[CANCER RESEARCH 41. 82-88, January 1981]0008-5472/81/0041-OOOOS02.00

Effects of Serum and Insulin on the Sensitivity of the Human BreastCancer Cell Line MCF-7 to Estrogen and Antiestrogens1

W. Barkley Butler,2 William H. Kelsey, and Nicolene Goran

Michigan Cancer Foundation, Department of Chemistry, Detroit. Michigan 48201

ABSTRACT

The effect of the antiestrogens tamoxifen and nafoxidine onthe growth of the human breast cancer cell line MCF-7 ismodified by both serum and insulin. Tamoxifen inhibition of thegrowth of MCF-7 cells in culture is reduced as the concentra

tion of serum in the medium is increased from 0.1% to 5 to10%. Estradiol does not stimulate cell growth over the samerange of serum levels.

Insulin changes the sensitivity of MCF-7 cells to both estrogen and antiestrogens. Cells growing in media containing insulin are less sensitive to inhibition by either tamoxifen ornafoxidine than are cells growing in its absence. In addition,higher concentrations of estradiol are required to stimulate theproduction of plasminogen activator when cells are grown inmedia containing insulin.

This effect of insulin can be accounted for by the finding thatinsulin lowers the level of estrogen receptor in MCF-7 cellswithout altering the binding constant for the hormone. Cellsgrown with insulin have an average of 21,000 ±4,700 (S.D.)estrogen binding sites/cell compared to 62,000 ± 9,700sites/cell in cells grown in the absence of insulin. This difference in receptor level is sufficient to account for the differencein the concentration of estradiol needed for equivalent induction of plasminogen activator in cultures with or without insulin.

These results indicate that the level of estrogen receptor inbreast cancer cells can be changed and that the sensitivity ofsuch cells, both to estrogen and to antiestrogens, is altered bychanges in the level of estrogen receptor. They also haveimplications concerning the mechanism by which antiestrogensact to inhibit the growth of mammary tumor cells.

INTRODUCTION

The role of estrogen in the development and growth ofmammary carcinomas in humans and animals has been studiedin a variety of systems. Some recent work has also focused onthe role that the peptide hormone insulin may play, not only inaffecting the growth of such tumors, but also in modifying theirresponsiveness to estrogen (11, 15, 32).

The human breast cancer cell line MCF-7 was derived fromthe pleural effusion of a patient with a metastatic mammarycarcinoma (33). In recent years, it has been used extensivelyas a model system for studying the effects of various hormoneson breast cancer (29). It is well suited to these studies since itcontains receptors for several hormones including estrogen(7), androgens, glucocorticoids, progesterone (17), and insulin(30).

' Supported by Grant CA-21606 from the National Cancer Institute and by an

institutional grant from the United Foundation of Greater Detroit.2 To whom requests for reprints should be addressed.

Received April 27, 1979; accepted September 29, 1980.

The antiestrogens tamoxifen and nafoxidine inhibit thegrowth of MCF-7 cells, leading eventually to cell death (19).We have noted that the inhibitory effect of tamoxifen is modifiedby serum and also by insulin (9) and in this paper show thatinsulin alters the level of estrogen receptor in MCF-7 cells in away which can account for their reduced sensitivity to tamoxifen and to estrogen.

MATERIALS AND METHODS

Cell Culture. The human breast cancer cell line MCF-7 (33)was obtained from Dr. H. D. Soule. Cells were grown in Falconplastic T-flasks (25 and 75 sq cm) at 37°in Eagle's minimumessential medium, Hanks' base, supplemented with nonessen-

tial amino acids and L-glutamine from KC Biological, Inc.,Lenexa, Kans. The medium contained 10% calf serum (KCBiological, Inc.), 12.5 /ig bovine insulin per ml (Sigma ChemicalCo., St. Louis, Mo.), and 100 units penicillin and 100 /ugstreptomycin per ml (Grand Island Biological Co., Grand Island,N. Y.). Cells were passed weekly using trypsiniEDTA (GrandIsland Biological Co.) when near confluence. Cultures were fedon Days 3 and 6 following transfer.

Most of the experiments reported here were performed oncells growing on the bottoms of sterile glass scintillation vials(2) in 2 ml of the same medium with Earle's base in a humidified

atmosphere of 5% CO2:95% air. Serum and insulin were addedas indicated for each experiment. Cells were plated at between50,000 and 100,000 cells/vial. Under these conditions, theyreach confluence at approximately 1,000,000 cells/vial.

The serum used in many of these experiments has beenstripped of endogenous steroids by two 30-min incubations at55°with a pellet of dextran-coated charcoal (1 ) (1 % activated

charcoal; 0.1 % dextran; Sigma Chemical Co.; prepared at one-half of this concentration by stirring overnight at 4°in 0.9%

NaCI and pelleted by centrifugation prior to use). Steroids wereadded as ethanol solutions to a final concentration of not morethan 0.2% ethanol.

