Tetrasomy 16 in the mouse: a more severe condition than the corresponding trisomy

While trisomy is the most commonly observed autosomal aneuploidy in man, only a few reports have been made of full or, more usually, partial autosomal tetrasomy in either mosaic or pure form (Balestrazzi et al. 1983; Batista, Vianna-Morgante & Richiere-Costa, 1983; Garcia-Cruz et al. 1982; Schinzel et al. 1981; Shapiro, Hansen & Littlefield, 1985; Wisniewski, Hassold, Heffelfinger & Higgins, 1979). Understanding the phenotypic similarities and differences between individuals trisomic and tetrasomie for the same chromosome or chromosome segment is of considerable theoretical interest with regard to attempts to relate the phenotypic features of aneuploid states to changes in the dosage of specific genes (Epstein, 1986). Unfortunately, the conclusions that can be drawn from the human tetrasomy data are quite limited because of the scarcity of cases, the difficulties in matching unrelated translocations (two translocations for a similar region do not necessarily involve exactly the same loci), and the problems associated with investigations of human prenatal development. The existence of these limitations in the study of tetrasomy in man suggests a need for an animal model system in which specific tetrasomies can be generated and analysed.

A method for generating embryos and foetuses trisomic for whole autosomes already exists (Gropp & Kolbus, 1974; White, Tjio, Van de Water & Crandall, 1974) and is based on the use of animals, usually males, doubly heterozygous for two Robertsonian translocation (metacentric) chromosomes that have monobrachial homology (i.e. share a chromosome arm in common) (Gropp, Kolbus & Giers, 1975). A high degree of non-disjunction occurs in such animals (Gropp & Winking, 1981) and results in the generation of both nullisomic and disomic gametes. The latter, upon fertilization of a normal gamete, produces trisomic progeny. Depending on the specific metacentric chromosomes used, all of the mouse autosomal trisomics can be produced, and this breeding system has been used in a variety of studies of aneuploidy in the mouse (Epstein, 1985a, b). Since the same combinations of chromosomes can be carried by both males and females, it would be predicted that a breeding scheme employing pairs of animals with identical sets of metacentric chromosomes should result in the conception of embryos tetrasomie or nullisomie for the common chromosome arm, as well as the already observed trisomics and monosomies (Fig. 1). Such a breeding scheme had previously been used to study trisomy 19 (White, Tjio, Van de Water & Crandall, 1972) and nondisjunction (Searle & Beechey, 1982) and had been suggested by Ford(1975) as a means to produce nullisomie embryos. A preliminary report of its use to generate embryos with tetrasomy 15 has recently appeared (Beechey, Kirk & Searle, 1985).

In this report, we describe the generation and phenotypic properties of mouse embryos tetrasomie for chromosome 16. These embryos, while able to implant and survive to midgestation, are much more severely affected than are siblings with trisomy 16. In addition, we describe some previously unreported aspects of the development of trisomy 16 embryos and of a possible relationship between intrauterine conditions and the survival of aneuploid embryos.

Female mice doubly heterozygous for two metacentric Robertsonian translocation chromosomes, Rb(11.16)2H and Rb(16.17)32Lub were mated with doubly heterozygous males of the same type to generate aneuploid embryos and diploid controls. Both homozygous strains were obtained from Dr Heinz Winking and the late Prof. Alfred Gropp, Lübeck, Germany, and were maintained in our laboratory as closed outbred colonies. The day the vaginal plug was observed was considered as day 0. At the appropriate day of gestation, the pregnant females were killed, the embryos removed and their appearance recorded. Some of them were photographed before the genotype was determined by karyotype analysis.

Embryos were karyotyped by modifications of the method described in Magnuson, Smith & Epstein (1982). Day-3 (preimplantation) embryos were removed from the uteri of superovulated females (4 to 5 weeks old) and incubated for 3–4 h in modified Whitten’s medium (Golbus & Epstein, 1974) containing 0·08μgml⁻¹ of vinblastine (Velban, GIBCO). After incubation, the mitotically arrested embryos were individually transferred to a microdrop of a hypotonic solution (0·75 % sodium citrate) in a well of a microtitre plate and incubated at room temperature for 5 min and then on ice for up to 1 h. One embryo at a time was then transferred to a new well containing a 7:3 methanol: acetic acid fixative. After a few seconds, the embryo was transferred to a microscope slide and immediately covered with a microdrop of a 1:4:5 lactic acid (85 % syrup solution) : distilled water: acetic acid solution. Once the fixed embryo was seen to break apart, it was washed with an excess of fixative and dried thoroughly. The slides were stained with 3 % Giemsa (Gurr) for 10 min. The genotype was determined by counting the number of acrocentric and metacentric chromosomes per metaphase plate (Fig. 1). Tetrasomie embryos have 42 chromosome arms distributed as 34 acrocentrics and 4 metacentrics while trisomics have 41 chromosome arms and 3 metacentrics and diploid embryos have 40 arms and 2 metacentrics.

