2025 Volume 48 Issue 3 Pages 267-278
Although accelerated growth is observed in both sexes upon reaching puberty, the growth of girls ceases around menarche (the average age at menarche is 12–13 years in Japan). However, the molecular basis of the action of estrogen remains unclear. In this study, we investigated whether estrogen is involved in the differences in growth between males and females while focusing on thyroid hormone metabolic enzymes. We analyzed the promoters of iodothyronine deiodinase (DIO)2 and DIO3 by 17β-estradiol (E2). ATDC5 cells (mouse chondrocytes cell line) were treated with E2, and the expression of DIO2 and DIO3 mRNA and proteins was evaluated. Sham-operated (sham) or ovariectomized (OVX) female mice were treated daily with E2 or vehicle for three consecutive weeks. Subsequently, the left femur was removed to evaluate the effect of E2 on DIO2/DIO3 gene and protein expression. E2 increased the transcriptional activity of DIO3 in a concentration-dependent manner. On the DIO3 promoter indicates the presence of an estrogen response element. DIO2 and DIO3 mRNA and protein expression in ATDC5 cells in the presence of E2 was significantly increased, while DIO2 expression was unchanged. In vivo, we used OVX mice and E2 supplementation as a model of amenorrhea for further investigation. DIO3 expression was significantly increased in mice treated with E2 in OVX compared to that in mice treated with vehicle in sham. E2 increases DIO3 expression in chondrocytes and long bone tissues, suggesting that E2 may affect bone growth and cause sexual dimorphism during puberty.
According to reports from the Ministry of Health, Labor, and Welfare in Japan, menarche typically occurs first between the ages of 10 and 16 years. The average age at menarche has been evaluated in various countries and is estimated to be 12–13 years, although it varies depending on the environment and country.1–3) Although 98% of women reach menarche by the age of 15 years, menarche tends to occur later among athletes.4–6) Delayed age at menarche is associated with physical maturation, while a later age at menarche is associated with taller stature. With the onset of menstruation, increased estrogen secretion promotes sexual maturity in the female body, including that of the female reproductive organs and secondary sexual characteristics.
As estrogen plays an important role in mediating the increase in growth hormone (GH) action during puberty, the estrogen receptor (ER) is also involved in the increase in GH activity during the same period. Liganded-ER stimulates GH-producing cells in the pituitary gland.7–9) Govoni et al. reported that prepubertal ovariectomized (OVX) mice exhibited increased levels of insulin-like growth factor (IGF)-1 mRNA in the serum, liver, and bones and that treatment of OVX mice with estrogen resulted in increased mRNA levels of IGF-1 in the bone and serum.9) These levels were comparable to those observed in sham-operated (sham) mice. Several studies have suggested that estrogen increases GH levels, resulting in bone elongation either directly or indirectly.7–9) Furthermore, estrogen deficiency during puberty increased IGF-1 levels, bone size, and bone mass9) and also increased bone elongation in ERα-deficient mice.10)
Previous reports have demonstrated that estrogen indirectly promotes growth by enhancing GH secretion at low concentrations7–9,11) but advances the closure of the epiphyseal line at high concentrations.10–12) Estrogen has been reported to exert both growth-promoting and growth-suppressing functions due to its bimodal effects on the growth plate. However, the molecular mechanisms by which estrogen promotes epiphyseal closure and attenuates cell growth remain unclear.
GH, IGF-1, and thyroid hormone (TH) are involved in skeletal growth during puberty regardless of gender.13,14) Insufficient TH levels in children lead to stunted growth and bone maturation, while hyperthyroidism promotes these processes.14) TH is essential for bone remodeling and maintenance, even in adults. Thyroxine (T4), a prohormone produced by the thyroid gland, is converted by iodothyronine deiodinase (DIO) into target cells in the peripheral tissues. DIOs are selenoproteins involved in thyroid hormone metabolism. There are three types of DIOs that include type 1, type 2, and type 3 (DIO1, DIO2, and DIO3, respectively). Prohormone (T4) is converted by DIO1 into active triiodothyronine (T3) and inactive reverse T3 (rT3), and DIO2 converts T4 to T3, DIO3 converts T4 to rT3 and T3 to T2. The major regulators of DIO expression and function are changes in the thyroid hormone status, several growth factors, cytokines, and pathophysiological conditions. Estrogen and progesterone have been suggested to regulate DIO2 and DIO3 expression in the rodent uterus.15) This is believed to control TH balance and plays an important role in fetal development.15)
This study aimed to determine for the first time the relationship between estrogen and TH in the context of bone growth. ATDC5 cells were used as a chondrocyte model.16–18) ATDC5 cells differentiate into chondrocytes in the presence of insulin, exhibit several differentiation markers of growth plate chondrocytes, and calcify, thereby reflecting the endochondral ossification process in vivo. Therefore, they are often used in in vitro experiments examining bone growth.18) Studies using ATDC5 and MC3T3-E1 cells as osteoblasts have demonstrated neither DIO1 expression nor involvement in T4 metabolism.19) Miura et al. measured the enzyme activity in ATDC5 cells and demonstrated that the reaction was not affected by 6-propyl-2-thiouracil, an inhibitor of DIO1 enzyme activity, and that DIO2 primarily catalyzed the reaction.17) In this study, we examined the regulation of DIOs expression by estrogen both in vitro and in vivo while focusing on types 2 and 3.
