This document contains text automatically extracted from a PDF or image file.

Formatting may have

been lost and not all text may have been recognized.
To remove this note, right-click and select "Delete table".
Neurobiology of Disease 5, 474–482 (1998)
Article No. NB980223
Role of Cannabinoid Receptors in Memory Storage
Robert E. Hampson and SamA. Deadwyler
Department of Physiology & Pharmacology and Center for the Neurobiological
Investigation of Drug Abuse, Wake Forest University School of Medicine,
Winston–Salem, North Carolina 27157-1083
Studies have shown that 9-tetrahydrocannabinol, the principal psychoactive ingredient in marijuana,
produces memory deficits similar to those produced by neurochemical lesions of the hippocampus.
Such lesions impair performance in short-term spatial memory tasks learned prior to the lesion.
Animals trained in the behavioral task following the lesion can still perform the task, but learn a
different behavioral strategy. Cannabinoid agonists impair behavioral performance in a delay-
dependent manner similar to that produced by lesions, but also shift the behavioral response strategy.
A possible role for cannabinoid receptors and endogenous cannabinoids may thus be to regulate the
storage (i.e., encoding) of information, aswell as themeans bywhichthat information is retrieved.
1998
Academic Press
The role of cannabinoids in memory processes can
be traced to early observations in humans that signifi-
cant disruption of short-term recall of events was the
most consistent disruptive effect in experimental tests
of marijuana intoxication (1). More recent investiga-
tions have confirmed that the memory disruptive
effect of 9-tetrahydrocannabinol (THC) in humans is
also accompanied by several other perceptual and
disorientation effects (2,3). Early studies on the effects
of high concentrations of 9-THC (i.e., 10 mg/kg) in
rodents showed profoundly impaired maze learning
and operant behavior during and after chronic THC
exposure (4–6). Initial studies of low doses of 9-THC
on hippocampal memory processes showed subtle but
consistent dose-dependent impairment (7,8). In subse-
quent years the laboratory has compiled considerable
data on the nature of these effects in many different
behavioral paradigms. The major overriding finding is
that effects on hippocampal-mediated processes are
consistent with respect to dose range and severity of
disruption (9). One of the most intriguing findings in
this regard has been the observation that cannabinoids
disrupt processing of sensory information within hip-
pocampal circuits (10,11). This disruption occurs pri-
marily in the dentate gyrus, where high densities of
cannabinoid receptors exist (12,13). Local application
of the potent cannabinoid agonist CP55,940 to the
hippocampus of rats has shown that the memory
impairment produced by cannabinoids can be sepa-
474
rated from collateral motor effects. Intrahippocampal
injection of CP55,940 produced a dose-dependent im-
pairment of maze performance, with no change in the
time to complete the task (14). Application of the
cannabinoid agonist directly to cells of the hippocam-
pus, impaired memory independent of other systemic
effects of the drug (15).
The cellular basis for these disruptions is consistent
with reports from this and other laboratories regarding
alterations in ionic currents and intracellular mecha-
nisms that would inhibit neural transmission within
the hippocampal formation (16–20). For example, it
was shown that cannabinoids inhibited cyclic-AMP
accumulation within neurons (21,22). Cannabinoid
receptor involvement was identified (23–25) and was
confirmed to be coupled to adenylate cyclase by an
inhibitory G-protein (26,27). It has subsequently been
demonstrated that cannabinoid receptor activation
results in inhibition of voltage-activated calcium cur-
rents (28,29) and enhancement of potassium currents
(16,30). Both of these effects would reduce the excitabil-
ity of hippocampal neurons, and henceneural transmis-
sion. In addition, cannabinoid agonists have been
shown to interfere with the mechanism of long-term
potentiation (LTP), a candidate mechanism for learn-
ing and memory (31–34). These results, coupled with
recent evidence that cannabinoids presynaptically al-
ter release of GABA (35) and glutamate (20) from
hippocampal neurons, points to a potentially critical
0969-9961/98 $25.00
Copyright 1998 by Academic Press
All rights of reproduction in any form reserved.
Role of Cannabinoid Receptors in Memory 475
role for the cannabinoid receptor within this critical
perirhinal and entorhinal cortex (Fig. 1), are also
substrate for memory processing.
shown. Lower densities of CB1 receptors are also
located in each of the cell layers of the hippocampus
(25). Receptors in these areas could potentially play a
THE IMPORTANCE OF THE
HIPPOCAMPUS IN MEMORY
role in differentially modifying information by affect-
ing only the entorhinal–dentate gyrus–CA3–CA1 pro-
jection, and not the entorhinal–CA1 projection, or by
The role of the mammalian hippocampus in the
processing of memory has been elaborated over many
years with reports showing memory deficits in hu-
mans following damage to the medial temporal lobe
and hippocampus (36–39). Early studies in rats showed
that lesions of hippocampus impaired performance in
delayed memory tasks (40–43). However; later studies
demonstrated that many of these deficits resulted from
retrohippocampal damage, and performance was not
impaired when the lesions were confined to the den-
tate gyrus and CA1/CA3 cell fields via precise injec-
tions of the neurotoxin ibotenic acid (44–46). In fact, it
appears in many cases that memory tasks such as
delayed-nonmatch-to-sample (DNMS) do not require
the hippocampus unless they involve complex associa-
tions or spatial memory (47–49). Lesions of retrohippo-
campal areas including pre- and parasubiculum, ento-
rhinal cortex, and perirhinal cortex will impair short-
