[CANCER RESEARCH 44, 3398-3402, August 1984]
Identification of a Chemoattractant for Fibroblasts Produced by Human Breast Carcinoma Cell Lines
Wayne E. Gleiber1 and Elliott Schiffmann
Laboratory of Developmental Biology ano Anomalies, National Institute of Dental Research, NIH, Bethesda, Maryland 20205
ABSTRACT human breast tumor cell line (11), as well as three other human breast tumor cell lines, secretes a potent Chemoattractant for Serum-free, conditioned media from the human breast tumor fibroblasts which appears to be a high-molecular-weight protein. cell line ZR-75-1 and three other breast tumor cell lines were each found to contain a potent Chemoattractant for fibroblasts. MATERIALS AND METHODS The Chemoattractant activity was characterized and found to reside in a high-molecular-weight (M, > 100,000) protein. This Materials activity was stable to heating to 56°and from pH 3 to 11, but it L-Glutamine, penicillin, streptomycin, and fetal bovine serum were was sensitive to trypsin and pepsin treatment and to reduction. obtained from Grand Island Biological Co. Bovine insulin, L-thyroxine, When crude attractant was placed in vivo in a slowly releasing human transferrin, 17/i-estradiol, BSA, aprotinin, iodoacetic acid, gelatin pump, a fibrous tissue mass that formed at the releasing site from swine skin, 2-mercaptoethanol, pepsin, and soybean trypsin inhib within 11 days was 3- to 5-fold larger than was that observed itor were obtained from Sigma Chemical Co. PDGF and fibronectin were with pumps containing control medium. It is likely that the pro from Collaborative Research, Inc. DNase, RNase, and trypsin were from duction of a Chemoattractant for fibroblasts by breast tumor cells Worthington Biochemical Corp. Dithiothreitol was from Schwarz/Mann. might contribute to the fibrotic involvement that is common in Formylmethionylleucylphenylalanine was obtained from Peninsula Labo breast carcinomas. ratories. Cell Lines INTRODUCTION Human embryonic skin fibroblasts CRL-1475 and 3T3 murine embry Fibrosis occurs in many diseases in which chronic inflammation onic fibroblasts were obtained from the American Type Culture Collec tion. Human breast tumor cell lines, ZR-75-1, MCF-7, T-47D, and ZR- is found, such as rheumatoid arthritis, tuberculosis, and liver 75-31 and human renal, Ewing's sarcoma, and mucoepidermoid tumor cirrhosis. In addition, fibrotic involvement is associated with some cell lines were obtained from Dr. Lance A. Liotta (National Cancer tumors. Fibrosis is associated with greater than 80% of human Institute, NIH). Murine melanoma BL/6 cells were obtained from Dr. breast carcinomas (24), especially the infiltrating duct carcinoma, Victor P. Terranova (National Institute of Dental Research, NIH). Cells which is characterized by a dense fibrous stroma (6). This fibrotic were grown in 75-sq cm tissue culture flasks (Falcon Plastics), incubated tissue may be formed as an attempt by the host at wound repair, at 37°in a 5% CO2 humidified atmosphere. The growth medium was since the area around the tumor is damaged by its growth (22, changed every 3 to 4 days, and cells were split at a ratio of 1:4 every 1 25). This damage causes an influx of inflammatory cells which to 3 weeks, depending upon the growth rate of the cell line. release mediators that attract fibroblasts from the surrounding Media tissue. As in wound repair, the fibroblasts synthesize new con nective tissue matrix components (20). Growth Medium. All cell lines were grown in DMEM with glucose (4.5 Alternatively, it may be possible that factors other than those g/liter) (NIH medium unit) with additions of 2 mw t-glutamine, 100 related to inflammation may be involved, such as factors pro penicillin (100 units/ml), streptomycin (100 fig/ml), and 7% (v/v) fetal duced by the tumor itself. Previous studies have demonstrated bovine serum. that fibroblasts can be attracted to various soluble factors by Defined Medium. DMEM containing high glucose was supplemented with insulin (3 ^g/ml), thyroxine (7.0 ng/ml), transferrin (1 ¿ig/ml),17/3- chemotaxis, the directed migration of cells toward an attractant (14, 17). Lymphokines (7, 11), fibronectin (12, 19, 23), PDGF,2 estradiol (2.7 ng/ml), L-glutamine, penicillin, and streptomycin. (9,18) complement-derived factors (15), and collagen fragments Conditioned Medium. Procedures for the isolation and characteriza tion of the activity required a defined medium without serum. Serum (4,13) have been found to be attractants for fibroblasts. Each of contains other chemotactic factors for fibroblasts, such as fibronectin, these components could be present at the site of a tumor, due complement factors, and PDGF. Cells were cultured in serum-containing to the inflammation and injury of normal tissue, and each could growth medium following passage to provide a rapid growth of cells. contribute to the formation of fibrotic tissue around the tumor. When the cultures were confluent, the growth medium was removed, We have investigated the possibility that breast tumor cells and the flasks were rinsed twice with phosphate-buffered saline. A produce a fibroblast Chemoattractant. Here we have used cell serum-free defined medium (10 ml) was added to the flasks, and the lines derived from human breast carcinomas (5, 21), grown in cells were incubated for 2 days. The supplements added to the defined serum-free defined medium (1-3). We find that the ZR-75-1 medium (insulin, thyroxine, transferrin, and estradiol), while not essential for production of the attractant, are beneficial for the continued growth of the cells. The hormones lack chemotactic activity and appeared to 1Support is provided by a USPHS National Research Service Award have no effect upon the chemotactic response. The CM was collected (3F32DE05299-03). and centrifuged at 10,000 x g for 10 min, and the supernatant fraction 2 The abbreviations used are: PDGF, platelet-derived growth factor(s); DMEM, was stored at -20°. Production of CM was repeated 3 to 5 times before Dulbecco's modified Eagle's medium; CM, conditioned medium (or media); BSA, bovine serum albumin. cells were split into new flasks. The concentration of protein in dialyzed Received August 8, 1983; accepted March 30, 1984. CM ranged from 70 to
3398 CANCER RESEARCH VOL. 44
Downloaded from cancerres.aacrjournals.org on September 24, 2021. © 1984 American Association for Cancer Research. Tumor-produced Fibroblast Chemoattractant
Chemotaxis Assay BSA (1.5 mg/ml). Indsions were made at the lower right or left flank of lightly anesthetized C57BL/6 mice (2 pumps/animal), and the pumps Chemotaxis assays on fibroblasts were performed in modified Boyden were inserted head first to insure that the releasing end of each pump chambers as described by Posttethwaite ef al. (14). Nudeopore filters was distal to the incision. Each incision site was closed with a single 9- (Neuroprobe Corporation) with 8-¿imdiameter pores were treated with mm stainless steel Autodip (Clay-Adams). On Days 4, 6,9, and 11 after 0.087 M acetic acid at 50°for 20 min, washed twice in distilled water, insertion of the pumps, the sites of implantation on each animal were placed in a gelatin solution (25 pg/ml) at 100° for 60 min, and dried opened, and any fibrous tissue mass which formed was excised, overnight at 37°.The test material in DMEM containing 0.1% BSA was weighed, and fixed in a 10% formaldehyde solution. placed in the lower well of the Boyden chamber, and a treated Nudeopore filter was placed over this. Cells freshly trypsinized from confluent cultures were washed first with DMEM containing soybean trypsin RESULTS inhibitor and then with DMEM containing 0.1% BSA; they were finally resuspended in DMEM plus BSA at a concentration of 3 x 105 cells/ml. Chemotactic Activity from ZR-75-1 CM for Fibroblasts. A A 0.2-ml aliquot of this cell suspension was added to the upper chamber. potent Chemoattractant for human embryonic, adolescent, and The chambers were incubated for 4 hr at 37°in a 5% CO2 humidified adult skin fibroblasts, as well as human lung and mouse skin atmosphere. The filters were then removed, fixed, and stained in Diff- fibroblasts (data not shown), was found in CM from the ZR-75- Quik staining solutions (Harleco), rinsed in water, and placed on micro 1 human breast tumor cell line. This activity was concentration scope slides, such that cells migrating through the filter and attached to dependent, with an ED50(that amount of material which produced the underside of the filter were against the surface of the glass. We did a half-maximal response) between