GENETICS OF SOMATIC CELLS 441 



specific activities." 1173 Such limitations may concern, for example, the enzymes of 

 nucleic-acid synthesis, and there is experimental evidence to indicate that different 

 tissues in the same species may differ in their ability to utilize preformed pyrimidines. 1378 

 If such differences can arise in the course of differentiation, there is no reason why they 

 could not arise in tumor cells by processes akin to those governing differentiation and 

 not necessarily corresponding to mutations at the genie level. If differentiation normally 

 operates by limiting genetic potentialities, some of the enzymatic losses accompanying 

 the development of resistance to drugs in tumors 137 ' 1048 might be epigenetic rather 

 than genetic. This question cannot be approached experimentally at present and 

 there is little likelihood of any substantial progress in this regard until some method of 

 genetic analysis can be worked out for somatic cells. 



Another field, closely related to drug sensitivity and, in addition, of considerable 

 relevance for the problems of tumorigenesis, cellular autonomy, and somatic variation, 

 deals with the dependence of neoplasms originating in endocrine organs or their target 

 tissues on the original hormonal imbalance that induced them, this being usually an 

 overstimulation by a given hormone or the lack of a normally inhibiting hormone. This 

 field has been reviewed repeatedly, 87 - 4&1, 412 ' 414 - 413, 421, 917 and only some points 

 relevant to the present discussion will be emphasized. It appears that prolonged stimu- 

 lation of certain target tissues with superimposed hormones often leads to neoplasms the 

 growth of which at first depends on continued stimulation. These forms are usually 

 called dependent or conditioned tumors. The same result can be achieved by the 

 removal of an inhibiting hormone ; the thyrotropic pituitary tumors in mice arising 

 after the removal of the thyroid illustrate this. When observed for extensive periods 

 of time and, if necessary, through serial transplantation, it is the rule rather than the 

 exception that such tumors change to a more autonomous state, or show a tendency to 

 throw off autonomous variant sublines. These no longer require the hormonal stimulus 

 or the absence of the hormonal inhibitor that was originally needed for tumor induction 

 and maintenance. This progression to autonomy is random, unpredictable, although 

 bound to happen sooner or later. The probability of its occurrence varies from system 

 to system, and, sometimes, from tumor to tumor within a single morphologic and etio- 

 logic group. In certain systems, a series of successive steps can be distinguished, from 

 full dependence, through decreasing levels of partial dependence, to full autonomy. 32, 413 

 There is a certain resemblance between these phenomena and resistance to drugs, and, 

 when viewed together, it is difficult to escape the conclusion that, whenever it is possible 

 to influence a population of tumor cells by natural or artificial, stimulating or inhibit- 

 ing, chemical or hormonal factors, it will exhibit a varying degree of plasticity, tending 

 to escape from the limiting factor. Here, as in the case of drug resistance, it is most 

 plausible to assume that evolutionary mechanisms of the variation-selection type are 

 responsible. Unfortunately, no case of hormonal dependence has yet been subjected 

 even to the preliminary type of analysis from the point of view of population dynamics 

 that is available for some cases of drug resistance. It is the opinion of the reviewer 

 that this is an important and potentially fertile field for investigation, particularly 



