Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

No animal model exactly duplicates clinically defined alcoholism, but many animal models for specific factors, such as the withdrawal syndrome, have been developed. When EtOH is eliminated, as its depressant effects disappear, characteristic symptoms of hyperexcitability first wax and then wane, including tremor, autonomic nervous system overactivity and, in extreme cases, convulsions, which can be lethal. These withdrawal symptoms define a preexisting state of physical dependence on the drug. Alcohol withdrawal convulsions have been reported to occur in all animal species studied, including humans (Friedman, 1980), and provide a convenient, quantitative index of the severity of withdrawal in mice (Goldstein and Pal, 1971).

In a series of studies in the early 1970s, Dr. Dora Goldstein developed a system for inducing dependence on alcohol in mice by administering EtOH vapor continuously to animals confined in an inhalation chamber. She also described and quantified the characteristic HIC displayed by withdrawing mice when they are picked up by the tail (Goldstein and Pal, 1971). HIC severity after inhalation is dose and duration dependent (Goldstein, 1972), and there is a clear genetic contribution to individual differences in this trait (Goldstein, 1973). We subsequently showed that there were substantial genetically determined differences among inbred mouse strains that were independent of EtOH dose (Crabbe et al., 1983a) (i.e., the differences were pharmacodynamic rather than pharmacokinetic; Kalant et al., 1971).

Knowing that there are genetically influenced differences in EtOH withdrawal severity does not help us to identify which genes are involved in determining those differences. Recently, however, >10,000 genetic polymorphisms in microsatellite DNA segments have been identified using the polymerase chain reaction, and each such genetic marker has been mapped to a specific location on mouse chromosomes. By ascertaining associations, based on genetic linkage, between alleles at particular markers and withdrawal severity, we can now identify the location of particular genes in mice that influence EtOH dependence. Because EtOH withdrawal is a continuously distributed trait in populations (rather than all-or-none, which would imply single-gene influence), such regions are called QTLs, and each QTL implies the presence of a gene or genes nearby that affects the trait under investigation.

The current experiments used a standard method of QTL mapping to identify 11 provisional QTLs in mice originally derived from the C57BL/6J and DBA/2J inbred strains. Given the limited number of genotypes available in this set of strains (25), the power to detect associations is necessarily limited as well, so further verification testing in subsequent studies will be required to confirm these provisional linkages.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Animals and husbandry. Adult male mice (n = 446, 3-58/strain for the EtOH-exposed groups, n = 268 mice, 3-24/strain for the control groups: table 1) from inbred strains C57BL/6J (B6), DBA/2J (D2) and 25 recombinant inbred strains derived by inbreeding from the F₂ cross of B6 X D2 (BXD RI strains) were bred in the Veterinary Medical Unit at the Portland VA Medical Center from breeding pairs initially purchased from The Jackson Laboratory (Bar Harbor, ME). Mice were housed two to four per cage with mice of the same strain. Food and water were available ad libitum, and lights were on from 6:00 a.m. to 6:00 p.m. in colony rooms maintained at 22.0 ± 1.0°C. All procedures were approved by the VA Institutional Animal Care and Use Committee in accordance with United States Department of Agriculture and United States Public Health Service guidelines.

