(CANCER RESEARCH 50, 6626-6631. October 15. 1990|

Mechanisms of Resistance of Confluent Human and Rat Colon Cancer Cells toAnthracyclines: Alteration of Drug Passive Diffusion1

HélènePelletier,2 Jean-Marc Millot, Bruno Chauffert, Michel Manfait, Philippe Genne, and FrançoisMartin

Research Group on Digestive Tumors, INSEKM U252, Faculty of Medicine, 7 Bd Jeanne d'Arc, 21033 Dijon Cedex [H. P., B. C., P. G., F. M.], and Laboratory of

Biomolecular Spectroscopy, Faculty of Pharmacy, 51 rue Cognacq Jay, 5I096 Reims Cedex fj. M. M., M. M.], France

ABSTRACT

Two colon cancer cell lines, HT-29 (human) and DHD/K12/TRb (rat),were grown as monolayer cultures to various confluence degrees. Thecytotoxic efficacies of doxorubicin and 4'-deoxydoxorubicin, evaluated

by a survival assay, and the nuclear drug concentrations, measured bymicrospectrofluorometry, were shown to progressively decrease with theaugmentation of confluence. This confluence dependent resistance (CDR)to anthracyclines was demonstrated independent of the multidrug resistance drug efflux mechanism. The cellular uptake of three compounds(sodium [*'Cr|chromate, D-l'^Cjalanine, i.-|14C)glucose) known to pas

sively diffuse across the cell membrane as anthracyclines do was alsoreduced in confluent cells. After trypsin cell detachment, the kinetics ofreversion of the sodium ("Crjchromate uptake decrease and that of CDRwere similar. Therefore, CDR may be attributed to a reduction ofanthra-cycline cell intake due to a general alteration of passive diffusion acrossthe cell membrane. However, CDR is only partly explained by thisphenomenon since a reduced sensitivity of confluent cells was observedcompared with nonconfluent cells for a similar amount of drug in theirnuclei. CDR could explain the high resistance to anthracyclines of somesolid tumors, such as colon tumors, in which cancer cells are tightlyaggregated.

INTRODUCTION

Anthracyclines are among the most active anticancer agents,widely used in treatment of solid tumors and leukemias. However, a natural or acquired anthracycline resistance of tumorsoften hinders the curative potentiality of these drugs.

One of the most investigated mechanisms of anthracyclineresistance consists of an increased active drug efflux out of thecancer cell supporting, at least partly, their M DR1 phenotype

(1-3). Verapamil (4), amiodarone (5), and some other compounds are able to inhibit this efflux and to circumvent anthracycline resistance in some experimental cancer cell systems.

However, in solid tumors other mechanisms than MDR havebeen put forward to explain their resistance to anthracyclines:alteration of vascularization; low drug penetration throughmultiple cell layers, as clearly demonstrated by experiments onspheroids; and deleterious effect of low pH and pO2 in tumordepth (reviewed in Refs. 6 and 7). Moreover, several in vitrostudies demonstrated the resistance of noncycling cells to anthracyclines comparing the drug sensitivities of log- and plateau-phase monolayer cultures (8-12). This resistance was evenmore pronounced when cells were cultured as spheroids compared to monolayers (13). The anthracycline resistance of non-cycling cells could be of first importance to explain the naturalresistance of slow growing solid tumors such as colon cancer,and mechanism(s) underlying this phenomenon must be precise.

Received 12/12/89; accepted 7/5/90.The costs of publication of this article were defrayed in part by the payment

of page charges. This article must therefore be hereby marked advertisement inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact.

' Supported by Contract 86023 with GLAXO-France Laboratories.2To whom requests for reprints should be addressed.3The abbreviations used are: MDR. multidrug resistance; DXR. doxorubicin;

deoDXR, 4'-deoxydoxorubicin; CDR, confluence dependent resistance; PBS.

phosphate buffered saline.

In the present study, we have showed a dramatic diminutionof nuclear accumulation and cytotoxicity of anthracyclines related to the increase of the cellular confluence of monolayercultures of two colon cancer cell lines. Our results have led usto suggest that the increase of confluence induces a generalalteration of passive diffusion across the cell membrane and, asa consequence, a reduction of the cell intake of anthracyclines.

