Cancer Cell Societies and lhmor Progression

Transcription

1 Cancer Cell Societies and lhmor Progression Gloria H. Heppner Breast Cancer Program, Michigan Cancer Foundation, Detroit, Michigan, USA Key Words. Tumor heterogeneity 9 Progression Cell interactions Chemotherapy Abstract. It is now well accepted that caners are heterogeneous and contain multiple subpopulations of cancer cells that differ in important behavioral properties, such as gmwth rate, ability to metastasize and sensitivity to treatment. This review describes results using a mouse mammary tumor system to model the effect of heterogeneity on tumor growth and response to chemotherapy. %mar subpopulations that differ in growth and therapy parameters are mixed together to simulate tumor heterogeneity. The mixtures are studied in a variety of protocols designed to detect interactions between the subpopulations and to determine the mechanisms of these interactions. The central conclusion of this work is that tumor subpopulations do not behave independently of each other but rather form a society of cells in which they influence each other s growth and treatment response. The mechanisms by which they do so depend on the unique characteristics of the interacting subpopulations, on the behavior being affected and on the circumstances of their interaction. These observations suggest that the maintenance of tumor heterogeneity is a consequence of the cancer cell society and that cancer behavior and progression depend on interactions among all the members of that society, not just on its most deviant variants. Introduction It is well recognized that cancers are heterogeneous and dynamic [ They are heterogeneous in that they contain mixtures of multiple, phenotypically and genotypically distinct populations of cells. These cells, and the subpopulations to which they belong, are generally thought Correspondence: Dr. Gloria H. Heppner, Breast Cancer Program, Michigan Cancer Foundation, 110 E. Warren Ave., Detroit, MI 48201, USA. Received March 10,1993; accepted for publication March 10, OAlphaMed Press / 93/$5.00/0 of as either cancerous or normal (that is, stromal cells, inflammatory infiltrates and nontransformed parenchymal cells). All of the cells, however, are participants in the pathologic process manifested as cancer. Cancer is a tissue, not a cellular, disease. To understand how cancers behave, it is necessary to analyze how the various cellular subpopulations behave together. I will here consider how subpopulations of cancer cells interact to affect growth and response to chemotherapy. The role of normal cells in these processes is beyond the scope of this review. Cancers are dynamic in that they change over time, a process called progression [4]. Change comes from two forces: alterations in genes of the neoplastic cells, or in the expression of genes, that lead to new neoplastic clones or variant subpopulations; and selection by host, microenvironmental and other mechanisms that act on the subpopulations to remodel the composition of the cancer. The result adds up to the behavior of the cancer at any given time. But in what way? What is the effect of heterogeneity on neoplastic behavior? For quite a few years my colleagues Drs. Fred and Bonnie Miller and I have been addressing the following hypothesis: Cancers are a cellular society; cancer cell subpopulations interact to affect each other s behavior; these interactions determine the behavior of the cancer as a whole. We have developed a number of experimental tools to test this hypothesis. Experimental Analysis of the Cancer Cell Society Our basic approach is to simulate the cancer society by mixing together subpopulations of cancer cells that differ in characteristics of interest. In these experiments we are asking two questions: 1) How does the mixed cancer behave, as compared with the behavior of the isolated STEM CELLS 1993;11:

2 He p p n e r 200 subpopulations? and [2] How does the subpopulation composition of the cancer change during growth and treatment? All of our work traces back to a single mouse mammary adenocarcinoma that arose spontaneously in a breeding female of the strain BALBIcfC3H (this strain was produced by inbreeding BALB/c mice which had been fosternursed on strain C3H mothers and thereby infected with the milk-transmitted mouse mammary tumor virus (MMTV). Some 15 years ago, Dc Dun Dexter, our colleague at that time, used morphological criteria to isolate four cancer cell subpopulations from a primary BALBkfC3H mammary tumor. These four subpopulations were termed subpopulation lines 66, 67, 168 