Psychopharmacology (1992) 108:417-431
Psychopharmacology © Springer-Verlag 1992
Original investigations Nicotinic systems and cognitive function Edward D. Levin Nicotine Research Laboratory, Department of Psychiatry, Duke University Medical Center, Box 3557, Durham, NC 27710, USA ReceivedApril 21, 1992 / Final version May 13, 1992
Abstract. Nicotinic acetylcholine receptors have been found to be important for maintaining optimal performance on a variety of cognitive tasks. In humans, nicotine-induced improvement of rapid information processing is particularly well documented. In experimental animals nicotine has been found to improve learning and memory on a variety of tasks, while the nicotinic antagonist mecamylamine has been found to impair memory performance. Nicotine has been found to be effective in attenuating memory deficits resulting from lesions of the septohippocampal pathway or aging in experimental animals. Nicotinic receptors are decreased in the cortex of patients with Alzheimer's disease. Preliminary studies have found that some aspects of the cognitive deficit in Alzheimer's disease can be attenuated by nicotine. Nicotine may prove to be useful therapeutic treatment for this and other types of dementia.
on tasks assessing attention, learning and memory. However, it is not clear which aspects of cognition actually underlie these improvements. In addition, there are studies which have failed to find improvements and others which have found nicotine-induced performance deficits. These results point to the limits of the beneficial effects of nicotine. They can also be used to help define the aspects of cognition which are affected by nicotine. Much of the impetus for this research is the finding that nicotinic receptors are substantially decreased in Alzheimer's disease. Investigation of nicotinic systems is important for improving the basic understanding of the neural substrates of cognitive function. It can also lead to new avenues of treatment for cognitive disorders such as Alzheimer's disease. Nicotinic agonist effects
Key words: Acetylcholine - Nicotinic - Memory - Nicotine Mecamylamine
Acetytcholine (ACh) systems have long been known to be important for accurate performance on cognitive tasks (see Bartus et al. 1987). The great majority of research concerning cholinergic mechanisms has focused on the muscarinic ACh receptor subtype. One of the most replicated findings in behavioral pharmacology is that the muscarinic blocker scopolamine impairs performance on memory tasks (see Levin 1988a). In contrast, the nicotinic cholinergic receptor subtype has received much less attention. Over the past 3 decades, there has been only a scattering of studies concerning nicotinic involvement in cognitive function. They have demonstrated that as has been shown with muscarinic antagonists, nicotinic antagonists effectively impair performance on memory tasks. In the past several years there has been a blossoming of research concerning the importance of nicotinic systems in cognitive function. Nicotine has been shown in many studies to improve performance
Acute effects While relatively less attention has been paid to nicotinic cholinergic systems compared to muscarinic systems, there is convincing evidence that they are also involved in cognitive processing. A variety of studies have reported acute, nicotine-induced facilitation of cognitive function in humans (Andersson and Post 1974; Houston et al. 1978; Wesnes and Warburton 1983; Wesnes et al. 1983; Peeke and Peeke 1984; Warburton etal. 1986; West and Hack 1991). As reviewed below, many studies have also seen beneficial effects of acute nicotine administration on cognitive performance in experimental animals. However, some investigators have not seen beneficial effects.
Nicotine effects in humans. Nicotine has been tbund to have a variety of cognitive enhancing properties in humans including improved attention and memory functioning (Warburton 1992). Nicotine effects in humans are clouded by the fact that a great majority of the research has been conducted on cigarette smokers. There
418 are several reasons why this presents a problem. First, the population of smokers is self-selected. Pre-existing differences between smokers and nonsmokers may have caused them to begin and continue to smoke. Test-related differences between smokers and nonsmokers may be related to nicotine exposure or to these pre-existing differences. Second, acute effects of nicotine in smokers are quite likely to be altered because of past chronic exposure. Third, since most smokers self-administer nicotine in a chronic fashion, it is difficult to have a proper control condition. Regular smokers may be either under the influence of nicotine or in a state of nicotine withdrawal. Some of the most valuable studies in this area have assessed the effects of nicotine in people who only occasionally smoke or nonsmokers. The most consistent finding of nicotine and cigarette smoking effects on human cognition is an improvement in vigilance and rapid information processing (Wesnes and Warburton 1983; Warburton and Wesnes 1984; Warburton 1992). A variety of studies have shown that smoking improves vigilance performance by attenuating the decline in performance which normally occurs over time (Wesnes and Warburton 1983). However, in some cases it is difficult to determine how