Biology and treatment of gliomas

Annals of Oncology 3: 423-433, 1992. O 1992 Kluwer Academic Publishers. Printed in the Netherlands.

Review Biology and treatment of gliomas T. J. Janus,2 A. P. Kyritsis,1 A. D. Forman1 & V. A. Levin1 'Department of Neuro-Oncology, The University of Texas M.D. Anderson Cancer Center, Houston, Texas, USA; ^Department of Neurology, Division of Neuro-Oncology, The University of Iowa, Iowa City, Iowa, USA

Summary. This review discusses some of the recent advances in glioma research and treatment. Our understanding of the characteristics of these tumors has been strengthened by the application of molecular biologic and genetic techniques to pathologic grading and therapy outcome. Newer attempts to correlate imaging modalities to pathologic grading are also discussed. It is anticipated that these developments will strengthen our ability to design improved treatment strategies, an essential goal inasmuch as current treatment

schemes have limited benefit. More work needs to be done to understand the biology of these rumors especially the complex interactions of their cytokine expression, multiplicity of genetic abnormalities, and their local environment. Only then will be able to develop improved therapeutic interventions. Key words: chemotherapy, glioma, astrocytoma, glioblastoma, ependymoma, oligodendroglioma, radiotherapy, surgery, cytokines

Recently, Daumas-Duport et al. [7, 8] proposed a modification of previous grading schemes, which relied The worldwide incidence of gliomas has increased over on histologic criteria that included nuclear atypia, the last few years, most noticeably in the elderly. This is mitosis, endothelial proliferation and necrosis as differlikely due to both an increase in our ability to detect entiating features. Daumas-Duport assigned a point these malignancies with newer imaging methods and an system to these histologic variables: Grade 1 tumors absolute rise in incidence. The dismal record of success had none of the above features, Grade 2 had one feain the treatment of these tumors makes glioma research ture, Grade 3 had two features and Grade 4 had three imperative if we are to see improvement. The advent of or more features. In their initial evaluation, this groupmolecular genetics brings hope as it increases our abil- ing led to distinct median survival curves [8], but a subity to study gliomas and thus gain insight into their sequent review of 251 cases at the Massachusetts genesis, phenotypic diversity, and behavior and, ulti- General Hospital found no statistical difference in surmately, to develop new treatment strategies. vival between grade 2 and 3 [9]. The only significant In this review we will discuss areas of progress in the predictor of survival for malignant gliomas was found laboratory and clinical management over the last sev- to be necrosis, agreeing with previous grading systems eral years and relate these findings to our current clini- [10]. A three tiered system of well differentiated astrocal understanding. Past reviews provide additional cytoma, anaplastic astrocytoma and glioblastoma detailed accounts regarding the state of clinical and multiforme represents the most practical means for basic research into gliomas [1-4]. grading these tumors. Any histologic grading system will suffer if insufficient or unrepresentative biopsy material is submitted Histology for evaluation; this is especially true for gliomas, which have considerable intratumoral heterogeneity [11-14]. The pathologic grading of gliomas has been the subject This heterogeneity (reflected in the name given to the of heated debate. Various classification schemes have most malignant member of the gliomas, glioblastoma been advanced with no universal standard yet adopted. multiforme) within a given tumor plagues their evaluaThis controversy leads to continued confusion regard- tion whether by histology, immunochemistry, molecular ing the interpretation of treatment response and malig- genetics and troubles clinical evaluation by diagnostic nant growth potential as tumors are not always placed imaging techniques. into uniform and comparable categories. Classically, the system advanced by Kernohan [5] divided these Radiology neoplasms into four grades. Later grading systems combined several features into a three-grade system Even with the advances of computed tomographic which has gained increasing favor [6]. (CT) scanning and nuclear magnetic resonance imaging Introduction

424

(MRI), the main advantage of scanning is to define the tumor boundaries and distinguishing tumor from edema and blood. While some inferences about the biologic behavior of tumors may be made from these studies, accurate evaluation demands biopsy. The previously held opinion that non-contrast enhancing tumors are uniformly of a less malignant histologic phenotype has been disputed [19]. When stereotactic frames are used in concert with these powerful imaging techniques, lesions may be sampled even when their location would have made any open surgical approach fraught with hazard [15—18]. The heterogeneity of gliomas, however, makes interpretation of the small samples obtained with stereotaxic biopsy unreliable and the wise clinician should be aware of sampling errors from such biopsies. Given the advantage of tumor bulk reduction, particularly if greater than 90% removal is feasible, an open procedure is preferable when non-eloquent regions of the brain are involved. Stereotaxic biopsy at the very least is warranted in the vast majority of cases to secure a diagnosis of malignancy. Several studies have found that radiographically defined tumor margins do not correlate with histologically confirmed tumor margins [11, 15]. In an autopsy study utilizing CT correlation, histologically confirmed tumor was found outside the radiographically defined contrast enhancing tumor and peritumoral 'edema' edge [13]. Greene et al. studied CT and NMR images with CT-guided sterotactic biopsies and found, at the time of initial surgery, tumor up to 1.5 cm outside radiographically abnormal areas [15]. Current neuroradiographic techniques remain only approximate in delineating the extent of disease. Newer techniques are currently under study to improve the non-surgical evaluation of tumor infiltration. Positron emmission tomography (PET) allows the study of glioma metabolic activity by evaluation of glucose (|l8F]-fluoro-2-deoxyglucose), amino acid (|"C]methionine), or oxygen metabolism [20-22]. PET scanning has been used to correlate areas of high metabolic activity (as indicated by increased glucose or methionine uptake) with extent and grade tumor [13, 23]. Indeed some argue that PET scanning is superior to CT evaluation [20, 24]. Patronas et al., for example, found that increased uptake of [18F]-fluoro-2-deoxyglucose (FDG) by tumor correspnded to decreased survival time [25]. There are also reports that changes in intensity of FDG uptake may correspond to alterations in the biology of the tumor in vivo [22]. These physiologic studies are presently hampered by poor resolution when compared to CT or NMR. Technical difficulties such as the expense of the hardware, the need for arterial blood sampling and on site cyclotron generated reagents limit PETs present clinical applications. Single photon emission computed tomography (SPECT) with 2nithallium is another physiologic scanning method which is believed to work by exploiting differences in NaVK+ pump activity between malig-

nant and benign tissues. Studies with this technique are in their infancy but they may play a role in assessing malignancy of tumors [33] and, like PET scanning, may prove useful in identifying radiation necrosis from tumor recurrence, but without the technical problems that hamper wide clinical use of PET scanning. Other reagents such as Tc99"1 MIBI (Cardiolite) are also being assessed for use with SPECT in identifying tumor.

Proliferation markers

Conventional imaging techniques and histologic review alone are insufficient measures of tumor growth potential and outcome. Tumor cell proliferative markers are being used increasingly to predict tumor growth potential. One of the more studied markers is bromodeoxyuridine (BUdR), a halopyrimidine that competes with thymidine for tumor cell uptake and incorporation by S-phase cells. In cell kinetic studies, BUdR is infused into patients shortly before surgery and the percentage of BUdR-labeled cells in surgical specimens is evaluated using specific monoclonal antibodies to BUdR incorporated into DNA [26-28]. This allows the clinician to establish which tumors have a faster rate of cell division. Hoshino et al. [26] found that histologically more malignant appearing tumors have a higher percentage of BUdR labeled cells. Hoshino et al. [26] LaBrousse et al. [29] found that BUdR labeling may correlate better with survival than histologic grade. Other techniques are currently under study, including a silver colloid staining technique that evaluates nuclear of DNA organization [30]. A correlation has been found between nuclear regions stained by this technique, appearance of histologic malignancy, and prognosis. Burger et al. have found that higher nuclear staining of gliomas with the Kl-67 antibody seems to correspond with more malignant histologic appearance and faster growth characteristics [31]. As a bridge between the evaluation of histologic labeling and neuroradiologic imaging, on group is studying the use of [18F]-fluoro-2-deoxyuridine, which can be used both as a histologic marker and a PET scan marker [32].

