Response of intracerebral human glioblastoma xenografts to multifraction radiation exposures

$: Response of intracerebral human glioblastoma xenografts to multifraction radiation exposures$
B

Ntemapt

fonu1anp

sn4

a

Int. J. Radiation Oncology Biol. Phys., Vol. 66, No. 1, pp. 263–270, 2006Copyright © 2006 Elsevier Inc.

Printed in the USA. All rights reserved0360-3016/06/$–see front matter

doi:10.1016/j.ijrobp.2006.05.010

IOLOGY CONTRIBUTION

RESPONSE OF INTRACEREBRAL HUMAN GLIOBLASTOMA XENOGRAFTSTO MULTIFRACTION RADIATION EXPOSURES

TOMOKO OZAWA, M.D., PH.D.,* BRUCE A. FADDEGON, PH.D.,† LILY J. HU, M.S.,*ANDREW W. BOLLEN, M.D., D.V.M.,‡ KATHLEEN R. LAMBORN, PH.D.,*

AND DENNIS F. DEEN, PH.D.*†

*Brain Tumor Research Center, Department of Neurological Surgery, †Department of Radiation Oncology,and ‡Department of Pathology, University of California San Francisco, San Francisco, CA

Purpose: We investigated the effects of fractionated radiation treatments on the life spans of athymic rats bearingintracerebral brain tumors.Methods and Materials: U-251 MG or U-87 MG human glioblastoma cells were implanted into the brains ofathymic rats, and the resulting tumors were irradiated once daily with various doses of ionizing radiation for5 consecutive days or for 10 days with a 2-day break after Day 5.Results: Five daily doses of 1 and 1.5 Gy, and 10 doses of 0.75 and 1 Gy, cured some U-251 MG tumors. However,five daily doses of 0.5 Gy increased the survival time of animals bearing U-251 MG tumors 5 days without curingany animals of their tumors. Ten doses of 0.3 Gy given over 2 weeks extended the lifespan of the host animals9 days without curing any animals. For U-87 MG tumors, 5 daily doses of 3 Gy produced an increased lifespanof 8 days without curing any animals, and 10 doses of 1 Gy prolonged lifespan 5.5 days without curing anyanimals. The differences in extension of life span between the 5- and 10-fraction protocols were minor for eithertumor type.Conclusion: The finding that the U-251 MG tumors are more sensitive than U-87 MG tumors, despite the factthat U-251 MG tumors contain many more hypoxic cells than U-87 MG tumors, suggests the intrinsiccellular radiosensitivities of these cell lines are more important than hypoxia in determining their in vivoradiosensitivities. © 2006 Elsevier Inc.

Human brain tumor xenografts, Fractionated radiation therapy, Athymic rats.

Rzlastit

astsittc

Ata

A

INTRODUCTION

ext to surgery, radiation therapy is the most beneficialherapy for patients with malignant gliomas (1–3). How-ver, despite surgery, radiation therapy and subsequent che-otherapy, most malignant gliomas recur, and patients usu-

lly die within a year (4–7). The overall prognosis foratients with malignant gliomas has not changed much inhe past three decades (8).

There have been continuing laboratory and clinical ef-orts to develop useful treatment strategies that combinether forms of treatment with radiation therapy for malig-ant gliomas. Various chemotherapeutic agents have beensed for this purpose, including chloroethylnitrosoureas (9–1), paclitaxel (12–14), cis-platinum (15, 16), and the poly-mine inhibitor �-difluoromethyl ornithine (17, 18). Yetone of these combined therapies have significantly im-roved the outcome provided by radiation therapy alone.

Reprint requests to: Dennis F. Deen, Ph.D., Brain Tumor Re-earch Center, University of California San Francisco, 505 Par-assus Avenue, U-378, San Francisco, CA 94143-0520. Tel: (415)76-4590; Fax: (415) 476-9687; E-mail: [email protected] by the National Institute of Health Grants NS-42927

nd CA-85356 and the Phi Beta Psi Sorority.

263

ecently, however, several clinical studies show that temo-olomide treatment combined with radiation therapy pro-ongs patient lifespan as compared with radiation therapylone (19–22), and this combination is rapidly becomingtandard treatment for patients with glioblastoma. Becausehese new combination therapies are being developed, it ismportant to establish a preclinical model with which to testhem.

Cell lines derived from human gliomas are readily growns subcutaneous tumors in athymic mice, and these modelystems have been used for studying the response of humanumors to fractionated radiation therapy (23–26). However,ubcutaneous models do not permit evaluation of treatment-nduced toxicity to the brain tissue that is normally adjacento brain tumors or the impact of alterations of normal brainissue on the response of the brain tumor to therapy. Toircumvent these shortcomings, investigators have used in-

cknowledgments—The authors thank Ms. Raquel A. Santos forechnical assistance and Ms. Ilona Garner for excellent editorialssistance.