much the improved performance after smoking is due to relief of the deficits shown during nicotine withdrawal. A study by Wesnes and Warburton (Wesnes and Warburton 1978) compared vigilance performance of deprived smokers and non-deprived smokers with non-smoking subjects. Both the deprived and non-smoking groups showed declines in performance over time, whereas the smoking group maintained good performance. Smoking-induced improvement in a complex signal detection task was found to be due to both an increase in the number of correct detections and response speed (Wesnes and Warburton 1978). There was no evidence that improvement in accuracy was due to a decrease in response speed or visa versa. Nicotine administration reduces color naming time on the Stroop test where the presence of an irrelevant cue (a color name) slows the naming of a color (Wesnes and Warburton 1978, 1983; Provost and Woodward 1991). This may reflect an effect of nicotine improving selective attention as suggested by Warburton and Wesnes 1984. However, the effect of nicotine is not apparent at the beginning of testing. Provost and Woodward (1991) pointed out that this effect is seen only during repeated testing. The nicotine-induced improvement emerges as a faster improvement in responding. There seems to be some learning or adaptation which takes place such that subjects more quickly process the information or more efficiently screen out irrelevant cues. Nicotine potentiates this process. The converse effect can be seen during nicotine withdrawal. One of the symptoms of nicotine withdrawal is decreased attentiveness. Hatsukami and wo-workers (Hatsukami et al. 1989) found increased errors of commission on a vigilance task in smokers after 24 h of smoking deprivation. Interestingly, no decrease in vigilance was seen with shorter periods of withdrawal. The effect of nicotine on memory function in humans
is less clear. Some investigators have found deficits, no facilitation or variable effects after nicotine administration (Andersson and Hockey 1977; Williams 1980; Peters and McGee 1982; Dunne et al. 1986). Other studies have documented nicotine-induced improvements in memory function (Warburton 1992). Cigarette smoking has been found to improve short-term verbal recall (Peeke and Peeke 1984). Nicotine-induced improvement of short-term memory was also seen in a study by Warburton and co-workers (Warburton et al. 1986). Memory search rate on Sternberg's memory search task is significantly faster in deprived smokers after smoking a nicotine-containing cigarette versus a placebo cigarette (West and Hack t991). This effect was seen both in regular and occasional smokers suggesting that the improvement was not merely due to amelioration of withdrawal effects in regular smokers. Roth and co-workers found (Roth et al. 1992) that cognitive effects of cigarette deprivation and smoking are different in different types of smokers. They studied the effects of smoking a cigarette in the early morning on memory performance on the Austin maze and a word recognition memory test. Subjects who typically smoked their first cigarette within the first hour after getting up performed the maze task and word recognition task significantly better after smoking than after deprivation. The reverse was true for subjects who did not usually smoke their first cigarette in the early morning.
Nicotine effects in experimental animals. Many of the problems associated with assessing acute nicotine effects in humans can be controlled for in animal models. Subjects can be randomly assigned to treatment conditions so that pre-existing individual differences would not affect group composition. Acute administration can be studied separately from chronic administration. Undrugged control tests can be separately assessed from withdrawal states. A variety of animal studies have shown that nicotine improves learning and memory performance (B/ittig 1970; Bovet-Nitti 1966; Buccafusco and Jackson 1991 ; Clarke 1987; Elrod et al. 1988; Erickson 1971; Evangelista and Izquirierdo 1972; Garg and Holland 1969; Haroutunian etal. 1985; Jackson and Buccafusco 1989; Jackson et al. 1989; Levin et al. 1990a, 1992a, b; Levin and Rose 1990, 1991; Nelsen 1978; Nordberg and Bergh 1985; Oliverio 1966, 1968; Orsingher and Fulginiti 1971 ; Pradhan 1970; Sansone et al. 1991; Sasaki et al. t991). However, some studies have not detected effects or have documented negative actions (Essman and Essman 1971 ; Hunter et al. 1977; Dunnett and Martel 1990; Bammer 1982; Mundy and Iwamoto 1988 a, b; Welzl et al. 1988). Several factors, including differences in dose, strain, task structure or level of training, may explain differences in the effects of nicotine on cognitive performance. Several older studies have shown that acute injections of nicotine significantly improved learning. Garg and Holland (1969) showed a nicotine-induced performance improvement in the Hebb-Williams closed field test and Oliverio (1968) found that post-trial nicotine caused an enhancement of learning transfer. Nicotine administra-
