J o u r n a l o f N e u r o - O n c o l o g y 4: 9 2 - 1 1 6 , 1986 9 M a r t i n u s N i j h o f f Publishers, B o s t o n - P r i n t e d in the N e t h e r l a n d s

1

PROTO-ONCOGENESAS REGULATORSOF GROWTHAND DIFFERENTIATION Russel E, Kaufman, Dept. of Medicine, Duke Unfversfty Medical Center, Durham~ NC

Malignant c e l l s can be distinguished in part by abnormalities in p r o l i f e r a t i o n and d i f f e r e n t i a t i o n , These processes are normal ly control led by e x t r a c e l l u l ar factors or signals which activate or modulate complex biologic pathways and regulate p r o l i f e r a t i o n , Specific i n t r a c e l l u l a r factors are probably essential in the coordinate and d i f f e r e n t i a l expression of multtgene f a m i l i e s necessary for progression through the eel 1 cycle. In the l a s t S years~ numerous genetic elements which partfctpate in these processes have been I d e n t i f i e d and characterized. C o l l e c t l v e l y they have been termed iproto)oncogenes. Individually they normally participate in a wide range of c e l l u l a r functions. In nearly a l l malignant tissues studted to datep investigators have found that certain oncogenes have a l t e r e d patterns of expression, either by excess a c t i v i t y , Inappropriate tlmlng of expression or altered function due to changes in the structure of the product. S i m i l a r i t i e s among these genes and t h e i r products al low c l a s s f f y i n g them into several groups. One group includes oncogenes whose normal gone product is involved in signal transductton. This group includes genes whtch encode growth factors (e,g,, PDGF=sis)p t h e i r receptors (EGF receptor=Erb B), or proteins which interact with receptors (receptor coupltng protelns=ras). Another group of oncogenes has gone products which are nuclear proteins and are suspected t o p a r t i c i p a t e in c e l l cycle regulation (e.g., myc, myb). A minorlty of i d e n t i f i e d oncogenes cannot be grouped using t hi s c l a s s i f i c a t i o n scheme. I w i l l present a model which attempts to unlfy the role of oncogenes in malignant transformation. Examples of clinical appllcations of oncogene typing wlll be presented,

3

MULTIPLE GROWTHFACTORSREGULATECELLULAR PROLIFERATION W.J. Pledger, Department of Cell Biology, Vanderbilt University School of Medicine, Nashville, Tennessee 37232

Several lines of evidence indicate that cellular proliferation is controlled by the actions of multiple growth factors. The growth factors required to regulate proliferation vary depending on cell type. Fibroblasts have been extensively studied and a model of in vitro growth regulation by platelet derived grow--t-h--ira-c-i~or, epidermal growth factor and somatomedin-C has been proposed. Each growth factor has a unique function that may be required during a specific cell cycle phase. The growth factors interact with the cell via specific receptors leading to modulation of gene expression. The specific sequence of biochemical events after growth factor-cell interaction has not been elucidated. It has been demonstrated that various growth factors interact with cells and modulate the ceils capacity to respond to other growth factors. This modulation may be brought about by alteration of receptor number and/or receptor a f f i n i t y for ligand. These types of interactions suggest possible mechanisms whereby growth factor-cellular interactions can be modulated to regulate proliferation. Abrogation of growth factor requirements by transformation can be caused by unscheduled growth factor production or modification of the processes controlled by growth factors. Finally, growth factors, their specific receptors, the biochemical processes these control; and the gene expression that growth factors regulate must be investigated in order that proliferative control can be understood.

2

CHROMOSOMES AND GENE AMPLIFICATION IN MALIGNANT HUMAN GLIOMAS. Sandra H. Bigner, M.D., ]oachim Mark, M.D., Kenneth W. Kinzler, Bert Vogelsteinp M.D., S. Clifford Schold, 3r., M.D. and Darell D. Bigner, M.D., Ph.D., Duke University Medical Center, Durham, North Carolina, Central Hospital, Sk~vde, Sweden, and The 3ohns Hopkins University School of Medicine, Baltimore, Maryland.

In contrast to karyotypic studies of most leukemias and lymphomas in which structural deviations predominate, the most consistent and probably earliest gross chromosomal abnormalities of malignant human gliomas consist of gains or losses of whole chromosomes and the presence of double minutes (DMs). We have been able to identify and characterize abnormal stemlines in 31 of the 38 MHG which we have karyotyped in direct preparation and/or short term culture. The most prevalent chromosomal deviations were gains of chromosome No. 7, usually accompanied by losses of chromosome No. 10. Eighteen tumors had one or both of these changes. Five of these tumors have been serially passaged in cultures or in athymic mice and have maintained the abnormalities of Nns. 7 and 10. Consistent structural abnormalities are less common. Structural abnormalities of No. 9 were seen in 9 cases with breakpoints in 9p or at the centromere. Two oi these tumors have maintained this abnormality in culture, in nude mice or both. Seven tumors contained structurally abnormal No. 6 but the breakpoints were variable. One of these tumors maintains this structural change as a serially passaged nude mouse tumor. DMs were identified in 15 tumors; they heavily involved nearly every cell in 5 of these cases. Four of the 5 tumors with abundant DMs have been propagated in culture or in athymic mice and have maintained the DMs. Molecular analysis performed on cells from 2 of these lines have confirmed that they contain amplified sequences. The amplified gone in D-g98MG was identified as erb B by Southern blotting. The amplified gene in D-2-SgMG did not correspond to any of the oncogene probes tested to date. Further studies to isolate and identify this gene are in progress. The ability to propagate ceils derived from malignant gliomas in culture and in athymic mice which maintain the specific karyotypic abnormalities seen originally provides an opportunity to define the significance of these chromosomal changes.

4

GROWTH FACTORS AND ONCOGENES IN HUMAN GLIOMA Bengt Westermark, Dept. of Pathology, University of Uppsala, Uppsala, Sweden

Studies on oncogenes in relation to recent information on the structure and function of polypeptide growth factors have yielded a unifying concept of the molecular mechanism of neoplastic transformation and normal mitogenesis. The f i r s t link was provided by the finding that the oncogene v-sis of simian sarcoma virus (SSV) is derived from cellu--1-ar sequences encoding the B chain of platelet-derived growth factor (PDGF). This suggests that a PDGF-like growth factor is effective in SSVinduced transformation by causing an autocrine growth stimulation. An involvement of the cellular homolog to the viral sis gene (c-sis) in human glioma is suggested by the f in l~Tng of c-sis messenger RNA and PDGF production in established glioma---cell lines. One such example is the clonal line U-343 MGa C12 that contains the 4.2 kilobases c-sis transcript and produces r e l a t i v e l y large quantitTes of PDGF or a closely related factor. An analysis of a large number of U-343 MGa clones has provided evidence f o r a considerable clonal variation in c-sis expression, PDGF production and apparent number o~t-I~DGF receptors. High passage clones in general tend to have a higher PDGF production than clones derived from low passage cultures, suggesting that production of a PDGF-Iike growth factor is of selective growth advantage. An endogenously produced growth factor such as gliomaderived PDGFmay not only be involved in an autocrine growth stimulation but also stimulate the proliferation of neighboring cells such as non-producing receptorbearing tumor cells and stroma cells.

93 MECHANISMS FOR TARGET ORGAN AND TARGET CELL SPECIFIC1-WIN CARCINOQENESES. James A. Swenberg, Departaent of Biochemical Toxicology and Pathobiology, CIIT, Research Triangle Park, NC 27709 Many carcinogens are highly s p e c i f i c f o r both the s i t e end c e l l type involved in chemical carcinogenesis. Factors involved in such c e l l and organotrophy, include the route of exposure, absorption and d i s t r i b u t i o n , s i t e of biotransformation, c e l l r e p l i c a t i o n and DNA repair. D i f f e r e n t chemicals give r i s e to e l e c t r o p h i l e s of varying degrees of hardness that react with cellular nueleophiles, leading to fingerprints of DNA adducts. DNA adducte d i f f e r in t h e i r a b i l i t y to base-pair during DNA synthesis. Some adducts are highly e f f i c i e n t in causing mispalring, leading to a high rate of mutation, while other DNA adducte do not a l t e r base-pairing from t h a t of the parent nucleotide and t h e r e f o r e do not r e s u l t in mutations. Repetative exposure to carcinogens f u r t h e r complicates the issue, since d i f f e r e n t DNA adducts are repaired at d i f f e r e n t rates. Furthermore, d i f f e r e n t cell types vary in t h e i r a b i l i t y to form and repair ~ A adducts. For example, hepatocytes r a p i d l y repair aJkylguanine, but are much less e f f i c i e n t at repairing O~-ethyldeoxythymidine. On chronic exposure the l a t t e r adduct accumulates to concentrations at least 50-fold greater than 0 -EtdC, even though only 1/4 as much is formed. Despite the above complexity there is general agreement t h a t the potential f o r chemically-induced mutations in a given cell or tissue type correlates well with c a r c i n o g e n l c i t y . Present studies are focusing on the development of u l t r a s e n s i t i v e methods f o r DNA adduct q u a n t i t a t i o n so t h a t comparisons of high to low dose exposure and species differences can be made. The same techniques should be applicable to monitoring DNA adducts of cancer chemotherapeutics. 5

6

SENSITiViTY AND RESISTANCE TO THE HALOETHYLNITROSOUREAS. David B. Ludlum, Department of Medicine, Albany Medical College, Albany, NY 12208.

The haloethyinitrosoureas are of special interest because of t h e i r a c t i v i t y against CNS malignancles. Not al1 such tumors respond, however, and some that do eventually develop resistance. Investigations of the action of these agents at a molecular level have suggested that the resistance problem may be related to DNA repair. The c y t o t o x i c i t y of the haloethyinitrosoureas is e v i d e n t l y caused by t h e i r chemical attack on DNA. Many d i f f e r e n t DNA modifications are produced, including monofunctiona] base s u b s t i t u t i o n s , phosphate a l k y ] a t i o n , and crossiink formation. Crosslinks result from the transfer of a haloethyi group to DNA followed by reaction of this group with the opposite DNA strand, and are especially important because t h e i r formation correlates with c y t o t o x i c i t y . One p a r t i c u l a r DNA crosslink, l - ( 3 - d e o x y c y t i d y l ) , 2 (I-deoxyguanosinyi)-ethane, is formed by a mechanism which involves i n i t i a l attack by g haloethyi group on the 6 position of guanine. The 0 -haloethyIguanine formed in this way is gradually converted into a crosslink unless it is repaired by 06-alkyltransferase, thus oFFering an explanation for the resistance of cell lines which have high levels of 06-aikyitransferase a c t i v i t y . Evidence will be presented that other DNA modifications are also cytotoxic and may be repaired by entirely different mechanisms. These studies offer the possibility that cell lines can be assayed for their repair a c t i v i t y and their potential sensitivity to the nitrosoureas determined before chemotherapy is instituted. Supported by grant CA ]2171 f r ~ the National Cancer Institute,

THE ROLE OF 0-6 ALKYLGUANINE MONOADDUCTREPAIR IN DETERMININGDNA CROSSLINKINGLEVELS IN HUMAN GLIOMA CELL LINES EXPOSEDTO CHLOROETHYLNITROSOUREAS (CENU). L.C. Erickson*, E. Sariban, C. Zlotogorski, N. Gibson, D. Yarosh, and K. Kohn. *Loyola Medical Center, Maywood, IL., and National Cancer Institute, NIH, Bethesda, MD. Recent studies from our lab have shown a strong correlation between the level of DNA interstrand crosslinking (ISC) produced by the CENU, cis-Pt, and AZQ, with the cytotoxicity in human t u m o r ~ l l s exposed to the drugs in v i t r o . The level of ISC induced by the CENU, ~ not cis-Pt or AZQ, also correlated with the a b i l i t y o~"-t-he cells to repair 0-6 methylguanine lesions in their DNA (R. Day and coworkers). In cells that actively repair 0-6 methylguanine lesions, no CENU-induced ISC was observed, and these cells were resistant to cell k i l l i n g when compared to cells deficient at this repair process. We have examined 16 human glioma cell lines exposed to CENU, and ISC was observed in only 6 cell lines. 15 ceil lines were exposed to cis-Pt and assayed for DNA ISC. There was no correel-at-ion betwen the ISC observed w i t h c i s - P t and the a b i l i t y to prevent CENUISC. 8 c e l l ~ n e s were studied following AZQ exposure. Again, there was no correlation with the a b i l i t y to prevent CENU ISC and the AZQ ISC. Collectively, the data show that the DNA repair protein 0-6 alkylguanine DNA alkyltransferase (see W. BodeII, this symposium) may protect a large proportion of human glioma tumor cells from the cytotoxic lesions induced by CENU, but not from the DNA lesions induced by cis-Pt or AZQ. DNA transfection experiments using o-d'6-~or DNA from repair proficient cells show that the resistant phenotype of DNA ISC prevention, and increased survival after CENU exposure, can be transferred to a repair deficient recipient glioma cell line.

7

INCREASED REPAIR OF O6-ALKYLGUANINE IN DNA OF CELLS RESISTANT TO 1,3 bis (2-CHLOROETHYL)-INITROSOUREA By: W.J. Bodell, IL Cheitlin, T. Aida, M.S. Berger, M.R. Rosenblum. Brain Tumor Researeh Center, University of California, San Francisco, CA Tne purpose of this investigation was to study cellular resistance to the cytotoxic effects of BCNU in human glioma-derived cell lines. Compared to SF-126 cells, SF-188 cells are approximately 3-fold more resistant to the eytotoxic effects of BCNU treatment. SF-188 and SF-253-2 cells are 5-14-fold more resistant than SF-126 cells to the induction of sister ohromatid exchanges by a 1 hr BCNU treatment. "[he formation of DNA interstrand erossllnks in cells 6 hr after treatment with 100 PM of BCNU was measured by alkaline elution. This treatment induced 80 crosslinks in SF-120, 30 in SF-188 and 57 in SF-253-2 cells. The role of DNA repair in cellular resistance to BCNU was analyzed b~ treating cells with I00 pM 3H-methylnitrosourea for 1 hr. The aH methylated base alkylation products were separated by HPLC. The O6-mGua/7-mGua ratio was 0.131 for SF-126, .067 for SF-188 and .043 for SF-253-2, indicating t h a t cells resistant to BCNU have an inereased eapability for repair of O6--alkylguanine derivatives in DNA. The enzyme which removes O6-mGua is 0 6 alkylguanine DNA alkyl transferase. The activity of the enzyme was measured using a DNA template containing 3H-OS-mGua-~. Treatment of SF-188 cells with MNU resulted in a dose-dependent inhibition of the transferase activity. P r e - t r e a t m e n t of SF-188 ceils with MNU followed by BCNU t r e a t m e n t potentiated the induetion of SCE's; suggesting t h a t inhibition of transferase activity increases the induction of SCEs by BCNU t r e a t m e n t . Our results suggest t h a t one of the moleeular mechanisms for the observed eenular resistanee to BCNU may be due to inereesed repair of O6-alkylguanine derivatives formed by BCNU t r e a t m e n t in treated eells. This repair process appears to result in fewer DNA interstrand erosslinks and SCEs induced and increased eellular survival a f t e r BCNU t r e a t m e n t of resistant human glioma cells. Supported by CA-13525. 8

94 HUMN JC POLYOMA VIRUS IN EXPERIMENTAL NEURO-ONCOLOGY G a b r i e l e N. ZuRhein,Oepartment o f P a t h o l o g y p U n i v e r s l t y o f Wisconsin,Nsdison~Wisconsin,USA Human JC polyoma v i r u s , a s m s l l naked DNA v i r u s , ~ s first Iso l a t e d in 1971 ( P a d g e t t ~ a l . L a n c e t l l 1 2 5 7 ) from a h u m n b r a i n w i t h p r o g r e s s l ~ m ~ m u l t l f o c e l l e u k o e n c e p h a l o p a t h y . I n t h i s rare usually s eslow virus e disease the virus kills the ollgo~ d e n d r o c y t e s l e a d i n g t o d e m y s l i n a t i o n . About 70~ o f a d u l t s in t h e USA have serum a n t i b o d i e s a g a i n s t J C V . ~ e S y r i a n hamster ( N e s o c r l c e t u s a u r a t u s ) i s t h e animal s p e c i e s most e x t e n s i v e l y s t u d i e d f o r J C V n e u r o - o n c o g e n e e i s . T~n d i f f e r e n t s e r i e s , a n t msls were I n o c u l a t e d during l a t e f e t a l l i f e , as n e u b o r n s , a t a g e 3 or 19 dayst as ~ e a n l i r g s or a d u l t s . I n o c u l a t i o n s i t e s variedl intreeerebral(i.c.), subcutaneous(s.c.)plus i . e . , s . c . p l u s i n t r a p e r i t o n e a l , i n t r a o e u l a r or intravenous.None o f t h e about 750 hamsteres were immunosuppressed. In a l l e x p e r i m e n t s , JCV i n o c u l a t e d hamstera~developed nervous s y s t e m tumors and a l l ( over 300) c o n t r o l a n i m l s were n e g a t i v e . In R e i d e n e e uas h i g h e s t (about 8 5 - 9 5 ~ ) i n animals i n o c u l a t e d i . e . as newborns and l o w e s t ( l e s s than 10~) in animals i n o c u l a t e d p a r e n t e r a l l y . Latency p e r i o d s f o r c e n t r a l nervous sytem t u o more i n c r e a s e d w i t h t h e age o f the animal a t t h e time o f i n o c u l a t i o n and e x t e n d e d from 3 t o 30 months. Favorite s i t e s for neuroectodermSl tumors were the c e r e b e l l u m , t h e t h a l a m i , the p y r i f o r m l o b e s , t h e o l f a c t o r y - f r o n t a l region,the ventricl e s , and the s p i n a l c o r d . R e s u l t s o f t h e f i r s t experiment (Walker e t e l . S c i e n c e 1 8 1 1 6 7 4 , 1 9 7 3 ) included serial transplan l a b i l i t y o f b r a i n t u m o r s , p r e s e n c e o f T a n t i g e n in c u l t u r e d tumor c e l l s and p r e s e n c e o f a n t i b o d y t o T a n t i g e n in about 60~ o f s e r e . The spectrum os JC~ induced t u n e r s o f the n e r vous s y s t e m i n c l u d e s medulloblastomas,mlignant a s t r o e y t i c tumors and g l i o m a t o s i s c e r e b r i , p r i m i t i v e neuroectodermal tumors, c e r e b r a l n e u r o b l a s t o m S s , m s l i g n a n t schwennomas, men i n g e a l sarcomas,malignant ependymomas, p i n e O e y t o r r ~ s , r e t i n o b l a s t o m a s , p e r i p h e r a l s e u r o b ~ a s t o ~ s and c e r e b r a l angiomas. In v i v o e x p e r i m e n t s with JCV have a l s o been conducted with non-human p r i m a t e s . A f t e r i . c . . s . c . a n d l.~.inoculation of a d u l t non-lmmunosppressed owl monkeya(Aotus t r l v i r g a t u s ) (London e t , ; a l . S c i e n c e 201:12~6,1978 and Z u g h e i n , u n p u b l . ) grade 3-4 a s t r o c y t o m s s d e v e l o p w i t h l a t e n c y periods from 1 5 - ~ months. E x t e n s i v e g l i o m a t o a i s and m o n s t r o c e l l u l a r t u mors a r e n o t e d . The s q u i r r e l , monkey r e a c t s s i m i l a r l y 9

9

Acrylonitrile Neurooncogenesis in F34# Rats. Bigner, D.D., Bigner, S.H., Burger, P.C., Sheiburne, 3.D., and Friedman, H.S.Duke University Medical Center, Durham, N.C. Acryionitrile (ACN, vinyl cyanide, CH=CHCN) is an important bulk commodity chemical in the textile, plastics and synthetic rubber industry. Because of the potential human contact in manufacture and use of the chemically active ACN monomer, potential environmental concerns, and existing and proposed use of plastic containing ACN polymers as beverage and food conraisers, chronic bioassays and genotoxicity evaluation of ACN began in the 1970s. ACN is mutagenic, teratogenic, and causes DNA damage in vitro. Several chronic bioassays with administration of doses of 100-:~00 PPM of ACN in drinking water or following inhalation have been performed. We have performed a lifetime study of exposure of Fischer-34~ rats to doses ol 500 and 100 PPM ACN. At the present time, 215 animals dying between 6 and 18 months after dosing with 500 PPM of ACN have been autopsied. Zymbal gland tumors, forestomach papillomas, subcutaneous papillomas, and primary brain tumors have been observed. OI the total of #9 brain tumors examined 22% were microscopic, 5g% were medium in size (up to 5 ram) and 2096 were greater than 5 mm in diameter. The brain tumors were highly infiltrative. The neoplastic cells were monomorphous with dark nuclei and scant cytoplasm. There was a marked propensity for neuronal satellitosis and frequent positioning of neoplastic ceils was around blbod vessels. With GFA immunohistochemistry reactive gila were abundant but no unequivocal staining of tumor cells occurred. The majority of attempted tumor transplantations were unsuccessful, but three large primary brain tumors could be transplanted intracerebraiiy in athymic mice or athymic rats. Multiple attempts at establishing permanent cell culture lines were unsuccessful. The majority of karyotypic studies in direct preparation showed normal rat stemlines with diploid karyotype and minimal numerical dr structural deviations. A trisomy-2 was detected in the tumor transplantable in athymic rats. Chronic oral administration of ACN induces a significant incidence of primary brain tumors in Fischer-344 rats. The tumors have been preliminarily classified as anaplastic astrocytomas but there is no definitive evidence for astrocytic histJogenesis from ultrastructural or immunohistochemical studies. 11

10

BRAIN TUHORS IN F344 RATS CHRONICALLY EXPOSED TO ETHYLENE OXIDE. R. H. Carman I, W. M. Snellings l, and R. R. Maronpot 2, iBushy Run Research Center, R.D. #4, Export, PA 15632 and 2NIEHS, P. O. Box 12233, Research Triangle Park, NC 27709.

Male and female F344 rats (120 rats/sex/exposure group) were exposed to either ethylene oxide (ETO) vapor (100, 33 or 10 ppm) or to room air (2 concurrent control groups, each with 120 rats/sex). Exposures were for 6 hours daily, 5 days/week for up to 2 years. Three representative sections of the brain were evaluated from each rat sacrificed at scheduled time points (6, 12, 18, and 24 months), as well as from those dying spontaneously or sacrificed in a moribund condition. Twenty-three primary brain tumors were found, representing three major diagnostic categories: granular cell to,more, glial cell tumors, and malignant retlculoses. Two tumors (I granular cell and 1 gllal cell tumor) were seen in control rats. Twenty-one tumors (5 granular cell, 14 gllal cell and 2 malignant reticuloses) were seen in ETO-exposed rats. Eleven of these tumors were in ETO-exposed rats dying while on study. At least 6 were considered to be the primary cause of death. The adjusted ratio frequencies (~.s number of rats with tumor/number alive at the time the first tumor in any group was observed) for primary brain tumors in the male rats were: 7/87 (100 ppm), 5/85 (33 ppm), 1/92 (10 ppm) and 1/181 (combined control groups). For the female rats, the frequencies were: 4/80 (lO0 ppm), 3/92 (33 ppm), 1/94 (10 ppm) and 1/188 (0 ppm). The brain tumor frequencies were, therefore, increased for both the 100 and 33 ppm groups (particularly for males), but no increased frequency was seen for the 10 ppm groups.

