- Email: [email protected]
Breakdown of the Meningeal Barrier Surrounding the Intraorbital Optic Nerve After Experimental Subarachnoid Hemorrhage
Breakdown of the Meningeal Barrier Surrounding the Intraorbital Optic Nerve After Experimental Subarachnoid Hemorrhage THOMAS BRINKER, MD, WOLF LUDEMANN, MD, DIRK BERENS YON RAUTENFELD, MD, FRIEDHELM BRASSEL, MD, HARTMUT BECKER, MD, AND MADJID SAMII, MD
• PURPOSE: The intraorbital optic nerve sheath meninges contain a perineural subarachnoid space lined by meningeal cell layers and intercellular fibrous tissue. We sought to determine whether functional or structural characteristics, or both, of the optic nerve sheath are influenced by the increased intracranial pressure after the rupture of cerebral aneurysms. • METHODS: We infused the great cisterns of cats with either x-ray contrast medium or autologous blood. The cisternal infusions were done under the experimental condition of a sudden 2.5-minute increase in intracranial pressure similar to that recorded after the rupture of cerebral aneurysms in humans. • RESULTS: Digital subtraction radiographs of the optic nerves taken during the cisternal infusion of contrast medium at the start showed the opacification of the optic nerve subarachnoid space. After 2 minutes, the contrast medium leaked into the orbit, indicating the breakdown of the meningeal fluid barrier. Ultrastructural investigation of the optic nerve sheath after high-pressure cisternal infusions showed the arachnoid cell layers scat tered. The flattened arachnoid cells displayed Accepted for publication March 10, 1997. From the Neurosurgical Department, Nordstadt Hospital (Drs Brinker and Sarmi), and the Departments of Anatomy (Drs Liidemann, von Rautenfeld, and Brassel) and Neuroradiology (Dr Becker), Medical School Hannover; and the Department of Neuroradiology/Radiology, Medical School, University of Greifswald (Dr Brassel), Germany. This study was supported in part by the Deutsche Forschungsgemeinschaft, Grant Br 1416/1-1. Reprint requests to Thomas Brinker, MD, Neurosurgical Department, Nordstadt Hospital, Haltenhoffstr 41, D-30167, Hannover, Germany; fax: 49-511-9701606; e-mail: [email protected]
VOL.124, No. 3
mainly intracellular and some intercellular, pore like openings. After infusion of blood into the great cistern, erythrocytes were found within pqrelike openings of the arachnoid cells. • CONCLUSIONS: The meningeal fluid barrier of the optic nerve sheath can be destroyed by pressure changes associated with subarachnoid hemor rhage. This disruption might be regarded as a natural optic nerve sheath fenestration that allows outflow of cerehrospinal fluid into the orbit to protect the optic nerve from increased intracranial pressure after aneurysmal rupture.
T
HE COINCIDENCE OF OPHTHALMIC DISORDERS with subarachnoid hemorrhage is well known. Vitreous hemorrhage, optic nerve sheath hem orrhage, and orbital hemorrhage can be observed after aneurysmal rupture and subarachnoid hemor rhage. It has been suggested that these complications of subarachnoid hemorrhage are related to the minute-long, marked increase in intracranial pressure after aneurysmal rupture.1,2 The meninges surrounding the intraorbital optic nerves are normally watertight except for a small retrobulbar zone in which tortuous channels con necting the perineural subarachnoid space and the retrobulbar soft tissue have been described in ham sters and rabbits.3,4 Experimentally, at increased intracranial pressure, cerebrospinal fluid and dye particles drain from the subarachnoid space through retrobulbar channels into the orbit. On the other hand, an increased intracranial pressure can destroy the meningeal fluid
© AMERICAN JOURNAL OF OPHTHALMOLOGY i997;i 24:373-380
373
barrier: examining convexity meninges of cats after a strong increase in intracranial pressure, Butler and associates5 reported structural changes of the menin ges (that is, openings of intercellular tight junctions) that showed a general breakdown of the barrier function. Our aim was to investigate the effects on the optic nerve sheath following the increase in intracranial pressure after the rupture of a cerebral aneurysm. We asked the following questions: Is a spontaneous out flow of cerebrospinal fluid into the orbit possible under the intracranial pressure conditions of subarachnoid hemorrhage in cats? Can this phenome non be related to ultrastructural findings? We performed cisternal infusions of x-ray contrast medi um or blood in cats during a sudden increase in intracranial pressure similar to that recorded after the rupture of cerebral aneurysms in humans.6,7 The meningeal barrier function was studied by radio graphs of the optic nerves during the infusion of contrast medium. Electron microscopic studies were done to analyze the ultrastructure of the optic nerve sheath after the increased intracranial pressure.
METHODS TEN CATS OF BOTH SEXES, EACH WEIGHING BETWEEN
3.6 and 4-2 kg, were used for the experiments. Four cats were injected with contrast medium and four with fresh autologous nonheparinized blood, and two underwent sham surgery for ultrastructure control examinations. The experiments were performed with institutional approval. Anesthesia was induced by intramuscular injection of ketamine, 30 mg/kg, supplemented by a small dose of barbiturate. The animals were intubated endotracheally and ventilated by a small-animal respirator. The femoral blood vessels were catheterized for con tinuous recording of arterial blood pressure and con tinuous administration of drugs. Physiological body temperature was maintained by a heating blanket. Anesthesia was continued with the infusion of keta mine at a rate of 30 mg per hour combined with pancuronium, 0.8 mg per hour. Arterial blood gas measurements were checked routinely and kept with in normal limits. A microcatheter was inserted through the atlanto-occipital membrane into the 374
great cistern of the posterior fossa. A pressure trans ducer was connected to the cisternal catheter by a three-way stopcock. Either 3.2 ml of x-ray contrast medium (iodine content 300 mg/ml) or 2.8 ml of fresh autologous blood was infused into the great cistern. The contrast medium was supplemented with Berlin blue solution 1% (10:1) for postmortem morphologic tracing exam inations. The cisternal infusions were performed using a computer-controlled pump with adjustable infusion speed. The infusion rate was adjusted to achieve a sudden, 2.5-minute-long increase in intracranial pressure up to the level of the arterial blood pressure.8 During the infusion of x-ray contrast medium, serial radiographs were taken by a biplane digital angiographic x-ray unit. Immediately after the cisternal injection with ei ther blood or contrast medium, an arterial perfusionfixation was done through a catheter introduced into the aortic arch after clamping the descending aorta. The cava vein was opened by slit incision. After a perfusion with 100 ml of heparinized saline, 200 ml of an aldehyde fixative (formaldehyde 2.0%, glutaraldehyde 2.5%, 0.1 M sodium cacodylate, and CaCl2 0.025% at pH 7.4) was infused within 10 minutes. After the perfusion, the eyeballs were removed and the optic nerves were exposed by microsurgical tech niques. The specimens were immersed in the alde hyde fixative for 24 hours. Samples were postfixed in osmium tetroxide 2% in cacodylate buffer for 2 hours and dehydrated in increasing concentrations of alco hol. For scanning electron microscopy, specimens of the optic nerve were dried in a critical point dryer with carbon dioxide and mounted on suitable stubs. The specimens were coated with gold and observed and photographed with a scanning electron micro scope. For transmission electron microscopy, 3- to 5-mm specimens of the optic nerve were embedded in epoxy resin. Ultrathin sections were stained with uranyl acetate 5% and lead citrate 0.4% and exam ined by electron microscope.
