[CANCER RESEARCH 45, 3322-3331, July 1985]

Changes in Stem Cell Populations of Rat Trachea! Epithelial Cell Cultures at an

Early Stage in Neoplastic Progression

David Thomassen,1'2 J. Carl Barrett, Diane K. Beeman, and Paul Nettesheim2

Epithelial Carcinogenesis Group, Laboratory of Pulmonary Pathobio/ogy, National Institute of Environmental Health Sciences, Research Triangle Park, North Carolina 27709

ABSTRACT

The development of transformed colonies and concomitantchanges in proliferative and nonproliferative cell compartmentswere studied in rat trachéalepithelial (RTE) cell cultures followingexposure to W-methyl-A/'-nitro-A/-nitrosoguanidine (MNNG). Pri

mary RTE cells were plated onto 3T3 feeder layers and treatedwith MNNG (0.25 /¿g/ml)or solvent. Seven days later, the feedercells were removed to select for enhanced growth variants, whichare the transformants of the RTE cell system, usually scored 5weeks after carcinogen exposure. Most of the RTE cell colonies,which originally formed during the first 7 days of culture, disappeared within 2 weeks after feeder cell removal in control andMNNG-treated cultures. In control cultures, about 3% of theoriginal colonies persisted, while in MNNG-treated cultures, alarger percentage (~9%) of the colonies persisted. These per

centages remained constant from 3 to 7 weeks. Based on colonysize, cell density, and cell morphology, the persistent colonieswere classified into transformed colonies (large colony size, highcell density, high nuclear:cytoplasmic ratio) and untransformedcolonies (small size, low cell density, low nuclearcytoplasmicratio). In the MNNG-treated cultures, about 50% of all persistent

colonies showed transformed morphology. Their frequency remained unchanged between 3 and 7 weeks of culture. In contrast, only 10 to 15% of the persistent colonies in control culturesshowed transformed morphology at 3 weeks, but that proportionincreased steadily between 3 and 7 weeks. These data suggestthat, in control cultures, transformed colonies developed spontaneously as a function of time within untransformed colonies.

Autoradiographic studies with [3H]thymidine showed that labeling indices in the early "normal" RTE cell colonies between

Days 4 and 7 of culture were very high, ranging between 75 and90%. In contrast, the labeling indices of persistent colonies, boththose without and those with transformed morphology, werelow, i.e., between 18 and 25%, indicating that a major proportionof cells was either noncycling or cycling very slowly.

The relative compartment sizes of cells with stem cell characteristics and of cells with characteristics of transformed stemcells were estimated before and after transformed colonies appeared. The data showed that (a) transformed colonies werecomposed of nonstem cells, nontransformed stem cells, andtransformed stem cells; (b) the proportion of stem cells withtransformed growth characteristics increased steadily between2 and 10 weeks, as EG-variant colonies developed; and (c)

transformed stem cells with a capacity to attach and grow onplastic appeared at 5 weeks but with a frequency of only ~1%

of the total cell population in transformed cultures. Cells with this

1Recipient of National Research Service Award 1F32 ES05224-01 from the

NIH. Present address: Laboratory of Experimental Pathology, Division of CancerEtiology, National Cancer Institute, Frederick, MD 21701.

z To whom requests for reprints should be addressed.

Received 12/11/84; revised 4/9/85; accepted 4/11/85.

phenotype comprised 15% or more of the total cell populationof transformed RTE cell cultures at passage six. These studiesdemonstrate the heterogeneity of the cell populations comprisingtransformed cultures and the continuing selection occurring inthe transformed cell population.

INTRODUCTION

To understand the cellular and molecular basis of carcinogen-esis, specific cellular phenotypic changes essential for neoplastictransformation need to be identified and characterized. Cellculture systems have proven useful in studying a variety ofneoplasia-related changes (1-3) and have provided informationon the nature of those changes (1,2,4-8) and on specific cellular

events responsible for their expression (see Ref. 1). We havepreviously described a cell culture model, which is useful forquantitative investigations of different stages of neoplastic progression of RTE3 cells following carcinogen treatment (9-12).

This system is attractive for performing investigations on themechanism of carcinogenesis because (a) respiratory epitheliumis an important target for environmental carcinogens in humans,(b) comparative in vitro and in vivo studies can be performed(13-15), and (c) quantitative analyses of the induction by carcin

ogens of early, preneoplastic changes (14,16,17) as well as lateneoplastic events can be made.

The first, detectable stage in the neoplastic progression ofRTE cells in culture is the formation of large colonies of alteredcells, termed EG variant colonies (10, 12). The cell populationswhich comprise these EG variant colonies exhibit an enhancedgrowth potential when compared to normal cells. Our quantitativeassay for the transformation of RTE cells is based on an expression period for the fixation and expression of carcinogen-inducedevents during which all RTE cells proliferate, followed by aselection period which allows the growth of EG variants. Replication of the normal cells during the expression period is stimulated by growing the cells at clonal density on 3T3 feeder layers,which permits RTE cells to form colonies after 7 days with aCFE of 2 to 10%. Removal of the feeder cells from the cultureresults in cessation of division of the normal RTE cells, whichenlarge and slough from the dish (12). EG variants are recognizedunder these conditions at 3 to 6 weeks as large colonies ofproliferating, small hyperchromatic cells with increased nuclearcytoplasmic ratios. These cells usually can be subculturedindefinitely. The cells are initially anchorage dependent and non-tumorigenic but progress to become anchorage independent andultimately neoplastic, forming mostly keratinizing squamous cell

3The abbreviations used are; RTE, rat trachéalepithelial; EG, enhanced growth;CFE, colony-forming efficiency; CFU, colony-forming unit; EGV-CFU, enhancedgrowth variant colony-forming unit; CFlf, colony-forming unit on plastic; EGV-CFlf, enhanced growth variant colony-forming unit on plastic; MNNG, N-methyl-W'-nitro-N-nitrosoguanidine; HEPES, W^-hydroxyethylpiperazine-A/'^-ethanesul-

fonic acid; CFl/, colony-forming unit on feeders.

CANCER RESEARCH VOL. 45 JULY 1985

3322

Research. on January 10, 2020. © 1985 American Association for Cancercancerres.aacrjournals.org Downloaded from

STEM CELL POPULATIONS OF PRENEOPLASTIC TRACHEAL CELLS

carcinomas upon injection into nude mice. We have recentlyreviewed the evidence for the preneoplastic nature of EG variants(10).

