Expression of the IL-11 Gene in Metastatic Cells Is Supported by Runx2-Smad and Runx2-cJun Complexes Induced by TGFβ1

CELL LINES AND CELL CULTURE

Three PC cell lines were used in these studies to model PC progression during tumor growth in bone. LNCaP cells that are lymph node, but not bone, metastatic were used as a control cell line; PC3-L cells that produce mixed osteolytic and osteblastic lesions and have low Runx2 levels; and PC3-H cells that have high Runx2 levels and exhibit aggressive osteolytic disease in mouse models, followed by mixed osteolytic and osteblastic lesions. Microsatellite analyses carried out by the University of Vermont DNA Analysis Facility were used to identify the genotype as authentic LNCaP and/or PC3 cells [Zhang et al., 2015].

LNCaP cells and PC3-L cells were cultured in RPMI 1640 with 10% FBS, 10 mM non-essential amino acids, 2 mM L-glutamine and 1 mM sodium pyruvate. PC3-H cells were cultured in T-medium with 5% fetal bovine serum (FBS) [Huang et al., 2005]. All media were supplemented with 100 U/ml penicillin and 100 µg/ml streptomycin. Cell culture media and supplements were obtained from Invitrogen, Carlsbad, CA, with the exception of FBS, which was obtained from Atlanta Biologicals, Norcross, GA.

TGFβ1 AND BMP2 TREATMENT

For experiments involving growth factor additions, sub-confluent cell layers were first cultured in 1% charcoal-stripped media (Life Technologies, Carlsbad, CA) for 24 h. Some cultures were treated with the TGFβ inhibitor SB431542 at 5 µM for 1 h pre-incubation prior to TGFβ1 treatment, where indicated. Treatment was for 24 h, with vehicle control (DMSO), porcine TGFβ1 (10 ng/ml), or BMP2 (100 ng/ml) (R&D Systems, Minneapolis, MN). Cells were then harvested for protein detection by Western blot and for mRNA levels by qPCR.

WESTERN BLOT ANALYSIS

Cells were lysed in RIPA buffer (50 mMTris-HCl [pH 7.5], 150 mM NaCl, 1 mM Na₂EDTA, 0.1% SDS, 1% NP-40, 0.5% sodium deoxycholate) containing 25 mM MG132, EDTA-free cOmplete Protease Inhibitor Cocktail (Roche, Nutley, NJ) and 1 mM PMSF. Proteins were resolved by SDS-PAGE and transferred to PVDF membranes (EMD Millipore, Billerica, MA). Membranes were incubated with mouse anti-Runx2 monoclonal (MBL International Corporation, Woburn, MA), rabbit anti-Smad2/3 (Cell Signaling Technology, Danvers, MA), or rabbit anti-phospho-Smad2, rabbit anti-phospho-Smad3, rabbit anti-c-Jun (Cell Signaling Technology), rabbit anti-cdk2 polyclonal antibody (Santa Cruz Biotechnology, Dallas, TX). Proteins were detected using Clarity^TM Western ECL Substrate (Bio-Rad Laboratories, Hercules, CA).

REVERSE TRANSCRIPTION-QUANTITATIVE PCR (QPCR)

Total RNA was isolated from cells using Trizol reagent according to the manufacturer's protocol (Invitrogen) and then further purified using a DNA-Free RNA kit (Zymo Research, Irvine, CA). cDNA was synthesized using Superscript First-Strand Synthesis System (Invitrogen). qPCR was performed using SYBR Green Master Mix (Bio-Rad) and gene-specific primers shown in Table S1 in an ABI Prism 7000 real-time PCR instrument (Life Technologies). Data were normalized using human glyceraldehyde 3-phosphate dehydrogenase (GAPDH).

