Multifactorial genetics of exencephaly in SELH/Bc mice.

BACKGROUND: The SELH/Bc mouse strain has 10-30% exencephaly and is an animal model for human neural tube closure defects. This study examined the number of causative genes, their dominance relationships, and linkage map positions. METHODS: The SELH/Bc strain (S) was crossed to the normal LM/Bc strain (L) and frequencies of exencephaly were observed in the F(1), BC(1), and F(2) generations. 102 F(2) males were individually testcrossed by SELH/Bc. The extremes, the 10 highest and 10 zero exencephaly-producing F(2) sires, were typed for 109 SSLP marker loci in a genome screen. Next, the resultant five provisional chromosomal regions were tested for linkage in 31 F(2) exencephalic embryos. Finally, 12 males, SS or LL for the Chr 13 region on an LM/Bc background, were testcrossed by SELH/Bc. RESULTS: The exencephaly frequencies in the F(1) (0.3%), BC(1) (4.4%), and F(2) (3.7%), and the distribution of F(2) males' testcross values (0-15.5%), indicated that the high risk of exencephaly in SELH/Bc is due to the cumulative effect of two or three loci. Linkage studies indicated the location of semidominant exencephaly-risk genes on Chr 13 near D13Mit13 (P < 0.001), Chr 5 near D5Mit168 (P < 0.025), and possibly Chr 11 near D11Mit10 (P < 0.07). The gene on Chr 13, Exen1, and the strong role of other loci were confirmed by the congenic males. CONCLUSIONS: The high risk of exencephaly in SELH/Bc mice is caused by the cumulative effect of two to three semidominant genes. Candidate genes include Msx2, Madh5, Ptch, and Irx1 (Chr 13) and Actb and Rac1 (Chr 5).

Unravelling the complex genetics of cleft lip in the mouse model.

Nonsyndromic cleft lip in "A" strain mice and humans is genetically complex and is distinct from isolated cleft palate. Cleft lip embryos recovered in 2.4% of 1485 first backcross (BC1) segregants from a cross of A/WySnJ (24% cleft lip) and C57BL/6J (no cleft lip) in A/WySnJ mothers, and in testcrosses of 10 recombinant inbred (RI) strains (AXB/Pgn or BXA/Pgn), were used for gene mapping and for inference of genetic architecture. The A/WySnJ maternal genotype increased cleft lip risk in reciprocal crosses; the relevant genetic difference between AXB-6/Pgn (8%) and A/WySnJ (24%) is entirely maternal. A combination of new mapping panels (325 meioses), new markers, and a recombinant cleft lip embryo redefined the location of a recessive factor essential to cleft lip risk, clf1, and candidate genes Itgb3 and Crhr, to between D11Mit146/360 and D11Mit166/147. A screen of 54 YACs for 46 genes and SSLP loci located Wnt15, Wnt3, Crhr, Mtapt, Itgb3, Dlx3, and Dlx7 within the clf1 candidate region. The clf2 locus was newly mapped to Chromosome (Chr) 13 by a genome screen of BC1 segregants, and further defined to a 4-cM region between D13Mit13/54 and D13Mit231 by strain distribution patterns of cleft lip liability and markers in testcrossed RI strains. Specific combinations of marker genotypes associated with cleft lip risk indicated that high risk in A/WySnJ mice is caused by epistatic interaction between clf1 and clf2 in the context of a genetic maternal effect. Human homologs of clf1 and clf2 are expected to be on 17q and 5q/9q.

Gaping lids, gp, a mutation on centromeric chromosome 11 that causes defective eyelid development in mice.

