Death receptor 5 (DR5) and a 5-gene apoptotic biomarker panel with significant differential diagnostic potential in colorectal cancer

High expression of Inhibitor of apoptosis proteins (IAPs) has been related to colorectal cancer (CRC) progression, resistance to treatment and poor prognosis. TRAIL (TNF-related apoptosis-inducing ligand) through its receptors DR4 (TRAIL-R1) and DR5 (TRAIL-R2) can selectively induce cancer cell apoptosis. The mRNA expression of DR4, DR5, c-IAP1, c-IAP2, XIAP and BIRC5/Survivin genes was examined in 100 paired (cancerous-normal) colorectal tissue specimens by real-time PCR, 50 of which were KRAS wild-type and 50 KRAS-mutant. DR5, XIAP and BIRC5/Survivin genes are significantly up-regulated (p < 0.0001, p = 0.012 and p = 0.0003, respectively), whereas c-IAP1 and c-IAP2 genes are significantly down-regulated at mRNA and protein levels in CRC (p < 0.0001 for both). ROC analyses showed that DR5, cIAP1 and cIAP2 expression has discriminatory value between CRC and normal tissue (AUC = 0.700, p < 0.0001 for DR5; AUC = 0.628, p = 0.011 for cIAP1; AUC = 0.673, p < 0.0001 for cIAP2). Combinatorial ROC analysis revealed the marginally fair discriminatory value of 5 genes as a panel (AUC = 0.685, p < 0.0001). Kaplan-Meier survival curves revealed significant association of cIAP2 down-regulation in CRC with lower overall survival probability of CRC patients (p = 0.0098). DR5, BIRC5/Survivin, XIAP, c-IAP1 and c-IAP2 mRNA expression are significantly deregulated in CRC and could provide a panel of markers with significant discriminatory value between CRC and normal colorectal tissue.

One of the most frequently mutated genes in CRC is the KRAS oncogene, component of the RAS/RAF/MEK/ERK signaling pathway downstream of EGFR (Epidermal Growth Factor Receptor), which is a significant regulator of cell growth, proliferation, differentiation and apoptosis. KRAS is found mutated in approximately 40% of CRC cases, mainly in the codons 12 (wild-type GGT) and 13 (wild-type GGC) of exon 1, with demonstrated capacity as major predictive markers of resistance to treatment with anti-EGFR monoclonal antibodies in metastatic CRC. Therefore, the effectiveness of treatment is dependent on the mutation status of KRAS in CRC [7][8][9] .
Apoptotic cell death is important for the homeostasis of normal colorectal epithelial cells, which are generated from stem cells at the base of the crypt region via their differentiation and migration to the epithelial surface, where they eventually die by apoptosis and are discarded in the lumen of the colon 10,11 . Imbalance in cell death signaling is implicated in CRC, since it is well known that the failure of apoptotic cell death pathways constitutes a crucial process for the survival of cancer cells and a leading cause of resistance to current therapeutic approaches 12,13 .
TRAIL (TNF-related apoptosis-inducing ligand) is a cytokine best known for its ability to selectively induce apoptosis in cancer cells while sparing most normal cells, which made it a rather attractive potential chemotherapeutic agent. TRAIL has been shown to induce apoptosis in malignant cells both in vitro and in pre-clinical models of cancer [14][15][16][17] . However, TRAIL therapy has a major limitation due to the fact that a large number of cancers become resistant to TRAIL and escape from immunosurveillance mechanisms 15,18,19 . The proapoptotic TRAIL receptors DR4 (TRAIL-R1) and DR5 (TRAIL-R2) not only trigger apoptosis in TRAIL-sensitive cells, but also activate survival pathways in tumor cells that resist the induction of cell death upon exposure to TRAIL 19 . Moreover, it has been reported that there is a marked increase in sensitivity of TRAIL-induced apoptosis during transition from colorectal adenoma to CRC 20,21 . It was recently shown that constitutive signaling from DR5 (TRAIL-R2) promotes migration and invasion in a cancer cell-autonomous manner 22 . As a biomarker, DR5 was found significantly up-regulated in stage II and III CRC by immunohistochemistry but without prognostic significance 23 . Bavi et al. reported that DR4 was also significantly up-regulated in CRC and adenomas by immunohistochemistry and was associated with a less aggressive phenotype characterized by early stage disease, whereas DR5 was associated with a microsatellite stable (MS-S/L) phenotype and with absence of KRAS mutations in a Middle-Eastern population 24 .
