EJSO 2003; 29: 379±382 doi:10.1053/ejso.2002.1415
Survivin expression in human osteosarcoma is a marker for survival K. Trieb*, R. Lehner², T. Stulnig³, I. Sulzbacher§ and K. R. Shroyer² Departments of *Orthopedics, ³Internal Medicine III and §Pathology, University of Vienna, Vienna and ²Department of Pathology, University of Colorado Health Sciences Center, Denver
Aims: Osteosarcoma is the most frequent malignant bone tumor with a peak incidence in the second and third decade of life. Evaluation of prognosis of patients with osteosarcoma is limited to clinical parameters whereas molecular markers of tumor aggression have not yet been identified. Inhibition of apoptotic cell death could play a role in the development or progression of neoplasia. Survivin is a member of the inhibitor of apoptosis (IAP) protein gene family and is expressed both during normal fetal development and in human cancer. Methods: The localization and distribution of survivin was investigated immunohistochemically in high-grade osteosarcomas by an indirect immunoperoxidase method. Results: Survivin was detected in the cytoplasm in 23/40 and in the nucleus in 20/40 cases of osteosarcoma. Nuclear localization of survivin expression was significantly correlated with a prolonged survival (P 0.0347) but cytoplasmic staining showed no correlation with patient outcome. Conclusions: The results of this study indicates that the evaluation of survivin expression might be a useful prognostic marker in osteosarcoma. Patients with osteosarcoma exhibiting nuclear survivin expression could potentially benefit from stratification of neoadjuvant chemotherapy. # 2003 Elsevier Science Ltd. All rights reserved. Key words: survivin; osteosarcoma; expression; survival.
INTRODUCTION Osteosarcoma is the most common malignant tumor of bone and shows a peak incidence in the the second and third decades of life. Although survival rates increased from 20 to 75% due to the combination of radical surgery and neoadjuvant chemotherapy in recent years, a preoperative stratification of chemotherapy in individual cases is not possible.1 Postoperatively, clinical stratification is determined by response to chemotherapy and the size of the tumor, which in the past, have been the only useful prognostic parameters. Survivin is a member of the inhibitor of apoptosis protein (IAP) gene family and is expressed both during normal fetal development and in a broad spectrum of human cancers.2±4 The regulation of apoptosis may play a central role in the development of neoplastic cellular proliferative processes and may exert a strong influence on tumor progression. During normal cell division, survivin interacts with microtubules of the mitotic
Correspondence to: Prof. Dr Klemens Trieb, Department of Orthopaedics, University of Vienna, WaÈhringer GuÈrtel 18-20, A-1090 Vienna, Austria. Tel: 43-1-40400-4070; Fax: 43-1-40400-4077; E-mail:
[email protected] 0748±7983/03/$30.00
spindle in a saturated reaction that is regulated by microtubular dynamics.5 Destruction of survivinmicrotubule interactions results in the loss of the antiapoptotic function and in the activation of caspase-3, an effector protease in the apoptotic cascade. Survivin expression in vitro has been shown to inhibit both caspases 3 and 7.3,5 It is the aim of this study to investigate survivin expression in osteosarcoma and to evaluate its value as prognostic marker.
MATERIAL AND METHOD Formalin-fixed, paraffin-embedded tissue specimens were obtained from 40 patients with high-grade osteosarcoma (grade III: 36, grade II: four, gender distribution 14 females, 26 males; age 23 + 18 years (mean + SD), range 7±68 years) from the archival collections of the Department of Pathology at the University of Vienna. All tissues were biopsy specimens that were collected prior to definitive surgical management or chemotherapeutic intervention. Following initial biopsy, all patients underwent multiagent, neoadjuvant chemotherapy according to the Cooperative Osteosarcoma study protocol (COSS 861) and definitive surgery. #
