| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
,1
* Laboratory of Clinical and Translational Studies, and
Laboratory of Physiologic Studies, NIAAA,
¶ Laboratory of Molecular Pharmacology, NCI, National Institutes of Health, Bethesda, USA;
Department of Clinical Neuroscience, Karolinska Institutet, Stockholm, Sweden;
Department of Mental Health and Alcohol Research, National Public Health Institute, Helsinki, Finland; and
|| Department of Experimental Medicine and Public Health, University of Camerino, Camerino, Italy
1Correspondence: Laboratory of Clinical and Translational Studies, NIAAA/NIH, 10 Center Dr, B 10, R 15330, Bethesda, MD, USA. E-mail: wolfgang.sommer{at}mail.nih.gov
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
2-fold higher Gsta4 transcript levels in several brain regions of ethanol-naive AA compared with ANA rats. Differences in mRNA levels were accompanied by differential levels of GSTA4 protein. We identified a novel haplotype variant in the rat Gsta4 gene, defined here as var3. Allele frequencies of var3 were markedly different between AA and ANA rats, 52% and 100%, respectively. Gsta4 expression was strongly correlated with the gene dose of var3, with 60% of the variance in expression accounted for by genotype at this locus. The contribution of glutathione S-transferase expression to the ethanol-preferring phenotype is presently unclear. It could, however, underlie observed differences in life span between AA and ANA lines, prompting a utility of this animal model in aging research.—Björk, K., Saarikoski, S. T., Arlinde, C., Kovanen, L., Osei-Hyiaman, D., Ubaldi, M., Reimers, M., Hyytiä, P., Heilig, M. Sommer, W. H. Glutathione-S-transferase expression in the brain: possible role in ethanol preference and longevity.
Key Words: alcoholism • aging • animal model • microarray • haplotype
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
. Several pairs of divergent rat lines have been independently generated. The most studied lines are the ethanol-preferring P and AA lines and their ethanol nonpreferring counterparts NP and ANA, respectively. Both high-drinking lines have been suggested to model various aspects of human alcoholism, including a high voluntary alcohol intake sufficient to produce reinforcement from ethanol attributed to its pharmacological effects (2
, 3)
. In addition, the alcohol-preferring lines show good predictive validity with respect to the effects of drugs used for treating alcohol-dependent humans, including naltrexone and acamprosate (4)
.
The AA line, which has now been maintained for more than 90 generations, shows some phenotypic traits thought to be characteristic of human early onset alcoholism (5)
. Recently it became evident that the nonpreferring ANA counterpart line has a significantly shorter life span than AA rats. Independent of ethanol exposure, the death rate by 24 months in AA rats was less then one-third of that in the ANA line (6)
. A correlation between ethanol avoidance and shortened life span has not been observed before and the underlying mechanism remains unclear.
In theory, the bidirectional selection pressure enriches the frequency of alleles with either positive or negative influences on ethanol preference in the two populations under selection. Thus, global profiling of gene expression patterns comparing each pair of lines may reveal putative susceptibility genes, provide a tool for identifying molecular mechanisms underlying differential gene expression, and ultimately point to novel targets for drug development. This strategy was used in two independent studies conducted in P- and NP-derived inbred lines (iP, iNP). Among the strongest difference found in both studies was differential expression of the class alpha glutathione-S-transferase isoform alpha 4 (Gsta4, RefSeq XM_217195) (7
, 8)
. iP rats had lower Gsta4 mRNA and protein levels in several brain regions, including hippocampus and cortex. Subsequent sequence analysis revealed four single nucleotide polymorphisms (SNPs) in the iP Gsta4 gene, suggesting a possible relation between genetic polymorphism of Gsta4 and the ethanol drinking behavior of these rats.
