Hedgehog signaling activates a positive feedback mechanism involving insulin-like growth factors to induce osteoblast differentiation Yu Shia, Jianquan Chena,1, Courtney M. Karnera, and Fanxin Longa,b,c,2 Departments of aOrthopaedic Surgery, bMedicine, and cDevelopmental Biology, Washington University School of Medicine, St. Louis, MO 63110
Hedgehog (Hh) signaling is essential for osteoblast differentiation in the endochondral skeleton during embryogenesis. However, the molecular mechanism underlying the osteoblastogenic role of Hh is not completely understood. Here, we report that Hh markedly induces the expression of insulin-like growth factor 2 (Igf2) that activates the mTORC2-Akt signaling cascade during osteoblast differentiation. Igf2-Akt signaling, in turn, stabilizes full-length Gli2 through Serine 230, thus enhancing the output of transcriptional activation by Hh. Importantly, genetic deletion of the Igf signaling receptor Igf1r specifically in Hh-responding cells diminishes bone formation in the mouse embryo. Thus, Hh engages Igf signaling in a positive feedback mechanism to activate the osteogenic program. Hedgehog
| Igf | Gli2 | Akt | osteoblast
T
he Hedgehog (Hh) proteins are secreted signaling molecules that regulate many developmental processes in animals ranging from Drosophila to humans (1, 2). In mammals, three members of the Hh family, i.e., Indian hedgehog (Ihh), Sonic hedgehog (Shh), and Desert hedgehog (Dhh), are expressed in diverse tissues. Upon binding to the cell-surface receptor Patched 1 (Ptch1), Hh triggers the relocation of the seven-pass transmembrane protein Smoothened (Smo) to the primary cilium, resulting in activation of the downstream events (3). Hh signaling can also be activated artificially by small molecules such as purmorphamine that directly targets Smo (4, 5). A major output of Hh signaling is the induction of target genes through the Gli family of transcription factors. Of the three family members, Gli2 and Gli3 are the primary effectors for Hh-induced transcription and, thus, critical for embryogenesis (6–9). Gli1, however, itself a target of Gli2 and Gli3, is dispensable in the mouse embryo (7, 10). Biochemically, Gli2 is predominantly present in the fulllength form that becomes a transcriptional activator in response to Hh, whereas Gli3 exists preferentially as a C-terminally truncated transcriptional repressor whose formation is inhibited by Hh (11–13). Overall, activation of Gli2 together with derepression of Gli3 constitutes the primary mechanism for Hh-induced gene activation, whereas Gli1 plays a secondary role in amplifying the transcriptional response. Indian hedgehog (Ihh) is indispensable for osteoblast development in the endochondral skeleton of the mouse embryo (14). Genetic studies of Smo have demonstrated that Ihh input is directly required for osteoblast differentiation from the perichondrial progenitors in the long bone (15). Furthermore, forced activation of Gli2 together with Gli3 deletion is sufficient to replace Ihh in inducing osteoblast formation in the mouse embryo, indicating that Gli2 activation and Gli3 derepression are primarily responsible for Ihh-induced osteoblastogenesis (16). The osteogenic effect of Gli2 and Gli3 is likely amplified by Gli1, which has been shown to induce expression of osteoblast marker genes Alpl and Ibsp (17). However, how Ihh-Gli signaling activates the entire osteogenic program remains unresolved. www.pnas.org/cgi/doi/10.1073/pnas.1502301112
Insulin-like growth factors (Igfs) have been implicated in bone formation (18). Igf1 and Igf2 both signal through Igf1 receptor (Igf1r) to activate PI3K/Akt/mTORC1 and Ras/Raf/Erk pathways (19). Igf signaling also activates mTORC2 via a mechanism that is not well understood, but requires PI3K and inversely correlates with mTORC1 activity (20, 21). The activity of Igfs is modulated by the Igf-binding proteins Igfbp1–6 in complex ways involving both inhibitory and stimulatory effects (22). Igf1 knockout mice exhibit defects in bone formation, whereas overexpression of Igf1 in osteoblasts enhanced their activity (23, 24). Similarly, deletion of Igf1r in the osteoblast lineage impaired bone formation (25, 26). Mechanistically, Igf signaling has been proposed to stimulate Runx2 activity downstream of either Akt or Erk (27, 28). We have shown that Igf and Ihh signaling control cartilage development largely independently, but it is not known whether the two signals intersect during osteoblast differentiation (29). Here, we discover that Hh-Gli2 signaling markedly increases Igf2 mRNA and activates the Igf-mTORC2-Akt signaling cascade. Akt, in turn, stabilizes full-length Gli2 via direct phosphorylation at Ser230, thus augmenting Hh-induced transcriptional activation. These findings identify a Hh-Igf positive feedback mechanism that operates during osteoblast differentiation. Results Hh Induces Igf Signaling During Osteoblast Differentiation. To in-
vestigate the molecular mechanism responsible for Hh-induced osteoblast differentiation, we first established an in vitro differentiation system where the murine bone marrow stromal-derived cell line M2-10B4 cells (hereafter M2 cells) was stimulated with purmorphamine (PM), a small molecule agonist of Hh signaling (5). The M2 cells were shown to undergo osteoblastogenesis in response to Hh (30). Through a series of dosage and time-course experiments, we found that 1 μM PM maximally induced the Significance Hedgehog (Hh) signaling is essential for embryonic bone formation in mammals. However, how Hh signaling executes the osteogenic function is not clear. We describe a positive feedback mechanism whereby Hh-Gli2 signaling induces expression of insulin-like growth factor 2 (Igf2), which signals via protein kinases mTORC2 and Akt to stabilize Gli2 and potentiate Hh induction of osteoblast differentiation. The Hh-Igf positive feedback loop may have general implications for Hh signaling in development and disease. Author contributions: F.L. designed research; Y.S. and J.C. performed research; Y.S., C.M.K., and F.L. analyzed data; and Y.S. and F.L. wrote the paper. The authors declare no conflict of interest. This article is a PNAS Direct Submission. 1
Present address: Department of Orthopaedic Surgery, Duke University School of Medicine, Durham, NC 27710.
2
To whom correspondence should be addressed. Email: [email protected].
This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10. 1073/pnas.1502301112/-/DCSupplemental.
PNAS Early Edition | 1 of 6
DEVELOPMENTAL BIOLOGY
Edited by Clifford J. Tabin, Harvard Medical School, Boston, MA, and approved February 27, 2015 (received for review February 10, 2015)
mRNA levels of Gli1 and Ptch1, both Hh transcriptional targets, at 48 or 72 h of treatment (SI Appendix, Fig. S1). Importantly, 1 μM PM activated expression of several common osteoblast markers including Sp7, Alpl, and Bglap at those time points (SI Appendix, Fig. S1). Staining for Alpl activity or matrix mineralization confirmed the induction of osteoblast differentiation by 1 μM PM (Fig. 1A). Knockdown of Gli2, the primary effector for Hh-mediated transcriptional activation, not only reduced Gli1 induction, but also notably diminished the expression of the osteoblast markers in response to PM (Fig. 1B). Thus, PM activates Hh-Gli signaling to induce osteoblast differentiation in M2 cells. To discover effector molecules downstream of Hh-Gli signaling, we performed RNA-seq (whole transcriptome shotgun sequencing) experiments to compare the transcriptome of M2 cells with or without PM stimulation for 24 or 72 h. By examining all genes expressed at >1 read per kilobase of transcript per million reads mapped in at least one condition, and affected by PM by >twofold, we identified Igf2 and Igfbp5 being induced at 24 h, and Igf1, Igf2, Igfbp6, and Igfbp7 increased at 72 h. RT-qPCR (reverse transcription-quantitative polymerase chain reaction) experiments confirmed that PM significantly increased Igf2 mRNA by 24 h and reached the maximal induction of >20 fold by 48 h. The induction of Igf1 mRNA was less pronounced, reaching only ∼twofold after 72 h (Fig. 2A). Igfbp5 showed a twofold induction by 6 h, but Igfbp6 and Igfbp7 were also notably increased by 48 h (>twofold) (SI Appendix, Fig. S2). Furthermore, knockdown of Gli2 abolished Igf2 induction by PM at 24 h (Fig. 2B). Thus, Hh signaling up-regulates the expression of Igf signaling components. To test whether Hh activates Igf signaling, we monitored the phosphorylation of insulin receptor substrate 1 (Irs1) at S632/635 (equivalent of S636/639 in human IRS1), which is stimulated by insulin or Igf signaling (31). In keeping with the maximal induction of Igf2 at 48 h, PM notably increased both total- and phospho-Irs1 at 48 and 72 h (Fig. 2C). Recombinant human N-Shh had a similar effect on the phosphorylation (SI Appendix, Fig. S3). Importantly, knockdown of Igf2, Igf1r, or Gli2 severely blunted the induction of P-Irs1 at 72 h by PM (Fig. 2 D and E). Taken
Fig. 1. Hh signaling induces osteoblast differentiation through Gli2 in M2 cells. (A) Alkaline phosphatase activity (ALP) in M2 cells treated with 1 μM PM or DMSO (Veh) for 72 h. Alizarin Red staining was performed after additional 8 d of incubation in mineralization medium. (B) Effects of shRNA knockdown of Gli2 (sh-Gli2) on mRNA levels of indicated genes as assayed by qPCR. Cells were infected with lentivirus expressing sh-Gli2 or sh-Lacz for 24 h before further incubation with or without 1 μM PM for 72 h. Ribosomal 18S rRNA was used as internal control for qPCR. *P < 0.05, n = 3.
2 of 6 | www.pnas.org/cgi/doi/10.1073/pnas.1502301112
Fig. 2. Hh induces Igf-mTORC2 signaling through Gli2. (A) qPCR analyses in M2 cells treated with 1 μM PM or DMSO (Veh) for 6–72 h. (B) Effect of Gli2 knockdown on Igf2 mRNA level in response to PM for 24 h as assayed by qPCR. Ribosomal 18S rRNA used as internal control for all qPCR. *P < 0.05, n = 3. (C) Western blot for phospho-Irs1(S636/639) (P-Irs1) after 1 μM PM treatment for 48 or 72 h. P-Irs1 is normalized to α-tubulin and fold changes were calculated from three independent experiments with SD indicated. (D and E) Effects of shRNA knockdown of Igf1r, Igf2 (D), or Gli2 (E) on P-Irs1 in response to PM treatment for 72 h. P-Irs1 is normalized to total Irs1, and fold changes were calculated from three independent experiments with SD indicated. (F and G) Effects of shRNA knockdown of Gli2 (F), Igf1r, or Igf2 (G) on P-Akt (S473) in response to PM treatment for 72 h. P-Akt (S473) is normalized to total Akt, and fold changes were calculated from three independent experiments with SD indicated. Sh-Lacz was used as control.
together, the data indicate that Hh activates Igf signaling through up-regulation of Igf signaling components. We next investigated the intracellular pathway downstream of Igf1r in response to Hh signaling. In time-course experiments, we found that PM consistently increased phosphorylation of Akt at S473 [P-Akt (S473)] by 48 h, indicating activation of mTORC2 (SI Appendix, Fig. S4). However, PM did not induce mTORC1 activity, because phosphorylation of S6 or Eif4ebp1 was not different from the control at any of the time points examined (SI Appendix, Fig. S4A). Likewise, N-Shh activated mTORC2 but not mTORC1 after 48 or 72 h of treatment (SI Appendix, Fig. S4B). Importantly, knockdown of Gli2, Igf1r, or Igf2 greatly suppressed the induction of mTORC2 activity by PM, as measured by P-Akt (S473) levels (Fig. 2 F and G). Thus, Hh signaling activates Igf-mTORC2 signaling. We then investigated whether Igf-mTORC2 signaling contributes to Hh-induced osteoblast differentiation in vitro. Knockdown of Igf2 abolished osteoblast differentiation in either M2 cells or primary bone marrow stromal cells (BMSCs) in response to PM, as indicated by the lack of Sp7, Alpl, and Bglap induction (Fig. 3A and SI Appendix, Fig. S5). Knockdown of Igf1r also suppressed PM-induced osteoblast differentiation, but knockdown Shi et al.
