Early embryonic lethality in genetically engineered mice: diagnosis and phenotypic analysis

Abstract

Embryonic lethality is a common phenotype that occurs in mice homozygous for genetically engineered mutations. These phenotypes highlight the time and place that a gene is first required during embryogenesis. Early embryonic lethality, i.e. before and up to mid-gestation, can be straightforward to analyze because the stage at which death occurs suggests why an embryo has failed. Here we summarize general strategies for analyzing early embryonic lethal phenotypes in genetically engineered mouse mutants.

Introduction

The intrauterine development of mammals provides a protective, nurturing environment for the complex processes of embryogenesis but presents an obstacle to the observation and study of early-acting mutations. Particularly in cases of lethality during gestation, analysis of a mutant phenotype may require specialized embryological and analytic tools stretching back to the time of fertilization. The earlier a mutant gene has its primary effect, the more likely the defect will ramify throughout development and lead to embryonic lethality. In addition, disruption of the early events of placentation can quickly lead to the demise of the embryo.

Although it may not always be the intended outcome, many mutations produced through genetic engineering affect developmental processes and compromise heterozygous or homozygous mutant embryos long before birth. Key to any analysis of the effects of a mutant gene is a thorough understanding of the expression pattern of the gene in question, but expression studies may not have extended to all embryonic stages and thus some surprises may be encountered when a mutation is produced. In this article, we outline a general strategy for investigating embryonic lethality caused by genetically engineered mutations from fertilization to just beyond mid-gestation (~12.5 days of gestation). It presupposes a general knowledge of mouse embryogenesis^1,3,4,8,9 and also that a detailed expression analysis is available or will be done in conjunction with the phenotypic analysis to guide the investigation and to fully understand the etiology of a mutant phenotype. We make the further assumption that a robust genotyping assay is available and that heterozygous mice are viable and fertile. The special case of a dominant lethal mutation is very rare and requires a different approach for analysis (discussed in detail in ref. 6).

Indications of embryonic lethality and establishing the time of death

Usually the first indication that a genetically engineered mutation has an embryonic lethal effect is a modified mendelian ratio of genotypes among offspring born from heterozygous crosses, and a smaller litter size compared to controls. If the mutation is fully penetrant, there will be no homozygous mutant offspring and there will be a 2:1 ratio of heterozygotes to wild-type mice (Fig. 1). However, if the phenotype is variable, some homozygous mutants may survive but in lower than expected numbers. Once it has been established by observation that the homozygous mutants are not dying between birth and weaning, then the time of embryonic or fetal death of the mutants can be determined by examining key stages of gestation from timed pregnant females.

Figure 1.

PCR genotyping suggests homozygous mutant lethality. Diagram showing the results of a cross between a heterozygous female (Mo, mother) and heterozygous male (Fa, father), yielding pups (1-6) that were genotyped at weaning. The litter is smaller than average size and there is a 2:1 ratio of heterozygous to wild-type pups. Controls include: m/+, known heterozygote as a positive control; +/+, known wild type as a negative control; and no DNA as a negative control. M, 100-bp ladder.

Embryonic day (E) 12.5 provides a useful starting point for analysis. Systematic dissection and examination of each implantation site and embryo (Fig. 2), followed by genotyping each embryo, will reveal the time of death of homozygous mutants. If no homozygous mutants are detected and all implantation sites contain normal, wild-type or heterozygous embryos, it is likely the mutants died some time during the 4 days prior to implantation and thus did not induce a decidual swelling in the uterus. Further analysis would then concentrate on preimplantation stages (next section). If all the embryos are wild type or heterozygous but there are implantation sites with no embryonic remains or a small amount of degenerating tissue (Fig. 2A,E,F), it is likely that the mutants died after implantation but prior to or shortly after establishment of the chorioallantoic placenta at about E9.5 and thus are in the process of being resorbed. Finally, if it is confirmed by genotyping that the mutants are among the embryos present at E12.5, whether they are normal or not, their phenotype can be assessed (Fig. 2B-D, G,H). If the mutants are indistinguishable from wild type at E12.5, a later time of death is indicated; if they are already dead or abnormal (Fig. 2G,H), the condition and stage of development of the mutants will provide information about the time of death or onset of the mutant phenotype⁶.