Copyright 1998 by Academic Press
All rights of reproduction in any form reserved.
modifying the hippocampal outputs via the subicu-
lum. Also, since there is evidence that these neural
pathways are involved in successful short-term
memory, it is likely that the inhibitory effects of CB1
receptor activation on hippocampal neural activity
during performance of the DNMS task will be similar
to the effects of hippocampal lesions. Recent observa-
tions have in fact shown that successful DNMS behav-
ioral performance can be attributed to the intensity of
encoding of trial-relevant information by hippocampal
CA3 and CA1 neurons (61–63), and that cannabinoids
disrupt this encoding phase (64,65). Thus, short-term
memory likely represents a process involving not only
the hippocampal cells, but also reciprocal projections
between entorhinal cortex and hippocampus. Based on
the distribution of CB1 receptors in this region, canna-
binoids likely play an important part in modifying the
processing of short-term memory.
term memory in nonspatial tasks, but whether this
results merely from interruption of hippocampal pro-
jections is unclear. In rodents, it is clear that restricted
hippocampal lesions impair spatially based short-term
memory (50–52). In other types of memory such as
COMPARISON OF CANNABINOID
ADMINISTRATION AND HIPPOCAMPAL
LESION EFFECTS ON DNMS
paired association, however, it is not clear whether
PERFORMANCE
hippocampal lesions impair the memory, or the ability
to represent the associations between stimuli (48,53–
Delay-dependent performance in the delayed-match-
55). It is apparent that the hippocampus is essential to
to-sample task was previously reported to be impaired
representing relationships between stimuli, and the
following exposure to 9-THC, the psychoactive ingre-
projections between hippocampusand retrohippocam-
dient in marihuana (11). The same is true for animals
pal areas are essential to the memory storage of these
performing the delayed-nonmatch-to-sample task
representations(49,56–58). Theseprojections, in particu-
(64,65). Figure 2 shows that administration of THC (2.0
lar the reciprocal connections between entorhinal (and
mg/kg, ip) produced a highly significant decrease
perirhinal cortices) and hippocampus proper (Fig. 1),
(F
(5,257)
are critical to the successful performance of spatial
DNMS tasks in rats.
It is essential to note that cannabinoid receptors
(CB1) are found in sufficient densities along this circuit
to suggest an important role in modifying the normal
flow of information into and out of the hippocampus.
Figure 1 diagrams the intrinsic and extrinsic connec-
tions of the hippocampus and retrohippocampal areas
(59,60) and the cell fields affected by the ibotenate
lesion. The greatest densities of CB1 receptors which
are found predominately in the molecular layer of the
dentate gyrus, as well as the superficial layers of
18.11, P 0.001) at all delays 5 s. Although
performance in the 1- to 5-s epoch was slightly reduced
(F
(5,257)
2.42, P 0.05), there wasno significant differ-
ence in performance at delays between 0 and 3 s. This
delay-dependent deficit could be blocked by pretreat-
ment of the same animals with the competitive canna-
binoid receptor antagonist SR141716A. DNMS perfor-
mance was assessed for 15 days following controlled
ibotenate lesions of hippocampus. Animals with total
loss of hippocampal tissue but minimal encroachment
on surrounding limbic structures showed delay-
dependent deficits in the DNMS task that approached
chance levels at the longest delays (Fig. 2). Mean
476 Hampson and Deadwyler
FIG. 1. Diagram of hippocampal and retrohippocampal neural circuitry required for short-term memory. A‘‘slice’’ of
hippocampus, subiculum
(SUB), and entorhinal cortex (EC) is shown to illustrate the neural projections between the various regions and distribution of
cannabinoid
receptors. Projections into lateral and medial entorhinal cortex (LEC and MEC, left) originate (off of the diagram) in perirhinal
cortex and
terminate in the superficial entorhinal layers. Cells within layer II give rise to the perforant path and project to the dentate gyrus
(DG). Dentate
granule cells project to CA3, CA3 cells to CA1, and then CA1 cells project to subiculum and deep layers of EC. Neurons in EC
project out to
perirhinal cortex and back to superficial EC, forming a closed loop with the hippocampal circuit. A second pathway projects from
EC via pre
(PRE) and parasubiculum (PARA) to subiculum and CA1, bypassing the dentate gyrus. The densest accumulation of cannabinoid
receptors is in
the molecular layer of the dentate gyrus (dark shading). Cannabinoid receptors are also found in superficial EC, with lesser
densities found in
CA3, CA1, and SUB. The diagonal-shaded region corresponds to the area removed by ibotenate lesions of hippocampus. The
high densities of
cannabinoid receptors in DG could allow cannabinoids to mimic the effects of hippocampal lesions by blocking information flow
through this
portion of the circuit. The remaining circuit from EC to SUB (and back to EC) provides a means for lesioned animals to perform
the same
behavioral tasks, but with different response strategies. The presence of cannabinoid receptors at the beginning of the latter
circuit (shaded
region in EC) could explain additional effects of cannabinoids on the behavioral response strategy.
performance on the DNMS task following the lesion
was significantly below prelesion levels (F
2,327
18.90,
P 0.001) and yet was not significantly different
(F
2,327
1.42, NS) from performance following admin-
istration of THC (Fig. 2). Thus CB1 receptor activation
produced behavioral effects similar to those produced
by the hippocampal lesion.
The hippocampal lesion impaired the ability of the
animal to correctly recall encoded information across
long delays, but did not impair the ability to perform
Copyright 1998 by Academic Press
All rights of reproduction in any form reserved.

