| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
,
* Laboratory of Clinical and Translational Studies, National Institute on Alcohol Abuse and Alcoholism, National Institutes of Health, Bethesda, Maryland, USA;
Department of Clinical Neuroscience, Karolinska Institutet, Stockholm, Sweden;
Department of Pharmacology, University of Bologna, Italy; and
Department of Mental Health and Alcohol Research, National Public Health Institute, Helsinki, Finland
1Correspondence: Laboratory of Clinical and Translational Studies, NIAAA/NIH, 10 Center Dr., B 10, R 15330, Bethesda, MD, USA 20892-1108. E-mail: wolfgang.sommer{at}mail.nih.gov
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: alcoholism • animal model • brain • gene expression • ethanol preference
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. In the brain, GPCR signaling is involved in slow synaptic neurotransmission, a process particularly important for long-term neuroadaptation and plasticity, both of which underlie fundamental cognitive and psychological processes, such as learning, memory, affective state, and reward (2)
. Among the four members of the arrestin family, beta-arrestins 1 and 2 are expressed throughout the brain. Although they have partly overlapping functions, it has also been shown that preferential recruitment of the respective arrestin is highly dependent on receptor and ligand (3
, 4)
. In contrast to beta-arrrestin 1, Arrb2 interacts with receptors that mediate rewarding properties of common drugs of abuse, such as µ-opioid and dopamine D2 receptors (5
6
7
8)
. For example, mice lacking Arrb2, but not mice without beta-arrestin 1, show increased sensitivity to morphine, a prototypical opiate that acts mainly through µ-opioid receptors (8)
. The effects of Arrb2 on µ-opioid and dopaminergic signaling are classically attributed to its role in ligand-induced receptor desensitization, but more recently, the expression of psychostimulant actions has been shown to be mediated through an Arrb2-dependent formation of a kinase/phosphate scaffolding complex at dopamine D2 receptors (6)
. Although these and other results suggest a role for beta-arrestin 2 in substance dependence, there are presently no data on how voluntary drug intake is modulated by variation in Arrb2 expression and function.
Ethanol is consumed to millimolar concentrations in humans and is reinforcing in experimental animals. Its use is a major cause of morbidity and mortality. Recent data indicate that alcohol use accounts for 
85,000 deaths/yr in the United States alone, making it the third externally modifiable cause of mortality (9)
. The most severe form of alcohol use disorder, alcohol dependence, or alcoholism, occurs in 5% of the adult population and has a 50–60% heritability. This is thought to be contributed by multiple loci, and the contribution of individual candidate genes accounts for not more than 5–10%, in a pattern common for complex polygenetic disorders resulting from gene-by-environment interactions (10
, 11)
. Identification of heritable susceptibility factors might offer improved and ultimately individualized pharmacological treatment for this disorder.
In contrast to most other drugs of abuse, ethanol does not have a defined site of pharmacological action, but it is currently thought to act on specific proteins rather than via nonspecific membrane perturbations (12)
. Ethanol’s reinforcing properties result, in part, from interactions with opioid and GABA neurotransmission within the mesolimbic dopaminergic pathway (13)
. Although Arrb2 is involved in signaling through both the opioid and the dopaminergic components of this cascade, its role in ethanol consumption and reward has not been addressed. A link between ethanol reward and Arrb2 was suggested by our prior microarray experiments involving the brains of experimental animals. These data pointed to genetic variation at this gene locus between two rat lines, the AA and ANA rats, selectively bred for high ethanol preference or ethanol avoidance, respectively (14
, 15)
.
To investigate a possible role of Arrb2 in ethanol reward, we resequenced the Arrb2 locus in AA and ANA rats and found complete segregation between the lines at this locus. We discovered a novel haplotype variant in AA rats that spans the entire Arrb2 locus. This genetic variant was associated with higher Arrb2 transcript levels in several brain regions known to control alcohol reward and preference. Western blot analysis demonstrated that the differential gene expression translates into increased Arrb2 protein levels in the AA line. To perform a functional validation of the genetic findings, we studied the consumption and psychostimulant action of ethanol in mutant mice lacking Arrb2.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. Rats were of the 95th generation, aged 3 mo and housed 4/cage with food and water ad libitum on a reversed 12 h light/dark cycle (lights on at 10 PM).
