(CANCER RESEARCH 48. 2904-2910 May 15, 1988)

Treatment of Intracranial Human Glioma Xenografts with 131I-labeledAnti-Tenascin Monoclonal Antibody 81C61

Yisheng Lee, Dennis E. Bulinai, Peter A. Humphrey, Edward V. Colapinto, Henry S. Friedman, Michael R. Zalutsky,R. Edward C(»lemán,and Dareil D. Bigner2

Departments of Pathology [Y. L., P. A. H., E. V. C., M. R. Z., D. D. B.], Surgery [D. E. B., D. D. B.], Pediatrics fH. S. F.], and Radiology [M. R. Z., R. E. C.J and thePreuss Laboratory for Brain Tumor Research /£>.D. B,], Duke University Medical Center, Durham, North Carolina 27710

ABSTRACT

Lack of tumor specificity renders current modalities for treating malignant glioma ineffective. The administration of '"¡-labeled monoclonal

antibody (Mah) 81C6, which reacts with the glioma-associated extracel

lular matrix antigen, tenascin, to nude mice carrying s.c. human gliomaxenografts has resulted in significant tumor growth delay and tumorregression. In this study, we evaluated the therapeutic efficacy of '"I-

labeled 81C6 in athymic rats bearing intracranial human glioma xenografts, a more appropriate model for human gliomas. Mah 81C6, anIgG2b immunoglobulin, and an isotype-matched control Mab, 45.6, werelabeled at 12.5-23.6 mCi/mg with chloramine-T. The Mabs were giveni.v. at 1.25 and 2.5 mC ¡/animalfor "'I-labeled 81C6, and 1.25 mCi for1"I-labeltd 45.6 control. Therapeutic response was evaluated by survival

prolongation using Wilcoxon rank sum analysis. Three experiments weredone. No significant survival prolongation was found in the trial in whichthe average tumor size at the time of Mab administration was 60 ±14mm ', two-thirds the size which causes animal death. In experiment 2,Mab was given at 16 ± 14 mm3 average intracranial tumor volume.Statistically significant ( /' £0.005) survival prolongation was found foranimals treated with 2.5 mCi ' "I-labcled 81C6. In that experiment, male

animals with intracranial xenografts had significantly shorter survivalthan females (/' < 0.005). When only female animals were used in the

analysis, the 1.25-mCi 81C6 group also was found to have longer survivalbenefit (/' £0.01). In the third experiment, only female animals wereused and the tumor size at the initiation of treatment was 20 ±9 mm3.Highly significant survival prolongation again was found in both 1.25 (/'= 0.001) and 2.5 mCi (/' < 0.001 ) ' "I-lahelcd 81C6 groups. The estimateddose to intracranial tumors from 1.25 mCi of I3ll-labeled Mab was 1585

rads for 81C6 and 168 rads for 45.6. Dose to other organs from 81C6and 45.6 was similar, ranging between 31 rads to the brain and 734 radsto the bone marrow. However, normocellularity was observed in mostmarrow tissue examined microscopically. Three animals receiving thelow dose (1.25 mCi 81C6) survived for more than 71 days with apparentcures. In conclusion, intracranial human glioma xenografts were treatedsuccessfully with ' "l-labvlcd 81C6 but not control Mab.

INTRODUCTION

The objective of targeting human gliomas with antibodieshas been attempted with conventional heteroantibodies for diagnostic and therapeutic purposes. Day et al. ( 1) demonstratedspecific localization of radiolabeled rabbit anti-human gliomaantibodies in human brain tumors in vivo by a qualitativeanalytical method based on tissue fraction distribution ingliomas. However, problems such as low specific antibody yield(approximately 0.0002% of whole antiserum globulin), normaltissue cross-reactivity, and heterogeneity of antibody-specificaffinity limit the application of labeled heteroantibodies. Incontrast to heteroantibodies, monoclonal antibodies operationally specific to tumors can be produced in high quantity withhigh homogeneity.

Received 8/26/87; revised 2/15/88; accepted 2/22/88.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 inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact.

' This work was supported by NIH Grants RO1-CA 42324, RO1-CA 43638,RO1-CA 11898, NIH-CA 32672, and NS-20023.

2To whom requests for reprints should be addressed, at Box 3156, Duke

University Medical Center, Durham, NC 27710.

Radiolabeled Mabs3 against tumor-associated antigens have

been investigated in human tumor xenograft models as well asin human trials. Various systems have been studied includinglymphoma, colorectal carcinoma, melanoma, ovarian carcinoma, and breast cancers (2-8). For human glioma, Kemsheadet al. (9) have demonstrated positive brain tumor images 12-14 days after injection of murine Mab UJ13A in patients,although in later paired-label studies there was no specificantibody localization observed with this antibody (10). Monoclonal antibodies reactive with primary brain tumors that localize specifically, but not with normal brain, would provide apotential means of delivering diagnostic and therapeutic agentsselectively to human malignant gliomas (11).

Mab 81C6, an IgG2b immunoglobulin, defines an epitope ofthe extracellular matrix antigen tenascin, present in many human glioma cell lines, primary human gliomas, and gliomaxenografts in nude mice and rats, but not in normal adult orfetal brain (10). In paired-label studies, radioiodinated 81C6has been shown to localize specifically in s.c. and i.e. HGLxenografts ( 12). The specificity of localization was high enoughto allow imaging of intracranial tumors as small as 20 mg. Incontrast, with radiolabeled control Mab 45.6, only tumors muchlarger in size (300 mg) could be imaged (13). In therapeutictrials, we have achieved significant growth delay and tumorregression of s.c. D-54 MG HGL xenografts in nude mice afterinjection of ' " l-labclcd 81C6 (14). In a preliminary paired-labelstudy of the utility of ml-labeled 81C6 administered in the

carotid artery for the specific delineation of gliomas in patients,increased tracer accumulation was noted in regions corresponding to computerized tomography-defined tumor sites, and antibody to control immunoglobulin ratios of up to 5:1 wereobserved in biopsies of the intracranial gliomas (15).

