Acta neurol. scandinav. 57, 405-417, 1978 Department of Neurology, Sahlgren Hospital, and Institute of Neurobiology, University of Gsoteborg, Goteborg, Sweden
Ultrastructural studies on blood-brain barrier dysfunction around cerebral stab wounds, aggravated by acute ethanol intoxication* LENNART I. PERSSON, LARSE. ROSENGREN AND HANS-ARNE HANSSQN Ethanol-intoxicated and non-intoxicated albino rats subjected to cerebral stab wounds were injected with Evans blue-labelled albumin (EBA) and horseradish peroxidase (HRP) 24 hours after the injury. The tracers were injected 30 and 2 minutes before perfusion fixation with glutaraldehyde and then prepared for light microscopy and electron microscopy. There was prominent leakage of peroxidase through chains of pinocytotic vesicles and transendothelial channels from blood, through the endothelial cells, into the brain, There was no difference in the way of peroxidase leakage through the endothelial cells between ethanol-intoxicated and control rats, but the area containing leaking blood vessels was greater in intoxicated rats. Furthermore, the number of trans-endothelial channels, vesicles and caveolae per unit length oh capillaries and venules was increased in ethanol-intoxicated rats. Injured neurons with a diffuse distribution of peroxidase in their cytoplasm were observed within a greater area around the stab wound in ethanol-intoxicated than in control rats.
Acute ethanol intoxication is known to cause a considerable increase in leakage of Evans blue-labelled albumin from blood into brain after small cerebral stab wounds (Persson .& Rosengren 1977) as well as after cerebral air embolisation (Rosengren et al. 1977). An increased leakage of macromolecular tracers after ethanol intoxication has also been observed around experimental cerebral and spinal contusions (De Crescito et al. 1974, Brodner et al. 1976, Weiss et al. 1976, Flamm et al. 1977). However, the exact cellular mechanisms underlying this aggravation of an already existent blood-brain barrier dysfunction have hithero remained unknown. The present work was performed as an attempt to further elucidate the way of tracer leakage due
*
A preliminary communication of some of the results in this paper was presented at The 11th World Congress of Neurology, Amsterdam, the Netherlands, September 1977.
406 to blood-brain barrier dyfunction around cerebral stab wounds, aggravated by acute ethanol intoxication. MATERIALS AND METHODS Twenty-four male albino rats of the Sprague-Dawley strain with a weight of 150-160g were used. All animals were fed pellets (Astra-Ewos, Sodertalje, Sweden) and water ad libitum. The operation was performed as described by Persson et al. (1976). In short, under ether anaesthesia, a hole with an inner diameter of 1 mm was drilled in the left frontal bone with a dental drill. Dura mater was incised and a steel needle with an outer diameter of 50 p m was inserted vertically 1 mm into area 10 (according to the nomenclature of Krieg 1946), immediately withdrawn, and the skin sutured. Serum ethanol determinations on selected intoxicated and control animals were made by gas chromatography. Twelve rats were injected with ethanol (20 % w/v, 1.5 m1/100 g b.w. in saline) intraperitoneally 1 day after injury, i.e. 1 h before sacrifice. Twelve rats used as controls were injected with the same amount of saline. Thirty minutes prior to sacrifice (1 day after the injury), all animals were injected with EBA (Evans blue 1 % w/v and bovine albumin 5 % in saline, 0.5 ml/100 g b.w.) in the right femoral vein. Eight rats in each group of animals were also injected with horseradish peroidase intravenously (HRP) (Type 11, Sigma Chem. Co., St. Louis, Mo., U.S.A., 50 mg/ml saline, 0.7 m1/100 g b.w.). The injection of EBA and/or HRP was repeated 2 min before perfusion fixation. The rats injected with EBA and HRP were fixed by intraaortal perfusion for 10 min with 3 % purified glutaraldehyde in 0.15 M cacodylate buffer, pH 7.2 a t 20" C. The rats injected with only EBA were fixed by perfusion with 4 % formaldehyde in 0.15 M cacodylate buffer, pH 7.2. After a delay of 15 min, the brain was carefully remowed and fixed by immersion in the same fixative for at least 6 h at 4" C. The injured area was dissected and sectioned in a Vibratome set (Oxford Laboratories, Calif., USA). Vibratome sections for fluorescent microscopy were mounted in 50 % glycerol in water and studied unstained in a Zeiss fluorescence microscope. The specimens of rats injected with HRP were incubated in a solution of 0.05 % 3.3-diaminobenzidine and 0.01 % hydrogen peroxide in 0.15 M cacodylate buffer, p H 7.5, a t 20" C for 45 min, rinsed in buffer and postfixed in a buffered solution of 2 % OsO,. These specimens were then rinsed, dehydrated in a graded series of ethanol and embedded in Epon. Sections for electron microscopy were prepared on an Ultrotome I11 ultramicrotome (LBK, Stockholm, Sweden) and observed unstained in a Siemens Elmiskop IA electron microscope or a Zeiss electron microscope.
Figure 1 . Cerebral stab wound in a control rat injected with EBA. The tracer is spread around the stab wound. Only a few neurons at the edge of the wound display a diffuse cytoplasmic distribution of EBA (arrow), while most neurons show a granular cytoplasmic uptake of EBA (arrow head). X 440. Figure 2. Cerebral stab wound in an ethanol-intoxicated rat injected with EBA. The tracer is spread in a greater area than in Figure 1 and there is a di#use cytoplasmic spread o f EBA in neurons in a large area around the wound (arrow). There is a wide area with a granular cytoplasmic uptake of EBA in neurons (arrow head). X 440.
