Prolactin receptors in primary cultures of carcinogen-induced rat-mammary tumors

Molecular and Cellular Endocrinology,

0 Elsevier/North-Holland

PROLACTIN

14 (1979) 81-97 Scientific Publishers, Ltd.

81

RECEPTORS IN PRIMARY CULTURES OF

CARCINOGEN-INDUCED

RAT-MAMMARY

M.E. COSTLOW, P.E. GALLAGHER

TUMORS

and Y. KOSEKI

Department of Biochemistry, St. Jude Children’s Research Hospital, 332 North Lauderdale, Memphis, TN 38101 (U.S.A.)

Received 25 September 1978;accepted

23 January 1979

7,12-Dimethylbenz[a]anthracene-induced rat mammary tumors were dissociated with collagenase and hyaluronidase and placed into primary culture. In most cultures, specific binding of 1251-labeled ovine prolactin was (i) lower than that for the original tumors unless bovine prolactin (1 pg/ml) had been added to the dissociation medium, and (ii) varied with the type of growth medium used. The level of prolactin binding in cultured cells was relatively constant for the first 7-10 days. Prolactin binding in cultured cell homogenates was maximal at pH 7.0, proportional to cell protein, specific for prolactin, and reached a steady state by 12 h at 22°C. The half-maximum inhibition of 12 5 I-labeled prolactin binding by unlabeled prolactin was 100 ng/ml for cells grown in 5-1000 ng of prolactin/ml. After prolactin was removed from the growth medium, the level of available binding sites progressively increased, reached a maximum at 48 h and then declined. At 48 h, the dissociation constant for prolactin binding (Kd - 1 X lo-lo M) was comparable to that in tumors. In some cultured tumors, a 48-h treatment with 0.5 or 1.0 ng of prolactin/ml caused an apparent increase in the level of prolactin binding. Prolactin increased DNA synthesis and its removal caused a reduction in [3H]estradiol and [3H]R5020 binding to cultured cell cytosols. Keywords:

estrogen receptors; progestin receptors; hormone receptors.

Hormone dependence in experimental mammary carcinoma is a complex phenomenon involving the concerted action of a number of hormones. Considerable evidence indicates that prolactin is a major hormone controlling mammary tumor growth, although other polypeptide and steroid hormones are clearly involved (Kim and Furth, 1976; McGuire et al., 1977; Welsch and Nagasawa, 1977). Since the first and requisite event in the mediation of the cellular response to prolactin is the interaction of the hormone with cell surface receptors (Shiu and Friesen, 1976), the relationship of prolactin receptors to hormone-dependent mammary tumor growth has been widely studied. In certain instances, prolactin-receptor levels correlate with hormone-dependent growth characteristics (Costlow et al., 1975, 1977; Kelly et al., 1974), while in others they do not (Costlow, 1976b; DeSombre et al., 1976; Holdaway and Friesen, 1976; Smith et al., 1976). The biological basis for these differences remains unclear,

M.E. Costlow, P.E. Gallagher, Y. Koseki

82

owing partly to the difficulty in defining hormonal factors in vivo which influence the level of prolactin binding. In the fo~owing study, methods were developed for placing 7,12-d~ethylbenz[~]anthracene (DMBA~~duced rat mammary tumors in primary monolayer cultures. This report gives some of the conditions, including those for tumor cell dissociation, that influence the growth of mammary tumor cells and the concentration of prolactin binding sites in vitro.

MATERIAL

AND METHODS

Materials HI-WOs/BAzooo Medium was obtained from International Scientific Industries. All other media, fetal calf serum, penicillin-streptomycin-fun~zone mixture and glutamine were obtained from Gibco. Gentamicin and perforated cellophane were from Microbiological Associates; ovine prolactin (NIH-P-S1 2) and all pituitary hormones were gifts from NIAMDD. Insulin, triiodothyronine, steroid hormones (except R5020), DMBA, lactoperoxidase and hyaluronidase Type I were from Sig ma Chemical Co. Collagenase Type III (Worthington Biochemical Co.) was from batches 46BOO7X and 46DO84 and had activities of 148 U/mg and 100 Ulmg, respectively. t3H]~stradiol was from Amersham; dimethyl-19-norpregna-4,9-diene3,20-dione (R5020) and t3H]R5020 were from New England Nuclear. Tissue culture labware was made by Falcon Products.

Preparation of mummap tumors Mammary tumors were induced in 50- to 55day-old virigin female SpragueDawley rats; each rat was treated with 20 mg of DMBA in peanut oil, given by gavage (Huggins et al., 1959). After 2-3 months, tumor size was measured with calipers twice weekly. When the tumor area reached 2-5 cm’, the animal was killed by cervical dislocation. Under aseptic conditions, the tumor was removed and freed of debris; the excised tumor was cut into 0.5 mm-thick sections with a Stadie Riggs Tissue Slicer (Arthur H. Thomas Co.) and then minced finely with scissors. One portion of the minced tissue was saved for histologic examination; another portion (0.5 g) was frozen in liquid Na and stored at -70°C. The remaining minced tissue was weighed and the cells were dissociated.