The generation of Arrb2 knockout mice was described previously (8)
. Mutants were generated on a C57BL/6 background. Genotyping of mice was performed using PCR from tail samples. Mutants and wild-type littermates were housed together in standard Plexiglas cages, with food and water ad libitum and on a 12 h light/dark cycle (lights on at 7 AM). Only males 2–3 mo of age were used in these experiments.
Sequence analysis
For sequence analysis, rats (n=15/line) were sacrificed, and blood was collected into EDTA tubes and stored at –20°C until DNA isolation. Genomic DNA was isolated from 3 ml of whole blood by using a Puregene DNA isolation kit (Gentra Systems, Minneapolis, MN, USA). Approximately 1.6 kb of the promoter region and all exons (except for exon 11, which we were unable to amplify) were searched for sequence variation in a subset of 8 AA and 8 ANA rats. Sequencing was done as described; for primer sequences, see Supplemental Table 1 (16)
. Where evidence for sequence variation was found, the whole sample set was analyzed.
In situ hybridization
For in situ hybridization, animals were sacrificed by decapitation; brains were quickly removed, snap-frozen in –40°C isopentane, and stored at –70°C until use. Arrb2 expression in AA and ANA rats was assessed in 8 animals/line. Brain sections were taken at bregma levels 1) +2.5 to + 1.7 mm, 2) –0.3 to –0.4 mm, 3) –1.7 to –2.0 mm, and 4) –2.3 to –3.3 mm, according to the atlas of Paxinos and Watson (17)
. The in situ hybridization protocol has been described previously (18)
. Sense and antisense riboprobes for the Arrb2 and c-fos transcripts correspond to nucleotides 1238–1679 (NM_012911.1) and 306–864 (NM_022197.2), respectively. Fuji BAS-5000 PhosphorImager plates (Fujifilm, Tokyo, Japan) were exposed to the hybridized sections, and PhosphorImager-generated digital images were analyzed using AIS Image Analysis Software (Imaging Research Inc., St. Catharines, ON, Canada). Regions of interest were defined using anatomical landmarks as described in the atlas. On the basis of the known radioactivity in the 14C standards, image values were converted to nCi/g. Expression values were compared regionwise by 1-way ANOVA followed by Holm’s sequentially rejective multiple-test procedure.
Western blot analysis
Tissue samples from hippocampus of AA and ANA rats (n=6/strain) were prepared as described previously (19)
. Samples containing 35 µg total protein were separated on a 10% tris-glycine gel and blotted onto nitrocellulose membranes using the Xcell II system (Invitrogen, Carlsbad, CA, USA). Arrb2 protein was detected using an anti-Arrb2 antibody (Santa Cruz Biochemicals, Santa Cruz, CA, USA) and a mouse-specific secondary antibody (Bio-Rad, Hercules, CA, USA) and alpha-tubulin using a primary antitubulin antibody (Cell Signaling Technology Inc., Beverly, MA, USA) and a secondary anti-rabbit antibody (Jackson ImmunoResearch Laboratories, West Grove, PA, USA). Detection and quantitation were performed using the Fuji LAS-3000 system and Multigauge software (Fujifilm). The Arrb2 and alpha-tubulin signal was normalized against total protein content as measured by Memcode staining (Pierce, Rockford, IL, USA) (19)
. Protein levels were compared by 1-way ANOVA.
