EGFR controls bone development by negatively regulating mTOR-signaling during osteoblast differentiation

EGFR controls bone development by negatively regulating mTOR-signaling during osteoblast... −/− Mice deficient in epidermal growth factor receptor (Egfr mice) are growth retarded and exhibit severe bone defects that are poorly understood. Here we show that EGFR-deficient mice are osteopenic and display impaired endochondral and intramembranous ossification resulting in irregular mineralization of their bones. This phenotype is recapitulated in mice lacking EGFR exclusively in osteoblasts, but not in mice lacking EGFR in osteoclasts indicating that osteoblasts are responsible for the bone phenotype. Experiments are presented demonstrating that signaling via EGFR stimulates osteoblast proliferation and inhibits their differentiation by suppression of the IGF-1R/mTOR-pathway via ERK1/2-dependent up- −/− regulation of IGFBP-3. Osteoblasts from Egfr mice show increased levels of IGF-1R and hyperactivation of mTOR- −/− pathway proteins, including enhanced phosphorylation of 4E-BP1 and S6. The same changes are also seen in Egfr bones. Importantly, pharmacological inhibition of mTOR with rapamycin decreases osteoblasts differentiation as well as rescues the −/− low bone mass phenotype of Egfr fetuses. Our results demonstrate that suppression of the IGF-1R/mTOR-pathway by EGFR/ERK/IGFBP-3 signaling is necessary for balanced osteoblast maturation providing a mechanism for the skeletal phenotype observed in EGFR-deficient mice. Introduction Markus Linder and Manfred Hecking contributed equally to this work. Skeletal development requires complex and coordinated Edited by E Gottlieb interplay between mesenchymal cells—chondrocytes and osteoblasts—at various stages of differentiation. The suc- Electronic supplementary material The online version of this article (https://doi.org/10.1038/s41418-017-0054-7) contains supplementary cession of events whereby osteoblast formation follows material, which is available to authorized users. chondrocyte differentiation resulting in bone formation in long bones of vertebras is termed “endochondral ossifica- * Maria Sibilia Sibilia-Office@meduniwien.ac.at tion” [1]. Differentiation of osteoblasts is induced by up- regulation of specific transcription factors accompanied by Department of Internal Medicine I, Institute of Cancer Research, the expression of factors that facilitate mineralized extra- Comprehensive Cancer Center, Medical University of Vienna, cellular matrix formation [2]. Vienna, Austria Genetic ablation of the epidermal growth factor receptor Department of Internal Medicine 3, Friedrich-Alexander- (EGFR) in mice revealed its intricate role during embryonic University Erlangen-Nürnberg (FAU), Universitätsklinikum Erlangen, Erlangen, Germany and postnatal development [3, 4]. Mice lacking the EGFR −/− (Egfr ) have major organ defects and die either in utero or Department of Internal Medicine III, Medical University of Vienna, Vienna, Austria shortly after birth, depending on the genetic background [3- 5]. Among other developmental abnormalities EGFR- Spanish National Cancer Research Center (CNIO), Madrid, Spain deficient mice are severely growth retarded and exhibit Institute of Comparative Molecular Endocrinology, University of skeletal defects [6]. We have previously reported that long Ulm, Ulm, Germany bones of EGFR-deficient mice display a greatly increased Present address: Department of Internal Medicine III, Medical zone of hypertrophic chondrocytes, suggesting that EGFR University of Vienna, Waehringer Guertel 18-20, Vienna A-1090, Austria negatively regulates condrocyte maturation [6]. Similar 1234567890();,: EGFR regulates osteoblast differentiation 1095 wt -/- wt -/- −/− Egfr Egfr Egfr Egfr Fig. 1 Egfr mice are osteopenic. a Alcian blue and ab Alizarin red whole body mount showing skeletal mineralization wt −/− of Egfr and Egfr mice on postnatal day 7 (P7). b Skeletal preparations of WT and KO mice: femur and spine. c µCT wt image of 7-day-old Egfr (left) -/- and Egfr (right) mice d Von wt Kossa staining of Egfr and −/− Egfr calvaria at P7; scales: wt -/- Egfr Egfr 100 μm for lower and 20 µm for higher magnification. e Von Kossa staining showing wt calcification of Egfr and Egfr −/− tibiae on P7; scales: 500 μm for lower and and 100 μm for higher magnification. f Histomorphometric analysis of wt −/− Egfr and Egfr tibiae from wt -/- 1mm 1mm Egfr Egfr P1 to P14: Quantification of bone volume/tissue volume (BV/TV), trabecular number wt -/- (Tb.N), trabecular separation Egfr Egfr (Tb.Sp), trabecular thickness (Tb.Th) and osteoblast number per bone perimeter (N.Ob/B. Pm). P1: n = 6. P7: n = 5. P14: n = 6 WT, 3 KO mice wt -/- Egfr Egfr 50 20 30 15 500 wt Egfr -/- Egfr ** * 400 40 ** 300 30 *** ** ** ** 20 10 5 100 10 0 0 0 0 0 P1 P7 P14 P1 P7 P14 P1 P7 P14 P1 P7 P14 P1 P7 P14 observations were made by Wang et al. [7], who in addition shortened long-bones with increased cartilage mineraliza- found delayed primary ossification with irregular distribu- tion [9]. −/− tion of osteoblasts in Egfr embryos [7]. EGF treatment of WT calvariae increased the prolifera- Moreover, EGFR knock-in mice where the murine tion of osteoprogenitor cells and maintained them in an −/− EGFR is replaced by the human counterpart display low undifferentiated state [10]. Accordingly, Egfr osteoblasts EGFR activity in the bone and show impaired endochondral show reduced proliferation but elevated differentiation ossification and an increased hypertrophic chondrocyte indicating that EGFR is essential during osteoblast zone [6]. Similarly, mice with reduced EGFR activity by maturation [8]. However, the underlying molecular combined expression of a dominant-negative Egfr Wa5 mechanisms has so far not been investigated. It is also allele and deletion of an Egfrfloxed allele using Col1a1-Cre unclear whether the bone defects observed in adult mice wa5/f mice (Col1a1-Cre Egfr ), display bone abnormalities result from developmental defects or arise later during bone starting around 3 months of age [8]. Mice lacking the remodeling. The mouse models employed so far have not membrane-anchored metalloproteinase ADAM17, respon- allowed to investigate this effect, since incomplete EGFR sible for cleavage of several membrane-bound cytokines deletion was observed using Col-Cre mice and osteoblast- and growth factors including EGFR ligands also develop independent effects of the ubiquitously expressed, expanded zones of hypertrophic chondrocytes, and dominant-negative Wa5 on other ErbB family members chondrocyte-specific deletion of ADAM17 results in cannot be excluded [8]. BV/TV (%) -1 Tb.N (mm ) Tb.Sp (μm) Tb.Th (μm) -1 N.Ob/B.Pm (mm ) 1096 M. Linder et al. wt -/- Egfr Egfr Fig. 2 EGFR deletion leads to reduced proliferation and pERK1/2 in bone-lining cells. Representative images and quantifications of IHC stainings on femur sections from 7-day- wt −/− wt -/- old Egfr and Egfr mice against (a) p-Histone H3 (n = b 5), (b) PCNA (n = 5), and (c) pERK1/2 (n = 3). Scales: 200 µm (lower magnification) and 20 µm (higher magnification) wt -/- ** wt -/- Here we investigated the bone phenotype occurring in columns of calcified extracellular matrix (ECM) with a −/− the first weeks of age in Egfr mice and in adult mice in clearly delineated border were observed in WTs, these −/− which Egfr is conditionally deleted in the osteoblast lineage structures were lacking in Egfr calvariae (Fig. 1d). f/f ΔOb using Egfr Runx2-Cre (Egfr ) mice. We found that Taken together our results show that Egfr deletion leads EGFR signaling in osteoblasts negatively regulates IGF-1R/ to impaired bone development in newborn mice with mTOR pathway via ERK1/2 dependent up-regulaion of defects in both endochondral and intramembranous IGFBP-3 to coordinate differentiation during embryonic ossification. −/− and postnatal bone formation. Egfr long-bones displayed a low-bone-mass pheno- type with less calcified bone and fewer bony trabeculae on P7 (arrowheads; Fig. 1e) and a thickened growth plate −/− RESULTS (arrows; Fig. 1e). Egfr tibiae exhibited thicker zones of ECM located at the cortical sides reaching into the center of −/− Egfr mice show impaired endochondral and the bone (Fig. S1b) indicating that the mineralization pro- −/− intramembranous ossification cess in Egfr bones was impaired due to misbalanced deposition of ECM by osteoblasts. −/− We first performed an analysis of the skeleton of Egfr Histomorphometric analyses confirmed that the ratio of mice that survived until postnatal day 7 (P7). Bones of Egfr bone volume over tissue volume (BV/TV) was significantly −/− −/− mice were less mineralized and reduced in length lower in Egfr mice (Fig. 1f). The trabecular number (Tb. compared to WT littermates (Fig. 1a). Whole-mount body N) was decreased while trabecular separation (Tb.Sp) was staining revealed reduced centers of secondary ossification increased at P7 and P14, although trabecular thickness (Tb. −/− in long bones and irregular calcification of vertebral end- Th) was not significantly changed (Fig. 1f). While Egfr plates in EGFR-deficient mice (Fig. 1b). Additionally, Egfr mice were born with osteoblast numbers (N.Ob) compar- −/− mice showed reduced mineralization of costal cartilage able to WT levels, their amount was significantly decreased (Fig. S1a). on P14 (Fig. 1f). While most bones develop by endochondral ossifica- tion, the lateral clavicles and parts of the skull are formed EGFR is essential for osteoblast proliferation and by intramembranous ossification, where mesenchymal ERK1/2 activation cells directly differentiate into osteoblasts without chondrocyte involvement [1, 11]. To determine whether As osteoblasts are essential for bone mineralization we next −/− the bone phenotype of Egfr mice can occur inde- focused on the role of EGFR during osteoblastogenesis. We pendently of the cartilage defects, we examined skulls of found decreased proliferation of primary pre-osteoblasts −/− Egfr mice. µCT analysis revealed an impaired cranial lacking the EGFR [6] (Fig. S2a), without any significant suture closure on day 14 (Fig. 1c), indicating that EGFR differences in the number of apoptotic cells (Fig. S2b). −/− also plays an important role during intramembranous Additionally, Egfr osteo-progenitors showed reduced ossification. Furthermore, while straight, well-organized BrdU and Cyclin D1 levels, indicating that EGFR deletion p-ERK1/2 PCNA p-Histone H3 PCNA pos / BPm p-H3 pos / BPm pERK1/2 pos / BPm EGFR regulates osteoblast differentiation 1097 ΔOb Fig. 3 Egfr mice phenocopy a −/− the bone phenotype of Egfr mice. a H&E stainings showing distal femurs with increased zone of hypertrophic chondrocytes in 6-day-old ΔOb Egfr mice; scales: 200 µm (lower magnification) and 100 µm (higher magnification). b wt ΔOb Egfr Egfr µCT image of femurs from 3- wt ΔOb months old Egfr and Egfr b c littermate; scale: 1 mm. c P6 P21 P90 Quantification of femur length 6.0 10 * **** wt ΔOb of Egfr and Egfr mice with indicated age 5.5 5.0 8 12 4.5 wt wt ΔOb ΔOb wt ΔOb Egfr Egfr Egfr Egfr Egfr Egfr −/− leads to cell autonomous proliferation defects without mice was significantly increased, comparable to Egfr ΔOb affecting apoptosis (Figs. S2c, d). mice (Fig. 3a). Importantly, Egfr mice showed reduced To confirm that the observed defects are also occurring length of long bones which was significant by P21 and in vivo, we evaluated the proliferation of bone-lining cells became more severe with age (Figs. 3b, c). These results −/− wt in femoral sections of Egfr and Egfr mice. The number demonstrate that EGFR signaling in osteoblasts is essential of cells positive for the mitosis marker p-Histone H3 for proper bone development. (Fig. 2a) and the S-phase related marker PCNA (Fig. 2b) ΔOb −/− were significantly reduced in Egfr mice indicating that Adult Egfr mice develop a low-bone-mass EGFR is crucial for proliferation during bone development. phenotype The ERK pathway, a major EGFR downstream signaling pathway, plays a central role in cell proliferation [12]. The augmented zone of hypertrophic chondrocytes was Therefore, we analyzed the phosphorylation of ERK1/2 in accompanied by increased expression of the hypertrophic −/− ΔOb bone lining cells at P7. Egfr mice exhibited significantly chondrocyte marker Col10a1 in long bones of Egfr mice reduced numbers of p-ERK1/2 positive cells on their tra- on P21 (Fig. 4a). Significantly elevated Runt-related tran- becular bone (Fig. 2c), suggesting that the proliferation scription factor-2 (Runx2) mRNA levels together with defects during bone development are based on impaired reduced Colagen1a1 (Col1a1) mRNA and reduced Osteo- ERK1/2 activation. calcin (Ocn) mRNA and protein levels (Figs. 4a, S4a) indicate that EGFR deletion in osteoblasts leads to impaired Osteoblast-specific deletion of EGFR leads to bone mineralization due to premature differentiation of osteo- defects progenitors. Histomorphological analysis revealed a pro- gressive, low-bone-mass phenotype