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".
Department of Pharmacology and Toxicology, Medical College of Virginia/Virginia Commonwealth University. Richmond,
Virginia
Accepted for publication August 19, 1994
ABSTRACT
Dose-effect

Administration to
Mice: Possible Mechanisms for Interaction with Morphine’
SANDRA P. WELCH, CHARITY THOMAS and GRAY S. PATRICK

THC and

tracerebroventricularly (icv.)] and compared with those previ-

ously generated after administration intrathecally
The ED,,
values after administration of

56,667 was significantly more potent after icv. administration

than
administration, and was nearly 10 times more potent
than CP 55,940 (icv.j. CP 55,940 and

for 7 days
totally abolished the antinociception produced by the cannabi-

ABBREVIATIONS:

sulfoxide.
310

manner nearly 1 O-fold

after icv. administration. The

or Cl-CAMP (10

the cannabinoids.
Recent work has shown that the cannabinoids produce
potent antinociceptive effects in mice and rats through both
spinal and supraspinal mechanisms (Welch and Stevens,
1992; Welch, 1993; Smith and Martin, 1992; Lichtman and
Martin,

the

salt; DMSO,

Modulation of Cannabinoid-Induced Antinociception After

56,667, which did not
produce greater than additive effects in combination with

shifted phine when the drugs were administered

the mor-
phine (icv.) dose-effect curve in a

nabinoids (icv.) were not blocked by ICI 174,864 (20

mouse),

significantly attenuated the antinociception produced

by the
and CP 55,940. However, when administered
icv., forskolin and Cl-CAMP produced antinociception, but did
not block or produce greater than additive effects with the

possible effect; THC, tetrahydrocannabinol;

monosodium salt; 4-AP, 4-aminopyridine; TEA, tetraethylammonium;

or naloxone (20 kg/mouse or

10

Received for publication September

This

CL, confidence limit; AD,,, dose producing 50% antagonism; Cl-c-AMP,

antinociception in the rat has been proposed to

(icv. and i.t.). Pretreatment of the mice with forskolin

the mechanism by which cannabinoids produce

supported by the Nationai

Pertussis toxin pretreatment

Although the spinal mechanism of cannabi-

(79
were generated for the canrrabinoids

. VS
icv. did not differ significantly.

intrathecally; icv, intracerebroventricularly;

and

receptor

i.t.), which produced no

1993.

of Drug Abuse Grants

Versus

and Martin,

55,940,

of the

intravenously;

tration of the cannabinoids. Conversely.

the antinociceptive effects of the cannabinoids. The AZ.

ues

antinociception produced by the cannabinoids

failed to alter the antinociception produced by the

which did not alter the antinociceptive effects of

Thapsigargin (icv.) blocked the antinociception

icv. The

CP 55,940 were 215

nmollmouse, respectively.

effect curves of

ATP- and apamin-sensitive potassium channels, these

intracellular systems may be common points of

nificantly reversed the calcium-induced blockade of

sium channels, produced a parallel rightward shift in the

apamin (5

nabinoids. Because acute administration of

tinociception has not been

tion of a specific cannabinoid receptor has been the to]

intense investigation leading to the cloning of a

have been shown to interact with

In receptor
addition. an endoge
ligand for the cannabinoid receptor has been identified
ane et al.,

tion-What are the transduction mechanisms

with the activation of the cannabinoid receptor’?

modulation of adenylate

sible transduction mechanisms have been proposed

calcium. prostaglandin
opioids. The role of adenylate cyclase, calcium

and potassium channel modulation, and to a lesser

decrease calcium entry to and content of neurons.

levels and produce hyperpolarization of neurons

dibutyryl cyclic
and CP 55,940. Apamin, blocker

for calcium-induced

nor-binaitorphimine; ICI 174,864.

administration of calcium and calcium

subcutaneously:

These reports led to the next logical

icv.) failed to block icv.

176 (122-253) and

and

determined. The

monophosphate

percentage of the

of

protein-coupled

55,940

(1
1995
role of
of the cannabinoids is the focus of this paper. These
mechanisms will be discussed individually, although the in-
terplay of these various systems in the transduction
the
effects of the cannabinoids is clearly a possibility.
C
AMP. The cloning of the cannabinoid receptor was per-
formed in
modula-
tion after receptor activation

Numerous studies have shown the involvement of the

cannabinoids in the modulation of

amide, the endogenous cannabinoid ligand, has also been

shown to inhibit

levels in cells of
various types in culture as well as in homogenates of brain
regions (Kelly and Butcher. 1979; Zimmerman et al..

et
al., 1986). The potency of numerous cannabinoids to inhibit

and Ng, 1984; Howlett and Fleming, 1984; Howlett, 1984;

Howlett,
The interaction of the cannabinoids with a
membrane protein

bosylated protein was identified as the

formation in the neuroblastoma cells was found to

correlate to the antinociceptive effects of the drugs in

The effects of the cannabinoids on

spinal
accumulation have not been examined. This
study utilizes modulators of adenylate cyclase or
levels
in
to evaluate the effects of such treatments on the
antinociceptive effects of the cannabinoids in the brain and in
the spinal cord.

transmission and an-

tinociception. Previous work has indicated that

An alteration of intracellular cal-

cium by the cannabinoids could be the trigger for a

events leading to decreased

de-
creases the release of acetylcholine presynaptically by a de-
crease in the influx of calcium into presynaptic nerve
terminals

studies of the involvement

of calcium in the antinociceptive effects of the cannabinoids
have not been reported previously. In addition, potassium
channel opening and the resultant hyperpolarization of

19821 found that cannabinoids decrease calcium uptake to

several brain regions, an effect which did not correlate to the
psychoactivity of the drugs. Direct measurement of the ef-
fects of cannabinoids on free intracellular calcium

has shown that depolarization-in-

duced rises in intracellular calcium are attenuated by

vitro studies indicate a role for calcium in the

effects of the cannabinoids, in

membrane could account for the antinociceptive

effects of cannabinoids. The modulation of potassium chan-
nels by the cannabinoids may differ in the brain and in the
spinal cord. The involvement of potassium channels in

nabinoid-induced effects has been reported in

although the nature of the interaction of the

drugs with potassium channels has not been determined.

