Pharmacology and Neurotoxicity of C/S-Diamminedichloroplatinum

[CANCER RESEARCH 43. 5278-5285, November 1983]

Pharmacology and Neurotoxicity of c/s-Diamminedichloroplatinum, Bleomycin, 5-Fluorouracil, and Cyclophosphamide Administration following Osmotic Blood-Brain Barrier Modification1

Edward A. Neuwelt,2 Peggy A. Barnett, Mark Glasberg, and Eugene P. Frenkel

Departments of Neurosurgery and Biochemistry, Oregon Health Sciences University, Portland, Oregon [E. A. N., P. A. B.]; Department of Medicine [E. P. F.], and Pathology [M. G.], University of Texas Southwestern Medical School, Dallas, Texas; and the Veterans' Administration Hospitals, Dallas, Texas and Portland, Oregon

ABSTRACT clinical studies of patients with malignant brain tumors, we have opened the BBB osmotically 245 times in 42 patients with little c/s-Platinum, bleomycin, 5-fluorouracil, and Cyclophosphamide toxicity and have preliminary evidence of some efficacy of chem were administered to rodent and canine models of osmotic blood- otherapy given in association with BBB opening (24-26). brain barrier modification to evaluate the relationship between The neurotoxicity of these drugs when administered by con tissue drug concentration in brain and the physiological and ventional routes is uncommon and minimal (5, 16). However, neuropathological effects of the drug. The toxicity studies were clinical and laboratory studies of these drugs suggest that the carried out in the canine, and the pharmacological studies were basis of this limited neurotoxicity is the result of the prevention carried out in the rodent. Even without the osmotic procedure, of drug entry into the brain by the BBB. For example, studies of intracarotid c/s-platinum, in contrast to the other drugs, produced Adriamycin (doxorubicin) delivery in a rodent and canine model modification of the blood-brain barrier with resultant severe showed no immunoreactive drug detected in the cerebrum, neurotoxicity. Osmotic blood-brain barrier modification was used whereas following BBB modification significant concentrations to increase delivery of bleomycin and 5-fluorouracil to the ipsilat- of drug and its metabolites were found in the brain (20, 32). eral brain region, but the increased delivery was associated with However, significant neurotoxicity was seen when Adriamycin evident neurotoxicity. Cyclophosphamide administration in as was detectable in brain tissue, and neuropathological examina sociation with osmotic blood-brain barrier opening did not cause tion showed correlative necrosis and hemorrhagic infarcts (28, significant neural toxicity. These studies indicate that some 32). Thus, Adriamycin (doxorubicin) produced significant neuro chemotherapeutic agents given in association with osmotic toxicity in dogs when even a small amount of drug was delivered blood-brain barrier opening can result in neurotoxicity. The cor to the brain tissue (32). The present study is a direct extension ollary of the known limited neurotoxicity when these chemother of our earlier studies with Adriamycin (32) and examined the apeutic agents are used in a conventional manner appears to be relationship between drug entry into brain tissue and neurotox due to the presence of the blood-brain barrier. icity when c/s-platinum, bleomycin, 5-FU, or Cyclophosphamide was administered in association with BBB opening. INTRODUCTION

Although a variety of chemotherapeutic agents such as c/s- MATERIALS AND METHODS platinum, bleomycin, 5-FU,3 and Cyclophosphamide individually Canine Model and in combination have shown efficacy in systemic cancers, none has provided clinically significant response rates in the BBB Modification. Adult conditioned dogs (20 to 25 kg) were used treatment of either primary or metastatic lesions in the CMS. except where noted (i.e., see c/s-platinum studies and Table 2). The Although ultrastructural defects in the vascular endothelium are dogs were anesthetized with sodium thiopental (20 mg/kg), intubated seen in some areas of malignant brain tumors, functional mea with an endotracheal tube, and ventilated with a Harvard animal respirator surements provide strong evidence that the BBB is at least (Harvard Apparatus Company, Inc., Millis, Mass.). Anesthesia was main partially intact in most tumors and may even be completely intact tained with a 60% nitrous oxide to oxygen mixture and supplemental in some (22,23, 25). Thus, inadequate drug delivery to the tumor sodium thiopental. Intraoperatively, atropine sulfate (0.015 mg/kg i.v.) in the CNS may be an important basis for their poor responsive was administered to prevent stimulation of the carotid body, and Lasix (5 mg i.v.) (furosemide; Hoechst-Roussel Pharmaceuticals, Inc., Sum- ness. One approach to improve drug delivery is to bypass the merville, N. J.) was given to counterbalance the 1.5% increase in brain BBB by administering the drug intravascularly after the BBB is transiently "opened" by osmotic means (13, 23). Since such an water associated with osmotic BBB opening (35). BBB modification was performed using the technique described pre approach is feasible (24-26), it is critical to determine the dis viously by this laboratory (22, 27, 30-32). The left internal carotid artery position and potential toxicity of the agents delivered. In our was cannulated with a 16-gauge catheter via the common carotid artery. Fifteen min before barrier opening, Evans blue (2%, 3 ml/kg) was 1These studies were supported by the Veterans' Administration, NIH Grants administered i.v. to provide a marker dye of barrier opening since it is CA31770 and CA23115, the Oregon Medical Research Foundation, the John Raaf Institute, the Meadows Foundation, and the Southwestern Medical Foundation- known to bind tightly, but reversibly, to plasma albumin, and it therefore Kinsler Williamson Brown Fund. does not normally penetrate the normal tight junctions between cerebral 2To whom requests for reprints should be addressed, at Oregon Health Sci endothelial cells (36, 37). Mannitol (25%) (Abbott Laboratories, North ences University, Division of Neurosurgery (L472), 3181 S.W. Sam Jackson Park Chicago, III.) at 37Â°was filtered (0.20-Mm pore diameter; Nalge Company, Road, Portland, Oreg. 97201. 3The abbreviations used are: 5-FU, 5-fluorouracil; CNS, central nervous system; Rochester, N. Y.) and then infused into the internal carotid artery at a BBB, blood-brain barrier; i.e., intracarotid; i.a., intraarterial. rate of 1.5 ml/sec over 30 sec. Five min after mannitol i.e. infusion, one Received December 9, 1982; accepted August 9, 1983. of the following chemotherapeutic agents was infused over a 15-min

