- Email: [email protected]
Expression of the Polycomb-Group Protein BMI1 and correlation with p16 in astrocytomas
ARTICLE IN PRESS
Pathology – Research and Practice 204 (2008) 625–631 www.elsevier.de/prp
ORIGINAL ARTICLE
Expression of the Polycomb-Group Protein BMI1 and correlation with p16 in astrocytomas An immunohistochemical study on 80 cases$ Roberto Tiraboscoa, Giovanna De Magliob, Miran Skrapc, Giovanni Falconierib,, Stefano Pizzolittob a
Department of Histopathology, The Royal National Orthopaedic Hospital, NHS Trust, Stanmore, Middlesex, UK Department of Pathology, Azienda Ospedale-Universita`, Udine, Italy c Department of Neurosurgery, Azienda Ospedale-Universita`, Udine, Italy b
Received 22 July 2007; accepted 25 February 2008
Abstract A number of studies on the putative relation between Polycomb-Group (PcG) proteins overexpression and carcinogenesis have been published recently. BMI1, the prototype PcG gene, is critically involved in cell cycle control and differentiation, and despite the regulatory role demonstrated in central nervous system (CNS) development, its implication in brain tumorigenesis is scarcely known. Moreover, to the best of our knowledge, large studies on human brain tumors tissue are lacking. To gain a new insight, we tested 80 primary brain astrocytomas for BMI1 expression using immunohistochemistry and established a correlation with the expression of p16, a negatively regulated target of BMI1 function. Fifty-four cases (72.5%) were BMI1+/p16, and 22 cases (27.5%) were BMI1+/p16+. Slight nonsignificant differences were noted in the expression profile between grades II, III, and IV astrocytomas. However, when the 22 BMI1+/p16+ tumors were examined cytologically, a substantial proportion contained a significant gemistocytic component, which is thought to be an adverse prognostic factor or to display a high degree of anaplasia, suggesting a common molecular mechanism of BMI1/p16 pathway disruption, which may have prognostic implications. r 2008 Elsevier GmbH. All rights reserved. Keywords: Brain tumors; Gliomas; Carcinogenesis; Polycomb-Group Protein
Introduction The link between the disruption of the molecular mechanisms involved in the control of the development $ Presented in abstract form at the 95th Annual Meeting of the United States and Canadian Academy of Pathology, Atlanta, 2006. Corresponding author. Department of Pathology, General Hospital S. Maria della Misericordia, I 33100 Udine, Italy. Tel.: +39 0432 552821; fax: +39 0432 552830. E-mail address: [email protected] (G. Falconieri).
0344-0338/$ - see front matter r 2008 Elsevier GmbH. All rights reserved. doi:10.1016/j.prp.2008.02.007
and differentiation and genesis of cancers has intrigued scientists for a long time, and is one of the most currently addressed issue in tumor biology. The various Polycomb-Group (PcG) gene members regulate several key developmental processes by affecting the expression of homeobox and cell cycle genes [22,30,31]. BMI1 was the first PcG gene discovered. It has extensively been studied in numerous types of lymphomas and carcinomas, where its altered expression was correlated with cancer progression and claimed to be an adverse prognostic factor [3–6,9,17,18,21,23,24,28,29,32].
ARTICLE IN PRESS 626
R. Tirabosco et al. / Pathology – Research and Practice 204 (2008) 625–631
Recently, BMI1 was shown to control the selfrenewal and the proliferative capabilities of neural stem and progenitor cells during development, suggesting a possible role in brain oncogenesis [16,22,26,36]. The aim of this study was the immunohistological assessment of BMI1/p16 expression in a large series of human gliomas in order to further support the experimental evidence of its role on CNS tumorigenesis and to gain an insight into its role as a potential prognosticator.
Table 1. Case series: age/sex and tumor grade distribution of 80 astroglial tumors
Materials and methods Eighty primary astrocytomas were retrieved from the files of a single Institution (Udine, Italy), independently reviewed by the authors, and graded according to the current WHO classification [10]. Tumors featuring at least 20% of gemistocytes were considered ‘‘gemistocyticrich’’. Immunohistochemistry was performed on 4 mm sections, deparaffinized in xylene, and rehydrated through graded alcohols using the following antibodies: Dako CINtecTM p16INK4a Histology kit (according to the manufacturer’s instructions); BMI1 monoclonal antibody (Upstate clone 229-F) at a dilution of 1:100, with heat-induced epitope retrieval in a pH 6.0 citrate buffer and Dako EnvisionTM visualization system. Based on tumor cell reactivity for p16, tumors were subdivided into pattern A (BMI1+/p16) and pattern B (BMI1+/ p16+). Intra-tumoral perivascular lymphocytic cuffing was used as an internal control for BMI1 and p16, the former being strongly and diffusely positive and the latter being diffusely negative.
Results Clinical and main pathologic data A summary of the clinical features is given in Table 1. Age ranged from 18 to 84 years, with 27 cases (34%) occurring during the seventh decade of life. There was a mild male predominance with an M:F ratio of 1.15:1 (42 cases versus 38 cases; 52.5% and 45.5%, respectively). Tumors were graded as follows: 16 diffuse astrocytomas grade II (13 fibrillary, 3 gemistocytic), 15 anaplastic astrocytomas grade III, and 49 astrocytomas grade IV (glioblastoma and variants, gliosarcoma). When stratified by tumor grade, there was no significant difference in age and sex distribution compared to overall data.