et al., 1973; Chesher and

Jackson, 1985; Sanders et al., 1979; Martin,

fails to block the effects of various parenterally admin-

istered cannabinoids

also failed to block the antinociception in-

duced by a variety of

al.,

is unclear. Many investigators have shown that

icv. or

in concentrations of 1

Welch,

using

et

The role of opioids in cannabinoid antinocicep-

as possible mechanisms in the antinoeiceptive

1981; Gilbert, 1981; Welch and Stevens,

It is intriguing that, despite the data

with an evaluation of

and

ADP-ribosylation was shown. The

production in

icv. or spinally administered

or higher

et al.,

protein

et

Naloxone

Stokes

1988

brain
311
suggesting that cannabinoids and opiates have distinct

of action, the effects of morphine have been

found to be enhanced by crude cannabis extract
and
Bhattacharya,

et al., 1984). Intrathecal administration

of several cannabinoids leads to synergism with

istered morphine in the production of antinociception in mice

binoids

intraventricular injections. Intrathecal in-

jections were performed according to the protocol of Hylden and
Wilcox

antinociception, but not

catalepsy,
or hypoactivity has been also re-
ported
Although we have evaluated the
effects of opioid antagonists on the antinociceptive effects of

(Welch and Stevens, 1992;

Welch,

with icv. administered opioids and opioid antagonists

has not been reported previously.
Thus, the research presented in this manuscript was de-
signed to evaluate the involvement of the cannabinoids

routes both
of administration) on the modulation
of
in the
production ofantinociception. In addition, we wanted to com-
pare and contrast the interaction of the carmabinoids with
morphine in the brain and in the spinal cord. Finally, from
our results we hoped to generate a hypothesis as to the
possible

Mice were lightly anesthetized with ether and an incision was made
in the scalp such that the bregma was exposed. Injections were
performed using a 26-gauge needle with a sleeve of PE

produced significant antinociceptive effects

alone in the tail flick test and were not used as the cannabinoid
vehicle when performing i.t. injections. Preliminary time course
studies of the various drugs (at the highest dose which exhibited no
toxicity) or their respective controls were initially performed in com-
bination with the ED,, dose of each cannabinoid. Once the time of
the peak effect of those drugs which altered cannabinoid

were administered. lntraventricular injec-

tions were performed according to the method of Pedigo et al.

tubing to
control the depth of the injection, An injection volume of 5

DMSO. Nor-BNI, ICI 174.864, naloxone hydrochloride, morphine

sulfate, apamin, pertussis toxin, calcium chloride, magnesium chlo-
ride.

and verapamil were dissolved in distilled

water. The cannabinoids or DMSO vehicle

min before determination of the response latency of the mice in the

tail flick test. This time point represents the peak

DMSO vehicle produced scratching behavior in mice

which lasted 2 min after injection. Other vehicles have been tested
previously in our laboratory.

needle.
and L6 area of the spinal cord with a
Injection volumes of 5

at a site
2-mm rostra1 and 2-mm
to the bregma at a depth of 2 mm was
administered to the mice. The cannabinoids,

charybdotoxin. TEA.

of the drugs
as determined in previous studies in our laboratory (Welch and
Stevens, 1992

spinal and supraspinal sites.

Methods
Intrathecal

administered cannabinoids, and the interaction of

by

K 8644, ryanodine and nimodipine were prepared in

was determined. all further testing was performed at that

and Stevens, 1992; Smith and Martin, 1992; Smith et

and

of the A”-THC-induced

Recently, the blockade of cannabinoid antinocicep-

kappa opioid antagonist,

with morphine

Unanesthetized mice were injected between the

been reported

et

the interaction of icv. administered

Cannabinoid-Induced

1:

interaction of the opiates and the

and potassium channels in

and by orally administered

Cl-CAMP, forskolin.

1993

were administered

The blockade by

and emulphor!
thapsigargin,

1975

and
Welch et al.
time point. With few exceptions, the time of the peak effects of the
drugs was 5 to 10
before the administration of the cannabinoids.
The exceptions include glyburide

pretreatment; thapsigargin, which produced peak

blockade after a 1-hr pretreatment and pertussis toxin, which was
administered i. t. (0.5

potas-
sium channel openers and morphine (Welch and

days before testing in combination

with the cannabinoids. Thapsigargin, Bay K 8644 and calcium pro-
duced antinociception when administered i,t. Forskolin,

1993)
using a 15

and Cl-CAMP and Bay K 8644 produced antinociception when

administered
To test for greater than additive effects of the
intrinsically antinociceptive drugs
in combination with
the cannabinoids, doses of the drugs were chosen which produced no
significant antinociception. These doses were then tested in combi-
nation with a wide range of doses of the cannabinoids. For the
studies of the enhancement of the effects of morphine

nabinoids
before
morphine. The mice were tested 10 min after the administration of
morphine using the tail flick test.
To study the effect of
on calcium-induced blockade of
the cannabinoids, the
was administered icv. at 5 min
before the calcium
which was administered at 10 min before the
cannabinoid
The mice were tested 15 min later using the tail
flick test.
The tail flick test. The tail flick procedure used was that of

and
a cut-off time of 10
were empioyed. Antinociception was quanti-
fied as the %MPE as developed by Harris and Pierson
using
the following formula:

Percent MPE was calculated for each mouse using at least 12 mice
per dose. Using the

modification of the

were
determined using the method of Litchfield and
as
well as the

1.
Using at least three doses of blocker,

for each mouse, the mean effect and

standard error of the mean
was calculated for each dose.
Dose-response curves were generated using at least three doses of
test drug.

method omitting doses pro-

ducing 100% or 0% MPE) and 95% confidence limits

et al., 1977).
The AD,, for blockade the cannabinoid-induced antinociception
was determined by calculation of the percent antagonism of

ciception (using a dose of the cannabinoid that resulted in at least

80% MPE) by the blockers according to the following formula:
% antagonism = 100

values were determined

by log-probit analysis (a modification of the

method omitting doses producing 100% or 0%

1949).
were
At least 12 mice per dose were used for all determinations.
Statistical analysis. Significant differences between treatment
and control groups were determined using analysis of variance and
the Dunnett’s
19551. The dose-response curves were
evaluated for the parallelism of shifts using the method of Tallarida
and Murray

Drugs. All of the cannabinoids were obtained from the National

Institute on Drug Abuse with the exception of CP 55.940, CP 56,667,
levonantradol and dextronantradol which were obtained from Dr.
Lawrence Melvin, Pfizer Central Research. Nor-BNI, Bay K 8644,

MA). chemicals, Inc.

