Cancer Cell Biology
Serum soluble Fas levels and prediction of response to platinum-based chemotherapy in epithelial ovarian cancer
Version of Record online: 18 DEC 2007
Copyright © 2007 Wiley-Liss, Inc.
International Journal of Cancer
Volume 122, Issue 8, pages 1716–1721, 15 April 2008
How to Cite
Chaudhry, P., Srinivasan, R., Patel, F. D., Gopalan, S. and Majumdar, S. (2008), Serum soluble Fas levels and prediction of response to platinum-based chemotherapy in epithelial ovarian cancer. Int. J. Cancer, 122: 1716–1721. doi: 10.1002/ijc.23213
- Issue online: 19 FEB 2008
- Version of Record online: 18 DEC 2007
- Manuscript Accepted: 17 AUG 2007
- Manuscript Received: 21 MAY 2007
- Indian Council of Medical Research. Grant Number: 3/1/JRF/37/MPD/02
- epithelial ovarian cancer;
- platinum-based chemotherapy;
Epithelial ovarian cancer (EOC) is treated mainly by platinum-based combination chemotherapy. Chemotherapy induces apoptosis in which the Fas/Fas ligand pathway is important. Serum soluble Fas (sFas) is a biomarker of this pathway and functionally inhibits Fas-/FasL-mediated apoptosis. In this study, we have investigated the role of sFas in prediction of response to chemotherapy in EOC. Thirty-five patients were recruited and their serum sFas levels were estimated by ELISA at 4 time points—preoperative (sFas1), postoperative (sFas2), midchemotherapy (sFas3) and at the end of chemotherapy (sFas4). The response to chemotherapy was documented clinically, radiologically and by CA-125 levels, based on which, 2 groups were identified: primary chemosensitive (n = 24) and primary chemoresistant (n = 11). Based on the disease status at last follow-up, 2 groups were identified: No Evidence of Disease (n = 15) and Evidence of Disease (n = 20). The primary chemoresistant tumors showed significantly higher median sFas2 levels (p = 0.033) with the sFas2/sFas1 ratio ≥1 (p = 0.001). A multivariate Cox proportional hazards regression model identified sFas2/sFas1 ratio as a significant factor for the prediction of response to platinum-based chemotherapy (p = 0.011). Receiver operating characteristic (ROC) analysis showed that at a ratio of 1.2, sFas2/sFas1 achieved a sensitivity of 82% and specificity of 100% for prediction of chemotherapeutic response. sFas2/sFas1 and sFas3/sFas1 ratio was also higher in patients with evidence of disease (p = 0.018 and p = 0.028, respectively). Progression-free survival rates in patients with sFas2/sFas1 ratio <1 exceeded those with ratio ≥1 (p = 0.004). In conclusion, serum sFas is a useful biomarker for predicting response to platinum-based chemotherapy in EOC. © 2007 Wiley-Liss, Inc.
Epithelial ovarian cancer (EOC) is the leading cause of death among the female genital tract malignancies.1 The high mortality of ovarian carcinoma is partly due to the fact that most patients are detected when the disease is at an advanced stage (III and IV) and partly because the current chemotherapeutic regimes are ineffective.2 The 5-year survival rate of these women is only 20–30%. Currently, the preferred treatment regimen for ovarian cancer is combination chemotherapy; usually a platinum-based drug, such as cisplatin or carboplatin, coupled with paclitaxel. While this treatment course shows promising effects in a high percentage of cases, the development of chemoresistance is a hurdle that significantly hinders successful treatment outcomes.3 The ability of a cancer cell to respond to a chemotherapeutic agent is believed to be due, in part, to its apoptotic capacity. For example, cisplatin has been shown to upregulate the proapoptotic factors p53, Fas, and Bax in a number of cell types.4–6 The extrinsic death receptor pathway is activated by members of the tumor necrosis factor (TNF) family and the receptors for these ligands (TNFR).7 Among the death receptor pathways, the Fas/FasL pathway is the best-characterized pathway.8
Fas is a Type I membrane protein, which upon triggering by Fas ligand (FasL), induces apoptosis through a series of interacting proteins. Fas is characterized by a highly conserved death domain (DD) that is responsible for activating the death signal upon activation by FasL.9, 10 A multimolecular complex of proteins called the death-inducing signaling complex (DISC) is triggered by receptor ligand crosslinking.11, 12 The apoptotic signaling is propagated by Procaspase-8 and further by active caspase-8 and caspase-3.13, 14 Caspase-mediated proteolysis of specific protein targets is central to the execution of apoptosis.
