Introduction

Top Abstract Introduction Methods Results Discussion References

Evidence indicates a heritable component to alcoholism (Goodwin et al., 1973; Bohman et al., 1981) and, in human (Schuckit, 1980, 1984; Wilson et al., 1984) and animal studies (McClearn and Kakihana, 1981), individual differences in sensitivity to ethanol are well established. Careful studies have shown that individuals at high risk for alcoholism differ from individuals at low risk in their subjective responses to ethanol and sensitivity to body sway (Schuckit, 1984). Recent findings (Schuckit, 1994) provide strong evidence that initial sensitivity to ethanol is a good predictor of the development of alcoholism. Thus, identification of genes regulating ethanol sensitivity has been the subject of much recent research.

Quantitative genetic studies of ethanol-related behaviors have shown that measures of hypnotic sensitivity, ST or BECRR, and hypothermia are continuously and normally distributed as a result of polygenic determination of the phenotypes (DeFries et al., 1989; Erwin et al., 1990a). Indeed, results obtained with the LSXSS RI strains indicate that hypnotic sensitivity and hypothermia are regulated by a minimum of seven and four genes, respectively (DeFries et al., 1989; Erwin et al., 1990a). These RI strains were derived from Long Sleep (LS/lbg) and Short Sleep (SS/lbg) lines of mice, selectively bred for differences in hypnotic sensitivity to ethanol and found to differ also in ethanol-induced hypothermia (DeFries et al., 1989).

Several studies have used QTL analyses to identify chromosomal regions linked to genes that regulate differences in ethanol-related behaviors. Using the BXD RI derived from the C57BL/6J and DBA/2J inbred strains of mice, investigators (Plomin et al., 1991; Rodriguez et al., 1995) found QTL at the P < .05 level for ethanol consumption and hypnotic sensitivity. Subsequently, other investigators (Buck et al., 1994; Crabbe et al., 1994a; Phillips et al., 1994, 1995) have reported finding provisional, P < .05, QTLs for ethanol-related behaviors including preference, locomotor activation, tolerance and acute withdrawal seizures. Markel et al. (1996) recently reported the use of SSLP markers to identify QTL for ethanol-induced sleep time in LSXSS RI strains of mice. These investigators found several provisional, P < .05, QTL for hypnotic sensitivity and have identified and confirmed significant QTL in a large panel of F₂ mice derived from the ISS and ILS inbred strains (Markel et al., 1996). One of the highly significant QTL was for the marker, D2MIT21, located at 80 cM on chromosome 2 (Markel et al., 1996), a locus linked to the region containing the high-affinity NT receptor gene, Ntsr (Laurent et al., 1994).

Consideration of Ntsr as a candidate gene for regulation of ethanol sensitivity is consistent with a number of pharmacological and behavioral effects of NT. Central administration of NT produces effects similar to those observed after acute ethanol administration (Erwin et al., 1990a) including hypothermia (Nemeroff et al., 1977; Jolicoeur et al., 1981; Hernandez et al., 1984) and locomotor activation (Kalivas et al., 1981; Kalivas et al., 1982) or inhibition (Nemeroff et al., 1977; Clineschmidt et al., 1979; Jolicoeur et al., 1981), depending on the dose and site of administration. Moreover, centrally administered NT markedly enhances hypnotic, locomotor inhibitory and hypothermic effects of ethanol (Luttinger et al., 1981; Erwin et al., 1987; Widdowson, 1987; Erwin and Su, 1989) in both rats and mice. Recent studies have shown a cross-tolerance between ethanol- and neurotensin-induced locomotor inhibition or hypothermia (Erwin et al., 1995).

Neurotensin receptors have been well-characterized from rat and mouse brain (Mazella et al., 1983; Kitabgi et al., 1985; Campbell et al., 1991). These studies revealed biphasic binding isotherms best described by two independent, NTR_H and NTR_L, binding sites. The H₁ histamine antagonist, levocabastine, selectively inhibits binding of NT to NTR_L (Kitabgi et al., 1985; Campbell et al., 1991), providing a valuable tool to distinguish between NTR_H and NTR_L. Levocabastine has been used to demonstrate differences in densities of NTR_H and NTR_L in brain regions from LS and SS mice (Campbell et al., 1991) and from the LSXSS RI strains of mice (Erwin et al., 1993). The results indicated that differences in NT receptor densities and in NT-ir levels were mediated by multiple genes with additive effect (Erwin and Jones, 1993; Erwin et al., 1993) and significant genetic correlations were observed between hypnotic sensitivity to ethanol and FC NTR_H density (Erwin and Jones, 1993). Genetic correlations between NT-ir levels, or NT receptor densities, and ethanol-related behaviors provides a rational basis for searching for common QTL between NT receptor densities or NT-ir levels and ethanol sensitivity.

