Expression of vascular endothelial growth factor and its possible relation with neovascularization in human brain tumors.

To examine which growth factors correlate with neovascularization in human brain tumors, the mRNA levels of transforming growth factor alpha, transforming growth factor beta, basic fibroblast growth factor, and vascular endothelial growth factor (VEGF) genes were determined by a Northern blot analysis in surgically obtained human gliomas and meningiomas. The vascular development was determined by counting the number of microvessels which were immunostained with von Willebrand factor. We normalized the growth factor mRNA levels versus the glyceraldehyde phosphate dehydrogenase mRNA level. In the 17 gliomas and 16 meningiomas examined, the mRNA of transforming growth factors alpha and beta, basic fibroblast growth factor, and VEGF were expressed at various levels. Among those 4 growth factors, the mRNA levels of VEGF, but not those of transforming growth factors alpha and beta and basic fibroblast growth factor, correlated significantly with vascularity in both gliomas (correlation coefficient r = 0.499; P < 0.05) and meningiomas (correlation coefficient r = 0.779; P < 0.001). These findings thus suggest that VEGF may be a positive factor in tumor angiogenesis in both human gliomas and meningiomas.

A herpes simplex virus type 1 mutant deleted for gamma34.5 and LAT kills glioma cells in vitro and is inhibited for in vivo reactivation.

To create an oncolytic herpes simplex virus type 1 (HSV-1) that is inhibited for reactivation, we constructed a novel herpes recombinant virus with deletions in the gamma34.5 and LAT genes. The LAT gene was replaced by the gene for green fluorescent protein, thereby allowing viral infection to be followed. This virus, designated DM33, is effective in killing primary and established human glioma cell lines in culture. DM33 is considerably less virulent following intracerebral inoculation of HSV-susceptible BALB/c mice than the wild-type HSV-1 strain McKrae. The safety of this virus is further supported by the retention of its sensitivity to ganciclovir and its relatively limited toxicity against cultured human neuronal cells, astrocytes, and endothelial cells. The ability of DM33 to spontaneously reactivate was tested in a rabbit ocular infection model that accurately depicts human herpes infection and reactivation. Following ocular infection of rabbits, spontaneous reactivation was detected in 83% (15/18) of the eyes infected with wild-type McKrae. In contrast, none of the eyes infected with DM33 had detectable reactivation. The efficacy of this virus in cultured human glioma cell lines, its safety, confirmed by its inability to reactivate, and its attenuated neurovirulence make DM33 a promising oncolytic agent for tumor therapy.

P-glycoprotein expression in brain capillary endothelial cells after focal ischaemia in the rat.

P-glycoprotein (P-gp) is expressed not only in tumour cells but also in some normal tissues including brain capillaries. We investigated whether or not P-gp was expressed in the capillary endothelial cells of a rat focal ischaemic brain. The brains were immunohistochemically studied for Factor VIII, glial fibrillary acidic protein (GFAP), and P-gp. Endothelial gamma-glutamyl transpeptidase (gamma-GTP) activity, which is thought to be induced by glial cells, was also studied histochemically. The P-gp positive endothelial cells disappeared in the ischaemic lesion by post-ischaemic day 3. Factor VIII-positive regenerating capillaries were first observed on day 3 without P-gp expression. The P-gp positive endothelial cells began to reappear on day 5, and were detected in all the endothelial cells by day 8. The P-gp expression in endothelial cells showed a similar pattern as that of gamma-GTP, and seemed to correlate with GFAP-positive reactive astrocytes. The newly-formed brain capillaries thus appeared to have a potential to express P-gp in abnormal pathogenic conditions as cerebral infarction, and our present study also suggested that P-gp in the brain capillaries might therefore be expressed in conjunction with glial cells.

Increased delivery of a new cisplatin analogue (254-S) in a rat brain tumor by an intracarotid infusion of bradykinin.

The intracarotid infusion of bradykinin has been shown to selectively increase capillary permeability in a brain tumor without affecting either normal brain capillary permeability or the systemic blood pressure. We examined whether the intracarotid infusion of bradykinin could selectively increase the delivery of a new watersoluble antitumor agent, cis-diammine glycolato-platinum (254-S, 303.2 mol. wt.), to transplanted RG2 glioma in rats. The platinum contents in the brain, tumor tissues and plasma were measured using an atomic absorption spectrophotometer. The transfer ratio of 254-S from plasma to the tissues was calculated and expressed as the volume of plasma containing platinum per g tissue (Dp, microliter g-1). Intracarotid bradykinin infusion at a rate of 20 micrograms kg-1 min-1 increased the delivery of 254-S in the tumor tissue by 1.3-fold when compared with intracarotid infusion of 0.9% saline (48.78 +/- 18.11 vs. 37.12 +/- 12.53; p < 0.05). In normal brain tissue including the ipsilateral cortex, the contralateral basal ganglia and the contralateral cortex, bradykinin did not significantly increase the delivery of 254-S in comparison with 0.9% saline (12.28 +/- 9.53 vs. 10.70 +/- 5.05, 4.96 +/- 3.54 vs 4.96 +/- 4.80, 7.64 +/- 4.10 vs. 13.07 +/- 11.38, respectively). These results indicate that the intracarotid infusion of bradykinin selectively increases the delivery of 254-S to the brain tumor without affecting the normal brain. This method may, therefore, enhance the antitumor effect of 254-S for the treatment of brain tumors and also reduce neurotoxicity in the normal brain.

