Accelerated Bone Loss in Transgenic Mice Expressing Constitutively Active TGF-β Receptor Type I

Transforming growth factor beta (TGF-β) is a key factor mediating the intercellular crosstalk between the hematopoietic stem cells and their microenvironment. Here, we investigated the skeletal phenotype of transgenic mice expressing constitutively active TGF-β receptor type I under the control of Mx1-Cre (Mx1;TβRICA mice). μCT analysis showed decreased cortical thickness, and cancellous bone volume in both femurs and mandibles. Histomorphometric analysis confirmed a decrease in cancellous bone volume due to increased osteoclast number and decreased osteoblast number. Primary osteoblasts showed decreased ALP and mineralization. Constitutive TβRI activation increased osteoclast differentiation. qPCR analysis showed that Tnfsf11/Tnfrsf11b ratio, Ctsk, Sufu, and Csf1 were increased whereas Runx2, Ptch1, and Ptch2 were decreased in Mx1;TβRICA femurs. Interestingly, Gli1, Wnt3a, Sp7, Alpl, Ptch1, Ptch2, and Shh mRNA expression were reduced whereas Tnfsf11/Tnfrsf11b ratio was increased in Mx1;TβRICA mandibles. Similarly, osteoclast-related genes were increased in Mx1;TβRICA osteoclasts whereas osteoblast-related genes were reduced in Mx1;TβRICA osteoblasts. Western blot analysis indicated that SMAD2 and SMAD3 phosphorylation was increased in Mx1;TβRICA osteoblasts, and SMAD3 phosphorylation was increased in Mx1;TβRICA osteoclasts. CTSK was increased while RUNX2 and PTCH1 was decreased in Mx1;TβRICA mice. Microindentation analysis indicated decreased hardness in Mx1;TβRICA mice. Our study indicated that Mx1;TβRICA mice were osteopenic by increasing osteoclast number and decreasing osteoblast number, possibly by suppressing Hedgehog signaling pathways.


Introduction
Transforming growth factor-beta (TGF-β) is a multi-functional growth factor that regulates proliferation, differentiation, apoptosis, and migration of many cell types including skeletal cells [1]. Three different isoforms of TGF-β-TGF-β1, TGF-β2 and TGF-β3-have been identified in mammals. However, TGF-β1 is the most abundant and widely expressed isoform, especially in bone. It plays a crucial roles in bone development and homeostasis, stimulates matrix protein synthesis, and is responsible for bone formation and resorption [2]. TGF-β is secreted as an inactive precursor which contains the signal peptide, the latency-associated protein (LAP), comprising the so-called latent TGF-β complex. Upon cleavage of the LAP by proteases, the active TGF-β is released. TGF-β acts through a heteromeric receptor complex of two types of transmembrane serine/threonine kinase receptors, type I (TβRI) and type II receptors (TβRII), at the cell surface. Canonical TGF-β signaling occurs when ligands bind to TβRII and phosphorylate TβRI. Phosphorylated TβRI initiates intracellular signaling through subsequent phosphorylation of downstream SMAD2 and SMAD3. The phosphorylated SMAD2 and SMAD3 form a complex with SMAD4 and translocate into the nucleus to regulate target gene transcription. TGF-β ligands also use non-canonical (non-SMAD2/3) signaling pathways to regulate cellular functions. TGF-β ligands can exert signals through activation of MAPK pathway responses, including activation of the p38, JNK, and ERK pathways, phosphoinositide 3-kinase/Akt, Rho GTPase, Wnt, and Notch signaling. The effects of TGF-β on MAPK signaling include both SMAD-dependent and SMAD-independent pathways [1]. Bone cells express TGF-β1, TGF-β2, and TGF-β3, and TGF-β induces both positive and negative effects on bone cells. TGF-β1 increases bone formation by recruiting osteoblast progenitors and stimulating their proliferation and differentiation in the early stages. On the other hand, it blocks the later phases of differentiation and mineralization [3]. Inhibition of TGF-β with the 1D11 antibody clone resulted in increased bone mineral density, trabecular thickness, and bone volume while decreasing active osteoclasts in bone marrow [4]. A biphasic effect of TGF-β1 on osteoclasts has also been reported. TGF-β1 stimulated osteoclast development at low doses by enhancing the Tnfsf11/Tnfrsf11b (RANKL/OPG) ratio while high concentrations of TGF-β1 downregulated Tnfsf11, leading to a reduced Tnfsf11/Tnfrsf11b ratio and a decrease in the numbers of osteoclast precursors [5].
Tgfb1 −/− Rag2 −/− mice had reduced cancellous bone density and osteoblast number [6]. Osteoblast-specific mutant Tgfb1 transgenic mice exhibited abnormal remodeling of mandibular condylar subchondral bone, including cartilage degradation and excessive chondrocyte apoptosis [7]. High concentrations of active TGF-β1 in subchondral bone induced abnormal subchondral bone formation and degeneration of articular cartilage, leading to osteoarthritis [8]. On the other hand, TβRII −/− nestin + mesenchymal progenitors decreased osteoarthritis development in an anterior cruciate ligament transection model. Deletion of TβRII in limbs using Prx1-cre resulted in the defects in joint development and a failure of chondrocyte hypertrophy [9]. Transgenic mice expressing a truncated kinase-defective TβRII, which acted as a dominant loss-of-function mutation, developed kyphoscoliosis and cartilage disorganization resembling osteoarthritis [10]. Moreover, Tgfb2-null mice caused defects in the skull base and vertebrae with reduced ossification, and delayed development [11]. Taken together, the effects of TGF-β or its receptors very much depend on both the target cells and biological context. However, no previous studies had been reported on the effects of constitutive TβRI activation in hematopoietic stem-cells-derived osteoclasts and osteoblasts on bone turnover.
We examined the impact of constitutive TβRI activation on bone turnover in 9-weekold transgenic mice with inducible expression of a constitutively active TβRI in hematopoietic stem-cells-derived osteoclasts and osteoblasts (Mx1;TβRI CA mice). Serum calcium levels were decreased whereas serum parathyroid hormone (PTH) levels were increased in Mx1;TβRI CA mice. Mx1;TβRI CA mice were osteopenic due to increased osteoclast numbers and decreased osteoblast numbers in both long bones and mandibles. Constitutive TβRI activation decreased osteoblast and increased osteoclast differentiation. Osteoclastrelated genes including the Tnfsf11/Tnfrsf11b ratio, Ctsk, Sufu, and Csf1 were increased, whereas Runx2, Ptch1, and Ptch2 osteoblast-related genes were decreased in femurs. In mandibles, qPCR analysis indicated that Gli1, Wnt3a, Sp7, Alpl, Ptch1, Ptch2, and Shh were decreased, but the Tnfsf11/Tnfrsf11b ratio was increased. In addition, osteoclast-related genes were increased in Mx1;TβRI CA osteoclasts, whereas osteoblast-related genes were reduced in Mx1;TβRI CA osteoblasts. pSMAD2/total SMAD2 and pSMAD3/total SMAD3 was enhanced in Mx1;TβRI CA osteoblasts, whereas pSMAD3/total SMAD3 was increased in Mx1;TβRI CA osteoclasts. CTSK protein levels were increased, while RUNX2 and PTCH1 protein levels were decreased in Mx1;TβRI CA mice. Mx1;TβRI CA mice displayed decreased bone hardness.

Generation of Mx1;TβRI CA Mice
TβRI CA lox/lox (TβRI CA ) mice were crossed with mice expressing Cre recombinase under the control of the Mx1 promoter induced by a type I interferon following injection of polyinosinic-polycytidylic acid (poly (I:C) to generate Mx1;TβRI CA mice. Using pRCA/pPA primers, PCR analysis of tail DNA demonstrated lox/lox bands in WT and transgenic mice ( Figure 1). As expected, WT (+/+) but not TβRI CA lox/lox mice showed a band using pHPRT primers, representing disruption of this locus via targeted insertion of the TβRI CA lox/lox transgene. However, Mx1-Cre primers detected bands only in Mx1;TβRI CA mice indicating effective constitutive expression of TβRI in hematopoietic stem-cells-derived osteoclasts and osteoblasts.

Generation of Mx1;TβRI CA Mice
TβRI CA lox/lox (TβRI CA ) mice were crossed with mice expressing Cre recombinase under the control of the Mx1 promoter induced by a type I interferon following injection of polyinosinic-polycytidylic acid (poly (I:C) to generate Mx1;TβRI CA mice. Using pRCA/pPA primers, PCR analysis of tail DNA demonstrated lox/lox bands in WT and transgenic mice ( Figure 1). As expected, WT (+/+) but not TβRI CA lox/lox mice showed a band using pHPRT primers, representing disruption of this locus via targeted insertion of the TβRI CA lox/lox transgene. However, Mx1-Cre primers detected bands only in Mx1;TβRI CA mice indicating effective constitutive expression of TβRI in hematopoietic stem-cells-derived osteoclasts and osteoblasts.

Mx1;TβRI CA Mice Decreases Serum Calcium and Increases Serum PTH Levels
To examine the changes in serum chemistries in Mx1;TβRI CA mice, serum phosphorus, calcium, and PTH were determined. There were no significant differences in serum phosphorus levels between the two groups ( Figure 2A). Serum calcium levels were decreased while serum PTH levels were increased in Mx1;TβRI CA mice compare to WT controls ( Figure 2B-C).

Mx1;TβRI CA Mice Decreases Serum Calcium and Increases Serum PTH Levels
To examine the changes in serum chemistries in Mx1;TβRI CA mice, serum phosphorus, calcium, and PTH were determined. There were no significant differences in serum phosphorus levels between the two groups ( Figure 2A). Serum calcium levels were decreased while serum PTH levels were increased in Mx1;TβRI CA mice compare to WT controls ( Figure 2B,C).

Generation of Mx1;TβRI CA Mice
TβRI CA lox/lox (TβRI CA ) mice were crossed with mice expressing Cre recombinase under the control of the Mx1 promoter induced by a type I interferon following injection of polyinosinic-polycytidylic acid (poly (I:C) to generate Mx1;TβRI CA mice. Using pRCA/pPA primers, PCR analysis of tail DNA demonstrated lox/lox bands in WT and transgenic mice ( Figure 1). As expected, WT (+/+) but not TβRI CA lox/lox mice showed a band using pHPRT primers, representing disruption of this locus via targeted insertion of the TβRI CA lox/lox transgene. However, Mx1-Cre primers detected bands only in Mx1;TβRI CA mice indicating effective constitutive expression of TβRI in hematopoietic stem-cells-derived osteoclasts and osteoblasts.

Mx1;TβRI CA Mice Decreases Serum Calcium and Increases Serum PTH Levels
To examine the changes in serum chemistries in Mx1;TβRI CA mice, serum phosphorus, calcium, and PTH were determined. There were no significant differences in serum phosphorus levels between the two groups ( Figure 2A). Serum calcium levels were decreased while serum PTH levels were increased in Mx1;TβRI CA mice compare to WT controls ( Figure 2B-C).

Mx1;TβRI CA Mice Have Cancellous and Cortical Bone Loss in Femurs and Mandibles
There was no significant difference in body weight between Mx1;TβRI CA mice and WT mice ( Figure 3A). Femur and mandible length was similar ( Figure 3B). To evaluate whether the constitutive activation of TβRI impacted femoral and mandibular cancellous and cortical bone in 9-week-old mice, µCT analysis was performed. µCT images of femoral and mandibular bone in WT and Mx1;TβRI CA mice are shown in Figure 3C. Mx1;TβRI CA mice were osteopenic in both femurs and mandibles. Bone microstructural parameters are shown in Figure 3D. Both femurs and mandibles in Mx1;TβRI CA mice showed a significant decrease in cancellous bone volume, trabecular thickness, and bone mineral density with a concomitant increase in trabecular separation and structural model index compared to WT controls. Trabecular number and connectivity density was decreased in the femur whereas it did not change in mandibles. The skeletal phenotype in mandibles is milder than femurs. Cortical thickness and bone mineral density were significantly reduced in both the femurs and mandibles of Mx1;TβRI CA mice ( Figure 3E). These findings demonstrated that femoral and mandibular bone loss was occurred in both cancellous and cortical bone in Mx1;TβRI CA mice.