and 68H [5]. Subsequently, DI: Fred Miller isolated a fifth subpopulation line, 410, from a spontaneous metastasis growing in the lung of a mouse bearing the tenth transplant generation of the original tumor [6]. Over the years all these lines have been characterized with regard to growth parameters in vivo and in vitro, immunogenicity, ability to metastasize and sensitivity to chemotherapy. Just as has been shown by other investigators using a wide variety of other cancers, the subpopulations of our tumor differ among themselves with regard to all these characteristics. In order to address the second of our above two questions, however, it is not enough to have a series of subpopulations that behave differently. We need to be able to quantify each of the subpopulations independently as they grow in mixtures. This required producing derivatives of our basic set of lines that maintain the behavior of interest but also express unique, quantifiable characteristics that allow us to tell them apart unequivocally. Dc Fred Miller has produced such a reference set of subpopulation lines by introducing drug-resistance markers into each member of the basic set. For example, we have subpopulation lines that are able to grow in concentrations of 6-thioguanine, 2,6-diaminopurine, or ouabain that kill other subpopulations (as well as the parental lines). We have developed methods to clone these lines quantitatively from mixed cultures in vitro and from both primary and metastatic deposits in vivo [7-91. Thus we are able to follow the fate of individual subpopulations within mixtures by means of clonigenic assays in selective media. We maintain frozen stocks of these lines and periodically recharacterize them in order to assure ourselves that we have a stable source of experimental material. We use a variety of protocols that allow us not only to determine whether there is an interac- tion between the mixed subpopulations but also to classify the mechanisms that are involved. Thus, by comparing the behavior of combinations of subpopulations in vivo and in vitro, we have demonstrated that an immune response [ 101 or metabolism of a drug by a host [ 111 may be responsible for an interaction. We also test the combinations in physical mixtures versus those separated within a common culture dish or growing at different sites on a mouse to see whether or not cell contact is necessary for an interaction. We have examples of both [ll, 121. We have used these protocols in experiments designed to see whether individual subpopulations can influence each other s growth rates or responses to chemotherapy and to measure the effect of any such interactions on the growth and therapeutic response of the reconstituted heterogeneous cancer. What have we found? Growth Interactions Within Heterogeneous Cancers One might postulate three outcomes of an experiment in which two subpopulations that differ in growth rate are mixed together: the mixture could grow as fast as the faster partner, the mixture could grow as slow as the slower partner, or the mixture could grow at some different rate, perhaps between the two extremes. We have found that, depending upon the participants, any of these three outcomes can occur. More interesting are the results of the analyses at the level of the individual subpopulations. Again, depending upon the participants, we have seen a number of outcomes. For some combinations, the subpopulation distribution in the growing mixture continues to reflect the initial distribution; the subpopulations grow at the same rate together even though they may be very different apart [ 131. For some combinations, a new equilibrium is reached, after which the subpopulations remain together, in balance. In yet other combinations one subpopulation may eventually overwhelm the tumor to the point where the second subpopulation is no longer detectable. This later outcome has been called clonal dominance and can occur in mixtures in which the eventual winner was originally less than 1% of the starting population and where the two populations actually grow by themselves at the same rate [12]. All of these outcomes are, of course, reproducible with the same mixtures and conditions. However, it is not possible to predict in advance which outcome will occur based on the growth

