Brain damage due to episodic alcohol exposure in vivo and in vitro: furosemide neuroprotection implicates edema-based mechanism MICHAEL A. COLLINS,*,1 JIAN-YUN ZOU,* AND EDWARD J. NEAFSEY† *Department of Molecular and Cellular Biochemistry, and †Department of Cell Biology, Neurobiology, and Anatomy, Stritch School of Medicine, Loyola University Chicago, Maywood, Illinois 60153, USA
Adult rats intubated with a single dose of ethanol (alcohol; Ç5 g/kg) for 5 to 10 successive days incur neurodegeneration in the entorhinal cortex, dentate gyrus, and olfactory bulbs accompanied by cerebrocortical edema and electrolyte (Na/, K/) accumulation. The brain damage is not lessened by cotreatment with the NMDA receptor antagonist MK801; also, as reported elsewhere, MK-801 as well as non-NMDA receptor and Ca2/ channel antagonists are not neuroprotective in a similar, but more compressed, intoxication protocol. However, cotreatment with the electrolyte transport inhibitor/diuretic furosemide reduces alcohol-dependent cerebrocortical damage by 75–85% while preventing brain hydration and electrolyte elevations; olfactory bulb neurodegeneration is not attenuated. In parallel in vitro studies, rat organotypic entorhinal/hippocampal slice cultures exposed to alcohol (50–200 mM) 15 h/day for 6 days, mirroring episodic intoxication in vivo, demonstrate concentration-related release of the cytotoxic indicator, lactate dehydrogenase. Analogous to the in vivo findings, furosemide blocks this alcohol-induced in vitro cytotoxicity. Our results showing neuroprotection by furosemide indicate that brain edema and swelling are essential events in the brain damage induced by episodic alcohol exposure. Furosemide and related agents might be useful as neuroprotective agents in alcohol abuse. We suggest that the neurodegeneration is elicited in part by edema-dependent oxidative stress, but the regional selectivity of the damage may be best explained by physical (mechanical) compression of the limbic cortex against the adjacent tympanic bulla and subsequent neuronal cytoskeletal collapse. A scheme for these apparently nonexcitotoxic metabolic and mechanical pathways initiated by repeated alcohol exposure is proposed.—Collins, M. A., Zou, J.-Y., Neafsey, E. J. Brain damage due to episodic alcohol exposure in vivo and in vitro: furosemide neuroprotection implicates edema-based mechanism. FASEB J. 12, 221–230 (1998) ABSTRACT
Key Words: ethanol · neurodegeneration · excitotoxicity · binge intoxication · entorhinal cortex · dentate gyrus · dizo-
cilpine maleate · olfactory bulb · cell swelling · intracranial pressure
EPISODIC ALCOHOL (ethanol) intoxications and withdrawals characteristic of binge alcoholism result in brain damage (1). The underlying mechanisms are uncertain, but as with other neurodegenerative conditions, interest has focused on synaptically based excitotoxic events involving excessive glutamatergic activity, increased intracellular Ca2/, and perhaps decreased g-aminobutyric acid (GABA)2 -ergic activity (2, 3). Also, a link between alcohol, excitotoxicity, and nitric oxide generation has been suggested (4). However, definitive in vivo evidence supporting a key role for synaptic excitotoxicity in the brain damage induced solely by alcohol is lacking. In vitro studies with neuronal preparations, so informative for other types of brain damage, have so far yielded limited mechanistic insights. In our studies of neurotoxicity subsequent to episodic alcohol exposure, we have used repetitive intoxication of adult rats via gastric intubation (5, 6). Neuronal damage is visualized and quantitated as argyrophilic cell counts with the de Olmos cupric-silver staining method (7). Recent studies document that cupric-silver staining is the method of choice to assess irreversible neurodegeneration induced by a range of toxic insults (8–10). Expanding on a study by Switzer et al. (11), we showed that rats given 8 to 12 g alcohol/kg/day in three or more fractional doses for 4–5 days (the Majchrowicz dependence model) display selective argyrophilia in olfactory and limbic cor1
Correspondence: Department of Molecular and Cellular Biochemistry, Loyola Medical School, Maywood IL 60153, USA. E-mail:
[email protected] 2 Abbreviations: ANOVA, analysis of variance; BAC, blood alcohol concentration; CNS, central nervous system; DG, dentate gyrus; DNQX, 6,7-dinitro-quinoxaline-2,3-dione; EC, entorhinal cortex; GABA, g-aminobutyric acid; ICP, intracranial pressure; LDH, lactate dehydrogenase; LSD, least significant difference; MEM, modified Eagle’s medium; NMDA, Nmethyl-D-aspartate; OB, olfactory bulb; SD, Sprague-Dawley; MK-801, dizocilpine maleate.