407
408 RESULTS
Light and fluorescence microscopy. There was leakage of Evans blue-labelled albumin and horseradish peroxidase within 50-100 pm from the edge of the stab wound in control rats (Figure 1). Ethanol-intoxicated rats displayed a leakage of both tracers within 200-250 pm from the edge of the stab wound (Figure 2). Scattered neurons with a diffuse distribution of EBA or peroxidase in their cytoplasm were observed in both intoxicated and non-intoxicated rats in the area with leaking blood vessels. The area observed containing neurons with a diffuse distribution of peroxidase in their cytoplasm was situated within half the radius of the leaking area. Thus, the area containing neurons with a diffuse distribution of peroxidase in their cytoplasm was larger in ethanol-intoxicated than in not intoxicated rats. Electron microscopy. There was a severe swelling of cells and cell processes within 50-100 pm from the edge of the stab wound in control rats and within 200-250 pm in ethanol-intoxicated rats. Rupture of cell membranes, especially of perivascular astrocytic end-feet, was observed within the immediate peritraumatic region (Figures 4-8). Swelling of endothelial cells and vascular basement membranes was observed within 50 pm from the injury in control rats and within 200 pm in ethanol-intoxicated rats. The endothelial swelling observed was much more pronounced in intoxicated than in control rats (Figures 4,5, 7, 8). There was a leakage of horseradish peroxidase in both intoxicated rats arid controls through trans-endothelial channels (diameter 30-70 nm) caveolae (vesicles with a diameter exceeding 70 nm) and pinocytotic vesicles (diameter 30-70 nm) from the lumina of blood vessels, through the endothelial cyto-
Figure 3. Capillary ( C ) 50 p m from the edge o f a stab wound in a control rat injected with HRP. There is a slight spread of the tracer into basement membranes as well as into extracellular spaces and into an injured astrocytic end-foot (A). Next to the stab wound is an accumulation of HRP in the extracellular space (P). There is a slight swelling of the endothelial cytoplasm and a few endothelial pinocytic vesicles are filled with HRP (arrow). Unstained. X 19,000. Figure 4. Capillary ( C ) 50 ,urn from the edge o f a stab wound in an ethanol-intoxicated rat injected with HRP. There is a considerable leakage o f H R P through trans-endothelial channels (arrow) and multiple endothelial pinocytotic vesicles (arrow head) into the capillary basement membrane (bm), which is infiltrated with large amounts of HRP. Next to the stab wound, there is an extravasated erythrocyte ( E ) and a diffuse cytoplasmic spread of H R P in ruptured neuronal ( N ) and astrocytic ( A ) processes. A pen’vascular pericyte ( P ) shows no uptake of HRP at this tracer circulation time. All cell processes as well as the endothelial cells are severely swollen. Unstained. X 29,000.
409
27 Acta neurol. scandinav. 5 7 5
4 10
plasm into the periendothelial vascular basement membrane. Pinocytotic vesicles often formed continuous chains from the ad-luminal cell membrane, through the endothelial cytoplasm, to the basement membrane. Serial sections revealed that many of these chains of vesicles comprised sections of tortuous trans-endothelial channels, filled with peroxidase in their entire length, from the vessel lumen to the basement membrane. The vascular basement membranes and surrounding extracellular space between neuronal and glial processes were filled with peroxidase in areas with peroxidase-containing transendothelial channels. A few endothelial cells, severely injured by the stab wound, displayed a diffuse distribution of peroxidase in the endothelial cytoplasm. In controls, the leakage was predominantly encountered in arterioles, met-arterioles, capillaries and scattered venules, while in ethanol-intoxicated rats a proportionally larger number of the leaking vessels was encountered among capillaries and venules. Additionally, in ethanol-intoxicated animaIs, the number of vesicles per unit vessel length was increased in capillaries and venules, compared to corresponding vessels in controls. Furthermore, the area containing trans-endothelial channels was much
Figure 5. Capillary (C) 100 ,um from the edge of the stab wound in an ethanol-intoxicared rat injected with H R P . There is a considerable leakage of H R P through chains of endothelial pinocytic vesicles and caveolae {arrow] into adjacent basement membranes and extracellular spaces. Several injured neuronal ( N ) and astrocytic ( A ) processes show a diffuse cytoplasmic spread of HRP in the cytoplasm. Most cell processes as well as the endothelium ( E ) are severely swollen. Occasional pinocytic uptake of H R P is observed in a f e w cell processes (arrow head) in spite o f the cell injury. Unstained. X 7,600. Figure 6. Venule 50 p m from the edge o f a stab wound in a control rat injected with H R P . Several aggregates of pinocytic vesicles and chains of vesicles (arrow) are observed in the endothelial cytoplasm (E). T w o caveolae filled with H R P joined to pinocytotic vesicles are observed. Small amounts of H R P are observed in the basement membranes. There is a diffuse cytoplasmic spread of H R P in an adjacent astrocytic process (A). Unstained. X 20,000. Figure 7. Venule 50 p n ~ from the edge of a stab wound in an ethanol-intoxicated rat injected with H R P . Multiple pinocytotic vesicles and chains of vesicles (arrow) are observed in the endothelial cytoplasm. There is a rnoderat infiltration of adjacent basement memebranes and some underlying cell processes with H R P . Both endothelium ( E ) and a pericyte ( P ) as well as surrounding astroglial processes ( A ) appear swollen. Unstained. X 22,100. Figure 8. Venule I00 ,urn f r o m the edge o f a stab wound in an ethanol-intoxicated rat injected with H R P . Several pinocytotic vesicles (arrow) and a trans-endothelial channel (arrow head) are observed in the endothelial cytoplasm. There is a slight infiltration of adjacent basement membranes and neuropil with H R P . Unstained. X 19,000.
27*
412 larger in ethanol-intoxicated than in non-intoxicated rats, as was the area containing neurons with a diffuse distribution of peroxidase in their cytoplasm. In both groups, most neurons surrounding the leaking area showed an uptake of peroxidase in pinocytotic vesicles as did a few astrocytes and oligodendrocytes. This majority of cells contained no diffuse distribution of peroxidase, and these cells were in general situated at a greater distance from the edge of the stab wound than those with a diffuse cytoplasmic distribution of peroxidase. No difference in the way of peroxidase leakage from blood, through the endothelial cells, into brain parenchyma was observed in ethanol-intoxicated compared to non-intoxicated rats. However, in intoxicated rats: 1) the area containing leaking blood vessels was greater, as was the area containing neurons with a diffuse distribution of peroxidase, 2) a proportionally larger number of the leaking vessels was encountered among capillaries and venules, and 3) the number of transendothelial channels per unit vessel length was increased. DISCUSSION