2003 Elsevier Science Ltd. All rights reserved.
380 The osteosarcomas were located in the femur (26 cases), the tibia (seven cases), the ileum, or sacrum (five cases) and the humerus (two cases). Response to neoadjuvant chemotherapy was analysed according to tumor cell necrosis (good response less than 10% viable tumor cells). Twelve patients had metastases at the time of diagnosis. The mean tumor size was 144 + 94 ml, the mean serum alkaline-phosphatase level was 250 + 131 U/ml and the mean follow-up was 64 + 40 months (median 62). Nineteen patients died of their disease 29 + 17 months (range 1±67 months) after diagnosis. The remaining 21 patients were followed clinically for a mean of 96 + 25 months (range 36±125). Survivin expression was detected by immunohistochemistry using a polyclonal rabbit serum antibody that was generated by immunization with the survivin peptide sequence PTLPPAWQPFLKDHRI (Genebank) as described previously, according to the originally described protocol.2 Western blot analysis of the survivin antibody revealed specificity for a single protein of approximately 16.5 kd, consistent with the size of the full-length survivin protein. Five micron tissue sections were collected on glass slides, deparaffinized in xylene, and rehydrated in graded alcohol baths. After baking at 60 C, rehydrating in distilled water, and quenching of endogenous peroxidase in 3% H2O2 for 15 min, antigen retrieval was performed by incubation in citrate buffer for 7 min in a decloaking chamber (Biocare Medicalâ ). Specimens were incubated with a purified immunoglobulin fraction of survivin antiserum at a dilution of 1:1350 (1.84 mg/ml) at room temperature for 20 h. Control slides were incubated with an equivalent dilution of the corresponding preimmune immunoglobulin fraction. Subsequently, the slides were incubated with a secondary peroxidase-labeled goat anti-rabbit antibody at a dilution of 1:500 (Biosource Internationalâ ) for 30 min at 37 C. Finally, staining was visualized using 2 0 4 0 diaminobenzidine/H2O2, followed by dehydration and counterstaining with hematoxylin. Staining of a colon adenoma showing a strong cytoplasmic and nuclear survivin reaction served as positive control. The slides were scored for nuclear and cytoplasmic staining blinded to clinical data independently by two of us. A minimum of 10 microscopic fields at 200 magnification were examined in each case to estimate the total percentage of stained cells. Staining of the nucleus or cytoplasm of more than 10% of tumor cells was recorded as positive for survivin and staining of 10% or less tumor cells was scored as negative for survivin.
Statistics All the data are given as mean SD unless stated otherwise and were analysed by SPSSâ software. Correlation was estimated by Spearman's rank
K. TRIEB ET AL. correlation coefficient (rho). The association of survivin expression was assessed by Kaplan±Meier analysis6 and differences were tested for significance by the logrank test. Cox's regression was used to analyse survivin expression in a multicentric model including the presence of metastases and response to neoadjuvant chemotherapy. A two-tail P value of ,0.05 was considered statistically significant.
RESULTS Survivin was detected in the cytoplasm of 23/40 and in the nucleus of 20/40 cases of osteosarcoma. Both nuclear and cytoplasmic staining of tumor cells was seen in 18 of 40 cases (45%, Fig. 1A). Two cases (5%) showed only nuclear staining, five cases (13%) showed only cytoplasmic staining and 15 cases (37%) were completely negative for survivin. Histologic control sections incubated with preimmune immunoglobulin did not show staining (Fig. 1B). Survivin localization in the nucleus
Figure 1 (A) Representative immunohistochemical data showing nuclear staining of survivin in a high-grade osteosarcoma. Note both nuclear and cytoplasmic staining. Original magnification 400. (B) Control histologic sections that were incubated with preimmune immunoglobulin did not show staining.
SURVIVIN EXPRESSION IN OSTEOSARCOMA
Figure 2 Kaplan±Meier survival analysis of osteosarcoma patients according to the expression of survivin in the nucleus (P 0.035) of tumor cells. The solid line represents positive staining, the dashed line negative staining.
(n 20) was correlated with tumor size (rho ÿ0.43, P 0.015; cytoplasmic: rho ÿ0.01, P 0.97), but neither nuclear nor cytoplasmic survivin expression was correlated with the presence of metastases at the time of diagnosis (nuclear: rho 0.11, P 0.52; cytoplasmic: rho 0.10, P 0.54) nor with response to preoperative chemotherapy (nuclear: rho ÿ0.03, P 0.88; cytoplasmic: rho ÿ0.08, P 0.64). Nuclear localization of survivin (n 20) exhibited a significant correlation with prolonged survival (P 0.035, Fig. 2). Among the 20 cases with nuclear localization of survivin, the cumulative survival rate at 5 years after diagnosis was 70 + 10% (mean + SEM) which was considerably higher than survival of patients with tumors lacking nuclear survivin expression (40 + 11%). No correlation was found between patient survival and cytoplasmic staining for survivin (P 0.411, 5-year survival rate: 61 + 10% for positive tumors vs 47 + 12%). Survivin expression in the nucleus was the only independent predictor of survival (P 0.045) in a multivariate Cox regression analysis including tumor size (P 0.20) and presence of metastases (P 0.09; whole model: P 0.005).