Immunohistochemical results The immunohistochemical results are summarized in Table 2. Overall, 58 cases (72.5%) presented pattern A
(Fig. 1) and 22 cases (27.5%) pattern B (Fig. 2). When analyzed according to grade, astrocytomas grade II expressed pattern A in 10 cases (62.5%) and pattern B in 6 cases (37.5%); astrocytomas grade III expressed pattern A in 12 cases (80%) and pattern B in 3 cases (20%); astrocytomas grade IV expressed pattern A in 36 cases (73.5%) and pattern B in 13 cases (26.5%). Two of the three grade III astrocytomas and all 13 cases of grade IV astrocytomas, which belong to pattern B, comprised tumors with a high degree of anaplasia including gliosarcoma, giant cell glioblastoma, and small cell glioblastoma or containing diffuse areas of gemistocytic neoplastic astrocytes. When subtyping astrocytomas grade II (16 cases) into fibrillary (13 cases) and gemistocytic (3 cases) variants, pattern A was expressed in 10 cases (77%) of fibrillary astrocytoma and in none of the gemistocytic astrocytomas, whereas pattern B was recognized in 3 cases (23%) of fibrillary astrocytoma and in 3 cases (100%) of gemistocytic astrocytoma. Staining for p16 in grade II gemistocytic astrocytes and gemistocytic-rich glioblastomas was statistically significant (Po0.05, Fisher test). Regarding the presence or absence of a gemistocytic component, grade III astrocytomas and glioblastomas were separated as detailed in Table 2. In essence, BMI1 was expressed by all tumors regardless of type and grade, whereas different profiles emerged for p16. The proliferative index as assessed by Ki67/Mib1 nuclear uptake was in keeping with tumor grade: up to 2% for grade II, up to 8–10% for grade III, and more than 10% for grade IV. The intensity of nuclear immunostaining was different, ranging from weak to very strong, but was randomly distributed within a single tumor in all glioma grade categories, with no significant differences noted. Clinical follow-up was available for 77 patients. Of the 52 cases belonging to pattern A of immunostaining (BMI1+/p16), 41 died of tumor: 2 cases were grade II, 6 cases were grade III, and 33 cases were grade IV. Of
ARTICLE IN PRESS R. Tirabosco et al. / Pathology – Research and Practice 204 (2008) 625–631
627
Table 2. Summary of immunohistochemical results and separation of BMI1 and p16 staining patterns with respect to histologic tumor type and the presence of a gemistocytic populationa P-value
Tumor type and grade: number and (%)
Pattern A BMI1+/ p16 case number (%)
Pattern B BMI1+/ p16+ case number (%)
All grades: 80 (100%)
58 (72.5%)
22 (27.5%)
Grade II: all 16 (20%) Grade II, fibrillary (13) Grade II, gemistocytic (3)
10 (62.5%) 10 0
6 (37.5) 3 3
0.03
Grade III: all 15 (18.7%) Non-gemistocytic (4) Gemistocytic (11)
12 (80%) 3 9
3 (20%) 1 2
N.S.
Grade IV (or glioblastoma): all 49 (61.2%) Non-gemistocytic (27) Gemistocytic (22)
36 (73.5%) 24 12
13 (26.5%) 3 10
–
0.007
N.S., not significant. a Patterns C (BMI1/p16) and D (BMI1/p16+) not documented in any neoplasm.
the 22 cases belonging to pattern B of immunostaining (BMI1+/p16+) 15 patients died of tumor (3 grade II; 1 grade III; and 11 grade IV).
Discussion The continuous search for prognostic factors in human cancers over the recent years has led to a growing number of clinical, pathological, and molecular studies. Many investigators now mainly focus on the search for molecular markers whose expression in cancers may predict the biological behavior, and which may eventually be used as targets for new therapeutic approaches. Remarkable examples of this so-called ‘‘translational research’’ are represented by the assessment of KIT and HER2 status in gastrointestinal stromal tumors – GIST and breast carcinomas, respectively. A promising new molecule that has recently gained popularity is BMI1, which is the prototype member of a large group of regulatory proteins collectively called PcG proteins [22,26,30]. The members of PcG are critically involved in the control of cell proliferation and differentiation, and are expressed throughout life. BMI1, in particular, has a main role in hematopoiesis, skeletal, and brain development [27]. It acts as a negative regulator of p16 expression, a well-known tumor-suppressor gene. An altered expression of BMI1 has been detected in some types of non-Hodgkin lymphomas, in neuroblastoma, and in various carcinomas including breast, lung, gastrointestinal tract, prostate, and head and neck carcinomas [9,17,18,21,23,24,28,29,32]. Analysis of BMI1 in astrocytoma was also mentioned by Glinsky et al. [7] and Nakahara et al. [17].
These studies on carcinomas have demonstrated that an increased level of BMI1 expression inversely correlates with that of p16. Interestingly, the same expression pattern has also been observed in severe dysplasia and carcinoma in-situ, leading investigators to speculate that this molecular alteration has an important role in carcinoma formation and progression, possibly providing a basis for developing a cancer-specific treatment targeting BMI1 [14]. Recently, BMI1 was shown to control the self-renewal and the proliferative capabilities of neural stem and progenitor cells, and a possible role in CNS tumorigenesis has been suggested. Indeed, expression of BMI1 protein has been reported in medulloblastoma, a highly aggressive embryonal tumor of the childhood, located preferentially in the cerebellum [13]. The same authors, however, did not detect a significant expression of BMI1 protein in glioblastoma, the most common malignant tumor of the CNS, located preferentially in the cerebral hemispheres. This line of evidence prompted us to check at the immunohistochemical level the expression of BMI1 protein in a large series of astrocytic tumors, and we correlated the results with the expression of p16 protein and tumor grade. Data from our series show that all the tumors regardless of the grade show diffuse nuclear staining with anti-BMI1 antibody. The intensity of the immunostaining was different, from weak to very strong, but randomly distributed in all glioma grades. Therefore, any attempt to score the result based on the intensity of the stain was in vain. Therefore, given this universal expression, the usefulness of BMI1 as a prognostic factor cannot be promised. In fact, other than in lymphomas and carcinomas, our observation suggests that there exist no differences in BMI1 expression
ARTICLE IN PRESS 628
R. Tirabosco et al. / Pathology – Research and Practice 204 (2008) 625–631
Fig. 1. Gemistocytic astrocytoma. H&E (a, 10). Evidence of perivascular cuffing of lymphocytes. Tumor cells are positive for both BMI-1 (b, 10 ) and p16 (c, 10 ). Note immunopositivity of small lymphocytes for BMI-1.
Fig. 2. Fibrillary astrocytoma. H&E (a, 10); tumor cells are positive for anti-BMI1 (b, 10); staining for p16 (c, 10) is negative.
