Clinical Significance of Osteoprotegerin Expression in Human Colorectal Cancer

Colorectal cancer (CRC) is one of the most prevalent cancers in the world. It is the third most common cancer and the second most common cause of cancer-related mortality worldwide (1). The mortality rate has changed little over the last 30 years despite advances that have been made in understanding the pathogenesis of this cancer at the molecular level. Among patients who undergo curative surgery, some patients develop local recurrence or distant metastases leading to shorter survival (2). The tumor node metastasis (TNM) classification proposed by the International Union Against Cancer remains the indicator of prognosis and provides the basis for therapeutic decision making (3). However, the current TNM classification system is limited in that it cannot predict prognosis for individual patients. To improve the prognosis of CRC, there is a crucial need to explore cancer-related genes that can serve as predictive biomarkers for individualization of therapy (4).

Microarray analysis allows the exploration of several thousands of cancer-related or cancer-specific genes (5). The potential of expression profiling using microarray analysis as a tool to predict the prognosis for different types of cancer has been realized (6, 7). Data obtained from microarray analysis can provide significant insight into biological differences between patients with good and poor prognoses and can be used as a screening tool to find individual molecules for individualization of therapy.

Inactivation of the apoptotic pathway is an important mechanism that leads to tumor progression and tolerance to antitumor drugs (8). Thus, clarifying the mechanisms of apoptotic dysfunction may help provide new approaches for CRC treatment. In this study, we identified osteoprotegerin (OPG; TNFRSF11B) as a gene related to distant metastases in CRC patients. OPG helps cancer cells survive by providing resistance to apoptosis induced by the TNF-related apoptosis-inducing ligand (TRAIL; ref. 9). OPG is a secreted member of the TNF receptor superfamily that acts as a decoy receptor binding to TRAIL and neutralizing its apoptotic function (10). This is the first study that revealed the clinicopathologic significance of OPG expression by using clinical tissue samples from patients with CRC.

Primary tumors from 378 consecutive patients who underwent surgical therapy for CRC between 2002 to 2007 at Tokyo Medical and Dental University Hospital were included in this study. The Institutional Review Board approved the study, and written informed consent was obtained from all patients. A total of 104 patients were assigned to the microarray study including 13 patients with stage I, 37 patients with stage II, 34 patients with stage III, and 20 patients with stage IV disease. Metastatic recurrence after surgery occurred in 18 patients with stage I to III disease. The median follow-up time was 22 months for these patients (range = 1–34 months). We assigned a microarray study set to these patients to compare the expression in CRC tissues from 19 patients with noncancerous tissues from 25 patients using homogenized tissues. As a validation study, we carried out a quantitative reverse transcriptase PCR (RT-PCR) assay in 77 CRC patients and an immunohistochemical study in 274 CRC patients including patients who underwent RT-PCR assay. The patients enrolled in microarray analysis were excluded from these validation studies. In 274 patients assigned to validation studies, there were 50 patients with stage I, 88 patients with stage II, 88 patients with stage III, and 48 patients with stage IV disease. The median follow-up time for these patients was 47 months (range = 1–92 months).

Cancer tissues for laser capture microdissection were immediately embedded in Tissue-Tek OCT compound medium (Sakura Finetek Japan) after resection. Serial frozen sections of 9-μm thickness were mounted onto a foil-coated glass slide, 90 FOIL-SL25 (Leica Microsystems). Laser capture microdissection (LCM) was carried out using the Application Solutions Laser Capture Microdissection System (Leica Microsystems). Total RNA from LCM and bulk samples of cancer tissues and adjacent noncancerous tissues were extracted using the RNeasy mini Kit (QIAGEN) according to the manufacturer's protocol. The integrity of the total RNA obtained was assessed using an Agilent 2100 BioAnalyzer (Agilent Technologies). Samples with an RNA integrity number greater than 5.0 were analyzed further.

cRNA was prepared from 100 ng total RNA using 2-cycle target labeling and a control reagents kit (Affymetrix). Hybridization and signal detection of the Human Genome U133 Plus 2.0 arrays (Affymetrix) were carried out according to the manufacturer's instructions. The gene expression data sets were submitted to Gene Expression Omnibus (http://www.ncbi.nlm.nih.gov/geo/; accession number GSE21510). The gene expression data were then analyzed to identify genes related to distant metastases. We defined patients with synchronous or metachronous distant metastases as the metastatic group and patients without any distant metastases as the nonmetastatic group. The criteria to select candidate genes were as follows: (i) a higher expression level in cancer cells of the metastatic group than in the nonmetastatic group and (ii) a higher expression level in cancer tissues than in adjacent noncancerous tissues. A higher expression was defined as a numerical value 1.5 times greater than in another group.

