Differential effects of selenium on the growth of mouse mammary cells in vitro

Cancer Letters, 13 (1981) o Elsevier/North-Holland

333-344 Scientific Publishers Ltd.

DIFFERENTIAL EFFECTS OF SELENIUM MOUSE MAMMARY CELLS IN VITRO

DANIEL

MEDINA

Department

and CAROL

333

ON THE GROWTH

OF

J. OBORN

of Cell Biology, Baylor College of Medicine, Houston,

Texas 77030 (U.S.A.)

(Received 21 April 1981) (Revised version received 18 May 1981) (Accepted 20 May 1981)

SUMMARY

The

of

10e8 M, stimulated the growth of primary cell cultures of normal mammary cells and C4 preneoplastic cells and the established cell line YN-4, but not the growth of D2 preneopiastic cells and tumors in primary cell cultures and of established cell lines CL-S1 and WAZ-2t. The differential responses of cells from preneoplastic outgrowth lines C4 and D2 and of D2 primary tumors in vitro correlated with the sensitivity of these same cell populations to selenium-mediated inhibition of growth and tumorigenesis in vivo. The differential responses among the 3 cell lines will allow further comparative studies on the cell biological basis of selenium function.

INTRODUCTION

A number of recent experiments have demonstrated that selenium supplementation to the diet can inhibit chemical carcinogen-induced tumorigenesis in skin [32], liver [6,13], colon [12,18] and mammary gland [14,28,36] and viral-induced tumorigenesis in mouse mammary gland [27,29,30]. The mechanisms postulated for chemoprevention include alterations in chemical carcinogen metabolism [9,19], immune response [33,34], lipid peroxide catabolism [17] and a general anti-oxidant function [7,18]. Whether any of these mechanisms are imvolved in all aspects of selenium chemoprevention has not been fully investigated. Two recent publications have shown that selenium chemoprevention is effective in the promotion stages of

334

carcinogenesis [ 12,361, a step which does not involve chemical carcinogen metabolism. Additionally, selenium is known to be an essential trace element for mammalian species [31,35] and an essential requirement for clonal growth of normal human fibroblasts in vitro [22]. Although seleniumdependent glutathione peroxidase is the best characterized function for selenium in the physiology of the mammalian cell, it is conceivable that other enzymes or proteins will be found to be dependent or influenced by selenium [ 5,201. One of the characteristics of mouse mammary tumorigenesis is the presence of preneoplastic cell populations which are direct precursors to mammary tumors [23] . Recent results have shown that selenium markedly inhibits the occurrence of preneoplastic alveolar and ductal hyperplasias, but has little, if any, effect on the growth of established primary mammary tumors [ 281. Additionally, selenium significantly inhibited the tumorproducing capabilities of preneoplastic nodule line C4, but not of 3 other nodule lines [27]. The ability to grow preneoplastic nodule lines in serumfree, chemically-defined medium [26] provided the means to examine the effects of selenium on the growth potential of normal, preneoplastic, and neoplastic mammary cells in vitro. Here we report on the differential effects of selenium on the growth of mammary cells in primary cell cultures and on established epithelial cell lines of mammary origin. MATERIALS

AND METHODS

Tissues All mammary tissues for primary monolayer cell cultures were obtained from BALB/cCrlMed mice bred and maintained in a closed conventional mouse colony in the Department of Cell Biology, Baylor College of Medicine. Normal mammary glands were taken from primiparous mice at 16-18 days of gestation. Preneoplastic mammary tissues were taken from hyperplastic alveolar nodule outgrowth lines D2 and C4 [ 24,251, which were propagated by serial transplantation in the mammary fat pads of syngeneic mice. Mammary adenocarcinomas developed as primary neoplasms in mice bearing transplants of the hyperplastic alveolar nodule lines. Cell lines The established mouse mammary cell lines CL-S1 and WAZ-2t were kindly provided by Dr. H. Hosick, Washington State University, Pullman, WA. The mouse mammary cell line YN-4 was provided by Dr. S. Nandi, University of California, Berkeley, CA. All the cell lines were routinely grown and maintained in.plastic flasks (Corning Labware, Corning, NY) in Dulbecco’s modified Eagle’s medium (DMEM) (H-21; Grand Island Biological Co., Grand Island, NY) with 10% fetal bovine serum (FBS) (Reheis Chemical Co., Kankakee, IL), 15 mM Hepes buffer, gentamicin and insulin (5 pg/rnl). Cultures were maintained at 37°C in a air/COz’

