Multiple quantitative trait locus (QTL) mapping studies designed to localize seizure susceptibility genes in C57BL/6 (B6, seizure resistant) and DBA/2 (D2, seizure susceptible) mice have detected a significant effect originating from midchromosome 5. To confirm the presence and refine the position of the chromosome 5 QTL for maximal electroshock seizure threshold (MEST), reciprocal congenic strains between B6 and D2 mice were created by a DNA marker-assisted backcross breeding strategy and studied with respect to changes in MEST. A genomic interval delimited by marker D5Mit75 (proximal to the acromere) and D5Mit403 (distal to the acromere) was introgressed for 10 generations. A set of chromosome 5 congenic strains produced by an independent laboratory was also studied. Comparison of MEST between congenic and control (parental genetic background) mice indicates that genes influencing this trait were captured in all strains. Thus, mice from strains having D2 alleles from chromosome 5 on a B6 genetic background exhibit significantly lower MEST compared with control littermates, whereas congenic mice harboring B6 chromosome 5 alleles on a D2 genetic background exhibit significantly higher MEST compared with control littermates. Combining data from all congenic strains, we conclude that the gene(s) underlying the chromosome 5 QTL for MEST resides in the interval between D5Mit108 (26 cM) and D5Mit278 (61 cM). Generation of interval-specific congenic strains from the primary congenic strains described here may be used to achieve high-resolution mapping of the chromosome 5 gene(s) that contributes to the large difference in seizure susceptibility between B6 and D2 mice.
- C57 and DBA mice
- maximal electroshock seizure threshold
previous studies have documented robust differences in susceptibility to experimental seizures between C57BL/6J mice, which are relatively seizure resistant, and DBA/2J mice, which are relatively seizure susceptible (5, 9, 20). Quantitative trait locus (QTL) mapping has been used to dissect genetic influences on differences in seizure-related traits between B6 and D2 mice, and a number of chromosomal regions have been identified as containing seizure-related genes using this strategy. Among the regions of the genome documented to harbor such QTL, distal chromosome 1 and midchromosome 4 may include genes that are fundamental for determining neuronal excitability, an hypothesis based on the fact that QTL in these regions were detected by multiple and diverse paradigms to elicit seizures (2, 3, 10, 11, 13). In addition to the convergence of data on chromosomes 1 and 4, midchromosome 5 has also yielded evidence for seizure QTL in multiple studies. Paradigms utilizing measures of seizure susceptibility based upon responses of hybrid populations (B6 × D2) of mice to kainic acid (KA) (13), pentylenetetrazol (PTZ) (11), and electrical stimulation (10) have all led to the detection of genetic effects originating from chromosome 5 in the interval between 25 and 55 cM.
Refinement of QTL to the level of single causative gene variants remains difficult (15); however, the task is facilitated by the use of congenic strains (1). Accordingly, congenic strains have proven to be valuable resources for reducing the critical genomic intervals containing the chromosome 1 and chromosome 4 seizure QTL in the B6-D2 model (6, 8). Extending the results of congenic strain studies, we have provisionally identified two genes have been from these QTL, Mpdz on midchromosome 4 (24) and Kcnj10 on distal chromosome 1 (7, 8). The map position of a chromosome 11 QTL for seizures induced by withdrawal from central nervous system depressant drugs was also confirmed and refined recently with congenic strains (17). Overall, the utility of approaches involving congenic strains for the analysis of QTL is well documented.
In keeping with proven strategies for analyzing seizure QTL in the B6-D2 model, we describe here the generation of a reciprocal pair of chromosome 5 congenic strains that is based on previous genome scan data. We also describe results obtained with related congenic strains created by an independent laboratory. Overall, these studies confirm the presence of a QTL on chromosome 5 that influences maximal electroshock seizure threshold (MEST) and define its map position between markers D5Mit108 (26 cM) and D5Mit278 (61 cM). Since the first seizure-related QTL to be mapped to this interval was termed Szs11 (13), we refer to congenic strains created to study the interval as B6.D2-Szs11 and D2.B6-Szs11.