Day-6 (early postimplantation) embryos were dissected from the uteri of hormonally induced females (8 to 12 weeks old) and incubated 4–6 h in Dulbecco’s modified Eagle’s medium supplemented with 10% foetal calf serum and 0–04μg ml⁻¹ of vinblastine. The karyotyping procedure was the same as that for preimplantation embryos, except that the hypotonic solution was 0·6 % sodium citrate.

Day-10 and day-12 (midgestation) embryos were dissected either from superovulated’ or naturally mated females (8 to 12 weeks old). Small pieces of tissue were incubated for 3–4h in the same medium as day-6 embryos, then treated in 1 ml of 0·56 % KC1 for 10 min, fixed in 1 ml of 3:1 methanol: acetic acid for 15 min, and transferred to 80 μlof 60 % acetic acid to allow the cells to break apart. After 5 min, the cell suspension was transferred to a microscope slide covered with a film of a 0·01 % Nonidet P40 solution. The slides were dried at 37°C and stained with Giemsa for 4min.

Forty-two superovulated Rb2H/Rb32Lub females were mated with Rb2H/ Rb32Lub males and sacrificed at various stages of development. The mean litter size (number of preirnplantation embryos or implantation sites) was 23·1, 14·7, 15·1, and 18·2 at, respectively, 3, 6, 10, and 12 days of gestation. Twenty-three additional naturally mated females were dissected at days 10 and 12 and had a mean of 9·8 and 10·2 implantation sites, respectively.

Among the 593 embryos successfully karyotyped, 20 were tetrasomie for chromosome 16 (3·4 %). The frequency in embryos 10 days or less in gestational age was 3·3 %. The frequency of tetrasomy per 100 diploid embryos ranged from 1·8 in day-12 superovulated females to 7·8 in day-10 superovulated females (Table 1).

Approximately equal frequencies of monosomy 16 and trisomy 16 embryos were found at the early blastocyst (day 3) stage (Table 1). However, the mean cell number was significantly lower (P< 0·005, Student’s t-test) in monosomie than in diploid or in trisomic littermates (38 ±3 [S.E.M.], 52 ±2 and 49 ±3 cells per embryo, respectively). The single tetrasomy 16 embryo karyotyped at this stage had 51 cells. At day 6, no monosomie embryos were found and the ratio of trisomics to disomics was lower, although not significantly, than before implantation.

On the tenth day of development, the proportion of trisomy 16 embryos from either superovulated or naturally mated females was the same as at day 6. Two days later, no change was found in the naturally mated females, but superovulated one carried significantly fewer (P<0·05, chi-squared test) trisomy 16 embryos (Table 1). The same difference was observed when the numbers of diploid and chromosomally abnormal embryos were related to the total number of implantation sites (Fig. 2).

Trisomic and tetrasomie embryos could not be distinguished by size from normal littermates at the 6th day of gestation. However, the tetrasomie embryos were very delayed in development at day 10 (Fig. 3, Table 2). Although a beating heart and onset of somite and allantois formation was noted, fusion of the anterior neuropore had not occurred (Fig. 4B) and the rotation of the embryo was incomplete (Fig. 3C, Fig. 4A,B). The most advanced day-10 tetrasomie embryo had only 17 somites (Fig. 4D), compared with 29·8 ±1·4 (S.D.) and 26·8 ±1·9 somites in diploid and trisomic siblings, respectively. At day 12, the tetrasomie embryos were either dying (Table 2, Fig. 5D,E,F) or were so resorbed that only extraembryonic membranes were found. At this stage, only one out of seven tetrasomies still had a beating heart.

An extremely deformed and retarded day-10 embryo (Fig. 3D) had 53 acrocentric and 4 metacentric chromosomes, giving a total of 61 chromosome arms. We assume that this triploid-tetrasomic embryo resulted from the combination of diploid and haploid-disomic gametes. Alternatively, it could have been generated by a dispermic fertilization, with one of the three gametes being disomic.

At day 10 about 60 % of the trisomics were smaller than the diploid controls (Table 2, Fig. 3). Two days later, the discrepancy between the normal-appearing trisomics (Fig. 5B) and the retarded ones (Fig. 5C) was even greater, the latter being in an advanced stage of resorption (Table 2). Surprisingly, no intermediate form of abnormal phenotype was observed, suggesting the existence of two distinct populations of trisomics. The existence of these populations was not correlated with the sizes of the litters or the ages of the females.

