Glial fibrillary acidic protein synthesized in vitro using messenger RNA from a human glioma cell line

Messenger RNA (mRNA) extracted from a continuous human glioma cell line grown in culture or as a solid tumor was translated in an mRNA-dependent reticulocyte lysate system. Translation products labeled with [35S]methionine were immunoprecipitated with antiserum specific for glial fibrillary acidic (GFA) protein, separated by one- and two-dimensional polyacrylamide gel electrophoresis and analyzed fluorographically. Immunoprecipitates from both cell culture and tumor mRNA translations had a molecular weight of 49,000 daltons, consistent with GFA protein extracted from human tissue. In two dimensions, the 49,000-dalton band resolved into two to three spots at pH 5.7-5.9, the isoelectric point of GFA protein. Minor lower molecular weight products were detected in fluorographs of heavily overloaded gels or in film exposed for extended periods of time. These data indicate that the GFA protein produced by this glioma cell line is chemically and immunologically similar to normal human GFA protein, which suggests that the primary phenotypic expression of GFA protein in this tumor cell line is not altered by the neoplastic process.

Immunohistochemical detection of factor VIII/von Willebrand factor in hyperplastic endothelial cells in glioblastoma multiforme and mixed glioma-sarcoma

The sarcomatous components of most glioma-sarcomas are thought to arise from the neoplastic transformation of hyperplastic endothelial and adventitial vascular cells in a preexisting glioblastoma multiforme. The expression of factor VIII/von Willebrand factor (FVIII/vWF), a marker for endothelial cells, and of glial fibrillary acidic protein (GFAP), a marker for glial cells, was examined in 10 glioblastomas and seven mixed glioma-sarcomas using the peroxidase-antiperoxidase immunohistochemical technique. Hyperplasia of small blood vessels was observed in all 10 glioblastomas; in five, the vascular proliferation had resulted in the formation of prominent glomeruloid structures. FVIII/vWF was detected in the endothelial cells in these vascular structures, but not in the adventitial cells. In the mixed glioma-sarcomas. FVIII/vWF was detected only in endothelial cells; there was no staining for FVIII/vWF in the neoplastic mesenchymal cells. The gliomatous components of the mixed tumors stained intensely for GFAP. These observations indicate that both endothelial and nonendothelial cell types contribute to the small vessel hyperplasia in glioblastomas, and that the sarcomatous components of mixed glioma-sarcomas are derived from either non-endothelial cells or endothelial cells that have undergone antigenic loss following transformation.

Human fetal brain antigen expression common to tumors of neuroectodermal tissue origin

The antigenic relationship between human tumors of neuroectodermal origin and fetal brain were further investigated by characterization of two hybridoma antibodies derived from a fusion of P3-NS1/1-Ag 4-1 (NSI) myeloma cells and splenocytes hyperimmunized to second trimester human fetal brain homogenate. Monoclonal antibodies (MAs) 1H8cl 2 and 1H8cl 3 were analyzed by cell surface radioimmunoassay (CS-RIA), quantitative absorption, indirect immunofluorescence, and peroxidase-antiperoxidase (PAP) immunohistology. MA 1H8cl 3 is the more broadly reactive, binding to 9/14 glioblastoma (GBM), 2/3 neuroblastoma, 1/2 melanoma, and 1 medulloblastoma cell line(s) by CS-RIA analysis, and to 12/15 GBM, fetal brain, spleen, and liver, and adult spleen by PAP analysis. MA 1H8cl 2 is more restricted, binding to 7/14 GBM, 2/3 neuroblastoma, 1 medulloblastoma, and 2/3 fetal skin fibroblast cell line(s) by CS-RIA, and to 9/15 GBM and fetal brain and spleen by PAP analysis. Control non-central nervous system tumors and normal adult tissue including brain, thymus, lymph node, liver, kidney, lung, skin, and pancreas, were unreactive with both 1H8cl 2 and 1H8cl 3 by CS-RIA, PAP, and absorption analysis. The data presented here establish the unique nature of the detected antigenic specificities as compared to previously described oncofetal and onconeural antigens, and define two immune reagents which are operationally specific for tumors of neuroectodermal origin within the adult central nervous system.

