Maternal hypothyroidism in mice influences glucose metabolism in adult offspring

Aims/hypothesis During pregnancy, maternal metabolic disease and hormonal imbalance may alter fetal beta cell development and/or proliferation, thus leading to an increased risk for developing type 2 diabetes in adulthood. Although thyroid hormones play an important role in fetal endocrine pancreas development, the impact of maternal hypothyroidism on glucose homeostasis in adult offspring remains poorly understood. Methods We investigated this using a mouse model of hypothyroidism, induced by administration of an iodine-deficient diet supplemented with propylthiouracil during gestation. Results Here, we show that, when fed normal chow, adult mice born to hypothyroid mothers were more glucose-tolerant due to beta cell hyperproliferation (two- to threefold increase in Ki67-positive beta cells) and increased insulin sensitivity. However, following 8 weeks of high-fat feeding, these offspring gained 20% more body weight, became profoundly hyperinsulinaemic (with a 50% increase in fasting insulin concentration), insulin-resistant and glucose-intolerant compared with controls from euthyroid mothers. Furthermore, altered glucose metabolism was maintained in a second generation of animals. Conclusions/interpretation Therefore, gestational hypothyroidism induces long-term alterations in endocrine pancreas function, which may have implications for type 2 diabetes prevention in affected individuals. Electronic supplementary material The online version of this article (10.1007/s00125-020-05172-x) contains peer-reviewed but unedited supplementary material, which is available to authorised users.


INTRODUCTION
Type 2 diabetes (T2D) and hypothyroidism are two major public health issues, affecting ~ 9 % and 2 %, respectively, of the population worldwide (1,2). These endocrine pathologies alter whole body metabolism and can be sometimes related, without presenting a common etiology (3,4). T2D arises from a complex interplay between genetic and environmental factors (5). In particular, the fetal environment plays a key role in the establishment of a functional beta cell mass (6). Changes in the intra-uterine milieu can modify beta cell differentiation and proliferation in the fetus, leading to long-term effects on glucose metabolism (7). Different maternal conditions alter the circulating levels of nutrients and hormones, which might impact beta cell development in utero. First, pre-existing metabolic disorders, such as malnutrition, obesity and diabetes, have been linked to increased susceptibility of the offsprings to chronic diseases, such as hypertension and diabetes (7,8). In mice, maternal diabetes induces fetal hyperglycemia and hyperinsulinemia through accelerated endocrine pancreas development, predisposing to T2D at later stages (9). Second, gestation itself leads to important metabolic and hormonal modifications. For instance, gestational diabetes, occurring in 13 % of pregnancies (10), alters endocrine pancreas maturation in the fetus and constitutes a risk factor for T2D in adulthood (9). In addition, gestation increases demand on thyroid hormones in the mother, leading to hypothyroidism in 0.5 % of pregnancies (11).
As master metabolic gatekeepers, the thyroid hormones thyroxine (T4) and triiodothyronine (T3) play an essential role in metabolism and fetal development. Maternal hypothyroidism is associated with deficits in fetal growth and cardiac, nervous and bone maturation (12,13). Such dramatic effects result from a complete dependence of the fetus on 4 maternal thyroid hormones until mid-gestation in mice and second trimester of pregnancy in humans (14,15), and a continued influence of maternal thyroid hormones at later stages (14).
At the level of the pancreas, different studies have demonstrated important effects of thyroid hormones on beta cell development and maturation (16,17). These effects can be direct, through specific interactions with cognate receptors on beta cells (18), or indirect, through modification of the availability of growth factors (16,19), thereby altering glucose metabolism and insulin resistance (3). Although a recent study showed that fetal hypothyroidism in sheep leads to increased beta cell proliferation and hyperinsulinemia in the fetus (17), the consequences of gestational hypothyroidism on beta cell function in adult offsprings remains unexplored. Thus, we sought to investigate the effects of gestational hypothyroidism on beta cell maturity and function, glucose metabolism, and susceptibility to metabolic stress such as high-fat diet (HFD) in adult mice offsprings and their descendants.

Maternal hypothyroidism alters glucose homeostasis in adult offsprings
Gestational hypothyroidism in female mice was induced from the first day post-coitus by administration of an iodine-deficient diet supplemented with propylthiouracil (PTU), known to block thyroid hormone synthesis and conversion (20). This approach led to severe hypothyroidism, shown by a marked post-partum decrease in total T4 hormone levels ( Fig.   1A), as previously described (20, 21). Although male mice born to hypothyroid mothers did not display significant changes in weight gain compared to control mice (Fig. 1B) and were euthyroid (Fig. 1C), glucose metabolism in adults (8-10 weeks) was altered. In particular, male offsprings of iodine-deficient mothers presented higher fasting blood glucose levels ( Fig. 1D) but improved glucose tolerance (Fig. 1E).Insulin sensitivity was also improved (Fig.   1F), depicted by increased fasting insulin levels ( Fig. 1G) and lower glucose-stimulated insulin release during intraperitoneal glucose tolerance test (IPGTT) (Fig. 1H).
In female offsprings, alterations in glucose metabolism were also observed, albeit less pronounced (Fig. 2). Female mice born to hypothyroid mothers presented similar weight gain from birth to adulthood compared to control mice ( Fig. 2A), and were euthyroid (Fig. 2B).
Although fasting blood glucose levels were similar to those in control mice (Fig. 2C), female offsprings of iodine-deficient mothers displayed improved glucose tolerance (Fig. 2D).

