The Mouse Model for Normalcy of Embryos Derived via Assisted Reproductive Technology.

Abstract

Superovulation; in vitro oocyte maturation, fertilization, and development of embryos; cryopreservation of oocytes and embryos, and embryo transfer are being applied in numerous species, especially in humans and cattle, since millions of offspring have been produced by these techniques in these two species. Many studies of normalcy of offspring resulting from assisted reproductive technology show a small increased incidence of abnormalities. Most such studies have limitations, e.g. the human babies so produced are primarily from a population of subfertile or infertile parents; thus, comparisons with offspring of fertile parents confound whether it is the compromised fertility or the assisted reproductive techniques that result in increased abnormalities. The bovine applications only rarely involve subfertile parents, but similarly, there often are confounding factors, e.g. more careful management of expensive pregnancies produced via assisted reproduction. Thus, offspring of mice (Mus musculus) often are used to assess effects of assisted reproductive techniques on offspring. Mice have many advantages: less expense per unit of data, rapid generation of data due to short gestations, genetic homogeneity, efficacious techniques, etc., but there also are limitations, some of which can be controlled, and some not. Examples of the former include use of eCG and hCG for superovulation. As used, due to long half lives, these hormones are more severe endocrine intrusions than procedures typically used in women and cattle. Embryos often are obtained from pre-pubertal mice, so one starts with an immature reproductive tract that likely is suboptimal for gametes and embryos. Also, there is the issue of strain differences, e.g. the well known benefits of C57 BL6 genetics to facilitate in vitro culture of embryos. There also is a need to control for litter size, as small litters can be problematic. An insidious problem is that the ability to study up to 100 or more endpoints economically often results in experiments with overly broad hypotheses. In such situations, probabilities increase that there will be some Type I statistical errors. A major intrinsic problem with the mouse model is the rapid developmental trajectory. For example, the maternal-embryonic transition in genetic control occurs at the 2-cell stage in mice, but at the 8- to 16-cell stage in humans and cattle. It is possible that a sub-optimal in vitro environment is more detrimental in the pre-implantation mouse embryo than for species in which cleavage stages involve more delayed differentiation. In conclusion, extrapolation of effects of assisted reproductive technologies on offspring of mice to other species may be misleading; such studies are best if controlled for superovulation, age, litter size, strain effects, and being stringent about Type I statistical errors.