Survivin Expression in Intracranial Ependymomas and Its Correlation ...

Anatomic Pathology / SURVIVIN EXPRESSION IN EPENDYMOMA

Survivin Expression in Intracranial Ependymomas and Its Correlation With Tumor Cell Proliferation and Patient Outcome Matthias Preusser, MD,1 Stefan Wolfsberger, MD,2 Thomas Czech, MD,2 Irene Slavc, MD,3 Herbert Budka, MD,1 and Johannes A. Hainfellner, MD1 Key Words: Apoptosis; Ependymoma; Proliferation; Prognosis; Survivin DOI: 10.1309/PP2G5GAAFKV82DTG

Abstract Survivin expression has been described as a prognostic factor in various tumor types and has been shown to correlate with cytologic anaplasia in ependymoma. We immunohistochemically studied survivin expression and its association with Ki-67 and topoisomerase IIα (TIIα) expression and outcome in 63 patients with intracranial ependymoma. Survivin is expressed in a fraction of nuclei of tumor and endothelial cells including mitotic figures. Survivin indices range from 0.6% to 43.2% and correlate with Ki-67 and TIIα indices. On average, 62.86% of Ki-67– expressing tumor cell nuclei coexpress survivin, whereas 92.2% of survivin-expressing nuclei coexpress Ki-67. High survivin, Ki-67, and TIIα indices univariately correlated with an unfavorable outcome. In multivariate analysis, only the Ki-67 index remained an independent prognostic factor. In ependymoma, survivin is expressed in a subset of proliferating cells. High survivin expression is associated with poor patient outcome. However, the Ki-67 index is a more accurate prognostic marker than the survivin or TIIα index.

Survivin is a recently identified protein that acts as an inhibitor of apoptosis by preventing caspase-9 activation within a functional apoptosome. In addition, survivin regulates mitosis1,2 by stabilizing microtubules in metaphase spindle formation and regulation of the spindle assembly checkpoint.3 High expression of survivin has been associated with an unfavorable prognosis in a large number of malignant neoplasms4-13 and also has been described as a marker of favorable prognosis in some tumor types.14-16 In a recent study, Preusser et al17 showed that survivin is expressed in proliferating tumor cells in glioblastoma. In ependymoma, loss of nuclear survivin expression was reported to associate with cytologic anaplasia.18 In a recent study, Wolfsberger et al19 showed that the Ki67 immunolabeling index is an accurate predictor of outcome in patients with intracranial ependymoma. Considering the role of survivin as regulator of mitosis, its expression might be associated with tumor cell proliferation and patient outcome in ependymoma. To test the prognostic accuracy of survivin in relation to other tumor cell proliferation markers, we revisited the ependymoma series of Wolfsberger et al19 and systematically analyzed survivin expression and its correlation with tumor cell proliferation markers Ki-67 and topoisomerase IIα (TIIα). Furthermore, we analyzed the impact of survivin expression on patient outcome.

Materials and Methods Patients In a previous study, Wolfsberger et al19 analyzed a cohort of 103 patients who underwent the first operation for intracranial Am J Clin Pathol 2005;124:543-549

