in vitro selection of murine b16 melanoma variants with … · new assay methods have been devised...

[CANCER RESEARCH 40. 1636-1644, May 1980]

In Vitro Selection of Murine B16 Melanoma Variants with EnhancedTissue-invasive Properties

G. Poste,1 2 J. Doll,1 I. R. Hart,3 and I. J. Fidler3

Department of Experimental Pathology. Roswell Park Memorial Institute, Buffalo. New York 14263 [G. P., J. D.I, and Cancer Biology Program, National CancerInstitute. Frederick Cancer Research Center. Frederick, Maryland 21701 Â¡l.R. H., I. J. F.]

ABSTRACT

New assay methods have been devised to quantitate tumorcell invasion of tissues of differing histological complexity maintained as organ cultures in vitro (chorioallantoic membrane ofchicken, mouse urinary bladder, and canine blood vessel). Inaddition to quantitating tumor cell invasion, these methods alsoallow recovery of invasive cells for comparison with noninvasivecells. These methods have been used to select variant sublinesfrom murine B16-F1 and B16-F10 melanoma lines that displaysignificantly greater tissue-invasive abilities than the parentlines. B16 variant sublines selected in vitro for increased in-

vasiveness through the bladder wall or vein also show a significant increase in their ability to form spontaneous and experimental mÃ©tastases in vivo. In contrast, cells from the sameparent cell line selected for increased invasiveness through thechorioallantoic membrane do not show significant alterationsin metastatic behavior. We conclude that invasive variants canbe isolated from the parent B16 tumor by several in vitromethods and that the level of expression of the invasive phe-

notype in vivo may be determined by the severity of the selection procedure in vitro.

INTRODUCTION

Invasion of lymphatics and blood vessels by malignant cellsprovides a major pathway for the dissemination of neoplasticcells within the body (for reviews, see Refs. 7, 8, and 27). Inaddition, following their arrest in the capillary beds of differentorgans, circulating tumor cells invade the wall of these vesselsand escape into the extravascular tissue(s) where they establish mÃ©tastases. Despite its obvious importance in the metastatic process, relatively little is known about the mechanism(s)of tumor invasion. This deficiency reflects the formidable technical problems encountered in studying invasion in vivo andthe lack of quantitative methods for studying invasion in vitroand for recovering invasive cells once invasion has taken place.

A substantial body of evidence has been assembled in thelast few years which indicates that primary malignant neoplasms are not homogeneous entities of cells with uniformproperties but instead contain subpopulations of tumor cellswith widely differing metastatic abilities (for reviews, see Refs.7, 8, and 27). This phenotypic heterogeneity dictates thatexamination of heterogeneous unselected tumor cell populations may offer little insight into the cellular properties responsible for invasion and/or metastasis if only very few cells within

the population express these behavioral traits. A more productive approach to the experimental analysis of the cellular properties needed for successful invasion is to isolate invasivetumor cell subpopulations and compare them with poorly invasive or noninvasive tumor cells derived from the same parentcell population. In this report, we describe methods for the invitro isolation of a series of murine B16 melanoma cell variantswith enhanced invasive properties and an initial characterization of their behavior in vivo.

MATERIALS AND METHODS

Cells. The origin and properties of the B16-F1 (low potentialfor lung colonization) and B16-F10 (high potential for lungcolonization) sublines of the B16 melanoma have been described in detail previously (5, 12). Cultures of mouse embryofibroblasts were prepared by trypsinization of 14- to 16-day-old C57BL/6 mouse fetuses as described elsewhere (29).Homogeneous peritoneal macrophage cultures were obtainedfrom C57BL/6 mice inoculated i.p. with sodium thioglycollate(Baltimore Biological Laboratories, Cockeysville, Md.) as described previously (29). All cells were incubated in plasticflasks or Retri dishes in CMEM4 (12, 29). The components of

CMEM were obtained from Flow Laboratories, Inc., Rockville,Md., and the Grand Island Biological Co., Grand Island, N. Y.Cell cultures were incubated at 37Â°in a humidified atmosphere

containing 5% CO2 in air. All cultures were free of Mycoplasmespecies and pathogenic murine viruses (12), and they remainedso throughout the experiments. In certain experiments, cellcultures (except macrophages) were labeled with [125l]ldUrd as

detailed in Ref. 5. Briefly, actively growing nonconfluent mono-layer cell cultures were incubated for 24 hr in CMEM containing[125l]ldUrd (0.2 juCi/ml; specific activity, 200 /iCi/mmol; New

England Nuclear, Boston, Mass.). This method labels morethan 95% of the cells and does not alter their metastaticbehavior in vivo (5).

Assay of Tumor Cell Invasion in CAM and Selection ofInvasive Cell Variants. Newly fertilized eggs were incubatedat 38Â°in a humidified atmosphere (relative humidity, approxi

mately 75%) in an egg incubator (Favorite Incubator; LeahyMfg. Co., Higginsville, Mo.) until Day 8 after fertilization, whenthe CAM was dropped using the false air sac method asdescribed elsewhere (28). Two days later, the CAM was harvested and dissected by a sterile technique into pieces approximately 0.625 inch square for eventual transfer to the

1 Recipient of USPHS Research Grant CA 18260 and NIH Core Support Grant

17609 to Roswell Park Memorial Institute Cancer Cell Center.2 To whom requests for reprints should be addressed.3 Recipient of National Cancer Institute Contract NO1-CO-75380 with Litton

Bionetics Inc.Received November 1. 1979; accepted February 7. 1980.

' The abbreviations used are: CMEM. Eagle's minimal essential medium

supplemented with 10% fetal bovine serum, sodium pyruvate, nonessential aminoacids, L-glutamine, and 2-fold concentrated vitamin solution; [125l]ldUrd,[125l]iododeoxyuridine; CAM, chorioallantoic membrane; o.d., outside diameter;i.d., internal diameter; MEM, Eagle's CMEM (see above) without serum; HBSS,Hanks' balanced salt solution.

1636 CANCER RESEARCH VOL. 40

on May 25, 2021. © 1980 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

http://cancerres.aacrjournals.org/

Selection of Invasive Variants

invasion chamber. The latter, shown in cross-section in Chart

1, comprises an upper and lower chamber constructed from 2sterile Teflon rings. The edges of these rings were renderedsmooth by heating before use. To prepare the chamber, weplaced the lower ring (o.d., 0.375 inch; i.d., 0.25 inch; height,0.375 inch) in a sterile plastic Retri dish (100-mm diameter),and a 0.25-inch-thick layer of sterile nutrient agar was poured

around the ring for support. Nutrient agar was prepared asdescribed previously (28). The inside of the lower ring wasthen filled with nutrient agar, and a dissected piece of CAMwas placed over the lower ring (ectodermal surface facing up)after which the larger upper ring (i.d., 0.375 inch; height, 0.5inch) was placed in position to trap the CAM. A second layerof semisolid support agar was then poured into the dish aroundthe upper ring so that the 2 rings, plus the free edges of theCAM, were held in a firm agar support (Chart 1). B16 melanomacells (1 x 106) labeled with 125l-ldUrd were then added to the

upper chamber as a single-cell suspension in CMEM, and thechamber was incubated at 37Â°for up to 7 days. The structural

integrity of the CAM was measured at the end of each assayby aspirating the culture medium from the upper chamber (thisis not discarded; see below) and replacing it by 0.4 ml Indiaink (Pelikan, Hannover, W. Germany). Only the chambers retaining the ink in the upper compartment for at least 30 minwere considered acceptable. Preparations showing leakage ofink to the bottom compartment were rejected. With completionof this test, the upper ring and the CAM were removed fromthe chamber. The ' 25Iradioactivity in the CAM and in the culture

medium aspirated from the upper compartment (see above)was measured with a Packard gamma counter. The 125Iradio

activity of the contents of the lower compartment was thenmeasured and expressed as a percentage of the total radioactivity in the upper and lower compartments and the CAM.This figure represents the percentage of cells in the originalcell inoculum which had successfully penetrated the full thickness of the CAM. Several chambers can be prepared in thesame dish (Fig. 1) permitting replicate cell populations to betested in parallel.

