Directory UMM :Data Elmu:jurnal:B:Biological Psichatry:Vol49.Issue2.2001:
Effects of Chronic Antidepressant Drug Administration
and Electroconvulsive Shock on Locus Coeruleus
Electrophysiologic Activity
Michael M. Grant and Jay M. Weiss
Background: The locus coeruleus (LC) is the major
noradrenergic cell body group in the brain. Although
previous studies have examined changes in electrophysiologic activity of LC neurons produced by antidepressant
drugs, only a small number have examined changes that
occur with chronic drug administration, which is the
therapeutically effective regimen, and only one group of
investigators has assessed effects on activated (or
“burst”) firing of LC neurons under such treatment
conditions. The present study assessed changes produced
in rats by effective antidepressant treatments—several
drugs given chronically (two tricyclic antidepressants, two
selective serotonin reuptake inhibitors, and a monoamine
oxidase inhibitor) as well as a series of electroconvulsive
shocks (ECSs)—in single-unit electrophysiologic activity
of LC neurons, measuring effects on spontaneous depolarization rate and also on sensory-evoked burst firing.
Methods: Drugs were administered via osmotic
minipumps for either 14 or 30 days; ECSs were administered five times, with a 72-hour interval between each
administration. Electrophysiologic recording of LC activity took place under halothane anesthesia on the last day
of drug treatment or following a delay of 1 or 5 days after
the final ECS.
Results: A common effect of all drugs tested and ECS
treatment was to decrease LC spontaneous and sensoryevoked burst firing.
Conclusions: The clinical efficacy of antidepressant medication and ECS may be mediated, in part, through
reduction of LC neural activity. The findings reported here
are consistent with recent indications that LC neurons are
hyperactive in depressed individuals and with suggestions
that some behavioral changes seen in depression can arise
from consequences of rapidly depolarizing LC terminals,
such as release of peptides. Biol Psychiatry 2001;49:
117–129 © 2001 Society of Biological Psychiatry
From the Department of Psychology, Georgia State University (MMG) and the
Department of Psychiatry and Behavioral Sciences, Emory University School
of Medicine (JMW), Atlanta, Georgia.
Address reprint requests to Jay M. Weiss, Ph.D., Emory University School of
Medicine, Emory West Campus, 1256 Briarcliff Road NE, Atlanta GA 30306.
Received January 22, 2000; revised May 2, 2000; accepted May 16, 2000.
© 2001 Society of Biological Psychiatry
Key Words: Antidepressant, locus coeruleus, electrophysiology, tricyclic, SSRI, electroconvulsive shock
Introduction
T
he suggestion that noradrenergic neurons in the brain
are involved in depression has been with us for many
years. The original catecholamine hypothesis of depression, advanced more than 30 years ago (Bunney and Davis
1965; Schildkraut 1965; Schildkraut and Kety 1967),
proposed that depression arose from a deficiency of
norepinephrine (NE) in the brain. But when various
predictions derived from this hypothesis were not confirmed (e.g., depressed individuals did not appear to have
lower levels of NE or metabolites in brain, and tricyclic
antidepressants required many days to reverse depression
despite elevating synaptic NE immediately after administration), initial enthusiasm for the catecholamine hypothesis gave way to attempts to develop formulations that did
not focus on NE.
Despite prodigious efforts to develop alternatives, evidence has continued to point to the involvement of NE in
depression and the action of antidepressant drugs. For
example, when effective antidepressant drugs were discovered that did not block NE reuptake or directly increase
NE release, this seemed to further compromise the catecholamine hypothesis until it was subsequently discovered
that these drugs downregulate beta-adrenergic receptors,
an action that could derive from, or was a consequence
equivalent to, prolonged action of NE in the synapse
(Vetulani et al 1976a, 1976b; Vetulani and Sulser 1975). A
survey of antidepressant drugs that appeared shortly thereafter concluded that almost all antidepressant drugs potentiated NE action in the synapse (Richelson and Pfenning
1984). Even drugs developed more recently that seem to
suggest the primacy of neurotransmitters other than NE
have been found to significantly affect noradrenergic
systems. For instance, the antidepressant drugs classified
as selective serotonin reuptake inhibitors (SSRIs) produce
marked increases in extracellular NE in the brain as well
0006-3223/01/$20.00
PII S0006-3223(00)00936-7
118
M.M. Grant and J.M. Weiss
BIOL PSYCHIATRY
2001;49:117–129
Table 1. Effects of Chronic Treatment (1 Week or Longer) with Antidepressant Drugs on
Spontaneous and Sensory-Evoked Depolarization Rate of Locus Coeruleus Neurons
Spontaneous
Tricyclic antidepressants
Desipramine
Imipramine
Monoamine oxidase inhibitor
Phenelzine
Huang et al 1980
McMillen et al 1980
– Valentino et al 1990
Svensson and Usdin 1978
Blier and de Montigny 1985
Valentino and Curtis 1991
Evoked
– Valentino et al 1990
? Valentino and Curtis 1991a
Selective serotonin reuptake inhibitor
Sertraline
– Valentino et al 1990b
– Valentino et al 1990
Atypical antidepressant
Mianserin
–? Curtis and Valentino 1991c
–? Curtis and Valentino 1991c
The depolarization rate shown by animals treated with drug in comparison with animals receiving no drug was , decreased,
or 2, unchanged.
a
Evoked response assessed, but results not reported for this measure. Report presents only the ratio of evoked to spontaneous
rate; because spontaneous rate was decreased by phenelzine, effect on evoked rate could not be determined from data presented.
b
Spontaneous firing rate of sertraline-treated rats was not different from untreated rats but was significantly lower than that
of rats that received a similar schedule of injections of vehicle.
c
Spontaneous and evoked firing rate of mianserin-treated rats was not different from untreated rats but was less than that of
rats that received a similar schedule of injections of vehicle; however, the statistical significance of the comparison of
mianserin-treated rats with vehicle-treated rats was not presented (see Curtis and Valentino 1991 [Figure 4, 334].
as serotonin when given systemically as occurs in
clinical use (Cosford 1995; Jordan et al 1994). Furthermore, the antidepressant drug bupropion blocks dopamine reuptake more effectively than any other antidepressant drug presently in widespread use, but
preclinical studies have found that antidepressantlike
behavioral changes produced by bupropion parallel
changes in brain NE produced by this drug rather than
changes in dopamine (Cooper et al 1994). Yet another
example implicating NE is provided by the Flinders
Sensitive rat strain that was selectively bred for sensitivity to cholinergic drugs and has been proposed as a
model for depression (reviewed in Overstreet 1993).
Although these animals were bred for sensitivity to
cholinergic drugs, their depression-related behavior has
been found to respond to drugs that affect brain NE
rather than acetylcholine (Schiller et al 1992). In
summary, despite various inconsistencies regarding the
role of NE in depression, research continues to produce
evidence that activity of noradrenergic neurons is important in this disorder.
The study described here examined changes in electrophysiologic single-unit activity of neurons of the major
noradrenergic cell-body group in the brain, the locus
coeruleus (LC). This study assessed the changes produced
by effective antidepressant treatments, both chronic administration of antidepressant drugs and electroconvulsive
shock. The terminals of LC neurons supply nearly 70% on
the total NE in the brain, giving rise to most of the NE in
the forebrain and all of the NE in cortex and hippocampus.
Studies have examined the effects of antidepressant drugs
on electrophysiologic activity of LC neurons, but most of
these studies have reported effects of acute drug administration. Because antidepressant drugs require prolonged or
chronic administration to have therapeutic effects, changes
of greater clinical interest would be those observed after
these drugs have been given for a therapeutically relevant
period of time. Also, LC neurons have two (or at least two)
modes of firing: they depolarize at a spontaneous rate but
also fire more rapidly in response to excitatory stimuli.
Few studies have examined effects of antidepressant
treatment on rapid firing (or “burst” firing) that occurs in
response to an excitatory stimulus. The previously reported effects of chronic administration of antidepressant
drugs on spontaneous depolarization rate and burst firing
to an excitatory stimulus are summarized in Table 1.
Examination of this table will reveal that the reported
effects have not been consistent. To clarify these effects,
we assessed spontaneous depolarization rate and sensoryevoked burst firing in rats treated chronically with several
different types of antidepressant drugs or a series of
electroconvulsive shocks.
Methods And Materials
Subjects
Male and female Sprague-Dawley rats (virus/antigen free, bred
in our laboratory from stock originally obtained from Charles
Antidepressant Effects on LC Electrophysiology
River [Wilmington, MA]) aged 5 to 6 months were used.
Animals were group housed (three to a cage) directly on bedding
in solid-bottom polycarbonate cages. In studies of antidepressant
drugs, animals in any single cage received the same drug, with
control animals (i.e., animals receiving vehicle) distributed
among drug groups and housed with them. Animals that received
ECS and control animals for this manipulation were similarly
housed. Food (lab chow) and water were available ad libitum. A
12:12 light:dark cycle and temperature of approximately 21°C
was maintained in the colony room.
Drugs and Administration
The effects of five antidepressants were examined: desipramine
HCl (DMI), a secondary amine dibenzazepine tricyclic NE
reuptake inhibitor (Sigma, St. Louis); imipramine HCl (IMI), a
tertiary amine dibenzazepine tricyclic NE reuptake inhibitor
(Sigma); phenelzine sulfate (PHE), a nonselective monoamine
oxidase inhibitor (Sigma); and fluoxetine HCl (FLU) and sertraline HCl (SER), both heterocyclic SSRIs (Lilly and Pfizer,
respectively). Antidepressants were administered via Alzet Osmotic Minipumps (Alza, Mountain View, CA), implanted subcutaneously in the animals, using Model 2ML2 pumps for 14-day
administration and Model 2ML4 pumps for 30-day administration. Minipumps, which provide continuous drug delivery beginning approximately 4 hours after pump implantation, were used
to insure the presence of drug in the animal throughout the study
and at the time that electrophysiologic recording took place. Use
of minipumps also eliminated the need for repeated handling and
injection of animals to administer drug chronically. Studies that
have assessed effects of daily injection of vehicle while studying
effects of antidepressant drugs have found that repeated injections constitute a stressful procedure. For example, compared
with no injections, repeated vehicle injections have been observed to produce 1) elevated brain norepinephrine release as
measured by microdialysis (E. Abercrombie, unpublished data)
and 2) increased stress-sensitive tumor growth (Garabal et al
1991). Consequences of daily handling and injection were
therefore avoided in the present study.
Each animal’s body weight was determined before pump
implantation, and the flow rate of mini-pumps (manufacturer’s
specifications) was used to compute the drug concentration
loaded in the pumps to achieve the desired dosage (i.e., DMI,
IMI, FLU, and SER, 10 mg/kg/day; PHE, 5 mg/kg/day; also, an
additional group was treated with SER 25 mg/kg/day to assess
effects of a higher dose of an SSRI). The doses for the different
drugs used in this study were chosen on the basis that the dose
used had been administered to rats via minipump and was found
to be effective in preclinical models for detection of antidepressant activity (West and Weiss, 1995, 1998). We prepared DMI,
IMI, and PHE in sterile 0.9% saline solution, whereas FLU and
SER were administered in 75% polyethylene glycol (PEG;
Sigma) and 25% saline because of their low solubility in water.
