Abstract

   Numerous studies have suggested potential associations between maternal
   subclinical hypothyroidism (SCH) and adverse metabolic outcomes in
   offspring, however, the underlying mechanism remains unclear. In this
   study, we generated a maternal SCH mouse model by administering 50 ppm
   6-propyl-2-thiouracil (PTU) in the drinking water of pregnant mice
   until delivery. This model was used to investigate the mechanisms
   influencing glucose metabolism in offspring. RNA sequencing (RNA-seq)
   revealed a substantial increase in ARRDC3 expression in the livers of
   the offspring of the SCH model mice, which may contribute to insulin
   resistance. Additionally, the phosphorylation levels of key proteins in
   the insulin signalling pathway, such as protein kinase B (Akt),
   glycogen synthase kinase 3 beta (GSK-3β), and Forkhead box protein O1
   (FoxO1), were correspondingly reduced in the SCH offspring. Moreover,
   overexpression of ARRDC3 in Hepa1‒6 cells suppressed the
   Akt/GSK-3β/FoxO1 signalling pathway and increased the expression of
   glucose-6-phosphatase (G6Pase) and phosphoenolpyruvate carboxykinase
   (PEPCK), which was consistent with the molecular changes observed in
   SCH offspring. Our results also indicated that the upregulation of
   ARRDC3 in SCH offspring may result from increased H3K27 acetylation of
   the ARRDC3 promoter region, driven by elevated expression of P300.
   Importantly, adequate L-T4 supplementation during pregnancy improved
   insulin sensitivity and reversed the molecular alterations in the
   insulin signalling pathway observed in SCH offspring. In conclusion,
   exposure to intrauterine SCH resulted in altering the P300-ARRDC3 axis
   in offspring and impaired insulin sensitivity by disrupting the
   Akt/GSK-3β/FoxO1 signalling pathway. Timely L-T4 supplementation during
   pregnancy is an effective strategy to prevent insulin resistance in
   offspring of SCH mothers. This study elucidates potential molecular
   mechanisms behind insulin resistance in SCH offspring and suggests
   novel therapeutic targets for treating metabolic disorders related to
   maternal SCH.

   Keywords: Hypothyroidism, Pregnancy, ARRDC3, H3K27Ac, Glucose metabolic
   dysfunction

Highlights

     * •
       Maternal subclinical hypothyroidism (SCH) affects glucose
       metabolism in offspring.
     * •
       RNA-seq revealed increased ARRDC3 expression in the liver of SCH
       offspring.
     * •
       Reduced phosphorylation levels of Akt, GSK-3β, and FoxO1 and
       elevated expression levels of G6Pase and PEPCK were observed in SCH
       offspring.
     * •
       Overexpression of ARRDC3 suppressed the Akt/GSK-3β/FoxO1 pathway in
       Hepa1-6 cells.
     * •
       Maternal SCH enhanced acetylation of the ARRDC3 promoter via
       increased P300 expression.
     * •
       P300-ARRDC3 pathway mediates decreased insulin sensitivity in
       offspring of maternal SCH by modulating the Akt/GSK-3β/FoxO1
       signaling pathway.

1. Introduction

   Dr. David Barker proposed the Foetal Origins of Adult Disease (FOAD)
   theory, which posits that foetuses and neonates are vulnerable to
   adverse intrauterine or early developmental environments, increasing
   their susceptibility to chronic diseases later in life [[39]1]. A
   detrimental maternal metabolic environment during pregnancy can have
   lasting effects on the offspring's health, including an elevated risk
   of developing chronic metabolic conditions such as obesity, type 2
   diabetes (T2DM), and cardiovascular disease [[40]2,[41]3]. Maternal
   subclinical hypothyroidism (SCH)—characterized by elevated serum
   thyroid-stimulating hormone (TSH) levels with normal serum free T4
   levels—occurs in 3.5 % to 18 % of pregnancies, depending on the
   diagnostic criteria used, and is more common than overt hypothyroidism
   [[42]4]. Recent evidence suggests that maternal hypothyroidism can
   induce the metabolic phenotype in offspring. Studies in rodents have
   shown that offspring exposed to maternal hypothyroidism exhibit
   increased weight, hyperglycaemia, hyperinsulinaemia, insulin
   resistance, hypertriglyceridaemia, hypercholesterolaemia, and
   hyperleptinaemia [[43][5], [44][6], [45][7], [46][8], [47][9]], these
   conditions may worsen when the offspring are fed a hypercaloric diet
   [[48]10,[49]11]. However, relatively few studies have explored the
   long-term effects of maternal SCH on the metabolic parameters of
   offspring. Therefore, this study aimed to investigate the impacts of
   maternal SCH on glucose metabolism in adult progeny and explore the
   molecular mechanisms underlying these metabolic changes in SCH
   offspring.