Glutathione-S-transferases (GSTs) are a large family of multifunctional proteins that are essential for disposal of exogenous toxic compounds and the adaptive, antioxidant response to reactive oxygen species (ROS). The GSTs are encoded by eight different gene families: alpha, chi (or omega), mu, pi, theta, zeta, and kappa (9)
. The main function of these enzymes is to catalyze the formation of glutathione-S-conjugates with electrophiles, whether of endo- or exogenous origin, which is crucial for inactivation and subsequent excretion of these molecules (10
11
12)
. The Gsta4 isoenzyme has the highest efficiency among all GSTs in conjugating GSH to 4-hydroxynonenal (4-HNE), a major end product of lipid peroxidation and one that is involved in stress-mediated signal transduction (13)
. In addition, the GST alpha class accounts for a large part of GSH peroxidase activity, and as such is an essential part of the cellular antioxidant defense (14)
. In the brain, most GSTs are located in glial cells, which are rich in GSH and may protect neurons with low GSH against oxidative insults. Neurons are particular vulnerable to oxidative stress: they are nonreplicating cells, have high metabolic rates, and contain relatively low levels of natural antioxidants. Thus, with age there is a general functional decline of neuronal cells. The importance of GSH for neuropsychological traits was highlighted recently by two reports linking impaired brain GSH utilization and metabolism to anxiety, memory impairment, and age-related neurodegeneration. First it was shown that glutathione reductase 1, which controls GSH levels, and glyoxalase 1, which uses GSH as a cofactor, regulate anxiety in mice (15)
. The other study demonstrated that mutant mice lacking the excitatory amino acid carrier-1 (EAAC1), which besides neurons is also located in blood-brain barrier endothelial cells (16)
, show neuronal GSH deficiency, and with aging develop brain atrophy, memory impairment, and aggressiveness (17)
.
Several lines of research provide evidence for a putative role of GSTs in the development of alcoholism. Oxidative stress is one of the probable mechanisms involved in alcohol-induced neural damage. Alcohol treatment has been shown to induce oxidative stress and deplete the levels of GSH in primary astrocytes (18
19
20)
. The metabolism of dopamine, a neurotransmitter important for alcohol reward, also produces oxidation metabolites that are eliminated by GST activity (21
, 22)
. Lipid peroxidation products, such as 4-HNE, can directly modulate the binding and functional properties of dopamine D1/D5 receptors, as well as effector proteins within their signaling pathway (23)
. Furthermore, human allelic variants of the GSTM1 gene (glutathione S-transferase mu1) have been associated with increased alcoholism and liver disease (24
, 25)
.
Because of stochastic processes and differences in founder populations, each selected line acquires a different combination of genes contributing to the given phenotype, making it difficult to identify the key genes contributing to the selected trait. However, any candidate genes that are found to be consistently enriched in different independently developed models of high or low ethanol preference are most likely to underlie the genetic susceptibility to drink alcohol. Here, we queried our emerging database of expression profiles from medial prefrontal cortex (mPFC) of ethanol naive AA and ANA rats (26
27
28)
for differential expression of GSH metabolizing enzymes in these lines. The mPFC plays an important role in behavioral inhibition and decision-making, and is viewed as a key region in the development of substance dependence. We recently demonstrated that genetic impairment of endocannabinoid metabolism in this brain region contributes to the high ethanol preference in the AA line (29)
. The microarray data indicated a differential expression pattern of several GSTs between AA and ANA rats, among them Gsta4 and the mu class genes. We verified the microarray screening finding in the mPFC by quantitative real-time RT-PCR, and added three other key brain regions involved in addictive behaviors: nucleus accumbens, amygdala, and hippocampus. We then sequenced the Gsta4 gene of the AA and ANA rats in search of a genetic variability that might underlie expression, and ultimately the phenotypic differences.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
: medial prefrontal cortex, nucleus accumbens, and amygdala were collected bilaterally, and the hippocampus was collected separately from both hemispheres. Samples (10–60 mg depending on region) were obtained under a magnifying lens using anatomical landmarks (30)
. Amygdala and hippocampus samples were prepared from a 2 mm-thick coronal slice taken in a Kopf brain slicer by placing the rostral blade on the caudal edge of the optic chiasm. For preparation of medial prefrontal cortex (mPFC, containing Cg1, Cg2, and PrL (30)
and nucleus accumbens, the blades were placed 3 and 5 mm rostral to this landmark, and a second 2 mm coronal slice was obtained. Cortical and hippocampal tissue was dissected out with a scalpel, while amygdala and accumbens tissue was obtained using a punch (2 mm diameter). Total RNA was extracted from individual rats as described.
Microarray analysis
Target preparation was done for individual samples from mPFC (n=8/line) and hybridization to RAE230A arrays, staining, washing, and scanning of the chips were performed according to the manufacturer’s technical manual (Affymetrix, Santa Clara, CA, USA). CEL files produced by Microarray Analysis Suite 5.0 (MAS5) were inspected for regional hybridization bias and quality control parameters, as recently described (31
, 32)
. Fifteen chips passed the quality control. The MAS5 recognized 60% of the 15,800 probe sets on the RAE230A array as present in our samples. CEL files were imported into the Genesifter platform (gs2.genesifter.net). Robust Multichip Average (RMA) expression values were obtained and tested for differential gene expression using Welch’s two-sample t test, assuming unequal variances. Correction for multiple testing was done by adjusting the P value for family-wise error rate according to Holm’s sequentially rejective multiple test procedure (33)
.
For meta-analysis of separate studies, we used established statistical tools that normalize the observed variance in independent studies (Cohen’s d statistic; ref. 34
, 35
). The advantage of Cohen’s d allows combining different data sets without the need for normalization of the raw data, and it can detect small differences that are consistent in change, even though these changes would not be detected in any single data set. The following formula was applied: d = 2t/
df, where t is the t value from each experiment and df is degree of freedom. The direction of change was post hoc coded into d; positive values indicate up-regulation in the preferring line.
Quantitative real-time RT-PCR
A separate set of rats was used for quantitative real-time RT-PCR analysis. Reverse transcription was done using 100 ng total RNA per animal. The following primers and probe (5'-3', final concentration in parentheses) were used: Gsta4: forward: GAA GTT CTA GTG ACA GCG TGC TTT A (900 nM), reverse: TGT AGC TGC TGC TGT GAT TGG (900 nM), Gsta4 probe: FAM-ACC CTT GCA GTA GCC A-MGB (250 nM) (7)
. Cyclophilin A (Cyca): forward: TGT GCC AGG GTG GTG ACT T (300 nM), reverse: TCA AAT TTC TCT CCG TAG ATG GAC TT (300 nM), Cyca probe: FAM-CCA CCA GTG CCA TTA TGG CGT GT-TAMRA (175 nM) actin (Actx): forward: CCC TGG CTC CTA GCA CCA T (200nM), reverse: GAG CCA CCA ATC CAC ACA GA (200 nM), Actx probe: FAM-ATC AAG ATC ATT GCT CCT CCT GAG CGC-TAMRA (200 nM). All probes were labeled with the reporter dye 6-carboxy-fluorescein (FAM). The primer/probe sets had comparable efficiency of amplification. Real-time RT-PCR was performed in TaqMan Universal Master Mix (Applied Biosystems Foster City, CA, USA) on an ABI Prism 7700, according to the instructions of the manufacturer (Applied Biosystems), and repeated at least twice. Data were expressed as ratio of Gsta to an endogenous control (Cyca or Actx). ANOVA was used for statistical evaluation.
Sequence analysis
The initial sample set analyzed included 20 AA and 20 ANA rats. The animals were sacrificed, and blood samples were collected to EDTA tubes and stored at –20°C until DNA isolation. Genomic DNA was isolated from 3 ml of whole blood using Puregene DNA isolation kit (Gentra Systems, Minneapolis, MN, USA). All exons and 1 kb of the promoter region were analyzed for sequence variation in a subset of 5 AA and 5 ANA rats. Since we only found evidence for major sequence variation in the promoter region and exon 6, we collected sequence data from these regions for the entire sample set.
The following primers were used (5' to 3'): Exon 1: CGG CTG TAA ATG GAG GCG A and CCC TCC CCT CCT TCG TAT CTA; Exon 2: CCC AGT CAC ATG CAT CCC A and TTT AAG GGA CTT GGG TGC TGA; Exon 3: CCC CTC TCT TTC ACA TCC CA and GCC CAA GCC ACA TCA ACT CTA; Exon4: AGC TGG TTT GGT CTG ACT TGC and CGG CAC CCG TCT TGT TTA TC; Exon 5: TTC GTG CTT GAG GGT TGG CT and CCC TGA CGG ACC CTA CAG TG; Exon 6: AGC CCC TCA TTT CAC CAT T and GTC GAT CTA CCA CCT CAT AAG G; promoter: TTC GTT TGG TGG TTT GTT GC and ATC GAC TCC ATC CTT CCC CT.
In addition to the protein coding regions, the PCR fragments covered the exon-intron splice sites as well as variable lengths of flanking sequences. PCR reactions were carried out in a total volume of 50 µl in the presence of 1 x PCR buffer (5 mM Tris-HCl, pH 8.0, 10 mM NaCl, 0.01 mM EDTA, 0.1 mM DTT, 5% glycerol, 0.1% TritonX-100), 0.2 mM of each dNTP, 1.0 mM of each primer, 100 ng genomic DNA as template, and 1.25 U of TaqDNA polymerase (Promega, Madison, WI, USA). After initial denaturation at 94°C for 2 min, 40 cycles of 30 s at 94°C, 30 s at 60°C, 30 s at 72°C, and a final extension of 10 min at 72°C were performed in a PTC-200 thermal cycler (MJ Research, Waltham, MA, USA). The PCR products were purified by QIAquick PCR Purification Kit (Qiagen, Valencia, CA, USA). Sequencing was done by employing primers used in the initial PCR reaction, and ABI PRISM Big Dye Terminator v3.0 sequencing kit according to manufacturer’s instructions. The analyses were performed on an ABI PRISM 3100 Genetic Analyzer (Applied Biosystems).
To compare genotype and gene expression, all animals used in the expression studies were genotyped for promoter and exon 6 using the same method as described above. Genomic DNA was isolated from brain tissue using an ABI PRISM Nucleic Acid Prepstation according to the manufacturer’s instructions.
Western blot
Hippocampus from the same animals as in the microarray study was used (n=8/line). Samples were prepared in lysis buffer consisting of 137 mM NaCl, 20 mM Tris-HCl ph 8.0, 1% Nonidet P-40, 10% glycerol, 1% protease inhibitor cocktail (Sigma) using 10 µl per mg brain tissue. Aliquots from each sample were resolved on 10% SDS-PAGE gels and transferred to nitrocellulose membranes. The membrane was blocked by incubation for 1 h in TPBS buffer (1x PBS, 0.1% Tween 20) containing 5% (w/v) nonfat dry milk at room temperature, then immunoblotted with GST-4 alpha primary antibody (Ab) (U.S. Biological, Swampcott, MA, USA; at 1:1000 dilution in TPBS plus 3% BSA) overnight at 4°C and horseradish peroxidase-conjugated secondary Ab (1:2000 in blocking buffer) for 1 h at room temperature. Immunoreactive bands were then detected using enhanced chemiluminescence (ECL) detection system (Amhersham Biosciences). The level of GST was compared with beta-actin running controls. Densitometry of protein band intensity was quantified using the NIH image-J software.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
15,800 oligonucleotide probe sets. Normalization and estimation was done using the RMA procedure. To maximize statistical power, we carried out a focused analysis of a subset of genes directly involved in glutathione metabolism.Searching Affymetrix (www.affymetrix.com) and other public databases, we found that the REA230A chip contains 29 probe sets representing genes directly involved in glutathione metabolism (Table 1
), i.e., glutathione S-transferases (EC 2.5.1.18), glutathione peroxidases (EC 1.11.1.9)/reductases (EC 1.8.1.7, 1.8.1.9)/synthetase (EC 6.3.2.3), and gamma-glutamyl transpeptidase/transferase (EC 2.3.2.2).
|
Welch t test, followed by Holm’s sequentially rejective multiple test procedure, demonstrate significant differences between AA and ANA in the expression of Gsta4 and Gpx3 genes, with both these transcripts expressed at a higher level in the AA line. In addition, glutathione S-transferases mu 1 and mu 3 (Gstm1 and Gstm3), kappa 1 (Gstk1), as well as glutathione peroxidase 1, show an uncorrected P < 0.05, providing suggestive evidence for differential expression in the mPFC.