Pumps containing 0.9% saline or saline/PEG vehicle were
installed in vehicle-treated control animals. Pumps were implanted subcutaneously under halothane anesthesia in the dorsal
rear flank region, and the wound was closed with stainless steel
clips; this surgery, details of which can be found in West and
BIOL PSYCHIATRY
2001;49:117–129
119
Weiss (1998), required approximately 10 min per animal. Following surgery, animals were returned to the home cage and not
disturbed until removed for the electrophysiologic recording
session.
Electroconvulsive Shock Procedure
Electroconvulsive shock was applied using a 5:1 AC step-up
transformer, with the primary voltage set to deliver 350 V
from the output. With no additional resistance in series other
than the animal, this output generated an approximately 50
mA current when administered to the animal transcranially
(current intensity verified by oscilloscope recording during
ECS delivery). For each ECS treatment, the animal was
anesthetized with halothane for 5 min, after which an ECS
was administered to the anesthetized animal for 500 msec
through ear clips moistened with conducting cream. Rats
treated with ECS immediately exhibited a tonic phase lasting
1 to 5 secs, followed by a clonic phase lasting 20 to 30 secs.
Animals received the ECS treatment five times, with 72 hours
between each treatment. Control animals for ECS treatment
were anesthetized in the same manner as the ECS-treated rats
for the five treatment sessions, but no ECS was given.
Electrophysiologic recording was then conducted either 1 day
or 5 days after the last ECS or control treatment.
Electrophysiologic Recording
Recording of single-unit electrophysiologic activity was carried out under halothane anesthesia as described in Borsody
and Weiss (1996); technical details of the recording process
can be found in that reference. Following anesthesia and
opening of the skull, a glass micropipette was slowly lowered
into the brain in the region of the LC until single-unit activity
was detected. Recording of LC neurons was verified by
criteria described in various references (Aston-Jones and
Bloom 1981; Borsody and Weiss, 1996; Foote et al 1980;
Graham and Aghajanian 1971; Korf et al 1974; Simson and
Weiss 1987). These criteria included 1) a positive–negative
waveform with a notch on the ascending limb (Figure 1, top),
2) a spontaneous firing pattern of 0.5 to 3.5 Hz, and 3)
elicitation of a rapidly occurring succession of depolarization
spikes (burst firing) followed by a period of quiescence
(poststimulus inhibition) in response to application of a salient
sensory stimulus (in this case, compression of the contralateral hind paw; Figure 1, bottom). When a single unit of stable
amplitude demonstrating these characteristics was isolated,
spontaneous rate of depolarization was recorded for 3 min to
obtain this rate. The magnitude of the sensory-evoked response of the neuron was then measured. For this procedure,
the paw was compressed for 1 sec between the ends of a pair
of 13-cm surgical forceps (Malony, curved end [Jarit Surgical
Instrument, Hawthorne, NY]) by applying pressure midway
along the forceps such that the opposite sides of the forceps,
at this midpoint, came into contact. Paw compression applied
in this manner reliably elicits “burst firing” in the anesthetized
rat, as shown in the bottom section of Figure 1. The magnitude
of burst firing (i.e., the number of depolarizations in a burst)
120
M.M. Grant and J.M. Weiss
BIOL PSYCHIATRY
2001;49:117–129
Measurement of Blood Levels of Antidepressant
Drugs
To confirm the presence of drug in circulation at the time of
electrophysiologic measurement, blood levels of the drugs used
were measured. Randomly selected male and female rats that had
undergone electrophysiologic recording were sacrificed by decapitation and truck blood collected. To permit comparison of
blood levels of drug achieved in the present study when drug was
administered by minipump with blood levels when drug is given
by repeated intraperitoneal injection, additional rats (female)
were given daily injections of drug, after which the animals were
sacrificed by decapitation (following anesthetization with halothane) and truck blood was collected. The injection procedure
used attempted to approximate the schedule of Valentino and
colleagues (Curtis and Valentino 1991; Valentino et al 1990;
Valentino and Curtis 1991); thus, drug was injected daily for 21
days, and sacrifice took place either 12 hours or 24 hours after
the final injection (to bracket the time of recording in those
studies, which is stated to have occurred between 12 and 20
hours after last injection). To provide additional information,
blood levels at these time intervals after a single intraperitoneal
(IP) injection of drug were also measured. Serum from samples
was maintained frozen at 280°C until analyzed. The levels of
desipramine and imipramine were determined by the method of
Mazhar and Binder (1989); the levels of fluoxetine, sertraline,
and their metabolites were determined by the method described
in Ritchie and Zhang (1996).
Statistical Analysis
Figure 1. (Top) The positive–negative waveform discharge of a
locus coeruleus (LC) neuron. Note the notch on the ascending
limb. Bar 5 1.0 msec. (Bottom) The typical response pattern of
an LC neuron to brief paw compression (PC) of the contralateral
hind paw. Note that before the PC, the cell fires at a slow, regular
rate in the halothane-anesthetized animal (spontaneous depolarization) and that an increase in depolarization spikes occurs when
PC is applied (sensory-evoked “burst” firing), followed by a
period of quiescence (poststimulus inhibition). Bar 5 1.0 sec;
also indicates period of application of the PC.
Statistical analysis, which was carried out on the activity of the
individual units, was conducted primarily using one-way analyses of variance (ANOVAs). If a significant main effect of
treatment (p , .05) was obtained, the significance of the
difference between each drug or ECS group and the control
condition was then determined using Dunnett’s test. In cases
where frequency distributions between conditions were compared, x2 test was used.
Results
Effects of Antidepressant Drugs
has been shown to be little affected by the intensity of the paw
compression; rather, a burst will occur when the intensity exceeds
the threshold needed to elicit a burst response, with the number of
depolarizations in a burst then being determined by factors unrelated
to paw compression intensity such as the activation of receptors on
the LC neuron, resting potential of the cell, and so forth (Simson and
Weiss 1989). To determine the amount of sensory-evoked burst
firing by a unit, several paw compressions were applied, each
spaced at least 10 sec apart; the average was calculated as the
response of the neuron. Following the last paw compression,
spontaneous activity was recorded for 1 min (to verify that the cell
returned to its normal baseline firing rate) after which the electrode
was slowly moved to isolate another unit. Whenever possible,
several units (3 to 5) were recorded in this manner from an animal.
Mean spontaneous and sensory-evoked burst firing rates
of LC neurons in male and female rats given antidepressant drugs or vehicle (saline or saline/PEG) for 14 and 30
days are shown in Figures 2 and 3. To compare LC firing
rates seen in the drug-treated groups with vehicle-treated
animals, the vehicle-treated condition is represented by the
combined data from all vehicle-treated animals because no
differences were found among any of the different vehicle
regimens (i.e., 14 vs. 30 days of vehicle administration or
use of physiologic saline vs. PEG/saline as vehicle).
As can be seen in Figure 2 and 3, chronic administration
of all antidepressant drugs tested caused a marked reduction in the rate of depolarization of LC neurons. Effects on
Antidepressant Effects on LC Electrophysiology
BIOL PSYCHIATRY
2001;49:117–129
121
Figure 2. Spontaneous depolarization rate (spikes/sec [Hz]) of locus coeruleus (LC) neurons in adult male and female rats treated
chronically with antidepressant drugs. Experimental groups were treated for 14 or 30 days (by subcutaneous minipump) with
desipramine (DMI; 10 mg/kg/day), imipramine (IMI; 10 mg/kg/day), phenelzine (PHE; 5 mg/kg/day), fluoxetine (FLU; 10 mg/kg/day),
or sertraline (SER; 10 mg/kg/day or 25 mg/kg/day). Distributed in the home cages amongst the drug-treated animals, control animals
that received vehicle (VEH) were measured throughout the study; all control conditions (i.e., 14-day and 30-day treatment) are
combined as there were no differences between them. Numbers of LC units represented in each bar of the graph (left to right) are as
follows: for male animals, 39, 14, 10, 12, 11, 11, 7, 15, 10, 9, 11, 11, and 10; for female animals, 34, 8, 14, 19, 8, 12, 9, 15, 6, 29,
16, 16, and 16. Numbers of individual animals from which LC units were recorded in each group (left to right) are as follows: for male
animals, 10, 3, 4, 5, 4, 4, 3, 6, 4, 4, 4, 4, and 4; for female animals, 10, 3, 6, 5, 3, 5, 4, 6, 3, 6, 6, 4, and 4. Means and standard errors
are shown. *Statistically significant (at least p , .05) decrease from the spontaneous rate of the VEH condition of the same gender.
spontaneous rate of depolarization are shown in Figure 2.
When the data for the groups shown in this figure were
analyzed by one-way ANOVA (one ANOVA for male rats
and one for female rats), a significant overall effect
Figure 3. Sensory-evoked “burst” firing (spikes/sec [Hz]) of locus coeruleus neurons. All descriptions as in Figure 2, except that in
this figure * denotes a statistically significant (at least p , .05) decrease from the sensory-evoked burst firing rate of the vehicle (VEH)
condition of the same gender. Because treatment with antidepressant drugs decreased burst firing, such firing did not persist beyond
the initial half-second of the 1.0-sec paw compression for these animals, with the second half-second evidencing “poststimulus
inhibition.” Therefore, the depolarization spikes in the first half-second of the paw compression were counted for all animals (including
VEH) to determine the rate (expressed in Hz) of sensory-evoked burst firing.
122
BIOL PSYCHIATRY
2001;49:117–129
attributable to drug treatment was obtained for both male
rats [F(12,156) 5 15.82, p , .001] and female rats
[F(12,190) 5 14.08, p , .001]. Post hoc comparison of
each individual group with the vehicle-treated condition
showed that every antidepressant drug at either time of
drug administration (i.e., 14 or 30 days) caused a significant suppression of spontaneous firing rate (at least p ,
.05) in both male and female rats.
Effects on the sensory-evoked burst firing are shown in
Figure 3. When the data for these groups were also
analyzed by one way ANOVA (one ANOVA for male rats
and one for female rats), a significant overall effect
attributable to drug treatment was obtained for both male
subjects [F(12,156) 5 11.43, p , .001] and female
subjects [F(12,190) 5 2.56, p , .004]. Post hoc comparison of each individual group with the vehicle-treated
condition showed that every antidepressant drug at either
time of administration (i.e., 14 or 30 days) caused a
significant suppression of sensory-evoked burst firing in
the male rats. In the female rats, however, this post hoc
analysis indicated that seven of the drug groups differed
significantly from the vehicle-treated condition, but five of
the groups did not; the conditions that did not differ
significantly were DMI and IMI at 14 days of administration and IMI, PHE, and SER 25 mg at 30 days. Nonetheless, close inspection of the data suggested that these five
conditions might not have differed significantly from
vehicle treatment because of the presence of a few
fast-firing cells in these conditions. This possibility was
tested by examining frequency distributions for the sensory-evoked burst firing rates of each of these groups, which
are shown in Figure 4. Chi-square analysis comparing the
frequency distribution for each of these drug groups with
that of the vehicle-treated animals showed that the sensory-evoked burst activity of the cells in four of these five
drug groups (IMI 14, IMI 30, PHE 30, and SER 25 mg 30)
differed significantly from that of the vehicle-treated
condition [x2(4) for these comparisons were, respectively,
9.6 (p , .05), 25.6 (p , .001), 18.6 (p , .001), and 11.0
(p , .05)]. As can be seen in Figure 4, these drug
treatments resulted in an appreciable number of cells that
had a low amount of burst firing but also a few cells that
had either a normal or high amount of burst firing that
apparently precluded reaching statistical significance using a parametric statistical test. For the DMI 14 group, the
frequency distribution did not differ significantly from the
vehicle-treated condition, but even the DMI 14 group had
more low-frequency burst firing cells than did the vehicletreated condition.
