IN VITRO Volume 14, No. 5, 1978 All rights reserved 9
EFFECT OF SERUM ON CELLS CULTURED FROM HUMAN MAMMARY TUMORS E. V. GAFFNEY ~ ANDD. PIGOTT
The Imperial Cancer Research Fund Laboratories, Lincoln's Inn Fields, London, England
SUMMARY
Clusters of cells derived from biopsy specimens of human mammary ductal carcinomas form two morphologically distinct epithelial colonies in culture, designated as E and E'. The proportion of E' cell clusters that attached and formed colonies ranged from 0.3 to 13.0% with different tumors. Attachment was independent of tumor grade. Microscopic observations revealed that the survival of E' cell colonies was limited to approximately 10 days with rapid cell degeneration commencing about 7 days. A comparison of sera showed that colony formation by cells from malignant tumors during the 1st week of culture was maximum in the presence of fetal bovine serum. Human serum alone was 70 to 100% less effective in promoting E' colonies. The most significant finding was that human serum from normal donors inhibited E' colony development in the presence of FBS. Although human serum was less effective than FBS in promoting colony formation by clusters of E cells, an inhibition was not observed. Inhibitory activity could not be attributed to either antagonistic hormones or the source of human serum. These results demonstrate that normal human serum contains a factor{s) that exhibits an inhibitory activity specific for human epithelial cells (E') derived from malignant tumors.
Key words: mammary; human; sera; colony formation. INTRODUCTION
Most studies of malignant human mammary cells are limited to the few lines isolated from both primary and metastatic breast tumors. These
studies provide important information on cell growth and morphology (1,2) as well as tumor etiology (3,4). However, it would be more meaningtul for the individual patient if investigators were routinely able to culture malignant cells from tumor specimens. Unfortunately, there has been limited success deriving even primary cultures from the majority of samples {5). This failure stems in part from our lack of knowledge concerning both the identity of the cell types observed in culture and their nutritional requirements essential for long-term maintenance. A recent report established that two distinct epithelial cell types, designated as E and E', can be observed in cultures established from breast
carcinomas (6). Both cell types (E and E') demonstrated ultrastructural properties typical of epithelium, but differed in general cell shape and growth pattern. Cells similar in morphology and arrangement to E cells also were isolated from lacteal secretions (7), benign dysplasias (8) and reduction mammoplasties (9). In contrast E' cells were isolated solely from malignant tumors. Thus it was suggested by Hallows et al. (6) that E cells may arise from residual normal or hyperplastic epithelium originating in tumor-associated breast tissue and that E' cells are derived from neoplastic epithelium. The current study serves two purposes: first, to quantitative cell yield from tumors and survival time in primary culture as a basis for investigating nutritional factors that may enhance division potential; and second, to establish the response of mammary cells to different sera. MATERIALS AND METHODS
Histological grading of ductal carcinomas, which predicts the relative aggressiveness of the tumor, was based on the criteria of Bloom and 451
~Request for reprints should be sent to: Dr. E. V. Gaffney, Pennsylvania State University, 5101 Freer Bldg., Philadelphia, Pa. 16802.
452
GAFFNEY AND PIGOTT
Richardson (10). Specimens of human breast carcinomas averaging 3 cm 3 were delivered to the laboratory in medium 199 supplemented with 1% fetal bovine serum and gentamycin (50 U per ml). Tissues were processed following surgery or stored overnight at 4 ~ C. Storage of tissue did not influence the recovery of viable cells. Specimens were placed in glass Petri dishes with medium and scraped with a scalpel. In one series of experiments tissue fragments were incubated with growth medium containing collagenase {1 mg per ml) at 37 ~ C for 1 to 3 hr. The number of clusters and cells per cluster obtained was determined by directly counting 0.01ml aliquots of the suspensions with the aid of a dissecting microscope. The total population was determined from hemacytometer counts of single cells obtained after 20 min incubation in a 0.05% trypsin-EDTA solution. Cell cultures were established in Limbro 35-ram multiwell tissue culture plates. Medium 199 was supplemented with a variety of sera during early studies. Sera selected included: three batches of virus-tested mycoplasma-free fetal bovine serum (Flow Laboratories, Rockville, Md.); two batches of mixed human serum and human serum-type A B - - b o t h negative for hepatitis-associated antigen (Grand Island Biological Co., Grand Island, N.Y.); and one batch each of agamma bovine (Grand Island Biological Co.L swine, horse, chicken and bovine (Microbiological Associates, Bethesda, Md.L Three batches of human serum also were obtained from blood-bank samples of normal donors. Sermn was separated from cells by decanting and sterilizing through a 0.22-/~m Millipore filter. These samples also were found negative for HA antigen. No attempt was made to separate serum based on age or sex of donor. In one series of experiments human sera were modified further in one of the following ways: (a} dialyzed for 3 days against phosphate buffered saline; (b) heated at 56 ~ C for 1 hr to inactivate complement; or (c) absorbed for 1 hr at 53 ~ C with 10 mg per ml activated charcoal and 1 mg per ml Dextran T 40 to reduce hormone concentrations. The number of cell colonies formed from different samples of ductal carcinoma varied considerably. Thus 1000 cell clusters or more were inoculated per replicate culture to ensure adequate colony number. The efficiency of cell cluster attachment was found from the average count of six 35mm dishes stained at 1 or 2 days after seeding.