with decreased bone −/− ΔOb To address whether the bone phenotype in Egfr mice is volume and trabecular number in adult Egfr mice f/f due to cell-autonomous defects in osteoblasts, Egfr mice (Figs. 4b, c). Additional trabecular bone markers further were crossed to an osteoblast-specific Cre line (Runx2-Cre), showed reduced trabecular thickness and increased spacing f/f ΔOb ΔOb to generate Egfr Runx2-Cre (Egfr ) mice [13]. Com- in Egfr mice (Fig. S4b). Less osteoblasts on the trabe- plete deletion of EGFR was confirmed by Western Blot in cular bone and reduced osteocalcin serum levels (Fig. 4d) cultured osteoblasts and by IHC in long bones indicate that the low-bone-mass is based on osteoblast (Figs. S3a, b). As shown by qRT-PCR from RNA isolated defects. wt ΔOb from bone and cartilage of Egfr and Egfr femurs, To exclude that EGFR in osteoblasts indirectly affects deletion of Egfr was restricted to bone tissue, but not car- osteoclastogenesis, osteoclast-specific markers were ana- ΔOb tilage (Fig. S3c). Egfr mice developed normally without lyzed in long bones and serum. No significant differences in any significant differences in overall body length (Fig. S3d). osteoclast number could be detected neither in young nor in ΔOb ΔOb ΔOb On P6 the zone of hypertrophic chondrocytes of Egfr adult Egfr mice. Furthermore, Egfr mice did not femur lenght (mm) femur lenght (mm) femur lenght (mm) 1098 M. Linder et al. ΔOb *** Fig. 4 Egfr mice show severe wt ab Egfr bone defects. a qRT-PCR 2.0 ΔOb Egfr analysis: RNA isolated from whole femurs of 21-day-old wt ΔOb 1.5 * Egfr and Egfr mice; n = 7 WT, 9 ΔOb. b H&E stainings **** **** wt Egfr showing distal femurs of 3- 1.0 wt ΔOb months-old Egfr and Egfr mice; scales: 200 µm. c Histomorphometric analysis of 0.5 WT/ΔOb long-bones at P6, P21 P90 and P210: Quantification of 0.0 bone volume/tissue volume (BV/TV) and trabecular number ΔOb Egfr (Tb.N). P6: n = 4 WT, 3 ΔOb. P21: n = 5. P90: n = 8 WT, 6 wt wt Egfr c 25 Egfr d ΔOb. P210: n = 4 WT, 6 ΔOb 25 30 1000 ΔOb ΔOb Egfr Egfr *** mice. d Osteoblast number (N. 20 800 Ob/B.Pm) and surface on the p=0.053 15 600 trabecular bone (Ob.S/BS) at 10 400 P21 (n = 7 WT, 10 ΔOb) and P210 (n = 6 WT, 8 ΔOb) and 5 5 200 Osteocalcin as measured by 0 0 0 0 ELISA in Serum at P21 (n = 4 P6 P21 P90 P21 0 age=P21 wt 6 Egfr 15 20 250 WT, 6 ΔOb) and P210 (n = 5 ΔOb Egfr WT, 9 ΔOb) 4 *** 10 2 5 0 0 0 0 P6 P21 P90 P21 0 age=P210 show any differences of the serum biomarker for bone [14], we investigated whether EGFR regulates bone resorption C-terminal telopeptide (CTX-1) (Fig. S4c). development by interacting with the IGF-1R signaling Additionally, we assessed whether EGFR directly affects pathway. We detected elevated levels of total and phos- f/f osteoclast development by breeding Egfr mice to LysM- phorylated IGF1Rβ in differentiated osteoblasts isolated −/− −/− Cre mice that express Cre recombinase in the myeloid from Egfr mice (Fig. 5a). Furthermore, Egfr osteo- ΔOc ΔOc lineage (Egfr ). Osteoclasts isolated from Egfr mice blasts showed increased total and phosphorylated protein showed reduced EGFR protein levels (Fig. S4d), but did not levels of the IGF-1R adapter protein insulin receptor sub- display any bone defects nor differences in the number of strate 1 (IRS-1) and its downstream target mTOR (Fig. 5a). osteoclasts in trabecular bones or serum CTX-1 (Fig. S4e). Importantly, IGF-1R/IRS1/mTOR up-regulation was ligand −/− Bone-marrow derived pre-osteoclasts from Egfr mice independent as the levels of IGF-1 and IGF-2 were not did not show any significant differences in their ability to altered (Fig. S5a). form osteoclasts in vitro (Fig. S4f). Finally, OC number in To investigate the kinetics of mTOR activation we next trabecular bones and serum CTX-1 levels were not altered analyzed multiple time points during osteoblast differ- -/- in Egfr mice (Fig. S4g) indicating that lack of EGFR does entiation. IGF-1R/mTOR-pathway proteins were con- not affect osteoclastogenesis. sistently present at higher levels and were hyper- −/− phosphorylated during differentiation in Egfr osteo- −/− Enhanced differentiation of Egfr osteoblasts blasts indicating that IGF-1R/mTOR-signaling remained correlates with IGF-1R/mTOR activation elevated throughout the whole culture period (Fig. 5b). IHC staining on femur sections of WT and EGFR- Once confirmed that the defects are primarily in the osteo- deficient mice at P7 revealed that the mTOR-signaling wt blast lineage, we employed primary osteoblasts from Egfr pathway was also altered in vivo. In line with the in vitro −/− and Egfr mice to investigate the underlying molecular findings, significantly increased phosphorylation of mTOR −/− mechanism. As osteoblasts from Egfr mice display and its main downstream targets 4E-BP1 and S6 protein −/− enhanced differentiation [6] and the IGF-1R pathway was were observed in Egfr long-bones (Figs. 5c, d). Addi- ΔOb shown to play a central role during osteoblast differentiation tionally, Egfr mice also showed reduced p-S6 protein Egf r Osx Run x2 Col 1a1 Col 1a2 Ocn Opn On Col 2a1 Col 10a1 -1 Tb.N (mm ) BV/TV (%) mRNA fold change N.Ob/B.Pm N.Ob/B.Pm Ob.S/BS (%) Ob.S/BS (%) Ocn (ng/ml) Ocn (ng/ml) EGFR regulates osteoblast differentiation 1099 a b Fig. 5 Enhanced differentiation kDa wt -/- correlates with mTOR-signaling. 175 EGFR wt wt wt kDa wt -/- -/- -/- wt -/- a Western blot analysis of Egfr p-IGF-1Rβ −/− EGFR and Egfr osteoblasts under 175 IGF-1Rβ differentiation conditions; IGF-1Rβ isolated on differentiation day Tubulin p-AKT 14. b Western blot analyses of wt −/− AKT Egfr and Egfr osteoblasts 180 p-IRS-1 under differentiation conditions p-S6 180 IRS-1 on days 6, 9, 12, and 15. c S6 50 Tubulin 32 Immunohistochemical staining p-4E-BP1 of p-mTOR, p-4E-BP1 and p-S6 289 p-mTOR on trabecular bone sections from 4E-BP1 wt distal femurs of P7 Egfr and mTOR Tubulin −/− 50 Egfr littermates; scales: 200 50 Tubulin µm (lower magnification) and 20 µm (higher magnification). d Quantification of IHC staining, wt -/- Egfr Egfr shown as positive cells per bone perimeter (B.Pm); n ≥ 3 c d ** wt -/- ** wt -/- p=0.052 wt -/- levels in bone-lining cells indicating that this activation the EGF-induced hypo-differentiation phenotype resulting in depends on osteoblastic EGFR signaling (Figs. S5b, c). normalized bone nodule formation comparable to untreated controls (Fig. 6i). Taken together our results show that IGF- Interplay between EGFR- and IGF-1R-pathways in 1R signaling enhances, whereas EGFR signaling inhibits osteoblast differentiation osteoblast differentiation and that EGFR signaling dom- inates by negatively regulating IGF-1R via ERK1/2. To analyze the cross-talk between EGFR and IGF-1R- To dissect the underlying molecular mechanism we signaling during osteoblast differentiation, WT osteoblasts analyzed the activation of EGFR and IGF-1R downstream were cultured under differentiation-inducing conditions proteins in differentiated WT osteoblasts cultured for together with IGF-1, EGF and/or the ERK1/2 inhibitor 21 days in the presence of EGF / IGF-1 and U0126. EGF U0126. At day 21 bone nodule formation was assessed as a treatment prevented phosphorylation of IGF-1Rβ with functional read-out for differentiation. Mineralization was reduced activation of the mTOR/S6/4E-BP1 pathway, enhanced by IGF-1 treatment (Figs. 6a, b) and completely whereas IGF-1 induced the phosphorylation of IGF-1Rβ/ abolished by EGF (Fig. 6c). Addition of EGF was able to mTOR/S6/4E-BP1 (Fig. 6j). When osteoblasts were cul- suppress IGF-1 induced differentiation in a dose-dependent tured with both growth factors, activation was again manner with complete inhibition at 100 ng/ml (Figs. 6d–f). reduced suggesting that EGFR signaling is able to block IGF-1 induced differentiation was further increased when differentiation via IGF-1Rβ inhibition. Importantly, EGF- ERK1/2 signaling was blocked by U0126 (Figs. 6g, h). ERK induced downregulation of the IGF-1Rβ pathway was partly inhibition together with EGF and IGF-1 stimulation rescued restored when ERK1/2 was blocked, indicating that EGFR p-S6 p-4E-BP1 p-mTOR p-S6 pos. p-4E-BP1 pos. p-mTOR pos. cells / BPm cells / BPm cells / BPm 1100 M. Linder et al. a d g j - + - ++ EGF k EGF (10ng/ml) untreated IGF-1 (100ng/ml) U0126 (10μM) - - IGF-1 + ++ kDa - -- +- U0126 - - EGF + ++ - - + ++ IGF-1 p-IGF-1Rβ - -- +- U0126 kDa 45 IGFBP-3 IGF-1Rβ 116 Vinculin p-ERK b e h EGF (50ng/ml) U0126 44 IGF-1 (100ng/ml) ERK IGF-1 (100ng/ml) IGF-1 (100ng/ml) HSP90 kDa 0 01 .1 10 100 1000 Afatinib (nM) 289 p-mTOR 175 p-EGFR 289 mTOR 175 EGFR p-S6 U0126 289 p-mTOR c f i EGF (100ng/ml) EGF (100ng/ml) S6 EGF (100ng/ml) 32 IGF-1 (100ng/ml) IGF-1 (100ng/ml) 289 mTOR HSP90 IGFBP-3 p-4E-BP1 50 Tubulin 4E-BP1 Tubulin m n p Serum Tibia 1250 10 1000 8 750 6 *** 500 4 wt -/- 250 2 Egfr Egfr 0 0 wt -/- wt -/- Egfr Egfr Egfr Egfr wt -/- Egfr Egfr kDa 45 IGFBP-3 wt -/- Egfr Egfr 116 Vinculin Fig. 6 Differentiation in WT osteoblasts is mediated by specific cultured in αMEM + 10%FCS after 4 h treatment with indicated components of the IGF-1R-pathway and inhibited by EGF. a–i Ali- concentrations of Afatinib. m IGFBP-3 protein levels in serum (n = 7) wt −/− zarin red staining of WT osteoblasts after 21 days (D21) in culture and n whole tibia lysates (n = 4) of 7-day old Egfr and Egfr under differentiation conditions (+AA, βGP) with EGF, IGF-1, and/or littermates. o Western Blot analysis of whole tibia protein lysates wt −/− U0126 over the whole culture period. Pictures taken from a 6-well isolated from 7-day old Egfr and Egfr littermates. p Alizarin red −/− plate. j, k Western blot analysis of differentiated WT osteoblasts (D21) staining of differentiated WT and Egfr osteoblasts (D21) cultured cultured with EGF (100 ng), IGF-1 (100 ng) and/or U0126 (10 µM). l with vehicle (DMSO) or Rapamycin (10 nM). Stained with alizarin red Western Blot analysis of undifferentiated osteoblast precursor cells negatively regulates differentiation by down-regulating ERK1/2 and at the same time reduced the phosphorylation IGF-1Rβ/mTOR signaling via ERK1/2 (Fig. 6j). No dif- of IGF-1Rβ whereas IGF-1 stimulation did not affect ferences in insulin receptor β (IRβ) phosphorylation could ERK1/2 signaling (Fig. S6b). be detected indicating that EGF stimulation exclusively To investigate the mechanism how EGFR signaling downregulates IGF-1R without affecting IRβ activation suppresses IGF-1R/mTOR signaling we next analyzed (Fig. S6a). IGFBP-3 levels, as IGFBP-3 is known to modulate and To prove that reduced activation of IGF-1Rβ is a direct repress IGF-1R signaling [15, 16]. Moreover, it has been consequence of EGF stimulation, we cultured WT osteo- shown that EGFR directly regulates IGFBP-3 in primary blasts under differentiation conditions for 21 days, starved esophageal cells [17]. We found elevated IGFBP-3 levels in them for 24 h and stimulated for 10 min with EGF or IGF-1. osteoblasts cultured together with EGF or with EGF and As expected, EGF treatment induced a strong activation of IGF-1 whereas IGF-1 alone had no effect (Fig. 6k). IGFBP-3 (ng/ml) IGFBP-3 (pg/μg protein) Rapamycin DMSO EGFR regulates osteoblast differentiation 1101 wt Egfr *** -/- E15.5 E16.5 E18.5 Egfr Rapamycin - 5mg/kg bw, 2x/d Analyze embryonic bones b 2 +Rapamycin wt Egfr Ocn 2.0 wt -/- -/- Egfr Egfr Egfr 1.5 1.0 0.5 wt -/- Egfr Egfr 0.0 +Rapamycin −/− Fig. 7 mTOR inhibition partially rescues bone phenotype of Egfr volume relative to total volume (BV/TV) after rapamycin/vehicle mice. a Timeline showing experiment outline for in utero Rapamycin treatment; n = 7 WT, 6 KO mice for vehicle and 6 WT, 8 KO mice for treatment. Pregnant mice were subcutaneously injected with 5 mg rapamycin treatment; 5 mothers per treatment group. d mRNA Rapamycin per kg bodyweight or Injection-vehicle twice a day on expression levels of Osteocalcin (Ocn) in femurs of E18.5 mice E15.5 and E16.5 b H&E staining of trabecular bone sections showing measured by qRT-PCR; n = 9 WT, 5 KO mice for vehicle and 7 WT, wt −/− distal femurs of Egfr and Egfr embryos on E18.5 after Rapa- 5 KO mice for rapamycin treatment mycin/vehicle treatment; scale: 200 µm. c Quantification of bone Importantly, IGFBP-3 up-regulation was a direct con- formation was strongly reduced in the presence of rapa- sequence of ERK1/2 signaling, as additional treatment with mycin (Fig. 6p). Upon rapamycin treatment, phosphor- the ERK1/2 inhibitor U0126 normalized EGF-induced ylation of the mTOR downstream proteins 4E-BP1 and S6 −/− IGFBP-3 levels (Fig. 6k). In contrast, EGFR inhibition was down-regulated in Egfr cultures similarly to WT with Afatinib led to a dose-dependent decrease in IGFBP-3 osteoblasts (Fig. S6e) demonstrating that the increased −/− protein levels along with increased p-mTOR phosphoryla- differentiation in Egfr osteoblasts can be prevented by tion in osteoblast precursors (Fig. 6l). In addition, IGFBP-3 mTOR-inhibition. was also strongly reduced in the supernatant of osteoblast Taken together our data provide evidence that EGFR precursor cells after 48 h