MATERIALS AND METHODS

Drugs. DXR was obtained from Roger Bellon Laboratories (Neuilly-sur-seine, France). deoDXR was a gift from Farmitalia Laboratories(Milan, Italy). Amiodarone was obtained from Labaz Laboratories(Bordeaux, France) and was used in this study for inhibiting the activeanthracycline efflux transport in PROb cells (5). Stock solutions (1 g/liter) of these drugs were prepared in distilled water. Final dilutionswere prepared in Ham's F-10 medium immediately before each exper

iment.Cell Lines. The HT-29 cell line of human colon adenocarcinoma was

originally established from a xenograft by Fogh and Trempe (14). TheDHD/K12/TRb cell line (referred to as PROb) was established in ourlaboratory from a transplantable colon adenocarcinoma induced by 1,2-dimethyl In (Ira/ine in syngeneic BD IX rats (15). PROb cells exhibit aprimary resistance to anthracyclines due partly to an active drug efflux(16). By immunoblotting, using MRK-16 antibody, we have shown theexistence of membrane P-180 glycoproteins in PROb cells but theirabsence in HT-29 cells (data not published). These two cell lines weremaintained as monolayers in tissue culture flasks using Ham's F-10

medium supplemented with 10% fetal bovine serum. Cells were detached by sequential treatment with EDTA and trypsin.

Survival Assay. To determine the cytotoxic effect of anthracyclinesa previously described colorimetrie test was used (17). Briefly, tumorcells were cultured for a constant time of 48 h in microtiter 96-wellflat-bottomed plates and the different stages of confluence were obtained by the variation of inoculum (nonconfluence, 10" cells/well;subconfluence, 5 x 10" cells/well, confluence, 15 x IO4cells/well). Inthe case of HT-29 cells, it was possible to obtain an additional stage ofhyperconfluence after 8 days of culture (15 x IO4 cells/well) without

cell detachment. Then, cells were treated with drugs for 1 h. Afterrinsing, cells were cultured for 72 h in complete medium. In someexperiments, tumor cells at different stages of confluence in cultureflasks were detached by trypsin and suspended for l h in anthracyclinesolutions at a same drug:cell ratio (5 x IO5cells/ml) whatever was the

previous confluence status of the cells. After washing, cells were seeded(IO4 cells/well) in microtiter 96-well flat-bottomed plates and cultured

for 72 h in complete medium. In all cases, after the posttreatmentincubation, cells were rinsed twice in PBS in order to remove non-adherent dead cells. Then, cells were fixed for 10 min in absolute ethylalcohol and stained with 1% méthylèneblue in 0.01 M borate buffer,pH 8.5. After abundant rinsings, the dye bound to residual cells waseluted with 0.1 N hydrochloric acid and its absorbance was measuredon an automatic photometer (Multiskan, Flow Laboratories, Irvine,United Kingdom) equipped with a 630 nm filter. Absorbance of theeluted dye was demonstrated to be proportional to the number ofresidual target cells. Each determination issued from a quadruplicate.Results were expressed as

Cell survival = Mean absorbance in treated wellsMean absorbance in control wells

x 100

6626

Research. on September 11, 2020. © 1990 American Association for Cancercancerres.aacrjournals.org Downloaded from

RESISTANCE TO ANTHRACYCLINES OF CONFLUENT COLON CANCER CELLS

Microspectrofluorometer. Fluorescence emission spectra from an in-tracellular microvolume were recorded with a prototype microspectro-fluorometer obtained from a modified OMARS 89 Raman Spectrometer (Dilor, Lille, France) as already described (18, 19). By means of aconventional optical microscope (Olympus BH2) equipped with a x 100water phase contrast immersion objective (Olympus), a laser beam wasfocused on a spot less than 1 ^m in diameter. Sample observation andcollection of fluorescence emission were obtained through the sameoptics. The actual fluorescence sampling was restricted by means of asuitable pinhole diaphragm on the image plane of the microscopeobjective. The emission light signal, spectrally dispersed with a diffraction grating, was detected with an optical multichannel analyzer, madeof a cooled 512-diode array, optically coupled to an image intensifier.Data were collected locally and then transferred to a Goupil G4computer for analysis with a specifically developed program (18). Atleast 20 spectra from the same intracellular location were accumulatedin order to obtain a signalmoise ratio of about 30. A Rhodamine Bsolution was used as a fluorescence standard in order to control laserpower and instrumental response day by day and to permit quantitativecomparisons between spectra recorded on different days. Sample heating, photobleaching, and photodamage were checked empirically andfound to be negligible under our experimental conditions:a 4-/iW laserpower on the sample; and an illumination time of 0.5-1 s. In particular,cells always remained viable after repeated fluorescence determinations,as checked by phase contrast microscopy.

