OPG Is Regulated by β-Catenin and Mediates Resistance to TRAIL-Induced Apoptosis in Colon Cancer

Loss of sensitivity to apoptosis is a hallmark of cancer. The mechanism of how cells lose the ability to undergo apoptosis is thought to occur through several mechanisms, which include loss of membrane-bound death receptors and impairment of components of the extrinsic and intrinsic apoptotic pathways. This is supported by in vitro studies from cell lines as well as clinical studies showing a positive correlation between loss of proapoptotic membrane receptors and dedifferentiation, tumor size, poor prognosis, and tumor recurrence (1). Sensitization of cancer cells to proapoptotic stimuli including administration of apoptosis-inducing ligands such as TRAIL has therefore become a new approach in cancer therapy and proved safe in first clinical trials (2). Understanding the specific mechanisms leading to resistance to apoptosis will therefore allow for the development of new targeted therapeutic options. In particular, well-characterized signaling pathways known to play important roles in carcinogenesis should be considered concerning their contribution to cell immortalization by causing resistance to apoptosis. The Wnt/β-catenin signaling pathway is tightly regulated and has important functions in development, tissue homeostasis, and regeneration. Deregulation of the Wnt/β-catenin pathway is frequently found in various human cancers, and its activation is involved in 80% of colorectal cancers (3). However, the multiplicity of the mechanisms by which deranged Wnt/β-catenin signaling participates in carcinogenesis is not fully understood yet.

Osteoprotegerin (OPG; TNFRSF11b) was initially found to contribute to homeostasis of bone turnover due to its capability of binding to receptor activator of nuclear factor κB. Subsequent investigations revealed that OPG also acts as a decoy receptor and binds to TRAIL, neutralizing its function. Several reports suggest a role of OPG mediating resistance to apoptosis in cancer (reviewed by Holen et al.; ref. 4). OPG is thought to exert a protective antiapoptotic action in OPG-expressing tumors by overcoming the physiologic mechanism of tumor surveillance exerted by TRAIL. Here, we report on the expression, regulation, and function of OPG in colorectal cancer and on OPG serum levels in patients with various stages of colorectal cancer. We show that the Wnt/β-catenin pathway regulates OPG in colorectal cancer cells, that OPG mediates resistance to TRAIL-induced apoptosis, and that its serum levels are highly elevated in patients with advanced colorectal cancers.

Analysis of OPG in medium and serum. OPG concentrations in medium and serum of 40 healthy individuals, 22 patients with colonic adenomas, and 127 patients with colorectal cancers were assayed by ELISA (Raybiotech) according to the manufacturer's instructions. The patients with colorectal cancers included 10 patients with International Union Against Cancer (UICC) I, 34 patients with UICC II, 43 patients with UICC III, and 40 patients with UICC IV stage disease. Of the patients with stage IV disease, only one had overt bone metastases.

Cell culture. Cell lines were from American Type Culture Collection and cultured under standard conditions as recommended. To assess OPG production, cells were seeded in 12-well plates and grown for 3 d in the same medium. Medium from three different wells for assessment of triplicate values were harvested every 24 h.

Short interfering RNAs were purchased from Dharmacon as on-target validated siRNAs (Standard siGENOME siRNATM). Transfection of siRNAs was done with Lipofectamine 2000 (Invitrogen); controls were done with noncoding sequences for β-galactosidase (sense, UUAUGCCGAUCGCGUCACAUU; antisense, PUGUGACGCGAUCGGCAUAAUU) purchased from Dharmacon. Cells (1.5 × 10⁵) were seeded in 12-well plates and, after overnight incubation, transfected with Oligofectamine (Invitrogen) according to the protocol of the manufacturer. Human recombinant TRAIL was purchased from Sigma. Recombinant OPG and OPG antibody was from R&D. DNA fragmentation was quantified as the fraction of cells with subdiploid DNA content using flow cytometry as described (5). For preconditioned medium, 3 × 10⁶ cells were seeded in a 10-cm dish and let in culture for 5 d. OPG-depleted cell culture medium were obtained essentially as described (6). Briefly, preconditioned sera were centrifuged and filtered (0.2 μm; Sartorius) to remove cells from the supernatant, and incubated with antibodies against OPG (R&D). OPG immunocomplexes were removed by centrifugation after overnight incubation with Protein G Agarose (Upstate).

For reporter gene assays in HeLa cells, cells were transfected with 200 ng of a luciferase reporter gene construct for OPG (kindly provided by Rudolf Grosschedl, Max-Planck Institute for Immunobiology, Freiburg, Germany), 200 ng of constitutively active β-galactosidase reporter plasmid pCH110 (Amersham Pharmacia Biosciences), and the pcDNA3-based plasmids S33Y-β-catenin, Tcf4, and VP16Tcf (Kolligs Genes Dev, 2000). Transfections were done with Nanofectin I (PAA Laboratories) as recommended by the manufacturer, and luciferase activity was measured 48 h later (Promega) according to the manufacturer's instructions using a luminescence reader (Turner). To normalize luciferase activity, β-galactosidase activity was determined using standard methods.

