- Email: [email protected]
Effect of body temperature on brain edema and encephalopathy in the rat after hepatic devascularization
GASTROENTEROLOGY
1989;96:885-91
Effect of Body Temperature on Brain Edema and Encephalopathy in the Rat After Hepatic Devascularization PETER ANDRES
TRABER, T.
MAURO
DALCANTO,
DANIEL
and
BLEI
Gastroenterology and Neuropathology Sections, Departments Lakeside Veterans Administration Hospital and Northwestern
Brain edema is a fatal complication of fulminant hepatic failure and its pathogenesis remains unclear. To determine its presence in a model of ischemic hepatic failure, rats were subjected to a portacaval anastomosis followed by hepatic artery ligation. Brain water was measured using the sensitive gravimetric method. Preliminary studies revealed marked hypothermia in devascularized animals kept at room temperature (26.9” f 28°C). An additional group of devascularized rats was kept in an incubator. As expected for hypothermia, such animals had a lower arterial pressure and heart rate; the duration of encephalopathy was markedly prolonged. Water content of the cortical gray matter was only increased in normothermic devascularized rats: 80.14% f 0.31%, normal: 80.06% + portacaval shunt only; 80.42% + 0.26%, 0.22% devascularized at room temperature: 81.29% f 0.38%, devascularized at controlled temperature (p < 0.001). Such differences could not be detected using the dry-weight technique in whole cerebral hemispheres. Astrocyte changes in the cortical gray matter were noted in both edematous and nonedematous devascularized groups, coupled with the presence of vesicles containing horseradish peroxidase in the endothelial capillary cell. This suggests that in this model, brain edema may be due to both a cytotoxic mechanism and changes in the permeability of the blood-brain barrier. Future studies with this widely used model will require strict control of temperature to allow interpretation of experimental results. A therapeutic role for hypothermia in the management of brain edema deserves further attention.
rain edema is frequently a fatal complication of fulminant hepatic failure (FHF) (1). Its pathogenesis, however, remains obscure; this reflects, in part,
B
GANGER,
of Medicine University,
and Pathology, Chicago, Illinois
the difficulty in performing clinical studies in critically ill patients. Furthermore, few studies have attempted to measure water content of the brain or the mechanisms responsible for water accumulation in experimental models of FHF. In the hepatectomized rat, Livingstone et al. (2) reported a 5.4% increase in water content of gray matter at the time of death coupled with gross changes in the permeability of the blood-brain barrier (BBB). Three groups have noted an increase in the water content of the brain in the galactosamine-treated rat (S-S), although in two studies this was noted only in the hindbrain (3,~). However, in all these studies water was measured using the dry-weight technique, an insensitive method to detect changes in brain water (6). Pathogenically, cytotoxic factors have been incriminated (4); evidence both for (7,8) and against (9,101 changes in the permeability of the BBB has been presented. We have recently shown in the rabbit that galactosamine-induced FHF was associated with a selective increase in the water content of cortical gray matter (11); such changes did not appear to be due to hypotension, hypoxia, acidosis, or hypoglycemia. Electron-microscopic evaluation revealed swollen astrocytes in cortical gray matter (12). No evidence of horseradish peroxidase after intravenous injection was present in capillary endothelial cells or brain parenchyma, indicating a grossly intact BBB. To measure brain water we used the gravimetry technique (IS), a sensitive method that can determine water content in small tissue specimens. We suggested the presence of a cytotoxic mechanism that may induce cellular swelling in this model of FHF.
Abbreviations used in this paper: BBB, blood-brain barrier; FHF, fulminant hepatic failure. 0 1989 by the American Gastroenterological Association 0016-5085/89/$3.50
886
GASTROENTEROLOGY
TRABER ET AL.
In humans, such a mechanism may also be present, as sera of patients with FHF in advanced coma was found to inhibit Na+, K+-adenosine triphosphatase activity of rat brain membranes (14). To gain further insight into the pathogenesis of brain edema, we studied an anhepatic model of FHF, namely, the rat after hepatic devascularization. This model has been used extensively to study the pathogenesis of hepatic encephalopathy (15). In a recent communication, Zieve (16)noted that the brains of such animals do not exhibit an increase in water content as measured with the dry-weight technique; this finding contrasts with the water accumulation seen in the hepatectomized rat (2). As in these latter studies the role of covariables that could influence water accumulation in the brain was not rigorously examined, we continuously monitored several factors that could affect any potential changes in brain water. Among them we noted the previous reports of (17). hypothermia after hepatic devascularization Our results suggest that this change in body temperature greatly affects the characteristics of this model, particularly those pertaining to cerebral edema.
Materials and Methods Adult male Sprague-Dawley rats, weighing 300400 g, were anesthetized with ether. An end-to-side portacaval anastomosis was constructed according to the method of Lee and Fisher (18), with a loose ligature left around the hepatic artery. Once the animals had recovered, they were housed in individual cages for 24-48 h. Under ether anesthesia, a laparotomy was then performed to tie off the hepatic artery. At the same time, a PE-50 catheter (Intramedic) was introduced into a femoral artery for measurement of arterial pressure and blood gases and the other end was tunneled under the skin, exteriorized over the dorsal aspect of the back, and sealed with a metal plug. A thermistor wire for temperature measurement was left between bowel loops and exteriorized at the same site; the abdomen was closed with double suture layering. Preliminary studies indicated that animals that had undergone hepatic devascularization developed a marked reduction in body temperature. Therefore, animals were divided into three groups: control (portacaval anastomosis but at the time of the second laparotomy no hepatic artery ligation), ischemic FHF kept at room temperature (IFHFA), and ischemic FHF kept at controlled temperature (IFHF-B). In the latter group, animals were kept in an incubator and temperature was maintained at 375°C with the use of an infrared heating lamp. Animals were continuously monitored after ligation of the hepatic artery. Blood glucose was measured every 2 h from tail vein samples with the use of glucose strips. Supplemental glucose [as DlOW), in volumes of up to 5 ml, was administered subcutaneously as needed to maintain a blood glucose level above 140 mg% (19). Blood pressure was measured by connecting the arterial line to a
Vol. 96, No. 3
P23 Gould-Statham transducer and the tracing was continuously inscribed in a Hewlett-Packard 7754A recorder. Heart rate was derived from the arterial pressure tracing. Temperature was monitored continuously by connecting the thermistor to a digital display unit (Bailey Instruments, N.J.). Animals were monitored until stage 4 encephalopathy developed, as defined by Zieve et al. (19). At this stage, animals are responsive to pain but are unable to assume a standing position when lying on their side. Arterial blood gases were determined and corrected for hypothermia when necessary. Animals were killed by decapitation and truncal blood was collected for chemical measurements. This study was approved by the Animal Care and Use Committee at Northwestern University.