For cDNA synthesis, 10 μg of total RNA was reverse transcribed into cDNA samples using High Capacity cDNA Reverse Transcription Kit (Applied Biosystems) according to the manufacturer's protocol. TaqMan gene expression assays (Applied Biosystems: TNFRSF11B; Hs00900360_m1, β-actin; Hs99999903_m1) were used to determine the expression of OPG. β-Actin was used as an internal control. The PCR reaction was carried out with the TaqMan Universal PCR Master Mix (Applied Biosystems). The thermal cycling conditions were as follows: 50°C for 2 minutes, 95°C for 10 minutes, and 40 cycles of 15-second denaturation at 95°C and 1-minute annealing at 60°C. All calculated concentrations of target genes were normalized by the amount of the endogenous reference with the comparative C_t method for relative quantification (ΔΔC_t method) using Relative Quantification Study Software (7300 Sequence Detection System version 1.2.1; Applied Biosystems). Each assay was done in duplicate.

Immunohistochemical studies of OPG were conducted on formalin-fixed, paraffin-embedded surgical sections. Antigen retrieval by microwave (Microwave-MI 77; Azuyama) pretreatment at 98°C was carried out for 30 minutes in a high pH target retrieval resolution (Dako). The endogenous peroxidase activity was stopped by 15-minute incubation in a mixture of 3% hydrogen peroxide solution in 100% methanol. After washing with PBS, the slide was incubated with goat polyclonal antibody at a 1:50 dilution against human OPG (Santa Cruz Biotechnology) for 20 minutes by microwave. Sections were incubated with labeled polymer [Histofine Simplestain Max PO (G); Nichirei] for 30 minutes at room temperature. Color development was carried out with DAB (0.02% 3,3′-diaminobenzidine tetrahydrochloride; Nichirei) for 40 minutes at room temperature. OPG expression was graded by 2 independent observers using a 4-point scale, where 0 = no staining, 1+ = weak staining, 2+ = moderate staining, and 3+ = strong staining.

Statistical analyses of microarray data were normalized using the robust multiarray average method with R 2.6.2 statistical software together with the BioConductor package (R Foundation for Statistical Computing), as described previously (11, 12). Differences between 2 groups were estimated using the Mann–Whitney U test for 104 LCM patients and for homogenized samples of whole tumor and tumor adjacent tissues. Biostatistical analyses of OPG expression were carried out with SPSS (version 17.0, SPSS Inc) for Windows software. Differences between groups were estimated using the Mann–Whitney U test and the χ² test. Correlation was estimated using the Spearman's rank correlation coefficient. Survival curves were estimated by the Kaplan–Meier method, and comparisons between curves were made by the log-rank test. Prognostic factors were examined by univariate and multivariate analyses (Cox proportional hazards model). A probability level of 0.05 was used for statistical significance.

In the microarray analysis, 6 genes (COL9A3, SERPINB5, HOXB8, OPG, STXBP1, and TWSC) were identified that fulfilled the specified criteria. Three genes were associated with human cancer: SERPINB5, HOXB8, and OPG. Among these genes, OPG was suspected to play a role in cell survival through its antiapoptotic action. The high expression of OPG has been shown to be associated with worse prognosis in several solid cancers (13–16), although not in CRC. Thus, we selected OPG as a candidate gene for further analysis.

Quantitative RT-PCR analyses of 77 patients with CRC showed that OPG mRNA expression levels were significantly higher in tumors than in corresponding normal tissues (P < 0.001; Fig. 1A). The expression of OPG mRNA was significantly higher in tumors in the metastatic group than in the nonmetastatic group (P = 0.035; Fig. 1B).

Figure 2 shows the results of immunohistochemical studies of OPG expression in representative cancer tissues. We examined the localization of the OPG protein in cancer tissues from the 77 cases that underwent RT-PCR analysis. OPG was detected in cytoplasm of cancer cells and lymphocytes and macrophages, whereas OPG protein was not detected in normal epithelial cells. All regions in the stroma also acquired negative staining except such as lymphocytes and macrophages. The staining was strong at the invasive tumor front of CRC, and a correlation was observed between OPG mRNA expression and protein expression at the invasive front in these cases (correlation coefficient r: 0.249, P = 0.029).

