Males Require Estrogen Signaling Too: Sexual Dimorphism in the Regulation of Glucose Homeostasis by Nuclear ERα

Females are more protected against insulin resistance and cardiovascular disease compared with males of the same age or BMI, and this relative protection is diminished in postmenopausal women, attributable to a loss in estrogen signaling (1–4). Evidence from genetic animal models further points to the beneficial effects of estrogen signaling via estrogen receptor-α (ERα) in regard to glucose homeostasis in health and metabolic diseases (5). Mice harboring whole-body deletion of ERα (ERαKO) reveal that estrogen signaling through this receptor regulates glucose homeostasis in part by modulating hepatic insulin sensitivity (6) and glucose uptake in the skeletal muscle and adipose tissue (7). Moreover, ERα signaling has been demonstrated to promote pancreatic insulin-producing β-cell function, survival, and proliferation, as well as protection from development of diabetes in mice of both sexes (8–10). Thus, estrogen is generally thought to positively regulate glucose homeostasis primarily through ERα, which is expressed in both male and female tissues, but the respective importance of nuclear and membrane ERα pools in the regulation of glucose homeostasis is not clear.

ERα is predominantly characterized as a nuclear receptor and thus noted for exclusively regulating transcription of target genes. However, extranuclear ERα action or membrane compartment–initiated estrogen signaling is now widely accepted to activate different pathways occurring in the cytoplasm and nucleus. For example, membrane localization of ERα (which accounts for ∼5–10% of the ERα pool depending on cell type) facilitates membrane-initiated signaling events important for reproduction and vascular physiology (11) as well as β-cell function and survival (12). Although it is clear that ERα positively regulates glucose homeostasis, the distinct and overlapping contributions of the extranuclear and nuclear pool of ERα remain unknown. Therefore, mouse models specifically engineered to have only nuclear or membrane ERα action are necessary to delineate specific ERα functions related to its subcellular localization, which will give us much needed insight into the mechanisms underlying the beneficial effects of estrogen on insulin resistance and diabetes observed in both clinical and animal studies.

In this issue of Diabetes, Allard et al. (13) aimed to interrogate the relative contribution of nuclear and membrane ERα in glucose homeostasis. This research endeavor is important because their findings refine our knowledge of the actions of ERα and may impact our ability to target specific action of ERα in critical tissues, a potential avenue for the development of sex-based therapy for diabetes. First, they show that female mice expressing exclusively nuclear ERα (NOER) or membrane ERα (MOER) developed glucose intolerance by 6 months of age. However, only male MOER mice, expressing membrane but not nuclear ERα, demonstrated hyperglycemia associated with glucose intolerance (assessed via intraperitoneal injection [i.p.] glucose tolerance test). It is important to note that both female and male mice with global ERα deficiency (ERαKO) are glucose intolerant and insulin resistant based on elevated insulin levels (14) or reduced glucose infusion rate via euglycemic-hyperinsulinemic clamp (15,16). Allard et al. also reported that female MOER mice, like female ERαKO mice, exhibited resistance to the hypoglycemic effect of insulin (via i.p. insulin tolerance test [ITT]), whereas female NOER mice exhibited normal insulin sensitivity (via ITT). By contrast, all male mice (MOER, NOER, and ERαKO) exhibited comparable and relatively normal responses to an ITT (13). The authors further supported their conclusion on insulin resistance by demonstrating that both female MOER mice and female NOER mice show a decreased glucose infusion rate in a hyperinsulinemic-euglycemic clamp study, a gold-standard technique to assess insulin sensitivity. Others have also reported that female ERαKO mice displayed insulin resistance via assessment of glucose infusion rate (15). The metabolic phenotypes of mice are summarized in Supplementary Table 1. The authors conclude that loss of nuclear ERα, and to a lesser extent membrane ERα, impairs glucose homeostasis in mice of both sexes.

