Search

Sixty-one glioblastomas have been studied, subdivided into the categories of classic glioblastomas (GBM) and glioblastomas with astrocytic (GBA) and oligodendroglial (GBO) differentiated areas. On surgical samples, TP53, Mdm2, CDKN2A/p16-p14 alterations were studied by molecular biology techniques and by immunohistochemistry. It has been found that Mdm2 amplification was more frequent in GBM than in GBA and GBO, that p14ARF was inactivated in a high percentage of cases in the three tumor categories. Both these and other alterations did not reach a statistical significance, with the exception of CDKN2A/p16 homozygous deletion which showed the highest frequency in GBO. The latter finding could be in line with the observation that CDKN2A/p16 inactivation is a step in the molecular pathway to tumor progression in oligodendrogliomas. TP53 mutations and Mdm2 amplifications were mutually exclusive, whereas TP53 mutations and CDKN2A/p14 inactivation coexisted in 5 cases. The alterations of the p53/Mdm2/p14ARF pathway occurred in 73% of cases and in 80% of cases if CDKN2A homozygous deletions were associated. All glioblastomas with gemistocytic areas showed p14ARF inactivation. Immunohistochemistry showed higher percentages of positivity in comparison with molecular genetics, but with similar variations.

The F-box protein Skp2 regulates G1-S transition by controlling p27/Kip.1. The deregulated expression of p27/Kip.1 plays a critical role in the pathogenesis of many human tumors. Its cellular levels depend on ubiquitin-mediated degradation. Recently, Skp2 has been demonstrated to mediate p27/Kip.1 degradation and to have oncogenic properties. In a series of astrocytic gliomas, immunohistochemistry and Western blot of p27/Kip.1 and Skp2 have been compared. p27/Kip.1 decreased with anaplasia and almost disappeared in glioblastomas (GBM), whereas Skp2 was absent or poorly expressed in well differentiated astrocytomas and it was diffusely or focally expressed in most GBM. Since the expression of Skp2 increases during G1-S transition, the correlation of Skp2 levels with malignancy might simply reflect the highest percentage of proliferating cells in anaplastic gliomas or alternatively be instrumental to p27/Kip.1 degradation.

Sixteen cases of ependymoma were studied for CDKN2A/p16 inactivation by immunohistochemistry using a p16 monoclonal antibody, by homozygous deletion (HD) assay and 5'CpG promoter methylation assay (methylation-specific PCR). Three out of 16 cases were p16 immuno-negative: two corresponded to grade II ependymomas and one to grade III. The latter ependymoma, characterized by a high Ki-67/MIB-1 LI, was the only one of the whole series to show CDKN2A HD. No promoter methylation was found in the two immuno-negative cases without CDKN2A HD. Alternative mechanisms, such as point mutations or alterations in p16 post-translational regulation, may be responsible for p16 inactivation. Since in our series just one out of eight anaplastic cases showed negative immunostaining and CDKN2A HD, p16/CDKN2A inactivation may not play an important role in the malignant transformation of ependymomas. Amplification of CCNDI and CDK4, p27/Kipl degradation and TP53 mutations were previously studied by other authors and were demonstrated not to correlate with anaplasia. Up to date, molecular genetic studies have not been useful in recognizing the anaplastic variant in ependymomas.

In brain tumors, the prognostic value of apoptosis is still debated, even though recent observations are rather negative. In 50 astrocytic gliomas, apoptotic nuclei were recognized by TUNEL technique and morphology and apoptotic index (AI), mitotic index (MI) and the MI/AI ratio were calculated in proliferative areas and in areas containing perinecrotic palisadings and hypercellular centres. In proliferative areas, a positive linear correlation between MI and AI and a MI/AI ratio > I were found. The latter was < I in perinecrotic palisadings and hypercellular centres. This suggests the possibility that apoptosis is triggered respectively by cell proliferation and hypoxia. A small number of apoptotic nuclei was positive for c-Jun and JNKI, suggesting the involvement of this pathway in apoptosis as well as that of sphingomyelinase/ceramide. No relationship was found between AI and labeling indices of Bcl-2, Bcl-x, Bax, p53, APO.I/Fas. The rationale for a prognostic value of apoptosis in gliomas is thus poor.

Activated caspase-3has been immunohistochemically studied in 30glioblastomas. Its distribution has been compared with that of apoptotic nuclei demonstrated by terminal dUTP nick-end labeling (TUNEL) and morphology. The best procedure for the demonstration of caspase-3 requires formalin fixation, followed by Carnoy fixation, with microwave irradiation. The number of positive cells is lower than that of apoptotic nuclei shown by TUNEL technique, especially in perinecrotic pseudo-palisadings, and there are also qualitative variations. Positive staining occurs in nuclei, cytoplasms or in both cell compartments. The interpretation of Caspase-3 positive staining is based on its crucial position in the final pathway to apoptosis and on the mechanisms by which it cleaves cytoplasmic and nuclear proteins among which inhibitory/caspase-activated DNase system is included.

