Thyroid hormone receptor Thra and Thrb knockout differentially affects osteoblast biology and thyroid hormone responsiveness in vitro

Thyroid hormones (TH) are important modulators of bone remodeling and thus, thyroid diseases, in particular hyperthyroidism, are able to compromise bone quality and fracture resistance. TH actions on bone are mediated by the thyroid hormone receptors (TR) TRα1 and TRβ1, encoded by Thra and Thrb, respectively. Skeletal phenotypes of mice lacking Thra (Thra0/0) and Thrb (Thrb−/−) are well‐described and suggest that TRα1 is the predominant mediator of TH actions in bone. Considering that bone cells might be affected by systemic TH changes seen in these mutant mice, here we investigated the effects of TR knockout on osteoblasts exclusively at the cellular level. Primary osteoblasts obtained from Thra0/0, Thrb−/−, and respective wildtype (WT) mice were analyzed regarding their differentiation potential, activity and TH responsiveness in vitro. Thra, but not Thrb knockout promoted differentiation and activity of early, mature and late osteoblasts as compared to respective WT cells. Interestingly, while mineralization capacity and expression of osteoblast marker genes and TH target gene Klf9 was increased by TH in WT and Thra‐deficient osteoblasts, Thrb knockout mitigated the responsiveness of osteoblasts to short (48 h) and long term (10 d) TH treatment. Further, we found a low ratio of Rankl, a potent osteoclast stimulator, over osteoprotegerin, an osteoclast inhibitor, in Thrb‐deficient osteoblasts and in line, supernatants obtained from Thrb−/− osteoblasts reduced numbers of primary osteoclasts in vitro. In accordance to the increased Rankl/Opg ratio in TH‐treated WT osteoblasts only, supernatants from these cells, but not from TH‐treated Thrb−/− osteoblasts increased the expression of Trap and Ctsk in osteoclasts, suggesting that osteoclasts are indirectly stimulated by TH via TRβ1 in osteoblasts. In conclusion, our study shows that both Thra and Thrb differentially affect activity, differentiation and TH response of osteoblasts in vitro and emphasizes the importance of TRβ1 to mediate TH actions in bone.

osteoblasts in vitro and emphasizes the importance of TRβ1 to mediate TH actions in bone.

| INTRODUCTION
Bone is a dynamic organ that undergoes constant remodeling to ensure bone integrity and stability. 1Bone remodeling is a tightly regulated process performed by bone-resorbing osteoclasts and bone-forming osteoblasts under the direction of the mechanosensors embedded in bone, called osteocytes. 1 Several osteoblast/osteocytesecreted cytokines and factors coordinate osteoblastosteoclast coupling.As such, receptor activator of NF-κB ligand (RANKL), an important osteoclast stimulator, and osteoprotegerin (OPG), a potent RANKL decoy receptor and thus osteoclast inhibitor, link osteoblast activity to bone resorption. 2 The thyroid hormones (TH) L-thyroxine (T 4 ) and the biologically more active form 3,3',5-triiodo-L-thyronine (T 3 ) are well-known regulators of the whole body's metabolism and as such, they also play a key role in bone remodeling. 3][18][19][20][21] Recently, osteocytes in bones of hyperthyroid mice have been reported to acquire osteoclast-like features and actively shape their lacunae, a process so-termed osteocytic osteolysis. 22o exert their physiological actions, TH must bind to thyroid hormone receptors (TR) of the nuclear receptor superfamily that are encoded by the two genes, Thra and Thrb. 3,23In addition to the functional receptors TRα1, TRβ1, and TRβ2, further isoforms are expressed that fail to bind T 3 and may act as antagonists. 3,23While TRβ2 is restricted to the hypothalamus and pituitary to control the hypothalamic-pituitary-thyroid axis, TRα1 and TRβ1 are widely expressed but may show a temporospatialdependent expression during developmental stages and in adulthood. 3,24Given that the TR-binding affinity of T 3 is 15-fold higher than that of T 4 , in the periphery, T 4 is mostly converted by deiodinase 1 and 2 into T 3 . 3In bone tissue, both TRα1 and TRβ1 are expressed, however, due its 10-fold higher mRNA expression, TRα1 is proposed to be the main mediator of T 3 actions in the skeleton. 25,26In interaction with their coactivators, corepressors and the coreceptor retinoid X receptor, ligand-bound or free TR bind to thyroid hormone response elements (TRE) in promotor regions of target genes to induce or inhibit their transcription. 3,27,28Of note, T 3 -bound versus non-liganded TRs can compete for TRE to start or repress a DNA response via chromatin remodeling. 3,27,28In general, additional nongenomic actions of TH aside from the transcriptional response have been reported mediated by TR or αVβ3 integrin via MAPK and PI3K pathways at the plasma membrane and in the cytoplasm. 28,29everal TR knockout mouse models have been studied to unveil the importance of TR in skeletal development and bone health in adulthood and a comprehensive review can be found elsewhere. 3Thra 0/0 mice that express no TRα isoforms, but are characterized by normal Thrb expression, show normal circulating T 4 , T 3 , and TSH levels. 3,30Although Thra 0/0 juveniles display a transient growth delay, delayed ossification and reduced mineral deposition, adult Thra 0/0 mice are characterized by an osteosclerotic phenotype with increased bone volume and impaired bone resorption due to the absence of TRα1-mediated T 3 action in bone ("local hypothyroidism"). 3,30Thrb −/− mice without TRβ and normal TRα expression suffer from resistance to thyroid hormones (RTH) syndrome with elevated T 4 (3-4-fold), T 3 (3-4-fold) and TSH (2.6-8-fold) concentrations in the serum as compared to WT mice. 3,30,31hrb −/− mutants are of persistent short stature and advanced bone age given the progressive endochondral and intramembranous ossification during growth. 3,30,31dult Thrb −/− mice have osteoporotic bones with low bone mass and impaired bone quality due an accelerated bone remodeling. 3,30,31Given these findings in Thrb −/− mice, systemic thyrotoxicosis might bias the cell-intrinsic effects of Thrb knockout on skeletal cells.
Here, we investigated the activity, differentiation, and T 3 responsiveness of Thra and Thrb deficient osteoblasts when taken out of their biased environment in vivo to unravel the direct effects of TR ablation at the cellular level.

