Abstract
Purpose: Colorectal cancer is one of the most common cancers. The tumor necrosis factor–related apoptosis inducing ligand (TRAIL) pathway transmits apoptotic signals and anticancer agents that activate this system, which are in clinical development. We sought to determine the prognostic value of the clinically most relevant members of this pathway in colorectal cancer patients.
Experimental Design: We used an arrayed panel of colorectal cancer tissue to assess the protein expression of the functional TRAIL receptors (TRAIL-R1 and TRAIL-R2) and both the long and short forms of FLICE inhibitory protein (FLIPL and FLIPS). Disease-free survival was examined by Kaplan-Meier estimates and the log-rank test. Prognostic factors were determined by Cox multivariate analysis.
Results: The TRAIL receptors and FLIPS were not associated with survival. On univariate analysis, strong FLIPL expression was associated with a significantly higher survival (P = 0.0082). On multivariate analysis using the Cox proportional hazards model, FLIPL phenotype was significantly associated with a poor prognosis in this series (hazard ratio, 2.04; 95% confidence interval, 1.18-3.56; P = 0.011).
Conclusions: Overexpression of FLIPL, but not TRAIL-R1 or TRAIL-R2, provides stage-independent prognostic information in colorectal cancer patients. This may indicate a clinically more aggressive phenotype and a subset of patients for whom more extensive adjuvant treatment would be appropriate.
- Colorectal Cancer
- Tissue Microarray
- Prognostic Factor
- TRAIL-R1
- TRAIL-R2
- FLIPL
Colorectal cancer kills ∼500,000 people worldwide each year (1) and new, more effective treatments are warranted. Tumor necrosis factor–related apoptosis inducing ligand (TRAIL) was identified as a unique death ligand with respect to its ubiquitous receptor expression and the lack of cytotoxic effects in normal tissue (2, 3). Five TRAIL receptors have been described. TRAIL-R1 and TRAIL-R2 carry cytoplasmic death domains and mediate apoptosis (4–7). The three other receptors (TRAIL-R3, TRAIL-R4, and osteoprotegrin) serve as decoy receptors (6, 8, 9). A key inhibitor of death receptor signaling is FLICE inhibitory protein (FLIP), which can efficiently block caspase-8 cleavage due to a structural similarity to caspase-8 (10). High expression of FLIP has been correlated with TRAIL resistance in various tumor types including colon cancer (11). Furthermore, increased expression of FLIP was found in stomach cancer patients (12). Alternative splicing generates two isoforms of FLIP: a long form (FLIPL) that contains a caspase-like domain but is devoid of caspase catalytic activity and a short form (FLIPS) lacking the caspase-like domain. Both FLIP variants are capable of protecting cells from TRAIL-R-induced apoptosis (10). However, overexpression of FLIPL but not FLIPS has been found to inhibit chemotherapy-induced colorectal cancer cell death (13). Several clinical trials, including phase II, using agonistic monoclonal antibodies to TRAIL-R1 and TRAIL-R2 to treat cancer, have been launched.3
Since its first description in 1998, tissue microarray analysis (14) has been used for the immunohistochemical analysis of target protein expression in a wide range of primary tumor types. Initial fears that the reduced amount of individual tumor tissue analyzed using this technique might not be representative of the tumor as a whole seem to be largely unfounded (15). The strengths of this approach lie in its ability to provide a rapid turnover of results from very large patient cohorts while reducing variability in experimental conditions and reducing costs (16). This tissue microarray of colorectal cancer patients has previously been validated with a p53−/Bcl-2+ phenotype, loss of histocompatibility leukocyte antigen, or overexpression of MHC class I –related chain A, all being independent markers of poor survival (17–19). To our knowledge, this is the largest study assessing the role for TRAIL-R1 and TRAIL-R2 as prognostic markers in colorectal cancer patients. We have also included FLIPL and FLIPS in our analyses because these markers play a key role in the TRAIL pathway. There seems to be no previous study addressing the potential role for FLIPL in colorectal cancer patients.
Materials and Methods
Patients and specimens. The study population was composed of 462 patients undergoing elective surgery for a single, nonmetachronous histologically proven primary colorectal cancer at University Hospital, Nottingham between 1st January 1994 and 31st December 2000. Data about tumor site, stage, histologic type, and tumor grade have been recorded prospectively for these patients. Only patients with lymph node–positive disease were routinely treated with adjuvant chemotherapy consisting of 5-flurouracil and folinic acid. The original histopathologic slide sets and pathologic reports for all cases were obtained from the hospital archives and were reviewed to confirm the diagnosis and the accuracy of existing data. However, in a small minority of cases, only very limited slide sets were available and the full clinicopathologic data set was incomplete. In addition, we did not attempt to ascertain the vascular invasion status of tumors where this information had not previously been recorded.
