[CANCER RESEARCH 52. 6168-6174, November 15, 1992)

Intrathecal 4-Hydroperoxycyclophosphamide: Neurotoxicity, Cerebrospinal Fluid

Pharmacokinetics, and Antitumor Activity in a Rabbit Model of VX2Leptomeningeal Carcinomatosisl

Peter C. Phillips,2 Thuy T. Than, Linda C. Cork, John Hilton, Benjamin S. Carson, O. Michael Colvin, and

Louise B. GrochowDepartments of Neurology [P. C. P.], Neurosurgery [T. T. T., B. S. C.], Oncology [J. H., O. M. C., L. B. C.], Pathology and Comparative Medicine [L. C. C.],The Johns Hopkins University School of Medicine, Baltimore, Maryland 21205

ABSTRACT

Dissemination of tumor to the leptomeninges and cerebrospinal fluidrepresents a common pattern of metastasis for many cancers; however,few chemotherapeutic agents are available for intrathecal (i.t.) use andtreatment results are often poor. We studied the neurotoxicity and phar-macokinetics of i.t. 4-hydroperoxycyclophosphamide (4-HC) in the rabbit and the activity of i.t. 4-HC in a VX2 rabbit model of leptomeningealcarcinomatosis to evaluate the potential use of 4-HC in the treatment ofleptomeningeal tumors. Toxicity studies examined 4-HC doses ranging

from 0.5 to 6.0 »imoladministered by intraventricular injection weeklyfor 4 to 8 weeks. Clinical or histológica! neurotoxicity was not observedin rabbits treated with <1.0 fimol 4-HC for 4 weeks. Clinical toxicity,

characterized by lethargy, weight loss, seizures, or death, was apparentat doses >2.0 MHIO!.Vasculitis of superficial arteries was observed inrabbits treated with >1.0 /imol 4-HC. In cerebrospinal fluid pharmaco-kinetic studies, the mean drug half-life after intraventricular or in-

tralumbar administration was 24.3 and 18.2 min. Regional inequities indrug exposure were apparent as area under the clearance curve valuesfor cerebrospinal fluid distant from the injection site were lower thanthose of proximate sites (P < 0.001). Weekly intraventricular treatmentof VX2 leptomeningeal tumor-bearing rabbits with 0.5 or 1.0 ftmol of4-HC resulted in an increased life span of 22.5 and 35%, respectively.These results indicate that i.t. 4-HC, at doses lower than those produc

ing neurotoxicity in the rabbit, is effective treatment for VX2 leptomeningeal carcinomatosis.

INTRODUCTIONThe leptomeninges and CSF3 are common sites of metastasis

for leukemias (1,2) and solid tumors (3). Despite the recognized clinical importance of this problem, current treatmentresults for leptomeningeal tumor remain poor. Effective therapy has been limited primarily by three factors: (a) inadequatepenetration of the CSF by conventional systemic chemotherapeutic agents; (b) unacceptable neurotoxicity associated withthe intrathecal administration of many chemotherapeuticdrugs; and (c) limited clinical activity against many solid tumorsfor drugs commonly available for i.t. therapy, including meth-otrexate, l-/3-o-arabinofuranosylcytosine, and thiotepa (4, 5).

Cyclophosphamide, an effective agent in the treatment ofsolid and hematological malignancies, is a prodrug; it requiresactivation by hepatic microsomal enzymes, which precludes intrathecal use. This problem may be circumvented by use of apreactivated Cyclophosphamide analogue such as 4-HC (6).4-HC is reduced rapidly to 4-hydroxycyclophosphamide and

Received 2/7/91; accepted 9/10/92.The costs of publication of this article were defrayed in part by the payment of

page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1This work was supported by NIH Grants NS01279 and CA39914.2 To whom requests for reprints should be addressed, at The Children's Hospital

of Philadelphia, 34th and Civic Center Boulevard, Philadelphia, PA 19104.3 The abbreviations used are: CSF, cerebrospinal fluid; 4-HC, 4-hydroperoxy

cyclophosphamide; ¡.t.,immillerai; IVent, intraventricular; AUC, area under theclearance curve; HPLC, high-performance liquid chromatography.

then converts spontaneously to the cytotoxic alkylating moiety,phosphoramide mustard (7). 4-HC has shown activity in vitroagainst a diverse group of human tumors, including medullo-blastoma (8), LI210 murine leukemia (9), Burkitt's lymphoma

(10), and breast carcinoma (11). In a recent study, Fuchs et al.(12) reported significant therapeutic activity for i.t. 4-HC inhuman rhabdomyosarcoma (TE-671) and malignant glioma(D-54 MG) xenograft models of leptomeningeal tumor in thenude rat.

To further evaluate the clinical potential of i.t. 4-HC in thetreatment of leptomeningeal tumors, we examined the behavioral and histológica! characteristics of neurotoxicity fromIVent 4-HC and compared the regional pharmacokinetics ofIVent versus intralumbar 4-HC administration in rabbits. Weused VX2 rabbit carcinoma, a well-characterized and reproducible model of leptomeningeal carcinomatosis in the rabbit (13),to further define the in vivo activity of i.t. 4-HC.

MATERIALS AND METHODS

Drug Synthesis

4-HC (Mr 229,000) was synthesized by the ozonization of Cyclophosphamide as described by Hohorst et al. (14). Purity of the samples wasassessed by HPLC (Waters Associates, Milford, MA). Samples showing greater than 90% purity were washed in ether, dried by evaporationunder a nitrogen stream, and stored at -20°C.

