Cerebral Aneurysms Arising at Nonbranching Sites
An Experimental Study
Background and Purpose The origin and pathogenesis of cerebral aneurysms arising at nonbranching sites are not clear. Using our animal model to induce cerebral aneurysms in rats, we examined induced aneurysms that developed at nonbranching sites.
Methods In 35 Sprague-Dawley rats, the left common carotid artery was ligated and renal hypertension was produced to induce cerebral aneurysms. Twelve months later, the circle of Willis was carefully examined under a dissecting microscope.
Results Other than cerebral aneurysms at branching sites of the circle of Willis, aneurysmal bulges developing at nonbranching sites were found in the proximal portion of the posterior cerebral artery (P1) on the side of carotid ligation, which supposedly acted as a major collateral pathway after the ligation, in 19 of 35 treated rats. A total of 30 lesions were found in these 19 rats, and they were classified into fusiform aneurysms (22 lesions) involving the entire vessel wall for a short distance and saccular aneurysms (8 lesions) involving only a part of the wall and expanding laterally from the vessel wall. These P1s became larger in caliber and more tortuous after ligation. Aneurysms developed more frequently in proportion to these changes in these vessels. Moreover, most aneurysms in these vessels developed at or near the curvatures. All of the lateral aneurysms were found on the lateral wall of the curvatures of the vessels.
Conclusions The present findings indicate that cerebral aneurysms at nonbranching sites and saccular aneurysms at branching sites can occur under the same etiologic conditions. The site of origin is strongly related to hemodynamic stress.
Saccular cerebral aneurysms have arisen at the apex of arterial bifurcations and have been considered a unique aspect of cerebral aneurysms. However, with recent advances in microsurgery and microanatomy, increased numbers of cerebral aneurysms, which originate without apparent relation to arterial bifurcation, have been recognized. Some are considered to be related to tiny branches such as the superior hypophysial artery, but there is no evidence to relate others to an arterial branch.1 2 3 4 5 Although such aneurysms are now well known in the clinical field, the origin and pathogenesis are not well elucidated. Intracranial fusiform aneurysms are considered to be atherosclerotic in origin.6 However, it is still not clear whether severe atherosclerotic changes in such fusiform aneurysms are the cause of the dilatation or the result of the dilatation.
We developed experimental animal models to induce cerebral aneurysms not only in rats7 but also in monkeys.8 Although aneurysms have been induced at nonbranching sites as well as branching sites, most studies have focused on aneurysms at branching sites.7 8 9 10 11 12 In this report we examine the development of experimentally induced cerebral aneurysms arising at nonbranching sites in rats.
Materials and Methods
Procedures for Induction of Aneurysms
In 35 5-week-old Sprague-Dawley rats (weight, 200 to 300 g), the left common carotid artery and posterior branches of both renal arteries were ligated to induce cerebral aneurysms. These procedures were performed in rats under intraperitoneal sodium pentobarbital anesthesia (40 mg/kg) with additional injections when necessary. One week after the operation, 1% saline was substituted for the drinking water to enhance the degree of hypertension. As a control group, 25 age-matched, untreated male rats were used.
Systolic blood pressure was measured by the tail cuff plethysmographic method in all rats just before the operation. All the animals were killed with an overdose of sodium pentobarbital anesthesia 12 months after the start of the experiment. A catheter was placed in the abdominal aorta, and the animals were perfused with heparinized saline, followed by a solution of 2.5% glutaraldehyde in 0.1 mol/L phosphate buffer (pH 7.4) or Zamboni solution (2% paraformaldehyde containing picric acid in 0.15 mol/L phosphate-buffered saline, pH 7.3). After perfusion fixation, the circle of Willis was carefully removed from the brain of each rat under a dissecting microscope.