Measurement
of Brain
Water
The content of brain water was measured by determining tissue specific gravity after flotation of tissue samples in a bromobenzene-kerosene gradient, as described (11). Four K,SO, solutions of different concentrations were used to calibrate the specific gravity gradients, with a linear coefficient of correlation >0.99. With this technique, brain tissue is gently placed in such a column and the equilibrium position is recorded after 2 min (13). The specific gravity can then be calculated from the previously determined linear equation and expressed as water content, using the following equation: Tissue water (%) = 1 -
(%t - 1) 1(1 - %,b!&
1x
100,
where sg, is the specific gravity of the tissue studied and sg, is the specific gravity of the solid component. The latter was calculated in five nonmanipulated control rats using the ratio of wet (W) to dry (D) weight obtained after surgical dissection of gray matter and freeze-drying for 48 h at -70°C using the following equation:
The values obtained are in agreement with our previous study (11) and other reports (6,13). As soon as the brain was carefully removed, it was placed in an airtight container. Using a Z-mm-diameter needle, four punch samples were taken from the cortical gray matter of the left hemisphere; subcortical white matter could not be accurately sampled due to the thinness of the tissue. Each sample was promptly placed in the calibrated column and tissue water percentage was calculated as previously described. The right cerebral hemisphere, brainstem, and cerebellum were weighed, placed in an oven, and reweighed until a stable dry weight was attained at 24 h. Tissue water could then be calculated as follows: Tissue water (%)
=
Wmitia~- ha,, Whitid
Electron Microscopy In 3 rats from each group, the integrity of the BBB was tested using horseradish peroxidase. Thirty minutes
March I!389
Table
BRAIN EDEMA AFTER HEPATIC DEVASCULARIZATION
1. Characteristics
-Temperature (“C) Time after HAL (h) MAP (mmHg) Heart rate Supplemental glucose (ml)
of the Model Control (n = 5)
37.5 t 0.4 11.9 C 4.2b 89 ? 12 430 k 48 21 5 8.2
IFHF-A (n = 5) 26.9 2 17.0 2 60 2 213 k 26 k
2.8" 3.1" 25" 102" 5.5
IFHF-B (n = 5) 38.4 2 1.6 9.4 t 1.0 140 2 18 412? 55 20 ?I 3.5
HAL, hepatic artery ligation; MAP, mean arterial pressure. Values are mean * SD. Control: portacaval anastomosis only. IFHF-A: portacaval anastomosis and hepatic artery ligation kept at room temperature. IFHF-B: devascularized rats kept at 37.5% a p < 0.01 vs. other two groups. b Sham hepatic artery ligation. ’ p < 0.01 vs. IFHF-B.
before killing the rats, 200 mg of horseradish peroxidase type II (Sigma Chemical Co., St. Louis, MO.) dissolved in phosphate-buffered saline was injected into a femoral vein catheter that had been previously placed at the time of the second operation. The rats were anesthetized with ketamine, a tracheostomy was performed, and a polyethylene catheter was placed in a femoral artery. Arterial blood gases were checked for adequate oxygenation. The chest was opened and chilled phosphate-buffered saline followed by 4% formaldehyde and 3% glutaraldehyde in phosphate-buffered saline was infused via a cannula placed in the left ventricle. A pulsatile pump assured a delivery of 200 ml/min; blood pressure was within normal values during the perfusion. After removing the brain, multiple s-mm-thick sections were removed from cortex, cerebellum, and basal ganglia and cut in a Vibratome into sections 15-30 pm thick. These sections were treated for 30 min in a solution of 0.05% diaminobenzidine in Tris buffer with added 0.01% H,O,. The sections were then washed, postfixed with 1% osmic acid, dehydrated through a graded series of ethyl alcohol, cleared in propylene oxide, and embedded in
Epon. The sections were mounted on glass slides before polymerization to be observed under light microscopy. Selected sections were then polymerized for both l-pm and ultrathin sectioning. Grids were studied with and without counterstaining in a Joel-100 CX II electron microscope.
Statistical
Analysis
Comparison among the three experimental groups was made with one-way analysis of variance; differences between means were analyzed with Tukey’s HSD test (20). All values are expressed as mean ? SD.
Results
887
with a mean temperature of 26.9”C and a range of 22.5”-30°C. No evidence of shivering was noted in spite of the marked reduction in temperature. Animals of group IFHF-B remained at 37”-38°C with the use of a heating lamp. In 2 rats, temperature rose suddenly to 38.2”C and 41.2%. Devascularized rats exhibited marked differences in the time elapsed to stage 4 encephalopathy. The range was 12.5-20 h in IFHF-A rats and 8.25-11 h in the IFHF-B group (p < 0.01). In 2 animals of the latter group, encephalopathy was associated with stimulus-induced seizurelike activity, characterized by rapid movements of head and limbs, without a tonic phase. The control rats were studied at times within the range of the devascularized animals (9-19 h1. Marked differences in heart rate and blood pressure were also seen in the two devascularized groups. As anticipated with this degree of hypothermia, IFHF-A rats exhibited lower arterial pressure and heart rate (Table 1). The animal with the lowest body temperature (23%) had a blood pressure of 37/ 16 mmHg and a pulse of 36 beats per minute. Rats of the IFHF-B group showed mild elevations in blood pressure when compared with controls; however, portacaval-shunted rats exhibited lower values of arterial pressure than normal awake animals in our laboratory (100-110 mmHg), a decrease that probably reflects the hyperdynamic circulation seen after this procedure (21). Laboratory
Tests
The amount of supplemental glucose administered to all three groups did not differ (Table 1). However, blood glucose was raised in the IFHF-A group (Table 2); this probably reflects impaired utiof glucose in the presence of a low body lization
temperature. Hematocrits were similar in all animals. Both devascularized groups exhibited hypocapnia, suggesting a component of respiratory alkalosis, mild hyponatremia, and a rise in creatinine. Marked elevations of transaminases were noted in these animals; however, these findings should be interpreted with caution as collection of blood by decapitation may be contaminated by leakage of enzymes of muscle origin. This was noted when truncal blood collections in a group of 5 normal decapitated rats still showed elevations of serum glutamic oxaloacetic transaminase when compared with values obtained from the inferior vena cava in normal rats.
General Aspects The body temperature of the control group was normal (Table 1). Rats that underwent hepatic devascularization (IFHF-A) developed hypothermia,
Water Content of the Brain cient
Using the gravimetry technique, the coeffiof variation for the four measurements of
888
GASTROENTEROLOGYVol. 96, No. 3
TRABER ET AL.
Table
2. Laboratory
Table
Results IFHF-A
Control Glucose (mg%) Hematocrit (%) PH PO, (torr) Pco, (torr) Creatinine (mg%l Sodium (mEq/L) Bilirubin (mgO/o) Alkaline phosphatase
164 f 39 c 7.47f 85 f 33 k 0.582
31 3 0.03 21 6 0.05
390 f 412 7.53+ 102 + 25 2 1.12
144 2 3
IFHF-B 153 + 119 42 2 5 7.49-+0.13 103 f 5 23 k 4'= 1.62 0.36"
234" 2 0.09 9 3= 0.55
127 +ll
128 2 12
2.1+ 0.5"
2.52 0.8"
0.132 0.17
8512 466b
169 2 95
649 + 32gb
(NJ/ml)
SGOT (U/ml]
793 ‘- 673
cortical gray matter within each animal ranged from 2.2% to 5.8%. Devascularized rats kept in a warm environment had a significant increase in the water content of the cortical gray matter [Figure 1). Control animals had a water content of 80.06% + O.22%, the IFHF-A group 80.42% + 0.26%, and the IFHF-B group 81.29% ? 0.38% (F = 24.28, p < 0.001). Results in the control group were comparable to 5 normal noninstrumented rats in which water content of the cortical gray matter was 80.14% ? 0.31%. The water content of the other cerebral hemisphere, brainstem, and cerebellum was also calculated using the dry-weight method (Table 3). Only the cerebellum showed a significant increase in the
82
.
81
: A
80
: .
:
79
0
C
Cerebral hemisphere Brainstem Cerebellum
Weight Method
Control (5)
IFHF-A (5)
IFHF-B (5)
80.372 0.51
79.62k 0.98
80.59+ 0.69
77.62+ 0.79 78.87C 1.23
77.55k 0.48 77.752 1.61
78.352 0.73 80.17t 0.54"
Control: portacaval anastomosis only. IFHF-A: portacaval anastomosis and hepatic artery ligation kept at room temperature. IFHF-B: devascularized rats kept at 37.5"C. Values are mean f SD. All results are percentage of water. a p < 0.05vs.IFHF-A.
water content with ischemic
of the IFHF-B group when compared rats kept at room temperature.