3 20 1 Cancer Cell Societies and Tumor Progression rates of the individual subpopulations, and the underlying shifts in subpopulation distribution do not reflect the growth of the mixture as a whole. The same sort of dichotomy between overall behavior and underlying subpopulation distribution is seen in experiments that test the effect of heterogeneity on response to chemotherapy. Therapeutic Interactions Within Heterogeneous Cancers Again, one might postulate three outcomes of an experiment in which two subpopulations that differ in sensitivity to a chemotherapeutic agent are mixed together and treated with that agent. The mixture could respond as the more sensitive population, respond as the less sensitive population, or exhibit a different level of sensitivity than either one alone. In these experiments we have used a variety of chemotherapeutic drugs, including cyclophosphamide, melphalan and methotrexate, and a variety of subpopulation combinations, and we have tested the effect of treatment in vivo and, except for cyclophosphamide, in vitro. As with the growth experiments, all three postulated outcomes occurred. Which outcome occurs is not predictable based on the direction or degree of sensitivity of the individual subpopulations. For some combinations, the overall sensitivity is as good as that of the sensitive subpopulation and for some combinations as poor as the resistant partner. We have one combination for which the mixture is less sensitive than either of the two participants alone [ 11, 14, 151. Clonal analysis of the subpopulation composition during therapy yields provocative results. Although for some combinations we do obtain the intuitively expected result, namely selection within the treated mixture of the subpopulation that is less sensitive to treatment, this is by no means always the case. A sizeable proportion of sensitive cells may yet remain in a cancer that overall shows a response to chemotherapy [ 141. Alternatively, several logs of killing of a sensitive subpopulation can be seen in a cancer that is not responding to treatment as measured by a reduction in overall growth rate or size [15]. Thus, even with cancers in which we have known sensitivities for known drugs and when we can monitor drug response on the clonal level, we are unable to predict the overall response to treatment. Tumor heterogeneity does markedly affect treatment outcome, but it does so in very complex ways. Mechanisms of the Cancer Cell Society We interpret the data showing that cancer cell subpopulations may behave differently when grown together than when grown alone to be evidence of a cancer cell society. In a society the behavior of the individual components contributes to the overall behavior of the entire population, but individual behaviors are modified by interactions among those components. What are the mechanisms of interaction in a cellular society? One of the key conclusions from our work is that the outcome of an encounter between two cell subpopulations is highly dependent upon the particular subpopulations involved and the circumstances of the encounter. This suggests that the mechanism underlying each interaction may also be unique to that interaction and, in fact, we have found that to be the case. As already indicated, some interactions require mediation by the host: subpopulation 168 grows slower in the presence of subpopulation 410 than it does alone because 410 induces an immune response to which 168 is sensitive, but 168 cannot induce the response itself [lo]. Subpopulation 410 is more sensitive to cyclophosphamide in the presence of 168 than it is by itself because mice bearing 168 metabolize the drug more effectively than do mice bearing 410 or normal mice [ 111. Neither of these interactions occur in vitro. On the other hand, some interactions can occur in vitro, as well as in vivo. Such is the case for the ability of subpopulation 4T07 to dominate line 168 [ 121. This latter example also illustrates another variable; 4T07 and 168 cells must be physically mixed for dominance to occur, suggesting that cell contact, or close proximity, is necessary for this interaction. In the above examples, interaction occurs even if the two subpopulations are growing on opposite sides of the same mouse. Similarly, line 67 and 168 cells are more sensitive to methotrexate when cocultured with line cells than they are by themselves, but this can occur when the cells are grown separately with only a common medium bathing the interacting subpopulations [ 111. Our experience with numerous other examples of subpopulation interactions leads to similar conclusions: Some interactions are host-mediated, and some only involve the interacting cancer cells. Some interactions require contact between the participants, and some appear to work by soluble paracrine factors released by one subpopulation and active on another.