Effects of adrenal cortical steroids and osmotic blood-brain barrier opening on methotrexate delivery to gliomas in the rodent: the factor of the blood-brain barrier

The effect of adrenal cortical steroids and osmotic blood-brain barrier modification on methotrexate delivery to normal and glioma-bearing rats was studied. In animals with the avian sarcoma virus-induced glioma, osmotic blood-brain barrier modification resulted in significantly increased delivery of methotrexate to the tumor-bearing hemisphere (including the tumor, the brain around the tumor, and the brain distant to the tumor), compared to the nonmodified hemisphere or to control animals. The administration of adrenal steroids, followed by intracarotid methotrexate, resulted in slightly decreased chemotherapeutic agent (methotrexate) delivery to the tumor, the brain around the tumor, and the brain distant to the tumor. When adrenal steroids were given prior to barrier modification and methotrexate therapy, the level of methotrexate was significantly less in the tumor. These studies provide evidence that the blood-brain barrier exists in tumors and is a factor in drug delivery to tumors. Steroid administration greatly interferes with the enhancement of drug delivery to tumors that can be achieved with osmotic blood-brain barrier modification.

Experimental allergic encephalomyelitis in Lewis rats bearing avian sarcoma virus-induced brain tumors

Lewis rats bearing avian sarcoma virus (ASV)-induced brain tumors were injected with guinea pig spinal cord emulsion and complete Freund's Adjuvant to determine if they remained susceptible to induction of experimental allergic encephalomyelitis (EAE). The incidence of EAE among rats with small, and moderate sized gliomas was similar to non-tumor-bearing controls (P less than 0.5; P less than 0.9) while 18 of 24 (75%) animals with large gliomas developed EAE as compared to 31 of 33 (93%) controls (P less than 0.05). The histologic features and geographical distribution of "ordinary' EAE were seen in controls and were maintained in tumor-bearing rats. The presence of an intracranial tumor did not significantly alter the ability of Lewis rats to develop EAE.

Specificity and sensitivity of immunohistochemical detection of factor VIII/von Willebrand factor antigen in formalin-fixed paraffin-embedded tissue

The immunohistochemical detection of factor VIII/von Willebrand factor antigen (FVIII/vWF-AG) in formalin-fixed paraffin-embedded tissues was investigated using the peroxidase-antiperoxidase (PAP) method. Highly purified human FVIII/vWF was used to raise rabbit anti-FVIII/vWF-AG serum. In addition to anti-FVIII/vWF-AG activity, the unabsorbed antiserum had anti-IgG, anti-IgM, and anti-alpha2-macroglobulin specificities. Following exhaustive absorption with these proteins, the antiserum reacted monospecifically for FVIII/vWF-AG in immunodiffusion, immunoelectrophoresis, and PAP immunohistochemistry. Sections of normal tissues from six patients and a total of 43 neoplasms were examined. Treatment of the tissue sections with trypsin prior to application of the antiserum markedly increased the sensitivity of FVIII/vWF-AG detection. The positive staining for FVIII/vWF-AG was restricted to endothelial cells in both neoplastic and nonneoplastic tissue. In general, the hyperplastic endothelia in neoplastic and reactive tissues stained more intensely than those in normal tissues. Expression of FVIII/vWF-AG by nonendothelial neoplastic cells was not observed. FVIII/vWF-AG is a reliable marker for endothelial cells.

Permeability of different experimental brain tumor models to horseradish peroxidase

