Do metastases arise from pre-existing subpopulations of cancer cells?

The hypothesis that metastases arise from pre-existing metastatic sub-populations of cancer cells with heritable metastasis-related characteristics, was tested by comparing the metastatic behaviour of cancer cells derived from pulmonary metastases with those from corresponding primary tumours, after implanting them subcutaneously in mice. In the case of KHT osteosarcomas and B16 melanomas, injected minces of metastases gave rise to more pulmonary metastases than cells derived from minces of the primary cancers generating them. However, in the case of 3LL and T241 cancers, the primary tumour minces gave rise to more pulmonary metastases than those derived from minced metastases. It is therefore concluded that the subpopulation hypothesis cannot be accepted as a general rule. When fragments of solid tumours were implanted into animals, no differences were detected between the metastatic behaviour of implants taken randomly from pulmonary metastases and the volume/age matched primary tumours generating them. These experiments thus provide no support for the hypothesis that metastases arise exclusively or predominantly from pre-existing metastatic subpopulations of cancer cells. Finally, implants of matched fragments from 3LL tumours of different volume and age, essentially produced no statistically significant differences in numbers of metastases. These observations do not therefore support the concept of a progressive evolution of subpopulations of cancer cells with heritable metastatic phenotypes during tumour growth.