| Animals
For our in vitro studies and collection of bones samples, we used adult male Thra 0/0 and Thrb −/− mice on C57BL/6 background as well as respective age and sexmatched wildtype (WT) littermates as controls that were a kind gift from Prof. Lars Möller (University Hospital Essen).Mice with a homozygous mutation of Thra 0/0 and Thrb −/− do not express any Thra isoforms or functional TRβ1 or TRβ2 variants, respectively, as described previously. 24,30,32Mice were maintained in groups up to four animals in a light-dark cycle of 12/12 h at room temperature in filter-top cages with cardboard houses for enrichment purposes and had ad libitum access to their respective drinking water and standard chow diet.Genotyping was performed by PCR analysis.At the age of 18-22 weeks, mice were euthanized using CO 2 .Bone marrow mesenchymal stromal cells (BMMSCs) for osteoblast culture were obtained by flushing long bones (femora, tibiae, and humeri) and the bone tissue was immediately frozen with liquid nitrogen and stored at −80℃ until further RNA isolation.
Each in vitro experiment was performed at least twice with a number of 3-4 animals per group.All subsequent analyses (RNA analysis, ALP activity measurement, Alizarin Red staining, tartrate-resistant acid phosphatase [TRAP] staining) were performed in a blinded manner.
Animal procedures and organ collection were approved by the institutional animal care committee of the Technische Universität Dresden and the Landesdirektion Sachsen (TVT 03/2023) and were performed according to the ARRIVE guidelines.

| ALP activity assay
ALP activity was evaluated as a marker of osteoblast differentiation.After the 72 h treatment in DMEM supplemented with 1% FCS and 1% P/S, osteoblasts were washed twice in PBS and then scraped in ALP lysis buffer (10 mM Tris-HCl pH 8.0, 1 mM MgCl 2 , and 0.5% Triton X-100) for isolation of total proteins.Subsequently, cell lysates were processed through a 24-gauge needle and centrifuged at 25,000 × g for 30 min at 4℃.The 10 µL of diluted supernatants (1:5 in distilled water) were incubated with 90 µL ALP substrate buffer (100 mM diethanolamine, 150 mM NaCl, 2 mM MgCl 2 , and 2.5 g/ml p-nitrophenylphosphate) for 30 min at 37℃.The enzymatic activity of ALP leads to a color change due to the hydrolysis of p-nitrophenylphosphate.Absorbance was measured at 410 nm using FluoStar Omega (BMG Labtech).Results were normalized to the total protein content that was quantified using Pierce™ BCA Protein Assay Kit (Thermo Fisher Scientific) following the manufacturer's protocol.

