Recovery of Impaired Endothelium-Dependent Relaxation After Fluid-Percussion Brain Injury in Cats

Transcription

1 911 Recovery of Impaired Endothelium-Dependent Relaxation After Fluid-Percussion Brain Injury in Cats Mary D. Ellison, PhD, Daniel E. Erb, MS, Hermes A. Kontos, MD, PhD, John T. Povlishock, PhD The effect of a moderate level of fluid-percussion brain injury on acetylcholine-induced cerebral arteriolar vasodilation was examined for 12 hours after trauma in anesthetized cats equipped with cranial windows. The cats were then perfused with aldehydes, and the pial arteries were prepared for electron microscopy. Immediately after brain injury, the normal vasodilator response to topical application of acetylcholine was converted to vasoconstriction. By 4 hours after trauma, the ability of small pial arterioles (diameters <100 jitm) to dilate after acetylcholine application had returned to the pretrauma level and was observed to be normal at both 8 and 12 hours after trauma (p<0.05). The vasodilator response of large caliber arterioles (diameters > 100 jum) at 4, 8, and 12 hours after injury was reduced relative to the pretrauma response but was significantly improved relative to their response at 30 minutes after trauma (/?<0.05). Moreover, the response of large vessels at 4, 8, and 12 hours in injured animals was equal to that observed in noninjured control animals assessed at 4, 8, and 12 hours after window implantation. At 12 hours after injury, the ultrastructural characteristics of both large and small vessels resembled their preinjury state. These data suggest that the impairment of acetylcholine-induced endothelium-dependent relaxation observed in cats after fluid-percussion brain injury is not irreversible but returns to normal (small arterioles) or exhibits significant recovery (large arterioles) within 4 hours after injury. (Stroke 1989;20: ) During the first 1 to 4 hours after traumatic brain injury in experimental animals, a number of cerebrovascular alterations have been described, including sustained vasodilation, 1-2 reduced vessel-wall oxygen consumption, 2 changes in cerebral blood flow, 1 altered blood-brain barrier permeability, 3 and abnormal vasomotor responsiveness to variations in systemic blood pressure and carbon dioxide tension. 2 Such functional alterations have been shown to be accompanied by discrete morphological lesions in the endothelium, notably in the absence of overt smooth muscle damage. 2 ' 4-6 The failure of cerebrai vessels to dilate after topical acetylcholine (ACh, 10~ 7 M) application has also been observed during the first hour after experimental traumatic brain injury. 7 ACh-induced dilation From the Departments of Anatomy and Medicine, Medical College of Virginia, Richmond, Virginia. Supported by Grants NS and NS from the National Institutes of Health and the R. Clifton Brooks Jr. Scholarship from the Sons of Confederate Veterans. Address for reprints: M.D. Ellison, PhD, Assistant Professor, Department of Anatomy, Medical College of Virginia, MCV Station, Box 709, Richmond, VA Received September 27, 1988; accepted January 19, occurs through the release of an endotheliumderived relaxing factor (EDRF), which, in turn, acts on the vascular smooth muscle. 8 Because traumatic injury simultaneously alters ACh-mediated dilation, endothelial morphology, and vascular function, there may be a direct correlation between these events. The key question with regard to these posttraumatic changes is whether they persist or resolve over time, because prolonged vascular impairment may pose additional risks. In the face of a secondary insult, altered vasomotor function in the resistance vessels could result in compromised cerebrai biood flow to regional vascular beds possibly resulting in central nervous system damage. Because of the potential importance of this problem, we have examined the course of altered endothelium-dependent relaxation (EDR) after concussive brain injury. We report here observations of ACh-induced dilation in pial arterioles examined at 30 minutes and 4, 8, and 12 hours after a moderate level of fluid-percussion brain injury in cats. Additionally, we provide ultrastructural data from vessels harvested after the completion of the functional studies.