treatment with EGFR inhibitor as controls osteoblasts differentiation via ERK-dependent compared to DMSO treated controls (Fig. S6c). IGFBP-3 up-regulation, which ensures proper osteoblast In line with our in vitro results, we also found sig- maturation by controlling IGF-1R/mTOR signaling. −/− nificantly reduced IGFBP-3 in the serum of Egfr and ΔOb Egfr mice (Figs. 6m, S6d) indicating that EGFR sig- mTOR inhibition partially rescues bone phenotype −/− naling in osteoblasts is essential for IGFBP-3 production. of Egfr embryos IGFBP-3 levels were also reduced in whole tibia protein −/− lysates of Egfr mice, as revealed by both ELISA and We next analyzed whether mTOR inhibition during western blot analysis (Figs. 6n, o). These results demon- embryonic development, when mineralization starts, can strate that EGFR is required for IGFBP-3 production and normalize the bone defects in EGFR-deficient mice. Phar- suppression of IGF-1R/mTOR activation thus providing a macological inhibition of mTOR during gestation has pre- mechanistic link between EGFR and IGF-1R signaling and viously been reported not to cause any bone-specific side osteoblast differentiation. effects in mice [18]. We injected pregnant females from To further show that the hyper-differentiation pheno- EGFR heterozygous intercrosses with rapamycin or vehicle −/− type of Egfr osteoblasts is indeed a consequence of twice a day on E15.5 and on E16.5 and analyzed embryonic elevated mTOR activation we next inhibited mTOR in bones at E18.5 (Fig. 7a). Rapamycin treatment was not differentiating osteoblasts using rapamycin. Bone nodule teratogenic nor did it affect litter size or viability of pups Rapamycin Vehicle mRNA fold change BV/TV (%) 1102 M. Linder et al. −/− (Fig. S7a). Inhibition of mTOR signaling pathway was that Egfr osteoblasts showed elevated mineralization −/− ΔOb confirmed by p-S6 IHC staining on femurs of fetuses in vitro, both Egfr and Egfr mice are osteopenic. This obtained from rapamycin or vehicle-treated mothers apparent discrepancy might be due to the fact that osteo- (Fig. S7b). progenitor cells lacking the EGFR, which display pro- We could not observe any effect on hypertrophic chon- liferation defects, cannot form sufficient numbers of −/− drocyte zone in embryonic Egfr bones after rapamycin osteoblasts to guarantee proper maturation and ossification treatment (Fig. S7c), which is in line with our hypothesis in vivo. that the hypertrophic chondrocyte phenotype is not We identified the mTOR-pathway as a positive regulator responsible for the impaired bone development. However, of osteoblast differentiation that is suppressed by EGFR chemical inhibition of mTOR increased the zone of signaling. In the absence of EGFR, IGF-1R/mTOR signal- hypertrophic chondrocytes in WT animals (Fig. S7c) ing is up-regulated due to reduced IGFBP-3 signaling without affecting Egfr expression levels in long-bones leading to accelerated osteoblast differentiation thus not (Fig. S7d). allowing a sufficient number of osteoprogenitor cells to −/− Importantly, after rapamycin treatment bones of Egfr accumulate to form proper bones. Under normal physiolo- embryos showed BV/TV comparable to WT mice (Figs. 7b, gical conditions EGFR/ERK-mediated IGFBP-3 is essential c). Furthermore, Osteocalcin mRNA levels in femurs of to suppress IGF-1R/mTOR in order to ensure efficient -/- Egfr embryos from rapamycin-injected mothers were also osteoblasts maturation. normalized (Fig. 7d). In addition, rapamycin treatment also Many possible interactions between IGF-1R and EGFR normalized the ratio between Runx2 and Osteocalcin have been identified [20]. Cancer cells acquire resistance −/− mRNA expression in bones of Egfr embryos against EGFR inhibitor treatment via loss of IGFBP-3, which (Figs. S7e, f) providing evidence that EGFR signaling activates the IGF-1R signaling pathway [21, 22]. A tight suppresses mTOR during bone formation to prevent early regulation of IGFBP-3 signaling is not only essential for maturation of osteoprogenitor cells to ensure the develop- cancer treatment but also during bone development as shown ment of functional osteoblasts. by both Igfbp3 transgenic and knock-out mouse models. Long-bones of Igfbp3 transgenic mice overexpressing human IGFBP-3 demonstrate reduced trabecular and cortical bone −/− Discussion density [23]. Igfbp3 mice, on the other hand, develop a low-bone-mass phenotype comparable to Egfr-deficient mice In the present study, we show that EGFR-deficient mice comprising reduced trabecular bone volume and number with suffer from a complex bone phenotype with decreased increased trabecular separation [24]. In agreement with our bone mass, which starts before birth and persists to data, a link between EGFR and IGFBP-3 has also been adulthood. Moreover, deleting EGFR specifically in the described for primary human esophageal cells and esopha- osteoblast or osteoclast lineage demonstrates that EGFR geal squamous cell carcinomas indicating that EGFR indeed in the osteoblast lineage is essential for adequate bone directly regulates IGFBP-3 [17]. development. The mTOR-pathway plays an important role during Histological analyses revealed an enlarged zone of development by regulating cell survival, growth, differ- hypertrophic chondrocytes, which could be the reason for entiation and autophagy [25]. Recently, rapamycin-induced the subsequent bone defects. However, we show that both autophagy was shown to increase the number of osteoblasts endochondral as well as intramembranous ossification is and the mineralized area in fracture calluses of rats during defective in the absence of EGFR. Since intramembranous bone fracture healing [26]. mTOR signaling has also been ossification does not involve chondrocyte differentiation linked to other bone-related diseases like osteoarthritis and cartilage formation, our results suggest that the osteo- (OA). Patients suffering from OA show increased mTOR blast and bone defects are unlikely to result from chon- protein and mRNA levels in affected joints [27]. Addi- drocyte defects. Therefore, EGFR signaling seems to be tionally, rapamycin treatment or deletion of mTOR in required cell-autonomously in osteoblasts. Long-bones of chondrocytes reduced the severity of experimental OA in mice with osteoblast-specific deletion of EGFR showed mice [27, 28]. Reduced EGFR signaling, on the other hand, elevated Runx2 with reduced Colagen1a1 and Osteocalcin leads to a worse progression of experimental OA due to expression levels revealing an important role of EGFR increased cartilage destruction in gefitinib-treated mice [29] during mineralization. This finding also reflects results from and subchondral bone plate thickening with increased joint wa5/f published in vitro experiments suggesting that a major pain in genetically modified (Egfr Col2-Cre) animals function of the EGFR is to maintain a pool of osteopro- [30]. These findings suggest that EGFR might not only genitor cells by downregulating Runx2 and Osterix in order negatively regulate the mTOR-pathway during bone to prevent premature differentiation [19]. Despite the fact development, but also during OA progression. Further EGFR regulates osteoblast differentiation 1103 studies are needed to investigate the impact of EGFR sig- guidelines. All animal experiments conducted were com- naling on mTOR activation in bone-related diseases. pliant with federal laws and guidelines of the Medical Mice with osteoblast-specific IGF-1R deletion display University of Vienna. mineralization defects [31, 32]. mTOR signaling pathway activation via IGF-1 has been reported to play a major role Whole mount stainings, histomorphometry, in bone development by regulating osteoblast differentiation immunohistochemistry in adult mice [32]. Moreover, osteoblast-specific deletion of TSC2, a negative regulator of the mTOR pathway, leads to Mice were sacrificed at indicated time points. Whole mount elevated mTOR signaling with increased bone formation stainings were performed as described previously [36]. For starting around 6 weeks after birth. Interestingly, three histological stainings, bones were fixed in 4% PBS-buffered weeks after birth these mice showed an osteopenic-like formaline and embedded either in paraffin or methylmeta- phenotype with significantly increased trabecular separa- crylate. 5 μm paraffin sections were used for H.E.-stainings tion, reduced bone volume to tissue volume and reduced after decalcification in 0.5 M EDTA or uncalcified for Von- −/− number of trabecles [33]. As Egfr mice also exhibit an Kossa stainings (calvaria); methylmetacrylate was used for osteopenic bone phenotype with elevated mTOR expression Von-Kossa stainings (long bone) and for Movat-stainings in osteoblasts, we hypothesize that up-regulation of mTOR (osteoid). Histomorphometry was performed with Movat pathway might inhibit bone formation during embryonic and/or H&E-stainings according to the standardized proto- and early postnatal development, whereas it induces bone cols of the American Society for Bone and Mineral mineralization in older animals. Consistently, treatment of Research [37] on the Osteo-measure system (Osteometrix) pregnant dams with rapamycin largely rescued the low bone in a blinded fashion. Immunohistochemistry was performed −/− mass phenotype of Egfr embryos. on 4 µm formalin-fixed paraffin embedded and decalcified In summary, we demonstrate that impaired prolifera- femur sections. Primary antibodies (for a full list see tion and enhanced differentiation of osteoblasts is Table S1) were incubated overnight at 4 °C followed by responsible for the osteopenia and irregular mineralization HRP-based immunoreactivity detection (CST). Non- −/− ΔOb in bones of Egfr and Egfr mice. The bone defects of specific binding was blocked by applying TBS-T contain- −/− Egfr mice are not restricted to endochondral ossifica- ing 2% BSA and 5% normal goat serum. Quantifications of tion, since mineralization defects are also apparent in IHC stainings were performed in a blinded fashion by −/− skulls of Egfr pups. Therefore, defective osteoblast counting positive cells on the trabecular bone surface and maturation very likely is the driving force for the miner- results are shown as positive cells per bone perimeter. −/− alization defects in Egfr mice. We identified the mTOR-pathway as a positive regulator of osteoblast dif- Primary osteoblast cultures ferentiation, suppressed by EGFR/ERK/IGFBP-3-signal- ing and hyper-activated in its absence via IGF-1R. Future Osteoblasts were cultured in α-MEM containing ribonu- studies will address whether the cross-talk between these cleosides and deoxyribonucleosides (GlutaMAX, Sigma) important signaling pathways is also operating in other and 10% FBS (Autogen Bioclear). Primary osteoblasts were tissues and under pathological conditions. isolated from calvariae of neonatal mice (P1-P7) as pre- viously described [38] and seeded at a density of 5.000 cells/cm . For differentiation, ascorbic acid (50 μg/ml) and Materials and methods β-glycerolphosphate (10 mM) were added to the culture medium. Bone nodules were stained at differentiation day Mice 21 using Alizarin Red (Sigma). For BrdU stainings, osteo- blasts were cultured until 70% confluency and incubated −/− ΔOc Egfr mice have been described previously [4]. Egfr with 10 µM BrdU (Roche) for 4 h, before fixation with 70% f/f mice were generated by breeding Egfr mice [34]to LysM- ethanol and staining with an anti-BrdU antibody according ΔOb Cre [35] transgenic mice. Egfr mice were generated by to the manufacturer’s instructions (Becton Dickinson). f/f crossing Egfr mice with Runx2-Cre [13] transgenic mice Rapamycin (Wyeth), EGF (Roche) and IGF-1 (Promega) (kindly provided by Jan Tuckermann, University Ulm). were used in concentrations indicated in the respective ΔOb f/f f/+ Only male Egfr and littermate controls (Egfr , Egfr or figure legends. f/+ Egfr Runx2-Cre) with a C57BL/6 genetic background were used for experiments. Genotyping was performed as Primary osteoclast cultures previously described [4, 35, 13]. Mice were kept in the animal facility of the Medical University of Vienna in For osteoclast isolation, bone marrow cells were harvested accordance with institutional policies and federal from long-bones of 8 week old mice. Cells were cultured 1104 M. Linder et al. overnight in α-MEM containing 10% FBS. Non-adherent collected from osteoblast cultures on differentiation day 14. cells were harvested, counted and seeded in 6-well plates Osteocalcin (Alfa Aesar) and CTX-1 Elisa (RatLaps, IDS (1.0 × 10 cells/well) with M-CSF (50 ng/ml). 