GENETICS OF SOMATIC CELLS 457 



We have carried out extensive experiments to trace its eventual occurrence in popu- 

 lations of neoplastic cells by appropriate isoantigenic markers, in systems where possible 

 recombinants were certainly demonstrable by highly selective methods. 716 The results 

 were wholly negative. If nuclear fusion were a regular occurrence in cellular popula- 

 tions and would lead to viable cells, the probability of segregation might perhaps be 

 increased by irradiation, known to accelerate somatic variegation in various species. 773 



DNA-mediated transformation may well be within the limits of reality. Since 

 strongly selective methods are required to demonstrate transformants even in most 

 cases of bacterial transformation, the H-2 -isoantigenic system of mice appears presently 

 most suitable among the markers available in vivo. Selectivity can be made very high, 

 so that 10 -7 — 10 -8 variants can be demonstrated. Another advantage is that the 

 genetic determination mechanism is known and possible transformations may be readily 

 distinguishable from spontaneous variation by selecting appropriate marker combina- 

 tions in which changes to new antigenic specificities do not occur spontaneously. To 

 avoid the problem of incompatibility of transformants with new antigenic specificities 

 and the enzymatic breakdown of DNA, the experiment must be carried out in vitro, 

 followed by retransplantation to appropriate selective hosts. It is hopeful that mam- 

 malian cells are known to be capable of taking up highly polymerized DNA 1214 

 within short periods of time after exposure. 



Although their mechanism of genetic determination is unknown, drug-resistance 

 markers may also be useful for transformation experiments, particularly in view of their 

 successful use in bacterial transformation. The selectivity of systems in vivo is usually not 

 high enough to demonstrate the presence of minute resistant fractions since the large 

 sensitive population may exhibit considerable residual background growth in spite of the 

 treatment. Experiments in vitro are therefore definitely preferable. Szybalski 1307 

 has recently reported that the study of drug resistance in vitro is not only feasible but 

 even readily amenable to quantitation. Multiple drug-resistance markers are probably 

 preferable to simple ones. Again, care must be taken to distinguish transformation 

 from spontaneous variation by appropriate controls. Another source of error, found 

 in some of our unpublished experiments when amethopterin-sensitive leukemic cells 

 were treated with DNA derived from resistant variants, lies in the possibility that 

 sensitive cells may be protected from the drug by some physiologic or chemical action 

 of DNA that may extend into the period subsequent to the removal of the nucleic 

 acid. Controls should therefore always include cells treated with DNA derived from 

 the sensitive type. 



It is impossible to predict the correct conditions that may permit the possible 

 occurrence of transformation. Obviously, care must be taken to minimize enzymatic 

 decomposition of DNA. On the cellular side, competence is a highly elusive property, 

 difficult to. control even in bacterial systems and exerting a profound influence on the 

 liability of the exposed cells to undergo transformation. The experimental conditions 

 necessary to secure competence are very critical and most of them have been established 

 empirically. For this reason, negative results should not discourage the vigorous 