EtOH dependence induction. Details of the basic inhalation exposure method have been published (Terdal and Crabbe, 1994). Mice were tested in one of 18 groups, each representing several RI strains, based on availability, over a period spanning >2 years. Mice from the B6 and D2 strains were included in almost all of the 18 groups. For each strain × treatment condition, mice from multiple litters from each RI strain were used. Mice in the EtOH groups were initially injected intraperitoneally with EtOH (1.5 g/kg) to raise BEC to ~1.50 mg of EtOH/ml of blood. An alcohol dehydrogenase inhibitor, pyrazole HCl (1 mmol/kg i.p.), was given each day to inhibit EtOH metabolism and stabilize BECs. EtOH physical dependence was induced by 72-hr exposure to EtOH vapor by inhalation. Because strains differ in EtOH elimination rates (Crabbe JC, unpublished findings), different levels of EtOH in vapor (3.0, 4.5, 6.0, 7.5, 9.0 or 10.5 mg of EtOH/liter of air) were selected by trial and error on a strain-specific basis (table 1) to achieve approximately equal blood EtOH levels in all strains. For example, the relatively slow-metabolizing D2 strain was exposed to vapor concentrations from 3 to 7.5 mg/liter, whereas the more rapidly metabolizing B6 strain was exposed to 6 to 9 mg/liter. Mice in the control groups were injected with saline on day 1, daily with pyrazole, and were placed in identical inhalation chambers where they were exposed to air alone.

Blood EtOH assessment. After 24- and 48-hr exposure, two or three mice from each of the B6 and D2 strain × vapor concentration groups tested during that week were removed from the chamber for blood sampling. These data served as an additional check on the efficacy of the inhalation procedures that week. At the time of withdrawal (72 hr), all mice were gently restrained, and a 20-µl sample was drawn from the end of the nicked tail with a capillary tube for BEC determinations using a previously published gas chromatographic assay (Terdal and Crabbe, 1994). Control mice had their tails nicked, but samples were not collected.

Withdrawal testing. After 72-hr exposure to EtOH vapor or air (24 hr after the last pyrazole injection), mice were removed from the inhalation chambers, a blood sample was drawn and all mice were scored for withdrawal HIC severity each hour for 10 hr and at hr 24 and 25. Handling-induced convulsions range from no response (score = 0) to severe tonic-clonic convulsions (scores of 4 or 5) that may continue after the mouse is released (scores of 6 or 7). The complete scale and its scoring have been published (Crabbe et al., 1991). As a general index of overall convulsion severity, the area under the entire 25-hr HIC curve was computed (AREA 25).

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Indexing withdrawal severity. Inbred strains of mice have previously been found to differ significantly in HIC scores after EtOH inhalation and after pyrazole injections with or without air inhalation (Crabbe et al., 1983a). To illustrate the withdrawal response, figure 1 shows the HIC scores during the withdrawal test for a randomly selected group of the EtOH-treated B6 and D2 mice and for the control groups from these strains. Preliminary analyses of the BXD RI data showed that mean strain AREA 25 score in the EtOH-treated groups was modestly, but significantly, correlated with the control group AREA 25 (r = .47, P < .05). To create an index of EtOH withdrawal severity that was independent of control HIC scores, the difference between the EtOH group and control group AREA 25 scores was calculated by subtracting the mean value for each strain's control group from each individual EtOH-treated animal's AREA 25 for that strain. These values, termed DELAREA 25, were genetically correlated (r = .80, P < .0001) with the EtOH group AREA 25 but were not significantly correlated with control group AREA 25 (r = .15, P > .10). We examined two other corrections (residual from linear and nonlinear regression of EtOH AREA 25 on control AREA 25) and found these strain mean values to correlate strongly with DELAREA 25 (r  .91, P < .0001). The individual scores on which the DELAREA 25 summary index was based are illustrated for the B6 and D2 strains in figure 2.

View larger version (19K): [in this window] [in a new window]   Fig. 1.   Mean ± S.E. hourly HIC scores for groups of 14 B6 and D2 mice starting immediately after treatment with EtOH vapor as described for 72 hr. These mice were randomly selected from the 52 and 28 mice indicated in table 1. Also shown are the HIC scores of 15 B6 and 11 D2 control mice randomly selected from the 15 and 24 in table 1. These mice received daily pyrazole but were exposed only to air.

View larger version (14K): [in this window] [in a new window]   Fig. 2.   Mean ± S.E. hourly HIC scores used to compute DELAREA 25 (see text). For each animal in the EtOH groups in figure 1, the mean of its control group at that hour was subtracted from its individual HIC score.

