Background

One of the clinical manifestations of cystic fibrosis (CF) not adequately predicted by the cystic fibrosis transmembrane conductance regulator (CFTR) genotype is the intestinal phenotype meconium ileus [1]. Cftr-knockout mouse models show intestinal plugging at birth, resembling the clinical phenotype, and the strain-dependent incidence of this lethal complication has been used to map a genetic modifier of CF intestinal disease, named cystic fibrosis modifier gene 1 [2].

We have mapped potential modifiers of the CF intestinal phenotype [3] in C57BL/6J (B6) × BALB F2 CF mice, using an approach based on genotypic differences between surviving CF mice and mice succumbing to intestinal disease. In that study, the genetic influence on CF mouse survival was demonstrated by the fact that more F2 CF mice were alive at weaning (54% of the expected number of CF mice lived to this time point) than were congenic B6 (22%) or BALB (13%) CF mice. We identified an X chromosome locus which may contribute to this survival difference, in that non-Mendelian inheritance favoring B6 alleles in the surviving CF mice and BALB alleles in mice of a control population, was observed. Two loci of transmission ratio distortion, independent of Cftr genotype, on chromosomes 3 & 5, were also detected in the female mice of this cross. Based on birth records the CF mice not surviving to weaning were presumed to have succumbed during both the prenatal and postnatal periods. We therefore concluded the phenotype of survival to weaning to be from either death due to intestinal distress in the perinatal period as in [2, 4], or from death in utero, which may not be due to intestinal distress specifically.

As in our previous study [3], the putative loci of transmission ratio distortion were identified in 96 control mice, in this report we typed mice of a larger B6 × BALB F2 cohort (n = 1208) to investigate the CF dependence of the potential X chromosome modifier locus, and to confirm or refute the non-Mendelian inheritance of genotypes at loci on chromosomes 3 & 5. In addition, to narrow the time frame of the influence of the potential cystic fibrosis modifier we evaluated embryonic mice of the same cross.

Results

To investigate whether previously identified regions of transmission ratio distortion (TRD) existed in a larger cohort of B6 × BALB F2 mice, we bred 1208 mice and genotyped the population for markers on chromosomes 3, 5, and X. As shown in Table 1, two of the loci of TRD persisted as the female mice of this cross had an excess of the homozygous BALB genotype at the tested markers on chromosomes 5 and X while the distortion was not evident in male mice nor at the chromosome 3 marker.

To determine whether the confirmed transmission ratio distortion of the chromosome 5 and X genotypes was dependent on Cftr genotype, we next typed the three-week old F2 mice for CF status (CF or non CF). Of the 1208 mice evaluated 137 were identified to be CF and this number is fewer than would be expected for an autosomal recessive trait (25% of the population sample of 1208 mice = 302 mice expected). We therefore conclude that pre-weaning lethality of CF mice occurred. The DXMit16 genotypes for the subset of our population which were identified to be CF mice were evaluated, and as shown in Table 1, X chromosome genotypes were not distorted from Mendelian expectations in this sample of F2 CF female and male mice. We next compared the set of DXMit16 genotypes of CF mice to that of non-CF littermates and observed a lack of genotype distortion in female CF mice. This makes the CF dataset distinct from that of female non-CF mice for DXMit16 genotypes (p = 8.7 × 10-6), due to the non-Mendelian inheritance of genotypes for this region in the latter group. To further investigate the CF dependence of the putative distortion, we grouped the non-CF female F2 mice by Cftr +/+ and +/- genotype and found the DXMit16 genotype to be independent of Cftr genotype (p = 0.12) in control mice. The genotypes at marker D5Mit239 of B6 × BALB F2 CF female mice were also not distorted from Mendelian expectations, as shown in Table 1, although an excess of BALB alleles may be present in this population (p = 0.06). The set of D5Mit239 genotypes of the female F2 CF mice is not different from those of the non-CF mice (p = 0.26), therefore the distortion at this marker is independent of Cftr genotype which is in keeping with our first report.