between grades II, III, and IV astrocytomas, probably reflecting a more complex cytological and genetical heterogeneity [8]. On the other hand, studies specifically
addressing the expression of BMI1 in non-neoplastic glial cells and, generally, in normal brain are not available. In addition, submission of non-tumoral brain
ARTICLE IN PRESS R. Tirabosco et al. / Pathology – Research and Practice 204 (2008) 625–631
adjacent to neoplastic foci is not a common practice in our experience; microscopic evaluation may also be hampered by tissue inadequacy, whether scant or artifactually fragmented. Nevertheless, BMI1 immunoreactivity could be documented in all cases suitable for microscopic evaluation. More interesting results, however, were achieved when BMI1 expression was correlated with that of p16. Fifty-eight out 80 (72.5%) cases show the expected BMI1+/p16, which we call expression pattern A, which is in keeping with the negative regulation of BMI1 against p16. p16 immunostaining of grade II gemistocytic astrocytomas and glioblastomas featuring a large number of gemistocytes was statistically significant (P ¼ 0.05). It is also controversial whether p16-positive cytoplasmic staining, like nuclear immunostaining, can be a potential prognosticator in high-grade astrocytoma [1]. Immunopositivity for p16 also correlated with poor prognosis in univariate study [19]. It is noteworthy that in our series, staining for p16 was both nuclear and cytoplasmic: however, a more intensive reaction was observed in the cytoplasm. The results are even more intriguing biologically with regard to the remaining 22 tumors (27.5%), which displayed expression pattern B, that is BMI1+/p16+. It is clear that this is a group of tumors that escape the BMI1-mediated p16 suppression, but nevertheless, they show highly anaplastic features on routinely stained sections, or they contain a significant proportion of gemistocytic neoplastic astrocytes (see Results). The issue of gemistocytes in astrocytomas is most controversial and enigmatic in neuropathology. They are defined as glial cells with voluminous, eosinophilic cytoplasm which contains a massive accumulation of glial fibrillary acidic protein [33]. They are believed to be in a non-proliferative status (G0 phase of the cell cycle), suggestive of terminal differentiation. When present in a significant proportion within an astrocytoma grade II (labeled gemistocytic astrocytoma), they are thought to be associated with a more rapid progression to higher grade gliomas, grade III and grade IV [2,11,12,20,25,34,35], although some authors do not believe that a high proportion of gemistocytes would necessarily entail a worse prognosis [15]. It is speculated that the same expression pattern, observed in several tumor types (including gemistocytic astrocytomas grade II, anaplastic astrocytoma grade III with gemistocytic component, glioblastomas with gemistocytic areas, or a high degree of anaplasia, including also giant cell glioblastoma, small cell glioblastoma, and gliosarcoma), apparently indicates a shared tumorigenic molecular pathway, although this cannot be confirmed by immunohistochemical testing only. Molecular insights might lead to further revelations, considering that the speculated p16 suppression mechanisms (such as hypermethylation of the promoter region, deletions, and intragenic mutations) and the BMI1-mediated
629
activity are dynamically regulated processes which, in turn, depend on the phosphorylative status of its promoter [31]. From a clinical point of view, follow-up data would not apparently confirm a significant dissimilarity between the two groups of gliomas (BMI1+/p16 and BMI1+/p16+). The results of our study would therefore suggest that BMI1 does not play a significant prognosticator role in astrocytic tumors. As a corollary, the consistent BMI1 positivity observed in the present series also raises some doubts regarding the claims postulating a prognostic role for BMI1 in neoplasms other than those of CNS. Again, additional investigations on larger series of solid tumors, including clinical follow-up and also based on novel molecular techniques, are needed to corroborate our conclusions.
References [1] M.T. Arifin, S. Hama, Y. Kajiwara, K. Sugiyama, T. Saito, S. Matsuura, F. Yamasaki, K. Arita, K. Kurisu, Cytoplasmic, but not nuclear, p16 expression may signal poor prognosis in high-grade astrocytomas, J. Neurooncol. 77 (2006) 273–277. [2] S. Avninder, M.C. Sharma, P. Deb, V.S. Mehta, A.K. Karak, A.K. Mahapatra, C. Sarkar, Gemistocytic astrocytomas: histomorphology, proliferative potential and genetic alterations – a study of 32 cases, J. Neurooncol. 78 (2006) 123–127. [3] S. Bea, F. Tort, M. Pinyol, X. Puig, L. Hernandez, S. Hernandez, P.L. Fernandez, M. van Lohuizen, D. Colomer, E. Campo, BMI-1 gene amplification and overexpression in hematological malignancies occur mainly in mantle cell lymphomas, Cancer Res. 61 (2001) 2409–2412. [4] O.P. Berezovska, A.B. Glinskii, Z. Yang, X.M. Li, R.M. Hoffman, G.V. Glinsky, Essential role for activation of the Polycomb group (PcG) protein chromatin silencing pathway in metastatic prostate cancer, Cell Cycle 5 (2006) 1886–1901. [5] R.H. Breuer, P.J. Snijders, G.T. Sutedja, R.G. Sewalt, A.P. Otte, P.E. Postmus, C.J. Meijer, F.M. Raaphorst, E.F. Smit, Expression of the p16(INK4a) gene product, methylation of the p16(INK4a) promoter region and expression of the polycomb-group gene BMI-1 in squamous cell lung carcinoma and premalignant endobronchial lesions, Lung Cancer 48 (2005) 299–306. [6] G.P. Dimri, J.L. Martinez, J.J. Jacobs, P. Keblusek, K. Itahana, M. Van Lohuizen, J. Campisi, D.E. Wazer, V. Band, The Bmi-1 oncogene induces telomerase activity and immortalizes human mammary epithelial cells, Cancer Res. 62 (2002) 4736–4745. [7] G.V. Glinsky, O. Berezovska, A.B. Glinskii, Microarray analysis identifies a death-from-cancer signature predicting therapy failure in patients with multiple types of cancer, J. Clin. Invest. 115 (2005) 1503–1521. [8] T. Homma, T. Fukushima, S. Vaccarella, Y. Yonekawa, P.L. Di Patre, S. Franceschi, H. Ohgaki, Correlation
ARTICLE IN PRESS 630
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
R. Tirabosco et al. / Pathology – Research and Practice 204 (2008) 625–631
among pathology, genotype, and patient outcomes in glioblastoma, J. Neuropathol. Exp. Neurol. 65 (2006) 846–854. J.H. Kim, S.Y. Yoon, C.N. Kim, J.H. Joo, S.K. Moon, I.S. Choe, Y.K. Choe, J.W. Kim, The Bmi-1 oncoprotein is overexpressed in human colorectal cancer and correlates with the reduced p16INK4a/p14ARF proteins, Cancer Lett. 203 (2004) 217–224. P. Kleihues, W.K. Cavenee, Pathology and Genetics of Tumours of the Nervous System, International Agency for Research on Cancer, IARC Press, Lyon. S. Kosel, B.W. Scheithauer, M.B. Graeber, Genotypephenotype correlation in gemistocytic astrocytomas, Neurosurgery 48 (2001) 187–193 (discussion 193-184). J.M. Kros, N. Waarsenburg, D.P. Hayes, W.C. Hop, H. van Dekken, Cytogenetic analysis of gemistocytic cells in gliomas, J. Neuropathol. Exp. Neurol. 59 (2000) 679–686. C. Leung, M. Lingbeek, O. Shakhova, J. Liu, E. Tanger, P. Saremaslani, M. Van Lohuizen, S. Marino, Bmi1 is essential for cerebellar development and is overexpressed in human medulloblastomas, Nature 428 (2004) 337–341. L. Liu, L.G. Andrews, T.O. Tollefsbol, Loss of the human polycomb group protein BMI1 promotes cancer-specific cell death, Oncogene 25 (2006) 4370–4375. D.C. Martins, S.M. Malheiros, L.H. Santiago, J.N. Stavale, Gemistocytes in astrocytomas: are they a significant prognostic factor?, J. Neurooncol. 80 (2006) 49–55. A.V. Molofsky, R. Pardal, T. Iwashita, I.K. Park, M.F. Clarke, S.J. Morrison, Bmi-1 dependence distinguishes neural stem cell self-renewal from progenitor proliferation, Nature 425 (2003) 962–967. Y. Nakahara, T. Shiraishi, H. Okamoto, T. Mineta, T. Oishi, K. Sasaki, K. Tabuchi, Detrended fluctuation analysis of genome-wide copy number profiles of glioblastomas using array-based comparative genomic hybridization, Neuro-oncology 6 (2004) 281–289. K. Nowak, K. Kerl, D. Fehr, C. Kramps, C. Gessner, K. Killmer, B. Samans, B. Berwanger, H. Christiansen, W. Lutz, BMI1 is a target gene of E2F-1 and is strongly expressed in primary neuroblastomas, Nucleic Acids Res. 34 (2006) 1745–1754. S.M. Ranuncolo, M. Varela, A. Morandi, J. Lastiri, S. Christiansen, E. Bal de Kier Joffe`, M.G. Pallotta, L. Puricelli, Prognostic value of Mdm2, p53 and p16 in patients with astrocytomas, J. Neurooncol. 68 (2004) 113–121. R.M. Reis, A. Hara, P. Kleihues, H. Ohgaki, Genetic evidence of the neoplastic nature of gemistocytes in astrocytomas, Acta Neuropathol. (Berl.) 102 (2001) 422–425. M. Sanchez-Beato, E. Sanchez, J. Gonzalez-Carrero, M. Morente, A. Diez, L. Sanchez-Verde, M.C. Martin, J.C. Cigudosa, M. Vidal, M.A. Piris, Variability in the expression of polycomb proteins in different normal and tumoral tissues. A pilot study using tissue microarrays, Mod. Pathol. 19 (2006) 684–694. O. Shakhova, C. Leung, S. Marino, Bmi1 in development and tumorigenesis of the central nervous system, J. Mol. Med. 83 (2005) 596–600.