Dose-effect relationships. Results for the B6 and D2 inbred strains that served as the source of the genetic variability represented in the BXD RI strains are shown in figures 3 and 4. Both B6 and D2 mice showed increasing withdrawal severity as the vapor concentration of EtOH was increased (fig. 3). These strains could be compared directly at the 6.0 and 7.5 mg of EtOH/liter air treatment levels, and D2 mice showed 3- to 4-fold greater DELAREA 25 scores. The greater sensitivity to EtOH withdrawal of D2 vs. B6 mice has been long established (Goldstein and Kakihana, 1974; Griffiths and Littleton, 1977; Crabbe et al., 1983a). However, some of the difference in withdrawal severity in the current data could be attributed to pharmacokinetic factors. For example, B6 mice exposed to 6 mg of EtOH/liter air had average BEC values on withdrawal of .57 mg/ml compared with D2 mice, which averaged 1.53 mg/ml. The 25 RI strains tested showed similar, but strain-specific, curves (data not shown) relating EtOH vapor concentration to achieved dose (BEC at withdrawal; see table 1).

View larger version (17K): [in this window] [in a new window]   Fig. 3.   Mean ± S.E. EtOH withdrawal severity (DELAREA 25) for B6 and D2 mice as a function of the vapor concentration they inhaled for 72 hr. DELAREA 25 is corrected for the control group mean AREA 25 scores, as described. Data for B6 mice exposed to 10.5 mg of EtOH/liter air are also shown but are excluded from table 1.

View larger version (18K): [in this window] [in a new window]   Fig. 4.   Mean ± S.E. EtOH withdrawal severity (DELAREA 25) for 52 B6 and 28 D2 mice as a function of the BEC assessed at withdrawal. Each symbol represents a vapor concentration, and the S.E. of the BEC is also given. DELAREA 25 is corrected for the control group mean AREA 25 scores, as described. Data for B6 mice exposed to 10.5 mg of EtOH/liter air are also shown but are excluded from table 1.

View larger version (26K): [in this window] [in a new window]   Fig. 5.   Mean ± S.E. EtOH withdrawal severity (DELAREA 25) for mice from five BXD RI strains are shown as a function of the BEC assessed at withdrawal. Each symbol represents a vapor concentration, and the S.E. of the BEC is also given. For some strains, vapor concentrations in addition to those given in table 1 are included. DELAREA 25 is corrected for the control group mean AREA 25 scores, as described.

View larger version (23K): [in this window] [in a new window]   Fig. 6.   Mean ± S.E. EtOH withdrawal severity (DELAREA 25) for mice from five BXD RI strains are shown as a function of the BEC assessed at withdrawal. Each symbol represents a vapor concentration, and the S.E. of the BEC is also given. For some strains, vapor concentrations in addition to those given in table 1 are included. DELAREA 25 is corrected for the control group mean AREA 25 scores, as described.

Indexing strain-specific withdrawal severity. The next issue, therefore, was to develop a way of equating the strains for dose of ethanol so a pharmacodynamic index of withdrawal severity could be devised that was independent of dose. Scrutiny of the dose-effect curves given as examples in figures 5 and 6 correctly suggests that attempting to develop an ED_x from linear regression would be highly inaccurate and ultimately fruitless for some strains. To develop a basis for comparing strains, a subset of mice was selected from each strain such that their average BEC values were near 1.50 mg/ml. All mice exposed to the highest dose, 10.5 mg of EtOH/liter air were first excluded from these analyses. By selecting between 3 and 58 mice/strain from the remaining concentrations, the range of strain mean BEC values was constrained to 1.34-1.59 (F_26,419 = .04, P > .10). It is these mice whose data are summarized in table 1.