Heterogeneity One difficulty in relating in vitro models of malignant gliomas to the in situ environment is the instability of cell lines [34]. Several changes in cell morphology after multiple passages have been found including changes in expression of protein receptors (platelet derived growth factor receptor (PGDF-R), epidermal growth factor receptor (EGF-R), etc.) chromosome loss (22, 10) and changes in cytokine expression (see below) [35-37]. Chromosomal changes may also affect the ability to establish a cell culture from a glioma biopsy [38-40]. The heterogeneity of these tumors also makes in

425

vivo laboratory evaluation extremely difficult. Early work found that tumors removed from patients and cloned by monolayer culture had multiple subpopulations after many cell passages, including alterations in both karyotype and histologic appearance [34]. However, cell lines with chromosomal similarity to the initial tumor cell culture were also found. Cell culture study of malignant gliomas must take into consideration their genetic instability and their changing phenotypic appearance [37, 39]. Studies also indicate that cells with wildly aberrant chromosomal numbers may be more susceptible to chemotherapy, while cells with diploid chromosome number may, in fact, be chemotherapy resistant [37]. It is possible that these 'less malignant' diploid cells are responsible for further tumor repopulation and growth after multi-modality treatment. Due to the histologically heterogeneous nature of malignant gliomas and their ability to transform phenotypic expression, studies using even the most sophisticated means of pathophysiologic analysis will never be entirely reliable in this most capricious family of tumors. The smaller samples obtained with stereotaxic biopsy favors erring the estimate of a falsely lower malignant potential [42] than would be obtained with more complete sampling obtained with open biopsy.

tion to neurons. There may be a coordinated biological interaction between GFAP, vimentin, and S-100 [51-53]. Recently, the human GFAP gene has been cloned and its chromosomal location identified as 17q21 [54, 55]. GFAP has also been found to be expressed in gliomas with low levels of fibronectin gene transcription and expression [52]. A correlation between GFAP mRNA levels and PDGFa receptor mRNA, but not PDGFp receptor mRNA, was found as well [54]. S-100, a 25 kd protein composed of two similar subunits, forms heterodimers and homodimers [56] and is believed to be a calcium-binding protein expressed in the cell nucleus. S-100 is expressed in a wide variety of CNS tumors as well as many non-central nervous system tumors, mostly schwanomas and neurofibromas. Because of S-100's wide expression in CNS tumors its usefulness lay mainly in its ability to identify the gliomatous portions of gliosarcomas and it may help to distinguish astrocytomas (GFAP and S-100 positive) from oligodendrogliomas (less strongly GFAP positive but still S-100 positive). Unfortunately, none of the glioma markers are sufficiently reliable to be used to grade gliomas, reflect differentiation, or indicate lineage of astrocytomas, oligodendrogliomas, or ependymomas. The lack of specific markers for most CNS tumors reflects their complex heterogenous biology.

Cell markers Molecular biology

The pathologic evaluation of brain tumors can frustrate the most experienced neuropathologist. The use of Little is understood regarding the generation of maligmarker proteins and immunohistochemical methods nant gliomas and the progression (if operant) from [43, 44] can assist in pathologic diagnosis, supplement- more 'benign' forms (such as the juvenile pilocytic ing traditional histologic stains and at times making un- astrocytoma) to astrocytoma and glioblastoma multiforme. Whether tumors are initiated from a normal cell necessary ultrastructural evaluations. Although tumors differ within and between each to form a malignant variant by sequential steps beginother with respect to these markers, there has been ning with the slower growing tumors and subsequently little attempt to correlate their level of expression with evolving to a glioblastoma or can be genetically altered clinical outcome [45]. The histochemical markers com- to an aggressive phenotype directly from the normal monly used at present include glial fibrillary acidic pro- cell remains to be determined [57, 58]; it seems likely to tein (GFAP), S-100 protein, synaptophysin, neuron- both scenarios are operant. The application of molecular biological techniques specific enolase, and vimentin [46, 47]. Synaptophysin and neuro-specific enolase are expressed primarily in to gliomas has greatly increased our understanding of primitive neuroectodermal tumors and have been said changes that lead to their creation. It is believed that oncogenes and tumor suppressor genes exert a balto differentiate these tumors from gliomas [43,44]. Of these marker, GFAP appears to be more specific anced control on cell growth [59-61]. For example, the for gliomas. GFAP is a 55-kd protein monomer that p53 gene system appears to suppress cell growth, while can form intrachain disulfide cross-links. It is present in aberrant forms of the gene allow cell proliferation and astrocytomas and oligodendrogliomas and less so in stimulate growth [1, 3, 62]. Several families of onconeurofibromas, medulloblastomas, and ependymomas genes may be involved in malignant glioma develop[44-48]. Its presence in primitive neuroectodermal ment. Many of the receptors for the factors believed tumors underscores the pluripotent differentiation of involved in glioma tumorigenesis [3, 4, 61, 63] such as these tumors. It is believed that GFAP is expressed in epidermal growth factor (EGF), platelet derived cells reacting to their local environment. In fact, the growth factor (PDGF), insulin growth factor (IGF-1) first isolation of GFAP was from a multiple sclerosis and fibroblast growth factor (FGF) have structural plaque, which perhaps represents a reaction to a local- homologies [64], including an extracellular binding ized insult [49]. Using antisense oligomers, Weinstein et domain, a transmembrane domain and an intracellular al. [50] found GFAP to be required for astrocytic reac- domain that possesses tyrosine kinase activity.

426 The EGF-receptor (EGF-R) protein has been found to be similar to the erb-B oncogene (vide infra). The EGF-R is overexpressed in a large number of gliomas, most of them phenotypically glioblastoma multiforme [65, 66]. This occurs through gene amplification and increased RNA transcription of the EGF-R There appear to be intra- and inter-tumoral differences [67], which suggests that EGF-R overexpression is related to a late evolution of a more malignant phenotype. The biological result of EGF-R overexpression is unknown, and EGF-R overexpression has not been found to correlate with survival. It is known that the EGF-R has an integral protein tyrosine kinase that become active with receptor stimulation [64]. This would be expected to increase protein phosphorylation and, in turn, cell division. The stimulation of the EGF-R could be via autocrine stimulation, altered protein confirmation, or increased protein production. In some cases over-expression of EGF-leads to cellular transformation, but only after receptor activation [66]. Furthermore, this amplification, while a heterogeneous response in tumors, appears to result from a loss of a specific coding sequence [68]. Recently, expression of a structurally altered EGF-R was found in a glioblastoma cell line [69]. This protein, termed pl90, appears to be structurally similar to the EGF-R and has similar phosphorylation sites. This analogue is distinct from the erb-B-oncogene. The erb-B-2 gene arises as a result of a point mutation [70] of the EGF-R gene; this alteration in the gene product leads to increased expression and cell growth [71]. Of further note is the finding that the EGF-R maps to chromosome 7, commonly overexpressed in glioblastoma. The EGF-R is bound by transforming growth factor alpha (TGFa). This protein is a single amino acid chain of molecular weight 5,600 [72]. TGFa appears to bind the EGF-R in an autocrine fashion and has been found to be present in glial tumors of various histologic appearances [73, 74]. The expression of TGFa may be the result of genetic alterations of glioma cells at an earlier stage of malignant evolution. Eventually the over-expression of EGF-R may then accelerate tumor growth. Recent work has demonstrated gene amplification of EGF-R in higher grade gliomas but not in lower grades [75]. Elevated levels of TGFa have been found in the urine of glioma patients [76] in amounts correlating with the degree of malignancy. One case of reduction of TGFa levels in the urine following cyto-reductive surgery has been reported [77]. Transforming growth factor beta (TGFP) is a 25 kd disulfide-linked protein with the novel feature that it can stimulate or inhibit cell growth and is secreted by glioblastoma cells [78, 79]. It has been shown that TGFp has suppressive effects on lymphocyte function, specifically the IL-2 mediated generation of tumorinfiltrating lymphocytes [80]. Thus TGFp may play a role in the suppression of the immune response to astrocytic tumors.