Determination of Nuclear Anthracyclines Concentrations in LivingCells. For anthracyclines uptake studies, cells cultured on glass coverslides were incubated in Ham's F-10 medium containing the appropri

ate drug concentration for 1 h. Then, they were washed free of drug incooled PBS at 4°Cand the glass cover slide was placed in a Petri dish

containing PBS. The fluorescence emission arising from the nucleus ofa cell treated with DXR or deoDXR can be expressed as a sum of thespectral contributions of free drug, DNA-bound drug and intranuclearauto-fluorescence. By studies in aqueous solutions, we showed in arecent paper (18), that each of these contributions has a characteristicspectral shape and that the fluorescence yield in the free form is 48times higher than that of the DNA-bound form. Thus, the drug concentration in the living cell nucleus was obtained from the determinedsurface spectral contributions and by means of the correspondingfluorescence yields (18). Intranuclear concentrations of drugs wereexpressed in //molar.

Radioactive Molecule Cell Uptake Measurements. Cells were trypsin-ized and adjusted to 5 x IO6 cells/ml. In a tube, 200 «¿1of the cellsuspension were mixed either with 70 n\ of a Na^'CKX solution (1mCi/ml; CEA, Gif/Yvette. France) or 40 p\ of a D-[uC]alanine solution(50 MCi/ml; Amersham, Buckinghamshire, England) or 50 ^1 of a L-['"'CJglucose solution (200 /iCi/ml; Amersham). After a 1-h incubationat 37°C,cells were washed and adjusted to IO6cells/ml. The radioac

tivity of aliquots of 100 n\ of the cell suspensions was measured througha gamma counter (1 KB 1272 Clini Gamma, Stockholm, Sweden) orthrough a beta counter (LKB 1214 Rackbeta). Each result was themean of eight determinations.

RESULTS

Sensitivity to Anthracyclines of HT-29 and PROb CellsTreated at Different Degrees of Confluence. The cell densitiescorresponding to the different degrees of confluence of bothHT-29 (human) and PROb (rat) colon cancer cell monolayercultures were: nonconfluence, 309 ±94 cells/mm2; subconflu-ence, 1805 ±206 cells/mm2; confluence, 3478 ±627 cells/mm2; hyperconfluence, 7000 ±851 cells/mm2. A general observation was that sensitivity of HT-29 and PROb cells to DXRand deoDXR progressively decreased with the augmentation ofcell confluence (Fig. 1). The different cell densities from non-confluence to confluence were obtained by different inocula ofcancer cells cultured for a constant time. In that way, we couldascertain that this resistance was independent of the duration

HT-29

PROb

DXR deoDXRFig. 1. Survival of HT-29 and PROb cells treated, at different degrees of

confluence, with DXR or deoDXR (1 h, 37'C). AT, nonconfluent; SC. subcon-

fluent; C. confluent: HC. hyperconfluent. Similar results were obtained in threeindependent experiments.

of culture but only linked to confluence. Thus, in the course ofthe following study, we often worked with two extreme celldensities of nonconfluence and hyperconfluence in HT-29 cells,this latter cell density being obtained after 8 days of culture.

Since a reduction in glucose or oxygen supply or a low pHwere described as causes of anthracycline resistance ( 12, 20) wechecked that similar results were obtained when medium wasrenewed every 12 h during the 48 h before treatment (data notshown).

Both cell lines were more sensitive to deoDXR than to DXR,but the confluence effect was potent with both anthracyclines.However, the cytotoxicity of deoDXR only clearly decreasedwith the higher level of confluence obtained with each cell line.Thus, the importance of anthracycline resistance arising withconfluence appeared to be modulated by the efficacy of thecytotoxic drug.

