Download Effects of ()-HA-966, CGS-19755, Phencyclidine, and Dizocilpine on

0022-3565/99/2891-0521$03.00/0
THE JOURNAL OF PHARMACOLOGY AND EXPERIMENTAL THERAPEUTICS
Copyright © 1999 by The American Society for Pharmacology and Experimental Therapeutics
JPET 289:521–527, 1999
Vol. 289, No. 1
Printed in U.S.A.
Effects of (1)-HA-966, CGS-19755, Phencyclidine, and
Dizocilpine on Repeated Acquisition of Response Chains in
Pigeons: Systemic Manipulation of Central Glycine Sites1,2
CORY M. CAMPBELL,3 EDUARDO R. BUTELMAN,4 and JAMES H. WOODS
Departments of Pharmacology (C.M.C., E.R.B., J.H.W.) and Psychology (J.H.W.), University of Michigan, Ann Arbor, Michigan
Accepted for publication December 6, 1998
This paper is available online at http://www.jpet.org
Activation of the N-methyl-D-aspartate (NMDA) subtype of
glutamate receptors requires the presence of glycine at a site
on the receptor that is distinct from the site at which NMDA
and glutamate bind (Corsi et al., 1996). The actions of glycine
at this site are not blocked by strychnine, which makes them
distinct from the inhibitory actions of glycine on glycinesensitive chloride channels (Vannier and Triller, 1997).
There are several similarities between compounds such as
(1)-3-amino-1-hydroxy-2-pyrrolidine (HA-966) and 7-chlorokynurenic acid (7CKA), which block binding of glycine at
its site on this glutamate receptor; compounds such as cis-4phosphonomethyl-2-piperidine carboxylic acid (CGS-19755)
and DL-(E)-2-amino-4-methyl-5-phospono-3-pentenoic acid,
which are competitive antagonists at the glutamate binding
site; and noncompetitive antagonists such as phencyclidine
and dizocilpine, which block the ion channel of the NMDA
Received for publication April 9, 1998.
1
This research was supported by U.S. Public Health Service Grant
DA05325 and National Research Service Award Fellowship DA032049
(C.M.C.).
2
A preliminary report of these findings was presented to the EBPS meeting
in Berlin, Germany, October 1994.
3
Present address: Case Western Reserve Medical School, Cleveland, OH
48109-0632.
4
Present address: Rockefeller University, Box 171, 1230 York Avenue, New
York, NY 10021.
pairment produced by a large dose of (1)-HA-966 (100 mg/kg)
were completely prevented by coadministration of the glycinesite agonist D-serine (560 mg/kg) but not by its enantiomer,
L-serine (1000 mg/kg). D-Serine, however, produced incomplete antagonism of the effects of dizocilpine and phencyclidine
and failed to alter those of CGS-19755. These findings provide
evidence that reducing the activity of the NMDA subtype of the
glutamate receptor through pharmacological action at any of
three sites produces similar decrements in acquisition, and
those produced through antagonism of the glycine site are
differentially sensitive to the glycine-site agonist D-serine.
receptor. All, for example, have protective effects against the
effects of central hypoxia and ischemia (McDonald et al.,
1989; McNamara and Dingledine, 1990; Priestley et al., 1990;
Wood et al., 1992), and all have anticonvulsant effects
(Croucher and Bradford, 1990, 1991; Singh et al., 1990; Meldrum, 1994). The fact that the anticonvulsant effects of (1)HA-966 and 7CKA are mediated through the glycine site was
demonstrated by the ability of i.c.v. administration of the
glycine-site agonist D-serine to prevent the anticonvulsant
effects of these antagonists but not those of an antagonist at
the glutamate site (Lu, 1994). A third common aspect is that
both the glutamate site and the glycine site on the NMDA
receptor have been implicated in learning and long-term
memory. Morris et al. (1986) found that i.c.v. administration
of aminophosphonovaleric acid, a competitive antagonist at
the NMDA site, selectively impaired the ability of the rats to
acquire spatial information and prevented the induction of
hippocampal long-term potentiation, which may be associated with synaptic plasticity relevant to acquisition of new
behavior. The competitive glutamate-site antagonists DL-(E)2-amino-4-methyl-5-phospono-3-pentenoic acid and DL-(E)-2amino-4-methyl-5-phospono-3-pentenoic acid carboxy-ethylester have also been reported to cause decrements in the
performance of rats in a radial arm maze (Butelman, 1989;
Bischoff and Tiedtke, 1992). In operant repeated acquisition
ABBREVIATIONS: CGS-19755, cis-4-phosphonomethyl-2-piperidine carboxylic acid; 7CKA, 7-chlorokynurenic acid; HA-966, 3-amino-1-hydroxy-2-pyrrolidine; NMDA, N-methyl-D-aspartate.