| Alizarin Red S staining
To assess the mineralization capacity, BMMSCs were differentiated and simultaneously treated with 100 nM T 3 in DMEM supplemented with 10% FCS and 1%P/S over 10 and 21 d, respectively, then fixed in 10% paraformaldehyde in PBS for 15 min and stained with 1% Alizarin Red S solution (pH 5.5, Sigma-Aldrich) for 30 min at room temperature.To remove excess stain, repeated washing steps with distilled water were performed.100 mM cetylpyridinium chloride solution (Sigma-Aldrich) was used to dissolve incorporated calcium that was then quantified by photometric measurement at a wavelength of 540 nm using FluoStar Omega.

| Indirect coculture experiments and TRAP staining
For indirect coculture experiments, supernatants were collected from over 10 d differentiated WT and Thrbdeficient osteoblasts that were cultured beforehand with or w/o 100 nM T 3 over 48 h in DMEM with 1% FCS and 1% P/S.To obtain osteoclasts, bone marrow cells from 12-week-old male C57BL/6JRj were cultured in a-MEM supplemented with 10% FCS, 1% P/S, and 10 ng/mL macrophage colony-stimulating factor (MCSF, from R&D Systems) over 24 h in a cell culture flask.Only the floating cells (mostly myeloid precursors) were collected and grown in medium supplemented with 25 ng/mL M-CSF over 48 h.Osteoclastogenesis was induced by additionally adding 50 ng/mL receptor activator of nuclear factor kappa-B ligand (RANKL, from R&D Systems) over 4 d.Early osteoclasts were then treated with fresh a-MEM supplemented with 1% FCS and 1% P/S and osteoblast supernatants (1:1 ratio) supplemented with 25 ng/mL M-CSF and 50 ng/mL RANKL over 48 and 72 h, respectively, and samples were used for subsequent RNA isolation and TRAP staining.As controls, we used osteoclasts treated with a-MEM (1% FCS, 1% P/S) and osteoblast medium (DMEM, 1% FCS, 1% P/S) (1:1 ratio) supplemented with 25 ng/mL M-CSF and 50 ng/mL RANKL (CO) and directly added 100 nM T 3 to the medium (T3).
The number of TRAP-positive osteoclasts was quantified by TRAP staining at d 7 after starting RANKL treatment.Therefore, cells were first fixated for 10 min in acetone/citrate buffer containing 37% paraformaldehyde at room temperature.After two washing steps with tap water, staining was performed over 20 min in the dark using TRAP-Kit (from Sigma-Aldrich).TRAP-positive cells with three or more nuclei were counted as osteoclasts.

| RNA isolation, real-time polymerase chain reaction (RT-PCR), and quantitative RT-PCR
Total RNA from primary osteoblasts, primary osteoclasts and bone tissue was extracted using ReliaPrep TM RNA Tissue Miniprep System (Promega) and TRIzol reagent (Invitrogen), respectively, according to the manufacturer's protocol and quantified using a Nanodrop spectrophotometer (Peqlab).
Five hundred nanograms of RNA were reversetranscribed using M-MLV reverse transcriptase (Promega) followed by GoTaq® qPCR Master Mixbased quantitative real-time PCR (Promega) following established protocols (ABI7500 Fast; Applied Biosystems).Primer sequences for mice are listed in Supporting Information Table S1.PCR conditions were: 95℃ for 2 min followed by 40 cycles with 95℃ for 15 s and 60℃ for 1 min.Melting curves were evaluated using the following scheme: 95℃ for 15 s, 60℃ for 1 min and 95℃ for 30 s. Results were calculated based on the ΔΔCT method, are normalized to β-actin mRNA levels and are shown as X-fold increase compared to respective untreated controls.

| Statistical analysis
Data are presented as mean ± standard deviation.Statistical analysis comparing two groups are based on a two-sided unpaired student's t-test using GraphPad Prism 9.0 (GraphPad).Statistical comparison of three groups was performed with a One-Way ANOVA followed by Bonferroni's multiple comparison post hoc test using GraphPad Prism 9.0.Two-way ANOVA was performed for experiments with more than three groups followed by Bonferroni's multiple comparison post-hoc test using GraphPad Prism 9.0.Values of p < 0.05 were considered statistically significant.Significant outliers were excluded based on Grubbs' test provided in GraphPad by Dotmatics (https://www.graphpad.com/quickcalcs/grubbs1/).