Somatostatin and other neuropeptides are expressed in tumors originating from neuronal precursors and paraganglia, namely medulloblastoma, central Primitive Neuro-Ectodermal Tumors (cPNETs), neurocytoma, gangliocytoma. olfactory neuroblastoma, paraganglioma. In medulloblastoma, the most common malignant tumor in childhood, there is an extensive expression of somatostatin in addition to somatostatin receptors (SSTR) type 2. Although density of SSTR-2 and intensity of expression of somatostatin genes have no prognostic significance in medulloblastoma. their presence may bring along important information on oncogenesis and relate medulloblastoma to cPNETs. Radio-labeled octreotide scintigraphy may be useful in the follow-up of these patients. allowing differentiation between scar and tumoral tissue. Moreover, on the basis of octreotide-induced inhibition of cell proliferation in medulloblastoma, a trial with octreotide in patients with recurrent or high-risk tumor is warranted. Meningiomas and low-grade astrocytic gliomas, even if not displaying a clear neuroendocrine phenotype, have high levels of SSTR-2. In meningiomas, SSTRs-scintigraphy is not part of the routine pre-operative assessment; moreover, a therapeutic trial with somatostatin-analogues in patients with recurrent or inoperable meningiomas should be carried-out with great caution, because somatostatin and octreotide slightly increase cell proliferation in cultured meningiomatous cells. Low-grade gliomas (WHO grade 2), and a smaller fraction of anaplastic astrocytomas, express SSTR-2, while glioblastomas usually do not. Unfortunately, radiolabeled-octreotide scintigraphy is not useful in the differential diagnosis of gliomas, because the results are altered by the disruption of the blood brain barrier (BBB); in addition, radionuclide-labeled somatostatin analogues are not useful in the therapy of low-grade gliomas, because the intact BBB prevents them from reaching the target SSTR-2. Recently, a pilot study in gliomas, has proposed the use of a radio-labeled somatostostatin analogue with a loco-regional approach in order to overcome the intact BBB.

The cell-cycle regulator p16 inhibits the complex cdk4-cyclin D1 and controls G1-S transition. In human tumors, p16 inactivation is often accomplished by homozygous deletion (HD) of its encoding gene, CDKN2A. Methylation of the 5' CpG island promoter has been proposed as an alternative mechanism of inactivation. Expression of p16, CDKN2A HD and 5' CDKN2A CpG island methylation was studied in 25 oligodendrogliomas by immunohistochemistry and PCR amplification. Ten oligodendrogliomas were p16-immunonegative, and CDKN2A HD was determined in 8 of these cases. In the 2 immunonegative cases without HD, no CpG island methylation was found. The absence of CpG island methylation in the p16-immunonegative cases without HD suggests either non-genetic regulation of p16 or different genetic changes. CDKN2A HD did not correlate with histological grading (p = n.s.); however, it showed a correlation with survival (p = 0.03), supporting an important role of CDKN2A in the prognosis of oligodendrogliomas., (Copyright 2000 Wiley-Liss, Inc.)

p27/kip1 regulates the G1-S transition of the cell cycle by inhibiting cyclinD-CDK4, cyclinE-CDK2 and cyclinA-CDK2 complexes. Regulation of p27 levels occurs mainly post-translationally by ubiquitin-mediated proteasomal proteolysis. Although genetic changes of p27/kip1 are extremely rare, in many human carcinomas p27 levels are reduced, correlate with histological malignancy, and are associated with poor prognosis. In astrocytic gliomas, p27 decreases with anaplasia and is almost absent in glioblastomas. p27/kip1 was immunohistochemically studied in 37 oligodendrogliomas, categorized according to WHO classification. In this series, the immunohistochemical reaction for p27 was confined to nuclei. p27 score showed a tendency to decrease with malignancy. When the p27 score was considered as high versus low expression (cut-off of p27 labeling index, LI, at 25%), it represented an independent prognostic factor in univariate (P = 0.02) and in multivariate analysis (P = 0.04). The risk ratio suggested that the p27 low expression group had a threefold increased possibility to show a reduced survival. Moreover, p27 levels did not correlate with MIB-1 LI, suggesting that p27 is not merely associated with the control of proliferation.