Macroscopic Study of the P1 Segments
The proximal portion of the posterior cerebral artery (P1) on the side of carotid ligation, which acts as the major collateral pathway after the ligation, and P1 of the nonligated side were examined in detail. The shape of an aneurysm was classified into two types by one of the authors (S.K.). One was a fusiform aneurysm, which is a dilatation involving the entire vessel wall of a parent artery (Fig 1⇓). The other was a lateral aneurysm, which is a saccular dilatation involving only a part of the wall (Fig 2⇓).
Dilatation and tortuosity of the P1 segments were also examined. Two of us (N.H. and I.N.) observed 50 P1 segments from 25 control rats to establish the criteria. No P1 segments from the control rats were 1.5 times larger in diameter than the contralateral P1 segments. The existence of more than three sharp curvatures (<45°) or a loop was significantly rare in a P1 segment of control rats (observed in only three P1s; P<.001, χ2 test). The P1 segment was judged to be dilated significantly when its diameter was at least 1.5 times larger than that of the contralateral P1. The degree of tortuosity was divided into two categories according to the number of sharp curvatures (<45°) and the existence of looping in vessels: “normal or mildly tortuous” when the number of curvatures was two or less than two (Fig 3a⇓) and “significantly tortuous” when at least three sharp curvatures or a minimum of one loop existed (Fig 3b⇓). All of the P1 segments were studied by two of us (N.H. and I.N.).
Induced aneurysms were excised for histological examination and immersed in the same fixatives as those for perfusion fixation at 4°C for 12 hours. The specimens were then washed, post-fixed in 1% osmium tetroxide in phosphate buffer (7.4) for 1 hour, dehydrated in a graded ethanol series, and embedded in acrylic resin. Semithin sections (1 μm) were stained with toluidine blue for the light microscopic study.
Specimens fixed with Zamboni solution were embedded in paraffin after dehydration. They were cut 3 μm in thickness and stained with elastica van Gieson stain.
The difference in blood pressure just before operation for induction of aneurysms and just before the animals were killed was estimated by the Wilcoxon signed rank test. The χ2 test was used to investigate whether aneurysms at nonbranching sites were induced with or without any relation to the curvature of the vessels and the degree of tortuosity of the P1 segments of control rats. All other statistical analyses were performed with the Mann-Whitney U test or Kruskal-Wallis test. A value of P<.05 was considered indicative of a statistically significant difference.
A summary of the study results is presented in the Table⇓.
Systolic Blood Pressure
The systolic blood pressure of the rats was 108.2±12.3 mm Hg (mean±SD) just before operation and 176.4±20.5 mm Hg just before they were killed. The difference between them was significant (P<.001, Wilcoxon signed rank test).
Types and Locations of Induced Aneurysms
No aneurysms were found in 25 rats in the control group. In 30 of the 35 experimental rats, a total of 55 saccular aneurysms were induced at the arterial bifurcations, mainly at the bifurcations of the anterior cerebral artery and the olfactory artery on the nonligated side.
In 19 of the 35 experimental rats, 30 aneurysms arising at nonbranching sites were induced in the P1 segments on the side of carotid ligation. Twenty-two were fusiform aneurysms involving the entire vessel wall for a short distance. In large fusiform aneurysms, the afferent and efferent vessels appeared to originate from a globular lesion (Fig 1⇑). Another 8 aneurysms were lateral aneurysms that were saccular dilatations from the wall of the artery (Fig 2⇑). The appearance of aneurysmal walls was various; some of them were fairly thick and others appeared to be very thin.
Fifteen aneurysms originated from branching sites of tiny arteries from the left P1. These were excluded from aneurysms at nonbranching sites. No aneurysms were found in the contralateral P1 segments of any rats.
In 16 of the 19 rats developing aneurysms at nonbranching sites in the left P1, saccular aneurysms arising at bifurcations also developed concomitantly (84.2%). In the other 3 rats, only aneurysms at nonbranching sites were induced.
Dilatation and Tortuosity of Proximal Portion of Posterior Cerebral Artery
The left P1 was significantly dilated or tortuous in 31 of the 35 experimental rats (88.6%). In 21 rats it showed both significant dilatation and tortuosity (60.0%), in 4 it was neither significantly dilated nor tortuous, and in 6 (17.1%) it was significantly tortuous. These findings indicated that the left P1s of experimental rats were significantly tortuous and dilated compared with the right P1s (P<.001, Mann-Whitney U test).
