- Email: [email protected]
Nicotinic modulation in an animal model of a form of associative learning impaired in Alzheimer’s disease
Behavioural Brain Research 113 (2000) 11 – 19 www.elsevier.com/locate/bbr
Nicotinic modulation in an animal model of a form of associative learning impaired in Alzheimer’s disease Diana S. Woodruff-Pak a,b,c,*, Isagani S. Santos c a
Department of Psychology, 1701 North 13th Street, Temple Uni6ersity, Philadelphia, PA 19122, USA b Philadelphia Geriatric Center, Philadelphia, PA, USA c Albert Einstein Healthcare Network, Philadelphia, PA, USA Accepted 31 January 2000
Abstract Eyeblink classical conditioning is a widely used associative learning paradigm that has striking behavioral and neurobiological parallels between humans and other mammals. Eyeblink conditioning is impaired in older organisms, and patients with Alzheimer’s disease (AD) are impaired beyond the normal aging deficit. The cholinergic system is of demonstrated involvement in eyeblink conditioning. Blockade of nicotinic cholinergic receptors with mecamylamine prolonged acquisition of conditioned responses (CRs) in young adult rabbits, and the nicotinic agonist, GTS-21 ameliorated conditioning deficits in older rabbits. Galantamine induces allosteric modulation of nicotinic cholinergic receptors to increase acetylcholine release as well as acting as an acetylcholinesterase inhibitor. Galantamine doses of 0.0, 1.0, 2.0, 3.0, and 4.0 mg/kg were tested in ten daily sessions in 40 retired breeder rabbits (mean age =29 months) in the 750 ms delay conditioning paradigm. A dose of 3 mg/kg galantamine was effective in improving conditioning in older rabbits, enabling them to achieve learning criterion rapidly and to produce a very high percentage of CRs. Control tests of rabbits in explicitly unpaired conditions demonstrated that non-associative factors could not account for the results. The efficacy of galantamine in a learning paradigm that shows severe impairment in AD indicates that the drug may be effective as a cognition-enhancer in AD. © 2000 Elsevier Science B.V. All rights reserved. Keywords: Cognition-enhancing drug; Galantamine; Acetylcholine; Nicotinic cholinergic receptor; Memory; Eyeblink; Nictitating membrane response; Classical conditioning
1. Introduction One of the earliest behavioral deficits manifested in patients in the beginning stages of Alzheimer’s disease (AD) is the impaired ability to learn and remember new facts and experiences. Associative learning and memory are affected early in the progression of AD. A prototypical form of associative learning is classical or Pavlovian conditioning. The most widely used Pavlovian conditioning paradigm is eyeblink classical conditioning in which a neutral stimulus such as a tone or light conditioned stimulus (CS) is presented about half a
* Corresponding author. Tel.: +1-215-2041258; fax: + 1-2156356490. E-mail address: [email protected] (D.S. Woodruff-Pak).
second before the onset of a corneal airpuff unconditioned stimulus (US). The learned or conditioned response (CR) is a blink to the CS before the onset of the US. At present, there is almost complete identification of the neural circuitry underlying this form of learning and memory. The essential site of the plasticity for learning resides in the cerebellum ipsilateral to the eye receiving the US. Data on normal aging in non-human mammals and humans support the hypothesis that agerelated changes in the cerebellum are associated with deficits in eyeblink classical conditioning [25,30,37]. In the delay eyeblink classical conditioning procedure, medial-temporal lobe circuitry including the hippocampus is not essential for learning, but disruption in this region can slow the rate of conditioning. Antagonism of the hippocampal cholinergic system is
0166-4328/00/$ - see front matter © 2000 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 6 - 4 3 2 8 ( 0 0 ) 0 0 1 9 6 - 0
12
D.S. Woodruff-Pak, I.S. Santos / Beha6ioural Brain Research 113 (2000) 11–19
one means of prolonging acquisition of conditioned eyeblink responses [21]. AD compromises the medialtemporal lobes and the cholinergic system, and delay eyeblink classical conditioning with a 400-ms interval between the CS and US is a classical conditioning procedure in which patients diagnosed with probable AD are profoundly impaired [20,27,28].
1.1. Eyeblink conditioning in Alzheimer’s disease and other dementing diseases It is our working hypothesis that selective loss of hippocampal pyramidal cells and disruption of the septo-hippocampal cholinergic system impairs acquisition of eyeblink classical conditioning in AD beyond the impairment observed in normal aging. We initially based the prediction of severe disruption of eyeblink classical conditioning in AD patients on neurobiological and behavioral data collected with non-human mammals on eyeblink classical conditioning. Subsequently, West et al. [24] reported selective loss of hippocampal pyramidal cells; the very cells in animals that fire in conjunction with the behavioral CR and unconditioned response (UR). The hypothesis that eyeblink conditioning would be severely impaired in patients with AD was confirmed [27,28], replicated [35], and independently replicated [20]. Using a criterion of producing 25% CRs in a session and combining data from several studies conducted during the last decade in our laboratory, eyeblink conditioning correctly identified 54 of 60 patients with probable AD [26]. In some cases eyeblink classical conditioning was effective in differentiating cerebrovascular dementia from probable AD [35]. In patients with Huntington’s disease [34] and Parkinson’s disease [7], eyeblink conditioning is relatively normal and clearly differentiated from eyeblink classical conditioning in AD. In addition to working with probable AD and other dementing neurological diseases of old age, we extended this work to adults with Down’s syndrome who inevitably develop AD-like neuropathology around the age of 35 years (called DS/AD). Patients with DS/AD perform eyeblink classical conditioning similarly to probable AD patients, but young adults with DS perform eyeblink classical conditioning in a manner comparable to normal older adults. That is, the eyeblink classical conditioning performance of younger adults with DS is not equal to the performance of normal young adults, but it is significantly better than adults with DS over the age of 35 years [36].
1.2. Predictions about eyeblink conditioning and Alzheimer’s disease based on an animal model It was demonstrated in animal studies that neuronal
unit activity in the hippocampus increased markedly within trials early in the eyeblink classical conditioning process. Activity recorded in the CA1 and CA3 regions of the hippocampus formed a predictive ‘model’ of the amplitude-time course of the learned behavioral response. Firing to the tone-CS and corneal airpuff-US occurred only under conditions when the stimuli were paired and produced behavioral learning [2]. When the CS and US were presented independently in the explicitly unpaired paradigm, there was no hippocampal response to the tones or airpuffs [5]. The hippocampal modeling of the behavioral CR and UR was generated largely by hippocampal pyramidal neurons in the CA1 and CA3 fields [4]. In neuropathological studies of human brains, West et al. [24] identified pyramidal cells in the CA1 field of hippocampus as the cells selectively lost in AD. The rabbit model system demonstrated that the hippocampus could play a modulatory role in conditioning [3]. Disruption or facilitation of hippocampal activity affected the rate of conditioning. In rabbits, disruption of muscarinic cholinergic receptors with scopolamine injections impaired acquisition of CRs, and this disruption occurred only when the hippocampus was intact [21]. Scopolamine also disrupted eyeblink conditioning in humans [1,19].
1.3. A role for nicotinic cholinergic receptors in eyeblink classical conditioning Artificial induction of behavioral deficits in the acquisition of CRs comparable to those observed with the muscarinic cholinergic antagonist, scopolamine have been reported using the nicotinic cholinergic antagonist, mecamylamine [29,32]. When mecamylamine was injected into young rabbits, acquisition was prolonged and resembled acquisition in older rabbits [32]. Mecamylamine also impairs non-verbal learning and memory in humans [12]. Nicotinic receptors are significantly reduced in cerebral cortex and hippocampal regions of the brain in AD [10,18]. In particular, the a4b2 nicotinic cholinergic receptor subtype is vulnerable in AD [23]. A drug in development for cognition-enhancement in AD, GTS21 is a selective agonist at the neuronal nicotinic a7 nicotinic cholinergic receptor subtype [8,9]. This drug improves acquisition of CRs in older rabbits so that they learn as well as young rabbits [33], and it reverses the effect of mecamylamine in the disruption of learning. Stimulation of remaining nicotinic receptors in the brains of patients with AD might alleviate some cognitive deficits due to cholinergic neuron dysfunction. Intravenous administration of nicotine to patients diagnosed as probable AD produced some improve-
D.S. Woodruff-Pak, I.S. Santos / Beha6ioural Brain Research 113 (2000) 11–19
ments in cognitive performance, but anxiety and other side effects were observed [13,14]. A novel nicotine derivative with highly selective binding to central nicotinic receptors, ABT-418 was effective in ameliorating memory deficits in six patients with moderate AD [16]. There is continuing interest in developing new brain-selective nicotinic agonists to inhibit the progressive loss of cognitive function in AD and other neurodegenerative diseases as evidenced by many of the articles in this Special Issue. A drug that acts by a dual mechanism of action that affects nicotinic cholinergic neurotransmission is galantamine. Galantamine is a tertiary alkaloid originally derived from the bulbs of the snowdrop and various Narcissus species [15]. It induces allosteric modulation of nicotinic acetylcholine receptors to increase acetylcholine release as well as acting as a reversible, competitive, selective inhibitor of acetylcholinesterase. Galantamine is a reversible acetylcholinesterase inhibitor that interacts selectively and competitively with the enzyme acetylcholinesterase [22]. Research also indicates that galantamine is an allosteric modulator at nicotinic cholinergic receptor sites [11,17]. It is proposed that galantamine increases cholinergic function in two ways: (a) increasing the concentration of acetylcholine through a competitive reversible inhibition of acetylcholine hydrolysis by the enzyme acetylcholinesterase; and (b) increasing the release of acetylcholine and other neurotransmitters such as glutamate through an allosteric modulation of acetylcholine effects at nicotinic cholinergic receptor sites. Because galantamine improved cognitive deficits in animal models involving cholinergic deficits, and because the cholinergic systems is of demonstrated involvement in eyeblink classical conditioning, we anticipated that galantamine would be effective in ameliorating conditioning deficits in older rabbits. Our initial aim was to identify a dose or doses of galantamine that showed measurable effects on the acquisition of CRs.
