[CANCERRESEARCH55, 360-368, January 15, 19951

ABSTRACT

Previous studies have shown that accumulation of tumor ascites fluid

results in large part from increased permeability of peritoneal lining

vessels (Nagy et aL, Cancer Res., 49: 5449-5458, 1989; Nagy et aL, Cancer

Res, 53: 2631—2643,1993). However, the specific microvessels renderedhyperpermeable have not been identified nor has the basis of peritoneal

vascular hyperpermeability been established. To address these questions,

TA3/St and MOT carcinomas, well-characterized transplantable murinetumors that grow in both solid and ascites form, were studied as model

systems. Ascites tumor cells ofeither type were injected ip. into syngeneicA/Jax and C3Heb/FeJ mice, and ascites fluid and plasma were collected atintervals thereafter up to 8 and 28 days, respectively. Beginning severaldays after twnor cell injection, small blood vessels located in tissues lining

the peritoneal cavity (mesentery, peritoneal wall, and diaphragm) becamehyperpermeable to several macromolecular tracers (IssI@humanserumalbumin, FITC-dextran, colloidal carbon, and Monastral Blue B). Increased microvascular permeability correlated with the appearance in

ascites fluid of vascular permeability factor (VPF), a tumor cell-secretedmediator that potently enhances vascular permeability to circulating macromolecules. VPF was measured in peritoneal fluid by both a functionalbioassayand a sensitiveimmunofluorometricassay. The VPF concentralion, total peritoneal VPF, ascites fluid volume, tumor cell number, andhyperpermeability of peritoneal lining microvessels were found to increase in parallel over time. The close correlation of peritoneal fluid VPFconcentration with the development of hyperpermeable peritoneal mlcrovessels in these two well-defined ascites tumors suggests that VPFsecretion by tumor cells is responsible, in whole or in part, for initialingand maintaining the ascites pattern of tumor growth.

INTRODUCTION

Injection or metastasis of tumor cells into body cavities often leadsto an “ascites―pattern of growth in which malignant cells grow as acell suspension in peritoneal fluid in the apparent absence of connective tissue stroma of the type that characterizes solid tumors (1—8).

Earlier studies from our laboratory have been concerned with themechanisms responsible for the ascites tumor phenotype and particularly for the fluid accumulation, often massive, that characterizes thispattern of tumor growth (9, 10). Studies with two syngeneic transplantable murine ascites-producing tumors, MOT mouse ovarian tumor and the TM/St mammary carcinoma, demonstrated that outflow(i.e., the disappearancerate from the peritoneal cavity of i.p.-injected1@I-HSA3 tracer, @d/min)was markedly (S-fold) impeded within 24 h

of i.p. injection of either tumor cell line. Reduced outflow precededany increase in tumor cell number but was not by itself sufficient toprovoke net peritoneal fluid accumulation; apparently, the remainingdrainage capacity of the peritoneal cavity (primarily via diapbragmatic lymphatics) was sufficient to counterbalance normal fluid inflow. In contrast to the rapid decrease in tracer outflow that followedi.p. tumor cell injection, inflow (i.e., the rate of appearance in theperitoneal cavity of i.v.-injected ‘@I-HSAtracer, pi/min) remainedunchanged for some days. However, at 5—7days, at a time whentumor cell number had increased >500-fold, 1251-HSA tracer inflowincreased dramatically (13- to 25-fold above control values). Onlyafter the inflow rate of tracer from plasma to peritoneal cavity hadincreased so that it exceeded the outflow rate was there net peritonealfluid accumulation. The fluid that accumulated in the peritonealcavities of ascites tumor-bearing animals had a protein concentrationseveralfold greater than that of normal pentoneal fluid and came toapproximate 85% of the protein level of plasma, including albumin ina proportion similar to that found in plasma (9). Thus, the fluidaccumulating in the peritoneal cavities of ascites tumor-bearing micewas a plasma exudate, a finding others have also reported (11—18).

While the mechanism(s) by which peritoneal tumor cells inducedan early and dramatic decline in peritoneal outflow remains obscure,we hypothesized a likely explanation for the increased tracer inflowand high protein content of the peritoneal fluid that accumulatedsubsequently. Studies from our laboratory have shown that a widevariety of tumor cells produce and secrete in vitro a cytokine, VPF;VPF is a Mr 34,00042,000 disulfide-bonded dimeric protein (12,19—22).In low nanomolar to picomolar concentrations, VPF (alsoknown as VEGF or vascular endothelial growth factor) increases thepermeability of venules and small veins to plasma proteins with apotency some 50,000 times that of histamine on a molar basis (12).Abundant VPF activity has been found in guinea pig (12) and, morerecently, in human tumor ascites fluids (23).

Taken together, these fmdings suggest that tumor ascites fluidaccumulation might result from increased permeability of the bloodvessels lining the peritoneal cavity and that the hyperpermeabiity ofsuch vessels might be mediated by VPF secreted by ascites tumorcells. To test this hypothesis, we assessed the permeability of theblood vessels of the peritoneal lining to macromolecules at successive

intervals after i.p. injection of either MOT or TA3/St tumor cells andcorrelated these fmdings with measurements of ascites fluid volumes,VPF concentrations, and numbers of peritoneal tumor cells.

Received 6/1/94; accepted 11/11/94.Thecostsof publicationof thisarticleweredefrayedin partby thepaymentof page

charges. This article must therefore be hereby marked advertisement in accordance with18 U.S.C. Section 1734 solely to indicate this fact.

1 This work was supported by NIH research Grants CA-40624 and CA-58845, under

terms of a contract from the National Foundation for Cancer Research, and by the BIHPathology Foundation, Inc. Part One of a series on “Pathogenesisof Ascites TumorGrowth.―

2 To whom requests for reprints should be addressed, at Department of Pathology, Beth

Israel Hospital, 330 Brookline Avenue, Boston, MA 02215.3 The abbreviations used are: HSA, human serum albumin; VPF, vascular permeabi

lilty factor; FITC-D, FITC-dextran; MBB, Monastral Blue B; @ME,@-mercaptoethanol;C-IgG, N-IgG, antibodies to C and N terminal peptides of human and rat VPF/vascularendothelial growth factor, respectively.

360

Pathogenesis of Ascites Tumor Growth: Vascular Permeability Factor, Vascular

Hyperpermeability, and Ascites Fluid Accumulation'

Janice A. Nagy2,Elizabeth M. Masse,Kemp T. Herzberg, Michelle S. Meyers, Kiang-Teck Yeo, Tet-Kin Yeo,Tracy M. Sioussat, and Harold F. Dvorak

Departments of Pathology, Beth Israel Hospital and Harvard Medical School, Boston, Massachusetts 02215

MATERIALS AND METHODS

Tumor Cells. TM/St and MOT tumor cells were the kind gifts of Drs.Barbara Sanford and Gerald Kolodny, respectively. The TM/St ascites tumoris an ascites subline originally derived from a spontaneous mouse mammaryadenocarcinoma found in the A/HeHa strain of mice (24—28).The MOT(mouse ovarian tumor) ascites cell line, an ovarian embryonal cell carcinoma,originated spontaneously in a C3H female mouse (29). It was originallypropagated by s.c. transplantation, where MOT grows in solid form. However,MOTcanalso be maintainedin ascitesformfollowing i.p. injection,exhibitinga behavior much like that of ascites human ovarian cancers with lack ofhematogenous dissemination (30, 31).