Mx1;TβRI CA Mice Have Cancellous and Cortical Bone Loss in Femurs and Mandibles
There was no significant difference in body weight between Mx1;TβRI CA mice and WT mice ( Figure 3A). Femur and mandible length was similar ( Figure 3B). To evaluate whether the constitutive activation of TβRI impacted femoral and mandibular cancellous and cortical bone in 9-week-old mice, µCT analysis was performed. µCT images of femoral and mandibular bone in WT and Mx1;TβRI CA mice are shown in Figure 3C. Mx1;TβRI CA mice were osteopenic in both femurs and mandibles. Bone microstructural parameters are shown in Figure 3D. Both femurs and mandibles in Mx1;TβRI CA mice showed a significant decrease in cancellous bone volume, trabecular thickness, and bone mineral density with a concomitant increase in trabecular separation and structural model index compared to WT controls. Trabecular number and connectivity density was decreased in the femur whereas it did not change in mandibles. The skeletal phenotype in mandibles is milder than femurs. Cortical thickness and bone mineral density were significantly reduced in both the femurs and mandibles of Mx1;TβRI CA mice ( Figure 3E). These findings demonstrated that femoral and mandibular bone loss was occurred in both cancellous and cortical bone in Mx1;TβRI CA mice.

Mx1;TβRI CA Mice Had Decreased Bone Formation and Increased Bone Resorption
Similar to µCT data, histomorphometric analysis of tibiae and mandibles indicated cancellous bone loss. As shown in Table 1, Mx1;TβRI CA mice had decreased cancellous bone volume, trabecular thickness, trabecular number, and increased trabecular separation compared to WT mice. Furthermore, osteoblast surface per bone surface, osteoblast number per bone perimeter, and osteoblast number per tissue area were reduced in Mx1;TβRI CA mice. In addition, TβRI CA mice had increased osteoclast surface per bone surface, osteoclast number per bone perimeter, osteoclast number per tissue area, and eroded

Mx1;TβRI CA Mice Had Decreased Bone Formation and Increased Bone Resorption
Similar to µCT data, histomorphometric analysis of tibiae and mandibles indicated cancellous bone loss. As shown in Table 1, Mx1;TβRI CA mice had decreased cancellous bone volume, trabecular thickness, trabecular number, and increased trabecular separation compared to WT mice. Furthermore, osteoblast surface per bone surface, osteoblast number per bone perimeter, and osteoblast number per tissue area were reduced in Mx1;TβRI CA mice. In addition, TβRI CA mice had increased osteoclast surface per bone surface, osteoclast number per bone perimeter, osteoclast number per tissue area, and eroded surface in Mx1;TβRI CA mice. Therefore, our data confirmed that Mx1;TβRI CA mice displayed altered bone turnover due to decreased bone formation and increased bone resorption leading to bone loss.

Mx1;TβRI CA Mice Have Decreased Osteoblast Differentiation and Increased Osteoclast Differentiation
Histomorphometry indicated that constitutive activation of TβRI led to a decrease in osteoblast number and increase in osteoclast number. To confirm that constitutive TβRI activation decreased osteoblast differentiation, we performed an in vitro assay using primary osteoblast progenitors derived from long bones and mandibles. ALP and mineralized bone nodules were decreased in osteoblasts derived from both Mx1;TβRI CA long bones and mandibles, indicating that exogenous mutated TβRI reduced osteoblast differentiation ( Figure 4). surface in Mx1;TβRI CA mice. Therefore, our data confirmed that Mx1;TβRI CA mice displayed altered bone turnover due to decreased bone formation and increased bone resorption leading to bone loss. Results are mean ± SEM. * p < 0.05 compared to WT.

Mx1;TβRI CA Mice Have Decreased Osteoblast Differentiation and Increased Osteoclast Differentiation
Histomorphometry indicated that constitutive activation of TβRI led to a decrease in osteoblast number and increase in osteoclast number. To confirm that constitutive TβRI activation decreased osteoblast differentiation, we performed an in vitro assay using primary osteoblast progenitors derived from long bones and mandibles. ALP and mineralized bone nodules were decreased in osteoblasts derived from both Mx1;TβRI CA long bones and mandibles, indicating that exogenous mutated TβRI reduced osteoblast differentiation ( Figure 4). We next determined whether Mx1;TβRI CA mice had increased osteoclast number. An in vitro osteoclast differentiation assay was performed using long bone and mandibular osteoclast precursors. Consistent with the increased bone resorption observed in vivo, constitutive TβRI activation increased TRAP positive osteoclasts derived from bone marrow macrophages (BMMs) of Mx1;TβRI CA mice compared to WT controls ( Figure 5). These data revealed a major role of TβRI in osteoblast and osteoclast differentiation. We next determined whether Mx1;TβRI CA mice had increased osteoclast number in vitro osteoclast differentiation assay was performed using long bone and mandib osteoclast precursors. Consistent with the increased bone resorption observed in v constitutive TβRI activation increased TRAP positive osteoclasts derived from bone m row macrophages (BMMs) of Mx1;TβRI CA mice compared to WT controls ( Figure 5). Th data revealed a major role of TβRI in osteoblast and osteoclast differentiation.

