S been applied to the study of the biology of normal

Transcription

1 Application of In Vitro Soft Agar Techniques for Growth of Tumor Cells to the Study of Colon Cancer RONALD N. BUICK, PHD,* STEPHEN E. FRY, MS,t AND SYDNEY E. SALMON, MD$ An in vitro assay to measure the clogenic or colony-forming capability of cancer cells present in biopsy samples has recently been applied to study the biology and drug-sensitivity -of a variety of human neoplasms. This approach appears to be suitable for study of the tumor stem or progenitor cells present in malignant effusions from patients with colonic carcima. In our preliminary studies, morphology of the tumor colonies by inverted microscopy and with Papanicolaou staining of dried agar plating layers as well as immufluorescent localization with a specific antiserum to human carciembrionic antigen have been used as markers of the neoplastic origin of colon tumor colony-forming cells. Successful application of this assay to colonic solid tumors will require improvement in techniques for disaggregation of viable clogenic cells. We anticipate that short term clonal assays will have increasing use for clinical and biological studies of human colon cancer. Cancer EMI-SOLID MEDIUM CLONING TECHNIQUES have S been applied to the study of the biology of rmal murine and human cell-renewal system progenitor/ stem cells since the work of Pluznik and SachsI4 and Bradley and Metcalf' in the middle 1960s. In addition, investigators at the Ontario Cancer Institute have shown that tumor stem cells responsible for population renewal and for the cloning property of a metastatic neoplasm, from murine lymphoma or myeloma, could be assayed in vivo or in vitro with tumor colony-forming assays*2,' 1.13 Classic cell kinetic parameters have long indicated that human tumor cell populations can also be regarded as cell renewal systems.ix Semi-solid growth conditions have been employed using these principles to quantitate powerful way of assessing proliferation of a mirity tumor progenitor cells in n-lymphocytic le~kemia.~.~~ The ability to form colonies in soft-agar from a variety of human cancers was recently demonstrated by Hamburger and Salmon.6 The progenitor cells in populations from a number of human tumors have w been quantitated using minimal modifications of these techniques, with regularity and reproducibility: myeloma,' ovarian carcima.* transitional cell carcima of the From the Department of Internal Medicine, Section of Hematology and Oncology, and The Cancer Center, University of Arizona, Tucson, Arizona. * Former Assistant Professor of Medicine at Univ. of Arizona, w with the Ontario Cancer Institute, Toronto, Ontario, Canada. t Research Assistant. f Professor of Medicine, Head, Section of Hematology and Oncology Director, The Cancer Center. Address for reprints: Sydney E. Salmon, MD, Departme.nt of internal Medicine, Section of Hematology and Oncology and the Cancer Center, University of Arizona, Tucson, AZ Accepted for publication June 20, X/80/0315/1238 $ American Cancer Society 1238 bladder,4 neuroblastoma," melama,'o,19 and breast carcima. The purpose of this paper is to review briefly the available information on stemiprogenitor cell-derived colony growth from various human tumors and to provide preliminary data on the development of a similar assay for colon carcima. Materials and Methods Since tumor progenitor cells might be expected to exist in mirity populations,18 selective assays are required for their study. Many attempts have been made to study such cells in suspension culture by the uptake of radioactive precursors; however, a much more population is with an assay of colony formation. In such assays, the semi-solid conditions of the media allow the progeny of single progenitor cells to remain localized. The number of such colonies provides an estimate (minimum) of progenitor cell frequency, the cellular composition gives an estimate (minimum) of both proliferative capacity and capacity for differentiation. The culture conditions required for colony formation represent part of the mechanisms regulating maintenance of the tumor cell population in culture and may be symbolic of events in vivo. Agar colony assays for progenitor cells were originally described for murine systems. Recently, a two-layer soft-agar technique has been described for human tumor stem cells.'j These basic procedures are used in all experiments described in this article. The conditioned medium derived from the adherent cells of oil-primed Balb/c mice, necessary for myeloma colony