2. Materials and methods
2.1. Subjects A total of 40 female retired breeder specific pathogen free (SPF) New Zealand white rabbits completed this experiment. Birth dates of the rabbits were recorded by the breeder (Covance, Inc.). Rabbits ranged in age from 24 to 52 months, and the mean age was 29.2 months (S.D.= 6.6). A one-way analysis of variance comparing the various groups by age indicated no differences among the groups. Mean weight of the 40 rabbits was 4.5 kg (S.D.=0.58), and the range was 3.0 – 5.7 kg with no significant group differences in weight. Rabbits were individually housed in stainless steel cages in an
13
AAALAC-accredited animal facility. They were fed rabbit chow ad lib. and tap water and had a 12/12 h light/dark cycle.
2.2. Apparatus and procedure At least 24 h after arrival at the animal facility, rabbits were adapted twice in the restraining apparatus in sessions separated by 24 h. They were placed in Plexiglas restrainers for a 1-h adaptation session. After the second 1-h adaptation session, rabbits were given a local ophthalmic anesthetic (proparacaine hydrochloride) in the left eye so that a 6-0 nylon suture loop could be placed in the temporal margin of the nictitating membrane (NM). A patch of fur approximately 3 cm2 was shaved on the back, just below the head to expose the skin for subcutaneous (s.c.) injections. The classical conditioning equipment attached to the rabbit’s head included elastic eyelid retractors and a platform holding a minitorque potentiometer (San Diego Instruments prototype model) for NM movement measurement that was secured under the animal’s muzzle and behind the ears. The potentiometer was attached by a lever and a thread to the nylon suture loop in the NM. Analog output from the potentiometer was digitized and read into and IBM-PC-compatible system described by Chen and Steinmetz [6]. This system also controlled the timing and presentation of conditioning stimuli. For classical conditioning, the CS was an 850-ms, 85 dB, 1-kHz tone, followed 750 ms after its onset by a 100-ms, 3-psi corneal airpuff US. The CS and US coterminated. The inter-trial-interval was random, ranging between 20 and 30 s at 1-s intervals. One training session lasted about 45 min. Rabbits were tested four at a time. Each training session was controlled by a program written in C+ + language [6] and run on an IBM-PCcompatible 386 computer. Data were collected about the position of the NM in 3 ms bins during the trials. NM responses were displayed on the monitor, and the amplitudes, areas, and latencies of the response relative to baseline were calculated. In addition, averages were calculated for every block of nine trials. A CR was scored if the NM was pulled back a minimum of 0.5 mm in the interval between 25 and 750 ms after CS onset. The dependent measure, learning criterion, was scored as the number of training trials it took the animal to produce eight CRs within nine consecutive trials. Data were collected in RAM and saved to a hard drive, and individual data summaries for each of the four rabbits run simultaneously were printed at the end of each session. Rabbits in the explicitly unpaired condition were treated in a fashion identical to rabbits tested in the paired condition with the exception that the 850-ms, 85
14
D.S. Woodruff-Pak, I.S. Santos / Beha6ioural Brain Research 113 (2000) 11–19
dB, 1-kHz tone CS and 100-ms, 3-psi corneal air puff US were never paired. In the unpaired condition, rabbits received at total of 90 stimuli, 45 tone CSs and corneal air puff 45 USs. Each stimulus was presented at an inter-trial-interval that was random and ranged between 20 and 30 s. Thus, the duration of the session and the inter-trial interval was identical in the paired and unpaired sessions, and all rabbits were tested four at a time.
2.4. Research design There were five groups of rabbits in the paired condition treated with 0 (vehicle), 1, 2, 3, and 4 mg/kg galantamine. The vehicle group in the paired condition had eight rabbits, and there were four rabbits tested at each dose of galantamine. In the unpaired groups there were eight rabbits tested with 3 mg/kg galantamine and eight rabbits tested with vehicle.
2.3. Drugs 3. Results Galantamine was purchased from Tocris Cookson, Inc. (Ballwin, MO). The vehicle solution was sterile saline. Galantamine was freshly prepared in solution each week and injected s.c. a minimum of 15 min before behavioral testing to ensure that peak blood levels of galantamine were attained during the testing session. Rabbits were tested between 15 and 30 min after injections, and behavioral testing was completed a maximum of 1 h and 15 min after injection. Equal volumes were injected into the animals, including vehicle-treated animals, based on weight. In this experiment we report on the behavioral results from ten daily injections of various doses of galantamine or vehicle.
Fig. 1. The number of paired trials of tone conditioned stimulus and corneal air puff unconditioned stimulus in the 750-ms delay eyeblink classical conditioning paradigm it took retired breeder rabbits to attain a learning criterion of eight conditioned responses in nine consecutive trials. There were four rabbits treated at each dose of galantamine, and eight rabbits treated with sterile saline vehicle. The difference between rabbits treated with the 3.0-mg/kg dose or vehicle was significant at the 0.01 level of confidence (**PB 0.01). Error bars are standard deviation.
The major aim of this experiment was to identify a dose of galantamine that was most effective in ameliorating the learning deficits in the 750 ms delay eyeblink classical conditioning procedure in older rabbits. A one-way analysis of variance (ANOVA) was used to compare the trials to learning criterion attained by the 24 rabbits in the paired condition treated with doses of 0.0, 1.0, 2.0, 3.0, or 4.0 mg/kg galantamine. The effect of dose was statistically significant, F (4, 19)=3.19; PB 0.05. Post hoc comparisons using the Dunnett Control Group Comparison test indicated that the difference in trials to criterion between rabbits treated with the 3.0 mg/kg dose of galantamine and rabbits treated with sterile saline vehicle was significant at the 0.01 level of confidence (Fig. 1). Among the 24 older rabbits treated with doses ranging from 0.0 to 4.0 mg/kg galantamine, six failed to achieve learning criterion and were assigned a score of 1400 (the 1350 paired trials they experienced plus 50, which is still likely an underestimate of the number of trials to criterion for these animals). There were four rabbits in the vehicle-treated group that did not achieve criterion, and one rabbit each in the 1.0- and 4.0-mg/kg galantamine groups. All rabbits in the groups treated with 2.0 and 3.0 mg/kg attained learning criterion within 1350 trials. To examine the effect of various doses of galantamine on eyeblink classical conditioning over training sessions, a five (dose of 0.0, 1.0, 2.0, 3.0, or 4.0 mg/kg galantamine) by ten (training sessions) repeated measures ANOVA was carried out on the dependent measure of percentage of CRs. The percentage of CRs was different depending on the dose of galantamine, F (4, 19)= 6.33, PB 0.01. Post hoc comparisons using the Dunnett Control Group Comparison test indicated that rabbits treated with a dose of 3 mg/kg galantamine produced significantly more CRs than did vehicletreated rabbits (PB 0.01; Fig. 2). The effect of training sessions was also significant, with an increase in the percentage of CRs over sessions, F (9, 171)= 14.48, PB 0.0001. The interaction effect indicating that dose level affected the change in percentage of CRs over training session was also significant, F (36, 171)=1.93,
D.S. Woodruff-Pak, I.S. Santos / Beha6ioural Brain Research 113 (2000) 11–19
15
Fig. 2. Left: Percentage of conditioned responses (CRs) for each of ten daily sessions of paired tone conditioned stimulus and corneal air puff unconditioned stimulus in the 750-ms delay eyeblink classical conditioning paradigm for retired breeder rabbits treated with daily doses of 0, 1.0, 2.0, 3.0, or 4.0 mg/kg galantamine. Right: Mean of the percentage of CRs over ten training sessions for the data shown on the left. Error bars are standard deviation. *P B 0.05; **PB 0.01; ***PB0.001.