0892-6638/98/0012-0221/$02.25 Q FASEB
/ 3820 0011 Mp 221 Wednesday Jan 07 12:03 PM LP–FASEB 0011
221
tical brain regions—notably, entorhinal cortex layer III pyramidal neurons, dentate gyrus granule cells, and olfactory nerve terminals in the olfactory bulb glomeruli—but little in hippocampal CA regions and none in the cerebellum (5, 6). Because of the evidence for excitotoxicity in alcohol-induced brain damage, we anticipated that blocking putative excitotoxic steps would reduce neurodegeneration. However, administration of glutamate receptor and calcium channel antagonists imparted little or no neuronal protection in the thrice-daily Majchrowicz model (12). Furthermore, inhibitors of nitric oxide synthase tended to increase, not decrease, the alcohol-dependent brain damage (6). We report here the use of a modified, milder ‘oncedaily’ intoxication paradigm, in which neurodegeneration in the same brain regions—mainly the entorhinal cortex, dentate gyrus, and olfactory bulbs—is induced by a single daily alcohol dose (Ç5 g/kg) in one-third of the rats after 5 consecutive treatment days and in all rats after 10 days. In addition, for complementary in vitro studies, we have used the rat organotypic entorhinal/hippocampal slice culture (13). This and similar cerebrocortical preparations are now frequently used to study neurodegenerative mechanisms in the relatively intact, mature central nervous system (CNS) (14, 15). In our case, the preparation encompasses the two limbic cortical regions that manifest the most damage during episodic alcohol intoxication. We find statistically significant increases in media lactate dehydrogenase (LDH) when the organotypic brain slice cultures are exposed to high, albeit physiological, concentrations of alcohol in an episodic, fluctuating pattern that mimics the in vivo situation. Widely applied with dispersed neuronal cell preparations, LDH release is likewise considered a valid measure of cytotoxic damage in organotypic slice cultures (16, 17). After confirming in the once-daily alcohol intoxication model that daily cotreatment with MK-801 (dizocilpine maleate), a potent antagonist of the glutamate/NMDA (N-methyl-D-aspartate) receptor, did not suppress the neurodegeneration, other possible mechanisms were considered. It is known that alcohol can cause glial swelling and brain hydration (18–20), and research on the genesis of status epilepticus, which can cause brain damage, has shown a link between nonsynaptic mechanisms of glial swelling and epileptiform discharge activity (21, 22). There have been indications that brain hydration may be a factor in alcoholic brain damage (23), but an association remained to be established experimentally. We therefore examined whether repetitive subchronic alcohol intoxication would induce brain hydration and, if so, whether preventing edema would reduce neuronal damage. Indeed, in rats intoxicated once daily for a period sufficient to result in argyrophilia, the brain was sig222
Vol. 12
February 1998
nificantly hydrated and had elevated electrolytes. Cotreatment with the electrolyte transport inhibitor/diuretic, furosemide, which counters brain cell swelling in epilepsy studies (24), prevented the alcohol-dependent brain water and electrolyte increases and dramatically reduced cerebrocortical neurodegeneration. Also, furosemide completely prevented the alcohol-induced LDH release from organotypic brain slice cultures. These corresponding studies in models that mirror one of the most common patterns of alcohol abuse, binge alcoholism (25), provide a convincing demonstration of significant pharmacological protection against a specific form of alcohol-induced brain damage. The data suggest that an apparently nonexcitotoxic, edema-based mechanism underlies early brain neurodegeneration due to episodic alcohol exposure.
MATERIALS AND METHODS Materials Furosemide and HEPES were obtained from Sigma Chemical (St. Louis Mo.) and MK-801 was from Research Biochemicals Int. (Natick, Mass.). Modified Eagle’s medium (MEM) media, Hank’s buffer, and horse serum were from Gibco (Gaithersburg, Md.). Tissue culture inserts were from Nunc Inc. (Naperville, Ill.) and plastic ware was from Fisher Scientific (Pittsburgh, Pa.). In vivo studies Male Sprague-Dawley (SD) rats, 270–330 g, from Sasco Laboratories (Wilmington, Mass.) were used after 4–5 days acclimation in the animal facilities. Water and lab chow were available throughout. They were intubated intragastrically with a single morning dose of 25% alcohol (v/v) in isotonic saline or saline only (controls). Alcohol doses ranged between 3.3 and 6.5 g/kg, and averaged 5 g/kg. To determine blood alcohol concentrations (BACs), tail blood samples were taken daily 2 h after alcohol administration, and also 1/2, 8, and 24 h in the study illustrated in Fig. 1A. The samples were analyzed spectrophotometrically with an alcohol dehydrogenase method described elsewhere (6). For the study with MK801, a neuroeffective cumulative dose (75 mg·kg01·day01 s.c.) was given in isotonic saline 8 h after daily alcohol or saline intubation. Furosemide in isotonic saline was administered (10 mg/kg12 i.p.) 1 and 8 h after daily alcohol or saline treatment. Rats were anesthetized with pentobarbital sodium (40 mg/ kg i.p.) 12 or 24 h after the last alcohol or saline dose and were perfused transcardially with 4% paraformaldehyde / 5% sucrose in 0.1 M sodium phosphate-buffered saline (pH 7.4). Brains and olfactory bulbs were removed and postfixed as described (5, 6). Six to 10 horizontal sections (50 mm) were taken at 6 to 8 mm below bregma and stained for degenerating neurons and processes according to de Olmos et al. (7). The maximum number of argyrophilic neurons in a region on one side on any one section reflected the highest amount (severity) of damage; the mean amount of overall damage was obtained from counting and averaging the number of argyrophilic neurons/glomeruli in all brain sections/side/region. Maximal or
The FASEB Journal
/ 3820 0011 Mp 222 Wednesday Jan 07 12:03 PM LP–FASEB 0011
COLLINS ET AL.
Figure 1. Characterization of the once-daily intoxication model in adult male rats: blood alcohol concentrations (BACs) and brain neurodegeneration. A) Daily mean BACs at 1/2, 2, 8, and 24 h during 6 consecutive days of alcohol treatment (3.3 to 6.7 g·kg01·day01, X axis; nÅ5 rats/group). B) Representative photomicrographs showing degenerating (argyrophilic) entorhinal cortical (EC) and dentate gyrus (DG) neurons and argyrophilic olfactory bulb (OB) glomeruli 24 h after 6 days of alcohol treatment. C) Average maximal (Max) and mean numbers ({ SEM) of argyrophilic neurons in the EC and DG and of argyrophilic OB glomeruli 24 h after 5, 6, or 10 days of once-daily alcohol treatment (3.5–6.5 g·kg01·day01; nÅ3–12 rats/group). Rats killed 24 h after last alcohol dose. Mean daily 2hrBAC range, 240–265 mg/dl.
mean counts in alcohol and alcohol plus MK-801 or furosemide groups were compared using nonparametric MannWhitney U tests. To determine brain hydration, brain electrolytes, and plasma osmolality, rats were killed by decapitation. Trunk blood was centrifuged in a clinical centrifuge, and plasma was taken for osmolality measurements by freezing point depression. Brains were removed, wet weights were taken, and dry weights were obtained after 48 h at 1027C (26). The percent of brain water was calculated according to (wet weight-dry weight)/wet weight 1 100. For electrolytes (Na/, K/, Cl0), dried brains were digested for 1 wk in 1 ml 1 N nitric acid. Aliquots (200 ml) were diluted with 2 ml of 3 mM CsCl2 in dionized H20 and analyzed by flame photometry for Na/ and K/; brain Cl0 was determined spectrophotometrically in diluted nitric acid aliquots (27). Percent water and electrolyte means (mEq/g dry brain; percent control shown) were compared with one-way analysis of variance (ANOVA), followed by the post-hoc Tukey least significant difference (LSD) test.
