The effects of hypoglycemia and ethanol on rat performance in the radial-arm maze

Physiology & Behavior 75 (2002) 243 – 250

The effects of hypoglycemia and ethanol on rat performance in the radial-arm maze Perry M. Duncan*, Monica A. Gaffney Psychology Department, Old Dominion University, Norfolk, VA 23259-0267, USA Received 8 September 2001; received in revised form 26 September 2001; accepted 31 October 2001

Abstract A deficit in blood glucose (BG) (hypoglycemia, HG) causes a characteristic array of physiological sequelae, symptoms and cognitive impairment in human subjects. However, the performance of hypoglycemic animal subjects in the standard paradigms of behavioral pharmacology has received little research attention. The primary purpose of these experiments was to determine the effect of insulinproduced HG on spatial working memory in rats trained in the radial-arm maze (RAM), using a noninterrupted (win-shift) procedure. A second aim was to further investigate possible interaction between HG and ethanol administration since potentiation between the drugs’ depressant effects has been reported for rat spontaneous motor activity (SMA). Insulin administration (resulting in BG levels approximately 65% of control levels) was combined factorially with ethanol treatment in two experiments. HG significantly increased time required to complete RAM trials in both experiments, but did not impair accuracy of arm choice. In the second experiment, ethanol was administered only once to minimize development of tolerance, and under these conditions, ethanol at 1500 mg/kg impaired arm-choice accuracy and marginally potentiated HG-produced slowing of running time. The current results appear somewhat similar to previously reported effects of HG on reaction time in human subjects. D 2002 Elsevier Science Inc. All rights reserved. Keywords: Ethanol; Hypoglycemia; Insulin; Rats; Radial-arm maze

1. Introduction Hypoglycemia (HG), a condition of inadequate levels of blood glucose (BG), is not an unusual occurrence in persons suffering from insulin-dependent diabetes mellitus [8]. Glucose is the only energy source ordinarily utilized by neurons [1,42], and when plasma glucose levels fall below a critical value, progressive degrees of neuronal malfunction result from the decreased utilization of energy [5,6]. Extreme degrees of HG can result in convulsions, coma, and death [8], but at less severe levels of HG, a characteristic array of behavioral and subjective effects occur. Some aspects of this syndrome result from counter-regulatory processes, including sympathetic nervous system activation [10]. Various types of cognitive dysfunction also appear during HG as brain function becomes progressively impaired. This cognitive impairment has been studied extensively in both diabetic and nondiabetic human subjects [9]. Deficits * Corresponding author. Tel.: +1-757-683-4447; fax: +1-757-683-5087. E-mail address: [email protected] (P.M. Duncan).

have been reported in reaction time [11,18,27], vigilance [25], visual information processing [17], and performance on several other neuropsychological tests such as the digit – symbol substitution test and trail making [23]. Changes in EEG indices often accompany these cognitive and behavioral deficits [25,38], further revealing the alteration in normal brain function. The behavioral correlates of HG in animal subjects have received relatively little attention in the research literature. Reduction in spontaneous motor activity (SMA) of HG rats has been described [12,14], and HG produces an interoceptive stimulus that can be discriminated by rats in the operant drug-discrimination paradigm [13,15,39]. Pavlovian conditioning of glycemic responses to insulin injection has been investigated [19,40,45], as has the conditioned suppression of motor activity in response to stimuli predicting insulin injection and the subsequent onset of HG [14]. However, it appears that the standard paradigms of behavioral pharmacology have not been used previously to investigate animal analogues of the cognitive or behavioral dysfunctions related to HG as seen in humans.

0031-9384/02/$ – see front matter D 2002 Elsevier Science Inc. All rights reserved. PII: S 0 0 3 1 - 9 3 8 4 ( 0 1 ) 0 0 6 5 9 - X