| Thra knockout benefits osteoblast differentiation and activity in vitro
While the bone phenotype of mice lacking either Thra or Thrb is well described, in vitro studies about TR-deficient osteoblasts and their activity and differentiation potential are still lacking.Early osteoblasts (d 7) derived from Thra 0/0 mice showed increased expression of early osteoblast marker genes Sp7 (2.8-fold), Runx2 (2.1-fold), Alpl (2.1-fold), and Col1a1 (2.6-fold) as compared to WT osteoblasts (Figure 1A).In line, more mature stages of Thra-deficient osteoblasts (d 14 and d 21) presented an upregulated expression of osteogenic genes such as Bglap (d 14: 9.9-fold; d 21: 12.6-fold) and Dmp1 (d 21: 7.3-fold) indicating that overall osteoblast differentiation is potentially enhanced with Thra knockout under cell culture conditions (Figure 1B,C).In addition, Thra deletion led to higher ALP activity at d 21 (1.6-fold) (Figure 1D) as well as advanced bone matrix mineralization at d 10 (1.4-fold) and d 21 (1.6-fold) (Figure 1E).To investigate the role of Thra in osteocytes, the predominant cell type in bone tissue, we used total RNA isolated from flushed long bone tissue of knockout and WT mice for expression analysis and found significantly decreased levels of Klf9 (−49.2%), a TH target gene, and Thrb (−35.6%) with Thra deletion (Figure S1A,B).Nevertheless, Thra-deficient osteocytes did not show an altered expression of common osteocyte marker genes (Figure S1C-E).In contrast to Thra knockout osteoblasts, early and mature Thrb-deficient osteoblasts did not display any significant changes of osteoblast marker gene expression, ALP activity or mineralization capacity as compared to WT cells (Figure 1F-J).In bone tissue from Thrb −/− mice, expression of Klf9 (1.9-fold) and Thra (1.6-fold) were upregulated (Figure S1F,G) reflecting systemic hyperthyroidism in these mice, while again osteocyte marker expression was not affected by TR knockout (Figure S1H-J).These findings elaborate the complex role of TRs in osteoblast versus osteocyte biology highlighting that a single receptor knockout does not hamper bone cell development but that a Thra knockout even improves osteoblast features.
To further establish that TH actions in osteoblasts are mainly mediated via TRβ1, we further used the TRβspecific agonist GC-1 that was previously reported to stimulate mouse and rat osteoblast-like cells (MC3T3-E1 and ROS 17/2.8 cells, respectively) in vitro. 33While Sp7 transcript levels were only significantly upregulated by T 3 (T 3 : 1.9-fold; GC-1: 1.6-fold, p = 0.089), both, T 3 and GC-1 treatment, enhanced the expression of Bglap (T 3 : 8.4-fold; GC-1: 6.1-fold) and Klf9 (T 3 : 2-fold; GC-1: 1.9-fold) in mature primary murine osteoblasts and further improved their mineralization capacity (T 3 : 1.7-fold; GC-1: 1.4-fold) as compared with untreated control cells (Figure 3A-D).Given the rather short treatment period of 48 h, we also challenged Thrbdeficient osteoblasts with T 3 for up to 10 d over the full period of their differentiation (Figure 3E-G).While expression of Bglap increased with the T 3 treatment period in WT cells, Thrb-deficient were not affected even after 10 d of combined treatment and differentiation (Figure 3E).Adding T 3 over 2 and 4 d significantly increased Klf9 expression (d 2: 1.7-fold, d 4: 1.5-fold), followed by upregulated expression of deiodinase 3 (Dio3), the TH inactivating enzyme, after 7 and 10 d with T 3 in WT cells only (d 7: 6.1-fold, d 10: 11.3-fold) (Figure 3F,G).In conclusion, our results indicate that Thrb expression is crucial for TH responsiveness of osteoblasts in vitro.

