In a study that's bound to attract considerable controversy, a pair of researchers are claiming that between 60 and 66 per cent of all cancer-causing mutations are the result of random DNA copying errors, making them essentially unavoidable. The new research is offering important insights into how cancer emerges, and how it should be diagnosed and treated -- but many questions remain.
For decades, the paradigm among cancer researchers has been that cancer arises as the result of heredity or environmental factors, or a combination of the two. New research from Johns Hopkins Kimmel Cancer Center suggests there's a third factor that causes mutated cells to start reproducing out of control -- random and unpredictable copying errors during DNA replication. Upsettingly, the research shows that these errors occur in people who don't have a family history of cancer, and among those who lead very healthy lives. This revelation suggests improved methods to detect cancer are needed, especially in the early stages when cancers are highly curable.
"The reason why [cancer researchers] have never seen this before is because no one has ever measured this before," said study co-author Cristian Tomasetti at a news conference held yesterday. "This is what we are presenting to you today."
As the authors conclude in the new paper -- published today in Science -- random mutations appear to be responsible for nearly two-thirds of all cancers. The researchers informally refer to these cancers as "bad luck" cancers because there's evidently nothing we can do to avoid them. But as the scientists go to great pains to point out, that's not an excuse to live one's life with reckless abandon. As this new paper and pre-existing studies show, approximately 40 per cent of cancers can still be prevented.
"It is well-known that we must avoid environmental factors such as smoking to decrease our risk of getting cancer. But it is not as well-known that each time a normal cell divides and copies its DNA to produce two new cells, it makes multiple mistakes," explained Tomasetti. "These copying mistakes are a potent source of cancer mutations that historically have been scientifically undervalued, and this new work provides the first estimate of the fraction of mutations caused by these mistakes."
Tomasetti, along with his colleague Bert Vogelstein, reached this conclusion by using a new mathematical model to evaluate DNA-sequencing data and epidemiological findings from around the world. Among the 32 cancer types studied, the researchers found that approximately 66 per cent of cancer mutations are the result of DNA copying errors, 29 per cent are caused by environmental factors, and just five per cent are hereditary.
A study published in 2015 by Tomasetti and Vogelstein reached a similar conclusion, but it kindled a firestorm of controversy by excluding certain cancers, and by only considering US cancer patients. Given the role environmental pollutants and other toxic agents play in cancer, it was not unreasonable to suggest that the causes of cancer may vary globally. What's more, the researchers failed to determine the fraction of the mutations in cancers caused by DNA copying errors; the researchers only compared cancer incidence to the total number of cell divisions in the organs in which the cancers occurred.
"The two studies are different in many ways, and includes things we hadn't done before," Tomasetti told Gizmodo.
This time, the researchers included breast and prostate cancer, which they had previously excluded before on account of insufficient data, and they collected cancer data from around the world.
"The fact that we were able to show a correlation [between stem cell division and tissue specific cancer risk]... on par with the United States is an important step forward, but it's not really the most important point [of the updated study]", said Tomasetti. "The big difference is that in the 2015 paper we found a correlation between stem cells mutations and cancer rates, but as we know correlation is not causation. This new paper is the first time that anyone has looked at proportions of mutations in any cancer type and assigned the three causal factors. I consider this a completely new result compared to our previous one."
Tomasetti says his team's data sits in perfect accord with epidemiological estimates showing that 42 per cent of cancers can be prevented -- a reminder that "everyone should adhere to prevention guidelines".
For the study, the researchers compared the number of stem cell divisions -- an important source of cell mutations -- among cancer patients from 68 countries listed in 423 registries monitored by the International Agency for Research on Cancer. Tomasetti and Vogelstein found strong correlations between cancer incidence and cell divisions not influenced by heritage or the environment among the cancer types studied, regardless of the country's environment or economic output. Looking at the mutations that drive abnormal cell growth, the researchers developed a statistical model using DNA sequencing data from the Cancer Genome Atlas and epidemiologic data from the Cancer Research UK database.