Received Feb 17, 2006, and in revised form April 13, 2006.ccepted for publication May 11, 2006.

tsscoset

vhicitaCeoMibrtgai

C

oBmdmactat

I

Lk�tskdhCwnefc

pwRms

I

pawbtacUgbbbna

R

SivstvCAtpoCa

(mecactfSetm

R

cMbto

264 I. J. Radiation Oncology ● Biology ● Physics Volume 66, Number 1, 2006

racerebral tumor models to evaluate preclinical therapeutictrategies. The intracerebral 9L rat brain tumor model is atandard model that has been used to investigate drugs inombination with single doses of radiation for the treatmentf gliomas (9, 27–29). The 9L model also has been used totudy chemotherapeutic agents such as 1.3-bis (2-chloro-thyl)-1-nitrosourea (30) and bleomycin (31) in combina-ion with fractionated radiation therapy.

Our use of human brain tumor xenograft models to in-estigate combined modalities involving radiation has beenindered by the need to maintain athymic animals in ansolated facility to prevent unwanted infections that couldonfound data, because these animals have compromisedmmune systems. We have developed a rat irradiation sys-em for athymic rats using a cesium irradiator in the isolatednimal facility at the University of California San Francisco,omprehensive Cancer Center, and we used this system tovaluate the radiation sensitivity of two relatively commonrthogonal human glioblastoma xenograft models, U-251G and U-87 MG, to multifraction radiotherapy. We first

nvestigated a 10-fraction protocol given over 2 weeksecause this simulates the first third of a typical humanadiation protocol and is a practical scheme to conduct inhe laboratory. We also studied a five-fraction protocoliven daily for 5 days because this requires animals to benesthetized fewer times, and it is more convenient for thenvestigators.

METHODS AND MATERIALS

ell culturesU-251 MG and U-87 MG human glioblastoma cell lines were

btained from the Department of Neurological Surgery Tissueank at the University of California, San Francisco. Cells wereaintained as exponentially growing monolayers in complete me-

ium consisting of Eagle’s minimal essential medium supple-ented with 10% fetal calf serum and 1% nonessential amino

cids. Cells were cultured at 37°C in a humidified atmosphereontaining 95% air and 5% CO2. Cells were harvested byrypsinization, washed once with Hanks’ Balanced Salt Solutionnd resuspended in Hanks’ Balanced Salt Solution for implanta-ion.

n vitro cell irradiationCells were seeded into T-75 flasks 24 h before irradiation.

og-phase cells were irradiated with 3, 4.5, or 6 Gy using a 150Vp X-ray machine (Philips, Hanover, Germany) at a dose rate of1.3 Gy/min. Immediately after irradiation, cells were rinsed and

rypsinized. Cell survival was determined using a previously de-cribed colony forming efficiency (CFE) assay (32). Briefly,nown numbers of cells were added to six-well plastic cultureishes (Falcon Plastic, Oxnard, CA) containing medium to whicheavily irradiated feeder cells had been added 24 h in advance.ells were then incubated for 2–3 weeks at 37°C, and coloniesere stained with methylene blue (0.66% solution in 95% etha-ol). Colonies containing at least 50 cells were counted and platingfficiencies were calculated as the ratio of the number of coloniesormed to the number of cells seeded. Surviving fraction (SF) was
alculated as the plating efficiency of treated cells divided by the d
lating efficiency of control cells. The 95% confidence intervalsere calculated by Fieller’s theorem, as described previously (33).adiation dose–response curves were fit to the linear quadraticodel: SF � exp(��D � �D2), where SF is the fraction of cells

urviving a dose D, and � and � are derived empirically (34).

mplantation of intracerebral tumors in athymic ratsFive-week-old male athymic rats (rnu/rnu, homozygous) were

urchased from Harlan (Indianapolis, IN) and were housed underseptic conditions, which included filtered air and sterilized food,ater, bedding, and cages. Tumor cells were implanted into therains of rats as previously described (35). Briefly, rats were anes-hetized with an intraperitoneal injection of ketamine (60 mg/kg)nd xylazine (7.5 mg/kg) and were injected slowly with 10 �L ofell suspension (2 � 106 cells for U-251 MG and 5 � 106 cells for-87 MG) into the right caudate-putamen using an implantableuide-screw system. This procedure resulted in 100% tumor takeased upon our historical data obtained from 264 untreated ratsearing U-251 MG intracerebral tumors and 181 untreated ratsearing U-87 MG intracerebral tumors. The University of Califor-ia, San Francisco, Institutional Animal Care and Use Committeepproved all protocols used in the animal studies.

at irradiation geometry and dosimetryA rat was placed 16.3 cm from a cesium-137 source (J. L.