| Thrb-deficient osteoblasts present a low Rankl/Opg ratio
Given that TH actions on osteoclasts might rather be indirectly mediated by cells of the osteoblast lineage, 3,[16][17][18][19][20][21] we assessed the effects of T 3 on the osteoblastic expression of Rankl and Opg, the two major regulators of osteoclastogenesis, in dependence on their Thra and Thrb expression, respectively.While F I G U R E 1 Thra knockout promotes osteoblast differentiation and activity in vitro.Primary osteoblasts derived from Thra 0/0 and Thrb −/− mice, respectively, as well as from respective WT littermates were cultured in DMEM supplemented with 10% FCS and 1% P/S and tested regarding their expression of osteoblast marker genes and activity at an early (d 7), mature (d 10-14) and mature/late (d 21) stage of their differentiation.Using quantitative real-time PCR analysis, mRNA expression of (A) osterix (Sp7), runt-related transcription factor 2 (Runx2), alkaline phosphatase (Alpl), and collagen type I alpha 1 chain (Col1a1) was measured in Thra 0/0 and WT osteoblasts at d7 of differentiation.(B) At d 14, mRNA levels of mature osteoblast marker genes osteocalcin (Bglap), Alpl and Col1a were quantified.expression of Rankl was not affected by T 3 in DMEM supplemented with 1% FCS and 1%P/S in WT osteoblasts (Figure 4A,D), Opg mRNA levels were significantly downregulated (Figure 4B,E) resulting in an overall increased Rankl/Opg ratio (Figure 4C,F).Thra deletion did not significantly alter the expression of Rankl, Opg, and the Rankl/Opg ratio (Figure 4A-C).In contrast, Thrb-deficient osteoblasts displayed reduced Rankl (−51.2%) and Opg (−32.2%)expression as well as a low Rankl/Opg ratio (−55.2%) as compared to respective, untreated WT osteoblasts (Figure 4D-F).Interestingly, T 3 treatment did not significantly affect either expression of Rankl and Opg or Rankl/Opg ratio in both Thra 0/0 and Thrb −/− osteoblasts as compared with untreated respective mutant cells (Figure 4A-F).Moreover, the increase of Rankl/Opg ratio persisted during a T 3 treatment period between two to 7 d in WT osteoblasts (d 2: 3-fold; d 4: 2.9-fold; d 7: 2.8-fold; d 10: 2.3-fold, p = 0.14), but not in Thrb knockout cells (Figure 4G).
Given that RANKL and OPG are soluble factors secreted by osteoblasts, we used supernatants derived from WT and Thrb −/− osteoblasts to treat primary murine osteoclasts.TRAP staining revealed reduced numbers of TRAP-positive multinucleated osteoclasts with supernatant from Thrb −/− osteoblasts as compared with osteoclasts with osteoblast medium only (−25.7% vs. CO) and osteoclasts treated with WT osteoblast supernatant (−24.7% vs. WT/CO) (Figure 5A).Nevertheless, expression of late osteoclast marker genes Trap and Ctsk, but not early osteoclast marker Nfatc1 was enhanced in osteoclasts subjected to supernatants from T 3 -treated WT osteoblasts (Trap: 1.5-fold; Ctsk: 1.3-fold vs. control osteoclasts [medium only]), while supernatants derived from both untreated and T 3 -treated Thrb-deficient osteoblasts did not affect osteoclast marker expression (Figure 5B-D).With regard to TH signaling, Klf9, Dio3, and Thrb expression was enhanced in osteoclasts with supernatants from both T 3 -treated WT as well as Thrb knockout osteoblasts but not osteoclasts directly treated with T 3 (Figure 5E-G).Osteoclastic expression of Thra remained unchanged between the groups (Figure 5H).Thus, our results from indirect coculture experiments emphasize that THs might be acting indirectly on osteoclasts via osteoblast secreted factors and that Thrb deletion in osteoblasts can affect osteoclast numbers and gene expression, respectively.