Bioinformatic analysis
Quantitative trait locus analysis to define genomic regions associated with the expression of distinct genes (eQTL) was done using the WebQTL program at www.genenetwork.org (20)
. Linkage with Arrb2 expression is expressed in logarithm of odds (LOD) scores and was plotted against chromosomal position. Seven independent data sets from recombinant inbred mouse lines between the C57BL/6J and DBA/2J parental lines (BXD RI strains were used): 1) UCHSC BXD whole brain, from the University of Colorado at Denver; 2) OHSU/VA B6D2F2 brain and 3) OHSU/VA B6D2F2 striatum, from The Oregon Health Sciences University; 4) HQF BXD striatum, from the University of Tennessee; 5) the Hippocampus Consortium; 6) VCU BXD prefrontal cortex, from Virginia Commonwealth University; and 7) GE-NIAAA cerebellum.
Voluntary ethanol consumption
Increasing concentrations of ethanol in a 0.2% saccharin solution were made available in a continuous two-bottle free choice between ethanol-saccharine and saccharine-alone solutions. Bottles and mice were weighed on Mondays, Wednesdays, and Fridays between 9 and 10 AM to follow consumption. Bottle positions were switched daily to avoid development of side preference. Ethanol concentration was increased gradually from days 2 to 4 (2% ethanol+0.2% saccharin), to days 4 to 6 (4% ethanol+0.2% saccharin), and to days 7 to 28 (6% ethanol+0.2% saccharin). To control for putative differences in taste preference, the same mice were allowed ad libitum access to either sucrose (5%) or quinine (0.02 mM) for 14 days, and taste preference ratios were measured.
Ethanol-induced locomotion
Locomotion was assessed after a single injection of ethanol (0.75 g/kg). Infrared locomotor cages were used (27.9x27.9 cm test environment with three 16-beam I/R arrays; Med-Associates, St. Albans, VT, USA), and activity was recorded for 30 min in 10-min intervals. Results were analyzed using 2-way ANOVA for treatment and strain and Tukey’s HSD post hoc test.
Ethanol metabolism and clearance
Blood levels of ethanol were measured after a single ethanol injection (3.5 g/kg, i.p.). Blood samples were obtained from the tail at 15, 60, 120, and 240 min after injection. The samples were subsequently analyzed using Analox analyzer (Analox Instruments, Luneburg, MA, USA). Data were analyzed using repeated-measures ANOVA.
Rotarod
Four-month-old mice, with and without Arrb2, were placed on a rotarod (Ugo Basile, Varese, Italy) turning at a fixed rate of 10 rpm. The mice learned to remain on the rod during three 60-s training periods. Saline or increasing doses of alcohol (0.5 mg/kg dose, range from 0.5 to 3 mg) were injected i.p. After 5 min, each mouse was tested 3 times in succession for its ability to remain on the rod. The cutoff time was 60 s. The ED50 was calculated using nonlinear regression implemented in GraphPad Prism v4 (GraphPad Software Inc., San Diego, CA, USA) and compared by 1-way ANOVA.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
A). The AA-specific, high ethanol preference-associated haplotype, herein referred to as var1, consisted of 7 polymorphisms (6 SNPs and 1 insertion), which were in complete linkage disequilibrium (LD). The sequence of Arrb2 in ANA animals did not differ from the published Rattus norvegicus sequence. Arrb2var1 is the only genetic variant discovered to date that segregates entirely between the AA and ANA lines.
|
In situ hybridization demonstrated that the presence of Arrb2var1 in the ethanol-preferring AA line was associated with significantly higher Arrb2 mRNA levels in the nucleus accumbens, dorsal striatum, and hippocampus (Fig. 1B
). A corresponding difference in hippocampal Arrb2 protein levels was detected using Western blot, while the housekeeping gene alpha-tubulin was unaltered between the two lines. (Fig. 1C
).