IL-11 ELISA ASSAY

Supernatants collected from cell cultures were analyzed for secreted IL-11 expression using the Human IL-11 Quantikine kit (R&D Systems), according to manufacturer's instructions. All samples were measured in duplicate and a biological replicate assay was carried out.

siRNA TRANSFECTION

Cells at approximately 60–70% confluency were transfected with 50 nM siRNA targeting Runx2 or with 50 nM non-silencing control siRNA (NS siRNA) using Oligofectamine™ Transfection Reagent (Life Technologies) according to manufacturer's instructions. “Control” transfections, in which cells were exposed to Lipofectamine 2000 only, were also included where indicated. mRNA and protein knockdown was assessed 48 h after transfection.

ADENOVIRUS INFECTION

For studies involving expression of Runx2 wild-type (WT) and mutant proteins, adenoviral vectors containing Runx2 WT-IRES-GFP, Runx2-HTY-IRES-GFP, and Runx2-ΔC-IRES-GFP, and GFP-only (infection-efficiency control) constructs under the control of the CMV5 promoter were delivered to low Runx2 expressing PC-L cells using previously described methods [Afzal et al., 2005]. Preparation and purification of virus were performed using the Adeno-X Maxi Purification Kit (Clonetech, Mountain View, CA), according to the manufacturer protocol. After infection, cells were washed twice with serum-free medium, incubated for 24 h, and prepared for downstream analysis or cultured for TGFβ1 treatment, as described above.

CHROMATIN IMMUMOPRECIPITATION (CHIP)

For ChIP experiments, approximately 1 × 10⁷ PC3-H cells with high endogenous Runx2 expression were either single-crosslinked in 1% formaldehyde for 10 min at room temperature (RT) or dual-crosslinked in 2 mM disuccinimidyl glutarate (DSG) for 30 min with agitation at RT, washed with PBS and then treated with 1% formaldehyde for 10 min at RT with rotation. Formaldehyde crosslinking was halted by addition of 2.5 M Glycine to a final concentration of 125 mM. Crosslinked cells were then washed with ice-cold PBS, harvested by scraping, pelleted, and resuspended in nuclei isolation buffer A (50 mM HEPES pH 7.5, 140 mM NaCl, 1 mM EDTA, 10% glycerol, 0.5% IGEPAL CA-630. 0.25% Triton-X100, Roche cOmplete protease inhibitor) for 10 min at 4°C with constant rotation. Nuclei were then pelleted by centrifugation, washed in washing buffer B (10 mM Tris-HCL pH 8.0, 200 mM NaCl, 1 mM EDTA, 1 mM EGTA, Roche cOmplete protease inhibitor), collected by centrifugation and resuspended in lysis buffer C (10 mM Tris-HCL pH 8.0, 100 mM NaCl, 1 mM EDTA, 1 mM EGTA, 0.1% sodium deoxycholate, 0.5% N-laurylsarcosine Roche cOmplete protease inhibitor). Samples were then sonicated for 7 min (formaldehyde crosslinking) or 15 min (dual-crosslinking) using a Covaris S220 focused-ultrasonicator (Covaris, Woburn, MA) to reduce chromatin to an average size of 0.3 kb. Chromatin was then incubated overnight with Runx2 antibody (M-70, Santa Cruz Biotechnology), rabbit c-Jun antibody (Cell Signaling Technology) or rabbit IgG (EMD Millipore) at 4°C followed by Protein A Dynabeads (Life Technologies). Immune complexes were washed three times with low salt buffer (20 mM Tris-HCl pH 8.0, 1 mM EDTA, 0.1% SDS, 1% Triton-X100, 150 mM NaCl) once with high salt buffer (20 mM Tris-HCl pH 8.0, 1 mM EDTA, 0.1% SDS, 1% Triton-X100, 500 mM NaCl) and once with TEN buffer (50 mM Tris-HCl pH 8.0, 1 mM EDTA, 50 mM NaCl) and eluted into elution buffer (50 mM Tris-HCl pH 8.0, 10 mM EDTA, 1% SDS) at 55°C for 30 min with agitation. Chromatin was then treated with RNase A for 2 h followed by Proteinase K for 2 h at 55°C. DNA was then recovered by phenol/chloroform/isoamyl alcohol extraction followed by ethanol precipitation. Recovered DNA was then used as a template in qPCR reactions using primer pairs specific for each gene of interest (Table S2).