To obtain counts of cells growing in vials, the medium wasremoved, and the cells were rinsed gently with 0.9% NaCI.After removing the 0.9% NaCI solution, 0.5 ml of buffer [0.01M 4-(2-hydroxyethyl)-1 -piperazineethanesulfonic acid (pH 7.4):1.5 mM MgCI2] was added followed by one or 2 drops of Zap-Isoton II (Coulter Diagnostics, Inc., Hialeah, Fla.). This lysedthe cells, leaving a suspension of nuclei which were suspendedto 10 ml in 0.9% NaCI:0.5% formalin and counted on a ModelZf Coulter counter. Counts obtained in this way were in goodagreement with those obtained by removing the cells withtrypsinrEDTA and counting on a hemocytometer.

Antiestrogens. Tamoxifen (irans-p-jß-dimethylaminoethox-phenyl-1 -1,2-diphenyl-but-1 -en), I.C.I. 46474, was a gift from

Lois Trench, I.C.I. United States, Inc., Wilmington, Del., andnafoxidine [1 -(2-p-(3,4-dihydro-6-methoxy-2-phenyl-1 -naph-

82 CANCER RESEARCH VOL. 41

on May 17, 2020. © 1981 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Insulin and Antiestrogens in Human Breast Cancer

tyl)phenoxyethyl)pyrrolidine hydrochlorlide], U-11.100A, wasa gift from Dr. Paul O'Connell, The Upjohn Company, Kalama-

zoo, Michigan.Thymidine Incorporation. Cells growing in glass scintillation

vials were labeled as described with [3H]thymidine (New Eng

land Nuclear, Boston, Mass.). At the end of the incubationperiod, the vials were treated for 5 min once with ice-cold 0.9%

NaCI and then 3 times with 5% trichloroacetic acid and oncewith ethanol. Ten ml of counting solution (toluene, PRO, PO-

POP) were added, and the radioactivity was counted in a SearleAnalytic 81 liquid scintillation counter.

Plasminogen Activator Induction and Assay. Cells wereplated in 25-sq cm T-flasks at a density of 500,000 cells/flask

in medium containing 10% calf serum. The next day, themedium was changed to a chemically defined serum-free me

dium described by Higuchi and Robinson (14) but lackinghydrocortisone and thyroxine and containing 5 |iig insulin perml as indicated. Growth in this medium is as rapid as it is inmedium with serum. Cells were grown for 5 days with a mediumchange on Day 2 or 3. They were then approximately 50%confluent and were changed to fresh medium with estradioladded, as indicated, in a final concentration of 0.1% ethanol.Two days later, the medium was collected, centrifuged, andstored frozen for subsequent assay.

Plasminogen activator was assayed by measuring the release of 3H-labeled peptides from [3H]fibrin-coated plates (3).A solution containing 100 /ig of [3H]fibrinogen (200,000 cpm)

was spread on 35-mm plastic Petri dishes and dried for 24 hrat 45°. The fibrinogen was then clotted by a 2-hr treatment

with purified bovine thrombin (plasminogen free; Miles Research Products, Miles Laboratories, Inc., Elkhart, Ind.). Thefinal assay was carried out at 37°for 4 hr in a volume of 1 ml

which contained 1% acid-treated dog serum as a source of

plasminogen. In addition to assaying an aliquot of the cellculture medium, usually 0.1 ml, a series of urokinase standards(0.01 to 0.1 units, Calbiochem-Behring Corp., LaJolla, Calif.)

was assayed, and the resulting standard curve was used tostandardize the assay. The final results are expressed asnumber of urokinase units per 1,000,000 cells after correctingfor the volume of medium assayed and the number of cells inthe flask.

Measurement of Estrogen Binding Capacity. The procedure used for measuring estrogen binding capacity in intactcells was similar to that described by Shafie and Brooks (31 ).We also found that the binding reaches equilibrium by 30 to 60min at 37°and have used an incubation time of 60 min as they

did.

RESULTS

Effects of Serum on the Growth and Tamoxifen Sensitivityof MCF-7 Cells. The sensitivity of MCF-7 cells to the antiestro-gen tamoxifen is altered by several factors. Depending on theconditions used, cells may be relatively unaffected, grow at areduced rate, or die when incubated with tamoxifen. One factoraffecting the sensitivity to tamoxifen is the concentration ofserum in the medium. MCF-7 cells grow well at serum concen

trations between 0.5% and 10%. At lower concentrations ofserum, they grow at a reduced rate and to a lower density(Chart 1A). Their sensitivity to 10~6 M tamoxifen is much greater

in media containing lower concentrations of serum (Chart 15).