Preparation and study of organotypic entorhinal/ hippocampal slice cultures Slice cultures were prepared from 8- to 10-day-old male and female SD rats (maternal source, Zivic-Miller Labs, Portersville, Pa.). Pups were lightly anesthetized with ethyl ether and decapitated. The entorhinal/hippocampal complex was removed and transverse slices (350 mm) were taken with a McIlwain tissue chopper. Adapting the procedure of Stoppini et al. (13), slices were placed on 0.22 mm tissue culture inserts (Nunc Anopor membranes), Ç4 slices per well, in 6-well Falcon plates with covers. They were cultured at 377C on MEM media containing 25% horse serum, 25% Hank’s buffer, 20 mM HEPES and 6.5 mg/ml glucose in an atmosphere of 95% O2/5% CO2. Slices were periodically examined and those appearing unhealthy (darkened appearance) under the phase microscope were discarded. After 2–3 wk in culture, alcohol was added to the media for 15 h each day (6 PM to 9 AM) for 6 successive days at initial concentrations of 50 to 200 mM, with complete media changes
ALCOHOL-INDUCED BRAIN DAMAGE
/ 3820 0011 Mp 223 Wednesday Jan 07 12:03 PM LP–FASEB 0011
223
at 3 days. Media alcohol concentrations, monitored spectrophotometrically with the procedure used for BACs, were observed to drop 35–50% due to evaporation during the 15-h incubation periods. The initial media concentrations were reestablished by alcohol addition before each daily incubation period. When used, furosemide (1.1 mM) was present in the media throughout the 6-day period. Samples of the pooled incubation and postincubation (9 h without alcohol) media were analyzed for LDH with a Sigma diagnostic kit. Results were standardized to tissue protein determined according to Lowry et al. (28), and means were compared by ANOVA, followed by post-hoc Tukey LSD tests.
RESULTS Characterization of alcohol-induced brain damage in vivo The once-a-day intoxication procedure was characterized, like the more severe thrice-daily Majchrowicz method (5, 6, 11), by determining the extent of brain damage in key regions with respect to treatment duration. The cyclicity of the BACs also was ascertained. Rats were killed 24 h after alcohol or saline (control) intubations lasting from 2 to 10 consecutive days. Figure 1A shows the BACs at four blood sampling times per day during the first 6 days, with alcohol doses ranging from 3.3 to 6.7 g/kg. The mean daily BACs were generally highest 2 h after alcohol administration (2hrBAC) and ranged between 150 and 380 mg/ dl, depending on alcohol dose. The overall mean 2hr BAC in this and other once-daily intoxication studies averaged Ç250 mg/dl, or 100–150 mg/dl below typical values in the 4-day Majchrowicz model (5, 9). Twenty-four hours after alcohol intubation, BACs were below detection for doses of less than 5 g/kg, but were still appreciable for higher doses. No overt convulsions were seen in any rats, and mortality was õ15%. There was no neuronal argyrophilia in any brain region of control rats, consistent with numerous previous stainings of controls (5, 6, 11). Nor was argyrophilia present after one, two, or three alcohol intubations, but scattered argyrophilic cells and/or processes were sometimes observed after four oncedaily treatments. After 5, 6, and 10 sequential daily alcohol doses, 33%, 67%, and 100% of the rats, respectively, displayed neuronal argyrophilia in the entorhinal cortex, dentate gyrus, and/or olfactory bulbs. Figure 1B shows photomicrographs of degenerating (argyrophilic) neurons in these regions of a representative rat after once-daily alcohol intubations for 6 successive days. Figure 1C illustrates the average maximal and mean counts for degenerating cerebrocortical neurons and glomeruli in olfactory bulbs after 5, 6, and 10 alcohol doses. Between the fifth and sixth alcohol doses, a marked increase in degeneration was apparent among entorhinal cortical pyram224
Vol. 12
February 1998
idal neurons and dentate granule cells, which was not seen in the olfactory bulb glomeruli. Alcohol-induced brain damage in vivo and MK-801 To examine the role of the synaptic NMDA receptor in alcohol-induced neurodegeneration, the noncompetitive NMDA receptor antagonist MK-801 was given daily for 8 days in a low daily dose range that suppresses brain damage due to soman intoxication or kainate treatment (29, 30). The results in Fig. 2 show that MK-801 did not significantly diminish the maximal and mean counts for neurodegeneration in the three most vulnerable brain regions, even though the mean 2hrBAC (inset, Fig. 2) was 20% lower in the surviving MK-801 cotreated rats. Cotreatment with MK801 greatly increased mortality (ú45% overall) in more intoxicated rats, resulting in a lower mean 2hr BAC in survivors. The failure of MK-801 to reduce selective neurodegeneration due to once-a-day alcohol treatment is in agreement with our findings when using a higher MK-801 dose (1 mg·kg01·day01 1 4– 5 days) in the more severe thrice-daily intoxication procedure (12). Alcohol, brain edema, brain electrolytes, and furosemide We studied whether the once-daily intoxication caused significant increases in brain hydration and
Figure 2. Effect of MK-801 (75 mg·kg01·day01 i.p.) on maximal and mean numbers ({ SEM) of argyrophilic neurons and/or glomeruli in the EC, DG, and OB caused by 8 consecutive days of once-daily alcohol intoxication (3.8–6.2 g·kg01·day01; nÅ7–9 rats/group). Rats were killed 24 h after the last alcohol dose, and neurodegeneration was quantitated as argyrophilia according to Materials and Methods. There were no significant differences between means for the two groups (nonparametric Mann-Whitney U test). Inset: Mean 2hrBACs in surviving rats with and without MK-801. *P õ 0.05 compared to rats without MK-801 (two-tailed Student’s t test).
The FASEB Journal
/ 3820 0011 Mp 224 Wednesday Jan 07 12:03 PM LP–FASEB 0011
COLLINS ET AL.
