by diverse endothelium-derived relaxing factors, are depressed during ischemia of 10 to 15 minutes’ duration. This cannot be a nonselective effect on vessel responsivity caused by constriction, reduced flow, or reduced intraluminal pressure during ischemia because under the same conditions dilation to endothelium-independent sodium nitroprusside is preserved. The selective endothelial dysfunction may play a role in exacerbating ischemia by precluding the ability of some dilators, released during ischemia, to dilate the resistance vessels.
The effects of ischemia and ischemia-reperfusion have been extensively studied in several organs, including brain and heart. It is surprising, therefore, that there have been very few publications concerning the effect of ischemia or of ischemia-reperfusion on the vasomotor properties of brain blood vessels.1 2 3 4 The following report describes an inhibitory effect of ischemia on some endothelium-dependent responses in resistance vessels on the brain surface (pial arterioles). The responses to three different endothelium-dependent dilators were tested because these dilators—ACh, BK, and calcium ionophore A-23187—each have a different endothelium-derived mediator by which they dilate pial arterioles.5 6 7 8 In addition, the effect of SNP, an endothelium-independent dilator, was tested. Finally, scavengers of oxygen-centered radicals were tested to see whether they altered the effects of ischemia or the responses to dilator.
Materials and Methods
All procedures were approved by the institution’s animal care and use committee.
Male ICR mice (Harlan–Sprague-Dawley; weight, 27 to 35 g) were used. They were anesthetized with urethan and the pial surface exposed as previously reported (eg, References 9 and 10). Briefly, a tracheostomy and craniotomy were performed, and the dura was stripped from the craniotomy site to expose the pial vessels beneath the transparent arachnoid. The mouse was maintained at 37°C with a heated mattress, and the pial surface was continuously suffused with mock cerebrospinal fluid (Elliott’s solution11 ) at 37°C. The pH of the suffusate was adjusted by bubbling CO2 through it and was always between 7.34 and 7.35, differing slightly within this range from mouse to mouse but always remaining constant for any individual mouse. All drugs were dissolved in the Elliott’s solution except for calcium ionophore. Stocks of the latter were made with the aid of DMSO added to the solution. The final dilutions were made in Elliott’s solution, and in preliminary studies identical amounts of DMSO did not affect the diameter or responsivity of the pial arterioles. All drug solutions, when suffused, had the same pH as the Elliott’s solution.
In each mouse a pial arteriole was selected for continuous monitoring with the use of incident light, microscope, television camera, and monitor. An image-splitting system was used to measure diameter.12 Such systems are capable of measuring changes less than 0.5 μm in size, as explained by Dyson.13 In our laboratory an object 10 μm wide was measured on 20 consecutive occasions with a standard error of 0.2 μm. With 10 μm as a standard, a 5-μm distance was measured and deviated by only 1% (0.05 μm) from the predicted size.
The only criterion for selection of the arteriole to be monitored was an internal diameter between 25 and 40 μm with a straight segment at least 100 μm long. In practice the mean and range of diameters selected for the different groups of mice were virtually identical.
Each mouse had a ligature of 4-0 suture material loosely placed around each common carotid artery. The artery was narrowed by pulling on both ends of the ligature. In these studies both arteries were ligated in this way except in sham-operated mice, in which the ligatures were not tightened.
A laser-Doppler flowmeter and probe14 15 16 were used to continuously monitor flow over the cerebral hemisphere. This method provides a very accurate record of flow expressed in arbitrary units.14 15 16 The probe was placed over the craniotomy at a position adjacent to the monitored vessel. However, preliminary data showed that flow was uniformly depressed over all points in the parietal cortex of both hemispheres (to approximately 10% of control) during bilateral carotid ligation.
