[CANCER RESEARCH 40, 4571-4580, December 1980]0008-5472/80/0040-OOOOS02.00

Development of an in Vitro and in Vivo Epithelial Tumor Model for theStudy of Invasion1

Bendicht U. Pauli,2 Steven N. Anderson, Vincent A. Memoli, and Klaus E. Kuettner

Departments of Pathology [B. U. P., S. N. A., V. A. M.¡,Orthopedic Surgery ¡K.E. K.J. and Biochemistry [K. E. K.I Rush Medical College and Rush-Presbyterian-St. Luke's Medical Center, Chicago, Illinois 60612

ABSTRACT

Three continuous cell lines were isolated from A/-[4-(5-nitro-2-furyl)-2-thiazolyl]formamide (FANFT)-induced carcinomas of

the Fischer rat urinary bladder by standard expiant techniques.RBTCC-2 carcinoma cells were derived from a noninvasiveFANFT tumor of Stage 0, RBTCC-5 carcinoma cells were froman invasive FANFT tumor of Stage B2, and RBTCC-8 carcinoma

cells were from a s.c. metastasis of a FANFT tumor of StageD2. Invasive and metastatic carcinoma cells were differentiatedfrom their noninvasive counterparts by cellular and nuclearpleomorphism, cell size, nuclearcytoplasmic ratio, number ofnucleoli, and abnormalities of occludens junctions. Using low(less than 10) and high (greater than 80) passages of thesecell strains, tumorigenicity experiments in syngeneic ratsshowed that the normal in vivo progression of FANFT tumorswas interrupted by the isolation of carcinoma cells to cellculture. Histological appearance and biological behavior oftumor isografts closely resembled those of the original FANFTtumors. This was best demonstrated when tumor cells wereinoculated adjacent to rat femurs. The destruction of bone,monitored radiographically and histologically, served as ameasure of the invasive potential of the tumor cells. Destructionand deep invasion were observed only with ¡sograftsof invasiveand metastatic carcinoma cells, presumably due to collagen-

olytic activity. Despite rapid degradation of bone by theseisografts, the natural resistance of cartilage to tumor invasioncould not be overcome.

These carcinoma cell lines, together with their normal epithelial counterparts and the major supporting cells of connective tissue characterized previously by our laboratory, providea unique system to study tumor invasion.

INTRODUCTION

The structural and functional integrity of normal organ systems is thought to be maintained, at least in part, by close-

range Interactions between epithelium and mesenchyme (8,37). These interactions are altered when epithelia undergomalignant transformation and when carcinoma cells becomeinvasive. Connective tissue components of the epithelial-

stromal junction such as basal lamina are degraded by tumorenzymes, and carcinoma cells then infiltrate the areas of destruction, eventually penetrating blood and/or lymph vessels(1, 11, 12, 16, 20, 21, 39, 41 ). The complexity of cell-to-celland cell-to-stroma interactions, which accompany these events

' This work was supported by NIH Grant CA-25034 and in part by Grant CA-

21566, both from the National Cancer Institute.2 To whom requests for reprints should be addressed, at Department of

Pathology, Rush-Presbyterian-St. Luke's Medical Center, 1753 West Congress

Parkway, Chicago, III. 60612.Received May 15, 1980; accepted August 27, 1980.

in spontaneous and experimental tumors, underscores theneed for the isolation of intimately interacting tumor and stromalcells from the host (15-17). In isolated cultures, host regulatory

systems such as hormone and immune systems can be eliminated, and interactions between selected cells and tissues cantherefore be studied (17). However, central to the understanding of tumor invasion are the physiological interactions whichunderlie epithelial mesenchymal integrity in normal organ systems. Our first report therefore focused on the isolation and invitro characterization of epithelial cells, endothelial cells, andfibroblasts as the normal constituents of the epithelial-stromal

junction of rat urinary bladders (29). This discussion of our invitro cell system deals with the culture characteristics of severaltumor cell strains derived from FANFT3-induced carcinomas in

the Fischer rat urinary bladder.The FANFT-induced tumor model is of particular relevance

to our studies. The compound FANFT induces carcinomas ofthe urinary bladder in nearly 100% of the exposed rats (6, 40).The tumors progress through the stages of epithelial hyperpla-sia and dysplasia to noninvasive, invasive, and metastatic-

transitional cell carcinomas (6, 35, 40). This multistep processof FANFT chemical carcinogenesis in rat urinary bladder involves the occurrence of sequential changes in the precursorcells (9, 25, 31). These intrinsic changes in tumor cells arebelieved to determine, at least in part, the biological behaviorof neoplasms (10, 23). If intrinsic changes are essential for theacquisition of the properties of invasiveness and metastasis,then it is reasonable to anticipate that such changes might firstappear when tumors pass from a noninvasive to an invasiveand metastatic stage of tumor development. Thus, the purposeof this study is 2-fold: (a) to establish in culture carcinoma celllines derived from noninvasive, invasive, and metastaticFANFT-induced tumors; and (b) to compare these cell lines onthe basis of culture characteristics, morphology, and in vivogrowth behavior. Based on our previous studies that tumorcells rapidly invade bone but not hyaline cartilage (15-17), the

in vivo growth behavior of these cell strains was analyzed byinoculating tumor cells adjacent to the femur or the rib cartilageof syngeneic rats.

MATERIALS AND METHODS

Materials. RPMI 1640 was purchased from Grand IslandBiological Co., Grand Island, N. Y. FBS was from ReheisChemical Co., Kankakee, III. 4-(2-Hydroxyethyl)-1-piperazine-ethanesulfonic acid was from Calbiochem, La Jolla, Calif. Gen-tamycin was from Schering Corp., Kenilworth, N. Y. Amphoter-

icin B was from E. R. Squibb & Sons, New York, N. Y. Trypsin

3 The abbreviations used are FANFT, N-[4-<5-nitro-2-furyl)-2-thiazolyl]form-

amide; RPMI 1640, Roswell Park Memorial Institute Medium 1640; FBS, fetalbovine serum.