[23] L.B. Song, M.S. Zeng, W.T. Liao, L. Zhang, H.Y. Mo, W.L. Liu, J.Y. Shao, Q.L. Wu, M.Z. Li, Y.F. Xia, L.W. Fu, W.L. Huang, G.P. Dimri, V. Band, Y.X. Zeng, Bmi-1 is a novel molecular marker of nasopharyngeal carcinoma progression and immortalizes primary human nasopharyngeal epithelial cells, Cancer Res. 66 (2006) 6225–6232. [24] K. Tateishi, M. Ohta, F. Kanai, B. Guleng, Y. Tanaka, Y. Asaoka, M. Tada, M. Seto, A. Jazag, L. Lianjie, M. Okamoto, H. Isayama, M. Tada, H. Yoshida, T. Kawabe, M. Omata, Dysregulated expression of stem cell factor Bmi1 in precancerous lesions of the gastrointestinal tract, Clin. Cancer Res. 12 (2006) 6960–6966. [25] T. Tihan, P. Vohra, M.S. Berger, G.E. Keles, Definition and diagnostic implications of gemistocytic astrocytomas: a pathological perspective, J. Neurooncol. 76 (2006) 175–183. [26] M.E. Valk-Lingbeek, S.W. Bruggeman, M. van Lohuizen, Stem cells and cancer; the polycomb connection, Cell 118 (2004) 409–418. [27] N.M. van der Lugt, J. Domen, K. Linders, M. van Roon, E. Robanus-Maandag, H. te Riele, M. van der Valk, J. Deschamps, M. Sofroniew, M. van Lohuizen, et al., Posterior transformation, neurological abnormalities, and severe hematopoietic defects in mice with a targeted deletion of the bmi-1 proto-oncogene, Genes Dev. 8 (1994) 757–769. [28] F.J. van Kemenade, F.M. Raaphorst, T. Blokzijl, E. Fieret, K.M. Hamer, D.P. Satijn, A.P. Otte, C.J. Meijer, Coexpression of BMI-1 and EZH2 polycomb-group proteins is associated with cycling cells and degree of malignancy in B-cell non-Hodgkin lymphoma, Blood 97 (2001) 3896–3901. [29] G.J. van Leenders, D. Dukers, D. Hessels, S.W. van den Kieboom, C.A. Hulsbergen, J.A. Witjes, A.P. Otte, C.J. Meijer, F.M. Raaphorst, Polycomb-group oncogenes EZH2, BMI1, and RING1 are overexpressed in prostate cancer with adverse pathologic and clinical features, Eur. Urol. (2006). [30] J.W. Voncken, D. Schweizer, L. Aagaard, L. Sattler, M.F. Jantsch, M. van Lohuizen, Chromatin-association of the Polycomb group protein BMI1 is cell cycleregulated and correlates with its phosphorylation status, J. Cell Sci. 112 (Part 24) (1999) 4627–4639. [31] J.W. Voncken, H. Niessen, B. Neufeld, U. Rennefahrt, V. Dahlmans, N. Kubben, B. Holzer, S. Ludwig, U.R. Rapp, MAPKAP kinase 3pK phosphorylates and regulates chromatin association of the polycomb group protein Bmi1, J. Biol. Chem. 280 (2005) 5178–5187. [32] S. Vonlanthen, J. Heighway, H.J. Altermatt, M. Gugger, A. Kappeler, M.M. Borner, M. van Lohuizen, D.C. Betticher, The bmi-1 oncoprotein is differentially expressed in non-small cell lung cancer and correlates with INK4A-ARF locus expression, Br. J. Cancer 84 (2001) 1372–1376. [33] K. Watanabe, O. Tachibana, Y. Yonekawa, P. Kleihues, H. Ohgaki, Role of gemistocytes in astrocytoma progression, Lab. Invest. 76 (1997) 277–284. [34] K. Watanabe, A. Peraud, C. Gratas, S. Wakai, P. Kleihues, H. Ohgaki, p53 and PTEN gene mutations in
ARTICLE IN PRESS R. Tirabosco et al. / Pathology – Research and Practice 204 (2008) 625–631
gemistocytic astrocytomas, Acta Neuropathol. (Berl.) 95 (1998) 559–564. [35] H.J. Yang, J.E. Kim, S.H. Paek, J.G. Chi, H.W. Jung, D.G. Kim, The significance of gemistocytes in astrocytoma, Acta Neurochir. (Wien) 145 (2003) 1097–1103 (discussion 1103).
631
[36] D. Zencak, M. Lingbeek, C. Kostic, M. Tekaya, E. Tanger, D. Hornfeld, M. Jaquet, F.L. Munier, D.F. Schorderet, M. van Lohuizen, Y. Arsenijevic, Bmi1 loss produces an increase in astroglial cells and a decrease in neural stem cell population and proliferation, J. Neurosci. 25 (2005) 5774–5783.