The permeability of different brain tumor models to horseradish peroxidase (HRP) was examined by determining the fraction of tumor that contained HRP after intravenous administration. The intracerebral tumor models studied were Avian Sarcoma Virus (ASV)-induced tumors and tumors from transplanted RG-2, S69-C1-5, and 9L cell lines. The average fraction of RG-2 tumors permeable to HRP was .95; of S69-C1-5 tumors, .699; of ASV-induced tumors. .63; and of 9L tumors, .52. Except for the RG-2 tumors, there was considerable regional variation in HRP permeability, which was most marked in the ASV-induced tumors. In ASV-induced tumors, HRP permeability did not correlate with tumor histological classification, size, or anatomic location within the brain. The subcutaneous tumor models studied were RG-2-, S69-C1-5, and 9L-transplanted tumors in rats, and human glioblastoma cell lines transplanted into nude mice. All were completely permeable to HRP. These results indicate that significant differences in permeability to HRP exist among brain tumor models when the tumors are intracerebral, and that all subcutaneous tumors from transplanted glial cell lines are completely permeable to HRP. These variables must be considered in future studies of permeability in experimental brain tumors. Care must be exercised in extrapolating results about permeability from one brain tumor model to another.

Localization of factor VIII/von Willebrand factor and glial fibrillary acidic protein in the hemangioblastoma: implications for stromal cell histogenesis

The histogenesis of hemangioblastoma stromal cells is unresolved. Ultrastructural observations suggest that the stromal cells, endothelial cells, and pericytes that compose this neoplasm are all derived from angiogenic mesenchyme. The expression of factor VIII/von Willebrand factor (FVIII/vWF), a specific marker for endothelial cells, and of glial fibrillary acidic protein (GFAP), a specific marker for glial cells, was examined in 16 hemangioblastomas using the peroxidase-antiperoxidase immunohistochemical method. Endothelial cell staining for FVIII/vWF was intense in 14 tumors, weak in one, and absent in another. There was no stromal cell staining in any of the neoplasms. Process-bearing, GFAP-positive cells were observed near the tumor margin in 13 cases, and deeper in the neoplasm in 8. In two of these tumors there were also occasional GFAP-positive cells that lacked processes and had a vacuolated cytoplasm. Virtually all of the GFAP-positive cells were interpreted as trapped astrocytes rather than stromal cells. The lack of expression of FVIII/vWF by the stromal cells indicates that they are antigenically distinct from endothelial cells. Several alternatives for stromal cell histogenesis remain open. The stromal undergone antigenic loss, or from angiogenic mesenchymal cells that do not express FVIII/vWF. Alternatively, the stromal cells may originate from non-angiogenic mesenchymal cells derived from the mesoderm or neuroectoderm.

Evaluation of the lipid-soluble diaminopyrimidines, metoprine and etoprine, in the avian sarcoma virus rat glioma model

Using the avian sarcoma virus (ASV) rat glioma model, we have evaluated the in vivo effectiveness of two lipid-soluble folate antagonists, metoprine (DDMP) and its 6-ethyl analog etoprine. When adult Fischer 344 rats, which were previously inoculated with ASV as newborns, received DDMP ip, high levels of this drug accumulated in both normal and gliomatous brain tissue, as well as in lung and liver. When animals bearing ASV-induced gliomas were treated with DDMP or etoprine with or without citrovorum factor (calcium leucovorin), there was no increase in survival, even though the drugs were administered at maximally tolerated doses.

Expression of human fetal brain antigens by human tumors of neuroectodermal origin as defined by monoclonal antibodies

The antigenic relationships between human tumors of neuroectodermal origin and fetal brain were investigated by the production of hybridoma antibodies derived from a fusion of P3-NS1/1-Ag 4-1 (NS1) myeloma cells with splenocytes from a mouse multiply immunized with an homogenate of second-trimester human fetal brain tissue. Two monoclonal antibodies (MAs), 4D2cl 6 and 7H10cl 4, were studied in detail by cell surface radioimmunoassay (CS-RIA), quantitative absorption, indirect immunofluorescence, and peroxidase-anti-peroxidase (PAP) immunohistology. MA 4D2cl 6 binds to 5 of 14 glioblastoma (GBM) cell lines, 1 of 2 melanoma cell lines, 1 of 3 neuroblastoma cell lines, and 1 of 5 fetal fibroblast lines by CS-RIA and to 13 of 13 GBM, 1 neuroblastoma, and fetal brain, liver, spleen, and adult spleen unfixed frozen tissue by PAP analysis. MA 7H10cl 4 binds to 13 of 14 GBM, 1 of 3 neuroblastoma, and 1 medulloblastoma cell line(s) by CS-RIA analysis and to 13 of 13 GBM, 1 neuroblastoma, fetal brain, liver, spleen, thymus, and adult spleen by PAP analysis. Control non-central nervous system tumors and normal adult tissue, including brain, were unreactive with both MAs by CS-RIA, PAP, and absorption analysis. Tissue distribution and localization analyses established that MAs 4D2cl 6 and 7H10cl 4 recognize specificities of shared fetal-neuroectodermal-lymphoid distribution which are operationally specific within the adult central nervous system and which are not related to previously described oncofetal or onconeural antigens.