244

P.M. Duncan, M.A. Gaffney / Physiology & Behavior 75 (2002) 243–250

The first purpose of the experiments reported here was to determine the effect of HG on rat performance in the radialarm maze (RAM). The RAM paradigm tests working memory and utilization of spatial cues in rodents [30], as well as locomotor activity as indicated by time required to complete a maze run. The effects of various drugs in the RAM have been studied extensively (see review, Ref. [30]), and the RAM is a useful procedure in behavioral toxicology [16,43]. However, there are no previous reports of effects of HG on RAM performance. The second purpose of the present study was to further investigate the combined effects of HG and ethanol. Clinical observations dating from the 1960s have indicated that combinations of HG and ethanol intoxication may cause extremely severe CNS depression which can result in permanent disability or death [2]. Both ethanol and HG produce CNS malfunction with resultant behavioral and cognitive impairment, and there is considerable clinical interest in the results of ethanol intoxication in combination with HG [26,31]. Interaction between the two conditions, especially potentiation of either depressant effect, would be of relevance to persons utilizing insulin therapy to achieve glycemic control. Kerr et al. [27] found that ethanol administration decreased both normal and diabetic human subjects’ ability to recognize the HG state (i.e., ethanol caused ‘HG unawareness’) even though some physiological responses associated with HG were potentiated by ethanol. Reaction time was slowed by both ethanol and HG, and the combination of the two experimental manipulations produced effects on reaction time which were similar to the separate treatments. Duncan [12] found that a moderate ethanol dose had a greater depressant effect on the SMA of HG rats in comparison to rats with normal BG levels (euglycemia). Ethanol at lower doses disinhibited HG-related SMA decreases, with the ethanol dose – effect depending on the level of HG and other experimental conditions. The present experiments were further investigations of ethanol effects on rats in the hypoglycemic state. Rats were trained to criterion performance in the RAM (noninterrupted, win-shift procedure), then administered ethanol over a range of doses in combination with insulin-produced HG. BG levels were monitored in order to determine whether any altered behavioral effect was due to exacerbation of HG, or to some other type of interaction between the two drug treatments.

2. Materials and methods, Experiment 1 2.1. Subjects and maintenance Twelve male rats of the Long– Evans strain, 100 days of age at the beginning of the experiment, were housed in individual cages in a room with light onset at 0800 h and darkness onset at 2000 h. Experimental procedures were

conducted between 1200 and 1400 h. Water was freely available, but food (Purina Rat Chow) was restricted as described below. Mean body weight prior to food deprivation was 350 g. Data from one animal which died during the experiment were not included in the analysis. Six rats of the same strain and gender, but previously used in a different experiment, served as subjects for the pilot study. The procedures used in this experiment were approved by the Institutional Animal Care and Use Committee of Old Dominion University, and are in compliance with guidelines provided by the Society for Neuroscience for ethical use of animal subjects. 2.2. Drugs Regular Iletin Insulin was diluted with physiological saline to a concentration of 2 units/ml. Ethanol doses (mg/kg) and concentrations (w/v) in 0.9% saline were as follows: 0; 250, 3%; 500, 6%; 1000, 12%; 1500, 12%. 2.3. Apparatus The eight-arm radial maze was constructed from plastic and wood, with arms measuring 79-cm long  10-cm high  14-cm wide extending from a central circular platform, diameter 27 cm. The maze was elevated 1 m above the floor and was located in a small room with stable visual cues (a door and overhead lights) plus a card with a different black geometric shape located above the end of each arm. Shallow cups for holding reinforcements were at the end of each arm. Reinforcements were 1/2 of a single piece of sweetened cereal (‘Tootie Fruitie,’ Malt O Meal). BG levels were measured with a Lifescan ‘One-Touch’ glucometer manufactured by Johnson and Johnson, and ‘First-Choice’ reagent strips were generously donated by Lifescan. Messier and Kent [32] have demonstrated that a similar BG measuring system, designed for self-monitoring of BG by diabetic persons, compares favorably in accuracy and reliability to a research-quality glucose analyzer manufactured by Beckman instruments. 2.4. Procedure 2.4.1. Pilot study A pilot study (n = 6) was conducted to establish procedures (including insulin dose, postinjection delays, and food-deprivation levels) suitable for an investigation of ethanol– HG interaction in the RAM paradigm. This pilot study also demonstrated that consumption of the reinforcements (as described above) did not detectably increase BG levels within the 5-min time limit allowed to complete a maze trial. 2.4.2. Training phase Daily food ration was restricted to 20 g/day, which eventually resulted in body weights 85% of predeprivation