Research. on August 24, 2020. © 1995 American Association for Cancercancerres.aacrjournals.org Downloaded from

VPF AND ASCITES TUMOR GROWTH

TA3/St and MOT tumor cells (1 X 106) were passaged weekly in theperitoneal cavities of syngeneic, 5- to 6-week-old female A/Jax and maleC3Heb/FeJ mice, respectively. The MOT tumor grows equally well in bothmale and female mice (29), and males were chosen for our experiments

because of their greater availability from the supplier (The Jackson Laboratory,

Bar Harbor, ME).Collection of Malignant Ascites fluid. At various intervals after i.p.

injection of 1.0 x 106 TA3/St or MOT tumor cells, tumor-bearing or controlanimals were anesthetized with ether, and blood samples were collected byretroorbital puncture into a known volume of heparin for preparation ofplatelet-poor plasma (centrifugation at 15,600 X g for 10 mm at 4°C).Animalswere sacrificed by CO2narcosis. For collection of ascites fluid, 2 ml of Hanks'balanced salt solution were injected i.p., and the contents of the peritonealcavity were mixed by kneading. Peritoneal fluid was then recovered, itsvolume recorded, and tumor cells were counted (9). Prior to centrifugation(160 X g for 20 mm at 4°C),protease inhibitors were added to the followingfinal concentrations: iodoacetamide, 0.37 mg/mI; N-ethylmaleimide, 0.25 mg/ml; PMSF, 0.35 mg/ml; and aprotinin, 210 KIU/ml. Cell-free ascites fluid andplasma were aliquoted and stored at —80°Cfor subsequent assay. The 2-mivolume of Hanks' balanced salt solution, injected to maximize tumor cell andfluid removal from the peritoneum, was subtracted from the total cell-free fluidvolume recovered from the peritoneal cavity at each time interval to yield theactual volume of tumor ascites fluid found in the peritoneal cavity as a functionof time after tumor cell injection. At early time intervals, the volume of tumorascites fluid was also determined by the indicator-dilution method as describedpreviously (9).

Inflow Studies Using lasI@Albuminas Tracer. HSA (Sigma ChemicalCo., St. Louis, MO) was iodinated (Iodogen; Pierce Chemical Co., Rockford,IL) to a specific activity of 0.02—0.05moles ‘@Iper mole HSA; >97% of‘@I-HSAwas precipitatedby 10%trichloroaceticacid. Inflow measurementswere performed by injecting 2.5 X 10@cpm of ‘@I-HSA(in 0.2 ml saline) i.v.At various times thereafter, ascites tumor-bearing or control animals weresacrificed by CO2 narcosis; then the amount of ‘@I-HSAin the plasma andascites fluid was determined, and the transcompartmental tracer inflow rateswere calculated as described previously (9).

In other experiments, cell-free ascites fluid and partially purified VPF weretested for their ability to increase the influx of circulating ‘@I-HSAinto theperitoneal cavity. Immediately following i.v. injection of ‘@I-HSAtracer(2.5 X 106 cpm) into normal nontumor-bearing mice, 1.0 ml of either MOT14-day cell-free ascites fluid (1:10 dilution in 5% BSA; Sigma), heparmnSepharose-purifled murine VPF (diluted in 5% BSA; Ref. 21), or a controlsolution of5% BSA was injected i.p. At various time points thereafter, animalswere sacrificed by CO2narcosis; then pentoneal fluid and blood samples werecollected as described above, and the amounts of ‘@I-HSAin the plasma andperitoneal fluid were determined by gamma spectrometry for calculation ofinflow rates (9).

Vessel Hyperpermeability Studies with Macromolecular FITC-D, Calloidal Carbon, and MBB as Tracers. FITC-D of average molecular weightof 150,000 (mean diameter, 17.4 am) was obtained from Sigma Chemical Co.FITC-D of Mr 5,000,000 (mean diameter, 40 nm) was prepared as described(10). Ascites tumor-bearing or normal mice were injected i.v. with —-0.2@molFITC-D in PBS. After various time periods (5 to 30 mm) following tracerinjection, animals were sacrificed, and the peritoneal fluid was collected. Formicroscopic visualization of FITC-D, the entire peritoneal wall, as well as thediaphragm and mesentery, were rapidly excised and immersed in a 70:30mixture of ethanol:10% formalin (32). Fixation proceeded for 4 h at roomtemperature. Subsequently, tissues were dehydrated in ascending grades ofethanol (70 to 100%) over 24 h and were cleared in xylene (10, 33). Fullthickness segments of the parietal peritoneal wall (up to 1 cm x 1 cm),diaphragm, and mesentery were mounted under coverslips on glass slideswith immersion oil and viewed in a Wild macroscope or in a fluorescencemicroscope.

Colloidal carbon and MBB are tracers that can be visualized in tissues byboth macroscopy and by high resolution light microscopy (34). Ascites tumorbearing or normal control mice were injected i.v. with 0.1 ml of a 1:5 dilution(in PBS) of either MBB (Sigma; Ref. 34) or colloidal carbon (C11/1431A;mean particle diameter, —50am; Guenther Wagner, Germany). One h later,animals were euthanized by CO2 narcosis, and after washing out the peritonealcavity with PBS, portions of the mesentery with attached bowel, parietal

361

peritoneum, and diaphragm were pinned out in 10%formalin for 2—4h at roomtemperature. Tissues were then washed in PBS, dehydrated through a series ofalcohols, cleared in methyl salicylate, and mounted in microscope immersion

oil on 4 x 5-inch glass lantern slides for viewing in a Wild macroscope. Tissuesfrom otheranimalswere similarlyfixed but in paraformaldehyde-glutaraldehyde, washed in cacodylate buffer, and processed for Giemsa-stained, 1-g.@mEpon sections (33, 35).

Immunohistochemical Staining with Antibodies to VPF. At the time thiswork was begun, mouse VPF had not been cloned; therefore, we synthesizeda peptide that corresponded to amino acid residues 1 through 25 of rat VPF,except that an amide replaced the carboxyl group at the COOH terminus:APTFEGEQKAHEVVKFMDVYQRSYC (36). Subsequently, mouse VPF hasbeen cloned (37), and the predicted murine amino acid sequence is 98%identical with that of the rat. Only two conservative substitutions occur withinthe NH2-terminal25 amino acids; Ala10and Va114in rat VPF are replaced bySer10 and 1le14in mouse VPF. Rabbits were immunized at multiple intradermalsites with an emulsion containing 1 mg of rat VPF peptide-keyhole limpet

hemocyanin conjugate (Pierce Chemical Co.; Ref. 21) in complete Freund'sadjuvant; animals were subsequently boosted at 4—6week intervals with anequivalent amount of the same conjugate in incomplete Freund's adjuvant.Collected serum was stored at —20°Cprior to affinity purification on a columnprepared by conjugating 5 mg of peptide to 1 g of CNBr-activated Sepharose(Pharmacia). Typically, 20 ml of antiserum was passed over the column, afterwhich the column was washed thoroughly with PBS. Bound antibody was theneluted with 15 ml of 0.1 Macid glycine buffer, pH 2.5. Eluted antibody wasimmediately diluted in 2 ml of 1 M Tris buffer, pH 8.0, and dialyzed againstPBS. Concentrations of affinity-purified antibody were typically —0.5—1.0mg/mi.