335

: 7.5) mixture. The cell lines were subcultured every 3-4 days on the basis of the density of the culture. Cell lines CL-S1 and YN-4 were derived from preneoplastic nodule outgrowth lines Dl and D2, respectively [l; Nandi, pers. comm.] . Line WAZ-2t was derived from a mammary tumor which arose in the nodule outgrowth line Dl [S 1. Cell lines CL-S1 and YN-4 maintained an epithelial morphology in cell culture and retained plasma membrane junctional specializations as seen with the electron microscope. Whereas CL-S1 does not produce tumors upon injection into syngeneic BALB/c mice [ 1; Medina, unpublished data] , lines YN-4 and WAZ-2t produce adenocarcinomas upon injection in BALB/c mice [l; Medina, unpublished data]. (92.5

Preparation

of cell cultures

The procedures for dissociation of mammary tissues and preparation of primary mammary cell cultures have been detailed previously [2]. Briefly, tissue fragments were dissociated in DMEM (pH 7.75) containing collagenase, bovine serum albumin, Hepes buffer without serum or bicarbonate for 60-90 min at 37°C in a shaker water bath. After enzymatic digestion, the cell suspension was vigorously pipetted and then centrifuged very briefly (800 rev./min, 3-4 min). The resulting cell pellet was resuspended in DMEM with 10% FBS and insulin (5 pgfml) and centrifuged briefly. After 2 washes in serum-containing medium, the cells were counted using a Coulter counter and plated at 1 X 105/cm2 into 35 or 60-mm Falcon petri dishes containing serum-free DMEM with chemicallydefined supplements of fibronectin (1 r.cg/ml) or fetuin (1 mg/ml), transferrin (10 pg/ml), epidermal growth factor (EGF) (10 &ml), insulin (5 pg/ml), 15 mM Hepes buffer; and gentamicin (50 yg/ml). Cell lines were collected from the parent flasks by trypsinization, washed twice in DMEM with serum and insulin and plated at 3 X 104/cm2 into 35-mm Falcon petri dishes containing serum-free, chemicallydefined DMEM. The cells were incubated at 37°C in air/CO, (92.5 : 7.5) for 24-48 h before changing the medium to selenium-containing medium. Measurement

of cell growth

The numerical estimation of cell growth was performed by several assays: (a) total cell number was calculated by counting cells with a Coulter counter after the cells were removed from the culture dish using trypsin versene solution or indirectly estimated by analyzing the amount of DNA per dish using the procedure of Burton [4] ; (b) estimation of DNA synthesis was done using a modification of the procedure of Hennings et al. [15]. Cells were exposed to r3H] thymidine (1 pCi/ml) for 1 h. Cells were washed in phosphate-buffered saline twice, treated with 1 ml of cold 0.5 N HC104, and then scraped from the dish adding 1 ml of cold 0.5 N HClO+ The precipitated material was centrifuged at 2000 rev./min, the pellet was resuspended in 0.5 N HC104, and DNA was hydrolyzed at 70°C for 30 min. An aliquot of the supernatant (100 ~1) was neutralized in 1 N NaOH, Bio-

336

fluor was added, and the mixture was counted on a Beckman liquid scintillation counter to determine [3H] thymidine incorporation into DNA. The remaining hydrolysate was analyzed for DNA content using the procedureof Burton [4], with calf thymus DNA used to produce a standard curve. Chemicals