For selection of tumor cell variants with increased invasiveproperties, cells added to the upper compartment are notradiolabeled. After cells were incubated in the chamber for 7days, the integrity of the CAM is tested using India ink as

Cells in fluidmedium

Teflon ringsnutrient agar

or

fluid medium

Chart 1. Cross-section of an invasion chamber system for quantitating tumorcell (O) invasion in the chick CAM and for the recovery of tumor cells thatsuccessfully penetrate the full thickness of the CAM. A detailed description ofthe structure and assembly of the chamber and the assay methods are given in"Materials and Methods."

Fig. 1. Retri dish containing 4 CAM invasion chambers (described in Chart 1)embedded in agar. Pigmented B16 melanoma cells can be seen growing in theindividual chambers.

Fig. 2. Individual CAM invasion chamber embedded in agar in which pig-mented colonies of B16 melanoma cells on the CAM can be seen. By focusing atdifferent levels, colonies could also be identified below the CAM in the agar inthe bottom chamber

described above, after which the upper ring and the CAM areremoved and discarded. Cells that had penetrated the CAMand grew as colonies in the agar in the bottom compartment(Fig. 2) were then harvested and cultured further in plasticflasks to obtain a sufficient number of cells, and 1 x 106 viable

cells were inoculated onto fresh CAM preparations. The cells

MAY 1980 1637



G. Posfe ef al.

penetrating the new CAM'S were, in turn, harvested and passaged again through fresh CAM'S up to 14 times. These cells

will be described by the prefix CP (CAM-passage) followed by

the number of times the cells had been passaged through CAM(e.g., B16-F1-CP10 = B16-F1 cells passaged 10 times

through the CAM).In most selection experiments, nutrient agar was used in the

lower chamber. Normal CAM cells were unable to grow in agar,but B16 tumor cells could form colonies in agar because oftheir loss of anchorage dependence. In certain experiments,fluid culture medium (CMEM) or factors which are chemotacticfor macrophages or leukocytes and/or certain tumor cells wereplaced in the bottom compartment to determine if penetrationof radiolabeled tumor cells was affected by such chemotacticstimuli. Bacterial culture filtrates containing leukotactic factorswere produced from cultures of Escherichia coli (WHO reference strain 16) as described by Ward ef al. (41). Serumfractions containing leukotactic factors were produced by zy-mosan activation of human serum in the presence of 1 M e-aminocaproic acid (40). Zymosan-activated sera were incu

bated with trypsin using the method described by Orref al. (25)to generate material containing factors from the third and fifthcomponents of complement which are chemotactic for sometumor cells (25). All of the above materials were added to MEMat a final concentration of 75 jul/ml. In other experiments, thebottom chamber was filled with cell-free culture supernatant

fluid harvested from the same tumor cell population added tothe upper chamber.

Assay of Tumor Cell Invasion in Mouse Urinary Bladderand Selection of Invasive Cell Variants. Adult male C57BL/6mice were killed by ether inhalation. The mice were submergedin a 5% Wescodyne solution (West Chemical Products, NewYork, N. Y.), immersed in 70% ethanol for 2 min, and placedinto a laminar flow hood. The abdominal cavity was opened ina sterile fashion, and the urogenital system was fully exposed(Fig. 3). Urine within the urinary bladder was expressed bygentle finger pressure (sterile surgical gloves were worn duringthese procedures), after which the urethra was occluded byapplication of a hemostat. One testis was displaced from thescrotal sac, and the ductus deferens was identified and spreadacross an open pair of forceps, with care being taken not torupture the duct. The ductus deferens was penetrated (Fig. 3)using a 27-gauge needle (Becton, Dickinson, and Company,

Rutherford, N. J.) attached to a tuberculin syringe, and 0.3 to0.4 ml of tumor cell suspension (0.9 to 1.2 x 106 cells) was

injected slowly. The injected cells reached the urethra via theemergence of the ductus deferens in the prostrate, and retrograde urethral flow carried suspended tumor cells to the bladder which gradually became filled with cells (Fig. 4). A ligatureof 2-0 chromic catgut (Detnatel, Queens Village, New York,

N. Y.) was placed around the neck of the bladder distal to theentrance of the ureters and tied tightly (Fig. 4). The bladderwas then excised from the abdominal cavity using hot scissors(flamed) in order to cauterize the cut tissue. The ligated isolatedbladder was then vigorously rinsed in 4 changes of HBSScontaining gentamicin sulfate (0.1 mg/ml; Schering Corporation, Kenilworth, N. J.). The excised bladders were then placedin sterile 60-mm plastic Petri dishes containing a preformed

base of semisolid agar produced by pouring 5 ml of agar (seebelow) into dishes followed by chilling at 4Â° for 10 min to

solidify the agar. The semisolid agar medium was prepared by

Fig. 3. Retrograde filling of murine urinary bladder with tumor cells via thevas deferens. The bladder has been emptied mechanically, and the ductusdeferens is cannulated by a 27-gauge needle.

Fig. 4. Tumor cell suspension (0.3 to 0 4 ml) has been injected into the ductusdeferens, and the urinary bladder is filled by retrograde flow. Hemostat occludesthe penile urethra and ligature is 2-0 catgut.

mixing 20 ml 2 x MEM, 5 ml fetal calf serum, 5 ml tryptosephosphate broth (all reagents from Grand Island BiologicalCo.), and 20 ml 1.25% agar(Difco Laboratories, Detroit, Mich.).The excised bladders were placed on the agar base, neckvertically up, and semisolid supporting agar, composed of 2volumes of agar base to 1 volume of CMEM, was added to apoint just below the ligature. Each bladder was restrained inthe vertical position by stretching the free ends of the ligatingcatgut at 180Â°to one another and placing the Petri dish lid on

the free ends. Thus, solidification of the agar maintained thebladders in the correct upright position. One bladder wasplaced in a single Petri dish, and between 5 and 10 bladderswere used to obtain cells at each passage level (see below).Individual bladders were removed from the agar at daily intervals after incubation at 37Â° in a humidified atmosphere con

taining 5% CO? for up to 5 days. The agar situated around thesite where the base of the bladder had rested was then pipettedup and added to 1 ml CMEM in a single well of a Costar 16-mmmultiwell dish (Costar, Cambridge, Mass.). These dishes wereincubated under the same conditions as described above, andthe wells were examined daily for tumor colonies. When tumor