EG variants are induced by carcinogens, such as MNNG, at ahigh frequency (^1 %/colony-forming cell). As described earlier(12), the linear dose-response curve is consistent with a one-hit

mechanism. EG variants can be distinguished in vitro from normalRTE cells on the basis of 3 phenotypic changes (10): (a) theyhave a reduced propensity to senesce; (b) they have an increasedproliferati ve potential; and (c) they have an altered substraterequirement; i.e., they are able to grow on plastic, which doesnot support the continued growth of normal cells. It is not knownwhether EG variants expressing these new properties developin a single step following carcinogen treatment of RTE cells orprogress through qualitatively distinct intermediates exhibitingone or more of these changes.

The purposes of the studies presented here are (a) to describethe appearance of EG variant colonies in control and carcinogen-treated cultures; (b) to analyze cell populations in transformedRTE cell cultures during the evolution of EG variant colonies andmeasure the sizes of nonstem cell, untransformed, and transformed stem cell compartments,4 and (c) to quantitate the emerg

ence in the transformed RTE cell cultures of new cell variantswhich have an altered substrate requirement allowing growth onplastic.

MATERIALS AND METHODS

RTE Cell Culture

RTE cells were obtained from the tracheas of normal 8-week-old maleFischer 344 rats (specific-pathogen-free) as previously described (9,12)

with the following modification. After removing the tracheas from thedonor animals, they were filled with a 1% Pronase solution (type XIV;Sigma Chemical Co., St. Louis, MO) in Ham's F-12 medium (Grand Island

Biological Co., Grand Island, NY) and incubated at 4°Cfor 16 to 20 h.

Cells were cultured on feeder layers of lethally irradiated 3T3 cells usingHam's F-12 medium supplemented with 5% fetal bovine serum, insulin

(1 i/g/ml), hydrocortisone (0.1 /¿g/ml),and antibiotics as described (9,12).

Transformation Assay

RTE cells (1 to 5 x 10*/60-mm dish) were plated onto lethally irradiated

3T3 feeder layers and exposed to MNNG 24 h later as previouslydescribed (9, 12) except for the following modifications. Stock solutionsof MNNG were prepared in distilled water, and cultures were exposedto MNNG for 4 h in 0.2 M HEPES (Calbiochem-Behring Corp., LaJolla,CA)-buffered F-12 (pH 6.8) without serum. Control cultures were exposed to F-12 medium buffered with HEPES. To determine the cytotox-

icity of MNNG treatment, the number of surviving colonies was countedat Day 7 of culture following ethanol fixation and staining with 10%aqueous Giemsa (Fisher Scientific Co.).

To follow the fate of colonies after selection for transformed EGvariants, RTE cell cultures were treated as above and grown for 4 or 7days to allow expression of the transformed cells, and then the feedercells were removed by treatment with a dilute EDTA solution (12). Setsof at least 5 cultures were fixed and stained at various times after feeder

' All cells able to form colonies within 8 days are considered to be "stem cells."This term is used interchangeably with the term "colony-forming cells" or CFU; cellswhich cannot form colonies within 8 days are termed "nonstem cells"; cells which

can form EG variant colonies during a 5-week period of selection are termed"transformed stem cells" or EGV-CFU.

cell removal, and the fraction of persistent colonies was calculatedrelative to the number of colonies present on Day 7 of culture.

Analysis of Cell Populations in Control and MNNG-treated RTE Cell

Cultures as a Function of Time

Cells from primary cultures were subcultured at different times afterselection (removal of feeders at Day 4 or Day 7 of culture) for EG variantsby first removing any contaminating rat trachéalfibroblasts with a brief(<1 min) treatment with 0.05% trypsin:0.02% EDTA (Grand Island Biological Co.) at 37°Cand then dissociating the RTE cell colonies by a 5-

min treatment using the same trypsin:EDTA concentration. RTE cellswere then collected and counted using a Model ZBI Coulter Counter(Coulter Electronics, Hialeah, FL) and were replated onto dishes with orwithout feeder cells. The cell populations thus obtained from MNNG andcontrol cultures were analyzed using 4 different colony-forming assays

designed to measure cells with different types of replicative ability.CFU Assay. This assay detects the colony-forming cells or CFU in a

culture, i.e., all cells which have sufficient replicative capacity to formcolonies of ^50 cells within 7 to 8 days after plating onto feeder layers(Fig. 1). All cells able to form colonies within 8 days are considered to be"stem cells." This term is used interchangeably with the term "colony-forming cells" or CFU; cells which cannot form colonies within 8 days aretermed "nonstem cells."

Determinations of the number or proportion of cells with CFU characteristics in a cell population, using the CFU assay, are always lowestimates. The values obtained are influenced by 2 factors: (a) the truenumber of CFU in the cell population and (b) the efficacy of the assay indetecting CFU.

EGV-CFU Assay. This assay detects only colony-forming cells whichare enhanced growth variants. These EGV-CFU are detected by replating

cells from primary cultures onto feeder layers, removing the feeder layers4 or 7 days later, and determining the number of colonies which have anEG variant morphology (Fig. 2) after a selection period of approximately4 weeks. Cells which can form EG variant colonies are termed "transformed stem cells" or EGV-CFU.

CFUP Assay. Cells from primary cultures are replated onto plastic,

and the number of colonies is determined 8 days later (stem cells ableto grow on plastic).

EGV-CFUP Assay. Cells from primary cultures are replated onto

plastic, and the number of colonies with EG variant morphology isdetermined 4 weeks later. Thus, this assay measures transformed stemcells able to grow on plastic.

Autoradiography

RTE cells were labeled with [mef/jy/-3H]thymidine (20 or 75 Ci/mmol)([3H]thymidine) (New England Nuclear, Boston, MA) for 24 h (0.1 /¿Ci/ml)

in complete medium containing 1 /JMunlabeled thymidine. Cultures werethen washed, fixed, dipped in nuclear track emulsion (type NTB-2;

Eastman Kodak Co., Rochester, NY), and processed as previouslydescribed (9).

RESULTS

Development of EG Variant Colonies. We examined the fateof RTE cell colonies after the feeder cells were removed on Day4 or Day 7 to select for EG variants and have monitored the timeof appearance of EG variant colonies (see Ref. 12). Normal RTEcells seeded at low density onto feeder layers formed colonieswithin 5 to 7 days, with a CFE of 2 to 10%. The morphology andsize of the cells in these colonies were fairly homogeneous (Fig.1). After the feeder cells were removed from the cultures, theRTE cells continued to divide for 2 to 3 days. Proliferation thenceased, and many cells enlarged and sloughed from the dish.