the basic DNMS task. However, in animals that re-

ceived lesions that encroached on areas outside the
hippocampus (i.e., including subiculum and entorhi-
nal cortex), performance was impaired at all delays,
including 0-s delay trials (66). Animals that received
lesions prior to training were still able to learn the
DNMS task, and in fact performed better at some
delays than rats that were fully trained before lesion-
ing. However, performance was not uniform for all
trials, suggesting that the animals were performing a

Can nabìnoid Receptors

PARA. PRE,
MEC SUB
lbotenate Lesion
Role of Cannabinoid Receptors in Memory 477
strategy that resulted in elevated performance, but did
not involve actual encoding and recall of the Sample
lever. The appropriate DNMS strategy was for the
animal to press a single (Sample) lever, left or right,
presumably encode and store that information, then
after a delay, recall the information and press the
opposite (Nonmatch) lever. However, some lesioned
animals consistently produced errors on trials that
required a particular ‘‘nonpreferred’’ lever press. Par-
ticularly on trials with delays 15 s, the animal would
preferentially press only one lever for the Nonmatch
response. Thus trials of one type were associated with
greatly increased errors at long delays while trials of
the opposite type had fewer errors at the same delays.
This error pattern was termed a lever ‘‘preference,’’
since the animal’s performance strategy at longer
delays was to press a ‘‘preferred’’ lever, rather than the
appropriate Nonmatch lever. This lever preference was
FIG. 2. Comparison of lesion and cannabinoid-induced deficits in
behavioral performance. Male Sprague–Dawley rats were trained to
perform a delayed-nonmatch-to-sample (DNMS) behavioral task for
present in the performance of all lesioned animals, but
was most consistent in animals that were not pre-
trained in the DNMS task prior to the lesion (66).
water reward (10). Each behavioral trial was initiated by the
Animals performed better on trials beginning with the
extension of one of two retractable levers (left or right). A press on
that lever was recorded as the Sample response and the lever was
immediately retracted, starting the interposed delay phase. The
delay was signaled by illumination of a cue light on the opposite
wall, and varied randomly on any given trial from 1 to 30 s. The
Sample lever in one particular position (the Preferred
trial type), and had at least 10% fewer correct re-
sponses (averaged across all delays) on trials begin-
ning with the opposite Sample lever position (the
animal performed a nosepoke response (breaking a photobeam)
Nonpreferred trial type). In nonlesioned, fully trained
continuously throughout the delay. The last nosepoke occurring
after the delay interval timed out, turned off the cue light, and
extended both levers into the chamber, designating the onset of the
Nonmatch or recognition phase of the task. The animal was required
to press the lever opposite to the Sample (i.e., a Nonmatch response)
animals, there was usually less than 10% difference
between Preferred and Nonpreferred trial types over
all trials (i.e., no Preference) with the greatest differ-
ence only on trials with delays 25 s.
for water reinforcement. Both levers were retracted immediately
Figure 3 illustrates the lever preference for pre-
following the response. Incorrect (Match) responses were followed
by an immediate 5-s dark period in which the chamber lighting was
turned off. A new trial was initiated 10 s after the previous lever
response (correct or error) by the extension of a randomly selected
trained lesioned animals, animals given 9-THC (THC,
pretrained), and nonpretrained lesioned animals. Note
that for the lesioned, pretrained animals, there was no
Sample lever. DNMS trial performance (percentage of correct Non-
significant difference (F
1,327
2.09, N.S.) in the perfor-
match responses) was sorted according to the delay interval of each
trial, plotted as mean ( SEM) percentage of correct Nonmatch
responses within 5-s increments of delay (61). Control curve (filled
circles) was averaged across 20 sessions each from 11 fully trained
male Sprague–Dawley rats prior to either the lesion or initiation of
mance between the Preferred and Nonpreferred trial
types on trials with delays 5 s, but that performance
on the Nonpreferred trial type decreased to chance
levels (i.e., 50%, F
1,327
11.47, P 0.001) on trials with
cannabinoid treatment. Lesion curve (open circles) was averaged
the longest delays (26–30 s). The DNMS performance
over 10 sessions each from 6 pretrained animals with histologically
confirmed ibotenate lesions confined to hippocampus (66). THC
curve (triangles) was averaged over 10 sessions each from 5
pretrained animals that received 9-tetrahydrocannabinol (THC, 2
mg/kg ip injection) 10 min before the start of a DNMS session. Two
curves for the THC, pretrained and lesioned, nonpre-
trained animals showed highly significant differences
between Preferred and Nonpreferred trial types at all
delays (all F
1,327
9.71, P 0.001). As stated above,
non-THC, vehicle-only sessions (days) were interposed between
this lever preference likely reflects a basic change in the
each THC session; performance on vehicle days was not signifi-
cantly different from control. Cannabinoid effects on DNMS perfor-
mance were also blocked by coadministration of the CB1 cannabi-
noid receptor antagonist SR141716A(Sanofi) (1.5 mg/kg ip injection).
strategy used by the animal to perform the DNMS
trial. Under control circumstances at all delays, and in
lesioned animals at short delays, the lever preference
does not influence performance of the DNMS trial.
However, in lesioned animals at longer delays, or on
all trials in animals given 9-THC, the animal appears
Copyright 1998 by Academic Press
All rights of reproduction in any form reserved.