2 912 Stroke Vol 20, No 7, July 1989 TABLE 1. Trauma (n=7) ph Paco 2 (mm Hg) Pao 2 (mm Hg) MABP (mm Hg) Control (n=4) ph Paco 2 (mm Hg) Pao 2 (mm Hg) MABP (mm Hg) Time Course of Mean Arterial Blood Pressure and Blood Gas Parameters Time after trauma or baseline measurement Preinjury baseline 30 min 4 hr 8hr 7.39± ±1.6 11O± ± ± ± ± ±5.4 98± ± ± ± ± ±8.9 95± ± ± ±6.4 lll± ± ±6.0 93± ± ± ± hr 7.36± ±2.3 97±6.8 92± ± ± ± ±13.5 Data are mean±sem. Trauma, cats with fluid-percussion brain injury; MABP, mean arterial blood pressure; Control, cats without fluid-percussion brain injury. Materials and Methods For these experiments, 25 adult cats weighing between 2.0 and 3.5 kg were used. Fourteen cats were excluded from the study due to a variety of factors (see "Results"). Cats were anesthetized with sodium pentobarbital (30 mg/kg i.v.), intubated, and, after surgery, paralyzed with gallamine triethiodide (4 mg/kg i.v.) and ventilated mechanically with room air supplemented by O 2 as needed. The right femoral artery and vein were cannulated for blood pressure recording and monitoring of arterial Paco 2, Pao 2, and ph levels and for pentobarbital supplementation throughout the experiment. Body temperature was maintained at 39 C with the aid of a heating pad. Under sterile conditions, two openings were made in the parietal bones, bilaterally in the skull, one for implantation of a cranial window and the other for implantation of a metal fitting that was later connected to the fluid-percussion apparatus. 9 The cranial window and the technique for implanting it have been described in detail previously. 10 The cranial window was implanted in the left parietal bone after incision of the dura and allowed visualization of the pial vessels overlying the suprasylvian and ectosylvian gyri. The space under the cranial window was filled with sterile artificial feline cerebrospinal fluid, and the two openings in the window were attached to stopcocks used as an inlet and outlet for flushing the space with various solutions. The pial microcirculation was observed with a Leitz/ Wild stereomicroscope (Heerbrugg, Switzerland) equipped with a Vickers image splitting device (Maiden, Massachusetts). The experimental design was as follows: With arterial blood gas values within normal limits, 11 each cat was placed under the microscope, and control diameters of selected pial arterioles were measured. In each cat, both small arterioles (diameters <100 fim) and large arterioles (diameters >100 fim) were selected for study. After baseline diameter measurements, pial arteriolar responses to topical application of 10 7 M ACh (actetylcholine hydrochloride, Sigma Chemical Co., St. Louis, Missouri) were then examined. The window was subsequently flushed with fresh cerebrospinal fluid. In one group of cats, after baseline measurements were made and responses to ACh were assessed, each cat was subjected to a moderate level of fluid-percussion brain injury. 9 The injury device consisted of a Plexiglas cylinder filled with isotonic saline. The cylinder was mounted on a metal frame and bounded at one end by a Plexiglas piston mounted on O rings. The opposite end was connected to a saline-filled transducer housing that was connected to the metal fitting on the cat's skull. From a predetermined height, a pendulum was allowed to fall on the piston, generating within the saline column a transient (18-23 msec) pressure pulse that impacted the intact dura mater at the site of the injury fitting. A moderate injury level was chosen because it is known to elicit functional vascular abnormalities without the subarachnoid hemorrhage seen with severe injury. After brain injury, responses to ACh were reexamined at 30- minute and 4-, 8-, and 12-hour intervals. Mean arterial blood pressure levels and arterial blood gas values were also recorded at these time points (Table 1). If any vessel failed to dilate after posttrauma ACh application, an additional test was performed to determine whether the vascular smooth muscle was maximally relaxed in these vessels and incapable of further vasodilation. For this purpose, vascular responses were examined after topical application of sodium nitroprusside (Sigma Chemical Co., 0.25 fig/m\ cerebrospinal fluid), a potent smooth muscle relaxant. Cats in a second group served as controls and were not subjected to fluid-percussion injury. These cats were followed for 12 hours after the implantation procedures. Baseline arteriolar measurements were made, and the same vessels were reexamined 30 minutes and 4, 8, and 12 hours later in order to assess the effects of the preparation and of extended barbiturate anesthesia on pial vascular responses to ACh.