Dissociation of tumor tissue Initi~y, previously described methods using collagenase, hyaluronidase, pronase and trypsin, alone or in various combinations (Builough and Wallis, 1974; Harmon and Hilf, 1976; Hosick and Nandi, 1974; Pitelka et al., 1951; Smith et al., 1976) were investigated but none were totally satisfactory for DMBA-induced mammary tumors. It was found that the duration of enzyme treatment needed to completely dissociate tumor tissue could not be predicted in advance unless the tissue had been

Prolactin receptors in cultured mammary tumors

83

uniformly sliced and minced. Second, dissociation time was prolonged if the minced tissue was not previously washed. Third, trypsin or pronase often caused extensive cell damage. For these reasons, alternative methods with collagenase and hyaluronidase were evaluated. Minced tumor tissue was suspended in 50-100 ml of Leibowitz L-15 medium containing 1% bovine serum albumin (BSA), 2 mM glutamine, gentamicin (25 yg/ ml), penicillin (100 U/ml), streptomycin (100 pug/ml) and fungizone (25 pg/ml). (Unless otherwise stated, bovine prolactin (1 pg/ml) was added to this and all other media used for the preparation of cells for culture.) The tissue s,uspension was then rocked gently for 40 min at 25’C. After the tissue fragments settled, the cloudy supernatant was discarded and various concentrations of collagenase or hyaluronidase (or both) in Medium 199 containing 1% BSA and 20 mM N-2-hydroxyethylpiperazine-N’-ethanesulfonic acid buffer (pH 7.4) plus antibiotics was added (5 ml/g of tissue). This suspension was incubated with shaking for 1 h at 37’C. The cloudy supernatant was then removed and discarded, and fresh enzyme solution was added. This suspension was drawn in and out of a sterile lo-ml disposable pipette until the tissue particles passed freely through the pipette tip. The suspension was incubated for another hour at 37’C, and again dispersed with a pipette, and filtered through organza cloth. Residual cells in the bottle and cloth were recovered and then filtered. The collective sample of filtered cells was sedimented at 200 g for 5 min, and washed twice with L15 medium. The cells were counted in a particle counter (Particle Data Co.) and cell viability was determined by trypan-blue dye exclusion. For some experiments, a sample of the preparation (about 2 X 10’ cells in pellet form) was frozen in liquid Nz and stored at -70°C. Procedure for culturing tumor cells

Suspended cells from the above procedure were distributed to culture flasks: 0.7-I X 10’ cells per flask with a growth area of 25 cm2 or 2-3 X 10’ cells per flask with a growth area of 75 cm’. Since medium had not been specifically formulated to support rat-mammary tumor growth in vitro, commercially available media were tried. Cells were grown in several different media supplemented with 2 mM glutamine, antibiotics, 1 PM insulin, 1 nM estradiol, 1 nM triiodothyronine, 0.1 PM progesterone and 0.1 PM corticosterone (basal media). Where indicated in the results, the media were also supplemented with ovine prolactin (500 ng/ml), and/or given amounts of fetal-calf serum, stripped of steroid hormones by incubation at 55’C for 30 min with dextran-coated charcoal (Zava et al., 1977). Stock solutions of steroid hormones (10 OOO-fold concentrated) were prepared in 95% ethanol and stored at -20°C. Insulin, triiodothyronine and prolactin (1 OOO-fold concentrated) were prepared in 25 mM sodium phosphate, 0.15 M NaCl, pH 7.4, and stored in small aliquots at -20°C. (Dilute NaOH was added to facilitate solution of the insulin and triiodothyronine.) All additions to the media were made just prior to use.

84

The cultures were incubated at 37°C in air (all other media) and the medium was changed or 3 days. At the time of harvest, monolayers scraped from the plastic flasks with perforated fugation at 200 g for 5 min. The pelleted cells in liquid Nz and stored at -7O’C.

M.E. Costlow, P.E. Gallagher, Y. Koseki

(L-15 medium) or air pfus 5% CO, 24 h after plating and then every 2 were washed twice in L-15 medium, cellophane and collected by centrifrom each culture flask were frozen

Prolactin binding assay Ovine prolactin was labeled with “‘1 in the presence of lactoperoxidase and purified by DEAE-cellulose chromatography (Costlow and Gallagher, 1977). The labeled hormone had a specific activity of 2325 -t 286 cpm/fmol (mean 5 S.E. for 9 samples; 70% counting efficiency). Iodinated prolactin prepared in this manner has the same binding properties as those of native prolactin (Costlow et al., 1976a); binding is selective for prolactin target cells (Costlow and McGuire, 1977a) and is inhibited by only lactogenic hormones (Costlow et al., 1975b). At assay, frozen minced tumors, dissociated cell suspensions and cultured cells were warmed to 4°C in 0.5-1.0 ml of incubation buffer (25 mM sodium phosphate, 10 mM MgCls, 0.1% BSA; pH 7.0) and homogenized in a glass-glass Dual homogenizer. Homogenates of tumors and cultured cells were used instead of partially purified membranes in order to minimize possible differences due to subcellular fractionation. Each 0.1 ml of homogenate was then added to a 12 X 75 mm polystyrene tube with0.1 ml of incubation buffer and ‘251-labeled prolactin ([“‘I] oPr1, 80 000 cpm) with or without unlabeled prolactin (200 ng). After incubation with shaking for 18 h at 22°C each sample was diluted with 2.5 ml of ice-cold 10 mM sodium phosphate (PI-I 7.0) and centrifuged at 20 OOOg for 10 min. The supernatant was discarded, each pellet was washed with 10 mM sodium phosphate buffer and the amounts of bound [ ‘251]oPr1, protein (Lowry et al., 1951) and DNA (Burton, 1956) were determined. Specific prolactin binding was defined as the difference in bound [ “‘1 ] oPr1 between samples incubated with vs. without excess unlabeled hormone. The binding of [ 1251]oPr1 to tumor- and cultured-cell homogenates was directly proportional to the amount of protein (20-200 ,ug) in the assay (data not shown). Nonspecific binding in some experiments was relatively high (>50%) and/or specific binding was low (
Prolacfin receptors in cultured mammary tumors

