This Article Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Law, L. W. Search for Related Content PubMed PubMed Citation Articles by Law, L. W.

Differences between Cancers in Terms of Evolution of Drug Resistance^*

( National Cancer Institute, National Institutes of Health, Bethesda, Md.)

Resistance to several carcinostatic drugs has been established in neoplastic cells. The resistant variants in the lymphocytic neoplasm L1210, developed in short periods of time through the use of antifolic and antipurine compounds, appear to arise in a discrete, stepwise fashion, resembling the penicillin pattern of microorganisms. In several other lymphocytic neoplasms resistance was developed but not always with ease. It is to be expected, however, that the ease with which the character develops will vary from one neoplasm to another and from one drug to another. Mutation and selection appear to constitute the mechanism involved. The changes are shown to be stable, irreversible, and heritable, and persist in the absence of the drug used in selection. The possibility should not be ignored that the pattern of resistance and the mechanism involved may be quite different in other neoplasms responding to these same compounds.

It is indicated that another pattern of resistance, developed through the use of azaserine, in a plasma cell neoplasm (70429) of C3H mice, occurs rapidly, probably in a single step, to a high level of resistance, resembling in this respect the streptomycin pattern of microorganisms. Details of this problem are now being worked out.

The possibility that physiologic adaptation plays a part in the development of nonsensitivity to drugs must be considered, though in our limited experience it has not been encountered. It is evident that these nonheritable, unstable changes in populations of cells, especially among bacteria, are more common than was formerly believed (see Ravin [73]). The mechanisms underlying this form of adaptation as distinct from the stable, irreversible, heritable types are undoubtedly quite dissimilar. Therefore, at the experimental level, when nonsensitivity to a drug arises, it should be determined to what extent the population of neoplastic cells has responded to the adaptive stimulus and, if the response is discontinuous, to what extent it is due to a heritable differentiation among the cells of that population with respect to ability to adapt. Successful therapy depends upon this knowledge. These distinctions can be made in experimental neoplasms, but it appears impossible, or at least not feasible, to do so presently in neoplasms of man.

By analogy, resistance in leukemic children to the agents known to be effective in the mouse is assumed to be a result of variables within the cell rather than in the host. There is no definitive information on this issue, however, and attempts should be made to obtain an answer.

Although mutation and selection appear to constitute an important mechanism through which resistance arises in neoplastic cells, it is impossible at the moment to define in genetic terms the precise mechanism(s) concerned, since neoplastic cells are somatic cells and, as such, are generally thought not to lend themselves to genetic analysis. The possibility, however, of the development of procedures analogous to sexual processes for a direct analysis of neoplasms by conventional crossbreeding studies is hardly less likely than it was for E. coli, in which an efficient selective method for the occurrence of genetic recombination has been established. Thus, it may be inferred that the hereditary mechanism in bacteria is similar to that operating in higher organisms. Fusion of Sarcoma 180 cells in the ascitic form of this neoplasm has been observed, and the likelihood of direct hereditary interaction at least may be suggested. Necessary suitable markers, which appear most promising at present in lymphocytic neoplasms, are drug resistance and histocompatibility genes. Thus, the possibility of distinguishing the process of genic recombination in somatic cells may be realized. The possibility exists that in addition to these conventional modes of genetic variation there may occur among leukemic cells genetic variants brought about by agencies similar to the transforming and transducing factors known to play a role in the evolution of bacterial types. In bacteria such agents are known to produce changes in genetic characteristics analogous to those characters used as markers in leukemic cells.

More information is desirable concerning cross-resistance and collateral sensitivity in an array of different neoplasms. The experimental situation necessary is one containing several variant sublines developed through selection by several drugs. This indeed may be most difficult and entails a fairly large scale screening of neoplasms in their sensitivities to drugs and the systematic development of resistant sublines. Inferences concerning mechanisms of resistance, the possible use of associations of drugs, and particularly the use of drugs in sequence can be gained in this manner. It is premature to transfer the knowledge gained concerning cross-resistance and collateral sensitivities in the neoplasm L1210, for example, even to other lymphocytic neoplasms of the mouse.

Knowledge that resistance may arise by mutation, that there appears no known method for decreasing mutation rates (although, as mentioned previously, some progress in this direction has been achieved), and that it is unlikely that the host is able to alter the process of spontaneous mutation suggests a familiar and feasible approach, the use of combinations of agents in attempts to eliminate mutants to one drug by means of a second drug simultaneously present. The principle of combined therapy is unusually successful against the L1210 lymphocytic neoplasm with two compounds known to have independence of action, A-methopterin and 8-azaguanine, and in L5178 with A-methopterin plus 6-mercaptopurine. However, it would appear most difficult to predict effective combinations without an understanding of drug interactions involving cross-resistance, collateral sensitivity, synergism, antagonism, etc. Moreover, as shown in bacterial studies, the bacterial species affect, to some extent, the cross-resistance and sensitivity patterns (86). This points up again the danger in generalizing from results gained with a single neoplasm.