Effects of Electroconvulsive Shock
Mean spontaneous depolarization rate and sensory-evoked
burst firing of LC neurons in male and female rats that
M.M. Grant and J.M. Weiss
Figure 4. Frequency distribution of sensory-evoked “burst” firing of locus coeruleus neurons in female rats that received
14-day treatment with desipramine (DMI) or imipramine (IMI)
or 30-day treatment with IMI, phenelzine (PHE), or sertraline
(SER; 25 mg/kg/day) compared with the control condition of
female rats treated with vehicle. Cells in each of these treatment
conditions were distributed according to sensory-evoked firing
frequency in bins of 0 –2.0, 2.01– 4.0, 4.01– 6.0, 6.01– 8.0, and
8.01 Hz. Each bar is the percentage of the total number of cells
in the group that fired in that frequency range. *Statistically
significant difference (at least p , .05) in the frequency distribution for the group when compared with vehicles.
received a series of electroconvulsive shocks or all preparative procedures without shock (control rats) are shown
in Figure 5. To compare LC firing rates in animals that
Antidepressant Effects on LC Electrophysiology
123
BIOL PSYCHIATRY
2001;49:117–129
Table 2. Blood Levels of Antidepressant Drugs (ng/mL)
Resulting from Administration by Minipump
Duration of Administration
Drug (dosage)
DMI (10 mg/kg/day)
IMI (10 mg/kg/day)
FLU (10 mg/kg/day)
SER 10 (10 mg/kg/day)
SER 25 (25 mg/kg/day)
14 days
30 days
669.7 6 100.6
553.2 6 152.2
1923.0 6 237.0
369.0 6 132.0
1684.0 6 715.5
1874.2 6 540.7
254.5 6 52.6
2618.8 6 479.0
619.8 6 154.6
2251.8 6 519.4
Blood was taken for measurement following electrophysiologic recording on
the day of administration indicated. Means and standard errors are shown. N 5 4 in
each cell except for 14-day administration of DMI and IMI (n 5 3) and 14-day
administration of FLU (n 5 2). The blood levels (ng/mL serum) reported above are
as follows: 2 for DMI, desipramine; for IMI, imipramine 1 desipramine; for FLU,
norfluoxetine 1 fluoxetine; for SER, des-sertraline 1 sertraline.
Figure 5. Effects of electroconvulsive shock (ECS) on spontaneous depolarization rate and sensory-evoked “burst” firing of
locus coeruleus (LC) neurons. Adult male and female rats
received ECSs (a series of five shocks under halothane anesthesia spaced 72 hours apart) whereas control rats received similar
treatment without shock being given; electrophysiologic recording then took place 1 day (11 day) or 5 days (15 days) after the
last ECS or control procedure. Control rats measured 1 day or 5
days after final treatment are combined as there were no
differences between them. Numbers of LC units represented in
each bar of the graph (left to right) are as follows: for male rats,
39, 10, and 8; for female rats, 49, 10, and 13. Numbers of
individual animals from which LC units were recorded in each
group (left to right) are as follows: for male rats, 8, 4, and 2; for
female rats, 9, 5, and 4. Means and standard errors are shown.
*Statistically significant (at least p , .05) decrease from control
rats of the same gender.
received ECS with the control animals, the data from all
non-ECS animals were combined because no differences
were found between such animals tested either 1 day or 5
days after having received preparatory procedures without
ECS.
As can be seen in Figure 5, a series of electroconvulsive
shocks caused a reduction in the rate of depolarization of
LC neurons. The effect on spontaneous depolarization rate
is shown in the top part of Figure 5. When the data for
these groups were analyzed by one-way ANOVA (one
ANOVA for male rats and one for female rats), a significant overall effect attributable to treatment was obtained
for male rats [F(2,53) 5 19.13, p , .001] and female rats
[F(2,68) 5 23.02, p , .001]. Post hoc comparison of each
ECS-treated group with the control condition showed that
ECS treatment resulted in a significant reduction of
spontaneous depolarization rate at each time of measurement (1- or 5-day delay) in both male and female rats.
The effect of ECS on sensory-evoked burst firing is
shown in the bottom part of Figure 5. When the data for
these groups were analyzed by one-way ANOVA (one
ANOVA for male rats and one for female rats), a significant overall effect attributable to treatment was obtained
for male rats [F(2,53) 5 29.13, p , .001] and female rats
[F(2,68) 5 26.06, p , .001]. Post hoc comparison of each
ECS-treated group with the control condition showed that
ECS treatment resulted in a significant suppression of
burst firing at each time of measurement (1- and 5-day
delay) in both male and female rats.
Blood Levels of Antidepressant Drugs
To establish that antidepressant drugs were in circulation
at the time electrophysiologic measures were taken and
also to make possible informed comparison of the blood
levels achieved by minipump administration with those
achieved by repeated IP injection, blood levels of different
antidepressant drugs were measured. Table 2 shows blood
levels of the antidepressant drugs used in this study
Table 3. Blood Levels of Antidepressant Drugs (ng/mL)
Resulting from Administration by Intraperitoneal Injection
Duration of administration
Time of measurement
after last injection
DMI (10 mg/kg)
SER (10 mg/kg)
12
24
12
24
hours
hours
hours
hours
Chronic
(i.e., daily
injection
for 21 days)
Acute
(i.e., a single
injection)
573.8 6 277.4
422.0 6 85.8
765.0 6 172.9
287.0 6 85.8
334.5 6 181.9
295.2 6 204.0
329.5 6 85.8
172.5 6 20.0
DMI, desipramine; SER, sertraline. For details, see Table 2. N 5 4 for each cell
except for chronic administration of SER (n 5 3).
124
BIOL PSYCHIATRY
2001;49:117–129
(except for phenelzine, which was not assessed because of
the unavailability of the method in our clinical laboratory).
The blood levels shown in Table 2 were those present after
administration of drug by osmotic minipump for 14 or 30
days. Blood levels were also determined after repeated IP
injection of drug to permit comparison of levels produced
by this technique with those resulting from minipump
administration. Table 3 shows blood levels of DMI and
SER following repeated IP injection. The blood levels
shown in Table 3 were those present at 12 and 24 hours
after the final injection of 21 days of daily drug injections
and also at these times after a single IP injection.
Discussion
Several effective antidepressant treatments— chronic administration of five antidepressant drugs (two tricyclics,
two SSRIs, and an MAO inhibitor) and a series of
electroconvulsive shocks—all decreased electrophysiologic activity of LC neurons in halothane-anesthetized
rats. Spontaneous depolarization of LC neurons was markedly decreased by all of the treatments tested. Sensoryevoked burst firing was also decreased, although in one of
the groups (female rats that received DMI for 14 days), the
change in this parameter did not reach statistical significance. It should be noted that despite the size and
consistency of these effects, the magnitude of the changes
shown in Figures 2 through 5 quite possibly underestimates the decrease in LC activity produced by the antidepressant treatments. This is because LC units were difficult to isolate and quantify in some of the animals that
received antidepressant treatments. As a result, some of
the animals that received an antidepressant treatment did
not contribute to the data reported here because, during
recording from these animals, repeated passes of the
microelectrode resulted in no units that fired consistently
or rapidly enough to be designated as LC neurons by the
electrophysiologic criteria used (described in Electrophysiologic Recording in Methods and Materials). In
several cases, a total absence of discernable single-unit
electrophysiologic activity was encountered. Specifically,
whereas LC single units were successfully isolated and
quantified in all 37 of the vehicle-treated and control
animals used during the course of the study, 23 out of 140
animals that received an antidepressant treatment were not
represented in the data because of failure to isolate and
measure any LC units in these animals. This difference
between control and antidepressant-treated conditions is
highly significant (p , .01 by x2). All antidepressanttreated conditions contributed to this result (i.e., animals in
which no units could be isolated/animals from which
recording was attempted 5 3/19 DMI; 5/22 IMI; 3/19
PHE; 4/23 FLU; 2/22 SER 10 mg; 2/18 SER 25 mg; and
M.M. Grant and J.M. Weiss
4/19 ECS). It can be suggested that the animals in which
no units could be isolated were ones in which electrophysiologic activity of LC neurons was so suppressed that
their LC neurons did not depolarize sufficiently to permit
identification of LC activity by the criteria employed. In
this event, exclusion of these subjects from the data set
means that the results shown above underestimate somewhat the extent to which LC activity was suppressed by
these treatments.
Mechanisms can be suggested by which the antidepressant treatments tested produced inhibition of LC activity.
Decreased LC activity caused by drugs that block NE
reuptake (DMI and IMI) is consistent with a strong
inhibitory action on LC firing that is exerted via stimulation of somatodendritic a2 receptors on LC neurons
(Aghajanian and Vandermaelen 1982; Cedarbaum and
Aghajanian 1976; Simson and Weiss 1987). Blockade of
NE reuptake and resultant elevation of NE in the LC
region will increase stimulation of these somatodendritic
a2 receptors to potently inhibit depolarization of LC
neurons. A similar influence can be suggested to account
for the effects of SSRIs on LC activity. Serotonergic
receptors are found on LC cell bodies (Pickel et al 1977),
and activation of these receptors also inhibits LC activity
(Segal 1979); thus, blockade of reuptake of 5-HT released
from serotonergic terminals in the LC region will increase
stimulation of these 5-HT receptors to inhibit LC depolarization. Because MAO inhibition increases the extracellular levels of both NE and 5-HT, the MAOI tested (PHE)
could decrease LC activity by increasing stimulation of
both inhibitory adrenergic and serotonergic receptors on
LC neurons. Whereas the effects of the drugs tested are not
surprising, it is noteworthy that ECS also produced a
decrease in activity of LC neurons at two time points after
the final ECS had been given (i.e., 24 hours and 5 days
afterward). The mechanism by which a series of electroconvulsive shocks causes a persisting inhibition of LC
activity is not evident at this time and remains to be
elucidated.
The findings reported here—namely, that antidepressant
treatments decrease LC activity—are consistent with what
other investigators have seen after acute administration of
antidepressant drugs (e.g., Blier and de Montigny 1985;
Nyback et al 1975; Scuvee-Moreau and Dresse 1979;
McMillen et al 1980; Valentino and Curtis 1991; Valentino et al 1990) and also with most findings of the more
limited number of studies that have assessed effects of
chronic administration (summarized in Table 1). The most
notable exception to the results reported here and the
findings of others is the observation by Valentino et al
(1990) that chronic administration of DMI did not decrease either spontaneous or sensory-evoked LC activity.
These investigators also did not find that chronic admin-
Antidepressant Effects on LC Electrophysiology
istration of the SSRI sertraline changed evoked activity of
LC neurons as was observed in the present study. They did
observe, however, that spontaneous activity of LC was
decreased by chronic administration of the MAO inhibitor
phenelzine and also sertraline if comparison is made with
rats that received a similar series of vehicle injections.