Anaplastic human gliomas grown in athymic mice. Morphology and glial fibrillary acidic protein expression

The morphologic and biochemical characteristics of human surgical biopsy specimens taken from 17 patients with anaplastic human gliomas and of athymic mouse-grown tumors derived from them were examined. Fourteen were categorized as glioblastoma multiforme, one as an anaplastic astrocytoma, one as a recurrent glioblastoma multiforme, and one as a gliosarcoma. Fifteen of 17 tumors stained positively immunohistochemically for glial fibrillary acidic protein (GFAP), a glial-specific marker. When portions of the 17 surgical biopsy specimens were injected into the flank subcutaneous space of athymic mice, 16 produced tumors; different portions of a single biopsy specimen were used to establish three separate tumor lines; in toto, 18 tumor lines were established. Mouse-borne tumors contained various proportions of fibrillary and protoplasmic astrocytes, gemistocytes, small anaplastic cells, and multinucleated giant cells. Some were more homogeneous than the human tumors from which they were derived, while others contained a mixed population similar to that of the original biopsy specimen. Of these initial 18 tumors, 16 were stained for GFAP and 14 contained from fewer than 5% to almost 100% GFAP-expressing cells. Ten of the tumor lines were studied in serial passage, several demonstrating increased cellularity with increased passage. GFAP expression was followed through serial passage, and 7 of 10 tumor lines continued to express it, often in reduced amounts, and 2 of 10 ceased expression, one (the gliosarcoma) never having expressed it. These data demonstrate that while athymic mouse-borne human anaplastic gliomas retained some features of the human tumors from which they were derived, they varied from one another morphologically. These mouse-borne tumors also continued to evolve, often changing their levels of GFAP and demonstrating increased cellularity with passage.

Lymphocyte subpopulations and responsiveness in rats bearing intracranial tumors induced by avian sarcoma virus

The mitogenic responsiveness of peripheral blood lymphocytes obtained from Fischer 344 rats inoculated with avian sarcoma virus was studied. In addition, quantitative alterations in lymphocyte subpopulations were determined in these animals. Only peripheral blood lymphocytes from rats bearing astrocytomas had significantly diminished responses to concanavalin A when compared to control responses. The percentage of lymphocyte subpopulations detected in either the peripheral blood or spleen of tumor-bearing rats did not differ from values obtained with control rats. However, rats bearing astrocytomas had a marked decrease in the absolute number of the various lymphocyte subpopulations as a result of lymphopenia. Neither the sarcoma-bearing rats nor the virus-inoculated rats that did not develop tumors exhibited this lymphopenia. In addition, sera from rats bearing astrocytomas diminished the concanavalin A reactivity of spleen cells obtained from normal rats. The results of this study establish the avian sarcoma virus-induced rat astrocytoma model as a useful immunological parallel for the human disease.

Hyperimmunization of non-human primates with BCG-CW and cultured human glioma-derived cells. Production of reactive antisera and absence of EAE induction

The efficacy and non-toxicity of BCG-cell wall preparation (BCG-CW) + cultured human glioma-derived cells as immunogen in non-human primates (Macaca fascicularis) were investigated to determine the optimal conditions for immunization without induction of experimental allergic encephalomyelitis (EAE). BCG-CW vaccines were prepared with non-irradiated and irradiated (10,000 rads) human glioma cells and used in a short-term hyperimmunization protocol; sera obtained from this regimen were compared to those obtained by long-term hyperimmunization of macaques with complete Freund's Adjuvant (CFA) and glioma cells. Immunization with BCG-CW + either irradiated or non-irradiated human glioma cells produced a specific antibody response in non-human primates without clinical, macroscopic, or microscopic evidence of allergic encephalomyelitis.

