Purpose: Valid molecular markers need to be implemented in clinical trials to fulfill the demand of a risk-adapted and more individualized multimodal therapy of locally advanced primary rectal cancer. In this study, the expression of the inhibitor-of-apoptosis (IAP) protein survivin was evaluated in pretreatment biopsies and corresponding posttreatment resection specimens, and was correlated to histo-pathological tumor characteristics and clinical follow-up.
Patients and Methods: One hundred sixteen patients with stage II/III rectal cancer treated with 5-FU–based neoadjuvant radiochemotherapy (RCT) at a single university medical centre within the German Rectal Cancer Trials were investigated. Survivin expression in pretreatment biopsies and surgical resection specimens were determined by immunohistochemistry by two independent institutions and correlated with histopathologic parameters, tumor recurrences, disease-free (DFS), and overall cancer-specific survival (CSS).
Results: In pretreatment biopsies, a higher survivin expression correlated with advanced ypT (P = 0.026) and ypUICC (P = 0.05) stage as well as DFS (P = 0.038) after preoperative RCT. High posttreatment survivin levels were associated with advanced ypT stage (P = 0.03) and residual lymph node metastases (P = 0.04). Moreover, neoadjuvant RCT resulted in a significant downregulation of survivin expression (P < 0.0001). A failure of RCT-induced downregulation was associated with development of distant metastases (P = 0.0056) and cancer-related death (P = 0.026), and correlated significantly with DFS (P = 0.011*/0.02**) and CSS (P = 0.0017*/0.01**) in uni-* and multivariate** analyses.
Conclusions: Survivin expression displays a marker with prognostic utility in rectal cancers. These results underline the potential of survivin to monitor individual response to RCT and encourage anti-survivin strategies in multimodal rectal cancer therapy within future randomized clinical trials. Clin Cancer Res; 17(6); 1623–31. ©2010 AACR.
Overexpression of survivin has been shown to be associated with poor clinical outcome and aggressive behavior in a variety of human tumors, indicating its relevance as a prognostic marker and its potential as a target for molecular cancer therapy. Our actual data indicate that rectal cancer patients with low pretreatment survivin levels show a significantly prolonged disease-free survival after neoadjuvant radiochemotherapy (RCT). Moreover, by comparison of pre- and posttreatment survivin levels, an elevated expression and failure of RCT-induced downregulation of the protein was related to a higher incidence of distant metastatic disease and cancer-related death. Thus, survivin seems to be relevant not only for response of the primary lesion, but also for tumor dissemination and systemic tumor control within multimodal treatment of rectal cancer. As a growing number of survivin antagonists recently entered clinical phase I/II studies, these results provide a further rationale for the implementation of anti-survivin strategies within future randomized clinical trials in rectal cancer patients.
Neoadjuvant radiochemotherapy (RCT) followed by total mesorectal excision (TME) has become standard treatment in locally advanced UICC (Union International contre le Cancer) II/III rectal cancer within the last decade (1). Individual treatment response, however, varies considerably. As shown in our previous studies, histopathological tumor regression ranges from a complete response with no viable tumor cells detectable to virtually no regression at all (2). Despite uniform treatment protocols, this observed variance in tumor response is most probably caused by differences in the genetic profile expressed in the individual lesions (3, 4), and might influence long-term survival (5). Reliable molecular markers may help to individually tailor multimodal treatment regimens by risk-adapted or novel therapeutic approaches.
Survivin, the smallest member of the inhibitor-of-apoptosis (IAP) family is a multifunctional protein that is reported to interact at the crossroads of disparate molecular networks of cellular division, intracellular signaling, and apoptosis (6, 7). In this context, one of the signature features of survivin is not only its ability to associate with multiple protein partners, but also to localize to disparate subcellular compartments (8, 9). Abundantly, overexpressed in nearly every tumor tested so far, survivin is associated with a more aggressive tumor behavior and worse clinical outcome (10). In addition, survivin is involved in resistance to both chemotherapy and ionizing irradiation (11). Several preclinical in vitro studies have demonstrated that targeting survivin's expression by the use of small interfering RNAs, antisense-oligonucleotides and small molecule repressors radiosensitize tumor cells and reduces tumor growth in vivo. Because of these properties, survivin has been proposed as a molecular target for anticancer therapies (12, 13).
