![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
|
|
|
|
|
||
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
NEUROPHARMACOLOGY
Departments of Pharmacology (W.H., L.S., S.V.B., P.L.H., B.T.) and Preventive Medicine and Biometrics (K.K.), University of Colorado School of Medicine, Aurora, Colorado
Received for publication
February 4, 2008
Accepted
June 9, 2008.
 |
Abstract |
|---|
 Top Abstract Materials and MethodsResultsDiscussionReferences |
|---|
). Most measures of alcohol sensitivity in humans and animals include a component that can be ascribed to acute functional tolerance (AFT). AFT is a form of tolerance that develops within a single period of alcohol exposure and has a genetic component. We used microarray technology as well as quantitative trait locus analysis of phenotypic and gene expression data across 30 BXD recombinant inbred strains of mice, 20 inbred strains of mice, and two replicate lines of mice selectively bred for differences in AFT, to identify differentially expressed candidate genes that contribute to predisposition to AFT. Eight candidate genes were identified by our statistical and filtering methods. The location of brain expression of these genes was mapped using the Allen Brain Atlas (http://www.brain-map.org), and the transcript location and molecular pathway analysis indicated that brain structures and biochemical pathways implicated in long-term potentiation and memory might also participate in the generation of acute functional alcohol tolerance.
).
Functional (pharmacodynamic) alcohol tolerance manifests itself in three different forms: chronic (Tabakoff et al., 1986), rapid (Khanna et al., 2002), or acute (Mellanby, 1919). Acute functional tolerance (AFT) refers to the tolerance that is manifested during a single session of alcohol drinking (Mellanby, 1919), and it can start to develop within minutes after an individual starts to imbibe alcohol. The phenomenon of AFT was initially described by Mellanby (1919) through experiments that showed that humans display a greater degree of intoxication at a given blood alcohol concentration on the rising phase of the blood alcohol curve than at the same alcohol concentration on the falling phase of the blood alcohol curve. Acute functional alcohol tolerance is a neuronal resistance to alcohol effects (Tabakoff et al., 1986); however, its molecular mechanism is still not well characterized.
Sons of alcoholics (SOA) are three to five times more likely to develop alcohol dependence than sons of nonalcoholics (Cotton, 1979). SOA also display less alcohol intoxication when measurements are made at least 1 h after drinking alcohol. The low level of response to alcohol has been found to predict future alcohol dependence (Schuckit et al., 2005), and both men and women with no prior history of alcohol dependence, but with a positive family history of alcohol dependence, display a lower response to alcohol than those with no family history of alcohol dependence (Schuckit et al., 2005). Newlin and Thomson (1990) reviewed numerous studies and concluded that many of the measures of response to alcohol could be attributed to more rapid development of acute functional tolerance in the SOA subjects. This result suggests that AFT makes an important contribution to measures of the level of response to alcohol as reported by Schuckit et al. (2005) and may be a predictor for development of alcohol dependence.
AFT to the locomotor incoordinating effect of alcohol in mice can be measured using a stationary dowel test. The genetic influence on AFT in this particular test is demonstrated by studies of inbred and recombinant inbred (RI) mouse strains (Kirstein et al., 2002), as well as by successful selective breeding of low AFT (LAFT) and high AFT (HAFT) mouse lines (Erwin and Deitrich, 1996). Heritability for AFT has been estimated to range from 0.25 to 0.39 (Kirstein et al., 2002; Bennett et al., 2007).
The genomic locations that contribute to AFT have been mapped using quantitative trait locus (QTL) analysis in 30 BXD RI strains and the progenitor strains (Kirstein et al., 2002). However, specific genes contributing to AFT are still unknown. It is posited that if the expression level of a gene affects a phenotype, the expression of the gene should be regulated by a genetic element located within the phenotypic QTL for the phenotype, such as AFT. Therefore, we have also mapped expression QTLs (eQTLs) using gene expression data generated by microarray experiments in 30 BXD RI mouse strains and the progenitor strains (Saba et al., 2006). We and others have previously used eQTLs to help identify candidate genes for complex phenotypes (Hubner et al., 2005; Wang et al., 2007). In the present study, we use this procedure as a starting point to identify candidate genes that contribute to AFT.
|
Materials and Methods |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
AFT Measurement. AFT to the locomotor incoordinating effect of alcohol was measured using the dowel test (Erwin and Deitrich, 1996). In brief, mice were trained to balance on a wooden dowel rod for 5 min. On the test day, a mouse was given an i.p. alcohol injection of 1.75 g/kg (10% w/v) and was put on the dowel. When it fell off, a blood sample of 20 µl was taken from the retro-orbital sinus for measurement of blood alcohol concentration (BEC) 0. A second blood sample was taken when the mouse regained the ability to balance on the dowel (BEC1). The mouse then received a second alcohol injection of 2 g/kg and was placed on the dowel. A third blood sample was taken when the mouse again regained the ability to balance on the dowel (BEC2). The difference between BEC1 and BEC2 was defined as the magnitude of AFT (Erwin and Deitrich, 1996). Blood alcohol concentration was measured by gas chromatography (Tabakoff et al., 1976).