2 No. 5 AGAR COLONY FORMATION IN COLON CANCER. Buick et a/ growth, was t used in these experiments. An underlayer of 0.5% agar in enriched McCoy's 5A medium containing 10% heat-inactivated fetal calf serum (H.I. FCS) was prepared (1 ml in a 35 mm plastic petri dish). Cells to be tested for colony formation were suspended in a plating layer of 0.3% agar in enriched CMRL 1066 medium with 15% horse serum. Specific enrichments have been detailed previous1y.l' Cells were routinely plated at a concentration of 5 x 1W cells/ml in a 1 ml plating layer. Cultures were incubated at 37 C in a 7.5% CO, humidified atmosphere. Cultures were scored with an inverted microscope at x 100 and x days after plating. Tumor Progenitor Assays: Current Status In contrast to clonal assays for hematopoiesis, hunian tumor colony forming techniques are in their infancy. The relationship of measured progenitors to, or identity with, tumor stem cells remains to be established. However, the importance of such assays has been amply demonstrated by Salmon and co-workersl6 who showed considerable progstic significance of progenitor drug sensitivity both in human myeloma and ovarian carcima. This evidence provides the basis for considering progenitor assays as important tools in the development and assessment of cancer therapy. Assays for other n-hemopoietic tumors are only w under development. The status of assays for various tumor cell types including colon cancer is summarized in Table 1. The growth requirements for many individual tumor types have t been rigorously established. For progenitors from myeloma-involved bone marrow, a conditioned medium from oil-primed mouse spleen adherent cells is r~quired.~ Colony growth from other tumor types has t demonstrated a requirement for this conditioned medium. The role of the individual constituents of the highly enriched media remains to be established. Optimal colony growth may be expected to depend on the discovery of cell-derived factors (conditioned media), and this area of research can be expected to adbance rapidly in the next few years. Colony assays for tumor progenitors can also be used to investigate cell-cell interactions important for the maintenance of the tumor cell population in culture. We have recently demonstrated a dose-dependent requirement for macrophages for colony formation from tumor progenitor cells in malignant effusions.3 Such studies point to the necessity for careful evaluation of frequency of colony formation from tumor samples containing heterogeneous cell types. The measured frequency of progenitors within tumor cell populations has proven relatively constant between tunior types, and is consistently within the range ex- TABLE 1. ProgenitoriStem Cell Assays Available for Human Neoplasms Use of Growth assay for conditions progstic Neoplasm References defined significance Hematopoietic AML Myeloma Lymphoma Solid tumor Ovarian carcima Bladder Melama Neuroblastorna Breast cancer Colon carcima (Ascites) Buick er d.? Spitzer L'I ul." Hamburger & Salmon' Jones er 111." Hamburger Pt 01.~ Buick er ii/,!j Meyskem el d. l " Von Hoff er d. I:' Von Hoff et u/.lv Von Hoff er Buick et ul. (this paper) Von Hoff er n/. '* ""' possible N.'h possible4 "' unkwn unkwn pected for a cell renewal system (approximately one clogenic cell per lo3- lo4 nucleated cells). As culture conditions are developed to approximate more closely the in vivo conditions, one would expect measured frequency determinations to show an increase from the presently available minimum estimate. Colon Cancer: Current Status All assays for clogenic cells depend on the assumption that the cells with which the culture is initiated are in a single cell suspension. We have observed that the treatments necessary to generate single cell suspensions from certain human solid tumors ((7.g. breast, colon) often result in a suspension of nviable cells. However, when tumor cells are collected from heparinized malignant effusions, a single cell suspension can almost always be obtained without compromising the cell viability. Our results show substantial difficulty in mechanically dissociating primary colon cancers and in obtaining a viable single cell suspension suitable for plating (Table 2). However, we have demonstrated clonal growth of colon cancer from all effusions so far studied. The frequency of TABLE 2. Colony/Cluster Formation from Cells Obtained from Patients with Colon Cancer Colonies/ Patient Source clusters1 Viability of cell. of cells S x 10" cells suspension (5%) 1 Solid Tumor Solid Tumor Solid Tumor Solid Tumor 0 ND 5 Effusion Effusion 28 ND I Effusion 229 > 99

