ICANCER RESEARCH 58. 25M4-26(X>, June 15, 1998]

Differential Inhibition of Fluid Accumulation and Tumor Growth in Two MouseAscites Tumors by an Antivascular Endothelial Growth Factor/PermeabilityFactor Neutralizing Antibody1

Jin Cai Luo, Masashi Toyoda, and Masabumi Shibuya2

Department (if Genetics, Institute of Medical Science, University of Tokyo. Minalo-ku, Tokyo 10f<.Japan

ABSTRACT

In the accompanying paper (Luo et al., Cancer Res., 58: 2652-2660,

1998), we demonstrated that vascular endothelial growth factor (VEGF),also designated vascular permeability factor (VPF), significantly accumulated in all mouse malignant ascites tested, suggesting its fundamental rolein ascites tumors. Removal of VEGF may inhibit the development ofascites tumors. In this study, using a goat antimouse VEGF-neutralizingantibody, we tested this hypothesis with two well-defined syngeneic mouseascites tumors: MM2 breast adenocarcinoma and OG/Gardner lym-

phoma 6C3HED (expressing moderate and low levels of VEGF, respectively). This antibody significantly inhibited MM2 and OG cell-free ascitesfluid-induced hyperpermeability of mouse peritoneal microvessels and in

vitro endothelial cell growth. Mice bearing tumors were administered i.p.daily with the antibody or normal goat IgG as controls for 8 days, at dosesof 20-fold (for MM2-bearing mice) or 40-fold (for OG-bearing mice) the

estimated amounts of VEGF that kinetically accumulated in the ascitesfluid after the tumor inoculation. The average volume of ascites fluid,number of tumor cells and leaked RBCs, and the peritoneal microvesselpermeability in MM2-bcaring mice that received the antibody treatmentwere significantly lower than those in the matched controls I/' < 0.01).

Unexpectedly, OG-bearing mice did not show satisfactory response to theanti-VEGF treatment. This discrepancy was not likely due to inadequate

doses or different host immune responses, but it was quite possibly to thedifferent characteristics of MM2 carcinoma and OG lymphoma tumors,the latter being strongly invasive, and/or the existence of an inflammatorymediator(s), such as bradykinin or cytokine(s) other than VEGF. Insummary, our results directly demonstrated, for the first time, differentialroles for VEGF in ascites tumors in vivo and suggest the potential ofVEGF inhibition as a specific therapy for ascites tumors of carcinomaorigin, which are the major cause of the malignant ascites in adulthumans.

INTRODUCTION

The accumulation of malignant ascites fluid is an important causeof morbidity and mortality in patients with intraabdominal tumors.However, the mechanisms involved are not well understood. At leastthree pathological features are supposed for ascites tumor formation:first, a reduced lymphatic recovery system, which was proposed fromthe evidence that tumor cells obstruct the draining lymphatics (Ref. l;this finding was later confirmed by both experimental and clinicalfindings that diaphragmatic lymphatic drainage from the peritonealcavity decreased in mice bearing ascites tumor; Refs. 2-4); second,

angiogenesis, which was associated with the accumulation of malignant ascites based on the observations that a marked peritoneal neo-

vascularization accompanied some ascites tumors and that angiogenesis inhibitor significantly reduced the ascites accumulation after the

Received 12/9/97; accepted 4/20/98.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 with18 U.S.C. Section 1734 solely to indicate this fact.

1This work was supported by Cirant-in-Aid for Special Project Research on Cancer-

Bioscicnce 04253204 from the Ministry of Education and by research grants from theYakult Bioscicnce Foundation and the Mitsubishi Foundation in Japan.

2 To whom requests for reprints should be addressed, at Department of Genetics.Institute of Medical Science, University of Tokyo. Minato-ku. Tokyo, 108. Japan. Phone:81-3-5449-5550; Fax: 81-3-5449-5425.

inhibition of vessel proliferation (5, 6); and finally, and probably themost importantly, hyperpermeability of the microvessels lining peritoneal cavity, which was thought to be essentially responsible for themalignant ascites accumulation and therefore has been extensivelyinvestigated (7-11).

The hyperpermeability of microvessels associated with tumors invivo has been shown to be mediated by a variety of factors, including:(a) many inflammatory mediators, such as prostaglandin (12), leuko-kinins (13), bradykinin (14-16), histamine (17) and nitric oxide (18);(b) several cytokines, such as tumor necrosis factor and interleukin-2(19), transforming growth factor-a (20), and so forth (c) VEGF,' also

known as VPF because it was originally identified as a tumor-secreted

protein that significantly increases vascular permeability (10), about50,000 times more potent than histamine on a molar basis. As abifunctional protein enhancing vascular permeability and stimulatingendothelial growth, VEGF is thought to be responsible for fluidaccumulation and angiogenesis in ascites tumors (21 ). Consistent withthis notion, abundant VEGF activity has been found in guinea pig(10), mouse (22), and human malignant ascites fluid (23), and VEGFprotein has been demonstrated to be concentrated in the leakymicrovessels lining the peritoneal cavities of ascites tumor-bearing

animals (24).In the paper immediately preceding this one (25), we reported that

VEGF significantly accumulated in the malignant ascites of all mousetumors tested, strongly suggesting that VEGF plays a fundamentalrole in ascites tumor formation. However, a vast difference in thelevels of VEGF accumulated in the ascites fluids was observed. Theascites tumors originated from sarcoma and carcinoma expressedVEGF strongly, whereas those of lymphocytic and hematologicaltumor origins expressed relatively low levels of VEGF. These observations suggest that the importance of VEGF varies according totumors origin and the exact role of VEGF in these tumors needs to befurther defined. In this study, using an antimouse VEGF-neutralizingantibody, we addressed this question in vivo. We selected two well-

defined tumors: MM2 (derived from mammary adenocarcinoma, secreting moderate levels of VEGF) and OG (derived from Gardnerlymphoma 6C3HED, secreting relatively low levels of VEGF) asmodels. Because the two tumors are also syngeneic, the experimentalresults of the anti-VEGF treatment may provide an insight into po

tential clinical application.