Passage of macromolecules from blood into brain under normal conditions is prevented by the vascular endothelial cells, which are joined to neighboring endothelial cells by a continuous rim of tight junctions. These cells, joined by tight junctions, permit water and solutes (ions, amino acids, glucose) to enter the brain with variable restriction during the passage across the bloodbrain barrier. Proteins do not pass the endothelial lining in significant amounts under normal conditions (Ehrlich 1885, Goldmann 1913, Broman 1949, Tschirgi 1950, Reese & Karnovsky 1967), due to the absence of transendothelial channels, the scarcity of endothelial pinocytosis in cerebral endothelial cells and the presence of interendothelial tight junctions (Reese & Karnovsky 1967). However, some pinocytosis is observed even in normal conditions in cerebral endothelium, especially in arterioles, but this pinocytosis allows only passage of small amounts of macromolecules from blood into brain (Westergaard & Brightman 1973, NGrtved & Westergaard 1974, Westergaard 1975). Under certain pathological conditions, there is a passage of macromolecules from blood into brain (Broman 1949, Brightman et al. 1970). Great effort has been made to show the way of passage of macromolecular tracers in various conditions (for review see Manz 1974). However, the exact route of macromolecules in many conditions still remains unclear. Possible ways of leakage of macromolecules across the blood-brain barrier are rupture of the endothelial cell membrane, widening of the interendothelial tight junctions, diffusion across the endothelial cell membranes and cytoplasm
413 or passage by pinocytosis or trans-endothelial channels (Reese & Karnovsky 1967. Brightman & Reese 1969, Hirano et al. 1969, Brigthman et al. 1970, Clasen et al. 1970). Pinocytotic transport of proteins across the blood-brain barrier is known to occur in a variety of conditions (e.g. Hansson et al. 1975, Beggs & Waggener 1976, Persson & Hansson 1976, Nag et al. 1977, Petito et al. 1977). The leakage of macromolecular tracers around cerebral wounds might be due to an increased pinocytotic transport and a formation of trans-endothelial channels. Pinocytotic vesicles, caveolae as well as trans-endothelial channels filled with peroxidase, have been shown around cerebral stab wounds (Persson & Hansson 1976) as well as in experimental spinal cord compression (Beggs & Waggener 1976). In the present work, the mechanism of extravasation of peroxidase was by passage through trans-endothelial channels and through chains of pinocytotic vesicles, forming channels. These channels and chains of vesicles, as well as scattered pinocytotic vesicles and caveolae, were observed within the area showing extravasation of peroxidase into vascular basement membranes and extracellular space around cells and cell processes. However, the number of trans-endothelial channels, pinocytotic vesicles and caveolae was enhanced in ethanol-intoxicated rats. The latter finding was consistent with the finding in a preceding study (Persson & Rosengren 1977), where it was shown that the amount of extravasated Evans blue-labelled albumin was increased around cerebral stab wounds in ethanol-intoxicated rats. Furthermore, it was shown that the area containing leaking blood vessels was greatly increased in ethanol-intoxicated compared to control rats. Such an increase of the leaking area has previously been observed to Evans blue-labelled albumin around cerebral contusions (DeCrescito et al. 1974, Flamm et al. 1977). The mechanisms behind this increased leakage of peroxidase around cerebral stab wound is unknown. Ethanol and its metabolite acetaldehyde affect biogenic amine metabolism (Duritze & Truitt 1966, Kalant 1975, von Wartenburg et al. 1975), and biogenic amines, especially serotonin, are known to affect the blood-brain barrier permeability to macromolecular tracers (Cotran & Karnowsky 1967, Westergaard 1975). Ethanol is also known to dilatate cerebral blood vessels (Weiss et al. 1976). This might give an increased pressure load on cerebral capillaries and venules, making them more sensitive to injury. Another mechanism is the effect of ethanol and acetaldehyde on biological membranes. Both ethanol and acetaldehyde inhibit Na+-K-ATPase in cell memebranes, which results in an accumulation of ATP, a decrease in the influx of K into cells and a decreased O2 consumption (Jiirnefelt 1961, Wallgren 1966, Israel & Salazar 1967, Israel 1970, Nielsen et al. 1975). This
414 effect is reflected by a reversible depressant effect of ethanol on the electrical activity of, e.g. squid giant axon (Armstrong 1964, Moore et al. 1964), rat cerebral cortex (Nikander et al. 1971) and mouse cerebral and cerebellar neuronal cultures (Seil et al. 1977). This depressant effect of ethanol on the activity of biological membranes can immediately be reversed by rinsing the biological membranes with ethanol-free solutions (Seil et al. 1977) and is thus likely to be a direct effect of ethanol on membranes rather than mediated by biogenic amines. One effect of ethanol and its principal metabolite acetaldehyde on biological membranes has been considered the ability of liberating free radicals in membranes (DiLuzio & Hartman 1967, 1969). These free radicals might interfere with membrane phospholipids, and thereby with Na+-K adenosine triphosphatase in cell membranes (Kimelberg & Papahadjopoulos 1974). This interaction of ethanol with biological membrane functions might explain the observation in this work that a greater number of injured neurons in the vicinity of the stab wound are unable to maintain their integrity to foreign proteins in ethanol-intoxicated rats, which results in a free passage of peroxidase through injured membranes into cytoplasm, where the peroxidase molecules are distributed in a diffuse pattern among the neuronal organelles. Functionally, this might mean that a smaller number of neurons have the possibility to recover after a moderate to severe injury in ethanol-intoxicated compared to control rats. That conclusion is further supported by the finding by Brodner et al. (1976) and by Flamm et al. (1977) that functional recovery after spinal contusions was rare in cats which were ethanol-intoxicated at injury, compared to non-intoxicated ones. Furthermore, the microscopical integrity of the spinal cord was fairly preserved at 6-8 weeks after injury in controls, while the ethanol-intoxicated group showed a severe atrophic scar at the impact area in the spinal cord (Flamm et al. 1977). There was a potentiation of the blood-brain barrier disturbances around cerebral stab wounds in ethanol-intoxicated rats, compared to controls. Endothelial cells showed an increase in trans-endothelial channels, pinocytotic vesicles and caveolae. Injured neurons in the vicinity of the stab wound showed a loss of their integrity to foreign proteins, with a trans-membraneous leakage of peroxidase into cytoplasm in a larger number of cells in ethanolintoxicated rats, compared to controls. Both the endothelial and the neuronal membrane effects of ethanol might be due to interference with biogenic amine metabolism, an increased formation of free radicals or due to other effects on biological membranes.
415 ACKNOWLEDGEMENTS Our sincere thank to Mrs. Barbro Eriksson and Mrs. Ulla Svedin for skilled technical assistance. Dr. Barbro Johansson in thanked for helpful discussions. This work was supported by grants from the Trygg-Hansa (Fylgia) Foundation, Tore Nilson's Foundation for Medical Research, the Swedish Medical Research Council and the Medical Faculty of the University of Goteborg.