Two cell lines were used for cell culture, including TSA 201 cells that are derived from human embryonic kidney 293 cells (HEK293 cells) and ATDC5 cells, a chondrogenic cell line from mouse embryonal carcinoma (Riken Cell Bank, Ibaraki, Japan). Briefly, TSA201 cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM) (Nissui Pharmaceutical Co., Ltd., Tokyo, Japan) containing 10% fetal bovine serum (FBS), 2 mM l-glutamine, 100 U/mL penicillin, and 100 μg/mL streptomycin (P/S). ATDC5 cells were maintained in DMEM/Ham’s F-12 1 : 1 (with l-glutamine and Phenol Red; FUJIFILM Wako Pure Chemical Corporation, Osaka, Japan) containing 5% fetal bovine serum (FBS) and P/S. ATDC5 cell differentiation was induced as previously described.16)
Construction of PlasmidsReporter plasmids were constructed in which the DIO2 or DIO3 promoters (designated hDIO2 pro-Luc and hDIO3 pro-Luc, respectively) were cloned upstream of the luciferase coding sequence.20,21) The expression plasmids for pEGFP-C1-ERα (RRID: Addgene_28230) and pEGFP-C1-ERβ (RRID: Addgene_28237) were purchased from Addgene (Watertown, MA, U.S.A.). The internal control plasmid was pGL4.70 (Promega, Madison, WI, U.S.A.).
Dual Luciferase AssayTSA201 cells were seeded at 2.0 × 105 cells/well into a 12-well culture plate. After 16 h, calcium phosphate transfection was used to introduce either hDIO2 pro-Luc, hDIO3 pro-Luc, or pGL4.10 (Luciferase vector) (20 ng) as the reporter plasmid, pEGFP-C1-ERα or -β (200 ng), green fluorescent protein (GFP)-Androgen Receptor (AR) (200 ng), and pGL4.70 (50 ng) as the internal control plasmid. We also measured each promoter activity without co-transfection or with GFP co-transfection only.22) Five hours after transfection, the culture medium was replaced with phenol red-free DMEM (FUJIFILM Wako Pure Chemical Corporation) containing 10% Charcoal/Dextran Treated FBS, 2 mM l-glutamine, and P/S and 17β-estradiol (E2) or dihydrotestosterone (DHT) (Nacalai Tesque Inc., Kyoto, Japan). After incubation for 20 h, the cells were harvested and measured for luciferase activity using the Dual-Luciferase® Reporter Assay System (Promega) according to the manufacturer instructions.
RNA Extraction for Quantitative RT-PCR and PCR Agarose ElectrophoresisDifferentiated ATDC5 cells were maintained in the presence of E2 (100 nM) for 6, 12, and 24 h. The resulting cells were harvested, and total RNA was extracted using RNAiso Plus (TaKaRa Bio Inc., Shiga, Japan). The extracted total RNA was reverse transcribed using iScript Advanced Reverse Transcriptase (Bio-Rad, Hercules, CA, U.S.A.) under the following conditions: 46 °C for 20 min and 95 °C for 5 min. The cells were stored at 4 °C until further use. RT-PCR was performed to evaluate the expression level of ERα/β with β-actin as the internal control using Tks Gflex™ DNA Polymerase (TaKaRa Bio Inc.) with 20 ng of cDNA and 0.2 μM of the relevant primers under the following conditions: 30 cycles of 98 °C for 10 s; 60 °C for 15 s; 68 °C for 4 s (30 s/kB). The resulting PCR products were electrophoresed on 2% agarose gels. Then, quantitative RT (qRT)-PCR was performed to verify the gene expression levels of DIO2/DIO3 using β-actin as an internal control, iTaq™ Universal SYBR™ Green Supermix (Bio-Rad) with 10 ng of cDNA, and 0.2 μM of the relevant primers under the following conditions: 95 °C for 25 s, 40 cycles of 95 °C for 5 s, and 60 °C for 30 s. Data were analyzed using the ΔΔCt method. The primers used in this study are listed in Table 1.