12

REGIONAL IHAGING OF BLOOD FLOW, GLUCOSE USE, p,,, ATP, G L U C O S E A~!D L A C T A T E C O N T E N T IN EXPERI~TENTAL

G. ~ s

RAT

lq. Paschen*,

Ir I~echsler * * *Hax-Planck-lnstltute

RRAIN

L. for

CLIOHAR

qzaho ~,

K-A.

Neurological

Hossmann*, Research,

Dept. of Expt. Neurology, D-5000 K51n 91 and ** D e p a r t m e n t of N e u r o p a t h o l o g y , U n i v e r s i t y of 1)Eisseldorf, p-4000 DUsseldorf, FEe

Experimental

brain tumors were produced In rats by s t e r e o t a c E i c i n o c u l a t i o n of glloma cell clones (RC2, F98) and the cell line RGI2.2. E s t a b l i s h e d and new i m a g l s g t e c h n i q u e s were a p p l i e d for measuring d i f f e r e n t p a r a m e t e r s of tumor patbop h y s i o l o g y : c e r e b r a l blood flow, g l u c o s e u t i l l s a tins and protein s y n t h e s i s by q u a n t i t a t i v e autor a d i o g r a p h y oslng labeled i o d o a n t i p y r i n e , deoxyg l u c o s e and amino acids, r e s p e c t i v e l y ; regional ATP, )~lucose and lactate by substrate-speclfic biolumlnescence; r e g i o n a l tissue pH by DHO a u t o r a d l o g r a p h y or by pal-dependent u m b e l l i f e r o n e fluorescence. The glloma t r a e s p l a n t a t i o n tumors showed d i f f e r e n c e s but also s i m i l a r i t i e s . In solid tumor regions blood flow was in the range of normal hrais tissue but g l u c o s e u t i l l s a t l o n and protein s y n t h e s i s were d i s t i n c t l y i n c r e a s e d . C l u c o s e use was u n r e l a t e d s ATP c o n t e n t but c o r r e l a t e d i n v e r s e l y w i t h tissue g l u c o s e contest. 14hen set g l u c o s e e x t r a c t i o n I n c r e a s e d about 30-40Z, ATP d e c l i n e d to below 30%. pl| was g e n e r a l l y more alkallne in tumors than tn s u r r o u n d i n g brain tissue w i t h the e x c e p t i o n of some regions w i t h low ATP w h e r e pH a p p e a r e d more acid. The results are cons i s t e n t with recent PET f i n d i n g s of a e r o h l c g l y c n lysls in human gllomas. Factors for the m a i n t e n a n c e of the energy state of tumors are g l u c o s e a v a i l a h i l l t y on the one hand, and tissue a c i d o s i s on the other. This I n d l e a t e s that tissue c l e a r a n c e of aeld e q u i v a l e n t s at reduced p e r f u s i o n is of equal i m p o r t a n c e as reduced s u b s t r a t e d e l i v e r y for the d e v e l o p m e n t of s p o n t a n e o u s tumor n e c r o s i s .

95 EVIDENCE THAT ONCOGENESAND GROWTHFACTORSPARTICIPATE IN CHEMOTHERAPY RESISTANCE IN HUMANMALIGNANTGLIOMAS. J.R. Shapiro and W.R. Shapiro, Memorial SloanKettering Cancer Center, New York, NY. Our previous studies on heterogeneity of human gliomas have increased our understanding of the natural evolution of such tumors and how therapy changes that evolution. Freshly resected and cloned gliomas are heterogeneous in their chromosomal complements and chemosensitivity. This heterogeneity is regional; adjacent regions often share the same clonal populations while distant regions d i f f e r remarkably. In contrast to the heterogeneity that follows untreated tumor evolution, our studies of intrinsic and acquired BCNUresistance suggest that therapy establishes a selective pressure that produces a substantially more homogeneous tumor. Cytogenetic studies indicate that the sensitive ceils are hyperdiploid with chromosome numbers greater than 60, while intrinsically resistant cells are small, neardiploid and fast growing. Continued BCMUexposure to heterogeneous populations in acquired resistance studies always produce resistant cells that are near-diploid. The regional studies demonstrate sensitive regions as hyperdiploid and resistant regions as near-diploid. We have found that resistance also correlates with the over-representation of chromosomes 7, 20 and 22. Chromosomes 7 and 22 contain the proto-oncogenes c-erb-B and c-sis, respectively coding for epidermal growth factor receptor (EGFR) and platelet derived growth factor (PDGF), while chromosome 20 contains the proto-oncogene c-src. PDGFis mitogenic to glial cells. Our recent studies demonstrate that BCNU-resistan~ human glioma ceils secrete a substance that stimulates Hthymidine incorporation into PDGFreceptor-positive ceils, but not into PDGFreceptor-negative ceils. The substance thus appears to be PDGFor PDGF-like in its action. Others have shown increased EGFRin gliomas. All three oncogenes may be interrelated. We now propose that as chemotherapy k i l l s off the sensitive ceils, regionally situated foci of resistant cells re-populate the tumor, using to their selective advantage the gene-products (growth factors) of their specific chromosomal complements. These observations provide clues to developing more specific therapy, perhaps directed against such growth factors and/or their receptors. 13

15

THE BIOLOGY OF INDOLENT ASTROCYTOMAS: LESSONS LEARNED FROM CHRONIC MYELOGENOUS LEUKEMIA (CML)HYPOTHESIS. J. Gregory Cairncross, Departments of Clinical Neurological Sciences and Radiation Oncology, University of Western Ontario and the London Regional Cancer Centre, Victoria Hospital, London, Canada. CML is an indolent, often asymptomaCic tumor in which the neoplastic cells resemble normal granulocytes. The cell of origin is a pleuripotential hematopoietic precursor which, although transformed, continues to differentiate along myeloid lines. There is increasing evidence that differentiation can occur or be induced in a transformed precursor cell. In the case of CML, the factors directing the preferred "myeloid tumor phenotype" are unknown. CML often evolves into an acute leukemia. The molecular events underlying "blast crisis" are unknown but transformed precursor cells begin to divide at an accelerated rate and no longer differentiate. The clinical picture is then indistinguishable from de novo acute myelogenous leukemia. Superficially at least, there are parallels between the behavior of CML and that of indolent astrocytomas. The following hypothesis is offered: astrocytomas arise from pleuripotential glial precursor cells; transformed gliel precursors then differentiate giving rise to the "astrocytoma tumor phenotype"; malignant change, like blast crisis, reflects accelerated growth and "maturation arrest" in the transformed previously differentiating precursor cell. The analogy between CML and astrocytoma is strengthened by several observations: glial precursor cells persist in adult brain; glial precursors are the probable ethylnitrosourea (ENU) target cell; astrocytic differentiation can be induced in certain murine embryonal carcinoma lines; differentiation occurs in primitive neuroectodermal tumors; and differentiation is reversible only in its earliest phases (ie. de-differentiation does not occur).

14

Vagaries of Clonogenie (Stem) Cell Analysis of Human ~,a~, T U | Mark L. Rosenblum, Dennis A. Emma, Henne Hoifodt, and Dolores V. Dougherty Brain Tumor Researeh Center, University o f California, San Francisco.

It is presumed that an understanding of stem cell biology is necessary to best comprehend tumor growth and to develop means of curing, rather than merely palliating, patients with malignant brain tumors. In vitro methods have been developed to grow end study tumor cells that can operationally be defined as stern ceils; the definition is based on a eell~ proliferative c a p a c i t y to develop into a colony using various culture systems. However, there are many potential problems a t t e n d a n t with the in vitro isolation end quantiation of this self-renewing eel1 populati'~'n. Clonogenic cells will develop into colonies in an anchorage dependent fashion in monolayer culture end in an anchorageindependent fashion using soft-gel matrix systems. We have evaluated the colony forming efficiency (CFE) o f tumor cultures using monolayar, Courteney, and Hamburger-Salmon (H-S) assays. Tne CFE in monolayer was always g r e a t e r than in ogar; growth in Courtenay was usually b e t t e r than in H--S systems. The s t a t e of proliferation of cells prior to plating influenced the CFE both for established cell lines and very early passage cells. Six tumor cell cultures were evaluated for the growth promoting.activity of 23 me~]ium constituents using a 3H-Tdr incorporation essay. Response to each faetor showed some variation from tumor to tumor. Six factors were considered "stimulatory", 8 "inhibitory", and 11 did not influence cell proliferation. All 4 "stimulatory" factors increased CFE or colony size for selected tumors in the Courtenay essay; elimination of "inhibitors" universally inereesed CFE using the H-S essay. Clonogenie cell growth was usually improved using the more physiological 5% 82 atmosphere end upon the addition of August RBCs. Therefore, it is apparent that the ceils we define as clonogenie will depend upon the culture conditions employed. As a consequence, conclusions derived from biochemical and molecular biological studies of the clonogenie cell population might differ from one laboratory to another. Identification of "optimal" culture conditions should beneficially influence investigations of the elonogenie cells derived from human brain tumors. Supported by CA 31882 and CA 13525

16

SYNCHRONIZATION OF HUMAN GLIOMA CELL LINES WITH HYDROXYUREA. Jaclyn A. Biegel, Ph.D., Sandra H. Bigner, M.D., David S. Leslie, B.A. and Darell D. Bigner, M.D., Ph.D., Departments of Pathology and Medicine, Duke University Medical Center, Durham, NC 27710

Karyotypic studies of human gliomas are often limited by a low mitotic index and the appearance of contracted chromosomes that do not band well. In the past few years, several methods for increasing the mitotic yield and chromosome length in hematopoietic cells have been established, most notably methotrexate synchronization. Preliminary attempts to synchronize established human glioma cell lines with methotrexate were unsuccessful. We therefore investigated the use of hydroxyurea (HU) as a synchronizing agent. Exponentially growing replicate cultures of four glioma lines U-251MG, D-245MG, D247MG and D-263MG were exposed to 2mM HU in zinc option medium containing 10% fetal calf serum for 12-24 hours. Glioma cells were released from the HU block with fresh unsupplemented medium. Cells were stained with 0.25 mg/ml propidiom iodide in 5% Triton X-IO0 and subjected to cell cycle analysis on the basis of DNA content with the use of an EPICS V cell sorter. An analysis of the proportions of cells in G1, S and G2/M in control and exposed cultures was made at 4-6 time points following release from the HU block. The rate of cell cycle traverse was different for the four cell lines, and could not be predicted from the population doubling time. The highest proportion of cells reached G2/M in B89 13, 1589 or 1689 hours following the HU release for D-Z47MG, U-251MG, D-245MG and D-263MG, respectively. Hydroxyurea is an effective agent for synchronizing glioma cell lines. Application of this procedure to biopsied human gliomas is a promising method to increase the mitotic index as well as chromosome length in these tumors.

96

17

DEVELOP~N~NT AND CHARACTERIZATION OF A HL~4AN EPENDYMOMA XENOGRAFT MODEL. Marc E. Horowltz, David M. Parham, Edwln C. Douglass, Larry E. Kun, Janet A.

Houghton, and P e t e r J . Houghton, St. Jude C h i l d r e n ' s Research H o s p i t a l , Memphis, TN 38101. C l i n i c a l and l a b o r a t o r y s t u d i e s of ependymoma have been c o n s t r a i n e d by the r e l a t i v e r a r i t y of the tumor and u n a v a i l a b i l i t y of a p p r o p r i a t e animal models. We r e p o r t the s u c c e s s f u l h e t e r o t r a n s p l a n t a t l o n of a human epesdymoma in CBA-CaJ mice Immune-deprived by i n f a n t thymectomy and whole body i r r a d i a t i o n . This x e n o g r a f t was e s t a b l i s h e d subcutaneously from e l o c a l l y r e c u r r e n t f o u r t h v e n t r l c u l a r ependymoma. The patient, an l l - y e a r - o l d g l r l , was t r e a t e d wlth 50 Gy to the posterior fosse at dlsgnosls. At local recurrence 20 months later she received 3 courses of MOPP chemotherapy without response and then tumor was resected. Flask tumor growth was first noted 8 months after heterotransplantatlon. The xenograft is Is its fifth passage and is histologically stable. It is a low-grade ependymoma with abundant true rosettes and perlvaseular pseudo-rosettes, moderate cellularity, flbrovascular trabeculae and so necrosis. These findings are consistent with the original tumor histology. RAP, NSE and neurofilament protein imunoperoxldase stains are negative both in the original tumor and the xenograft while SI00 protein Is positive Is both, Electron microscopic features include mlcrovilll and occasional cilia on the apical sides of cells bordering the lumen of a true rosette. The karyotype of the xesograft is human, 44xx, -5, -6, -ll, +14, -17 t(l:lb)(pl2:II), del(3q21) der(6) t(l:6)(q21:q21), del(10q24) + del(12pll) which is consistent with a DNA index of 1.0 from cytofluorometrlc analysis of the original tumor. The xenograft tumor volumedo~bllng time Is 48 days. Cell proliferation kinetic data derived in the xenograft wlll be presented. This tumor will provide a valuable model for the study of the blology end therapy of low-grade ependymoma. Supported by American Cancer Society award IN-99K, CA23099 from the National Cancer Institute, and by Americas, Lebanese, Syrian, Associated Charities.

1 9

c e n Kinetic Studies of M Situ Human Brain Tumors With lt-omodeoxyuridine (BrdU)

18

The Histochemical Localization of Glial Fibrillary A.cidic Protein (GFAP) in Human Central Nervous System (CN5) Neoplasms by Monoclonal Antibodies (MCAs). RE McLendon, PC Burger, C3 Wikstrand, CN Pegram, LF Eng, + and DD Bigner. Duke Medical Center, Durham, North Carolina and +Veterans Administration Hospital, Palo Alto, California.

GFAP has proved to be an important astrocytic marker in diagnostic neuropathology. However, the variability in the staining patterns that has been seen with the poiyvalent antisera (PVA) from institution to institution as well as from lot to lot of antisera has emphasized the need for an identical reagent for comparative immunohistochemical studies. We have prepared three MCAs designated IB4, 2El, and /~AII which are monospecific to GFAP by radioimmunoassay, immunoblot analysis ot cytoskeletal proteins, and immunohistochemistry of normal human brain. Utilizing a well characterized PVA against GFAP as the reference serum, a comparative assay of GFAP distribution was undertaken on formalin-fixed, paraffin-embedded sections of primary and metastatic CN5 tumors with each MCA individually at concentrations of 25 iJg/ml as well astogetherin a cocktail preparation of 75 IJg/ml total protein. Trypsinization prior to primary incubation was found to improve the staining patterns of all antisera. Although the staining patterns were not equivalent, each MCA positively stained I8/20 astrocytomas and 3/4 oligodendrogliomas. The cocktail preparation, however, positively stained 19/20 astrocytomas and 4/4 oligodendrogliomes. Further testing of the cocktail revealed staining in 27/28 GFAP positive astrocytomas, 5/5 gangliogiiomas, 6/6 oligodendrogliomas, 2/2 ependymomas. No GFAP was identified by any antisera in 2 choroid plexus papiUomas, /~ meningiomas, 4 schwannomes, I craniopharyngioma, 2 medulloblastomas, 2 chordomas~ or I/~ metastatic carcinomas. In no case did the cocktail solution stain cells that were not also stained by the PVA. As IBm, t a l l , and 2El identify epitopes restricted to GFAP and are available in an infinite supply and in a quantifiable and reproducibile titer, these MCAs, in a cocktail preparation, represent an identical reagent which could be used in an infinite number of institutions to standardize comparative studies.

20

A STUDY OF BRAIN TUMOR ASSOCIATED PROTEINS USING TWO-DIMENSIONAL GEL ELECTROPHORESIS

T. Hoshino

Brain Tumor Research Center, University of California, San Francisco, CA 94143 BrdU, like 3H-thymidine, is specifieally taken up by cells during DNA synthesis (S-phase), and ean be detected by anti-BrdUmonoclonal antibody (MAb). Over 100 patients with various brain tumors reeeived I.V. BrdU (200 mg/m 2) over 30-60 min at the time of eraniotomy. Biopsied materials were fixed in 70% ethanol, embedded in paraffin and deparaffinized seetions were reaeted with anti-BrdU MAb following denaturation of DNA with 2 N HCI for 30 rain. BrdU-labeled nuelei were satisfaetorily visualized in tissue seetions by the indireet peroxidase method, thus the percentage of BrdU-labeled nuelei per total number of cells (labeling index: LI) was obtained for eaeh tumor. LIs were 5-20% in glioblastoma multiforme, meduUoblastomas, and the majority of highly anaplastie astrocytomas. Most moderately anaplastic astroeytomas and ependymomas showed LIs of less than 1%. Pituitary adenomas showed Lie of less than 1% exeept for two eases whieh had Nelson's syndrome. Meningiomas also showed very low LIs (< 1%) exeept for 3 cases whinh were diagnosed as malignant meningiomas. These figures eoineide with the LIs obtained from autoradiographie studies done in the past decade and appeared to correlate well with the proliferative potential of the neoplasm based on its elinieal behavior. As BrdU is non-radioactive, non-toxie at the dosage used, and is detected readily by MAb, this method may be useful in estimating the biologieal malignaney of human brain tumors i.~n

si tu___. (Supported in part by grants PDT-159 from ACS and CA 13525 from NCI.)

Raj K. Narayan, M.D., William E. Heydorn, Ph.D., G. Joseph Creed, B.A. and David M. Jacobowitz, Ph.D. Department of Neurosurgery, Baylor College of Medicine, Houston, TX 77030 & Laboratory of Clinical Science, National Institute of Mental Health, Bethesda, MD 20205. Two-dimenslonal gel electrophoresis is being increasingly employed for protein separation. We have been interested in the application of this technique to the study of human brain tumor proteins. These studies have shown that each type of brain tumor has a fai;ly characteristic protein profile that could be developed as a diagnostic and prognostic adjunct. We have taken a step in this direction by semiquantltatlvely grading a large number of proteins seen in gels of several malignant brain tumors. By comparing these patterns with those seen in normal human cerebral cortex, certain tumor associated protein spots have been identified. In addition, using electroi~Inunoblotting and comigration techniques, we have identified twelve of the major spots seen on these gels. These proteins include albumin, actin, alpha and beta tubulin, neuron specific enolase (NSE), non-neuronal enolase (NNE), soluble glutamic oxaloacetic transaminase (sGOT), glial flbrillary acidic protein (GFAP), vlmentln, a 66-kD intermediate filament protein, the 68-kD neurofilament protein and the beta subunit of the guanine nucleotide regulatory proteins (G-protelns). Our current state of knowledge in this area will be reviewed and potential avenues for further investigation discussed.

97

21

ARACHIDONIC ACID METABOLISM AND gRAIN TUMORS. I 2 2 2 C.Chlabr~ndo , R.K~erich , G.B~tti , P.Gaetani ~ M.G. i Castelli , E.Cozzl , V.Silvani , and P.Paoletti . Laboratory of Environmental Pharmacology and Toxicolo~, 2Mario Negri Institute, Milan, Italy. Dept. of Surgery, Neurosurgical Sect., University of Pavia, Italy. Prostaglandins consist of a large group of araehldonate metabolites and have been detected In all tissues. They are involved in a large variety of biological effects and appear to play s regulatory role in many phisiological and pathological conditions. Authors have studied five stable metabolites of erachidonic acid (AA) via the cyelooxygenase pathway (PGE POD PGF , Prostacyelin 2' 2' 2alpha and TxB2) by high-resolution gas ehromatbgraphy mass-spectrometry in human brain tumor homogenates. Specimens of 25 intracrsnial tumors (8 Meningiomas, 9 Malignant Gliomas, 4 Blow-Growing Astrocytomas, 1 Medulloblastoma, I Acoustic Neurinoma, 2 Dermoids) were obtained at surgery, immediately frozen in liquid nitrogen and then analyzed for their potential capacity to synthetize prostaglandins and thromboxsnes from endogenous AA. Most tumors showed a very active AA metabolism. When available, normal brain tissue revealed a lower synthesis capacity as compsred to adjacent tumoral tissue. Similar AA metabolic profiles were observed in tumor specimens with similar histological characteristics. Many tumors showed high TxB and/or PGE synthesis. Low Prostscyclin synthesis was 2 observed i~ most tumors, while P G D and P G F . was produced in ~alpna variable amount. The aim Of this study iS to verify prostaglandln profiles and arachidonic acid metabolism in brain tumor and relate it to biological characteristics of the tumor.

Phenotypic expression of glutamine synthetase and pyruvate carboxylase in human brain tumor ceil lines. E;. LeM. Campbell, Dept. of Neurology, The Graduate Hospital, Philadelphia, PA 191~.6 and D.D. Bigner, Dept. of Pathology, Duke University Medical Center, Durham NC 27710.

23

Astrocyte specific markers are limited. However, two such markers are metabolic enzymes, glutamine synthetase and pyruvate carboxylase. These enzymes regulate the pools of glutamate and glutamine in normal astrocytes. E;lutamine has been shown in several tumor systems to be the major substrate for energy metabolism, in addition to its role in pyrimidine and purine synthesis. Recent investigations have examined the influence of glutamine on the growth of human rnedullobiastoma and glioma cell lines in vitro, as well as the effect of metabolites which interfere with glutamine synthesis or utilization. In view of the potential efficacy of such antimetabolites for new approaches to the chemotherapy of brain tumors, it is of interest to determine the phenotypic heterogeneity of enzymes associated with glutamine synthesis and utilization, particularly if these enzymes are restricted in CN5 to astrocytes. We have examined the expression Of two such enzymes, glutamine synthetase and pyruvate carboxyiase in six ceil lines, 2T, a human osteosarcoma cell line, TE-67i, a human meduiloblastoma celr line, U-i IS, U-2.~l, U373, and D-3q ME; human glioma cell lines. The osteosarcoma ceil line showed no evidence of either enzyme based on immunocytochemicai localization techniques. Pyruvate carboxyiase is expressed only in U-373, and TE-671. Three different phenotypes were identified, lines containing both enzymes, e.g., TE-671, cell lines only expressing pyruvate carboxylase, e.g., D-Stj ME;, and ceil lines expressing neither, e.g., U-115. To date we have not identified a cell line only expressing glutamine synthetase. The significance of these results will be discussed in terms of energy metabolism and giutamine utilization.

22

Patterns of Viral Genome Integration In Invasive and Noninvasive SV40 Produced Hamster grain Tumors John W. Walsh, M.D., Ph.D., Gregory Dulgou, Stephen G. Zimmer and Jean Oeltgen

Thls report presents an investigation using Southern blot-hybrldlzatlon of the patterns of integration of SV40 viral DNA In a series of invaslve and nonlnvasive primary intracranial tumors. The tumors were produced in 1-2 day old Syrian hamsters by iatracerebral inoculatlon of SV40 transformed hamster cerebral cortex cells. Soma tumors were twice reestabllshed in cell culture and inoculated intracerebrally to produce third generation tumors (T 3) and then cloned by dilution plating for anchorage dependence.These cloned tumors diffusely invaded adjacent brain and resembled human malignant astrocytle series tumors. Other tumors (T 1) were not passed but were cloned in soft agar for anchorage independence. They were uniformly noninvaslve. DNA from tumor cells of each group was then extracted, digested separately with BglI, geoRI, BamNI and Xbal restriction endonueleases and analyzed by blot-transfer hybridization. The results indicate that development or emergence of the invasive phenotype is the result of a dynamic process of rearrangement and selection. This conclusion is supported by the fact that different integration patterns are foand in the T T cell lines and their derived clones, and 1 3 that subclones of the cloned invasive cell lines show a genomle instability leading to rearrangement. This rearrangement seems to Occur primarily in the flanking cellular (genomie) sequences rather than within the viral DNA itself.