REFERENCES Armstrong, C . L. L. Binstock (1964): The effects of several alcohols on the properties of the squid giant axon. J. Gen. Physiol. 48, 265-277. Beggs, J. L. & J. D. Waggener (1976): Transendo~thelialvesicular transport of pro'tein following compression injury to the spinal cord. Lab. Invest. 34, 428-439. Brightman, M. W. (&T. S. Reese (1969): Junctions between intimately opposed cell membranes in the vertebrate brain. J. Cell. Biol. 40, 648-677. Brightman, M. W., I. Klatzo, Y. Olsson Lk T. S. Reese (1970): The blood-brain barrier to protein tracers under normal and pathological conditions. J. Neurol. Sci. I 0 , 2 15-239. Brodner, R. A,, J. C. Van Gilder, W. F. Collin0 Jr, ,& B. Hamill (1976): Experimental spinal cord trauma: The efficacy of treatment. 44th Annual Meeting of the American Association of Neurological Surgeons. San Francisco. Broman, T. (1949): The permeability of the cerebrospinal vessels in normal and pathological conditions. Munksgaard, Copenhagen. Clasen, R. A., S. Pandolfi & G. M. Hass (1970): Vital staining, serum albumin and the blood-brain barrier. J. Neuropath. Exp. Neurol. 29, 266-284. Cotran, R. S. & M. J. Karnovsky (1976): Vascular leakage induced by horseradish peroxidase in the rat. Proc. SOC.Exp. Biol. (N.Y.) 126, 557-561. I. J. Ransohoff (1974): Ethanol poDeCrescito, V., H. G . Demopoulos, E. S. Flamm tentiation of traumatic cerebral edema. Surg. Forum 25, 438-440. DiLuzio, N. R. A. D. Hartman (1967): Role of lipid peroxidation in the pathogenesis of the ethanol-induced fatty liver. Fed. Proc. 26, 1436-1442. DiLuzio, N. R. .& A. D. Hartman (1969): The effect of ethanol and carbon tetrachloride administration on hepatic lipid-soluble antioxidant activity. Exp. Mol. Pathol. II, 38-52. Duritz, G. ,& E. B. Truitt, Jr (1966): Importance of acetaldehyde in the action of ethanol on brain norepinephrine and 5-hydroxytryptamine. Biochem. Pharmacol. 15, 711-721. Ehrlich, P. (1885): Das Sauerstoff-Bediirfnis des Organismus. Eine farbenanalytische Studie. Hirschwald, Berlin. Goldmann, E. E. (1913): Vitalfarbung am Zentralnervensystem. Beitrag zur PhysioPathologie des Plexus choroideus und der Hirnhaute. Abh. K8nigl. Preuss. Akad. Wissensch. No. 1. Hanssm, H.-A., B. Johansson & C. Blomstrand (1975): Ultrastructural studies o'n cerebrovascular permeability in acute hypertension. Acta Neuropath. (Berl). 32, 187-198. A. M. Zimmerman (1969): Pathological alterations in the Hirano, A,, N. H. Becker cerebral endothelial cell barrier to peroxidase. Arch. Neurol. (Chic.) 20, 300-308. Flamm, E. S., H. B. Demopoulos, M. L. Seligman, J. J. Tomasula, V. DeCrescito .& J. Ransohoff (1977): Ethanol potentiation of central nervous system trauma. J. Neurosurg. 46, 328-335. (&
(&
I&
416 Israel, Y . (1970): Cellular effects of alcohol. Quart. J . Stud. Alcohol. 31, 293-315. Israel, Y . (& I. Salazar (1967): Inhibition of brain microsomal adenosine triphosphatases by general depressants. Arch. Biochem. Biophys. 122, 310-317. JLrnefelt, J. (1961): Inhibition of the brain microsomal adenosine-triphosphatase by depolarizing agents. Biochem. Biophys. Acta 48, 111-116. Kalant, H. (1975): Direct effects of ethanol on the central nervous system. Fed Proc. 34, 1930-1941. Kimelberg, H. #&D. Papahadjopoulos (1974): Effects of phospholipid acyl chain fluity, phase transitious and cholesterol on (Na' K+)- stimulated adenosine triphosphatase. J. Biol. Chem. 249, 1071-1080. Krieg, W. J. S. (1946): Connection of the cerebral cortex. I. The albino rat. A. Topography of the cortical areas. J. Comp. Neurol. 84, 221-275. Majno, G. #&G. E. Palade (1961): Studies on inflammation. Part. 1. (The effects of histamine and serotosnin on vascular permeability: an electron microscopic study). J. Biophys. Biochem. Cytol. I I , 571-606. Manz, H. J. (1974): The pathology of cerebral edema. Human Path. 5, 291-313. Moore, J. W., W. Ulbricht ,& M. Takata (1964): Effect of ethanol on sodium and potassium conductances of the squid axon membrane. J. Gen. Physiol. 48, 279-295. Nag, S., D . M . Robertson & H.B. Dinsdale (1977): Cerebral cortical changes in acute experimental hypertension. An ultrastructural study. Lab. Invest. 36, 150-161. Nielsen, R. H., R. A. Hawkins J&R. L. Veech (1975): The effects of acute ethanol intoxication on cerebral energy metabolism. Adv. Exp. Med. Biol. 59, 93-109. Nikander, P., P. von Boguslawsky & H. Wallgren (1971): Actions of ethanol and tertiary butanol on net movements of sodium and potassium in electrically stimulated rat brain tissue in vitro. Scand. J. Clin. Lab. Invest. Suppl. 27, 116, 67. NGrtved, J. ,& E. Westergaard (1974) Transport of horseradish peroxidase across cerebral arteroles and venules. Anat. Rec. 178, 428. Persson, L. & H.-A. Hansso'n (1976): Reversible blood-brain barrier dysfunction to' peroxidase after a small stab wound in the rat cerebral cortex. Acta Neuropath. (Berl.) 35, 333-342. Persson, L. & L. Rosengren (1977): Increased blood-brain barrier permeability around cerebral stab wounds, aggravated by acute ethanol intoxication. Acta Neurol. Scand. 56, 7-16. Petito, C. K., J. A. Schaefer & F. Plum (1977): Ultrastructural characteristics of the brain and blood-brain barrier in experimental seizures. Brain Research. 127, 251-267. Raimondi, A. J., J. P. Evans ,& S. Mullan (1962): Studies on cerebral edema. Part 3 (Alterations in the white matter. An electron microscopy study using ferritin as a labeling compound). Acta Neuropath. (Berl.) 2, 177-197. Reese, T. S. & M. J. Karnovsky (1967): Fine structural localization of a blo'od-brain barrier to exogenous peroxidase. J. Cell. Biol. 34, 207-217. Rosengren, L., L. Persson ,& B. Johansson (1977): Enhanced blood-brain barrier leakage to Evans blue-labelled abumin after air embolism in ethanol-intoxicated rats. Acta Neuropath. (Berl.) 38, 149-152. Seil, F. J., A. L. Leiman, M. M. Herman & R. A. Fisk (1977): Direct effects of ethanol on central nervous system cultures: An electrophysiological and morphological study. Exp. Neurol. 55, 390-404. Tschirgi, R. D. (1950): Protein complexes and the impermability of the blood-brain barrier to dyes. Amer J. Physiol. 163, 756. Wallgren, H. (1966): Effects of alcohol on biochemical processes in the central nervous system. Psychosom. Med. 28, 431-442.