The reagents used in these experiments were purchased from the following vendors: DMEM, Grand Island Biological Co., Grand Island, NY; insulin, transferrin, collagenase, and Hepes buffer, Sigma Chemical Co., St. Louis, MO; fetuin, Colorado Serum Co., Denver, CO: EGF, fibronectin, Collaborative Research Co., Waltham, MA; Na,Se03, Pfaltz and Bauer, Inc., Stamford, CT; FBS, Reheis Chemical Co., Kankakee, IL. Statistics

The data for each experiment were analyzed by the Student’s t-test. RESULTS

Selenium

and growth

of normal mammary

cells

The effect of selenium on the growth of normal mammary gland cells in primary monolayer cell culture is shown in Fig. 1. These experiments were performed at different times during the course of these investigations, thus they reflect evolving experimental protocols. In order to compare the results among the 3 experiments, the data are plotted as lug DNA/cm*. In the first trial, the cells were grown in DMEM + 13% FBS + insulin (5 pg/ml) for 48 h in SO-mm petri dishes, then switched to serum-free, chemically-defined DMEM with and without 5 X 10e8 M selenium (selenium supplied as Na,SeO,: the molarity refers to the molarity of selenium). The cultures were

7

0

5X

10-8MSe

6

Fig. 1. This graph shows the effect of selenium on the growth of normal mammary cells grown in primary monolayer cell culture. Three separate experiments grown under different conditions are shown in the bar graphs (see text for specific conditions). The data are given as the mean pg DNA/cm* of petri dish + S.E.M. The open bars represent cells grown at 5 x lo-& M selenium and the hatched bars are cells grown in the absence of selenium. The numbers represent the number of petri dishes/group.

331

analyzed 7 days after the switch to serum-free medium. In the second trial, the cells were grown in serum-free, chemically-defined DMEM supplemented with fetuin for 48 h in 35-mm petri dishes, then switched to serum-free, chemicallydefined DMEM with and without 5 X lo-’ M selenium. The cultures were analyzed 5 days after the switch to serum-free medium. The third trial was done similarly to the previous trial except the cells were grown for 24 h before the switch to the selenium-containing medium and were analyzed 4-days subsequently. The qualitative results in all 3 experiments were similar. Normal mammary cells grew significantly better in medium containing selenium. There were 56, 72 and 91% more cells in the seleniumcontaining medium than in medium without selenium for the 3 experiments, respectively. Selenium

dose and growth

of mammary

cells in primary

cell culture

The effect of selenium dose on the growth of normal, preneoplastic and neoplastic cells in primary cell cultures is shown in Fig. 2. In these experiments, the freshly dissociated cells were grown in 35-mm petri dishes in 200

Normal(l) --II-

C4-HAN@)

D2-HAN(2) DP-Tumor(S)

180 160 3 : 5

140-

0 120t E E

loo-

s 6

80.

f t

6040 20 -

0' ABCDE

LBCDE *am

ABCDE

ABCE

Fig. 2. This figure shows the effect of graded doses of selenium on the growth of different mammary cell populations in primary mammary cell cultures. The concentrations of selenium in the different groups are A, 5 x 10e8 M; B, 5 X 10.’ M; C, 5 X 10e6 M; D, 1 X 10.’ M; E, 5 X 10.’ M. The numbers in parentheses represent the number of different experiments. The data are plotted as percent of control (Jc rg DNA/dish) in order to compare the different groups.