P B0.01. Post hoc analysis of the significant dose by training session interaction effect indicated that there were significant differences between groups in Sessions 3, 4, 6, 7, 9, 10 as shown in Fig. 2. With the exception of Session 9 in which the significance level exceeded 0.05, the differences exceeded the 0.01 level of confidence. A five (dose of 0.0, 1.0, 2.0, 3.0, or 4.0 mg/kg galantamine) by ten (training sessions) repeated measures ANOVA was carried out on the dependent measure of CR amplitude. The main effects of dose and session were statistically significant, but the dose by session interaction effect did not attain significance at the 0.05 level of confidence. CR amplitude was different depending on the dose of galantamine, F (4, 19) = 5.76, PB 0.01. The effect of training sessions was also significant, with an increase in CR amplitude over sessions, F (9, 171) =2.48, PB 0.01 (Fig. 3). The magnitude of the motor or unconditioned response was not affected by drug dose or training. A five (dose of 0.0, 1.0, 2.0, 3.0, or 4.0 mg/kg galantamine) by ten (training sessions) repeated measures ANOVA was carried out on the dependent measure of UR amplitude, but neither the effect of dose (Fig. 3) nor the training session or dose by training session interaction effects were statistically significant. To control for the possibility that galantamine affected the blink response to the tone CS or to the corneal air puff US, we tested the 3.0-mg/kg dose of
galantamine in the explicitly unpaired condition in which rabbits received randomly presented CS and US stimuli identical to the stimuli experienced by the rabbits in the paired condition. A two (dose of 0.0 or 3.0 mg/kg galantamine) by ten (training session) repeated measures ANOVA comparing percentage of blink responses in the CS period revealed no significant effects for dose, training session, or dose by training session interaction. A two (dose of 0.0 or 3.0 mg/kg galantamine) by ten (training session) repeated measures ANOVA comparing the amplitude of the blink response in the CS period also revealed no significant effects (Fig. 4). A two (dose of 0.0 or 3.0 mg/kg galantamine) by ten (training session) repeated measures ANOVA comparing the amplitude of the UR revealed no significant dose or dose by training session interaction effect. The effect of training session was significant, and as can be seen in Fig. 4, this effect was a trend to increased UR amplitude in later sessions by both groups.
4. Discussion Testing five groups of rabbits at four doses of galantamine and vehicle, we were impressed by the magnitude and consistency of differences in eyeblink classical conditioning produced by the optimal dose of the drug, 3.0 mg/kg. Trials to learning criterion, a measure that is
16
D.S. Woodruff-Pak, I.S. Santos / Beha6ioural Brain Research 113 (2000) 11–19
larger when learning is poorer, revealed a classical U-shaped response curve with doses of 1.0 and 2.0 mg/kg galantamine producing non-significant effects in comparison to rabbits treated with vehicle, a dose of 3.0 mg/kg galantamine reducing the number of trials to learning criterion to a very low mean significantly dif-
ferent from vehicle-treated rabbits, and 4.0 mg/kg galantamine producing a non-significant effect. Total percentage of CRs over the ten training sessions, a measure that is higher with better performance, produced an inverted-U shaped response curve. Galantamine doses of 1.0 and 2.0 mg/kg were under-doses,
Fig. 3. Amplitude of the conditioned response for each of ten daily sessions of paired tone conditioned stimulus and corneal air puff unconditioned stimulus in the 750-ms delay eyeblink classical conditioning paradigm for retired breeder rabbits treated with daily doses of 0, 1.0, 2.0, 3.0, or 4.0 mg/kg galantamine. Right: Mean of the percentage of unconditioned responses over ten training sessions for the same rabbits whose data are shown on the left. Error bars are standard deviation.
Fig. 4. Left: Amplitude of the UR for each of ten daily sessions of 90 explicitly unpaired tones or corneal air puffs for retired breeder rabbits treated with daily doses of 0 or 3.0 mg/kg galantamine. Right: Mean of the percentage of blink responses to a tone over ten training sessions for the same rabbits whose data are shown on the left.
D.S. Woodruff-Pak, I.S. Santos / Beha6ioural Brain Research 113 (2000) 11–19
the dose of 3.0 mg/kg was optimal, and the dose of 4.0 mg/kg was an over-dose for the acquisition of CRs in the 750-ms eyeblink classical conditioning procedure. Older rabbits treated with 3 mg/kg galantamine produced a high percentage of CRs early in training and attained a high performance level of over 90% CRs by Session 10 typically seen only in young rabbits. In our laboratory for a period covering almost a decade, hundreds of rabbits of various ages have been treated with drugs or vehicle and tested with the 750-ms delay eyeblink classical conditioning procedure. Young adult rabbits treated with vehicle attain learning criterion in around 400 trials, and retired breeder rabbits treated with vehicle attain learning criterion in around 1000 trials, with standard deviations of 200 – 300 trials. Older rabbits receiving daily injections of a dose of 3.0 mg/kg galantamine achieved learning criterion in a mean of 233.0 trials (S.D.=176.9). The identification of a dose level of galantamine that enables older rabbits to acquire CRs at a rate that is approaching half the number of trials to criterion required by young rabbits is a striking result. The high percentage of CRs maintained by older rabbits treated with 3 mg/kg galantamine is also exceptional by our laboratory standards for older rabbits. In independent studies using the eyeblink classical paradigm with the 750-ms delay procedure in older rabbits, we have tested the effects of a nicotinic agonist (GTS-21) and the effects of cholinesterase inhibitors (physostigmine, donepezil). For each drug, we tested the effect of at least two doses on learning in comparison to sterile saline vehicle. There were statistically significant effects on trials to learning criterion and percentage of CRs in all three of these drugs. However, the magnitude of the effects of the most optimal dose of GTS-21, physostigmine, or donepezil were not as large as the effect of a dose of 3.0 mg/kg galantamine observed in the present experiment. Trials to learning criterion in rabbits treated with 3.0 mg/kg galantamine were less than half the trials to learning criterion in rabbits treated with either a nicotinic agonist or either one of the cholinesterase inhibitors. Because galantamine induces allosteric modulation of nicotinic cholinergic receptors to increase acetylcholine release as well as acting as an acetylcholinesterase inhibitor, it may ameliorate learning impairment in older rabbits more effectively than drugs with one mechanism of action. Of course, additional tests with a larger sample size are required to confirm these initial results with galantamine. The result that 3.0 mg/kg galantamine caused a reduction in the number of trials to achieve learning criterion and increased the percentage of CRs did not occur because galantamine affected blink rate or blink amplitude. In the paired condition, UR amplitude was not different among the five groups of rabbits treated
17
with different doses of galantamine or vehicle. The learned response (the CR), but not the motor response (the UR) was affected by galantamine at a dose of 3.0 mg/kg. The explicitly unpaired control procedure testing rabbits’ response to independently presented tone and air puff stimuli demonstrated no group differences between rabbits treated with 3.0 mg/kg galantamine or vehicle. The percentage of eyeblinks during the CS period, blink amplitude during the CS period, and the amplitude of the eyeblink to the air puff were similar in galantamine- or vehicle-treated animals. The effect of a dose of 3.0 mg/kg galantamine appeared to be on mechanisms of associative learning. The one statistically significant effect in the analyses of data in the explicitly unpaired conditioning sessions was a significant effect of Session. Over sessions there was a tendency for UR amplitude to increase in both the group treated with vehicle and in the group treated with 3.0 mg/kg galantamine. However, the increase in UR amplitude was not accompanied by an increase in blink amplitude during the CS period. Blink amplitude during the CS period remained stable through the ten training sessions. Rabbits tested in the explicitly unpaired condition showed some increase in the magnitude of their reflexive blink to the US, but they did not change their rate of blinking as indexed by the percentage of blinks during the CS period. They did not increase the magnitude of the blink during the CS period, either. In addition, this trend of a change in UR amplitude in the unpaired condition was consistent across the galantamine- and vehicle-treated groups. There was no non-associative learning effect selectively present in galantamine-treated rabbits. The present experiment with the aim of identifying an optimal dose of galantamine for testing in the eyeblink classical conditioning paradigm is a first step in a program of research to identify the means by which galantamine affects associative learning and memory. One of the advantages of using eyeblink classical conditioning in pharmacological testing is that the neural circuitry and sites of plasticity for learning are described. The cerebellum is the essential brain structure for the acquisition and retention of CRs, and it is possible that galantamine acts in the cerebellum of older rabbits to ameliorate learning. However, experiments have demonstrated that the disruptive effect of scopolamine on eyeblink conditioning [21] and the ameliorating effect of nefiracetam on eyeblink conditioning [31] occur only when the hippocampus is intact. The hippocampus, with its higher concentration of nicotinic cholinergic receptors may be the site where the dual mechanisms of galantamine are effective in improving the rate of learning. Galantamine induces allosteric modulation of nicotinic cholinergic receptors to increase acetylcholine release as well as acting as an acetylcholinesterase inhibitor. Using this dual mecha-
18
D.S. Woodruff-Pak, I.S. Santos / Beha6ioural Brain Research 113 (2000) 11–19
nism of action, galantamine appears to ameliorate learning impairment in older rabbits more effectively than drugs with one mechanism of action.
