p73 regulates DRAM-independent autophagy that does not contribute to programmed cell death.

Evading programmed cell death is a common event in tumour development. The p53 family member, p73, is a potent inducer of death and a determinant of chemotherapeutic response, but different to p53, is rarely mutated in cancer. Understanding cell death pathways downstream of p53 and p73 is therefore pivotal to understand both the development and treatment of malignant disease. Recently, p53 has been shown to modulate autophagy--a membrane trafficking process, which degrades long-lived proteins and organelles. This requires a p53 target gene, DRAM, and both DRAM and autophagy are critical for p53-mediated death. We report here that TA-p73 also regulates DRAM and autophagy, with different TA-p73 isoforms regulating DRAM and autophagy to varying extents. RNAi knockdown of DRAM, however, revealed that p73's modulation of autophagy is DRAM-independent. Also, p73's ability to induce death, again different to p53, is neither dependent on DRAM nor autophagy. In contrast to TA-p73, deltaN-p73 is a negative regulator of p53-induced and p73-induced autophagy, but does not affect autophagy induced by amino-acid starvation. These studies, therefore, represent not only the first report that p73 modulates autophagy but also highlight important differences in the mechanism by which starvation, p53 and p73 regulate autophagy and how this contributes to programmed cell death.

Discovery of a perinecrotic 60 kDa MDM2 isoform within glioma spheroids and glioblastoma biopsy material.

Necrosis in glioblastoma is often associated with high levels of Fas (APO-1), HIF-1alpha and PARP expression. The presence of such molecules suggests a regulative element to cell death within this tissue, which may involve p53. We aimed to establish whether p53 and its downstream targets Bax, MDM2 and p21 play a role in perinecrotic cell death in glioblastoma. Following sequencing of the p53 gene in U87 and U373 glioma cell lines, p53 was found to be reactive in the p53 wild-type line U87 in response to hypoxia but not in the p53 mutant line, U373. Although no increase in perinecrotic p53 expression was detected in spheroid cultures derived from these lines, a 60 kDa MDM2 isoform lacking a C-terminal domain showed perinecrotic localization, irrespective of p53 status. Similar findings were observed surrounding regions of necrosis in 80% of glioblastoma biopsies examined. Increasing levels of wild-type p53 did not affect cell death in U87 spheroid cultures but killed all U373 cells 3 days post transfection. Dominant negative p53 did not affect cell death in U373 and U87 spheroid cultures. Although p53 accumulation appeared not to be important for the onset of cell death both in spheroid and biopsy cases, high levels of perinecrotic 60 kDa MDM2 may have implications for glioma cell death susceptibility in both p53 mutant and wild-type tumour cell populations.

Highly unsaturated fatty acid induced tumour regression in glioma pharmacodynamics and bioavailability of gamma linolenic acid in an implantation glioma model: effects on tumour biomass, apoptosis and neuronal tissue histology.

Highly unsaturated fatty acids (HUFAs) are naturally occurring anti-tumour agents. HUFAs act as intracellular signalling molecules in cell proliferation and death. In human glioma, HUFAs may stimulate tumour regression and apoptosis. An implantation glioma model, using the C6 glioma cell line, was used to investigate the bioactivity of locally infused n-6 HUFA gamma linolenic acid (GLA). Rat brains (15 normal and 37 C6 tumour bearing) were infused with vehicle or GLA 200 microM-2 mM. The most active local concentration of GLA for anti-tumour activity was 2 mM, infused at 1 microl/h over 7 days. Tumour regression, increased apoptosis and decreased proliferation were observed in tumours of rats infused with this concentration of GLA. Little effect on normal neuronal tissue was detected. The intraparenchymal route was an effective method of GLA administration in the treatment of glioma. These studies provide further insights into the potential role of HUFAs as anti-glioma agents.

Antitumour and pro-apoptotic actions of highly unsaturated fatty acids in glioma.

