Relationships between Morphology, Dissemination, Migration ... · supernatant or the plasma prior to analysis by RIA since, at the concen trations used, there was no nonspecific interference

[CANCER RESEARCH 45, 3918-3923, August 1985)

Relationships between Morphology, Dissemination, Migration, and

Prostaglandin E2 Secretion by Cloned Variants of Lewis LungCarcinoma1

M. Rita Young,2 Marilyn Newby, and JoAnn Meunier

Center for Medical Education, Ball State University, Muncie, Indiana 47306

ABSTRACT

Lewis lung carcinoma cells (LLC) were isolated and clonedfrom metastatic lung nodules of C57BL/6 mice or from a culturedparental LLC line. Their dissemination abilities were defined invivo by their capacities to disseminate to the lungs following s.c.or i.v. injection into mice and in an in vitro model for tumordissemination by their ability to migrate out of glass capillarytubes. Cloned LLC cells with an enhanced dissemination capacityexhibited a rounded morphology, were nonadherent during invitro culture, and readily migrated out of capillary tubes. Incontrast, clones not capable of dissemination were adherent andspread during in vitro culture and did not migrate out of capillarytubes. Production of prostaglandin E2 by these clones wasmeasured by a radioimmunoassay. An inverse relationship wasobserved between the extent of migration and dissemination ofclones and their ability to secrete prostaglandin E2 In vitro and invivo. However, the prostaglandin E2 did not regulate the migration-dissemination capacities of LLC clones as inhibiting prosta

glandin synthesis did not alter their capacity to migrate in vitroor to lodge in the lungs following i.v. inoculation into mice.

INTRODUCTION

Tumors differ from normal cells and from each other withregard to adherence, migration, and dissemination capabilities(1,11,15,18,19). For example, tumors isolated from a primary3-methylcholanthrene-induced fibrosarcoma migrated in vitro out

of agarose to a lesser extent than cells isolated from a lungmetastasis (25). Similarly, metastatic MAT 13762 rat mammaryadenocarcinoma cells migrated out of glass capillary tubes to agreater degree than a nonmetastatic variant (1). Altered adhesiveness of tumor cells may contribute to the migration anddissemination capabilities of tumor cells, and, indeed, tumor cellshave been shown to be less adherent to artificial surfaces thannormal cells (16, 17, 22). In addition, B16-F1 cells grown as

spheroid cells were more metastatic than when grown in a flatconfiguration (20). However, more metastatic B16 melanomavariants were more adherent to other cells than less metastaticvariants (7, 14). Metastatic tumor cells were also shown tospecifically bind in vitro to the organs they selectively colonize invivo (13). Varani et al. (26) showed a direct relationship betweencell adhesiveness and tumorigenicity of a mouse fibrosarcoma.

Among the factors secreted by tumor cells in elevated amountsare prostaglandins, particularly of the E series. Rolland ef al. (21)demonstrated elevated prostaglandin levels in malignant breasttissue, primarily during the early stages of tumor development.Favalli ef a/. (6) reported that B16 melanoma tissue synthesized

1This work was supported in part by the DelawareCounty CancerSociety, Inc.,Muncie, IN, and by the American Lung Association.

2To whom requests for reprints should be addressed, at Center for MedicalEducation, Maria Bingham Hall, Ball State University, Muncie, IN 47306.

Received10/18/84; revised 4/3/85; accepted 5/8/85.

more PGE3 than analogous normal tissue. Elevated concentra

tions of PGE are present in culture supernatants of LLC cells(29) and in the circulation of tumor bearers (12, 29). The metastatic mammary tumor lines 4526 and 410.4 were shown toproduce more PGE than the nonmetastatic tumor line 410 (9).Although the role of PGE2 in subversion of immune competencehas been well studied (23, 27, 28), the direct effects of prostaglandins on tumor growth and dissemination are less certain.

The current studies were conducted to determine the relationships between in vitro morphologies of LLC variants, their capacities to disseminate in vivo, migrate in vitro, and to secretePGE2. Also the effect of indomethacin, a prostaglandin synthesisinhibitor, on the migration and dissemination abilities of thesecells was examined.