ICI 174,864 was purchased from
Cambridge Research

and Smith

and forskolin were purchased from Research

test

the cannabinoids were administered 10

values were determined by log-probit analysis

control\
%MPE = 100 x

using the method of Litchfield and

program

as well as the

Control reaction times of

(test

(Valley Stream, NY). Calcium

of antagonist + cannabinoid)

of vehicle + cannabinoid)

which blocks both

or

control)

program
and 95%

to 4

by

et

monosodium

was provided by Miles, Inc. Pharmaceutical

Division (West Haven, CN). Thapsigargin was purchased from LC

GVIA was purchased from Pen-

insula (Belmont, CA). Ryanodine and pertussis toxin were purchased
from Calbiochem (San Diego, CA).
Results

values for a series of cannabinoids are listed in

table 1. The ED,, values of the cannabinoids after
admin-
istration were taken from Welch and Stevens
and
were compared with the ED,, values generated for the

icv. administration. The dose-effect curves

used to generate the

of the cannabinoids were full

agonists in the tail flick test

precluded but the solubility of the

the use of higher
doses. The

did not differ significantly after adminis-

tration

after icv. administration. Conversely,

the

us. and CP
icv. were nearly identical. The ED,,
values for

or icv.. although a trend toward higher potency of

the drug was
did not differ significantly after
administration
or icv., although a trend toward higher
potency of the drug was observed after i.t. administration. CP
56,667 was significantly more potent after icv. administra-
tion than
administration, and was nearly 10 times more
potent than CP 55,940
and CP 56,667 are
stereoisomers and CP 55,940 has been described to be the
active isomer in many tests (Johnson and Melvin, 1986;
Little

icv.
administration. However, 25 &mouse produced 44% MPE
in the tail flick test. Dextronantradol was inactive at 25

and icv. administration.

Interaction with opioids. Table 2 lists the ED,, values
for morphine

(taken from Welch and Stevens, 1992) and the ED,, values
for morphine (icv.)

and levonantradol
produced greater than additive effects in combination with
TABLE 1
ED, values

and 95% confidence limits for various

cannabinoids in the tail flick test in mice
The cannabinoids were administered erther

using the tail flick test. The ED,, values and 95%

2.89 (1.6-5.1)
CP 56,667 4.17

Levonantradol 0.04

Dextronantradol Inactive at 25

11

shown in figure 1.

55,940 2.28
IN).

Drug

which produced only an average of 68% MPE at

under “Methods.”

magnesium chloride, verapamil. glyburide.

al.,

to complete a dose-effect curve of the drug

and apamin were purchased from Sigma

values for

after both

We tried to administer higher doses of

was

MA).

As-THC,

values after administration of

obtained from Boehringer Mannheim

after pretreatment with cannabinoids

We were unable to obtain enough

44.97 (22.96-88.09) 16.4 (1 l-24.8)

72.07 (36.06-l 44) 125.8

14.67

pretreatment with cannabinoids

values for the cannabinoids

CP

both

or

and icv.

at

the exception of

0.02

0.18

were determined

25

min before

= 44%

4-U
icv.)
Fig. 1. Dose-effect
icv. administration.
Cannabinoids were injected icv. and tested for the production of

tinociception using the tail flick test and the ED,, values were deter-
mined as described under “Methods.” The following cannabinoids were
tested: A = levonantradol; B

At least 12 mice were used per dose.

morphine when the drugs were administered

nantradol) failed to produce greater than additive effects

with morphine

CP 55,940,
CP 56,667 and dextronantradol

and CP 56,667
produced greater than additive effects with morphine when
the drugs were administered icv. The
values for mor-
phine were shifted from 2.4

produced greater than additive effects with morphine when

the drugs were administered either i.t. or icv. The shifts in
the morphine dose-effect curve in combination with the

mouse by CP

nabinoids were parallel.

Various opioid antagonists were administered
before
icv. administered cannabinoids and did not alter

noid-induced antinociceptive effects. These drugs are listed

in Table 3 and include the delta opioid antagonist. ICI
174,864 (20
the kappa
opioid antagonist, nor-BNI (70

(Takemori et al.,
1988) and naloxone (20

Interaction with
C

for 7 days, which ribosylates

doses used blocks mu, delta and kappa receptors

et al., 1987

E=

0.001 0.01 0.1

10 100 1000

Dose

AMP. Pertussis toxin pretreatment

Conversely, CP

for the cannabinoids

and CP 56,667, respectively.

CP 56,667: C = CP 55,940; D =

or 10

et al.,

and/or

inactive isomer of

to 0.28 and 0.65

proteins (for a

which at

313
TABLE 2
ED, values
for morphine various cannabinoids in the tail
in combination with
and icv.
administration
Mice were injected with cannabinoids either
binoid pretreatment or vehicle pretreatment. Cannabinoids were administered 10
min before morphine and the mice were tested using the tail flick test at 10 min
after morphine administration. The

were determined as
described under “Methods.”
Pretreatment
ED,, of morphine

review see Brown and Birnbaumer, 19901, totally abolished

the antinociception produced by the cannabinoids

and significantly attenuated or blocked the antinocicep-

tion produced by the cannabinoids

administered i.t. was least affected by the pertussis toxin

pretreatment. Pretreatment with pertussis toxin
did not
alter the baseline latencies for the mice in the tail flick test
and did not produce any overt toxicity in the mice. However,
pretreatment with pertussis toxin icv. (0.5
for 7
days) led to loss in body weight, and lethality in greater than
30% of the animals. Pretreatment of the mice with 5 and 25
&mouse forskolin
which produced no antinociception,
significantly attenuated the antinociception produced by the

A similar effect was ob-

served using CP 55,940. Forskolin (25 &mouse
pre-
treatment significantly reduced the antinociceptive effects of
CP 55,940

forskolin MPE. However, when administered

produced
antinociception. The ED,, of forskolin

which produced
no antinociceptive effects, did not block or produce greater
than additive effects with the antinociception produced by
the cannabinoids administered icv. Administration of

partially, but significantly,

blocked the antinociceptive effects produced by the
ad-
ministration of ED,, doses the three cannabinoids tested (fig.

failed to alter the antinocicep-

tion produced by the cannabinoids administered i.t. How-
ever,

produced antinociceptive effects when ad-

ministered icv. [ED,, values = 0.5

was
obtained using higher doses of Cl-CAMP. Both Cl-CAMP and
dibutyryl

and
2.3)
dibutyryl
respectively) did
not block or produce greater than additive effects with the
antinociception produced by the cannabinoids administered
icv.
Interaction with calcium. Various calcium modulators
were tested in combination with the cannabinoids (table

The

1.6)

None 1 (0.6-l

2.7 (1.7-4.3)
DMSO 0.61 (0.26-I

Levonantradol 0.06 (0.015.24) 0.66

N.T.
11
Dextronantradoi 0.51

morphine. The dose-effect curve for morphine was determined after

55.940 0.3

However, no greater blockade of the

(3.13
(6.25

(25

administration of calcium
administered

(10

Forskolin (0.1

Cannabinoid-Induced Antinociception

respectively]. Inactive doses of

(0.02 and 0.05

0.15
0.05

0.05

0.08

0.26

(fig.

from 83

values and 95%

test in mice after

before

morphine or icv. before

MPE to 16

(fig.