An interesting discovery is that it takes approximately two gene mutations for cell division to get out of control and become cancerous -- but not all mutations results in cancer.
"Every time a normal cell divides, it makes several mistakes called mutations," explained Vogelstein at the press conference. Because they're random, these mutations aren't usually harmful, affecting genes unrelated to cancers, he said. "But every once in a while a mutation will occur in a cancer driver gene, and that's bad luck."
While across the board, these "bad luck" cancers account for about two-thirds of all cancers, each cancer type has its own story to tell. In pancreatic cancer, for example, 77 per cent of cases are the result of random DNA copying errors, 18 per cent are due to environment, and the remaining five per cent to hereditary. For prostate, brain and bone cancers, the "random" (R) factor jumps to 95 per cent (the researchers defined R as mutations caused by normal DNA replication, and not those influenced by inherited genetic factors and/or the environment). Most of the mutations in breast cancer were attributed to R, but some were due to health factors (obesity being a contributing factor). Lung cancer is quite different, with the environment accounting for 65 per cent of cases (attention smokers). Around 35 per cent of lung cancer is the result of copying errors, while heredity plays a virtually non-existent role.
The reason for the causal discrepancy among the 32 cancers studied, say the researchers, is that not all tissues are created equally. Certain cells, for reasons that are still to be determined, are simply more susceptible to mutation. It could be that some cells replicate with greater frequency, yielding more opportunities for a mutation. "The more stem cells, the more mistakes," says Vogelstein. "That's why colorectal cancers are more frequent than, say, brain cancers. It's not that the nature of the mutations are different."
Vogelstein hopes that this work will inspire researchers to better understand the risk factors involved, and to devise novel prevention measures. He also says that secondary prevention of cancer -- that is, treating cancer as soon as it appears -- is a grossly undervalued and under-researched component of cancer research, and that this study should serve as a wake-up call in that regard.
Like their 2015 study, this research is guaranteed to attract attention -- and a hefty amount of scrutiny.
"It is an interesting paper," said Song Wu, an Associate Professor at Stony Brook University's Department of Applied Mathematics and Statistics, in an interview with Gizmodo. "However, just like their first one published in 2015, I feel their conclusions are somewhat too bold. They are now backtracking from claiming that two-thirds of all cancers are due to intrinsic processes to claiming that two-thirds of mutations are due to unavoidable intrinsic processes. They may backtrack their conclusion again two years from today."
Wu is also concerned that the paper relies too much on one epidemiological study from Cancer Research UK, while ignoring a large body of epidemiological studies. "For example, early studies on immigrants showed that prostate [cancer] clearly have significant environmental component, but Tomasetti and Vogelstein simply claim prostate cancer has almost zero environmental effect."
Wu's colleague and collaborator, Yusuf Hannun from the Stony Brook Cancer Center, also has misgivings about the updated research. "The flaw with their logic from the previous study persists: The correlation between stem cell division and tissue specific cancer risk," he told Gizmodo. Hannun points out the correlation doesn't imply a causative role for stem cells in cancer, and that even if a causative role was shown, it hardly means that stem cells are the only factor involved. "Generalising this to geographically disperse data does not undo this shortcoming," he added.
Last year, Wu and Hannun published a study in Nature that came to a vastly different conclusion. The researchers showed that cancer risk is heavily influenced by environmental factors, accounting for upwards of 70 to 90 per cent of all cancers.
The Tomasetti and Vogelstein study may be flawed, and it's undoubtedly incomplete, but the idea that certain cancers are beyond our control is certainly an intriguing one. It makes sense that our bodies, outside of heredity and the environment, will on an unfortunate occasion produce a nasty cancerous cell that churns out more copies of its twisted self. Moving forward, scientists have the daunting tasks of figuring out how to predict these instances of cancer -- if that's even possible.
In the meantime, you should still do your best to avoid cancer: Eat healthy foods, maintain a healthy weight, quit smoking, and avoid carcinogenic environments.