85

all cultures were fed basal medium with 1% serum (+prolactin) that lacked estradiol and progesterone. 4-6 flasks (growth area 7.5 cm2 each) were used for each assay point (in triplicate). The culture in each flask was washed twice in Hank’s balanced salt solution and then in 2 ml of phosphate buffer (5 mM sodium phosphate, 1 mM thioglycerol, 10% glycerol, pH 7.4, at 4’C). Cells were scraped from the flasks with perforated cellophane, pooled, homogenized and centrifuged at 10.5 000 g for 50 min. The clear supernatant (cytosol) was removed, diluted to 1 .O mg of protein/ ml, and incubated at 4°C for 2 h with 2 X lo-’ M estradiol or 4 X lo-* M R.5020. Each of the cytosols (200 ~1) was added to 200-400 ~1 of ice-cold hydroxylapatite suspension (50% v/v), then vortexed for 15 min before the receptor-containing precipitate was sedimented at 800 g for 5 min. Each precipitate was incubated for 17 h at 30°C in phosphate buffer with 10 nM [3H]estradiol with or without 1 PM estradiol or 20 nM [3H]R5020 with or without 2 PM R5020. (This step permits complete ligand exchange; Koseki and Costlow, unpublished results.) The precipitates were then washed 3 times at 4°C in 50 mM Tris, 1 mM potassium phosphate (monobasic), and 1% Tween 80 (v/v), pH 7.5. The bound radioactivity was extracted with 2.0 ml of ethanol (23”C, overnight). Each extract was mixed with 5 ml of ACS scintillation fluid (Amersham) and counted. Specific binding was the difference in bound radioactivity between samples incubated with and without excess unlabeled steroid.

RESULTS Dissociation of tumor cells for culture

A combination of collagenase and hyaluronidase (1 mg/ml each) gave the highest cell yield; increased enzyme concentrations (2 mg/ml) or collagenase alone (2 mg/ ml) reduced the yield (Table 1). Hyaluronidase alone at 2 mg/ml was nearly as effective as both enzymes (1 mg/ml each) although cell yields were not consistently high. In 2 other experiments, the yield of viable cells treated with hyaluronidase (2 mg/ml) was 50 and 27 X lo6 per g of tissue compared to 140 and 64 X lo6 cells per g of tissue separated with collagenase and hyaluronidase (1 mg/ml each). Specific prolactin binding was 26 and 30 cpm/ng protein (hyaluronidase alone) and 33 and 38 cpm/pg protein (collagenase/hyaluronidase), respectively. The combination of collagenase and hyaluronidase (1 mg/ml each) was adopted for the routine preparation of tumor cells for culture. With this procedure, cell viability was consistently 85-90%, with more than 90% dissociation of tumor (by wet weight). Yields from 1 g of tissue varied from 6 X IO6 to 1.4 X lo8 cells with a mean yield of 5 X 10’ viable cells (n = 28 tumors). Typical cell suspensions contained many single cells as well as small clumps (Fig. 1a). Tumor cell culture

Tumor cells showed poor plating efficiency

and little growth in Ham’s F-l 2 Me-

M.E. Costlow, F.E. ~a~la~~er, Y. Koseki

86

Table 1 Prolactin binding and cell yield fram mammary tumors dissociated with collagenase and hyaluronidase Sliced, minced tumor tissue was divided into portions, weighed and dissociated into cell suspensions by enzymes at the indicated concentrations (without unlabeled prolactin). The viable-cell yield and specific prolactin binding activity were then determined. Nonspecific binding was 5 1 cpm/pg of protein. Protein averaged 78 pg/assay. Enzyme comb~ation Collagenase (mglml)

Hyaluronidase (mslml)

1.0 2.0 2.0 0.0 0.0 0.5 1.0 0.0

2.0 0.0 2.0 1.0 1.0 0.5 0.0

1.0

Cell yield (lo6 X cells/g)

Viability (%}

Viable cell yield (IO6 X cells/g tissue)

Specific prolactin binding (cpmltig protein)