% Correct
Control Lesion THC 2mg/kg
1-5 6-10 11-15 16-2O 21-25 26-
DELAY (sec)
478 Hampson and Deadwyler
FIG. 3. DNMS performance curves comparing lever position biases for lesioned animals vs animals exposed to 9-THC. DNMS
trial
performance was sorted according to the Delay interval of each trial, in 5-s increments as in Fig. 2. Data were additionally sorted
by Sample lever
position within each session. The trial type with higher performance was termed the Preferred lever type (filled circles), and the
trials with lower
performance were labeled as Nonpreferred lever type (open circles). Mean (and SEM) within each delay and trial type data are
plotted. Left:
Performance on trials with Preferred vs Nonpreferred Sample lever positions from 6 animals (15 sessions) that were fully trained
in the DNMS
task prior to lesions. Preferred trials averaged 13% more correct responses than Nonpreferred trials, across all delays. Center:
Preferred vs
Nonpreferred trial performance from 5 animals (10 sessions) that were fully trained in the DNMS task, but received 2.0 mg/kg
(ip) injections of
9-THC 10 min prior to the DNMS session. Preferred trials averaged 39% more correct responses than Nonpreferred trials
following THC
administration. Right: Performance on Preferred vs Nonpreferred trials from 3 animals (15 sessions) that received hippocampal
lesions prior to
training in the DNMS task. Preferred trials averaged 20% more correct responses than Nonpreferred trials.
to respond to a preferred lever on trials in which the
nonpreferred response is appropriate. As stated above,
this implies that without a hippocampus, animals do
not perform the DNMS task in the same manner as
animals with an intact hippocampus. In this respect,
activation of cannabinoid receptors in intact animals
influences more than just the retention of trial specific
events, it also affects the strategy by which the task is
performed.
BOTH HIPPOCAMPAL LESIONS
AND EXPOSURE TO CANNABINOIDS
ALTER SEQUENTIAL DEPENDENCY
IN THE DNMS TASK
The similarity in behavioral effects of both cannabi-
noids and hippocampal lesions shown in Fig. 2 sug-
gested that cannabinoid receptors in the hippocampus
may ‘‘disrupt’’information processing inthe hippocam-
pus by acting as a ‘‘reversible’’ hippocampal lesion
(64,65). The analysis of the types of errors per-
formed within the DNMS task following CB1 recep-
Copyright 1998 by Academic Press
All rights of reproduction in any form reserved.

tor activation was consistent with performance in

lesioned animals. A recent study (67) identified non-
delay-dependent (or ‘‘miscode’’) errors in which infor-
mation from the preceding trial was retained, and
proactively interfered with the current trial (61,62,67).
This proactive interference was shown to occur only
following errors on trials with long delays (i.e., delay-
dependent errors). Following an error, if the next trial
‘‘type’’ (i.e., Sample lever position) was the same,
performance was improved; however, if the trial after
the error was different, performance was decreased.
In lesioned animals, in contrast, proactive interfer-
ence occurs following trials of any length delay (Fig. 4)
not only long-delay errors. This implies that the pres-
ence of the hippocampus provides a time-dependent
‘‘protection’’ from proactive interference not present in
lesioned animals. These same measures of proactive
interference were examined both in lesioned animals
and animals exposed to 9-THC (and plotted as a
function of preceding trial delay in Fig. 4). THC-
treated and lesioned nonpretrained animals showed
similar patterns of proactive interference (Fig. 4).
Figure 4 shows that both groups of animals showed

Lesion, Pretrained THC, Pretrained Lesion, Not Pretrained 100

100 100
% Correct CD O 4
40-

40 - U 40 f 0 Preferred Lever
o NonPreferred Lever
1-5 6-10 11-15 16-20 21-25 26-30 1-5 6-10 11-15 16-20 21-25 26-30 1-5 6-10 11-15 16-20 21-25 26-

Current Trial Delay Current Trial Delay Current Trial Delay

Role of Cannabinoid Receptors in Memory 479
FIG. 4. Proactive influences vs lever position bias on THC and lesioned nonpretrained animals. DNMS performance for each
group was
plotted as a function of length of delay on the preceding trial. Trials were sorted by Preferred (Pref) and Nonpreferred (Non) trial
types (see Fig.
3), as well as whether the preceding trial was the same (i.e., Pref–Pref and Non–Non) or different (Pref–Non and Non–Pref).
Delays were sorted
into 5-s increments according to length of delay on the previous trial (for any length delay on the current trial) (62). Open
symbols represent
mean (and SEM) for pairs of same trials, closed symbols represent mean (and SEM) for pairs of different trials. Left: Curves for
5 animals (10
sessions) given 2.0 mg/kg (ip) injections of 9-THC 10 min prior to the session. Right: Curves for 3 animals (15 sessions) that
received
hippocampal lesions prior to training in the DNMS task. Note that Preferred vs Nonpreferred current trials produced a greater
difference in
performance than proactive influences (same vs different).
only aslightdifference (F
1,327
5.71, P 0.02) in perfor-
mance between pairs of same vs different DNMS trials,
but a much greater difference (F
1,327
10.13, P 0.001)
for Preferred vs Nonpreferred trials. This reflected not
only an increase in the number of trials with the
‘‘delay-dependent’’ type of error (even at short delays,
see Fig. 4) but also a corresponding increase in the
number of trials in which Nonmatch information
proactively interfered with the next trial. Since the
delay-dependent errors were previously demonstrated
to correspond to trials in which Sample phase informa-
tion was ‘‘weakly’’ encoded by the hippocampus, and
both lesions and cannabinoids impair that encoding, it
is likely that the increase in these errors in the lesioned
animals could be directly attributed to the lack of the
hippocampal encoding to reinforce this information
for retention across long delay intervals. The increase
in the influence of lever preference, however, may
Copyright 1998 by Academic Press
All rights of reproduction in any form reserved.

indicate retrohippocampal effects of cannabinoids that

alter the strategy by which the animal performs the
task.
THE ROLE OF CANNABINOID
RECEPTORS IN THE STORAGE AND
RETRIEVAL OF SHORT-TERM MEMORY
The similarities in DNMS behavior between animals
with hippocampal lesions and intact animals exposed
to 9-THC suggested that cannabinoids would act in a
manner similar to a reversible hippocampal lesion. By
virtue of their placement along key parts of the neural
circuit underlying short-term memory, cannabinoid
receptors could act to interrupt information flow dur-
ing either the storage or the retrieval processes of
short-term memory. However, precise comparison of