| DISCUSSION
TH homoeostasis is indispensable for bone health and thus, thyroid disease may lead to impaired bone quality and increased bone fragility.Nevertheless, the cellular function of TRs that are indispensable for TH actions on bone has not been studied in depth yet.Given that osteoblasts are the primary TH target cell in bone so far, 3,21 we analyzed effects of Thra and Thrb knockout on osteoblast differentiation, TH responsiveness and osteoblast-osteoclast coupling in vitro.
Although TRα1 is suggested to be the primary mediator of T 3 effects on bone, 25,26 Thra knockout promoted osteoblast development function from early to late osteoblasts under cell culture conditions.Given the high complexity of TH signaling in the nucleus involving transcriptional activators and repressors, chromatin remodeling as well as co-receptors, we can only speculate on the underlying molecular mechanisms.Unbound TR can act as transcriptional repressors and the unbalanced intracellular ratio of TRα:TRβ can change the rate of TH target gene transcription, 3,27,28,34 thus Thra deletion might unlock advanced osteoblast features by releasing its transcriptional repression.Also, the predominance of TRβ in Thra knockout osteoblasts might benefit their differentiation and activity under cell culture conditions given that TRβ, but not TRα was essential to mediate T 3 actions in osteoblasts in vitro (Figure 2).Despite the pro-osteogenic effects of Thra knockout in vitro, Thra 0/0 mice show impaired bone mineral deposition at a juvenile age and more importantly low bone resorption in adulthood leading to high | 1955 bone mass. 30While pituitary production of growth hormone and circulating levels of estrogen, two major drivers of bone remodeling besides TH, were not affected in Thra 0/0 mice, 35,36 effects on other mainly TRα1-expressing tissues 37 and their production of osteomodulatory cytokines and hormones such as glucocorticoids, irisin, insulin-like growth factor 1 and adiponectin have not been studied yet.Interestingly, we found a decreased expression of Klf9 and Thrb in bone tissue of adult Thra 0/0 mice (Figure S1A,B) implying local TH-deprivation, but not in Thra 0/0 osteoblasts in vitro (Figure 2C,E).Further, osteocyte marker gene expression in bone was not changed by Thra knockout despite the local hypothyroidism (Figure S1C-E).Thus, discrepancies between the in vitro versus in vivo performance of Thra 0/0 osteoblasts and osteocytes need further comprehensive investigation, but may suggest indirect systemic actions on bone in vivo.
On the contrary to Thra knockout, Thrb loss did not affect osteoblast physiology, but impeded TH responsiveness of osteoblasts in vitro.The preclinical findings are contradictory.Due to a dysbalance of the hypothalamicpituitary-thyroid axis and thus, systemic thyrotoxicosis, Thrb −/− mice display high bone turnover-mediated bone loss as seen in hyperthyroid mice. 30Interestingly, Monfoulet and colleagues show that adult hypothyroid-rendered Thrb -/-mice, but not Thra 0/0 mice, were protected against rapid bone loss with pharmacologic T 4 treatment over 3 weeks. 38Given that a longer hyperthyroxinemic treatment (5 weeks) also led to a moderate osteoporotic phenotype in F I G U R E 4 Expression of Rankl and Opg is differentially regulated by TRs.Primary mature osteoblasts derived from Thra 0/0 and Thrb −/− mice, respectively, as well as from respective WT littermates were analyzed after 48-h treatment with 100 nM T 3 in DMEM supplemented with 1% FCS and 1% P/S.Based on quantitative real-time PCR analysis, mRNA expression of (A) receptor activator of NF-κB ligand (Rankl) and (B) osteoprotegerin (Opg) was quantified and (C) the Rankl/Opg ratio was calculated in Thra 0/0 and WT osteoblasts.In addition, (D) Rankl, (E) Opg and (F) Rankl/Opg ratio were assessed in Thrb −/− and respective WT osteoblasts after T 3 treatment in DMEM supplemented with 1% FCS and 1% P/S.(G) Thrb −/− and respective WT osteoblasts were treated with 100 nM T 3 in DMEM supplemented with 10% FCS and 1% P/S over 2, 4,  Thrb −/− mice, the authors concluded that TRβ mediates acute response to transient TH alterations in bone, while TRα controls the response to chronic high TH exposure. 38onsidering our findings that TRβ mediates T 3 actions in osteoblasts in vitro, these rapid versus slow TH actions on bone remodeling might be mediated by indirect pathways 38 through other TH-stimulated cells or factors within the bone marrow cavity or circulating in the periphery.To further establish the role of TRβ as a mediator of T 3 action in osteoblasts, we treated WT osteoblasts with the TRβ-specific thyromimetic GC-1.In accordance to a previous study using mouse and rat osteoblast-like cell lines (MC3T3-E1 and ROS 17/2.8 cells, respectively), 33 we found increased expression of osteoblast marker genes and improved mineralization capacity in primary murine osteoblasts treated with GC-1.A preclinical study using adult rats reported that GC-1 did not lead to bone loss as seen in hyperthyroid animals, 39 while another study showed that the same compound had osteoanabolic effects in young hypothyroid rats corroborating our in vitro findings. 40In conclusion, TRβ mediates TH actions in osteoblasts in vitro und thus, potentially on bone, however, skeletal consequences of TRβ-targeting therapies might vary based on the preclinical model and treatment period.
As hyperthyroid mice elicit bone loss due to a predominant bone resorption, 21,[41][42][43] we analyzed the ratio of Rankl/Opg expressed by TR-deficient osteoblasts as a link to osteoclastogenesis and bone resorption.Thrb −/− osteoblasts showed a lower Rankl/Opg ratio as compared with WT cells and correspondingly, we found mildly reduced numbers of TRAP-positive osteoclasts when treated at an early differentiation stage (d 4) with supernatants from mutant osteoblasts.Furthermore, osteoclasts treated with supernatants from T 3 -stimulated WT osteoblasts, but not T 3 -treated Thrb-deficient osteoblasts exhibited elevated expression of late osteoclast marker genes.However, due to the rather mild effects, additional experiments such as bone resorption assays as well as different time points for treatment might be needed to fully understand the nature of indirect TH actions on osteoclasts.In addition, multiple other osteoblast-derived factors could contribute to impaired osteoclastogenesis such as M-CSF, Wnt ligands 44 or miRNA 45 and thus, proteomic or miRNA array-based analyses of supernatants could help to better distinguish the impact of RANKL and OPG.In vivo, hyperthyroidism can stimulate other RANKL-producing cells such as immune cells 46,47 and in RTH patients increased levels of the proinflammatory and osteomodulatory cytokine tumor necrosis factor-α (TNF-α) have been reported [48][49][50] and therefore, increased bone turnover in Thrb −/− mice might not be based on direct TH actions in osteoblasts only.
Despite its strengths, this study has potential limitations.Due to the lack of specific TR antibodies, only mRNA transcript levels but not effects on protein expression are shown. 3Further, truncated TR and TR isoforms unable to bind T 3 might also play a major role in osteoblast physiology and thus, future studies should focus on unraveling the role of distinct TR isoforms.Downstream of TR, not only TRinduced transcriptional response but also activated MAPK, BMP/SMAD, and PI3K signaling were reported in different TH stimulated tissue and cell types 21,28,29 that might be also affected in TR-deficient osteoblasts and bone tissue in general.Given that systemic TR knockouts can affect circulating THs levels and the whole body's metabolism and thus, cell-intrinsic effects of TR are potentially masked, bone-specific Thra/Thrb-floxed knockout mouse models would present a useful tool to decipher the particular role of each TR in skeletal cells and bone remodeling in vivo.
In conclusion, this study highlights the autonomous and nonredundant functions of TR in the development and TH responsiveness of osteoblasts in vitro.While Thra knockout promoted osteoblast differentiation and activity under basal cell culture conditions, Thrb loss hampered the TH response of osteoblasts and affected osteoblast-osteoclast crosstalk in vitro, corroborating the importance of TRβ1-mediated signaling in bone.
boards from Alexion, Amgen, Kyowa Kirin International, and UCB to his institution and himself.Elena Tsourdi reports honoraria for lectures from Amgen, UCB, Takeda, Alexion and Ascendis and educational grants from UCB and Takeda.Franziska Lademann has nothing to disclose.
Further, (C) d 21 Thra 0/0 and WT osteoblasts were tested with regard to their Bglap, Col1a1 and dentin matrix acidic phosphoprotein 1 (Dmp1) expression.(D) In addition, ALP activity was assessed at d 10 and d21 in Thra 0/0 and WT osteoblasts.(E) Mineralization capacity was evaluated after differentiation over 10 and 21 d.Representative images of Alizarin Red stained bone matrix are shown for both Thra 0/0 and WT osteoblasts.Moreover, expression of osteoblast marker genes in (F) d 7, (G) d 14, and (H) d 21 Thrb −/− and WT osteoblasts was measured by quantitative real-time PCR.With regard to osteoblasts activity, (I) ALP activity and (J) mineralization capacity of Thrb −/− and WT osteoblasts were assessed at d 10 and d 21.Representative images of Alizarin Red stained bone matrix are shown for both Thrb −/− and WT osteoblasts.Each dot indicates an individual mouse.The horizontal lines represent the mean +/− SD.Thra WT: N = 9, Thra 0/0 : N = 8; Thrb WT: N = 8, Thrb −/− : N = 10.Statistical analysis was performed by two-sided student's t-test versus respective WT control and significant p values are shown within the graph.

