Editorials Christmas 2015

Cancer’s Darwinian dilemma: an evolutionary tale in three acts

BMJ 2015; 351 doi: https://doi.org/10.1136/bmj.h6581 (Published 15 December 2015) Cite this as: BMJ 2015;351:h6581
  1. Mel Greaves, director
  1. 1Centre for Evolution and Cancer, Institute of Cancer Research, Sutton, UK
  1. mel.greaves{at}icr.ac.uk

If enough cells are playing the game, then somewhere there will be a winner, no matter how harsh the challenge

“Nothing in biology makes sense except in the light of evolution”

Theodosius Dobzhansky, 1973

There are upwards of 10 million new cancer cases a year in the world. One in three of us can expect that unwelcome diagnosis, and around one in four will succumb to metastatic, drug resistant disease. The big questions are, why are humans so vulnerable to cancer? what exactly is cancer as a biological process? and why is drug resistance the norm for advanced disease? Evolutionary biology has something to say about each of these grand challenges.

Act 1. Why are we so vulnerable?

Species in almost all animal phyla, including within the invertebrates, can develop cancer.1 But there is one stand out group here, and it’s Homo sapiens. Our cancer rates are through the roof. It’s sometimes posited that this reflects the inevitable price of surviving into old age. Cancer risk is certainly age associated, and cancer in other mammals (such as horses) increases with age. But it’s not driven by the degenerative effect of ageing. This is evident from the highly variable incidences of the major cancers over time and in different geographical settings. So what is it? Is it industrial pollution, electricity pylons, your parents’ genes, divine retribution, or just “bad luck?” Or none of these.

Each cancer has a degree of inherited genetic susceptibility and an element of chance or bad luck. The luck arises because cancers evolve by Darwinian clonal selection (see below) contingent on random mutations in the genome. But the epidemiology of cancer in humans has long taught us that many, and probably most, common cancers are associated with lifestyle or linked to social conditions in particular societies.

My contention is that cancer risk is inherent to the design of multicellular animals but that our modern lifestyles greatly ratchet up the risk through a mismatch with our evolutionary genetics and previous environmental adaptations.2 3

One obvious example is skin cancer. Migrants from Africa to Europe some 50 000 years ago evolved depigmented skin from our black ancestors for good adaptive reasons, probably to secure sufficient vitamin D from the diminished solar ultraviolet B radiation in cloudy northern climes. That’s fine as long as we stay there. But we don’t.

Another potent example is breast cancer. Human females are unique among primates in having year round, or non-seasonal, regular oestrus. It was a good idea, increasing fertility. But if pregnancy is absent or delayed and breast feeding is not protracted (it is two to five years in hunter-gatherer societies), then each month between the ages of about 13-50 the breast (and ovary) receives a pulse of proliferative stress.

Much of our vulnerability might therefore be attributable to our rapid and biologically exotic social evolution. This has occurred at a speed far removed from the sluggish pace of evolutionary adaptability. And, moreover, the trade-off penalty of lethal cancer hits primarily after we’ve ceased reproduction and when evolutionary selection is largely indifferent to our fate.

A large fraction of cancers are, however, preventable, and we don’t need to adopt a Stone Age lifestyle to do so. Prudent avoidance, better balanced diets, and prophylactic vaccination would take care of much of the worldwide burden.

Act 2. Survival of the fittest cell

We’ve known for a long time that cancer is driven by acquired genetic changes but now have a vivid portrait of the complexities of the process.4 Cancers are initiated by mutations in single cells, primarily immature stem or progenitor cells in tissues, the progeny of which proliferate and acquire additional mutations that are advantageous to the cell but at the expense of the individual. It’s as if cancer cells are incipient parasites. This is now seen as a classic Darwinian process of genetic diversification coupled with phenotypic change in cells filtered by natural selection within the tissue ecosystems of our bodies. It’s survival of the fittest cell.5

When we get down to single cell genetics, it’s clear that the dynamic evolution of a cancer clone mirrors that of species evolution as imagined by Darwin (figure).6 Distinct subclones have a branching evolutionary relation with variegated genetics. Every patient has a unique cancer evolutionary tree with multiple subclones, linked by descent to a common ancestral founder. The clinical implications of this space-time diversity for biopsy based prognostication and therapy are considerable.7 8


Darwin’s iconic speciation tree diagram illustrates how the variants A, B, C, and D might evolve and diverge from their common ancestor

This evolutionary process can take anything from a few months (paediatric cancers) to decades (common epithelial cancers). The real problem is with migrants. As tumours expand, they solicit new blood supplies—angiogenesis—and become diffuse in the tissue of origin. In the face of increasing competition for space and resources, cells seek refuge elsewhere, by escaping into the blood circulation or lymphatics to establish residence as a diaspora in other tissues. Cancer then becomes a territorial hijack of essential tissue functions.

There are fragilities we can exploit. All cancers depend upon rewired, self-renewing stem cell activity.9 This provides a therapeutic target. Catch cancer early enough, before dissemination, and it’s considerably easier to eradicate. The evolutionary process of clonal evolution is contingent upon ecosystem selective pressures—hormonal drive, inflammation, and hypoxia, for example.10 These provide alternative targets for therapy that might slow down or halt clonal progression towards malignancy.

Act 3. Resistance wins out

Cells have had three or more billion years to develop survival skills through random shuffling of their genomes—the substrate for all evolutionary change. If enough cells are playing the game, then somewhere, by serendipity, there will be a winner no matter how harsh the challenge.

When cancers evolve to the late stage of metastasis they are robust and equipped with genetic, or epigenetic, tricks that can thwart most cancer drugs.11 Given mutation rates, often enhanced by genetic instability, and the number of cells spawned—1012 or more—it is inevitable that rare cells will exist that are intrinsically drug resistant. It’s similar in principle to the problem we face with drug resistance for malaria, tuberculosis, and HIV and with pesticides and herbicides12: the intrinsic liability of natural selection within diversifying, pathogenic species.

But this too is a tractable problem. Aside from prevention and early intervention, it should be possible rationally to design cocktails of drugs (or drugs plus immunotherapeutics) and scheduling that effectively degrade cancer cells’ survival skills.13 14 We are in an evolutionary arms race.


Cite this as: BMJ 2015;351:h6581


  • Competing interests: I have read and understood BMJ policy on declaration of interests and declare support from the Wellcome Trust.

  • Provenance and peer review: Commissioned; not externally peer reviewed.


View Abstract

Sign in

Log in through your institution

Free trial

Register for a free trial to thebmj.com to receive unlimited access to all content on thebmj.com for 14 days.
Sign up for a free trial