Total RNA Extraction. Mice were killed by exposure to CO2 for 10 to 15 s. The whole brain was removed rapidly (<2 min), and total RNA was isolated using the RNeasy Lipid Tissue Midi Kit (QIAGEN, Valencia, CA). RNA samples from each mouse brain were analyzed using a separate Affymetrix MOE430V2 array (Affymetrix, Santa Clara, CA).
Microarray Analysis. The quantitation of mRNA using the Affymetrix arrays was performed according to the protocol provided by Affymetrix and as previously described in our laboratories (Saba et al., 2006). All arrays that were used for analyses passed the quality control procedures available on http://phenogen.uchsc.edu. Six microarrays were eliminated from the BXD panel and 15 microarrays were eliminated from the inbred panel because they did not meet the quality control requirements.
Some concern has been raised recently about the effect of single nucleotide polymorphisms (SNPs) within the 25mer represented by each probe on Affymetrix arrays on the hybridization of particular mRNAs. Walter et al. (2007) identified 13,292 probes within 6590 probesets that had a sequence polymorphism between C57BL/6J and DBA/2J mice. These probes were eliminated from all analyses, and if less than four probes within a probeset contained no SNPs, the entire probeset was eliminated. The Robust Multichip Average method, which includes a log base 2 transformation, was used to normalize the expression values of the perfect match probes and to summarize these values in probesets (Irizarry et al., 2003). Each of the four experiments was normalized separately. For the BXD RI strains, only gene array data from the 28 strains, including the two parental strains, that had acute functional tolerance data available, were included in the normalization. All 20 inbred strains had acute functional tolerance data available, and expression data from all inbred strains were therefore included in the normalization.
meta-Analysis. Correlation coefficients of gene expression values and AFT measurements in mice were calculated for each probeset in each of the four populations of animals. Correlation analysis between gene expression and AFT was performed with SAS version 9.1.3 (SAS Institute, Cary, NC). For the BXD RI panel and the inbred panel of mice, Pearson product-moment correlation coefficients were calculated using strain means for both gene expression values and AFT. For the two lines of HAFT and LAFT mice, a point biserial correlation was calculated for each line. Correlation coefficients were then transformed to their corresponding Z values (Field, 2001). These transformed effect size values (Z values) were combined for each probeset using a weighted average, where weights were based on the variance of the particular experiment. Following the method outlined by Hedges and Vevea (1998) for a fixed effects model, a Z-score and its corresponding p value were calculated from the weighted average of the effect size values and its standard error. Raw p values were adjusted for multiple comparisons using the method for false discovery rates (FDR) outlined by Benjamini and Hochberg (1995).
Expression QTL and AFT Behavioral QTL Analyses. Using gene expression as the quantitative trait of interest, eQTLs were calculated similarly to traditional QTLs. For this particular analysis, eQTLs were calculated for each probeset individually using a weighted marker regression analysis. The analysis is weighted to account for difference in sample size within strain because strain means were used for the analysis (Saba et al., 2006). Strain means were regressed against a set of 943 markers with unique strain distributions that were derived from the set of 3795 markers available at http://www.genenetwork.org/dbdoc/BXDGeno.html. Permutation was used to calculate p values associated with the maximal logarithm of the odds (LOD) score for each probeset, to account for the multiple comparisons across markers. The number of permutations per transcript was increased until the maximal LOD score from the true data was no longer in the top 10 LOD scores from the permutation, or until 1,000,000 permutations were calculated. A bootstrap method (see Saba et al., 2006), using 1000 bootstrap samples, was used to determine 95% confidence limits for location of eQTLs with empirical p values less than 0.10 (see Supplemental Table 1 and Supplemental Fig. 1 for a discussion of this methodology). All eQTL calculations were carried out with the freeware QTL-Reaper, which can be downloaded at http://sourceforge.net/projects/qtlreaper/.
To generate the behavioral QTL (bQTL) data for AFT, we used our previously published AFT data on the 30 BXD RI strains and the two progenitor strains. QTLs were identified using marker regression on strain means and 1774 markers. QTL intervals were identified as a region covered by neighboring markers with a LOD score greater than 1.44 (p < 0.01), with a distal extension of 5 Mb on either side of the region.
Heritability. Narrow sense heritability (h2) of each probeset was calculated within the BXD RI panel and the inbred panel separately using the method outlined by Hegmann and Possidente (1981). This method assumes that the variance between inbred strains is twice the variance that would be seen in a random mating population [h2 = [1/2]VA/([1/2]VA + VE)], where VA is the variance between strain means and VE is the variance within strains.
Creation of Candidate Gene List. A list of candidate genes for AFT was created using multiple filters. The analysis was initially limited to probesets that had an eQTL that overlapped an AFT bQTL. The eQTL had to be significant or suggestive (empirical p value 
0.10), with a 95% confidence interval for location that overlapped the location of one of the AFT bQTL regions, to pass this filter. For those probesets that passed the filter, an FDR was calculated based on the raw p values from the meta-analysis of correlation between expression level and AFT. Only probesets with an FDR less than 0.01 were retained. The third filter was implemented to ensure that the genes represented in the candidate gene list not only showed a strong genetic correlation with AFT, but also had a high heritability (above the overall median level of heritability) of expression intensities. Probesets were eliminated if the mRNA expression heritability in either the BXD RI panel or the inbred strain panel was lower than the median heritability of all probesets. Finally, the sequence alignment for the remaining probesets was examined using the Ensembl Genome Browser program (http://www.ensembl.org), and probesets, the sequence of which did not map appropriately to the transcript sequences they are purported to represent, were eliminated from consideration.