Figure 2.

Dissection of embryonic day (E) 12.5 normal and abnormal embryos. (a) Composite figure of a gravid uterus with 4 implantation sites on the right side and 9 sites on the left side including one small blood filled site (arrowhead). (b-d) Successive stages of dissection of a normal conceptus with giant cells and Reichert's membrane intact and forming a band around the yolk sac (b), giant cells and Reichert's membrane reflected to the edge of the placenta (c), and with the placenta removed and the yolk sac reflected (d). The embryo is still enclosed in the amnion. (e, f) Successive stages of dissection of the conceptus indicated by the arrowhead in a. In addition to placental structures, only membranes and embryonic debris (arrow) remain in this early resorption. (g,h) Successive stages of dissection of a Tbx2 homozygous mutant embryo suffering from cardiac insufficiency. The mutant was still vital as evidenced by a heartbeat, but there is a reduced amount of blood in the yolk sac vasculature, pooling of blood in the heart (g), and pericardial edema (h). Development is also somewhat retarded compared with the wild type control judging by developmental landmarks ⁴,⁸. c, cervix; e, eye; fp, fat pad, gc, giant cells; h, heart; m, mesentery; o, ovary; p, placenta; ys, yolk sac. Bar in panel a=0.5 mm; bar in panel b=0.3mm and indicates scale for b-h.

It is essential to examine a sufficient number of embryos and to genotype each embryo to establish statistical significance of a correlation between genotype and phenotype. Variability in a phenotype requires a larger number of embryos for analysis to firmly establish the time or range of time of death. It should be noted that embryonic death is not uncommon among wild-type mice and may vary with genetic background. Consequently, not every empty decidua or abnormal embryo is necessarily attributable to the mutant effect.

Preimplantation lethality

The fertilized oocyte or zygote undergoes regular cleavage divisions to form the blastocyst during preimplantation development, making the morphological assessment of preimplantation stages straightforward. At E3.5, blastocysts flushed from the uterus, or earlier stages flushed from the oviduct, will still be enclosed within the zona pellucida and can be recovered, scored for viability and stage of development (Fig. 3), and genotyped by PCR. Identification of the mutants by genotype will allow correlation with any developmental arrest, delay or abnormality observed. Embryos recovered from the oviducts at E0.5 shortly after fertilization, or at later stages from the oviduct or uterus can be cultured in vitro through all preimplantation stages⁵. Cultured embryos can hatch from the zona pellucida and will attach to a culture dish, a proxy for implantation, at the appropriate time (Fig. 4). Thus, for a dynamic picture of a preimplantation defect, embryos can be recovered at the one-cell stage at E0.5, observed during culture throughout preimplantation development, hatching, attachment and outgrowth and can then be genotyped. In addition, proliferation, cell death, and expression of specific genes that are diagnostic of developmental stages or cell types can be assessed. Analysis of RNA, proteins, histological and ultrastructural details are also possible ^5,6,9. However, because destructive techniques may preclude subsequent genotyping of embryos, reliance on mendelian ratios may require a large number of embryos to accurately attribute any defects observed to a mutation. More sophisticated experimental approaches are also possible, including the use of fluorescent protein markers that are becoming a more common element of many gene targeting experiments for microscopic analysis of living embryos during culture².

Figure 3.

Stages of preimplantation development including unfertilized oocytes and zygotes from embryonic day (E) 0.5 to E3.5 as well as abnormal embryos. All are surrounded by the non-living zona pellucida. (a) Ovulated (unfertilized) oocytes with a single polar body. b) One cell zygotes with two polar bodies and two haploid pronuclei. (c-e) Two- to eight-cell cleavage stages with discrete blastomeres. (f) Compact morulae: individual blastomeres are no longer distinguishable due to tighter cell-cell adhesion. (g) Blastocysts consisting of an outer layer of trophoblast surrounding the blastocoelic cacity and an acentrically located inner cell mass. The trophoblast overlying the inner cell mass is the polar trophoblast and the rest is the mural trophoblast. (h) Abnormal embryos fragmenting (*), shrunken (arrow) or degenerate. bc, blastocyst cavity; bl, blastomere; icm, inner cell mass; mt, mural trophoblast; pb, polar body; pn, pronucleus; pt, polar trophoblast; zp, zona pellucida. Bar=0.1 mm. (Figure adapted with permission from reference ⁶.)