dilutions of 1:8 and 1:32, not observe any cells detached from the filter and present in the lower depending on the preparation (Chart 1); the concentration of chamber. The cells on the upper surface of the membrane were wiped away with a wet cotton-tipped applicator, and the number of cells protein ranged from 70 to 100 ¿«g/ml.Whencompared to the migrating to the lower surface of the membrane was determined by chemotactic activity of PDGF (18) or fibronectin (12, 19, 23), known attractants for fibroblasts, ZR-75-1 CM induced a 2-fold counting the number of stained nuclei in randomly selected microscopic fields encompassing an area of 1 sq mm (Vi? of the total area of greater number of cells to migrate. Attempts to concentrate the migration). Nonstimulated migration of cells, determined in control cham medium resulted in the formation of a precipitate with a loss of bers, with defined medium as the attractant, accounted for less then some activity from the soluble fraction. Therefore, it was not 10% of the maximal response and was subtracted from all values. All possible to measure an increase, if any, in chemotactic activity assays were performed in duplicate. These values did not differ by more at concentrations beyond that at which a precipitate formed. than 10%, and all results represent means of 2 or more experiments. We wanted to determine whether this activity was different Treatment of CM from other fibroblast chemoattractants such as PDGF and fibro nectin. Addition of maximally stimulating concentrations of either Stability to pH. CM was adjusted to pH 2 to 12 with either 1 N HCI or PDGF or fibronectin to cells in the upper chamber with CM in 1 N NaOH and allowed to stand for 1 hr at room temperature. CM was the lower chamber did not significantly decrease (10%) the adjusted back to pH 7.0 and then centrifuged at 10,000 x g for 10 min. migration of fibroblasts below that produced by CM alone (Table Volumes were adjusted with DMEM such that they were all equal, and 1). Addition of either PDGF or fibronectin to the cells in the upper assays for Chemotaxis were performed on the samples. Enzymatic Treatment. CM was treated with either DNase (2 //g/ml). RNase (50 /¿g/ml),or trypsin (100 ng/m\) at 37°for 1 hr. Trypsin was 300 J inhibited by the addition of aprotinin (1 mg/ml). The samples to be treated with pepsin were adjusted to pH 3 with 1 N HCI, and then pepsin in 0.01 N HCI was added to a final concentration of 100 pg/ml. Following incubation at 37°for 1 hr, pepsin was inactivated by raising the pH to 7.0 with 1 N NaOH. Treated samples were then assayed for their ability to stimulate Chemotaxis. Reduction and Alkylation. CM was treated with either 20 mw 2- \ 200 mercaptoethanol or 1 mM dithiothreitol for 2 hr at room temperature, at neutral pH. One-half of each sample was further treated for 1 hr at 4° tr with iodoacetic acid (15 mg/ml) (0.08 M), and solid Tris was added to C3 bring the pH above 7.5. All samples were dialyzed against DMEM and CO then tested for chemotactic activity. _l Gel Filtration. CM from ZR-75-1 breast tumor cells was dialyzed against Tris-buffered saline (0.05 M Tris-HCI: 0.15 M NaCI, pH 7.4) and 0 100 then concentrated 10-fold in an Amicon ultrafiltration cell. Medium con centrated above a PM-10 membrane retained chemotactic activity. The concentrated CM was applied to a Sephadex G-200 (Pharmacia Fine Chemicals) column (28 x 1.5 cm) and eluted with Tris-buffered saline. Fractions (1.2 ml) were dialyzed against DMEM and assayed for chem otactic activity. 1:32 1:8 1:2
In Vivo Delivery of CM DILUTIONS Chart 1. Dose-response curve of human skin fibroblasts to ZR-75-1 CM, PDGF, ZR-75-1 CM was concentrated 10-fold in an Amicon Ultrafiltration cell and fibronectin. Two-day ZR-75-1 CM (•),PDGF(O), or fibronectin (A) were added over a PM-10 membrane. Alzet osmotic minipumps (Model 1702; Alza to the lower well of the Boyden chamber. The undiluted concentration for PDGF was 2 units/ml and, for fibronectin, it was 80 iigfrii. The number of fibroblasts Corp.) which release approximately 0.5 /¿l/hrfor up to 2 weeks, were migrating through the Nudeopore filter per sq mm was determined. Each point loaded with the 10x CM, defined medium, or defined medium containing represents the mean of 4 experiments; öars,S.E.
