Peritumoral vasogenic brain edema (PVBE) is a common accompaniment of malignant gliomas. It results from microvascular extravasation of plasma fluid and proteins through the interendothelial spaces. Tumor-associated cysts (TACs) are observed more commonly with benign gliomas that are not associated with PVBE. This study investigates the hypothesis that these morphologically distinct epiphenomena of microvascular extravasation are linked by a common pathophysiological mechanism involving vascular endothelial growth/permeability factor (VEG/PF), which has been implicated in vascular leak phenomena including ascites, malignant effusions, and brain edema. Furthermore, VEG/PF has been isolated from cultured glioma cells, and both VEG/PF protein and messenger RNA transcripts are expressed in brain tumor tissue. To further elucidate the relationship of VEG/PF to PVBE and TACs, the authors examined 34 pathological specimens for VEG/PF expression. Nineteen primary low-grade tumors, 11 primary high-grade tumors, and four gliosis controls were immunostained with a polyclonal anti-VEG/PF immunoglobulin G antibody. Magnetic resonance imaging was used to quantitate PVBE and to determine the presence of TACs and tumor enhancement. The study revealed that eight VEG/PF-negative specimens exhibited no significant edema, whereas 26 VEG/PF-positive tumors exhibited either significant PVBE or TACs. Notably, eight of nine benign TACs that were not associated with PVBE immunostained positive for VEG/PF. These data indicate a high degree of correlation between VEG/PF expression by gliomas and the occurrence of PVBE or TACs, irrespective of tumor grade, thus supporting VEG/PF's pivotal role as the common pathophysiological link between these processes.

Metastatic brain tumors are almost always associated with vasogenic brain edema, which in turn plays a pivotal role in the evolution of neurological morbidity associated with these lesions. Attention has recently focused on a group of proteinaceous vascular permeability factors (VPF's) that are capable of inducing angiogenesis and promoting increased capillary permeability. To test the hypothesis that metastatic brain tumors expressing VPF's are associated with peritumoral cerebral edema, a rabbit polyclonal immunoglobulin (Ig) G anti-VPF was used to immunostain pathological specimens of metastatic cerebral tumors obtained from 22 patients who underwent surgery at Yale-New Haven Hospital. Magnetic resonance (MR) imaging was used to correlate VPF staining in tumor tissue with the occurrence of peritumoral brain edema. A histological study of the microvasculature was then conducted by immunostaining the specimens for endothelial cell factor VIII surface antigen, using two gliosis specimens as controls. Results revealed 21 of 22 tumors stained positively for VPF's; the negative-VPF tumor was a melanoma that exhibited no peritumoral edema. Twenty of 22 tumors had MR imaging-evident vasogenic edema. The presence and intensity of VPF immunostaining of microvascular features were noted and compared. Factor VIII staining demonstrated tumor vascularity to be most abundant in VPF-rich regions of tumor. The authors therefore report a high correlation between the presence of VPF's and the occurrence of peritumoral brain edema associated with cerebral metastases.

A factor in normal serum that selectively and reversibly inhibits proliferation of cells in culture has been enriched 160-fold from calf serum by sequential ammonium sulfate precipitation, gel filtration, and lectin-affinity chromatography. DNA synthesis of normal (but not transformed) rat hepatocytes, human lymphoblast lines, and mitogen-stimulated murine spleen cells is inhibited by >90%, and Vero, murine myeloma, MELC, and a human colon carcinoma cell line to a lesser extent. Growth of other cell lines tested was not affected.Responsive cells are arrested apparently in GI by this inhibitor, the effect of which is maximal by 24 hr and is spontaneously reversible thereafter unless it is renewed. The active fraction is a protein that migrates with the a2-globulins; it is not a lipoprotein, and it is of high apparent molecular weight.Negative signals inhibiting proliferation may, in balance with mitogens, control the growth of at least some cell types (1). Some humoral inhibitors of proliferation have been designated (2-6), but they remain less well characterized than the hormonal growth factors of serum (7)(8)(9)(10)(11)(12). Owing to its content of these mitogens, increasing amounts of serum enhance the rate of proliferation of most cell types in a concentration-dependent manner. In contrast, we found that the proliferation of normal (i.e., nontransformed) hepatocytes in primary culture or in established lines was depressed in proportion as the concentration of serum in the medium was increased above a minimal 2%. This negative effect on growth was found to be due to the presence of an active inhibitory principle in the a2-globulin fraction of normal sera (13) that can also suppress the proliferation of certain other cell types. In this report we describe the occurrence and some characteristics of this serum inhibitor, its partial purification from calf serum, and its effects on growth or mitogenesis of various cell types. The selectivity, reversibility, nontoxicity, and kinetics of growth inhibition by this serum component suggest that it may be of physiologic signifi- Assay of Cell Proliferation and Inhibition of Growth. Assays of inhibitor activity were performed at a concentration of serum (5%) that permitted logarithmic growth without, however, overcoming possible negative effects on growth of putative inhibitory fractions. In the assay system used, a known amount of test fraction in phosphate-buffered saline (pH 7.2) was included in medium containing 5% FCS, to a final concentration of 10% (vol/vol). Control medium was identical but contained 10% phosphate-buffered saline alone. At intervals, as indicated, cells were counted or, more usually, [3Hjthymidine (New England Nuclear, NET 027E, 20 Ci/mol) was added in a final concentration of 1 ,uCi/ml (1 Ci = 3.7 X 1010 becquerels). After 1 or 2 hr, cultures were rinsed with Hanks' balanced salt solution and the label was chased with medium containing excess nonradioactive thymidine. After the cells were harvested and centrifuged, the pellets were disso...