© American Society for Clinical Pathology 543

DOI: 10.1309/PP2G5GAAFKV82DTG

543 543

Preusser et al / SURVIVIN EXPRESSION IN EPENDYMOMA

ependymoma between December 1957 and October 2000. Of the 103 cases, 40 were excluded from the present study because material was no longer available (5 cases) or owing to methodological difficulties (35 cases; for more detailed information, see “Statistical Analysis” in the “Results” section). Patients with survival times less than 30 days (n = 5) were excluded from survival analyses to exclude deaths due to postoperative complications. The patients underwent the first operation between 1957 and 2000 at the University Hospital of Vienna, Vienna, Austria. For statistical comparisons (see “Statistical Analysis”), we subdivided our cohort into 3 time-related groups: (1) operation in 1979 or earlier; (2) operation from 1980 through 1990; and (3) operation in 1991 or later. All tumor specimens were reevaluated histologically by a neuropathologist (J.A.H.) and classified as ependymoma (grade II, 44 cases; grade III, 19 cases) according to the current World Health Organization definition.20,21 In the patient subgroup younger than 2 years old, 8 tumors were infratentorial and 1 was supratentorial. In patients between 2 and 18 years, 16 tumors were infratentorial and 11 were supratentorial. In patients 18 years or older, 8 tumors were infratentorial and 19 were supratentorial. Of the 63 patients, 37 (59%) underwent gross total resection and 25 patients (40%), subtotal tumor resection. In 1 case (2%), no information on the extent of resection was available. For 16 patients (25%), adjuvant chemotherapy and radiotherapy were given; 25 patients (40%) received only radiotherapy and 1 patient (2%) received only chemotherapy. No adjuvant therapy was given to 20 patients (32%), and for 1 patient (2%), insufficient information about adjuvant therapy was available. Immunohistochemical and Immunofluorescent Staining Specimens were fixed routinely in 4% formalin, embedded in paraffin, and cut at a thickness of 3 µm. Sections were deparaffinized in xylene. For antisurvivin immunostaining, sections were incubated for 1 hour with S2367 TargetRetrieval Solution, pH 9.0 (DAKO, Glostrup, Denmark) in a steamer. For anti–Ki-67 and anti-TIIα immunostaining, slides were heated in 0.01 mol/L of citrate buffer (pH 6.0) for 30 minutes in a microwave oven at 600 W. We used the following antibodies for immunostaining: antisurvivin FL-142 (polyclonal rabbit, sc-10811, dilution 1:300, overnight incubation, 4°C; Santa Cruz Biotechnology, Santa Cruz, CA), anti–Ki-67 (monoclonal mouse, clone MIB-1, dilution 1:50, 25 minutes, room temperature; DAKO), and anti-TIIα (monoclonal mouse, clone JH2.7, dilution 1:20, 60 minutes, room temperature; Neomarkers, Fremont, CA). Detection of immunostaining was performed using the EnVision kit (DAKO), and diaminobenzidine was used as the chromogen. A tissue sample from a brain metastasis of malignant melanoma was used as a positive control sample for antisurvivin immunostaining. 544 544

Am J Clin Pathol 2005;124:543-549 DOI: 10.1309/PP2G5GAAFKV82DTG

Negative control samples included omission of the primary antisurvivin antibody and nonneoplastic central nervous system tissue samples. For Ki-67–survivin double immunolabeling, we used the EnVision kit and diaminobenzidine as the chromogen for survivin. For Ki-67, we used biotinylated antimouse antibody (DAKO) and streptavidin alkaline phosphatase (DAKO) with fast blue (0.5%, 15-30 minutes) as the chromogen. For Ki-67–survivin–nucleic acid and TIIα–survivin– nucleic acid triple immunofluorescent staining, the fluorescent-labeled secondary antibody for anti–Ki-67 and antiTIIα was Alexa Fluor 488 goat antimouse IgG (dilution 1:200; Molecular Probes, Eugene, OR). For survivin, we used Alexa Fluor 633 goat antimouse IgG (dilution 1:200; Molecular Probes). Nuclei were labeled using the Sytox Orange Nucleic Acid Stain (S-11368, Molecular Probes) according to the manufacturer’s instructions. For Ki67–TIIα–survivin triple immunofluorescent staining, the secondary antibody for TIIα was Alexa Fluor 546 goat antimouse IgG (dilution 1:200; Molecular Probes). Immunofluorescent labeling was evaluated using a Zeiss LSM 510 confocal laser microscope (Carl Zeiss, Vienna, Austria). We used argon 488-nm and helium/neon 546-nm and 633-nm lasers to elicit fluorescence. Evaluation For the assessment of survivin, Ki-67, and TIIα immunolabeling, 500 tumor cell nuclei were evaluated in each specimen in fields showing the highest density of immunopositive nuclei. The fraction of labeled tumor cell nuclei was expressed as a percentage (survivin, Ki-67, or TIIα index). In 5 cases, double staining for survivin and Ki-67 was performed for evaluation of the fraction of survivin-coexpressing nuclei per 500 Ki-67–expressing nuclei and of the fraction of Ki-67–coexpressing nuclei per 500 survivinexpressing nuclei. Statistical Analysis The Spearman coefficient of correlation, χ2 test, Wilcoxon signed rank test, and Fisher exact test were used as appropriate. Survival probabilities were computed as outlined by Kaplan and Meier.22 The log-rank test and Cox proportional hazards model were used for univariate and multivariate analyses of overall survival. Overall survival was defined from the day of initial surgery until death of the patient. Death due to a cause other than ependymoma or survival until the end of the observation period were considered as censoring events in statistical survival analysis. The median was defined as the cutoff value between low and high survivin, Ki-67, and TIIα indices. The survivin, Ki-67, and TIIα indices were entered into Cox regression. For all tests, a 2-tailed P value of .05 or less was considered significant. © American Society for Clinical Pathology