Correlation Between Development of Aneurysms and Dilatation and/or Tortuosity of Vessel
All 30 aneurysms at nonbranching sites in the left P1 were found in significantly tortuous or dilated vessels. No aneurysms were found in P1s that were neither tortuous nor dilated. There was a significant correlation between development of aneurysms at nonbranching sites and dilatation and/or tortuosity of the left P1 in experimental rats (P<.001, Mann-Whitney U test).
Correlation Between Development of Aneurysms and Curvatures in Proximal Portion of Left Posterior Cerebral Artery
All of the 8 lateral aneurysms developed at or near a curvature, as shown in Fig 2⇑. Among them, 5 were on top of the curvature, 2 were proximal, and the other was distal. However, there were no significant differences between the incidence of each type of aneurysm (P=.197, χ2 test). All of these lateral aneurysms developed at the lateral side of a curvature. No aneurysms developing at the medial side of a curvature were found.
With respect to 22 fusiform aneurysms, 19 aneurysms developed at or near a curvature (10 on top of the curvature, 2 proximal, and 7 distal). Only 3 aneurysms arose on portions with no relation to a curvature. There were no significant differences between the incidence of each type of aneurysm (P=.059, χ2 test).
These aneurysms had a significant tendency to be induced at or near the vessel curvatures (P<.001, χ2 test).
Correlation Between Development of Aneurysms and Degree of Blood Pressure Elevation
The mean blood pressure elevation of 19 rats with aneurysms at nonbranching sites in the left P1 was 78.7±17.6 mm Hg. This was significantly higher than that of rats without aneurysms (55.3±16.8 mm Hg) (P<.001, Mann-Whitney U test). When blood pressure elevation was less than or equal to 50 mm Hg, aneurysms at nonbranching sites were not induced. In rats that had multiple aneurysms at nonbranching sites, blood pressure elevation was more than or equal to 70 mm Hg with the exception of one rat. The degree of blood pressure elevation was significantly correlated with the number of induced aneurysms arising at nonbranching sites (P=.0069, Kruskal-Wallis test). In contrast, there was no significant correlation between the degree of blood pressure elevation and location in terms of curvature (P=.781, Kruskal-Wallis test) or type (P=.266, Mann-Whitney U test) of induced aneurysm at nonbranching sites.
The dilated and tortuous P1 segment proximal and distal to an aneurysm generally had a thinner wall than the P1 segment of the control animals. This was mainly caused by thinning of the medial muscle layer. The internal elastic lamina was mostly continuous, but small disruption of the lamina was seen in some places.
Despite the macroscopic difference between fusiform and lateral aneurysms, their histopathological features were similar. At the beginning of aneurysmal change, the medial layer became thinner and tapered to nothing at the dome of the aneurysm (Fig 4a⇓). Smooth muscle cells in the medial layer of the aneurysmal wall varied in size and shape and were arranged in a disorderly fashion. Shrunken cells, whose cytoplasms or nuclei exhibited fragmentation, were sometimes observed. The number of smooth muscle cells decreased, and the media was replaced by connective tissue in large aneurysms. Although the internal elastic lamina was continuous in small aneurysms, it had a tendency to be tapered or fragmented or to disappear at the dome (Fig 4b⇓). In large aneurysms, the internal elastic lamina had almost completely disappeared. Such changes were also seen at the end of the fusiform or lateral aneurysms. In the inner side of the aneurysmal wall, endothelial cells almost always existed. They tended to be flat along the axis of the blood flow (Fig 4c⇓). In a fusiform aneurysm, endothelial cells on intimal thickenings were taller than those in other parts. No apparent intimal thickening was seen in lateral aneurysms, but intimal thickening was seen in some fusiform aneurysms (Fig 4d⇓). The adventitia of an aneurysm was continuous to the proximal and distal segments of the parent vessel. Thickening of the adventitia was sometimes seen in fusiform aneurysms but not in lateral aneurysms. Thus, fusiform aneurysms tended to have thick walls, and lateral aneurysms generally had very thin walls.
In a saccular aneurysm arising at bifurcations, thinning or disappearance of the medial smooth muscle layer and fragmentation or absence of the internal elastic lamina were common findings (Fig 4e⇑).
In all cases in the present study, neither signs of rupture nor histological changes indicated dissection of the wall.
By ligating one common carotid artery and producing hypertension with or without β-aminopropionitrile feeding, we experimentally induced saccular aneurysms arising from branching sites in rats7 and monkeys8 and demonstrated that hemodynamic stress,11 hypertension,12 and abnormal metabolism of connective tissue played important etiologic roles in their development. The macroscopic and microscopic findings and their natural course are in accordance with those arising spontaneously in humans.9 10
Most saccular cerebral aneurysms arise at the apex of arterial bifurcations in humans. Cerebral aneurysms without relation to bifurcations are unusual. With recent advances in microsurgery and microanatomy of cerebral aneurysms, aneurysms at nonbranching sites have been found more frequently. It is not clear whether the etiology and pathogenesis of saccular aneurysms at bifurcations and those at nonbranching sites are the same. Stehbens13 noted that saccular aneurysms unrelated to arterial divisions can be caused by arteriosclerosis. Ohara et al14 also interpreted them as sclerotic. In contrast, Nutik15 considered that several paraclinoid aneurysms were congenital berry aneurysms histologically. Fusiform aneurysms, which develop more frequently on the vertebral artery than on the internal carotid artery, have been speculated to be atherosclerotic.16 However, it is not clear whether atherosclerotic changes in such aneurysms are the result or the cause of the dilatation. On the other hand, saccular aneurysms at bifurcations are not considered to be generally caused by atherosclerosis, even if they show a severely sclerotic appearance.
Patients with aneurysms unrelated to bifurcations often have other intracranial saccular aneurysms.1 2 15 17 18 In the present study cerebral aneurysms at nonbranching sites were induced concomitantly with those arising at branching sites by the same generating factors: increases of hemodynamic stress and hypertension.
Hypertension is associated with microaneurysms of penetrating arteries that are related to intracerebral hemorrhage.19 20 21 However, the relationship of hypertension to the development of aneurysms of leptomeningeal vessels, including those arising at nonbranching sites, is not clear.22 23 24 Previously, we reported that it takes longer to induce aneurysms at branching sites in animals by ligation of the carotid artery without hypertension.25 In the present study, aneurysms arising at nonbranching sites did not tend to develop when the degree of blood pressure elevation was not high. Furthermore, there was a significant correlation between the degree of blood pressure elevation and the number of induced aneurysms. However, hypertension in itself may not cause such aneurysms. If hypertension had been the main cause of aneurysms, they would also have been induced in various portions other than the left P1 segments in the circle of Willis of experimental rats. Aneurysms at nonbranching sites could not be induced at any portions other than the left P1 segments. We have considered that further amplification of blood flow on collateral pathways due to blood pressure elevation plays an important role rather than hypertension itself in the development of aneurysms at nonbranching sites.
The P1 segment in nontreated Sprague-Dawley rats is generally smaller than any other segment of the circle of Willis except the anterior communicating arterial complex, which has considerable variations. When one carotid artery in the neck is ligated, the P1 segment on the side of carotid ligation acts as a major collateral pathway of the circle of Willis. Because this segment is small, it is supposedly more sensitive to increased blood flow than any other portion of the circle, and because the P1 segment is fairly long, it may be more prone to be affected.