MATERIALS AND METHODS

Animals, Tumors, and Other Reagents. MM2 and OG tumors are twosyngeneic transplantable mouse ascites tumors that have been well defined(26-30). MM2 tumor originated from a spontaneous mammary adenocarcinoma (26-28), whereas OG tumor is one of the three forms of Gardnerlymphoma 6C3HED, which is sensitive to antitumor agent i.-asparaginase (29,30). They were maintained in the peritoneal cavities of syngeneic, 6-week-oldC3H/He male mice (SLC. Hamamatsu. Japan). MM2 is not particularly inva-

1The abbreviations used are: VPF. vascular permeability factor; VEGF, vascular

endothelial growth factor; r.m-VEGF. recombinant mouse VEGF; pN2. affinity-purifiedantibodies to NH,-terminal peptide of mouse VEGF; HUVEC, human umbilical vein

endothelial cell.

2594

Research. on September 10, 2021. © 1998 American Association for Cancercancerres.aacrjournals.org Downloaded from

INHIBITION 01 ASCIlhS Tl MORS BY AN] I VHlil- ANTIBODY

sive (28) and is grown ¡nascites form. The mice inoculated i.p. with 1.5 X IO6

MM2 cells/mouse died, with a mean survival period of 22 ± 0.5 daysaccompanying peritoneal hemorrhage. In contrast, OG is strongly invasive intothe peritoneal lining tissues (25). The mice inoculated i.p. with 3 x IO7 OG

cells/mouse died, with a mean survival period of IO ±0.5 days, without anaccompanying significant peritoneal bleeding. r.m-VEGF|M and antimouseVEGF-neutralizing antibody AF-493-NA. which has the ability to neutralizethe bioactivity of 10 ng/ml r.m-VEGF at the lowest dosage of 0.35 fig/ml with

a low level of endotoxin, were obtained from R&D Systems (Minneapolis,MN). FITC-D (M, of 71,200) was from Sigma Chemical Co. (St. Louis. MO).When used, FITC-D was dissolved (30 mg/ml) in PBS (pH 7.4) and passed

over a Sephadex 50 column (Amersham Pharmacia Biotech. Uppsala. Sweden)to remove any low molecular weight contaminants in advance. The bradykinin-

quantifying kit Markit A was from Dainippon Seiyaku (Osaka. Japan).Collection of Ascites Fluid. Immediately after a mouse was sacrificed, 2

ml of cold PBS (pH 7.4) were injected i.p.. the contents of the peritoneal cavitywere mixed, and the ascites fluid was harvested to the fullest extent possible.The exact volume of the ascites fluid was recorded by subtracting the 2 ml ofPBS injected from the total fluid volume recovered. The total number of tumorcells and RBCs were counted. After protease inhibitors (final concentrations:phenylmethylsulfonyl fluoride, 0.35 mg/ml: aprotinin. 200 kallidinogenase-inactivating units/ml) were added to the ascites fluid, cell-free ascites fluid wasobtained by centrituging at 180 X # for 20 min at 4°C.For the quantitation of

ascites bradykinin, instead of the above protease inhibitors, an inhibitor cocktail (16) containing 1 mM phenylmethylsulfonyl fluoride, 0.05% polybrene.1.5% EDTA, 1 mM Captopril, and l mM carboxypeptidase N inhibitor wasadded to the ascites fluid. One-mi aliquots of the cell-free ascites fluids werecentrifuged at 15,000 x g for 5 min at 4°C.and the supernatants were storedat -80°C for subsequent assays. Bradykinin was quantified according to themanufacturer's instructions (Dainippon Seiyaku. Osaka. Japan).

Cytotoxicity Test of MM2 Cells with Neutralizing Antimouse VEGFAntibody. The fractionated MM2 cells, prepared as described before, weregrown in MEM-2%- PCS in a humidified atmosphere of 5% CO, and 95% airat 37°C.To determine the effect of neutralizing antimouse VEGF antibodyAF-493-NA on the proliferation of MM2 cells in vitro, 2 X IO4 cells/well in

triplicate were cultured in a 24-well microtiter plate (Sumilon. Tokyo, Japan)in 0.5 ml/well fresh MEM-2% PCS either containing or not containing 10

jig/ml AF-494-NA neutralizing antibody for 72 h. Thereafter, cell number and

viability were determined by counting nonstained cells using a Coulter counterafter the addition of 0.1 ml of 0.5% trypan blue to each well.