[13]
Acknowledgements
[14]
A number of individuals assisted at various stages of this project. We would like to thank Tara Orlando for assistance with data collection, Richard Vogel for assistance with data entry, and Michelle Pak for assistance at the data collection, entry, and analyses phases. Dan Lyons, Manager of the Central Animal Facility at Albert Einstein Healthcare Network facilitated our work in many ways. This research was supported by a grant from Janssen Pharmaceutica, Inc.
[15]
[16]
[17]
[18]
References [19] [1] Bahro M, Schreurs BG, Sunderland T, Molchan SE. The effects of scopolamine, lorazepam, and glycopyrrolate on classical conditioning of the human eyeblink response. Psychopharmacology 1995;122:395 – 400. [2] Berger TW, Alger B, Thompson RF. Neuronal substrate of classical conditioning in the hippocampus. Science 1976;192:483 – 5. [3] Berger TW, Berry SD, Thompson RF. Role of the hippocampus in classical conditioning of aversive and appetitive behaviors. In: Isaacson RL, Pribram KH, editors. The Hippocampus. New York: Plenum, 1986:203–39. [4] Berger TW, Rinaldi PC, Weisz DJ, Thompson RF. Single-unit analysis of different hippocampal cell types during classical conditioning of rabbit nictitating membrane response. J Neurophysiol 1983;50:1197–219. [5] Berger TW, Thompson RF. Neuronal plasticity in the limbic system during classical conditioning of the rabbit nictitating membrane response. I. The hippocampus. Brain Res 1978;145:323 – 46. [6] Chen G, Steinmetz JE. A general-purpose computer system for behavioral conditioning and neural recording experiments. Behav Res Meth Instruments Computers 1998;30:384–91. [7] Daum I, Schugens MM, Breitenstein C, Topka H, Spieker S. Classical eyeblink conditioning in Parkinson’s disease. Movement Disorders 1996;11:639–46. [8] De Fiebre CM, Meyer EM, Henry JC, Muraskin SI, Kem WR, Papke RL. Characterization of a series of anabaseine-derived compounds reveals that the 3-(4)-dimethylaminocinnamylidine derivative is a selective agonist at neuronal nicotinic alpha7/125alpha-bungarotoxin receptor subtypes. Mol Pharmacol 1995;47:164 – 71. [9] Kem WR, Mahnir VM, Papke RL, Lingle CJ. Anabaseine is a potent agonist on muscle and neuronal alpha-bungarotoxin-sensitive nicotinic receptors. Pharmacol Exp Ther 1997;283:979 – 92. [10] London ED, Ball MJ, Waller SB. Nicotinic binding sites in cerebral cortex and hippocampus in Alzheimer’s dementia. Neurochem Res 1989;14:745–50. [11] Maelicke A, Schrattenholz A, Periera EFR, Albuquerque EX. Nicotinic receptors in the central nervous system: Implications for Alzheimer’s disease. Abstract no. 12: 6th International Conference on Alzheimer’s Disease and Related Disorders, 18 – 23 July 1998, Amsterdam. [12] Newhouse PA, Potter A, Corwin J, Lenox R. Age-related effects
[20]
[21]
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
[30]
of the nicotinic antagonist mecamylamine on cognition and behavior. Neuropsychopharmacology 1994;10:93 – 107. Newhouse PA, Potter A, Lenox R. The effects of nicotinic agents on human cognition: possible therapeutic applications in Alzheimer’s and Parkinson’s diseases. Med Chem Res 1993;2:628 – 42. Newhouse P, Sunderland T, Tariot P. Intravenous nicotine in AD. Psychopharmacology 1988;95:171 – 5. Parys W. Development of Reminyl (galantamine), a novel acetylcholinesterase inhibitor, for the treatment of Alzheimer’s disease. Alzheimer’s Rep 1998;1:S19 – 20. Potter A, Corwin J, Lang J, Piasecki M, Lenox R, Newhouse PA. Acute effects of the selective cholinergic channel activator (nicotinic agonist) ABT-418 in Alzheimer’s disease. Psychopharmacology 1999;142:334 – 42. Schrattenholz A, Pereira EFR, Roth U, et al. Agonist responses of neuronal nicotinic acetylcholine receptors are potentiated by a novel class of allosterically acting ligands. Mol Pharmacol 1996;49:1 – 6. Schroder G, Giacobini E, Struble RG, Zilles K, Maelicke A. Nicotinic cholinergic receptive neurons of the frontal cortex are reduced in Alzheimer’s disease. Neurobiol Aging 1991;12:259– 62. Solomon PR, Groccia-Ellison M, Flynn D, Mirak J, Edwards KR, Dunehew A, Stanton ME. Disruption of human eyeblink conditioning after central cholinergic blockade with scopolamine. Behav Neurosci 1993;107:271 – 9. Solomon PR, Levine E, Bein T, Pendlebury WW. Disruption of classical conditioning in patients with Alzheimer’s disease. Neurobiol Aging 1991;12:283 – 7. Solomon PR, Solomon SD, Vander Schaaf E, Perry HE. Altered activity in the hippocampus is more detrimental to classical conditioning than removing the structure. Science 1983;220:329– 31. Thomsen T, Kewitz H. Selective inhibition of human acetylcholinesterase by galantamine in vitro and in vivo. Life Sci 1990;46:1553 – 8. Warpman U, Nordberg A. Epibatidine and ABT 418 reveal selective losses of alpha 4 beta 2 nicotinic receptors in Alzheimer brains. NeuroReport 1995;6:2419 – 23. West MJ, Coleman PD, Flood DG, Troncoso JC. Differences in the pattern of hippocampal neuronal loss in normal aging and Alzheimer’s disease. Lancet 1994;344:769 – 72. Woodruff-Pak DS. Eyeblink classical conditioning in normal aging and Alzheimer’s disease. In: Woodruff-Pak DS, Steinmetz JE, editors. Eyeblink Classical Conditioning: Volume I, Applications in Humans. Amsterdam: Kluwer, 2000:163 – 189. Woodruff-Pak DS. Insights about learning in Alzheimer’s disease from the animal model. In: Carroll ME, Overmier JB, editors. Linking Animal Research and Human Psychological Health. Washington, DC: American Psychological Association, in press. Woodruff-Pak DS, Finkbiner RG, Katz IR. A model system demonstrating parallels in animal and human aging: extension to Alzheimer’s disease. In: Meyer EM, Simpkins JW, Yamamoto J, editors. Novel Approaches to the Treatment of Alzheimer’s Disease. New York: Plenum, 1989:355 – 71. Woodruff-Pak DS, Finkbiner RG, Sasse DK. Eyeblink conditioning discriminates Alzheimer’s patients from non-demented aged. NeuroReport 1990;1:45 – 8. Woodruff-Pak DS, Hinchliffe RM. Scopolamine- or mecamylamine-induced learning impairment: reversed by nefiracetam. Psychopharmacology 1997;131:130 – 9. Woodruff-Pak DS, Jaeger ME. Predictors of eyeblink classical conditioning over the adult age span. Psychol Aging 1998;13:193 – 205.
D.S. Woodruff-Pak, I.S. Santos / Beha6ioural Brain Research 113 (2000) 11–19 [31] Woodruff-Pak DS, Li Y-T, Hinchliffe RM, Port RL. Hippocampus in delay eyeblink classical conditioning: essential for nefiracetam amelioration of learning in older rabbits. Brain Res 1997;747:207 – 18. [32] Woodruff-Pak DS, Li Y-T, Kazmi A, Kem WR. Nicotinic cholinergic system involvement in eyeblink classical conditioning in rabbits. Behav Neurosci 1994;108:486–93. [33] Woodruff-Pak DS, Li Y-T, Kem WR. A nicotinic agonist (GTS-21), eyeblink classical conditioning, and nicotinic receptor binding in rabbit brain. Brain Res 1994;645:309–17. [34] Woodruff-Pak DS, Papka M. Huntington’s disease and eye-
.
19
blink classical conditioning: normal learning but abnormal timing. J Int Neuropsychol Soc 1996;2:323 – 34. [35] Woodruff-Pak DS, Papka M, Romano S, Li Y-T. Eyeblink classical conditioning in Alzheimer’s disease and cerebrovascular dementia. Neurobiol Aging 1996;17:505 – 12. [36] Woodruff-Pak DS, Papka M, Simon EW. Eyeblink classical conditioning in Down’s Syndrome, Fragile X Syndrome, and normal adults over and under age 35. Neuropsychology 1994;8:14 – 24. [37] Woodruff-Pak DS, Trojanowski JQ. The older rabbit as an animal model: implications for Alzheimer’s disease. Neurobiol Aging 1996;17:283 – 90.
Nicotinic modulation in an animal model of a form of associative learning impaired in Alzheimer’s disease Diana S. Woodruff-Pak a,b,c,*, Isagani S. Santos c a
Department of Psychology, 1701 North 13th Street, Temple Uni6ersity, Philadelphia, PA 19122, USA b Philadelphia Geriatric Center, Philadelphia, PA, USA c Albert Einstein Healthcare Network, Philadelphia, PA, USA Accepted 31 January 2000
Abstract Eyeblink classical conditioning is a widely used associative learning paradigm that has striking behavioral and neurobiological parallels between humans and other mammals. Eyeblink conditioning is impaired in older organisms, and patients with Alzheimer’s disease (AD) are impaired beyond the normal aging deficit. The cholinergic system is of demonstrated involvement in eyeblink conditioning. Blockade of nicotinic cholinergic receptors with mecamylamine prolonged acquisition of conditioned responses (CRs) in young adult rabbits, and the nicotinic agonist, GTS-21 ameliorated conditioning deficits in older rabbits. Galantamine induces allosteric modulation of nicotinic cholinergic receptors to increase acetylcholine release as well as acting as an acetylcholinesterase inhibitor. Galantamine doses of 0.0, 1.0, 2.0, 3.0, and 4.0 mg/kg were tested in ten daily sessions in 40 retired breeder rabbits (mean age =29 months) in the 750 ms delay conditioning paradigm. A dose of 3 mg/kg galantamine was effective in improving conditioning in older rabbits, enabling them to achieve learning criterion rapidly and to produce a very high percentage of CRs. Control tests of rabbits in explicitly unpaired conditions demonstrated that non-associative factors could not account for the results. The efficacy of galantamine in a learning paradigm that shows severe impairment in AD indicates that the drug may be effective as a cognition-enhancer in AD. © 2000 Elsevier Science B.V. All rights reserved. Keywords: Cognition-enhancing drug; Galantamine; Acetylcholine; Nicotinic cholinergic receptor; Memory; Eyeblink; Nictitating membrane response; Classical conditioning
1. Introduction One of the earliest behavioral deficits manifested in patients in the beginning stages of Alzheimer’s disease (AD) is the impaired ability to learn and remember new facts and experiences. Associative learning and memory are affected early in the progression of AD. A prototypical form of associative learning is classical or Pavlovian conditioning. The most widely used Pavlovian conditioning paradigm is eyeblink classical conditioning in which a neutral stimulus such as a tone or light conditioned stimulus (CS) is presented about half a
* Corresponding author. Tel.: +1-215-2041258; fax: + 1-2156356490. E-mail address: [email protected] (D.S. Woodruff-Pak).
second before the onset of a corneal airpuff unconditioned stimulus (US). The learned or conditioned response (CR) is a blink to the CS before the onset of the US. At present, there is almost complete identification of the neural circuitry underlying this form of learning and memory. The essential site of the plasticity for learning resides in the cerebellum ipsilateral to the eye receiving the US. Data on normal aging in non-human mammals and humans support the hypothesis that agerelated changes in the cerebellum are associated with deficits in eyeblink classical conditioning [25,30,37]. In the delay eyeblink classical conditioning procedure, medial-temporal lobe circuitry including the hippocampus is not essential for learning, but disruption in this region can slow the rate of conditioning. Antagonism of the hippocampal cholinergic system is
0166-4328/00/$ - see front matter © 2000 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 6 - 4 3 2 8 ( 0 0 ) 0 0 1 9 6 - 0
12
D.S. Woodruff-Pak, I.S. Santos / Beha6ioural Brain Research 113 (2000) 11–19
one means of prolonging acquisition of conditioned eyeblink responses [21]. AD compromises the medialtemporal lobes and the cholinergic system, and delay eyeblink classical conditioning with a 400-ms interval between the CS and US is a classical conditioning procedure in which patients diagnosed with probable AD are profoundly impaired [20,27,28].
1.1. Eyeblink conditioning in Alzheimer’s disease and other dementing diseases It is our working hypothesis that selective loss of hippocampal pyramidal cells and disruption of the septo-hippocampal cholinergic system impairs acquisition of eyeblink classical conditioning in AD beyond the impairment observed in normal aging. We initially based the prediction of severe disruption of eyeblink classical conditioning in AD patients on neurobiological and behavioral data collected with non-human mammals on eyeblink classical conditioning. Subsequently, West et al. [24] reported selective loss of hippocampal pyramidal cells; the very cells in animals that fire in conjunction with the behavioral CR and unconditioned response (UR). The hypothesis that eyeblink conditioning would be severely impaired in patients with AD was confirmed [27,28], replicated [35], and independently replicated [20]. Using a criterion of producing 25% CRs in a session and combining data from several studies conducted during the last decade in our laboratory, eyeblink conditioning correctly identified 54 of 60 patients with probable AD [26]. In some cases eyeblink classical conditioning was effective in differentiating cerebrovascular dementia from probable AD [35]. In patients with Huntington’s disease [34] and Parkinson’s disease [7], eyeblink conditioning is relatively normal and clearly differentiated from eyeblink classical conditioning in AD. In addition to working with probable AD and other dementing neurological diseases of old age, we extended this work to adults with Down’s syndrome who inevitably develop AD-like neuropathology around the age of 35 years (called DS/AD). Patients with DS/AD perform eyeblink classical conditioning similarly to probable AD patients, but young adults with DS perform eyeblink classical conditioning in a manner comparable to normal older adults. That is, the eyeblink classical conditioning performance of younger adults with DS is not equal to the performance of normal young adults, but it is significantly better than adults with DS over the age of 35 years [36].
1.2. Predictions about eyeblink conditioning and Alzheimer’s disease based on an animal model It was demonstrated in animal studies that neuronal
unit activity in the hippocampus increased markedly within trials early in the eyeblink classical conditioning process. Activity recorded in the CA1 and CA3 regions of the hippocampus formed a predictive ‘model’ of the amplitude-time course of the learned behavioral response. Firing to the tone-CS and corneal airpuff-US occurred only under conditions when the stimuli were paired and produced behavioral learning [2]. When the CS and US were presented independently in the explicitly unpaired paradigm, there was no hippocampal response to the tones or airpuffs [5]. The hippocampal modeling of the behavioral CR and UR was generated largely by hippocampal pyramidal neurons in the CA1 and CA3 fields [4]. In neuropathological studies of human brains, West et al. [24] identified pyramidal cells in the CA1 field of hippocampus as the cells selectively lost in AD. The rabbit model system demonstrated that the hippocampus could play a modulatory role in conditioning [3]. Disruption or facilitation of hippocampal activity affected the rate of conditioning. In rabbits, disruption of muscarinic cholinergic receptors with scopolamine injections impaired acquisition of CRs, and this disruption occurred only when the hippocampus was intact [21]. Scopolamine also disrupted eyeblink conditioning in humans [1,19].
1.3. A role for nicotinic cholinergic receptors in eyeblink classical conditioning Artificial induction of behavioral deficits in the acquisition of CRs comparable to those observed with the muscarinic cholinergic antagonist, scopolamine have been reported using the nicotinic cholinergic antagonist, mecamylamine [29,32]. When mecamylamine was injected into young rabbits, acquisition was prolonged and resembled acquisition in older rabbits [32]. Mecamylamine also impairs non-verbal learning and memory in humans [12]. Nicotinic receptors are significantly reduced in cerebral cortex and hippocampal regions of the brain in AD [10,18]. In particular, the a4b2 nicotinic cholinergic receptor subtype is vulnerable in AD [23]. A drug in development for cognition-enhancement in AD, GTS21 is a selective agonist at the neuronal nicotinic a7 nicotinic cholinergic receptor subtype [8,9]. This drug improves acquisition of CRs in older rabbits so that they learn as well as young rabbits [33], and it reverses the effect of mecamylamine in the disruption of learning. Stimulation of remaining nicotinic receptors in the brains of patients with AD might alleviate some cognitive deficits due to cholinergic neuron dysfunction. Intravenous administration of nicotine to patients diagnosed as probable AD produced some improve-
D.S. Woodruff-Pak, I.S. Santos / Beha6ioural Brain Research 113 (2000) 11–19
ments in cognitive performance, but anxiety and other side effects were observed [13,14]. A novel nicotine derivative with highly selective binding to central nicotinic receptors, ABT-418 was effective in ameliorating memory deficits in six patients with moderate AD [16]. There is continuing interest in developing new brain-selective nicotinic agonists to inhibit the progressive loss of cognitive function in AD and other neurodegenerative diseases as evidenced by many of the articles in this Special Issue. A drug that acts by a dual mechanism of action that affects nicotinic cholinergic neurotransmission is galantamine. Galantamine is a tertiary alkaloid originally derived from the bulbs of the snowdrop and various Narcissus species [15]. It induces allosteric modulation of nicotinic acetylcholine receptors to increase acetylcholine release as well as acting as a reversible, competitive, selective inhibitor of acetylcholinesterase. Galantamine is a reversible acetylcholinesterase inhibitor that interacts selectively and competitively with the enzyme acetylcholinesterase [22]. Research also indicates that galantamine is an allosteric modulator at nicotinic cholinergic receptor sites [11,17]. It is proposed that galantamine increases cholinergic function in two ways: (a) increasing the concentration of acetylcholine through a competitive reversible inhibition of acetylcholine hydrolysis by the enzyme acetylcholinesterase; and (b) increasing the release of acetylcholine and other neurotransmitters such as glutamate through an allosteric modulation of acetylcholine effects at nicotinic cholinergic receptor sites. Because galantamine improved cognitive deficits in animal models involving cholinergic deficits, and because the cholinergic systems is of demonstrated involvement in eyeblink classical conditioning, we anticipated that galantamine would be effective in ameliorating conditioning deficits in older rabbits. Our initial aim was to identify a dose or doses of galantamine that showed measurable effects on the acquisition of CRs.
2. Materials and methods
2.1. Subjects A total of 40 female retired breeder specific pathogen free (SPF) New Zealand white rabbits completed this experiment. Birth dates of the rabbits were recorded by the breeder (Covance, Inc.). Rabbits ranged in age from 24 to 52 months, and the mean age was 29.2 months (S.D.= 6.6). A one-way analysis of variance comparing the various groups by age indicated no differences among the groups. Mean weight of the 40 rabbits was 4.5 kg (S.D.=0.58), and the range was 3.0 – 5.7 kg with no significant group differences in weight. Rabbits were individually housed in stainless steel cages in an
13
AAALAC-accredited animal facility. They were fed rabbit chow ad lib. and tap water and had a 12/12 h light/dark cycle.
2.2. Apparatus and procedure At least 24 h after arrival at the animal facility, rabbits were adapted twice in the restraining apparatus in sessions separated by 24 h. They were placed in Plexiglas restrainers for a 1-h adaptation session. After the second 1-h adaptation session, rabbits were given a local ophthalmic anesthetic (proparacaine hydrochloride) in the left eye so that a 6-0 nylon suture loop could be placed in the temporal margin of the nictitating membrane (NM). A patch of fur approximately 3 cm2 was shaved on the back, just below the head to expose the skin for subcutaneous (s.c.) injections. The classical conditioning equipment attached to the rabbit’s head included elastic eyelid retractors and a platform holding a minitorque potentiometer (San Diego Instruments prototype model) for NM movement measurement that was secured under the animal’s muzzle and behind the ears. The potentiometer was attached by a lever and a thread to the nylon suture loop in the NM. Analog output from the potentiometer was digitized and read into and IBM-PC-compatible system described by Chen and Steinmetz [6]. This system also controlled the timing and presentation of conditioning stimuli. For classical conditioning, the CS was an 850-ms, 85 dB, 1-kHz tone, followed 750 ms after its onset by a 100-ms, 3-psi corneal airpuff US. The CS and US coterminated. The inter-trial-interval was random, ranging between 20 and 30 s at 1-s intervals. One training session lasted about 45 min. Rabbits were tested four at a time. Each training session was controlled by a program written in C+ + language [6] and run on an IBM-PCcompatible 386 computer. Data were collected about the position of the NM in 3 ms bins during the trials. NM responses were displayed on the monitor, and the amplitudes, areas, and latencies of the response relative to baseline were calculated. In addition, averages were calculated for every block of nine trials. A CR was scored if the NM was pulled back a minimum of 0.5 mm in the interval between 25 and 750 ms after CS onset. The dependent measure, learning criterion, was scored as the number of training trials it took the animal to produce eight CRs within nine consecutive trials. Data were collected in RAM and saved to a hard drive, and individual data summaries for each of the four rabbits run simultaneously were printed at the end of each session. Rabbits in the explicitly unpaired condition were treated in a fashion identical to rabbits tested in the paired condition with the exception that the 850-ms, 85
14
D.S. Woodruff-Pak, I.S. Santos / Beha6ioural Brain Research 113 (2000) 11–19
dB, 1-kHz tone CS and 100-ms, 3-psi corneal air puff US were never paired. In the unpaired condition, rabbits received at total of 90 stimuli, 45 tone CSs and corneal air puff 45 USs. Each stimulus was presented at an inter-trial-interval that was random and ranged between 20 and 30 s. Thus, the duration of the session and the inter-trial interval was identical in the paired and unpaired sessions, and all rabbits were tested four at a time.
2.4. Research design There were five groups of rabbits in the paired condition treated with 0 (vehicle), 1, 2, 3, and 4 mg/kg galantamine. The vehicle group in the paired condition had eight rabbits, and there were four rabbits tested at each dose of galantamine. In the unpaired groups there were eight rabbits tested with 3 mg/kg galantamine and eight rabbits tested with vehicle.
2.3. Drugs 3. Results Galantamine was purchased from Tocris Cookson, Inc. (Ballwin, MO). The vehicle solution was sterile saline. Galantamine was freshly prepared in solution each week and injected s.c. a minimum of 15 min before behavioral testing to ensure that peak blood levels of galantamine were attained during the testing session. Rabbits were tested between 15 and 30 min after injections, and behavioral testing was completed a maximum of 1 h and 15 min after injection. Equal volumes were injected into the animals, including vehicle-treated animals, based on weight. In this experiment we report on the behavioral results from ten daily injections of various doses of galantamine or vehicle.
Fig. 1. The number of paired trials of tone conditioned stimulus and corneal air puff unconditioned stimulus in the 750-ms delay eyeblink classical conditioning paradigm it took retired breeder rabbits to attain a learning criterion of eight conditioned responses in nine consecutive trials. There were four rabbits treated at each dose of galantamine, and eight rabbits treated with sterile saline vehicle. The difference between rabbits treated with the 3.0-mg/kg dose or vehicle was significant at the 0.01 level of confidence (**PB 0.01). Error bars are standard deviation.
The major aim of this experiment was to identify a dose of galantamine that was most effective in ameliorating the learning deficits in the 750 ms delay eyeblink classical conditioning procedure in older rabbits. A one-way analysis of variance (ANOVA) was used to compare the trials to learning criterion attained by the 24 rabbits in the paired condition treated with doses of 0.0, 1.0, 2.0, 3.0, or 4.0 mg/kg galantamine. The effect of dose was statistically significant, F (4, 19)=3.19; PB 0.05. Post hoc comparisons using the Dunnett Control Group Comparison test indicated that the difference in trials to criterion between rabbits treated with the 3.0 mg/kg dose of galantamine and rabbits treated with sterile saline vehicle was significant at the 0.01 level of confidence (Fig. 1). Among the 24 older rabbits treated with doses ranging from 0.0 to 4.0 mg/kg galantamine, six failed to achieve learning criterion and were assigned a score of 1400 (the 1350 paired trials they experienced plus 50, which is still likely an underestimate of the number of trials to criterion for these animals). There were four rabbits in the vehicle-treated group that did not achieve criterion, and one rabbit each in the 1.0- and 4.0-mg/kg galantamine groups. All rabbits in the groups treated with 2.0 and 3.0 mg/kg attained learning criterion within 1350 trials. To examine the effect of various doses of galantamine on eyeblink classical conditioning over training sessions, a five (dose of 0.0, 1.0, 2.0, 3.0, or 4.0 mg/kg galantamine) by ten (training sessions) repeated measures ANOVA was carried out on the dependent measure of percentage of CRs. The percentage of CRs was different depending on the dose of galantamine, F (4, 19)= 6.33, PB 0.01. Post hoc comparisons using the Dunnett Control Group Comparison test indicated that rabbits treated with a dose of 3 mg/kg galantamine produced significantly more CRs than did vehicletreated rabbits (PB 0.01; Fig. 2). The effect of training sessions was also significant, with an increase in the percentage of CRs over sessions, F (9, 171)= 14.48, PB 0.0001. The interaction effect indicating that dose level affected the change in percentage of CRs over training session was also significant, F (36, 171)=1.93,
D.S. Woodruff-Pak, I.S. Santos / Beha6ioural Brain Research 113 (2000) 11–19
15
Fig. 2. Left: Percentage of conditioned responses (CRs) for each of ten daily sessions of paired tone conditioned stimulus and corneal air puff unconditioned stimulus in the 750-ms delay eyeblink classical conditioning paradigm for retired breeder rabbits treated with daily doses of 0, 1.0, 2.0, 3.0, or 4.0 mg/kg galantamine. Right: Mean of the percentage of CRs over ten training sessions for the data shown on the left. Error bars are standard deviation. *P B 0.05; **PB 0.01; ***PB0.001.