48 h later Immunodiagnostic Systems) were performed according RANKL (50 ng/ml) was added to induce differentiation for to the manufacturer’s instructions with serum isolated from additional 96 h. male mice at p21 and p210. Mouse IGFBP-3 Elisa (R&D Systems) was performed with 48 h-old supernatants Total RNA isolation, Real-time qRT-PCR analysis collected from osteoblast cultures on differentiation day 21 or from undifferentiated osteoblast precursors. Serum Total RNA from osteoblasts and whole bone was isolated IGFBP-3 levels where analyzed in serum isolated from p7 using peqGOLD TriFast reagent (Peqlab) or RNeasy Kit and p210 mice. For IGFBP-3 quantification in whole (Qiagen). cDNA synthesis was performed with ProtoScript tibia protein lysates from p7 mice, 20 µg protein/well II Reverse Transcriptase (NEB) according to the manu- were applied after Bradford-based protein measurement facturer’s instructions. Real-time qRT-PCR was performed (Bio-Rad). using the Power SYBR Green Master Mix (Thermo Fisher Scientific) together with the Applied Biosystems 7500 Fast Rapamycin treatment Real-Time PCR System (Thermo Fisher Scientific) using the following primers: Collagen type 1 alpha 1 (Col1a1) 5′- Rapamycin (Sigma) was diluted in injection vehicle con- ACCTGGTCCACAAGGTTTCC-3′ and 5′-GACCCATT taining 10% PEG-400 and 17% Tween-80 in 1 × PBS. GGACCTGAACCG-3′; Collagen type 1 alpha 2 (Col1a2) Mice were randomly assigned into two groups and injected 5′-GGTCCAAGAGGAGAACGTGG-3′ and 5′-TGGGAC every 12 h between E15.5 and E16.5 either with 5 mg CTCGGCTTCCAATA-3′; Collagen type 2 alpha 1 Rapamycin per kg bw in 200 µl injection vehicle or with (Col2a1)5′-GGCCAGGATGCCCGAAAATTA-3′ and 5′- 200 µl injection vehicle alone according to a published CGCACCCTTTTCTCCCTTGT-3′; Collagen type 10 alpha protocol [18].The investigators were not blinded during the 1(Col10a1)5′-CATCTCCCAGCACCAGAATC-3′ and 5′- experiment. GCTAGCAAGTGGGCCCTTTA-3′; Epidermal growth factor receptor (Egfr)5′-TTGGAATCAATTTTA- Statistical methods CACCGAAT-3′ and 5′-GTTCCCACACAGTGACACCA- 3′; Osteocalcin (Ocn) 5′-AGACTCCGGCGCTACCTT-3′ Sample size calculation: For in vivo treatment experiments and 5′-CTCGTCACAAGCAGGGTTAAG-3′; Osteonectin a minimum of six embryos per group were considered, (On) 5′-TCTCAAAGTCTCGGGCCAAC-3′ and 5′-ATG- which ensures a 90% power to detect a difference in means CAAATACATCGCCCCCT-3′; Osteopontin (Opn)5′- of 2 standard deviations at the significance level of 0.05. CTGGCTGAATTCTGAGGGACT-3′ and 5′-TTCTGT Based on the central limit theorem, we can assume a normal GGCGCAAGGAGATT-3′; Osterix (Osx)5′-TGCCTGAC distribution of mean values even if the underlying variable TCCTTGGGACC-3′ and 5′-TAGTGAGCTTCTTCCTCA is not perfectly normally distributed. Unless otherwise sta- AGCA-3′; Runt-related transcription factor 2 (Runx2) ted experiments were performed at least 2 times and data are 5′-GCCGGGAATGATGAGAACTA-3′ and 5′-GGACCGT shown as mean ± s.e.m. For analyses of IHC and qRT-PCR CCACTGTCACTTT-3′; Expression levels were standar- data, univariable comparisons of expression values between dized to the primer set specific for TATA-binding protein groups were analyzed by unpaired two-tailed Student’s t-test (Tbp):5′-GGGGAGCTGTGATGTGAAGT-3′ and 5′- with f-test to ensure comparable variances between the CCAGGAAATAATTCTGGCTCAT-3′. groups. For analysis of hypertrophic chondrocyte zone, BV/ TV and qRT-PCR analysis after Rapamycin treatment, one- Western blot analysis way ANOVA was applied. A p-value below 0.05 was considered statistically significant and was marked with a Western blot analysis was performed as previously descri- star (*), p < 0.01 with 2 stars (**), p < 0.001 with 3 stars bed [39]. For a full list of the antibodies used, please see (***) and p < 0.0001 with 4 stars (****). For analyses, SAS Table S1. for Windows 9.1.3 (The SAS Institute, Inc., Cary, North Carolina, USA) and Prism 6 (GraphPad) were used. Enzyme-linked immunosorbent assay (ELISA) Acknowledgements We thank Mruniya Vaibhavkumar Gawali, Theresia Lengheimer and Martina Hammer for maintaining mouse Mouse IGF-1 (Quantikine, R&D Systems) and IGF-2 colonies, Sarah Bardakji and Malgorzata Tryniecki for genotyping, (RayBiotech) Immunoassays were performed according Temenuschka Baykuscheva-Gentscheva for technical assistance and to manufacturer’s instructions with 48 h-old supernatants Alexander Kainz for helping with statistical analysis. The authors also EGFR regulates osteoblast differentiation 1105 thank Reinhold Erben for helpful discussions. This work was 12. Yarden Y. The EGFR family and its ligands in human cancer. supported by the Austrian Science Fund (FWF) (DK W1212), the signalling mechanisms and therapeutic opportunities. Eur J Can- FWF-grant I764-B13, the Deutsche Forschungsgemeinschaft (CRC cer. 2001;37:S3–8. 1149, Priority Program Immunobone Tu220/6 to J.T.) and “Fonds der 13. Rauch A, Seitz S, Baschant U, Schilling AF, Illing A, Stride B, Stadt Wien für innovative interdisziplinäre Krebsforschung” et al. Glucocorticoids suppress bone formation by attenuating (AP00288OFF). M.S. is supported by an ERC-Advanced grant osteoblast differentiation via the monomeric glucocorticoid (ERC-2015-AdG TNT-Tumors 694883). receptor. Cell Metab. 2010;11:517–31. 14. Guntur AR, Rosen CJ. IGF-1 regulation of key signaling path- ways in bone. Bone Rep. 2013;2:437. Compliance with ethical standards 15. Mohseni-Zadeh S, Binoux M. Insulin-like growth factor (IGF) binding protein-3 interacts with the type 1 IGF receptor, Conflict of interest The authors declare that they have no conflict of reducing the affinity of the receptor for its ligand: an alternative interest. mechanism in the regulation of IGF action. Endocrinology. 1997;138:5645–48. 16. Ricort JM, Binoux M. Insulin-like growth factor (IGF) Open Access This article is licensed under a Creative Commons binding protein-3 inhibits type 1 IGF receptor activation Attribution 4.0 International License, which permits use, sharing, independently of its IGF binding affinity. Endocrinology. adaptation, distribution and reproduction in any medium or format, as 2001;142:108–13. long as you give appropriate credit to the original author(s) and the 17. Takaoka M, Harada H, Andl CD, Oyama K, Naomoto Y, source, provide a link to the Creative Commons license, and indicate if Dempsey KL, et al. Epidermal growth factor receptor regulates changes were made. The images or other third party material in this aberrant expression of insulin-like growth factor-binding protein article are included in the article’s Creative Commons license, unless 3. Cancer Res. 2004;64:7711–23. indicated otherwise in a credit line to the material. If material is not 18. Liu KJ, Arron JR, Stankunas K, Crabtree GR, Longaker MT. included in the article’s Creative Commons license and your intended Chemical rescue of cleft palate and midline defects in conditional use is not permitted by statutory regulation or exceeds the permitted GSK-3beta mice. Nature. 2007;446:79–82. use, you will need to obtain permission directly from the copyright 19. Zhu J, Shimizu E, Zhang X, Partridge NC, Qin L. EGFR signaling holder. To view a copy of this license, visit http://creativecommons. suppresses osteoblast differentiation and inhibits expression of org/licenses/by/4.0/. master osteoblastic transcription factors Runx2 and Osterix. J Cell Biochem. 2011;112:1749–60. References 20. Jones HE, Gee JM, Hutcheson IR, Knowlden JM, Barrow D, Nicholson RI. Growth factor receptor interplay and resistance in 1. Wagner EF, Karsenty G. Genetic control of skeletal development. cancer. Endocr Relat Cancer. 2006;13:S45–51. Curr Opin Genet Dev. 2001;11:527–32. 21. Chakravarti A, Loeffler JS, Dyson NJ. Insulin-like growth factor 2. Cohen MM Jr.. The new bone biology: pathologic, molecular, and receptor I mediates resistance to anti-epidermal growth factor clinical correlates. Am J Med Genet A. 2006;140:2646–706. receptor therapy in primary human glioblastoma cells through 3. Sibilia M, Kroismayr R, Lichtenberger BM, Natarajan A, continued activation of phosphoinositide 3-kinase signaling. Hecking M, Holcmann M. The epidermal growth factor Cancer Res. 2002;62:200–7. receptor: from development to tumorigenesis. Differentiation. 22. Guix M, Faber AC, Wang SE, Olivares MG, Song Y, Qu S, et al. 2007;75:770–87. Acquired resistance to EGFR tyrosine kinase inhibitors in cancer 4. Sibilia M, Wagner EF. Strain-dependent epithelial defects in mice cells is mediated by loss of IGF-binding proteins. J Clin Invest. lacking the EGF receptor. Science. 1995;269:234–38. 2008;118:2609–19. 5. Threadgill DW, Dlugosz AA, Hansen LA, Tennenbaum T, Lichti 23. Silha JV, Mishra S, Rosen CJ, Beamer WG, Turner RT, Powell U, Yee D, et al. Targeted disruption of mouse EGF receptor: DR, et al. Perturbations in bone formation and resorption in effect of genetic background on mutant phenotype. Science. insulin-like growth factor binding protein-3 transgenic mice. J 1995;269:230–34. Bone Miner Res. 2003;18:1834–41. 6. Sibilia M, Wagner B, Hoebertz A, Elliott C, Marino S, Jochum W, 24. Yakar S, Rosen CJ, Bouxsein ML, Sun H, Mejia W, Kawashima et al. Mice humanised for the EGF receptor display hypomorphic Y, et al. Serum complexes of insulin-like growth factor-1 mod- phenotypes in skin, bone and heart. Development. ulate skeletal integrity and carbohydrate metabolism. FASEB J. 2003;130:4515–25. 2009;23:709–19. 7. Wang K, Yamamoto H, Chin JR, Werb Z, Vu TH. Epidermal 25. Kim YC, Guan KL. mTOR: a pharmacologic target for autophagy growth factor receptor-deficient mice have delayed primary regulation. J Clin Invest. 2015;125:25–32. endochondral ossification because of defective osteoclast recruit- 26. Yang GE, Duan X, Lin D, Li T, Luo D, Wang L, et al. ment. J Biol Chem. 2004;279:53848–56. Rapamycin-induced autophagy activity promotes bone fracture 8. Zhang X, Tamasi J, Lu X, Zhu J, Chen H, Tian X, et al. Epidermal healing in rats. Exp Ther Med. 2015;10:1327–33. growth factor receptor plays an anabolic role in bone metabolism 27. Zhang Y, Vasheghani F, Li YH, Blati M, Simeone K, Fahmi H, in vivo. J Bone Mineral Res. 2011;26:1022–34. et al. Cartilage-specific deletion of mTOR upregulates autophagy 9. Hall KC, Hill D, Otero M, Plumb DA, Froemel D, Dragomir CL, and protects mice from osteoarthritis. Ann Rheum Dis. et al. ADAM17 controls endochondral ossification by regulating 2015;74:1432–40. terminal differentiation of chondrocytes. Mol Cell Biol. 28. Carames B, Hasegawa A, Taniguchi N, Miyaki S, Blanco FJ, Lotz 2013;33:3077–90. M. Autophagy activation by rapamycin reduces severity of 10. Chandra A, Lan S, Zhu J, Siclari VA, Qin L. Epidermal growth experimental osteoarthritis. Ann Rheum Dis. 2012;71:575–81. factor receptor (EGFR) signaling promotes proliferation and sur- 29. Zhang X, Zhu J, Liu F, Li Y, Chandra A, Levin LS, et al. Reduced vival in osteoprogenitors by increasing early growth response 2 EGFR signaling enhances cartilage destruction in a mouse (EGR2) expression. J Biol Chem. 2013;288:20488–98. osteoarthritis model. Bone Res. 2014;2:14015. 11. Karsenty G, Wagner EF. Reaching a genetic and molecular 30. Jia H, Ma X, Tong W, Doyran B, Sun Z, Wang L, et al. EGFR understanding of skeletal development. Dev Cell. 2002;2:389–406. signaling is critical for maintaining the superficial layer of 1106 M. Linder et al. articular cartilage and preventing osteoarthritis initiation. Proc 35. Clausen BE, Burkhardt C, Reith W, Renkawitz R, Forster I. Natl Acad Sci USA. 2016;113:14360–65. Conditional gene targeting in macrophages and granulocytes using 31. Zhang M, Xuan S, Bouxsein ML, von Stechow D, Akeno N, LysMcre mice. Transgenic Res. 1999;8:265–77. Faugere MC, et al. Osteoblast-specific knockout of the insulin-like 36. Wallin J, Wilting J, Koseki H, Fritsch R, Christ B, Balling R. The growth factor (IGF) receptor gene reveals an essential role of IGF role of Pax-1 in axial skeleton development. Development. signaling in bone matrix mineralization. J Biol Chem. 2002;277: 1994;120:1109–21. 44005–12. 37. Parfitt AM, Drezner MK, Glorieux FH, Kanis JA, Malluche H, 32. Xian L, Wu X, Pang L, Lou M, Rosen CJ, Qiu T, et al. Matrix Meunier PJ, et al. Bone histomorphometry: standardization of IGF-1 maintains bone mass by activation of mTOR in mesench- nomenclature, symbols, and units. Report of the ASBMR Histo- ymal stem cells. Nat Med. 2012;18:1095–1. morphometry Nomenclature Committee. J Bone Miner Res. 33. Riddle RC, Frey JL, Tomlinson RE, Ferron M, Li Y, Digirolamo 1987;2:595–10. DJ, et al. Tsc2 is a molecular checkpoint controlling osteoblast 38. Jochum W, David JP, Elliott C, Wutz A, Plenk H Jr., Matsuo K, development and glucose homeostasis. Mol Cell Biol. 2014; et al. Increased bone formation and osteosclerosis in mice 34:1850–62. overexpressing the transcription factor Fra-1. Nat Med. 34. Lanaya H, Natarajan A, Komposch K, Li L, Amberg N, Chen L, 2000;6:980–84. et al. EGFR has a tumour-promoting role in liver macrophages 39. Sibilia M, Fleischmann A, Behrens A, Stingl L, Carroll J, Watt FM, during hepatocellular carcinoma formation. Nat Cell Biol. et al. The EGF receptor provides an essential survival signal for 2014;16:972–81. 971–977 SOS-dependent skin tumor development. Cell. 2000;102:211–20. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Cell Death & Differentiation Springer Journals

EGFR controls bone development by negatively regulating mTOR-signaling during osteoblast differentiation

Free
13 pages
Loading next page...