We do not yet, however, understand the activation of TGFp. In vitro it is cleaved from a protopolypeptide by the use of plasmin or treatment in an acidic environment (pH 3.1 or lower) [81, 82]. Furthermore we do not understand how this activation can occur in the CNS environment. Perhaps it is secreted by glioma cells and activated by contact with vascular endothelial cells [83]. The interaction of TGFp with the vascular system may cause significant alteration of immune responses via changes in the blood coagulation cascade [84-86]. Work has begun on blocking the production of TGFp with in vitro glioma cell cultures to test these theories [87]. TGFp has also been shown to alter the expression of the c-sis oncogene [88]. This gene encodes the P chain of platelet derived growth factor (PDGF) which is composed of two heterologous dimers (AB). Each chain has at least 80% homology with the other and heterodimers and homodimers have been found. PDGF is involved in glial evolution [89], and two receptors, PDGF-R-A and PDGF-R-B, are known. Malignant gliomas secrete the PDGF protein in all of its dimer forms (AA, BB, AB) with PDGF-A chain having the highest level of expression [73]. The half life of the PDGF-B-mRNA is not prolonged in glioma cells, indicating gene transcript stability may not play a role in glioma initiation or progression [90]. Fibroblast growth factors (FGF) are heparin-associating proteins with mitogenic and angiogenic activity. The basic form (bFGF) has been found in glioma cells along with high-affinity bFGF receptors [91, 92]. In addition, tumors of increasing malignant grade appear to possess relatively higher levels of bFGF, apparently localized to subadjacent vascular areas (93, 94]. However, other researchers have found increased levels of PDGF-A chain and PDGF-R in this same cell membrane region [95]; and further studies have demonstrated increased levels of TGFa in these areas [96]. Addition of antisense oligonucleotide primers to block transcription of bFGF were found to inhibit growth of a glioma cell culture line, but not growth of a nontransformed human glia culture line [97].

Treatment Malignant gliomas can present as a single lesion or as multifocal lesions, but rarely will either metastasize outside the CNS. The incidence of multifocal gliomas is higher than previously believed and ranges from 5% in anaplastic gliomas to approximately 13% in glioblastoma multiforme. Moreover, multifocal gliomas appear to be associated with a high incidence of secondary malignancies, indicating a genetic predisposition (A. P. Kyritsis, personal communication, 1992). The prognosis of the multifocal gliomas as a group is very poor with a median survival time of only 6 months after diagnosis. Most clinical trials are directed towards treatment of

427

the malignant forms of brain tumors, e.g. anaplastic astrocytoma and glioblastoma. It is not completely established at present if treatment of the low-grade astrocytomas improves survival. The median survival time of patients with glioblastoma multiforme is between 8 and 13 months and for anaplastic gliomas between 20 and 36 months depending on treatment [98]. Covariates that influence survival are histologic type and grade tumor, proliferative index, post-surgical tumor volume, age of patient, performance status, and type of chemotherapy used. In general, the prognosis is favorable if the patient is young, histologic necrosis is absent, the proliferative (labeling) index is low, and the tumor volume is minimal [98-103). Current therapy of malignant gliomas includes surgery, radiation therapy and chemotherapy.

Surgery and radiotherapy

Although the conclusion is still questioned, complete or near complete surgical extirpation of the tumor prior to radiotherapy or chemotherapy seems to improve survival [101-104]. From a practical viewpoint, surgical reduction of at least 90% of tumor volume (more than 1 log of cells) is required for an increase in survival attributable to surgery. Radiotherapy is one of the more effective treatment modalities for malignant gliomas. Originally it consisted of whole-brain radiation but later it became evident that this approach resulted in unnecessary CNS toxicity, since the tumor tends to recur within 3 cm of its original margins in most instances [104-108]. Whether radiation is given by whole brain or limited field portals, conventional fractionation is 1.7-2.0 Gy fractions per day,fivedays a week to a total of 60 Gy. In current practice, this treatment covers the tumor bed and a 2-3 cm margin based on CT enhancement and probably 2 cm based on MRI enhancement. Other fractionation schedules evaluated include hyperfractionation to a higher than normal total dose and accelerated fractionation to a more conventional total radiation dose in a shorter period of time. The former has been used with moderate success against brainstem gliomas [109]. However, a randomized study by the Brain Tumor Cooperative Group in 557 patients with cerebral gliomas showed no advantage of hyperfractionated radiotherapy plus carmustine over conventional radiotherapy plus carmustine [110]. Accelerated fractionation is believed to enhance the radiation effect since the tumor cells are less resistant to frequent radiation fractions than normal brain cells Another recent modification is the use of radiosensitizers or radiopotentiators in order to enhance the radiation effect. Included are hypoxic cell sensitizers such as misonidazole and fluosol, halogenated pyrimidines such as bromodeoxyuridine and iododeoxyuridine, and platinum anlogues. The jury is still out on the

benefit of these agents with radiotherapy in patients with malignant gliomas. For selected patients with recurrent malignant gliomas interstitial brachytherapy with stereotactic implantation of radioisotope seeds prolongs survival but decreases the quality of life since a high percentage of these patients become steroid dependent and need reoperation for removal of radiation necrosis [113]. As adjuvant therapy to extend beam radiotherapy, brachytherap/s benefits are limited to patients with glioblastoma multiforme [114]. In the adjuvant setting chronic steroid usage and reoperation may be responsible for of the benefit attributed to brachytherapy.

Chemotherapy

Adjuvant chemotherapy following surgery can be given before, during or after radiotherapy [115, 116]. Most reports are of chemotherapy after radiotherapy. Most of the drugs used have produced either no benefit or short-lasting response. Drug treatment failures have been attributed to poor drug delivery, a low tumor growth fraction, low immunogenicity of the tumor cells, and cellular resistance and heterogeneity. In general, drug delivery is poor in the normal brain area adjacent to the tumor (an area that exhibits microscopic tumor cell infiltration), greater in the periphery of the tumor bulk, and low in central necrotic areas. Attempts to increase drug concentration in the tumor bed using intraarterial delivery with or without blood-brain barrier disruption with osmotic agents have been disappointing: these have increased brain and retinal toxicity and produced little or unconfirmed benefits in survival. Currently there are efforts to enhance drug or radioisotope delivery to brain tumors using monoclonal antibodies as carriers. The limitation of this approach is nonspecific binding to other tissues and the immunogenicity of the antibodies, which will limit the number of times they can be used in a multicourse therapy regimens. Even polylysine conjugates of radionuclides have been considered as possible carriers to brain tumors [117]. This targeting technique is limited by concomitant high uptake to other vital organs such as kidneys, liver and spleen. The heterogeneity of malignant gliomas and selection of resistant clones is another major limiting factor in the effectiveness of chemotherapy. For instance, it has been shown that BCNU-resistant cells demonstrate amplification of the multidrug resistant gene and increase in polyamine biosynthesis [118]. There is recent evidence that TGFp1 may modulate multidrug transport in glioma cells [119]. Calcium channel blockers and calmodulin inhibitors can overcome multidrug resistance. The action of calcium channel blockers is obscure and unrelated to inhibition of calcium processes [120]. Verapamil, a calcium channel blocker, produced inhibition of glioblastoma growth in nude mice [121]. Human studies employing verapamil to augment chemother-