P B0.01. Post hoc analysis of the significant dose by training session interaction effect indicated that there were significant differences between groups in Sessions 3, 4, 6, 7, 9, 10 as shown in Fig. 2. With the exception of Session 9 in which the significance level exceeded 0.05, the differences exceeded the 0.01 level of confidence. A five (dose of 0.0, 1.0, 2.0, 3.0, or 4.0 mg/kg galantamine) by ten (training sessions) repeated measures ANOVA was carried out on the dependent measure of CR amplitude. The main effects of dose and session were statistically significant, but the dose by session interaction effect did not attain significance at the 0.05 level of confidence. CR amplitude was different depending on the dose of galantamine, F (4, 19) = 5.76, PB 0.01. The effect of training sessions was also significant, with an increase in CR amplitude over sessions, F (9, 171) =2.48, PB 0.01 (Fig. 3). The magnitude of the motor or unconditioned response was not affected by drug dose or training. A five (dose of 0.0, 1.0, 2.0, 3.0, or 4.0 mg/kg galantamine) by ten (training sessions) repeated measures ANOVA was carried out on the dependent measure of UR amplitude, but neither the effect of dose (Fig. 3) nor the training session or dose by training session interaction effects were statistically significant. To control for the possibility that galantamine affected the blink response to the tone CS or to the corneal air puff US, we tested the 3.0-mg/kg dose of
galantamine in the explicitly unpaired condition in which rabbits received randomly presented CS and US stimuli identical to the stimuli experienced by the rabbits in the paired condition. A two (dose of 0.0 or 3.0 mg/kg galantamine) by ten (training session) repeated measures ANOVA comparing percentage of blink responses in the CS period revealed no significant effects for dose, training session, or dose by training session interaction. A two (dose of 0.0 or 3.0 mg/kg galantamine) by ten (training session) repeated measures ANOVA comparing the amplitude of the blink response in the CS period also revealed no significant effects (Fig. 4). A two (dose of 0.0 or 3.0 mg/kg galantamine) by ten (training session) repeated measures ANOVA comparing the amplitude of the UR revealed no significant dose or dose by training session interaction effect. The effect of training session was significant, and as can be seen in Fig. 4, this effect was a trend to increased UR amplitude in later sessions by both groups.
4. Discussion Testing five groups of rabbits at four doses of galantamine and vehicle, we were impressed by the magnitude and consistency of differences in eyeblink classical conditioning produced by the optimal dose of the drug, 3.0 mg/kg. Trials to learning criterion, a measure that is
16
D.S. Woodruff-Pak, I.S. Santos / Beha6ioural Brain Research 113 (2000) 11–19
larger when learning is poorer, revealed a classical U-shaped response curve with doses of 1.0 and 2.0 mg/kg galantamine producing non-significant effects in comparison to rabbits treated with vehicle, a dose of 3.0 mg/kg galantamine reducing the number of trials to learning criterion to a very low mean significantly dif-
ferent from vehicle-treated rabbits, and 4.0 mg/kg galantamine producing a non-significant effect. Total percentage of CRs over the ten training sessions, a measure that is higher with better performance, produced an inverted-U shaped response curve. Galantamine doses of 1.0 and 2.0 mg/kg were under-doses,
Fig. 3. Amplitude of the conditioned response for each of ten daily sessions of paired tone conditioned stimulus and corneal air puff unconditioned stimulus in the 750-ms delay eyeblink classical conditioning paradigm for retired breeder rabbits treated with daily doses of 0, 1.0, 2.0, 3.0, or 4.0 mg/kg galantamine. Right: Mean of the percentage of unconditioned responses over ten training sessions for the same rabbits whose data are shown on the left. Error bars are standard deviation.
Fig. 4. Left: Amplitude of the UR for each of ten daily sessions of 90 explicitly unpaired tones or corneal air puffs for retired breeder rabbits treated with daily doses of 0 or 3.0 mg/kg galantamine. Right: Mean of the percentage of blink responses to a tone over ten training sessions for the same rabbits whose data are shown on the left.
D.S. Woodruff-Pak, I.S. Santos / Beha6ioural Brain Research 113 (2000) 11–19
the dose of 3.0 mg/kg was optimal, and the dose of 4.0 mg/kg was an over-dose for the acquisition of CRs in the 750-ms eyeblink classical conditioning procedure. Older rabbits treated with 3 mg/kg galantamine produced a high percentage of CRs early in training and attained a high performance level of over 90% CRs by Session 10 typically seen only in young rabbits. In our laboratory for a period covering almost a decade, hundreds of rabbits of various ages have been treated with drugs or vehicle and tested with the 750-ms delay eyeblink classical conditioning procedure. Young adult rabbits treated with vehicle attain learning criterion in around 400 trials, and retired breeder rabbits treated with vehicle attain learning criterion in around 1000 trials, with standard deviations of 200 – 300 trials. Older rabbits receiving daily injections of a dose of 3.0 mg/kg galantamine achieved learning criterion in a mean of 233.0 trials (S.D.=176.9). The identification of a dose level of galantamine that enables older rabbits to acquire CRs at a rate that is approaching half the number of trials to criterion required by young rabbits is a striking result. The high percentage of CRs maintained by older rabbits treated with 3 mg/kg galantamine is also exceptional by our laboratory standards for older rabbits. In independent studies using the eyeblink classical paradigm with the 750-ms delay procedure in older rabbits, we have tested the effects of a nicotinic agonist (GTS-21) and the effects of cholinesterase inhibitors (physostigmine, donepezil). For each drug, we tested the effect of at least two doses on learning in comparison to sterile saline vehicle. There were statistically significant effects on trials to learning criterion and percentage of CRs in all three of these drugs. However, the magnitude of the effects of the most optimal dose of GTS-21, physostigmine, or donepezil were not as large as the effect of a dose of 3.0 mg/kg galantamine observed in the present experiment. Trials to learning criterion in rabbits treated with 3.0 mg/kg galantamine were less than half the trials to learning criterion in rabbits treated with either a nicotinic agonist or either one of the cholinesterase inhibitors. Because galantamine induces allosteric modulation of nicotinic cholinergic receptors to increase acetylcholine release as well as acting as an acetylcholinesterase inhibitor, it may ameliorate learning impairment in older rabbits more effectively than drugs with one mechanism of action. Of course, additional tests with a larger sample size are required to confirm these initial results with galantamine. The result that 3.0 mg/kg galantamine caused a reduction in the number of trials to achieve learning criterion and increased the percentage of CRs did not occur because galantamine affected blink rate or blink amplitude. In the paired condition, UR amplitude was not different among the five groups of rabbits treated
17
with different doses of galantamine or vehicle. The learned response (the CR), but not the motor response (the UR) was affected by galantamine at a dose of 3.0 mg/kg. The explicitly unpaired control procedure testing rabbits’ response to independently presented tone and air puff stimuli demonstrated no group differences between rabbits treated with 3.0 mg/kg galantamine or vehicle. The percentage of eyeblinks during the CS period, blink amplitude during the CS period, and the amplitude of the eyeblink to the air puff were similar in galantamine- or vehicle-treated animals. The effect of a dose of 3.0 mg/kg galantamine appeared to be on mechanisms of associative learning. The one statistically significant effect in the analyses of data in the explicitly unpaired conditioning sessions was a significant effect of Session. Over sessions there was a tendency for UR amplitude to increase in both the group treated with vehicle and in the group treated with 3.0 mg/kg galantamine. However, the increase in UR amplitude was not accompanied by an increase in blink amplitude during the CS period. Blink amplitude during the CS period remained stable through the ten training sessions. Rabbits tested in the explicitly unpaired condition showed some increase in the magnitude of their reflexive blink to the US, but they did not change their rate of blinking as indexed by the percentage of blinks during the CS period. They did not increase the magnitude of the blink during the CS period, either. In addition, this trend of a change in UR amplitude in the unpaired condition was consistent across the galantamine- and vehicle-treated groups. There was no non-associative learning effect selectively present in galantamine-treated rabbits. The present experiment with the aim of identifying an optimal dose of galantamine for testing in the eyeblink classical conditioning paradigm is a first step in a program of research to identify the means by which galantamine affects associative learning and memory. One of the advantages of using eyeblink classical conditioning in pharmacological testing is that the neural circuitry and sites of plasticity for learning are described. The cerebellum is the essential brain structure for the acquisition and retention of CRs, and it is possible that galantamine acts in the cerebellum of older rabbits to ameliorate learning. However, experiments have demonstrated that the disruptive effect of scopolamine on eyeblink conditioning [21] and the ameliorating effect of nefiracetam on eyeblink conditioning [31] occur only when the hippocampus is intact. The hippocampus, with its higher concentration of nicotinic cholinergic receptors may be the site where the dual mechanisms of galantamine are effective in improving the rate of learning. Galantamine induces allosteric modulation of nicotinic cholinergic receptors to increase acetylcholine release as well as acting as an acetylcholinesterase inhibitor. Using this dual mecha-
18
D.S. Woodruff-Pak, I.S. Santos / Beha6ioural Brain Research 113 (2000) 11–19
nism of action, galantamine appears to ameliorate learning impairment in older rabbits more effectively than drugs with one mechanism of action.