Hyperpermeability Analysis of Microvessels Lining Mouse PeritonealCavity with FITC-D as Tracer. For in vivo experiments, normal or ascitestumor-bearing mice were injected i.v. with 10 mg/mouse FITC-D (M, of

71.200); 2 h thereafter, they were sacrificed with dry ice. and the ascites fluidwas harvested as described above. For the transient permeability experiment ofmouse peritoneal lining vessels. 2 ml of suitably diluted cell-free ascites orr.m-VEGF diluted in 2% BSA/PBS were incubated with anti-VEGF antibodyAF-494-NA or normal goat IgG of the same amount at 25°Cfor 30 min. The

incubated solutions or the controls of 2% BSA/PBS solution were injected i.p.immediately following i.v. injection of FITC-D (10 mg/mouse) into normal

mice; 2 h thereafter, mice were sacrificed, and the peritoneal fluid wasrecovered. The quantitation of the FITC-D in the peritoneal fluid was per

formed with a fluorometric assay. All samples were read at an excitationwavelength of 499 nm and emitted wavelength of 519 nm at pH 7.4-7.8. The

assay accurately detects concentrations in biological fluid as low as 2 ng/ml.For the microscopic visualization of FITC-D in tissue, the excised peritoneal

wall was fixed and mounted after dehydration and clearance, according to aprocedure described before (22).

Endothelial Cell Growth Assay. HUVECs (Morinaga. Kanagawa, Japan)were grown in HCM (Nissui, Tokyo, Japan) containing the cell-free ascites

fluids (without the addition of protease inhibitors), using an cndothelial growthassay as described before (31).

Radio-Receptor Binding Assay. This assay, in which an affinity purifiedantimouse VEGF antibody pN2 was used as capture antibody and I25l-labeled

7N Fit-1 as detection reagent, is described in detail in the accompanying article

(25).In Vivo Treatment with the Anti-VEGF-neutraiizing Antibody. To re

duce the possible interference by humoral immune response due to the administration of the goat protein, we limited the period of the antibody treatment to8 days. The mice injected i.p. with 1.5 X 10" MM2 cells or 3 X IO7 OG cells

were administered i.p. with anti-VEGF antibody AF-493-NA at doses of20-fold (for MM2-bearing mice) or 40-fold (for OG-bearing mice) the VEGF

concentration kinetically accumulated in the peritoneal cavity, as shown in Fig.\A, from the day of tumor transplantation (designated day 0) to day 7. The

10000

a>_ 1000

IIII1

TOO

10]

Fig. 1. Kinetics of VEGF production, tumor expansion,and ascites accumulation in mice inoculated i.p. with MM2or OG ascites tumors. Tumor-bearing mice were killed atsuccessive intervals after i.p. injection of 1.5 x ]Qh MM2or 3 x IO7 OG cells per mouse. After the ascites fluid was

totally harvested, the volume of ascites fluid, the number oftumor cells, and the amount of VEGF were determined. A,VEGF; B, tumor cells; C, ascites fluid, all expressed as theaverage value per mouse (n = 8 for MM2-bearing mice,n = 4 for OG-bearing mice, at each time point). Dala

points, mean; bars, ±SD.

1

1000

U ÓT501

3

iiKi^

I!

100

10

0 5 10 15 20

Days after Tumor Cells Injection

0 5 10 15 20

Days after Tumor Cells Injection

100TI3

I

10

I

0 5 10 15 20

Days after Tumor Cells Injection

2595

Research. on September 10, 2021. © 1998 American Association for Cancercancerres.aacrjournals.org Downloaded from

INHIBITION OF ASCITES TUMORS BY ANTI-VEGF ANTIBODY

BSA VEGF MM2

I

VEGF MM2 OGFig. 2. Inhibitory effects of anti-VEGF antibody on the cell-free ascites-induced vascular hyperpermeahility of the peritoneal cavity and the stimulation of endothelial cell growth.

A, the reduced extravasation of FITC-D into the peritoneal cavity, induced by the cell-free ascites fluids, after the antibody treatment. The experiments were performed as describedin "Materials and Methods." The results are expressed as the percentage of the i.V. injected dose in the peritoneal cavity (bars, SD). B, HUVECs were cultured in HCM medium with

l()ng/ml of r.m-VEGF. with 10 ng of VEGF-containing MM2 orOG cell-free asciles fluid, or these preincubated with I /¿gof antibody AF-493-NA or normal goat IgG. respectively.After 4-day culture, the cell growth was measured by MTT assay (Sigma). The results are the mean of 5 wells for each experiment (bar. SD). Day 12 MM2 or day 8 OG cell-freeascites fluid was used. Asterisks represent significant differences between the results of nonpreincubation (NT) and those of preincubation with normal goat IgG (nig) or the anti-VEGFantibody (NAb; »,P < 0.05; **, P < 0.01).

mice injected i.p. with tumors were treated with an amount of normal goat IgGequal to the amount of the antibody or left intact to serve as the controls. Onday 8, mice were sacrificed or left untreated until natural death.

Protein G Column Filtration. Clarified ascites diluted in cold PBS (pH7.4) were passed twice over a protein G column (Pharmacia) equilibrated withPBS, and the void volume was collected.

Immunoprecipitation Western Blot Analysis. Samples of ascites fluidwere immunoprecipitated with pN2 antibody, and then protein A-Sepharose

beads (Pharmacia Biotech, Uppsala. Sweden) were added. Western blot analysis was performed as described elsewhere (25).