DECEMBER 1980 4571

on March 2, 2021. © 1980 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

B. U. Pauli et al.

and Bacto-agar were from Difco Laboratories, Detroit, Mich.FANFT was from Saber Laboratories, Inc., Morton Grove, III.Falcon Petri dishes (35 mm and 60 mm) and Falcon flasks (25sq cm) were from Falcon Plastics, Oxnard, Calif. Multiplate "-8

well containers and Thermanox coverslips (24 x 30 mm) werefrom Lux Scientific Corp., Thousand Oaks, Calif. Discs (25-mm) prepared from Nitex nylon cloth (HC-103) were from

TETKO, Inc., Elmsford, N. Y.Growth Medium. RPMI 1640 (growth medium) contained

10% FBS, 25 rriM 4-(2-hydroxyethyl)-1-piperazineethanesul-

fonic acid 50 fig gentamycin per ml, and 5 /ig amphotericin Bper ml. The medium was made from commercially suppliedpowder and was passed through a Millipore filter (0.22-fim

pore size) for sterilization. Complete medium osmolality was304 mosM at pH 7.2 to 7.4.

FANFT-induced Tumors. FANFT-induced tumors were obtained from Dr. S. M. Cohen, St. Vincent's Hospital, Worcester,

Mass. The moderately well-differentiated transitional cell car

cinomas were produced by feeding to weanling male Fischer344 rats (Charles River Breeding Laboratories, Inc., Boston,Mass.) the carcinogen FANFT at a dose of 0.2% of their dietfor 26 weeks (6, 35). Noninvasive tumors (Stage 0) (24) developed by 26 weeks, and by 63 weeks, tumors were invasive(Stages A to C) (24) or metastatic (Stage D) (24). Many aspectsof the development and the morphology of FANFT-induced rat

urinary bladder carcinoma have been described in detail elsewhere (6, 13, 30, 31, 35). For this study, a noninvasive FANFTtumor of Stage 0, an invasive FANFT tumor of Stage B2, anda s.c. metastasis (neck) of a FANFT tumor of Stage D2 wereprepared for cell culture.

Cell Culture Techniques. Noninvasive, invasive, or metastatic FANFT-induced tumors were removed from anesthetizedanimals, washed in culture medium, and hemidissected; onehalf was immediately processed for cell culture, and the otherhalf was for routine histopathological examination.

Cell cultures from the solid portions of FANFT-induced tu

mors were prepared by mincing the neoplastic tissue intofragments of approximately 1-mm diameter. The fine fragments

were placed directly into tissue culture flasks and barely covered with growth medium to ensure their attachment to theplastic surface. Cultures were incubated at 37°in a humidified

5% CO2:air atmosphere. After outgrowth began, an additional3 ml of the growth medium was added to the flasks. The growthmedium renewed every 3 to 4 days until the monolayers became subconfluent. The cells were then removed from thegrowth surface with 0.25% trypsin and 0.02% EDTA in Dul-becco's phosphate-buffered saline (pH 7.3; 15 min; 37°)and

split at a ratio of 1:3 for continuous cultures. Cells of variouspassages were frozen and stored in liquid nitrogen as described by Coriell (7). Growth curve, doubling time, and colony-

forming efficiency of each frozen cell passage were determinedaccording to standard procedures (14).

Primary cultures contaminated with fibroblasts were subjected to a differential trypsinization according to the methodof Owens ef al. (28). After elimination of fibroblastic contaminants, tumor cells were frozen and stored as described above.

Growth in Soft Agar. The ability of cultured cells to growand produce colonies in soft agar medium has been studied bythe method of MacPherson (22). In brief, 7 ml of warm RPMI1640 (culture medium) (44°)containing 10% FBS, 10% tryp-

tose phosphate broth, and 0.5% Bacto-agar (Difco) were pi-

peted into 60-mm Falcon plastic dishes and left to set at roomtemperature. This agar base layer was then topped with 1.5 mlof a mixture of one volume of tumor cell suspension in RPMI1640 at 37° (2 x 10s cells) and 2 volumes of 0.5% agarmedium at 44°. The cultures were incubated in a humidified

5% CO2:air atmosphere. Cells were refed by adding 1 to 2 mlculture medium on the agar top layer every seventh day.Unstained cultures were examined for the presence of cellcolonies at intervals of 7 days using a Zeiss ICM 405 invertedmicroscope equipped with graded eyepieces. Colonies werecounted and their diameters measured.

Light and Electron Microscopy. Cultured cells to be examined by light and electron microscopic techniques were sub-cultured on either 24- x 30-mm Thermanox plastic coverslipsor collagen-coated Nitex discs as described previously (4, 5,29). The procedures for both light and electron microscopypreparations (thin section, freeze fracture) have been described in detail elsewhere (32, 36).

Tumorigenicity. The tumorigenicity of cultured cells wastested by s.c. or i.p. inoculation of 1 x 105 tumor cells into 10

syngeneic Fischer 344 rats aged 6 to 9 days. From each cellline, a low passage number (< 10) and a high passage number(> 80) were tested. Inoculated animals were examined biweekly for tumor growth during a period of 3 to 6 months. Acomplete autopsy was performed on rats with advanced neoplastic disease. Tissue samples were examined by routine lightand thin-section electron microscopy (35).

Periosseus and Perichondrial Growth. The periosseus andperichondrial growth of tumor cells was tested by injecting 1x 105 tumor cells (low and high passages) adjacent to the

metaphysis of the femur or the rib cartilage of 10 syngeneicFischer 344 rats aged 9 to 10 days. Inoculated rats wereexamined for tumor growth as described above. Tumor iso-grafts, measuring 2 to 3 cm in diameter, were X-rayed and then

excised together with the femur or the ribs. Tissue samplescontaining the tumor-bone or tumor-cartilage junction wereprocessed for routine light- and electron-microscopic examination.