CANCER RESEARCH VOL. 45 JULY 1985

3323

Research. on January 10, 2020. © 1985 American Association for Cancercancerres.aacrjournals.org Downloaded from

STEM CELL POPULATIONS OF PRENEOPLASTIC TRACHEAL CELLS

Entire colonies were even lost with time after feeder cell removal.During the first 2 weeks of the experiment, all colonies in

control and treated cultures were morphologically similar. However, at 3 weeks, the persistent colonies (persistent colonies areall colonies, regardless of morphology, remaining 3 or moreweeks after plating) in control and treated cultures becamemorphologically heterogeneous and could be classified as transformed (EG variant colonies) or untransformed colonies. Asdescribed (26), EG variant colonies were composed mostly ofsmall, hyperchromatic cells with increased nuclearcytoplasmicratios (Fig. 2A). These cells were found in large numbers in all ofthe transformed colonies. Transformed colonies ranged in sizefrom about 5 to 20 mm in diameter when scored at 5 weeks. Incontrast, untransformed colonies were composed of large palestaining cells with greatly decreased nuclearcytoplasmic ratios-and of cells which appeared similar to normal RTE cells in 1-week-old cultures (Fig. 26). The size of the persistent colonies

classified as untransformed ranged from about 2 to 7 mm indiameter at 5 weeks. Due to differences in cell sizes betweentransformed and nontransformed colonies, transformed colonieshad approximately 5 to 10 times more cells than nontransformedcolonies of equal area. Chart 1 summarizes the changes in thenumber and type of colonies occurring between the time offeeder removal and Day 50 when the experiment was terminated.The data are expressed as the number of colonies present inthe cultures at various times, relative to the number of coloniespresent on Day 7. On Day 18, only 3% of the colonies presenton Day 7 remained in the cultures not treated with carcinogen(Chart 1/\). After Day 18, the number of colonies persistingremained constant until the termination of the experiment. Theabsolute frequency of persistent colonies in control culturesvaried greatly from experiment to experiment. The data illustrated in Chart 14 represent an experiment with a high frequencyof persistent colonies in control cultures. In many other experiments, this frequency was s*10-fold lower.

In MNNG-treated cultures, the rate of loss of colonies between

oz

i

1.0

01

O 0.01

0001

1.0

0.1

. 0.01

10 20 30DAYS

40 500.001

10 20 30DAYS

40 50

Chart 1. Loss of normal colonies and appearance of transformed colonies incontrol and carcinogen-treated primary RTE cell cultures. RTE cells were platedon feeder cells at 5000 cells/dish and were treated with F-12 medium buffered withHEPES (control cultures) or with MNNG (0.25 (¿g/ml)in F-12 medium buffered withHEPES (treated cultures). Feeder cells were removed on Day 4. Sets of 5 disheswere fixed and stained at the times indicated. The fraction of colonies observedrelative to the number of colonies present at Day 7 was determined at various timeintervals. A, persistent colonies (all types) in MNNG-treated (A) and control (O)cultures; 6, transformed colonies (EG variant colonies) in MNNG-treated (A) andcontrol (O) cultures.

Days 7 and 18 after feeder cell removal was very similar to thatobserved in control cultures (Chart 1X\); however, the relativenumber of colonies persisting between Days 20 and 50 wasabout 3-fold higher than that observed in controls. In 3 differentexperiments, the fraction of persistent colonies in MNNG-ex-posed cultures was quite reproducible, varying less than 2-foldcompared to a 10-fold variation in control cultures of the same

experiments. Thus, the magnitude of the increase in persistentcolonies in MNNG-exposed cultures over that in controls was

variable, due mostly to the variability in the controls.In the MNNG-exposed cultures, the frequency of EG variant

colonies was 3.4% at 3 weeks (i.e., 3.4% of the original coloniespresent at Day 7 after plating showed the morphological characteristics of EG variant colonies) and remained constant at avalue of about 4 to 5% for the subsequent weeks (Chart 1B). Incontrast, the frequency of transformed colonies in control cultures rose from 0.6% at 3 weeks to ~2% at 7 weeks. Aboutone-half of all persistent colonies in MNNG-exposed cultures

between 3 and 7 weeks was EG variant colonies. In controlcultures, this proportion was initially much lower and reachedthe 50% level only at about 7 weeks. These findings werereproduced in a subsequent experiment, which gave almostidentical results (data not shown). In a series of 10 transformationexperiments conducted previously over the period of about 1year, the mean frequency of MNNG-induced transformation

(MNNG, 0.25 to 0.3 ¿ig/ml)was 2.8 ±0.7%, while the meanfrequency of spontaneous transformation was 0.3 ±0.1%. Therelatively high spontaneous transformation frequency in the experiment reported here was unusual but made the comparisonof the development of EG variants in control and MNNG-exposed

cultures feasible.To determine the growth fraction in early as well as persistent,

transformed, and untransformed colonies, control and MNNG-treated cultures were labeled for 24 h with [3H]thymidine. A high

labeling index (s=75%) was observed following 24-h labeling with[3H]thymidine of colonies at Day 7 of culture (Table 1). The

fraction of labeled cells in colonies of both treated and controlcultures decreased to about 40% on Day 14. At 40 to 46 daysof culture, the mean labeling index in untransformed and transformed colonies of control as well as MNNG-treated cultures

was 18 and 25%, respectively. This indicates that the number ofcycling cells in both types of colonies was rather low comparedto the early "normal" RTE cell colonies. The mean labeling index

of transformed colonies was about 1.4 times that of untransformed colonies, and there was considerable overlap. No significant differences existed between control and MNNG-treated

cultures in the labeling indices of the different types of colonies;therefore, the data were combined. No obvious differences wereobserved in the distribution of labeled cells within transformedand nontransformed persistent colonies. The distribution of labeled cells within persistent colonies was generally random,although focal labeling was occasionally seen.