Identification of common QTL regulating NTR binding capacities, NT-ir levels and ethanol actions would support the hypothesis that genetic differences in ethanol sensitivity are mediated, in part, by differences in these NTergic processes. Thus, the research presented herein had three aims: 1) to identify, in LSXSS RI strains, QTL for BECRR and hypothermia; 2) to determine QTL for NT receptor densities and NT-ir levels and 3) to determine if any QTL for hypnotic or hypothermic effects of ethanol are in common with QTL for NT receptor densities and NT-ir levels in specific brain regions.

	Methods

Top Abstract Introduction Methods Results Discussion References

Animals. Male LSXSS RI strains of mice were obtained from the Institute for Behavioral Genetics, University of Colorado, Boulder, CO. All experiments were conducted with mice (60-80 days of age) which were maintained in a constant temperature (22°C), humidity (20%) and light (12L/12D) environment. Separate groups of mice were used for each phenotype determination.

Hypnotic and hypothermic sensitivity to ethanol. Hypnotic sensitivity to ethanol was measured by determining the duration of loss of righting response (sleep time) and the blood ethanol concentration, in mg ethanol per dl blood (mg/100 ml), at regaining righting response following 4.2 g/kg ethanol. Hypothermia was measured as the difference in rectal temperature immediately before, and at 60 min after administration of ethanol (4.2 g/kg i.p.). The procedures and the mean values for these behavioral phenotypes in each LSXSS RI strain have been published previously (DeFries et al., 1989; Erwin et al., 1990a) and the data from those studies were used to determine genetic correlations and perform QTL analyses described in the present paper.

Neurotensin extraction and radioimmunoassay. Naive mice, housed in groups of two to six per cage, were killed by cervical dislocation and decapitation. Brains were dissected quickly (<1 min) and specific regions punched according to the anatomical guidelines of Slotnick and Leonard (1975). Neurotensin was extracted and assayed as previously described (Erwin and Jones, 1989; Erwin et al., 1990b). Regions were weighed and homogenized in 10 to 20 volumes of 0.01 N HCl, and placed in a boiling water bath for 5 min. Homogenates were centrifuged at 20,000 × g for 20 min, and the resulting supernatant extracts were lyophilized and stored at 70°C for less than 2 wk before radioimmunoassay. Standard double antibody radioimmunoassays (were used to measure NT concentrations (Erwin and Jones, 1989; Erwin et al., 1990b). Details of the procedures and the mean values for NT levels in specific brain regions including hypothalamus, nucleus accumbens and ventral midbrain from LSXSS RI strains of mice, have been published previously (Erwin et al., 1993).

Neurotensin binding assays. Animals were killed and brains dissected as described above; dissected and/or punched regions were rapidly chilled in cold (4°C) 50 mM Tris buffer, pH 7.4, containing 40 mg/ml bacitracin and 1 mM EDTA. Pooled regions from two to four brains were homogenized in ten volumes of buffer; the homogenates were centrifuged at 100,000 × g for 30 min. The resulting membrane pellet was rehomogenized and centrifuged; this wash procedure was repeated twice.

