Quantitative Genetic Study of Maximal Electroshock Seizure Threshold in Mice: Evidence for a Major Seizure Susceptibility Locus on Distal Chromosome 1

doi:10.1006/geno.2001.6577, available online at http://www.idealibrary.com on IDEAL

Article

Quantitative Genetic Study of Maximal Electroshock Seizure Threshold in Mice: Evidence for a Major Seizure Susceptibility Locus on Distal Chromosome 1 Thomas N. Ferraro,1,2,* Gregory T. Golden,1,4 George G. Smith,1,4 Ryan L. Longman,1 Robert L. Snyder,1,4 Denis DeMuth,1,4 Ivanna Szpilzak,1 Nicole Mulholland,1 Elaine Eng,1 Falk W. Lohoff,1 Russell J. Buono,1 and Wade H. Berrettini1,3 Departments of 1Psychiatry, 2Pharmacology, and 3Genetics, University of Pennsylvania, Philadelphia, Pennsylvania 19104, USA 4Research Service, Veterans Affairs Medical Center, Coatesville, Pennsylvania 19320, USA *To whom correspondence and reprint requests should be addressed. Fax: 215-573-2041. E-mail: [email protected]

We conducted a quantitative trait locus (QTL) mapping study to dissect the multifactorial nature of maximal electroshock seizure threshold (MEST) in C57BL/6 (B6) and DBA/2 (D2) mice. MEST determination involved a standard paradigm in which 8- to 12-week-old mice received one shock per day with a daily incremental increase in electrical current until a maximal seizure (tonic hindlimb extension) was induced. Mean MEST values in parental strains were separated by over five standard deviation units, with D2 mice showing lower values than B6 mice. The distribution of MEST values in B6 × D2 F2 intercrossed mice spanned the entire phenotypic range defined by parental strains. Statistical mapping yielded significant evidence for QTLs on chromosomes 1, 2, 5, and 15, which together explained over 60% of the phenotypic variance in the model. The chromosome 1 QTL represents a locus of major effect, accounting for about one-third of the genetic variance. Experiments involving a congenic strain (B6.D2-Mtv7a/Ty) enabled more precise mapping of the chromosome 1 QTL and indicate that it lies in the genetic interval between markers D1Mit145 and D1Mit17. These results support the hypothesis that the distal portion of chromosome 1 harbors a gene(s) that has a fundamental role in regulating seizure susceptibility. Key Words: mice, quantitative trait, polygenic inheritance, seizure, electroshock, epilepsy

INTRODUCTION Susceptibility to experimental seizures in common strains of laboratory mice is determined by environmental and genetic factors. Elucidating the nature of genetic influences in these models can provide a focus for translating basic research to clinical studies and lead to greater general understanding of the biological mechanisms that underlie seizure disorders. Genetic studies of seizure susceptibility in mice frequently use response to convulsant drugs to determine phenotype. Thus, drugs such as kainic acid [1], pentylenetetrazol (PTZ) [2], β-carbolines [3], and cocaine [4] have been used to map seizure-related mouse genes. Nonpharmacological models of experimental seizures are a useful and important complement to those involving convulsant drugs in that they circumvent interpretational issues associated with pharmacokinetic variables. Electroconvulsive shock (ECS) is a classic method for

studying experimental seizures, particularly in rodents [5]. A variety of endpoints can be measured in ECS paradigms, including qualitative phenotypes related to seizure pattern as well as quantitative minimal and maximal seizure thresholds [6]. Several applications of this procedure have been described with greatest emphasis placed on its use in the development of anticonvulsant drugs [7–9]. Different laboratories also may introduce some variation, albeit subtle, because the specific instruments, settings, electrode application, and other parameters of the procedure may influence results [6,10,11]. We report here the genetic dissection of maximal electroshock seizure threshold (MEST) in a cross between C57BL/6 (B6) and DBA/2 (D2) mice, two common inbred strains that, because of their inherently divergent seizure phenotypes, have been the focus of numerous studies on the genetic basis of experimental seizures [12–17]. Our results document the location of genes that influence electrical seizure threshold in mice and confirm the existence of a

GENOMICS Vol. 75, Numbers 1–3, July 2001 Copyright © 2001 by Academic Press. All rights of reproduction in any form reserved.

35

Article

doi:10.1006/geno.2001.6577, available online at http://www.idealibrary.com on IDEAL

MEST values for parental and hybrid mice are shown in Fig. 1. Of the two parental strains, D2 mice exhibit a dramatically lower threshold (24.6 + 2.5 mA, mean + S.D., n = 50; 25 male, 25 female) than B6 mice (75.1 + 6.0, n = 50; 25 male, 25 female), which confirms earlier work [15,18]. We observed a significant gender difference in the B6 strain, with females exhibiting a significantly lower MEST (71.3 + 4.0 mA, n = 25) than males (77.6 + 5.8, n = 25, P < 0.001, unpaired t test assuming unequal variance). F1 and F2 progeny exhibit intermediate thresholds (44.1 + 5.2, n = 30; 42.3 + 11.6, n = 294, respectively). F2 female mice (40.2 + 11.1) are more sensitive than males (46.0 + 12.7); however, there is no effect of

