- Email: [email protected]
Expression of mos in astrocytic tumors and its potential role in neoplastic progression
Expression of mos in Astrocytic Tumors and its Potential Role in Neoplastic Progression BRANKO PERUNOVIC, ATHANASIOS ATHANASIOU, ROBERT D. QUILTY, VASSILIS G. GORGOULIS, CHRISTOS KITTAS, AND SETH LOVE The c-mos gene and its protein product mos, components of the mitogen-activated protein kinase transduction pathway, are known to be involved in the control of meiosis and mitosis. Apart from a study on lung carcinomas, there is little information about its role in human neoplasia. The aim of this study was to investigate expression of mos in astrocytic tumors and to correlate it with accumulation of p53. We studied expression of mos in 62 cases of supratentorial astrocytic tumor. Intracytoplasmic immunostaining for mos was found in 28 (45%) cases: 3 of 20 (15%) grade 2 astrocytomas, 9 of 20 (45%) grade 3 anaplastic astrocytomas, and 16 of 22 (73%) glioblastomas. Immunopositivity for mos correlated significantly (P < 0.01) with tumor grade but not with p53 expression. In contrast to the findings in relation to lung tumors, immunopositivity for mos in astrocytic tumors did not predict recurrence-free or overall survival time. Cyto-
plasmic immunostaining was observed in scattered large cortical neurons adjacent to tumors, possibly due to stress-induced abortive entry into the cell cycle. The correlation of mos immunopositivity with tumor grade may reflect the expansion of more malignant mospositive clones. This study provides evidence that mos may be involved in the neoplastic progression of a proportion of astrocytic tumors. HUM PATHOL 33:703-707. Copyright 2002, Elsevier Science (USA). All rights reserved. Key words: astrocytoma, astrocytic neoplasms, c-mos gene, immunohistochemistry, brain neoplasms, lymphokines, p53, proto-onko gene proteins mos. Abbreviations: MAPK, mitogen-activated protein kinase; WHO, World Health Organization.
During last decade, much has been learned about the genetic alterations that cause the initiation and progression of gliomas.1 These alterations include structural and functional abnormalities in a range growth factors, growth factor receptors and tumor suppressor genes, many of which are known to be involved in controlling progression through the cell cycle, cell death and cell differentiation. A pathway that is involved in the control of meiosis and probably mitosis but about which little information is available in relation to glial tumors is the mitogen-activated protein kinase (MAPK) transduction pathway and, in particular, the c-mos gene and its protein product, mos. C-mos was the first proto-oncogene of the serine/threonine protein kinase super-family to be discovered.2 In man it is located on chromosome 8q11-12 and encodes mos, a 39-kDa protein.3 All of the known biological roles of mos are mediated through its MAPK activity.3 Much c-mos–related research has concerned its role in meiosis, and several extensive reviews are available on this topic.3-5 Mos is an important regulator of meiotic maturation and is essential for both meiosis I and II, as has been shown in amphibian6 and mammalian7 experimental systems. It is a component of a large multiprotein complex, known as cytostatic factor, that is involved in metaphase II arrest of eggs in verte-
brates.8 There is also evidence of a permissive role in the organization of microtubules and assembly of meiotic spindle.3,9 C-mos knockout mice have severely reduced fertility, due to the premature release of the eggs from cell cycle arrest before their fertilization. The parthenogenetically activated germ cells that result may give rise to ovarian teratomas at an early age.10,11 The role of c-mos in somatic tissues is poorly understood. Mos or its coding mRNA appears to be expressed at low but unequal levels in most somatic tissues, although there is some disparity between the data on mRNA and protein levels.12,13 Mos has been found to be associated with several other proteins in both normal and transformed cells, including vimentin,14 tubulin,15 hsp70,16 p34cdc2,17 p35cdk,18 and myoD.19 Increased expression of c-mos in somatic cells has been linked to neoplastic transformation.20 Two mechanisms have been proposed to explain this: (1) the mos/MAPK pathway interferes with normal cyclin D1–Cdk4 –pRb– E2F signaling3 and (2) increased expression of c-mos leads to acquisition of a meiosis-like phenotype that may compromise the mitotic checkpoints that normally ensure the integrity of genetic material and symmetry of cell division during mitosis.21 High levels of mos have been reported to cause somatic cell death22 due to p53-mediated growth arrest and apoptosis.21 The authors of the latter study suggested that loss of p53 cell checkpoint activity may be a key permissive event in c-mos–induced neoplastic transformation. In previous studies we have shown that increased transcription of c-mos and elevated levels of immunostainable mos are demonstrable in a high proportion of human lung carcinomas23,24 in the absence of mutation or abnormal expression of p53. The aim of the present study was to investigate the expression of mos in astrocytic tumors. In addition, because mutations of the p53
From the Departments of Neuropathology and Neurosurgery, Frenchay Hospital, Bristol, United Kingdom, and the Department of Histology and Embryology, School of Medicine, University of Athens, Athens, Greece. Accepted for publication March 27, 2002. Address correspondence and reprint requests to Branko Perunovic, MD, Department of Cellular Pathology, Medical School, University of Birmingham, Vincent Drive, Birmingham B151TT, UK. Copyright 2002, Elsevier Science (USA). All rights reserved. 0046-8177/02/3307-0005$35.00/0 doi:10.1053/hupa.2002.125377
703
HUMAN PATHOLOGY
Volume 33, No. 7 (July 2002)