521
Downloaded from jpet.aspetjournals.org at ASPET Journals on May 3, 2017
ABSTRACT
The effects of i.m. injections of (1)-HA-966, a glycine-site antagonist at the N-methyl-D-aspartate (NMDA) subtype of the
glutamate receptor, its enantiomer (2)-HA-966, the competitive
glutamate antagonist CGS-19755, the uncompetitive glutamate
antagonists phencyclidine and dizocilpine, and the m opioid
agonist morphine were evaluated in a repeated acquisition task
in pigeons. All of the drugs produced dose-dependent decreases in rates of responding. The NMDA receptor and channel blockers and (1)-HA-966 appeared to have a greater effect
on acquisition than did morphine at doses that did not fully
suppress responding. The rate suppression and learning im-
522
Campbell et al.
Materials and Methods
Subjects. Eight experimentally naive White Carneau pigeons
(Palmetto, Sumter, SC), maintained at approximately 80% of their
free-feeding weight, constituted the study population. The pigeons
were individually housed in cages, with water and grit freely available, in a colony room with a 12-h light/dark schedule (lights on at
7:00 a.m.). All animals had previously received phencyclidine-like
compounds. Animals in these studies were maintained in accordance
with the University Committee on the Use and Care of Animals,
University of Michigan and “Guidelines of the Committee on the
Care and Use of Laboratory Animals” of the Institute of Laboratory
Animal Resources, National Health Council (Department of Health,
Education and Welfare, publication no. NIH 85-23, revised 1983).
Apparatus. Experiments were conducted in ventilated, soundattenuating chambers measuring 36 3 28 3 33 cm. Operant conditioning boxes were located within each chamber and had three translucent response keys, 2.4 cm in diameter, located on the middle of
one wall, 25 cm from the floor. The keys were 5 cm apart and could
be transilluminated by blue, red, green, or amber (red and green
lights simultaneously illuminated) 7-W lights located behind the
wall. Mixed grain was made available by means of a hopper that
could be raised to a position below the center response key and 10 cm
from the floor of the chamber. During food availability, the hopper
was illuminated by a 7-W white light. Execution of the experiments
and data collection were accomplished using an IBM model 70 PC
and Med-State software (Med Associates, Inc., East Fairfield, VT).
Repeated Acquisition Procedure. Pigeons were trained to acquire a four-link response chain under a second-order FR5 schedule
(Thompson, 1973). The terminal schedule was one in which all three
keys were illuminated red at the beginning of a trial. The pigeon
could respond on any of the three keys but only one was designated
“correct.” A response on that key turned all three keys green. A peck
on one of the two “incorrect” red keys resulted in a 4-s timeout during
which the chamber was dark and key pecks had no programmed
consequences. After the timeout, the three lights were again illuminated red; the same key was designated correct, and a response on
this key turned all the keys green. When all keys were green, a peck
on the one correct green key turned all the keys amber. A peck on one
of the two incorrect keys darkened the chamber for 4 s, and then the
three green keys were again illuminated. In the presence of amber
key lights, a correct response turned all the keys blue and a correct
response in the presence of the blue key lights produced a 0.5-s flash
of the hopper light and restarted the cycle in red. On the fifth time
through the cycle (FR5), a correct response in the presence of the
blue light resulted in 4-s access to mixed grain. The sequence of light
colors was constant (red, green, amber, blue) each day. However, the
key that was designated correct in the presence of each light color
was changed from day to day, and the response-sequence made by
the pigeon had to be altered each day and match the newly programmed key sequence to produce food.
Training the pigeons on this schedule was initiated by illuminating the center key blue and presenting food as a consequence of a
peck on this key. Once the key peck response had been made and food
presented, all three keys were illuminated blue, and the pigeon was
required to peck the correct key to produce access to grain. Incorrect
selection resulted in a 4-s timeout followed by reillumination of the
blue key lights. Once the pigeon had pecked the correct blue key, the
training was continued by illuminating all keys amber and reinforcing correct amber key selection by turning the keys blue. The key
that had been correct during previous exposure to the blue lights was
still correct, and a response on this key resulted in grain presentation. The training continued to step backward in the key color presentation until a correct response in the presence of the red key
lights turned all keys green, a correct response in the presence of the
green key lights turned all the keys amber, and a correct response in
the presence of the amber key lights turned all the keys blue. A
correct response in the presence of the blue key light resulted in
grain presentation during the initial steps of training; the number of
times the pigeon was required to complete the four-color links in the
chain was gradually increased to five. Incorrect responses in any key
color produced a 4-s timeout and the subsequent reillumination of
three keys in the color in which the error had been made.