428

apy's effect are currently underway in several centers. Patients with gliomas have generally depressed cellmediated immunity [122], and their circulating T-lymphocytes have been found to be reduced [123]. There is evidence that the humoral responses are affected as well, but their exact implication is less clearly defined [124,125]. Peripheral blood lymphocytes from patients with gliomas produced lower levels of IL-2. However, after incubation with IL-2, the lymphocytes exhibited higher natural killer activity and strong cytotoxicity, suggesting that IL-2 may be useful in adoptive immunotherapy [126]. Chief among the drugs used for brain tumor chemotherapy are the nitrosoureas. Most effective members of this family are drugs that are generally lipid soluble, less than 400 daltons, and can cross the blood-brain barrier readily. Carmustine (BCNU) administered as a single agent after surgery and radiotherapy was the first drug to improve survival of patients with malignant gliomas [99]. The combination of lomustine (CCNU), procarbazine, and vincristine (PCV) has been reported to be superior in treating anaplastic gliomas producing a median survival time of 36 months comparing with 19 months for carmustine [127]. PCV appeared superior to carmustine, however, the increased survival was not significantly different in glioblastoma patients. Unfortunately, in most glioblastoma trials, the survival curves for all but the longest surviving 30% of patients has not changed over the past two decades of clinical trials. A large randomized trial found no advantage of streptozotocin, a 1-methyl-1-nitrosourea combined with glucose, over BCNU [110]. In another randomized study, there were no significant differences among BCNU, BCNU alternating with procarbazine, or BCNU with hydroxyurea alternating with procarbazine and VM-26 (teniposide) [128]. Simultaneous use of ifosfamide during radiation therapy followed by BCNU in a non-randomized trial of 75 patients resulted in a median survival time of 15.4 months for glioblastomas and 21.4 months for anaplastic astrocytomas [129]. Aggressive therapeutic regimens have been disappointing in malignant gliomas. The 'eight drugs in one day' chemotherapy, which involves administration of methylprednisolone, vincristine, CCNU, procarbazine, hydroxyurea, cisplatin, cytosine arabinoside and the imidazole carboxamide the same day, failed to improve survival compared with BCNU alone and in addition had increased toxicity [130]. Administration of high doses of BCNU (900-1050 mg/m2 IV) combined with autologous bone marrow transplantation [131-133] also resulted in increased pulmonary and liver toxicity with only marginal additional survival benefit. Similar disappointing results followed intracarotid administration of either BCNU or ACNU for the treatment of glioblastoma multiforme [134, 135]. Patients with recurrent or progressive malignant glioma have been shown to respond to a variety of anticancer agents but few respond with complete resolution and disappearance of radiographically demon-

strated tumor. Some patients respond to several successive and different forms of chemotherapy before succumbing to the tumor. Of the many reasons for chemotherapy failure tumor cell resistance and limitations in drug delivery are likely to be the most significant. Treatment with high-dose cyclophosphamide and vincristine of 5 patients with recurrent glioblastoma and 6 with recurrent anaplastic astrocytoma produced partial response or stable disease in 3 and 5 patients, respectively. The median time to progression (MTP) was 4 months in the glioblastoma and 28 months in the anaplastic astrocytoma groups [136J. Addition of benznidazole to CCNU produced no advantage in 42 patients with anaplastic astrocytoma [137]. Combination of trifluoperazine, a calmodulin inhibitor with bleomycin was tried against recurrent malignant glioma in phase II trial but showed no benefit [138]. Two chemotherapeutic agents, eflornithine (DFMO) and procarbazine, have shown activity against recurrent gliomas. Eflornithine, a specific polyamine inhibitor, has been tested in several trials and shown to be effective in over 50% of recurrent astrocytomas in combination with either mitoguazone or carmustine. The effect of glioblastomas was less pronounced, with only 20% of patients showing stable disease [139-140]. Eflornithine was currently tested as a single agent for recurrent gliomas. Preliminary evaluation indicated that approximately 45% of anaplastic glioma patients either responded or stablized for a median duration of nearly a year (48 weeks); results for glioblastomas were less convincing (V. A. Levin et al., submitted). Procarbazine, a methylhydrazine derivative, has been active as a single agent against recurrent glioma [141, 142]. In a recent study procarbazine produced partial response or stable disease in approximately 50% of patients. The median time to progression for responding patients was 13 months [143]. Carboplatin, a new generation heavy metal compound that inhibits DNA synthesis, had a 40% response rate in glioblastomas and 57% in anaplastic astrocytomas with median time to progression of 20 and 21 weeks, respectively [144]. A combination of mechlorethamine, vincristine and procarbazine appeared also to be active in recurrent glioma with 52% response and stable disease and median time to progression of 42 weeks [145]. Therapy with interferon-beta resulted in 50% response and stable disease in both glioblastomas and anaplastic astrocytomas with median time to progression of 18 and 16 weeks, respectively [146]. The interferon response rate is impressive, but the lack of durable responses makes its continued use problematic. Interferon-alpha with BCNU is being evaluated currently as initial post-radiotherapy chemotherapy by the North Central Cooperative Group. Finally, intratumoral infusion of chemotherapy is another approach being tested. Continuous infusion of methotrexate intratumorally via multiple catheters was tested in a pilot study and found to be technically fea-

429

sible without significant side effects [147]. Interstitial chemotherapy with biodegradable polymers containing BCNU after maximal surgery is currently being evaluated as a means of delivering high drug concentration to tumor with limited toxicity [148].

Conclusions Malignant cerebral gliomas, like most solid cancers, are variably resistant to treatment because of inadequate drug delivery, systemic toxicity, inherent tumor cell resistance to drugs, and limited immunologic surveillance. Specific considerations for gliomas are that: 1) the blood-brain barrier restricts the delivery of chemotherapy to certain areas of tumor, 2) cellular heterogeneity is widespread with prominent drug and radiation resistant clones appearing to survive treatment; and 3) defects in immune surveillance and response limit live and dead tumor cell removal. Gross total resection (greater than one log reduction) of the malignant glioma when possible has favorable effects on survival and quality of life. Radiotherapy is, in general, a beneficial form of palliation. In addition to conventional dose fractionation, schedules of hyperfractionated, accelerated hyperfractionated, brachytherapy, brachytherapy and hyperthermy, and stereotactic radiosurgery are under investigation. Chemotherapy of malignant gliomas, although failing to produce cure or substantial numbers of longterm survivors, has played an important role in the treatment of primary brain tumors. Many chemotherapeutic agents have been tried alone or in combination in malignant gliomas. BCNU has been the main drug used for both glioblastomas and anaplastic astrocytomas. Further studies will likely validate the superiority of the combination of CCNU, procarbazine, and vincristine (PCV) for the treatment of anaplastic gliomas and replace BCNU monotherapy for these tumors. In recurrent tumors, chemotherapy is less effective, although procarbazine, eflornithine, high-dose cyclophosphamide, vincristine, carboplatin, and interferonbeta are known active agents in addition to the nitrosoureas. Finally, other approaches are currently being tested to: 1) enhance drug delivery to tumor with monoclonal antibodies; 2) counteract drug resistance with use of calcium channel blockers; 3) develop drugs to block unique tumor-specific enzymes; and 4) potentiate the immune response. Intense laboratory research is currently being conducted to investigate the genetic abnormalities of malignant gliomas, the interaction of oncogenes wih suppressor genes, and the specific protein cascades that lead to oncogenesis. The first steps in understanding the transformation of glial cells to neoplastic gliomas has been taken. As we learn more about oncogenes and suppressor genes and the subsequent cas-

cades of stimulatory and inhibitory proteins and cytokines they control, we will be able to pharmacologically manipulate these interactions and slow tumor growth. Examples recently advocated are to use biologic differentiation agents [149] to induce the reversion of the malignant cell back to a more normal phenotype and the insertion of modified gene elements utilizing retroviruses that will alter glioma growth and phenotype [150]. Other factors that would help in the treatment of these tumors are an improved ability to define the extent of disease, which would allow a more exact evaluation of tumor volume, cell proliferation, and early tumor recurrence. Comparisons of histologic parameters to growth control factors and other cytokines that may alter the ability of the body to respond to these tumors may lead to further improvements in survival. At this point, the histologic grading and radiographic appearance of these tumors remains a poor approximation of their true phenotypic appearance and malignant potential. We must understand the genetic changes that allow glioma genesis and support glioma tumor growth. We do not, as yet, understand how the currently defined clinical methods of evaluation will be affected by our increasing knowledge of cellular mechanisms of these tumors and their ability to escape detection. As we learn more through laboratory research, the evaluation of these tumors through markers of cellular proteins and proliferative potential become increasingly important, while the histologic and radiographic appearance of glioma may well assume a lesser role. Eventually we should be able to design better chemotherapy programs which will stop the great suffering and loss of life due to this malignancy.