Deletion of Igf1r in Hh-Responsive Cells Impairs Bone Formation in Vivo. To investigate the contribution of Igf signaling to Hh-induced Fig. 3. Igf2-mTORC2-Akt signaling is required for Hh-induced osteoblastogenesis. (A) Effects of Igf2 knockdown on mRNA levels of indicated genes in response to PM for 72 h, as assayed by qPCR. (B and C) Effect of Rictor knockdown (B) or MK2206 (C) on P-Akt (S473) assayed by Western blot, and on mRNA levels of osteoblast markers assayed by qPCR. (D) Effects of MK2206 on ALP or Alizarin Red staining in response to PM. Ribosomal 18S rRNA used as internal control for all qPCR. *P < 0.05, n = 3.
osteoblast differentiation in vivo, we sought to disrupt Igf signaling in the Hh-responsive osteogenic cells in the mouse embryo by using the Gli-CreERT2 mouse line. To establish the effectiveness of strategy, we first crossed Gli-CreERT2 with the mT/mG reporter mouse and exposed the embryos to tamoxifen (TM) from embryonic day (E)13.5 until E16.5 when they were harvested. In the long bones of the Gli-CreERT2; mT/mG embryo, numerous green cells were observed within the osteogenic perichondrium and the cartilage, indicating that Gli-CreERT2 effectively targets the
of insulin receptor (Insr) had no effect (SI Appendix, Fig. S6). Furthermore, knockdown of the mTORC2-specific component Rictor greatly diminished PM-induced osteoblast differentiation (Fig. 3B). Pharmacological inhibition of all mTOR activity with Torin1 exerted even greater inhibition on differentiation (SI Appendix, Fig. S7). Finally, inhibition of the mTORC2 target Akt with MK2206 abolished PM-induced mRNA expression of Sp7, Alpl, and Bglap, as well as Alpl staining and matrix mineralization (Fig. 3 C and D). Thus, Hh signaling activates the Igf2-mTORC2Akt axis to induce osteoblast differentiation. Igf2 Synergizes with Hh Signaling via Akt-Dependent Gli2 Stabilization.
We next pursued how Igf2-mTORC2-Akt signaling mediates Hh-induced osteoblastogenesis. Igf2 by itself did not induce the expression of Sp7, Alpl, or Bglap in M2 cells, but synergistically increased the mRNA levels of these markers when used in combination of suboptimal levels (e.g., 0.5 μM) of PM (Fig. 4A). Interestingly, in such settings, Igf2 also boosted the expression levels of Hh target genes Gli1 and Ptch1 compared with PM alone, although Igf2 itself did not induce these genes (Fig. 4A). Conversely, knockdown of Igf2 or Igf1r reduced the mRNA levels of Gli1 and Ptch1 in response to PM (Fig. 4B and SI Appendix, Fig. S8A). Likewise, inhibition of mTORC2 with either Torin1 or by knocking down Rictor severely diminished the induction of Gli1 by PM (SI Appendix, Fig. S8 B and C). Finally, inhibition of Akt with MK2206 also blunted the Hh transcriptional response (Fig. 4C). Thus, Igf2mTORC2-Akt signaling stimulates osteoblast differentiation by augmenting the transcriptional response to Hh signaling. To determine how Igf2-mTORC2-Akt enhances transcriptional activation by Hh, we examined the potential effect of Igf signaling on Gli2, the primary effector for Hh-induced gene activation. To this end, we infected M2 cells with lentiviruses Shi et al.
Fig. 4. Igf2 synergizes with Hh signaling via Akt-dependent Gli2 stabilization. (A) Synergistic induction of genes by 0.5 μM PM or 100 ng/mL Igf2, as assayed by qPCR after 72 h of treatment. (B and C) Effects of Igf2 knockdown (B) or MK2206 (C) assayed by qPCR after 72 h of PM treatment. *P < 0.05, n = 3. (D) Effects of Igf2 and MK2206 on Flag-Gli2 turnover, as measured by Western blot. *P < 0.05, n = 3. (E) Effect of Igf2 on Flag-Gli2-S230A turnover. (F) Effect of Igf2 on endogenous Gli2 stability. All quantification is from three independent experiments. CHX, cyclohexamide.
PNAS Early Edition | 3 of 6
DEVELOPMENTAL BIOLOGY
expressing Flag-tagged full-length Gli2 (Flag-Gli2) in response to doxycycline (Dox), and then treated the cells with Igf2. Igf2 had no effect on Gli2 mRNA levels (including both endogenous Gli2 and exogenous Flag-Gli2), but notably increased the protein level of Flag-Gli2, indicating that Igf2 may suppress Gli2 turnover (SI Appendix, Fig. S9 A and B). In contrast, Igf2 did not affect the protein level of Flag-Gli3 in a similar experiment (SI Appendix, Fig. S9C). To determine the potential effect of Igf2 on the half-life of Gli2, we blocked de novo protein synthesis with cyclohexamide after the induction of Flag-Gli2 expression with Dox, and monitored the abundance of Flag-Gli2 as a function of time with or without Igf2 treatment. These experiments revealed that Igf2 prolonged the half-life of Flag-Gli2 from 24 h, and that the stabilizing effect of Igf2 was completely abolished by the Akt inhibitor MK2206 (Fig. 4D). Examination of the mouse Gli2 protein sequence revealed a consensus phosphorylation site for Akt, R225KRALS230 (32). Alanine substitution of S230 abolished the stabilizing effect of Igf2 on Gli2 (Fig. 4E). Finally, to confirm the effect of Igf2 on endogenous Gli2, we generated an antibody against a peptide within the C-terminal half of the mouse Gli2 (SI Appendix, Fig. S10) and found that Igf2 stabilized the endogenous full-length Gli2 protein (Fig. 4F). Thus, Igf2 markedly suppresses Gli2 turnover likely through direct phosphorylation by Akt.
cells known to respond to Ihh during long bone development (Fig. 5A). We next generated Gli-CreERT2; Igf1rf/f embryos (CKO) and induced Igf1r deletion with the same TM regimen. Immunofluorescence staining confirmed that Igf1r was notably reduced in both perichondrial cells and chondrocytes in the long bones of E16.5 embryos (SI Appendix, Fig. S11). Importantly, whole-mount skeletal staining revealed that the bone collars of all long bones in the CKO embryo were notably shorter than their normal counterparts, although the total length of each element appeared to be similar across genotypes (Fig. 5B). Quantification indicated that the relative length of bone collar versus the entire element was reduced by 10–20% in the CKO embryo, indicating that ossification was suppressed (Fig. 5C). The bone collar deficit persisted at E18.5 (SI Appendix, Fig. S12). Staining of the long bone sections with the von Kossa method identified a marked decrease in bone collar mineralization in the CKO embryo (Fig. 5D). Furthermore, in situ hybridization on sections showed that the mRNA levels of osteoblast markers Runx2, Sp7, Alpl, and the Hh target gene Ptch1 were diminished in CKO, although Ihh expression was relatively normal (Fig. 5E). However, chondrocyte proliferation or hypertrophy was relatively normal in E14.5 CKO embryos after 3 d of TM induction (SI Appendix, Fig. S13). Western blot analyses confirmed that in the limbs of the E14.5 CKO embryos, the levels of Igf1r, Gli2, and P-Akt (S473) were notably reduced compared with those in the
Fig. 5. Deletion of Igf1r in Hh-responsive cells impairs bone formation in vivo. (A) Detection of GFP by immunofluorescence on a tibial section from an E16.5 Gli1-CreERT2; mT/mG mouse embryo that received daily tamoxifen beginning at E13.5. H, hypertrophic chondrocytes; P, perichondrium; PH, prehypertrophic chondrocytes; S, primary spongiosa. (B and C) Skeletal staining of long bones and quantification of relative bone collar length in E16.5 wildtype (WT) versus Gli1-CreERT2; Igf1rf/f (CKO) littermate embryos. Vertical lines demarcate the ends of the bone collar; the horizontal lines denote the deficit in the mutant (CKO). *P < 0.05, n = 5. F, femur; H, humerus; R, radius; T, tibia; U, ulna. (D) Von Kossa staining on longitudinal tibial sections in WT versus CKO littermates at E16.5. Green boxed areas shown at a higher magnification in Insets. (E) In situ hybridization on tibial sections of E16.5 embryos. Signal is in red. (F and G) Western blot analyses of protein extracts from limbs of E14.5 WT versus CKO littermates with TM treatments beginning at E11.5. Each lane is from a separate embryo. Intensity normalization is against α-tubulin, except for P-Akt (S473) against total Akt.
