Upregulation of Voluntary Alcohol Intake, Behavioral
Sensitivity to Stress, and Amygdala Crhr1 Expression
Following a History of Dependence
Wolfgang H. Sommer, Roberto Rimondini, Anita C. Hansson, Philip A. Hipskind, Donald R. Gehlert,
Christina S. Barr, and Markus A. Heilig
hormone (CRH) has been implicated in this transition, but underlying molecular mechanisms remain unclear.
Methods: A postdependent state was induced using intermittent alcohol exposure. Experiments were carried out following ?3 weeks of
recovery to eliminate contributions of acute withdrawal. Voluntary alcohol consumption was assessed in a two-bottle, free choice proce-
forced swim stress on voluntary alcohol intake were examined as a function of exposure history. Expression of Crh, Crhr1, and Crhr2
transcripts was analyzed by in situ hybridization histochemistry.
Results: Alcohol drinking was upregulated long-term following a history of dependence. Fear suppression of behavior was selectively
3-(4-Chloro-2-morpholin-4-yl-thiazol-5-yl)-8-(1-ethylpropyl)-2,6-dimethyl-imidazo[1,2-b]pyridazine (MTIP) (10 mg/kg). Forced swim stress
increased alcohol intake in postdependent animals but not in control animals. Behavioral changes were paralleled by an upregulation of
In contrast, Crhr2 expression was down in the BLA.
Conclusions: Neuroadaptations encompassing amygdala CRH signaling contribute to the behavioral phenotype of postdependent
Key Words: Alcoholism, animal model, conflict test, in situ hybrid-
elevated anxiety, low mood, and increased sensitivity to stress,
here collectively labeled negative affect, has been postulated as
being critical for the transition from a nondependent to a
dependent state (Breese et al. 2005a; Heilig and Egli 2006;
Valdez and Koob 2004). Clinically, negative affect is most
prominent during acute alcohol withdrawal but persists into
protracted abstinence (Hershon 1977), as shown, for example,
by potentiated startle responses (Krystal et al. 1997) and
increased frequency of panic attacks (George et al. 1990).
Environmental stressors constitute a major category of stimuli
capable of triggering relapse in humans and experimental
animals (Brownell et al. 1986; Shaham et al. 2003). Thus,
increased sensitivity to stress in the postdependent state is
likely to contribute to maintaining alcohol dependence. Iden-
tifying its neural substrates, therefore, is critical to developing
novel alcoholism treatments.
lcoholism develops over years and requires prolonged
periods of brain exposure to intoxicating levels of alco-
hol. Over the course of this process, recruitment of
A history of alcohol dependence has been modeled in
laboratory rats using prolonged exposure to alcohol vapor,
which triggers long-lasting neural and behavioral plasticity that
appear relevant for modeling human alcoholism. This type of
manipulation produces persistently increased alcohol intake in
genetically nonselected rats (Rimondini et al. 2002; Roberts et al.
2000). Exposure to repeated cycles of intoxication and with-
drawal, which mimics the course of the clinical condition, is most
effective for inducing increased alcohol drinking (O’Dell et al.
2004; Rimondini et al. 2002). Similar to the human condition, a
minimum duration of dependence is required for lasting upregu-
lation of alcohol preference (Rimondini et al. 2003). Elevated
alcohol intake in postdependent rats is sensitive to the clinically
effective compound, acamprosate, while alcohol intake of nonde-
pendent rats is unaffected by the same treatment (Egli 2005; Heyser
et al. 1998; Rimondini et al. 2002; Spanagel and Zieglgansberger
1997). Furthermore, the postdependent state is characterized by a
persistently upregulated behavioral sensitivity to stress (Valdez et al.
2002, 2003). Together, these findings indicate that neuroadaptive
processes induced by a prolonged exposure to cycles of intoxica-
tion and withdrawal parallel those in human alcoholism and might
be able to shed light on underlying neural mechanisms.
Corticotropin-releasing hormone (CRH) mediates behavioral
stress responses through extrahypothalamic mechanisms. These
actions of CRH are primarily mediated through the CRH-R1
receptor subtype, which has been proposed as an attractive
target for medication development in anxiety, depression, and
addiction (Heinrichs and Koob 2004; Holsboer 2003; Reul and
Holsboer 2002; Sarnyai et al. 2001). Brain regions of particular
importance for drug reward, including the medial prefrontal
cortex (mPFC), the nucleus accumbens (NAcc), the bed nucleus
of the stria terminalis (BNST), and several nuclei within the
amygdala, including central (CeA), medial (MeA), and basolat-
From the Laboratory of Clinical and Translational Studies (WHS, ACH, CSB,
MAH), National Institute on Alcohol Abuse and Alcoholism/National
Institutes of Health, Bethesda, Maryland; Department of Pharmacology
DRG), Lilly Research Laboratories, Indianapolis, Indiana.
Address reprint requests to Markus A. Heilig, M.D., Ph.D., Laboratory of
Clinical and Translational Studies, NIAAA/NIH, 10 Center Drive, B 10, R
15330, Bethesda, MD 20892-1108; E-mail: email@example.com.