P.M. Duncan, M.A. Gaffney / Physiology & Behavior 75 (2002) 243–250

values. Rats were handled daily for a week, and given three sweetened cereal reinforcements in their home cage daily during this period. A 3-day maze-adaptation period followed during which each rat was placed in the RAM for 10 min daily and allowed to consume several reinforcements placed along the maze arms and in the cups at the end of each arm. A 15-day training phase followed. On each of these days, each rat was placed in the central hub of the RAM and allowed to retrieve the reinforcements at the end of each maze arm (noninterrupted procedure). Behavior was recorded by an observer who, during the testing phase, was not aware of experimental treatment for each trial. Two observers were used on some tests, and comparison of their observations indicated a high degree of consistency in recording of both performance measures. Time to complete the maze (find and consume all eight reinforcements) was recorded, as were total errors and entries before the first error. An error was defined as a full entry (hind feet across the arm threshold) into a maze arm already visited on that day/trial. At the end of the training phase, all rats had fulfilled the following performance criteria during the final five trials: (1) completed the maze run in a maximum of 5 min, (2) made at least six correct arm entries prior to the first error on four of the five trials, and (3) made no more than two total errors per trial. During the final 3 days of the training phase, saline was injected intraperitoneally (1 ml/kg) 20 min prior to the maze trial. The tail was pricked with a stylet and a drop of blood drawn for a BG measurement immediately before the trial started.

245

jected once to all experimental conditions (combinations of insulin and ethanol doses). The sequence of treatments was counterbalanced to control for tolerance to ethanol, adaptation to experimental procedures, and possible Pavlovian conditioning of drug effects to the test apparatus. The counterbalancing control sequence of treatment combinations followed a modified Latin-square design (for ethanol doses), with insulin and saline treatments nested within each ethanol dose and the order of saline and insulin for each ethanol dose alternated among subjects. 2.4.5. Statistical analysis All data are presented as the mean ± S.E.M. Total time to complete a maze trial was divided by the number of arms entered during that trial to provide a time-per-arm (TPA) performance measure. Accuracy of maze performance was indicated primarily by number of arm entries prior to the first error (reentry into a previously entered arm). A secondary index of maze accuracy was the total number of such errors made during each maze trial. These results, plus the BG readings, were analyzed for significance by means of factorial analyses of variance (ANOVAs), repeated measures on both factors (insulin and ethanol dose). The significance level for rejection of the null hypothesis was P < .05.

3. Results, Experiment 1 3.1. Maze performance

2.4.3. Testing phase Procedures were as described for the final training days, but two intraperitoneal injections (one for insulin and one for ethanol as described below) were administered 20 min prior to the start of the maze trial. Trials were run at 48-h intervals. Twenty grams of food was given in the home cage immediately after each maze run, and the same amount given again 24 h later (24 h before the next maze trial). Eighteen hours before the next maze run, any food remaining in the home cage was removed. The pilot study indicated that rats were likely to be too debilitated to run in the maze at BG levels below 28 mg/dl. Therefore, rats with BG below this criterion level when blood was drawn 20 min postinjection were returned to the home cage, fed immediately, and not run in the maze on that day. This extreme degree of HG occurred on five occasions after various insulin – ethanol combinations (including ethanol dose = 0). Upon these occasions that experimental condition (drug combination) was repeated at the end of the experiment, after all other ethanol–insulin combinations had been tested. 2.4.4. Experimental design Each subject received all insulin –ethanol dose combinations in a counterbalanced sequence. A 2 (0, 2 units/kg insulin)  5 (0, 250, 500, 1000, 1500 mg/kg ethanol) withinsubject factorial design was employed. Each rat was sub-

In the HG state, time required to complete the maze run and subsequent TPA increased significantly, and to a similar extent in all ethanol conditions, including ethanol dose 0. As Fig. 1 illustrates, mean TPA (no ethanol) was

Fig. 1. Mean ( ± S.E.M.) seconds per arm for RAM trials under each experimental condition. Insulin dose = 2 units/kg, n = 11. Insulin main effect significant, P < .05.

246

P.M. Duncan, M.A. Gaffney / Physiology & Behavior 75 (2002) 243–250

3.2. Blood glucose Fig. 3 illustrates that insulin decreased BG levels at all ethanol doses, including ethanol dose 0. Mean BG level in the insulin and no-ethanol condition was 65% of the saline and no-ethanol condition, and similar degrees of HG resulted after insulin injection in the other experimental conditions. ANOVA revealed a significant main effect of insulin treatment [ F(1,90) = 291.21, P < .001]. The main effect of ethanol was not significant [ F(4,90) = 1.90, P >.05] and the Insulin  Ethanol interaction effect was also not significant [ F(4,90) = 1.13, P >.05].