72 41 21 61 51 36 32 11

87

63 35 9.5 55 37 26 24 2.8

330 11.5 170 256 141 101 121 192

85 35 90 72 72 7.5 23

dium, Serumless Medium or ~I-WO~~B~~~~ which all lacked serum but had prolactin. In contrast, cells placed in Minimum Essential Medium, Medium 199, Leibowitz L15 medium, Ham’s F-12 or RPM1 1640 (each with prolactin and 1, 4 or 10% charcoal-stripped fetal calf serum) adhered to the bottom of the flasks within 6-8 h. After the first 24 h of culturing, the plating efficiency was about 50% at any serum concentration; the adhering cells were predominantly epithelial-like and were spread out (Fig. lb), covering lo-20% of the total surface area. All media appeared to support growth to a greater or lesser extent, dependiirg on the tumor specimen used; a monolayer of actively growing cells formed within 3-5 days (Fig. 1c). Prolactin binding The enzymes used to dissociate mammary tumors markedly altered the level of prolactin binding. The level of specific binding was highest for uncultured cells dissociated with collagenase and hyaluronidase (1 mg/ml each), but was not directly related to the yield of viable cells (Table I). Autoradiographs of cultured cells incubated with [ rz51]oPr1 showed grains over epithelial-like (tumor) cells but not over fibroblasts (Fig. Id). The number of grains over tumor cells exposed to [1251]oPrl and excess unlabeled prolactin was comparable to the background (not shown). The level of prolactin binding in cultured tumor cells varied with the type of culture medium (Table 2). Independent of the serum concentration, the level of prolactin receptors in cells grown on L-l 5 medium was much higher than that for Minimum Essential Medium, Medium 199 or Ham’s F-12 but only slightly higher than

Prolactin receptors in cultured mammary tumors

87

Fig. 1. DMBA-induced rat-mammary tumor cells in culture. (a) Phase-contrast photomicrograph of collagenase/hyaluronidase-dissociated tumor cells. (b) Phase-contrast photomicrograph of tumor cells after 24 h in culture. (c) DNA synthesis in cultured tumor cells grown on 1 X 3-inch glass slides for 3 days in basal L-15 medium with 1% serum and prolactin. After an incubation of 4 h in L-15 medium with [ 3H]thymidine (0.5 &i/ml), the cells were washed, fixed in Carnoy’s fixative, and processed for autoradiography. (d) Prolactin binding to cultured tumor cells grown on 1 X 3-inch glass slides for 7 days in basal L-15 medium and 1% serum. 2 days after prolactin was removed from the medium, the cells were incubated in L-15 medium with 0.1% BSA, 10mM MgClz, and [ 1251]oPrl (400 000 cpm/ml). After 4 h at 22”C, the cells were washed, fixed in Carnoy’s fixative and processed for autoradiography. Note the difference in appearance between epithelial- and fibroblast-like cells in the sparse area of the culture. (Phase contrast; line = 100 firn)

that for RPM1 1640 or Waymouth’s 752/l (not shown). Although tumor cells dissociated with collagenase and hyaluronidase and cultured in L15 medium retained specific-binding ability, prolactin binding in most cultured cells was less than that in the original tumor (Fig. 2a). This loss was independent of the original level of prolactin binding and was prevented by the addition of a saturating amount (1 pg/ml) of unlabeled bovine prolactin to the medium during the washing- and enzyme-dissociation steps (Fig. 2b). On the average, the level of prolactin binding in both cell suspensions and cultured cells decreased by

Tumors:

Ham’s F-12

Leibowitz L-15 Minimum essential medium Medium 199

Growth medium

186 A.9 97.4 ). 0.5

81.4 f 4.0

22.3 f 0.2 8.9 * 0.1 49.8 f 2.0

C

B

A

1% serum

Specific prolactin binding (cpm/,ug protein)

74.1 i: 4.3 33.3 e 0.8 26.9 f 0.5

43.9 r 0.2 19.9 f 0.3 21.3 f 1.7 19.9 t 2.0

E

D

4% serum

.p

50.0 * 2.1

2

Q ::

70.0 i: 3.0 32.5 f 2.3 40.0 c 0.7

F

10% serum

Table 2 Effect of growth medium on prolactin binding in primary cultures of mammary tumors 6 tumors were dissociated and cultured in the indicated basal medium supplemented with prolactin and serum. After 7-9 days, prolactin was removed from the cultures and 2 days later, the cel.ls were harvested. The specific [ f25110Pri binding per pg of protein was determined. AU cultures except those in L-15 medium were incubated in air with 5% Cot. The amount of protein in each sample for tumors A-F was 38,33,65,23, 35 and 50 yg, respectively. The level of nonspecific binding averaged 46.5 cpm/pg of protein. Values are the mean f S.E. for triplicate samples. PIOlactin binding in tumors cultured in L-15 medium vs. each of the other media was significantly different @ < 0.02; Student’s t test).

89

Prolactin receptors in cultured mammary tumors

T UMOR

c”dily

1‘UMOR

J “%LY”L”s’”

Fig. 2. Effect of prolactin during enzyme dissociation on prolactin binding in cultured mammary-tumor cells. Prolactin binding was measured for untreated minced tumors (0 and l; panels a and b) and for cultured cells that had been dissociated without (0; panel a) or with (o; panel b) bovine prolactin (1 gg/ml). Dissociated tumor cells were grown for 7-10 days in basal L-15 medium with 1% serum and prolactin. 2 days prior to harvest, prolactin was removed from the medium. Each point indicates the mean specific binding for triplicate assays of tumors before dissociation or after culture. A line was drawn between data points for each tumor.