Commenting on why decreased LC firing was not seen
after chronic administration of DMI whereas it was observed after acute DMI, Valentino and colleagues suggested that “tolerance” develops to the acute effect of the
drugs. In the case of DMI that potently blocks NE
reuptake, a likely mechanism for “tolerance” would be
downregulation of the inhibitory somatodendritic a2-receptors on LC neurons so that firing rate of LC neurons
returns to normal despite elevation of NE in synapses in
the LC region. The same rationale could be applied to the
SSRI; chronic administration could be thought to downregulate the inhibitory 5-HT receptors on LC neurons to
reduce inhibitory effects. It was with this hypothesis in
mind that we administered drug for 30 days as well as for
14 days in our studies. The rationale was as follows: if
drug administration for 14 days resulted in a decrease in
LC single-unit activity, it might be that drug administration for this length of time was not sufficient to cause
“tolerance” to occur (e.g., was not sufficient to produce
enough downregulation of presynaptic inhibitory receptors
to bring about recovery of LC activity). Were this the case,
however, administering drug for a longer period (i.e., 30
days) should either produce the expected recovery or, at
the least, cause some recovery from the amount of
decrease seen with 14 days of drug administration as an
indication that the proposed “tolerance” process was
underway. The results found in the present study indicated
that drug administration for 30 days did not bring about
recovery of LC activity. Moreover, inhibition of LC
activity was often more pronounced with 30 days of drug
administration than it was with 14 days, so that there was
no clear indication that a recovery process was underway
that would eventually return LC activity to a normal level.
It should be noted that we do not contend that downregulation of inhibitory presynaptic receptors does not occur
with chronic stimulation of these receptors; rather, our
conclusion is that when one delivers drug to rats via
minipump so that drug is clearly in circulation at time of
electrophysiologic measurement (see the following two
paragraphs), the findings indicate that downregulation is
not sufficient to cause return of LC activity to a normal
level, and, consequently, antidepressant drug administration results in an ongoing and pervasive decrease of LC
activity.
To try to explain why we and other researchers observed decreases in LC activity with chronic administration of certain antidepressant drugs whereas Valentino and
BIOL PSYCHIATRY
2001;49:117–129
125
colleagues did not, we considered that blood levels of drug
might be a significant factor. In the present study, antidepressant drugs were administered by minipump not only to
avoid subjecting the animals to repeated injections but also
to have constant infusion of drug to insure that drug was in
circulation at the time electrophysiologic recordings were
made. The latter consideration was thought to be important
because clinical practice indicates that sufficient circulating levels of drug should be attained to produce a
therapeutic result (Orsulak 1986; Preskorn 1989; Preskorn
and Fast 1991; Van Brunt 1983). In contrast to minipump
administration, Valentino and colleagues (i.e., Curtis and
Valentino 1991; Valentino and Curtis 1991; Valentino et
al 1990) gave antidepressant drugs by daily intraperitoneal
injection for 21 days (usually 10 mg/kg) and then recorded
LC activity 12 to 20 hours after the final injection. We
therefore measured blood levels of drug to assess the
possibility that these different methods of drug administration resulted in different circulating levels of drug.
Table 2 shows blood levels of several of the antidepressants used in the present study (DMI, IMI, FLU, and SER)
resulting from minipump administration for 14 and 30
days. Relative to clinically effective levels reported for
humans (Orsulak 1986; Preskorn 1989; Van Brunt 1983),
the levels produced by minipump administration to the rats
were high. Nonetheless, the data shown in Table 2 make
clear that the antidepressant drugs used certainly were in
circulation at the time of electrophysiologic recording. (It
should be noted that although blood levels were high, no
rat in the study died during drug administration or was
observed to have seizures or become ataxic. This included
the 60 rats that received a tricyclic or MAOI, so that the
high blood levels were not associated with evidence of
cardiotoxicity that occurs in humans when abnormally
elevated levels of TCA and MAOIs are present.) In
comparison with these values produced by minipump
administration, Table 3 shows blood levels of DMI and
SER in rats that received drug by IP injection (10 mg/kg)
daily for 21 days, with measurements made at 12 and 24
hours after the final injection (to bracket the time during
that Valentino and colleagues reported they took their
electrophysiologic measures). (Valentino and colleagues,
as well as other previous studies of electrophysiologic
responses of LC neurons after antidepressant administration, did not include blood levels of antidepressant drugs
in the publications.) Comparison of the values in the left
column of Table 3 (i.e., blood levels after chronic administration; levels after a single injection were also measured
and are shown in the right column) with those for the same
drugs in Table 2 indicate that daily injection for 21 days
produced circulating levels of drug between 12 and 24
hours postinjection that were similar to what was obtained
with minipump administration. Thus, differences between
126
M.M. Grant and J.M. Weiss
BIOL PSYCHIATRY
2001;49:117–129
Table 4. Summary of the Effects of Chronically Administered Antidepressants on Levels of Cerebrospinal Fluid MHPG
Drug type
Drug
Tricyclic
Amitryptiline
Amitryptiline
Amitryptiline
Chlorimipramine
Monoamine
oxidase inhibitor
Effect
Study citation
75–150 mg/day for 2 weeks
Up to 250 mg/day for 4 weeks
250 mg/day for 4 weeks
150 mg/day for 3 weeks
Decrease
Decrease
Decrease
Decrease
Chlorimipramine
Desipramine
Desipramine
Desipramine
Desipramine
Desipramine
(reserpine
augmented)
Imipramine
Nortryptiline
For 3 and 6 weeks
150 mg/day for 6 weeks
181 mg/day for 5 weeks
100 –300 mg/day for 3– 4 weeks
100 –300 mg/day for 4 – 6 weeks
2.5 mg/kg/day for 4 weeks;
reserpine 5 mg/b.i.d.
Decrease
Decreasea
Decrease
Decrease
Decrease
Decrease
Mendelwicz et al 1982
Maas et al 1984
Bowden et al 1985
Bertilsson et al 1974;
Thoren et al 1980;
Traskman et al 1979
Martensson et al 1991
Dahl et al 1982
Sharma et al 1994
Potter et al 1981
Potter et al 1985
Price et al 1987
250 mg/day for 4 weeks
150 mg/day for 3 weeks
Decrease
Decrease
Clorgyline
20 –30 mg/day
for 3 weeks
5–10 mg/day
for 5– 8 weeks
75–100 mg/day
for 3 weeks
60 mg/day for 3 weeks
600 mg/day for 6 weeks
Decrease
Bowden et al 1985
Bertilsson et al 1974;
Thoren et al 1980
Major et al 1979
Decrease
Potter et al 1985
Decrease
Major et al 1979
Decrease
Decreasea
Sunderland et al 1994
Dahl et al 1982
50 –300 mg/day
for 3 weeks
200 mg/day for 3 weeks
Up to 300 mg/day
for 4 – 6 weeks
200 mg/day for 3– 4 weeks
Unstated
30 – 60 mg/day for 2 weeks
Decrease
Bertilsson et al 1980
Decrease
Decreasea
Wallinder 1981
Potter 1981
Decrease
Decreasea
Decrease
Potter et al 1985
Hsiao et al 1987
Mendelwicz et al 1982
Clorgyline
Pargyline
Selective serotonin
reuptake inhibitor
Selegiline
Femoxetine
Zimelidine
Zimelidine
Zimelidine
Aminoketone
Atypical
a
Zimelidine
Buproprion
Mianserin
Dosage
Not statistically significant; all effects not so noted are statistically significant.
the electrophysiologic findings of Valentino and colleagues compared with those of the present study and
other investigators do not appear explainable on the basis
of blood levels of drug at the time of electrophysiologic
recording. At present, we are unable to account for the
difference between results of Valentino and colleagues in
comparison with others.
Although measuring blood levels of drug failed to
clarify why one group of investigators obtained results that
differ from the findings of the others, there is significant
additional information that bears on the question of
whether antidepressant treatment decreases LC activity. A
number of human studies have attempted to assess noradrenergic activity in human brain by measuring the noradrenergic metabolite MHPG in cerebrospinal fluid (CSF) in
conjunction with antidepressant treatment. These studies,
summarized in Table 4, have consistently found decreased
MHPG in CSF of human patients taking antidepressant
drugs. Although MHPG in CSF does not exclusively
reflect NE release in brain because MHPG in blood can
diffuse freely into CSF (Kopin 1985), Scheinin reviewed
this issue and concluded, “Drug induced alterations in the
concentration of monoamine metabolites in human CSF
probably reflect quite accurately the effects of several
classes of drugs [inhibitors of MAO and antidepressant
drugs included in this list] on monoamine turnover in the
CNS” (Scheinin 1985, 10). One study (Sharma et al 1994)
utilized a correction of CSF MHPG for potential contribution by plasma MHPG that was recommended by Kopin
(1985) and reached the same conclusion as other studies
that CSF MHPG was reduced as a result of antidepressant
drug treatment. Considering that approximately 70% of
the NE in brain derives from the LC, the findings shown
in Table 4 are consistent with the influence of antidepressant drugs in humans decreasing LC neural activity, as was
observed in the rats in the present study. Moreover, these
data from humans also suggest that despite the high blood
levels of antidepressant drug found in the animals of the
Antidepressant Effects on LC Electrophysiology
present study, the results reported here are relevant to what
occurs in humans undergoing antidepressant treatment.
The finding reported here that all antidepressant treatments tested cause a marked decrease in LC neuronal
activity leads to the question of how such a change might
be important for effective antidepressant treatment. Recently, other investigators have suggested that LC neuronal activity is abnormally elevated in depression and that
antidepressant action consequently may occur through
reduction of LC neuronal activity (Gold and Chrousos
1999; Zhu et al 1999). Of course, in considering the
mechanism by which reduced LC activity might produce
an antidepressant effect, it should be stated at the outset
that decreased LC neuronal activity resulting from antidepressant treatment could be an epiphenomenon unrelated
to antidepressant action. For example, LC neuronal activity might be decreased simply as a homeostatic adjustment
because antidepressants increase NE in synapses throughout the brain and that potentiation of NE action throughout
the brain (or resulting downregulation of B-adrenergic
receptors) is what actually underlies antidepressant action.
Also, the linkage of decreased LC activity to depression is
possibly, or even probably, not a simple one, but, rather,
the result of complex influences on neural transmission
and information processing (e.g., Mongeau et al 1997). On
the other hand, the study described in this article was
prompted by a hypothesis we have offered recently that
links LC activity directly to certain behavioral changes
and processes seen in depression (Weiss et al 1996, 1998).