| Gene | Forward primer(5'–3') | Reverse primer(5'–3') | Product length |
|---|---|---|---|
| Mouse ERα | GGGCGACATTCTTCTCAAGC | GAGAGAGGAACACGGGATGTG | 124 bp |
| Mouse ERβ | ACCAAGTCCGCCTCTTGGAA | CCTCATCCCTGTCCAGAACG | 123 bp |
| Mouse DIO2 | TTGGGGTAGGGAATGTTGGC | TCCGTTTCCTCTTTCCGGTG | 99 bp |
| Mouse DIO3 | GCTCGAAACAGCGCCTAAAG | TGGTCTCGAAGTCCATCCT | 108 bp |
| Mouse β-actin | CCTTCCAGCAGATGTGGATCA | CTCAGTAACAGTCCGCCTAGAA | 102 bp |
Proteins were extracted either from differentiated ATDC5 cells treated with E2 (100 nM) for 24 h or from the bones of experimental mice. The protein samples were electrophoresed in TGX FastCast Acrylamide at 12% (Bio-Rad) and transferred onto an Amersham™ Hybond™ PVDF membrane (GE Healthcare Life science, Piscataway, NJ, U.S.A.). The membrane was blocked with the blocking reagent WB-Chemically (Cosmo Bio Co., Ltd., Tokyo, Japan) before incubating with primary and secondary antibodies. (i) The primary antibodies included DIO2 (ab77779; Abcam, Cambridge, U.K.) and/or DIO3 (NBP1-05767; Novus Biologicals, Littleton, CO, U.S.A.). (ii) The secondary antibodies included anti-rabbit immunoglobulin G, HRP-linked antibody (7074S; Cell Signaling Technology, Danvers, MA, U.S.A.), and β-actin (C4) (sc47778; Santa Cruz Biotechnology, Dallas, TX, U.S.A.). IMMUNO Shot-Platinum (Cosmo Bio Co. Ltd.) was used for all reactions. Immune complexes were visualized using Clarity Western ECL Substrate (Bio-Rad). Densitometric analysis of the bands was performed using an ATTO WSE-6200 LuminoGraph II (ATTO, Tokyo, Japan).
Animal ExperimentsEight-week-old female C57BL/6J mice were purchased from SLC, Inc. (Shizuoka, Japan). They were acclimated to an animal facility for 4 weeks under a 12-h light/dark cycle prior to surgery. All mice were housed individually. The mice were divided into three groups that included (i) sham operation with vehicle (Sham + V: n = 8), (ii) OVX with vehicle (OVX + V: n = 8), and (iii) OVX with E2 (OVX+ E: n = 8). Mice were anesthetized with isoflurane (3–4%) (FUJIFILM Wako Pure Chemical Corporation) for ovariectomy (OVX). Similarly, a sham operation was performed in which the ovaries were exposed but not removed. Three days after surgery, E2 (5 μg/kg/d) was administered intravenously daily for three consecutive weeks. Body weight was measured daily. The control group received an equivalent volume of saline as vehicle. Upon sacrifice, blood was sampled from the posterior vena cava and replaced with saline from the left ventricle. The left femur was sampled. Uterine weight was measured at sacrifice to determine the effectiveness of the OVX surgery and E2 treatment (Supplementary Fig. 1).
Ethical ApprovalThe animals were handled and cared for in accordance with the Guidelines for the Care and Use of Laboratory Animals at Mukogawa Women’s University that are compatible with the National Institutes of Health Guide for the Care and Use of Laboratory Animals (NIH Publication No. 8023, revised 1978). The experimental protocols were approved by the Institutional Animal Care and Use Committee of the Mukogawa Women’s University (Approval Nos. P-15-2022-03-A2 and P-15-2023-02-A).
Statistics AnalysisStatistical analyses were performed using the IBM SPSS Statistics software (SPSS Inc., NY, U.S.A.). Data are presented as the mean ± standard deviation (S.D.) from at least three in vitro experiments. Statistical analysis was performed using a one-way ANOVA followed by multiple comparisons using the Bonferroni method or uncorrelated t-test. Differences were considered statistically significant at p < 0.05.