hepherd & Associates, San Fernando, CA) and shielded with anron collimator that limited radiation to a narrow 1-cm wideertical beam (Fig. 1A, B). A 5-cm thick T-shaped Cerrobendhield abutted the iron shield and was used to reduce the dose tohe rat’s throat. This Cerrobend shield was placed between the twoertical halves of the iron collimator with the upper edge of theerrobend shield 13 cm above the floor of the irradiation chamber.n anesthetized rat was placed onto a piece of Styrofoam set on

op of a laboratory jack. The rat’s head rested on a 5-mm thickiece of plastic, placed tightly against the shield, and the bottomsf the eyes were aligned with the horizontal edge of the T-shapederrobend shield (Fig. 1C). The posterior edges of both eyes wereligned 2 mm to the right of the vertical slit in the iron shield.

Dose was measured 1 cm from the T-shaped Cerrobend shieldat the approximate center of the intracerebral tumor). Measure-ents were done with Kodak XV film (Eastman Kodak Co., Roch-

ster, NY). The film was calibrated against a Farmer air-ionizationhamber, which in turn was calibrated for exposure in a cesium beamt an accredited dosimetry calibration laboratory. Accuracy of the filmalibration was �5%. Figure 2 shows the dose rate profile relativeo the top edge of the T-shaped Cerrobend shield (0 cm) and 1 cmrom the T-shaped Cerrobend shield. When a rat was positioned ontyrofoam set on top of the laboratory jack, the tumor’s extremedges were located between approximately 0.6 and 1.6 cm abovehe top edge of the Cerrobend shield. The dose rate at the tumor’sidpoint (1.1 cm) was calculated to be 247 cGy/min.

at irradiationWe conducted experiments in which athymic rats bearing intra-

erebral human tumors were irradiated once daily for 5 days,onday through Friday, or once daily for 10 days with a 2-day

reak after Day 5. All rats received the first radiation dose whenumors were approximately 50 mg in size for the 5-day treatmentr 20 mg in size for the 10-day treatment. The radiation dose was
elivered to alternating sides of the head on each day.

aaatoe

HU

n

FdshwareC2head is placed tightly against the Cerrobend shield.

FcrmG0

F(baG

265Fractionated radiotherapy of human glioblastoma xenografts ● T. OZAWA et al.

The rats were injected intraperitoneally with 15 mg/kg of ket-mine and 1.9 mg/kg of xylazine. Immediately after injection,nimals were placed into an anesthesia chamber for 5 min to benesthetized with 2% isoflurane. Next, the rats were removed fromhe chamber and positioned in the irradiator as described previ-usly. Estimated tumor size was determined from previously gen-rated growth curves (35).

istopathology and immunochemistry in intracerebral-251 MG and U-87 MG tumorsWe examined the brains of animals that did not exhibit physical/

eurologic symptoms by Day 90 for evidence of intracerebral

ig. 3. In vitro survival curves for U-251 MG cells and U-87 MGells. The error bars represent a 95% confidence interval. Theadiation dose–response curves were fit to the linear quadraticodel: SF � exp(��D � �D2). The � and � values were 0.52y�1 and 0.0058 Gy�2 for U-251 MG cells and �0.034 Gy�1 and

ig. 2. Vertical profile of dose rate 1 cm from the T-shaped shieldwhere the approximate center of the intracerebral tumor woulde) as measured with calibrated, radiographic film. The shieldingdequately protects the animal, while permitting a dose rate of 247y/min to be delivered to the center of the tumor.

ig. 1. (A) Schematic of a cross-section through the cesium irra-iation device showing the general relationship of the animal to theource. (B) The radiation set-up showing the laboratory jack thatolds the animal, the iron shielding, the T-shaped Cerrobend shieldith its upper edge 13 cm above the floor of the cesium irradiator,

nd the 5-mm thick plastic head rest. (C) View of an anesthetizedat placed on laboratory jack. The anterior-bottom edges of bothyes are aligned with the top horizontal edge of the T-shapederrobend shield and the posterior edges of the eyes are positionedmm to the right of the horizontal irradiation slit. The animal’s

.0085 Gy�2 for U-87 MG cells, respectively.

