Karl Rössler

Expression of hypoxia-inducible factor-1α in oligodendrogliomas : Its impact on prognosis and on neoangiogenesis

BACKGROUND Hypoxia-inducible factor (HIF)1 alpha is considered to play a key role in the adaptation of cells to hypoxia by stimulating angiogenesis via regulation of vascular endothelial growth factor and by metabolic adaptation to O(2) deprivation. METHODS Expression of HIF-1 alpha protein and p53 was investigated by immunohistochemistry in 51 specimens of supratentorial pure oligodendrogliomas. Microvessels density (MVD) was determined by anti-CD34 immunostaining. The influence of HIF-1 alpha expression on survival was investigated using univariate and multivariate analysis. RESULTS Strong expression of HIF-1 alpha was observed in 12 (23.5%) specimens, moderate in 21 (41.2%) specimens, and weak in 8 (15.7%) cases, and no expression was found in 10 samples (19.6%). There was no correlation of HIF-1 alpha expression with histologic grading (P = 0.428, Mann-Whitney test). Hypoxia-inducible factor-1 alpha expression and MVD showed a strong correlation (P < 0.001, r = 0.735, Spearman coefficient of correlation). Overexpression of p53 was observed in only two cases. Patients with strong or moderate expression of HIF-1 alpha had a significantly shorter overall survival rate compared with those with low or no expression in univariate (P = 0.0434; log-rank test) and multivariate analysis (P = 0.0187). CONCLUSIONS Overexpression of HIF-1 alpha indicates a diminished prognosis in oligodendrogliomas, independent of p53 status. This finding may be explained by the strong vascularization of these tumors that prevents hypoxia and allows O(2) diffusion and henceforth tumor progression.

Vascular Patterns in Glioblastoma Influence Clinical Outcome and Associate with Variable Expression of Angiogenic Proteins: Evidence for Distinct Angiogenic Subtypes

No data exist on angiogenic patterns and their prognostic impact in human glioblastoma. Such data are relevant for translation of antiangiogenic therapies into clinical applications. Using immunohistochemistry for CD34, we assessed vascular patterns in 114 primary glioblastomas. Vascular patterns comprised unevenly distributed glomeruloid/garland‐like/clustered bizarre vascular formations and evenly distributed delicate capillary‐like microvessels (“classic” vascular pattern). The combination of low content of bizarre vascular formations and prominent classic vascular pattern (n=29) was an independent factor for longer survival (p= 0.006, Cox regression), as well as postoperative high Karnofsky performance status (p=0.005). In patients with a prominent classic vascular pattern, there was no difference of MIB1 labeling index whereas microvessel density and apoptotic index (TUNEL) were significantly higher as compared to all other patients (p<0.05). In addition, diffuse expression of hypoxia‐inducible factor (HIF)‐1α and strong expression of vascular endothelial growth factor were more common (p<0.05, Chi‐square test). FISH revealed loss of chromosomes 1p and 19q only in 1/7 long‐time survivors with classic pattern. We conclude that vascular patterns in primary glioblastoma influence clinical outcome and associate with variable expression of angiogenic proteins. Our findings denote for the first time distinct angiogenic subtypes of human glioblastoma which may prove relevant for anti‐angiogenic therapy approaches.

Expression of Adhesion Molecules and Histocompatibility Antigens at the Blood-Brain Barrier

The migration of inflammatory cells through the blood‐brain barrier in health and disease involves complex interactions between hematogenous cells, endothelial cells, the basement membrane and the perivascular glia limitans. Recent evidence is presented, suggesting that part of these interactions involve antigen independent mechanisms, mediated by cellular adhesion molecules. The accessibility of endothelial adhesion molecules in the intact blood‐brain barrier is lower compared to vessels in other organs. This may account for the low traffic of hematogenous cells through the normal blood‐brain barrier. However, in inflammatory conditions the expression of endothelial adhesion molecules is upregulated, which may lead to recruitment of inflammatory cells into the lesions. Antigen specific activation of T‐ceiis at the blood‐brain barrier apparently takes place mainly at perivascular monocytes and microglia cells. In severe inflammatory lesions, however, astrocytes may be additionally involved in antigen presentation.