Figure 4.

In vitro attachment and outgrowth of a blastocysts. (a,b) Embryos explanted at embryonic day (E) 3.5 after removal of the zona pellucida and grown for 2 days in vitro. The embryo in a has attached to the culture dish, the blastocysts cavity has collapsed, and the trophoblast cells are beginning to outgrow on the plastic surface. In b, the trophoblast cells form a monolayer on the culture dish surrounding the central clump of cells composed of ICM derivatives. (c,d) Embryos explanted at E3.5 after 7-8 days of in vitro growth. The large clear nuclei with prominent nucleoli are the giant trophoblast cell nuclei (arrow) in the monolayer surrounding the ICM. ICM derivatives form a compact central mass and have also begun to migrate away from the outgrowth (black arrowheads) as isolated presumptive endoderm cells. A few rounded dead cells are evident (white arrowhead). (Figure adapted with permission from reference ⁶.)

Peri-implantation lethality, E4.5-E5.5

The detection of empty implantation sites upon dissection at E12.5, or sites with only trophoblast giant cells indicates lethality shortly after implantation. Embryo implantation is accompanied by decidualization of the uterine tissue, which results in visible swellings or implantation sites as early as E5.0. This occurs in response to the presence of an embryo, but can be triggered by a highly compromised embryo, as long as it has shed the zona pellucida and has differentiated a trophoblast layer. Isolating living embryos from the uterus during the implantation process is possible, but impractical because they are small and becoming attached to the epithelial lining of the uterus. Consequently, assessment of a mutation causing lethality at this stage is best approached by examining embryos in the late preimplantation period as indicated in the previous section to determine the phenotype of the embryos prior to implantation. In vitro culture starting at E3.5 (Fig. 4) can be used to assess the implantation process itself, and in addition, histological and immunohistochemical evaluation of the implantation site can be performed in serial sections of the uterus⁶.

Postimplantation lethality, E5.5-E12.5

Successful development through implantation depends on the development of extraembryonic tissues, the trophoblast (also called trophectoderm or TE), primitive endoderm and their derivatives, which mediate the initial communication between the fetus and mother. At the same time, the embryo is proceeding through gastrulation, forming the primary germ layers (ectoderm, mesoderm, and endoderm) and mesodermally-derived extraembryonic structures that will contribute to the yolk sac and chorioallantoic placenta (see article by Ward, this issue). Coordination of the establishment of the chorioallantoic placenta with the onset of fetal circulation (E8.5-E9.5) is a critical point in development. Failure results in degeneration of the embryo, although extraembryonic structures can persist for some time (Fig. 2E,F). Because of the interdependence of embryonic and extraembryonic structures, a primary defect in one tissue can have secondary effects on other tissues and on embryo survival. Thus, knowing the normal expression pattern of the mutated gene will greatly simplify analysis of a lethal phenotype by pinpointing the tissue in which the mutation is likely to have its primary effect.

Morphological assessment of embryos during the postimplantation period can be accomplished by dissection (Figs. 2,5) and also by fixation and sectioning of embryos within the uterus^3,6. The trade-off is that integrity of the placental and extraembryonic structures is compromised by dissection and is best evaluated with the embryo in utero. However, dissection of embryos allows for the evaluation of features such as the presence or absence of a heartbeat, presence of blood and blood circulation and extravascular blood. Features such as the direction of heart looping and embryo turning as well as the accurate assessment of developmental stage by somite number is more easily established in freshly dissected embryos than in sections⁶. Following dissection, embryonic and placental samples can be fixed for further analysis either as whole mounts or sections for cell death and proliferation assays, gene or protein expression, or evaluation of detailed morphological and histological features. Samples for genotyping are easy to obtain during dissection, as even degenerating tissue will usually yield sufficient DNA for PCR analysis, whereas embryos sectioned within the uterus will have to be genotyped by recovery of embryonic tissue after sectioning. With either method, abnormal embryos can be assessed in comparison with their control littermates which helps check for differences in developmental timing within and between litters.