 
/lp/springer_journal/egfr-controls-bone-development-by-negatively-regulating-mtor-signaling-32zevXckju
Publisher
Nature Publishing Group UK
Copyright
Copyright © 2018 by ADMC Associazione Differenziamento e Morte Cellulare
Subject
Life Sciences; Life Sciences, general; Biochemistry, general; Cell Biology; Stem Cells; Apoptosis; Cell Cycle Analysis
ISSN
1350-9047
eISSN
1476-5403
D.O.I.
10.1038/s41418-017-0054-7
Publisher site
See Article on Publisher Site

Abstract

−/− Mice deficient in epidermal growth factor receptor (Egfr mice) are growth retarded and exhibit severe bone defects that are poorly understood. Here we show that EGFR-deficient mice are osteopenic and display impaired endochondral and intramembranous ossification resulting in irregular mineralization of their bones. This phenotype is recapitulated in mice lacking EGFR exclusively in osteoblasts, but not in mice lacking EGFR in osteoclasts indicating that osteoblasts are responsible for the bone phenotype. Experiments are presented demonstrating that signaling via EGFR stimulates osteoblast proliferation and inhibits their differentiation by suppression of the IGF-1R/mTOR-pathway via ERK1/2-dependent up- −/− regulation of IGFBP-3. Osteoblasts from Egfr mice show increased levels of IGF-1R and hyperactivation of mTOR- −/− pathway proteins, including enhanced phosphorylation of 4E-BP1 and S6. The same changes are also seen in Egfr bones. Importantly, pharmacological inhibition of mTOR with rapamycin decreases osteoblasts differentiation as well as rescues the −/− low bone mass phenotype of Egfr fetuses. Our results demonstrate that suppression of the IGF-1R/mTOR-pathway by EGFR/ERK/IGFBP-3 signaling is necessary for balanced osteoblast maturation providing a mechanism for the skeletal phenotype observed in EGFR-deficient mice. Introduction Markus Linder and Manfred Hecking contributed equally to this work. Skeletal development requires complex and coordinated Edited by E Gottlieb interplay between mesenchymal cells—chondrocytes and osteoblasts—at various stages of differentiation. The suc- Electronic supplementary material The online version of this article (https://doi.org/10.1038/s41418-017-0054-7) contains supplementary cession of events whereby osteoblast formation follows material, which is available to authorized users. chondrocyte differentiation resulting in bone formation in long bones of vertebras is termed “endochondral ossifica- * Maria Sibilia Sibilia-Office@meduniwien.ac.at tion” [1]. Differentiation of osteoblasts is induced by up- regulation of specific transcription factors accompanied by Department of Internal Medicine I, Institute of Cancer Research, the expression of factors that facilitate mineralized extra- Comprehensive Cancer Center, Medical University of Vienna, cellular matrix formation [2]. Vienna, Austria Genetic ablation of the epidermal growth factor receptor Department of Internal Medicine 3, Friedrich-Alexander- (EGFR) in mice revealed its intricate role during embryonic University Erlangen-Nürnberg (FAU), Universitätsklinikum Erlangen, Erlangen, Germany and postnatal development [3, 4]. Mice lacking the EGFR −/− (Egfr ) have major organ defects and die either in utero or Department of Internal Medicine III, Medical University of Vienna, Vienna, Austria shortly after birth, depending on the genetic background [3- 5]. Among other developmental abnormalities EGFR- Spanish National Cancer Research Center (CNIO), Madrid, Spain deficient mice are severely growth retarded and exhibit Institute of Comparative Molecular Endocrinology, University of skeletal defects [6]. We have previously reported that long Ulm, Ulm, Germany bones of EGFR-deficient mice display a greatly increased Present address: Department of Internal Medicine III, Medical zone of hypertrophic chondrocytes, suggesting that EGFR University of Vienna, Waehringer Guertel 18-20, Vienna A-1090, Austria negatively regulates condrocyte maturation [6]. Similar 1234567890();,: EGFR regulates osteoblast differentiation 1095 wt -/- wt -/- −/− Egfr Egfr Egfr Egfr Fig. 1 Egfr mice are osteopenic. a Alcian blue and ab Alizarin red whole body mount showing skeletal mineralization wt −/− of Egfr and Egfr mice on postnatal day 7 (P7). b Skeletal preparations of WT and KO mice: femur and spine. c µCT wt image of 7-day-old Egfr (left) -/- and Egfr (right) mice d Von wt Kossa staining of Egfr and −/− Egfr calvaria at P7; scales: wt -/- Egfr Egfr 100 μm for lower and 20 µm for higher magnification. e Von Kossa staining showing wt calcification of Egfr and Egfr −/− tibiae on P7; scales: 500 μm for lower and and 100 μm for higher magnification. f Histomorphometric analysis of wt −/− Egfr and Egfr tibiae from wt -/- 1mm 1mm Egfr Egfr P1 to P14: Quantification of bone volume/tissue volume (BV/TV), trabecular number wt -/- (Tb.N), trabecular separation Egfr Egfr (Tb.Sp), trabecular thickness (Tb.Th) and osteoblast number per bone perimeter (N.Ob/B. Pm). P1: n = 6. P7: n = 5. P14: n = 6 WT, 3 KO mice wt -/- Egfr Egfr 50 20 30 15 500 wt Egfr -/- Egfr ** * 400 40 ** 300 30 *** ** ** ** 20 10 5 100 10 0 0 0 0 0 P1 P7 P14 P1 P7 P14 P1 P7 P14 P1 P7 P14 P1 P7 P14 observations were made by Wang et al. [7], who in addition shortened long-bones with increased cartilage mineraliza- found delayed primary ossification with irregular distribu- tion [9]. −/− tion of osteoblasts in Egfr embryos [7]. EGF treatment of WT calvariae increased the prolifera- Moreover, EGFR knock-in mice where the murine tion of osteoprogenitor cells and maintained them in an −/− EGFR is replaced by the human counterpart display low undifferentiated state [10]. Accordingly, Egfr osteoblasts EGFR activity in the bone and show impaired endochondral show reduced proliferation but elevated differentiation ossification and an increased hypertrophic chondrocyte indicating that EGFR is essential during osteoblast zone [6]. Similarly, mice with reduced EGFR activity by maturation [8]. However, the underlying molecular combined expression of a dominant-negative Egfr Wa5 mechanisms has so far not been investigated. It is also allele and deletion of an Egfrfloxed allele using Col1a1-Cre unclear whether the bone defects observed in adult mice wa5/f mice (Col1a1-Cre Egfr ), display bone abnormalities result from developmental defects or arise later during bone starting around 3 months of age [8]. Mice lacking the remodeling. The mouse models employed so far have not membrane-anchored metalloproteinase ADAM17, respon- allowed to investigate this effect, since incomplete EGFR sible for cleavage of several membrane-bound cytokines deletion was observed using Col-Cre mice and osteoblast- and growth factors including EGFR ligands also develop independent effects of the ubiquitously expressed, expanded zones of hypertrophic chondrocytes, and dominant-negative Wa5 on other ErbB family members chondrocyte-specific deletion of ADAM17 results in cannot be excluded [8]. BV/TV (%) -1 Tb.N (mm ) Tb.Sp (μm) Tb.Th (μm) -1 N.Ob/B.Pm (mm ) 1096 M. Linder et al. wt -/- Egfr Egfr Fig. 2 EGFR deletion leads to reduced proliferation and pERK1/2 in bone-lining cells. Representative images and quantifications of IHC stainings on femur sections from 7-day- wt −/− wt -/- old Egfr and Egfr mice against (a) p-Histone H3 (n = b 5), (b) PCNA (n = 5), and (c) pERK1/2 (n = 3). Scales: 200 µm (lower magnification) and 20 µm (higher magnification) wt -/- ** wt -/- Here we investigated the bone phenotype occurring in columns of calcified extracellular matrix (ECM) with a −/− the first weeks of age in Egfr mice and in adult mice in clearly delineated border were observed in WTs, these −/− which Egfr is conditionally deleted in the osteoblast lineage structures were lacking in Egfr calvariae (Fig. 1d). f/f ΔOb using Egfr Runx2-Cre (Egfr ) mice. We found that Taken together our results show that Egfr deletion leads EGFR signaling in osteoblasts negatively regulates IGF-1R/ to impaired bone development in newborn mice with mTOR pathway via ERK1/2 dependent up-regulaion of defects in both endochondral and intramembranous IGFBP-3 to coordinate differentiation during embryonic ossification. −/− and postnatal bone formation. Egfr long-bones displayed a low-bone-mass pheno- type with less calcified bone and fewer bony trabeculae on P7 (arrowheads; Fig. 1e) and a thickened growth plate −/− RESULTS (arrows; Fig. 1e). Egfr tibiae exhibited thicker zones of ECM located at the cortical sides reaching into the center of −/− Egfr mice show impaired endochondral and the bone (Fig. S1b) indicating that the mineralization pro- −/− intramembranous ossification cess in Egfr bones was impaired due to misbalanced deposition of ECM by osteoblasts. −/− We first performed an analysis of the skeleton of Egfr Histomorphometric analyses confirmed that the ratio of mice that survived until postnatal day 7 (P7). Bones of Egfr bone volume over tissue volume (BV/TV) was significantly −/− −/− mice were less mineralized and reduced in length lower in Egfr mice (Fig. 1f). The trabecular number (Tb. compared to WT littermates (Fig. 1a). Whole-mount body N) was decreased while trabecular separation (Tb.Sp) was staining revealed reduced centers of secondary ossification increased at P7 and P14, although trabecular thickness (Tb. −/− in long bones and irregular calcification of vertebral end- Th) was not significantly changed (Fig. 1f). While Egfr plates in EGFR-deficient mice (Fig. 1b). Additionally, Egfr mice were born with osteoblast numbers (N.Ob) compar- −/− mice showed reduced mineralization of costal cartilage able to WT levels, their amount was significantly decreased (Fig. S1a). on P14 (Fig. 1f). While most bones develop by endochondral ossifica- tion, the lateral clavicles and parts of the skull are formed EGFR is essential for osteoblast proliferation and by intramembranous ossification, where mesenchymal ERK1/2 activation cells directly differentiate into osteoblasts without chondrocyte involvement [1, 11]. To determine whether As osteoblasts are essential for bone mineralization we next −/− the bone phenotype of Egfr mice can occur inde- focused on the role of EGFR during osteoblastogenesis. We pendently of the cartilage defects, we examined skulls of found decreased proliferation of primary pre-osteoblasts −/− Egfr mice. µCT analysis revealed an impaired cranial lacking the EGFR [6] (Fig. S2a), without any significant suture closure on day 14 (Fig. 1c), indicating that EGFR differences in the number of apoptotic cells (Fig. S2b). −/− also plays an important role during intramembranous Additionally, Egfr osteo-progenitors showed reduced ossification. Furthermore, while straight, well-organized BrdU and Cyclin D1 levels, indicating that EGFR deletion p-ERK1/2 PCNA p-Histone H3 PCNA pos / BPm p-H3 pos / BPm pERK1/2 pos / BPm EGFR regulates osteoblast differentiation 1097 ΔOb Fig. 3 Egfr mice phenocopy a −/− the bone phenotype of Egfr mice. a H&E stainings showing distal femurs with increased zone of hypertrophic chondrocytes in 6-day-old ΔOb Egfr mice; scales: 200 µm (lower magnification) and 100 µm (higher magnification). b wt ΔOb Egfr Egfr µCT image of femurs from 3- wt ΔOb months old Egfr and Egfr b c littermate; scale: 1 mm. c P6 P21 P90 Quantification of femur length 6.0 10 * **** wt ΔOb of Egfr and Egfr mice with indicated age 5.5 5.0 8 12 4.5 wt wt ΔOb ΔOb wt ΔOb Egfr Egfr Egfr Egfr Egfr Egfr −/− leads to cell autonomous proliferation defects without mice was significantly increased, comparable to Egfr ΔOb affecting apoptosis (Figs. S2c, d). mice (Fig. 3a). Importantly, Egfr mice showed reduced To confirm that the observed defects are also occurring length of long bones which was significant by P21 and in vivo, we evaluated the proliferation of bone-lining cells became more severe with age (Figs. 3b, c). These results −/− wt in femoral sections of Egfr and Egfr mice. The number demonstrate that EGFR signaling in osteoblasts is essential of cells positive for the mitosis marker p-Histone H3 for proper bone development. (Fig. 2a) and the S-phase related marker PCNA (Fig. 2b) ΔOb −/− were significantly reduced in Egfr mice indicating that Adult Egfr mice develop a low-bone-mass EGFR is crucial for proliferation during bone development. phenotype The ERK pathway, a major EGFR downstream signaling pathway, plays a central role in cell proliferation [12]. The augmented zone of hypertrophic chondrocytes was Therefore, we analyzed the phosphorylation of ERK1/2 in accompanied by increased expression of the hypertrophic −/− ΔOb bone lining cells at P7. Egfr mice exhibited significantly chondrocyte marker Col10a1 in long bones of Egfr mice reduced numbers of p-ERK1/2 positive cells on their tra- on P21 (Fig. 4a). Significantly elevated Runt-related tran- becular bone (Fig. 2c), suggesting that the proliferation scription factor-2 (Runx2) mRNA levels together with defects during bone development are based on impaired reduced Colagen1a1 (Col1a1) mRNA and reduced Osteo- ERK1/2 activation. calcin (Ocn) mRNA and protein levels (Figs. 4a, S4a) indicate that EGFR deletion in osteoblasts