gender among F1 mice. Using the formula of Sewell and Wright [19], we estimated broad sense heritability to be 0.75. Data were log10 transformed to normalize parental variances before heritability calculations. The distribution of threshold values across the F2 population is depicted in Fig. 2. Fig. 2A shows raw ECS data and Fig. 2B shows log-transformed data. Both data sets diverge from a normal distribution (KS distance 0.1825, P < 0.001 for raw data; KS distance 0.1405, P < 0.001 for log-transformed data; Kolmogorov–Smirnov test). We observed no significant effect of body weight, but there was a significant effect of gender as described above, with F2 females exhibiting lower MEST values than males (F = 7.228, P < 0.01; ANOVA). We noted no significant interactions between gender and genotype at any marker locus. Univariate mapping results are given in Table 1 for all markers reaching an arbitrary significance level of P < 0.02. Seven chromosomes are represented, with chromosome 1 exhibiting the strongest statistical evidence for linkage. In all but one case (chromosome 11), the D2 allele is associated with lower MEST. Table 2 shows the results of Mapmaker/QTL analyses for genomic intervals in which lod scores exceed 2.8 on firstpass scanning, a proposed threshold for declaring suggestive linkage in QTL studies involving F2 intercross-based mapping [20]. Multipoint mapping results generally confirm univariate χ2 analyses, providing suggestive or significant evidence for QTLs on 7 different chromosomes with evidence for two QTLs each on chromosomes 2 and 15. We detected the locus of greatest effect on distal chromosome 1 and it may be considered a QTL of major effect, explaining nearly one-quarter of the total phenotypic variance. We detected other QTLs yielding a lod score > 4.3 [20] on chromosomes 2, 5 and 15. Mapmaker/QTL LOD plots for chromosomes exhibiting the highest scores are shown (Fig. 3). Analysis with the Mapmaker NP program yields Z scores for these genomic intervals that also exceed the suggested threshold for declaring linkage [20] (data not shown). When

A

B

FIG. 1. MEST in B6, D2, and B6 × D2 hybrid mice. Each experimental group was composed of equal numbers of male and female mice and all mice were 8–12 weeks old when testing began. The top of each bar represents mean (+S.D.) MEST (mA) for each group.

gene(s) on distal chromosome 1 that has a fundamental role in determining differences in seizure sensitivity between C57 and DBA inbred strains.

RESULTS

FIG. 2. Distribution of MEST values among F2 mice. The top of each bar indicates the number of F2 mice in data bin. (A) Raw data. (B) Log10-transformed data.

36

GENOMICS Vol. 75, Numbers 1–3, July 2001 Copyright © 2001 by Academic Press. All rights of reproduction in any form reserved.

Article

doi:10.1006/geno.2001.6577, available online at http://www.idealibrary.com on IDEAL

TABLE 1. Univariate Contingency Analysisa Marker (cM2)b

χ2

D1Mit157 (58.7)

20.76

3 × 10–5

D2

D1Mit30 (70.0)

20.79

3 × 10–5

D2

D1Mit110 (88.1)

31.19

2 × 10–7

D2

D1Mit150 (100.0)

18.68

8 × 10–5

D2

D1Mit17 (106.3)

10.21

0.010

D2

D2Mit1 (1.0)

9.02

0.015

D2

D2Mit521 (15.3)

15.48

4.3 × 10–4

D2

D2Mit156 (32.0)

13.42

1.2 × 10–3

D2

P value

Susceptibility allelec

D2Mit285 (86.0)

8.97

0.015

D2

D2Mit200 (106.0)

10.01

6.7 × 10–3

D2

D5Mit294 (8.0)

15.66

4.0 × 10–4

D2

D5Mit15 (39.0)

15.16

5.1 × 10–4

D2

er, the difference is not statistically significant under the conditions of this analysis. To better characterize the extent of the introgressed D2 genomic interval in Mtv7 mice, we typed polymorphic Mit markers in the region. Results indicate that the Mtv7 strain harbors D2 genetic material bounded by markers D1Mit206 (95.8 cM, http://www.informatics.jax.org/) and D1Mit221 (102 cM, http://www.informatics.jax.org/) on a B6 background (Table 3). Contrary to the public database, D1Mit221 mapped distal to D1Mit273 in our crosses. We obtained additional information on the interval upon typing a polymorphism in Atp1a2 (94.2 cM, http://www.informatics.jax.org/), which showed that the Mtv7 strain harbors the D2 allele. Thus, there is a proximal region (between D1Mit145 and Atp1a2) and a smaller distal region (between D1Mit221 and D1Mit17) where the origin of the genetic material (B6 or D2) remains unknown. Additional informative DNA markers and gene polymorphisms are required to better define the limits of these boundaries.

D8Mit271 (57.0)

8.96

0.015

D2

D10Mit186 (40.0)

11.01

4.0 × 10–3

D2

DISCUSSION

D11Mit228 (2.0)

9.31

9.5 × 10–3

B6

D11Mit236 (20.0)

10.03

6.6 ×10–3

B6

The results of this study are consistent with previous reports on dissection of seizure susceptibility in B6 and D2 mice in that seizure-related traits in these strains have been found to be highly heritable with evidence for the influence of multiple genes as well as environmental variables. Whereas previous studies on adult B6 and D2 mice focused on pharmacologically induced susceptibility to seizures [1,2,22,23], the experiments reported here involve a nonpharmacological seizure paradigm, an important consideration because variables associated with strain-specific pharmacokinetic effects are obviated. In spite of the substantial difference in experimental paradigm, several areas of overlap exist between loci that are involved in determining MEST and loci that influence drug-related seizure sensitivities. The most notable area of overlap is the distal region of chromosome 1, which was documented previously to contribute the largest genetic effect in chemoconvulsant seizure screens involving kainic acid [1] and PTZ [2]. In fact, the same marker, D1Mit150, resides at or near the peak lod score in the MEST screen (Fig. 3A) and the chemoconvulsant screens [1,2]. This same region of the genome also contributes large effects in seizure paradigms involving withdrawal from ethanol [22] and pentobarbital [23]. Taken together, the coincidence of mapping results in studies that invoke such a wide variety of neurochemical and physiological mechanisms indicates that distal chromosome 1 harbors a gene (or genes) that has a fundamental role in controlling neuronal excitability. Not all studies that involve genetic mapping of seizure susceptibility in B6 and D2 mice have found evidence for a chromosome 1 locus, however. Such studies include reports of susceptibility to acute cocaine-induced seizures [4], PTZ-induced seizures [24], and audiogenic seizures [25,26], although the latter model uses very young (~3-week-old) mice.