RESULTS USING THE INFUSION RATES DEPICTED IN FIGURE 1, LEFT,
the computer-controlled cisternal infusion of the four
AMERICAN JOURNAL OF OPHTHALMOLOGY
SEPTEMBER 1997
4.0
Contrast Medium 80
iJV
5 min.
0.0 0.5 1.0
Blood
5 min.
1.5 2.0 2.5 3.0
Time (min.) FIGURE 1. (Left) The flow rates of the cisternal infusion of x-ray contrast medium (dotted line) or blood (solid line) are plotted against time. They are computer-calculated using a mathematical model16 to achieve a 2.5-minute-long increase in intracranial pressure of up to 80 mm Hg. The flow rates of blood and x-ray contrast medium differ, considering that the intracranial volume buffering mechanism is impaired during experimental subarachnoid hemorrhage but not during the cisternal infusion of contrast medium.8 Therefore, in comparison to the contrast medium, a lower infusion rate of blood is necessary to produce the same increase in intracranial pressure. (Right) Intracranial pressure recordings during the cisternal infusions (arrowheads = start and stop of the infusion). There is a sharp increase to the level of the arterial blood pressure, then a plateau for 2.5 minutes. Afterward, the intracranial pressure declines to baseline levels.
cats with 3.2 ml of contrast medium raised the intracranial pressure within seconds from a mean ± SD of 8.3 ± 3.2 mm Hg to 85.4 ± 7.9 mm Hg. The intracranial pressure remained at this level during the infusion (2.5 minutes), then declined to baseline levels within minutes. The same pattern of intracra nial pressure was achieved by the computer-controlled infusion of 2.8 ml of autologous fresh blood in four cats. The intracranial pressure increased suddenly to 81.8 ± 2.9 mm Hg and remained at this level (91.5 ± 25.8 mm Hg) until the end of infusion. Then it declined within minutes to baseline levels (Figure 1, right). The radiographs, taken during the first minute of the cisternal infusion with x-ray contrast medium, showed filling of the spinal dural sac and the basal cisterns. Then the contrast medium approached the chiasmatic cistern and opacified the optic nerve subarachnoid space. After 2 minutes of infusion, the contrast medium leaked from the optic nerve sub arachnoid space into the orbit (Figure 2, left). At the end of the cisternal infusion, the x-rays showed a large deposit of contrast medium surrounding the VOL. 124, NO. 3
whole intraorbital portion of the optic nerve (Figure 2, right). After the high-pressure infusion of contrast medi um or blood, vitreous hemorrhage was not observed. Only when blood had been cisternally infused did each animal show optic nerve sheath hemorrhage and leakage of blood into the orbit. Optic nerve sheath hemorrhage or intraorbital blood was not found in animals that had been injected with contrast medium. However, because the contrast medium was supplemented with Berlin blue, the optic nerve sub arachnoid space was stained, and a blue fluid wheel was seen in the retrobulbar tissue surrounding the optic nerve. The normal ultrastructural appearance of the optic nerve sheath was studied in the two cats that underwent sham surgery. In transmission electron microscopy, the outermost, thickest layer of the meninges, the dura mater, was seen to be mainly composed of crisscrossed bundles of collagen fibers with few fibrocytic cells. The outer arachnoid cell layer next to the dura consists of flattened cells; these were densely packed. In the direction of the subarach-
BREAKDOWN OF THE MENINGEAL BARRIER
375
FIGURE 2. (Left) Digital subtraction image at the beginning of the infusion. Both optic nerves (arrowheads) are visible by opacification of their subarachnoid spaces (submentovertex view of the optic nerves, with the orbital portion indicated by the double-headed arrow, including the chiasm [asterisk]). Intracranial pressure during the infusion = 85 mm Hg. (Right) Digital subtraction image 2 minutes after start of the infusion: contrast medium leaked into the orbit and formed a deposit ( + ) surrounding the intraorbital portion of both nerves (double-headed arrow; submentovertex view of the optic nerves [arrowheads] and the chiasm [asterisk]). Intracranial pressure during the infusion = 85 mm Hg.
noid space, the arachnoid cells were more loosely arranged and less flattened. Between these innermost arachnoid cell layers, small interstitial spaces contain ing mainly collagenous fibers and some elastic fibers were observed (Figure 3, left). In scanning electron microscopy, the outer arach noid cell layers can be seen as a smooth surface of cells closely joined to each other (Figure 3, right). After high-pressure cisternal infusion, the arachnoid cell layers were scattered and the arachnoid cells flattened. The intercellular space between the arach noid cell layers was widened, and particles of either Berlin blue or blood cells were seen in the subarach noid space and the intercellular spaces of the arach noid cell layers (Figure 4, left). Scanning electron microscopic examination after high-pressure cisternal infusion showed the arachnoid cells of all arachnoid layers to have many 3- to 7-|xm (median, 5 |JLm), intracellular (and some intercellu lar), porelike openings (Figure 4, right), in which particles of either Berlin blue or blood cells were seen. Electron microscopy after cisternal infusion with blood disclosed blood clots within the arachnoid cell layers rather than clots in the perioptic subarachnoid space (Figure 5, bottom). Transmission electron mi croscopy (Figure 5, top left) and scanning electron microscopy (Figure 5, bottom) showed red blood cells 376
crossing the arachnoid cells through porelike open ings and accumulation in front of the dural fibers. Optic nerve sheath hemorrhage was found only in the cats infused with blood, not in those infused with contrast medium.