18353 f 5511b 28276k 7577b
SGOT, serum glutamic oxaloacetic transaminase. Control: portacaval anastomosis only. IFHF-A: portacaval anastmosis and hepatic artery ligation kept at room temperature. IFHF-B: devascularized rats kept at 37.5”C. Values are mean * SD. ’ p < 0.05 vs. control. b p < 0.01 vs. other two groups.
0n z
3. Water Content/Dry
IFHF-A
IFHF-B
Figure 1. Water content of cortical gray matter in individual experiments using the gravimetric technique. A clear increase in brain water was seen in devascularized rats kept at controlled temperature (IFHF-B) as compared with animals with portacaval anastomosis (C) or devascularized kept at room temperature (IFHF-A).
Electron
Microscopy
In all three groups, horseradish peroxidase was present in brain capillary endothelial cells, mainly as endothelial vesicles, without accumulation in the interstitium or cellular elements [Figure 2). In all three groups, astrocytes of the cortical gray matter appeared with variable degrees of cytoplasmic vacuolation and swelling of pericapillary foot processes. Qualitative assessment of white matter astrocytes did not reveal such changes.
Discussion In the present study, we observed that hypothermia after hepatic devascularization affected several physiologic parameters in this model while prolonging the period of encephalopathy. Furthermore, maintenance of a normal body temperature in devascularized rats was associated with an increase in brain water. Studies of the encephalopathy of FHF in the experimental animal have proceeded via one of two experimental models: toxic injury or the anhepatic state. Thus, the significance of experimental findings could be enhanced by determining their presence in different preparations rather than selecting a single model for study. Each model can be criticized for its shortcomings. The one used in this study, for example, does not have a potential for recovery, limiting the assessment of therapeutic maneuvers to the possible prolongation of encephalopathy rather than to survival. Marked differences are present in metabolic disturbances, even within different species. Such is the case for brain edema in FHF, a complication whose pathogenesis is still not completely understood. Our previous study in the rabbit with galactosamine-induced FHF suggested that a cytotoxic mechanism could account for the increase in cortical gray matter water, as detected with the gravimetry method (11). Neuropathologically, we
March 1989
Figure
BRAIN EDEMA AFTER HEPATIC DEVASCULARIZATION
889
2. Electron micrographs from cortical vessels derived from (a) rats with portacaval anastomosis, lb) rats with portacaval anastomosis and hepatic artery ligation, and (c) devascularized rats kept at controlled temperature. A modest number of vesicles containing horseradish peroxidase (arrows) are present in endothelial cells in all three instances. Increased vacuolation of underlying astrocytic foot processes can also be observed. Magnification, ~14,000.
noted swelling of gray matter astrocytes and the absence of a gross breakdown of the BBB (12). Determining these parameters in another model of FHF would be important to support the presence of a cytotoxic mechanism. The rat after hepatic devascularization developed marked hypothermia. A decrease in body temperature has been previously noted in anhepatic models of FHF (2,17). One possible explanation lies in the contribution of hepatic metabolism to the generation of body heat, estimated to be -12% under resting conditions in the adult rat (22). To explain the precipitous drop in body temperature, it is possible that compensatory mechanisms after devascularization, via shivering or nonshivering thermogenesis, were not activated; alternatively, adequate central regulation of body temperature may have been absent. Regardless of the mechanism, severe hypothermia can be associated with encephalopathy and changes in the BBB (231, and thus become a confounding factor in the interpretation of experimental results. In our hypothermic rats, arterial pressure and heart rate were reduced as an appropriate physiologic response to a decrease in body temperature. The moderate elevation of plasma glucose values could reflect its diminished utilization, a common finding in hypothermia (24). Devascularized animals maintained at normal body temperature had a shorter duration of encephalopathy; in 2 animals,
seizure activity was observed. Although it has been stated that the decrease in body temperature simply slows the progression of hepatic encephalopathy (16), our studies suggest multiple biological differences. Brain water was increased in the cortical gray matter of “warm,” devascularized rats. No overlap was seen with the portacaval anastomosis group and with 6 normal noninstrumented rats. The unchanged values of water content of the whole cerebral hemisphere with the dry-weight method may again reflect its relative insensitivity for studies of brain edema in FHF. With the gravimetry technique, a mean increment of 1.2% in brain water was seen at an intermediate stage during the development of encephalopathy, corresponding to an -8% rise in brain volume (251.
To explain this increase in brain water, several factors need to be considered. Differences in cerebral hemodynamics may be present; a “protective” role for hypothermia in “cold” animals could be mediated by a decrease in cerebral blood flow (26). Although seizure activity could have potentially contributed to the rise in brain water, water content in the 2 affected animals was similar to that of the entire group. Our morphologic studies suggest the presence of both cellular changes and an increased permeability to horseradish peroxidase. As in the galactosamine rabbit, the increase in water probably resides in
890
TRABER
GASTROENTEROLOGY
ET AL.
astrocytes of the cortical gray matter. It should be noted that animals with portacaval anastomosis alone also exhibited astrocytic changes. Marked cytoplasmic vacuolation of astrocytes in the first week after portacaval shunting in the rat has been previously reported (27,281. On the other hand, all three groups showed vesicles containing horseradish peroxidase within the capillary endothelial cell; this suggests lysosomal activation and increased pinocytotic transfer across the BBB (29). Such endothelial vesicles have also been reported after portacaval anastomosis alone (30). As horseradish peroxidase was not seen in the basement membrane or within the brain tissue, it is unlikely that a gross breakdown of the BBB had occurred. The overall neuropathological similarity between groups suggests that a similar pathogenic mechanism may be present. Unfortunately, quantitative estimations of astrocyte swelling are hampered by the small cellular soma and the multiple projections represented by the foot processes. Still, the net accumulation of water in the “warm” group could result from a higher concentration of a putative cytotoxin, such as ammonia. Tenfold increases of ammonia levels reported in the whole brain of “warm” devascularized rats (31) contrast with the doubling of values seen after portacaval anastomosis alone (32) and the threefold rise seen in “cold” animals (19). Of note, the neurotoxicity of ammonia is reduced in hypothermia (33). A further link between astrocyte swelling and ammonia in FHF is the fact that glutamine synthetase is located solely in astrocytes, which remove ammonia by conversion to glutamine; the ability to increase ammonia detoxification via induction of glutamine synthetase is limited (34). Coupled with the observations of ammonia-induced astrocyte swelling in the normal primate (35), similar astrocyte changes seen in rats with portacaval anastomosis given an ammonia load (36), cytoplasmic swelling of astrocytes in vitro after the addition of ammonia (37), and ammonia-induced changes in the BBB (38), we hypothesize that ammonia may be pathogenically related to the increase in brain water seen in FHF. In summary, the devascularized rat model of FHF exhibits major differences in the duration of encephalopathy, neurologic signs, and accumulation of brain water according to the temperature at which the animal is maintained. Brain edema is only seen at normal body temperature; its pathogenesis may be related to both cytotoxic changes and to changes in BBB permeability. These findings suggest that future studies with this model will require strict control of temperature to allow interpretation of experimental results. In view of the clinical use of hypothermia in the treatment of vasogenic edema (39), a role for such
Vol. 96, No. 3
therapy in the management of encephalopathy and brain edema in human FHF may deserve further attention.