1638 CANCER RESEARCH VOL 40




cells were identified in the wells, they were grown to con-

fluency, harvested, and cultured further in plastic flasks toobtain sufficient numbers of cells for reinjection into freshbladder preparations. This process was repeated 6 times; thetumor cell subline established from cells recovered from theagar after one pass was designated B16-F10-BL1 (BL1), andthe subline derived after 6 passages was designated B16-F10-BL6 (BL6). The B16-F10-BL1 subline was obtained from bladders incubated for 5 days, whereas the B16-F10-BL6 subline

was recovered from agar only 1 day after explanting bladders.In other experiments, bladders were filled with tumor cells

via the vas dÃ©feronsand excised as described above but wereused to prepare organ cultures rather than whole organ ex-

plants. Excised bladders were placed in sterile plastic Retridishes and covered with CMEM containing gentamicin sulfate(0.1 mg/ml) and incubated at 37Â° for 12 hr. Trial-and-error

histological studies established that this time is sufficient toallow for tumor cells in the lumen of the bladder to attach toand/or penetrate into the transitional epithelium lining of thebladder. Following this initial incubation period, the bladderswere cut into organ culture blocks as described by Reese efal. (31). The organ cultures were placed on sterile Falcon No.3014 stainless steel grids (Falcon Plastics, Oxnard, Calif.) inFalcon No. 3037 organ culture dishes. The central well of theorgan culture dishes was then filled with either nutrient agar orCMEM. The peripheral wells in the dishes were filled withsemisolid support agar to a level sufficient to trap the edges ofthe steel grid. After solidification of the agar, CMEM was addedto cover the epithelial surface of the cultures. For quantitativeinvasion assays, 125l-ldUrd-labeled tumor cells were injected

into the bladders. Organ cultures containing radiolabeled cellswere incubated for 7 days at 37Â° after which the culture

medium, the organ cultures, and the supporting grids were

BY

PE

1C

Chart 2. Cross-section of a perfusion invasion chamber (1C)system for quan-

titating tumor cell (â€¢)invasion in blood vessels (BV) and for the recovery of tumorcells that successfully invade and penetrate the vessel wall. The invasion chamberand the components of the perfusion system are not drawn to scale relative toeach other. Detailed descriptions of the chamber and the assay methods aregiven in "Materials and Methods." IP, injection port; PE, ultrahigh-molecular-

weight polyethylene; R, culture medium reservoir; P, peristaltic pump; GÃˆ, gasexchange reservoir.

Fig. 5. Perfusion apparatus for studying tumor cell invasion of blood vesselsin vitro. The components of the apparatus are shown in cross-section in Chart 2.Invasion chamber containing segment of vein surrounding a central ultrahigh-molecular-weight polyethylene tube; IP, injection ports for introduction of cellsinto outer chamber; P, peristaltic pump; PC, pump speed control; H. culturemedium reservoir (in this figure the medium contains large numbers of pigmentedcells which have successfully penetrated the vein from the outer chamber toreach the inner perfusion circuit); GÃˆ,gas exchange reservoir.

removed and their 125Iradioactivity was measured. 125Iradio

activity associated with tumor cells, which had penetrated thebladder organ cultures to reach the central well, was thenmeasured and expressed as a percentage of the total radioactivity. For selection experiments, organ cultures were prepared from bladders given injections of nonradiolabeled cells.Cultures were incubated for 7 days at 37Â° as above, after

which cells growing in the central well were harvested andcultivated further to generate sufficient cells for repassaging innew bladder organ cultures. Invasive cells recovered from thecentral well were passaged up to 10 times in bladder organcultures. The cells obtained in these experiments are designated by the prefix BP (bladder passage) followed by thenumber of passages the cells have undergone.

Assay of Tumor Cell Invasion in Veins and Selection ofInvasive Cell Variants. The principle of the culture systemused to isolate tumor cell variants capable of invading the wallof veins is shown in Chart 2. Portions of dog femoral vein werefitted onto a central tube of porous ultrahigh-molecular-weight

polyethylene with a pore size of 20 /im (Glassrock Products,Fairburn, Ga.). Culture medium was perfused through the center of this tube from a reservoir by a variable-speed peristalticpump with flow rates at 2 to 4 ml/min. Tumor cells wereinjected into the outer chamber surrounding the blood vesselvia injection ports fitted with self-sealing rubber caps. The

injected cells then attached to the surface of the vein. Segments of vein can be placed around the central tube with eitherthe endothelium facing inwards or as everted segments inwhich the endothelium is facing the injection ports. Cells thatare able to invade the wall of the vein will then penetrate thepores of the central tube and be recovered in the internalperfusion circuit.

The actual equipment is shown in Fig. 5. The main chamberis a modified Amicon Vitafiber Capillary Perfusion System(Amicon, Lexington, Mass.) in which the network of anisotropic

MAY 1980 1639



G. Posfe eÃal.

hollow fiber capillary bundles have been replaced by a 15-cmlength of ultrahigh-molecular-weight polyethylene tube. Thevein is fitted onto this tube before being placed into the chamber. Segments of dog femoral vein were perfused in situ withice-cold 0.9% NaCI solution before removal from the animal.Collateral vessels were tied off or occluded by a diathermyneedle before removing the main portion of vein. Veins werefitted onto the tube by inserting the tube into the lumen of thevessel for a sufficient distance to completely cover the tube.Two encircling catgut ties were then made at each end of thetube, and the excess vein was trimmed with a scalpel. The tubeand vessel were then placed into the polycarbonate outerchamber and sealed with screw-fit end pieces and siliconrubber sealing O-rings. Both ends of the tube were then connected to 0.132 inch i.d. x 0.183 inch o.d. gas-permeable

medical grade silicone tubing (Dow Corning Corp., Midland,Mich.) to create a closed perfusion circuit with an interveningreservoir of culture medium. The tubing between the reservoirand the input side of the chamber was passed through a 500-

ml conical flask containing a mixture of 5% CO2 in air to ensurecorrect gaseous exchange. Tumor cells (1 x 106) in CMEM

(1.5 ml) were then injected into the outer chamber, and theentire apparatus was incubated at 37Â° in a walk-in hot room.For invasion assays, 125l-ldUrd-labeled cells were injected into

the outer chamber, and the number of cells penetrating thewall of the vessel was determined by measuring the radioactivity recovered in the inner perfusion circuit. For selection experiments, nonradiolabeled cells were introduced, and invasivecells harvested from the internal circuit were reinjected intochambers containing fresh vessels. The invasive cell populations recovered by serial passage through veins will be referredto by the prefix BV (blood vessel) followed by the number oftimes the cells have been passaged through veins.

Animals. Specific-pathogen-free C57BL/6 mice were obtained from the Animal Production Area, Frederick CancerResearch Center, and the West Seneca Laboratories, RoswellPark Memorial Institute.