QTL analysis. LSXSS RI strains were genotyped as previously published (Markel et al., 1996) using 120 simple sequence length polymorphism (SSLP) markers (Research Genetics, Huntsville, AL) found to be polymorphic in LS and SS parental strains. These markers covered the mouse genome at an average marker interval of 15 cM. Strain distribution patterns were established for all 20 linkage groups in 24 RI strains (Markel et al., 1996). Because the LS and SS parental lines were not completely inbred, more than two alleles exist for some markers among the RI strains. Ten percent (12 of 120) markers gave three or more alleles and as a result of being unable to know the exact genotype (frequency of alleles) in the outbred LS and SS progenitors of the Ris, we were unable to use the interval mapping method described by Markel et al. (1996) in identifying QTL. Therefore, one-way analysis of variances were carried out with the phenotypic measure as the dependent variable and the RI strains were grouped by allele type. In approximately 6% (8 of 120) of the markers, a given allele was represented only by one RI strain. Because the unique strain may be the result of a new microsatellite mutation, the resulting QTL are suspect; those instances are indicated in the tables. Analyses were performed with SPSS version 6.0 for Windows. Because this is not only an exploratory study to identify provisional QTL for NT measures, but is intended to identify common QTL between independent phenotypes, levels of significance, P < .05, is used. It is recognized that this level of significance, particularly with limited numbers of RI strains (22-26, depending on the phenotype), will produce type I errors (Lander and Schork, 1994). However, the goal of this study is similar to those of Crabbe et al. (1994a) and Phillips et al. (1995), i.e., to identify any potentially important QTL by using an  level of 0.05 because it reduces the type II error rate. Protection against type I error depends on future LS and SS inbred strains (Belknap et al., 1996) and on finding QTL in common with those reported by other investigators using independent panels of RI strains. Candidate genes that are associated with a linkage group ±15 cM (the average map interval for markers) were taken from 1994 Chromosome Committee Report (Committee on the Mouse Genome, 1994).

	Results

Top Abstract Introduction Methods Results Discussion References

Genetic correlations between ethanol sensitivity and NT measures. The genetic correlations between NTR_H and NTR_L and hypnotic sensitivity to ethanol (ST) have been reported previously (Erwin and Jones, 1993). In our study, additional NT receptor binding data were included for some RI strains so that mean values were based on at least four degrees of freedom for those strains. Therefore, in table 1, we have included the correlations between NTR densities and ST as well as between all NT measures and BECRR and hypothermia. Among the LSXSS RI strains, ST and BECRR are negatively correlated measures of hypnotic sensitivity (DeFries et al., 1989) with a high BECRR being related to a short ST, i.e., insensitivity to ethanol. Thus, as observed in table 1, NT measures positively correlated with ST are negatively correlated with BECRR and vice versa. Correlations between NT-ir in HYP or NA and ST were not significant, but showed a consistent trend with r values of 0.38 and 0.24, respectively. Significant positive correlations were observed between ST and NTR_H densities in frontal cortex (FC) and STR, but not in VMB. The results showed 4 correlations of 27 with values of P < .05 and the cumulative Poisson distribution was used to determine the probability of observing 4 correlations at P < 0.05 when only 1.35 (0.05 × 27) are expected by chance. The probability was calculated to be p = 0.048 [P(4 or more) = 1-0.9517]. Thus, it is highly likely that one or more of these genetic correlations indicate common underlying mechanisms regulating hypnotic sensitivity to ethanol and NT measures.

Provisional QTL for ethanol actions. It should be recognized that QTL analyses allow only a rough identification of genetic map locations of genes that exert modest effects on continuously distributed phenotypes. Results presented in table 2 show provisional QTL identified by specific SSLP markers and map distance (cM), for ST, BECRR and hypothermia. QTL for ST in the LSXSS RI strains have been published elsewhere (Markel et al., 1996) and are presented for comparison with other phenotypes and for identification of common QTL. As indicated in table 2, two of the six QTL for ST have been confirmed (Markel et al., submitted) in a large panel of F₂ mice derived from inbred strains of LS and SS. The two confirmed QTL for ST were on chromosomes 1 at 58 cM (D1MIT45 marker) and 2 at 80 cM (D2MIT21 marker). There were three common QTL for ST and BECRR, an expected result, because these phenotypes both measure hypnotic sensitivity at the same endpoint. The QTL for ST on chromosome 2 was observed also for BECRR and there were two other, as yet unconfirmed, QTL in common for ST and BECRR on chromosomes 1 and 18. The five QTL for hypothermia were on chromosomes 4, 6, 11, 14 and 15. 