Fas can be present both as a cell surface protein and as a soluble protein called soluble Fas (sFas). sFas is a splice variant generated by alternative mRNA splicing and lacks a transmembrane domain.15 Papoff et al.16 have reported that sFas suppresses Fas-mediated apoptosis by competitive binding with FasL. Increased sFas levels in hepatocellular carcinoma, bladder, breast, renal cell carcinoma and melanoma have been documented.17–21 sFas has been investigated as a prognostic marker in gynecological malignancies by some workers who have correlated it with more invasive, advanced stage tumors.22, 23
Involvement of sFas in prediction of response to therapy has not been documented as yet. So the aim of our study was to evaluate the serum sFas levels in primary EOC patients using ELISA and examine the relationship between sFas levels and response to platinum-based chemotherapy. We hypothesized that the levels of serum sFas in EOC may be used as a predictor of response to platinum-based chemotherapy and may add useful information along with CA-125, an established tumor marker.
Material and methods
Patients and samples
This study was approved by the Institutional Review Board of the Postgraduate Institute of Medical Education and Research (PGIMER, Chandigarh, India) and all ethical guidelines were followed. A formal written informed consent for participation in this study was obtained from all the patients. We included only those patients in the study who had no history of prior ovarian surgery or chemotherapy of any kind. Thus, a total of 35 patients with primary epithelial ovarian cancer served as subjects for this study. The patients underwent exploratory laparotomy in the Department of Obstetrics and Gynaecology, between July 2003 and July 2005. They were subjected to detailed histopathologic analysis at the Department of Cytology and Gynaecological Pathology; the diagnosis was confirmed according to the World health Organization (WHO), criteria and the tumors were staged as per the Federation International Gynecological Oncologists classification (FIGO) system, respectively.24 The median age at the time of diagnosis of ovarian cancer was 48 years (range 26–61). Optimal debulking of <2 cm residual disease was achieved in 8 patients, and 27 women had gross residual disease after surgery.
From each patient, 5 ml of blood was collected for serum sFas estimation at 4 time points—(i) preoperatively (sample 1 or sFas1), (ii) postoperatively (4 weeks after surgery) (sample 2 or sFas2), (iii) midchemotherapy, i.e., after completion of 3rd cycle of chemotherapy (sample 3 or sFas3) and (iv) postchemotherapy, i.e., after completion of 6 cycles of chemotherapy (sample 4 or sFas4). The serum was separated and stored at −80°C until examination. In addition to sFas, the serum CA-125 levels were also evaluated.
Estimation of soluble Fas
For measuring serum sFas levels, a commercially available enzyme-linked immunosorbent assay (ELISA) was used. The sFas levels in serum of primary epithelial ovarian cancer patients were determined using BD OptEIA™ Human sFas kit (BD Biosciences, CA) as per manufacturer's instructions. The assay employs the quantitative sandwich enzyme immunoassay technique. All serum sFas analyses were done at the same time, in the same batch, and in duplicate. Two-fold diluted serum was used for determination, and the standard curve was constructed using serial dilutions of recombinant sFas/CD95 at concentrations from 31.25 to 2,000 pg ml−1.