4. Experiment 2 Fig. 2. Mean ( ± S.E.M.) arm entries before first error under each experimental condition. Insulin dose = 2 units/kg, n = 11.

approximately 26 s in the HG condition, compared to 13 s in the euglycemic condition. Although some ataxia and slightly increased mean TPA were evident at the two highest doses, ethanol had no significant effect on this dependent variable. ANOVA results revealed a significant main effect from insulin treatment [ F(1,90) = 46.7, P < .05], but no significant main effect from ethanol [ F(4,90) = 0.86, P > .05], and none from Ethanol  Insulin interaction [ F(4,90) = 1.14, P >.05]. Neither the HG state nor ethanol at any dose had significant effects on maze performance as indicated by number of correct entries prior to the first error (Fig. 2). ANOVA of this measure revealed no significant main or interaction effects. For the insulin main effect [ F(1,90) = 3.65], ethanol main effect [ F(4,90) = 1.23], interaction effect [ F(4,90) = 2.03], P >.05 for all F values. Mean total errors per trial are presented in Table 1, which shows that although few errors occurred in any treatment condition, performance accuracy as indicated by this measure was slightly better in the HG condition at four of the five ethanol doses. The significant main HG effect [ F(1,90) = 8.83, P < .05], indicated the statistical reliability of this small effect. The main ethanol effect for this variable was not significant [ F(4,90) = 1.66, P >.05], nor was the interaction effect [ F(4,90) = 1.79, P >.05].

Ethanol had no effect on either maze running time or accuracy, and HG did not degrade maze accuracy in Experiment 1. However, several ethanol behavioral effects can be attenuated by tolerance [4,20,24], and there is some evidence of adaptation to certain effects of HG [7] after repeated HG episodes. Since the previous experiment involved repeated administration of both drugs, a second experiment was designed to preclude tolerance effects for ethanol and to minimize such effects of HG. Insulin was administered on two test days to one of two subject groups, and ethanol (1500 mg/kg) was administered only once to both groups. With exceptions described below, subjects, procedures, and materials were very similar to those of the first experiment. 4.1. Materials and methods 4.1.1. Subjects and maintenance Subjects were 20 naive male rats, identical in strain and age to those used in Experiment 1. Maintenance procedures were also as described above.

Table 1 Total errors per trial (mean ± S.E.M.), n = 11 Experimental condition Ethanol dose (mg/kg)

Saline

Insulin, 2 units/kg

0 250 500 1000 1500

1.55 0.91 1.00 2.27 1.36

0.40 0.91 0.82 0.67 0.55

( ± 0.62) ( ± 0.37) ( ± 0.38) ( ± 0.96) ( ± 0.96)

( ± 0.12) ( ± 0.46) ( ± 0.44) ( ± 0.20) ( ± 0.16)

Fig. 3. Mean ( ± S.E.M.) BG levels 20 min after drug or saline injection. Insulin dose = 2 units/kg, n = 11. Insulin main effect significant, P < .001.

P.M. Duncan, M.A. Gaffney / Physiology & Behavior 75 (2002) 243–250

247

4.1.2. Drugs, equipment, and procedure Humulin insulin and ethanol were diluted with saline as described above. Equipment, training, and test procedures were as described for Experiment 1. 4.1.3. Experimental design The 20 rats were trained to criterion performance in the RAM and then divided into two groups of n = 10, matched for time required to complete RAM trials. The HG group was given insulin treatment (2 units/kg) on each of the two test days and the saline group was given saline injections on both test days. Five rats of each group were given intraperitoneal injections of 1500 mg/kg ethanol on the first test day, with the other five rats receiving saline injections. The saline – ethanol treatments were reversed on the second test day. Thus, ethanol was a within-subjects treatment manipulation and insulin (HG) was a betweensubjects manipulation. The subgroups receiving different ethanol or saline treatments on a given test day were also matched for RAM completion time in the nondrugged state. Forty-eight hours elapsed between the first and second RAM test days. 4.1.4. Statistical analysis A mixed ANOVA was used to determine main and interaction effects on maze accuracy (errors before first reentry, and total errors per trial), TPA, and BG levels.

5. Results, Experiment 2 5.1. Maze performance Fig. 4 illustrates that in the HG state, mean TPA was increased to more than twice that of the euglycemic state,

Fig. 4. Mean ( ± S.E.M.) seconds per arm for RAM trials under each experimental condition. Insulin dose = 2 units/kg, n = 10 for insulin, and for no-insulin groups. Insulin main effect significant, P < .001.

Fig. 5. Mean ( ± S.E.M.) arm entries before first error under each experimental condition. Insulin dose = 2 units/kg, n = 10 for insulin, and for noinsulin groups. Ethanol main effect significant, P < .002.

and this HG-produced slowing of RAM completion time occurred both with and without the concurrent effects of ethanol. ANOVA indicated that the main effect of HG was significant [ F(1,18) = 21.47, P < .001], but the main ethanol effect was not significant [ F(1,18) = 3.91, P >.05]. Although Fig. 4 shows that ethanol produced a marked difference in the mean rate of the HG group (and an ethanol effect was much less apparent in the saline group), the HG  Ethanol interaction was only marginally significant [ F(1,18) = 3.23, P < .10]. Fig. 5 shows that ethanol treatment produced a significant degradation in maze accuracy as indicated by number of arm entries prior to the first erroneous reentry, but HG did not decrease maze accuracy. Main effect for ethanol was [ F(1,18) = 13.48, P < .002], for HG [ F(1,18) = 1.37, P >.05]. Potentiation between the two effects did not occur; interaction effect [ F(1,18) = 1.49, P >.05]. Analysis of total error scores (not reported here) yielded differences with a pattern quite similar to those of the primary indicator of performance accuracy. However, neither main nor interaction effects were significant for the total error measure. Means for all performance indicators were similar between the subgroups which received HG (and no ethanol) on Day 1 vs. Day 2. 5.1.1. Blood glucose Mixed ANOVA of BG levels indicated a highly significant main effect of insulin [ F(1,18) = 47.23, P < .0001], no significant main effect of ethanol [ F(1,18) = 2.31, P >.05], and no significant interaction effect [ F(1,18) = 1.12, P >.05]. In the no-ethanol condition, the mean BG level of the saline group was 56.7 mg/dl, and for the insulin group this value was 36.3 (64% of the control value). Ethanol-condition BG levels for the two groups were quite similar to their no-ethanol values.

248

P.M. Duncan, M.A. Gaffney / Physiology & Behavior 75 (2002) 243–250

6. Discussion In both experiments, insulin-produced HG caused a marked increase in time to complete maze trials, while not impairing accuracy of arm selection. The increased maze time, which was adjusted for any differences in number of arms entered, appeared to result from slowed locomotor activity during both parts of maze trials, i.e., involving decreased speed of running down the radial arms, and greater latency to select the subsequent arm to be entered. This slowing of running speed in the RAM may be a behavioral expression of a generalized decrease in vigor of motor activity, which apparently caused the similar reduction in levels of spontaneous locomotor activity seen in HG rats [12,14]. Meaningful comparison between human and rat performance in different paradigms presents obvious difficulties. However, one of the most reliable behavioral effects of HG in humans is increased reaction time, thought to be due to slowed information processing [9,11], and a similar effect could well be partly responsible for the increase in time to complete maze trials seen here. The absence of impairment of choice accuracy (and some evidence of improved accuracy) resulting from HG in the RAM paradigm is somewhat surprising, given the definite HG-related slowing of maze running and the variety of cognitive impairments associated with HG in human subjects [9]. At least two factors could be involved in these rats’ ability to maintain maze accuracy in spite of the decreased glucose availability. First, cognitive impairment in humans at lesser degrees of HG is seen most clearly for complex tasks which have not been practiced extensively and which require sustained attention [9,11]. These rats were trained to a high criterion of maze accuracy before testing, and the noninterrupted RAM paradigm might be considered a relatively simple task for natural foraging animals such as rodents [35]. Second, adaptation to HG impairment of reaction time [28] and other types of HG effects [7] has been reported for human subjects. Rats have been found to increase transport of glucose into the brain under conditions of HG [37], which might prevent or attenuate some aspects of cognitive impairment. In that earlier experiment, the adaptation occurred with sustained HG, in contrast to the brief periods of HG in the current study. The repeated exposure to the HG state, and especially maze running in the state of glucose deprivation, might have facilitated such adaptation in the first experiment of the current study. However, similar HGrelated effects on rate, and absence of choice accuracy impairment also occurred in Experiment 2 in which insulin was administered only twice. Additionally, performance of subgroups receiving insulin on the first or second day in Experiment 2 was similar, further suggesting that adaptation to HG did not have much effect here. Driesen et al. [11] found that under some HG conditions choice accuracy of human subjects was maintained even though reaction time