3 Ellison et al Injury and Endothelium-Dependent Relaxation 913 After completion of the functional studies, the pial vessels were examined by transmission electron microscopy. Brains were fixed by transcardial perfusion with aldehydes. 12 After fixation, the cranial window was removed, and the associated pial vessels were dissected out in accordance with a protocol described previously. 12 After removal, the vessels were processed in accordance with one of two selected strategies. Some vessels were processed for routine transmission electron microscopy. Others were processed for the visualization of endogenous immunoglobulin G (IgG) within the vascular endothelium by use of biotinylated anti-cat IgG. 13 This immunocytochemical approach was used to determine if any endothelial IgG flooding had occurred because such an event has been linked with impending endothelial cell death. 14 These morphological studies were performed in both control and injured cats. Only vessels whose responses to topical ACh application had been measured throughout the experiment were examined so that the morphological status of the endothelium could be correlated with the recovery of EDR. Results Of the 25 cats prepared for injury or control studies, 11 were included in the analysis of data. For varied reasons, the remaining 14 experiments could not be followed to completion. Such reasons included posttraumatic brain swelling, subarachnoid hemorrhage, coverglass breakage, and white blood cell accumulation under the window. Several cats did not exhibit consistent pial arteriolar AChinduced vasodilation before injury. Four of the six control cats were followed for 12 hours and were included in the study. Seven cats subjected to fluid-percussion injury were included. The mean injury level was 2.65 atm. After trauma, the mean arterial blood pressure was transiently elevated from 97±6.5 to 199±8.6 mm Hg (mean±sem) but returned within 10 minutes to the pretrauma level. Dilation of pial arterioles was observed after injury and was more pronounced in small arterioles than in large ones. Small arterioles exhibited a 32% increase in the mean diameter at 30 minutes after trauma (compared with the pretrauma mean diameter), whereas the large vessel mean diameter increased by 11%. These findings are consistent with our previously published observations. 2 Over 12 hours, the time course of changes in ACh-induced EDR was examined in two groups of cats. In the injured group, the normal vasodilation observed after ACh application was converted to vasoconstriction at 30 minutes after trauma in both large and small vessels (Figure 1). After sodium nitroprusside application at 30 minutes after trauma, dilation was observed in all vessels except in one cat. In this cat, five of the eight vessels with impaired EDR also failed to dilate when exposed to sodium nitroprusside. Repeated-measures analysis of variance showed that the response of small < o < CM ^ b i 11 s N OJ H* ^ SMALL ARTERIOLES D LARGE ARTERIOLES H «P = o. «g - To * * T» K* ^L ft CONTROL 30 MIN 4HR 8HR 12 HR FIGURE 1. Histogram of percent changes in cerebral arteriolar diameters caused by 10~ 7 M acetylcholine application immediately after cranial window implantation and at 30 minutes and 4, 8, and 12 hours after moderate level of fluid-percussion brain injury. The mean±sem of the baseline diameters, from which the percent changes caused by ACh were calculated, are given in \im. * Significantly different relative to control (p<0.05). vessels was significantly reduced, relative to pretrauma responses, only at 30 minutes after trauma (p<0.05). By 4 hours after injury, the response of small arterioles had returned to pretrauma level at which the group continued to respond 8 and 12 hours after trauma. Repeated-measures analysis of variance showed that the response of large arterioles at 30 minutes and 4, 8, and 12 hours after injury was reduced significantly relative to the pretrauma level (p<0.05). At 4 hours, however, the response was significantly improved compared with that observed at 30 minutes (p<0.05). In the control group, not subjected to fluid-percussion injury, there was a reduction, over time, in the ability of large vessels to dilate after ACh application (Figure 2). Repeated-measures analysis of variance showed that responses of large vessels were significantly lower relative to