_o>X

o>O

IO

Time (days)Chart 1. Effect of serum concentration on the growth and sensitivity to

tamoxifen of MCF-7 cells. Cells were plated in glass scintillation vials in mediumcontaining 2% charcoal-extracted calf serum and insulin (12.5 /ig/ml); 50.000cells were plated per vial. After 24 hr, the medium was changed to mediumcontaining insulin and the indicated concentrations of charcoal-treated calf serum(Day 0). Those in A received 0.1% ethanol and those in B an ethanol solution oftamoxifen at a final concentration of 10~6 M tamoxifen and 0.1% ethanol. All

counts were taken from duplicate vials on the days indicated, and the remainingvials were changed to fresh media at the same time.

Growth is prevented by tamoxifen at serum concentrationsbelow 1% and inhibited at higher serum concentrations. As hasbeen reported by others (24, 29), we have also seen thattamoxifen inhibition can be overcome by estradiol.

It has been reported that the growth of MCF-7 cells is

stimulated by estradiol (24, 29). While we have consistentlybeen able to demonstrate growth inhibition of MCF-7 cells by

tamoxifen, we have seldom seen a significant stimulation ofgrowth by estradiol. Estradiol did not stimulate growth at anyconcentration of serum tested between 0.1% and 10%, although there was some inhibition at low serum levels (Chart26). In an attempt to get a consistent and substantial stimulationof growth with estradiol, we have varied the concentration ofestradiol and the concentration, type, and treatment of serum.We have also used MCF-7 cells which had been grown first for

several passages in medium with estradiol in order to enrichthe culture for cells with a requirement for estradiol. Finally,since MCF-7 is an uncloned population containing cells of a

variety of phenotypes based on appearance (36), we selectedseveral clones of various types at random and tested them.None of the clones tested was more responsive than thegeneral culture. Others have also reported that growth of theirMCF-7 cultures was not stimulated by estrogen3 (16, 18, 23).

For experiments with tamoxifen, we have used both calf andfetal calf serum either untreated or charcoal treated. Whileresults differ from lot to lot, they follow the same pattern. Thesensitivity to tamoxifen always decreases with increasing serum concentration, and charcoal treatment of the serum usuallyresults in somewhat greater sensitivity to tamoxifen.

Effects of Insulin on Tamoxifen Inhibition. At serum levelsabove 1%, the addition of insulin reduces the inhibition pro-

3 H. D. Soule and C. M. McGrath Estrogen responsive proliferation of clonal

human breast carcinoma cells in athymic mice. 11 th Meeting on MammaryCancer in Experimental Animals and Man, 1978.

JANUARY 1981 83

on May 17, 2020. © 1981 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

W. B. Butler et al.

Time (days)Chart 2. Effect of serum concentration on the growth of MCF-7 cells ±

estradiol. The procedure was the same as that described in Chart 1. No insulinwas used, and the serum was fetal calf serum.

A- -Insulin

io6

_o>

UIO9

IO*

B4 Insulin

% Strum Control Tamoxifen

° 0.25** OS *•30 *»100 *

-i -

Time (days)Chart 3. Effect of insulin on the sensitivity of MCF-7 cells to tamoxifen. The

procedure was the same as that described in Chart 1. Charcoal-treated fetal calfserum was used. A, no insulin ±1CT6 M tamoxifen; ß,12.5 /ig insulin per ml ±10~6 M tamoxifen.

duced by 10 6 M tamoxifen (Chart 3). Most of the following

experiments were carried out using 2% charcoal-treated serum. Cells grown in this medium with 10~6 M tamoxifen die

after 1 to 2 weeks, as is evident in Chart 3/4 where charcoal-treated fetal calf serum was used. The addition of insulinprevents this and allows the cells to continue growing. Thesame result is seen with another antiestrogen, nafoxidine.Insulin decreases the inhibition due to 10~6 M nafoxidine (Chart

4) and prevents the eventual death of the cells as describedabove for tamoxifen. The effect of insulin on tamoxifen inhibition of thymidine incorporation is shown in Chart 5. Thymidineincorporation may not be equivalent to growth, since estradiolhas been shown to stimulate and tamoxifen to inhibit thymidinekinase (29). However, incorporation of thymidine is inhibited

D'>

01O

-Insulin•fInsulin-Insulin +Tamoxifen+lnsulin t Tamoxifen

o -Insulin•+lnsulin»-Insulin +Nofoxidine»tlnsulin* Nafoxidine

TIME (days)Chart 4. Effect of insulin on the sensitivity of MCF-7 cells to tamoxifen and

nafoxidine. The procedure was the same as that described in Chart 1 except thatcells were plated directly in the test media. Charcoal-extracted calf serum wasused. Insulin was used at 12.5/ig/ml and tamoxifen or nafoxidine at 10~6 M.