The following topically applied drugs were obtained from Sigma Chemical Co: ACh, BK, calcium ionophore A-23187, SNP, SOD, and catalase (thymol free). Only one drug was tested per mouse. Cumulative dose-response curves were obtained with the use of two ascending doses. Each was applied for 2 minutes (a plateau was reached). The application began after 10 minutes of ischemia, and the ligation was maintained throughout.
Neither arterial gases nor systemic blood pressure was monitored during the ischemic period. It was thought that systemic blood pressure was irrelevant since intraluminal pressure must be very low in the ischemic brain with flow only 10% of baseline levels. Similarly, arterial CO2 alters pial arteriolar diameter only insofar as the CO2 controls local pH.17 In cases in which flow is greatly diminished, as in the present study, local pH must be strongly affected by metabolic changes, such as local acidosis, that are not reflected in arterial CO2 levels. Nevertheless, in a separate series of 7 mice, we continuously measured expired CO2 (a function of Paco2) using a microcapnometer (Columbus Instruments), and we continuously monitored blood pressure via the femoral artery. After 10 minutes of bilateral carotid occlusion, the expired CO2 had fallen dramatically in every mouse; the systemic blood pressure rose in 2 and fell in 5 of the 7. In these 5 it was reduced by approximately 40%. For the reasons noted above, we do not believe that such systemic measurements of CO2 or blood pressure are of much value in interpreting events in a local microvascular bed cut off from the systemic blood supply.
Two-way ANOVA was used to analyze the data. Major treatments were before versus during ischemia. Minor treatments were the two doses of the drug being tested.
As indicated below, bilateral carotid ligation uniformly reduced flow to approximately 10% of baseline. This was accompanied by narrowing of the arterioles, and “trickle flow” (barely moving and/or intermittently moving columns of red blood cells) was observed through the microscope. Separate experiments are reported below investigating responses to ACh, SNP, BK, and calcium ionophore A-23187, respectively, before and during bilateral occlusion. The mean flows in each study, as expressed in absolute laser-Doppler flow units, were remarkably similar (66±25, 65±13, 70±11, and 61±14 flow units [mean±SD], respectively), as were flows after 10 minutes of bilateral occlusion (6±2, 5±3, 7±2, and 4±1 flow units, respectively).
Effect of Ischemia on Response to ACh
Fig 1⇓ shows the dose-dependent (P<.001) dilation produced by the endothelium-dependent dilator ACh before ischemia and the significant reduction (P<.01) in magnitude of the response and loss of dose response at the end of a 10-minute period of ischemia. During ischemia the dilation produced by ACh was reduced or zero in 9 of the 10 mice; in 2 mice ACh produced constriction during ischemia. At the time the ischemic response was tested, flow was only 9±6% (mean±SD) of preischemic value; diameter was also reduced, to 72±13% (mean±SD) of its preischemic value of 30±3 μm.
Effect of Ischemia on Response to SNP
Since ischemia reduced flow and diameter, it is possible that one or both of these actions caused a nonspecific loss in dilating ability that would explain the diminished response to ACh. Therefore, SNP, an endothelium-independent dilator, was tested in a separate experiment (n=10). Fig 2⇓ shows absolutely no effect on response at the end of a 10-minute ischemic period, even though flow was reduced to the same degree as in the ACh study (8±5% of base), and diameter was reduced to 78±16% of its preischemic level of 30±2 μm.
Effect of Ischemia on Response to BK or Calcium Ionophore A-23187
Each of these endothelium-dependent dilators was tested in its own, separate experiment. Fig 3⇓ shows the dose-dependent (P<.001) dilation produced by BK before ischemia and the dramatic reduction (P<.01) in response after a 10-minute period of ischemia. The dilation produced by BK was reduced or zero during ischemia in 9 of the 10 mice, and constriction rather than dilation was produced in 3 of them. The flow was reduced to 10±3% of baseline at that time, and diameters were 75±13% of the preischemic value of 31±3 μm.