Mouse tissues (parietal peritoneal wall, diaphragm, and mesentery) fromnormal control and ascites tumor-bearing animals were excised, incubated in29 m@tI3MEin PBS for 15 mm, fixed in 10% buffered formalin-29 m@(3MEfor 4 h at room temperature,rinsed in PBS (two times for 30 mm each),dehydrated, and embedded in paraffin. Reduction of tissues with @MEintensified VPF staining without reducing specificity. This fmding is not surprisingin that the antibodies used were directed against a relatively short peptidecorresponding to the NH2 terminus of VPF. The analogous region in intactVPF might be expected to be more accessible to these antibodies followingcleavage of interchain and possibly intrachain disulfide bonds within the VPFmolecule. Immunohistochemistrywas performed on 5-7-gm paraffin sections,using as primary antibody the affinity-purified polyclonal antibody to rat VPFpeptide and an avidin-biotin peroxidase sequence (Vector Lab, Burlingame,CA; Ref. 38). Normal rabbit IgG at an equivalent protein concentration wassubstituted for the primary antibody as a control.

Immunofluorometric Assay for VPF. A sensitive two-site, time-resolvedimmunofluorometric assay has been developed for measuring guinea pig andhuman VPF, using appropriate rabbit antibodies prepared against the VPFCOOH- and NH2-terminal peptides (23, 39). Because the COOH-terminalregion of VPF is highly conserved, we adapted this assay to quantify mouseVPF in tumor ascites fluids by using anti-peptide antibodies directed againstthe human VPF COOH-terminal peptide (C-IgG) as the “capture―antibody.Briefly, microtiter wells were coated with C-lgG as previously described (39).Fifty pi of sample and 50 @tlofassay buffer were added per well and incubatedovernight at 4°Cwith gentle shaking. After washing, 0.5 g@g(in 100 @l)ofbiotinylated anti-peptide antibodies directed against the rat VPF NH2-terminalpeptide (N-IgG) were added per well and mixed for 4 h at room temperature.Next, 100 pi of Eu3@-labeled streptavidin (100 ng/ml; —P7atoms of Eu3@/molecule of streptavidin; Pharmacia LKB Nuclear, Gaithersburg, MD) wereadded and mixed for an additional 2 h at room temperature. After binding andwashing, Eu3@was dissociated from the “sandwichcomplex―with an enhancement buffer containing @3-diketone(Pharmacia LKB Nuclear) to yield a highlyfluorescent chelate that was measured in a time-resolved fluorimeter. Sincerecombinant mouse VPF was not available, we arbitrarily used dilutions ofrecombinant human VPF (prepared as described; Ref. 23) as the calibrators forthe mouse VPF assay. Thus, the amounts and the concentrations of VPF inmouse ascites fluids are expressed as “humanVPF equivalents―(in pmol,nmol/liter, or nmol/tumor cell).

Statistics. Because our data did not conform to bell-shaped distributionswith equal SDs, comparisons were made with the nonparametric KruskaliWallis test and with Dunn's variation of the Bonferroni test (40).

Research. on August 24, 2020. © 1995 American Association for Cancercancerres.aacrjournals.org Downloaded from

‘/PFAND ASCITES TUMOR GROWTH

RESULTS

Growth Patterns of MOT and TA3/St Tumors Injected into thePeritoneal Cavity. After a short lag period, MOT and TA3/St tumorcells grew for a time in logarithmic phase in the peritoneal cavity;thereafter, the growth rate plateaued (Fig. 1, A and B; Ref. 9).Significant net fluid accumulation did not begin for 3—4days (TA3/St) or 6—8days (MOT) but thereafter increased exponentially (Fig. 1,A and B). Fluid accumulation peaked in TM/St animals at 2—3ml byday 7, a day or two prior to animal death; in contrast, fluid accumulation increased progressively in MOT-bearing mice for nearly amonth and to a volume of —25ml. At all times, and in both tumorsystems, tumor cells accounted for 99% of nucleated cells in thepentoneal fluid. However, TM/St and MOT ascites fluids typicallybecame bloody, containing 0.5 to 2.0 X i0@ RBC by days 7 and 27,

respectively; blood loss into the peritoneal cavity was, therefore,significant and contributed importantly to animal death.

Quantitative Measurement of Peritoneal Vessel Hyperpermeability Using 1asI@HSAas Tracer. Consistent with our earlier findings with other macromolecular tracers (9, 10), the inflow rate (p1/min) of ‘@I-HSAincreased dramatically (—10-fold)between days 3and 7 in TM/St ascites tumor-bearing mice and between days 5 and10 in MOT ascites tumor-bearing mice (Fig. 1, C and D).

Macroscopic Identification of Hyperpermeable Peritoneal Lin

IngBloodVesselsUsing ColloidalCarbonand MBBas Tracers.Within a few days of i.p. injection of either MOT or TM/St tumorcells, the peritoneal lining tissues became thickened and hyperemicwith loss of their normal glistening appearance. Macroscopic examination revealed carbon- or MBB-labeled blood vessels beginning by3—4days; at first, labeling was focal, but over the course of the nextseveral days, it came to involve the entire peritoneal surface (Fig. 2;Table 1).

Vessel labeling appeared first and most intensely in the mesentery,where it was confmed to fatty portions that carry the major bloodvessels to the intestines and that are themselves normally well vascularized (Fig. 2, A-F; Table 1). The intervening, fat-free mesenteryremained normally thin, relatively clear, and avascular. Vessel labeling developed slightly later in the peritoneal wall and diaphragm (Fig.2, I-L; Table 1). The blood vessels that labeled initially with carbon or

MBB were of relatively large caliber (Fig. 2, B and E; Ref. 41).

However, at later intervals (e.g., after day 10 in MOT-bearing mice),tracer deposits were noticeably finer and more numerous and appeared predominantly to involve vessels of smaller size (Fig. 2, C, D,and F).

In addition to forming a suspension of tumor cells in the peritonealcavity, both MOT and TM/St tumor cells commonly formed solitarymacroscopic nodules of solid tumor in the peritoneal wall at the siteof the needle track through which tumor cells had been injected i.p.;the blood vessels supplying these solid “needletrack―tumors labeledearlier (33) and more intensively with carbon than did the surroundingperitoneal wall (Fig. 2G). At later stages, TM/St tumor cells alsoformed nodular solid tumor implants in the mesentery and in theparietal peritoneum away from tumor injection sites; these also became vascularized and labeled more intensively with carbon than thesurrounding mesentery (Fig. 2H).