   The liver plays a critical role in glucose homeostasis, producing
   glucose through glycogenolysis and gluconeogenesis during fasting and
   converting glucose into glycogen in response to feeding [[50]12].
   Insulin insufficiency disrupts glucose homeostasis, resulting in
   elevated blood glucose levels during fasting and the depletion of
   glycogen stores in the liver [[51]13,[52]14]. The insulin signalling
   pathway consists primarily of two classical branches: the Ras/MAP
   kinase pathway, which regulates cell growth and differentiation, and
   the PI3-kinase/Akt pathway, which mediates most metabolic effects
   [[53]15,[54]16]. Activation of the insulin receptor (INSR) by insulin
   triggers phosphorylation of insulin receptor substrates (IRS) on
   tyrosine residues, which serve as binding sites for the p85 regulatory
   subunit of PI3K [[55]17]. PI3K phosphorylates phosphatidylinositol
   4,5-bisphosphate (PI45P2) to generate phosphatidylinositol
   3,4,5-trisphosphate (PIP3), while PIP3 is a key molecule that recruits
   and activates Akt [[56]18]. Activated Akt phosphorylates Forkhead box
   protein O1 (FoxO1), leading to its exclusion from the nucleus and a
   consequent reduction in the transcriptional activity of
   phosphoenolpyruvate carboxykinase (PEPCK) and glucose-6-phosphatase
   (G6Pase) [[57]19]. Moreover, AKT-mediated phosphorylation of glycogen
   synthase kinase 3 beta (GSK-3β) at Ser9 alleviates its inhibitory
   effect on glycogen synthase, promoting glycogen synthesis [[58]20].

   Recently, a novel group of arrestins known as α-arrestins has been
   identified [[59]21,[60]22]. These proteins play important roles in
   maintaining membrane protein homeostasis and global cellular metabolism
   [[61]23]. The α-arrestin family includes arrestin domain-containing
   proteins (ARRDC) 1 through 5 and thioredoxin-interacting protein
   (TXNIP) [[62]21,[63]22]. Among these, ARRDC3 represents a newly
   discovered α-arrestin involved in inflammation, cancer, obesity, and
   glucose and lipid metabolism [[64]24,[65]25]. ARRDC3-null rodents
   exhibit increased insulin sensitivity, a lower propensity to gain
   weight, and reduced hepatic steatosis [[66]25]. In hyperinsulinaemic
   states, ARRDC3 expression in the liver is upregulated and is dependent
   on insulin receptor signalling [[67]26]. Studies have shown that ARRDC3
   likely interacts with phosphatidylinositol 4,5-bisphosphate (PI45P2)
   [[68]27], which is involved in the recruitment and activation of Akt.
   While it has been shown that ARRDC3 reduces insulin sensitivity by
   binding directly to the insulin receptor in the liver [[69]26], the
   precise role of ARRDC3 in regulating the insulin signalling pathway and
   inducing insulin resistance in SCH offspring remains unclear [[70]24].

   Epigenetics refers to the modification of heritable gene expression
   patterns without altering the underlying DNA sequence [[71]28].
   Numerous studies have suggested that exposure to environmental factors
   such as chemicals, drugs, stress, or infections during specific
   sensitive periods of intrauterine development or early childhood can
   influence gene expression and increase susceptibility to adult diseases
   [[72][29], [73][30], [74][31]]. Epigenetic modifications include DNA
   methylation, histone modification, and noncoding RNAs [[75]32,[76]33].
   Among these, histone acetylation is a reversible post-translational
   modification that regulates gene expression by altering chromatin
   structure, making it more accessible for transcription factors to bind
   [[77]34]. A well-studied epigenetic mark is the acetylation of histone
   H3 at lysine 27 (H3K27Ac), which has been extensively characterized for
   its role in gene regulation. Chen et al. reported that a chronic
   maternal high-fat diet (HFD) leads to changes in the methylation and
   acetylation of H3K27, thereby influencing the expression of genes
   involved in osteoblast differentiation [[78]35].

   In this study, we established a mouse model of maternal SCH to
   investigate the alterations in glucose metabolism and the underlying
   epigenetic regulatory mechanisms in offspring. The primary objectives
   were to elucidate the impact of maternal SCH on insulin sensitivity and
   identify key molecules involved in the induction of insulin resistance
   in offspring. Our findings offer new insights and potential therapeutic
   targets for the prevention and treatment of metabolic abnormalities in
   offspring affected by maternal SCH.

2. Materials and methods

2.1. Animals

   C57BL/6J mice were obtained from the Experimental Animal Center at
   Soochow University. All animal experiments were approved by the Soochow
   University Institutional Animal Care and Use Committee (approval
   number: SYXK-2022-0043). The mice were housed at 22 ± 2 °C in a room
   with a 12-h light/dark cycle, with chow and water ad libitum.
   Appropriate anaesthesia was administered prior to all invasive
   procedures. Pregnant mice were randomly divided into two groups: the
   SCH group, receiving 6-propyl-2-thiouracil (PTU) at 50 μg/ml (50 ppm,
   P3755, Sigma, USA) in drinking water, and a control group receiving
   standard water [[79]36,[80]37]. PTU supplementation was ceased at pup
   birth. Pups' body weights were measured weekly from weaning. Levels of
   serum TSH, free T3 (FT3), and free T4 (FT4) were determined via ELISA
   kits (YJ063200, YJ301038, YJ057763, Elabscience, China). Given that
   estrogen may contribute to sex-specific differences in glucose
   tolerance related to fetal hypothyroidism [[81]38], only male offspring
   were selected for this study.

2.2. LT4 treatment

   L-T4 is a standard treatment for hypothyroidism. The pregnant mice were
   randomly assigned to three groups: the CON + VEH, SCH + VEH, and
   SCH + LT4 groups. LT4 (0.1 μg/ml, S2599, Selleck Chemicals, USA) or
   vehicle (VEH) was added to their drinking water on the basis of
   previously reported dosing until delivery [[82]39].