This hypothesis suggests that certain salient behavioral
changes seen in depression, particularly psychomotor
retardation and failure to appreciate rewarding stimuli, can
arise from abnormally high activity of LC neurons. The
proposed mechanism is as follows: when LC neurons are
highly active, the peptide galanin, which is colocalized in
LC neurons, will be released from the LC terminals in
addition to NE. When galanin release occurs from LC
terminals in the ventral tegmentum, this peptide will exert
a hyperpolarizing influence on dopaminergic cell bodies in
this brain region, resulting in decreased release of dopamine in the forebrain. This in turn is hypothesized to cause
behavioral changes that occur in depression, particularly
psychomotor retardation and anhedonia. With respect to
the findings described here, this hypothesis suggests that
decreasing LC activity will reduce galanin release from
LC terminals in the ventral tegmentum, thereby releasing
dopaminergic cell bodies in this region from galaninmediated inhibition. A reduction in LC activity, possibly
through the mechanism described here, is proposed to
underlie, at least in part, the e
and Electroconvulsive Shock on Locus Coeruleus
Electrophysiologic Activity
Michael M. Grant and Jay M. Weiss
Background: The locus coeruleus (LC) is the major
noradrenergic cell body group in the brain. Although
previous studies have examined changes in electrophysiologic activity of LC neurons produced by antidepressant
drugs, only a small number have examined changes that
occur with chronic drug administration, which is the
therapeutically effective regimen, and only one group of
investigators has assessed effects on activated (or
“burst”) firing of LC neurons under such treatment
conditions. The present study assessed changes produced
in rats by effective antidepressant treatments—several
drugs given chronically (two tricyclic antidepressants, two
selective serotonin reuptake inhibitors, and a monoamine
oxidase inhibitor) as well as a series of electroconvulsive
shocks (ECSs)—in single-unit electrophysiologic activity
of LC neurons, measuring effects on spontaneous depolarization rate and also on sensory-evoked burst firing.
Methods: Drugs were administered via osmotic
minipumps for either 14 or 30 days; ECSs were administered five times, with a 72-hour interval between each
administration. Electrophysiologic recording of LC activity took place under halothane anesthesia on the last day
of drug treatment or following a delay of 1 or 5 days after
the final ECS.
Results: A common effect of all drugs tested and ECS
treatment was to decrease LC spontaneous and sensoryevoked burst firing.
Conclusions: The clinical efficacy of antidepressant medication and ECS may be mediated, in part, through
reduction of LC neural activity. The findings reported here
are consistent with recent indications that LC neurons are
hyperactive in depressed individuals and with suggestions
that some behavioral changes seen in depression can arise
from consequences of rapidly depolarizing LC terminals,
such as release of peptides. Biol Psychiatry 2001;49:
117–129 © 2001 Society of Biological Psychiatry
From the Department of Psychology, Georgia State University (MMG) and the
Department of Psychiatry and Behavioral Sciences, Emory University School
of Medicine (JMW), Atlanta, Georgia.
Address reprint requests to Jay M. Weiss, Ph.D., Emory University School of
Medicine, Emory West Campus, 1256 Briarcliff Road NE, Atlanta GA 30306.
Received January 22, 2000; revised May 2, 2000; accepted May 16, 2000.
© 2001 Society of Biological Psychiatry
Key Words: Antidepressant, locus coeruleus, electrophysiology, tricyclic, SSRI, electroconvulsive shock
Introduction
T
he suggestion that noradrenergic neurons in the brain
are involved in depression has been with us for many
years. The original catecholamine hypothesis of depression, advanced more than 30 years ago (Bunney and Davis
1965; Schildkraut 1965; Schildkraut and Kety 1967),
proposed that depression arose from a deficiency of
norepinephrine (NE) in the brain. But when various
predictions derived from this hypothesis were not confirmed (e.g., depressed individuals did not appear to have
lower levels of NE or metabolites in brain, and tricyclic
antidepressants required many days to reverse depression
despite elevating synaptic NE immediately after administration), initial enthusiasm for the catecholamine hypothesis gave way to attempts to develop formulations that did
not focus on NE.
Despite prodigious efforts to develop alternatives, evidence has continued to point to the involvement of NE in
depression and the action of antidepressant drugs. For
example, when effective antidepressant drugs were discovered that did not block NE reuptake or directly increase
NE release, this seemed to further compromise the catecholamine hypothesis until it was subsequently discovered
that these drugs downregulate beta-adrenergic receptors,
an action that could derive from, or was a consequence
equivalent to, prolonged action of NE in the synapse
(Vetulani et al 1976a, 1976b; Vetulani and Sulser 1975). A
survey of antidepressant drugs that appeared shortly thereafter concluded that almost all antidepressant drugs potentiated NE action in the synapse (Richelson and Pfenning
1984). Even drugs developed more recently that seem to
suggest the primacy of neurotransmitters other than NE
have been found to significantly affect noradrenergic
systems. For instance, the antidepressant drugs classified
as selective serotonin reuptake inhibitors (SSRIs) produce
marked increases in extracellular NE in the brain as well
0006-3223/01/$20.00
PII S0006-3223(00)00936-7
118
M.M. Grant and J.M. Weiss
BIOL PSYCHIATRY
2001;49:117–129
Table 1. Effects of Chronic Treatment (1 Week or Longer) with Antidepressant Drugs on
Spontaneous and Sensory-Evoked Depolarization Rate of Locus Coeruleus Neurons
Spontaneous
Tricyclic antidepressants
Desipramine
Imipramine
Monoamine oxidase inhibitor
Phenelzine
Huang et al 1980
McMillen et al 1980
– Valentino et al 1990
Svensson and Usdin 1978
Blier and de Montigny 1985
Valentino and Curtis 1991
Evoked
– Valentino et al 1990
? Valentino and Curtis 1991a
Selective serotonin reuptake inhibitor
Sertraline
– Valentino et al 1990b
– Valentino et al 1990
Atypical antidepressant
Mianserin
–? Curtis and Valentino 1991c
–? Curtis and Valentino 1991c
The depolarization rate shown by animals treated with drug in comparison with animals receiving no drug was , decreased,
or 2, unchanged.
a
Evoked response assessed, but results not reported for this measure. Report presents only the ratio of evoked to spontaneous
rate; because spontaneous rate was decreased by phenelzine, effect on evoked rate could not be determined from data presented.
b
Spontaneous firing rate of sertraline-treated rats was not different from untreated rats but was significantly lower than that
of rats that received a similar schedule of injections of vehicle.
c
Spontaneous and evoked firing rate of mianserin-treated rats was not different from untreated rats but was less than that of
rats that received a similar schedule of injections of vehicle; however, the statistical significance of the comparison of
mianserin-treated rats with vehicle-treated rats was not presented (see Curtis and Valentino 1991 [Figure 4, 334].
as serotonin when given systemically as occurs in
clinical use (Cosford 1995; Jordan et al 1994). Furthermore, the antidepressant drug bupropion blocks dopamine reuptake more effectively than any other antidepressant drug presently in widespread use, but
preclinical studies have found that antidepressantlike
behavioral changes produced by bupropion parallel
changes in brain NE produced by this drug rather than
changes in dopamine (Cooper et al 1994). Yet another
example implicating NE is provided by the Flinders
Sensitive rat strain that was selectively bred for sensitivity to cholinergic drugs and has been proposed as a
model for depression (reviewed in Overstreet 1993).
Although these animals were bred for sensitivity to
cholinergic drugs, their depression-related behavior has
been found to respond to drugs that affect brain NE
rather than acetylcholine (Schiller et al 1992). In
summary, despite various inconsistencies regarding the
role of NE in depression, research continues to produce
evidence that activity of noradrenergic neurons is important in this disorder.
The study described here examined changes in electrophysiologic single-unit activity of neurons of the major
noradrenergic cell-body group in the brain, the locus
coeruleus (LC). This study assessed the changes produced
by effective antidepressant treatments, both chronic administration of antidepressant drugs and electroconvulsive
shock. The terminals of LC neurons supply nearly 70% on
the total NE in the brain, giving rise to most of the NE in
the forebrain and all of the NE in cortex and hippocampus.
Studies have examined the effects of antidepressant drugs
on electrophysiologic activity of LC neurons, but most of
these studies have reported effects of acute drug administration. Because antidepressant drugs require prolonged or
chronic administration to have therapeutic effects, changes
of greater clinical interest would be those observed after
these drugs have been given for a therapeutically relevant
period of time. Also, LC neurons have two (or at least two)
modes of firing: they depolarize at a spontaneous rate but
also fire more rapidly in response to excitatory stimuli.
Few studies have examined effects of antidepressant
treatment on rapid firing (or “burst” firing) that occurs in
response to an excitatory stimulus. The previously reported effects of chronic administration of antidepressant
drugs on spontaneous depolarization rate and burst firing
to an excitatory stimulus are summarized in Table 1.
Examination of this table will reveal that the reported
effects have not been consistent. To clarify these effects,
we assessed spontaneous depolarization rate and sensoryevoked burst firing in rats treated chronically with several
different types of antidepressant drugs or a series of
electroconvulsive shocks.
Methods And Materials
Subjects
Male and female Sprague-Dawley rats (virus/antigen free, bred
in our laboratory from stock originally obtained from Charles
Antidepressant Effects on LC Electrophysiology
River [Wilmington, MA]) aged 5 to 6 months were used.
Animals were group housed (three to a cage) directly on bedding
in solid-bottom polycarbonate cages. In studies of antidepressant
drugs, animals in any single cage received the same drug, with
control animals (i.e., animals receiving vehicle) distributed
among drug groups and housed with them. Animals that received
ECS and control animals for this manipulation were similarly
housed. Food (lab chow) and water were available ad libitum. A
12:12 light:dark cycle and temperature of approximately 21°C
was maintained in the colony room.
Drugs and Administration
The effects of five antidepressants were examined: desipramine
HCl (DMI), a secondary amine dibenzazepine tricyclic NE
reuptake inhibitor (Sigma, St. Louis); imipramine HCl (IMI), a
tertiary amine dibenzazepine tricyclic NE reuptake inhibitor
(Sigma); phenelzine sulfate (PHE), a nonselective monoamine
oxidase inhibitor (Sigma); and fluoxetine HCl (FLU) and sertraline HCl (SER), both heterocyclic SSRIs (Lilly and Pfizer,
respectively). Antidepressants were administered via Alzet Osmotic Minipumps (Alza, Mountain View, CA), implanted subcutaneously in the animals, using Model 2ML2 pumps for 14-day
administration and Model 2ML4 pumps for 30-day administration. Minipumps, which provide continuous drug delivery beginning approximately 4 hours after pump implantation, were used
to insure the presence of drug in the animal throughout the study
and at the time that electrophysiologic recording took place. Use
of minipumps also eliminated the need for repeated handling and
injection of animals to administer drug chronically. Studies that
have assessed effects of daily injection of vehicle while studying
effects of antidepressant drugs have found that repeated injections constitute a stressful procedure. For example, compared
with no injections, repeated vehicle injections have been observed to produce 1) elevated brain norepinephrine release as
measured by microdialysis (E. Abercrombie, unpublished data)
and 2) increased stress-sensitive tumor growth (Garabal et al
1991). Consequences of daily handling and injection were
therefore avoided in the present study.
Each animal’s body weight was determined before pump
implantation, and the flow rate of mini-pumps (manufacturer’s
specifications) was used to compute the drug concentration
loaded in the pumps to achieve the desired dosage (i.e., DMI,
IMI, FLU, and SER, 10 mg/kg/day; PHE, 5 mg/kg/day; also, an
additional group was treated with SER 25 mg/kg/day to assess
effects of a higher dose of an SSRI). The doses for the different
drugs used in this study were chosen on the basis that the dose
used had been administered to rats via minipump and was found
to be effective in preclinical models for detection of antidepressant activity (West and Weiss, 1995, 1998). We prepared DMI,
IMI, and PHE in sterile 0.9% saline solution, whereas FLU and
SER were administered in 75% polyethylene glycol (PEG;
Sigma) and 25% saline because of their low solubility in water.