The large and small intestines also contained small blood vesselsthat labeled with carbon (Fig. 2, A-D), but the labeling pattern wasvery different from that observed in the free mesentery, panetalperitoneum, and diaphragm. Carbon labeling of bowel wall vesselswas evident, even in normal mice; in tumor-injected mice, the patternof bowel vessel labeling with carbon was irregular and did not evolvein parallel with that of ascites tumor growth as did labeling in theother tissues lining the peritoneal cavity (Table 1). The relativehyperpermeability (e.g., —-10times that of skeletal muscle vessels) ofnormal alimentary tract vessels to circulating macromolecules is wellknown and is thought to reflect the presence of open endothelial cellfenestrae (42, 43).

Microscopic Identification of Hyperpermeable Blood Vessels ofthe Peritoneal Lining with FITC-D. Within minutes of i.v. injection, FITC-D was detected leaking from scattered blood vessels thatsupplied the mesentery, peritoneal wall, and diaphragm of ascitestumor-bearing animals. F!TC-D extravasation was observed as earlyas 2 days after i.p. tumor cell injection and became much moreextensive at later intervals (Fig. 3, B and C). In contrast, FITC-D didnot leak detectably from comparable vessels in normal control animals over a periodof hours(Fig. 3A).

Microscopic Identification of Hyperpermeable Peritoneal Liiiing Blood Vessels with COllOidal Carbon. Microscopic study of1-@Mfl Epon sections taken from animals injected i.v. with colloidal

carbon confirmed and extended these observations. Rare carbonlabeled vessels were identified microscopically in the mesentery,parietal peritoneum and diaphragm as early as 2 days after i.p. tumorcell injection and thereafter increased dramatically in both numberand labeling intensity (Fig. 4, A-C; Fig. 5, A-D). Accompanying thesemicrovascular changes was significant interstitial edema, particularlyevident in the peritoneal wall (Fig. SB). Surface mesothelial cellsbecame activated (Fig. 5B) and sometimes displayed mitotic figures.At later times, tumor cells attached to the peritoneal and diaphragmatic surfaces as individual cells or as cell clumps (Fig. 5D). mdividual carbon-labeled blood vessels were also observed in the bowelmusculature and submucosa (Fig. 4D); however, as noted by macroscopy (see above), labeling did not differ significantly in number orappearance in tumor-bearing versus control animals.

Immunoperoxidase Staining for VPF. Blood vessels lining theperitoneal walls stained intensely for VPF by day 14 in MOT ascitestumor-bearing mice; however, staining was not evident at earlierintervals up to day 7 (Fig. 5, E and F). Peritoneal wall tumor vesselstaining also became evident in TM/St-bearing mice on day 8 (datanot shown). VPF staining was not observed in peritoneal liningvessels of control mice.

6

Ii

I

it

Fig. 1. Kinetics of tumor cell proliferation, ascites fluid accumulation, and ‘@‘I-HSAinflow rates into the peritoneal cavities of syngeneic mice bearing TA3/St (A and C) orMOT(B andD) ascites tumors.O, tumorcell number;0, volumeof ascitesfluid(ml);A,inflow rates (pi/min) of ‘@I-HSAtracer into the peritoneal cavity.

4 6 8 8 12

Days After ip Tumor Cell In!ection

362

Research. on August 24, 2020. © 1995 American Association for Cancercancerres.aacrjournals.org Downloaded from

VPF AND ASCI1'ES TUMOR GROWTH

Fig. 2. Low power macrographs of leaky blood vessels of mesentery (A-F and H), peritoneal wall (G and 1-K)and diaphragm (L) in mice injected i.p. with either MOT (A-C, E,F, andi-IC)orTA3/St (D, G. H, andL) ascitestumorcells Colloidalcarbon(A-K)or MBB (L) were usedas tracers.A-C, mesenterywith attachedbowel in a normalcontrolC3Heb/FeJmouse(A)as comparedwiththemesenteriesof othermicesacrificedon days6 (B)and 12(C) afteri.p. injectionof 1 X 10@MOTtumorcells.Notelackof labelingof mesentericvessels in (A) but progressively intense vessel labeling at subsequent stages of ascites tumor growth. D, mesentery of an A/Jax mouse 8 days after i.p. injection of 1 X 1O@TA3/Sttumor cells showing extensive vessel labeling, comparable in appearance to that of a 12-day MOT-bearing animal (C). E and F, higher magnification views of carbon-labeled bloodvessels in 6- and 15-day MOT ascites tumor-bearing mice, respectively. At 6 days, relatively large caliber vessels are labeled, whereas at 15 days, many more vessels are labeled, butthey are of smaller size. G, small solid TA3/St tumor which developed in the peritoneal wall at the tumor cell injection site, 6 days. Extensive vessel labeling of this solid tumor nodulewith relative lack of labeling in surrounding peritoneal wall vessels H. three small solid tumor implants in the mesentery 5 days after i.p. injection ofTA3/St tumor cells. Such implantswere fairlycommonin TA3/St-bcaringmice andalwaysdisplayedmorenumerousleakyblood vessels thanthe surroundingmesentery.I-K,peritonealwall vessel labelingin a controlanimal(1)andin othermiceharvestedat 9 and12days,respectively,afteri.p.injectionof MOTtumorcells.Vessellabelingwasnotobservedin controlanimals(1)buthadbecomeintenseby 9 and12 days(I andK).1, MBB labelingof thediaphragmof a TM/ST ascitestumor-bearinganimal,7 days.A-D, X 6-14; E, X 27; F andG, X 15;H, X 30; I-K, X 13—20;1, X 13.

. - .

. % ,1@

..

B@

\ @,;

.‘@@ :1

?@@@

.....:@ •@ 1@. ‘@,@

. •@:@ @.@;

@ S:. •:;‘@@@@

:;‘@: ,@‘1@@

,. @,... .@.;_. —‘S

I-I

363

. @.

. ;@ .

.@,;

_\

Research. on August 24, 2020. © 1995 American Association for Cancercancerres.aacrjournals.org Downloaded from

MOTTA3/StTime

(days)MesenteryDiaphragmPeritoneal wallBowelMesenteryDiaphragmPeritonealwallBowel0

1234678

121520280,0,0

0,0,00,0,0

0,1+,1+2+,3+,2+

4+,4+,4+4+,4+,4+4+,4+,4+3+,4+,4+

3+0,f,0

0,0,00,0,0

f,0,0f,f,f

0,f,04+,4+,4+4+,4+,4+4+,4+,4+

4+0,0,0

0,0,00,0,0

0,f,ff,f,f

f,f,f4+,4+,4+4+,4+,4+4+,4+,4+

4+2+,3+,3+

±,±,±±,0,±

2+,2+,2+3+,3+,3+

1+,4+,3+±,±,±±,±,±±,0,±

1+0,0,0

0,0,0±,±,±2+,±,±3+,±,3+

3+,4+,3+4+,4+,3+0,0,0

0,0,01+,1+,2+

±,±,±2+,±,3+

4+,4+,3+4+,4+,4+0,f,0

0,0,00,0,f±±,0

2+,1+,2+4+,4÷,2+4+,4+,4+±,0,1+

3+,O,2+±,±,1+

2+,2+,±3+,±,1+±,±,3+

±,±,±a

Each data pointrepresentsa separate individual animal. Semiquantitative scaleof 0 to 4+; f, focal±.

@TTT@@@@;-:@ .@- I@@

y@..