control at 30 minutes and 4, 8, and 12 hours after injury (p<0.05). The same analysis showed that in small vessels the degree of AChinduced vasodilation that was observed at each application remained statistically equal to the baseline level. Analysis of covariance showed that the responses of large vessels in injured cats did not differ significantly from those in control cats at 4, 8, and 12 hours after baseline measurements. Responses of small vessels differed significantly between control and injured cats at 8 hours only, at which time control cats showed an elevated response relative to injured cats at the same time point. Arterial Paco 2, Pao 2, ph, and mean blood pressure levels remained relatively constant throughout the experiment and were comparable at each time point in the control and the injured groups (Table 1). o

4 914 Stroke Vol 20, No 7, July SMALL ARTERIOLES O LARGE ARTERIOLES CONTROL 30 MIN 4 HR 8 HR 12 HR FIGURE 2. Histogram of percent changes in cerebral arteriolar diameters caused by 10' 7 M acetylcholine application immediately after cranial window implantation and at 30 minutes and 4, 8, and 12 hours thereafter. The mean±sem of the baseline diameters, from which the percent changes caused by acetylcholine were calculated, are indicated in fim. * Significantly different relative to control (p<0.05). In the injured cats, the macroscopic morphological features of the brain were the same as those reported previously by investigators using the same technique. 4 " 6 ' 9 In the cortex, there was focal contusion confined to the site directly incident to the fluid-percussion pulse. There was no evidence of subarachnoid bleeding at the injury site. Control cats showed no visible cortical lesions. Ultrastructural analyses of vessels harvested 12 hours after injury revealed little evidence of change when compared with the noninjured controls (Figure 3). Only occasional smooth muscle and endothelial abnormalities were identified. The abnormalities seen within the arteriolar smooth muscle consisted primarily of electron lucent inclusions similar to those previously described in vitro in mechanically stretched smooth muscle. 15 Endothelial change consisted of occasional cytoplasmic vacuoles and foci of perinuclear swelling. In contrast to our previous observations, the conspicuous luminal membrane blebs and concavities that were consistently observed during the early posttraumatic period were not commonly observed at 12 hours after trauma. Isolated examples of necrotic endothelium were seen dispersed among other unaltered endothelial cells, and occasional IgG-laden endothelial cells were also identified in two of 12 vessels examined for the immunocytochemical recognition of the endogenous immunoglobulin (Figure 4). Discussion The results of the present study show that the loss of ACh-induced endothelium-dependent relaxation in pial arterioles after brain injury is not irreversible but returns to normal levels in small vessels within 4 hours after trauma. At 4 hours after injury, the response in large vessels is greatly improved in comparison with that observed at 30 minutes after trauma. Previously, we have demonstrated the loss of EDR during the first hour after traumatic brain injury. 7 In the present study, we extend these observations, demonstrating that the posttraumatic loss of normal ACh-induced vasodilation does not persist indefinitely. Normally, cerebral vessels, 16 like vessels in other vascular beds, 8 dilate by an indirect mechanism when exposed to certain agents such as ACh. The mechanism involves the action of the dilating agent on the endothelium and causes the secondary release of an EDRF, which induces vascular smooth muscle relaxation. When the normal vasodilating action of ACh is converted to vasoconstriction after fluid-percussion brain injury (Figure 1), the direct effect of ACh on vascular smooth muscle (vasoconstriction) becomes unmasked as the vasodilating influence of EDRF is eliminated. The causative mechanisms involved in the transient elimination of EDR after trauma are unknown. Data from a number of previous studies show that free oxygen radicals can be involved in the impairment of EDR.-> > Studies in our laboratory have shown that the hypertensive episode resulting from the fluid-percussion brain injury is directly linked to the generation of free oxygen radicals, which results in a number of morphological and functional pial vascular abnormalities.