The experimental samples that underwent immunohistochemical analysis were divided into 2 groups: the low-expression group (staining score of 0 and 1+, n = 135) and the high-expression group (staining score of 2+ and 3+, n = 139). Table 1 shows the expression of OPG protein in cancer samples and clinicopathologic data of the 274 patients. The high-expression group had significant correlation deeper tumors (P = 0.002), lymphatic invasion (P = 0.026), lymph node metastasis (P = 0.001), high carcinoembryonic antigen (CEA; P < 0.001), and distant metastases (P < 0.001) than the low-expression group. OPG expression in stage—III to IV CRC patients was significantly higher than in stage—I to II patients in this study (P < 0.001). The overall survival (OS) rate was significantly lower (P < 0.001) in patients with high OPG protein expression than in those with low OPG expression (Fig. 3A). Table 2 shows the expression of OPG protein in cancer samples and the clinicopathologic data of 226 patients with stage I, II, and III disease. The high-expression group had trends toward higher CEA levels. There were no other significant differences in clinicopathologic features between the low- and high-expression groups among these 226 patients. The relapse-free survival rate (RFS) was significantly lower (P = 0.001) in patients with high OPG protein expression than in those with low OPG expression (Fig. 3B). Univariate analysis showed that the following factors were significantly associated with worse RFS: gender, histology, tumor depth, lymphatic invasion, venous invasion, lymph node metastasis, CEA level, and high OPG expression (Table 3). Multivariate analysis using the Cox proportional hazards model indicated that histology, tumor depth, lymph node metastasis, CEA level, and high OPG expression were associated with worse RFS.

OPG was initially identified as a protein that regulates bone resorption (17–20). It is a member of the TNF receptor superfamily, but it differs from other members due to a lack of a transmembrane domain. Because of the lack of this domain, OPG acts as a decoy receptor, binding to TRAIL and neutralizing its function (9, 10). TRAIL is the principle mediator of the extrinsic apoptotic pathway, which has tumor-killing activity, and is being investigated for its therapeutic potential to induce tumor cell death (21). TRAIL triggers apoptosis by binding to death receptors 4 and 5 on the surface of cells (22). OPG produced by tumor cells binds to TRAIL and prevents the activation of death receptors 4 and 5 (17). Thus, OPG is thought to protect against the antiapoptotic action in OPG-expressing cells by overcoming the mechanism induced by TRAIL. Indeed, some studies suggest that OPG production and release into the serum of patients is likely to exert an antiapoptotic effect on tumor cells during the invasion and metastasis process (23–25). In addition to the roles of OPG in tumor cell survival described above, OPG has been reported to be a positive regulator of tumor microvessel formation (26, 27). If OPG does have a significant function in the vasculature, this may have implications for tumor angiogenesis, a key process in cancer development and metastasis. OPG production is suggested to be part of a tumor cell survival strategy, and a number of different tumor cells, such as prostate, breast, and gastric cancer cells, have been found to produce OPG (13, 15, 16). However, the OPG expression status in CRC has only been analyzed in the cell line and serum (28, 29). It is reported that serum OPG levels were increased in patients with advanced CRC (28), but no correlation with prognosis has been shown. This is the first study to reveal the localization of OPG in CRC tissue and the relationship between OPG expression and prognosis in CRC patients. Some evidence suggests that the balance between pro- and antitumor factors at the invasive front of CRC may be decisive in determining tumor progression and the clinical outcome of patients with CRC (30, 31). Our results support the idea that overexpression of OPG at the invasive tumor front might play a critical role in the initiation of progression and metastasis of CRC.

Despite curative surgery for localized disease, approximately 40% of CRC patients will eventually relapse (32). Recently, some adjuvant chemotherapies for patients with CRC have been shown to be associated with improved cure rates after surgery (33). However, even though these chemotherapies have been shown to be effective, they are costly and are associated with severe adverse effects. Therefore, there exists a need to identify patients who would benefit from adjuvant chemotherapy prior to treatment; molecular markers may give us important insight into this challenge (34, 35), as these markers can be taken into account when selecting the appropriate chemotherapy. Our study suggests that high expression of OPG protein is a novel marker for recurrence after curative surgery for CRC. OPG might be a useful biomarker to decide whether to use adjuvant chemotherapy in patients with CRC after curative surgery.

In the last decade, the development of novel therapies that target critical biological pathways has greatly expanded treatment options for patients with CRC and has shown substantial improvement in survival (36). Antagonizing OPG by neutralizing antibodies might increase the sensitivity of many tumors to TRAIL. Moreover, because OPG is overexpressed only in CRC cells but not expressed in normal cells, it may be a novel target for cancer treatment without harming normal colorectal cells. This result provides hope regarding the development of novel target therapies without adverse effects. Death receptor agonists present a new class of therapeutics that selectively target apoptosis. Exogenous administration of TRAIL and agonistic antibody has been shown to be a rational and promising tumor therapy that overcomes resistance to apoptosis and has proven to be safe in early clinical trials (37, 38). Although several cancers are sensitive to TRAIL-induced apoptosis, others are resistant to TRAIL or acquire resistance during therapy (39). Overexpression of OPG might increase tolerance to TRAIL-induced apoptosis in cancer cells. Therefore, OPG blocking therapies might be a new strategy in the treatment of CRC.

In conclusion, high expression of OPG in cancer tissues is associated with biological aggressiveness and recurrence in patients with CRC after curative resection. OPG may become a new biomarker and a target of treatment for patients with CRC.

No potential conflicts of interest were disclosed.

The authors thank Y. Takagi and M. Itoda for excellent technical assistance.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.