DISCUSSION CONCERNING THE TECHNICAL ASPECTS OF THE EXPERI-
ments, it is to be questioned whether the increase in intracranial pressure during cisternal infusion reliably simulates the clinical conditions of aneurysmal rup ture and subarachnoid hemorrhage in patients. The changes of intracranial pressure as recorded after aneurysmal rupture in humans are unique: the intra cranial pressure jumps within seconds to the level of the arterial blood pressure. After some minutes, the pressure returns spontaneously to baseline levels in those patients surviving this critical phase of sub arachnoid hemorrhage. Comparing the published intracranial pressure recordings6,7 with those of the present experiments, it is obvious that the model closely simulates the intracranial pressure increase after aneurysmal rupture in humans. 8 This is true for the injections of the animals with both x-ray contrast medium and blood. Therefore, our model is valuable
AMERICAN JOURNAL OF OPHTHALMOLOGY
SEPTEMBER
1997
FIGURE 3. (Left) Transmission electron microscopic image of the normal optic nerve sheath meninges of a control cat shows the thickest and outermost dural layer (D) and the arachnoid cell layer (A). The dural layer consists predominantly of collagenous fibers. The arachnoid portion is formed by cells that are densely packed in the outer layer and more loosely arranged in the direction of the subarachnoid space (X 6,000). (Right) With scanning electron microscopy, the surface of normal arachnoid cell layers can be observed. The intact arachnoid layer has a smooth cell surface with cells closely joined. The cell borders (arrows) are visible. N = nucleus (X 7,200).
FIGURE 4. (Left) Transmission electron microscopic image showing the appearance of the meninges after high-pressure cisternal infusion with contrast medium. The arachnoid cells (asterisk) are flattened and scattered, thus widening the intercellular spaces (arrows, extracellular spaces); F = collagenous fibers (X 16,000). (Right) Appearance of the meninges after high-pressure cisternal infusion with contrast medium. Scanning electron microscopy discloses numerous round, porelike openings (asterisks) of the arachnoid cell surface (diameter, 5 to 10 |JLm). Following the cell borders (arrows), it is obvious that these porelike openings are intracellular (i) as well as intercellular (asterisks). N = nucleus (X 7,000). VOL.124, No. 3
BREAKDOWN OF THE MENINGEAL BARRIER
377
FIGURE 5. (Top left) Optic nerve sheath after cisternal infusion with blood. Transmission electron microscopic image of an erythrocyte (E) within the arachnoid cell layers (arrowheads) passing between collagenous filaments (F) (X 20,000). (Top right) Optic nerve sheath after cisternal infusion with blood. Transmission electron miscroscopic image of an erythrocyte (E) passing through the porelike intercellular opening of arachnoid cells. The irregular edges of the pore indicate a mechanical disruption of the arachnoid cell layer as a result of the sudden increase in intracranial pressure, with consecutive stretching of the optic nerve sheath. Cell borders are indicated by arrows (X 14,000). (Bottom) Optic nerve sheath after cisternal infusion with blood. After crossing the arachnoid layers, the blood cells accumulate in front of the filterlike fiber system of the dura (D) and form a "subdural" blood clot (C) lying within and between the arachnoid cell layers in the subdural space. SAS = subarachnoid space (X800).
for the interpretation of clinical findings after subarachnoid hemorrhage. The radiographs show an outflow of cerebrospinal fluid containing contrast medium into the orbit in every animal. The outflow is not limited to the retrobulbar portion but comprises the complete intraorbital portion of the nerve. Therefore, the experi ments prove the breakdown of the meningeal barrier surrounding the intraorbital optic nerve in the living animal 2 minutes after starting the cisternal infusion. Until now, clinical cisternographies had shown only the opacification of the optic nerve sheath, 378
not leakage of contrast medium into the orbit.9'11 This discrepancy can be explained by the increased intracranial pressure in our experiments. All the previous studies mentioned were done at normal intracranial pressure. Therefore, we suggest that the leakage of contrast medium into the orbit can be attributed to an increased intracranial pressure during cisternography. Anatomically, the optic nerve sheath is continuous with the dura mater and extends the intracranial subarachnoid space into the orbit. The role of the meninges as a barrier to the exchange of fluid and of
AMERICAN JOURNAL OF OPHTHALMOLOGY
SEPTEMBER 1997
particles is well known and has been examined by electron microscopic and ultrastructural horseradishperoxidase histochemical techniques. The meningeal barrier has been shown to exist at the "arachnoid barrier layer" located within multiple layers of cells forming the arachnoid membrane between the dura and subarachnoid space. The arachnoid barrier layer prevents the transfer of fluid because it consists of cells connected by a dense and continuous system of tight junctions.12 Our observation confirms the exper iments of Butler and associates,5 who studied convexi ty meninges of cats with the transmission electron microscopy technique after an increased intracranial pressure of up to 50 mm Hg. The authors injected saline and horseradish-peroxidase solutions into the subarachnoid space, and the tracer was found beyond the arachnoid barrier layer in the extracellular space of the adjacent dura. Corresponding to the present results, a widening of the extracellular spaces between arachnoid cell layers was found secondary to the intracranial pressure increase. The ultrastructural cellular mechanism of the suggested meningeal barri er disruption, however, was not shown in that investi gation.5 In the present experiments, intercellular and mainly intracellular, porelike openings were found to be the cellular mechanism allowing the transmeningeal passage of fluid and particles. Similar pores were found in cells at the angle of the anterior chamber for the outflow of aqueous humor in the eye and in cells of the arachnoid villi for cerebrospinal fluid outflow. Those pores, described as openings of transcellular channels occurring at in creased intraocular or intracranial pressure,13 are in contrast to those in our findings. They are considered to be the ultrastructural correlate of a pressuredependent transcellular bulk flow mechanism for the outflow of aqueous humor or cerebrospinal fluid.13 Because channels related to the observed porelike openings of the cells could not be demonstrated in our experiments, it is unlikely that the present ultrastructural findings can be compared with the ultrastructure of the aqueous chamber or of the arachnoid villi. Considering the radiologic findings, the hypothesis we favor is that the observed porelike openings result from the intravital rupture of arachnoid cells, proba bly because of the distention of the optic nerve sheath caused by increased intracranial pressure. The VOL.124, No. 3