Table 1 Genotype frequency in 3 week old B6 × BALB F2 CF and non-CF mice
Full size table

The distortion of genotypes at DXMit16 in female mice of the current study is consistent with what we observed in a previous cross of the B6 and BALB strains [3], as shown in Table 2, while the genotypes of the additional F2 CF males do not support the trend identified in the previous dataset. In the prior study, F2 CF male mice were found to have an excess of B6 alleles at DXMit16, and, as such distortion is not evident in this dataset, the original finding may have been due to insufficient sample size. The previous study also differs from the current work in that, previously, BALB Cftr +/- mice were crossed with B6 Cftr +/- mice to generate the F1 mice while in the current study the BALB parents were Cftr +/+. This change results in the CF mutation being solely inherited from the B6 chromosome in the present study. The consequence of this change would be the non-Mendelian inheritance of genotypes (favouring B6 alleles) in the chromosome 6 region of Cftr in the CF mice of this cross. As the change in breeding strategy would not affect the inheritance of alleles on other chromosomes, we combined the datasets of the two investigations for further analysis. Comparing the genotypes of the total non-CF mice of both studies, to the combined genotypes of CF mice, reveals a significant difference in DXMit16 genotype ratios, by CF status, in both male (p = 9.0 × 10-4) and female mice (p = 3.4 × 10-16). From these data we conclude the non-Mendelian inheritance of the X chromosome marker in F2 mice of a B6 × BALB Cftr +/- intercross to depend on CF status.

Table 2 DXMit16 Genotype frequency in 3 week old CF and non-CF mice of replicate studies
Full size table

Our analyses also confirmed the significant incidence of pre-weaning lethality in B6 × BALB F2 CF mice (137/302 = 45% of the expected number of CF mice were identified at weaning). To narrow the time frame in which this lethality occurred, we assayed day E15.5 embryos of nineteen litters and determined CF embryo survival at this age. The Cftr genotypes of the embryos were 40:104:42 (Cftr +/+: +/-: -/-) which is consistent with Mendelian expectations (p = 0.27) and which also indicates that the lethality of the CF mice of this cross occurred between embryonic day 15.5 and the age of three weeks. Finally, to further investigate the potential transmission distortion of markers D5Mit239 and DXMit16 in the mice of this cross, the embryos were typed for these loci and for a Y chromosome marker to indicate their sex. At marker D5Mit239 the inheritance of B6/BALB genotypes was consistent with Mendelian expectations in both female (p = 0.41) and male (p = 0.35) embryos, data not shown. For marker DXMit16 the transmission ratio distortion identified in female mice at weaning was evident in the embryos, as the genotypes of the females were (35 BALB/BALB: 44 BALB/B6: 13 B6/B6, p = 0.0048, compared to an expected ratio of 1:2:1) and of males were (51 BALB: 37 B6, p = 0.14, compared to an expected ratio of 1:1). The genotype ratios for the subset of embryos identified as CF were consistent with those reported for the larger group (9:14:1 in CF embryonic females and 11:4 in males).

Discussion

Herein we confirm the existence of transmission ratio distortion at specific loci on chromosomes 5 and X in the female mice of a B6 × BALB Cftr +/- intercross. In this replicate study we also show the non-Mendelian inheritance of X chromosome genotypes to influence the survival to weaning of Cftr +/+ and Cftr +/- mice, but not Cftr -/- mice, which suggests the mechanism leading to TRD in this cross may involve functional Cftr.

With the completion of this second study, our observation of X chromosome transmission ratio distortion as influenced by the CF status of the mice is now based on genotypes of 395 CF mice and 1184 non-CF mice, compared to the 96 non-CF mice evaluated in the original report. These numbers of mice are in line with those used to identify X chromosome transmission ratio distortion, or non-Mendelian inheritance, in offspring of different mouse crosses [5, 6] and, further, the marker of distortion we report on maps to one of these previously defined intervals. We also provide evidence for a second locus of non-Mendelian inheritance, in this cross, on chromosome 5 which affected females only, and we illustrate the importance of sufficient sampling in transmission ratio distortion studies as no distortion of the locus on chromosome 3, now tested in 367 females compared with 45 in our prior study [3] was evident.

The previous studies of transmission ratio distortion have attributed distortion to prenatal lethality associated with particular allelic combinations but the causative genes producing the non-Mendelian inheritance have not been uncovered. Our embryonic data supports a prenatal effect as the distortion of DXMit16 genotypes was evident at day E15.5. This effect did not, however, influence the survival of CF embryos to E15.5, which was consistent with Mendelian expectations. As the survival of CF mice to weaning was clearly below expectations (45% of the expected number of one fourth of the litters), and as the strong distortion of DXMit16 genotypes, evident in control mice, was absent in CF mice, the mechanisms leading to these observations may be linked and likely act in mice between the ages of E15.5 and 3 weeks. Further study is needed to identify the X-linked modifier which may enhance the survival of CF mice in the perinatal period.

