- Email: [email protected]
A model of embolic cerebral ischemia in the rat
EXPERIMENTAL
NEUROLOGY
96,393-405
(1987)
A Model of Embolic Cerebral
lschemia in the Rat
PAUL L. PENAR’ Section
OfNeurosurgery,
Received
Yale University
August
School
2 7, 1986; revision
ofMedicine,
received
New Haven,
December
Connecticut
06510
9, 1986
Experimental studies of embolic cerebral &hernia using the rat are limited by variability in the location, size, and frequency of lesions produced. A technique is described herein which improves the reliability of an established model. Eight male Sprague-Dawley rats underwent injection of the cervical internal carotid artery with 0.1 ml of 1-h-old fragmented autologous blood clot through an external carotid artery cannula. The pterygopalatine artery was ligated prior to embolization. At killing 2 h after embolization, clot was observed in the proximal middle cerebral and posterior cerebral arteries in all animals. Areas of reduced blood flow at 2 h postembolization were assessedby digital image processing of iodo-[‘4C]antipyrine autoradiographic images. No-flow and low-flow areas were measured for each of approximately 25 serial brain sections with a computerized bit-pad. Volumes were calculated and lesions localized by anatomical reconstructions. No animal sustained a hemorrhagic lesion. One animal sustained only a very small area of &hernia in the internal capsule. Of the remaining seven, all had large regions of ischemia in the middle cerebral distribution involving cortex and basal ganglia. Posterior cerebral involvement was observed in six of the seven animals as well. The contralateral hemisphere was unaffected. Volume values could be calculated for primary vascular distributions. Most variability occurred in the pattern of posterior cerebral involvement. The technique described produces a relatively consistent region of &hernia in the middle and posterior cerebral artery distributions in the rat and is a useful model for the study of cerebral &hernia. 0 1987 Academic Press, Inc.
Abbreviations: BW-body weight; ECA-external carotid artery; LF, NF-low flow, no flow. ’ This research was supported in part by grants NS- 10 174 and NS- 19430 from the National Institutes of Health. Presented in poster format at the American Association of Neurological Surgeons Annual Meeting, April 13 to 17, 1986 at Denver, Colorado. The author thanks Dr. Carole C. LaMotte for the Quandens imaging system, and acknowledges the assistance of Dr. Charles Greer.
393 0014-4886/87 $3.00 Copyright 0 1987 by Academic Press, Inc. All rights of reproduction in any form reserved.
394
PAUL L. PENAR
INTRODUCTION Cerebral embolization with homologous blood clot in the rat has been shown to be a practical and inexpensive model of focal cerebral ischemia (11). Recently, in modification of Kudo’s procedure (1 l), Kaneko et al. (8) described embolization into the internal carotid artery of the rat with the pterygopalatine artery ligated. Variable results occurred, including bilateral middle cerebral artery occlusions. Distal migration of emboli was seen after initial proximal occlusion of an involved vessel. In an effort to improve the reproducibility of the embolic model of cerebral ischemia in the rat, we introduced a modification ofthe procedure previously described. In particular, clot preparation and volume of injection were altered. We then undertook to define the volume and distribution of regions in the rat brain and mesencephalon affected by internal carotid embolization with this preparation of homologous blood clot, as defined by iodo[‘4C]antipyrine autoradiographic techniques. METHODS
Experimental Procedure. Eight male Sprague-Dawley rats weighing 3 14 to 459 g were anesthetized lightly with ether and then received i.p. injections of 40 mg/kg sodium pentobarbital (Nembutal, 50 mg/ml, Abbott Laboratories). A tracheostomy was carried out, followed by insertion of right femoral arterial and venous polyethylene catheters. Arterial pressure was continually recorded via transducer and chart recorder (Grass Instrument Co.). Temperature was monitored by rectal probe and maintained at approximately 37°C. Anesthesia was maintained with intermittent intravenous and intraperitoneal doses of pentobarbital. Arterial blood gasses were measured by loo-p1 capillary tube sampling prior to embolization and just prior to killing. With the aid of a Zeiss operating microscope and electrocautery, the right digastric and stemohyoid muscles were transected, and the wing of the hyoid bone removed. The external maxillary artery, superior thyroid artery, and occipital artery (overlying the carotid bifurcation) were coagulated and severed. Without dissection of the carotid bifurcation, the right pterygopalatine artery was ligated with 4-O silk suture and coagulated. The external carotid artery (ECA) was ligated distal to the bifurcation and cannulated with polyethylene catheter (PE-50, Clay Adams; i.d. 0.58 mm, o.d. 0.965 mm). The catheter was advanced to the carotid bifurcation and secured in place with a ligature and tissue adhesive at its point of entry into the ECA. The blood clot was prepared in the following fashion: 0.6 to 0.8 ml blood was removed from the animal 60 to 90 min prior to embolization. Serum
RAT EMBOLIC
STROKE
MODEL
395
was nearly completely extruded from the syringe in which the blood had clotted, and the remaining material was passed once through a 20-gauge needle and then through a 22-gauge needle. The clot (with a volume of approximately 0.3 to 0.4 cc) was loaded into a l-ml syringe attached to a 22-gauge needle and connected to the ECA catheter. Air bubbles were eliminated by careful insertion of the needle into the catheter. Microscopic examination of sample clot preparations showed that fragments of clot ranged from 0.2 to 2.0 mm in length, the majority being less than 0.4 mm. With a temporary aneurysm clip occluding the common carotid artery at the base of the neck, 0.1 cc of clot was injected by hand into the external carotid catheter during 2 min. If all this clot volume could not be injected (due to resistance), the injection was considered to be inadequate. The common carotid clip was removed and the catheter withdrawn. The ECA was coagulated distally. Prior to killing at 120 min postembolization, 0.9 ml blood was withdrawn for chemical assay. A dose of 240 &/kg of iodo-4-[Nmethyl’4C]antipyrine in saline (100 &i/ml, American Radiolabeled Chemicals Inc., St. Louis, Missouri) was injected i.v. followed 10 s later by 2 ml saturated KC1 solution. This short delay allowed for distribution of the tracer along major vascular channels, reducing the significance of tracer distribution through collateral pathways. Preparation of Tissue. The brain and brain stem were removed within 10 min of death and frozen in isopentane cooled to approximately -70°C in dry ice. The carotid system in the neck was removed and placed in Formalin. Sectioning of the brain was carried out within 20 h of freezing on a cryostat (Damon IEC) in 32-pm sections after mounting the specimen with EM 1 embedding matrix (Lipshaw Manufacturing Corp., Detroit, Michigan). Two sections were saved on a glass coverslip after discarding 20 sections. Starting from the olfactory bulbs, sections were taken until occipital pole tissue was no longer visible, which included the mesencephalon to the level of the superior colliculus but below the interpeduncular nucleus. The specimens were then dried at 70°C for approximately 30 min, mounted on cardboard, and incubated 6 days with Kodak SB-5 film. Histologic sections were prepared with cresyl violet stains of mounted specimens. Volumetric Analysis. Of the two sections saved, the higher quality specimen was chosen for photographic enlargement to 100 times original size (Durst Laborator S-45 EM) as a negative image on PI-1 Rapitone paper. A frequency analysis of the distribution of optical intensities from a representative non-specimen-containing portion of each individual autoradiographic plate was obtained with the Quandens image analysis system (FHC Inc., Brunswick, Maine). Local tissue concentrations of the carbon- 14 isotope are roughly proportional to autoradiographic density. Blood flow within the
396
PAUL L. PENAR
FIG. 1. Sample autoradiographic images (left) and corresponding processed images (right) demonstrating region of hypoperfusion. Dark band represents range of densities from greaterthan-background to 20% of maximum density.