gene are among the earliest detectable alterations in many human astrocytomas, and abnormal accumulation of p53 can be demonstrated immunohistochemically in a high proportion of these tumors,1,25 we were interested in determining whether or not in astrocytic tumors, unlike in carcinomas of the lung, elevated mos levels are associated with p53 accumulation. MATERIALS AND METHODS Cases Studied We studied 62 cases of supratentorial astrocytic tumor diagnosed in the Department of Neuropathology, Frenchay Hospital, Bristol between 1988 and 2000. These tumors were from 38 males (61%) and 24 females (39%), ranging in age from 16 to 74 years (mean, 45 years) at diagnosis. The original slides were reviewed and the grading according to World Health Organization (WHO) criteria1 was confirmed in all cases. The cases consisted of 20 astrocytomas (WHO grade 2), 20 anaplastic astrocytomas (grade 3), and 22 glioblastomas (grade 4). Other types of astrocytic tumor (e.g., pilocytic astrocytoma, pleomorphic xanthoastrocytoma) and infratentorial tumors were not included in the study. Relevant clinical and follow-up data were obtained by examining the patients’ hospital and outpatient records, by contacting the patients’ general practitioners and, when necessary, by enquiry to the Local Health Authority. The study was approved by the Frenchay Hospital Research Ethics Committee. In most cases (55 of 62; 89%) the initial surgery was described as achieving complete or almost-complete macroscopic resection of the tumor, but in 7 cases (11%) the initial surgery was an incisional biopsy. Forty-three patients (69%) had received additional treatment: radiotherapy only (32 patients), radiotherapy in combination with chemotherapy (4 patients), radiotherapy followed by further surgery (3 patients), chemotherapy only (1 patient), or chemotherapy followed by further surgery (1 patient). The patients in the last 2 categories both had glioblastomas. For 1 patient with a grade 2 tumor and 1 patient with a grade 3 tumor, the only additional treatment was further surgical debulking of the lesion. In the 3 cases of astrocytoma and 2 cases of anaplastic astrocytoma for which further surgery was performed, histology of the resulting biopsies revealed a higher-grade tumor than had been present in the original biopsy. In 2 other cases of anaplastic astrocytoma and 1 case of astrocytoma, there was radiologic evidence of progression to higher grade in the form of prominent contrast enhancement, a feature that had not been present at the time of surgery. In the remaining 32 cases of grade 2 and grade 3 tumors, no unequivocal assessment of progression of tumor grade could be made. Three cases were lost to follow-up. At the time of data analysis, 42 patients (68%) had died; for the remaining 17 patients, the median duration of follow-up was 53 months (mean, 47 months; standard deviation, 32). The median survival was 22 months (mean, 32; standard deviation, 34) for the entire cohort. The clinical data are summarized in Table 1.
Immunohistochemistry Immunocytochemistry was performed on 5-m-thick sections cut from the paraffin blocks originally used for diagnostic purposes. Sections were dewaxed and rehydrated, then immersed for 30 minutes in methanol containing 0.9% hydrogen peroxide. Antigen retrieval was performed by micro-
TABLE 1. Clinicopathologic Data Tumors Characteristics Grade (WHO) Grade 2 Grade 3 Grade 4 Site Frontal lobe Temporal lobe Parietal lobe Occipital lobe Thalamus Treatment Type of surgery Macroscopic resection/debulking Biopsy Additional treatment No Yes Outcome Evidence of recurrence No Yes Progression No Yes Survival status Alive Dead
Number
20 20 22 30 17 12 1 2 55 7 19 43 45 17 57 5 17 42
waving the sections in 0.01 M sodium citrate buffer (pH 6.0), with the buffer being brought to boiling point twice over a 5-minute period. The sections were incubated overnight at room temperature in either anti-mos (P-19) goat polyclonal antiserum (Santa Cruz Biotechnology, Santa Cruz, CA) at 1:400 dilution or mouse monoclonal anti-p53 antibody (clone DO-7; Dako, Glostrup, Denmark) at 1:4000 dilution. Bound antibody was visualized by incubation with biotinylated secondary antibody (either Vectastain Universal Elite or Goat Vectastain Elite, as appropriate; Vector, CA) and avidin-biotin horseradish peroxidase, followed by reaction with 0.01% H2O2 and diaminobenzidine. We used 3 lung carcinoma specimens previously shown to express c-mos23 as positive controls for mos immunostaining and 1 astrocytoma specimen previously shown to have abnormal nuclear accumulation of p53 as a positive control for this antigen. Negative controls consisted of sections immunostained as above apart from the omission of primary antibody. We assessed the staining by first surveying the entire section under a ⫻10 objective, then counting labeled cells by examining at least 500 cells under a ⫻40 objective in the area where the density of labeled cells was greatest. Staining for mos was cytoplasmic, and staining for p53 was nuclear. A positive result was defined as staining of ⬎10% of the tumor cells for the relevant antigen (mos or p53).
Statistical Analysis We assessed mos expression in relation to sex, age, tumor grade, and p53 accumulation. For tumors of grades 2 and 3, we also examined the relationship between mos expression and subsequent evidence of progression of tumor grade. We analyzed the data using Pearson’s 2 test with Yates’ correction, or Fisher’s exact test when appropriate. Fifty-nine cases with follow-up data were subjected to Kaplan–Meier survival analysis. We used the log-rank test for comparisons of survival in different populations. Disease-related survival was defined
704
EXPRESSION OF mos IN ASTROCYTIC TUMORS (PERUNOVIC ET AL)
as the time interval between the date of the surgery and the date of death due to astrocytic tumor, whereas recurrencefree survival was defined as the time interval between the date of surgery and the date of clinical presentation with evidence of recurrence. SPSS for Windows R10 software was used for the statistical calculations. P values ⬍0.05 were regarded as significant.
RESULTS Immunohistochemical Findings Immunohistochemical expression of mos was found in 28 of the 62 cases (45%). The proportion of the positive cases varied with the tumor grade: 3 of the 20 (15%) grade 2 astrocytomas, 9 of the 20 (45%) grade 3 anaplastic astrocytomas, and 16 of the 22 (73%) glioblastomas. In most cases the staining had a distinct punctuate pattern, with single or multiple small cytoplasmic inclusions standing out clearly against the immunonegative surrounding cytoplasm (Fig 1). Only very occasional weak membranous or nuclear staining was seen and was disregarded for the purposes of scoring. In most cases the expression was patchy, and the density of positive cells varied in different parts of the tumor. With the exception noted below, immunolabeling was confined
FIGURE 2. Membranous staining for mos in cortical neurons adjacent to the tumor (arrows). (Anti-mos; original magnification ⫻200.)