DISCUSSION This study based on the assumption that there could be a direct correlation between patient survival and survivin expression in human osteosarcoma. So far, the only generally accepted prognostic factors for osteosarcoma are tumor size and response to chemotherapy. Up to now, molecular markers are not yet established in daily routine. The introduction of a new prognostic marker could enable development of new strategies in the treatment of osteosarcoma, e.g., a preoperative stratification
381 of neoadjuvant chemotherapy, in the treatment of osteosarcoma. The expression of markers indicating a poor prognosis already at biopsy could facilitate a preoperative stratification of neoadjuvant chemotherapy. To date, this stratification is done postoperatively because of the lack of reliable markers. Therefore, we evaluated the role of survivin as a potential new prognostic marker in high grade osteosarcomas. Survivin is thought to exert its anti-apoptotic role through interaction with the terminal effector proteases of the apoptotic cascade, caspases 3 and 7. Thus, the nuclear sequestration of survivin could inhibit the interaction with caspases in the cytoplasm and thereby, prevent the anti-apoptotic effect of survivin expression. It is unknown if the survivin molecule includes nuclear translocation or export signal sequences. It is possible that pathways that regulating the transport of survivin across the nuclear membrane could be disrupted during the processes of malignant transformation or tumor progression. Thus, changes in the subcellular localization of survivin could represent epiphenomena rather than reflecting a functional role for survivin expression in osteosarcomas. Alternatively, the subcellular localization of survivin could be involved mechanistically involved in the regulation of tumor growth. Survivin has been shown to exist in distinct subcellular pools with 80% in the cytosol and 20% in the nucleus. Intracellular antibody targeting did not interfere cytokinesis, but resulted in microtubular defects.7 The current study showes a dramatic and statistically significant difference in patient survival depending on whether or not nuclear staining for survivin was observed. Whereas 7/10 patients with absent nuclear staining had died of their disease within 36 months postoperatively, 12/15 patients with nuclear positivity for survivin were alive at an average follow-up of 97 months. Thus, osteosarcomas that lack localization of survivin appear to be more aggressive from a biological as well as from a clinical point of view. These results might be surprising, because one could expect survivin to be an adverse prognostic factor. But members of the IAP family have been recognized to have several, sometimes controversial functions. For instance, nuclear survivin expression in gastric cancer is associated with a more favorable prognosis8 and with a greater period of disease free survival in transitional cell carcinoma of the bladder.9 These results are in concordance with our findings. On the other hand survivin expression is correlated with an adverse outcome in hepatocellular and in oesophageal carcinoma.10,11 The results of the current study indicate that the immunohistochemically verified localization of survivin could be an important prognostic marker for the prediction of the survival of patients with high grade osteosarcoma. Thus, survivin analysis could facilitate preoperative stratification of neoadjuvant chemotherapy towards the utilization of a more aggressive or
382 individualized therapy regimen for patients with the most aggressive forms of osteosarcoma.
ACKNOWLEDGEMENTS The authors thank Mrs H. Hitschmann for correcting the manuscript. The MAX KADE Foundation N.Y provided fellowship support for RL and this work was supported by the `Medizinisch wissenschaftlicher FoÈrderungsfonds des BuÈrgermeisters der Stadt Wien' (N. 1847) to KT.
REFERENCES 1. Bieling B, Rehan N, Winkler P et al. Tumor size and prognosis in aggressively treated osteosarcoma. J Clin Oncol 1996; 14: 848±58. 2. Ambrosini G, Adida C, Altieri DC. A novel anti-apoptosis gene, survivin, expressed in cancer and lymphoma. Nat Med 1997; 3: 917±21. 3. Lu CD, Altieri DC, Tanigawa N. Expression of a novel antiapoptosis gene, survivin, correlated with tumor cell apoptosis and p53 accumulation in gastric carcinomas. Cancer Res 1998; 58: 1808±12.
K. TRIEB ET AL. 4. Kappler M, Kohler T, Kampf C. Increased survivin transcript levels: an independent negative predictor of survival in soft tissue sarcoma patients. Int J Cancer 2001; 95: 360±3. 5. Li F, Ambrosini G, Chu EY et al. Control of apoptosis and mitotic spindle checkpoint by survivin. Nature 1998; 396: 580±4. 6. Kaplan EL, Meier P. Nonparametic estimation from incomplete observations. J Am Stat Assoc 1958; 53: 457±61. 7. Fortugno P, Wall N, Giodini A et al. Survivin exists in immunohistochemically distinct subcellular pools and is involved in spindle microtuble function. J Cell Sci 2002; 115: 575±85. 8. Okada E. Murai Y, Matsui K et al. Survivin expression in tumor cells nuclei is predictive of a favorable prognosis in gastric cancer patients. Cancer Res 2001; 163: 109±16. 9. Lehner R. Lucia MS, Jarboe EA et al. Immunohistochemical localization of the IAP protein survivin in bladder mucosa and transitional cell carcoma. Appl Immun Mol Morphol 2002; 10: 134±8. 10. Ikeguchi M, Ueta T, Yamane Y et al. Inducible nitric oxide synthetase and survivin messenger RNA expression in hepatocellular carcinoma. Clin Cancer Res 2002; 8: 3131±6. 11. Ikeguchi M, Kaibara N. Survivin messenger RNA expression is a good prognostic biomarker for oesophageal carcinoma. Br J Cancer 2002; 87: 883±7.
Accepted for publication 14 November 2002