BIOL PSYCHIATRY 2008;63:139–145
© 2008 Society of Biological Psychiatry
eral amygdala (BLA), are rich in CRH receptors, specifically of the
R1 subtype (Potter et al. 1994; Van et al. 2000). Most of the CRH
neurons targeting these regions originate from cortical interneu-
rons or CeA (Swanson et al. 1983), a structure that mediates fear
and anxiety (Davis et al. 1997; LeDoux et al. 1988; Möller et al.
1997). Anxious responding during acute alcohol withdrawal is
attenuated by CRH antagonists, administered either systemically
or directly into the CeA (Baldwin et al. 1991; Funk et al. 2007;
Knapp et al.. 2004; Overstreet et al. 2004; Rassnick et al. 1993;
Valdez et al. 2002). Furthermore, CRH antagonists block both
elevated alcohol self-administration and potentiated anxiety-like
responses to stressors seen during protracted abstinence follow-
ing a history of dependence (Breese et al. 2005b; Valdez et al.
Recruitment of central CRH signaling thus underlies two key
features of the postdependent state, namely, long-term increased
voluntary consumption of alcohol and persistently upregulated
behavioral sensitivity to stress. The molecular mechanisms for
these changes are largely unknown. Recently, we reported that
in the genetically selected Marchigian-Sardinian alcohol prefer-
ring (msP) rat line (Ciccocioppo et al. 2006), high alcohol
preference has cosegregated with increased behavioral sensitiv-
ity to stress, creating a phenocopy of the postdependent pheno-
type. A screen for differential gene expression identified an
innate upregulation of the Crhr1 transcript in several brain
regions of the msP line. Administration of the selective CRH-R1
antagonist, antalarmin, demonstrated a causal role of upregu-
lated CRH-R1 receptors in the behavioral phenotype of msP rats
(Hansson et al. 2006).
Here, we examined alcohol drinking, stress sensitivity, and
expression of Crh and its receptors Crhr1 and Crhr2. Our
hypothesis was that, similar to what we have found in msP rats,
an upregulation of the Crhr1 transcript might be present in
genetically nonselected rats following a history of dependence
and that this upregulation would contribute to their behavioral
Methods and Materials
Male Wistar rats (Mollegard, Denmark), weighing 225 g to
250 g at outset of experiments, were housed four per cage at
20oC to 22oC, 45% to 55% controlled humidity, and reverse 12:12
hour light/dark cycle (lights off at 11:00 AM) and tested during the
dark phase. All procedures followed the European Commission
Council Directive for Care and Use of Laboratory Animals (ethics
permit S84/98, Stockholm South).
Animals were removed from alcohol or sham exposure after 4
or 7 weeks. After recovering for a period of 3 weeks, animals
from each group were randomized to one of the three following
1. Voluntary Alcohol Drinking (7-week exposed: n ? 10 vs.
n ? 8; 4-week exposed: n ? 7 vs. n ? 8, exposed and
sham, respectively). After initial assessment of drinking, all
7-week animals continued voluntary alcohol drinking on
the 24-hour access two-bottle free choice for 29 days, until
assessed for effects of forced swim stress.
2. Stress Sensitivity in the Vogel Conflict Test (7-week only:
n ? 20 for exposed and sham, respectively). This experi-
ment was carried out directly after the 3-week recovery
period. Subjects were then kept for an additional 10 weeks,
after which postdependent and control subjects were ran-
domized to pretreatment with vehicle or CRH-R1 antagonist
and retested (n ? 10 per group).
3. Analysis of Gene Expression. Gene expression was as-
sessed by in situ hybridization. Animals were sacrificed
directly after the 3-week recovery period (n ? 7 for both
exposed and sham).
Alcohol Vapor Exposure
Exposure was as described (Rimondini et al. 2002) in glass/
steel chambers (1 ? 1 ? 1 m). High-performance liquid chro-
matography (HPLC) pumps (Knauer, Berlin, Germany) delivered
alcohol into electrically heated stainless steel coils (60°C) con-
nected to an airflow of 18 L per minute. Alcohol concentration
was adjusted by changing pump flow and monitored via a spec-
trometer (Wilks, South Norwalk, Connecticut). Exposure was for
17 hours during each 24-hour period (on 4:00 PM; off 9:00 AM).
Rats were allowed to habituate to the chambers for 1 week, then
exposed to low alcohol concentration for 1 week, and finally
exposed to alcohol vapor yielding blood alcohol concentrations
as shown in Table 1. Control animals were kept in identical
chambers with normal airflow. Weekly, rats were weighed and
blood was collected from the lateral tail vein, serum extracted,
and assayed for ethanol using an nicotinamide adenine dinucle-
otide phosphate dehydrogenase/spectrophotometric assay kit
(Sigma Aldrich Inc., St. Louis, Missouri) according to the manu-
Alcohol Consumption and Its Modulation by Stress
Alcohol consumption was measured as 24-hour access two-
bottle free choice between 6% alcohol (wt/vol) in .2% saccharin
solution, or vehicle, .2% saccharin solution only, as described
(Rimondini et al. 2002). One week was used to fade in alcohol,
and consumption was measured over the following 2 weeks.