to the tumor cells: non reactive astrocytes in adjacent brain tissue, endothelium, inflammatory cells, and most neurons were completely lacking in detectable mos. The exception was an intense cytoplasmic and membranous staining of scattered large cortical neurons in the neocortex adjacent to the tumor in 31 of the 42 cases (74%) in which adjacent brain tissue was present (Fig 2); the immunolabeled neurons appeared otherwise normal, without associated reactive or artefactual change. Thirty-six cases (58%) were immunopositive for p53. These included 11 (55%) grade 2 astrocytomas, 14 (70%) grade 3 anaplastic astrocytomas, and 11 (50%) glioblastomas. Relationship Between mos and Clinicopathologic Features Immunopositivity for mos correlated significantly (P ⬍ 0.01) with tumor grade, whether the tumors were subdivided according to WHO grade (i.e., as grades 2, 3, or 4) or simply into low-grade (grade 2) and highgrade (grades 3 and 4 combined) categories (Table 2). The proportion of tumors with mos expression was not significantly different in those grade 2 and 3 tumors that subsequently progressed to a higher grade than in those for which there was no evidence of progression. As illustrated in the survival plots in Fig 3, immunopositivity for mos did not significantly affect recurrence-free survival (log rank, 0.37, P ⫽ 0.54) or overall survival time (log rank, 0.91; P ⫽ 0.34). There was also no significant association between the presence of detectable mos and patient sex or age. Mos expression showed no significant association with p53 expression by the tumor cells. DISCUSSION
FIGURE 1. Single or multiple small punctate cytoplasmic inclusions, standing out clearly against the immunonegative surrounding cytoplasm. (Anti-mos; original magnification [A] ⫻200, [B] ⫻600.)
To the best of our knowledge, no previous studies have examined the expression of mos in astrocytic tumors. Apart from our previous investigation of lung
705
HUMAN PATHOLOGY
Volume 33, No. 7 (July 2002)
TABLE 2. Correlation of c-mos Expression with Clinicopathological Features
Age 0–55 years ⬎55 years Sex F M p53 expressuion Yes No No Tumor grade Grade 2 Grade 3 Grade 4 Tumour grade (adjusted) Grade 2 Grades 3 and 4 Tumor progression Yes No Tumor recurrence Yes No Survival status Alive Dead
mos (⫹)
mos (⫺)
2 P value
20 8
27 7
Not significant
14 14
10 24
Not significant
18 10 3
8 16 12
Not significant
3 9 16
17 11 6
0.01
3 25
17 17
0.01
9 24
3 4
Not significant
5 21
12 21
Not significant
7 19
10 23
Not significant
different mechanisms are involved in the genesis of astrocytic tumors and epithelial tumors of the lung. The significance of the strong immunostaining for mos in scattered large neocortical neurons in brain tissue adjacent to tumors is unclear. C-mos mRNA transcripts were previously detected in tissue extracts from mouse and monkey brains,12,32 and progressive neuronal degeneration was found in transgenic mice overexpressing c-mos.33 Recent studies have shown that neurons may respond to stress by abortive entry into the cell cycle.34,35 Large cortical neurons may be particularly vulnerable in this regard, and it is possible that the expression of c-mos is related to this cell cycle transition, to local hypoxia, or to products secreted by or in response to the neoplastic cells. The present study found a relationship between immunopositivity for mos and tumor grade, although the immunohistochemical findings did not correlate significantly with survival. These findings are in contrast with those in lung carcinomas, in which immunopositivity for mos was associated with significantly longer survival.23 The apparent discrepancy may reflect the modest size of our series or real differences in the pathogenesis of malignant transformation in these dif-
carcinomas,23,24 a review of literature has revealed only 5 studies of c-mos in human tumors. Stenman et al26 and Rommel et al27 analyzed the q11-12 region of chromosome 8 in pleomorphic adenomas of the salivary gland; in the former study, mutations of c-mos gene were detected in 2 of 23 tumors. Elevated levels of c-mos mRNA were found in 1 germ cell tumor and 3 carcinomas of the ovary28 and in a single case of medullary carcinoma of the thyroid.29 However, in a recent study of ovarian teratomas, de Foy et al30 did not identify mutations in the coding region of the c-mos gene. Finally, low levels of c-mos expression were detected in vitro in human cervical carcinoma-derived cell lines.13 In the present study, almost half of the astrocytic tumors that we examined were immunopositive for mos, and the proportion was much higher in glioblastomas. As in carcinomas of the lung,23 there was considerable heterogeneity in mos expression within tumors, probably reflecting clonal diversity rather than variation related to the stage in the cell cycle.5,13 The positive correlation between tumor grade and mos detection may be due to the expansion of more malignant mos-positive clones of tumor cells. As might be expected in relation to a protein kinase with cytoplasmic activity, immunostaining for mos was confined to the cytoplasm of the tumor cells. But this is in contrast to our previous study23 in which many cells showed nuclear immunopositivity. The combination of nuclear mos and frequent aneuploidy in mos-positive tumors24 was felt to be related to the previous observation that high levels of mos interfere with assembly of the mitotic spindle in somatic cells.31 The findings in our current study may reflect the fact that
FIGURE 3. Survival plots. (A) Mos expression and survival in astrocytic tumours (log rank, 0.37; P ⫽ 0.54). (B) Mos expression and recurrence in astrocytic tumors (log rank, 0.91; P ⫽ 0.34).