Examination of the Promoter Region of the Candidate Genes. Transcription factor binding sites within the upstream promoter region of AFT candidate genes were studied by using oPOSSUM (http://burgundy.cmmt.ubc.ca/oPOSSUM). Over-representation of transcription factor binding sites in a 2-kb upstream region (before the transcription start site) of the AFT candidate genes versus a precompiled background set of genes was determined using the Z-scores and Fisher's exact test p values generated by oPOSSUM.
Western Blot Analysis. Mouse brains were homogenized in 0.32 M sucrose, 50 mM Tris-HCl, 1 mM EDTA, and a protease inhibitor cocktail [4-(2-aminoethyl) benzenesulfonyl fluoride, pepstatinA, E-64, bestatin, leupeptin, and aprotinin (Sigma-Aldrich, St. Louis, MO)], pH 7.4, followed by centrifugation at 3,000g for 10 min to remove nuclei and debris. The supernatant was centrifuged at 45,000g for 1 h to pellet membranes. Membranes were resuspended in a buffer consisting of 1% SDS, 5% glycerol, 5% β-mercaptoethanol, 1 mM EDTA, 30 mM Tris-HCl, and 0.1% bromphenol blue, pH 6.8. Protein concentration was measured with the BCA assay (Thermo Fisher Scientific, Rockford, IL). After electrophoresis through 4 to 12% Bis-Tris gels (Invitrogen, Carlsbad, CA), proteins were transferred to nitrocellulose membranes. Anti-Kv2.1 antibody (Sigma-Aldrich) was used to probe the membrane at 1:250 dilution followed by goat anti-rabbit horseradish peroxide-coupled secondary antibody (Bio-Rad, Hercules, CA) at 1:5000 dilution. The chemiluminescence reagent was purchased from PerkinElmer Life and Analytical Sciences (Waltham, MA). β-Actin was used as a loading control.
Localization of Expression of the Candidate Genes in the Brain. We used the Allen Brain Atlas (http://www.brain-map.org) to obtain the expression levels of mRNA produced from our candidate genes in various brain areas. The Allen Brain Atlas contains expression levels for more than 20,000 genes in numerous brain regions of C57BL6/J mice. The data displayed by the Allen Brain Atlas were obtained using in situ hybridization as described in Lein et al. (2007). The gene expression in coronal and sagittal sections is available, and 3-dimensional images for gene expression throughout brain areas can be obtained by using the software (Brain Explorer 1.4.1) available on the Allen Brain Atlas website (see Fig. 3). Because the results obtained from the Allen Brain Atlas can be considered semiquantitative, we chose to express the results in Table 3 in a format indicating relative levels of expression rather than absolute values. The symbols that we assigned to the expression values found on the Allen Brain Atlas site are as follows: 0 when the Atlas notes expression levels to be 0 to 15; + when the Atlas notes expression levels to be 15.1 to 30; ++ when the Atlas notes expression levels to be 30.1 to 50; +++ when the Atlas notes expression levels to be 50.1 to 70; and ++++ when the Atlas notes expression levels to be 70.1 to 100. In Fig. 3, the images are color-coded. Red indicates high expression, and green indicates low expression.
|
|
Biochemical Pathway Analysis. PathwayAssist software (Stratagene, La Jolla, CA) was used to examine the biochemical/functional interaction of the AFT candidate genes. The software uses a database (ResNet Allen Brain Atlas; Stratagene) that contains more than 1.25 million events of regulation, interaction, and modification between and among proteins, cellular processes, and small molecules. ResNet is built from information extracted from PubMed abstracts and 47 full-text journals, and it is regularly updated.
|
Results |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
) and in two replicate selectively bred lines of mice (HAFT1 and LAFT1, HAFT2 and LAFT2) (Saba et al., 2006). In the present study, we measured AFT in the 20 inbred strains of mice using the same dowel test as used previously. The experiment on inbred strains was conducted at three time periods over 2 years. To control for systematic, chronologic bias, DBA/2J mice were included during each period as an internal control. Data in each period were standardized, i.e., adjusted for variation across each period, by an adjustment factor that was the average DBA/2J value across all three periods divided by the average DBA/2J value for a specific period. The data are shown in Fig. 1. Analysis of variance showed a significant strain effect on AFT (p < 0.001). The CAST/EiJ strain was the strain with highest AFT, and the 129/P3J strain had the lowest AFT. These strains were 9-fold different in AFT. Strain accounted for 43% of the total variance of AFT in the inbred mouse strains.
|
Reanalysis of AFT bQTLs Using BXD RI Mice. AFT bQTLs were previously determined in our laboratories using 30 BXD strains and two parental strains (Kirstein et al., 2002). We were able to use more markers together with the behavioral AFT data that we published previously to redetermine AFT bQTLs. Marker locations that are used currently are from NCBI Mouse Build 36 (Ensembl). Six putative QTLs were detected that were located on chromosomes 1, 2, 4, 5, and 6 (Table 1). QTLs on chromosomes 4 and 5 were not detected in our prior analysis. One QTL on chromosome 6 (27.6–92 Mb) was identified as two separate QTLs in the previous analysis (Table 1). The marker D14Byu1 used to identify an AFT-QTL on chromosome 14 in our prior work (Kirstein et al., 2002) is currently no longer assigned to that location, and flanking markers are not associated with AFT. Therefore, no QTL for AFT is now reported on chromosome 14.