It was previously demonstrated that PROb cells resist anthracyclines in part through an active drug efflux pump (5)identified as a p 180 glycoprotein whereas HT-29 cells did notpossess this mechanism of resistance. When PROb cells weretreated in presence of amiodarone, an inhibitor of this pump,the cytotoxic effects of DXR and of deoDXR were increased(Fig. 2). A maximal effect of amiodarone was obtained at adose of 5 ng/ml. However, even in presence of amiodarone, theconfluence effect on PROb cell sensitivity to DXR and deoDXRpersisted. For HT-29 cells, amiodarone had no effect on thecytotoxicity of the two anthracyclines at any stage of cellconfluence (data not shown). Thus, the anthracycline resistanceoccurring with the augmentation of confluence of the two celllines appeared independent of an active drug efflux mechanismsince this phenomenon was observed when the efflux was inhibited or did not exist. We have further used the term CDR forconfluence dependent resistance. In all the subsequent experiments, we treated PROb cells in the presence of amiodaronewhereas HT-29 cells were treated in the absence of this effluxinhibitor.

After trypsinization, the CDR of hyperconfluent HT-29 cellsto anthracyclines was stable at least up to 6 h but had disappeared within 24 h (Fig. 3).

Nuclear Concentrations of Anthracyclines in HT-29 and PRObCells Treated at Different Degrees of Confluence. The nuclearaccumulation of DXR or deoDXR progressively decreased withthe augmentation of confluence for both HT-29 cells and PROb

6627

Research. on September 11, 2020. © 1990 American Association for Cancercancerres.aacrjournals.org Downloaded from

RESISTANCE TO ANTHRACVCLINES OF CONFLUENT COLON CANCER CELLS

(pg/ml)

ILIO 20

10 ()jg/ml)

deoDXRFig. 2. Survival of nonconfluent (•)and confluent (•)PROb cells treated

with DXR or deoDXR ( 1 h. 37"C) in the presence of 5 ..i; ml amiodarone ( )or not (—).

cells (Fig. 4). In the particular case of PROb cells, treated with10 Mg/ml of deoDXR in presence of amiodarone, the nucleardrug concentrations were not significantly different betweennonconfluent and confluent cells, suggesting that saturation ofDNA was reached under these conditions of treatment. Thecytotoxicity under these conditions was actually complete forboth nonconfluent and confluent PROb cells (Fig. 2).

We observed a high discrepancy between the different stagesof confluence concerning the relation between the level ofanthracycline nuclear accumulation and the cytotoxicity induced. For example, for HT-29 cells, DXR at a dose of 1 jig/ml led to a nuclear accumulation of about 50 /¿Mand a cellmortality of 50% for nonconfluent cells whereas for hypercon-fluent cells, treated with 10 Mg/ml DXR, a nuclear accumulationof 230 MMinduced a cell mortality of only 5%. Thus, a nearly5-fold higher DXR nuclear accumulation in hyperconfluentHT-29 cells than in nonconfluent ones induced a smaller cyto-

toxic effect. These results clearly demonstrated that CDR toanthracyclines had two causes: (a) a decreased accumulation ofdrug in the nucleus of confluent cells; and (b) a nuclear resistance of the confluent cell to the cytotoxic effect of the drug.

For PROb cells, similar ratios of nuclear anthracycline concentrations in nonconfluent cells to those in confluent cellswere obtained whether they were treated in presence of amiodarone or not (Table 1). These results confirmed that CDR wasindependent of an active drug efflux mechanism.

We wondered whether anthracycline influx in confluent cellscould be altered by the reduction of cell surface in contact withthe drug and/or by a reduced drug:cell ratio. When HT-29 cellsof different confluence degrees were trypsinized adjusted to

100

80-

60-

40-

20-

0

1 h

o 10 12

100

80-

60-

40

20-

0

6 h

12

luU

100

80 -

60 -

40-

20 •

24 h

8 10 DXR(ug/ml)

Fig. 3. Reversibility of the CDR to DXR of hyperconfluent HT-29 cells.Nonconfluent (A'O and hyperconfluent (HC~)cells were trypsinized and incubatedat 37'C in identical conditions of cell density and culture. One. 6. and 24 h aftertheir trypsinization they were detached, treated in suspension (1 h. 37°C),washed,and seeded under identical conditions. Cell survival was assessed after a 72-hincubation (37'C).

similar cell densities, and immediately treated with anthracyclines in similar conditions, it still appeared that former hyperconfluent cells were more resistant to anthracyclines than former nonconfluent cells (Fig. 5A). We observed for deoDXR,but not for DXR, that HT-29 cells treated in suspension hadan altered capacity to adhere again during the survival assay.Thus, we had to reduce the deoDXR concentrations in thetreatment of cells in suspension, as compared to that of cells inmonolayer. The nuclear concentrations of DXR or deoDXR informer nonconfluent cells were higher than in former hyperconfluent cells (Fig. 5B) as observed with cell monolayers (Fig.4). Thus, the reduced nuclear anthracycline accumulation inconfluent cells may likely be related to a change of the confluentcell compared with the nonconfluent one.