RESULTS
Growth characteristics of E' cells. The initial studies included a series of 33 ductal carcinomas representing different histological tumor grades and five samples of carcinoma metastatic to the lymph nodes. Cell suspensions were inoculated to 35-ram plastic dishes in medium 199 supplemented with 10% FBS. All samples of metastatic tumor, 14 of 16 samples of tumor classified as malignancy grade I I I , 5 of 7 samples classified as grade II, and 3 of 10 samples of grade I tumor gave large numbers of refractile cell clusters after cutting and scraping with a scalpel. The cultures established from cell clusters remained fibroblast free. Epithelial cells in colonies were classified on the morphological criteria of Hallowes et al. (6). E' cell colonies were observed to form within 24 hr of seeding, whereas colonies of E cells were identified after 48 to 72 hr. Colonies of E cells were irregularly shaped, but the edges were continuous showing no intercellular spacing when viewed under light microscope. In contrast, E' colonies were spherical and composed of cells more heterogeneous in shape. These lacked the close cell to cell contacts observed with E cells (Fig. 1). Cultures from 20 tumors were composed solely of E' cell colonies. A gradual loss of the E' colonies was noted beyond the 1st week of culture, and the few epithelial-like colonies present at 2 weeks were composed of the type E cells. In a subsequent study the efficiency of cell cluster attachment was determined from the average number of colonies observed in six replicate cultures 24 or 48 hr after inoculation of cells of seven tumors. The tumors represented the three histological grades with variations in inocula from 1.5 x 102 to 6 x 103 cell clusters per 35-ram dish. Table 1 shows that the proportion of cell clusters capable of attaching varied greatly among samples from a low of 0.3% to a high of 13.0%. The attachment of cell clusters was related to variations among tumor samples rather than differences in the average number of cells per cluster or the tumor grade. The relationship between the method of tissue disruption and the types of cell colonies subsequently observed in culture was determined from experiments with three tumor specimens. Approximately equal amounts of tissue were either scraped with a scalpel (spillage technique) or cut into 4-ram cubes and incubated in collagenase for
SERUM AND MAMMARY TISSUE
FI6.1. Morphology of the two epithelial-like cell colonies observed in 7-day cultures established from biopsy samples of human mammary ductal carcinoma. Cells designated as E'. • Cells designated as E. x130.
453
454
GAFFNEY AND PIGOTT TABLE 1 ATTACHMENT OF C ELL C LUSTERS FROM MALIGNANT T ISSUES Tumor
Grade
luncula per Dish
Average No. Cells per Cluster
% Attachment
140 177 278 235 343 522 523
I II II III III III II
4.2 x I(P 4.0x 102 6.0 x 103 4.6 x 102 8.0 x 102 3.0 x 103 1.5 x 102
18 20 34 32 16 11 --
0.5 6.5 0.3 12.0 13.0 1.7 0.7
I hr. The cell suspensions were washed and inoculated into replicate dishes. The distribution of colonies composed of the E', E or F (fibroblast) cell types was found by observing nine cultures resuiting from each treatment after maintenance for I week. The data for one representative experiment (Table 2) show that the greatest number of E' cell colonies resulted from the spillage technique and that these cultures are fibroblast free. Collagenase increased the number of E and F cell colonies. E' colonies were seen to increase in size during the ]st week of culture and a few mitotic figures were seen in most dishes. However, attempts to quantitate growth by an increase in the total number of cells in replicate cultures from eight tumors were unsuccessful. Cell counts following removal of cells from dishes with trypsin and E D T A were not significantly different except that a decrease was consistently observed by I0 days. Daily microscopic monitoring of the cultures revealed that after a few days many E' colonies were degenerating. This loss of cell colonies was accompanied by an increase in the amount of debris in the supernatant fluids.
Serum requirements of mammary epithelium. Various sera were examined to determine which would provide optimal growth of E' cells derived from ductal carcinomas. Suspensions of cells from 11 tumors were seeded in triplicate to medium containing 5 or 15% fetal bovine, agamma bovine, whole bovine, human, swine, horse or chicken serum. Sufficient cell colonies were ob-
TABLE 2 COLONY FORMATION BY MAMMARY TUMOR CELLS Cell Type
E' E F
Average Colony Count Spillage Conagenase (1 hr)
38_+3 4.+1 0
19+_2 11_+1 14+3
served after 7 days in cultures from four tumors to permit an accurate assessment of the sera. Colony formation occurred in the presence of fetal bovine, human and agamma bovine sera. However, colony enlargement was only observed in FBS and these cultures contained three times as many colonies as seen in those with human serum. The average numbers of colonies for one tumor following an inocuium of 2 x 103 cell clusters were: FBS, 21; agamma bovine serum, 9; human, 7. These results and those of previous studies (6,8) that reported the growth of mammary cells in a combination of FBS and human sera prompted further experiments. Scraped cells from several samples of primary ductal carcinoma were seeded in triplicate to medium containing 15% FBS and 20% human serum singly or in combination. Thirty-five percent FBS was included with some cultures as a control for serum toxicity at high concentrations. All cultures were fluid changed at 2 days. The resuits for three experiments (Table 3) are expressed as the average number of colonies with deviation among triplicate samples observed after I week. Several important facts emerged from these studies. First, approximately the same number of E' cell colonies were seen whether 15 or 35% FBS was present. Second, based on growth pattern, the majority of colonies were composed of E ' cells. Third, cultures established in FBS contained 4 to I0 times more E' cell colonies than observed with human serum. Fourth, colony formation was not enhanced by FBS in the presence of human serum. This implied that human serum not only lacked factors necessary for colony formation but, more importantly, inhibited cells from responding to the active components present in FBS.