24

PRELIMINARY STUDY OF BRAIN TUMORS CELL KINETICS INVESTIGATED BY SERIAL STEREOTACTIC BIOPSIES. G. BregEi*, A. Costa*', A. Franzini**. * Ist. Neurologico "C. Bests", Milano (Italy). *" Ist. Nazionale Tumori, Milano (Italy).

Between July 1983 and December 1984, 33 consecutive patients affected from brain gliomas underwent serial stereotaetie biopsies to obtain the histological grading. The serial stereotactic biopsies was planned on mathematical transposition of the neu~oradiologieal (CT, NMR, angiography) lesion images. Multiple tissue sampling (2-4 specimens) was guided by intraoperative neurophysiologieal monitoring. The series includes 24 males, 9 femal~s, average age 37 years (6-70 ys). In these cases H thymidine in vitro testing of cell kinetics (Label Index; L.I.) has been performed. The cell kinetics analysis matched with WHO histological grading showed that L.I. ranges between 1.4% and 21.5% in glioblastomas; between 2.4% and 17.7% in anaplastic astrocytomas; between 0.6% and ll.O~ in mature astro= eytomas. Moreover when L.I. has been studied in the multiple specimens, an high intratumoral heterogeneity has been found. The value and the feasibility of cell kinetics studies applied to stereotactic biopsies as well as the possibility of identifying prognostic classes within brain glial tumors will be discussed. This study has been supported by

Supported in part by W.W. Smith Charitable Trust and NIH NSCA 200023-1, CA 1189g and CA 1t~236.

84.00474.44.

C.N.R.

grant N ~

98 SYSTEMIC METASTASES FROM OLIGODENDROGLIOMA Macdonald DR, Cairncross JG, Gilbert JJ London, Canada

25

Extra-cerebral

metastasis

of

(ODG) has rarely been reported.

oligodendroglioma

In 4 years we have

seen 2 men and 2 women (age 21-47 years, with systemic metastases from ODG.

mean 35.8)

All patients had

had multiple craniotomies (2-5) for resection of tumor and all had received cerebral irradiation.

Metastases

developed 1-26 months (mean 13.3) following most recent surgery and 35-102 months (mean 53.8) following initial tumor

surgery.

lymph

nodes

patients,

Metastases included (ipsilateral

scalp

(2),

appendicular skeleton;

to bone

regional

cervical

craniotomy)

in

(3,

spine

both

all

4 and

I with spinal cord compression),

and lung (1). The pathology was malignant ODG in both the

cerebral

and

systemic

tumors

in

all

patients.

Symptomatic systemic metastases responded well to local irradiation months with

in

2/2.

(median

12)

progressive

Three after

cerebral

died,

4-18

developing metastases,

patients have

all

and

One

systemic

tumor.

patient is still alive on BCNU chemotherapy.

Multiple

craniotomies,

survival

from initial

malignant

histology

and

long

surgery may be the predisposing factors

to systemic metastasis of ODG.

TEMPORAL DEVELOPMENT OF PERITUMORAL BRAIN EDEMA IN RABBIT VX2 BRAIN TUMORS. David Weissman, Stuart Grossman, The Johns Hopkins Ontology Center, Baltimore, MD 21205 The temporal development of perltumoral brain edema and its relationship to the development of neurologlc deficits were studied following implantation of VX2 carcinoma ceils into the frontal lobe of New Zealand White rabbits. Tumor implantation results in an intracerebral tumor, peritumoral brain edema, neurologic signs, and death in 13 ~ I days. Brain edema was quantltated by admlnisteriug Evan's Blue (EB) intravenously 24 hours before sacrificing the rabbits and removing the brain. EB was extracted from the tmmor containing hemisphere with dimethylformamide and q u a n t l t s t e d u s i n g s p e c t r o p h o t o m e t r l c analysis. S t u d i e s were performed in r a b b i t s a t d i f f e r e n t t i m e s f o l l o w i n g tumor i m p l a n t a t i o n to determine the temp o r a l development of p e r l t u m o r a l edema. The b r a i n s of c o n t r o l animals w i t h o u t tumor c o n t a i n e d an a v e r a g e of 5.1 0.9 ug of EB per c e r e b r a l h e m i s p h e r e . Rabbits i n j e c t e d w i t h s t e r i l e media w i t h o u t tumor c o n t a i n e d an e q u i v a l e n t amount of Evans Blue. Three days a f t e r tumor i m p l a n t a t i o n t h e r e was an a v e r a g e of 25.7 • 5.0 ug EB in the tumor cont a i n i n g h e m i s p h e r e . An e x p o n e n t i a l r i s e in EB e x t r a v a s a t i o n i n t o t h e tumor c o n t a i n i n g hemisphere was noted ~ x t h e time of tumor i n n o o u l a t l o n u n t i l death (y=14.3 e r=.97). No r a b b l t s developed u e u r o l o g i c d e f i c i t s u n t i l days f o l l o w i n g tumor i m p l a n t a t i o n . All symptomatic r a b b i t s had greater than 150 ug of EB in the tumor containing hemisphere. Intramuscularly administered dexamethasone (I0 mg b i d ) s i g n i f i c a n t l y reduces EB e x t r a v a s a t i o n i n t o t h e tumor c o n t a i n i n g hemisphere f o r a t l e a s t 10 days f o l l o w i n g tumor implantantlon. From t h i s s t u d y , i t a p p e a r s t h a t the o n s e t of n e u r o l o g i c d e f i c i t s i n t h i s b r a i n tumor model i s r e l a t e d to e x p o n e n t i a l l y i n c r e a s i n g p e r i t u m o r a l b r a i n edema. Dexamethasone r e d u c e s the e a r l y accumulation of p e r i t u m o r a l b r a i n edema.

27

~EXPRESSION OF EPIDERMAL GROWTH FACTOR RECEPTOR IN fq.J CULTURED HUMAN BRAIN TUMOR CELLS. W.K.Alfred Yung, G.E. Ga11ick, M.D.Waterfleld, R.P.Moser and P.A. Steck Depts of Neuro-Oncology and Tumor Biology, The Unlverslty of Texas M.D.Aederson Hospital and Tumor Institute, Houston, Texas 77030 Protein Chemistry Laboratory, Imperlal Cancer Research Fund, Lincoln's Inn Fields, London WC2A 3PX. Aberrant expression of epidermal growth factor receptor (EGF-R) gene has been recently described in some human glloblastomas. Due to EGF-R's postulated roles in growth regulation and its partial similarity to the erb-B oncogene and the ne____uuuoncogene,altered expression of EGF-R gene product was analyzed in terms of development and progression of gllomas. We have examined the expression of EGF-R on normal glial and glioma cells grown in vitro by several independent assays. We have identified EGF-R by: (a) iumunopreclpitatlon of metabolllcally radiolabeled cell extracts with monoclonal antibody R1 (b) phosphorylation of immunopreelpitated functional EGF-R-kinase complexes; (c) immunoblottlng of cell extracts with an antiserum against a yT@rb-B s y n t h e t i c p o l y p e p t i d e ; and q u a n t i t a t i v e b i n d i n g of L = ~ I - l a b e l l e d ; (d) epidermal growth factor (EGF) and (e) RI monoclonal antibodies to cells, lumunoprecipitates and immunoblots of the various glloma cell extracts r e s o l v e d by SDS-polyacrylamlde gel electrophoresis revealed a protein of M r- 170,000 which Identical to the EGF-R of A431 cells. In addition a Mr- 190,000 protein was also observed in the majority of glloma cell extracts by both immunologic methods. Phosphorylatlon of both the 170KD and 190KD proteins were observed. The majority of gllal and glloma cells bound 20,000-60,000 EGF molecules per cell as determined by Scatchard analysis. Similar results were obtained using the RI antibody. These results suggest the expression of EGF-R and an EGF-R-Iike (190KD) protein is common on cultured human glioma cells, and further blochemical characterization of the 170KD and 190KD proteins is being done to determine if the 190KD protein is specific to human gllomas. (Supported by grants NCI CA33027 and from John S. Dunn Research Foundation to W.K.A.Yung, and NIH RR5511-23 to P.A. Steck, and G.E. Galllck.)

~i o A N A L Y S I S OF EPIDERMAL GROWTH FACTOR RECEPTOR AND I.~ ASSOCIATED GLYCOPROTEIN ON HUMAN GLIOMA CELLS IN VITRO. P.A.Steck, G.E.Gallick, S.A.Maxwell, end W.K.A.Yung. Depts of Neuro-Oncology and Tumor Biology, The University of Texas M.D.Anderson Hospital and Tumor Institute, Houston, TX 77030. The expression of epidermal growth factor receptor (EGF-R; Mr- 170,000) and a closely related glycoproCein (Mr- 190,000) have been previously identlfied on cultured human glioma cells by several independent techniques. We have examined these two molecules biochemically to ascertain similarities and differences between the 170KD and 19OKD proteins. A kinetic pulse-chase experiment revealed that these two proteins did not represent precursorproducts of each other. After a 4h chase period the 170KD and 190KD protein were both present even though they exhibited different half llves. The two proteins generated relatively similar tryptic polypeptlde fragments as determined by Cleveland analysis. Similar phospboamino acids were seen after pbosphorylatlon of functional Immunoprecipltated EGF-R-kinase complexes and protein hydrolysates. The major phosphorylated residue was l~entifled as phosphotyrosine. Metabolic incorporation of [~H]mannose and binding of radlolabeled lecties indicated the 170KD and 190KD proteins expressed related carbohydrate moieties. However, t~e 190KD protein incorporated significantly less [ R]mannose than the 170KD, suggesting the 190KD glycoprotein does not represent an ever glycosylated form of EGF-R. Furthermore, both proteins were demonstrated to bind to epidermal growth factor covalently linked to an insoluble matrix. These results suggest a close similarity of the 190KD glycoproteln to EGF-R both structurally and functlonally expressed by cultured human glioma cells. The relationshlp between the 190KD glyeoproteln and the ne_._uuoncogeneprotein (Mr- 185,000) is being i n v e s t i g a t e d . (Supported by g r a n t s from NCI CA33027 and the John S. Dunn Research Foundation to W.K.A.Yung and NIH RR5511-23 to P . A . S t e c k and G.E. G a l l i e k . )

99 CNS T U M O R S IN AIDS IN A H O M O S E X U A L POPULA rION Richard L. Davis, Yuen T. So, and Dale E. Bredeson, D e p a r t m e n t s of P a t h o l o g y and N e u r o l o g y , U n i v e r s i t y of C a l i f o r n i a , School of M e d i c i n e , San Francisco, CA 94143. Since 1981, there have been an i n c r e a s i n g number of ~IDS cases in San F r a n c i s c o m a i n l y in male h o m o s e x u a l s . The m a j o r m a n i f e s t a t i o n of this s y n d r o m e in the CNS is i n f e c t i o u s but a s i g n i f i c a n t number of cases of n e o p l a s t i c i n v o l v e m e n t of this organ system have been found. We here report 13 cases, two with m e t a s tatic K a p o s i ' s sarcoma, and e l e v e n with lymphomas, the latter mostly primary in the CNS. Four cases p r e s e n t e d with AIDS and their CNS disease. In five cases, the CNS d i a g n o s i s was made at autopsy. All of the l y m p h o m a s were thought to be of B-cell origin, and there was s i m u l t a n e o u s i n v o l v e m e n t of s y s t e m i c o r g a n s in only one case. T h e r e were cases of i m m a n o b l a s t i c sarcoma, large cell lymphome, l y m p h o b l a s t i c sarcoma, and five cases were not s u b c a t e g o r i z e d . Three cases had m u l t i p l e lesions on CT scans. Multicentric i n v o l v e m e n t was seen in all a u t o p s y cases. The f i n d i n g s with CT and MRI were not c o n s i s t e n t , and though most of the l e s i o n s were enhancing, some were of low density and did not show a barrier defect. It is not p o s s i b l e to d i s t i n guish t o x o p l a s m o s i s from l y m p h o m a on the basis of the CT findings. All of our p a t i e n t s were male, and all but one were a d m i t t e d h o m o s e x u a l s . The age range was 26-60, with 2 being less than 30, 4 being 30-39, and 6 being 40-49. The mean time to death after d i a g n o s i s was 1.9 months. P o s s i b l e m e c h a n i s m s for the d e v e l o p m e n t of B-cell l y m p h o m a s in i m m u n o s u p p r e s s e d i n d i v i d u a l s in the s p e c i a l e n v i r o n m e n t of the CNS will be discussed.

29

30

D{~sseldorf, 4000 Oiisseldorf

I,

FaG

The RC2 clone was isolated from a s u b c u t a n e o u s t r a n s p l a n t a t i o n tudor (D74), the first passaRe of a Rlioma of the spinal cord of an inbred CDF offs p r i n R (C619), the mother of w h i c h received a single i n t r a v e u o u s injection of 50 mK/k R ENU on the 21st day of gestation. Tumor logy,

Rrowth a n d m o r p h o l o g y

were studied by h i s t o imlnuncytochemistry and m o r p h o m e t r y . RC2 clone, o r ~ i n a l l y isolated in 1973, presented i n t e r e s t i n g characteristics for a neuroectodermal cell type, such as SIO0 p r o t e i n and high content o f CNPase. S a t u r a t i o n density, d o u b l i n g time, c l o n i n g e f f i c i e n c y , k a r y o t p y e in the near diploid region, e x p r e s s i o n of b i o c h e m i c a l d i f f e r e n t i a t i o n and the morpholo~ly of the RC2 clone will be summ a r i s e d and c o m p a r e d with data after 5 and lO years. RC2 a p p e a r s as a cell type with a low p h e n o t y p i c and g e n o t y p l c v a r i a b i l i t y . RG2 is s u i t a b l e for tumor s p h e r o i d a g g r e g a t i o n in culture and can easily be t r a n s p l a n t e d to adult rodents.

I n t r a c e r e b r a l s t e r e o t a c t i e i n o c u l a t i o n of IO 2 to 105 cells into the right c a u d a t e nuclei of adult CDF and U l s t e r rats, leads to initial tumor R r o w t h in all animals. Tumor r e ~ r e s s i o n in n l l o g e n i c rats is i n f l u e n c e d by ~he number of cells inoculated. Tumor morphology, v a s c e l a r i s a t i o n and the d e v e l o p ment of n e c r o s i s have been d e t e r m i n e d by m o r p h o m e t r i c analysis, since these data are relevant for the study of p a t h o p h y s l o l o g l c a l p a r a m e t e r s .

of Human Malignant gliomas

into the 8rain of kormsl Adult Rnts. Salford t.G. ~ 8run A. and Str~mbled L-G. Dept. o f Neurosurgery and Neuropathology, Lund U n i v e r s i t y H o s p i t a l , Lund, Sweden. Aim: To develop a model for the study o f the behaviour o f human malignant gliomas in the brain o f normal, non immunosuppressed r a t s . Method: Specimen o f the tumor f r o n t o f astroeytomas grade I l l - I V were taken from 8 p a t i e n t s . H a l f o f the tissue was transformed i n t o a suspension o f tumor c e l l s and the other h a l f w a s studied pathoenatomically. Transplantation took place w i t h i n 2 hours a f t e r tumor r e moval from the p a t i e n t s . V i a b i l i t y studies proved the t u morcells to survive at high pereend;age f o r several hours in glucose solution and 40 C. Two sites in each hemisphere of adult Sprague-Dawley rats were injected stereotactically with each 5 p l of the suspension. After survival 1 - 3 weeks the animals were sacrificed by perfusion-fixstion. The brains were studied pathoanatomieally. Preliminary results: Specimen from 4 of the 8 patients resulted in tumor growth and one additional patient yielded a suspected take. The original glioma picture tended to adopt a more homogenous pattern, though still with vascular proliferation and mitoses. Conclusion: The model gives a high take rate for a nonimmunosuppressed host. Contrary to our previous method (ref) with solid transplants, this method allows for transplantation to any site in the brain. The model may constitute a useful model for the study of biology of human malignant gliomas. Ref: A model for Xenotransplantation of human malignant astrocytomas into the brain of normal adult rats. Str~mblad L-G., 8run A., Salford L.0. and Stenevi UIf. Acta Neurochirurgica 65, 217-226. 1982.

TIIE HAI.ICNANT RC2 GLIO2'IA CLONE: DEVELOP)lENT, CHARACTERISATIO~, AND INTRACEREBRAL SYNCENIC A~ZD AI.I.OOENIC TRANSPLANTATION TUMORS. N. W e c h s l e r , G. R e i f e n h e r R e r , .]. P l u m b a u m , U. P. T e s k e Department of Neurupathology, U n i v e r s i t y of 31

Xanotransplantation

32

Perspectives on Biological Response Modifiers for Tumor Therapy. G. Yancey Gillespie, Ph.D., Div. of Neurosurgery, Univ. North Carolina, Chapel H i l l , NC

Even with multi-modal application of the 3 traditional approaches to therapy (surgery, radiotherapy, chemotherapy) for patients with brain tumors, within the last few years we appear to have reached a relative plateau in treatment efficacy. While new surgical techniques and improved methods of radiotherapy and chemotherapy continue to evolve, primary intracranial neoplasms remain an intractable cancer for which novel approaches to therapy must be sought. Immunotherapy, from a historical perspective, has been envisaged as a fourth "Horseman" in the multi-modal approach to therapy. However, efforts to specifically and/or non-specifically stimulate antimtumor immune responses have been largely disappointing. Lack of adequate quantities of pure, defined reagents together with an incomplete understanding of the complex series of actions each immunotherapeutic compound has on the individual's overall biologic response have contributed to the general lack of efficacy reported for immunotherapy. The advent of molecular biology techniques in monoclonal antibody production and mammalian gene cloning holds the promise of providing ultrapure biological compounds in sufficient quantities for meaningful clinical trials.

Among these reagents are cytokines, lymphokines and monokines such as the Interferons and the Interleukins. I t is becoming increasingly apparent that these very potent biological response modifiers (BRMs) may have both direct and indirect anti-neoplastic activities in the tumor bearing host. As this area of inquiry broadens, biological response mechanisms are now being viewed as encompassing more than those classically associated with the immune system. Besides activation of the immune system, other important mechanisms for controlling neoplastic growth include factors that control growth and differentiation of tumor cells as well as those that control its vascular supply. Correlation of in v i t r o effects of these agents with in vivo responses in cancer patients is shedding new l i g h t on the potential impact of BRM therapy.

100 INTEPJ~CTIONS AMON~ INTERFERON, GLu NATURAL KILLER CELLS AND HUMAN GLIOMA$ Allan J. Yates, Ralph E. Stephens and Ronald L. Whlsler, Department of Pathology (Neuropathology) and Medicine (Rheumatology), The Ohio State University, Columbus, Ohio 43210 33

Human beta interferon (IFN-~) inhibited the growth of 3 cell lines cultured from human glloblastoma (GH) but not human fetal brain cells. For the growth of GM cells to be inhibited IFN-~ (1000 unlts/ml medium) had to be present in the medium for at least 24 to 48 hours but the growth inhibitory effect persisted for up to 3 weeks. The earlier IFN- 8 was added, the greater was growth inhibition. Although it has been suggested that ganglloside may be a part of IFN-Breceptor, preincubation with GM2 or a mixture of human brain gangllosides had no effect on growth inhibition due to IFN-B . Cells cultured from I0 human neural tumors and 3 human brains were tested in a 4 hour chromium release assay for sensitivity to natural killer (NK) cytolysis. Only 2 tumor but both fetal brain cell lines tested were sensitive. Results of target binding and single cell lysis assays indicated that resistance could be due either to low levels of target binding or lysis separately, or to a moderately low level of both in combination. Exposure of 4 glloma and 2 fetal brain cells lines to IFN-B prior to NK assay in general decreased the sensitivity of these targets to NK lysls and increased their total neutral glycollpid and ganglloaide contents. Therefore, glycollplds may participate in the mechanism of resistance to NK lysls. Exposure of effector cells to human IFN-O or IFN-51ncreased their lytic activity towards these targets. The results of these studies indicate the complexity of the biological effects that IFN therapy for human gllomas may have: (a) decrease glloma growth; (b) increase NK activity against glioma targets; (c) increase r e s i s t a n c e of gliomas to NK lysls.

34

Interferon Modulation of Glioma Cells N. de Tribolet, A.C. Diserens, S. Carrel, J.P. Much Neurosurgical Service, CHUV, I011 Lausanne, and Ludwig Institute for Cancer Research, 1066 Epalinges, Switzerland

The modulation of HLA-DR and HLA-ABC antigens by recombinant IFN-gamma was studied on 15 malignant glioma cell lines established in our laboratory and 8 clones or subclones derived from 2 of these lines. Comparative studies were performed with recombinant IFN-gamma. The results not only confirm the selective a c t i v i t y of IFN-gamm~ on the modulation of HLA-DR antigens expression, but also demonstrate a marked heterogeneity in the response of glioma cell lines and their clones to the two types of IFN tested. This heterogeneity does not seem to be due to the absence of the receptor for IFN-gamma on the surface of these cells since almost all of the cell lines or clones tested responded to IFN-gamma by the induction or enhancement of either HLA-DR or HLA-ABC antigens expression (or both). The heterogeneity of induction was also demonstrated between clones derived from a glioma line which did not express HLA-DR after IFN-gamma treatment. The production of HLA-DR antigens by one of the clones was abundant enough to be confirmed by immunoprecipitation and SDS-PAGE analysis. I t thus might be speculated that enhancement or induction of HLA-DR expression by injection of IFN-gamma could help malignant astrocytes to present their own tumor associated antigens to T cells and thus enhance their immunogeneity.

This work was supported by grants from PHS CA-31564 and 5P 30-CA-1605B, The Department of Pathology and College of Hediclne~ Ohio State University.

35

The glioblastoma cell derived T cell suppressor factor (G-TsF) Fontana,

A. and Bodmer St.

Section of Clinical Neurosurgery University

Hospital,

Immunology

and Department of

Zurich, Switzerland

Previous studies have suggested that either antibodies or not yet defined soluble factors may induce a cellular immunodeficiency state in patients with glioblastoma. The results of detailed analyses of glioblastoma cell derived factors provide strong evidence that g]ioblastoma cells release factors (G-TsF) which inhibit T ceil activation. The G-TsF leads to inhibition of lectin stimulation of mouse thymocytes or human peripheral blood mononuclear cells. Neither interleukin -I nor interleukin -2 are able to overcome the G-TsF efFects. The G-TsF appears to be equivalent to a glioblastoma cell derived Factor which inhibits the growth of mouse neuroblastoma cells in-vitro. This conclusion is drawn from biochemical studies in which G-TsF copurifles with the glioblastoma cell derived neuroblastoma growth suppressor factor. The availability of pure G-TsF will allow detailed studies of both the interactions of the factor with its target cells and the demonstration of analogous factors in glioblastoma patients.

36

IMMUNOTHERAPY OF PATIENTS WITH PRIMARY GLIOMAS M. S. Mahaley, Jr., M.D., Ph.D., Division of Neurological Surgery, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina.

The author will attempt to summarize the various clinical trials which have been conducted referable to Immunotherapy of patients with primary malignancies of the brain. The summary will include trials with general immune stimulation, adoptive i~m~unotherapy, serotherapy, and active immunotherapy as well as the use of interferon.