60% when prolactin was absent during tumor cell dissociation (Table 3). (The level of prolactin binding in cell suspensions prepared with prolactin could not be determined since all available binding sites were occupied by unlabeled hormone.) Similar results were obtained when prolactin binding levels were based on DNA (rather than protein) content of the tumor and cultured cells. In a series of 8 tumors dissociated in the presence of bovine prolactin, prolactin binding was 133 f 38 cpm

Table 3 Effect of prolactin during enzyme dissociation on specific prolactin binding in mammary-tumor cell suspensions and cultures Experimental details are the same as in Fig. 2. Values are the means f S.E. of specific binding per @g protein for N tumors. -, not determined. Significance of differences 07 values, paired data) between binding in tumors vs. cell suspensions (without prolactin),
Without prolactin With prolactin

Specific prolactin binding (cpm/pg protein) Tumors

Cell suspensions

Cultured cells

N tumors

142 t 19 126 + 16

73 f 12 -

662 9 115 * 15

12 16

M.E. Costlow, P.E. Gallagher, Y. Koseki

90

per pg of protein and 834 + 141 cpm per 1.18of DNA; after culturing, the binding was 652 + 112 cpm per pg of DNA and 111 + 18 cpm per bg of protein. Thus, prolactin-binding levels in primary cultures remained comparable to that in the original minced tumor tissue as long as prolactin was included during tumor dissociation. Cells cultured for 7-8 days retained 94 +_13% of tumor prolactin binding (n = 15 tumors) and cells cultured for 9-10 days retained 106 f 16% (n = 5 tumors). Effect of the prolactin level in vitro on prolactin binding in cultured mammary tumors Many studies suggest that the level and/or the availability of prolactin receptors in vivo are altered by changes in the level of circulating prolactin (Costlow et al., 197.5; Holdaway et al., 1977; Kelly et al., 1974; Posner et al., 1974, 1975). When cultured tumor cells were grown with increasing concentrations of prolactin, the availability of binding sites progressively decreased (Fig. 3). The half-maximum inhibition of [ 1251]oPr1 binding by unlabeled prolactin was 100 ng/ml for cells grown in 5-1000 ng of prolactin/ml. Except for quantitative differences, these results are similar to those of Holdaway et al. (1977), who found that prolactin binding in mammary tumors and liver was significantly inhibited when circulatinghormone concentrations exceeded 300 ng/ml. After prolactin removal from the growth medium, the number of available binding sites progressively increased, reached a maximum at 48 h and then declined

ng PROLACTIN

/ml

Fig. 3. Effect of prolactin on prolactin binding to cultured mammary tumor cells. Dissociated mammary tumor cells were grown for 6 or 7 days in L-15 medium with 1% serum and prolactin (200 ng/ml). The cultures were then switched to medium containing the indicated concentrations of prolactin. The medium was replaced after 48 and 96 h and cells were then harvested on day 12 or 13. Homogenates of these cells were assayed for specific prolactin binding. The data are plotted as the mean percent inhibition (G.E.) in specific binding by the indicated concentrations of unlabeled prolactin. The number of tumors assayed is indicated in parentheses.

Prolactin receptors in cultured mammary tumors

91

(Fig. 4). In the continued presence of prolactin (500 ng/ml), the level of prolactin binding remained low over the 48-h period. However, when a low level of ovine prolactin (0.5 or 1 ng/ml) was included in the medium, the level of binding in 3 of 4 cultured tumors rose above that of controls incubated without the hormone (Table 4). Characteristics of prolactin binding to cultured cells

Fig. 5 shows that a steady state in specific prolactin binding was achieved by 12 h of incubation at 22°C and was maintained for up to 22 h. [The same time-course of binding was noted for tumor homogenates (data not shown).] Optimum binding was at a pH of 7.0 and decreased by about 30% at pH extremes of 6.6 and 8.0 (data not shown). Binding sites in cultured tumor cells were specific for hormones with lactogenic prop&ties (ovine, bovine and rat prolactin, and human placental lactogen) (Table 5). Scatchard analysis of prolactin binding (Scatchard, 1949) indicated that cultured tumor cells had a single class of prolactin binding sites with a Kd of 1 .l X 10V1’ M (Fig. 6).

TIME

(hours)

Fig. 4. Effect of prolactin on prolactin binding in cultured mammary tumors. Dissociated mammary-tumor cells were grown on basal L-15 medium supplemented with 1% serum and prolack Then, at 0 h, the prolactin in some cultures was removed while in others, prolactin remained. At the indicated intervals, cells were harvested and the amount of specific prolactin binding per rg of protein determined. So that data from a number of tumors could be compared, prolactin binding for cells cultured without prolactin for 48 h was given an arbitmry value of 100 and the other data points were normalized accordingly. (Specific [‘251]oPrl binding for these tumors at 48 h was 107, 43.9, 125 and 120 cpm/pg of protein). The figure shows the mean f S.E. of the normalized data for the number of tumors (indicated in parentheses) tested at each interval. Nonspecific binding averaged 67 cpm/yg protein.