3 1240 CANCER Murch Supplement 1980 Vol. 45 FIG. 1A. Typical morphology of a colon cancer colony. Cells are tightly packed (~500). 1B. Morphologic appearance of cells from colon cancer colony fixed and dried in agar layer (15). Stained with Papanicolou stain (~1250). colon cancer colonyicluster formation is consistent with the measurement of a neoplastic progenitor cell compartment within a renewing tumor cell population. The values of the three effusions studied range from 28 to 229 colonies/5 x 10; cells plated. This has also been the experience of Von Hoff,19 who has generated clonal colon cancer growth from 4/4 effusions. Figure 1 shows a typical cluster and cellular morphology within a cluster. Cluster/colony size under the growth conditions so far used has rarely been more than 40 cells. Cells within clusters have relatively tight intercellular junctions and form dense colonies. The time to reach maximum size has been three to four weeks. Cellular morphology was assessed by Papanicolou staining of a dried agar layeri5 containing clusters. Cells are large and vacuolated with heterogeneous nuclear size. To define more clearly the cells responsible for cluster formation in agar, we have ascertained the presence of CEA on the surface of cells within clusters. Indirect immufluorescence was performed on slides containing intact agar layers, which had been air-dried after glutaraldehyde fi~ation. ~ Agar layers were incubated for 30 minutes at room temperature with rabbit anti-human CEA (kindly provided by Dr. Charles Todd, City of Hope National Medical Center, Duarte, CA). They were then rinsed three times with PBS, and a second staining for 30 minutes was performed with FITC-conjugated goat, anti-rabbit IgG (Cappel Laboratories, Inc., PA). After thorough rinsing in PBS, the preparations were examined for fluorescence. Negative controls for CEA were agar layers containing mature granulocyte colonies (CFU-C) from rmal human bone marrow. Figure 2 shows a typical cluster photographed by light and immufluorescent microscopy. The light microscopic field shows the cluster, post-stained by Papanicolou stain. Prior to plating, the cells from the malignant effusion of this patient were also positive for CEA. Discussion Studies of tumor stem/progenitor cells in populations of colon cancer cells are obviously in a preliminary stage. By analogy to other tumor types and on theoretical grounds, such studies are likely to be vital for investigating the biology of the tumor for measuring anticancer drug and radiosensitivity, for developing new agents, and for assessing currently available therapeutic protocols. The inability to perform clogenic experiments from solid tumors is a severe limitation. One major area of effort will be to extend our experience to solid tumors by improving the means of developing viable single cell suspensions. Our present

4 No. 5 AGAR COloNY FORMATION IPU COLON CANCER. Buick el d FIG. 2A. Indirect immufluorescence of a colon cancer colony fixed and dried in agar layer. Staining as detailed in text (~500). 2B. Colony aligned as in Figure?A, stained with Papanicolou stain ( ~500). technique involves mincing the tumor to approximately I rnm square pieces, teasing the pieces apart with the tips of 21 gauge needles, and then passing the cell suspension through needles of decreasing size to 23 gauge. This technique has proven satisfactory for ovarian (8) and bladder carcima (4) and melama (10). Combination of aspects of these techniques with enzymatic procedures may allow individualized ge.neration of viable single cell suspensions from different tumor types. Comparing the stemiprogenitor cell assays with other systems for studying human tumor cell biology is appropriate. Two such systems, which have been discussed at length in this issue O~CNIZL~OY, are the use of human tumor xegrafts in immune-deprived mice and the development and use of continuous cell lines. Figure 3 compares these approaches to the study of tumor biology through three important characteristics: relevance to individual patient characteristics, relevance to type of disease in general, and experimental versatility. The direction of the arrows depicts an increase (i.c. improvement) in the particular characteristic. All three systems have been relevant for overall disease characteristics. In particular patients, the short-term progenitor cell assays would be expected to provide more relevant information. This expectation has been borne out in recent years by studies of the progstic significance of progenitor cell characteristics, for individual patients with leukemia, ovarian cancer and myeloma. Experimental versatility, the cell line approach has advantages over short-term colony assays. The ability to perform repeated experiments, to prepare large quantities of cellular material, and to have absolute control over cellular growth conditions has allowed a multidisciplinary approach to be applied using cell lines and, to a somewhat lesser degree, using xegrafts. Obviously, all these approaches are complementary. None have absolute advantage over ather, besides suiting the particular skills of the investigator. StemIProgenitor Assays Human Xegrafts Cell Lines Relevance to Relevance to Subcellular Individual Patient Disease Experimental Characteristics Characteristics Versatility FIG. 3. Models for study of human tumor cell biology.