101 Mouse Monoclonal Antibodies to Human Glioma Ceils. Bigner, D.D., Wikstrand, C.3., Bullard, D.E. and Humphrey, P.A., Duke University Medical Center, Durham, N.C. A group of murine monoclonal antibodies made against or reactive with human gliomas have been evaluated in vitro and in human glioma xenografts prior to initial clinical investigation. Mab 81C6 which reacts with a novel extracellular matrix glycoprotein has been most extensively investigated. In multiple experiments, large amounts of total percent injected dose of radioiodinated glC6 have localized specifically in human glioma xenografts. Up to 1696 of total injected dose has been localized to tumor with tumor-to-brain ratios as high as 235:1. Comparative imaging of intracerebral xenogralts has shown that higher quality scans can be obtained with B i t 6 compared to nonspecific IgG2b immunoglobulin. Moreover, tumors as small as 20 mg can be detected with $1C6 whereas tumor weights of at least 3it0 mg were necessary for detection with nonspecific IgG2b. glC6 specifically imaged intracerebral tumors in 16117 animals injected with antibody whereas myeloma IgG2b only imaged 3/9 successfully. Comparing routes of administration there was no detectable difference in intracarotid versus intravenous injection of glC6 localization in intracerebral xenografts. Intratumoral injection compared to intravenous injection gave both higher initial localized doses and higher maintained localized doses. Fab fragments of 81C6 localized faster than whole immunoglobulin with tumor-to-brain ratios 2.5 times greater than whole immunoglobulin, although total percent localized dose with Fab was much less than with whole immunoglobulin. Statistically significant growth delays were obtained in mice treated intravenously with 1 mCi 1311 81C6 bearing subcutaneous glioma xenografts. In addition to Mab 81C6, significant in vivo localization has been obtained with Mel-14, U313A, C12, Di2 and F9, murine Mabs that are cross-reactive with human gliomas. Now that successful in vivo localization, imaging, and therapeutic efficacy with minimal to no toxicity have been shown in human xenograft bearing animals with routine Mabs reactive with human gliomas, Phase I clinical trials are being developed.

37

CLINICAL PHARMACOLOGY OF ANTINEOPLASTIC AGENTS IN THE CNS. V.A. Levin. Brain Tumor Research Center, University of California Medical School, San Francisco, CA 94143 39

The application of pharmacologic principles to the use of anticancer agents in patients is appropriately called the clinical pharmacology of anticancer agents. The goal of clinical pharmacology is the design of dosage regimens to achieve the therapeutic objective, the control of cancer growth. The means by which clinical pharmacology attempts to achieve this objective is by i) measurement of drug and metabolite levels in body fluids (i.e. plasma, urine, cerebrospinal fluid, etc.); 2) extrapolation of drug effects in culture and in animals to humans; 3) computer simulation and mathematical modeling of drug level measurements in these body fluids; and 4) the optimization of drug delivery to the cellular site(s) of drug action that minimizes systemic toxicity. To improve drug delivery to tumors without increasing systemic toxicity, a number of approaches have been taken. Increased tumor drug levels can be achieved if the amount of drug entering the tumor is increased or the amount leaving is decreased. The latter is impractical, but the former is possible. To increase the amount entering CNS tumors, the following regional approaches have been advocated and tried:.l) intra-arterial injection, infusion, and ,effusion; 2) intra-tumoral injection and infusion; and 3) circulation of drug over the tumor bed using intrathecal or intraventricular injection, infusion, and perfusion for leptomeningeal neoplasia. Targeting of drugs with antibody-coupled drugs, liposome encapsulated drugs, and antibody-liposome complexes are additional approaches that have been considered.

CLINICAL EXPERIENCEWITH ANTI-GLIOMA MONOCLONALANTIBODIES. H.B. Coakham, J.T. Kemshead, A.G. Davies, R.B. Richardson, S.P. Bourne, D.H. Jones. Department of Neurosurgery, Frenchay Hospital, Bristol, UK and Imperial Cancer Research Fund Laboratory, Institute of Child Health, London, UK. Monoclonal antibody UJl3A recognises a g lycolipid surface antigen which is neuroectoderm-associated. Radiolabelled UJI3A given intravenously will cross the blood brain barrier (BBB) in various types of tumour (12 cases). Scintigraphic, biokinetic and serological studies have been carried out. The majority of tumours were imaged and MCA uptake generally correlated with 99mTc glucoheptonate uptake. However, a variation in BBB permeability for 13l-I UJI3A was seen, with one glioma (low grade) showing no uptake of MCA. The average T~ values were tumour 179 hr., and blood T~ (1)=I.8 hr., T~ (2)=53 hr. Naturally occurring antimouse lgG was detected in the pre-injection serum of one patient which resulted in antibody aggregation in vivo. Analysis of resected tissue revealed tumour:brain ratios of over 12:l, with higher values in necrotic tumours. However, using paired label analysis with 131-1UJI3A and 125-I HMFG(control), specific uptake could not be demonstrated despite success in an animal model. In contrast, good in vivo specificity was achieved in a case of neoplastic meningitis from pineocytoma when the intrathecal route was used. An anti-neuroblast MCA, 'UJl81.4 (13l-I) was given together with a control MCA (125-I) producing specificity ratios of over 20:I (external counting) and lO:l (CSF cell pellet). A therapeutic dose of 131-1 uJ]gl.4 given intrathecally resulted in marked clinical improvement and a continuing remission (10 months). Our data suggests that further clinical t r i a l s of MCAtargeted intrathecal therapy may now be considered but that the problems of antibody delivery to solid tumours require continuing experimental study, particularly with regard to vascular permeability and the non-specific accumulation of proteins. 38

40

PET Measurements of Brain Tumour Blood Flow and Transcapillary Transport.

D.G.T. Thomas, D. Brooks, R. Beaney, T. Jones and E. Yoshimo. The National Hospitals for Nervous Diseases and MRC Cyclotron Unit, Hammersmith Hospital, London. The ECAT-II tomograph has been employed to study cerebral gliomas by using positron emitting isotopes labelled tracers of biological interest. The oxygen-15 steady state method has been used to measure blood flow and oxygen consumption in the tumour and surrounding brain and to observe the abnormalities before and the changes after surgery, steroid treatment and irradiation. Rubidium, Methyl glucose and Albumen have been used as tracers to study the transcapillary transport characteristics of human brain tumours in vivo.

The results of these investigations

will be described.

The practical problems

of such studies in human subjects as well as problems in design and interpretation of quantitative tracer models will be discussed.

102 41

6BGALLIUM-EDTA POSITRON SCANNING IN MALIGNANT BRAIN TUMORS: BLOOD-BRAIN AND BLOOD-TUMOR TRANSPORT KINETICS AND IMPLICATIONS FOR CHI~OI~ERAPY

D.C. Wright, C.A. K a ~ k i , P.G. Baldwin, R.E. Carson, C.S. Patlak, R.G. Blasberg. National Institutes of Health, Bethesda, Md. 20205 The transport kinetics across tumor capillaries were studied in patients harboring glioblastoma multiforme using dynamic positron computed tomcgraphy (PCT). Timed arterial blood samples and sequential (OCT scans) tissue data were collected over a two hour period following a 2.5 mCi bolus of intravenous 68-GalIiL~-EDTA. The tissue-blood distribution of the tracer could be described by a two compartment model and non-linear least squares analysis of the data was performed to yield the following values in a typical patient:

Whole tturor TLu~or rim Tumor center Brain

~/tOOg

K1 ul/g/min

Ve ml/lOOmg

tl/2 sin

2.1+.2 2.2~.2 2.2~.4 1.8_~.1

14.8+.4 16.6~.5 18.9_~.9 <0.1

38.9+1.0 41.4u 52.4_~4.2 -

18 17 19 -

where Vp is the tissue plamna volume, K] is the blood-totissue or influx constant, Ve is the tigsue distribution vol~ne and is equivalent to the extracellular space, and tl/2 is the equilibration half-time. The scans demonstrate considerable variability in the above parameters for tLmlors in different patients as well as within a single ttFaor of an individual patient. Edematous brain regions had very low influx constants which were indistinguishable from normal brain. Anatomical correlation was also made with CT and MRI, and companion fluoro-deoxyglucose OCT scans. These studies clearly demonstrate the feasibility of using the kinetic information obtained from GalIium-EDTA I=CT scanning in the development of optimal therapeutic strategies and in determining the efficacy of therapy.

4 3

Comparison of I n t r a v e n o u s (IV) V e r s u s [ n t r a e a r o t i d (IC) T h e r a p y with 1 , 3 - B i s - ( 2 - e h l o r o e t h y l ) - 1 - N i t r o s o u r e a (BCNU) in a Rat B r a i n - T u m o r Model

Dennis E. B u l l s r d , M . D . , Duke Univ. Medical C e n t e r We h a v e e v a l u a t e d the dose r e s p o n s e c u r v e for the i n t r a v e n o u s ( i . v . ) and i . e . administration of BCNU in a commonly utilized e x p e r i m e n t a l b r a i n t u m o r model, the 9L r a t giiosarcome. An initial toxicity trial utilizing the i n t r a p e r i t o n e a l ( i . p . ) LD10 of BCNU b y the i . v . and i . c . r o u t e s failed to d e m o n s t r a t e a n y significance in toxicity b e t w e e n the two r o u t e s . T u m o r - b e a r i n g animals were t h e n t r e a t e d on Day 15-16 a f t e r tumor inoculations with 1, 10, 25, 50, 75, and 100% of the LD10 dose by e i t h e r the i . v . or the i . e . r o u t e . Both i . v . and i . e . BCNU gave maximum survival increases at 75%-100% LDI0 doses and there was no therapeutic advantage was seen from i.c. delivery. However, at 50% of the LD10 dose (6.65 mg/kg), triplicate experiments demonstrated that the i.c., but not the i.v., dose maintained maximum efficacy equivalent to 100% of the LDI0 given either i.v. or i.c. W h e n the dose was reduced to 25% of the LD10 close (3.33 mg/kg), two of three experiments showed efficacy of the i.e. delivery of this lower drug dosage to be equivalent to 100% of the LD10 given i.v. or i.e. The i.v. dosage resulted in a significant reduction in survival in all three trials. At 10% of the LD10 dose (1.30 mg/kg), neither the i.v. nor the i.e. administration retained equivalent efficacy to 100% of the LD10. However, in one of two trials, the i.c. groups bad statistically better survival than controls, while in neither experiment was any advantage over controls seen in the i.v. treated groups. At 1% of the LD10 dose, neither the i.v. nor the i.c. route demonstrated any therapeutic efficacy. From our data, the advantage of the i.e. delivery of B C N U in the intracranial 9L rat giiosarcoma appears to be in the fact that significantly lower dosages than that given i.v. may be utilized to achieve equivalent survival with potentially less systemic toxicity.

THE QUANTITATIVE MEASUREMENT OF CAPILLARY TRANSFER CONSTANTS AND VASCULAR VOLUME OF IODINATED COMPOUNDS IN HUMAN BRAIN TUMORS WITH COMPUTED TOMOGRAPHIC METHODS. Dennis Oroothuis, Department of Neurology, N o r t h w e s t e r n U n i v e r s i t y Medical School, Evanston Hospital, Evanston IL. We have d e v e l o p e d a method to m e a s u r e the bidirectional capillary transfer constants of iodinated c o m p o u n d s in brain tumors with the CT scanner. The method uses a two c o m p a r t m e n t model (plasma vascular space as one c o m p a r t m e n t and extravascular tissue space as the other compartment), and d e p e n d s upon being able to m e a s u r e the amount of iodine in tissue (Am) and plasma (An) at d i f f e r e n t points in time. Am is r e l a t e d to the e x t r a v a s c u l a r amount of iodine by the expression Am = Ae + VoAm, where A e is the amount of iodine i n the e x ~ r a v a s c u l a r tissue, and V o is the tissue plasma vascular volume. A differential e q u a t i o n that d e s c r i b e s the rate of c h a n g e in e x t r a v a s c u l a r tissue is: dA = KIA p - k 2 A e

42

These two e x p r e s s i o n s can be used to o b t a i n v a l u e s of the transfer c o n s t a n t s K I (tissue-toblood) and k 2 (blood-to-tissue) and Vp. To do the studies, m e g l u m i n e iothalamate is infused at a c o n s t a n t rate over 5 minutes; serial CT scans and arterial blood samples are obtained at timed intervals over a 30 m i n period. The pairs of timed data: (Am, t) and (A , ) were fitted to the equation above with noPnlinear least squares methods. R e c o n s t r u c t i o n of the entire CT d a t a m a t r i x was p e r f o r m e d by using pixels from the timed series of CT scans, r e s u l t i n g in images of K I, k 2 and Vo, permitting regional anallysis of each of these parameters. This type of study has been performed on dogs with A S V - i n d u c e d g l l o m a s and on human brain tumor patients, and demonstrates the m a r k e d regional v a r i a b i l i t y in all three parameters that occur in brain tumors.

ESTIMATING ENTRY RATEAND EXPOSUREOF CELL POPULATIONS EXPERIMENTAL RAT BRAINTUMORSUSING QUANTITATIVE AUTORADIOGRAPHY(QAR). W.R. Shapiro, Memorial Sloan-Kettering Cancer Center, New York, NY. A major cause of chemotherapy failure in brain tumors is thought to be the emergence of resistant cell populations because of inadequate drug entry. In most models of drug entry into brain tumor, total drug content is measured in pieces of tumor rather than at a macrocellular level. QAR quantifies regional drug entry, thus permitting in vivo estimates of the proportionate exposure of cells~o given drug concentrations. I~C-MTXentry was measured in rats bearing intracerebral C$ gliomas after unilateral intracarotid (IC) adminlstration (25-50 uCi, 0.56-1.13 mg) without or with hyperosmolar mannitol to open the BBB; the contralateral side represented the systemic drug phase. Drug entry was determined at 5, 10, 20, 40, 60 and 90 min. The rat@ of drug entry KI in ml/g-min and drug exit K2 in min-• and the K1/k~ ratio, lambda, was estimated using a least-squares curve f l t t i n g program. There was considerable variation in the values, but i t was possible to identify only some of the sources for the variation. For example, previous studies with 14C-AIB demonstrated a direct relationship between the tumor size (area) and the K1 value. Because the tumors varied in size among the rats, we used only the highest 10% of the K values in the MTX experiments. The i n i t i a l estimate for the K1 on the side contralateral t 9 the infusion was 0.015 ml/g-min and for k2 i t was 0.14 min-• yielding a lambda of 0.7g ml/g. Tumor cell exposure was estimated as follows: In the cross section of the tumor digitized by QAR, the frequency distribution of picture elements (pixels) contained within a 0.0-5.0 ug/g range of drug concentrations was determined. For these measurements the percent exposure to several drug concentrations as well as the ED50 concentration was determined. Both the percentage of cells exposed and the EDso concentration were lower on the l e f t sided brain tumor exposed to systemically administered 14C-MTX than they were on the right sided which was exposed to IC drug administration. Depending on the technique, 5-90% of the cells would not be exposed in high enough concentrations of MTX to produce adequate cell k i l l . Such cells are the likely precursors of resistant cells.

44T0 14C-MTX IN

103 Cellular Transport and Accumulation of Methotrexate (MTX). Kamen, B.A. and Kolhouse, J.F. Dept. of Pediatr/Pharmacol, Univ. of Texas Health Sci. Center at Dallas, TX and Dept. of Int. Med. Univ. of Colorado Health Sci. Center, Denver CO. MTX, a folate analogue, has been in clinical use for 35 years. Its mechanism of transport and more recently its metabolism to a polyglutamate derivative in vitro has been extensively studied. Multiple phTrmaco~etic studies have also been done. MTX transport is competitive with folate in vitro and both have a Kt of 0.5-10 pM in several huma-n-an-6-4--rodent leukemia and hepatocyte cell lines. Most of these studies have been done in cells grown in medium containing 100 to 1000 times more f o l i c acid than folate present in serum. This results in an intracellular folate pool that is higher in vitro than in vivo. Our recent studies in vivo demon%-trate-t-e-4-a rapTd equilibrat~on o~ plasma MTX-and--l-lver tissue over a 4 log range (10- -10" M) but after efflux the remaining intracellular metabolite(s) of MTX only increased by 2-3 fold despite the marked increase in plasma concentration. This suggested i n i t i a l transport was not rate limiting with regard to intracellular metabolism. Transport studies of MTX and the naturally occurring, main serum folate, 5CH~tetrahydrofo~ate, in cells growing at physiological fol~te (5-10x10"=M) in vitro reveal at Kt of 2-5 nM and the presence of a spe-61iClC~turable receptor not identified in presence of high f o l i c acid. This data, derived in vitro, coupled with the previous data in vivo, allows s-6vera-'raT-suggestions: a) prior models of MTXand folate transport in vitro have given a r t i f i c i a l l y high, possibly non-physioi-ogi--c-~--,transport constants;b) a specific, high a f f i n i t y receptor for MTX ~nd fol%te exists, functional at physiologic folate (10-" - 10- M) and, c) treatment plans for obtaining specific MTX concentrations in situ which were extrapolated from in vitro data w i l l ~ i l ~ result in higher than needed concentration of MTX for maximal intracellular metabolism in vivo. Clinical and in vivo t r i a l s of low-dose MTX are supporting these con--/cepts.

Applications and Limitations of in vivo Measurements R.G. Blasberg, M.D., NIH, Bethesda, Maryland 20205. Drug concentrations in blood and tissue have been measured under various experimental conditions in order to assess the delivery to and exposure of tumor and various organs to drugs used in the treatment of cancer. In most cases the measurement of the activity-time course of drug (as well as active metabolites) in blood is reported; less frequently these measurements are made in tumor and organ tissues of experimental animals. Infrequently, single measurement values in human tumor tissue are reported. I t is now possible to measure the activity-time course of selected test molecules and some drugs in human tumors using positron emission tomography (PET). This data can then be analyzed using various pharmacokinetic models to yield appropriate parameters that can be correlated with therapeutic response and efficacy. In the near future i t may be possible to design brain tumor chemotherapy protocols that are based on measurable parameters such as tumor blood flow, vascular permeability, and the rates of drug clearance from blood and tissue. Thus, chemotherapy may be "individualized" on the basis of measurable parameters in a given tumor. The limitations of this approach relate to the technology required to obtain such measurements and to the kinetic models that are used to analyze the data. PET is expensive and technically complex. Although activity-time course data is readily obtai,ed, issues such as resolution (SxSxlOmm for the latest machines) and the rapid appearance of radiolabeled metabolites must be considered. Often the system we wish to measure is considerably more complex than the kinetic model used to describe the actual process. Thus, i t is important that rigorous validation of a particular measurement be insisted upon. Finally, do PET measurements t e l l us anything about the microenvironment of brain tumors? Yes, but we must recognize that the parameter estimates are approximations of the actual condition of the tumor averaged over the volume of measurement.

MORPHOMETRIC STUDIES OF NORMAL BRAIN AND BRAIN TUMOR CAPILLARIES: IMPLICATIONS FOR BRAIN TUMOR THERAPY. Peter Molnar*, Dennis R. Groothuist, and Nicholas A. Vick,. *Dept. of Pathology, University Medical School of Debrecen, Debrecen, Hungary, and ,Dept. of Neurology, Northwestern University Medical School, Evanston Hospital, Evanston, IL. The amount of drug that will enter a brain tumor is dependent upon the permeability (P) and total capillary surface area (S) of the capillaries available for exchange. In addition, the distance that drugs must travel to reach tumor cells once having crossed the capillary wall is dependent upon the intercapillary distance (ICD). We have developed computer-assisted methods for obtaining measurements of capillary structure at the light (LM) and electron microscopic (EM) levels. At the LM level we have measured total capillary surface area (S), diameter, density, and intercapillary distance (ICD). At the EM level we have measured capillary wall thickness, and number, size and density of fenestrations, interendothelial junctions, endothelial vesicles and endothelial discontinuities. These data were obtained from several normal brain regions and from rat brain tumors from the RG-2, ethylnitrosourea (ENU), and D54-MG brain tUmor models. In comparison to values from normal brain, our resuits indicate a marked increase in capillary diameter, decrease in capillary wall thickness and capillary density, and increase in ICD in these brain tumor models. Increases in "permeability" in these tumors is best explained by structural changes (e.g., fenestrations), which increase P, and not by increased capillary surface area, S. Changes in those capillary structures which explain an increase in P, particulary fenestrations, correlate closely with previous measurements of blood-to-tissue transport in these brain tUmor models.

4 8 ADMINSTERED CREMOTHERAPEUTIC AGENTS: A QUANTITATIVE

45

47

46of Drug Delivery.

THE INTRACEREBRAL PENETRATION OF INTRAVENTRICULARLY ADTORADIOGRAI~IC STUDY. Stuart A. Grossman aud Carla S. Reinhard. The Johns Hopkins Ontology Center, Baltimore, Md. 21205. Intrathecal chemotherapy is associated with significant neurologic toxicity which is poorly understood. This study was performed to determine the intracerebral penetration of iutraventrioularly admlnstered methotrexate. Subcutaneous reservoirs and ventricular catheters (SRVC) were placed in New Zealand White rabbits. Proper catheter placement was documented by ventriculography. Tritiated methotrexate was injected through the SRVC and the rabbits were sacrificed one hour later. The brains were removed, frozen, sectioned and placed on autoradiography film. The images were digitized and computer assisted image analysis was used to quantitate regional methotrexate concentrations. Several rabbits were injected with trltiated inulin for comparison. Approximately 50 percent of the area of each brain section was exposed to the radiolabelled m e t h o t r e x a t e or i n u l i n one hour a f t e r i n t r a v e n t r l c u l a r administration. Gray m a t t e r a d j a c e n t to t h e c e r e b r o s p i n a l

fluid (hippocampue, thalamus, caudate nucleus, and periaquaductal gray) contained the highest concentrations of tritiated methotrexate. Large white matter tracts were poorly penetrated by the tracers and appeared to act as a p h y s i c a l b a r r i e r to t h e d i f f u s i o n of the r a d l o l a b e l l e d compounds. The r a p i d and e x t e n s i v e p e n e t r a t i o n of intraventrlcularly administered methotrexate provides insight i n t o t h e p a t h o g e n e s i s of m e t h o t r e x a t e - i n d u c e d neurotoxiclty. The absence of a s i g n i f i c a n t c e r e b r o s p i n a l fluld-ependyma b a r r i e r f o r m e t h o t r e x a t e and f o r l a r g e compounds, such as i n u l i n , should be c o n s i d e r e d i n p l a n n i n g the d e l i v e r y of a n t l - n e o p l a s t i c agents into t h e b r a i n .

104 PHARMACOKINETICS OF IgM MONOCLONALANTIBODY DELIVERY TO NORMALBRAIN WITH AND WITHOUTOSMOTICBLOOD-BRAIN BARRIER OPENING. E. Neuwelt, P. Barnett, C. McCormick, J. Minna, and E. Frenkel. OregonHealth Sciences University, Portland, Oregon 97201, University of Texas Southwestern Hed School, Dallas, TX 75235 and MIH, Bethesda, MD 20814 Pharmacokinetic parameters of iodinated monoclonal antibody (MAb) delivery to normal rat brain were evaluated. The PA (permeability X capillary surface area) to IgM MAb (MW 10,000,000) was O. Employing osmotic blood-brain barrier (BBB) disruption, using intra-carotid 25% mannitol, the PA increases to 5.51. The intravenous (I.V.) and intracarotid (I.C.) routes of MAb administration and/or osmotic BBB disruption were studied. There was not a significant effect of the route of administration. However, osmotic disruption significantly (p < .0005) increased MAb delivery to brain which was independent of route of administration. Following BBB disruption and I.C. MAb administration, the maximum concentration in brain at 1 hour was 0.72% of the total administered dose and decreased by 70% over 24 hours. For 6 hours following barrier disruption, ipsilateral brain levels were 25100 fold greater than either in the contralateral hemisphere or non-barrier disrupted controls. The delivery of radiolabeled MAb to ipsilateral brain after BBB disruption was linear over a dose range of O.5-5.0pg IgM MAb while the percentage of the total administered dose delivered remained unchanged. Following disruption, the barrier was maximally open to MAb for I-2 minutes. Bolus administration 120 minutes after disruption, decreased MAb delivery 89%. When a thyroid blocking agent (which reduced thyroid uptake of 1251MAb by < 95%) was administered prior to BBB disruption and MAb, the brain levels at 3 hours were unaffected though the serum levels were elevated; whereas at 24 hours, the ipsilateral brain levels were doubled. The type of anesthesia used was found to significantly effect MAb delivery to brain after barrier disruption. Thus, these studies have provided valuable pharmacokinetic information which can be applied to the study of MAb delivery to brain.