p27/kip1 regulates the G1-S transition of the cell cycle by inhibiting cyclin D-CDK4, cyclin E-CDK2, and cyclin A-CDK2. Modulation of p27 cellular abundance occurs mainly at post-translational level by the ubiquitin-proteasome proteolysis. Although rearrangements and mutations of p27/kip1 are extremely rare events, p27 levels are reduced and associated with a poor prognosis in many human carcinomas. In astrocytic tumors, p27 decreases with advancing anaplasia and is almost absent in glioblastomas. To verify whether the degradation of p27 protein was responsible for its reduced levels in malignant gliomas, p27 degradation activity was tested in 22 tissue extracts that represented high, low, and absent p27 protein levels. p27 protein expression was detected by immunohistochemistry and immunoblot analysis and comparable results between the 2 methods were obtained. Low or undetectable p27 degradation activity was found in samples that displayed high levels of p27, i.e. all 4 normal brain biopsies, and 4 out of 6 grade II astrocytomas. Enhanced degradation activity resulted in malignant gliomas with low or absent p27 protein levels. The proteasome inhibitor LLnL abolished p27 degradation, demonstrating that it occurs in a proteasome-dependent manner. These data suggest that proteasome degradation of p27 may be instrumental in the deregulation of the cell cycle and to the malignant transformation of gliomas.

Cyclin D1 (cycD1) expression was defined immunohistochemically using monoclonal antibody DCS-6 and polyclonal antiserum H-295 in 50 glioma biopsies. The number of positive nuclei was higher for H-295 than for DCS-6, with a ratio of 3:1. The labelling index (LI) was compared to the grade of histological malignancy and to Ki-67 MIB-1 LI. The LI for cycD1 increased with histological malignancy, in parallel with the increase in MIB-1 LI. In most tumours, the maximum LI for cycD1 and MIB-1 were found in the same areas. The mean MIB-1 LI: mean cycD1 LI ratio does not vary in the three grades of astrocytic tumours. However, in this study the correlation between the two LIs was not statistically significant. Staining for cycD1 antigen does not necessarily imply that the gene is overexpressed since other molecular mechanisms can also be responsible for cell cycle deregulation. In invasive areas, the cycD1 LI is frequently higher than in solid tumour, either because more tumour cells are positive or because reactive astrocytes and activated microglia express cycD1. The relative contribution of neoplastic and reactive cells remains to be defined.

CDKN2/p16 inactivation is the most frequent alteration in the molecular regulation of G1-S transition. CDKN2/p16 homozygous deletions was studied in paraffin-embedded sections of 45 astrocytic tumours by multiplex PCR. Immunohistochemistry for p16 and proliferation marker Ki-67 MIB-1 was performed in adjacent sections; their labelling index (LI) have been calculated. CDKN2/p16 gene was not deleted in astrocytomas, while homozygous deletion was found in 26.7% anaplastic astrocytomas, and in 55.0% of glioblastomas. Analysis of CDKN2/p16 homozygous deletion in discrete areas of the same tumour, showed that the deletion occurred independently of the phenotypic aspect of the areas. Nevertheless a genotypic and phenotypic heterogeneity is present in few cases. p16 immunohistochemistry mostly corresponds to the genotypic pattern. No correlation was found between CDKN2/p16 homozygous deletion and MIB-1 LI.

Background: A reliable marker for tumor oligodendroglial cells is not yet available, so that the histological recognition of the tumor still encounters uncertainties. There is no general agreement also on prognostic factors in oligodendroglioma. The inconsistency concerns mainly the histopathological factors. The aim of the study was recognition of prognostic factors in oligodendroglioma., Methods: In a series of ninety-eight oligodendrogliomas, including twenty mixed oligoastrocytomas, clinical [sex, age at surgery, tumor location, symptoms at presentation], therapeutic [extent of resection, year of surgery, post-operative Karnofsky score, post-operative radiotherapy, post-operative chemotherapy], histological [cell density, nuclear pleomorphism, vascular endothelial proliferation, necrosis, microcysts, mitoses, mitotic index (MI), apoptosis, apoptotic index (AI)] and immunohistochemical parameters [MIB-1 and PCNA Labeling Indexes (LIs), staining for GFAP, positivity for p53] were correlated with survival in uni- and multivariate analysis in order to identify their prognostic significance., Results: Age at surgery, extent of surgical resection, year of surgery, post-operative Karnofsky score and MIB-1 LI were associated with survival in both uni- and multivariate analysis. Location, symptoms at presentation, mitoses, MI, AI, and PCNA LI showed a significant correlation with survival in uni- but not in multivariate analysis. The twenty cases of oligoastrocytomas did not show any difference in survival from pure oligodendrogliomas., Conclusions: Some clinical and therapeutic factors together with MIB-1 LI play a prognostic role. MIB-1 LI is prognostic with a cutoff of 8%. Histology gives a limited contribution to the prognosis. Oligoastrocytomas had the same outcome and prognostic factors as pure oligodendrogliomas.