In mammals, during fetal development, the eyelids grow and flatten over the eyes and temporarily fuse closed. Failure of this normal developmental process in mice leads to the defect, open-eyelids-at-birth. Nearly all newborns of the GP/Bc strain, homozygous for the spontaneous recessive mutation, gaping lids (gp), have bilateral open eyelids at birth, with essentially no fusion between the upper and lower eyelids. Histological sections and scanning electron microscopy of GP/Bc eyes during the normal period of eyelid growth and fusion indicate that gp/gp mutant fetuses have deficient upper and lower eyelids; surface periderm cells that appear to have some role in eyelid growth and fusion are present, but lack a normal "streaming pattern toward the fusion zone. No other defects due to the gaping lids mutation were detected. A genetic analysis based on outcrosses of GP/Bc to various linkage marker stocks and to CBA/J and ICR/Bc normal strains was done. Penetrance in F(2) segregants, but not in BC1 segregants, was usually significantly less than 100%, was strongly affected by the identity of the normal strain used, ranging from 44% to 92%, and indicated a potential complexity of modifiers. Forty-one affected F(2) and 120 BC(1) segregants from the outcross of GP/Bc to CBA/J, and 23 affected F(2) segregants from the outcross to ICR/Bc, were used to map gp to proximal Chr 11 between the centromere and D11Dal1 (Camk2b), an interval previously defined as less than 1 cM. Sets of whole F(2) litters from the crosses to CBA/J (n = 106) and ICR/Bc (n = 65) strains were typed for informative SSLPs near gp (D11Mit62 and D11Mit74, respectively) and demonstrated that the segregation ratios in the region are Mendelian. The known genes in the interval, Nf2 and Lif, do not seem to be obvious candidate genes for gp. An Egfr-null allele was used to confirm the previously reported map position of the potential candidate locus, Egfr, to a more distal interval, between D11Mit62/226 and D11Mit151, from which gp had been excluded. Tests for allelism showed that the Egfr mutation and the gp mutation complement each other, and therefore also indicate that they are at different gene loci. Open-eyelids-at-birth is associated with several mutations at other loci with variable penetrance owing to modifiers and in other more complex genetic liabilities in inbred strains, and the genetics of this trait is a model for other genetically complex developmental threshold traits. The gaping lids mutation identifies a previously unknown locus on proximal Chromosome (Chr) 11 that has a strong role in fetal eyelid growth.

Mouse models for neural tube closure defects.

Neural tube closure defects (NTDs), in particular anencephaly and spina bifida, are common human birth defects (1 in 1000), their genetics is complex and their risk is reduced by periconceptional maternal folic acid supplementation. There are > 60 mouse mutants and strains with NTDs, many reported within the past 2 years. Not only are NTD mutations at loci widely heterogeneous in function, but also most of the mutants demonstrate variable low penetrance and some show complex inheritance patterns (e.g. SELH/Bc, Abl / Arg, Mena / Profilin1 ). In most of these mouse models, the NTDs are exencephaly (equivalent to anencephaly) or spina bifida or both, reflecting failure of neural fold elevation in well defined, mechanistically distinct elevation zones. NTD risk is reduced in various models by different maternal nutrient supplements, including folic acid ( Pax3, Cart1, Cd mutants), inositol ( ct ) and methionine ( Axd ). Lack of de novo methylation in embryos ( Dnmt3b -null) leads to NTD risk, and we suggest a potential link between methylation and the observed female excess among cranial NTDs in several models. Some surprising NTD mutants ( Gadd45a, Terc, Trp53 ) suggest that genes with a basic mitotic function also have a function specific to neural fold elevation. The genes mutated in several mouse NTD models involve actin regulation ( Abl/Arg, Macs, Mena/Profilin1, Mlp, Shrm, Vcl ), support the postulated key role of actin in neural fold elevation, and may be a good candidate pathway to search for human NTD genes.

Whiskers amiss, a new vibrissae and hair mutation near the Krt1 cluster on mouse chromosome 11.

Whiskers amiss (wam) is a new spontaneous recessive mutation in the SELH/Bc strain of mice that causes a phenotype of askew, sometimes kinked or curled, breakable whiskers and disheveled-appearing body hair, apparently owing to disoriented guard hairs. Heterozygotes on three genetic backgrounds are indistinguishable from normal. Using informative SSLPs in the F2 generation after crosses to two normal strains, we have mapped wam to the region of the type I keratin cluster on Chromosome (Chr) 11, within an approximately 6-cM segment according to the current Mouse Genome Database (MGD) map position of flanking SSLPs. Although several other hair mutations also map to the Krt1 region (Re, Rim3, Bdai, Bsk), none has a hair and whisker phenotype similar to that of wam, and, because all are transmitted as dominants, interpretable complementation tests could not be done. Scabbing and tissue loss occur on the rims of the pinnae and tail tip in some aging wam homozygotes, suggesting that wam may be an animal model of a genetic ectodermal disorder. The SELH/Bc strain background appears to have an unusually high rate of spontaneous mutation; wam is the sixth mutation to be described.

Mini-review: toward understanding mechanisms of genetic neural tube defects in mice.