2.4 (1.7-3.3)

0.65

and 0.3

was 0.5
8

N.T.

(fig.
itself did not alter the effects of the cannabinoids, it did
significantly block calcium-induced blockade of the cannabinoids.
3
gargin (0.1 &mouse), which increases intracellular calcium

314 Welch et al.

TABLE 3

cannabinoids after

before
cannabinoids
and
the time

pools
of the endoplasmic reticulum by blocking
activity
(Bian et al., 1991; Koshiyama and Tashjian, 1991); Bay K
8644 (0.3 &mouse), which predominantly increases calcium
entry through dihydropyridine-sensitive calcium channels;
verapamil (30

&mouse),
which block voltage-gated

which alters calcium flux from

the intracellular caffeine-sensitive (but not
sensitive)
calcium pool (Thayer et al., 1988;

cal-
toxin (1 &mouse) which blocks both
cium channels

and ryanodine

et al., 1989) failed to

produce antinociception or to alter the antinociception

administration of the cannabinoids. The

doses listed represent the highest non-antinociceptive dose
for each of the drugs. At higher doses, calcium, thapsigargin
and BAY K 8644, which increase intracellular calcium, pro-
duced antinociception after

blocked the antinociceptive effects of the

cannabinoids (fig. 4, panel A). The AD,, values
for
calcium-induced block of
mouse, respectively. Higher doses of calcium (greater than
450

because we could not use higher doses of calcium. This diffi-

culty is reflected in the large confidence limits for the AD,, of
calcium us.
which did
not alter the antinociceptive effects of

and CP 55,940
were 215

produced hyperactivity in the mice. It

was difficult to generate an

icv.
Opioid antagonists
Naloxone No block No block
ICI 174,864 No block No block
nor-BNI
Blocked No block
Calcium modulators
Calcium No block Blocked
Verapamil No block No block

No block No block
Nimodipine No block No block
Thapsigargin No block Blocked
BAY-K No block No block
Ryanodine No block No block
Modulators of

Blocked No block
Forskolin Blocked No block
Magnesium Chloride No block No block
Modulators of potassium channels

No block No block
Apamin Blocked No block
l

administration.
Calcium

No block No block
TEA No block No block
Glyburide No block No block

blockade of entry of calcium into the

by the

et al.

were administered

of blockade of the antinociceptive effects of

toxin Blocked Blocked

as described under “Methods”

1986;

Antinociception was determined using the tail flick

Drugs

and icv. administration

176 (122-253) and 123

et al., 1987; Reynolds et al.,

et al., 1987; Zemig,

and nimodipine

before

(1 &mouse icv.

calcium channels (God-

for calcium us.

No block No block

and

significantly

or
Fig. 2. The effect of pertussis toxin pretreatment (i.t.) on the antinoci

(panel A) c
icv. (panel

7 vehicle was administered

days before the administration of th
cannabinoids

or icv.) in distilled water-pretreated mice (the vehicle for the

At least 12 mice per treatment group were used.

l

vehicle. Th
mice were tested for antinociceptive effects after either

administration of the cannabinoids or DMSO. The clear bar denote

6% (icv.), respectively. Bars denoted

represent the antinociceptive effects of those cannabinoids in mic

pretreated for 7 days i.t. with distilled water vehicle. The bars denote

PANEL B
100
80

group.
respective distilled
reversed the calcium-induced blockade of

blocked the antinociception produced by

gargin us.

panel B!.
Pretreatment of the mice icv. with thapsigargin for 1

55,940, data not shown) administered icv.

Other calcium modulators were tested including:

or tion after either

icv. administration. All of these dru:
were tested at the highest nontoxic doses in combination
the cannabinoids administered either i.t. or icv. and failed

alter cannabinoid-induced antinociception (table 3).

PANEL A

represents the antinociception produced by the administration

40

vehicle in PT-treated mice. Not shown is the effect of DMSO

indicate the effects of the cannabinoids in PT-pretreated

generated in those treatment groups were 5

effects of the cannabinoids administered either

and nimodipine

omega conotoxin

None of these drugs produced

= 0.003

toxin

or CP 55,940) or

(0.5 &mouse) or distilled

and magnesiu:

and “CP

verapamil

3%

55940
for

0.01

(fig.

and 8

(fig.

or

Th
PANEL A
Fig. 3. The effect of forskolin pretreatment (i.t.) (panel A) or Cl-c-AMP
[panel B) on the antinociceptive effects of the cannabinoids adminis-
tered
or distilled water vehicle
was administered
or DMSO
vehicle. The mice were tested for antinociceptive effects after

represents the antinociception produced by the admin-

istration of distilled water

in distilled water-pretreated mice (the vehicle for the forskolin).

The

55,940” represent the antinociceptive effects of those

Cannabinoids in mice pretreated

which block voltage-gated potassium

channels (North, 1992;

mouse), which blocks ATP-gated potassium channels

mouse), which blocks the large

with distilled water vehicle. The

bars denoted

with potassium channels. Various potas-

sium channel blockers were tested: TEA

sett et al., 1988; Gaines et al., 1988;

at these doses, which are the highest nontoxic doses of

the drugs, failed to alter cannabinoid-induced

mice per treatment group were used.

were used for all treatment groups. Panel

and

However. apamin. blocker of

Panel A, Forskolin

potassium channels (Smith et al., 1986;

et al., 1988). These drugs administered either

effects of distilled water administration before DMSO. At least

and Lazdunski,

generated in that treatment group was

of the cannabinoids or

i.t.)-pretreated mice. “VEH” denotes the

before the administration of

indicate the effects of the cannabinoids in

before

1 5 25

@mouse,

and charybdotoxin

fast) conductance

1992); glyburide

Not shown is the effect of

The dark bar denoted

0.01.