Repeated, short-term ischemia augments bradykinin-mediated opening of the blood-tumor barrier in rats with RG2 glioma.

The objective of this study was to investigate the effects of repeated, short-term ischemia on bradykinin-mediated permeability of the blood-brain barrier (BBB) and the blood-tumor barrier (BTB). The mechanism by which bradykinin transiently opens the BTB, involves B2 receptors, Ca2+ flux, nitric oxide (NO) and cyclic GMP (cGMP). Since global and focal cerebral ischemia are known to increase levels of brain nitric oxide synthase (bNOS) and endothelial nitric oxide synthase (eNOS) we tested the hypothesis that bradykinin may increase the BTB permeability to a greater extent under ischemic rather than nonischemic conditions. The vertebral arteries in female Wistar rats were coagulated immediately after intracerebral implantation of RG2 glioma. Short-term ischemia was produced in some rats by a modification of the four-vessel occlusion procedure for incomplete forebrain ischemia, in which the common carotid arteries were clamped daily for 15 min on days 7, 8 and 9 after tumor implantation, after which reperfusion was allowed. On day 10 after tumor implantation, bradykinin (10 microg kg(-1) min(-1)) or phosphate-buffered saline (PBS) was infused for 15 min into the right carotid artery of anesthetized, sham-operated (nonischemic controls) and ischemic rats, followed by an intravenous bolus (100 microCi kg(-1)) each of [14C]-iodo-antipyrine (IAP), [14C]-dextran or [14C]-aminoisobutyric acid (AIB) to measure regional cerebral blood flow (rCBF), blood volume, or unidirectional transfer constant Ki, respectively, by quantitative autoradiography. A single 15-min ischemic episode significantly decreased rCBF in the tumor center (158.9 +/- 17.33 in control vs. 58.78 +/- 24.45 ml 100 g(-1) min(-1) in ischemic group; p < 0.01) and in the tumor periphery (106.82 +/- 7.34 in control vs. 70.55 +/- 26.66 ml 100 g(-1) min(-1) in ischemic group; p < 0.05). Respective mean blood volume in tumors (11.7 +/- 13.3, 12.7 +/- 14.0, and 13.3 +/- 14.5 microl g(-1)) from ischemic-PBS, nonischemic-bradykinin, and ischemic-bradykinin groups, respectively, was not significantly different; mean blood volume in normal brain (3.7, 3.1 and 3.8 microl g(-1)) was not significantly different among these groups either. Intracarotid infusion of bradykinin following repeated ischemia significantly increased mean Ki, as compared to bradykinin infusion in nonischemic controls, in both the tumor center (36.60 +/- 8.4 vs. 22.90 +/- 4.61 microl g(-1) min(-1), p < 0.05) and in tumor periphery (17.70 +/- 5.93 vs. 8.50 +/- 4.42 microl g(-1) min(-1), p < 0.05). Mean Ki values for tumor center and tumor periphery of ischemic rats receiving intracarotid bradykinin were 3-fold greater than those of nonischemic rats infused with PBS. Immunohistochemical and Western blot analyses showed that repeated, short-term ischemia significantly increased the levels of bNOS in tumor cells and eNOS in tumor capillaries, but neither induced iNOS nor affected B2 receptor levels in tumor cells in vivo, as compared with nonischemic controls. Taken together, these results demonstrate for the first time that repeated, short-term ischemia augments bradykinin-mediated opening of the BTB. We conclude that the elevated intratumoral levels of bNOS and eNOS may 'prime' the NO generating capacity of tumor cells. Consequently, increased de novo synthesis and a correspondingly elevated concentration of NO within the tumor, therefore, may be one mechanistic explanation for the significantly increased, bradykinin-mediated BTB opening under ischemic conditions, reported here.

Correlation between bradykinin-induced blood-tumor barrier permeability and B2 receptor expression in experimental brain tumors.

Localization of B2 receptors in brain tumor cells and microvessel endothelial cells of the brain tumors was investigated to study the differential sensitivity of brain tumors to bradykinin. The present study shows that B2 receptor expression levels vary in cultured RG2, C6 and 9L glioma cells as well as in the intracerebral tumors established with these cell lines in rats. The double immunohistochemical data indicate that B2 receptors are localized to tumor cells and not to the tumor capillaries. Immunostaining and Western blot analysis for B2 receptor showed that the B2 receptor expression was in the order C6 > RG2 > 9L. The permeability studies on RG2, C6 and 9L tumors in rats showed that C6 tumor had the highest increase (178%) in Ki (unidirectional transport across blood-brain barrier (BBB)/blood-tumor barrier (BTB)), while 9L tumor had the least increase of Ki (35%) over the control group, following intracarotid infusion of bradykinin. We found a positive correlation (r = 0.965, p < 0.001) between B2 receptor levels and bradykinin-induced increase in BTB permeability. We conclude that B2 receptors are localized to tumor cells and not to normal or tumor capillary endothelial cells. C6 tumor with highest B2 receptor expression was most responsive to bradykinin, while RG2 and 9L tumors with lower B2 receptor expression level were less sensitive to bradykinin with regard to BTB permeability.