tbttccy

ith(TtmwcbmpptFomegi

S

r

rrmpctdtta

I

MCtfc

HM

fst

FUpmfit

U

U

tt


umors. Whole brains were removed and immersion fixed in 10%uffered formalin for 24 to 48 h. After fixation, a single 5-mmhick section of brain containing the tumor was placed in a plasticissue holder and processed for paraffin embedding. Eight-mi-rometer thick coronal sections were cut and placed on polylysine-oated glass microscope slides, which were stained with hematox-lin and eosin using standard staining methodology.To evaluate tumor hypoxia in both U-251 MG and U-87 MG

ntracerebral xenografts, untreated rats were euthanized at selectedime points during the tumor growth. Tumors were labeled with theypoxic cell indicator pimonidazole. Pimonidazole hydrochlorideHypoxyprobe-1; Natural Pharmacia International, Inc., Researchriangle Park, NC) was dissolved in normal saline to a concen-

ration of 20 mg/mL. Rats were injected intraperitoneally with 60g/kg of pimonidazole 1.5 h before euthanasia. Anesthetized ratsere then perfused with 10% buffered formalin and whole brains

ontaining a tumor were removed. Brains were kept in 10%uffered formalin overnight and then embedded in paraffin. Eight-icrometer sections were deparaffinized, incubated at room tem-

erature for 40 min with a 1:50 dilution of mouse anti-hy-oxyprobe-1 antibody (Natural Pharmacia International, Inc.), andhen incubated for 10 min with a biotin-sp-conjugated anti-mouse(ab=)2 antibody. Antigens were visualized with horseradish per-xidase and diaminobenzidine (DAKO, Carpinteria, CA) and he-atoxylin was used as a final counterstain. A neuropathologist

valuated the immunohistopathology including the estimated de-ree of necrosis and the percentage of pimonidazole-positive cellsn the tumor.

tatistical analysisFor each experiment and each radiation dose, the Wilcoxon

ank-sum test was used to compare survival between rats that

Table 1. Histopathologic evaluation of tumor hypoxia in humanglioblastoma xenografts*

Dayspost-implantation

Intracerebral tumor Pimonidazole

Size (mm) Necrosis (%) Positive cells (%)

-251 MGDay 18 2.0 � 2.0 10 10Day 21 2.5 � 2.5 10 10Day 25 4.5 � 3.0 25 20Day 31 5.0 � 4.0 40 30Day 31 8.5 � 5.5 40 30Day 34 4.5 � 4.0 10 20Day 34 7.0 � 5.5 25 25

-87 MGDay 13 1.5 � 1.0 0 0Day 13 2.0 � 2.0 0 0Day 15 4.0 � 3.0 0 0Day 15 4.0 � 3.0 0 0Day 17 3.0 � 2.5 0 0Day 17 4.0 � 4.0 0 0Day 20 2.5 � 2.0 0 0Day 20 4.0 � 4.0 0 0Day 22 7.5 � 5.0 0 0Day 22 6.5 � 4.0 0 0Day 22 8.5 � 7.0 0 0

* Intracerebral tumor-bearing rats were euthanized during theumor growth on the day indicated, and all brains containing brain

tumors were collected and sectioned for immunohistochemistry.

eceived radiation and a control group. Because only rats in theadiation treatment groups were cured of their tumors, no adjust-ent for censoring was required. These results are included to

rovide an indication of the degree of difference between theontrol and treated groups and no adjustment was made for mul-iple comparisons. The experiments were not designed to proveefinitively that any one treatment group was statistically superioro its control. Because the control results were highly reproducible,he number of control animals was kept small, as consistent withnimal use guidelines.

RESULTS

n vitro radiosensitivity of human glioblastoma cellsFigure 3 shows the in vitro cell survival curves for U-251G and U-87 MG cells that were irradiated and assayed byFE. The U-251 MG cells were more sensitive to X-rays

han U-87 MG cells; the dose required to reduce survivingraction to 0.1 was approximately 3.4 Gy for U-251 MGells and approximately 5.4 Gy for U-87 MG cells.

ypoxia in intracerebral U-251 MG and U-87G tumorsWe have previously demonstrated marked histologic dif-

erences between U-251 MG and U-87 MG tumors (35),uggesting that the distribution of hypoxic cells throughouthese two tumor types may be different. In the current study,

ig. 4. Histopathology of typical intracerebral U-251 MG and-87 MG tumors at death. (a) Relatively well-circumscribed neo-lasms in the right caudate-putamen. Hematoxylin and eosin at 1�agnification. (b) Hematoxylin and eosin staining at 10� magni-cation. (c) Pimonidazole staining at 10� magnification. T �

umor, N � necrosis, NB � normal brain.

umor necrosis was present in all U-251 MG tumors, and

rwirmptpTcob

T

aedx3

tarcesvoTMrdso

F

aas

E

M

I

E

M

I

Db

f

a


anged from 10% to 40%; no evidence of tumor necrosisas seen in U-87 MG tumors (Table 1). Immunohistochem-

cal staining with the hypoxic cell marker pimonidazoleevealed that tumors derived from U-251 MG cells had pi-onidazole-positive cells surrounding necrotic areas. The

ercentage of pimonidazole-positive cells in the U-251 MGumors ranged from 10% to 30%. We found no evidence ofimonidazole-positive cells in the U-87 MG tumors (Table 1).he pimonidazole staining in U-251 MG tumors stronglyorrelated with necrosis, producing a staining pattern thatutlined the regions of necrosis, but not those of normalrain tissue (Fig. 4).

