One-Cell Zygote Transfer from Diabetic to Nondiabetic Mouse Results in Congenital Malformations and Growth Retardation in Offspring

Abstract

Fetuses of type 1 and 2 diabetic women experience higher incidences of malformations and fetal death as compared with nondiabetics, even when they achieve adequate glycemic control during the first trimester. We hypothesize that maternal diabetes adversely affects the earliest embryonic stage after fertilization and programs the fetus to experience these complications. To test this hypothesis, we transferred either one-cell mouse zygotes or blastocysts from either streptozotocin-induced diabetic or control mice into nondiabetic pseudopregnant female recipients. We then evaluated the fetuses at embryonic d 14.5 to assess fetal growth and the presence or absence of malformations. We found that fetuses from the diabetic mice transferred at the blastocyst stage but also as early as the one-cell zygote stage displayed significantly higher rates of malformations consistent with neural tube closure problems and abdominal wall and limb deformities. In addition, both these groups of fetuses were significantly growth retarded. To determine if this phenomenon was due to high glucose concentrations, two-cell embryos were cultured to a blastocyst stage in 52 mm d-glucose or l-glucose as an osmotic control, transferred into nondiabetic pseudopregnant mice, and examined at embryonic d 14.5. These embryos did not demonstrate any evidence of malformations, however, they did experience significantly higher rates of resorptions, lower implantation rates, and they were significantly smaller at embryonic d 14.5. In summary, exposure to maternal diabetes during oogenesis, fertilization, and the first 24 h was enough to program permanently the fetus to develop significant morphological changes.

FETUSES OF TYPE 1 and 2 diabetic women experience a higher incidence of malformations, mostly neural tube defects (NTDs) and skeletal/cardiovascular abnormalities, and fetal death compared with nondiabetic pregnant women (1,2). Most diabetic rodent studies focus on development after implantation and during organogenesis, at embryonic d 9–11. However, in humans these complications still occur at rates 4- to 10-fold higher than nondiabetic patients despite the fact that these women obtain prenatal care and adequate glycemic control during the first trimester and often within days of implantation (1). Due to these clinical observations, we hypothesize that maternal diabetes adversely affects the mammalian zygote at the earliest stages, before implantation, and that these insults manifest later in development as a malformation, growth retardation, or spontaneous resorption. Our data support this hypothesis and suggest that in vivo metabolic insults can permanently affect future development as early as a one-cell zygote.

Materials and Methods

The animal experiments were all conducted within the acceptable standard of humane animal care, and the protocols were accepted by the Animal Study Committee of Washington University. To test our hypothesis, we used embryo transfer studies in which we transferred either one-cell zygotes or blastocyst stage murine embryos from superovulated streptozotocin-induced type 1 diabetic mice (B6XSJL F1 mice) vs. control mice into nondiabetic pseudopregnant female recipients. Induction of diabetes, superovulation, recovery of embryos, and transfers were all described separately in prior publications (3,4,5). The embryo cultures from the two-cell to blastocyst stage embryos were conducted in potassium simplex optimized media supplemented with either 52 mm d-glucose or 52 mm l-glucose as an osmotic control. This concentration was chosen because we had demonstrated in earlier manuscripts that blastocysts cultured at 52 mm d-glucose most closely mimicked the phenotype of the diabetic blastocysts with the same degree of apoptosis, and decreased glucose transport and glucose transporter 1 expression (4,6,7,8). Because equal concentrations of l-glucose had no effect on the embryos, it is unlikely that this concentration is toxic. The atmosphere that the embryos were cultured in was 5% CO₂ in air at 37 C. The diabetic mice were hyperglycemic for at least 7 d before recovery of the embryos, confirmed by blood glucose levels of more than 250 mg/dl. The transferred fetuses were then evaluated by a blinded observer at embryonic d 14.5 (equivalent to late trimester in humans) by killing the recipient mice and performing hysterotomies to recover the fetuses, assess fetal growth, and detect the absence or presence of gross malformations by inspection based on an adaptation of the morphological assessment of diabetic embryos established by Wentzel et al. (9). Total implantation sites and resorption sites were recorded per mouse. The embryos were dissected free of membranes and then fixed in Bouin’s solution at room temperature for 24 h. Subsequently, the embryos were dehydrated in graded ethanol, transferred to xylene, embedded in paraffin, and serially sectioned at 5 μm. Sections were rehydrated and stained with hematoxylin-eosin, and read by a pathologist blinded to the origin of the embryos. Major and minor malformations were noted, as well as any other developmental abnormalities were recorded. Transfer experiments were done at least five times for each group and treatment condition. Embryo measurements were compared using Student’s t tests and ANOVA when appropriate. Significance was defined as P < 0.05.