was significantly increased. Several other investigators have also described a similar pattern of HG-related impairment (cf. Ref. [9]). It should be emphasized that the rats in the present study were working at the lower limit of BG level required to sustain locomotion in the RAM. This basic performance limit of HG was demonstrated by the pilot study and was confirmed by debilitation of animals with BG levels below 28 mg/dl. Ethanol did not degrade either aspect of RAM performance (running time or choice accuracy) in Experiment 1. In the second experiment, maze accuracy was significantly impaired by a single ethanol treatment (1500 mg/kg), suggesting that the repeated exposure to ethanol in Experiment 1 produced tolerance (including behavioral tolerance) to this effect on maze performance. Ethanol has been reported to impair spatial working memory as indicated by RAM performance at the moderate doses used here [21,22,44], although variations of the basic RAM paradigm make comparisons among experiments somewhat difficult. During the two highest-dose conditions rats were moderately ataxic, although maintaining running speed in both experiments. In Experiment 1, ethanol did not potentiate the increased running time caused by HG. However, Fig. 5 illustrates an apparent tendency for potentiation in the TPA measure of RAM completion time in the second experiment. Although the interaction effect failed to reach statistical significance, the slower mean running speed of the HG group was further decreased by ethanol treatment, whereas ethanol had no detectable effect on the euglycemic group. Development of tolerance to ethanol in the first experiment is the most likely explanation for the tendency toward potentiation of running speed impairment in the second, but not the first experiment. The degree of HG was very similar for the two experiments, thus eliminating a difference in BG levels as being responsible for the somewhat different expression of HG – ethanol potentiation. A somewhat surprising finding was the apparent improvement of maze accuracy by HG. This effect was significant as indicated by the total error measurement in the first experiment, and can be seen as a nonsignificant tendency in the first-error index of the second experiment (Fig. 5). This small but unexpected effect occurred mainly in the HG –ethanol combination conditions of both experiments, and a possible explanation would involve the slowed running of HG rats which may have counteracted some of the disinhibiting effects of ethanol and thereby decreased incorrect arm choice. Insulin can improve cognitive performance in some behavioral paradigms and under certain conditions [36]. The mechanism underlying such facilitation has not been identified, but would seem to be an insulin effect not closely related to its hypoglycemic action. The failure to clearly demonstrate a significant degree of potentiation between HG and ethanol effects differs from the results of the previous study of the combined drug effects on rat behavior [12]. In that investigation, potentia-

P.M. Duncan, M.A. Gaffney / Physiology & Behavior 75 (2002) 243–250

tion of SMA decreases occurred when HG rats were administered ethanol at 1200 mg/kg. Insulin dose (and degree of HG) and ethanol doses used in the SMA experiment were identical or similar to those of the current study. The most obvious differences, which may account for the discrepancy, are the high levels of motivation and training of the rats in the RAM paradigm, as opposed to the SMA experiment which did not involve trained, food-motivated behavior. Another difference is the duration of the behavioral tests, which was 30 min in the SMA study in comparison with 5 min in the current RAM experiment. The study with human subjects in which ethanol did not potentiate the HG-produced slowing of reaction time [27] is another example of a lack of HG – ethanol interaction. Apparently, HG – ethanol interaction may be demonstrated convincingly in only some types of behaviors and occurs readily only under certain experimental conditions. Ethanol caused no significant change in BG levels, a result consistent with previous reports of BG levels in rats after acute ethanol administration [33,34,41]. In these previous studies, ethanol did produce HG in rats, but only when greater degrees of food deprivation (48 h) were combined with higher ethanol doses (3 g/kg) than were employed here. Certain combinations of food deprivation and ethanol dose can influence the depth and time-course of HG, particularly in diabetic persons [3,31], but such potentiation was not seen in the present experiment. The absence of such an interaction is not inconsistent with previous studies of acute ethanol effects on insulin-produced HG (e.g., Ref. [29]). In summary, insulin treatment caused marked reduction in BG levels and increased time to complete maze trials in two experiments using a noninterrupted RAM procedure. HG did not impair accuracy of correct arm choice, but in one experiment slightly increased RAM accuracy as indicated by one performance indicator. Ethanol at these low to moderate doses did not slow running time, but when administered acutely to preclude tolerance development tended to potentiate the slowing effect of HG on maze completion time. Ethanol degraded choice accuracy at a moderate dose only when administered acutely and not after chronic administration. Ethanol did not cause HG, nor did it potentiate insulin’s reduction of BG levels. The relatively weak ethanol– HG interaction may limit the generality of potentiation of behavioral effects resulting from combination of the two experimental treatments as previously seen in rat SMA. References

[4]

[5]

[6]

[7]

[8]

[9]

[10]

[11]

[12] [13]

[14]

[15] [16]

[17]

[18]

[19]

[20]

[21] [22]

[1] Amiel SA. Organ fuel selection: brain. Proc Nutr Soc 1995;54:151 – 5. [2] Arky R, Veverbrants E, Abramson EA. Irreversible hypoglycemia: a complication of alcohol and insulin. J Am Med Assoc 1968;206: 575 – 8. [3] Avogaro A, Beltramello P, Gnudi L, Maran A, Valerio A, Miola M, Marin N, Crepaldi C, Confortin L, Costa F, MacDonald I, Tiengo A.