3. Metabolic analyses

3.1. Glucose, insulin, and pyruvate tolerance test

   Metabolic assessments were conducted at 16 weeks of age. For the
   intraperitoneal glucose tolerance test (IPGTT) and intraperitoneal
   pyruvate tolerance test (IPPTT), mice were fasted overnight (>16 h) and
   injected intraperitoneally with glucose (ST1228, Beyotime, China,
   2 g/kg body weight) [[83]40] or pyruvate (P4562, Sigma, USA, 2 g/kg
   body weight) [[84]41]. For the intraperitoneal insulin tolerance test
   (IPITT), mice were fasted for 2 h and injected intraperitoneally with
   insulin (P3376, Beyotime, China, 1 IU/kg body weight) [[85]42]. Blood
   glucose levels were measured at 0, 15, 30, 60, 90, and 120 min. Blood
   samples were collected from the tail. A single droplet of blood was
   applied onto a glucose test strip and analysed via glucometer
   (Onetouch, Johnson & Johnson). The area under the curve (AUC) was
   calculated.

3.2. Plasma insulin measurement

   Blood samples were collected from the abdominal aorta. The measurement
   of the serum insulin level was conducted via an enzyme-linked
   immunosorbent assay (ELISA) kit (F2579A, Fankew, China).

3.3. Tissue morphological analysis

   Liver tissues were fixed in 4 % formaldehyde and embedded in paraffin.
   Paraffin blocks were cut into 6 μm slices and subjected to hematoxylin
   and eosin (HE) staining and periodic acid-Schiff (PAS) staining. Images
   were captured with a TE300 microscope (Nikon, Japan). Histopathological
   analysis of whole liver sections was performed in a blinded manner. The
   extent of hepatic steatosis was calculated as a percentage of the total
   hepatic steatosis area in HE-stained liver sections via ImageJ software
   [[86]43]. PAS staining was performed according to the manufacturer's
   instructions (C0142M, Beyotime, China), with purple-stained cells
   identified as positively stained cells. Hepatic glycogen was quantified
   as the percentage of the PAS-positive area relative to the total
   section area by ImageJ [[87]44].

3.4. RNA sequencing (RNA-seq)

   Total RNA was extracted from liver tissues of two groups of mice
   (n = 5/group) using TRIzol reagent (A33251, TaKaRa, Japan). RNA was
   quantified via a NanoDrop spectrophotometer (NanoDrop, USA) and an
   Agilent 2100 bioanalyzer (Agilent, USA). RNA-seq was conducted via the
   Illumina HiSeq 4000 platform (Genewiz, China). GRCm38 was used as the
   reference genome. Expression levels of each transcript were determined
   via the fragments per kilobase of transcript sequence per million base
   pairs (FRKM) approach. The DEseq2 program was used for the analysis of
   differentially expressed genes (DEGs). DEGs between two groups were
   determined by a log2-fold change <−1 or >1 and a P value ≤ 0.05. Kyoto
   Encyclopedia of Genes and Genomes (KEGG) enrichment analysis of the
   DEGs was performed with the “clusterProfiler” R package [[88]45].

3.5. Real-time quantitative RT-PCR

   Total RNA was prepared with TRIzol reagent (A33251, TaKaRa, Japan) and
   used for cDNA synthesis via the PrimeScript RT reagent kit (K1622,
   Thermo Scientific, USA). Real-time PCR was performed via SYBR Green PCR
   Master Mix (CN830A, TaKaRa, Japan), and relative gene expression was
   calculated via the 2^−ΔΔCT method, with 18S as an internal control. The
   primer sequences are listed in [89]Supplementary Table 1.

3.6. Cell culture and treatment

   The Hepa1-6 cells were cultured in high-glucose Dulbecco's modified
   Eagle's medium (SH30243, HyClone, USA) supplemented with 10 % foetal
   bovine serum (FBS) (SV 30208, HyClone, USA) and 100 μg/mL
   penicillin‒streptomycin solution (SV30010, Cytiva, USA) at 37 °C in a
   5 % CO2 atmosphere. Transfection of pcDNA3.1 (empty vector) or
   pcDNA3.1-ARRDC3 was performed via Lipofectamine 2000 Transfection
   Reagent (11668, Invitrogen, USA) according to the reagent protocol.
   Following a 6-h incubation with the transfection mixture, cells were
   returned to fresh complete medium for gene expression. After overnight
   starvation in low glucose-DMEM (SH30021, HyClone, USA) containing 0.2 %
   bovine serum albumin (ST025, Beyotime, China), the cells were treated
   for 40 min with insulin (25 nM) in high-glucose DMEM (a schematic
   diagram of the protocol used to study ARRDC3 overexpression-induced
   insulin resistance is shown in [90]Supplementary Fig. 1).

3.7. Antibodies and immunoblot analysis

   The liver tissue and Hepa1-6 cells were dissolved in lysis buffer
   (P0013C, Beyotime, China) containing 1 % PMSF (P001, NCM Biotech,
   China). After homogenization, the lysates were clarified by
   centrifugation (16,000×g, 30 min, 4 °C). The resulting supernatant was
   then collected and used for immunoblotting. The protein concentration
   was quantified via a BCA protein assay kit (P0010, Beyotime, China).
   Protein samples were separated via sodium dodecyl
   sulphate-polyacrylamide gel electrophoresis (SDS‒PAGE) and transferred
   onto nitrocellulose membranes (Millipore, Germany). The membranes were
   incubated with 5 % nonfat milk (P0216, Beyotime, China) in TBST for 2 h
   at room temperature, followed by overnight incubation with primary
   antibodies at 4 °C. After three washes with TBST, the membranes were
   incubated with a secondary antibody conjugated to HRP for 1 h at room
   temperature. An enhanced chemiluminescence (ECL) system (Tanon, China)
   was used to visualize the immunoreactive bands, and the ImageJ software
   was used to quantify the results. The primary antibodies were as
   follows: ARRDC3 (1:1000, DF3517, Affinity Biosciences, China), PEPCK
   (1:1000, 16754-1-AP, Proteintech, China), G6Pase (1:1000, 22169-1-AP,
   Proteintech, China), p-Akt (Thr308) (1:1000, #2965, Cell Signaling
   Technology, USA), Akt (1:1000, #4691, Cell Signaling Technology, USA),
   p-FoxO1 (Ser253) (1:500, AF2305, Affinity Biosciences, China), FoxO1
   (1:4000, 18592-1-AP, Proteintech, China), p-GSK3β (Ser9) (1:1000,
   #9322, Cell Signaling Technology, USA), GSK3β (1:4000, A5007, Bimake,
   USA), H3K27Ac (1:1000, #39133, Active Motif, China), H3 (1:1000, #9717,
   Cell Signaling Technology, USA), and P300 (1:1000, AF5360, Affinity
   Biosciences, China).