IOO

O>-o- 80

60

O) •*-C O

to' IO" IO"

Tamoxifen Concentration (M)io

Chart 5. Effect of insulin on the inhibition of [3H]thymidine incorporation by

tamoxifen. Cells were grown in vials for 5 days in 2% charcoal-treated calfserum ±12.5 /ig insulin per ml. The medium was changed on Day 3. Tamoxifenat the indicated concentrations was then added to the vials in fresh medium ±insulin, and 2 days later, the cultures were labeled for 1 hr with [3H]thymidine

(0.2 /iCi/ml) and the incorporation into trichloroacetic acid-insoluble materialwas measured. Point, average of triplicate samples; bars, S.D.

by tamoxifen. As shown in Chart 5, the effect of insulin is notto make the cells completely resistant to tamoxifen but, rather,to increase the level of tamoxifen required to give the samedegree of inhibition.

Effect of Insulin on the Induction of Plasminogen Activatorby Estrogen. While we have been unable to demonstrate much,if any, stimulation of the growth of MCF-7 cells by estradiol, we

have shown that estradiol stimulates the production of theserine protease plasminogen activator in a chemically definedserum-free medium4 (10). The stimulatory effect of estradiol on

4 W. B. Butler, W. L. Kirkland and T. L. Jorgensen. Steroid stimulation of

fibrinolytic activity in a human breast cancer cell line (MCF-7). 11th Meeting onMammary Cancer in Experimental Animals and Man, 1978.

84 CANCER RESEARCH VOL. 41

on May 17, 2020. © 1981 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Insulin and Antiestrogens in Human Breast Cancer

04

03

•u

O.I

o Control

•+ Insulin

Estradici Concentration (M)Chart 6. Effect of insulin on the induction of plasminogen activator by

estradici. Estradici at the indicated concentrations was added to cells grown withor without insulin (see "Materials and Methods"). After 2 days, the medium was

collected and assayed for plasminogen activator. Bars, S.D.

the production of this enzyme is also modified by insulin. As isshown in Chart 6, cells grown in the presence of insulin areless sensitive to estradiol than are those grown in its absenceand require 5- to 10-fold higher concentrations of estradiol to

achieve comparable levels of induction. Thus, the sensitivity ofMCF-7 cells to both this stimulatory effect of estradiol and the

inhibitory effects of tamoxifen and nafoxidine is reduced byinsulin.

Effect of Insulin on the Level of Estrogen Receptor in MCF-7 Cells. In an attempt to explain the effect of insulin on thesensitivity of MCF-7 cells to estrogen and antiestrogens, wemeasured the level of estrogen receptor in intact MCF-7 cells

growing attached to the bottom of glass scintillation vials. Weused this assay in order to measure the total binding capacityand binding constant of live cells under conditions similar tothose in which their sensitivity to estrogen and tamoxifen hasbeen measured. Cells were grown in 2% charcoal-treated calfserum plus or minus insulin and then incubated with severalconcentrations of [3H]estradiol for 1 hr at 37° in the same

medium lacking serum. Nonspecific binding was measured ina second set of vials which received a 200-fold excess of

unlabeled estradiol. The results of one such experiment areshown in Chart 7. The receptor levels and binding constantswere determined from Scatchard plots of the specific binding.Cells grown with insulin averaged 21,000 ± 4,700 (S.D.)estradiol binding sites/cell compared to 62,000 ±9,700 sites/cell in cells grown in its absence. MCF-7 cells contain 10 pg

DNA per cell (31), and thus these levels of bound estradiol areequivalent to 3.5 and 10.5 pmol/mg DNA, respectively. Thebinding constants were not significantly different and averaged2.6 ±0.8 x 10~10 M and 2.3 ±1 x 10~10 M for cells grown

in the presence or absence of insulin, respectively. The estrogen receptor levels reported here are in the range reportedpreviously for total binding in MCF-7 cells; from 3 to 4 pmol

(19) to 14 to 18 pmol/mg DNA (31). Thus, cells grown in thepresence of insulin have one-third as much receptor as do

those grown in its absence. This difference is sufficient to

0 02 04 06 08 IO I.2 I.4 I6 IB 20

[5H] Estradici (nM)Chart 7. Binding of [3H)estradiol to MCF-7 cells. See "Materials and Meth

ods" for the procedure. Cells were grown in 2% charcoal-treated calf serum ±

12.5 /ig insulin per ml for 5 to 6 days prior to measuring estradiol binding. Whatis shown is specific binding, obtained by subtracting nonspecific binding fromtotal binding. Bars. S.D.; insert, Scatchard plot of the data from this experiment.

account for their reduced sensitivity to both estrogen andantiestrogens, as is discussed below.

DISCUSSION

In addition to increased growth, a number of other responsesto estrogen have been reported for MCF-7 cells. We have

shown that estrogen stimulates the production of plasminogenactivator (10)." Others have reported that estrogen increases

the level of progesterone receptor in the cells (18), elevatesthe intracellular levels of lactate dehydrogenase (8) and thy-

midine kinase (29), stimulates amino acid incorporation (25),and induces the synthesis of proteins of various molecularweights which are released into the culture medium (35).