AUGUST 1984 3399
Downloaded from cancerres.aacrjournals.org on September 24, 2021. © 1984 American Association for Cancer Research. W. E. Gleiber and E. Schiff mann
Table 1 is also stable to more than 10 freezing and thawing cycles. Chemotactic response of fibroblasts to various attrae!anÃs The stability to pH of the attractant was determined by assay ing CM for chemotaxis following acid or base treatment (Table % of maximal LowerchamberZR-75-1 (cells)Fibronectinchamber response"100 3fl). The CM did not lose activity over a wide range of pH (3 to CMFibronectin ±7" 11). However, chemotactic activity was inhibited by treatment of (20 M9/ml) 37 ±4 CM at very acidic pH (<2.5) or very basic pH (>11.5). ZR-75-1 CM (20 ug/ml) 92 ±5 The CM was treated with various enzymes to determine the PDGF(0.5 unit/ml) 53 ±4 ZR-75-1 CMUpper PDGF(0.5 unit/ml)Chemotaxis: 89 ±7 nature of the chemotactic activity (Table 3C). DNase or RNase * The maximal chemotactic response was 372 ceiis/sq mm = 100%. treatment did not reduce activity appreciably, while treatment Mean ±S.E.of 3 experiments. with either trypsin or pepsin reduced chemotactic activity by 80 to 90%. In addition, treatment of CM with the reducing agents chamber did not stimulate motility in the absence of CM in the 2-mercaptoethanol or dithiothreitol also inhibited the activity by lower chamber (not shown). When gelatin-coated membranes about 80 to 90%. Alkylation with iodoacetic acid following reduc were coated additionally with fibronectin or when purified antifi- tion inhibited the chemotactic activity to about the same extent bronectin was added to the lower well, fibroblast chemotaxis to as did reduction alone. These studies suggest that the che- ZR-75-1 CM was not inhibited. As expected, these treatments moattractant is a protein. inhibited fibroblast chemotaxis to fibronectin (7) (data not shown). These conclusions are tentative, since CM and not purified Table 2 Checkerboard analysis of Zfì-75-l CM attractant was used. It is conceivable, however, that the CM Dilution from the tumor cells may contain proteases which degrade both of ZR-75- fibronectin and PDGF. In such a case, the effects of these 1 CM No. of fibroblasts migrating through nucleoporefilters/sq mm at attractants in the migration assay obviously would be diminished. placed in upper chamber ZR-75-1 CM dilutions of: Furthermore, the antibody to fibronectin used here was prepared tower chamber 0 1:8 1:4 1:2 1 against plasma fibronectin and may not react with the cell-derived protein. However, as is detailed below, the tumor-derived attrac 27 ±1.2 18 ±4.5 tant differs from fibronectin with respect to molecular weight and stability to a broad pH range and from PDGF with respect to molecular weight and stability to heating. In addition, the CM attractant produced a greater maximally chemotactic response than did either fibronectin or PDGF. Also, in a previous study (7), 1:4 both plasma and cell-derived fibronectin produced essentially equivalent responses in fibroblasts. However, while not conclu 1:2 226±7.1 170±5.9 116±2.3\62 ±1.8\4 ±0.7 sive, the results suggest that breast carcinoma cells produce a factor distinct from PDGF and fibronectin that induced greater 1 272±11 198±11 cell migration than did these 2 components (Chart 1). The attractant is not, apparently, a growth factor, because CM does * Mean±S.E.of 2 experiments. not stimulate DNA synthesis, protein synthesis, or doubling times of human embryonic fibroblasts in culture.3 Table 3 A "checkerboard analysis" (27) was performed (Table 2) to Characterization of ZR-75-1 CM determine whether the CM from ZR-75-1 cells had true chemo of maximalchem CMStability of ZR-75-1 otacticresponse*100±8t>64±216 tactic activity (directed migration) as well as chemokinetic activity A.B.C.Treatmentof chemotactic activity to heating (random migration) for fibroblasts. Different