References 1. Ware AJ, D’Agostino AN, Combes BC. Cerebral edema: a major complication of massive hepatic necrosis. Gastroenterology 1971;61:887-4. 2. Livingstone AS, Potvin M, Goresky CA, et al. Changes in the blood-brain barrier in hepatic coma after hepatectomy in the 1977;73:697-704. rat. Gastroenterology 3. Groflin R, Tholen H. Brain edema in the rat with galactosamine-induced severe hepatitis. Experientia 1978;34:1501-3. 4. Ede RJ, Gove CD, Hughes RD, Marshall W, Williams R. Reduced brain Na+-K+ ATPase activity in rats with galactosamine-induced hepatic failure: relationship to encephalopathy and cerebral oedema. Clin Sci 1987;72:365-71. 5. Dixit V, Chang TMS. Effects of prostaglandin E2 on brain edema and liver histopathology in a galactosamine-induced fulminant hepatic failure rat model. Biomater Artif Cells, Artif Organs 1987;15:559-73. 6 Shigeno T, Brock M, Shigeno S, et al. The determination of brain water content: microgravimetry versus drying-weighing method. J Neurosurg 1983;59:803-9. 7 Horowitz ME, Schafer DF, Molnar P, et al. Increased bloodbrain transfer in a rabbit model of fulminant hepatic failure. Gastroenterology 1983;84:1003-11. 8. Zaki AEO, Ede RJ, Davis M, et al. Experimental studies of blood-brain Hepatology
barrier
permeability
in
acute
hepatic
failure.
1984;4:359-63.
9. Lo WD, Ennis SR, Goldstein GW, Betz AL. The effects of galactosamine-induced hepatic failure upon blood-brain bar1987;7:452-6. rier permeability. Hepatology 10. Knudsen GN, Poulsen HE, Paulsen DB. Blood-brain barrier permeability in galactosamine-induced hepatic encephalopathy. No evidence for increased GABA transport. J Hepatol 1988;6:187-92. 11. Traber PG, Ganger DR, Blei AT. Brain edema in rabbits with galactosamine-induced fulminant hepatic failure. Regional differences and effects on intracranial pressure. Gastroenterology 1986;91:1347-56, 12. Traber PG, DalCanto M, Ganger D, Blei AT. Electron microscopic evaluation of brain edema in rabbits with galactosamine-induced fulminant hepatic failure. Hepatology 1987;7: 1257-61. 13. Marmarou A, Poll W, Shulman K, et al. A simple gravimetric technique for measurement of cerebral edema. J Neurosurg 1978;49:530-7. 14. Seda HWM, Hughes RD, Gove CD, et al. Inhibition of rat brain Na+-K+ ATPase activity by serum from patients with fulminant hepatic failure. Hepatology 1984;4:74-9. 15. Zieve L. Pathogenesis of hepatic encephalopathy. Metabol Brain Dis 1987;2:147-66. 16. Zieve L. Reply to Muller KD. Hepatology 1987;7:618. 17. Peignoux M, Bernuau J, Benhamou JP. Total hepatectomy and hepatic vascular exclusion in the rat. A comparison, with special reference to the influence of body temperature. Clin Sci 1982;62:273-7. 18. Lee SH, Fisher B. Portacaval shunt in the rat. Surgery 1961;50: 668-72. 19. Zieve L, Doizaki WM, Lyftogt C. Brain methanethiol and ammonia concentrations in experimental hepatic coma and coma induced by injection of combination of these substances. J Lab Clin Med 1984;104:665-74.
March 1989
20. Haber Mass.: 21. Benoit portal 22. Webster
BRAIN EDEMA
A, Runyon
RP. General
Addison-Wesley, J, Granger hypertension. AJF.
statistics.
2nd
ed. Reading,
hemodynamics
in chronic
1973.
DN. Splanchnic Semin
Energetics
of maintenance
Girardier L, Stock MJ, eds. Mammalian don: Chapman and Hall, 1983:178-207.
and
growth.
thermogenesis.
In: Lon-
23. Wells LA. Alteration of the blood-brain barrier system by hypothermia: critical time period vs critical temperature. Comp Biochem Physiol 1973;44A:293-6. 24. Porte D Jr, Halter JB. The endocrine pancreas and diabetes mellitus. In: Williams RH, ed. Textbook Philadelphia: WB Saunders, 1981:776.
of endocrinology.
25. Nelson SR, Mantz ML, Maxwell JA. Use of specific gravity in the measurement of cerebral edema. J Appl Physiol 1971;30: 268-71. 26. Werten B, Sehgal N, Assali NS. Regional blood flows and vascular resistance 1961;188:473-6.
during
33.
Liver Dis 1986;6:287-98.
hypothermia
in dogs. Am J Physiol
27. Zamora AJ, Cavanagh JB, Kyu MH. Ultrastructural responses of the astrocyte to portacaval anastomosis in the rat. J Nemo1 Sci 1973;18:25-45. 28. Laursen H, Diemer NH. Morphometry of astrocyte and oligodendrocyte ultrastructure after portacaval anastomosis in the rat. Acta Neuropathol (Berl) 1980;51:65-70. 29. Broadwell RD, Salzman M. Expanding the definition of the blood-brain barrier to protein. Proc Nat1 Acad Sci USA 1981;78:7820-4. 30. Laursen H, Westergaard E. Enhanced permeability to horseradish peroxidase across cerebral portacaval anastomosis. Neuropathol
vessels in the rat after Appl Neurobiol 1975;3:
2943. SJ, Kirsch RE, Biebuyck JF. Correlation 31. Mans AM, Saunders of plasma and brain amino acid and putative neurotransmitter alterations during acute hepatic coma in the rat. J Neurothem 1979;32:285-92. 32. Erlich M, Plum F, Duffy T. Blood
and brain
ammonia
concen-
34.
35.
36.
37.
38.
39.
AFTER
HEPATIC
DEVASCULARIZATION
891
trations after portacaval anastomosis. Effects of acute ammonia loading. J Neurochem 1980;34:1538-42. Schenker S, Warren KD. Effect of temperature variation on toxicity and metabolism of ammonia in mice. J Lab Clin Med 1962;60:291-301. Butterworth RF, Girard G, Giguere JF. Regional differences in the capacity for ammonia removal by brain following portacaval anastomosis. J Neurochem 1988;51:486-90. Voorhies TM, Erlich ME, Duffy TE, Petit0 CK, Plum F. Acute hyperammoniemia in the young primate: physiologic and neuropathologic correlates. Pediatr Res 1983;17:970-5. Norenberg MD. A light and electron microscopic study of experimental portal-systemic (ammonia] encephalopathy. Progression and reversal of the disorder. Lab Invest 1977;36: 618-27. Gregorios JB, Mozes LW, Norenberg MD. Morphologic effects of ammonia on primary astrocyte cultures. II. Electron microscopic studies. J Neuropathol Exp Neurol 1985;44:404-14. Zaki AEO, Wardle EN, Canalese J, et al. Potential toxins in acute liver failure and their effects on blood brain barrier permeability. Experientia 1983;39:988-91. Shapiro HM, Wyte SR, Loeser J. Barbiturate-augmented hypothermia for reduction of persistent intracranial hypertension. J Neurosurg 1974;40:90-100.
October 7, 1988. Received May 26, 1988. Accepted Address requests for reprints to: Andres T. Blei, M.D., Lakeside Veterans Administration Medical Center, Department of Medicine, 333 East Huron Street, Chicago, Illinois 60611. This study was supported by a Merit Review of the Veterans Administration Research Service and the OS. Sprague fund at Northwestern University, Chicago, Illinois. Dr. Peter Traber was a recipient of an American Liver Foundation Fellowship Award. The authors thank Dr. Roger Butterworth for critical review of the manuscript. This work was presented at the III Simposio Hispanoparlante de Hepatologia, Buenos Aires, May 1988.