D11Mit157 (34.2)

8.42

0.015

B6

D13Mit9 (45.0)

13.02

1.5 × 10–3

D2

D13Mit148 (59.0)

8.00

0.02

D2

D15Mit13 (6.7)

12.14

2.3 × 10–3

D2

2.3 × 10–3

D15Mit34 (52.2) 12.10 D2 × 3 contingency table was formulated for each of 90 microsatellite marker loci to

aA 2

determine the probability that the distribution of genotypes between the groups of F2 mice exhibiting the lowest (n = 48) and highest (n = 41) MEST values occurred by random assortment of alleles. bMarker positions were taken from databases at http://www.genome.wi.mit.edu/ and http://www.informatics.jax.org/. cAllele associated with low MEST.

all QTLs (chromosomes 1, 2, 5, and 15) were fixed and mapped simultaneously with Mapmaker/QTL, 62% of the phenotypic variance was explained. Thus, these QTLs explain most of the genetic variance in the model because heritability was estimated to be about 75%. We further investigated the effect of the chromosome 1 QTL using the B6.D2-Mtv-7a/Ty (Mtv7) congenic strain, which contains a limited portion of D2 genome from distal chromosome 1 on a B6 genetic background [21]. We tested Mtv7 mice for MEST as described above (Fig. 4). Comparison of MEST values in parental strains with homozygous and heterozygous (with respect to the introgressed segment of D2 genome) Mtv7 mice reveals significant differences between D2 mice and both Mtv7 genotypes (P < 0.01 versus heterozygous Mtv7; P < 0.05 versus homozygous Mtv7; Kruskal–Wallis ANOVA). B6 mice are significantly different from D2 mice and homozygous Mtv7 mice (P < 0.01; Kruskal–Wallis ANOVA). There is a trend for lower values in heterozygous Mtv7 mice than in B6; howev-

GENOMICS Vol. 75, Numbers 1–3, July 2001 Copyright © 2001 by Academic Press. All rights of reproduction in any form reserved.

37

Article

doi:10.1006/geno.2001.6577, available online at http://www.idealibrary.com on IDEAL

A

B

C

D

FIG. 3. Genetic maps showing interval lod scores for linkage to loci controlling MEST in B6 × D2 F2 mice. Maps were derived with the Mapmaker/QTL software program. Markers (cM) used for interval mapping include the following: (A) chromosome 1, D1Mit427 (8.4), D1Mit411 (18.5), D1Mit156 (32.8), D1Mit7 (41), D1Mit157 (58.7), D1Mit30 (70), D1Mit110 (88.1), D1Mit150 (100), D1Mit17 (106.3), and D1Mit155 (112); (B) chromosome 2, D2Mit175 (2), D2Mit521 (15.3), D2Mit156 (32), D2Mit102 (52.5), D2Mit133 (66.8), D2Mit285 (86), and D2Mit200 (107); (C) chromosome 5, D5Mit294 (8), D5Mit352 (20) D5Mit15 (39), D5Mit278 (61), and D5Mit168 (78); (D) chromosome 15, D15Mit13 (6.7), D15Mit229 (26.2), D15Mit158 (46.9), D15Mit34 (52.2), and D15Mit15 (64.8). Results are shown for model independent as well as fixed model analyses and were generated with genotypes from all F2 mice or from high and low MEST groups only. Horizontal line at LOD = 4.3 indicates a proposed threshold for declaring significant linkage in F2 intercross populations tested with constrained genetic models [20].

The ultimate goal of QTL studies is to identify the underlying genetic mutation or variation that results in functional alteration and contributes to a definable phenotype; however, highly refined genetic maps of a QTL are a prerequisite to gene identification. One favored means of more precisely defining the map position of a QTL involves using congenic strains [27]. In this study, we took advantage of a commercially available chromosome 1 congenic strain, B6.D2Mtv7a/Ty (Mtv7), that was produced to study genetic control of immunologic responses [21]. When we used PCRbased markers to characterize the extent of the introgressed

38

interval in this strain, we found that it contains a very limited tract of D2-origin genetic material including marker D1Mit150 (Table 3). Subsequent seizure testing revealed the Mtv7 strain is D2-like in that it exhibits a relatively low MEST (Fig. 4) and is highly susceptible to PTZ-induced maximal seizures (data not shown). These data suggest that the gene(s) with the greatest influence on overall seizure sensitivity in mature D2 mice has been captured in the Mtv7 congenic strain. The magnitude of the reduction of MEST in the Mtv7 strain relative to the B6 background strain suggests that the phenotype is sufficiently robust to

GENOMICS Vol. 75, Numbers 1–3, July 2001 Copyright © 2001 by Academic Press. All rights of reproduction in any form reserved.

doi:10.1006/geno.2001.6577, available online at http://www.idealibrary.com on IDEAL