Western blot analysis. Cell lysates were prepared with radioimmunoprecipitation assay lysis buffer (TBS, 0.5% deoxycholate, 0.1% SDS, and 1% Nonidet P-40), and equal amounts of protein were separated by electrophoresis in discontinuous SDS-polyacrylamide gels. The antibody incubation steps were done in TBS containing 0.05% Tween 20 and 5% nonfat dry milk. The anti–β-catenin antibody (Transduction Labs) and the anti–β-actin antibody (Sigma-Aldrich) were used at 1:2,500 and 1:10,000 dilutions, respectively. The secondary horseradish peroxidase–conjugated antibody goat antimouse (Amersham Biosciences) was used at 1:10,000. The blots were subjected to enhanced chemiluminescence (Amersham) and exposed to hyperfilm enhanced chemiluminescence (Amersham).

Quantitative reverse transcription-PCR. RNA was extracted from cells and tissues using Trizol reagent (Invitrogen). RNA was reverse transcribed after DNase treatment with random hexamer primers (final concentration, 2.5 ng/μL) using SuperScript II reverse transcriptase as described by the manufacturer (Invitrogen). The oligonucleotide primers used will be made available on request. PCR reactions were run on a Realplex4 Mastercycler Epgradient S (Eppendorf). Resulting PCR products were monitored by melting point analysis and agarose gel electrophoresis. PCR efficiency was determined by analyzing serial dilutions of cDNA.

Overexpression of OPG in colorectal cancer cells. OPG, a secreted member of the tumor necrosis factor receptor superfamily of receptors, has previously been reported to act as a soluble decoy receptor for receptor activator of nuclear factor κB ligand (RANKL; for review, see Theoleyre et al.; ref. 7). OPG prevents binding of RANKL to its receptor, receptor activator of nuclear factor κB, and inhibits osteoclast differentiation and, hereby, contributes to bone homeostasis. Subsequent studies showed that OPG also functions as a soluble receptor for TRAIL and exerts antiapoptotic functions in prostate and breast cancer cells (6, 8). As only very little is known about expression and function of OPG in colorectal cancer (9), we initially studied the expression of OPG in colorectal cancer cells. For this, we assessed OPG expression by reverse transcription-PCR in microdissected colorectal cancers and matched normal epithelium of the same patients (Fig. 1A) and in several colorectal cancer cell lines (Fig. 1B). OPG was expressed at higher levels in the four primary colorectal cancers than in matched normal colon epithelium and expressed in all colorectal cancer cell lines tested with the exception of RKO cells. Interestingly, in contrast to the other cell lines, RKO cells do not contain a mutation of downstream components of the Wnt-pathway. Correspondingly, OPG, secreted by colorectal cancer cells lines carrying mutations of the adenomatous polyposis coli gene, SW480, or β-catenin gene, HCT116, accumulated in cell culture medium (Fig. 1C).

OPG expression is regulated by β-catenin in colon cancer cells. As the Wnt/β-catenin pathway is deregulated in the majority of colorectal cancers and it has previously been reported that expression of OPG is regulated by β-catenin/Tcf in osteoblasts (10), we assessed whether Wnt/β-catenin contributes to the regulation of OPG in colorectal cancers. For this purpose, SW480 cells were transiently transfected with siRNA directed against β-catenin. This resulted in effective silencing of β-catenin expression on both protein and mRNA levels (Fig. 2A and B). Reduction of cellular β-catenin levels coincided with a strong reduction of OPG mRNA expression and expression of well-known target genes of β-catenin, including c-Myc and Dickkopf-1 (Fig. 2B; ref. 3). Correspondingly, only very little OPG accumulated in the supernatants of SW480 cells transfected with siRNA directed against β-catenin compared with cells transfected with control siRNA (Fig. 2C). Luciferase reporter assays confirmed that transcription from the human OPG promoter in epithelial cells is dependent on the presence of β-catenin and Tcf-4 (Fig. 2D). Strongest induction of reporter activity was observed upon transfection of a construct expressing a fusion of Tcf4 to the transactivation domain of VP16, further supporting the role of Tcf-4 in the regulation of OPG.