5278 CANCER RESEARCH VOL. 43

Downloaded from cancerres.aacrjournals.org on September 28, 2021. © 1983 American Association for Cancer Research. Toxicity of CNS Chemotherapy period or given as a bolus over 2 min: i.e. c/s-platinum (0.25 mg/kg to 2 noticeable staining; grade 2+, moderate blue staining; and grade 3+, mg/kg); i.v. c/s-platinum (2 mg/kg); i.e. bleomycin (3 to 12.5 units/sq m); dark blue staining. i.e. 5-FU (1 to 15 mg/kg); or i.v. cyclophosphamide (20 mg/kg). In control Determination of Tissue and Serum Drug Concentrations. Bleo animals, only chemotherapeutic drugs were infused, since in previous mycin content in serum and tissue was determined by radioimmunoassay studies the i.e. administration of 0.9% sodium chloride solution rather (3). than mannitol had no effect (27), and this regimen is more analogous to Radiolabeled [r/ng-4-14C]cyclophosphamide (New England Nuclear, clinical i.e. chemotherapeutic drug trials. As a result, the catheter was in Boston, Mass.) and 5-fluoro[6-14C]uracil (Amersham Corporation, Arling the internal carotid artery for 25 min in experimental animals and for 20 ton Heights, III.) were used to determine tissue and serum content of min in control animals. Even though the catheter occludes the internal these drugs. Radiochemical purity as determined by thin-layer Chromat carotid artery during this period, this is physiologically acceptable since ographie analysis was >93%. Ten /Â¿Ciofradiolabeled drug were mixed the internal carotid artery forms an anastomosis in the cavernous sinus with the administered unlabeled drug and given as described above. with the anastomotic artery of the external ophthalmic artery (10). The Samples for analysis were weighed in tared scintillation vials and then animals were permitted to awaken after the procedure and then exam digested with a tissue solubilizer (Soluene-350; Packard Instruments, ined for evidence of toxicity. The animals were observed for approxi Downers Grove, III.). Scintillation cocktail (Dimilume-30; Packard) was mately 29 to 52 days prior to sacrifice unless toxicity necessitated earlier added, and the radioactivity was measured in a Beckman LS5000 beta sacrifice. counter. Quench corrections were made, and the drug and metabolite Toxicity Evaluation in the Canine. Evidence of toxicity included the content (^g/g or MQ/ml) was determined based on the specific activity of following parameters: (a) signs of neurological dysfunction; (b) signs of the labeled drug. Therefore, in the case of bleomycin, immunoreactive visual dysfunction; (c) changes in gross motor function or behavior; (d) drug was measured which is primarily the parent drug due to the neuropathological evaluation of the brain at necropsy. specificity of the antibody (3). Quantitation of 5-FU and cyclophospha At sacrifice, the brain was removed, evaluated for Evans blue staining, mide was less specific since tracer amounts of each was added to and then fixed in 10% neutralized formalin for histopathological exami varying doses of unlabeled drug and both agents are metabolized in vivo nation. The degree of barrier opening was evaluated by the amount of (7,11). Attempts at assay of tissue and serum c/s-platinum levels have staining with the Evans blue dye; the animals infused with mannitol all been made with inconsistent results and therefore are not included in demonstrated staining of parenchyma characteristic of an altered BBB. this paper. Because of the extended and variable time to sacrifice, the staining was graded as either present (+) or absent (-). Following fixation, for at least RESULTS 1 week, 0.5-cm coronal sections were prepared and inspected grossly. Histological sections of any gross lesions were prepared and stained Control Studies in the Canine Model. Four conditioned dogs with hematoxylin and eosin. In addition, we routinely prepared and were given an infusion of mannitol into the internal carotid artery reviewed histological sections of cerebral cortex, white matter, basal at a rate of 1.5 ml/sec for 30 sec to evaluate the effect of such ganglia, and brain stem. One section was taken from the frontal lobe in infusions (Table 1). No animal had sequelae from the procedure, the anterior and middle cerebral artery distribution including the white observation of the animals from 29 to 52 days demonstrated no matter of the corona radiata extending to the angle of the lateral ventricle. subsequent neurological abnormalities, and neuropathological A second section included posterior hippocampus, adjacent temporal examination at sacrifice showed no abnormalities. The degree lobe, thalamus, putamen, and adjacent structures. A third section was obtained through the pons or medulla and the adjacent half of the of barrier opening was evaluated by the amount of staining with cerebellar hemisphere because several animals demonstrated hemiation the Evans blue dye; the animals infused with mannitol all dem and secondary brain stem changes associated with supratentorial path onstrated staining of parenchyma characteristic of an altered ological changes. BBB (22, 27, 30-32). Because of the extended and variable time to sacrifice, the staining was graded as either present (+) or Rodent Model absent (0). For each drug studied, the initial dose was the same as that BBB Modification. Adult Sprague-Dawley rats weighing 250 to 300 g commonly employed for systemic (i.v.) administration. The max were used for the determination of drug concentration in tissues following imum conventional single doses are: 5-FU, 15 to 20 mg/kg; c/s- BBB disruption. Blood brain-barrier disruption was performed using the technique of Rapoport with minor modifications (30, 32, 33). Animals platinum, 2 to 3 mg/kg; cyclophosphamide, 20 to 40 mg/kg; were anesthetized with sodium pentobarbital (40 to 50 mg/kg i.p.). A bleomycin, 20 units/sq m. When toxicity followed this i.a. dose, catheter filled with sodium heparin and 0.9% sodium chloride solution subsequent doses were serially reduced. Cyclophosphamide was tied into the right external carotid artery for retrograde infusion. Five was given i.v. since hepatic activation is required for its function min before BBB modification, Evans blue (Chroma-Gesellschaft, Stutt (11, 42). In our studies, we evaluated the pharmacology and gart, West Germany) was administered i.v. (2%, 2 ml/kg). Mannitol (25%) toxicity in a range of 10 to 100% of these doses. If toxicity was (Abbott Laboratories), warmed to 37Â°,was infused cephalad for 30 sec observed at about 10% of the conventional i.v. dose, BBB into the right internal carotid artery via the external carotid artery catheter at a rate of 0.12 ml/sec. In control studies, 0.9% sodium chloride solution Table 1 instead of mannitol was infused at an identical rate and volume (0.12 ml/ CNS effects of i.e. mannitol infusion in the canine sec for 30 sec). Five min after the i.e. infusion of 0.9% sodium chloride Time (days) solution or mannitol, one of the following chemotherapeutic agents was of sacrifice infused over 15 min or as a bolus over 2 min: i.e. bleomycin (3 to 6 units/ Intracarotid infu Neurological Pathological after BBB BBB dis sq m); i.e. 5-FU (1 to 15 mg/kg), or i.v. cyclophosphamide (20 mg/kg). sion deficits findings disruption ruptkxi All animals were sacrificed 45 min after i.e. 0.9% sodium chloride solution MannitolMannitolMannitolMannitolNoneNoneNoneNoneNoneNoneNoneNone29+40 or mannitol infusions. Serum and tissue samples were collected and the +52 +37 extent of Evans blue staining was evaluated. + Evaluation of the Degree of BBB Disruption. The effectiveness of " Mannitol was infused into the internal carotid artery at a rate of 1.5 ml/sec for the barrier modification procedure was estimated by the degree of 30 sec. staining by the previously administered Evans blue. The degree of barrier "The success of BBB modification was determined by the presence or absence opening was graded as follows: grade 0, no staining; grade 1+, just of Evans blue staining at sacrifice.