We confirmed that the expression plasmid vector (pEGFP-C1) did not alter the hDIO2 and/or hDIO3 promoters (Fig. 1a), and the effects of ERα, and ERβ on luciferase vectors (Figs. 1b, 1c). To determine whether ERα or ERβ affected the activity of the DIO2/DIO3 promoter, the dose-dependent effects of E2 on the DIO2 and DIO3 promoters were evaluated. Liganded-ERα suppressed the transcription level of hDIO2 pro-Luc by 0.71-, 0.60-, and 0.41-fold in a dose-dependent manner, respectively, compared to that of the control (Fig. 1b, left panel). In the case of ERβ, the activities were decreased by 0.69-, 0.57-, and 0.59-fold, respectively, in a dose-dependent manner compared to that of the control (Fig. 1b, center panel).
(a) The effect of the expression plasmid vector (pEGFP-C1) on the hDIO2 and hDIO3 promoters was confirmed. (b, the upper row panels) and (c, the lower row panels) examined the effect of increasing doses of 17β-estradiol (E2) on the transcriptional activity of DIO2/DIO3 using TSA201 cells, which are easily transfected. TSA201 cells were transfected with either hDIO2 or hDIO3 promoter as a reporter, and pEGFP-C1-estrogen receptor a or pEGFP-C1-ERα. To compare the effect of ER on transcription, both ERα/β were used. pGL4.70 (50 ng) was used as internal control. Data are presented as mean ± S.D. from three transfections performed in triplicate. Error bar = mean ± S.D.. *p < 0.05, ***p < 0.005, ****p < 0.001, (n = 3) vs. vehicle. DIO: iodothyronine deiodinase, ER: estrogen receptor, S.D.: standard deviation.
ERα increased the transcription levels of hDIO3 pro-Luc by 1.41-, 1.94-, and 2.20-fold in a dose-dependent manner in response to the E2 concentrations, respectively, compared to that of the control (Fig. 1c, left panel). Similarly, ERβ increased in transcriptional activities by 2.78-, 3.72-, and 5.44-fold in a dose-dependent manner in response to E2, respectively, compared to that of the control (Fig. 1c, center panel). The promoter activities of hDIO2 and hDIO3 increased slightly in the presence of androgen receptor (AR) and DHT, without statistical significance (Figs. 1b, 1c, right panel).
Estrogen Response Elements in the Human DIO2/DIO3 PromotersWe used TFBIND software23) to search for candidate estrogen response elements (ERE) (AGGTCAnnnTGACCT) in the hDIO2 and hDIO3 promoters (Table 2). Transcriptional activity of hDIO2 promoter cut constructs (–1792, –722, –322, –83 bp) and hDIO3 promoter cut constructs (–1487, –524 bp) were examined using E2-ERα and ERβ Next, we examined transcriptional activity when the predicted ERE was deleted (Figs. 2c, 2d). White squares indicate existing consensus sequences and black squares indicate deleted consensus sequences.
| a) ERE consensus sequence on the hDIO2 promoter | ||
| ERE consensus sequence | AGGTCAnnnTGACCT | |
| Estimated ERE consensus sequence on the hDIO2 promoter | A | GACGTCCTCAATGAGCCAA |
| B | GGAGGTAGCAGTGAACAAA | |
| C | TCAGCTATCTATTACCCAG | |
| D | GGGGGTCATGCCATTCTGC | |
| E | AATGGTCAATGTGGCACCT | |
| F | ATGTGGCACCTTGAACCAA | |
| G | CTAGGTCACAGATCTTACA | |
| H | CTGTCAAGGGTATTA | |
| I | CCTTCTAAAAAAAAA | |
| J | ATTCACTGCAATCCT | |
| K | AGGCGTCAGGGAGACTCAC | |
| b) ERE consensus sequence on the hDIO3 promoter | ||
| ERE consensus sequence | AGGTCAnnnTGACCT | |
| Estimated ERE consensus sequence on the hDIO3 promoter | A | CTAGGTCAAGGCACCTCCC |
| B | CCGGCGGCCGCTGACCCAG | |
The consensus sequence for ERE is reported as AGGTCAnnnTGACCT (Table 2). Open squares indicate existing consensus sequences and solid squares indicate deleted consensus sequences. pEGFP-C1-ER or pEGFP-C1-ERα, or as reporter plasmids, the indicated DIO2 or DIO3 pro-Luc constructs or constructs deleted possible EREs, and the internal control plasmid pGL4.70 were simultaneously transfected. Each bar is expressed as a relative value indicating the basal transcriptional activity of each reporter plasmid; the concentration of 17β-estradiol (E2) was 10 nM. (a, d) The effect of vector (pEGFP-C1) was confirmed in all promoter cut constructs used in the study. (b, c) Rectangles indicate predicted ERE consensus sites, designated A-K. DG-DIO2 pro-Luc, DH-DIO2 pro-Luc, DI-DIO2 pro-Luc, and DJ-DIO2 pro-Luc indicate deletion of consensus sequences G, H, I, and J, respectively. Data represent relative values indicating the basal transcriptional activity of each reporter plasmid. (e, f) Rectangles indicate predicted ERE consensus sites, designated A and B. DA-DIO3 pro-Luc and DB-DIO3 pro-Luc indicate deletion of consensus sequences A and B, respectively. Data represents relative values indicating the basal transcriptional activity of each reporter plasmid. Data are shown as mean ± S.D. of at least three transfections performed in triplicate. Error bars = mean ± S.D. ****p < 0.001, (n = 3) vs. vehicle. DIO: iodothyronine deiodinase, ER: estrogen receptor, ERE: estrogen response element, S.D.: standard deviation.