+
417 von Wartenburg, J. P., D. Berger, M. M. Ris & B. Tabakoff (1975): Enzymes of biogenic aldehyde metabolism. Adv. Exp. Med. Biol. 59, 119-138. Weiss, M.H., J.S. Heiden, M.L. J. Apuzzo & T. Kurze (1976): The effects of acute ethanol intoxication on intracranial physical dynamics. 44th Annual Meeting of the American Association of Neurological Surgeons. San Francisco. Westergaard, E. (1975): Enhanced vesicular transport of exogenous peroxidase across cerebral vessels induced by serotonin. Acta Neuropath. (Berl.) 32, 27-42. Westergaard, E. M. W. Brightman (1973): Transport of proteins across normal cerebral arterioles. J. Comp. Neur. 152, 17-44. (&
Received March 20, accepted April 5, 1978
Dr. Lennart I . Persson Department of Neurology SahIgren Hospital S-413 45 Goteborg Sweden
Ultrastructural studies on blood-brain barrier dysfunction around cerebral stab wounds, aggravated by acute ethanol intoxication* LENNART I. PERSSON, LARSE. ROSENGREN AND HANS-ARNE HANSSQN Ethanol-intoxicated and non-intoxicated albino rats subjected to cerebral stab wounds were injected with Evans blue-labelled albumin (EBA) and horseradish peroxidase (HRP) 24 hours after the injury. The tracers were injected 30 and 2 minutes before perfusion fixation with glutaraldehyde and then prepared for light microscopy and electron microscopy. There was prominent leakage of peroxidase through chains of pinocytotic vesicles and transendothelial channels from blood, through the endothelial cells, into the brain, There was no difference in the way of peroxidase leakage through the endothelial cells between ethanol-intoxicated and control rats, but the area containing leaking blood vessels was greater in intoxicated rats. Furthermore, the number of trans-endothelial channels, vesicles and caveolae per unit length oh capillaries and venules was increased in ethanol-intoxicated rats. Injured neurons with a diffuse distribution of peroxidase in their cytoplasm were observed within a greater area around the stab wound in ethanol-intoxicated than in control rats.
Acute ethanol intoxication is known to cause a considerable increase in leakage of Evans blue-labelled albumin from blood into brain after small cerebral stab wounds (Persson .& Rosengren 1977) as well as after cerebral air embolisation (Rosengren et al. 1977). An increased leakage of macromolecular tracers after ethanol intoxication has also been observed around experimental cerebral and spinal contusions (De Crescito et al. 1974, Brodner et al. 1976, Weiss et al. 1976, Flamm et al. 1977). However, the exact cellular mechanisms underlying this aggravation of an already existent blood-brain barrier dysfunction have hithero remained unknown. The present work was performed as an attempt to further elucidate the way of tracer leakage due
*
A preliminary communication of some of the results in this paper was presented at The 11th World Congress of Neurology, Amsterdam, the Netherlands, September 1977.
406 to blood-brain barrier dyfunction around cerebral stab wounds, aggravated by acute ethanol intoxication. MATERIALS AND METHODS Twenty-four male albino rats of the Sprague-Dawley strain with a weight of 150-160g were used. All animals were fed pellets (Astra-Ewos, Sodertalje, Sweden) and water ad libitum. The operation was performed as described by Persson et al. (1976). In short, under ether anaesthesia, a hole with an inner diameter of 1 mm was drilled in the left frontal bone with a dental drill. Dura mater was incised and a steel needle with an outer diameter of 50 p m was inserted vertically 1 mm into area 10 (according to the nomenclature of Krieg 1946), immediately withdrawn, and the skin sutured. Serum ethanol determinations on selected intoxicated and control animals were made by gas chromatography. Twelve rats were injected with ethanol (20 % w/v, 1.5 m1/100 g b.w. in saline) intraperitoneally 1 day after injury, i.e. 1 h before sacrifice. Twelve rats used as controls were injected with the same amount of saline. Thirty minutes prior to sacrifice (1 day after the injury), all animals were injected with EBA (Evans blue 1 % w/v and bovine albumin 5 % in saline, 0.5 ml/100 g b.w.) in the right femoral vein. Eight rats in each group of animals were also injected with horseradish peroidase intravenously (HRP) (Type 11, Sigma Chem. Co., St. Louis, Mo., U.S.A., 50 mg/ml saline, 0.7 m1/100 g b.w.). The injection of EBA and/or HRP was repeated 2 min before perfusion fixation. The rats injected with EBA and HRP were fixed by intraaortal perfusion for 10 min with 3 % purified glutaraldehyde in 0.15 M cacodylate buffer, pH 7.2 a t 20" C. The rats injected with only EBA were fixed by perfusion with 4 % formaldehyde in 0.15 M cacodylate buffer, pH 7.2. After a delay of 15 min, the brain was carefully remowed and fixed by immersion in the same fixative for at least 6 h at 4" C. The injured area was dissected and sectioned in a Vibratome set (Oxford Laboratories, Calif., USA). Vibratome sections for fluorescent microscopy were mounted in 50 % glycerol in water and studied unstained in a Zeiss fluorescence microscope. The specimens of rats injected with HRP were incubated in a solution of 0.05 % 3.3-diaminobenzidine and 0.01 % hydrogen peroxide in 0.15 M cacodylate buffer, p H 7.5, a t 20" C for 45 min, rinsed in buffer and postfixed in a buffered solution of 2 % OsO,. These specimens were then rinsed, dehydrated in a graded series of ethanol and embedded in Epon. Sections for electron microscopy were prepared on an Ultrotome I11 ultramicrotome (LBK, Stockholm, Sweden) and observed unstained in a Siemens Elmiskop IA electron microscope or a Zeiss electron microscope.
Figure 1 . Cerebral stab wound in a control rat injected with EBA. The tracer is spread around the stab wound. Only a few neurons at the edge of the wound display a diffuse cytoplasmic distribution of EBA (arrow), while most neurons show a granular cytoplasmic uptake of EBA (arrow head). X 440. Figure 2. Cerebral stab wound in an ethanol-intoxicated rat injected with EBA. The tracer is spread in a greater area than in Figure 1 and there is a di#use cytoplasmic spread o f EBA in neurons in a large area around the wound (arrow). There is a wide area with a granular cytoplasmic uptake of EBA in neurons (arrow head). X 440.
407
408 RESULTS
Light and fluorescence microscopy. There was leakage of Evans blue-labelled albumin and horseradish peroxidase within 50-100 pm from the edge of the stab wound in control rats (Figure 1). Ethanol-intoxicated rats displayed a leakage of both tracers within 200-250 pm from the edge of the stab wound (Figure 2). Scattered neurons with a diffuse distribution of EBA or peroxidase in their cytoplasm were observed in both intoxicated and non-intoxicated rats in the area with leaking blood vessels. The area observed containing neurons with a diffuse distribution of peroxidase in their cytoplasm was situated within half the radius of the leaking area. Thus, the area containing neurons with a diffuse distribution of peroxidase in their cytoplasm was larger in ethanol-intoxicated than in not intoxicated rats. Electron microscopy. There was a severe swelling of cells and cell processes within 50-100 pm from the edge of the stab wound in control rats and within 200-250 pm in ethanol-intoxicated rats. Rupture of cell membranes, especially of perivascular astrocytic end-feet, was observed within the immediate peritraumatic region (Figures 4-8). Swelling of endothelial cells and vascular basement membranes was observed within 50 pm from the injury in control rats and within 200 pm in ethanol-intoxicated rats. The endothelial swelling observed was much more pronounced in intoxicated than in control rats (Figures 4,5, 7, 8). There was a leakage of horseradish peroxidase in both intoxicated rats arid controls through trans-endothelial channels (diameter 30-70 nm) caveolae (vesicles with a diameter exceeding 70 nm) and pinocytotic vesicles (diameter 30-70 nm) from the lumina of blood vessels, through the endothelial cyto-
Figure 3. Capillary ( C ) 50 p m from the edge o f a stab wound in a control rat injected with HRP. There is a slight spread of the tracer into basement membranes as well as into extracellular spaces and into an injured astrocytic end-foot (A). Next to the stab wound is an accumulation of HRP in the extracellular space (P). There is a slight swelling of the endothelial cytoplasm and a few endothelial pinocytic vesicles are filled with HRP (arrow). Unstained. X 19,000. Figure 4. Capillary ( C ) 50 ,urn from the edge o f a stab wound in an ethanol-intoxicated rat injected with HRP. There is a considerable leakage o f H R P through trans-endothelial channels (arrow) and multiple endothelial pinocytotic vesicles (arrow head) into the capillary basement membrane (bm), which is infiltrated with large amounts of HRP. Next to the stab wound, there is an extravasated erythrocyte ( E ) and a diffuse cytoplasmic spread of H R P in ruptured neuronal ( N ) and astrocytic ( A ) processes. A pen’vascular pericyte ( P ) shows no uptake of HRP at this tracer circulation time. All cell processes as well as the endothelial cells are severely swollen. Unstained. X 29,000.