Fig 4⇓ shows the effect of ischemia on the response to calcium ionophore A-23187. In this experiment the design was changed so the mice did not serve as their own controls. Instead, an ischemic group and a sham-operated group were used. This was done to avoid the possible toxic effects to the vessels of many applications of ionophore. The figure shows the dose-dependent (P<.01) dilation produced by calcium ionophore A-23187 in the absence of ischemia and the large reduction in this response (P<.01) when tested after 10 minutes of ischemia. Flow was reduced to 7±2% of baseline and diameter to 74±15% of the basal value of 29±1 μm.
In addition to the SNP study already presented, we performed time controls in two separate series of sham-operated mice to be certain that responses to either ACh or BK did not simply diminish with the passage of a 10-minute interval between tests. Responses to ACh were identical at the two test points (dilations of 6±2% and 14±1% before and 5±3% and 14±2% after the 10-minute interval; n=5). The response to BK was not affected by the passage of time (dilations of 7±9% and 14±4% before and 7±2% and 13±2% after the 10-minute interval).
Effects of SOD and Catalase
Because superoxide and/or hydroxyl radical may be generated during ischemia4 and because radicals may injure endothelium and/or destroy the endothelium-derived mediator of the dilation produced by ACh, the pial surface of 10 mice was suffused with mock cerebrospinal fluid containing 80 U/mL SOD and 50 U/mL catalase beginning at the time of bilateral carotid occlusion and continuing for 14 minutes until the end of the test of ACh during the occluded period. Before occlusion, dilations of 5±3% and 13±1% were produced in arterioles 29±1 μm wide by 5×10−5 and 4×10−4 mol/L ACh, respectively. After occlusion and in the presence of SOD plus catalase, there was still a profound depression of response (dilations of only 0±3% and 3±2%, respectively). This depression of response during ischemia was similar to and indeed greater (not less) than that produced in a parallel, contemporary control group of 10 mice not treated with the radical scavengers. These mice displayed dilations of 6±3% and 12±4% to the two doses of ACh before ischemia and only 3±3% and 4±3% at the end of the 10-minute ischemic period.
The new findings are the dramatic reduction and/or virtual loss of response to endothelium-dependent dilators during an ischemic period. The period was 10 minutes long, and the tests of responsiveness took place while ischemia was still present, during a period of several more minutes. Thus, this was not a study of responsivity during reperfusion. In fact, we are aware of only three earlier studies of response to agonists during ischemia.1 2 3 Two of these1 2 were before the time when endothelium-dependent responses had been discovered, and both concern responses to constrictors applied after 2 or more hours of ischemia. One of the two studies indicated enhanced responsiveness of pial arteries or arterioles to the constrictor serotonin.1
In the past decade there have been very few reports of responsivity during reperfusion. Dilation by ACh was abolished and replaced by constriction in cats reperfused for 30 minutes after 10 minutes of total cessation of flow to the underlying cortex.4 The authors provided evidence that this might be due to prolonged production of superoxide or hydroxyl reduced by the brain during and for a prolonged period after ischemia.4
The EDRF for dilation by ACh may contain and possibly gives off nitric oxide, but its precise chemical composition has not been determined.18 Because BK and calcium ionophore A-23187 each have their own EDRF and these differ in nature from the EDRF for ACh, we have adopted the nomenclature EDRFACh, EDRFBK, and EDRFionophore. Oxygen-centered radicals destroy EDRFACh.19 In the aforementioned study concerning ischemia-reperfusion,4 no data were provided that would enable one to distinguish between damage of endothelium by the radicals or inactivation of EDRFACh.