MATERIALS AND METHODS

Mice. All of the studies utilized 6- to 8-week-old C57BL/6 mice which

were obtained from our own colony.Tumor Cells. LLC, a spontaneous lung carcinoma derived from mouse

strain C57BL/10, was kindly donated by Or. G. Sundharadas, Universityof Wisconsin. The cells were maintained by in vitro culture at 37°Cin a

humidified atmosphere containing 5% C02. Culture medium for LLC cellsconsisted of RPMI 1640 supplemented with penicillin (100 units/ml),streptomycin (100 MQ/ml), and 10% FBS (KC Biological, Inc., Lenexa,KS). In some experiments, cells were also cultured for 24 h in mediumcontaining 10"6 M indomethacin (Sigma Chemical Co., St. Louis, MO).

Variant subpopulations of LLC were derived from either metastaticlung nodules or from parental LLC maintained by in vitro culture. C57BL/6 mice received dorsal s.c. implants of 106 parental LLC. After 4 weeks,

mice were sacrified, and the lungs were aseptically removed. Metastaticlung nodules were dissected from each lung, individually meshed througha screen, and cultured in 25-sq cm tissue culture flasks containing 10 ml

of RPM11640 plus 10% FBS. These cells were then cloned by limitingdilution and screened for their ability to disseminate to the lungs of micereceiving s.c. or i.v. implants of tumor cells. Several clones derived fromthe metastatic lung nodules were selected and termed C3, C4, and E3.

In addition, subpopulations of cells were derived from parental LLCwhich had been maintained by in vitro culture for greater than 100passages. Cultured LLC were cloned by limiting dilution, and the cloneswere screened for their minimal ability to disseminate to the lungs ofmice receiving s.c. or i.v. implants of tumor cells. Two clones derivedfrom the cultured parental LLC were selected and termed C8 and E5.

Assay for Quantitating PGE2 Secreted by LLC Variants into CultureSupernatants. Into each of 4 round-bottomed polypropylene tubes wereseeded 10e cloned LLC variants in 2 ml of RPMI 1640 plus 10% FBS.

After 4 h, cells were pelleted by centrifugaron, and the supernatantswere removed and frozen. PGE2 concentrations in the supernatants weredetermined by a RIA.

Assay for Quantitating PGE: Levels in Plasma of Tumor-bearing

Mice. To determine the amounts of PGE2 in plasma of tumor bearers,

3The abbreviations used are: PGE2, prostaglandin E2 (other prostaglandins

defined similarly);LLC, Lewis lung carcinoma; RIA, radioimmunoassay;FBS, fetalbovine serum.

CANCER RESEARCH VOL. 45 AUGUST 1985

3918

on May 21, 2020. © 1985 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

HETEROGENEITY OF CLONED LLC

mice were exsanguinated 4 weeks after s.c. implantation of 106 LLC

variants. Each ml of blood was mixed with 0.7 mg of EDTA andindomethacin to yield 1(T6 M to prevent prostaglandin synthesis in vitro.Blood was then centrifugad at 4"C. Plasma PGE2 levels were determined

by a RIA.RIA for PGE2. Analysis of PGE2 levels in the supematants of in vitro-

cultured LLC variants and plasma of tumor-bearing mice was performed

with a commercially available RIA kit (New England Nuclear, Boston,MA). To each RIA tube were added 0.1 ml of anti-PGE2, 0.1 ml of 125I-

PGEa, and 0.1 ml of either known amounts of PGE2, undiluted cell culturesupernatant, or plasma diluted 1:20. After 24 h of incubation at 4°C,

immune complexes were precipitated with 1 ml of a 16% polyethyleneglycol solution (M, 6000), and the radioactivity in the precipitate wasdetermined. Using this procedure, the extent of cross-reactivity between

PGE2 and PGE, was less than 1%, and the sensitivity of the assay was0.25 pg of PGE2. It was not necessary to extract PGE2 from either thesupernatant or the plasma prior to analysis by RIA since, at the concentrations used, there was no nonspecific interference in the assay bycomponents of these samples. In prior control studies which we conducted, known amounts of PGE2 were added to either medium or to a5% solution of plasma. Using the described RIA, an additive amount ofPGE2 was quantitated (PGE2 in the sample plus PGE2 added).

Dissemination of Tumor Cells to the Lungs of Mice. Four weeksafter s.c. implantation of 10s LLC variants, lungs of tumor-bearing mice

were removed and washed in phosphate-buffered saline. Lungs were

disrupted into fragments (less than 1 mm in size) by mincing lungs ofeach mouse with scissors and pressing against a wire mesh into 10 mlof medium. Lung fragments from each mouse were then cultured in a25-sq cm flask for 1 week, during which time growth of tumor cells was

microscopically monitored. Lung cultures were determined to containtumor cells if growing cells which morphologically resembled LLC variantscould be propagated by serial culture. In additional studies, 2 x 105

cloned LLC variant cells were injected i.v. into mice. Four weeks later,the mice were sacrificed, and the lungs were exposed. The number ofmetastatic lung nodules was quantitated after filling the lungs with 15%India ink and then washing in buffer..

