GENETICS OF SOMATIC CELLS 427 



tumor were still highly sensitive to the second-set reaction initiated by one previous 

 injection of heavily irradiated cells of their own kind. Thus, while loss and weakening 

 of isoantigenic components doubtless occurs, as will be discussed below, and can give 

 rise to variant tumor sublines with specifically changed compatibilities, homotrans- 

 plantability cannot be generally explained by these phenomena. Homotransplantable 

 tumors are still isoantigenic and can be so to the same extent as comparable non- 

 homotransplantable tumors. Thus, the phenomenon of tumor homotransplantability 

 remains unexplained but perhaps more attention should be given to other characteristics 

 of transplanted tumors than isoantigenicity or resistance to isoantibodies, as pointed 

 out by Snell. 1239 The speed of vascularization of the graft 868 and its subsequent 

 growth rate are important factors that may profoundly influence the delicate balance 

 between the time elapsing before the host can develop an efficient homograft reaction 

 and the speed with which the tumor reaches irreversible size. Also, the fascinating 

 paradox of immunologic enhancement, characterized by isoantibody facilitating rather 

 than inhibiting tumor growth in a foreign genotype, may be of considerable relevance 

 and intensified studies are urgently needed on the cellular mechanisms involved. 674 



In suitable experimental systems it is possible to separate true and specific isoantigenic 

 variation from nonspecific increase of homotransplantability. We have been particu- 

 larly interested in exploiting for this purpose some of the possibilities offered by the iso- 

 genic resistant (IR) lines of mice, developed by Snell. 1242 - 1238 These IR lines have been 

 bred with the intention of establishing a coisogenic background while maintaining an 

 allelic difference at one of the histocompatibility loci. Theoretically, they are homo- 

 zygous with regard to their entire genome, except the histocompatibility gene in 

 question and a chromosomal section of undefined length around it. This expectation 

 is not completely fulfilled in reality; 795, 1238, 1245 the existing IR lines are good approxi- 

 mations at best, and it can only be stated that the number of major histocompatibility 

 differences which influence the fate of tumor transplants is usually not more than one. 

 With a more sensitive indicator, such as skin, numerous other weak differences could 

 be detected in the lines we have used; this was probably exceptional, however, and due 

 to the particular history of these lines. 1245 



Tumors originating in the F 1 hybrid outcross of two IR lines are especially suitable 

 to detect isoantigenic variation at the cellular level, 545 ' 716 ' 717> 723> 722 ' 774 - 883, 884 

 since variant cells arising by mutation or stable phenotypic changes that lead to the 

 inactivation or loss of isoantigenic components, specifically determined by a strong 

 histocompatibility factor, such as H-2, derived from one of the parental strains, may 

 be expected to become selectively compatible with the other parental strain. Model 

 experiments with artificial mixtures of tumor cells derived from known genotypes have 

 shown 723 that compatible cells can grow out selectively from populations of incompatible 

 cells even, if their proportion is very small (4 x 1 " 7 in the actual experiments) and 

 even if compatibility is mostly a matter of a single difference in a major histocompati- 

 bility barrier. 



A fairly large number of such F 1 tumors have been studied in our laboratory, 