Sessions were typically run 6 days per week. Each session began
with a blackout period lasting approximately 5 min. During this
blackout period, key pecks had no programmed consequence. Sessions terminated after pigeons received 60 reinforcers or 60 min had
elapsed. Correct key locations for each link color were changed on a
daily basis according to a schedule of three sets of six series (18
distinct orders). The series were selected to be equivalent in several
ways with restrictions in their ordering across sessions (Thompson,
1973). An example of a set of six chains is LRCR, CLRL, LRLC,
RCRL, CLCR, and RCLC, where L is left, R is right, and C is center.
For all chains, the first response requirement was in the presence of
red stimulus lights, the second was in the presence of green stimulus
lights, the third was in the presence of amber stimulus lights, and
the last was in the presence of blue stimulus lights.
Pigeons were trained until a steady state of acquisition was
reached. This required that animals consistently receive at least 54
(90%) of the available 60 food reinforcers while responding more
rapidly than 0.5 responses/s, with a maximum of 33% errors during
the entire session. (Errors were designated as the number of incorrect key pecks divided by the total number of key pecks multiplied by
100 and are referred to as percent errors.) When an animal achieved
this level of acquisition, baseline data were collected for 8 to 10 days,
and mean values were calculated for the percent errors and response
rate.
Drug evaluation was initiated when criteria were met for 5 consecutive days. The error criterion was that the animal make no more
or fewer errors than 15% of its baseline error number. To meet the
response rate criterion, the number 0.18 was added and subtracted
from the session’s response rate, and the established baseline rate
had to fall within these values. The 0.18 value is 15% of the typical
high rate in this procedure of 1.2 responses/s. After each drug test,
criterion-level performance had to be reached for 2 consecutive days
before subsequent testing.
Design. Four or five pigeons were tested with each drug and with
saline. All drugs were administered in a single i.m. injection in a
Downloaded from jpet.aspetjournals.org at ASPET Journals on May 3, 2017
procedures, the uncompetitive glutamate antagonists phencyclidine and dizocilpine impaired acquisition in rats (Cohn
and Cory-Slechta, 1992), monkeys (Moerschbaecher et al.,
1985; France et al., 1991), and pigeons (Thompson and Moerschbaecher, 1982). Recent reports have identified learning
deficits produced by glycine-site antagonists. Intracoronary
administration of 7CKA blocked the ability of chicks to learn
passive avoidance (Steele and Stewart, 1993). This same
glycine-site antagonist impaired the working memory of rats
in a three- runway task, and this impairment was prevented
by the intrahippocampal administration of D-serine (Ohno et
al., 1994). The present experiments were designed to compare the glycine-site antagonist (1)-HA-966 with the glutamate-site competitive antagonist CGS-19755 and the uncompetitive antagonists dizocilpine and phencyclidine with
respect to their ability to modify complex behavior in pigeons
after parenteral (i.m.) administration. A repeated acquisition
task was used, similar to those used in previous studies, to
evaluate drug interactions with learning and memory. In
addition, the ability of parenteral administration of the glycine-site agonist D-serine to reverse the deficits produced by
each of these antagonists was evaluated to determine
whether their similar effects on acquisition and performance
were mediated through a similar or different site.
Vol. 289
1999
Glycine-Site Antagonist and Agonist on Repeated Acquisition
523
Results
Pigeons typically reached criterion-level performance of
the repeated acquisition procedure after 6 to 8 months of
training. Values of baseline percent errors and response
rates were very similar across animals. In baseline sessions,
most errors were made as the first few reinforcers were
earned (i.e., reinforcers 1–3, or 5–15 correct sequences).
Comparison of NMDA Antagonists. A detailed illustration of the effects of the glycine-site NMDA antagonist (1)HA-966 on accuracy and rate of responding can be observed
over the first 28 reinforcers earned within a session by a
single animal (Fig. 1). Although each dose caused this pigeon
to make more errors compared with saline control (Fig. 1,
top), only 56 mg/kg was associated with a large reduction in
response rate (Fig. 1, bottom). Dose of both 10 and 32 mg/kg
had modest response rate-reducing effects that were restricted to the early portion of the session (Fig. 1, bottom).
After the administration of 10 mg/kg (1)-HA-966, error accumulation declined by the 28th reinforcer (Fig. 1, top). This
is evident by the fact that only two errors were made in
earning the 24th through 28th reinforcer, indicating that the
sequence was probably acquired at this point in the session.
In contrast, the animal made 23 errors in earning the 24th
through 28th reinforcer after the administration of 32 mg/kg
(1)-HA-966. Similarly, no reduction in error accumulation
was observed after the administration of 56 mg/kg (1)-HA966; the animal stopped responding after making 174 errors,
earning six reinforcers. Thus, no evidence of acquisition of
the sequence was observed after the administration of 32 or
56 mg/kg (1)-HA-966. Although this animal’s rate of responding was different with each dose of (1)-HA-966 for the
first six reinforcers, cumulative errors per reinforcer for the
first six reinforcements were quite similar across doses.