4 of 6 | www.pnas.org/cgi/doi/10.1073/pnas.1502301112
wild-type littermates (Fig. 5 F and G). In contrast, the level of fulllength Gli3 was largely normal in the CKO samples (Fig. 5G). In summary, Igf signaling is necessary to potentiate osteoblastogenesis in Ihh-responsive cells during long bone development. Discussion We have uncovered a positive feedback mechanism between Hh and Igf signaling during osteoblast differentiation. In particular, Hh signaling induces transcription of multiple members of the Igf system, most notably Igf2, resulting in activation of the mTORC2-Akt signaling axis. Akt, in turn, stabilizes full-length Gli2, likely through phosphorylation of S230, thereby amplifying the transcriptional output of Hh signaling. Supporting the physiologic relevance of the Hh-Igf positive feedback loop, cellautonomous removal of Igf signaling in Ihh-responding cells diminishes osteoblast differentiation in the mouse embryo. This finding not only sheds light on how Hh signaling induces osteoblastogenesis, but also adds another dimension to the role of Igf signaling in bone formation. The mechanism whereby Hh activates osteoblast differentiation remains to be fully elucidated. Our data have demonstrated an important role for Igf signaling in Hh-induced osteoblastogenesis, but it is also clear that Igf signaling cannot replace Hh to initiate the osteogenic program. In contrast to Ihh deletion that abolishes bone formation altogether in the endochondral skeleton, removal of Igf1r impairs but does not prevent the process (33). Data from the present study support the model wherein Igf signaling amplifies Gli-mediated transcriptional induction to promote osteoblast differentiation. This model predicts that certain osteogenic genes are induced not immediately by Hh but only after subsequent activation of Igf signaling, a process occurring between 24 and 48 h of PM treatment in M2 cells. Indeed, our RNA-seq experiments identified several transcription factors being induced after 72 but not 24 h of PM treatment; these factors include Dlx5, Dlx6, Sp7, Msx1, and Ets2, all of which have been implicated in osteoblast differentiation (34–37). Future studies are necessary to determine the contributions of those factors to osteoblast differentiation in response to Hh signaling. The mechanism whereby Hh signaling induces Igf2 expression is not clear at present. We show that Gli2 is necessary for Hh to induce Igf2 mRNA. This result, along with the previous report that Gli1 overexpression induces Igf2 in C3H10T1/2 cells, raises the possibility that Igf2 might be a direct target of Gli activators (38). However, unlike the known Gli target genes such as Gli1 that show activation by 6 h of Hh stimulation, Igf2 is not induced until after 24 h (SI Appendix, Fig. S1). Moreover, although a putative Gli binding site is identified in the 3′UTR of the Igf2 gene, chromatin immunoprecipitation experiments with cells overexpressing Flag-tagged Gli2 have failed to reveal Gli2 binding to the site. We have considered the possibility that other secreted factors such as BMP, Wnt, or PTHrP may mediate the induction of Igf2 by Hh, but the RNA-seq experiment did not detect any significant increase in their expression after 24 h of PM stimulation. Thus, the induction of Igf2 by Gli2 appears to be through an indirect and yet unknown mechanism. We have identified S230 of Gli2 as a critical mediator for Igf2 to potentiate Hh signaling. In our model, upon activation by Igf2, Akt phosphorylates and stabilizes full-length Gli2 that requires Hh signaling for activation. Previous work has identified clusters of adjacent serine residues whose sequential phosphorylation by PKA and Gsk3 controls Gli2 turnover (39). That study also reported that alanine substitutions of the PKA, but not Gsk3 or Akt phosphorylation sites, abolished the inhibitory effect of dominant negative Akt on Gli2-mediated transcription, although it did not examine the stability of Gli2. It is possible that Aktmediated S230 phosphorylation precludes PKA-mediated phosphorylation, thus stabilizing Gli2. Moreover, the full effect of Akt on Gli2 stability may involve both the direct regulation as reported Shi et al.
Cell Culture. M2-10B4 cells (CRL-1972; ATCC) were maintained in RPMI (catalog no. 11875-085, Gibco) with 10% FBS (Gibco) as per instructions. Unless otherwise indicated, M2 cells were seeded at the density of 1 × 104 cells per cm2, 48 h before experiments. Primary cultures of BMSCs were harvested from the femur and the tibia of 6-wk-old mice and cultured in α-MEM containing 10% FBS (Gibco), with a change of medium at day 3. Cells were dissociated at day 7 and reseeded at 1 × 105 cells per cm2 for qPCR, ALP staining or mineralization assays at appropriate time points. To express Flag-tagged Gli2 proteins in M2 cells, the mouse Gli2 full-length cDNA, either wild type or the S230A mutant form, was cloned into the Fuw-tetO vector (plasmid 20723; Addgene) and retrovirus was produced according to standard protocol. A retrovirus producing Flag-Gli3 (full-length) was made in the same way. Then M2-10B4 cells were infected with FuW-rtTA (plasmid 20342; Addgene) virus, together with Fuw-tetO-FlagGli2 or Fuw-tetO-FlagGli3 virus. Gli2 expression was induced with addition of doxycycline (100 nM) to the media. The recombinant human Shh (464-SH-025) and Igf2 (792-MG-050) were purchased from R&D Systems. Purmorphamine (540223; Millipore), MK2206 (S1078; Selleckchem), Torin1 (4247; Tocris Biosciences) were purchased. Gli2 shRNA plasmid (Mission shRNA SHCLND-NM_001081125) was purchased from Sigma. Lentiviral shRNA targeting vectors were purchased from Washington University shRNA consortium. Targeted sequences are listed in SI Appendix, Table S1.
Materials and Methods
Western Blot Analyses. Protein extracts from cells were prepared in radioimmunoprecipitation assay buffer containing phosphatase and proteinase inhibitors (Roche). Total protein content was determined by the bicinchoninic acid method (Pierce). Membranes were imaged with Molecular Imager ChemiDoc XRS+ System (Bio-Rad). Quantification of Western blots was performed with ImageLab software. The monoclonal antibodies against phospho-IRS1(S636/S639) (2388), total-IRS1 (3407), phospho-AKT(S473) (9271), total-AKT (9272), phospho-S6 (2215), total-S6 (2217), phosphoEif4ebp1 (T37/46) (2855), total-Eif4ebp1 (9644), Rictor (2140) and β-actin (4970) were purchased from Cell Signaling Technology. The antibody against Gli3 (ab6050) was from Abcam, and the M2 antibody (catalog no. F1804) against Flag was from Sigma. The monoclonal antibody against α-tubulin (sc8035) and the polyclonal antibody against Igf1r (sc-713) were obtained from Santa Cruz Biotechnology. HRP-conjugated anti-rabbit (NA934V) or antimouse secondary antibody (sc-2005) was purchased from GE Healthcare or Santa Cruz Biotechnology, respectively. The Alexa Fluor 488 donkey antirabbit IgG (H+L) antibody was obtained from Invitrogen.
High-Throughput RNA Sequencing. M2 cells were treated with or without 1 μM PM for 24 or 72 h before total RNA was isolated by using the RNeasy kit with on-column DNase treatment (Qiagen). Poly-A RNA was isolated from 10 μg of total RNA by using Oligo-dT beads, followed by fragmentation and reverse transcription. The double-stranded cDNA was used to generate a library with sequencing adapters added to the ends of DNA fragments, and the resulting library was sequenced by using the Illumina HiSeq-2000 as single reads extending 42 bases (Genome Technology Access Center, Washington University School of Medicine). The raw data were demultiplexed and aligned to the reference genome (mm9) by using TopHat. Partek Genomic Suite was used to assemble transcripts and analyze expression. Mouse Studies. Igf1rf/f, mT/mG, and Gli1-CreERT2 mice have been reported (26, 47, 48). The Animal Studies Committee at Washington University in St. Louis reviewed and approved all mouse procedures used in this study. The Gli1-CreERT2;Igf1rf/f or Gli1-CreERT2;mT/mG pregnant female mice were administered tamoxifen dissolved in corn oil (4 mg/30 g of body weight) by oral gavage daily for durations as indicated. Whole-mount skeletal staining was performed with a protocol based on McLeod (49). To assess skeletal phenotypes on sections, embryonic limbs were harvested in PBS and fixed in 10% (vol/vol) buffered formalin overnight at room temperature. The fixed limbs were then processed and embedded in paraffin before sectioning at 6 μm. Limbs from E18.5 embryos were decalcified in 14% (wt/vol) EDTA/PBS (pH 7.4) for 48 h after fixation and before processing. In situ hybridization was performed by using 35S-labeled riboprobes as described (15, 42, 50, 51). The chondrocyte proliferation rate in embryos was assessed by immunodetection of BrdU incorporation on limb sections of the embryos (Invitrogen) (52). The percentage of labeled cells was calculated from medial sections of three pairs of wild-type versus mutant littermates. Igf1r immunofluorescence staining was performed on paraffin sections after antigen retrieval with citrate buffer and blocking with 5% (vol/vol) normal goat serum in Tris buffered saline with Tween. Sections were incubated with Igf1r antibody (sc-713; Santa Cruz Biotechnology) at 1:100 dilution overnight at 4 °C, and then Alexa Fluor 488 donkey anti-rabbit IgG (H+L) antibody (A11371; Invitrogen) at 1:300 dilution at room temperature for 1 h.
1. Ingham PW, McMahon AP (2001) Hedgehog signaling in animal development: Paradigms and principles. Genes Dev 15(23):3059–3087. 2. Huangfu D, Anderson KV (2006) Signaling from Smo to Ci/Gli: Conservation and divergence of Hedgehog pathways from Drosophila to vertebrates. Development 133(1):3–14. 3. Goetz SC, Ocbina PJ, Anderson KV (2009) The primary cilium as a Hedgehog signal transduction machine. Methods Cell Biol 94:199–222.
Shi et al.
Osteoblast Differentiation Assays. Alkaline phosphatase staining was as described (53). When inhibitors were used, cells were pretreated with the inhibitor in normal growth medium for 1 h. When shRNA lentiviruses were used, cells were infected with respective shRNA lentivirus for 24 h before subsequent treatments. For mineralization assays, confluent cells were incubated in the presence of 50 mg/mL ascorbic acid and 50 mM β-glycerophosphate for 8 d and then stained with Alizarin Red S.
Quantitative PCR. RNA was isolated from whole cells with QIAGEN RNeasy kit (74104) and was transcribed into cDNA by using iScript cDNA synthesis kit (Bio-Rad). Fast-start SYBR Green (Bio-Rad) and 0.1 mM primers were used in each reaction. Ribosomal 18S RNA was used for normalization. PCR was performed with Step-One machine (ABI). The primer sequences are provided in SI Appendix, Table S2. Production of Gli2 Antibody. A monoclonal antibody specific for full-length Gli2 was generated against a peptide within the C-terminal half of mouse Gli2 according to proprietary techniques (Abmart). The purified antibody was validated by Western blot analyses for detection of overexpressed Gli2, and for loss of endogenous Gli2 following shRNA knockdown. Statistical Analyses. All quantitative data are presented as mean ± SD with a minimum of three independent samples. Statistical significance is determined by Student’s t test. ACKNOWLEDGMENTS. This work is supported by NIH Grant R01 DK065789 (to F.L.).
4. Wu X, Walker J, Zhang J, Ding S, Schultz PG (2004) Purmorphamine induces osteogenesis by activation of the hedgehog signaling pathway. Chem Biol 11(9):1229–1238. 5. Sinha S, Chen JK (2006) Purmorphamine activates the Hedgehog pathway by targeting Smoothened. Nat Chem Biol 2(1):29–30. 6. Sasaki H, Nishizaki Y, Hui C, Nakafuku M, Kondoh H (1999) Regulation of Gli2 and Gli3 activities by an amino-terminal repression domain: Implication of Gli2 and Gli3 as primary mediators of Shh signaling. Development 126(17):3915–3924.