Bars denoted

et al., 1989; and

4%. At
conductance

and

12

or

Fig. 4. The effect of calcium pretreatment (icv.) (panel A) or w-conotoxin

CP 55,940 or
DMSO vehicle. The mice were tested for antinociceptive effects after
icv. administration of the cannabinoids or DMSO. The bar denoted

represents the antinociception produced by the administration

of distilled water icv. before DMSO icv. The AD,, values for calcium
blockade of the cannabinoids were determined as described under
“Methods.” At least 12 mice per treatment group were used.

Mice were pretreated

with dual injections icv. of vehicle (v), calcium (Ca, 300

binoids were shifted from

or distilled water vehicle was administered

icv. before the icv. administration of

et al., 1982; Cas-

tle et

as described
under “Methods.” The mice were tested 15 min after the injection of

At least 12 mice per treatment group were used. w-Conotoxin

significantly reversed the calcium-induced blockade of

calcium-gated potassium channels

produced a parallel rightward

shift in the dose-effect curves of

was shifted
from 6

failed respectively. However, apamin (5

to
block icv. administered cannabinoids.
and CP
55,940 (fig. 6, panels A-C). The ED,, values for the

to 33-fold by administration of
apamin

120, 300

315
PANEL A

(OC) icv. before calcium pretreatment (panel B) on the

--
Calcium Calcium Calcium
PANEL
100
80

0.01 from

60

40
20

effects of the cannabinoids administered icv. Panel A. Calcium

antinociception.

and CP 55,940 were shifted from 21

1989; Strong,

&mouse to 123

to 82

Cannabinoid-Induced

P
The ED,, of

and/or

0.01 from

Panel

and 20

or

The ED,, values for

0.05 from

and 0.6

0.05;
to
greater than lo-fold by the pretreatment with cannabinoids
at doses which did not have any antinociceptive effects. W’hen
we compared the enhancement of morphine-induced

administration ciception by the cannabinoids

with
those data previously generated after i.t. administration, we
found that some cannabinoids enhanced morphine in the

1981; Gilbert, 1981; Lichtman and Martin, 1991, a and b;

Welch and Stevens, 1992). Intrathecally

nabinoids seem to act at predominantly spinal sites in the

production of antinociception (Smith and Martin, 1992). It is
not clear whether the icv. administration of the cannabinoids
activates both spinal and supraspinal sites in the production
of antinociception, Due to the lypophilicity of the drugs, pas-
sage to the spinal cord is very likely.
In summary, the cannabinoids seem to produce potent
antinociceptive effects at both spinal and supraspinal sites,
However, the mechanism by which the cannabinoids produce
antinociceptive effects at these two sites seemed to differ and
is the focus of the remainder of this manuscript.
Interaction with opiates. We have
previously
that the ED,, values for morphine were shifted from

316 Welch et al.

The mice were tested 15 min later in the tail flick

test. The clear bar denoted
represents the antinociceptive ef-
fects of DMSO icv. administered 1 hr before

binoids produced antinociception after both i.t. and icv. ad-

ministration and are approximately
at either site
with the exception of CP 56,667. These results are consistent
with previous studies which indicate that the cannabinoids
are active as antinociceptive drugs when injected

icv.
Fig. 5. Blockade of

gargin (icv.). Thapsigargin, at the doses listed, or vehicle was injected at

1 hr before

The remain-
ing bars represent thapsigargin administration before
At least
12 mice were used per dose.
l
P

represents

20

VEH 0.001
0.005 0.01
THAPSIGARGIN

vehicle

Discussion

(icv.)-induced antinociception by

of the cannabinoids.

at 1 hr before

0.05 from

icv. The bar labeled

0.01 from

been shown to be dense in the striatum and the

which are associated with dense binding of the opiatt

et al., 1988; Gamse et al., 1979). Thus, neuroanatom

cal location of the receptors does not seem to account for

differences observed between the effects of CP 55,940

were determined 15 min later using the tail flick test. The ED
values and the parallelism of the shifts in the dose-effect curves we
determined as described under “Methods” using 12 mice per dose.
brain whereas others enhanced the effects of morphine in
bility that subtypes of cannabinoid receptors may exist

brain and spinal cord and differ in their interactions

opiate-sensitive pathways either via their neuroanatomic:

proximity to opiate receptors or differ in the convergence

second messenger systems with those of the opiates. Hot;

ever, in the case of one cannabinoid, CP 55,940, binding

gelatinosa of the spinal cord

are administered

spinal cord. Our data are suggestive of the intriguing

Fig. 6. Blockade of the antinociceptive effects of

min before the cannabinoids. The dose-effect curves of the

PANEL A

i.t.)
PANEL C

(panel

and CP 55,940 (panel C) by apamin when all

10

Apamin (5

1000
A8

1 10

doses

doses

doses
was administered at

et

A9 +
apamin
ED50 = 82

(panel
in the
spinal cord.
opioid or the direct interaction with opiate receptors.
Thus, the enhancement of opiate antinociception by

binoids icv. may be due to interactions with common intra-

cellular second messenger systems. Some possible points of
interaction include the modulation of calcium,

sensitive, high-voltage-activated calcium channel, an effect

that is blocked by the administration of pertussis toxin and
independent of the formation of
C
AMP. Because the L-type
calcium channel blocker, nitrendipine, fails to alter such an
effect, the cannabinoids were hypothesized to interact with
an N-type calcium channel (Mackie and Hille, 1992). A sim-
ilar study produced similar results. The cannabinoids were
found to inhibit I,, current in newoblastoma cells, an effect
which was not. dose related, but was pertussis toxin- and

administration of calcium
to mice

administration of the cannabinoids is not sensitive to

calcium modulation and thus may not directly involve cal-
cium modulation.
Conversely, the icv. administered cannabinoids were sen-
sitive to blockade by calcium icv. and the AD,, values for
calcium icv. us. the cannabinoids

of the cannabinoids at supraspinal sites are not me-

diated

ciceptive effects of icv. administered cannabinoids. These

data suggest that the icv. administration of the cannabinoids
does not result in the release of an antinociceptive

and
potassium flux.
Modulation of calcium and

enhance or fail

blastoma cells indicate cannabinoids inhibit an

1986) or by
modulators of intracellular pools of calcium such as

et al., 1988; Welch et al., 1992; Smith and

Dewey, 19923, as well as Bay K 8644 and thapsigargin
(Damaj et

before the cannabinoids did not

produce greater than additive effects with the cannabinoids.
These data indicate that the antinociception produced by the

did not differ signifi-

cantly from that previously shown for calcium blockade of

calcium entry to
a variety of tissues. Electrophysiological studies in

All such
previous studies were performed in
Based on the re-
sults of such studies we hypothesized that the antinocicep-

produced by the cannabinoids might be sensitive to

calcium, or modulators of calcium.