Data representing the effects of the enantiomers of HA-966
Fig. 1. Effects of (1)-HA-966 (10 –56 mg/kg) on error accumulation within
a session: a record from a single animal (pigeon 8607). The first 28
reinforcers are shown. Ordinate, cumulative number of errors (top) and
response rates in responses per second per reinforcement (bottom). Abscissa, number of reinforcers. Individual points are represented for the
first through the sixth reinforcer (before the break) and are represented
in groups of two reinforcers thereafter.
on percent errors and response rates for each of five birds and
their average in the repeated acquisition program are shown
in Fig. 2. Both enantiomers produced dose-dependent increases in percent errors and decreases in rates of responding. The (2)-enantiomer was more potent in decreasing rates
of responding (Fig. 2, right bottom) and in increasing errors
(Fig 2, right top). There were substantial individual differences among the pigeons with respect to their response to the
HA-966 enantiomers, but there was a general association
between decreased rates of responding and increased percent
errors for the individual animals for both enantiomers.
In the repeated acquisition schedule, each of the three
NMDA antagonists, dizocilpine, phencyclidine, and
CGS-19755, as well as the m opioid morphine, produced dosedependent increases in percent errors and reduced response
rates (Fig. 3). Morphine, however, produced fewer errors at
rate-decreasing doses, suggesting less effect of this drug on
acquisition behavior. At doses that produced a 50% decrease
in rates of responding, dizocilpine produced 62% errors,
phencyclidine produced 39% errors, and CGS-19755 produced 50% errors. Morphine, in contrast, at a dose that
caused a 50% decrease in response rates, produced only 27%
errors.
Coadministration of D-Serine With NMDA Antagonists. A dose of 100 mg/kg (1)-HA-966 produced marked
Downloaded from jpet.aspetjournals.org at ASPET Journals on May 3, 2017
volume of 1 ml/kg b.wt. Dizocilpine (18 –180 mg/kg) was given 20 min,
phencyclidine (0.56 –1.8 mg/kg) was given 10 min, CGS-19755 (0.32–
3.2 mg/kg) was given 90 min, morphine (0.32–10 mg/kg) was given 10
min, and (1)-HA-966 (10 –100 mg/kg) and its enantiomer (2)-HA966 (3.2–17.8 mg/kg), as well as D-serine (10 –560 mg/kg) and its
enantiomer L-serine (1000 mg/kg), were given 30 min before the start
of the corresponding test session. The dose of L-serine (1000 mg/kg)
and the highest dose of D-serine (560 mg/kg) used were chosen from
pilot studies that indicated that these were the largest behaviorally
inactive doses in these procedures.
Data. The effects of drugs on individual animals’ rates of responding and percent errors are presented for each condition. In addition,
these values were averaged and are presented as mean 6 S.E.M.
Rates of responding were individually calculated as the total number
of responses per time in the session, and percent errors represent the
number of errors divided by the total number of responses multiplied
by 100 for one session. When no reinforcers were obtained in a test
session, no percent error value is reported for the individual animal.
Drugs. Dizocilpine maleate (MK-801; Merck, Sharp and Dohme,
West Point, PA); morphine sulfate (Mallinckrodt, Inc., St. Louis,
MO); phencyclidine, (1)-HA 966, and its enantiomer (2)-HA 966
(National Institute on Drug Abuse); and D-serine and its enantiomer
L-serine (Sigma Chemical Co., St. Louis, MO) were dissolved in
physiological saline. Doses of dizocilpine and morphine are expressed
in their salt forms. CGS-19755 (Novartis Corp., Summit, NJ) was
dissolved in physiological saline and a minimum quantity of 0.1 N
NaOH. The solution was titrated to pH 6 to 8 with small quantities
of lactic acid.
524
Campbell et al.
Vol. 289
decreases in rates of responding, and marked increases in
percent errors, as shown in Fig. 2. Administration of increasing doses of D-serine produced dose-dependent reversals in
the percent errors and response rate reductions caused by
administration of 100 mg/kg (1)-HA-966 (Fig. 4, left). At the
largest dose of D-serine (560 mg/kg) coadministered with 100
mg/kg (1)-HA-966, the mean percent errors (Fig. 4, left top)
and response rates (Fig. 4, left bottom) were not substantially
different from saline control values. This dose of D-serine
administered alone had no effect on percent errors but
slightly reduced response rates. Unlike D-serine, L-serine did
not completely reverse the effects of (1)-HA-966 on percent
errors and response rates in the repeated acquisition procedure (Fig. 4, right). Although percent errors after the coadministration of 1000 mg/kg L-serine with 100 mg/kg (1)-HA966 was moderately lower than with 100 mg/kg (1)-HA-966
alone, this value was still much greater than saline control
(Fig. 4, right top). In addition, L-serine only slightly reversed
the profound reduction in response rates caused by 100
mg/kg (1)-HA-966 (Fig. 4, right bottom). Thus, the reversal
by D-serine of the behavioral effects caused by (1)-HA-966
was stereoselective.