electrolytes, and if so, whether the imbalances would be corrected by the diuretic, furosemide. Figure 3A reveals that brain water in rats killed 12 h after the eighth once-daily alcohol treatment was significantly elevated compared to controls. Calculated according to Elliott and Jasper (31), the edema represented 2.5% brain swelling. As shown, daily cotreatment with 20 mg/kg furosemide in two fractional doses completely inhibited alcohol-induced brain edema. As depicted in Fig. 3B, brain Na/ and K/ concentrations were significantly elevated 12 h after the eighth daily alcohol dose, and the daily furosemide cotreatment also prevented the alcohol-induced
brain accumulation of these two electrolytes. In an experiment in which rats were killed 2 rather than 12 h after the eighth alcohol dose (not shown), only brain K/ was increased significantly, suggesting a differential or temporal pattern of electrolyte accumulation that requires further study. Brain Cl0 was not increased at either the 2 or 12 h time points, but in the 12 h experiment (Fig. 3B), cotreatment with furosemide reduced Cl0 levels significantly below control. In results not shown, furosemide did not alter the overall mean 2hrBAC (mg/dl { SEM: alcohol, 262{14; alcohol plus furosemide, 260{6; tÅ0.07, PÅ0.94) nor did it prevent the well-documented increase in mean plasma osmolality (33) due to the presence of alcohol (mOsm { SEM: control, 306{2; furosemide, 311{3; alcohol, *340{10; alcohol plus furosemide, *341{13; FÅ5.09, PÅ0.0072, *Põ0.01 vs. control). An analysis of the daily 2hrBAC pattern (not shown) revealed 10% higher mean values on days 1 and 2 for the furosemide/alcohol group, but no differences between the two alcohol-treated groups on other days. The diuretic was well tolerated, and mortality was the same for the two alcohol-treated groups. Furosemide protection of alcohol-induced brain damage in vivo
Figure 3. A) Daily alcohol intoxication significantly increases brain water in adult rats, and the increase is prevented by furosemide cotreatment (20 mg·kg01·day01 i.p. in two doses). Rats were killed 12 h after the eighth alcohol dose (3.5–6.2 g·kg01·day01; nÅ7 rats/group). Mean 2hrBACs (262{14 vs. 260{6 mg/dl) were not different between the two intoxicated groups (Student’s t test). Symbols in each column denote individual rat values. All results were compared with one-way ANOVA and the post-hoc Tukey LSD tests. *P õ 0.01 compared to control; **P õ 0.01 compared to alcohol only (FÅ13.18, PÅ0.00001). B) Daily alcohol intoxication significantly increases brain Na/ (FÅ5.81, PÅ0.004) and brain K/ (FÅ6.77, PÅ0.002) in adult rats in Fig. 3A, and the increases (expressed as percent change from control { SEM) are prevented by the furosemide cotreatment. Brain Cl0 concentrations were not significantly changed by the alcohol treatment, but were reduced significantly below control by furosemide cotreatment (FÅ3.71, PÅ0.02). *P õ 0.01 compared to control, **P õ 0.01 compared to alcohol group.
The experimental question was whether furosemide would protect against the brain damage concomitant with preventing brain water and electrolyte increases in the alcohol-treated rats. Figure 4 shows that in the two vulnerable cortical regions, furosemide provided a large degree of neuroprotection during a 10-day regimen of daily alcohol intoxications that normally results in brain argyrophilia in all rats. Compared to the alcohol-only group, maximal and mean entorhinal cortical neurodegeneration was 85% less in the furosemide-alcohol group, and the percentage of rats showing any alcohol-induced entorhinal cortical damage was lowered by furosemide from 100% to 55% (not shown). In the dentate gyrus, furosemide significantly reduced the amount of alcohol-dependent neurodegeneration by Ç75% (Fig. 4). Also, the percentage of rats with any granule cell damage dropped from 100% to 78% (not shown). In the olfactory bulbs, furosemide was not neuroprotective nor did it lower the percentage of rats with neuronal glomerular damage in that structure. Consistent with experiments on brain water and electrolytes, neither the overall mean 2hrBACs for 10 days (inset, Fig. 4) nor the daily mean 2hrBACs (not shown) differed between the two alcohol-treated groups. Alcohol-induced cytotoxicity in organotypic brain slice cultures and prevention by furosemide The intact organotypic rat brain slice in prolonged culture was examined as an in vitro model for alco-
ALCOHOL-INDUCED BRAIN DAMAGE
/ 3820 0011 Mp 225 Wednesday Jan 07 12:03 PM LP–FASEB 0011
225
Furosemide’s blockade of the alcohol-dependent LDH elevations during the postincubation period was observed in three replicate experiments.
Figure 4. Effect of furosemide (20 mg·kg01·day01 i.p.) on the average maximal and mean numbers ({ SEM) of argyrophilic EC and DG neurons and OB glomeruli of adult rats caused by once-daily intoxication with alcohol. Rats were killed 24 h after the 10th alcohol dose (nÅ9–12 rats/group). *P õ 0.05 compared to alcohol (nonparametric Mann-Whitney U test). Inset: Mean 2hrBACs (mg/dl { SEM) were not different between alcohol- and furosemide/alcohol-treated groups.
hol-induced damage to mature brain and possible neuroprotection by furosemide. Nissl staining of selected organotypic rat brain slices after 2–3 wk in vitro revealed well-defined entorhinal cortical, hippocampal CA, and dentate gyrus regions (not shown). In preliminary studies with these cultures, maintaining alcohol concentrations at Ç100 mM in a closed system with no alcohol-free episodes for 4 to 6 consecutive days failed to promote cytotoxicity as indicated by media LDH activity, in general agreement with others (32). However, as shown in Fig. 5A, exposure to alcohol (initial concentrations of 50, 100, 150, or 200 mM) in 15 h episodes, followed by 9-h ‘withdrawal’ (no alcohol) periods for 6 consecutive days—simulating episodic intoxication in vivo— elicited concentration-related increases in LDH activity in the incubation (alcohol present) and postincubation (alcohol withdrawn) media pooled from days 4 to 6. During each 15-h interval, alcohol concentrations dropped to 30, 65, 80, or 95 mM, respectively; these levels were replenished to the appropriate initial concentration for the subsequent incubation period. As shown in Fig. 5B, the increases in LDH release (35–50% above control) caused by episodic exposure to alcohol (150 mM initial concentration, dropping to 80 mM during each repetitive 15 h incubation) were completely prevented by continuous cotreatment with 1.1 mM furosemide, indicating effective cytoprotection. 226
Vol. 12
February 1998
Figure 5. A). Rat organotypic entorhinal/hippocampal slice cultures treated with alcohol (50, 100, 150, or 200 mM, initial concentrations) 15 h daily for 6 consecutive days show significantly increased release of LDH (percent from control { SEM). Media pooled from days 4–6, means from six cultures/ group. Incubation Å alcohol present; postincubation Å 9 h without alcohol. *P õ 0.05, incubation compared to control (Student’s t test). B) Furosemide (1.1 mM) blocks induced LDH release from rat organotypic entorhinal/hippocampal slice cultures exposed episodically for 6 days to alcohol (150 mM, initial concentration), as in Fig. 5A. Medium pooled from days 4–6; values (means { SEM) are from a representative experiment with n Å 6 well cultures/group. *P õ 0.05 compared to control; **P õ 0.001 compared to alcohol (one-way ANOVA, followed by post-hoc Tukey LSD tests).
