A quantitative study on the growth variability of tumour cell clones in vitro

Objectives: In this study, we quantify the growth variability of tumour cell clones from a human leukemia cell line. Materials and methods: We have used microplate spectrophotometry to measure the growth kinetics of hundreds of individual cell clones from the Molt3 cell line. The growth rate of each clonal population has been estimated by fitting experimental data with the logistic equation. Results: The growth rates were observed to vary among different clones. Up to six clones with a growth rate above or below the mean growth rate of the parent population were further cloned and the growth rates of their offsprings were measured. The distribution of the growth rates of the subclones did not significantly differ from that of the parent population thus suggesting that growth variability has an epigenetic origin. To explain the observed distributions of clonal growth rates we have developed a probabilistic model assuming that the fluctuations in the number of mitochondria through successive cell cycles are the leading cause of growth variability. For fitting purposes, we have estimated experimentally by flow cytometry the maximum average number of mitochondria in Molt3 cells. The model fits nicely the observed distributions of growth rates, however, cells in which the mitochondria were rendered non functional (rho-0 cells) showed only a 30% reduction in the clonal growth variability with respect to normal cells. Conclusions: A tumor cell population is a dynamic ensemble of clones with highly variable growth rate. At least part of this variability is due to fluctuations in the number of mitochondria.

💡 Research Summary

The study set out to quantify how much individual tumor‑cell clones differ in their proliferative capacity, using the human T‑cell leukemia line Molt3 as a model system. Hundreds of single‑cell‑derived clones were isolated in 96‑well plates and their growth was monitored by periodic microplate spectrophotometry, which records optical density as a proxy for cell number. For each clone the authors fitted the time‑course data to the logistic growth equation, extracting a maximum specific growth rate (r) and a carrying capacity (K). The distribution of r values was broad: some clones grew markedly faster, others considerably slower than the mean rate of the parental population.

To test whether this variability was genetically fixed or epigenetically driven, the authors selected six extreme clones (three above, three below the mean) and performed a second round of cloning. The growth‑rate distributions of the second‑generation sub‑clones were statistically indistinguishable from the original population, indicating that the high‑ or low‑growth phenotype does not persist across generations. This result strongly supports an epigenetic or stochastic origin of the observed heterogeneity.

The authors then proposed a mechanistic hypothesis: fluctuations in mitochondrial number from one cell division to the next are the primary source of growth‑rate variability. Mitochondria are essential for ATP production and metabolic regulation, and during cytokinesis they are partitioned randomly between daughter cells. Using flow cytometry, the team measured the average maximal mitochondrial content of Molt3 cells, which served as a key parameter in a probabilistic model. The model assumes that mitochondrial number follows a normal distribution at each division and that growth rate scales linearly with mitochondrial content. When the model was fitted to the empirical r‑distribution, it reproduced the observed spread with a high coefficient of determination, suggesting that stochastic mitochondrial partitioning can indeed generate the measured heterogeneity.

To validate the hypothesis experimentally, the authors generated rho‑0 Molt3 cells, which lack functional mitochondrial DNA and therefore have severely compromised oxidative phosphorylation. If mitochondrial number were the sole driver of variability, rho‑0 clones should display a dramatic reduction in growth‑rate spread. Instead, the variability was only reduced by about 30 %, indicating that additional sources—such as fluctuations in cell‑cycle regulators, transcription‑translation efficiency, or other metabolic pathways—contribute to the phenotype. Moreover, technical limitations (incomplete mitochondrial inactivation, variability in fluorescent staining efficiency, and the assumption of a linear r‑mitochondria relationship) may also affect the quantitative conclusions.

In summary, the paper demonstrates that a tumor cell population is a dynamic ensemble of clones with highly variable proliferation rates. A substantial fraction of this variability can be explained by stochastic differences in mitochondrial content inherited during cell division, but mitochondria are not the exclusive factor. The findings have important implications for cancer biology: clonal growth heterogeneity can influence tumor evolution, therapeutic resistance, and disease progression. Future work should integrate mitochondrial dynamics with other epigenetic, metabolic, and genomic variables to build a more comprehensive model of clonal fitness, which could ultimately inform personalized treatment strategies that account for intra‑tumoral growth diversity.