Pathology – Research and Practice 204 (2008) 625–631 www.elsevier.de/prp
ORIGINAL ARTICLE
Expression of the Polycomb-Group Protein BMI1 and correlation with p16 in astrocytomas An immunohistochemical study on 80 cases$ Roberto Tiraboscoa, Giovanna De Magliob, Miran Skrapc, Giovanni Falconierib,, Stefano Pizzolittob a
Department of Histopathology, The Royal National Orthopaedic Hospital, NHS Trust, Stanmore, Middlesex, UK Department of Pathology, Azienda Ospedale-Universita`, Udine, Italy c Department of Neurosurgery, Azienda Ospedale-Universita`, Udine, Italy b
Received 22 July 2007; accepted 25 February 2008
Abstract A number of studies on the putative relation between Polycomb-Group (PcG) proteins overexpression and carcinogenesis have been published recently. BMI1, the prototype PcG gene, is critically involved in cell cycle control and differentiation, and despite the regulatory role demonstrated in central nervous system (CNS) development, its implication in brain tumorigenesis is scarcely known. Moreover, to the best of our knowledge, large studies on human brain tumors tissue are lacking. To gain a new insight, we tested 80 primary brain astrocytomas for BMI1 expression using immunohistochemistry and established a correlation with the expression of p16, a negatively regulated target of BMI1 function. Fifty-four cases (72.5%) were BMI1+/p16, and 22 cases (27.5%) were BMI1+/p16+. Slight nonsignificant differences were noted in the expression profile between grades II, III, and IV astrocytomas. However, when the 22 BMI1+/p16+ tumors were examined cytologically, a substantial proportion contained a significant gemistocytic component, which is thought to be an adverse prognostic factor or to display a high degree of anaplasia, suggesting a common molecular mechanism of BMI1/p16 pathway disruption, which may have prognostic implications. r 2008 Elsevier GmbH. All rights reserved. Keywords: Brain tumors; Gliomas; Carcinogenesis; Polycomb-Group Protein
Introduction The link between the disruption of the molecular mechanisms involved in the control of the development $ Presented in abstract form at the 95th Annual Meeting of the United States and Canadian Academy of Pathology, Atlanta, 2006. Corresponding author. Department of Pathology, General Hospital S. Maria della Misericordia, I 33100 Udine, Italy. Tel.: +39 0432 552821; fax: +39 0432 552830. E-mail address: [email protected] (G. Falconieri).
0344-0338/$ - see front matter r 2008 Elsevier GmbH. All rights reserved. doi:10.1016/j.prp.2008.02.007
and differentiation and genesis of cancers has intrigued scientists for a long time, and is one of the most currently addressed issue in tumor biology. The various Polycomb-Group (PcG) gene members regulate several key developmental processes by affecting the expression of homeobox and cell cycle genes [22,30,31]. BMI1 was the first PcG gene discovered. It has extensively been studied in numerous types of lymphomas and carcinomas, where its altered expression was correlated with cancer progression and claimed to be an adverse prognostic factor [3–6,9,17,18,21,23,24,28,29,32].
ARTICLE IN PRESS 626
R. Tirabosco et al. / Pathology – Research and Practice 204 (2008) 625–631
Recently, BMI1 was shown to control the selfrenewal and the proliferative capabilities of neural stem and progenitor cells during development, suggesting a possible role in brain oncogenesis [16,22,26,36]. The aim of this study was the immunohistological assessment of BMI1/p16 expression in a large series of human gliomas in order to further support the experimental evidence of its role on CNS tumorigenesis and to gain an insight into its role as a potential prognosticator.
Table 1. Case series: age/sex and tumor grade distribution of 80 astroglial tumors
Materials and methods Eighty primary astrocytomas were retrieved from the files of a single Institution (Udine, Italy), independently reviewed by the authors, and graded according to the current WHO classification [10]. Tumors featuring at least 20% of gemistocytes were considered ‘‘gemistocyticrich’’. Immunohistochemistry was performed on 4 mm sections, deparaffinized in xylene, and rehydrated through graded alcohols using the following antibodies: Dako CINtecTM p16INK4a Histology kit (according to the manufacturer’s instructions); BMI1 monoclonal antibody (Upstate clone 229-F) at a dilution of 1:100, with heat-induced epitope retrieval in a pH 6.0 citrate buffer and Dako EnvisionTM visualization system. Based on tumor cell reactivity for p16, tumors were subdivided into pattern A (BMI1+/p16) and pattern B (BMI1+/ p16+). Intra-tumoral perivascular lymphocytic cuffing was used as an internal control for BMI1 and p16, the former being strongly and diffusely positive and the latter being diffusely negative.
Results Clinical and main pathologic data A summary of the clinical features is given in Table 1. Age ranged from 18 to 84 years, with 27 cases (34%) occurring during the seventh decade of life. There was a mild male predominance with an M:F ratio of 1.15:1 (42 cases versus 38 cases; 52.5% and 45.5%, respectively). Tumors were graded as follows: 16 diffuse astrocytomas grade II (13 fibrillary, 3 gemistocytic), 15 anaplastic astrocytomas grade III, and 49 astrocytomas grade IV (glioblastoma and variants, gliosarcoma). When stratified by tumor grade, there was no significant difference in age and sex distribution compared to overall data.
