Morphologic effect of dimethyl sulfoxide on the blood-brain barrier.

Dimethyl sulfoxide (DMSO) opens the blood-brain barrier of mice to the enzymatic tracer horseradish peroxidase. A single injection of horseradish peroxidase in 10 to 15 percent DMSO into the tail vein along with 10 to 15 percent DMSO delivered intraperitoneally allowed horseradish peroxidase to fill the extracellular clefts throughout the brain within 2 hours. In the absence of DMSO, peroxidase failed to enter brain parenchyma except through the circumventricular organs. Opening of the blood-brain barrier by DMSO is reversible. Dimethyl sulfoxide stimulated the pinocytosis of horseradish peroxidase by the cerebral endothelium; the peroxidase was then directed to lysosomal dense bodies for degradation. Vesicular transport of horseradish peroxidase from the luminal to the abluminal wall of the endothelial cell was not observed. Dimethyl sulfoxide did not alter the morphology of endothelial cells or brain parenchyma.

Intravenous recombinant interferon beta in patients with recurrent malignant gliomas: a phase I/II study.

A multicenter phase I/II trial of a human recombinant interferon beta (Betaseron; Triton Biosciences, Alameda, CA) was conducted in patients with recurrent glioblastoma and anaplastic astrocytoma in six centers between 1986 and 1988. Betaseron was given intravenously three times per week, starting at 90 x 10(6) IU per dose and escalating by 90 x 10(6) IU every 2 weeks up to a maximum dose of 540 x 10(6) per treatment. All patients had failed prior radiotherapy, and most had failed one or more courses of chemotherapy. Of the 72 patients entered into the protocol, 65 were considered assessable. Of 65 patients, 41 had glioblastoma, and 24 had anaplastic astrocytoma. Of the 65 assessable patients, 15 (23%) had an objective response (R), and 18 (28%) had stable disease (S), with a combined R and S rate of 51%. The Kaplan-Meier median time to progression was 24 weeks for the responders, 10 weeks for the nonresponders, and 23 weeks for the whole group. These results suggest that Betaseron has definite activity in recurrent gliomas, with an R + S rate of 51%. The maximum-tolerated dose (MTD) is between 180 and 360 x 10(6) IU, with neurotoxicity being the most troublesome toxicity at higher doses. Two patients died of treatment-related complication. Since most responders showed responses at the 180 x 10(6 IU dose range, further studies using a lower dose of Betaseron aimed at decreasing toxicity and allowing chronic maintenance therapy are merited.

Cyclophosphamide plasma and cerebrospinal fluid kinetics with and without dimethyl sulfoxide.

Ten patients with brain tumors and indwelling ventricular reservoirs were pretreated with 5% to 10% dimethyl sulfoxide (DMSO) (intravenous, oral, or both) and were then treated with 1.0 to 1.25 gm/m2 cyclophosphamide (CYC). All patients were also on anticonvulsants and dexamethasone. CYC and alkylating activity (alk act) in plasma and concomitant ventricular cerebrospinal fluid (CSF) were measured by gas chromatography and p-nitrobenzyl pyridine assay. CYC entered the CSF without difficulty and was lost from CSF more slowly than from plasma. Alk act did not enter CSF as well as did CYC. DMSO did not alter any measured aspect of CYC or alk act disposition. Specifically, it did not alter the CYC plasma half-life (t1/2), CSF t1/2, peak CSF: peak plasma CYC concentration ratio, or the urinary excretion of CYC. DMSO did not alter the plasma t1/2 or urinary excretion of alk act or the peak CSF:peak plasma concentration ratio of alk act. Our data show reduced plasma t1/2 of CYC and increased plasma and urinary alk act. This may reflect tht effect of long-term therapy with anticonvulsants or steroids.

Avenues for entry of peripherally administered protein to the central nervous system in mouse, rat, and squirrel monkey.

Pathways traversed by peripherally administered protein tracers for entry to the mammalian brain were investigated by light and electron microscopy. Native horseradish peroxidase (HRP) and wheat germ agglutinin (WGA) conjugated to peroxidase were administered intranasally, intravenously, or intraventricularly to mice; native HRP was delivered intranasally or intravenously to rats and squirrel monkeys. Unlike WGA-HRP, native HRP administered intranasally passed freely through intercellular junctions of the olfactory epithelia to reach the olfactory bulbs of the CNS extracellularly within 45-90 minutes in all species. The olfactory epithelium labeled with intravenously delivered HRP, which readily escaped vasculature supplying this epithelium. Blood-borne peroxidase also exited fenestrated vessels of the dura mater and circumventricular organs. This HRP in the mouse, but not in the other species, passed from the dura mater through patent intercellular junctions within the arachnoid mater; in time, peroxidase reaction product in the mouse brain was associated with the pial surface, the Virchow-Robin spaces of vessels penetrating the pial surface, perivascular clefts, and with phagocytic pericytes located on the abluminal surface of superficial and deep cerebral microvasculature. Blood-borne HRP was endocytosed avidly at the luminal face of the cerebral endothelium in all species. WGA-HRP and native HRP delivered intraventricularly to the mouse were not endocytosed appreciably at the abluminal surface of the endothelium; hence, the endocytosis of protein and internalization of cell surface membrane within the cerebral endothelium are vectorial. The low to non-existent endocytic activity and internalization of membrane from the abluminal endothelial surface suggests that vesicular transport through the cerebral endothelium from blood to brain and from brain to blood does not occur. The extracellular pathways through which probe molecules enter the mammalian brain offer potential routes of passage for blood-borne and air-borne toxic, carcinogenic, infectious, and neurotoxic agents and addictive drugs, and for the delivery of chemotherapeutic agents to combat CNS infections and deficiency states. Methodological considerations are discussed for the interpretation of data derived from application of peroxidase to study the blood-brain barrier.

Angioarchitecture of the CNS, pituitary gland, and intracerebral grafts revealed with peroxidase cytochemistry.

Blood vessels of the fetal, neonatal, and adult subprimate and primate CNS, including circumventricular organs (e.g., median eminence, pituitary gland, etc.), and of solid CNS and nonneural (anterior pituitary gland) allografts placed within brains of adult mammalian hosts were visualized with peroxidase cytochemistry applied in three ways: to tissues from animals injected systemically with native horseradish peroxidase (HRP) or peroxidase conjugated to the lectin wheat germ agglutinin (WGA) prior to perfusion fixation; to tissues from animals infused with native HRP into the aorta subsequent to perfusion fixation; and to tissues from animals fixed by immersion and incubated for endogenous peroxidase activity in red cells retained within blood vessels. In neonatal and adult animals receiving native HRP intravascularly, non-fenestrated vessels contributing to a blood-brain barrier were outlined with HRP reaction product when tetramethylbenzidine (TMB) as opposed to diaminobenzidine (DAB) was used as the chromogen; fenestrated vessels of circumventricular organs were not discernible due to the density of extravascular reaction product. Fenestrated and non-fenestrated cerebral and extracerebral blood vessels exposed to bloodborne WGA-HRP were visible when incubated in TMB and DAB solutions. Native HRP infused into the aorta of fixed animals likewise labeled non- fenestrated vessels throughout the brain upon exposure to TMB or DAB but obscured fenestrated vessels of the circumventricular organs. Endogenous peroxidase activity of red cells, seen equally well with TMB and DAB, outlined blood vessels throughout the cerebral gray and white matter and all circumventricular organs in fetal, neonatal, and adult animals. Application of the three peroxidase cytochemical approaches to study the development or absence of a blood-brain barrier in intracerebral allografts demonstrated that the vascularization of day 16-19 fetal/1 day neonatal CNS allografts is not well defined prior to 7 days following intracerebral placement of the grafts. CNS allografts secured from donor sites expected to possess a blood-brain barrier exhibited blood vessels that were not leaky to HRP injected intravenously in the host. Fenestrated blood vessels associated with anterior pituitary allografts were evident prior to 3 days posttransplantation within the host brain and permitted blood-borne HRP in the host to enter the graft and surrounding host brain parenchyma.