In the first report that estrogen stimulated the growth of MCF-7 cells (24), the stimulation was less than 2-fold, and even in

the presence of estradiol, growth was slow. McGuire reportedthat the growth of MCF-7 cells was stimulated either only

slightly (27) or not at all (18) by estrogen. More recently, 3other laboratories have reported that estrogen did not stimulatethe growth of MCF-7 cells in culture (16, 23).3 Our results are

in agreement with these later reports. Even under conditionswhere we can demonstrate the stimulation of plasminogenactivator production by estradiol (Chart 6), we see no effect ongrowth rate (10). Thus, the lack of growth response to estradiolseems unlikely to be due to the presence of residual estradiolin the cells or media.

We have described 2 factors which modify the response ofthe human breast cancer cell line MCF-7 to antiestrogens. The

first, serum concentration, is of concern primarily to thoseworking with MCF-7 and other steroid-sensitive cell lines. The

effect of increasing concentrations of serum in reducing tamoxifen inhibition could be due to estradiol in the serum. Theconcentration of estradiol in calf and fetal calf serum rangesbetween 10"10 and 10~9M (13), and thus the concentrationspresent in 10% serum (10"" to 10~10M) would be in the range

over which estradiol stimulates these cells in serum-free me

dium (Chart 6). It is less likely, however, that the level of

JANUARY 1981 85

on May 17, 2020. © 1981 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

W. B. Butler et al.

estradiol present in 2% charcoal-treated serum would still be

sufficient to protect against tamoxifen. In our hands, charcoaltreatment removes over 90% of an added dose of [3H]estradiol,

although it is reported to be less effective in removing conjugated steroids such as estrogen sulfates (34). It is also possiblethat steroid binding proteins in serum reduce the effectiveconcentration of tamoxifen or that factors in serum act, as doesinsulin, to alter the level of estrogen receptor in the cells. Sucha change, as discussed below, could account for the alteredsensitivity to tamoxifen. What is important is that the conditionsused will affect the response observed and that this will varywith the type, treatment, and level of serum.

The action of insulin in altering the level of estrogen receptorin MCF-7 cells and reducing their sensitivity to both estrogen

and antiestrogens demonstrates 2 important points: (a) that thelevel of estrogen receptor need not be a fixed property of a cellbut can be modified by other factors, in this case, the peptidehormone insulin; and (b) that the level of estrogen receptor incells can be related to their sensitivity to estrogen and antiestrogens.

Insulin has been reported to stimulate the growth of MCF-7

cells when they are grown in the absence of serum (28, 31).Our growth experiments were carried out in media with serum,and while we occasionally observed some insulin stimulationof growth, we usually did not. The effect of insulin in loweringthe estrogen receptor level and reducing the sensitivity of thecells to antiestrogens was seen in either case. We reportedpreviously (9) that cells grown for many passages in the absence of added insulin contained the same level of estrogenreceptor as cells maintained with insulin. The present resultsdo not contradict those since the receptor level was not measured in the media in which the cells had been growing. In thoseexperiments, cells from both cultures were grown under thesame conditions (without insulin) for the receptor measurement. Similarly, in the experiments reported here, all cells aretaken from cultures which have been maintained for manypassages in the presence of added insulin. The increasedreceptor level is seen when they are grown in media withoutinsulin.

The effect of insulin and diabetes on mammary tumors induced in rats by 7,12-dimethylbenz(a)anthracene has beenstudied extensively by Hilf ef al. (15) and Shafie and Hilf (32).About 60% of these tumors regress in diabetic animals (15).When animals are made diabetic, the level of estrogen receptordeclines in tumors which regress but increases in tumors whichcontinue to grow (15, 32). By the criteria of growth and estrogen receptor level, MCF-7 cells in culture are similar to the rat

mammary tumors which continue to grow in diabetic animals,that is, their growth is not dependent on insulin and the level ofestrogen receptor is higher in the absence of insulin. Whetherother hormones can also alter the level of estrogen receptor inMCF-7 cells is not known, although it has been reported thatthe level is increased by prolactin (31).

The mechanism by which insulin lowers the level of estrogenreceptor in MCF-7 cells is not known, although preliminaryexperiments have indicated that the effect may be rapid. It hasbeen suggested recently that phosphorylation and dephos-

phorylation of the receptor may play a role in the binding ofestrogen to the cytosolic receptor and the translocation of thereceptor to the nucleus (4). If this proves to be the case, insulinmight act by influencing such a system.