The FASEB Journal
/ 3820 0011 Mp 226 Wednesday Jan 07 12:03 PM LP–FASEB 0011
COLLINS ET AL.
DISCUSSION These studies show that episodic alcohol intoxication of adult rats for a restricted period—either several times daily (5, 6, 12) or, in this report, simply once a day—produces irreversible neurodegeneration in selected olfactory and limbic cortical regions, and the damage is accompanied or preceded by electrolyte accumulation and brain edema. Also, when exposed to alcohol discontinuously rather than continuously, organotypic cerebrocortical/hippocampal slice cultures exhibit statistically significant LDH release that is directly dependent on initial alcohol concentrations, thus providing a new model for in vitro studies of alcohol-induced brain damage. The alcohol concentrations producing cytotoxicity in this brain slice culture model, although high, are not unusual during alcohol detoxification (33); however, they are still lower than concentrations often needed to obtain significant neurodegeneration in dispersed neuronal tissue cultures (34, 35). We suspect that alcohol’s induction of cytotoxicity in the mature brain slice culture is a consequence not only of the repetitive daily exposures, but of the declining alcohol concentrations during each exposure, i.e., conditions occurring in binge intoxication. A ramification of both models’ findings, consistent with observations in human subjects (36, 37), is that only a limited number of severe intoxication bouts may cause selective and lasting brain deficits in alcohol abusers or heavy social drinkers. With many kinds of brain damage, neuroprotection by MK-801 is taken as evidence of involvement of the NMDA receptor. In our report showing no reductions in the entorhinal cortical damage induced by thrice-daily alcohol intoxication for 4–5 days using adult male rats supplied from the former Holtzman Laboratories, the MK-801 dose was 1 mg·kg01·day01 (12). In contrast, in a preliminary experiment none of the virus-free Sasco Laboratories SD rats (used for studies herein) survived beyond 5 days during oncedaily cotreatment with alcohol and a similar MK-801 dose. However, most Sasco rats did survive alcohol treatment with MK-801 at 75 mg·kg01·day01 (a total of 600 mg/kg) over an 8-day period. Although low, the latter dose is nevertheless in a neuroprotective range; comparable acute doses (30–100 mg/kg) of the potent NMDA receptor antagonist effectively reduced damage to the adult brain in other situations— for example, soman intoxication in guinea pigs (29) and kainic acid treatment in rats (30). We conclude that the lack of neuroprotection by MK-801 in both the once-daily or thrice-daily intoxication models and by DNQX or nimodipine in the thrice-daily model (12) rules against a primary pathophysiological mechanism for alcohol that involves synaptic glutamate (ionotropic) receptors or voltage-gated Ca2/ channels.
Considering alternative mechanisms, we were intrigued by a possible connection between neurodegeneration and brain hydration (edema), a condition we determined was present in the alcoholtreated rats. Brain edema is a known phenomenon in alcohol abuse that was suggested more than a decade ago to be causatively linked to CNS damage (23). The edema could be promoted by the total body water and plasma volume expansion observed in alcoholics during withdrawal (38, 39), possibly secondary to inappropriate vasopressin secretion (39, 40). However, alcohol also can act directly to cause glial swelling and hydration through nonsynaptic induction of ionic and osmolar imbalances (18, 19). In experimental status epilepticus, which can result in brain damage, neuronal hyperexcitability depends mainly on synaptic receptor/ion channel events, but a second underlying feature, neuronal hypersynchronization, may be linked to evoked increases in brain (especially glial) cell volume (22, 24, 41). In those studies, furosemide prevented hippocampal cell swelling and derived neuronal hypersynchronization, both in the presence and absence of Ca2/ (24). This apparently nonsynaptic action is thought to involve the furosemide-sensitive electrolyte (Na,K,2Cl) cotransporter on glial cells (24, 42). In primary spinal cord cultures, furosemide and related electrolyte transport blockers has been found to inhibit the NMDA receptor (43); however, NMDA receptor antagonism clearly was insignificant in the furosemide-exposed adult hippocampus (24), and thus may be unique to receptor systems in the primary fetal neurons. In our alcohol experiments, furosemide provided a high degree of neuroprotection in the entorhinal cortex and hippocampal dentate gyrus of alcoholintoxicated adult rats, while concomitantly preventing the cerebrocortical electrolyte increases and edema. Also consistent with the possibility that alcohol-induced cell hydration is a critical event is furosemide’s reversal of alcohol-induced cytotoxicity in the organotypic rat brain slice cultures, particularly during the postincubation (‘alcohol withdrawn’) period. To our knowledge, extensive protection from alcohol-induced neurodegeneration has not been previously reported. In rats fed alcohol for a year, Paula-Barbosa et al. (44) considered the suppressive action of piracetam, a GABA nootropic, on the accumulation of aging pigment (lipofuscin) as neuroprotection, but reduced neuronal loss was not shown.The lack of furosemide neuroprotection in the olfactory bulb glomeruli in vivo requires further study. One possibility is that the olfactory nerves, confined by the narrow channels of the cribiform plate, may be exquisitely sensitive to even minimal alcohol-induced swelling, which was not arrested by the diuretic at the dose used (G. Shepherd, private communication).
ALCOHOL-INDUCED BRAIN DAMAGE
/ 3820 0011 Mp 227 Wednesday Jan 07 12:03 PM LP–FASEB 0011
227
Figure 6. Scheme showing proposed mechanical and metabolic pathways leading to the selective neurodegeneration that is initiated nonsynaptically by brain edema secondary to episodic alcohol intoxication and withdrawal. Brain seizure activity is anticipated, but (as discussed) may not be a precursor of the observed brain damage. ICP Å intracranial pressure.