Histology. The peritoneal wall was fixed in 4% paraformaldehyde-PBS.

processed, and embedded in paraffin for routine H&E staining.Statistics. Comparisons of the results with and without the treatment were

performed using an unpaired Wilcoxon Mann-Whitney test. A value of

P < 0.05 was considered significant.

RESULTS

Kinetics of VEGF Accumulation in MM2 and OG Ascites. Ourpreliminary tests found that i.p. inoculation with 1.5 X IO6 MM2 or3 x IO7 OG cells for 7 days led to significant ascites accumulation,

which was used as a marker for evaluating the therapeutic effects ofthe anti-VEGF antibody in the two tumor models. To assess the doseof the anti-VEGF antibody to be tested, we first examined the kinetics

of VEGF accumulation in MM2 and OG ascites and the relationshipamong VEGF and tumor expansion and ascites accumulation. VEGFwas found when net malignant ascites fluid appeared, as early as 4days after i.p. injection of 1.5 X IO6 MM2 or 3 X IO7 OG cells (Fig.

IA and 1C). In both tumor models, the accumulation of total VEGF inthe peritoneal cavity increased progressively, accompanying the tumor expansion and ascites accumulation in a roughly parallel way(Fig. Iß).The amount of VEGF in OG ascites was relatively small,reaching a peak around day 8, approximately 2 days prior to the death

Table I The changes of bradykinin concentration in ascites fluids of tumors aftertreatment

Tumor TreatmenlBradykinin concentration

(ng/mouse) No. of samples

MM2OGNT"nigNAhNTnigNAb2.3 ±0.15*2.51.40.720.660.210.120.120.180.78

0.21777444" NT, no treatment; nig. treated with normal goat IgG; Nab. treated with the anti-

VEGF antibody." Mean ±SD.

of the animal, whereas that in MM2 ascites fluid was quite large,reaching a peak about day 14, approximately 1 week prior to death,and falling slightly thereafter.

Inhibitory Effects of Anti-VEGF Antibody on Cell-free AscitesFluid-induced Peritoneal Vascular Hyperpermeability and in

Vitro Endothelial Cell Growth. The effect of the antibody on endogenous VEGF action in cell-free ascites of MM2 and OG tumorswas first evaluated by testing its capacity to inhibit cell-free ascitesfluid-induced hyperpermeability of microvessels lining the peritoneal

cavity of normal syngeneic mice, using a transient experiment withFITC-D (Mr of 71,200) as tracer, i.p. injection of 2 ml of PBS

solution, including 240 ¿i!of MM2 ascites fluid or 1100 /j.1 of OGascites fluid (each contained 20 ng of VEGF as measured by radio-receptor binding assay) or 20 ng of r.m-VEGF into normal mice for

2 h led to the extravasation of 5.64, 3.44, and 3.20% of i.v. injectedFITC-D tracer into the peritoneal cavity, respectively. In contrast, the

injection of BSA solution with the same amount of protein and thesame volume only led to the extravasation of 1.04% of i.v. FITC-D

(Fig. 2A). Prior to the i.p. injection, preincubation of these solutionswith 2 /ig of AF-493-NA antibody, which well neutralizes the activityof 20 ng of VEGF. abolished OG ascites fluid-induced extravasationof FITC-D and significantly reduced that induced by MM2 ascites

fluid, from 5.64 to 2.24% (P < 0.001). Preincubation with normalgoat IgG did not significantly affect the extravasation of i.v. FITC-D

by these solutions (Fig. 2A).The effect of the anti-VEGF antibody on the action of native VEGF

in cell-free ascites of MM2 and OG tumors was then estimated by

using in vitro cultures of endothelial cells. Its effect on the action ofnative VEGF in mouse tumor ascites had been examined previously(25) by using rat liver sinusoidal endothelial cell. Because the effectof the antibody AF-493-NA on r.m-VEGF was originally estimated by

using in vitro cultures of HUVECs (R&D Systems), in this study weused the same cells to evaluate its effect on the action of native VEGFin cell-free ascites fluids of MM2 and OG tumors. Stimulation ofHUVECs with HCM culture medium including 10 ng of VEGF-

containing MM2 or OG ascites for 4 days produced a significantincrease in cellular DNA contents compared with the unstimulatedcells. MM2 ascites fluid stimulated endothelial cells more stronglythan did 10 ng of r.m-VEGF did, whereas OG ascites fluid stimulatedendothelial cells at the same strength as did 10 ng of r.m-VEGF. Prior

to the addition into the cultures, incubation of these solutions with 1l¿gof the antibody significantly reduced the growth stimulation ofHUVECs induced by these solutions; it completely blocked the 10 ngr.m-VEGF or OG ascites fluid-induced growth stimulation of endo-

2596

Research. on September 10, 2021. © 1998 American Association for Cancercancerres.aacrjournals.org Downloaded from

INHIBITION OF ASCITES TUMORS BY ANTI-VEGF ANTIBODY

•5 1200B

l l

MM2-nu/nu

ÃŒli40

NT nig NAb NT nig NAb NT nig NAb

„c

p20I!o

r

NT nlg NAb

¡L

o

l

£a

2000

1500

1000

500

0

D

l

OG-C3H/H6

Ili100

75

SO

NAb NT nlg NAb NT nlg NAb

Fig. 3. Differential inhibition of MM2 and OG ascites tumors by the anti-VEGF antibody. Mice i.p. inoculated with MM2 or OG tumor cells were administered i.p. daily antibodyAF-493-NA or normal goat IgG from day 0 to day 7 as described in "Materials and Methods." Anti-VEGF treatment significantly inhibited MM2 tumor inoculated both in C3H/He