419 tion improved performance of a shuttle box conditioned response in rats (Evangelista et al. 1970). Curiously, hexamethonium also improved performance. More recent studies have found that acute pre-trial nicotine found that acute nicotine administration (0.125 mg/kg) improves passive avoidance learning (Sansone et al. 1991 ; Sasaki et al. 1991) and retention (Nordberg and Bergh 1985). Nicotine also improves passive avoidance learning (Sansone et al. 1991). Dose of nicotine can be critically important. In general, low doses facilitate while high doses have no effect or impair memory. Haroutunian and co-workers (Haroutunian et al. i985) found that low but not high doses of nicotine improve passive avoidance. A variety of investigators have found that low doses of nicotine (0.5 mg/kg or less) improve active avoidance performance (Oliverio 1966; Evangelista et al. 1970; Erickson 1971; Orsingher and Fulginiti 1971), while higher doses (above 0.8 mg/kg) impair performance (Oliverio 1966; Essman and Essman 1971; Gilliam and Schlesinger 1985). Genetic factors could explain differential effects of nicotine on cognitive function. Nicotinic binding patterns in a variety of subcortical regions including the hippocampus haves been found to significantly vary in different mouse strains (Marks et al. 1989a). Strain differences in nicotine sensitivity has been seen for a variety of neurobehavioral responses such as thermoregulation, seizure sensitivity and locomotor activity (Marks et al. 1989b; Miner and Collins 1989; Dilsaver et al. 1991). Critical to the current discussion, differential effects of nicotine on memory function have been seen in various strains of rats or mice. Water escape light-dark discrimination learning in a Y-maze is improved in C57 strain mice but impaired in DBA strain mice by pre- or posttrial application of nicotine (Castellano 1976). The DBA strain also is impaired by nicotine in terms of retention of active avoidance of shock, whereas the C57 strain shows a less general deficit (Gilliam and Schlesinger 1985). Particular inbred strains of mice are more sensitive to the enhancing effects of nicotine on shock avoidance and visual discrimination conditioning (Bovet et al. 1966; Bovet-Nitti 1969). In general the strains which showed the poorest performance were aided by nicotine to the greatest degree. Post-trial application of nicotine improves performance on Hebb-Williams maze performance and deficits on shuttle-box avoidance in the same strains of rats (Roman low and high avoidance strains) (Garg 1969). With another strain of rats (Maudsley reactive strain), post-trial nicotine caused an improvement in Hebb-Williams performance (Garg and Holland 1968), but had no effect in an underwater discrimination test (Wraight et al. 1967). Sex differences may underlie some of the heterogeneous effects of nicotine. Gilliam and Schlesinger (1985) found that nicotine caused a deficit in reacquisition of an active avoidance response in C57 strain mice. The effect was most pronounced in females. It did not seem to be due to state-dependent learning effects. Buccafusco and co-workers (Elrod et al. 1988; Jackson and Buccafusco 1989; Jackson et al. 1989; Buccafusco and Jackson 1991) have shown that nicotine improves
choice accuracy of monkeys on delayed matching to sample. In one study (Elrod et al. 1988), five young adult fascicularis monkeys were trained on a delayed matching to sample working memory task in an operant chamber. The animals learned the task and established a stable baseline of performance of 95--100% correct at the 0 s delay, 80-85% correct at the 5 4 0 s delays and 65-75% correct at the 15-60 s delays. Each monkey was given a series of doses of nicotine from 0.625 to 10 gg/kg. There was heterogeneity in the dose response of the monkeys. The best dose for facilitating choice accuracy ranged from 0.625 to 7.5 lag/kg for the different monkeys. The beneficial effect was seen with performance after the long (15--60 s) delays. In a follow-up study, four young adult fascicularis monkeys were trained on delayed matching to sample task in their home cages. Approximately a five percentage point improvement was seen after 7.5 gg/kg. Significant improvement was seen both 10 rain and 24 h after injection. Nicotine most effectively improved accuracy at the longest delays. Nicotine was most effective in improving choice accuracy performance on the most difficult stimulus configurations (Jackson and Buccafusco 1989). This group found pronounced individual differences in dose-response to nicotine. They addressed this problem by determining the most facilitating dose of nicotine for each individual and basing further experimentation on this dose. Individual differences in sensitivity to nicotine together with an inverted U-shaped function for facilitation could explain some of the heterogeneity in findings concerning nicotine induced facilitation of memory performance. Determination of the "best dose" for each subject is a good solution to this problem. However, one must be careful to replicate the effects of the "best dose" before analysis. Otherwise, selective assessment of random fluctuations in performance could give the false impression of a beneficial effect. Rupniak and co-workers (Rupniak et al. 1989) assessed the effectiveness of acute nicotine treatment of improving memory for long lists of objects or reversing scopolamine-induced deficits on spatial delayed response. As with the studies discussed above, they found heterogeneity in response to nicotifie. In the case of this study nicotine treatment caused a trend towards improvement, but this was not statistically significant