A detailed illustration of the reversal by D-serine of the
effects caused by 100 mg/kg (1)-HA-966 on accuracy and rate
of responding in the repeated acquisition procedure is shown
over the first 28 reinforcers averaged over all five animals
(Fig. 5). The largest doses of D-serine (560 mg/kg) restored
both percent errors and response rates to those observed
after saline administration. The next smallest dose of Dserine (320 mg/kg) also returned response rates to predrug
baseline levels (Fig. 5, bottom) but did not greatly modify the
rate of error accumulation compared with administration of
100 mg/kg (1)-HA-966 alone (Fig. 5, top).
Administration of a range of doses of D-serine along with a
dose of dizocilpine (180 mg/kg) that produced marked increases in percent errors and decreases in response rates
Fig. 3. Effects of dizocilpine (18 –180 mg/kg, top left), phencyclidine (0.56 –
1.8 mg/kg, bottom right), CGS-19755 (0.32–3.2 mg/kg, bottom left), and
morphine (0.32–10 mg/kg, bottom right) on percentage of errors (top for
each treatment) and response rates (bottom for each treatment) in the
repeated acquisition procedure. The large circles represent the mean 6
S.E.M. from five animals, and the smaller symbols represent the individual pigeon data. Some individual data points are hidden within the larger
average symbols. These symbols indicate data from the same animals
shown in Fig 2. The points above Sal indicate percent errors and response
rates after saline administration.
resulted in the partial attenuation of the dizocilpine-induced
impairments (Fig 6). The largest dose of D-serine (560 mg/kg)
was most effective in this regard (Fig. 6, left). These effects,
although modest and with considerable interanimal variability, were reliably observed within animals over three determinations per animal. The two smaller doses of D-serine (100,
320 mg/kg) had very little effect on the dizocilpine-induced
impairments. Coadministration of 1000 mg/kg L-serine with
180 mg/kg dizocilpine did not modify the effects of dizocilpine
(Fig. 6, right). D-Serine was less effective in reversing deficits
produced by another NMDA channel blocker, phencyclidine
(1.8 mg/kg) in the repeated acquisition procedure (Fig. 7, left)
and entirely ineffective against comparable deficits caused by
the competitive NMDA antagonist CGS-19755 (3.2 mg/kg;
Fig. 7, right).
Discussion
Intramuscular administration of (1)-HA-966, an antagonist at the glycine binding site of the NMDA subtype of the
Downloaded from jpet.aspetjournals.org at ASPET Journals on May 3, 2017
Fig. 2. Effects of (1)-HA-966 (10 –100 mg/kg, left) and (2)-HA-966 (3.2–
17.8 mg/kg, right) on percentage of errors and response rates in the
repeated acquisition procedure. Ordinate, percent errors (top) and response rates (bottom). Abscissae, dose of drug (mg/kg). Larger open
circles represent mean values (6 S.E.M.). Smaller symbols are singleanimal data and represent an animal’s behavior during an entire session.
The points above Sal indicate percentage of errors and response rates
after saline administration.
1999
Glycine-Site Antagonist and Agonist on Repeated Acquisition
525
glutamate receptor, caused a dose-dependent reduction in
accuracy of responding and rates of responding in a repeated
acquisition task in pigeons. A similar decrement was produced by dizocilpine and phencyclidine, drugs that block the
NMDA channel, and by CGS-19755, a competitive antagonist
at the glutamate binding site. Although a reduction in response rates was often observed at doses that caused an
impairment in acquisition, rate-decreasing effects were not
always accompanied by acquisition impairments. Only the
largest doses of (1)-HA-966, CGS-19755, phencyclidine, and
dizocilpine reduced response rates below the criterion level of
0.5 responses/s. Intermediate doses of each NMDA antagonist did not profoundly affect rates of responding, yet each
increased percent errors to greater-than-maximum criterion
levels of 33%. The (2)-enantiomer of HA-966, which is relatively less active as a glycine-site antagonist (Pullan et al.,
1990; Singh et al., 1990), was 0.5 to 0.75 log unit more potent
in suppressing rates of responding, and nearly as effective in
blocking acquisition of behavior at doses just below those
that stopped all responding. Morphine, which has no clear
interaction with NMDA receptor neurotransmission, produced considerably less decrement in acquisition at doses
that profoundly suppressed rates, a result similar to that
found by Thompson and Moerschbaecher (1981).