Mx1;TβRI CA Mice Have Decreased Osteoblast and Increased Osteoclast-Related Gene Ex pression
To investigate the mechanisms by which constitutive activation of TβRI decrea bone formation and increased bone resorption, qPCR was performed to determine expression of osteoblast-and osteoclast-associated transcripts in femurs and mandib The expression of Runx2, a key transcription factor for osteoblast differentiation, was nificantly reduced in Mx1;TβRI CA femurs compared to WT controls ( Figure 6A). There no significant difference in the expression of a number of genes influencing osteobla genesis, including Gli1, Wnt3a, Fgf23, Sp7, Alpl, Ibsp, Col1a1, Bglap, and Shh. For mand lar bones, the levels of osteoblast-related genes, Sp7, and Alpl were decreased Mx1;TβRI CA mice ( Figure 6B). Fgf23, Runx2, Ibsp, Col1a1, and Bglap were not altered, ar ing that the reduction in Runx2 expression in femurs observed reflected a specific do regulation of Runx2 itself as opposed to reflecting a general paucity of osteoblasts. Hed hog and Wnt signaling are essential regulators of osteoblast proliferation and differen tion, through the expression of Gli1, Shh, Ptch1, Ptch2, and Wnt3a. Gli1, Shh, and W were reduced in Mx1;TβRI CA mandibles but not femurs. However, Ptch1 and Ptch2 w decreased in both Mx1;TβRI CA mandibles and femurs. Bone loss depends on the rati Tnfsf11/Tnfrsf11b, as this ratio is a major regulator of osteoclast differentiation and vival. Tnfsf11/Tnfrsf11b gene expression was significantly increased in both femurs mandibles of Mx1;TβRI CA mice compared to WT controls ( Figure 6A,B). Moreover, C Sufu, and Csf1, regulators of bone resorption, were increased in femoral bone  To investigate the mechanisms by which constitutive activation of TβRI decreased bone formation and increased bone resorption, qPCR was performed to determine the expression of osteoblast-and osteoclast-associated transcripts in femurs and mandibles. The expression of Runx2, a key transcription factor for osteoblast differentiation, was significantly reduced in Mx1;TβRI CA femurs compared to WT controls ( Figure 6A). There was no significant difference in the expression of a number of genes influencing osteoblastogenesis, including Gli1, Wnt3a, Fgf23, Sp7, Alpl, Ibsp, Col1a1, Bglap, and Shh. For mandibular bones, the levels of osteoblast-related genes, Sp7, and Alpl were decreased in Mx1;TβRI CA mice ( Figure 6B). Fgf23, Runx2, Ibsp, Col1a1, and Bglap were not altered, arguing that the reduction in Runx2 expression in femurs observed reflected a specific downregulation of Runx2 itself as opposed to reflecting a general paucity of osteoblasts. Hedgehog and Wnt signaling are essential regulators of osteoblast proliferation and differentiation, through the expression of Gli1, Shh, Ptch1, Ptch2, and Wnt3a. Gli1, Shh, and Wnt3a were reduced in Mx1;TβRI CA mandibles but not femurs. However, Ptch1 and Ptch2 were decreased in both Mx1;TβRI CA mandibles and femurs. Bone loss depends on the ratio of Tnfsf11/Tnfrsf11b, as this ratio is a major regulator of osteoclast differentiation and survival. Tnfsf11/Tnfrsf11b gene expression was significantly increased in both femurs and mandibles of Mx1;TβRI CA mice compared to WT controls ( Figure 6A,B). Moreover, Ctsk, Sufu, and Csf1, regulators of bone resorption, were increased in femoral bone of Mx1;TβRI CA mice ( Figure 6A). These findings demonstrated that Mx1;TβRI CA mice exhibited femoral and mandibular bone loss associated with reduced osteoblast-and increased osteoclast-related gene expression. Mx1;TβRI CA mice ( Figure 6A). These findings demonstrated that Mx1;TβRI CA mice exhibited femoral and mandibular bone loss associated with reduced osteoblast-and increased osteoclast-related gene expression.  To confirm these gene expression changes observed in femurs and mandibles, osteoblasts and osteoclasts derived from long bones were studied to determine the expression of osteoblast-and osteoclast-related genes by qPCR analysis in vitro. The gene expression levels of TβRI significantly increased in Mx1;TβRI CA osteoblasts and osteoclasts compared to WT controls, confirming that Mx1;TβRI CA cells had an overexpression of TβRI as expected ( Figure 7A). Tgfb1 expression significantly increased in osteoclasts but not osteoblasts. However, there was no difference in TβRII expression in osteoblasts and osteoclasts between Mx1;TβRI CA and WT mice. The expression of Runx2, a key transcription factors for osteoblast differentiation, was significantly reduced in Mx1;TβRI CA mice compared to WT controls ( Figure 7B). Sp7 and Alpl were decreased in Mx1;TβRI CA mice. However, expression of Ibsp, Col1a1, and Bglap was not altered. Interestingly, Hedgehogrelated genes, Gli1, Ptch1, Ptch2, and Wnt signaling-related gene, Wnt3a, were decreased in Mx1;TβRI CA mice compared to WT controls ( Figure 7B). Tnfsf11/Tnfrsf11b gene expression was significantly increased in Mx1;TβRI CA osteoclasts compared to WT controls ( Figure 7B). Moreover, Ctsk, Sufu, Acp5, and Tnf were increased in Mx1;TβRI CA osteoclasts as well ( Figure 7B). These results demonstrated that Mx1;TβRI CA mice exhibited bone loss associated with reduced osteoblast-and increased osteoclast-related gene expression.  To explore the role of constitutive activation of TβRI on the TGF-β signaling pathway, Western blot analysis for phosphorylated SMAD 2 and 3 (pSMAD 2 and 3) was performed alongside examining the expression of osteoblast and osteoclast-specific proteins. We

Mx1;TβRI CA Mice Have Increased TGF-β Signaling Leading to Decreases in RUNX2 and PTCH1 and Increases in CTSK Protein Levels
To explore the role of constitutive activation of TβRI on the TGF-β signaling pathway, Western blot analysis for phosphorylated SMAD 2 and 3 (pSMAD 2 and 3) was performed alongside examining the expression of osteoblast and osteoclast-specific proteins. We found that the ratio of pSMAD2/total SMAD 2 and pSMAD3/total SMAD 3 was significantly increased in osteoblasts from Mx1;TβRI CA mice compared to WT controls ( Figure 8A). Constitutive activation of TβRI increased the ratio of pSMAD3/total SMAD3 but not pSMAD2/total SMAD2 in osteoclasts ( Figure 8B). In addition, a key transcription factor for osteoblast formation, RUNX2 protein levels, were significantly reduced in Mx1;TβRI CA mice compared to WT control ( Figure 8A). Consistently, levels of PTCH1, transmembrane protein receptor of Hedgehog signaling pathway, were decreased ( Figure 8A). Furthermore, levels of CTSK, an osteoclast-specific protein, were increased in Mx1;TβRI CA mice ( Figure 8B). These results confirmed that constitutive activation of TβRI cause bone loss by affecting osteoblast and osteoclast protein levels. ( Figure 8B). These results confirmed that constitutive activation of TβRI cause bone loss by affecting osteoblast and osteoclast protein levels.