administration of calcium chloride or a

series of modulators of calcium flux across voltage-gated
calcium channels

ciceptive effects result from the

1993). Administration of non-antinociceptive

doses of these drugs

effects of the cannabinoids administered i.t. were not

nor-BNI administered icv. failed to block the

have previously reported that the kappa antagonist,

and ICI 174,864

data support the hypothesis that the antinociceptive

1990;

by the

produced by the cannabinoids administered

interaction with nor-BNI-sensitive receptors.

Thayer et al., 1988;

et al., 1991; Koshiyama and Tashjian, 1991

(Godfraind et

but not other opioid antagonists, blocks

alter

antinociception in the brain

et

et

(Caulfield and Brown

antinociception (Welch, 1993).

1988; Harris and Stokes,

1987; Reynolds et

also failed to block the

et al.,

nimodipine, Bay K 8644 or

1986;

Cannabinoids either

the antinocicep-

1989

et al., 1987;

or

morphine icv. (Harris et al., 1975; Chapman and Way, 1980;

Guerrero-Munoz and

and stimulates calcium influx into cells

(for a review see Irvine,
also blocked the antinocicep-
tion produced by

which increases intracellular calcium (Koshiyama

and Tashjian,

which increases calcium

entry to cells via the voltage-gated
calcium channel,
as well as via nonspecific mechanisms (Bouchelouche et

nabinoids is due to calcium entry through voltage-gated cal-

cium channels because

induced blockade (fig.

alone did not
alter cannabinoid-induced antinociception. Thus, it appears
that the entry of calcium to neurons must be stimulated by
the administration of calcium to observe effects of omega

calcium chan-
nels, and found no reversal of the effects of icv. calcium (data
not shown). Thus, the calcium-induced blockade of

a similar logic we attempted to reverse the

calcium-induced blockade of cannabinoids with both

pamil and nimodipine, which block

calcium noid-induced antinociception seems to be due

en-
try through the

tinociception by thapsigargin, which raises intracellular cal-

cium levels, indicates that the cannabinoids may decrease

and Stokes (1982) and Martin et al. (1988). In

addition, the stimulation of calcium entry to neurons after
thapsigargin treatment would further increase calcium con-
centrations intsacellularly, an effect insensitive to blockade
by classic calcium channel blockers (for review, see Irvine?
1992; Jackson et al., 1988; Takemura et al., 1991) and con-
sistent with our data. Our data therefore suggest that,

formation follows a structure-activity relationship and

stereoselectivity similar to that observed for behavioral mea-
sures (Howlett, 1985; Dill and Howlett, 1988; Howlett et al.,
1986; Howlett and Fleming, 1984; Howlett, 1987; Howlett et
al., 1990; Devane et

1989

calcium channel. These results are

consistent with reported effects observed in vitro (Mackie
and Hille, 1992). The blockade of cannabinoid-induced

entry to neurons which is consistent with the in vitro

work

nabinoids produce antinociception at supraspinal, but not

spinal, sites by decreasing intracellular calcium levels via
blockade of calcium entry through
calcium chan-
nels.
Consistent with the report of Mackie and Hills

antinociception resulting from supraspinal administra-

tion of the cannabinoids seems to involve interaction with a

due
to the blockade of antinociception by pertussis toxin pretreat-
ment which inactivates

lation (for a review, see Brown and

proteins, nabinoids with

but indicating a lack of involve-
ment of

induced antinociception or other behavioral effects of the

cannabinoids (Hillard and Bloom, 1983; Little and Martin,
1991; Dolby and Kleinsmith, 1974; Dolby and Kleinsmith,
1977; Askew and Ho, 1974). In
studies indicate a some-
what different profile of cannabinoid action. Studies in

1986; Bidaut-Russell et al., 1991;

Howlett et al.,

Our
data agree with studies indicating an interaction of the

did not alter cannabinoid antinociception.

It is likely that the calcium-induced blockade of the
or

doses of BAY K 8644

protein and be independent of changes in

cells show that cannabinoid-induced inhibition of

in the brain in the modulation of

Cannabinoid-induced Antinociception

and is mediated via coupling to the

although

and CP 55,940.

and

1982;

proteins via.

reversed the

et al.,

in
al.
protein because pertussis toxin attenuates binding of CP
55,940

These studies, along with supporting work evaluating

the stereoselectivity and the Hill coefficient for binding of the
synthetic

ylation of calcium channels

nabinoid receptor linked through a

our studies indicate that modula-

tion of

tion by alteration of calcium entry to neurons

protein to the modu-

lation of
C
AMP.

is not directly involved in the production of

antinociception, we can not rule out the possibility that

and Birnbaumer, 1990;

Schultz et
cannabinoids
could decrease calcium entry to neurons.
In the spinal cord, the cannabinoids seem to produce

protein tinociception
in conjunction with the
modulation of
because the antinociceptive effects ofi.t.
cannabinoids are blocked by pertussis toxin, forskolin and
Cl-CAMP, The lack of blockade of antinociception by

may be due to the lack of penetration of the

Gimenez-Gallego
et

Morphine has been shown to modulate

ratios, High ATP levels close these channels leading to a less

negative membrane potential and depolarization. Opening-of
these channels leads to hyperpolarization. Pharmacologi-
cally, these channels are blocked by the oral hypoglycemic
agents, such as glyburide, which stimulate the entry of
to the
site of action. The effects of
on antinociception are the
opposite in the brain us. in the spinal cord, which is similar to
the opposite effects of calcium observed in brain

administration may reflect differences in the function

of
in the brain us. in the spinal cord in the- modulation
of antinociception.
Modulation of potassium.
basic classes of potas-
sium channels exist in the central nervous system (for a
review, see Aronsen, 1992). This classification is clearly a
generalization and the drugs discussed often interact with
more than one
of
the channels. In general these potassium
channels are either voltage-activated, calcium-gated,

gated or ligand-activated. The

nels are either opened or closed in response to membrane

electrical activity and are blocked pharmacologically by TEA
and

proteins or other second messenger systems,

Calcium-gated potassium channels have been studied using
the bee venom
apamin, which blocks the small (low)
conductance (SK or
et al., 1982;
Castle et al., 1989; Strong, 1990). An endogenous ligand for
the apamin receptor has been found in brain

tive calcium-gated potassium channels in peripheral sensory

neurons

channels are gated by changes in intracellular

et al.,
1988: Howlett et al..

spinal
cord

The calcium-gated potassium channels are activated

by alteration in intracellular calcium and are thought to be
coupled to

et al.,
1987). Charybdotoxin, a scorpion venom, blocks predomi-
nantly the large

1988). Charybdotoxin does not interact with

apamin-sensitive sites in rat brain

as well as several toxins (North, 1992; Aronsen,

1992

indirectly alters icv. cannabinoid-induced

in cannabinoid-induced antinociception following icv.