Although the role of survivin as a prognostic marker is consistent within the literature, its value as a marker for the prediction of treatment response toward radiation therapy or RCT remains controversial (14). In rectal cancer, a high survivin expression in pretreatment biopsies of patients treated with 5-FU–based preoperative RCT or short-course radiotherapy was significantly associated with a higher risk of local tumor recurrences (15) and decreased survival (16), respectively. However, other studies with smaller and more heterogeneous patient cohorts reported that there were no correlations between survivin mRNA or protein expression and histopathological tumor regression grades as well as clinical outcome (17, 18). This discrepancy may in part arise from the use of different methods of survivin evaluation, the use of different clinical endpoints or different quantitative scoring systems of the expression characteristics, as well as variable stratification of the patient groups.
Recent data indicate that in oesophageal and ovarian cancer intratumoral survivin expression decreased significantly during neoadjuvant chemotherapy or RCT (19, 20). In contrast, elevated postoperative survivin levels after neoadjuvant treatment were significantly associated with a higher tumor stage, poor histopathological response, and reduced cancer-specific survival.
In this study, survivin expression was—for the first time—evaluated in both pretreatment biopsies and corresponding surgical specimens from patients with locally advanced rectal cancer, treated with standardized preoperative RCT within randomized clinical phase III trials. Methodical and interobserver accordance were evaluated between 2 independent institutions, and survivin expression levels were correlated with clinicopathologic characteristics and clinical follow-up.
Patients and Methods
One hundred and sixteen patients with locally advanced UICC II/III rectal adenocarcinoma were included in this study. All patients underwent standardized preoperative RCT and quality-assessed curative surgery within the randomized phase III German Rectal Cancer Trials CAO/ARO/AIO-94 and CAO/ARO/AIO-04 (1). Approval from the ethics committee of the University of Göttingen and informed consent for additional translational research projects were obtained from all patients before enrollment into this study. Staging procedures including rigid rectoscopy with endorectal ultrasonography; contrast enhanced CT scans of thorax, abdomen, and pelvis; and MRI imaging of the pelvis were performed in all cases to confirm locally advanced tumor stage and to exclude patients with evidence of distant metastatic disease at the time of initial rectal cancer diagnosis.
Treatment and pathological staging procedures
Treatment procedures consisted of neoadjuvant RCT according to the protocol of the German Rectal Cancer Trials (CAO/ARO/AIO-94 and -04). All patients received a total irradiation dose of 50.4 Gy in 28 fractions of 1.8 Gy with concomitant application of 5-fluorouracil (73 patients;63%) or combined 5-flourouracil/oxaliplatin (43 patients; 37%) chemotherapy. Surgery was scheduled within 6 weeks after completion of RCT. Surgical procedures comprised 83 low anterior resections (72%), 32 abdomino-perineal resections (28%), and 1 discontinuous resection (Hartmann procedure), all including quality-assessed TME.
Pretreatment biopsy material was taken from different representative luminal tumor sites and stored within the prospective translational biobank of the Department of General and Visceral Surgery of the University Medical Center Göttingen. Pathological staging in posttreatment tumor resections was performed according to the respective TNM classification (21). Histopathological tumor regression was denoted on the basis of a semiquantitative 5-point grading system described by Dworak and colleagues (22): Grade 0: no regression; Grade 1: dominant tumor mass with obvious fibrosis or mucin; Grade 2: dominantly fibrotic or mucinous changes with few tumor cells or cell groups; Grade 3: very few tumor cells in fibrotic or mucinous tissue; and Grade 4: no tumor cells, complete fibrotic or mucinous regression.