distention of the optic nerve sheath during an in crease in intracranial pressure is well known 14 and might stretch the arachnoid cells, leading to a disrup tion in the observed manner. The pathway that we describe for the outflow of cerebrospinal fluid through the optic nerve is different from that reported in previous ultrastructural studies on the optic nerve sheath architecture.3 Our investi gation demonstrated a zone of increased fluid perme ability of the meninges in the retrobulbar portion of the optic nerve where the optic nerve sheaths join with the sclera. In this region of the nerve, numerous tortuous channels within the meningeal nerve sheath were described, providing a physiological outflow pathway for cerebrospinal fluid. The cerebrospinal fluid outflow pathway as described here is not related to channels. Furthermore, the porelike openings are not limited to the end of the nerve but were found along the entire intraorbital portion of the optic nerve. Considering the experiments of Butler and associates,5 we suggest that meninges, independent of their location inside or outside the intracranial space, have the same behavior when exposed to a significant pressure gradient: the arachnoid cells become flat tened, the extracellular space is widened, and finally the cells are disrupted, showing intracellular and intercellular pores. The pores are the expression of a mechanically disrupted meningeal fluid barrier. In contrast to the results of Butler and associates,5 we found mainly intracellular pores. Whether these pores stay open after increased pressure and how they influence the outflow resistance in the long term remain unclear. According to our experiments, the hypothesis that optic nerve sheath hemorrhage after subarachnoid hemorrhage results from the disruption of intradural and bridging nerve sheath vessels because of the increased intracranial pressure1,2 should be ques tioned. In our experiment, optic nerve sheath hemor rhage is related to the extension of intracranial blood, not to the disruption of meningeal vessels, because no optic nerve sheath hemorrhage was found in those animals infused with contrast medium. Our experimental observations explain why, after subarachnoid hemorrhage in humans, blood clots have often been found to be localized in the subdural rather than in the subarachnoid space of the optic nerve. 2 The subdural blood clots result from the
BREAKDOWN OF THE MENINGEAL BARRIER
379
disruption of the arachnoid barrier layer, which allows blood cells to penetrate from the optic nerve subarachnoid space into cell layers of the nerve sheath to accumulate in front of the filterlike duralfibersystems. The sudden increase in intracranial pressure after aneurysmal rupture and subarachnoid hemorrhage opens the arachnoid barrier layer to enable an outflow of cerebrospinal fluid and even of blood cells across the optic nerve sheath into the orbit. Opening of the optic nerve sheath barrier during subarachnoid hem orrhage may have a protective effect on the optic nerve, similar to surgical nerve sheath fenestration in pseudotumor cerebri patients.15
6. 7. 8.
9.
10. 11.
REFERENCES
12.
1. Garfinkle AM, Danys I, Nicolle D, Colohan A, Brem S. Terson's syndrome: a reversible cause of blindness following subarachnoid hemorrhage. J Neurosurg 1992;76:766-771. 2. Muller P, Deck ]. Intraocular and optic nerve sheath hemorrhage in cases of sudden intracranial hypertension. J Neurosurg 1974;41:160-166. 3. Erlich S, McComb JS, Weiss M. Ultrastructure of the orbital pathway for cerebrospinal fluid drainage in rabbits. ] Neuro surg 1989;70:926-931. 4- Shen J-Y, Kelly D, Hyman S, McComb ]. Intraorbital cerebrospinal fluid outflow and the posterior uveal compart ment of the hamster eye. Cell Tissue Res 1985;240:77-87. 5. Butler A, Van Landingham K, McComb J. Pressure facilitat ed CSF flow across the arachnoid membrane. In: Ishii S,
13. 14. 15.
16.
Nagai H, Brock M, editors. Intracranial pressure. Volume V. Berlin: Springer Verlag, 1983:598-601. Nornes H. Monitoring of patients with intracranial aneurysms. Clin Neurosurg 1975;22:321-331. Volby B, Enevoldsen E. Intracranial pressure changes follow ing aneurysm rupture, 3: recurrent hemorrhage. ] Neurosurg 1982;56:784-789. Brinker T, Dietz H. Computer simulation of the dynamics of the cerebrospinal fluid system and its experimental applica tion. Informatik, Biometrie und Epidemiologie in Medizin und Biologie 1993;24:129-133. Stuck R, Reeves D. Dangerous effects of thorotrast use intracranially with special reference to experimental produc tion of hydrocephalus. Arch Neural Psychiatry 1938;40: 86-115. Chambers E, Manelfe C, Cellerier P. Metrizamide CT cisternography and perioptic subarachnoid space imaging. J Comput Assist Tomogr 1981;5:875-880. Davis J, Haughton V, Harris G, Eldevik O, Gager W. Cat optic nerve imaging with metrizamide. Invest Ophthalmol Vis Sci 1979;18:273-277. Nabeshima S, Reese T, Landis D, Brightman M. Junctions in the meninges and marginal glia. J Comp Neurol 1993;164: 127-170. Tripathi B, Tripathi R. Vacuolar transcellular channels as a drainage pathway for cerebrospinal fluid. J Physiol 1974;239: 195-206. Liu D, Kahn M. Measurement and relationships of subarach noid pressure of the optic nerve to intracranial pressures in fresh cadavers. Am ] Ophthalmol 1993;116:548-556. Spoor TC, Ramocki JM, Madion MP, Wilkinson MJ. Treat ment of pseudotumor cerebri by primary and secondary optic nerve sheath decompression. Am ] Ophthalmol 1991;112: 177-185. Brinker T, Seifert V, Stolke D. Acute changes of cerebrospi nal fluid system dynamics during experimental subarachnoid hemorrhage. Neurosurgery 1990;27:369-372.