[13]
Acknowledgements
[14]
A number of individuals assisted at various stages of this project. We would like to thank Tara Orlando for assistance with data collection, Richard Vogel for assistance with data entry, and Michelle Pak for assistance at the data collection, entry, and analyses phases. Dan Lyons, Manager of the Central Animal Facility at Albert Einstein Healthcare Network facilitated our work in many ways. This research was supported by a grant from Janssen Pharmaceutica, Inc.
[15]
[16]
[17]
[18]
References [19] [1] Bahro M, Schreurs BG, Sunderland T, Molchan SE. The effects of scopolamine, lorazepam, and glycopyrrolate on classical conditioning of the human eyeblink response. Psychopharmacology 1995;122:395 – 400. [2] Berger TW, Alger B, Thompson RF. Neuronal substrate of classical conditioning in the hippocampus. Science 1976;192:483 – 5. [3] Berger TW, Berry SD, Thompson RF. Role of the hippocampus in classical conditioning of aversive and appetitive behaviors. In: Isaacson RL, Pribram KH, editors. The Hippocampus. New York: Plenum, 1986:203–39. [4] Berger TW, Rinaldi PC, Weisz DJ, Thompson RF. Single-unit analysis of different hippocampal cell types during classical conditioning of rabbit nictitating membrane response. J Neurophysiol 1983;50:1197–219. [5] Berger TW, Thompson RF. Neuronal plasticity in the limbic system during classical conditioning of the rabbit nictitating membrane response. I. The hippocampus. Brain Res 1978;145:323 – 46. [6] Chen G, Steinmetz JE. A general-purpose computer system for behavioral conditioning and neural recording experiments. Behav Res Meth Instruments Computers 1998;30:384–91. [7] Daum I, Schugens MM, Breitenstein C, Topka H, Spieker S. Classical eyeblink conditioning in Parkinson’s disease. Movement Disorders 1996;11:639–46. [8] De Fiebre CM, Meyer EM, Henry JC, Muraskin SI, Kem WR, Papke RL. Characterization of a series of anabaseine-derived compounds reveals that the 3-(4)-dimethylaminocinnamylidine derivative is a selective agonist at neuronal nicotinic alpha7/125alpha-bungarotoxin receptor subtypes. Mol Pharmacol 1995;47:164 – 71. [9] Kem WR, Mahnir VM, Papke RL, Lingle CJ. Anabaseine is a potent agonist on muscle and neuronal alpha-bungarotoxin-sensitive nicotinic receptors. Pharmacol Exp Ther 1997;283:979 – 92. [10] London ED, Ball MJ, Waller SB. Nicotinic binding sites in cerebral cortex and hippocampus in Alzheimer’s dementia. Neurochem Res 1989;14:745–50. [11] Maelicke A, Schrattenholz A, Periera EFR, Albuquerque EX. Nicotinic receptors in the central nervous system: Implications for Alzheimer’s disease. Abstract no. 12: 6th International Conference on Alzheimer’s Disease and Related Disorders, 18 – 23 July 1998, Amsterdam. [12] Newhouse PA, Potter A, Corwin J, Lenox R. Age-related effects
[20]
[21]
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
[30]
of the nicotinic antagonist mecamylamine on cognition and behavior. Neuropsychopharmacology 1994;10:93 – 107. Newhouse PA, Potter A, Lenox R. The effects of nicotinic agents on human cognition: possible therapeutic applications in Alzheimer’s and Parkinson’s diseases. Med Chem Res 1993;2:628 – 42. Newhouse P, Sunderland T, Tariot P. Intravenous nicotine in AD. Psychopharmacology 1988;95:171 – 5. Parys W. Development of Reminyl (galantamine), a novel acetylcholinesterase inhibitor, for the treatment of Alzheimer’s disease. Alzheimer’s Rep 1998;1:S19 – 20. Potter A, Corwin J, Lang J, Piasecki M, Lenox R, Newhouse PA. Acute effects of the selective cholinergic channel activator (nicotinic agonist) ABT-418 in Alzheimer’s disease. Psychopharmacology 1999;142:334 – 42. Schrattenholz A, Pereira EFR, Roth U, et al. Agonist responses of neuronal nicotinic acetylcholine receptors are potentiated by a novel class of allosterically acting ligands. Mol Pharmacol 1996;49:1 – 6. Schroder G, Giacobini E, Struble RG, Zilles K, Maelicke A. Nicotinic cholinergic receptive neurons of the frontal cortex are reduced in Alzheimer’s disease. Neurobiol Aging 1991;12:259– 62. Solomon PR, Groccia-Ellison M, Flynn D, Mirak J, Edwards KR, Dunehew A, Stanton ME. Disruption of human eyeblink conditioning after central cholinergic blockade with scopolamine. Behav Neurosci 1993;107:271 – 9. Solomon PR, Levine E, Bein T, Pendlebury WW. Disruption of classical conditioning in patients with Alzheimer’s disease. Neurobiol Aging 1991;12:283 – 7. Solomon PR, Solomon SD, Vander Schaaf E, Perry HE. Altered activity in the hippocampus is more detrimental to classical conditioning than removing the structure. Science 1983;220:329– 31. Thomsen T, Kewitz H. Selective inhibition of human acetylcholinesterase by galantamine in vitro and in vivo. Life Sci 1990;46:1553 – 8. Warpman U, Nordberg A. Epibatidine and ABT 418 reveal selective losses of alpha 4 beta 2 nicotinic receptors in Alzheimer brains. NeuroReport 1995;6:2419 – 23. West MJ, Coleman PD, Flood DG, Troncoso JC. Differences in the pattern of hippocampal neuronal loss in normal aging and Alzheimer’s disease. Lancet 1994;344:769 – 72. Woodruff-Pak DS. Eyeblink classical conditioning in normal aging and Alzheimer’s disease. In: Woodruff-Pak DS, Steinmetz JE, editors. Eyeblink Classical Conditioning: Volume I, Applications in Humans. Amsterdam: Kluwer, 2000:163 – 189. Woodruff-Pak DS. Insights about learning in Alzheimer’s disease from the animal model. In: Carroll ME, Overmier JB, editors. Linking Animal Research and Human Psychological Health. Washington, DC: American Psychological Association, in press. Woodruff-Pak DS, Finkbiner RG, Katz IR. A model system demonstrating parallels in animal and human aging: extension to Alzheimer’s disease. In: Meyer EM, Simpkins JW, Yamamoto J, editors. Novel Approaches to the Treatment of Alzheimer’s Disease. New York: Plenum, 1989:355 – 71. Woodruff-Pak DS, Finkbiner RG, Sasse DK. Eyeblink conditioning discriminates Alzheimer’s patients from non-demented aged. NeuroReport 1990;1:45 – 8. Woodruff-Pak DS, Hinchliffe RM. Scopolamine- or mecamylamine-induced learning impairment: reversed by nefiracetam. Psychopharmacology 1997;131:130 – 9. Woodruff-Pak DS, Jaeger ME. Predictors of eyeblink classical conditioning over the adult age span. Psychol Aging 1998;13:193 – 205.
D.S. Woodruff-Pak, I.S. Santos / Beha6ioural Brain Research 113 (2000) 11–19 [31] Woodruff-Pak DS, Li Y-T, Hinchliffe RM, Port RL. Hippocampus in delay eyeblink classical conditioning: essential for nefiracetam amelioration of learning in older rabbits. Brain Res 1997;747:207 – 18. [32] Woodruff-Pak DS, Li Y-T, Kazmi A, Kem WR. Nicotinic cholinergic system involvement in eyeblink classical conditioning in rabbits. Behav Neurosci 1994;108:486–93. [33] Woodruff-Pak DS, Li Y-T, Kem WR. A nicotinic agonist (GTS-21), eyeblink classical conditioning, and nicotinic receptor binding in rabbit brain. Brain Res 1994;645:309–17. [34] Woodruff-Pak DS, Papka M. Huntington’s disease and eye-
.
19
blink classical conditioning: normal learning but abnormal timing. J Int Neuropsychol Soc 1996;2:323 – 34. [35] Woodruff-Pak DS, Papka M, Romano S, Li Y-T. Eyeblink classical conditioning in Alzheimer’s disease and cerebrovascular dementia. Neurobiol Aging 1996;17:505 – 12. [36] Woodruff-Pak DS, Papka M, Simon EW. Eyeblink classical conditioning in Down’s Syndrome, Fragile X Syndrome, and normal adults over and under age 35. Neuropsychology 1994;8:14 – 24. [37] Woodruff-Pak DS, Trojanowski JQ. The older rabbit as an animal model: implications for Alzheimer’s disease. Neurobiol Aging 1996;17:283 – 90.