Next, we asked whether genetic variation at the Arrb2 locus could affect a cis-regulatory mechanism for expression and regional control of the gene. Bioinformatic analysis in seven independent microarray expression data sets, including samples from whole brain, prefrontal cortex, hippocampus, striatum, or cerebellum of BXD RI strains revealed a single, distinct, and highly significant eQTL for Arrb2, which peaks above its own locus on mouse chromosome 11. The highest linkage at the Arrb2 locus was found in samples from hippocampus and striatum (LOD scores 9.8 and 8.2, respectively. LOD scores above 3.3 are considered significant, Fig. 2
). Prefrontal cortex and two whole brain data sets showed less significant or suggestive eQTLs, with LOD scores of 5.1, 4.3, and 2.5, respectively. No evidence for an eQTL at the Arrb2 locus was found in the samples from cerebellum. Similar eQTLs are not found at Arrb2’s closest neighbor genes, demonstrating that the effect is specific for Arrb2. Furthermore, no eQTL exists for beta-arrestin 1 in any of the data sets.
|
To establish a functional relationship between Arrb2 and ethanol consumption, we used mutant mice lacking Arrb2 (8)
. Both homozygote and heterozygote Arrb2 mutants (Arrb2–/– and Arrb2+/–) consumed significantly less and had lower preference for ethanol than wild-type littermate controls when ethanol was available in a two-bottle free-choice paradigm (6% vol/vol ethanol/0.2% saccharine or 0.2% saccharine only in tap water; Fig. 3
A–C). There was a reciprocal effect on intake from the nonethanol bottle, and total liquid intake did not differ between the genotypes, suggesting that nonspecific differences in regulation of liquid intake and thirst did not account for the reduced ethanol consumption in the mutant mice. There were no significant weight differences at the end of the experiment: 27.6 ± 0.68, 27.1 ± 0.57, and 27.4 ± 0.75 for Arrb2–/–, Arrb2+/–, and Arrb2+/+ animals, respectively.
|
Pharmacological specificity for ethanol was further supported by the taste preference experiments. Alcohol’s taste is thought to be a composite of sweet and bitter. An absence of differences in taste preferences for sucrose and quinine between the genotypes demonstrated that the differences in ethanol intake between genotypes are not likely to be driven by altered taste preference (Fig. 3D, E
).
The ethanol metabolism and elimination experiment demonstrated that drinking differences between genotypes are not driven by pharmacokinetic differences. Arrb2–/– mice did show somewhat lower blood ethanol levels after systemic administration of 3.5 g/kg ethanol compared to wild-type mice, whereas Arrb2+/–mice did not (Fig. 4
A). Since reduced voluntary ethanol intake was found in both Arrb2+/– and Arrb2–/– mice, differences in ethanol metabolism and elimination cannot account for the drinking findings. Similarly, sensitivity to ethanol’s motor-impairing effects was somewhat reduced in heterozygous mice only, and is therefore unlikely to account for the reduced ethanol drinking found both in Arrb2+/– and Arrb2–/– mice (Fig. 4B
). Similarly, the previously reported lower spontaneous locomotor activity of Arrb2 mutant mice (5
, 6)
is unlikely to account for the observed differences in ethanol drinking and sensitivity.
|
Psychomotor stimulation, resulting in increased locomotor activity, is a proxy measure of mesolimbic dopaminergic system activation. Most drugs of abuse, including ethanol, can induce locomotor activity in experimental animals. In line with the findings of reduced voluntary ethanol drinking, Arrb2 mutants were insensitive to the stimulant effects of ethanol (0.75 g/kg; Fig. 5
A). Consistent with the absence of ethanol-induced locomotion in Arrb2–/– mice, ethanol-induced c-fos expression in the shell of nucleus accumbens was also absent in homozygous mutants (Fig. 5B
). No differences in c-fos expression were found in the core of the nucleus accumbens. Other regions were not studied.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
A number of rodent lines have been successfully bred for different ethanol preference and are used in work aimed at identifying targets for novel pharmacological treatments for alcohol dependence (21
, 22)
. The alcohol-preferring AA and the alcohol-avoiding ANA rat lines have been bred bidirectionally for high and low alcohol consumption, respectively, for more than 90 generations (15)
. AA rats display some interesting behavioral traits characteristic of early onset alcoholism in humans, which has a distinct heritability and pharmacogenetic profile (23
, 24)
. The AA and ANA lines have diverged with respect to a variety of behavioral and neurochemical measures, but the factors causing their differential ethanol preference remain largely unknown (15)
.