Suprasellar meningioma with expression of glial fibrillary acidic protein: a peculiar variant

A 24-year-old female presented with a 3-year history of a suprasellar and intraventricular solid midline process measuring about 3×4 cm. At surgery, this tumour was sharply delineated and of stone-like firmness and was removed completely. Histology suggested meningioma, featuring nests and cords of epithelium-like cells with prominent cytoplasm amidst abundant fibrous stroma with prominent lymphoplasmocellular infiltration. Immunocytochemically, the tumour cells expressed vimentin, S-100 protein, epithelial membrane antigen, cytokeratins, and most surprisingly, glial fibrillary acidic protein (GFAP). Ultrastructural investigation revealed abundant intermediate filaments and occasionally dense secretory granules in tumour cells with short, finger-like cytoplasmic processes joined by very rare small, but well-developed desmosomes. This tumour most likely represents a peculiar variant of meningioma with prominent production of GFAP, as previously described [Budka H (1986) Acta Neuropathol (Berl) 72: 43–54].

Proliferation and DNA fragmentation in meningioma subtypes

H. Maier, J. Wanschitz, R. Sedivy, K. Rössler, D. Öfner and H. Budka (1997) Neuropathology and Applied Neurobiology23, 496–506

Feasibility of Long-Term Intraventricular Therapy with Mafosfamide (n = 26) and Etoposide (n = 11): Experience in 26 Children With Disseminated Malignant Brain Tumors

Treatment options for leptomeningeal disseminated brain tumors are limited by the lack of effective drugs for intrathecal therapy of non-hematologic malignancies. We report on our experience with an intraventricular therapy consisting of mafosfamide, a preactivated cyclophosphamide derivative, and etoposide. Between May 1994 and 2002, 26 patients aged 2–19 years with various intensely pretreated disseminated brain tumors received intraventricular mafosfamide via an indwelling subcutaneous reservoir. Twenty-three of them received a dose of 20mg. Mafosfamide was administered once or twice weekly until remission was achieved and every 2–6 weeks thereafter as maintenance therapy for a total of 736 administrations (2–63/patient). Since March 1998, two patients were switched to receive intraventricular etoposide and nine received etoposide alternating with mafosfamide. Etoposide was given at a dose of 0.5mg×5d every 3–6 weeks for a total of 122 courses (1–29/patient).Immediate toxicities such as transient headaches, nausea, and vomiting occurred with mafosfamide but were manageable with premedication. Etoposide did not cause any discomfort. No long-term toxicities attributable to intrathecal therapy as evidenced by magnetic resonance imaging or neurologic evaluation were observed. Since all patients received some sort of concurrent anti-cancer therapy, the efficacy of intrathecal therapy cannot be assessed independently. However, seven of 13 patients evaluable for response by cerebrospinal fluid (CSF) cytology developed CSF dissemination under systemic chemotherapy and cleared their CSF only after administration of intrathecal mafosfamide. In conclusion, intraventricularly administered mafosfamide at a dose of 20mg and etoposide at a dose of 0.5mg×5d for patients over 2 years of age are feasible and safe and may produce responses.

Fluorescent In Situ Hybridization on Isolated Tumor Cell Nuclei: A Sensitive Method for 1p and 19q Deletion Analysis in Paraffin-Embedded Oligodendroglial Tumor Specimens

In oligodendroglial neoplasms, losses of chromosomal material at 1p and 19q associate with chemosensitivity and prolonged survival. Thus, 1p/19q testing is increasingly proposed for use in brain tumor diagnosis and prognostic assessment. Fluorescent in situ hybridization (FISH) is a classic technique for investigation of 1p/19q status in paraffin-embedded tissues. A major limitation of this method is truncation of tumor cell nuclei complicating assessment of hybridization results. In our study, we analyzed 1p and 19q status in a series of 79 oligodendroglial neoplasms (49 oligodendrogliomas, 30 oligoastrocytomas, WHO: 57 Grade II, 22 Grade III tumors) and controls (gliotic brain tissue: n = 4, diffuse low-grade astrocytoma: n = 4) using FISH on isolated whole tumor cell nuclei, prepared as cytospin preparations, thus bypassing the problem of nuclear truncation. For interpretation of FISH results, we used consensus criteria as defined by the SIOP—Europe Neuroblastoma Study Group for analysis of peripheral neuroblastic tumors. FISH yielded interpretable results in 98.7% for 1p and 92.1% for 19q. Chromosome 1p/19q alterations comprised deletions (1p: 79.5%, 19q: 80%) and imbalances (1p: 11.5%, 19q: 12.9%). 1p aberrations were more frequent in oligodendroglioma than in oligoastrocytoma (100% versus 75.9%, P = .001). The frequency of 1p/19q alterations was not significantly different in WHO Grade II or Grade III tumors or in primary and recurrent tumors. We conclude that FISH on isolated cell nuclei, with application of the SIOP Europe Neuroblastoma consensus criteria, is a sensitive method for detection and interpretation of 1p and 19q aberrations in paraffin-embedded tissue specimens of oligodendroglial neoplasms.