Follow-up data about the date and cause of death for this cohort of patients have been provided prospectively by the UK Office for National Statistics. Follow-up was calculated from the date of resection of the primary tumor, and all surviving cases were censored for survival analysis on 31st December 2003. Disease-specific survival was used as the primary end point. The Local Research Ethics Committee granted approval for the study.
Preparation of the tissue microarray. Tissue microarrays were constructed as previously described (14).
Five-micrometer H&E-stained slides were used to identify and mark out representative areas of viable tumor tissue. Needle core biopsies (0.6 mm) from the relevant areas of corresponding paraffin-embedded blocks were then placed at defined coordinates in the recipient paraffin array blocks using a manual arrayer (Beecher Instruments). Array blocks were constructed at a density of 80 to 150 cores per array. Analysis of each marker was done on a single core from each tumor. This typically shows >90% concordance with conventional whole-section analysis of tumor markers and has previously been validated (15).
Immunohistochemical methods. Immunohistochemistry was done with the use of a standard avidin-biotin peroxidase method with TRAIL-R1 (H130) anti-human mouse monoclonal antibody clone (1:20 dilution; Santa Cruz Biotechnology, Inc.), TRAIL-R2 anti-human goat polyclonal antibody clone (1:100 dilution; Calbiochem), and FLIPL (C-19) anti-human goat monoclonal antibody (1:125 dilution; Santa Cruz Biotechnology) and FLIPS (1:250 dilution; Sigma). Briefly, 5-μm array sections were deparaffinized with xylene (twice for 10 min), rehydrated through graded alcohol (thrice for 10 s), and immersed in methanol containing 0.3% hydrogen peroxide for 15 min to block endogenous peroxidase activity. Heat-induced epitope retrieval consisting of 15-min microwave treatment (10 min at high power and 5 min at low power) in pH 6.0 citrate buffer was necessary with the FLIPL monoclonal antibody only. Endogenous avidin/biotin binding was blocked using an avidin/biotin blocking kit (Vector Labs) and sections were treated with 100 μL of normal swine serum for 10 min to block nonspecific binding of the primary antibody.
Sections were incubated with 100 μL of primary antibody for 1 h at room temperature except for FLIPL, which was incubated overnight at 4°C. Primary antibody was omitted from negative control sections, which were incubated in normal swine serum. After washing with TBS, sections were incubated with 100 μL of biotinylated antimouse immunoglobulin (DAKO Ltd.) diluted 1:100 in normal swine serum for TRAIL-R1 and biotinylated antigoat immunoglobulin (Vector Labs) diluted 1:100 or 1:200 in normal swine serum for 30 min for TRAIL-R2 and FLIPL, respectively, followed by 100 μL of preformed streptavidin-biotin/horseradish peroxidase complex (DAKO) for 60 min at room temperature. Staining was finally developed using 3,3′-diaminobenzidine tetrahydrochloride (DAKO) and enhanced with crystal violet staining. The slides were finally counterstained with hematoxylin and washed before passing through serial alcohol and xylene baths. Slides were finally dried and mounted with DPX (a mixture of disterene, plasticizer, and xylene; Sigma).
Evaluation of staining. Evaluation of the staining was carried out by two observers, who were blinded to the clinicopathologic data, with a consensus decision in all cases. For each antigen, tumors were classified into four groups with regard to staining intensity: negative, weakly positive, moderately positive, and strongly positive.
Statistical methods. All statistical analyses were done using the SPSS package (version 14 for Windows, SPSS, Inc.). Associations between categorical variables were examined using cross-tabulation and the Pearson χ2 test for categorical variables. Kaplan-Meier curves were derived to assess disease-specific survival, and the significance of differences in disease-specific survival between groups was calculated using the log-rank test. Complete outcome data were available for all but one patient in the study. Patients whose deaths related to their colorectal cancer were considered in the disease-specific survival calculations. However, patients who died from postoperative complications (deaths within 1 month of the date of surgery) were excluded from the survival analyses. Deaths resulting from non–colorectal cancer–related causes were censored at the time of death. Multivariate analysis using the Cox proportional-hazards model was used to determine hazard ratios and identify variables with independent prognostic significance in this cohort. In all cases, P < 0.05 was considered statistically significant.