Immunohistochemical results The immunohistochemical results are summarized in Table 2. Overall, 58 cases (72.5%) presented pattern A
(Fig. 1) and 22 cases (27.5%) pattern B (Fig. 2). When analyzed according to grade, astrocytomas grade II expressed pattern A in 10 cases (62.5%) and pattern B in 6 cases (37.5%); astrocytomas grade III expressed pattern A in 12 cases (80%) and pattern B in 3 cases (20%); astrocytomas grade IV expressed pattern A in 36 cases (73.5%) and pattern B in 13 cases (26.5%). Two of the three grade III astrocytomas and all 13 cases of grade IV astrocytomas, which belong to pattern B, comprised tumors with a high degree of anaplasia including gliosarcoma, giant cell glioblastoma, and small cell glioblastoma or containing diffuse areas of gemistocytic neoplastic astrocytes. When subtyping astrocytomas grade II (16 cases) into fibrillary (13 cases) and gemistocytic (3 cases) variants, pattern A was expressed in 10 cases (77%) of fibrillary astrocytoma and in none of the gemistocytic astrocytomas, whereas pattern B was recognized in 3 cases (23%) of fibrillary astrocytoma and in 3 cases (100%) of gemistocytic astrocytoma. Staining for p16 in grade II gemistocytic astrocytes and gemistocytic-rich glioblastomas was statistically significant (Po0.05, Fisher test). Regarding the presence or absence of a gemistocytic component, grade III astrocytomas and glioblastomas were separated as detailed in Table 2. In essence, BMI1 was expressed by all tumors regardless of type and grade, whereas different profiles emerged for p16. The proliferative index as assessed by Ki67/Mib1 nuclear uptake was in keeping with tumor grade: up to 2% for grade II, up to 8–10% for grade III, and more than 10% for grade IV. The intensity of nuclear immunostaining was different, ranging from weak to very strong, but was randomly distributed within a single tumor in all glioma grade categories, with no significant differences noted. Clinical follow-up was available for 77 patients. Of the 52 cases belonging to pattern A of immunostaining (BMI1+/p16), 41 died of tumor: 2 cases were grade II, 6 cases were grade III, and 33 cases were grade IV. Of
ARTICLE IN PRESS R. Tirabosco et al. / Pathology – Research and Practice 204 (2008) 625–631
627
Table 2. Summary of immunohistochemical results and separation of BMI1 and p16 staining patterns with respect to histologic tumor type and the presence of a gemistocytic populationa P-value
Tumor type and grade: number and (%)
Pattern A BMI1+/ p16 case number (%)
Pattern B BMI1+/ p16+ case number (%)
All grades: 80 (100%)
58 (72.5%)
22 (27.5%)
Grade II: all 16 (20%) Grade II, fibrillary (13) Grade II, gemistocytic (3)
10 (62.5%) 10 0
6 (37.5) 3 3
0.03
Grade III: all 15 (18.7%) Non-gemistocytic (4) Gemistocytic (11)
12 (80%) 3 9
3 (20%) 1 2
N.S.
Grade IV (or glioblastoma): all 49 (61.2%) Non-gemistocytic (27) Gemistocytic (22)
36 (73.5%) 24 12
13 (26.5%) 3 10
–
0.007
N.S., not significant. a Patterns C (BMI1/p16) and D (BMI1/p16+) not documented in any neoplasm.
the 22 cases belonging to pattern B of immunostaining (BMI1+/p16+) 15 patients died of tumor (3 grade II; 1 grade III; and 11 grade IV).
Discussion The continuous search for prognostic factors in human cancers over the recent years has led to a growing number of clinical, pathological, and molecular studies. Many investigators now mainly focus on the search for molecular markers whose expression in cancers may predict the biological behavior, and which may eventually be used as targets for new therapeutic approaches. Remarkable examples of this so-called ‘‘translational research’’ are represented by the assessment of KIT and HER2 status in gastrointestinal stromal tumors – GIST and breast carcinomas, respectively. A promising new molecule that has recently gained popularity is BMI1, which is the prototype member of a large group of regulatory proteins collectively called PcG proteins [22,26,30]. The members of PcG are critically involved in the control of cell proliferation and differentiation, and are expressed throughout life. BMI1, in particular, has a main role in hematopoiesis, skeletal, and brain development [27]. It acts as a negative regulator of p16 expression, a well-known tumor-suppressor gene. An altered expression of BMI1 has been detected in some types of non-Hodgkin lymphomas, in neuroblastoma, and in various carcinomas including breast, lung, gastrointestinal tract, prostate, and head and neck carcinomas [9,17,18,21,23,24,28,29,32]. Analysis of BMI1 in astrocytoma was also mentioned by Glinsky et al. [7] and Nakahara et al. [17].
These studies on carcinomas have demonstrated that an increased level of BMI1 expression inversely correlates with that of p16. Interestingly, the same expression pattern has also been observed in severe dysplasia and carcinoma in-situ, leading investigators to speculate that this molecular alteration has an important role in carcinoma formation and progression, possibly providing a basis for developing a cancer-specific treatment targeting BMI1 [14]. Recently, BMI1 was shown to control the self-renewal and the proliferative capabilities of neural stem and progenitor cells, and a possible role in CNS tumorigenesis has been suggested. Indeed, expression of BMI1 protein has been reported in medulloblastoma, a highly aggressive embryonal tumor of the childhood, located preferentially in the cerebellum [13]. The same authors, however, did not detect a significant expression of BMI1 protein in glioblastoma, the most common malignant tumor of the CNS, located preferentially in the cerebral hemispheres. This line of evidence prompted us to check at the immunohistochemical level the expression of BMI1 protein in a large series of astrocytic tumors, and we correlated the results with the expression of p16 protein and tumor grade. Data from our series show that all the tumors regardless of the grade show diffuse nuclear staining with anti-BMI1 antibody. The intensity of the immunostaining was different, from weak to very strong, but randomly distributed in all glioma grades. Therefore, any attempt to score the result based on the intensity of the stain was in vain. Therefore, given this universal expression, the usefulness of BMI1 as a prognostic factor cannot be promised. In fact, other than in lymphomas and carcinomas, our observation suggests that there exist no differences in BMI1 expression
ARTICLE IN PRESS 628
R. Tirabosco et al. / Pathology – Research and Practice 204 (2008) 625–631
Fig. 1. Gemistocytic astrocytoma. H&E (a, 10). Evidence of perivascular cuffing of lymphocytes. Tumor cells are positive for both BMI-1 (b, 10 ) and p16 (c, 10 ). Note immunopositivity of small lymphocytes for BMI-1.
Fig. 2. Fibrillary astrocytoma. H&E (a, 10); tumor cells are positive for anti-BMI1 (b, 10); staining for p16 (c, 10) is negative.