The highly unsaturated fatty acids (HUFA) of the n-6 and n-3 series are involved in cell signalling in normal and transformed cells and have recently been associated with pathways leading to tumour cell death. The antitumour activity of three HUFA (arachidonic acid, gamma linolenic acid and eicosapentaenoic acid) were studied in glioma cells and tissue. Using five glioma models, including primary cell suspensions prepared from 46 human glioma samples and an in vivo rat C6 glioma model, we obtained evidence that, following exposure to HUFA, either administered into the medium surrounding human glioma cells or in 16 preparations of multicellular spheroids derived from human and rodent glioma cell lines (C6, MOG, U87, U373) or administered intra-tumourally by infusion using osmotic mini-pumps in 48 rats, glioma regression and apoptosis were detected. Additionally, synergy between gamma irradiation and HUFA administration was observed in 13 experiments analyzing C6 glioma cell apoptosis in vitro. These pro-apoptotic and antiproliferative activities were observed using both C18 and C20 fatty acids of the n-6 and n-3 series, but not when saturated and monounsaturated C18 and C20 fatty acid preparations were used. In the glioma infusion model, in addition to the apoptosis detected in glioma tissue infused with HUFA for 3-7 days, preservation of normal neural tissue and vasculature in adjacent brain was observed. Also, there was little evidence of acute inflammatory infiltration in regressing tumours. Our findings suggest that intraparenchymal infusion of HUFA may be effective in stimulating glioma regression.

Glioma tumourgenicity is decreased by iNOS knockout: experimental studies using the C6 striatal implantation glioma model.

Nitric oxide synthase (NOS) has recently been shown to be an important pathophysiological regulator in experimental implantation glioma since manipulation of NOS can significantly alter tumoural blood flow and inhibit tumour growth. In this study we investigated the role of iNOS (inducible NOS) in glioma tumourogenisis using the rodent C6 striatal implantation model. We produced genetically engineered C6 clones that do not express iNOS activity even after stimulation with a mixture of lipopolysaccaride (LPS) and tumour necrosis factor (TNF)-alpha. These iNOS knockout cells showed a similar growth rate to control cells in vivo at 5 days. We then performed an in vivo implantation glioma study using either the iNOS knockout clone or two genetically engineered control C6 clones. There was a significant reduction (p < 0.01) of tumour mass with the iNOS knockout clone 28 days after the implantation. Immunocytochemistry indicated infiltrates of CD3 positive T cells and macrophages in the controls and the iNOS knockout group. These studies indicate that iNOS expression by tumour parenchymal cells is a critical factor for tumour growth with this model. The mechanisms that cause failure of tumour growth need clarification prior to considering that specific iNOS inhibitors might be candidates for adjuvant treatment of malignant glioma.

Heme oxygenase (HO) isoforms in experimental C6 glioma: an immunocytochemical study.

Abstract Because of their potential to regulate tumoural blood flow and interactions with nitric oxide the expression of the type 1 and 2 isoforms of heme oxygenase (HO-1 and HO-2) were evaluated in implanted C6 striatal gliomas. Immunocytochemistry using antibodies specific for HO-1 and HO-2 were used in 20 C6 glioma tumours. The bulk of the tumour parenchyma and endothelium was negative for both HO isoforms. Isolated, but weak staining for HO-1 was seen in most tumours with focally increased expression in perinecrotic regions. Cells morphologically resembling macrophages stained with both HO-1 and HO-2, but were not numerous. These findings suggest that carbon monoxide, unlike nitric oxide, does not have a major role in regulating tumoural blood flow in this experimental glioma model. These findings once again demonstrate the differences between human malignant glioma and experimental implantation glioma models.

Gliosarcoma with areas of primitive neuroepithelial differentiation and extracranial metastasis.

We report a case of gliosarcoma with areas of primitive neuroepithelial differentiation arising in the temporal lobe of a 53-year-old man. The sarcomatous component of this tumor was perivascular in its distribution and showed expression of factor VIII-related antigen, smooth muscle actin and CD34. The primitive neuroepithelial component possessed a small cell morphology and showed expression of neuronal antigens. Strong expression of p53 was demonstrated throughout the tumor with only focal weak expression of epidermal growth factor receptor. The tumor developed widespread extraneural metastases 5 months after surgical resection of the primary tumor. Histological examination of the liver metastases showed them to consist predominantly of the primitive neuroepithelial component. We believe this to be a novel pattern of differentiation in a gliosarcoma which in this case was associated with an aggressive metastatic potential.

The development of necrosis and apoptosis in glioma: experimental findings using spheroid culture systems.