specimen is related to 14C concentration by means of a polynomial relationship which takes into account the partition coefficient ofthe tracer (15). A value was established for background intensity as the minimum background intensity extrapolated from the large background peak. This value was used to delineate regions of absent perfusion in the specimen. Thus density and not actual blood flow was used to define regions of “no flow.” The optical intensity corresponding to 20% of the maximum density (in the maximumto-background density range) was calculated to obtain an additional boundary in the specimen. This value was less sensitive to changes in maximum density values (maximal blood flows). All autoradiographic images were then reviewed using these values to define areas of nonlabeling and of reduced labeling with tracer (termed “noflow” and “low-flow” regions, respectively, see Fig. l), which were then delineated on the photographic enlargements. The areas of the left and right halves of the hemispheres, brain stem, and diencephalon were measured for
RAT EMBOLIC
STROKE
397
MODEL
TABLE 1 Mean Arterial Pressure and Blood Gas Values MAP” Preembolization Prekilling
134* 7 131+18
PO2
80.7 + 10 88.6 + 16
PC02
PH
39.6 f 3.7 36.4 f 4.9
7.390 f 0.076 7.412 f 0.052
a MAP = mean arterial blood pressure. MAP, PO,, and PCOZreported as mean + SD mm Hg.
each image using a microcomputer bit pad (Summagraphics Bit Pad One) interfaced with a PDP- 11 computer, as were the areas designated as no-flow and low-flow regions. Each measurement of area was repeated and the results averaged for that section. Correction for magnification in the process of enlargement was accomplished with measurements of an enlarged 10 X 10 mm standard. The volumes were then calculated using the average areas of the two contiguous saved sections to represent the entire sequence of 20 discarded slices; the more posterior sections were added into the total for the sequence. The most anterior and most posterior discarded sections were represented by the nearest saved section. Corrections were made for lost sections or irregularities in the cutting process. RESULTS Physiologic Measures, Anesthesia, and Fluids. Animal weight ranged from 314 to 495 g (405 + 53, mean + SD). Table 1 lists values for mean arterial pressure, arterial PO* and Pco~, and pH preembolization and prekilling. Animals received an average of 0.06 + 0.01 mg pentobarbitai sodium/g body weight (BW) during the course of the experiment, and 0.02 + 0.009 ml saline/g BW. Neither the barbiturate dose or fluid dose nor the arterial pressure or blood gas values were significantly correlated with no-flow (NF) or low-flow (LF) volumes. Observations. The cervical internal carotid arteries of animals 1,2,3,4,5, and 8 appeared by visual inspection to contain significant clot at the time of killing, with dark blue discoloration and weak or absent pulsation. The common carotid arteries of animals 2 and 3 had a similar appearance. The cervical carotid systems of rats 6 and 7, in contrast, appeared normal in pulsation and coloration throughout the experiment. With the sole exception of rat 8, all animals harbored blood clot in the internal carotid, proximal middle cerebral, and proximal posterior cerebral vessels at the base of the brain.
398
PAUL L. PENAR TABLE 2 Brain Values, “No Flow” (NF) and “Low Flow” (LF)” Volume (mm3)
Rat No.
Total brain
Right brain
NF
LF
NF%of r. brain
LF%of r. brain
1 2 3 4 5 6 7 -
1800 1655 1823 1813 1692 1628 1618
921 922 898 906 895 908 904
342 216 476 421 118 286 178
569 389 657 589 265 472 331
37.2 23.5 53.0 46.5 13.2 31.5 19.7
61.8 42.2 73.2 65.0 29.6 51.9 36.6
sz,
1718 91
908 10
291 131
467 145
32.1 14.5
51.5 16.1
’ Raw values for total brain volume are given, followed by volumes corrected for total brain volume.
Animal 8 demonstrated only a small proximal middle cerebral clot. The brains of rats 2,5, and 7 showed hemispheric hyperemia on gross inspection prior to freezing. V&metric Analysis. Table 2 shows the volume values measured for seven animals, corrected for a mean total brain volume of 17 18 f 9 1 mm3. Animal 8 had only a very small region of low blood flow in the region of the internal capsule; pentobarbital dose (0.07 g/g BW), saline dose (0.02 ml/g BW), mean arterial pressures, and blood gas values were not noticeably different for this animal. Mean NF volume was 29 1 + 13 1 mm3 for the remaining seven animals; LF volume was 467 + 145 mm3. NF volume occupied 32 f 14.5% of the right brain. Low-flow volumes were comparatively less variable than no-flow volumes. Figure 2 demonstrates a reconstruction based on the atlas of Kiinig and Klippel(l0) of the anatomic regions affected for rat 6, an animal with NF and LF volumes of approximately average size. The middle cerebral artery distribution was most involved. The head of the caudate, globus pallidus, and putamen were in no-flow zones, as was much of the adjoining cortex. The hippocampus was nearly completely involved, except its most superior extent. The occipital cortex and lateral brain stem, both in the posterior cerebral artery distribution, were also affected. The anterior cerebral distribution was spared. The nucleus accumbens was unaffected, as were the more medial
RAT
EMBOLIC
STROKE
MODEL
399
400
PAUL
L. PENAR
thalamic nuclei. In the case of animal 3 (with the largest volume affected) a similar pattern of involvement was seen, but with a larger amount of NF region in the posterior thalamus, occipital cortex, and lateral brain stem (posterior cerebral distribution). By contrast, three animals (2, 5, and 7) showed relatively smaller regions of involvement. In this group of animals (all of which had observable hyperemia of the cortex), the anterior cerebral distribution was for the most part as equally unaffected as in the group with larger NF volumes. There was greater sparing of the medial portions of the caudate nucleus, lenticular nuclei, and lateral thalamus, together with relative sparing in the posterior cerebral circulation. The low-flow region surrounding the no-flow region in the middle cerebral distribution was broader. The hippocampus was largely unaffected in rat 5; the brain of rat 2 did show involvement of the hippocampus and lateral thalamic nuclei. The regions of low flow around areas of no flow were larger in the three animals with smaller NF volumes, relative to the total region affected (LF volume). The NF:LF ratios of animals 2,5, and 7 were all less than 0.60; NF: LF ratios for the remaining four animals ranged from 0.60 to 0.72. Sequential histogram plots of right brain volume, NF volume, and LF volume were constructed for each animal based on identification within autoradiographic images of specific nuclear groups and major fiber tracts (Fig. 3). These plots illustrate a consistent involvement in the distribution of the middle cerebral artery, with variable involvement of structures supplied by the posterior cerebral artery. Volume sections 10 to 16 (just caudal to the head of the caudate, at the genu of the corpus callosum, through to the midthalamus) are more specific for middle cerebral distribution. For these sections, NF volume averaged 124 f 52 mm3 for the seven animals; LF volume mean was 190 + 44 mm3. NF volume occupied 40 f 15% of a mean right brain volume of 3 10 f 20 mm3 for these sections. LF volume comprised 6 1 + 12% of right brain volume. The maximal NF area in a given section for each animal ranged from 35 to 7 1% and corresponded well with the trend in total NF and LF volumes. These maximal points were located from sections 15 to 19 in the seven animals, a region supplied by both middle and posterior cerebral arteries. DISCUSSION Experimental cerebral &hernia produced by vascular occlusion is typically accomplished by embolization techniques or by external direct vascular occlusion. Although the lesions produced by embolization have not been quite as consistent or reproducible as those produced by direct vascular oc-
RAT EMBOLIC
STROKE
MODEL
IO 7Or
IX
50 30 IO
Ld Rostrol
CoudaI
Rostra1
Caudal
Right brain volume Low flow dume Noflow volume FIG. 3. Sequential histogram plots of involvement for individual volume sections from rostra1 to caudal. Maximal NF volumes were located from sections 15 to 19, supplied by both middle and posterior cerebral arteries. Anatomic figure from atlas of Kijnig and Klippel, used with permission.
402
PAUL
L. PENAR
elusion, embolization involves no direct invasion of the cranial vault and therefore carries less risk of alteration of the local response to injury, changes in blood-brain barrier permeability, and direct cortical damage (4,7). Internal carotid artery embolization in the rat has several advantages as a model of cerebral ischemia. This inexpensive animal possessesa cerebral circulation similar to that of man (24), and an identical cervical carotid system but for the pterygopalatine artery, a large branch of the internal carotid artery. Furthermore, the use of homologous clot provides the opportunity to investigate a variety of clinically relevant situations, including the effects of thrombolytic substances following acute arterial occlusion by thrombus or a thrombogenic substance ( 13, 17). In spite of the small number of animals studied in this preliminary investigation, we have shown that a reproducible region of severe hypoperfusion can be produced in the rat brain ipsilaterally by embolization into the internal carotid artery with a fixed amount of homologous blood clot. This region of ischemia lies within the middle and posterior cerebral distributions, and results from proximal occlusion of these vessels at the base of the brain. The greatest variability occurred in the degree of posterior cerebral involvement. The principal differences between our experimental technique and that of Kaneko et al. (8) is in the preparation of the clot used for embolization: fragmentation of clot through a 27-gauge needle followed by dilution with 0.3 ml saline resulted in a clot size of 100 to 200 pm in a total volume of 0.4 ml (8). Our technique resulted in much larger clot fragments (to 2 mm in length) which produced no detectable contralateral regions of reduced perfusion, and more consistent proximal vessel occlusion. A further difference is our use of Sprague-Dawley rather than Wistar rats, the significance of which is not clear. When using an embolic model, analysis of the entire cerebrum is desirable if not mandatory. Volumetric techniques are well suited to this purpose and have been used to define other models of cerebral ischemia (1, 18), albeit. with a limited number of specimens. Our use of a large number of sections helps to reduce variability due to measurement techniques. Our findings of occipital cortex involvement may be explained on the basis of posterior cerebral artery occlusion, as attested by the observable lateral brain stem and posterior thalamic regions of ischemia. However, studies of selective external occlusion of the middle cerebral artery in the rat (19, 20) have shown that by autoradiographic techniques, the occipital cortex sustained a decrease in blood flow to 20 to 40% of control values, without histological changes. Histologic infarction occurred in the frontal, parietal, and temporal cortex, and lateral neostriatum. Decrements in flow were also seen in other studies (2 1) in the rat in a distribution similar to that we observed.
RAT EMBOLIC
STROKE
MODEL
403
We chose for our experiment a 2-h time course postembolization as a clinically relevant interval which also minimizes spontaneous clot dissolution, better insures animal survival, and entails smaller anesthetic doses. However, complete documentation of our technique requires histological analysis of affected regions. A region which is ischemic for a limited period of time (dependent on the species of animal, overall metabolic conditions, degree of nonperfusion, and anesthetic agent used) may later manifest the histological changes of infarction. Definitive histological delineation of an infarcted area by light microscopy may require a maturation period of at least 4 to 6 h for objectively verifiable changes to take place (2, 12). In addition, changes due to regional ischemia are quite heterogeneous soon after injury: vacuolation of the neuropil may be seen as early as 5 h in boundary zones while the appearance of neuronal cell bodies may vary along a spectrum from normal to pyknotic (5). For area determinations digital processing of autoradiographs can be more comprehensive and objective than visual determination of ischemic neuronal changes. Nevertheless, an additional limitation of our study is that the degree of ischemia corresponding to the LF region was not assessed. Although histologically infarcted tissue is associated with blood flow values of less than 24 ml/ 100 g/min 4 h after middle cerebral occlusion in the rat (2 I), with maximal normal flows of approximately 100 ml/ 100 g/ min, the limit of 20% of maximal autoradiographic density chosen for our experiment was really rather arbitrary because of the indirect relationship between density and flow, and the fact that nonischemic white matter may have a flow less than 20% of peak gray matter flows. The chosen anesthetic agent may affect the size and distribution of the ischemic region. Pentobarbital reduced infarct volume in cats subjected to middle cerebral artery clipping (I), and in certain doses reduced blood flow in cortical regions much more than in deep white matter ( 12). Pentobarbital may prevent infarction in an ischemic region. The cortical hyperemia observed exclusively in the animals with small NF volumes may be due to greater collateral flow (22) or to thrombolysis and distal migration of emboli (9, 14,23). It is notable that the state of the cervical internal carotid artery as observed at the time of killing in our experiment bore no particular relationship to the presence of an ischemic region. The one animal (rat 8) without a significant lesion appeared to have a small amount of blood clot in the proximal middle cerebral artery (with an occluded cervical internal carotid artery), and most likely benefited from significant patency of primary or collateral vessels. The circle of Willis in the rat is usually complete, and common carotid occlusion usually fails to bring about ischemia unless combined with hypoxia (6) or hypotension (3, 16). Under these conditions, similar regions are affected
404
PAUL L. PENAR
(neocortex, hippocampus, and lateral striatum) as we observed, but subcortical involvement is more variable. In summary, we have described a model of cerebral ischemia in the rat which produces large regions of hypoperfusion by means of embolization of homologous blood clot into the middle and posterior cerebral arteries. This model should prove useful in the investigation of a wide range of therapeutic strategies, particularly those involving thrombolytic agents. REFERENCES 1. BLACK, K. L., D. J. WEIDLER, N. S. JALLAD, T. M. SODEMAN, AND G. D. ABRAMS. 1978. Delayed pentobarbital therapy of acute focal cerebral &hernia. Stroke% 245-249. 2. BROWN, A. W., AND J. B. BRIERLY. 1968. The nature, distribution, and earliest stages of anoxic-ischemic nerve cell damage in the rat brain as defined by the optical microscope. Br. .I. Exp. Pathol. 49: 87-106. 3. BUSTO, R., AND M. D. GINSBERG. 1985. Graded focal cerebral ischemia in the rat by unilateral carotid artery occlusion and elevated intracranial pressure: hemodynamic and biochemical characterization. Stroke 16: 466-476. 4. DIAZ, F. G., AND J. I. AUSMAN. 1980. Experimental cerebral &hernia. Neurosurgery 6: 436-445. 5.
GARCIA, J. H., A. S. LOSSINSKY, F. C. KAUFFMAN, AND K. A. CONGER. 1978. Neuronal ischemic injury: light microscopy, ultrastructure, and biochemistry. Acta Neuropathol.
6.
GINSBERG, M. D., R. MEDOFF, AND M. REIVICH. 1976. Heterogeneities of regional cerebral blood flow during hypoxia-ischemia in the rat. Stroke 7: 125- 13 1. HUDGINS, W. R., AND J. H. GARCIA. 1980. Transorbital approach to the middle cerebral artery of the squirrel monkey. Stroke 1: 107- 111. KANEKO, D., N. NAKAMURA, AND T. GGAWA. 1985. Cerebral infarction in rats using homologous blood emboli: development of a new experimental model. Stroke 16: 76-84. KANGSTROM, E., M. L. SMITH, AND B. K. SIESIO. 1983. Recirculation in the rat brain following incomplete &hernia. J. Cereb. Blood Flow Metab. 3: 183- 192. KOENIG,J. F., AND R. A. KLIPPEL. 1963. The Rat Brain: A Stereotaxic Atlas ofthe Forebrain and Lower Parts of the Brain Stem. The Williams & Wilkins Co., Baltimore, Maryland. KUDO, M., A. AOYAMA, S. ICHIMORI, AND N. FUKUNAGA. 1982. An animal model of cerebral infarction: homologous blood clot emboli in rats. Stroke 13: 505-508. LAURENT, J. P., P. LAWNER, F. A. SIMEONE, AND E. FINK. 1982. Pentobarbital changes compartmental contribution to cerebral blood flow. J. Neurosurg. 56: 504-5 10. LYFORD, C. L., W. E. CONNOR, J. C. HOAK, AND E. D. WARNER. 1967. The coagulant and thrombogenic properties of human atheroma. Circulation 36: 284-293. OLSEN, T. S., AND N. A. LASSEN. 1984. A dynamic concept ofmiddle cerebral artery occlusion and infarction in the acute state based on interpreting severe hyperemia as a sign of embolic migration. Stroke 15: 458-468. SAKURADA, O., C. KENNEDY, J. JEHLE, J. D. BROWN, G. L. CARBIN, AND L. SOKOLOFF. 1978. Measurement of local cerebral blood flow with iodo[ “C]antipyrine. Am. J. Physiol. 234: H59-H66. SMITH, M., G. BENEDEK, N. DAHLGREN, I. ROSEN, T. WIELOCH, AND B. K. SIESJO.1984. Models for studying long-term recovery following forebrain &hernia in the rat. A 2vessel occlusion model. Acta Neural. Stand. 69: 385-40 I.
43: 85-95.
7. 8. 9. 10. 11. 12. 13. 14. 15. 16.
RAT
EMBOLIC
STROKE
MODEL
405
17. SNYDER, H. E., AND J. L. SHAPIRO. 196 I. A correlative study of atheromatous embolism in human beings and experimental animals. Surgery 49: 195-204. 18. SUNDT, T. M., JR., AND A. G. WALTZ. 1966. Experimental cerebral infarction: retro-orbital, extradural approach for occluding the middle cerebral artery. Pm. Mayo Clin. 41: 159-168. 19. TAMURA, A., D. I. GRAHAM, J. MCCULLOCH, AND G. M. TEASDALE. 198 1. Focal cerebral ischemia in the rat. 2. Regional cerebral blood flow determined by [‘4C]iodoantipyrine autoradiography following middle cerebral artery occlusion. J. Cereb. Blood Flow Metab. 1: 61-69. 20. TAMURA, A., D. I. GRAHAM, J. MCCULLOCH, AND G. M. TEASDALE. 198 1. Focal cerebral ischemia in the rat. 1. Description of technique and early neuropathological consequences following middle cerebral artery occlusion. J. Cereb. Blood Flow Metab. 1: 5360. 2 1. TYSON, G. W., G. M. TEASDALE, D. I. GRAHAM, AND J. MCCULLOCH. 1984. Focal cerebral ischemia in the rat: topography of hemodynamic and histopathological changes. Ann. Neural. 15: 559-567. 22. VANDER EECKEN, H. M., AND R. D. ADAMS. 1953. The anatomy and functional significance of the meningeal arterial anastomoses of the human brain. J. Neuropathol. Exp. Neurol. 12: 132-157. 23. WALTZ, A. G., AND T. M. SUNDT. 1967. The microvasculature and microcirculation of the cerebral cortex after arterial occlusion. Brain 90: 68 l-696. 24. YAMORI, Y., T. HORIE, H. HANDA, M. SATO, AND M. FUKASE. 1967. Pathogenic similarity of strokes in stroke-prone spontaneously hypertensive rats and humans. Stroke 7: 46-53.