The consensus sequence for ERE is reported as AGGTCAnnnTGACCT (Table 2). Open squares indicate existing consensus sequences and solid squares indicate deleted consensus sequences. pEGFP-C1-ER or pEGFP-C1-ERα, or as reporter plasmids, the indicated DIO2 or DIO3 pro-Luc constructs or constructs deleted possible EREs, and the internal control plasmid pGL4.70 were simultaneously transfected. Each bar is expressed as a relative value indicating the basal transcriptional activity of each reporter plasmid; the concentration of 17β-estradiol (E2) was 10 nM. (a, d) The effect of vector (pEGFP-C1) was confirmed in all promoter cut constructs used in the study. (b, c) Rectangles indicate predicted ERE consensus sites, designated A-K. DG-DIO2 pro-Luc, DH-DIO2 pro-Luc, DI-DIO2 pro-Luc, and DJ-DIO2 pro-Luc indicate deletion of consensus sequences G, H, I, and J, respectively. Data represent relative values indicating the basal transcriptional activity of each reporter plasmid. (e, f) Rectangles indicate predicted ERE consensus sites, designated A and B. DA-DIO3 pro-Luc and DB-DIO3 pro-Luc indicate deletion of consensus sequences A and B, respectively. Data represents relative values indicating the basal transcriptional activity of each reporter plasmid. Data are shown as mean ± S.D. of at least three transfections performed in triplicate. Error bars = mean ± S.D. ****p < 0.001, (n = 3) vs. vehicle. DIO: iodothyronine deiodinase, ER: estrogen receptor, ERE: estrogen response element, S.D.: standard deviation.
We confirmed that the expression vector (pEGFP-C1) did not exert any significant effect on the activity of the truncated constructs of hDIO2 and hDIO3 promoter (Figs. 2a, 2d).
Transcriptional activity at –83/ + 1 bp hDIO2 pro-Luc in ERα and –322/ + 1 bp hDIO2 pro-Luc in ERβ was almost completely abolished (Fig. 2b). We observed that ERα failed to suppress the deleted construct of the designated H box at [–240 bp] in –322/ + 1 hDio2 pro-Luc, and ERβ also failed to suppress the deleted designated as G box construct at [–502 bp] in –722/ + 1 hDio2 pro-Luc (Figs. 2b, 2c). These results suggest that binding to the consensus sequences located at –240 bp for ERα and at –502 bp for ERβ, respectively, results in negative regulation.
In the cut construct of hDIO3 promoter, both ERα and ERβ lost both A and B, the candidate sequences for ERE, and the transcriptional activity of –524/ + 1 hDIO3 pro-Luc was completely reduced (Fig. 2e). Boxes A and B were deleted from the putative ERE of the hDIO3 promoter. In particular, transcriptional activity was reduced in ΔB hDIO3 pro-Luc with the B box deleted (Fig. 2f).
These results indicate that the ERE located at –854 bp of hDIO3 pro-Luc contributes to DIO3 transcription.
Estrogen Induces DIO3 mRNA Expression in Differentiated ATDC5 CellsWe evaluated DIO2 and DIO3 expression in the presence of E2 in differentiated ATDC5 cells. After 6, 12, and 24 h of E2 (100 nM) stimulation, DIO2 mRNA expression did not change, regardless of E2 treatment at either point. DIO3 mRNA expression increased 2.18-fold compared to that in the vehicle control after 12 h of E2 treatment, although no statistical differences between the control and E2-treated cells were observed after 6 and 24 h (Fig. 3a).
After ATDC5 cells were treated with 100 nM of 17β-estradiol (E2) for 6, 12, and 24 h, relative mRNA expressions of iodothyronine deiodinase (DIO) 2/DIO3 were assessed by qRT-PCR. Relative expression levels of DIO2/DIO3 by E2 are shown in summary. Data were analyzed using the ΔΔCt method. The data are presented as the mean ± S.D. from three independent experiments performed in triplicate. Error bar = mean ± S.D. *p < 0.05, **p < 0.01, (n = 3) vs. vehicle. N.S.: not statistically significant, qRT: quantitative RT, S.D.: standard deviation.