Provisional QTL for neurotensin receptors. The densities of NTR_H and NTR_L in the FC or STR have been shown to differ up to 1.6- and 1.9-fold, respectively, among the LSXSS RI strains (Erwin et al., 1993). These differences are continuously and normally distributed and show a significant effect of RI strain (genetic effect) on NT receptor density in all brain regions. Estimates indicate a minimum of six and four genes controlling the differences in densities of NTR_H and NTR_L, respectively (Erwin et al., 1993) for STR and FC. In VMB, four and three genes were estimated to account for differences in NTR_H and NTR_L, respectively. Provisional QTL for genes regulating differences in NTR_L and NTR_H densities are shown in tables 3 and 4. For NTR_H, there were 4, 9 and 8 QTL for VMB, STR and FC, respectively. This number of QTL for each brain region is reasonable considering the number of genes estimated to regulate densities of these receptors. There are common QTL for NTR_H in STR and FC on chromosomes 2 at 80 cM and 8 at 8 cM and one QTL in the same linkage group on chromosome 8 for VMB, STR and FC. Several candidate genes are located in the region of the QTL on chromosome 2 including Ntsr, NTR_H; Avp, arginine vasopressin; and Pdyn, prodynorphin genes. Previous studies showed genetic correlations for NTR_H density across brain regions were relatively low, r = 0.24 to 0.37 (Erwin et al., 1993). Similarly, our study shows a limited number of common QTL for NTR_H across brain regions indicating differences in regulation of densities of the receptors in these brain regions.

Provisional QTL for NT-ir levels. Levels of NT-ir in HYP, NA or VMB have been shown to differ by 4.7-, 5.4- and 3.0-fold, respectively, among the LSXSS RI strains (Erwin et al., 1993). The levels of NT-ir were normally and continuously distributed with estimates of five, four and three genes regulating the levels in HYP, NA and VMB, respectively. The QTL for NT-ir levels in HYP, NA and VMB are shown in table 5; there are 7, 7 and 10 QTL for those brain regions, respectively. Previous studies showed that NT-ir levels in various brain regions were only moderately correlated (Erwin et al., 1993) indicating a mostly region-specific regulation of NT-ir levels. Those observations are consistent with results of the present study showing only three of the 24 NT-ir QTL in common among brain regions; two of those three are in common for NT-ir in HYP and VMB. Some of the QTL in table 5 are linked to candidate genes of interest including: mme, neutral metallo endopeptidase on chromosome 3 at 34 cM; Pkcg, protein kinase C, gamma on chromosome 7 at 2 cM and Gria3, glutamate receptor AMPA3, alpha 3 on chromosome X at 5 to 12.5 cM. Neurotensin has been shown to be degraded by metallo endopeptidases (Checler et al., 1988; Kitabgi et al., 1992) and potential catalytic differences would be expected to regulate, in part, differences in levels of the peptide. The chromosomal location of the NT precursor gene for mouse has not been determined; therefore, it remains to be seen whether this gene might be a candidate for an NT-ir QTL. Whether there are polymorphisms in mme, Ntsr or the NT precursor genes in the LSXSS RI strains is unknown, but is the subject of current investigation.

Common QTL for NT measures and ethanol sensitivity. Common QTL for these phenotypes are summarized in table 6. Four QTL for NTR_H density were found in common with confirmed QTL for hypnotic sensitivity on chromosomes 1 (58 cM), 2 (80 cM), 11 (56 cM) and 15 (61 cM) and one was in common with an unconfirmed QTL on chromosome 3 (19 cM). As noted in tables 2 and 6, the QTL for ST on chromosomes 11 and 15 were not observed in the LSXSS RI strains, but, were identified in a study with F₂ mice derived from LS and SS inbred strains (Markel et al., submitted). These QTL for ST are shown here because of they are in common with QTL for NTR_H. Only one common QTL was found for NTR_L and hypnotic sensitivity to ethanol on chromosome 1 at 103 cM. There were no common QTL for NTR_L or NTR_H and ethanol-induced hypothermia. Candidate genes linked to these common QTL are listed in table 6.

	Discussion

Top Abstract Introduction Methods Results Discussion References

Our study involved new determinations of genetic correlations between ethanol-induced hypnotic effects or hypothermia and NT measures. These observations were followed by the use of QTL analyses to identify tentative chromosomal loci that regulate ethanol effects and levels of NTR and NT-ir expression. The second step in these ongoing studies will be to confirm the QTL for NT measures and the common QTL for ethanol effects and NT measures in an independent investigation using a panel of F₂ mice derived from the LS and SS inbred strains.