Most importantly, the results presented here provide evidence that the level of estrogen receptor in cells can becorrelated with their responsiveness both to estradiol and toantiestrogens. The 3-fold decrease in estrogen receptor level,

produced by growing cells in the presence of insulin, is sufficient to account for their reduced sensitivity to estradiol, asmeasured by the induction of plasminogen activator (Chart 6).This is shown in Chart 8 where data on the induction ofplasminogen activator (derived from Chart 6) have been combined with a curve showing the number of sites per cell withbound estradiol at different concentrations of estradiol. Bindingcurves were calculated from the average levels of receptor percell and the binding constants measured in experiments of thetype shown in Chart 7. In both cultures (plus or minus insulin),50% induction of plasminogen activator occurs at concentrations of estradiol which give similar levels of receptor-boundestradiol (10,000 molecules/cell). However, it requires over 4times as high a concentration of estradiol to achieve this levelin the cells with a lower receptor level (plus insulin).

If receptors, activated by bound estrogen, achieve theireffects by binding to specific sites in the nucleus, then theirbinding to any one site should be dependent on the level of thereceptor-estrogen complex rather than directly on the level of

estrogen. The relationship described in Chart 8 is consistentwith such a model.

The possibility that the action of steroids on cells or tissuesis modified by altering the level of steroid receptors has beensuggested in other systems. In several tissues, progesteroneantagonizes the action of estradiol and appears to do this bylowering the level of estrogen receptor (6, 20, 21). Reducingthe estrogen receptor level in a tissue may prevent estrogenaction even when the level of circulating hormone is unchanged(6). Another example is cell lines sensitive to dexamethasonein which lines with decreasing levels of receptor are increasingly resistant to dexamethasone (5).

It has been suggested that the level and not just the presenceor absence of estrogen receptor in human breast tumors should

oï>

O IO"" IO"10 IO"* IO"'

Estradiol Concentration [M]

Chart 8. Comparison of estradiol binding and plasminogen activator inductionas a function of estradiol concentration ±insulin. The data shown were obtainedfrom Chart 6 and from the average binding constant and receptor levels per cellfrom experiments of which Chart 7 is an example. Bars, S.D.

86 CANCER RESEARCH VOL. 41

on May 17, 2020. © 1981 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Insulin and Antiestrogens in Human Breast Cancer

be considered in designing therapy (26). The results presentedhere lend support to that suggestion by providing evidence thatthe sensitivity of responsive cells to both estrogen and anties-

trogens changes as a function of the estrogen receptor level.The effect of insulin on tamoxifen inhibition, shown in Chart

5, can also be explained by the lower level of estrogen receptorin cells grown with insulin. This observation has implicationsconcerning the mode of action of the antiestrogens tamoxifenand nafoxidine in inhibiting growth and thymidine incorporationin these cells. While the mechanism by which antiestrogensinhibit the growth of mammary tumors is unclear (19), at least3 models can be suggested whereby antiestrogens, by bindingto the estrogen receptor, could inhibit the growth of mammarytumor cells such as MCF-7. (a) If estrogen acts through its

receptor to stimulate growth, then antiestrogens, by competingwith estrogen for receptor, would block the stimulatory effectof estrogen, (to) It was reported that estrogen-free receptorsare present in the nucleus of MCF-7 cells (37), though a more

recent report from the same laboratory indicates that may notbe so (12). If, in MCF-7 cells, estrogen-free receptors can

stimulate the cells, as was suggested (37), then antiestrogens,by binding to free receptors, could make them unavailable andprevent their stimulatory activity. And (c) the receptor, withantiestrogen bound to it, could act in the nucleus to alter geneexpression in a way which inhibits cell division or tumor growth.In any of these 3 models, added estrogen would be expectedto overcome the inhibition caused by antiestrogens, as it doesin MCF-7 cells. However, the effect of lowering the level ofestrogen receptor in the cells on their sensitivity to antiestrogens would not be the same in all models. In either of the first2 models, growth is stimulated by estrogen receptors, andantiestrogens, by binding to receptors, make them unavailableto the cells. Thus, lowering the total level of receptor in thecells should increase their sensitivity to antiestrogens. In thethird model, however, lowering the level of receptor wouldmake the cells less sensitive to the action of antiestrogens, justas it makes them less sensitive to the induction of plasminogenactivator by estradiol (Charts 6 and 8). The increased resistance to these antiestrogens produced by lowering the receptorlevel is most easily explained if receptor with bound antiestrogens is actively inhibitory. If antiestrogens only act by competing with estrogen or by titrating out free receptor, then cellswith a lower level of receptor should be more sensitive to them,not less sensitive as seen here.

Such an active model of antiestrogen action has a furtherimplication. Tumors containing estrogen receptor might looseeither their requirement for estrogen or their sensitivity toantiestrogens independently. It might be expected, therefore,that not all tumors would be equally responsive to proceduressuch as ovariectomy or adrenalectomy, which are designed tolower the level of estrogens on one hand and to therapy withantiestrogens on the other. There is some evidence that this isindeed the case. In one report on patients with metastaticbreast cancer who had previously undergone adrenalectomyand were then treated with nafoxidine, the type of response toadrenalectomy did not necessarily correlate with the type ofresponse to nafoxidine (22). Some responded to both, some toneither, and some to only one or the other, a finding consistentwith the model proposed above in which antiestrogens canactively inhibit the growth of mammary tumors cells rather thanor in addition to interfering with the action of estrogen.