Computed tomography of virally induced canine brain tumors: a preliminary report

Seven neonatal dogs were inoculated intracerebrally with Avian Sarcoma Virus (ASV) and studied by computed tomography (CT) for intracranial tumors. The tumor yield was five anaplastic astrocytomas and two sarcomas, with an average latency of 57 days. Computed tomography accurately detected all tumors over 5 mm in diameter and predicted the size of the tumors within 3 mm. The intensity of contrast enhancement was directly related to the dose of Conray-60. The area of enhancement on the CT scan correlated precisely with tumor permeability as determined histologically with horseradish peroxidase (HRP). Edema seen in CT scans correlated well with edema present in histological sections. Peritumoral edema was impermeable to both Conray-60 and HRP. The ASV-induced canine brain tumor model appears well suited for future CT studies.

Relationship of in vitro morphologic and growth characteristics of established human glioma-derived cell lines to their tumorigenicity in athymic nude mice

Fifteen permanent cell lines derived from human gliomas which are individually distinct by immunologic and biochemical criteria were evaluated to determine if morphologic or cell biologic parameters distinguished the 4 lines which were tumorigenic in athymic nude mice. By subjective morphologic appraisal, the 4 tumorigenic lines were considered "malignant" or "borderline," but 4 of the non-tumorigenic lines were also classified in this way. By objective criteria, these 15 lines varied markedly in percentage of piled-up cells, chromatin pattern, pleomorphism, nuclear to cytoplasmic ratio, number of bizarre multinucleate giant cells, presence of abnormal mitotic figures, percentage of colony formation in soft agar, saturation density, population doubling time, and absolute plating efficiency. Among these criteria, percentage of colony formation in soft agar had the highest correlation coefficiency with tumorigenicity, and when this parameter was held constant the only additional characteristic which correlated significantly (p less than .05) was the number of bizarre multinucleate giant cells. When the 11 non-tumorigenic lines were ranked by these 2 criteria, 1 non-tumorigenic line (U-251 MGsp) had greater than .95 predicted probability of tumorigenicity. Although further tumorigenicity testing may increase the number of tumorigenic lines, the lines with few "malignant" characteristics may correspond to the population resembling cells of low grade astrocytomas seen within glioblastomas. The histologic pleomorphism of human gliomas is reflected in their morphologic and cell biologic diversity in culture.

Growth and chemotherapeutic response in athymic mice of tumors arising from human glioma-derived cell lines

Fifteen permanent cell lines derived from human gliomas were subcutaneously transplanted into athymic nude mice (nu/nu genotype, NIH Swiss and BALB/c backgrounds). Four were tumorigenic. Three of the four (D-54 MG, U-118 MG, and U-251 MG) produced progressively growing, solid, noncystic tumors. Subcutaneous volume measurement of these tumors, which correlated directly with tumor weight, was a reliable method for monitoring growth. All three cell lines which produced progressively growing subcutaneous tumors were also tumorigenic when cells were inoculated intracerebrally. These grew as well-circumscribed, intraparenchymal brain tumors. After initial implantation, each of the progressively growing, solid, subcutaneous tumors was histologically similar to the permanent cell lines from which it was derived. Tumors could be reliably passed, and stabilization of latency periods and growth rates developed. Tumors became morphologically less distinct in later passages, though some individual features remained. Mice bearing subcutaneous tumors from each of these cell lines were treated with a single ip dose of 25 mg/kg BCNU and compared to controls receiving only drug vehicle. A significant, but different, amount of reduction in tumor mass occurred among each of the three tumor lines. This model allows cell lines derived from human gliomas to be grown in animal hosts, thereby providing a potential means for evaluating growth parameters and chemotherapeutic responsiveness of tumors derived from individual human gliomas or cell lines.