CO-IMUNOPRECIPITATION

PC3-H cells (TGFβ1 treated or not treated) were scraped into ice-cold PBS and pelleted by centrifugation. Cells were then lysed with lysis buffer (150 mM NaCl, 50 mM Tris, pH 8.0, 1% IGEPAL CA-630, EDTA-free cOmplete protease inhibitor cocktail [Roche], 25 mM MG132) on ice for 20 min and centrifuged to pellet insoluble cellular debris. Runx2 M70 antibody (Santa Cruz Biotechnology) or rabbit IgG (EMD Millipore) was added to supernatant and incubated at 4°C overnight with agitation. Fifty microliter protein A/G PLUS Agarose beads (Santa Cruz Biotechnology) were then added and incubated at 4°C for 3 h. Protein-antibody complexes were then collected by centrifugation, beads were washed four times with lysis buffer, resuspended in 2 × SDS sample buffer (Bio-Rad Laboratories), containing 5% β-mercaptoethanol and boiled for 5 min. immunoprecipated proteins were separated by SDS–PAGE and transferred to PVDF membranes for Western blot analysis.

IMMUNOFLUORESCENCE MICROSCOPY

Cells were grown on gelatin-coated coverslips and processed for immunofluorescence microscopy as described previously [Young et al., 2007]. Antibodies were diluted in 1% BSA in PBS. The following primary antibodies were used in immunofluorescence experiments: mouse monoclonal RUNX2 (MBL International Corporation), goat polyclonal Smad3 (Santa Cruz Biotechnology), and rabbit polyclonal c-Jun (described above). Secondary antibodies used were goat anti-mouse-IgG conjugated to Alexa Fluor 488, donkey anti-goat-IgG conjugated to Alexa Fluor 594, and goat anti-mouse-IgG conjugated to Alexa Fluor 594, all from Life Technologies. ProLong gold antifade reagent with DAPI (Life Technologies) was added to the cells before they were visualized. The cells were examined by fluorescence microscopy and images were captured and processed using Zen pro2011 (Carl Zeiss Microscopy, Thornwood, NY). Immunofluorescence signals were quantified by using the ImageJ software package (National Institutes of Health, Bethesda, MD).

STATISTICAL ANALYSIS

Data were expressed as the mean ± standard error of the mean (SEM). Statistically significant differences were determined using the Student's t-test with Welch's correction. P < 0.05 was considered statistically significant.

IL-11 EXPRESSION RELATED TO TGFβ1 AND RUNX2 SIGNALING

To begin to elucidate mechanisms involving both Runx2 and IL-11 in mediating prostate cancer metastatic disease, we initially characterized the relationship between Runx2 and expression of TGFβ1 and IL-11 among three prostate cancer cell lines, PC3-L, PC3-H, and LNCaP (Fig. 1). qPCR results showed high levels and a positive correlation between Runx2, IL-11 and TGFβ1 expression in PC3-H cells (known to have aggressive osteolytic properties in vivo). In contrast, PC3-L cells have significantly lower Runx2 and TGFβ1 levels and nearly undetectable IL-11 expression compared to PC-3H cells (Fig. 1A). Notably, LNCaP, a non-bone metastatic cell, exhibited no detectable levels of all three genes. Expression of Runx2 is 7-fold, IL-11 ∼25-fold, and TGFβ1 4-fold higher in PC3-H than in PC3-L cells. We also confirmed expression of known osteolytic genes in these cells, RANKL and MMP2 (data not shown, as reported in [Dutta et al., 2014]), as well as PTHrP, CSF2, and CTGF (Supplementary Fig. S1). These genes also correlated with IL-11 and Runx2 expression in PC3-H cells.