Cell proliferation is characteristic, though non specific, of tumors. The cell cycle is regarded as a clock with ordered activation of protein complexes triggering initiation and advancement through checkpoints. The cell cycle and its control mechanisms are briefly described herein focusing on neuroepithelial tumors. The assessment of cell proliferation in brain tumors is a very important tool in diagnosis and especially in prognosis. It can be performed by different methods: counting mitoses, calculating the labeling index (Ll) of [3H] Thymidine, BrdU, Ki-67, MIB.1, PCNA, cytometry and AgNORs. Each method has its advantages and disadvantages. The main obstacles to the usefulness of different Lls in establishing prognosis in individual cases are sampling error, heterogeneity of the proliferation potential of brains tumors and the wide overlapping of Ll ranges between classic and anaplatic variants. The principal findings are critically described and commented upon.

Well-differentiated astrocytomas may transform into malignant astrocytomas in time. In surgical specimens, when the histological picture strictly corresponds to that of grade II glioma, the transformation is unpredictable. Clinically, the bad outcome of a quota of astrocytomas is a well known phenomenon. The use of proliferation markers, and recently of MIB-1 LI, for detecting the proliferation potential comes out to be a useful tool for prognosis. A survival analysis of fifty astrocytomas grade II according to the WHO classification was performed with univariate and multivariate analysis of a series of clinical and histological parameters. MIB-1 LI was calculated and compared with all the other parameters. A cut-off of 8% of MIB-1 LI divided the astrocytomas in two groups with significantly different survival (p = 0.0066): median survival time of 1062 versus 1686 days. According to multivariate analysis MIB-1 LI resulted to be an independent factor (p = 0.002) along with extension of surgical removal (partial versus total), postoperative Karnofsky status (> or = 70 versus < 70) and age (< or = 30 versus > 30). The interpretation of well-differentiated astrocytomas with high MIB-1 LI is that the increasing number of cycling cells precedes phenotypic transformation. MIB-1 LI can be used as a prognostic factor.

Prognostic factors in oligodendrogliomas are not well defined, even considering the labeling index of proliferation markers. As in other neuroepithelial tumors, the difficulty in calculating cell loss may contribute to this uncertainty. Proliferation markers Ki-67/MIB.1 and PCNA, mitoses, apoptotic nuclei, p53 and bcl-2 expression were investigated in 98 oligodendrogliomas. Apoptosis was assessed by the aspect of nuclei, by in situ end-labeling (ISEL) technique and by c-Jun immunohistochemical demonstration. The Bcl-2 also was immunohistochemically studied for its anti-apoptotic role. Mitotic index (MI), labeling index (LI) for MIB.1 and PCNA and apoptotic index (AI) were calculated and compared among themselves and with histology and survival. It was found that AI correlated with MI (p = 0.001) and was significantly higher in anaplastic than in classic oligodendrogliomas (p = 0.001). Apoptosis occurred only slightly more frequently in cases with high LIs for proliferation markers (MIB.1 and PCNA) (p = non-significant) and it was definitely higher in p53-positive cases (p = 0.008). It did not correlate with bcl-2 which was poorly expressed in oligodendrogliomas, with the exception of cells with astrocytic features. Apoptotic index correlated very weakly with survival (p = 0.05); therefore, it cannot be considered a highly reliable prognostic factor in oligodendrogliomas.

Cell proliferation and invasion were studied in sixty biopsies of malignant gliomas selected to reproduce the spreading modalities identified in ninety autopsy cases of glioblastoma. Cell proliferation was studied by the immunohistochemical demonstration of PCNA and MIB-1 and by the calculation of their labeling indexes (LI). The main finding was that cell proliferation and cell invasion are not necessarily associated. The interface between the solid tumor and the adjacent brain was represented either by a gradient of tumor cell density or by a clearcut demarcation of the tumor. In the first case the LI either did not change in the infiltration area in comparison with solid tumor or it was much lower, whereas in the second case there was a ring with a high density of labeled nuclei at the tumor periphery. Perineuronal satellites were usually positive for proliferation markers. Cells accumulated in the outer cortical layers, from a deeply located tumor, were almost negative, whereas those originating from subarachnoidal or subpial invasion showed a high LI. High LIs were also found in subarachnoidal and subpial growths, and in a cell population descending into the brain parenchyma around meningeal penetrating vessels. The relationship between cell proliferation and invasion from in vivo studies is not a direct and a simple one.