Immunohistochemical staining of survivin
Immunhistochemical staining of survivin was performed on rectal cancer biopsies and on the corresponding surgical specimens. In 11 patients (9.5%) with pathologically confirmed complete response (pCR) after neoadjuvant RCT staining for survivin was only accessible on the pretreatment biopsies. Immunhistochemical staining was performed by alternative methods in 2 institutions (Göttingen and Frankfurt) on consecutive tissue slides using the polyclonal rabbit antihuman survivin antibody AF886 (R&D Systems).
Staining procedure Göttingen
Tissue sections were mounted on microscope slides (Star Frost) and subjected to an automatic staining procedure with standardized tissue preparation, epitope retrieval (98°C for 60 minutes) and antibody incubation (42°C for 90 minutes, 1:750) using a Benchmark XT Autostainer (Ventana Medical Systems SA). Alkaline phosphatase (Red Detection Kit, Ventana Medical Systems) was used for color development and visualization of the epitope–antibody reaction product.
Staining procedure Frankfurt
Staining of survivin was performed as described previously (23). Briefly, before labeling, tissue sections heated for 20 minutes in a pressure cooker, rinsed in Tris-puffered saline (TBS) and nonspecific binding sites were blocked with 5% nonfat dry milk in TBS. Primary anti-survivin antibodies were applied at a 1:50 dilution and incubated at 4°C overnight. The slices were next incubated with biotinylated goat anti-rabbit secondary antibody (Dianova, 1:50 dilution) followed by a strepavidin/biotinylated horseradish perioxidase complex (Dako). Finally, 3-amino-9-ethyl-carbazole (AEC) solution was used as chromogen and hematoxylin (37%) for counterstaining.
Scoring of survivin expression
Survivin expression was analyzed by an individual labeling score considering percent of positive cells and staining intensity. Intensity was scored as: 1+ (weak), 2+ (moderate), and 3+ (intense). The fraction of tumor cells with survivin positivity was assigned to: 1 (0–25%), 2 (25–50%), 3 (50–75%), and 4 (>75%). The individual labeling score for survivin results from the multiplication of staining intensity and number of positive cells, and ranges from 0 (no positive tumor cells) to 12 (>75% of tumor cells with intense staining). Initial scoring was independently performed in each institution. Next, results of both institutions were correlated to evaluate intratumor and interobserver variances as well as availability and comparability of the survivin labeling score (Supplementary Fig. S1). Subsequently, a consensual score was performed based on collaborative re-evaluation of all cases by 2 investigators of each institution (T.S., F.R.). This score served for correlation of survivin expression with clinicopathological parameters and survival.
Clinical follow-up and criteria for relapse
All patients were re-evaluated at 3-month intervals for 2 years and every 6 months thereafter. Evaluations consisted of pertinent medical history, physical examination, blood cell counts, and blood chemistry including carcinoembryonic antigen (CEA) levels at every follow-up visit. Rigid recto-sigmoidoscopy (in patients treated with anterior resection) was performed at 3-month intervals in the first year, at 6-month intervals in the second year, and once per year thereafter. A follow-up schedule for abdominal ultrasound, computerized tomography studies of the abdomen and pelvis, and chest X-rays was defined at regular intervals within the protocol of the German Rectal Cancer Trials. Histological confirmation of local recurrence and distant relapse (defined as tumor manifestation outside the pelvis) was encouraged. Alternate acceptable criteria included sequential enlargement of a mass in radiologic studies with simultaneous increase of serum CEA levels. DFS was defined as the time from surgery to detection of any tumor relapse. Accordingly, CSS was defined as the time from surgery to death due to malignant disease. Median follow-up time was 66.5 months.
Statistical analysis was performed using the open-source statistical computing software R and the packages exactRankTests and survival. Correlation of the survivin labeling score with clinical outcome variables was computed by the Spearman Rank correlation using the function cor.test. Two group comparisons were performed using a Wilcoxon Rank Test. Time to event data were visualized using Kaplan-Meyer analysis and significance was computed based on Cox proportional hazards regression models using the function coxph when using the ordinal survivin score (0–12) or based on a log-rank test using the function survdiff when using the dichotomized survivin score. In addition, the influence on the following variables: gender, age, TRG, ypT, ypN, and ypUICC stage on survival was assessed in univariate analysis, because of these variables only gender and ypN reached significance, multivariate analysis was performed using these variables and posttreatment survivin score in a Cox model. The level of significance was set to α = 5% for all tests.