Metastasis Formation. The ability of cells to form experimental mÃ©tastasesafter i.v. injection was measured by injecting5 x 10" to 5 x 105 viable cells as a single-cell suspension in

0.2 ml HBSS into the tail vein of unanesthetized adult C57BL/6 mice matched for age, weight, and sex. Mice were killed 18or 21 days later, and the number of lung mÃ©tastases wasdetermined with a dissecting microscope as described previously (6, 9).

The following methods were used to assay the ability of cellsto form spontaneous mÃ©tastases:(a) C57BL/6 mice were givens.c. injections in the external ear of 5 x 104 cells in 0.1 ml

HBSS, and the number of lung mÃ©tastaseswas determined 4weeks later; (b) C57BL/6 mice were given s.c. injections inthe ear as above. The ear plus the growing tumor were thenamputated 10 days later (7), and formation of regional lymphnode and lung mÃ©tastaseswas determined 4 to 6 weeks later;or (c) C57BL/6 mice were given i.m. injections of 2.5 x 104

viable cells in 0.5 ml HBSS in the footpad of the hind leg. Insome animals, the "primary" tumor was amputated after 4

weeks, and metastasis formation was determined 4 weeksthereafter. In other animals, the "primary" tumor was not

amputated, and metastasis formation was measured 6 weeksafter initial tumor cell injection. In all of the protocols, micewere matched for age, weight, and sex.

RESULTS

In Vitro Selection of B16 Melanoma Cell Variants withIncreased Ability to Invade CAM. The ability of B16-F1 mela

noma cells to invade and penetrate CAM maintained in vitroand the invasive behavior of variants selected from B16-F1

cells by serial passage of cells that successfully penetrate theCAM are shown in Table 1. The results indicate that theproportion of cells in the parent cell population that penetratethrough the CAM to reach the lower chamber (Fig. 2) is verylow. Serial passage of the invasive cell fractions recoveredfrom the lower chamber produced a significant increase in theproportion of cells penetrating the CAM. Maximum efficiencyof invasion was achieved after 10 passages through the CAM(Table 1).

Confirmation that invasion in this system involves selection(enrichment) of subpopulations with increasing invasive properties was provided by the finding that the "nonpenetrating"cell fractions recovered from CAM'S inoculated with early-pas

sage populations contained significantly fewer invasive cellsthan did the starting inoculum (Table 1). With increasing CAMpassage, however, the proportion of invasive cells recoveredin the nonpenetrating fraction increased. After 10 passages,the efficiency of invasion by the reinoculated nonpenetratingfraction was similar to that of the original inoculum, suggestingthat the cell population was now composed almost exclusivelyof invasive cells. Thus, in early passages, invasion was anonrandom process which selects for invasive cell subpopula-

tions. With subsequent passaging, the proportion of invasivecells in the population was enriched; by 10 passages, invasionhad become a random process and the population consistedof cells with invasive properties.

The invasive behavior of variants at any given passage levelis highly reproducible, and significant variation has not beendetected using different CAM preparations. In addition, assays

Table 1Invasiveness of B16-FT mouse melanoma cells during serial passage in chick

CAM in vitroAliquots of 1 x 106 viable B16-F1 melanoma cells in CMEM were added to

the ectodermal surface of portions of CAM in invasion chambers. Cells penetrating the CAM were recovered from the bottom chamber of the invasion apparatus7 days later and reinoculated onto fresh CAM preparations. This was repeatedup to 14 times. At each passage level, [I25l]ldUrd-labeled cells (penetration

assay) and nonradiolabeled cells (serial passaging) derived from a commonparent culture were used.

Passageno.1

246

81014%

of inoculated cells penetrating theCAM2.4

Â± 0.6Â°

4.1 Â± 1.1

11.6 Â± 2.317.4 Â± 3.7

23.7 Â± 4.4

31.6 Â±4.930.6 Â± 4.2%of

"nonpenetrating

fraction penetratingCAM6Not

detectable

1.3 Â±0.4

3.9 Â±0.813.7 Â± 2.1

18.4 Â± 2.727.6 Â± 4.333.4 Â± 4.6"

cell

new

" Cell-associated [125l]ldUrd radioactivity recovered in the bottom chamber as

a percentage of total radioactivity in the original cell inoculum. Inocula containedbetween 3.5 and 6 x 10e cpm/106 cells.

b Nonradiolabeled cells (1 x 106) were added to CAM invasion chambersand incubated for 7 days at 37Â°after which "nonpenetrating" cells which failed

to reach the bottom chamber were recovered from the culture medium in theupper chamber and from the CAM (see "Materials and Methods"). These werecultured in vitro to generate sufficient numbers of cells, labeled with ['25l]ldUrd.

and reinoculated onto fresh CAM to assay cell penetration as described inFootnote a.

0 Mean Â±S.E. derived from measurements on 8 replicate chambers at each

passage.





of the invasiveness of variants cultivated in vitro for up to 4months after initial isolation or stored in liquid nitrogen for upto 6 months after initial isolation revealed no significantchanges in behavior from that shown in Table 1.

The percentage of cells penetrating the CAM at any givenpassage level could not be altered significantly by substitutingthe semisolid agar in the bottom chamber by: (a) fluid culturemedium (CMEM) containing different concentrations of fetalcalf serum (2, 5, 10, or 20%); (b) cell-free culture supernatant

fluids from homologous B16 cell populations; (c) homogenizedextract of B16 tumor; or (d) MEM containing components whichare chemotactic for leukocytes and macrophages (zymosan-

activated serum; E. coli culture fluids) and certain types oftumor cells (trypsin-digested zymosan-activated serum) (re

sults not shown).Penetration of the CAM requires cells to be viable. I25l-ldUrd-

labeled B16 cells killed by heating at 80Â° for 15 min before

seeding onto the CAM were unable to penetrate. Histologicaland radioautographic studies revealed that these cells werenot able to infiltrate the CAM although they adhered to thesurface of ectodermal epithelium. Successful penetration ofthe CAM also requires more than active cell motility. Studiesusing macrophages, which are known to infiltrate a wide rangeof normal tissues, revealed that these cells readily infiltratedthe mesoderm of the CAM but did not penetrate the full thickness of the CAM to reach the lower chamber (results notshown).

In Vitro Selection of B16 Melanoma Cell Variants withIncreased Ability to Invade Urinary Bladder Tissue and Veins.Structurally, the CAM is a relatively simple tissue, and successful invasion does not require tumor cells to breach substantial mechanical barriers such as thick basement membranes, compact fibrous stroma, or muscle layers. Formidablebarriers of this kind are routinely encountered by invadingtumor cells in vivo (17-22). It was therefore considered worth

while to see if tumor cell variants could be selected for invasionin tissues containing mechanical barriers of this kind. Urinarybladder and vein were selected since their walls contain all ofthe barriers listed above. Invasion of blood vessels is alsorelevant to the pathogenesis of hematogenous mÃ©tastases inwhich malignant cells must first penetrate into blood vesselsand, after dissemination and arrest in the circulation, cross thewall of blood vessels to reach the extravascular tissues andform mÃ©tastases(17, 18).