Mx1;TβRI CA Mice Have Decreased Bone Hardness in Tibia
Mx1;TβRI CA mice display bone loss. To determine whether, in addition to a reduction in bone volume, the mechanical properties of the cortical bone are compromised, we performed microindentation hardness testing of the tibial midshaft. The tibial cortical bone was tested with five indents ( Figure 9A) and the hardness was determined. Mx1;TβRI CA mice displayed significantly reduced cortical bone hardness compared to WT mice ( Figure  9B).

Mx1;TβRI CA Mice Have Decreased Bone Hardness in Tibia
Mx1;TβRI CA mice display bone loss. To determine whether, in addition to a reduction in bone volume, the mechanical properties of the cortical bone are compromised, we performed microindentation hardness testing of the tibial midshaft. The tibial cortical bone was tested with five indents ( Figure 9A) and the hardness was determined. Mx1;TβRI CA mice displayed significantly reduced cortical bone hardness compared to WT mice ( Figure 9B).

Discussion
TGF-β, a pleiotropic growth factor, binds to TβRII and then activates TβRI leading to induction of downstream signaling pathways and ultimately regulation of bone metabolism [1]. It was previously demonstrated that TβRI and TβRII are expressed on osteoblasts and osteoclasts [12]. Although the effects of TβRII on bone and cartilage homeostasis are well documented, the function of TβRI in bone function was poorly defined. Here, we Results are mean ± SEM. * p < 0.05 compared to WT. HV; Vickers hardness.