NOVEMBER 1983 5279

Tabte2 CA/Seffects after varying doses of cis-platinum following BBB modification disruption in the canine Time of sacrifice after BBB BBB dis- doseaControlc/s-Platinum deficitsNone findingsNone animals 2 mg/kg i.v. days 1 mg/kgi.e.1 Tendency to the left Small multifocal regions of ne 36days3 for 48 hr crosis in left hemisphere mg/kg i.e.Neurological Head tilt to the left, Swollen left hemisphere,hemor- daysruption"++ died at 3 daysPathologicalrhagic necrosisdisruption34

Experimentalanimals (mannitol in fused) 1 mg/kgi.e.0.25 seizuresat softening and rÃ© days+48 3daysSeizures ipsilateralhemisphereNecrosissorption; swollen

mg/kgi.e.0.25 hrSeizuresat 48 rÃ©sorptionAmmon'sand hr+20 hornsclerosisNO"NDNDNDNDND3 mg/kgi.e.2 daysNoneSeizuresat 20 days+28 mg/kgi.e.02 days+48 mg/kgi.c.c1 hrNever at 24 hr+5hr mg/kgi.e.02 regainedcon +48 sciousnessNoneAtaxiaNever mg/kgi.vc2 hr+24 mg/kgi.v.c2 hr+24 mg/kg i.vcTorticollis; regained con hr + sciousnessHemorrhagic ac;s-Platinum,at the indicated dose, was infused into the internal carotid artery (i.e.)over 15 min beginning 5 min after BBB opening or i.v. over 15 min beginning 5 min after BBB opening. In control animals,c/s-platinum only was infused into the internal carotid artery or i.v. 6The success of BBB modification was determined by the presence or absenceof Evans blue staining at sacrifice. cThese animals were nonconditioned (i.e., general health status unknown) and were intended for only short-term toxicity study. Therefore, pathological specimens were not submitted for evaluation. "NO, not done. opening was not thought to be worthwhile since i.a. administra bolus routes of administration. tion alone would suffice. This is because either i.a. infusion (13, Correlative pharmacological studies were performed in the rat 27) or osmotic BBB opening (27, 33) increases drug delivery 5- model. As shown in Table 4, animals were given bleomycin at to 10-fold, and combined they potentially can increase delivery either of 2 doses, 3 and 6 units/sq m. One group had i.e. 0.9% to the CNS 50- to 100-fold (27, 33). sodium chloride solution, and the other had i.e. mannitol prior to c/s-Platinum Studies. In initial canine studies, c/s-platinum at the bleomycin. In the 0.9% sodium chloride solution-treated a dose of 1 mg/kg (28 mg/sq m) was administered via the animals, no evidence of barrier disruption was found, and no internal carotid artery but without osmotic BBB opening. In these drug was measurable in the brain. Drug concentrations were 2 animals, neurological deficit, mild in one and severe in the measurable in serum, liver, and kidney. In the animals given second, was seen (Table 2). Evans blue staining in these animals mannitol prior to the bleomycin, the concentrations of drug in showed that barrier modification resulted from the i.e. adminis liver, kidney, and serum were not significantly different from tration of c/s-platinum (1 mg/kg; 28 mg/sq m). To rule out emboli those found in the 0.9% sodium chloride solution (control) group. in the drug as the source of parenchymal damage, an additional In the mannitol-treated group, barrier disruption was documented dog was given filtered (0.2-Min filter) c/s-platinum i.e. BBB open by Evans blue staining and measurable, but variable concentra ing and parenchymal damage were still seen. tions of bleomycin were found in the brain. For unclear reasons, After osmotic BBB opening in a second group of animals, c/s- after BBB opening the ipsilateral brain levels after either 3 or 6 platinum was administered over a range of 0.25 to 2 mg/kg units/sq m are about the same, but much higher than in both either i.v. or into the internal carotid artery. As shown in Table 2, control groups. Similarly, liver and kidney levels are about the neurological deficits were seen in 7 of 9 dogs. Careful neuro- same at both doses in the two 0.9% sodium chloride solution- pathological studies in conditoned dogs given c/s-platinum into infused groups and the two mannitol-infused groups. The con the internal carotid artery at doses from 0.25 to 1 mg/kg showed centration in the ipsilateral barrier-modified cerebral hemisphere gross and microscopic hemorrhagic lesions as the evidence of was increased to levels well in the range found in liver and injury. kidney. In only 2 of the mannitol animals was drug identified in Bleomycin. Bleomycin was administered (3 to 12.5 units/sq the contralateral hemisphere due to mannitol crossing the circle m) by either slow infusion or bolus injection into the internal of Willis via the anterior communicating artery. This provides carotid artery. As shown in Table 3, 5 of 8 animals demonstrated support for the selective distribution of drug delivery following neurological deficits and 4 of the 8 animals had neuropathological barrier modification. abnormalities at necropsy performed 7 hr to 42 days after 5-FU. 5-FU was administered i.a. over a dose range of 1 to 15 therapy. All 8 animals had effective barrier modification from the mg/kg, and when given at doses of 5 to 15 mg/kg i.a. after evident Evans blue-albumin staining. Even at very low doses of mannitol induced barrier modification, significant neurological bleomycin (3 units/sq m), neurological functional deficits and lesions were seen (Table 5). At lower doses of 5-FU (1 mg/kg), neuropathological changes were seen in 2 of 3 animals. Unfor 3 of the 4 dogs had no neurological toxicity or evidence of tunately, the toxicity was seen by both slow infusion and rapid pathological change.