706
EXPRESSION OF mos IN ASTROCYTIC TUMORS (PERUNOVIC ET AL)
ferent types of tumors. Overexpression of c-mos in astrocytic tumors may be of primary pathogenetic importance, but could also simply be a reaction to disturbances of cell cycle regulation involving, for example, the cyclin D1–Cdk4 –pRb–E2F signaling pathway.3 The proportion of astrocytic tumors with abnormal p53 accumulation is in keeping with previous studies.1,25 It has been suggested that disturbances of p53 may be important in enabling c-mos–induced neoplastic transformation.21,23 However, we have found no evidence that this is the case for astrocytic tumors. In particular, mos was equally likely to be detectable in tumors with and tumors without abnormal nuclear accumulation of p53, irrespective of tumor grade. In conclusion, this study provides evidence that mos is involved in the neoplastic progression of a substantial proportion of astrocytic tumors. It remains to be determined whether or not mos has a causal role in this process and how other components of the MAPK pathway are involved. REFERENCES 1. Cavenee WK, Fumari FB, Nagane M, et al: Astrocytic tumours, in P Kleihues, WK Cavenee (eds): Pathology and Genetics of Tumours of the Nervous System. Lyon, France, IARC Press, 2000, pp 9-54 2. Kloetzer WS, Maxwell SA, Arlinghaus RB: P85gag-mos encoded by ts110 Moloney murine sarcoma virus has an associated protein kinase activity. Proc Natl Acad Sci U S A 80:412-416, 1983 3. Singh B, Arlinghaus RB: Mos and the cell cycle. Prog Cell Cycle Res 3:251-259, 1997 4. Sagata N: What does Mos do in oocytes and somatic cells? Bioessays 19:13-21, 1997 5. Yew N, Strobel M, Vande Woude GF: Mos and the cell cycle: The molecular basis of the transformed phenotype. Curr Opin Genet Dev 3:19-25, 1993 6. Sagata N, Oskarsson M, Copeland T, et al: Function of c-mos proto-oncogene product in meiotic maturation in Xenopus oocytes. Nature 335:519-525, 1988 7. O’Keefe SJ, Kiessling AA, Cooper GM: The c-mos gene product is required for cyclin B accumulation during meiosis of mouse eggs. Proc Natl Acad Sci U S A 88:7869-7872, 1991 8. Sagata N, Watanabe N, Vande Woude GF, et al: The c-mos proto-oncogene product is a cytostatic factor responsible for meiotic arrest in vertebrate eggs. Nature 342:512-518, 1989 9. Choi T, Fukasawa K, Zhou R, et al: The Mos/mitogen-activated protein kinase (MAPK) pathway regulates the size and degradation of the first polar body in maturing mouse oocytes. Proc Natl Acad Sci U S A 93:7032-7035, 1996 10. Colledge WH, Carlton MB, Udy GB, et al: Disruption of c-mos causes parthenogenetic development of unfertilized mouse eggs Nature 370:65-68, 1994 11. Hashimoto N, Watanabe N, Furuta Y, et al: Parthenogenetic activation of oocytes in c-mos-deficient mice. Nature 370:68-71, 1994 12. Propst F, Rosenberg MP, Iyer A, et al: C-mos proto-oncogene RNA transcripts in mouse tissues: Structural features, developmental regulation, and localization in specific cell types. Mol Cell Biol 7:1629-1637, 1987 13. Li CC, Chen E, O’Connell CD, et al: Detection of c-mos proto-oncogene expression in human cells. Oncogene 8:1685-1691, 1993 14. Singh B, Arlinghaus RB: Vimentin phosphorylation by
p37mos protein kinase in vitro and generation of a 50-kDa cleavage product in v-mos–transformed cells. Virology 173:144-156, 1989 15. Zhou RP, Oskarsson M, Paules RS, et al: Ability of the c-mos product to associate with and phosphorylate tubulin. Science 251:671675, 1991 16. Liu H, Vuyyuru VB, Pham CD, et al: Evidence of an interaction between Mos and Hsp70: A role of the Mos residue serine 3 in mediating Hsp70 association. Oncogene 18:3461-3470, 1999 17. Bai WL, Singh B, Karshin WL, et al: Phosphorylation of v-mos Ser 47 by the mitotic form of p34cdc2. Oncogene 6:1715-1723, 1991 18. Zhou R, Daar I, Ferris DK, et al: pp39mos is associated with p34cdc2 kinase in c-mosxe-transformed NIH 3T3 cells. Mol Cell Biol 12:3583-3589, 1992 19. Lenormand JL, Benayoun B, Guillier M, et al: Mos activates myogenic differentiation by promoting heterodimerization of MyoD and E12 proteins. Mol Cell Biol 17:584-593, 1997 20. Blair DG, McClements WL, Oskarsson MK, et al: Biological activity of cloned Moloney sarcoma virus DNA: Terminally redundant sequences may enhance transformation efficiency. Proc Natl Acad Sci U S A 77:3504-3508, 1980 21. Fukasawa K, Vande Woude GF: Synergy between the Mos/ mitogen-activated protein kinase pathway and loss of p53 function in transformation and chromosome instability. Mol Cell Biol 17:506518, 1997 22. Papkoff J, Verma IM, Hunter T: Detection of a transforming gene product in cells transformed by Moloney murine sarcoma virus. Cell 29:417-426, 1982 23. Athanasiou A, Gorgoulis VG, Zacharatos P, et al: c-mos immunoreactivity is an indicator of good prognosis in lung cancer. Histopathology 37:45-54, 2000 24. Gorgoulis VG, Zacharatos P, Mariatos G, et al: Deregulated expression of c-mos in non-small cell lung carcinomas: Relationship with p53 status, genomic instability, and tumor kinetics. Cancer Res 61:538-549, 2001 25. Watanabe K, Sato K, Biernat W, et al: Incidence and timing of p53 mutations during astrocytoma progression in patients with multiple biopsies. Clin Cancer Res 3:523-530, 1997 26. Stenman G, Sahlin P, Mark J, et al: Structural alterations of the c-mos locus in benign pleomorphic adenomas with chromosome abnormalities of 8q12. Oncogene 6:1105-1108, 1991 27. Rommel B, Bullerdiek J, Bartnitzke S, et al: No rearrangement of c-mos in salivary gland pleomorphic adenomas with 8q12 aberrations. Cancer Genet Cytogenet 49:165-169, 1990 28. Xerri L, Charpin C, Hassoun J, et al: Mos oncogene expression in human ovarian tumors. Anticancer Res 11:1629-1634, 1991 29. Parkar MH, Seid JM, Stringer BM, et al: Abnormal expression of the MOS proto-oncogene in human thyroid medullary carcinoma. Cancer Lett 43:185-189, 1988 30. de Foy KA, Gayther SA, Colledge WH, et al: Mutation analysis of the c-mos proto-oncogene in human ovarian teratomas. Br J Cancer 77:1642-1644, 1998 31. Wang XM, Yew N, Peloquin JG, et al: Mos oncogene product associates with kinetochores in mammalian somatic cells and disrupts mitotic progression. Proc Natl Acad Sci U S A 91:8329-8333, 1994. 32. Heikinheimo O, Dong KW, Lanzendorf SE, et al: Primate reproductive organs reveal a novel pattern of proto-oncogene c-mos and transcription factor Oct-3 mRNA expression. Mol Reprod Dev 42:397-406, 1995. 33. James ND, Davis DR, Sindon J, et al: Neurodegenerative changes including altered tau phosphorylation and neurofilament immunoreactivity in mice transgenic for the serine/threonine kinase Mos. Neurobiol Aging 17:235-241, 1996 34. Park DS, Obeidat A, Giovanni A, et al: Cell cycle regulators in neuronal death evoked by excitotoxic stress: implications for neurodegeneration and its treatment. neurobiol Aging 21:771-781, 1998 35. van Lookeren C, Gill R: Cell cycle-related gene expression in the adult rat brain: selective induction of cyclin G1 and p21WAF1/ CIP1 in neurons following focal cerebral ischemia. Neuroscience 84:1097-1112, 1998
707
plasmic immunostaining was observed in scattered large cortical neurons adjacent to tumors, possibly due to stress-induced abortive entry into the cell cycle. The correlation of mos immunopositivity with tumor grade may reflect the expansion of more malignant mospositive clones. This study provides evidence that mos may be involved in the neoplastic progression of a proportion of astrocytic tumors. HUM PATHOL 33:703-707. Copyright 2002, Elsevier Science (USA). All rights reserved. Key words: astrocytoma, astrocytic neoplasms, c-mos gene, immunohistochemistry, brain neoplasms, lymphokines, p53, proto-onko gene proteins mos. Abbreviations: MAPK, mitogen-activated protein kinase; WHO, World Health Organization.