Induction of lethal experimental allergic encephalomyelitis in nonhuman primates and guinea pigs with human glioblastoma multiforme tissue

✓ The introduction of active specific immunotherapy as an adjunct to conventional therapy of the brain-tumor patient creates the risk of the concomitant induction of experimental allergic encephalomyelitis (EAE). The lack of resolution concerning the total group of central nervous system (CNS) antigens which may be encephalitogenic, and the lack of definition of the necessary conditions for the induction of an anti-CNS myelin response complicate the design of an immunotherapeutic regimen for brain-tumor patients. We report here the ready induction of EAE in four of four guinea pigs and both of two nonhuman primates (Macaca fascicularis) with human glioblastoma multiforme (GBM) tissue injected with either complete or incomplete Freund's adjuvant (CFA, IFA). Immunization protocols utilizing encephalitogenic GBM tissue and adjuvant which did not result in EAE induction were established in both of two macaques, and the production of significant levels of antibodies specifically reactive with immunizing GBM-derived cultured cell lines in all of 12 macaques without EAE induction was demonstrated. As the lower detection limit of the sodium dodecyl sulfate-polyacrylimide gel electrophoresis (SDS-PAGE) assay for human myelin basic protein (HBP) was 0.6 µg HBP/gel, and an extract prepared from WR-GBM tumor tissue contained less than 0.6 µg of detectable HBP/25 µg of pH 3 extractable protein, and as 100 to 1000 µg of purified human basic protein (HBP) failed to induce EAE in three of three macaques, it was hypothesized that 1) GBM tissue may act as an adjuvant and markedly lower myelin basic protein (MBP) threshold doses for EAE induction, that 2) MBP encephalitogenic fragments capable of EAE induction may be present in GBM tissue but difficult to quantitate in precipitates by in vitro methods, or that 3) secondary encephalitogenic antigens unrelated to MBP may be present in GBM tissue. The threat of EAE induction and the potential difficulty of its detection in the deteriorating brain-tumor patient receiving active specific immunotherapy warrants a biological screen in immunizing CNS material in experimental animals prior to administration to patients in immunotherapy protocols.

Heterogeneity of Genotypic and phenotypic characteristics of fifteen permanent cell lines derived from human gliomas

Six new permanent cell lines were established from human gliomas and compared to nine other cell lines from human gliomas. All fifteen lines had individually distinct HLA phenotypes and all but two, which were from a black patient, had type B glucose-6-phosphate-de;hydrogenase isoenzymes. Morphologically, the lines could be classified into four patterns descriptively designated as fibroblastic, fascicular, epithelial, or glial. Four of the lines grew progressively and could be serially transplanted when injected into athymic mice; two others grew initially and then regressed. From none to 100% of cells developed elongated tapering processes and showed reduction in nuclear-cytoplasmic ratio in the presence of 1 mM cyclic AMP and theophylline. Levels of 2'-3' cyclic nucleotide 3'-phosphohydrolase activity ranged from nondetectable to 12.78 +/- 1.49 micromoles 2' AMP formed per hr mgm total protein. None of the lines had detectable S-100 protein, but two had readily demonstrable glial fibrillary acidic protein in indirect immunofluorescence. Fibronectin levels in spent culture supernatants ranged from undetectable levels to 21.4 micrograms/ml/10(5) cells. All but one line shared surface antigens with normal human adult or fetal brain, as detected in absorption analyses with nonhuman primate antiserum raised against glioblastoma multiforme tissue or cell line U-251 MG. Although there were many common properties of the lines, each line had a unique profile of the parameters evaluated. This heterogeneity most likely reflects the individuality of the tumors of origin and individual genotypes and capacity for a range of phenotypic expression of cells.