PNAS Early Edition | 5 of 6
DEVELOPMENTAL BIOLOGY
here and indirect control via phosphorylation and inactivation of Gsk3 (40). Future studies are necessary to determine whether and how Akt-mediated phosphorylation of Gli2 affects its phosphorylation by PKA or Gsk3. It is worth noting that deletion of Igf1r in Hh-responsive cells did not grossly affect cartilage development. This observation is consistent with our previous finding that Ihh and Igf control the linear growth of the skeleton largely independently (29). In addition, previous studies have demonstrated that Ihh controls cartilage development chiefly through derepression of Gli3, but induces osteoblastogenesis via both Gli3 derepression and Gli2 activation (16, 41, 42). Here, we show that Igf signaling stabilizes full-length Gli2 but not Gli3. Thus, the relative contribution of Gli2 versus Gli3 likely determines the functional relevance of the Hh-Igf feedback mechanism in specific physiological contexts. Beyond osteoblastogenesis, Hh and Igf may functionally interact in other processes. Most notably, in the Ptch1+/− mouse model for medulloblastoma and rhabdomyosarcoma, Igf2 was upregulated and necessary for tumor formation (43). Similarly, cooverexpression of Igf2 or activated Akt with Shh synergistically induces medulloblastoma in mice (44). Mechanistically, Igf-PI3K and Hh signaling appeared to converge on Nmyc1 to promote proliferation of cerebellar neuronal precursors (45). Moreover, Igf synergized with Shh in promoting somite myogenesis in explant cultures (46). Overall, Hh and Igf signaling exhibit synergistic interactions both in normal development and during pathogenesis.
7. Bai CB, Auerbach W, Lee JS, Stephen D, Joyner AL (2002) Gli2, but not Gli1, is required for initial Shh signaling and ectopic activation of the Shh pathway. Development 129(20):4753–4761. 8. Mo R, et al. (1997) Specific and redundant functions of Gli2 and Gli3 zinc finger genes in skeletal patterning and development. Development 124(1):113–123. 9. Litingtung Y, Dahn RD, Li Y, Fallon JF, Chiang C (2002) Shh and Gli3 are dispensable for limb skeleton formation but regulate digit number and identity. Nature 418(6901):979–983. 10. Park HL, et al. (2000) Mouse Gli1 mutants are viable but have defects in SHH signaling in combination with a Gli2 mutation. Development 127(8):1593–1605. 11. Wang B, Fallon JF, Beachy PA (2000) Hedgehog-regulated processing of Gli3 produces an anterior/posterior repressor gradient in the developing vertebrate limb. Cell 100(4):423–434. 12. Pan Y, Bai CB, Joyner AL, Wang B (2006) Sonic hedgehog signaling regulates Gli2 transcriptional activity by suppressing its processing and degradation. Mol Cell Biol 26(9):3365–3377. 13. Pan Y, Wang B (2007) A novel protein-processing domain in Gli2 and Gli3 differentially blocks complete protein degradation by the proteasome. J Biol Chem 282(15): 10846–10852. 14. St-Jacques B, Hammerschmidt M, McMahon AP (1999) Indian hedgehog signaling regulates proliferation and differentiation of chondrocytes and is essential for bone formation. Genes Dev 13(16):2072–2086. 15. Long F, et al. (2004) Ihh signaling is directly required for the osteoblast lineage in the endochondral skeleton. Development 131(6):1309–1318. 16. Joeng KS, Long F (2009) The Gli2 transcriptional activator is a crucial effector for Ihh signaling in osteoblast development and cartilage vascularization. Development 136(24):4177–4185. 17. Hojo H, et al. (2012) Gli1 protein participates in Hedgehog-mediated specification of osteoblast lineage during endochondral ossification. J Biol Chem 287(21):17860–17869. 18. Kawai M, Rosen CJ (2012) The insulin-like growth factor system in bone: Basic and clinical implications. Endocrinol Metab Clin North Am 41(2):323–333, vi. 19. Pollak M (2008) Insulin and insulin-like growth factor signalling in neoplasia. Nat Rev Cancer 8(12):915–928. 20. Liu P, et al. (2013) Sin1 phosphorylation impairs mTORC2 complex integrity and inhibits downstream Akt signalling to suppress tumorigenesis. Nat Cell Biol 15(11): 1340–1350. 21. Dalle Pezze P, et al. (2012) A dynamic network model of mTOR signaling reveals TSCindependent mTORC2 regulation. Sci Signal 5(217):ra25. 22. Rechler MM, Clemmons DR (1998) Regulatory actions of insulin-like growth factorbinding proteins. Trends Endocrinol Metab 9(5):176–183. 23. Bikle D, et al. (2001) The skeletal structure of insulin-like growth factor I-deficient mice. J Bone Miner Res 16(12):2320–2329. 24. Zhao G, et al. (2000) Targeted overexpression of insulin-like growth factor I to osteoblasts of transgenic mice: Increased trabecular bone volume without increased osteoblast proliferation. Endocrinology 141(7):2674–2682. 25. Xian L, et al. (2012) Matrix IGF-1 maintains bone mass by activation of mTOR in mesenchymal stem cells. Nat Med 18:1095–1101. 26. Zhang M, et al. (2002) Osteoblast-specific knockout of the insulin-like growth factor (IGF) receptor gene reveals an essential role of IGF signaling in bone matrix mineralization. J Biol Chem 277(46):44005–44012. 27. Yang S, et al. (2011) Foxo1 mediates insulin-like growth factor 1 (IGF1)/insulin regulation of osteocalcin expression by antagonizing Runx2 in osteoblasts. J Biol Chem 286(21):19149–19158. 28. Qiao M, Shapiro P, Kumar R, Passaniti A (2004) Insulin-like growth factor-1 regulates endogenous RUNX2 activity in endothelial cells through a phosphatidylinositol 3-kinase/ERK-dependent and Akt-independent signaling pathway. J Biol Chem 279(41): 42709–42718. 29. Long F, Joeng KS, Xuan S, Efstratiadis A, McMahon AP (2006) Independent regulation of skeletal growth by Ihh and IGF signaling. Dev Biol 298(1):327–333. 30. Dwyer JR, et al. (2007) Oxysterols are novel activators of the hedgehog signaling pathway in pluripotent mesenchymal cells. J Biol Chem 282(12):8959–8968.
6 of 6 | www.pnas.org/cgi/doi/10.1073/pnas.1502301112
31. Copps KD, White MF (2012) Regulation of insulin sensitivity by serine/threonine phosphorylation of insulin receptor substrate proteins IRS1 and IRS2. Diabetologia 55(10):2565–2582. 32. Obata T, et al. (2000) Peptide and protein library screening defines optimal substrate motifs for AKT/PKB. J Biol Chem 275(46):36108–36115. 33. Liu JP, Baker J, Perkins AS, Robertson EJ, Efstratiadis A (1993) Mice carrying null mutations of the genes encoding insulin-like growth factor I (Igf-1) and type 1 IGF receptor (Igf1r). Cell 75(1):59–72. 34. Robledo RF, Rajan L, Li X, Lufkin T (2002) The Dlx5 and Dlx6 homeobox genes are essential for craniofacial, axial, and appendicular skeletal development. Genes Dev 16(9):1089–1101. 35. Nakashima K, et al. (2002) The novel zinc finger-containing transcription factor osterix is required for osteoblast differentiation and bone formation. Cell 108(1):17–29. 36. Zhang Z, et al. (2003) Msx1/Bmp4 genetic pathway regulates mammalian alveolar bone formation via induction of Dlx5 and Cbfa1. Mech Dev 120(12):1469–1479. 37. Raouf A, Seth A (2000) Ets transcription factors and targets in osteogenesis. Oncogene 19(55):6455–6463. 38. Ingram WJ, Wicking CA, Grimmond SM, Forrest AR, Wainwright BJ (2002) Novel genes regulated by Sonic Hedgehog in pluripotent mesenchymal cells. Oncogene 21(53):8196–8205. 39. Riobó NA, Lu K, Ai X, Haines GM, Emerson CP, Jr (2006) Phosphoinositide 3-kinase and Akt are essential for Sonic Hedgehog signaling. Proc Natl Acad Sci USA 103(12): 4505–4510. 40. Cross DA, Alessi DR, Cohen P, Andjelkovich M, Hemmings BA (1995) Inhibition of glycogen synthase kinase-3 by insulin mediated by protein kinase B. Nature 378(6559):785–789. 41. Koziel L, Wuelling M, Schneider S, Vortkamp A (2005) Gli3 acts as a repressor downstream of Ihh in regulating two distinct steps of chondrocyte differentiation. Development 132(23):5249–5260. 42. Hilton MJ, Tu X, Cook J, Hu H, Long F (2005) Ihh controls cartilage development by antagonizing Gli3, but requires additional effectors to regulate osteoblast and vascular development. Development 132(19):4339–4351. 43. Hahn H, et al. (2000) Patched target Igf2 is indispensable for the formation of medulloblastoma and rhabdomyosarcoma. J Biol Chem 275(37):28341–28344. 44. Rao G, et al. (2004) Sonic hedgehog and insulin-like growth factor signaling synergize to induce medulloblastoma formation from nestin-expressing neural progenitors in mice. Oncogene 23(36):6156–6162. 45. Kenney AM, Widlund HR, Rowitch DH (2004) Hedgehog and PI-3 kinase signaling converge on Nmyc1 to promote cell cycle progression in cerebellar neuronal precursors. Development 131(1):217–228. 46. Pirskanen A, Kiefer JC, Hauschka SD (2000) IGFs, insulin, Shh, bFGF, and TGF-beta1 interact synergistically to promote somite myogenesis in vitro. Dev Biol 224(2): 189–203. 47. Yu W, McDonnell K, Taketo MM, Bai CB (2008) Wnt signaling determines ventral spinal cord cell fates in a time-dependent manner. Development 135(22):3687–3696. 48. Muzumdar MD, Tasic B, Miyamichi K, Li L, Luo L (2007) A global double-fluorescent Cre reporter mouse. Genesis 45(9):593–605. 49. McLeod MJ (1980) Differential staining of cartilage and bone in whole mouse fetuses by alcian blue and alizarin red S. Teratology 22(3):299–301. 50. Hu H, et al. (2005) Sequential roles of Hedgehog and Wnt signaling in osteoblast development. Development 132(1):49–60. 51. Long F, Zhang XM, Karp S, Yang Y, McMahon AP (2001) Genetic manipulation of hedgehog signaling in the endochondral skeleton reveals a direct role in the regulation of chondrocyte proliferation. Development 128(24):5099–5108. 52. Chen J, Holguin N, Shi Y, Silva MJ, Long F (2015) mTORC2 signaling promotes skeletal growth and bone formation in mice. J Bone Miner Res 30(2):369–378. 53. Shi Y, et al. (2012) Uniaxial mechanical tension promoted osteogenic differentiation of rat tendon-derived stem cells (rTDSCs) via the Wnt5a-RhoA pathway. J Cell Biochem 113(10):3133–3142.