All genes of the Gst mu class reside in close physical proximity (2q34). We hypothesized that such closely located paralogs may supplement each others’ function or might be commonly regulated in the associated phenotype, although no one gene shows strong effects consistently. This was tested by adding the intensity values for all five genes to give an overall score for the locus. We then applied a Student’s t test to these scores. We found a significant up-regulation in the AA line for the Gstm1–5 cluster (P<0.015).
Additional statistical power may be obtained by combining expression data from several genetic models of ethanol drinking. For this meta-analysis we used the standardized effect size, Cohen’s d. Using the published data on iP/iNP rats for Gstm2 and Gstm3 (8)
, we find an average d of 1.43 and 1.90, respectively. According to Cohen, these values can be interpreted as 68% and 80%, respectively, of nonoverlap of the preferring group’s scores with those of the nonpreferring group (35)
.
Quantitative real-time RT-PCR for Gsta4
Real-time RT-PCR was used for a targeted secondary analysis of the initial microarray findings in the medial prefrontal cortex and to test Gsta4 expression in three other key brain regions of dependence: the nucleus accumbens, amygdala, and hippocampus. No difference in Cyca gene expression was found between AA and ANA rats in these regions (mPFC: F[1,12]=0.86, P>0.05; Acc: F[1,14]=0.27, P>0.05; Amy: F[1,14] =1.16, P>0.05; Hip: F[1,14]=1.28, P>0.05), and this gene could thus be used as an endogenous reference. After normalization to Cyca, Gsta4 expression was 2- to 2.5-fold higher in all four regions of the AA rats compared with ANA, and these differences were highly significant (mPFC: F[1,12]=8.15, P<0.05; Acc: F[1,14]=56.55, P<0.001; Amy: F[1,14]=37.88, P<0.001; Hip: F[1,14]=18.26, P<0.001, Fig. 1
). Similar results were obtained using Actx for normalization (data not shown).
|
Western blot
Western analysis demonstrated that the differential Gsta4 mRNA expression found between the AA and ANA lines is paralleled by similar differences in protein levels (Fig. 2
).
|
Sequence analysis
Sequence analysis of Gsta4 in AA and ANA rats revealed that both strains carry a novel allelic variant. SNPs were found at nucleotide positions –807, –385, +624, +710, +723, and + 752 (Fig. 3
A). We used the Haploviewprogram (36)
to create a linkage disequilibrium (LD) map of these markers and to carry out a statistical haplotype reconstruction. All of the six SNPs except –807 were in strong LD with each other and belong to a common haplotype block (Fig. 3B
). Two different haplotypes were identified, but since these were identical except for the T/C variation at position –807, they are referred to here as var3 (Fig. 3B
, Table 2
). This variant is similar to that found in iP but differs at positions +777 and +792–793 (Table 2)
. In the AA population we found both the var3 haplotype and the wild-type (WT) sequence (i.e., the one originating from Rattus norvegicus represented by the GeneBank reference sequence XM_217195 and the locus ENSRNOG00000030449 at the ENSEMBL Genome Browser). Despite the size of the haplotype block, spanning 14 kb, there is no evidence for recombination between WT and var3 in the AA line. Both haplotypes are equally common in the AA population, and seem to follow a mendelian pedigree (Fig. 3C
). The ANA rats carry only the var3 haplotype. However, in the ANA population there is variation for the –807 allele. Whether SNP –807 belongs to or in fact determines the boundary of the haplotype block remains to be seen. Because of the sparse information on genomic variation in the rat, the average size of a haplotype block in this species is not estimated, and thus it is difficult to predict how far the present haplotype block does extend.
|
|
Haplotype gene expression comparison
To determine the contribution of Gsta4 haplotype to mRNA expression at this locus, we first tested whether the variation at SNP –807 in the ANA population affected Gsta4 mRNA levels. There was no difference in gene expression between the two alleles (one-way ANOVA, F[1,22]=1.07, n.s.). Each individual irrespective of line was then assigned a gene dose of var3 (var3/var3=2, var3/wt=1, wt/wt=0) and a linear regression of mRNA expression vs. gene dose was carried out (Fig. 4
). mRNA expression correlated closely and in a linear manner with var3 gene dose. Regression analysis indicated that the genotype at this locus contributes to 60% of the variance in expression levels (P <0.00001).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
2-fold increase of Gsta4 expression in the AA rat was detected in all four of the brain regions examined. The difference in Gsta4 mRNA levels is accompanied by corresponding changes in protein levels.