Invasive cell variants have been selected from the B16-F1

melanoma line by selection and passaging of cells which successfully penetrate organ cultures of full-thickness murine uri

nary bladder wall (Table 2). In contrast to the emergence ofinvasive variants from B16-F1 cells during passaging with CAM

in which the proportion of invasive cells increased relativelyslowly with successive passages, the invasive variants selectedin bladder emerged much more rapidly. After only 4 passages,the variant population was composed almost exclusively ofcells with invasive potential. Although only approximately 25 to35% of the cells passaged 4 or more times penetrated thebladder wall in any given assay (Table 2), the noninvasivefraction in these populations contained large numbers of cellswith invasive properties. This was demonstrated by assayingthe invasive potential of tumor cells recovered from the lumenof injected bladders. These cells have failed to penetrate thetransitional epithelium during initial incubation of cells within

Table 2Invasiveness of B16-F1 mouse melanoma cells during serial passage in bladder

organ cultures from C57BL/6 miceBladders of C57BL/6 mice were given injections of 1 x 106 viable B16-F1

melanoma cells and incubated for 12 hr at 37Â°after which organ cultures wereprepared. Organ cultures were incubated for 7 days at 37Â°after which any cells

that had penetrated the full thickness of the organ culture to reach the centralwell of the organ culture dish were recovered. After further short cultivation invitro, these cells were reinjected into bladders under the same conditions, andcells penetrating the organ culture were recovered once again. This process wasrepeated up to 10 times. Replicate populations of (125l]ldllrd-labeled cells and

nonlabeled cells from the same parent culture were used at each passage.[125l]ldUrd-labeled cells were used to quantitate cell penetration, and nonlabeled

cells were used for serial passaging.

Passageno.123468%

of inoculated cells penetrating bladderwall"1.3

Â±0.4Â°16.4

Â±2.729.7Â±4.532.5

Â±5.234.8Â±4.330.7Â± 4.2%

of"nonpenetrating1cellfraction

penetrating bladderwall*NO"4.9

Â±2.415.4Â±2.723.4Â±3.629.3Â±4.032.6Â±5.1

a Cell-associated [l25l]ldUrd radioactivity recovered in the central well as a

percentage of the total radioactivity in the original cell inoculum.6 Bladders were given injections of ['25l]ldUrd-labeled cells as described in

Footnote a and incubated for 12 hr at 37Â°. The contents of the bladder lumen

were then aspirated. Cells recovered in the aspirates from several bladders werepooled, aliquots containing 1 x 106 cells were reinjected into fresh bladders,

and their ability to penetrate organ cultures prepared from the injected bladderswas measured as described in Footnotes a and b.

c Mean Â±S.E. The results was derived from measurements on 10 organ

cultures derived from bladders given injections of the same cell population.d ND, not done.

the closed bladder. However, when reinoculated into bladders,large numbers of invasive cells could still be detected in thispopulation. As shown in Table 2, after cells have been passaged 4 or more times through the bladder, the invasivenessof "non-penetrating" cells recovered from the lumen is equivalent to that of the "penetrating" cell fraction (Table 2). This

indicates that the population has been enriched for invasivecells which now constitute the majority of cells.

Variant B16 melanoma sublines with increased ability toinvade mouse bladder have also been isolated from the B16-

F10 line by selecting for cells which are able to penetrate thewall of intact bladders explanted into agar in vitro. The properties of one variant, designated BL6, selected by 6 serialpassages through intact bladder will be discussed in moredetail later in this paper.

The same strategy used to select B16 melanoma cell variantsfor increased invasiveness in CAM and bladder has also beensuccessful in selecting variants from the B16-F1 line that

display increased abilities to invade the wall of segments ofdog femoral vein maintained in a perfusion culture system. Incommon with invasive variants selected in bladder organ cultures, only a few passages were needed through this tissue toisolate variant cell populations which contain a high proportionof invasive cells (Table 3). After only 4 passages, the variantpopulation was composed almost exclusively of invasive cells,as shown by the fact that recovery and reinoculation of the"noninvasive" fraction revealed the presence of large numbers

of invasive cells (Table 3).As in the case of CAM-selected variants, the invasive behav

ior of bladder- and blood vessel-selected variants (Tables 2

and 3) is highly reproducible in separate assays and remainsstable in cells grown in vitro for up to 3 months or stored inliquid nitrogen for up to 6 months following initial isolation.

MAY 1980 1641



G. Poste ef al.

Invasion of Different Tissues by Variants Selected in CAM,Bladder, and Vein. Selection of cells for Â¡nvasivenessin onetissue may not necessarily enable them to invade other tissues.This was demonstrated by showing that CAM-passaged inva

sive variants did not differ from unselected parent F1 cells intheir ability to invade bladder or vein (Table 4). In contrast,bladder- and vein-passaged variants were highly invasive in

the CAM, urinary bladder, and vein preparations (Table 4).Formation of Spontaneous and Experimental MÃ©tastases

by Invasive Variants. Assay of the ability of CAM-passagedinvasive cell variants to form experimental mÃ©tastasesafter i.v.injection or spontaneous mÃ©tastases after s.c. injection intothe external ear revealed that their metastatic activity did notdiffer significantly from parental B16-F1 cells (Table 5). In

contrast, cells from invasive variants selected by passaging inbladder organ culture (BP series) or perfused vein (BV series)produced significantly more pulmonary mÃ©tastases than didparent B16-F1 cells following both i.v. (experimental metas

tasis) and s.c. (spontaneous metastasis) injection (Table 5).Indeed, the metastatic activity of these invasive variants iscomparable to that of the highly metastatic B16-F10 line whichwas selected from the B16-F1 line for its ability to produce lung

mÃ©tastasesfollowing i.v. injection (5).

Table 3Invasiveness of B16-Ft mouse melanoma cells during serial passage in

segments of dog femoral vein maintained in a perfusion apparatus in vitro

Segments of dog femoral vein in a perfusion culture apparatus were injectedwith 1 x 106 B16-F1 cells and maintained for 2 weeks at 37Â°. Cells that

penetrated the blood vessel wall were recovered from the inner perfusion circuit(see "Materials and Methods") and reinjected onto fresh vein preparations. Thisprocess was repeated up to 10 times. Replicate populations of [1D6l]ldUrd-labeled

cells (penetration assay) and nonlabeled cells (serial passaging) from the sameparent culture were used at each passage.

% of inoculated cells pÃ©nÃ©trÃ¢t-% of "nonpenetrating" cellPassage no. ing vessel wall3 fraction penetrating newvessel61234681

.6 Â±0.4C19.3

Â±4.232.4Â±4.936.9Â±5.635.8Â±3.731.6Â± 4.4011.3

Â±2.920.7Â±3.630.7Â±4.831.3Â±4.232.7Â± 5.0

"Cell-associated [I25l]ldllrd radioactivity recovered in the inner perfusion

circuit as a percentage of the total radioactivity in the original cell inoculum.b Cells were injected into the outer chamber as in Footnote a, and the

chamber was incubated for 4 days at 37Â°.Cells in the outer chamber which failed

to attach to the vessel were aspirated, cultivated in vitro for a short period toprovide sufficient cells, labeled with [125l]ldUrd, and reinjected onto fresh vessels,

and their penetration was assayed as in Footnote b.c Mean Â±S.E. derived from measurements on 2 chambers at each passage.