Article

TABLE 2. Summary of Multipoint Linkage Analysisa

region is considerably smaller [28,29]. Most of the genes that map Peak LOD Peak LOD Peak LOD Variance to this region in mice have homo(dominant) (recessive) (additive) explainedc logues in the human genome on chromosome 1q21–q23. A smaller D1Mit110–D1Mit150 12.6 2.9 9.2 0.24 fraction of mouse genes from the Mtv7 interval have homologues on D2Mit1–D2Mit521 3.7 2.7 4.4 0.08 1q41–q44. Several genes that reside D2Mit285–D2Mit200 2.9 1.0 2.6 0.05 in this interval are compelling in D5Mit294–D5Mit15 6.8 2.0 5.3 0.16 terms of their potential for influD10Mit44–D10MIt186 2.2 1.7 2.9 0.06 encing neuronal excitability and seizure threshold including Kcnj9 D11Mit236–D11Mit157 3.2 0.2 1.8 0.08 (94.2 cM; http://www.informatD13Mit9–D13Mit148 1.5 2.3 2.9 0.07 ics.jax.org/) and Kcnj10 (93.5 cM; D15Mit13–D15Mit229 5.0 0.5 3.5 0.09 http://www.informatics. jax.org/), which encode potassium D15Mit158–D15Mit34 5.0 0.7 3.4 0.09 aMultipoint analysis was conducted with the Mapmaker/QTL software package [45,46]. Analyses on chromosomes 1, 2, 5, and ion channel proteins, and Atp1a2, 15 were conducted with phenotype information from all 294 F2 progeny and genotype data from F2 progeny from the low (n = which encodes a subunit of Na,K48), high (n = 41), and midrange (n = 187) MEST groups. Analyses for chromosomes 10, 11, and 13 were conducted with genoATPase. In spite of their provocatypes from low and high MEST groups only. Phenotype data were log10-transformed before analysis. bIntervals are included if lod scores exceed the Lander and Kruglyak [20] threshold (LOD = 2.8) for declaring suggestive evidence tive nature, however, these genes for a QTL in an F2 intercross with 2 degrees of freedom (three genetic models tested: dominant, recessive, and additive). are only several among many that cFraction of total phenotypic variance explained by individual QTL. reside in the critical interval; thus, further systematic reduction of this interval is required. Consistent with a polygenic mode of inheritance for genpermit straightforward characterization of recombinant offeral seizure susceptibility, we also detected loci of signifispring and systematic reduction of the critical interval. The cant effect on chromosomes 2, 5, and 15, with chromosome phenotype is readily detectable in heterozygotes as well as 15 yielding evidence for two QTLs, one near the acromere homozygotes and a gene dosage effect was observed (Fig. and another at the telomere. Genetic intervals that define 4). It will be useful to confirm this phenotype in an indesome of these QTLs overlap broadly with intervals identipendent congenic strain and also to examine the phenotype fied in previous studies. The chromosome 5 locus maps to a of a reciprocal D2.B6 congenic strain. These studies are region harboring a QTL for susceptibility to both kainic acid under way in our laboratory. [1] and PTZ seizures [2], and the proximal chromosome 15 The relative lack of known informative DNA markers locus maps to a region where a QTL for kainic acid seizure results in flanking regions of unknown (B6 or D2) origin so sensitivity was detected [1]. Although it is possible that the that the extent of the introgressed genomic region in the proximal chromosome 15 QTL identified in this study and Mtv7 strain genome cannot be calculated with certainty. Discrepancies in marker–marker recombination frequencies between publicly accessible databases also complicate delineation of the Mtv7 congenic interval so that the introgressed region may be estimated to be as small as 7 cM or as large as 17 cM. Further insight into the extent of this chromosome 1 region of interest comes from recent work on the looptail mutation in which a physical map of the looptail critical interval indicates that the distance between markers D1Mit113 and D1Mit150 is about 2.6 Mb [28,29]. These data contrast with genetic map data that place D1Mit113 at 93.3 cM and D1Mit 150 at 100 cM (http://www. informatics.jax.org/). In our laboratory, analysis of 306 progeny derived from crosses of heterozygous Mtv7 mice and parental B6 mice has revealed only 11 recombinants between informative markers D1Mit206 (95.8 cM; FIG. 4. MEST in B6, D2, and Mtv7 congenic mice. Mice (male, 8–12 weeks old) http://www.informatics.jax.org/) and D1Mit273 (102 cM; were tested for MEST. The top of each bar represents mean (+S.D.); B6 http://www.informatics.jax.org/). These data are in gener(n = 25), D2 (n = 25), Mtv/Mtv (n = 16), Mtv/B6 (n = 14). *P < 0.01 versus D2, Mtv/Mtv; **P < 0.01 versus B6, Mtv/B6; P < 0.05 versus Mtv/Mtv al agreement with public database information, and under(Kruskall–Wallis ANOVA; Neuman Keuls post hoc). Mtv/Mtv, B6.D2score the limitations of gene mapping by genetic recombiMtv7a/Ty; Mtv/B6, B6 x B6.D2-Mtv7a/Ty. nation alone since physical mapping data indicate that the Marker intervalb

GENOMICS Vol. 75, Numbers 1–3, July 2001 Copyright © 2001 by Academic Press. All rights of reproduction in any form reserved.