Pumps containing 0.9% saline or saline/PEG vehicle were
installed in vehicle-treated control animals. Pumps were implanted subcutaneously under halothane anesthesia in the dorsal
rear flank region, and the wound was closed with stainless steel
clips; this surgery, details of which can be found in West and
BIOL PSYCHIATRY
2001;49:117–129
119
Weiss (1998), required approximately 10 min per animal. Following surgery, animals were returned to the home cage and not
disturbed until removed for the electrophysiologic recording
session.
Electroconvulsive Shock Procedure
Electroconvulsive shock was applied using a 5:1 AC step-up
transformer, with the primary voltage set to deliver 350 V
from the output. With no additional resistance in series other
than the animal, this output generated an approximately 50
mA current when administered to the animal transcranially
(current intensity verified by oscilloscope recording during
ECS delivery). For each ECS treatment, the animal was
anesthetized with halothane for 5 min, after which an ECS
was administered to the anesthetized animal for 500 msec
through ear clips moistened with conducting cream. Rats
treated with ECS immediately exhibited a tonic phase lasting
1 to 5 secs, followed by a clonic phase lasting 20 to 30 secs.
Animals received the ECS treatment five times, with 72 hours
between each treatment. Control animals for ECS treatment
were anesthetized in the same manner as the ECS-treated rats
for the five treatment sessions, but no ECS was given.
Electrophysiologic recording was then conducted either 1 day
or 5 days after the last ECS or control treatment.
Electrophysiologic Recording
Recording of single-unit electrophysiologic activity was carried out under halothane anesthesia as described in Borsody
and Weiss (1996); technical details of the recording process
can be found in that reference. Following anesthesia and
opening of the skull, a glass micropipette was slowly lowered
into the brain in the region of the LC until single-unit activity
was detected. Recording of LC neurons was verified by
criteria described in various references (Aston-Jones and
Bloom 1981; Borsody and Weiss, 1996; Foote et al 1980;
Graham and Aghajanian 1971; Korf et al 1974; Simson and
Weiss 1987). These criteria included 1) a positive–negative
waveform with a notch on the ascending limb (Figure 1, top),
2) a spontaneous firing pattern of 0.5 to 3.5 Hz, and 3)
elicitation of a rapidly occurring succession of depolarization
spikes (burst firing) followed by a period of quiescence
(poststimulus inhibition) in response to application of a salient
sensory stimulus (in this case, compression of the contralateral hind paw; Figure 1, bottom). When a single unit of stable
amplitude demonstrating these characteristics was isolated,
spontaneous rate of depolarization was recorded for 3 min to
obtain this rate. The magnitude of the sensory-evoked response of the neuron was then measured. For this procedure,
the paw was compressed for 1 sec between the ends of a pair
of 13-cm surgical forceps (Malony, curved end [Jarit Surgical
Instrument, Hawthorne, NY]) by applying pressure midway
along the forceps such that the opposite sides of the forceps,
at this midpoint, came into contact. Paw compression applied
in this manner reliably elicits “burst firing” in the anesthetized
rat, as shown in the bottom section of Figure 1. The magnitude
of burst firing (i.e., the number of depolarizations in a burst)
120
M.M. Grant and J.M. Weiss
BIOL PSYCHIATRY
2001;49:117–129
Measurement of Blood Levels of Antidepressant
Drugs
To confirm the presence of drug in circulation at the time of
electrophysiologic measurement, blood levels of the drugs used
were measured. Randomly selected male and female rats that had
undergone electrophysiologic recording were sacrificed by decapitation and truck blood collected. To permit comparison of
blood levels of drug achieved in the present study when drug was
administered by minipump with blood levels when drug is given
by repeated intraperitoneal injection, additional rats (female)
were given daily injections of drug, after which the animals were
sacrificed by decapitation (following anesthetization with halothane) and truck blood was collected. The injection procedure
used attempted to approximate the schedule of Valentino and
colleagues (Curtis and Valentino 1991; Valentino et al 1990;
Valentino and Curtis 1991); thus, drug was injected daily for 21
days, and sacrifice took place either 12 hours or 24 hours after
the final injection (to bracket the time of recording in those
studies, which is stated to have occurred between 12 and 20
hours after last injection). To provide additional information,
blood levels at these time intervals after a single intraperitoneal
(IP) injection of drug were also measured. Serum from samples
was maintained frozen at 280°C until analyzed. The levels of
desipramine and imipramine were determined by the method of
Mazhar and Binder (1989); the levels of fluoxetine, sertraline,
and their metabolites were determined by the method described
in Ritchie and Zhang (1996).
Statistical Analysis
Figure 1. (Top) The positive–negative waveform discharge of a
locus coeruleus (LC) neuron. Note the notch on the ascending
limb. Bar 5 1.0 msec. (Bottom) The typical response pattern of
an LC neuron to brief paw compression (PC) of the contralateral
hind paw. Note that before the PC, the cell fires at a slow, regular
rate in the halothane-anesthetized animal (spontaneous depolarization) and that an increase in depolarization spikes occurs when
PC is applied (sensory-evoked “burst” firing), followed by a
period of quiescence (poststimulus inhibition). Bar 5 1.0 sec;
also indicates period of application of the PC.
Statistical analysis, which was carried out on the activity of the
individual units, was conducted primarily using one-way analyses of variance (ANOVAs). If a significant main effect of
treatment (p , .05) was obtained, the significance of the
difference between each drug or ECS group and the control
condition was then determined using Dunnett’s test. In cases
where frequency distributions between conditions were compared, x2 test was used.
Results
Effects of Antidepressant Drugs
has been shown to be little affected by the intensity of the paw
compression; rather, a burst will occur when the intensity exceeds
the threshold needed to elicit a burst response, with the number of
depolarizations in a burst then being determined by factors unrelated
to paw compression intensity such as the activation of receptors on
the LC neuron, resting potential of the cell, and so forth (Simson and
Weiss 1989). To determine the amount of sensory-evoked burst
firing by a unit, several paw compressions were applied, each
spaced at least 10 sec apart; the average was calculated as the
response of the neuron. Following the last paw compression,
spontaneous activity was recorded for 1 min (to verify that the cell
returned to its normal baseline firing rate) after which the electrode
was slowly moved to isolate another unit. Whenever possible,
several units (3 to 5) were recorded in this manner from an animal.
Mean spontaneous and sensory-evoked burst firing rates
of LC neurons in male and female rats given antidepressant drugs or vehicle (saline or saline/PEG) for 14 and 30
days are shown in Figures 2 and 3. To compare LC firing
rates seen in the drug-treated groups with vehicle-treated
animals, the vehicle-treated condition is represented by the
combined data from all vehicle-treated animals because no
differences were found among any of the different vehicle
regimens (i.e., 14 vs. 30 days of vehicle administration or
use of physiologic saline vs. PEG/saline as vehicle).
As can be seen in Figure 2 and 3, chronic administration
of all antidepressant drugs tested caused a marked reduction in the rate of depolarization of LC neurons. Effects on
Antidepressant Effects on LC Electrophysiology
BIOL PSYCHIATRY
2001;49:117–129
121
Figure 2. Spontaneous depolarization rate (spikes/sec [Hz]) of locus coeruleus (LC) neurons in adult male and female rats treated
chronically with antidepressant drugs. Experimental groups were treated for 14 or 30 days (by subcutaneous minipump) with
desipramine (DMI; 10 mg/kg/day), imipramine (IMI; 10 mg/kg/day), phenelzine (PHE; 5 mg/kg/day), fluoxetine (FLU; 10 mg/kg/day),
or sertraline (SER; 10 mg/kg/day or 25 mg/kg/day). Distributed in the home cages amongst the drug-treated animals, control animals
that received vehicle (VEH) were measured throughout the study; all control conditions (i.e., 14-day and 30-day treatment) are
combined as there were no differences between them. Numbers of LC units represented in each bar of the graph (left to right) are as
follows: for male animals, 39, 14, 10, 12, 11, 11, 7, 15, 10, 9, 11, 11, and 10; for female animals, 34, 8, 14, 19, 8, 12, 9, 15, 6, 29,
16, 16, and 16. Numbers of individual animals from which LC units were recorded in each group (left to right) are as follows: for male
animals, 10, 3, 4, 5, 4, 4, 3, 6, 4, 4, 4, 4, and 4; for female animals, 10, 3, 6, 5, 3, 5, 4, 6, 3, 6, 6, 4, and 4. Means and standard errors
are shown. *Statistically significant (at least p , .05) decrease from the spontaneous rate of the VEH condition of the same gender.
spontaneous rate of depolarization are shown in Figure 2.
When the data for the groups shown in this figure were
analyzed by one-way ANOVA (one ANOVA for male rats
and one for female rats), a significant overall effect
Figure 3. Sensory-evoked “burst” firing (spikes/sec [Hz]) of locus coeruleus neurons. All descriptions as in Figure 2, except that in
this figure * denotes a statistically significant (at least p , .05) decrease from the sensory-evoked burst firing rate of the vehicle (VEH)
condition of the same gender. Because treatment with antidepressant drugs decreased burst firing, such firing did not persist beyond
the initial half-second of the 1.0-sec paw compression for these animals, with the second half-second evidencing “poststimulus
inhibition.” Therefore, the depolarization spikes in the first half-second of the paw compression were counted for all animals (including
VEH) to determine the rate (expressed in Hz) of sensory-evoked burst firing.
122
BIOL PSYCHIATRY
2001;49:117–129
attributable to drug treatment was obtained for both male
rats [F(12,156) 5 15.82, p , .001] and female rats
[F(12,190) 5 14.08, p , .001]. Post hoc comparison of
each individual group with the vehicle-treated condition
showed that every antidepressant drug at either time of
drug administration (i.e., 14 or 30 days) caused a significant suppression of spontaneous firing rate (at least p ,
.05) in both male and female rats.
Effects on the sensory-evoked burst firing are shown in
Figure 3. When the data for these groups were also
analyzed by one way ANOVA (one ANOVA for male rats
and one for female rats), a significant overall effect
attributable to drug treatment was obtained for both male
subjects [F(12,156) 5 11.43, p , .001] and female
subjects [F(12,190) 5 2.56, p , .004]. Post hoc comparison of each individual group with the vehicle-treated
condition showed that every antidepressant drug at either
time of administration (i.e., 14 or 30 days) caused a
significant suppression of sensory-evoked burst firing in
the male rats. In the female rats, however, this post hoc
analysis indicated that seven of the drug groups differed
significantly from the vehicle-treated condition, but five of
the groups did not; the conditions that did not differ
significantly were DMI and IMI at 14 days of administration and IMI, PHE, and SER 25 mg at 30 days. Nonetheless, close inspection of the data suggested that these five
conditions might not have differed significantly from
vehicle treatment because of the presence of a few
fast-firing cells in these conditions. This possibility was
tested by examining frequency distributions for the sensory-evoked burst firing rates of each of these groups, which
are shown in Figure 4. Chi-square analysis comparing the
frequency distribution for each of these drug groups with
that of the vehicle-treated animals showed that the sensory-evoked burst activity of the cells in four of these five
drug groups (IMI 14, IMI 30, PHE 30, and SER 25 mg 30)
differed significantly from that of the vehicle-treated
condition [x2(4) for these comparisons were, respectively,
9.6 (p , .05), 25.6 (p , .001), 18.6 (p , .001), and 11.0
(p , .05)]. As can be seen in Figure 4, these drug
treatments resulted in an appreciable number of cells that
had a low amount of burst firing but also a few cells that
had either a normal or high amount of burst firing that
apparently precluded reaching statistical significance using a parametric statistical test. For the DMI 14 group, the
frequency distribution did not differ significantly from the
vehicle-treated condition, but even the DMI 14 group had
more low-frequency burst firing cells than did the vehicletreated condition.