We review the data from studies of mouse mutants that lend insight to the mechanisms that lead to neural tube defects (NTDs). Most of the 50 single-gene mutations that cause neural tube defects (NTDs) in mice also cause severe embryonic-lethal syndromes, in which exencephaly is a nonspecific feature. In a few mutants (e.g., Trp53, Macs, Mlp or Sp), other defects may be present, but affected fetuses can survive to birth. Multifactorial genetic causes, as are present in the curly tail stock (15-20% spina bifida), or the SELH/Bc strain (15-20% exencephaly), lead to nonsyndromic NTDs. The mutations indicate that "spina bifida occulta," a dorsal gap in the vertebral arches over an intact neural tube, is usually genetically and developmentally unrelated to exencephaly or "spina bifida" (aperta). Almost all exencephaly or spina bifida aperta of genetic origin is caused by failure of neural fold elevation. The developmental mechanisms in genetic NTDs are considered in terms of distinct rostro-caudal zones along the neural folds that likely differ in mechanism of elevation. Failure of elevation leads to: split face (zone A), exencephaly (zone B), rachischisis (all of zone D), or spina bifida (caudal zone D). The developmental mechanisms leading to these genetic NTDs are heterogeneous, even within one zone. At the tissue level, the mutants show that the mechanism of failure of elevation can involve, e.g., (1) slow growth of adjacent tethered tissue (curly tail), (2) defective forebrain mesenchyme (Cart1 or twist), (3) defective basal lamina in surface ectoderm (Lama5), (4) excessive breadth of floorplate and notochord (Lp), (5) abnormal neuroepithelium (Apob, Sp, Tcfap2a), (6) morphological deformation of neural folds (jmj), (7) abnormal neuroepithelial and neural crest cell gap-junction communication (Gja1), or (8) incomplete compensation for a defective step in the elevation sequence (SELH/Bc). At the biochemical level, mutants suggest involvement of: (1) faulty regulation of apoptosis (Trp53 or p300), (2) premature differentiation (Hes1), (3) disruption of actin function (Macs or Mlp), (4) abnormal telomerase complex (Terc), or (5) faulty pyrimidine synthesis (Sp). The NTD preventative effect of maternal dietary supplementation is also heterogeneous, as demonstrated by: (1) methionine (Axd), (2) folic acid or thymidine (Sp), or (3) inositol (curly tail). The heterogeneity of mechanism of mouse NTDs suggests that human NTDs, including the common nonsyndromic anencephaly or spina bifida, may also reflect a variety of genetically caused defects in developmental mechanisms normally responsible for elevation of the neural folds.

Nonallelic noncomplementation models in mice: the first arch and lidgap-Gates mutations.

We tested for complementation between two Mendelian mutations in mice, Far (first arch) and lgGa (lidgap-Gates). Each of these mutations gives greater than 70% risk of the birth defect, open eyelids, in homozygotes and gives little or no risk in heterozygotes. Far and lgGa are known to not be alleles; Far maps to Chr 2 and lgGa maps to Chr 13. However the cross between +/Far (on the BALB/cGaBc strain) and lgGa/lgGa (on the LGG/Bc strain) gave 32% (48/149) of progeny affected with open eyelids at birth: 63% (45/71) of the double heterozygote, +/Far, +/lgGa, compared with 4% (3/78) of the +/+, +/lgGa progeny. That is, the complementation test suggests that Far and lgGa are alleles, whereas the mapping data show that they are not. We interpret the result of the Far by lgGa test as an example of nonallelic noncomplementation (or "false allelism") in mammals, and suggest that this phenomenon might be expected because open eyelids at birth involves a developmental threshold. Our data also show that both the embryonic and the maternal background genotypes strongly influence the risk of open eyelids in the Far by lgGa crosses. The risk to the double heterozygote (+/Far, +/lgGa) is highest (77%) with Far from the BALB/cGaBc rather than the ICR/Bc (0%) strain and in a BALB/cGaBc (77%) rather than an LGG/Bc (50%) dam in the reciprocal cross. This effect of genetic context on risk is also predicted by the threshold model. Based on our data on open eyelids at birth, we suggest that false allelism may be common in mammalian birth defects that result from failure to meet developmental thresholds, even when the "causal" mutations are Mendelian.

Genetic landmarks for defects in mouse neural tube closure.

Many mutations cause neural tube closure defects (NTDs, exencephaly or spina bifida) in mice and the gene loci are widely distributed in the mouse genome. This compilation summarizes the map position of 40 mouse NTD mutations and the corresponding human linkage homology of each. It includes the nature of the gene product where known, and whether the NTD is part of a syndrome involving other developmental systems. Also listed are the several mouse strains known to have genetic susceptibility to exencephaly, with multifactorial genetic cause in at least one case. The purposes of this mouse NTD compilation are to enable recognition of patterns in genetic causes of NTDs, of molecular pathways essential for closure of specific regions of the mammalian neural tube, and of candidate regions for mapping loci contributing to human multifactorial NTDs.