(A) and BALB/c (nu/nu) mice (B) and also reduced the leakage of RBCs in the peritoneal fluid (C). In contrast, it did not effectively inhibit OG tumor (£)).The values are the averageof eight mice (MM2-C3H/He) or four mice [MM2-BALB/C (nu/nu) and OG-C3H/He] (bars. SD). Asterisks represent significant differences between the results of nontreatment (NT}and those of treatment with normal goat IgG (nlg) or the anti-VEGF antibody (NAb; *, P < 0.05; **, P < 0.01).

thelial cells and inhibited the effect elicited by MM2 ascites fluid to23% of the level of the non-antibody-incubated controls. Preincuba-

tion with normal goat IgG did not significantly affect the growthstimulation by these solutions (Fig. 2B).

Differential Inhibition of Ascites Accumulation and TumorGrowth in Vivo. By in vitro growth assay, we did not find anycytotoxicity of the anti-VEGF antibody AF-493-NA on MM2 and OG

cells at the dose used in this study (data not shown). The antibodyAF-493-NA was administered i.p. daily for 8 days, at doses of 20-fold(for MM2-bearing mice) or 40-fold (for OG-bearing mice) that of

VEGF, which kinetically accumulated in the peritoneal cavity asshown in Fig. \A. This treatment markedly inhibited MM2 tumor but,unexpectedly, did not significantly inhibit OG tumor (Fig. 3).

The average volume of ascites fluid, the numbers of tumor cells andRBCs in MM2-bearing mice that received the anti-VEGF antibodytreatment (0.51 ml; 2.50 X IO8 tumor cells and 1.61 X IO7 RBCs per

mouse) were significantly lower than those in untreated mice or micetreated with the same doses of normal goat IgG (1.16 ml; 5.90 X IO8tumor cells and 1.55 X IO8 RBCs per mouse; P < 0.01). Such

differences were not observed between OG-bearing mice that received

the same treatment and their matched controls (P > 0.05).Consistent with the above results, the vascular permeability of

microvessels lining the peritoneal cavity of mice bearing MM2 tumordecreased significantly in the anti-VEGF antibody-treated animals

compared with that of the matched controls (P < 0.01), as assessedthrough the quantitative assay of the extravasated FITC-D tracer from

circulation into peritoneal cavity (Fig. 4) and the microscopic visualization of hyperpermeable microvessels in the peritoneal wall withFITC-D (Fig. 5). Furthermore, the appearances of pathological mor

phology correlated well with the decreased vascular permeability ofmicrovessels lining the peritoneal cavity. As shown in Fig. 6, themuscular bundles of peritoneal wall from mouse treated with anti-

VEGF antibody are still tightly apposed; in contrast, those from the

control animal have interstitial edema with wide separate muscularbundles cells. However, these effects were not observed in OG-

bearing mice that received the antibody treatment (Fig. 4).The peritoneal lining tissues generally did not become angiogenic

until day 10 after the MM2 tumor was transplanted (data not shown),but local angiogenesis had appeared. However, the local angiogenesiswas hardly visible in the anti-VEGF treated animals (Fig. 5).

To elucidate whether short-term anti-VEGF treatment has a pro

found effect, mice that had received the antibody or normal IgGtreatment for 8 days were left untreated until their natural death. Theanti-VEGF treatment prolonged the life span of MM2-bearing miceby 3-4 days compared to the matched controls. No significant prolongation was observed in OG-bearing mice that received the anti-

VEGF treatment.

MM2 OG

Fig. 4. Reduced microvascular hyperpermeability of the peritoneal cavities in MM2-bearing mice after the anti-VEGF treatment. Normal mice and tumor bearing mice thatreceived 8-day treatment were injected i.V. with 10 mg/mouse FITC-D; 2 h later, theywere killed, and the leaked FITC-D in the peritoneal cavity was measured (see "Materialsand Methods"). The values are the average of 8 mice (MM2) or 4 mice (OG. normal

controls; bara, SD). Asterisks represent significant differences between the results ofnontreatment (NT) and those of treatment with normal goat IgG (nlg) or the anti-VEGFantibody (NAb; *, P < 0.05; **, P < 0.01).

2597

Research. on September 10, 2021. © 1998 American Association for Cancercancerres.aacrjournals.org Downloaded from

INHIBITION OF ASCITES TUMORS BY ANTI-VEOF ANTIBODY

Fig. 5. Microphotographs of reduced microvas-cular hyperpermeability of the peritoneal cavitiesof MM2-hearing mice after the anti-VEGF treat

ment. Normal mice and MM2 lumor hearing micethat received 8-day treatment were injected i.v.with 10 mg/mouse FITC-D. and 2 h later, theperitoneal wall was excised, fixed, and mounted.Unlike in normal controls (Ai, in which FITC-D isalmost confined to vessels, a leakage of tracer, tooextensive to define the vessels, was observed innormal goat IgG-treated mice (B). In contrast, inthe peritoneal wall of antibody-treated mice, thetracer, although leaked lo some extent, partly remained in the vessels and was easily seen (C). D.an ordinary photograph which helps identify thevessels in the peritoneal wall of normal goat IgG-trealed mouse.