Effects of Electroconvulsive Shock
Mean spontaneous depolarization rate and sensory-evoked
burst firing of LC neurons in male and female rats that
M.M. Grant and J.M. Weiss
Figure 4. Frequency distribution of sensory-evoked “burst” firing of locus coeruleus neurons in female rats that received
14-day treatment with desipramine (DMI) or imipramine (IMI)
or 30-day treatment with IMI, phenelzine (PHE), or sertraline
(SER; 25 mg/kg/day) compared with the control condition of
female rats treated with vehicle. Cells in each of these treatment
conditions were distributed according to sensory-evoked firing
frequency in bins of 0 –2.0, 2.01– 4.0, 4.01– 6.0, 6.01– 8.0, and
8.01 Hz. Each bar is the percentage of the total number of cells
in the group that fired in that frequency range. *Statistically
significant difference (at least p , .05) in the frequency distribution for the group when compared with vehicles.
received a series of electroconvulsive shocks or all preparative procedures without shock (control rats) are shown
in Figure 5. To compare LC firing rates in animals that
Antidepressant Effects on LC Electrophysiology
123
BIOL PSYCHIATRY
2001;49:117–129
Table 2. Blood Levels of Antidepressant Drugs (ng/mL)
Resulting from Administration by Minipump
Duration of Administration
Drug (dosage)
DMI (10 mg/kg/day)
IMI (10 mg/kg/day)
FLU (10 mg/kg/day)
SER 10 (10 mg/kg/day)
SER 25 (25 mg/kg/day)
14 days
30 days
669.7 6 100.6
553.2 6 152.2
1923.0 6 237.0
369.0 6 132.0
1684.0 6 715.5
1874.2 6 540.7
254.5 6 52.6
2618.8 6 479.0
619.8 6 154.6
2251.8 6 519.4
Blood was taken for measurement following electrophysiologic recording on
the day of administration indicated. Means and standard errors are shown. N 5 4 in
each cell except for 14-day administration of DMI and IMI (n 5 3) and 14-day
administration of FLU (n 5 2). The blood levels (ng/mL serum) reported above are
as follows: 2 for DMI, desipramine; for IMI, imipramine 1 desipramine; for FLU,
norfluoxetine 1 fluoxetine; for SER, des-sertraline 1 sertraline.
Figure 5. Effects of electroconvulsive shock (ECS) on spontaneous depolarization rate and sensory-evoked “burst” firing of
locus coeruleus (LC) neurons. Adult male and female rats
received ECSs (a series of five shocks under halothane anesthesia spaced 72 hours apart) whereas control rats received similar
treatment without shock being given; electrophysiologic recording then took place 1 day (11 day) or 5 days (15 days) after the
last ECS or control procedure. Control rats measured 1 day or 5
days after final treatment are combined as there were no
differences between them. Numbers of LC units represented in
each bar of the graph (left to right) are as follows: for male rats,
39, 10, and 8; for female rats, 49, 10, and 13. Numbers of
individual animals from which LC units were recorded in each
group (left to right) are as follows: for male rats, 8, 4, and 2; for
female rats, 9, 5, and 4. Means and standard errors are shown.
*Statistically significant (at least p , .05) decrease from control
rats of the same gender.
received ECS with the control animals, the data from all
non-ECS animals were combined because no differences
were found between such animals tested either 1 day or 5
days after having received preparatory procedures without
ECS.
As can be seen in Figure 5, a series of electroconvulsive
shocks caused a reduction in the rate of depolarization of
LC neurons. The effect on spontaneous depolarization rate
is shown in the top part of Figure 5. When the data for
these groups were analyzed by one-way ANOVA (one
ANOVA for male rats and one for female rats), a significant overall effect attributable to treatment was obtained
for male rats [F(2,53) 5 19.13, p , .001] and female rats
[F(2,68) 5 23.02, p , .001]. Post hoc comparison of each
ECS-treated group with the control condition showed that
ECS treatment resulted in a significant reduction of
spontaneous depolarization rate at each time of measurement (1- or 5-day delay) in both male and female rats.
The effect of ECS on sensory-evoked burst firing is
shown in the bottom part of Figure 5. When the data for
these groups were analyzed by one-way ANOVA (one
ANOVA for male rats and one for female rats), a significant overall effect attributable to treatment was obtained
for male rats [F(2,53) 5 29.13, p , .001] and female rats
[F(2,68) 5 26.06, p , .001]. Post hoc comparison of each
ECS-treated group with the control condition showed that
ECS treatment resulted in a significant suppression of
burst firing at each time of measurement (1- and 5-day
delay) in both male and female rats.
Blood Levels of Antidepressant Drugs
To establish that antidepressant drugs were in circulation
at the time electrophysiologic measures were taken and
also to make possible informed comparison of the blood
levels achieved by minipump administration with those
achieved by repeated IP injection, blood levels of different
antidepressant drugs were measured. Table 2 shows blood
levels of the antidepressant drugs used in this study
Table 3. Blood Levels of Antidepressant Drugs (ng/mL)
Resulting from Administration by Intraperitoneal Injection
Duration of administration
Time of measurement
after last injection
DMI (10 mg/kg)
SER (10 mg/kg)
12
24
12
24
hours
hours
hours
hours
Chronic
(i.e., daily
injection
for 21 days)
Acute
(i.e., a single
injection)
573.8 6 277.4
422.0 6 85.8
765.0 6 172.9
287.0 6 85.8
334.5 6 181.9
295.2 6 204.0
329.5 6 85.8
172.5 6 20.0
DMI, desipramine; SER, sertraline. For details, see Table 2. N 5 4 for each cell
except for chronic administration of SER (n 5 3).
124
BIOL PSYCHIATRY
2001;49:117–129
(except for phenelzine, which was not assessed because of
the unavailability of the method in our clinical laboratory).
The blood levels shown in Table 2 were those present after
administration of drug by osmotic minipump for 14 or 30
days. Blood levels were also determined after repeated IP
injection of drug to permit comparison of levels produced
by this technique with those resulting from minipump
administration. Table 3 shows blood levels of DMI and
SER following repeated IP injection. The blood levels
shown in Table 3 were those present at 12 and 24 hours
after the final injection of 21 days of daily drug injections
and also at these times after a single IP injection.
Discussion
Several effective antidepressant treatments— chronic administration of five antidepressant drugs (two tricyclics,
two SSRIs, and an MAO inhibitor) and a series of
electroconvulsive shocks—all decreased electrophysiologic activity of LC neurons in halothane-anesthetized
rats. Spontaneous depolarization of LC neurons was markedly decreased by all of the treatments tested. Sensoryevoked burst firing was also decreased, although in one of
the groups (female rats that received DMI for 14 days), the
change in this parameter did not reach statistical significance. It should be noted that despite the size and
consistency of these effects, the magnitude of the changes
shown in Figures 2 through 5 quite possibly underestimates the decrease in LC activity produced by the antidepressant treatments. This is because LC units were difficult to isolate and quantify in some of the animals that
received antidepressant treatments. As a result, some of
the animals that received an antidepressant treatment did
not contribute to the data reported here because, during
recording from these animals, repeated passes of the
microelectrode resulted in no units that fired consistently
or rapidly enough to be designated as LC neurons by the
electrophysiologic criteria used (described in Electrophysiologic Recording in Methods and Materials). In
several cases, a total absence of discernable single-unit
electrophysiologic activity was encountered. Specifically,
whereas LC single units were successfully isolated and
quantified in all 37 of the vehicle-treated and control
animals used during the course of the study, 23 out of 140
animals that received an antidepressant treatment were not
represented in the data because of failure to isolate and
measure any LC units in these animals. This difference
between control and antidepressant-treated conditions is
highly significant (p , .01 by x2). All antidepressanttreated conditions contributed to this result (i.e., animals in
which no units could be isolated/animals from which
recording was attempted 5 3/19 DMI; 5/22 IMI; 3/19
PHE; 4/23 FLU; 2/22 SER 10 mg; 2/18 SER 25 mg; and
M.M. Grant and J.M. Weiss
4/19 ECS). It can be suggested that the animals in which
no units could be isolated were ones in which electrophysiologic activity of LC neurons was so suppressed that
their LC neurons did not depolarize sufficiently to permit
identification of LC activity by the criteria employed. In
this event, exclusion of these subjects from the data set
means that the results shown above underestimate somewhat the extent to which LC activity was suppressed by
these treatments.
Mechanisms can be suggested by which the antidepressant treatments tested produced inhibition of LC activity.
Decreased LC activity caused by drugs that block NE
reuptake (DMI and IMI) is consistent with a strong
inhibitory action on LC firing that is exerted via stimulation of somatodendritic a2 receptors on LC neurons
(Aghajanian and Vandermaelen 1982; Cedarbaum and
Aghajanian 1976; Simson and Weiss 1987). Blockade of
NE reuptake and resultant elevation of NE in the LC
region will increase stimulation of these somatodendritic
a2 receptors to potently inhibit depolarization of LC
neurons. A similar influence can be suggested to account
for the effects of SSRIs on LC activity. Serotonergic
receptors are found on LC cell bodies (Pickel et al 1977),
and activation of these receptors also inhibits LC activity
(Segal 1979); thus, blockade of reuptake of 5-HT released
from serotonergic terminals in the LC region will increase
stimulation of these 5-HT receptors to inhibit LC depolarization. Because MAO inhibition increases the extracellular levels of both NE and 5-HT, the MAOI tested (PHE)
could decrease LC activity by increasing stimulation of
both inhibitory adrenergic and serotonergic receptors on
LC neurons. Whereas the effects of the drugs tested are not
surprising, it is noteworthy that ECS also produced a
decrease in activity of LC neurons at two time points after
the final ECS had been given (i.e., 24 hours and 5 days
afterward). The mechanism by which a series of electroconvulsive shocks causes a persisting inhibition of LC
activity is not evident at this time and remains to be
elucidated.
The findings reported here—namely, that antidepressant
treatments decrease LC activity—are consistent with what
other investigators have seen after acute administration of
antidepressant drugs (e.g., Blier and de Montigny 1985;
Nyback et al 1975; Scuvee-Moreau and Dresse 1979;
McMillen et al 1980; Valentino and Curtis 1991; Valentino et al 1990) and also with most findings of the more
limited number of studies that have assessed effects of
chronic administration (summarized in Table 1). The most
notable exception to the results reported here and the
findings of others is the observation by Valentino et al
(1990) that chronic administration of DMI did not decrease either spontaneous or sensory-evoked LC activity.
These investigators also did not find that chronic admin-
Antidepressant Effects on LC Electrophysiology
istration of the SSRI sertraline changed evoked activity of
LC neurons as was observed in the present study. They did
observe, however, that spontaneous activity of LC was
decreased by chronic administration of the MAO inhibitor
phenelzine and also sertraline if comparison is made with
rats that received a similar series of vehicle injections.