5280 CANCER RESEARCH VOL. 43

Tables CNS effects after varying doses of I.e. bleomycin following BBB modification disruption in the canine Time of sacri fice after BBS BBB disaip- disrup- Bleomycin dose* Neurologicaldeficits Pathologicalfindings tion tion

Control animals 12.5 units/sq m None Petechialhemorrhagesin me 35 days dulla representative of termi nal change

Experimental animals (mannitol in fused)12.5 units/sqm(slow days+36 infusion)12.5 units/sqm(slow hemiparesisPathologicalinfarctionSubarachnoidmiddle cerebral days+42 infusion)12.5 units/sqm(slow eatingNoneAtaxic hemorrhage,mi days+39 infusion)6 infarctionNoneNoneAcutecroscopic units/sqm(slow days+16 infusion)6 units/sqm(slow for 6days;seizures days+12 infusion)3 16daysSeizuresat

units/sqm(bolus)3 daysNoneNeverat 12 microhemorrhages,me days+43days dullaNonePetechial units/sqm(bolus)3 +7hr

units/sqm(slow regainedconsciousnessNoneLeftde-myelinationhemorrhagesand + infusion)NoneLeft in left hemi sphere, petechial hemor rhages in pons and medulla42 " Bleomycin, at the indicated dose, was infused into the internal carotid artery by slow infusion (over 15 min beginning 5 min after BBB opening with mannitol) or by bolus infusion (over 30 sec beginning immediately after BBB opening with mannitol). In control animals, bleomycin only was infused into the internal carotid artery. The success of BBB modification was determined by the presence or absence of Evans blue staining at sacrifice.

The drug concentration in the CNS achieved with i.a. 5-FU Adriamycin (32), the studies here were carried out in both the administration following i.e. 0.9% sodium chloride solution (con canine model (22,26,27,30-32) and the rat model (21,32). The trol) or i.e. mannitol (BBB modification) in rats is shown in Table rodent was used for the pharmacological studies because of its 6. Relative to the toxidty seen at 15-mg/kg doses of 5-FU, it is economy and simplicity. The canine model is more appropriate noteworthy that after barrier modification a nearly 10-fold in for the serial "clinical" observations of possible toxicity than is crease in drug delivery was seen (from 4.55 to 37.4 /Â¿g/g). the rat model. Studies of these models using methotrexate have Cyclophosphamide. Studies of the effect of cyclophospha- provided evidence that a good correlative relationship exists mide were carried out with the i.v. administration of the drug between these 2 models. Thus, comparing pharmacology in one prior to barrier modification to permit hepatic activation. As species with toxicity in a second species is not ideal, but we shown in Table 7, of 7 dogs given cyclophosphamide (20 mg/ believe that our studies with methotrexate justify the present kg) prior to osmotic barrier modification, 3 had evidence of economy and design. At a dose of 4 mg/kg, all tissue and serum neurological abnormalities. In 2, these were transient and without methotrexate levels are about 4 times higher in the canine (27) pathological sequelae. One animal did not regain consciousness than in the rat (21), but the increased brain delivery after osmotic after the barrier modification procedure; necropsy at 7 hr re BBB opening is the same in both species (5- to 10-fold). The vealed an infarction of the cerebrum and secondary hemorrhages difference between a 25-kg dog (0.957 sq m) and a 250-g rat of the pons and medulla. No other animal had neuropathological (0.036 sq m) based on weight is 100-fold, whereas the difference changes at sacrifice. based on body surface area is 26-fold. The resultant difference Correlative pharmacological studies with cyclophosphamide (4 times) in body surface area would account for the 4-fold (at 20 mg/kg) were done in the rat as shown in Table 8. Barrier increase in serum and tissue drug levels in the dog versus those modification did not enhance the delivery of the radiolabeled drug in the rat (21, 27). to the barrier modified cerebral hemisphere; labeled drug con The drugs examined (c/s-platinum, bleomycin, 5-FU, and cy centrations found in the contralateral hemisphere were the same clophosphamide) were selected because of their recognized as in the barrier-modified hemisphere. clinical efficacy in a variety of systemic neoplasms. Thus, they represented potentially important agents in the management of DISCUSSION selected metastatic neoplasms to the CNS as well as in the exploration of alternative therapeutic strategies for primary CNS The present studies examined the neurological sequelae of tumors. the delivery of a variety of chemotherapeutic agents after BBB In the present studies, i.e. c/s-platinum was found to have very modification and related these findings to the drug concentra significant neurotoxicity, even in the absence of osmotic BBB tions achieved in the brain. As in our previous studies with opening. The administration of c/s-platinum alone (filtered or

NOVEMBER 1983 5281

Downloaded from cancerres.aacrjournals.org on September 28, 2021. © 1983 American Association for Cancer Research. E. A. Neuwelt et al. Table4 Effect of BBB modification on bleomycin concentration In brain of the rat achieved by i.e. administration of drug (/iunits/g)Bleomycin Tissue bleomycin

dis hemi (^units/ (i.e.)"0.9% dose ruption0003+3+3+3+00002+2+3+3+Contra-lateralsphere000000000000009250610Â±5Ipsilateralhemisphere00008455302194159 mO1890127013301497 NaCIsolution3 units/sqm3 units/sqm3 mMannitol3units/sq