During last decade, much has been learned about the genetic alterations that cause the initiation and progression of gliomas.1 These alterations include structural and functional abnormalities in a range growth factors, growth factor receptors and tumor suppressor genes, many of which are known to be involved in controlling progression through the cell cycle, cell death and cell differentiation. A pathway that is involved in the control of meiosis and probably mitosis but about which little information is available in relation to glial tumors is the mitogen-activated protein kinase (MAPK) transduction pathway and, in particular, the c-mos gene and its protein product, mos. C-mos was the first proto-oncogene of the serine/threonine protein kinase super-family to be discovered.2 In man it is located on chromosome 8q11-12 and encodes mos, a 39-kDa protein.3 All of the known biological roles of mos are mediated through its MAPK activity.3 Much c-mos–related research has concerned its role in meiosis, and several extensive reviews are available on this topic.3-5 Mos is an important regulator of meiotic maturation and is essential for both meiosis I and II, as has been shown in amphibian6 and mammalian7 experimental systems. It is a component of a large multiprotein complex, known as cytostatic factor, that is involved in metaphase II arrest of eggs in verte-
brates.8 There is also evidence of a permissive role in the organization of microtubules and assembly of meiotic spindle.3,9 C-mos knockout mice have severely reduced fertility, due to the premature release of the eggs from cell cycle arrest before their fertilization. The parthenogenetically activated germ cells that result may give rise to ovarian teratomas at an early age.10,11 The role of c-mos in somatic tissues is poorly understood. Mos or its coding mRNA appears to be expressed at low but unequal levels in most somatic tissues, although there is some disparity between the data on mRNA and protein levels.12,13 Mos has been found to be associated with several other proteins in both normal and transformed cells, including vimentin,14 tubulin,15 hsp70,16 p34cdc2,17 p35cdk,18 and myoD.19 Increased expression of c-mos in somatic cells has been linked to neoplastic transformation.20 Two mechanisms have been proposed to explain this: (1) the mos/MAPK pathway interferes with normal cyclin D1–Cdk4 –pRb– E2F signaling3 and (2) increased expression of c-mos leads to acquisition of a meiosis-like phenotype that may compromise the mitotic checkpoints that normally ensure the integrity of genetic material and symmetry of cell division during mitosis.21 High levels of mos have been reported to cause somatic cell death22 due to p53-mediated growth arrest and apoptosis.21 The authors of the latter study suggested that loss of p53 cell checkpoint activity may be a key permissive event in c-mos–induced neoplastic transformation. In previous studies we have shown that increased transcription of c-mos and elevated levels of immunostainable mos are demonstrable in a high proportion of human lung carcinomas23,24 in the absence of mutation or abnormal expression of p53. The aim of the present study was to investigate the expression of mos in astrocytic tumors. In addition, because mutations of the p53
From the Departments of Neuropathology and Neurosurgery, Frenchay Hospital, Bristol, United Kingdom, and the Department of Histology and Embryology, School of Medicine, University of Athens, Athens, Greece. Accepted for publication March 27, 2002. Address correspondence and reprint requests to Branko Perunovic, MD, Department of Cellular Pathology, Medical School, University of Birmingham, Vincent Drive, Birmingham B151TT, UK. Copyright 2002, Elsevier Science (USA). All rights reserved. 0046-8177/02/3307-0005$35.00/0 doi:10.1053/hupa.2002.125377
703
HUMAN PATHOLOGY
Volume 33, No. 7 (July 2002)
gene are among the earliest detectable alterations in many human astrocytomas, and abnormal accumulation of p53 can be demonstrated immunohistochemically in a high proportion of these tumors,1,25 we were interested in determining whether or not in astrocytic tumors, unlike in carcinomas of the lung, elevated mos levels are associated with p53 accumulation. MATERIALS AND METHODS Cases Studied We studied 62 cases of supratentorial astrocytic tumor diagnosed in the Department of Neuropathology, Frenchay Hospital, Bristol between 1988 and 2000. These tumors were from 38 males (61%) and 24 females (39%), ranging in age from 16 to 74 years (mean, 45 years) at diagnosis. The original slides were reviewed and the grading according to World Health Organization (WHO) criteria1 was confirmed in all cases. The cases consisted of 20 astrocytomas (WHO grade 2), 20 anaplastic astrocytomas (grade 3), and 22 glioblastomas (grade 4). Other types of astrocytic tumor (e.g., pilocytic astrocytoma, pleomorphic xanthoastrocytoma) and infratentorial tumors were not included in the study. Relevant clinical and follow-up data were obtained by examining the patients’ hospital and outpatient records, by contacting the patients’ general practitioners and, when necessary, by enquiry to the Local Health Authority. The study was approved by the Frenchay Hospital Research Ethics Committee. In most cases (55 of 62; 89%) the initial surgery was described as achieving complete or almost-complete macroscopic resection of the tumor, but in 7 cases (11%) the initial surgery was an incisional biopsy. Forty-three patients (69%) had received additional treatment: radiotherapy only (32 patients), radiotherapy in combination with chemotherapy (4 patients), radiotherapy followed by further surgery (3 patients), chemotherapy only (1 patient), or chemotherapy followed by further surgery (1 patient). The patients in the last 2 categories both had glioblastomas. For 1 patient with a grade 2 tumor and 1 patient with a grade 3 tumor, the only additional treatment was further surgical debulking of the lesion. In the 3 cases of astrocytoma and 2 cases of anaplastic astrocytoma for which further surgery was performed, histology of the resulting biopsies revealed a higher-grade tumor than had been present in the original biopsy. In 2 other cases of anaplastic astrocytoma and 1 case of astrocytoma, there was radiologic evidence of progression to higher grade in the form of prominent contrast enhancement, a feature that had not been present at the time of surgery. In the remaining 32 cases of grade 2 and grade 3 tumors, no unequivocal assessment of progression of tumor grade could be made. Three cases were lost to follow-up. At the time of data analysis, 42 patients (68%) had died; for the remaining 17 patients, the median duration of follow-up was 53 months (mean, 47 months; standard deviation, 32). The median survival was 22 months (mean, 32; standard deviation, 34) for the entire cohort. The clinical data are summarized in Table 1.