Authors Interactive™ Comments and questions regarding this article may be addressed to authors via THE JOURNAL'S Web Site. Selected comments, questions, and author responses are posted on Authors Interactive™ at http://www.ajo.com/
380
AMERICAN JOURNAL OF OPHTHALMOLOGY
SEPTEMBER 1997
• PURPOSE: The intraorbital optic nerve sheath meninges contain a perineural subarachnoid space lined by meningeal cell layers and intercellular fibrous tissue. We sought to determine whether functional or structural characteristics, or both, of the optic nerve sheath are influenced by the increased intracranial pressure after the rupture of cerebral aneurysms. • METHODS: We infused the great cisterns of cats with either x-ray contrast medium or autologous blood. The cisternal infusions were done under the experimental condition of a sudden 2.5-minute increase in intracranial pressure similar to that recorded after the rupture of cerebral aneurysms in humans. • RESULTS: Digital subtraction radiographs of the optic nerves taken during the cisternal infusion of contrast medium at the start showed the opacification of the optic nerve subarachnoid space. After 2 minutes, the contrast medium leaked into the orbit, indicating the breakdown of the meningeal fluid barrier. Ultrastructural investigation of the optic nerve sheath after high-pressure cisternal infusions showed the arachnoid cell layers scat tered. The flattened arachnoid cells displayed Accepted for publication March 10, 1997. From the Neurosurgical Department, Nordstadt Hospital (Drs Brinker and Sarmi), and the Departments of Anatomy (Drs Liidemann, von Rautenfeld, and Brassel) and Neuroradiology (Dr Becker), Medical School Hannover; and the Department of Neuroradiology/Radiology, Medical School, University of Greifswald (Dr Brassel), Germany. This study was supported in part by the Deutsche Forschungsgemeinschaft, Grant Br 1416/1-1. Reprint requests to Thomas Brinker, MD, Neurosurgical Department, Nordstadt Hospital, Haltenhoffstr 41, D-30167, Hannover, Germany; fax: 49-511-9701606; e-mail: [email protected]
VOL.124, No. 3
mainly intracellular and some intercellular, pore like openings. After infusion of blood into the great cistern, erythrocytes were found within pqrelike openings of the arachnoid cells. • CONCLUSIONS: The meningeal fluid barrier of the optic nerve sheath can be destroyed by pressure changes associated with subarachnoid hemor rhage. This disruption might be regarded as a natural optic nerve sheath fenestration that allows outflow of cerehrospinal fluid into the orbit to protect the optic nerve from increased intracranial pressure after aneurysmal rupture.
T
HE COINCIDENCE OF OPHTHALMIC DISORDERS with subarachnoid hemorrhage is well known. Vitreous hemorrhage, optic nerve sheath hem orrhage, and orbital hemorrhage can be observed after aneurysmal rupture and subarachnoid hemor rhage. It has been suggested that these complications of subarachnoid hemorrhage are related to the minute-long, marked increase in intracranial pressure after aneurysmal rupture.1,2 The meninges surrounding the intraorbital optic nerves are normally watertight except for a small retrobulbar zone in which tortuous channels con necting the perineural subarachnoid space and the retrobulbar soft tissue have been described in ham sters and rabbits.3,4 Experimentally, at increased intracranial pressure, cerebrospinal fluid and dye particles drain from the subarachnoid space through retrobulbar channels into the orbit. On the other hand, an increased intracranial pressure can destroy the meningeal fluid
© AMERICAN JOURNAL OF OPHTHALMOLOGY i997;i 24:373-380
373
barrier: examining convexity meninges of cats after a strong increase in intracranial pressure, Butler and associates5 reported structural changes of the menin ges (that is, openings of intercellular tight junctions) that showed a general breakdown of the barrier function. Our aim was to investigate the effects on the optic nerve sheath following the increase in intracranial pressure after the rupture of a cerebral aneurysm. We asked the following questions: Is a spontaneous out flow of cerebrospinal fluid into the orbit possible under the intracranial pressure conditions of subarachnoid hemorrhage in cats? Can this phenome non be related to ultrastructural findings? We performed cisternal infusions of x-ray contrast medi um or blood in cats during a sudden increase in intracranial pressure similar to that recorded after the rupture of cerebral aneurysms in humans.6,7 The meningeal barrier function was studied by radio graphs of the optic nerves during the infusion of contrast medium. Electron microscopic studies were done to analyze the ultrastructure of the optic nerve sheath after the increased intracranial pressure.
METHODS TEN CATS OF BOTH SEXES, EACH WEIGHING BETWEEN
3.6 and 4-2 kg, were used for the experiments. Four cats were injected with contrast medium and four with fresh autologous nonheparinized blood, and two underwent sham surgery for ultrastructure control examinations. The experiments were performed with institutional approval. Anesthesia was induced by intramuscular injection of ketamine, 30 mg/kg, supplemented by a small dose of barbiturate. The animals were intubated endotracheally and ventilated by a small-animal respirator. The femoral blood vessels were catheterized for con tinuous recording of arterial blood pressure and con tinuous administration of drugs. Physiological body temperature was maintained by a heating blanket. Anesthesia was continued with the infusion of keta mine at a rate of 30 mg per hour combined with pancuronium, 0.8 mg per hour. Arterial blood gas measurements were checked routinely and kept with in normal limits. A microcatheter was inserted through the atlanto-occipital membrane into the 374
great cistern of the posterior fossa. A pressure trans ducer was connected to the cisternal catheter by a three-way stopcock. Either 3.2 ml of x-ray contrast medium (iodine content 300 mg/ml) or 2.8 ml of fresh autologous blood was infused into the great cistern. The contrast medium was supplemented with Berlin blue solution 1% (10:1) for postmortem morphologic tracing exam inations. The cisternal infusions were performed using a computer-controlled pump with adjustable infusion speed. The infusion rate was adjusted to achieve a sudden, 2.5-minute-long increase in intracranial pressure up to the level of the arterial blood pressure.8 During the infusion of x-ray contrast medium, serial radiographs were taken by a biplane digital angiographic x-ray unit. Immediately after the cisternal injection with ei ther blood or contrast medium, an arterial perfusionfixation was done through a catheter introduced into the aortic arch after clamping the descending aorta. The cava vein was opened by slit incision. After a perfusion with 100 ml of heparinized saline, 200 ml of an aldehyde fixative (formaldehyde 2.0%, glutaraldehyde 2.5%, 0.1 M sodium cacodylate, and CaCl2 0.025% at pH 7.4) was infused within 10 minutes. After the perfusion, the eyeballs were removed and the optic nerves were exposed by microsurgical tech niques. The specimens were immersed in the alde hyde fixative for 24 hours. Samples were postfixed in osmium tetroxide 2% in cacodylate buffer for 2 hours and dehydrated in increasing concentrations of alco hol. For scanning electron microscopy, specimens of the optic nerve were dried in a critical point dryer with carbon dioxide and mounted on suitable stubs. The specimens were coated with gold and observed and photographed with a scanning electron micro scope. For transmission electron microscopy, 3- to 5-mm specimens of the optic nerve were embedded in epoxy resin. Ultrathin sections were stained with uranyl acetate 5% and lead citrate 0.4% and exam ined by electron microscope.