In this study, we have evaluated the therapeutic efficacy of'"I-laliclcd 81C6 in intracranial tumors because this model

possesses some unique biological characteristics resulting fromthe intracranial environment. These include the presence of theblood-brain barrier, which can decrease the accessibility ofantigenic target sites for Mabs, and altered antigen expressionand distribution; all influence the result of treatment (16, 17).We have demonstrated significant survival benefit in animalscarrying intracranial D-54 MG xenografts treated i.v. with I3II-

labeled 81C6 compared to animals receiving labeled controlMab.

MATERIALS AND METHODS

Intracranial Xenograft Model. HGL D-54 MG, the Duke Universitysubline A-172 established by G. Todaro (18), is serially transplantableas a s.c. tumor in athymic mice. For i.e. experiments, athymic ratsobtained from a colony maintained at Animal Laboratory and IsolationFacility, Duke University, were inoculated with 10 n\ of D-54 MGhomogenate, prepared from s.c. tumors, in the right hemisphere at adepth of 5 mm using a Hamilton syringe fitted with a No. 23 Luer stub

3The abbreviations used are: Mab, monoclonal antibody, HGL, human glioma

cell line; i.e., intracranial.

2904

on August 9, 2019. © 1988 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

RADIOIMMUNOTHERAPY OF INTRACRANIAL GLIOMA XENOGRAFT

adapter needle (Becton Dickinson, Parsippany, NJ) and plastic depthgauge. The homogenate was prepared by mixing an equal volume of1% methylcellulose (Fluka, Buchs, Switzerland) in zinc option mediumwith homogenized tumor cell pellets. Intracranial tumor growth wasmonitored by both animal survival and direct tumor size measurementof brain/tumor sections from animals sacrificed at different times. D-54 MG HGL grows as a well circumscribed intracranial tumor inathymic rats. The length of the shortest and longest diameters wasmeasured. Estimated tumor volume was determined according to theformula: [(short dimension)2 x (long dimension)]/2.

Monoclonal Antibody. Mab 81C6 and a control Mah of the sameIgG2b isotype, 45.6 TG 1.7 (45.6), were purified from culture supernatant using a protein A-sepharose affinity column. Mab 45.6 is a myeloma IgG2b with no known specificity. The purity of Mabs was examined by sodium dodecyl sulfate-polyacrylamide gel electrophoresis andhigh pressure liquid chromatography. Mabs were labeled with ' "1 using

the chloramine-T method (19) and passed through a 10-ml SephadexG-25 fine column to separate labeled Mabs from free iodide. The ' "Iactivity in the void-volume peak of these preparations was 95-98%precipitatile in trichloroacetic acid and demonstrated a single peakprofile on a size exclusion high pressure liquid chromatography columnwith a retention time corresponding to intact IgG. All Mabs were usedwithin 4 h of labeling.

Immunoreactivity Assay and Mab Affinity Determination. The bindingcharacteristics of purified radiolabeled Mab 81C6 were analyzed byimmunoreactivity and affinity assays. For immunoreactivity assay, 100ng of 13ll-labeled 81C6 or 45.6 Mab were incubated overnight with 300

mg of washed D 54 MG s.c. tumor homogenate in 1 ml 0.15 Mphosphate saline buffer. The amount of 45.6 bound to the tumorhomogenate was considered to represent nonspecific binding and wassubtracted from the percentage of 81C6 bound in order to calculate thespecific binding percentage of "'I-labcled 81C6. For measurement of81C6 Mab binding affinity, a high antigen-expressing HGL, U251MG-clone 3, was used and an antigen-negative human osteogenic sarcomacell line, 2T, was used to determine nonspecific binding. For each line,5 x IO4cells/well were plated in 96-well plates in 10% fetal calf serum

zinc option medium 3 days before assay. Serial dilutions of batches ofml-labeled 81C6 Mab used in the therapy experiments were added to

both cell lines in quadruplicate and incubated overnight. Binding datawere analyzed using the Equilibrium Binding Data Analysis programobtained from the Biomedicai Computing Technology InformationCenter, Vanderbilt Medical Center, Nashville, TN, and modified foran IBM PC computer by McPherson (20). Binding affinity was compared to previous data with 81C6 obtained at the lower specific activities(0.4-5.6 mCi/mg) used in radiolocalization experiments.

Therapeutic Trials. Three experiments were performed. Animalscarrying intracranial D-54 MG tumors were randomized according tobody weight the day before therapy. In the first two experiments, bothmale and female animals were used with weights varying from 100 to400 g. Only female athymic rats weighing between 150 and 200 g wereused in the third experiment. Lugol's solution was also given, starting

1 day prior to treatment and continuing throughout the experimentalperiod, to block iodine uptake in the thyroid. Ten animals per treatmentgroup were used routinely except in experiment 2; five animals wereused for the 2.5 mCi/200 /ig treatment group. In each experiment, fiveadditional animals were sacrificed on the day of Mab injection toestablish the existence of i.e. tumors and to obtain the average tumorvolume at the initiation of therapy. The animals were kept in anisolation environment (Animal Laboratory and Isolation Facility, DukeUniversity) and checked twice daily for survival. Each animal wasexamined after death to establish the existence of an intracranial tumor.

Animal groups in experiment 1 were given buffer (phosphate salinebuffer, 2% bovine serum albumin), unlabeled 81C6 (100 /ig), 131I-labeled81C6 (1.39 mCi/100 /¿g),or l31I-45.6 (1.61 mCi/100 /ig) via the femoral

vein 14 days after tumor inoculation. In experiment 2, buffer, unlabeled81C6 (100 Mg),131I-81C6 (1.25 mCi/100 /ig and 2.5 mCi/200 /<g),orl31I-45.6 (1.25 mCi/100 jig) was given to animals 10 days after tumor

transplantation. A similar protocol was used in experiment 3 withunlabeled 81C6 (100 /ig), '"I-81C6 (1.25 mCi/53 /ig and 2.5 mCi/106/ig), or l3'I-45.6 (1.25 mCi/46 /ig) given 11 days after tumor inoculation.