49

51

Interferon gamma (IFN gamma) can induce the expressien of MHC Class I I antigens on cultured human astrocytes. M. Pulver, E. Frank, A.C. Diserens, S. Carrel, N. de Tribolet Neurosurgical Service, CHUV, 1011 Lausanne and Ludwig Institute for Cancer Research, 1066 Epalinges Switzerland

Antigen presentation by accessory cells is restricted by the presence of Class I I major histocompatibility (MHC) antigens HLA-DR or la. IFN gammamodulates the expression of these antigens providing a regulatory mechanism for local immune reactivity. Within the CNS, HLA-DR expression has been recognised on reactive astrocytes and on glioma cells. In order to investigate the modulation of HLA-DRon normal astrocytes, two cell lines were grown from a 20 week old human fetal brain. The brain did not express HLA-DRwhen studied by inwunohistology. The two cell lines, FB I and FB I f , derived from i t , express GFAP. FB I was DR negative from the f i r s t passages, but can be induced to express DR when treated with 500 U/ml IFN gamma. FB I I was DR positive in the early passages, lost this expression at I I passages, and is also DR inducible with IFN gamma, as demonstrated by RIA and SDS-PAGEusing DI-IZ MAB. The DR phenotype could be determined by double immuno fluorescence cytotoxicity only after IFN gammatreatment, and was shown to be DR4, DR6, DRw52, DRw53and DQwI. In addition, FB I I has been shown to secrete Interleukin I. These results show that human fetal astrocytes can be induced by IFN gamma toexpress HLA-DR in vitro, although these cells were DR negative in vivo. The capacity of FB I and I I cells to function as antigen presenting cells is currently under investigation.

Immunocytochemical Identification of Glioma Cells With a Set of Three Monoclonal Antibodies B.H. Liwnicz, G, Archer, D. Herlyn, H. Koprowski; University of Cincinnati Medical Center and the Wistar Institute Stereotactic small biopsies and CSF evaluation often require diagnosis based on individual tumor cells. Cytologic identification of glioma cells is severely hampered by the tumor's marked pleumorphism. In addition, i t has been shown (Wikstrand et a l . , J. Neuropath. Exp. Neurol. 44:229, 1985) that gliomas express a marked antigenic heterogeneity. We have developed over 900 mouse monoclonal antibodies (MCAs) by immunization with human gliomas and glioma-derived cell lines. Screening against glial and non-glial neoplasms, normal CNS and nonneoplastic extraneuraxial tissue revealed three groups of MCAs: 1. binding to normal and neoplastic glial tissue, 2. binding to a variety of neoplastic tissues, but not to normal CNS and, 3. binding to proliferative tissue regardless of its origin. I t is our impression that the f i r s t group represents MCAs binding to tissue specific glial markers, the second group of MCAs bind to neoplastic transformation markers and the third - to markers of proliferation. Confirmation of these three types of specificity could lead to a development of sets of MCAsfor immunocytochemical identification of viable, proliferating glioma cells. 50

PRELIMINARY CHARACTERIZATION OF ANTI-MEOULLOBLASTOMAMONOCLONAL ANTIBODY REACTIVITY WITH HUMAN NEUR/g_TUHORS. Zeltzer P.M., Schneider S.L., Edwards D.P., McGuire W.L., Univ. Tx. H1th. Sci. Ctr., San Antonio, Tx 78284, USA

~

A

~'

Monoclonal antibodies (MOAB) 283/B3, 275/F2 and 215/D6 were raised against human medulloblastoma cell line TE671. After three subclonings MOABreactivity was tested by ELISA on I0 neuroectodermal and 5 nonneuroectodermal cell lines; and by immunofluoresence on normal human adult and fetal tissue and on neuroectodermal and non-neuroectodermal tumors. These MOAB react with antigenic determinants expressed on normal tissue, tumor cell lines and neuroectodermal derived tumors.

Brain Fetal Brain Adult Medulloblastoma Astrocytuma Glioma Neuroblastoma Wilms' Ewing's

282/B3

275/F2

215/96

0/7 +/4 8-721 4/11 4/21 1/7 2/8 2/3

2/7 0/4 15/21 5/11 8/21 3/7 3/8 2/3

0/7 4/4 15/21 7/11 13/21 5/7 6/8 2/3

B3 (IgG2a) and F2 (IgM) antigens are common to glial and neuronal tumors. F2 recognizes an antigen on 14-16 week fetal brain and has potential for differentiating tumor cells from normal brain. D6 (IgM) recognizes surface and cytoplasmic actin-llke moiety preferentially on transformed cells. These MOABhave limited reactivity on normal tissue and may be used to identify and isolate subsets of normal and malignant neuroectodermal cells. These MOAB may be useful reagents to define developmental stages of neural development and tumor prognosis as has been done for malignant lymphoid disorders. Funded by the Robert J. and Helen C. Kleberg Foundation.

105

53

ANTIGENSDETECTED BY MONOCLONALANTIBODIESTO HUMAN ASTROCYTOMAS by Beresford HR, Rettig WJ, Walker S, Jennings MT, Old LJ. Memorial Sloan-Kettering Cancer Center, New York City, and North Shore University Hospital, Manhasset, NY.

While immunotherapy has been proposed for malignant astrocytomas, evidence that human astrocytomas regularly express unique or restricted antigens is limited. Hybridoma technology provides a basis for further attempts to identify such antigens. Utilizing serological, immunochemical and immunostaining methods, we analyzed antigens detected by 28 monoclonal antibodies (McAbs) produced after immunizing mice with cells from 5 human astrocytomas. Most of the detected antigens are widely distributed among neoplastic and non-neoplastic cells and tissues, but 7 have restricted and distinctive patterns of reactivity. The previously described McAbs AJ8 and AOlO detect lO0 and llO-Kilodalton (KD) glycsproteins, respectively. McAb Gl71 detects a 85-KD glycoprotein preferentially expressed by astrocytomas, neuroblastomas and sarcomas and not found on normal brain tissues. McAbG184 detects a ll5-KD glycoprotein that is found on some astrocytomas, various other neoplastic cells and normal brain. McAbG336 detects an 80-KD glycoprotein found on many astrocytomas, other neoplasms and fibroblasts but not on normal brain. McAb Kll7 detects a 25-KD glycoprotein that closely resembles the putative human thy-l determinant and that is preferentially expressed by astrocytomas, neuroblastomas, sarcomas and fibroblasts. McAb$5 does not immunoprecipit a t e a detectable component, but recognizes a heat-labile antigen that is found on many neoplastic cells, fibroblasts and fetal brain.

Administration of Perfluorochemicals to Enhance Chemotherapeutic Drug Effects. 5. Hiraga, P. Klubes, E.S. Owens and R.6. Blasberg National institutes of Health, Bethesda, liD 20205 It has been suggested that oxygen carrying blood substitutes, perf]uorochemicals (PFC), could increase blood flow and oxygen delivery to ischemic and hypoxic tumor tissue because of their small particle size and low viscosity compared to red blood cells. We examined blood flow in awake WL-Z56 tumorbearing mate Wlstar rats under conditions of l ) Isovolemic PFC-blood exchange which reduced the hematocrlt to less than 4X, 2) PFC-blood partial exchange (hematocrlt 20-25X) an~:l,3) control animals. All animals were breathing i00Z 02; '"Clodoantlpyrine (lAP) and quantitative autoradiograph[c techniques were used. The tissue/blood (or PFC) partition coefficients of lAP were determined in separate groups of animals. Subcutaneous tumor Brain tumor ~earino rats bearlno rats Control Complete Exchange Control Partial Exchange (n=4) (n=5) (n=5) (n=5) A 2.00+_.10 4.22• 1.44• 2.33• B 0.53• 1.14• 0.38• 0.59• C 0.26+_.06 0.25+_.04 O.90• 1.27• D 0.85+_.05 1.40Z.05*=e 0.78+_.05 0.gg• Blood flow (A-C) ml/min/g, A=cortex, B=corpus csliosum, C= tumor; tumor/blood partition coefficient (D) ml/g; /lean, +SEll; * p(.05, ** p<.02, , s , p<.01 These results demonstrate a higher tumor/blood partition coefficient of lAP wRh PFC-blood exchange. A significant blood flow increase-was demonstrated in each brain structure following complete and partial exchange. An Increase In blood flow was also measured in intracerebral tumors whereas no change in flow was observed in subcutaneous tumors. These findings indicate that the effect of PFC-blood exchange on drug delivery to the tumor ls variable and may depend on the particular tumor and the site of tumor growth. The enhanced therapeutic effect that has been reported in some experimental tumor models may result from a higher tissue/blood distribution ratio and altered body clearance for some drugs following PFC administration. 5 5

54

ANTIGENIC PHENOTYPESOF ASTROCYTOMASUBSETSby Rettig WJ, Beresford HR, Jennings MT, Cohen J, Feickert HJ, Oettgen HF, Old LJ. Memorial SloanKettering Cancer Center, New York City, and University College of London, England.

In testing a large panel of mouse monoclonsl antibodies (McAb) against cultured human astrocytomas, we identified 5 McAb that distinguish subsets of astrocytomas on the basis of their antigenic phenotypes. McAbAJ8 detects a lO0-Kilodalton (KD) glycoprotein that is closely related or identical to the CALLAantigen, f i r s t described on human leukemic cells but since demonstrated on melanomas, astrocytomas and normal astrocytes. McAbFl9 detects a 140/90-KD glycoprotein, and McAb G253 detects a 95-KD glycoprotein. Both these antigens are widely distributed among cultured human cells but are not found on GFAP+ astrocytomas. McAbAOlO detects a llO-KD glycoprotein that is restricted in distribution and found among betk GpAB+ and GFAP- &~ocy~omas. McAbA4(C5) detects a nonprecipitating heat-labile antigen that is restricted to CNS neurons and some GFAP+astrocytomas. When these 5 McAbswere repeatedly tested against 23 established human astrocytomas (16 GFAP-, 7 GFAP+), the following five phenotypic patterns were observed: I - AJ8+/FIg+/ G253-/AOlO-/A4-/GFAP- (3 lines); I I - AJ8+/F19+/G253+/ AOIO-/A4-/GFAP- (8 lines); I I I - AJ8+/Flg+/G253+/AOlO+/ A4-/GFAP- (4 lines); IV - AJ8+/Flg+/G253+/AOlO+/A4+/ GFAP+ (4 lines); V - AJ8-/Flg-/G253+/AOlO+/A4+/GFAP+ (4 lines). These findings indicate that human astrocytomas are antigenically heterogeneous and are consistent with other findings of phenotypic heterogeneity in this class of tumor. They also suggest that human astrocytes exist as subsets of cells that are developmentally related but antigenicslly distinct. Implications of these findings for neoplastic transformation of astrocytes will be discussed.

CAPILLARY PERMEABILITY OF EXPERIMENTAL BRAIN TUMORS AFTER HYPEROSMOTIC BLOOD-BRAIN BARRIER DISRUPTION. Peter Warnke*, Dennis Groothuis*, Hide NakagawaT, Ronald G. Blasbergt, Michael A. Mikhaelw Depts. of Neurology* and Radiologyw Northwestern Univ. Med. School, Evanston Hosp., Evanston IL, and Nuclear Medicine Dept.t, NIH, Bethesda, MD. Hyperosmotic BBB disruption (HBBBD) has been used to try and increase water-soluble drug delivery to brain tumors. We have studied the effect of HEBBD in three different tumor models: RG-2, ASV, and ENU. These models represent a spectrum of the changes in permeability that can be seen in experimental brain tumors: ENU tumors consistently have permeability values similar to those of normal brain, RG-2 tumor values that are 15 times higher, and ASV tumors have values that span the range from normal brain to 100 times higher. In ASV-induced canine tumors we made measurements of KI, the blood-to-tissue transfer constant, of meglumine iothalamate (Conray-60) with CT scan methods before and after disruption. In the RG-2 and ENU rat glioma models, quantitative autoradiography was used to measure K 1 of AIB. In tumor-free brain, HBBBD consistently increased K 1 values, with a pattern of regional variation that was consistently higher in grey than white maZter. Although small increases in K 1 were occasionally observed in individual tumors, on the whole HEBBD did not produce a consistent increase in K 1 in any of the tumor models. In the ASV-induced canine tumors, which were studied with multiple-time point sampling, there was no increase in the tissue concentration integral when compared to normal brain. Our data strongly question the efficacy of HBBBD as a method to increase the delivery of water-soluble drugs to brain tumors. In fact, it seems to be counterproductive since it needlessly exposes normal brain to the neurotoxicity of chemotherapeutic drugs. 56

106 THE EFFECT OF CORTICOSTEROIDSON DRUG DELIVERYPARAMETERSIN EXPERIMENTAL BRAIN TUMORS. Peter Warnke*, Dennis Groothuis*, AbrahamKuruvilla*, Michael Mikhaelw Hide Nakagawa?, Ronald Blasbergt. Depts. of Neurology* and Radiologyw Northwestern Univ. Medical School, Evanston Hosp., Evanston IL, and Nuclear Medicine Dept., National Inst. of Health, Bethesda, MD. Corticosteroids, principally dexamethasone, are a mainstay of therapy of malignant brain tumors, primarily because of their ability to reduce tumorassociated edema. However,the mechanism of steroid action, and possible steroid effects on capillary permeability have remained areas of speculation. Here, we describe the results of several experiments to study the effects of dexamethasone on RG-2 and ASV experimental brain tumors. In the RG-2 tumors we have made double-lable quantitative autoradiographic (QAR) measurements of blood flow and blood-to-tissue transfer constant, K1, of AIB, and single-lable QAR measurements of K1 of RISA, and measurements of change in vascular and extracellular space. In ASV-induced canine gliomas we used CT methods to measure K1, k2 (blood-to-tissue transfer constant), and Vp (vascular plasma space), of meglumine iothalamate (Conray-6O) before and afer the administration of dexamethasone. Dexamethasone has complex effects on all of these parameters in the RG-2 and ASV tumor models, to effectively produce a reduction in K1, blood flow, and extracellular space. There was a less prominent effect on vascular space, and k2. Detailed results from each of these studies will be presented, and the implications of these results on the drug delivery of lipid- and water-soluble chemotherapeutic drugs will be discussed.

57

Nerve Growth Factor as a potential reverse transfor5 Q marion agent. V*Koestner, A., **Vinores, S.A. and *Love]l, K., *Department of Pathology, Michigan State University, E. Lansing, MI 48824 and **Department of Ophthalmology, University of Virginia Medical School, Charlottesville, VA 22908. Nerve Growth Factor (NGF) is, in fact, a maturation factor essential for the differentiation of the autonomic and sensory nervous system. Its capability of inducing maturation in immature cells deriving from the neural crest suggests that NGF may also be a potential reverse transformation agent of anaplastic tumor cells of neural crest derivation. NGF was capable of reducing the in vitro growth of anaplastic glioma cells and initia~ng---BiTferentiation. These cells were shown to contain NGF receptors by several methods including the avidin-bioton peroxidase complex method. Rats treated with NGF following implantation of glioma cells derived from this anaplastic glioma clone had a decreased tumor growth rate and tumor volume and a significantly increased survival time. I t was shown in previous studies (Swenberg et al. J. Natl. Cancer Inst. 55:147, 1975) that transplacentally exposed rats to ENU (ethylnitrosourea) had a nearly 100% incidence of neoplastic proliferation of the trigeminal nerve at go days post partum but only 30-35% of the rats developed clinically detectable neurinomas peaking at 7 mos. of age. NGF treatment of pregnant rats before ENU exposure or of their offspring following transplacental ENU exposure resulted in a significantly reduced incidence of trigemihal neurinomas at 90 days post partum as compared to ENU exposed but NGF untreated rats. Until the mode of action of NGF is better understood its importance as a suppressing agent of neoplastic proliferation and as a prospective tumor therapeutic can not be f u l l y assessed.

CLINICALSTUDIES ON INTERFERONTHERAPYFOR 14ALIGNANT BRAIN TUMORS--- SPECIAL REFERRENCE TO THE EFFECTOF GAMMA-INTERFERON Masakatsu Nagai, M.D. and Toshimoto Arai, M.D. Dept.of Neurosurgery, Dokkyo University School of Medicine. Phase I and phase I I clinical t r i a l s of the therapy with ~- and B-type interferon (IFN) on primary malignant brain tumors (mainly, glioblastoma) were conducted since several years (18 cases treated with ~-IFN and 30 cases with 8IFN). The highest effectiveness (33%) was obtained in the cases treated with locally administered human fibroblast IFN. This result suggested that the anti-tumor effect of IFN might depend mainly on its direct action. The study has been proceeding to the therapy using y-type IFN. 6 cases treated with human leukocyte IFN-y (GreenCross Co.,Japan)(i.v. of IO-25• 2, daily for 4 weeks) and 6 cases treated with recombinant IFN-V (Roche)(i.v. of O.25-2.OxlO6U./m~, daily for 4 weeks) were evaluated, and l case of partial remission and 2 cases of minor remission were obtained. Except for pyrexia, the side effects were milder than those of ~- or 8-1FN. Changes of various i~unological parameters during the treatment were examined and the correlation between those parameters and the effectiveness of the therapy was investigated. Natural k i l l e r activity was remarkably elevated in most cases of ~-IFN treated patients after the therapy (p
PLASMA AND CENTRAL NERVOUS SYSTEM PHAP~IACOLOGY OF TIAZOFURIN. D.J. Stewart, R.M. Green, J.A. Maroun, M. Thlbault. The Ontario Cancer Treatment and Research Foundation Ottawa Regional Cancer Centre (General Hospital Division) and the University of Ottawa Faculty of Health Sciences. A pharmacokinetlc study of tiazofurln was carried out in 13 patients treated in a phase I clinical trial of the drug and 7 patients undergoing surgical resection of brain tumor. Tiazofurin was found to be rapidly eliminated from plasma and red blood cell fractions of both groups with kinetics consistent with a two-compartment model of elimination with tl/2e 0.137 + .051 hr, ti/28 5.95 + 3.27 hr. VD 27.42 + 5.91 L/m 2 and CLp 4.42 + 2.38 L/m2/hr f~r plasma. Corresponding values for the RBC were .278 + .229 hr, 8.42 + 4.36 hr, 35.51 + 12.63 L/m2 and 3.73_+--1,99 L/m2/hr. ~ne tl/2a and VD were significantly greater in the RBC fraction (p = .05). Operative conditions did not significantly change the phsrmaeokineties of tlazofurln in the CNS patients. The kinetics were linear over the dose range 500-2700 mg/m 2. The data suggested that there is only a small degree of tissue binding of drug and that the drug is not concentrated by tissues. These conclusions were supported by studies on the penetration of tiazofurin into central nervous system tumors. Tissue samples were obtained from 16 patients who received a non-toxlc dose (500 mg/m 2) of tiazofurln prior to or during surgical resection of tumor. Samples were obtained at periods ranging from 0.1 to 5.7 hr post infusion. Tiazofurin was readily detected in 15 of the 16 tumors, and uptake was rapid with the drug being found as early as 0.I hr after the end of the infusion, The drug was not concentrated by tissues: in 32 of 34 samples from the 16 patients the tissue level was less than the concurrent plasma concentration (46.6% ~ 24.4 for these 32 samples). Tissue concentrations of tlazofurln ranged from 1.3-14.2 ~g/g, 0.6-4.0 ~g/mL, 5.6-i0.I ug/g, and 5.9-22.1 ,g/g in tumor, tumor cyst fluid, brain adjacent to tumor and muscle/fascla, respectively. 60

107 Utilization of purified, radiolabeled monoclonal antibody Fab fragments for in vivo localization in malignant gliomas. Humphrey, P.A., Pegram, C.N, Wikstrand, C.3., Bullard, D.E., and Bigner, D.D., Duke University Medical Center, Durham, N.C. Monocional antibody (Mab) g l C 6 is directed against a unique glioma-associated extracellular matrix glycoprotein. Mab g i C 6 was purified from nude mouse ascltes by protein A affinity chromatography; intact Mab chromatographed as a single peak on size exclusion HPLC and was 97% homogeneous on SDS-PAGE. Proteolytic fragmentation of 1(~.6 mg of g l C 6 with 4% w/w pepsin for 20 hours at 37~ resulted in the generation of Fab and Fc species. Removal of Fc fragments by protein A chromatography resulted in a 27% yield of pure 70 Kd Fab fragments, as assessed by SDS-PAGE and HPLC. Isolated Fab fragments retained immunoreactlvity as demonstrated by comparative indirect ceil surface radioimmunoassay (CS-RIA). 127I labeling of 81C6 Fabs via the chioramine T method followed by passage over G - t 7 yielded a single symmetrical peak by HPLC and Sephadex G-100 chromatography; neither aggreRates nor breakdown products were observed. By direct CS-RIA 12"SI-Fabs bound to t a r g e t O - 2 7 ! MG ceil culture monolayers, but at a reduced level compared to intact Mab. Purified, 125I-Fab fragments were utilized in a dual label human giioma localization experiment in nude mice. Localization of intravenously administered t27I-gic6 Fab was rapidly detectable (within 6 hr), progressing to a peak localization index (Lt) of 7.5 (x of three animals) by day ~, and maintaining an LI of 6.g at day 7. In contrast, intact Mab gIC6 did not localize significantly until day I (LI = 2.6), attaining.Eeak localization at days 7-9 (LI = 15.0-16.77). The half-life of l z ) I - g l C 6 Fab in the blood was 1.09 days in contrast to 3.3 days for intact 1251-gIC6, indicating rapid systemic clearance of i'25I-gIC6 Fab. Tumor:brain ratios of localized Fab peaked at 6 hrs (ratio = 101) and rapidly declined, as opposed to the slower accumulation of intact 12)l-glC6 (peak of 39 at 7 days). The observed rapid localization of 1251-81C6 Fab, high tumor to normal brain ratios and rapid systemic clearance suggest giC6-Fab will be a good imaging agent; intact Mab glC6 with greater percent localization of total injected dose and longer maintenance of localization may be superior for therapy. 61

63

SIXTH ANNUALSTANLEYN. GOREMEMORIAL LECTURE G1ioma Cytogeny and Differentiation Viewed Through the Window of Neoplastic Vulnerability

L.J. Rubinstein, M.D. Division of Neuropathology, Department of Pathology, University of Virginia School of Medicine, Charlottesville, Virginia This lecture examines the concept of the window of neoplastic vulnerability as applicable to the clinical and experimental incidence of nervous system tumors and of their different cell populations. The window results from the interaction of several factors: the existence of a reserve population of stem cells, the capability of differentiated cells to re-enter the kinetic cycle, the number of replicating cells at risk at a particular time, the length of time during which a particular cell population remains in the cycle, the state of differentiation and the further differentiation potential of that population, and the speed at which various steps of differentiation are achieved in successive cell generations. This concept explains many aspects of neuroepithelial, especially CNS, tumor incidence, and the relationship of central neuroepithelial embryonal tumors to tumors of adult cell type. The incidence of different neuroepithel i a l tumors can be correlated with the width of the window. The application of immunohistochemicaI markers has been of much help in exploring this concept. Examplesof tumor entities illustrating its applicability will be reviewed.

