Clonal interaction in tumours

The development of cancer is contingent on the emergence of at least one clone of transformed cells. One method used to investigate whether human tumours are monoclonal depends on the mosaicism in the normal tissues of women heterozygous for the two forms of the enzyme glucose-6-phosphate dehydrogenase (G-6-PD) 1 . This mosaicism results from the inactivation of one X chromosome in all somatic cells and should not exist in a monoclonal population. Following the discovery 2 in feral mice of an electrophoretic variant (A) of the X-coded enzyme phosphoglycerate kinase (PGK-1) which differs from the form (B) found in common laboratory mouse strains it was reported 3 that fibrosarcomas induced chemically in hybrids of feral and laboratory-bred mice expressed both enzyme phenotypes, but the conclusion that both were expressed by neoplastic cells was based solely on morphological evidence. The development of histocompatible substrains of mice homozygous for one or other alloenzyme has made it possible to study the clonal composition of tumours under experimental conditions in which the neoplastic status of subpopulations of cells can be verified by transplantation. The experiments we now report, while confirming that murine fibrosarcomas are often pleoclonal, show that the clonal composition may change markedly during tissue culture and on transplantation to con-genie hosts. These changes presumably reflect changes in the growth kinetics of differentiating subpopulations of the tumour. Cloned sublines are less readily transplantable than uncloned tumour cell populations, and some sublines are less readily transplantable than others; this suggests that sublines resistant to a host's attack are selected on transplantation or that some sublines require the cooperation of others to survive. We postulate that changes in clonal composition occur also during tumour development, metastasis and recurrence.