Colony formation in different human sera. The following series of experiments were performed to determine the nature of the inhibitory activity associated with human serum. Cultures were established with suspensions of cells obtained by scraping biopsy specimens of ductal carcinoma. Media
455
SERUM AND MAMMARY TISSUE TABLE 3 E FFECT OF F ETAL B OVINE AND H UMAN S ERA ON C OLONY FORMATION BY MAMMARY TUMOR C ELLS No. Colonies Tumor
Grade
Serum
E
177
II
15 FBS 35 FBS 15 FBS + 20 HuS 20HuS
28 -+ 2 27 -+ 2 5 _+1 5_+2
-----
235
II
15 FBS 35 FBS 15 FBS + 20HuS 20HuS
23 +- 1 25 _+3 5__.3 2_+1
1 +_1 2 _+1 7_+1 4_+0
522
III
15 FBS 15 FBS + 20 HuS 20 HuS
51 _+12 14 _+2 12 -+ 7
----
were renewed 2 days after inoculation a n d experim e n t s were t e r m i n a t e d on d a y 7. T h e n u m b e r of colonies was d e t e r m i n e d f r o m o b s e r v a t i o n s on triplicate cultures for each e x p e r i m e n t a l variable. TABLE 4 E FFECT OF D IFFERENT H UMAN S ERA ON COLONY FORMATION Serum
Average No. E' Cell Colonies
Exp. 1 10% 20% 10% 10% 10% 10% 10% 10%
E'
Exp. 2
FBS 388-+ 1 245-+ 17 FBS 396-+ 11 239 -+ 14 FBS + 1 0 % blood bank HuS 52-+ 7 35-+ 7 Blood-bank HuS 3 _+ 4 2 _+ 2 FBS -4- 10% GIBCO HuS 58_+ 20 33 _+10 GIBCO HuS 6_+ 3 5_+ 4 FBS -4- 10% patient HuS 84-+ 16 73 _+ 1 Patient HuS .4_+ 3 0
Cells o b t a i n e d from two samples of tissue were inoculated to m e d i u m with 10% or 2 0 % F B S a n d / o r 10% . h u m a n serum. C o m m e r c i a l h u m a n serum, s e r u m p r e p a r e d f r o m b l o o d - b a n k samples, a n d s e r u m pooled from p a t i e n t s a t a local hospital were included. D i f f e r e n t h u m a n s e r u m b a t c h e s were used in each e x p e r i m e n t . T h e results in T a b l e 4 were o b t a i n e d f r o m two large t u m o r sampies. T h e d a t a illustrate t h a t h u m a n s e r u m alone or in c o m b i n a t i o n with F B S i n h i b i t e d cell colony formation. However, pooled h o s p i t a l s e r u m samples p e r m i t t e d t h e a t t a c h m e n t a n d s u b s e q u e n t m i g r a t i o n of cells from clusters more readily t h a n h u m a n s e r u m f r o m o t h e r sources. T h e possibility t h a t this inhibitory activity s t e m m e d from the presence of h o r m o n e s a n t a g o n istic to E ' cells was reduced by f u r t h e r experim e n t s using c h a r c o a l - a b s o r b e d h u m a n serum.
TABLE 5 E FFECT OF TREATED HUMAN SERA ON COLONY FORMATION Serum
Average No. Colonies
Experiment I 10% 20% 10% 10% 10% 10%
FBS FBS FBS + 10% HuS FBS + 10% charcoal-absorbed HuS HuS Absorbed HuS
E'
E
21-2-_1 20 _+1 4 _+3 3 _+1 2 _+2 3 -+ 1
1_+0 1_+1 5_+2 1_+1 0 1_+0
23 -+ 1 22 _+ 1 5 -+ 3 2_+1 3 -+ 1 0 4+2 0 4 _+0 2_ 2
1 +_0
Experiment 2 10% FBS 20% FBS 10% FBS + 10% 10% 10% FBS + 10% 10% 10% FBS -4- 10% 10% 10% FBS + 10% 10%
HuS HuS dialyzed HuS dialyzed HuS heat-inactivated HuS heat-inactivated HuS HuS - - type AB HuS - - type AB
1 _+1 7 -+ 1
4_+0 1 _+1 1 _+0 1 +_1 1 _+0 3 -+ 0 2 _+2
456
GAFFNEY AND PIGOTT
The results (Experiment 1, Table 5) show that absorbed serum alone or in combination with FBS was as effective as whole serum in inhibiting E' cell colony formation. However, higher numbers of E cell colonies were observed in FBS supplemented with untreated human serum, suggesting that the removal of hormones diminishes the colony-promoting activity for E cells. The involvement of small molecular-weight constituents, complement-dependent and bloodtype specific antibodies was examined in a series of experiments with four carcinoma samples. The results of one experiment are shown in Table 5 (Experiment 2). Treatment of human serum by dialysis or heat inactivation failed to eliminate the inhibitory factors. These were also present in type AB serum. As observed in previous experiments, untreated human serum promoted the development of E cell colonies.