[23]

249

Alcohol intake impairs glucose counterregulation during acute insulininduced hypoglycemia in IDDM patients. Diabetes 1993;42:1626 – 34. Bell R, McKinzie D, Murphy J, McBride W. Sensitivity and tolerance to the motor impairing effects of moderate doses of ethanol. Pharmacol, Biochem Behav 2000;67:583 – 6. Bendtson I. Neurophysiological changes of hypoglycaemia. In: Frier BM, Fisher BM, editors. Hypoglycemia and diabetes: clinical and physiological aspects. London: Edward Arnold, 1993. pp. 72 – 9. Benzi G, Agnoli A. Influence of aging on cerebral carbohydrate, amino acid and lipid metabolism during severe hypoglycemia recovery. In: Andreani D, Marks V, Lefebvre P, editors. Hypoglycemia. New York: Raven Press, 1987. pp. 149 – 68. Boyle PJ, Nagy EJ, O’Connor AM, Kempers SF, Yeo RA, Qualls C. Adaptation in brain glucose uptake following recurrent hypoglycemia. Proc Natl Acad Sci USA 1994;91:9352 – 6. Cryer PE, Binder C, Bolli GB, Cherrington AD, Gale EAM, Gerich JE, Sherwin RS. Hypoglycemia in IDDM. Diabetes 1989;38: 1193 – 9. Deary IJ. Effects of hypoglycemia on cognitive function. In: Frier BM, Fisher BM, editors. Hypoglycemia and diabetes: clinical and physiological aspects. London: Edward Arnold, 1993. pp. 80 – 91. Deary LJ, Hepburn DA, MacLeod KM, Frier BM. Partitioning the symptoms of hypoglycemia using multi-sample confirmatory factor analysis. Diabetologia 1993;36:771 – 7. Driesen NR, Cox DJ, Gonder-Frederick L, Clarke W. Reaction time impairment in insulin-dependent diabetes: task complexity, blood glucose levels, and individual differences. Neuropsychology 1995;9: 246 – 54. Duncan P. The effect of ethanol on spontaneous motor activity in hypoglycemic rats. Pharmacol, Biochem Behav 2001;69:291 – 8. Duncan PM, Hooker EH. The role of sympathetic arousal in discrimination of insulin-produced hypoglycemia. Behav Pharmacol 1997;8: 389 – 95. Duncan PM, Koontz K. Pavlovian conditioning of blood glucose and motor activity responses. Paper presented at the Fifth International Meeting of the European Behavioral Pharmacology Society, Berlin, 1994. Duncan PM, Lichty W. Discrimination of insulin-produced hypoglycemia in rats. Physiol Behav 1993;54:1099 – 102. Eckerman DA, Gordon WA, Edwards JD, MacPhail RC, Gage MI. Effects of scopolamine, pentobarbital, and amphetamine on radial arm maze performance in the rat. Pharmacol, Biochem Behav 1980;12: 595 – 602. Ewing FME, Deary IJ, McCrimmon RJ, Strachan MWJ, Frier BM. Effect of acute hypoglycemia on visual information processing in adults with type 1 diabetes mellitus. Physiol Behav 1998;64:653 – 60. Fanelli C, Pampanelli S, Epifano LE, Rambotti AM, Ciofetta M, Modarelli F, DiVincenzo A, Annibale B, Lepore M, Lalli C, DelSindaco P, Brunetti P, Bolli GB. Relative roles of insulin and hypoglycemia on induction of neuroendocrine responses to, symptoms of, and deterioration of cognitive function in hypoglycemia in male and female humans. Diabetologia 1994;37:797 – 807. Flaherty CF, Grigson PS, Brady AB. Relative novelty of conditioning context influences directionality of glycemic conditioning. J Exp Psychol (Anim Behav Proc) 1987;13:144 – 9. Gauvin DV, Baird TJ, Briscoe RJ. Differential development of behavioral tolerance and the subsequent hedonic effects of alcohol in AA and NA rats. Psychopharmacology 2000;151:335 – 43. Gibson WS. Effects of alcohol on radial maze performance in rats. Physiol Behav 1985;35:1003 – 5. Givens BS. Low doses of ethanol impair spatial working memory and reduce hippocampal theta. Alcohol: Clin Exp Res 1995;19: 763 – 7. Gold AE, Deary IJ, MacLeod KM, Thomson KJ, Frier BM. Cognitive function during insulin-induced hypoglycemia in humans: short-term cerebral adaptation does not occur. Psychopharmacology 1995;119: 325 – 33.