VPF AND @scrras TUMOR GROWTH

Table 1 Macroscopic scoring of carbon-labeled blood vessels in the peritoneal lining at various intervals after i.p. injection into syngeneic mice ofMOT orTA3/St ascites tumor cells0

Fig. 3. A-C, distribution of i.v.-injectedFITC-D in the mesenteries of control (A) or ascites tumor-bearing(B and C) mice. A. controlf3Heb/FeJ mouse harvested 5 min after i.v. injection of M, 150,000 FITC-D; tracer remainsconfined to blood vessels. B, 8-day TA3/St ascitestumor-bearing A/Jax mouse harvested 5 mm afteriv. injection of Mr 150,000 FITCD. C, 11-dayMOT ascites tumor-bearingC3HebIFeJmouseharvested 15 mm after i.v. injection of M, 5000FITC-D. Note extravasation of FITC-D frommesenteric vessels in (B) and (C) but not in (A).Not unexpectedly, M, 5000 FITC-D (C) diffusedless extensively from leaky vessels than M,150,000FITC-D(B). D, leakageof M, 5000FITC-D into the mesentery of an otherwise normal @3Heb/FeJmouse 15 mm afteri.p. injectionof 0.5 ml cell-free ascites fluid (diluted 1:1 insaline)harvested from a 12-day MOT tumor-bearing mouse. A-D, X 34.

Vascular Permeability Enhancing Activity of MOT and TA3/StCell-free Ascites Fluid. Experiments were undertaken to determinewhether the ascites fluid induced by peritoneal tumor cells containedmicrovessel permeability enhancing activity, and, more particularly,VPF. We had shown previously that both TM/St and MOT tumorcells synthesized and secreted VPF in vitro and that, like the solubleVPFs secreted by guinea pig, rat, and human tumor cells (12, 21, 44,45), the VPFpresentin mousetumorculturemediumwascompletelyabsorbed by hepann-Sepharose.4

Preliminary experiments demonstrated that cell-free peritoneal fluids from 14-day MOT ascites exhibited substantial vascular permeability-enhancing activity when tested in normal mice by the Milesassay (Refs. 12 and 46; data not shown). Cell-free tumor ascites fluidwas also tested for its capacity to induce increased permeability in theperitoneal lining vessels of normal syngeneic mice (Fig. 6). When14-day cell-free MOT ascites fluid was injected i.p. into normal mice,circulating 1@I-HSA extravasated into the peritoneal cavity 10-timesmore rapidly than in mice that had been injected i.p. with a volumeand protein-matched control solution (CBSA group). Furthermore, 2 hafter i.p. injection, more than 15% of i.v. tracer had accumulated in theperitoneal cavities of mice injected i.p. with MOT cell-free tumorascites fluid as compared with —2%in the CBSA group (P < 0.001;

Fig. 6). All of the permeability-enhancing activity present in bothTM/St and MOT tumor ascites fluid bound to heparmn-Sepharose(data not shown). Mice injected i.p. with partially purified (21) mouseVPF accumulated ‘@I-HSAtracer in the peritoneal cavity in amountsand with kinetics similar to those of mice injected with cell-freeascites fluid (data not shown). Finally, when MOT cell-free ascitesfluid was injected i.p., circulating macromolecular FITC-D tracer wasobserved to extravasate from mesenteric and peritoneal wall microvessels by fluorescence microscopy (Fig. 3D).

Immunofluorometric Assay of VPF in Cell-free Ascites Fluid asa Function of Time after i.p. Injection of MOT or TA3/St TumorCells. VPF was not detectable with our sandwich immunoassay in thesmall (<0.5 ml) amounts of fluid that could be recovered from thepentoneal cavities of normal mice (assay sensitivity, <5 pM). However, VPF was readily measured in cell-free tumor ascites fluid asearly as 2 or 5 days after injection of 106 TA3/St or MOT tumor cells(Fig. 7). Total VPF (pmol) within the pentoneal cavity increasedprogressively through day 7—8(TA3/St) or day 20 (MOD, eventuallyachieving peak levels equivalent to 14—32pmol (Fig. 7, A and B). Theconcentration of VPF, expressed as nmol/liter ascites fluid (or asnmol/tumor cell), reached peak levels on day 3—5 (or day 8) in

TM/St-bearing mice at —10—12nM(or —4X 10 “nmol/tumor cell)and on day 7 (or day 16) in MOT-bearing animals at —-4—5aM (or—-1.0X 10@―nmol/tumor cell; Fig. 7, C and D).4 Unpublished data.

364

Research. on August 24, 2020. © 1995 American Association for Cancercancerres.aacrjournals.org Downloaded from

VPF AND ASCITF.S TUMOR GROWTH

C

. _;3,.@ .‘@;5.-,@ --@

@ , . @:a.0@ ;j.: .,@.@

-@ ,“@

.- .@ 0',@s ,

@ £@. ,@@@

Fig. 4. Giemsa-stained, 1-@.un-thickEpon sections illustrating the microscopic appearance of colloidal carbon labeling of mesenteric (A-C) and large bowel (D) blood vessels incontrol (A) or in TA3/St (B-D) ascites tumor-bearing mice. A, control vessels are neither prominent nor carbon-labeled. B and C, hyperpermeable vessels in mesenteries of A/hz mice7 days after i.p. injection ofTA3/St ascites tumor cells. In (B), several carbon-labeled vessels are indicated with arrows. In (C), a carbon-labeled vessel (V) and two nearby unlabeledvessels (v) are illustrated. D, wall of large intestine of an A/Jax mouse injected i.p. 2 days previously with TA3/St tumor cells. Note carbon-labeled vessel (arrow) in submucosa.Identical vascular labeling was common in control mice, indicating that the hyperpermeabiity of this vessel subset is not tumor-dependent. A and B, X 270; C and D, X 540.

A

@. ‘. @@_av .;.

0 ,,..c..@o@ fa@Jràl@tT@iøC .@,.@

@ EFig. 5. A-D, 1-@.tm-thick,Giemsa-stained Epon sections illustrating microvascular hyperpermeabiity in the peritoneal wall (A-C) and diaphragm (D) of mice bearing MOT (A) or

TA3/St (B-D) ascites tumors. A, normal peritoneal wall with muscle bundies tightly apposed. This normal appearance persisted for up to —6days in C3Heb/FeJ mice injected withMOTascites tumorcells. B, peritonealwall of a 7-day TM/St ascites tumor-bearingmouse exhibits mesotheial cell activation,interstitialedemawith wide separationof individualmuscle bundles, and microvascular hyperpermeability (arrowheads, carbon leakage). Vertical distance in (B) corresponds to that portion of (A) indicated by double-headed arrow. C,higher magnification micrograph of a leaky vessel deeper in the peritoneal wall of another 7-day TA3/St ascites tumor-bearing mouse. D. diaphragm of a TA3/St ascites tumor-bearingmouse at 8 days. Note suspended tumor cells, monolayer of adherent tumor cells (some displaying mitotic figures), and numerous carbon-labeled (arrowheads) superficial blood vessels.E and F, peritoneal walls of MOT ascites tumor-bearing mice stained for VPF by immunohistochemistry. VPF was undetectable at 7 days (E) but was strongly positive by 14 days(F) and thereafter. A, X 220; B, X 430 (bar, 16 p@m); C, X 630; D, X 520; E and F, X 275.