1989;96:885-91
Effect of Body Temperature on Brain Edema and Encephalopathy in the Rat After Hepatic Devascularization PETER ANDRES
TRABER, T.
MAURO
DALCANTO,
DANIEL
and
BLEI
Gastroenterology and Neuropathology Sections, Departments Lakeside Veterans Administration Hospital and Northwestern
Brain edema is a fatal complication of fulminant hepatic failure and its pathogenesis remains unclear. To determine its presence in a model of ischemic hepatic failure, rats were subjected to a portacaval anastomosis followed by hepatic artery ligation. Brain water was measured using the sensitive gravimetric method. Preliminary studies revealed marked hypothermia in devascularized animals kept at room temperature (26.9” f 28°C). An additional group of devascularized rats was kept in an incubator. As expected for hypothermia, such animals had a lower arterial pressure and heart rate; the duration of encephalopathy was markedly prolonged. Water content of the cortical gray matter was only increased in normothermic devascularized rats: 80.14% f 0.31%, normal: 80.06% + portacaval shunt only; 80.42% + 0.26%, 0.22% devascularized at room temperature: 81.29% f 0.38%, devascularized at controlled temperature (p < 0.001). Such differences could not be detected using the dry-weight technique in whole cerebral hemispheres. Astrocyte changes in the cortical gray matter were noted in both edematous and nonedematous devascularized groups, coupled with the presence of vesicles containing horseradish peroxidase in the endothelial capillary cell. This suggests that in this model, brain edema may be due to both a cytotoxic mechanism and changes in the permeability of the blood-brain barrier. Future studies with this widely used model will require strict control of temperature to allow interpretation of experimental results. A therapeutic role for hypothermia in the management of brain edema deserves further attention.
rain edema is frequently a fatal complication of fulminant hepatic failure (FHF) (1). Its pathogenesis, however, remains obscure; this reflects, in part,
B
GANGER,
of Medicine University,
and Pathology, Chicago, Illinois
the difficulty in performing clinical studies in critically ill patients. Furthermore, few studies have attempted to measure water content of the brain or the mechanisms responsible for water accumulation in experimental models of FHF. In the hepatectomized rat, Livingstone et al. (2) reported a 5.4% increase in water content of gray matter at the time of death coupled with gross changes in the permeability of the blood-brain barrier (BBB). Three groups have noted an increase in the water content of the brain in the galactosamine-treated rat (S-S), although in two studies this was noted only in the hindbrain (3,~). However, in all these studies water was measured using the dry-weight technique, an insensitive method to detect changes in brain water (6). Pathogenically, cytotoxic factors have been incriminated (4); evidence both for (7,8) and against (9,101 changes in the permeability of the BBB has been presented. We have recently shown in the rabbit that galactosamine-induced FHF was associated with a selective increase in the water content of cortical gray matter (11); such changes did not appear to be due to hypotension, hypoxia, acidosis, or hypoglycemia. Electron-microscopic evaluation revealed swollen astrocytes in cortical gray matter (12). No evidence of horseradish peroxidase after intravenous injection was present in capillary endothelial cells or brain parenchyma, indicating a grossly intact BBB. To measure brain water we used the gravimetry technique (IS), a sensitive method that can determine water content in small tissue specimens. We suggested the presence of a cytotoxic mechanism that may induce cellular swelling in this model of FHF.
Abbreviations used in this paper: BBB, blood-brain barrier; FHF, fulminant hepatic failure. 0 1989 by the American Gastroenterological Association 0016-5085/89/$3.50
886
GASTROENTEROLOGY
TRABER ET AL.
In humans, such a mechanism may also be present, as sera of patients with FHF in advanced coma was found to inhibit Na+, K+-adenosine triphosphatase activity of rat brain membranes (14). To gain further insight into the pathogenesis of brain edema, we studied an anhepatic model of FHF, namely, the rat after hepatic devascularization. This model has been used extensively to study the pathogenesis of hepatic encephalopathy (15). In a recent communication, Zieve (16)noted that the brains of such animals do not exhibit an increase in water content as measured with the dry-weight technique; this finding contrasts with the water accumulation seen in the hepatectomized rat (2). As in these latter studies the role of covariables that could influence water accumulation in the brain was not rigorously examined, we continuously monitored several factors that could affect any potential changes in brain water. Among them we noted the previous reports of (17). hypothermia after hepatic devascularization Our results suggest that this change in body temperature greatly affects the characteristics of this model, particularly those pertaining to cerebral edema.
Materials and Methods Adult male Sprague-Dawley rats, weighing 300400 g, were anesthetized with ether. An end-to-side portacaval anastomosis was constructed according to the method of Lee and Fisher (18), with a loose ligature left around the hepatic artery. Once the animals had recovered, they were housed in individual cages for 24-48 h. Under ether anesthesia, a laparotomy was then performed to tie off the hepatic artery. At the same time, a PE-50 catheter (Intramedic) was introduced into a femoral artery for measurement of arterial pressure and blood gases and the other end was tunneled under the skin, exteriorized over the dorsal aspect of the back, and sealed with a metal plug. A thermistor wire for temperature measurement was left between bowel loops and exteriorized at the same site; the abdomen was closed with double suture layering. Preliminary studies indicated that animals that had undergone hepatic devascularization developed a marked reduction in body temperature. Therefore, animals were divided into three groups: control (portacaval anastomosis but at the time of the second laparotomy no hepatic artery ligation), ischemic FHF kept at room temperature (IFHFA), and ischemic FHF kept at controlled temperature (IFHF-B). In the latter group, animals were kept in an incubator and temperature was maintained at 375°C with the use of an infrared heating lamp. Animals were continuously monitored after ligation of the hepatic artery. Blood glucose was measured every 2 h from tail vein samples with the use of glucose strips. Supplemental glucose [as DlOW), in volumes of up to 5 ml, was administered subcutaneously as needed to maintain a blood glucose level above 140 mg% (19). Blood pressure was measured by connecting the arterial line to a
Vol. 96, No. 3
P23 Gould-Statham transducer and the tracing was continuously inscribed in a Hewlett-Packard 7754A recorder. Heart rate was derived from the arterial pressure tracing. Temperature was monitored continuously by connecting the thermistor to a digital display unit (Bailey Instruments, N.J.). Animals were monitored until stage 4 encephalopathy developed, as defined by Zieve et al. (19). At this stage, animals are responsive to pain but are unable to assume a standing position when lying on their side. Arterial blood gases were determined and corrected for hypothermia when necessary. Animals were killed by decapitation and truncal blood was collected for chemical measurements. This study was approved by the Animal Care and Use Committee at Northwestern University.
