Epigenetic Drift and Cancer

Epigenetics and Aging (16)

In a provocative paper published in 2015, Cristian Tomasetti and Bert Vogelstein argued that only about 35% of cancers can be attributed to inherited genes and environmental carcinogens combined. The vast majority of cancers, they argued, were the result of bad luck (10.1126/science.1260825). They based their argument on the large variation in carcinogenesis among cell types. There are as many types of cancer as there are cell types (over 200). So Tomasetti and Vogelstein had a large data set to work with. The found that the incidence of any particular cancer was highly correlated with the number of cell divisions in the stem cell population that replenished the differentiated cells. Why would the number of stem cell divisions correlate with cancer in the corresponding cells? Because, they argued, with each replication event there is a small but nonzero risk of a somatic mutation. As such the more replications in a stem cell population the higher the risk of a somatic mutation that will initiate carcinogenesis.

This paper created quite a stir in the community of cancer researchers, and it was attacked from multiple directions. Some argued that it underestimated environmental factors, others that it underestimated inherited genetic factors. Some questioned the overarching framework of carcinogenesis—the somatic mutation theory (SMT) that the authors assumed (10.1016/j.pbiomolbio.2016.07.004). SMT has been the dominant paradigm in cancer research since then President Richard Nixon declared the “War on Cancer” over 50 years ago. The manifest lack of progress in this battle has reduced, somewhat, the ranks of SMT supporters but it remains the dominant paradigm, and the official view promulgated by the National Institute of Health.

As the name implies, the somatic mutation theory is genocentric. Aside from its appeal to geneticists, SMT is attractive for the oft emphasized parallels with biological evolution: a mutation/selection dynamic. According to SMT carcinization begins with a somatic mutation that disinhibits, if only to a slight degree, the proliferative tendencies inherent in all cells (https://doi.org/10.1073/pnas.1501713112). As such, the mutated cell will have a selective advantage over non-mutants in its tissue environment and produce more mutant cells as a consequence. Within the mutant cell population, further somatic mutations will arise, some conferring an even greater selective advantage than the original mutation. And so on, until only the most virulently multi-mutated cell populations are left to metastasize.

From the outset, some detractors noted that somatic mutations occur in all cells, not just cancer cells, and that the vast majority have no effect at all (DOI: 10.1126/science.aab4082). This led to a search for somatic mutations that were drivers of cancer development as opposed to mere passengers. The mutated drivers were labeled oncogenes and tumor suppressor genes.

SMT was formulated and entrenched before epigenetic research exploded. So it’s not surprising that one family of major critiques came from an epigenetic direction. In the initial assault it was emphasized that somatic mutations could be as much effects as causes of cancer, for example the effects of epigenetic genomic dysregulation. Moreover, for most cancers the heterogeneity of cancer cells is more epigenetic than genetic throughout the multiple stages. Additionally, a characteristic feature of cancer cells is dedifferentiation, a return to a more stem-like state, which would explain their proliferative tendencies. Dedifferentiation, like cellular differentiation during development, is an epigenetic process. Somatic mutations may increase its likelihood but would not seem up to the role of sufficient cause. There is a growing trend to invoke some combination of genetic and epigenetic alterations in explaining the onset and development of cancer.

What about that correlation between stem cell divisions and particular cancer types? Well, it’s not only the probability of somatic mutations that increase with increased replication events, epimutations would also increase and at a much higher rate, as epigenetic fidelity during mitosis is lower than genetic fidelity.

It is also worth noting that spontaneous remission, though rare, is well documented even when carcinogenesis is well advanced. This is impossible to explain from the SMT perspective, as it would require successive reverse mutations, which would be borderline miraculous. The improbability is recognized by the Catholic Church which has awarded posthumous sainthood to candidates, prayers to which, such remissions are attributed. (For details, see my Epigenetics: How Environment Shapes Our Genes:  https://a.co/d/cN5Uro6  ).

Epigenetic Drift

The biggest risk factor for cancer is age. Like, diabetes, cardiovascular diseases, and neurodegenerative diseases, cancers are diseases of aging. There are of course childhood and adolescent cancers—cause by inherited (not somatic) mutations—but they constitute a miniscule proportion of cancers diagnosed each year. Cancer rates remain low through early adulthood (25 per 100,000 people for ages 25-29), but the rate steadily increases through middle age (350 per 100,000 for ages 45-49), then rapidly accelerates after 50 years of age and even more so after age 60 (1,200 per 100,000 for ages 60-64; 1,600 per 100,000 for ages 65-69; nearly 2,000 per 100K for ages 70-74). On average cancer risk increases about 5% per year after 60 years of age, before leveling off somewhat after 80 years of age

So cancer should be considered in the general context of aging. Of all the theories of aging considered previously, somatic mutations do not figure prominently in any, except, perhaps, the stem cell theory but it is not foregrounded even there. Despite the flaws in their theoretical framework, Tomasetti and Vogelstein were right to forcefully argue that chance had been given insufficient due as a factor in cancer etiology. Chance is an important factor in all aging processes. It is stochasticity, not anything programmatic, or even deterministic, that undergirds the effectiveness of epigenetic clocks as crude measures of aging, as averaged over multiple tissues and multiple individuals. But aging is tissue-specific to a significant degree. And individual biological aging varies much more widely than the baseline variation expected by chance alone, as measured by chronological epigenetic clocks. The tissue-specific nature of aging is being addressed with the development of tissue-specific epigenetic clocks. And environmental effects on biological aging are beginning to be addressed in next generation epigenetic clocks. One even measures the effects of smoking in packs per day ( 10.18632/oncotarget.9795).

But epigenetic drift is potentially a more effective estimator of biological age, and hence cancer risk, than even the most sophisticated epigenetic clocks. Epigenetic drift is inherently tissue specific. Epigenetic drift, properly conceived, is also an indicator of biological, not just chronological age. in addition to the stochastic age-associated background changes measured by standard epigenetic clocks, epigenetic drift incorporates deviations caused by environmental factors and heritable genomic factors.

Most importantly, epigenetic drift incorporates more epigenetics than epigenetic clocks. Epigenetic clocks measure only one dimension of the epigenetic landscape, DNA methylation. Epigenetic drift, properly conceived, should incorporate all epigenetic dimensions of the epigenetic landscape, including histone modifications, loss of heterochromatin and alterations in non-coding RNA levels. In the next post of this series, I will consider cancer in the context of this multi-dimensional epigenetic landscape.