39

Article

doi:10.1006/geno.2001.6577, available online at http://www.idealibrary.com on IDEAL

TABLE 3. Characterization of Chromosome 1 Markers in B6.D2-Mtv7a/Ty Congenic Mice Marker

D1Mit16

D1Mit145

Atp1a2

D1Mit206

D1Mit150

D1Mit360

D1Mit221

D1Mit273

D1Mit17

MITa

(cM) b MGD (cM)

84.2

87.4

—

94.0

99.5

102.7

104.9

106.0

110.4

87.2

89.0

94.2

95.8

100.0

101.2

102.0

102.5

106.3

Allele

B6

B6

D2

D2

D2

D2

D2

D2

B6

aGenetic distances were taken from the database at http://www.genome.wi.mit.edu/. bGenetic distances were taken from the database at http://www.informatics.jax.org/.

the one identified in the kainic acid study [1] may reflect the same genetic influence, it is not likely that the same is true for the chromosome 5 QTL because the kainic acid [1] and PTZ [2] sensitivity loci on chromosome 5 are derived from the B6 strain, whereas the QTL for low MEST is derived from the D2 strain. Finally, there is a potential overlap between the chromosome 2 QTL and a seizure frequency QTL mapped previously in the EL mouse strain [30]. Studies of seizure susceptibility in mice continue to define a research area marked by great complexity with regard to the nature of the components involved as well as the most useful models. Spontaneous mutations of single genes that alone are sufficient to cause seizure phenotypes have so far been most amenable to identification [31–34], although the relationship of these models to most cases of human idiopathic epilepsy is uncertain given that most common forms of the disease are multifactorial and are usually inherited in non-mendelian fashion [35–39]. Polygenic seizure models may have more relevance to common forms of human epilepsy; however, the feasibility of isolating susceptibility genes in such models has not been proven. One polygenic model that has been intensively studied with regard to dissection of the contributing genetic factors is represented by the EL mouse strain [40]. Systematic genetic mapping using a “tossing”induced seizure phenotype [30] and congenic strain studies [41] localized the EL QTL of greatest effect to chromosome 2. Attempts to further reduce the critical QTL interval through the use of derivative recombinant congenic strains provided evidence that the effect represents the combined action of at least two linked genes [42], although this conclusion is tempered by discrepancies between the two seizure paradigms used to screen the congenic strains and the recombinant progeny. Differences in susceptibility to tossing-induced seizure frequency and PTZ-induced tonic-clonic seizure susceptibility extended to specific gender effects as well as to specific strain effects [42], which calls into question the relationship between these two experimental paradigms. Also, because initial QTL mapping was done in EL mice with the tossing phenotype alone [30], the extent to which PTZ susceptibility in particular, and chemoconvulsant sensitivity in general, is associated with the El2 locus is not clear. Without establishing full trait parameters for PTZ sensitivity, it is difficult to fairly evaluate EL congenic testing results [42]. Seizure phenotypes in mice are characterized by substantial variability, and this is particularly true in studies that involve chemoconvulsants [15]. It is possible that the use of MEST may provide a more

40

useful measure of general seizure sensitivity because it is characterized by substantially less variability [15,18] and that application of this test to EL mice may help resolve the underlying genetics of that model. We have mapped the chromosomal location of genes that mediate the dramatic difference in MEST between B6 and D2 mice and we have established that the distal region of chromosome 1 contains a gene(s) of major effect. The critical genomic interval overlaps with major QTLs for susceptibility to seizures induced by kainic acid, PTZ, and withdrawal from central nervous system depressants, which indicates the presence of a gene of fundamental importance in controlling neuronal excitability. Evaluation of a commercially available congenic strain indicates that the critical interval is flanked by markers D1Mit145 and D1Mit17 and that it may be less than 10 cM. The robust nature of the congenic strain phenotype will facilitate further systematic reduction of the interval so that positional cloning of a single causative gene may ultimately become a reality. Although it is possible that multiple genes within the Mtv7 interval contribute to the seizure phenotype, no established precedent conclusively indicates that this possibility is any more likely than that in which the phenotype is the result of a single gene. Only further assessment of the model will answer this question.

MATERIALS AND METHODS Animals. We purchased C57BL/6J (B6), DBA/2J (D2), and B6.D2-Mtv-7a/Ty (Mtv7) mice of both genders from the Jackson Laboratory (Bar Harbor, ME) and bred them in house to propagate mice for these studies. We crossed female B6 mice with male D2 mice to produce F1 progeny. F1 intercrosses produced F2 progeny. Mice were weaned at about 4 weeks of age and group housed (three or four mice per cage) by gender. We maintained the mice on a 12-h light/12-h dark schedule with food and water freely available. Experiments were approved by Animal Care and Use Committees overseeing each participating laboratory. Seizure testing. MEST was determined as described [18]. We tested mice (8–12 weeks old) once per day. The current level initially was set at 20 mA and increased by 1 mA (5 mA for F2 mice) with each successive trial until a maximal seizure was observed. The ECS apparatus (model 7801, Ugo Basile, Varese, Italy) was equipped with earclip electrodes and the pulse train generator was used in a square-wave mode. We delivered shocks with a frequency of 60 Hz, a pulse width of 0.4 ms, and a duration of 0.2 s. A feedback circuit within the ECS unit determined a resistance value for each animal and delivered current at a constant level. After we attached the electrodes, we placed the mice in a Plexiglas cage and administered the shock immediately. Each shock induced at least a partial clonic seizure. The following sequence of events characterizes a session in which a maximal seizure occurred: tonic forelimb flexion, tonic hindlimb flexion, tonic hindlimb extension, and

GENOMICS Vol. 75, Numbers 1–3, July 2001 Copyright © 2001 by Academic Press. All rights of reproduction in any form reserved.