Genetic polymorphism has been described in many genes of the GST gene family (37)
. In search of genetic variability potentially underlying the differential Gsta4 expression in AA and ANA rats, sequence analysis of the gene was performed. In addition to the WT allele and the exon 6 variants described by Liang et al. (7)
, analysis of AA and ANA lines revealed the existence of a Gsta4 haplotype variant not previously described.
Our results, together with the previously published iP/iNP data, suggest a correlation between Gsta4 haplotype and gene expression. The var3 haplotype ubiquitous in ANA rats resembles the iP variant; the ANA rats, as do iP rats, demonstrate lower levels of the Gsta4 transcript. In contrast, the allelic variant found in the iNP line exhibits the most similarity to the WT allele that was present in the AA population, suggesting that these alleles are associated with higher expression of Gsta4. The genetic heterogeneity of the AA population may explain the higher variance observed in the Gsta4 mRNA levels in this line. Of the six SNPs, two reside in the promoter and are not in LD. SNP –807 is located in the core region of an interleukin (IL) Il-6 response element (38)
. The functional relevance of this element in brain cells remains to be demonstrated. However, strong up-regulation of Gsta4 mRNA by Il-6 was found in cultured hepatocytes (39)
. Noteworthy, IL-6 plays an important role in the endothelial-astroglial interactions regulating blood-brain barrier permeability (16)
. The SNP –385 is located in a mouse-rat conserved binding motif for the transcription factor Pax2 (www.cisreg.ca). However, the expression of Pax2 seems to be regionally restricted to the cerebellum, the ventral midbrain, and the hindbrain and may therefore not account for the global Gsta4 expression differences in the brain of AA and ANA rats. Of the four SNPs on exon 6, one is coding but synonymous, and the others occur within the untranslated region (3'-UTR) of the mRNA. SNP +710 and +723 are located within binding motifs for microRNAs. Of the >200 known rodent microRNA species, only 10 have predicted binding sites in the Gsta4 3'-UTR; among these, miR-17 and miR-1 are located at SNPs +710 and +723 miR-17 and miR-1, respectively (40)
. Notably, var3 increases Watson-Crick base pairing to these microRNAs. MicroRNAs are small, highly conserved, 22–25-mer RNA molecules that bind to the 3'-UTR of specific mRNAs. They are increasingly recognized as important tissue specific regulators of mammalian gene expression. Whereas miR-17 seems to be restricted to embryonic tissue, miR-1 is expressed in the brain (41
, 42)
. Thus, the differences in brain Gsta4 mRNA levels between AA and ANA rats could reflect a post-transcriptional miRNA-mediated mechanism. Whether the SNPs are in fact directly involved in the regulation of mRNA levels or are merely in linkage disequilibrium with other functional polymorphisms remains to be determined.
The difference in expression of the Gsta4 gene between AA and ANA rats is opposite in direction to previous results obtained by comparing iP alcohol-preferring and iNP nonpreferring rats, in which Gsta4 mRNA levels were elevated throughout the brain in the nonpreferring line (iNP) and decreased in the preferring iP line. The phenotypic selection criteria used for developing the P/NP and AA/ANA lines were identical, but the genetic composition of the foundation stocks were different, with the P/NPs selected from an outbred Wistar colony and the AA/ANAs developed using genetic material from several strains, including Wistar, Sprague-Dawley, Lewis and Brown Norwegian strains (43
, 44)
. Thus, differential expression of Gsta4 between the AA/ANA and iP/iNP rats, respectively, is not consistently associated with the alcohol-preferring phenotype of these lines. There are several possible explanations for this finding. One possibility is that the Gsta4 gene is situated in the vicinity of loci affected by the selective breeding for alcohol preference and that its variants have cosegregated with those under selection pressure without themselves being functionally related to this phenotype. Possibly supporting this notion, it has been shown that the Gsta4 gene (chromosomal location: 8q31) is located near a suggestive quantitative trait locus (QTL) region for alcohol preference in the iP rats (45)
. Alternatively, the reduced diversity of the Gsta4 locus in the selected lines could reflect founder effects inherent in the selection procedure, followed by random segregation of the represented alleles. Finally, the possibility should not be excluded that selection for high and low alcohol preference, respectively, has led to selection of the respective variants in a manner not understood at this time.