DIO2 and DIO3 protein expression was evaluated by Western blotting in differentiated ATDC5 cells at 24 h after the addition of 100 nM E2. Although E2 failed to induce the expression of DIO2, DIO3 protein levels increased by 2.27-fold compared to that of the control in the presence of E2, with statistical significance (Fig. 4).
After treatment with 100 nM of E2 for 24 h, proteins were extracted and Western blotting analysis of ATDC5 cells were performed to assess DIO2/DIO3 protein expression. The representative data are shown as the mean ± S.D. Error bar = mean ± S.D. *p < 0.05, **p < 0.01, (n = 3) vs. vehicle. N.S.: not statistically significant, S.D.: standard deviation.
We investigated the effects of ovariectomy and E2 supplementation on the expression of DIO2 and DIO3 (Fig. 5a). Mice were divided into three groups that included OVX and vehicle administration as a model of prepuberty (OVX + V), OVX with E2 supplementation (OVX+ E), and sham operation and vehicle administration (Sham + V). DIO2 protein levels increased 2.55-fold in OVX + V mice compared to that in Sham + V mice; however, in OVX+ E2 mice, the levels were reduced to the same level as that in Sham + V mice (Fig. 5a). DIO3 protein levels were also slightly increased in OVX + V mice compared to Sham + V mice. It was also increased 2.59-fold in OVX + E2 mice (Fig. 5a). Initially, we confirmed the expression of both ER isoforms in both sexes (Fig. 5b). We compared the expression levels of ERα and β using qPCR in OVX + E and Sham + V mice, which had significantly higher protein expression of DIO3, but found no difference in expression levels (Fig. 5c).
(a) After OVX or sham-operation were performed, E2 or vehicle injected intravenously daily for 3 weeks. The right femoral bone was sampled to proceed Western blotting analysis. DIO2/DIO3 protein expressions were evaluated. The representative data are shown as the mean ± S.D. Error bars = mean ± S.D. *p < 0.05, (n = 5) vs. Sham + V. (b) RT-PCR analysis of ERa and ERα expression in ATDC5 cells, mouse left femur (male or female) using β-Actin as an endogenous control. (c) After OVX or sham surgery, E2 or vehicle was injected intravenously daily for 3 weeks, and relative mRNA expression of ERα or ERα was assessed by quantitative RT-PCR. The data are presented as the mean ± S.D. from three independent experiments performed in triplicate. Error bar = mean ± S.D. *p < 0.05, **p < 0.01, (n = 5) vs. vehicle. DIO: iodothyronine deiodinase, N.S.: not statistically significant, ER = estrogen receptor, OVX: ovariectomized, S.D.: standard deviation.
The regulation of longitudinal growth, particularly in the context of sexual dimorphism, is highly complex. Nutritional conditions are established factors in the regulation of health status, although the cellular mechanisms of action are largely unknown in both sexes.
Postnatal growth is predominantly influenced by the GH-IGF-1 axis and mixed endocrine, paracrine, and autocrine secretion patterns. Furthermore, TH, glucocorticosteroids, and sex steroid hormones are closely associated with somatic growth. GH and TH cooperatively play pivotal roles in teenage growth.
TH plays a pivotal role in endochondral ossification and longitudinal bone growth and is essential for skeletal elongation during the neonatal period through puberty. The direct action of TH on chondrocytes is crucial for bone elongation, and this has been confirmed in several in vitro and in vivo studies, including clinical cases of cretinism and/or resistance to thyroid hormones and mouse genetic model studies.24)
Intracellular T3 levels are critical for various physiological and biological processes, and T3 requirements may vary widely depending on T3 demand. Consequently, the expression levels and balance of DIO2 and DIO3 determine local concentrations of T3.25)
Steroid hormones and growth factors form complex interacting networks. They are also involved in longitudinal bone growth in both children and adolescents. As mentioned in the Introduction, the bimodal effects of estrogen that include growth promotion and attenuation have been previously investigated. For example, for growth promotion, a low dose of estrogen equivalent to a serum estradiol concentration of approximately 4 pg/mL (15 pmol/L) in both prepubertal boys and girls has been demonstrated to increase prepubertal growth rates by more than 60% in both boys and girls.7,26) Juul reported that estrogen induces GH secretion in combination with IGF-1, thereby stimulating growth. Consequently, estrogen exerts a biphasic effect on longitudinal growth in both sexes, such that very low levels of estrogen directly increase bone elongation in the growth plate. Conversely, high estrogen levels promote epiphyseal fusion.7)
After the onset of menarche and the subsequent increase in estrogen secretion, the growth spurt gradually decreases, eventually reaching a plateau in teenagers. Furthermore, in support of the notion that estrogen is responsible for this reduction in the growth rate, studies have reported that patients with aromatase deficiency and males with mutations in ERα continue to grow taller, as the epiphyseal line does not fuse.27–29) In recent years, estrogen supplementation therapy has been employed to promote the closure of the epiphyseal line to stop the growth of patients with the aforementioned diseases.28–30) However, the molecular mechanism underlying the estrogen therapy-induced closure of the epiphyseal line remains unclear.