by their means of evaluation. The nature of the task can explain some of the differential effects of nicotine on memory performance. Nicotine has been found to impair choice accuracy in some studies. Dunnett and Martel (1990) found that nicotine (0.1 and 0.3 mg/kg) impaired choice accuracy in a delayed matching to position task. The deficit was principally seen after trials in which the previous response was to the currently incorrect position. This effect was interpreted as nicotine causing a deficit by increasing proactive interference. Thus, increased mnemonic ability by nicotine may have caused the observed choice accuracy deficits. This interpretation is supported by a recent study by Bushnell (Bushnell, unpublished observations). He used the same delayed matching to position procedure that Dunnett and Martel used except that he had slightly longer intertrial intervals (10 s versus 5 s) and
420
had visual discrimination trials interspersed among the delayed matching trials to concurrently assess reference and working memory. This procedure reduced the influence of proactive interference. Dunnett and Martel found that substituting 15 s intertrial intervals for 5 s intervals virtually eliminated the proactive interference effect. Bushnell administered doses of nicotine (0.18 and 0.3 mg/kg) similar to the doses found by Dunnett and Martel to impair performance accuracy. He did not find any effect of nicotine in control animals using his procedure, supporting the hypothesis that the nicotine deficit in the Dunnett and Martel study was due to increased proactive interference and not to impaired working memory. Mundy and Iwamoto (1988a) conducted a study where the nature of the task may lend insight into the critical effects of nicotine. They found that nicotine impaired acquisition of a food rewarded lever touch operant task. Dose-related deficits were seen when nicotine was injected before or just after daily testing. Mundy and Iwamoto did not seem to be missing a low dose facilitation; they tested very low doses of 0.025 and 0.05 mg/kg nicotine but found no evidence for facilitation. The effect of nicotine was greater when given after daily testing, suggesting that nicotine was affecting consolidation processes. Central administration of chlorisondamine, a nicotinic antagonist, blocked the effect of nicotine, but interestingly it had no effect of its own. This task shows quite different properties in response to brain lesions compared to other memory tasks. Nanry et al. (1989) found that colchicine lesions of the dentate granule cells in the hippocampus caused a dramatic impairment of acquisition in the Morris water maze which tests spatial memory but caused improved performance on this touch task. Level of training on a particular task can also be important for the effects of nicotine. We have found that direct ICV infusion of nicotine will improve choice accuracy in untrained rats (Brucato, Levin, Rose and Swartzwelder, unpublished data). Similar facilitation was not seen in a subsequent experiment where rats were highly trained on the task. However, in that study, ICV nicotine was found to effectively reverse the impairment caused by mecamylamine. Interestingly, with systemic injection opposite effects have been found. Mundy and Iwamoto (1988b) found that pretrial injection of 0.45 mg/kg nicotine impaired choice accuracy in the radial-arm maze in rats with little or no training but had no effect in trained rats. We have found that a lower dose of 0.2 mg/kg will improve radial-arm maze working memory performance in trained rats (Levin and Rose 1991).
of the radial-arm maze. They were then implanted subcutaneously with glass and Silastic pellets containing nicotine base which delivered approximately 12-14 mg/kg/ day nicotine. Control rats were implanted with empty pellets. As shown in Fig. 1, the rats with nicotine-containing pellets significantly improved relative to control rats over the 3 week period of administration (Levin and Rose 1990; Levin et al. t990a). No tolerance to the cognitive enhancing effects of nicotine was seen. The onset of the effect was delayed. It appeared during the second week of administration and persisted during the third week. In a follow-up study, the nicotine-induced facilitation lasted for 4 weeks of administration (Levin et at. 1992a). Bfittig (1970) showed that either pre- or post-training trial injection of nicotine improved performance in a Hebb-Williams maze. The rats were injected every other day with nicotine. Interestingly, they only began to show consistent improvement relative to controls after they had been injected over a period of t0 days. Nelsen (1978) found that performance on a go-no go memory paradigm was improved by chronic nicotine administration but only after a few days of dosing. Welzl et al. (1988) also studied the effect of chronic nicotine administration but did not find facilitation of memory performance. A possible reason for this lack of facilitation is that the chronic blood levels achieved in their study (1656 ng/ml) were considerably lower than those achieved by the glass and Silastic pellet used in our studies (400-500 ng/ml). The lower levels more closely approximate those seen in humans, however, like many other drugs, effective doses of nicotine seem to be about an order of magnitude higher in the rat than in the human.