Cytochemical localization of glucose-6-phosphatase activity in cerebral endothelial cells.

Glucose-6-phosphatase (G6Pase) activity, with glucose-6-phosphate and mannose-6-phosphate as substrates, was examined by cytochemistry in capillary and arteriole endothelial cells of the mouse brain. G6Pase activity was observed ultrastructurally in the lumen of the nuclear envelope and endoplasmic reticulum (ER) of these cells. The reactive ER and nuclear membrane appeared to be in continuity. Nucleoside diphosphatase activity, also a marker for the ER in some cell types, was not seen within the ER of the cerebral microvasculature. The ER of arterioles and capillaries did not bind lead nonspecifically when incubated in a substrate-free medium. Speculation is raised concerning the involvement of G6Pase in glucose metabolism of cerebral endothelial cells and in making blood-borne glucose available to brain parenchyma.

Brain tumors and the geriatric patient.

The prognosis, site of occurrence, and histologic type of primary brain tumors are age-dependent phenomena. In general, the incidence of meningiomas, acoustic Schwannomas, and glioblastomas increases with advancing age until the end of the eighth decade. Of 99 patients consecutively admitted to an aggressive multimodality treatment program for glioblastoma multiforme, 18 per cent were in the 61-70 age group and 4 per cent in the 71-80 group; the oldest was 85. The operative mortality was only 4 per cent. In 16 patients over 65, the 6- and 12-month calculated survival probabilities were 0.65 and 0.31, respectively. The Kaplan-Meier survival curve for these patients was significantly different from that for 26 patients under the age of 40. Grade 4 astrocytomas were present in 62 per cent of patients under 40 but in 83 per cent of patients over 61. In all glioblastoma populations, age is the most significant prognostic variable. The incidence of metastic brain tumors also increases with age, and all of the usual primary sites are represented. The prognosis for elderly patients with metastatic brain tumor is uniformly worse than that for younger patients, even though modern diagnostic and operative techniques carry virtually the same morbidity and mortality rates in older patients as in younger ones. It is vitally important, therefore, that the clinical effects of treatable intracranial tumors in the elderly are not ascribed to dementia, the aging process, the systemic effects of cancer, or the side effects of cancer therapy, without suitable diagnostic investigation.

Aggressive multimodality therapy based on a multicompartmental model of glioblastoma.

Glioblastoma multiforme is composed of multiple cellular compartments with different morphologic, kinetic, metabolic, vascular, and genetic properties. Optimal therapy may consist of a variety of therapeutic strategies designed for individual compartments, administered in close temporal relation. These concepts may turn out to be valid for other solid tumors as well. Microwave-induced hyperthermia can be used to treat metabolically quiescent, relatively hypoxic, nondividing cells (Go) otherwise resistant to radiation and chemotherapy. Similarly, polychemotherapy can treat a broad spectrum of cell types if the blood-brain barrier can be circumvented. Radical surgery, repetitively applied, can be safely used to "set up" experimental agents if the operation microscope and laser are employed. A consecutive series of 74 adult patients with malignant astrocytoma were treated with primary resection, radiation therapy, and 1,3,-bis(2 chloroethyl) 1 nitrosourea chemotherapy. At recurrence, all patients were offered reoperation with the microscope and the laser prior to administration of phase-I agents--hyperthermia via an implantable miniature microwave antenna (6 cases); aziridinylbenzoquinone chemotherapy (13 cases); and blood-brain barrier reversal with dimethyl sulfoxide (DMSO) and polychemotherapy (9 cases). It was concluded that temperatures of 45 degrees C could be safely achieved and human tumors could not efficiently dissipate heat; that DMSO plus drug therapy could be tolerated but blood-brain barrier reversal demonstrated by us in animals could not be shown in humans; and that aggressive multimodality therapy and reoperation could produce a 40% 2-year survival rate for patients younger than 40 years.