Inhibitor of apoptosis proteins (IAPs) are eight anti-apoptotic human proteins characterized by the presence of one to three baculoviral IAP repeats (BIR) domains. The most extensively studied IAP protein family members are cIAP1, cIAP2, XIAP, ML-IAP and BIRC5/Survivin 11,25 . Functionally, IAPs have been shown to interact with multiple regulators of both the extrinsic (death receptor) and the intrinsic (mitochondrial) apoptotic pathways and inhibit the activation of caspases (caspases 3,7,8,9) directly (XIAP) or indirectly (cIAP1, cIAP2) via binding to SMAC/DIABLO (Second mitochondria-derived activator of caspase, SMAC) endogenous inhibitor of XIAP. In addition, cIAP1 and cIAP2 are crucial activators of the canonical NF-κ B (Nuclear factor-kappa B, NF-κ B) signaling pathway by TNF (Tumor necrosis factor, TNF) and negative regulators of the non-canonical NF-κ B signaling 26,27 . Therefore, IAPs act as modulators of cell viability by inducing pro-survival signaling pathways.
In cancer, IAPs are found to be overexpressed in several malignancies, including CRC, and their action has been reported to enhance cancer cell resistance to apoptotic death stimuli. Increased IAP expression has been related to tumorigenesis, progression of cancer, resistance to treatment and poor prognosis 12,28,29 . Hence, IAP proteins could represent promising biomarkers in CRC and/or targets for therapeutic intervention.
The aim of the present study is to investigate the mRNA expression profile of the death receptor genes DR4 (TRAIL-R1) and DR5 (TRAIL-R2), as well as IAP genes cIAP1, cIAP2, XIAP and BIRC5/Survivin in CRC and normal tissue samples and to evaluate their potential diagnostic and prognostic significance, as well as their putative association with KRAS mutation status in CRC. DR5, BIRC5/Survivin, XIAP, cIAP1 and cIAP2 gene expression is found significantly deregulated in CRC at the mRNA level and could provide a 5-gene panel of markers with significant discriminatory value between CRC and normal colorectal tissue.

Results
In the current study we examined 100 cancerous and 100 normal paired colorectal tissues for the mRNA expression levels of DR5, DR4 and BIRC5/Survivin, XIAP, cIAP1, cIAP2 (IAPs) genes, using quantitative real-time PCR analysis. Data were normalized to GAPDH and HPRT mRNA expression levels. All samples yielded analyzable data. Out of the 100 CRC tissues included in the study, 50 were KRAS wild-type and 50 were KRAS-mutant tissues, harboring the activating KRAS-G13D, KRAS-G12D and KRAS-G12V mutations. The descriptive statistics regarding the expression status of all genes under study in patients with primary CRC are presented in Table 1. Follow-up information (ranging 1-99 months) was available for 65 patients and included survival status (alive or deceased) along with the dates of the events and cause of death. During respective follow-up period, 23 patients (35.4%) had passed away.
DR5 is significantly up-regulated in CRC. Death receptor gene DR5 (TRAIL-R2) was found to be over-expressed in CRC tissues compared to normal pairs. DR5 showed very significant higher mRNA expression levels in CRC tissues (mean ± SEM: 1.915 ± 0.16 fold expression units) compared to normal counterparts (mean ± SEM: 1.174 ± 0.11 fold expression units) (p < 0.0001), as calculated by the Wilcoxon Signed Ranks test (Tables 1 and 2). As presented in Table 1, DR5 mRNA expression was found to be 1.58-fold higher in CRC specimens than in normal ones, according to median -fold expression value. We also observed that the increase of DR5 mRNA expression was significantly higher in low TNM stage (I/II) colorectal tumors (p = 0.0293) compared to high TNM stage (III/IV) ones ( Fig. 1A) with the use of Mann Whitney U test. The increase of DR5 mRNA expression in CRC was not significantly different between groups of patients classified according to lymph node invasion status (p = 0.089) (Fig. 1B). In order to evaluate the discriminatory value of DR5 mRNA expression profile between cancer and normal tissue, we performed ROC analysis. ROC curves were constructed for DR5 expression levels by plotting sensitivity versus (1-specificity). The ROC curve ( Fig. 2A) illustrates fair discriminatory power of DR5 mRNA expression levels in distinguishing cancer from normal tissues, with strong statistical significance (AUC = 0.700; 95%CI = 0.631-0.763; p < 0.0001). Survival curves determined by the Kaplan-Meier method for overall survival (OS) demonstrated that DR5 expression is not significantly associated with OS of CRC patients (p = 0.140) (Fig. 2B).