Clinical staging results of all 116 patients are presented in Table 1. Therapy- and surgery-related 30-day morbidity and mortality were 22% and 0%, respectively. In initial staging, 85 patients (73%) had evidence of mesorectal lymph node involvement (cN+). After preoperative RCT, 44 patients (38%) presented with residual nodal metastases (ypN+). Within the follow-up period, cancer relapse occurred in 26 patients (22%). Isolated distant metastatic disease occurred in 22 patients (19%), whereas simultaneous local and distant recurrence was observed in an additional 4 patients (3%). There was no case of isolated local cancer recurrence. Cancer-related death within the follow-up period occurred in 13 patients (11%). Pathologic staging results and tumor recurrences are summarized in Table 1.
Immunostaining of survivin on pretreatment biopsies and surgical resections
Survivin immune-reactivity was observed in all investigated tumor biopsies and the corresponding surgical specimens. Survivin positivity was predominately homogeneous within biopsy as well as posttreatment tumor samples. The fraction of tumor cells with positive survivin immunoreactivity ranged between 25% and 100% in all but 1 case. Staining intensity varied between different cases as well as between the matched biopsy–resection specimen pairs of individual patients. Staining of noncancerous tissue became apparent as slight positivity of mucosal epithelium in the apex of the rectal crypts and as lymphocytic nuclear positivity in a few cases. Table 2 displays percentages of immunopositive cells, staining intensity, and individual labeling score in all 116 rectal cancer biopsies and the corresponding 105 surgical resection specimens.
There was a high concordance of the intratumor staining characteristics (not shown) and interobserver variability results obtained by 2 different institutions. For the survivin score in pretreatment biopsies, the interobserver correlation coefficient was 0.742 (P < 0.0001) and 0.831 (P < 0.0001) for the corresponding posttreatment tumors (Supplementary Fig. S1).
Association of pretherapeutic survivin expression with clinicopathologic parameters and survival
The survivin score in pretherapeutic biopsies was significantly correlated with posttherapeutical tumor stage (ypT) (P = 0.026) as well as ypUICC stage (P = 0.05), indicating that high pretreatment survivin expression was associated with unfavorable histopathological characteristics following preoperative RCT. Moreover, patients with low pretreatment survivin levels (individual labeling score: ≤6) showed increased DFS-rates compared with patients with high initial (>6) survivin scores (P = 0.038, Fig. 1).
Downregulation of survivin expression following neoadjuvant radiochemotherapy
As compared with pretherapeutic biopsies, a highly significant downregulation of survivin expression (P < 0.0001) was observed after neoadjuvant RCT (Fig. 2A). Figure 2B shows an example of an intensively stained pretreatment biopsy with more than 75% of positive tumor cells (individual labeling score: 12) and its corresponding surgical resection specimens, showing decreased survivin staining intensity (individual labeling score: 4). Stratified by neoadjuvant treatment arm 50.4 Gy + 5-FU (n = 73) versus 50.4 Gy + 5-FU + oxaliplatin (n = 43) no significant difference in downregulation of survivin expression was observed (P = 0.3828).
Association of posttherapeutic survivin levels with clinicopathologic parameters and survival
In rectal cancer specimens after neoadjuvant RCT, high expression levels of survivin were significantly associated with advanced ypT stage (P = 0.03) and the presence of residual nodal metastases (ypN+) (P = 0.04). In addition, a high survivin level correlated to the development of distant metastases (P = 0.0056) and cancer-related death (P = 0.026) (Supplementary Table). In the subgroup of patients with a high posttreatment labeling score of 12, distant metastases were evident in 5 of 7 patients (71%) within the follow-up period, whereas there were only 15% (7 of 47 patients) developing metastases in the subgroup with low survivin levels ≤6, indicating that survivin expression may display a useful marker to distinguish patients at high risk for distant cancer relapse.