Â±21"15211517165Â±155140173442398288Â±1981170189011309741291

units/sqm(bolus)3

units/sqm3 units/sqm3 m0.9%units/sq

Â±5600000100158108231149Â±31588124048Â±29787275Â±77304139703846498 Â±20423404230192015702515

NaCIsolution6 units/sqm6 units/sqm(bolus)6

units/sqm6 mMannitol6units/sq

Â±166292283288 Â±59319101650392057503308

units/sqm6 units/sqm6 units/sqm(bolus)6

units/sqm(bolus)BBB

Â±30Liver1807130Â± 3Kidney174596673481Â± 5Serum Â±959 '' Bleomycin, at the indicated dose, was infused into the internal carotid artery by slow infusion (over 15 min beginning 5 min after i.e. 0.9% NaCIsolution or manmtol)or by bolus (over 30 sec beginning immediatelyafter i.e. 0.9% NaCI solution or mannitol). Animals were sacrificed 45 min after 0.9% NaCI solution or mannito! infusion. Â°MeanÂ±S.E.

Tables CNSeffects after varying doses of i.e. 5-FU following BBB modification in the canine of sacrifice afterBBBdisruption36 (mg/kg)"Control5-FU dose deficitsNoneSeizures findingsNoneNone tion5+ animals 15Experimental days26

animals (mannitol infused) 15 at 18 hr hr 151051 SeizureshrSeizures at 4 Petechiaein the pons and me 24hr18 ++ dulla at 18 hr Left middle cerebral artery infarc hr tionSmall Seizures at 18hrNone microinfarcts in the left 26hr43 ++ frontal lobe None days 11 None None 36 days + None None 36 days + +"5-FU,1Neurological Never regained Pontine hemorrhage,hemor- 6hrBBBdisrup consciousness, rhagic necrosis of the left par died at 6 hrPathologicalietal lobe, pons, and medullaTime at the indicated doses, was infused into the internalcarotid artery over 15 min, beginning5 min after

The success of BBB modification was determined by the presence or absence of Evans blue staining at sacrifice.

5282 CANCER RESEARCH VOL. 43

Downloaded from cancerres.aacrjournals.org on September 28, 2021. © 1983 American Association for Cancer Research. Toxicity of CNS Chemotherapy unfiltered) by the La. route actually resulted in "opening" of the Bleornycin has been studied in patients with brain tumors. BBB as documented by the Evans blue staining. In addition to Bioassay of brain tissue was done in 67 patients given 1 mg/kg this physiological alteration, c/s-platinum administration was as i.v. (36 to 37 units/sq m); the drug concentration in the tumors sociated with significant neurological deficit and parenchyma! ranged from 0 to 1.5 Â¿Â¿g/goftissue (14). In the present studies, damage. c/s-Platinum i.v. had no effect on the BBB and was i.e. doses of bteomycin between 3 and 12.5 units/sq m were without subsequent neuropathological findings. It is of interest studied. Unfortunately, even at 3 units/sq m, 2 of 3 dogs had that a clinical counterpart of these observations was recently significant neurological complications. Corresponding pharma reported by Stewart ef al. (39), where neurotoxicity occurred cological studies confirmed the fact that bleomycin did not cross with the i.e. infusion of c/s-platinum into patients with recurrent the BBB in normal animals, but when given i.e. after barrier malignant intracerebral tumors. Such clinical observations (1, 4, modification the levels of drug in the brain (Table 4) approximated 6, 12, 16) plus the evidence presented in the current studies those in liver and kidney. Systemic tissue levels did not decrease suggest that i.a. c/s-platinum after BBB modification has at best after BBB modification since the brain of a 250-g rat weighs only a limited role in the treatment of primary or metastatic brain 1 g and represents only a very small portion of total mass. From tumors, i.e. c/s-Platinum in the absence of osmotic BBB opening the present studies and previously reported data (14), it appears appears to be better tolerated clinically than in our canine model that any enhanced drug levels that can be achieved by i.e. drug (28). administration and/or barrier modification may be associated

Tabte6 Effect of BBB modification on [ÃŒ4C]-S-FUconcentration in brain of the rat achieved by i.e. administration of drug Tissue 5-FU(xg/g)

5-FUdose (mg/kg dis i.e.)"0.9% ruption00Contralateralhemisphere0.13 hemisphere0.49 (jig/ml)0.61 NaCIsolution 1 1BBB 0.12Ipsilateral 0.59Liver6.59 5.06Kidney12.98 16.20Muscle0.250.32Serum 0.66

0.125 Â±0.005Â° 0.54 Â±0.05 5.83 Â± 0.77 14.59 Â± 1.61 0.283 Â±0.035 0.64 Â±0.03

Mannitol 1 2+ 0.10 1.70 7.52 5.96 0.29 0.63 1 3+ 0.12 1.71 12.80 7.70 0.31 0.69 1 3+ 0.12 1.13 12.30 8.57 0.26 1.15

0.113 Â±0.007 1.51 Â±0.19 10.87 Â± 1.68 7.41 Â± 0.77 0.29 Â±0.01 0.82 Â±0.16

0.9% NaCIsolution 15 1.63 4.55 70.30 234.30 4.20 13.10 15 1.34 4.55 132.60 180.60 3.95 8.30

1.49 Â±0.15 4.55 Â±0 101.45Â±31.24 207.45 Â±26.93 4.08 Â±0.13 10.70 Â±2.41

Mannitol 15 3+ 1.82 43.50 103.70 140.20 5.70 12.00 15 3+ 2.93 31.30 130.70 143.40 4.30 14.50

2.38 Â±0.56 37.40 Â±6.12 117.20 Â±13.54 141.80 Â± 1.60 5.00 Â±0.70 13.25Â±1.25 'Tracer 5-FU (10 /iCi) plus the indicated dose of 5-FU was administered i.e. over 15 min beginning 5 min after i.e. 0.9% NaCI solution or mannitol.Animals were sacrificed 45 min after 0.9% NaCIsolution or mannitol infusion. "Mean Â±S.E.