Changing Stem Cell Populations during Transformation ofRTE Cell Cultures. As previously described (10-13), transfor

mation of RTE cells is characterized by a profound change intheir growth capacity. Normal primary RTE cells have only a verylimited replicative potential, while transformed cells proliferate forextended periods of time. The purpose of the next series ofexperiments was to trace the evolution of this increased growthcapacity during the development of EG variant colonies by

CANCER RESEARCH VOL. 45 JULY 1985

3324

Research. on January 10, 2020. © 1985 American Association for Cancercancerres.aacrjournals.org Downloaded from

STEM CELL POPULATIONS OF PRENEOPLASTIC TRACHEAL CELLS

Table 1Fraction of cells labeled with pHJthymidine in colonies from treated and control

cultures of RTEcells with time after treatmentCultures of RTE cells were incubated with [3H]thymidine(0.1 (¿Ci/mi)for 24 h at

various times after treatment. Cultures were then fixed and processed for autora-diography, and the fraction of labeledcells in individualcolonieswas determined.

Time afterplating(days)4

71440-4640-46Labeled

cellsColonytype"Normal

(10)"

Normal (10)Normal(23)Nontransformed(26)Transformed (33)%»90

»7540±12C

18 ±1225 ±11Range15-50

2-505-50

BTransformed colonies are not distinguishable up to Day 14. All colonies up tothis point have the same morphology and are considered "normal." Since there

were no differences between control and MNNG-treated cultures in terms of labelingindices in the different types of colonies, the data from the 2 groups were combined.

6 Numbers in parentheses, number of colonies by type.°Mean±SD.

Table 2Number and frequency of CFUin control and MNNG-treatedRTEcell cultures as

a function of timeBetween 2 and 6 x 103 RTE cells were plated into 60-mm dishes and were

exposed to MNNG (0.25 Mg/ml) or solvent only (0.02 M HEPES) as a control (see"Materials and Methods" for detail). The relative survival in the MNNG-treated

primary culture was 50%; there were between 68 and 152 Day-7 colonies per

culture. One to 5 weeks later, the primary cultures were dissociated, the numberof cells per dish was determined (7 to 10 cultures per group and time point), andthe cells were replated on feeders to determine the number of CFU per primaryculture.

TreatmentControlMNNGControlMNNGControlMNNGDaysafter

treatment7714143535Av.

no. ofcellsrecovered/dish(x

ioy1.61.26.2±2.1C6.6

±42.7

±1.523.2±11.5CFE

on feedersofRTEcellsreplatedfromprimary cul

tures4.4x10-24.1x10~24±1tf

X10-41.5±0.6X10-32.3

±1.1X10-»1±0.4 x10'2Av.

no.ofCFUpresentin

primary cultures*7044922599622320

8 Different numbers of colonies survived in control and treated cultures. In order

to make direct comparisons of cell number, CFU, and EGV-CFU between controland MNNG-exposed cultures, the measurements were normalized for differencesin the number of colonies at Day 7 in the treated and control primary cultures.

b The number of CPU/primary culture was calculated with the equation: CFU/

dish = (normalized number of cells recovered per dish) x (CFE of replated cells).0 Mean ±SO." Mean ±SE.

quantitatively analyzing the RTE cell cultures for subpopulationsendowed with different growth potentials. We used several clonalgrowth assays described in "Materials and Methods" to distinguish nonstem cells, stem cells, and transformed stem cells.4

In the experiments summarized in Tables 2 to 4, the transformation frequency (i.e., number of EG variant colonies per surviving Day-7 colony), determined at Day 35 after treatment of theprimary cultures, was 2.9% in MNNG-treated cultures and 0.8%

in control cultures. As shown in Table 2, marked differenceswere observed in the growth pattern of MNNG-treated cultures

compared to control cultures. The total number of cells in controlcultures increased nearly 4-fold from Day 7 to Day 14 and then

decreased by 53% during the subsequent 3 weeks. These valuesare consistent with previously published growth curves, whichshowed that normal primary RTE cell cultures grow for 10 to 14days, become stationary, and then slowly decrease in cell number (11, 16). The growth of MNNG-treated cultures was quite

different. At 1 week, the number of cells per culture was slightly

lower than in controls. This is consistent with our previousobservations that MNNG reduces not only the CFE of the cellsbut also the growth of the surviving colonies (9). The cell numberin treated cultures increased >5-fold from Day 7 to Day 14 andincreased >3-fold during the subsequent 3 weeks. As a resultof this difference in growth pattern, the cell number in MNNG-exposed cultures was >8 times that of control cultures at 5\nrCCKS.

To assess the relative sizes of the proliferative cell pools incultures as a function of time and MNNG exposure, the primarycultures were subcultured at 1, 2, and 5 weeks onto irradiatedfeeder cells to determine the number of CFU per culture (seeTable 2). At 1 week, the number of CFU per culture was lower(30%) in MNNG-treated cultures than in control cultures. This

was due to the reduced cell number in the cultures, since theCFE of the cells from both treated and control cultures was thesame. Between 1 and 2 weeks, CFE decreased drastically incontrol as well as treated cultures, but less so in the latter. Thenumber of CFU in the treated cultures at 2 weeks was nearly 4times greater than the number of CFU in the controls. BetweenWeeks 2 and 5, a 23-fold increase in the number of CFUs per

culture occurred in treated cultures, which resulted from anincrease in both cell number and CFE. This suggests that a stemcell population was selectively expanding during this time. Incontrol cultures, on the other hand, the number of CFU increasedonly 2.5-fold (the number of cells per culture decreased 2.3-fold,and the CFE increased 5.8-fold). Comparison of the size of theCFU compartments indicates that, at 5 weeks, there are approximately 37 times as many stem cells in the MNNG-treated as inthe control cultures.

These data suggest that the number of cells with stem cellcharacteristics decreased dramatically between 1 and 2 weeksin both control and MNNG-treated cultures. It then increasedrapidly in the MNNG-treated cultures during the subsequent 3

weeks, during which time EG variant colonies developed andenlarged, although the CFE in 5-week treated cultures was <1 %.

In contrast, it increased slowly in the control group in which thetransformation frequency was about one-fifth of that in MNNG-

treated cultures.MNNG-treated as well as control cultures were further exam

ined for the presence of transformed stem cells (EGV-CFU) at 2and 5 weeks after treatment (see "Materials and Methods" for

EGV-CFU assay). As shown in Table 3, the number of EGV-CFUat 14 days in MNNG-treated cultures was 10-fold greater than

in the untreated cultures. However, the frequency of the totalcells present on Day 14 which can form EG variant colonies onreplating was very low in control as well as treated cultures (3 x10"5 and 3 x 10~4, respectively). Twenty % of the CFU present

at 2 weeks in MNNG-treated cultures had transformed characteristics (ratio of EGV-CFU to total CFU) as compared to only

8% in the control cultures.From Day 14 to Day 35, the number of EGV-CFU had in

creased 27-fold in control cultures and 58-fold in treated cultures,

indicating the extent of proliferation of transformed stem cells inboth types of cultures (Table 3). The frequency of cells whichhave EGV-CFU properties was similar (2 x 10~3versus 5 x 10~3)

at 5 weeks in control and treated cultures, although the totalnumber of EGV-CFU was now 21-fold greater in treated than in

control cultures. It is important to notice that, even at 5 weeks,only s;0.5% of all cells in control as well as treated cultures was

CANCER RESEARCH VOL. 45 JULY 1985

3325

Research. on January 10, 2020. © 1985 American Association for Cancercancerres.aacrjournals.org Downloaded from

STEM CELL POPULATIONS OF PRENEOPLASTIC TRACHEAL CELLS

Table3Number and frequency ol EGV-CFU in control and MNNG-treated RTE cell cultures as a function of time

These were the same primary cultures as those described in Table 1.

TreatmentControlMNNGControlMNNGDaysafter

treatment14143535Av.

no.ofEGV-CFUpres

ent in primarycultures*220541160Frequency

ofEGvariantcolonies/re-

platedcell3±2CX10-"3

±1.5X10-*2

±0.7X10-35±2 xio-*Ratio

ofEGV-CFUtoCFU

in primarycultures"0.080.200.870.50Estimated

no.ofnontransformedCFU

in primaryculture*237981160

" Normalized total numbers of EGV-CFU and nontransformed CFU per dish based on normalized total

number of cells per dish from Table 2. EGV-CFU/dish = (normalized number of cells/dish) x (frequency of EGvariant cdonies/replated cell). Nontransformed CFU/dish = (total CFU/dish from Table 2) - (EGF-CFU/dish).

6 CFU/culture from Table 2.0 Mean ±SE.

identifiable as transformed stem cells by the EGV-CFU assay.The fact that the total cell number was sS-fold greater in treated

than in the control cultures (Table 2) indicated that remnant,normal cells cannot account for this large proportion of nontransformed and nonstem cells in the treated (or the untreated)cultures. This result therefore suggests that many of the cells inthe EG variant colonies are terminal and nonproliferating, whichis consistent with the results of the autoradiographic studies(see Table 1).

Fifty to 90% of the stem cells in both treated and controlcultures at 5 weeks were transformed as indicated by the ratioof EGV-CFU to CFU (Table 3). While this represents a largepercentage of the total stem cells, nontransformed stem cellswere still present at Day 35. The calculated number of nontransformed CFU (;.e., the number of EGF-CFU subtracted from thetotal number of CFU) increased -15-fold in treated cultures from

Day 14 to Day 35, suggesting that the cells in EG variant coloniesgenerate both transformed and nontransformed stem cells.

Development of Transformed Stem Cells with ReducedSubstratum Requirements. Preneoplastic RTE cell transform-ants have the capability to grow on plastic substratum (16-18),

but it is not known when during the course of transformationthis capability is acquired. We therefore examined the timecourse of development of the altered substrate requirement andits relationship to the development of EG variant colonies. It hasbeen our experience that normal RTE cells have a very low CFEon plastic when cultured in the medium described here. Thenumber of cells capable of forming Day-7 colonies and EG variantcolonies on plastic (CFLf and EGV-CFLf0, see "Materials andMethods") was determined at Days 14 and 35 from the same

primary cultures described above in Tables 2 and 3. The CFE onplastic was significantly less than the CFE on feeder cellsthroughout the course of the experiment (Table 4). At 2 weeks,the CFE on plastic was 3 x 10~B for control cultures and 7 x1CT5for MNNG-treated cultures compared to 4 x 10~4 and 1.5

x 1(T3 for CFE on feeder layers (see Table 2). At 5 weeks, theCFE on plastic had increased to 3 x 10~4and 6 x 10~4 in control

and MNNG-treated cultures, respectively, which was still 10-foldless than CFE on feeder layers. The average number of CFUP at

2 weeks was 0.2 per control culture and 4.6 per MNNG-treatedculture. At 5 weeks, approximately 8 CFLr°were found per

control culture compared to 140 per MNNG-treated culture (Ta

ble 4).At Day 14 and Day 35, the total number of CFlf0 in the cultures

was similar to the total number of EGV-CFLf0 in both control and

treated cultures (Table 4). In fact, the ratio of EGV-CFUP to CFLf

was near and in some cases above unity. It is possible that somecolony-forming cells grew initially very slowly on plastic and were

therefore not scored as colonies (i.e., <50 cells) on Day 7 butgrew sufficiently thereafter to be scored as EG variant colonieson Day 35. It is equally likely that the values above unity simplyreflect experimental error.

Since the number of transformants which are able to grow onplastic is about one-fifth of the number of transformants able to

grow on feeder layers, the cells capable of growth on plasticmay represent a subpopulation of transformed stem cells. Wetherefore extended the analysis of the growth properties of theEG variants beyond 35 days by analyzing the CFE of 70-day-old

cultures containing primary EG variant colonies and of an established EG variant cell line (EGV-1) at passage 6, i.e., about 150

days after isolation from a primary culture treated with MNNG(12). The average CFE of cells from 35-day-old MNNG-exposed

cultures when replated onto feeder cells and plastic was 0.74%and 0.04%, respectively (Table 5); the ratio of the CFE on plasticto the CFE on feeder cells was 0.05. At 70 days after treatment,the average CFE of cells from primary cultures upon replatingwas 1.8% on feeder cells (~2.5-fold increase from Day 35) and1.2% on plastic (~30-fold increase from Day 35). The ratio of the

CFE on plastic to the CFE on feeder cells (0.67) had increasedmore than 10-fold between Day 35 and Day 70 in primary culture.Finally, EGV-1 at passage 6 had a CFE on feeder layers which

was about 10 times higher than that of the EG variants at Day70. Furthermore, the CFE on plastic nearly equaled the CFE onfeeder cells (ratio = 0.88).

DISCUSSION

The purposes of our experiments were to study the development of transformed colonies from normal RTE cells and toanalyze the transformed cell populations in terms of their growthpotential. As described in previous publications (10-12, 17, 18),

the first observed phenotypic alteration of the RTE cell transformant is a markedly enhanced growth capacity resulting in "immortalization"; thus, the RTE cell transformants were termed EG

variants. We previously demonstrated in a number of studiesthat this EG variant is preneoplastic (11-13).