Primary normal epithelial cells (1 x 106) prepared from

enzymatically released epithelial sheets of Fischer 344 raturinary bladders (29) were also inoculated adjacent to boneand cartilage of 10 syngeneic rats aged 9 to 10 days. Theirgrowth behavior was compared with that of tumor cells.

RESULTS

Cell Lines

Three continuous cell lines designated RBTCC-2, RBTCC-5,and RBTCC-8 have been established from FANFT-induced

tumors of the rat urinary bladder. These 3 cell lines weredifferentiated by the histopathologically defined stages of neoplastic transformation of their original tumors and by theirbiological behavior in the syngeneic host (Table 1). RBTCC-2

cells were derived from a noninvasive FANFT tumor of Stage0. These tumor cells formed multiple, low-grade transitional

cell carcinomas with expansive growth at the inoculation siteof the syngeneic host. RBTCC-5 cells were developed from an

invasive FANFT tumor of Stage B2. Inoculated into syngeneichosts, RBTCC-5 cells formed transitional cell carcinomas ofhistopathological Grade 2 that grew slowly infiltrative. RBTCC-

4572 CANCER RESEARCH VOL. 40

on March 2, 2021. © 1980 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Tumor Model for Invasion in Vitro and in Vivo

Table 1Comparisons between original tumors and tumor isografts of carcinoma cell lines RBTCC-2. fíBTCC-5. and

RBTCC-8

RBTCC-2Tumor

locationAsymmetricunitmembranePleomorphic

microvilliTonofilamentsNuclear

pleomorphismSquamousdifferentiationOccludens

junctionsNormalHypertrophieAttenuatedDiscontinuousGap

junctions (PF-1)DesmosomesHemidesmosomesBasal

laminaTumorigenicityInvasionMetastasisOriginal

NRBatumor-"

B+

+ + ++

++++++

++

+ +c +++

+ + ++±±

+±±+++

+ + + ++++ ++++CON

CON--NoNoTumor

xenograftDMP-+

++++++

+++++

+++++±++

+++-t-+CON8/10NoNoRBTCC-5Tumor

Original xeno-tumorgraftB

DMHIP_+

+ ++++++ + +++/++

++++ + +++

+ +++++/+++

+ ++++++

++++

+ ++++ ++FRA

FRA10/10Yes

YesNoNoRBTCC-8Original

Tumor xen-tumorograftS

DMHIPLS-+++

++++++++

++/++++/++++++

+/+++/+++/+++

+ ++++ ++++ +++

+++ +++

±——10/10Yes

YesYesYes (5/10)

NRB, normal rat bladder: B, bladder; D, diaphragm; M, mesenterium; P. peritoneum; H, liver; I, intestine; S,subcutis; L, lung; CON, continuous; FRA, fragmented.

b —,missing; ±.variable; + to + + +. few (mild) to many (severe).c Occludens junctions occur between apical lateral membranes of all superfical cells (+ + +).d Approximately 6 desmosomes/100-/im cell perimeter (+ + +) (30).

8 carcinoma cells were derived from a s.c. metastasis (neck)of an invasive FANFT tumor of Stage D2. At autopsy, thisanimal also had métastasesto the lungs, liver, and peritoneum,which histopathologically resembled that found in the bladderand the subcutis. The tumor isografts of RBTCC-8 were tran

sitional cell carcinomas of Grade 2 to 3 and rapidly infiltratedhost tissues, causing métastasesin the lungs of 5 of 10 Fischerrats. The métastaseswere randomly distributed in both halvesof the lungs and consisted of 5 to 23 tumor nodules ranging 1to 3 mm in diameter. Tumor nodules in abdominal organs, e.g.,the liver and the kidneys (Table 1), were not considered as truemétastases, since they originated presumably from implantations of i.p. injected tumor cells. There was no difference in thehistological appearance and the biological behavior of tumorisografts induced with low or high passages of carcinoma cells.Morphology and growth of the tumor isografts of RBTCC-2,RBTCC-5, and RBTCC-8 carcinoma cells were compared with

their original tumors in Table 1.

In Vitro Growth Characteristics

In vitro growth characteristics of the 3 carcinoma cell strainswere summarized in Table 2. RBTCC-2, RBTCC-5, andRBTCC-8 carcinoma cells showed similar growth patterns onplastic surfaces and on collagen matrices. On plastic growthsurfaces, the doubling times were 19, 25, and 17 hr, and thecolony-forming efficiencies were 17.22, 18.50, and 12.40%.In soft agar, the 3 carcinoma cell lines were differentiated bythe size of their colonies. The maximal colony diameters were160 ¿imfor RBTCC-2, 320 /xm for RBTCC-5, and 170 p.m forRBTCC-8 carcinoma cells. Maximal colony sizes were reachedafter an incubation period of 14 days for RBTCC-2 and RBTCC-8 and 28 days for RBTCC-5 tumor cells. Although tumor cells

were kept viable for several additional weeks under optimal

culture conditions, after these time periods, colonies remainedconstant in size (Table 2).

Morphology of Cell Cultures

Light Microscopic Findings. On plastic surfaces, all 3 carcinoma cell lines grew contact uninhibited and focally piled upto 3 and 4 cell layers (Fig. 1). The monolayers were composedof polygonal cells which varied in size and shape. The averagesurface area occupied by one carcinoma cell increased from291 sq firn for the noninvasive cell strain RBTCC-2 to 422 sqfirn for the metastatic cell strain RBTCC-8. The nuclei were

spherical or ellipsoid in shape and hyperchromatic and contained several prominent nucleoli. Multinucleated giant cellswere occasionally observed (Fig. 2).