Hitchings and colleagues (33) have shown combinations of drugs, with suitably related mechanisms of action, to be synergistic, regularly and predictably, in inhibiting the growth particularly of L. casei and S. faecalis. The compounds used were analogs of purine and pyrimidine bases and folic acid analogs, all involved in a major biochemical pathway leading to nucleic acid synthesis. This pathway for mammalian cells may be represented as follows:

X, Y, and F represent intracellular forms. PGA is supplied exogenously and the transformation from X to Y (leucovorin, citrovorum factor) is followed by modifications leading probably to several coenzymes F concerned with the incorporation of one carbon (1C) fragments into the products shown. The incorporation of preformed purine and pyrimidine bases is shown as an alternative pathway.

Our studies with lymphocytic neoplasms, and other lymphomas, have been limited arbitrarily to antimetabolites related to this biochemical pathway.

The probability of success appears equally good with other limited systems, for example, the B-6-dependent system. Recently certain pyrimidine analogs, especially 6-azauracil and 6-uracil methyl sulfone, have been shown to be effective inhibitors of several lymphocytic neoplasms (24). The mechanism of inhibition of growth in bacterial systems of this group of compounds has been detailed by Welch and colleagues (see Welch [93]), and cogent reasons exist for the development of effective uracil and orotic acid antagonists for chemotherapeutic application.

It should be pointed out here that the emphasis placed on this whole group of compounds as nucleic acid inhibitors is not because they are unique in this respect or that other tissue components are not equally affected. There are, however, many known reactions concerned with the biosynthesis of the basic components of nucleic acids, and the possibilities of elucidating the mechanisms of action of these compounds in mammalian tissues seem especially encouraging, particularly for the antifolic compounds and for 6-mercaptopurine and azaserine.

One answer to the danger of resistance in microorganisms is the employment of associations of drugs. This practice has been unusually successful in certain situations. Clearly, without knowledge of the pattern of resistance or the mechanism of drug action, predictability of success of combinations of drugs is uncertain. In practice this is borne out by observations of a range from true synergism to simple additive effects, indifference, or even antagonism of drugs. Where resistance in neoplasms is shown to be analogous to the penicillin pattern of resistance in microorganisms, that is, developing in discrete, stepwise fashion, and presumably the result of an influence of a polygenic system, it may be predicted that simultaneous administration of such drugs as antifolics and antipurines (or antipyrimidines) will show potentiation of effects. This possibility can be determined in experimental neoplasms with the advent of new drugs where treatment may be initiated early and continued at effective concentrations. The host limitations may be so serious, however, as to preclude adequate tests.

In consideration of the effective use of drugs in the treatment, for example, of acute lymphocytic leukemia, a drug known to select for resistance in a discrete, stepwise fashion should be used at an effective concentration and continually. Decrease of the concentration below the effective level will permit the accumulation of resistant first-step mutants and allow the occurrence of higher levels of resistance, and so on. Control of these neoplastic cells is made more difficult. In view of the toxic effects of commonly used antileukemic compounds on rapidly regenerating epithelium of oral surfaces and intestine and on rapidly dividing cells of the bone marrow, it is questionable that effective levels of such drugs are ever attained. Further, the difficulties in maintaining adequate levels of the drugs commonly used are apparent. It is unlikely at present that a regimen of drug therapy, designed for effective elimination of resistant neoplastic cells, can be attained in chronic neoplastic conditions.

The use of combinations of drugs, one of which might be expected to select neoplastic cells exhibiting a streptomycin pattern of resistance, occurring in some cells to a high level in a single step, clearly would not be too efficacious.

Resistance in S. faecalis to A-methopterin, on the basis of present information, appears to be the result of a mechanism different from that determining A-methopterin resistance in leukemic cells. The definitive experiment of a comparison of "sonicates" of resistant and sensitive leukemic cells with intact cells in their abilities to convert PGA to CF, however, has not been feasible. Resistance to sulfonamides within a single species of bacteria, Staphylococcus aureus, is known to be associated with more than one type of change related, apparently, to changes in enzyme patterns (1). Also, a number of growth and enzymatic differences have been noted in S. faecalis among variant lines resistant to the same antimetabolite (36). The relationship of such difference to actual mechanisms of resistance is not yet clear. It remains to be determined whether resistance, for example, in acute lymphocytic neoplasms of the mouse, results in each case from the same mechanism. This information is necessary if one is prepared to transfer knowledge gained in studies of experimental neoplasms to analogous situations in man.

In summary, our current knowledge of leukemia (neoplasms) at the cellular level leads to the consideration of leukemic (neoplastic) cells as a population free to vary genetically within the limits set by point mutation, recombination, changes in ploidy, and possibly by transformation and transduction. A comprehensive understanding of leukemia such as to enable its control therapeutically requires that all these parameters be defined and their influences assessed quantitatively in terms of therapeutic response. In contrast to in vitro bacterial systems, complex regulatory mechanisms of the host are imposed upon this population of leukemic cells, probably leading to variations in the population which are little understood at present. It would appear that primary attention, however, should be directed toward the genetic status of the leukemic cell both as regards mechanisms of drug resistance and response to therapy. A concomitant problem is the exploitation of the host's limitations in any attempt to control leukemic cells by therapeutic measures.

^* Presented at the meeting of the Scientific Review Committee of the American Cancer Society, held at the Westchester Country Club, Rye, N.Y., March 23–25, 1956.