RESULTS USING THE INFUSION RATES DEPICTED IN FIGURE 1, LEFT,
the computer-controlled cisternal infusion of the four
AMERICAN JOURNAL OF OPHTHALMOLOGY
SEPTEMBER 1997
4.0
Contrast Medium 80
iJV
5 min.
0.0 0.5 1.0
Blood
5 min.
1.5 2.0 2.5 3.0
Time (min.) FIGURE 1. (Left) The flow rates of the cisternal infusion of x-ray contrast medium (dotted line) or blood (solid line) are plotted against time. They are computer-calculated using a mathematical model16 to achieve a 2.5-minute-long increase in intracranial pressure of up to 80 mm Hg. The flow rates of blood and x-ray contrast medium differ, considering that the intracranial volume buffering mechanism is impaired during experimental subarachnoid hemorrhage but not during the cisternal infusion of contrast medium.8 Therefore, in comparison to the contrast medium, a lower infusion rate of blood is necessary to produce the same increase in intracranial pressure. (Right) Intracranial pressure recordings during the cisternal infusions (arrowheads = start and stop of the infusion). There is a sharp increase to the level of the arterial blood pressure, then a plateau for 2.5 minutes. Afterward, the intracranial pressure declines to baseline levels.
cats with 3.2 ml of contrast medium raised the intracranial pressure within seconds from a mean ± SD of 8.3 ± 3.2 mm Hg to 85.4 ± 7.9 mm Hg. The intracranial pressure remained at this level during the infusion (2.5 minutes), then declined to baseline levels within minutes. The same pattern of intracra nial pressure was achieved by the computer-controlled infusion of 2.8 ml of autologous fresh blood in four cats. The intracranial pressure increased suddenly to 81.8 ± 2.9 mm Hg and remained at this level (91.5 ± 25.8 mm Hg) until the end of infusion. Then it declined within minutes to baseline levels (Figure 1, right). The radiographs, taken during the first minute of the cisternal infusion with x-ray contrast medium, showed filling of the spinal dural sac and the basal cisterns. Then the contrast medium approached the chiasmatic cistern and opacified the optic nerve subarachnoid space. After 2 minutes of infusion, the contrast medium leaked from the optic nerve sub arachnoid space into the orbit (Figure 2, left). At the end of the cisternal infusion, the x-rays showed a large deposit of contrast medium surrounding the VOL. 124, NO. 3
whole intraorbital portion of the optic nerve (Figure 2, right). After the high-pressure infusion of contrast medi um or blood, vitreous hemorrhage was not observed. Only when blood had been cisternally infused did each animal show optic nerve sheath hemorrhage and leakage of blood into the orbit. Optic nerve sheath hemorrhage or intraorbital blood was not found in animals that had been injected with contrast medium. However, because the contrast medium was supplemented with Berlin blue, the optic nerve sub arachnoid space was stained, and a blue fluid wheel was seen in the retrobulbar tissue surrounding the optic nerve. The normal ultrastructural appearance of the optic nerve sheath was studied in the two cats that underwent sham surgery. In transmission electron microscopy, the outermost, thickest layer of the meninges, the dura mater, was seen to be mainly composed of crisscrossed bundles of collagen fibers with few fibrocytic cells. The outer arachnoid cell layer next to the dura consists of flattened cells; these were densely packed. In the direction of the subarach-
BREAKDOWN OF THE MENINGEAL BARRIER
375
FIGURE 2. (Left) Digital subtraction image at the beginning of the infusion. Both optic nerves (arrowheads) are visible by opacification of their subarachnoid spaces (submentovertex view of the optic nerves, with the orbital portion indicated by the double-headed arrow, including the chiasm [asterisk]). Intracranial pressure during the infusion = 85 mm Hg. (Right) Digital subtraction image 2 minutes after start of the infusion: contrast medium leaked into the orbit and formed a deposit ( + ) surrounding the intraorbital portion of both nerves (double-headed arrow; submentovertex view of the optic nerves [arrowheads] and the chiasm [asterisk]). Intracranial pressure during the infusion = 85 mm Hg.
noid space, the arachnoid cells were more loosely arranged and less flattened. Between these innermost arachnoid cell layers, small interstitial spaces contain ing mainly collagenous fibers and some elastic fibers were observed (Figure 3, left). In scanning electron microscopy, the outer arach noid cell layers can be seen as a smooth surface of cells closely joined to each other (Figure 3, right). After high-pressure cisternal infusion, the arachnoid cell layers were scattered and the arachnoid cells flattened. The intercellular space between the arach noid cell layers was widened, and particles of either Berlin blue or blood cells were seen in the subarach noid space and the intercellular spaces of the arach noid cell layers (Figure 4, left). Scanning electron microscopic examination after high-pressure cisternal infusion showed the arachnoid cells of all arachnoid layers to have many 3- to 7-|xm (median, 5 |JLm), intracellular (and some intercellu lar), porelike openings (Figure 4, right), in which particles of either Berlin blue or blood cells were seen. Electron microscopy after cisternal infusion with blood disclosed blood clots within the arachnoid cell layers rather than clots in the perioptic subarachnoid space (Figure 5, bottom). Transmission electron mi croscopy (Figure 5, top left) and scanning electron microscopy (Figure 5, bottom) showed red blood cells 376
crossing the arachnoid cells through porelike open ings and accumulation in front of the dural fibers. Optic nerve sheath hemorrhage was found only in the cats infused with blood, not in those infused with contrast medium.