NEUROLOGY
96,393-405
(1987)
A Model of Embolic Cerebral
lschemia in the Rat
PAUL L. PENAR’ Section
OfNeurosurgery,
Received
Yale University
August
School
2 7, 1986; revision
ofMedicine,
received
New Haven,
December
Connecticut
06510
9, 1986
Experimental studies of embolic cerebral &hernia using the rat are limited by variability in the location, size, and frequency of lesions produced. A technique is described herein which improves the reliability of an established model. Eight male Sprague-Dawley rats underwent injection of the cervical internal carotid artery with 0.1 ml of 1-h-old fragmented autologous blood clot through an external carotid artery cannula. The pterygopalatine artery was ligated prior to embolization. At killing 2 h after embolization, clot was observed in the proximal middle cerebral and posterior cerebral arteries in all animals. Areas of reduced blood flow at 2 h postembolization were assessedby digital image processing of iodo-[‘4C]antipyrine autoradiographic images. No-flow and low-flow areas were measured for each of approximately 25 serial brain sections with a computerized bit-pad. Volumes were calculated and lesions localized by anatomical reconstructions. No animal sustained a hemorrhagic lesion. One animal sustained only a very small area of &hernia in the internal capsule. Of the remaining seven, all had large regions of ischemia in the middle cerebral distribution involving cortex and basal ganglia. Posterior cerebral involvement was observed in six of the seven animals as well. The contralateral hemisphere was unaffected. Volume values could be calculated for primary vascular distributions. Most variability occurred in the pattern of posterior cerebral involvement. The technique described produces a relatively consistent region of &hernia in the middle and posterior cerebral artery distributions in the rat and is a useful model for the study of cerebral &hernia. 0 1987 Academic Press, Inc.
Abbreviations: BW-body weight; ECA-external carotid artery; LF, NF-low flow, no flow. ’ This research was supported in part by grants NS- 10 174 and NS- 19430 from the National Institutes of Health. Presented in poster format at the American Association of Neurological Surgeons Annual Meeting, April 13 to 17, 1986 at Denver, Colorado. The author thanks Dr. Carole C. LaMotte for the Quandens imaging system, and acknowledges the assistance of Dr. Charles Greer.
393 0014-4886/87 $3.00 Copyright 0 1987 by Academic Press, Inc. All rights of reproduction in any form reserved.
394
PAUL L. PENAR
INTRODUCTION Cerebral embolization with homologous blood clot in the rat has been shown to be a practical and inexpensive model of focal cerebral ischemia (11). Recently, in modification of Kudo’s procedure (1 l), Kaneko et al. (8) described embolization into the internal carotid artery of the rat with the pterygopalatine artery ligated. Variable results occurred, including bilateral middle cerebral artery occlusions. Distal migration of emboli was seen after initial proximal occlusion of an involved vessel. In an effort to improve the reproducibility of the embolic model of cerebral ischemia in the rat, we introduced a modification ofthe procedure previously described. In particular, clot preparation and volume of injection were altered. We then undertook to define the volume and distribution of regions in the rat brain and mesencephalon affected by internal carotid embolization with this preparation of homologous blood clot, as defined by iodo[‘4C]antipyrine autoradiographic techniques. METHODS
Experimental Procedure. Eight male Sprague-Dawley rats weighing 3 14 to 459 g were anesthetized lightly with ether and then received i.p. injections of 40 mg/kg sodium pentobarbital (Nembutal, 50 mg/ml, Abbott Laboratories). A tracheostomy was carried out, followed by insertion of right femoral arterial and venous polyethylene catheters. Arterial pressure was continually recorded via transducer and chart recorder (Grass Instrument Co.). Temperature was monitored by rectal probe and maintained at approximately 37°C. Anesthesia was maintained with intermittent intravenous and intraperitoneal doses of pentobarbital. Arterial blood gasses were measured by loo-p1 capillary tube sampling prior to embolization and just prior to killing. With the aid of a Zeiss operating microscope and electrocautery, the right digastric and stemohyoid muscles were transected, and the wing of the hyoid bone removed. The external maxillary artery, superior thyroid artery, and occipital artery (overlying the carotid bifurcation) were coagulated and severed. Without dissection of the carotid bifurcation, the right pterygopalatine artery was ligated with 4-O silk suture and coagulated. The external carotid artery (ECA) was ligated distal to the bifurcation and cannulated with polyethylene catheter (PE-50, Clay Adams; i.d. 0.58 mm, o.d. 0.965 mm). The catheter was advanced to the carotid bifurcation and secured in place with a ligature and tissue adhesive at its point of entry into the ECA. The blood clot was prepared in the following fashion: 0.6 to 0.8 ml blood was removed from the animal 60 to 90 min prior to embolization. Serum
RAT EMBOLIC
STROKE
MODEL
395
was nearly completely extruded from the syringe in which the blood had clotted, and the remaining material was passed once through a 20-gauge needle and then through a 22-gauge needle. The clot (with a volume of approximately 0.3 to 0.4 cc) was loaded into a l-ml syringe attached to a 22-gauge needle and connected to the ECA catheter. Air bubbles were eliminated by careful insertion of the needle into the catheter. Microscopic examination of sample clot preparations showed that fragments of clot ranged from 0.2 to 2.0 mm in length, the majority being less than 0.4 mm. With a temporary aneurysm clip occluding the common carotid artery at the base of the neck, 0.1 cc of clot was injected by hand into the external carotid catheter during 2 min. If all this clot volume could not be injected (due to resistance), the injection was considered to be inadequate. The common carotid clip was removed and the catheter withdrawn. The ECA was coagulated distally. Prior to killing at 120 min postembolization, 0.9 ml blood was withdrawn for chemical assay. A dose of 240 &/kg of iodo-4-[Nmethyl’4C]antipyrine in saline (100 &i/ml, American Radiolabeled Chemicals Inc., St. Louis, Missouri) was injected i.v. followed 10 s later by 2 ml saturated KC1 solution. This short delay allowed for distribution of the tracer along major vascular channels, reducing the significance of tracer distribution through collateral pathways. Preparation of Tissue. The brain and brain stem were removed within 10 min of death and frozen in isopentane cooled to approximately -70°C in dry ice. The carotid system in the neck was removed and placed in Formalin. Sectioning of the brain was carried out within 20 h of freezing on a cryostat (Damon IEC) in 32-pm sections after mounting the specimen with EM 1 embedding matrix (Lipshaw Manufacturing Corp., Detroit, Michigan). Two sections were saved on a glass coverslip after discarding 20 sections. Starting from the olfactory bulbs, sections were taken until occipital pole tissue was no longer visible, which included the mesencephalon to the level of the superior colliculus but below the interpeduncular nucleus. The specimens were then dried at 70°C for approximately 30 min, mounted on cardboard, and incubated 6 days with Kodak SB-5 film. Histologic sections were prepared with cresyl violet stains of mounted specimens. Volumetric Analysis. Of the two sections saved, the higher quality specimen was chosen for photographic enlargement to 100 times original size (Durst Laborator S-45 EM) as a negative image on PI-1 Rapitone paper. A frequency analysis of the distribution of optical intensities from a representative non-specimen-containing portion of each individual autoradiographic plate was obtained with the Quandens image analysis system (FHC Inc., Brunswick, Maine). Local tissue concentrations of the carbon- 14 isotope are roughly proportional to autoradiographic density. Blood flow within the
396
PAUL L. PENAR
FIG. 1. Sample autoradiographic images (left) and corresponding processed images (right) demonstrating region of hypoperfusion. Dark band represents range of densities from greaterthan-background to 20% of maximum density.