338

serum-free, chemically-defined DMEM supplemented with fetuin or fibronectin for 24-48 h, then were switched to the serum-free, chemicallydefined DMEM supplemented with graded doses of selenium (5 X 1O-8-5 X lo-’ M). The cells were assayed for their DNA content 4-5 days after the medium switch. Three to 5 petri dishes were assayed for each dose. The data are expressed as the percent of control values with the control values representing cells grown in the absence of selenium. Normal cells from both experiments for each selenium concentration exhibited a significant increase in growth at 5 X 10W8M; this effect decreased with increasing doses of selenium. At 1 X 10m5M (group D), selenium significantly inhibited the net growth of normal mammary cells. The response of mammary cells from the preneoplastic nodule line C4 was qualitatively and quantitatively similar to normal mammary cells. There was growth stimulation at 5 X 1Om8M and growth inhibition at 1 X lo-’ M. The response of these cells to 5 X 10e6 M was varied with 2 experiments exhibiting no significant inhibition compared to controls (82%, 80%) and 1 experiment exhibiting a specific inhibition of growth (54%). There were no obvious differences in the culture conditions between the 3 experiments. The response of mammary cells from the preneoplastic nodule line D2 differed from the response of normal mammary cells and C4 preneoplastic cells. Selenium did not stimulate the growth of these cells at the lower concentrations (5 X 10e8-5 X 10e7 M), but did inhibit the growth of these cells at 1 X 10W5M. The response of tumor cells derived from the D2 cell populations varied slightly from its precursor population. Selenium at concentrations of 5 X lo-’ M and 5 X lo-’ M did not stimulate growth of these tumor cells, however, at 5 X 10e6 M, there was a significant inhibition of growth (55%) which was not seen in the other cell populations. This experiment was repeated 5 times with the same result each time. The inter-experiment variability in the means at 5 X 10W6M for the D2 tumor cells was only 2% (normalizing the values of each experiment to the controls which were considered as 100%). The inter-experiment variability in the means of all the doses for all the groups ranged from 2--8% within each tissue type, with the 1 exception noted above for the C4 preneoplastic cells at 5 X 1O-6 M. The uptake of [ 3H] thymidine into DNA during the log phase of the growth curves for the various mammary cell populations is illustrated in Fig. 3. The peak periods of growth occurred at different times after plating into serum-free, selenium-containing DMEM. Normal and D2 HAN cultures were assayed 48 h after the switch. C4 HAN cultures were assayed 24 h, and D2 tumor cultures 96 h, after the switch. The qualitative results in each culture paralleled the results seen in the DNA content assays. In all cases, 5 X lo-’ M (group E) selenium inhibited uptake of [3H] thymidine, whereas 5 X 10m8M and 5 X 10m7M (groups A and B) selenium stimulated uptake of f3H] thymidine only in C4 and normal mammary cells. Selenium

and growth

of established

mammary

cell lines

The effect of selenium dose on the growth of established non-neoplastic

339

Normal ---I

C4-HAN

DP-HAN

DP-Tumor

140? z 0

120-

eE loog 5 L E 6

r

80BO-

6 E8

40-

z n

20.

ABCE

A BCE ABCE Groups

11,

ABCE

Fig. 3. This figure illustrates the same experiment as Fig. 2, but the parameter measured was uptake of [‘HI thymidine during peak periods of DNA synthesis. The data are expressed as percent of control (5 cpm/pg DNA) in order to compare the different groups.

and neoplastic mammary cell lines is shown in Fig. 4. In these experiments, cells were plated and grown in serum-free, chemically-defined DMEM with fibronectin for 24 h, then switched to the serum-free, chemically-defined DMEM with various doses of selenium. The cultures were analyzed 4 days later for DNA content and by cell counts. Three to 5 petri dishes were analyzed for each point in each experiment. Both types of analysis gave the same relative results. Selenium at 5 X 10T5 M (group E) inhibited the growth of all 3 cell populations, whereas the growth of only YN-4 was stimulated by low concentrations of selenium (5 X lo-‘, 5 X lo-’ M; groups A and B). The response of cell line YN-4 was similar to that of normal mammary and C4 preneoplastic mammary cells in primary cell cultures whereas cell lines CL-S1 and WAZ-2t responded like D2 preneoplastic mammary cells in primary cell cultures. Since YN-4 and CL-S1 represented apparently selenium responsive and non-responsive mammary cell populations, we examined the response of these 2 populations further, under conditions where the cells had almost saturated the petri dish and their growth rates had decreased, but not stopped (Medina, unpublished data). If selenium still exerted.a differential effect on the 2 cell populations with this protocol, then future experiments examining uptake and localization of selenium isotopes would be easier to perform and analyze. The cells were plated at 3 X 104/cm2 in petri dishes containing serum-free, chemicallydefined DMEM supplemented with fibronectin and grown for 3 days. At 72 h, the medium was switched to serum-