Hypnotic sensitivity to ethanol was correlated with NTR_H densities in the STR or FC (r = 0.40 or 0.39, P < .05, respectively), but ethanol-induced hypothermia did not correlate with NTR_H density. Significant positive correlations between hypnotic sensitivity to ethanol and NTR_H density in these brain regions are consistent with the findings of motor inhibitory effects of NT when the peptide is administered into the NA area (Erwin and Su, 1989). Similarly, NT, injected into the NA, inhibits the locomotor activating effects of dopaminergic drugs (Kalivas et al., 1984) and, when administered intracerebro ventricularly, NT markedly potentiates hypnotic effects of ethanol (Erwin et al., 1987). Another study has shown a positive association between NT receptor density and ethanol sensitivity (Erwin and Jones, 1993). Chronic ethanol administration produced a decrease in NT receptor density in the NA, which was associated with development of tolerance to locomotor inhibitory effects of ethanol (Erwin et al., 1992). Thus, there are substantial data linking the inhibitory effects of ethanol to NTR_H densities.

Significant correlations were observed between NT-ir levels in the VMB and hypnotic sensitivity to ethanol (r = 0.62, P < .002) (table 1). The negative correlation is consistent with the observations that NT administered into the VMB produces locomotor activation in rats and mice (Kalivas et al., 1982; Erwin and Jones, 1989).

Analyses of associations between molecular markers and ethanol sensitivity, NT receptor densities and NT-ir levels were conducted using 24 LSXSS RI strains. Clearly, there is limited statistical power with only 24 RI strains. Thus, except for those confirmed QTL for ST, the QTL shown in tables 2, 3, 4 must be considered provisional. There were no common QTL for the hypnotic and hypothermic effects of ethanol. These results are consistent with the absence of significant genetic correlations between hypothermia and ST in the LSXSS RI strains (Erwin and Jones, 1993) and with other studies showing different QTL for hypothermia and hypnotic sensitivity (Crabbe et al., 1994, a and b). Comparison of QTL for hypothermia induced by 4.2 g/kg ethanol (table 2) with those obtained by Crabbe et al. (1994a) at 4.0 g/kg indicate one possible common QTL, located on chromosome 11 in the region of 19 cM, for the LSXSS and BXD RI panels. Our results could be considered a confirmation of a QTL for hypothermia in this region. This linkage group contains several candidate genes that might be mechanistically linked to ethanol effects, including the gamma-aminobutyric acid receptor subunits alpha-1 (Gabra1) and gamma 2 (Gabrg2), and the adrenergic receptor alpha-1 (Adra1) genes. One might expect some overlap in QTL for ethanol effects between the LSXSS RI and the BXD RI panels, because the C57BL/6J and DBA/2J mice were two of the eight inbred strains that comprised the genetically heterogeneous stock from which the LS and SS mice were derived. However, because there are undoubtedly allelic differences in the parental stocks used in generating the RI panels, it is to be expected that some QTL would differ between the two sets of RI strains.

Results in tables 3 and 4 show provisional QTL for NT receptor densities and NT-ir levels in specific brain regions. Common QTL for either NT receptor densities or NT-ir levels were not always observed across brain regions, an expected result if these NT processes are regulated by multiple factors that differ by brain region. Alternatively, it is probable that some of the QTL for each brain region are fortuitous, the result of type I and type II errors. However, the finding of common QTL for NTR_H densities in both FC and STR and for ST and BECRR on chromosome 2 (80 cM), a region containing the Ntsr locus, makes this a strong candidate gene. Certainly, the observation suggests pleiotropic effects of a locus linked to the region containing the Ntsr gene. The presence of a QTL for ethanol-induced sleep time on chromosome 2 (80 cM) has been confirmed in a panel of 186 F₂ mice derived from the ILS/lbg and ISS/lbg strains. This QTL accounts for 25% of the genetic variance with a LOD score of 6.1 (Markel et al., submitted). It is important to confirm the QTL for NTR_H on this chromosome. Other common QTL for hypnotic sensitivity to ethanol and NT receptor densities were found on chromosomes 1 (D1MIT45, 58 cM), a QTL also confirmed in ILSXISS F₂ generations, and on chromosome 3 (D3MIT21, 19 cM).