All candidate genes previously screened for allelic variation between AA and ANA rats have shown incomplete segregation between the two lines, making it difficult to establish a link between genetic findings and a behavioral or expression phenotype (16)
. The complete segregation of AA and ANA animals at the Arrb2 locus into two haplotype variants is a unique finding for noninbred lines and suggests biological significance. The length of the novel haplotype block spans the entire Arrb2 locus, but it is unclear whether a high degree of LD extends beyond this locus. The average length of a haplotype block in rats is currently not known. Furthermore, it is difficult to obtain block size estimates for selected populations because block size is strongly dependent on the genetic diversity of the founder population and will be severely skewed by any random fixation effects that occur during selection. Although this leaves open the possibility for an involvement of neighboring genes in the biological effects of the variant allele, we find several lines of evidence that favor an important role for Arrb2.
First, the in situ hybridization experiment reveals differences in Arrb2 expression between AA and ANA rats in the hippocampus and striatum, two brain regions known to be important in substance dependence (25)
. Furthermore, albeit the differences in Arrb2 expression are modest and restricted to distinct brain regions, their coexistence with widespread genetic variation at the gene locus points to cis-acting mutations, probably altering regulatory elements involved in regional control of gene expression.
Second, a cis-regulatory mechanism for expression and regional control of Arrb2 is supported by bioinformatic analysis. Combination of expression profiling from recombinant inbred mouse lines of the BXD RI panel and genome-wide allelic marker association allow identification of genomic regions that control the expression of individual genes (26
, 27)
. This approach was recently employed to identify novel candidate genes involved in ethanol preference and tolerance (28)
. Here, we found a single, distinct, and highly significant eQTL for Arrb2 at its own locus in hippocampal, striatal, and prefrontocortical tissue but not in cerebellum. These findings strongly suggest that variation at the Arrb2 locus has region-specific impact, and it is important for Arrb2 expression in the hippocampus and striatum, the regions that show the largest Arrb2 expression differences between AA and ANA rats. Furthermore, no eQTL exists for beta-arrestin 1 in any of the data sets, suggesting that Arrb1 and Arrb2 are regulated through independent mechanisms. It remains unknown for now how the polymorphic alleles of Arrb2 contribute to the regulation of its gene expression.
Third, an involvement of a candidate gene in alcohol preference can be supported by a consistent pattern of differential gene expression across different rodent lines with differing ethanol preference. Notably, significant and consistent up-regulation of brain Arrb2 expression was found by microarray analysis in selected, noninbred mouse lines with high ethanol preference compared to their controls (29)
. However, differential Arrb2 expression was not found in the brains of another widely studied pair of rat lines, the Indiana ethanol-preferring P and nonpreferring NP rats, nor does the Arrb2 gene reside in any of the QTLs for ethanol preference that have been identified in rodents so far (reviewed in ref. 28
).
Finally, the most straightforward functional validation for a differentially expressed candidate gene can be obtained by interventions that disrupt the function of its product. When available, we have used pharmacological tools to investigate the role of candidate genes identified on the basis of their differential expression between high- and low-preferring animals, such as endocannabinoid genes in AA rats (30)
. However, in the absence of suitable pharmacological tools or transgenic approaches for probing the function of Arrb2 directly in AA rats, we used a genetic strategy using Arrb2 knockout mice. Ethanol preference is affected in numerous mutant lines, suggesting that the control of this behavior is complex and may also be affected by behaviorally nonspecific effects of the mutations (31)
. Consequently, we tested for and excluded a number of possible confounding factors that might contribute to the reduced ethanol consumption in mutants lacking Arrb2. Neither taste preference nor ethanol metabolism or sedation seemed to account for the reduced ethanol consumption in Arrb2 knockout mice. Importantly, both mice homozygous and heterozygous for the Arrb2 null mutation exhibited markedly reduced ethanol drinking. This finding underscores that ethanol preference and reward are sensitive to changes in Arrb2 expression and function.