Injection of "Cr-labeled LLC Variants i.v. Single cell suspensionscontaining 107 of each of the LLC variants were exposed to 250 nC\ ofNa251CrO4 (ICN, Irvine, CA). After 1 h of incubation at 37°C in an

atmosphere of 5% C02, the cells were washed 3 times with phosphate-buffered saline. Each of the 51Cr-labeled variants (106 cells) was injected

i.v. into the tail veins of C57BL/6 mice (4 mice/group/experiment). Thetotal amount of radioactivity in the injected cells was determined bycounting a sample of the cell suspension used for injection in a gammacounter. After 4 h, mice were killed, and their lungs were removed. Any

residual blood was cleaned off the surface of the lungs. The lungs ofeach mouse were then placed in a tube, and the radioactivity wascounted in a gamma counter.

Migration of LLC Variants in Wfro. The assay for measuring the invitro migration of LLC variants out of capillary tubes was carried out aspreviously described (4,12) with modifications. Briefly, cells were drawninto sterile 1.5- x 100-mm glass capillary tubes and sealed at one end

with a 60:40 mixture of sterile paraffin and vaseline. The capillary tubeswere centrifugea (200 x g) for 5 min and cut at the cell-fluid interface.The cell-containing stubs were placed into wells of a migration plate

(York Scientific, Inc., Ogdensburg, NY) and secured with sterile vaseline.

The wells were filled with RPMI 1640 plus 10% FBS. In some experiments, 10"6 M indomethacin was added to the migration wells. After

incubation at 37°C, the migration areas were projected onto paper,

traced, and measured by weight. Results are expressed as arbitraryunits of migration area.

Analysis of Data. Data are expressed as the mean ±SD or SE. TheStudent f test was used to analyze significance of differences of meanvalues.

RESULTS

Morphological Characteristics of LLC Variants. The morphologies of LLC variants which had been derived from metastatic lung nodules of mice or from in wfro-cultured parental LLC

were compared. Both the primary cultures and clones (C3, C4,and E3) of tumor cells derived from lung nodules exhibited arounded morphology when grown in RPM11640 plus 10% FBS.Furthermore, these cells were not dependent on adherence to atissue culture surface for growth and remained nonadherentthroughout culture. In contrast, the parental LLC and clones (C8and E5) derived from LLC which had been maintained by in vitroculture were adherent and spread on the tissue culture surfacethroughout in vitro culture. Shown in Fig. 1 is the typical roundedmorphology of a clone (C3) which was derived from a metastaticlung nodule and the typical spread morphology of a clone (C8)which was derived from in w'fro-cultured parental LLC. These

characteristics have remained stable for over 100 passages invitro.

Secretion of PGE, by LLC Variants in Vitro and in Vivo.Since we (12, 29) and others (6, 21) have previously shown thattumor cells secrete greater amounts of prostaglandins, especiallyof the E series, than normal cells, the amount of PGE2 secretedby LLC variants was measured. The results of Table 1 showthat PGE2 production in vitro did not correspond with a metastatic origin of the cells. Both the primary cultures and clones oftumor cells derived from lung nodules secreted less PGE2 in vitrothan the parental LLC and clones derived from LLC which hadbeen maintained by in vitro culture (P < 0.02). Using trypan blueexclusion, the viability of all LLC variants was determined to begreater than 95%. Therefore, the PGE2 in the culture superna-

tants was attributed to the secretion of PGE2 by the cells ratherthan cell lysis. Furthermore, the PGE2-secreting cells were con

sidered to be LLC variants and not contaminating cell types,such as macrophages, since the LLC cells had either beenmaintained by in vitro culture for over 100 passages (parentalLLC and thus clones C8 and E5) or were nonadherent (lungnodule cultures and clones C3, C4, and E3).