If blood flow increases, wall shear stress, which is given as 4Q/πr3 (expressed in dynes per square centimeter), assuming laminar flow, where Q is blood flow (expressed in milliliters per second) and r is the internal radius (expressed in centimeters), is also elevated in proportion to it. Many studies on arteriovenous fistulas and the collateral circulation have shown that increased blood flow induces blood vessel dilatation and tortuosity.4 26 27 28 29 Kamiya and Togawa30 demonstrated autoregulation, which can keep wall shear stress constant, in arterial walls. As a result of this regulation, increases in blood flow may induce adaptive enlargement of the vessel radius, which acts as negative feedback to reduce the stress itself. The most essential histological findings in this process were severe fragmentation of internal elastic lamina and degeneration of medial smooth muscle layer,31 which were consistent with those of aneurysms. These changes may have been caused by the direct mechanical force of increased blood flow, but the authors proposed that active arterial wall restructuring was responsible for keeping wall shear stress constant.31 Furthermore, the significance of an endothelial cell in this restructuring as a mechanosensor that can perceive increased wall shear stress was also proposed.32 In our aneurysms as well, endothelial cells constantly exist in even very thin walls whether they developed at branching or nonbranching sites. From this point of view, an aneurysmal formation can be one of active restructuring of arterial wall against increased blood flow. Although the mediators between endothelial cells and medial smooth muscle cells or internal elastic lamina are not clear, some types of endothelium-derived factor such as nitric oxide and prostaglandin I2 may be candidates.33 34
In our study all lateral aneurysms developed near or on the outside wall of the curvature of the vessel. The development of fusiform aneurysms was also related to vessel curvature. With a glass model of curved vessels with lateral aneurysms, Niimi et al35 showed that wall shear stress at the outer side wall is much higher than that at the inner side wall. Blood flow into the curvature may enhance tortuosity and cause lateral aneurysms. When such a change occurs in the entire circumference, the segment may enlarge, thereby causing fusiform aneurysms in our model. Lateral aneurysms may enlarge to involve the entire circumference, finally growing as a fusiform aneurysm.
Cerebral arterial bifurcations are sites where increased hemodynamic stress causes degeneration of the arterial wall, which contributes to aneurysmal bulging, while, as shown in the present study, morphological and histological changes can also occur at nonbranching sites of the circle of Willis through hemodynamic changes. A morphological change such as tortuosity may further aggravate hemodynamic stress on a certain portion, such as the corner of the curvatures. This vicious cycle may be the basis of aneurysmal formation at sites where no branches emerge.
In conclusion, the present findings indicate that cerebral aneurysms at nonbranching sites and saccular aneurysms at branching sites can occur under the same etiologic conditions. The sites of origin are strongly related to hemodynamic stress. Further studies in which this animal model is used should provide new information regarding cerebral aneurysms arising at nonbranching sites.
Reprint requests to Soichiro Kondo, MD, Department of Neurosurgery, Nagahama City Hospital, 313 Oinui-cho, Nagahama-shi, Shiga, 526 Japan.
- Received April 4, 1996.
- Revision received September 6, 1996.
- Accepted September 6, 1996.
- Copyright © 1997 by American Heart Association
Liebow AA. Situations which lead to changes in vascular patterns. In: Hamilton WF, Dow P, eds. Handbook of Physiology, Section 2: The Cardiovascular System. Washington, DC: American Physiological Society; 1963;2:1251-1276.
Pia HW, Langmaid C, Zierski J, eds. Cerebral Aneurysms. New York, NY: Springer-Verlag; 1979:5-19.
Stehbens WE. The pathology of intracranial arterial aneurysms and their complications. In: Fox JL, ed. Intracranial Aneurysms. New York, NY: Springer-Verlag; 1983;1:272-357.
Ohara H, Sakamoto T, Suzuki J. Sclerotic cerebral aneurysms. In: Suzuki J, ed. Cerebral Aneurysms. Tokyo, Japan: Neuron Publishing Co; 1979:673-682.
Sano K. Internal carotid posterior communicating anterior choroidal region aneurysms: microsurgical treatment. In: Pia HW, Langmaid C, Zierski J, eds. Cerebral Aneurysms. Advances in Diagnosis and Therapy. New York, NY: Springer-Verlag; 1979:252-259.
