GENETICS OF SOMATIC CELLS 461 



could be obtained from the same tumor; the isoantigenic losses were stable and irre- 

 versible after return to the F x genotype of origin; and different tumors of similar origin 

 and histology were individually different with regard to variant formation. Somatic 

 crossing over was considered as a possible explanation of this phenomenon. A certain 

 parallel could be drawn between isoantigenic variant formation and tumor progression. 

 The former can be viewed as the loss of isoantigens preventing growth in a foreign 

 host genotype, while the latter is essentially a gradual loss of responsiveness to various 

 growth-controlling forces in the autochthonous host. As a model system, variant 

 formation demonstrated that irreversible loss of certain genetically determined cellular 

 constituents may, under appropriate conditions, convey upon the cell a new ability 

 to grow under circumstances in which its progenitor would have been checked by a 

 superimposed, systemic mechanism. 



At the population level, the dynamics of some changes in population have been 

 analyzed. While a variation-selection mechanism has been demonstrated for some 

 cases of drug resistance, certain changes in antigenic and transplantation characteristics, 

 ascites conversion, shift of predominant karyotype, and the establishment of amelanotic 

 melanomas, there is also one instance of a host-induced adaptive modification and an 

 automatic change in responsiveness due to increase in population size. Specific 

 pairs of tumor sublines, selected from a common original population, but differing 

 with regard to well-defined unit characteristics related to progression, appear particu- 

 larly suitable for the study of the cellular mechanisms determining the phenotype in 

 question. 



The recent striking developments in the field of viral tumors have introduced the 

 concept of infective heredity into the tumor field. Some form of integration between 

 viral and cellular genomes probably exists, at least with certain tumor viruses. In 

 some cases, new cellular characteristics appear concurrently with viral oncogenesis, 

 not directly related to neoplasia itself, suggesting a possible analogy with lysogenic 

 conversion. 



Tissue-culture systems provide excellent models to study somatic variation. 

 Available markers comprise drug, virus, and radiation resistance and nutritional, 

 morphologic, and chromosomal characteristics. Fluctuation tests and cloning tech- 

 niques can be directly applied. Unfortunately, the problem of representativeness is 

 still overwhelming and it seems as if only a limited number of cellular types could be 

 propagated in present tissue-culture media. The development of new media that will 

 permit a larger variety of cells to grow in vitro and the application of cloning techniques 

 to cells removed directly from the body are urgently needed before tissue-culture 

 methods will become directly useful for the study of the genetics of somatic cells as they 

 exist in the body. 



The question of malignization of tissue-culture cells of normal origin is presently 

 rather confused due to the discovery that many of the so-called "altered" lines have 

 been contaminated with established and highly virulent lines of malignant cells. 

 There is no doubt that malignization in vitro does occur, but nothing can be said about 