Inhibitory Effect on MM2 Tumor by Anti-VEGF TreatmentIndependent of Host Immune Response. To exclude the possibilitythat the significant therapeutic response of the MM2 tumor and theweak response of the OG tumor to the anti-VEGF antibody treatment

was due to different immune responses, using essentially the sameprotocol, we performed the anti-VEGF treatment in athymic mice

[BALB/c (nu/nu), SLC] and obtained comparable results (Fig. 3B).Therefore, we concluded that the therapeutic effect of antimouseVEGF antibody on MM2 ascites tumor is due to the direct binding andabsorption of VEGF, not to the host immune response.

To clarify why the anti-VEGF antibody did not completely abolish

MM2 tumor and why it did not significantly inhibit OG tumor, wefirst investigated whether the dose of the antibody used was inadequate by examining the existence of free VEGF in the ascites fluid,namely those not bound to the anti-VEGF antibody AF-493-NA

administered. As shown in Fig. 7, VEGF markedly decreased in leveland completely disappeared in the passthrough fractions of protein Gcolumns, derived from the ascites fluid of MM2 and OG-bearing mice

given the antibody treatment, respectively. In contrast, VEGF in theascites fluid of MM2 or OG tumor-bearing mice in the controls

basically remained unchanged after passing through the columns.These results indicate a slightly inadequate dose in the antibodytreatment of MM2 but not in OG tumor.

We then tested whether a non-VEGF permeability factor was

present in the two tumor ascites fluids. Because bradykinin, one of thevascular permeability mediators, has been shown to express in theascites of both mouse and human tumors (14-17), we measured its

concentration in the two tumor ascites fluids by using an enzymeimmunoassay. We found that both MM2 and OG tumor ascites fluidcontained some amounts of bradykinin, the level of which did notchange significantly after the antibody treatment (Table 1). Thisfinding suggests that, in addition to VEGF, inflammatory mediators,such as bradykinin play a role, to some extent, in these two ascitestumors.

DISCUSSION

To understand the role of endogenous VEGF in ascites tumor invivo, we chose to study two well-defined mouse ascites tumors MM2

and OG, which are derived from different origins: carcinoma andlymphoma. Previously, we found that mouse ascites tumors of sarcoma or carcinoma origin expressed VEGF at high levels; in contrast,those of lymphocytic or hematological tumor origins expressed VEGFat relatively low levels, indicating that the role of VEGF may differ inthe two tumor types. This was confirmed by the present study.

In accordance with its effect of inhibiting the transient permeabilityof microvessels lining the peritoneal cavity and in vitro endothelialcell growth stimulated by the cell-free ascites fluid, antimouse VEGFantibody AF-493-NA markedly inhibited the MM2 tumor, reducing

the ascites fluid accumulation, RBC leakage, and tumor cell growth(Fig. 3).

Although the antibody completely abolished the transient vascularpermeability and endothelial cell growth in vitro stimulated by OGcell-free ascites fluid, unexpectedly it could not significantly inhibit

OG tumor in vivo (Fig. 3). At least three possibilities may, separatelyor together, account for this contradiction. Two are associated withhighly invasive characteristics of OG lymphoma. First, as a result ofits invasion, the VEGF secreted by tumor cells may become too easilyaccessible to the blood vessels in the peritoneal tissues to be effectively blocked by the anti-VEGF antibody. Second, the invaded tumor

cells may also lead to the accumulation of ascites fluid by causing theobstruction of draining lymphatic vessels.

The third possibility is that inflammatory mediators, such as bradykinin and/or non-VEGF cytokines, are involved in the development

of ascites tumor, especially after the removal of VEGF by the antibody. This is supported by the finding of some amounts of bradykininaccumulated in OG ascites, because such a level of bradykinin mayhave biological function in vivo (15). However, considering thatVEGF is about 40 times more potent than bradykinin at the molar

2598

Research. on September 10, 2021. © 1998 American Association for Cancercancerres.aacrjournals.org Downloaded from

INHIBITION OF ASCITES TUMORS BY ANTl-VEGF ANTIBODY

level and that the concentration of VEGF in OG ascites is about 18ng/ml, nearly 30-fold higher than that of bradykinin (0.62 ng/ml),

bradykinin may only play a minor role, if any, in the ascites formationof OG tumor.

The significant inhibition of tumor by the anti-VEGF treatment

directly suggests a causal role of VEGF for MM2 ascites tumor invivo. Because angiogenesis of the peritoneal wall had not been statistically significant until day 10 after the i.p. injection of 1.5 X IO6

MM2 cells, this inhibitory effect can be mainly attributed to theblockage of the vascular permeability-enhancing activity of endoge

nous VEGF secreted by the tumor cells. In addition, as shown in Fig.5, the treatment delayed the subsequent angiogenesis that may accelerate the process of the disease.