3.8. ChIP assays

   Chromatin immunoprecipitation (ChIP) assays were conducted with a
   SimpleChIP Plus Enzymatic Chromatin IP Kit (#9005, Cell Signaling
   Technology, USA) according to the manufacturer's instructions. Liver
   tissue fragments were fixed with 37 % formaldehyde (8.18708, Sigma,
   USA) solution for 30 min at room temperature to facilitate chromatin
   cross-linking. The crosslinking reaction was quenched by adding glycine
   at room temperature. Subsequently, the samples were washed three times
   with precooled PBS buffer (SH30256, HyClone, USA) containing a protease
   inhibitor cocktail (PIC). After washing, tissue fragments were
   homogenized into single-cell suspensions with a homogenizer. Each
   sample was supplemented with 0.5 μL of micrococcal nuclease and
   incubated at 37 °C for 20 min to fragment the DNA into 150–900 bp
   fragments. After digestion, sonication was used to disrupt the nuclear
   membrane (20 s on/30 s off for 13 cycles). Next, the fragmented
   chromatin samples were centrifuged at 16,000×g for 10 min at 4 °C. The
   supernatant was carefully collected and appropriately diluted with ChIP
   buffer. A total of 10 μL of the diluted supernatant was saved as input
   DNA, while the remainder was incubated at 4 °C overnight on a rotator
   with anti-H3K27Ac (#39133, Active Motif, China). Rabbit IgG (#2729,
   Cell Signaling Technology, USA) served as the isotype control. Protein
   A/G was added, and the mixture was incubated for 2 h at 4 °C with
   rotation. Afterward, elution buffer was added, and the mixture was
   incubated at 65 °C for 30 min to elute the protein-DNA complexes from
   the beads. The samples were then incubated at 65 °C with 40 μg/mL
   proteinase K for 2 h and subsequently purified via spin columns
   following the manufacturer's protocol. Finally, the purified DNA was
   assayed via quantitative polymerase chain reaction (qPCR). The primer
   sequences can be found in [91]Supplementary Table 1.

3.9. DNA isolation and targeted bisulphite sequencing assay

   Genomic DNA from liver tissue was extracted using the conventional
   phenol/chloroform method. Sodium bisulphite treatment was carried out
   on the DNA samples via the EZ DNA Methylation-GOLD Kit (D5006, Zymo
   Research, USA), following the methods recommended by the manufacturer.
   The DNA concentration was measured and then diluted to 20 ng/μL in
   preparation for bisulphite amplicon sequencing (BSAS). The CpG island
   situated in the proximal promoter region of ARRDC3 was chosen based on
   the following criteria: (1) a GC content of more than 50 %, (2) a
   minimum length of 200 base pairs, and (3) an observed/expected ratio
   over 60 %. BSAS primers were meticulously constructed and are listed in
   [92]Supplementary Table 1. Following the completion of PCR
   amplification and library creation, the products were sequenced on the
   Illumina MiSeq platform. The methylation level at the CpG site was
   determined by calculating the ratio of methylated cytosines to the
   total number of cytosines that were tested.

3.10. Statistical analysis

   SPSS 20.0 (IBM SPSS, Armonk, NY, USA) was used to perform all the data
   analyses. All the data are presented as the means ± SEMs. Student's
   t-test was used for comparisons between two groups. Comparisons among
   multiple datasets were performed by one-way analysis of variance
   (ANOVA), and Bonferroni correction was used to adjust P values for
   multiple comparisons. Significant differences are displayed as follows:
   ∗P < 0.05, ∗∗P < 0.01, and ∗∗∗P < 0.001. Statistical analysis was
   performed with GraphPad Prism 8 (GraphPad, San Diego, California, USA).
   The sample size (n) for statistical tests representing biological
   replicates.

4. Results

4.1. Maternal subclinical hypothyroidism alters body weight, liver profile,
and plasma insulin levels in offspring

   To assess the impact of maternal SCH on offspring development, we
   established an SCH mouse model during pregnancy. Thyroid hormone
   testing confirmed significantly elevated TSH concentrations in the SCH
   mice, while serum FT4 and FT3 levels remained unchanged on the day of
   delivery ([93]Fig. 1A), indicating the successful establishment of the
   maternal SCH model. To explore this further, we measured thyroid
   hormones in 16-week-old offspring, finding similar concentrations of
   TSH, FT4, and FT3 between the SCH and control groups ([94]Fig. 1B).
   Additionally, the SCH offspring exhibited similar food and water intake
   ([95]Supplementary Figs. 2A and B) but had higher body weights compared
   to controls at 15 and 16 weeks ([96]Fig. 1C). Histological analysis
   revealed that SCH offspring exhibited mild but noticeable hepatic
   steatosis and reduced hepatic glycogen content compared to normal
   offspring ([97]Fig. 1D). Moreover, fasting serum insulin concentrations
   were significantly higher in the SCH offspring ([98]Fig. 1E). In
   summary, while maternal SCH does not affect the thyroid hormone levels
   of offspring, it has negative effects on their development,
   particularly on body weight, liver profile, and insulin levels.