To assess effects of stress on voluntary alcohol consumption,
two-bottle free-choice drinking was continued for 29 days.
Baseline data were obtained over a 3-day block on days 30 to 32
after initiation of drinking. The forced swim stress was carried
out daily over a 3-day block on days 33 to 35 as described
(Vengeliene et al. 2003). Briefly, around 3:00 PM on each test day,
animals were removed from their home cage and placed for 10
min in a plastic cylinder (45 ? 20 cm) filled up to 35 cm with
19°C water. After completion of the forced swim, subjects were
returned to their home cages. Following completion of the stress
block, poststress drinking measures were obtained over a final
3-day block (days 36 to 38).
Punished Drinking Test
A modified Vogel drinking test was used as described (Som-
mer et al. 2001), and punished licks during the 8-min conflict
Table 1. Blood Alcohol Concentrations Resulting from the Alcohol Vapor
Exposure Used to Induce Dependence
Exposure Week BAC (mg/dL; mean ? SEM)Range
194.4 ? 29.2
313.9 ? 15.0
237.3 ? 22.7
204.4 ? 14.6
309.8 ? 21.6
388.1 ? 38.6
303.0 ? 15.5
BAC, blood alcohol concentration.
140 BIOL PSYCHIATRY 2008;63:139–145
W.H. Sommer et al.
interval were recorded. Unpunished licks during the preceding
4-min control interval were also recorded, as a control for
possible effects on thirst or motor performance.
To assess whether potentiated fear suppression of behavior in
the conflict test is mediated by CRH-R1 activation, a novel
imidazopyridazine CRH-R1 antagonist, 3-(4-Chloro-2-morpholin-
pyridazine (MTIP) (Lilly Research Laboratories, Indianapolis,
Indiana) was used. The MTIP binds CRH-R1 receptors with
nanomolar affinity, with no detectable activity at the CRH-R2
receptor or other common drug targets, and is highly brain
penetrant. A 10 mg/kg was chosen because median effective
dose (ED50) has been determined to approximately 1.5 mg,
while 10 mg/kg produces a more than 90% blockade of several
stress-induced behaviors (Gehlert et al., 2007). Vehicle (10%
Tween [Sigma Aldrich Inc.] 80 in distilled water) or MTIP in
vehicle were administered intraperitoneally (IP) 30 min prior to
In Situ Hybridization
Procedures were performed as described previously (Hans-
son et al. 2003, 2006). Rats were decapitated in the inactive phase
(11:00 PM to 3:00 AM), and brains were removed, snap frozen in
?40°C isopentane, and stored at ?70°C until use. The 10-?m
brain sections were taken at Bregma levels 1) ?2.5 to ?1.7 mm,
2) ?.3 to ?.4 mm, 3) ?1.7 to 2.0 mm, and 4) ?2.3 to ?3.3 mm
(Paxinos and Watson 1998). Following hybridization with ribo-
probes for Crh, Crhr1, or Crhr2 messenger RNA (mRNA),
sections were exposed to Fuji BAS-5000 Phosphorimager plates
(Fujifilm, Tokyo, Japan). Digital images were analyzed using AIS
Image Analysis Software (Imaging Research Inc., St. Catharines,
Ontario, Canada). Regions of interest were chosen based on
available functional data with microinjections of CRH receptor
ligands within the amygdala and BNST. Values were converted to
nCi/g using carbon 14 (14C) standards.
Daily alcohol intake was averaged over 2 weeks. The 4-week
and 7-week sham-exposed control groups did not differ from
each other and were, therefore, pooled. Following one-way
analysis of variance (ANOVA), each of the exposed groups was
compared against the pooled control group by Dunnett’s post
hoc test. In the swim-stress experiment, daily intake was aver-
aged over each of the three blocks (baseline, stress, poststress),
each of which was 3 days in duration. Drinking data were
analyzed using two-way ANOVA, with history of dependence as
between-subjects and the three drinking blocks as a within-
subjects factor. Post hoc comparison was performed using the
Punished licks violated assumptions of homogenous vari-
ances and were rank-transformed prior to analysis. In the initial
experiment, a one-way ANOVA was carried out, with history of
dependence as between-subjects factor. In the pharmacological
experiment, a two-way analysis was carried out, with history of
dependence and treatment (MTIP vs. vehicle) as factors. To
assess whether any fear suppression by the conflict remained
following MTIP treatment, unpunished and punished licks
were both recalculated as rates (licks/min) to be directly
comparable, and equality was tested using the confidence
interval method. In each experiment, unpunished licks were
separately analyzed as a control for nonspecific effects on
thirst or motor performance.
Gene Expression. Data were compared by region-wise one-
way ANOVAs, followed by the reverse Holm-Bonferroni correc-
tion (Holm 1979; Dow 2003).
Increased Alcohol Consumption Following a History of
Dependence and Selective Increase of Alcohol Intake by
Stress in Postdependent Rats
Average daily intake of 7-week and 4-week exposed animals
was compared with that of their pooled control animals, since
the latter did not differ. A more than twofold increase was
observed in 7-week but not 4-week exposed animals when
compared with control animals [3.8 ? .35 and 1.12 ? .42,
respectively, vs. 1.44 ? .28; F(2,30) ? 17.5, p ? .001, Dunnett’s
post hoc test p ? .001, 7-week exposed vs. control animals;
Figure 1]. Vehicle intake was unaffected (data not shown).