19883 by insertion into the mouth of the channel

preparations

et al,, 1988). Thus, differences in the

19901. By decreasing

et

cannabinoid. CP 55,940

et

fast) conductance

or

1988; Bidaut-Russell and Howlett,

1992). The ATP-gated

support the hypothesis of a

et al.,

channel

potassium

et al.,

channel in

of

et

Anatomically, these channels exist in areas where high levels

of opiate binding occur

and Lazdunski, 1990). ATP-gated potassium channels have

been found in the brain, with particularly high levels in the
substantia nigra, hippocampus, basal ganglia and thalamus.
and to a lesser degree in the spinal cord, although the func-
tion of such channels remains unknown

et al., 1988). Opiates also

interact with l&and-activated potassium channels, a diverse
collection of channels which are modulated by a variety of
neurotransmitters and drug classes. These channels are
most

effects produced by the cannabinoids, although the

is blocked by potassium channel blockers and con-

versely, that potassium channel openers administered

produce antinociception that is blocked by opiate antagonists

(Welch and
The administration of apamin. a
blocker of calcium-gated potassium channels, produced a
complete blockade of morphine-induced antinociception.
Evaluation of several potassium channel blockers

blocked the antinociceptive ef-

fects of the cannabinoids
producing parallel rightward
shifts in the dose-effect curves of the cannabinoids.

the parallel shifts could be coincidental and the

could be nonspecific, this is unlikely because
parallel shifts were observed with three cannabinoids.

it is likely that cannabinoids interact spinally with a

gated, apamin-sensitive potassium channel in the production

of antinociception. Cannabinoids administered icv. appear to
produce antinociception which does not involve

on cannabinoid-induced antinociception can not

be attributed to lack of binding of the apamin.
Summary. Cannabinoids produce antinociception after
administration to both spinal and supraspinal sites; how-
ever, modulation of antinociception by calcium,
and
potassium seems to differ in the brain
in the spinal cord.
Additionally, the cannabinoid-induced modulation of opiate
antinociception seems to differ in the brain and the spinal
cord. Based on our data, we hypothesize that in the brain,
cannabinoids produce antinociception
decreasing calcium
levels in neurons through receptor linkage to
proteins
involved in the modulation of the function of the

calcium channel. Cyclic

seem to not be involved in the production of antinociceptior

induced by icv. administered cannabinoids. In the

ing in a decreased

calcium-gated potassium channels. The involvement

of

tion with G proteins has the potential to alter both

and potassium channel function.

Because acutely administered
have beer
shown to interact with

of
apamin in

sitive potassium channels. Since apamin binding has been

shown in the brain
the lack of effects of
apamin

or to play an indirect role. In addition, potassium

cord, our data indicate that the cannabinoids produce

ciceptive

1978; Childers and Snyder,

We have shown previously that opiate-induced

in combination with the cannabinoids

into the ceils of both the pancreas and the brain

appears to not play a direct role in the

1988; Gaines

that only apamin

coupled to a G protein (Brown and Birnbaumer.

via the interaction with G,, proteins

1988;

protein-coupled receptors

production and interaction

et

t.

seems to either not be

decrease calcium entry

et al.,

and
et

and

cal
and content of neurons (Harris et al., 1976; Cardenas and
R
OSS
,
1976; Welch and Olson,

levels
(Collier and Roy,
1974;
1978) and
produce hyperpolarization of neurons

these three intracellular systems

may be common points of interaction with the cannabinoids.
For example, it has been hypothesized that the end result of

bination may be the phosphorylation of similar proteins

which are proposed to be synapsins I and II which are

volved in the release of

(Childers et al.,
1992). Finally, the data suggest that cannabinoids interact
spinally with nor-BNI-sensitive receptors, This could indi-
cate some interaction of the cannabinoids directly with opi-
ate-sensitive pathways. Kappa receptor-selective opioids

have not been shown to be displaced by the cannabinoids.

Thus, although we can hypothesize as

be done to
delineate conclusively the mechanisms underlying the
greater than additive effects observed using cannabinoids
and opiates in combination.
Acknowledgments
The author wishes to thank Dr. Bill R. Martin, Department of

the points of inter-

action ofcannabinoids and opiates based on the in
mech-
anisms of action of the drugs indicated by this work and that

and Toxicology, Medical College of Virginia for his

assistance and collaboration on this project.
References

levels in rat brain areas.

calcium-activated potassium channels: Effects of channel gating, volt-

age, and ionic strength. J. Gen. Physiol. 91: 317333, 1988.
of intracellular
calcium pools. Selective modification by thapsigargin. J. Biol. Chem.

A. C.: Cannabinoid receptors

and modulation of cyclic AMP accumulation in the rat brain. J. Neurochem.
57: 21-26. 1991.

membranes:
Regulation by ions and nucleotides.

Effect of BAY K 8644

on
free calcium in isolated rabbit gall-bladder epithelial cells. Cell
Calcium

Protein subunits.

morphine treatment. Br. J. Pharmacol.

toxin-sensitive
mechanism. Br. J. Pharmacol.

Rev.
Pharmacol. Toxicol.

G. A.: The effect of cannabinoids on intestinal

motility and their

others, clearly much more research needs

modulation of

11-14, 1992.

W. E.

accumulation in

A. M.

G. B.
J.

of potassium channels. Trends Pharmacol.

G. B..

D.

in

S.

C.,

H. L.
AND

et al., 1987) and

A.,

1973.

Interaction of ligands with opiate receptors

1991.

37-46. 1989.