The findings in the present study suggest that loss of response to ACh begins during ischemia and is not dependent on reperfusion. Moreover, the result may not be a simple consequence of destruction of EDRFACh by radicals. This is because the responses to BK and calcium ionophore A-23187 were also markedly impaired. Hydroxyl radical is itself the EDRF for BK,6 7 and ionophore is mediated by an eicosanoid8 and is not thought to be destroyed by oxygen-centered radicals. Since responses mediated by three different EDRFs were all greatly impaired during ischemia and only one of the EDRFs is known to be destroyed by radicals, production of the radicals cannot be the sole explanation for our data. Indeed, since we found no protective action of SOD plus catalase on the deleterious effect of ischemia when ACh was the dilator, we cannot even ascribe the impaired response to ACh to either superoxide, hydroxyl ion, or hydrogen peroxide.
The release of EDRFACh and probably that of EDRFBK are receptor mediated. The response to ionophore is not receptor mediated. The effect of ischemia on dilation by all three agonists suggests that ischemia has an action that reduces release of diverse EDRFs whether or not that release is receptor coupled. One obvious factor that might explain the results is the reduction of flow and shear that parallels ischemia. The spontaneous or basal release of at least two different EDRFs, EDRFACh20 21 22 and a prostanoid,23 appears to depend on and/or increase with increasing flow and shear. It may be that profound reduction of shear also prevents or greatly decreases agonist-stimulated release of EDRFs. Basal release of EDRFACh does occur from mouse pial arterioles.24
The responses to agonists in this study were modest to moderate in size, and high doses were required to elicit them. This might be due in part to barriers to penetration through the pia-arachnoid and/or to rapid breakdown of one or the other agonist as it diffused to the endothelium. Unless the steady-state concentration of dilator reaching the endothelium is known, one cannot conjecture about the physiological significance to the mouse of the responses. However, it should be pointed out that in the microcirculation small changes in diameter have much larger effects on resistance because of the cubic relationship between the two variables.25 It should also be pointed out that ACh and BK may each be released by damaged brain tissue and so are available to get to the vessels during or after ischemia.26 27 Calcium ionophore is, of course, neither a physiological nor pathophysiological agonist. However, it directly tests the release of an endothelium-derived mediator triggered by a rise in intracellular calcium, an important vasoregulatory mechanism for release of EDRFs, probably including EDRFs released by shear.21 It has been shown that EDRFACh is released by shear.20 21 22 Consequently, apart from either their physiological or pathophysiological importance to the mouse, the agents and species used here provide a test of mechanisms for cerebrovascular control that may be important in a variety of species and that are called into play and found wanting during ischemia.
A word should be added concerning the use of the term “ischemia” in the present model. Flow was not reduced to zero as in another study, cited earlier.4 However, reductions to approximately 10% of baseline were consistently achieved. Some may wish the term ischemia to refer only to zero flow and to use another term, such as oligemia, to describe lesser degrees of compromise. Such usage might be thought particularly advisable when the laser-Doppler technique is used to measure flow since the technique, although quantitative, gives only relative rather than absolute values.14 15 16 However, ischemia is the term commonly used to refer to severe degrees of oligemia, especially those that produce symptoms. In the present study direct microscopic observations of the pial surface accompanied laser-Doppler measurement of flow in the underlying cortex and confirmed the conclusion that flow was severely impaired. Thus, these observations consistently showed slowing of red cells and trickle flow.
As reported in “Results,” the average preischemic flow in a group of mice tested with one dilator was remarkably similar to that in any of the other groups of mice, with means ranging from 61 to 70 flow units with SDs of 13 to 25 flow units. The decline in flow during bilateral ligation brought flows down to less than 10% of baseline in all four studies. Since the responses to ACh, BK, and calcium ionophore A-23187 were all impaired, it is reasonable to conclude that flows reduced by at least 90% will result after 10 minutes in a general impairment of endothelium-dependent relaxation. A similar conclusion can be drawn if one looks at variability within groups as opposed to between groups. Thus, as noted in “Results,” in cases in which mice served as their own control (ACh and BK groups), the response during ischemia was impaired or eliminated in virtually every mouse (9 of 10 in each study). Therefore, it is clear that essentially all levels of flow reduction within the studied range resulted in impaired endothelium-dependent responses. It should also be pointed out that no mice were discarded because of insufficient flow reduction. That is, all mice were used, and the uniformity of preligation and postligation flows and flow reductions was as just described. No attempt was made to define a threshold of reduction above which responses would be unimpaired. This would require studies with less severe carotid ligation.