Effect of multifactorial genetic liability to exencephaly on the teratogenic effect of valproic acid in mice.

The present study shows that the multifactorial genetic liability to spontaneous exencephaly in the SELH/Bc mouse strain (10-20% of embryos) also confers an elevated risk of exencephaly induced by valproic acid. Treatment of pregnant dams (600 mg/kg sodium valproate in distilled water, i.p.) during the critical period on day 8 (D8) of gestation resulted in D14 exencephaly frequencies of 69% in SELH/Bc contrasted with 39% in each of the SWV/Bc and ICR/Bc strains. Analysis of these data under the assumptions of the threshold model indicated that the valproic acid-induced-shift in mean liability was similar for all three strains, and therefore the effects of genotype and teratogen were additive, not synergistic. A similar exencephaly response pattern for the same three strains was observed previously with retinoic acid [Tom et al. (1991) Teratology 43:27-40], a pattern that, combined with the data of Finnell et al. [(1988) Teratology 38:313-320], argues that strain differences in exencephaly response are not due to strain differences in teratogen metabolism. SWV/Bc and ICR/Bc embryos differ in location of the Closure 2 initiation site of cranial neural tube closure [Juriloff et al. (1991) Teratology 44:225-233], but the observation that they do not differ in risk of exencephaly produced by either valproic acid or retinoic acid contradicts the hypothesis that this particular morphological difference underlies strain differences in exencephaly risk. The high exencephaly response of SELH/Bc to two teratogens predicts that human conceptuses with a genetically determined elevated risk for neural tube defects could be easily tipped into high risk by mild teratogens.

Exencephaly and cleft cerebellum in SELH/Bc mouse embryos are alternative developmental consequences of the same underlying genetic defect.

SELH/Bc inbred mice have ataxia in 5-10% of young adults and exencephaly in 10-20% of newborns. SELH/Bc mice also have a high rate of spontaneous mutation and therefore it could not be assumed that these two abnormalities share the same genetic cause. Previously, we have shown that the liability to exencephaly in SELH/Bc mice is multifactorial, involving two to three loci, and that all the ataxics have a midline cleft cerebellum. The purpose of the present study was to resolve the genetic relationship between liability to exencephaly and liability to cleft cerebellum. We tested whether these traits were transmitted together by segregating F2 males; cotransmission would indicate that both traits are probably caused by the same genes. Approximately 100 embryos from each of 25 F2 sires from a cross between SELH/Bc and the normal LM/Bc strain were scored for exencephaly and the non-exencephalic embryos were scored for cleft cerebellum. The range of exencephaly production by these 25 F2 sires was 0% to 16%; the sires had been selected to represent the extremes of the range of exencephaly production. We found that the 10 sires that produced no exencephaly also produced no cleft cerebellum and 12 of the 15 sires that produced some exencephaly also produced some cleft cerebellum. This indicated strongly that the two traits are transmitted together (Fisher's exact test, P < 0.0002). Furthermore, within exencephaly-producing sires, the specific frequencies of the two traits were significantly positively correlated (Spearman rs = 0.58; P < 0.05), indicating that the same multifactorial risk factors influence both traits. All SELH/Bc embryos omit one normal initiation site of cranial neural tube closure, Closure 2. In a previous study, absence of the Closure 2 initiation site of cranial neural tube closure has been shown to be genetically correlated with liability to exencephaly. In the second part of the present study, the same Closure 2 data from eight of the F2 sires were observed to be significantly positively correlated with liability to cleft cerebellum (Spearman rs = 0.83; P < 0.05). The results of this genetic approach have supported the hypothesis, based on observation of embryos, that one basic multifactorial genetic defect in SELH mice leads to an abnormal cranial neural tube closure mechanism, to exencephaly to cleft cerebellum, and to ataxia.

The lidgap-Gates (lgGa) mutation for open eyelids at birth maps to mouse chromosome 13.

Complex nonadditive interactions between specific alleles at multiple loci may underlie many so-called multifactorial threshold birth defects. The open-eyelids-at-birth defect in mice is a good model for these defects, and an understanding of its genetic complexity begins with mapping the participating loci. The open-eyelids defect can be part of a syndrome or can occur with no other obvious phenotypic effects. Of the latter nonsyndromic forms, the lidgap series includes four extant mutations that are considered to be alleles based on complementation tests. All show genetic complexity in segregation ratios. None has been mapped previously. On the basis of a strategy of mapping the mutation with the simplest inheritance pattern first, we generated an extensive exclusion map for lidgap-Gates, lgGa, using morphological and protein polymorphisms. We then screened the non-excluded regions in a congenic strain, AEJ.LGG-lgGa, for SSLP markers and located the differential chromosome segment containing the lgGa locus in a region near the distal end of mouse Chromosome (Chr) 13. This linkage was confirmed and refined by typing SSLPs in 64 F2 and 74 BC1 progeny of a cross of LGG/Bc (lgGa/lgGa) to SWV/Bc. The lgGa mutation maps to a 1- to 2-cM region between D13Mit76 and D13Mit53. Integrin alpha 1 and integrin alpha 2, which map to the same general region, are possible candidate loci, based on their embryonic expression and cellular function. Evidence is also presented for a common unlinked recessive suppressor of the open eyelids trait caused by lgGa.

Genetically determined absence of an initiation site of cranial neural tube closure is causally related to exencephaly in SELH/Bc mouse embryos.

The SELH/Bc mouse strain (SELH) has a high frequency of the lethal neural tube closure defect, exencephaly, in newborns and embryos. Previous work has shown that all SELH mouse embryos have an abnormal mechanism of rostral neural tube closure. They lack initiation of contact and fusion of the cranial neural tube at the prosencephalon/mesencephalon boundary [Closure 2), and undergo closure by extension of a more rostral site of fusion. This process fails in 10-20% of embryos, where the mesencephalic folds remain unelevated, resulting in exencephaly. Previous work has also shown that the cause of liability to exencephaly in SELH mice is multigenic, involving a small number of loci. The purpose of the present study was to test the hypothesis that the genes causing the lack of Closure 2 also cause the liability to exencephaly in SELH, by observation of their joint transmission from genetically segregating animals. A concurrent mapping study provided the necessary genetic material, a segregating F2 generation from a cross of SELH with the normal LM/Bc strain. The genetic liability to exencephaly transmitted by individual F2 sires had been measured by the frequencies of exencephalic day 14 embryos they produced in test-crosses with SELH females. A selected subset of 13 of these test-crossed F2 sires was bred with a second set of SELH females, and the embryos were examined earlier, during the period of neural tube closure, on days 8 and 9 of gestation, to determine the presence of Closure 2. Six F2 sires were among the highest exencephaly producers (6-11%), six were among the lowest (0%), and one was intermediate (5%). Among embryos at the appropriate stage for scoring, the presence of Closure 2 was observed to be inversely correlated with the later risk of exencephaly, being present in 93% (71/76) from the low-risk sires and 35% (36/103) from the high-risk sires. In each case, the remaining embryos had a closure mechanism like that of SELH embryos. Among the individual intermediate- and high-risk sires, there was also a clear correlation between the frequency of exencephaly in older embryos and the frequency of lack of Closure 2 in early embryos (rs = 0.88; P < 0.05). This study demonstrates that high liability to exencephaly and absence of Closure 2 are genetically transmitted together. That is, the cause of the lack of Closure 2 in SELH mice is shown to be also the probable cause of the high liability to exencephaly.

Deficient and delayed primary palatal fusion and mesenchymal bridge formation in cleft lip-liable strains of mice.

During mammalian primary palate formation, the facial prominences enlarge around the nasal pit, fuse and then merge to give rise to the tissue of the upper lip and premaxillary region. The mechanisms involved in successful primary palate formation and how they are affected in the cleft lip genotype remain poorly understood. The purpose of this study was to compare morphometrically internal development and growth of the primary palate in five different strains of mice. Two of the strains, BALB/cByJ, and C57BL/6J, have normal primary palate development, and three of the strains, A/J, A/WySn, and CL/Fr, have stable frequencies of cleft lip associated with genotype. In the present study, frequencies of 4, 23, and 24%, respectively, were observed on day 13. For palatal growth analysis, embryos were collected on days 10 and 11, staged by number of tail somites (TS), and the heads were photographed and serially sectioned for measurement of primary palate components. The heights of the epithelial seam and the mesenchyme bridge between the facial prominences were measured on serial sections and areas of contact were calculated. The position or depth of the maxillary prominence was determined from the number of frontal sections from its tip to the rostral end of the nasal fin. Analysis of measurements showed that in cleft lip strains enlargement of the epithelial seam and replacement of epithelia by a mesenchymal bridge were both delayed relative to somite stages. Measurements from day 11 embryos with complete failure of contact were excluded from the growth analyses. The mesenchymal bridge formed at 12--13 TS in noncleft strains, 14 TS in the A/J strains with higher cleft lip frequency, and 15--17 TS in A/WySn and CL/Fr strains with higher cleft lip frequency. Forward growth of the maxillary prominence was highly correlated with the primary palate measurements and mesenchymal bridge formation in all strains. In both cleft and noncleft strains, the primitive choanae open at 18--20 TS and the medial nasal region narrows with advancing embryonic development. As a result, cleft lip-liable strains have a narrower window in development in which a robust mesenchymal bridge must form, thus increasing the liability to cleft lip.