In patients with a downregulation of survivin during preoperative RCT, a significant decrease of distant cancer relapse (P = 0.022) was observed. Moreover, a significant correlation between high survivin expression levels in residual tumor after RCT and shortened DFS (P = 0.011, Fig. 3A) as well as CSS (P = 0.0017, Fig. 3B) was evident. In multivariate analyses intratumoral survivin expression was confirmed as an independent prognostic factor with significant influence on DFS (P = 0.02) and CSS (P = 0.01, Table 3).
In this study, we investigated the predictive and prognostic relevance of survivin expression both in pretreatment biopsies and corresponding surgical specimens of 116 patients with locally advanced rectal cancer treated with standardized 5-FU based neoadjuvant RCT. Comparing pretreatment biopsies and surgical resection specimens we observed a significant downregulation of survivin in cancer cells following RCT (Fig. 2). This finding is in line with a significant downregulation of survivin expression, as recently described in ovarian and oesophageal cancer treated with neoadjuvant chemotherapy or RCT (19, 20). In the latter study, patients with elevated survivin score after RCT and deficient downregulation were reported to have minor histomorphological tumor regression and decreased survival rates.
In our present investigation, a low survivin expression in pretreatment biopsies was significantly related to lower tumor (ypT) and ypUICC stages after preoperative RCT. As these parameters display strong prognostic factors in recent investigations (2, 24), survivin expression is assumed to be not only a predictive marker for local tumor control but also for disease-free survival after preoperative RCT and surgical resection. The predictive relevance of survivin expression in rectal cancer, however, is controversial as conflicting data exist concerning the relationship between survivin expression and response to RCT. Most investigations, including our own previous studies, reported a significant association between a low survivin expression in pretreatment biopsies and decreased distant and local recurrence rates as well as increased survival (15, 16). Although our present results further strengthen a predictive value of pretreatment survivin levels for patients treated with neoadjuvant RCT, other studies with smaller study cohorts did not report on any relevance of intratumoral survivin expression (17, 18).
The discrepancies between results of different studies may be associated with variable methods of evaluation of survivin expression (mRNA vs. protein expression), the use of antibodies with divergent staining characteristics, or may be due to variable study collectives and treatment regimens, scoring criteria as well as follow-up periods. However, these are well-known methodical problems in quantitative immunhistochemical evaluations. Thus, we tried to minimize these effects by implementation of 2 different staining techniques in different institutions with independent interpretation of the results and a consensual re-evaluation of all cases. Statistical analysis of interobserver variability further strengthened a high concordance of the results exposing a methodically solid basis for the performed analyses.
High survivin levels after neoadjuvant RCT were correlated to advanced ypT and lymph node positive disease (ypN+). Given the generally tight associations between (y)pT, (y)pN, and (y)pUICC stage it is surprising that the posttreatment survivin level correlated with ypT and ypN, but not ypUICC stage (P = 0.11). This discrepancy might be based on a statistical problem as 11 patients (9.5%) with pCR (representing ypUICC 0) were not included into statistical analyses due to a lack of cancer tissue for immunhistochemistry and Spearman rank correlation was performed only with ypUICC I to IV stages.
Individual tumor response to multimodal therapy in rectal cancer is highly heterogeneous and to date unpredictable. Although 8% to 20% of patients have pCR, others patients show intrinsic or acquired resistance for which, at least in part, elevated intratumoral survivin levels might play a substantial role. This hypothesis is strengthened by in vitro results from pancreatic and colorectal tumor cell lines with divergent intrinsic radio-responsiveness: a high constitutive survivin as well as radiation-induced upregulation of survivin was associated with a more radioresistant phenotype (25, 26). Notably, in previous studies, we have shown an inverse relationship between survivin expression and spontaneous apoptosis in rectal cancer biopsies (15) as well as radiation-induced apoptosis in a panel of colorectal cancer cell lines (26). Moreover, an elevated level of apoptosis displays an significant prognosticator for increased histopathologic response and lower rates of local tumor recurrences in rectal cancer patients both treated with or without radiation therapy (23, 27–29) further indicating an impact of apoptotic pathways on clinical outcome in rectal cancer.