Shi et al.
Hedgehog (Hh) signaling is essential for osteoblast differentiation in the endochondral skeleton during embryogenesis. However, the molecular mechanism underlying the osteoblastogenic role of Hh is not completely understood. Here, we report that Hh markedly induces the expression of insulin-like growth factor 2 (Igf2) that activates the mTORC2-Akt signaling cascade during osteoblast differentiation. Igf2-Akt signaling, in turn, stabilizes full-length Gli2 through Serine 230, thus enhancing the output of transcriptional activation by Hh. Importantly, genetic deletion of the Igf signaling receptor Igf1r specifically in Hh-responding cells diminishes bone formation in the mouse embryo. Thus, Hh engages Igf signaling in a positive feedback mechanism to activate the osteogenic program. Hedgehog
| Igf | Gli2 | Akt | osteoblast
T
he Hedgehog (Hh) proteins are secreted signaling molecules that regulate many developmental processes in animals ranging from Drosophila to humans (1, 2). In mammals, three members of the Hh family, i.e., Indian hedgehog (Ihh), Sonic hedgehog (Shh), and Desert hedgehog (Dhh), are expressed in diverse tissues. Upon binding to the cell-surface receptor Patched 1 (Ptch1), Hh triggers the relocation of the seven-pass transmembrane protein Smoothened (Smo) to the primary cilium, resulting in activation of the downstream events (3). Hh signaling can also be activated artificially by small molecules such as purmorphamine that directly targets Smo (4, 5). A major output of Hh signaling is the induction of target genes through the Gli family of transcription factors. Of the three family members, Gli2 and Gli3 are the primary effectors for Hh-induced transcription and, thus, critical for embryogenesis (6–9). Gli1, however, itself a target of Gli2 and Gli3, is dispensable in the mouse embryo (7, 10). Biochemically, Gli2 is predominantly present in the fulllength form that becomes a transcriptional activator in response to Hh, whereas Gli3 exists preferentially as a C-terminally truncated transcriptional repressor whose formation is inhibited by Hh (11–13). Overall, activation of Gli2 together with derepression of Gli3 constitutes the primary mechanism for Hh-induced gene activation, whereas Gli1 plays a secondary role in amplifying the transcriptional response. Indian hedgehog (Ihh) is indispensable for osteoblast development in the endochondral skeleton of the mouse embryo (14). Genetic studies of Smo have demonstrated that Ihh input is directly required for osteoblast differentiation from the perichondrial progenitors in the long bone (15). Furthermore, forced activation of Gli2 together with Gli3 deletion is sufficient to replace Ihh in inducing osteoblast formation in the mouse embryo, indicating that Gli2 activation and Gli3 derepression are primarily responsible for Ihh-induced osteoblastogenesis (16). The osteogenic effect of Gli2 and Gli3 is likely amplified by Gli1, which has been shown to induce expression of osteoblast marker genes Alpl and Ibsp (17). However, how Ihh-Gli signaling activates the entire osteogenic program remains unresolved. www.pnas.org/cgi/doi/10.1073/pnas.1502301112
Insulin-like growth factors (Igfs) have been implicated in bone formation (18). Igf1 and Igf2 both signal through Igf1 receptor (Igf1r) to activate PI3K/Akt/mTORC1 and Ras/Raf/Erk pathways (19). Igf signaling also activates mTORC2 via a mechanism that is not well understood, but requires PI3K and inversely correlates with mTORC1 activity (20, 21). The activity of Igfs is modulated by the Igf-binding proteins Igfbp1–6 in complex ways involving both inhibitory and stimulatory effects (22). Igf1 knockout mice exhibit defects in bone formation, whereas overexpression of Igf1 in osteoblasts enhanced their activity (23, 24). Similarly, deletion of Igf1r in the osteoblast lineage impaired bone formation (25, 26). Mechanistically, Igf signaling has been proposed to stimulate Runx2 activity downstream of either Akt or Erk (27, 28). We have shown that Igf and Ihh signaling control cartilage development largely independently, but it is not known whether the two signals intersect during osteoblast differentiation (29). Here, we discover that Hh-Gli2 signaling markedly increases Igf2 mRNA and activates the Igf-mTORC2-Akt signaling cascade. Akt, in turn, stabilizes full-length Gli2 via direct phosphorylation at Ser230, thus augmenting Hh-induced transcriptional activation. These findings identify a Hh-Igf positive feedback mechanism that operates during osteoblast differentiation. Results Hh Induces Igf Signaling During Osteoblast Differentiation. To in-
vestigate the molecular mechanism responsible for Hh-induced osteoblast differentiation, we first established an in vitro differentiation system where the murine bone marrow stromal-derived cell line M2-10B4 cells (hereafter M2 cells) was stimulated with purmorphamine (PM), a small molecule agonist of Hh signaling (5). The M2 cells were shown to undergo osteoblastogenesis in response to Hh (30). Through a series of dosage and time-course experiments, we found that 1 μM PM maximally induced the Significance Hedgehog (Hh) signaling is essential for embryonic bone formation in mammals. However, how Hh signaling executes the osteogenic function is not clear. We describe a positive feedback mechanism whereby Hh-Gli2 signaling induces expression of insulin-like growth factor 2 (Igf2), which signals via protein kinases mTORC2 and Akt to stabilize Gli2 and potentiate Hh induction of osteoblast differentiation. The Hh-Igf positive feedback loop may have general implications for Hh signaling in development and disease. Author contributions: F.L. designed research; Y.S. and J.C. performed research; Y.S., C.M.K., and F.L. analyzed data; and Y.S. and F.L. wrote the paper. The authors declare no conflict of interest. This article is a PNAS Direct Submission. 1
Present address: Department of Orthopaedic Surgery, Duke University School of Medicine, Durham, NC 27710.
2
To whom correspondence should be addressed. Email: [email protected].
This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10. 1073/pnas.1502301112/-/DCSupplemental.
PNAS Early Edition | 1 of 6
DEVELOPMENTAL BIOLOGY
Edited by Clifford J. Tabin, Harvard Medical School, Boston, MA, and approved February 27, 2015 (received for review February 10, 2015)
mRNA levels of Gli1 and Ptch1, both Hh transcriptional targets, at 48 or 72 h of treatment (SI Appendix, Fig. S1). Importantly, 1 μM PM activated expression of several common osteoblast markers including Sp7, Alpl, and Bglap at those time points (SI Appendix, Fig. S1). Staining for Alpl activity or matrix mineralization confirmed the induction of osteoblast differentiation by 1 μM PM (Fig. 1A). Knockdown of Gli2, the primary effector for Hh-mediated transcriptional activation, not only reduced Gli1 induction, but also notably diminished the expression of the osteoblast markers in response to PM (Fig. 1B). Thus, PM activates Hh-Gli signaling to induce osteoblast differentiation in M2 cells. To discover effector molecules downstream of Hh-Gli signaling, we performed RNA-seq (whole transcriptome shotgun sequencing) experiments to compare the transcriptome of M2 cells with or without PM stimulation for 24 or 72 h. By examining all genes expressed at >1 read per kilobase of transcript per million reads mapped in at least one condition, and affected by PM by >twofold, we identified Igf2 and Igfbp5 being induced at 24 h, and Igf1, Igf2, Igfbp6, and Igfbp7 increased at 72 h. RT-qPCR (reverse transcription-quantitative polymerase chain reaction) experiments confirmed that PM significantly increased Igf2 mRNA by 24 h and reached the maximal induction of >20 fold by 48 h. The induction of Igf1 mRNA was less pronounced, reaching only ∼twofold after 72 h (Fig. 2A). Igfbp5 showed a twofold induction by 6 h, but Igfbp6 and Igfbp7 were also notably increased by 48 h (>twofold) (SI Appendix, Fig. S2). Furthermore, knockdown of Gli2 abolished Igf2 induction by PM at 24 h (Fig. 2B). Thus, Hh signaling up-regulates the expression of Igf signaling components. To test whether Hh activates Igf signaling, we monitored the phosphorylation of insulin receptor substrate 1 (Irs1) at S632/635 (equivalent of S636/639 in human IRS1), which is stimulated by insulin or Igf signaling (31). In keeping with the maximal induction of Igf2 at 48 h, PM notably increased both total- and phospho-Irs1 at 48 and 72 h (Fig. 2C). Recombinant human N-Shh had a similar effect on the phosphorylation (SI Appendix, Fig. S3). Importantly, knockdown of Igf2, Igf1r, or Gli2 severely blunted the induction of P-Irs1 at 72 h by PM (Fig. 2 D and E). Taken
Fig. 1. Hh signaling induces osteoblast differentiation through Gli2 in M2 cells. (A) Alkaline phosphatase activity (ALP) in M2 cells treated with 1 μM PM or DMSO (Veh) for 72 h. Alizarin Red staining was performed after additional 8 d of incubation in mineralization medium. (B) Effects of shRNA knockdown of Gli2 (sh-Gli2) on mRNA levels of indicated genes as assayed by qPCR. Cells were infected with lentivirus expressing sh-Gli2 or sh-Lacz for 24 h before further incubation with or without 1 μM PM for 72 h. Ribosomal 18S rRNA was used as internal control for qPCR. *P < 0.05, n = 3.
2 of 6 | www.pnas.org/cgi/doi/10.1073/pnas.1502301112
Fig. 2. Hh induces Igf-mTORC2 signaling through Gli2. (A) qPCR analyses in M2 cells treated with 1 μM PM or DMSO (Veh) for 6–72 h. (B) Effect of Gli2 knockdown on Igf2 mRNA level in response to PM for 24 h as assayed by qPCR. Ribosomal 18S rRNA used as internal control for all qPCR. *P < 0.05, n = 3. (C) Western blot for phospho-Irs1(S636/639) (P-Irs1) after 1 μM PM treatment for 48 or 72 h. P-Irs1 is normalized to α-tubulin and fold changes were calculated from three independent experiments with SD indicated. (D and E) Effects of shRNA knockdown of Igf1r, Igf2 (D), or Gli2 (E) on P-Irs1 in response to PM treatment for 72 h. P-Irs1 is normalized to total Irs1, and fold changes were calculated from three independent experiments with SD indicated. (F and G) Effects of shRNA knockdown of Gli2 (F), Igf1r, or Igf2 (G) on P-Akt (S473) in response to PM treatment for 72 h. P-Akt (S473) is normalized to total Akt, and fold changes were calculated from three independent experiments with SD indicated. Sh-Lacz was used as control.
together, the data indicate that Hh activates Igf signaling through up-regulation of Igf signaling components. We next investigated the intracellular pathway downstream of Igf1r in response to Hh signaling. In time-course experiments, we found that PM consistently increased phosphorylation of Akt at S473 [P-Akt (S473)] by 48 h, indicating activation of mTORC2 (SI Appendix, Fig. S4). However, PM did not induce mTORC1 activity, because phosphorylation of S6 or Eif4ebp1 was not different from the control at any of the time points examined (SI Appendix, Fig. S4A). Likewise, N-Shh activated mTORC2 but not mTORC1 after 48 or 72 h of treatment (SI Appendix, Fig. S4B). Importantly, knockdown of Gli2, Igf1r, or Igf2 greatly suppressed the induction of mTORC2 activity by PM, as measured by P-Akt (S473) levels (Fig. 2 F and G). Thus, Hh signaling activates Igf-mTORC2 signaling. We then investigated whether Igf-mTORC2 signaling contributes to Hh-induced osteoblast differentiation in vitro. Knockdown of Igf2 abolished osteoblast differentiation in either M2 cells or primary bone marrow stromal cells (BMSCs) in response to PM, as indicated by the lack of Sp7, Alpl, and Bglap induction (Fig. 3A and SI Appendix, Fig. S5). Knockdown of Igf1r also suppressed PM-induced osteoblast differentiation, but knockdown Shi et al.