Characterization of Localizing Glioma-reactive Monoclonal Antibodies (Mabs). Wikstrand, C.1., Humphrey, P.A., Bullard, D.E. and Bigner, D.D., Duke University Medical Center, Durham, N.C. Multiple fusions following immunization of athymic mice with the human glioma ceil line D-5~ MG resulted in the selection of 7 Mabs highly reactive with neuroectodermal tumors and unreactive with normal nervous system tissue. Three Mabs, C12, DI2, and E9, performed well in athymic mouse-human glioma xenograft localization assays. Mabs C12 and DI2 are IgG 3 antibodies; E9 is an IBGI; all bind readily to Staphylococcal protein A in column purification and radioimmunoprecipitation (RIP) procedures. All iodinate via the chloramine-T method yielding 1251-immunoreactive product by direct cell-surface radiolmmunoassay (CS-RIA) and absorption assay. By indirect CS-RIA, a cultured cell line panel consisting of gliomas (17), medulloblastomas (3), neuroblastornas (2), melanomas (2), and fetal (2) and adult brain (2) was examined; the 3 Mabs were distinct in their reactivity profiles. Each was + with > ~0% of the gliomas tested (C12, 9/17, DI2, 7/17; Eg, 1~/17); and with 1/2 medulloblastomas, 1/2 melanomas, and cell lines derived from 12 and 16 week gestation human fetal brain. No reactivity was observed with neuroblastoma or adult brain-derived cell lines. Notable extraneuroectodermal reactivity included DI2 with splenic trahecula and spermatids, and E9 with fetal human gut. By CS-RIA and immunohistology, Mab E9 bound to an antigen fag) on extraceilular matrix material (ECM); Mabs C12 and Di2 stained dispersed cells in 6/7 and 7/7 glioma tissue samples, respectively. Repeated attempts to RIP Ag from surfaoe-iodihated cell membranes were inconsistent; in the case of Mab E9, this is not unexpected due to predominant antigen localization in ECM. Current antigen purification methods involve endogenous label incorporation and polyethylene glycol precipitation. When used in paired label localization experiments in subcutaneous D-50 MG tumor-hearing athymic mice, Mabs Cl2, Di2 and E9 demonstrated similar localization patterns, reaching peak localization at day 3 (DI2, E9) or ~ (C12); maximum percent of injected Mab bound to tumor ranged from 2.0% (E9) to g% (C12). Peak tumor:braln localization ratios (167- 185) were attained by all 3 Mabs at day 1-2 followed by rapid clearance. Mabs C12, D12, and E9 are suitable for further investigation including possible clinical trials for imaging and therapy.

62

SIR~4ARY OF CLINICAL TRIALS OF CHEMOTHERAPY AND CONVENTIONAL RADIATION THERAPY IN THE TREATMENT OF PATIENTS WITH PRIMARY MALIGNANT BRAIN TUMORS. M. S. Mahaley, Jr., M.D., Ph.D., Division of Neurological Surgery, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina.

64

The author will attempt to summarize the results of the clinical trials of various cooperative groups and others with reference to chemotherapy delivered through various routes and conventional radiotherapy in the treatment of patients with primary malignant brain tumors. This summary will be organized in accord with specific treatment modallties rather than chronologically or by study group. Particular mentlonwill be made of intraarterial chemotherapy and associated toxicltles.

108 BRAIN TUMORRESEARCHCENTERAND NORTHERN CALIFORNIAONCOLOGYGROUP: NEWTREATMENT REGIMENS FOR MALIGNANTGLIOMASAND OTHER PRIMARY BRAIN TUMORS. V.A. Levin, P. Gutin, C.B. Wilson. Neuro-Oncology Service, Dept. of Neurological Surgery, University of California, San Francisco, CA 94143.

6~

Four new programs will be discussed. The f i r s t two protocols, NCOG6G82-I and 6G82-2 were initiated in 1982 for untreated glioblastoma multiforme (GM) and other anaplastic gliomas (AG). Both are modifications of BTRC6G61, where hydroxyurea (HU) was given during irradiation (RT) followed by PCV chemotherapy. In NCOG6G82-1, HU was replaced by bromodeoxyuridine (BrdU), a potentiator of RT cell k i l l . In 6G82-2, brachytherapy with 1251 is given after RT-HU and before PCV. As of June 1, 86 pts were entered on 6G82-1 and 56 pts to BG82-2. The third program, designed for both newly diagnosed and recurrent tumors (BTRC 8422 and 8522), represents the f i r s t overt attempt to reduce cellular resistance to a chloroethylnitrosourea (CENU). The program uses 6-thioguanine, procarbazine, dibromodulcitol, CCNU, and fluorouracil (FU) and HU. 39 pts without prior chemotherapy and 22 with prior CENUchemotherapy were entered by June 1. The fourth program, BTRC8424, was designed to offer a chemotherapy alternative to pts with mildly anaplastic astrocytomas and non-enhancing moderately anaplastic astrocytomas. Hoshino has proposed that these tumors may have few GO cells capable of re-entry from the non-cycling into the cycling pool. The treatment consists of cycles of the synergistic combination FU-HU followed by 6-TG and HU. 13 pts were accrued as of June 1.

J~l~It~,LSfA'IIJSOF ~l~] CIIILL~EI~SCa~C~ SnJDY GRU(/P (CL~) ~ N ru~uR 'IILLALS. Finlay JL, l~_imrtJnentof Pediatrics, University of Wisconsin, Hmdison; for tl}e C1~ildrens C~alcor Study Group. ~rain tt~ors rltFtkseoo~I only to Icukemias as the ieajor class of cancers in cilildhocxl.~x~Jt IU years aI~, t]~ C L ~ e ~ r k e d upon three r,%ndclnized trials for children With newly-diagnosed (a) high-grade astroc3rtomas , (b) brain stFan tt~rors (BS'r), and (c) medulloblastc~nas. Patients received elthel ladidLiun ti~;tapy 0~(i') aloltu ,Jr >~i' dloI~ ~rith aJK] follo~d by chranotl)et~pywith vincristine, CCMJ and p~dJ1isone. XKr involved local therapy for the astrocylomas and Ks'r, or cranio-spinal therapy with a posterior fossa boost for medulloblastnmas. For the astI~)c3rtc~nstudy, with all living patients now followed for over 4 years, the 5-year relapse-free survival (RFS) is 44.5% for patients who received X}{r + cl~notherapy, versus 18.2% for patients who received X~T alone. Only the s~nall numbers of biopsy-only patients failed to sl)aw any benefit from dlemotherapy. Ketl~OhOJlgrades III or IV patients henefited equally frr~n chemotherapy. Patients with necrosis or mitosis benefited from chemotherapy, but these were still poor prognostic features for the 6 7

ch(~1~)thorapy ffroup. For the brain stemtumorstudy, there~ no difference in the 5-yearRFS between those treated wath xKr aloi~ and these also receiving chemotherapy. For the medulloblast~a stuay, a mkr-popalati(~l of patients was Uelineated r~iir~,trn~q~cctiveanalysis ~ i c h benefited from the adjuvant cherootherapy; p a t i e n t s with s i z e a b l e tumors (~r~ang Stages T3-T4) o r with m e t a s t a s e s '3"~-~'B) ~I~-M = ~i~ptifit~utLlyi,~uvL~ 5-ycat R}~ wi~lt Et~tLC~ Wi~h

chemotherapyc~paredwith X1ffalone. [58.4%versus41.6%for T3-T4 cases). '[his groupof patients l ~ e f i t ~ from:hc~notherapyconstitutes t~co-thirds of rnedulloblast~rnapatients. Based upon a study designed at the Childrees Ot'thepedic Hospital, Seattle, the CCSG opened a F,~ase II Study of recurrent brain rtmDrs in 1983, using an "8-drugs-in-l-day" co~)ination of CC~, vincristine, merJlylprednisolone, hyde.>e/urea, procarbazine, cis-platin, cytosine arabinoside, and eitller cyelophosphamide or uric. Preliminary data from these studies, as wall as data on" this regimen in select patients with newly-diagnosed tumors pre.~{r, has led to the development by CCSG of studies for patients with newly|iagnosed brain tumors, in which the 3-drug containing regimen of the prim itudies will be compared with the "8-in-l" regimen (delivered both preand post-Xl~). The high-grade astrocytc~na study laas been opened since April 1985. '[he study for the "standard-risk" medulloblastomas (those benefiting from chemotherapy) will hepefully be opened within the next 6 months. Other studies,of hyperfractionated x){r for BST, and an XRT only study for 'Tx)od-risk" medulloblastomas, are in preparation.

66

COOPERATIVE STUDIES OF CHEMOTHERAPYFOR GLIOMAS AND MEDULLOBLASTOMAS Schold SC Jr, Mahaley MS Jr, Friedman HS, Vick NA, Bullard DE, Burger PC, Cairncross JG, Falletta JM, Halperin EC, Khandekar dD, Macdonald DR, Bigner DD. DukeUniversity, University of North Carolina, Evanston Hospital, University of Western Ontario. We are conducting phase I t and phase I l l t r i a l s of chemotherapy in patients with primary anaplastic central nervous system neoplasms. In all studies, pathology is reviewed and classified by a single neuropathologist (PCB), CT interpretations are made by one neuroradiologist at each institution and confirmed independently, and radiation parameters are reviewed by a single radiotherapist (EH). We define response and failure using both clinical and CT c r i teria. We have completed 2 phase I t studies of AZQ-based chemotherapy combinations in patients with primary anaplastic brain tumors and I phase It study of intermediate dose cyclophosphamide (CPA) and vincristine (VCR) in patients with recurrent medulloblastoma (MB). Both AZQ + BCNUand AZQ + procarbazine produced response rates of approximately 30% with occasional long-term responders in patients with recurrent gliomas. CPA + VCR produced responses in 8 of 12 patients with recurrent MB. A phase I l l study comparing AZQ with BCNUis nearing completion and has thus far shown no difference between the two arms. Current phase I I studies include intravenous melphalan alone in patients who have not previously received chemotherapy and a study comparing AZQ with fludarabine as single agents. The choice of agents and routes to be tested in future studies will be based on results in an athymic mouse system using human gliomas and MB cell lines and on studies of intracarotid chemotherapy in athymic rats using these lines.

CHILDHOOD BRAIN TUMORINVESTIGATIONS: PEDIATRIC ONCOLOGYGROUP, L.E. Kun, O.G. Norris, J. Boyett for the POG Brain Tumor Committee The Pediatric Oncology Group has initiated a number of clinical investigations in childhood brain tumors. Primary treatment protocols include a Phase I l l t r i a l of irradiation + MOPPchemotherapy for meduIloblastoma, a phase I / I I tTial of hyperfractionated irradiation for brain stem gliomas, and a prospective natural history/ treatment response study of intracranial ependymomas. Phase I t studies for recurrent tumors incorporate standardized objective response c r i t e r i a . A t r i a l of cisplatinum has been concluded; studies of AZQ and DBD are near completion. Current effects focus on a randomized Phase I I comparison of carboplatin vs. CHIP.

68

Projects nearing groupwide initiation include a randomized t r i a l of reduced-dose neuraxis irradiation for favorable ("standard risk") meduIloblastoma, a pilot study of preirradiation chemotherapy (cis-platinum, vincristine, high dose cyclophosphamide, moderate dose methotrexate) for "high risk" medu]loblastoma, and a major effort in the management of infants and young children with malignant CNS tumors. The latter study will assess the f e a s i b i l i t y of using 12 or 24 months of chemotherapy (high dose cyclophosphamide, vincristine, cis-platinum, procarbazine) postoperatively, delaying irradiation to allow further brain maturation. Standard doses of irradiation will be utilized after chemotherapy; a second surgical resection will be advised for children with residual or recurrent disease at initiation or irradiation. A surgical protocol to prospectively study the impact of extent of resection and histology for cerebellar astrocytomas is being planned. All POG CNS protocols have included systematic neuropathology review. Primary treatment protocols incorporate prospective, serial studies of neuropsychologic function. Future studies will build upon phase I I data incorporating newer platinum derivitives, hyperfractionated irradiation, and specific irradiationchemotherapy sequences.

109 IN VITRO AND IN VIVO CHEMOTHERAPY - AN UPDATE

69 John L.

Darling and David G.T. Thomas

Dept. of Neurological Surgery, Institute of Neurology, Queen Square, London, WCIN 3BG, U.K. An association between relapse free interval (RFI) and in vitro chemosensitivity to two alkyfating agents, CCNUand PCB has been demonstrated in patients with malignant glioma. However, response to an unrelated agent, VCR, did not seem to influence RFI. Cultures were designated responders or non responders in vitro by comparison with a "training set" of cultures derived from tumours of similar histology. Those cultures whose IDso's were below the median of the group were designated responders and those cultures who-IDso's were above the median of the group were designated as nonresponders. The cut-off point used remains arbitary, although further analysis using the 75th or 25th quartile has been carried out to determine a rational endpoint for clinical studies. The increase in RFI observed in chemosensitive patients was not due to imbalances in prognostic factors, although grade of tumour and age did influence RFI and chemosensitivity. Further analysis of prognostic signs in relation to in vitro chemosens i t i v i t y are in progress. In particular, the relationship between age and sensitivity to nitrosoureas and other alkylating agents such as AZQ is being investigated in a panel of 200 glioma cultures. I t has now been possible to expand the correlation between RFI and in vitro chemosensitivity to a total of 60 patients. This extra data will allow the subdivision of grade I l l and IV astrocytomas to be made during statistical analysis in order to elucidate the exact relationship between histological grade and in vitro chemosensitivity.

1

Photodynamie Cell Killing by Bromodeoxyuridine and Ultraviolet Light: A Potential New Therapy for Brain Tumors Corey Raffel, M. D., Ph.D., Dennis F. Deen, Ph.D., and Michael S. B. Edwards, M. D. Incorporation of the thymidine analog, bromodeoxyuridine (BrdU), into the DNA of dividing ceils sensitizes the cells to ultraviolet light (UV). The potential of BrdU and UV as an adjuvant therapy in the treatment of malignant brain tumors has been investigated using the 9L rat gliosareoma model i..nn vit.r.o. The effects of this combined treatment have been assessed by two methods= 1) colony forming efficiency, an assay of cytotoxieity, 2) alkaline elution of DNA from treated cells, an assay of DNA damage. We have found that sequential treatment of 9L monolayers with 10 pm BrdU for twenty-four hours followed by 30 seconds UV (259 nm, 18j/m z) leads to greater than three logs (>99.9%) cell kill when assayed by colony forming efficiency. Little cytotoxieity is seen with BrdU or ultraviolet light alone. Alkaline elution analyses indicate that DNA from sequentially treated cells contains two components, a quickly eluting component containing multiple single strand breaks, and a slower component containing no new single strand breaks. Increasing the time the ceils are exposed to BrdU increases the percent of DNA in the first component. We have demonstrated that the portion of DNA containing the large number of single strand breaks is the DNA synthesized after the addition of BrdU. We have also shown that significant numbers of DNA-protein crosslinks are formed by sequential treatment with BrdU and UV. Our data indicate that the toxieity exhibited by BrdU and UV may be caused by direct DNA effects, both single strand breaks and DNA-protein erosslinks. The potential for treatment of patients is being investigated in an in viva animal model.

E~ERI_MENTAL CHEMOTHERAPY OF HUMAN GLIOMA AND MEDULLOBLASTOMA HS Friedman, SC Schold, Jr=, SH Bigner, aM Colvln, SM Ludeman, GB Ellen, VL Boyd, PC Burger, CJ Wikstrand, JQ Trojanowskl, EC Halperin, PF Jacobsen, PA Papadlmitrlou, DD Bigner, Durham, N.Co, Baltimore, MD., Washington, DEC., Perth Australia. We have established a panel of human glioma and medulloblastoma cell lines to study their biology and therapeutic sensitivities. The human glloma lines included D-54 MG, U-If8 MG, U-251MG, N-456, N-519 and N-735. The human medulloblastoma cell lines studied were rE-671, D283 Med and Daoy. In vitro chemosensitivity has been studied by a clonogenic assay in soft sgar and growth inhibition in plastic dishes. In vlvo r has been studied by treating athy~le mice hearing subcutaneous or intracranial xenografts with chemotherapeutic agents administered by i.p. injection at the 10% lethal dose. The most active agents against the gllomm lines were PCNU, melphalan, cyclophosphamlde, proearbazine, fludarablne and AZQ. The most active agents against the single medulloblastoma line presently tested, TE-671, were melphalan, eyclophosphamlde, Asta Z 7557, iphosphamide and phenylketocyelophosphamide and phenylketolphosphamlde, novel ketonic derivatives of cyclophosphamlde and iphosphamide, respectively. Two additional medullohlastoma lines have now been established and their therapeutic sensitivities will be similarly eharaeterlzed, D283 Med, derived from the peritoneal implant and ascitic fluid of a child with metastatic medulloblastoma, grows in vitro in suspension culture with spontaneous macroscopic spheroid formation and is tumorigenic in nude athymie mice. Daoy, derived from a cerebellar medulleblastoma, grows in vitro in plastic dishes or soft agar, and is similarly tumorigenic in vlveo Combination chemotherapy with vincristlne and cyelophosphamide has successfully been utilized in our phase II protocol for patients with recurrent medullohlastoma, and melphalan, the single most active compound in the laboratory, is now in clinical trial. These models will allow further analysis of the neurobiology and therapeutic sensitivity of glioma and medulloblastoma, with the rational design of therapy, 70

72

PHOTOCHEMOTHERAPY OF ~LIGNANT GLIOMAS USING MITOCHONDRIAL-SPECIFIC DYES

Stephen K. Powers, M.D. - Univ. of North Carolina This report reviews our experience with the use of Rhodamine-123 (Rh-123), a xanthine d e r i v a t i v e , to photosensitize malignant gliomas. Rh-123 is a mitochondrialspecific staining dye which is selectively retained by hath carcinoma and glioma cells and is expelled from normal cells. Selective staining of both the avian sarcoma virus induced glioma in rat and multiple other carcinoma cell lines in small animals has been demonstrated in our laboratory. Due to the specific staining of glioma cells offered by the selective retention of Rh-123 and its intracallular location, this drug has great potential as a photosensitizer. U-251 MG cells that have been incubated in Rh-123 and then exposed to blue-green light from the argon laser at subthermal energy dosages exhibit logarithmic cell killing over time. Rh-123 is easily isolated from tissues to which it is bound with and can be quantified by HPLC. Extent of tissue fluorescence of Rh-123 closely correlates with the concentration of drug as determined by HPLC. The principle difficulty with using Rh-123 as a photosensitizer resides in the-fact that it is only activated by blue-green light and has an absorption maximum of 510 nm. Thus, in vascular tissues and in heavily pigmented tissues where the penetrance of blue-green light is minimal, the phetochemotherapeutie effect using this agent would be localized near the light source for short light exposure times. We are currently investigating other classes of compounds whose absorption maxima lie in the near infrared wavelength range. These agents tend to be mitochendrialspecific like Rh-123 because they are also cationic and lipophilic. Further investigation involving the toxicology, tumor and normal tissue drug distribution, and phototherapeutic efficiency are required before these drugs can be exploited for use in photochemotherapy.

110 73

Drug Streaming During Intraarterial Chemotherapy J- Bob Blacklock, M.D., Donald C. Wright, M.D. Robert L. Dedrick, Ph.D., Ronald G. Blasberg, M.D., Robert L. Lutz, Ph.D., John L. Doppman, M.D., Bdward H. Oldfield, M.D. National Institutes of Health, Bethesda, F~)

To investigate nonunifrom drug delivery as the cause of focal toxicity follo~ing intraarterial (]CA) ch~Totherapy we examined the distribution of drug delivery after internal carotid artery (ICA) infusion of (14C)-iodoantipyrine in 8 rhesus monkeys. Infusions were delivered at slow infusion rates (1-2% of ICA flow, comparable to those being used clinically) or at fast infusion rates (20% of ICA flow) combined with additional tecahniques to promote mixing with ICA blood. Two monkeys received intravenous (IV) (14C)-antipyrine, to assess normal cerebral blood flow and uniformity of drug delivery to the brain. Uniformity of delivery was assessed by comparing high-to-low ratios of isotope concentration in four brain regions by quantitative autoradiography. There was striking nonuniformity of drug delivery in the slow IA group with as ~ c h as 13-fold differences in drug concentration in anatcmically contiguous areas. The values of high-to-l~w concentration ratios (meaP + 1 SD) in individual autoradiographic planes were: (I) frontoparietal cortex: slow IA 4.54 + 2.07, fast IA 1.71-+0.31. IV 1.30-+0.174; (2) frontoparietal white matter: slow IA slow IA 2.94-+1.45, fast IA 1.59-+0.41, IV 1.34_+0.21; (3) temporal cortex: slow IA 5.43+3.57, fast IA 1.69+0.24, IV 1.67-+0.25; (4) basal ganglia: slow IA 3.6-+2.9, fast IA 1.18-+0.10, IV 1.09-+0.04. Differences between slow IA and fast IA are significant (p<0.01) ; those between fast IA and IV are not sionificant. Druo streamino durinq intraarterial delivery persists through several levels of bifurcation and results in variable drug concentrations in the hemisphere. Areas of high concentration may result in focal toxicity and areas of low concentration may result in therapeutic failure.

75

Supraophthalmic Carotid Infusion of Cisplatin and BCNU for Malignant Gliomas. JP Kapp, State University of New York at Buffalo, and RB Vance, University of Mississippi Medical Center.

The combination of Cis-diammedichloroplatinum I I (cisplatin) and 1,3 bis(2-chloethyl)-l-nitrosourea (BCNU) in an arterial infusion regime is attractive because these agents d i f f e r in time of onset of myelosuppression, in target organ t o x i c i t y , and in penetration of the blood brain barrier. Forty-two patients with malignant gliomas have been treated in a protocol consisting of two internal carotid infusions of Cisplatin and BCNU, followed by maintenance 1-(2-chIoroethyl)-3-cyclohexyl-l-nitrosourea CCNU (100 mg/m~) bimonthly. Twenty-two patients with symptomatic progressive disease after radiotherapy received high dose cisplatin (120-200 mg total dose) and BCNU(300 mg total dose). Nine patients received low dose cisplatin (110 mg total dose) with 300 mg BCNUafter radiotherapy. Eleven patients received cisplatin 120-150 mg and BCNU 240-300 mg total dose with no radiotherapy until disease progression. The radiographic response rate for high dose infusion for progressive disease after radiotherapy was 84%; median survival was 10+ months with the longest survival being 25+ months and 41~ of this group surviving. In the patients receiving low dose cisplatin and BCNU, the radiographic response rate was 25% and median survival was 11+ months with 45% of the patients surviving. In the group infused before radiotherapy, 75% of evaluable patients demonstrated radiographic response; no responders in this group have progressed; all patients in this group are living. Performancestatus increased in 71% of responders in all groups. Infusing drugs distal to the ophthalmic artery has prevented ocular complications. Five patients (12%) have developed hemiparesis or hemiplegia. Reducing drug dose has decreased this incidence although no patient receiving less than 69 mg/m~ cisplatin has responded. Since many of the patients are s t i l l living, the median survivals are increasing. Presentation will include updated survival statistics and representative CT scans.