|
Using the recommendations of assigning the nomenclature of suggestive (LOD >1.9) and significant (LOD >3.3) to QTLs (Lander and Kruglyak, 1995), all of the AFT QTLs that we identified should be considered suggestive. One QTL [chromosome (chr)6, 27.6–92.0 Mb; LOD 3.1] is approaching significance.
AFT Candidate Genes. Our gene expression data obtained on Affymetrix arrays from all three types of animals are publicly available on http://phenogen.uchsc.edu, and an illustration of the distribution of levels of mRNA for Mtch2 and Kcnb1 across the 20 strains of inbred mice used in our studies is contained in Fig. 2. Among 45,021 probesets on the Affymetrix MOE430V2 arrays (64 control probesets and the 16 probesets that had fewer than four probes without a SNP between C57BL/6J and DBA/2J mice were removed before analysis), 5468 probesets were found to have a significant or suggestive eQTL (p 0.1). A total of 1210 probesets had an eQTL overlapping one of the AFT bQTLs. Fifty of these 1210 probesets were significantly correlated with AFT, based on the meta-analysis of the four populations of animals (FDR <0.01). Among these 50 probesets, 13 had above median heritability in both the 20 inbred strains of mice (median = 0.49) and the 32 strains of mice in the BXD panel (median = 0.40). Sequence alignment on Ensembl showed that eight probesets aligned to the genes they were designed to represent, and the remaining five probesets did not. The eight genes were considered to be candidate genes for AFT (Table 2). Some of the eight candidate genes were represented by more than one probeset on the Affymetrix arrays. In each case, one probeset passed all of the filters, whereas the additional probesets were eliminated at particular steps in the filtering procedure. Supplemental Table 2 provides information on the performance of each probeset targeted to each of the candidate gene mRNAs.
|
|
Examination of Brain Expression Patterns for Candidate Genes. The mRNA for two of the eight candidate genes (Hlf and Cutl1) showed a broad distribution in brain (Table 3). This is not surprising, because the products of these genes act as transcription factors. The mRNA for WNK lysine-deficient protein kinase 1 (Wnk1), the serine-threonine kinase, also demonstrated a broad distribution indicative of the general importance of its function. On the other hand, the mRNA products of Epb4.1l2 and Kcnb1 were expressed together but localized at higher levels in cortical and hippocampal areas (Table 3; Fig. 3). Kcnb1 mRNA was also plentiful in the striatum (motor control function?) and thalamus (signal relay function). Mtch2 mRNA (Table 3) was highly expressed in the cerebellum and hippocampus, which are areas related to motor control and learning and memory, respectively.
Promoter Analysis of AFT Candidate Genes. Five transcription factor binding sites are over-represented within the 2000-basepair 5'-upstream region of the eight AFT candidate genes. The transcription factors are Elk1, Arnt-Ahr, Irf1, Creb1, and E2f1 (Z-score >4, Fisher's exact test, p < 0.05) (Table 4).
|
Western Blot Confirmation for Kcnb1 Protein Product. Kcnb1 is a promising candidate gene for AFT based on past (Lewohl et al., 1999) and current studies. KCNB1 protein levels were determined in HAFT/LAFT mice and in two strains of BXD mice. We chose BXD13 and BXD38 mice for protein quantitation because these two strains display significantly different levels of AFT (Kirstein et al., 2002). In our current study, we also noted significantly different Kcnb1 mRNA levels between HAFT1, HAFT2 and LAFT1, LAFT2 mice and between BXD 13 and BXD 38 mice. The results of Western blot analysis for KCNB1 protein are shown in Table 5 and Fig. 4. BXD 13 mice display high AFT levels and low Kcnb1 mRNA levels (9.19 ± 0.06) compared with BXD 38 mice (log 2-transformed intensity value = 9.95 ± 0.05, p = 0.00002). KCNB1 protein levels are also lower in the brain areas of BXD 13 mice (densitometric value = 1.96 ± 0.13) than in BXD 38 mice (densitometric value = 2.52 ± 0.11, p < 0.01). Kcnb1 mRNA is lower in HAFT1 mice (log 2-transformed intensity value = 8.39 ± 0.04) than in LAFT1 mice (9.13 ± 0.13, p = 0.002). KCNB1 protein is also lower in brain of HAFT1 mice (densitometric value = 1.01 ± 0.07) than in LAFT1 mice (densitometric value = 1.22 ± 0.07, p < 0.05).
|
|
Pathway Analysis of the AFT Candidate Genes. Using PathwayAssist software, a relational pathway shown in Fig. 5 was generated. Five of the eight candidate gene products were included in this pathway, and there were a total of 16 gene products in the pathway.
|
|
Discussion |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
), atherosclerosis (Wang et al., 2007), and hypertension (Hubner et al., 2005) from larger panels of differentially expressed genes.