% Correct
100 100
THC, Pretrained Lesion, Not Pretrained
Non-Pref
1-5 6-10 11-15 16-20 21-25 26-30 1_5 6-10 11-15 16-20 21-25 26-30 Preœding Trial Preceding Trial Delay
480 Hampson and Deadwyler
Copyright 1998 by Academic Press
All rights of reproduction in any form reserved.
cannabinoid effects with those of ibotenate lesions
tion prevents stored information from being ‘‘overwrit-
revealed a more complex interaction. Recent results
ten’’ by extraneous stimuli. It seems likely in either
indicate that cannabinoid agonistsattenuate hippocam-
case that cannabinoids would not affect memory re-
pal neural activity during the Sample (storage) phase,
trieval, but could completely block retention of infor-
but not the Nonmatch (retrieval) phase of the DNMS
mation. The presence of endogenous cannabinoids
task (65). However, cannabinoid receptors are located
within thehippocampalcircuitcould thusvary continu-
on neurons that should be involved in both the
ously to regulate or ‘‘tune’’ the short-term memory
mnemonic storage and retrieval processes (Fig. 1).
storage process. Another prospective role for endog-
In-depth studies of hippocampal lesions in pretrained
enous cannabinoids may be to switch the hippocam-
and nonpretrained animals indicate that animals with-
pus in or out of the circuit for short-term memory
out an intact hippocampus learn a different strategy to
processing. The shift in behavioral performance strate-
perform the DNMS task than control animals (66). In
gies following THC administration may thus indicate
trained animals, successful DNMS behavior likely
that not only is processing interrupted within the
depends on the ‘‘learning’’ of a hippocampal neural
hippocampus, but also that the whole hippocampal/
circuit to store Sample phase information across the
retrohippocampal circuit for retention of DNMS trial
delay, and following the lesion, animals attempt to
relevant stimuli was inactive. This is confirmed by
continue this behavior, but are unsuccessful at delays
preliminary studies which indicate that a combination
10 s. Untrained lesioned animals, on the other hand,
of hippocampal lesions plus THC administration pro-
can learn the DNMS task better without a hippocam-
duces effects similar to those of hippocampal retro-
pus, primarily because the retrohippocampal areas
hippocampal lesions (64). Thus, the role of central
that are spared have not been modified by prior
nervous system cannabinoid receptors (and hence of
training (66). It seems likely that cannabinoid receptors
endogenous cannabinoids) in memory processing
in retrohippocampal areas are therefore also involved
in moderating the behavioral shift, since DNMS behav-
ior in animals with more extensive hippocampal and
retrohippocampal lesions do not show the same delay-
dependent deficits (66).
seems likely to involve ‘‘switching’’ of sections of the
hippocampal memory circuit both to regulate the
storage of information and to initiate other response-
based strategies to deal with delay-dependent contin-
gencies when the hippocampus is taken ‘‘off-line.’’
Recent studies of the effects of cannabinoid agonists
in vitro indicate that cannabinoid receptors may be
located presynaptically on interneurons in many brain
ACKNOWLEDGMENTS
regions such as substantia nigra (68,69), globus palli-
dus (70,71), and hippocampus (35). Application of
The authors thank Doug Byrd, Joanne Konstantopoulos, Janet
potent cannabinoid agonists such as WIN 55,212-2
produced a prolongation of GABA responses and
decreased GABA reuptake (69,70). This effect was
Brooks, Justin Rawley, and John Simeral for assistance. This work
was supported by NIH Grants DA08549 to R.E.H. and DA03502 and
DA00119 to S.A.D.
completely blocked by the antagonist, SR141716A (71).
Thus, cannabinoid receptors may subsequently modify
other neurotransmitter systems via their effects on
REFERENCES
interneurons. A similar role for cannabinoid modula-
tion of neurotransmission has been suggested by
recent studies that show that cannabinoids inhibit
glutamate release in spinal cord (72) and in cultured
1. Miller, L. L., & Branconnier, R. J. (1983) Cannabis: Effects on
memory and the cholinergic limbic system. Psychol. Bull. 93,
441–456.
2. Chait, L. D., & Pierri, J. (1992) Effects of smoked marijuana on
hippocampal neurons (20). Given the critical role for
human performance: A critical review. In: Marijuana/Cannabi-
NMDA receptor in long-term potentiation (LTP) and
memory processing, this modulation may also help
explain the memory suppressive effects of exogenous
noids: Neurobiology and Neurophysiology (L. Murphy & A. Bartke,
Eds.), pp. 387–423. CRC Press, Boca Raton, FL.
3. Hall, W., Solowij, N., & Lemon, J. (1994) The Health and Psychologi-
cal Consequences of Cannabis Use. Australian Govt. Publ. Service,
