Tuesday, January 12, 2016

Cancer--luck or environment? Part II: Nothing to food-fight over

Yesterday we commented on the 'controversy' over whether cancer is mainly due to environmentally (lifestyle) or inherently (randomly) arising mutations.  This is a tempest in a teaspoon.

Mutations, whatever their individual cause, must accumulate among dividing cells until one cell has the bad luck to accumulate a set of changes that 'transforms' it into a misbehaving cancer cell.  The set of changes varies even among tumors of the same organ, because many different genes and their expression-regulation contribute to the growth, or restraint of growth, even within the same tissue. That is, not all breast, colon, or lung cancers are caused by the same set of mutations.   It then proliferates, rapidly dividing and thus indubitably acquiring more mutational changes that enable it to do things like metastasize to other parts of the body, or develop resistance to drug treatment.  The more rapidly it grows and spreads, the more rapidly such things can happen.

Even if the first transformational cause were due entirely to environmentally-induced mutations, the real dangers that ensue during the tumor's lifespan are relatively rapid additions to the original tumorigenesis process, and so in a sense the main dangers of cancer are primarily, if not nearly exclusively, due to inherent mutation among cancer cells.  If you get lung cancer and then stop smoking, your lung cancer will still evolve. Indeed, if environment contributes, it may make things worse--if that "environment" is radiation or chemotherapy: radiation definitely causes mutations, and chemotherapy weeds out cells that haven't experienced resistance mutations, leaving or even making room for tumor lineage cells that do have resistance mutations.  Finally, things that stimulate cell division can facilitate new mutations or even just make a tumor spread more rapidly.

So clearly cancer is not all due to environmental, nor to inherently occurring changes.  These and other factors comprise multiple, interacting causative effects.  Attributing cause to environment or inherency is misleading.  But what if cancer were in fact even entirely due to lifestyle factors that stimulate cell division or directly cause mutation?  Of course this would be very good for the Big Data epidemiologists and their studies, and threatening to industries and so on that produce mutagenic waste or products etc.  But suppose epidemiologists were to continue to find major carcinogenic environmental factors (that is, that the major ones, like smoking, aren't already known). Let us further suppose that avoidance behavior were to follow the announcement of the risks (not an obvious thing to assume, actually; the tobacco industry is still thriving, after all).  Then what?

Epidemiologists would say their work has prevented cancer and would claim victory over the to-them strange idea that cancer is due to inherent mistakes in DNA replication and is inevitable if one were to live long enough.  A lifestyle-change-based reduction in cancer would be clearly a very good thing.  But in fact, it would not be an unalloyed victory: one thing it would do is keep the non-exposed person alive (because s/he didn't get cancer!) and that in turn means that s/he would be at higher risk of (1) other age-related deteriorative diseases that dying of cancer would have precluded, many of which are waiting in the wings at ages when cancers arise, and (2) eventually getting cancer at some older age.  In the first case, the rates of other diseases like stroke and diabetes etc. would necessarily go up.  The risk of slowly petering out in increasingly bad shape in an intensive nursing unit would go up.  That would, of course, lower the lifetime cancer risk, but not in a very pleasant way.

In other words, lifestyle changes can delay cancer, but even assuming that the per-year exposure to environmental mutagens were reduced, the consequently longer exposure to those mutagens might mean their lifetime total would go up), so whether or not it decreased the lifetime risk of cancer would be an open question.  However, what this would do would be, by removing environmental causes, to raise the fraction of cancers that are due to inherent mutation, strengthening the fraction of Vogelstein-Tomasetti cases!

It's undoubtedly good to get cancer later rather than earlier in life, but not an unalloyed good.  In any case, what these points show is that the argument over the particular fraction of cancers that are due to environment vs inherent mutation is rather needless.  At most it might be relevant to ask how much of funding investment in big epidemiological studies is going to pay off, rather than spending on some other clearer issues (especially if the major environmental mutagens are already known).  There have already been scads of massive long-term studies of almost anything you can name, to identify carcinogenic exposures. With some very important exceptions, that are by now well known, these studies have largely come up empty, or with now-it-is/now-it-isn't conclusions, in the sense that risk factors are either weak, or if strong are rare and hard to find embedded in the broad mix of chronic disease risk factors.  Environments are always changing with new possible carcinogenic exposures arising, but basically those with strong effects usually show up on their own such as by multiple cases of a particular cancer type in some specific location or among workers in a particular industry or in vitro mutagenesis studies and the like.

If causation is too generic, don't get your hopes up
If comparisons among countries, for example, show that the same cancer can have very different age patterns or incidence rates, this may suggest lifestyles as major risk differences. But that's far from saying that the causal elements are individually strong or simple enough to be enumerated by the usual Big Study epidemiological approach. One can be extremely doubtful that this would be the case.

Saying something is 'environmental' because, for example, it varies among populations is like saying something is 'genetic' because it varies among relatives.  If it's like genetic factors as documented by countless GWAS studies, there are many different, correlated or even independent contributors, then each person's cancer will be due to a different set or complex set of experiences and the luck of the mutational draw.  As with GWAS and related approaches, it is far from clear that large, long-term environmental studies, more mega than we've already had for decades, will be the appropriate way to approach the problem.

Indeed, to a considerable extent, if each case is causally unique, by some different combination of factors and their respective strengths in that individual, then it's epistemologically not very different from saying that cancer occurs randomly, which, though for a different sort of reason, is what V and T said.  There won't, for example, be a specific environmental change you can make, any more than a specific gene you can re-engineer, to make the disease go away or even to change much in frequency or age of onset.

Food fights like this one are normal in science and often have to do with egos, investment in one 'paradigm' or another, how research is supported or advice from experts are conveyed to the public.  But such disputes, though very human, are rather off the point.  We often basically ignore risks we know, as in the proliferation of CT and other radiation-based scanning and medical testing which can be carcinogenic.  Life is mutagenic, one way or another.  So while you have life, enjoy your food--don't waste it by throwing it at each other!  There are better questions to argue about.

2 comments:

Allan Hotti said...

Are you aware of any information that can incriminate accumulated mutations (non-cancerous) as at least a partial contributor to "aging"?

Ken Weiss said...

This is a legitimate and difficult question.

There are various theories about this. Mitochondrial mutations have been shown to accumulate and some have suggested this plays a role in mental deterioration with age. Age spots have, I think, been show to be related to somatic mutation. Telomere loss, a common 'candidate' for the essential aging process is another. Response to infection would have to be another, of a separate kind, including perhaps autoimmune disease. Greying of hair may be another. Epigenetic changes may be involved as well.

It is also the case, perhaps rather instructively, that in theory it simply must be true: somatic mutations occur, not all are detected, and this has to contribute to the age-related degradation of our systems. One can debate whether this is related to the correlation of body size with lifespans because of the number of at-risk cells involved etc. The key issue is one I once termed 'phenotype amplification', that mutations in individual cells have to have proliferative effects in some cumulative way, to appear as organismal disease or other traits. Cancer is of course the classic case. A few years ago, in Trends in Genetics, I have written that I think epilepsy may be one case where such a mechanism could be suspected. But that was speculation, and in fact, I argued that more attention should be paid to the question.

I expect there are others, probably, many in both plants and animals including humans--but I can't offhand cite them, and haven't tried to scope them out. I worked in this area for many years--but not recently. The technological challenges have been great but suitable cell-specific methods are becoming available. I think much would easily be turned up (if it exists) by PubMed searching. So I certainly must acknowledge that I develop my view largely as one in principle, based on the nature of cells and tissues, and any such view must be open to question.