Chronic effects
Persisting effects
Administration of nicotine over days or weeks at a time has also been found to improve memory performance. We have conducted a series of studies to examine the behavioral effects of chronic nicotine administration in rats. Rats were trained on a working memory version
Nicotine administration has been found to have persisting effects on memory function. Buccafusco and Jackson (1991) have found that acute injections of nicotine improves delayed matching to sample accuracy in monkeys not only 10 min after injection but also 24 h after injec-
8,0 '
7.5' 7.0. m 6 . 5 -'
~, 6'02 LI.I 5.5' 5.0
,
, 2
1 Pre
, 1
, .... 2 Nicotine
, 3
, 1
, 2 Post
Fig. 1. Chronic nicotine and withdrawal effects on working memory performancein the radial-arm maze (entries to repeat, mean±
SEM). (-O-) Control; (--e-) nicotine
421
£
16
1-
742
13'-18
SESSIONS
Fig. 2. Persisting effects of chronic nicotine administration on learning in the radial-arm maze (entries to repeat, mean 4-SEM) * P
Psychopharmacology © Springer-Verlag 1992
Original investigations Nicotinic systems and cognitive function Edward D. Levin Nicotine Research Laboratory, Department of Psychiatry, Duke University Medical Center, Box 3557, Durham, NC 27710, USA ReceivedApril 21, 1992 / Final version May 13, 1992
Abstract. Nicotinic acetylcholine receptors have been found to be important for maintaining optimal performance on a variety of cognitive tasks. In humans, nicotine-induced improvement of rapid information processing is particularly well documented. In experimental animals nicotine has been found to improve learning and memory on a variety of tasks, while the nicotinic antagonist mecamylamine has been found to impair memory performance. Nicotine has been found to be effective in attenuating memory deficits resulting from lesions of the septohippocampal pathway or aging in experimental animals. Nicotinic receptors are decreased in the cortex of patients with Alzheimer's disease. Preliminary studies have found that some aspects of the cognitive deficit in Alzheimer's disease can be attenuated by nicotine. Nicotine may prove to be useful therapeutic treatment for this and other types of dementia.
on tasks assessing attention, learning and memory. However, it is not clear which aspects of cognition actually underlie these improvements. In addition, there are studies which have failed to find improvements and others which have found nicotine-induced performance deficits. These results point to the limits of the beneficial effects of nicotine. They can also be used to help define the aspects of cognition which are affected by nicotine. Much of the impetus for this research is the finding that nicotinic receptors are substantially decreased in Alzheimer's disease. Investigation of nicotinic systems is important for improving the basic understanding of the neural substrates of cognitive function. It can also lead to new avenues of treatment for cognitive disorders such as Alzheimer's disease. Nicotinic agonist effects
Key words: Acetylcholine - Nicotinic - Memory - Nicotine Mecamylamine
Acetytcholine (ACh) systems have long been known to be important for accurate performance on cognitive tasks (see Bartus et al. 1987). The great majority of research concerning cholinergic mechanisms has focused on the muscarinic ACh receptor subtype. One of the most replicated findings in behavioral pharmacology is that the muscarinic blocker scopolamine impairs performance on memory tasks (see Levin 1988a). In contrast, the nicotinic cholinergic receptor subtype has received much less attention. Over the past 3 decades, there has been only a scattering of studies concerning nicotinic involvement in cognitive function. They have demonstrated that as has been shown with muscarinic antagonists, nicotinic antagonists effectively impair performance on memory tasks. In the past several years there has been a blossoming of research concerning the importance of nicotinic systems in cognitive function. Nicotine has been shown in many studies to improve performance
Acute effects While relatively less attention has been paid to nicotinic cholinergic systems compared to muscarinic systems, there is convincing evidence that they are also involved in cognitive processing. A variety of studies have reported acute, nicotine-induced facilitation of cognitive function in humans (Andersson and Post 1974; Houston et al. 1978; Wesnes and Warburton 1983; Wesnes et al. 1983; Peeke and Peeke 1984; Warburton etal. 1986; West and Hack 1991). As reviewed below, many studies have also seen beneficial effects of acute nicotine administration on cognitive performance in experimental animals. However, some investigators have not seen beneficial effects.
Nicotine effects in humans. Nicotine has been tbund to have a variety of cognitive enhancing properties in humans including improved attention and memory functioning (Warburton 1992). Nicotine effects in humans are clouded by the fact that a great majority of the research has been conducted on cigarette smokers. There