Inverting papilloma of the sphenoid sinus: report of two cases.

Two patients with sphenoid sinus inverting papilloma who were treated either by transcranial or sublabial trans-septal approach are reported. Inverting papillomas arising from the sphenoid sinus are exceedingly rare. The clinical and neuro-imaging features, as well as surgical treatment, for sphenoid sinus tumours are also briefly discussed.

P-glycoprotein expression in brain capillary endothelial cells after focal ischemia in rat.

We investigated the time kinetics of P-glycoprotein (P-gp), a membrane bound drug efflux pump for many anti-cancer drugs in multidrug resistant cells, using a rat ischemic brain model. Frozen sections of the brain were studied immunohistochemically with anti-Factor VIII antibody for endothelial cells, with anti-glial fibrillary acidic protein (GFAP) antibody for reactive astrocytes, and with MC6-4 monoclonal antibody for P-gp. A putative blood-brain barrier (BBB) marker, gamma-glutamyl transpeptidase (gamma-GTP), and the progression of the brain edema were also studied. P-gp positive endothelial cells disappeared in the ischemic lesion by post-ischemic Day 3. Factor VIII-positive regenerating capillaries were first observed on Day 3 without P-gp expression when the brain edema reached a maximum. P-gp positive endothelial cells began to reappear on Day 5, and were detected in all endothelial cells by Day 8. The time kinetics of P-gp expression in the endothelial cells showed a similar pattern as that of gamma-GTP, and its induction is associated with GFAP-positive reactive astrocytes. These results suggest that P-gp might play an important role in maintaining the BBB function in conjunction with glial cells.

VEGF and flt. Expression time kinetics in rat brain infarct.

BACKGROUND AND PURPOSE: Vascular endothelial growth/vascular permeability factor (VEGF) is a candidate for an angiogenic and hyperpermeability inducing factor in an infarct because it is a secretable mitogen specific for endothelial cells and is upregulated by hypoxia. Our study attempts to clarify the chronological expression of VEGF and its receptor (flt) system in experimental cerebral infarction. METHODS: With the use of a reproducible middle cerebral artery occlusion model in rats, VEGF expression was identified by Western blotting with anti-VEGF antibody. The chronological expression of the VEGF/flt system was analyzed semiquantitatively by immunohistochemical means in infarcts with different time courses from 3 hours to 3 weeks. RESULTS: VEGF and flt were expressed exclusively in the ischemic brain. The bands obtained on the immunoblot at 38 and 45 kD are related to those of VEGF121 and VEGF165 isoforms. Macrophages, neurons, and glial cells chronologically expressed VEGF immunoreactivity in a different fashion. Both VEGF (bound) and flt were detected in endothelial cells along with the development of angiogenesis. CONCLUSIONS: In the ischemic brain the macrophages, neurons, and glial cells appear to contain VEGF. The VEGF receptor flt was induced in endothelial cells along with the progression of angiogenesis in infarct. The VEGF/flt system is thus considered to be involved in the healing process of brain infarct.

Selective transvascular delivery of oligodeoxynucleotides to experimental brain tumors.

Optimal therapeutic strategy for malignant brain tumors is controversial. Recent studies of viral or nonviral gene therapy in rats emphasize the need for a selective delivery system. We examined whether phosphorothioate oligodeoxynucleotides (lacZ 2157, 5'-GTGGCGTCTGGCGGAAAACC-3') could be selectively delivered transvascularly into experimental brain tumors following intracarotid infusion of bradykinin, a specific blood-tumor barrier opener. The specificity of 32P-labeled complementary antisense lacZ 2157 and the stability of lacZ 2157 in vivo were confirmed using slot-blotting hybridization method and polyacrylamide gel electrophoresis. Concentrations of lacZ 2157 after intracarotid injection (2 mg/kg, 10 microg/kg/min) with or without bradykinin were determined in the brain, tumor tissue, liver, kidney, and plasma. The transfer ratio of lacZ 2157 from the plasma to the tissues was calculated and expressed as tissue content relative to plasma content of lacZ 2157 per mg tissue (Do, microl/mg). Delivery of lacZ 2157 to tumor tissue increased 3.24 times with bradykinin over delivery in controls (0.0243 +/- 0.0176 vs. 0.00750 +/- 0.00389; p < 0.05). Delivery of lacZ 2157 to ipsilateral and contralateral cerebral cortex to the tumor, and delivery to the contralateral basal ganglia, did not increase significantly with bradykinin. These results indicate that such transvascular delivery with bradykinin can deliver a relatively large amount of oligodeoxynucleotide selectively to brain tumors without affecting normal brain.