Drug Assay

CSF and plasma samples were placed immediately in an acidifiedm-aminophenol-hydroxylamine solution and stored in the dark at 4°Cfor no more than 7 days until assayed. Samples were heated to 100°C

for 15 min and centrifuged at 3000 rpm for 5 min, and the supernatantwas stored at 4°Cuntil assayed for 4-HC concentration. The 4-HC

assay was based on the method of Alarcon (15) and has been describedpreviously (16). Briefly, methyl vinyl ketone was added to the acidifiedw-aminophenol-hydroxylamine solution. This produces a fluorescent4-methyl-7-hydroxyquinoline which is detected as an internal standard.Acrolein, released from 4-HC and 4-hydroxycyclophosphamide byacidification, reacts with m-aminophenol to produce a fluorescent 7-hydroxyquinoline. The fluorescent products were identified by HPLCequipped with a fluorescent detector (330 nm excitation, 418 nm emission). The limit of sensitivity of the assay was 0.1 MM-

Rabbits

Intraventricular Catheter. Male New Zealand White (nzw) rabbits(Bunnyville, Littlestown, PA) were used for all studies. General anesthesia was induced in 3.0- to 3.6-kg rabbits with i.v. injections ofacepromazine maléate(10 mg) and ketamine hydrochloride (150 mg),followed by i.v. sodium thiamylal (Bio-Ceutic, St. Joseph, MO) toeffect. Under sterile conditions, a 20-gauge Silastic catheter (DowCorning, Midland, MI) was inserted into the right lateral ventriclethrough a parietal skull burr hole according to previously describedtechniques (17). The catheter was then connected to an injection reservoir which was buried under the skin overlying the forehead. The

6168

on May 12, 2020. © 1992 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

i.t. 4-HC IN LEPTOMENINGEAL CARCINOMATOSIS

catheter position and patency were evaluated by fluoroscopy after theinjection of metrizamide into the reservoir. At least 4 days were allowedfor wound healing prior to the injection of drug or tumor. No neurological abnormalities were observed in any rabbits after catheterimplantation.

Cisterna Magna and Lumbar Catheter Placement. Rabbits weighing4.5 to 5.5 kg were anesthetized as described above and placed in a proneposition with the back slightly flexed. The distance from the cisternamagna to the seventh lumbar vertebra (L7) was measured and a sterilepolyethylene catheter (PE 10; Intramedic, Parsippany, NJ) was markedaccordingly. The lumbosacral area was then shaved, a midline skinincision (1 cm long) was made over the L7 and Si vertebrae, and alumbar puncture was made in the L7-S! interspace with a 20-gauge,1.5-inch spinal needle (Becton-Dickinson, Rutherford, NJ). Once CSFwas aspirated, the PE 10 catheter was threaded through the lumen ofthe spinal needle into the subarachnoid space and then advanced to themidlumbar region or the cisterna magna for subsequent drug injectionor CSF sampling (see "Pharmacokinetic Studies"). After correct place

ment of the subarachnoid catheter tip, the spinal needle was removed.Since we did not use the same catheter for drug injection and CSFsampling. Group 4 rabbits in the pharmacokinetic study had two lumbar catheters placed at the L6-L7 interspace by the procedures describedabove. The position of the catheter tips was verified at autopsy.

Neurotoxicity Studies

Rabbits (3-3.6 kg) with chronic indwelling lateral ventricular cathe-

the data were best fit by a one compartment model with first orderdelivery of drug to the sampling site, first order elimination, and a lagtime (when present) (18). The following kinetic parameters were estimated by use of nonlinear least squares analysis by using PCNONLIN(Statistical Consultants, Lexington, KY): the rate constant for drugdelivery (kj); the elimination rate constant (A); the volume of distribution (V)\ and the delivery lag time (T). The area under the clearancecurve (AUQ and clearance rate (Cl) were calculated by using the parameter estimates.

VX2 Leptomeningeal Tumor Model

VX2 rabbit carcinoma was obtained from the American Type Culture Collection (Rockville, MD) and maintained as a s.c. tumor inrabbits by passage every 2 to 3 weeks. For the VX2 leptomeningealmodel, approximately 0.5 g of freshly harvested tumor was coarselysectioned and then incubated for 18 h (5% CO2, 37°C)in 10 M!of

dissociation medium consisting of 50 units/ml collagenase (Worthing-ton, Freehold, NJ) and 0.005% DNase (Sigma Chemical Co., St. Louis,MO) in RPMI 1640 containing 10% fetal bovine serum and 60 mivi(pH7.3) 4-(2-hydroxyethyl)-l-piperazineethanesulfonic acid. The tumorwas further dissociated by repeated pipeting, washed twice with RPMI1640/10% fetal bovine serum, and then centrifuged and resuspended inElliott's B solution. Viable cells, determined by trypan blue exclusion,

were counted with a hemocytometer.Tumor cells, 3.0 x IO3, were suspended in 100 M>of Elliott's B

solution and injected into the s.c. reservoir of lateral ventricular cath-ters were treated with IVent injections of 4-HC at 0.5-, 1.0-, 2.0-, 3.0-, eters that had been implanted in 3.0- to 3.6-kg rabbits at least 5 days4.0-, 5.0-, or 6.0-Mmol doses. 4-HC was diluted in a 0.9% NaCl solutionimmediately before injection and 100 ^1 of drug were followed by200 M'of Elliott's B CSF solution over 1 min. Control rabbits received100 n\ of a 0.9% NaCl solution followed by 200 ß\of Elliott's B

solution. IVent injections were repeated weekly for 4 or 8 weeks, duringwhich time the rabbits were observed closely for weight loss, lethargy,seizures, ataxia, paralysis, or other signs of neurotoxicity. Rabbitsshowing evidence of severe neurotoxicity (e.g., paralysis or persistentseizures) were sacrificed before the completion of the fourth or eighthweek; all other animals were sacrificed 1 week after the fourth or eighthinjection. The brain and spinal cord were removed and fixed in 10%buffered formalin. Paraffin-embedded sections of the brain and cordwere stained with Luxol fast blue or hematoxylin and eosin for histológica!examination by light microscopy.