DISCUSSION Our results and those of others 16,8,11) have shown that clusters of cells from breast samples routinely form discrete colonies of epithelium during short-term culture. Two types of epithelial cells have been recognized from light and electron microscopic studies: E', isolated only from malignant tissues; and E, observed in cultures established from all breast tissues. However, the homogeneity of the populations of E cells from normal, benign and malignant sources has not been established ~6). The data in Table 1 illustrate that the frequency with which cell clusters from carcinomas attach to culture dishes is very low, and microscopic observations revealed that the survival of E' cell colonies is limited to a few days. These restrictions of low cell number and limited in vitro growth capacity hinder attempts to quantitate the effects of factors that may alter growth. This study takes advantage of the colony-forming characteristic of E' cells as the only early culture event by which to measure cell survival. Measurements of colony number account for the attachment of cell clusters and spreading. A similar approach utilizing colony number and size has been used to study the effects of sera on E cells derived from fibroadenomas and lacteal secretions (8,11 ). The rate of division by cells can be altered in response to changes in the external environment, such as the addition of serum (12) or growthpromoting factors t13). Certain serum components appear to interact with and aid the expression of growth regulators, such as fibroblast
growth factor, hydrocortisone and pituitary hormones i14). The effects of serum and regulating substances on the growth of cultured human mammary cells are not clear. Some studies have recommended that medium containing fetal bovine serum, human serum, insulin and hydrocortisone provides the best environment for the growth of epithelial cells derived from normal sources and benign tumors (6,8). Our data demonstrate that human serum alone does not promote colony development by either cell type derived from carcinomas. This difference in results appears to stem from the absence of exogenous hormones in our medium rather than differences among sources of human serum (Table 5). Thus certain factors, such as hormones, may interact with some component of serum to enhance mammary cell proliferation. For example, epidermal growth factor was shown to elicit a maximum mitogenic response only in the presence of serum in mouse mammary (15~ and human mammary epithelial cell cultures (8). Unpublished observations from experiments in this laboratory also have shown that colony formation and enlargement by E cells derived form fibroadenomas are promoted by hydrocortisone added to medium containing certain sera. The current investigation demonstrated that normal human serum inhibits colony development with E' cells, but not with E cells cultured from malignant tissues. This could result from the presence of hormones, such as testosterone, antagonistic to E' cell survival in culture. However, dialysis or charcoal absorption failed to diminish serum activity. This inhibition also could reflect the presence in normal human serum of antibodies against E' cells or factors antagonistic to cell surface proteins necessary for attachment (16). The inhibition of colony formation in cultures established from carcinoma cells illustrates a specific, but as yet unexplainable, difference among mammary cell types. An initial approach to a better understanding of this phenomenon is a study comparing colony formation in the presence of normal human serum with that in serum from patients with breast cancer. Further explanation of the differences in colony formation may provide insight into the interaction of hormones, serum factors and mammary cells. REFERENCES 1. Lasfargues, E. Y., and L. Ozzelo. 1958. Cultivation of human breast carcinomas. J. Nat. Cancer Inst. 21: 1131-1147.
SERUM AND MAMMARY TISSUE 2. Cailleau, R., R. Young, M. Olive, and W. R. Reeves. 1974. Breast tumor cell lines from pleural effusions. J. Nat. Cancer Inst. 53: 661-674. 3. Soule, D., J. Vazquez, A. Long, S. Albert, and M. Brennan. 1973. A human cell line from a pleural effusion derived from a breast carcinoma. J. Nat. Cancer Inst. 51: 1409-1416. 4. Lasfargnes, E. Y,, and B. E. Moore. 1974. Search for a viral etiology of human breast cancer. J. Invest. Dermatol. 63: 147-159. 5. Lasfargues, E. Y. 1975. New approaches to the cultivation of human breast carcinomas. In: J. Fogh {Ed.), Human Tumor Cells In Vitro. Plenum Press, London, pp. 51-77. 6. Hallowes, R. C., R. Millis, D. Pigott, M. Shearer, M. G. P. Stoker, and J. Taylor-Papadimitriou. 1977. Results on a pilot study of lacteal secretions and benign and malignant breast tumors. Clin. Oncol. 3: 81-90. 7. Gaffney, E.V., F.P. Polanowski, S.E. Blackbum, and J. T. Lambiase. 1976. Origin, concentration and structural features of human mammary cells cultured from breast secretions. Cell Tissue Res. 172: 269-279. 8. Stoker, M. G.P., D. Pigott, and J. TaylorPapadimitriou. 1976. Response to epidermal growth factors of cultured human mammary epithelial cells from benign tltmors. Natue ~London} 264: 764-767.
457
9. Gaffney, E.V., F . P . Polanowski, S. E. Blackbum, J. T. Lambiase, and R. E. Burke. 1976. Cultures of normal human mammary cells. Cell Differ. 5: 69-81. 10. Bloom, H., and W. Richardson. 1957. Histological grading and prognosis in breast cancer. Br. J. Cancer 11: 359-377. 11. Taylor-Papdimitriou, J., M. Shearer, and R. Tilly. 1977. Some properties of cells cultured from early lactation human milk. J. Nat. Cancer Inst. 58: 1563-1571. 12. Temin, H. M., and M. Howard. 1971. Stimulation by serum of multiplication of stationary chicken cells. J. Cell. Physiol. 78: 161-170. 13. Rudland, P. S., W. E. Siebert, and D. Gospodarowicz. 1974. Growth control in cultured mouse fibroblasts: Induction of the pleiotypic and mitogenie responses by a purified growth factor. Proc. Nat. Aead. Sci. U.S.A. 70: 2600-2604. 14. Thrash, C. R., and D. Cunningham. 1973. Glucocorticoid stimulation of division in density inhibited fibroblasts. Nature (Londonj 242: 399-400. 15. Turkington, R.W. 1969. The role of epithelial growth factor in mammary gland development in vitro. Exp. Cell Res. 57: 79-85. 16. All, I.U., V. Mantner, R. Lanza, and R.O. Hynes. 1977. Restoration of normal morphology, adhesion and cytoskeleton in transformed cells by addition of a transformation-sensitive surface protein. Cell l h 115-126.
Supported by NCI Contract CB-33898 and a Fellowship from the Imperial Cancer Research Fund, London, England. The helpful advice of Drs. J. Taylor-Papadimitriou and M. Stoker is greatly appreciated. The surgical specimens and their histopathology were provided by Dr. Rosemary MiUis, Consultant Pathologist, Guy's Hospital, London, England, and Leonard Patton, Geisinger Medical Center, Danville, Pennsylvania.