DISCUSSION CONCERNING THE TECHNICAL ASPECTS OF THE EXPERI-
ments, it is to be questioned whether the increase in intracranial pressure during cisternal infusion reliably simulates the clinical conditions of aneurysmal rup ture and subarachnoid hemorrhage in patients. The changes of intracranial pressure as recorded after aneurysmal rupture in humans are unique: the intra cranial pressure jumps within seconds to the level of the arterial blood pressure. After some minutes, the pressure returns spontaneously to baseline levels in those patients surviving this critical phase of sub arachnoid hemorrhage. Comparing the published intracranial pressure recordings6,7 with those of the present experiments, it is obvious that the model closely simulates the intracranial pressure increase after aneurysmal rupture in humans. 8 This is true for the injections of the animals with both x-ray contrast medium and blood. Therefore, our model is valuable
AMERICAN JOURNAL OF OPHTHALMOLOGY
SEPTEMBER
1997
FIGURE 3. (Left) Transmission electron microscopic image of the normal optic nerve sheath meninges of a control cat shows the thickest and outermost dural layer (D) and the arachnoid cell layer (A). The dural layer consists predominantly of collagenous fibers. The arachnoid portion is formed by cells that are densely packed in the outer layer and more loosely arranged in the direction of the subarachnoid space (X 6,000). (Right) With scanning electron microscopy, the surface of normal arachnoid cell layers can be observed. The intact arachnoid layer has a smooth cell surface with cells closely joined. The cell borders (arrows) are visible. N = nucleus (X 7,200).
FIGURE 4. (Left) Transmission electron microscopic image showing the appearance of the meninges after high-pressure cisternal infusion with contrast medium. The arachnoid cells (asterisk) are flattened and scattered, thus widening the intercellular spaces (arrows, extracellular spaces); F = collagenous fibers (X 16,000). (Right) Appearance of the meninges after high-pressure cisternal infusion with contrast medium. Scanning electron microscopy discloses numerous round, porelike openings (asterisks) of the arachnoid cell surface (diameter, 5 to 10 |JLm). Following the cell borders (arrows), it is obvious that these porelike openings are intracellular (i) as well as intercellular (asterisks). N = nucleus (X 7,000). VOL.124, No. 3
BREAKDOWN OF THE MENINGEAL BARRIER
377
FIGURE 5. (Top left) Optic nerve sheath after cisternal infusion with blood. Transmission electron microscopic image of an erythrocyte (E) within the arachnoid cell layers (arrowheads) passing between collagenous filaments (F) (X 20,000). (Top right) Optic nerve sheath after cisternal infusion with blood. Transmission electron miscroscopic image of an erythrocyte (E) passing through the porelike intercellular opening of arachnoid cells. The irregular edges of the pore indicate a mechanical disruption of the arachnoid cell layer as a result of the sudden increase in intracranial pressure, with consecutive stretching of the optic nerve sheath. Cell borders are indicated by arrows (X 14,000). (Bottom) Optic nerve sheath after cisternal infusion with blood. After crossing the arachnoid layers, the blood cells accumulate in front of the filterlike fiber system of the dura (D) and form a "subdural" blood clot (C) lying within and between the arachnoid cell layers in the subdural space. SAS = subarachnoid space (X800).
for the interpretation of clinical findings after subarachnoid hemorrhage. The radiographs show an outflow of cerebrospinal fluid containing contrast medium into the orbit in every animal. The outflow is not limited to the retrobulbar portion but comprises the complete intraorbital portion of the nerve. Therefore, the experi ments prove the breakdown of the meningeal barrier surrounding the intraorbital optic nerve in the living animal 2 minutes after starting the cisternal infusion. Until now, clinical cisternographies had shown only the opacification of the optic nerve sheath, 378
not leakage of contrast medium into the orbit.9'11 This discrepancy can be explained by the increased intracranial pressure in our experiments. All the previous studies mentioned were done at normal intracranial pressure. Therefore, we suggest that the leakage of contrast medium into the orbit can be attributed to an increased intracranial pressure during cisternography. Anatomically, the optic nerve sheath is continuous with the dura mater and extends the intracranial subarachnoid space into the orbit. The role of the meninges as a barrier to the exchange of fluid and of
AMERICAN JOURNAL OF OPHTHALMOLOGY
SEPTEMBER 1997
particles is well known and has been examined by electron microscopic and ultrastructural horseradishperoxidase histochemical techniques. The meningeal barrier has been shown to exist at the "arachnoid barrier layer" located within multiple layers of cells forming the arachnoid membrane between the dura and subarachnoid space. The arachnoid barrier layer prevents the transfer of fluid because it consists of cells connected by a dense and continuous system of tight junctions.12 Our observation confirms the exper iments of Butler and associates,5 who studied convexi ty meninges of cats with the transmission electron microscopy technique after an increased intracranial pressure of up to 50 mm Hg. The authors injected saline and horseradish-peroxidase solutions into the subarachnoid space, and the tracer was found beyond the arachnoid barrier layer in the extracellular space of the adjacent dura. Corresponding to the present results, a widening of the extracellular spaces between arachnoid cell layers was found secondary to the intracranial pressure increase. The ultrastructural cellular mechanism of the suggested meningeal barri er disruption, however, was not shown in that investi gation.5 In the present experiments, intercellular and mainly intracellular, porelike openings were found to be the cellular mechanism allowing the transmeningeal passage of fluid and particles. Similar pores were found in cells at the angle of the anterior chamber for the outflow of aqueous humor in the eye and in cells of the arachnoid villi for cerebrospinal fluid outflow. Those pores, described as openings of transcellular channels occurring at in creased intraocular or intracranial pressure,13 are in contrast to those in our findings. They are considered to be the ultrastructural correlate of a pressuredependent transcellular bulk flow mechanism for the outflow of aqueous humor or cerebrospinal fluid.13 Because channels related to the observed porelike openings of the cells could not be demonstrated in our experiments, it is unlikely that the present ultrastructural findings can be compared with the ultrastructure of the aqueous chamber or of the arachnoid villi. Considering the radiologic findings, the hypothesis we favor is that the observed porelike openings result from the intravital rupture of arachnoid cells, proba bly because of the distention of the optic nerve sheath caused by increased intracranial pressure. The VOL.124, No. 3