Fig. 1.

   [99]Fig. 1
   [100]Open in a new tab

   Maternal subclinical hypothyroidism alters body weight, liver profile,
   and plasma insulin levels in offspring. (A) Serum levels of TSH, FT4,
   and FT3 of F0 on the day of delivery, (B) and F1 at postnatal 16 weeks
   (n = 7 per group). (C) Postnatal growth curve of offspring from weaning
   to 16 weeks (CON-F1: n = 15; PTU-F1: n = 22). (D) Representative images
   of hematoxylin-eosin (HE) and periodic acid Schiff staining (PAS) of
   the liver (magnification: × 400, scale bar = 25 μM) and the
   quantitative analysis of HE and PAS staining. (E) Serum insulin levels
   of CON-F1 and SCH-F1 mice were measured after fasting overnight (or
   16h) (n = 7 per group). ∗p < 0.05, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001. Data
   are mean ± SEM.

4.2. Maternal subclinical hypothyroidism alters glucose metabolism in
offspring

   Previous studies have demonstrated that maternal hypothyroidism can
   lead to hyperglycaemia, hyperinsulinaemia, and insulin resistance in
   offspring. However, limited research has been conducted on the
   metabolic effects of maternal SCH. To investigate this further, we
   examined glucose metabolism and insulin sensitivity in SCH offspring.
   During the IPGTT, SCH offspring exhibited a higher blood glucose
   response at 60 and 90 min, indicating significantly reduced glucose
   tolerance compared to the control group ([101]Fig. 2A). Insulin
   resistance and hepatic gluconeogenesis were evaluated using IPITTs and
   IPPTTs, respectively. In both tests, SCH offspring showed elevated
   blood glucose levels at 60 and 90 min ([102]Fig. 2B and C).
   Additionally, the expression levels of key gluconeogenic enzymes,
   G6Pase and PEPCK, were significantly higher in SCH offspring ([103]Fig.
   2D). Overall, these results indicate that maternal SCH causes a
   decrease in glucose tolerance and insulin sensitivity, and increases
   gluconeogenesis in progeny.

Fig. 2.

   [104]Fig. 2
   [105]Open in a new tab

   Maternal subclinical hypothyroidism alters glucose metabolism in
   offspring. (A) Glucose tolerance test (2 g/kg body weight) and AUC of
   CON-F1 and SCH-F1 mice (n = 10 per group). (B) Insulin tolerance test
   (1 IU/kg body weight) and AUC of CON-F1 and SCH-F1 mice (n = 10 per
   group). (C) Pyruvate tolerance test (2 g/kg body weight) and AUC of
   CON-F1 and SCH-F1 mice (n = 10 per group). (D) G6pase and PEPCK protein
   expression in liver tissue from CON-F1 and SCH-F1 mice was detected by
   Western blot analysis (n = 7 per group). ∗p < 0.05, ∗∗p < 0.01,
   ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001. Data are mean ± SEM.

4.3. Maternal subclinical hypothyroidism decreases insulin sensitivity in
offspring via the ARRDC3-Akt/GSK-3β/FoxO1 pathway

   To explore the molecular mechanisms underlying glucose metabolism
   dysregulation caused by maternal SCH, we performed RNA sequencing to
   identify differences in gene expression in liver tissues of SCH
   offspring. We found 41 differentially expressed genes (DEGs) between
   the SCH offspring and control groups, including ARRDC3, which showed
   significantly increased expression in the SCH-F1 (SCH offspring) group
   ([106]Fig. 3A). A heatmap of DEGs ([107]Fig. 3B) illustrates changes in
   gene expression, with red indicating increased and blue indicating
   decreased expression. Gene Ontology (GO) analysis indicated that
   maternal SCH may affect liver development in terms of biological
   processes ([108]Fig. 3C). The KEGG pathway enrichment analysis revealed
   that the predicted target genes were involved mainly in the PI3K-Akt
   and MAPK signalling pathways ([109]Fig. 3D). RT‒PCR and Western blot
   analysis confirmed the RNA-seq findings, showing elevated ARRDC3
   expression in SCH-F1 liver tissues ([110]Fig. 3E‒F). Since ARRDC3 is
   known to regulate insulin receptor signalling [[111]26] and the
   PI3K-Akt pathway is essential for insulin signalling [[112]46], we
   examined key genes in the Akt pathway in liver tissues of both groups.
   SCH-F1 exhibited lower phosphorylation levels of Akt (Thr308), FoxO1
   (Ser256), and GSK-3β (Ser9), although total Akt, FoxO1, and GSK-3β
   levels remained unchanged ([113]Fig. 3G). Similar reductions in
   phosphorylation were observed during fasting and refeeding in the SCH
   group compared to controls ([114]Supplementary Fig. 3A). These findings
   suggest that the reduction in insulin sensitivity in SCH offspring may
   be linked to the ARRDC3-Akt/GSK-3β/FoxO1 pathway.

Fig. 3.