DR4 (TRAIL-R1) mRNA expression was also elevated in CRC (mean ± SEM: 1.926 ± 0.13 fold expression units) compared to normal tissues (mean ± SEM: 1.661 ± 0.08 fold expression units) ( Table 1), but not in a statistically significant manner (p = 0.548), as calculated by the Wilcoxon Signed Ranks test (Table 2). DR4 mRNA expression was found 1.15-fold higher in CRC specimens than in normal ones, according to median -fold expression value ( Table 1). The increase of DR4 mRNA expression in CRC was not significantly different between groups of patients classified according to TNM stage or lymph node invasion status (Suppl. Figure 1A,B, respectively). As expected, ROC curve analysis for DR4 mRNA expression levels (  (Tables 1 and 2). Differential mRNA expression of BIRC5/Survivin was found 1.35-fold higher in CRC specimens than in normal ones, according to median -fold expression value ( Table 1). The increase of BIRC5/Survivin mRNA expression in CRC was not significantly different between groups of patients classified according to TNM stage or lymph node invasion status (Suppl. Figure 1C,D, respectively). ROC curve analysis for BIRC5/Survivin expression levels ( Fig. 2E) demonstrates marginal discriminatory power in distinguishing cancer from normal tissues with statistical significance (AUC = 0.598; 95%CI = 0.526-0.666; p = 0.015). Kaplan-Meier survival curves for OS demonstrated that BIRC5/ Survivin expression is not significantly associated with OS of CRC patients (p = 0.110) (Fig. 2F).
XIAP gene, also of the IAP gene family, is also significantly up-regulated in the 100 CRC tissues examined (mean ± SEM: 1.299 ± 0.10 fold expression units) compared to normal counterparts (mean ± SEM: 1.090 ± 0.09 fold expression units) at the mRNA level (p = 0.012), as calculated by the Wilcoxon Signed Ranks test (Tables 1  and 2). Differential mRNA expression of XIAP was found 1.18-fold higher in CRC specimens than in normal ones, according to median -fold expression value ( Table 1). The increase of XIAP mRNA expression in CRC was not significantly different between groups of patients classified according to TNM stage or lymph node invasion status (Suppl. Figure 1E,F, respectively). ROC curve analysis for XIAP mRNA expression levels (  (mean ± SEM: 1.509 ± 0.09 fold expression units) at the mRNA level (p < 0.0001), as calculated by the Wilcoxon Signed Ranks test (Tables 1 and 2). Differential mRNA expression of cIAP1 was 0.19-fold lower in CRC specimens than in normal ones, according to median -fold expression value ( Table 1). The decrease of cIAP1 mRNA expression in CRC was significantly observed in high TNM stage (III/IV) colorectal tumors compared to low TNM stage (I/II) ones (p = 0.033) (Fig. 1C), whereas it was not significantly different between groups of patients classified according to lymph node invasion status (p = 0.736) (Fig. 1D). ROC curve analysis for cIAP1 mRNA expression levels ( Fig. 2I) revealed poor discriminatory power in distinguishing cancer from normal tissues with strong statistical significance (AUC = 0.628; 95%CI = 0.558-0.695; p = 0.0011). Kaplan-Meier survival curves for OS demonstrated that cIAP1 down-regulation in CRC is not significantly associated with OS patients (p = 0.140) (Fig. 2J). Similarly, cIAP2 gene was found significantly down-regulated in CRC (mean ± SEM: 5.652 ± 0.53 fold expression units) compared to normal tissue (mean ± SEM: 9.208 ± 0.93 fold expression units) at the mRNA level (p < 0.0001), as calculated by the Wilcoxon Signed Ranks test (Tables 1 and 2). Differential mRNA expression of cIAP2 was 0.43-fold lower in CRC specimens than in normal ones, according to median -fold expression value ( Table 1). The decrease of cIAP2 mRNA expression in CRC was not significantly different between groups of patients classified according to TNM stage or lymph node invasion status (Suppl. Figure 1G,H, respectively). ROC curve analysis for cIAP2 mRNA expression levels (Fig. 2K) revealed poor discriminatory power between cancer and normal tissues, although with very strong statistical significance (AUC = 0.673; 95%CI = 0.603-0.737; p < 0.0001). Kaplan-Meier survival curves for OS demonstrated that cIAP2 down-regulation in CRC is significantly associated with lower survival probability and shorter OS of CRC patients (p = 0.0098) (Fig. 2L).