CSF Pharmacokinetic Studies

The disposition of 4-HC in the CSF after IVent or intralumbaradministration was studied in four groups of rabbits. Group 1 (n = 7),4-HC administered through a lateral ventricular catheter, CSF sampledthrough a cisterna magna catheter; Group 2 (n = 5), 4-HC administeredthrough a lateral ventricular catheter, CSF sampled through a lumbarcatheter; Group 3 (n = 6), 4-HC administered through a lumbar catheter, CSF sampled through a cisterna magna catheter; Group 4 (n = 6),4-HC administered through a lumbar catheter, CSF sampled through aseparate lumbar catheter (see "Cisterna Magna and Lumbar CatheterPlacement," above). All pharmacokinetic studies were performed with

the rabbits under general anesthesia, and all animals were hydratai withan i.v. solution of 0.9% NaCl throughout the study. The 4-HC dose forall rabbits was 2.0 Mmol diluted in 100 M'of 0.9% NaCl solution immediately prior to injection for all rabbits, followed by a 200-^1 flushwith Elliott's B solution. A small sample of the 4-HC to be injected was

reserved for drug assay. After IVent 4-HC injection (Groups 1 and 2),CSF samples (10 to 50 /il) were obtained at 2, 10, 30, 60, 90, and 120min. After intralumbar injection (Groups 3 and 4), CSF samples wereobtained 2, 10, 30, 45, 60, 90, and 120 min after drug injection. The4-HC concentrations in CSF samples were measured as describedabove.

Pharmacokinetic Calculations

4-HC disposition curves for each experiment were evaluated for suitable pharmacokinetic models by visual inspection. For all experiments.

earlier. The reservoir was then flushed with 200 n\ of Elliott's B solution

through the same needle. Previous studies in our laboratory have shownthat signs of leptomeningeal carcinomatosis (e.g., anorexia, lethargy,cranial nerve palsies, or paralysis) are apparent within 20 days afterIVent injection of 3000 viable VX2 cells (19). Animals were sacrificedwhen severe neurological deficits developed (e.g., ataxia, seizures, orparaplegia), inasmuch as previous studies showed that these signs werefollowed invariably by death within 24 h.

Survival Studies

Rabbits with leptomeningeal VX2 were randomly assigned to 4-HCtreatment or control groups. At 24 h after VX2 injection, rabbits weretreated with 0.5 /¿mol(0.11 mg; n = 18) or 1.0 Mmol(0.23 mg; n = 13)of 4-HC in 100 M'of 0.9% NaCl solution administered by IVent injection. Control animals received 100 v\ of 0.9% NaCl solution. After allinjections, the intraventricular reservoirs were flushed with 200 n\ ofElliott's B solution. 4-HC or NaCl solution injections were adminis

tered weekly for 4 weeks. Rabbits were observed daily for signs ofleptomeningeal carcinomatosis. Upon the appearance of severe neurological deficits (see above), the rabbits were sacrificed by injection ofi.m. acepromazine and ketamine followed by i.v. injection of euthanasiasolution T-61 (Taylor Pharmacal, Decatur, IL). Rabbits with no signsof leptomeningeal carcinomatosis were sacrificed on day 60 after thefirst 4-HC treatment and were considered long-term survivors. The dayof death or sacrifice was recorded as the study end point. For all rabbits,the brain and spinal cord were removed and fixed in 10% bufferedformalin for at least 10 days. Paraffin sections of the brain and spinalcord were stained with Luxol fast blue or hematoxylin and eosin andexamined by light microscopy. Differences in survival between treatment and control groups were analyzed by Student's t test.

RESULTS

Neurotoxicity Studies. The maximum tolerated dose of4-HC in the rabbit as defined by the absence of behavioral orclinical signs of neurotoxicity was 1.0 ¿¿molweekly for 4 weeks(Table 1). However, in all rabbits treated at this dose, mild tomoderate arteritis was apparent on histológica! examination.No behavioral or histológica! abnormalities were observed inrabbits treated with 0.5 /umol 4-HC for 4 weeks. All rabbitstreated with 2 Mmol of 4-HC weekly developed clinical signs of

6169

on May 12, 2020. © 1992 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

i.t. 4-HC IN LEPTOMENINGEAL CARCINOMATOSIS

Table 1 Intralhecal 4-hydroperoxycyclophosphamide neuroloxicity

4-HC dose(mmol)0.5

(n = 6)1.0 (n = 6)2.0 (n = 4)3.0 (n = 2)4.0 (n = 1)Clinical

toxicitymedian onset

(days)NS"

NS14.017.59.0Observed

neurotoxicity(rabbits)0

6421

" NS, not seen.

toxicity: all had progressive weight loss; one had seizures; andanother had lower extremity weakness. One rabbit died fromtoxicity 3 days after the third 4-HC injection. All rabbits treatedwith >2 /¿molof 4-HC weekly developed severe neurotoxicityand died before the fourth 4-HC dose.

On histológica! examination, a reactive gliosis that rangedfrom mild to severe was noted at the catheter implantation site

in all rabbits. The most consistent histológica! abnormality wasa dose-dependent vasculitis. Vasculitis was not observed inrabbits treated with 0.5 Mmol of 4-HC for 4 or 8 weeks. Inrabbits treated with 1.0 /¿molof 4-HC, the arterial lesions weremost prominent at the base and lateral surfaces of the medullaand were characterized by granulocytic and mononuclear infiltrates, mineralization, and fibrosis (Fig. \,A and B). At higherdoses, the severity and distribution of the vascular injury weregreater. Vessels on the ventral aspect of the cerebellum, pons,medulla, and spinal cord were surrounded by a dense sheetof reactive connective tissue resulting in the entrapment of exiting cranial nerves. Similar vascular changes were observedin the spinal cords of these rabbits. In all rabbits treated withdoses >2.0 //mol, the superficial blood vessels in the cerebellumand medulla were severely compromised by transmural andcircumferential fibrinoid necrosis accompanied by small foci ofhemorrhage.

. •¿�> - -*«.W , 1^ I -^W.

*•- ^ . X'

A •¿�«"^>.

SO/'

Fig. 1. Vascular and cerebellar injury in a rabbit treated with three weekly intraventricular 2.0-/¿moldoses of 4-HC. A, large vessel at the base of the brain containingmultiple areas of fibrinoid necrosis and infiltrations of inflammatory cells, x 400. B, small artery at the base of the brain with an intense arteritis characterized bykaryorrhexis. endothelial necrosis, and infiltration of a mixture of inflammatory cells, x 400. In C, the Purkinje cells in the cerebellum are degenerating and shrunken,and there is a mild gliotic response, x 160.

6170

on May 12, 2020. © 1992 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

i.t. 4-HC IN LEPTOMENINGEAL CARCINOMATOS1S

In addition to the vascular lesions, vacuolization of the chor-oid plexus was noted in the majority of the rabbits treated with>1.0 /¿molof 4-HC. At doses >2.0 ^mol, parenchyma! injurywas also observed, predominantly in brainstem and cerebellarregions that appeared to be secondary to the vascular lesions.Leukomalacia, vacuolization of white matter, focal inflammation, edema, and hemorrhage were noted in the dorsal medulla.In the cerebellum, reactive gliosis and foci of inflammation inthe molecular layer and Purkinje cell degeneration were found(Fig. 1C). Axonal degeneration, demyelination, and gliosis inthe subjacent subpial region of the cerebellum and pons werealso evident. Similar changes were observed in the spinal cord atdoses >2.0 jumolof 4-HC.

Histological examination of 7 of 24 rabbits treated with4-HC and 1 of 4 untreated controls revealed lesions consistentwith infection by Encephalitozoon cuniculi, a ubiquitous andclinically indolent protozoal infection in rabbits which causes acharacteristic granulomatous meningoencephalitis.

Pharmacokinetic Studies. For rabbits treated at the highesti.t. 4-HC dose (4 /imol), the concentration of 4-HC in plasmawas below the limits of HPLC detection (1 JIM).Plasma 4-HCvalues were not evaluated at lower i.t. 4-HC doses. Fig. 2 showsthe mean 4-HC concentrations in CSF after IVent or intralum-bar administration in Groups 1 to 4. Peak 4-HC concentrations

A3,000

1.000

oE loo_2_G"

5 »10

Cisterna magnan=7

Lumbarn=5

B3.000

1.000

300

O

S^tr

100

30

10

60 80

MINUTES

Lumbarn=6

Cisterna magnan=6

40 60 80

MINUTES100 120 140

Fig. 2. CSF concentration of 4-HC after intraventricular (A) or intralumbar (B)injection of 2.0 j<mol (0.45 mg). •¿�.mean values for CSF sampled from thecisterna magna; O, mean values for CSF sampled from the lumbar space; bars,SD.

were attained within 30 min of injection for all groups exceptGroup 2 (ventricular injection, lumbar sampling) in which thepeak value, attained at 45 min, was the lowest of all groups. InGroup 4 (lumbar injection, lumbar sampling), the peak 4-HCconcentration was highest and was attained within 2 min ofinjection. Table 2 summarizes the pharmacokinetic parametersfor all treatment groups.

Survival Studies. Treatment of rabbits with VX2 leptom-eningeal carcinomatosis with IVent 4-HC resulted in a dose-dependent increased life span compared to controls (Table 3).One cohort of rabbits did not have successful tumor implantation, evidenced by survival of all control and treated rabbits toDay 45, and was not included in the survival data analysis. Themedian survival for rabbits treated with 0.5 ^mol of 4-HC was24.5 days, representing a 22.5% increased life span over controls (P = 0.002). A longer median survival (27.0 days) wasfound in rabbits treated at the 1.0-Mmoldose, resulting in a 35%increase compared with controls (P = 0.0004). IVent 4-HCtreatment produced two prolonged survivors (>60 days) at the0.5-Mmoldose and one at the 1.0-Mmoldose. These prolongedsurvivals were not due to failure of tumor implantation, becausesmall foci of tumor were observed upon histological examination of the spinal cords. Although we cannot exclude the possibility of an error in the number of tumor cells injected, thesurvival data for control animals in this treatment cohort weresimilar to those for other control groups.

The types of clinical abnormalities observed in control and4-HC-treated groups were different, suggesting different sites oftreatment failure. Thirteen of 19 controls had signs that wereexclusively or predominantly due to intracranial tumor (e.g.,seizures, anorexia, focal cranial nerve abnormalities) and 6 hadparaplegia as the only abnormal neurological sign. By contrast,20 of 31 rabbits in the 4-HC-treated groups developed paraplegia as the only neurological abnormality. Histological examination at the time of death or sacrifice confirmed that thegreatest tumor burden in control animals was found in intracranial regions close to the site of tumor injection, the cerebralleptomeninges and ventricles. In 4-HC-treated rabbits, however, tumor burden was greatest in regions furthest from the siteof tumor and 4-HC injection, the thoracic and lumbar spinalcord (Fig. 3).

DISCUSSION

Leptomeningeal metastasis is an important cause of morbidity and mortality in cancer patients. Prior to the introduction ofcentral nervous system prophylactic therapy in childhood acutelymphoblastic leukemia, the incidence of meningeal métastasesexceeded 80% (2). For extraneural solid tumors, in which he-matogenous spread represents the most probable mechanism ofmeningeal métastases(20), the incidence of leptomeningeal tumor may reach 46% in melanoma (21), 11 to 18% in small celllung carcinoma (22), and 5% in breast carcinoma (23). Directextension via CSF pathways is responsible for metastatic tumorspread in primary brain tumors, where leptomeningeal métastases have been identified in 19% of childhood brain tumorpatients (24). Reported increases in the frequency of leptomeningeal metastasis (25) may be due to improved diagnostictests, greater surveillance, or longer patient survival (3).