Passive Diffusion Analysis. It is now admitted that anthracyclines passively diffuse across the cell membrane as neutralcompounds (21). We studied the confluence effect on the celluptake of three radioactive molecules known to cross the membrane independently of active transport mechanisms. Thesemolecules were Na25'CrO4 and n-['4C]alanine and L-['4C]glu-

cose, the two isomers of the natural forms of these components.The passive diffusion of these molecules was checked by linear

6628

Research. on September 11, 2020. © 1990 American Association for Cancercancerres.aacrjournals.org Downloaded from

RESISTANCE TO ANTHRACYCLINES OF CONFLUENT COLON CANCER CELLS

HT-29

PROb

g< 800oc

01 600O

ff<ai_U

200

Z2 800

li)O

ouce

O

z

10 DXR (Mg/ml) deoDXR (pg ml)

DXR (|jg mi) + amiodarone 5 ug/ml1 10

deoDXR (pg mi) + amiodarone 5 jjg mi

DXR deoDXRFig. 4. DXR and deoDXR uptake in the nucleus of HT-29 and PROb cells treated (l h, 37'C) at different degrees of confluence: nonconfluent (D); confluent (D);

hyperconfluent (•).Drug concentration in the nucleus was quantified by microspectrofluorometry.

Table 1 Nuclear concentrations (¡IM)of DXR in nonconfluent (NC) and confluent(C) PROb cells treated for l h in the presence or not of amiodarone

CellconfluenceNCCNCCDXR

doses"(jig/ml)(5

Mg/ml) 19.1±2*2.6

±1(3.5)'+

22±4+5.6 ±2(3.9)1056

±9I7±5(3.3)230

±7069±21(3.3)

" Nuclear drug concentrations were determined by microspectrofluorometry

immediately after treatment.* Mean ±SD.f Numbers in parentheses, ratio of the nuclear drug concentration in non

confluent cells to that in confluent cells.

timed uptake curves (data not shown). Their uptake by HT-29cells, trypsinized and incubated for l h with these molecules,was in all cases lower when cells had been previously hyperconfluent than when cells had been previously nonconfluent (Fig.6). With Na25lCrO4, we demonstrated that the decreased uptake

induced by confluence was stable at least up to 6 h aftertrypsinization but disappeared after 24 h (Fig. 7). This kineticsof reversion was similar to that of CDR to anthracyclines.

These results led us to suggest that parameters generallyconditioning passive diffusion at the cell level were modifiedwith the increase of confluence, affecting the passive diffusionof anthracyclines.

DISCUSSION

Using two colon cancer cell lines, cultured as monolayers,this study clearly established a progressive development of

resistance to anthracyclines when confluence increased. Wepartly related this phenomenon to a decreased accumulation ofanthracyclines in the nucleus as measured at this site by microspectrofluorometry, a recently developed method (18).

The best-known mechanism of resistance to anthracyclinesin cancer cells is an active drug efflux pump preventing anthracyclines from accumulating in their nuclei and conferring onthem, at least partly, the multidrug resistance (MDR) pheno-type (1-3). The use of an inhibitor of this efflux pump, amiodarone (5), did not suppress the confluence dependent resistance (CDR) and demonstrated that this phenomenon was independent of the MDR mechanism. We supposed that thereduced drug accumulation in nuclei of confluent cells could bedue to an alteration of the anthracycline uptake by cells. Itcould not be attributed to differences in drug:cell ratio amongnonconfluent, confluent, and hyperconfluent monolayers whichcould induce differences in effective drug concentrations delivered per cell. Indeed, when nonconfluent or hyperconfluentHT-29 cells were trypsinized and immediately treated in suspension at a same drug:cell ratio, the same difference of nucleardrug accumulation was observed as when cells were treated asmonolayers. In the same way, the reduction of DXR or deoDXRuptake on account of a decreased membrane surface in contactwith the drug in confluent cells was excluded by our resultssince the detachment by trypsin of hyperconfluent HT-29 cellsbefore treatment did not reverse either their low drug nuclearaccumulation or their resistance. These results demonstratedthat the decreased anthracycline nuclear accumulation in confluent cells was not artifactual but linked to a modification ofthe confluent cell compared with the nonconfluent one. Becauseanthracyclines passively diffuse across the cell membrane (21),