340 Y N-4(3)

160 g 0 9 P

s

140 , 2.

i

WAZ-2T(3) CL-S l(3)

t l-l loo-

E g 0

80-

6 E Q

60-

g

40-

. A

20 0.

1 ABC

E

1

ABCE

ABCE QlOUpS

Fig. 4. This figure shows the effect of different doses of selenium on the growth of established mammary cell lines in monolayer cell culture. The numbers in parentheaea represent the number of experiments for each cell line. The data are expressed as percent of control (ccg DNA/dish) with the controls representing cells grown in the absence of selenium. The groups A-E are the same as those given in Fig. 2.

free, chemicallydefined DMEM with various doses of selenium. The cells were assayed for DNA content (4 petri dishes/group) at 0 time (72 h of culture) and for DNA content and uptake of [ 3H] thymidine after 72 h in the selenium-containing DMEM (144 h of culture), the time when both cell lines exhibited maintenance levels of DNA synthesis. The results are shown in Fig. 5. The cells doubled in number between days 3 and 6 of culture +-CL-s1w

+YN-4----( 100 80

ABCDEF

ABCDEF

Qroups Fig. 5. This figure illustrates the effect of graded doses of selenium on the growth of cell lines YN-4 and CL-S1 The cells were grown for 3 days in the absence of selenium (group A), then grown for 3 more days in either the absence (group B) or presence of graded doses of selenium (group C, 5 x 10eR M; group D, 5 X lo-’ M; group E, 5 X 10m6M; group F, 5 X 10.’ M). The data are given as the mean fig DNA/dish + S.E.M.

341

(group A vs. B, respectively). Low doses of selenium (5 X lo-* M; group C) stimulated the growth of YN-4, but not CL-Sl, whereas high doses (5 X lo-” M; group F) completely inhibited, but did not kill the YN-4 and CL-S1 cells. At 144 h of culture, the uptake of [3H] thymidine was depressed >99% at 5 X 10m5 M selenium for both populations, whereas control and 5 X 10-s M groups exhibited similar levels of [ 3H] thymidine uptake. DISCUSSION

The results reported herein document some of the direct effects of a trace element, selenium, on the growth in vitro of different mammary epithelial cell phenotypes. Several recent reports have demonstrated .that selenium supplementation in the diet inhibits experimentally-induced carcinogenesis in skin [32], liver [6,13], colon [12,18] and mammary gland [14,2730,361. Selenium inhibits both the initiation and promotion stages [12,36] in chemical carcinogenesis as well as viral-induced mouse mammary tumorigenesis. Several mechanisms have been proposed to explain this seleniummediated inhibition in vivo. In contrast, there are only a few reports on the direct effect of selenium in vitro. McKeehan and McKeehan showed that selenium was essential for the clonal growth of normal human fibroblasts WI-38 [21,22] ; Gaisuddin and Diplock showed that selenium stimulates the growth of BALB/c 3T3 cells in vitro [lo], and Sato and coworkers use selenium in their serum-free medium for optimal growth of a variety of cell lines [3]. Several of the results reported herein are noteworthy. The addition of selenium to the serum-free medium resulted in differential effects on growth of the target cells depending on the dose of selenium and the target cell phenotype. Selenium, at 5 X low8 enhanced the growth of normal and C4 preneoplastic mammary cell populations in primary monolayer cell cultures and of the established cell line YN-4. This coincides with the observations of McKeehan and McKeehan who demonstrated that 5 X 10e8 M enhances the growth of WI-38 cells [22]. Of more interest is the correlation between inhibition of tumorigenesis in vivo and response to selenium in vitro. Earlier studies demonstrated that preneoplastic line C4, but not line D2, was responsive to selenium-mediated inhibition of tumorigenesis in vivo [27]. In the results reported herein, cells from line C4, but not from line D2, were responsive to low doses of selenium. Toxic levels of selenium were manifested in all tissues examined at lo-’ M. In addition to the primary cell cultures and cell lines reported herein, we have examined 2 other cell lines of non-mammary origin (3T3, SV-3T3) and 3 other cell lines of mammary origin (BESD, C57NMG and MTV-L). Selenium at 10T5 M markedly inhibited (>80%) the growth of all of these cell lines (Medina, unpublished data). This non-specific toxicity of lo-’ M selenium under in vitro conditions was reported by McKeehan and McKeehan for WI-38 cells [22] and was probably the reason for the inhibition of ascites tumor growth recently reported by Greeder and Milner [ 111.