On collagen-coated Nitex discs, tumor cells grew in a mannersimilar to that seen in the animal host. Areas of papillary-liketumor cell clusters alternated with areas of single cell layers(Fig. 3). RBTCC-2 tumor cells differentiated into keratinocytesas evidenced by their ability to form keratohyaline pearls.RBTCC-8 cells occasionally formed cell clusters within the

collagen coat.Electron Microscopic Findings. The surface membrane of

the tumor cells consisted of a 9-nm-thick symmetric unit membrane. Regions of 11-nm-thick asymmetric unit membrane,typical for normal rat bladder epithelium in vivo (26, 35), werenot observed. The cell membranes were covered by a sparseglycocalyx. There were numerous pleomorphic microvilli at thefree surface (Fig. 4). The tumor cell cytoplasm had a relativelysimple cytoarchitecture and was characterized by the presenceof a few clustered mitochondria and numerous free ribosomes.Smooth and rough endoplasmic reticulum was poorly developed. The Golgi apparatus was usually prominent and oftenmulticentric (Fig. 5). Secondary lysosomes were extremelyplentiful in some cells and missing in others. Intracytoplasmic

DECEMBER 1980 4573

on March 2, 2021. © 1980 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

B. U. Pau// ef al.

MSg181CM

51

S(0cj1ffl2"§o0,1'5.çäSs

j:g(93n

matrixIue0COm0)o(0--;sIQ.Co1oO•-

_ e1 à"a?aliili2

E o .2 - -oo-oSHIM"--

wfe•••'euOo-oz|oo*~f

exliso

*Eo_2"n\

9°

oïz»£

EiooeO

Q.'¿aC

^ffloZf31"(0,

^ÈgO0)>.

-^£oco "^

'Soí0

C,aC*-

£•5,

'cinöS

&C

».3o

Oo0

=0

Xc73•

3>Cfeiil

jCI)CO(073a•-=

ozÉlS«1o00a.COCO"co

O-C

Cì

«»-raO

a.2Q•CM•Hs0Z-.j•LILS0

O+100ooq!•+1CM

CMNO»CM

1'Ooo^j.qCMCM

O0+1)•^0_

fjCO3

CO.2E±±COCO3CO

CACD

OJ•H,_S•aa1eC0sXI

oOCM8mEcoCNO

CMCOco•Hco(M0Ztoia.u.2SOco0CDCM+1O

IOcdIOCMCM|

-Cio.coooCOqCMCMo0•H0)cod_

tflCO3

C0.2E.•^COco

tn1—ai•Hcoco

co73>,

£It

coCO

COra

OoE—IO0

UH03ozÕoIO+1CMto1a.ILSo60.28

±,_O3*-•HOCM1^*-CM^jC9i0.IOo+1CMCMSó

-HCM^01509C

COK0)•HCMCM•Õ5iÌ9cQ<DZpÜCQ0

Uh-EocsT.2S^.

oQ0W|•H

E«u."IÉ«

«

lumina were observed occasionally (Fig. 6) (2). Squamousdifferentiated cells found in all 3 tumor cell strains had prominent bundles of tonofilaments. The nuclei varied in size andchromatin distribution. The nuclear envelope typically hadmany shallow infoldings. Similar cytological observations weremade when tumor cells were grown on collagen-coated Nitex

discs (Fig. 6). Basal tumor cells were lacking a basal laminaand grew in direct contact with the collagenous growth surface.There were no hemidesmosomes anchoring the basal plasmamembrane to collagenous fibers.

Intercellular junctions of noninvasive and invasive transitionalcarcinoma cell lines have been described in previous publications from our laboratory (32, 33). In brief, noninvasive(RBTCC-2) and invasive (RBTCC-5 and RBTCC-8) carcinoma

cells have zonulae occludentes at the apices of their lateralmembranes (Figs. 4 and 7). Some occludens junctions ofRBTCC-5 and especially RBTCC-8 carcinoma cells were fo-

cally attenuated, consisting of only one or 2 strands. Occasionally, the strands were discontinuous, thus forming fasciaeor maculae occludentes. Expanded occludens junctions wereobserved in cultures of RBTCC-8 tumor cells (Fig. 8). A fewgap junctions, type PF-1, as well as numerous desmosomeswere present in all 3 carcinoma cell lines.

Periosseus and Pericartilagenous Growth of Tumor Cells

Primary Normal Epithelial Cells. Normal rat bladder epithelial cells were unable to grow at their periosseus and perichon-drial injection sites. Four months after injection, neither scartissue as a reaction to the inoculum nor abnormal developmentof bone or cartilage was observed.

RBTCC-2 Carcinoma Cells. Periosseus and pericartilagen-ous injection of RBTCC-2 carcinoma cells caused the formation

of transitional cell carcinomas after 4 to 6 weeks measuring 1to 3 cm in diameter. The tumors were noninvasive and showedmoderate squamous differentiation. They were encapsulatedby a complete basal lamina and dense fibrous connectivetissue. Periosteum and perichondrium as well as bone andcartilage were free of tumor and appeared normal by radio-

graphical and histological examination (Fig. 9).RBTCC-5 Carcinoma Cells. RBTCC-5 carcinoma cells

formed palpable periosseus and pericartilagenous tumors 8weeks postinoculation. Tumors were moderately well differentiated, and there were focal areas of squamous differentiationwith pearl formation. Many tumor cells broke through an incomplete fragmented basal lamina into surrounding connectivetissues. The metaphyseal compact bone of the femur waspenetrated and destroyed in several places by invasive bladdercarcinoma cells (Figs. 10, 12, 13). Tumor nodules formedwithin the marrow cavity completely digested the bone spicules.At the epiphyseal growth plate, the tumor extended as far intothe hyaline cartilage as the vascular loops, namely to the areaof the last hypertrophie chondrocyte and its calcified matrix(Fig. 12). The noncalcified cartilage was not penetrated bytumor cells.