Acknowledgements

We wish to thank Bonnie Buhler for secretarial assistance and Walter Pagel for editorial assistance. References 1. Levine AJ, Momand J. Tumor suppressor genes: The p53 and retinoblastoma sensitivity genes and gene products. Biochem Biophys Acta 1990; 1032:119-36. 2. Bigner SH, Bigner MS. Cytogenetics of human brain tumors. Cancer Genet Cytogenet 1990; 47: 141-54, 3. Steck PA, Saya H. Pathways of oncogenesis in primary brain tumors. Curr Opin Oncol 1991; 3:476-84. 4. Westphal M, Herrmann HD. Growth factor biology and oncogene activation in human gliomas and their implications for specific therapeutic concepts. Neurosurgery 1989; 25 (5): 681-94. 5. Kernohan JW, Mabon RF, Suien HS et al. A simplified classification of gliomas. Proc Staff Meeting Mayo Clinic 1949; 24: 71-5. 6. Fulling KH, Nelson JS. Cerebral astrocytic neoplasms in the adult Contribution of histologic examination to the assessment of prognosis. Semin Diagn Pathol 1984; 1: 152-63. 7. Daumas-Duport C, Scheithauer BW, Kelly PJ. A histologic

430

8. 9. 10. 11. 12.

13.

14. 15. 16. 17.

18.

19. 20.

21. 22. 23.

24.

25. 26. 27.

and cytologic method for the spatial definition of gliomas. Mayo Clin Proc 1987; 62:435-49. Daumas-Duport C, Scheithauer B, O'Fallon J et al. Grading of astrocytomas, a simple and reproducible method. Cancer 1988; 62: 2152-65. Kim TS, Halliday AL, Hedley-Whyte ET et al. Correlates of survival and the Daumas-Duport grading system for astrocytomas. J Neurosurg 1991; 74: 27-37. Nelson JS, Tsukada Y, Schoenfeld D et al. Necrosis as a prognostic criterion in malignant supratentorial, astrocytic gliomas. Cancer 1983; 52: 550-4. Taratuto AL, Sevlever G, Piccardo P. Clues and pitfalls in stereotactic biopsy of the central nervous system. Arch Pathol Lab Med 1991; 115: 595-602. Kelly PJ, Daumas-Duport C, Scheithauer BW et al. Stereotactic histologic correlations of computed tomography - and magnetic resonance imaging-defined abnormalities in patients with glial neoplasms. Mayo Clin Proc 1987; 62:450-9. Halperin EC, Bentel G, Heinz ER et al. Radiation therapy treatment planning in supratentorial glioblastoma multiforme: An analysis based on post mortem topographic anatomy with CT correlations. Int J Radiat Oncol Biol Phys 1989; 17:134750. Paulus W, Peiffer J. Intratumoral histologic heterogeneity of gliomas; A quantitative study. Cancer 1989; 64 (2): 442-7. Greene GM, Hitchon PW, Schelper RL et al. Diagnostic yield in CT-guided stereotactic biopsy of gliomas. J Neurosurg 1989; 71: 494-7. Heppner GH, Miller BE. Tumor heterogeneity: Biological implications and therapeutic consequences. Cancer Metast Rev 1983; 2: 5-23. Vecht CJ, Avezaat CJJ, van Putten WLJ et al. The influence of the extent of surgery on the neurological function and survival in malignant glioma; A retrospective analysis in 243 patients; J Neurol, Neurosurg Psych 1990; 53: 466-71. Winger MJ, MacDonald DR, Cairncross GJ. Supratentorial anaplastic gliomas in adults, the prognostic importance of extent of resection and prior low-grade glioma. J Neurosurg 1989; 71: 487-93. Chamberlain MC, Murovic JA, Levin VA. Absence of contrast enhancement on CT brain scans of patients with supratentorial malignant gliomas. Neurology 1988; 38: 1371—4. Mosskin M, Von Hoist H, Bergstrom M et al. Positron emission tomography with 11 C-methionine and computed tomography of intracranial tumours compared with histopathologic examination of multiple biopsies. Acta Radiol 1987; 28 (3): 673-81. Derlon JM, Bourdet C, Bustany P et al. |"C]L-methionine uptake in gliomas. Neurosurgery 1989; 25 (5): 720-8. Rozental JM, Levine RL, Nickles RJ et al. Glucose uptake by gliomas after treatment. Arch Neurol 1989; 46: 1302-7. Mosskin M, Hindmarsh ET, Von Hoist H et al. Positron emission tomography compared with magnetic resonance imaging and computed tomography in supratentorial gliomas using multiple stereotactic biopsies as reference. Acta Radiol 1989; 30 (3): 225-32. Mosskin M, Ericson K, Hindmarsh T et al. Positron emission tomography compared with magnetic resonance imaging and computed tomography in supra tentorial gliomas using multiple stereotactic biopsies as reference. Acta Radiol 1989; 30: 225-32. Patronas NJ, di Chiro G, Kufta C et al. Prediction of survival in glioma patients by means of positron emission tomography. J Neurosurg 1985; 62:816-22. Hoshino T, Prados M, Wilson CG et al. Prognostic implications of the bromodeoxyuridine labeling index of human gliomas. J Neurosurg 1989; 71: 335-41. Nishizaki T, Orita T, Kajiwara K et al. Correlation of in vitro bromodeoxyuridine labeling index and DNA aneuploidy with survival or recurrence in brain-tumor patients. J Neurosurg 1990; 73: 396-400.

28. Cho KG, Nagashima T, Barnwell S et al. Flow cytometric determination of modal DNA population in relation to proliferative potential of human intracranial neoplasms. J Neurosurg 1988; 69: 588-92. 29. Labrousse F, Daumas-Duport C, Batorski L et al. Histological grading and bromodeoxyuridine labeling index of astrocytomas. J Neurosurg 1991; 75: 202-5. 30. Kajiwara K, Nishizaki T, Orita T et al. Silver colloid staining technique for analysis of glioma malignancy. J Neurosurg 1990; 73: 113-7. 31. Burger PC, Shibata T, Kleihues P et al. 18F-fluoro-2'-deoxyuridine as a tracer of nucleic acid metabolism in brain tumors. J Neurosurg 1990; 72: 110-3. 32. Tsurumi Y, Kameyama M, Ishiwata K et al. l8F-fluoro-2'deoxyuridine as a tracer of nucleic acid metabolism in brain tumors. J Neurosurg 1990; 72:110-3. 33. Black KL, Hawkins RA, Kim KT et al. Use of thallium-201 SPECT to quantitate malignancy grade gliomas. J Neurosurg 1989; 71: 342-6. 34. Shapiro JR, Yung WKA, Shapiro WR. Isolation, karyotype, and clonal growth of heterogeneous subpopulations of human malignant gliomas. Cancer Res 1981; 41:2349-59. 35. Nister M, Wedell B, Betsholtz C et al. Evidence of progressional changes in the human malignant glioma line U-343 MGa: Analysis of karyotype and expression of genes encoding the subunit chains of platelet-derived growth factor. Cancer Res 1987; 47:4953-60. 36. Livhtor T, Dohrmann GJ, Gurney ME. Cytokine gene expression by human gliomas. Neurosurgery 1990; 26 (5): 788-93. 37. Shapiro JR, Shapiro WR. The subpopulations and isolated cell types of freshly resected high grade human gliomas: Their influence on the tumor's evolution in vivo and behavior and therapy in vitro. Cancer Met Rev 1985; 4: 107-24. 38. Onda K, Tawaka R, Washiyama K et al. Correlation of DNA ploiding and morphological features of human glioma cell cultures with the establishment of cell lines. Acta Neuropathol (Berl) 1988; 76: 433-40. 39. Giangaspero F, Burger PC. Correlations between cytologic composition and biologic behavior in the glioblastoma multiforme; A post-mortem study of 50 cases. Cancer 1983; 52: 2320-33. 40. Pollack IF, Randall MS, Kristofik MP et al. Response of lowpassage human malignant gliomas in vitro to stimulation and selective inhibition of growth factor-mediated pathways. J Neurosurg 1991; 75: 284-93. 41. Bigner SH, Bjerkvig LOD et al. DNA content and chromosomes in permanent cultured cell lines derived from malignant human gliomas. Anal Quant Cytol Histol 1987; 9:435-44. 42. Coffey RJ, Lunsford LD, Taylor FH. Survival after stereotactic biopsy of malignant gliomas. Neurosurgery 1988; 22 (3): 465-73. 43. Bonnin JM, Rubinstein LJ. Immunohistochemistry of central nervous system tumors; Its contributions to neurosurgical diagnosis. J Neurosurg 1984; 60: 1121-33. 44. Kleihues P, Kiessling M, Janzer RC. Morphological markers in neuro-oncology. Curr Top Pathol 1988; 77: 307-38. 45. Kros JM, Van Eden CG, Stefanko SZ et al. Prognostic implications of glial fibrillary acidic protein containing cell types in oligodendrogliomas. Cancer 1990; 66 (6> 1204-12. 46. Nakopoulou L, Kerezoudi E, Thomaides T et al. An immunocytochemical comparison of glial fibrillary acidic protein, S-lOOp and vimentin in human glial tumors. J Neuro-oncol 1990; 8 (1>. 33-40. 47. Cosgrove M, Fitzgibbons PL, Sherrod A et al. Intermediate filament expression in astrocytic neoplasms. Am J Surg Pathol 1989; 13(2): 141-5. 48. Rettig WJ, Chesa PG, Beresford HR et al. Differential expression of cell surface antigens and glial fibrillary acidic protein in human astrocytoma subsets. Cancer Res 1986; 46:6406-12. 49. Eng LF, Vanderhaegen JJ, Bignam A et al. An acidic protein isolated from fibrous astrocytes. Brain Res 1971; 28: 351-4.