Commenting on why decreased LC firing was not seen
after chronic administration of DMI whereas it was observed after acute DMI, Valentino and colleagues suggested that “tolerance” develops to the acute effect of the
drugs. In the case of DMI that potently blocks NE
reuptake, a likely mechanism for “tolerance” would be
downregulation of the inhibitory somatodendritic a2-receptors on LC neurons so that firing rate of LC neurons
returns to normal despite elevation of NE in synapses in
the LC region. The same rationale could be applied to the
SSRI; chronic administration could be thought to downregulate the inhibitory 5-HT receptors on LC neurons to
reduce inhibitory effects. It was with this hypothesis in
mind that we administered drug for 30 days as well as for
14 days in our studies. The rationale was as follows: if
drug administration for 14 days resulted in a decrease in
LC single-unit activity, it might be that drug administration for this length of time was not sufficient to cause
“tolerance” to occur (e.g., was not sufficient to produce
enough downregulation of presynaptic inhibitory receptors
to bring about recovery of LC activity). Were this the case,
however, administering drug for a longer period (i.e., 30
days) should either produce the expected recovery or, at
the least, cause some recovery from the amount of
decrease seen with 14 days of drug administration as an
indication that the proposed “tolerance” process was
underway. The results found in the present study indicated
that drug administration for 30 days did not bring about
recovery of LC activity. Moreover, inhibition of LC
activity was often more pronounced with 30 days of drug
administration than it was with 14 days, so that there was
no clear indication that a recovery process was underway
that would eventually return LC activity to a normal level.
It should be noted that we do not contend that downregulation of inhibitory presynaptic receptors does not occur
with chronic stimulation of these receptors; rather, our
conclusion is that when one delivers drug to rats via
minipump so that drug is clearly in circulation at time of
electrophysiologic measurement (see the following two
paragraphs), the findings indicate that downregulation is
not sufficient to cause return of LC activity to a normal
level, and, consequently, antidepressant drug administration results in an ongoing and pervasive decrease of LC
activity.
To try to explain why we and other researchers observed decreases in LC activity with chronic administration of certain antidepressant drugs whereas Valentino and
BIOL PSYCHIATRY
2001;49:117–129
125
colleagues did not, we considered that blood levels of drug
might be a significant factor. In the present study, antidepressant drugs were administered by minipump not only to
avoid subjecting the animals to repeated injections but also
to have constant infusion of drug to insure that drug was in
circulation at the time electrophysiologic recordings were
made. The latter consideration was thought to be important
because clinical practice indicates that sufficient circulating levels of drug should be attained to produce a
therapeutic result (Orsulak 1986; Preskorn 1989; Preskorn
and Fast 1991; Van Brunt 1983). In contrast to minipump
administration, Valentino and colleagues (i.e., Curtis and
Valentino 1991; Valentino and Curtis 1991; Valentino et
al 1990) gave antidepressant drugs by daily intraperitoneal
injection for 21 days (usually 10 mg/kg) and then recorded
LC activity 12 to 20 hours after the final injection. We
therefore measured blood levels of drug to assess the
possibility that these different methods of drug administration resulted in different circulating levels of drug.
Table 2 shows blood levels of several of the antidepressants used in the present study (DMI, IMI, FLU, and SER)
resulting from minipump administration for 14 and 30
days. Relative to clinically effective levels reported for
humans (Orsulak 1986; Preskorn 1989; Van Brunt 1983),
the levels produced by minipump administration to the rats
were high. Nonetheless, the data shown in Table 2 make
clear that the antidepressant drugs used certainly were in
circulation at the time of electrophysiologic recording. (It
should be noted that although blood levels were high, no
rat in the study died during drug administration or was
observed to have seizures or become ataxic. This included
the 60 rats that received a tricyclic or MAOI, so that the
high blood levels were not associated with evidence of
cardiotoxicity that occurs in humans when abnormally
elevated levels of TCA and MAOIs are present.) In
comparison with these values produced by minipump
administration, Table 3 shows blood levels of DMI and
SER in rats that received drug by IP injection (10 mg/kg)
daily for 21 days, with measurements made at 12 and 24
hours after the final injection (to bracket the time during
that Valentino and colleagues reported they took their
electrophysiologic measures). (Valentino and colleagues,
as well as other previous studies of electrophysiologic
responses of LC neurons after antidepressant administration, did not include blood levels of antidepressant drugs
in the publications.) Comparison of the values in the left
column of Table 3 (i.e., blood levels after chronic administration; levels after a single injection were also measured
and are shown in the right column) with those for the same
drugs in Table 2 indicate that daily injection for 21 days
produced circulating levels of drug between 12 and 24
hours postinjection that were similar to what was obtained
with minipump administration. Thus, differences between
126
M.M. Grant and J.M. Weiss
BIOL PSYCHIATRY
2001;49:117–129
Table 4. Summary of the Effects of Chronically Administered Antidepressants on Levels of Cerebrospinal Fluid MHPG
Drug type
Drug
Tricyclic
Amitryptiline
Amitryptiline
Amitryptiline
Chlorimipramine
Monoamine
oxidase inhibitor
Effect
Study citation
75–150 mg/day for 2 weeks
Up to 250 mg/day for 4 weeks
250 mg/day for 4 weeks
150 mg/day for 3 weeks
Decrease
Decrease
Decrease
Decrease
Chlorimipramine
Desipramine
Desipramine
Desipramine
Desipramine
Desipramine
(reserpine
augmented)
Imipramine
Nortryptiline
For 3 and 6 weeks
150 mg/day for 6 weeks
181 mg/day for 5 weeks
100 –300 mg/day for 3– 4 weeks
100 –300 mg/day for 4 – 6 weeks
2.5 mg/kg/day for 4 weeks;
reserpine 5 mg/b.i.d.
Decrease
Decreasea
Decrease
Decrease
Decrease
Decrease
Mendelwicz et al 1982
Maas et al 1984
Bowden et al 1985
Bertilsson et al 1974;
Thoren et al 1980;
Traskman et al 1979
Martensson et al 1991
Dahl et al 1982
Sharma et al 1994
Potter et al 1981
Potter et al 1985
Price et al 1987
250 mg/day for 4 weeks
150 mg/day for 3 weeks
Decrease
Decrease
Clorgyline
20 –30 mg/day
for 3 weeks
5–10 mg/day
for 5– 8 weeks
75–100 mg/day
for 3 weeks
60 mg/day for 3 weeks
600 mg/day for 6 weeks
Decrease
Bowden et al 1985
Bertilsson et al 1974;
Thoren et al 1980
Major et al 1979
Decrease
Potter et al 1985
Decrease
Major et al 1979
Decrease
Decreasea
Sunderland et al 1994
Dahl et al 1982
50 –300 mg/day
for 3 weeks
200 mg/day for 3 weeks
Up to 300 mg/day
for 4 – 6 weeks
200 mg/day for 3– 4 weeks
Unstated
30 – 60 mg/day for 2 weeks
Decrease
Bertilsson et al 1980
Decrease
Decreasea
Wallinder 1981
Potter 1981
Decrease
Decreasea
Decrease
Potter et al 1985
Hsiao et al 1987
Mendelwicz et al 1982
Clorgyline
Pargyline
Selective serotonin
reuptake inhibitor
Selegiline
Femoxetine
Zimelidine
Zimelidine
Zimelidine
Aminoketone
Atypical
a
Zimelidine
Buproprion
Mianserin
Dosage
Not statistically significant; all effects not so noted are statistically significant.
the electrophysiologic findings of Valentino and colleagues compared with those of the present study and
other investigators do not appear explainable on the basis
of blood levels of drug at the time of electrophysiologic
recording. At present, we are unable to account for the
difference between results of Valentino and colleagues in
comparison with others.
Although measuring blood levels of drug failed to
clarify why one group of investigators obtained results that
differ from the findings of the others, there is significant
additional information that bears on the question of
whether antidepressant treatment decreases LC activity. A
number of human studies have attempted to assess noradrenergic activity in human brain by measuring the noradrenergic metabolite MHPG in cerebrospinal fluid (CSF) in
conjunction with antidepressant treatment. These studies,
summarized in Table 4, have consistently found decreased
MHPG in CSF of human patients taking antidepressant
drugs. Although MHPG in CSF does not exclusively
reflect NE release in brain because MHPG in blood can
diffuse freely into CSF (Kopin 1985), Scheinin reviewed
this issue and concluded, “Drug induced alterations in the
concentration of monoamine metabolites in human CSF
probably reflect quite accurately the effects of several
classes of drugs [inhibitors of MAO and antidepressant
drugs included in this list] on monoamine turnover in the
CNS” (Scheinin 1985, 10). One study (Sharma et al 1994)
utilized a correction of CSF MHPG for potential contribution by plasma MHPG that was recommended by Kopin
(1985) and reached the same conclusion as other studies
that CSF MHPG was reduced as a result of antidepressant
drug treatment. Considering that approximately 70% of
the NE in brain derives from the LC, the findings shown
in Table 4 are consistent with the influence of antidepressant drugs in humans decreasing LC neural activity, as was
observed in the rats in the present study. Moreover, these
data from humans also suggest that despite the high blood
levels of antidepressant drug found in the animals of the
Antidepressant Effects on LC Electrophysiology
present study, the results reported here are relevant to what
occurs in humans undergoing antidepressant treatment.
The finding reported here that all antidepressant treatments tested cause a marked decrease in LC neuronal
activity leads to the question of how such a change might
be important for effective antidepressant treatment. Recently, other investigators have suggested that LC neuronal activity is abnormally elevated in depression and that
antidepressant action consequently may occur through
reduction of LC neuronal activity (Gold and Chrousos
1999; Zhu et al 1999). Of course, in considering the
mechanism by which reduced LC activity might produce
an antidepressant effect, it should be stated at the outset
that decreased LC neuronal activity resulting from antidepressant treatment could be an epiphenomenon unrelated
to antidepressant action. For example, LC neuronal activity might be decreased simply as a homeostatic adjustment
because antidepressants increase NE in synapses throughout the brain and that potentiation of NE action throughout
the brain (or resulting downregulation of B-adrenergic
receptors) is what actually underlies antidepressant action.
Also, the linkage of decreased LC activity to depression is
possibly, or even probably, not a simple one, but, rather,
the result of complex influences on neural transmission
and information processing (e.g., Mongeau et al 1997). On
the other hand, the study described in this article was
prompted by a hypothesis we have offered recently that
links LC activity directly to certain behavioral changes
and processes seen in depression (Weiss et al 1996, 1998).
This hypothesis suggests that certain salient behavioral
changes seen in depression, particularly psychomotor
retardation and failure to appreciate rewarding stimuli, can
arise from abnormally high activity of LC neurons. The
proposed mechanism is as follows: when LC neurons are
highly active, the peptide galanin, which is colocalized in
LC neurons, will be released from the LC terminals in
addition to NE. When galanin release occurs from LC
terminals in the ventral tegmentum, this peptide will exert
a hyperpolarizing influence on dopaminergic cell bodies in
this brain region, resulting in decreased release of dopamine in the forebrain. This in turn is hypothesized to cause
behavioral changes that occur in depression, particularly
psychomotor retardation and anhedonia. With respect to
the findings described here, this hypothesis suggests that
decreasing LC activity will reduce galanin release from
LC terminals in the ventral tegmentum, thereby releasing
dopaminergic cell bodies in this region from galaninmediated inhibition. A reduction in LC activity, possibly
through the mechanism described here, is proposed to
underlie, at least in part, the e