GENETICS OF SOMATIC CELLS 425 



and were perpetuated from one cell generation to another, they were interpreted as 

 probable mutations. 794 Cloudman 208 has also found that three different adenocarci- 

 nomas of the breast, arising at essentially the same time in a single individual mouse, 

 grew in individually different frequencies in segregating hybrids. This was interpreted 

 as indicative of differences in the genetic constitution of the three tumors. 



Little 794 has pointed out that transplantation in known and controlled genetic 

 material provides a more delicate test of physiologic and biologic differences between 

 certain neoplasms than any other available test. In his view, transplanation experi- 

 ments demonstrating "somatic mutational changes in the genetic constitution of a 

 tumor . . . afford a most helpful avenue of investigation on the nature and incidence of 

 somatic mutation as a process of importance in cancer research. " 



Since different primary tumors of the same mice may differ in their transplantation 

 characteristics, it is conceivable that different cell clones of the same tumor may also 

 differ in this respect. This has been demonstrated experimentally by Axelrad and 

 Klein 42 who compared primary tumors with their metastases, the latter presumably 

 originating from one or a few cells, and found them different in their transplantability 

 to segregating F 2 hybrids, and by Hauschka and Levan 532 who studied the transplanta- 

 tion behavior of different single-cell clones isolated by micromanipulation from the 

 Ehrlich and Krebs 2 ascites tumors, and found clear-cut differences in their homo- 

 transplantability to certain foreign genotypes. 



The differences between primary tumors have been assumed to reflect differences 

 in their genetic constitution. It has been argued that since multiple primary tumors 

 differ from each other, at least all but one of them must differ from the tissue of origin 

 at the genetic level and a mutation or mutations must have been involved in their 

 origin ; in fact, this is one type of experiment on which the mutation theory of carcino- 

 genesis has been built. 



This interpretation of changes in the host range of tumors has been complicated 

 by several findings in recent years. In particular, it has been realized increasing- 

 ly 20, 1237 that tumors may grow and kill their hosts in spite of weak histocompatibility 

 differences. When F 2 or backcross tests of the conventional type are being carried 

 out with tumors arising in inbred strains, the ratios of animals killed by progressively 

 growing tumors to survivors are usually in good agreement with Mendelian expectation, 

 indicating a difference of only a few relevant histocompatibility genes between the two 

 strains used for the outcross, seldom more and usually less than 4 or 5 (if n is the number 

 of histocompatibility genes that have to be identical in tumor and host to permit pro- 

 gressive growth, (3/4) n expresses the Mendelian expectation for the proportion of 

 animals killed by tumor in an F 2 and (l/2) n in a backcross test). Skin grafting has, on 

 the other hand, revealed a much larger number of histocompatibility differences, 

 indicating the existence of not less and probably more than 15 loci. 52, 789 Is this 

 due to the antigenic simplification which, according to Gorer, occurs frequently in 

 transplanted lines of tumor cells? 456 Not always and not necessarily. Preimmuniza- 

 tion of the recipient hybrids with normal or neoplastic cells isogenic with the animal in 