Table? CNSeffects of i.v. cyclophosphantideprior to BBB modification in the canine Cyclophosphamide,20 mg/kg, was infused i.v. over 5 min, 10 to 15 min prior to BBB modification.Becauseof the minimaltoxicity in experimental(BBS-modified) animals, experiments with control animals (cydophosphamide infusion only) were not done. of sacri fice after BBB tion*: deficitsSlightNeurological findingsNoneNone disruption32 tendency to days15 left for 48 hr Seizures at 15 days days None None 35 days None None 54 days None None 35 days Never regainedcon Left cerebral infarction; 7hr38 sciousnessNonePathologicalsecondary hemor rhages in the pons and medulla NoneTime daysBBBdisrup "The success of BBB modification was determined by the presence or absence of Evans blue staining of brain at sacrifice.

NOVEMBER 1983 5283

Tablea Effect of BBB modification on Â¡"CJcyclophosphamideconcentrationin brain of the rat achieved by i.v. administrationof drug Tracer cydophosphairtde (10 /.Ci) plus 20 mg/kg was administered i.v. over 2 min, 5 min prior to i.e. 0.9% NaCI solution or mannitol. Animals were sacrificed 45 min after 0.9% NaCI solution or mannitol infusion. Tissue cyciophosphamide (^g/g) Contralat- eral hemi- Ipsilateral Serum (/ig/ BBB disruption sphere hemisphere Liver Kidney Muscle ml) 0.9% NaCI solution i.e. 0 13.3 12.6 25.0 30.9 19.3 20.0 0 13.9 13.1 23.4 27.5 18.3 18.5 0 12.6 13.0 24.8 30.5 20.2 13.2 Â±0.4" 12.9 Â±0.2 24.4 Â±0.5 29.6 + 1.1 18.8 Â±0.5 19.6 Â±0.5

Mannitoli.e.3+3+3+8.612.811.611.0Â±1.39.614.614.412.9

Â±1.622.528.127.726.1Â±1.819.131.234.628.3Â±4.717.218.017.6Â±0.413.818.818.317.0Â±1.6 3Mean Â±SE. with significant neurotoxicity. ued to proliferate. Finally, it is still not clear what the metabolic Previous reports have suggested that 5-FU can cross the BBB fate, distribution, and biological significance are of the various (2, 7, 18), although probably only to a limited degree (15). metabolites of cyciophosphamide (42). Even so, based on these Certainly, the occurrence of cerebellar degeneration has provided animal studies, cyciophosphamide has been given in association support for such a circumstance (17,19, 38). Unlike most other with barrier modification in 30 brain tumor patients without chemotherapeutic drugs, 91% of Â¡sotopically labeled 5-FU is evidence of neurotoxicity. cleared from the vascular compartment very rapidly, within 5 min The present studies suggest that the limited degree of neuro (2). Such a clearance rate supports the validity of comparison toxicity resulting from the use of some chemotherapeutic agents studies between brain and other organs rather than serum. In relates to the presence of an intact blood-brain barrier. When the present studies, 30 min after the completion of 5-FU admin this barrier is breached, several chemotherapeutic agents have istration, liver and kidney levels were 10 to 20 times the serum significant neurotoxicity. The marked neurotoxicity noted with levels. When barrier modification was followed by the i.e. admin even very low doses of Adriamycin (32) and c/s-platinum makes istration of 5-FU (1 mg/kg), the ipsilateral brain level averaged it unlikely that increased delivery of these agents when admin only 17% of the liver and kidney levels. Tissue concentrations of istered in association with BBB opening will have an important 5-FU at this dose were approximately 1.5 /Â¿gof 5-FU per g of role in the therapy of tumors in the CNS. The toxicity of bleomycin brain which was only 1% of the liver or the kidney content found and 5-FU may also preclude their use in association with BBB after a conventional clinical dose (15 mg/kg) of 5-FU. Neverthe opening. Cyciophosphamide, like methotrexate, appears to be less, at doses of 5 to 15 mg/kg, neurotoxicity was seen. These reasonably tolerated and provides the potential for the develop observations of neurotoxicity and tissue concentration suggest ment of a multiagent therapy program in association with BBB that 5-FU after BBB opening may have a limited role in the modification. treatment of CNS tumors. Previous studies have indicated that the neurotoxicity of Â¡.a. The capability of cyciophosphamide to enter the CNS is not 1,3-bis(2-chloroethyl)-1-nitrosourea is greater in the canine than dear. It has been reported to have both limited entry potential in primates or humans (8, 34), and initial clinical studies of i.a. (7) and effective entry (9,11, 40, 42). Egorin ef al. (9) found that c/s-platinum at 2 to 3 times the maximum dose reported here the peak cerebrospinal fluid:plasma intact drug ratio as measured (39) indicate less severe neurotoxicity than encountered in the by gas chromatography was 0.5, although the cerebrospinal dog. Thus, the i.a. administration of the drugs reported here in fluid:plasma activity ratio relative to alkylation was only 0.2. Graul association with osmotic BBB modification may be better toler ef al. (11), used [3H]cyclophosphamide and showed that the drug ated in primates or humans than in the canine. The internal penetrated the BBB fairly well. In the present studies, we were carotid artery of the dog is, on a relative basis, much smaller also able to show that radiolabeled cyciophosphamide and its than in the human or in primates (34). metabolites appeared to enter the brain parenchyma with tissue These observations provide evidence that our therapeutic levels that were 40 to 50% of those found in liver and kidney. armamentarium for the management of primary and secondary The drug was given i.v. to permit hepatic activation (42). The tumors in the CNS cannot simply be expanded by virtue of our concentration in brain tissue of i.v. radiolabeled cyciophospha present ability to deliver drug across the BBB. Parenthetically, mide did not appreciably increase when followed with barrier the toxicity described in the present studies further suggests modification. Nevertheless, evidence from animal tumor models that with many chemotherapeutic drugs a significant barrier must suggests that the cytotoxic metabolites of cyciophosphamide exist, else severe neurotoxicity would be more commonly en enter the CNS less well. For instance, Thomas ef al. (41) showed countered in the clinical arena. Hopefully, the present technology that cyciophosphamide given to animals bearing intracerebral which has provided the capability of increased drug delivery to L1210 leukemia destroyed the leukemic cells in the liver, spleen, the brain will provide the opportunity to further explore and bone marrow, and dura, while leukemic cells in the CNS contin identify chemotherapeutic or antitumor agents (i.e., tumor-spe-