As expected, tetrasomie embryos were actually generated by an intercross between mice doubly heterozygous for two metacentric chromosomes that share one arm in common (Fig. 1). Nullisomie embryos, also predicted by the breeding scheme to be generated in equal numbers, were not detected at the early blastocyst stage. Preliminary results from the karyotyping of earlier stages suggest that nullisomies might undergo one, perhaps, at a maximum, two divisions, before they die (S. Debrot, A. Zweig, C. J. Epstein, unpublished observations). In previous experiments (Magnuson et al. 1985) we found that the Rb2H/Rb32Lub males, when mated with ICR females, generate 17 % monosomies and 17 % trisomics when analysed after 2 days of development, giving a male non-disjunction rate of 34%. Using these data, we calculated that a 36% rate of non-disjunction in the doubly heterozygous females best fits with our day-3 results. We therefore conclude that superovulated Rb2H/Rb32Lub females have a rate of non-disjunction similar to that of doubly heterozygous males and that when inter-crossed Rb2H/Rb32Lub heterozygotes will generate about 3 % tetrasomie embryos (or 6 tetrasomies per 100 diploids). The frequency of tetrasomies found at day 10, 7·8 per 100 diploids (Table 1), is consistent with this expectation and indicates that all tetrasomies probably implant and survive to midgestation.

The findings with the tetrasomie embryos make it clear that four copies of chromosome 16 have a greater adverse effect on embryonic development than do three copies. Van der Hoeven & de Boer (1984) found a similar situation when the translocation T(l;13)70H was used to generate tetrasomy for the small regions of chromosome 1 and 13 involved in the l¹³ translocation chromosome. This tetrasomy (actually a double partial tetrasomy) is a prenatal lethal, while mice that are trisomic for the same regions survive and are fertile. Similarly, Beechey, Kirk & Searle (1985) found that tetrasomy 15 results in death by day 9–9·5, whereas trisomy 15 embryos at the same stage were still alive although they were about a third smaller than normal in size and had exencephaly. Therefore, for the two whole chromosome tetrasomies and one double partial tetrasomy studied to date, the tetrasomie state results in developmental abnormalities more severe than do the corresponding trisomics. On the other hand, Beechey & Speed (1981) described a male tetrasomie for the extra small 5¹² translocation chromosome of T(5;12)31H. This animal, although viable, was infertile, smaller than normal, and had a shortened head and small testes. A slightly retarded foetus with a twisted spine and the same genotype was also found at 12·5 days of gestation (Beechey, Kirk & Searle, 1980). These abnormalities are not much more severe than were found with the equivalent trisomy which resulted in male sterility and mild growth retardation. Overall, the situation in the mouse is reminiscent of what has been reported in man, with the effects of aneuploidy ranging from little or no specific phenotypic differences between tetrasomies and trisomics to the existence of clearly distinguishable syndromes with the tetrasomies being more severely affected (Epstein, 1986).

In addition to its usefulness for producing tetrasomie (and, potentially, nullisomie) embryos, the breeding system described here, with both parents bearing doubly heterozygous chromosomes, may also be of value in the study of trisomy (White et al. 1972). Trisomic embryos will be generated in a higher proportion than in the traditional mating of a doubly heterozygous male with a normal ‘all acrocentric’ female. For example, with the Rb2H/Rb32Lub heterozygotes used in this study, 47 rather than 27 trisomics per 100 diploids can be expected when two double heterozygotes are intercrossed rather than just one being mated to a wild-type animal.

Our data do not reveal any differences in the incidence of aneuploidy between superovulated and naturally mated females in embryos analysed at 10 days of gestation (Fig. 2A,B). However, a significant loss of aneuploid embryos is observed two days later in superovulated females (Fig. 2C). We interpret this result as indicating that a selection process may be operating within the uterus. In a crowded situation (20·5 implants/female after superovulation versus 10·2 after natural mating), a greater competition for either nutrients or space could occur at midgestation which could place trisomic and tetrasomie embryos at a proliferative disadvantage. This conclusion would be consistent with the suggestion of Boué, Philippe, Giroud & Boué (1976) and Gropp (1981) that the time of death of aneuploid foetuses could be determined in part by the ability of the aneuploid placenta to serve the metabolic needs of the embryo or foetus. In a marginal situation, with an already hypoplastic placenta (Gearhart, Oster-Granite & Hatzidimitriou, 1985), trisomy 16 embryos could be quite vulnerable to the local effects of uterine crowding.

The separation of the trisomy 16 embryos into two populations, some being grossly normal or only mildly abnormal while the others are severely delayed, could suggest a critical developmental threshold about or shortly before the 10th day of development. Some trisomics may be so delayed in development at this stage that they do not recover and are resorbed, while the others are able to cross this threshold and continue through a programme of fairly normal development up to and beyond day 12. Trisomy 16 foetuses have been reported to survive until or even shortly beyond birth (Miyabara, Gropp & Winking, 1982). It is of interest to note, in the context of a barrier to embryonic development, that the 9th day of gestation is often considered in the mouse to be a critical one, since it appears to be the stage most sensitive to various exogenous insults (Rugh, 1968). Furthermore, it also represents the time at which the most lethal of the mouse trisomics result in embryonic death (Gropp, 1978).

We are grateful to Ms Theodosia Zamora, Ms Estrella Lamella and Mr Karl Meneses for valuable technical assistance. This work was supported by grants from the National Institutes of Health (GM-24309 and HD-17001). S.D. was a recipient of a postdoctoral research fellowship (83.111.0.83) from the Swiss National Science Foundation.