cannabinoids.
Canberra.
The role of endogenous cannabinoids in hippocam-
4. Stiglick, A., & Kalant, H. (1982) Residual effects of prolonged
pus is somewhat less certain. The above results indi-
cate that cannabinoids may play an essential role in
‘‘forgetting’’ of stored stimuli following retrieval. An-
cannabis administration on exploration and DRL performance in
rats. Psychopharmacology 77, 124–128.
5. Stiglick, A., & Kalant, H. (1982) Learning impairment in the
radial-arm maze following prolonged cannabis treatment in rats.
other possibility is that cannabinoid receptor activa-
Psychopharmacology 77, 117–123.
Role of Cannabinoid Receptors in Memory 481
6. Stiglick, A., & Kalant, H. (1985) Residual effects of chronic
lites of delta-9-tetrahydrocannabinol, and synthetic analogs
cannabis treatment on behavior in mature rats. Psychopharmacol-
having psychoactivity. NIDA Res. Monogr. 79, 148–157.
ogy 85, 436–439.
22. Howlett, A. C., Johnson, M. R., Melvin, L. S., & Milne, G. M.
7. Campell, K. A., Foster, T. C., Hampson, R. E., & Deadwyler, S. A.
(1988) Nonclassical cannabinoid analgetics inhibit adenylate
(1986) Effects of delta-9-tetrahydrocannabinol on sensory-
cyclase: development of a cannabinoid receptor model. Mol.
evoked discharges of granule cells in the dentate gyrus of
Pharmacol. 33, 297–302.
behaving rats. J. Pharmacol. Exp. Ther. 239, 941–945.
23. Devane, W. A., Dysarz, F. A., Johnson, M. R., Melvin, L. S., &
8. Campbell, K. A., Foster, T. C., Hampson, R. E., & Deadwyler,
Howlett, A. C. (1988) Determination and characterization of a
S. A. (1986) Delta-9-tetrahydrocannabinol differentially affects
cannabinoid receptor in rat brain. Mol. Pharmacol. 34, 605–613.
sensory-evoked potentials in the rat dentate gyrus. J. Pharmacol.
24. Herkenham, M., Lynn, A. B., Little, M. D., Johnson, M. R.,
Exp. Ther. 239, 936–940.
Melvin, L. S., de Costa, B. R., & Rice, K. C. (1990) Cannabinoid
9. Deadwyler, S. A., Heyser, C. J., Michaelis, R. C., & Hampson,
receptor localization in brain. Proc. Natl. Acad. Sci. USA 87,
R. E. (1990) The effects of delta-9-THC on mechanisms of
1932–1936.
learning and memory. In: Neurobiology of Drug Abuse: Learning
25. Herkenham, M., Lynn, A. B., Johnson, M. R., Melvin, L. S., de
and Memory (Erinoff, L., Ed.), pp. 79–93. National Institute on
Costa, B. R., & Rice, K. C. (1991) Characterization and localiza-
Drug Abuse, Rockville, MD.
tion of cannabinoid receptors in rat brain: A quantitative in vitro
10. Hampson, R. E., Heyser, C. J., & Deadwyler, S. A. (1993)
autoradiographic study. J. Neurosci. 11, 563–583.
Hippocampal cell firing correlates of delayed-match-to-sample
26. Howlett, A. C., Bidaut-Russell, M., Devane, W. A., Melvin, L. S.,
performance in the rat. Behav. Neurosci. 107, 715–739.
Johnson, M. R., & Herkenham, M. (1990) The cannabinoid
11. Heyser, C. J., Hampson, R. E., & Deadwyler, S. A. (1993) The
receptor: Biochemical, anatomical and behavioral characteriza-
effects of delta-9-THC on delayed match to sample performance
tion. Trends Neurosci. 13, 420–423.
in rats: Alterations in short-term memory produced by changes
27. Bidaut-Russell, M., Devane, W. A., & Howlett, A. C. (1990)
in task specific firing of hippocampal neurons. J. Pharmacol. Exp.
Ther. 264, 294–307.
12. Hampson, R. E., Foster, T. C., & Deadwyler, S. A. (1989) Effects of
delta-9-tetrahydrocannabinol on sensory evoked hippocampal
activity in the rat: Principal components analysis and sequential
dependency. J. Pharmacol. Exp. Ther. 251, 870–877.
13. Deadwyler, S. A., Hampson, R. E., & Childers, S. R. (1995)
Functional significance of cannabinoid receptors in brain. In:
Cannabinoid Receptors (Pertwee, R. G., Ed.), pp. 205–231. Aca-
demic Press, London.
14. Lichtman, A. H., Dimen, K. R., & Martin, B. R. (1995) Systemic or
intrahippocampal cannabinoid administration impairs spatial
memory in rats. Psychopharmacology 119, 282–290.
15. Lichtman,A. H., & Martin, B. R. (1996) Delta 9-tetrahydrocannabi-
nol impairs spatial memory through a cannabinoid receptor
mechanism. Psychopharmacology 126, 125–131.
16. Deadwyler, S. A., Hampson, R. E., Bennett, B. A., Edwards, T. A.,
Mu, J., Pacheco, M. A., Ward, S. J., & Childers, S. R. (1993)
Cannabinoids modulate potassium current in cultured hippocam-
pal neurons. Receptors Channels 1, 121–134.
17. Deadwyler, S. A., Hampson, R. E., Mu, J., Whyte, A., & Childers,
S. R. (1995) Cannabinoids modulate voltage-sensitive potassium
A-current in hippocampal neurons via a cAMP-dependent
process. J. Pharmacol. Exp. Ther. 273, 734–743.
18. Hampson, R. E., Mu, J., Evans, G. J. O., Zhuang, S.-Y., Childers,
S. R., & Deadwyler, S. A. (1995) Role of cAMP-dependent protein
kinase in cannabinoid receptor modulation of potassium A-
Copyright 1998 by Academic Press