418 are several reasons why this presents a problem. First, the population of smokers is self-selected. Pre-existing differences between smokers and nonsmokers may have caused them to begin and continue to smoke. Test-related differences between smokers and nonsmokers may be related to nicotine exposure or to these pre-existing differences. Second, acute effects of nicotine in smokers are quite likely to be altered because of past chronic exposure. Third, since most smokers self-administer nicotine in a chronic fashion, it is difficult to have a proper control condition. Regular smokers may be either under the influence of nicotine or in a state of nicotine withdrawal. Some of the most valuable studies in this area have assessed the effects of nicotine in people who only occasionally smoke or nonsmokers. The most consistent finding of nicotine and cigarette smoking effects on human cognition is an improvement in vigilance and rapid information processing (Wesnes and Warburton 1983; Warburton and Wesnes 1984; Warburton 1992). A variety of studies have shown that smoking improves vigilance performance by attenuating the decline in performance which normally occurs over time (Wesnes and Warburton 1983). However, in some cases it is difficult to determine how much the improved performance after smoking is due to relief of the deficits shown during nicotine withdrawal. A study by Wesnes and Warburton (Wesnes and Warburton 1978) compared vigilance performance of deprived smokers and non-deprived smokers with non-smoking subjects. Both the deprived and non-smoking groups showed declines in performance over time, whereas the smoking group maintained good performance. Smoking-induced improvement in a complex signal detection task was found to be due to both an increase in the number of correct detections and response speed (Wesnes and Warburton 1978). There was no evidence that improvement in accuracy was due to a decrease in response speed or visa versa. Nicotine administration reduces color naming time on the Stroop test where the presence of an irrelevant cue (a color name) slows the naming of a color (Wesnes and Warburton 1978, 1983; Provost and Woodward 1991). This may reflect an effect of nicotine improving selective attention as suggested by Warburton and Wesnes 1984. However, the effect of nicotine is not apparent at the beginning of testing. Provost and Woodward (1991) pointed out that this effect is seen only during repeated testing. The nicotine-induced improvement emerges as a faster improvement in responding. There seems to be some learning or adaptation which takes place such that subjects more quickly process the information or more efficiently screen out irrelevant cues. Nicotine potentiates this process. The converse effect can be seen during nicotine withdrawal. One of the symptoms of nicotine withdrawal is decreased attentiveness. Hatsukami and wo-workers (Hatsukami et al. 1989) found increased errors of commission on a vigilance task in smokers after 24 h of smoking deprivation. Interestingly, no decrease in vigilance was seen with shorter periods of withdrawal. The effect of nicotine on memory function in humans
is less clear. Some investigators have found deficits, no facilitation or variable effects after nicotine administration (Andersson and Hockey 1977; Williams 1980; Peters and McGee 1982; Dunne et al. 1986). Other studies have documented nicotine-induced improvements in memory function (Warburton 1992). Cigarette smoking has been found to improve short-term verbal recall (Peeke and Peeke 1984). Nicotine-induced improvement of short-term memory was also seen in a study by Warburton and co-workers (Warburton et al. 1986). Memory search rate on Sternberg's memory search task is significantly faster in deprived smokers after smoking a nicotine-containing cigarette versus a placebo cigarette (West and Hack t991). This effect was seen both in regular and occasional smokers suggesting that the improvement was not merely due to amelioration of withdrawal effects in regular smokers. Roth and co-workers found (Roth et al. 1992) that cognitive effects of cigarette deprivation and smoking are different in different types of smokers. They studied the effects of smoking a cigarette in the early morning on memory performance on the Austin maze and a word recognition memory test. Subjects who typically smoked their first cigarette within the first hour after getting up performed the maze task and word recognition task significantly better after smoking than after deprivation. The reverse was true for subjects who did not usually smoke their first cigarette in the early morning.
Nicotine effects in experimental animals. Many of the problems associated with assessing acute nicotine effects in humans can be controlled for in animal models. Subjects can be randomly assigned to treatment conditions so that pre-existing individual differences would not affect group composition. Acute administration can be studied separately from chronic administration. Undrugged control tests can be separately assessed from withdrawal states. A variety of animal studies have shown that nicotine improves learning and memory performance (B/ittig 1970; Bovet-Nitti 1966; Buccafusco and Jackson 1991 ; Clarke 1987; Elrod et al. 1988; Erickson 1971; Evangelista and Izquirierdo 1972; Garg and Holland 1969; Haroutunian etal. 1985; Jackson and Buccafusco 1989; Jackson et al. 1989; Levin et al. 1990a, 1992a, b; Levin and Rose 1990, 1991; Nelsen 1978; Nordberg and Bergh 1985; Oliverio 1966, 1968; Orsingher and Fulginiti 1971 ; Pradhan 1970; Sansone et al. 1991; Sasaki et al. t991). However, some studies have not detected effects or have documented negative actions (Essman and Essman 1971 ; Hunter et al. 1977; Dunnett and Martel 1990; Bammer 1982; Mundy and Iwamoto 1988 a, b; Welzl et al. 1988). Several factors, including differences in dose, strain, task structure or level of training, may explain differences in the effects of nicotine on cognitive performance. Several older studies have shown that acute injections of nicotine significantly improved learning. Garg and Holland (1969) showed a nicotine-induced performance improvement in the Hebb-Williams closed field test and Oliverio (1968) found that post-trial nicotine caused an enhancement of learning transfer. Nicotine administra-