Anatomic Pathology / ORIGINAL ARTICLE

Results Immunohistochemical and Immunofluorescent Staining Antisurvivin immunostaining was performed on a total of 98 cases in which anti–Ki-67 and anti-TIIα yielded satisfactory staining results. In 35 of 98 cases, survivin could not be visualized properly (no immunoreactive structures including unstained mitotic figures), apparently owing to material or methodological reasons. (These were mainly older blocks; for more details, see the following “Statistical Analysis” section). These cases were excluded from further analyses. In 63 of 98 cases used for further analysis, survivin expression was prominent in a fraction of nuclei of tumor cells ❚Image 1A❚ and endothelial cells ❚Image 1B❚ throughout the specimen. Antisurvivin labeled all mitotic figures ❚Image 1C❚, ❚Image 1D❚, ❚Image 1E❚, and ❚Image 1F❚. Faint cytoplasmic expression was seen in a few tumor cells. There was no antisurvivin staining of surrounding nonneoplastic central nervous system tissue. Quantitative results for survivin, Ki-67, and TIIα expression are summarized in ❚Table 1❚. Triple immunofluorescent staining and confocal microscopy showed that Ki-67, TIIα, and survivin were coexpressed in a fraction of tumor cell nuclei (Image 1C-1F) ❚Image 1G❚, ❚Image 1H❚, ❚Image 1I❚, and ❚Image 1J❚. Double immunolabeling analysis of 5 cases showed that a mean of 62.9% (range, 43.8%-76.6%) of Ki-67–expressing tumor cell nuclei coexpressed survivin. In contrast, a mean of 92.2% (range, 86.6%-96.8%) of survivin-expressing nuclei coexpressed Ki-67. Statistical comparison of Ki-67 and survivin index levels including all cases showed that the Ki-67 index was consistently higher than the survivin index (P < .0001; Wilcoxon signed rank test). Statistical Analysis The median patient age at surgery was 11.3 years (range, 0.1-74.4 years). The mean survival time was 7.7 years for patients with grade II tumors and 3.2 years for patients with grade III tumors. Of 63 patients, 31 (49%) died of disease, and 32 patients (51%) were lost to follow-up or died of causes other than ependymoma. There was a significantly higher number of cases that had to be excluded for material or methodological reasons (see also earlier text) during the period before 1979 than in the later periods (P = .001; χ2 test). So survivin immunodetectability seems to be affected by tissue processing factors in earlier times or duration of tissue sample storage. High survivin (P = .003; Fisher exact test), high Ki67 (P = .018; Fisher exact test), and high TIIα (P = .028; Fisher exact test) indices are associated significantly with tumor grade III. The survivin index showed significant correlation with the Ki-67 and TIIα indices ❚Figure 1❚. Statistical comparison of the Ki-67 and survivin index levels showed that

the Ki-67 index was consistently higher than the survivin index (P < .0001; Wilcoxon signed rank test). The survivin, Ki-67, and TIIα indices correlated significantly with overall survival in univariate analysis (Table 1, Figure 1). In multivariate analysis, only the Ki-67 index remained an independent prognostic factor (Table 1). There was no statistically significant difference in survival times between patients with supratentorial vs infratentorial tumors (P = .8357; log-rank test), patients in different age groups (≤2 years, 2-18 years, >18 years; P = .3135; log-rank test), or patients treated during different periods (P = .3534; logrank test).