concentrations of 37°1hr56" CM were placed both above and below the Nuclepore membrane 1hr100° 1hrStability in order to present varying concentration gradients to the fibro ±826 blasts. The fibroblasts respond chemotactically to the ZR-75-1 pHc 3400 CANCER RESEARCH VOL. 44 Downloaded from cancerres.aacrjournals.org on September 24, 2021. © 1984 American Association for Cancer Research. Tumor-produced Flbroblast Chemoattractant The ZR-75-1 CM was chromatographed by gel filtration on a simply produce less of a material like that of the ZR-75-1 cells. Sephadex G-200 column (Chart 2). The chemotactic activity was In Vivo Assay of CM. Since the ZR-75-1 CM attracted fibro eluted from the column over a broad range of molecular weight blasts in vitro, we determined what effect the CM had in vivo. (Mr 80,000 to 150,000). The major portion of biological activity Mice were used in these studies, since the CM was a potent was associated with, at most, 20% of the protein, as measured attractant for murine fibroblasts as well as for human fibroblasts by A280. (Table 4). For up to 6 days, there was no difference in the fibrous Chemotactic Activity of CM from Other Tumor and Normal reaction to pumps implanted with CM or control medium (Chart Cell Lines. We examined CM from other breast tumor cell lines 3). Beyond Day 6, the fibrous mass did not increase at pumps to determine whether this fibroblast Chemoattractant was unique implanted with medium alone or with medium containing a foreign to the ZR-75-1 cell line. CM obtained from the human breast protein, BSA. However, the mass of fibrous tissue at the CM tumor cell lines MCF-7, T-47D, and ZR-75-31 contained attrac- pumps increased markedly (200 mg). Examination (not shown) tants as potent as were those found in ZR-75-1 CM (Table 4). of histological sections taken from these tissue masses revealed Dose-response curves and preliminary characterization indicate large numbers of fibroblasts, inflammatory cells and, in some a chemotactic factor with properties similar to that of ZR-75-1 sections from CM masses, areas of fibroplasia. The fibrotic mass (data not shown). In all cases, dilution of the CM resulted in a which formed at the heads of the pumps containing bovine serum reduced chemotactic response. Nonfibrotic tumor cell lines and albumin at a concentration (1.5 mg/ml) twice the protein concen normal cells were also tested for the production of attractants. trations of CM did not differ in both appearance and weight from CM from these cell lines (Table 4) contained chemotactic activity, that which formed at the heads of pumps loaded with medium but the response in each case for the undiluted medium does alone. Since bovine serum albumin is not an attractant for not exceed 50% of that seen for the breast tumor CM. However, fibroblasts, our results are in accord with the suggestion that the it cannot be ruled out that normal and nonfibrotic tumor cells CM contains an attractant for fibroblasts which is acting in vivo. DISCUSSION We have investigated whether CM from breast tumor cells contains fibroblast chemoattractants. Such factors may contrib ute to the development of fibrosis, which occurs in a majority of human breast tumors. CM from 4 different human breast tumor cell lines contain a potent Chemoattractant for fibroblasts. This chemotactic activity is 2 to 5 times greater than that found for CM from several nonfibrotic tumor cell lines or normal cell lines. The chemotactic activity for fibroblasts in the CM appears to be a true Chemoattractant causing directed migration of the fibro blasts and not merely random migration of the cells due to increased chemokinetic activity. This was indicated by a "check VOLUME (ml) erboard analysis" (27), which showed a positive correlation be Chart 2. Gel filtration of ZR-75-1 CM. Concentrated ZR-75-1 CM in Tris-buffered tween migration of fibroblasts