Surgical management of lung cancer with solitary cerebral metastasis.

Between 1964 and 1986, 19 patients underwent resection of both a primary lung cancer and the associated brain metastasis. One patient underwent resection of 2 separate primary lung cancers and the associated metastases. The 12 men and 7 women ranged in age from 42 to 67 years (mean, 54.6 years). The cell type was adenocarcinoma in 12 tumors, squamous or adenosquamous cell in 5, large cell undifferentiated or anaplastic in 2, and malignant carcinoid in 1 tumor. The types of resection were as follows: lobectomy for 12 neoplasms, pneumonectomy for 5, bilobectomy for 2, and wedge resection for 1 neoplasm. Radiotherapy to the brain was given in connection with sixteen of the twenty craniotomies. The patient with 2 separate primary neoplasms survived 19 years before dying 5 months after the second craniotomy. The mean survival is 8.0 +/- 2.1 years (+/- the standard error), and the median survival is 1.67 years. Survival at 1 year was 65 +/- 10.7% and at 5 years, 45 +/- 11.1%. On univariate analysis, the following factors were found to correlate significantly with longer survival: a lung tumor in Stage I or II; negative mediastinal nodes; curative rather than palliative resection of the lung tumor; and age younger than 55 years. However, on multivariate analysis, only curative resection was a significant factor (p less than 0.01). We believe these results justify continued application of this combined surgical approach to patients having limited-stage lung cancer with a solitary brain metastasis.

CT characteristics of a transplantable canine glioma model: preliminary kinetic analysis.

Experimental brain tumors can be produced in dogs through the intracerebral injection of 3 X 10(6) live tumor cells in either neonates or adult animals. Tumors are visible by computed tomography on day 8 postinjection. Most tumors appear as ring lesions with central lucencies and shaggy borders. By postinjection day 12, tumor volumes increase more than 10 times; the cell cycling time is about 1-3 days. The initial doubling time is about 1-2 days and corresponds to the in vitro doubling time of about 24 hr. The use of computed tomography to perform noninvasive kinetic analysis deserves further study. The transplantable canine glioma model would appear to be ideal for this purpose.

Correlation of absorption coefficients with intracranial fluid protein concentrations and specific gravities.

Samples of ventricular cerebrospinal fluid, tumor cyst fluid, and subdural fluid were obtained from 30 patients at operation. The protein concentration and the specific gravity of each sample were measured and the corresponding mean absorption numbers were calculated from the numerical printout of the preoperative computerized tomogram. For fluids with a specific gravity greater than 1.005, a linear relationship was demonstrated between protein concentration and specific gravity. For protein concentrations greater than 300 mg/dl, there was a linear relationship between protein concentration and the mean absorption number. As the precision of present instrumentation improves, it is expected that a noninvasive technique for estimating intracranial protein concentration will have a number of clinical application.

Probability and the brain.

Probability theory asserts the lawfulness of seemingly random events in large populations and seems to be a reasonable approach to a general understanding of the structure and function of the nervous system. The brain, by virtue of the number of its components, the multiplicity of their possible interconnections, and the range and rapidity of their outputs, is almost implausably complex in its over-all design. Probability theory, therefore, is usually applied to (a) descriptions of the behavior of large neuronal populations, (b) statistical analysis of neuronal spike trains, and (c) theoretical models of neuronal interaction. A consideration of each of these subjects is presented, as is a discussion of the most fundamental level of application of the theory to the nervous system: (d) the assertion that the neuron and/or brain is inherently nondeterministic. In practical terms this is shown to be a "nonissue," the uncertainty principle that follows has rather definite philosophical implications.