Treatment response was evaluated using Wilcoxon rank sum test bycomparing the animal survival time in groups given Mabs 81C6, ' "1-81C6, and l3'I-45.6 with that of animals treated with buffer control

(21). The mortality difference between male and female rats in allcontrol groups (buffer, 81C6, l3'I-45.6) was also compared.

Antibody Localization and Radiation Dosimetry. In experiment 3, 30additional animals were given injections of 1.25 mCi of either '"I

labeled 45.6 or 81C6 and were used for antibody biodistribution anddosimetry analysis. Three animals from each group were sacrificed ondays 1, 2, 4, 8, and 12 after Mab injection. Blood, femur, intracranialtumor, and normal brain samples were taken immediately after animaldeath. The rest of the animal carcass was stored in a —20°Cfreezeruntil 30 days after Mab injection to allow for decay of '"I activity.

Tumor, blood, and various tissues were weighed and counted in agamma counter (Packard Instruments, Downers Grove, IL). Bonemarrow was derived from the red marrow of femurs. The levels of I3'I

in tissue samples were compared to injection standards of appropriatedilution in order to calculate the percentage of the injected Mab dose/g of tissue.

For calculation of radiation dosimetry, tissue '"I levels measured on

day 30 after injection were corrected for radiation decay to days 1, 2,4, 8, and 12 postinjection. Cumulative absorbed radiation doses werecalculated using the trapezoid integration method as described (14).Cumulative activity in each tissue was calculated from the area underthe percentage of dose/g tissue uptake curves using the trapezoidintegration method as described (14). Distribution of '"I in each tissue

was assumed to be homogeneous. These data were then converted to/iCi •h/g by multiplying the percentage of dose •h/g data by the /iCiinjected in each treatment group. Radiation doses in rads were calculated using the standard method of the Medical Internal RadiationDose Committee by multiplying the integrated /iCi-h/g by the g •rads/nCi-h factor for I31I,0.4313 (22).

RESULTS

Intracranial Model. Following i.e. transplantation of D-54MG tumor homogenate, 100% of control animals died within14-22 days. There was no correlation between the presence ofneurological signs and symptoms and the time of animal death.Randomization to treatment and control arms was thus doneaccording to animal weight at the time of Mab injection. Anaverage tumor volume of 93 ±31 mm3 at death was seen (Fig.

1). All tumors were well circumscribed with minimal invasionof surrounding brain (Fig. 2/4). The average tumor volumes atthe time of Mab administration in the three treatment experiments were 60 ±14 (SD), 16 ±14, and 20 ±9 mm3, respectively

(Table 1).

10080______nE1«^__j9

40_tO=>

20i-

/\~•.i^d.t/

i(20

±9)/T_p~rIP

Jt(16H4),

, , /l ,/'

\'

-1-n*^//f(60t

14)í

r\AVERAGE"VOLUME

ATDEATHINITIATION

OFU1I-81C6-RADIOIMMUNOTHERAPY8

10 12 1416DAYS

AFTER TUMORINJECTIONINTRACEREBRALLY

Fig. 1. D-S4 MG intracranial growth curve, tumor volume at animal death,and treatment tumor volumes. Tumor doubling time is 2.5 days.

2905

on August 9, 2019. © 1988 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

RADIOIMMUNOTHERAPY OF INTRACRANIAL GLIOMA XENOGRAFT

B

r

Fig. 2. Whole brain section for tumors at animal death (A) and at the beginningof treatment for experiment 2 (A), and for brain of "cured" animals surviving

255 days after tumor inoculation (C). Well circumscribed tumors were seen in Aand B. No evidence of tumor was observed in C, with original needle tract seen.T, tumors; arrow, needle tract. H & E.

Characterization of '"I-81C6. The binding characteristics of'"I-labeled 81C6 used in the therapeutic studies were analyzed

by both percentage of immunoreactivity as determined by ab

sorption with D-54 MG tumor homogenate and Scatchardanalysis of binding affinity to an antigen-expressing cell line invitro. After correction for nonspecific binding, 40-50% bindingactivity was observed for 81C6 used in treatments after overnight incubation in the immunoreactivity assay. A range of 35-70% immunoreactivity was obtained in previous localizationexperiments using lower specific activity 81C6. The bindingaffinity of therapeutic level I31I-81C6 (23.6 mCi/mg) was 3.03x IO8 M"' in experiment 3 (Fig. 3), even though the I31I-81C6

used had the highest specific activity used in the three experiments. This compared favorably to a range of 3.3 x 107-3.0 xIO8 M~" found in previous experiments using lower specificactivity 13ll-labeled 81C6 (0.5-5.0 mCi/mg). These data indi

cate that, at the higher specific activities used in these studies,the binding affinity and immunoreactivity of I31l-labeled 81C6were preserved and permitted a direct comparison of the bio-distribution data presented herein with the data obtained previously in localization and imaging experiments.

Antibody Biodistribution. The Mab blood clearance and thelocalization profiles expressed as percentage of dose/g tissue innormal tissues and tumor are shown in Fig. 4. Because controlMab 45.6 failed to produce any survival benefit, only one animalin the 45.6-treated group was alive at 8 days and available formeasurement of antibody biodistribution. Two animals in the81C6 treatment group remained alive at day 12. After i.v.injection, a biphasic clearance from the blood was observed,similar to that observed previously in trace-labeled localizationexperiments. No difference was seen between 81C6 and thecontrol Mab, 45.6; both cleared from the blood with a half-lifeof 2.8 days. No difference in 81C6 and 45.6 antibody localization in all other normal tissues examined was observed, withthe concentration of 13II reaching a peak 24 h after Mab

injection and decreasing slowly afterward. There was minimallocalization in the brain and muscle. A relatively high ' "I level,

however, was observed in the bone marrow.The level of 13ll-labeled 81C6 in tumors did not reach a

plateau (2.9% injected dose/g) until 48 h after Mab injection;it decreased to 1% injected dose/g at 12 days. The localizationindex, the average ratio of specific 81C6 Mab to control 45.6Mab in the tumor normalized to blood level, increased continuously from 6.2 at 24 h to 15 at day 8 after Mab injection.Tumor '"I-labeled 81C6 levels remained higher than those in

any other tissue at all time points. The concentration of controlMab '"I-labeled 45.6 in tumors was similar to the level achieved

in normal tissues, basically reflecting the blood level at eachtime point. These data are similar to those obtained in tracelabel localization experiments.

Intracranial Tumor Therapy. In experiment 1 (Fig. 5A), tumorsize at the initiation of treatment was 60 ±14 mm3 (Fig. 1).

All buffer control animals died within 4 days. No significantdifference between buffer control and treatment groups (bufferversus 81C6, P = 0.5; buffer versus '"1-45.6, P = 0.398, bufferversus '"I-81C6, P = 0.140) was found, presumably because of

the large average initial tumor size. However, four animals inthe I3'I-81C6 treatment group had slightly longer survival times(20-22 days), suggesting possible benefit in animals carryingsmall tumors at the beginning of treatment.

In experiment 2 (Fig. 5B), the treatment was initiated 10days after tumor injection with a 16 ±14 mm3 average tumor

size (Fig. 1). The buffer control animals died between days 14and 19 postinoculation. Survival data again were analyzed withthe following significances of difference observed: buffer versus81C6, P = 0.5; buffer versus '-"1-45.6, P = 0.456; buffer versus1.25 mCi 131I-81C6,P= 0.095; buffer versus 2.5 mCi 131I-81C6,

2906

on August 9, 2019. © 1988 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

RADIOIMMUNOTHERAPY OF INTRACRANIAL GLIOMA XENOGRAFT

Table 1 Statistical analysis ofD-54 MG i.e. xenograft treatment response

Initial tumor AnimalExperiment volume"no.1

60 ±14mm3 10

1010

102

16±14mmJ 10

1010

1053

20 ±9 mm3 10

1010

10TreatmentBuffer

81C6131I-45.6

131I-81C6Buffer

81C6131I-45.6

'"I-81C61311-81C681C6

131I-45.6131I-81C6

131I-81C6Dose100

^g1.61

mCi/lOOf.g1.35mCi/100^g100

Mg1.25mCi/100ííg

1.25mCi/100Mg2.5 mCi/200».g100,1g

1.25mCi/45^g1.25mCi/53,ig

2.5 mCi/106^gMedian

survival(days)16

16161716

1516

202216

182326Survival

range(days)15-18

15-2315-20

15-2214-19

13-2414-20

14-2518-3914-22

15-2315-71

23-33P»NSCNS

NSNSNS

(0.095)''

0.004NS0.001

<0.001°Mean ±SD.b Statistical /"-value of treatment group compared to control group using Wilcoxon sign rank test.c NS, not significant.! /' > 0.05)d P £0.03 when only female animals were used for statistical analysis." Significant as P < 0.05.

1.2

g io<ociti 0.8

§0.6Io

§04-ouS 0.2

(b)

Krfl.89

1.0 2.0 3.081C6 BOUND nM

4.0

Fig. 3. Scatchard analysis comparing binding affinity of trace-labeled (a) andtherapeutic-ally labeled (b) Mab 81C6 on cell line U251-C13). Radioactive specificactivity; trace-label, 3.8 mCi/mg; experiment 3 label, 23.6 mCi/mg.

P = 0.004. Animals receiving 2.5 mCi 131I-81C6 also had

significantly longer survival than the unIaheled 81C6 (P = 0.01)and the l311-45.6 treatment animals (P = 0.01) but had therapeutic results similar to lower dose (1.25 mCi) '"I-81C6 ani

mals (P = 0.339). Such data would predict a significant survivalbenefit (P < 0.05) for 1.25 mCi l311-81Co-treated versus buffer(P = 0.095), unlabeled 81C6 (P = 0.14), and 131I-45.6 (P =

0.116). This incongruity led us to examine the animal survivaldata more closely. In eight animals that died in the first 3 days(days 13-15), regardless of the group to which the animalbelonged, seven were male athymic rats. The two animals withapparent cures in experiment 2 and the last four animals of the13II-81Co-treated group in experiment 1 were all female.

We proceeded to analyze the sex dependence of animalsurvival by comparing the survival data of male and femaleathymic rats pooled from buffer control, unlabeled 81C6, and131I-45.6-treated groups in experiment 2. Male athymic ratswith intracranial D-54 MG tumors died significantly earlierthan their female counterparts (P < 0.005). The effect of bodyweight may also have contributed to this apparent sex differencesince most male animals (13 of 20) were larger than 300 g,whereas all females were between 150 and 300 g. When onlymale animals with weight less than 300 g were used in comparison, less significant results were obtained (0.1 > P > 0.05).

The survival data in experiment 2 were «evaluated with onlyfemale animal data. In agreement with our expectation, 1.25

3

2

I

ui 3D

i/ï;= 2EO

48 96 144 192 240 288

BRAIN HEART LUNG

LIVER KIDNEY SPLEEN

BONE MARROW

•4-

BLOOD

48 96 144 192 240 288 48 96 144 192 240 288 48 96 144 192 240 288

HOURS AFTER Mob INJECTION

Fig. 4. Biodistribution of '"l-lubeled Mabs 81C6 and 45.6 in intracranial D-54 MG tumor-bearing animals.

mCil311-81 Co-treated animals now showed significantly longersurvival than the buffer controls (P < 0.03). This result becameeven more obvious when both buffer and unlabeled 81C6 groupswere used as controls (P< 0.01). Whatever the cause, it seemedjudicious to use only one sex in subsequent survival experiments. As female athymic rats attained their maximal adultweight of approximately 250 g earlier than male animals, weelected to use only female animals for the third experiment.

In experiment 3 (Fig. 5C), the treatment was begun on day11, when the average tumor size was 20 ±9 mm3 (Fig. 1).