These data suggest that blockade of ion transport through
the NMDA channel by pharmacological action at any of these
three relevant binding sites can produce a very similar reduction of acquisition of learning in a relatively selective
manner. Similar conclusions were reached by Xu et al. (1995)
in studies of the anterograde amnesia, fear conditioning, and
anticonvulsant effects of (1)-HA-966 and dizocilpine. These
data also provide further support for the notion that activation of the NMDA subtype of glutamate receptors is an important aspect of at least some types of learning.
It is unlikely that (1)-HA-966 was acting at a site on the
NMDA receptor other than the glycine site because systemic
administration of D-serine, an agonist at this glycine site,
Fig. 5. Effects of D-serine (320 or 560 mg/kg), administered alone or with
100 mg/kg (1)-HA-966, on error accumulation within a session. Numbers
directly above the data points indicate the number of animals contributing to the data after the administration of 100 mg/kg (1)-HA-966
(squares) when this number was less than five. These animals did not
earn more than five reinforcers in the 60-min session due to low rates of
responding. Diamonds represent data when 560 mg/kg D-serine was
administered alone, and open circles indicate rates and error accumulation after saline administration.
Downloaded from jpet.aspetjournals.org at ASPET Journals on May 3, 2017
Fig. 4. Effects of D-serine (56 –560 mg/kg,
left) and L-serine (1000 mg/kg, right) coadministered with 100 mg/kg (1)-HA-966 on
percent errors and response rates in the repeated acquisition procedure. For L-serine,
the data represent the mean 6 S.E.M. from
five animals. For D-serine, the large circles
represent the mean 6 S.E.M. from five animals, and the smaller symbols represent
the individual pigeon data. These symbols
indicate data from the same animals shown
in Fig 2. Some individual data points are
hidden within the larger average symbols.
The points above Sal indicate percentage of
errors and response rates after saline administration. The points above Ser indicate
percentage of errors and response rates after administration of 560 mg/kg D-serine
alone. The points above (1)-HA indicate
percentage of errors and response rates after the administration of 100 mg/kg (1)-HA966 given alone. Ordinate, percentage of errors (top) and response rates (bottom).
Abscissae, treatment condition.
526
Campbell et al.
Vol. 289
Fig. 7. The effects of D-serine (56 –560
mg/kg) coadministered with 1.8 mg/kg
PCP (left) and 3.2 mg/kg CGS-19755
(right) on percentage of errors and response rates in the repeated acquisition procedure. The larger symbols indicate the mean 6 S.E.M. for five
animals, and the smaller symbols indicate individual data points. These
symbols represent data from the same
animals shown in Fig. 2. Some individual data points are hidden within
the larger average symbols. The
points above Sal indicate percentage
of errors (top) and response rates (bottom) after saline administration. The
points above Ser indicate the percent
errors and response rates after the administration of 560 mg/kg D-serine
alone. The points above Phen (left) indicate the effects of 1.8 mg/kg phencyclidine given alone. The points
above CGS (right) indicate the effects
of 3.2 mg/kg CGS-19755 given alone.
produced a dose-dependent reversal of both the rate and
learning impairments caused by (1)-HA-966. At the largest
dose of D-serine, the reversal of (1)-HA-966-induced impairments was complete. When coadministered with an effective
dose of the uncompetitive antagonist dizocilpine, the largest
dose of D-serine returned both acquisition and rate of responding to approximately 50% of control values. Smaller
doses had little effect. Modest improvements in phencyclidine-impaired responding were observed but were not related
to D-serine dose, and D-serine had no effect on impairments
produced by CGS-19755. This indicates a specific interaction
of (1)-HA-966 on the glycine site to produce the decrements
in acquisition reported here.
Other investigators have found an interaction between
D-serine and glutamate channel blockers. Tanii et al. (1994)
reported that i.c.v. administration of D-serine produced a
dose-dependent but incomplete reversal of hyperactivity and
ataxia induced by phencyclidine in the rat; this reversal
could be prevented by administration of 7CKA. This suggests
that our observation of D-serine-induced partial antagonism
of the antiacquisition effects of phencyclidine and dizocilpine
might be a general phenomenon. The mechanism of this
Downloaded from jpet.aspetjournals.org at ASPET Journals on May 3, 2017
Fig. 6. Effects of D-serine (100 –560 mg/
kg, left) and L-serine (1000 mg/kg, right)
coadministered with 0.18 mg/kg dizocilpine on percentage of errors and response rates in the repeated acquisition
procedure. The large symbols represent
the mean 6 S.E.M. for five animals, and
the smaller symbols represent individual
animal data. These symbols represent
data from the same animals shown in Fig
2. When fewer than five smaller symbols
are shown, the data are hidden by the
larger average symbol. Single-animal
data points for 560 mg/kg D-serine coadministered with 0.18 mg/kg dizocilpine
represent the average of three sessions.