The mechanism of action of Arrb2 in modulation of ethanol reward remains unknown. Dopamine-opioid system interactions contribute to ethanol reward (13
, 32)
. Endogenous opioids released by ethanol can act at µ-opioid receptors located on midbrain GABAergic interneurons, leading to disinhibition of dopaminergic neurons in the ventral tegmental area, and dopamine release into their terminal regions (33
34
35
36)
. In addition, similar to morphine, ethanol might act directly on the target neurons of the mesocorticolimbic pathway (37)
, since inactivation of the mesocorticolimbic dopamine neurons does little to modulate ethanol self-administration (38
, 39)
, while pharmacological blockade or gene knockout of µ-opioid receptors substantially reduces ethanol consumption and ethanol-induced locomotion (40
41
42)
.
Arrb2 expression and function appear to be negatively correlated with analgesic effects of opioids, while its role in opioid-mediated reward is less consistent. Deletion of Arrb2 in mice potentiates and prolongs morphine analgesia (7
, 8)
, an effect partially rescued by local overexpression of Arrb2 in brain regions involved in pain control (43)
. However, although both conditioned place preference for opiates, a measure of reward, and exracellular dopamine release in response to morphine are increased in the Arrb2 mutants, morphine-induced locomotor activity is decreased in these animals (5)
. No drug self-administration data are previously available from Arrb2 mutants, and the conditioned place preference data are therefore the only ones that directly address how opioid reward is modulated by Arrb2. These results suggest that, similar to opioid analgesia, opioid reward is inversely related with Arrb2 expression and function. If that indeed is the case, the positive correlation between Arrb2 expression and voluntary ethanol intake observed in AA rats and Arrb2 mutants in the present study cannot be attributed to opioid mechanisms.
It has recently been shown that beta-arrestins can mediate receptor signaling independent of G proteins and the cyclic AMP cascade via a scaffolding complex with certain kinases and/or phosphatases (44)
. This mechanism appears to be crucial for psychomotor stimulation following activation of dopamine D2 receptors. Psychostimulant-induced locomotor activity depends on the formation of a signaling complex at the D2 receptor, which includes Arrb2 and the kinases PP2A and Akt (6)
. It has been suggested that this mechanism underlies the profound reduction in locomotor response to amphetamine and morphine in Arrb2 mutants (5
, 6
, 45)
. Here, we show that ethanol-stimulated activity is also reduced, and disruption of noncanonical D2 signaling could account for this observation. Activation of the ventral striatum plays a crucial role in psychostimulant-induced locomotor behavior (46)
. The actions of psychostimulants, ethanol, and opiates, are all associated with induction of the immediate early gene c-fos in these regions (47
48
49
50
51)
. Although this gene is induced by a variety of stimuli, the c-fos response has been causally linked to locomotor response to psychostimulants (52
53
54)
. Consistent with these findings, c-fos induction after acute systemic challenge with a low dose of ethanol was absent in Arrb2 knockout mice. A possible implication of this finding is that ventral striatal c-fos induction by ethanol is downstream of the recently postulated noncanonical Arrb2 signaling cascade.
Also, since the most robust differences in Arrb2 expression between AA and ANA animals are found in the hippocampus, the possibility exists that this region is involved in ethanol preference in these lines and in Arrb2 mutant mice. In fact, ANA rats have impaired learning and memory compared to AA animals. However, the differences in alcohol drinking reported here were assessed in a simple task that does not require learning. Furthermore, the large body of literature on these lines does not give any indication that memory impairment is a key factor explaining the divergence in ethanol and drug-related behaviors between the lines (15)
.
In summary, we report that Arrb2 expression is positively correlated with ethanol preference, suggesting a role for this protein in mediation of ethanol reward within the mesocorticolimbic system.
|
ACKNOWLEDGMENTS |
|---|
Received for publication November 21, 2007. Accepted for publication February 21, 2008.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
B system. PLoS ONE 2,e930[CrossRef]This article has been cited by other articles:
|
|
 |
|
  R. Spanagel Alcoholism: A Systems Approach From Molecular Physiology to Addictive Behavior Physiol Rev, April 1, 2009; 89(2): 649 - 705. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