en-fraction protocolAll animals tolerated 10 fractions of anesthesia and radi-

tion therapy well. Rats were monitored daily and wereuthanized when they exhibited neurologic symptoms in-icative of impending death (35). Rats with U-251 MGenografts that did not receive radiation survived between3 and 43 days (median, 34 days; Table 2). Median survival

Table 2. Days post-tumor implantation to

U-251 MG xenografts

0 Gy 0.3 Gy 0.5 Gy 0

xperiment 1n � 8 n � 8 n � 8

Day 33 Day 39 Day 34Day 33 Day 43 Day 39Day 34 Day 43 Day 39Day 34 Day 43 Day 47Day 34 Day 43 Day 47Day 36 Day 46 Day 49Day 37 Day 48 Day 61Day 37 Day 61 Day 63edian survival and p value†

Day 34 Day 43 (p � 0.01) Day 47 (p � 0.01)ncrease in median animal survival relative to control

9 days 13 days

xperiment 2n � 4

Day 34 DDay 34 DDay 34 DDay 43 D

D

edian survival and p value†

Day 34 Day 43ncrease in median animal survival relative to control

* Tumor-bearing animals received the first radiation dose whenay 18 for U-251 MG xenografts and on Day 12 for U-87 MG xreak after Day 5.

† For each experiment and each radiation dose, survival was comor that experiment using the Wilcoxon rank-sum test.

‡ Rats showing no symptoms by Day 90 were euthanized. No evi
nd therefore, we considered them to be cured.
imes were 43, 47, 43, and 70 days for the 0.3-, 0.5-, 0.75-,nd 1-Gy treatment groups, respectively. Two animals thateceived 0.75 Gy per treatment and three animals that re-eived 1 Gy per treatment lived to Day 90, when they wereuthanized. These rats did not show any adverse neurologicymptoms, and histologic examination of their brains re-ealed no evidence of either tumor formation at the site ofriginal implantation or toxicity in normal brain tissue.herefore, we considered these rats to be cured. For U-87G xenografts, survival times of unirradiated animals

anged from 19 to 23 days, with a median survival of 23ays. Tumor-bearing rats that were given 10 1-Gy dosesurvived between 26 to 36 days and had a median survivalf 28.5 days (Table 2).

ive-fraction protocolAll animals receiving five daily doses of radiation toler-

ted the radiation therapy well and were monitored dailynd euthanized as described previously. Table 3 showsurvival times for intracerebral tumor-bearing rats that re-

ization after 10 fractions of radiotherapy*

U-87 MG xenografts

1 Gy 0 Gy 1 Gy

n � 6 n � 8Day 19 Day 26Day 22 Day 27Day 23 Day 28Day 23 Day 28Day 23 Day 29Day 23 Day 30

Day 33Day 36

Day 23 Day 28.5 (p � 0.01)

5.5 days

n � 8Day 39Day 48Day 48Day 70Day 70Cured‡

Cured‡

Cured‡

0.045) Day 70 (p � 0.01)

36 days

erebral tumors were approximately 20 mg in size, which was onfts. Animals were irradiated once daily for 10 days with a 2-day

etween those animals that received radiation and the control group

of tumor was observed histopathologically in any of these animals,

euthan

.75 Gy

n � 7ay 36ay 39ay 43ay 43ay 48

Cured‡

Cured‡

(p �

9 days

intracenogra

pared b

dence

cxUrTw11gwfrnefct

p

itttiaortdtottxf

pgtc

E

M

I

E

M

I

D

f

a


eived graded daily doses from 0 to 1.5 Gy (U-251 MGenografts) or 0, 2, and 3 Gy (U-87 MG xenografts). For-251 MG xenografts, median survival times for treated

ats were longer than those of rats in the 0-Gy control group.he median increases in survival time relative to controlsere 5, 13, 15.5, and �52.5 days for the 0.5-, 0.75-, 1-, and.5-Gy treatment groups, respectively. Two animals in theGy treatment group and 3 animals in the 1.5-Gy treatment

roup did not show any adverse neurologic symptoms andere euthanized on Day 90. There was no evidence of tumor

ormation at the site of original implantation in any of theseats, nor was there any histologic evidence of toxicity inormal brain tissue. Therefore, these animals were consid-red cured. For U-87 MG xenografts, median survival timesor treated animals were prolonged 2 and 8 days (relative toontrols) for the 2-Gy and 3-Gy treatment groups, respec-ively.

