“I had always gone to so much trouble to live healthily – so now this hits me, I wasn’t really ready for it”. Doctors are not entirely unfamiliar with this statement whenever they bring the bad news to the patient that he or she is suffering from cancer. As recently as 20 or 30 years ago everything seemed to be quite straight forward. Whoever smoked a lot, fed oneself on unhealthy meals or stayed too long in the sun had to expect that it would “strike” him or her. There were a number of risk factors that increased the risk for a particular tumour. Sometimes an increased incidence in the family was also present, which was able to be attributed to inherited “susceptibility genes”.
Risk factor cell division
Bert Vogelstein from the renowned Howard Hughes Medical Institute had made a respected name for himself with his work in molecular oncology and seemed by way of one article in “Science” a year ago to suddenly put all these findings in question. He and his colleague Cristian Tomasetti claimed that the development of cancer depends in the vast majority of cases on chance and is nothing more than “tough luck”. The total number of stem cell divisions in a particular tissue over a lifetime correlates very strongly with the incidence of cancer in this organ. Each division of this tissue progenitor cells is able to lead to damage in the DNA. One damage event with the risk that out of a “normal” cell one without growth inhibition emerges.
The publication of these findings was a major setback for cancer prevention. Should things really come down to only one-third of all cancer cases being due by an avoidable risk, as the authors calculated? Numerous critical voices quickly made themselves heard. The IARC (International Agency for Research on Cancer) also responded quickly and with harsh criticism. Two of the most important tumour types, carcinoma of the breast and the prostate, had not been factored in by the authors. In addition, according to the press release of the IARC, “the comparison of different populations would have given different results”.
In actual fact, the cancer rates for single tumour forms in different regions of the world in some instances differ completely. Breast cancer occurs in Europe about five times more often than in East Asia or Central Africa. In Australia, the risk of developing prostate cancer is about 20 times as high as in Central Asia. Emigrants adapt to the externalities of their new home, something which also allows us to conclude that a strong influence of extrinsic factors is present, even if these risk factors are often not known yet.
Epidemiology against randomness
For many months nobody could hold out proof that Vogelstein and Tomasetti had made mistakes in their work. In fact, the different cancer rates in different tissues are not really able to be explained solely in terms of influence via environmental and familial risk. However, in December an investigation appeared in “Nature” [Paywall] which seemed to come up with facts challenging the pot-luck random development of events. Yusuf Hannun and his team at New York Stony Brook University attempted by looking at 30 types of cancer to quantify the influence of known extrinsic factors.
Even taking into account the total number of stem cell divisions, with individual types of tumours there are still significant differences in the frequency to be seen – more than can be explained by the Vogelstein-Tomasetti model. Accordingly, the share of responsibility carried by external influences must probably be far higher. Epidemiological data for individual types of tumours do not correlate with the “Science” article’s calculations, which attaches much more weight to cell division-derived accidental error than to other influences.
Reading from signatures
Furthermore, certain changes can be observed in many tumour genomes, so called “tumour signatures”, which are typical for the respective tumour. Tumours due to extensive sun exposure look quite different from those which are associated with greater risks from tobacco smoke. The authors looked at these genetic fingerprints more closely and found only two with a clear age correlation. The rest appear to manifest independent of age. This in turn confirms the hypothesis that external factors – independent of the “bad luck” principle – play a much stronger role. If one ultimately also precisely observes the error rate in cell divisions, one ends up calculating far lower rates for the possibility of transformation into a malignant tumour cell than those figures put forward by Vogelstein/Tomasetti in their thesis as the origin of most tumours being due to chance. The authors even ventured with their analyses to go so far as to make the statement: in 70 to well over 90 percent of cases external factors and unfavourable genetic constellations are what lead to the “cancer” dignosis.
It’s questionable that 95 percent of brain tumours, more than 99 percent of prostate and at least 98 percent of thyroid carcinomas – figures given in Hannun’s account – are due to extrinsic factors, Vogelstein and Tomasetti state in their reply to their challengers. Until now, epidemiologist have not been able to define the real external danger source attached to any of these types of tumours.
Are there external factors present of which we not know anything yet? The renowned British oncologist Mel Greaves assumes that “90 percent of the common cancers could be avoided or prevented”, judging by the documented differences in incidence rates, it says in a statement from the Institute of Cancer Research.
Birth of a tumour cell
But how does the birth and early growth of a tumour cell look? An interesting report [Paywall] in the New England Journal of Medicine by pathologists from San Francisco appeared in November, to which the aforementioned Berti Vogelstein responded in a detailed commentary [Paywall] in the same issue. The team led by Boris Bastian and Hunter Shain analysed changes to nearly 300 genes in the early stages of a melanoma cell right up to to the invasive advancing tumour stage. Which of the mutations are critical in securing the growth advantage in competition with its environment and in doing so not be brought back to the start by the body’s own guards?
Driver mutations occur as characteristic mutations in tumour cells and especially in genes for mitogen activated kinases. They appear in the actual precursor melanoma cells as precursor lesions. In advanced stages, the DNA sequences in NRAS and TERT genes had then altered. Invasive tumour cells are characterised by biallelic CDKN2A aberrations. TP53 mutations are in turn a typical feature of advanced melanoma.
Three strikes and cancer is dished out
Similar findings exist with cervical, pancreatic and colon cancer, albeit each with different driver mutations – suggesting that three changes in such “drivers” convert the normal tissue cell into an aggressive predator that constantly multiplies. Only about 200 of our 20,000 genes are susceptible to such changes. All the others are so-called “passengers”; passengers who ride along, but do not affect the choice of route. It appears that, depending on the tissue type, various changes are the key to malignancy.
Three specific changes in a given order to get to the specific “Achilles heel” of the cell in its lifetime? Somewhat unlikely. Rather, it seems that the first two strikes lead to a slight growth advantage and altered cell architecture, which is also typical of benign tumours. Especially at this early stage it seems very much to come down to whether the strike lands on a predetermined “sensitive” point. The third strike often brings the wounded cell completely out of balance some two or three generations later, and is probably far less specific. It’s this event that initiates the invasive life course of the tumour cell.
Tracking down shot cells
But this also means there exist good opportunities for future cancer diagnosis and prevention. In a full-blown tumour early “breakthrough” and expansion phases are difficult to find because they are overgrown by cells in the later stage. If one can discover the cells in the early phase, this would possibly offer a good chance to take them out of the system before the third strike hits.
Whether all this happens as a series of random events is disputable, even after the Nature publication in December appeared. External influences surely affect the error rate during replication of stem cells. Should the repair system not manage to quickly discover the alterations in the genome and repair these, this may mean the first step has been taken towards having a tumour. It pays to keep the strike rate as low as possible.