431 50. Weinstein DE, Shelanski ML, Liem RK. Suppression by antisense mRNA demonstrates a requirement for the glial fibrillary acidic protein in the formation of stable astrocytic processes in response to neurons. J Cell Biol 1991; 112 (6): 1205-13. 51. Backhovens H, Gheuens J, Siegers H. Expression of glial fibrillary acidic protein in rat C6 glioma relates to vimentin and is independent of cell-cell contact. J Neurochem 1987; 49 (2): 348-54. 52. Paetau A. Glial fibrillary acidic protein, vimentin and fibronectin in primary cultures of human glioma and fetal brain. Acta Neuropathol (Berl) 1988; 75 (5): 448-55. 53. Kimura T, Budka H, Soler-Federsppiel S. An immunocytochemical comparison of the glia-associated proteins glial fibrillary acidic protein (GFAP) and S-100 protein (S100P) in human brain tumors. Clin Neuropathol 1986; 5 (1): 21-7. 54. Bongcam-Rudloff E, Nister M, Betsholtz C et al. Human glial fibrillary acidic protein: Complementary DNA cloning, chromosome localization, and messenger RNA expression in human glioma cell lines of various phenotypes. Cancer Res 1991; 51 (5): 1553-60. 55. Reeves SA, Helman LJ, Allison A et al. Molecular cloning and primary structure of human glial fibrillary acidic protein. Proc Natl Acad Sci USA 1989; 86: 5178-82. 56. Donato R. S-100 Proteins: Cell Calcium 1986; 7:123-45. 57. Riddler LD, Calliauw L. Invasion of human brain tumors in vitro: Relationship to clinical evolution. J Neurosurg 1990; 72: 589-93. 58. Cairncross JG. The biology of astrocytoma: Lessons learned from chronic myelogenous leukemia - hypothesis. Journal of Neuro-Oncol 1987; 5:99-104. 59. Weingerg RG. Positive and negative controls on cell growth. Biochemistry 1989; 28 (21): 8264-9. 60. Wang JL, Hsu YM. Negative regulators of cell growth. TIBS 1986; 11: 24-6. 61. Bigner SH, Burger PC, Wong AJ et al. Gene amplification in malignant human gliomas: Clinical and histopathologic aspects. J Neuropathol Exp Neurol 1988; 47 (3): 191-205. 62. Hishi T, Saya H. Neurofibromatosis 1 (NF1) gene: Implication in neuroectodermal differentiation and genesis of brain tumors. Cancer Met Rev (in press 1992). 63. Maruno M, Kovach JS, Kelly PJ et al. Transforming growth factor-a epidermal growth factor receptor, and proliferating potential in benign and malignant gliomas. J Neurosurg 1991; 75: 97-102. 64. Cooper G. Oncogenes. Jones and Bartlett Publishers: Boston, Massachusetts 1991; pi 79. 65. Liberman TA, Nusbaum HR, Razon HR et al. Amplification, enhanced expression and possible rearrangement of EGF receptor gene in primary human brain tumours of glial origin. Nature (Lond.) 1985; 313:144-7. 66. Wong AJ, Bigner SH, Bigner DD et al. Increased expression of the epidermal growth factor gene is invariably associated with gene amplification. Proc Natl Acad Sci USA 1987; 84: 6899903. 67. Strommer K, Hamou MF, Diggelmann H et al. Cellular and tumoural heterogeneity of EGFR gene amplification in human malignant gliomas. Acta-Neurochir (Wien) 1990; 107 (3-4): 82-7. 68. Sugawa N, Ekstrand AJ, James DC et al. Identical splicing of aberrant epidermal growth factor receptor transcripts from amplified rearranged genes in human glioblastomas. Proc Natl Acad Sci USA 1990; 87: 8602-6. 69. Steck PA, Lee P, Hung MC et al. Expression of an altered epidermal growth factor receptor by human glioblastoma cells. Cancer Res 1988; 48:5433-9. 70. Hung MC, Yan DH, Zhao X. Amplification of the proto-neu oncogene facilitates oncogenic activation by a single point mutation. Proc Natl Acad Sci USA 1989; 86: 2545-8. 71. Gerosa MA, Talarico D, Fognani C et al. Overexpression of N-ras oncogene and epidermal growth factor receptor

72. 73.

74. 75.

76. 77.

78. 79.

80. 81. 82. 83.

84.

85. 86. 87.

88.

89. 90. 91. 92.