4 Heppner 202 Some (therapeutic) interactions are unique to the particular chemotherapeutic drug-subpopulation combination. Others are more broadly displayed (4T07-168,4T07-66 combinations with either methotrexate or melphalan [ 14, 151. In short, the mechanisms of interaction are multiple and depend on the specific characteristics of the interacting subpopulations and on the behavior that is being influenced. Every subpopulation combination does not interact with regard to every behavioral characteristic, but the overall picture is that of mutual accommodation by whatever mechanisms are available and relevant. Implications of the Cancer Cell Society Cancer research, indeed medical research in general, is becoming increasingly more reductionist in approach. Great efforts are expended toward identifying the molecular mechanism(s) by which a cancer cell might gain ascendancy and free itself from the restraints of communal life. Cancer cells are viewed as the outlaws of the cellular world. Yet not even in our own society is the situation quite so simplistic. There are very few truly lone criminals, deviants never touched by the influences of others, for good or bad. We would know even less than we do about how our society works if we only studied those who live in the isolation of deathrow. A similarly broad perspective is also necessary if we are going to learn to understand-and control-neoplasia. Cancer processes are too much like normal processes and cancer cells too much like normal cells for us to make a useful distinction between them. Instead, we must pay attention to the entire cancer cell society and learn how the various components are able to communicate with and control each other in order to be able to intervene in the process of cancer progression. The work of my colleagues and myself suggests that the circumstances of these interactions are at least as important as their particular mechanisms to the overall outcome. The behavior of a society is not the sum of its parts, and knowing bow groups of similar individuals behave does not necessarily tell you how the heterogeneous society will behave in its complexity and richness. Population heterogeneity is the basis of biology. It is what allows for flexibility of response, for evolution, for adaptation. Accordingly, there are strong forces to maintain heterogeneity, and it is unlikely that these forces are any less active in cancer than in normal pop- ulations. Some of these forces manifest themselves by maintaining the cancer cell society by encouraging a balance among its member subgroups and by protecting each of them from total destruction through adverse selective pressures such as chemotherapy. The work described in this concise review illustrates some of these principles. The application of these principles is a big challenge. Acknowledgment This study was supported by US Public Health Grant CA References 1 Nowell PC. The clonal evolution of tumor cell populations. Science 1976;194: Fidler IJ, Hart IR. Biological diversity in metastatic neoplasms: origins and implications. Science 1982;217: Heppner GH. Tumor heterogeneity. Cancer Res 1984;44: Foulds L. Neoplastic Development, Vols. I and II. New York: Academic Press, Inc., 1969, Dexter DL, Kowalski HM, Blazar BA, Fligiel Z, Vogel R, Heppner GH. Heterogeneity of tumor cells from a single mouse mammary tumor. Cancer Res 1978;38: Heppner GH, Dexter DL, DeNucci T, Miller FR, Calabresi P. Heterogeneity in drug sensitivity among tumor cell subpopulations of a single mammary tumor. Cancer Res 1978;38: Miller BE, Miller FR, Wilburn D, Heppner GH. Analysis of tumor cell composition of tumors composed of paired mixtures of mammary tumor cell lines. Brit J Cancer 1987;56: Miller BE, Aslakson, CJ, Miller FR. Efficient recovery of clonogenic stem cells from solid tumors and occult metastatic deposits. Invasion and Metastasis 1990; 10: Aslakson CJ, Miller FR. Selective events in the metastatic process defined by analysis of the sequential dissemination of subpopulations of a mouse mammary tumor. Cancer Res 1992;52: Miller BE, Miller FR, Leith J, Heppner GH. Growth interactions in vivo between tumor subpopulations derived from a single mouse mammary tumor. Cancer Res 1980;40: Miller BE, Miller FR, Heppner GH. Interactions between tumor subpopulations affecting their

5 203 Cancer Cell Societies and Tumor Progression sensitivity to the antineoplastic agents cyclophosphamide and methotrexate. Cancer Res 1981;41: Miller BE, Miller FR, Wilburn D, Heppner GH. Dominance of a mammary tumor subpopulation line in mixed, heterogeneous tumors. Cancer Res 1988;48: Heppner CH, Miller B, Cooper DN, Miller FR. Growth interactions between mammary tumor cells. In: Brennan MJ, Rich MA, McCrath CM, eds. Cell Biology of Breast Cancer. New York: Academic Press, Inc., 1980: Miller BE, Miller FR, Heppner GH. Therape '- tic perturbation of the tumor ecosystem in reco - strutted heterogeneous mouse mammary tumo!,s. Cancer Res 1989;49: Miller BE, Machemer T, Lehotan M, Heppner CH. Tumor subpopulation interactions affecting melphalan sensitivity in palpable mouse mammary tumors. Cancer Res 1991;51: