Insulin-like growth factor I suppresses bone morphogenetic protein signaling in prostate cancer cells by activating mTOR signaling

Abstract

Insulin-like growth factor (IGF) I and bone morphogenetic proteins (BMP) are critical regulators of prostate tumor cell growth. In this report, we offer evidence that a critical support of IGF-I in prostate cancer is mediated by its ability to suppress BMP4-induced apoptosis and Smad-mediated gene expression. Suppression of BMP4 signaling by IGF-I was reversed by chemical inhibitors of phosphoinositide 3-kinase (PI3K), Akt, or mTOR; by enforced expression of wild-type PTEN or dominant-negative PI3K; or by small hairpin RNA-mediated silencing of mTORC1/2 subunits Raptor or Rictor. Similarly, IGF-I suppressed BMP4-induced transcription of the Id1, Id2, and Id3 genes that are crucially involved in prostate tumor progression through PI3K-dependent and mTORC1/2-dependent mechanisms. Immunohistochemical analysis of non-malignant and malignant prostate tissues offered in vivo support for our model that IGF-I-mediated activation of mTOR suppresses phosphorylation of the BMP-activated Smad transcription factors. Our results offer the first evidence that IGF-I signaling through mTORC1/2 is a key homeostatic regulator of BMP4 function in prostate epithelial cells, acting at two levels to repress both the proapoptotic and pro-oncogenic signals of BMP-activated Smads. We suggest that deregulation of this homeostatic control may be pivotal to the development and progression of prostate cancer, providing important implications and new potential targets for the therapeutic intervention of this malignancy.

Figure 1. Biological activity of TGF-β superfamily ligands on prostate epithelial cell lines

A, NRP-152 cells were treated ±TGF-β1 (0-10 ng/ml), Activin (A, B, AB) (10 ng/ml), BMP4 (0-10 ng/ml), or MIS (10 ng/ml), Nodal (10 ng/ml), Cripto (10 ng/ml) for 24 h treatment and analyzed for Smad activation by Western blot using antibodies against the two c-terminal serines of phospho-Smads 1, 3, 5, 8 (Ab #1), phospho-Smad1/5/8 (Ab #2), and phospho-Smad2. B, NRP-152, DP-153, LNCaP, PC3, and DU145 cells treated with BMP4 (0-20 ng/ml) for 72 h and total adherent cells were enumerated using a Coulter Electronic counter. Values represent averages of triplicate determinations ± S.E.

Figure 2. LR³-IGF-I blocks BMP4-induced cell death in non-tumorigenic (NRP-152 and DP-153) and tumorigenic(LNCaP and VCaP) prostate epithelial cancer cell lines

A, NRP-152 cells treated with ±BMP4 (10 ng/ml) for 24-72 h and total adherent cells were measured using a Coulter counter. B, NRP-152 cells were treated with ±2 nM or 10 nM LR³-IGF-I for 24 h followed by ±BMP4 (10 ng/ml) for 72 h after which total adherent cells were counted using a Coulter Electronic counter. C, NRP-152 cells were treated with ±10 nM LR³-IGF-I for 24 h followed by ±5 ng/ml BMP4 for an additional 72 h and stained with Hoechst dye (left), Trypan Blue (middle) Flow Cytometry (right). D, NRP-152, DP-153, LNCaP, and VCaP cells were treated as C and examined cell number using Coulter Electronic counter or examined by phase-contrast microscopy (200X) for changes in morphology. Columns (A-D) are the average of triplicate determinants or three independent experiments ± S.E. * P<0.001.

Figure 3. LR³-IGF-I abrogates BMP4-induced activation of Smad1/5/8, and Id-1, -2, and Id-3 expression

A, NRP-152 cells were treated with ±LR³-IGF-I (10 nM) for 24 h followed by ±BMP4 (10 ng/ml) for 4 h and cell lysates were analyzed by Western blot (top), or NRP-152 cells were co-transfected with 25 ng of CMV-Renilla reporter construct and 1 μg of Id-1-luciferase reporter element 24 h prior ±LR³-IGF-I (10 nM, 24 h) and then ±BMP4 (10 ng/ml, 4 h) treatments. Dual luciferase activity was then assayed and relative values of firefly luciferase were normalized to renilla luciferase (bottom). Columns, average of triplicate determinations; bar, ±S.E. B, Expression of Id-1, Id-2 and Id-3 mRNAs in NRP-152 (B) or LNCaP (C) cells treated with ±LR³-IGF-I (10 nM) for 24 h followed by ±BMP4 (10 ng/ml) for 4 h. C, RWPE-1 and DU-145 were treated as specified in 3a and cell lysates were analyzed by Western blot for Phospho-Smad1/5/8 activation and ID-1 expression. D, Real-time quantitiative PCR (RT-q-PCR) examined expression of ID-1 mRNA in NRP-152 cells ±LR³-IGF-I (10 nM) for 24 h followed by ±BMP4 (10 ng/ml) for a total of 48 h and semi-quantitative PRC is below. Data is representative of three independent experiments.