Our transformation assay is designed to select for the EGvariants by removing from the cultures the 3T3 feeder layerswhich are required for the continued growth of normal RTE cells.When the feeder layers were removed on Day 4 or 7 of culture,

CANCER RESEARCH VOL. 45 JULY 1985

3326

Research. on January 10, 2020. © 1985 American Association for Cancercancerres.aacrjournals.org Downloaded from

STEM CELL POPULATIONS OF PRENEOPLASTIC TRACHEAL CELLS

Table4Number and frequency of CFUf and EGV-CFUf growing on a plastic substratum in control and MNNG-treated

RTE cell cultures at different times

These were the same primary cultures as described in Table 1.

TreatmentControlMNNGControlMNNGDaysaftertreatment14143535CFE

onplasticofcellsreplatedfrom

primarycultures3±3"x

1fr«7±5xirr53±3x1fr46±2X10-4No.

ofCPU"inprimarycultures'0.24.68.1139.2FrequencyofEGvariant

coloniesonplastic/replatedcell7

±7"xirr65±3 x10-'5±3

X10-41.4±0.5 X10-3No.

ofEGV-CFlfon

plasticinpnmarycultures'0.43.313.5324.8Ratio

ofEGV-CFU"to CFU"onplastic

ofcellsfromprimary

culture2.00.71.72.3

' CFUP = cells forming 7-day colonies on plastic; EGV-CFlf* = cells forming EG variant colonies on plastic.

Calculated from normalized total number of cells/dish (Table 2) and either CFE on plastic or frequency of EGvariant colonies on plastic per replated cell.

' Mean ±SE.

Tables

Progressive changes in substratum requirements of transformed RTE cells duringlate primary culture and at early subculture: CFE on feeder layers and on plastic

Cells recovered on Day 35 or Day 70 from 7 MNNG-treated primary cultureswere replated to determine CFE on feeder layers or on plastic. EGV-1 is atransformed RTE cell line and was tested at passage 6.

TransformantsPrimary

EG variantsPrimary EG variantsCell line EGV-1Days

aftertreatment35

70-150Av.

CFE onfeeder cells

(%)0.74

1.817Av.

CFE onplastic(%)0.04

1.215Ratio

ofCFE on

plastic toCFE onfeeders0.05

0.670.88

the vast majority of colonies ceased to grow; the cells enlarged,became squamous like, and entire colonies sloughed from thedishes. This process appears to be analogous to changes observed in epidermoid cells and human bronchial cells undergoingterminal differentiation (19-24). The EG variant colonies were

identified at 3 to 5 weeks primarily based on their morphology(i.e., cell type and cell density) and their size. In addition to theEG variant colonies, a second colony type persisted in thecultures after feeder cell removal. These colonies were distinguished from EG variant colonies based on their smaller size,their lower cell density, and the low nuclearcytoplasmic ratio ofmost of their cells (see Fig. 2B). The number of these nontrans-formed persistent colonies in control cultures was highly variablefrom experiment to experiment. Initially, we assumed that thecells comprising these nontransformed colonies were nondivid-ing. However, in this study, we observed that most of thecolonies contained cells which synthesize DNA as measured byincorporation of [3H]thymidine. Furthermore, we found that, in

control cultures, the fraction of persistent colonies with transformed morphology increased with time, suggesting that someof the nontransformed persistent colonies progressed to thetransformed state. In contrast, the number of EG variant coloniesin carcinogen-treated cultures remained relatively constant with

time. \These findings are important for understanding the variables

in the RTE cell transformation assay. A problem in this assayhas been the variability of the spontaneous transformation frequency. Although this frequency can be as high as 1%, theaverage of 12 experiments carried out over the last 18 monthshas been 0.15%.5 We have reported that the spontaneous

5 D. Thomassen, J. C. Barrett, D. K. Beeman, and P. Nettesheim, unpublished

data.

transformation frequency increases when normal RTE cells aregrown on 3T3 feeder cells for increasing periods of time (12).We have also found that certain lots of sera which have stronggrowth-stimulatory activity increased the spontaneous transformation frequency.5 These observations suggest that sponta

neous transformation arises as a result of events occurringduring periods of sustained cell replication. There are also unknown factors which affect spontaneous transformation, sincewe have observed considerable variability even under presumably similar conditions.

Transformed EG variant cell populations, whether spontaneous or MNNG induced, shared several characteristics. First,the EG variant colonies contained heterogeneous cell populations of dividing and nondividing cells. This was deduced from[3H]thymidine labeling studies which showed that between 5 and

50% of the cells within different transformed colonies werelabeled, following a 24-h pulse as compared to >75% labeledcells in 7-day colonies of normal RTE cells grown on feeders.

Interestingly, only slight differences in the labeling indices ofnontransformed persistent colonies and EG variant colonies wereobserved. The fact that the EG variant colonies rapidly increasedin cell number and size and the persistent untransformed coloniesremained small therefore suggests that the rate of differentiation,cell death, and/or exfoliation may be different in the persistentnontransformed colonies from those in the EG variant colonies.Further studies are required to define more precisely the kineticsof growth and differentiation in persistent nontransformed andtransformed colonies.

A second feature of the EG variant population was that only afraction of its cells exhibited stem cell characteristics as measured by the CFU assay (on feeder cells). This supports theconclusions derived from the autoradiographic studies, namely,that the composition of the cell population in transformed colonies is heterogeneous and that the majority of its cells arenondividing. Even at 5 weeks, «1% of the cells in the MNNG-treated cultures had colony-forming potential. However, the"true" number of stem cells cannot be determined with the CFU

assay, since its efficacy of detecting such cells is not known. Wealso do not know what proportion of cells is rendered nonreplicative by the dissociation procedures and other manipulationsused.