The points above Sal indicate percentage
of errors (top) and response rates (bottom)
after saline administration. The points
above Ser indicate the percent errors and
response rates after the administration of
560 mg/kg D-serine alone. The points
above Dizo indicate percent errors and
response rates after the administration of
180 mg/kg dizocilpine alone.
1999
Glycine-Site Antagonist and Agonist on Repeated Acquisition
References
Bischoff C and Tiedtke PI (1992) Competitive and non-competitive NMDA receptor
antagonists in spatial learning tasks. Eur J Pharmacol 213:269 –273.
Cohn J and Cory-Slechta DA (1992) Differential effects of MK-801, NMDA and
scopolamine on rats learning a four-member repeated acquisition paradigm. Behav
Pharmacol 3:403– 413.
Corsi M, Fina P and Trist DG (1996) Co-agonism in drug-receptor interactions:
Illustrated by the NMDA receptors. Trends Pharmacol Sci 17:220 –222.
Croucher MJ and Bradford HF (1990) 7-Chlorokynurenic acid, a strychnineinsensitive glycine receptor antagonist, inhibits limbic seizure kindling. Neurosci
Lett 118:29 –32.
Croucher MJ and Bradford HF (1991)The influence of strychnine-insensitive glycine
receptor agonists and antagonists on generalized seizure thresholds. Brain Res
543:91–96.
France CP, Moerschbaecher JM and Woods JH (1991) MK-801 and related compounds in monkeys: Discriminative stimulus effects and effects on a conditional
discrimination. J Pharmacol Exp Ther 257:727–734.
Huettner JE (1989) Indole-2-carboxylic acid: A competitive antagonist of potentiation by glycine at the NMDA receptor. Science (Wash DC) 243:1611–1613.
Johnson JW and Ascher P (1987) Glycine potentiates the NMDA response in cultured mouse brain neurons. Nature (London) 324:529 –531.
Kushner L, Lerma J, Zukin RS and Bennett MVL (1988) Coexpression of N-methylD-aspartate and phencyclidine receptors in Xenopus oocytes injected with rat brain
mRNA. Proc Natl Acad Sci USA 85:3250 –3254.
Lerma J, Zukin RS and Bennett MVL (1990) Glycine decreases desensitization of
N-methyl-D-aspartate (NMDA) receptors expressed in Xenopus oocytes and is
required for NMDA responses. Proc Natl Acad Sci USA 87:2354 –2358.
Lu Y (1994) In vivo pharmacological characterization of glycine site ligands of the
NMDA receptor complex. Unpublished PhD thesis dissertation, University of
Michigan.
Mayer ML, Vycklicky F and Clements J (1989) Regulation of NMDA receptor desensitization in mouse hippocampal neurones by glycine. Nature (London) 338:425–
427.
McDonald JW, Uckele J, Silverstein FS and Johnston MV (1989) HA-966 (1-hydroxy3-aminopyrrolidone-2) selectively reduces N-methyl-D -aspartate (NMDA)mediated brain damage. Neurosci Lett 104:167–170.
McNamara D and Dingledine R (1990) Dual effect of glycine on NMDA-induced
neurotoxicity in rat cortical cultures. J Neurosci 10:3970 –3976.
Meldrum BS (1994) The role of glutamate in epilepsy and other CNS disorders.
Neurology 44:S14 –S23.
Moerschbaecher JM, Thompson DM and Winsauer PJ (1985) Effects of opioids and
phencyclidine in combination with naltrexone on the acquisition and performance
of response sequences in monkeys. Pharmacol Biochem Behav 22:1061–1069.
Morris RGM, Anderson E, Lynch GS and Baudry M (1986) Selective impairment of
learning and blockade of long-term potentiation by an N-methyl-D-aspartate receptor antagonist, AP5. Nature (London) 319:774 –776.
Ohno M, Yamamoto T and Watanabe S (1994) Intrahippocampal administration of a
glycine site antagonist impairs working memory performance of rats. Eur J Pharmacol 253:183–187.
Parsons CG, Zong X and Lux HD (1993) Whole cell and single channel analysis of the
kinetics of glycine-sensitive N-methyl-D-aspartate receptor desensitization. Br J
Pharmacol 109:213–221.
Priestley T, Horne AL, McKernan RM and Kemp JA (1990) The effect of NMDA
receptor glycine site antagonists on hypoxia-induced neurodegeneration of rat
cortical cell cultures. Brain Res 531:183–188.