Immunohistochemistry Immunocytochemistry was performed on 5-m-thick sections cut from the paraffin blocks originally used for diagnostic purposes. Sections were dewaxed and rehydrated, then immersed for 30 minutes in methanol containing 0.9% hydrogen peroxide. Antigen retrieval was performed by micro-
TABLE 1. Clinicopathologic Data Tumors Characteristics Grade (WHO) Grade 2 Grade 3 Grade 4 Site Frontal lobe Temporal lobe Parietal lobe Occipital lobe Thalamus Treatment Type of surgery Macroscopic resection/debulking Biopsy Additional treatment No Yes Outcome Evidence of recurrence No Yes Progression No Yes Survival status Alive Dead
Number
20 20 22 30 17 12 1 2 55 7 19 43 45 17 57 5 17 42
waving the sections in 0.01 M sodium citrate buffer (pH 6.0), with the buffer being brought to boiling point twice over a 5-minute period. The sections were incubated overnight at room temperature in either anti-mos (P-19) goat polyclonal antiserum (Santa Cruz Biotechnology, Santa Cruz, CA) at 1:400 dilution or mouse monoclonal anti-p53 antibody (clone DO-7; Dako, Glostrup, Denmark) at 1:4000 dilution. Bound antibody was visualized by incubation with biotinylated secondary antibody (either Vectastain Universal Elite or Goat Vectastain Elite, as appropriate; Vector, CA) and avidin-biotin horseradish peroxidase, followed by reaction with 0.01% H2O2 and diaminobenzidine. We used 3 lung carcinoma specimens previously shown to express c-mos23 as positive controls for mos immunostaining and 1 astrocytoma specimen previously shown to have abnormal nuclear accumulation of p53 as a positive control for this antigen. Negative controls consisted of sections immunostained as above apart from the omission of primary antibody. We assessed the staining by first surveying the entire section under a ⫻10 objective, then counting labeled cells by examining at least 500 cells under a ⫻40 objective in the area where the density of labeled cells was greatest. Staining for mos was cytoplasmic, and staining for p53 was nuclear. A positive result was defined as staining of ⬎10% of the tumor cells for the relevant antigen (mos or p53).
Statistical Analysis We assessed mos expression in relation to sex, age, tumor grade, and p53 accumulation. For tumors of grades 2 and 3, we also examined the relationship between mos expression and subsequent evidence of progression of tumor grade. We analyzed the data using Pearson’s 2 test with Yates’ correction, or Fisher’s exact test when appropriate. Fifty-nine cases with follow-up data were subjected to Kaplan–Meier survival analysis. We used the log-rank test for comparisons of survival in different populations. Disease-related survival was defined
704
EXPRESSION OF mos IN ASTROCYTIC TUMORS (PERUNOVIC ET AL)
as the time interval between the date of the surgery and the date of death due to astrocytic tumor, whereas recurrencefree survival was defined as the time interval between the date of surgery and the date of clinical presentation with evidence of recurrence. SPSS for Windows R10 software was used for the statistical calculations. P values ⬍0.05 were regarded as significant.
RESULTS Immunohistochemical Findings Immunohistochemical expression of mos was found in 28 of the 62 cases (45%). The proportion of the positive cases varied with the tumor grade: 3 of the 20 (15%) grade 2 astrocytomas, 9 of the 20 (45%) grade 3 anaplastic astrocytomas, and 16 of the 22 (73%) glioblastomas. In most cases the staining had a distinct punctuate pattern, with single or multiple small cytoplasmic inclusions standing out clearly against the immunonegative surrounding cytoplasm (Fig 1). Only very occasional weak membranous or nuclear staining was seen and was disregarded for the purposes of scoring. In most cases the expression was patchy, and the density of positive cells varied in different parts of the tumor. With the exception noted below, immunolabeling was confined
FIGURE 2. Membranous staining for mos in cortical neurons adjacent to the tumor (arrows). (Anti-mos; original magnification ⫻200.)
to the tumor cells: non reactive astrocytes in adjacent brain tissue, endothelium, inflammatory cells, and most neurons were completely lacking in detectable mos. The exception was an intense cytoplasmic and membranous staining of scattered large cortical neurons in the neocortex adjacent to the tumor in 31 of the 42 cases (74%) in which adjacent brain tissue was present (Fig 2); the immunolabeled neurons appeared otherwise normal, without associated reactive or artefactual change. Thirty-six cases (58%) were immunopositive for p53. These included 11 (55%) grade 2 astrocytomas, 14 (70%) grade 3 anaplastic astrocytomas, and 11 (50%) glioblastomas. Relationship Between mos and Clinicopathologic Features Immunopositivity for mos correlated significantly (P ⬍ 0.01) with tumor grade, whether the tumors were subdivided according to WHO grade (i.e., as grades 2, 3, or 4) or simply into low-grade (grade 2) and highgrade (grades 3 and 4 combined) categories (Table 2). The proportion of tumors with mos expression was not significantly different in those grade 2 and 3 tumors that subsequently progressed to a higher grade than in those for which there was no evidence of progression. As illustrated in the survival plots in Fig 3, immunopositivity for mos did not significantly affect recurrence-free survival (log rank, 0.37, P ⫽ 0.54) or overall survival time (log rank, 0.91; P ⫽ 0.34). There was also no significant association between the presence of detectable mos and patient sex or age. Mos expression showed no significant association with p53 expression by the tumor cells. DISCUSSION
FIGURE 1. Single or multiple small punctate cytoplasmic inclusions, standing out clearly against the immunonegative surrounding cytoplasm. (Anti-mos; original magnification [A] ⫻200, [B] ⫻600.)
To the best of our knowledge, no previous studies have examined the expression of mos in astrocytic tumors. Apart from our previous investigation of lung
705
HUMAN PATHOLOGY
Volume 33, No. 7 (July 2002)
TABLE 2. Correlation of c-mos Expression with Clinicopathological Features
Age 0–55 years ⬎55 years Sex F M p53 expressuion Yes No No Tumor grade Grade 2 Grade 3 Grade 4 Tumour grade (adjusted) Grade 2 Grades 3 and 4 Tumor progression Yes No Tumor recurrence Yes No Survival status Alive Dead
mos (⫹)
mos (⫺)
2 P value
20 8
27 7
Not significant
14 14
10 24
Not significant
18 10 3
8 16 12
Not significant
3 9 16
17 11 6
0.01
3 25
17 17
0.01
9 24
3 4
Not significant
5 21
12 21
Not significant
7 19
10 23
Not significant
different mechanisms are involved in the genesis of astrocytic tumors and epithelial tumors of the lung. The significance of the strong immunostaining for mos in scattered large neocortical neurons in brain tissue adjacent to tumors is unclear. C-mos mRNA transcripts were previously detected in tissue extracts from mouse and monkey brains,12,32 and progressive neuronal degeneration was found in transgenic mice overexpressing c-mos.33 Recent studies have shown that neurons may respond to stress by abortive entry into the cell cycle.34,35 Large cortical neurons may be particularly vulnerable in this regard, and it is possible that the expression of c-mos is related to this cell cycle transition, to local hypoxia, or to products secreted by or in response to the neoplastic cells. The present study found a relationship between immunopositivity for mos and tumor grade, although the immunohistochemical findings did not correlate significantly with survival. These findings are in contrast with those in lung carcinomas, in which immunopositivity for mos was associated with significantly longer survival.23 The apparent discrepancy may reflect the modest size of our series or real differences in the pathogenesis of malignant transformation in these dif-
carcinomas,23,24 a review of literature has revealed only 5 studies of c-mos in human tumors. Stenman et al26 and Rommel et al27 analyzed the q11-12 region of chromosome 8 in pleomorphic adenomas of the salivary gland; in the former study, mutations of c-mos gene were detected in 2 of 23 tumors. Elevated levels of c-mos mRNA were found in 1 germ cell tumor and 3 carcinomas of the ovary28 and in a single case of medullary carcinoma of the thyroid.29 However, in a recent study of ovarian teratomas, de Foy et al30 did not identify mutations in the coding region of the c-mos gene. Finally, low levels of c-mos expression were detected in vitro in human cervical carcinoma-derived cell lines.13 In the present study, almost half of the astrocytic tumors that we examined were immunopositive for mos, and the proportion was much higher in glioblastomas. As in carcinomas of the lung,23 there was considerable heterogeneity in mos expression within tumors, probably reflecting clonal diversity rather than variation related to the stage in the cell cycle.5,13 The positive correlation between tumor grade and mos detection may be due to the expansion of more malignant mos-positive clones of tumor cells. As might be expected in relation to a protein kinase with cytoplasmic activity, immunostaining for mos was confined to the cytoplasm of the tumor cells. But this is in contrast to our previous study23 in which many cells showed nuclear immunopositivity. The combination of nuclear mos and frequent aneuploidy in mos-positive tumors24 was felt to be related to the previous observation that high levels of mos interfere with assembly of the mitotic spindle in somatic cells.31 The findings in our current study may reflect the fact that
FIGURE 3. Survival plots. (A) Mos expression and survival in astrocytic tumours (log rank, 0.37; P ⫽ 0.54). (B) Mos expression and recurrence in astrocytic tumors (log rank, 0.91; P ⫽ 0.34).
706
EXPRESSION OF mos IN ASTROCYTIC TUMORS (PERUNOVIC ET AL)
ferent types of tumors. Overexpression of c-mos in astrocytic tumors may be of primary pathogenetic importance, but could also simply be a reaction to disturbances of cell cycle regulation involving, for example, the cyclin D1–Cdk4 –pRb–E2F signaling pathway.3 The proportion of astrocytic tumors with abnormal p53 accumulation is in keeping with previous studies.1,25 It has been suggested that disturbances of p53 may be important in enabling c-mos–induced neoplastic transformation.21,23 However, we have found no evidence that this is the case for astrocytic tumors. In particular, mos was equally likely to be detectable in tumors with and tumors without abnormal nuclear accumulation of p53, irrespective of tumor grade. In conclusion, this study provides evidence that mos is involved in the neoplastic progression of a substantial proportion of astrocytic tumors. It remains to be determined whether or not mos has a causal role in this process and how other components of the MAPK pathway are involved. REFERENCES 1. Cavenee WK, Fumari FB, Nagane M, et al: Astrocytic tumours, in P Kleihues, WK Cavenee (eds): Pathology and Genetics of Tumours of the Nervous System. Lyon, France, IARC Press, 2000, pp 9-54 2. Kloetzer WS, Maxwell SA, Arlinghaus RB: P85gag-mos encoded by ts110 Moloney murine sarcoma virus has an associated protein kinase activity. Proc Natl Acad Sci U S A 80:412-416, 1983 3. Singh B, Arlinghaus RB: Mos and the cell cycle. Prog Cell Cycle Res 3:251-259, 1997 4. Sagata N: What does Mos