ACKNOWLEDGMENTS

We are grateful to Dr. Sam C. Brooks of the Wayne State University School ofMedicine for helpful discussions and to Dr. Willis Kirkland and Trudy Gargala forthe assay of plasminogen activator. We also thank Lois Trench of I.C.I. UnitedStates, Inc., and Dr. Paul O'Connell. The Upjohn Company, for their gifts of

tamoxifen and nafoxidine, respectively.

REFERENCES

1. Armelin. H. A.. Nishikawa. K., and Sato. G. H. Control of mammalian cellgrowth in culture: the action of protein and steroid hormones as effectorsubstance. In: B. Clarkson and R. Baserga (eds.). Control of Proliferation inAnimal Cells, pp. 97-104. Cold Spring Harbor, N. Y.: Cold Spring HarborLaboratory. 1973.

2. Ball, C. R.. Poynter, R. W., and Van Den Berg, H. W. A novel method formeasuring incorporation of radioactive precursors into nucleic acids andproteins of cells in monolayer culture. Anal. Biochem.. 46: 101-107, 1972.

3. Barret, J. C.. Crawford, B. D., and Ts'o, P. O. P. Quantitäten of fibrinolyticactivity of Syrian hamster fibroblasts using 3H-labeled fibrinogen prepared

by reductive alkylation. Cancer Res.. 37: 1182-1185. 1977.4. Bodwin. U.S.. Cho-Chung, Y. S., and Schneider, W. C. Nuclear translocation

of estrogen receptor and cyclic AMP-binding protein in MTW9 mammarytumor. Proc. Am. Assoc. Cancer Res., 21: 36, 1980.

5. Bourgeois, S., Newby, R. F., and Huet, M. Glucocorticoid resistance inmurine lymphoma and thymoma lines. Cancer Res., 38. 4279-4284, 1978.

6. Brenner. R. M., Resko, J. A., and West, N. B. Cyclic changes in oviductalmorphology and residual cytoplasmic estradiol binding capacity induced bysequential estradiol-progesterone treatment of spayed rhesus monkeys.Endocrinology, 95. 1094-1104, 1974.

7. Brooks, S. C., Locke, E. R., and Soule, H. D. Estrogen receptor in a humancell line (MCF-7) from breast carcinoma. J. Biol. Chem.. 248: 6251-6253.

1973.8. Burke, R. E., Harris, S. C., and McGuire, W. L. Lactate dehydrogenase in

estrogen-responsive breast cancer cells. Cancer Res., 38. 2773-2776,

1978.9. Butler, W. B., Brooks, S. C., and Goran, N. Factors affecting the growth and

hormone responsiveness of a breast cancer cell line (MCF-7). J. Cell Biol.,75. 186a. 1977.

10. Butler. W. B., Kirkland, W. L., and Jorgenson, T. L. Induction of plasminogenactivator by estrogen in a human breast cancer cell line (MCF-7). Biochem.Biophys. Res. Commun., 90. 1328-1334, 1979.

11. Cohen, N. D., and Hilf, R. Effect of estrogen treatment on DMBA-inducedmammary tumor growth and biochemistry in intact and diabetic rats. Proc.Soc. Exp. Biol. Med., 748. 339-343. 1975.

12. Edwards, D. P., Martin, P. M., Horwitz, K. B.. Chamness, G. C.. and McGuire,W. L. Subcellular compartmentalization of estrogen receptors in humanbreast cancer cells. Exp. Cell Res., J27. 197-213, 1980.

13. Esber, H. J., Payne, I. J., and Bogden. A. E. Variability of hormone concentrations and ratios in commercial serum used for tissue culture. J. Nati.Cancer Inst., 50. 559-562, 1973.

14. Higuchi, K., and Robinson, R. C. Studies on the cultivation of mammaliancell lines in a serum-free, chemically defined medium. In Vitro (Rockville). 9:114-121, 1973.

15. Hilf. R.. Hissin, P. J., and Shafie, S. M. Regulatory interrelationships forinsulin and estrogen action in mammary tumors. Cancer Res., 38: 4076-4085, 1978.

16. Hokanson, K. M. Human breast tumor cell lines and their use of steroidhormones. Diss. Abstr. Int. B Sei. Eng.. 39: 2271, 1978.

17. Horwitz, K. B., Costlow, M. E., and McGuire. W. L. MCF-7: A human breastcancer cell line with estrogen, androgen. progesterone, and glucocorticoidreceptors. Steroids, 26. 785-795, 1975.

18. Horwitz, K. B., Koseki. Y., and McGuire, W. L Estrogen control of progesterone receptor in human breast cancer: role of estradiol and antiestrogen.Endocrinology, Õ03: 1742-1751, 1978.