Cell death in gliomas may occur either by apoptosis, or, in the case of high grade tumours, by necrosis, but questions remain as to the pathogenesis and relationship between these processes. The development of cell death was investigated in multicellular glioma spheroid cultures. Spheroids model the development of cell death due to diffusion gradients in a three-dimensional system without confounding influences of immune response, pressure gradients, etc. Spheroid cultures were established from four malignant glioma cell lines: U87, U373, MOG-G-CCM and A172; harvested from culture at weekly intervals and stained with Haematoxylin and Eosin (H&E), TdT-mediated dUTP-X nick end labelling (TUNEL) and by immunohistochemistry for vimentin, Glial Fibrillary Acidic Protein (GFAP) and Ki67. Annexin V flow cytometry and counts of apoptotic cells on H & E stained sections were performed to assess levels of apoptosis. Modes of cell death were also characterized by electron microscopy. Spatially separate zones of proliferation, differentiation and central cell death developed with increasing spheroid diameter. Central cell death developed at a predictable radius (300-400 microm) for each cell line. Ultrastructural examination showed this to be necrotic in type. Apoptosis was most reliably assayed by morphological counts using H & E. Basal levels of apoptosis were low (< 0.5%), but increased with increasing spheroid diameter (> 2% in U87). In particular, levels of apoptosis rose following development of central necrosis and apoptoses were most abundant in the peri-necrotic zone. There were quantitative differences in the levels of apoptosis and necrosis between glioma cell lines. The predictable onset of necrosis in the spheroids will allow us to investigate the pathogenesis of necrosis and events in prenecrotic cells. There is a relationship between the development of necrosis and apoptosis in this model and these processes can be separately assayed. Further in vitro and genetic studies will enable us to study these events and interactions in greater detail than is possible using other cell culture and in vivo systems.

Expression of poly(ADP-ribose) polymerase and distribution of poly(ADP-ribosyl)ation in glioblastoma and in a glioma multicellular tumour spheroid model.

Development of necrosis is a characteristic feature of glioblastoma but its pathogenesis remains poorly understood. The process of poly(ADP-ribosyl)ation in response to DNA damage is mediated by poly(ADP-ribose) polymerase (PARP) and results in NAD+ depletion. The consequent ATP and energy depletion may result in cell necrosis. Therefore PARP activation is a potential candidate for a regulatory role in the pathogenesis of necrosis in glioblastoma. This study investigated whether there might be a relationship between both PARP expression and poly(ADP-ribosyl)ation, and necrosis in glioblastoma. The pattern of expression of PARP and of poly(ADP-ribose) groups in an archival series of glioblastoma was examined using immunohistochemistry. These parameters were also studied in multicellular tumour spheroids, derived from human glioma cell lines in which central necrosis develops with increasing spheroid diameter. Poly(ADP-ribose) groups were expressed in peri-necrotic tumour cells in glioblastoma. In the spheroid model poly(ADP-ribosyl)ation was seen centrally in pre-necrotic and necrotic cells with increasing spheroid diameter. PARP was widely expressed in viable tumour cells in the glioblastoma sections. In the spheroids, PARP expression, which was initially diffuse, became confined to the outer proliferative zone with increasing diameter. The pattern of expression of poly(ADP-ribose) groups in the spheroids and in glioblastoma raises the possibility that poly(ADP-ribosyl)ation may play a role in the development of necrosis in glioma. The high basal PARP expression in both glioblastoma and the spheroids suggests that this enzyme may have additional roles in glioma cell biology.

The effects of chronic nitric oxide synthase suppression on glioma pathophysiology.