5284 CANCER RESEARCH VOL. 43

Biopharm., 4: 499-519,1976. ciÃ±emonoclonal antibodies) capable of significant cell kill with 19. Lynch, H. T., Droszcz. C. P., Albano, W. A., and Lynch, J. F. "Organic brain little or no related neurotoxicity. To date, in clinical studies of syndrome" secondary to 5-fluorouracil toxicity. Dis. Colon Rectum, 24: 130- chemotherapy administration in association with osmotic BBB 131,1981. opening, tumor regression has been documented in patients with 20. Merker, P. C., Lewis, M. R., Walker, M. D., and Richardson, E. P., Jr. Neurotoxicity of Adriamycin (doxorubicin) perfused through the cerebrospinal microglioma, ependymoblastoma, medullablastoma, glioblas- fluid spaces of the rhesus monkey. Toxico!. Appi. Pharmacoi., 44: 191-205, toma, and metastatic carcinoma of the breast. After 245 proce 1978. 21. Neuwelt, E. A., Bigner, D., Frenkel, E. P., Bamett, P., and Page), M. The dures in 42 patients, the limited toxicity encountered and the effects of osmotic modification of the blood-brain barrier and adrenal steroid responses observed are very encouraging (24, 29). administration on methotrexate delivery of gliomas in rats: The blood-brain barrier is a factor. Proc. Nati. Acad. Sei. U. S. A., 79:4420-4423,1982. 22. Neuwelt, E. A., Diehl, J., Frenkel, E. P., and Bamett, P. Osmotic blood-brain barrier disruption in posterior fossa of the dog. J. Neurosurg., 55: 742-748, ACKNOWLEDGMENTS 1981. 23. Neuwelt, E. A., and Frenkel, E. Is there a therapeutic rote for blood-brain The radtoimmunoassays of bteomyÃ¤n levels were kindly done courtesy of Dr. barrier disruption. Ann. Intern. Med., 93:147-149,1980. Robert Smyth, Bristol Laboratories, Syracuse, N. Y. 24. Neuwelt, E. A., Frenkel, E., Diehl. J., Hid, S.. and Balaban. E. Successful treatment of primary central nervous system lymphomas (microgliomas) with chemotherapy after osmotic blood-brain barrier opening. Neurosurgery, 72: REFERENCES 662-671,1983. 25. Neuwelt, E. A., Frenkel, E. P., Diehl, J. T., Vu, L. H., and Hill, S. A. Reversible 1. Berman. I. J., and Mann, M. P. Seizures and transient cortical blindness osmotic blood-brain disruption in humans: implications for the chemotherapy associated with c/s-platinum (11) diamminedichloride (PDD) therapy in a thirty- of malignant brain tumors. Neurosurgery, 7: 44-52,1980. year-old man. Cancer (Phila.), 45: 764-766,1980. 26. Neuwelt, E. A., Frenkel, E. P., Diehl, J. T., Vu, L. H., and Hill, S. A. Monitoring 2. Bourke, R. S., West, C. R., Chheda, G., and Tower, D. B. Kinetics of entry of methotrexate delivery in patients with malignant brain tumors after osmotic and distribution of 5-fluorouracil in cerebrospinal fluid and brain following blood-brain barrier disruption. Ann. Intern. Med., 94: 449-454,1981. intravenous injection in a primate. Cancer Res., 33:1735-1746,1973. 27. Neuwelt, E. A., Frenkel, E. P., Rapoport, S. I., and Bamett, P. A. Effect of 3. Broughton, A., and Strong, J. E. Radioimmunoassay of bteomycin. Cancer osmotic blood-brain barrier disruption on methotrexate pharmacokinetics in Res., 36: 1418-1421,1976. the dog. Neurosurgery, 7:36-43,1980. 4. Calvo, D. B., Patt, Y. Z., WaBace, S., Chuang, V. P., Benjamin, R. S., Pritchard, 28. Neuwelt, E. A., Glasberg, M., Frenkel, E., and Barnen, P. Neurotoxicity of J. 0., Hersh, E. M., Bodey, G. P., and Mavligit, G. M. Phase 1-11 Trial of chemotherapeutic agents after blood-brain barrier modification: neuropatho- percutaneous intra-arterial c/s-diamtninedichloroplatinum (11) for regionally logic studies. Ann. Neurol., 14: 316-324,1983. confined malignancy. Cancer (Phila.), 45:1278-1283,1980. 29. Neuwelt, E. A., Howteson, J., Specht, D., Haynes, J., and Hill, S. Comparison 5. Chabner, B. A., Myers, C. E., Coteman, C. N., and Johns, D. G. The clinical of enhanced computerized tomography and radtonudide brain scan to monitor pharmacology of antineoplastic agents (first of two parts). N. Engl. J. Med., osmotic blood-brain barrier disruption. Am. J. Neuroradiol.. 4: 907-913,1983. 292:1107-1113,1975. 30. Neuwelt, E. A., Maravilla, K. R., Frenkel, E., Bamett, P.. Hill, S., and Moore, 6. Clark, A. W., Parhad, I. M., Griffin, J. W., and Price, D. L. Neurotoxicity of c/s- R. Use of enhanced computerized tomography to evaluate osmotic blood- platinum: pathology of the central and peripheral nervous systems. Neurology, brain barrier disruption. Neurosurgery, 6: 49-56, 198oO. 30: 429,1980. 