Statistically significant survival prolongation was found in comparison to the unlabeled 81C6 control for both 1.25 (P= 0.001)and 2.5 mCi (P < 0.001) 131I-labeled 81C6 groups, but not for

2907

on August 9, 2019. © 1988 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

RADIOIMMUNOTHERAPY OF INTRACRANIAL GLIOMA XENOGRAFT

100

a«20

ot

D-54MGinoculation

100

"10 12 14 16 18 2"0 22 "24 26 28 30 32 34 36 38 40

0 2 46 8 10 12 14 16 18 20 22 24 26Days

a«5

Buffer81C6 (100 Mg)45.6 (1.25 mCi/100pg)81C6 (1.25 mCi/100 yg)

81C6 (2.5 mti/JOOjjg)

n I ,1,1 i—i—i—i—i—i—i—i—i—i_g—i—i—i—a—i—i—i—i—i—i—i—i—i—i—i'ii ."~ 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40

02 46 8 10 12 14 16 18 20 22 24 26 28 30D-54MG Dav«inoculation UaVs

B

™ »ic»uoo^g)™ 45.6 (1.25 mCi/45jjg)'///. 8IC6 (1.25 mCi/5J/jg)*m 8IC6 (2.5 tnli/luo WJ

T 10 12 14 16 18 20 22 24 26 28 30 32 340 2 4 6 8 10 12 14 16 18 20 22

D-54MGinoculation Days

Fig. 5. '"I-81C6 therapy of intracranial D-54 MG in nude rats. Mortality

distribution of three different experiments.. I. experiment 1; B, experiment 2; andC, experiment 3.

the "'I-labeled 45.6 group (P = 0.102). Again there was nosignificant difference between 1.25 and 2.5 mCi "'I-81C6-treated groups (P = 0.124) despite the apparent early survivaladvantage and longer median survival (26 versus 23 days) in the2.5 mCi-treated group.

In addition, three animals with apparent cures were observedin the last two experiments following treatment with "'I-labeled

81C6. Interestingly, all were in animals receiving the lower doseof 1.25 mCi "'I-labeled 81C6. In experiment 2, the two long-

term survival animals died 147 and 255 days after tumorinoculation. The apparently cured animal in experiment 3 liveduntil 71 days postinoculation. All animals were examined atthe time of death and no tumor was found (Fig. 2C). Previousintracranial injection was evidenced by the presence of a needletract in the brain.

Dosimetry. Biodistribution data for "'I-81C6 and 45.6 in

various tissues and tumors were used to calculate the cumulative/iCi-h/g for the tissues of interest. Radiation-absorbed dosecalculations were performed using standard Medical InternalRadiation Dose Committee formalism. For i.e. tumors inathymic rats, 1585 rads were delivered by 1.25 mCi "'I-81C6

over a period of 12 days, whereas only 168 rads were deliveredby '-"1-45.6. These data were similar to the radiation dosespredicted from trace-labeled localization studies. Lower radiation doses were delivered to most normal organs (Table 2),ranging from the lowest to the brain (31 rads) to the highest tothe bone marrow (734 rads). No significant difference wasobserved between I3'l-labeled 81C6 and 45.6 dose in all normal

organs. Except for the three athymic rats with long-term survival, all animals were presumed to have died from the documented presence of intracranial D-54 MG tumors. Theoretically, immunologie-ai suppression may be problematic as 600-

700 rads were delivered to athymic rat bone marrow. In general,though, radiation damage was not evident by morphologicalexamination of femur marrow tissue sections. Normocellularmarrows were observed in animals euthanized at days 1, 2, and4 after Mab injection for both 81C6- and 45.6-treated groups.Marrow examined for the day 8 animal for 45.6 and the day 12animal for 81C6 was also normal. Only one animal (81C6, day8) had a hypocellular marrow.

DISCUSSION

The therapeutic effect of various radiolabeled tumor-associated Mabs has been demonstrated in many animal models;however, the distinctive tumor response for each system andthe occasional disparity between in vitro cytotoxicity and negative therapeutic results in vivo indicate the uniqueness of eachMab-antigen system (23-31). The effectiveness of Mab tumortherapy depends not only on immunological factors such asMab specificity, binding affinity, immunoreactivity, antigendensity, antigen heterogeneity, and Mab and antigen in vivostability, but also on nonimmunological factors such as tumorlocation, tumor blood flow, vascularization, permeability, antigen availability, extracellular fluid circulation, and differentialradiation sensitivity (11, 29, 32-34). Of these, accessibility isparticularly important for brain tumors as the existence of theblood-brain barrier may exclude Mab from the tumor parenchyma.

In order to optimize therapeutic results with minimal toxicity, it is important to characterize the above-mentioned factorsin each system in detail. The in vivo kinetics and tumor localization of the antiglioma extracellular matrix Mab 81C6 inboth s.c. and intracranial xenografts of a human glioma cellline D-54 MG have been well established (11,13, 35). The highconcentration of '"I-81C6 localized in s.c. tumors led us to

perform initial therapeutic experiments in s.c. tumor models innude mice. A clear dose-related antitumor effect for '"I-labeled

81C6 was demonstrated (14). Significant tumor growth delay

Table 2 Dosimetry of Mabs in D-54 MG ¡ntracranialtumor-bearing nude rats

TumorBrainHeartLungLiverKidneySpleenBone

marrowMuscle131I-labeled

81C6158533282442307287223606161"'I-labeled45.616831202389233235206734109

2908

on August 9, 2019. © 1988 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

RADIOIMMUNOTHERAPY OF INTRACRANIAL GLIOMA XENOGRAFT

and regression were achieved in animals receiving 250, 500,and 1000 ¿tCiof 131I-labeled 81C6. Biodistribution data per

formed in parallel with survival studies indicated a radiationdose delivered to tumor of 9719 rads, at 1000 /¿Ci,a dose whichwas 35% higher than the dose predicted by extrapolation fromprior trace-labeled localization studies.

Compared to the s.c. model, intracranial D-54 MG xeno-grafts in athymie rats showed a much lower radiolabeled 81C6peak level (2.5% versus 15% injected dose/g tumor), with amarked intratumoral Mab distribution difference (35). In intracranial xenografts, the expression and availability of antigenfor binding was greatest within central as compared to peripheral regions (35). Also found was the regional variation in bothcapillary permeability and capillary surface area within theintracranial D-54 MG gliomas. In contrast, peripheral regionsof s.c. xenografts have the highest level of antibody uptake.These data suggest a variability of antigen expression betweenthe two systems and/or the possibility of persistent blood-brainbarrier in the intracranial tumor periphery, reflecting host-tissue effects. With such information at hand, it is imperativeto pursue therapeutic approaches in the intracranial model.