Despite advances in the treatment of various primary tumors,the chemotherapeutic options for patients with leptomeningealmétastasesare limited and results are generally poor. Treatment failures after standard-dose systemic chemotherapy may

6171

on May 12, 2020. © 1992 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

i.I. 4-HC IN LEPTOMENINGEAL CARCINOMATOSIS

Table 2 Pharmacokinetic parameters of 4-HC in cisterna magna and lumbar subarachnoid space

Treatmentgroup01

(n = 7)2 (n = 5)3 (n = 6)4 (n = 6)Injection

siteVentricular

VentricularLumbarLumbarSampling

siteCisterna

magnaLumbarCisterna magnaLumbarPeak

4-HC(><mol/liter)500

62223

3684AUC(mmol/min/liter)18

811

100t'/2(min)21

272313Vi(ml)5.0

23.017.00.5Clearance(ml/min)0.17

0.530.220.03

" lntr.nliLV.il 4-HC dose of 2.0 »imolwas used for all treatment groups.

Table 3 Survival following intralhecal treatment for rabbits with VX2leptomeningeal carcinomatosis

IntrathecaltreatmentNaCI

(0.9%) (n = 19)4-HC (0.5 «mol)(n = 18)4-HC (1.0 «mol)(n = 13)Median

Survival(days)20.0

±0.924.5 ±2.6°27.0 ±1.6*Survivors

(>8wk)0

21

" P = 0.002, Student's t test.* P = 0.0004. Student's i test.

•¿�

Fig. 3. VX2 leptomeningeal tumor encasing and infiltrating spinal cord andnerve roots in a rabbit treated with 0.5 «mol4-HC weekly for 4 weeks, x 10.

be due, in part, to blood-CSF barrier restrictions resulting ininadequate drug concentrations in the CSF. This problem maybe circumvented by regional (i.e., intrathecal) drug administration, where higher drug concentrations can be achieved than byconventional systemic routes. Unfortunately, few chemothera-peutic drugs are available for intrathecal use because of neuro-toxicity. Intrathecal methotrexate is the most active drug forthe treatment of central nervous system leukemia (1, 26) andmay be useful in leptomeningeal carcinomatosis from breastcancer (27, 28); however, intrathecal methotrexate is associatedwith significant neurotoxicity (29-31). Intrathecal l-/3-o-ara-binofuranosylcytosine may be less neurotoxic than methotrexate but is ineffective against many solid tumors. Thiotepa maybe active against solid tumors (32), but the rapid afflux ofthiotepa from the CSF may limit its effectiveness (5).

Efforts to improve treatment results for patients with leptomeningeal tumor include the use of intraventricular (Ommaya)reservoirs (26, 33), the development of animal models necessaryfor the preclinical evaluation of intrathecal drug neurotoxicityand efficacy (12, 19, 34, 35), and the introduction of new anti-neoplastic agents for potential intrathecal use (36-38). Pharmacological studies have indicated advantages for intrathecal4-HC over i.v. cyclophosphamide for the treatment of leptomeningeal disease (39). Fuchs et al. (12) reported that 4-HC isactive against leptomeningeal rhabdomyosarcoma (TE-671)

and malignant glioma (D-54) human xenografts at doses that

do not cause neurotoxicity in nude rats. To extend the preclinical evaluation of intrathecal 4-HC, we modified several of theexperimental conditions used by these investigators: (a) we usedimmunocompetent rabbits instead of nude rats, inasmuch as thefull expression of neurotoxicity may require a functional immune system; (b) we administered 4-HC by an indwelling intraventricular rather than intralumbar catheter, because previous intrathecal drug studies have suggested a therapeuticadvantage for this route (26, 40); (c) we used a weekly scheduleof 4-HC administration instead of a single dose, as multiple

doses more closely approximate the anticipated human clinicalschedule; finally (d) to expand the range of preclinical activitystudies, we used a rabbit carcinoma tumor model (VX2) withsensitivity to 4-HC in vitro that is 4-fold less than reportedvalues for TE-671 (41) (90% inhibitory dose = 50 MMversus12.8 MMfor VX2 and TE671, respectively).

In the present study, weekly intraventricular 4-HC at doses>1.0 /¿molproduced a dose-dependent vasculitis that most often involved the superficial arteries of the pons and medulla. Athigher 4-HC doses (>2.0 Mmol), axonal degeneration, gliosis,and focal necrosis and hemorrhage were also observed. Demy-elination and other signs of parenchymal injury have been reported following intrathecal administration of methotrexate(42, 43) or 1-0-o-arabinofuranosylcytosine (44) and may represent a nonspecific response to chemical injury. However, thedevelopment of vasculitis as the earliest histological evidence ofneurotoxicity has not been reported in CSF studies for anyantineoplastic drugs. It is possible that the observed vasculitis iscaused by acrolein, the metabolite that may be responsible forcyclophosphamide-induced hemorrhagic cystitis (45). Alternatively, the higher rate of cell division for cerebral blood vessels(46, 47) may predispose the superficial arteries to injury fromalkylating agents. The pontomedullary distribution of arteritismay occur because these vessels are exposed to a high concentration of 4-HC as CSF exits from the foramina of the fourthventricle. The absence of detectable 4-HC plasma concentrations in animals treated at the highest intrathecal doses(6 MHiol)excludes the possibility that the observed vasculitis isdue to high intravascular drug concentration.