M.E. Costlow, P.E. Gallagher, Y. Koseki

92

Table 4 Cells were grown for 7-10 days on basal L-15 medium with 1% serum and prolactin 2 days prior to harvest, the cells were fed medium with the indicated concentrations of prolactin. Each point is the mean f SE. for triplicate determinations. Tumors Es. Is and T3 has significantly more prolactin binding 0, < 0.05, two-tailed Student’s t test) when cultured with prolactin. Nonspecific binding averaged 20 cpm/ug protein. Tumor

Prolactin level (nglml)

Specific prolactin binding (cpm/ng protein)

Percent control (no prolactin)

E3

0 0.5 1.0

27.0 f 1.6 40.0 + 0.7 39.0 f 2.8

148 144

0 0.5 1.0

12.9 f 0.3 30.3 f 1.7 23.1 f 0.9

234 179

0 0.5 1.0

78.0 + 5.6 68.3 f 3.5 75.0 f 0.1

87 96

0 0.5

35.7 f 0.7 46.4 f 0.9

130

I3

J3

T3

-

60

-

TIME

(hours)

Fig. 5. Time course of prolactin binding to cultured rat-mammary tumor cells. Tumor cells were incubated for 7 days in L-15 medium with 1% serum and prolactin. 2 days prior to harvest, prolactin was removed from the medium. Homogenates from these cells were incubated with [1251]oPrl (400 000 cpm/ml), with or without unlabeled prolactin (1000 ng/ml). At the indicated intervals, aliquots (0.2 ml, 139 pg of protein) were removed, and specific prolactin binding per wg of protein determined. Each point represent results for a single assay. Nonspecific binding averaged 37.1 cpm/pg protein.

Prolactin receptors in cultured mammary tumors

93

Table 5 Hormone specificity of prolactin binding to primary cultures of mammary tumors Tumor cells were grown in basal L-15 medium with prolactin and 1% serum. Homogenates of these cells (84 ng of protein per sample) were then incubated for 16 h with [ ‘251]oPrl with or without the indicated unlabeled hormones. Nonspecific binding was 2209 cpm and total binding was 6081 cpm. Competition by unlabeled ovine prolactin was taken as 100%. Each value is the mean f SE. for triplicate samples. Competitor (2 r*g/ml)

Percent competition

Ovine prolactin Rat prolactin Bovine prolactin Human placental lactogen Rat follicle-stimulating hormone Insulin Rat growth hormone Rat lutropin

100 f 4 88 * 1 92 f 2 87 f 5 Ok1 Ok1 Ok1 0+2

Kd = I.1 X IO-“M SITES = 52 fmoles/mg

5 PROLACTIN

IO

BOUND

protein

I3

(nM)

Fig. 6. Scatchard analysis of prolactin binding to cultured mammary-tumor cells. Dissociated tumor cells were grown for 7 days in basal L-15 medium with 1% serum and prolactin. 2 days prior to cell harvest, prolactin was removed from the medium. Homogenates of these cells were then incubated with 20 000, 40 000 cpm [‘251]oPrl, and with 40 000 cpm [‘251]oPrl plus 0.125, 0.25, 0.5, 1.0 or 2.0 ng of unlabeled prolactin. Nonspecific binding (36% of total) was determined by incubating the homogenate with 40 000 cpm [1251]oPrl and 256 ng of unlabeled prolactin. Following all additions, the incubation tubes (in triplicate for each point) were counted for total radioactivity. The data were corrected for nonspecific binding according to the method of Chamness and McGulre (1975) and then plotted according to the method of Scatchard (1949). The lines were fitted by least-squares regression analysis. The correlation coefficient (r) for the line is 0.972.

M.E. Costlow, P.E. Gallagher, Y. Koseki

94

Table 6 Effect of prolactin on estrogen and progestin binding to cultured mammary-tumor cytosol Each value is the mean + S.E. for triplicate samples of 1 lday-old cultures, which had (+) and had not (-) been deprived of prolactin for the last 4 days of culture. Specific procedures for culturing the cells for these assays are given in “Materials and Methods”. Tumor

Prolactin

Specific binding (fmol/mg cytosol protein) [3H]Estradiol

x4

+

y4

+

-

71.5 t 3.6 28.2 f 0.5 80.6 f 6.7 64.7 +-2.0

13H]R5020 78.0 + 2.8 16.6 t 0.3 332 189

+ 33 + 54

Response of cultured cells to prolactin Initial assessments of this culture system in studies of prolactin regulation of tumor growth and function have yielded the following results. 2 days after the addition of prolactin to 7-day-old cultures, the percentage of 10 000 nuclei that were labeled with t3H]thymidine had increased 3-fold over that for parallel cultures incubated without prolactin (2.7% and 0.89%, respectively; P < 0.02, two-tailed Student’s t test). Removal of prolactin from cultures for 4 days resulted in a decrease in specific binding for [3H]estradiol and [3H] R5020 (Table 6).