365

t, S

@“

@ ,;

,@@ @t

.3t

#. -.-@ -@@ ‘

;@‘@ , @‘‘-@@@

F

Research. on August 24, 2020. © 1995 American Association for Cancercancerres.aacrjournals.org Downloaded from

0 2 4 6

vpF AND ASCI1'ESTUMORGROWTH

peritoneal lining vessels (Fig. 6). Several independent pieces of cvidence indicate that most, and perhaps all, of this vascular permealizing activity is attributable to VPF: (a) immuno-fluorometric assayrevealed that VPF accumulated in substantial concentrations in bothMOT and TA3/St cell-free ascites fluids (Fig. 7); these concentrationsare well in excess of those needed to induce increased microvascularpermeability in vivoor to promote the division of cultured endotheialcells in vitro (12, 19, 47, 48); (b) all of the vascular permeabiizingactivity present in MOT ascites fluid was absorbed by heparinSepharose, a property of guinea pig and human VPFs (12, 44).Although the possibility of additional heparmn-binding,vascular permeabiizing factors in tumor ascites fluid has not been excluded, ourdata indicate that tumor-secreted VPF is present in amounts more thansufficient to account for accumulation of tumor ascites fluid; and (c)as previously described in solid and ascites guinea pig tumors (38),VPF protein accumulated in the leaky blood vessels that line theperitoneal cavities of mice bearing both MOT and TA3/St ascitestumors and in amounts that were readily detected by immunohistochemistry (Fig. SF).

Vascular hyperpermeability developed more slowly in ascites tomors than when these same and other tumors grew in solid form (49).Hyperpermeable blood vessels have been documented within a fewhours of tumor cell transplant to the s.c. or intradermal spaces, sites atwhich they form solid tumors (38). In contrast, when MOT or TM/Stcells were injected i.p. to generate ascites tumors, peritoneal liningvessels did not become detectably leaky for 2—3days. Moreover, foran additional several days, the density of hyperpermeable vesselsremained significantly lower in peritoneal vessels associated withascites tumors than was found in their solid tumor counterparts,whether growing in the s.c. space, in the peritoneal wall at tumor cellinjection sites, or in mesentery as tumor implants (Fig. 2, G and H;Refs. 33 and 38).

Several explanations could account for the slower kinetics withwhich vascular hyperpermeabiity developed in ascites as comparedwith solid tumors. One possibility is that the mesothelial cell layer thatlines the peritoneal cavity and separates ascites tumor cells from theunderlying peritoneal lining blood vessels might serve as a barrier toVPF, preventing it from reaching blood vessels of the peritoneallining. However, this explanation is unlikely because i.p. injection ofVPF into normal mice4 or guinea pigs (12) elicited a prompt (within

Fig. 6. Inflow of circulating ‘@I-HSAfrom the plasma into the peritoneai cavities ofotherwise normal mice at various intervals after i.p. injection of 1.0 ml of 1:10 diluted14-day MOT cell-free ascites fluid (•)or a 5% solution of BSA (0). Fourteen-day MOTcell-free ascites fluid elicited a striking and statistically significant (a, P < 0.05; b.P < 0.001) increase in ‘@I-HSAinflow into the peritoneal cavity, whereas the volume andprotein-matched solution of BSA did not. Data expressed as mean ±SEM. n = 3 miceat each time point.

DISCUSSION

MOT100

10

.01

0.75

0.50

0.25

4!-'

2@

A8 0 4 8 12 16

Days After Ip Tumor Cell injection

Fig. 7. Kinetics of appearance of immunofluorometric assay-detectable VPF in theperitoneal cavities of mice injected i.p. with TA3/St (A and C) or MOT (B and D) ascitestumorcells. A andB, VPF is expressedas totalpmolofVPF presentwithinthe peritonealcavity at various intervals after i.p. tumor cell injection (A). C and D, concentration ofVPF expressed as 10―X nmol/tumor cell (0) or as nmol/liter of ascites fluid (•)atsuccessive intervals after i.p. tumor cell injection.5 A. M. Dvorak, unpublished data.

366

0 30 60 90 120Minutes

Extravasation of circulating molecules from blood vessels is afunction of both local blood flow and microvascular permeability.Microvessels lined by a continuous endothelium, such as those thatsupply the mouse peritoneum (42), normally retain high-molecularweight proteins almost quantitatively. Therefore, whereas extravasation of water and other small circulating molecules into tissues may besignificantly affected by blood flow, extravasation of proteins andother macromolecules is, within limits, almost entirely dependent onmicrovascular pore size. Like the normal peritoneal microvasculature,that supplying the peritoneal walls of mice bearing either TA3/St orMOT tumors is largely of the continuous type;5 however, thesetumor-supplying vessels extravasated substantial amounts of circulating ‘@I-HSA,FITC-D, colloidal carbon, and MBB. This increasedextravasation cannot be attributed to vessel disruption. Vessels wereintact by electron microscopy,5 and significant leakage of erythrocytesdid not occur until after day 7 in the case of TM/St tumors and afterday 28 in the case of MOT tumors (9, 41); i.e., at times well after ourmeasurements of tracer extravasation (up to day 7 in the case ofTM/St tumors and up to day 10 in the case of MOT tumors).Therefore, the extensive extravasation of macromolecular tracers described here in ascites tumor-bearing mice is largely attributable toalterations in peritoneal microvessel permeability (9, 10).

Vascular hyperpermeability was quantified by measurement of theincreased inflow rate of 1@I-HSA into the pentoneal cavity (Fig. 1).Extravasation of FITC-D, colloidal carbon, and MBB was visualizeddirectly by macroscopy and/or microscopy (Figs. 2—5).Small vesselsof the free mesentery were the first to become leaky to these tracers,followed shortly thereafter by vessels in the pentoneal wall anddiaphragm (Table 1). Subsequently, as will be discussed in an accompanying paper (41), new blood vessels were generated in pentoneallining tissues, and these also were found to be hyperpermeable tomacromolecular tracers.

The hyperpermeability of peritoneal lining microvessels associatedwith ascites tumor growth correlated closely with the accumulation ofperitoneal fluid and also with the appearance in cell-free ascites fluidof an activitythatenhancedthepermeabilityof normalskinand

Research. on August 24, 2020. © 1995 American Association for Cancercancerres.aacrjournals.org Downloaded from

\‘PFAND ASCITES TUMOR GROWTH

a few minutes) increase in the permeabililty of pentoneal liningvessels, a response with kinetics similar to that when VPF wasinjected intradermally in the Miles assay.

A second possibility is that tumor cells expressed little VPF immediately after i.p. injection but, after a time, began to express more.Against this possibility, however, are the results of Northern analysesperformed on both TM/St and MOT tumor cells that were recoveredfrom the peritoneum at successive intervals after i.p. transplant; theseanalyses showed no significant changes in VPF mRNA levels overtime (data not shown).