E. cuniculi is a ubiquitous, endemic parasitic infection oflaboratory rabbits. This parasitic infection is typically found inimmunocompetent rabbits and is not specifically associatedwith a compromised host. The incidence of E. cuniculi infectionwas apparently higher in the 4-HC-treated rabbits than in controls and it is possible that i.t. 4-HC treatment activated orfacilitated the development of this clinically indolent, nonfatalinfection. Although we cannot exclude a contributory role, it isunlikely that the arteritis was due to the coexistent E. cuniculiinfection for the following reasons: (a) rabbits without evidenceof encephalitozoonosis treated with 4-HC at doses >1.0 Mmoldeveloped arteritis; (b) rabbits with encephalitozoonosis nottreated with 4-HC and those treated at doses <1.0 Mmoldid notdevelop arteritis; (c) the histological appearance of arteritis inanimals with and without encephalitozoonosis was identical;

6172

on May 12, 2020. © 1992 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

i.t. 4-HC IN LEPTOMEN1NGEAL CARCINOMATOSIS

Table 4 Inlralhecal 4-hydroperoxycyclophosphamide interstudy comparisons

Ref.Arndt

el al. (16)Fuchs et al. (12)Present studyPhase I trialSpeciesMonkey

RatRabbitHumanInjected

dose1.7

«mol(0.40 mg)0.4 «mol(0.09 mg)*1.0 «mol(0.23 mg)ft4.4 «mol(1.00 mg)'CSFvolume

(ml)12.0

0.42.0

140.0Estimated

14-HClcs,,"

(DIM)0.142

1.0000.5000.031

" Estimated |4-HC]<-si-, calculated as injected dose (mmol) - CSF volume,

assumes instantaneous drug distribution and no clearance.* Maximum tolerated dose used in survival studies.' Phase I (rial in progress. Injected dose indicates starting dose in dose escalation

schedule.

and (d ) the granulomatous changes characteristic off. cuniculiwere not observed in blood vessels.

Our results contrast with intrathecal 4-HC neurotoxicity

studies reported in the nude rat (12): (a) the maximum tolerateddose in the nude rat was 2-fold higher than in the rabbit (Table4); (b) 4-HC neurotoxicity in the nude rat was characterizedhistologically by demyelination, most prominently affecting thedorsal columns of the thoracic and cervical spinal cord, andthere was no evidence of arteritis or parenchymal inflammatorychanges as noted in the rabbit. Differences in maximum tolerated dose and neuropathology between the nude rat and rabbitmodels may be due to the weekly for 4 weeks dose schedule inthe rabbit compared with a single dose in the rat. Alternatively,the ability to mount an inflammatory response may be generallyreduced in immunodeficient nude rats or the drug clearance ratemay be higher in the rat.

In the present study, the pharmacokinetic parameters forintraventricular or intralumbar 4-HC administration in the rabbit were closely comparable to those reported by Arndt et al. inthe rhesus monkey (16). Differences in peak 4-HC concentration and AUC between these studies reflect the smaller CSFvolume in the rabbit. Following an intraventricular injection of4-HC, the mean AUC for lumbar CSF was 55% lower than thatfor the cisterna magna. However, a greater difference in AUCwas observed; the AUC for the cisterna magna was 89% lowerthan the lumbar AUC value. These results, suggesting that intraventricular 4-HC may provide better drug exposure to theentire CSF space than intralumbar administration, are similarto findings reported by Bleyer and Poplack (26) and Shapiroet al. (40) for intraventricular methotrexate.

Treatment of VX2 leptomeningeal tumor-bearing rabbitswith four weekly doses of 4-HC produced a dose-dependentincreased survival. Although survival in three rabbits was prolonged, these rabbits were not "cured" of tumor, as evidenced by

minute foci of residual leptomeningeal tumor identified at elective postmortem examination. It is beyond the scope of thisstudy to identify the optimal treatment schedule for clinical use.It is possible that continuation of weekly treatments, an increase in dose intensity, or other alterations may further prolong survival or yield cures. The improvement in survival afterfour 4-HC treatments in VX2 tumor-bearing rabbits was similar to that reported after single-dose 4-HC treatment in nuderats bearing human rhabdomyosarcoma or malignant glioma,both of which show greater in vitro sensitivity to 4-HC thandoes VX2 (12). An evaluation of leptomeningeal tumor at thetime of death or sacrifice showed that the greatest residualtumor burden was located in intracranial and upper cervicalspinal cord regions for untreated rabbits and in the lumbarregion for 4-HC-treated rabbits. These findings suggest that theregional inequities in 4-HC distribution demonstrated in pharmacokinetic studies may have significant implications for clin

ical outcome. Recognizing that CSF dynamics are usually abnormal in patients with leptomeningeal metastasis and thatCSF transit is often delayed, it is possible that effective therapymay require alternating intraventricular and lumbar administration routes.

Our results indicate that i.t. 4-HC, at doses lower than thoseassociated with significant neurotoxicity, is an effective treatment for VX2 leptomeningeal carcinomatosis in the rabbit.This and other preclinical studies strongly suggest that i.t.4-HC may provide an important addition to the small numberof chemotherapeutic agents available for intrathecal use. Recently opened clinical trials for adults and children with leptomeningeal tumor will define the toxicity and clearance ratesfor i.t. 4-HC and provide additional information for optimal i.t.4-HC dosing schedules.

ACKNOWLEDGMENTS

The authors thank Carol Hartke for technical assistance, BettyTyson for manuscript preparation, and Pamela Talalay for editorialadvice.

REFERENCES

12.

13.

14,

15

16.

17.