Gestational hypothyroidism alters beta cell proliferation
Morphometric analyses of pancreatic sections showed that islet area was similar in offsprings born to hypothyroid and euthyroid mothers ( Fig. 3A-B). However, Ki67 labeling in beta cells, a marker of proliferation, was found to be ~2-fold higher both in males and females born to hypothyroid mothers (Fig. 3A, C).
Suggesting that defects in insulin secretion are distal to the triggering pathway, glucose-and KCl-stimulated Ca 2+ rises were unchanged in islets from male animals born to hypothyroid mothers ( Fig. 3D-F). Furthermore, the islet mRNA abundance of genes responsible for maintenance of beta cell identity and maturity remained unchanged (Mafa, Pdx1, Rfx6, Pax6, Ins1, Ins2, Slc2a2, Gck and Glp1r), with the exception of increased Nkx6-1 mRNA levels (Fig. 3G).

Gestational hypothyroidism renders offsprings more susceptible to metabolic stress
We next explored whether gestational hypothyroidism would influence compensatory responses to metabolic stress on adult male offsprings. Animals born to hypothyroid mothers displayed increased weight gain following high fat diet feeding ( Fig. 4A-B), whereas weight gain was comparable when animals were on normal chow for the same duration ( Fig. 4A-B).
While HFD increased fasting glucose (Fig. 4C), and induced glucose intolerance (Fig. 4D) in all animals examined, the defect was most severe in offsprings born to hypothyroid mothers, despite similar insulin sensitivity to age-matched controls (Fig. 4E). A similar trend was seen in fasting insulin levels (Fig. 4F), with pronounced hyperinsulinemia present in offspring born to hypothyroid mothers, despite the absence of change in insulin levels following glucose challenge ( Fig. 4G).
At the morphological level, HFD increased islet size and beta cell proliferation to a similar extent in offsprings from both hypothyroid or euthyroid mothers ( Fig. 5A-C), as 7 expected (22). Using multicellular live Ca 2+ imaging of intact islets (23), no differences in glucose-or KCl-stimulated Ca 2+ rises were detected between the two groups of animals ( Fig.   5D-F), suggesting that HFD does not affect preferentially ionic fluxes or glucoseresponsiveness in males born to hypothyroid mothers.

The effects of gestational hypothyroidism on glucose homeostasis are transgenerational
We finally investigated whether alterations in glucose metabolism persisted in a second generation of animals. To do this, glucose tolerance, insulin resistance, fasting insulin levels and beta cell proliferation were assessed in adult offspring (8-10 weeks of age) originating from male mice born to hypothyroid mothers. While the second generation of males presented similar weight and fasting blood glucose compared to control aged-matched animals (