leads to impaired Osteoblast-specific deletion of EGFR leads to bone mineralization due to premature differentiation of osteo- defects progenitors. Histomorphological analysis revealed a pro- gressive, low-bone-mass phenotype with decreased bone −/− ΔOb To address whether the bone phenotype in Egfr mice is volume and trabecular number in adult Egfr mice f/f due to cell-autonomous defects in osteoblasts, Egfr mice (Figs. 4b, c). Additional trabecular bone markers further were crossed to an osteoblast-specific Cre line (Runx2-Cre), showed reduced trabecular thickness and increased spacing f/f ΔOb ΔOb to generate Egfr Runx2-Cre (Egfr ) mice [13]. Com- in Egfr mice (Fig. S4b). Less osteoblasts on the trabe- plete deletion of EGFR was confirmed by Western Blot in cular bone and reduced osteocalcin serum levels (Fig. 4d) cultured osteoblasts and by IHC in long bones indicate that the low-bone-mass is based on osteoblast (Figs. S3a, b). As shown by qRT-PCR from RNA isolated defects. wt ΔOb from bone and cartilage of Egfr and Egfr femurs, To exclude that EGFR in osteoblasts indirectly affects deletion of Egfr was restricted to bone tissue, but not car- osteoclastogenesis, osteoclast-specific markers were ana- ΔOb tilage (Fig. S3c). Egfr mice developed normally without lyzed in long bones and serum. No significant differences in any significant differences in overall body length (Fig. S3d). osteoclast number could be detected neither in young nor in ΔOb ΔOb ΔOb On P6 the zone of hypertrophic chondrocytes of Egfr adult Egfr mice. Furthermore, Egfr mice did not femur lenght (mm) femur lenght (mm) femur lenght (mm) 1098 M. Linder et al. ΔOb *** Fig. 4 Egfr mice show severe wt ab Egfr bone defects. a qRT-PCR 2.0 ΔOb Egfr analysis: RNA isolated from whole femurs of 21-day-old wt ΔOb 1.5 * Egfr and Egfr mice; n = 7 WT, 9 ΔOb. b H&E stainings **** **** wt Egfr showing distal femurs of 3- 1.0 wt ΔOb months-old Egfr and Egfr mice; scales: 200 µm. c Histomorphometric analysis of 0.5 WT/ΔOb long-bones at P6, P21 P90 and P210: Quantification of 0.0 bone volume/tissue volume (BV/TV) and trabecular number ΔOb Egfr (Tb.N). P6: n = 4 WT, 3 ΔOb. P21: n = 5. P90: n = 8 WT, 6 wt wt Egfr c 25 Egfr d ΔOb. P210: n = 4 WT, 6 ΔOb 25 30 1000 ΔOb ΔOb Egfr Egfr *** mice. d Osteoblast number (N. 20 800 Ob/B.Pm) and surface on the p=0.053 15 600 trabecular bone (Ob.S/BS) at 10 400 P21 (n = 7 WT, 10 ΔOb) and P210 (n = 6 WT, 8 ΔOb) and 5 5 200 Osteocalcin as measured by 0 0 0 0 ELISA in Serum at P21 (n = 4 P6 P21 P90 P21 0 age=P21 wt 6 Egfr 15 20 250 WT, 6 ΔOb) and P210 (n = 5 ΔOb Egfr WT, 9 ΔOb) 4 *** 10 2 5 0 0 0 0 P6 P21 P90 P21 0 age=P210 show any differences of the serum biomarker for bone [14], we investigated whether EGFR regulates bone resorption C-terminal telopeptide (CTX-1) (Fig. S4c). development by interacting with the IGF-1R signaling Additionally, we assessed whether EGFR directly affects pathway. We detected elevated levels of total and phos- f/f osteoclast development by breeding Egfr mice to LysM- phorylated IGF1Rβ in differentiated osteoblasts isolated −/− −/− Cre mice that express Cre recombinase in the myeloid from Egfr mice (Fig. 5a). Furthermore, Egfr osteo- ΔOc ΔOc lineage (Egfr ). Osteoclasts isolated from Egfr mice blasts showed increased total and phosphorylated protein showed reduced EGFR protein levels (Fig. S4d), but did not levels of the IGF-1R adapter protein insulin receptor sub- display any bone defects nor differences in the number of strate 1 (IRS-1) and its downstream target mTOR (Fig. 5a). osteoclasts in trabecular bones or serum CTX-1 (Fig. S4e). Importantly, IGF-1R/IRS1/mTOR up-regulation was ligand −/− Bone-marrow derived pre-osteoclasts from Egfr mice independent as the levels of IGF-1 and IGF-2 were not did not show any significant differences in their ability to altered (Fig. S5a). form osteoclasts in vitro (Fig. S4f). Finally, OC number in To investigate the kinetics of mTOR activation we next trabecular bones and serum CTX-1 levels were not altered analyzed multiple time points during osteoblast differ- -/- in Egfr mice (Fig. S4g) indicating that lack of EGFR does entiation. IGF-1R/mTOR-pathway proteins were con- not affect osteoclastogenesis. sistently present at higher levels and were hyper- −/− phosphorylated during differentiation in Egfr osteo- −/− Enhanced differentiation of Egfr osteoblasts blasts indicating that IGF-1R/mTOR-signaling remained correlates with IGF-1R/mTOR activation elevated throughout the whole culture period (Fig. 5b). IHC staining on femur sections of WT and EGFR- Once confirmed that the defects are primarily in the osteo- deficient mice at P7 revealed that the mTOR-signaling wt blast lineage, we employed primary osteoblasts from Egfr pathway was also altered in vivo. In line with the in vitro −/− and Egfr mice to investigate the underlying molecular findings, significantly increased phosphorylation of mTOR −/− mechanism. As osteoblasts from Egfr mice display and its main downstream targets 4E-BP1 and S6 protein −/− enhanced differentiation [6] and the IGF-1R pathway was were observed in Egfr long-bones (Figs. 5c, d). Addi- ΔOb shown to play a central role during osteoblast differentiation tionally, Egfr mice also showed reduced p-S6 protein Egf r Osx Run x2 Col 1a1 Col 1a2 Ocn Opn On Col 2a1 Col 10a1 -1 Tb.N (mm ) BV/TV (%) mRNA fold change N.Ob/B.Pm N.Ob/B.Pm Ob.S/BS (%) Ob.S/BS (%) Ocn (ng/ml) Ocn (ng/ml) EGFR regulates osteoblast differentiation 1099 a b Fig. 5 Enhanced differentiation kDa wt -/- correlates with mTOR-signaling. 175 EGFR wt wt wt kDa wt -/- -/- -/- wt -/- a Western blot analysis of Egfr p-IGF-1Rβ −/− EGFR and Egfr osteoblasts under 175 IGF-1Rβ differentiation conditions; IGF-1Rβ isolated on differentiation day Tubulin p-AKT 14. b Western blot analyses of wt −/− AKT Egfr and Egfr osteoblasts 180 p-IRS-1 under differentiation conditions p-S6 180 IRS-1 on days 6, 9, 12, and 15. c S6 50 Tubulin 32 Immunohistochemical staining p-4E-BP1 of p-mTOR, p-4E-BP1 and p-S6 289 p-mTOR on trabecular bone sections from 4E-BP1 wt distal femurs of P7 Egfr and mTOR Tubulin −/− 50 Egfr littermates; scales: 200 50 Tubulin µm (lower magnification) and 20 µm (higher magnification). d Quantification of IHC staining, wt -/- Egfr Egfr shown as positive cells per bone perimeter (B.Pm); n ≥ 3 c d ** wt -/- ** wt -/- p=0.052 wt -/- levels in bone-lining cells indicating that this activation the EGF-induced hypo-differentiation phenotype resulting in depends on osteoblastic EGFR signaling (Figs. S5b, c). normalized bone nodule formation comparable to untreated controls (Fig. 6i). Taken together our results show that IGF- Interplay between EGFR- and IGF-1R-pathways in 1R signaling enhances, whereas EGFR signaling inhibits osteoblast differentiation osteoblast differentiation and that EGFR signaling dom- inates by negatively regulating IGF-1R via ERK1/2. To analyze the cross-talk between EGFR and IGF-1R- To dissect the underlying molecular mechanism we signaling during osteoblast differentiation, WT osteoblasts analyzed the activation of EGFR and IGF-1R downstream were cultured under differentiation-inducing conditions proteins in differentiated WT osteoblasts cultured for together with IGF-1, EGF and/or the ERK1/2 inhibitor 21 days in the presence of EGF / IGF-1 and U0126. EGF U0126. At day 21 bone nodule formation was assessed as a treatment prevented phosphorylation of IGF-1Rβ with functional read-out for differentiation. Mineralization was reduced activation of the mTOR/S6/4E-BP1 pathway, enhanced by IGF-1 treatment (Figs. 6a, b) and completely whereas IGF-1 induced the phosphorylation of IGF-1Rβ/ abolished by EGF (Fig. 6c). Addition of EGF was able to mTOR/S6/4E-BP1 (Fig. 6j). When osteoblasts were cul- suppress IGF-1 induced differentiation in a dose-dependent tured with both growth factors, activation was again manner with complete inhibition at 100 ng/ml (Figs. 6d–f). reduced suggesting that EGFR signaling is able to block IGF-1 induced differentiation was further increased when differentiation via IGF-1Rβ inhibition. Importantly, EGF- ERK1/2 signaling was blocked by U0126 (Figs. 6g, h). ERK induced downregulation of the IGF-1Rβ pathway was partly inhibition together with EGF and IGF-1 stimulation rescued restored when ERK1/2 was blocked, indicating that EGFR p-S6 p-4E-BP1 p-mTOR p-S6 pos. p-4E-BP1 pos. p-mTOR pos. cells / BPm cells / BPm cells / BPm 1100 M. Linder et al. a d g j - + - ++ EGF k EGF (10ng/ml) untreated IGF-1 (100ng/ml) U0126 (10μM) - - IGF-1 + ++ kDa - -- +- U0126 - - EGF + ++ - - + ++ IGF-1 p-IGF-1Rβ - -- +- U0126 kDa 45 IGFBP-3 IGF-1Rβ 116 Vinculin p-ERK b e h EGF (50ng/ml) U0126 44 IGF-1 (100ng/ml) ERK IGF-1 (100ng/ml) IGF-1 (100ng/ml) HSP90 kDa 0 01 .1 10 100 1000 Afatinib (nM) 289 p-mTOR 175 p-EGFR 289 mTOR 175 EGFR p-S6 U0126 289 p-mTOR c f i EGF (100ng/ml) EGF (100ng/ml) S6 EGF (100ng/ml) 32 IGF-1 (100ng/ml) IGF-1 (100ng/ml) 289 mTOR HSP90 IGFBP-3 p-4E-BP1 50 Tubulin 4E-BP1 Tubulin m n p Serum Tibia 1250 10 1000 8 750 6 *** 500 4 wt -/- 250 2 Egfr Egfr 0 0 wt -/- wt -/- Egfr Egfr Egfr Egfr wt -/- Egfr Egfr kDa 45 IGFBP-3 wt -/- Egfr Egfr 116 Vinculin Fig. 6 Differentiation in WT osteoblasts is mediated by specific cultured in αMEM + 10%FCS after 4 h treatment with indicated components of the IGF-1R-pathway and inhibited by EGF. a–i Ali- concentrations of Afatinib. m IGFBP-3 protein levels in serum (n = 7) wt −/− zarin red staining of WT osteoblasts after 21 days (D21) in culture and n whole tibia lysates (n = 4) of 7-day old Egfr and Egfr under differentiation conditions (+AA, βGP) with EGF, IGF-1, and/or littermates. o Western Blot analysis of whole tibia protein lysates wt −/− U0126 over the whole culture period. Pictures taken from a 6-well isolated from 7-day old Egfr and Egfr littermates. p Alizarin red −/− plate. j, k Western blot analysis of differentiated WT osteoblasts (D21) staining of differentiated WT and Egfr osteoblasts (D21) cultured cultured with EGF (100 ng), IGF-1 (100 ng) and/or U0126 (10 µM). l with vehicle (DMSO) or Rapamycin (10 nM). Stained with alizarin red Western Blot analysis of undifferentiated osteoblast precursor cells negatively regulates differentiation by down-regulating ERK1/2 and at the same time reduced the phosphorylation IGF-1Rβ/mTOR signaling via ERK1/2 (Fig. 6j). No dif- of IGF-1Rβ whereas IGF-1 stimulation did not affect ferences in insulin receptor β (IRβ) phosphorylation could ERK1/2 signaling (Fig. S6b). be detected indicating that EGF stimulation exclusively To investigate the mechanism how EGFR signaling downregulates IGF-1R without affecting IRβ activation suppresses IGF-1R/mTOR signaling we next analyzed (Fig. S6a). IGFBP-3 levels, as IGFBP-3 is known to modulate and To prove that reduced activation of IGF-1Rβ is a direct repress IGF-1R signaling [15, 16]. Moreover, it has been consequence of EGF stimulation, we cultured WT osteo- shown that EGFR directly regulates IGFBP-3 in primary blasts under differentiation conditions for 21 days, starved esophageal cells [17]. We found elevated IGFBP-3 levels in them for 24 h and stimulated for 10 min with EGF or IGF-1. osteoblasts cultured together with EGF or with EGF and As expected, EGF treatment induced a strong activation of IGF-1 whereas IGF-1 alone had no effect (Fig. 6k). IGFBP-3 (ng/ml) IGFBP-3 (pg/μg protein) Rapamycin DMSO EGFR regulates osteoblast differentiation 1101 wt Egfr *** -/- E15.5 E16.5 E18.5 Egfr Rapamycin - 5mg/kg bw, 2x/d Analyze embryonic bones b 2 +Rapamycin wt Egfr Ocn 2.0 wt -/- -/- Egfr Egfr Egfr 1.5 1.0 0.5 wt -/- Egfr Egfr 0.0 +Rapamycin −/− Fig. 7 mTOR inhibition partially rescues bone phenotype of Egfr volume relative to total volume (BV/TV) after rapamycin/vehicle mice. a Timeline showing experiment outline for in utero Rapamycin treatment; n = 7 WT, 6 KO mice for vehicle and 6 WT, 8 KO mice for treatment. Pregnant mice were subcutaneously injected with 5 mg rapamycin treatment; 5 mothers per treatment group. d mRNA Rapamycin per kg bodyweight or Injection-vehicle twice a day on expression levels of Osteocalcin (Ocn) in femurs of E18.5 mice E15.5 and E16.5 b H&E staining of trabecular bone sections showing measured by qRT-PCR; n = 9 WT, 5 KO mice for vehicle and 7 WT, wt −/− distal femurs of Egfr and Egfr embryos on E18.5 after Rapa- 5 KO mice for rapamycin treatment mycin/vehicle treatment; scale: 200 µm. c Quantification of bone Importantly, IGFBP-3 up-regulation was a direct con- formation was strongly reduced in the presence of rapa- sequence of ERK1/2 signaling, as additional treatment with mycin (Fig. 6p). Upon rapamycin treatment, phosphor- the ERK1/2 inhibitor U0126 normalized EGF-induced ylation of the mTOR downstream proteins 4E-BP1 and S6 −/− IGFBP-3 levels (Fig. 6k). In contrast, EGFR inhibition was down-regulated in Egfr cultures similarly to WT with Afatinib led to a dose-dependent decrease in IGFBP-3 osteoblasts (Fig. S6e) demonstrating that the increased −/− protein levels along with increased p-mTOR phosphoryla- differentiation in Egfr osteoblasts can be prevented by tion in osteoblast precursors (Fig. 