Deletion of Igf1r in Hh-Responsive Cells Impairs Bone Formation in Vivo. To investigate the contribution of Igf signaling to Hh-induced Fig. 3. Igf2-mTORC2-Akt signaling is required for Hh-induced osteoblastogenesis. (A) Effects of Igf2 knockdown on mRNA levels of indicated genes in response to PM for 72 h, as assayed by qPCR. (B and C) Effect of Rictor knockdown (B) or MK2206 (C) on P-Akt (S473) assayed by Western blot, and on mRNA levels of osteoblast markers assayed by qPCR. (D) Effects of MK2206 on ALP or Alizarin Red staining in response to PM. Ribosomal 18S rRNA used as internal control for all qPCR. *P < 0.05, n = 3.
osteoblast differentiation in vivo, we sought to disrupt Igf signaling in the Hh-responsive osteogenic cells in the mouse embryo by using the Gli-CreERT2 mouse line. To establish the effectiveness of strategy, we first crossed Gli-CreERT2 with the mT/mG reporter mouse and exposed the embryos to tamoxifen (TM) from embryonic day (E)13.5 until E16.5 when they were harvested. In the long bones of the Gli-CreERT2; mT/mG embryo, numerous green cells were observed within the osteogenic perichondrium and the cartilage, indicating that Gli-CreERT2 effectively targets the
of insulin receptor (Insr) had no effect (SI Appendix, Fig. S6). Furthermore, knockdown of the mTORC2-specific component Rictor greatly diminished PM-induced osteoblast differentiation (Fig. 3B). Pharmacological inhibition of all mTOR activity with Torin1 exerted even greater inhibition on differentiation (SI Appendix, Fig. S7). Finally, inhibition of the mTORC2 target Akt with MK2206 abolished PM-induced mRNA expression of Sp7, Alpl, and Bglap, as well as Alpl staining and matrix mineralization (Fig. 3 C and D). Thus, Hh signaling activates the Igf2-mTORC2Akt axis to induce osteoblast differentiation. Igf2 Synergizes with Hh Signaling via Akt-Dependent Gli2 Stabilization.
We next pursued how Igf2-mTORC2-Akt signaling mediates Hh-induced osteoblastogenesis. Igf2 by itself did not induce the expression of Sp7, Alpl, or Bglap in M2 cells, but synergistically increased the mRNA levels of these markers when used in combination of suboptimal levels (e.g., 0.5 μM) of PM (Fig. 4A). Interestingly, in such settings, Igf2 also boosted the expression levels of Hh target genes Gli1 and Ptch1 compared with PM alone, although Igf2 itself did not induce these genes (Fig. 4A). Conversely, knockdown of Igf2 or Igf1r reduced the mRNA levels of Gli1 and Ptch1 in response to PM (Fig. 4B and SI Appendix, Fig. S8A). Likewise, inhibition of mTORC2 with either Torin1 or by knocking down Rictor severely diminished the induction of Gli1 by PM (SI Appendix, Fig. S8 B and C). Finally, inhibition of Akt with MK2206 also blunted the Hh transcriptional response (Fig. 4C). Thus, Igf2mTORC2-Akt signaling stimulates osteoblast differentiation by augmenting the transcriptional response to Hh signaling. To determine how Igf2-mTORC2-Akt enhances transcriptional activation by Hh, we examined the potential effect of Igf signaling on Gli2, the primary effector for Hh-induced gene activation. To this end, we infected M2 cells with lentiviruses Shi et al.
Fig. 4. Igf2 synergizes with Hh signaling via Akt-dependent Gli2 stabilization. (A) Synergistic induction of genes by 0.5 μM PM or 100 ng/mL Igf2, as assayed by qPCR after 72 h of treatment. (B and C) Effects of Igf2 knockdown (B) or MK2206 (C) assayed by qPCR after 72 h of PM treatment. *P < 0.05, n = 3. (D) Effects of Igf2 and MK2206 on Flag-Gli2 turnover, as measured by Western blot. *P < 0.05, n = 3. (E) Effect of Igf2 on Flag-Gli2-S230A turnover. (F) Effect of Igf2 on endogenous Gli2 stability. All quantification is from three independent experiments. CHX, cyclohexamide.
PNAS Early Edition | 3 of 6
DEVELOPMENTAL BIOLOGY
expressing Flag-tagged full-length Gli2 (Flag-Gli2) in response to doxycycline (Dox), and then treated the cells with Igf2. Igf2 had no effect on Gli2 mRNA levels (including both endogenous Gli2 and exogenous Flag-Gli2), but notably increased the protein level of Flag-Gli2, indicating that Igf2 may suppress Gli2 turnover (SI Appendix, Fig. S9 A and B). In contrast, Igf2 did not affect the protein level of Flag-Gli3 in a similar experiment (SI Appendix, Fig. S9C). To determine the potential effect of Igf2 on the half-life of Gli2, we blocked de novo protein synthesis with cyclohexamide after the induction of Flag-Gli2 expression with Dox, and monitored the abundance of Flag-Gli2 as a function of time with or without Igf2 treatment. These experiments revealed that Igf2 prolonged the half-life of Flag-Gli2 from 24 h, and that the stabilizing effect of Igf2 was completely abolished by the Akt inhibitor MK2206 (Fig. 4D). Examination of the mouse Gli2 protein sequence revealed a consensus phosphorylation site for Akt, R225KRALS230 (32). Alanine substitution of S230 abolished the stabilizing effect of Igf2 on Gli2 (Fig. 4E). Finally, to confirm the effect of Igf2 on endogenous Gli2, we generated an antibody against a peptide within the C-terminal half of the mouse Gli2 (SI Appendix, Fig. S10) and found that Igf2 stabilized the endogenous full-length Gli2 protein (Fig. 4F). Thus, Igf2 markedly suppresses Gli2 turnover likely through direct phosphorylation by Akt.
cells known to respond to Ihh during long bone development (Fig. 5A). We next generated Gli-CreERT2; Igf1rf/f embryos (CKO) and induced Igf1r deletion with the same TM regimen. Immunofluorescence staining confirmed that Igf1r was notably reduced in both perichondrial cells and chondrocytes in the long bones of E16.5 embryos (SI Appendix, Fig. S11). Importantly, whole-mount skeletal staining revealed that the bone collars of all long bones in the CKO embryo were notably shorter than their normal counterparts, although the total length of each element appeared to be similar across genotypes (Fig. 5B). Quantification indicated that the relative length of bone collar versus the entire element was reduced by 10–20% in the CKO embryo, indicating that ossification was suppressed (Fig. 5C). The bone collar deficit persisted at E18.5 (SI Appendix, Fig. S12). Staining of the long bone sections with the von Kossa method identified a marked decrease in bone collar mineralization in the CKO embryo (Fig. 5D). Furthermore, in situ hybridization on sections showed that the mRNA levels of osteoblast markers Runx2, Sp7, Alpl, and the Hh target gene Ptch1 were diminished in CKO, although Ihh expression was relatively normal (Fig. 5E). However, chondrocyte proliferation or hypertrophy was relatively normal in E14.5 CKO embryos after 3 d of TM induction (SI Appendix, Fig. S13). Western blot analyses confirmed that in the limbs of the E14.5 CKO embryos, the levels of Igf1r, Gli2, and P-Akt (S473) were notably reduced compared with those in the
Fig. 5. Deletion of Igf1r in Hh-responsive cells impairs bone formation in vivo. (A) Detection of GFP by immunofluorescence on a tibial section from an E16.5 Gli1-CreERT2; mT/mG mouse embryo that received daily tamoxifen beginning at E13.5. H, hypertrophic chondrocytes; P, perichondrium; PH, prehypertrophic chondrocytes; S, primary spongiosa. (B and C) Skeletal staining of long bones and quantification of relative bone collar length in E16.5 wildtype (WT) versus Gli1-CreERT2; Igf1rf/f (CKO) littermate embryos. Vertical lines demarcate the ends of the bone collar; the horizontal lines denote the deficit in the mutant (CKO). *P < 0.05, n = 5. F, femur; H, humerus; R, radius; T, tibia; U, ulna. (D) Von Kossa staining on longitudinal tibial sections in WT versus CKO littermates at E16.5. Green boxed areas shown at a higher magnification in Insets. (E) In situ hybridization on tibial sections of E16.5 embryos. Signal is in red. (F and G) Western blot analyses of protein extracts from limbs of E14.5 WT versus CKO littermates with TM treatments beginning at E11.5. Each lane is from a separate embryo. Intensity normalization is against α-tubulin, except for P-Akt (S473) against total Akt.