and an increased gradient of CM. saline (0.05 M Tris-HCI:0.15 M NaCI, pH 7.4) was applied to a Sephadex G-200 column and eluted with Tris-buffered saline. Absorbance at 280 nm ( ) was The Chemoattractant in the CM appears to be a protein with determined, and samples were then dialyzed against OMEM and assayed for a molecular weight between 80,000 and 150,000. It is stable to chemotaxis. The number of tibroblasts migrating through the Nudepore filter per heating up to 56° and over a wide pH range (3 to 11). A sq mm (•)wasdetermined. Molecular weight markers indicated are: 758K, aldolase; 65K, albumin; and 23K, ovalbumin. Chemotaxis assays were performed in triplicate, and the S.E. did not exceed 6%. Chemotaxis ofTable fibroblasts 4 to CM Chemotaxis (% of maximal Source of CM response)3 Human breast tumor cell lines ZR-75-1 100±6" MCF-7 97 ±9 T-47D 96±7 ZR-75-31 90±2 0 2 4 6 8 10 12 14 Tumor cell lines DAYS Human renal 53 ±4 Human mucoepidermoid Chart 3. Formation of fibrotic tissue in vivo in response to ZR-75-1 CM. Con Human Ewing's sarcoma 52 ±4 centrated ZR-75-1 CM, defined medium, or BSA at 1.5 mg/ml was loaded into 29 ±2 Murine melanoma 37 ±1 Alzet osmotic minipumps. These were implanted s.c. into the lower flank of mice. On Days 4, 6,9, and 11 after insertion of the pump, the fibrous tissue mass which Normal cell lines formed at the head of the pump was excised and weighed. Fibrous tissue masses Human embryonic fibroblasts 47 ±2 formed at the head of pumps implanted for 11 days containing BSA in defined medium (O), defined medium (•),orZR-75-1 CM (A). Each point tor ZR-75-1 CM Bovine aortic endothelial 33±1 Murine embryonic fibroblasts 28 ±1 and defined medium represent the mean weight of fibrous tissue at 2 pumps, one * The maximal chemotactic response of undiluted medium was 358 cells/sq mm placed singly in a mouse and the other from a mouse with a ZR-75-1 CM pump on one side and a control defined medium pump on the other side. BSA-containing " Mean ±S.E. of 3 experiments. pumps were placed singly in mice, and each point represents the fibrous tissue weight from a single mouse. AUGUST 1984 3401 Downloaded from cancerres.aacrjournals.org on September 24, 2021. © 1984 American Association for Cancer Research. ••:.-f» W. E. GleiberandE. Schiff mann secondary structure involving disulfide bond formation appears mediators, hormones, or growth factors (8,10, 26). to be necessary, since reduction and alkylation results in inacti- Further studies of the role of this fibroblast chemoattractant vation of the fibroblast chemotactic activity. The stability of the will not only help to define the biology of the breast carcinoma activity at 56°but not above 80°,and an apparent molecular but also could give insight into the involvement of fibroblasts in weight greater than 100,000, indicate that it is not a complement- normal wound repair processes. derived factor (15) or PDGF (16). In this latter regard, PDGF placed together with cells did not inhibit their chemotactic re REFERENCES sponse to tumor-conditioned medium. Also taken together, the 1. Allegra, J. C., and Lippman, M. E. Growth of a human breast cancer cell line in serum-free hormone-supplemented medium. Cancer Res., 38: 3823-3829, failure of attempts to inactivate CM with antibody to fibronectin, 1978. the inability of fibronectin to compete with the tumor-derived 2. Barnes, D., and Sato, G. Growth of a human mammary tumor cell line in a serum-free medium. Nature (Lond.), 281: 388-389,1979. attractant during chemotaxis, the apparent molecular weight of 3. Barnes, D. and Sato, G. Serum-free cell culture: a unifying approach. Cell, 22: the CM attractant, and the apparently greater capacity of CM to 649-655,1980. induce a chemotactic response compared with those of fibro 4. Chiang, T. M . Postlethwaite. A. E., Beachey. E. H., and Seyer. J. M. Binding of chemotactic