6l). In addition, IGFBP-3 mTOR-inhibition. was also strongly reduced in the supernatant of osteoblast Taken together our data provide evidence that EGFR precursor cells after 48 h treatment with EGFR inhibitor as controls osteoblasts differentiation via ERK-dependent compared to DMSO treated controls (Fig. S6c). IGFBP-3 up-regulation, which ensures proper osteoblast In line with our in vitro results, we also found sig- maturation by controlling IGF-1R/mTOR signaling. −/− nificantly reduced IGFBP-3 in the serum of Egfr and ΔOb Egfr mice (Figs. 6m, S6d) indicating that EGFR sig- mTOR inhibition partially rescues bone phenotype −/− naling in osteoblasts is essential for IGFBP-3 production. of Egfr embryos IGFBP-3 levels were also reduced in whole tibia protein −/− lysates of Egfr mice, as revealed by both ELISA and We next analyzed whether mTOR inhibition during western blot analysis (Figs. 6n, o). These results demon- embryonic development, when mineralization starts, can strate that EGFR is required for IGFBP-3 production and normalize the bone defects in EGFR-deficient mice. Phar- suppression of IGF-1R/mTOR activation thus providing a macological inhibition of mTOR during gestation has pre- mechanistic link between EGFR and IGF-1R signaling and viously been reported not to cause any bone-specific side osteoblast differentiation. effects in mice [18]. We injected pregnant females from To further show that the hyper-differentiation pheno- EGFR heterozygous intercrosses with rapamycin or vehicle −/− type of Egfr osteoblasts is indeed a consequence of twice a day on E15.5 and on E16.5 and analyzed embryonic elevated mTOR activation we next inhibited mTOR in bones at E18.5 (Fig. 7a). Rapamycin treatment was not differentiating osteoblasts using rapamycin. Bone nodule teratogenic nor did it affect litter size or viability of pups Rapamycin Vehicle mRNA fold change BV/TV (%) 1102 M. Linder et al. −/− (Fig. S7a). Inhibition of mTOR signaling pathway was that Egfr osteoblasts showed elevated mineralization −/− ΔOb confirmed by p-S6 IHC staining on femurs of fetuses in vitro, both Egfr and Egfr mice are osteopenic. This obtained from rapamycin or vehicle-treated mothers apparent discrepancy might be due to the fact that osteo- (Fig. S7b). progenitor cells lacking the EGFR, which display pro- We could not observe any effect on hypertrophic chon- liferation defects, cannot form sufficient numbers of −/− drocyte zone in embryonic Egfr bones after rapamycin osteoblasts to guarantee proper maturation and ossification treatment (Fig. S7c), which is in line with our hypothesis in vivo. that the hypertrophic chondrocyte phenotype is not We identified the mTOR-pathway as a positive regulator responsible for the impaired bone development. However, of osteoblast differentiation that is suppressed by EGFR chemical inhibition of mTOR increased the zone of signaling. In the absence of EGFR, IGF-1R/mTOR signal- hypertrophic chondrocytes in WT animals (Fig. S7c) ing is up-regulated due to reduced IGFBP-3 signaling without affecting Egfr expression levels in long-bones leading to accelerated osteoblast differentiation thus not (Fig. S7d). allowing a sufficient number of osteoprogenitor cells to −/− Importantly, after rapamycin treatment bones of Egfr accumulate to form proper bones. Under normal physiolo- embryos showed BV/TV comparable to WT mice (Figs. 7b, gical conditions EGFR/ERK-mediated IGFBP-3 is essential c). Furthermore, Osteocalcin mRNA levels in femurs of to suppress IGF-1R/mTOR in order to ensure efficient -/- Egfr embryos from rapamycin-injected mothers were also osteoblasts maturation. normalized (Fig. 7d). In addition, rapamycin treatment also Many possible interactions between IGF-1R and EGFR normalized the ratio between Runx2 and Osteocalcin have been identified [20]. Cancer cells acquire resistance −/− mRNA expression in bones of Egfr embryos against EGFR inhibitor treatment via loss of IGFBP-3, which (Figs. S7e, f) providing evidence that EGFR signaling activates the IGF-1R signaling pathway [21, 22]. A tight suppresses mTOR during bone formation to prevent early regulation of IGFBP-3 signaling is not only essential for maturation of osteoprogenitor cells to ensure the develop- cancer treatment but also during bone development as shown ment of functional osteoblasts. by both Igfbp3 transgenic and knock-out mouse models. Long-bones of Igfbp3 transgenic mice overexpressing human IGFBP-3 demonstrate reduced trabecular and cortical bone −/− Discussion density [23]. Igfbp3 mice, on the other hand, develop a low-bone-mass phenotype comparable to Egfr-deficient mice In the present study, we show that EGFR-deficient mice comprising reduced trabecular bone volume and number with suffer from a complex bone phenotype with decreased increased trabecular separation [24]. In agreement with our bone mass, which starts before birth and persists to data, a link between EGFR and IGFBP-3 has also been adulthood. Moreover, deleting EGFR specifically in the described for primary human esophageal cells and esopha- osteoblast or osteoclast lineage demonstrates that EGFR geal squamous cell carcinomas indicating that EGFR indeed in the osteoblast lineage is essential for adequate bone directly regulates IGFBP-3 [17]. development. The mTOR-pathway plays an important role during Histological analyses revealed an enlarged zone of development by regulating cell survival, growth, differ- hypertrophic chondrocytes, which could be the reason for entiation and autophagy [25]. Recently, rapamycin-induced the subsequent bone defects. However, we show that both autophagy was shown to increase the number of osteoblasts endochondral as well as intramembranous ossification is and the mineralized area in fracture calluses of rats during defective in the absence of EGFR. Since intramembranous bone fracture healing [26]. mTOR signaling has also been ossification does not involve chondrocyte differentiation linked to other bone-related diseases like osteoarthritis and cartilage formation, our results suggest that the osteo- (OA). Patients suffering from OA show increased mTOR blast and bone defects are unlikely to result from chon- protein and mRNA levels in affected joints [27]. Addi- drocyte defects. Therefore, EGFR signaling seems to be tionally, rapamycin treatment or deletion of mTOR in required cell-autonomously in osteoblasts. Long-bones of chondrocytes reduced the severity of experimental OA in mice with osteoblast-specific deletion of EGFR showed mice [27, 28]. Reduced EGFR signaling, on the other hand, elevated Runx2 with reduced Colagen1a1 and Osteocalcin leads to a worse progression of experimental OA due to expression levels revealing an important role of EGFR increased cartilage destruction in gefitinib-treated mice [29] during mineralization. This finding also reflects results from and subchondral bone plate thickening with increased joint wa5/f published in vitro experiments suggesting that a major pain in genetically modified (Egfr Col2-Cre) animals function of the EGFR is to maintain a pool of osteopro- [30]. These findings suggest that EGFR might not only genitor cells by downregulating Runx2 and Osterix in order negatively regulate the mTOR-pathway during bone to prevent premature differentiation [19]. Despite the fact development, but also during OA progression. Further EGFR regulates osteoblast differentiation 1103 studies are needed to investigate the impact of EGFR sig- guidelines. All animal experiments conducted were com- naling on mTOR activation in bone-related diseases. pliant with federal laws and guidelines of the Medical Mice with osteoblast-specific IGF-1R deletion display University of Vienna. mineralization defects [31, 32]. mTOR signaling pathway activation via IGF-1 has been reported to play a major role Whole mount stainings, histomorphometry, in bone development by regulating osteoblast differentiation immunohistochemistry in adult mice [32]. Moreover, osteoblast-specific deletion of TSC2, a negative regulator of the mTOR pathway, leads to Mice were sacrificed at indicated time points. Whole mount elevated mTOR signaling with increased bone formation stainings were performed as described previously [36]. For starting around 6 weeks after birth. Interestingly, three histological stainings, bones were fixed in 4% PBS-buffered weeks after birth these mice showed an osteopenic-like formaline and embedded either in paraffin or methylmeta- phenotype with significantly increased trabecular separa- crylate. 5 μm paraffin sections were used for H.E.-stainings tion, reduced bone volume to tissue volume and reduced after decalcification in 0.5 M EDTA or uncalcified for Von- −/− number of trabecles [33]. As Egfr mice also exhibit an Kossa stainings (calvaria); methylmetacrylate was used for osteopenic bone phenotype with elevated mTOR expression Von-Kossa stainings (long bone) and for Movat-stainings in osteoblasts, we hypothesize that up-regulation of mTOR (osteoid). Histomorphometry was performed with Movat pathway might inhibit bone formation during embryonic and/or H&E-stainings according to the standardized proto- and early postnatal development, whereas it induces bone cols of the American Society for Bone and Mineral mineralization in older animals. Consistently, treatment of Research [37] on the Osteo-measure system (Osteometrix) pregnant dams with rapamycin largely rescued the low bone in a blinded fashion. Immunohistochemistry was performed −/− mass phenotype of Egfr embryos. on 4 µm formalin-fixed paraffin embedded and decalcified In summary, we demonstrate that impaired prolifera- femur sections. Primary antibodies (for a full list see tion and enhanced differentiation of osteoblasts is Table S1) were incubated overnight at 4 °C followed by responsible for the osteopenia and irregular mineralization HRP-based immunoreactivity detection (CST). Non- −/− ΔOb in bones of Egfr and Egfr mice. The bone defects of specific binding was blocked by applying TBS-T contain- −/− Egfr mice are not restricted to endochondral ossifica- ing 2% BSA and 5% normal goat serum. Quantifications of tion, since mineralization defects are also apparent in IHC stainings were performed in a blinded fashion by −/− skulls of Egfr pups. Therefore, defective osteoblast counting positive cells on the trabecular bone surface and maturation very likely is the driving force for the miner- results are shown as positive cells per bone perimeter. −/− alization defects in Egfr mice. We identified the mTOR-pathway as a positive regulator of osteoblast dif- Primary osteoblast cultures ferentiation, suppressed by EGFR/ERK/IGFBP-3-signal- ing and hyper-activated in its absence via IGF-1R. Future Osteoblasts were cultured in α-MEM containing ribonu- studies will address whether the cross-talk between these cleosides and deoxyribonucleosides (GlutaMAX, Sigma) important signaling pathways is also operating in other and 10% FBS (Autogen Bioclear). Primary osteoblasts were tissues and under pathological conditions. isolated from calvariae of neonatal mice (P1-P7) as pre- viously described [38] and seeded at a density of 5.000 cells/cm . For differentiation, ascorbic acid (50 μg/ml) and Materials and methods β-glycerolphosphate (10 mM) were added to the culture medium. Bone nodules were stained at differentiation day Mice 21 using Alizarin Red (Sigma). For BrdU stainings, osteo- blasts were cultured until 70% confluency and incubated −/− ΔOc Egfr mice have been described previously [4]. Egfr with 10 µM BrdU (Roche) for 4 h, before fixation with 70% f/f mice were generated by breeding Egfr mice [34]to LysM- ethanol and staining with an anti-BrdU antibody according ΔOb Cre [35] transgenic mice. Egfr mice were generated by to the manufacturer’s instructions (Becton Dickinson). f/f crossing Egfr mice with Runx2-Cre [13] transgenic mice Rapamycin (Wyeth), EGF (Roche) and IGF-1 (Promega) (kindly provided by Jan Tuckermann, University Ulm). were used in concentrations indicated in the respective ΔOb f/f f/+ Only male Egfr and littermate controls (Egfr , Egfr or figure legends. f/+ Egfr Runx2-Cre) with a C57BL/6 genetic background were used for experiments. Genotyping was performed as Primary osteoclast cultures previously described [4, 35, 13]. Mice were kept in the animal facility of the Medical University of Vienna in For osteoclast isolation, bone marrow cells were harvested accordance with institutional policies and federal from long-bones of 8 week old mice. Cells were cultured 1104 M. Linder et al. overnight in α-MEM containing 10% FBS. Non-adherent collected from osteoblast cultures on differentiation day 14. cells were harvested, counted and seeded in 6-well plates Osteocalcin (Alfa Aesar) and CTX-1 Elisa (RatLaps, IDS (1.0 × 10 cells/well) with M-CSF (50 ng/ml). 