View larger version (27K): [in this window] [in a new window]   Fig. 7.   Mean ± S.E. AREA 25 for the mice indicated in table 1. Open bars, control groups. Dark bars, EtOH-treated groups. See text for methods and analysis. Numbers under bar refer to BXD strain number.

View larger version (17K): [in this window] [in a new window]   Fig. 8.   Mean ± S.E. DELAREA 25. These numbers represent the data in dark bars in figure 7, corrected for control data (see fig. 7), and represent the strain values used for the QTL analyses.

QTL analyses: EtOH withdrawal. Mean DELAREA 25 values for each strain were compared with 1533 genetic markers in our database. These markers and their chromosomal positions have been published (Silver and Nadeau, 1997). In addition, control AREA 25 values were analyzed as an index of central nervous system excitability independent of EtOH withdrawal state. The methods for determining association of withdrawal with particular chromosomal regions has been discussed in detail elsewhere (e.g., Buck et al., 1997; Grisel et al., 1997; Crabbe and Belknap, 1992). Briefly, allelic status at each marker in the database is coded as 1 if the RI strain has been shown to possess two copies of the D2 allele and as 0 if it possesses two copies of the B6 allele. There are no heterozygotes because these strains are all inbred. The point-biserial correlation of each strain's DELAREA 25 with its marker values yields a set of correlations. When a QTL influences withdrawal, a significant association is seen with nearby markers. A set of linked markers that were consistently correlated with withdrawal was taken as provisional evidence of an association with a QTL, implying the presence of a nearby linked gene.

View this table: [in this window] [in a new window]   TABLE 2 Provisionally mapped QTLs for EtOH withdrawal Name, position (chromosome number and centiMorgans distal to the telomere), correlation of RI strain means with marker values and significance are given. Negative correlations indicate that the B6 allele is associated with more severe withdrawal. Candidate genes (cM position) are given in the last column. Marker and gene positions were from Silver and Nadeau, 1997.

View larger version (20K): [in this window] [in a new window]   Fig. 9.   Schematic representation of the 12 mouse chromosomes bearing significant associations (P < .01) for DELAREA 25 (given as boxes containing "ET") or control AREA 25 (given as boxes containing "CON"). Boxes are placed in the approximate location of the most highly significant association. To the left of each chromosome, candidate genes are positioned. Scale refers to the distance in centiMorgans from the telomere of the chromosome, indicated with a dark dot at the top. Gene symbols are defined in tables 2 and 3, which give the markers supporting this map, their location, and the value and significance of the association. a, Chromosomes 1, 2, 3, 4, 9 and 10. b. Chromosomes 11, 12, 13, 15, 16 and 18.

QTL analyses: Control HIC scores. Concurrent analyses of the strain mean values for control group AREA 25 scores revealed the presence of 8 QTLs: on chromosomes 2, 4, 10, 11, 15, 16 (2 QTLs) and 19 (see table 3 and fig. 9). Some of these also were in regions near potential candidate genes. Probably because EtOH withdrawal severity was defined in a way that corrected for strain differences in control HIC scores, resulting in a nonsignificant genetic correlation between strain means for these two variables, there were no cases where the control and EtOH withdrawal QTLs were found in the same chromosomal regions.

View this table: [in this window] [in a new window]   TABLE 3 Provisionally mapped QTLs for control HICs See caption to table 2. The largest correlation for the QTL on distal chromosome 16 was with the marker D16Rik32 (r = .59, P = .01). Because this marker has only been positioned by synteny, the association with App is shown.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

From the analysis of variance result of DELAREA 25, it is possible to estimate that the narrow-sense heritability of this trait is h² = .55. This suggests that about half the total variability among individual animals is attributable to genetic influences. This estimate is reasonably consistent with the value derived from realized response to selection in lines of mice selectively bred for severe (WSP) or mild (WSR) chronic EtOH withdrawal (h² = .28; Crabbe et al., 1985). Heritability estimates from RI strains will almost always be higher than those from genetically segregating populations due to the absence of heterozygotes, whose intermediate scores contribute relatively little to the heritability (Belknap et al., 1996).