doi:10.1006/geno.2001.6577, available online at http://www.idealibrary.com on IDEAL

hindlimb clonus. The maximal seizure event was scored positive upon observation of tonic hindlimb extension. Microsatellite marker genotyping. We constructed a genotyping panel of 90 DNA markers from published maps [43] and an online database (http://www.informatics.jax.org/). The mean (+S.D.) distance between markers was 16.7 (+4.4) cM, with the largest gap estimated at 33 cM between markers D10Mit28 and D10Mit186. Other regions of the genome where distances between markers were estimated to be greater than 22 cM include D3Mit29–D3Mit200, D6Mit159–D6Mit263, D12Mit105–D12Mit114, and D16Mit130–D16Mit13. We tried to use markers with alleles that differed by more than 8 bp in length between strains. For such markers, we analyzed PCR amplicons by agarose gel electrophoresis with ethidium bromide staining [44]. After electrophoresis, two independent scorers recorded genotypes from Polaroid 667 (3000 ISO) black and white prints and entered them into a database for subsequent error checking. Alternatively, gels were analyzed with a digital gel documentation system (BioRad, Hercules, CA), which provides thermal prints of digitized images from gels stained with ethidium bromide. A third scorer resolved discrepancies in genotype readings. Irreconcilable discrepancies were discarded from analysis and the genotype was scored as unknown. For marker loci with alleles with less than an 8-bp difference in length between strains, genotyping involved using 32P-end-labeled primers, polyacrylamide gel electrophoresis, and autoradiography as described [1]. Two independent individuals scored genotypes from autoradiograms, which were checked for errors as described above. Atp1a2 polymorphism detection and genotyping. B6 and D2 mice were anesthetized and killed by cervical dislocation. Brains were removed and frozen at –80ºC. Frozen tissue (~100 mg) was homogenized in Ultraspec (Biotecx Labs, Inc., Houston, TX) and RNA was isolated according to the manufacturer’s protocols. We estimated the integrity and quantity of RNA by formaldehyde-agarose gel electrophoresis and ethidium bromide staining. RNA was reverse transcribed with Superscript (Life Technologies, Inc., Gaithersburg, MD). No information on the mouse gene is publically available, so we used primers designed against the rat Atp1a2 sequence (GenBank acc. no. NM012505) for PCR. The primer sequences are as follows: forward, 5′ACCTTTCCAGCCTAGGTC-3′; reverse, 5′-TGCTGAGGCACCATGTTC-3′. The forward primer was targeted to a DNA sequence in the 5′ untranslated region of the gene, whereas the reverse primer was targeted to coding sequence. Together they yielded a 537-bp product. We subcloned PCR amplicons into pTOPOII vector (In Vitrogen, Carlsbad, CA) and propagated them in Escherichia coli strain JM109. We isolated DNA from recombinant clones using the Wizard plasmid DNA isolation kit (Promega, Madison, WI) and sequenced it using vector primers T7 and SP6. Sequencing was performed in the University of Pennsylvania Genetics core facility with ABI automated sequencing methodology. We derived cDNA pools from at least two animals per strain and used each of these pools to perform PCR. PCR amplicons were generated from multiple pools, subcloned, and sequenced at least three times with at least one of the sequencing analyses carried out on the opposite (noncoding) strand. To genotype a single nucleotide polymorphism detected in the sequence of Atp1a2 (T354C), we incubated the 537-bp PCR amplicon with NcoI. The B6 amplicon contains a single restriction site for NcoI, resulting in 384- and 153-bp fragments. The D2 polymorphism introduces a second NcoI site so that fragments of 240, 153, and 144 bp are generated. Fragments were analyzed by electrophoresis on 0.8% agarose gels stained with ethidium bromide and visualized by ultraviolet irradiation. Data analysis. Complementary statistical tests used to map loci that influence MEST were similar to those used in previous seizure QTL studies [1,2]. We calculated trait heritability by the method of Sewall and Wright, using response variance of parental and hybrid populations [19]. For purposes of selective genotyping used in the first-pass genome scan, we segregated F2 mice based on MEST values into low-threshold (25–30 mA; n = 49), midrangethreshold (35–55 mA; n = 207), and high-threshold (60–90 mA; n = 41) groups. Initial statistical mapping procedures utilized genotypes derived solely from the low- and high-threshold groups. Cutoff values for selecting extreme responding mice (15% highest and lowest values) were based on achieving at least 90% power to detect a QTL, explaining 20% of the phenotypic variance [45]. After a first-pass screening, we collected genotypes from a large portion (n = 187) of the midrange group at selected markers that defined regions giving greatest evidence for the presence of QTLs. Specifically, these markers

Article

were as follows: chromosome 1, D1Mit30, D1Mit110, and D1Mit150; chromosome 2, D2Mit1, D2Mit521, D2Mit285, and D2Mit200; chromosome 5, D5Mit15, D5Mit352, and D5Mit294; chromosome 15, D15Mit13, D15Mit229, D15Mit158, and D15Mit34. The initial statistical mapping strategy involved evaluating associations between marker genotypes and MEST by χ2 contingency analysis with genotype (B6/B6, D2/D2, B6/D2) and MEST category (low, high) taken as the grouping variables. We assessed the influence of other phenotypic variables on MEST by multiway analysis of variance (ANOVA) that included gender, body weight, and marker genotypes as independent variables and MEST as the outcome variable. We identified regions of the genome where χ2 analysis indicated a significant association between phenotype and genotype with a liberal threshold (P < 0.02), and we analyzed chromosomes containing those regions with standard Mapmaker software [45,46]. This suite of programs generates linkage maps (Mapmaker/Exp) and enables multipoint linkage analysis through calculation of maximum likelihood ratios (Mapmaker/QTL). Output of the Mapmaker/QTL program includes an estimate of the phenotypic variance explained by individual QTLs. We also used Mapmaker/QTL to evaluate the mode of inheritance by fitting data to each of three genetic models (dominant, recessive, and additive). After initial QTL detection, we refined genetic mapping by fitting multiple QTLs simultaneously. QTLs with logarithm of odds (lod) scores greater than 4.3 were entered into this analysis. The threshold was based on the guidelines proposed by Lander and Kruglyak [20] for declaring significant linkage in a QTL screen involving an F2 intercross population evaluated with an unconstrained statistical genetic model. Fixing QTLs allows control over part of the variance, thus reducing unexplained noise in the analysis and facilitating evaluation of other QTLs. Finally, we analyzed data from regions associated with significant QTLs with Mapmaker NP, the nonparametric version of Mapmaker/QTL [47], to assess the possibility that the nonnormal distribution of F2 phenotypes yielded statistical artifacts.