19. Horwitz, K. B., and McGuire, W. L. Antiestrogens: mechanisms of action andeffects in breast cancer. In: W. L. McGuire (ed.), Breast Cancer—Advancesin Research and Treatment. Vol. 2. pp. 155-204. New York: PlenumPublishing Corp. 1978.

20. Hsueh. A. J. W.. Peck, E. J.. Jr., and Clark. J. H. Control of uterine estrogenreceptor levels by progesterone. Endocrinology. 98: 438-444, 1976.

21. Huag, E. Progesterone suppression of estrogen-stimulated prolactin secretion and estrogen receptor levels in rat pituitary cells. Endocrinology, )04:429-437. 1979.

22. Jain, J., Samal, B., Singhakowinta, A., and Vaitkevicius, V. K. Clinical trialof nafoxidine in adrenalectomized patients with advanced breast cancer.Cancer (Phila.). 40: 2063-2066, 1977.

23. Jozan, S., Moure, C., Gillois, M., and Bayard, F. Effects of estrone on cellproliferation of a human breast cancer (MCF-7) in long-term tissue culture.J. Steroid Biochem., 10: 341-342, 1979.

24. Lippman, M., Bolán, G., and Huff, K. The effects of estrogens and antiestrogens on hormone-responsive human breast cancer in long-term tissue

JANUARY 1981 87

on May 17, 2020. © 1981 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

W. B. Butler et al.

culture. Cancer Res., 36. 4595-4601. 1976.25. Lippman. M., Monaco. M. E., and Bolán, G. Effects of estrone, estradiol,

and estriol on hormone-responsive human breast cancer in long-term tissueculture. Cancer Res., 37. 1901-1907, 1977.

26. McGuire. W. L. Steroid receptors in human breast cancer. Cancer Res., 38.4289-4291, 1978.

27. McGuire. W. L.. Horwitz, K. B.. Zava, D. T., Carola, R. E., and Chamness.G.C. Hormones in breast cancer: update 1978. Metab. Clin. Exp., 27 487-

501. 1978.28. Osborne, C. K., Bolán, G., Monaco, M. E., and Lippman, M. E. Hormone

responsive human breast cancer in long-term tissue culture: effect of insulin.Proc. Nati. Acad. Sei. U. S. A., 73 4536-4540, 1976.

29. Osborne, C. K., and Lippman. M. E. Human breast cancer in tissue culture:the effects of hormones. In: W. L. McGuire (ed.). Breast Cancer—Advancesin Research and Treatment, Vol. 2, pp. 103-154. New York: PlenumPublishing Corp., 1978.

30. Osborne, C. K., Monaco, M. E., Lippman, M. E., and Kahn, C. R. Correlationamong insulin binding, degradation, and biological activity in human breastcancer cells in long-term tissue culture. Cancer Res.. 38. 94-102, 1978.

31. Shafie, S.. and Brooks. S. C. Effects of prolactin on growth and the estrogen

receptor level of human breast cancer cells (MCF-7). Cancer Res., 37. 792-799, 1977.

32. Shafie, S. M., and Hilf, R. Relationship between insulin and estrogen bindingto growth response in 7,12-dimethylbenz(a)anthracene-induced rat mammary tumors. Cancer Res., 38. 759-764, 1978.

33. Soule, H. D., Vazquez, J., Long, A., Albert. S.. and Brennan, M. A humancell line from a pleural effusion derived from a breast carcinoma. J. Nati.Cancer Inst., 5Õ. 1409-1415, 1973.

34. Vignon, F., Terqui, M., Westley, B.. Derocq, D., and Rochefort, H. Effects ofplasma estrogen sulfates in mammary cancer cells. Endocrinology, 106:1079-1086. 1980.

35. Westley, B.. and Rochefort, H. Estradiol induced proteins in the MCF-7human breast cancer cell line. Biochem. Biophys. Res. Commun., 90. 410-416, 1979.

36. Yang, N-S., Soule. H. D., and McGrath, C. M. Expression of murine mammarytumor virus-related antigens in human breast carcinoma (MCF-7) cells. J.Nati. Cancer Inst.. 59. 1357-1367, 1977.

37. Zava, D. T., Chamness, G. C., Horwitz, K. B., and McGuire, W. L. Humanbreast cancer: biologically active estrogen receptor in the absence ofestrogen. Science (Wash. D. C.), Õ96:663-664. 1977.

88 CANCER RESEARCH VOL. 41

on May 17, 2020. © 1981 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

1981;41:82-88. Cancer Res   W. Barkley Butler, William H. Kelsey and Nicolene Goran  Breast Cancer Cell Line MCF-7 to Estrogen and AntiestrogensEffects of Serum and Insulin on the Sensitivity of the Human

  Updated version

  http://cancerres.aacrjournals.org/content/41/1/82

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/41/1/82To request permission to re-use all or part of this article, use this link

on May 17, 2020. © 1981 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from