between grades II, III, and IV astrocytomas, probably reflecting a more complex cytological and genetical heterogeneity [8]. On the other hand, studies specifically
addressing the expression of BMI1 in non-neoplastic glial cells and, generally, in normal brain are not available. In addition, submission of non-tumoral brain
ARTICLE IN PRESS R. Tirabosco et al. / Pathology – Research and Practice 204 (2008) 625–631
adjacent to neoplastic foci is not a common practice in our experience; microscopic evaluation may also be hampered by tissue inadequacy, whether scant or artifactually fragmented. Nevertheless, BMI1 immunoreactivity could be documented in all cases suitable for microscopic evaluation. More interesting results, however, were achieved when BMI1 expression was correlated with that of p16. Fifty-eight out 80 (72.5%) cases show the expected BMI1+/p16, which we call expression pattern A, which is in keeping with the negative regulation of BMI1 against p16. p16 immunostaining of grade II gemistocytic astrocytomas and glioblastomas featuring a large number of gemistocytes was statistically significant (P ¼ 0.05). It is also controversial whether p16-positive cytoplasmic staining, like nuclear immunostaining, can be a potential prognosticator in high-grade astrocytoma [1]. Immunopositivity for p16 also correlated with poor prognosis in univariate study [19]. It is noteworthy that in our series, staining for p16 was both nuclear and cytoplasmic: however, a more intensive reaction was observed in the cytoplasm. The results are even more intriguing biologically with regard to the remaining 22 tumors (27.5%), which displayed expression pattern B, that is BMI1+/p16+. It is clear that this is a group of tumors that escape the BMI1-mediated p16 suppression, but nevertheless, they show highly anaplastic features on routinely stained sections, or they contain a significant proportion of gemistocytic neoplastic astrocytes (see Results). The issue of gemistocytes in astrocytomas is most controversial and enigmatic in neuropathology. They are defined as glial cells with voluminous, eosinophilic cytoplasm which contains a massive accumulation of glial fibrillary acidic protein [33]. They are believed to be in a non-proliferative status (G0 phase of the cell cycle), suggestive of terminal differentiation. When present in a significant proportion within an astrocytoma grade II (labeled gemistocytic astrocytoma), they are thought to be associated with a more rapid progression to higher grade gliomas, grade III and grade IV [2,11,12,20,25,34,35], although some authors do not believe that a high proportion of gemistocytes would necessarily entail a worse prognosis [15]. It is speculated that the same expression pattern, observed in several tumor types (including gemistocytic astrocytomas grade II, anaplastic astrocytoma grade III with gemistocytic component, glioblastomas with gemistocytic areas, or a high degree of anaplasia, including also giant cell glioblastoma, small cell glioblastoma, and gliosarcoma), apparently indicates a shared tumorigenic molecular pathway, although this cannot be confirmed by immunohistochemical testing only. Molecular insights might lead to further revelations, considering that the speculated p16 suppression mechanisms (such as hypermethylation of the promoter region, deletions, and intragenic mutations) and the BMI1-mediated
629
activity are dynamically regulated processes which, in turn, depend on the phosphorylative status of its promoter [31]. From a clinical point of view, follow-up data would not apparently confirm a significant dissimilarity between the two groups of gliomas (BMI1+/p16 and BMI1+/p16+). The results of our study would therefore suggest that BMI1 does not play a significant prognosticator role in astrocytic tumors. As a corollary, the consistent BMI1 positivity observed in the present series also raises some doubts regarding the claims postulating a prognostic role for BMI1 in neoplasms other than those of CNS. Again, additional investigations on larger series of solid tumors, including clinical follow-up and also based on novel molecular techniques, are needed to corroborate our conclusions.
References [1] M.T. Arifin, S. Hama, Y. Kajiwara, K. Sugiyama, T. Saito, S. Matsuura, F. Yamasaki, K. Arita, K. Kurisu, Cytoplasmic, but not nuclear, p16 expression may signal poor prognosis in high-grade astrocytomas, J. Neurooncol. 77 (2006) 273–277. [2] S. Avninder, M.C. Sharma, P. Deb, V.S. Mehta, A.K. Karak, A.K. Mahapatra, C. Sarkar, Gemistocytic astrocytomas: histomorphology, proliferative potential and genetic alterations – a study of 32 cases, J. Neurooncol. 78 (2006) 123–127. [3] S. Bea, F. Tort, M. Pinyol, X. Puig, L. Hernandez, S. Hernandez, P.L. Fernandez, M. van Lohuizen, D. Colomer, E. Campo, BMI-1 gene amplification and overexpression in hematological malignancies occur mainly in mantle cell lymphomas, Cancer Res. 61 (2001) 2409–2412. [4] O.P. Berezovska, A.B. Glinskii, Z. Yang, X.M. Li, R.M. Hoffman, G.V. Glinsky, Essential role for activation of the Polycomb group (PcG) protein chromatin silencing pathway in metastatic prostate cancer, Cell Cycle 5 (2006) 1886–1901. [5] R.H. Breuer, P.J. Snijders, G.T. Sutedja, R.G. Sewalt, A.P. Otte, P.E. Postmus, C.J. Meijer, F.M. Raaphorst, E.F. Smit, Expression of the p16(INK4a) gene product, methylation of the p16(INK4a) promoter region and expression of the polycomb-group gene BMI-1 in squamous cell lung carcinoma and premalignant endobronchial lesions, Lung Cancer 48 (2005) 299–306. [6] G.P. Dimri, J.L. Martinez, J.J. Jacobs, P. Keblusek, K. Itahana, M. Van Lohuizen, J. Campisi, D.E. Wazer, V. Band, The Bmi-1 oncogene induces telomerase activity and immortalizes human mammary epithelial cells, Cancer Res. 62 (2002) 4736–4745. [7] G.V. Glinsky, O. Berezovska, A.B. Glinskii, Microarray analysis identifies a death-from-cancer signature predicting therapy failure in patients with multiple types of cancer, J. Clin. Invest. 115 (2005) 1503–1521. [8] T. Homma, T. Fukushima, S. Vaccarella, Y. Yonekawa, P.L. Di Patre, S. Franceschi, H. Ohgaki, Correlation
ARTICLE IN PRESS 630
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
R. Tirabosco et al. / Pathology – Research and Practice 204 (2008) 625–631
among pathology, genotype, and patient outcomes in glioblastoma, J. Neuropathol. Exp. Neurol. 65 (2006) 846–854. J.H. Kim, S.Y. Yoon, C.N. Kim, J.H. Joo, S.K. Moon, I.S. Choe, Y.K. Choe, J.W. Kim, The Bmi-1 oncoprotein is overexpressed in human colorectal cancer and correlates with the reduced p16INK4a/p14ARF proteins, Cancer Lett. 203 (2004) 217–224. P. Kleihues, W.K. Cavenee, Pathology and Genetics of Tumours of the Nervous System, International Agency for Research on Cancer, IARC Press, Lyon. S. Kosel, B.W. Scheithauer, M.B. Graeber, Genotypephenotype correlation in gemistocytic astrocytomas, Neurosurgery 48 (2001) 187–193 (discussion 193-184). J.M. Kros, N. Waarsenburg, D.P. Hayes, W.C. Hop, H. van Dekken, Cytogenetic analysis of gemistocytic cells in gliomas, J. Neuropathol. Exp. Neurol. 59 (2000) 679–686. C. Leung, M. Lingbeek, O. Shakhova, J. Liu, E. Tanger, P. Saremaslani, M. Van Lohuizen, S. Marino, Bmi1 is essential for cerebellar development and is overexpressed in human medulloblastomas, Nature 428 (2004) 337–341. L. Liu, L.G. Andrews, T.O. Tollefsbol, Loss of the human polycomb group protein BMI1 promotes cancer-specific cell death, Oncogene 25 (2006) 4370–4375. D.C. Martins, S.M. Malheiros, L.H. Santiago, J.N. Stavale, Gemistocytes in astrocytomas: are they a significant prognostic factor?, J. Neurooncol. 80 (2006) 49–55. A.V. Molofsky, R. Pardal, T. Iwashita, I.K. Park, M.F. Clarke, S.J. Morrison, Bmi-1 dependence distinguishes neural stem cell self-renewal from progenitor proliferation, Nature 425 (2003) 962–967. Y. Nakahara, T. Shiraishi, H. Okamoto, T. Mineta, T. Oishi, K. Sasaki, K. Tabuchi, Detrended fluctuation analysis of genome-wide copy number profiles of glioblastomas using array-based comparative genomic hybridization, Neuro-oncology 6 (2004) 281–289. K. Nowak, K. Kerl, D. Fehr, C. Kramps, C. Gessner, K. Killmer, B. Samans, B. Berwanger, H. Christiansen, W. Lutz, BMI1 is a target gene of E2F-1 and is strongly expressed in primary neuroblastomas, Nucleic Acids Res. 34 (2006) 1745–1754. S.M. Ranuncolo, M. Varela, A. Morandi, J. Lastiri, S. Christiansen, E. Bal de Kier Joffe`, M.G. Pallotta, L. Puricelli, Prognostic value of Mdm2, p53 and p16 in patients with astrocytomas, J. Neurooncol. 68 (2004) 113–121. R.M. Reis, A. Hara, P. Kleihues, H. Ohgaki, Genetic evidence of the neoplastic nature of gemistocytes in astrocytomas, Acta Neuropathol. (Berl.) 102 (2001) 422–425. M. Sanchez-Beato, E. Sanchez, J. Gonzalez-Carrero, M. Morente, A. Diez, L. Sanchez-Verde, M.C. Martin, J.C. Cigudosa, M. Vidal, M.A. Piris, Variability in the expression of polycomb proteins in different normal and tumoral tissues. A pilot study using tissue microarrays, Mod. Pathol. 19 (2006) 684–694. O. Shakhova, C. Leung, S. Marino, Bmi1 in development and tumorigenesis of the central nervous system, J. Mol. Med. 83 (2005) 596–600.