419 tion improved performance of a shuttle box conditioned response in rats (Evangelista et al. 1970). Curiously, hexamethonium also improved performance. More recent studies have found that acute pre-trial nicotine found that acute nicotine administration (0.125 mg/kg) improves passive avoidance learning (Sansone et al. 1991 ; Sasaki et al. 1991) and retention (Nordberg and Bergh 1985). Nicotine also improves passive avoidance learning (Sansone et al. 1991). Dose of nicotine can be critically important. In general, low doses facilitate while high doses have no effect or impair memory. Haroutunian and co-workers (Haroutunian et al. i985) found that low but not high doses of nicotine improve passive avoidance. A variety of investigators have found that low doses of nicotine (0.5 mg/kg or less) improve active avoidance performance (Oliverio 1966; Evangelista et al. 1970; Erickson 1971; Orsingher and Fulginiti 1971), while higher doses (above 0.8 mg/kg) impair performance (Oliverio 1966; Essman and Essman 1971; Gilliam and Schlesinger 1985). Genetic factors could explain differential effects of nicotine on cognitive function. Nicotinic binding patterns in a variety of subcortical regions including the hippocampus haves been found to significantly vary in different mouse strains (Marks et al. 1989a). Strain differences in nicotine sensitivity has been seen for a variety of neurobehavioral responses such as thermoregulation, seizure sensitivity and locomotor activity (Marks et al. 1989b; Miner and Collins 1989; Dilsaver et al. 1991). Critical to the current discussion, differential effects of nicotine on memory function have been seen in various strains of rats or mice. Water escape light-dark discrimination learning in a Y-maze is improved in C57 strain mice but impaired in DBA strain mice by pre- or posttrial application of nicotine (Castellano 1976). The DBA strain also is impaired by nicotine in terms of retention of active avoidance of shock, whereas the C57 strain shows a less general deficit (Gilliam and Schlesinger 1985). Particular inbred strains of mice are more sensitive to the enhancing effects of nicotine on shock avoidance and visual discrimination conditioning (Bovet et al. 1966; Bovet-Nitti 1969). In general the strains which showed the poorest performance were aided by nicotine to the greatest degree. Post-trial application of nicotine improves performance on Hebb-Williams maze performance and deficits on shuttle-box avoidance in the same strains of rats (Roman low and high avoidance strains) (Garg 1969). With another strain of rats (Maudsley reactive strain), post-trial nicotine caused an improvement in Hebb-Williams performance (Garg and Holland 1968), but had no effect in an underwater discrimination test (Wraight et al. 1967). Sex differences may underlie some of the heterogeneous effects of nicotine. Gilliam and Schlesinger (1985) found that nicotine caused a deficit in reacquisition of an active avoidance response in C57 strain mice. The effect was most pronounced in females. It did not seem to be due to state-dependent learning effects. Buccafusco and co-workers (Elrod et al. 1988; Jackson and Buccafusco 1989; Jackson et al. 1989; Buccafusco and Jackson 1991) have shown that nicotine improves
choice accuracy of monkeys on delayed matching to sample. In one study (Elrod et al. 1988), five young adult fascicularis monkeys were trained on a delayed matching to sample working memory task in an operant chamber. The animals learned the task and established a stable baseline of performance of 95--100% correct at the 0 s delay, 80-85% correct at the 5 4 0 s delays and 65-75% correct at the 15-60 s delays. Each monkey was given a series of doses of nicotine from 0.625 to 10 gg/kg. There was heterogeneity in the dose response of the monkeys. The best dose for facilitating choice accuracy ranged from 0.625 to 7.5 lag/kg for the different monkeys. The beneficial effect was seen with performance after the long (15--60 s) delays. In a follow-up study, four young adult fascicularis monkeys were trained on delayed matching to sample task in their home cages. Approximately a five percentage point improvement was seen after 7.5 gg/kg. Significant improvement was seen both 10 rain and 24 h after injection. Nicotine most effectively improved accuracy at the longest delays. Nicotine was most effective in improving choice accuracy performance on the most difficult stimulus configurations (Jackson and Buccafusco 1989). This group found pronounced individual differences in dose-response to nicotine. They addressed this problem by determining the most facilitating dose of nicotine for each individual and basing further experimentation on this dose. Individual differences in sensitivity to nicotine together with an inverted U-shaped function for facilitation could explain some of the heterogeneity in findings concerning nicotine induced facilitation of memory performance. Determination of the "best dose" for each subject is a good solution to this problem. However, one must be careful to replicate the effects of the "best dose" before analysis. Otherwise, selective assessment of random fluctuations in performance could give the false impression of a beneficial effect. Rupniak and co-workers (Rupniak et al. 1989) assessed the effectiveness of acute nicotine treatment of improving memory for long lists of objects or reversing scopolamine-induced deficits on spatial delayed response. As with the studies discussed above, they found heterogeneity in response to nicotifie. In the case of this study nicotine treatment caused a trend towards improvement, but this was not statistically significant by their means of evaluation. The nature of the task can explain some of the differential effects of nicotine on memory performance. Nicotine has been found to impair choice accuracy in some studies. Dunnett and Martel (1990) found that nicotine (0.1 and 0.3 mg/kg) impaired choice accuracy in a delayed matching to position task. The deficit was principally seen after trials in which the previous response was to the currently incorrect position. This effect was interpreted as nicotine causing a deficit by increasing proactive interference. Thus, increased mnemonic ability by nicotine may have caused the observed choice accuracy deficits. This interpretation is supported by a recent study by Bushnell (Bushnell, unpublished observations). He used the same delayed matching to position procedure that Dunnett and Martel used except that he had slightly longer intertrial intervals (10 s versus 5 s) and