collagen-derived peptides to fibroblasts. J. Clin. Invest., 62: nectin and PDGF make it unlikely that the attractant is fibronectin 916-922,1978. or PDGF. Therefore, the tumor-derived chemoattractant appears 5. Engel. L. W., Young, N. A„Tralka,T. S., üppman,M. E., O'Brien, S. J., and to be different from any of the known fibroblast attractants. Joyce, M. J. Establishment and characterization of three new continuous cell lines derived from human breast carcinomas. Cancer Res., 38: 3352-3364, The CM appeared to act as a chemoattractant in vivo, pro 1978. ducing a fibrotic response when it was implanted in a mouse 6. Foster, C. S., and Neville, A. M. Histopathology of breast cancer. In: R. C. with the aid of a slow-release device. Since the CM is a foreign Coombes, T. J. Powtes, H. T. Ford, and J. C. Gazet (eds.), Breast Cancer Management, pp. 17-76. New York: Academic Press, Inc., 1981. antigen to the animal, it was possible that the fibrous tissue 7. Gauss-Mülter,V., Kteinman, H. K., Martin, G. R., and Schiffmann, E. Rote of mass which formed was the result of an immunological reaction. attachment factors and attractants in fibroblast chemotaxis. J. Lab. Clin. Med., However, when bovine serum albumin was placed in the pump, 96: 1071-1080,1980. 8. Gospodarowicz, D., Greenburg, G., and Birdwell. C. R. Determination of cellular it did not produce a greater fibrotic response than did control shape by the extracellular matrix and its correlation with the control of cellular medium alone. This suggests that the mass formed with the CM growth. Cancer Res., 38: 4155-4171,1978. was not the result of a nonspecific immunological or inflammatory 9. Grotendorst, G. R., Seppa, H. E. J., Kteinman, H. K., and Martin, G. R. Attachment of smooth músete cells to collagen and their migration toward reaction. In addition to its direct chemotactic effects, CM may platelet derived growth factor. Proc. Nati. Acad. Sci. USA, 78: 3669-3672, induce a fibrotic reaction indirectly. The presence of CM in vivo 1981. might play a role in the production of host-derived attractants. 10. Guzmar, R. C., Osbom, R. C., Yang, J., DeOme, K. B., and Mandi, S. Transplantation of mouse mammary epithelial cells grown in primary collagen While we have no evidence that the attractant alone in vitro is a gel cultures. Cancer Res., 42:2376-2383,1982. 11. Postlethwaite, A. E., and Kang, A. H. Characterization of a guinea pig lympho growth factor for fibroblasts, it is conceivable that an effect of cyte-derived chemotactic factor for fibroblasts. J. Immunol., 124:1462-1466, the CM in vivo might generate a mitogenic activity. Additionally, 1980. compared with the effects of CM, the growing tumor may be 12. Postlethwaite, A. E., Keski-Oja, J., Balian, G., and Kang, A. H. Induction of fibroblast chemotaxis by fibronectin. J. Exp. Med., 753: 494-499,1981. more effective, causing a desmoplastic response by producing 13. Postlethwaite, A. E., Seyer, J. M., and Kang, A. H. Chemotactic attraction of both mitogenic and chemotactic factors. human fibroblasts to type I, II, and III collagens and collagen-derived peptides. We have found, in CM from a number of other cell lines derived Proc. Nati. Acad. Sci. USA, 75: 871-875,1978. 14. Postlethwaite. A. E., Snyderman, R., and Kang, A. H. The chemotactic attrac from fibrotic breast carcinomas, levels of chemotactic activity tion of human fibroblasts to a lymphocyte-derived factor. J. Exp. Med., 144: similar to that from the ZR-75-1 cell line. However, CM from 1188-1203,1976. nonfibrotic tumor cell lines and normal cell lines gave a response 15. Posttethwaite, A. E., Snyderman, R., and Kang, A. H. Generation of a fibroblast chemotactic factor in serum by activation of complement. J. Clin. Invest., 64: 50% or less than that of the fibrotic tumor CM. This suggests 1379-1385,1979. that the production of a potent fibroblast attractant is