Perhaps the most likely explanation for differences in the kineticsof vascular hyperpermeabiity relates to the very different microenvironments that tumor cells experience in solid and ascites tumors. Atearly intervals after injection of equivalent numbers of cells, tumorcell density is considerably higher in the s.c. space than in theperitoneal cavity. When injected s.c., tumor cells are largely confinedto the injection site because their dispersion is physically restrained bysurrounding tissues. As a consequence, the VPF secreted by injectedtumor cells is also confined, and its local concentration will, therefore,be relatively high. By contrast, when injected i.p., tumor cells maydistribute throughout the entire peritoneal cavity, and the VPF thatthey secrete will be similarly dispersed. Therefore, if individual tumorcells synthesize and secrete equivalent amounts of VPF over timeafter s.c. or i.p. injection, then it follows that local concentrations ofVPF will be initially much higher in developing solid tumors than inascites tumors. Only after a period of days, when the pool of peritoneal tumor cells has expanded, will peritoneal VPF concentrationsreach levels that were achieved much earlier at local sites of solid

tumor implantation. Also to be considered, of course, is clearance ofboth tumor cells and VPF from s.c. or i.p. injection sites. We have notmade such clearance measurements for either solid or ascites tumors.However, we have previously shown that macromolecular outflow ismarkedly reduced within hours of i.p. tumor cell injection; however,

the capacity of the peritoneal cavity to clear macromolecules is suchthat the outflow rates of molecules the size of VPF and smaller werenot restricted for a period of at least several days (9).

In summary, our data indicate that tumor cells planted in theperitoneal cavity secrete VPF and thereby render hyperpermeable themicrovasculature supplying peritoneal lining tissues (mesentery, peritoneal wall, and diaphragm). Hyperpermeability of vessels lining theperitoneal cavity, in turn, leads to extravasation of a plasma-richexudate which extends beyond the peritoneal lining tissues and intothe peritoneal cavity itself. With respect to vascular hyperpermeability, therefore, ascites tumors resemble their solid tumor counterparts.However, in solid tumors, plasma fibrinogen extravasates and isrendered insoluble by clotting to form a cross-linked fibrin gel; inturn, this gel induces angiogenesis and provokes the generation of

connective tissue stroma. The possibility that similar events mightalso occur in ascites tumors is investigated in the accompanyingpapers (41, 50).

REFERENCES

1. Sato, T., Takusagawa, IC, Asoo, N., Kumano, N., Hasuike, M., and Konno,K. Conversion of Lewis lung carcinoma into ascitic form. Oncology, 46: 188—192,1989.

2. Hagmar, B. MCG1O1-AA, a new ascites tumor in C57 mice: induction proceduresand some cytological and physiochemical characteristics. Acts PathoL Microbiol.Scand. Sect. A Pathol., 82: 358—368, 1974.

3. Klein,0., andKlein,E.Conversionof solidneoplasmsintoascitestumors.Ann.NYAcad. Sci., 63: 640—661,1956.

4. Klein, 0. Conversion of solid into ascites tumours. Nature (Land.), 171: 398—399,1953.

5. Patt,H.Growthcharacteristicsofthe Krebsascitestumor.Proc.Soc.Exp.Biol.Med.,83: 520—524,1953.

6. Klein, G., and Klein, E. The transformation of a solid transplantable mouse carcinomainto an “ascitestumor.―Cancer Res., 11: 466-469, 1951.

7. Goldie, H., and Felix, M. D. Growth tharacteristics of free tumor cells transferredserially in the peritoneal fluid of the mouse. Cancer Res., 11: 73—81,1951.

8. Yoshida, T. The Yoshida sarcoma, an ascites tumor. Gann, 40: 1-21, 1949.9. Nagy, J. A., Herzberg, K. T., Dvorak, J. M., and Dvorak, H. F. Pathogenesis of

malignant ascites formation: initiating events that lead to fluid accumulation. CancerRes., 53: 2631—2643,1993.

10. Nagy,J. A., Herzberg,K. T., Masse,E. M., Zientara,0. P., and Dvorak,H. F.Exchange of macromolecules between plasma and peritoneal cavity in ascites tumorbearing, normal, and serotonin-injected mice. Cancer Res, 49: 5448-5458, 1989.

11. Garrison, R. N., Galloway, R. H., and Heuser, L S. Mechanisms of malignant ascitesproduction. J. Surg. Res., 42: 126—132,1987.

12. Senger, D. R., Galli, S. J., Dvorak, A. M., Peruzzi, 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.

13. Markowitz, H. Proteins in ascites and ascites fluid. in: Proteins in Body Fluids,Amino ACIdS,and Tumor Markers: Diagnostic and Qinwal Aspects, pp. 221-226.New York: Alan R. Lisa, 1983.

14. Lifshitz, S., Clarke, W., and Buchsbaum, H. J. Serum and ascitic fluid proteinelectrophoresis and isnmunoglobulin levels in gynecologic cancer. Gynecol. Oncol.,8: 31—36,1980.

15. Fastaia, J., and Dumont, A. E. Pathogenesis of ascites in mice with peritonealcarcinomatosis. J. Nail. Cancer Inst., 56: 547—549,1976.

16. McGowan, L, Davis, R. H., and Bunnag, B. The biochemical diagnosis of ovariancancer. Am. J. Obstet. Gynecol., 116: 760—768,1973,

17. Hirabayashi, K., and Graham, J. Genesis of ascites in ovarian cancer. Am. J. Obstet.Gynecol., 106: 492—497,1970.

18. Straube, R. L Fluid accumulation during initial stages of ascites tumor growth.Cancer Res., 18: 57—65,1958.

19. Senger, D. R., Van De Water, L, Brown, L F., Nagy, J. A., Yeo, K-T., Yeo, T-K.,Berse, B., Jackman, R. W., Dvorak, A. M., and Dvorak, H. F. Vascular permeabilityfactor (VPF, VEGF) in tumor biology. Cancer Metastasis Rev., 12: 303—324,1993.

20. Berse, B., Brown, L F., Van De Water, L, Dvorak, H. F., and Senger, D. R. Vascularpermeabilityfactor(vascularendothelialgrowthfactor)geneis expresseddifferentially in normal tissues, macrophages, and tumors. Mol Biol. Cell, 3: 211—220,1992.

21. Senger, D. R., Connolly, D. T., Van De Water, L, Feder, J., and Dvorak, H. F.Purificationand NH2-terminalamino acid sequence of guinea pig tumor-secretedvascular permeability factor. Cancer Res., 50: 1774—1778, 1990.

22. Senger, D. R., Peruzzi, C. A., Feder, J., and Dvorak, H. F. A highly conservedvascular permeability factor secreted by a variety of human and rodent tumor celllines. Cancer Res., 46: 5629—5632, 1986.

23. Yea, K-T., Wang, H. H., Nagy, J. A., Sioussat, T. M., Ledbefler, S. R., Hoogewerf,A. J., Thou, Y., Masse, E. M., Senger, D. R., Dvorak, H. F., and Yeo, T-K. Vascularpermeability factor (vascular endothelial growth factor) in guinea pig and humantumor and inflammatory effusions. Cancer Res., 53: 2912—2918, 1993.

24. Sanford, B. H., Codington, J. F., Jeanloz, P.. W., and Palmer, P. D. Transplantabilityand antigenicityof two sublinesof the TA3tumor.J. Immunol.,110: 1233—1237,1973.