Nitric oxide synthase (NOS) is strongly expressed in glioma and has an important role in tumour blood flow (TBF) regulation. Whether manipulation of NOS function within a tumour can have any therapeutic effect is unknown. This study therefore evaluated the pathophysiological effects of chronic systemic NOS inhibition on experimental rodent glioma blood flow, growth and necrosis. To determine the duration and pathophysiological effects of systemic NOS inhibition, Ng-nitro-L-arginine methyl ester (L-NAME) was given to rats bearing C6 glioma acutely (single dose i.v., 30 mg kg) or for either 4 or 7 days (i.p. 75 mg kg day) prior to study. TBF and local cerebral blood flow (LCBF) were measured using C14-iodoantipyrine quantitative autoradiography. Tumour volume, tumoural necrosis and tumoural NOS were measured using conventional neuropathology and immunocytochemistry. Acute and 4-day L-NAME administration produced significant TBF reductions (-48 and -39%, respectively) with less marked changes in LCBF (-35 and -15%, respectively). Seven-day L-NAME administration reduced tumour volume (p = 0.12), increased tumoural necrosis (p < 0.05), but immunohistochemistry showed no difference in tumoural NOS expression. These results confirm that NOS has a significant role in the pathophysiology of experimental glioma, and that in this glioma model the effects of chronic systemic NOS inhibition are, for the period under study, predominately anti-tumoural. Whether chronic NOS inhibition is useful as an adjunct in glioma therapy or provides the opportunity for novel therapeutic approaches requires further study.

Effects of N-6 essential fatty acids on glioma invasion and growth: experimental studies with glioma spheroids in collagen gels.

OBJECT: Intracranial infusions of gamma-linolenic acid (GLA), an essential fatty acid, have been used as an adjuvant therapy following malignant glioma resection; however, little is known about the dose response of glioma cells to this therapy. In this in vitro study the authors address this important pharmacological question. METHODS: Glioma spheroids derived from U87, U373, MOG-G-CCM, and C6 cell lines were grown in collagen gel and exposed to a range of GLA concentrations (0-1 mM) for 5 days. The diameter of glioma spheroids was measured, the apoptotic index was assessed using both the terminal deoxynucleotidyl transferase-mediated deoxyuridine triphosphate nick-end labeling technique and cell morphological testing, and the levels of proliferating cell nuclear antigen were also measured. CONCLUSIONS: The dose-response patterns were similar for all four glioma spheroids. Low concentrations of GLA (<100 microM) increased both apoptosis and proliferation with a net increase in tumor growth and invasion, whereas high-dose GLA (>100 microM) significantly impaired spheroid cell growth. The proliferative effects of low-dose GLA could be a hazard in the clinical treatment of malignant glioma; however, because of the low toxicity of GLA against normal cells, local delivery of millimolar doses of GLA could significantly reduce tumor size.

Glioma cells transduced with selection transgenes may not form gliomas in vivo and can also inhibit glioma formation by admixed wild glioma cell lines.

Following in vitro lipofection transfection of the rat glioma cell line A15A5 with the plasmid transgene CMV/HyTK, which confirms hygromycin resistance and ganciclovir sensitivity, a series of experiments was planned in which the "bystander" phenomenon would be evaluated using the rodent implantation glioma model. However examination of the brain of rodents in which the A15A5HyTK cells were implanted showed no evidence of glioma growth. Furthermore, rodents having intracerebral implantation of (i) wild A15A5 and A15A5HyTK cells in a 50/50 mix, (ii) wild A15A5 and A15A5HyTK cells in a 90/10 mixture and (iii) wild C6 and A15A5HyTK cells in a 50/50 mix all failed to grow macroscopic tumours by 15-17 days irrespective of whether the animals had been administered ganciclovir (GCV) in the week before sacrifice. Neuropathological and immunocytochemical analysis of the implantation sites showed no difference between the GCV and saline treated groups of animals for any implantation cell mix. These observations confirm previous results that suggest transduction of malignant rodent glioma cell lines with a variety of selection, oncogenic and marker genes significantly impairs their in vivo tumorigenic potential compared to the wild type cell lines. This study also demonstrates that even without GCV treatment the transduced cells inhibit, by an unknown mechanism(s), the tumorigenicity of other non transfected malignant cells. The implications of this study for gene therapy of human malignant glioma are discussed.

Terminal ventriculostomy for syringomyelia.

The clinical course of 12 patients who underwent terminal ventriculostomy for syringomyelia is presented. Opening the central canal at the tip of the conus medullaris is a relatively benign procedure that improves the symptoms of syringomyelia and syringobulbia. This canal normally terminates at the tip of the conus, but in each of the 12 surgical specimens it continued into the filum terminale for distances up to 8 cm. In most cases the tip of the conus was located more caudally than normal, indicating some degree of tethering in fetal life. This belief is supported by the fact that the newborn, whose conus is tethered to a lipoma at the sacral level, may develop syringomyelia in adult life.