The correlation between male pubertal growth rate and the effects of GH has been ascribed to androgens. Conversely, pubertal growth velocity has been attributed to estrogen and/or adrenal androgens in female. Previous reports have suggested that estrogen may be the main hormone that accelerates the pubertal growth spurt in both sexes. These actions are mediated by ERs in the long bone growth plates. The case of an individual with a disruptive mutation in the ERα gene who did not experience a pubertal growth spurt but continued to grow into adulthood was attributed to a lack of epiphyseal fusion.7,10,27) Using cartilage-specific disruption of ERα in mice, Börjesson et al. demonstrated that ERα in growth plate cartilage is not important for skeletal growth during early sexual maturation. They concluded that ERα is indispensable for high-dose estradiol to reduce the growth plate height in adult mice and for the reduction of longitudinal axis growth in the bones of 1-year-old mice.10) Thus, estrogen promotes linear skeletal growth during the early stages of sexual maturation with high estrogen levels during the later stages of puberty, resulting in growth plate fusion in humans. In rodents, a reduction in growth plate height has been observed after treatment with high doses of estradiol, although growth plates do not fuse directly after sexual maturation. Currently, it remains unknown whether the effects of estrogen on skeletal growth are mediated directly via ERs in the growth plate cartilage and/or indirectly via other mechanisms such as the TH, GH, and/or IGF-1 axis. To the best of our knowledge, reports detailing the skeletal sexual dimorphism of the relationships between TH-TRs and sex steroids are scarce in the literature. Sex-dependent skeletal differences may emerge during early puberty and depend on the secretion of GH-IGF-1 and thyroid hormones and their effects on the skeletal system. Sex steroids such as androgens and estrogens exert stimulatory effects on bone elongation during puberty. In females, estrogen appears to restrict bone size during early puberty, although bone mineral density increases.
The differentiation and proliferation of chondrocytes and elongation of the linear growth of the epiphyseal plate are initiated in the embryo, continue throughout the postnatal and adult stages, and are tightly regulated by the hormonal environment. Extracellular stimulatory factors such as TH, GH, IGF-1, and estrogen execute their functions spatiotemporally, either directly or indirectly. However, these findings raise the question of how estrogen and ER allocate their roles to other extracellular factors and whether estrogen and ER act as initiators or executors of growth plate extension.
To elucidate the interaction of ERβ in the regulation of DIOs, we applied ERβ and compared it to ERα. We measured transcriptional activity in various combinations of E2 + ERα or ERβ and DIO2/DIO3 for the purpose of examining the effect on individual transcription. E2 + ERα and ERβ increased the transcriptional activity of DIO3 (Fig. 1c). These results indicate that the magnitude of ERβ on DIO3 transcription is different from that of ERα, but it functions positively similar to ERα. In mouse models, although the growth plates do not fuse even after adulthood, persistent treatment with a physiologically excessive dose of E2 leads to their fusion. In contrast, in humans the growth plate becomes completely fused after sexual maturation. A study in ERβ knockout (KO) mice also indicates that ERβ inhibits skeletal growth in young adult female mice.31) We confirmed that the expression of both genes was constant in differentiating ATDC5 cells (data not shown). Although ERβ activity increased in the DIO3 promoter, whether ERα or ERβ was favorable for a higher level of DIO3 expression in ATDC5 cells or in mouse bone was not determined. Furthermore, we examined AR and DHT that failed to induce the transcription of both DIO promoters (Figs. 1b, c).