EFFECT OF SERUM ON CELLS CULTURED FROM HUMAN MAMMARY TUMORS E. V. GAFFNEY ~ ANDD. PIGOTT
The Imperial Cancer Research Fund Laboratories, Lincoln's Inn Fields, London, England
SUMMARY
Clusters of cells derived from biopsy specimens of human mammary ductal carcinomas form two morphologically distinct epithelial colonies in culture, designated as E and E'. The proportion of E' cell clusters that attached and formed colonies ranged from 0.3 to 13.0% with different tumors. Attachment was independent of tumor grade. Microscopic observations revealed that the survival of E' cell colonies was limited to approximately 10 days with rapid cell degeneration commencing about 7 days. A comparison of sera showed that colony formation by cells from malignant tumors during the 1st week of culture was maximum in the presence of fetal bovine serum. Human serum alone was 70 to 100% less effective in promoting E' colonies. The most significant finding was that human serum from normal donors inhibited E' colony development in the presence of FBS. Although human serum was less effective than FBS in promoting colony formation by clusters of E cells, an inhibition was not observed. Inhibitory activity could not be attributed to either antagonistic hormones or the source of human serum. These results demonstrate that normal human serum contains a factor{s) that exhibits an inhibitory activity specific for human epithelial cells (E') derived from malignant tumors.
Key words: mammary; human; sera; colony formation. INTRODUCTION
Most studies of malignant human mammary cells are limited to the few lines isolated from both primary and metastatic breast tumors. These
studies provide important information on cell growth and morphology (1,2) as well as tumor etiology (3,4). However, it would be more meaningtul for the individual patient if investigators were routinely able to culture malignant cells from tumor specimens. Unfortunately, there has been limited success deriving even primary cultures from the majority of samples {5). This failure stems in part from our lack of knowledge concerning both the identity of the cell types observed in culture and their nutritional requirements essential for long-term maintenance. A recent report established that two distinct epithelial cell types, designated as E and E', can be observed in cultures established from breast
carcinomas (6). Both cell types (E and E') demonstrated ultrastructural properties typical of epithelium, but differed in general cell shape and growth pattern. Cells similar in morphology and arrangement to E cells also were isolated from lacteal secretions (7), benign dysplasias (8) and reduction mammoplasties (9). In contrast E' cells were isolated solely from malignant tumors. Thus it was suggested by Hallows et al. (6) that E cells may arise from residual normal or hyperplastic epithelium originating in tumor-associated breast tissue and that E' cells are derived from neoplastic epithelium. The current study serves two purposes: first, to quantitative cell yield from tumors and survival time in primary culture as a basis for investigating nutritional factors that may enhance division potential; and second, to establish the response of mammary cells to different sera. MATERIALS AND METHODS
Histological grading of ductal carcinomas, which predicts the relative aggressiveness of the tumor, was based on the criteria of Bloom and 451
~Request for reprints should be sent to: Dr. E. V. Gaffney, Pennsylvania State University, 5101 Freer Bldg., Philadelphia, Pa. 16802.
452
GAFFNEY AND PIGOTT
Richardson (10). Specimens of human breast carcinomas averaging 3 cm 3 were delivered to the laboratory in medium 199 supplemented with 1% fetal bovine serum and gentamycin (50 U per ml). Tissues were processed following surgery or stored overnight at 4 ~ C. Storage of tissue did not influence the recovery of viable cells. Specimens were placed in glass Petri dishes with medium and scraped with a scalpel. In one series of experiments tissue fragments were incubated with growth medium containing collagenase {1 mg per ml) at 37 ~ C for 1 to 3 hr. The number of clusters and cells per cluster obtained was determined by directly counting 0.01ml aliquots of the suspensions with the aid of a dissecting microscope. The total population was determined from hemacytometer counts of single cells obtained after 20 min incubation in a 0.05% trypsin-EDTA solution. Cell cultures were established in Limbro 35-ram multiwell tissue culture plates. Medium 199 was supplemented with a variety of sera during early studies. Sera selected included: three batches of virus-tested mycoplasma-free fetal bovine serum (Flow Laboratories, Rockville, Md.); two batches of mixed human serum and human serum-type A B - - b o t h negative for hepatitis-associated antigen (Grand Island Biological Co., Grand Island, N.Y.); and one batch each of agamma bovine (Grand Island Biological Co.L swine, horse, chicken and bovine (Microbiological Associates, Bethesda, Md.L Three batches of human serum also were obtained from blood-bank samples of normal donors. Sermn was separated from cells by decanting and sterilizing through a 0.22-/~m Millipore filter. These samples also were found negative for HA antigen. No attempt was made to separate serum based on age or sex of donor. In one series of experiments human sera were modified further in one of the following ways: (a} dialyzed for 3 days against phosphate buffered saline; (b) heated at 56 ~ C for 1 hr to inactivate complement; or (c) absorbed for 1 hr at 53 ~ C with 10 mg per ml activated charcoal and 1 mg per ml Dextran T 40 to reduce hormone concentrations. The number of cell colonies formed from different samples of ductal carcinoma varied considerably. Thus 1000 cell clusters or more were inoculated per replicate culture to ensure adequate colony number. The efficiency of cell cluster attachment was found from the average count of six 35mm dishes stained at 1 or 2 days after seeding.