The withdrawal QTL mapped to distal chromosome 1 is in the same region as one previously definitively mapped (LOD = 5.6; Buck et al., 1997) for acute alcohol withdrawal based on experiments that were also performed in mice derived from the B6 and D2 genotypes. In the acute withdrawal experiments, a single high dose of EtOH was administered, and the withdrawal HIC followed for several hours. Evidence derived from the WSP and WSR lines also suggests that there is substantial genetic codetermination of acute and chronic EtOH withdrawal. WSP mice show exacerbated acute EtOH withdrawal as compared to WSR mice (Kosobud and Crabbe, 1986).

The QTL region on distal chromosome 1 was also recently mapped in a study of acute pentobarbital withdrawal using BXD RI strains and other B6, D2-derived populations (LOD = 6.5; Crabbe JC, unpublished data). In addition, using standard inbred strains of mice, there was a significant genetic correlation between sensitivity to acute alcohol and pentobarbital withdrawal (Metten and Crabbe, 1994). WSP and WSR mice also differ in the severity of both acute pentobarbital and chronic phenobarbital withdrawal (Belknap et al., 1988; Crabbe et al., 1991). Thus, these results strongly suggest that a gene or genes located distally on chromosome 1 modulates genetic predisposition to both acute and chronic alcohol withdrawal severity and, more generally, predisposes mice to withdrawal from drugs that depress the central nervous system. Common genetic determination of responses to central nervous system depressant drugs of abuse has been discussed elsewhere (Crabbe et al., 1994).

Of the three QTLs mapped for acute ethanol withdrawal by Buck et al. (1997), one was detected in the current analysis of chronic withdrawal in the RI strains. This reflects the fact that there is not complete identity of genes determining these two aspects of EtOH withdrawal. Plausible candidate genes for the distal QTL on chromosome 1 include Atp1a2 and Atp1b1 (see fig. 9). These genes code for Na⁺/K⁺-ATPase alpha-2 and beta-1 subunits, respectively. Given the strong evidence implicating GABA_A receptor function in modulating EtOH and pentobarbital effects, and the highly similar pharmacology of these drugs, it is also possible that the chromosome 1 gene affects GABA function, directly or indirectly.

More than 7 years before the polymerase chain reaction, Dr. Benjamin Taylor of the Jackson Laboratories generously provided us with BXD RI strains from his personal colonies, allowing us to perform a primitive experiment similar to the one reported here (Crabbe et al., 1983). A total of 82 mice representing 16 of the RI strains plus B6 and D2 were first tested for either acceptance of EtOH in their drinking fluid, ambulatory ataxia after EtOH or activity in an open field after EtOH. Four weeks later, they were rendered physically dependent using essentially the inhalation method described here. Strains were classified in withdrawal severity as "B6-like" (score = 0) or "D2-like" (score = 1). That is, the assumption of a single-gene effect was made, and point-biserial correlations were calculated between withdrawal severity and allelic status at all markers that had been typed in the BXD RI strains. There were only 90 markers, mostly coat color, tissue histocompatibility and enzyme activity genes. Results indicated that a marker, Car2, on chromosome 3, was possibly linked with withdrawal, but three of the 16 RI strains had the "wrong" allele. This experiment did not control for, or even measure, BEC during inhalation and did not control for strain differences in control levels of HIC. It is interesting that the current study detected a provisional linkage with Car2 (P < .05) and several other markers on chromosome 3. The gene coding for corticotropin-releasing hormone, Crh, maps near this QTL.