Survival in glioblastoma: historical perspective.

The accumulated retrospective experience with glioblastoma multiforme was reviewed. Data were extracted from 17 reports in the literature, comprising 2532 patients. Survival curves were constructed for 1561 selected cases that did not include limited surgery and/or astrocytoma Grade III. The median survival after operation for the 1561 patients was 6 months, and only 7.5% lived 2 years. The survival curve is exponential in shape; calculation of its rate constant as well as extrapolation on a least squares regression line for the semilogarithmic plot of the data both predict 3- and 5-year survival rates of 1.7 and 0%, respectively. Curves were also plotted for patients who were treated by operation alone (6 study groups, 349 cases), operation plus radiation (11 groups, 568 cases), and operation plus radiation plus chemotherapy (5 groups, 146 cases); the median survival times for the three groups were 4, 9.25, and 10 months, respectively. It was concluded that (a) all curves converge at 18 to 24 months, irrespective of treatment; (b) radiotherapy is the decisive treatment during the first 18 months; (c) survival can be predicted by adopting an exponential model; and (d) prospective studies are required to detect the marginal benefits of current therapies.

Intramedullary subependymoma of the cervical spinal cord: case report.

The third case of intramedullary subependymoma of the cervical spinal cord is reported. This is the first such case for which intraoperative views are available. It was possible to remove most of this lesion with the aid of the operating microscope, the carbon dioxide laser, and the use of somatosensory evoked potentials. The diagnosis was clouded by the coexistence of Holmes-Adie pupils and absent reflexes in the lower extremities. All three reported intramedullary subependymomas have been found in the cervical spinal cord of middle-aged patients. Tumors in this clinical setting should not be assumed to be unresectable astrocytomas without careful histopathological and intraoperative evaluation.

Occurrence of glioblastoma multiforme in three generations of a cancer family.

We report a cancer family in which glioblastoma multiforme occurred in each of three successive generations and strictly in the paternal lineage. The pedigree is consistent with an autosomal-dominant pattern of inheritance, and this conclusion is supported by the incidence of other tumors within the family. Ten of the 23 blood relations of the proband have developed cancers (43.5%), and 5 of these have been breast carcinomas. The clinical importance of the cancer family syndrome is discussed, as is the relevance of genetic factors in the expression of the neoplastic phenotype.

Brain stem hemorrhage: historical perspective.

Brain stem hemorrhage was first recorded in 1812. Its origin has been the subject of much study and experimentation. We review the studies that stress mechanical forces because dynamic axial distortion of the brain stem is a major factor in the physiology and pathology of the central nervous system.

Dynamic axial brain stem distortion as a mechanism in the production of brain stem hemorrhages: role of the carotid arteries.

Dynamic axial distortion of the brain stem has long been recognized as occurring in the presence of supratentorial mass lesions. Although brain stem hemorrhages occur under these circumstances, the pathophysiology of these hemorrhages is still a matter of controversy. In the present study, the role of the carotid arteries in tethering the brain stem was evaluated during dynamic axial distortion of the neuraxis of the dog. Brain stem hemorrhages occurred in 32 of 48 dogs in which the carotid arteries were left intact (67%) and in only 9 of 46 animals in which the carotid arteries were sectioned (20%). The difference between the two groups is statistically significant (chi 2 = 19.3; P less than 0.001). In this study, we can eliminate intracranial pressure as the primary cause of the lesions because a single variable is present in one series and absent in the other, whereas intracranial pressure and its duration are kept constant. It is concluded that the tethering effect of the carotid arteries is a significant factor in the production of brain stem hemorrhages in the dog by dynamic axial distortion of the brain stem. A selective review of the literature on production of brain stem lesions is provided to support these findings.

Value of sequential computed tomography in the multimodality treatment of glioblastoma multiforme.