AND

P. N.,

et al.,

T.

M.,

Ho, B. T.: The effects of

W
AY

J
ACKSON

MING

Ross, R. H.: Calcium depletion of

, E. I,.: Metal ion interactions with opiates.

potassium channels (Welch and

cannabidiol and

channels in nervous tissue.

G.
AND

553-579, 1980.

, D. M.: The quasi-morphine withdrawal syndrome:

D. A.: Cannabinoid

J. C.

rat

W. A.

S
MITH

G.

L.: Ionic channels and their regulation by

by cannabinoids and opioids in

Physiol. 52: 197-213, 1990.

A. C.: Cannabinoid receptor-regulated cyclic

,
C..

G
ILL

cells via a

effect in mice. Br. J.

231-232, 1992.

opioids

J. Neurochem. 57:

, D. L.:

MILLER. C.:

M.,

Natl.

1975;

521-526, 1976.

D.,

H.: Toxins in the

reduce

1974.

both ATP- and

D. M.,

M., S
EXTON

U.S.A. 75:
59-65, 1989.

et al.,

Pharmacol.

on cyclic

Pharmacol.

inhibit

block

, T.

49:

of

C
OLLIER

M.: Regulation of the ATP-sensitive potassium

channel. In Ion Channels, ed. by T. Narahashi, pp. 205-222, Plenum Press,
New York, 1990.
D
ILL
, J. A.

1157-
1163, 1988.
D
OLBY
, T.

Pharmaeol. 23: 1817-1825, 1974.

D
OLBY
, T. W.
AND

81: 7728-7732. 1987.

G
AINES
, K L.,

275,
1979.
G
ILBERT
,

1975.
H
ARRIS
, R. A.
AND

, AND
S,: Opioid and cannabinoid receptor inhibition of adenylyl cyclase
in brain. Ann N.Y.

des of opiate agonist and antagonist receptor interactions. J. Neurochem. 34:

Morphine-like drugs inhibit the stimulation by

E

A. C.: Determination and characterization of a

of cyclic AMP formation by rat brain homogenates. Nature

248: 24-27, 1974.

B. R.: Involvement of calcium and

L-type calcium channels in nicotine-induced antinociception. J. Pharmacol.
Exp. Ther., 286: 1330-1338, 1993.

logical dose-response curves. Am. J. Physiol. 235:

receptor in rat
brain. Mol. Pharmacol. 34: 605-613, 1988.

sensation. J. Pharmacol. Exp. Ther. 72:

D.: Simultaneous analysis of families

of sigmoidal curves: Application

,
G.,
R.: Isolation
and structure of a brain constituent that binds to the cannabinoid receptor.
Science 258: 19461949, 1992.

A. C.: An assessment
of the role of opioid receptors in the response
cannabimimetic drugs. J.
Neurochem.

by chronic
exposure

on the
levels of cyclic

Physiol. and phosphodiesterase activities of mouse brain. Can.

Pharmacol.
55: 934-942, 1977.

control action potential properties

in heart cells via high

The presence in pig brain of an endogenous equivalent of

W.: A multiple comparison procedure for comparing several treat-

ments with a control.

receptors that are linked to ATP-dependent

potassium channels. J. Biol. Chem.

the bee
venom

F.: Indirect evidence for presynaptic

location of opiate receptors in chemosensitive primary sensory

that specifically blocks calcium-dependent potassium chan-

nels.

urea receptor on beta cell membranes. J.

of mor-
phine analgesia in rat-Role of brain

E.: A comparison of THC, nantradol, nabilone, and morphine in the

chronic spinal dog. J.

cations. cation
chelators, and an ionophore on morphine analgesia and tolerance. J.

G.

selective inhibitor of calcium-activated potassium

channels.
M.: Calcium antagonism and calcium
entry blockade. Pharmacol. Rev.

analgesic response mod-

ified by calcium. Life Sci. 31:

phan series. J. Pharmacol. Exp. Ther. 143: 141-145, 1964.

E. L. In Opiates and
Endogenous Opioid Peptides, ed. by H. W.

New York, 1976.

F.: Opioid delta

receptor involvement in supraspinal and spinal
in mice.
Brain

in the effects of
the cannabinoids on

W
ARD

, H. 0. J.

M. I.,

R., H
OLZER

L. S.

W. A.,

R. A.,

Natl.

R. A.,

W. A.,

W. A., S
PAIN
A..

M.,

F. E.

J.

Exp. Ther. 195:

M., D
EWEILLE

C.

R.
AND

B. R.

R.

Natl.

M., L
YNN

D,,

M.: Antidiabetic

420:

AND

M
ILLER

cannabimimetic drugs. J. Pharmacol.

J. Neurosci. 2: 443-447, 1982.

F.

S
NYDER

Natl.

. P.

G.,

1929-1935, 1986.

S
MITH

H. H.

Cannabinoid-Induced

M. L.: Purification. sequence and model structure of

B
LOOM
, A. S.: Possible role of

P. J.

, J. W.,

R
ICE

R
OY

S. P.
AND

L., B
REWER

, R.

, A. B.,
I. F. A.,

U.S.A. 87: 1932-1936. 1990.

, A.

, J. R., G
REEN

S.

, D. L,: A method for determining loss of

Sci.

J. A.: Cannabinoids inhibit calcium uptake by brain

A C.: Regulation of

A. K.: Some narcotic antagonists in the

H.,

Am. Stat. Assoc.

Sci.

A.,

1987.

C.: Cannabinoid receptor localization in brain.

M. A..

L. J.: Cannabinoid effects on adenylate

Sci. U.S.A.

Pharmacol. 21:
L. J.: Effects of

H.: Differential regulation by

W
AY

, A.,

K.: Cannabis-induced

A.,

Pharmacol.

Z.:

, E. L.: Effects of

H. H. AND

C.

H
UGHES

33-51, 1992.

bioassay,

A. E.: Characterization of the

, R.

H. I.
AND

activity. Eur. J. Pharmacol.

M. R., M
ELVIN

D., J
OHNSON

7933-7936, 1988.
, M..

1982.

J.

3329-3333, 1988.

R. G., S
TEVENSON

A.

10961121, 1955.

1986.

1941.

281-285, 1979.

in mouse brain.

pp. 361-368,

G.

J. Med. Res.

, L.

G. M..

1981.

assay, and

R.,

2589-2592.1988.

, L. S., G
RIFFIN

1977.

L. S.,

AND

21-