Table 4

Invasiveness of B / 6 melanoma cell variants in different tissues in vitro

% of cellpenetrationCell"Parent

B16-F1B16-F1-CP10B16-F1-BP8B16-F1-CP10. BP4B16-F1-BV6CAM2.3

Â±0.5C

29.8 Â±4.338.2 Â±5.130.9 Â±4.047.2 Â±6.8Bladder1.3

Â±0.41.9 Â±0.6

30.7 Â±4.322.6 Â±4.626.5 Â±4.1Vein1.6

Â±0.42.3 Â±0.5

24.7 Â±3.914.3 Â±3.133.2 Â±4.5

The suffix designations for B16-F1 cells are: CP10, 10 passages in CAM;BP8, 8 passages in bladder organ cultures; CP10, BP4, 10 passages in CAMfollowed by 4 passages in bladder organ cultures; BV6, 6 passages through veinpreparations.

Preparations of the indicated cell types labeled with [125l]ldUrd were assayed

for their ability to penetrate CAM, mouse bladder organ cultures, and perfuseddog femoral vein as described in Tables 1, 2, and 3, respectively.

c Mean Â±S.E. derived from measurements on 4 replicate assay preparations

for CAM and bladder and 2 for vein.

The BL6 variant selected from B16-F10 cells by passagingin intact bladder is also more efficient than the parental B16-

F10 cells in producing spontaneous mÃ©tastases. However,unlike the BP variants passaged in bladder organ culture, theBL6 variant is less efficient than the parent B16-F10 line in

producing experimental mÃ©tastases after i.v. injection (Table6).

DISCUSSION

A variety of in vitro systems have been used to study invasionof normal tissues by malignant tumor cells. Attempts to studyinvasion by allowing normal and tumor cell populations tointeract in monolayer culture (1 -3, 13, 15) are probably over

simplified since tumor cells are not required to invade anorganized tissue matrix. Invasion of tumor cells into aggregatesof normal cells (10, 16) suffers from s'Tiilar disadvantages. In

addition, doubts must be expressed about the functional significance of tumor cell infiltration into aggregates in view of

Table 5Metastasis formation by cell variants selected from B16-F1 melanoma cells

MetastasisformationCell"B16-F1

B16-F10B16-F1-CP10B16-F1-BP8B16-F1-CP10. BP4B16-F1-BV4B16-F1-BV8Spontaneous2/20(10)"

12/20(60)2/19(11)

14/20(70)10/19(53)13/18(72)17/20(85)Median

no. of pulmonarycolonies09

(2-27)e

111 (23-196)7(0-31)

92(18-210)47(12-129)

104(31-224)123(27-185)

The suffix designations for B16-F1 cells are as in Table 4, Footnote a.Number of animals with lung mÃ©tastasesand total number of animals given

injections. C57BL/6 mice were given injections of 5 x 104 viable cells in the

external ear (s.c.). the injected ear was amputated 10 days later, and metastasisformation was assayed 4 weeks after amputation.

c C57BL/6 mice given i. v. injections of 5 x 104 cells, and metastasis

formation was assayed 18 days later. Results derived from measurements on 10animals/group.

d Numbers in parentheses, percentage.e Numbers in parentheses, range.

Table 6Metastasis formation by an invasive cell variant (BL6) derived from B16-F10

melanoma cells

Route of injection Cell Experiment 1 Experiment 2

No. of mice bearing lung tumornoduless.c.a

(externalear)i.m.

(footpad)B16-F10

B16-BL6B16-F10

B16-BL63/9

(33)"6/8(75)"2/7

(29)b'c7/9 (78)b'c1/12(17)6

7/10(70)1

/9 (22)"-d4/5 (SO)6'"

Median no. of lung colonies

i.v. B16-F10B16-BL6

441 (170->500)148(43-274)"''

231 (66-481 )e'983(16-258)e9

Mice were given injections of 5 x 10" viable cells in the external ear s.c.

and autopsied 4 weeks after injection.Numbers in parentheses, percentage.

c Mice given injections of 2.5 x 10" viable cells in the footpad; primary lesion

amputated after 4 weeks and mice killed 4 weeks later.Autopsied 6 weeks after injection without amputation.Number of lung nodules formed 3 weeks after i.v. injection of 5 x 105 cells

median. Numbers in parentheses, range.Experiment 1. 8 animals/group.

9 Experiment 2, 15 animals/group.





recent data showing that identical infiltrative behavior is displayed by inert particles such as Sephadex beads (42). In vitrostudies on the ability of tumor cells to invade explanted tissuessuch as chick blastoderm (21), CAM (4, 12, 14, 24), heart(22), mesonephros (30), skin (24), wingbud (39), murine lung(35), or human decidual tissue (36) are more relevant to tumorinvasion in vivo, but most of these assays do not permit accurate quantitation of invasion or recovery of the invasive cells.The assay methods reported in this paper satisfy both theserequirements.

We describe here new methods for the selection of tumorcell variants with increased abilities to invade a variety oftissues in vitro. By comparing sublines selected in differenttissues, information can be obtained on whether a common setof cellular properties are selected or if the properties expressedby invasive cells depends on the tissue in which the selectionwas made. Our results indicate that the invasive variants selected in vitro in different tissues vary significantly in theirproperties, despite being derived from the same parent cellline. The ability to isolate cells with different invasive pheno-types from a common parent cell population is consistent withthe growing body of evidence that many malignant tumors areheterogeneous and contain subpopulations of cells with widelydiffering invasive and metastatic properties (for reviews, seeRefs. 7 to 9 and 27).

Our results demonstrate that B16 melanoma variants selected for invasiveness in the chick CAM do not exhibit comparable invasive behavior in bladder tissue or veins. In contrast,invasive variants selected in the latter 2 tissues are highlyinvasive in all 3 tissues. Successful penetration of the CAM bytumor cells might be considerably less demanding since theCAM does not offer a formidable mechanical barrier comparable to the bladder wall or the medial and adventitial elementsof veins. Invasive variants selected by passaging in bladder orvein have thus probably been subjected to a more demandingset of selection pressures than were CAM-passaged variants.

This may also explain the more rapid selection (lower numberof passages) of highly invasive cells in bladder and vein compared with CAM. If bladder and vein do indeed constitute moredemanding selection systems than the CAM, the phenotypesthat befit cells for invasion in bladder and vein would beexpected to be more specialized and thus probably would bepresent in the parent cell population at a lower frequency thanthe phenotypes suitable for invasion of CAM. If this interpretation is correct, variant cell populations selected in bladder andvein would from the outset be more uniform than their counterparts selected in CAM. The restricted diversity of the invasivephenotype in bladder- and vein-passaged cells would mean

that cell subpopulations with highly invasive properties wouldquickly become dominant. In contrast, the greater heterogeneity of invasive phenotypes in CAM selected cells would meanthat overgrowth and dominance of a single or a few highlyinvasive phenotype(s) could take much longer.