In the stress experiment, animals with a history of dependence
continued to consume higher amounts of alcohol [main effect of
exposure history: F(1,16) ? 12.7, p ? .003]. There was a significant
overall effect of stress exposure [main effect: F(2,32) ? 9.0, p ?
.0008], but this affected the two groups differentially, as shown
by a significant interaction term [F(2,32) ? 4.1, p ? .027]. Post
hoc analysis revealed that within the postdependent group,
drinking both during the stress and the poststress block was
significantly higher than during the prestress baseline block (p ?
.001). In contrast, within the sham-exposed control group,
drinking during the stress and the poststress block was indistin-
guishable from that measured during the prestress baseline block
(Figure 2). Vehicle intake did not differ between groups and was
unaffected by stress exposure (data not shown).
Potentiated Fear Suppression of Behavior Following a History
of Dependence and Its Reversal by the Selective CRH-R1
After 3-week recovery, fear suppression followed a history of
dependence, as shown by markedly lower rates of punished
responding observed in 7-week exposed animals compared with
their corresponding control group [F(1,36) ? 14.3, p ? .0006;
Figure 3]. Unpunished responding did not differ between the
Potentiation of fear suppression by a history of dependence
persisted when tested 13 weeks after recovery [main effect of
Figure 1. Long-lasting increases in voluntary alcohol consumption in rats
with a history of dependence. Two-bottle, free-choice, continuous access
alcohol in .2% saccharin versus .2% saccharin only was assessed after a
3-week resting period that followed the last exposure cycle. One week was
testing was over the following 2 weeks. Left panel: Consumption over the
2-week period. Right panel: Average daily consumption over the same
7-week exposed group compared with their sham-exposed control group.
Values are expressed as mean daily consumption ? SEM of 6% (vol/vol)
alcohol consumption. ***p ? .001. For detailed statistics, see Results.
W.H. Sommer et al.
BIOL PSYCHIATRY 2008;63:139–145 141
exposure history: F(1,35) ? 5.4, p ? .025]. At this time,
pretreatment with the CRH-R1 antagonist, MTIP, had a robust
anticonflict effect [main effect: F(1,35) ? 13.2, p ? .0009]. The
MTIP eliminated the difference in conflict behavior related to
exposure history. In fact, in both animals with and without a
history of dependence, MTIP fully eliminated any fear sup-
pression by conflict (equality between punished and unpun-
ished lick rates within a 10% indifference interval: p ? .05;
Long-Lasting Increase ofCrhr1 and Crh Gene Expression in the
Amygdala Following a History of Dependence
Postdependent animals had robustly elevated levels of Crh1
transcript in BLA [F(1,12) ? 11.3, p ? .01] and MeA [F(1,12) ?
25.6, p ? .001] but not in CeA or BNST (Figure 5C). The Crhr1
transcript was not reliably detected in the hypothalamic paraven-
tricular nucleus (PVN).
Expression of the Crhr2 transcript was less affected in the
extrahypothalamic regions studied, with the exception of a
moderate decrease in the BLA [F (1,11) ? 6.7, p ? .025; Figure
5D]. Postdependent animals did not differ from control animals
in Crhr2 expression within the PVN (22.5 ? 1.43 vs. 23.3 ? 1.17
nCi/g, mean ? SEM, respectively).
Expression of the Crh transcript was elevated in CeA of
postdependent animals [F(1,12) ? 7.1, p ? .02], while no
difference was seen within the BNST (Figure 5B). Postdependent
animals did not differ from control animals within the PVN
(103.5 ? 3.5 vs. 107.4 ? 4.7).
We report elevated voluntary alcohol consumption and po-
tentiated stress sensitivity during protracted abstinence following
a history of alcohol dependence. Postdependent animals were
selectively sensitive to upregulation of alcohol consumption by a
stressor. Long-term upregulation of the transcript encoding the
CRH-R1 receptor was found within MeA and BLA of postdepen-
dent animals, while CRH-R1 antagonism eliminated the increased
behavioral sensitivity to stress in the postdependent state.
Our drinking data replicate prior reports (O’Dell et al. 2004;
Rimondini et al. 2002; Rimondini et al. 2003; Roberts et al. 2000).
In the present study, alcohol intake more than doubled in
animals with a history of dependence. This increase was seen
after a 3-week recovery following completion of alcohol expo-
sure and persisted for more than an additional month. It is
therefore related to long-term neuroadaptations rather than acute
withdrawal. A previously reported temporal threshold was rep-
licated, so that a 7-week exposure upregulated alcohol intake,
while a 4-week exposure did not. We have independently found
that increased consumption in postdependent animals is not due
to altered alcohol metabolism (Sommer et al., in preparation).
