Maternal hypothyroidism in mice influences glucose metabolism in adult offspring

Animal studies were conducted according to the European guidelines for animal welfare (2010/63/EU). Protocols were approved by the Institutional Animal Care and Use Committee (CEEA-LR-1434) and the French Ministry of Agriculture (APAFIS#13044). Mice were housed in a conventional facility on a 12 h light/12 h dark cycle and were given chow and water ad libitum. FVB/NJ mice were purchased from Janvier-SAS (Le Genest-St-Isle, France). Hypothyroidism in gestating mice was induced by feeding animals with an iodine-deficient diet (low-iodine diet, LID) supplemented with 0.15% propylthiouracil (PTU) (LID +0.15% PTU diet TD.95125, Envigo, USA) [19] from the first day post-coitus. Offspring were subsequently fed with normal diet (ND) until age 8–10 weeks and then fed with either ND or HFD (63% energy from fat) (Safe Diets, France) for 8 weeks. Analysed offspring were from at least three independent breeding pairs per group. Offspring numbers were evenly distributed between dams for analysis and allocated to treatment groups in a randomised manner to ensure that all states were represented in the different experiment arms. Second generation animals originated from two different male or three different female breeders born to two different hypothyroid mothers. The two males were each bred to three different females, and the three females each bred to one male, specifically purchased for this purpose. Experimenters were blind to group assignment during data analysis. No data were excluded unless animals died during experimentation. IPGTT, ITT and glucose-stimulated insulin secretion in vivo (GSIS) tests were as described [20, 21]. We chose a glucose dose of 3 g/kg body weight to ensure full beta cell challenge and generation of an insulin peak [22]. Mice were euthanised by decapitation after asphyxiation with CO₂. Total trunk blood was collected and total T₄ was measured in plasma in duplicate using a total T₄ Elisa kit (EIA-1781, DRG International). T₄ concentrations were measured in plasma (50 μl) isolated from blood collected at the tail for a limited number of dams, to ensure proper nursing of pups. To measure T₄ concentrations in neonates, pooling of blood from a minimum of four animals at day 0 postpartum (P0) was necessary for each measurement. Litters from two to three different dams, corresponding to a total of 15 to 17 P0 pups were used. A flow chart of experimental design and mouse selection for the different assays in depicted in electronic supplementary material (ESM) Fig. 1.

Pancreas preparation and antibody labelling were as described [21]. Briefly, pancreases were fixed overnight in 4% paraformaldehyde and sliced (100 μm slices) on a vibratome (Leica, Germany) before immunostaining. Antibodies used were: rabbit anti-Ki67 (1:200, CliniSciences, France), guinea pig anti-insulin (1:400, Abcam, UK) and mouse anti-glucagon (1:200, Sigma, USA). Nuclei were labelled using DAPI (Sigma). Images (top 20 μm at the surface of slices) were acquired using a Zeiss (Germany) LSM 780 confocal microscope. Images were analysed using Imaris (Bitplane, UK), Volocity (Perkin Elmer, USA) and ImageJ (NIH, USA). For quantifications, four slices were randomly selected from at least three animals/group, and all islets present analysed. A priori, this is sufficiently-powered to detect a minimum 1.2-fold difference with an SD of 40%, a power of 0.9, and α = 0.05 (G*Power 3.1, Germany). The proportion of proliferative beta cells was obtained by dividing the number of Ki67⁺ nuclei by the total number of nuclei of insulin⁺ cells in islets, as described [21]. Beta cell mass was measured by dividing the area occupied by insulin⁺ cells by the total area of the tissue.

Islets were hand-picked after collagenase digestion of the whole pancreas [23], and cultured (5% CO₂, 37°C) in RPMI medium containing 10% FCS, 100 units/ml penicillin, and 100 μg/ml streptomycin. Islets were loaded with Fluo-8 (AAT Bioquest, USA; Cat#21082) dissolved in DMSO containing 20% pluronic acid. Islets were then imaged using an X-Light spinning disk system (Crest, Italy) coupled to a Ti-E base (Nikon, Japan) and ×10/0.4/air objective (Nikon, Japan). Excitation was delivered at λ = 458–482 nm using a Spectra X-light engine (Lumencor, USA), with emitted signals detected at λ = 500–550 nm using a Delta Evolve EM-CCD (Teledyne Photometrics, USA). Imaging buffer contained (in mmol/l) 120 NaCl, 4.8 KCl, 24 NaHCO₃, 0.5 Na₂HPO₄, 5 HEPES, 2.5 CaCl₂, 1.2 MgCl₂, and 3–17 d-glucose. Data were analysed using ImageJ, with traces presented as F/F_min where F = fluorescence at any timepoint and F_min = minimum fluorescence. No data were excluded unless the cells displayed a clear non-physiological state (i.e. impaired viability).

Pancreatic islets were hand-picked after collagenase digestion of whole pancreas, as described [23]. Total RNA from mouse islets was extracted using RNeasy microkit (Qiagen, Germany) following the manufacturer’s instructions. Reverse transcription was carried out using random hexamer oligonucleotides and SuperScriptIII Reverse Transcriptase (2000 U; Invitrogen, LifeTechnologies, USA). The reverse transcription product was diluted according to the efficiency curve and submitted in duplicates to real-time quantitative PCR using LightCycler 480 SYBR Green I Master (Roche, Switzerland) in 7500 System (Applied Biosystems, USA). Selection of housekeeping genes was performed using NormFinder software (Denmark, https://​moma.​dk/​normfinder-software) [24]. PCR reactions were performed following the conditions: 95°C for 5 min, followed by 45 cycles of 95°C for 10 s and 72°C for 30 s, and the melting curve was performed from 65°C up to 97°C for 1 min. C_t values were recorded for each gene (listed in ESM Table 1) and normalised to the geometric mean of Ppia and Mrpl32. C_t values were then expressed relative to offspring from ND-fed animals. A list of primers used is shown in ESM Table 1.