Pullan LM, Britt M, Chapdelaine MJ, Keith RA, Lamonte D, Mangano TJ, Patel J,
Powel RJ, Stumpo RJ, Warwick PJ, Zinkand WC and Salama AI (1990) Stereoselectivity for the (R)-enantiomer of HA-966 (1-hydroxy-3-aminopyrrolidone-2) at the
glycine site of the N-methyl-D-aspartate receptor complex. J Neurochem 55:1346 –
1351.
Singh L, Donald AE, Foster AC, Hutson PH, Iversen LL, Iversen SD, Kemp JA,
Leeson PD, Marshall GR, Oles RJ, Priestley T, Thorn, L, Tricklebank MD, Vass
CA and Williams BJ (1990) Enantiomers of HA-966 (3-amino-1-hyroxypyrrolid-2one) exhibit distinct central nervous system effects: (1)-HA-966 is a selective
glycine/N-methyl-D-aspartate receptor antagonist, but (2)-HA-966 is a potent
gamma-butyrolactone-like sedative. Proc Natl Acad Sci USA 87:347–351.
Steele RJ and Stewart MG (1993) 7-Chlorokynurenate, an antagonist of the glycine
binding site on the NMDA receptor, inhibits memory formation in day-old chicks
(Gallus domesticus). Behav Neurobiol 60:89 –92.
Tanii Y, Nishikawa T, Hashimoto A and Takahashi K (1994) Stereoselective antagonism by enantiomers of alanine and serine of phencyclidine-induced hyperactivity, stereotypy and ataxia in the rat. J Pharmacol Exp Ther 269:1040 –1048.
Thompson DM (1973) Repeated acquisition as a behavioral baseline for studying
drug effects. J Pharmacol Exp Ther 184:506 –514.
Thompson DM and Moerschbaecher JM (1981) Selective antagonism of the errorincreasing effect of morphine by naloxone in a repeated-acquisition task. J Exp
Anal Behav 36:371–380.
Thompson DM and Moerschbaecher JM (1982) Phencyclidine in combination with
pentobarbital: Supra-additive effects on complex operant behavior in pigeons.
Pharmacol Biochem Behav 17:353–357.
Vannier C and Triller A (1997) Biology of the postsynaptic glycine receptor. Inter Rev
Cytol 176:201–244.
Wood ER, Bussey TJ and Phillips AG (1992) A glycine antagonist reduces ischemiainduced CA1 cell loss in vivo. Neurosci Lett 1:10 –14.
Xu X, Klinger P and Davis R (1995) Comparative anterograde amnestic and anticonvulsant effects of two types of NMDA receptor antagonists: MK-801 and HA966. Psychopharmacology 117:333–339.
Send reprint requests to: James H. Woods, Ph.D., Department of Pharmacology, University of Michigan, 1301 Medical Science Research Building III,
Ann Arbor, MI 48109-0632. E-mail [email protected]
Downloaded from jpet.aspetjournals.org at ASPET Journals on May 3, 2017
effect is unclear but might involve D-serine-induced increases
in the number of NMDA channels that are operational, even
in the presence of the channel blockers.
Although all of the NMDA antagonists produced similar
impairments in the repeated acquisition procedure, their differential sensitivity to D-serine provides evidence that the
learning impairments observed were produced by activity at
different ligand-binding sites. Similarly selective pharmacological interactions have been reported previously after the
central administration of D-serine. For example, i.c.v. administration of D-serine reversed the inhibition of NMDA-induced convulsions caused by i.c.v. administration of 7CKA
and i.c.v. or i.p. administration of (1)-HA-966 in mice, but
was ineffective against the anticonvulsant effects of CGS19755 (Lu, 1994). In pigeons, the response-rate decreasing
effects of i.c.v. administration of (1)-HA-966 or 7CKA, but
not i.c.v. administration of CGS-19755, were also reversed by
i.c.v. administration of D-serine (Lu, 1994).
Pharmacological characterization of the electrophysiological (Johnson and Ascher, 1987; Kushner et al., 1988; Huettner, 1989; Vycklicky et al., 1990) and in vitro binding (Mayer
et al., 1989; Lerma et al., 1990; Parsons et al., 1993) characteristics of glycine-site antagonists has demonstrated that
these effects are selectively mediated at glycine sites associated with the NMDA receptor. Pharmacological characterization of glycine-site antagonists has not been well established
using behavioral measures. In addition, systemic bioavailability of glycine-site antagonists has not been previously
demonstrated using subconvulsive doses. The present findings provide robust evidence that central glycine sites can be
pharmacologically manipulated by systemic administration
of a glycine-site agonist and antagonist and that the effects of
blocking the NMDA receptor with a glycine-site antagonist
are similar to those produced by blocking this receptor at
either of two other sites.
527