After applying the eQTL/bQTL overlap filter and other filters, we identified eight genes as predisposing elements for AFT. Erythrocyte membrane protein band 4.1-like 2 (EPB41l2, or 4.1G) is a member of the cytoskeletal protein family that also includes 4.1R, 4.1N, and 4.1B. It is present in the postsynaptic density in neurons as well as in microglia in the brain. The 4.1G protein interacts with AMPA receptor glutamate receptor (GluR)1 and GluR4 subunits, and it may support their surface expression (Coleman et al., 2003). It also interacts with D2 and D3 dopamine receptors (Binda et al., 2002), metabotropic GluR subtype 1
(Lu et al., 2004a), and A1 adenosine receptors (Lu et al., 2004b).
The voltage-gated potassium channel, Shab-related subfamily, member 1 (Kcnb1) is found on neuronal soma and proximal dendrites (reviewed in Trimmer and Rhodes, 2004). Kcnb1 contributes significantly to the negative membrane potential, which dampens neuronal excitability. Another potassium channel, the calcium-gated potassium channel, has already been found to be necessary for the development of rapid tolerance to alcohol sedation in Drosophila (Cowmeadow et al., 2005). Kcnb1 mRNA levels obtained using the Affymetrix MOE 430V2 arrays, across the 20 inbred strains of mice, are illustrated in Fig. 2. Our measures of KCNB protein (Fig. 4; Table 5) substantiated the inverse quantitative relationship between Kcnb1 gene products and AFT in the selectively bred HAFT and LAFT mice and in BXD recombinant inbred mice that differed in AFT. It is of interest that the genes related to scaffolding for the GluRs, and related to membrane polarization that controls N-methyl-Daspartate (NMDA) receptor function, are differentially expressed and have a negative correlation with AFT. Expression of Kcnb1 and Epb4.1l2 also are located in the same brain areas (cortex and hippocampus).
Wnk1 is a serine/threonine kinase. Wnk1 phosphorylates synaptotagmin 2 (Syt2), a Ca2+ sensor, and increases the amount of Ca2+ required for Syt2 binding to synaptic vesicles (Lee et al., 2004). Syt2 has been shown to be involved in the release of glutamate and other neurotransmitters (Pang et al., 2006). Wnk1 also activates Erk5 (Xu et al., 2004), and Erk5 in turn activates calcineurin and down-regulates NMDA receptor activity. Thus, Wnk1 can modulate both presynaptic and postsynaptic components of glutamatergic transmission.
Cut-like 1 (Cutl1) generally functions as a transcription repressor, but it may also be a transcription activator (Nepveu, 2001). It is mainly expressed in neurons in the upper layers of the murine brain cortex (Nieto et al., 2004) and may determine the neuronal phenotype of these neurons.
Hepatic leukemia factor (Hlf) protein is a basic-leucine zipper transcription factor. Reports on the function of Hlf protein in brain have linked the expression of Hlf to synaptogenesis in the cortex and thalamus (Hitzler et al., 1999).
Mitochondrial carrier homolog 2 (Mtch2) is a mitochondrial membrane protein. Proteins such as Mtch2 have also been linked to promotion of neurite outgrowth (Lepagnol-Bestel et al., 2008). The high expression of Mtch2 mRNA in the hippocampus and cerebellum may indicate a role for Mtch2 in rearrangement of neuronal architecture in brain areas important in spatial and motor learning and memory. Overall, it is of interest that a number of the candidate genes, including Epb4.1l2, Wnk1, Hlf, Kcnb1, and Mtch2, show high levels of expression in the brain areas implicated in learning and performance of motor skills, particularly the hippocampus, cerebellum, and cerebral cortex (Fig. 3; Table 3).
The interactions noted in the literature among the products of the candidate genes we identified and other neuronal proteins indicate the importance of regulation of NMDA receptors with regard to AFT. Membrane depolarization upon AMPA receptor activation is generally required for NMDA receptors to open. Kcnb1 contributes to a negative membrane potential and therefore increases the difficulty of activating NMDA receptors. The negative correlation between Kcnb1 mRNA levels and AFT would suggest that NMDA receptors may be more active in animals that develop higher levels of AFT. The AMPA receptors undergo rapid trafficking (Shepherd and Huganir, 2007). Normally, Epb4.1l2 supports AMPA receptors at the synapse. However, the mRNA levels of the splice variant of Epb4.1l2 that is negatively correlated with AFT in our studies does not contain a domain that binds actin and spectrin and thus may act as a "dominant negative" modulator of other Epb4.1 molecules. In essence, higher levels of Epb4.1l2 may decrease AMPA receptor surface expression and reduce the propensity for AFT development. Although not themselves differentially expressed, six receptor proteins (i.e., the products of genes Adora1, Drd2, Drd3, Grm1, Gria1, and Gria4) are coupled to the differentially expressed candidate genes through pathway analysis (Fig. 5). The products of these genes, including the subunits of the AMPA glutamate receptor, the metabotropic glutamate receptor 1, the D2 and D3 dopamine receptors, and the adenosine receptor 1, play a major role in excitatory and inhibitory signaling in the brain. On the presynaptic side of neurotransmission, the interaction of Wnk1 and Syt2 may be related to glutamate release (Pang et al., 2006). The candidate gene product Wnk1 phosphorylates Syt2 and may reduce glutamate release, which could explain the negative correlation between Wnk1 mRNA and AFT in our analysis. It is of interest to note that our earlier microarray analysis of brains of HAFT and LAFT mice only, also implicated NMDA receptor systems in AFT, although different candidate genes were identified (Saba et al., 2006).