Results and Discussion

Fetuses that developed from the transferred one-cell diabetic zygotes displayed significantly higher rates of malformations consistent with neural tube closure problems, and higher rates of hydrocephaly, skeletal disorders, and growth retardation compared with zygotes transferred at the same stage from nondiabetic controls into controls (Fig. 1). Exposure of the ovulated oocytes through the fertilized one-cell zygote stage to the maternal diabetic condition in vivo for 24 h was enough to program the zygote to go on to develop into a growth retarded and/or malformed fetus (size: control, 11.4 ± 0.09 mm vs. diabetic, 10.6 mm ± 0.12, P < 0.007; malformation rate: control, 0.00% vs. diabetic, 9.7 ± 1.6%). The one-cell transfers from the diabetic mice did not display higher rates of resorptions and had no decrease in implantation rates compared with controls.

Figure 1.

In vivo diabetic one-cell zygote transfers. One-cell zygotes recovered 24 h after human chorionic gonadotropin injection and mating from either streptozotocin-induced diabetic mice or control mice were transferred into nondiabetic pseudopregnant female recipients. The fetuses were evaluated at embryonic d 14.5 to assess fetal growth and the absence or presence of malformations. A and B, The fetuses from the diabetic mice displayed significantly higher rates of malformations consistent with neural tube closure problems and abdominal wall and limb deformities (six transfer experiments for each condition). In addition, these fetuses from the one-cell zygotes were significantly growth retarded. Arrows show neural tube abnormalities and hydrocephaly, both more common in the diabetic zygotes.

In contrast, the embryo transfers at the diabetic blastocyst stage experienced significantly higher rates of detectable resorptions and lower implantation rates (Fig. 2, A and B). Similar to the one-cell zygote transfers, the diabetic transferred blastocysts developed into fetuses with higher malformation rates, including anterior and posterior NTDs, limb deformities, and growth retardation (Fig. 2, C and D). Exposure to the maternal diabetic conditions for 96 h in vivo, during the period of ovulation through cleavage stage and compaction to the blastocyst stage of development, was enough to program the embryo to go on to develop not only into a growth retarded fetus or malformed fetus like the 24-h exposure but also into a failed pregnancy manifesting as a resorption or miscarriage (Fig. 2, A and B).

Figure 2.

In vivo diabetic blastocyst transfers. Blastocyst stage embryos, recovered 72 h after human chorionic gonadotropin and mating from either streptozotocin-induced diabetic mice or control mice were transferred into nondiabetic pseudopregnant female recipients. A and B, Resorption or failed pregnancy rates and malformation rates were higher in diabetic blastocyst transfers compared with control (six transfer experiments for each condition). In addition, the diabetic blastocyst transfers were significantly smaller than controls. C and D, Diabetic blastocyst transferred embryos experienced higher rates of anterior and posterior neural tube closures, as well as skeletal and limb abnormalities. *, P < 0.01.

To determine if this phenomenon could be replicated in vitro, two-cell embryos from control mice were cultured to a blastocyst stage (72 h) in 52 mm d-glucose to mimic the maternal diabetic condition or l-glucose as an osmotic control, transferred into nondiabetic pseudopregnant female mice, and examined at embryonic d 14.5. These embryos cultured in d-glucose experienced significantly higher rates of resorptions, lower implantation rates, and were significantly smaller at embryonic d 14.5, however, no malformations were seen (Fig. 3).

Figure 3.

In vitro cultured embryos and blastocyst transfers. Two-cell embryos from control mice were cultured for 72 h in 52 mm d-glucose (D) or 52 mm l-glucose (L) and transferred into pseudopregnant controls at a blastocyst stage. A and B, Embryos recovered at embryonic d 14.5 were significantly smaller in the d-glucose exposed embryos (five transfer experiments). C, These embryos also experience increased resorptions and decreased normal implantation rates. *, P < 0.001; **, P < 0.05.