ACKNOWLEDGMENTS This work was supported by National Institutes of Health grant NS33243. Thanks to Wayne Frankel (The Jackson Laboratory, Bar Harbor, ME) for bringing the existence of mouse strain B6.D2-Mtv-7a/Ty to our attention and to Lisa Tarantino (Novartis Research Institute, La Jolla, CA) for helping with the Mapmaker NP program. RECEIVED FOR PUBLICATION FEBRUARY 20, 2001; ACCEPTED APRIL 11, 2001.

REFERENCES 1. Ferraro, T. N., et al. (1997). Mapping murine loci for seizure response to kainic acid. Mamm. Genome 7: 200–208. 2. Ferraro, T. N., et al. (1999). Mapping loci for pentylenetetrazol-induced seizure susceptibility in mice. J. Neurosci. 19: 6733–6739. 3. Martin, B., Clement, Y., Venault, P., and Chapouthier, G. (1995). Mouse chromosomes 4 and 13 are involved in b-carboline-induced seizures. J. Hered. 86: 274–279. 4. Hain, H. S., Crabbe, J. C., Bergeson, S. E., and Belknap, J. K. (2000). Cocaine-induced seizure thresholds: quantitative trait loci detection and mapping in two populations derived from the C57BL/6 and DBA/2 mouse strains. J. Pharmacol. Ther. 293: 180–187. 5. Toman, J. E .P., Swinyard, E. A., and Goodman, L.S. (1946). Properties of maximal seizures and their alteration by anticonvulsant drugs and other agents. J. Neurophysiol. 9: 231–239. 6. Swinyard, E.A. (1972). Electrically induced convulsions. In Experimental Models of Epilepsy—A Manual for the Laboratory Worker (D. Purpura, J. Penry, D. Tower, D. Woodbury, and R. Walter, Eds.), pp. 433–458. Raven Press, New York. 7. Friese, A., Seiller, E., Quack, G., Lorenz, B., and Kreuter, J. (2000). Increase of the duration of the anticonvulsive activity of a novel NMDA receptor antagonist using poly(butylcyanoacrylate) nanoparticles as a parenteral controlled release system. Eur. J. Pharm. Biopharm. 49: 103–109. 8. Sills, G. J., Butler, E., Thompson, G. G., and Brodie, M. J. (1999). Vigabatrin and tiagabine are pharmacologically different drugs. A pre-clinical study. Seizure 8: 404–411. 9. White, H. S., Brown, S. D., Woodhead, J. H., Skeen, G.A., and Wolf, H. H. (1997). Topiramate enhances GABA-mediated chloride flux and GABA-evoked chloride currents in murine brain neurons and increases seizure threshold. Epilepsy Res. 28: 167–179. 10. Browning, R. A., and Nelson, D. K. (1985). Variation in threshold and pattern of electroshock-induced seizures in rats depending on site of stimulation. Life Sci. 37: 2205–2211. 11. Kitano, Y., Usui, C., Takasuna, K., Hirohashi, M., and Nomura, M. (1996). Increasing-

GENOMICS Vol. 75, Numbers 1–3, July 2001 Copyright © 2001 by Academic Press. All rights of reproduction in any form reserved.

41

Article

12. 13.

14. 15. 16. 17.

18. 19. 20. 21. 22.

23.

24.

25. 26.

27. 28.

doi:10.1006/geno.2001.6577, available online at http://www.idealibrary.com on IDEAL