Discussion
TGF-β, a pleiotropic growth factor, binds to TβRII and then activates TβRI leading to induction of downstream signaling pathways and ultimately regulation of bone metabolism [1]. It was previously demonstrated that TβRI and TβRII are expressed on osteoblasts and osteoclasts [12]. Although the effects of TβRII on bone and cartilage homeostasis are well documented, the function of TβRI in bone function was poorly defined. Here, we generated Mx1;TβRI CA mice in which TβRI was conditionally activated in hematopoietic stem-cells-derived osteoclasts and osteoblasts in order to determine the physiological role of TGF-β in skeletogenesis. Mx1-Cre is silent in normal mice but can be induced in interferon-responsive cells, including monocytes and macrophages, following administration of poly (I:C) [13,14]. We confirmed the genotype of study mice by PCR, demonstrating that mice showed the Mx1-Cre band after poly (I:C) injection to induce TβRI expression. Hematopoietic stem cells are capable of differentiating into macrophages and osteoclasts [15]. In this study, we found that gene expression and protein levels of osteoclast markers were increased in Mx1;TβRI CA mice. These increased osteoclast markers indicated that the overexpression of TβRI in hematopoietic stem-cells-derived osteoclasts and osteoblasts caused an increase in the differentiation of hematopoietic stem cells to osteoclasts. Likewise, previous evidence found that TGF-β stimulates the formation of osteoclast from hematopoietic stem cells [16].
Our results showed that Mx1;TβRI CA mice were osteopenic through increased osteoclastmediated bone resorption and decreased osteoblastic bone formation. Furthermore, osteoblast formation was decreased by suppressing Hedgehog signaling pathways. As expected, we found that calcium levels were lower in Mx1;TβRI CA mice than WT controls. Abnormalities of calcium and phosphate metabolism can in turn impact secretion of PTH. Interestingly, the present study observed that serum PTH levels were increased in Mx1;TβRI CA mice. High levels of PTH can induce osteoclastic bone resorption [17]. The PTH receptor is expressed on osteoblasts and activates the PKA pathway, leading to increased secretion of Tnfsf11 [18]. Administration of PTH in male rat increased the levels of TGF-β1 [19]. PTH enhanced TGF-β1 expression and secretion through the PKC pathway and increased TGF-β2 expression and secretion through the PKA pathway [20]. These findings may indicate that the increased PTH levels in Mx1;TβRI CA mice are due to low serum calcium levels, thus increasing bone resorption.
µCT analysis indicated that cancellous bone volume, trabecular thickness, cortical thickness, and BMD in both femurs and mandibles were significantly decreased in Mx1;TβRI CA mice, indicating the presence of osteopenia. Several studies have reported that increasing TGF-β signaling results in bone defects. TGF-β stimulated signaling via formation of heteromeric complexes of type I and type II serine/threonine kinase receptors and initiated intracellular signaling through both SMAD-dependent and SMAD-independent signaling pathways. Overexpression of TβRII-B, a spliced form of TβRII, in chondrocytes was associated with increases in SMAD2 phosphorylation, SMAD3 signaling, and type II collagen production [21]. Overexpression of Tgfb1 caused abnormal bone remodeling and mandibular condylar cartilage degradation in osteoblast-specific Tgfb1 mutants [7]. Osteoblast-specific overexpression of Tgfb2 using Ocn promoter caused a mineralization defect and severe hypoplasia of clavicles similar to cleidocranial dysplasia [22]. This present study found that the constitutive activation of TβRI in mice resulted in osteopenia.
Bone homeostasis is maintained by the process of bone formation by osteoblast and bone resorption by osteoclast. Histomorphometric analysis indicated that cancellous bone volume was decreased in Mx;TβRI CA tibiae and mandibles compared to WT controls. Osteoblast numbers in tibial and mandibular bone in Mx;TβRI CA mice were reduced. Microindentation testing revealed that Mx1;TβRI CA mice had decreased bone hardness consistent with osteopenia from µCT and histomorphometric analysis. The TβRI kinase inhibitor, SD-208, increased bone mass by increasing osteoblast differentiation and suppressing osteoclast differentiation [23]. Lian et al. demonstrated that TGF-β inhibited osteoblast differentiation by suppressing ATF4-dependent Ocn transcription [24]. TGF-β stimulates the early stages of osteoblast differentiation but inhibited later stages of osteoblast maturation and mineralization [3,25]. Hematopoietic stem cells serve as the source of osteoclast precursors. Mandibular and long bones of Mx1;TβRI CA mice had increased osteoclast numbers and differentiation status compared to those of WT controls. A previous study reported that TGF-β1 stimulated osteoclast maturation in fetal mouse long bones [26]. TGF-β1 exerted dual effects on hematopoietic stem cell proliferation depending on concentration [27]. High concentrations of TGF-β1 induced phosphorylation of SMAD3 and activated c-Jun and ATF-2. In contrast, low TGF-β1 concentrations activated STAT signaling through an SMAD3-independent pathway. Moreover, TGF-β induced the activation of an SMAD2/3-associated TRAF6-TAB1-TAK1 complex leading to stimulation of osteoclastogenesis in response to Tnfsf11 stimulation [28]. TβRI deficiency using Dermo1-cre has been shown to have shortened long bones, decreased cancellous bone and bone collar mineralization, produced abnormalities in the perichondrium, and reduced osteoblast proliferation and differentiation [29]. Our in vitro assay indicated that constitutive activation of TβRI decreased osteoblast and increased osteoclast differentiation. Mx1;TβRI CA mice had decreased ALP activity and mineralization in long bones and mandibles compare to WT controls. In addition, the numbers of TRAP positive osteoclasts in both long bones and mandibles of Mx1;TβRI CA mice were increased. Therefore, the present study is the first to report that gain of function of TβRI decreased osteoblast numbers and differentiation together with increasing osteoclast numbers and differentiation, ultimately leading to bone loss.
Constitutive TβRI activation promoted osteoclast differentiation and formation by controlling genes involved in tumor necrosis factor (TNF) family signaling including Tn-fsf11 and Tnfrsf11b. Upregulation of the ratio of Tnfsf11/Tnfrsf11b expression in femurs and mandibles and apparent increases in osteoclasts derived from long bones were observed. Tnfsf11 is synthesized and secreted from osteoblast-lineage cells and binds with RANK on osteoclasts, leading to increases in osteoclast formation, function, and survival. In addition, osteoblasts also secrete Tnfrsf11b, which acts as a soluble Tnfsf11 decoy receptor for preventing Tnfsf11 binding to RANK, thereby inhibiting osteoclast development [30]. Our result correlated with previous findings that increased TGF-β1 stimulation increased Tnfsf11 gene expression in primary human osteoblasts cell culture via SMAD2/3 stimulation and that the TβRI inhibitor, SB431542, blocked these effects [31]. In addition, low concentrations of TGF-β increased the Tnfsf11/Tnfrsf1b ratio and Csf1 expression, leading to induce osteoclastogenesis [5]. A previous study revealed that after Tnfsf11 binds with its receptor, RANK, on osteoclasts, it leads to the activation of TRAF6, ERK, JNK, and p38, ultimately resulting in osteoclast proliferation and differentiation [32]. Csf1 expression was increased in femoral bone of Mx1;TβRI CA compared to WT controls. Moreover, Csf1 bound to Csf1r and stimulated PI3K/Akt and ERK, resulting in osteoclast proliferation and survival [32]. In addition, our qPCR analysis in femoral bone indicated that the expression of specific osteoclastogenesis-associated