Among the eight candidate genes, the expression of Epb4.1l2, Wnk1, and Hlf is trans-regulated from within bQTLs; the remaining genes, Kcnb1, Cutl1, Glcci1, Mtch2, and A130022J15Rik, seem to be cis-regulated (the eQTL information is available on http://phenogen.uchsc.edu). In particular, A130022J15 Rik is located within the bQTL on chromosome 6 that has a LOD score of 3.1. Although this gene is currently not well annotated, our results suggest that this gene would be of value to explore in terms of understanding the determinants of AFT. The eQTLs for all three trans-regulated genes are located on chromosome 2. The overlapping genomic area of these eQTLs (chromosome 2: 11.7–25.9 Mb) contains seven transcription factor genes (Bmyc, Nrarp, Sohlh1, Tbpl2, Bmi1, Ptf1a, and Pax8). It would be appropriate to consider that one or more products of these transcription factor genes could influence the expression of the three trans-regulated candidate genes. However, the current lack of definitive data regarding DNA sequences that comprise binding sites for any of the seven transcription factors precludes a test of such a hypothesis. We did note that a Creb1 binding site was present in six of the eight candidate gene promoter regions, including all of the trans-regulated genes. The expression level of Creb1 was negatively correlated with AFT in our studies, but Creb1 was not included in our candidate gene list because it did not pass the eQTL/bQTL overlap filter. Although it is enticing to speculate about the involvement of Creb in neuroadaptive phenomena such as tolerance, the relation of Creb to the expression levels of our candidate genes is tenuous at present.
A critical issue in evaluating our findings is whether they can be extrapolated across species. The human genome and mouse genome share a high level of synteny (Waterston et al., 2002). The syntenic areas of mouse AFT bQTLs on chromosomes 2 (19.8–30.4 Mb) and 6 (0–17.4 Mb) overlap with two human alcohol dependence susceptibility loci (chromosome 7, 82.4–94.9 Mb and chromosome 10, 26–52 Mb) identified in the Collaborative Study on the Genetics of Alcoholism (Agrawal et al., 2008). One of the AFT bQTLs (chromosome 1, 152.5–169.8 Mb) identified in mice is also syntenic with one of the human genomic loci (LOD >2 in 238 sibling pairs; marker at chromosome 1, 163 Mb) for a low alcohol response as measured by body sway and/or Subjective High Assessment Scale (Schuckit et al., 2005). In addition, the mouse AFT QTL on chromosome 6 (27.6–92.0 Mb; LOD = 3.1) is syntenic with a human locus (chromosome 1, 71.9 Mb) that was reported to affect alcohol tolerance, as indicated by the quantity of alcohol that an individual could consume (Kuo et al., 2006). We cannot be certain whether the human orthologs of our candidate genes for AFT play a role in human alcohol dependence susceptibility or the propensity for humans to develop AFT, but the candidate genes we identified in mice may well, with further investigation, help in understanding the molecular basis of acute alcohol tolerance in humans as well as rodents.
|
Footnotes |
|---|
Parts of this work were previously presented as follows: Wei H (2008) Genomic determinants of alcohol effects, Ph.D. thesis, University of Colorado, Denver, CO.
Article, publication date, and citation information can be found at http://jpet.aspetjournals.org.
ABBREVIATIONS: AFT, acute functional tolerance; SOA, sons of alcoholics; RI, recombinant inbred; HAFT, lines of mice selectively bred for high AFT; LAFT, lines of mice selectively bred for low AFT; QTL, quantitative trait locus; eQTL, expression QTL; BEC, blood alcohol concentration; SNP, single nucleotide polymorphism; FDR, false discovery rates; LOD, logarithm of the odds; bQTL, behavioral QTL; chr; chromosome; Wnk1 and Wnk1, WNK lysine-deficient protein kinase 1; GluR, glutamate receptor; NMDA, N-methyl-D-aspartate; Syt2 and Syt2, synaptotagmin 2; Cutl1 and Cutl1, cut-like 1 (Drosophila); Hlf and Hlf, hepatic leukemia factor; Mtch2, mitochondrial carrier homolog 2; Drd2 and Drd2, dopamine receptor D2; Drd3 and Drd3, dopamine receptor D3; Grm1 and Grm1, glutamate receptor, metabotropic 1; Kcnb1 and Kcnb1, voltage-gated potassium channel; Adora1 and Adora1, adenosine A1 receptor; Gria1 and Gria1, glutamate receptor, ionotropic, AMPA 1; Gria4 and Gria4, glutamate receptor, ionotropic, AMPA 4; Epb4.1l2 and Epb4.1l2, erythrocyte protein band 4.1-like 2; G protein, guanine nucleotide-binding protein.
The online version of this article (available at http://jpet.aspetjournals.org) contains supplemental material. 