TGFβ1 and BMPs are secreted during bone resorption and increase tumor cell secretion of osteolytic factors and osteoblastic factors that stimulate stromal cells to cause pathologic bone formation. The response of IL-11 and Runx2 to TGFβ1 and BMP2 was examined in PC3-L cells using molecular markers: PAI-1 for TGFβ1 activity and ID1, a target gene of BMP2 stimulation (Fig. 1B). TGFβ1 modestly increased Runx2 expression in PC3-L cells (1.5-fold). In striking contrast, IL-11 gene expression was induced significantly by TGFβ1 (7-fold), although not as high as the PAI-1 marker gene of TGFβ activity (20-fold upregulated). BMP2 treatment had no significant effect on Runx2 or IL-11 compared to non-treated PC3-L cells with the exception of ID1, as expected (Fig. 1B). We find that TGFβ-induced expression of IL-11 and PAI-1 is almost completely inhibited by the TGFβ1 inhibitor SB431542 (in the presence or absence of exogenous TGFβ1 treatment) (Fig. 1C). In contrast, the same inhibitor had no significant effect on Runx2 expression. In addition, Western blot results showed that TGFβ1 induced phosphorylation of its co-receptor Smad2, which was completely inhibited by SB431542 (Fig. 1D). Furthermore, an ELISA assay showed secretion of IL-11 in response to TGFβ1 was prevented by the inhibitor (Fig. 1E). Together these findings show a positive correlation between IL-11, Runx2 and TGFβ1 expression, with robust induction of IL-11, but only a modest increase in Runx2 in response to TGFβ1.

EXPRESSION OF OSTEOLYTIC GENES PTHrP AND RANKL RESPOND TO CHANGES IN IL-11

To further investigate the role of IL-11 in the context of the “vicious cycle” of osteolytic bone disease, we addressed whether changes in IL-11 affected the expression of PTHrP and RANKL. Consistent with the finding that IL-11 is induced by TGFβ1 in PC3-L cells (Fig. 1B), both PTHrP and RANKL expression can also be induced by TGFβ1 in the same cells (Fig. 2A). To determine the relationship between IL-11 expression and these osteolytic factors, IL-11 was selectively inhibited by RNA interference in PC3-H cells, which express high endogenous IL-11 levels (Fig. 2B). Effective knockdown (60–80%) of IL-11 was confirmed by assaying secreted IL-11 (Fig. 2B, left panel) and IL-11 mRNA expression (Fig. 2B, right panel). As a result, the expression of the osteolytic factors PTHrP and RANKL were significantly decreased upon IL-11 depletion (P < 0.05) (Fig. 2C). Thus, in support of the hypothesized functional relationship between Runx2 and IL-11, we find that the same osteolytic genes increased by Runx2 transcription, PTHrP and RANKL, are also responsive to IL-11 knockdown. These results suggest a functional cooperation between Runx2 and IL-11 mediated by TGFβ1 in advanced prostate cancer cell lines.

IL-11 EXPRESSION IS RELATED TO RUNX2-SMAD SIGNALING IN PROSTATE CANCER PC3 CELLS

To address the functional relationship between Runx2 and IL-11 gene expression, we first determined if Runx2 regulates IL-11. Runx2 was exogenously expressed in PC3-L cells (having low endogenous Runx2 levels) by adenovirus infection (Fig. 3A). The results showed that Runx2 can activate IL-11 gene expression in PC3-L cells (Fig. 3B). We then selectively inhibited Runx2 expression by RNA interference in PC3-H cells (Fig. 3C). Accordingly, Runx2 protein expression was almost completely depleted, along with an 80% reduction in Runx2 mRNA levels. This knockdown of Runx2 reduced IL-11 gene expression by 40–50%, again supporting a potential role for Runx2 in regulating IL-11 (Fig. 3D).