25. Friber& S. J Comparison of an immunoresistant and an immunosusceptible ascitessubline from murinetumorTA3. 1. Transplantability,morphologyand some physicochemical characteristics. J. Nati. Cancer. Inst., 48: 1463—1476,1972.

26. Levan,A.Thesignificanceof polyploidyfortheevolutionof mousetumorstrainsofthe TA3 mammary adenocarcinomawith different ploidy. Exp. Cell Res., 11:613—629,1956.

27. Hauschka, T. S. Cell population studies on mouse ascites tumors. Trans. NY Acad.Sci., 16: 64—73,1953.

28. Klein, U. Comparativestudies of mouse tumorswith respect to their capacity forgrowth as “ascitestumors―and their average nucleic acid content per cell. Exp. CellRes., 2: 518—573,1951.

29. Fekete, E., and Farringno, M. A. Studies on a transplantable teratoma of the mouse.Cancer Res., 12: 438—440,1952.

30. Ozols, R. F., Grotzinger, K. R., Fisher, R. I., Myers, C. E., and Young, R. C. Kineticcharacterization and response to chemotherapy in a transplantable murine ovariancancer. Cancer Res,, 39: 3202-3208, 1979.

31. Order, S. E., Donahue, V., and Knapp, R. Immunotherapy of ovarian carcinoma: anexperimentalmodel. Cancer(Phila.), 32: 573—579,1973.

32. Tripathi, R. C., and Tripathi, B. J. A new method for light and electron microscopiclocalization offluorescein-labelled dextran in ocular tissue using epoxy-resin embedding. Exp. Eye Res., 25: 259—264,1977.

33. Dvorak,H. F., Nagy,J. A., Dvorak,J. T., and Dvorak,A. M. Identificationandcharacterization of the blood vessels of solid tumors that are leaky to circulatingmacromolecules. Am. J. Pathol., 133: 95—109,1988.

34. Joris, I., DeGirolami, U., Wortham, K, and Majno, 0. Vascular labelling withmonastral blue B. Stain Technol., 57: 177—183,1982.

35. Dvorak,H. F., Dickersin,0. R., Dvorak,A. M., Manseau,E. J., and Pyne, K.Humanbreastcarcinoma:fibrin deposits and desmoplasia, inflammatorycell typeand distribution, microvasculature and infarction. J. NatI. Cancer Inst., 67:335—345,1981.

36. Conn, 0., Bayne, M. L, Soderman, D. D., Kwok, P. W., Sullivan, K. A., Palisi, T. M.,Hope, D. A., and Thomas, K. A. Amino acid and cDNA sequences of a vascularendotheial cell mitogen that is homologous to platelet-derivedgrowthfactor.Proc.Nail. Acad. Sci. USA, 87: 2628—2632,1990.

37. Claffey, K. P., Wilkison, W. 0., and Spiegelman, B. M. Vascular endothelial growthfactor:regulationby cell differentiationandactivatedsecond messengerpathways.J.Biol. Chem., 267: 16317—16322,1992.

38. Dvorak,H. F., Sioussat,T. M.,Brown,L F., Nagy,J. A., Sotrel,A., Manseau,E.,Van De Water,L, and Senger,D. R. Distributionof vascularpermeabilityfactor

367

Research. on August 24, 2020. © 1995 American Association for Cancercancerres.aacrjournals.org Downloaded from

VPF AND ASCITE5 TUMOR GROWTH

(vascular endothelial growth factor) in tumors: concentration in tumor blood vessels.J. Exp. Med., 174: 1275—1278,1991.

39. Yeo, T. K., Sioussat, T. M., Faix, J. D., Senger, D. R., and Yeo, T. K. Developmentof time-resolved immunofluorometric assay of vascular permeability factor. Gin.Chem.,38: 71—75,1992.

40. Zar, J. H. Biostatistical analysis. in: W. D. McElroy and C. P. Swanson (eds.),Prentice-Hall Biological Sciences Series, p. 620. Englewood Cliffs, NJ: Prentice-Hall,Inc., 1974.

41. Nagy, J., Morgan, E., Manseau, E., Dvorak, A., and Dvorak, H. Pathogenesis ofascites tumor growth: angiogenesis, vascular remodeling and stroma formation intissues of the peritoneal lining. Cancer Res., 55: 376—385,1995.

42. Granger, D. N., and Barrowman, J. A. Microcirculation of the alimentary tract. I.Physiology of transcapillary fluid and solute exchange. Gastroenterology, 84:846—868,1983.

43. Simonescu, N., Simonescu, M., and Palade, 0. Permeability of intestinal capillanes. Pathway followed by dextrans and glycogens. J. Cell Biol., 53: 365—392,1972.

44. Ferrara, N., Leung, D. W., Cachianes, 0., Winer, J., and Henzel, W. J. Purification and cloning of vascular endothelial growth factor secreted by pituitary

folliculostellate cells. Methods Enzymol., 198: 391—405,1991.45. Coon, 0., Soderman, D., Schaeffer, M. T., Wile, M., Hatcher, V. B., and Thomas,

K. A. Purification of a glycoprotein vascular endothelial cell mitogen from a ratglioma-derived cell line. Proc. NatI. Acad. Sci. USA, 87: 1323—1327,1990.

46. Miles, A. A., and Miles, E. M., Vascular reaction to histamine, histamine-liberator,and leukotaxinein the skin of guineapigs. J. Physiol. (Lond)., 118: 228—257,1952.

47. Ferrara, N., Houck, K. A., Jakeman, L B., Winer, J., and Leung, D. W. The vascularendotheial growth factor family of polypeptides.I. Cell Biochem., 47: 211—218,1991.

48. Connolly, D. T., Heuvelman, D. M., Nelson, R., Olander, J. V., Eppley, B. L,Delfino, J. J., Siegel, N. R., Leimgruber,R. M., and Feder, J. Tumor vascularpermeability factor stimulates endotheial cell growth and angiogenesis. J. CbsInvest., 84: 1470—1478,1989.

49. Brown,L F.,Asch,B.,Harvey,V. S.,Buchinski,B.,andDvorak,H. F. Fibrinogeninflux and accumulation of crosslinked fibrin in mouse carcinomas. Cancer Res., 48:1920—1925,1988.

50. Nagy,J. A., Meyers,M. S., Masse,E. M., Herzberg,K. T., and Dvorak,H. F.Pathogenesisofascites tumor growth: fibrinogen influx and fibrin accumulation in theperitoneallining of ascites tumor-bearingmice. CancerRes., 55: 369—375,1995.

368

Research. on August 24, 2020. © 1995 American Association for Cancercancerres.aacrjournals.org Downloaded from

1995;55:360-368. Cancer Res   Janice A. Nagy, Elizabeth M. Masse, Kemp T. Herzberg, et al.   AccumulationFactor, Vascular Hyperpermeability, and Ascites Fluid Pathogenesis of Ascites Tumor Growth: Vascular Permeability

  Updated version

  http://cancerres.aacrjournals.org/content/55/2/360

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/55/2/360To request permission to re-use all or part of this article, use this link

Research. on August 24, 2020. © 1995 American Association for Cancercancerres.aacrjournals.org Downloaded from