---- Additional studies of acute hypertension have demonstrated that after a hypertensive insult ACh-induced vasodilation is converted to vasoconstriction in pial arterioles and that the vasodilicationofreradicalscavengers.lating influence of ACh can be restored immediately upon topical application of free radical scavengers. 18 Therefore, it seems reasonable to suggest that free oxygen radicals, generated by the posttraumatic hypertensive episode, participate in the posttraumatic loss of ACh-induced EDR. Furthermore, since the action of EDRF is rapidly restored on application of radical scavengers after hypertension, it seems likely that the radicals directly interfere with the EDRF itself. Alternately, it might be proposed that free radicals, or other phenomena, interfere not with EDRF but with ACh, ACh receptors, or mechanisms of smooth muscle relaxation through cyclic guanosine monophosphate (cgmp). However, the normal direct action of ACh on vascular smooth muscle (vasoconstriction) is observed after injury; this suggests that neither ACh itself nor its receptors are destroyed. To investigate the possibility that normal ACh-induced smooth muscle relaxation was hampered at the level of cgmp, which normally mediates ACh-induced relaxation, sodium nitroprusside (0.25 /ig/ml artificial feline cerebrospinal fluid) was superfused under the window when vessels failed to dilate after ACh application. At 30 minutes after injury in every cat but one, vessels exhibited vasodilation after nitroprusside application. Because sodium nitroprusside and EDRF both effect vasodilation through cgmp, it appears that this mechanism is intact in the present study. In the

5 Ellison et al Injury and Endothelium-Dependent Relaxation FIGURE 3. Electron 915 micro- graph revealing normal ultrastructural detail in arteriole harvested 12 hours after injury. Note that both the endothelium (arrow) and underlying smooth muscle (SM) appear unremarkable with exception of small luminal endothelial membrane defect only rarely observed in the present study. Original magnification, x 5,000. I l.ctym cat that did not show vasodilation to nitroprusside, it is possible that the vessels were already maximally dilated as a consequence of the injury. A number of possible explanations can be advanced to explain the return of EDR over time after injury. It is notable that the various endothelial morphological alterations immediately after injury that have consistently been observed and described in detail2-4-6 were not observed in the present study. It therefore appears that the acutely observed alterations had resolved by 12 hours after trauma. Perhaps the return of endothelial morphological integrity not only parallels but also accounts for the return of EDR. However, this seems unlikely in view of the rapid restoration of EDR that has been demonstrated in hypertension studies after radical scavenger application.18 An alternative explanation for the recovery of EDR involves the role that posttraumatic free oxygen radical production may play in its impairment. Support for this hypothesis lies in preliminary data, collected during the present study, regarding posttraumatic prostaglandin synthesis. Because prostaglandin synthesis has been linked with the metabolism of arachidonate and the production of oxygen radicals, it seemed reasonable to consider that pros-

6 916 Stroke Vol 20, No 7, July 1989 FIGURE 4. In some cases, electron microscopic examination reveals isolated endothelia innundated with IgG (arrow). Note that despite IgG flooding, adjacent endothelial cells (asterisks) and adjacent smooth muscle are unaltered. Original magnification, x.8,000. to v - * taglandin levels of cerebrospinal fluid would reflect levels of free radical production.4'7'17-18 In two of the traumatized cats, cerebrospinal fluid samples were collected from the cranial window before and at 30 minutes and 4, 8, and 12 hours after trauma. Data from the two cats suggest a positive correlation between prostaglandin E2 levels and the number of pial arterioles with compromised EDR. In one cat, both were highest at 30 minutes after trauma. In the second cat, both were highest at 4 hours after injury. In both cases and for the remainder of the experiment, prostaglandin E2 levels dropped as the