Our initial interpretation was that the mechanism of alcohol-induced neurotoxicity may resemble that which is considered to function in status epilepticus, e.g., after glial swelling, a nonsynaptically mediated reduction in extracellular space that fosters neuronal hypersynchronization, excessive seizure activity, and neuronal death. Repetitive alcohol intoxication is known to promote kindling and augmented severity of brain seizures, possibly in a site-specific fashion (32, 45, 46). However, since seizure activity does have a prominent excitatory synaptic component (47), the negligible neuroprotection by glutamate receptor antagonists as well as the lack of degeneration in the hippocampal CA regions, which are highly susceptible to alcohol-induced kindling phenomena (32) and seizure damage (47, 48), argues against a causative role for brain seizures in the degenerative process initiated by subchronic episodic alcohol treatment. In place of a fundamental role for seizures, we speculate that the alcohol-dependent brain damage in the episodically exposed models is engendered through a metabolic route involving polyunsaturated fatty acid release, lipid peroxidation, and associated oxidative stress; a physical (mechanical) route of compression trauma and neuronal cytoskeletal collapse; or a combination of both. Further clarifying these potential routes, edema and associated cell stress can result in excessive arachidonic acid release from membrane phospholipid stores (49, 50). Arachidonic acid can stimulate lipid peroxidation and a deleterious free radical cascade through its oxidative metabolism (51); it also may cause a cycle of increased brain edema (52, 53). Lipid peroxidation and reactive free radicals have been implicated in alcoholic neurotoxicity (54), but specifically how they are promoted by alcohol has been unclear. The notion of a mechanically induced (parallel) degenerative process arises from clinical studies showing associations between edema-induced increases in intracranial pressure (ICP) and regional brain damage (55). During brain swelling and poten228
Vol. 12
February 1998
tially increased ICP due to episodic alcohol intoxication in the rat, the brain regions incurring substantial neurodegeneration are in fact those that are subject to physical compression against the prominent protrusion into the cranial cavity made by the tympanic bulla. Our suggestion is that recurrent (or sustained) brain edema and cellular compression provoked by alcohol intoxication and withdrawal could produce a mechanical collapse of neuronal ‘tensional integrity’ (56, 57). Figure 6 presents a testable scheme showing these proposed events that would link episodic alcohol-induced brain hydration to the early neurodegeneration. In the aggregate, the results using a diuretic and selected synaptic antagonists imply that nonsynaptic regulation of brain cell volume may be more important for the occurrence of early neuronal degeneration due to binge alcohol abuse than activation of synaptic excitatory receptors/Ca2/ channels. However, since the hippocampal CA and cerebellar regions have high densities of excitatory receptors, and more prolonged or chronic alcohol regimens do cause neuronal loss in those regions (2, 58–60), it is likely that synaptic glutamate receptors, perhaps in concert with GABA systems and Ca2/ channels, have important roles in long-term alcoholism. Indeed, accumulating in vitro evidence indicates that chronic alcohol exposure can sensitize neurons to excitotoxic insult (61, 62). Analogous to epilepsy (24), another implication from our results is that therapeutic agents that counter alcohol intoxication/withdrawal-induced brain swelling and edema, judiciously applied, could lessen brain neurodegeneration and resultant dementia. Parenthetically, furosemide as adjuvant therapy in alcoholism would not be novel; based on their previous findings (38), Knott and Beard (63) used the diuretic nearly three decades ago to improve physiological measures of recovery during alcohol withdrawal. In contrast, excitotoxic antagonists may not be helpful as neuroprotectants in the early stages of binge alcohol abuse. Gluco-
The FASEB Journal
/ 3820 0011 Mp 228 Wednesday Jan 07 12:03 PM LP–FASEB 0011
COLLINS ET AL.
corticoid intervention to counter edema would be problematic. Anti-inflammatory steroids can reduce induced brain swelling (64), but endogenous corticoid may have a role in brain damage during conditions of status epilepticus (65) and has been suggested as a factor in alcohol-induced brain damage (66).
16.
The support of the Loyola Neuroscience and Aging Institute (NSAI) and the Banes Fund is gratefully acknowledged. We appreciate the assistance of Dr. Earl Holmes (LUMC Clinical Labs) and Ralston Reid (Hines VA Hospital) with measurements of osmolalities and electrolytes, respectively. Criticisms of the manuscript by Dr. Mary Druse (Biochemistry) are also appreciated. A portion of this work was reported in abstract form: Soc. Neurosci. Abstr. 22, p. 471 (1996).
19.
17.
18.
20. 21. 22.
REFERENCES 1. 2. 3. 4. 5.
6. 7. 8. 9. 10.
11. 12.
13. 14.
15.
Hunt, W. A., and Nixon, S., eds (1993) Alcohol-Induced Brain Damage, NIH Publication No. 93–3549, National Institute of Health, Bethesda, Md. Lovinger, D. M. (1993) Excitotoxicity and alcohol-related brain damage. Alcohol Clin. Exp. Res. 17, 19–27 Tsai, G., Gastfriend, G. R., and Coyle, J. T. (1995) The glutamatergic basis of human alcoholism. Am. J. Psychiat. 152, 332– 340 Lancaster, F. E. (1992) Alcohol, nitric oxide and neurotoxicity: is there a connection? Metab. Brain Dis. 10, 125–133 Collins, M. A., Corso, T. D., and Neafsey, E. J. (1996) Neuronal degeneration in rat cerebrocortical and olfactory regions during subchronic ‘binge’ intoxication with ethanol: Possible explanation for olfactory deficits in alcoholics. Alcohol Clin. Exp. Res. 20, 284–292 Zou, J.-Y., Martinez, D. B., Neafsey, E. J., and Collins, M. A. (1996) Binge ethanol-induced brain damage: effect of inhibitors of nitric oxide synthase. Alcohol Clin. Exp. Res. 20, 1406–1411 de Olmos, J. S., Ebbesson, S., and Heimer, L. (1981) In Neuroanatomical Tracing Methods (Heimer, L., and Robards, M., eds) pp. 117–170, Plenum, New York Switzer, R. C., III (1993) Silver staining methods: their role in detecting neurotoxicity. Ann. N.Y. Acad. Sci. 679, 341–348 Beltramino, C. A., de Olmos, J. S., Gallyas, F., Heimer, L., and Zaborszky, L. (1993) Silver staining as a tool for neurotoxic assessment. NIDA Res. Monogr. 136, 101–126 Fix, A. S., Ross, J. F., Stitzel, S. R., and Switzer, R. C., III (1996) Integrated evaluation of CNS lesions: stains for neurons, astrocytes, and microglia reveal the spatial and temporal features of MK-801-induced neuronal necrosis in the rat cerebral cortex. Toxicol. Pathol. 24, 291–304 Switzer, R. C., III, Majchrowicz, E., and Weight, F. F. (1982) Ethanol-induced argyrophilia in entorhinal cortex of rat. Anat. Rec. 202, 186 Corso, T. D., Collins, M. A., and Neafsey, E. J. (1997) Brain neuronal degeneration caused by episodic alcohol intoxication in rats: effects of nimodipine, DNQX and MK-801. Alcohol Clin. Exp. Res. In press Stoppini, L., Buchs, P. A., and Muller, D. (1991) A simple method for organotypic cultures of nervous tissue. J. Neurosci. Methods 37, 173–182 Diekmann, S., Nitsch, R., and Ohm, T. G. (1994) The organotypic entorhinal-hippocampal complex slice culture of adolescent rats. A model to study transcellular changes in a circuit particularly vulnerable in neurodegenerative disorders. J. Neural Trans. 44 (Suppl.), 61–71 Bahr, B. A. (1995) Long-term hippocampal slices: a model system for investigating synaptic mechanisms and pathologic processes. J. Neurosci. Res. 42, 294–305
23. 24.