Address correspondence to: Dr. Boris Tabakoff, Department of Pharmacology, University of Colorado School of Medicine, Mail Stop 8303, P.O. Box 6511, Aurora, CO 80045-0511. E-mail: boris.tabakoff{at}ucdenver.edu
|
References |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
Agrawal A, Hinrichs AL, Dunn G, Bertelsen S, Dick DM, Saccone SF, Saccone NL, Grucza RA, Wang JC, Cloninger CR, et al. (2008) Linkage scan for quantitative traits identifies new regions of interest for substance dependence in the Collaborative Study on the Genetics of Alcoholism (COGA) sample. Drug Alcohol Depend 93: 12-20.[CrossRef][Medline]
American Psychiatric Association (1994) Diagnostic and Statistical Manual of Mental Disorders. American Psychiatric Association, Washington DC.
Benjamini Y and Hochberg Y (1995) Controlling the false discovery rate: a practical and powerful approach to multiple testing. J R Statist Soc B 57: 289-300.
Bennett B, Downing C, Carosone-Link P, Ponicsan H, Ruf C, and Johnson TE (2007) Quantitative trait locus mapping for acute functional tolerance to ethanol in the L x S recombinant inbred panel. Alcohol Clin Exp Res 31: 200-208.[CrossRef][Medline]
Binda AV, Kabbani N, Lin R, and Levenson R (2002) D2 and D3 dopamine receptor cell surface localization mediated by interaction with protein 4.1N. Mol Pharmacol 62: 507-513.
Coleman SK, Cai C, Mottershead DG, Haapalahti JP, and Keinanen K (2003) Surface expression of GluR-D AMPA receptor is dependent on an interaction between its C-terminal domain and a 4.1 protein. J Neurosci 23: 798-806.
Cotton NS (1979) The familial incidence of alcoholism: a review. J Stud Alcohol 40: 89-116.[Medline]
Cowmeadow RB, Krishnan HR, and Atkinson NS (2005) The slowpoke gene is necessary for rapid ethanol tolerance in Drosophila. Alcohol Clin Exp Res 29: 1777-1786.[CrossRef][Medline]
Erwin VG and Deitrich RA (1996) Genetic selection and characterization of mouse lines for acute functional tolerance to ethanol. J Pharmacol Exp Ther 279: 1310-1317.
Field AP (2001) Meta-analysis of correlation coefficients: a Monte Carlo comparison of fixed- and random-effects methods. Psychol Methods 6: 161-180.[CrossRef][Medline]
Hedges LV and Vevea JL (1998) Fixed- and random-effects models in meta-analysis. Psychol Rev 3: 486-504.
Hegmann JP and Possidente B (1981) Estimating genetic correlations from inbred strains. Behav Genet 11: 103-114.[CrossRef][Medline]
Hitzler JK, Soares HD, Drolet DW, Inaba T, O'Connel S, Rosenfeld MG, Morgan JI, and Look AT (1999) Expression patterns of the hepatic leukemia factor gene in the nervous system of developing and adult mice. Brain Res 820: 1-11.[CrossRef][Medline]
Hubner N, Wallace CA, Zimdahl H, Petretto E, Schulz H, Maciver F, Mueller M, Hummel O, Monti J, Zidek V, et al. (2005) Integrated transcriptional profiling and linkage analysis for identification of genes underlying disease. Nat Genet 37: 243-253.[CrossRef][Medline]
Irizarry RA, Bolstad BM, Collin F, Cope LM, Hobbs B, and Speed TP (2003) Summaries of Affymetrix GeneChip probe level data. Nucleic Acids Res 31: e15.
Khanna JM, Morato GS, and Kalant H (2002) Effect of NMDA antagonists, an NMDA agonist, and serotonin depletion on acute tolerance to ethanol. Pharmacol Biochem Behav 72: 291-298.[CrossRef][Medline]
Kirstein SL, Davidson KL, Ehringer MA, Sikela JM, Erwin VG, and Tabakoff B (2002) Quantitative trait loci affecting initial sensitivity and acute functional tolerance to ethanol-induced ataxia and brain cAMP signaling in BXD recombinant inbred mice. J Pharmacol Exp Ther 302: 1238-1245.