6173

Blcyer, W. A., and Poplack. D. G. Prophylaxis and treatment of leukemia inthe central nervous system and other sanctuaries. Scmin. Oncol., 12: 131-148. 1988.Evans, A. W.. Gilbert, E. S., and Zandstra. R. The increasing incidence ofcentral nervous system leukemia in children. Cancer (Phila.). 26: 404-409,1970.Wasserstrom. W. R.. Glass. J. P., and Posner, J. B. Diagnosis and treatmentof leptomeningeal métastasesfrom solid tumors: experience with 90 patients.Cancer (Phila.), 49: 759-772, 1982.deVisser, B. W. O., Somers, R.. Nooyen, W. H., van Heerde, P.. Hart, A. A.M.. and McVie, J. G. Intraventricular methotrexate therapy of leptomeningeal metastasis from breast carcinoma. Neurology, 33:1565-1572, 1983.Trump, D. L., Grossman. S. A.. Thompson, G., Murry. K.. and Wharam, M.Treatment of ncoplastic meningitis with intraventricular thiotepa and methotrexate. Cancer Treat. Rep., 66: 1549-1551, 1982.Peter. G.. Wagner, T.. and Hohorst, H. J. Studies on 4-hydroperoxycyclophosphamide (NSC-181815): a simple preparation method and its application for the synthesis of a new class of "activated" sulfur-containing cyclo-phosphamide (NSC-26271) derivatives. Cancer Treat. Rep.. 60: 429-430,1976.Colvin, M., and Hilton. J. Pharmacology of cyclophosphamidc and metabolites. Cancer Treat. Rep., 65: 89-95, 1981.Friedman. H. S.. Colvin, O. M., Skapek. S. X., Ludeman. S. M.. Elion. G. B.,Schold. S. C., Jacobsen, P. F., Muhlbaier. L. H.. and Bigner. D. D. Experimental chemotherapy of human medulloblastoma cell lines and transplant-able xcnografts with bifunctional alkylating agents. Cancer Res.. 48: 4189-4195, 1988.Redwood. W.. Hilton. J., Colvin, M.. and Owens, A. The cellular uptake of4-hydroperoxycyclophosphamide. Proc. Am. Assoc. Cancer Res., 23: 169,1982.DeFabritiis, P.. Bregni, M., Lupton, J., Greenberg, J., Nadler, L., Rothstein.L., Korbling, M.. Ritz, J., and Bast. R. Elimination of clonogenic Burkitt'slymphoma cells from human bone marrow using 4-hydroperoxycyclophos-phamidc in combinations with monoclonal antibodies and complement.Blood. 65: 1064-1070, 1985.Kubota. T., Hanatani, Y., Tsuyuk, K., Nakada, M., Ishibiki. K., Abe, O.,Kamataki. T., and Kalo, R. Antitumor effect and metabolic activation ofhydroperoxycyclophosphamide in the human breast adcnocarcinoma (MX-1)nude mouse system. Gann, 74: 437-444, 1983.Fuchs, H. E., Archer, G. E., Colvin, O. M., Bigner, S. H., Schuster, J. M.,Füller,G. N.. Muhlbaier. L. H„Schold, S. C.. Friedman, H. S., and Bigner,D. D. Activity of intrathecal 4-hydroperoxycyclophosphamide in a nude ratmodel of human neoplastic meningitis. Cancer Res.. 50: 1954-1959, 1990.Mirée.J., Jr.. and Gold, S. Relationship of survival with number of V-2carcinoma cells implanted in the subarachnoid space of rabbits. J. Nail. Med.Assoc. 65:407-409, 1973.Hohorst. H.. Peter. G., and Struck, R. Synthesis of 4-hydroperoxy derivativesof ifosfamide and trofosfamide by direct ozonation and preliminary antitu-mor evaluation in vivo. Cancer Res., 36: 2278-2281, 1976.Alarcon, R. A. Fluorometric determination of acrolein and related compounds with m-aminophcrol Anal. Chem.. 40: 1704-1708. 1968.Arndt, C. A. S.. Colvin, O. M., Balis, F. M., Lester, C. M., Johnson. G., andPoplack. D. G. Intrathecal administration of 4-hydropcroxycyclophospha-mide in rhesus monkeys. Cancer Res.. 47: 5932-5934, 1987.Samphilipo, M. A., Jr.. Hasscnbusch, S. J., Grochow, L. B., Starr. F. L., Ill,

on May 12, 2020. © 1992 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

i.I. 4-HC IN LEPTOMENINGEAL CARCINOMATOS1S

and Anderson, J. H. A reservoir model for repeated CSF access in the rabbit.J. Neurosci. Methods. 22: 47-52. 1987.

18. Gibaldi. W., and Perrier, D. Pharmacokinetics. New York: Marcel Dekker,.Inc., 1982.

19. Carson. B. S.. Anderson, J.. Grossman, S., Hilton, J., White, C, Colvin, O.M.. Clark. A., Grochow. L., Kahn. A., and Murry, K. An improved rabbitbrain tumor model available for radiographie diagnostic procedures. Neuro-surgery. //: 603-609. 1982.

20. Kokkoris. C. P. Leptomeningeal carcinomatosis. How does cancer reach thepia-arachnoid. Cancer (Phila.), 51: 154-160, 1983.

21. de la Monte, S. M.. Moore. G. W., and Hutchins. G. M. Patterned distribution of métastasesfrom malignant melanoma in humans. Cancer Res.. 43:3427-3433, 1983.

22. Balducci. L., Little, D. D., Khansur, T.. and Steinberg. M. H. Carcinomatousmeningitis in small cell lung cancer. Am. J. Med. Sci., 287: 31-33, 1984.

23. Yap. H. Y., Yap, B. S., Rasmussen, S., Leavens, M. E., Hortobagyi, G. N.,and Blumenschien. G. R. Treatment for meningeal carcinomatosis in breastcancer. Cancer (Phila.), 49: 219-222. 1982.

24. Packer, R. J.. Siegel. K. R.. Sutton, L. N., I limami. P., Bruce. D. A., andSchul. L. Leptomeningeal dissemination of primary central nervous systemtumors of childhood. Ann. Neurol., 18: 217-221, 1985.

25. Nugent. J. L., Bunn, P. A., and Matthews, M. J. CNS métastasesin small cellbroncogenic carcinoma —¿�increasing frequency and changing patterns withlengthening survival. Cancer (Phila.). 44: 1885-1993, 1979.