409
27 Acta neurol. scandinav. 5 7 5
4 10
plasm into the periendothelial vascular basement membrane. Pinocytotic vesicles often formed continuous chains from the ad-luminal cell membrane, through the endothelial cytoplasm, to the basement membrane. Serial sections revealed that many of these chains of vesicles comprised sections of tortuous trans-endothelial channels, filled with peroxidase in their entire length, from the vessel lumen to the basement membrane. The vascular basement membranes and surrounding extracellular space between neuronal and glial processes were filled with peroxidase in areas with peroxidase-containing transendothelial channels. A few endothelial cells, severely injured by the stab wound, displayed a diffuse distribution of peroxidase in the endothelial cytoplasm. In controls, the leakage was predominantly encountered in arterioles, met-arterioles, capillaries and scattered venules, while in ethanol-intoxicated rats a proportionally larger number of the leaking vessels was encountered among capillaries and venules. Additionally, in ethanol-intoxicated animaIs, the number of vesicles per unit vessel length was increased in capillaries and venules, compared to corresponding vessels in controls. Furthermore, the area containing trans-endothelial channels was much
Figure 5. Capillary (C) 100 ,um from the edge of the stab wound in an ethanol-intoxicared rat injected with H R P . There is a considerable leakage of H R P through chains of endothelial pinocytic vesicles and caveolae {arrow] into adjacent basement membranes and extracellular spaces. Several injured neuronal ( N ) and astrocytic ( A ) processes show a diffuse cytoplasmic spread of HRP in the cytoplasm. Most cell processes as well as the endothelium ( E ) are severely swollen. Occasional pinocytic uptake of H R P is observed in a f e w cell processes (arrow head) in spite o f the cell injury. Unstained. X 7,600. Figure 6. Venule 50 p m from the edge o f a stab wound in a control rat injected with H R P . Several aggregates of pinocytic vesicles and chains of vesicles (arrow) are observed in the endothelial cytoplasm (E). T w o caveolae filled with H R P joined to pinocytotic vesicles are observed. Small amounts of H R P are observed in the basement membranes. There is a diffuse cytoplasmic spread of H R P in an adjacent astrocytic process (A). Unstained. X 20,000. Figure 7. Venule 50 p n ~ from the edge of a stab wound in an ethanol-intoxicated rat injected with H R P . Multiple pinocytotic vesicles and chains of vesicles (arrow) are observed in the endothelial cytoplasm. There is a rnoderat infiltration of adjacent basement memebranes and some underlying cell processes with H R P . Both endothelium ( E ) and a pericyte ( P ) as well as surrounding astroglial processes ( A ) appear swollen. Unstained. X 22,100. Figure 8. Venule I00 ,urn f r o m the edge o f a stab wound in an ethanol-intoxicated rat injected with H R P . Several pinocytotic vesicles (arrow) and a trans-endothelial channel (arrow head) are observed in the endothelial cytoplasm. There is a slight infiltration of adjacent basement membranes and neuropil with H R P . Unstained. X 19,000.
27*
412 larger in ethanol-intoxicated than in non-intoxicated rats, as was the area containing neurons with a diffuse distribution of peroxidase in their cytoplasm. In both groups, most neurons surrounding the leaking area showed an uptake of peroxidase in pinocytotic vesicles as did a few astrocytes and oligodendrocytes. This majority of cells contained no diffuse distribution of peroxidase, and these cells were in general situated at a greater distance from the edge of the stab wound than those with a diffuse cytoplasmic distribution of peroxidase. No difference in the way of peroxidase leakage from blood, through the endothelial cells, into brain parenchyma was observed in ethanol-intoxicated compared to non-intoxicated rats. However, in intoxicated rats: 1) the area containing leaking blood vessels was greater, as was the area containing neurons with a diffuse distribution of peroxidase, 2) a proportionally larger number of the leaking vessels was encountered among capillaries and venules, and 3) the number of transendothelial channels per unit vessel length was increased. DISCUSSION
Passage of macromolecules from blood into brain under normal conditions is prevented by the vascular endothelial cells, which are joined to neighboring endothelial cells by a continuous rim of tight junctions. These cells, joined by tight junctions, permit water and solutes (ions, amino acids, glucose) to enter the brain with variable restriction during the passage across the bloodbrain barrier. Proteins do not pass the endothelial lining in significant amounts under normal conditions (Ehrlich 1885, Goldmann 1913, Broman 1949, Tschirgi 1950, Reese & Karnovsky 1967), due to the absence of transendothelial channels, the scarcity of endothelial pinocytosis in cerebral endothelial cells and the presence of interendothelial tight junctions (Reese & Karnovsky 1967). However, some pinocytosis is observed even in normal conditions in cerebral endothelium, especially in arterioles, but this pinocytosis allows only passage of small amounts of macromolecules from blood into brain (Westergaard & Brightman 1973, NGrtved & Westergaard 1974, Westergaard 1975). Under certain pathological conditions, there is a passage of macromolecules from blood into brain (Broman 1949, Brightman et al. 1970). Great effort has been made to show the way of passage of macromolecular tracers in various conditions (for review see Manz 1974). However, the exact route of macromolecules in many conditions still remains unclear. Possible ways of leakage of macromolecules across the blood-brain barrier are rupture of the endothelial cell membrane, widening of the interendothelial tight junctions, diffusion across the endothelial cell membranes and cytoplasm
413 or passage by pinocytosis or trans-endothelial channels (Reese & Karnovsky 1967. Brightman & Reese 1969, Hirano et al. 1969, Brigthman et al. 1970, Clasen et al. 1970). Pinocytotic transport of proteins across the blood-brain barrier is known to occur in a variety of conditions (e.g. Hansson et al. 1975, Beggs & Waggener 1976, Persson & Hansson 1976, Nag et al. 1977, Petito et al. 1977). The leakage of macromolecular tracers around cerebral wounds might be due to an increased pinocytotic transport and a formation of trans-endothelial channels. Pinocytotic vesicles, caveolae as well as trans-endothelial channels filled with peroxidase, have been shown around cerebral stab wounds (Persson & Hansson 1976) as well as in experimental spinal cord compression (Beggs & Waggener 1976). In the present work, the mechanism of extravasation of peroxidase was by passage through trans-endothelial channels and through chains of pinocytotic vesicles, forming channels. These channels and chains of vesicles, as well as scattered pinocytotic vesicles and caveolae, were observed within the area showing extravasation of peroxidase into vascular basement membranes and extracellular space around cells and cell processes. However, the number of trans-endothelial channels, pinocytotic vesicles and caveolae was enhanced in ethanol-intoxicated rats. The latter finding was consistent with the finding in a preceding study (Persson & Rosengren 1977), where it was shown that the amount of extravasated Evans blue-labelled albumin was increased around cerebral stab wounds in ethanol-intoxicated rats. Furthermore, it was shown that the area containing leaking blood vessels was greatly increased in ethanol-intoxicated compared to control rats. Such an increase of the leaking area has previously been observed to Evans blue-labelled albumin around cerebral contusions (DeCrescito et al. 1974, Flamm et al. 1977). The mechanisms behind this increased leakage of peroxidase around cerebral stab wound is unknown. Ethanol and its metabolite acetaldehyde affect biogenic amine metabolism (Duritze & Truitt 1966, Kalant 1975, von Wartenburg et al. 1975), and biogenic amines, especially serotonin, are known to affect the blood-brain barrier permeability to macromolecular tracers (Cotran & Karnowsky 1967, Westergaard 1975). Ethanol is also known to dilatate cerebral blood vessels (Weiss et al. 1976). This might give an increased pressure load on cerebral capillaries and venules, making them more sensitive to injury. Another mechanism is the effect of ethanol and acetaldehyde on biological membranes. Both ethanol and acetaldehyde inhibit Na+-K-ATPase in cell memebranes, which results in an accumulation of ATP, a decrease in the influx of K into cells and a decreased O2 consumption (Jiirnefelt 1961, Wallgren 1966, Israel & Salazar 1967, Israel 1970, Nielsen et al. 1975). This
414 effect is reflected by a reversible depressant effect of ethanol on the electrical activity of, e.g. squid giant axon (Armstrong 1964, Moore et al. 1964), rat cerebral cortex (Nikander et al. 1971) and mouse cerebral and cerebellar neuronal cultures (Seil et al. 1977). This depressant effect of ethanol on the activity of biological membranes can immediately be reversed by rinsing the biological membranes with ethanol-free solutions (Seil et al. 1977) and is thus likely to be a direct effect of ethanol on membranes rather than mediated by biogenic amines. One effect of ethanol and its principal metabolite acetaldehyde on biological membranes has been considered the ability of liberating free radicals in membranes (DiLuzio & Hartman 1967, 1969). These free radicals might interfere with membrane phospholipids, and thereby with Na+-K adenosine triphosphatase in cell membranes (Kimelberg & Papahadjopoulos 1974). This interaction of ethanol with biological membrane functions might explain the observation in this work that a greater number of injured neurons in the vicinity of the stab wound are unable to maintain their integrity to foreign proteins in ethanol-intoxicated rats, which results in a free passage of peroxidase through injured membranes into cytoplasm, where the peroxidase molecules are distributed in a diffuse pattern among the neuronal organelles. Functionally, this might mean that a smaller number of neurons have the possibility to recover after a moderate to severe injury in ethanol-intoxicated compared to control rats. That conclusion is further supported by the finding by Brodner et al. (1976) and by Flamm et al. (1977) that functional recovery after spinal contusions was rare in cats which were ethanol-intoxicated at injury, compared to non-intoxicated ones. Furthermore, the microscopical integrity of the spinal cord was fairly preserved at 6-8 weeks after injury in controls, while the ethanol-intoxicated group showed a severe atrophic scar at the impact area in the spinal cord (Flamm et al. 1977). There was a potentiation of the blood-brain barrier disturbances around cerebral stab wounds in ethanol-intoxicated rats, compared to controls. Endothelial cells showed an increase in trans-endothelial channels, pinocytotic vesicles and caveolae. Injured neurons in the vicinity of the stab wound showed a loss of their integrity to foreign proteins, with a trans-membraneous leakage of peroxidase into cytoplasm in a larger number of cells in ethanolintoxicated rats, compared to controls. Both the endothelial and the neuronal membrane effects of ethanol might be due to interference with biogenic amine metabolism, an increased formation of free radicals or due to other effects on biological membranes.
415 ACKNOWLEDGEMENTS Our sincere thank to Mrs. Barbro Eriksson and Mrs. Ulla Svedin for skilled technical assistance. Dr. Barbro Johansson in thanked for helpful discussions. This work was supported by grants from the Trygg-Hansa (Fylgia) Foundation, Tore Nilson's Foundation for Medical Research, the Swedish Medical Research Council and the Medical Faculty of the University of Goteborg.