F
I G U R E 2 T 3 responsiveness of osteoblasts depends on Thrb expression.Primary mature osteoblasts derived from Thra 0/0 and Thrb −/− mice, respectively, as well as from respective WT littermates were tested regarding their T 3 response in vitro.Using quantitative real-time PCR analysis, mRNA expression of (A) osteocalcin (Bglap), (B) osterix (Sp7), and (C) krueppel-like factor 9 (Klf9), a thyroid hormone target gene, was quantified in Thra 0/0 and WT osteoblasts after 48-h treatment with 100 nM T 3 in DMEM supplemented with 1% FCS and 1%P/S.In addition, (D) mineralization capacity was assessed in Thra 0/0 and WT osteoblasts after 10 d of differentiation and simultaneous T 3 treatment in DMEM supplemented with 10% FCS and 1%P/S.Also, (E) expression of Thrb was quantified in Thra-deficient and WT osteoblasts treated with T 3 over 48 h.Moreover, mRNA levels of (F) Bglap, (G) Sp7, and (H) Klf9 were measured in Thrb −/− and respective WT osteoblasts after 48 h with 100 nM T 3 in DMEM supplemented with 1% FCS and 1%P/S.(I) Mineralization capacity of Thrb knockout and WT osteoblasts after simultaneous differentiation and T 3 treatment in DMEM supplemented with 10% FCS and 1%P/S over 10 d. (J) Expression of Thra in Thrb −/− and WT osteoblasts determined by quantitative real-time PCR.Each dot indicates an individual mouse.The horizontal lines represent the mean +/− SD.Thra WT: N = 10, Thra 0/0 : N = 10; Thrb WT: N = 10, Thrb −/− : N = 9.Statistical analysis was performed by two-way ANOVA and significant p values are shown within the graph.