RBTCC-5 carcinoma cells were also unable to invade the

cartilagenous matrix of the ribs. Tumor cells destroyed andpenetrated the dense fibrous perichondrial connective tissueonly at the periphery where numerous blood vessels normallywere present. Tumor cells and perichondrium were generallyseparated by a complete basal lamina. This was in contrast to

4574 CANCER RESEARCH VOL. 40

on March 2, 2021. © 1980 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Tumor Model for Invasion in Vitro and in Vivo

the fragmented basal lamina that surrounded the periosseusand i.p. tumor nodules.

Mitotic figures were common in all tumors but were rare intumor areas adjacent to the perichondrium.

RBTCC-8 Carcinoma Cells. RBTCC-8 carcinoma cellsformed palpable tumors 3 to 4 weeks postinoculation and killedthe animals within 6 to 10 weeks. Tumor cells rapidly invadedadjacent connective tissues such as muscle and bone. Thefemoral bone showed both vast areas of osteolysis and osteo-genesis (Fig. 11). The metaphyseal compact bone was twice asthick as that of the contralateral femur and was penetrated byislands of bladder carcinoma cells (Fig. 14). At the periphery,newly formed bone extended radially from the thickened compact bone. Clefts between bony spicules housed clusters ofcarcinoma cells. The spicule surface was either in direct contact with tumor cells or was lined by one or 2 layers ofosteoblasts. Where tumor cells confront the bone, lacunarinvaginations were observed. The marrow cavity was replacedby bulky tumor masses and newly formed bone (Fig. 15). Therewere few leukopoietic and erythropoietic cells remaining.

RBTCC-8 tumors which grew adjacent to rib cartilageshowed a behavior similar to that of the RBTCC-5 carcinomas.

The dense pericartilagenous connective tissue and the matrixof the hyaline cartilage resisted invasion by RBTCC-8 carci

noma cells (Fig. 16). Tumors were separated from the perichondrium by an incomplete basal lamina.

DISCUSSION

The in vitro FANFT tumor model described in this reportconsists of 3 continuous carcinoma cell lines. These 3 celllines were obtained from FANFT-induced carcinomas of the

Fischer rat urinary bladder by standard expiant techniques(32). RBTCC-2 carcinoma cells were derived from a noninva-sive FANFT tumor of Stage 0, RBTCC-5 carcinoma cells werefrom an invasive FANFT tumor of Stage B2, and RBTCC-8

carcinoma cells were from a s.c. metastasis of a FANFT tumorof Stage D2. In monolayer cultures, all 3 carcinoma cell linesdisplayed occludens junctions, desmosomes, pleomorphic mi-crovilli, and bundles of tonofilaments (19, 32). Invasive andmetastatic carcinoma cells were differentiated from their non-

invasive counterparts by cellular and nuclear pleomorphism,cell size, nuclearcytoplasmic ratio, number of nucleoli, andabnormalities of occludens junctions, which were associatedwith increased tumor cell surface protease activities (34). Tu-morigenicity experiments with these cell strains confirmed thateach line was a true representative of its original tumor. Thebiological behavior and the histological appearance of the 3tumor isografts closely resembled those of the original FANFT-induced tumors. This behavior in syngeneic hosts was unchanged whether ttie tumor-cell inoculum was selected from alow-passage number (<10) or a high-passage number (>80).

We interpret this to mean that the in vivo progression of FANFTtumors could be interrupted when carcinoma cells were isolated in culture.

Colony formation of these cell lines in soft agar coincidedfavorably with tumorigenicity experiments performed in s.c.sites (27, 38). However, these techniques failed to demonstratedifferences between noninvasive and invasive cell lines. Similardifficulties were encountered with i.p. tumor grafts, althoughthe histopathological evaluation at autopsy was more informa

tive than that of the s.c. counterpart. Using our tumor cell lines,we have introduced a novel method for the evaluation of tumorinvasiveness. This method relies on the degradation of developing bone by adjacent tumor masses in vivo. The degree ofdestruction of rat femoral bone was monitored by radiograph-

ical and histological techniques and served as a measure ofthe invasive potential of the tumor cell lines. Destruction anddeep invasion were observed only with isografts of the invasiveRBTCC-5 and the metastatic RBTCC-8 carcinoma cell line,while the noninvasive RBTCC-2 carcinomas cell line formed

expansively growing tumors with no effect on the normal development of femoral bones. Presumably, tumor-induced os

teolysis in this system was associated with collagenolytic activity (15-17). These findings suggest that our tumor cell lines

may elaborate different levels of collagenolytic activity relativeto the degree of invasiveness (1). This hypothesis was supported by observations correlating depth of invasion (stage) ofhuman urinary bladder transitional cell carcinomas with collagenolytic activity (41). In Stage A human transitional cell carcinomas, which corresponded to the noninvasive isografts ofRBTCC-2, collagenolytic activity was not detected. Stage C

and D human bladder tumors, which corresponded to theinvasive isografts of RBTCC-5 and RBTCC-8, exhibited high

levels of collagenolytic activity. Wirl and Frick (41) concludedthat the observed collagenolytic activities in human bladdertumors might be products of the host in response to the tumorand not products of the tumor itself (3). Although our presentstudies were not designed to evaluate this hypothesis, preliminary studies in vitro show that the invasive rat bladder tumorcell lines were capable of degrading devitalized bone matrix.4

In these experiments with rat femoral bones, we also observed that rat bladder carcinoma cells were unable to penetrate the hyaline cartilage of the growth plate. In previousstudies, we have documented that the natural resistance ofcartilage to tumor invasion may at least in part be due to aspectrum of protease inhibitors which are endogenous components of cartilage matrix (15, 16, 18). These protease inhibitors may be capable of inhibiting basal lamina-degrading tumor

collagenase. The collagenase activity has been correlated withthe invasive and metastatic potential of various tumors (21).This hypothesis was supported by our morphological observations that the failure of invasive rat bladder tumors to penetrate the hyaline cartilage of the ribs was associated with theformation of a complete basal lamina, in contrast to basallamina fragmentation seen in association with bone degradation.