Since the above study showed that cells derived from in w'fro-

cultured LLC secreted more PGE2 during culture than LLCvariants derived from metastatic lung nodules, it seemed likely

Table 1Secretion of PGE¡by LLC variantsduring 4 h of culture

Experiment123LLC variantC3C4C8Parental

LLCC3C4Lung

nodulecultureParentalLLCC4E3C8E5Parental

LLCPGE2

in supernatant(pg/ml)29.4

±2.9"11

9.7±3.980.7±5.6190.1±2.9187.5

±0.727.9

±0.7160.0±1.1107.7±13.499.2

±0.3224.3±5.424.3

±0.31?3.8±5.4131.4

+10.5221.1±3.2182.9±8.1209.0

±5.3* Mean ±SD of quadruplicates.

CANCER RESEARCH VOL. 45 AUGUST 1985

3919

on May 21, 2020. © 1985 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

HETEROGENEITY OF CLONED LLC

that a similar relationship might exist in vivo. Therefore, C57BL/6 mice were implanted s.c. with LLC variants, and the plasmaPGE2 concentrations were quantitated after 4 weeks of tumorgrowth. As shown in Table 2, the levels of PGE2 in the plasmaof mice implanted with cloned C8 or E5 cells (derived from inw'fro-cultured parental LLC) were elevated (P < 0.02 and P <

0.01, respectively). In contrast, plasma PGE2 concentrations ofmice bearing cloned C3, C4, or E3 tumor cells (derived frommetastatic lung nodules) were similar to those of normal mice.

Dissemination of LLC Variants to the Lungs 4 Weeks afterImplantation. The capacity of LLC variants to disseminate tothe lungs was measured by implanting the cells s.c. or i.v. intomice and 4 weeks later determining the frequency of mice whoselungs contained LLC cells or the number of metastatic noduleson the lungs. As shown in Table 3, mice implanted s.c. or i.v.with LLC variants derived from metastatic lung nodules (C3, C4,or E3) more frequently had tumor cells in their lungs (P < 0.01 )and had a greater quantity of metastatic lung nodules (P < 0.05)than did mice implanted with cells derived from LLC which hadbeen maintained by in vitro culture (E5 and C8 clones).

Dissemination of LLC Variants to the Lungs 4 h after i.v.Implantation. This study examined if the increased disseminationability of cells derived from metastatic lung nodules might be dueto an increased ability of these cells to lodge in the lungs, thepreferred organ for growth. Table 4 shows that i.v. injection of51Cr-labeled LLC clones yielded results after 4 h which were

similar to those in Table 3 in which the dissemination abilities ofLLC clones in mice 4 weeks after implantation with tumor weredescribed. Lungs of mice implanted with labeled cells derivedfrom metastatic lung nodules (C3, C4, and E3) contained agreater amount of radioactivity after 4 h than lungs of miceimplanted with labeled cells derived from in w'fro-cultured parental

Table 2Plasma PGEt levels in mice bearing LLC variants

Each experiment consisted of groups of 5 mice, each of which received s.c.implants of 10" cloned variant LLC cells. After 4 weeks, primary tumor weights andplasma PGE2levels of each mouse were measured.

LLCvariantC3

C4E3C8E5Primary

tumorwt,(g)2.6

±0.63.4 ±0.92.2 ±0.72.3 ±0.80.4 ±1.1PGE2

in plasma(pg/ml)469

±31"

424 ±46429 ±26481 ±36

1078 ±148997 ±83P

value3NSC

NSNS

<0.02<0.01

aThe Student t test was used to analyze significance of difference of PGE2levels in plasma from tumor-bearing and control mice.

" Mean ±SE of 4 experiments.c NS, not significant.

Table 3Disseminationof LLC variants to the lungs 4 weeks after i.v. implantation

In each experiment, groups of 5 mice each received dorsal s.c. implants of 10°variant LLC or i.v. implants of 2 x 10sLLC. The frequency of mice with tumor cellsin their lungs 4 weeks following s.c. LLC implantation or the number of metastaticlung nodules 4 weeks following i.v. LLC implantation was determined.

LLC variant

% of mice with LLC inlungs following s.c. im

plantation

No. of metastatic lung nodules/mouse following i.v. im

plantation

C3C4E3C8E51001001000037 ±10a40

±1636±32±69±7

Table 4Disseminationof LLC variants to the lungs 4 h after i.v. implantation

In each experiment, s'Cr-labeled LLC variants were injected i.v. into groups of

4 mice each. After 4 h, the radioactivity in the lungs was counted.