Briefly, the microplate was coated with diluted capture antibody (antihuman Fas monoclonal antibody), and the plate was incubated overnight at 4°C. After washing, the microplate was blocked with assay diluent (PBS with 10% FBS, pH 7.0) for 1 hr at room temperature. Standards and samples were then pipetted into the wells. The immobilized antibody in the well binds any sFas present. The microplate was kept for 2 hr at room temperature. After washing away any unbound substances, an enzyme-linked antibody specific for sFas (biotinylated antihuman Fas monoclonal antibody + avidin HRP conjugate) was added to the wells. Following a wash to remove any unbound antibody-enzyme reagent, TMB substrate (tetramethylbenzidine and H2O2) (BD PharmingenTM TMB Substrate Reagent Set, BD Biosciences, CA) was added to the wells. The color developed is in proportion to the amount of sFas bound in the initial step. The color development was stopped by adding stop solution (2 N H2SO4), and the intensity of the color was measured at 450 nm within 30 min of stopping the reaction with λ correction of 570 nm.
To determine the normal range for this kit, sFas was measured in 10 normal age-matched healthy women. The sFas levels in these healthy controls ranged from 0.4 to 0.9 ng ml−1 (Median, 0.6 ng ml−1).
Evaluation of response to chemotherapy
Following surgery, all patients were treated with platinum-based chemotherapy at the Department of Radiotherapy, PGIMER, Chandigarh. Out of 35 patients included, 28 (80%) patients received a combination of cisplatinum and cyclophosphamide and 7 (20%) patients received cisplatinum and paclitaxel. The reason for the choice of 2 types of treatment depended on affordability, since pacitaxel is an expensive drug and most of the patients recruited in this study could not afford paclitaxel. All patients received 6 cycles of chemotherapy. Prior to each cycle, patients were assessed clinically and radiological examinations were ordered if necessary. In addition, the patients were followed up and the course documented. At the end of 6 cycles of chemotherapy (∼6–7 months from initiation of therapy), based on their response using previously described criteria,25, 26 the patients were divided into 2 groups: (i) primary chemosensitive (n = 24)—if there was regression in size of the tumor as determined clinically or radiologically after the initiation of first-line platinum-based combination chemotherapy and no relapse or progression noted; (ii) primary chemoresistant (n = 11)—if the patient had persistent or progressive disease after the initiation of first-line platinum-based combination chemotherapy. Patients were further followed up ranging from 12 to 33 months, with a median follow-up of 18 months, on the basis of which 2 groups were identified: (i) No Evidence of Disease (n = 15) and (ii) Evidence of Disease including Stable/Progressive Disease (n = 20).
Mann–Whitney U test was done for comparison of the sFas levels in different groups. Multivariate analysis was performed using the Cox proportional hazards regression model. Receiver operating characteristic (ROC) curves were constructed to determine the optimal values of sFas, which provided high sensitivity and specificity. The log-rank test was applied and the progression-free survival curves were obtained by the Kaplan–Meier method. Correlation between postoperative sFas levels (sFas2) and postoperative CA-125 was done using Pearson's correlation. A probability value of p < 0.05 was considered to be significant. All statistical analyses were carried out using SPSS version13.0.
The characteristics of the patients included in the study are detailed in Table I.
|Characteristic||No. of patients (N = 35)|
|Age (range)||26–61 years|
|Postoperative residual tumor|
|Clinical response to chemotherapy|
|Disease status at last follow-up|
|No Evidence of Disease (NED)||15|
|Evidence of Disease (ED)||20|
The patients were stratified according to the chemotherapeutic combination received (cisplatinum/cyclophosphamide vs. cisplatinum/paclitaxel) and compared for the primary chemotherapeutic response as well as disease-status at last follow-up; there was no statistically significant difference in the 2 groups (Fisher's exact test, p = 0.619 and p = 0.668, respectively).