specimen is related to 14C concentration by means of a polynomial relationship which takes into account the partition coefficient ofthe tracer (15). A value was established for background intensity as the minimum background intensity extrapolated from the large background peak. This value was used to delineate regions of absent perfusion in the specimen. Thus density and not actual blood flow was used to define regions of “no flow.” The optical intensity corresponding to 20% of the maximum density (in the maximumto-background density range) was calculated to obtain an additional boundary in the specimen. This value was less sensitive to changes in maximum density values (maximal blood flows). All autoradiographic images were then reviewed using these values to define areas of nonlabeling and of reduced labeling with tracer (termed “noflow” and “low-flow” regions, respectively, see Fig. l), which were then delineated on the photographic enlargements. The areas of the left and right halves of the hemispheres, brain stem, and diencephalon were measured for
RAT EMBOLIC
STROKE
397
MODEL
TABLE 1 Mean Arterial Pressure and Blood Gas Values MAP” Preembolization Prekilling
134* 7 131+18
PO2
80.7 + 10 88.6 + 16
PC02
PH
39.6 f 3.7 36.4 f 4.9
7.390 f 0.076 7.412 f 0.052
a MAP = mean arterial blood pressure. MAP, PO,, and PCOZreported as mean + SD mm Hg.
each image using a microcomputer bit pad (Summagraphics Bit Pad One) interfaced with a PDP- 11 computer, as were the areas designated as no-flow and low-flow regions. Each measurement of area was repeated and the results averaged for that section. Correction for magnification in the process of enlargement was accomplished with measurements of an enlarged 10 X 10 mm standard. The volumes were then calculated using the average areas of the two contiguous saved sections to represent the entire sequence of 20 discarded slices; the more posterior sections were added into the total for the sequence. The most anterior and most posterior discarded sections were represented by the nearest saved section. Corrections were made for lost sections or irregularities in the cutting process. RESULTS Physiologic Measures, Anesthesia, and Fluids. Animal weight ranged from 314 to 495 g (405 + 53, mean + SD). Table 1 lists values for mean arterial pressure, arterial PO* and Pco~, and pH preembolization and prekilling. Animals received an average of 0.06 + 0.01 mg pentobarbitai sodium/g body weight (BW) during the course of the experiment, and 0.02 + 0.009 ml saline/g BW. Neither the barbiturate dose or fluid dose nor the arterial pressure or blood gas values were significantly correlated with no-flow (NF) or low-flow (LF) volumes. Observations. The cervical internal carotid arteries of animals 1,2,3,4,5, and 8 appeared by visual inspection to contain significant clot at the time of killing, with dark blue discoloration and weak or absent pulsation. The common carotid arteries of animals 2 and 3 had a similar appearance. The cervical carotid systems of rats 6 and 7, in contrast, appeared normal in pulsation and coloration throughout the experiment. With the sole exception of rat 8, all animals harbored blood clot in the internal carotid, proximal middle cerebral, and proximal posterior cerebral vessels at the base of the brain.
398
PAUL L. PENAR TABLE 2 Brain Values, “No Flow” (NF) and “Low Flow” (LF)” Volume (mm3)
Rat No.
Total brain
Right brain
NF
LF
NF%of r. brain
LF%of r. brain
1 2 3 4 5 6 7 -
1800 1655 1823 1813 1692 1628 1618
921 922 898 906 895 908 904
342 216 476 421 118 286 178
569 389 657 589 265 472 331
37.2 23.5 53.0 46.5 13.2 31.5 19.7
61.8 42.2 73.2 65.0 29.6 51.9 36.6
sz,
1718 91
908 10
291 131
467 145
32.1 14.5
51.5 16.1
’ Raw values for total brain volume are given, followed by volumes corrected for total brain volume.