In this report, we have established an in vitro tumor cellsystem comprising the spectrum of noninvasive, invasive, andmetastatic transitional cell carcinomas of the rat urinary bladder. Defined cell lines have been used to characterize a sensitive method which quantifies the degree of invasion based onbone destruction. We have also shown that cell lines, despitetheir invasive potential, were unable to invade cartilage. Theseneoplastic cell lines, together with the cells derived from theirnormal epithelial counterparts, and the major supporting cellsof the connective tissue stroma (endothelial cells and fibro-

blasts) (29) form a unique in vitro system for the study of tumorinvasion and its local regulation.

' B. U. Pauli, J. E. Morton and K. E. Kuettner. manuscript in preparation.

DECEMBER 1980 4575

on March 2, 2021. © 1980 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

8. U. Pau// ef al.

ACKNOWLEDGMENTS

We take pleasure in thanking Dr. R. S. Weinstein for his interest, encouragement, and helpful discussion and Dr. G. H. Frieden, Dr. S. M. Cohen, and Dr. J.B. Jacobs for providing us with FANFT tumors. We gratefully acknowledge theexpert technical assistance of Shu-Yuan Chi, Jayne Adams, and William Leonard,as well as Sandra Velasco for preparing the manuscript.

REFERENCES

1. Abramson, M., Schilling, R. W.. Huang, C., and Salome, R. G. Collagenaseactivity in epidermoid carcinoma of the oral cavity and larynx. Ann. Otol.Rhinol. Laryngol., 84: 158-163, 1975.

2. Alroy, J., Pauli, B. U.. Hayden, J. E.. and Gould, V. E. Intracytoplasmiclumina in bladder carcinomas. Hum. Pathol., 10: 549-555, 1979.

3. Bauer, E. A., Gordon, J. M., Reddick, M. E., and Eisen, A. Z. Quantitationand immunocytochemical localization of human skin collagenase in basalcell carcinoma. J. Invest. Dermatol., 69. 363-367, 1977.

4. Cerejido. M., Robbins, E. S., Dolan, W. J., Rotunno, C. A., and Sabatini, D.A. Polarized monolayers formed by epithelial cells on a permeable andtranslucent support. J. Cell Biol., 77. 853-880, 1978.

5. Chlapowski, F. J., and Haynes. L. The growth and differentiation of homogenous transitional epithelium in vitro. J. Cell Biol., 83. 605-614, 1977.

6. Cohen, S. M., Jacobs, J. B., Arai. M., Johannsson, S., and Frieden, G. H.Early lesions in experimental bladder cancer: experimental design and lightmicroscopic findings. Cancer Res., 36: 2508-2511, 1976.

7. Coriell, L. L. Preservation, storage, and shipment. Methods Enzymol., 58.29-36, 1979.

8. Croissant, R., Guenther, H., and Slavkin, H. C. How are embryonic prea-meloblasts instructed by odontoblasts to synthesize enamel? In: H. C.Slavkin and R. C. Greulich (eds.), Extracellular Matrix Influences on GeneExpression, pp. 515-521. New York: Academic Press, Inc., 1975.

9. Farber, E. Carcinogenesis. Cellular evolution as a unifying thread: presidential address. Cancer Res., 33. 2537-2550. 1973.

10. Fidler. I. J. Selection of successive tumour lines for metastasis. Nat. NewBiol., 242. 148-149, 1973.

11. Fidler, l. J., Gersten, D. M., and Hart, I. R. The biology of cancer invasionand metastasis. Adv. Cancer Res., 28. 149-250, 1978.

12. Hashimoto, K., Yamanishi, Y., Maeyens, E., Dabbous. M. K., and Kanzaki,T. Collagenolytic activities of squamous cell carcinoma of the skin. CancerRes., 33. 2790-2801, 1973.

13. Jacobs, J. B., Arai, M., Cohen, S. M., and Friedeil, G. H. Early lesions inexperimental bladder cancer: scanning electron microscopy of cell surfacemarkers. Cancer Res., 36. 2512-2517, 1976.

14. Kruse, R. F., Jr., and Patterson, M. K., Jr. (eds.). Tissue Culture: Methodsand Applications. New York: Academic Press, Inc., 1973.

15. Kuettner, K. E., and Pauli, B. U. Resistance of cartilage to normal andneoplastic invasion. In: J. E. Morton, T. M. Tarpley. Jr., and W. E. Davis(eds.), Mechanisms of Localized Bone Loss. Supplement to Calcified TissueAbstracts, pp. 251-278. Arlington, Va.: Information Retrieval, Inc., 1978.

16. Kuettner, K. E., and Pauli, B. U. Resistance of cartilage to invasion. In: H. A.Gilbert, L. Weiss, and D. C. G. Monsen (eds.), Bone Metastasis. Boston: G.K. Hall & Co., in press, 1980.

17. Kuettner, K. E., Pauli, B. U., and Soble, L. Morphological studies on theresistance of cartilage to invasion by osteosarcoma cells in vitro and in vivo.Cancer Res., 38. 277-287, 1978.

18. Kuettner, K. E., Soble, L., Croxen, R. L., Marxzynska, B., Hiti, J., andHarper, E. Tumor cell collagenase and its inhibition by a cartilage derivedprotease inhibitor. Science (Wash. D. C.), Õ96:653-654, 1977.

