Znt. J . Cancer: 22, 258-265 (1978)

INHIBITION BY LYMPHOBLASTOID INTERFERON OF GROWTH OF CELLS DERIVED FROM THE HUMAN BREAST F. BALKWILL, D. WATL~NG and J. TAYLOR-PAPADIMITRIOU Imperial Cancer Research Fund, Lincoln's Inn Fields, London W.C.2, England

Human lymphoblastoid interferon inhibited the growth in vitro of fibroblasts and epithelial cells from normal, hyperplastic and neoplastic human breast tissue. A t an interferon concentration of lo3 international units (IU) per mi, the inhibitory effects on monolayer growth were completely reversible but the growth potential of cells at lower density in colony-forming cultures was not completely recovered. Studies on the cell cycle distribution of interferon-treated cells demonstrated that the growth-inhibitory effect was not due to an effect on one specific phase of the cell cycle. but to a lengthening of all phases.

The interferons, originally discovered and conceived as antiviral agents (Isaacs and Lindenmann, 1957), have been shown over the past decade to have multiple biologic effects on cells (Gresser, 1977). The effectiveness of interferon as an anti-cancer agent, well documented in experimental animal systems (Gresser and Bourali, 1970, 1972; Lieberman et al., 1971), may depend on a growth-inhibitory effect on the tumour cells (Paucker etal., 1962; Gresser, 1970), an alteration of the surface of the tumour cells (Lindahl et al., 1974) and on an ability to enhance some specialized functions of host cells, in particular those of the immune system (Lindahl et al., 1972; Brown and Levy, 1972; Shultz et a[., 1977). In considering interferon as a potential anti-cancer agent in man, it would be useful to have information concerning the effect of human interferon o n the growth and function of human cells in vitro. Some studies have been reported on the growth-inhibitory effects of interferon on certain human cell strains and lines (Strander and Einhorn, 1977) and on primary cultures of normal, immune. and malignant haemopoietic cells (Balkwill and Oliver, 1977). In the present paper we describe the influence of human lymphoblastoid interferon preparations on the growth of normal, hyperplastic and neoplastic cells from the human breast. This interferon, derived from lymphoblastoid Namalva cells, contains primarily leukocyte interferon (Have11 et al., 1977; Strander et al., 1975) and may soon be available in large enough quantities for clinical trials. MATERIAL AND METHODS

Cells All cells were grown in Nunc 30-mm plastic dishes (Gibco Biocult, Paisley, UK), in a humidified 10% CO, atmosphere at 37" C and were free of PPLO as assessed by staining and growth on agar (Fogh and Fogh, 1964; Lemcke, 1965). Malignant cell lines from pleural effusions of breast cancer patients. MCF-7 cells (Soule et al., 1973) were kindly supplied by Dr. Marvin Rich of the Michigan

Cancer Foundation and had been through approximately 10 passages in our laboratories before use in the experiments reported here. The cells were grown in RPMI 1640 medium (Gibco Biocult)+lO% foetal bovine serum, FBS (Gibco Biocult). MDA 157 and 231 cell lines (Young et a[., 1974) were kindly sent to us by Dr. R. Cailleau (M. D. Anderson Hospital, Texas, USA) and had been through approximately three subcultures in our laboratories. These cells were grown in medium RPMI 1640+10% FBS+lO ,ug/mI insulin. Breast fibroblasts. HumF cells were derived from explants of chopped tissue from the nipple and areola area of a breast removed because of a carcinoma. The fibroblasts were diploid and were used between the 3rd and 5th passages. NIsF cells were obtained by enzyme digestion of tissue obtained from normal breast by reduction mammoplasty. N,,F cells were used between the 3rd and 6th passages. Both these cell strains were grown in RPMI 1640 medium+10% FBS, and exhibited a fibroblastic morphology. Human mammary epithelial cells (HumE) from milk. Cultures prepared from unfractionated milk aliquots were grown in 199 medium containing 10 ,ug/ml insulin, 5 ,ug/ml hydrocortisone, 15 % FBS and 20 % human serum (Taylor-Papadimitriou et al., 1977a). Alternatively, the epithelial cells were separated from foam cells by differential adhesion, followed by storage in liquid nitrogen, and cultured with mitomycin-treated 3T6 feeders (TaylorPapadimitriou et al., 1977b), in the same medium as above but without human serum. Epithelial cells from fibroadenomas. Cells were released from biopsy specimens by the spillage technique of Lasfarges (Lasfarges, 1975), and grown with 3T6 feeders in the same way as the fractionated HumE cells from milk. Cell and colony counts Cells were trypsinized, resuspended in " lsoton " (Coulter Electronics, Harpenden, UK) 1 % FBS and counted with a Coulter counter model ZB1 (Coulter Electronics). In MCF-7 and MDA 157 and 231 monolayer cultures, between 2 and 10% of the viable cells remained unattached to the monolayer and these were also counted. The sum of the supernatant and adherent cells is expressed in all figures. Colony sizes were measured using a slide with lines 1 mm apart on which the dishes were placed and examined under a low-power microscope.