[23] L.B. Song, M.S. Zeng, W.T. Liao, L. Zhang, H.Y. Mo, W.L. Liu, J.Y. Shao, Q.L. Wu, M.Z. Li, Y.F. Xia, L.W. Fu, W.L. Huang, G.P. Dimri, V. Band, Y.X. Zeng, Bmi-1 is a novel molecular marker of nasopharyngeal carcinoma progression and immortalizes primary human nasopharyngeal epithelial cells, Cancer Res. 66 (2006) 6225–6232. [24] K. Tateishi, M. Ohta, F. Kanai, B. Guleng, Y. Tanaka, Y. Asaoka, M. Tada, M. Seto, A. Jazag, L. Lianjie, M. Okamoto, H. Isayama, M. Tada, H. Yoshida, T. Kawabe, M. Omata, Dysregulated expression of stem cell factor Bmi1 in precancerous lesions of the gastrointestinal tract, Clin. Cancer Res. 12 (2006) 6960–6966. [25] T. Tihan, P. Vohra, M.S. Berger, G.E. Keles, Definition and diagnostic implications of gemistocytic astrocytomas: a pathological perspective, J. Neurooncol. 76 (2006) 175–183. [26] M.E. Valk-Lingbeek, S.W. Bruggeman, M. van Lohuizen, Stem cells and cancer; the polycomb connection, Cell 118 (2004) 409–418. [27] N.M. van der Lugt, J. Domen, K. Linders, M. van Roon, E. Robanus-Maandag, H. te Riele, M. van der Valk, J. Deschamps, M. Sofroniew, M. van Lohuizen, et al., Posterior transformation, neurological abnormalities, and severe hematopoietic defects in mice with a targeted deletion of the bmi-1 proto-oncogene, Genes Dev. 8 (1994) 757–769. [28] F.J. van Kemenade, F.M. Raaphorst, T. Blokzijl, E. Fieret, K.M. Hamer, D.P. Satijn, A.P. Otte, C.J. Meijer, Coexpression of BMI-1 and EZH2 polycomb-group proteins is associated with cycling cells and degree of malignancy in B-cell non-Hodgkin lymphoma, Blood 97 (2001) 3896–3901. [29] G.J. van Leenders, D. Dukers, D. Hessels, S.W. van den Kieboom, C.A. Hulsbergen, J.A. Witjes, A.P. Otte, C.J. Meijer, F.M. Raaphorst, Polycomb-group oncogenes EZH2, BMI1, and RING1 are overexpressed in prostate cancer with adverse pathologic and clinical features, Eur. Urol. (2006). [30] J.W. Voncken, D. Schweizer, L. Aagaard, L. Sattler, M.F. Jantsch, M. van Lohuizen, Chromatin-association of the Polycomb group protein BMI1 is cell cycleregulated and correlates with its phosphorylation status, J. Cell Sci. 112 (Part 24) (1999) 4627–4639. [31] J.W. Voncken, H. Niessen, B. Neufeld, U. Rennefahrt, V. Dahlmans, N. Kubben, B. Holzer, S. Ludwig, U.R. Rapp, MAPKAP kinase 3pK phosphorylates and regulates chromatin association of the polycomb group protein Bmi1, J. Biol. Chem. 280 (2005) 5178–5187. [32] S. Vonlanthen, J. Heighway, H.J. Altermatt, M. Gugger, A. Kappeler, M.M. Borner, M. van Lohuizen, D.C. Betticher, The bmi-1 oncoprotein is differentially expressed in non-small cell lung cancer and correlates with INK4A-ARF locus expression, Br. J. Cancer 84 (2001) 1372–1376. [33] K. Watanabe, O. Tachibana, Y. Yonekawa, P. Kleihues, H. Ohgaki, Role of gemistocytes in astrocytoma progression, Lab. Invest. 76 (1997) 277–284. [34] K. Watanabe, A. Peraud, C. Gratas, S. Wakai, P. Kleihues, H. Ohgaki, p53 and PTEN gene mutations in
ARTICLE IN PRESS R. Tirabosco et al. / Pathology – Research and Practice 204 (2008) 625–631
gemistocytic astrocytomas, Acta Neuropathol. (Berl.) 95 (1998) 559–564. [35] H.J. Yang, J.E. Kim, S.H. Paek, J.G. Chi, H.W. Jung, D.G. Kim, The significance of gemistocytes in astrocytoma, Acta Neurochir. (Wien) 145 (2003) 1097–1103 (discussion 1103).
631
[36] D. Zencak, M. Lingbeek, C. Kostic, M. Tekaya, E. Tanger, D. Hornfeld, M. Jaquet, F.L. Munier, D.F. Schorderet, M. van Lohuizen, Y. Arsenijevic, Bmi1 loss produces an increase in astroglial cells and a decrease in neural stem cell population and proliferation, J. Neurosci. 25 (2005) 5774–5783.