MATERIALS AND METHODS
Mice used in these studies were produced and/or maintained in-house at the Department of Veteran's Affairs Medical Center (DVAMC) in Coatesville, PA. Founder mice for inbred C57BL/6J (B6) and DBA/2J (D2) colonies used to generate reciprocal congenic strains were obtained from the Jackson Laboratory (Bar Harbor, ME). Additionally, mice from two chromosome 5 genome-tagged congenic strains (B6.D2-Chr5-P, B6.D2-Chr5-M) were a gift from Dr. Richard Davis at the University of California at Los Angeles (UCLA) (18). All mice are maintained on a 14 h/10 h light-dark schedule and are provided food and water ad libitum. Standard breeding practice in the facility involves housing one male mouse with two female mice. Females showing vaginal plugs are given their own cages where litters are delivered and nursed. At 3 wk of age, pups are weaned and tail-clipped to extract DNA for genotyping purposes. Mice are group-housed by sex and are used for breeding as early as 5 wk of age or for seizure testing beginning at 8 wk of age. All studies were approved by the Institutional Animal Care and Use Committees governing participating laboratories.
Congenic strain production and maintenance.
We created congenic strains in a reciprocal fashion (14) starting with an initial intercross (B6 × D2) followed by an F1 backcross to both parental strains. N1 generation breeding was conducted such that both F1 males and females were mated to parental strains. We genotyped backcross progeny and selected mice for subsequent rounds of breeding based upon heterozygosity at five DNA markers (see below). A total of 10 backcross generations was produced for each of the pair of reciprocal strains over the course of ∼3 yr. N10 mice heterozygous for introgressed alleles were crossed to produce an N10F1 generation. Males and females homozygous for introgressed alleles were crossed to establish the new strains, which are perpetuated by brother-sister mating. To generate congenic mice for seizure testing, mice homozygous for introgressed alleles were crossed with inbred mice of the background strain. Progeny heterozygous for markers across the introgressed region were intercrossed, and their offspring were tested blindly with respect to genotype. Phenotypes and genotypes were correlated subsequently. An attempt was made to study equal numbers of males and females. Data on mice from generations N10F6 through N10F8 are reported in this article. Mice carrying appropriate alleles were also intercrossed for preliminary seizure testing at generations N4 and N7; however, those data are not presented.
Introgression intervals and DNA marker genotyping.
A panel of five microsatellite DNA markers defining a 36-cM region of introgression was used for generation of B6.D2-Szs11 and D2.B6-Szs11 strains. The panel consisted of the following markers: D5Mit75 (20 cM), D5Mit108 (26 cM), D5Mit15 (39 cM), D5Mit398 (45 cM), and D5Mit403 (56 cM) (Fig 1). No genotyping was done for analysis of background DNA markers (either flanking the region defined by the markers used for introgression or elsewhere in the genome) during creation of the strains; however, additional markers were analyzed to better define the breakpoints of introgression after the strains were generated (N10 generation). Genotyping was also conducted to better define the breakpoints of the introgressed intervals in the chromosome 5 genome-tagged strains, which were created by an independent laboratory using a similar microsatellite DNA marker-assisted strategy (18).
MEST was determined as described previously (8–10, 12) using a constant current electroshock unit (model 7801; Ugo Basile, Varese, Italy). Mice were tested with a single electric shock delivered via ear-clip electrodes once per day beginning at age 8–9 wk. Initially, current level was set at 20 mA, and it was increased by 2 mA with each successive daily trial until a maximal seizure was observed. Other parameters of the stimulus were held constant (60 Hz, 0.4 ms pulse width, 0.2 s duration). Seizures were elicited at all current intensities utilized: lower intensities produced facial and forelimb clonus, whereas higher intensities produced generalized seizures. The sequence of responses that characterized a trial in which a maximal seizure was observed is as follows: tonic forelimb flexion, tonic hindlimb flexion, tonic hindlimb extension followed sometimes by hindlimb clonus. Mice were euthanized by cervical dislocation under CO2 anesthesia immediately after a trial in which a maximal seizure was elicited.
All phenotypic data including strain, age, sex, weight, and MEST are maintained in Excel spreadsheet format. Group averages for MEST were compared by a two-way ANOVA with strain and sex as independent variables. Post hoc analyses to examine specific group relationships were conducted by the Newman-Keuls test. All statistical analyses were performed using the Truepistat software program (Richardson, TX).