Our conflict data show a long-lasting decrease in punished
water drinking. This effect is not due to altered thirst, as shown
by the unaffected rates of unpunished licks. Furthermore, it is
a history of dependence, induced by a 7-week exposure to daily cycles of
of acute withdrawal, fear-induced suppression of behavior in the punished
drinking test was assessed after a 3-week resting period that followed the
last exposure cycle. Punished licks recorded over an 8-min conflict period
motor performance. ***p ? .001. For detailed statistics, see Results.
Figure 4. Persistent increased behavioral sensitivity to stress following a
history of dependence as measured in the conflict test, when animals were
retested 13 weeks after completion of alcohol exposure (main effect of
history of dependence: p ? .025). The selective, brain penetrant CRH-R1
antagonist MTIP (10mg/kg) produced a robust overall anticonflict effect
(main treatment effect: p ? .001). The MTIP eliminated the increased sensi-
tivity to stress in the postdependent group and, in fact, eliminated any fear
val of 10%: p ? .05 for both animals without a history of dependence and
postdependent subjects. CRH, corticotropin-releasing hormone; MTIP, 3-
Figure 2. Increase in voluntary alcohol consumption in response to forced
swim stress in animals with a history of dependence but not in nondepen-
tested for voluntary alcohol drinking for 2 weeks (Figure 1), subjects were
duration. During the stress block, animals were subjected to daily forced
swim stress. **p ? .01, ***p ? .001 versus prestress baseline value.
142 BIOL PSYCHIATRY 2008;63:139–145
W.H. Sommer et al.
also unlikely to be caused by altered nociception, since we have
independently established that a history of dependence does not
affect pain thresholds in the hot plate test (Heilig et al., unpub-
lished data). Central administration of CRH has previously been
shown to produce potentiated fear suppression in the Geller-
Seifter test, while both prototypical benzodiazepine anxiolytics
and intracerebroventricular administration of the nonselective
CRH receptor antagonist ?-helical-CRH9-41reversed this potenti-
ation (Britton et al. 1985, 1986). More recently, potent anticonflict
effects were shown in the classical Vogel test with the nonpep-
tide CRH-R1 antagonists antalarmin and SSR125543A (Griebel et
al. 2002). The potentiated suppression of punished drinking
observed in the present study is therefore likely to reflect a
specific, CRH-mediated increase in behavioral sensitivity to
stress. This finding is in line with prior reports that have
demonstrated increased behavioral stress responses in postde-
pendent rats (Breese et al. 2005b; Overstreet et al. 2002; Rasmus-
sen et al. 2001; Valdez et al. 2002, 2003). A history of multiple
withdrawals, paralleling the clinical course of alcoholism, seems
most effective in inducing anxiety-like behavior. For instance,
rats subjected to three cycles of withdrawal from an alcohol diet,
but not those exposed to a single withdrawal, showed reduced
social interaction during early abstinence (Overstreet et al. 2002).
Although this acute withdrawal effect subsided within 48 hours,
a remaining recruitment of negative affect systems was demon-
strated by the observation that following a history of three
withdrawal episodes, a single future withdrawal from reexposure
to chronic ethanol in control animals was anxiogenic. Interestingly,
in this model, administration of a CRH antagonist during the first
two cycles of withdrawal blocked the increase in anxiety behavior
(Knapp et al. 2004; Overstreet et al. 2004). Elevated anxiety has also
been reported on the elevated plus-maze 4 weeks after a liquid
alcohol diet (Rasmussen et al. 2001) or alcohol vapor exposure
(Valdez et al. 2002). In the latter case, no overt phenotype was seen
in postdependent rats, but when testing was preceded by a restraint
stress challenge, an exaggerated anxiogenic response was seen
(Valdez et al. 2003). Our present conflict data are similar in nature,
as the conflict model itself is a stressor sufficient to recruit CeA
activity and, in fact, relies on an activation of cell bodies in this
structure for fear suppression of behavior (Möller et al. 1997). The
fact that a stress-sensitive phenotype persists for more than 3
months after completion of alcohol exposure has not been reported
previously, and the duration of this state is remarkable. It may be
noted that all anxiety-like behaviors discussed here are uncondi-
tioned. In contrast, impaired associative fear learning has recently
been reported in young human binge drinkers, presumably as a
result of repeated cycles of intoxication and withdrawal (Stephens et
al. 2005). This report proposed a “saturation” mechanism, whereby
synaptic plasticity induced by a history of dependence reduces
capacity for future learning, while allowing unconditioned stimuli
access to neuronal pathways underlying conditioned fear.
Elevated alcohol consumption and increased sensitivity to
stress have previously been reported in postdependent rats, but
little data exist to link these phenomena. A transient increase in
voluntary alcohol intake has been reported in nonselected Wistar
rats in response to the forced swim stress used here, but these
animals had been consuming alcohol for 4 months and had
experienced alcohol deprivation during the course of this his-
tory, features that set the scene for neuroadaptive changes
(Vengeliene et al. 2003). In the present study, nondependent
Wistar rats did not increase their alcohol intake in response to
repeated forced swim stress. In contrast, postdependent animals
started out with elevated alcohol intake and showed a highly
significant further increase in response to the stressor. Thus, the
increased sensitivity to stress in the postdependent state trans-
lates into elevated motivation to consume alcohol in response to
Our data provide a putative molecular mechanism for the
postdependent behavioral phenotype. Recently, we reported an
innate upregulation of Crhr1 expression in the amygdala of msP
rats that causes a behavioral phenotype similar to that seen in
postdependent rats. This work also demonstrated that elevated
Crhr1 transcript within the amygdala is strongly correlated with
increased binding density (Hansson et al. 2006). Here, we found
that the postdependent phenotype is accompanied by a very
similar upregulation of Crhr1 receptors within the BLA and MeA.