19. Lavin. P.. and Koss, L. G. Studies of experimental bladder carcinoma inFischer 344 female rats. II. Characterization of 3 cell lines derived frominduced urinary bladder carcinomas. J. Nati. Cancer. Inst., 46: 597-614,

1971.20. 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. Sci. U. S. A., 76. 2268-2272, 1979.

21. Liotta, L. A.. Tryggvason, K., Garbisa, S., Hart, I., Foltz, C. H., and Shafie,S. Metastatic potential correlates with enzymatic degradation of basementmembrane collagen. Nature (Lond.). 284. 67-68, 1980.

22. MacPherson, I. Soft agar technique. In: P. F. Kruse, Jr., and H. K. Patterson,Jr. (eds.), Tissue Culture: Methods and Applications, pp. 276-280. NewYork: Academic Press, Inc., 1973.

23. Malech, H. L., and Lentz, T. L. Microfilaments in epidermal cancer cells. J.Cell Biol., 60: 473-482, 1974.

24. Marshall, V. F. The relation of the preoperative estimate to the pathologicdemonstration of the extent of vesical neoplasms. J. Urol., 68: 714-723,

1952.25. Medina, D. Tumor progression. In: F. B. Becker (ed.), Cancer—Comprehen

sive Treatise, Vol. 3, pp. 99-119. New York: Plenum Publishing Corp.,1975.

26. Merk, F. B., PauM, B. U., Jacobs, J. B., Alroy, J., Fiedell, G. H., andWeinstein, R. S. Malignant transformation of urinary bladder in humans andin fVX4-(5-nitro-2-furyl)-2-thiazolyl) formamide-exposed Fischer rats: ultra-

structure of the major components of the permeability barrier. Cancer Res.,37. 2843-2853, 1977.

27. Montesano, R., Drevon, C., Kuroki, T., Saint Vincent. L., Handleman, S.,Sandford. K. K.. DeFeo, D., and Weinstein. I. B. Test for malignant transformation of rat liver cells in culture: cytology, growth in soft agar, andproduction of plasminogen activator. J. Nati. Cancer Inst., 59: 1651-1658,

1977.28. Owens, R. B., Smith, H. S., and Hackett, A. J. Epithelial cell cultures from

normal glandular tissue of mice. J. Nati. Cancer Inst., 53: 261-269, 1974.29. Pauli, B. U., Anderson, J. N., Memoli, V. A., and Kuettner, K. E. The isolation

and characterization in vitro of normal epithelial cells, endothelial cells, andfibroblasts from rat urinary bladder. Tissue Cell, 12: 419-435. 1980.

30. Pauli, B. U., Cohen, S. M., Alroy, J., and Weinstein, R. S. Desmosomeultrastructure and the biological behavior of chemical carcinogen-inducedurinary bladder carcinomas. Cancer Res., 38: 3276-3285, 1978.

31. Pauli, B. U., Frieden, G. H., and Weinstein, R. S. Topography and numericaldensities of ¡ntramembrane particles in chemical carcinogen-induced urinarybladder carcinomas in Fischer rats. Lab. Invest., 39: 565-573, 1978.

32. Pauli, B. U., Kuettner, K. E., and Weinstein, R. S. Intercellular junctions inFANFT-induced carcinomas of rat urinary bladder in tissue culture, in situthin-section, freeze-fracture, and scanning electron microscopy studies. J.Microsc. (Oxf.), 115: 271-282, 1979.

33. Pauli. B., and Weinstein, R. S. Intercellular junctions in FANFT-inducedtransitional cell carcinomas in tissue culture, in situ freeze-fracture studies.Proc. Electron Microsc. Soc. Am., 35: 594-595, 1977.

34. Pauli. B. U.. and Weinstein, R. S. Abnormalities in zonulae occludentes ofrat bladder carcinomas in vitro. J. Cell Biol., in press. 1980.

35. Pauli, B. U., Weinstein, R. S., Alroy, J., and Arai, M. Ultrastructure of celljunctions in FANFT-induced urothelial tumors in urinary bladder of Fischerrats. Lab. Invest., 37: 509-621, 1977.

36. Pauli. B. U.. Weinstein, R. S., Sobel. L. W., and Alroy. J. Freeze-fracture ofmonolayer cultures. J. Cell. Biol., 72: 763-769, 1977.

37. Propper, A. Y. Epithelial-mesenchymal interaction between developing chickepidermis and rabbit embryo mammary mesenchyme. In: H. C. Slavkin andR. S. Greulich (eds.), Extracellular Matrix Influences on Gene Expression,pp... 541-547. New York: Academic Press, Inc., 1975.

38. San, R. H. C., Caspia, M. E., Soiefer, A. L., Maslansky, C. J., Rice, J. M.,and Williams, G. M. A survey of growth in soft agar and cell surfaceproperties as markers for transformation in adult rat liver epithelial like cellcultures. Cancer Res., 39: 1026-1034, 1979.

39. Strauch, L. The role of collagenases in tumor invasion. In: D. Tarin (ed.),Tissue Interactions in Carcinogenesis, pp. 399-433. New York: AcademicPress, Inc.. 1972.