A third important feature of the EG variant cell population wasthat not all of its stem cells appeared to be transformed, since<100% of them were able to form transformed colonies as

CANCER RESEARCH VOL. 45 JULY 1985

3327

Research. on January 10, 2020. © 1985 American Association for Cancercancerres.aacrjournals.org Downloaded from

STEM CELL POPULATIONS OF PRENEOPLASTIC TRACHEAL CELLS

measured by the EGV-CFU assay (only 50 and 87% in treatedand control cultures, respectively, at 5 weeks). One of thecharacteristics of EG variants is that they can be subculturedand can give rise to immortal cell lines (10). This provides anopportunity for studying the mechanism of cellular immortality,which has been proposed by some as a key event in neoplasticprogression (5,10,17,24,25). Our studies indicate that unlimitedproliferative capacity or immortality is a population characteristicand that both mortal and immortal subpopulations exist within atransformed colony. The important point of our results which wewish to emphasize is that, in carcinogen-treated cultures, not

only did the number of transformed stem cells increase with timebut also the number of nontransformed stem cells increasedsignificantly. The only explanation is that some of the progenyof transformed stem cells are committed to a limited life span.

Transformed colonies in 5-week-old cultures of RTE cells

therefore seem to be composed of several different cell types:transformed stem cells, untransformed stem cells, and nonstemcells or terminal cells, the latter representing the bulk of the cellpopulation. In this regard, the preneoplastic EG variant colonies-

may not be too dissimilar from tumor cell populations of differentiating carcinomas in which tumor stem cells produce terminally differentiating cells (e.g., Ref. 26; for review, see Ref. 27).If the interpretation of our data is correct, then one would predictthat one of the crucial factors determining the rate of progressionin cell populations generated from a transformed stem cell is theproportion of cells among its offspring with unlimited replicativepotential. Factors which can tip the balance of cell replication infavor of nontransformed terminal cells should reduce the rate ofprogression.

An additional component of EG variant cell populations is asubpopulation of cells with a reduced substrate requirement, i.e.,the ability to grow on plastic. Growth on plastic has previouslybeen used as a marker of cell transformation (19), and we havefound in earlier studies that epithelial cell populations isolatedfrom tracheas at various times after carcinogen exposure containcells capable of proliferating on plastic, while normal trachea!epithelial cells are not able to do so (16-18). In the present

experiments, we found this phenotype to be extremely rare inearly control cultures as well as early MNNG-exposed cultures(<10~4 of the cells). At 35 days, the fraction of stem cells able to

grow on plastic was quite low (CFl/iCFI/ ~ 0.05), but it in

creased with time, and at 70 days after MNNG exposure, 67%of all stem cells (CFU) in exposed primary cultures were capa-

bleof growth on plastic. This suggests that a new cell variantemerged, which has a selective growth advantage. We do notbelieve that the appearance of this new phenotype is simply theresult of adaptation of the transformed cells to the cultureconditions, because we previously showed that transformed cellvariants able to grow on plastic can be found in cell populationsfreshly isolated from tracheas exposed to carcinogen in vivo(16-18).

These studies are the first attempt to measure the sizes ofdifferent cell compartments in cultures of RTE cells undergoingspontaneous and induced transformation. Four cell compartments were identified: the compartment of terminal cells; thecompartment of nontransformed stem cells; the compartment oftransformed stem cells requiring feeder cell support for initialgrowth; and finally the compartment of transformed stem cellsable to grow on plastic. The 2 types of transformed stem cells

appeared to be 2 distinct phenotypic variants developing sequentially in these RTE cell cultures. Thus, the development ofRTE cell variants seems to be a multistep process which followssimilar pathways in carcinogen-treated and control cultures, al

though carcinogen treatment increases the frequency and decreases the time required for the detection of EG variants. Thedata further suggest that the majority of cells present in transformed colonies do not exhibit transformed characteristics; however, the proportion of transformed to nontransformed cellsrapidly increases with time in the transformed cultures. In theEG variant cell line EGV-1,15% of the cells exhibited transformedcharacteristics compared to only ~1% in the 70-day-old MNNG-treated primary cultures. Thus, qualitative and quantitativechanges continue to occur in MNNG-exposed RTE cell culturesfor many weeks. Future experiments will be conducted with cellsfrom isolated EG variant colonies to determine more preciselychanges in the transformed colonies as a function of time and ofexposure to modulators of neoplastic progression.

ACKNOWLEDGMENTS

We are indebted to Alma Gonzalez for her patience and diligence in thepreparation of this manuscript.

REFERENCES

1. Barrett, J. C., and Thomassen, D. G. Use of quantitative cell transformationassays in risk estimation. In: V. B. Vouk, G. C. Butler, D. G. Hoel, and D. B.Peakall (eds.), Methods of Estimating Risk and Human Chemical Damage inNon-Human Biota and Ecosystems, SCOPE, SGOMSEC 2, IPCS Joint Symposium 3. Chichester: John Wiley & Sons, in press, 1985.

2. Barrett, J. C., and Ts'o, P. O. P. Mechanistic studies of neoplastic transformation of cells in culture. In: H. Gelboin and P. 0. P. Ts'o (eds.), Polycyclic

Hydrocarbons and Cancer, pp. 235-267. New York: Academic Press, Inc.,1978.

3. Heidelberger, C., Freeman, A. E., Pienta, R. J., Sivak, A., Bertram, J. A.,Casto, B. C., Dunkel, V. C., Francis, M. C., Kakunaga, T., Little, J. B., andSchectman, L. M. Cell transformation by chemical agents: a review andanalysis of the literature. Mutât.Res., 114: 283-385,1983.

4. Barrett, J. C., Hesterberg, T. W., and Thomassen, D. G. Use of cell transformation systems of carcinogenicity testing and mechanistic studies of carcino-genesis. Pharmacol. Rev., 36: 53-70,1984.

5. Barrett, J. C. and Ts'o, P. O. P. Evidence for the progressive nature of

neoplastic transformation in vitro. Proc. Nati. Acad. Sci. USA, 75: 3761-3765,1978.

6. Franks, L. M., and Wigley, C. B. Neoplastic Transformation in DifferentiatedEpithelial Cell Systems in Vitro. New York: Academic Press, Inc., 1979.

7. McGrath, C. M., Brennan, M. J., and Rich, M. A. Cell Biology of Breast Cancer.New York: Academic Press, Inc., 1980.

8. Yuspa, S. H., Hawley-Nelson, P., Koehler, B., and Stanley, J. R. A survey oftransformation markers in differentiating epidermal cell lines in culture. CancerRes., 40: 4694-4703,1980.

9. Gray, T. E., Thomassen, D. G., Mass, M. J., and Barrett, J. C. Quantitation ofcell proliferation, colony formation, and carcinogen-induced cytotoxicity of rattrachea) epithelial cells grown in culture on 3T3 feeder layers. In Vitro (Rock-ville), 19: 559-570,1983.

10. Nettesheim, P., and Barrett, J. C. Trachéalepithelial cell transformation: amodel system for studies on neoplastic progression. CRC Crit. Rev. Toxico).,72:215-239,1984.

11. Pai, S. B., Steete, V. E., and Nettesheim, P. Neoplastic transformation ofprimary trachéalepithelial cell cultures. Carcinogenesis (Lond.), 4: 369-374,1983.

12. Thomassen, D. G., Gray, T., Mass, M. J., and Barrett, J. C. A high frequencyof carcinogen-induced early, preneoplastic changes in rat trachéalepithelialcells in culture. Cancer Res., 32: 5956-5963,1983.

13. Marchok, A. C., Rhoton, J. C., and Nettesheim, P. In vitro development ofoncogenicity in cell lines established from trachéalepithelium pre-exposed invivo to 7,12-dimethylbenz(a)anthracene. Cancer Res., 38: 2030-2037,1978.

14. Nettesheim, P., and Griesemer, R. A. Experimental models for studies ofrespiratory tract Carcinogenesis. In: O. Lenfant and C. C. Harris (eds.), Path-ogenesis and Therapy of Lung Cancer: Lung Biology in Health and Disease,Vol. 10, pp. 75-188. New York: Marcel Dekker, Inc., 1978.

15. Terzaghi, M., Nettesheim, P., and Williams, M. L. Repopulation of denudedtrachea! grafts with normal, preneoplastic, and neoplastic epithelial cell populations. Cancer Res., 38: 4546-4553,1978.

CANCER RESEARCH VOL. 45 JULY 1985

3328

Research. on January 10, 2020. © 1985 American Association for Cancercancerres.aacrjournals.org Downloaded from

STEM CELL POPULATIONS OF PRENEOPLASTIC TRACHEAL CELLS

16. Pai, S. B., Steele, V. E., and Nettesheirn, P. Identification of early carcinogen-induced changes in nutritional and substrate requirements in cultured trachéalepithelial cells. Carcinogenesis (Lond.), 3:1201-1206,1982.

17. Terzaghi, M, and Nettesheirn, P. Dynamics of neoplastic development incarcinogen-exposed trachéalmucosa. Cancer Res., 39:4003-4010,1979.

18. Marchok, A. C., Rhoton, J. L., Griesemer, R. A., and Nettesheirn, P. Increasedin vitro growth capacity of trachéal epithelium exposed in vivo to 7,12-dimethylbenz(a)anthracene. Cancer Res., 37:1811-1821,1977.

19. Bertolero, F., Kaighn, M. E., Gonda, M. A., and Saffiotti, U. Mouse epidermalkeratinocytes: donai proliferation and response to hormones and growth factorin serum-free medium. Exp. Cell Res., 755: 64-80, 1984.

20. Kutesz-Martin, M. F., Koehter, B., Hennings, H., and Yuspa, S. H. Quantitativeassay for carcinogen-altered differentiation in mouse epidermal cells. Carcinogenesis (Lond.), 1: 995-1006,1980.

21. Lechner, J. F., Haugen, A., McClendon, I. A., and Shamsuddin, A. M. Inductionof squamous differentiation of normal human bronchial epithelial cells by small

amounts of serum. Differentiation, 25: 229-237, 1984.22. Pera, M. F., and Gorman, P. A. In vitro analysis of multistage epidermal

carcinogenesis: development of indefinite renewal capacity and reducedgrowth factor requirements in colony forming keratinocytes precedes malignant transformation. Carcinogenesis (Lond.), 5: 671-682,1984.

23. Rheinwald, J. G. Serial cultivation of normal human epidermal keratinocytesMethods Cell Btol.. 21: 229-254,1980.

24. Sun, N-C., Sun, C., Chao, L., Fung, W., Tennant, R., and Hsie, A. In vitrotransformation of Syrian hamster epidermal cells by W-methyl-W'-nitro-W-nitro-

soguanidine. Cancer Res., 41:1669-1676,1981.25. NewbokJ, R. F., Overall, R. W., and Connell, J. R. Induction of immortality is

an early event in malignant transformation of mammalian cells by carcinogens.Nature (Lond.), 299: 633-635,1982.

26. Pierce, G. B., and Wallace, C. Differentiation of malignant to benign cells.Cancer Res., 3Õ: 127-134,1971.

27. Buick, R. N., and Pollack, M. N. Perspectives of clonogenic tumor cells, stemcells, and oncogenes. Cancer Res., 44: 4909-4918,1984.

CANCER RESEARCH VOL. 45 JULY 1985

3329

Research. on January 10, 2020. © 1985 American Association for Cancercancerres.aacrjournals.org Downloaded from

STEM CELL POPULATIONS OF PRENEOPLASTIC TRACHEAL CELLS

Fig. 1. Day-7 RTE cell colonies. A, several well-circumscribed epithelial colonies on 3T3 feeder layer, x 56. B, individual colony composed of basophilic cells withmoderately high nuclearcytoplasmic ratio, typical of normal RTE cells in culture, x 160.

CANCER RESEARCH VOL. 45 JULY 1985

3330

Research. on January 10, 2020. © 1985 American Association for Cancercancerres.aacrjournals.org Downloaded from

STEM CELL POPULATIONS OF PRENEOPLASTIC TRACHEAL CELLS

L*,t:••Vi«v•/;d'3v;««v*'!W*-Vft.JV»•/!'•.*•.••••>•>•*•rf

f /.t.•'»*

.-*.

BFig. 2. Photomicrograph of transformed (A) and untransformed (S) persistent colonies of RTE cells. Colonies were fixed with 70% ethanol and stained with 10%

aqueous Giemsa 5 weeks after treatment with MNNG. x 80.

CANCER RESEARCH VOL. 45 JULY 1985

3331

Research. on January 10, 2020. © 1985 American Association for Cancercancerres.aacrjournals.org Downloaded from

1985;45:3322-3331. Cancer Res   David Thomassen, J. Carl Barrett, Diane K. Beeman, et al.   Cultures at an Early Stage in Neoplastic ProgressionChanges in Stem Cell Populations of Rat Tracheal Epithelial Cell

  Updated version

  http://cancerres.aacrjournals.org/content/45/7/3322

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/45/7/3322To request permission to re-use all or part of this article, use this link

Research. on January 10, 2020. © 1985 American Association for Cancercancerres.aacrjournals.org Downloaded from