Deletion of IGF-I receptor (IGF-IR) in primary osteoblasts reduces GH-induced STAT5 signaling

Abstract

GH promotes longitudinal growth and regulates multiple cellular functions in humans and animals. GH signals by binding to GH receptor (GHR) to activate the tyrosine kinase, Janus kinase 2 (JAK2), and downstream pathways including signal transducer and activator of transcription 5 (STAT5), thereby regulating expression of genes including IGF-I. GH exerts effects both directly and via IGF-I, which signals by activating the IGF-I receptor (IGF-IR). IGF-IR is a cell surface receptor that contains intrinsic tyrosine kinase activity within its intracellular domain. In this study, we examined the potential role of IGF-IR in facilitating GH-induced signal transduction, using mouse primary calvarial osteoblasts with Lox-P sites flanking both IGF-IR alleles. These cells respond to both GH and IGF-I and in vitro infection with an adenovirus that drives expression of Cre recombinase (Ad-Cre) dramatically reduces IGF-IR abundance without affecting the abundance of GHR, JAK2, STAT5, or ERK. Notably, infection with Ad-Cre, but not a control adenovirus, markedly inhibited acute GH-induced STAT5 activity (more than doubling the ED(50) and reducing the maximum activity by nearly 50%), while sparing GH-induced ERK activity, and markedly inhibited GH-induced transactivation of a STAT5-dependent luciferase reporter. The effect of Ad-Cre on GH signaling was specific, as platelet-derived growth factor-induced signaling was unaffected by Ad-Cre-mediated reduction of IGF-IR. Ad-Cre-mediated inhibition of GH signaling was reversed by adenoviral reexpression of IGF-IR, but not by infection with an adenovirus that drives expression of a hemagglutination-tagged somatostatin receptor, which drives expression of the unrelated somatostatin receptor, and Ad-Cre infection of nonfloxed osteoblasts did not affect GH signaling. Notably, infection with an adenovirus encoding a C-terminally truncated IGF-IR that lacks the tyrosine kinase domain partially rescued both acute GH-induced STAT5 activity and GH-induced IGF-I gene expression in cells in which endogenous IGF-IR was reduced. These data, in concert with our earlier findings that GH induces a GHR-JAK2-IGF-IR complex, suggest a novel function for IGF-IR. In addition to its role as a key IGF-I signal transducer, this receptor may directly facilitate acute GH signaling. The implications of these findings are discussed.

Figure 1

Characterization of IGF-IR-floxed primary osteoblasts. A, GH-stimulated JAK2 and STAT5 phosphorylation. Primary osteoblasts were harvested from newborn mice bearing lox-P sites flanking both IGF-IR alleles. Serum-starved cells were treated with vehicle or GH (500 ng/ml) for 15 min. Detergent extracts were resolved by SDS-PAGE and immunoblotted with anti-pTyr JAK2 and anti-pTyr STAT5, respectively. B, Osteoblasts harvested and serum starved as in A were treated with vehicle or GH (500 ng/ml) for 15 min. Detergent extracts were immunoprecipitated with either anti-GHR or nonimmune (NI) serum. Precipitated proteins were resolved by SDS-PAGE and blotted with antiphosphotyrosine. The blots shown are representative of two independent experiments. C, IGF-IR-floxed osteoblasts were infected with either 800 MOI Ad-GFP or 800 MOI Ad-Cre for 48 h. After overnight serum starvation, detergent extracts were resolved by SDS-PAGE and immunoblotted with anti-IGF-IRβ and anti-β-actin, respectively.

Figure 2

Effect of IGF-IR deletion on GH signaling in IGF-IR-floxed osteoblasts. A and B, GH-induced STAT5 activity. Primary osteoblasts were infected with Ad-Cre vs. Ad-GFP, as indicated, as in Fig. 1C. Serum-starved cells were treated with the indicated concentrations of GH for 10 min, after which detergent extracts were resolved by SDS-PAGE and serially immunoblotted with anti-pSTAT5, anti-STAT5, and anti-IGF-IRβ. A, Representative immunoblots. B, Densitometric quantitation of pSTAT5/STAT5 signals from two (5, 25, 50, and 125 ng/ml) or six (250 and 500 ng/ml) independent experiments (including that shown in A). In each experiment, the maximum signal was considered 100%. Data are plotted as mean ± se. Asterisk indicates P < 0.02 for comparison of Ad-GFP- vs. Ad-Cre-infected cells at each GH concentration. C, GH-induced STAT5-dependent reporter gene activity. Floxed IGF-IR cells treated with Ad-GFP or Ad-Cre were infected with Ad-GHRE-luc and then stimulated with GH (50 ng/ml; 6 h). Luciferase activity was measured in cell extracts. Mean ± se for triplicates is shown, normalized to maximum response. D and E, GH-induced ERK activity. D, Blots in A were reprobed with anti-pERK and anti-ERK. E, Densitometric quantitation of pERK/ERK signal from the same experiments as in B above.