The survival advantage conferred to control mice by inheritance of the X chromosome homozygous BALB genotype was not evident in CF mice of replicate populations thus, the mechanism leading to transmission ratio distortion in the mice of this cross may involve Cftr protein function. One potential explanation for this may be in Cftr expression as Larson et al. [7] have shown Cftr gene therapy in utero to lead to death of Cftr +/+ embryos, while Cftr -/- embryos were spared. The F2 non-CF mice that inherited B6/BALB heterozygous genotypes on chromosome X may have expressed a level of Cftr that is incompatible with survival to weaning. X chromosome candidate genes for such an interaction can be taken from protein complexes reported to regulate Cftr, although its regulation is not completely understood [8]. This list includes signalling molecules, kinases, transport proteins, PDZ-domain-containing proteins, myosin motors, Rab GTPases, and SNAREs [8], as does the distorted region of the X chromosome. For example, transcription factors, such as Aristaless-related homeobox gene, Arx, which has a crucial role in pancreas development [9], and Dax-1 which is involved in epithelial differentiation [10], map to the distorted region on X and could influence Cftr-affected pathways. One such pathway, for which mutations in both Dax-1 [11] and Cftr [12] have been implicated, is male infertility. Elucidation of an X chromosome factor which influences Cftr activity, if such exists, could be an important advance in understanding cystic fibrosis disease development.

Conclusion

We have provided evidence for two loci of transmission ratio distortion in the female mice of a B6 × BALB F2 intercross. One locus, on the X chromosome, was shown to have non-Mendelian inheritance in the non-CF mice of this cross, but not to be distorted in the in the subset of mice identified to be CF. The CF status of the mouse may therefore be sparing the effect produced by the distorted inheritance of the X chromosome. Such an interaction, if confirmed, could provide important information on Cftr regulation in the developing mouse.

Methods

Mice

The mice of the C57BL/6 Cftr+/tm1UNC (Cftr+/-) strain were provided by Dr. Danuta Radzioch of the Montreal General Hospital, and the BALBc/J (BALB) strain was purchased from the Jackson Laboratory. F2 CF mice were created from these strains through two generations of breeding. Specifically, the B6 Cftr+/- and BALB mice were crossed to create F1 mice and the resultant F1 Cftr+/- mice were intercrossed to create 1208 F2 non-CF (Cftr+/+ and Cftr+/-) and CF (Cftr-/-) mice which were identified by their Cftr genotypes at weaning (~3 weeks of age). All mice were housed in micro-isolator cages in a Specific Pathogen Free room and handled according to the standard husbandry of the Animal Resource Centre of McGill University. Animal use and handling was approved by the McGill University Committee for Animal Care and Use, which follows the guidelines and regulations of the Canadian Council on Animal Care.

Embryos

F1 Cftr +/- mice were intercrossed and 186 embryos were dissected from the uteri of 19 sacrificed females on day E15.5. Embryonic day 15.5 was chosen as individual embryos are of sufficient size for genotyping at this time, and as data from Bonner et al. [13] show the embryonic expression of Cftr, in the lungs, to have reached a plateau at this age. At sacrifice the amniotic sac from each intact embryo was removed and used as a source of genomic DNA.

Genotyping

DNA was prepared from the tail clips of the three week-old F2 mice and from the amniotic sacs of the F2 embryos. Polymorphic mouse genetic markers, described by Dietrich et al. [14], were amplified by PCR for the genotyping of three specific loci on chromosomes 3, 5 and X (see Table 1). The PCR products were separated by electrophoresis on 3% agarose gels and visualized with ethidium bromide staining. The sex of the embryos was determined by amplifying and detecting a portion of the Sry1 gene, a Y chromosome-specific marker, as in [15]. The Cftr genotypes of the three week-old F2 mice and of the E15.5 embryos were determined by performing a previously described PCR assay [16] using the isolated genomic DNA.

Inheritance of B6 and BALB alleles

Deviations in genotypic ratios compared to the expected Mendelian ratio of 1:2:1 (B6/B6:B6/BALB:BALB/BALB) and in allelic ratios compared to the expected 1:1 ratio (B6 alleles: BALB alleles), were assessed by χ2 analysis (as in [6]) using Microsoft Excel software.