Measurement
of Brain
Water
The content of brain water was measured by determining tissue specific gravity after flotation of tissue samples in a bromobenzene-kerosene gradient, as described (11). Four K,SO, solutions of different concentrations were used to calibrate the specific gravity gradients, with a linear coefficient of correlation >0.99. With this technique, brain tissue is gently placed in such a column and the equilibrium position is recorded after 2 min (13). The specific gravity can then be calculated from the previously determined linear equation and expressed as water content, using the following equation: Tissue water (%) = 1 -
(%t - 1) 1(1 - %,b!&
1x
100,
where sg, is the specific gravity of the tissue studied and sg, is the specific gravity of the solid component. The latter was calculated in five nonmanipulated control rats using the ratio of wet (W) to dry (D) weight obtained after surgical dissection of gray matter and freeze-drying for 48 h at -70°C using the following equation:
The values obtained are in agreement with our previous study (11) and other reports (6,13). As soon as the brain was carefully removed, it was placed in an airtight container. Using a Z-mm-diameter needle, four punch samples were taken from the cortical gray matter of the left hemisphere; subcortical white matter could not be accurately sampled due to the thinness of the tissue. Each sample was promptly placed in the calibrated column and tissue water percentage was calculated as previously described. The right cerebral hemisphere, brainstem, and cerebellum were weighed, placed in an oven, and reweighed until a stable dry weight was attained at 24 h. Tissue water could then be calculated as follows: Tissue water (%)
=
Wmitia~- ha,, Whitid
Electron Microscopy In 3 rats from each group, the integrity of the BBB was tested using horseradish peroxidase. Thirty minutes
March I!389
Table
BRAIN EDEMA AFTER HEPATIC DEVASCULARIZATION
1. Characteristics
-Temperature (“C) Time after HAL (h) MAP (mmHg) Heart rate Supplemental glucose (ml)
of the Model Control (n = 5)
37.5 t 0.4 11.9 C 4.2b 89 ? 12 430 k 48 21 5 8.2
IFHF-A (n = 5) 26.9 2 17.0 2 60 2 213 k 26 k
2.8" 3.1" 25" 102" 5.5
IFHF-B (n = 5) 38.4 2 1.6 9.4 t 1.0 140 2 18 412? 55 20 ?I 3.5
HAL, hepatic artery ligation; MAP, mean arterial pressure. Values are mean * SD. Control: portacaval anastomosis only. IFHF-A: portacaval anastomosis and hepatic artery ligation kept at room temperature. IFHF-B: devascularized rats kept at 37.5% a p < 0.01 vs. other two groups. b Sham hepatic artery ligation. ’ p < 0.01 vs. IFHF-B.
before killing the rats, 200 mg of horseradish peroxidase type II (Sigma Chemical Co., St. Louis, MO.) dissolved in phosphate-buffered saline was injected into a femoral vein catheter that had been previously placed at the time of the second operation. The rats were anesthetized with ketamine, a tracheostomy was performed, and a polyethylene catheter was placed in a femoral artery. Arterial blood gases were checked for adequate oxygenation. The chest was opened and chilled phosphate-buffered saline followed by 4% formaldehyde and 3% glutaraldehyde in phosphate-buffered saline was infused via a cannula placed in the left ventricle. A pulsatile pump assured a delivery of 200 ml/min; blood pressure was within normal values during the perfusion. After removing the brain, multiple s-mm-thick sections were removed from cortex, cerebellum, and basal ganglia and cut in a Vibratome into sections 15-30 pm thick. These sections were treated for 30 min in a solution of 0.05% diaminobenzidine in Tris buffer with added 0.01% H,O,. The sections were then washed, postfixed with 1% osmic acid, dehydrated through a graded series of ethyl alcohol, cleared in propylene oxide, and embedded in
Epon. The sections were mounted on glass slides before polymerization to be observed under light microscopy. Selected sections were then polymerized for both l-pm and ultrathin sectioning. Grids were studied with and without counterstaining in a Joel-100 CX II electron microscope.
Statistical
Analysis
Comparison among the three experimental groups was made with one-way analysis of variance; differences between means were analyzed with Tukey’s HSD test (20). All values are expressed as mean ? SD.
Results
887
with a mean temperature of 26.9”C and a range of 22.5”-30°C. No evidence of shivering was noted in spite of the marked reduction in temperature. Animals of group IFHF-B remained at 37”-38°C with the use of a heating lamp. In 2 rats, temperature rose suddenly to 38.2”C and 41.2%. Devascularized rats exhibited marked differences in the time elapsed to stage 4 encephalopathy. The range was 12.5-20 h in IFHF-A rats and 8.25-11 h in the IFHF-B group (p < 0.01). In 2 animals of the latter group, encephalopathy was associated with stimulus-induced seizurelike activity, characterized by rapid movements of head and limbs, without a tonic phase. The control rats were studied at times within the range of the devascularized animals (9-19 h1. Marked differences in heart rate and blood pressure were also seen in the two devascularized groups. As anticipated with this degree of hypothermia, IFHF-A rats exhibited lower arterial pressure and heart rate (Table 1). The animal with the lowest body temperature (23%) had a blood pressure of 37/ 16 mmHg and a pulse of 36 beats per minute. Rats of the IFHF-B group showed mild elevations in blood pressure when compared with controls; however, portacaval-shunted rats exhibited lower values of arterial pressure than normal awake animals in our laboratory (100-110 mmHg), a decrease that probably reflects the hyperdynamic circulation seen after this procedure (21). Laboratory
Tests
The amount of supplemental glucose administered to all three groups did not differ (Table 1). However, blood glucose was raised in the IFHF-A group (Table 2); this probably reflects impaired utiof glucose in the presence of a low body lization
temperature. Hematocrits were similar in all animals. Both devascularized groups exhibited hypocapnia, suggesting a component of respiratory alkalosis, mild hyponatremia, and a rise in creatinine. Marked elevations of transaminases were noted in these animals; however, these findings should be interpreted with caution as collection of blood by decapitation may be contaminated by leakage of enzymes of muscle origin. This was noted when truncal blood collections in a group of 5 normal decapitated rats still showed elevations of serum glutamic oxaloacetic transaminase when compared with values obtained from the inferior vena cava in normal rats.
General Aspects The body temperature of the control group was normal (Table 1). Rats that underwent hepatic devascularization (IFHF-A) developed hypothermia,
Water Content of the Brain cient
Using the gravimetry technique, the coeffiof variation for the four measurements of
888
GASTROENTEROLOGYVol. 96, No. 3
TRABER ET AL.
Table
2. Laboratory
Table
Results IFHF-A
Control Glucose (mg%) Hematocrit (%) PH PO, (torr) Pco, (torr) Creatinine (mg%l Sodium (mEq/L) Bilirubin (mgO/o) Alkaline phosphatase
164 f 39 c 7.47f 85 f 33 k 0.582
31 3 0.03 21 6 0.05
390 f 412 7.53+ 102 + 25 2 1.12
144 2 3
IFHF-B 153 + 119 42 2 5 7.49-+0.13 103 f 5 23 k 4'= 1.62 0.36"
234" 2 0.09 9 3= 0.55
127 +ll
128 2 12
2.1+ 0.5"
2.52 0.8"
0.132 0.17
8512 466b
169 2 95
649 + 32gb
(NJ/ml)
SGOT (U/ml]
793 ‘- 673
cortical gray matter within each animal ranged from 2.2% to 5.8%. Devascularized rats kept in a warm environment had a significant increase in the water content of the cortical gray matter [Figure 1). Control animals had a water content of 80.06% + O.22%, the IFHF-A group 80.42% + 0.26%, and the IFHF-B group 81.29% ? 0.38% (F = 24.28, p < 0.001). Results in the control group were comparable to 5 normal noninstrumented rats in which water content of the cortical gray matter was 80.14% ? 0.31%. The water content of the other cerebral hemisphere, brainstem, and cerebellum was also calculated using the dry-weight method (Table 3). Only the cerebellum showed a significant increase in the
82
.
81
: A
80
: .
:
79
0
C
Cerebral hemisphere Brainstem Cerebellum
Weight Method
Control (5)
IFHF-A (5)
IFHF-B (5)
80.372 0.51
79.62k 0.98
80.59+ 0.69
77.62+ 0.79 78.87C 1.23
77.55k 0.48 77.752 1.61
78.352 0.73 80.17t 0.54"
Control: portacaval anastomosis only. IFHF-A: portacaval anastomosis and hepatic artery ligation kept at room temperature. IFHF-B: devascularized rats kept at 37.5"C. Values are mean f SD. All results are percentage of water. a p < 0.05vs.IFHF-A.
water content with ischemic
of the IFHF-B group when compared rats kept at room temperature.