4 of 6 | www.pnas.org/cgi/doi/10.1073/pnas.1502301112
wild-type littermates (Fig. 5 F and G). In contrast, the level of fulllength Gli3 was largely normal in the CKO samples (Fig. 5G). In summary, Igf signaling is necessary to potentiate osteoblastogenesis in Ihh-responsive cells during long bone development. Discussion We have uncovered a positive feedback mechanism between Hh and Igf signaling during osteoblast differentiation. In particular, Hh signaling induces transcription of multiple members of the Igf system, most notably Igf2, resulting in activation of the mTORC2-Akt signaling axis. Akt, in turn, stabilizes full-length Gli2, likely through phosphorylation of S230, thereby amplifying the transcriptional output of Hh signaling. Supporting the physiologic relevance of the Hh-Igf positive feedback loop, cellautonomous removal of Igf signaling in Ihh-responding cells diminishes osteoblast differentiation in the mouse embryo. This finding not only sheds light on how Hh signaling induces osteoblastogenesis, but also adds another dimension to the role of Igf signaling in bone formation. The mechanism whereby Hh activates osteoblast differentiation remains to be fully elucidated. Our data have demonstrated an important role for Igf signaling in Hh-induced osteoblastogenesis, but it is also clear that Igf signaling cannot replace Hh to initiate the osteogenic program. In contrast to Ihh deletion that abolishes bone formation altogether in the endochondral skeleton, removal of Igf1r impairs but does not prevent the process (33). Data from the present study support the model wherein Igf signaling amplifies Gli-mediated transcriptional induction to promote osteoblast differentiation. This model predicts that certain osteogenic genes are induced not immediately by Hh but only after subsequent activation of Igf signaling, a process occurring between 24 and 48 h of PM treatment in M2 cells. Indeed, our RNA-seq experiments identified several transcription factors being induced after 72 but not 24 h of PM treatment; these factors include Dlx5, Dlx6, Sp7, Msx1, and Ets2, all of which have been implicated in osteoblast differentiation (34–37). Future studies are necessary to determine the contributions of those factors to osteoblast differentiation in response to Hh signaling. The mechanism whereby Hh signaling induces Igf2 expression is not clear at present. We show that Gli2 is necessary for Hh to induce Igf2 mRNA. This result, along with the previous report that Gli1 overexpression induces Igf2 in C3H10T1/2 cells, raises the possibility that Igf2 might be a direct target of Gli activators (38). However, unlike the known Gli target genes such as Gli1 that show activation by 6 h of Hh stimulation, Igf2 is not induced until after 24 h (SI Appendix, Fig. S1). Moreover, although a putative Gli binding site is identified in the 3′UTR of the Igf2 gene, chromatin immunoprecipitation experiments with cells overexpressing Flag-tagged Gli2 have failed to reveal Gli2 binding to the site. We have considered the possibility that other secreted factors such as BMP, Wnt, or PTHrP may mediate the induction of Igf2 by Hh, but the RNA-seq experiment did not detect any significant increase in their expression after 24 h of PM stimulation. Thus, the induction of Igf2 by Gli2 appears to be through an indirect and yet unknown mechanism. We have identified S230 of Gli2 as a critical mediator for Igf2 to potentiate Hh signaling. In our model, upon activation by Igf2, Akt phosphorylates and stabilizes full-length Gli2 that requires Hh signaling for activation. Previous work has identified clusters of adjacent serine residues whose sequential phosphorylation by PKA and Gsk3 controls Gli2 turnover (39). That study also reported that alanine substitutions of the PKA, but not Gsk3 or Akt phosphorylation sites, abolished the inhibitory effect of dominant negative Akt on Gli2-mediated transcription, although it did not examine the stability of Gli2. It is possible that Aktmediated S230 phosphorylation precludes PKA-mediated phosphorylation, thus stabilizing Gli2. Moreover, the full effect of Akt on Gli2 stability may involve both the direct regulation as reported Shi et al.
Cell Culture. M2-10B4 cells (CRL-1972; ATCC) were maintained in RPMI (catalog no. 11875-085, Gibco) with 10% FBS (Gibco) as per instructions. Unless otherwise indicated, M2 cells were seeded at the density of 1 × 104 cells per cm2, 48 h before experiments. Primary cultures of BMSCs were harvested from the femur and the tibia of 6-wk-old mice and cultured in α-MEM containing 10% FBS (Gibco), with a change of medium at day 3. Cells were dissociated at day 7 and reseeded at 1 × 105 cells per cm2 for qPCR, ALP staining or mineralization assays at appropriate time points. To express Flag-tagged Gli2 proteins in M2 cells, the mouse Gli2 full-length cDNA, either wild type or the S230A mutant form, was cloned into the Fuw-tetO vector (plasmid 20723; Addgene) and retrovirus was produced according to standard protocol. A retrovirus producing Flag-Gli3 (full-length) was made in the same way. Then M2-10B4 cells were infected with FuW-rtTA (plasmid 20342; Addgene) virus, together with Fuw-tetO-FlagGli2 or Fuw-tetO-FlagGli3 virus. Gli2 expression was induced with addition of doxycycline (100 nM) to the media. The recombinant human Shh (464-SH-025) and Igf2 (792-MG-050) were purchased from R&D Systems. Purmorphamine (540223; Millipore), MK2206 (S1078; Selleckchem), Torin1 (4247; Tocris Biosciences) were purchased. Gli2 shRNA plasmid (Mission shRNA SHCLND-NM_001081125) was purchased from Sigma. Lentiviral shRNA targeting vectors were purchased from Washington University shRNA consortium. Targeted sequences are listed in SI Appendix, Table S1.
Materials and Methods
Western Blot Analyses. Protein extracts from cells were prepared in radioimmunoprecipitation assay buffer containing phosphatase and proteinase inhibitors (Roche). Total protein content was determined by the bicinchoninic acid method (Pierce). Membranes were imaged with Molecular Imager ChemiDoc XRS+ System (Bio-Rad). Quantification of Western blots was performed with ImageLab software. The monoclonal antibodies against phospho-IRS1(S636/S639) (2388), total-IRS1 (3407), phospho-AKT(S473) (9271), total-AKT (9272), phospho-S6 (2215), total-S6 (2217), phosphoEif4ebp1 (T37/46) (2855), total-Eif4ebp1 (9644), Rictor (2140) and β-actin (4970) were purchased from Cell Signaling Technology. The antibody against Gli3 (ab6050) was from Abcam, and the M2 antibody (catalog no. F1804) against Flag was from Sigma. The monoclonal antibody against α-tubulin (sc8035) and the polyclonal antibody against Igf1r (sc-713) were obtained from Santa Cruz Biotechnology. HRP-conjugated anti-rabbit (NA934V) or antimouse secondary antibody (sc-2005) was purchased from GE Healthcare or Santa Cruz Biotechnology, respectively. The Alexa Fluor 488 donkey antirabbit IgG (H+L) antibody was obtained from Invitrogen.
High-Throughput RNA Sequencing. M2 cells were treated with or without 1 μM PM for 24 or 72 h before total RNA was isolated by using the RNeasy kit with on-column DNase treatment (Qiagen). Poly-A RNA was isolated from 10 μg of total RNA by using Oligo-dT beads, followed by fragmentation and reverse transcription. The double-stranded cDNA was used to generate a library with sequencing adapters added to the ends of DNA fragments, and the resulting library was sequenced by using the Illumina HiSeq-2000 as single reads extending 42 bases (Genome Technology Access Center, Washington University School of Medicine). The raw data were demultiplexed and aligned to the reference genome (mm9) by using TopHat. Partek Genomic Suite was used to assemble transcripts and analyze expression. Mouse Studies. Igf1rf/f, mT/mG, and Gli1-CreERT2 mice have been reported (26, 47, 48). The Animal Studies Committee at Washington University in St. Louis reviewed and approved all mouse procedures used in this study. The Gli1-CreERT2;Igf1rf/f or Gli1-CreERT2;mT/mG pregnant female mice were administered tamoxifen dissolved in corn oil (4 mg/30 g of body weight) by oral gavage daily for durations as indicated. Whole-mount skeletal staining was performed with a protocol based on McLeod (49). To assess skeletal phenotypes on sections, embryonic limbs were harvested in PBS and fixed in 10% (vol/vol) buffered formalin overnight at room temperature. The fixed limbs were then processed and embedded in paraffin before sectioning at 6 μm. Limbs from E18.5 embryos were decalcified in 14% (wt/vol) EDTA/PBS (pH 7.4) for 48 h after fixation and before processing. In situ hybridization was performed by using 35S-labeled riboprobes as described (15, 42, 50, 51). The chondrocyte proliferation rate in embryos was assessed by immunodetection of BrdU incorporation on limb sections of the embryos (Invitrogen) (52). The percentage of labeled cells was calculated from medial sections of three pairs of wild-type versus mutant littermates. Igf1r immunofluorescence staining was performed on paraffin sections after antigen retrieval with citrate buffer and blocking with 5% (vol/vol) normal goat serum in Tris buffered saline with Tween. Sections were incubated with Igf1r antibody (sc-713; Santa Cruz Biotechnology) at 1:100 dilution overnight at 4 °C, and then Alexa Fluor 488 donkey anti-rabbit IgG (H+L) antibody (A11371; Invitrogen) at 1:300 dilution at room temperature for 1 h.
1. Ingham PW, McMahon AP (2001) Hedgehog signaling in animal development: Paradigms and principles. Genes Dev 15(23):3059–3087. 2. Huangfu D, Anderson KV (2006) Signaling from Smo to Ci/Gli: Conservation and divergence of Hedgehog pathways from Drosophila to vertebrates. Development 133(1):3–14. 3. Goetz SC, Ocbina PJ, Anderson KV (2009) The primary cilium as a Hedgehog signal transduction machine. Methods Cell Biol 94:199–222.
Shi et al.
Osteoblast Differentiation Assays. Alkaline phosphatase staining was as described (53). When inhibitors were used, cells were pretreated with the inhibitor in normal growth medium for 1 h. When shRNA lentiviruses were used, cells were infected with respective shRNA lentivirus for 24 h before subsequent treatments. For mineralization assays, confluent cells were incubated in the presence of 50 mg/mL ascorbic acid and 50 mM β-glycerophosphate for 8 d and then stained with Alizarin Red S.
Quantitative PCR. RNA was isolated from whole cells with QIAGEN RNeasy kit (74104) and was transcribed into cDNA by using iScript cDNA synthesis kit (Bio-Rad). Fast-start SYBR Green (Bio-Rad) and 0.1 mM primers were used in each reaction. Ribosomal 18S RNA was used for normalization. PCR was performed with Step-One machine (ABI). The primer sequences are provided in SI Appendix, Table S2. Production of Gli2 Antibody. A monoclonal antibody specific for full-length Gli2 was generated against a peptide within the C-terminal half of mouse Gli2 according to proprietary techniques (Abmart). The purified antibody was validated by Western blot analyses for detection of overexpressed Gli2, and for loss of endogenous Gli2 following shRNA knockdown. Statistical Analyses. All quantitative data are presented as mean ± SD with a minimum of three independent samples. Statistical significance is determined by Student’s t test. ACKNOWLEDGMENTS. This work is supported by NIH Grant R01 DK065789 (to F.L.).
4. Wu X, Walker J, Zhang J, Ding S, Schultz PG (2004) Purmorphamine induces osteogenesis by activation of the hedgehog signaling pathway. Chem Biol 11(9):1229–1238. 5. Sinha S, Chen JK (2006) Purmorphamine activates the Hedgehog pathway by targeting Smoothened. Nat Chem Biol 2(1):29–30. 6. Sasaki H, Nishizaki Y, Hui C, Nakafuku M, Kondoh H (1999) Regulation of Gli2 and Gli3 activities by an amino-terminal repression domain: Implication of Gli2 and Gli3 as primary mediators of Shh signaling. Development 126(17):3915–3924.