In our study, E2 increased the promoter activity of DIO3 via the ER. In ATDC5 cells, E2 treatment increases DIO3 mRNA and protein expression. However, the results of this study differ from those of the reporter assay with respect to the analysis of DIO2 expression. While the reporter assay with TSA201 allows the evaluation of effects on individual promoters, this transient transfection does not allow for all cofactors in each target cell or tissue factors and other factors in each target cell that could their behavior. A previous report demonstrated that estrogen increased DIO2 expression in the rodent uterus.15) In support of the ATDC5 cell results, estrogen treatment also induced DIO3 expression in mouse bones. The concentration of estradiol added to the ATDC5 cells was determined based on the previous reports, while the dose of estradiol to TSA201 cells was limited to 10 nM to avoid cytotoxicity at higher concentrations.32,33)The expression of DIO2 and DIO3 in the femurs of OVX mice increased markedly compared to that in the sham group (Sham + V in Fig. 5a); however, DIO2 expression was not altered in the presence of E2 in Sham + V with sexual cycle and treatment with daily E2 administration (Fig. 5a, left panel). These results indicate that E2 reciprocally affects the expression of DIO2/3 via ERs both in vitro and in vivo. Furthermore, we observed that DIO3 expression in OVX mice without E2 (OVX + V in Fig. 5a, right panel) was mildly elevated compared to that in control mice, possibly due to the involvement of other factors. As the expression of DIO3 changes depending on the TH status,34,35) it is considered that the increase in DIO3 expression is reciprocally accompanied by an increase in DIO2 expression to regulate the metabolic balance of TH in situ. Although DIO3 expression can be increased by increased expression of DIO2 and thyroid hormone status, the results suggest that in chondrocytes, E2 is an upregulate of DIO3 expression without changes in DIO2.
We compared the expression levels of ERα and β using qPCR in OVX + E2 and Sham + V mice, which had significantly higher protein expression of DIO3, but found no difference in expression levels (Fig. 5c). We could not compare ERα and ERβ throughout this study. Consequently, even after reviewing past literatures, it is currently challenging to reach a consistent conclusion regarding the contribution of the two receptors regarding bone growth.
Skeletal formation and development are also associated with the expression of DIO2 and DIO3 that are responsible for local TH.34,36) T4 secreted from the thyroid gland is converted by DIO into target tissues such as long bones. DIOs is a selenoprotein involved in thyroid hormone metabolism that belongs to the selenoprotein family and contains a selenocysteine (Sec) insertion sequence (SECIS). As SECIS binding protein-2 (SECISBP2) is epistatic to selenoproteins, the mutation of SECISBP2 leads to a decrease in the expression level of selenoproteins.37) To the best of our knowledge, there have been no reports of DIOs mutations in mammals. However, DIOs deficiency caused by mutations in SECISBP2 can lead to insufficient TH activity, resulting in short stature and transient growth retardation.36,38–40) Consequently, a balance between DIO2/DIO3 expression and TH conversion is involved in skeletal growth.36) Although the treatment of short-stature patients with GH alone partially restored their height but did not achieve bone maturation, the combined treatment with GH and TH resulted in nearly normal thyroid function tests and improvement in both longitudinal bone growth and maturation.36) Some studies examining DIO KO mice have demonstrated that DIO2 KO mice exhibit no changes, while KO of monocarboxylate transporter 8 or DIO1 causes growth retardation.41) DIO3 KO mice exhibit growth retardation that is thought to be caused by thyrotoxicosis.41,42) Conversely, the catabolic effect of DIO3 overexpression on TH results in growth retardation. Charalambous et al. reported that mice overexpressing DIO3 exhibited significant growth retardation compared to wild-type mice.43)
Therefore, it has been concluded that maintaining the TH balance in local tissues through DIO2 and DIO3 promotes normal skeletal development.41)
In summary, in the present study, supplementation with E2 increased the expression of DIO3 in chondrocytes and long bone tissues such as the femur. Conversely, the expression of DIO2 mRNA and protein induced by E2 was consistent in chondrocytes and mouse femur bones. While the expression of DIO2 and DIO3 has been suggested to be involved in skeletal growth to a small extent, the present study indicates that only DIO3 expression is consistently increased by estrogen in chondrocytes, and this may contribute to sex differences in the skeleton. These results suggest that T4 to rT3 metabolism may be promoted by E2 in the peripheral tissues. Evidence suggests that longitudinal skeletal sexual dimorphism is established during early puberty and primarily depends on sex steroid hormones and GH-IGF-1. Better insights into the mechanisms of epiphyseal fusion may ultimately help develop new strategies for the treatment of cartilage and growth disorders.
We would like to thank for the English language editing.
This study was supported by Grants from KAKENHI (23K07980 to T. T.) and the Foundation for Growth Science (M. M.). The funders had no role in the study design, data collection and analysis, decision to publish, or manuscript preparation.
The authors declare no conflict of interest.
This article contains supplementary materials.