31. Neuwelt, E. A., Maravilla, K. R., Frenkel, E. P., Rapoport, S. I., HiÂ»,S.A., and 7. Cline, M. J., and Haskeil, L. M. Cancer Chemotherapy, pp. 32-33. Philadelphia: Bamett, P. A. Osmotic blood-brain barrier disruption: computerized tomo- W. B. Saunders Co., 1980. graphic monitoring of chemotherapeutic agent delivery. J. Clin. Invest., 64: 8. DeWys, W. D., and Fowler, E. H. Report of vasculitis and blindness after 684-688,1979. mtracarotid injection of 1,3-bis(2-chloroethyl)-1-nitrosourea (BCNU; NSC- 32. Neuwelt, E. A., Pagel, M., Bamett. P., Glasberg, M., and Frenkel, E. P. 409962) in dogs. Cancer Chemother. Rep., 57: 33-40,1972. Pharmacology and toxicity of mtracarotid Adriamycin administration following 9. Egorin, M. J., Kaplan, R. S., Salcman. M., Aisner, J., Colvm. M., Wiernik, P. osmotic blood-brain barrier modification. Cancer Res., 41:4466-4470,1981. H., and Bachur. N. R. Cyclophosphamide plasma and cerebrospinal fluid 33. Ormo, K., Fredericks, W. R., and Rapoport, S. I. Osmotic opening of the blood- kinetics with and without dimethyl sulfoxide. Clin. Pharmacoi. Ther., 32: 122- brain barrier to methotrexate in the rat. Surg. Neurol., 12:323-328,1979. 128,1982. 34. Omojola. M. F., Fox, A. J., Auer. R. N., and Vinuela. F. V. Hemorrhagic 10. Evans, H. E., and Christiansen, G. C. (eds.). Miller's Anatomy of the Dog, Ed. encephalitis produced by selective non-occlusive mtracarotid BCNU injection 2, pp. 676-679. Philadelphia, W. B. Saunders Co., 1979. in dogs. J. Neurosurg., 57: 791-796,1982. 11. Graul, E. H., SchaumlÃ¶ffel, E., Hundehagen. H., Wilmanns, H., and Simon, G. 35. Pappius, H. M., Savake, H. E., Feichi, C., Rapoport, S. I., and Sokolof. L. Metabolism of radioactive cydophosphamide. Cancer (Phila.), 20: 896-899, Osmotic opening of the blood-brain barrier and local glucose utilization. Ann. 1967. Neurol., 5:211-219,1979. 12. Hadley, D., and Herr, H. W. Peripheral neuropathy associated with c/s-dichlo- 36. Rapoport, S. I. Blood-brain Barrier in Physiology and Medicine, pp. 229-232. rodiammineplatinum (11) treatment. Cancer (Phila.), 44:2026-2028,1979. New York: Raven Press, 1976. 13. Hasegawa, H., Allen. J. C., Menta. B. M., Shapiro, W. R., and Posner, J. B. 37. Rapoport, S. I., Fredericks, W. R., Ohno, K., and Pettkjrew, K. D. Quantitative Enhancement of CNS penetration of methotrexate by hyperosmolar intracar- aspects of reversible osmotic opening of the blood-brain barrier. Am. J. otid mannitol or carcinomatous meningitis. Neurology, 29: 1280-1286,1979. Physiol., 235: 421-431,1980. 14. Hayakawa, T., Ushio, Y., Morimoto, K., Hasegawa, H., Mogami, H., and 38. Rtehl, J. L., and Brown, W. J. Acute cerebellar syndrome secondary to 5-FU Horibaba, K. Uptake of bteomycin by human brain tumors. J. Neural. Neuro- therapy. Neurology, 14: 961-967.1964. surg. Psychiat., 39: 341-349,1976. 39. Stewart, D. J., Wallace, S., Feun. L, Leavens, M., Young, S. E., Handel, S., 15. Hombeck, C. L., Floyd, R. A., ByfiekJ, J. E., Griffiths, J. C., Frankel, S., and Mavhgit. G., and Benjamin, R. S. A phase 1 study of mtracarotid artery infusion Sharp, T. R. Cerebrospinal fluid versus serum concentrations of 5-FU, allopu- of c/s-diamminedichloroplatinum in patients with recurrent malignant intracer- rinol, and oxypurinol during treatment of metastatic brain cancer with 5-FU ebral tumors. Cancer Res., 42: 2059-2062,1982. infusion, altopurinol, and radiation. Cancer Treat. Rep., 66: 571-573,1982. 40. Talha, M. R. Z., Rogers, H. J., and Trounce, J. R. Distribution and pharmaco 16. Kaplan, R., and Wiernik. P. H. Neurotoxicity of antineoplastic drugs. Semin. kinetics of cydophosphamide in the rat. Br. J. Cancer., 41:140-143,1980. Oncol., 9. 103-109,1982. 41. Thomas, L. B., Chirigos, M. A., Humphreys, S. R., and Goldin, A. Pathology 17. Koenig. H., and Patel, A. Biochemical basis for fluorouracil neurotoxicity. Arch. of the spread of L1210 leukemia in the central nervous system of mice and Neurol., 23:155-160,1970. effect of treatment with cydophosphamide. J. Nati. Cancer Inst., 28: 1355- 18. Levin, V. A., Landahl, H. D., and Freeman-Dove, M. A. The application of brain 1368,1962. capillary permeability coefficient measurements to pathological conditions and 42. Torkelson. A. R., LaBudde, J. A., and Weikel, J. G., Jr. The metabolic fate of the selection of agents which cross the blood-brain barrier. J. Pharmacokinet. cydophosphamide. Drug Metab. Rev., 3:131-165,1974.

NOVEMBER 1983 5285

Downloaded from cancerres.aacrjournals.org on September 28, 2021. © 1983 American Association for Cancer Research. Pharmacology and Neurotoxicity of cis -Diamminedichloroplatinum, Bleomycin, 5-Fluorouracil, and Cyclophosphamide Administration following Osmotic Blood-Brain Barrier Modification

Edward A. Neuwelt, Peggy A. Barnett, Mark Glasberg, et al.

Cancer Res 1983;43:5278-5285.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/43/11/5278

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/43/11/5278. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.