Discussion Immunohistochemical analysis of survivin expression in ependymoma was performed in 2 previous studies on small series of cases.18,23 In 1 study, survivin expression was described in the cytoplasm but not in the nuclei of tumor cells.23 In contrast, we observed that immunostaining signals are particularly prominent in tumor cell nuclei and barely visible in the cytoplasm. In vitro studies have shown the presence of survivin in the cytoplasm, cell nuclei, and mitotic apparatus.2 Different antisurvivin antibodies bind more strongly to cytoplasmic survivin and others to nuclear survivin. We used antibody sc-10811 (Santa Cruz Biotechnology), which has been proven to specifically detect survivin in brain tissue by immunohistochemical analysis and Western blotting.18 It has been described to bind predominantly to nuclear survivin.18 So one explanation for the discrepancies of our results with previous findings in ependymoma might be the use of different antibodies. Another study on pediatric ependymomas and choroid plexus tumors reported that nuclear expression of survivin correlated with morphologic tumor grade.18 Cases with high tumor grade had a lower fraction of survivin-expressing tumor cells than did low-grade tumors. In contrast, we found in our larger series a significant correlation of a high survivin index with high tumor grade. In our hands, confocal laser scanning microscopy showed colocalization of Ki-67, TIIα, and survivin in tumor cell nuclei including mitotic figures. These findings confirm expression of survivin in proliferating cells. Tumor cell proliferation commonly is assessed by anti–Ki-67 immunostaining.24-30 To assess survivin as a potential alternative tumor cell proliferation marker, we compared the relation of Ki-67 and survivin expression in tumor cell nuclei. We found by double immunolabeling studies and statistical analysis that anti–Ki67 stains a significantly larger fraction of tumor cells than does antisurvivin. These findings are in line with previous in vitro studies showing that Ki-67 (expression in G1, S, G2, and Am J Clin Pathol 2005;124:543-549



545 545


A

B

C

D

E

F

G

H

I

J

❚Image 1❚ Antisurvivin labels nuclei of tumor cells (A, antisurvivin, original magnification ×600) and endothelial cells (B, antisurvivin, original magnification ×200). C, D, E, and F, Topoisomerase IIα and survivin are coexpressed in a mitotic figure (nucleic acids, blue signal; topoisomerase IIα, red signal; survivin, green signal; original magnification ×1,000). F is a merged image of C-E. G, H, I, and J, Ki-67, topoisomerase IIα, and survivin are coexpressed in a tumor cell nucleus (Ki-67, blue signal; topoisomerase IIα, red signal; survivin, green signal; original magnification ×1,000). J is a merged image of G-I. 546 546


© American Society for Clinical Pathology


M phases)31,32 is expressed in a larger segment of the cell cycle than survivin (G2 and M phases)2 or TIIα (S, G2, and M phases).33 We conclude that anti–Ki-67 identifies proliferating tumor cells more sensitively than does antisurvivin.

Univariate survival analysis showed significant correlation of Ki-67, TIIα, and survivin indices with patient outcome. With multivariate analysis including all 3 factors, only the Ki67 index remained an independent prognostic parameter. So

❚Table 1❚ α Expression and of Univariate and Multivariate Survival Analyses Quantitative Results of Survivin, Ki-67, and Topoisomerase IIα Range (%)

P

Index

Overall

Grade II

Grade III

Ki-67

2.8-78.6

2.8-56.2

2.8-78.6

Topoisomerase IIα

1.0-59.2

1.0-56.4

1.6-56.4

Survivin *

0.6-43.2

0.6-35.8

Median Cutoff (%)

No. of Cases

Log-Rank Test

Cox Regression

20 38 24 34 30 28

.0001*

.019*

.0016*

.570

.0032*

.886

≤20.5 >20.5 ≤9.4 >9.4 ≤6.4 >6.4

2.8-43.2

Statistically significant.

B

A

80

0.8 60 Low survivin index (n = 30)

0.6

Ki-67 Index

Cumulative Overall Survival

1.0

0.4 High survivin index (n = 28)

40

20

0.2

0.0

0

10

20 Survival Time (y)

30

40

0

0

10

20 30 Survivin Index

40

50

❚Figure 1❚ A, Kaplan-Meier curve showing significantly better outcome for patients with a low survivin index (P = .0032). B and C, Scatter plots showing significant correlation of the survivin index with the Ki-67 index (B, r = 0.586; P < .0001) and the topoisomerase IIα (TIIα) index (C, r = 0.733; P < .0001).