Kuo PH, Neale MC, Riley BP, Webb BT, Sullivan PF, Vittum J, Patterson DG, Thiselton DL, van den Oord EJ, Walsh D, et al. (2006) Identification of susceptibility loci for alcohol-related traits in the Irish Affected Sib Pair Study of Alcohol Dependence. Alcohol Clin Exp Res 30: 1807-1816.[CrossRef][Medline]
Lander E and Kruglyak L (1995) Genetic dissection of complex traits: guidelines for interpreting and reporting linkage results. Nat Genet 11: 241-247.[CrossRef][Medline]
Lee BH, Min X, Heise CJ, Xu BE, Chen S, Shu H, Luby-Phelps K, Goldsmith EJ, and Cobb MH (2004) WNK1 phosphorylates synaptotagmin 2 and modulates its membrane binding. Mol Cell 15: 741-751.[CrossRef][Medline]
Lein ES, Hawrylycz MJ, Ao N, Ayres M, Bensinger A, Bernard A, Boe AF, Boguski MS, Brockway KS, Byrnes EJ, et al. (2007) Genome-wide atlas of gene expression in the adult mouse brain. Nature 445: 168-176,[CrossRef][Medline]
Lepagnol-Bestel AM, Maussion G, Boda B, Cardona A, Iwayama Y, Delezoide AL, Moalic JM, Muller D, Dean B, Yoshikawa T, et al. (2008) SLC25A12 expression is associated with neurite outgrowth and is upregulated in the prefrontal cortex of autistic subjects. Mol Psychiatry 13: 385-397.[CrossRef][Medline]
Lewohl JM, Wilson WR, Mayfield RD, Brozowski SJ, Morrisett RA, and Harris RA (1999) G-protein-coupled inwardly rectifying potassium channels are targets of alcohol action. Nat Neurosci 2: 1084-1090.[CrossRef][Medline]
Lu D, Yan H, Othman T, and Rivkees SA (2004a) Cytoskeletal protein 4.1G is a binding partner of the metabotropic glutamate receptor subtype 1 alpha. J Neurosci Res 78: 49-55.[CrossRef][Medline]
Lu D, Yan H, Othman T, Turner CP, Woolf T, and Rivkees SA (2004b) Cytoskeletal protein 4.1G binds to the third intracellular loop of the A1 adenosine receptor and inhibits receptor action. Biochem J 377: 51-59.[CrossRef][Medline]
Mellanby E (1919) Alcohol: its absorption into and disappearance from the blood under different conditions. Medical Research Committee Special Report 31: 1-48.
Nepveu A (2001) Role of the multifunctional CDP/Cut/Cux homeodomain transcription factor in regulating differentiation, cell growth and development. Gene 270: 1-15.[CrossRef][Medline]
Newlin DB and Thomson JB (1990) Alcohol challenge with sons of alcoholics: a critical review and analysis. Psychol Bull 108: 383-402.[CrossRef][Medline]
Nieto M, Monuki ES, Tang H, Imitola J, Haubst N, Khoury SJ, Cunningham J, Gotz M, and Walsh CA (2004) Expression of Cux-1 and Cux-2 in the subventricular zone and upper layers II-IV of the cerebral cortex. J Comp Neurol 479: 168-180.[CrossRef][Medline]
Pang ZP, Sun J, Rizo J, Maximov A, and Sudhof TC (2006) Genetic analysis of synaptotagmin 2 in spontaneous and Ca2+-triggered neurotransmitter release. EMBO J 25: 2039-2050.[CrossRef][Medline]
Saba L, Bhave SV, Grahame N, Bice P, Lapadat R, Belknap J, Hoffman PL, and Tabakoff B (2006) Candidate genes and their regulatory elements: alcohol preference and tolerance. Mamm Genome 17: 669-688.[CrossRef][Medline]
Schuckit MA, Wilhelmsen K, Smith TL, Feiler HS, Lind P, Lange LA, and Kalmijn J (2005) Autosomal linkage analysis for the level of response to alcohol. Alcohol Clin Exp Res 29: 1976-1982.[CrossRef][Medline]
Shepherd JD and Huganir RL (2007) The cell biology of synaptic plasticity: AMPA receptor trafficking. Annu Rev Cell Dev Biol 23: 613-643.[CrossRef][Medline]
Tabakoff B, Anderson RA, and Ritzmann RF (1976) Brain acetaldehyde after ethanol administration. Biochem Pharmacol 25: 1305-1309.[CrossRef][Medline]
Tabakoff B, Cornell N, and Hoffman PL (1986) Alcohol tolerance. Ann Emerg Med 15: 1005-1012.[CrossRef][Medline]
Trimmer JS and Rhodes KJ (2004) Localization of voltage-gated ion channels in mammalian brain. Annu Rev Physiol 66: 477-519.[CrossRef][Medline]
Walter NA, McWeeney SK, Peters ST, Belknap JK, Hitzemann R, and Buck KJ (2007) SNPs matter: impact on detection of differential expression. Nat Methods 4: 679-680.[CrossRef][Medline]
Wang SS, Schadt EE, Wang H, Wang X, Ingram-Drake L, Shi W, Drake TA, and Lusis AJ (2007) Identification of pathways for atherosclerosis in mice: integration of quantitative trait locus analysis and global gene expression data. Circ Res 101: e11-e30.
Waterston RH, Lindblad-Toh K, Birney E, Rogers J, Abril JF, Agarwal P, Agarwala R, Ainscough R, Alexandersson M, An P, et al. (2002) Initial sequencing and comparative analysis of the mouse genome. Nature 420: 520-562.[CrossRef][Medline]
Xu BE, Stippec S, Lenertz L, Lee BH, Zhang W, Lee YK, and Cobb MH (2004) WNK1 activates ERK5 by an MEKK2/3-dependent mechanism. J Biol Chem 279: 7826-7831.
This article has been cited by other articles:
|
|
 |
|
  M. C. Havrda, B. T. Harris, A. Mantani, N. M. Ward, B. R. Paolella, V. C. Cuzon, H. H. Yeh, and M. A. Israel Id2 Is Required for Specification of Dopaminergic Neurons during Adult Olfactory Neurogenesis J. Neurosci., December 24, 2008; 28(52): 14074 - 14087. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
 |
 |
|
 |
 |
 |