PHAI~%CO-KINETIC STODIES OF THIO-TEPA IN IXX3S FOLLOWING DELrVI~{Y BY ~ . nmay ~,1 ~ipple j 1 ll~ski p,2 levin A,3 . ~k~ston L,2 Roze~tal j,4 and Egorin M. ~ D e ~ t s of Pediatrics,t RadiologY, Neuresurgery and Nel~'olog~ UniversitY5of Wiscensin, F~dison, WI, and The Division of Developmental Therapeutics, University of b~_ryland C~cer Center, Baltimore, ~fi).

4

v~

We have measured the concentrations of 'll~io-tepa (TP) in plasmm, CSF aml normal brain parenchyma of dogs following a~,inistration by (a) intravenous (IV) bolus, 0.Smg/Kg, (b) intra-arterial (IA) bolus, 0.Smg/Kg, (c) IA bolus with mannitol-induced blood brain barrier disruption (BBBD)/" and (d) continuous intra-venous infusion over 1 hour (IV-IFN). TP doses administered by IV-I~ ranged from 1.3 to 2.<~g/Kg. Foll~ing drug delivery, ~le animals l~=re sacl-ificed at times ranging fr~, 15 to 180 minutes. CSF and plasma samples were drawn simultaneeusly at 15 to 30 minute intervals both during and following IV-IFN. TP wes wi~as~ed by gas liquid chromatography (Greel*m~). A sif~tificant finding following IV-IFN is the cousistently higher TP concentration in the brain parenchy~ than in e~multane~ly sampled CSF and plasma, both at the end of the 13' infusion ('peak') and 180 minutes after the end of the TP infusion ('wasl~-out'). The brain:platoon ratios were 2.24:1 and 2.76:1 at 'peak', and 9.72:1 ar~ 3.28:1 at 'wnsh-out'. ~ e brsin:CSF Icatios were 2.~Q:I and 3.35:1 at 'peak', and 9.52:1 end 3.5:1 at '~nsl~-out'. lhe ~0aet of the route of %? ad~st-cation upon brain parenchymal levels of TP can he seen in the following table:

p~IGm.

IV BOU]5(n=5) IA BOLUS(n=5) IA+BBBO(n=5) IV-IFN(n=4)

O- 30 rains. Post-injeetn.

0.086 0.018

0.252 O.138

2.721

30-180 rains. }bst-injeetn.

O O 0

0.077 0.057 0.05

0.68 0.269 0.201

1.721

3.88 4.28 3.18 1.51

]he IV-IFN produces a 'peak' TP concentration of ~]~e same m~itude as D%+BBBI). Adjusting for c ~ dose differences, ~ 'peak' concentrations after IV-IFN are 7 to 16 times hi~ler ~lan after IV boh~. ~le 'wBsh-out' concentrations of TP are racrL-esustained following IAr-IFN ~%n follo~ng the other delivery teclmiques. ~l~eee studies have implications for the optimal mode of deli%~ry of TP to bz~in tumors in the clinical setting.

: O u ~ l ~ xP).

76

INTERSTITIAL RADIOTHERAPY OF GLIOMAS Christoph B. Ostertag, M.D.

The goal or i n t e r s t i t i a l radiotherapy (brachyt h e r a p y ) i s to e f f e c t a local tumor necrosis while c a r e f u l l y avoiding r a d i a t i o n damage to the normal surrounding b r a i n , which is of p a r t i c u l a r importance in long-term s u r v i v e r s . However, recent c l i n i c o - p a t h o l o g i c a l data have shown, t h a t morphological changes in the normal brain tissue may extent f a r beyond the radionecrosis and and adversely a f f e c t the c l i n i c a l course. To study the b i o l o g i c a l e f f e c t s , local radionecroses were e x p e r i m e n t a l l y induced by implanting r a d i o a c t i v e sources (iodine-125 and iridium-192 permanent low a c t i v i t y and temporary high a c t i v i t y s o u r c e s ) i n t o healthy dog brains and avian-sarcoma-virus induced dog brain tumors. The experimental data demonstrated the importance of the -though r e s t r i c t e d - d i s r u p t i o n of normal b l o o d - b r a i n - b a r r i e r f u n c t i o n which generates a chronic vasogenic edema. Vasogenic edema is the most important single f a c t o r f o r demyelination and r e a c t i v e g ] i o s i s beyond the i m p l a n t a t i o n s i t e . These experimental f i n d i n g s provide a basis f o r the c l i n i c a l dosimetry. For the time being, i n t e r s t i t i a l radiotherapy with iodine-125 permanent implants is recommended f o r slowly p r o l i f e r a t i n g , n o n - r e s e c t a b l e , d i f f e r e n t i a t e d tumors such as low grade gliomas in f u n c t i o n a l l y c r i t i c a l areas. Combined i n t e r s t i t i a l and e x t e r n a l beam r a d i o t h e r a p y , which allows the a p p l i c a t i o n of more moderate doses compared with external beam radiotherapy alone, has proved useful f o r the treatment of anap l a s t i c gliomas.

111 Stereotaxie Interstitial Braehytherapy of Malignal~t Brain Tumorswith Removable High Activity Iodine-125 (1251) Philip H. Outin, M. D. and Steven A. Leibel, M. D. Brain Tumor Research Center of the Department of Neurological Surgery and the Department of Radiation Oneology, University of California School of Medieine, Sen Francisco, California Sinee 1977, over 170 patients with malignant brain tumors have undergone CT-direeted sterotaxie implantation of radioactive isotopes at UCSF. Of these, 147 patients have reeeived removable high aetivity 125I implants and 41 patients of this sub-group were t r e a t e d as of September 1, 1984 for reeurrenees of anaplastie astroeytomas (23 patients), or glioblastomas (18 patients). All had received previous teletherapy and most had also received ehemotherapy. Patients received a minimum tumor dose of 8,000 to 12,000 red a t a dose rate of 30-40 rad/hr from sources implanted in one or more c a t h e t e r s using primarily the Brown-Roberts-Wells stereotaxie apparatus. The e a t h e t e r s were removed when the desired tumor dose was achieved. Response was documented in 23 patients for 3-25 § months. Stabilization of disease was achieved in 6 patients for 4-17 + months, and continued progression ensued in 14 eases, most often from focal radiation necrosis. Forty-four percent of the patients remain alive with a median survival of 22.5 months a f t e r reeurrenee. Approximately 4096 have required re-operation for tumor regrowth or focal radiation necrosis. In a matched-pair analysis, patients undergoing implantation for malignant glioma recurrence had a significantly longer median survival than those receiving chemotherapy. Based on our findings, a phase II study was initiated integrating interstitial implantation into the primary t r e a t m e n t of patients with malignant gliomas. F o r t y eight patients have been t r e a t e d in this protoeol, reeeiving 6000 red to the tumor volume by external technique, followed by an implant boost to deliver an additional minimum tumor dose of 6900 red. This study is not ready for analysis.

7 7

79

RADIATION EFFECTS AND RADIATION NECROSIS IN PATIENTS WITH INTRACRANIAL GLIOMAS. P.C. Burger, M.D., Dept. of Pathology, Duke University Medical Center, Durham, NC

The effects of therapeutic irradiation on gliomas and the surrounding brain are often considered as either (1) a desired direct tumoricidal or tumoristatic effect on the neoplastic ceils, or (2) an undesired e f f e c t on the surrounding brain presumably mediated by an irradiation-induced vasculopathy. The morphology of the tumoricidai effects has not been fully studied in human material, but, as a depopulation and morphological alteration of neoplastic cells, presumably occur within days or weeks a f t e r therapy. The second form of radiation e f f e c t is best exemplified by the uncommon, but potentially severe, late delayed radiation necrosis t h a t occurs months or years a f t e r the irradiation. This lesion has a number of distinctive features including a predominant white m a t t e r localization, an evolution by the coalescence of multiple smaller areas of necrosis, and the association of vascular changes including fibrinoid necrosis, telangiectases, hyalinization, and endothelial cell atypia. Postmortem studies of irradiated gliomas, however, have suggested that the above distinctions between desired direct and undesired indirect effects of irradiation are not always clear since histologic findings characteristic of delayed radionecrosis are common findings within the tumor bed of irradiated gliomas. This suggests that necrosis mediated by radiation's effect on vascular walls may account for some of its therapeutic effect, and that biopsy of the irradiated tumor bed may not be a reliable way to distinguish recurrent tumor from radionecrosis.

78

CLINICALEXPERIENCEWITH INTERSTITIAL IRRADIATION Robert G. S e l k e r , M.D. Our experience with i n t e r s t i t i a l irradiation involves both the permanent and temporary implantation of 1251 in patients harboring glioblastomas, ependymomas, sarcomas, and metastatic tumors. Using a semistereotactic (permanent volume implant) and stereotactic (temporary implant) procedure, all patients undergo CT scanning and dosimetry planning as generated by the GE treatment planner. Temporary . . implantation involves the use of up to 7, 20 mCi iz~I seeds, followed as a rule by the removal of the treated mass at the completion of the 5-8 day radiation period. Sevento eight thousand rads are delivered to the tumor periphery as determined by CT, 5-6,000 rads to a 1.5 cm cuff of normal tissue. Permanent implantation, d~signed for large volume recurrent tumors ( 50 cm~) utilizes large numbers of 0.5 mCi seeds (mean 50 seeds), restrained in Lexan tubes and implanted via a lucent template containing I cm spaced holes. I t is designed to homogenously deliver 13 rads/hour for approximately 10 cell cycles (85 hour median) of the average glioma, 20,000 rads over approximately 12 months. Whereas local disease seems controllable with either technique, the biology of the glioblastoma is such as to make recurrence, either at the perimeter or elsewhere in the brain, the cause of failure of the technique. Pilot studies designed to i n i t i a t e i n t e r s t i t i a l irradiation "up front," rather than at recurrence, are underway in the hope of expanding the disease control interval.

I251 Radiation Damage of Normal Brain: Effects of Dose, 8 O D o s e Rate and Irradiated Volume. K Turowski, J'R Fike, CE Cann, RJ Higgins, RL Davis, PH Gutin, TL Phillips and KA Weaver. Brain Tumor Research Center, University of California, San Francisco, CA 94zr Interstitial brachytherapy using IZ51 is currently being used in the t r e a t m e n t of malignant brain tumors; however few quantitative d a t a are available concerning the response of normal brain to this type of irradiation. We have used the canine brain as a model for studying radiation damage. High activity Iz51 sources were surgically placed into the frontal white m a t t e r of normal beagle dogs and later removed. Reference doses, calculated at a point o.75 mm from the source ranged from Io-4o Gy. The extent of necrosis, tissue contrast enhancement and edema were quantified using computed tomography; absolute minimum doses and dose r a t e s required to induce necrosis and contrast enhancement were derived from volumetric analyses. The extent of tissue necrosis was related to irradiated volume and a minimum effective dose averaging iBo-zoo Gy was required to induce this type of damage. Vascular related damage including contrast enhancement and endothelial cell proliferation was less dependent on irradiated volume or total dose. The extent of radiation-induced edema was directly related to the volumes of necrosis plus contrast enhancement. The studies performed here showed t h a t noninvasive serial studies in a well characterized in vivo model can address specific clinically-related questions r-~'gar-/-dl'ngdamage to normal tissue a f t e r interstitial irradiation. Supported by PHS Grant CA-3o445 and NIH 135z5

Center Grant

112 ga~diation neer~is v e r s u s t u m o r r e c t e r e n e e following 8 1 ~nterst~tial brachytherapy; utility o f tissue culture studies Mark L Rosenblum, Dolores V. Dougherty, Riehard Davis, Michael Edwards, Philip Gutin. Brain Tumor Research Center, University of California, San Freneisco. interstitial braehytherapy with 1251 appears to have improved the outcome of patients with malignant gliomas. To date favorable responses have been noted in approx. ?5% of more than 100 eases treated. In approx. 25% of eases reoperations for inereasing mass lesions were performed 4-12 months after implantation. Pathological evaluation of specimens from the latter operation has demonstrated the presence of abnormal cells consistent with tumor in the vast majority of patients, although many have not demonstrated elinieal recurrence. The inability of routine pathologleal studies to differentiate between clinically recurrent and clinieally non-recurrent patients (putatively pure radiation necrosis) prompted us to study the growth of such speeimens in the laboratory.

Specimens were obtained a t the post-seed operation from 14 sequential patients who had received 5-12,000 red of braehytherapy given adjuvantly or at tumor recurrence following 4-6,000 red of orthovoltage therapy (total tumor dose 10-18,000 red). All patients were followed a minimum of 6 additional months; 6 demonstrated clinical recurrence and 8 showed no tumor growth over 10-40 months, without further therapy in 6/8 cases. Histological study of all 14 specimens showed abnormal cells consistsnt with tumor. By contrast, 4 tumors grew in culture of the 6 with clinical recurrence, and only I tumor (an ependymoma) grew in culture of the 8 with no evidence of clinical growth. The predictive value (P.V.) for clinical growth was 43% for pathological evalutation and 80% for culture growth; the P.V. for clinical growth was 0% for pathology and 78% for culture studies. Although statistical analysis of the culture test showed pc0.06 (one-tailed Fisher Exact Test), the difference from standard methods of tumor evaluation is apparent. Future confirmation of clinical status will determine i f such culture studies could help determine which patients should receive further chemotherapy. Supported by CA 31882 and CA 13525

EARLY RESULTS OF HYPERFRACTIONATED RADIOTHERAPY FOR THE TREATMENT OF PRIHARY BRAIN STEM CLIOMAS OF CHILDHOOD. Roger J. Packer, H.D., Phillip A. Littman, M.D., Derek A. Bruce, M.D., Leslie N. Sutton, M.D. and Luis Schut, M.D. Divisions of Neurology, Radiation Therapy and Neurosurgery, Children's Hospital of Philadelphia, University of Pennsylvania With current means of treatment, the median length of survival in children with brain stem gliomas (BSG) is between g and 15 months, and 5-year survival is uncommon. Radiotherapy (RT), the only effective treatment for BSG, usually results in clinical improvement which may be dose dependent. However, the dose of RT which can be delivered Is limited by radiation tolerance of the normal surrounding brain. Hyperfractionation of RT theorectically allows for the delivery of higher doses of RT with less detrimental effects on normal brain. In August 1984, a limited institution, single-arm trial of hyperfractionated RT in children wlth BSG was begun. Patients, twice daily, received 120 rad separated by 4 to 8 hours. They were treated 5 days a week to a total dose of 6480 rad (conventional RT: 5000-5500 red). To date, II patients, median age at diagnosis -- 6 years (range 1.5 to 15 years) have been treated. All patients have had some degree of improvement at the completion of treatment -- 8 are alive and free of progressive disease, a median of 5 months post-diagnosis. Three patients developed progressive disease 3, 5 and 5 months after treatment. Autopsy was performed in one patient and there was no evidence of radiation necrosis. Treatment has been well-tolerated and there has been no increase over expected skin or mucosal irritation. No patient neurologically deteriorated during treatment and most were off glucocorticoids before the end of therapy. These preliminary results suggest that hyperfractionatlon RT of the brain stem in childhood is well-tolerated. Although therapy is delivered at higher than conventional doses to a swollen brain stem, no signs or symptoms of brain swelling have been seen. It is too early to discuss efficacy of this treatment. Outcome and evidence of late sequelae will be updated. 83

82

8romodeoxyuridine (BUdR): A Rationa| Intra-arterial Radiosens]tlzing Treatment for M a H g n a n t Astrocytomas of the Centra| Nervous System. Greenberg H5, Chandler WF, Ensminger WD, Lichter AS, Oiaz RF, Junck L, Page H. University of Hichigan, Ann Arbor, HI

Radiation therapy has provided the most effective adjunct treatment for patients with m a H g n a n t astrocytomas; however, brain into]erance to potentia]]y curative doses of radiation therapy has limlted the radiation dose de]iverab]e. BUdR is a ha]ogenated pyrimidine ana]o9 that is incorporated into the DNA of dividing cel]s and sensitizes these ce]Is to radiation therapy. BUdR wi]] not be incorporated into neurons or g H a to any significan~ degree because of their very ]ow mitotic rate. Differential sensitization wi]] be obtained between tumor where radiation therapy effect wi]] be enhanced and norma] brain which wi]1 be unenhanced. 8UdR is delivered intra-arteria]ly because it Is rapid]y deha]ogenated in the ]iver to an inactive form. Russo et a] (Cancer Res. 44:1702-05, i084), ca]culated intra-arteria] carotid administration wou]d give a Ii to 16-fold increase in tumor exposure than intravenous administration of an identical dose. At the University of Hichigan we have developed a permanent]y imp]antable Infusaid infusion system for safe intra-arterial carotid 8UdR de]ivery {Phi]]ips e t a ] , Neurosurgery ll:Z13-Z18, i082) and wi]] implant this system lO days following primary surgica] resection and/or biopsy. BUdR infusion wil] begin two weeks prior to starting radiation therapy and continue throughout radiation. The BUdR dose wi]| be 500 mg/m2/day at a pump f]ow rate of 6 to 8 cc/day. Fo|]owing radiation therapy patients wi|| be fo||owed to time of treatment fai|ure and the resu|ts will be compared wlth radiated historica| controls of grade il[ and IV astrocytomas of the Brain Tumor Study Group. We have imp|anted 3 patients and infused BUdR in doses of 250-350 mg/cc at f|ow rates of 2.3 - 3.2 co/day for up to one month. The only side effect to date has been mye|osuppression not requiring BUdR dose reduction.

COHBINATION CHEMOTHERAPY WITH CIS-PLATINUM (CPDD), LOHUSTINE (CCNU) AND VINCRISTINE (VCR) IN CHILDREN WITH RECURRENT AND NEWLY-DIAGNOSED PRIMITIVE NEUROECTODERmaL TUHORS -- HEDULLOBLASTOMA (PNET-HB) OF CHILDHOOD. Roger J. Packer, M.D., Kathy R. Siegel, P.A., Leslie N. Sutton, H.D., Derek A. Bruce, M.D. and Luis Schut, M.D. Children's Hospital of Philadelphia, University of Pennsylvania 84

Many drugs have been at least partially effective in the treatment of PNET-MB, but response is usually transient. Three of the most effective drugs have been cls-platinum (CPDD), lomustine (CCNU), and vincristine (VCR). We have employed these drugs in combination: CPDD - 90 mg/m 2, CCNU iOO mg/m 2, end VCR - 1.5 mg/m 2 in 7 patients with recurrent PNET-HB and in 3 newly-diagnosed patients under age 2, prior to radiation therapy. Eight (80%) have had an objective response - 4 a complete response; 2 a partial response; and 2 improvement. The median time for length of remission has not yet been reached, and 6 patients remain in remission, ranging from 2 to 18 months. Fourteen other patients with newlydiagnosed P N E T - ~ with staging studies predicting a high rate of relapse after radiotherapy alone (disease dissemination at the tlme of diagnosis, partial surgical resection and/or cellular differentiation) have been treated with the same drugs, 6 weeks after the completion of radiotherapy. Eight 6-week cycles of therapy are given. Thirteen (93%) remain in disease-free remission, a median of 7 months after diagnosis; 4 of whom have completed all therapy. Ototoxiclty was present in 88 percent of patients by the fifth course of treatment and necessitated dosage modification in 75 percent of patients by the sixth cycle of treatment. Grade III renal toxicity was present in the majority of patients by the sixth cycle of therapy. CPDD, CCNU and VCR is an effective, relatively well-tolerated therapy for children with recurrent PNET-MB and results to date compare favorably to response rates obtained with single agents or more complex drug regimens (including g-drugs-in-l-day therapy). This combination is effective in at least delaying relapse in children with "poor-risk" PNET-HB, although more time is needed to see if it will favorably impact on long-term disease-free survival.

113 85

INTRACAVITARY PHOTO-DYNAMIC THERAPY (PDT) OF MALIGNANT PRIMARY BRAIN TUMORSUSING A LASER COUPLED INFLATABLE BALLOON Muller P.J.,

Wilson B.C. & Yanche J.C.

PDT in the treatment of malignant tumors consists of the administration of a photosensitizer followed by photoillumination of the malignant tissue. The cytotoxic effect of PDT is related to the Light dose and the tissue concentration of the photosensitizer. The cytotoxic effect is Limited by the attenuation of Light as i t is scattered and absorbed by the tissue. We f e l t i t important to develop a method were by an entire tumor resection cavity could be uniformly illuminated so that the surface dose (irradiance; J/sq.cm.) could be accurately calculated. To this end a Laser coupled inflatable balloon which is f i l l e d with a dispursion medium was designed. The device consists of a stainless steel cylinder 16 cm Long and 1.6 cm in diameter; i t contains 3 channels - a central channel carries the optical fiber and two side channels for f i l l i n g the balloon. The balloon is secured to the cylinder by a rigid nylon flange. Using a Photodyne 88XL photometer with #150 sensor head )measurements of the balloon surface irradiance were made at O, 45, 90, and 135 degrees with respect to the forward direction of the optic fiber for balloon diameters of 3, 4, and 5 cm. These revealed uniformity of surface irradiance when the dispursion medium (Nutralipid) concentration was 0.1%. Light absorption by the medium was calculated for diameters of 3, 4, S, and 6 cm; the % light loss was 11, 14, 18, and 21%, respectively. We have performed intracavitary PDT in 9 patients with malignant primary brain tumors. HPD (Photofrin I) was administered 24-48 hours preoperatively. The light dose range was 8-68 J/sq.cm. In spite of our light dose escalation no acute post-operative neurological deterioration was identified. The technique and results will be presented.

87

The in v i t r o and in v i v o c h e m o s e n s i t i v i t y V M d K murine astrocytoma c e l l l i n e s .

R.BRADFORD

J.L.DARLING and

of

D.G.T.THOMAS

Dept. of N e u r o l o g i c a l Surgery, I n s t i t u t e of Neurology, Queen Square, London, WCIN 3BG, U.K. 3 c e l l l i n e s d e r i v e d from the VMdK m u r i n e astrocytoma have been assessed as a therapeutic model of human glioma. Lines P.560 and P.540 were poorly tumourigenic in syngeneic mice. P.497 was u n p r e d i c t a b l e in i t s degree of t u m o u r i g e n i c i t y and t i m e t o death o f mice bearing i n t r a c r a n i a l tumours was v a r i a b l e (22-60 days). Cell l i n e s from control and drug treated mice bearing P.497 i n t r a c r a n i a l ( i c ) tumours were d e r i v e d . Two of these l i n e s 4 9 7 - c ( I ) and 4 9 7 - p ( I ) were h i g h l y tumouri genic both subcutaneously and i n t r a c r a n i a l l y and the time to death of untreated mice bearing ic t~mours ranged between 11 and 14 days. Using the ~OS-methionine uptake assay the spectrum of in v i t r o c h e m o s e n s i t i v i t y to ten c y t o t o x i c agents has been constructed f o r parent 497 and f o r s i x of the d e r i v e d c e l l l i n e s . When compared t o the IDso'S o f human b r a i n tumour short term cultures, 497-p(I) was predicted to be s e n s i t i v e t o CCNU and VCR but r e s i s t a n t t o PCB The in vivo chemosensitivity o f 497-p(1) has been assessed by growth delay and increased l i f e span in mice bearing sc and ic tumours r e s p e c t i v e l y . VCR, CCNU and BCNU caused l i m i t e d growth delay but s i g n i f i c a n t volume r e d u c t i o n of sc tumours. These agents also caused a small but s i g n i f i c a n t increase in l i f e span of mice bearing ic tumours. PCB had no s i g n i f i c a n t e f f e c t on e i t h e r sc or i c tumours. Cell line 497-p(I) provides a reproducible model of glioma which responds to a l i m i t e d extent to chemotherapeutic agents which have been shown to be e f f e c t i v e in human glioma.