Significant discriminatory value of a five-gene panel between CRC and normal colorectal tissue.
Quantitative PCR demonstrated differential mRNA expression of 5 genes DR5, XIAP, BIRC5/Survivin, cIAP1 and cIAP2 (except DR4) between CRC and normal tissue. Data were combined using logistic regression and 5-gene panel ROC analysis was performed on all samples in order to evaluate overall discriminatory value of the 5 genes expression levels between CRC and normal tissues. As shown in Fig. 3, there is very significant marginally fair discriminatory value of the 5-gene panel between CRC and normal tissues (AUC = 0.685, 95%CI = 0.616-0.749, p < 0.0001). Thus, a 5-gene assay of DR5 and IAPs could have differential diagnostic perspectives.

Differential mRNA expression profile according to KRAS mutation status of CRC tissues. Since
KRAS mutation status in CRCs is correlated with resistance to treatment with anti-EGFR monoclonal antibodies, the effort for novel therapeutic approaches, such as TRAIL-related therapeutics for KRAS-mutated CRCs is crucial. In the present study we investigated the association of mRNA expression levels of TRAIL receptors (DR4    Fig. 4C. Therefore, the over-expression of BIRC5/Survivin observed in CRC is particularly attributed to KRAS wild-type tumors. cIAP2 mRNA expression was significantly decreased in KRAS-mutant CRC tissues (p = 0.0001), but not in KRAS wild-type ones (p = 0.055) compared to normal pairs (Fig. 4D).
On the contrary, the use of Mann Whitney U test did not reveal statistically significant overexpression of XIAP or DR4 in either group of cancerous tissues classified according to KRAS mutation status compared to normal pairs (Fig. 4E,F, respectively).

Protein expression of apoptotic factors confirms the pattern of their mRNA expression in CRC:
Protein expression of DR4, DR5, BIRC5/Survivin, is significantly upregulated in CRC. cIAP1 and cIAP2 protein levels are low in human colon cancer tissues. In the present study, protein expression of DR4, DR5, BIRC5/Survivin, cIAP1 and cIAP2 was determined immunohistochemically in twenty five colorectal cancer tissues that showed high levels of mRNA differential expression. All samples were found to express DR4, DR5 and BIRC5/Survivin showing either mild (+ ), moderate (+ + ) or strong (+ + + ) reactivity, whereas  for cIAP1 and cIAP2, staining was either negative (− ) or mild (+ ) ( Table 3, Fig. 5). 23/25 (92%) colorectal cancer tissues were found to overexpress DR4, 19/25 (76%) were found to overexpress DR5 and 18/25 (72%) were found to overexpress BIRC5/Survivin (Table 3). For statistical analysis, the specimens that showed negative (− ) or mild (+ ) staining were considered as 0 and those that showed moderate (+ + ) or strong (+ + + ) as 1. Normal mucosa for each respective tumor tissue was also considered as 0, since it showed negative reactivity. Paired samples t-tests were performed to evaluate the statistical significance of these results and in each case overexpression of the protein was found to be statistically significant (p < 0.0001). cIAP1 and cIAP2 protein expression was low (Table 3). In addition to protein expression analysis, sections from the tumor specimens were stained with haematoxylin-eosin to confirm the percentage of tumor. In all cases, the confirmed percentage of tumor ranged between 80 and 100 percent of the specimen, thus excluding cross-contamination from normal mucosa (Fig. 5F).