No prognostic impact of survivin expression in glioblastoma

Survivin is a member of a novel protein family of inhibitors of apoptosis, and also plays a role as a potent regulator of mitosis. In semiquantitative Western blot analysis of glioblastomas, survivin expression was shown to be a prognostically significant factor. In the present study we investigated the immunohistochemical expression of survivin and its prognostic impact in a large glioblastoma series comprising 104 consecutive adult patients undergoing a first operation for glioblastoma. We analyzed survivin, Ki-67, and topoisomerase-II-alpha expression in paraffin-embedded tissue, and correlated patient age, Karnofsky performance score, vascular pattern and survivin-, Ki-67-, topoisomerase-II-alpha-, and apoptotic indices with patient outcome using univariate and multivariate survival analysis. Survivin was expressed in all glioblastoma samples, and was prominent in a fraction of nuclei of tumor cells and vascular cells. Further, survivin labeled spindle- and chromosomal material of mitotic figures. Faint cytoplasmic expression was also seen. The survivin index showed significant correlation with Ki-67 and Topo-II-alpha indices. On average, 58.85% of Ki-67 and 91.08% of survivin-expressing nuclei co-expressed Ki-67 and survivin. The survivin index did not correlate significantly with overall survival, whereas patient age, Karnofsky performance score, vascular pattern, and Ki-67 and topoisomerase-II-alpha indices were associated with patient outcome. In summary, in glioblastoma, survivin is expressed predominantly in proliferating tumor cell nuclei. In contrast to Ki-67 and topoisomerase-II-alpha, survivin expression does not influence patient outcome. So, in contrast to Ki-67, survivin does not seem to be useful as prognostic factor in the clinical setting.

High expression of DNA topoisomerase IIα and Ki-67 antigen is associated with prolonged survival in glioblastoma patients

Assessment of tumour cell proliferation in glioblastoma (GB) has been a topic of considerable research interest over the past decade. However, the correlation of tumour proliferation and patient outcome has yielded controversial results. In this study, we examined immunohistochemically, using paraffin-embedded tissue, the expression of the proliferation-related markers DNA topoisomerase IIalpha (TIIalpha) and Ki-67 antigen in a cohort of 114 GB patients treated consecutively with surgery and radiochemotherapy, and correlated the expression with patient outcome. The TIIalpha labelling index (LI) ranged between 5.2 and 87.2% (median: 25.6%). Survival analysis disclosed an association between high TIIalpha expression levels and prolonged survival (P=0.040, log-rank test). TIIalpha expression correlates closely with Ki-67 labelling index (R=0.927, P<0.001), which itself is predictive of patient survival (P=0.044). However, in multivariate analysis, only the Karnofsky performance status remained predictive of patient survival. We conclude that high expression of TIIalpha and Ki-67 appears to be associated with a prolonged survival in our cohort of GB patients.

Fibrous meningeal tumours with extensive non-calcifying collagenous whorls and glial fibrillary acidic protein expression: the whorling-sclerosing variant of meningioma.

Meningiomas comprise a wide range of morphological patterns. We describe unusual fibrous meningeal tumours in two patients, composed of extensive non‐calcifying collagenous whorls of varying size, resembling non‐calcified psammoma bodies, while interposed tumour cells are sparse. Immunohistochemistry showed expression of S‐100, vimentin and glial fibrillary acidic protein, whereas only single tumour cells stained for epithelial membrane antigen. Electron microscopy detected desmosomes or desmosome‐like structures in both specimens. We conclude that these tumours represent a peculiar whorling‐sclerosing variant of fibrous meningioma. Recognition of this meningioma variant is important in the differential diagnosis of meningioma versus other fibrous tumours of the meninges, including solitary fibrous tumours of the meninges, unusual forms of desmoplastic gliomas or chondroid tumours.