RESULTS
Growth characteristics of E' cells. The initial studies included a series of 33 ductal carcinomas representing different histological tumor grades and five samples of carcinoma metastatic to the lymph nodes. Cell suspensions were inoculated to 35-ram plastic dishes in medium 199 supplemented with 10% FBS. All samples of metastatic tumor, 14 of 16 samples of tumor classified as malignancy grade I I I , 5 of 7 samples classified as grade II, and 3 of 10 samples of grade I tumor gave large numbers of refractile cell clusters after cutting and scraping with a scalpel. The cultures established from cell clusters remained fibroblast free. Epithelial cells in colonies were classified on the morphological criteria of Hallowes et al. (6). E' cell colonies were observed to form within 24 hr of seeding, whereas colonies of E cells were identified after 48 to 72 hr. Colonies of E cells were irregularly shaped, but the edges were continuous showing no intercellular spacing when viewed under light microscope. In contrast, E' colonies were spherical and composed of cells more heterogeneous in shape. These lacked the close cell to cell contacts observed with E cells (Fig. 1). Cultures from 20 tumors were composed solely of E' cell colonies. A gradual loss of the E' colonies was noted beyond the 1st week of culture, and the few epithelial-like colonies present at 2 weeks were composed of the type E cells. In a subsequent study the efficiency of cell cluster attachment was determined from the average number of colonies observed in six replicate cultures 24 or 48 hr after inoculation of cells of seven tumors. The tumors represented the three histological grades with variations in inocula from 1.5 x 102 to 6 x 103 cell clusters per 35-ram dish. Table 1 shows that the proportion of cell clusters capable of attaching varied greatly among samples from a low of 0.3% to a high of 13.0%. The attachment of cell clusters was related to variations among tumor samples rather than differences in the average number of cells per cluster or the tumor grade. The relationship between the method of tissue disruption and the types of cell colonies subsequently observed in culture was determined from experiments with three tumor specimens. Approximately equal amounts of tissue were either scraped with a scalpel (spillage technique) or cut into 4-ram cubes and incubated in collagenase for
SERUM AND MAMMARY TISSUE
FI6.1. Morphology of the two epithelial-like cell colonies observed in 7-day cultures established from biopsy samples of human mammary ductal carcinoma. Cells designated as E'. • Cells designated as E. x130.
453
454
GAFFNEY AND PIGOTT TABLE 1 ATTACHMENT OF C ELL C LUSTERS FROM MALIGNANT T ISSUES Tumor
Grade
luncula per Dish
Average No. Cells per Cluster
% Attachment
140 177 278 235 343 522 523
I II II III III III II
4.2 x I(P 4.0x 102 6.0 x 103 4.6 x 102 8.0 x 102 3.0 x 103 1.5 x 102
18 20 34 32 16 11 --
0.5 6.5 0.3 12.0 13.0 1.7 0.7
I hr. The cell suspensions were washed and inoculated into replicate dishes. The distribution of colonies composed of the E', E or F (fibroblast) cell types was found by observing nine cultures resuiting from each treatment after maintenance for I week. The data for one representative experiment (Table 2) show that the greatest number of E' cell colonies resulted from the spillage technique and that these cultures are fibroblast free. Collagenase increased the number of E and F cell colonies. E' colonies were seen to increase in size during the ]st week of culture and a few mitotic figures were seen in most dishes. However, attempts to quantitate growth by an increase in the total number of cells in replicate cultures from eight tumors were unsuccessful. Cell counts following removal of cells from dishes with trypsin and E D T A were not significantly different except that a decrease was consistently observed by I0 days. Daily microscopic monitoring of the cultures revealed that after a few days many E' colonies were degenerating. This loss of cell colonies was accompanied by an increase in the amount of debris in the supernatant fluids.
Serum requirements of mammary epithelium. Various sera were examined to determine which would provide optimal growth of E' cells derived from ductal carcinomas. Suspensions of cells from 11 tumors were seeded in triplicate to medium containing 5 or 15% fetal bovine, agamma bovine, whole bovine, human, swine, horse or chicken serum. Sufficient cell colonies were ob-
TABLE 2 COLONY FORMATION BY MAMMARY TUMOR CELLS Cell Type
E' E F
Average Colony Count Spillage Conagenase (1 hr)
38_+3 4.+1 0
19+_2 11_+1 14+3
served after 7 days in cultures from four tumors to permit an accurate assessment of the sera. Colony formation occurred in the presence of fetal bovine, human and agamma bovine sera. However, colony enlargement was only observed in FBS and these cultures contained three times as many colonies as seen in those with human serum. The average numbers of colonies for one tumor following an inocuium of 2 x 103 cell clusters were: FBS, 21; agamma bovine serum, 9; human, 7. These results and those of previous studies (6,8) that reported the growth of mammary cells in a combination of FBS and human sera prompted further experiments. Scraped cells from several samples of primary ductal carcinoma were seeded in triplicate to medium containing 15% FBS and 20% human serum singly or in combination. Thirty-five percent FBS was included with some cultures as a control for serum toxicity at high concentrations. All cultures were fluid changed at 2 days. The resuits for three experiments (Table 3) are expressed as the average number of colonies with deviation among triplicate samples observed after I week. Several important facts emerged from these studies. First, approximately the same number of E' cell colonies were seen whether 15 or 35% FBS was present. Second, based on growth pattern, the majority of colonies were composed of E ' cells. Third, cultures established in FBS contained 4 to I0 times more E' cell colonies than observed with human serum. Fourth, colony formation was not enhanced by FBS in the presence of human serum. This implied that human serum not only lacked factors necessary for colony formation but, more importantly, inhibited cells from responding to the active components present in FBS.