In this study, a therapeutic effect was seen for 81C6 only inanimals carrying intracranial xenografts with an average tumorvolume approximately one-fifth the size (16-20 mm3 versus 93mm3) that would cause animal death, or about two tumor

doubling time periods (5 days) ahead of control animal mediansurvival day (16 days). This result correlated well with previouss.c. therapy data, which showed that the tumor growth delaycaused by localized 13IIradiation will not be apparent until 3-

4 days after treatment initiation. The radiation delivered byradiolabeled Mab is cumulative rather than instantaneous, aswhen given by external beam irradiation. Tumor size in experiment 1 was apparently too large since only one-half of a tumordoubling time (1.8 day) was available for the radiation effect tobe evident. The four animals with slightly longer survival timeswere not enough to change the statistical significance. Muchimprovement was observed in experiments 2 and 3 in whichsufficient time was given for the accumulating radiation to takeeffect.

Another intriguing explanation is the size dependence oftumor uptake and/or response which has been observed previously not only in our s.c. therapy experiment (D-54 MG tumorregression consistently correlated with smaller initial tumorsize) (14) but also by others. For example, in localizationexperiments, Moshakis et al. (36) demonstrated that specificlocalization occurred only in tumors smaller than 200-250 mg.This may be a tumor model-dependent phenomenon since, inother systems, higher antibody concentration was found inlarger tumors (23). However, the distribution of Mabs in tumorsis highly heterogeneous as demonstrated by autoradiography.Perivascular and peripheral areas usually have the highest accumulation in most tumor model systems (29, 35). With largertumor volumes, the increased possibility of hypoxia and necrosis in the hypovascular area will further undermine Mab accessibility to target sites (24, 35) and diminish the effectiveness ofradiation. "Cold regions" of low Mab binding in larger tumors

represent possible trouble spots for Mab therapy (37). Thus,using the average percentage of injected Mab dose per g datafor dosimetry calculations which assume a homogeneous distribution of radionuclides in the tumor may be an oversimplification.

For D-54 MG glioma intracranial xenografts, the combination of small tumor size, the choice of a medium-range ßparticleemitter such as I3II, and the use of antiextracellular matrix Mab

may have synergistic effects for radiation response. Significantsurvival benefit and three apparent cures were obtained despitethe low average dose of 1585 rads delivered by I3ll-labeled 81C6

through the experimental period of 12 days. The average radiusof the 20-mm3 tumors (experiments 2 and 3) is only 1.7 mm,

which falls well within the average and maximal ranges of 0.6and 2 mm for the 13IIßparticles. Despite central localizationof I31l-labeled 81C6 in the intracranial D-54 MG xenografts(35), the peripheral area, even if protected by the residual blood-brain barrier, may still be exposed to sufficient therapeuticradiation. In addition, previous studies have shown that ant ¡extracellular matrix Mab 81C6 is retained for long periods in thetumors (38).

In experiments 2 and 3, a dose of 2.5 mCi '"I-labeled 81C6

resulted in no significant improvement in animal survival compared to a dose of 1.25 mCi. The radiation dose delivered tothe tumor at the higher 13IIdose, if proportional to the injectedMab, should be approximately 3000 rads, which had a dose-dependent tumor growth delay effect in the s.c. xenografts (14).However, the calculated radiation dose doubling may not havebeen possible if saturation of antigen sites had already beenachieved with 1.25 mCi (53 ¿ig)I31l-labeled 81C6. Moreover, if

the radiation level of 1500 rads is on the asymptotic portion ofthe response curve of intracranial D-54 MG tumor, a largeincrease in radiation dose could have minimal benefit (39).

Another significant factor is the possibility of bone marrowtoxicity. Both tumor-specific Mab 81C6 and control Mab 45.6delivered approximately 600-700 rads to the bone marrowduring the experimental period of 12 days. Death from marrowablation in AKR mice was observed with 500-600 rads externalbeam irradiation (24). However, direct examination of bonemarrow in these animals failed to show significant marrowsuppression. The apparent high doses may be due to bloodcontamination during marrow sampling. Nevertheless, by doubling our injected Mab dose (2.5 mCi), a higher toxic dose willbe delivered to the rat marrow. The possible therapeutic benefitof tumor treatment, as demonstrated by the apparently earlyrightward shift of survival curve in the 2.5 mCi I31l-labeled81Co-treated group, may be negated by the acute bone marrowtoxicity that supervened later. Previous studies in our laboratoryhave shown that, by intracarotid injection, 20% more Mabradiation dose can be delivered to the intracranial D-54 MGxenografts with constant normal tissue exposure, including thebone marrow (40). Along with such an approach, we are alsostudying the feasibility of using immunoglobulin Fab andF(ab')2 fragments and blood-brain barrier disruption to further

increase tumor delivery without increasing unnecessary normaltissue exposure. In situations in which large amounts of radio-labeled Mab are needed for curative therapy, posttherapy au-

tologous bone marrow transplantation may be required.In summary, we have demonstrated a therapeutic response

in the D-54 MG intracranial human glioma xenograft modelwith i.v. injection of '" I labeled antiglioma extracellular matrix

antigen Mab 81C6. Predicted dose to intracranial tumor using1 mCi 13ll-labeled 81C6 ranged from 1000 to 2500 rads. No

response with a labeled control antibody indicates that thetherapeutic effect is related to specific localization of 81C6 ingliomas. Despite the fact that a much lower dose was deliveredto intracranial xenografts (1565 rads) than their s.c. counterparts (9719 rads) (14), the much smaller tumor size (20 versus200 mm3) and the use of medium-range ßemitter 13II may

contribute to the survival prolongation and apparent cures inthe intracranial xenografts. Moreover, the lack of dose-dependent response in animals treated with the double dose (2.5 mCi)

2909

on August 9, 2019. © 1988 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

RADIOIMMUNOTHERAPY OF INTRACRANIAL GLIOMA XENOGRAFT

highlighted the importance of extensive analysis of all parameters, immunological as well as nonimmunological, involved inthe determination of Mab localization kinetics. Although weregard these results as encouraging, a word of caution is appropriate with regard to developing from them any general hypothesis of treatment of intracranial tumors with radiolabeled monoclonal antibodies. D-54 MG intracranial xenografts are a relatively permeable i.e. tumor system (17, 35) compared to otherbrain tumor models and may be more permeable to water-soluble molecules than most human brain tumors in patients.Moreover, this xenograft system, because of human speciesspecificity of the Mab used, offered an absolutely tumor-specificmodel system, a phenomenon not yet obtained with Mabscurrently in clinical evaluation in patients. Nevertheless, theseresults do demonstrate the ability to treat intracranial humanglicinias successfully with a single injection of a radiolabeledMab. This model system may allow the study of principlesnecessary to achieve such a goal in patients with malignantgliomas.

ACKNOWLEDGMENTS

The authors would like to thank Thomas G. Hooyman for his superbtechnical assistance.

REFERENCES

1. Day, E. D., Lassiter, S., Woodhall, B., Mahaley, J. L., and Mahaley, M. S.The localization of radioantibodies in human brain tumors. I. Preliminaryexploration. Cancer Res., 25: 773-778, 1965.

2. Miller, R. A., Maloney, D. G., and Warnke, R. Treatment of B cell lymphomawith monoclonal anti-idiotype antibody. N. Engl. J. Med., 306: 517-522,1982.

3. Mach, J-P., Buchegger, F., Forni, M., Ritschard, J., Berche, C., Lumbroso,J. D., Schreyer, M., Girardet, C., Accolla, L. S., and Carrel, S. Use ofradiolabelled monoclonal anti-CEA antibodies for the detection of humancarcinomas by external photoscanning and tomoscintigraphy. Immun»].Today, 2: 239-249, 1982

4. Farrands, P. A., Perkins, A. C., Pimm, M. V., Hardy, J. D., Embleton, M.J., Baldwin, R. W., and Hardcastle, J. D. Radioimmunodetection of humancoloréela!cancers by an anti-tumour monoclonal antibody. Lancet, //.- 397-

400, 1982.5. Larson, S. M., Brown, J. P., Write, P. W., Carrasquillo, J. A., Hellstrom, l.,

and Hellstrom, K. E. Imaging of melanoma with M 31-labeled monoclonalantibodies. J. Nucí.Med., 24:123-129, 1983.

6. Arklie, J., Taylor-Papadimitriou, J., Bodmer, W., Egan, M., and Millis, R.Differentiation antigens expressed by epithelial cells in the lactating breastare also detectable in breast cancers. Int. J. Cancer, 28: 23-29, 1981.

7. Epenetos, A. A., Britton, K. E., Mather, S., Shepherd, J., Granowska, M.,Taylor-Papadimitriou, J., Nimmon, C. C., Durbin, H., Hawkins, L. R.,Malpas, J. S., and Bodmer, W. F. Targeting of iodine-123-labeled tumour-associated monoclonal antibodies to ovarian, breast, and gastrointestinaltumours. Lancet, //.-999-1005, 1982.

8. Hammersmith Oncology Group and Imperial Cancer Research Fund. Antibody-guided irradiation of malignant lesions: three cases illustrating a newmethod of treatment. Lancet, /: 1441-1444, 1984.

9. Kemshead, J. T., Jones, D. II., Goldman, A., Richardson, R. B., andCoakham, H. B. Is there a role for radioimmunolocalization in diagnosis ofintracranial malignancies? Discussion paper. J. R. Soc. Med., 77: 847-854,1984.

10. Davies, A. G., Richardson, R. B., Bourne, S. P., Kemshead, J. T., andCoakham, H. B. Immunolocalization of human brain tumors. In: N. Bleehan(ed.). Tumors of the Brain, pp. 83-99. New York: Springer-Verlag, 1986.

11. Bourdon, M. A., Coleman, R. E., and Bigner, D. D. The potential ofmonoclonal antibodies as carriers of radiation and drugs for immunodetec-tion and therapy of brain tumors. Prog. Exp. Tumor Res., 28: 79-101, 1984.

12. Bourdon, M. A., Coleman, R. E., Blasberg, R. G., Groothuis, D. R., andBigner, D. D. Monoclonal antibody localization in subcutaneous and infra-cranial human xenografts: paired-label and imaging analysis. AnticancerRes., 4: 133-140, 1984.

13. Bullard, D. E., Adams, C. J., Coleman, R. E., and Bigner, D. D. In vivoimaging of intracranial human glioma xenografts comparing specific withnonspecific radiolabeled monoclonal antibodies. J. Neurosurg., 64:257-262,1986.

14. Lee, Y. S., Bullard, D. E., Zalutsky, M. R., Coleman, R. E., Friedman, H.

S., Colapinto, E. V., and Bigner, D. D. Therapeutic efficacy of murinemonoclonal antibody 8IC6 in a human glioma xenograft model. CancerRes., 48: 559-566, 1988.

15. Moseley, R., Zalutsky, M. R., Coakham, H. B., Coleman, R. E., and Bigner,D. D. Distribution of 1-131 81C6 monoclonal antibody (Mab) administeredvia carotid artery in patients. J. Nucí.Med., 28: 603-604, 1987.

16. Lee, Y. S., and Bigner, D. D. Aspects of immunobiology and immunotherapyand uses of monoclonal antibodies and biologic immune modifiers in humangliomas. In: Nicholas A. Vick and Darrell D. Bigner (eds.), NeurologicClinics of North America, pp. 901-917. Philadelphia: W. B. Saunders Co.,1985.

17. Vriesendorp, F. J., Pegram, C. N., Bigner, D. D., and Groothuis, D. R.Concurrent measurements of blood flow and transcapillary transport inxenotransplanted human gliomas in immunosuppressed rats. J. Nati. CancerInst., 79: 123-130, 1987.