PNAS Early Edition | 5 of 6
DEVELOPMENTAL BIOLOGY
here and indirect control via phosphorylation and inactivation of Gsk3 (40). Future studies are necessary to determine whether and how Akt-mediated phosphorylation of Gli2 affects its phosphorylation by PKA or Gsk3. It is worth noting that deletion of Igf1r in Hh-responsive cells did not grossly affect cartilage development. This observation is consistent with our previous finding that Ihh and Igf control the linear growth of the skeleton largely independently (29). In addition, previous studies have demonstrated that Ihh controls cartilage development chiefly through derepression of Gli3, but induces osteoblastogenesis via both Gli3 derepression and Gli2 activation (16, 41, 42). Here, we show that Igf signaling stabilizes full-length Gli2 but not Gli3. Thus, the relative contribution of Gli2 versus Gli3 likely determines the functional relevance of the Hh-Igf feedback mechanism in specific physiological contexts. Beyond osteoblastogenesis, Hh and Igf may functionally interact in other processes. Most notably, in the Ptch1+/− mouse model for medulloblastoma and rhabdomyosarcoma, Igf2 was upregulated and necessary for tumor formation (43). Similarly, cooverexpression of Igf2 or activated Akt with Shh synergistically induces medulloblastoma in mice (44). Mechanistically, Igf-PI3K and Hh signaling appeared to converge on Nmyc1 to promote proliferation of cerebellar neuronal precursors (45). Moreover, Igf synergized with Shh in promoting somite myogenesis in explant cultures (46). Overall, Hh and Igf signaling exhibit synergistic interactions both in normal development and during pathogenesis.
7. Bai CB, Auerbach W, Lee JS, Stephen D, Joyner AL (2002) Gli2, but not Gli1, is required for initial Shh signaling and ectopic activation of the Shh pathway. Development 129(20):4753–4761. 8. Mo R, et al. (1997) Specific and redundant functions of Gli2 and Gli3 zinc finger genes in skeletal patterning and development. Development 124(1):113–123. 9. Litingtung Y, Dahn RD, Li Y, Fallon JF, Chiang C (2002) Shh and Gli3 are dispensable for limb skeleton formation but regulate digit number and identity. Nature 418(6901):979–983. 10. Park HL, et al. (2000) Mouse Gli1 mutants are viable but have defects in SHH signaling in combination with a Gli2 mutation. Development 127(8):1593–1605. 11. Wang B, Fallon JF, Beachy PA (2000) Hedgehog-regulated processing of Gli3 produces an anterior/posterior repressor gradient in the developing vertebrate limb. Cell 100(4):423–434. 12. Pan Y, Bai CB, Joyner AL, Wang B (2006) Sonic hedgehog signaling regulates Gli2 transcriptional activity by suppressing its processing and degradation. Mol Cell Biol 26(9):3365–3377. 13. Pan Y, Wang B (2007) A novel protein-processing domain in Gli2 and Gli3 differentially blocks complete protein degradation by the proteasome. J Biol Chem 282(15): 10846–10852. 14. St-Jacques B, Hammerschmidt M, McMahon AP (1999) Indian hedgehog signaling regulates proliferation and differentiation of chondrocytes and is essential for bone formation. Genes Dev 13(16):2072–2086. 15. Long F, et al. (2004) Ihh signaling is directly required for the osteoblast lineage in the endochondral skeleton. Development 131(6):1309–1318. 16. Joeng KS, Long F (2009) The Gli2 transcriptional activator is a crucial effector for Ihh signaling in osteoblast development and cartilage vascularization. Development 136(24):4177–4185. 17. Hojo H, et al. (2012) Gli1 protein participates in Hedgehog-mediated specification of osteoblast lineage during endochondral ossification. J Biol Chem 287(21):17860–17869. 18. Kawai M, Rosen CJ (2012) The insulin-like growth factor system in bone: Basic and clinical implications. Endocrinol Metab Clin North Am 41(2):323–333, vi. 19. Pollak M (2008) Insulin and insulin-like growth factor signalling in neoplasia. Nat Rev Cancer 8(12):915–928. 20. Liu P, et al. (2013) Sin1 phosphorylation impairs mTORC2 complex integrity and inhibits downstream Akt signalling to suppress tumorigenesis. Nat Cell Biol 15(11): 1340–1350. 21. Dalle Pezze P, et al. (2012) A dynamic network model of mTOR signaling reveals TSCindependent mTORC2 regulation. Sci Signal 5(217):ra25. 22. Rechler MM, Clemmons DR (1998) Regulatory actions of insulin-like growth factorbinding proteins. Trends Endocrinol Metab 9(5):176–183. 23. Bikle D, et al. (2001) The skeletal structure of insulin-like growth factor I-deficient mice. J Bone Miner Res 16(12):2320–2329. 24. Zhao G, et al. (2000) Targeted overexpression of insulin-like growth factor I to osteoblasts of transgenic mice: Increased trabecular bone volume without increased osteoblast proliferation. Endocrinology 141(7):2674–2682. 25. Xian L, et al. (2012) Matrix IGF-1 maintains bone mass by activation of mTOR in mesenchymal stem cells. Nat Med 18:1095–1101. 26. Zhang M, et al. (2002) Osteoblast-specific knockout of the insulin-like growth factor (IGF) receptor gene reveals an essential role of IGF signaling in bone matrix mineralization. J Biol Chem 277(46):44005–44012. 27. Yang S, et al. (2011) Foxo1 mediates insulin-like growth factor 1 (IGF1)/insulin regulation of osteocalcin expression by antagonizing Runx2 in osteoblasts. J Biol Chem 286(21):19149–19158. 28. Qiao M, Shapiro P, Kumar R, Passaniti A (2004) Insulin-like growth factor-1 regulates endogenous RUNX2 activity in endothelial cells through a phosphatidylinositol 3-kinase/ERK-dependent and Akt-independent signaling pathway. J Biol Chem 279(41): 42709–42718. 29. Long F, Joeng KS, Xuan S, Efstratiadis A, McMahon AP (2006) Independent regulation of skeletal growth by Ihh and IGF signaling. Dev Biol 298(1):327–333. 30. Dwyer JR, et al. (2007) Oxysterols are novel activators of the hedgehog signaling pathway in pluripotent mesenchymal cells. J Biol Chem 282(12):8959–8968.
6 of 6 | www.pnas.org/cgi/doi/10.1073/pnas.1502301112
31. Copps KD, White MF (2012) Regulation of insulin sensitivity by serine/threonine phosphorylation of insulin receptor substrate proteins IRS1 and IRS2. Diabetologia 55(10):2565–2582. 32. Obata T, et al. (2000) Peptide and protein library screening defines optimal substrate motifs for AKT/PKB. J Biol Chem 275(46):36108–36115. 33. Liu JP, Baker J, Perkins AS, Robertson EJ, Efstratiadis A (1993) Mice carrying null mutations of the genes encoding insulin-like growth factor I (Igf-1) and type 1 IGF receptor (Igf1r). Cell 75(1):59–72. 34. Robledo RF, Rajan L, Li X, Lufkin T (2002) The Dlx5 and Dlx6 homeobox genes are essential for craniofacial, axial, and appendicular skeletal development. Genes Dev 16(9):1089–1101. 35. Nakashima K, et al. (2002) The novel zinc finger-containing transcription factor osterix is required for osteoblast differentiation and bone formation. Cell 108(1):17–29. 36. Zhang Z, et al. (2003) Msx1/Bmp4 genetic pathway regulates mammalian alveolar bone formation via induction of Dlx5 and Cbfa1. Mech Dev 120(12):1469–1479. 37. Raouf A, Seth A (2000) Ets transcription factors and targets in osteogenesis. Oncogene 19(55):6455–6463. 38. Ingram WJ, Wicking CA, Grimmond SM, Forrest AR, Wainwright BJ (2002) Novel genes regulated by Sonic Hedgehog in pluripotent mesenchymal cells. Oncogene 21(53):8196–8205. 39. Riobó NA, Lu K, Ai X, Haines GM, Emerson CP, Jr (2006) Phosphoinositide 3-kinase and Akt are essential for Sonic Hedgehog signaling. Proc Natl Acad Sci USA 103(12): 4505–4510. 40. Cross DA, Alessi DR, Cohen P, Andjelkovich M, Hemmings BA (1995) Inhibition of glycogen synthase kinase-3 by insulin mediated by protein kinase B. Nature 378(6559):785–789. 41. Koziel L, Wuelling M, Schneider S, Vortkamp A (2005) Gli3 acts as a repressor downstream of Ihh in regulating two distinct steps of chondrocyte differentiation. Development 132(23):5249–5260. 42. Hilton MJ, Tu X, Cook J, Hu H, Long F (2005) Ihh controls cartilage development by antagonizing Gli3, but requires additional effectors to regulate osteoblast and vascular development. Development 132(19):4339–4351. 43. Hahn H, et al. (2000) Patched target Igf2 is indispensable for the formation of medulloblastoma and rhabdomyosarcoma. J Biol Chem 275(37):28341–28344. 44. Rao G, et al. (2004) Sonic hedgehog and insulin-like growth factor signaling synergize to induce medulloblastoma formation from nestin-expressing neural progenitors in mice. Oncogene 23(36):6156–6162. 45. Kenney AM, Widlund HR, Rowitch DH (2004) Hedgehog and PI-3 kinase signaling converge on Nmyc1 to promote cell cycle progression in cerebellar neuronal precursors. Development 131(1):217–228. 46. Pirskanen A, Kiefer JC, Hauschka SD (2000) IGFs, insulin, Shh, bFGF, and TGF-beta1 interact synergistically to promote somite myogenesis in vitro. Dev Biol 224(2): 189–203. 47. Yu W, McDonnell K, Taketo MM, Bai CB (2008) Wnt signaling determines ventral spinal cord cell fates in a time-dependent manner. Development 135(22):3687–3696. 48. Muzumdar MD, Tasic B, Miyamichi K, Li L, Luo L (2007) A global double-fluorescent Cre reporter mouse. Genesis 45(9):593–605. 49. McLeod MJ (1980) Differential staining of cartilage and bone in whole mouse fetuses by alcian blue and alizarin red S. Teratology 22(3):299–301. 50. Hu H, et al. (2005) Sequential roles of Hedgehog and Wnt signaling in osteoblast development. Development 132(1):49–60. 51. Long F, Zhang XM, Karp S, Yang Y, McMahon AP (2001) Genetic manipulation of hedgehog signaling in the endochondral skeleton reveals a direct role in the regulation of chondrocyte proliferation. Development 128(24):5099–5108. 52. Chen J, Holguin N, Shi Y, Silva MJ, Long F (2015) mTORC2 signaling promotes skeletal growth and bone formation in mice. J Bone Miner Res 30(2):369–378. 53. Shi Y, et al. (2012) Uniaxial mechanical tension promoted osteogenic differentiation of rat tendon-derived stem cells (rTDSCs) via the Wnt5a-RhoA pathway. J Cell Biochem 113(10):3133–3142.
Shi et al.