LLCvariantC3

C4E3C8C8E5E5Cells

precultured in10"* M ¡ndomethacin+

+%

of injected radioactivity inlungs24.2

±1.6a

19.6 ±0.617.0 ±2.17.1 ±1.69.3 ±1.48.6 ±1.48.3 ±0.9

" Mean ±SE of 3 experiments.

Table 5In vitro migration of LLC variants

Eachexperiment consisted of quadruplicate determinations.

LLCvariantC3C3C3C3C404E3E3C8C8C8C8ESE5Parental

LLCCells

preculturedin ¡ndo

methacin-+———+—+—+-——+-MigrationmediumMediumIndomethacinC3

supernatantC8supernatantMediumIndomethacinMediumIndomethacinMediumIndomethacinC8

supernatantC3supernatantMediumIndomethacinMediumMigration

area4.07±0.28a4.19

±0.273.52±0.393.46±0.434.10

±0.364.22±0.364.1

3±0.523.80±0.410.88

±0.110.77±0.191.30

±0.351.14±0.241.40

±0.351.19±0.170.65

±0.03

°Mean ±SE of 3 experiments.

8 Mean ±SE of 3 experiments.

LLC (C8 and E5)(P< 0.01).The above studies showed that LLC variants with a low

capacity to disseminate secrete greater amounts of PGE2 thanLLC variants derived from metastatic lung nodules. Consequently, a study was designed to determine if tumor-derivedPGE2 regulates the tumor cells' ability to disseminate to thelungs. The capacity of C8 or E5 cells cultured in 10~6 M indo-methacin and labeled with 51Crto lodge in the lungs 4 h after i.v.

injection was measured. Inhibiting prostaglandin synthesis of C8or E5 cells did not influence their capacity to disseminate andremain in the lungs (Table 4).

in Vitro Migration of LLC Variants. Migration of LLC variantsout of glass capillary tubes was used as an in vitro model fortumor dissemination. As shown in Table 5, the in vitro migrationof cells derived from metastatic lung nodules (C3, C4, and E3)was enhanced compared to migration by parental LLC or theclones derived from in w'fro-cultured LLC (P < 0.05). The en

hanced migration by the C3, C4, and E3 cells was detectablewithin 4 h (data not shown). Parental LLC and cloned C8 and E5cells derived from in w'fro-cultured LLC migrated out of the

capillary tubes only to a minimal degree even after 24 h ofincubation.

The effect of inhibiting prostaglandin synthesis of the LLCvariants on their in vitro migration was also measured. To accomplish this, LLC variants were cultured from 24 h in the presenceor absence of 10~6 M ¡ndomethacin, a prostaglandin synthesis

inhibitor. We have previously (29) shown that indomethacin isnot toxic to LLC, although it completely inhibits prostaglandin

CANCER RESEARCH VOL. 45 AUGUST 1985

3920

on May 21, 2020. © 1985 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

HETEROGENEITY OF CLONED LUC

synthesis. Incubation with 10r6 M ¡ndomethacinwas sufficient to

inhibit prostaglandin synthesis, since subsequent culture of thesecells (4 h) in medium did not result in any addition of PGE2 to themedium. The medium- or indomethacin-precultured cells were

then allowed to migrate out of glass capillary tubes into mediumalone or medium containing 10~6 M indomethacin. As shown in

Table 5, inhibition of prostaglandin synthesis by LLC variants didnot alter their in vitro migration.

To determine if any other soluble factors secreted by theclones might be responsible for their enhanced or minimal migration capacities, the C3 and C8 cells were allowed to migrate into48-h culture supernatants of the LLC clones. As shown in Table

5, the enhanced migration of C3 cells was not reduced by C8culture supernatants. Likewise, the C8 migration was not enhanced by C3 culture supernatants. This suggests that themigration capabilities of these cells are inherent characteristicsof the cells and not mediated by secreted soluble factors.

DISCUSSION

Multiple properties have been attributed to tumor cells withvarying capacities to disseminate. Some of the properties whichdiffer between variants of even the same tumor line includemetastatic capacity (7, 24), growth rate (9), enzyme secretion(25), and adherence (26). In this study, we have cloned LLCvariants and have compared their morphologies and their capacities to disseminate in vivo and to secrete PGE2. In addition,migration of the LLC variants out of glass capillary tubes wasused as an in vitro model for tumor dissemination in a host.

