Hyperglycemia-induced TET3 insufficiency is responsible for maternal transmission of glucose intolerance

Gestational hyperglycemia is detrimental to both the mother and fetus as it causes pre-eclampsia, premature birth, macrosomia, fetal trauma, low Apgar score, and fetal hypoglycemia [1, 2]. Increasing lines of evidence have shown that gestational diabetes predisposes the developing fetus to future metabolic disorders, including obesity, diabetes, hypertension, and renal disease [1, 2]. Based on the observation that obese and diabetic mothers tend to give birth to children susceptible to metabolic disorders, it has been proposed that pregestational hyperglycemia may have adverse effects on the epigenome of the oocytes, leading to the predisposition of offspring to glucose intolerance [3, 4]. However, this hypothesis has never been tested, let alone the underlying molecular mechanism.

A recent paper published in Nature reports that mothers with hyperglycemia indeed tend to have offspring with metabolic disorders characterized by glucose intolerance in both mice and humans [5]. To distinguish the effects of gestational from pregestational hyperglycemia, oocytes were collected from adult female mice with hyperglycemia to perform in vitro fertilization followed by two-cell embryo transfer to foster mothers. This approach excludes the contribution of gestational hyperglycemia to the phenotype in offspring, thus allowing for the focused dissection of the impact of pregestational hyperglycemia on offspring health. However, because in vitro gamete manipulation and embryo culture are associated with the introduction of epigenetic errors [6, 7], a potential synergistic effect of maternal hyperglycemia and IVF procedure cannot be completely excluded. Glucose tolerance tests revealed that male F1s exhibited glucose intolerance starting from 16 weeks of age, whereas female F1s exhibited glucose intolerance starting from 1 year of age. The delayed onset of glucose intolerance in females is consistent with previous reports showing the protective effects of prior sex hormones [8]. When challenged with a high-fat diet, both male and female F1s showed glucose intolerance with an earlier onset, at 12 weeks of age. Interestingly, this metabolic phenotype was detected only in F1s. Further breeding using either an incrossing or an outcrossing scheme produced no similar metabolic phenotypes in F2s, suggesting an intergenerational rather than transgenerational effect of pregestational hyperglycemia on offspring health.

Given that only oocytes are exposed to a hyperglycemic environment, the metabolic phenotype observed in F1s likely originates from the oocytes. Two possibilities exist: one is that hyperglycemia may cause epigenetic alterations, i.e., epimutations, which then lead to disrupted maternal epigenome and consequently dysregulated gene networks critical for the regulation of glucose metabolism. The second is that some unknown maternal factor(s) may be affected by hyperglycemia, which, in turn, may disrupt the epigenome of the offspring post-fertilization, leading to dysregulated glucose metabolism. To test these hypotheses, the investigators profiled the mRNA transcriptome of oocytes from female mice with hyperglycemia and the oocytes from diabetic women. Among numerous dysregulated genes detected, Tet3 mRNA levels were found to be significantly downregulated in oocytes from the hyperglycemic mice and women. The Tet3 encodes a DNA 5mC-modifying enzyme responsible for converting 5mC to 5-hydroxymethylcytosine (5hmC) and other higher oxidation products [9]. The TET3 is abundantly expressed in mature oocytes and acts as a maternal factor essential for DNA demethylation of the male pronucleus in the pronuclear stages of zygotes [9]. To further determine whether the reduced TET3 levels in the oocytes of females with hyperglycemia affect the post-fertilization genome-wide DNA methylation reprogramming, the investigators isolated both paternal/sperm and maternal/oocyte pronuclei and performed whole-genome bisulfite sequencing and APOBEC-coupled epigenetic sequencing to determine the 5mC and 5hmC profiles, respectively. Indeed, many regions in the paternal genome are hypermethylated in the zygotes, and these differentially hypermethylated regions (hyper-DMRs) persist in the pancreatic islets from fetal development to adulthood. More interestingly, genes whose promoters display hyper-DMRs are enriched in the insulin secretion pathway. For example, Gck gene, with a hyper-DMR in its pancreas-specific promoter, encodes the rate-limiting glucokinase enzyme that regulates glucose-stimulated insulin secretion and acts as an essential component of glucose sensing machinery in pancreatic β cells [10, 11].

To go beyond correlation, the investigators generated Tet3 knockout mice and found that offspring derived from the oocytes of Tet3^+/− and Tet3^−/− females displayed glucose intolerance and DNA methylation abnormalities similar to those from the oocytes of hyperglycemic mice. Moreover, the supplementation of exogenous Tet3 mRNA to mature MII oocytes from hyperglycemic mice through microinjection totally abolished the metabolic phenotype in offspring. These data unequivocally demonstrate that TET3 insufficiency accounts for hypermethylation at the paternal alleles of several genes involved in the insulin secretion pathway (e.g., Gck) that persists from zygote to adulthood, thus promoting impaired glucose homeostasis due to the defect in glucose-stimulated insulin secretion.

In summary, this study revealed a causative relationship between pregestational hyperglycemia and offspring predisposition to metabolic disorders. Hyperglycemia appears to inhibit TET3 expression and TET3 insufficiency causes defects in global DNA methylation reprogramming, leading to dysregulation of genes involved in the insulin secretion pathway, which explains the impaired glucose intolerance in offspring (Figure 1). Future experiments are required to determine the mechanism by which hyperglycemia inhibits TET3 expression and to ascertain the extent of DNA hypermethylation and other physiological perturbations. Finally, given that the oocytes from women with hyperglycemia may have compromised TET3 levels, supplementation of TET3 may represent an effective means to correct the hyper-DMRs on the paternal genome, thus re-establishing normal expression profiles of the genes involved in the insulin secretion pathway and preventing offspring from developing metabolic disorders. Because the paternal DNA methylation reprogramming seems sensitive to TET3 levels, it would be critical to determine the optimal TET3 levels that would not lead to glucose intolerance.

Figure 1.

Schematic summary of the mechanism underlying maternal epigenetic inheritance of glucose intolerance due to TET3 insufficiency induced by hyperglycemia in the oocytes. Filled lollipops indicate DNA methylation. Light pink and green colors represent normal and hyperglycemic conditions, respectively.

Conflict of interest

The authors declare no conflict of interest.