Figure 5.

Postimplantation stages of mouse development between embryonic day (E) 5.5 and E16.5. Lateral views with anterior to the left. Arrowheads in E5.5-7.5 images indicate junctions between the embryonic (above) and extraembryonic (below) regions. The trophoblast and Riechert's membrane have been removed from the E5.5-7.5 embryos. The placenta, Reichert's membrane, yolk sac and amnion have been removed from all other embryos. Bars, 0.1 mm, E5.5-7.5, 0.5 mm, E8.5-10.5, 1mm, E12.5-16.5. (Figure adapted with permission from reference ⁶.)

The expression pattern of the gene will inform the detailed analysis of the mutant embryos, but there are several main causes of embryonic lethality during the early postimplantation period. Placental failure or placental insufficiency is primary among these and can lead to growth delays or death of the fetus. The underlying cause can be defects in the early-functioning yolk sac placenta, defects in trophoblast differentiation, proliferation or function, or failure of development or fusion of the allantois to the chorion to form the chorioallantoic placenta.

A second major cause of lethality is cardiovascular insufficiency. This can result from defects in development of blood, heart or vessels. Hematopoietic defects may affect the earliest hematopoiesis, which takes place in the yolk sac, or later hematopoiesis in the fetal liver. Pale embryos or an obvious lack of blood or a small liver could indicate hematopoietic defects. Morphogenetic defects in heart or abnormalities in vasculogenesis or angiogenesis can severely compromise development. Coordination of the extraembryonic circulation in yolk sac and allantois with the developing fetal heart and circulation is essential to establish the fetal-maternal link to maintain the embryo.

Whatever the defect, there are a myriad of molecular markers available that can be used to define particular stages of development or specific tissue types. These are essentially changes in gene expression that represent signposts of development. Their identification supplements the morphological analysis with a molecular backup that defines developmental stages or the differentiation of specific cell types and the functioning of specific signaling pathways. The onset of a molecular phenotype will usually precede a morphological phenotype and can lead to a molecular understanding of the mutant effect. These markers can be detected using in situ hybridization for mRNA, immunohistochemistry for proteins, and genetic reporters, if available. Molecular marker analysis is second in importance only to morphological analysis for the characterization of developmental defects.

Beyond the uterus – circumventing a lethal phenotype

The foregoing analysis is aimed at pinpointing the time and cause of early embryonic death, but will not necessarily provide complete information on gene function in all tissues. If, for example, a mutation compromises implantation through an effect on trophoblast, but the gene is also expressed in the brain, peri-implantation death will preclude study of a potential later role in the brain. There are a number of ways of circumventing this type of problem to determine developmental potential of different embryonic tissues: Short term culture either of whole embryos or of organ rudiments can circumvent placental insufficiency or other early developmental problems such as hematopoietic defects by providing the fetus or organ with essential metabolic needs. Developmental potential of mutant tissue can be tested in vivo by transplantation into ectopic sites in histocompatible or immune compromised hosts for the formation of teratomas⁵. The composition of the resulting tumor growths provides an indication of the developmental potential of the mutant tissue, when compared with wild-type tissue, although the organization of the embryos will be lost. Cell lines, such as fibroblasts, embryonic stem (ES) cells, trophoblast stem (TS) cells, or primitive endoderm stem cells (XEN) can be derived from mutant embryonic tissue to study the mutant effect outside the confines of gestational development⁷. Finally, mutant cells can be tested in the context of a developing embryo by making chimeras with mutant and wild-type tissue^5,6. This technique may circumvent specific developmental problems if the wild-type tissue assumes critical functions, thus rescuing the mutant phenotype, while allowing mutant cells to be carried along to contribute (or not) to organs and tissues.

Conclusions and future directions

Early embryonic lethality caused by genetically engineered mutations have led to great insights into the genetic control of fundamental biological processes, including cell division, cell death, cell fate, differentiation, and morphogenesis. Thus, the discovery of an early embryonic lethal phenotype is an exciting opportunity for novel insights into the genetic control of biological processes with important implications for biomedical research.