DISCUSSION These studies show that DMBA-induced rat-mammary tumors can be grown in primary monolayer cultures, retain prolactin-receptor sites and respond to prolactin. It is clear from our results that the conditions used for both dissociating and culturing these tumors have a profound effect on the expression of prolactin-receptor sites in vitro. A number of enzyme combinations were capable of dissociating solid tumors into viable cell suspensions, but even with the most effective combination in our study, prolactin-receptor levels were decreased in cultured cells unless a saturating amount of prolactin was present. This finding indicates that some receptor-degrading proteolytic activity may be present in the crude collagenase and hyaluronidase. Two explanations could have accounted for the sustained reduction of receptor in cultured cells that had been dissociated in the absence of prolactin. With some of the receptor sites destroyed, the tumor cells might have been incapable of resynthesizing receptors under our culture conditions. If this were true, the level of receptors in cells cultured for longer periods (10 days) should have been lower than that for cells cultured for shorter periods (3 days). This, however, was not the case (unpublished results). Alternatively, since some of these tumors comprise cells with and without receptors (Costlow and McGuire, 1977b), it is likely that cell cultures

Prolactin receptors in cultured mammary tumors

95

of such tumors remain heterogeneous. Reducing the level of receptors and/or omitting prolactin during dissociation may reduce the rate of growth of prolactindependent cells relative to that of autonomous cells. Under such circumstances, the apparent level of prolactin binding for the entire culture would remain low despite the synthesis of new receptor sites. A direct comparison of receptor-containing cells in tumors and cultured monolayers will be necessary to substantiate this notion. The varied effects of different stock media on the number of prolactin receptors in cultured tumor cells suggests that nutrients as well as hormones are important for the expression of prolactin receptors in culture. Since the growth media tested represent varied concentrations of complex mixtures of amino acids, vitamins, salts and other nutrients, we do not yet know which components, whether singly or in combination, are responsible for the effects on receptor levels. Possibly a general class of nutrient components is involved, since the amino acid concentrations of L15 medium are 2- to IO-fold higher than those in many of the other media tested. We chose to supplement the growth medium with several hormones (in addition to prolactin) which are known to affect monolayer cultures of a continuous cell line of human breast cancer (MCF-7) (Burke et al., 1977; Lippman et al., 1976a, b; Rillema and Linebaugh, 1977). At present, we are determining if all these hormones are required or are at optimal concentrations in the growth medium for the maintenance of DMBA-induced mammary tumors. The half-time occupancy of prolactin receptor has been estimated in two reports. Birkinshaw and Falconer (1972) found that [ *2sI]oPr1 injected into rabbitmammary glands was lost at a half-time rate of 52 h. Shiu and Friesen (1976) on the other hand, found that 5 h at 37’C was required for excess unlabeled prolactin to displace 50% of the receptor-bound [ ‘251]oPr1 from plasma membranes of rabbit mammary glands. After prolactin was removed from cultured cells, we found that 18 h was required for 50% of the occupied prolactin binding sites to become available. However, after the concentration of available sites reached a maximum at 48 h, receptor content declined, consistent with a dependence of prolactin receptors on the continued availability of prolactin (Costlow et al., 1975; Posner et al., 1974, 1975). Thus, the increasing availability of receptor sites following the removal of prolactin from mammary tumors in culture may be the combined effect of(i) the release of endogenously bound hormone, (ii) the loss of receptors in response to the paucity of exogenous hormone, and (iii) the synthesis of new receptors. That the synthesis of new receptors occurs under this condition is supported by the observation that a 48-h treatment with very low levels of prolactin in the growth medium results in an apparent increase in the level of binding. Further experiments with metabolic inhibitors will be necessary to determine whether the apparent increase during culture is due to de novo synthesis of receptors. The characteristics of prolactin binding with respect to pH optimum, hormone specificity, time-to-steady-state in binding, and to dissociation constants are essentially the same as those for rat tumors assayed directly after excision (Costlow et

96

ME.

Costlow, P.E. Gallagher, Y. Koseki

al., 1976a). Further, the wide variation in receptor levels in cultured cells is comparable to that in tumors taken directly from rats (Costlow et al., 1976a; DeSombre et al., 1976; Smith et al., 1976). With respect to the effect of prolactin on DNA synthesis in cultured tumor cells, our results, although not as extensive, confirm the findings of Hallowes et al. (1977) and Rudland et al. (1977). In addition, since prolactin can stimulate estrogen receptors in these tumors in vivo (V&non and Rochefort, 1976; Hawkins et al., 1977) and since progesterone-receptor synthesis in DMBA-induced tumors requires estrogen (Horwitz and McGuire, 1977) the loss of t3H]estradiol and [3H]R5020 binding following prolactin removal suggests that this model will be useful in defining the role of various hormones (and receptors) that control tumor growth and function.

ACKNOWLEDGEMENTS This research was supported by American Cancer Society Grants BC 247, IN 99C and IN 99E; Biomedical Research Support Grant RR05584 from NIH; Cancer Center Support (CORE) Grant CA 21765 from NCI; and by ALSAC. We thank Ms. Amie Hample for technical assistance, Ms. Patricia Nicholas for typing the manuscript and Ms. Jane Seifert for editing the manuscript.