18353 f 5511b 28276k 7577b
SGOT, serum glutamic oxaloacetic transaminase. Control: portacaval anastomosis only. IFHF-A: portacaval anastmosis and hepatic artery ligation kept at room temperature. IFHF-B: devascularized rats kept at 37.5”C. Values are mean * SD. ’ p < 0.05 vs. control. b p < 0.01 vs. other two groups.
0n z
3. Water Content/Dry
IFHF-A
IFHF-B
Figure 1. Water content of cortical gray matter in individual experiments using the gravimetric technique. A clear increase in brain water was seen in devascularized rats kept at controlled temperature (IFHF-B) as compared with animals with portacaval anastomosis (C) or devascularized kept at room temperature (IFHF-A).
Electron
Microscopy
In all three groups, horseradish peroxidase was present in brain capillary endothelial cells, mainly as endothelial vesicles, without accumulation in the interstitium or cellular elements [Figure 2). In all three groups, astrocytes of the cortical gray matter appeared with variable degrees of cytoplasmic vacuolation and swelling of pericapillary foot processes. Qualitative assessment of white matter astrocytes did not reveal such changes.
Discussion In the present study, we observed that hypothermia after hepatic devascularization affected several physiologic parameters in this model while prolonging the period of encephalopathy. Furthermore, maintenance of a normal body temperature in devascularized rats was associated with an increase in brain water. Studies of the encephalopathy of FHF in the experimental animal have proceeded via one of two experimental models: toxic injury or the anhepatic state. Thus, the significance of experimental findings could be enhanced by determining their presence in different preparations rather than selecting a single model for study. Each model can be criticized for its shortcomings. The one used in this study, for example, does not have a potential for recovery, limiting the assessment of therapeutic maneuvers to the possible prolongation of encephalopathy rather than to survival. Marked differences are present in metabolic disturbances, even within different species. Such is the case for brain edema in FHF, a complication whose pathogenesis is still not completely understood. Our previous study in the rabbit with galactosamine-induced FHF suggested that a cytotoxic mechanism could account for the increase in cortical gray matter water, as detected with the gravimetry method (11). Neuropathologically, we
March 1989
Figure
BRAIN EDEMA AFTER HEPATIC DEVASCULARIZATION
889
2. Electron micrographs from cortical vessels derived from (a) rats with portacaval anastomosis, lb) rats with portacaval anastomosis and hepatic artery ligation, and (c) devascularized rats kept at controlled temperature. A modest number of vesicles containing horseradish peroxidase (arrows) are present in endothelial cells in all three instances. Increased vacuolation of underlying astrocytic foot processes can also be observed. Magnification, ~14,000.
noted swelling of gray matter astrocytes and the absence of a gross breakdown of the BBB (12). Determining these parameters in another model of FHF would be important to support the presence of a cytotoxic mechanism. The rat after hepatic devascularization developed marked hypothermia. A decrease in body temperature has been previously noted in anhepatic models of FHF (2,17). One possible explanation lies in the contribution of hepatic metabolism to the generation of body heat, estimated to be -12% under resting conditions in the adult rat (22). To explain the precipitous drop in body temperature, it is possible that compensatory mechanisms after devascularization, via shivering or nonshivering thermogenesis, were not activated; alternatively, adequate central regulation of body temperature may have been absent. Regardless of the mechanism, severe hypothermia can be associated with encephalopathy and changes in the BBB (231, and thus become a confounding factor in the interpretation of experimental results. In our hypothermic rats, arterial pressure and heart rate were reduced as an appropriate physiologic response to a decrease in body temperature. The moderate elevation of plasma glucose values could reflect its diminished utilization, a common finding in hypothermia (24). Devascularized animals maintained at normal body temperature had a shorter duration of encephalopathy; in 2 animals,
seizure activity was observed. Although it has been stated that the decrease in body temperature simply slows the progression of hepatic encephalopathy (16), our studies suggest multiple biological differences. Brain water was increased in the cortical gray matter of “warm,” devascularized rats. No overlap was seen with the portacaval anastomosis group and with 6 normal noninstrumented rats. The unchanged values of water content of the whole cerebral hemisphere with the dry-weight method may again reflect its relative insensitivity for studies of brain edema in FHF. With the gravimetry technique, a mean increment of 1.2% in brain water was seen at an intermediate stage during the development of encephalopathy, corresponding to an -8% rise in brain volume (251.
To explain this increase in brain water, several factors need to be considered. Differences in cerebral hemodynamics may be present; a “protective” role for hypothermia in “cold” animals could be mediated by a decrease in cerebral blood flow (26). Although seizure activity could have potentially contributed to the rise in brain water, water content in the 2 affected animals was similar to that of the entire group. Our morphologic studies suggest the presence of both cellular changes and an increased permeability to horseradish peroxidase. As in the galactosamine rabbit, the increase in water probably resides in
890
TRABER
GASTROENTEROLOGY
ET AL.
astrocytes of the cortical gray matter. It should be noted that animals with portacaval anastomosis alone also exhibited astrocytic changes. Marked cytoplasmic vacuolation of astrocytes in the first week after portacaval shunting in the rat has been previously reported (27,281. On the other hand, all three groups showed vesicles containing horseradish peroxidase within the capillary endothelial cell; this suggests lysosomal activation and increased pinocytotic transfer across the BBB (29). Such endothelial vesicles have also been reported after portacaval anastomosis alone (30). As horseradish peroxidase was not seen in the basement membrane or within the brain tissue, it is unlikely that a gross breakdown of the BBB had occurred. The overall neuropathological similarity between groups suggests that a similar pathogenic mechanism may be present. Unfortunately, quantitative estimations of astrocyte swelling are hampered by the small cellular soma and the multiple projections represented by the foot processes. Still, the net accumulation of water in the “warm” group could result from a higher concentration of a putative cytotoxin, such as ammonia. Tenfold increases of ammonia levels reported in the whole brain of “warm” devascularized rats (31) contrast with the doubling of values seen after portacaval anastomosis alone (32) and the threefold rise seen in “cold” animals (19). Of note, the neurotoxicity of ammonia is reduced in hypothermia (33). A further link between astrocyte swelling and ammonia in FHF is the fact that glutamine synthetase is located solely in astrocytes, which remove ammonia by conversion to glutamine; the ability to increase ammonia detoxification via induction of glutamine synthetase is limited (34). Coupled with the observations of ammonia-induced astrocyte swelling in the normal primate (35), similar astrocyte changes seen in rats with portacaval anastomosis given an ammonia load (36), cytoplasmic swelling of astrocytes in vitro after the addition of ammonia (37), and ammonia-induced changes in the BBB (38), we hypothesize that ammonia may be pathogenically related to the increase in brain water seen in FHF. In summary, the devascularized rat model of FHF exhibits major differences in the duration of encephalopathy, neurologic signs, and accumulation of brain water according to the temperature at which the animal is maintained. Brain edema is only seen at normal body temperature; its pathogenesis may be related to both cytotoxic changes and to changes in BBB permeability. These findings suggest that future studies with this model will require strict control of temperature to allow interpretation of experimental results. In view of the clinical use of hypothermia in the treatment of vasogenic edema (39), a role for such
Vol. 96, No. 3
therapy in the management of encephalopathy and brain edema in human FHF may deserve further attention.