Discussion
The use of molecular markers in cancer medicine is still evolving and it is clear that there is much yet to be clarified. In the last decades there is increasing interest in the use of molecular staging in CRC prognosis and treatment. The choice of the molecular markers remains the most prominent parameter. Given the fact that there is no such thing as the perfect marker, the best choice would include a combination of sensitive and specific markers, in order to achieve maximum clinical relevance.

DR5 gene is significantly up-regulated in CRC, particularly in TNM stage I/II tumors, including KRAS-mutant tumors. The TNF-Related Apoptosis Inducing Ligand, TRAIL, and its receptors in cancer
have the role of a potent inducer of apoptosis in tumor cells while most normal cells are resistant. In the present study, expression of DR4 and DR5 was assessed in tumors from colorectal cancer patients in correlation to each patient's normal mucosa in the absence of adjuvant chemotherapy, in association with colon cancer staging. DR5 mRNA expression was found significantly higher in colorectal tumor specimens, as compared to physiological mucosa, especially in low TNM stage (I/II) CRC.
The identification of DR5 receptor up-regulation from early colon cancer stages may ensure better diagnosis for patients diagnosed early. Furthermore, DR5 expression suggests that, in addition to a potentially important role as a diagnostic but not prognostic marker, it might also have an important implication in therapy by administration of rhTRAIL and agonistic antibodies against DR5, already available.
Since KRAS mutation status in CRCs is correlated with resistance to treatment with anti-EGFR monoclonal antibodies, the effort for novel therapeutic approaches, such as TRAIL-related therapeutics, for KRAS mutated CRCs is crucial. Here, DR5 mRNA expression was found significantly elevated in both KRAS mutant and KRAS wild-type CRC tissues as compared to their normal pairs, with greater significance in KRAS mutant specimens. Notably, oncogenic KRAS can regulate Death Receptors during metastasis in colorectal cancer cells 30 . The findings of this study, clearly propose that appropriate combinatorial therapeutic schemes involving recombinant TRAIL and/or agonistic antibodies against DR5 may be efficient against colorectal metastatic tumors, including those bearing KRAS mutations.   has been proposed as a biomarker with diagnostic and prognostic potential as well as a therapeutic target 6,27,31 . BIRC5/Survivin has been shown to be up-regulated by oncogenic c-H-Ras 32 . Moreover, the ratio between survivin and TRAIL receptors can be predictive of recurrent disease in neuroblastoma 33 . In this study, significant over-expression of BIRC5/Survivin in CRC is reported, which is particularly observed in KRAS wild-type tumors. The individual mechanism of differential regulation of BIRC5/Survivin by either mutant or wild-type KRAS needs to be further examined. Even when the mRNA levels not always translate to protein expression, there is good chance that tumors with up-regulated BIRC5/Survivin will be responsive to therapies based on specific inhibitors targeting their promoter and mRNA expression. Notably, a randomized open-label phase II study evaluating antitumor activity of the survivin antisense oligonucleotide LY2181308 (LY) in combination with docetaxel (DO) for second-line treatment of patients with non-small cell lung cancer (NSCLC) was established 34 . Moreover, a phase I/II study of sepantronium bromide (YM155, survivin suppressor) with paclitaxel and carboplatin in patients with advanced non-small-cell lung cancer has been reported 35 . cIAP1 and cIAP2 are significantly down-regulated in CRC and only cIAP2 is dependent on KRAS mutation status; cIAP2 down-regulation is significantly associated with lower survival. IAP family members have been shown to play an important role in apoptosis. Different IAP family members may be differentially involved in tumor progression, diagnosis and therapy, indicating the purpose of studies which examine the importance of individual IAP members. Moreover, individual IAP family members have been shown to be differentially regulated by RAS oncogene either directly or by a RAS-induced autocrine manner 36 . Here, cIAP1 and cIAP2 were found significantly down-regulated in CRC, and low expression of cIAP2 was significantly correlated with detection of KRAS mutations in the colorectal tumors under study. Moreover, cIAP2 down-regulation in CRC was found significantly associated with lower survival probability and shorter OS of CRC patients, thus demonstrating a potential prognostic usefulness. Relative expression of IAPs in a particular tumour can have a major impact on their diagnostic and/or prognostic capacity, as well as to therapies targeting apoptosis in cancer.