Common QTL for hypnotic sensitivity and NT-ir levels were observed on chromosomes 3 (D3MIT21, 19 cM), 9 (D9MIT12, 55 cM) and 18 (D18MIT7, 50 cM). Candidate genes that might regulate ethanol actions and NT receptor or NT-ir precursor gene expression and that are linked to these common QTL include: Mme, membrane metalloendopeptidase on chromosome 3; Gnai2, guanine nucleotide binding protein, alpha inhibiting-2 and Acra3, acetylcholine receptor alpha-3 neural on chromosome 9 and Adrb2, adrenergic receptor, beta-2 and G-protein coupled receptor 7 on chromosome 18 (Committee on the Mouse Genome, 1994). These candidate genes are consistent with reports that the promoter region of the NT precursor gene contains a cyclic AMP responsive element, a glucocorticoid responsive element, and an AP-1 site (Kislauskis and Dobner, 1990; Dobner et al., 1992). The gene for preproNT/neuromedin N (preproNT) has been cloned from rat, dog, cow and human and in rats, levels of mRNA in specific brain regions, except for frontal cortex, correspond reasonably well, with levels of the peptide (Alexander et al., 1989). This gene has been mapped to human chromosome 12 (Gerhard et al., 1989), but to our knowledge has not been mapped in the mouse genome. Kislauskis and Dobner (1990) have provided evidence for cooperative regulation of preproNT gene expression by transcription factors including c-fos, CREB (cAMP responsive element binding protein), and glucocorticoids. In brain tissues, drugs that elevate levels of NT-ir, e.g., haloperidol, produce increases in c-fos mRNA before increases in preproNT mRNA (Merchant and Miller, 1994). Ethanol administration alters most, if not all of these factors that regulate preproNT expression. For example, it is well known that ethanol causes a rapid increase in plasma glucocorticoids in mice (Zgombick and Erwin, 1987) and in cells and brain tissues ethanol stimulates cAMP production via dopamine-coupled adenylyl cyclase (Rabin and Molinoff, 1981; Rabe et al., 1990). Ethanol increases dopamine overflow in nucleus accumbens and striatum (Lai et al., 1979; Imperato and Dichiara, 1986) and evidence has been reported indicating that chronic ethanol exposure alters c-fos mRNA levels (Dave et al., 1989). To date, little is known regarding regulation of expression of the high-affinity NT receptor gene.

There were no common QTL for ethanol-induced hypothermia and NT receptor densities, but there were two for NT-ir levels and hypothermia: on chromosomes 4 (43 cM) and 6 (41 cM). Potential candidate genes for these common QTL include the c-jun oncogene (Jun) on chromosomes 4 and inositol 1,4,5-trisphosphate receptor (ltpr1) on chromosome 6. A number of investigators have shown that centrally administered NT produces marked hypothermia (Martin et al., 1980; Bissette et al., 1982; Prange and Nemeroff, 1982; Erwin and Jones, 1990); thus, these common QTL suggest a possible mechanistic link between endogenous levels of NT in the HYP or NA and ethanol-induced hypothermia. This hypothesis is supported by data showing a genotype-and dose-dependent effect of ethanol on NT-ir levels as well as on thermoregulation in the LS and SS mice (Erwin et al., 1990b).

Previous studies showed no significant genetic correlation between NT-ir levels and NTR_H or NTR_L in the VMB (Erwin et al., 1993). This finding is consistent with results in tables 3, 4, 5 showing only one common QTL for NT-ir and NTR_H and one in common for NTR_L and NT-ir for VMB. Because the other brain regions used in NT-ir and NTR assays were not the same, only the VMB can be compared for common QTL.

Our results suggest, as might be expected, that for genetically uncorrelated phenotypes there are few, if any, common QTL. On the contrary, a number of common QTL would be expected for correlated phenotypes such as ethanol-induced ST and NTR_H density or NT-ir levels and this is what we have found. Candidate loci for the common QTL represent genes that might directly (transcription factors) or indirectly (neurotransmitter receptors and second messenger processes) regulate gene expression. Common QTL suggest pleiotropy for some loci regulating ethanol-induced hypothermia and NT-ir levels and for genes controlling hypnotic sensitivity to ethanol and NTR_H densities. The findings support the hypothesis that some of the effects of ethanol are mediated by NT and NTR_H and that genetic differences in sensitivity to ethanol are governed, in part, by differences in levels of NTR_H and NT-ir in specific brain regions.