It is presently unclear what functional consequences the altered expression of the Gsta4 gene may have. The great diversity of the GST family and the absence of a general effect on overall glutathione-S-transferase or peroxidase activities (data not shown), despite different protein levels between the lines, may point to highly specific functions of the GSTA4 isoform. Available immunohistochemical data on GSTA4 distribution in the brain demonstrate a predominant location of this immunoreactivity in ependymal cells of the chorioid plexus, endothelial cells, and perivascular endfeet of astrocytes (46)
. These data point to a particular role of this isoenzyme in the blood-brain barrier that is thus strategically positioned in the defense of the immediate microenvironment of brain cells (16)
. In tissue culture, Gsta4 overexpressing cells have lower steady-state levels of 4-HNE (13)
. Furthermore, the demands on Gsta4 seem to increase with aging, since its expression in the cerebral cortex of old rats is nearly doubled compared with young adult animals (47)
.
There is no known association of the GST alpha class with alcoholism or other neuro- or psychopathology. However, a number of such findings are reported for the mu class, which appears to be differentially expressed between the AA and ANA animals as well. Human GSTM1 null mutation has been associated with an increased risk of alcoholism (24
, 25)
but not with severe alcohol withdrawal syndrome (48)
. This appears to be opposite to the findings in the ethanol-preferring animal, where the mu class seems to be up-regulated (see below), and might point to important differences between the ethanol-preferring phenotype in animals and alcoholism as a human psychopathological state. GSTM1 absence has been repeatedly associated with an increased risk for schizophrenia and tardive dyskinesia (49
50
51)
. Conversely, mood stabilizers such as lithium or valproate induce Gstm1 and Gsta4 expression, inhibit oxidative damage to lipids and proteins, and consequently produce a neuroprotective effect against excitotoxicity (52
, 53)
.
Considering the wide array of symptoms pointing to progressive and severe brain impairment associated with multiple GST dysfunction, a behavioral outcome reflecting the differential GST expression between AA and ANA rats would be expected. In fact, it was recently found that ANA had a significantly shorter life span than AA, and apparently also that of common laboratory rats (6)
. This effect was independent of lifelong ethanol exposure. It is conceivable that an altered efficiency of the ROS defense system might influence processes important for longevity and subsequently affect the life span of these animals. This hypothesis is in agreement with studies of Gsta4 –/– mice, which are more susceptible to oxidative stress and show degenerative changes earlier in life than wild-type controls (54)
, in long-living Ames dwarf mice that show increased GST activity (55)
, as well as in fruit flies where disruption of GST genes reduces life span (56)
. Furthermore, we noted decreased activity, increased anxiety, and impairment in memory performance in the Morris-Water Maze test in the ANA rats (5
, 43)
and unpublished data). Whether selection for ethanol nonpreferring phenotypes is commonly linked with decreased GSH utilization and associated neuropathology needs to be established. Across a range of rat and mice models of ethanol preference, the expression of Gstm isoforms seems to be consistently decreased in the nonpreferring lines studied in the hippocampus of iP and iNP rats (8)
, as well as in total brain extracts from selected as well as inbred mouse lines with high and low ethanol consumption (57
, 58)
. Together, these data emphasize that the bidirectional selection procedure generates two phenotypes, both deviating from "normal, " rather than one being the control for the other. Thus, whereas AA rats are in fact a useful model of ethanol preference, the ANA line has not yet been exploited as a potential model of aging or neurodegenerative processes.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
Received for publication February 10, 2006. Accepted for publication April 17, 2006.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  K. Bjork, R. Rimondini, A. C. Hansson, A. Terasmaa, P. Hyytia, M. Heilig, and W. H. Sommer Modulation of voluntary ethanol consumption by beta-arrestin 2 FASEB J, July 1, 2008; 22(7): 2552 - 2560. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