48 h later Immunodiagnostic Systems) were performed according RANKL (50 ng/ml) was added to induce differentiation for to the manufacturer’s instructions with serum isolated from additional 96 h. male mice at p21 and p210. Mouse IGFBP-3 Elisa (R&D Systems) was performed with 48 h-old supernatants Total RNA isolation, Real-time qRT-PCR analysis collected from osteoblast cultures on differentiation day 21 or from undifferentiated osteoblast precursors. Serum Total RNA from osteoblasts and whole bone was isolated IGFBP-3 levels where analyzed in serum isolated from p7 using peqGOLD TriFast reagent (Peqlab) or RNeasy Kit and p210 mice. For IGFBP-3 quantification in whole (Qiagen). cDNA synthesis was performed with ProtoScript tibia protein lysates from p7 mice, 20 µg protein/well II Reverse Transcriptase (NEB) according to the manu- were applied after Bradford-based protein measurement facturer’s instructions. Real-time qRT-PCR was performed (Bio-Rad). using the Power SYBR Green Master Mix (Thermo Fisher Scientific) together with the Applied Biosystems 7500 Fast Rapamycin treatment Real-Time PCR System (Thermo Fisher Scientific) using the following primers: Collagen type 1 alpha 1 (Col1a1) 5′- Rapamycin (Sigma) was diluted in injection vehicle con- ACCTGGTCCACAAGGTTTCC-3′ and 5′-GACCCATT taining 10% PEG-400 and 17% Tween-80 in 1 × PBS. GGACCTGAACCG-3′; Collagen type 1 alpha 2 (Col1a2) Mice were randomly assigned into two groups and injected 5′-GGTCCAAGAGGAGAACGTGG-3′ and 5′-TGGGAC every 12 h between E15.5 and E16.5 either with 5 mg CTCGGCTTCCAATA-3′; Collagen type 2 alpha 1 Rapamycin per kg bw in 200 µl injection vehicle or with (Col2a1)5′-GGCCAGGATGCCCGAAAATTA-3′ and 5′- 200 µl injection vehicle alone according to a published CGCACCCTTTTCTCCCTTGT-3′; Collagen type 10 alpha protocol [18].The investigators were not blinded during the 1(Col10a1)5′-CATCTCCCAGCACCAGAATC-3′ and 5′- experiment. GCTAGCAAGTGGGCCCTTTA-3′; Epidermal growth factor receptor (Egfr)5′-TTGGAATCAATTTTA- Statistical methods CACCGAAT-3′ and 5′-GTTCCCACACAGTGACACCA- 3′; Osteocalcin (Ocn) 5′-AGACTCCGGCGCTACCTT-3′ Sample size calculation: For in vivo treatment experiments and 5′-CTCGTCACAAGCAGGGTTAAG-3′; Osteonectin a minimum of six embryos per group were considered, (On) 5′-TCTCAAAGTCTCGGGCCAAC-3′ and 5′-ATG- which ensures a 90% power to detect a difference in means CAAATACATCGCCCCCT-3′; Osteopontin (Opn)5′- of 2 standard deviations at the significance level of 0.05. CTGGCTGAATTCTGAGGGACT-3′ and 5′-TTCTGT Based on the central limit theorem, we can assume a normal GGCGCAAGGAGATT-3′; Osterix (Osx)5′-TGCCTGAC distribution of mean values even if the underlying variable TCCTTGGGACC-3′ and 5′-TAGTGAGCTTCTTCCTCA is not perfectly normally distributed. Unless otherwise sta- AGCA-3′; Runt-related transcription factor 2 (Runx2) ted experiments were performed at least 2 times and data are 5′-GCCGGGAATGATGAGAACTA-3′ and 5′-GGACCGT shown as mean ± s.e.m. For analyses of IHC and qRT-PCR CCACTGTCACTTT-3′; Expression levels were standar- data, univariable comparisons of expression values between dized to the primer set specific for TATA-binding protein groups were analyzed by unpaired two-tailed Student’s t-test (Tbp):5′-GGGGAGCTGTGATGTGAAGT-3′ and 5′- with f-test to ensure comparable variances between the CCAGGAAATAATTCTGGCTCAT-3′. groups. For analysis of hypertrophic chondrocyte zone, BV/ TV and qRT-PCR analysis after Rapamycin treatment, one- Western blot analysis way ANOVA was applied. A p-value below 0.05 was considered statistically significant and was marked with a Western blot analysis was performed as previously descri- star (*), p < 0.01 with 2 stars (**), p < 0.001 with 3 stars bed [39]. For a full list of the antibodies used, please see (***) and p < 0.0001 with 4 stars (****). For analyses, SAS Table S1. for Windows 9.1.3 (The SAS Institute, Inc., Cary, North Carolina, USA) and Prism 6 (GraphPad) were used. Enzyme-linked immunosorbent assay (ELISA) Acknowledgements We thank Mruniya Vaibhavkumar Gawali, Theresia Lengheimer and Martina Hammer for maintaining mouse Mouse IGF-1 (Quantikine, R&D Systems) and IGF-2 colonies, Sarah Bardakji and Malgorzata Tryniecki for genotyping, (RayBiotech) Immunoassays were performed according Temenuschka Baykuscheva-Gentscheva for technical assistance and to manufacturer’s instructions with 48 h-old supernatants Alexander Kainz for helping with statistical analysis. The authors also EGFR regulates osteoblast differentiation 1105 thank Reinhold Erben for helpful discussions. This work was 12. Yarden Y. The EGFR family and its ligands in human cancer. supported by the Austrian Science Fund (FWF) (DK W1212), the signalling mechanisms and therapeutic opportunities. Eur J Can- FWF-grant I764-B13, the Deutsche Forschungsgemeinschaft (CRC cer. 2001;37:S3–8. 1149, Priority Program Immunobone Tu220/6 to J.T.) and “Fonds der 13. Rauch A, Seitz S, Baschant U, Schilling AF, Illing A, Stride B, Stadt Wien für innovative interdisziplinäre Krebsforschung” et al. Glucocorticoids suppress bone formation by attenuating (AP00288OFF). M.S. is supported by an ERC-Advanced grant osteoblast differentiation via the monomeric glucocorticoid (ERC-2015-AdG TNT-Tumors 694883). receptor. Cell Metab. 2010;11:517–31. 14. Guntur AR, Rosen CJ. IGF-1 regulation of key signaling path- ways in bone. Bone Rep. 2013;2:437. Compliance with ethical standards 15. Mohseni-Zadeh S, Binoux M. Insulin-like growth factor (IGF) binding protein-3 interacts with the type 1 IGF receptor, Conflict of interest The authors declare that they have no conflict of reducing the affinity of the receptor for its ligand: an alternative interest. mechanism in the regulation of IGF action. Endocrinology. 1997;138:5645–48. 16. Ricort JM, Binoux M. Insulin-like growth factor (IGF) Open Access This article is licensed under a Creative Commons binding protein-3 inhibits type 1 IGF receptor activation Attribution 4.0 International License, which permits use, sharing, independently of its IGF binding affinity. Endocrinology. adaptation, distribution and reproduction in any medium or format, as 2001;142:108–13. long as you give appropriate credit to the original author(s) and the 17. Takaoka M, Harada H, Andl CD, Oyama K, Naomoto Y, source, provide a link to the Creative Commons license, and indicate if Dempsey KL, et al. Epidermal growth factor receptor regulates changes were made. The images or other third party material in this aberrant expression of insulin-like growth factor-binding protein article are included in the article’s Creative Commons license, unless 3. Cancer Res. 2004;64:7711–23. indicated otherwise in a credit line to the material. If material is not 18. Liu KJ, Arron JR, Stankunas K, Crabtree GR, Longaker MT. included in the article’s Creative Commons license and your intended Chemical rescue of cleft palate and midline defects in conditional use is not permitted by statutory regulation or exceeds the permitted GSK-3beta mice. Nature. 2007;446:79–82. use, you will need to obtain permission directly from the copyright 19. Zhu J, Shimizu E, Zhang X, Partridge NC, Qin L. EGFR signaling holder. To view a copy of this license, visit http://creativecommons. suppresses osteoblast differentiation and inhibits expression of org/licenses/by/4.0/. master osteoblastic transcription factors Runx2 and Osterix. J Cell Biochem. 2011;112:1749–60. References 20. Jones HE, Gee JM, Hutcheson IR, Knowlden JM, Barrow D, Nicholson RI. Growth factor receptor interplay and resistance in 1. Wagner EF, Karsenty G. Genetic control of skeletal development. cancer. Endocr Relat Cancer. 2006;13:S45–51. Curr Opin Genet Dev. 2001;11:527–32. 21. Chakravarti A, Loeffler JS, Dyson NJ. Insulin-like growth factor 2. Cohen MM Jr.. The new bone biology: pathologic, molecular, and receptor I mediates resistance to anti-epidermal growth factor clinical correlates. Am J Med Genet A. 2006;140:2646–706. receptor therapy in primary human glioblastoma cells through 3. Sibilia M, Kroismayr R, Lichtenberger BM, Natarajan A, continued activation of phosphoinositide 3-kinase signaling. Hecking M, Holcmann M. The epidermal growth factor Cancer Res. 2002;62:200–7. receptor: from development to tumorigenesis. Differentiation. 22. Guix M, Faber AC, Wang SE, Olivares MG, Song Y, Qu S, et al. 2007;75:770–87. Acquired resistance to EGFR tyrosine kinase inhibitors in cancer 4. Sibilia M, Wagner EF. Strain-dependent epithelial defects in mice cells is mediated by loss of IGF-binding proteins. J Clin Invest. lacking the EGF receptor. Science. 1995;269:234–38. 2008;118:2609–19. 5. Threadgill DW, Dlugosz AA, Hansen LA, Tennenbaum T, Lichti 23. Silha JV, Mishra S, Rosen CJ, Beamer WG, Turner RT, Powell U, Yee D, et al. Targeted disruption of mouse EGF receptor: DR, et al. Perturbations in bone formation and resorption in effect of genetic background on mutant phenotype. Science. insulin-like growth factor binding protein-3 transgenic mice. J 1995;269:230–34. Bone Miner Res. 2003;18:1834–41. 6. Sibilia M, Wagner B, Hoebertz A, Elliott C, Marino S, Jochum W, 24. Yakar S, Rosen CJ, Bouxsein ML, Sun H, Mejia W, Kawashima et al. Mice humanised for the EGF receptor display hypomorphic Y, et al. Serum complexes of insulin-like growth factor-1 mod- phenotypes in skin, bone and heart. Development. ulate skeletal integrity and carbohydrate metabolism. FASEB J. 2003;130:4515–25. 2009;23:709–19. 7. Wang K, Yamamoto H, Chin JR, Werb Z, Vu TH. Epidermal 25. Kim YC, Guan KL. mTOR: a pharmacologic target for autophagy growth factor receptor-deficient mice have delayed primary regulation. J Clin Invest. 2015;125:25–32. endochondral ossification because of defective osteoclast recruit- 26. Yang GE, Duan X, Lin D, Li T, Luo D, Wang L, et al. ment. J Biol Chem. 2004;279:53848–56. Rapamycin-induced autophagy activity promotes bone fracture 8. Zhang X, Tamasi J, Lu X, Zhu J, Chen H, Tian X, et al. Epidermal healing in rats. Exp Ther Med. 2015;10:1327–33. growth factor receptor plays an anabolic role in bone metabolism 27. Zhang Y, Vasheghani F, Li YH, Blati M, Simeone K, Fahmi H, in vivo. J Bone Mineral Res. 2011;26:1022–34. et al. Cartilage-specific deletion of mTOR upregulates autophagy 9. Hall KC, Hill D, Otero M, Plumb DA, Froemel D, Dragomir CL, and protects mice from osteoarthritis. Ann Rheum Dis. et al. ADAM17 controls endochondral ossification by regulating 2015;74:1432–40. terminal differentiation of chondrocytes. Mol Cell Biol. 28. Carames B, Hasegawa A, Taniguchi N, Miyaki S, Blanco FJ, Lotz 2013;33:3077–90. M. Autophagy activation by rapamycin reduces severity of 10. Chandra A, Lan S, Zhu J, Siclari VA, Qin L. Epidermal growth experimental osteoarthritis. Ann Rheum Dis. 2012;71:575–81. factor receptor (EGFR) signaling promotes proliferation and sur- 29. Zhang X, Zhu J, Liu F, Li Y, Chandra A, Levin LS, et al. Reduced vival in osteoprogenitors by increasing early growth response 2 EGFR signaling enhances cartilage destruction in a mouse (EGR2) expression. J Biol Chem. 2013;288:20488–98. osteoarthritis model. Bone Res. 2014;2:14015. 11. Karsenty G, Wagner EF. Reaching a genetic and molecular 30. Jia H, Ma X, Tong W, Doyran B, Sun Z, Wang L, et al. EGFR understanding of skeletal development. Dev Cell. 2002;2:389–406. signaling is critical for maintaining the superficial layer of 1106 M. Linder et al. articular cartilage and preventing osteoarthritis initiation. Proc 35. Clausen BE, Burkhardt C, Reith W, Renkawitz R, Forster I. Natl Acad Sci USA. 2016;113:14360–65. Conditional gene targeting in macrophages and granulocytes using 31. Zhang M, Xuan S, Bouxsein ML, von Stechow D, Akeno N, LysMcre mice. Transgenic Res. 1999;8:265–77. Faugere MC, et al. Osteoblast-specific knockout of the insulin-like 36. Wallin J, Wilting J, Koseki H, Fritsch R, Christ B, Balling R. The growth factor (IGF) receptor gene reveals an essential role of IGF role of Pax-1 in axial skeleton development. Development. signaling in bone matrix mineralization. J Biol Chem. 2002;277: 1994;120:1109–21. 44005–12. 37. Parfitt AM, Drezner MK, Glorieux FH, Kanis JA, Malluche H, 32. Xian L, Wu X, Pang L, Lou M, Rosen CJ, Qiu T, et al. Matrix Meunier PJ, et al. Bone histomorphometry: standardization of IGF-1 maintains bone mass by activation of mTOR in mesench- nomenclature, symbols, and units. Report of the ASBMR Histo- ymal stem cells. Nat Med. 2012;18:1095–1. morphometry Nomenclature Committee. J Bone Miner Res. 33. Riddle RC, Frey JL, Tomlinson RE, Ferron M, Li Y, Digirolamo 1987;2:595–10. DJ, et al. Tsc2 is a molecular checkpoint controlling osteoblast 38. Jochum W, David JP, Elliott C, Wutz A, Plenk H Jr., Matsuo K, development and glucose homeostasis. Mol Cell Biol. 2014; et al. Increased bone formation and osteosclerosis in mice 34:1850–62. overexpressing the transcription factor Fra-1. Nat Med. 34. Lanaya H, Natarajan A, Komposch K, Li L, Amberg N, Chen L, 2000;6:980–84. et al. EGFR has a tumour-promoting role in liver macrophages 39. Sibilia M, Fleischmann A, Behrens A, Stingl L, Carroll J, Watt FM, during hepatocellular carcinoma formation. Nat Cell Biol. et al. The EGF receptor provides an essential survival signal for 2014;16:972–81. 971–977 SOS-dependent skin tumor development. Cell. 2000;102:211–20.

Journal

Cell Death & DifferentiationSpringer Journals

Published: Feb 14, 2018

References

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off