gene in human glioblastoma. J Natl Cancer Inst 1989; 81 (1), 63-7. Derynck R, Roberts AB, Eaton DH et al. Human transforming growth factor-a: Precursor sequence, gene structure, and heterologous expression. Cancer Cells 1991; 3: 79-86. Nister M, Libermann TA, Betsholtz C et al. Expression of messenger RNAs for platelet-derived growth factor and transforming growth factor-a and their receptors in human malignant glioma cell lines. Cancer Res 1988; 48: 3910-8. Samuels V, Barrett JM, Bockman S et al. Immunocytochemical study of transforming growth factor expression in benign and malignant gliomas. Am J Pathol 1989; 134 (4> 895-902. Ekstrand AJ, James CD, Cavenee WK et al. Genes for epidermal growth factor receptor, transforming growth factor alpha, and epidermal growth factor and their expression in human gliomas in vivo. Cancer Res 1991; 58 (8): 2164-72. Kanno H, Kuwabara T, Yasumitsu H et al. Transforming growth factors in urine from patients with primary brain tumors. J Neurosurg 1988; 68 (5): 775-80. Stromberg K, Hudgins WR, Dorman LS et al. Human brain tumor-associated urinary high molecular weight transforming growth factor A high molecular weight form of epidermal growth factor. Cancer Res 1987; 47 (4> 1190-6. Sporn MB, Roberts AB, Wakefield LM et al. Transforming growth factor-p": Biological function and chemical structure. Science 1986; 233: 532-4. Bodmer S, Strommer K, Frei K, Siepl C, de Tribolet N, Heid I, Fontana A. Immunosuppression and transforming growth factor-p in glioblastoma: Preferential production of transforming growth factor-f$2.J Immunol 1989; 143 (10>. 3222-9. Kuppner MC, Hamou MF, Sawamura Y et al. Inhibition of lymphocyte function by glioblastoma-derived transforming growth factor-p. J Neurosurg 1989; 71: 211-7. Lyons RM, Keski-Oja J, Moses HL. Proteolytic activation of latent transforming growth factor-p from fibroblast-conditioned medium. J Cell Biol 1988; 106:1659-65. Weidenham KM, Campbell WG Jr. Perineural cell tumor. Virchows Arch 1986; 408: 375-83. Antonelli-Orlidge A, Saunders KB, Smith SR et al. An activated form of transforming growth factor-p is produced by cocultures of endothelial cells and pericytes. Cell Biol 1989; 86: 4544-8. Helseth E, Dalen A, Unsgaard G et al. Type beta transforming growth factor and epidermal growth factor suppress the plasminogen activator activity in a human glioblastoma cell line. Journal of Neuro-Oncology 1988; 6: 277-83. Sawaya R, Ramo OJ, Shi ML et al. Biological significance of tissue plasminogen activator content in brain tumors. J Neurosurg 1991; 74:480-6. Sawaya R. The fibrinolytic enzymes in the biology of brain tumors. Fibrinolysis and the Central Nervous System. Hanley & BeL (US, INc): Philadelphia, Pennsylvania 1990; p 102-26. Behl C, Apfel R, Schlingensiepen KH et al. Transforming growth factor-p-antisense-oligonucleotides inhibit a human melanoma cell line under serum-enriched and stimulate under serum-free culture conditions (Abstract 2541) Cancer Res 1991; 32:427. Press RD, Misra A, Gillaspy G et al. Control of the expression of c-sis mRNA in human glioblastoma cells by phorbol ester and transforming growth factor-p. Cancer Res 1989; 49: 2914-20. Raff MC, Lillien LE, Richardson WD et al. Platelet-derived growth factor from astrocytes drives the clock the times oligodendrocyte development in culture. Nature 1988; 333: 562-5. Press RD, Samols D, Goldthwait DA. Expression and stability of c-sis mRNA in human glioblastoma cells. Biochemistry 1988; 27 (15): 5736-41. Gross JL, Morrison RS, Eidsvoog K et al. Basic fibroblast growth factor A potential autocrine regulator of human glioma cell growth. J Neurosci Res 1990; 27 (4): 689-96. Morrison RS, Gross JL, Herblin WF et al. Basic fibroblast

432

93.

94. 95.

96.

97.

98. 99. 100. 101. 102.

103.

104. 105. 106. 107. 108. 109. 110.

111. 112. 113.

growth factor-like activity and receptors are expressed in a human glioma cell line. Cancer Res 1990; 50: 2524-9. Paulus W, Grothe C, Sensenbrenner M et al. Localization of basic fibroblast growth factor, a mitogen and angiogenic factor in human brain tumors. Acta-Neuropathol (Berl) 1990; 79 (4): 418-23. Courty J, Dauchel MC, Mereau A et al. Presence of basic fibroblast growth factor receptors in bovine brain membranes. J Biol Chem 1988; 263 (23): 11217-20. Hermansson M, Nister M, Betsholtz C et al. Endothelial cell hyperplasia in human glioblastoma: Coexpression of mRNA for platelet-derived growth factor (PDGF) B chain and PDGF receptor suggests autocrine growth stimulation. Proc Natl Acad Sci USA 1988; 85 (20): 7748-52. Maxwell M, Naber SP, Wolfe HJ et al. Expression of angiogenic growth factor genes in primary human astrocytomas may contribute to their growth and progression. Cancer Res 1991; 51 (4> 1345-51. Morrison RS. Suppression of basic fibroblast growth factor expression by antisense, oligodeoxynucleotides inhibits the growth of transformed human astrocytes. J Biol Chem 1991; 266 (2): 728-34. Walker MD, Green SN, Byar DP et al. Randomized comparisons of radiotherapy and nitrosoureas for malignant glioma after surgery. N Engl J Med 1980; 303:1323-9. Levin VA. Chemotherapy of primary brain tumors. Neurol Clin 1985; 3: 855-66. Nazzaro JM, Neuwelt EA. The role of surgery in the management of supratentorial intermediate and high-grade astrocytomas in adults. J Neurosurg 1990; 73: 331-44. Levin VA, Hofiman W, Heilbron DC et al. Prognostic significance of the pretreatment CT scan on time to progression for patients with malignant gliomas. J Neurosurg 1980; 52: 642-7. Wood JR, Green SB, Shapiro WR. The prognostic importance of tumor size in malignant gliomas: A computed tomographic scan study by the Brain Tumor Cooperative Group. J Clin Oncol 1988; 6: 338-43. Ammirati M, Vick N, Liao Y et al. Effect of the extend of surgical resection on survival and quality of life in patients with supratentorial glioblastomas and anaplastic astrocytomas. Neurosurgery 1987; 21: 201-6. Salazar OM, Rubin P. The spread of glioblastoma multiforme as a determining factor in the radiation treated volume. Int J Radiat Oncol Biol Phys 1976; 1:627-37. Shibamoto Y, Yamashita J, Takahashi M et al. Supratentorial malignant glioma: An analysis of radiation therapy in 178 cases. Radiother Oncol 1990; 18:9-17. Hochberg FH, Pmit A. Assumptions in the radiotherapy of glioblastoma. Neurology 1980; 30: 907-11. Scanlon PW, Taylor WF. Radiotherapy of intracranial astrocytoma: Analysis of 417 cases treated from 1960 through 1969. Neurosurgery 1979; 5: 301-8. Wallner KE, Galicich JH, Krol G et al. Patterns of failure following treatment for glioblastoma multiforme and anaplastic astrocytoma. Int J Radiat Oncol Biol Phys 1989; 16: 1405-9. Edwards MSB, Wara WM, Uratsun RC et al. Hyperfractionated radiation therapy for brainstem glioma. A phase I-II trial. J Neurosurg 1989; 70: 691-700. Deutsch M, Green SB, Strike TA et al. Results of a randomized trial comparing BCNU plus radiotherapy, streptozotocin plus radiotherapy, BCNU plus hyperfractionated radiotherapy, and BCNU following misonidazole plus radiotherapy in the post-operative treatment of malignant glioma. Int J Radiat Oncol Biol Phys 1989; 16: 1389-96. Thames HD, Peters LJ, Rodney H et al. Accelerated fractionation vs hyperfractionation: rational for several treatments per day. Int J Radiat Oncol Biol Phys 1983; 9: 127-38. Zeman EM, Bedford JS. Changes in early and late effects with dose per fraction: Alpha, beta, redistribution and repair. Int J Radiat Oncol Biol Phys 1984; 10: 1039-47. Gutin PH, Leibel SA, Wara WM et aL Recurrent malignant

114.

115.

116. 117.

118. 119. 120. 121.

122. 123. 124. 125.

126. 127.

128.

129. 130.

131.

132.

133.