distention of the optic nerve sheath during an in crease in intracranial pressure is well known 14 and might stretch the arachnoid cells, leading to a disrup tion in the observed manner. The pathway that we describe for the outflow of cerebrospinal fluid through the optic nerve is different from that reported in previous ultrastructural studies on the optic nerve sheath architecture.3 Our investi gation demonstrated a zone of increased fluid perme ability of the meninges in the retrobulbar portion of the optic nerve where the optic nerve sheaths join with the sclera. In this region of the nerve, numerous tortuous channels within the meningeal nerve sheath were described, providing a physiological outflow pathway for cerebrospinal fluid. The cerebrospinal fluid outflow pathway as described here is not related to channels. Furthermore, the porelike openings are not limited to the end of the nerve but were found along the entire intraorbital portion of the optic nerve. Considering the experiments of Butler and associates,5 we suggest that meninges, independent of their location inside or outside the intracranial space, have the same behavior when exposed to a significant pressure gradient: the arachnoid cells become flat tened, the extracellular space is widened, and finally the cells are disrupted, showing intracellular and intercellular pores. The pores are the expression of a mechanically disrupted meningeal fluid barrier. In contrast to the results of Butler and associates,5 we found mainly intracellular pores. Whether these pores stay open after increased pressure and how they influence the outflow resistance in the long term remain unclear. According to our experiments, the hypothesis that optic nerve sheath hemorrhage after subarachnoid hemorrhage results from the disruption of intradural and bridging nerve sheath vessels because of the increased intracranial pressure1,2 should be ques tioned. In our experiment, optic nerve sheath hemor rhage is related to the extension of intracranial blood, not to the disruption of meningeal vessels, because no optic nerve sheath hemorrhage was found in those animals infused with contrast medium. Our experimental observations explain why, after subarachnoid hemorrhage in humans, blood clots have often been found to be localized in the subdural rather than in the subarachnoid space of the optic nerve. 2 The subdural blood clots result from the
BREAKDOWN OF THE MENINGEAL BARRIER
379
disruption of the arachnoid barrier layer, which allows blood cells to penetrate from the optic nerve subarachnoid space into cell layers of the nerve sheath to accumulate in front of the filterlike duralfibersystems. The sudden increase in intracranial pressure after aneurysmal rupture and subarachnoid hemorrhage opens the arachnoid barrier layer to enable an outflow of cerebrospinal fluid and even of blood cells across the optic nerve sheath into the orbit. Opening of the optic nerve sheath barrier during subarachnoid hem orrhage may have a protective effect on the optic nerve, similar to surgical nerve sheath fenestration in pseudotumor cerebri patients.15
6. 7. 8.
9.
10. 11.
REFERENCES
12.
1. Garfinkle AM, Danys I, Nicolle D, Colohan A, Brem S. Terson's syndrome: a reversible cause of blindness following subarachnoid hemorrhage. J Neurosurg 1992;76:766-771. 2. Muller P, Deck ]. Intraocular and optic nerve sheath hemorrhage in cases of sudden intracranial hypertension. J Neurosurg 1974;41:160-166. 3. Erlich S, McComb JS, Weiss M. Ultrastructure of the orbital pathway for cerebrospinal fluid drainage in rabbits. ] Neuro surg 1989;70:926-931. 4- Shen J-Y, Kelly D, Hyman S, McComb ]. Intraorbital cerebrospinal fluid outflow and the posterior uveal compart ment of the hamster eye. Cell Tissue Res 1985;240:77-87. 5. Butler A, Van Landingham K, McComb J. Pressure facilitat ed CSF flow across the arachnoid membrane. In: Ishii S,
13. 14. 15.
16.
Nagai H, Brock M, editors. Intracranial pressure. Volume V. Berlin: Springer Verlag, 1983:598-601. Nornes H. Monitoring of patients with intracranial aneurysms. Clin Neurosurg 1975;22:321-331. Volby B, Enevoldsen E. Intracranial pressure changes follow ing aneurysm rupture, 3: recurrent hemorrhage. ] Neurosurg 1982;56:784-789. Brinker T, Dietz H. Computer simulation of the dynamics of the cerebrospinal fluid system and its experimental applica tion. Informatik, Biometrie und Epidemiologie in Medizin und Biologie 1993;24:129-133. Stuck R, Reeves D. Dangerous effects of thorotrast use intracranially with special reference to experimental produc tion of hydrocephalus. Arch Neural Psychiatry 1938;40: 86-115. Chambers E, Manelfe C, Cellerier P. Metrizamide CT cisternography and perioptic subarachnoid space imaging. J Comput Assist Tomogr 1981;5:875-880. Davis J, Haughton V, Harris G, Eldevik O, Gager W. Cat optic nerve imaging with metrizamide. Invest Ophthalmol Vis Sci 1979;18:273-277. Nabeshima S, Reese T, Landis D, Brightman M. Junctions in the meninges and marginal glia. J Comp Neurol 1993;164: 127-170. Tripathi B, Tripathi R. Vacuolar transcellular channels as a drainage pathway for cerebrospinal fluid. J Physiol 1974;239: 195-206. Liu D, Kahn M. Measurement and relationships of subarach noid pressure of the optic nerve to intracranial pressures in fresh cadavers. Am ] Ophthalmol 1993;116:548-556. Spoor TC, Ramocki JM, Madion MP, Wilkinson MJ. Treat ment of pseudotumor cerebri by primary and secondary optic nerve sheath decompression. Am ] Ophthalmol 1991;112: 177-185. Brinker T, Seifert V, Stolke D. Acute changes of cerebrospi nal fluid system dynamics during experimental subarachnoid hemorrhage. Neurosurgery 1990;27:369-372.
Authors Interactive™ Comments and questions regarding this article may be addressed to authors via THE JOURNAL'S Web Site. Selected comments, questions, and author responses are posted on Authors Interactive™ at http://www.ajo.com/
380
AMERICAN JOURNAL OF OPHTHALMOLOGY
SEPTEMBER 1997