As shown in this study, tumor cells may downregulate survivin expression during RCT in rectal cancer. The mechanisms that are responsible for impaired survivin expression or a failure in downregulation of survivin during (radio-) chemotherapy remain elusive. However, efficient downregulation of survivin in tumors likely to respond to DNA-damaging radiation and chemotherapy may result in increased tumor regression and improved survival. Accordingly, this study demonstrates that tumors with a sufficient RCT-induced downregulation of survivin and lower posttreatment expression levels developed significantly fewer cancer relapse and distant metastases, which constitutes an important clinical endpoint in curatively intended treatment of primary rectal cancer. On the contrary, a failure of survivin downregulation was associated with an increased risk to develop distant metastases, that was in line with recent reports on a correlation between a high survivin expression and distant recurrence in T1/T2 prostate (30) and colon cancer (31). This underlines the potential role of survivin in metastatic pathways—beside its antiapoptotic abilities. More recently, a intermolecular cooperation between X-linked inhibitor of apoptosis protein (XIAP) and survivin was reported to stimulate cell invasion and to promote metastases (32). Thus, both IAPs are considered to be metastasis genes, possibly orchestrating a cellular network of transcription factor nuclear factor kappa B (NF-KB) dependent expression of fibronectin, β1-integrin signaling, and activation of the cell motility kinases focal adhesion kinase (FAK) and Src (32). To the best of our knowledge, this is the largest evaluation of survivin expression in rectal cancer undergoing neoadjuvant RCT based on a well documented patient cohort and tumor material available. Thus, we consider the results as representative for a larger group of patients. Nevertheless, a prospective validation of the impact of survivin expression in rectal cancers has to be performed on the basis of a clinical trial including a suitable and cost-efficient immunhistochemical evaluation of the protein in both pretreatment biopsies and resection specimens.
In summary, survivin is relevant not only for individual response of the primary lesion and local control, but also for tumor dissemination and systemic tumor control after multimodal treatment of rectal cancers. By comparison of pre and posttherapeutic intratumoral survivin, our results implicate that therapeutic strategies to downregulate survivin expression during preoperative therapy or inhibit survivin mediated pathways may further increase individual local and distant tumor response and, as a consequence, patient's prognosis (12, 13). Concretely, patients with low pretreatment survivin levels (individual labeling score ≤6) show a significantly prolonged DFS compared with patients with high initial survivin expression (<6). Based on these results, patients with high baseline survivin levels might benefit from additional anti-survivin strategies to decrease the number of patients with impaired downregulation of intratumoral survivin expression, being correlated with higher incidences of distant metastatic disease and cancer-related death as well as decreased survival. Because of a variety of survivin antagonists including antisense-oligonucleotides (Eli-Lilly compound LY2181308) and transcriptional repressors (Ym155) recently entered phase I/II clinical trials (33, 34), our data display a further rationale to combine survivin inhibitors with radio(chemo)therapy in future clinical trials for rectal cancer.
Disclosure of Potential Conflicts of Interest
No potential conflicts of interest were disclosed.
This study was supported by the Deutsche Forschungsgemeinschaft (KFO 179 and DFG RO 3482/1-1) and the German Federal Ministry of Education and Research (01ES0808 and 01EX1021).
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.
The authors thank Birgit Jünemann for excellent technical support and high-quality performance of laboratory procedures as well as Johanna Kreutzer, MA, and Elke Stauffer (study nurse) for organizational support. In addition, the authors thank David M. Goldenberg, ScD, MD (Belleville, NJ) for his critical and helpful comments on the manuscript.
Note: Supplementary data for this article are available at Clinical Cancer Research Online (http://clincancerres.aacrjournals.org/).
- Received September 27, 2010.
- Revision received November 10, 2010.
- Accepted November 18, 2010.
- ©2010 American Association for Cancer Research.