Animal 8 demonstrated only a small proximal middle cerebral clot. The brains of rats 2,5, and 7 showed hemispheric hyperemia on gross inspection prior to freezing. V&metric Analysis. Table 2 shows the volume values measured for seven animals, corrected for a mean total brain volume of 17 18 f 9 1 mm3. Animal 8 had only a very small region of low blood flow in the region of the internal capsule; pentobarbital dose (0.07 g/g BW), saline dose (0.02 ml/g BW), mean arterial pressures, and blood gas values were not noticeably different for this animal. Mean NF volume was 29 1 + 13 1 mm3 for the remaining seven animals; LF volume was 467 + 145 mm3. NF volume occupied 32 f 14.5% of the right brain. Low-flow volumes were comparatively less variable than no-flow volumes. Figure 2 demonstrates a reconstruction based on the atlas of Kiinig and Klippel(l0) of the anatomic regions affected for rat 6, an animal with NF and LF volumes of approximately average size. The middle cerebral artery distribution was most involved. The head of the caudate, globus pallidus, and putamen were in no-flow zones, as was much of the adjoining cortex. The hippocampus was nearly completely involved, except its most superior extent. The occipital cortex and lateral brain stem, both in the posterior cerebral artery distribution, were also affected. The anterior cerebral distribution was spared. The nucleus accumbens was unaffected, as were the more medial
RAT
EMBOLIC
STROKE
MODEL
399
400
PAUL
L. PENAR
thalamic nuclei. In the case of animal 3 (with the largest volume affected) a similar pattern of involvement was seen, but with a larger amount of NF region in the posterior thalamus, occipital cortex, and lateral brain stem (posterior cerebral distribution). By contrast, three animals (2, 5, and 7) showed relatively smaller regions of involvement. In this group of animals (all of which had observable hyperemia of the cortex), the anterior cerebral distribution was for the most part as equally unaffected as in the group with larger NF volumes. There was greater sparing of the medial portions of the caudate nucleus, lenticular nuclei, and lateral thalamus, together with relative sparing in the posterior cerebral circulation. The low-flow region surrounding the no-flow region in the middle cerebral distribution was broader. The hippocampus was largely unaffected in rat 5; the brain of rat 2 did show involvement of the hippocampus and lateral thalamic nuclei. The regions of low flow around areas of no flow were larger in the three animals with smaller NF volumes, relative to the total region affected (LF volume). The NF:LF ratios of animals 2,5, and 7 were all less than 0.60; NF: LF ratios for the remaining four animals ranged from 0.60 to 0.72. Sequential histogram plots of right brain volume, NF volume, and LF volume were constructed for each animal based on identification within autoradiographic images of specific nuclear groups and major fiber tracts (Fig. 3). These plots illustrate a consistent involvement in the distribution of the middle cerebral artery, with variable involvement of structures supplied by the posterior cerebral artery. Volume sections 10 to 16 (just caudal to the head of the caudate, at the genu of the corpus callosum, through to the midthalamus) are more specific for middle cerebral distribution. For these sections, NF volume averaged 124 f 52 mm3 for the seven animals; LF volume mean was 190 + 44 mm3. NF volume occupied 40 f 15% of a mean right brain volume of 3 10 f 20 mm3 for these sections. LF volume comprised 6 1 + 12% of right brain volume. The maximal NF area in a given section for each animal ranged from 35 to 7 1% and corresponded well with the trend in total NF and LF volumes. These maximal points were located from sections 15 to 19 in the seven animals, a region supplied by both middle and posterior cerebral arteries. DISCUSSION Experimental cerebral &hernia produced by vascular occlusion is typically accomplished by embolization techniques or by external direct vascular occlusion. Although the lesions produced by embolization have not been quite as consistent or reproducible as those produced by direct vascular oc-
RAT EMBOLIC
STROKE
MODEL
IO 7Or
IX
50 30 IO
Ld Rostrol
CoudaI
Rostra1
Caudal
Right brain volume Low flow dume Noflow volume FIG. 3. Sequential histogram plots of involvement for individual volume sections from rostra1 to caudal. Maximal NF volumes were located from sections 15 to 19, supplied by both middle and posterior cerebral arteries. Anatomic figure from atlas of Kijnig and Klippel, used with permission.
402
PAUL
L. PENAR
elusion, embolization involves no direct invasion of the cranial vault and therefore carries less risk of alteration of the local response to injury, changes in blood-brain barrier permeability, and direct cortical damage (4,7). Internal carotid artery embolization in the rat has several advantages as a model of cerebral ischemia. This inexpensive animal possessesa cerebral circulation similar to that of man (24), and an identical cervical carotid system but for the pterygopalatine artery, a large branch of the internal carotid artery. Furthermore, the use of homologous clot provides the opportunity to investigate a variety of clinically relevant situations, including the effects of thrombolytic substances following acute arterial occlusion by thrombus or a thrombogenic substance ( 13, 17). In spite of the small number of animals studied in this preliminary investigation, we have shown that a reproducible region of severe hypoperfusion can be produced in the rat brain ipsilaterally by embolization into the internal carotid artery with a fixed amount of homologous blood clot. This region of ischemia lies within the middle and posterior cerebral distributions, and results from proximal occlusion of these vessels at the base of the brain. The greatest variability occurred in the degree of posterior cerebral involvement. The principal differences between our experimental technique and that of Kaneko et al. (8) is in the preparation of the clot used for embolization: fragmentation of clot through a 27-gauge needle followed by dilution with 0.3 ml saline resulted in a clot size of 100 to 200 pm in a total volume of 0.4 ml (8). Our technique resulted in much larger clot fragments (to 2 mm in length) which produced no detectable contralateral regions of reduced perfusion, and more consistent proximal vessel occlusion. A further difference is our use of Sprague-Dawley rather than Wistar rats, the significance of which is not clear. When using an embolic model, analysis of the entire cerebrum is desirable if not mandatory. Volumetric techniques are well suited to this purpose and have been used to define other models of cerebral ischemia (1, 18), albeit. with a limited number of specimens. Our use of a large number of sections helps to reduce variability due to measurement techniques. Our findings of occipital cortex involvement may be explained on the basis of posterior cerebral artery occlusion, as attested by the observable lateral brain stem and posterior thalamic regions of ischemia. However, studies of selective external occlusion of the middle cerebral artery in the rat (19, 20) have shown that by autoradiographic techniques, the occipital cortex sustained a decrease in blood flow to 20 to 40% of control values, without histological changes. Histologic infarction occurred in the frontal, parietal, and temporal cortex, and lateral neostriatum. Decrements in flow were also seen in other studies (2 1) in the rat in a distribution similar to that we observed.
RAT EMBOLIC
STROKE
MODEL
403
We chose for our experiment a 2-h time course postembolization as a clinically relevant interval which also minimizes spontaneous clot dissolution, better insures animal survival, and entails smaller anesthetic doses. However, complete documentation of our technique requires histological analysis of affected regions. A region which is ischemic for a limited period of time (dependent on the species of animal, overall metabolic conditions, degree of nonperfusion, and anesthetic agent used) may later manifest the histological changes of infarction. Definitive histological delineation of an infarcted area by light microscopy may require a maturation period of at least 4 to 6 h for objectively verifiable changes to take place (2, 12). In addition, changes due to regional ischemia are quite heterogeneous soon after injury: vacuolation of the neuropil may be seen as early as 5 h in boundary zones while the appearance of neuronal cell bodies may vary along a spectrum from normal to pyknotic (5). For area determinations digital processing of autoradiographs can be more comprehensive and objective than visual determination of ischemic neuronal changes. Nevertheless, an additional limitation of our study is that the degree of ischemia corresponding to the LF region was not assessed. Although histologically infarcted tissue is associated with blood flow values of less than 24 ml/ 100 g/min 4 h after middle cerebral occlusion in the rat (2 I), with maximal normal flows of approximately 100 ml/ 100 g/ min, the limit of 20% of maximal autoradiographic density chosen for our experiment was really rather arbitrary because of the indirect relationship between density and flow, and the fact that nonischemic white matter may have a flow less than 20% of peak gray matter flows. The chosen anesthetic agent may affect the size and distribution of the ischemic region. Pentobarbital reduced infarct volume in cats subjected to middle cerebral artery clipping (I), and in certain doses reduced blood flow in cortical regions much more than in deep white matter ( 12). Pentobarbital may prevent infarction in an ischemic region. The cortical hyperemia observed exclusively in the animals with small NF volumes may be due to greater collateral flow (22) or to thrombolysis and distal migration of emboli (9, 14,23). It is notable that the state of the cervical internal carotid artery as observed at the time of killing in our experiment bore no particular relationship to the presence of an ischemic region. The one animal (rat 8) without a significant lesion appeared to have a small amount of blood clot in the proximal middle cerebral artery (with an occluded cervical internal carotid artery), and most likely benefited from significant patency of primary or collateral vessels. The circle of Willis in the rat is usually complete, and common carotid occlusion usually fails to bring about ischemia unless combined with hypoxia (6) or hypotension (3, 16). Under these conditions, similar regions are affected
404
PAUL L. PENAR
(neocortex, hippocampus, and lateral striatum) as we observed, but subcortical involvement is more variable. In summary, we have described a model of cerebral ischemia in the rat which produces large regions of hypoperfusion by means of embolization of homologous blood clot into the middle and posterior cerebral arteries. This model should prove useful in the investigation of a wide range of therapeutic strategies, particularly those involving thrombolytic agents. REFERENCES 1. BLACK, K. L., D. J. WEIDLER, N. S. JALLAD, T. M. SODEMAN, AND G. D. ABRAMS. 1978. Delayed pentobarbital therapy of acute focal cerebral &hernia. Stroke% 245-249. 2. BROWN, A. W., AND J. B. BRIERLY. 1968. The nature, distribution, and earliest stages of anoxic-ischemic nerve cell damage in the rat brain as defined by the optical microscope. Br. .I. Exp. Pathol. 49: 87-106. 3. BUSTO, R., AND M. D. GINSBERG. 1985. Graded focal cerebral ischemia in the rat by unilateral carotid artery occlusion and elevated intracranial pressure: hemodynamic and biochemical characterization. Stroke 16: 466-476. 4. DIAZ, F. G., AND J. I. AUSMAN. 1980. Experimental cerebral &hernia. Neurosurgery 6: 436-445. 5.