On the other hand, as discussed in the accompanying article (25),high-level VEGF may cause bleeding secondary to the abnormalcoagulation. In this experiment, we found that the anti-VEGF treat

ment markedly reduced the RBC leakage, which is lethal for mouse insome ascites tumors (21). Interestingly, the treatment indeed prolonged the life span of MM2-bearing mice compared to the matched

controls. Because a little free VEGF seemingly remained in the ascitesfluid of MM2-bearing mice that received the anti-VEGF treatment,

increasing the antibody dose may improve the therapeutic effects.However, anti-VEGF treatment may not completely inhibit MM2tumor, because the anti-VEGF antibody could not abolish the stimu-

nlg NAb

>*.»'.i**«*/*fr^

**4f~

iT'ft-S^T^"-5** -

Column

MM2VEGF

9 10 11 12 13 14 15 16

OGVEGF

.-—•-**"* '"* T*- i*

Fig. 6. Histological changes of peritoneal walls in MM2-bearing mice after theAnti-VEGF antibody treatment. A, the microscopic appearance of normal peritoneal wallclose to the peritoneal cavity. B, peritoneal wall from mouse treated with normal goat IgG,showing locally new vessels (arrows), interstitial edema with wide separate muscularbundles, and thickened layer of mesothelial cell. C. peritoneal wall from mouse treatedwith anti-VEGF antibody, showing no evident new vessels, muscular bundles still tightlyapposed. and activated mesothelial cells. All samples were stained with H&E. X 66 (bar,100 /im).

2599

Fig. 7. Efficient absorption of VEGF accumulated in the tumor ascites fluids by theantibody. Samples of ascites fluid and its passlhrough fractions over protein G columnwere subjected to immunoprecipitation Western blot analysis using antibody pN2. Forsamples from MM2 tumor, 0.5 ml of ascites fluid and 1 ml of passthrough were used; forthose from OG tumor, 1 ml of ascites tluid and 3 ml of passthrough were used, nig, normalgoat IgG: NAb, antibody AF-493-NA. Arrows, positions of VEGF molecules (top,

VEGF164; bottom, VEGFI20).

latory effects of the vascular endothelial cells induced by the cell-free

MM2 ascites (Fig. 2) and also because a substantial level of bradykinin was found in the MM2 ascites.

In summary, our findings directly prove, for the first time, thatVEGF plays an important role in one ascites tumor of carcinomaorigin, suggesting that the inhibition of VEGF activity can be used asa specific therapy for this type of tumor, the major cause of adulthuman malignant ascites. Our results also emphasize that the roles ofnon-VEGF pathological factors, such as the characteristics of tumors,

inflammatory mediators and cytokines other than VEGF, and so forth,need to be determined, especially for the malignant ascites of certainlymphoma and hematological tumor origin.

REFERENCES

1. Holm-Nielsen. P. Pathogenesis of ascites in peritoneal carcinomatosis. Acta Palhol.Microbiol. Scand., 33: 10-21, 1953.

2. Feldman, G. B., Knapp, R. C., Order, S. E., and Hellman. S. The role of lymphaticobstruction in the formation of ascites in a murine ovarian carcinoma. Cancer Res.,32: 1663-1666. 1972.

3. Coates, G., Bush, R. S., and Aspin. N. A study of ascites using lymphoscintigraphywith "'""Tc-sulfur colloid. Radiology, 107: 577-583, 1973.

4. Feldman. G. B.. Knapp, R. C. Lymphatic drainage of the peritoneal cavity and itssignificance in ovarian cancer. Am. J. Obstet. Gynecol.. 119: 991-994. 1974.

5. Brown, H. R. Kinetics of angiogenesis in small vessels related to mouse parietalperitoneum. Anal. Ree., 184: 364. 1976.

6. Heuser. L. S., Taylor, S. H.. and Folkman, J. Prevention of carcinomatosis and bloodymalignant ascites in the rat by an inhibitor of angiogenesis. J. Surg. Res., 36:244-250, 1984.Sträube.R. L. Fluid accumulation during initial stages tumor growth. Cancer Res., IX:57-65, 1958.Hirabayashi. K., and Graham, J. Genesis of ascites in ovarian cancer. Am. J. Obstet.Gynecol., 106: 492-497, 1970.Senger, D. R., Galli, S. J.. Dvorak, A. M., Perruzzi, C. A., Harvey. V. S., and Dvorak,H. F. Tumor cells secrete a vascular permeability factor that promotes accumulationof ascites fluid. Science (Washington DC), 219: 983-985, 1983.

Garrison, R. N., Kaelin, L. D.. Heuser, L. S., and Galloway. R. H. Malignant ascitesclinical and experimental observations. Ann. Surg.. 203: 644-649, 1986.Garrison. R. N., Galloway. R. H.. and Heuser, L. S. Mechanisms of malignant ascitesproduction. J. Surg. Res.. 42: 126-132, 1987.Sykes, A. C. Pharmacologically active substances in malignant ascites fluid. Br. J.Pharmacol., 40: 595P-596P, 1970.Greenbaum, L. M., Grebow, P.. Johnston. Prakash, A., and Semente, G. Pepstatin. aninhibitor of leukokinin formation and ascitic fluid accumulation. Cancer Res., 35:706-710, 1975.Maeda, H., Matsumura, Y., and Kalo, H. Purification and identification of[hydroxyproly!3)bradykinin in ascitic fluid from a patient with gastric cancer. J. Biol.Chem., 26.?.- 16051-16054, 1988.

15. Matsumura. Y.. Kimura, M., Yamamoto, T., and Macda, H. Involvement of thekinin-generating cascade in enhanced vascular permeability in tumor tissue. Jpn. J.Cancer Res., 79: 1327-1334, 1988.

16. Matsumura, Y.. Maruo. Kimura. M., Yamamoto, T.. Konno, T.. and Maeda. H.Kinin-generating cascade in advanced cancer patients and in vitro study. Jpn. J.Cancer Res., 82: 732-741, 1991.