Our previous studies showed that Runx2 enhances the TGFβ1 response in cancer cells through formation of a Runx2-SMAD complex (reviewed in Introduction). To determine whether this mechanism is also involved in IL-11 gene regulation, we utilized constructs expressing either wild-type Runx2 or two mutations (Runx2-HTY and Runx2-ΔC) that abrogate Runx2-Smad signaling [Zhang et al., 2015]. The two Runx2 mutant proteins retain the DNA binding domain, but are not fully functional [Zhang et al., 2015]. Importantly, TGFβ1 treatment efficiently induced phosphorylation of Smad2 in cells expressing WT Runx2 and the mutant proteins (Fig. 3E). When the Runx2 mutant proteins were overexpressed in PC3-L cells, IL-11 gene expression was reduced by 50% compared to Runx2 WT (Supplementary Fig. S2 and Fig. 3F). In these same cells, in the presence of exogenous expression of WT Runx2, IL-11 expression increased ∼eightfold. However, exogenous expression of Runx2-HTY and -ΔC mutant proteins exhibited a 50% reduction in TGFβ1-stimulated activation of the IL-11 gene when compared with controls (Fig. 3F). These data strongly implicate TGFβ-Runx2-Smad signaling in regulation of IL-11 gene expression in advanced prostate cancer cells, in addition to the known TGFβ-AP1 pathway that increases IL-11 expression in numerous cell types.

RUNX2 AND C-JUN FORM A COMPLEX FOR UPREGULATION OF IL-11 GENE EXPRESSION

The above findings prompted an examination of direct binding of Runx2 in the IL-11 gene promoter, but only a single Runx2 site resides in the proximal promoter in the vicinity of two previously characterized AP-1 sites (Fig. 4A). Multiple Smad binding elements and several additional AP-1 consensus sites are located within 5 kb of the IL-11 promoter region. Even after optimizing several experimental conditions (e.g., testing of two lots of M-70 Runx2 antibody from Santa Cruz Biotechnology, lot #s E2410 and F0112), we could not detect direct Runx2 binding to the IL-11 gene. Supplementary Figure S3 displays the absence of Runx2 binding at several promoter regions enriched in Smad sites. Therefore, we considered the possibility of a Runx2-c-Jun complex involved in regulating IL-11 gene expression. Several members of the AP-1 family of transcription factors are proven regulators of the IL-11 gene dependent on the cell type [Yang and Yang, 1994; Tang et al., 1998b; Bamba et al., 2003; Tohjima et al., 2003]. Previous studies showed that Runx2 can augment c-Jun activation of the AP-1 binding site in the collagenase-3 promoter [Hess et al., 2001; D'Alonzo et al., 2002] and also regulates PTH [Selvamurugan et al., 1998]. Chromatin immunoprecipitation (ChIP) was therefore pursued with a c-Jun antibody which confirmed c-Jun binding within the proximal, as well as in the distal IL-11 promoter regions (Fig. 4B).

We further interrogated whether Runx2 regulated the IL-11 gene through direct interaction with c-Jun under TGFβ1 stimulation. In response to TGFβ1, protein levels of Runx2 and c-Jun remained unchanged; however, levels of phosphorylated Smad were noticeably increased demonstrating activation of TGFβ1 signaling (Fig. 5A). In PC3-H cells, with and without TGFβ1 treatment, we examined formation of a Runx2-c-Jun protein-protein complex. In Fig. 5B, results of a co-immunoprecipitation show an interaction between Runx2 and c-Jun in PC3-H cells that is amplified in the presence of TGFβ1. Validation that the Runx2-c-Jun complex contributes to IL-11 expression is shown by a decrease in IL-11 protein levels upon Runx2 depletion in the presence of constitutively high c-Jun protein (Fig. 5C). Data derived from immunofluorescence (IF) microscopy studies revealed nuclear co-localization of Runx2 and c-Jun in PC3-H cells, which was diminished by Runx2 knockdown (Fig. 5D). Quantifying the IF intensity corroborated these observations, with the presence of Runx2, as well as c-Jun, significantly decreasing after Runx2 depletion (Fig. 5E). Together these data provide compelling evidence for formation of a nuclear Runx2-c-Jun complex.

The summation of our findings supported by biochemical, molecular and in situ cellular approaches, is that a circuitry exists in prostate tumor cells whereby TGFβ1 signals to: (1) AP-1 activity for induced levels of IL-11 by Runx2-c-Jun; and (2) the formation of Runx2-Smad complexes that induce osteolytic factors PTHrP, RANKL. These findings reveal the cooperativity of multiple transcriptional regulators of the IL-11 gene induced by TGFβ1 that contribute to metastatic bone disease.