current electroshock seizure test: a new method for assessment of anti- and pro-convulsant activities of drugs in mice. J. Pharmacol. Toxicol. Meth. 35: 25–29. Hall, C. S. (1947). Genetic differences in fatal audiogenic seizures between two inbred strains of house mice. J. Hered. 38: 2–6. Schlesinger, K., Boggan, W. O., and Griek, B. J. (1968). Pharmacogenetic correlates of pentylenetetrazole and electroconvulsive seizure thresholds in mice. Psychopharmacol. (Berl) 13: 181–188. Taylor, B. A. (1976). Genetic analysis of susceptibility to isoniazid-induced seizures in mice. Genetics 83: 373–377. Engstrom, F., and Woodbury, D. M. (1988). Seizure susceptibility in DBA and C57 mice: the effects of various convulsants. Epilepsia 29: 389–395. Kosobud, A. E., and Crabbe, J. C. (1990). Genetic correlations among inbred strain sensitivities to convulsions induced by 9 convulsant drugs. Brain Res. 526: 8–16. Ferraro, T. N., Golden, G. T., Smith, G. G., and Berrettini, W. H. (1995). Differential susceptibility to seizures induced by systemic kainic acid treatment in mature DBA/2J and C57BL/6J mice. Epilepsia 36: 301–307. Ferraro, T. N., et al. (1998). Genetic influences on electrical seizure threshold. Brain Res. 813: 207–210. Lande, R. (1981). The minimum number of genes contributing to quantitative variation between and within populations. Genetics 99: 541–553. Lander, E. S., and Kruglyak, L. (1995). Genetic dissection of complex traits: guidelines for interpreting and reporting linkage results. Nat. Genet. 11: 241–247. Taylor, B. A., and Frankel, W. N. (1993). A new strain congenic for the Mtv-7/Mls-1 locus of mouse chromosome 1. Immunogenetics 38: 235–237. Buck, K. J., Metten, P., Belknap, J. K., and Crabbe, J. C. (1997). Quantitative trait loci involved in genetic predisposition to acute alcohol withdrawal in mice. J. Neurosci. 17: 3946–3955. Buck, K. J., Metten, P., Belknap, J. K., and Crabbe, J. C. (1999). Quantitative trait loci affecting risk for pentobarbital withdrawal map near alcohol withdrawal loci on mouse chromosomes 1, 4 and 11. Mamm. Genome 10: 431–437. Wakana, S., et al. (2000). Gene mapping of SEZ group genes and determination of pentylenetetrazol susceptible quantitative trait loci in the mouse chromosome. Brain Res. 857: 286–290. Neumann, P. E., and Collins, R. L. (1991). Genetic dissection of susceptibility to audiogenic seizures in inbred mice. Proc. Natl. Acad. Sci. USA 88: 5408–5412. Neumann, P. E., and Seyfried, T. N. (1990). Mapping of two genes that influence susceptibility to audiogenic seizures in crosses of C57BL/6J and DBA/2J mice. Behav. Genet. 20: 307–323. Flaherty, L. (1981). Congenic strains. In The Mouse in Biomedical Research, Vol. 1 (H. Foster, J. Small, and J. Fox, Eds.), pp. 215–222. Academic Press, New York. Doudney, K., et al. (2001). Comparative physical and transcript maps of ~1 Mb around looptail, a gene for severe neural tube defects on distal mouse chromosome 1 and human chromosome 1q22-q23. Genomics 72: 180–192.

42

29. Eddleston, J., Murdoch, J. N., Copp, A. J., and Stanier, P. (1999). Physical and transcriptional map of a 3-Mb region of mouse chromosome 1 containing the gene for the neural tube defect mutant loop-tail (Lp). Genomics 56: 149–159. 30. Frankel, W. N., et al. (1995). New seizure frequency QTL and the complex genetics of epilepsy in EL mice. Mamm. Genome 6: 830–838. 31. Letts, V.A., et al. (1998). The mouse stargazer gene encodes a neuronal Ca2+-channel g subunit. Nature Genet. 19: 340–346. 32. Burgess, D.L., Jones, J. M., Meisler, M. H., and Noebels, J.L. (1997). Mutation of the Ca2+ channel b subunit gene Cchb4 is associated with ataxia and seizures in the lethargic (lh) mouse. Cell 88: 385–392. 33. Cox, G. A., et al. (1997). Sodium/hydrogen exchanger gene defect in slow-wave epilepsy mutant mice. Cell 3: 139–148. 34. Fletcher, C. F., et al. (1996). Absence epilepsy in tottering mutant mice is associated with calcium channel defects. Cell 87: 607–617. 35. Callenbach, P. M., et al. (1998). Familial occurrence of epilepsy in children with newly diagnosed multiple seizures: Dutch Study of Epilepsy in Childhood. Epilepsia 39: 331–336. 36. Jain, S., Padma, M. V., Puri, A., and Maheshwari, M. C. (1998). Juvenile myoclonic epilepsy: disease expression among Indian families. Acta Neurol. Scand. 97: 1–7. 37. Berkovic, S. F. (1997). Epilepsy genes and the genetics of epilepsy syndromes: the promise of new therapies based on genetic knowledge. Epilepsia 38 (suppl. 9): S32–S36. 38. Sander, T. (1996). The genetics of idiopathic epilepsy: implications for the understanding of its aetiology. Mol. Med. Today 8: 173–180. 39. Johnson, W. G., et al. (1996). Pedigree analysis in families with febrile seizures. Am. J. Med. Genet. 61: 345–352. 40. Suzuki, J., and Nakamoto, Y. (1977). Seizure patterns and electroencephalograms of El mouse. Electroencephalogr. Clin. Neurophysiol. 43: 299–311. 41. Frankel, W. N., Johnson, E. W., and Lutz, C. M. (1995). Congenic strains reveal effects of the epilepsy quantitative trait locus, El2, separate from other El loci. Mamm. Genome 6: 839–843. 42. Legare, M. E., Bartlett, F. S., and Frankel, W. N. (2000). A major effect QTL determined by multiple genes in epileptic EL mice. Genome Res. 10: 42–48. 43. Dietrich, W., et al. (1996). A comprehensive genetic map of the mouse genome. Nature 380: 149–152. 44. Ferraro, T. N., et al. (1998). Genotyping microsatellite polymorphisms by agarose gel electrophoresis with ethidium bromide staining: application to quantitative trait loci analysis of seizure susceptibility in mice. Psychiatr. Genet. 8: 227–233. 45. Lander, E. S., and Botstein, D. (1989). Mapping Mendelian factors underlying quantitative traits using RFLP linkage maps. Genetics 121: 185–199. 46. Lander, E. S., et al. (1987). MAPMAKER: An interactive computer package for constructing primary genetic linkage maps of experimental and natural populations. Genomics 1: 174–181. 47. Kruglyak, L., and Lander, E. S. (1995). A nonparametric approach for mapping quantitative trait loci. Genetics 139: 1421–1428.

GENOMICS Vol. 75, Numbers 1–3, July 2001 Copyright © 2001 by Academic Press. All rights of reproduction in any form reserved.