25. 26. 27.
28. 29.
30.
31. 32. 33. 34.
35.
36. 37. 38.
Vornov, J. J., Tasker, R. C., and Coyle, J. T. (1991) Direct observation of agonist-specific regional vulnerability to glutamate, NMDA and kainate neurotoxicity in organotypic hippocampal cultures. Exp. Neurol. 114, 11–22 Bruce, A. J., Malfroy, B., and Baudry, M. (1996) Beta-amyloid toxicity in organotypic hippocampal cultures: protection by EUK-8, a synthetic catalytic free radical scavenger. Proc. Natl. Acad. Sci. USA 93, 2312–2316 Sato, N., Wang, X. B., and Greer, M. A. (1991) Hormone secretion stimulated by ethanol-induced cell swelling in normal rat adenohypophysial cells. Am. J. Physiol. 260, E946–950 Kimelberg, H. K. Cheema, M., O’Connor, E. R., Tong, H., Goderie, S. K., and Rossman, P. A. (1993) Ethanol-induced aspartate and taurine release from primary astrocyte cultures. J. Neurochem. 60, 1682–1689 Snyder, A. K. (1996) Responses of glia to ethanol. In The Role of Glia in Neurotoxicity (Aschner, M., and Kimelberg, H. K., eds) pp. 111–135, CRC Press, Boca Raton, Fla. Andrew, R. D. (1991) Seizure and acute osmotic change: clinical and neurophysiological aspects. J. Neurol. Sci. 101, 7–18 Dudek F. E., Obenaus, A., and Tasker, J. G. (1990) Osmolalitydependent changes in extracellular volume alter epileptiform bursts independent of chemical synapses in the rat: importance of non-synaptic mechanisms in hippocampal epileptogenesis. Neurosci. Lett. 120, 267–270 Lambie, D. G. (1985) Alcoholic brain damage and neurological symptoms of alcohol withdrawal—manifestations of overhydration. Med. Hypoth. 16, 377–388 Hochman, D. W., Baraban, S. C., Owens, J. W., and Schwartzkroin, P. A. (1995) Dissociation of synchronization and excitability in furosemide blockade of epileptiform activity. Science 270, 99–102 Hunt, W. A. (1993) Are binge drinkers more at risk of developing brain damage? Alcohol 10, 559–561 Olson, J. E., Mishler, L., and Dimlich, R. (1990) Brain water content, brain blood volume, blood chemistry and pathology in a model of cerebral edema. Ann. Emerg. Med. 19, 1113–1121 Murphy, V. A. (1987) Method for determination of sodium, potassium, calcium, magnesium, chloride and phosphate in the rat choroid plexus by flame atomic absorption and visible spectroscopy. Anal. Biochem. 161, 144–151 Lowry, O. H., and Rosenbrough, N. J., Farr, A. L., and Randall, R. (1951) Protein measurement with Folin phenol reagent. J. Biol. Chem. 193, 265–275 Sparenborg, S., Brenneck, L. H., Jaax, N. K., and Braitman, D. J. (1992) Dizocilpine (MK-801) arrests status epilepticus and prevents brain damage induced by soman. Neuropharmacol. 31, 357–368 Clifford, D. B., Olney, J. W., Benz, A. M., Fuller, T. A., and Zorumski, C. F. (1990) Ketamine, phencyclidine, and MK-801 protect against kainic acid-induced seizure-related brain damage. Epilepsia 31, 382–390 Elliott, K. A. C., and Jasper, H. (1949) Measurement of experimentally induced brain swelling and shrinkage. Am. J. Physiol. 157, 122–129 Becker, H. C., and Littleton, J. M. (1996) The alcohol withdrawal ‘kindling’ phenomenon: clinical and experimental findings. Alcohol Clin. Exp. Res. 20 (Suppl.), 121A–124A Snyder, H., Williams, D., Zink, B., and Reilly, K. (1992) Accuracy of blood ethanol determination using serum osmolality. J. Emerg. Med. 10, 129–133 Heaton, M. B., Paiva, M., Swanson, D., and Walker, D. W. (1994) Responsiveness of cultured septal and hippocampal neurons to ethanol and neurotrophic substances. J. Neurosci. Res. 39, 305– 318 Pantazis, N. J., Dohrman, D., Goodlet, C. R., Cook, R. T., and West, J.R. (1993) Vulnerability of cerebellar granule cells to alcohol-induced cell death diminishes with time in culture. Alcohol Clin. Exp. Res. 17, 1014–1021 Nichols, J. M., and Martin, F. (1993) P300 in heavy social drinkers: the effect of lorazepam. Alcohol 10, 269–274 Gross, L. (1985) How much is too much? Effects of social drinking, Random House, New York. Beard, J. D., and Knott, D. H. (1968) Fluid and electrolyte balance during acute withdrawal in chronic alcoholic patients. J. Am. Med. Assn. 204, 135–137
ALCOHOL-INDUCED BRAIN DAMAGE
/ 3820 0011 Mp 229 Wednesday Jan 07 12:04 PM LP–FASEB 0011
229
39. 40. 41. 42. 43. 44.
45.
46. 47. 48. 49.
50. 51.
52.
53.