Colony formation in different human sera. The following series of experiments were performed to determine the nature of the inhibitory activity associated with human serum. Cultures were established with suspensions of cells obtained by scraping biopsy specimens of ductal carcinoma. Media
455
SERUM AND MAMMARY TISSUE TABLE 3 E FFECT OF F ETAL B OVINE AND H UMAN S ERA ON C OLONY FORMATION BY MAMMARY TUMOR C ELLS No. Colonies Tumor
Grade
Serum
E
177
II
15 FBS 35 FBS 15 FBS + 20 HuS 20HuS
28 -+ 2 27 -+ 2 5 _+1 5_+2
-----
235
II
15 FBS 35 FBS 15 FBS + 20HuS 20HuS
23 +- 1 25 _+3 5__.3 2_+1
1 +_1 2 _+1 7_+1 4_+0
522
III
15 FBS 15 FBS + 20 HuS 20 HuS
51 _+12 14 _+2 12 -+ 7
----
were renewed 2 days after inoculation a n d experim e n t s were t e r m i n a t e d on d a y 7. T h e n u m b e r of colonies was d e t e r m i n e d f r o m o b s e r v a t i o n s on triplicate cultures for each e x p e r i m e n t a l variable. TABLE 4 E FFECT OF D IFFERENT H UMAN S ERA ON COLONY FORMATION Serum
Average No. E' Cell Colonies
Exp. 1 10% 20% 10% 10% 10% 10% 10% 10%
E'
Exp. 2
FBS 388-+ 1 245-+ 17 FBS 396-+ 11 239 -+ 14 FBS + 1 0 % blood bank HuS 52-+ 7 35-+ 7 Blood-bank HuS 3 _+ 4 2 _+ 2 FBS -4- 10% GIBCO HuS 58_+ 20 33 _+10 GIBCO HuS 6_+ 3 5_+ 4 FBS -4- 10% patient HuS 84-+ 16 73 _+ 1 Patient HuS .4_+ 3 0
Cells o b t a i n e d from two samples of tissue were inoculated to m e d i u m with 10% or 2 0 % F B S a n d / o r 10% . h u m a n serum. C o m m e r c i a l h u m a n serum, s e r u m p r e p a r e d f r o m b l o o d - b a n k samples, a n d s e r u m pooled from p a t i e n t s a t a local hospital were included. D i f f e r e n t h u m a n s e r u m b a t c h e s were used in each e x p e r i m e n t . T h e results in T a b l e 4 were o b t a i n e d f r o m two large t u m o r sampies. T h e d a t a illustrate t h a t h u m a n s e r u m alone or in c o m b i n a t i o n with F B S i n h i b i t e d cell colony formation. However, pooled h o s p i t a l s e r u m samples p e r m i t t e d t h e a t t a c h m e n t a n d s u b s e q u e n t m i g r a t i o n of cells from clusters more readily t h a n h u m a n s e r u m f r o m o t h e r sources. T h e possibility t h a t this inhibitory activity s t e m m e d from the presence of h o r m o n e s a n t a g o n istic to E ' cells was reduced by f u r t h e r experim e n t s using c h a r c o a l - a b s o r b e d h u m a n serum.
TABLE 5 E FFECT OF TREATED HUMAN SERA ON COLONY FORMATION Serum
Average No. Colonies
Experiment I 10% 20% 10% 10% 10% 10%
FBS FBS FBS + 10% HuS FBS + 10% charcoal-absorbed HuS HuS Absorbed HuS
E'
E
21-2-_1 20 _+1 4 _+3 3 _+1 2 _+2 3 -+ 1
1_+0 1_+1 5_+2 1_+1 0 1_+0
23 -+ 1 22 _+ 1 5 -+ 3 2_+1 3 -+ 1 0 4+2 0 4 _+0 2_ 2
1 +_0
Experiment 2 10% FBS 20% FBS 10% FBS + 10% 10% 10% FBS + 10% 10% 10% FBS -4- 10% 10% 10% FBS + 10% 10%
HuS HuS dialyzed HuS dialyzed HuS heat-inactivated HuS heat-inactivated HuS HuS - - type AB HuS - - type AB
1 _+1 7 -+ 1
4_+0 1 _+1 1 _+0 1 +_1 1 _+0 3 -+ 0 2 _+2
456
GAFFNEY AND PIGOTT
The results (Experiment 1, Table 5) show that absorbed serum alone or in combination with FBS was as effective as whole serum in inhibiting E' cell colony formation. However, higher numbers of E cell colonies were observed in FBS supplemented with untreated human serum, suggesting that the removal of hormones diminishes the colony-promoting activity for E cells. The involvement of small molecular-weight constituents, complement-dependent and bloodtype specific antibodies was examined in a series of experiments with four carcinoma samples. The results of one experiment are shown in Table 5 (Experiment 2). Treatment of human serum by dialysis or heat inactivation failed to eliminate the inhibitory factors. These were also present in type AB serum. As observed in previous experiments, untreated human serum promoted the development of E cell colonies.