Research. on September 10, 2021. © 1998 American Association for Cancercancerres.aacrjournals.org Downloaded from

INHIBITION OF ASCITES TUMORS BY ANTI-VEOF ANTIBODY

17. White. M. J., Miller, F. N.. Heuser. L. S., and Pietsch, C. G. Human malignant ascitesand histamme-induced protein leakage from the normal microcirculation. Microvasc.Res., 35: 63-72, 1988.

18. Maeda, H.. Noguchi, Y., Sato, K., and Akaike, T. Enhanced vascular permeabilityin solid tumor is mediated by nitric oxide and inhibited by both new nitric oxidescavenger and nitric oxide synthase inhibitor. Jpn. J. Cancer Res., 85: 331-334,

1994.19. Ettinghausen, S. E., Puri, R. J., and Rosenberg, S. A. Enhanced vascular perme

ability in organs mediated by the systemic administration of lymphokine-activated killer cells and recombinant interIeukin-2 in mice. J. Nati. Cancer Inst.,80: 177-187, 1988.

20. Ohmura, E., Tsushima. T., Kamiya. Y., Okada, M.. Onoda, N., Shizume, K., andDcmura. H. Epidermal growth factor and transforming growth factor alpha induceascilic fluid in mice. Cancer Res.. SO: 4915-4917, 1990.

21. Nagy. J. A.. Morgan, E. S., Herzberg, K. T., Manseau. E. J., Dvorak, A. M., andDvorak, H. F. Pathogenesis of ascites tumor growth: angiogenesis. vascularremodeling, and stroma formation in the peritoneal lining. Cancer Res., 55:376-385. 1995.

22. Nagy, J. A., Masse, E. M., Herzberg, K. T., Meyers, M. S., Yeo, K. T., Yeo. T. K..Siossat, and Dvorak, H. F., Pathogenesis of ascites tumor growth: vascular permeability factor, vascular hyperpermeability, and ascites fluid accumulation. CancerRes., 55: 360-368. 1995.

23. Yeo, K. T., Wang, H. H., Nagy. J. A., Scioussat, T. M., Ledbetter, S. R., Hoogewerf,A. J., Zhou, Y., Masse, E. M.. Senger, D. R., Dvorak, and Yeo. T. K. Vascularpermeability factor (vascular endolhelial growth factor) in guinea pig and humantumor and inflammatory effusions. Cancer Res., 53: 2912-2918. 1993.

24. Dvorak, H. F., Sioussat, T. M., Brown, L. F., Berse, B., Nagy, J. A.. Ana, S.,Manseau, E. J., Water, L. VD, and Senger, D. R. Distribution of vascular permeabilityfactor (vascular endothelial growth factor) in tumors: concentration in tumor bloodvessels. J. Exp. Med., ¡74: 1275-1278. 1991.

25. Luo, J. C., Yamaguchi, S., Shinkai, A., Shitara. K.. and Shibuya. M. Significantexpression of vascular endothelial growth factor/vascular permeability factor inmouse ascites tumors. Cancer Res., 58: 2652-2660, 1998.

26. Takeuchi, S., Ine, R. F., Inoue, M., Ine, K., Izumi. R., and Nishioka, K. Immuno-logical studies on mouse mammary tumors. II. Characterization of tumor-specificantibodies against mouse mammary tumors. Int. J. Cancer, 4: 130-138, 1969.

27. Nishioka. K., Ine, R. F., Takashi, K., Takeuchi, S. Immunological studies on mousemammary tumors. III. Surface antigens reacting with tumor-specific antibodies inimmune adherence. Int. J. Cancer. 4: 139-149, 1969.

28. Tanino, T., and Tsumita, T. Studies on the tumor-host relationship. Jpn. J. Exp. Med..42: 401-413, 1972.

29. Davies, M. R., Lambert. L. P., Marshall. R. D. L-Asparaginyl-tRNA symhetaseand L-asparaginyl-tRNA synthetase activities of L-asparaginase-sensitive and-resistant forms of the mouse Gardner lymphoma 6C3HED. Clin. Sci., 56:539-545, 1979.

30. Kamisaki, Y., Wada, H., Yagura, T.. Nishimura, H., Matsushima. A., and Inada. Y.Increased antitumor activity of Escherichia coli. L-asparaginase by modification withmonomethoxypolyethylene glycol. Gann, 73: 470-474. 1982.

31. Sawano, A., Takahashi, T.. Yamaguchi. S., Aonuma. M., and Shibuya, M. Fit- 1,but not KDR/Flk-1 tyrosine kinase is a receptor for placenta growth factor, whichis related to vascular endothelial growth factor. Cell Growth Differ.. 7: 213-221,

1996.

2600

Research. on September 10, 2021. © 1998 American Association for Cancercancerres.aacrjournals.org Downloaded from

1998;58:2594-2600. Cancer Res   Jin Cai Luo, Masashi Toyoda and Masabumi Shibuya  Growth Factor/Permeability Factor Neutralizing Antibodyin Two Mouse Ascites Tumors by an Antivascular Endothelial Differential Inhibition of Fluid Accumulation and Tumor Growth

  Updated version

  http://cancerres.aacrjournals.org/content/58/12/2594

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/58/12/2594To request permission to re-use all or part of this article, use this link

Research. on September 10, 2021. © 1998 American Association for Cancercancerres.aacrjournals.org Downloaded from