Results
Clinicopathologic data. The arrayed tumors were found to have similar clinicopathology to colorectal cancer encountered in the United Kingdom (Table 1 ), with 53% tumor-node-metastasis (TNM) stage I or II and 45% stage III or IV. The median age at the time of surgery was 72 years, consistent with the median age at diagnosis of colorectal cancer of 70 to 74 years in the United Kingdom (20). At the time of censoring for data analysis, 49% of patients had died from their disease, 37% were still alive, and 14% were deceased from non–colorectal cancer–related causes. Thirty-eight deaths occurred within 1 month of surgery. The median length of follow-up available for surviving patients was 75 months (range, 36-116 months). The overall median 5-year disease-specific survival for the cohort was 58 months, which is comparable with the ∼45% 5-year survival seen for colorectal cancer in the United Kingdom.
Clinicopathologic variables for the patient cohort (n = 462)
Antigen expression. Although it has previously been stated that the number of cores uninterpretable due to tissue loss or damage in tissue microarray studies may exceed 20% to 30% of the total, this was not observed in the current study. The highest number of uninterpretable cases (68 cases; 14.7% of total) was seen for TRAIL-R1. Representative examples of negative, weakly positive, moderately positive, and strongly positive staining for each antigen are shown in Fig. 1 . The normal mucosa adjacent to some tumors showed weak or moderate levels of FLIPL expression, suggesting that the pathology in the tumor is increased or decreased expression. However, because most tissue microarray samples did not contain normal mucosa, it was not possible to assess tumor expression relative to their adjacent normal mucosa. The relative expression of the markers is therefore assessed among the tumors.
Cores from colorectal cancer tumors showing negative, weak, moderate, and strong staining for TRAIL-R1 (A), TRAIL-R2 (B) FLIPL (C), and FLIPS (D). Original magnification, ×100.
TRAIL-R1 expression. Expression of TRAIL-R1 protein was detected at the cell surface or in the cytoplasm of 352 of 394 (89.3%) evaluable tumors, with only 42 cores being negative (Table 2 ). In contrast, <2% the cores showed nuclear expression of TRAIL-R1. In univariate analysis, no strong associations between TRAIL-R1 expression and clinicopathologic variables were noted. On Kaplan-Meier analysis, no association was found between TRAIL-R1 expression and survival (P = 0.5491; Fig. 2 ).
Number of tumors in each staining intensity group for the three markers investigated in the current study
Kaplan-Meier plot for disease-specific survival in colorectal cancer patients. TRAIL-R1 strongly staining versus moderately staining versus weakly staining versus negative staining tumors (n = 394).
TRAIL-R2 expression. Expression of TRAIL-R2 protein was detected at the cell surface or in the cytoplasm of 380 of 401 (94.8%) evaluable tumors, with only 21 cores being negative (Table 2). In contrast, no nuclear expression of TRAIL-R2 was observed. No strong associations between TRAIL-R2 expression and clinicopathologic variables, except from TNM stage (P = 0.023), were noted. On Kaplan-Meier analysis, no association was found between TRAIL-R2 expression and survival (P = 0.5547; Fig. 3 ).
Kaplan-Meier plot for disease-specific survival in colorectal cancer patients. TRAIL-R2 strongly staining versus moderately staining versus weakly staining versus negative tumors (n = 401).
FLIPL expression. Expression of FLIPL protein was detected in the cytoplasm of 375 of 396 evaluable tumors (94.7%), with only 21 cores being negative (Table 2). In contrast, <1% the cores showed nuclear expression of FLIPL. No strong associations between FLIPL expression and clinicopathologic variables were noted. On Kaplan-Meier analysis of all subgroups, the group of patients whose tumors overexpressed FLIPL seemed to do worse from 12 months postsurgery (Fig. 4A ). When this group was analyzed compared with all patients, a significant association was found between overexpression of FLIPL and survival (P = 0.0082; Fig. 4B). In contrast, although the patients whose tumors failed to express FLIPL seemed to do badly at later time points, this group did not show a significantly different association with survival from patients whose tumors expressed weak or moderate expression.
A, Kaplan-Meier plot for disease-specific survival in colorectal cancer patients. FLIPL strongly staining versus moderately staining versus weakly staining versus negative tumors (n = 396). B, Kaplan-Meier plot for disease-specific survival in colorectal cancer patients. FLIPL strongly staining versus low (negative, weakly, and moderately) staining tumors (n = 396).
FLIPS expression. Expression of FLIPS protein was detected in the cytoplasm of all 379 evaluable tumors. In contrast, <1% the cores showed nuclear expression of FLIPS. Twenty-seven percent of cores showed weak staining, 47% showed moderate staining, and 25%, strong staining. No strong associations between FLIPS expression and clinicopathologic variables were noted. On Kaplan-Meier analysis, no significant association was found between expression of FLIPS and survival (Fig. 5 ).
Kaplan-Meier plot for disease-specific survival in colorectal cancer patients. FLIPS strongly staining versus moderately staining versus weakly staining tumors (n = 379).