Protein expression of apoptotic factors confirms the pattern of their mRNA expression in CRC.
In an attempt to correlate mRNA expression levels with protein expression, immunohistochemistry was performed in a number of representative samples that were selected based on the levels of mRNA expression for each gene. Protein expression was confirmed in the investigated samples.   DR5 was found to be significantly up-regulated in 76/100 CRC tissues and protein expression was also high in 19/25 (76%) samples analysed. DR4 mRNA expression was elevated in CRC (59/100 patients), although not in a statistically significant manner, and protein expression was high in 23/25 (92%) investigated tumors, with a statistical significance. BIRC5/Survivin was significantly up-regulated in the 100 CRC tissues examined (61/100) and protein expression was also significantly high in 18/25 (72%) CRC tissues. cIAP1 and cIAP2 were found to be significantly down-regulated in CRC (67/100 and 69/100 respectively) and protein expression was low in all samples analysed. BIRC5/Survivin, XIAP, cIAP1 and cIAP2) with significant differential diagnostic potential. Differential mRNA expression of 5 genes DR5, XIAP, BIRC5/Survivin, cIAP1 and cIAP2 between CRC and normal colorectal tissue is reported here. Data were combined using logistic regression and 5-gene panel ROC analysis was performed on all samples in order to evaluate overall discriminatory value of the 5 genes expression levels between CRC and normal tissues. There is very significant fair discriminatory value of the 5-gene panel between CRC and normal tissues which could have significant differential diagnostic value.

5-genes biomarker panel (DR5,
This study confirms the central role of apoptosis pathways in CRC, despite the fact that the molecular mechanism remains unclear. The mRNA expression profiling of 5 apoptosis genes could contribute to CRC diagnosis, in cooperation with known molecular, clinical, and histologic diagnostic factors. In case that this promising result can be valid in a much larger patient cohort, the apoptotic signature could not only enhance the diagnostic accuracy to identify early stage CRC, but could also provide good candidate cancer diagnostic/predictive biomarkers and/or putative drug targets for tumor therapies. Therefore, further evidence is necessary towards this direction.

Methods
Study group. In this study, 200 fresh frozen colorectal tissues obtained from 100 patients with sporadic CRC collected at the G. Gennimatas General Hospital of Athens were examined. CRC specimens, 0.5-2 cm in diameter, were acquired from patients who underwent surgery for primary infiltrating CRC. Tumor specimens were taken from the most central part of the mass, thus ruling out the possibility of cross-contamination from normal mucosa or adenomatous tissue. Sections from the tumor specimens were stained with haematoxylin-eosin to confirm the percentage of tumor. In all cases, the confirmed percentage of tumour ranged between 80 and 100 percent of the specimen, thus excluding cross-contamination from normal mucosa (Fig. 5F). At the same time, adjacent normal colonic mucosa, sufficiently separated from the tumor, was also collected. Normal colonic tissue was taken from the histologically cancer-free margins of the resected colon to ensure absence of cancerous cells. The remaining biopsy section was prepared for routine histology. All specimens were processed immediately after surgical resection. Patients with low-lying locally advanced rectal cancers who received neoadjuvant chemo-/radiotherapy were excluded from the study to avoid possible bias deriving from unknown radiation-mediated effects on tumor gene status. The mean age of patients was 69.5 ± 1.2 years (ranging from 42 to 95). Informed consent was obtained from all patients. The study protocol was approved by the Medical Ethical Review Committee of the G. Gennimatas General Hospital of Athens (Institutional Review Board number: 4/06). Clinical and histological information for the tumor samples included in the present study (Table 4)  Cancerous and normal colon tissues were homogenized in 2 ml TRIzol reagent (Invitrogen by Life Technologies, Carlsbad, CA, USA) on ice using an electric tissue grinder (ULTRA-TURRAX, type T-25, Junke and Kunkel). Total RNA was extracted from homogenized tissue according to the manufacturer's instructions. Five micrograms of total RNA were reverse transcribed into first-strand cDNA using oligo(dT) primer, with the SuperScript TM II Reverse-Trascriptase pre-amplification system (Invitrogen by Life Technologies, Carlsbad, CA, USA), following the manufacturer's instructions.