86

DEVELOPMENT OF CNS "METASTASES" I N PATIENTS W I ~ GLIOBLASTONA NULTIFORNE AND OTHER ANAPLASTIC

GLIOMAS M I L E UNDERGOING THERAPY. V.A. Lcvin, A. Choucair, R.L. Davis, P. Silver, P. Gutln, M. Edwards, and C.B. Wilson. Brain Tumor Research Center, University of California Medical School, San Francisco, CA 94143 The purpose of this study was to determine the frequency of glioma "metastases" in patients undergoing radiation and chemotherapy. Patients were sought where there were neurodlagnostic studies (CT or myelogrem), autopsy, or surgical verification of a second lesion or demonstration that the tumor extended into a different lobe(s) from that originally established a t tumor presentation. To determine the frequency of metastasis, the Neuro-oncology Service database was searched for patients who were seen between 12/2/76 and 11/30/84, who had second lesions, a n d who had orlglnal diagnoses of glloblastoma multlforme (GM), highly anaplastic astrocytoma, gemistocytic astrocytema, moderately anaplastic astrocytoma, and mixed malignant astrocytoma. The last four tumor types were included categorically as other anaplastlc gliomas fAG). Of 249 GM patients, there were I0 (4%) cases of CNS metastasis, 5 of which were spinal cord metastasis. The median age of the GM patients was 49 years (range 29-70) and the median weeks between first diagnosis and metastasis was 54 weeks (range 13-110). Of 488 AG patients, there were 26 (5%) cases of CNS metastasis, 7 of which were spinal cord metastasis. Only I cas~ of extra-CNS metastasis was found. The median age of the AG patients was 31 years (range 11-60) and the median weeks between first diagnosis and metastasis was 88 weeks (range 1L-387). The importance of these observations for radiation therapy will be discussed.

88

P r e p a r a t i o n of m u l t t c e l l u l a r spheroids from cell lines derived from the VHdK spontaneous murine astrocytoma(SHA)

R.Bradford, D.G.T.Thomas

N.Hendry,

B.MittaI,

J.L.Darling

Dept. o f N e u r o l o g i c a l Surgery, I n s t i t u t e Neurology, Queen Square, London WCIN 3BG, UK

of

Multicellular s p h e r o i d s have s i m i l a r t h r e e dimensional microstructures to solid tumours and develop m i c r o e n v i r o m e n t s and r e d i s t r i b u t i o n o f c e l l growth k i n e t i c s s i m i l a r t o tumours in v i v o yet they are amenable to controlled experimentation in v i t r o . Cell lines derived from the VMdK SMA have been used to assess the degree to which t h i s model system of glioma is capable o f f o r m i n g spheroids. Using the l i q u i d o v e r l a y technique we have prepared spheroids from l i n e s P.497 and 497-p(1). Measurements o f spheroid volume have shown that exponential growth occurs when spheroids are placed on s o f t agar. Spheroids prepared from 497-p(1) have been used to assess the e f f e c t o f CCNU on growth by c a l c u l a t i n g the volume growth delay index. Exposure t o CCNU f o r 72hrs at 0.2, I , and 5 ug/ml caused volume growth at each dose. The calculated growth d e l ~ indices f o r each dose o f CCNU were 1.27, 1.41, 1.90 respectively. R e s u l t s o b t a i n e d from t h e s e e x p e r i m e n t s h ~ e been compared w i t h t h o s e obtained from oOS-methionine uptake assay and f r o m i n v i v o e x p e r i m e n t s w i t h mice b e a r i n g subcutaneous and i n t r a c r a n i a l t u m o u r s . The e f f e c t s o f VCR, PCB, and BCNU on m u l t i c e l l u l a r spheroids are presently being studied.

114 INTRAKRTERIAL PLUS SYSTEMIC CHEMOTHERAPY IN THE TREATMENT OF INTRACEREBRAL TUMORS, D.J. Stewart, Z. Crahovae, H. Hugenholtz, B. Benolt, M. Richard, N. Russell, J. Maroun, J. Dennery, E. Peterson, J. Nabwangu. The Ontario Cancer Treatment and Research Foundation Ottawa Regional Cancer Centre, the University of Ottawa Faculty of Health Sciences, and the Centre Hospital Regional de l'Outaouals. A study of combined intraarterlal and systemic chemotherapy was initiated for brain tumors. The carotid or vertebral artery was catheterized temporarily using a transfemoral approach. A 0.2 ~m in-llne filter was used. Doses (mg/m 2) of letraarterlal BCNU, clsplatln, and 9%4-26 were I00, 60, and 150, respectively, for carotid and 100 (no ethanol), 40, and 100, respectively, for vertebral artery infusions. Each drug (total final volume i00 ml using 0.45Z saline) was given separately over 20 min. Dexamethasone 50 mg was given IV once before and once after treatment. Mannltol 50 g was given IV with cisplatin. Intraarterlal infusions were repeated q 7 wk x 3. Systemic drugs included: bleomycio 30 units IV q 1 wk x 10; vlnerlstine 2 mg IV day i, course i, then days 8 and 36 each course; VM-26 50 mg/m 2 IV on day 8 each course, preceded by glycerol 500 mg/kg p.o. q 6 h x 4 doses; methotrexate 200 mg/m 2 IV day 8 and 36 each course followed in 24 hr by cltrovorum 20 mg p.o. q 6 h x 8 doses; procarbazlne I00 mg/m2/day p.o. days 22-28 each course. 22 patients with glloblastomas, 1 patient with brain metastases, and 1 patient with a primary germ cell tumor of brain have been entered. Prior therapy included XRT (23 patients) and chemotherapy (12 patients). 18 are evaluable for response (including 3 who failed to complete course i, 1 died early (pulmonary embolus), 3 are too early. Ten (55Z) responded by CT, including 10 of 16 evaluable glloblastomas (63Z). This response rate is comparable to that we previously obtained with the same intracarotid doses without systemic chemotherapy. Toxicity included ipsilateral bllndness/ophthalmoplegia (i), pulmonary toxicity (2), possible hepatic toxicity (i), stomatltls (2), myelosuppression (frequently dose-limltlng). Further different intraarterial plus systemic chemotherapy studies are planned. 89

91

PHASE I STUDY OF INTRACAROTID PCNU. D.J. Stewart, Z. Grahovac, S. Gupta, L, Coumnerova, M. Richard, H. Hugenholtz, J. Maroun. The Ontario Cancer Treatment and Research Foundation Ottawa Regional Cancer Centre and the University of Ottawa Faculty of Health Sciences. Based on observations that PCNU 30-60 mg/m2 caused no retinal or neurological toxicity when injected into the common carotid arteries of dogs (L. Goumnerova et el, submitted for publication), a phase I study of Intracarotid PCNU has been initiated in humans. Four patients have been treated with PCNU 60 mg/m 2 in 250 ml D5W infused into the internal carotid artery over 1 hr using an Infraophthalmlc catheter inserted via the femoral route. Drug was profiltered using a 0.5 micron filter, and a 0,2 micron in-line filter was also used. Patients were treated with dexamethasone 50 mg IV once before and once after the intracarotld treatments. No heparin was used. Two patients had eonjunetlval erythema in the ipsilateral eye and periocular moderate discomfort, one patient had mild ipsilateral forehead discomfort, and one patient had absolutely no pain during the treatments. In both of the first 2 patients, discomfort cleared rapidly as soon as the infusion ended. None of the patients were pretreated with analgesics. No decrease in vision or neurological toxicity has been noted, but followup time has been short. It is too early to evaluate patients for response. Further experience will be necessary before we will be able to judge whether intracarotld PCNU offers any advantage over intracarotld BCNU, although the ocular pain experienced by patients has been less than that seen with BCNU I00 mg/m 2. The study continues. Subsequent patients will he treated at a PCNU dose of 75 mg/m 2 if no late toxicity is seen in these first four patients. 90

MITOXANTRONEHYDROCHLORIDE: UPTAKE INTO HLrHANBRAIN

TUMORS AND PHASE II STUDY IN GLIOMAS. D.J, Stewart, H. Bugenholtz, R. Green, H. Richard, B. genoit, N. Russell, J. Marouo, M. Thlbault. The Ontario Cancer Treatment and Research Foundation Ottawa Regional Cancer Centre and the University of Ottawa Faculty of Health Sciences. Eight consenting patients undergoing surgical resection of intracerebral t%~mors were given mltoxantrone 6 mg/m 2 (a subtoxle dose) IV preoperatively to determine the ability of this drug to penetrate into human brain tumors. Resected samples were homogenlzedj extracted and assayed by a high pressure liquid chromatographic technique. Tumor mltoxantrone concentrations in the 8 patients varied from 0.74161.15 ng/g at times ranging from 0.25 to 25.75 hr after drug administration. Tumor mltoxantroee concentration was higher than concurrent plasma mitoxantrone concentration in some patients. Our results indicate that mltoxantrone attains and maintains potentially cytotoxle coneentratloas in human intracerebral tumors despite previous documentation of only low concentrations in normal nervous system. A phase II study of mltoxantrone 16 mg/m2 IV q 3 wk was initiated in adults with gllomas. Sixteen patients were treated. One asymptomatic patient had tumor shrinkage on his CT scan. Two symptomatic patients had apparent tumor shrinkage on CT scan but this was not accompanied by neurological improvement. Hence, mltoxantrone has a small degree of activity against gliomas. Toxicology studies of intracarotld infusion of mitoxantrone have been initiated In dogs. Supported by Cyanamid Canada, Inc.

92

DNA SINGLE STRAND BREAKS AND CELL SURVIVAL IN 9L RAT BRAIN TUMOR CELLS BY ACRIDINE ORANGE PHOTOACTIVATION WITH ARGON LASER. D.G. Gordon, D.F. Deen, and M.S. Edwards; Brain Tumor Research Center, University of California, San Francisco, CA. 94143.

Acridine orange (AO) is a diaminoacridine dye with a strong affinity for RNA and DNA, and interealates into double stranded nucleic acids. AO has metaehromatie properties with a spectral range of 514.5 to 454.5 nm making photoradiation therapy (PRT) ideal using the argon laser which has a frequency peak at 488 nm. We have measured single strand breaks using alkaline elutioo to study DNA damage produced by AO in cultured 9L cells after exposure to low energy (1 Joule) argon laser light. Cells kept in iced phosphate buffered saline after exposure to AO (0.005 to 3.7 I~M) for 2 hr were irradiated with 30 mW for 35 see. Laser exposure alone produced very few strand breaks and was not distinquishable from untreated controls or 0 reds equivalent of X rays (RE). AO (3.7 IJM) treated cells kept in the dark produced strand breaks equivalent to 100 RE. The highest dose of AO (3.7 gM) exposed to laser light produced strand breaks greater than 400 RE. Lower concentrations o f AO (0,005 and 0.01 ~JM) and laser exposure did not produce more damage than AO (3.7 IJM) kept in the dark (100 RE). Colony forming assays revealed little cell kill for varying doses of AO (0.1 to 3.7 ]JM) kept in the dark and for laser exposure alone. PRT with AO produced nearly 1 log cell kill over laser (1 joule) alone or AO (3.7 IJM) in the dark. Considering the amount of strand breaks seen with identieal treatment, a greater cell kill would have been expeeted. These results suggest that DNA single strand breaks may be repaired by the cells. The rate of this repair as compared with repair of X-ray damage is currently being investigated. Supported by A C S PDT-233.

115

93

INITIAL CHEMOTHERAPY FOR CNS MALIGNANCY IN PATIENTS UNDER 3 YRS. STRAUSSL.C., KILLMOND T., MARIA B.L., CARSON B., WHARAM M.D.; JOHNS HOPKINS ONCOLOGY CTR.,. BALTO., M.D. 21205.

The poor prognosis of patients presenting very early in life with intracranial malignancy is aggravated by the long-term anticipated toxicity of irradiation (RT) on neuropsychologic development, endocrine function, and spinal vertebral growth. Initial chemotherapy has potential to delay RT until maturation has advanced and to debulk tumor, possibly increasing efficacy of RT. We are treating five unirradiated infants ( < 36 mos. at dx.) with the combination of cisplatin (20 mg/M 2 IV) and VP-16 (etoposide, 75 mg/M 2 IV) q.d. x 5; courses initially q. 3-4 wks. for intracranial tumor: I PNET with glial differentation previously failing MOPP, [ medulloblastoma, and 3 posterior fossa ependymoma (2 anaplastic, I invasive). 3/5 had positive CSF cytopathology, 4/5 had macroscopic leptomeningeal dissemination prior to chemotherapy. Toxicity is moderate (mean 4.2 courses/pc, so far). Nausea/vomlting persists 2-7 days post-Rx, especially in patients with postoperative G.I. dysfunction. Serial BAERs show stable auditory function during treatment, except for tumor-related brainstem transmission deficits. Serial GFRs have sho~al no significan~ renal functional decrement, but patients may require Mg +g supplementation. Neutropenla is moderately severe (mean nadir neutrophil ct. = 380/mm 3 at day 14 + 4) and thrombocxtopenia is mild (mean nadir platele[ ct. = 9 8 , 0 0 0 / ~ = at day 15 ~ 3). No neuroCoxicity of therapy has been detected. Four patients had radiographically-evaluable disease (2 primary site residual, 2 intradural metastases). All four demonstrated rapid and marked tumor shrinkage after two courses (> 50% reduction in two dimensions). Excellent progressive neurologic and general clinical improvement has been noted since surgery in all 5 patients; one relapsed after 5 courses. Eight courses are planned; all patients will receive RT at age 36 mos. Ultimate efficacy and late effects, including psychological development, of thls Rx are being studied. Future regimens should evaluate primary chemotherapy with delayed RT in this setting.

Intrathecal ACNUagainst leptomeningeal dissemination 9 5 o f tumor: Experimental study. Y. Ushio, N. Arita, S. Ken, M. Nagatani, T. Hayakawa, H. Mogami. Osaka University Medical School, Osaka, Japan. Incidence of CSF dissemination of brain tumors is relatively high, however, no sufficient treatment has been established. We studied intrathecal (IT) ACNUwith respect to drug distribution, toxicity and efficacy using an animal model of leptomeningeal tumor induced in Sp~ague-Dawley rats by intracisternal inoculation of 1 x 10~ Walker 256 tumor cells. Radioautographic study revealed that 14C-ACNU (1 ,Ci/O.lml) administered into the cisterna magna rapidly distributed in the CSF space and was cleared from the CSF. 14C-ACNUdistributed to the tumor in the CSF space as well until tumor growth was limited to 10-20 layers of cells. As tumor grew, spread of 14C-ACNU in the CSF was blocked. Once tumor formed a mass, lqC-ACNU could reach only to the marginal area of the mass consisting of 20-30 layers of tumor cells. 14C-ACNUalso distributed in brain tissue to a depth of I mmfrom the cortical surface. Normal rats which received more than 3.0 mg/kg ACNUinto the CSF continued to lose weight, however, those receiving less than 1.5 mg/kg gained weight about the same as control animals. Immunohistological examination using anti-GFAP serum revealed gliosis approximately I mm in width along the cortical surface in the animals receiving more than 3.0 mg/kg ACNUinto the CSF. This gliosis was minimum or mild in the animals receiving less than 1.5 mg/kg. ACNU at a single IT dose of 1.5 mg/kg on Day 2 or Day 5 after tumor inoculation increased median survival time of animals with leptomeningeal tumor more than 350%. These results suggest that IT ACNU is promising in the treatment of leptomeningeal dissemination of tumor in the early stage and warrants clinical t r i a l s .

94

PILOT S T U D Y CHEMOTHERAPY

OF EIGHT-DRUGS-IN-ONE-DAY FOR HIGH-GRADE ASTROCYTOMAS

Jaek Rozental, Donald Trump, Jonathan Finlay, lan Robins, Richard Steeves, Per Langeland, and Henry Sehutta Nine adults and four ehildren with high-grade astroeytomee have been treated with eight-dru~2-in-one-dny chem.. therapy: methylprednisolone, 300 mg/m ; VCR 1.5 mg/m ; CCNU 75 mg/m2; proearbazine 75 mg/m2; hydroxyureu 3000 mg/m2; eisplatin 90 mg/m2; Ara-C 300 mgJm2; and DTIC 150 mg/m 2. This eombinee call eyele active and independent agents with lipid- and wateP-soluble ones. The four pediatric patients have reeeived a total of 18 eotmsee and the nine adults 32 courses of therapy. All except one adult (92%) have responded as judged by CT and MRI. Two adults have completed therapy and at 12 and 16 months of follow-up show no signs of residual tumor. Three of four pediatrie patients reeeived the regimen as salvage therapy. The fourth patient, diagnosed st age 6 months, eompieced l year of therapy and is now undergoing radiotherapy. Toxieity of the regimen has been low. The percentage of eou~ee assoeiated with HCT <: 25 was 2%; PLT < 60 k wee 6%; WBC < 2 k wee i0%. Three eouvses have been eomplieated by infeetions. Renal and ototoxieity were eueh seen in 1 adult. These results are very encouraging. We expeet that this high response rate and low toxieity will translate into prolonged survival.

96 D F

TIIE ~TBROCBNEITY IN TUMOR CBLL CHBMCSBNSITIVITY AS DBTEP~II~D BY SISTER CIROi~TID BXCH~IGBS (SC~s) Deen, h g Kendall, L J Mart0n and P J Tofilon,

Brain Tumor Research Center, University of California, San Francisco, CA 94143 The various tumor cell types that comprise many human tumors are thought to be heterogeneous with respect to chemosensitivity~ Therefore, assays that are used to indicate clinical s e n s i t i v i t y of tumors to specific drugs can provide misleading r e s u l t s if the assays are based on the average s e n s i t i v i t y of the cells within the tumor. Ve have found that the heterogeneity of response of the ceils within human tumor cell lines to BCNU can be detected using the SCB assay, Human brain tumor cell lines were exposed for 1 hr to 10 uU BCNU. After treatment, medium containing bromodeoxyuridine was added to the cultures and the cells were alloyed to replicate for two cell cycles. Mitotic cells were collected using colcemide and analysed for SCBs using standard techniques. The background level of $Cgs for all cell lines was -0.1. Frequency histograms of SeEs per chromosome induced by the BCNU revealed considerable heterogeneity in drug s e n s i t i v i t y within individual cell lines. Based on colony forming efficiency, cells were judged to be r e s i s t a n t to the drug if fever than 0.5 SCSs psi chromosome were induced and sensitive i f more than 0.5 SC~s per chromosome were induced by the BCNU treatment. Application of t h i s method as a predictive clinical assay is suggested for those drugs where a strong correlation exists between cell killing and SCE induction.

Supported by NIH grant CA-34351.

116

97

PCNU and Recurrent Childhood Brain Tumors. Jeffrey Allen, Russell Walker, Charlotte Tan. Memorial Sloan-Kettering Cancer Center, NY.

PCNU, the latest nitrosourea analogue to be subjected to clinical t r i a l s , h e l d promise as a superior chemotherapy agent for brain tumors because of more favorable biochemical and cytotoxic characteristics in laboratory studies. Thirty-nine children with evaluable recurrent primary CNS tumors participated in our phase I I PCNU t r i a l . Their mean age was 9.7 years. PCNUwas administered as a 2 hour IV infusion in one of 2 dose schedules at 6-7 weeks intervals; I00-~25 mg/m2 for minimally treated patients and 70-90 mg/m for heavily treated patients. Response was assessed after 2 courses of chemotherapy after attempting to taper the steroid dose. The overall response rate (complete plus partial response) was 18% (7/39) for a mean of 5.9 months (2+ - 1 2 ) . Diseasespecific responses rates were: brainstem glioma - 18% (3/17); cerebral glioma - 27% (3/11); ependymoma I / I ; and primitive neuroectodermal tumors - (0/9). Significant thrombocytopenia (< 50,000) was encountered in 30/38 (79%) patient t r i a l s . Activity of PCNU in recurrent childhood gliomas is confirmed. Our response rates for childhood gliomas, using objective CT c r i t e r i a , are similar to those -eported for CCNU and BCNU. Becauseof comparable hematologic toxicity and efficacy, there doesn't appear to be an advantage to continuing clinical t r i a l s with PCNU in children with recurrent brain tumors using a 2 hour IV infusion schedule. Supported in part by NCI-POI-CA 2964

99 v

INDUCTIVE ~

RADIOFRKQUENCY HEATING TUMOR NODEL

IN

A RAT BRAIN

Mark B e r n s t e i n , M.D., Rodney D ' S i l v a , and John W. Hunt, Ph.D., D i v i s i o n o f N e u r o s u r g e r y , T o r o n t o Western H o s p i t a l , University of Toronto, and the Ontario Cancer Institute, and the Department of Medical Biophysics, University of Toronto, Toronto, Canada To study the synergistic antineoplastic effects o f radiation and hyperthermie, a safe and effective method of deep local heating is essential. In this experimental model, 6 mm-diameter 9L tumors grown in the brains of F-344 rats are heated by the application of an external radiofrequency (RF) inductive heater. The heating is generated by the rapidly changing ~30 MHz RF magnetic fields produced by an 8 mm-square 4-turn copper solenoid. The side of the solenoid is placed 1.5 imn from a thin, silicone membrane which is applied to the rat's scalp, and the magnetic fields produce eddy c u r r e n t s in the rat's brain, and in turn, produce h e a t t h r o u g h ohmic l o s s e s . The magnetic field strength decreases gradually i n t o t h e d e p t h o f the t i s s u e ; t h e r e f o r e the a p p l i c a t i o n Of surface cooling by circulating liquid is needed, so that a better heating distribution is obtained. Temperatures are measured with intraparenchymal fine ( ~ " 4 0 p-diameter) chromel-alumel thermocouples supported by 5-0 silk sutures. The thermocouples are inserted laterally through the skull and brain by a fine hollow needle, and during the heat treatment, these are transversely scanned mechanically so that the heating distribution can be o b s e r v e d a c r o s s t h e f i e l d . Experiments with conducting gel phantoms indicate t h a t h e a t i n g to t h e r a p e u t i c temperatures at a depth of 6 mm is feasible with this system. Preliminary studies in rats with intact skulls suggest that reasonable heating distributions in the brain can be achieved although technical problems (such as interference between the RF currents and the thermometry system) remain to be solved.

MISONIDAZOLE AND CCNU CHEMOTHERAPYFOR RECURRENTPRIMARYMALIGNANTBRAIN TUMOR. Dorcas S. Fulton, M.D., Raul C. Urtasun, M.D., Cross Cancer Institute, University of Alberta, Edmonton, Alberta. CCNU chemotherapy prolongs survival of patients with primary malignant brain tumor when given at the time of tumor progression following radiation therapy. Used as a single agent, response rates of 30 to 80 percent have been reported with median response durations of 5 to 6 months. Experimentally, tumor cytotoxicity is enhanced using the combination of misonidazole and CCNU, without increasing myelotoxicity. In this phase I / I f study, 18 patients with primary malignant brain tumor which recurred following radiation therapy were treated with combined CCNU and misonidazole. CCNU120 mg/ML was given four hours following misonidazole 3.5 g/M2 every six to eight weeks, with dosage adjustments for myelotoxicity. Treatment was continued for one year or until tumor progression. Of the 16 patients in the study for one year or more, 11 (69%) survived one year, and six (38%) remained free of tumor progression for one year. Median time to tumor progression was 26.5 weeks and median survival was 86 weeks. No severe complications resulted from myelotoxicity. One patient developed mild peripheral neuropathy which disappeared following discontinuation of misonidazole. 98

This study was begun as an in st it u t io n a l p i l o t study and became study #83-16 of the RTOG. Supported by AHSTF (ACR) #6236 and ACB, Alberta, and CA 32768-02 US Public Health Service.