The QTL in midchromosome 10 maps near a pair of Shaw-related potassium channel genes (Kcn2, Kcn3), whose protein products function to reestablish resting membrane potentials after action potentials. A significant linkage in midchromosome 9 is in a region containing Htr1b (serotonin 1B receptor subtype) and El1 (a mouse epilepsy gene; Rise et al., 1991), as well as QTLs previously provisionally mapped for EtOH-conditioned taste aversion (Risinger and Cunningham, in press) and EtOH preference drinking (Phillips et al., 1994; Belknap et al., 1997). The Htr1b locus seems an unlikely candidate because studies with mice engineered for a null mutant of this gene have shown that they have equivalent acute and chronic EtOH withdrawal HIC scores to their congenic wild-type background strain (Crabbe et al., 1996). More distally on chromosome 9, a second QTL maps near Scn5a, a type V Na⁺ channel gene. The QTL on chromosome 12 lies near Pomc1, which codes for proopiomelanocortin, the prohormone for ACTH and a number of endorphins.

The QTL on chromosome 13 maps at the gene Srd5a1, coding for steroid 5- reductase (Jenkins et al., 1991). Steroid 5- reductase is required for the reduction of progesterone to 5-dihydroprogesterone, which is further metabolized to 3,5-P, one of a group of cholesterol metabolites with neuroactivity at the -aminobutyric acid_A receptor complex (Lambert et al., 1995). Steroid 5-reductase is present in brain, and the 5-reduced metabolites of the different steroids have been suggested to play a role in myelination (Celotti et al., 1992). Many neuroactive steroids have anticonvulsant activity (Paul and Purdy, 1992). Naive WSP mice were found to be more sensitive than WSR to the anticonvulsant effects of exogenously-administered 3,5-P, and during withdrawal from chronic EtOH inhalation, WSP mice were more sensitive to this effect than control-treated WSPs (Finn et al., 1995). Several studies have reported effects of chronic EtOH on neurosteroid systems (see Finn and Crabbe, 1997, for review). There was no hint of linkage in the region of chromosome 3 where the family of genes encoding 3-hydroxysteroid dehydrogenases, the other main synthetic enzyme in the steroid pathway, is mapped (Abbaszade et al., 1995). This chromosome 13 QTL also maps near the genes coding for the dopamine D₁ receptor and the dopamine transporter, as well as for corticotropin-releasing hormone-binding protein.

The chromosome 15 QTL mapped near three potential candidate genes: Prlr, coding for the prolactin receptor; Trhr, coding for the thyrotropin-releasing hormone receptor; and the gene stg. The stargazer recessive mutant displays frequent, prolonged generalized spike-wave cortical discharges (Noebels et al., 1990), and hippocampal mossy fiber sprouting does not occur (Qiao and Noebels, 1993). Interestingly, WSP mice show a large reduction in dorsal hippocampal mossy fiber zinc content with respect to WSR mice (Feller et al., 1990). Finally, the chromosome 18 QTL lies near ax, an ataxia mutation producing symptoms that include tremor, and the gene coding for synaptotagmin 4.

Candidate genes were also found in the regions mapped for control HIC responses. One marker with high association with control AREA 25 was Comt, the gene coding for catechol O-methyl transferase, located on chromosome 16, and Grin2a, coding for the N-methyl-D-aspartate 2A receptor. Genes for the Clc-2 chloride receptor and somatostatin also map nearby. More distally on chromosome 16, a provisional linkage with the amyloid precursor protein gene is also near genes coding for the serotonin 1F receptor and Grik1, an ionotropic glutamate receptor gene.

Several candidate genes are near the QTL linked to Ntp on chromosome 2, including the gene for glutamic acid decarboxylase 65 and a QTL with suggestive linkage (LOD = 2.3) to acute EtOH withdrawal severity (Buck et al., 1997). A cluster of candidates on chromosome 4 include Shaker-related potassium channel genes, a -carboline-induced seizure gene, Bis1, and the gene for the zinc transporter 2. Associations on chromosome 11 are in the vicinity of a QTL definitively mapped by Buck et al. (1997) for acute EtOH withdrawal severity, and this regions contains several -aminobutyric acid-related and other candidate genes. Finally, a region of chromosome 10 maps a QTL near two potassium channel genes.