References 1. Ware AJ, D’Agostino AN, Combes BC. Cerebral edema: a major complication of massive hepatic necrosis. Gastroenterology 1971;61:887-4. 2. Livingstone AS, Potvin M, Goresky CA, et al. Changes in the blood-brain barrier in hepatic coma after hepatectomy in the 1977;73:697-704. rat. Gastroenterology 3. Groflin R, Tholen H. Brain edema in the rat with galactosamine-induced severe hepatitis. Experientia 1978;34:1501-3. 4. Ede RJ, Gove CD, Hughes RD, Marshall W, Williams R. Reduced brain Na+-K+ ATPase activity in rats with galactosamine-induced hepatic failure: relationship to encephalopathy and cerebral oedema. Clin Sci 1987;72:365-71. 5. Dixit V, Chang TMS. Effects of prostaglandin E2 on brain edema and liver histopathology in a galactosamine-induced fulminant hepatic failure rat model. Biomater Artif Cells, Artif Organs 1987;15:559-73. 6 Shigeno T, Brock M, Shigeno S, et al. The determination of brain water content: microgravimetry versus drying-weighing method. J Neurosurg 1983;59:803-9. 7 Horowitz ME, Schafer DF, Molnar P, et al. Increased bloodbrain transfer in a rabbit model of fulminant hepatic failure. Gastroenterology 1983;84:1003-11. 8. Zaki AEO, Ede RJ, Davis M, et al. Experimental studies of blood-brain Hepatology
barrier
permeability
in
acute
hepatic
failure.
1984;4:359-63.
9. Lo WD, Ennis SR, Goldstein GW, Betz AL. The effects of galactosamine-induced hepatic failure upon blood-brain bar1987;7:452-6. rier permeability. Hepatology 10. Knudsen GN, Poulsen HE, Paulsen DB. Blood-brain barrier permeability in galactosamine-induced hepatic encephalopathy. No evidence for increased GABA transport. J Hepatol 1988;6:187-92. 11. Traber PG, Ganger DR, Blei AT. Brain edema in rabbits with galactosamine-induced fulminant hepatic failure. Regional differences and effects on intracranial pressure. Gastroenterology 1986;91:1347-56, 12. Traber PG, DalCanto M, Ganger D, Blei AT. Electron microscopic evaluation of brain edema in rabbits with galactosamine-induced fulminant hepatic failure. Hepatology 1987;7: 1257-61. 13. Marmarou A, Poll W, Shulman K, et al. A simple gravimetric technique for measurement of cerebral edema. J Neurosurg 1978;49:530-7. 14. Seda HWM, Hughes RD, Gove CD, et al. Inhibition of rat brain Na+-K+ ATPase activity by serum from patients with fulminant hepatic failure. Hepatology 1984;4:74-9. 15. Zieve L. Pathogenesis of hepatic encephalopathy. Metabol Brain Dis 1987;2:147-66. 16. Zieve L. Reply to Muller KD. Hepatology 1987;7:618. 17. Peignoux M, Bernuau J, Benhamou JP. Total hepatectomy and hepatic vascular exclusion in the rat. A comparison, with special reference to the influence of body temperature. Clin Sci 1982;62:273-7. 18. Lee SH, Fisher B. Portacaval shunt in the rat. Surgery 1961;50: 668-72. 19. Zieve L, Doizaki WM, Lyftogt C. Brain methanethiol and ammonia concentrations in experimental hepatic coma and coma induced by injection of combination of these substances. J Lab Clin Med 1984;104:665-74.
March 1989
20. Haber Mass.: 21. Benoit portal 22. Webster
BRAIN EDEMA
A, Runyon
RP. General
Addison-Wesley, J, Granger hypertension. AJF.
statistics.
2nd
ed. Reading,
hemodynamics
in chronic
1973.
DN. Splanchnic Semin
Energetics
of maintenance
Girardier L, Stock MJ, eds. Mammalian don: Chapman and Hall, 1983:178-207.
and
growth.
thermogenesis.
In: Lon-
23. Wells LA. Alteration of the blood-brain barrier system by hypothermia: critical time period vs critical temperature. Comp Biochem Physiol 1973;44A:293-6. 24. Porte D Jr, Halter JB. The endocrine pancreas and diabetes mellitus. In: Williams RH, ed. Textbook Philadelphia: WB Saunders, 1981:776.
of endocrinology.
25. Nelson SR, Mantz ML, Maxwell JA. Use of specific gravity in the measurement of cerebral edema. J Appl Physiol 1971;30: 268-71. 26. Werten B, Sehgal N, Assali NS. Regional blood flows and vascular resistance 1961;188:473-6.
during
33.
Liver Dis 1986;6:287-98.
hypothermia
in dogs. Am J Physiol
27. Zamora AJ, Cavanagh JB, Kyu MH. Ultrastructural responses of the astrocyte to portacaval anastomosis in the rat. J Nemo1 Sci 1973;18:25-45. 28. Laursen H, Diemer NH. Morphometry of astrocyte and oligodendrocyte ultrastructure after portacaval anastomosis in the rat. Acta Neuropathol (Berl) 1980;51:65-70. 29. Broadwell RD, Salzman M. Expanding the definition of the blood-brain barrier to protein. Proc Nat1 Acad Sci USA 1981;78:7820-4. 30. Laursen H, Westergaard E. Enhanced permeability to horseradish peroxidase across cerebral portacaval anastomosis. Neuropathol
vessels in the rat after Appl Neurobiol 1975;3:
2943. SJ, Kirsch RE, Biebuyck JF. Correlation 31. Mans AM, Saunders of plasma and brain amino acid and putative neurotransmitter alterations during acute hepatic coma in the rat. J Neurothem 1979;32:285-92. 32. Erlich M, Plum F, Duffy T. Blood
and brain
ammonia
concen-
34.
35.
36.
37.
38.
39.
AFTER
HEPATIC
DEVASCULARIZATION
891
trations after portacaval anastomosis. Effects of acute ammonia loading. J Neurochem 1980;34:1538-42. Schenker S, Warren KD. Effect of temperature variation on toxicity and metabolism of ammonia in mice. J Lab Clin Med 1962;60:291-301. Butterworth RF, Girard G, Giguere JF. Regional differences in the capacity for ammonia removal by brain following portacaval anastomosis. J Neurochem 1988;51:486-90. Voorhies TM, Erlich ME, Duffy TE, Petit0 CK, Plum F. Acute hyperammoniemia in the young primate: physiologic and neuropathologic correlates. Pediatr Res 1983;17:970-5. Norenberg MD. A light and electron microscopic study of experimental portal-systemic (ammonia] encephalopathy. Progression and reversal of the disorder. Lab Invest 1977;36: 618-27. Gregorios JB, Mozes LW, Norenberg MD. Morphologic effects of ammonia on primary astrocyte cultures. II. Electron microscopic studies. J Neuropathol Exp Neurol 1985;44:404-14. Zaki AEO, Wardle EN, Canalese J, et al. Potential toxins in acute liver failure and their effects on blood brain barrier permeability. Experientia 1983;39:988-91. Shapiro HM, Wyte SR, Loeser J. Barbiturate-augmented hypothermia for reduction of persistent intracranial hypertension. J Neurosurg 1974;40:90-100.
October 7, 1988. Received May 26, 1988. Accepted Address requests for reprints to: Andres T. Blei, M.D., Lakeside Veterans Administration Medical Center, Department of Medicine, 333 East Huron Street, Chicago, Illinois 60611. This study was supported by a Merit Review of the Veterans Administration Research Service and the OS. Sprague fund at Northwestern University, Chicago, Illinois. Dr. Peter Traber was a recipient of an American Liver Foundation Fellowship Award. The authors thank Dr. Roger Butterworth for critical review of the manuscript. This work was presented at the III Simposio Hispanoparlante de Hepatologia, Buenos Aires, May 1988.