18. Giard, D. J., Aaronson, S. A., Todaro, G. J., Arnstein, P., Kersey, J. H.,Dosile, H., and Parks, W. P. In vitro cultivation of human tumors: establishment of cell lines derived from a series of solid tumors. J. Nati. Cancer Inst.,51: 1417-1423, 1973.

19. Greenwood, F. C., Hunter, W. M., and Glover, J. S. The preparation of '"I-labeled human growth hormones of high specific radioactivity. Biochem. J.,«9:114-120,1963.

20. McPherson, G. EBDA/Ligand. Nashville, TN: Biomedicai Computing Technology Information Center, MED-79, 1984.

21. Conover, W. J. Practical Nonparametric Statistics, Ed. 2. New York: JohnWiley and Sons, 1980.

22. Loevinger, R., and Berman, M. A. A schema for absorbed-dose calculationsfor biologically-distributed radionuclides. J. Nucí.Med., / (Suppl.): 7-14,1968.

23. Badger, C. C., Krohn, K. A., Peterson, A. V., Shulman, H., and Bernstein, I.D. Experimental radiotherapy of murine lymphoma with '"l-labelcd anti-Thy 1.1 monoclonal antibody. Cancer Res., 45: 1536-1544, 1985.

24. Badger, C. C., Krohn, K. A., Shulman, H., Flournoy, N., and Bernstein, I.D. Experimental radioimmunotherapy of murine lymphoma with ' " I-labeledanti-T-cell antibodies. Cancer Res., 46:6223-6228, 1986.

25. Bernstein, I. D., Tarn, M. R., and Nowinski, R. C. Mouse leukemia: therapywith monoclonal antibodies against a thymus differentiation antigen. Science(Wash. DC), 207: 68-71, 1980.

26. Drebin, J. A., Link, V. C., Weinberg, R. A., and Greene, M. I. Inhibition oftumor growth by a monoclonal antibody reactive with an oncogene encodedtumor antigen. Proc. Nati. Acad. Sci. USA, 83:9129-9133, 1986.

27. Herlyn, D. M., Steplewski, Z., Herlyn, M. F., and Koprowski, H. Inhibitionof growth of colorectal carcinoma in nude mice by monoclonal antibody.Cancer Res., 40: 717-721, 1980.

28. Kirch, M. E., and Hammerlin, U. Immunotherapy of murine leukemias bymonoclonal antibody. I. Effect of passively administered antibody on growthof transplanted tumor cells. J. Immunol., 127:805-810, 1981.

29. Sands, H., Jones, P. L., Neacy, W., Shah, S. A., and Gallagher, B. M. Siterelated differences in the localization of the monoclonal antibody OX7 inSL1 and SL2 lymphomas. Cancer Immunol. Immunother., 22: 169-175,1986.

30. Young, W. W., and Hakomori, S. I. Therapy of mouse lymphoma withmonoclonal antibodies to glycolipid: selection of low antigenic variants invivo. Science (Wash. DC), 211: 487-489, 1981.

31. Zalcberg, J. R., Thompson, C. H., Lichtenstein, M., and McKenzie, I. F. C.Tumor immunotherapy in the mouse with the use of '"l-lahcled monoclonalantibodies. J. Nati. Cancer Inst., 72: 697-704, 1984.

32. Houghton, A. N., and Scheinberg, D. A. Monoclonal antibodies: potentialapplications to the treatment of cancer. Semin. Oncol., 13: 196-179, 1986.

33. Larson, S. M. Radiolabeled monoclonal anti-tumor antibodies in diagnosisand therapy. J. Nucí.Med., 26: 538-545, 1985.

34. Larson, S. M. A tentative biological model for the localization of radiolabelledantibody in tumor: the importance of immunoreactivity. Nucí.Med. Biol.,13:393-399, 1986.

35. Blasberg, R. G., Nakagawa, H., Bourdon, M. A., Groothuis, D. R., Patlak,C. S., and Bigner, D. D. Regional localization of a glioma-associated antigendefined by monoclonal antibody 81C6 in vivo: kinetics and implications fordiagnosis and therapy. Cancer Res., 47:4432-4443, 1987.

36. Moshakis, V., Mcllhinney, R. A. J., Raghaven, D., and Neville, A. M.Localization of human tumor xenografts after IV administration of radiolabeled monoclonal antibodies. Br. J. Cancer, 44: 91-99, 1981.

37. Humm, J. L. Dosimetrie aspects of radiolabeled antibodies for tumor therapy.J. Nucí.Med., 27: 1490-1497, 1986.

38. Wikstrand, C. J., Mclendon, R. E., Carell, S., Kemshead, J. D., Mach, J-P.,Coakham, H. B., de Tribolet, N., Bullard, D. E., Zalutsky, M. R., and Bigner,D. D. Comparative localization of glioma-reactive monoclonal antibodies invivo in an athymic mouse human glioma xenograft model. J. Neuroimmunol.,/5: 37-56, 1987.

39. Chem, H.-S. G., and Gross, J. F. Intra-arterial infusion of anticancer drugs:theoretic aspects of drug delivery and review of responses. Cancer Treat.Rep., 64: 31-40, 1980.

40. Lee, Y. S., Bullard, D. E., Wikstrand, C. J., Zalutsky, M. R., Muhlbaier, L.H., and Bigner, D. D. Comparison of monoclonal antibody delivery tointracranial glioma xenografts by intravenous and intracarotid administration. Cancer Res., 47: 1941-1946, 1987.

2910

on August 9, 2019. © 1988 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

1988;48:2904-2910. Cancer Res   Yisheng Lee, Dennis E. Bullard, Peter A. Humphrey, et al.   I-labeled Anti-Tenascin Monoclonal Antibody 81C6

131Treatment of Intracranial Human Glioma Xenografts with

  Updated version

  http://cancerres.aacrjournals.org/content/48/10/2904

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/48/10/2904To request permission to re-use all or part of this article, use this link

on August 9, 2019. © 1988 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from