   [115]Fig. 3
   [116]Open in a new tab

   Maternal subclinical hypothyroidism downregulated insulin sensitivity
   in offspring via ARRDC3-Akt/GSK-3β/FoxO1 pathway. (A)Volcano plot
   showing the differentially expressed genes between CON-F1 and SCH-F1
   mice, red points denote the upregulated genes and blue points indicate
   the downregulated ones in SCH-F1 compared to CON-F1 (n = 5 per group).
   (B) Heat map of differential expression genes across all 10 samples.
   (C) Gene Ontology (GO) classification for predicted target genes in
   three categories (cellular component and molecular function and
   biological process). (D) Kyoto Encyclopedia of Genes and Genomes (KEGG)
   functional enrichment analysis for the predicted target genes. (E)
   ARRDC3 mRNA and (F) protein expression levels in liver from 16-week-old
   CON-F1 and SCH-F1 mice (n = 7 per group). (G) CON-F1 and SCH-F1 mice
   were fasted overnight and then injected intraperitoneally with either
   insulin (2 IU/mouse) or PBS (phosphate-buffered saline) for 5 min.
   Representative western blots and the associated quantifications show
   phosphorylation of Akt (Thr308), FoxO1 (Ser256), and GSK-3β (Ser9)
   levels in liver (n = 6 per group). ∗p < 0.05, ∗∗p < 0.01,
   ∗∗∗∗p < 0.0001. Data are mean ± SEM.

4.4. ARRDC3 overexpression inhibits Akt/GSK-3β/FoxO1 signalling in
Hepa1-6 cells

   To further investigate whether ARRDC3 contributes to abnormal glucose
   metabolism in SCH offspring through the Akt/GSK-3β/FoxO1 signalling
   pathway, we measured protein expression in Hepa1-6 cells overexpressing
   ARRDC3. The results showed that overexpression of ARRDC3 significantly
   increased the levels of G6Pase and PEPCK ([117]Fig. 4A). Additionally,
   ARRDC3 overexpression significantly reduced the phosphorylation of Akt
   (Thr308), FoxO1 (Ser256), and GSK-3β (Ser9) in Hepa1-6 cells([118]Fig.
   4B). These findings suggest a negative correlation between ARRDC3
   expression and Akt/GSK-3β/FoxO1 signalling.

Fig. 4.

   [119]Fig. 4
   [120]Open in a new tab

   ARRDC3 overexpression inhibits Akt/GSK-3β/FoxO1 signalling in
   Hepa1-6 cells. (A) Representative Western blots and the associated
   quantifications show ARRDC3, G6pase, and PEPCK levels in hepatocyte
   lysates (n = 3 per group). (B) Representative Western blots and the
   associated quantifications show phosphorylation of Akt (Thr308), FoxO1
   (Ser256), and GSK-3β (Ser9) levels in hepatocyte lysates (n = 4 per
   group). ∗p < 0.05, ∗∗∗p < 0.001. Data are mean ± SEM.

4.5. Maternal subclinical hypothyroidism promotes ARRDC3 expression via
increased promoter H3K27 acetylation

   Growing evidence suggests that epigenetic modifications play a crucial
   role in how adverse pregnancy factors affect offspring health
   [[121]47]. To determine whether the upregulation of ARRDC3 in SCH
   offspring is regulated by epigenetic modifications, we used the
   Cistrome Data Browser (DB) to predict that ARRDC3 may be regulated by
   H3K27Ac ([122]http://cistrome.org/db). Additionally, P300, a well-known
   histone acetyltransferase (HAT), is primarily responsible for promoting
   H3K27 acetylation [[123]48]. We quantitatively determined the protein
   expression of H3K27Ac and P300 in the liver and found significant
   increases in H3K27Ac and P300 in the SCH offspring compared with the
   normal offspring ([124]Fig. 5A–C). Chromatin immunoprecipitation assays
   revealed higher enrichment of H3K27Ac in the promoter region of ARRDC3
   in SCH offspring ([125]Fig. 5D). We also evaluated the DNA methylation
   status of the CpG islands of ARRDC3 between the two groups. We selected
   the CpG island containing 18 CpG sites located within 1000 bp upstream
   of the transcription start site (TSS) of the ARRDC3 gene
   ([126]Supplementary Fig. 4A). Quality control procedures were
   implemented to validate the data quality ([127]Supplementary Fig. 4B).
   Bisulphite sequencing revealed that DNA methylation in the promoter
   region of ARRDC3 in the liver did not significantly differ between the
   two groups. ([128]Supplementary Figs. 4C and 4D). These findings
   suggest that maternal SCH promotes the H3K27 acetylation of the ARRDC3
   promoter region by increasing P300.

Fig. 5.

   [129]Fig. 5
   [130]Open in a new tab

   Maternal subclinical hypothyroidism promotes ARRDC3 expression via
   increased promoter H3K27 acetylation. (A–C) The protein expression of
   H3K27Ac and p300 in the liver of 16-week-old CON-F1 and SCH-F1 mice
   (n = 6 per group). (D) Enrichment of H3K27Ac in the ARRDC3 promoter
   region of liver by ChIP-PCR (n = 3 per group). ∗p < 0.05, ∗∗p < 0.01,
   ∗∗∗p < 0.001. Data are mean ± SEM.