The major locus for multifactorial nonsyndromic cleft lip maps to mouse chromosome 11.

Cleft lip with or without cleft palate, CL(P), a common human birth defect, has a genetically complex etiology. An animal model with a similarly complex genetic basis is established in the A/WySn mouse strain, in which 20% of newborns have CL(P). Using a newly created congenic strain, AEJ.A, and SSLP markers, we have mapped a major CL(P)-causing gene derived from the A/WySn strain. This locus, here named clf1 (cleft lip) maps to Chromosome (Chr) 11 to a region having linkage homology with human 17q21-24, supporting reports of association of human CL(P) with the retinoic acid receptor alpha (RARA) locus.

Genetic analysis of the construction of the AEJ.A congenic strain indicates that nonsyndromic CL(P) in the mouse is caused by two loci with epistatic interaction.

The A/WySn mouse strain has a high frequency (20%) of the craniofacial defect, cleft lip with or without cleft palate, CL(P), in fetuses or newborns. Previous genetic studies have indicated that the genetic control of the risk of CL(P) is complex, but the cause of the liability of the embryo involves only one or two loci. The genes that cause the liability to CL(P) in A/WySn have been transferred by 12 generations of selective breeding to a normal strain background, AEJ/RkBc, to form a congenic strain pair. The new strain, AEJ.A/Jur, has been used to map a major CL(P) locus, clf1. Analysis of the genetic data from the process of constructing the AEJ.A/Jur strain indicates that the cause of CL(P) in A/WySn mice is the joint action of two recessive loci, clf1 and clf2 with unequal duplicate epistasis. That is, the normal allele at clf1 is a dominant suppressor of the recessive phenotype at clf2, and the normal allele at clf2 is a semidominant suppressor of the recessive phenotype at the clf1 locus. To be at risk for CL(P), embryos must be homozygous for clf1 mutations and at least heterozygous for clf2 mutations, and the risk is higher if they are homozygous for mutations at both loci. The delineation of this epistatic model involving a locus, clf1, that has linkage homology with a region implicated in human CL(P), and that has three paralogous regions in both species, supports the following arguments: and that has three paralogous regions in both species, supports the following arguments: that the CL(P) of the A/WySn mouse strain is a homologue of human CL(P); that the genetic complexity of CL(P) in both species originates in epistasis, not in polygenic modifiers; that human linkage analyses should specify epistatic models for CL(P); that the paralogous linkage groups should be considered candidate regions for other CL(P) loci.

Development of the cerebellar defect in ataxic SELH/Bc mice.

In SELH/Bc mice, 5-10% of young adults are ataxic, due to a midline cleft in the cerebellum. An additional 10-20% of SELH/Bc embryos have exencephaly and die at birth. All SELH/Bc embryos omit a normal step in cranial neural tube closure, initiation of fusion at Closure 2. In the 80-90% that complete cranial neural tube closure, the last region of closure, on late D9, is the region of the prospective cerebellum, and its closure is late. We postulated that the cleft cerebellum in ataxic SELH/Bc mice derives from this delay in neural tube closure and predicted that we would see evidence of a cerebellar midline cleft in all earlier stages after cranial neural tube closure is normally complete. In the present study we show that the cerebellum is cleft in a 7-9% proportion of SELH/Bc D16 fetuses (2/28) and D11 embryos (15/167), and that the defect is detectable on D10. In these abnormal D16 fetuses, D11 and D10 embryos, there is a gap in midline continuity of cerebellar neuroepithelium, a finding consistent with our hypothesis that the neuroepithelium in this region fails to complete fusion in those embryos. We also show that cerebella of adult SELH/Bc ataxic mice have no obvious deficiency of lobules, or disorganization of tissue as in the Wnt-1 mutants.