342

The response of D2 mammary tumors grown in primary cell cultures to 5 X 10T6 M selenium was puzzling since this population appeared responsive to selenium-mediated inhibition in vitro. Previous results demonstrated that selenium (5 mg/l) did not inhibit the growth of primary tumors when assayed by subcutaneous transplantation into syngeneic mice [ 281. Several reasons could explain the different results in vivo and in vitro. The cell growth kinetics may be sufficiently different between the 2 growth states to obviate any selenium effect in vivo. Alternatively, tumor cells growing in vivo may not encounter and incorporate enough selenium to be inhibited in their growth. However, tumorigenesis in C4 preneoplastic cells are inhibited in vivo, which indicates that selenium at 3-5 mg/l in vivo is sufficient to inhibit some mammary populations. The ability of normal, preneoplastic and neoplastic mammary cells to grow in serum-free, chemicallydefined medium facilitated the analysis of the effects of selenium. McKeehan and McKeehan [22] showed that selenium bound to serum proteins which altered the amount of selenium available to the target cells. Furthermore, Gaisuddin and Diplock [lo] showed that the selenium content of serum is considerable enough to alter the intended concentrations in experimental designs. This problem, in addition to the variability in selenium concentration in different batches of serum can provide problems in experiments where low and physiological levels are being examined for their actions. The use of serum-free medium overcomes these variables and in the experiments reported here, it allowed the stimulatory effects of selenium to be easily and reproducibly demonstrated. The availability of selenium responsive and non-responsive cell lines will facilitate the analysis of the mechanism of selenium action. The relationship between cell growth and selenium function can be analyzed with respect to selenium uptake, localization, and content; glutathione peroxidase induction and content, and the general ultrastructure of the cell. The relationship between the effects of selenium at 5 X lo-’ M and the inhibitory effects on tumorigenesis in vivo are not obvious at this time. On the one hand, high doses of selenium show non-specific toxicity and one can argue that selenium-mediated inhibition of tumorigenesis is the consequence of this toxicity. However, only certain stages in tumorigenesis in vivo are inhibited by selenium [27,28] and the host exhibited no significant signs of general toxicity [28]. Alternatively, the stimulatory effects of responsive mammary cell populations, i.e., normal cells, C4 preneoplastic cells and YN-4 cell lines, may be coincidental to other effects of selenium within the cell. Finally, C4 and YN-4 may contain heterogeneous cell populations which have different responses to selenium. If a low or non-tumorigenic cell subpopulation is responsive to selenium-mediated stimulation of growth, these cells may eventually comprise a majority of the total population, thus resulting in a significant delay in tumor formation in vivo. At this time, the study of selenium chemoprevention and possible mechanisms of action are in their infancy and the significance of phenomenological observations can only be elucidated by further in-depth experiments.