The mechanisms involved in tissue invasion by tumor cellsare not known. Release of tissue destructive enzymes fromtumor cells, particularly lysosomal hydrolases and collageno-

lytic enzymes (for reviews, see Refs. 26, 31, and 38), has beenproposed as one possible mechanism. Examination of theenzyme profiles of invasive variants of the type isolated in thepresent study may be useful in defining what role lytic enzymesplay in tissue invasion. Two reservations must be expressed,

however, about attempts to correlate cellular enzymic activitywith invasive potential: (a) secretion of enzymes in vitro maymerely reflect the physiological potential of cells in culture andhave little relevance to events in vivo; (b) the relative importanceof different enzymes may not only vary between different typesof tumor cells but also be influenced by the nature of thetissue(s) encountered by the tumor cells (see below). Forexample, enhanced production and release of plasminogenactivator accompanies neoplastic transformation of a diverserange of cell types (33), and it has been speculated that thisenzyme might facilitate tumor cell invasion (32, 37). Definitiveevidence to support this proposal is yet to be presented.Studies in this (11 ) and other laboratories (23) indicate that, atleast for B16 melanoma cells, increased production and secretion of plasminogen activator is not correlated with enhancedinvasive or metastatic behavior.

The ability of invasive tumor cells to invade into, and escapefrom, blood vessels may be correlated with their ability todegrade the basement membrane of capillaries and venules (8,17, 18, 27). The predominant form of collagen in the capillarybasement membrane is type IV collagen and it is, therefore, ofinterest that the B16-BL6 variant described in this paper hasbeen found to produce high levels of type IV collagenasecompared with the parent B16-F10 line (19). The B16-BP8 and

BV8 variants isolated in the present study also show a significant increase in collagenolytic activity compared with the F1parent cells.5 CAM-passaged variants lack collagenolytic activ

ity, but they rapidly acquire this property when passaged ineither bladder or vein.

The present results indicate that variant sublines selected invitro for increased invasiveness in bladder or vein also displaya greater ability to form spontaneous mÃ©tastasesin vivo thandid their parental cells. This suggests that the factors responsible for the isolation of cells with the invasive phenotypecoincidentally select for cellular properties required for thesuccessful completion of other steps in the metastatic process.The variants selected in vein or organ cultures of bladder (BPseries) are also more metastatic than are parent B16-F1 cells

when injected i.v. In contrast, the variant selected in intactbladder (BL6) is less metastatic than is its parent line (B16-

F10) when injected i.v. We consider that these differences inthe ability of bladder-selected variants to form experimental

mÃ©tastasesafter i.v. injection results from the fact that they arederived from different parent cells. In the case of organ culturepassaged variants derived from the B16-F1 line, selection for

invasion results in cells with enhanced metastatic ability. Detection of the latter is facilitated, however, by the low metastaticactivity of the parent cells. In contrast, the BL6 variant wasselected from the F10 line which is already highly metastaticwhen injected i.v. (6). Formation of experimental mÃ©tastasesby cells injected directly into the circulation eliminates the needfor primary invasion, and the outcome of metastasis in thissituation may depend more on factors such as cell survival inthe bloodstream, deformability, embolization, arrest, extravasation, and multiplication in organ parenchyma rather thaninvasive capacity. Under these conditions, it is perhaps notsurprising that the parent F10 cells, selected for preciselythese properties, are more successful in producing experimental mÃ©tastases.However, when the BL6 variant is placed s.c.

5 G. Poste, unpublished observations.

MAY 1980 1643



G. Posfe et al.

or Â¡.m.,where invasiveness is a prerequisite for metastasis, itis able to exploit its advantage over the F10 cells and metas-

tasize more effectively.The importance of the selection procedure in determining

the invasive phenotype is also illustrated by the present resultsin the properties of CAM-passaged cells. These variants do notdiffer from parental B16-F1 cells in their ability to invade other

tissue (i.e., bladder or vein). In addition, invasive variantsselected in the CAM do not differ from the B16-F1 parent cellsin their metastatic behavior in vivo. Comparison of CAM-se-

lected variants, in which invasiveness and metastatic ability arefunctionally dissociated, with variant sublines selected in bladder and vein in which these traits are coupled could provideinsight into the contribution of specific cellular alterations tothese particular behavior properties.

As emphasized elsewhere (8, 27), the search for propertiesuniform to all invasive cells may well be unproductive. Weconsider, however, that the likelihood of identifying propertieswhich are expressed with a relatively high degree of consistency in invasive tumor cells will be substantially increased bystudying variant cell populations selected specifically for thisphenotype.

REFERENCES

1. Abercrombie, M., and Heaysman, J. E. M. Invasive behavior between sarcoma and fibroblast populations in cell culture. J. Nati. Cancer Inst.. 56.561-569, 1976.

2. Armstrong, P. B., and Lackie. J. M. Studies on intercellular invasion in vitrousing rabbit peritoneal neutrophil granulocytes (PMNS). I. Role of contactinhibition of locomotion. J. Cell Biol., 65: 439-462, 1975.

3. Barski. G., Leon, B., Le Francois, D., and Belehradek, J., Jr. Cell movementas related to oncognesis. Blood Cells, 2. 453-466, 1976.

4. Easty. D. M., and Easty, G. C. Measurement of the ability of cells to infiltratenormal tissues in vitro. Br. J. Cancer, 29: 36-49, 1974.

5. Fidler, I. J. Metastasis: quantitative analysis of distribution and fate of tumoremboli labeled with '25l-5-iodo-2'-deoxyuridine. J. Nati. Cancer Inst., 45:

775-782, 1970.6. Fidler, I. J. Selection of successive tumor lines for metastasis. Nat. New

Biol., 242: 148-149, 1973.7. Fidler. I. J. Tumor heterogeneity and the biology of cancer invasion and

metastasis. Cancer Res., 38: 2651-2660, 1978.8. Fidler, l. J., Gersten, D. M.. and Hart, I. R. The biology of cancer invasion

and metastasis. Adv. Cancer Res., 28: 149-250, 1978.9. Fidler, I. J., and Kripke, M. L. Metastasis results from preexisting variant

cells within a malignant tumor. Science (Wash. D.C.), 797: 893-895. 1977.10. Gershman, H., and Rosen, J. J. Cell adhesion and cell surface topography

in aggregates of 3T3 and SV40-virus-transformed 3T3 cells: visualization ofinterior cells by scanning electron microscopy. J. Cell Biol., 76 639-651,

1978.11. Hart, I. R. The selection and characterization of an invasive variant of the

B16 melanoma. Am. J. Pathol.. 97: 587-600, 1979.12. Hart, I. R., and Fidler, I. J. An in vitro quantitative assay for tumor cell

invasion. Cancer Res., 38: 3218-3224, 1978.13. Knyrim, K., and Paweletz, N. Cell interactions in a "bilayer" of tumor cells:

a scanning electron microscope study. Virchows Arch. B Cell Pathol., 25.309-325, 1977.