Feldmann E. Intracerebral hemorrhage. Stroke.. 1991;22:684-691.
Ross Russel RW. Observations on intracerebral aneurysms. Brain.. 1963;86:425-442.
Kwak R, Mizoi K, Katakura R, Suzuki J. The correlation between hypertension in past history and the incidence of cerebral aneurysms. In: Suzuki J, ed. Cerebral Aneurysms: Experience with 1000 Directly Operated Cases. Tokyo, Japan: Neuron Publishing Co; 1979:20-24.
Nagata I. Hemodynamic stress and developmental mechanism in experimental cerebral aneurysms in rats. Arch Jpn Chir.. 1982;51:44-58.
Shenk WG, Martin JW, Leslie MB, Portin BA. The regional hemodynamics of chronic experimental arteriovenous fistulas. Surg Gynecol Obstet.. 1961;110:44-50.
Weibel E. Early stages in the development of collateral circulation to the lung in the rat. Circ Res.. 1960;8:353-376.
Kamiya A, Togawa T. Adaptive regulation of wall shear stress to flow change in the canine carotid artery. Am J Physiol.. 1980;239:H14-21.
Masuda H, Bassiouny H, Glagov S, Zarins CK. Artery wall restructuring in response to increased flow. Surg Forum.. 1989;40:285-286.
Frangos JA, Eskin SG, McIntire LV, Ives CL. Flow effects on prostacyclin production by cultured human endothelial cells. Science.. 1985;227:1477-1479.
Rubanyi GM, Romero JC, Vanhoutte PM. Flow-induced release of endothelium-derived relaxing factor. Am J Physiol.. 1986;250:1145-1149.
Kondo and colleagues show that 1 year after unilateral carotid ligation and induction of hypertension, 89% of rats developed increased size and/or tortuosity of the posterior cerebral artery through which blood flow is presumably increased to compensate for the carotid occlusion. Eighty-six percent developed typical saccular aneurysms at arterial bifurcations, usually on the anterior circulation, and 54% developed fusiform or saccular aneurysms that were not associated with branches on the dilated and/or tortuous posterior cerebral arteries. There is a remarkably high incidence of aneurysm formation in this model, and aneurysm formation seems to be related to a combination of hypertension and changes in blood flow, although the latter were not actually measured in this study. This model in rats would seem to provide an excellent opportunity to investigate the pathogenesis of aneurysms from both a molecular biological and biophysical point of view, and some experiments have been published showing changes in fibronectin in rat bifurcation-type aneurysms.1R 2R The interpretation of these results and their generalization to humans would have to include consideration of how this model relates to aneurysm genesis in humans. Similarities include the association of dolichoectasia of intracranial arteries in humans with hypertension and the vertebrobasilar circulation and bifurcation-type saccular aneurysms occurring on arteries experiencing increased flow after carotid occlusion.3R The model seems to differ in that hypertension is not as prominent a factor in saccular aneurysm formation in humans as it seems to be in the rat model, and none of the aneurysms in rats ruptured over an albeit short course of observation. It is difficult to envision any clinical importance to the observations of Kondo et a1 at this point, but it has generally been hoped that with better understanding of the pathogenesis and hemodynamics of aneurysms, some therapeutic advance will be derived.
Futami K, Yamashita J, Tachibana O, Kida S, Higashi S, Ikeda K, Yamashima T. Basic fibroblast growth factor may repair experimental cerebral aneurysms in rats. Stroke.. 1995;26:1649-1654.
Futami K, Yamashita J, Tachibana O, Higashi S, Ikeda K, Yamashima T. Immunohistochemical alterations of fibronectin during the formation and proliferative repair of experimental cerebral aneurysms in rats. Stroke.. 1995;26:1659-1664.
Schwartz A, Rautenberg W, Hennerici M. Dolichoectatic intracranial arteries: review of selected aspects. Cerebrovasc Dis. 1993;3;273-279.