Endothelial cell surface expression and binding of factor VIII/von Willebrand factor

Factor VIII/von Willebrand factor (FVIII/vWF), a glycoprotein molecular complex found in human plasma, has been demonstrated by cell membrane fluorescence to be present on the surface of cultured human umbilical cord vein endothelial cells. The endothelial nature of these cells was established by electron-microscopic studies that revealed the presence of Weibel-Palade bodies in virtually all cells cultured. A newly developed radioreceptor assay was used to detect FVIII/vWF in the medium taken from these endothelial cell cultures; FVIII/vWF concentration in the medium samples increased with time in culture. FVIII/vWF binding studies showed no significant FVIII/vWF-specific binding to endothelial cell surfaces and did not corroborate a previous report suggesting a FVIII/vWF-specific receptor on human umbilical cord vein endothelium. The presence of FVIII/vWF on endothelial cell membranes and the lack of receptor-mediated binding suggests that the FVIII/vWF either has been absorbed non-specifically to the cell surface or is an integral part of the endothelial cell membrane.

The morphologic response of cell lines derived from human gliomas to dibutyryl adenosine 3':5'-cyclic monophosphate

The addition of exogenous cyclic nucleotides to cultured neoplastic cells has been reported by others to cause changes in growth properties and cellular morphology. We have studied the in vitro morphologic response to exogeneous dibutyryl adenosine 3':5'-cyclic monophosphate (DBcAMP) of twelve permanent cell lines derived from human gliomas, thirteen sublines and clonal lines derived from human gliomas, and nine cell lines derived from neoplastic and non-neoplastic human tissue from sources other than gliomas. Replicate samples from each cell line were evaluated during log phase growth by counting cells with cytoplasmic extensions greater than two times the diameter of the cell body and at saturation density for the percentage of cells morphologically responding to DBcAMP. The mean percentage of glioma-derived cells responding ranged from 5.0 to 97.3 percent during log phase growth and from 0 to 100 percent at saturation density; from 0 to 95.0 percent of non-glioma derived cells responded. The percentages of responding cells from gliomas and controls for any given line was reproducible within and between triplicate observations. Further evaluation of the twelve permanent human glioma-derived cell lines during log phase growth and at confluence delineated two clearly separate groups of cell lines. During log phase growth, eight of twelve lines had between 19.7 and 88.0 percent responding cells. At confluence, these same cell lines had between 25 and 100 percent responding cells. The other four cell lines had, respectively, 5.7 to 7.7 percent and 0 to 10 percent responding cells. The spectrum of percentage of responding cells and the variable nature of the response elicited provide evidence for the heterogeneity of the cell populations present within human glioma-derived cultured cell lines. The in vitro morphologic response to DBcAMP was reproducible and could be quantitated, but the nature and mechanism of the response to DBcAMP, whether "toxic," "differentiating," or otherwise, could not be determined from these studies.

Treatment of autochthonous rat brain tumors with fractionated radiotherapy. The effects of graded radiation doses and of combined therapy with BCNU or steroids

The avian sarcoma virus-induced glioma model in rats was used to evaluate radiation dose response (survival curves), giving fractionated treatments to the whole head. Groups of 14 to 20 rats received total doses of 2300, 4600, or 5750 rads over 2, 4, or 5 weeks, respectively. Median group survival times were compared to controls and to each other. All doses of radiation significantly prolonged survival with reference to control. The 4600- and 5750-rad doses were significantly more effective than the 2300-rad dose. In another series of experiments, BCNU chemotherapy (10 mg/kg) was combined with 2300 and 4600 rads radiation therapy. Synergism of therapies was demonstrated. Methylprednisolone acetate (2 mg/kg twice weekly over 4 weeks) alone did not affect survival curves, and its combination with 4600 rads radiation therapy negated the prolongation of survival achieved with 4600 rads alone.

Experimental gliomas: an autoradiographic study of the endothelial component

Autochthonous gliomas were induced in rats by intracerebral inoculation of avian of avian sarcoma virus and studied by 3H-thymidine autoradiography. Parenchymal glial tumor cells had a 3H-labeling index (LI) of 3.0 to 13.6%. Endothelial cells in tumor blood vessels had an LI of 2.6 to 34.3%, independent of and in most instances higher than the LI of the glial tumor. Endothelial cells of normal blood vessels had an average LI of 0.3%. This study documents the high proliferative rate of the endothelial cells in anaplastic experimental gliomas, and emphasizes the necessity for seeking direct, incontrovertible evidence to determine whether or not the rapidly proliferating endothelial cells are malignant.