343

In summary, the results reported herein document the feasibility of examining the direct effects of selenium with cell cultur& in vitro and describes 2 cell lines, 1 responsive and 1 non-responsive to selenium, which can be used as model systems for selenium studies. REFERENCES 1 Anderson, L.W. Danielson, K.G. and Hosick, H.L. (1979) Epithelial cell line and subline established from premalignant mouse mammary tissue. In Vitro, 15, 841-843. 2 Asch, B.B. and Medina, D. (1978) Concanavalin A-induced agglutinability of normal, preneoplastic and neoplastic mouse mammary cells. J. Natl. Cancer Inst., 61, 14231430. 3 Bottenstein, J.E., Hayaski, I., Hutchings, S., Masui, H., Mather, J., McClure, D.B., Ohasa, S., Rizzina, A., Sato, G.H., Serrrero, G., Wolfe, R. and Wu, R. (1979) The growth of cells in serum-free hormone-supplemented media. Methods Enzymol., 58,94-106. 4 Burton, K. (1956) A study of the conditions and mechanism of the diphenylamine reaction for the calorimetric estimation of deoxyribonucleic acid. Biochem. J., 62, 315-323. 5 Calvin, HI. (1978) Selective incorporation of selenium-75 into a polypeptide of the rat sperm tail. J. Exp. Zool., 204,445-452. 6 Clayton, CC. and Baumann, C.A. (1949) Diet and azo dye tumors: Effect of diet during a period when the dye is not fed. Cancer Res., 9, 575-582. 7 Combs, Jr., G.F., Noguchi, T. and Scott, M.L. (1975) Mechanisms of action of selenium and vitamin E in protection of biological membranes. Fed. Proc., 34, 20902095. 8 Danielson, K.G., Anderson, L.W. and Hosick, H.L. (1980) Selection and characterization in culture of mammary tumor cells with distinctive growth properties in uivo. Cancer Res., 40, 1812-1819. 9 Daoud, A.H. and Griffin, AC. (1978) Effects of selenium and retinoic acid on the metabolism of N-acetylaminofluorene and N-hydroxyacetylaminofluorene. Cancer Letters, 5, 231-237. 10 Giasuddin, A.SM. and Diplock, A.T. (1979) The influence of vitamin E and selenium on the growth and plasma membrane permeability of mouse fibroblasts in culture. Arch. Biochem. Biophys., 196, 270-280. 11 Greeder, G.A. and Mimer, J.A. (1980) Factors influencing the inhibitory effect of selenium on mice inoculated with Ehrlich ascites tumor cells. Science, 209, 825-827. 12 Griffin, A.C. (1979) Role of selenium in the chemoprevention of cancer, Adv. Cancer Res., 29, 419-442. 13 Griffin, A.C. and Jacobs, M.M. (1977) Effects of selenium on azo dye hepatocarcinogenesis. Cancer Letters, 3,177-181. 14 Harr, J.R., Exon, J.H., Weswig, P.H. and Whanger, P.D. (1973) Relationship of dietary selenium concentration, chemical cancer induction, and tissue concentration of selenium in rats. Clin. Toxicol., 6, 287-293. 15 Hennings, H., Michael, D., Cheng, C., Steinert, P., Holbrook, K. and Yuspa, S.H. (1980) Calcium regulation of growth and differentiation of mouse epidermal cells in culture. Cell, 19, 245-254. 16 Hoekstra, W.G. (1975) Biochemical function of selenium and its relation to vitamin E. Fed. Proc., 34, 2083-2089. 17 Ip, C. and Sinha, D.K. (1981) Enhancement of mammary tumorigenesis by dietary selenium deficiency in rats with a high polyunsaturated fat intake, Cancer Res., 41, 31-34.