either 16. Ross, R. and Vogel, A. The platelet-derived growth factor. Cell, 14: 203-210, 1978. unique to the fibrotic tumor cells or that these cells are producing 17. Schiffmann, E., and Gallin, J. I. Biochemistry of phagocyte chemotaxis. Curr. greater amounts of a common attractant. Production of antibody Top. Cell. Regul., 15: 203-261, 1979. specific for the ZR-75-1 breast tumor chemoattractant will permit 18. Seppä, H., Grotendorst, G., Seppä, S., Schiffmann, E., and Martin, G. R. Platelet-derived growth factor is chemotactic for fibroblasts. J. Cell BioL, 92: both qualitative and quantitative determinations of similar attrac 584-588,1982. tants in other cells. We could also determine the levels of 19. Seppa, H. E. J., Yamada, K. M., Seppä,S. T., Silver, M. H., Kteinman, H. K., and Schiffmann, E. The cell binding fragment of fibronectin is chemotactic for attactants produced in both primary breast tumor and normal fibroblasts. Cell Bid. Int. Rep., 5: 813-819,1981. tissue and determine whether the amount produced varies with 20. Shoshan, S. Wound healing. Int. Rev. Connect. Tissue Res., 9:1-26,1981. the degree of fibrosis seen in the disease. In addition, we have 21. Soute, H. D., Vazquez, J., Long, A., Albert, S., and Brennan, M. A human cell found that the CM contains chemoattractant activities for endo- line from a pleural effusion derived from a breast carcinoma. J. Nati. Cancer Inst., 51: 1409-1416,1973. thelial and smooth muscle cells but not for neutrophils. Whether 22. Spector, W. G. The fibroblast and inflammation. In: E. Kulonen and J. Pikka- it is a single material that affects both cell types and whether it rainen (eds.). Biology of the Fibroblasts, pp. 525-528. New York: Academic Press, Inc., 1973. has additional activity for these other cell types need to be 23. Tsukamoto, Y., Helsel, W. E., and Wahl, S. M. Macrophage production of determined. fibronectin: a chemoattractant for fibroblasts. J. Immund., J27: 673-678, Our data suggest that, in addition to the role of inflammation 1981. 24. Vorherr, H. Breast Cancer: Epktemtology, Endocrinology, Biochemistry, and in fibrotic disease, the production of a chemoattractant for fibro Pathobiology, pp. 347-373. Baltimore: Urban & Schwarzenberg, 1980. blasts by a tumor could be a major contributor to fibrosis. It is 25. Wilkinson, P. C. Chemotaxis and Inflammation, pp. 33-53. London: Churchill- Livingstone, 1974. not clear whether the induction of fibrosis benefits the tumor or 26. Yang, J., Richard, J., Bowman, P., Guzmar, R., Enami, J., McCormick. K., the host. Fibrotic involvement of the tumor could enhance growth Hamamoto, S., Pitelka. D., and Nandi, S. Sustained growth and three-dimen and metastasis by providing a connective tissue matrix for the sional organization of primary mammary tumor epithelial cells embedded in collagen gels. Proc. Na«.Acad. Sci. USA, 76: 3401-3405,1979. cells. This may result in a favorable 3-dimensional organization 27. Zigmond. S. H. and Hirsch, J. G. Leukocyte locomotion and chemotaxis. J. for the tumor cells to respond to or interact with adjacent cells, Exp. Methods, 137: 387-410,1973. 3402 CANCER RESEARCH VOL. 44 Downloaded from cancerres.aacrjournals.org on September 24, 2021. © 1984 American Association for Cancer Research. Identification of a Chemoattractant for Fibroblasts Produced by Human Breast Carcinoma Cell Lines Wayne E. Gleiber and Elliott Schiffmann Cancer Res 1984;44:3398-3402. Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/44/8/3398 E-mail alerts Sign up to receive free email-alerts related to this article or journal. Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected]. Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/44/8/3398. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site. Downloaded from cancerres.aacrjournals.org on September 24, 2021. © 1984 American Association for Cancer Research.