14. Leighton. J. Invasive growth and metastasis in tissue culture systems.Methods Cancer Res., 4: 85-124, 1968.

15. Leung, T.-K., and Babai, F. Etude ultrastructurale de l'invasion des cultures

primaires du foie par des cellules cancÃ©reuses. Virchows Arch. B CellPathol.. 29: 267-280, 1979.

16. Liebrich, W., and Paweletz, N. Ultrastructure of aggregates of tumor andnormal cells and of mixed aggregates. Virchows Arch. B Cell Pathol., 22:39-53, 1976.

17. Liotta, L. A., Abe, S., Robey, P. G., and Martin, G. R. Preferential digestion

of basement membrane collagen by an enzyme derived from a metastaticmurine tumor. Proc. Nati. Acad. Sei. U. S. A., 76: 2268-2272, 1979.

18. Liotta, L. A., Kleinerman, J., Catanzaso, P., and Rynbrandt. D. Degradationof basement membrane by murine tumor cells. J. Nati. Cancer Inst., 58.1420-1431, 1977.

19. Liotta, L. A., Tryggrason, K., Garbisa, S.. Hart, I., and Folts, C. M. Metastaticpotential correlates with enzymatic degradation of basement membranecollagen. Nature (Lond.), 284: 67-68, 1980.

20. Mareel, M., de Bruyne, G., and de Ridder, L. Invasion of malignant cells into5'Cr-labeled host tissues in organotypical culture. Oncology (Basel), 34: 6-

9, 1977.21. Mareel, M., de Ridder, L., de Brabander, M., and Vakaet, L. Characterization

of spontaneous, chemical and viral transformants of a C3H/3T3 type mousecell line by transplantation into young chick blastoderms. J. Nati. CancerInst., 54: 923-929, 1975.

22. Mareel, M., Kint, J., and Meyvisch, C. Methods of study of the invasion ofmalignant C3H-mouse fibroblasts into embryonic chick heart in vitro. Virchows Arch. BCell Pathol.,30: 95-111, 1979.

23. Nicolson, G. L., Winkelhake, J. L., and Nussey, A. C. An approach tostudying the cellular properties associated with metastasis. In: L. Weiss(ed.). Fundamental Aspects of Metastasis, pp. 291 -305. Amsterdam: NorthHolland Publishing Co., 1976.

24. Noguchi, P. D., Johnson, J. B., O'Donnell, R.. and Petricciani. J. C. Chick

embryonic skin as a rapid organ culture assay for cellular neoplasia. Science(Wash. D.C.), 199: 980-983, 1978.

25. Orr, W., Varani, J., and Ward, P. A. Characteristics of the chemotacticresponse of neoplastic cells to a factor derived from the fifth component ofcomplement. Am. J. Pathol., 93: 405-420, 1978.

26. Poole, A. R. Tumor lysosomal enzymes and invasive growth. In: J. T. Dingle(ed.), Lysosomes in Biology and Pathology, Vol. 3, pp. 303-337. Amsterdam: North Holland Publishing Co., 1973.

27. Poste, G., and Fidler, I. J. The pathogenesis of cancer metastasis. Nature(Lond.), 283: 139-146. 1980.

28. Poste, G., and Flood, M. K. Cells transformed by temperature-sensitivemutants of avian sarcoma virus cause tumors in vivo at permissive andnonpermissive temperatures. Cell, Ã•7: 789-800, 1979.

29. Poste, G., and Kirsh, R. L. Rapid decay of tumoricidal activity and loss ofresponsiveness to lymphokines in inflammatory macrophages. Cancer Res.,39: 2582-2590, 1979.

30. Pourreau-Schneider, N., Felix, H., Haemmerli, G.. and Stranii. P. The role of

cellular locomotion in leukemic infiltration: an organ culture study on penetration of L5222 rat leukemia cells into the chick embryo mesonephros.Virchows Arch. B Cell Pathol., 23. 257-264, 1977.

31. Reese, D. H., Friedman, R. D., Smith, J. M., and Sporn, M. B. Organ cultureof normal and carcinogen-treated rat bladder. Cancer Res., 36: 2525-2527,1976.

32. Reich, E. Tumor-associated fibrinolysis. Fed. Proc. 32: 2174-2175, 1973.33. Roblin, R., Chou, I. N., and Black, P. H. Proteolytic enzymes, cell surface

charge and viral transformation. Adv. Cancer Res.. 22: 203-260, 1975.34. Scher, C. D., Haudenschild, C., and Klagsbrun, M. The chick chorioallantoic

membrane as a model system for the study of tissue invasion by viraltransformed cells. Cell, 8: 373-382, 1976.

35. Schirrmacher. V., Shantz, G., Clauer, K., Komitowski, D., Zimmermann, H.-P., and Lohmann-Matthes, M.-L. Tumor mÃ©tastasesand cell-mediated immunity in a model system in DBA/2 mice. I. Tumor invasiveness in w'fro and

metastasis formation in Vivo. Int. J. Cancer, 23: 233-244, 1979.36. Schleich, A. B., Frick, M., and Mayer. A. Patterns of invasive growth in vitro.

Human decidua graviditatis confronted with established human cell linesand primary human expiants. J. Nati. Cancer Inst., 56: 221-237. 1976.

37. Sherman. M. I., Strickland, S.. and Reich, E. Differentiation of early mouseembryonic and teratocarcinoma cells in vitro: Plasminogen Activator Production. Cancer Res., 36: 4208-4216, 1976.

38. SylvÃ©n,B. Biochemical and enzymatic factors involved in cellular detachment. In: S. Garattini and G. Franchi (eds.), Chemotherapy of CancerDissemination and Metastasis, pp. 129-138. New York: Raven Press, 1973.

39. Tickle, C., Crawley. A., and Goodman, M. Cell movement and the mechanismof invasiveness: a survey of the behaviour of some normal and malignantcells implanted into the developing chick wing bud. J. Cell Sci., 31: 293-322, 1978.

40. VallÃ³la, E. H., and MÃ¼ller-Eberhard, H. J. Formation of C3a and C5aanaphylatoxins in whole human serum after inhibition of the anaphylatoxininactivator. J. Exp. Med., 737: 1109-1123, 1973.

41. Ward, P. A.. Lepow, I. H.. and Newman, L. J. Bacterial factors chemotacticfor polymorphonuclear leukocytes. Am. J. Pathol., 52: 725-736, 1968.

42. Wiseman, L. L. Contact inhibition and the movement of metal, glass andplastic beads within solid tissues. Experientia (Basel), 34: 26-28, 1976.




1980;40:1636-1644. Cancer Res G. Poste, J. Doll, I. R. Hart, et al. Enhanced Tissue-invasive Properties

Selection of Murine B16 Melanoma Variants withIn Vitro

Updated version

http://cancerres.aacrjournals.org/content/40/5/1636

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/40/5/1636To request permission to re-use all or part of this article, use this link



http://cancerres.aacrjournals.org/cgi/alerts

mailto:[email protected]



in vitro selection of murine b16 melanoma variants with … · new assay methods have been devised...

Documents