degree of EDR improved. This occurrence supports the suggestion that free radicals, produced concomitantly with prostaglandin E2, interfere with EDRF. The possible explanations for the reduced responsiveness to ACh exhibited over time by large arterioles are more elusive. The effect apparently is not related to the fluid-percussion injury because the noninjured control cats also showed reduced responsiveness over time in the large vessels. It appears that some as yet undetermined aspect of the preparation and/or extended anesthesia has an effect on their endothelium-dependent responsiveness to ACh that is perhaps due to a specific effect on smooth muscle, which is more abundant in large arterioles. Although it is difficult to speculate as to the overall physiologic consequences of altered EDR after traumatic brain injury, it most likely constitutes one of multiple vascular perturbations that

7 Ellison et al Injury and Endothelium-Dependent Relaxation 917 potentially predispose the central nervous system to damage after traumatic brain injury. These studies in cats suggest that this altered response recovers over time after injury; therefore, one can argue that clinically, after a comparable injury, a posttraumatic window exists for a time during which the injured brain may be more highly sensitive to insult. Acknowledgments We thank Enoch Wei for his many helpful comments, Ms. Susan Walker for her technical assistance, and Ms. Elaine Pollard for her help in preparing the manuscript. References 1. Dewitt DS, Jenkins LW, Wei EP, Lutz H, Becker DP, Kontos HA: Effects of fluid-percussion brain injury on regional cerebral blood flow and pial arteriolar diameter. / Neurosurg 1986;64: Wei EP, Dietrich WD, Povlishock JT, Navari RM, Kontos HA: Functional, morphological, and metabolic abnormalities of the cerebral mirocirculation after concussive brain injury in cats. Circ Res 1980;46: Povlishock JT, Becker DP, Sullivan HG, Miller JD: Vascular permeability alterations to horseradish peroxidase in experimental brain injury. Brain Res 1978;153: Wei EP, Kontos HA, Dietrich WD, Povlishock JT, Ellis EF: Inhibition by free radical scavengers and by cyclooxygenase inhibitors of pial arteriolar abnormalities from concussive brain injury in cats. Circ Res 1981;48: Povlishock JT, Becker DP, Sullivan HG, Miller JD: Vascular permeability alterations to horseradish peroxidase in experimental brain injury. Brain Res 1978;153: Povlishock JT, Kontos HA, Wei EP, Rosenblum WI, Becker DP: Changes in the cerebral vasculature after hypertension and trauma: A combined scanning and transmission electron microscopic analysis, in Eisenberg HM, Suddith RL (eds): The Cerebral Microvasculature. New York, Plenum Publishing Corp, 1980, pp Kontos HA: Oxygen radicals in experimental brain injury, in Proceedings of the International Symposium on Intracranial Pressure and Brain Injury, (in press) 8. Furchgott RF: Role of endothelium in responses for vascular smooth muscle. Circ Res 1983;53: Sullivan HG, Martinez J, Becker DP, Miller JD, Griffith R, Wist AO: Fluid-percussion model of mechanical brain injury in the cat. / Neurosurg 1976;45: Levasseur JE, Wei EP, Raper AJ, Kontos HA, Patterson JL: Detailed description of a cranial window technique for acute and chronic experiments. Stroke 1975;6: Herbert DA, Mitchell RA: Blood gas tensions and acid-base balance in awake cats. JAppl Physiol 1971;30: Dietrich WD, Wei EP, Povlishock JT, Kontos HA: Method for morphophysiological study of specific pial microvessels. Am J Physiol 1980;238:H172-H Ellison MD, Povlishock JT, Merchant RE: Blood-brain barrier dysfunction in cats following recombinant interleukin- 2 infusion. Cancer Res 1987;47: Hansson GK, Schwartz SM: Evidence for cell death in the vascular endothelium in vivo and in vitro. Am JPathol 1983; 112: Somlyo AP, Somlyo VA, Devine CE: Aggregation of thick filaments into ribbons in mammalian smooth muscle. Nature [NewBiol] 1971;231: Rosenblum WI: Endothelial dependent relaxation demonstrated in vivo in cerebral arterioles. Stroke 1986;12: Kontos HA, Wei EP: Superoxide production in experimental brain injury. J Neurosurg 1986;64: Wei EP, Kontos HA, Christman CW, Dewitt DS, Povlishock JT: Superoxide generation and reversal of acetylcholine-induced cerebral arteriolar dilation after acute hypertension. Circ Res 1985;57: KEY WORDS acetylcholine endothelium cats