420
had visual discrimination trials interspersed among the delayed matching trials to concurrently assess reference and working memory. This procedure reduced the influence of proactive interference. Dunnett and Martel found that substituting 15 s intertrial intervals for 5 s intervals virtually eliminated the proactive interference effect. Bushnell administered doses of nicotine (0.18 and 0.3 mg/kg) similar to the doses found by Dunnett and Martel to impair performance accuracy. He did not find any effect of nicotine in control animals using his procedure, supporting the hypothesis that the nicotine deficit in the Dunnett and Martel study was due to increased proactive interference and not to impaired working memory. Mundy and Iwamoto (1988a) conducted a study where the nature of the task may lend insight into the critical effects of nicotine. They found that nicotine impaired acquisition of a food rewarded lever touch operant task. Dose-related deficits were seen when nicotine was injected before or just after daily testing. Mundy and Iwamoto did not seem to be missing a low dose facilitation; they tested very low doses of 0.025 and 0.05 mg/kg nicotine but found no evidence for facilitation. The effect of nicotine was greater when given after daily testing, suggesting that nicotine was affecting consolidation processes. Central administration of chlorisondamine, a nicotinic antagonist, blocked the effect of nicotine, but interestingly it had no effect of its own. This task shows quite different properties in response to brain lesions compared to other memory tasks. Nanry et al. (1989) found that colchicine lesions of the dentate granule cells in the hippocampus caused a dramatic impairment of acquisition in the Morris water maze which tests spatial memory but caused improved performance on this touch task. Level of training on a particular task can also be important for the effects of nicotine. We have found that direct ICV infusion of nicotine will improve choice accuracy in untrained rats (Brucato, Levin, Rose and Swartzwelder, unpublished data). Similar facilitation was not seen in a subsequent experiment where rats were highly trained on the task. However, in that study, ICV nicotine was found to effectively reverse the impairment caused by mecamylamine. Interestingly, with systemic injection opposite effects have been found. Mundy and Iwamoto (1988b) found that pretrial injection of 0.45 mg/kg nicotine impaired choice accuracy in the radial-arm maze in rats with little or no training but had no effect in trained rats. We have found that a lower dose of 0.2 mg/kg will improve radial-arm maze working memory performance in trained rats (Levin and Rose 1991).
of the radial-arm maze. They were then implanted subcutaneously with glass and Silastic pellets containing nicotine base which delivered approximately 12-14 mg/kg/ day nicotine. Control rats were implanted with empty pellets. As shown in Fig. 1, the rats with nicotine-containing pellets significantly improved relative to control rats over the 3 week period of administration (Levin and Rose 1990; Levin et al. t990a). No tolerance to the cognitive enhancing effects of nicotine was seen. The onset of the effect was delayed. It appeared during the second week of administration and persisted during the third week. In a follow-up study, the nicotine-induced facilitation lasted for 4 weeks of administration (Levin et at. 1992a). Bfittig (1970) showed that either pre- or post-training trial injection of nicotine improved performance in a Hebb-Williams maze. The rats were injected every other day with nicotine. Interestingly, they only began to show consistent improvement relative to controls after they had been injected over a period of t0 days. Nelsen (1978) found that performance on a go-no go memory paradigm was improved by chronic nicotine administration but only after a few days of dosing. Welzl et al. (1988) also studied the effect of chronic nicotine administration but did not find facilitation of memory performance. A possible reason for this lack of facilitation is that the chronic blood levels achieved in their study (1656 ng/ml) were considerably lower than those achieved by the glass and Silastic pellet used in our studies (400-500 ng/ml). The lower levels more closely approximate those seen in humans, however, like many other drugs, effective doses of nicotine seem to be about an order of magnitude higher in the rat than in the human.
Chronic effects
Persisting effects
Administration of nicotine over days or weeks at a time has also been found to improve memory performance. We have conducted a series of studies to examine the behavioral effects of chronic nicotine administration in rats. Rats were trained on a working memory version
Nicotine administration has been found to have persisting effects on memory function. Buccafusco and Jackson (1991) have found that acute injections of nicotine improves delayed matching to sample accuracy in monkeys not only 10 min after injection but also 24 h after injec-
8,0 '
7.5' 7.0. m 6 . 5 -'
~, 6'02 LI.I 5.5' 5.0
,
, 2
1 Pre
, 1
, .... 2 Nicotine
, 3
, 1
, 2 Post
Fig. 1. Chronic nicotine and withdrawal effects on working memory performancein the radial-arm maze (entries to repeat, mean±
SEM). (-O-) Control; (--e-) nicotine
421
£
16
1-
742
13'-18
SESSIONS
Fig. 2. Persisting effects of chronic nicotine administration on learning in the radial-arm maze (entries to repeat, mean 4-SEM) * P