C 60

50

TIIα Index

40

30

20

10

0 0

10

20 30 Survivin Index

40

50

Am J Clin Pathol 2005;124:543-549



547 547


the Ki-67 index more accurately predicts patient outcome. This finding might be explained by more sensitive identification of proliferating cells by anti–Ki-67. Thus, anti–Ki-67 should be the first choice as a tumor cell proliferation marker in the clinical setting. The prognostic impact of survivin expression has been analyzed in various tumor types. In most tumor types, high survivin expression correlated with an unfavorable patient outcome,4-13 whereas in a few tumor types (eg, gastric cancer, breast cancer) it correlated with a favorable outcome.14-16 To our knowledge, only 1 previous study on mantle cell lymphoma compared the prognostic influence of survivin and Ki67.9 In that study, survivin and Ki-67 correlated with patient outcome in univariate analysis. With multivariate analysis, only the Ki-67 index remained an independent predictor of survival. By comparing the impact of survivin and Ki-67 in ependymoma, we obtained analogous results. Thus, the Ki-67 index seems to predict patient outcome more accurately than does the survivin index in different cancer types. Survivin is expressed in ependymoma predominantly in a subset of proliferating tumor cell nuclei. High survivin expression is associated with an unfavorable patient outcome in intracranial ependymoma. However, the Ki-67 index predicts the outcome for patients with ependymoma more accurately than does the survivin index. From the 1Institute of Neurology, 2Department of Neurosurgery, and 3Department of Pediatrics, Medical University of Vienna, Vienna, Austria. The study was performed within the Austrian NeuroOncology Network (http://www.ann.at), Vienna. Address reprint requests to Dr Hainfellner: Institute of Neurology, Medical University of Vienna, AKH 4J, Waehringer Guertel 18-20, POB 48, A-1097 Vienna, Austria. Acknowledgments: We thank Judith Ludwig, Gerda Ricken, and Helga Flicker, Institute of Neurology, Medical University of Vienna, and Evelin Simek, Medical Media Services, Medical University of Vienna for excellent technical assistance.

References 1. Ambrosini G, Adida C, Altieri DC. A novel anti-apoptosis gene, survivin, expressed in cancer and lymphoma. Nat Med. 1997;3:917-921. 2. Li F, Ambrosini G, Chu EY, et al. Control of apoptosis and mitotic spindle checkpoint by survivin. Nature. 1998;396:580584. 3. Altieri DC. Molecular circuits of apoptosis regulation and cell division control: the survivin paradigm. J Cell Biochem. 2004;92:656-663. 4. Adida C, Haioun C, Gaulard P, et al. Prognostic significance of survivin expression in diffuse large B-cell lymphomas. Blood. 2000;96:1921-1925. 5. Grabowski P, Kuhnel T, Muhr-Wilkenshoff F, et al. Prognostic value of nuclear survivin expression in oesophageal squamous cell carcinoma. Br J Cancer. 2003;88:115-119.