These provisional linkages will now be confirmed or rejected using additional genetically segregating populations. In one completed mapping project, Buck et al. (1997) first used the BXD RI strains to identify provisional linkages such as those reported here. We then performed similar association analyses using two genetically segregating populations, where every individual had a unique genotype. These populations were the F₂ cross from B6D2F₁ hybrid mice, and mice genetically selected for increased or decreased acute EtOH withdrawal HIC scores for two generations, starting with the F₂ generation. Combining the probabilities of linkage from these three experiments allowed definitive establishment of linkage for three of the 6 QTLs provisionally identified in the RI strains, with combined LOD scores of 5.6, 5.6 and 4.1. All three exceeded the statistical criteria for significant linkage established by Lander and Kruglyak (1995), LOD = 3.5 in this instance (Belknap et al., 1997). These LOD scores correspond approximately to P < .00005. None of the 10 QTLs mapped in the current study achieve this level of significance, but with only 25 genotypes available for mapping (each RI strain is a genotype), this is not surprising, and results for 3 QTLs were nevertheless very highly associated (P < .001). The criterion suggested for linkage by Lander and Kruglyak (1995) is P < .002, under assumptions that would lead to detecting only one false-positive association in a genome-wide analysis. Both QTLs that showed associations in the RI data at P < .001 were confirmed in the study of Buck et al. (1997) when additional genotypes were tested (LOD = 5.6 and 4.1). Thus, we are confident that some of the QTL linkages reported here will prove to represent the actions of nearby genes.

We currently plan to attempt verification of the 10 withdrawal QTLs provisionally mapped by using other genetic populations. One such population will be to examine the WSP and WSR mouse lines, genetically selected for divergence on the same trait mapped here. If WSP mice are found to have significantly different allelic frequencies for markers in the vicinity of the provisional QTLs than their corresponding WSR mice, this will support the linkage for that QTL. Another method for testing the hypothesis of linkage is to use what has variously been termed "marker-based selection" (Dudek and Tritto, 1995), "genotypic selection" (Plomin and McClearn, 1993; McClearn et al., 1997) or "segregating congenics" (Bennett et al., 1997). This method produces a pair of strains, one in which a small region of the genome containing the QTL under examination is fixed homozygous by mating individual mice homozygous for D2 alleles at markers that flank the QTL region, whereas alleles at other positions of that and other chromosomes are allowed to remain polymorphic. Another strain is developed in parallel for the B6 allele. If comparison of these two strains on the trait being mapped reveals a significant difference, additional evidence for the effect of the QTL is generated. This technique has been used to provide evidence for EtOH-induced loss of righting reflex QTLs (Bennett et al., 1997) and for an EtOH consumption QTL (McClearn et al., 1997).

The usefulness of mapping EtOH withdrawal QTLs in mice is 2-fold. On the one hand, if provisionally mapped QTLs can be verified, and the interval containing them reduced, it will ultimately be possible to identify the relevant genes whose influence they reveal. This should advance our basic understanding of the neurobiological basis for EtOH dependence. The second use for mapped QTLs derives from the fact that the same genes can be found close together in both mouse and humans (although often on different chromosomes) to a substantial degree: mouse and human genomes are estimated to be ~80% linkage homologous (Copeland et al., 1993). Thus, mapping QTLs in mice often leads us to the location of the analogous genes in humans without the need for any human studies. For example, the steroid 5-reductase gene maps to human chromosome 5p15 (Silver and Nadeau, 1997). Current methods for identification of individuals at high risk for alcoholism rely solely on degree of biological relatedness and are not modulated by knowledge about any particular genes. Knowledge about an individual's status with regard to risk markers could, in turn, allow assignment of relative risk with greater precision, thereby facilitating interventions before neurotoxic levels of drinking are reached.