230
Mander, A. J., Young, A., Merrick, M. V., and Morton, J. J. (1989) Fluid balance, vasopressin and withdrawal symptoms during detoxication from alcohol. Drug Alcohol Depend. 24, 233–237 Trabert, W., Caspari, D., Bernhard, P., and Biro, G. (1992) Inappropriate vasopressin secretion in severe alcohol withdrawal. Acta Psychiatr. Scand. 85, 376–379 Roper, S. N., Obenaus, A., and Dudek, F. E. (1992) Osmolality and nonsynaptic epileptiform bursts in rat CA1 and dentate gyrus. Ann. Neurol. 31, 81–85 MacVicar, B. A., and Hochman, D. A. (1991) Imaging of synaptically evoked intrinsic optical signals in hippocampal slices. J. Neurosci. 11, 1458–1469 Lerma, J., and del Rio, R. M. (1992) Chloride transport blockers prevent N-methyl-D-aspartate receptor-channel complex activation. J. Pharmacol. Exp. Ther. 41, 217–222 Paula-Barbosa, M. M., Brandao, F., Pinho, M., Andrade, J., Madeira, M., and Cadete-Leite, A. (1991) The effects of piracetam on lipofuscin of the rat cerebellar and hippocampal neurons after long-term alcohol treatment and withdrawal: a quantitative study. Alcohol Clin. Exp. Res. 15, 834–838 McCown, T. J., and Breese, G. R. (1990) Multiple withdrawals form chronic ethanol ‘kindles’ inferior collicular seizure activity: evidence for kindling of seizures associated with alcoholism. Alcohol Clin. Exp. Res. 14, 394–399 Veatch, L. M., and Gonzalez, L. P. (1996) Repeated ethanol withdrawal produces site-dependent increases in EEG spiking. Alcohol Clin. Exp. Res. 20, 262–267 Meldrum, B. S. (1994) The role of glutamate in epilepsy and other CNS disorders. Neurol. 44 (Suppl. 8), S14–23 Wasterlain, C. G., Fujikawa, D. G., Penix, L., and Sankar, R. (1993) Pathophysiological mechanisms of brain damage from status epilepticus. Epilepsia 34, S37–S53 Bazan, N. G., and Turco, E. B. (1980) Membrane lipids in the pathogenesis of brain edema: phospholipids and arachidonic acid, the earliest membrane components changed at the outset of ischemia. Adv. Neurol. 28, 197–205 Unterberg, A., Polk, T., Ellis, E., and Marmarou, A. (1990) Enhancement of infusion-induced brain edema by mediator compounds. Adv. Neurol. 52, 355–358 Asano, T., Koide, T., Gotoh, O., Joshita, H., Hanamura, T., Shigeno, T., and Takakura, K. (1989) the role of free radicals and eicosanoids in the pathogenetic mechanism underlying ischemic brain edema. Mol. Chem. Neuropathol. 10, 101–133 Chan, P. H., Fishman, R. A., Caronna, J., Schmidley, J. W., Prioleau, G., and Lee, J. (1983) Induction of brain edema following intracerebral injection of arachidonic acid. Ann. Neurol. 13, 625– 632 Staub, F., Winkler, A., Peters, J., Kempski, O., Kachel, V., and Baethmann, A. (1994) Swelling, acidosis and irreversible damage
Vol. 12
February 1998
54.
55.
56.
57.
58. 59. 60. 61. 62.
63. 64. 65.
66.
of glial cells from exposure to arachidonic acid in vitro. J. Cerebr. Blood Flow Metab. 14, 1030–1039 Vallett, M., Tabatabaie, T., Briscoe, R. J., Baird, T., Beatty, W. W., Floyd, R. A., and Gauvin, D. V. (1997) Free radical production during ethanol intoxication, dependence and withdrawal. Alcohol Clin. Exp. Res. 21, 275–285 Adams, H., and Graham, D. I. (1972) The relationship between ventricular fluid pressure and the neuropathology of raised intracranial pressure. In Int. Symp. Intracranial Pressure, Experimental and Clinical Aspects (Brock, M., and Dietz, H., eds) pp. 250–253, Springer-Verlag, New York Gallyas, F., Zoltay, G., and Dames, W. (1992) Formation of ‘dark’ (argyrophilic) neurons of various origin proceeds with a common mechanism of biophysical nature (a novel hypothesis). Acta Neuropathol. 83, 504–509 Maniotis, A. J., Chen, C. S., and Ingber, D. E. (1997) Demonstration of mechanical connections between integrins, cytoskeletal filaments, and nucleoplasms that stabilize nuclear structure. Proc. Natl. Acad. Sci. USA 94, 849–854 Lundqvist, C., Alling, C., Knoth, R., and Volk, B. (1995) Intermittent ethanol exposure of adult rats: hippocampal cell loss after one month of treatment. Alcohol Alcohol. 30, 737–748 Walker, D. W., Barnes, J., Riley, J. N., Hunter, B. E., and Zornetzer, S. (1980) Neuronal loss in the hippocampus induced by prolonged ethanol consumption in rats. Science 209, 711–713 Tavares, M., and Paula-Barbosa, M. (1982) Alcohol-induced granule cell loss in the cerebellar cortex of the adult rat. Exp. Neurol. 78, 574–580 Chandler, L. J., Newsom, H., Sumners, C., and Crews, F. (1993) Chronic ethanol exposure potentiates NMDA excitotoxicity in cerebrocortical neurons. J. Neurochem. 60, 1578–1581 Ahern, K. V., Lustig, H. S., and Greenberg, D. A. (1994) Enhancement of NMDA toxicity and calcium responses by chronic exposure of cultured cortical neurons to ethanol. Neurosci. Lett. 165, 211–214 Knott, D. H., and Beard, J. D. (1968) A diuretic approach to acute withdrawal from alcohol. South. Med. J. 62, 485–489 Unterberg, A., Schmidt, W., Dautermann, C., and Baethmann, A. (1990) The effect of various steroid treatment regimens on cold-induced brain swelling. Acta Neurochir. 51 (Suppl.), 104–106 Smith-Swintosky, V. L., Pettigrew, L. C., Sapolsky, R. M., Phares, C., Craddock, S. D., Brooke, S. M., and Mattson, M. P. (1996) Metyrapone, an inhibitor of glucocorticoid production, reduces brain injury induced by focal and global ischemia and seizures. J. Cerebr. Blood Flow Metab. 16, 585–598 Menzano, E., and Carlen, P. L. (1994) Zinc deficiency and corticosteroids in the pathogenesis of alcoholic brain dysfunction— A review. Alcohol Clin. Exp. Res. 18, 895–901
The FASEB Journal
/ 3820 0011 Mp 230 Wednesday Jan 07 12:04 PM LP–FASEB 0011
Received for publication June 16, 1997. Accepted for publication October 20, 1997.
COLLINS ET AL.
