Abstract
Determining the degree of diversity in therapeutic sensitivity exhibited by a tumour population is of considerable clinical importance. In addition to being a contributor to radiation resistance, diversity is the basis for variation in sensitivity over the course of treatment. To study intrapopulation diversity in radiosensitivity following γ-irradiation (2 Gy), distributions of the number of micronuclei/binucleate cell were obtained for human cervix carcinoma sandwich populations. Cell-to-cell diversity in radioresponse (micronucleus expression) was quantified using the overdispersion index ((variance/mean) − 1). As measured by this index, the radioresponse diversity of sandwich cultures sharply increased after introduction of oxygen/nutrients to the cultures, mimicking tumour reperfusion. In addition, a strong correlation was found between this measure of diversity and the extent to which the fraction of cells without micronuclei exceeds that expected from a Poisson distribution. This correlation indicates that for a diverse population there can be a significant departure of the aggregate population sensitivity (determined, for instance, by log-survival in a clonogenic assay) from that inferable from simply averaging per-cell sensitivities (reflected, e.g. by mean number of chromosome aberrations/cell). Our experimental results suggest a model attributing diversity in a population to its being a mixture of distinct subpopulations, each biologically homogeneous with respect to micronucleus expression, and each contributing an individual Poisson-distributed micronucleus response. We demonstrate how such radiodiversity may be quantified and show that reoxygenation of a microenvironmentally heterogeneous population leads to an increase in its radiobiological diversity.