Values are represented as mean ± SEM. Statistical tests were performed using GraphPad Prism (USA). 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. Multiple comparisons were made using one-way or two-way ANOVA followed by Bonferroni’s post hoc test. A p value was considered significant at *p < 0.05, **p < 0.01, ***p < 0.001.

Gestational hypothyroidism in female mice was induced from the first day post-coitus by administration of an LID supplemented with PTU, known to block thyroid hormone synthesis and conversion [19]. This approach led to severe hypothyroidism, shown by a marked postpartum decrease in total T₄ hormone concentrations (Fig. 1a), as previously described [19, 25, 26]. Since the model is well established, confirmation of T₄ hormone concentration, which requires large amounts of plasma, was limited to a reduced number of animals to avoid stressing dams and ensure proper nursing of pups. Analysis of total T₄ concentrations in pooled blood from newborn pups showed that concentrations were comparable between animals born to euthyroid and hypothyroid mothers, suggesting normal thyroid function at birth in both groups (Fig. 1b). Although male mice born to hypothyroid mothers did not display significant changes in weight gain compared with control mice (Fig. 1c) and were euthyroid (Fig. 1d), glucose metabolism in adults (8–10 weeks) was increased. In particular, male offspring of iodine-deficient mothers presented higher fasting blood glucose concentration (Fig. 1e) but improved glucose tolerance (Fig. 1f). Insulin sensitivity was altered (Fig. 1h), fasting insulin concentrations were increased (Fig. 1g) and glucose-stimulated insulin release during IPGTT was unchanged (Fig. 1i).

In female offspring (8–10 weeks of age), 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 with control mice (Fig. 2a), and were euthyroid (Fig. 2b). Although fasting blood glucose concentrations were similar to those in control mice (Fig. 2c), female offspring of iodine-deficient mothers displayed improved glucose tolerance (Fig. 2d). However, insulin sensitivity (Fig. 2f), fasting insulin concentration (Fig. 2e) and glucose-stimulated insulin release (Fig. 2g) remained unchanged.

Morphometric analyses of pancreatic sections showed that islet area was similar in offspring born to hypothyroid and euthyroid mothers (Fig. 3a–g). However, Ki67 labelling in beta cells, a marker of proliferation, was consistently found to be ~twofold higher both in males and females born to hypothyroid mothers (Fig. 3a, d, e), without changes in islet size, alpha to beta cell ratio or beta cell mass (Fig. 3b, c, f–i). Frequency distribution of islet sizes did not show a clear shift in islet sizes, since only a small increase in the percentage of very large islets in animals born to hypothyroid mothers could be evidenced (ESM Fig. 2).

Suggesting that defects in insulin secretion are distal to the triggering pathway, glucose- and KCl-stimulated Ca²⁺ rises were unchanged in islets from male animals born to hypothyroid mothers (Fig. 3j–l). Furthermore, the abundance of selected 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 expression levels (Fig. 3m).

We next explored whether gestational hypothyroidism would influence compensatory responses to metabolic stress in adult male offspring. Compared with controls, animals born to hypothyroid mothers displayed increased weight gain following HFD feeding (Fig. 4a, b). However, weight gain was comparable when the same animals were fed ND for an equivalent 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 offspring born to hypothyroid mothers, despite similar insulin sensitivity to age-matched controls (Fig. 4f). An analogous effect was seen in fasting insulin concentrations (Fig. 4e), with pronounced basal hyperinsulinaemia present in offspring born to hypothyroid mothers, despite the absence of any changes in insulin concentrations following glucose challenge (Fig. 4g). Interestingly, although maternal hypothyroidism altered glucose homeostasis in young adult offspring, differences in insulin sensitivity and glucose intolerance were not significant in older animals (16–18 weeks of age) kept on standard chow (Fig. 4). We note however that seven animals were required to reach statistical significance for female animals, so the lack of effect in male animals might reflect a sample size issue. At the morphological level, HFD increased islet size and beta cell proliferation to a similar extent in offspring from both hypothyroid and euthyroid mothers (Fig. 5a–c), as expected [27]. No differences in glucose- or KCl-stimulated Ca²⁺ rises were detected between the two groups of animals (Fig. 5d–f), suggesting that altered ionic fluxes were unlikely to contribute to the impaired glucose tolerance observed in HFD-fed offspring from hypothyroid mothers.

We finally investigated whether alterations in glucose metabolism persisted in a second generation of animals. To do this, glucose tolerance, insulin resistance, fasting insulin concentrations and beta cell proliferation were all 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 with control age-matched animals (Fig. 6a, b), glucose tolerance remained improved (Fig. 6c). This change was despite normal insulin sensitivity (Fig. 6d), decreased fasting insulin concentration (Fig. 6e) and normal insulin secretion in response to intraperitoneal glucose (Fig. 6f). Second generation female offspring also presented similar weight compared with control aged-matched animals (Fig. 6g). By contrast to males, glucose tolerance was unchanged (Fig. 6i), despite increased fasting blood glucose (Fig. 6h) and decreased insulin sensitivity (Fig. 6j), decreased fasting insulin concentration (Fig. 6k) and normal glucose-stimulated insulin release Fig. 6l). Finally, islet size and beta cell proliferation were unchanged in both male and female offspring (Fig. 6m–p). Comparable results were obtained in adult offspring (8–10 weeks of age) originating from female mice born to hypothyroid mothers (ESM Fig. 3). Thus, gestational hypothyroidism induces transgenerational changes in metabolism in both paternal and maternal lines, with differential effects in male and female offspring.