344 18 Jacobs,M.M., Jansson, B. and Griffin, A.C. (1977) Inhibitory effects of selenium on 1,2_dimethylhydrazine and methylazoxymethanol acetate induction of colon tumors. Cancer Letters, 2,133-137. 19 Marshall, M.V., Arnott, M.S., Jacobs, M.M. and Griffin, A.C. (1979) Selenium effects on the carcinogenicity and metabolism of 2-acetylaminofluorene. Cancer Letters, 7, 331-338. 20 McConnell, K.P., Burton, R.M., Kute, T. and Higgins, P.J. (1979) Selenoproteins from rat testis cytosol. Biochim. Biophys. Acta, 588, 113-119. 21 McKeehan, W.L. and McKeehan, K.A. (1980) Serum factors modify the cellular requirement for Ca++, K’, Mg’+, phosphate ions, and 2-oxocarboxylic acids for multiplication of normal human fibroblasts. Proc. Natl. Acad. Sci. U.S.A., 77, 3417-3421. 22 McKeehan, W.L., Hamilton, W.G. and Ham, R.G. (1976) Selenium is an essential trace nutrient for growth of WI-38 diploid human fibroblasts. Proc. Natl. Acad. Sci. U.S.A., 73, 2023-2027. 23 Medina, D. (1973) Preneoplastic lesions in mouse mammary tumorigenesis. Methods Cancer Res., 7,3-53. 24 Medina, D. (1976) Mammary tumorigenesis in chemical carcinogen-treated mice. VI. Tumor-producing capabilities of mammary dysplasias in BALB/cCrgl mice. J. Natl. Cancer Inst., 57,1185-1189. 25 Medina, D. and DeOme, K.B. (1970) Effects of various oncogenic agents on tumorproducing capabilities of D series BALB/c mammary nodule outgrowth lines. J. Natl. Cancer Inst., 45, 353-363. 26 Medina, D. and Oborn, C.J. (1980) Growth of preneoplastic mammary epithelial cells in serum-free medium. Cancer Res., 40, 3982-3987. 27 Medina, D. and Shepherd, F. (1980) Selenium-mediated inhibition of mouse mammary tumorigenesis. Cancer Letters, 8, 241-245. 28 Medina, D. and Shepherd, F. (1981) Selenium-mediated inhibition of 7,12dimethylbenzanthracene-induced mouse mammary tumorigenesis. Carcinogenesis, in press. 29 Schrauzer, G.N. and Ishmael, D. (1974) Effects of selenium and of arsenic on the genesis of spontaneous mammary tumors in inbred C3H mice, Ann. Clin. Lab. Sci., 4,441-447. 30 Schrauzer, G.N., White, D.A. and Schneider, C.J. (1976) Inhibition of the genesis of spontaneous mammary tumors in C3H mice: Effects of selenium and of seleniumantagonistic elements and their possible role in human breast cancer. Bioinorg. Chem., 6,265-270. 31 Schwarz, K. and Foltz, C.M. (1957) Selenium as an integral part of factor 3 against necrotic liver degeneration. J. Am. Chem. Sot., 79, 2392-3293. 32 Shamberger, R.J. (1970) Relationship of selenium to cancer. I. Inhibitory effect of selenium on carcinogenesis. J. Natl. Cancer Inst., 44, 931-936. 33 Spallholz, J.E., Martin, J.L., Gerlach, M.L. and Heinzerling, R.H. (1973) Immunologic response of mice fed diets supplemented with selenite selenium. Proc. Sot. Exp. Biol. Med., 143, 685-689. 34 Spallholz, J.E., Martin, J.L. Gerlach, M.L. and Heinzerling, R.H. (1975) Injectable selenium: Effect on the primary immune response of mice. Proc. Sot. Exp. Biol. Med., 148, 37-40. 35 Stadtman, T.C. (1980) Selenium-dependent enzymes. Annu. Rev. Biochem., 49, 93-l 10. 36 Thompson, H.J. and Becci, P.J. (1980) Selenium inhibition of N-methyl-Nnitrosourea-induced mammary carcinogenesis in the rat. J. Natl. Cancer Inst., 65, 1299-1301.