OPG protects colorectal cancer cells from TRAIL-induced apoptosis. TRAIL triggers apoptosis by binding to the death receptors DR4 and DR5 (11). Although TRAIL is present in tumors where it is produced by monocytes and is known as the principal mediator of acquired antitumor activity (12), many colorectal cancers show resistance to TRAIL due to various mechanisms including expression of antagonistic decoy receptors (for review, see van Geelen et al.; ref. 13). Thus far, three antagonistic TRAIL receptors, DcR1, DcR2, and OPG have been described (14). It has been reported that OPG serves as a survival factor for human prostate and mammary cancer cells (6, 8). We therefore assessed the role of OPG in mediating resistance of colorectal cancer cells to TRAIL-induced apoptosis. Four colorectal cancer cell lines were treated with increasing amounts of TRAIL, and a dose-dependent increase in apoptosis was observed (data not shown; Fig. 3A-C). Concomitant addition of OPG decreased TRAIL-induced apoptosis dependent on the dose of recombinant OPG added (data not shown; Fig. 3C). OPG was able to inhibit TRAIL-induced apoptosis in cells showing high (i.e., HCT116), intermediate (i.e., LoVo and SW480), and low (i.e., DLD1) sensitivity to TRAIL-induced apoptosis. In agreement with OPG specifically antagonizing TRAIL-induced apoptosis, apoptosis induced by treatment of HCT116 cells with increasing concentrations of etoposid could not be inhibited by OPG (data not shown). Moreover, removal of endogenous OPG secreted into the cell culture medium from HCT116 and DLD1 cells led to an increase in sensitivity to TRAIL-induced apoptosis by 18% to 19% (Fig. 3D). Finally, to link β-catenin expression to sensitivity to TRAIL-induced apoptosis in colorectal cancer cells, the expression of β-catenin was silenced by siRNA. This resulted in an increase in sensitivity to TRAIL-induced apoptosis (Fig. 3E). This supports the role of β-catenin as a regulator of OPG and inhibitor of TRAIL-induced apoptosis.

Serum OPG concentrations are increased in patients with metastasized colorectal cancers. Serum OPG concentrations have been reported to be increased in patients with prostate cancer, which progressed after androgen ablation therapy (15), and in patients with colorectal and pancreatic cancers compared with healthy individuals (16). To better understand the expression of OPG in patients with colorectal cancers, we assessed OPG concentration in sera from healthy individuals, patients with colonic adenomas, and patients with various stages of colorectal cancers. There was no significant difference in age between healthy individuals and cancer patients. The mean serum concentration in healthy individuals was 512 pg/mL; in patients with colorectal cancer, serum OPG levels were found to be in the range between 700 and 1,453 pg/mL (Fig. 4). We observed a significant increase in serum OPG concentrations in patients with UICC stage III and IV cancer compared with healthy individuals (P = 0.0359 and P < 0.0001, respectively). Moreover, a significant increase in OPG serum levels was observed in patients with metastatic disease compared with patients with locally advanced disease (UICC stage IV versus stage III; P < 0,00001). The majority of the stage IV patients had liver metastases, only one patient had overt bone metastases, ruling out that bone metastasis might be the underlying cause for elevated OPG serum levels.

The fact that OPG is regulated by β-catenin/Tcf-4 in colorectal cancer underlines the importance of this pathway in providing cancer cells with growth as well as survival signals allowing for proliferation and evasion from apoptosis (17, 18). Among the factors regulated by Wnt/β-catenin that function in the regulation of apoptosis, e.g., Fas antigen, Caspase 3, Clusterin, Survivin, and Cyclooxygenase-2 (for review, see Ilyas; ref. 18), OPG as an extracellular factor has the potential to become a therapeutic target. Targeting of deregulated β-catenin/Tcf signaling in cancer has thus far been hampered by the fact that the Wnt-pathway physiologically functions in tissue homeostasis and regeneration of several tissues including intestinal epithelium, skin, and hematopoetic cells (3, 17, 19). Therefore, an approach targeting central signaling components such as β-catenin and Tcf-4 might not be feasible. A strategy interfering with target genes of the Wnt-pathway, which have critical functions in cancer, might prove more promising. Exogenous administration of TRAIL has been shown to be a rational and promising tumor therapy that overcomes resistance to apoptosis and has proven to be safe in first clinical trials (2). Although several cancers are sensitive to TRAIL-induced apoptosis, others are resistant to TRAIL or acquire resistance during therapy (14). Antagonizing antiapoptotic factors such as OPG by neutralizing antibodies might therefore extend the number of tumors sensitive to TRAIL. Moreover, determination of OPG serum concentrations in patients could serve as a tool to screen for OPG-producing tumors and might be useful as a predictive marker for TRAIL resistance.

Our results indicate that production and release of OPG into the serum of patients affected by late stage metastatic tumors is likely to exert a protective antiapoptotic effect to tumor cells when these invade and metastasize. This systemic increase of OPG in patients affected by late-stage colorectal cancers is in accordance with data showing that loss of receptor-mediated apoptosis is a hallmark of cancer correlating with recurrence rates and poor prognosis (20). This notion is further supported by the fact that TRAIL receptor knockout mice or mice in which TRAIL receptors were blocked exhibited enhanced tumor formation and formation of metastases (21). Moreover, expression of OPG correlated with depth of tumor invasion, nodal metastases, and advanced tumor stage in patients with gastric carcinoma (22). In conclusion, we provide first evidence that Wnt/β-catenin signaling contributes to resistance to apoptosis by regulating OPG secretion in colorectal cancer. Our data suggest that inhibition of OPG expression might sensitize colorectal cancers to endogenous TRAIL produced by monocytes as well as exogenous administration of TRAIL. Determination of OPG serum levels might prove a useful biomarker that helps identify patients with TRAIL-resistance due to OPG overexpression.

No potential conflicts of interest were disclosed.

We thank Rudolf Grosschedl for providing the OPG reporter construct.