Figure 3

Impact of Ad-Cre infection on GH signaling does not reflect global effects on cell integrity or other pathways. A, Adenoviral infection of nonfloxed primary osteoblasts. Primary osteoblasts from mice without floxed IGF-IR genes were serum starved and treated with GH at the indicated concentrations for 10 min, after which detergent extracts were resolved by SDS-PAGE and sequentially immunoblotted with the indicated antibodies. Note the lack of effect of Ad-Cre vs. Ad-GFP on GH-induced signaling. The blots shown are representative of two independent experiments. B, Ad-Cre-mediated IGF-IR deletion did not affect PDGF signaling. IGF-IR-floxed primary osteoblasts were infected with Ad-GFP vs. Ad-Cre, as in Figs. 1 and 2. Serum-starved cells were treated with PDGF at the indicated concentrations for 10 min. Detergent extracts were resolved by SDS-PAGE and sequentially immunoblotted with anti-pAkt, anti-pERK, and anti-IGF-IRβ. The blots shown are representative of two independent experiments.

Figure 4

Reconstitution with IGF-IR rescues GH-induced STAT5 activation in Ad-Cre-infected IGF-IR-floxed osteoblasts. A, Floxed-IGF-IR primary osteoblasts were infected with either Ad-GFP only (1600 MOI), Ad-GFP plus Ad-Cre (800 MOI each simultaneously), or Ad-IGF-IR plus Ad-Cre (800 MOI each simultaneously), as in Figs. 1–3. Serum-starved cells were treated with vehicle or GH (250 ng/ml) for 10 min, after which detergent extracts were resolved by SDS-PAGE and sequentially immunoblotted with anti-pSTAT5, anti-STAT5, and anti-IGF-IRα. Positions of IGF-IR precursor and IGF-IRα chain are indicated. The blots shown are representative of two separate experiments. Note the rescue of GH-induced STAT5 activity by IGF-IR expression. B, The experiment in A was repeated, except that Ad-SSR-HA was used instead of Ad-IGF-IR for infection. Immunoblotting was with anti-pSTAT5, anti-STAT5, and anti-HA. The blots shown are representative of three separate experiments. Note the lack of rescue of GH-induced STAT5 activity by SSR-HA expression.

Figure 5

IGF-IR tyrosine kinase inhibitor does not prevent GH-induced STAT5 activation. A and B, IGF-IR-floxed primary osteoblasts were serum starved and pretreated with vehicle or NVP-AEW-541 (5 mm, as indicated) for 16 h before stimulation with IGF-I (A, 100 ng/ml) or GH (B, 500 ng/ml) for 10 min. Detergent extracts were resolved by SDS-PAGE and immunoblotted with anti-pAkt and anti-Akt (A) or anti-pSTAT5 and anti-STAT5 (B). The blots shown are representative of two separate experiments. Note inhibition of IGF-I, but not GH, signaling by NVP-AEW-541.

Figure 6

IGF-IR structure. A and B, Diagram of the full-length IGF-IR (A) and C-terminally truncated IGF-IR_Δ950 mutant (B). IGF-IR is a disulfide-linked heterotetramer consisting of an α₂β₂ assemblage in which the α-chain is entirely extracellular and the β-chain is a transmembrane protein that harbors a tyrosine kinase in its intracellular domain. L, leucine-rich repeat domain; Cys-rich, cysteine-rich; FnIII, fibronectin III; TK, tyrosine kinase domain.

Figure 7

Partial rescue of GH-induced STAT5 signaling by an IGF-IR lacking the intracellular domain of its β-chain. A, Upper panel, Floxed-IGF-IR primary osteoblasts were infected with either Ad-GFP only (1600 MOI), Ad-GFP plus Ad-Cre (800 MOI each simultaneously), or Ad-IGF-IR_Δ950 plus Ad-Cre (800 MOI each simultaneously). Serum-starved cells were treated with vehicle or GH (250 ng/ml) for 10 min, after which detergent extracts were resolved by SDS-PAGE and sequentially immunoblotted with anti-pSTAT5, anti-STAT5, and anti-IGF-IRα. Lower panel, Densitometric quantitation of pSTAT5 signals from three independent experiments. In each experiment, the maximum signal was considered 100%. Data are plotted as mean ± se. P values are indicated. B, The experiment in A was repeated, except that Ad-IGF-IR_Δ950 was used at 8, 80, or 400 MOI. Note that rescue of GH-induced STAT5 activity was afforded by Ad-IGF-IR_Δ950, despite the reduced level of expression of the truncation mutant achieved with 8–400 MOI compared with 800 MOI used in A. C, Floxed IGF-IR primary osteoblasts were infected as in A. Serum-starved cells were treated with vehicle or GH (50 ng/ml) for 6 h. Total RNA was extracted, and quantitative real-time PCR for IGF-I (normalized for β-actin) was performed. Pooled data from two separate experiments (one in triplicate and one in duplicate) are represented as mean ± se. In each case, the value of the vehicle-treated samples is considered 100%. P values are indicated.

Figure 8

GH, IGF-I, and IGF-IR interrelationships (see text for details). A, Schematized somatomedin hypothesis of GH action. GH interacts with GHR, causing JAK2/STAT5 activation and IGF-I production. IGF-I then acts in an endocrine fashion by stimulating IGF-IR signaling. B, Schematization of our findings. GH promotes GHR-JAK2-IGF-IR association. IGF-IR acts as a component of GH signaling. In the presence of IGF-I, GH action is augmented.