This set of experiments suggests that the timing of exposure to an abnormal maternal milieu is critical to the developing oocyte and preimplantation embryo. Moreover, it is the first study to show that in vivo metabolic changes, before implantation and as early as the first 24 h after fertilization, can result in congenital malformations and growth retardation. In addition, we demonstrate that the effect of maternal diabetes has different long-lasting effects on embryo outcome, depending on developmental exposure, specifically whether the insult occurs during ovulation, fertilization, cleavage stage, compaction, or blastocyst formation (Fig. 4). These findings support the fetal origins of adult disease, a hypothesis first described in 1993 by Barker (23) to explain how suboptimal maternal and fetal nutrition can have a sustained impact on the adult phenotype.

Figure 4.

Cartoon of three transfer groups and treatments. Time line on top bar represents d 0 or day of streptozotocin injection to induce hyperglycemia. After confirming induction of hyperglycemia by checking blood glucose levels greater than 250 mg/dl (average 4 d), mice were superovulated with pregnant mares’ serum gonadotropin (pmsg), followed 48 h later by human chorionic gonadotropin (hcg) to induce polyovulation. The mice were mated with control males overnight (1). In the one-cell zygote transfers, in vivo embryos were removed 24 h after mating from the diabetic mice and transferred to nondiabetic ICR control mice (2). In the blastocyst transfers, in vivo embryos were recovered 72 h after mating and transferred into nondiabetic recipients (3). In the in vitro transfers, two-cell embryos were recovered from superovulated control mice 48 h after mating and cultured in 52 mm d-glucose or 52 mm l-glucose for 72 h. All embryos reaching a blastocyst stage by 72 h were transferred into control recipients. Recipients from all three groups were killed on embryonic d 14.5, and observed/examined for malformations and measured. Resorption sites were counted.

Recent studies have support that preimplantation exposure can have an even earlier impact on offspring. Watkins et al. (10) demonstrated that in vitro culture of mouse embryos from a two-cell to a blastocyst stage without adding a protein source resulted in offspring with elevated systolic blood pressure and increased activity of angiotensin-converting enzyme and hepatic phosphoenolpyruvate carboxykinase. They speculated that postnatal cardiovascular physiology may be more sensitive to maternal dietary changes during the earliest period of development, such as a low protein diet, than postnatal growth. Moreover, exposure to environmental xenobiotics during the earliest stages of postimplantation mammalian development has also been linked to adult diseases (11), and more recently to epigenetic modifications such as DNA methylation of imprinting regulatory elements and transposable elements (12,13). In one report, epigenetic changes were found in the F2 and F3 generation germ line as a result of male embryonic exposure of the F1 generation to the endocrine disruptor vinclozolin at the time of gonadal sex determination (14). Similar imprinting abnormalities, a gain in methylation of H19 and demethylation of PEG1, have been linked with the female germ line in humans undergoing ovulation induction, as well as mice undergoing superovulation (15). It is also clear that metabolic perturbations, specifically the redox state of a cell, can regulate protein deacetylation of transcription factors and gene expression via the seven mammalian homologs of Sir2 (16). This family of proteins catalyzes nicotinamide adenine dinucleotide-dependent protein deacetylation, and by doing so, regulates longevity, apoptosis, DNA repair, and cellular metabolism. It is possible that metabolic changes in the early zygote or oocyte activate these factors to permanently alter development. In this study we are assuming that hyperglycemia is the insult that leads to these changes. Because we are using a streptozotocin model, it is possible that the insult could be low insulin or IGF-I, elevated free fatty acids, or hyperglycemia. Whatever the inciting factor, we propose the existence of embryonic and now zygotic origins of the postnatal phenotype, in this case congenital malformations, as the result of maternal metabolic aberrations.

Previously, it has been held that metabolic insults occurring during organogenesis were the only events resulting in pregnancy abnormalities, specifically birth defects (17). Although this study was done in a mouse model, these data establish “programming” of the one-cell zygote and cleavage stage/early blastocyst stage embryos to various environmental challenges. Prior studies have demonstrated that high glucose concentrations led to decreased glucose uptake and expression of glucose transporters during mammalian preimplantation development (4,6,18). This cellular change in glucose utilization at these earliest stages may trigger the cytoplasmic event leading to the long-term sequelae of malformations and miscarriages (8,19). Alternatively, nuclear changes as a result of maternal diabetes may be responsible for epigenetic phenomenon as described with other in vitro preimplantation studies (20,21,22). Because only 20% of NTDs in humans are due to folic acid deficiencies, these findings may also shed light on other causes of malformations and poor pregnancy outcomes due to early environmental exposure before maternal recognition of pregnancy.