genes, including Ctsk and Sufu, was significantly increased in Mx1;TβRI CA mice. In the process of bone resorption, osteoclasts utilize cathepsin K, a lysosomal cysteine protease, to degrade collagen and other matrix proteins [33]. CTSK protein was also increased in Mx1;TβRI CA osteoclasts. The expression of Ctsk is regulated by Tnfsf11 signaling and Tnfsf11 activates the transcriptional factor, Nfatc1, to mediate stimulation of Ctsk gene expression in osteoclasts [34]. Sufu, a cytoplasmic protein, is a negative regulator of Hedgehog signaling pathway that binds to GLI family protein and suppress the differentiation of early skeletal progenitors. Deletion of Gli2 and Sufu in the cranial neural crest enhanced calvarial bone formation [35]. Osteoporotic mice overexpressed Sufu [36]. Sufu deletion suppressed Tnfsf11-induced TRAP positive osteoclast differentiation, osteoclast gene expression, and osteoclast activity in vitro [37]. In addition, Acp5 and Tnf significantly increased in Mx1;TβRI CA osteoclasts but not in femurs and mandibles. Femurs and mandibles are tissues that have multiple factors and signaling networks. Thus, those factors may have been implicated in the levels of these genes. In addition, there are many cell types in femurs and mandibles. Therefore, our findings indicated that Mx1;TβRI CA mice displayed an increase in the expression of several pathways regulating osteoclast differentiation, leading to increases in osteoclast formation and subsequently increasing bone resorption.
Wnt3a expression was significantly reduced in Mx1;TβRI CA mice. Wnt3a activates the β-catenin phosphorylation via the canonical Wnt signaling pathway and then stimulates the downstream target of osteoblast-related genes, leading to osteoblast differentiation and bone formation [38]. Runx2 is a DNA binding transcription factor that plays a key role in bone formation [39]. Interestingly, we found a significant reduction in Runx2 gene and protein levels in Mx1;TβRI CA osteoblasts compared to WT controls. As expected, we demonstrated that levels of osteoblast-related genes, Sp7 and Alpl, in Mx1;TβRI CA mandibles were significantly lower than those in WT controls. TGF-β activates ERK1/2 and JNK, which in turn leads to negatively regulate SMAD3-induced ALP activity and mineralization in MC3T3-E1 cells [40]. The Hedgehog signaling pathway has an essential role for the differentiation and development of osteoblasts. The multitransmembrane protein Ptch1 and Ptch2 are the receptors for Hedgehog while also being Hedgehog target genes. Ptch1 and Ptch2 expression was dramatically decreased in Mx1;TβRI CA femurs and mandibles whereas Gli1 and Shh expression was reduced in Mx1;TβRI CA mandibles. PTCH1 protein levels also decreased in Mx1;TβRI CA osteoblasts by western blot analysis. Shh binds with Ptch receptors, leading to release and conformational change of Smo and activates transcriptional factors Gli1, 2 and 3 [41]. Gli translocates into the nucleus to induce transcription of Hedgehog target genes in osteoblasts. In addition, a reduction in bone mass caused by decreased osteoblast differentiation, and increased osteoclastogenesis were found in Gli1-haploinsufficient mice [42]. These findings suggest a crosstalk between TGF-β and Hedgehog signaling in TβRI activation as one of the mechanistic insights resulting from study of constitutive TβRI activation-induced bone loss. Furthermore, our study demonstrated that Mx1;TβRI CA mice induced SMAD 3 phosphorylation in osteoclasts, demonstrating the TβRI CA transgene was active. A previous study found that excessive TGF-β signaling increased SMAD 2 phosphorylation, leading to bone loss; conversely, the same study reported that inhibiting TGF-β signaling using the 1D11 neutralizing antibody improved the bone phenotype [43]. After Tnfsf11 binds with RANK, this leads to the activation of Smad 2/3 to form the complex with TRAF6-TAB1-TAK1 that causes increased osteoclast differentiation [44].
In addition, expression of Tgfb1 and TβRI was increased in osteoclast cells, confirming overexpression of TβRI. However, previous studies found that mesenchymal stromal cells (MSCs) identified as a group of heterogeneous cells, comprising multipotent stem cells [45]. Bone-marrow MSCs are targeted by Mx1-cre upon induction [46,47]. In addition, the interferon-inducible Mx1 promoter drives the activation of Cre after administration of poly (I:C). Therefore, our results from primary osteoblasts found the increased TβRI gene expression and the ratio of pSMAD2/total SMAD2 and pSMAD3/total SMAD3 in osteoblasts. Previous studies reported that TGF-β stimulates the early stages of osteoblast differentiation but inhibits later stages of osteoblast differentiation [3,25]. Interestingly, this may explain the reduced osteoblast numbers seen in Mx1;TβRI CA mice in our study. Taken together, our results indicated that Mx1;TβRI CA mice displayed decreased osteoblast formation associated with reductions in several osteoblast-related genes, including Wnt3a, Runx2, Sp7, Alpl, and Hedgehog target genes ( Figure 10).
In summary, constitutive TβRI activation induces osteopenia by both increasing osteoclast differentiation and decreasing osteoblast differentiation, possibly through the suppression of the Hedgehog signaling pathway. Serum calcium were decreased whereas serum PTH was increased. Nowadays, there are many pharmacologic treatments for osteoporosis such as bisphosphonates, estrogen-related therapies, parathyroid hormone analogs, and RANK-ligand inhibitory antibodies [48]. However, our study revealed that constitutive TβRI activation is able to produce an osteopenic phenotype. Inhibitors of TGF-β and their receptors are currently under clinical development in phase I/II trials, especially in cancer treatment [49]. Thus, this study indicates that the bone mass effects of these drugs should be monitored for possible increases. Moreover, TβRI inhibitors may represent an adjuvant therapy for the treatment of osteoporosis or other disorders of low bone mass.
cells [45]. Bone-marrow MSCs are targeted by Mx1-cre upon induction [46,47]. In addition, the interferon-inducible Mx1 promoter drives the activation of Cre after administration of poly (I:C). Therefore, our results from primary osteoblasts found the increased TβRI gene expression and the ratio of pSMAD2/total SMAD2 and pSMAD3/total SMAD3 in osteoblasts. Previous studies reported that TGF-β stimulates the early stages of osteoblast differentiation but inhibits later stages of osteoblast differentiation [3,25]. Interestingly, this may explain the reduced osteoblast numbers seen in Mx1;TβRI CA mice in our study. Taken together, our results indicated that Mx1;TβRI CA mice displayed decreased osteoblast formation associated with reductions in several osteoblast-related genes, including Wnt3a, Runx2, Sp7, Alpl, and Hedgehog target genes ( Figure 10). In summary, constitutive TβRI activation induces osteopenia by both increasing osteoclast differentiation and decreasing osteoblast differentiation, possibly through the suppression of the Hedgehog signaling pathway. Serum calcium were decreased whereas serum PTH was increased. Nowadays, there are many pharmacologic treatments for osteoporosis such as bisphosphonates, estrogen-related therapies, parathyroid hormone analogs, and RANK-ligand inhibitory antibodies [48]. However, our study revealed that constitutive TβRI activation is able to produce an osteopenic phenotype. Inhibitors of TGF-β and their receptors are currently under clinical development in phase I/II trials, especially in cancer treatment [49]. Thus, this study indicates that the bone mass effects of these drugs should be monitored for possible increases. Moreover, TβRI inhibitors may represent an adjuvant therapy for the treatment of osteoporosis or other disorders of low bone mass.