+

Received: May 23, 1978 and in revised form July 4, 1978.

259

INTERFERON AND HUMAN BREAST CELLS

Interferon Human lymphoblastoid interferon, produced by Namalva cells, was prepared and partially purified by Dr. K. Fantes (Wellcome Research Labs., Beckenham, UK). The preparations used here had specific activities ranging from 1.2 x los IU-3 x lo7IU/mg protein. Units refer to British Standard Units calibrated against Std B69/19 (Nat. Inst. Biol. Stds. Controls, London, UK).

Cytofluorimetry studies Cells were trypsinized from the monolayer, washed, passed through a 27-gauge needle to ensure single-cell dispersion and then incubated in lysis buffer (Terada et af., 1977) for 1 rnin at room temperature. The nuclei were then incubated for 30 min at 37" C with ribonuclease A 0.5 mg/ml (Sigma Chemical Co., Kingston, UK) in SMT buffer (0.25 M sucrose containing 5 m M MgClz and 20 mM Tris. HCI PH 7.4), then stained with propidium iodide (Sigma Chemical Co.), 0.005 mg/ml, in 0.1 % tri-sodium citrate. The DNA content of the cells was measured on a Fluorescence Activated Cell Sorter (Becton Dickinson, Calif., USA) FACS-I using 488 nm excitation. Histograms were recorded on graph paper and analyzed (Cullen and Capellaro, 1977).

Interferon antiserum This was prepared in rabbits injected with a partially purified preparation of Namalva interferon and was kindly provided by Dr. K. Fantes of Wellcome Research Labs. Ten pl of undiluted serum neutralized lo3units of Namalva interferon. Autoradiography Cultures were incubated with [3H]methyl thymidine (sp. act. approx. 20 Cifmmole), at a dilution of 10 pCi/ml for 30 min, and processed for autoradiography as previously described (TaylorPapadimitriou et af., 1971).

10

RESULTS

Efect of interferon on cells of human breast origin The effect of interferon has been tested on a variety of cell lines and types derived originally from

Hum F

N13F

2 5

I

'-

1

3

5

7

I

1

I

1-

1

3

5

7

1

3

5

7

Days FIGURE1 - The influence of human lymphoblastoid interferon on the growth of human breast cells in monolayer cultures. Cells were plated at 1 x loKcells/dish and interferon (sp. act. 1.2 x los IU/mg protein for MCF-7, MDA 157, NIsF and HumF cells, sp. act. 4.2 x 106IU/mg protein for MDA 231) was added from the start of the experiment. control; 0-0, Medium with or without interferon was renewed three times weekly in all the cultures. 0-0, 10 1U/ml interferon; v-v, lo* IU/ml interferon; 0--0, 105IU/ml interferon.

260

BALKWILL ET AL.

the human breast. Fibroblasts and cell lines derived from pleural effusions in breast cancer patients (Soule et al., 1973; Young et al., 1974) can be grown in monolayer and subcultured, enabling quantitative measurements of growth to be made. Epithelial cells which can be cultured from milk (Taylor-Papadimitriou et al., 1977a, b) or from fibroadenomas (Stoker et al., 1976) are available in small numbers and can only be cultured for approximately 20 days when their growth is assessed by colony size and number. Our results, summarized in Figures 1 to 3, indicate that interferon inhibited the growth of all the cells which were tested, although the breast fibroblasts appeared to be somewhat less sensitive than either normal or malignant epithelial cells. Figure 1 shows the kinetics of replication of two strains of breast fibroblasts and three cell lines from pleural effusions. The cell lines have been shown to be of mammary epithelial origin by their positive reaction with a specific antiserum prepared against components of the membrane of the milk fat globule (Ceriani et al., 1977) and two of them have been reported to synthesize a-lactalbumin (Rose and McGrath, 1975). The inhibitory effect of interferon is obvious at 1031U/ml and this is sufficient to prevent any increase in number of MDA 157 cells. Detectable effects can also be observed at lo2IU/ml in all cells tested and at 10 IU/ml in MDA 231 and HumF cells after 7 days’ culture. The inhibitory effect of lo9 IU/ml interferon, added at various times, on the number and size of colonies developing in milk cell cultures is shown in Figure 2. Culturing the cells with interferon from days 12-14 has an effect on colony size although the number of colonies developing is reduced only slightly. Addition of interferon earlier than 12 days following seeding significantly reduces the number as well as the size of the colonies developing. The concentration of interferon required to affect colony formation by breast epithelial cells was then tested by adding interferon 5 days after seeding and culturing for 9 further days in the presence of interferon. Figure 3 shows cultures of HumE cells from unfractionated milk aliquots (Taylor-Papadimitriou et al., 1977a) and from a fibroadenoma treated in this way with interferon. For comparison, MCF-7 cells were grown at low density as discrete colonies, to see if the growth-inhibitory effect of interferon is more pronounced on cells grown at low density. In all three cases, there is a marked decrease in colony number at loz IU/ml interferon and a detectable inhibition of colony size at 10 IU/ml. A comparison of the data of Figure 1 and of Figure 3 for MCF-7 shows that the colony-forming assay is more sensitive for detection of the growth-inhibitory effect of interferon. The experiment illustrated in Figure 2 was done using unfractionated milk aliquots which contain a large excess of foam cells. These cells have properties characteristic of macrophages and also feed the HumE cells (Taylor-Papadimitriou, 19776). That the interferon is inhibiting directly the growth of the epithelial cells and not acting indirectly by antagonizing the feeder effect, is shown by the fact that fractionated HumE cells grown alone with human

serum (data not shown) or with 3T6 feeders (Fig. 3) are inhibited to the same degree. Reversibility of the growth-inhibitory eflect of interferon The reversibility of the growth-inhibitory effects of the interferon preparation was assessed on the highdensity monolayer cultures and low-density colonyforming cultures: MCF-7 cells were grown in monolayer with lo3IU/ml interferon for various times, and then the growth potential of the inhibited cells was assessed by cloning and counting colonies after 14 days. Table I shows that treatment with 103JU/ml for up to 7 days has no effect on the subsequent growth potential of the cells in terms of cloning efficiency. Also, growth inhibition by interferon did not affect the subsequent proliferation of the cells as measured by colony size (data not shown). However, the inhibitory effect of the interferon preparations on low-density colony-forming HUM t

1

1

IF 5- 14d ays

IF 2-f4days

1 201

IF7-Ildays

1

I F O-lddays

Colony size - m m

FIGURE 2 - The effect of adding interferon at various times during the development of HurnE colonies in vifru. Unfractionated milk cells were plated as described by Taylor-Papadirnitriouet al. (1977~)and interferon (sp. act. 1.2 x lo6IU/mg protein) losIU/ml was added at various times as shown in the Figure. Medium, with or without interferon, was renewed every 2-4 days. All cultures were fixed and stained for counting after 14 days. The number of colonies shown in the histograms represents the sum of two dishes.

INTERFERON AND HUMAN BREAST CELLS TABLE I

THE EFFECT OF PRIOR EXPOSURE TO HUMAN INTERFERON PREPARATIONS ON THE CLONING EFFICIENCY OF MCF-7 CELLS No. of colonies developing from lo* cells Time of pre-exposure of monolayer culture to interieron (108 IU/ml) before cloning

1 day 2 days 4 days 7 days

Interferon pre-exposure

>20 cells

146 103 63 120

i 20 cells

32 25 17 29

>20 cells