All rights of reproduction in any form reserved.
Cannabinoid receptors and modulation of cyclicAMP accumula-
tion in the rat brain. J. Neurochem. 55, 21–26.
28. Mackie, K., & Hille, B. (1992) Cannabinoids inhibit N-type
calcium channels in neuroblastoma-glioma cells. Proc. Natl.
Acad. Sci. USA 89, 3825–3829.
29. Twitchell, W., Brown, S., & Mackie, K. (1997) Cannabinoids
inhibit N- and P/Q-type calcium channels in cultured rat
hippocampal neurons. J. Neurophysiol. 78, 43–50.
30. Mu, J., Zhuang, S.-Y., Kirby, M. T., Hampson, R. E., & Deadwyler,
S. A. (1997) Cannabinoid receptors have differential effects on
potassium D-current and A-current channels, but are linked via
the same cAMP-dependent processes. Soc. Neurosci. Abstr. 23,
1480.
31. Hampson, A. J., Bornheim, L. M., Scanziani, M., Yost, C. S., Gray,
A. T., Hansen, B. M., Leonoudakis, D. J., & Bickler, P. E. (1998)
Dual effects of anandamide on NMDA receptor-mediated re-
sponses and neurotransmission. J. Neurochem. 70, 671–676.
32. Stella, N., Schweitzer, P., & Piomelli, D. (1997) A second endog-
enous cannabinoid that modulates long-term potentiation. Na-
ture 388, 773–778.
33. Terranova, J. P., Michaud, J. C., Le Fur, G., & Soubrie, P. (1995)
Inhibition of long-term potentiation in rat hippocampal slices by
anandamide and WIN55212-2: reversal by SR141716 A, a selec-
tive antagonist of CB1 cannabinoid receptors. Naunyn Schmiede-
bergs Arch. Pharmacol. 352, 576–579.
34. Collins, D. R., Pertwee, R. G., & Davies, S. N. (1995) Prevention
by the cannabinoid antagonist, SR141716A, of cannabinoid-
mediated blockade of long-term potentiation in the rat hippocam-
current in cultured rat hippocampal neurons. Life Sci. 56,
pal slice. Br. J. Pharmacol. 115, 869–870.
2081–2088.
35. Hampson, R. E., Mu, J., Kirby, M. T., Zhuang, S.-Y., & Deadwyler,
19. Sim, L. J., Hampson, R. E., Deadwyler, S. A., & Childers, S. R.
S. A. (1998) Role of cannabinoid receptors on GABAergic
(1996) Effects of chronic treatment with delta-9-tetrahydrocan-
neurons in the rat hippocampus. Soc. Neurosci. Abst. 24, 1245.
nabinol on cannabinoid stimulated [35S]GTP-gamma-S autoradi-
36. Scoville, W. B., & Milner, B. (1957) Loss of recent memory after
ography in rat brain. J. Neurosci. 16, 8057–8066.
bilateral hippocampal lesions. J. Neurol. Neurosurg. Psychol. 20,
20. Shen, M., Piser, T. M., Seybold, V. S., & Thayer, S. A. (1996)
11–12.
Cannabinoid receptor agonists inhibit glutamatergic synaptic
37. Zola-Morgan, S., Squire, L. R., & Amaral, D. G. (1986) Human
transmission in rat hippocampal cultures. J. Neurosci. 16, 4322–
amnesia and the medial temporal region: Enduring memory
4334.
impairment following bilateral lesion limited to field CA1 of the
21. Howlett, A. C. (1987) Regulation of adenylate cyclase in a
hippocampus. J. Neurosci. 6, 2950–2967.
cultured neuronal cell line by marijuana constituents, metabo-
38. Squire, L. R., Zola-Morgan, S., & Chen, K. S. (1988) Human
482 Hampson and Deadwyler
Copyright 1998 by Academic Press
All rights of reproduction in any form reserved.
amnesia and animal models of amnesia: Performance of amnesic
57. Leonard, B. W., Amaral, D. G., Squire, L. R., & Zola-Morgan, S.
patients on tests designed for the monkey. Behav. Neurosci. 102,
(1995) Transient memory impairment in monkeys with bilateral
210–221.
lesions of the entorhinal cortex. J. Neurosci. 15, 5637–5659.
39. Squire, L. R., & Cave, C. B. (1991) The hippocampus, memory,
58. Jerusalinsky, D., Ferreira, M. B., Walz, R., Da Silva, R. C.,
and space. Hippocampus 1, 269–271.
40. Olton, D. S., & Feustle, W. A. (1981) Hippocampal function
required for nonspatial working memory. Exp. Brain Res. 41,
380–389.
41. Raffaele, K. C., & Olton, D. S. (1988) Hippocampal and amygda-
loid involvement in working memory for nonspatial stimuli.
Behav. Neurosci. 102, 349–355.
42. Aggleton, J. P., Hunt, P. R., & Rawlins, J. N. (1986) The effects of
hippocampal lesions upon spatial and non-spatial tests of
working memory. Behav. Brain Res. 19, 133–146.
Bianchin, M., Ruschel, A. C., Zanatta, M. S., Medina, J. H., &
Izquierdo, I. (1992) Amnesia by post-training infusion of gluta-
mate receptor antagonists into the amygdala, hippocampus, and
entorhinal cortex. Behav. Neural Biol. 58, 76–80.
59. Jones, R. S. G., & Lambert, J. D. C. (1990) Synchronous discharges
in the rat entorhinal cortex in vitro: Site of initiation and the role
of excitatory amino acid receptors. Neuroscience 34, 657–670.
60. Amaral, D. G., & Witter, M. P. (1989) The three-dimensional
organization of the hippocampal formation: Areview of anatomi-
43. Yee, B. K., & Rawlins, J. N. (1994) The effects of hippocampal
cal data. Neuroscience 31, 571–591.
formation ablation or fimbria-fornix section on performance of a