do in oocytes and somatic cells? Bioessays 19:13-21, 1997 5. Yew N, Strobel M, Vande Woude GF: Mos and the cell cycle: The molecular basis of the transformed phenotype. Curr Opin Genet Dev 3:19-25, 1993 6. Sagata N, Oskarsson M, Copeland T, et al: Function of c-mos proto-oncogene product in meiotic maturation in Xenopus oocytes. Nature 335:519-525, 1988 7. O’Keefe SJ, Kiessling AA, Cooper GM: The c-mos gene product is required for cyclin B accumulation during meiosis of mouse eggs. Proc Natl Acad Sci U S A 88:7869-7872, 1991 8. Sagata N, Watanabe N, Vande Woude GF, et al: The c-mos proto-oncogene product is a cytostatic factor responsible for meiotic arrest in vertebrate eggs. Nature 342:512-518, 1989 9. Choi T, Fukasawa K, Zhou R, et al: The Mos/mitogen-activated protein kinase (MAPK) pathway regulates the size and degradation of the first polar body in maturing mouse oocytes. Proc Natl Acad Sci U S A 93:7032-7035, 1996 10. Colledge WH, Carlton MB, Udy GB, et al: Disruption of c-mos causes parthenogenetic development of unfertilized mouse eggs Nature 370:65-68, 1994 11. Hashimoto N, Watanabe N, Furuta Y, et al: Parthenogenetic activation of oocytes in c-mos-deficient mice. Nature 370:68-71, 1994 12. Propst F, Rosenberg MP, Iyer A, et al: C-mos proto-oncogene RNA transcripts in mouse tissues: Structural features, developmental regulation, and localization in specific cell types. Mol Cell Biol 7:1629-1637, 1987 13. Li CC, Chen E, O’Connell CD, et al: Detection of c-mos proto-oncogene expression in human cells. Oncogene 8:1685-1691, 1993 14. Singh B, Arlinghaus RB: Vimentin phosphorylation by
p37mos protein kinase in vitro and generation of a 50-kDa cleavage product in v-mos–transformed cells. Virology 173:144-156, 1989 15. Zhou RP, Oskarsson M, Paules RS, et al: Ability of the c-mos product to associate with and phosphorylate tubulin. Science 251:671675, 1991 16. Liu H, Vuyyuru VB, Pham CD, et al: Evidence of an interaction between Mos and Hsp70: A role of the Mos residue serine 3 in mediating Hsp70 association. Oncogene 18:3461-3470, 1999 17. Bai WL, Singh B, Karshin WL, et al: Phosphorylation of v-mos Ser 47 by the mitotic form of p34cdc2. Oncogene 6:1715-1723, 1991 18. Zhou R, Daar I, Ferris DK, et al: pp39mos is associated with p34cdc2 kinase in c-mosxe-transformed NIH 3T3 cells. Mol Cell Biol 12:3583-3589, 1992 19. Lenormand JL, Benayoun B, Guillier M, et al: Mos activates myogenic differentiation by promoting heterodimerization of MyoD and E12 proteins. Mol Cell Biol 17:584-593, 1997 20. Blair DG, McClements WL, Oskarsson MK, et al: Biological activity of cloned Moloney sarcoma virus DNA: Terminally redundant sequences may enhance transformation efficiency. Proc Natl Acad Sci U S A 77:3504-3508, 1980 21. Fukasawa K, Vande Woude GF: Synergy between the Mos/ mitogen-activated protein kinase pathway and loss of p53 function in transformation and chromosome instability. Mol Cell Biol 17:506518, 1997 22. Papkoff J, Verma IM, Hunter T: Detection of a transforming gene product in cells transformed by Moloney murine sarcoma virus. Cell 29:417-426, 1982 23. Athanasiou A, Gorgoulis VG, Zacharatos P, et al: c-mos immunoreactivity is an indicator of good prognosis in lung cancer. Histopathology 37:45-54, 2000 24. Gorgoulis VG, Zacharatos P, Mariatos G, et al: Deregulated expression of c-mos in non-small cell lung carcinomas: Relationship with p53 status, genomic instability, and tumor kinetics. Cancer Res 61:538-549, 2001 25. Watanabe K, Sato K, Biernat W, et al: Incidence and timing of p53 mutations during astrocytoma progression in patients with multiple biopsies. Clin Cancer Res 3:523-530, 1997 26. Stenman G, Sahlin P, Mark J, et al: Structural alterations of the c-mos locus in benign pleomorphic adenomas with chromosome abnormalities of 8q12. Oncogene 6:1105-1108, 1991 27. Rommel B, Bullerdiek J, Bartnitzke S, et al: No rearrangement of c-mos in salivary gland pleomorphic adenomas with 8q12 aberrations. Cancer Genet Cytogenet 49:165-169, 1990 28. Xerri L, Charpin C, Hassoun J, et al: Mos oncogene expression in human ovarian tumors. Anticancer Res 11:1629-1634, 1991 29. Parkar MH, Seid JM, Stringer BM, et al: Abnormal expression of the MOS proto-oncogene in human thyroid medullary carcinoma. Cancer Lett 43:185-189, 1988 30. de Foy KA, Gayther SA, Colledge WH, et al: Mutation analysis of the c-mos proto-oncogene in human ovarian teratomas. Br J Cancer 77:1642-1644, 1998 31. Wang XM, Yew N, Peloquin JG, et al: Mos oncogene product associates with kinetochores in mammalian somatic cells and disrupts mitotic progression. Proc Natl Acad Sci U S A 91:8329-8333, 1994. 32. Heikinheimo O, Dong KW, Lanzendorf SE, et al: Primate reproductive organs reveal a novel pattern of proto-oncogene c-mos and transcription factor Oct-3 mRNA expression. Mol Reprod Dev 42:397-406, 1995. 33. James ND, Davis DR, Sindon J, et al: Neurodegenerative changes including altered tau phosphorylation and neurofilament immunoreactivity in mice transgenic for the serine/threonine kinase Mos. Neurobiol Aging 17:235-241, 1996 34. Park DS, Obeidat A, Giovanni A, et al: Cell cycle regulators in neuronal death evoked by excitotoxic stress: implications for neurodegeneration and its treatment. neurobiol Aging 21:771-781, 1998 35. van Lookeren C, Gill R: Cell cycle-related gene expression in the adult rat brain: selective induction of cyclin G1 and p21WAF1/ CIP1 in neurons following focal cerebral ischemia. Neuroscience 84:1097-1112, 1998
707