Circulating factors in utero can influence fetal endocrine pancreas development and lead to
life-long alterations in glucose metabolism. Since gestation modulates thyroid hormone levels which are known to play an important role in beta cell development and maturation (12,17), we sought to investigate whether maternal hypothyroidism influences glucose homeostasis in adult offspring. We found that gestational hypothyroidism induced by iodine-deficient diet increased beta cell proliferation, altered glucose metabolism and increased severity of high-fat diet-induced obesity in the offspring, without altering beta cell maturity and functional responses. Furthermore, alterations in glucose metabolism were maintained in a second generation of adults. We thus provide evidence that maternal hypothyroidism exerts transgenerational effects on metabolism, manifested by glucose intolerance in response to metabolic stress.
Hypothyroidism is one of the most common endocrine diseases during pregnancy and is mainly linked to dietary iodine deficiency, especially in low-middle income countries (3).
Thus, iodine deficiency in diet constitutes a reliable model to induce congenital hypothyroidism through severe decreases in circulating total T4 levels during gestation (24).
However, corresponding studies in mice remain scarce. While exposure to hypothyroidism in utero has been reported to influence growth in other rodents (25, 26), we could not detect significant differences in body mass between mice born to euthyroid or hypothyroid mothers.
While this may reflect the model used, we note that intrauterine growth restriction does not necessarily correlate with altered body weights in neonates (27). Indeed, in sheep, hypothyroidism in utero induced pancreatic beta cell proliferation and hyperinsulinaemia in the fetus (17), which would be expected to maintain growth rate.
1 0 a combination of modifications in different organs. For instance, congenital hypothyroidism has been shown to alter liver development (32), and modify glucose transporter expression, impairing glucose sensing in glucose-sensitive organs, including the liver and metabolic brain (33). Further studies will be needed to explore both these possibilities.
In adults, beta cell proliferation is triggered in response to increased metabolic demand such as gestation and high-fat diet feeding (22, 34). Although triiodothyronine stimulates proliferation of rat beta cell lines (35), whether thyroid hormones contribute to beta cell proliferation in response to demand remains unclear. During fetal development, the prepartum surge in thyroid hormone is thought to induce a switch from beta cell proliferation to functional maturation (16,36), thus explaining the maintenance of beta cell proliferation in islets of hypothyroid sheep fetuses (17). While similar mechanisms are likely at play in offspring of hypothyroid mothers, we cannot exclude an increase in beta cell proliferation due to increased metabolic demand. The source of such increased demand is however unclear, especially since normal diet-fed offspring displayed increase insulin sensitivity. Since T3/T4 are pre-requisite for cell maturation (37), and because in vivo insulin response to glucose was decreased, we analyzed overall gene expression of key markers defining adult beta cell functional identity (38). However, we could not detect major changes in adult offspring from hypothyroid mothers, suggesting that beta cell de-differentiation/or lack of maturation is not a feature here.
In addition to altered glucose metabolism and increased beta cell proliferation, maternal hypothyroidism increased susceptibility to HFD-induced metabolic stress in adult male offspring. This is in line with previous data showing that alterations in endocrine pancreas development can induce long-term consequences for glucose metabolism (7,39).
The results here support that maternal hypothyroidism may increase risk of type 2 diabetes development in later life. This increased susceptibility may be linked to exacerbated HFD- 1 1 induced hyperinsulinemia, which has previously been shown to drive insulin resistance and diet-induced obesity (40). Again, changes were independent of Ca 2+ channel activity, suggesting that insulin secretory defect may lie distal to the triggering pathway. Although female offspring were not analyzed, similar results may likely be obtained, since congenital hypothyroidism also affected glucose metabolism in female offspring.
Finally, we saw that altered glucose metabolism persisted in a second generation of offspring, albeit to a lesser extent, suggesting the presence of epigenetic changes. Such changes are likely to be imprinted due to thyroid hormone deprivation during fetal development, , since epigenetic reprogramming occurs during gametogenesis and early embryogenesis (41), before being transmitted to the next generation. Since both liver and pancreas are affected by similar signaling pathways during development and both organs display remarkable plasticity following insult in adults, epigenetic markers likely affect other organs than the endocrine pancreas (42). We concede, however, that identification of these epigenetic markers is needed, and that a multitude of other mechanisms may also be involved in altered glucose homeostasis following changes in thyroid hormone levels.
In summary, we show that gestational hypothyroidism induces trans-generational effects on glucose metabolism in the offspring, which may affect predisposition to T2D development in response to metabolic stress. Offspring were subsequently fed with normal diet (ND) until age 8-10 weeks and then fed with either normal diet (ND) or high fat diet (HFD, 63% calories from fat) (Safe Diets) for 8 weeks. Analyzed offspring were from at least three independent breeding pairs per group.
Second generation animals originated from two different male breeders born to hypothyroid mothers. Intra-peritoneal glucose tolerance tests (IPGTT), insulin tolerance tests (ITT) and glucose-stimulated insulin secretion in vivo (GSIS) tests were as described (43, 44). Mice were euthanized by decapitation after asphyxiation with CO 2 . Total trunk blood was collected and total T4 was measured in plasma in duplicate using a total T4 Elisa kit (EIA-1781, DRG International).

Confocal imaging and image analysis
Pancreas preparation and antibody labeling were as described (44). Briefly, pancreas were fixed overnight in 4% paraformaldehyde and sliced on a vibratome (Leica) before immunestaining. Antibodies used were: rabbit anti-Ki67 (1:200, CliniSciences), guinea-pig antiinsulin (1:400, Abcam), mouse anti-glucagon (1:200, Sigma). Nuclei were labeled using dapi 1 3 (Sigma). Images were acquired using a Zeiss LSM 780 confocal microscope. Images were analyzed using Imaris (Bitplane), Volocity (Perkin Elmer) and ImageJ (NIH). For quantifications, four slices were randomly selected from at least three animals/group, and all islets present analyzed. A priori, this is sufficiently-powered to detect a minimum 1.2-fold difference with a SD of 40%, a power of 0.9, and alpha = 0.05 (G*Power 3.1). The proportion of proliferative beta cells was obtained by dividing number of Ki67+ nuclei by total number of nuclei of insulin+ cells in islets, as described (44).

Islet isolation and real-time quantitative RT-PCR
Pancreatic islets were hand-picked after collagenase digestion of whole pancreas, as described (45). Total RNA from mouse islets was extracted using RNeasy microkit (Qiagen) following the manufacturer's instructions. Reverse transcription was carried out using random hexamer oligonucleotides and SuperScriptIII Reverse Transcriptase (2,000U; Invitrogen, 1 4 LifeTechnologies, EUA  Table S1) and normalized to the geometric mean of Ppia and Mrlp32. Ct values were then expressed relative to offspring from normal diet-fed animals (ND). A list of primers used is shown in Table S1.

Statistical analysis
Values are represented as mean ± SEM. Statistical tests were performed using GraphPad Prism. Normality was tested using D'Agostino-Pearson test, and comparisons were made using either unpaired Student's t-test, or two-tailed Mann-Whitney U-test, as appropriate.