Serum soluble Fas levels and primary response to chemotherapy
sFas levels were determined in each patient at 4 time points indicated in the Material and Methods section. The sFas levels were variable and ranged from 0.8 to 7.8 ng ml−1 in patients with EOC measured at 4 different time points. In general, peak sFas levels were observed at the midchemotherapy point. The mean ± S.D and median values of sFas levels in the primary chemosensitive and primary chemoresistant groups are shown in Table II. The distribution of sFas levels in these 2 groups is shown in Box plots (Fig. 1). The postoperative (sFas2) sFas levels were significantly lower in primary chemosensitive cases compared to the primary chemoresistant patients (Mann–Whitney U test, p = 0.033). There was no significant difference in sFas levels at the other time points, viz., prechemotherapy, midchemotherapy or at the end of chemotherapy. We also calculated the fold changes in sFas levels, i.e., sFas2/sFas1, sFas3/sFas1 and sFas2/sFas3. It was found that higher sFas2/sFas1 and sFas2/sFas3 ratios correlated to primary chemoresistant phenotype (p = 0.001 and p = 0.021, respectively). A multivariate Cox proportional hazards regression model demonstrated that sFas2/sFas1 ratio was a significant factor for prediction of response to platinum-based chemotherapy (p = 0.011). No relationship was observed between serum sFas level measured at four different time points to the age of the patient (p = 0.936, 0.206, 0.054, 0.825, respectively) or to the presence of gross residual disease (p = 0.142, 0.325, 0.693, 0.365, respectively). The sFas2/sFas1 ratio also did not correlate with either the age (p = 0.371) or the presence of gross residual disease (p = 0.584).
|Primary chemosensitive||Primary chemoresistant||p value*|
|Mean ± S.D.||Median||Mean ± S.D.||Median|
|sFas1 (ng ml−1)||2.64 ± 0.91||2.84||2.29 ± 1.06||2.15||0.214|
|sFas2 (ng ml−1)||2.11 ± 0.88||1.89||3.11 ± 1.27||3.09||0.033|
|sFas3 (ng ml−1)||3.54 ± 1.42||3.24||3.72 ± 1.86||3.39||0.736|
|sFas4 (ng ml−1)||3.06 ±1.46||2.90||2.89 ± 1.56||2.98||0.859|
|sFas2/sFas1||0.80 ± 0.17||0.83||1.43 ± 0.50||1.38||0.001|
|sFas3/sFas1||1.43 ± 0.59||1.31||1.87 ± 1.34||1.52||0.477|
|sFas2/sFas3||0.67 ± 0.43||0.60||1.01 ± 0.61||0.85||0.021|
As sFas2 and sFas2/sFas1 ratio achieved significance, ROC curves were constructed to determine the optimal values, which provide high sensitivity and specificity (Fig. 2). The area under curve (AUC) was significant for the sFas2 values alone (AUC = 73%, p = 0.033) as well as for the sFas2/sFas1 ratio (AUC = 92%, p = 0.001). The sensitivity and specificity were calculated for each possible threshold value of estimated probability for primary chemosensitive and primary chemoresistant group. The sFas2 level, which achieved an optimal sensitivity of 82% and specificity of 67%, was 2.3 ng ml−1. Similarly, sFas2/sFas1 ratio achieved an optimal sensitivity of 82% and specificity of 100% at 1.2 for prediction of chemotherapeutic response.
Serum soluble Fas levels and progression-free survival
The sFas levels in No Evidence of Disease versus the Evidence of Disease groups are depicted graphically in Box plots (Fig. 3) and in Table III. Statistically no correlation of sFas levels to progression-free survival could be demonstrated. However, the fold change in sFas levels, i.e., sFas2/sFas1 and sFas3/sFas1 was higher in patients with progressive disease as compared to those with No Evidence of Disease (Mann–Whitney U test, p = 0.018 and p = 0.028, respectively). ROC curves were constructed to determine the optimal values, which provide high sensitivity and specificity (Fig. 4). The AUC was not significant for sFas2 levels, but was significant for the sFas2/sFas1 ratio (AUC = 74%, p = 0.018). Sensitivity and specificity were calculated for each possible threshold value of estimated probability for No Evidence of Disease and Evidence of Disease groups. At a ratio of 0.89, sFas2/sFas1 achieved a sensitivity of 70% and specificity of 74% for prediction of evidence of disease. Further, applying the log-rank test, the progression-free survival rates in patients with sFas2/sFas1 ratio <1 (18 months) exceeded those with ratio ≥1 (11 months) (p= 0.004, Log rank test) as shown in Figure 5.