Quantitative real time PCR. Singleplex fluorescent-based quantitative real-time PCR assay (qPCR) was performed for the determination of DR4, DR5, cIAP1, cIAP2, XIAP and BIRC5 mRNA expression levels in the colon specimens on an iQ5 Real Time PCR Detection System (Bio-Rad Laboratories, Inc., USA). The quantitative PCR analysis was performed in 96-well plates using iQ SYBR Green Supermix (Bio-Rad Laboratories, Inc., USA). The 10 ml reaction mixtures contained 75 ng cDNA, 50 nM primers, and 5 μ l iQ SYBR Green Supermix. The thermal protocol consists of a 10 min polymerase activation step at 95 °C, followed by 40 cycles of denaturation at 95 °C for 20 sec and the primer annealing and extension at 60 °C for 1 min. Each sample was amplified twice in duplicates and the average C T values were calculated for the expression analysis. Following the amplification, a dissociation curve was generated in order to verify the PCR products of interest from any non-specific ones or any primer-dimer production, by their specific melting temperature (Tm). Relative quantification analysis for the DR4, DR5, cIAP1, cIAP2, XIAP and BIRC5/Survivin mRNA expression was performed with the comparative 2 −ΔΔCT method. mRNA expression of all genes was normalized with the use of the GAPDH and HPRT reference genes, while the Caco-2 cell line was used as a calibrator. The primers used for each gene were the following: GAPDH (NM_002046) F: Immunohistochemistry. Protein expression of DR4, DR5, BIRC5/Survivin, cIAP1 and cIAP2 in human colon tumors was determined immunohistochemically in twenty five representative specimens, in order to confirm the respective findings from the mRNA expression analysis. Samples for this part of the analysis were selected based on the levels of mRNA expression for each gene under study. For histology, 4-μ m sections were cut from formalin-fixed, paraffin-embedded tissues and stained with haematoxylin-eosin. For the detection of DR4, a rabbit polyclonal antibody (ProSci Inc., Poway, CA, USA) was applied on serial paraffin sections at 5 μ g/ml for 1hr at room temperature, whereas for the detection of DR5 a rabbit polyclonal antibody (ProSci Inc., Poway, CA, USA) was applied at 1 μ g/ml for 1 hr at room temperature. BIRC5/Survivin was detected using a rabbit monoclonal antibody (Cell Signaling Technology, Danvers, MA, USA) on serial paraffin sections at 1 μ g/ml for 1hr at room temperature. For the detection of cIAP1 and cIAP2, a rabbit polyclonal antibody (Santa Cruz Biotechnology, Inc., Dallas, TX, USA) was applied separately on serial paraffin sections at 4 μ g/ml for 1 hr at room temperature, after antigen unmasking using PT Link (pH 6) for Pre-Treatment (DAKO, Glostrup, Denmark) for 15 mins. The Envision polymer detection kit (DAKO, Glostrup, Denmark) was utilized and the 3′ ,3′ -diaminobenzidine was used as chromogen to reveal the antigen binding sites in all cases. Staining intensity was divided in three classes (+ , + + , + + + ). Sections from the same specimen, where the primary antiserum was substituted by buffer solution, served as negative controls. Positive controls were used following manufacturer's instructions. Normal mucosa was also stained for each corresponding tumor specimen. WHO histological criteria were used for the histological classification of adenocarcinomas.
Statistical analysis. Differences in the mRNA expression profile of the six genes between CRC and normal tissue specimens were evaluated with the non-parametric Wilcoxon Signed Ranks test. CRC tissues were classified according to their clinico-pathological features (TNM stage, lymph node invasion status) and statistical analysis was performed with the use of Mann-Whitney U test. Receiver-operating characteristic (ROC) curves were constructed for quantitative PCR data sets using a single-gene test in CRC versus normal tissues, by plotting sensitivity versus (1-specificity). Quantitative data demonstrating differential mRNA expression between CRC and normal tissue were combined using logistic regression and the performance of a multi-gene panel ROC analysis was performed on all samples. The areas under the ROC curves (AUC) were analysed by Hanley and McNeil method. Moreover, differential mRNA expression profile of the six genes between CRC and normal tissues was examined in subgroups of patients according to KRAS mutation status (mutant or wild-type CRC tissue) using