Elimination of fear suppression in the conflict test by the CRH-R1
antagonist MTIP strongly suggests that, similar to what we found
in the msP line, the upregulated Crhr1 expression in the post-
dependent state is causally related to the behavioral phenotype.
Figure 5. (A) Distribution of Crh transcript, encoding the
CRH precursor, and the Crhr1 and Crhr2 transcripts, en-
coding the respective receptor subtype. Representative
sections from the amygdala of rats without a history of
mm. Quantification of expression levels (nCi/g, mean ?
SEM) for the respective transcript in postdependent rats
versus rats without a history of dependence is shown in
panels (B) through (D). (B) Crh expression was upregu-
lated within CeA, which was the only amygdala region
where measurable levels of this transcript were present.
(C) Crhr1 message was robustly upregulated within BLA
and MeA but not in CeA or BNST. (D) Expression of Crhr2
with the exception of BLA, where a moderate decrease
was seen. For all panels, *p ? .05, **p ? .01, ***p ? .001
corrected for multiple tests. For detailed statistics, see
Results. BLA, basolateral amygdala; BNST, bed nucleus of
stria terminalis; CeA, central amygdala; CRH, corticotrop-
in-releasing hormone; MeA, medial amygdala.
W.H. Sommer et al.
BIOL PSYCHIATRY 2008;63:139–145 143
This conclusion is further supported by our recent findings, in a
separate study, that MTIP fully and dose-dependently blocks the
increase in alcohol intake found in the postdependent state
(Gehlert et al., 2007). These data in are in agreement with
previous studies, which have demonstrated that both stress
responses and elevated voluntary alcohol consumption in the
postdependent state are blocked by CRH antagonism or specif-
ically blockade of CRH-R1 receptors (Breese et al. 2005b; Valdez
et al. 2002, 2003). Our findings are also in line with a recent study
that showed a recruitment of local amygdala CRH circuits but not
those in BNST or NAcc following a history of dependence (Funk
et al. 2006). Collectively, these findings suggest that a long-
lasting upregulation of CRH-R1 receptors within the amygdala
constitutes a shared neurobiological substrate underlying the
behavioral phenotype of msP and postdependent rats. It remains
to be determined to what extent CRH-R1 receptors in CeA, MeA,
or BLA are involved.
In the postdependent subjects, the upregulated Crhr1 expres-
sion was accompanied by an increase of Crh mRNA in the CeA.
This finding confirms and expands on a prior report of increased
amygdala CRH immunoreactivity in postdependent rats 6 weeks
after termination of a liquid alcohol diet (Zorrilla et al. 2001).
Elevated tissue peptide content can result either from increased
synthesis or decreased release. The observation of increased
peptide levels in the previous report, together with elevated
transcript in the present study, jointly establish that CRH synthe-
sis is increased in CeA in the postdependent state. Thus, a
presynaptic and a postsynaptic signaling component may act in
concert to recruit the CRH system in postdependent animals.
Finally, the Crhr2 transcript, although less affected, was oppo-
sitely regulated to Crhr1 within the BLA, in which a decrease was
seen. This is in line with prior data showing that activation of
CRH-R2 receptors is also capable of reversing the increased
sensitivity to stress and elevated alcohol self-administration
during the postdependent state (Valdez et al. 2004).
Our expression analysis suggests that the dysregulation of
CRH systems found during the postdependent state is relatively
restricted to extrahypothalamic systems, while the PVN and
control of the hypothalamic-pituitary-adrenal (HPA) axis seem
unaffected. A limitation of our study is that we were not able to
analyze pituitary expression of CRH receptor transcripts. How-
ever, our findings are in agreement with prior reports that stable,
long-term changes in basal serum corticosterone levels are not
found in postdependent rats (Rimondini et al. 2002; Zorrilla et al.
2001). The precise mechanism by which a parallel increase in
CRH and CRH-R1 expression occurs within the amygdala in the
postdependent state is unclear. The increased vulnerability of
this structure may result from its unique organization, character-
ized by close proximity or intertwining of CRH synthesis and
target sites (Swanson and Petrovich 1998). Also, CRH in CeA
mimics actions of alcohol to potentiate local gamma-aminobu-
tyric acid (GABA) transmission through a CRH-R1 mediated
mechanism, possibly mediating an autoregulatory loop through
recurrent collaterals (Nie et al. 2004), while GABA responses to
alcohol in CeA are upregulated in the postdependent state (Nie
et al. 2004; Roberto et al. 2004). Intermittent exposure to alcohol
intoxication and withdrawal may be particularly effective in
driving an allostatic shift of the amygdala CRH system to a higher
functional set point (Valdez and Koob 2004), since Crhr1 gene
expression regulation is expected to be on a slower scale than
Crh itself. Ultimately, mechanisms regulating transcription must
be recruited. An intriguing possibility is that these might be
similar to those suggested in the phenomenon recently labeled
“incubation of craving,” which refers to increased propensity for
relapse-like behavior over time and which has been shown to
rely on signaling through the extracellular signal-regulated ki-
nase (ERK)-pathway, known to be capable of regulating tran-
scription (Grimm et al. 2001; Lu et al. 2005).