|No evidence of disease||Evidence of disease||p value*|
|Mean ± S.D.||Median||Mean ± S.D.||Median|
|sFas1 (ng ml−1)||2.77 ± 0.92||2.84||2.35 ± 0.97||2.15||0.177|
|sFas2 (ng ml−1)||2.23 ± 0.87||1.92||2.57 ±1.26||2.61||0.605|
|sFas3 (ng ml−1)||3.17 ± 1.21||2.81||3.92 ± 1.72||3.47||0.172|
|sFas4 (ng ml−1)||2.70 ± 1.09||2.45||3.23 ± 1.69||2.98||0.424|
|sFas2/sFas1||0.81 ± 0.17||0.83||1.14 ± 0.50||0.99||0.018|
|sFas3/sFas1||1.22 ± 0.54||1.04||1.83 ± 1.03||1.60||0.028|
|sFas2/sFas3||0.79 ± 0.50||0.65||0.76 ± 0.53||0.61||0.641|
CA-125 and presence/absence of gross residual disease do not predict response to chemotherapy
The levels of CA-125, an established tumor marker, was evaluated postoperatively in the patients and analyzed. There was no significant change in the postoperative CA-125 levels in the primary chemosensitive and primary chemoresistant groups (p = 0.925). Neither were the levels different in the No Evidence of Disease and Evidence of Disease groups (p = 0.737). Further, using Pearson's correlation coefficient, we observed that there was no correlation between postoperative CA-125 and sFas levels, respectively (p = 0.863). There was also no correlation of outcome (Evidence of or No Evidence of Disease) to the presence or absence of gross residual disease by the Pearson's χ2 test (p = 0.642).
The Fas family of proteins are key members of the extrinsic pathway of apoptosis. In the serum, the levels of sFas, an alternatively spliced variant of Fas, may be measured. It has been evaluated as a tumor marker in several recent studies, which show increased levels correlating to a poor outcome in patients with cancer of the liver, urinary bladder, breast, kidney, ovary and in non-Hodgkin's lymphoma.17–20, 22, 23, 27 Nevertheless, some studies found elevated sFas serum levels lacking a prognostic relevance in cancer patients.28, 29 In ovarian cancer, there are only a few studies regarding serum sFas. An increase in sFas levels in malignant ovarian cancer as compared to benign neoplasms and normal control subjects has been reported by some authors22, 23 but refuted by others.30
Epithelial ovarian cancer (EOC) is mainly treated by platinum-based chemotherapy, and so we evaluated sFas as a marker of chemotherapeutic responsiveness in such patients. This is the first study where sFas levels were estimated during the course of chemotherapy and the levels were correlated to response to chemotherapy. The present study demonstrated a highly significant correlation of elevated serum sFas level with primary resistance to platinum-based chemotherapy in advanced epithelial ovarian cancer. It is also an independent predictive factor for response to platinum-based chemotherapy by multivariate analysis.
In general, a strong increase in serum sFas concentration was observed upon treatment with chemotherapeutic drugs with peak levels in the midchemotherapy point of time (sFas3) in all the groups irrespective of the response to therapy. A possible explanation for this might be the upregulation of Fas expression in tumor cells induced by cytotoxic agents, as described previously for colon carcinoma cells31 and its release into the circulation during chemotherapy induced apoptosis. Ugurel et al.21 have shown that increased sFas serum concentration correlated with the stage of disease, which suggests that tumor tissue itself might be one possible source of sFas production. Hence, it is possible that the increased sFas serum levels in primary chemoresistant patients observed in this study could reflect the tumor burden present in these patients; however, statistically, there was no correlation to the chemotherapeutic response to either the amount of gross residual disease estimated clinically or to the CA-125 levels. Therefore, sFas may be derived from both tumor and nontumor sources. The possible nontumor sources of sFas could be the stromal tissue surrounding the tumor or even the peripheral blood lymphocytes and cells of the activated immune system.32 It is conceivable that cytotoxic chemotherapy can lead to activation of the immune system consequently to tumor apoptosis/necrosis and release of tumor antigens in the circulation. Stumm et al.33 in their study on breast cancer-derived cell lines observed that paclitaxel downregulated both sFas and surface Fas receptors by unexplained mechanisms. This is refuted by another study on breast cancer patients treated with chemotherapy who displayed higher sFas levels as compared to untreated patients or patients treated with hormonal therapy/chemoendocrine therapy.19 The differential ability of different therapeutic agents to regulate Fas is highlighted by yet another study, wherein melanoma patients who received IFN-γ treatment showed increased sFas levels as compared to those who received IFN-α treatment.21
It was reported that sFas levels increase with age.30 There was no correlation of sFas levels with the age of the patient in the present study. Further, no significant difference was observed in the median age of primary chemosensitive (47 years) versus primary chemoresistant patients (51 years) (p = 0.156).