F
I G U R E 3 TRβ-specific agonist GC-1 stimulates osteoblast differentiation and function in vitro.Mature primary osteoblasts obtained from WT mice were treated over 48 h with 100 nM T 3 and 100 nM GC-1, respectively, in DMEM supplemented with 1% FCS and 1%P/S and the expression of (A) osteocalcin (Bglap), (B) osterix (Sp7), and (C) krueppel-like factor 9 (Klf9) was quantified.(D) Mineralization capacity of WT osteoblasts differentiated and treated with T 3 and GC-1 over 10 d was assessed by Alizarin red staining.Further, Thrb −/− and respective WT osteoblasts were treated with 100 nM T 3 in DMEM supplemented with 10% FCS and 1%P/S over 2, 4, 7, and 10 d during their differentiation until the end of the experiment at d 10.Using quantitative real-time PCR, expression of (E) Bglap, (F) Klf9, and (G) deiodinase 3 (Dio3), a thyroid hormone inactivating enzyme, was measured after the indicated treatment period at d 10.Each dot indicates an individual mouse.The horizontal lines represent the mean +/− SD.Panel A-D: N = 6-10 per group.Statistical analysis was performed by one-way ANOVA.Panel E-G: Thrb WT: N = 5-9, Thrb −/− : N = 4-9.Statistical analysis was performed by two-way ANOVA.Significant p values are shown within the graph.LADEMANN ET AL.
7, and 10 d during their differentiation until the end of the experiment at d 10 and the Rankl/Opg ratio was assessed after the indicated treatment periods.Each dot indicates an individual mouse.The horizontal lines represent the mean +/− SD.Panel A-F: Thra WT: N = 10, Thra 0/0 : N = 12; Thrb WT: N = 10, Thrb −/− : N = 12.Statistical analysis was performed by two-way ANOVA.Panel G: Thrb WT: N = 5-8, Thrb −/− : N = 4-8.Statistical analysis was performed by two-way ANOVA.Selected p values are shown within the graph.

F
I G U R E 5 Supernatants of T 3 -treated wildtype osteoblasts increase late osteoclast marker gene expression in wildtype osteoclasts in vitro.Primary early osteoclasts were treated over 48 and 72 h, respectively, with supernatants from WT and Thrb-deficient osteoblasts that remained either untreated (WT/CO and Thrb −/− /CO) or were challenged with 100 nM T 3 (WT/T3 and Thrb −/− /T3).In addition, we used osteoclasts only (CO) and osteoclasts directly treated with 100 nM T 3 (T3) as controls.(A) Number of tartrate-resistant acid phosphatase (TRAP)-positive osteoclasts was quantified after a 72-h treatment period using TRAP staining at d 7 of osteoclast differentiation.Based on real-time quantitative PCR, expression of osteoclast marker genes (B) Trap, (C) cathepsin K (CtsK), and (D) nuclear factor of activated T cells 1 (Nfatc1) as well as (E) Klf9, (F) Dio3, (G) Thrb, and (H) Thra was assessed in osteoclasts after 48-h treatment.Each dot indicates an individual mouse.The horizontal lines represent the mean +/− SD.N = 7-9 per group.Statistical analysis was performed by one-way ANOVA.Significant p-values are shown within the graph.