61. Deadwyler, S. A., Bunn, T., & Hampson, R. E. (1996) Hippocam-
nonspatial radial arm maze task by rats. J. Neurosci. 14, 3766–
pal ensemble activity during spatial delayed-nonmatch-to-
3774.
sample performance in rats. J. Neurosci. 16, 354–372.
44. Rawlins, J. N., Lyford, G. L., Seferiades, A., Deacon, R. M., &
62. Hampson, R. E., & Deadwyler, S. A. (1996) Ensemble codes
Cassaday, H. J. (1993) Critical determinants of nonspatial work-
involving hippocampal neurons are at risk during delayed
ing memory deficits in rats with conventional lesions of the
performance tests. Proc. Natl. Acad. Sci. USA 93, 13487–13493.
hippocampus or fornix. Behav. Neurosci. 107, 420–433.
63. Hampson, R. E., Rogers, G., Lynch, G., & Deadwyler, S. A. (1998)
45. Sinden, J. D., Rawlins, J. N., Gray, J. A., & Jarrard, L. E. (1986)
Selective cytotoxic lesions of the hippocampal formation and
DRL performance in rats. Behav. Neurosci. 100, 320–329.
46. Jarrard, L. E. (1993) On the role of the hippocampus in learning
and memory in the rat. Behav. Neural Biol. 60, 9–26.
47. Otto, T., & Eichenbaum, H. (1992) Neuronal activity in the
hippocampus during delayed non-match to sample perfor-
mance in rats: Evidence for hippocampal processing in recogni-
tion memory. Hippocampus 2, 323–334.
48. Eichenbaum, H., Otto, T., & Cohen, N. J. (1992) The hippocam-
pus—What does it do? Behav. Neural Biol. 57, 2–36.
49. Bunsey, M., & Eichenbaum, H. (1993) Critical role of the
Facilitative effects of the ampakine CX516 on short-term memory
in rats: Correlations with hippocampal ensemble activity. J.
Neurosci. 18, 2748–2763.
64. Hampson, R. E., Byrd, D. R., Konstantopoulos, J. K., Bunn, T., &
Deadwyler, S. A. (1996) Delta-9-tetrahydrocannabinol influences
sequential memory in rats performing a delayed-nonmatch-to-
sample task. Soc. Neurosci. Abstr. 22, 1131.
65. Hampson, R. E., Rawley, J., Simeral, J. D., Byrd, D. R., Brooks,
J. K., & Deadwyler, S. A. (1997) Disruption of encoding but not
recognition by cannabinoids via differential action on hippocam-
pal memory circuits. Soc. Neurosci. Abstr. 23, 509.
parahippocampal region for paired-associate learning in rats.
66. Hampson, R. E., Jarrard, L. E., & Deadwyler, S. A. (1998) Effects
Behav. Neurosci. 107, 740–747.
of ibotenate hippocampal destruction on delayed matching and
50. Angeli, S. J., Murray, E. A., & Mishkin, M. (1993) Hippocampec-
nonmatching to sample behavior in rats. J. Neurosci. in press.
tomized monkeys can remember one place but not two. Neuropsy-
67. Deadwyler, S. A., & Hampson, R. E. (1997) The significance of
chologia 31, 1021–1030.
neural ensemble codes during behavior and cognition. In:
51. Cho, Y. H., Beracochea, D., & Jaffard, R. (1993) Extended
Annual Review of Neuroscience (W. M. Cowan, E.M. Shooter, C. F.
temporal gradient for the retrograde and anterograde amnesia
Stevens, R. F. Thompson, Eds.) Vol. 20, pp. 217–244, Annual
produced by ibotenate entorhinal cortex lesions in mice. J.
Reviews, Palo Alto, CA.
Neurosci. 13, 1759–1766.
68. Sanudo-Pena, M. C., & Walker, J. M. (1997) Role of the subtha-
52. O’Keefe, J., & Burgess, N. (1996) Geometric determinants of the
place fields of hippocampal neurons [see comments]. Nature 381,
425–428.
53. Eichenbaum, H., & Cohen, N. J. (1988) Representation in the
hippocampus: What do hippocampal neurons code? Trends
Neurosci. 11, 244–248.
54. Eichenbaum, H., Otto, T., & Cohen, N. J. (1994) Two functional
components of the hippocampal memory system. Behav. Brain
Sci. 17, 449–518.
55. Dusek, J. A., & Eichenbaum, H. (1997) The hippocampus and
memory for orderly stimulus relations. Proc. Natl. Acad. Sci. USA
94, 7109–7114.
lamic nucleus in cannabinoid actions in the substantia nigra of
the rat. J. Neurophysiol. 77, 1635–1638.
69. Tersigni, T. J., & Rosenberg, H. C. (1996) Local pressure applica-
tion of cannabinoid agonists increases spontaneous activity of
rat substantia nigra pars reticulata neurons without affecting
response to iontophoretically-applied GABA. Brain Res. 733,
184–192.
70. Glass, M., Brotchie, J. M., & Maneuf, Y. P. (1997) Modulation of
neurotransmission by cannabinoids in the basal ganglia. Eur. J.
Neurosci. 9, 199–203.
71. Maneuf, Y. P., Crossman, A. R., & Brotchie, J. M. (1996) Modula-
56. Otto, T., Schottler, F., Staubli, U., Eichenbaum, H., & Lynch, G.
tion of GABAergic transmission in the globus pallidus by the
(1991) Hippocampus and olfactory discrimination learning:
synthetic cannabinoid WIN 55,212-2. Synapse 22, 382–385.
Effects of entorhinal cortex lesions on olfactory learning and
72. Richardson, J. D., Aanonsen, L., & Hargreaves, K.M. (1998)
memory in a successive-cue, go-no-go task. Behav. Neurosci. 105,
Hypoactivity of the spinal cannabinoid system results in NMDA-
111–119.
dependent hyperalgesia. J. Neurosci. 18, 451–457.