In the present study, it was observed that the postoperative (sFas2) sFas levels were significantly lower in primary chemosensitive cases compared to the primary chemoresistant patients. A multivariate Cox proportional hazards ratio model also showed that the sFas2/sFas1 ratio to be an independent predictor of the primary or initial response to chemotherapy. Receiver operating curve (ROC) analysis showed that at a value of 2.3 ng ml−1, sFas2 achieved a sensitivity of 82% and specificity of 67%, and at a ratio of 1.2, sFas2/sFas1 achieved a sensitivity of 82% and a specificity of 100% for prediction of chemotherapeutic response. We also found a correlation of increased postoperative prechemotherapy sFas (sFas2) levels with the progression-free survival. The ratio of the postoperative to the preoperative sFas (sFas2/sFas1 ratio) of ≥1 was associated with evidence of disease and therefore a poor progression-free survival. The ROC analysis revealed a higher sensitivity and specificity of this ratio rather than the sFas2 levels alone in prediction of disease status. Higher sFas levels have correlated with poorer survival previously in a study by Konno et al. (22) who suggested a cut-off sFas level of ≥1.5 ng ml−1. On the other hand, Hefler et al.23 showed that a cut-off value of 3.7 ng ml−1 for pretreatment sFas was associated with a shortened disease-free survival and overall survival. However, the pretreatment sFas levels did not correlate with primary chemosensitivity or to progression-free survival, although the sFas2/sFas1 ratio did predict progression-free survival implying its importance in prognosis of epithelial ovarian cancer. Our study is in broad agreement with these studies although optimal sensitivity and specificity were achieved only at 2.3 ng ml−1. Overall, our observations suggest that the sFas levels may serve as a useful marker in predicting the primary response to chemotherapy rather than progression-free survival. Again, these initial observations are very encouraging and need confirmation on a larger sample size.
Fas-mediated apoptosis has been demonstrated to be involved in anticancer drug-induced apoptosis, and a poor response to platinum-based chemotherapy in patients with higher sFas levels is reported previously.31, 34 The biological role of sFas is thought to involve the binding and neutralization of either soluble or cell-surface FasL. It is possible that sFas forms complexes with cell-surface Fas to prevent or alter signal transduction35, 36 and thus plays an important role in the regulation of apoptosis as an inhibitor of Fas-mediated apoptosis.15, 16, 35 Therefore, it is conceivable that elevated sFas production may promote tumorigenesis and disease progression. It is also possible that high level of serum sFas may enable the tumor cells to evade from immunosurveillance by tumor-infiltrating lymphocytes15 and thereby correlating to more aggressive disease.
In conclusion, this study demonstrates that sFas is a useful marker in prediction of the primary response to platinum-based chemotherapy as well as to the disease outcome. As the number of cases analyzed is small, the findings in this pilot study need to be corroborated by large-scale multicentric studies to firmly establish the clinical significance of sFas in advanced epithelial ovarian cancer.
Parvesh Chaudhry was funded by the Indian Council of Medical Research Fellowship under the grant no. 3/1/JRF/37/MPD/02.