4.6. Levothyroxine treatment decreases ARRDC3 expression and improves insulin
sensitivity in SCH offspring

   Levothyroxine (LT4) therapy is a common treatment for SCH and has been
   shown to improve pregnancy outcomes of patients with SCH [[131]49]. To
   assess whether LT4 supplementation could improve glucose metabolism in
   SCH offspring, pregnant mice were divided into three groups: the
   CON + VEH, SCH + VEH, and SCH + LT4 (LT4 0.1 μg/ml/day) groups. Results
   showed that serum TSH levels in the SCH-F0 group were significantly
   higher than in the SCH + LT4-F0 group, while FT3 and FT4 levels
   remained unchanged ([132]Supplementary Figs. 5A and 5C), confirming
   that LT4 supplementation restored thyroid function in SCH-F0 mice. LT4
   treatment also improved glucose tolerance and insulin sensitivity in
   SCH offspring ([133]Fig. 6A and B), as serum insulin levels were
   significantly lower in the SCH + LT4-F1 mice compared to SCH + VEH-F1
   mice ([134]Fig. 6C). Moreover, H3K27Ac and P300 levels were
   significantly reduced in SCH + LT4-F1 mice ([135]Fig. 6D), and ARRDC3
   expression was also significantly lower in this group ([136]Fig. 6E).
   LT4 supplementation increased the phosphorylation levels of Akt
   (Thr308), FoxO1 (Ser256), and GSK-3β (Ser9) ([137]Fig. 6E), indicating
   enhanced Akt signalling and improved insulin sensitivity in SCH
   offspring. These results indicate that LT4 alleviates glucose
   intolerance and insulin resistance in SCH offspring through the
   P300-ARRDC3-Akt/GSK-3β/FoxO1 pathway.

Fig. 6.

   [138]Fig. 6
   [139]Open in a new tab

   Levothyroxine treatment decreases ARRDC3 expression and improves
   insulin sensitivity in SCH offspring. (A) Blood glucose levels during
   GTT and (B) ITT were detected in CON + VEH, SCH + VEH, PTU + LT4 mice.
   The area under the curve (AUC) is shown (n = 10 per group). (C) Serum
   insulin levels in CON + VEH, SCH + VEH, PTU + LT4 mice under fasting
   conditions (n = 5 per group). (D) The expression of H3K27Ac and P300 in
   the liver of three groups was detected by the Western bolt (n = 4 per
   group). (E)Protein levels of ARRDC3, phosphorylated and total Akt,
   FoxO1, GSK3β in liver of three groups after insulin injection (2
   IU/mouse, 5min) (n = 4 per group). ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001.
   Data are mean ± SEM.

5. Discussion

   Gestational subclinical hypothyroidism is more common than overt
   hypothyroidism and autoimmune thyroid disease during pregnancy, with
   its effect on foetal outcomes and long-term offspring health remaining
   controversial [[140]50]. To investigate this, we established a maternal
   SCH mouse model by administering PTU during pregnancy, which resulted
   in elevated serum TSH concentrations without significantly affecting
   serum FT4 and FT3 levels in the pregnant mice. Accumulating evidence
   suggests that maternal hypothyroidism may disrupt abnormal glucose
   metabolism in offspring. For instance, Lucaccioni L et al. reported
   that children born to mothers with hypothyroidism are more likely to
   develop metabolic syndrome, have higher body mass indices (BMI), and
   exhibit greater waist circumferences [[141]10]. Offspring of
   hypothyroid mothers may be at higher risk for type 2 diabetes mellitus
   (T2DM), although potential confounding factors, such as preterm birth
   or low birth weight, cannot be entirely excluded [[142]7]. Previous
   studies in rodents have also shown that maternal hypothyroidism leads
   to glucose intolerance in offspring, increasing their susceptibility to
   diabetes in adulthood [[143]9,[144]38,[145]51]. In our study, we found
   that maternal SCH predisposes offspring to glucose intolerance, insulin
   resistance, and impaired liver glucose regulation compared to offspring
   from euthyroid conditions. Early supplementation with L-T4 mitigated
   glucose intolerance and reduced insulin resistance in offspring of SCH
   mothers, indicating that even mild intrauterine thyroid hormone
   deficiency has significant long-term metabolic implications for
   offspring.

   However, the molecular mechanisms by which maternal SCH impairs glucose
   metabolism and induces insulin resistance in offspring remain poorly
   understood. To explore this, we conducted RNA sequencing of liver
   tissues from two groups and identified ARRDC3 as one of the most
   differentially expressed genes. ARRDC3 is closely associated with
   glucose and lipid metabolism. A previous study demonstrated that ARRDC3
   (also known as thioredoxin-binding-protein-2-like inducible membrane
   protein [TLIMP]) regulates peroxisome proliferator-activated receptor
   (PPAR) activity, which is involved in glucose metabolism [[146]52]. In
   another study, global ARRDC3 deletion in rodents prevented age-related
   adiposity, alleviated hepatic steatosis, and improved insulin
   sensitivity [[147]25]. Adipocyte-specific ARRDC3 deletion also enhanced
   glucose tolerance. However, the effects of ARRDC3 deletion in
   adipocytes alone did not fully explain the phenotypes observed in these
   mice, suggesting that other tissues may also contribute to the observed
   outcomes [[148]25]. Recent studies have reported that ARRDC3 is one of
   the top upregulated transcripts in the liver during a
   hyperinsulinemic-euglycemic clamp, and its loss in the liver increases
   insulin sensitivity [[149]26]. Based on these findings, we speculate
   that the upregulation of ARRDC3 may be a key mechanism underlying the
   abnormal glucose metabolic phenotype observed in SCH offspring.