In summary, a recruitment of intra-amygdala but not hypo-
thalamic CRH systems seems to be driving the postdependent
phenotype. An increased stress sensitivity in this state may not be
overt but is pronounced following a stress challenge. Postdepen-
dent upregulation of the CRH system gives rise to excessive rates
of alcohol self-administration. Together, these data provide
compelling evidence that a blockade of hyperactive signaling at
CRH-R1 receptors in the postdependent state inhibit heavy
drinking and reduced relapse risk.
This research was supported by intramural National Institute
on Alcohol Abuse and Alcoholism (NIAAA) funding, funding
from the Swedish Medical Research Council, and the Karolinska
Institute. This research has, in part, been carried out under a
standard U.S. Government collaborative research and development
agreement (CRADA) between the NIAAA and Eli Lilly Research
Laboratories. Authors so listed are employees of Eli Lilly and Co.
Remaining authors have no competing financial interest.
Baldwin HA, Rassnick S, Rivier J, Koob GF, Britton TK (1991): CRF antagonist
reverses the “anxiogenic” response to ethanol withdrawal in the rat.
etiology of alcoholism: A “ kindling”/stress hypothesis. Psychopharma-
Breese GR, Overstreet DH, Knapp DJ, Navarro M (2005b): Prior multiple
receptor agonist. Neuropsychopharmacology 30:1662–1669.
Britton KT, Lee G, Vale W, Rivier J, Koob GF (1986): Corticotropin releasing
factor (CRF) receptor antagonist blocks activating and ’anxiogenic’ ac-
tions of CRF in the rat. Brain Res 369:303–306.
Britton KT, Morgan J, Rivier J, Vale W, Koob GF (1985): Chlordiazepoxide
attenuates response suppression induced by corticotropin-releasing
factor in the conflict test. Psychopharmacology (Berl) 86:170–174.
Brownell KD, Marlatt GA, Lichtenstein E, Wilson GT (1986): Understanding
and preventing relapse. Am Psychol 41:765–782.
ring (msP) rats: An animal model to study the neurobiology of alcohol-
ism. Addict Biol 11:339–355.
Davis M, Walker DL, Lee Y (1997): Amygdala and bed nucleus of the stria
terminalis: Differential roles in fear and anxiety measured with the
detection of transcriptional changes induced in rat neuroblastoma
(NG108) cells by mefloquine. Malar J 2:4.
Egli M (2005): Can experimental paradigms and animal models be used to
discover clinically effective medications for alcoholism? Addict Biol 10:
Funk CK, O’Dell LE, Crawford EF, Koob GF (2006): Corticotropin-releasing
factor within the central nucleus of the amygdala mediates enhanced
ethanol self-administration in withdrawn, ethanol-dependent rats.
factor 1 antagonists selectively reduce ethanol self-administration in
ethanol-dependent rats. Biol Psychiatry 61(1):78–86.
Gehlert DR, Cippitelli A, Thorsell A, Le DA, Hipskind PA, Hamdouchi C, et al.
dimethyl-imidazo[1,2-b]pyridazine: A novel brain-penetrant, orally
available corticotropin-releasing factor receptor 1 antagonist with effi-
cacy in animal models of alcoholism. J Neurosci 27:2718-2726.
144 BIOL PSYCHIATRY 2008;63:139–145
W.H. Sommer et al.
George DT, Nutt DJ, Dwyer BA, Linnoila M (1990): Alcoholism and panic Download full-text
disorder: Is the comorbidity more than coincidence? Acta Psychiatr
Griebel G, Simiand J, Steinberg R, Jung M, Gully D, Roger P, et al. (2002):
amine hydrochloride (SSR125543A), a potent and selective corticotro-
phin-releasing factor(1) receptor antagonist. II. Characterization in ro-
dent models of stress-related disorders. J Pharmacol Exp Ther 301:333–
Grimm JW, Hope BT, Wise RA, Shaham Y (2001): Neuroadaptation. Incuba-
tion of cocaine craving after withdrawal. Nature 412:141–142.
Hansson AC, Cippitelli A, Sommer WH, Fedeli A, Bjork K, Soverchia L, et al.
(2006): Variation at the rat Crhr1 locus and sensitivity to relapse into
c-fos reduces corticosterone-mediated effects on neurotrophic factor
Heilig M, Egli M (2006): Pharmacological treatment of alcohol dependence:
Heinrichs SC, Koob GF (2004): Corticotropin-releasing factor in brain: A role
in activation, arousal, and affect regulation. J Pharmacol Exp Ther 311:
Hershon HI (1977): Alcohol withdrawal symptoms and drinking behavior. J
Heyser CJ, Schulteis G, Durbin P, Koob GF (1998): Chronic acamprosate
baseline responding for ethanol in rats. Neuropsychopharmacology 18:
Holm S (1979): A simple sequentially rejective multiple test procedure.