548 548


6. Kajiwara Y, Yamasaki F, Hama S, et al. Expression of survivin in astrocytic tumors: correlation with malignant grade and prognosis. Cancer. 2003;97:1077-1083. 7. Kami K, Doi R, Koizumi M, et al. Survivin expression is a prognostic marker in pancreatic cancer patients. Surgery. 2004;136:443-448. 8. Kappler M, Kotzsch M, Bartel F, et al. Elevated expression level of survivin protein in soft-tissue sarcomas is a strong independent predictor of survival. Clin Cancer Res. 2003;9:1098-1104. 9. Martinez A, Bellosillo B, Bosch F, et al. Nuclear survivin expression in mantle cell lymphoma is associated with cell proliferation and survival. Am J Pathol. 2004;164:501-510. 10. Sarela AI, Macadam RC, Farmery SM, et al. Expression of the antiapoptosis gene, survivin, predicts death from recurrent colorectal carcinoma. Gut. 2000;46:645-650. 11. Sarela AI, Scott N, Ramsdale J, et al. Immunohistochemical detection of the anti-apoptosis protein, survivin, predicts survival after curative resection of stage II colorectal carcinomas. Ann Surg Oncol. 2001;8:305-310. 12. Schlette EJ, Medeiros LJ, Goy A, et al. Survivin expression predicts poorer prognosis in anaplastic large-cell lymphoma. J Clin Oncol. 2004;22:1682-1688. 13. Span PN, Sweep FC, Wiegerinck ET, et al. Survivin is an independent prognostic marker for risk stratification of breast cancer patients. Clin Chem. 2004;50:1986-1993. 14. Kennedy SM, O’Driscoll L, Purcell R, et al. Prognostic importance of survivin in breast cancer. Br J Cancer. 2003;88:1077-1083. 15. Okada E, Murai Y, Matsui K, et al. Survivin expression in tumor cell nuclei is predictive of a favorable prognosis in gastric cancer patients. Cancer Lett. 2001;163:109-116. 16. Trieb K, Lehner R, Stulnig T, et al. Survivin expression in human osteosarcoma is a marker for survival. Eur J Surg Oncol. 2003;29:379-382. 17. Preusser M, Gelpi E, Matej R, et al. No prognostic impact of survivin expression in glioblastoma. Acta Neuropathol (Berl). 2005;109:534-538. 18. Altura RA, Olshefski RS, Jiang Y, et al. Nuclear expression of survivin in paediatric ependymomas and choroid plexus tumours correlates with morphologic tumour grade. Br J Cancer. 2003;89:1743-1749. 19. Wolfsberger S, Fischer I, Hoftberger R, et al. Ki-67 immunolabeling index is an accurate predictor of outcome in patients with intracranial ependymoma. Am J Surg Pathol. 2004;28:914-920. 20. Kleihues P, Cavenee W, eds. Pathology and Genetics of Tumours of the Nervous System. Lyon, France: IARC Press; 2000. World Health Organization Classification of Tumours. 21. Kleihues P, Louis DN, Scheithauer BW, et al. The WHO classification of tumors of the nervous system. J Neuropathol Exp Neurol. 2002;61:215-229. 22. Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc. 1985;53:457-481. 23. Sasaki T, Lopes MB, Hankins GR, et al. Expression of survivin, an inhibitor of apoptosis protein, in tumors of the nervous system. Acta Neuropathol (Berl). 2002;104:105-109. 24. Bredel M, Piribauer M, Marosi C, et al. High expression of DNA topoisomerase IIalpha and Ki-67 antigen is associated with prolonged survival in glioblastoma patients. Eur J Cancer. 2002;38:1343-1347.

© American Society for Clinical Pathology


25. Chiesa-Vottero AG, Rybicki LA, Prayson RA. Comparison of proliferation indices in glioblastoma multiforme by whole tissue section vs tissue microarray. Am J Clin Pathol. 2003;120:902-908. 26. Dictor M, Ehinger M, Mertens F, et al. Abnormal cell cycle regulation in malignancy. Am J Clin Pathol. 1999;112(1 suppl 1):S40-S52. 27. Ho DM, Hsu CY, Ting LT, et al. MIB-1 and DNA topoisomerase II alpha could be helpful for predicting longterm survival of patients with glioblastoma. Am J Clin Pathol. 2003;119:715-722. 28. Hsu DW, Louis DN, Efird JT, et al. Use of MIB-1 (Ki-67) immunoreactivity in differentiating grade II and grade III gliomas. J Neuropathol Exp Neurol. 1997;56:857-865. 29. Prayson RA. Cyclin D1 and MIB-1 immunohistochemistry in ependymomas: a study of 41 cases. Am J Clin Pathol. 1998;110:629-634.

30. Rickert CH, Paulus W. Prognosis-related histomorphological and immunohistochemical markers in central nervous system tumors of childhood and adolescence. Acta Neuropathol (Berl). 2005;109:69-92. 31. Gerdes J, Lemke H, Baisch H, et al. Cell cycle analysis of a cell proliferation-associated human nuclear antigen defined by the monoclonal antibody Ki-67. J Immunol. 1984;133:1710-1715. 32. Scholzen T, Gerdes J. The Ki-67 protein: from the known and the unknown. J Cell Physiol. 2000;182:311-322. 33. Agostinho M, Rino J, Braga J, et al. Human topoisomerase IIalpha: targeting to subchromosomal sites of activity during interphase and mitosis. Mol Biol Cell. 2004;15:2388-2400.

Am J Clin Pathol 2005;124:543-549



549 549