Impaired glucose intolerance in male MOER and ERαKO mice can be explained in part by impairment in pancreatic β-cell insulin secretion, as these mice display normal insulin sensitivity. Allard et al. (13) show that when male MOER mice were challenged for glucose-stimulated insulin secretion (GSIS) in vivo, they demonstrated a blunted response only during the first 5 min post–i.p. glucose injection (first phase of insulin secretion) but increased insulin levels after 30 min postglucose (second phase). They go on to demonstrate that the first-phase insulin secretion defect in male MOER mice is not observed in static incubation of islets in vitro; the authors concluded that the in vivo defect in first-phase insulin secretion in GSIS is independent of the loss of nuclear ERα in β-cells and secondary to the loss of ERα in extraislet tissues, possibly impairing islet function via a neural factor yet to be defined. Allard et al. (13) did not elaborate on possible explanations for the increased insulin levels observed during the second phase of insulin secretion in male MOER mice. Next, the authors show that female MOER mice demonstrated normal insulin secretion during the first phase of GSIS in vivo. Like male MOER mice (with normal insulin sensitivity), female MOER mice (insulin resistant) demonstrated hyperinsulinemia in the second phase of insulin secretion. Female MOER fed and fasted insulin levels, as well as basal plasma insulin prior to hyperinsulinemic-euglycemic clamp, were significantly elevated, suggesting β-cell hypersecretion. As previously mentioned, female MOER mice demonstrated significant insulin resistance. Thus, the β-cells appear to be responding to insulin resistance by increasing insulin levels. Total pancreatic insulin content in male and female MOER mice was normal, suggesting a lack of change in β-cell mass. The estimated β-cell mass between male MOER and control mice was also comparable. This is a surprise finding given the reported effects of ERα on insulin gene expression (17,18) and on β-cell survival and proliferation (9,19). It is interesting to also point out that female NOER showed normal glucose tolerance despite demonstrating insulin resistance. In this model, β-cell hypersecretion of insulin appears also to compensate for insulin resistance in NOER mice. Future studies should be directed toward identifying mechanisms by which ERα modulates temporal regulation of insulin secretion. It would be ideal in the future to use an islet perfusion technique to assess temporal dynamics (first and second phases of insulin secretion) compared to a static incubation test.

So what are the central mechanisms of impaired glucose homeostasis in the MOER mice? The central nervous system (CNS) integrates information regarding peripheral nutrient and hormonal changes and processes this information to regulate energy homeostasis. Recent findings indicate that some of the neural circuits and mechanisms underlying energy balance are also essential for the regulation of glucose homeostasis. Allard et al. (13) propose that ERα activity in the CNS stimulates pancreatic insulin secretion in males but promotes liver insulin sensitivity in females. Female MOER mice exhibited central insulin resistance, as determined by the failure of a central insulin infusion to activate Akt in the hippocampus or IL-6–STAT3 in the liver, leading to unsuppressed hepatic glucose production. Female ERαKO mice demonstrated hepatic insulin resistance just like female MOER and to a lesser degree female NOER (20). Hepatocyte-specific ablation of ERα (LERKO mice) has also been shown to induce insulin resistance in females (20,21). Therefore, ERα in the liver is important to whole-body and liver insulin sensitivity in female mice. While agouti-related protein (AgRP) neurons in the hypothalamus are thought to regulate hepatic glucose production (22), ERα has been shown to be completely excluded from AgRP and neuropeptide Y neurons (23). ERα is expressed in mouse proopiomelanocortin in hypothalamic (POMC) neurons (24), where its deletion leads to insulin resistance just like in female MOER mice. Therefore, it is likely that loss of ERα in POMC neurons can alter neuronal excitability of AgRP neurons, leading to changes in hepatic glucose production. Nevertheless, it is worth noting that the phenotype of the POMC ERαKO mice is not as striking as the MOER mice, suggesting that ERα action in POMC neurons is not sufficient to explain the full phenotype of the MOER mice. Thus, additional studies are needed to substantiate the role of ERα POMC neurons in the context of hepatic glucose production and insulin secretion.

In summary, Allard et al. (13) bring an important and timely first insight on the distinct role of nuclear and membrane ERα on glucose homeostasis (Fig. 1). Overall, the study brings a compelling finding that global deletion of nuclear ERα, and to a lesser extent membrane ERα, alters the central control of hepatic glucose production in female mice. However, in male mice, lack of nuclear ERα predominantly impairs the central regulation of insulin secretion. One of the key unanswered questions emerging from this study is which neuronal population is mediating these sexually dimorphic effects of ERα on glucose homeostasis? It will also be critical to show direct CNS manipulation of MOER and NOER to confirm that this is the central mechanism of ERα action in this context. Moreover, it will be important in the future to assess the phenotypes of mice with temporal and conditional tissue deletion of MOER and NOER to demonstrate direct causal effects. Subsequently, we can pinpoint viable tissue-specific targets for sex-based therapies for diabetes.