Because the laser-Doppler technique does not provide absolute flow values, we cannot provide the flow values before and during ischemia. However, for the reasons stated in the preceding paragraph, it is reasonable to assume they were similar in all groups. Since we are not attempting to provide a threshold, in absolute milliliters per 100 g per minute of flow, for the effect observed, the laser-Doppler technique as used here seems adequate to draw the conclusions we have made, namely, that reductions of flow approximating 90% or greater lead to impaired endothelium-dependent responses.
We have not attempted to define in absolute flow values an ischemic threshold for the appearance of impaired endothelium-dependent dilation. However, this may not be the critical value necessary to predict the effect. During the first 10 minutes it may not be ischemia per se or associated metabolic changes that accounted for the effect but rather the abrupt profound reduction of shear. This reduction in shear may impair the ability of endothelium to synthesize and/or release relaxing factors. In this regard, it is relevant to comment on the marked narrowing of arterioles observed during the 10 minutes of ischemia. Initially, we conceived of this phenomenon as “passive” collapse caused by the presence of tone in the face of greatly diminished distending pressure. Moreover, we thought that because of this great diminution in distending pressure vessels would not dilate to any dilator. This expectation was clearly incorrect since dilation by SNP was not compromised. Since of the four dilators tested only SNP was not compromised, the dilation produced by SNP cannot simply represent passive distension of the arteriole but must represent a metabolic effect in the vascular smooth muscle. This effect is most probably activation of guanylate cyclase, and the resultant decrease in tone permits even minimal intraluminal pressure to dilate the vessel. However, if there is sufficient distending pressure to dilate the arterioles when their tone is decreased by SNP, why isn’t there enough distending pressure to prevent narrowing before drug application? The explanation may reside, at least in part, in the same phenomenon we used to explain the diminished response to all three endothelium-dependent dilators, namely, reduced flow and shear. This will reduce basal release of EDRFACh and prostacyclin. Basal release of EDRFACh has been demonstrated from pial arterioles.24 Consequently, it is suggested that reduced basal release of EDRFACh, perhaps in concert with reduced distending pressure, accounts for the narrowing demonstrated during ischemia. This then would not be mere passive collapse but rather constriction occurring because tone increased as a result of a loss of one or more tone-reducing paracrine mediators.
The preserved response to SNP also indicates that impaired responses to ACh, BK, and calcium ionophore A-23187 were not a nonspecific consequence of reduced flow or of testing narrowed or constricted vessels. The arterioles in the study of SNP were constricted during ischemia by an amount equal to or greater than the constriction seen in the other experiments. Flow was also reduced to comparable degrees in all experiments. In fact, the response to SNP during ischemia was somewhat larger than the response to SNP before ischemia. The difference was not statistically significant. However, the trend is in keeping with the hypothesis that basal release of EDRFACh was reduced. This would lead to reduction of basal levels of cGMP in the vascular smooth muscle. It is well known that the response to activators of guanylate cyclase, like SNP, may increase when basal levels of cGMP have been reduced.
Selected Abbreviations and Acronyms
|EDRF||=||endothelium-derived relaxing factor|
This study was supported by grant HL-35935 from the National Heart, Lung, and Blood Institute.
Reprint requests to William I. Rosenblum, MD, Department of Neuropathology, Medical College of Virginia/Virginia Commonwealth University, Box 980017, Richmond, VA 23298-0017.
Review of this manuscript was directed by Guest Editor Richard J. Traystman, PhD.
- Received March 13, 1995.
- Revision received June 16, 1995.
- Accepted July 6, 1995.