Figure 4. LR³-IGF-I inhibits BMP4-mediated responses through a PI3K/Akt/mTOR-dependent mechanism

A (left), NRP-152 cells were transfected with 0.8 μg of expression constructs for control (pSG5), DN-PI3K or CA-PI3K and co-transfected with Id-1-luciferase as described above for 24 h then treated with ±LR³-IGF-I (10 nM) or vehicle for 24 h prior to ±BMP4 (5 ng/ml), and luciferase activity was measured after 24 h. A (right), NRP-152 cells were co-transfected with 20 ng of CMV-Renilla reporter and 1 μg of Id-1-luciferase constructs, and 24 h later cells were incubated with ±LY294002 (10 μM) for 2 h, followed by ±LR³-IGF-I (10 nM) or vehicle for 24 h. Cells were then treated ±BMP4 (5 ng/ml) and luciferase activity measured 24 h. B and C, NRP-152 cells were transfected with Id-1-luciferase reporter element as described in (B) then incubated with either ±perifosine (10 nM) or ±rapamycin (200 nM) for 2 h, followed by ±LR³-IGF-I (10 nM) for 24 h. Cells were then treated with ±BMP4 and assayed for luciferase 2h later. D, NRP-152 cells were pre-treated with 10μM LY294002 or 200 nM rapamycin for 2 h followed by ±LR³-IGF-I (10 nM) or vehicle for 24 h, then treated ±BMP4 (5 ng/ml) for 4 h. Western blot analysis was conducted for P-Smad1/5/8 (Ab#1 or Ab#2) or total Smad1/5/8 expression. Data are representative of two to three independent experiments. Columns, average of triplicate determinants; bar, ±S.E.

Figure 5. Raptor, Rictor and mTOR mediate the IGF-I suppression of BMP-induced Id-1 promoter expression in NRP-152 prostate epithelial cells

A, Raptor, Rictor and mTOR were effectively silenced individually as indicated at the protein level in NRP-152 cells. B and C, NRP-152-tTR-sh-LacZ, NRP-152-tTR-sh-Raptor or stably silenced NRP-152-sh-Raptor, NRP-152-sh-mTOR cells were treated with LR³-IGF-I (10 nM) 24 h prior to BMP4 (5 ng/mL) for an additional 4 h, and cells were then lysed for Western blot analysis of phospho- and/or total- Smads. D, NRP-152-tTR-sh-LacZ (Sh-LacZ), NRP-152-tTR-sh-mTOR (Sh-mTOR), NRP-152-tTR-sh-Raptor (Sh-Raptor), NRP-152-tTR-sh-Rictor (Sh-Rictor) stably silenced cells were transfected with Id-1-promoter construct 24 h prior to treatment with LR³-IGF-I (10 nM). Then after 2h cells were treated with BMP-4 (5 ng/ml) or vehicle and luciferase activity was measured 24 h later. Columns, average of triplicate determinants; bar, ±S.E.

Figure 6. IGF-I mediated inhibition of BMP-induced genes microarray analysis and In vivo examination of mTOR-mediated inhibition of Smad1/5/8 in advanced human prostate adenocarcinoma

A, Microarray analysis of NRP-152 cell treated with vehicle control, LR³-IGF-I (10 nM, 24 h) +/- BMP-4 (5 ng/ml) for total 48 h and analyzed to determine fold change relative tocontrol and biological process identified with Pathway Studio 5.0. B, Immunohistochemistry of normal prostate hyperplasia (top) or advanced prostate adenocarcinoma stage III stained with H&E, Phospho-Smad1/5/8 or Phospho-S6. C, Matched human prostate cancer cores (34 total cores) H score plotted, R²=0.431 and P=<0.0001 (left) or bar chart depicting each sequential core expression of P-Smad1/5/8 or P-S6. D, A schematic model of IGF-I regulation of BMP signaling and its implication in prostate cancer.