gliomas: Survival following interstitial brachytherapy with high-activity iodine-125 sources. J Neurosurg 1987; 67: 864-73. Gutin PH, Prados MD, Phillips TL et al. External irradiation followed by an interstitial high activity iodine-125 implant 'boost' in the initial treatment of malignant gliomas: NCOG Study 6G-82-2. Int J Radiat Oncol Biol Phys 1991; 21:601 -6. Levin VA, Sheline GE, Gutin PH. Neoplasms of the central nervous system. In De Vita VT Jr, Hellman S, Rosenberg SA (eds): Cancer. Principles and Practice of Oncology, 3rd Edition. J. B. Lippincott: Philadelphia 1989; pp 1557-79. Kyritsis AP, Levin VA. Chemotherapeutic approaches to the treatment of malignant gliomas. Advances in Oncol 1992; 8: 9-13. Komguth S, Kyritsis A, Lohmiller S et al. Biodistribution of polylysine-guided radionuclides. In Epenetos A (ed): Monoclonal antibodies. Applications in Clinical Oncology. Chapman and Hall Medical: London 1991; pp. 133-47. Freeman Al, Fenstermacher J, Shapiro W et al. Select Cancer Ther 1990; 6:104-18. Schluesener HJ, Meyermann R. Spontaneous multidrug transport in human glioma cells is regulated by transforming growth factor type beta Acta Neuropathol 1991; 81:641-8. Cano-Gauci DF, Riordan JR. Action of calcium antagonists on multidrug-resistant cells. Biochem Pharmacol 1987; 36: 2115-23. Bowles Jr AP, Pantazis CG, Wansley W. Use of verapamil to enhance the antiproliferative activity of BCNU in human glioma cells: An in vitro and in vivo study. J Neurosurg 1990; 73: 248-53. Brooks WH, Caldwell HD, Mortara RH. Immune responses in patients with gliomas. Surg Neurol 1974; 2: 419-23. Brooks WH, Roszway TL, Rogers AS. Impairment of rosetteforming T-lymphocytes in patients with primary intercranial tumors. Cancer 1976; 37: 1869-73. Mahaley MS Jr, Gillespie GY. Immunologic considerations of patients with brain tumors. In: Oncology of the nervous system. Boston: Martinus Nijhoff Publishers 1983; 151-65. Lee Y, Bigner DD. Aspects of immunobiology and the immunotherapy and uses of monoclonal antibodies of biologic immune modifiers in human gliomas. Neurol Clin 1985; 3: 901-15. Yoshida S, Takai N, Tanaka R. Functional analysis of interleukin-2 in immune surveillance against brain tumors. Neurosurgery 1987; 21:627-30. Levin VA, Silver P, Hannigan J et al. Superiority of postradiotherapy adjuvant chemotherapy with CCNU, procarbazine, and vincristine (PCV) over BCNU for anaplastic gliomas: NCOG 6G61 final report. Int J Radiat Oncol Biol Phys 1990; 18: 321-4. Shapiro WR, Green SB, Burger PC et al. Randomized trial of three chemotherapy regimens and two radiotherapy regimens in post-operative treatment of malignant glioma. Brain Tumor Coop Group Trial 8001. J Neurosurg 1989; 71: 1-9. Lange OF, Haase DK, Scheef W. Simultaneous radio- and chemotherapy of inoperable brain tumors. Radiother Oncol 1987; 8: 309-14. Rozental JM, Robins HI, Finlay J et al. 'Eight-drugs-in-oneday' chemotherapy administered before and after radiotherapy to adult patients with malignant gliomas. Cancer 1989; 63: 2475-81. Wolff SN, Phillips GL, Herzig GP. High dose carmustine with autologous bone marrow transplantation for the adjuvant treatment of high grade gliomas of the central nervous system. Cancer Treat Rep 1987; 71: 183-5. Mbidde EK, Selby PJ, Perreu TJ et al. High dose BCNU chemotherapy with autologous bone marrow transplantation and full dose radiotherapy for grade IV astrocytoma. Br J Cancer 1988; 58: 779-82. Hochberg FH, Parker LM, Takvorian T et al. High dose BCNU with autologous bone marrow rescue for recurrent

433 glioblastoma multiforme. J Neurosurg 1981; 54: 455-60. 134. Hochberg FH, Pruitt AA, Beck DO et al. The rationale and methodology for intra-arterial chemotherapy with BCNU as treatment for glioblastoma. J Neurosurg 1985; 63: 876-80. 135. Roosen N, Kiwit JCW, Lins E et al. Adjuvant intra-arterial chemotherapy with nimustine in the management of World Health Organization grade IV gliomas of the brain. Cancer 1989; 64: 1984-94. 136. Longee DC, Friedman HS, Albright RE et al. Treatment of patients with recurrent gliomas with cyclophosphamide and vincristine. J Neurosurg 1990; 72: 583-8. 137. Bleehen NM, Freedman LS, Stenning SP. A randomized study of CCNU with and without benznidozole in the treatment of recurrent grades 3 and 4 astrocytoma. J Radiat Oncol Biol Phys 1989; 16: 1077-81. 138. Hait WN, Byrne TN, Piepmeier J et al. The effect of calmodulin inhibitors with bleomycin on the treatment of patients with high-grade gliomas. Cancer Res 1990; 50:6636-40. 139. Levin VA, Chamberlain MC, Prados MD et al. Phase I-II study of eflornithine and mitoguazone combined in the treatment of recurrent primary brain tumors. Cancer Treat Rep 1987; 71:459-64. 140. Prados M, Rodriguez L, Chamberlain M et al. Treatment of recurrent gliomas with l,3-bis(z-chloroethyl)-l-nitrosourea and alpha-difluoromethylomithine. Neurosurgery 1989; 24: 806-9. 141. Kumar ARV, Renaudin J, Wilson CB et al. Procarbazine hydroghloride in the treatment of brain tumors. Phase 2 study. J Neurosurg 1974; 40: 365-71. 142. Rodriguez LA, Prados M, Silver P et al. Reevaluation of procarbazine for the treatment of recurrent malignant central nervous system tumors. Cancer 1989; 64: 2420-3. 143. Newton HB, Junck L, Bromberg J et al. Procarbazine chemotherapy in the treatment of recurrent malignant astrocytomas

Book review

144.

145.

146.

147.

148.

149.

150.

after radiation and nitrosourea failure. Neurology 1990; 40: 1743-6. Yung WKA, Mechtler L, Gleason MJ. Intravenous carboplatin for recurrent malignant glioma. A phase II study. J Clin Oncol 1991; 9: 860-4. Coyle T, Baptista J, WinGeld J et al. Mechlorethamide, vincristine, and procarbazine chemotherapy for recurrent high-grade glioma in adults: A phase II study. J Clin Oncol 1990; 8: 2014-8. Yung WKA, Castelanos AM, Van Tassel P et al. A pilot study of recombinant interferon beta in patients with recurrent glioma. J Neuro-oncol 1990; 9:29-34. Nierenberg D, Harbaugh R, Maurer LH et al. Continuous intratumoral infusion of methotrexate for recurrent glioblastoma: A pilot study. Neurosurg 1991; 28: 752-61. Brem H, Mahaley MS Jr, Vick NA et al. Interstitial chemotherapy with drug polymer implants for the treatment of recurrent gliomas. J Neurosurg 1991; 74:441-6. Yung WKA, Lotan R, Lee P et al. Modulation of growth and epidermal growth factor receptor activity by retinoic acid in human glioma cells. Cancer Res 1989; 49: 1014-9. Short MP, Choi BC, Lee JK et al. Gene delivery to glioma cells in rat brain by grafting of a retrovirus packaging cell line. J Neurosci Res 1990; 27: 427-33.

Received 9 March 1992; accepted 10 March 1992.

Correspondence to: Dr. Victor A. Levin University of Texas M.D. Anderson Cancer Center Department of Neuro-Oncology, Box 100 1515 Holcombe Blvd. Houston, Texas 77030-4596, USA

Annals of Oncology 3:433, 1992.

on these topics, ranging from basic genetic and biochemical mechanisms to clinical applications. Specific mechanisms of resistance to antimetabolites are not reviewed in detail; only one chapter covers some aspects of the phenomenon of gene amplificaThis book is part of a series, entitled Cancer Treatment tion, one which is certainly relevant in the mechanisms and Research, which provides annual or biannual up- of resistance to most antimetabolites. dates on current oncology literature concerning speAll of the authors are prominent scientists who have cific topics. This volume is essentially focused on made major contributions to this field, and many remechanisms of resistance to alkylating agents including views by the same authors are already available in cisplatinum and natural products. specialized oncology journals. Nevertheless, a collecCertainly the study of drug resistance is in the proc- tion of these reviews in one volume can be useful to ess of rapid evolution and a frequent updating of rel- students, scientists and clinicians involved in this area evant molecular and biological findings and results of of research. clinical applications is of great interest. With respect to M. D'lncalci basic mechanisms, this volume is particularly interesting on DNA repair, DNA-topoisomerases enzymes Milan and multi-drug resistance. It contains several chapters Molecular and clinical advances in anticancer drug

resistance. R. F. Ozols (ed). Kluwer Academic Publishers, Boston/Dordrecht/London, 1991. 320 pp, $ 125.00, £82.75, Dfl. 280.00.