Holsboer F (2003): Corticotropin-releasing hormone modulators and de-
Knapp DJ, Overstreet DH, Moy SS, Breese GR (2004): SB242084, flumazenil,
and CRA1000 block ethanol withdrawal-induced anxiety in rats. Alcohol
Krystal JH, Webb E, Grillon C, Cooney N, Casal L, Morgan CA, et al. (1997):
Evidence of acoustic startle hyperreflexia in recently detoxified early
onset male alcoholics: Modulation by yohimbine and m-chlorophe-
nylpiperazine (mCPP). Psychopharmacology (Berl) 131:207–215.
LeDoux JE, Iwata J, Cicchetti P, Reis DJ (1988): Different projections of the
central amygdaloid nucleus mediate autonomic and behavioral corre-
lates of conditioned fear. J Neurosci 8:2517–2529.
Lu L, Hope BT, Dempsey J, Liu SY, Bossert JM, Shaham Y (2005): Central
amygdala ERK signaling pathway is critical to incubation of cocaine
craving. Nat Neurosci 8:212–219.
Möller C, Wiklund L, Sommer W, Thorsell A, Heilig M (1997): Decreased
ing central but not basolateral amygdala lesions. Brain Res 760:94–101.
Nie Z, Schweitzer P, Roberts AJ, Madamba SG, Moore SD, Siggins GR (2004):
Ethanol augments GABAergic transmission in the central amygdala via
CRF1 receptors. Science 303:1512–1514.
O’Dell LE, Roberts AJ, Smith RT, Koob GF (2004): Enhanced alcohol self-
Overstreet DH, Knapp DJ, Breese GR (2002): Accentuated decrease in social
interaction in rats subjected to repeated ethanol withdrawals. Alcohol
withdrawal-induced anxiety-like behavior by CRF and CRF1 receptors.
Paxinos G, Watson C (1998): The Rat Brain in Stereotaxic Coordinates, 4th ed.
San Diego: Academic Press.
Potter E, Sutton S, Donaldson C, Chen R, Perrin M, Lewis K, et al. (1994):
and withdrawal: 2. Behavioral changes during prolonged abstinence.
Rassnick S, Heinrichs SC, Britton KT, Koob GF (1993): Microinjection of a
corticotropin-releasing factor antagonist into the central nucleus of the
amygdala reverses anxiogenic-like effects of ethanol withdrawal. Brain
in anxiety and depression. Curr Opin Pharmacol 2:23–33.
brain after intermittent exposure to alcohol. FASEB J 16:27–35.
of persistent alcohol preference: Behavioral evidence in a rat model of
Roberto M, Madamba SG, Stouffer DG, Parsons LH, Siggins GR (2004): In-
creased GABA release in the central amygdala of ethanol-dependent
rats. J Neurosci 24:10159–10166.
Roberts AJ, Heyser CJ, Cole M, Griffin P, Koob GF (2000): Excessive ethanol
drinking following a history of dependence: Animal model of allostasis.
factor in drug addiction. Pharmacol Rev 53:209–243.
Shaham Y, Shalev U, Lu L, de Wit H, Stewart J (2003): The reinstatement
model of drug relapse: History, methodology and major findings. Psy-
(2001): Local 5,7-dihydroxytryptamine lesions of rat amygdala: Release
in punished drinking, unaffected plus-maze behavior and ethanol con-
sumption. Neuropsychopharmacology 24:430–440.
Spanagel R, Zieglgansberger W (1997): Anti-craving compounds for etha-
nol: New pharmacological tools to study addictive processes. Trends
(2005): Repeated ethanol exposure and withdrawal impairs human fear
hippocampus. Biol Psychiatry 58:392–400.
Swanson LW, Petrovich GD (1998): What is the amygdala? Trends Neurosci
Valdez GR, Koob GF (2004): Allostasis and dysregulation of corticotropin-
opment of alcoholism. Pharmacol Biochem Behav 79:671–689.
Valdez GR, Roberts AJ, Chan K, Davis H, Brennan M, Zorrilla EP, et al. (2002):
Increased ethanol self-administration and anxiety-like behavior during
acute ethanol withdrawal and protracted abstinence: Regulation by
Valdez GR, Sabino V, Koob GF (2004): Increased anxiety-like behavior and
ethanol self-administration in dependent rats: Reversal via corticotrop-
in-releasing factor-2 receptor activation. Alcohol Clin Exp Res 28:865–
Valdez GR, Zorrilla EP, Roberts AJ, Koob GF (2003): Antagonism of cortico-
tropin-releasing factor attenuates the enhanced responsiveness to
Van PK, Viau V, Bittencourt JC, Chan RK, Li HY, Arias C, et al. (2000): Distribu-
tion of mRNAs encoding CRF receptors in brain and pituitary of rat and
and stress phases on voluntary alcohol intake. Alcohol Clin Exp Res 27:
immunoreactivity and plasma corticosterone during protracted drug
W.H. Sommer et al.
BIOL PSYCHIATRY 2008;63:139–145 145