REFERENCES Armstrong, C . L. L. Binstock (1964): The effects of several alcohols on the properties of the squid giant axon. J. Gen. Physiol. 48, 265-277. Beggs, J. L. & J. D. Waggener (1976): Transendo~thelialvesicular transport of pro'tein following compression injury to the spinal cord. Lab. Invest. 34, 428-439. Brightman, M. W. (&T. S. Reese (1969): Junctions between intimately opposed cell membranes in the vertebrate brain. J. Cell. Biol. 40, 648-677. Brightman, M. W., I. Klatzo, Y. Olsson Lk T. S. Reese (1970): The blood-brain barrier to protein tracers under normal and pathological conditions. J. Neurol. Sci. I 0 , 2 15-239. Brodner, R. A,, J. C. Van Gilder, W. F. Collin0 Jr, ,& B. Hamill (1976): Experimental spinal cord trauma: The efficacy of treatment. 44th Annual Meeting of the American Association of Neurological Surgeons. San Francisco. Broman, T. (1949): The permeability of the cerebrospinal vessels in normal and pathological conditions. Munksgaard, Copenhagen. Clasen, R. A., S. Pandolfi & G. M. Hass (1970): Vital staining, serum albumin and the blood-brain barrier. J. Neuropath. Exp. Neurol. 29, 266-284. Cotran, R. S. & M. J. Karnovsky (1976): Vascular leakage induced by horseradish peroxidase in the rat. Proc. SOC.Exp. Biol. (N.Y.) 126, 557-561. I. J. Ransohoff (1974): Ethanol poDeCrescito, V., H. G . Demopoulos, E. S. Flamm tentiation of traumatic cerebral edema. Surg. Forum 25, 438-440. DiLuzio, N. R. A. D. Hartman (1967): Role of lipid peroxidation in the pathogenesis of the ethanol-induced fatty liver. Fed. Proc. 26, 1436-1442. DiLuzio, N. R. .& A. D. Hartman (1969): The effect of ethanol and carbon tetrachloride administration on hepatic lipid-soluble antioxidant activity. Exp. Mol. Pathol. II, 38-52. Duritz, G. ,& E. B. Truitt, Jr (1966): Importance of acetaldehyde in the action of ethanol on brain norepinephrine and 5-hydroxytryptamine. Biochem. Pharmacol. 15, 711-721. Ehrlich, P. (1885): Das Sauerstoff-Bediirfnis des Organismus. Eine farbenanalytische Studie. Hirschwald, Berlin. Goldmann, E. E. (1913): Vitalfarbung am Zentralnervensystem. Beitrag zur PhysioPathologie des Plexus choroideus und der Hirnhaute. Abh. K8nigl. Preuss. Akad. Wissensch. No. 1. Hanssm, H.-A., B. Johansson & C. Blomstrand (1975): Ultrastructural studies o'n cerebrovascular permeability in acute hypertension. Acta Neuropath. (Berl). 32, 187-198. A. M. Zimmerman (1969): Pathological alterations in the Hirano, A,, N. H. Becker cerebral endothelial cell barrier to peroxidase. Arch. Neurol. (Chic.) 20, 300-308. Flamm, E. S., H. B. Demopoulos, M. L. Seligman, J. J. Tomasula, V. DeCrescito .& J. Ransohoff (1977): Ethanol potentiation of central nervous system trauma. J. Neurosurg. 46, 328-335. (&
(&
I&
416 Israel, Y . (1970): Cellular effects of alcohol. Quart. J . Stud. Alcohol. 31, 293-315. Israel, Y . (& I. Salazar (1967): Inhibition of brain microsomal adenosine triphosphatases by general depressants. Arch. Biochem. Biophys. 122, 310-317. JLrnefelt, J. (1961): Inhibition of the brain microsomal adenosine-triphosphatase by depolarizing agents. Biochem. Biophys. Acta 48, 111-116. Kalant, H. (1975): Direct effects of ethanol on the central nervous system. Fed Proc. 34, 1930-1941. Kimelberg, H. #&D. Papahadjopoulos (1974): Effects of phospholipid acyl chain fluity, phase transitious and cholesterol on (Na' K+)- stimulated adenosine triphosphatase. J. Biol. Chem. 249, 1071-1080. Krieg, W. J. S. (1946): Connection of the cerebral cortex. I. The albino rat. A. Topography of the cortical areas. J. Comp. Neurol. 84, 221-275. Majno, G. #&G. E. Palade (1961): Studies on inflammation. Part. 1. (The effects of histamine and serotosnin on vascular permeability: an electron microscopic study). J. Biophys. Biochem. Cytol. I I , 571-606. Manz, H. J. (1974): The pathology of cerebral edema. Human Path. 5, 291-313. Moore, J. W., W. Ulbricht ,& M. Takata (1964): Effect of ethanol on sodium and potassium conductances of the squid axon membrane. J. Gen. Physiol. 48, 279-295. Nag, S., D . M . Robertson & H.B. Dinsdale (1977): Cerebral cortical changes in acute experimental hypertension. An ultrastructural study. Lab. Invest. 36, 150-161. Nielsen, R. H., R. A. Hawkins J&R. L. Veech (1975): The effects of acute ethanol intoxication on cerebral energy metabolism. Adv. Exp. Med. Biol. 59, 93-109. Nikander, P., P. von Boguslawsky & H. Wallgren (1971): Actions of ethanol and tertiary butanol on net movements of sodium and potassium in electrically stimulated rat brain tissue in vitro. Scand. J. Clin. Lab. Invest. Suppl. 27, 116, 67. NGrtved, J. ,& E. Westergaard (1974) Transport of horseradish peroxidase across cerebral arteroles and venules. Anat. Rec. 178, 428. Persson, L. & H.-A. Hansso'n (1976): Reversible blood-brain barrier dysfunction to' peroxidase after a small stab wound in the rat cerebral cortex. Acta Neuropath. (Berl.) 35, 333-342. Persson, L. & L. Rosengren (1977): Increased blood-brain barrier permeability around cerebral stab wounds, aggravated by acute ethanol intoxication. Acta Neurol. Scand. 56, 7-16. Petito, C. K., J. A. Schaefer & F. Plum (1977): Ultrastructural characteristics of the brain and blood-brain barrier in experimental seizures. Brain Research. 127, 251-267. Raimondi, A. J., J. P. Evans ,& S. Mullan (1962): Studies on cerebral edema. Part 3 (Alterations in the white matter. An electron microscopy study using ferritin as a labeling compound). Acta Neuropath. (Berl.) 2, 177-197. Reese, T. S. & M. J. Karnovsky (1967): Fine structural localization of a blo'od-brain barrier to exogenous peroxidase. J. Cell. Biol. 34, 207-217. Rosengren, L., L. Persson ,& B. Johansson (1977): Enhanced blood-brain barrier leakage to Evans blue-labelled abumin after air embolism in ethanol-intoxicated rats. Acta Neuropath. (Berl.) 38, 149-152. Seil, F. J., A. L. Leiman, M. M. Herman & R. A. Fisk (1977): Direct effects of ethanol on central nervous system cultures: An electrophysiological and morphological study. Exp. Neurol. 55, 390-404. Tschirgi, R. D. (1950): Protein complexes and the impermability of the blood-brain barrier to dyes. Amer J. Physiol. 163, 756. Wallgren, H. (1966): Effects of alcohol on biochemical processes in the central nervous system. Psychosom. Med. 28, 431-442.
+
417 von Wartenburg, J. P., D. Berger, M. M. Ris & B. Tabakoff (1975): Enzymes of biogenic aldehyde metabolism. Adv. Exp. Med. Biol. 59, 119-138. Weiss, M.H., J.S. Heiden, M.L. J. Apuzzo & T. Kurze (1976): The effects of acute ethanol intoxication on intracranial physical dynamics. 44th Annual Meeting of the American Association of Neurological Surgeons. San Francisco. Westergaard, E. (1975): Enhanced vesicular transport of exogenous peroxidase across cerebral vessels induced by serotonin. Acta Neuropath. (Berl.) 32, 27-42. Westergaard, E. M. W. Brightman (1973): Transport of proteins across normal cerebral arterioles. J. Comp. Neur. 152, 17-44. (&
Received March 20, accepted April 5, 1978
Dr. Lennart I . Persson Department of Neurology SahIgren Hospital S-413 45 Goteborg Sweden