Variants of LLC derived from metastatic lung nodules wererounded and nonadherent during in vitro culture, migrated to agreater extent out of glass capillary tubes in vitro, and disseminated to the lungs of implanted mice more readily. In contrast,cells derived from in w'fro-maintained parental LLC were adher

ent, spread during culture, and did not readily migrate out ofglass capillary tubes or disseminate in vivo. We do not fullyunderstand the relationship between these characteristics, butthe lack of adherence of tumor cells may contribute to anincreased dissemination capacity. Our results are consistent withstudies or other investigators which describe a good correlationbetween spontaneous dissemination of cells to the lungs from aprimary tumor and artificial dissemination to the lungs followingi.v. injection of cells (10, 24). Since the rounded cells derivedfrom metastatic lung nodules readily disseminated to the lungsregardless of whether they were implanted s.c. or i.v., it is unlikelythat their enhanced dissemination was simply due to an increased ability to detach from the primary tumor mass. It ispossible that the rounded cells have a greater capacity to invadethrough the vascular endothelium. The spread cells derived fromin w'fro-cultured parental LLC were less capable of dissemination.

This is also not simply due to a reduced capacity to detach fromthe primary tumor, since these cells were less capable of lodgingin the lungs following i.v. implantation.

In this study, in vitro migration of LLC variants out of glasscapillary tubes corresponded with their ability to spontaneouslyand artificially disseminate to the lungs in vivo. Investigations byothers have also shown a direct relationship between a metastatic origin of tumor cells and their capacity to migrate in vitroout of either agarose (25) or capillary tubes (1). Since in vitromigration of tumor cells out of capillary tubes is a rapid method

for studying variables which might influence tumor behavior invivo, we used this model system to ascertain if there was acorrelation between tumor migration and PGE2 secretion andwhat the effect of inhibiting prostaglandin synthesis might be.Our study showed an inverse relationship between the amountof PGE2 secreted in vitro by cloned LLC variants and theircapacity to migrate. Similarly, plasma PGE2 concentrations ofmice bearing variants derived from metastatic lung nodules werelower than of mice bearing LLC variants with minimal dissemination capacity. Although the amount of PGE2 produced by LLCvariants correlated inversely with their migration-metastatic capacity, the tumor-derived PGE2 did not directly regulate these

characteristics, as inhibiting prostaglandin synthesis did not alterthe tumor cells' capacity to migrate in vitro or to lodge in the

lungs following i.v. injection into mice. In fact, it did not appearthat any soluble factor secreted by variants with an enhanced orminimal migration capacity was responsible for these characteristics, since culture supernatants of high- or low-migrating LLC

variants did not stimulate or inhibit cell migration.A variety of tumors have previously been shown to secrete

elevated amounts of PGE, although the amounts reported in thisstudy and in previous studies by us and others to be present intumor bearers have varied (6, 9, 12, 29). Much of the variationmay be attributed to the methods used to collect the blood andthe methods used to quantitate PGE2 levels. For example, prostaglandin concentrations are substantially greater in serum thanin plasma (5). In this study, we added EDTA and indomethacin(to reduce ex vivo prostaglandin synthesis) to the collected blood.Thus, the concentrations of PGE2 which we presently show tobe present in the plasma of mice bearing any of the LLC variantclones are less than the PGE levels which we previously reportedto be in the sera of mice bearing the parental LLC (29). Also, inthis study, we specifically quantitated PGE2 in the plasma oftumor bearers by using an antibody in the RIA which does notcross-react with other prostaglandins.

Our data are consistent with the study of Fitzpatrick andStringfellow (8), who demonstrated that arachidonic acid metabolism to PGD2 is less in the metastatic B16F10 than in themoderately metastatic B16F1. In our study, we showed aninverse relationship between the metastatic capacity of LLC cellsand PGE2 production. In several respects, our results are alsoconsistent with those of Rolland ef al. (21), who evaluated therelationship between prostaglandin production and metastaticcapacity of breast cancers. They reported a direct relationshipbetween cellular adhesiveness and prostaglandin production.Likewise, we have presently shown that the adherent nonmetas-

tatic LLC variants secrete more PGE2 than the nonadherentmetastatic LLC. In contrast, their study concluded a direct relationship between prostaglandin production and metastatic capability. This conclusion was based on results which were representative of the amount of prostaglandin produced by the cellswhich compose a tumor mass. Our conclusion of an inverserelationship between PGE2 production and metastatic LLC capacity was based on studies using cloned LLC variants withoutcontaminating mononuclear phagocytes, which are also knownto produce PGE2. Finally, the differences in the relationshipbetween prostaglandin production and tumor metastatic capability may be due to the differences in tumor types examinedsince, in other studies, a direct relationship has been observedbetween metastatic ability of mammary tumors and PGE pro-

CANCER RESEARCH VOL. 45 AUGUST 1985

3921

on May 21, 2020. © 1985 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

HETEROGENEITY OF CLONED LLC

duction (9). In reviewing the relationships between prostaglan-

dins and tumors, Bennett and coworkers (2, 3) have showndirect, inverse, or no relationship and, thus, point out manysources of error in comparative studies including the heterogeneity and tissue origin of tumors, the amount of cellular infiltrate,and necrosis.

This study did not exclude the possibility that PGE2 mightinfluence the ability of cells to detach from the primary tumor. Inaddition, our results did not preclude the significance of other invivo properties of PGE2 such as its role in inflammation, bonerésorption, cell growth, and immune modulation. In fact, ourprevious studies (12, 28, 29) have shown that reducing PGE2concentrations in tumor-implanted mice minimizes development

of the primary tumor and of métastases. We have attributedthese results to the prevention of PGE2-mediated immuno-

suppression which typically accompanies tumor growth. Thestudies described presently have evaluated the direct influenceof inhibiting prostaglandin synthesis on tumor migration-metas

tasis without interference of the immune modulating effects ofPGE2. However, we are currently expanding these studies tomeasure the immune antitumor competence of mice implantedwith the cloned LLC variants presently described.

ACKNOWLEDGMENTS

We are grateful to Glenna Hedge for her help in preparationof this manuscript.

REFERENCES

1. Badenoch-Jones,P., and Ramshaw, I. A. Spontaneouscapillarytube migrationof metastatic rat mammary adenocarcinomacells. Invasion & Metastasis, 4:98-110,1984.

2. Bennett, A. Prostaglandins and tumour metastasis. In: H. Thaler-Dao, ef a/,(eds), Icosanoids and Cancer, pp. 223-228. New York: Raven Press, 1984.

3. Bennett, A., DelTacca, M., Stamford, l. F., and Zebro, T. Prostaglandinsfromtumors of human large bowel. Br. J. Cancer, 35: 881-884, 1977.

4. Cheung, H. T., Cantarow, W. D., and Sundharadas, G. Characteristics of alow-molecular-weightfactor extracted from mouse tumors that affects in vitroproperties of macrophages. Int. J. Cancer, 23: 344-352, 1979.

5. Dray, F.,Charbonnel,B., and Maclouf,J. Radioimmunoassayof prostaglandinsE„,E,, and E2in human plasma. Eur. J. Clin. Invest., 5: 311-318,1975.

6. Favalli,C., Garaci,E., Etheredge,E., Santoro, M. G., andJaffe, B. M. Influenceof PGEon the immune response in melanoma-bearingmice. J. Immunol., 725:897-902, 1980.

7. Fidler, I. J., and Nicolson, G. L. Tumor cell and host properties affecting theimplantationand survivalof blood-bornemetastatic variants of B16 melanoma.Isr. J. Med. Sci., 14: 38-50,1978.

8. Fitzpatrick, F. A., and Stringfellow, D. A. Prostaglandin D2 formation bymalignantmelanomacellscorrelates inverselywith cellularmetastatic potential.Proc. Nati. Acad. Sci. USA 76: 1765-1769,1979.

9. Fulton, A. M. Effects of indomethacin on the growth of cultured mammarytumors. Int. J. Cancer,33: 375-379,1984.

10. Janik, P., Assaf, A., and Bertram, J. S. Inhibition of growth of primary andmetastatic Lewis lung carcinoma cells by the phosphodiesterase inhibitorisobutylmethylxanthene.Cancer Res., 40:1950-1954,1980.

11. Kripke, M. L., Gruyo, L., and Fidler, I. J. Metastatic heterogeneity of cells froman ultraviolet light-inducedmurine fibrosarcoma of recent origin. Cancer Res.,38: 2962-2967, 1978.

12. Mahan, M., Meunier, J., Newby, M., and Young, M. R. Prostaglandin E2production by EL4 leukemiacells: mechanismfor tumor dissemination.J. Nati.Cancer Inst., 74: 191-195,1985.

13. Netland, P. A., and Zetter, B. R. Organ-specific adhesion of metastatic tumorcells in vitro. Science(Wash. DC),224: 1113-1115,1984.