DISCUSSION

We have successfully developed fractionated irradiation

Table 3. Days post-tumor implantation to

U-251 MG xenografts

0 Gy 0.5 Gy 0.75 Gy 1 Gy

xperiment 1n � 4 n � 7 n � 7

Day 35 Day 39 Day 47Day 42 Day 41 Day 47Day 42 Day 47 Day 55Day 42 Day 47 Day 55

Day 47 Day 55Day 47 Day 63Day 49 Day 67

edian survival and p value†

Day 42 Day 47(p � 0.10)

Day 55(p � 0.01)

ncrease in median animal survival relative to control5 days 13 days

xperiment 2n � 6 n � 5

Day 31 Day 33Day 33 Day 33Day 35 Day 52Day 38 Cured‡

Day 39 Cured‡

Day 39edian survival and p value†

Day 36.5 Day 52(p � 0.31

ncrease in median animal survival relative to control15.5 days

* Tumor-bearing animals received the first radiation dose whenay 24 for U-251 MG xenografts and on Day 17 for U-87 MG x† For each experiment and each radiation dose, survival was com

or that experiment using the Wilcoxon rank-sum test.‡ Rats showing no symptoms by Day 90 were euthanized. No evi

nd therefore, we considered them to be cured.

rotocols for athymic rats bearing two different types of M

ntracerebral tumors. The shielding devised for the irradia-ions was effective in permitting radiation to reach the brainumors in the experimental animals while protecting otherissues, including the eyes and oropharynx. None of therradiated animals showed any symptoms normally associ-ted with radiation exposure such as eye damage, diarrhea,r skin/mucosa ulcerations. Seven of 76 (�10%) irradiatedats showed a temporary weight loss of 5 to 7 g between thehird and fifth day of irradiation, regardless of radiationose, and all these rats regained the weight shortly thereaf-er. However, this temporary weight loss was also seen in 4f 34 (�12%) unirradiated rats. We concluded that thisemporary weight loss was likely due to anesthesia and noto irradiation. Overall, the rats tolerated repeated ketamine,ylazine, and isoflurane anesthesia well during the course ofractionated radiation therapy.

One goal of these studies was to determine a radiationrotocol for each tumor type that causes reproducible tumorrowth delay of approximately 1 week without curing anyumors. We plan to use these protocols in future studies thatombine radiation with other therapeutic agents. For U-251

ization after 5 fractions of radiotherapy*

U-87 MG xenografts

1.5 Gy 0 Gy 2 Gy 3 Gy

n � 6 n � 6 n � 6Day 21 Day 24 Day 31Day 24 Day 26 Day 32Day 24 Day 26 Day 32Day 24 Day 26 Day 32Day 24 Day 26 Day 32Day 25 Day 36 Day 37

Day 24 Day 26(p � 0.01)

Day 32(p � 0.01)

2 days 8 days

n � 6Day 39Day 48Day 88Cured‡

Cured‡

Cured‡

�Day 89(p � 0.01)

�52.5 days

erebral tumors were approximately 50 mg in size, which was onfts. Animals were irradiated once daily for 5 consecutive days.etween those animals that received radiation and the control group

of tumor was observed histopathologically in any of these animals,

euthan

)

intracenograpared b

dence

G tumors, a growth delay of approximately 1 week was

afadprsmdti

mUctgpoi�mpt

d

totbcotiihcs

cpUttmCtTts

1


chieved by using either a daily dose of 0.3 Gy after 10ractions or 0.5 Gy after 5 fractions. For U-87 MG tumors,daily dose of 1 Gy after 10 fractions produced a growth

elay of 5.5 days, and a daily dose of 3 Gy after 5 fractionsroduced a growth delay of 8 days. We did not conductepeat experiments for U-87 MG tumors because animalurvival times were relatively constant in the first experi-ent. Our data did not show much difference in growth

elay between 5 fractions and 10 fractions for either tumorype, possibly owing to the different tumor sizes that wererradiated.

In this study, we found that U-251 MG tumors are muchore sensitive to fractionated radiation therapy than are-87 MG tumors, using either the 5- or 10-fraction proto-

ol. For example, using the 10-fraction protocol, U-251 MGumors irradiated with daily doses of 1 Gy produced arowth delay of 36 days, whereas the same irradiationrotocol used in U-87 MG tumors resulted in a growth delayf only 5.5 days (Table 2). Moreover, a daily dose of 1.5 Gyn the 5-fraction protocol, produced a growth delay of

52.5 days with 50% tumor cure in the U-251 MG tumorodel, whereas a daily dose of 3 Gy in the 5-fraction

rotocol produced a growth delay of only 8 days and noumor cure in the U-87 MG tumor model.

Our previous studies have shown that there are marked
ifferences in the histopathologic features of these two a
REFEREN

Chem Med 1980;37:169–181.

1

1

1

1

1

1

1

1

umor types (35). U-251 MG–derived tumors are composedf a mixture of malignant spindle and epithelioid cells, andhese tumors exhibit a large area of necrosis and an irregularorder with adjacent brain tissue. In contrast, U-87 MGells form relatively well-circumscribed tumors composedf malignant spindle cells in compact fascicles, and theseumors exhibit little or no necrosis. When we measuredntracerebral tumor hypoxia for these tumors using a mod-fied comet assay (36), U-251 MG tumors showed moreypoxia than did U-87 MG tumors of similar size. This wasonfirmed using the hypoxia-specific immunohistochemicaltain pimonidazole (Table 1 and Fig. 4).

Because it has been known for many years that hypoxicells are more resistant than their well-oxygenated counter-arts to ionizing radiation (37), we expected to find that-251 MG tumors were more resistant to radiation therapy

han U-87 MG tumors. However, this turned out not to behe case, possibly because U-251 MG cells are inherentlyore sensitive that U-87 MG cells to radiation. Our in vitroFE assay indicated that U-251 MG cells are more sensitive

han U-87 MG cells to X-rays by a factor of �1.6 (Fig. 3).hese findings suggest that the sensitivity of these in vivo

umors to radiation reflects their intrinsic cellular radiosen-itivity more than their differences in tumor hypoxia. This is
n interesting finding that merits further study.
CES

1. Taillibert S, Pedretti M, Sanson M. [Therapeutic strategies andprospects of gliomas]. Presse Med 2004;33:1278–1283.

2. Shibamoto Y, Yamashita J, Takahashi M, et al. Supratentorialmalignant glioma: An analysis of radiation therapy in 178cases. Radiother Oncol 1990;18:9–17.

3. Shrieve DC, Loeffler JS. Advances in radiation therapy forbrain tumors. Neurol Clin 1995;13:773–793.

4. Tsien C, Gomez-Hassan D, Ten Haken RK, et al. Evaluatingchanges in tumor volume using magnetic resonance imagingduring the course of radiotherapy treatment of high-gradegliomas: Implications for conformal dose-escalation studies.Int J Radiat Oncol Biol Phys 2005;62:328–332.

5. Prestwich RJ, Sivapalasunrtharam A, Johnston C, et al. Sur-vival in high-grade glioma: A study of survival in patientsunfit for or declining radiotherapy. Clin Oncol (R Coll Radiol)2005;17:133–137.

6. Ammirati M, Vick N, Liao YL, et al. Effect of the extent ofsurgical resection on survival and quality of life in patientswith supratentorial glioblastomas and anaplastic astrocytomas.Neurosurgery 1987;21:201–206.

7. Barker FG 2nd, Prados MD, Chang SM, et al. Radiationresponse and survival time in patients with glioblastoma mul-tiforme. J Neurosurg 1996;84:442–448.

8. Oertel J, von Buttlar E, Schroeder HW, et al. Prognosis ofgliomas in the 1970s and today. Neurosurg Focus 2005;18:e12.

9. Barker M, Deen DF, Baker DG. BCNU and X-ray therapy ofintracerebral 9L rat tumors. Int J Radiat Oncol Biol Phys1979;5:1581–1583.

0. Leenhouts HP, Chadwick KH, Deen DF. An analysis of theinteraction between two nitrosourea compounds and X-radia-tion in rat brain tumour cells. Int J Radiat Biol Relat Stud Phys

1. Nelson DF, Curran WJ Jr., Scott C, et al. Hyperfractionatedradiation therapy and bis-chlorethyl nitrosourea in the treat-ment of malignant glioma—possible advantage observed at72.0 Gy in 1.2 Gy B.I.D. fractions: Report of the RadiationTherapy Oncology Group Protocol 8302. Int J Radiat OncolBiol Phys 1993;25:193–207.

2. Schuck A, Muller SB, Kohler A, et al. Combined radioche-motherapy with paclitaxel in the treatment of malignant gli-oma. Strahlenther Onkol 2002;178:486–490.

3. Fountzilas G, Karavelis A, Capizzello A, et al. Radiation andconcomitant weekly administration of paclitaxel in patientswith glioblastoma multiforme. A phase II study. J Neurooncol1999;45:159–165.

4. Gupta N, Hu LJ, Deen DF. Cytotoxicity and cell-cycle effectsof paclitaxel when used as a single agent and in combinationwith ionizing radiation. Int J Radiat Oncol Biol Phys 1997;37:885–895.

5. Ferrer S. [Radiation therapy with synchrotron radiation andcisplatin-based chemotherapy as a treatment of gliomas]. MedClin (Barc) 2005;124:271–273.

6. Wolff JE, Wagner S, Sindichakis M, et al. Simultaneousradiochemotherapy in pediatric patients with high-grade gli-oma: A phase I study. Anticancer Res 2002;22:3569–3572.