6629

Research. on September 11, 2020. © 1990 American Association for Cancercancerres.aacrjournals.org Downloaded from

RESISTANCE TO ANTHRACYCLINES OF CONFLUENT COLON CANCER CELLS

A)

100<

E 80-

I 60-1

OC

W 40-

HO 20-

B)10 (ug/ml) 0,0 0,1 0,2 0,3 0,4 0,5 (pg/ml)

200

(ug/ml) 0.5 (ug/ml)

DXR deoDXRFig. 5. Survival (A) and drug uptake (B) of HT-29 cells treated in suspension, but formerly nonconfluent (NC) or hyperconfluent (HC). Cells were trypsinized and

adjusted to identical cell densities. They were immediately exposed to drug for l h at 37T, washed, and seeded under identical conditions. Cell survival was assessedafter a 72-h incubation (37'C). Drug uptakes were measured immediately after treatment, by microspectrofluorometry. Bars, SD.

»8

10000 -,

8000

6000

^ 4000 -

u2000-

8000l

D-ALA Na2Cr04 L-Gluc

( 90 ) ( 161 ) ( 182 ) ( MW )Fig. 6. Cell uptake of D-[MC]alanine, (D-ALA). sodium [!1Cr]chromate, and L-

|"C]glucose (L-Gluc) by nonconfluent (D) and confluent (D) HT-29 cells. Immediately after their trypsinization. cells were incubated for l h with the radioactivemolecules (see "Materials and Methods"). Then, cells were washed and radioactivity was measured from aliquots of 10s cells. Each result corresponds to eightaliquots. Bars, SD. MW, molecular-weight in thousands.

we checked the effect of confluence on the intake of nonmeta-bolizable molecules known to penetrate into the cell by passivediffusion because of the absence of a specific transport mechanism (sodium (5'Cr]chromate) or of the stereospecificity ofexisting transport mechanisms (D-['4C]alanine; L-[MC]glucose).

Since these three molecules were less uptaken by confluent thanby nonconfluent HT-29 cells, we concluded that parametersgenerally conditioning passive diffusion were altered with con-

6 24 hoursFig. 7. Sodium ["Cr]chromate uptake by nonconfluent (D), confluent (D), and

hyperconfluent (•)HT-29 cells at different times after their trypsinization. Cellswere trypsinized and adjusted to identical cell densities. They were incubated at37"C under similar conditions of culture. After 1. 6, or 24 h they were detachedand incubated for l h with Na251CrO4 (350 ^Ci/5 x IO6 cells/ml). Cells werewashed, and then radioactivity was measured from aliquots of 10* cells. Each

result corresponds to eight aliquots. Bars, SD.

fluence. Consequently, a reduced passive diffusion of anthra-

cyclines across the membranes of confluent cells could explainthe low nuclear drug accumulation in these cells and thus, atleast partly, the CDR. The observation that the confluencerelated decrease in Na251CrO4uptake and that the anthracycline

CDR had strictly similar kinetics of reversion, reinforced apossible link between the level of sensitivity to anthracyclines

6630

Research. on September 11, 2020. © 1990 American Association for Cancercancerres.aacrjournals.org Downloaded from

RESISTANCE TO ANTHRACYCLINES OF CONFLUENT COLON CANCER CELLS

of the cell and the general level of passive diffusion across itsmembrane. Experiments to determine whether confluence induced modifications could be related to membrane fluidity andthus to membrane composition are in progress.

However, the reduction of anthracycline passive diffusionand the low nuclear accumulation of these drugs in confluentcells cannot entirely explain CDR. Indeed, we observed that,for a similar nuclear anthracycline accumulation, confluent cellsappeared to be more resistant than nonconfluent cells. Thus,confluence appeared to induce nuclear changes in addition tothe membrane modifications suggested above.