   The Akt signaling pathway, which is activated by INSR, plays a central
   role in liver metabolism, particularly in balancing processes like
   gluconeogenesis and glycogen synthesis [[150]53]. Our results
   demonstrated that overexpression of ARRDC3 inhibited the
   insulin-induced activation of Akt at Thr308, GSK-3β at Ser9, and FoxO1
   at Ser256. This led to increased expression of key gluconeogenic
   enzymes, G6Pase and PEPCK, which are known to promote hepatic glucose
   production, further suggesting that ARRDC3 plays a negative regulatory
   role in insulin sensitivity and glucose homeostasis. Recent studies
   have proposed that ARRDC3 may interact with phosphatidylinositol
   4,5-bisphosphate (PI45P2) [[151]27], a phospholipid that plays a
   crucial role in the recruitment and activation of Akt at the plasma
   membrane [[152]54], which is essential for insulin-mediated signalling
   [[153]55]. By interacting with PI45P2, ARRDC3 could potentially disrupt
   Akt's recruitment to the membrane, thereby attenuating its activation.
   This would inhibit downstream signaling required for insulin's effects
   on glucose metabolism.

   Maternal hypothyroidism has been shown to influence epigenetic
   mechanisms, such as DNA methylation and histone modifications, which in
   turn alter the expression of various genes and affect processes like
   embryonic astrocyte maturation and cerebellar function in rodents
   [[154]56,[155]57]. In a similar context, elevated H3K27Ac levels have
   been reported in the offspring of mothers fed a high-fat diet (HFD) and
   treated with methimazole (MMI) [[156]58]. Consistent with these
   findings, we observed increased H3K27Ac expression in the liver tissue
   of SCH offspring. The ChIP results demonstrated that increased
   enrichment of H3K27Ac in the promoter region of ARRDC3 in the liver
   tissue of the SCH offspring result in the upregulation of ARRDC3.
   Moreover, the upregulation of P300 further supports the role of H3K27Ac
   in regulating ARRDC3 expression. In addition, our study found that this
   upregulation did not result from changes in the methylation status of
   the ARRDC3 promoter region ([157]Supplementary Fig. 3).

   The results of this study provide valuable insights into the long-term
   metabolic health risks associated with mild intrauterine thyroid
   hormone deficiency. Nevertheless, there are several limitations to
   consider. We observed slight lipid accumulation in the livers of the
   SCH offspring, suggesting that ARRDC3 is involved in lipid metabolism.
   However, we did not investigate the specific effects of ARRDC3 on lipid
   metabolism, nor could we elucidate the relationship between abnormal
   lipid metabolism and insulin resistance in SCH offspring. Further
   studies are needed to explore these connections. Additionally, we
   focused exclusively on the role of ARRDC3 in the liver, but given its
   multiple physiological functions, this molecule may also contribute to
   abnormalities in other organs of SCH offspring. Regarding the
   regulation of ARRDC3, although we concentrated on P300 as a HAT, we
   recognize that other HATs may also play a role in its acetylation.
   Future studies should explore a broader range of HATs and their
   regulatory mechanisms to provide a more comprehensive understanding of
   acetylation in the ARRDC3 regulatory network.

   In summary, our study revealed that maternal SCH causes insulin
   resistance and glucose intolerance in adult offspring. The potential
   molecular mechanism might involve the upregulation of P300-ARRDC3 axis,
   which suppresses the Akt pathway, eventually resulting in increased
   hepatic glucose production and the suppression of gluconeogenesis
   ([158]Fig. 7). Notably, when mild in utero thyroid hormone deficiency
   was corrected, both the insulin resistance phenotype and related
   molecular alterations, such as ARRDC3 expression, were reversed in SCH
   offspring. This suggests that timely L-T4 supplementation during
   pregnancy is crucial for ensuring the metabolic health of offspring
   exposed to maternal SCH. Our findings contribute to a deeper
   understanding of the prolonged detrimental effects of maternal SCH on
   metabolic health and provide a foundation for preventing abnormal
   metabolic function in SCH offspring.

Fig. 7.

   [159]Fig. 7
   [160]Open in a new tab

   Fig 7.

CRediT authorship contribution statement

   Ming Huo: Writing – review & editing, Writing – original draft,
   Methodology, Funding acquisition, Data curation. Xi Yu: Formal
   analysis, Data curation. Xianbin Yuan: Software, Investigation, Data
   curation. Jun Guo: Software, Methodology. Bin Wei: Formal analysis,
   Data curation. Yajun Shi: Methodology, Investigation. Yannan Gu: Data
   curation. Xuehong Zhang: Visualization, Validation, Supervision. Miao
   Sun: Writing – review & editing, Writing – original draft, Supervision,
   Resources, Project administration, Funding acquisition.

Ethics approval statement

   The experiment was approved by the Committee and Animal Ethics
   Committee of the Soochow University (approval number SYXK-2022-0043).

Ethics approval and consent to participate

   All animal experiments were performed in accordance with the guidelines
   of the University Committee on Animal Care of Soochow University.

Consent for publication

   All authors have approved for publication.

Data availability statement

   Data will be made available on request.

Funding

   This work was supported by the National Key R&D Program of China
   [2022YFC2703700 and 2019YFA0802600]; the National Natural Science
   Foundation of China [81974244 and 81570960]; the Gansu Natural Science
   Foundation [21JR1R097]; and the Intra-Hospital Fund of the First
   Hospital of Lanzhou University [ldyyyn2021-82].

Declaration of competing interest

   The authors declare that they have no known competing financial
   interests or personal relationships that could have appeared to
   influence the work reported in this paper.

Footnotes

   ^Appendix A

   Supplementary data to this article can be found online at
   [161]https://doi.org/10.1016/j.heliyon.2024.e39259.

Contributor Information

   Xuehong Zhang, Email: zhangxueh@lzu.edu.cn.

   Miao Sun, Email: miaosunsuda@163.com.

Appendix A. Supplementary data

   The following is the Supplementary data to this article:
   Multimedia component 1
   [162]mmc1.docx^ (7.5MB, docx)

References