DISCUSSION Our results and those of others 16,8,11) have shown that clusters of cells from breast samples routinely form discrete colonies of epithelium during short-term culture. Two types of epithelial cells have been recognized from light and electron microscopic studies: E', isolated only from malignant tissues; and E, observed in cultures established from all breast tissues. However, the homogeneity of the populations of E cells from normal, benign and malignant sources has not been established ~6). The data in Table 1 illustrate that the frequency with which cell clusters from carcinomas attach to culture dishes is very low, and microscopic observations revealed that the survival of E' cell colonies is limited to a few days. These restrictions of low cell number and limited in vitro growth capacity hinder attempts to quantitate the effects of factors that may alter growth. This study takes advantage of the colony-forming characteristic of E' cells as the only early culture event by which to measure cell survival. Measurements of colony number account for the attachment of cell clusters and spreading. A similar approach utilizing colony number and size has been used to study the effects of sera on E cells derived from fibroadenomas and lacteal secretions (8,11 ). The rate of division by cells can be altered in response to changes in the external environment, such as the addition of serum (12) or growthpromoting factors t13). Certain serum components appear to interact with and aid the expression of growth regulators, such as fibroblast
growth factor, hydrocortisone and pituitary hormones i14). The effects of serum and regulating substances on the growth of cultured human mammary cells are not clear. Some studies have recommended that medium containing fetal bovine serum, human serum, insulin and hydrocortisone provides the best environment for the growth of epithelial cells derived from normal sources and benign tumors (6,8). Our data demonstrate that human serum alone does not promote colony development by either cell type derived from carcinomas. This difference in results appears to stem from the absence of exogenous hormones in our medium rather than differences among sources of human serum (Table 5). Thus certain factors, such as hormones, may interact with some component of serum to enhance mammary cell proliferation. For example, epidermal growth factor was shown to elicit a maximum mitogenic response only in the presence of serum in mouse mammary (15~ and human mammary epithelial cell cultures (8). Unpublished observations from experiments in this laboratory also have shown that colony formation and enlargement by E cells derived form fibroadenomas are promoted by hydrocortisone added to medium containing certain sera. The current investigation demonstrated that normal human serum inhibits colony development with E' cells, but not with E cells cultured from malignant tissues. This could result from the presence of hormones, such as testosterone, antagonistic to E' cell survival in culture. However, dialysis or charcoal absorption failed to diminish serum activity. This inhibition also could reflect the presence in normal human serum of antibodies against E' cells or factors antagonistic to cell surface proteins necessary for attachment (16). The inhibition of colony formation in cultures established from carcinoma cells illustrates a specific, but as yet unexplainable, difference among mammary cell types. An initial approach to a better understanding of this phenomenon is a study comparing colony formation in the presence of normal human serum with that in serum from patients with breast cancer. Further explanation of the differences in colony formation may provide insight into the interaction of hormones, serum factors and mammary cells. REFERENCES 1. Lasfargues, E. Y., and L. Ozzelo. 1958. Cultivation of human breast carcinomas. J. Nat. Cancer Inst. 21: 1131-1147.
SERUM AND MAMMARY TISSUE 2. Cailleau, R., R. Young, M. Olive, and W. R. Reeves. 1974. Breast tumor cell lines from pleural effusions. J. Nat. Cancer Inst. 53: 661-674. 3. Soule, D., J. Vazquez, A. Long, S. Albert, and M. Brennan. 1973. A human cell line from a pleural effusion derived from a breast carcinoma. J. Nat. Cancer Inst. 51: 1409-1416. 4. Lasfargnes, E. Y,, and B. E. Moore. 1974. Search for a viral etiology of human breast cancer. J. Invest. Dermatol. 63: 147-159. 5. Lasfargues, E. Y. 1975. New approaches to the cultivation of human breast carcinomas. In: J. Fogh {Ed.), Human Tumor Cells In Vitro. Plenum Press, London, pp. 51-77. 6. Hallowes, R. C., R. Millis, D. Pigott, M. Shearer, M. G. P. Stoker, and J. Taylor-Papadimitriou. 1977. Results on a pilot study of lacteal secretions and benign and malignant breast tumors. Clin. Oncol. 3: 81-90. 7. Gaffney, E.V., F.P. Polanowski, S.E. Blackbum, and J. T. Lambiase. 1976. Origin, concentration and structural features of human mammary cells cultured from breast secretions. Cell Tissue Res. 172: 269-279. 8. Stoker, M. G.P., D. Pigott, and J. TaylorPapadimitriou. 1976. Response to epidermal growth factors of cultured human mammary epithelial cells from benign tltmors. Natue ~London} 264: 764-767.
457
9. Gaffney, E.V., F . P . Polanowski, S. E. Blackbum, J. T. Lambiase, and R. E. Burke. 1976. Cultures of normal human mammary cells. Cell Differ. 5: 69-81. 10. Bloom, H., and W. Richardson. 1957. Histological grading and prognosis in breast cancer. Br. J. Cancer 11: 359-377. 11. Taylor-Papdimitriou, J., M. Shearer, and R. Tilly. 1977. Some properties of cells cultured from early lactation human milk. J. Nat. Cancer Inst. 58: 1563-1571. 12. Temin, H. M., and M. Howard. 1971. Stimulation by serum of multiplication of stationary chicken cells. J. Cell. Physiol. 78: 161-170. 13. Rudland, P. S., W. E. Siebert, and D. Gospodarowicz. 1974. Growth control in cultured mouse fibroblasts: Induction of the pleiotypic and mitogenie responses by a purified growth factor. Proc. Nat. Aead. Sci. U.S.A. 70: 2600-2604. 14. Thrash, C. R., and D. Cunningham. 1973. Glucocorticoid stimulation of division in density inhibited fibroblasts. Nature (Londonj 242: 399-400. 15. Turkington, R.W. 1969. The role of epithelial growth factor in mammary gland development in vitro. Exp. Cell Res. 57: 79-85. 16. All, I.U., V. Mantner, R. Lanza, and R.O. Hynes. 1977. Restoration of normal morphology, adhesion and cytoskeleton in transformed cells by addition of a transformation-sensitive surface protein. Cell l h 115-126.
Supported by NCI Contract CB-33898 and a Fellowship from the Imperial Cancer Research Fund, London, England. The helpful advice of Drs. J. Taylor-Papadimitriou and M. Stoker is greatly appreciated. The surgical specimens and their histopathology were provided by Dr. Rosemary MiUis, Consultant Pathologist, Guy's Hospital, London, England, and Leonard Patton, Geisinger Medical Center, Danville, Pennsylvania.