26. Bleyer, W. A., and Poplack, D. G. Intraventricular versus intralumbar meth-otrexate for central nervous system leukemia: prolonged remission with theOmmaya reservoir. Med. Pediatr. Oncol., 6: 207-213, 1979.

27. Ongerboer de Visser. B. \V„Somers. R., Nooyen, W. H., Van Heerde. P..Hart. A. A. M., and McVie. J. G. Intraventricular methotrexate therapy ofleptomeningeal metastasis from breast carcinoma. Neurology, 33: 1565-1572. 1983.

28. Schabet. M., Kloeter. I.. Adam. T.. Heidemann. E., and Wietholter. H.Diagnosis and treatment of meningeal carcinomatosis in ten patients withbreast cancer. Eur. Neurol., 25: 403-411, 1986.

29. Shapiro, \V. R. Necrotizing encephalopathy following intraventricular instillation of methotrexate. Arch. Neurol., 2*.-96-102. 1973.

30. Norrell, H.. and Wilson, C. B. Leukoencephalopathy following the administration of methotrexate into the cerebrospinal fluid in the treatment of primary brain tumors. Cancer (Phila.), 33: 923-932. 1974.

31. Hoeve. R. R. A., and Twijnstra. A. A lethal neurotoxic reaction after intraventricular methotrexate administration. Cancer (Phila.). 62: 2111-2113.1988.

32. Gulin. P. H., Levi, J. A., Wiernik. P. H., and Walker. M. D. Treatment ofmalignant disease with intrathecal thioTEPA: a phase II study. Cancer Treat.Rep.. 61: 885-887. 1977.

33. Haaxma-Reiche, H., Daencn. S.. and Witteveen. R. J. W. Experiences with

the Ommaya reservoir for prophylaxis and treatment of the central nervoussystem in adult acute myelocytic leukemia. Blut, 57: 351-355, 1988.

34. Ushio, Y., Posner, J. B., and Shapiro. W. R. Chemotherapy of experimentalmeningeal carcinomatosis. Cancer Res., 37: 1232-1237, 1977.

35. Kooistra, K. L., Rodriguez, M., and Powis, G. Toxicity of intrathecallyadministered cytotoxic drugs and their antitumor activity against an intrathecal Walker 256 carcinosarcoma model for meningeal carcinomatosis inthe rat. Cancer Res., 49: 977-982. 1989.

36. Moseley. K. Intrathecal administration of '"I radiolabelled monoclonal antibody as a treatment for neoplastic meningitis. Br. J. Cancer, 62: 637-642,1990.

37. Kim, S., Kim, D. J., Geyer, M. A., and Howell, S. B. Multivesicular lipsomescontaining l-A-o-arabinofuranosylcytosine for slow-release intrathecal therapy. Cancer Res. 47: 3935-3937, 1987.

38. Levin, V. A., Chamberlain, M., Silver, P., Rodriguez, L., and Prados, M.Phase I/II study of intraventricular and intratheacal ACNU for leptomeningeal neoplasia. Cancer Chemother. Pharmacol., 23: 301-307, 1989.

39. Arndt, C. A. S., Balis, F. M., Lester, C., McCully, L., Colvin, O. M. andPoplack. D. G. Cerebrospinal fluid penetration of active metabolites of cy-clophosphamide and ifosfamide in rhesus monkeys. Cancer Res., 48: 2113-2115, 1988.

40. Shapiro, W. R.. Young, D. R., and Metha, B. M. Methotrexate: distributionin cererospinal fluid after intravenous, ventricular and lumbar injections.N. Engl. J. Med., 293: 161-166, 1975.

41. Friedman. H. S., Colvin, O. M., Ludeman, S. M., Schold, S. C, Boyd, V. L.,Mulhbaier, L. H., and Bigner, D. D. Experimental chemotherapy of humanmedulloblastoma with classical alkylators. Cancer Res. 46:2827-2833. 1986.

42. Skullerud. K.. and Halvorsen, K. Encephalomyelopathy following intrathecalmethotrexate treatment in a child with acute leukemia. Cancer (Phila.), 42:1211-1215.1978.

43. Bates. S. E., Raphaelson. M. I., Price, R. A.. McKeever, P., Cohen, S., andPoplack, D. G. Ascending myelopathy after chemotherapy for central nervous system acute lymphoblastic leukemia: correlation with cerebrospinalfluid myelin basic protein. Med. Pediatr. Oncol., 13: 4-8, 1985.

44. Breuer, A. C.. Pittman, S. W.. Dawson, D. M., et al. Paraparesis followingintrathecal cytosine arabinoside. A case report with ncuropathological findings. Cancer (Phila.), 40: 2817-2822, 1977.

45. Cox, P. J. Cyclophosphamide cystitis —¿�identification of acrolein as thecausative agent. Biochem. Pharmacol.. 28: 2045-2049. 1979.

46. Korr. H.. Schultz, B., and Maurer. W. Autoradiographic investigations ofglial proliferation in the brain of adult mice. J. Comp. Neurol.. ISO: 169-176,1971.

47. Herndon, R. M., Price, D. L., and Weiner, L. P. Regeneration of oligoden-droglia during recovery from demyelinating disease. Science (WashingtonDC), /95: 693-694. 1977.

6174

on May 12, 2020. © 1992 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

1992;52:6168-6174. Cancer Res   Peter C. Phillips, Thuy T. Than, Linda C. Cork, et al.   in a Rabbit Model of VX2 Leptomeningeal CarcinomatosisCerebrospinal Fluid Pharmacokinetics, and Antitumor Activity Intrathecal 4-Hydroperoxycyclophosphamide: Neurotoxicity,

  Updated version

  http://cancerres.aacrjournals.org/content/52/22/6168

Access the most recent version of this article at:

   

   

   

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

  Subscriptions

Reprints and

  .pubs@aacr.orgDepartment at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

  Permissions

  Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/52/22/6168To request permission to re-use all or part of this article, use this link

on May 12, 2020. © 1992 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from