14. Nicolson,G. L. Cell surface and blood-borne tumor metastasis. In: S. B. Day,P. Stansly, W. P. Laird Myers, S. Garattini, and M. G. Lewis (eds.), CancerInvasion and Metastasis: Biologic Mechanisms and Therapy, Progress inCancer Researchand Therapy, Vol. 5, pp. 163-174. New York: RavenPress,1977.

15. Nicolson, G. L., and Custead, S. E. Tumor metastasis not due to adaptationof cells to a new organ environment. Science (Wash. DC), 2/5: 176-178,1982.

16. Pastan,L,andJohnson,G. S. CyclicAMP and the transformationof fibroblasts.Adv. Cancer Res., )9: 303-330,1974.

17. Pastan, I., and Willingham, M. Cellular transformation and the "morphologicphenotype" of transformed cells. Review article. Nature (Lond.), 274: 645-

650,1978.18. Poste, G., Tzeng, J., Doll, J., Greig, R., Rieman,D., and Zeidman, I. Evolution

of tumor cell heterogeneity during progressive growth of individual lung métastases. Proc. Nati. Acad. Sci. USA, 79: 6574-6578, 1982.

19. Poste, G., and Weiss, L. Some considerations on cell surface alteration inmalignancy.In: L. Weiss (ed.), FundamentalAspects of Metastasis, pp. 25-47. Amsterdam: North-HollandPublishingCo., 1976.

20. Raz, A., and Ben-Ze'ev, A. Modulation of the metastatic capability in B16melanomaby cell shape. Science(Wash. DC),22J: 1307-1310,1983.

21. Rolland, P. H., Martin, P. M., Jacquemier, J., Rolland, A. M., and Toga, M.Prostaglandinin human breast cancer. Evidence suggesting that an elevatedprostaglandinproduction is a marker of high metastatic potential for neoplastiacells. J. Nati. Cancer Inst., 64:1061-1070,1980.

22. Sanford, K., Barker, B. E., Woods, M. W., Parshad,R., and Law, L. W. Searchfor "indicators" of neoplasticconversionin vitro. J. Nati. Cancer Inst., 39: 705-

733,1967.23. Ting, C. C., and Hargrove, M. E. Activation of natural killer-derived cytotoxic

T lymphocytes. I. Regulationby macrophageand prostaglandins.J. Immunol.,73?: 1734-1741, 1983.

24. van Lamsweerde, A. L., Henry, N., and Vaes, G. Metastatic heterogeneity ofcells from Lewis lung carcinoma.Cancer Res., 43: 5314-5320, 1983.

25. Varani, J., Orr, W., and Ward, P. A. Hydrolytic enzyme activities, migratoryactivity, and in vivo growth and metastatic potential of recent tumor isolates.Cancer Res., 39: 2376-2380,1979.

26. Varani, J., Orr, W., and Ward, P. A. Adhesive characteristics of tumor cellvariants of high and low tumorigenic potential. J. Nati. Cancer Inst., 64:1173-1178,1980.

27. Walker, C., Kristensen, F., Bettens, F., and deWeck, A. L. Lymphokineregulationof activated (G1) lymphocytes. I. ProstaglandinE2induces inhibitionof interieukin2 production. J. Immunol., 730:1770-1773,1983.

28. Young, M. R., and Dizer, M. Enhancement of immune function and tumorgrowth inhibition by antibodies against prostaglandin E2.Immunol. Commun.,72:11-23,1983.

29. Young, M. R., and Knies, S. Prostaglandin E production by Lewis lungcarcinoma: mechanismfor tumor establishment in vivo. J. Nati. Cancer Inst.,72:919-922,1984.

CANCER RESEARCH VOL. 45 AUGUST 1985

3922

on May 21, 2020. © 1985 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

HETEROGENEITY OF CLONED LLC

Fig. 1. Morphology of cloned C3 (left) and C8 (right) variants during in vitro culture.

CANCER RESEARCH VOL. 45 AUGUST 1985

3923

on May 21, 2020. © 1985 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

1985;45:3918-3923. Cancer Res   M. Rita Young, Marilyn Newby and JoAnn Meunier  Lung Carcinoma

Secretion by Cloned Variants of Lewis2and Prostaglandin ERelationships between Morphology, Dissemination, Migration,

  Updated version

  http://cancerres.aacrjournals.org/content/45/8/3918

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

  [email protected] 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/45/8/3918To request permission to re-use all or part of this article, use this link

on May 21, 2020. © 1985 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from