7. Tsukahara T, Tamura M, Yamazaki H, et al. The additiveeffect of alpha-difluoromethylornithine (DFMO) and radiationtherapy on a rat glioma model. J Cancer Res Clin Oncol1992;118:171–175.

8. Prados MD, Wara WM, Sneed PK, et al. Phase III trial ofaccelerated hyperfractionation with or without difluorometh-ylornithine (DFMO) versus standard fractionated radiotherapywith or without DFMO for newly diagnosed patients withglioblastoma multiforme. Int J Radiat Oncol Biol Phys 2001;
49:71–77.

1

2

2

2

2

2

2

2

2

2

2

3

3

3

3

3

3

3

3


9. Balana C, Lopez-Pousa A, Berrocal A, et al. Phase II study oftemozolomide and cisplatin as primary treatment prior toradiotherapy in newly diagnosed glioblastoma multiforme pa-tients with measurable disease. A study of the Spanish Med-ical Neuro-Oncology Group (GENOM). J Neurooncol 2004;70:359–369.

0. Chang SM, Lamborn KR, Malec M, et al. Phase I study oftemozolomide and thalidomide with radiation therapy fornewly diagnosed glioblastoma multiforme. Int J Radiat OncolBiol Phys 2004;60:353–357.

1. Raizer JJ, Malkin MG, Kleber M, et al. Phase I study of28-day, low-dose temozolomide and BCNU in the treatmentof malignant gliomas after radiation therapy. Neuro-oncol2004;6:247–252.

2. Stupp R, Mason WP, van den Bent MJ, et al. Radiotherapyplus concomitant and adjuvant temozolomide for glioblas-toma. N Engl J Med 2005;352:987–996.

3. Eshleman JS, Carlson BL, Mladek AC, et al. Inhibition of themammalian target of rapamycin sensitizes U87 xenografts tofractionated radiation therapy. Cancer Res 2002;62:7291–7297.

4. Hasegawa M, Niibe H, Mitsuhashi N, et al. Hyperfractionatedand hypofractionated radiation therapy for human malignantglioma xenograft in nude mice. Jpn J Cancer Res 1995;86:879–884.

5. Schueneman AJ, Himmelfarb E, Geng L, et al. SU11248maintenance therapy prevents tumor regrowth after fraction-ated irradiation of murine tumor models. Cancer Res 2003;63:4009–4016.

6. Sun LQ, Coucke PA, Mirimanoff RO, et al. Fractionatedirradiation combined with carbogen breathing and nicotin-amide of two human glioblastomas grafted in nude mice.Radiat Res 2001;155:26–31.

7. Douple EB. Treatment of an intracerebral rat brain tumor with

Corynebacterium parvum and radiation. Natl Cancer InstMonogr 1977;46:179–184.

8. Kimler BF, Vats TS, Morantz RA, et al. Response of the 9Lrat brain tumor to combination treatment with radiation andbleomycin. Int J Radiat Oncol Biol Phys 1981;7:1069–1074.

9. Wierowski JV, Thomas RR, Wheeler KT. DNA repair kineticsin mammalian cells following split-dose irradiation. RadiatRes 1984;98:242–253.

0. Wheeler KT, Kaufman K. Influence of fractionation scheduleson the response of a rat brain tumor to therapy with BCNU andradiation. Int J Radiat Oncol Biol Phys 1980;6:845–849.

1. Kimler BF, Martin DF, Evans RG, et al. Combination ofradiation therapy and intracranial bleomycin in the 9L ratbrain tumor model. Int J Radiat Oncol Biol Phys 1990;18:1115–1121.

2. Ozawa T, Lu RM, Hu LJ, et al. Radiopotentiation of humanbrain tumor cells by sodium phenylacetate. Cancer Lett 1999;142:139–146.

3. Gupta N, Lamborn K, Deen DF. A statistical approach foranalyzing clonogenic survival data. Radiat Res 1996;145:636–640.

4. Fertil B, Malaise EP. Inherent cellular radiosensitivity as abasic concept for human tumor radiotherapy. Int J RadiatOncol Biol Phys 1981;7:621–629.

5. Ozawa T, Wang J, Hu LJ, et al. Growth of human glioblas-tomas as xenografts in the brains of athymic rats. In Vivo2002;16:55–60.

6. Wang J, Klem J, Wyrick JB, et al. Detection of hypoxia inhuman brain tumor xenografts using a modified comet assay.Neoplasia 2003;5:288–296.

7. Gray LH, Conger AD, Ebert M, et al. The concentration ofoxygen dissolved in tissues at the time of irradiation as a factor
in radiotherapy. Br J Radiol 1953;26:638–648.

Response of intracerebral human glioblastoma xenografts to multifraction radiation exposures

Documents

Transcript of Response of intracerebral human glioblastoma xenografts to multifraction radiation exposures