Mice
Mice from Dr. Laurent Bartholin were transferred from the Department of Anatomy, Faculty of Science, Mahidol University, Bangkok, Thailand. TβRI CA mice were prepared according to the methods described by Vincent et al. [50]. Briefly, constitutively active TβRI was knocked into the X chromosome-linked hypoxanthine phosphoribosyl transferase (HPRT) locus for generating TβRI CA mice. Therefore, female mice were used in this

Mice
Mice from Dr. Laurent Bartholin were transferred from the Department of Anatomy, Faculty of Science, Mahidol University, Bangkok, Thailand. TβRI CA mice were prepared according to the methods described by Vincent et al. [50]. Briefly, constitutively active TβRI was knocked into the X chromosome-linked hypoxanthine phosphoribosyl transferase (HPRT) locus for generating TβRI CA mice. Therefore, female mice were used in this study. TβRI CA mice were crossed with mice expressing Cre recombinase under the control of the interferon-inducible Mx1 promoter to generate Mx1;TβRI CA mice to target hematopoietic stem cells. The genotype of 3-week-old female mice was determined by PCR analysis of their tail DNA. Genotyping PCR was performed using 3 primers, pRCA/pPA, pHPRTf/pHPRT-r, and Mx1-Cre/Internal control. Mx1;TβRI CA mice were injected intraperitoneally with 1 mg/mL of poly (I:C) in PBS. Two genotypes; TβRI CA lox/lox as a WT control and Mx1;TβRI CA lox/lox (Mx1;TβRI CA ) as mice that overexpressed TβRI were examined. Mice were housed at the Faculty of Medicine, Chulalongkorn University under controlled temperature (25 ± 2 • C) and with a 12 h light/dark cycle and free access to standard rodent chow (C.P. Mice Feed, Perfect Companion Group Co., Ltd., Bangkok, Thailand) and water. The animal protocol was approved by the Institutional Animal Care and Use Committee (IACUC) at the Faculty of Medicine, Chulalongkorn University and conducted according to the guidelines of the Animal Research: Reporting In Vivo Experiments (ARRIVE, Singapore). Nine-week-old female Mx1-cre;TβRI CA mice and their WT controls were used. At the end of the experiment, mice were anesthetized with isoflurane. Blood samples were collected and serum was kept at −80 • C for determination of serum chemistries. The left mandibles, femurs, and tibiae were removed and fixed in 70% alcohol for µCT analysis, histomorphometry, and microindentation analysis. The right mandibles and femurs were frozen in liquid nitrogen and kept at −80 • C for RNA isolation and qPCR analysis.

qPCR Analysis
Femur metaphysis and whole mandibles were harvested and snap-frozen in liquid nitrogen. The bones were placed in a pre-cooled mortar and grounded into fine powder using a pestle with liquid nitrogen. RNA was extracted using TRIzol (Invitrogen, Waltham, MA, USA), followed by RNA clean up using an RNeasy Mini kit (Qiagen, Germantown, MD, USA). The RNA yields were then determined using a NanoDrop 1000 (Thermo Fisher Scientific, Waltham, MA, USA). First-strand cDNA was synthesized using 1 µg of total RNA with SuperScript VILO cDNA synthesis kit (Invitrogen, Carlsbad, CA, USA) for reverse transcription. The qPCR was performed in Luna Universal qPCR master mix (New England Biolabs, Ipswich, MA, USA) using CFX96™ Optics Module (Bio-Rad, Hercules, CA, USA). The qPCR was conducted at 57 • C for 39 cycles. Gene expression signals were normalized to gapdh expression. All primer sequences are listed in Supplementary Table S1.

Osteoclast Culture
Briefly, bone marrow was flushed from mandibles and long bones using a needle (25 G) and with α-MEM medium. Bone marrow cells were passed through a 40 µm filter. Cells were then cultured in α-MEM medium containing 10% FBS, 100 units/mL penicillin, and 100 ug/mL streptomycin overnight. Nonadherent cells were placed on coverslips in the same medium containing 20 ng/mL M-CSF (R&D Systems, Inc., Minneapolis, MN, USA) for 2 days to generate BMMs. After that, BMMs were cultured in α-MEM medium containing 20 ng/mL M-CSF, and 3.3 ng/mL RANKL (R&D Systems, Inc., Minneapolis, MN, USA) for 6 days. TRAP-positive osteoclast number per total area (N.Oc/Ar) was quantified using OsteoMeasure software version 4.2.0.1 (Decatur, GA, USA).

Osteoblast Culture
The isolation of primary osteoblasts was prepared according to the methods described by Chevalier et al. [52]. After flushing the bone marrow out from mandibles and long bones, bones were minced with sterile scissors into small pieces and placed in α-MEM medium containing 1 mg/mL collagenase type II (Worthington Biochemical Corporation, Lakewood, NJ, USA) for 2 h at 37 • C with shaking. Then, cells were centrifuged at 5000 rpm for 8 min, and the supernatant was removed. Bone fragments were transferred to 75 cm 2 flasks containing α-MEM, 20% FBS, 100 units/mL penicillin, and 100 µg/mL streptomycin until cells were confluent. Cells were cultured in α-MEM medium containing 20% FBS, 5 mM β-glycerophosphate, 50 µg/mL ascorbic acid, and 10 µM dexamethasone for 7 and 25 days. Osteoblasts were fixed with 3.7% formaldehyde and stained with Fast Blue RR (Sigma, St. Louis, MO, USA) for ALP on day 7. For mineralized bone nodules, osteoblasts were stained with 2% alizarin red (Sigma, St. Louis, MO, USA) on day 25. Mineralization was determined by destained alizarin red using 10% cetylpyridinium chloride in 10 mM sodium phosphate and concentration of alizarin red (mM) was quantified.

Microindentation Analysis
Tibiae were embedded without demineralization in methyl methacrylate. Tibiae were polished with 600, 800, 1200 grit abrasive sandpaper followed by 0.5 µm de-agglomerated alumina powder. The hardness of tibial cortical bone was determined using microhardness tester, FM-810 (Future-Tech Corp., Kawasaki, Japan). A Vickers diamond indenter with an applied load of 25 gf for 10 s was used. The cortical bone was tested at midshaft for five indents each 50 µm apart. The Vickers hardness equation was 1.854 L/D 2 . The average length of two diagonals (D) was measured in millimeter and the load (L) was input kilograms.

Measurement of Bone Length
Bone length was determined using vernier calipers, measuring from the proximal to distal ends of femurs and the most anterior point of the dentary to the most posterior point of the articular condyle for mandibles [54,55].

Serum Chemistry
Levels of serum phosphorus and calcium were measured according to the manufacturer's instructions (Standbio Laboratory, Boerne, TX, USA). Serum PTH levels were determined using an ELISA kit (Quidel, San Diego, CA, USA).

Statistically Analysis
All statistical analyses were performed using SPSS 24 (IBM, Armonk, NY, USA). Independent t-test was used to compare the differences between two groups. Data are expressed as mean ± SEM. p < 0.05 were considered statistically significant.