Previous assessments of the value of sequential computed tomographic (CT) scanning in brain tumor patients have suffered from the heterogeneous nature of the study populations in regard to (a) pathology, (b) treatment plan, and (c) time of scanning. This report is based on the first 21 of 30 consecutive cases of glioblastoma multiforme entered into a cumulative high dose chemotherapy study. Each patient received a maximal surgical resection, 5800 to 6300 rads of radiation therapy, and BCNU (1,3-bis(2-chloroethyl)-1-nitrosourea) chemotherapy (beginning at 100 mg/m2/day X 3 days) every 8 to 10 weeks for the life of the patient. CT scans were obtained pre- and postoperatively at each admission for chemotherapy; the scan at the time of the first BCNU course served as the postradiotherapy scan. Edema was present on 94% of the initial scans, and 94% of the tumors demonstrated contrast enhancement. In 70 instances it was possible to compare clinical status and an enhanced scan at the time of adjuvant treatment. When the scan was improved or unchanged so was the patient (46 of 46), but worsening of the scan was accompanied by worsening of the patient only 62% of the time (15 of 24). Sequential CT scanning proved useful in (a) detecting the positive effect of a treatment plan in clinically stable patients; (b) detecting non-tumor related causes of clinical deterioration; (c) detecting early treatment failure as a prelude to reoperation and/or a change in drug protocol before clinical deterioration; and (d) detecting asymptomatic complications of the treatment plan (i.e., a 20% incidence of ventricular enlargement).

Hyperthermia for brain tumors: biophysical rationale.

Hyperthermia has great potential as an antineoplastic agent because: (a) it is effective against relatively radioresistant hypoxic cells and cells in S phase; (b) unlike most chemotherapeutic agents, it is effective against poorly vascularized and metabolically quiescent tissues; (c) as a physical agent, its biological effect is related to the duration and intensity of its application; (d) it seems to have no cumulative toxicity; and (e) it potentiates the effects of both chemotherapy and ionizing radiation at the cellular level. The use of hyperthermia for malignant brain tumors is constrained by a relatively narrow therapeutic index and the considerable thermal sensitivity of normal neural tissue. Glioblastoma multiforme, by virtue of its low growth fraction and heterogeneous cell populations, seems to be an ideal candidate for hyperthermia administered as part of a combined modality treatment program. Focal hyperthermia can be produced by a number of energy sources, including those utilizing ultrasound, microwave, and radiofrequency generators. The clinical safety and feasibility of a miniature microwave radiator/sensor system for direct implantation have been demonstrated. In comparison to normal feline brain, malignant brain tumors in humans are unable to dissipate heat efficiently.

In vitro response of human glioblastoma and canine glioma cells to hyperthermia, radiation, and chemotherapy.

Little is known about the sensitivity of human glioblastoma cells to hyperthermia alone and in combination with other therapies. We carried out in vitro cell survival studies on the human glioblastoma cell line U-87MG and our model canine glioma canine brain tumor (CBT) cells after multimodality treatment. Ionizing radiation was administered to flasks of cells in logarithmic growth at 500 rads (5 Gy) with consecutive treatment by hyperthermia, 1,3-bis-(2-chloroethyl)-1-nitrosourea (BCNU), or cisplatin. Cells were treated with single doses of BCNU at 5 microM with sequentially added radiation or hyperthermia and at 1 to 2 micrograms/ml of cisplatin with hyperthermia. Hyperthermia was administered in a precision controlled water bath at 44 degrees C for 30 minutes in combination with chemotherapy or radiation. In general, the sensitivity of U-87MG and CBT cells was similar for all test regimens. For example, colony formation efficiency decreased by 64% in CBT cells and by 64.4% in U-87MG cells after hyperthermia alone at 44 degrees C for 60 minutes. All combinations of BCNU, hyperthermia, and radiation administered in vitro produced enhanced cell killing, but the effects of multiple modalities were generally additive in both cell lines.(ABSTRACT TRUNCATED AT 250 WORDS)