Multivariate analysis. A multivariate analysis of factors influencing survival was done using the Cox proportional hazards model. Analysis included all available cases. Of the conventional clinicopathologic variables analyzed, significant independent prognostic value was shown for stage and vascular invasion. In addition, strong expression of FLIPL alone was found to be an independent marker of prognosis conferring a significant survival disadvantage as compared with the other patients (hazard ratio, 2.04; 95% confidence interval, 1.18-3.56, P = 0.011; Table 3 ).
Cox multivariate analysis
Discussion
In this study, we assessed the expression of TRAIL-R1, TRAIL-R2, FLIPS, and FLIPL on a large cohort of primary colorectal cancer specimens and evaluated the association between expression of these markers and colorectal cancer–specific survival. Using widely validated antibodies and established techniques, our results indicate that patient stratification by FLIPL phenotype provides tumor stage–independent prognostic information. In contrast, FLIPS showed no prognostic information. In this context, it is of interest that although down-regulation of either FLIPS or FLIPL enhanced chemotherapy-induced apoptosis, only overexpression of FLIPL but not FLIPS inhibited apoptosis, suggesting that FLIPL was the most important splice variant in mediating chemoresistance (13). This study would suggest that FLIPL is also more important as a prognostic indicator in colorectal cancer patients in the absence of chemotherapy because only 5 of 19 patients whose tumors overexpress FLIPL received chemotherapy. At the same time, expression of the TRAIL receptors was not significantly associated with survival in this series. These results suggest that colorectal tumors become resistant to TRAIL-mediated apoptosis by overexpressing FLIPL, and that therapy with TRAIL ligand or agonistic antibodies would be ineffective in this group of patients. A novel form of adjuvant therapy is therefore urgently required for this group of patients. Perhaps the new monoclonal antibody therapies, bevacizumab or cetuximab, which do not mediate killing via the TRAIL pathway, may be useful.
The therapeutic potential of the TRAIL system is supported by the fact that malignant cells are more sensitive to TRAIL-induced apoptosis than their benign counterparts (2, 3, 21). Expression of TRAIL-R1 and TRAIL-R2 has been assessed by tissue microarray in several cancers. In a study on non–small-cell lung cancer patients, it was concluded that the majority of the tumors expressed at least one of the two death receptors for TRAIL (22). In early-stage breast cancer patients, high TRAIL-R2 expression was found to be an independent prognostic marker (23). The expression of FLIPL was investigated, in addition to TRAIL-R1 and TRAIL-R2, in patients with ovarian cancer. Higher TRAIL expression in the surrounding tissue was significantly linked with favorable overall survival in advanced-stage patients (24). In a previous study on a limited number (n = 129) of colon cancer patients, TRAIL-R1 was found to be an independent prognostic factor for stage II/III disease (25). We did not confirm this association in the present study, in which a much greater number of colorectal cancer patients were assessed. In line with our results, they did not find a significant association between TRAIL-R2 and survival. In the present study, we also included FLIPL in our analyses because there is plenty of evidence to support a key role for FLIPL in the TRAIL system (10). A clear correlation between the increase in sensitivity to TRAIL and decreased FLIP levels was observed in human melanoma cell lines (21). Despite the observation of primary keratinocytes being 5-fold less sensitive to TRAIL than human transformed keratinocytes, both cell types exhibited similar TRAIL receptor surface expression. However, when the transformed human keratinocytes were transfected with a FLIPL expression vector, the resistance to TRAIL-mediated apoptosis of these cells increased. These studies indicate that FLIPL is a more important regulator of apoptosis than the TRAIL receptors (26). Our results are in line with these preclinical findings. High expression of FLIPL was indeed found to be an independent prognostic factor in the present study whereas TRAIL-R1 and TRAIL-R2 were not. Because FLIP blocks TRAIL-induced apoptosis, the positive correlation between strong FLIPL expression and poor survival we observed makes sense. To our knowledge, this is the first study investigating the expression of FLIP and its association to survival in colorectal cancer patients.
In summary, we investigated the expression of the key members of the TRAIL pathway in tumors from colorectal cancer patients. The functional TRAIL receptors were not correlated to survival. However, we show for the first time that high expression of FLIPL is associated with a poor prognosis in colorectal cancer patients. Colorectal cancer patients strongly expressing this marker could be considered for more aggressive treatment.
Acknowledgments
We thank R. Moss for technical assistance.
Footnotes
-
Grant support: Special Trustees of Nottingham Hospitals grant STR/03/M and The Research Foundation Stiftelsen Onkologiska Klinikens i Uppsala Forskningsfond.
-
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.
- Accepted June 27, 2007.
- Received October 30, 2006.
- Revision received June 7, 2007.
References
Volume 13, Issue 17