250

P.M. Duncan, M.A. Gaffney / Physiology & Behavior 75 (2002) 243–250

[24] Holloway FA, Bird DC, Holloway JC, Michaelis RC. Behavioral factors in development of tolerance to ethanol’s effects. Pharmacol, Biochem Behav 1988;29:105 – 13. [25] Howorka K, Heger G, Schabmann A, Anderer P, Tribl G, Zeitlhofer J. Severe hypoglycemia unawareness is associated with an early decrease in vigilance during hypoglycemia. Psychoneuroendocrinology 1996;21:295 – 312. [26] Kerr D. Drugs and alcohol. In: Frier BM, Fisher BM, editors. Hypoglycemia and diabetes: clinical and physiological aspects. London: Edward Arnold, 1993. pp. 326 – 8. [27] Kerr D, Macdonald IA, Heller SR, Tattersall RB. Alcohol causes hypoglycemic unawareness in healthy volunteers and patients with type 1 (insulin-dependent) diabetes. Diabetologia 1990;33:216 – 21. [28] Kerr D, MacDonald IA, Tattersall RB. Adaptation to mild hypoglycemia in normal subjects despite sustained increases in counter-regulatory hormones. Diabetologia 1989;32:249 – 54. [29] Kolaczynski JW, Ylikahri R, Harkonen M, Koivisto VA. The acute effect of ethanol on counterregulatory response and recovery from insulin-induced hypoglycemia. J Clin Endocrinol Metab 1988;67: 384 – 8. [30] Levin ED. Psychopharmacological effects in the radial-arm maze. Neurosci Biobehav Rev 1988;12:169 – 75. [31] Meeking D, Cavan D. Alcohol ingestion and glycemic control in patients with insulin-dependent diabetes mellitus. Diabetic Med 1997;14:279 – 83. [32] Messier C, Kent P. Repeated blood glucose measures using a novel portable glucose meter. Physiol Behav 1995;57:807 – 11. [33] Oliveira de Souza ML, Masur J. Blood glucose and body temperature alterations induced by ethanol in rats submitted to different levels of food deprivation. Pharmacol, Biochem Behav 1981;15:551 – 4.

[34] Oliveira de Souza ML, Masura J. Does hypothermia play a relevant role in the glycemic alterations induced by ethanol? Pharmacol, Biochem Behav 1982;16:903 – 8. [35] Olton DS. The radial-arm maze as a tool in behavioral pharmacology. Physiol Behav 1987;40:793 – 7. [36] Park CR. Cognitive effects of insulin in the central nervous system. Neurosci Biobehav Rev 2001;25:311 – 23. [37] Pelligrino DA, Segil LJ, Albrecht RF. Brain glucose utilization and transport and cortical function in chronic vs. acute hypoglycemia. Am J Physiol 1990;259:E729 – 35. [38] Pramming S, Thorsteinsson B, Stigsby B, Binder C. Glycemic threshold for changes in electroencephalograms during hypoglycemia in patients with insulin dependent diabetes. Br Med J 1988;296:665 – 7. [39] Schuh KJ, Schaal DW, Thompson T, Cleary JP, Billington CJ, Levine AS. Insulin, 2-deoxy-D-glucose, and food deprivation as discriminative stimuli in rats. Pharmacol, Biochem Behav 1994;47:317 – 24. [40] Siegel S. Conditioning insulin effects. J Comp Physiol Psychol 1975; 89:189 – 99. [41] Tramill JL, Turner PE, Harwell G, Davis SF. Alcoholic hypoglycemia as a result of acute challenges of ethanol. Physiol Psychol 1981;9: 114 – 6. [42] Wahren J, Ekerg K, Fernqvist-Forbes E, Nair S. Brain substrate utilization during acute hypoglycemia. Diabetologia 1999;42:812 – 8. [43] Walsh T, Chrobak J. The use of the radial-arm maze in neurotoxicology. Physiol Behav 1987;40:799 – 803. [44] White AM, Simson PE, Best PJ. Comparison between the effects of ethanol and diazepam on spatial working memory in the rat. Psychopharmacology 1997;133:256 – 61. [45] Woods SC. Conditioned hypoglycemia. J Comp Physiol Psychol 1976;90:1164 – 8.