Chromosome 5 regions of introgression for all congenic strains studied are shown in Fig. 1. Reciprocal Szs11 congenic strains harbor introgressed intervals of different sizes. Strain B6.D2-Szs11 is defined by a homozygous region of D2 alleles from D5Mit75 (20 cM) to D5Mit403 (61 cM). Relatively small uncharacterized regions of the genome in this congenic strain (with respect to parental strain of origin) flank the introgressed interval both proximal and distal to the acromere. On the proximal side, the uncharacterized region is ∼4 cM in size and lies between markers D5Mit294 (8 cM) and D5Mit251 (12 cM). On the distal side, the uncharacterized region is ∼4 cM in size and lies between markers D5Mit278 (61 cM) and D5Mit425 (65 cM). Strain D2.B6-Szs11 is characterized by a large “passenger” (linked) fragment of the genome from B6 at the end of the region of introgression proximal to the acromere relative to its partner reciprocal congenic strain described above. Thus, alleles from B6 are fixed nearly all the way to the acromere in congenic strain D2.B6-Szs11 with heterozygosity detected at the most acromeric marker examined (D5Mit345), indicating that a chromosomal fragment of >30 Mb in size remains linked to the selected interval on its 5′-flank. The end of the region of introgression distal to the acromere extends to marker D5Mit403 (56 cM) with a partially uncharacterized region between D5Mit403 (56 cM) and D5Mit425 (65 cM) including heterozygosity at D5Mit278 (61 cM). We also analyzed markers in the “genome-tagged” congenic strains obtained from UCLA (18) to better characterize introgression breakpoints. Genomic organization of chromosome 5 in strain B6.D2-Chr5-P reveals that D2 alleles are fixed from marker D5Mit346 (1 cM) to marker D5Mit403 (56 cM). Somewhat surprisingly, background strain B6 alleles were detected at two markers proximal to the acromere (D5Mit345 and D5Mit146), suggesting that recombination was suppressed in this region during creation of the strain. The flanking interval distal to the acromere is uncharacterized between D5Mit403 (56 cM) and D5Mit278 (61 cM). Analysis of genome-tagged strain B6.D2-Chr5-M showed D2 alleles fixed from D5Mit183 (29 cM) to D5Mit98 (78 cM). There is an uncharacterized 3-cM genomic interval at the end of introgression proximal to the acromere between D5Mit108 (26 cM) and D5Mit183 (29 cM) and an uncharacterized 4-cM interval at the end distal to the acromere between D5Mit98 (78 cM) and D5Mit222 (82 cM).
Seizure testing data for chromosome 5 congenic strains with a B6 genetic background are depicted in Fig. 2. Two-way ANOVA documents statistically significant effects of strain (P < 2 × 10−5, F = 9.05) and sex (P < 1 × 10−8, F = 79.3). The interaction of these effects was not statistically significant. Posthoc analysis (Newman-Keuls test) documents significantly lower MEST (means ± SD) in male mice from all three congenic strains (58.0 ± 5.4 mA, 59.8 ± 5.9 mA, and 59.9 ± 6.5 mA, for B6.D2-Chr5-P, B6.D2-Chr5-M, and B6.D2-Szs11, respectively) compared with control mice (65.3 ± 7.4 mA). Analysis of data from female mice indicates that only congenic strain B6.D2-Chr5-P (48.2 ± 5.9 mA) is significantly different from control (55.4 ± 6.9 mA). Between-group P values and numbers of mice tested are shown in Fig. 2.
Seizure testing data for the D2.B6-Szs11 congenic strain are depicted in Fig. 3. Two-way ANOVA documents statistically significant effects of strain (P < 1 × 10−8, F = 74.28) and sex (P < 1 × 10−5, F = 23.8), as well as a statistically significant interaction effect (P < 0.03, F = 5.0). Post hoc analysis documents significantly higher MEST (means ± SD) in congenic mice of both sexes compared with control (congenic males: 29.7 ± 2.8 mA, control males: 25.3 ± 2.3 mA; congenic females: 28.0 ± 3.2, control females: 21.1 ± 1.3). Between-group P values and numbers of mice tested are shown in Fig. 3.
Results of the present study demonstrate that transfer of a segment of midchromosome 5 between B6 and D2 mice through the creation of congenic strains has a significant effect on the electrical threshold for maximal seizures. Similar findings were obtained for three independent congenic strains in which D2 alleles were transferred to a B6 genetic background and for one reciprocal congenic strain having B6 alleles on a D2 background. The seizure trait studied, MEST, is a quantitative measure of seizure susceptibility that exhibits low variability within inbred strains of mice (9, 12). Methodological differences exist between laboratories in the conduct of the test; however, results of inbred strain surveys have been comparable with regard to the rank order of strains, particularly with respect to B6 and D2 mice (9, 12, 16), suggesting that the test is reliable and resistant to large effects from the environment.