Confluence, and hyperconfluence even more, corresponds tothe stationary phase of growth of PROb and HT-29 cells. Adecreased efficacy of anthracyclines on tumor cells in a stationary phase of growth compared with exponentially growing cellshave been widely reported for monolayer cultures (8-12) as wellas for suspension cultures (22). The most common assumptionin these previous studies to explain the resistance of stationaryphase cells was their noncycling state. For 4'-(acridinylam-

ino)methanesulfon-w-anisidide, another intercalating agent, itwas recently demonstrated that the resistance of noncyclingcells could lie in an alteration of the drug-DNA topoisomeraseII interaction (23, 24). In cultured human cells an importantdecrease of the intracellular DNA topoisomerase II level whencell density increased was demonstrated and primarily relatedto Go-G, (25). In our study, 83% of confluent and 69% ofnonconfluent HT-29 cells were accumulated in G,,-G] as observed by flow cytometry (data not shown). This difference incell cycle repartition was significant even weak. Whether amembrane modification and/or a defect of the nuclear target ofanthracyclines (DNA topoisomerase II) could be linked to thisincrease of cells in G,,-Gi phase has to be checked. However, a1-h exposure to 10 ng/m\ of DXR or deoDXR killed more than80% of nonconfluent HT-29 cells, suggesting that these drugscould be cytotoxic even for cells in G0-G, if they were nonconfluent. At the same concentration, the two anthracyclines killedno more than 20% of hyperconfluent HT-29 cells. Moreover,microspectrofluorometry experiments revealed that the difference of anthracycline nuclear accumulation between the different cell densities concerned the totality of the cell populationand not a subpopulation as expected if it was proliferationdependent. Indeed, as an example, we observed for nonconfluent and hyperconfluent HT-29 cells a correct homogeneityof measures (from 30 cells) inasmuch as the standard deviationswere, respectively, 11 and 22% of the mean level of deoDXRnuclear concentrations. Thus, these results prevent us to strictlyrelate the CDR to a lack of cell cycle progression.

As a conclusion, the anthracycline CDR phenomenon, demonstrated in this work from two colon cancer lines in vitro,could be of first importance to explain the high resistance ofcolon tumors to these drugs. Indeed, CDR could be the mainfactor leading cancer cells to resist anthracyclines when theyare tightly aggregated as in solid tumor.

ACKNOWLEDGMENTS

The authors would like to thank Anne Biais for her valuable andhelpful advice, Denis Bosson for revising the manuscript, and MarieFrance Michel for her secretarial assistance.

REFERENCES

1. Daño,K. Active outward transport of daunomycin in resistant Ehrlich ascitestumor cells. Biochim. Biophys. Acta, 323: 466-483, 1973.

2. Skovsgaard, T. Mechanisms of cross-resistance between vincristine and dau-norubicin in Ehrlich ascites tumor cells. Cancer Res.. 38: 4722-4727. 1978.

3. Inaba. M.. Kobayashi. H.. Sakurai. Y.. and Johnson. R. K. Active efflux ofdaunomycin in sensitive and resistant sublines of P388 leukemia. CancerRes.. 39: 2200-2203. 1979.

4. Tsuruo, T.. Lida. H.. Tsukagoshi. S.. and Sakurai. V. Potentiation of vincristine and Adriamycin effects in human hematopoietic tumor cell lines bycalcium antagonists and calmodulin inhibitors. Cancer Res.. 43: 2267-2272,1983.

5. Chauffert, B., Martin, M.. Hammann. A., Michel, M. F., and Martin, F.Amiodarone-induced enhancement of doxorubicin and 4'-deoxydoxorubicin

cytotoxicity to rat colon cancer cells in vitro and in vivo. Cancer Res., 46:825-830, 1986.

6. Kerr. D. J.. and Kaye. S. B. Aspects of cytotoxic drug penetration, withparticular reference to anthracyclines. Cancer Chemother. Pharmacol., 19:1-5, 1987.

7. Sutherland, R. M. Cell and environmental interactions in tumor microre-gions: the multiceli spheroid model. Science (Washington DC), 240: 177-184, 1988.

8. Barranco, S. C., and Novak. J. K. Survival responses of dividing and nondi-viding mammalian cells after treatment with hydroxyurea. arabinosylcyto-sine, or Adriamycin. Cancer Res., 34: 1616-1618, 1974.

9. Bhuyan, B. K., Fraser, T. J.. and Day. K. J. Cell proliferation kinetics anddrug sensitivity of exponential and stationary' populations of cultured LI 210cells. Cancer Res., 37: 1057-1063, 1977.

10. Epifanova. O. I., Smolcnskaya. I. N., and Polunovsky, V. A. Responses ofproliferating Chinese hamster cells to cytotoxic agents. Br. J. Cancer. 37:377-385, 1978.

11. Drewinko. B.. Patchen. M.. Yang, L. Y., and Barlogie. B. Differential killingefficacy of twenty antitumor drugs on proliferating and non proliferatinghuman tumor cells. Cancer Res.. 41: 2328-2333, 1981.

