Wednesday, 7 January 2015

Cancers and Tissue Cell Divisions: Yes…But…

A recent paper in Science* was widely covered in the national and international press. By gathering data on the total number of cell divisions in human tissues (x) and the incidence of cancer in those tissues (lifetime risk)(y), the authors emerged with a correlation coefficient (r) of 0.81 between the variables. In other words, the more renewing cell divisions, the greater the incidence of cancer. These findings then suggest that 65% (i.e. r2) ‘of the differences in cancer risk among different tissues can be explained by the total number of stem cell divisions in those tissues’.

from Tomasetti & Vogelstein. Science 347, 78-81
The authors then go on to suggest that this finding suggests a mainly stochastic effect of DNA replication as the cause of most human cancers. Thus the newspaper headlines: Cancer risk mainly bad luck, etc. To a great extent the findings confirmed my own prejudices, at least for the initiation of an error in replication. However, there is the major consideration of whether that initial error is eliminated and how rapidly and if at all the initial error progresses to a cancerous growth.

The authors also attempted to distinguish the effects of the chance effects of replication from other possible causative agents such as environmental factors and inherited mutations. Essentially their method seems a sophisticated version of looking at those cancers falling above the regression line and classifying them as extra risk. As might be expected the lung of smokers fell into this category.

The authors also suggested that their findings indicate that most cancers, the 65%, do not arise from environmental factors. And then I began to think a bit and then I began to look at their data. On the first point, I thought of the effect of cosmic radiation, and, indeed of a whole-body x-ray. Does not that affect all dividing cells equally? Whether one of your cells gets hit or not is down to chance, bad luck, but it is still an environmental effect that cannot be distinguished from a simple replicative error. With any non-tissue-specific environmental effect, the correlation coefficient could be the same, but the intercept of the regression line with the y-axis would be shifted upwards.

I then realised that data from some key organs of high cancer incidence were missing from the analysis: mammary gland and prostate—not missing because they had been missed out but missing because the necessary data do not exist. But as the following diagram from Cancer Research UK shows, breast and prostate cancer are so important in the overall incidence (the former accounting for 30% of all cancers in women and the latter for 25% in men) that it would be highly informative to see where they fit in the picture (including and excluding the BRCA mutations).

The 20 Most Common Cancers in 2011. Number of New Cases, UK
Interesting as the present data are they are in one species, the only one that counts for most cancer researchers. However, if the relation between number of cell divisions and lifetime incidence of cancer in a tissue is universal then the same correlation should emerge in other species. Can the necessary data be extracted from the current literature for mammals…or birds…or reptiles…?

Finally, my eyebrows started a slight twitch when I read in the paper: Most cells in tissues are partially or fully differentiated cells that are typically short-lived and unlikely to be able to initiate a tumor. Only the stem cells—those that can self-renew and are responsible for the development and maintenance of the tissue's architecture—have this capacity. Well, up to a point Lord Copper, and the general point is taken, but I can never get out of my head Peter Wooding’s electron micrographs showing fully differentiated, secreting mammary cells in the process of division.



*Tomasetti C,  Vogelstein, B. 2015. Variation in cancer risk among tissues can be explained by the number of stem cell divisions. Science 347, 78-81.