The impetus for the present work derives from previous QTL mapping studies involving experimental seizure paradigms in B6 and D2 mice and hybrid offspring populations. A seizure severity QTL of modest effect was first detected on chromosome 5 at marker D5Mit11 (26.0 cM) in an F2 intercross population using a paradigm involving systemic administration of KA and was termed Szs11 (13). Interestingly, this locus did not discriminate between those F2 mice that did and did not experience a KA-induced behavioral seizure, but rather it was associated with the severity of response among those mice that did experience a behavioral seizure (13). Also of note was the fact that the high-severity allele at the Szs11 locus was derived from the seizure-resistant B6 strain and was inherited in a recessive mode (13). A seizure QTL of modest effect was next detected on chromosome 5 between markers D5Mit294 (8.0 cM) and D5Mit398 (45.0 cM) in a study involving behavioral seizure response of B6 × D2 F2 mice to systemic administration of PTZ (11). Alleles in this interval distinguished between those F2 mice that exhibited a maximal seizure in response to PTZ and those that did not (11). Consistent with the characteristics of Szs11, the allele associated with the more severe seizure response was derived from the B6 strain and was inherited in a recessive mode (11). In a third QTL study of F2 mice from a B6 and D2 intercross, MEST was used as the mapping trait, and a locus of moderate phenotypic effect was again detected on chromosome 5 (10). The MEST QTL mapped to the interval between D5Mit294 (8.0 cM) and D5Mit15 (39.0 cM), similar to the previous studies using KA and PTZ; however, the allele associated with low MEST (high seizure susceptibility) was derived from the D2 strain (10). The marker panel used to create the reciprocal congenic strains described in the present study defined an approximate two-logarithm of odds support interval based on the MEST results (10). The genome-tagged strains were used to provide independent confirmation of the effect and to narrow of the interval based on their overlapping coverage. Thus, based on the overlap between donated segments in all of the congenic strains described in this study, the gene(s) underlying the chromosome 5 MEST QTL may be localized to the 35-cM interval defined by markers D5Mit108 (26 cM) and D5Mit278 (61 cM).
Although typical for single-pass QTL mapping experiments, the critical genomic interval defined by the strains used in the present study is large and requires refinement prior to intensive analysis of candidate genes and their strain-specific variation. Nonetheless, preliminary examination of this region identifies many interesting candidate genes, the most compelling being the cluster of GABA-A receptor subunit genes located at 40.0 cM. This cluster includes Gabra2 (GABA-A receptor alpha 2 subunit, 71240192–71374984 bp), Gabra4 (GABA-A receptor alpha 4 subunit, 71848869–71937443 bp), Gabrb1 (GABA-A receptor beta 1 subunit, 71979151–72416379), and Gabrg1 (GABA-A receptor gamma 1 subunit, 71030182–71121752). Despite the large number of other genes in the interval, the well-established role for GABA receptors in susceptibility to seizures and epilepsy confers special relevance upon this group. Examination of single nucleotide polymorphism databases indicates that there are no nonsynonymous sequence variations in transcripts for any of the GABA-A receptor subunit genes between B6 and D2 strains of mice (http://www.ensembl.org/Mus_musculus/index.html). This is consistent with previous studies that reported no strain differences in brain mRNA levels and no alternative transcripts for Gabrb1 (19) and Gabrg1 (25). Similarly, there are no reported cDNA sequence differences or differences in message levels for Gabra4 between B6 and D2 (4). An alternative transcript has been detected for Gabra1 in mice (23); however, it is not known if there is a strain difference in expression. A caveat of the studies cited above is that B6 and D2 mice were compared only at the age of 3 wk and not into adulthood when differences in seizure susceptibility have been documented to persist. Furthermore, levels of transcripts were measured only in whole brain, which can mask localized expression differences. Nonetheless, prior studies on potential strain-specific variations between B6 and D2 mice do not provide support for the possibility that a GABA-related gene underlies the effect of the chromosome 5 seizure QTL, and thus other genes in the interval deserve consideration. An expanded list of potential candidate genes that map between markers D5Mit108 and D5Mit278 includes Slc34a2 (solute carrier family 34 member 2, 31.0 cM, 53337605–53359912 bp), Cckar (cholecystokinin A receptor, 34.0 cM, 53986735–53996350 bp), Txk (TXK tyrosine kinase, 40.0 72975117–73015737 bp), Chrna9 (cholinergic receptor, nicotinic, alpha polypeptide 9, 41.0 cM, 66214056–66256461 bp), Cnga1 (cyclic nucleotide gated channel alpha 1, 41.0 cM, 72883250–72916981 bp), Gnrhr (gonadotropin-releasing hormone receptor, 44.0 cM, 87256547–87272431 bp), and Adrbk2 (adrenergic receptor kinase, beta 2, 60.0 cM, 113150781–113255817 bp).