40. Tiltman. A. J.. and Frieden. G. H. The histiogenesis of experimental bladdercancer. Invest Urol., 9. 218-226, 1971.

41. Wirl. G., and Frick. J. Collagenase—a marker enzyme in human bladdercancer? Urol. Res., 7: 103-108, 1979.

4576 CANCER RESEARCH VOL. 40

on March 2, 2021. © 1980 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Tumor Model for Invasion in Vitro and in Vivo

Fig. 1. RBTCC-5 carcinoma cells grow contact uninhibited and focally pile up to 3 and 4 cell layers. Phase contrast, x 260.Fig. 2. RBTCC-5 carcinoma cells are incubated for 7 days. The polyhedral tumor cells grow a cobblestone-like monolayer. There are few multinucleated giant

cells. Phase contrast, x 280.Fig. 3. Papillary growth pattern of RBTCC-5 carcinoma cells on collagen-coated Nitex filters (N) The cells and nuclei are moderately pleomorphic. The nuclei are

slightly hyperchromatic and contain several prominent nucleoli. Toluidine blue, x 520.Fig. 4. A junctional complex [consisting of a zonula occludens (ZO), a zonula adherens (ZA), and a macula adherens (MA)} joins 2 RBTCC-2 carcinoma cells at

their apical lateral membranes. There are numerous pleomorphic microvilli covered with abundant glycocalyx (arrow) at the free surface, x 65.000.Fig. 5. The cytoplasm of RBTCC-8 carcinoma cells contains scattered mitochondria, a multicentric Golgi complex (G). few strands of smooth and rough

endoplasmic reticulum. and numerous free ribosomes. Clusters of myelin figures and a few bundles of tonofilaments (arrows) are present in some cells. Tumor cellsare joined by desmosomes (MA) and numerous interdigitating pleomorphic microvilli. x 11.700.

Fig. 6. RBTCC-8 carcinoma cells grow in multiple layers on collagen-coated Nitex filters. Lumina (arrows) within the multilayers are sealed by occludens junctions.

The cytoplasm contains clusters of secondary lysosomes (LY) and mitochondria. The Golgi complex is hypotrophic (G). The nuclei vary in size and shape, and thenuclear envelope displays several shallow infoldings. x 3,400.

Fig. 7. Zonula occludens consists of a network of intramembrane fibrils in the P-face (arrow) and corresponding grooves in the E-face (double arrow) of thefreeze-fractured plasma membrane of RBTCC-2 carcinoma cell, x 80.000.

Fig. 8. Greatly expanded occludens junction in the E-face membrane of freeze-fractured RBTCC-8 carcinoma cell, x 114,000.

Fig. 9. X-ray shows a noninvasive RBTCC-2 tumor (T) growing in juxtafemoral position of a syngeneic rat 10 weeks postinoculation of tumor cells. The femoralbone appears normal, x 2.3.

Fig. 10. X-ray shows destruction (arrow) and penetration (double arrow) of the femoral compacta by an invasive RBTCC-5 carcinoma 12 weeks postinoculationof tumor cells. The tumor has penetrated into the marrow cavity and grows in the proximal metaphysis. x 2.3.

Fig. 11. X-ray shows rat femur 10 weeks postinoculation of RBTCC-5 carcinoma cells. The tumor causes osteolysis (arrow) and extensive osteogenesis of thefemoral bone. Newly formed bone surrounds the entire metaphyseal compacta, x 2.3.

Fig. 12. Epiphyseal hyaline cartilage resists invasion by RBTCC-5 carcinoma cells, whereas bone is rapidly destroyed, x 160.

Fig. 13. Cytoplasmic processes of an RBTCC-5 carcinoma cell have penetrated the bony matrix of the femoral compacta. At the tumor-bone junction, there arefragments of basal lamina with attached hemidesmosomes (arrows). Tumor cells are joined by desmosomes (circle), x 9,900.

Fig. 14. Light micrograph of the femur shown in Fig. 11 displays extensive new bone formation in the marrow cavity, the metaphyseal compacta, and theperiosteum. There are clusters of RBTCC-8 carcinoma cells (asterisk) between spicules of woven bone, x 90,

Fig. 15. RBTCC-8 carcinoma cells have penetrated the metaphyseal compacta and grow in the marrow cavity adjacent to the epiphyseal growth plate (C). Originalbone spicules have been destroyed and are replaced by woven bone, x 140.

Fig. 16. KBTCC-8 carcinoma cells grow adjacent to the rib hyaline cartilage. Tumor cells are unable to penetrate the dense perichondrial connective tissue andthe cartilagenous matrix, x 140.

DECEMBER 1980 4577

on March 2, 2021. © 1980 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

B. U. Pau// ef al.

4578 CÕNCER RESEARCH VOL. 40

on March 2, 2021. © 1980 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Tumor Model for Invasion in Vitro and in Vivo

gpgyc. a*

n i^^fe!"* "ífflk •.

**

r©

•*/-

!-',%;K.Ate

-

DECEMBER 1980 4579

on March 2, 2021. © 1980 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

8. U. Pau// eíal.

'•

...^ •' •. epiphyseal O•

. "V / -.--

. .•-••>'••^s: * ••

o«X- •: ^;:^:;$¥;';i;^;?-,; >^, ^S f" /«•¿i^i¿í"/?»:s'w^,; -«

S¥:lfií"- íVJ? »-JÕu' " ' , '•'••' "' ' •. " •- t=*lB

,'•,', '' ' :-' .$* &_¿!% i!1' •''.'.'•'•'''.*•',' ''•'.'\-•O •^rO.'^fí';^«,'i.'' ;>.r;.•,-'^•''•"''^ürvT

4580 CANCER RESEARCH VOL. 40

on March 2, 2021. © 1980 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

1980;40:4571-4580. Cancer Res   Bendicht U. Pauli, Steven N. Anderson, Vincent A. Memoli, et al.   for the Study of Invasion

Epithelial Tumor Modelin Vivo and in VitroDevelopment of an

  Updated version

  http://cancerres.aacrjournals.org/content/40/12/4571

Access the most recent version of this article at:

   

   

   

  E-mail alerts related to this article or journal.Sign up to receive free email-alerts

  Subscriptions

Reprints and

  .pubs@aacr.orgDepartment at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

  Permissions

  Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/40/12/4571To request permission to re-use all or part of this article, use this link

on March 2, 2021. © 1980 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from