GARCIA, J. H., A. S. LOSSINSKY, F. C. KAUFFMAN, AND K. A. CONGER. 1978. Neuronal ischemic injury: light microscopy, ultrastructure, and biochemistry. Acta Neuropathol.
6.
GINSBERG, M. D., R. MEDOFF, AND M. REIVICH. 1976. Heterogeneities of regional cerebral blood flow during hypoxia-ischemia in the rat. Stroke 7: 125- 13 1. HUDGINS, W. R., AND J. H. GARCIA. 1980. Transorbital approach to the middle cerebral artery of the squirrel monkey. Stroke 1: 107- 111. KANEKO, D., N. NAKAMURA, AND T. GGAWA. 1985. Cerebral infarction in rats using homologous blood emboli: development of a new experimental model. Stroke 16: 76-84. KANGSTROM, E., M. L. SMITH, AND B. K. SIESIO. 1983. Recirculation in the rat brain following incomplete &hernia. J. Cereb. Blood Flow Metab. 3: 183- 192. KOENIG,J. F., AND R. A. KLIPPEL. 1963. The Rat Brain: A Stereotaxic Atlas ofthe Forebrain and Lower Parts of the Brain Stem. The Williams & Wilkins Co., Baltimore, Maryland. KUDO, M., A. AOYAMA, S. ICHIMORI, AND N. FUKUNAGA. 1982. An animal model of cerebral infarction: homologous blood clot emboli in rats. Stroke 13: 505-508. LAURENT, J. P., P. LAWNER, F. A. SIMEONE, AND E. FINK. 1982. Pentobarbital changes compartmental contribution to cerebral blood flow. J. Neurosurg. 56: 504-5 10. LYFORD, C. L., W. E. CONNOR, J. C. HOAK, AND E. D. WARNER. 1967. The coagulant and thrombogenic properties of human atheroma. Circulation 36: 284-293. OLSEN, T. S., AND N. A. LASSEN. 1984. A dynamic concept ofmiddle cerebral artery occlusion and infarction in the acute state based on interpreting severe hyperemia as a sign of embolic migration. Stroke 15: 458-468. SAKURADA, O., C. KENNEDY, J. JEHLE, J. D. BROWN, G. L. CARBIN, AND L. SOKOLOFF. 1978. Measurement of local cerebral blood flow with iodo[ “C]antipyrine. Am. J. Physiol. 234: H59-H66. SMITH, M., G. BENEDEK, N. DAHLGREN, I. ROSEN, T. WIELOCH, AND B. K. SIESJO.1984. Models for studying long-term recovery following forebrain &hernia in the rat. A 2vessel occlusion model. Acta Neural. Stand. 69: 385-40 I.
43: 85-95.
7. 8. 9. 10. 11. 12. 13. 14. 15. 16.
RAT
EMBOLIC
STROKE
MODEL
405
17. SNYDER, H. E., AND J. L. SHAPIRO. 196 I. A correlative study of atheromatous embolism in human beings and experimental animals. Surgery 49: 195-204. 18. SUNDT, T. M., JR., AND A. G. WALTZ. 1966. Experimental cerebral infarction: retro-orbital, extradural approach for occluding the middle cerebral artery. Pm. Mayo Clin. 41: 159-168. 19. TAMURA, A., D. I. GRAHAM, J. MCCULLOCH, AND G. M. TEASDALE. 198 1. Focal cerebral ischemia in the rat. 2. Regional cerebral blood flow determined by [‘4C]iodoantipyrine autoradiography following middle cerebral artery occlusion. J. Cereb. Blood Flow Metab. 1: 61-69. 20. TAMURA, A., D. I. GRAHAM, J. MCCULLOCH, AND G. M. TEASDALE. 198 1. Focal cerebral ischemia in the rat. 1. Description of technique and early neuropathological consequences following middle cerebral artery occlusion. J. Cereb. Blood Flow Metab. 1: 5360. 2 1. TYSON, G. W., G. M. TEASDALE, D. I. GRAHAM, AND J. MCCULLOCH. 1984. Focal cerebral ischemia in the rat: topography of hemodynamic and histopathological changes. Ann. Neural. 15: 559-567. 22. VANDER EECKEN, H. M., AND R. D. ADAMS. 1953. The anatomy and functional significance of the meningeal arterial anastomoses of the human brain. J. Neuropathol. Exp. Neurol. 12: 132-157. 23. WALTZ, A. G., AND T. M. SUNDT. 1967. The microvasculature and microcirculation of the cerebral cortex after arterial occlusion. Brain 90: 68 l-696. 24. YAMORI, Y., T. HORIE, H. HANDA, M. SATO, AND M. FUKASE. 1967. Pathogenic similarity of strokes in stroke-prone spontaneously hypertensive rats and humans. Stroke 7: 46-53.