The ability to refine the map position of genes that influence complex (quantitative) traits following QTL analysis is dependent upon reliable and precise phenotyping as well as upon minimizing the influence of uncontrolled environmental variables. The magnitude of the effect of a QTL is also important in relation to the amount of environmentally induced phenotypic variability, and together these factors determine the “signal-to-noise” of the trait. Unfavorable values for either factor, either a weak QTL effect or strong uncontrollable environmental effects on the phenotype, hinder, or even obviate, the ultimate identification of causative gene polymorphisms. For studies of seizure-related traits, the use of electroshock seizure threshold as an endpoint is reproducible and relatively resistant to uncontrollable effects from the environment such as operator idiosyncrasies. Thus, traits based upon electroshock seizure thresholds in mice show relatively low variability as measured in genetically homogeneous populations including inbred strains (5, 9, 16), congenic strains (8), and transgenic lines (7).
QTL analysis of MEST in B6 and D2 mice suggests the existence of phenotypically relevant genes on seven different chromosomes (10). The locus of strongest effect maps to distal chromosome 1 and has been confirmed subsequently in congenic strains (8) and further analyzed with transgenic lines to characterize causative gene variation (7). After the chromosome 1 QTL, the locus having the next strongest effect on the difference in MEST between B6 and D2 mice maps to the Szs11 region of chromosome 5 and accounts for ∼16% of the total trait variance (10). The phenotypic effect of the chromosome 5 QTL, as captured in the congenic strains described here, ∼10% difference from mean control values, is consistent with the previous estimation of QTL strength. If the chromosome 5 QTL confirmed in the present study with congenic strains is defined by a single gene, the magnitude of the effect should be great enough to permit high-resolution mapping with interval-specific (“subcongenic”) strains. It is important to note that some seizure QTL captured initially in congenic strains have been lost when the critical interval is dissected into smaller segments (21, 22), suggesting that the QTL was mediated by the influence of polymorphism in multiple genes in the original interval. Although such results may be common, previous studies with seizure-related QTL on chromosomes 1, 4, and 11 in B6 and D2 mice showed that phenotypic effects are retained in interval-specific congenic strains harboring genomic intervals of <10 million base pairs (6, 8, 17). Thus, one rational approach to identify the specific genetic variation underlying the seizure-related QTL on chromosome 5 begins with the derivation of interval-specific congenic strains from strains reported here.
In summary, we have created a reciprocal pair of strains congenic between B6 and D2 and have used them, in conjunction with congenic strains created by an independent laboratory, to confirm the presence of a gene(s) on chromosome 5 that influences MEST. Based on overlapping regions of D2 genomic introgression among three strains with a B6 genetic background, we conclude that the gene(s) responsible for the observed phenotypic effect is located in the 35-cM interval flanked by markers D5Mit108 and D5Mit278. While this genomic region is still relatively large, previous studies of a chromosome 1 seizure QTL demonstrate that systematic evaluation of such intervals is possible and can provide high-resolution genetic maps to facilitate identification of specific gene influences (7, 8). Thus, creation of interval-specific congenic strains from the strains described in the present study may allow refinement of the QTL interval on chromosome 5 that influences MEST and lead ultimately to the discovery of causative gene variation at this locus.
This work was supported by National Institute of Neurological Disorders and Stroke Grant NS-40554 (T. N. Ferraro), the University of Pennsylvania Center for Neurobiology and Behavior, and the DVAMC, Coatesville, PA.
We thank James Martin and Girvin McAllister for excellent technical assistance and Dr. Glenn Doyle for critical reading of the manuscript. We also thank Dr. Richard Davis and the staff at the UCLA animal facility for providing strains of genome-tagged mice.
Article published online before print. See web site for date of publication (http://physiolgenomics.physiology.org).
Address for reprint requests and other correspondence: T. N. Ferraro, Ctr. for Neurobiology and Behavior, Dept. of Psychiatry, Univ. of Pennsylvania, 125 S. 31st St., Rm. 2209, Philadelphia, PA 19104-3404 (e-mail:).
- Copyright © 2007 the American Physiological Society