- Copyright © 1995 by American Heart Association
Rosenblum WI. Contractile responses of pial arterioles in gerbils with unilateral carotid ligation. Stroke. 1981;12:83-85.
Mayhan WG, Amundsen SM, Faraci FM, Heistad DD. Responses of cerebral arteries after ischemia and reperfusion in cats. Am J Physiol. 1988;255:H879-H889.
Nelson CW, Wei EP, Povlishock JT, Kontos HA, Moskowitz MA. Oxygen radicals in cerebral ischemia. Am J Physiol. 1992;263:H1356-H1362.
Rosenblum WI, Nelson GH, Shimizu T. L-arginine suffusion restores response to acetylcholine in brain arterioles with damaged endothelium. Am J Physiol. 1992;262:H961-H964.
Rosenblum WI. Hydroxyl radical mediates the endothelium dependent relaxation produced by bradykinin in mouse cerebral arterioles. Circ Res. 1987;61:601-603.
Kontos HA, Wei EP, Kukreja RC, Ellis EF, Hess ML. Differences in endothelium dependent cerebral vasodilation by bradykinin and acetylcholine. Am J Physiol. 1990;258:H1261-H1266.
Rosenblum WI, Nelson GH. Endothelium dependence of dilation of pial arterioles in mouse brain by calcium ionophore. Stroke. 1988;9:1379-1382.
Rosenblum WI, Nelson GH, Povlishock JT. Laser induced endothelial damage inhibits endothelium dependent relaxation in the cerebral microcirculation of the mouse. Circ Res. 1987;60:169-176.
Rosenblum WI, Zweifach BW. Cerebral microcirculation in the mouse brain. Arch Neurol. 1963;9:414-423.
Baez S. Recording of microvascular dimensions with an image splitter television microscope. J Appl Physiol. 1966;21:299-301.
Dyson J. Precise measurement by image splitting. J Opt Soc Am. 1960;50:754-757.
Haberl RL, Heizer ML, Marmarou A, Ellis EF. Laser-Doppler assessment of brain microcirculation: effect of systemic alterations. Am J Physiol. 1989;256:H1247-H1254.
Heistad D, Kontos HA. Cerebral circulation. In: Shepherd JT, Aboud F, eds. Handbook of Physiology: Section 2, Volume 3, Part 1. Bethesda, Md: American Physiological Society; 1983: chap 5.
Rosenblum WI. Endothelium dependent relaxing factor in brain blood vessels is not nitric oxide. Stroke. 1992;23:1527-1532.
Kuchan MJ, Frangos JA. Role of calcium and calmodulin in flow-induced nitric oxide production in endothelial cells. Am J Physiol. 1994;266:C628-C636.
Pohl U, Herlan K, Huang A, Bassenge E. EDRF mediated shear induced dilation opposes myogenic vasoconstriction in small rabbit arteries. Am J Physiol. 1991;261:H2016-H2023.
McIntire LU, Frangos JA, Eskin SG, Hall ER. Effect of hemodynamic shear on arachidonic acid metabolism of vascular endothelium. In: Hartmann A, Kuschinsky W, eds. Cerebral Ischemia and Hemorrhealogy. Berlin, Germany: Springer Publishing Co, Inc; 1987:280-289.
Rosenblum WI, Nishimura H, Nelson GH. Endothelium dependent L-arginine and L-NMMA sensitive mechanisms regulate tone of brain microvessels. Am J Physiol. 1990;259:H1396-H1401.
Mayrovitz HN, Roy J. Microvascular blood flow: evidence indicating a cubic dependence on arteriolar diameter. Am J Physiol. 1983;245:H1031-H1038.
Ellis EF. Initiation of eicosanoid and free radical formation following brain injury: the role of the kallikrein-kinin system. In: Bazan NG, ed. Lipid Mediators in Ischemic Brain Damage and Experimental Epilepsy. New Trends in Lipid Mediators Research. 1990; 4:129-145.