Introduction early training phase e.g. see Ref. [8]. Somewhat con-
Brain Research 887 2000 144–156 www.elsevier.com locate bres
Research report
Intra-cerebellar infusion of NMDA receptor antagonist AP5 disrupts classical eyeblink conditioning in rabbits
a a,b ,
Gengxin Chen , Joseph E. Steinmetz
a
Program in Neural Science , Indiana University, Bloomington, IN 47405-7007, USA
b
Department of Psychology , Indiana University, 1101 E. 10th Street, Bloomington, IN 47405-7007, USA
Accepted 19 September 2000
Abstract
Rabbits were infused with AP5, an NMDA receptor antagonist, into the region of the cerebellar interpositus nucleus during classical eyeblink conditioning with a tone conditioned stimulus and an air puff unconditioned stimulus. Acquisition of the conditioned eyeblink
response was delayed in rabbits infused with AP5 but the NMDA receptor antagonist had little effect on conditioned responses when these same rabbits were infused a second time after reaching asymptotic responding levels. Some rabbits that received AP5 infusions for the
first time after the conditioned response was well learned showed temporary alterations in response timing. These data indicate that NMDA receptor activity is involved in the acquisition of classically conditioned eyeblink response and may also be involved in regulating
cellular processes involved in response timing and other aspects of conditioned response execution.
2000 Elsevier Science B.V. All
rights reserved.
Theme : Neural basis of behavior
Topic : Learning and memory: systems and functions
Keywords : Plasticity; Learning and memory; Interpositus nucleus; Cerebellar cortex; Response timing
1. Introduction early training phase e.g. see Ref. [8]. Somewhat con-
sistent, Caramanos and Shapiro [10] found that systemic It has been well established that the NMDA receptor
injections of MK-801, another NMDA antagonist, and plays a critical role in the induction of long-term potentia-
intracerebral ventricular infusions of APV impaired spatial tion LTP for review see Refs. [5,16,35]. Administration
working memory or reference memory of the radial maze of selective NMDA receptor blockers, such as APV,
tasks depending on the drug dose, familiarity with the prevents the induction of LTP but appears to have little
environment, and training procedure. effect on normal synaptic transmission [20,22] but see
Other studies have also suggested that NMDA receptors Refs. [19,44]. In an attempt to associate spatial learning
are involved in a variety of learning and memory tasks. with LTP, Morris et al. [34] infused APV into the rat
Intra-amygdala infusion of APV blocked the acquisition hippocampus and found that the same drug dose and
and consolidation, but not expression of, auditory con- infusion protocol that blocked LTP in vivo also caused
ditioned fear-potentiated startle in rats [9,33]. Similarly, impairment of spatial learning in the Morris water-maze.
Kim et al. [24] found that ventricular infusions of APV However, their finding is complicated by the observation
completely blocked acquisition, but not expression, of that APV may have caused sensorimotor disturbance in the
Pavlovian fear conditioning, while the same dose of APV appeared to have no effect on pain sensitivity. Further-
more, it has been shown that infusion of APV into the basolateral amygdala also blocked acquisition of con-
Corresponding author. Tel.: 11-812-855-3991; fax: 11-812-855-
textual Pavlovian fear conditioning [17]. Mathis et al. [30]
4691. E-mail address
: steinmetindiana.edu J.E. Steinmetz.
showed that post-training ventricular infusion of AP5 and
0006-8993 00 – see front matter
2000 Elsevier Science B.V. All rights reserved. P I I : S 0 0 0 6 - 8 9 9 3 0 0 0 3 0 0 5 - 5
G . Chen, J.E. Steinmetz Brain Research 887 2000 144 –156
145
another competitive NMDA antagonist, gamma-LGLA, part of the memory trace for eyeblink conditioning is
impaired the retention of the temporal component but not localized within the cerebellum.