REFERENCES Aragona, C., Bohnet, H.G. and Friesen, H.G. (1976) Endocrinology 99, 1017-1022. Birkinshaw, M., and Falconer, I.R. (1972) J. Endocrinol. 55, 323-334. Bullough, W.A., and Wallis, M. (1974) J. Endocrinol. 62,463-472. Burke, R.E., Zava, D.T., and McGuire, W.L. (1977) Clin. Res. 25,461a. Burton, K. (1956) Biochem. J. 62,315-323. Chamness, G.C., and McGuire, W.L. (1975) Steroids 26, 538-542. Costlow, M.E., and Gallagher, P.E. (1977) Biochem. Biophys. Res. Commun. 77, 90.5-911. Costlow, M.E., and McGuire, W.L. (1977a) J. Endocrinol. 75,221-226. Costlow, M.E., and McGuire, W.L. (1977b) J. Natl. Cancer Inst. 58, 1173-1175. Costlow, M.E., Buschow, R.A., and McGuire, W.L. (1975a) Life Sci. 17,1457-1465. Costlow, M.E., Buschow, R.A., Richert, N.J., and McGuire, W.L. (1975b) Cancer Res. 35, 970-974. Costlow, M.E., Buschow, R.A., and McGuire, W.L. (1976a) Cancer Res. 36, 3324-3329. Costlow, M.E., Buschow, R.A., and McGuire, W.L. (1976b) Cancer Res. 36, 3941-3943. Costlow, M.E., Sluyser, M., and Gallagher, P.E. (1977) Endocrine Res. Commun. 4, 285-294. DeSombre, E.R., Kledzik, G., Marshall, S., and Meites, J. (1976) Cancer Res. 36,354-358. Hallowes, R.C., Rudland, P.S., Hawkins, R.A., Lewis, D.J., Bennett, D., and Durbin, H. (1977) Cancer Res. 37, 2492-2504. Harmon, J.T., and Hilf, R. (1976) Biochim. Biophys. Acta 443,114-125. Hawkins, R.A., Hill, A., Freedman, B., Killen, E., Buchan, P., Miller, W.R., and Forrest, A.P.M. (1977) Eur. J. Cancer 13,223-228. Holdaway, I.M., and Friesen, H.G. (1976) Cancer Res. 36,1562-1567.

Prolactin receptors in cultured mammary tumors

97

Holdaway, I.M., Deegan, M., and Friesen, H.G. (1977) Can. J. Physiol. Pharmacol. 55, 193195. Horwitz, K.B., and McGuire, W.L. (1977) Cancer Res. 37, 1733-1738. Hosick, H.L., and Nandi, S. (1974) Exp. Cell Res. 84,419-425. Huggins, C., Briziarelli, G., and Sutton, Jr., H. (1959) J. Exp. Med. 109, 25-42. Kelly, P.A., Bradley, C., Shiu, R.P., Meites, J., and Friesen, IX. (1974) Proc. Sot. Exp. Biol. Med. 146, 816-819. Kim, U., and Furth, J. (1976) Vitam. Horm. (N.Y.) 34, 107-136. Kledzik, G.S., Bradley, C.J., Marshall, S., Campbell, G.A., and Meites, J. (1976) Cancer Res. 36, 3265-3268. Lippman, M.E., Bolan, G., and Huff, K. (1976a) Cancer Res. 36, 4595-4601. Lippman, M.E., Bolan, G., and Huff, K. (1976b) Cancer Res. 36,4602-4609. Lowry, O.H., Rosebrough, N.J., Farr, A.L., and Randall, R.J. (1951) J. Biol. Chem. 193, 265275. McGuire, W.L., Chamness, G.C., Costlow, M.E., and Horwitz, K.B. (1977) In: HormoneReceptor Interactions: Molecular Aspects, Ed.: G.S. Levey, (Marcel-Dekker, New York) pp. 265-299. Pitelka, D.R., Kerkof, P.R., Gangne, H.T., Smith, S., and Abraham, S. (1969) Exp. Cell Res. 57, 43-62. Posner, B.I., Kelly, P.A., and Friesen, H.G. (1974) Proc. Natl. Acad. Sci. (U.S.A.), 71, 24072410. Posner, B.I., Kelly, P.A., and Friesen, H.G. (1975) Science 188, 57-59. Rillema, J.A., and Linebaugh, B.E. (1977) Biochim. Biophys. Acta, 475, 74-80. Rudland, P.S., Hallowes, R.C., Durbin, H., and Lewis, D. (1977) J. Cell Biol. 73, 561-577. Scatchard, G. (1949) Ann. N.Y. Acad. Sci. 51,660-672. Shiu, R.P., and Friesen, H.G. (1974) Biochem. J. 140, 301-311. Shiu, R.P., and Friesen, H.G. (1976) Science 192, 259-261. Smith, R.D., Hilf, R., and Senior, A.E. (1976) Cancer Res. 36, 3726-3731. Vignon, F., and Rochefort, H. (1976) Endocrinology 98, 722-729. Welsch, C.W., and Nagasawa, H. (1977) Cancer Res. 37, 951-963. Zava, D.T., Chamness, G.C., Horwitz, K.B., and McGulre, W.L. (1977) Science 196, 663-664.