5 1242 CANCER March Supplement 1980 Vol. 45 Much effort is being applied in these three systems. In human, solid tumor biology, the weight of scientific endeavor is applied to cell line work, with progressively less applied to human xegrafts and virtually ne to short-term clonal assays. In contrast, herpoietic tumor biology is founded on the reverse situation: a large percentage of investigators apply short-term clonal growth to investigate biological processes. The future should see more consideration of human tumors as renewal tissues using short-term clonal assays of tumor progenitor cells. REFERENCES 1. Bradley, T. K., and Metcalf, D.: The growth of mouse bone marrow cells in vilro. Aic~t. J. Exp. Biol. Med. Sci. 44: , Bruce. W. K., Meeker, B. E., and Valeriote, F. A.: Comparison of the sensitivity of rmal hematopoietic and transplanted lymphoma colony-forming cells to chemotherapeutic agents administered in rivo. J. Null. Cuncer Insr. 37: , Buick, R. N., Fry, S. E., and Salmon, S. E.: Effect of host-cell interactions on clogenic carcima cells in human malignant effusions. Br. J. Cancer (in press). 4. Buick, K. N., Fry, S. E., Salmon, S. E., and Stanisic, T.: Development of an agar-methylcellulose clogenic assay for cells in transitional cell carcima of the bladder. Cancer Rrs. 39: , Buick, R. N., Till, J. E., and McCulloch, E. A.: Colony assay for proliferative blast cells circulating in myeloblastic leukemia. Lancc,f 1: , Hamburger, A. W., and Salmon, S. E.: Primary bioassay of human tumor stem cells. Science 197: , Hamburger, A. W. and Salmon, S. E.: Primary bioassay of human myeloma stem cells. J. Clin. Invest. 60: , Hamburger, A. W., Salmon, S. E., Kim, M. B., Trent, J. M., Soehnlen, B. J., Alberts, D. S., and Schmidt, H. J.: Direct cloning of human ovarian carcima cells in agar. Cuncer Res. 38: , Jones, S. E., Hamburger, A. W., Kim, M. B., and Salmon, S. E.: Development of a bioassay for putative human lymphoma stem cells. Blood 53: , Meyskens, F. L., Salmon, S. E.: Studies on human melama colony forming units in soft agar. In Human Tumor Stem Cells: Clinical and Biological Studies, S. E. Salmon, ed. Alan Liss 1980, in press. 11. Ogawa, M., Bergsagel, D., and McCulloch, E. A,: Chemotherapy of mouse myeloma: Quantitative cell culture predictive of response in rzivo. Blood 4:7-15, Papanicolou, G. N.: Atlas of Exfoliative Cytology, Cambridge, Harvard University Press, 1954: p Park, C. H., Bergsagel. D., and McCulloch. E. A.: Mouse myeloma tumor stem cells; a primary cell culture assay. J. Nutl. Cancer Inst. 46: , Pluznik, D. H., and Sachs, L.: The cloning of rmal "mast" cells in tissue culture. J. Cell. Cornp. Physio[. 66: , Salmon, S. E., and Buick, R. N.: Preparation of permanent slides of intact soft agar colony cultures of hematopoietic and tumor stem cells. Cuncer Res. 39: , Salmon, S. I?., Hamburger, A. W., Soehnlen, B. S., Durie, B. G. M., Alberts, D. S., and Moon, T. E.: Quantitation of differential sensitivity of human-tumor stem cells to anticancer drugs. N. Engl. J. Med. 298: , Spitzer, G., Verma, D. S., Dicke, K. A., and McCredie, K. B.: Culture studies in vitro in human leukemia. Semin. in Hewtatology 15: , Steel, G. G.: Cytokinetics of Neoplasia. In Cancer Medicine, Holland and Frei Eds. Lea and Febiger, 1974; pp Von HOE, D. D., Johnson, G. E., and Glaubiger, D. L.: Secretion of tumor markers in the human tumor stem cell system. Proceedings of Am. Assoc. C'unc. Res., 1979.