the spatial discrimination component of a Y-maze active It is generally agreed that neurons in the cerebellar deep
avoidance task with mice. The involvement of NMDA nuclei express NMDA receptors [1,3,4,45] but whether or
receptors has also been demonstrated for step-through not Purkinje cells express NMDA receptors is somewhat
inhibitory avoidance tasks [26,29], step-down inhibitory controversial. Expression analysis by in situ hybridization
learning [23], discriminative approach response learning has revealed that Purkinje cells express NMDA receptor
[7], and taste-potentiated odor conditioning [21]. subunit NR1 but may not express, or express in relatively
It has been demonstrated that systemic administration of low levels, the NR2 family subunits [1,45]. However,
the non-competitive NMDA antagonist, MK-801 or PCP, pharmacological and electrophysiological studies failed to
and the competitive NMDA antagonist, CGP-39551, im- observe functional evidence of NMDA receptors in Pur-
paired acquisition of classical eyeblink conditioning in kinje cells [3,4]. These observations may be reconciled if
rabbits [37,42] and rats [39]. In all cases, the NMDA one considers that the functional properties of the
antagonists appeared to have no effect on retention and heteromeric NMDA receptor-channel complex are critical-
performance of previously acquired eyeblink conditioned ly determined by the constituting NR2 subunits.
responses, nor did they appear to affect sensory reactivity In cerebellum, there are cell type-specific expressions of
or unconditioned responses. While the above studies did NMDA receptor subunits [1,45]. In adult rats or mice, the
not ascertain the target of NMDA antagonists, one study cerebellar granule cells express subunit NR2A and, more
showed that memantine, a non-competitive NMDA re- abundantly, NR2C. Deep nuclei express subunit NR2A but
ceptor antagonist with higher affinity to cerebellar tissue little NR2C. And, Purkinje cells are thought to express
than forebrain tissue, also impaired eyeblink conditioning neither of those subunits. This information combined with
in humans [38], suggesting the involvement of cerebellar the gene knockout technique provide some clues of the
NMDA receptors. Interestingly, in Robinson’s study [37], functions of NMDA receptors in specific cells. Kishimoto
MK-801 appeared to have no effect on conditioning- et al. [27] showed that mutant mice lacking NMDA
related potentiation of perforant path-granule cell re- receptor subunit NR2A acquired eyeblink conditioning
sponses in hippocampus, suggesting its effective target more slowly than wild-type animals but could attain the
might be elsewhere. same asymptotic performance as the wild-type. In contrast,
A number of permanent and reversible lesion studies mutants lacking subunit NR2C did not exhibit significant
have demonstrated that the cerebellum, especially the impairment. This evidence suggests an important role of
interpositus nucleus, is critical for classical eyeblink NMDA receptors in the deep nuclei in the acquisition of
conditioning [11,15,28,36,41]. For example, Krupa et al. eyeblink conditioning.
[28] showed that infusion of the GABA agonist muscimol This study was an attempt to assess the role of cerebellar
A
into the region of the interpositus nucleus abolished NMDA receptors in classical eyeblink conditioning. As in
conditioned responses that were established before infu- a number of previous studies, we microinjected AP5
sion. More importantly, muscimol infusion during acquisi- solution into the brain region of interest, i.e. the cere-
tion training prevented the formation of conditioned re- bellum, and examined its effect on the acquisition and
sponses and no evidence of learning i.e. savings could be performance of conditioned responses. In brief, our data
discerned when training was instituted after the infusion. A show that AP5 impaired conditioned response acquisition,
recent study by Bracha et al. [6] provided additional and in some animals, had a temporary effect on con-
evidence that cellular mechanisms in the cerebellum are ditioned response timing once learning had occurred. Some
important for learning the eyeblink conditioned response of these results have been presented elsewhere in prelimin-
CR. Microinjections of anisomycin, a protein synthesis ary form [14].
inhibitor, into the intermediate cerebellum near the inter- positus nuclei impaired the acquisition of the conditioning
and appeared to have no effect on the expression of CRs in