12. Born. R., and Eichholtz-Wirth, H. Effect of different physiological conditionson the action of Adriamycin on Chinese hamster cells in vitro. Br. J. Cancer,•«.•241-246.1981.

13. Soranzo, C., and Ingrosso, A. A comparative study of the effects of anthracycline derivatives on a human adenocarcinoma cell line (LoVo) grown as amonolayer and as spheroids. Anticancer Res., 8: 369-373, 1988.

14. Fogh. J., and Trempe, G. New human tumor cell lines. In: J. Fogh (ed.).Human Tumor Cells in Vitro, pp. 115-140. New York: Plenum PublishingCorp., 1975.

15. Martin. F.. Caignard, A., Jeannin, J. F., Ledere, A., and Martin. M. S.Selection by trypsin of two sublines of rat colon cancer cells forming progressive or regressive tumors. Int. J. Cancer, 32: 623-627, 1983.

16. Chauffert, B.. Martin, F.. Caignard. A., Jeannin. J. F., and Ledere, A.Cytofluorescence localization of Adriamycin in resistant colon cancer cells.Cancer Chemother. Pharmacol., 13: 14-18. 1984.

17. Martin. F.. Caignard. A., Olsson. O., Jeannin. J. F.. and Ledere. A. Tumor-icidal effect of macrophages exposed to Adriamycin in vivoor in vitro. CancerRes.. «.-3851-3857. 1982.

18. Gigli. M.. Doglia. S. M.. Millot, J. M., Valentin!. L., and Manfait, M.Quantitative study of doxorubicin in living cell nuclei by microspectrofluorometry'. Biochim. Biophys. Acta. 950: 13-20. 1988.

19. Ginot. L.. Jeannesson, P., Angiboust, J. F., Jardillier, J. C.. and Manfait, M.Interactions of Adriamycin in sensitive and resistant leukemic cells: a comparative study by microspectrofluorometry. Studia Biophys., 104: 145-153,1984.

20. Smith. E., Stratford, I. J., and Adams. G. E. Cytotoxicity of Adriamycin onaerobic and hypoxic Chinese hamster V79 cells in vitro. Br. J. Cancer, 41:568-573. 1980.

21. Siegfried, J. M., Burke, T. G., and Tritton, T. R. Cellular transport ofanthracyclines by passive diffusion: implications for drug resistance.Biochem. Pharmacol.. 34: 593-598. 1985.

22. Ohnuma, T., Arkin, H., and Holland. J. F. Effects of cell density on drug-induced cell kill kinetics in vitro (inoculum effect). Br. J. Cancer. 54: 415-421, 1986.

23. Robbie, M. A., Baguley, B. C., Denny, W. A., Gavin, J. B., and Wilson, W.R. Mechanism of resistance of noncycling mammalian cells to 4'-(9-acridi-nylamino)mcthancsulfon-/n-anisidide: comparison of uptake, metabolism,and DNA breakage in log- and plateau-phase Chinese hamster fibroblast cellcultures. Cancer Res., •**:310-319, 1988.

24. Schneider. E., Darkin. S. J., Robbie. M. A., Wilson. W. R.. and Ralph. R.K. Mechanism of resistance of non-cycling mammalian cells to 4'-(9-acridi-nylamino)methanesulphon-m-anisidide: role of DNA topoisomerase II inlog- and plateau-phase CHO cells. Biochim. Biophys. Acta, 949: 264-272,1988.

25. Hsiang, Y. H.. Wu. H. Y., and Liu, L. F. Proliferation-dependent regulationof DNA topoisomerase II in cultured human cells. Cancer Res.. 48: 3230-3235. 1988.

6631

Research. on September 11, 2020. © 1990 American Association for Cancercancerres.aacrjournals.org Downloaded from

1990;50:6626-6631. Cancer Res   Hélène Pelletier, Jean-Marc Millot, Bruno Chauffert, et al.   DiffusionCancer Cells to Anthracyclines: Alteration of Drug Passive Mechanisms of Resistance of Confluent Human and Rat Colon

  Updated version

  http://cancerres.aacrjournals.org/content/50/20/6626

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/50/20/6626To request permission to re-use all or part of this article, use this link

Research. on September 11, 2020. © 1990 American Association for Cancercancerres.aacrjournals.org Downloaded from