One more time: no, epigenetics is not Lamarckism

The standard model, so to speak, in evolutionary theory is known as the Modern Synthesis (MS), which was shaped during the 1920s, ‘30s and ‘40s by the founders of modern population genetics (Ronald Fisher, J.B.S. Haldane, and Sewall Wright) and by some other legendary names of modern biology (the geneticist Theodosius Dobzhansky, the naturalist Ernst Mayr, the paleontologist George Gaylord Simpson, the botanist Ledyard Stebbins, and the eclectic Julian Huxley). As is well known, the MS was the result of the unification between the original theory of evolution by natural selection, proposed by Charles Darwin and Alfred Russel Wallace in 1858, and the then emerging discipline of genetics, resulting from the re-discovery of the work of Gregor Mendel.

The MS marked the death knell of Lamarckism, an obscure and confused set of ideas arching all the way back to the 18th century, and proposed by Jean-Baptiste de Lamarck. Lamarck, obviously writing in pre-Darwinian as well as pre-Mendelian terms, proposed two fundamental concepts: (i) that living organisms are somehow capable of positively (we would say, adaptively) react to changes in the external environment, through some sort of internal vital impulse; and (ii) that the new traits resulting from the action of this force could be passed from one generation to the other. At the least, that’s the most plausible interpretation of the rather obscure and unclear stuff Lamarck actually wrote. (Note that almost nobody writing today — myself included — has actually read Lamarck, in the original 18th century French. I have read parts of his work in English, and spoken to a couple of people who are actually familiar with the original.)

But Lamarckism has an uncanny ability to be the zombie of evolutionary biology: no matter how many times one thinks he has killed it, it keeps coming back and wreak havoc. This has happened again in recent years because of the emerging field of epigenetic studies, particularly epigenetic inheritance. Epi-genetic literally means “above” or “beyond” the gene, and it refers to non-genetic phenomena that take place inside living cells and that alter organismal development. Epigenetic effects have been known for decades, are well studied, for instance, in the field of cancer biology, and are entirely uncontroversial.

What is controversial, however, is the subset of epigenetic phenomena that appear to be heritable from one generation to the next. Initial reports of epigenetic inheritance, decades ago, where greeted with skepticism by the scientific community, precisely because they “smelled” rather too suspiciously of Lamarckism.

By now, overwhelming scientific evidence of the reality of epigenetic inheritance is available, and the concept is one of the pillars of the Extended Evolutionary Synthesis (EES), the updated version of the MS that is taking shape and to which I made modest contributions — including by way of experimental studies of epigenetic inheritance.

Epigenetic inheritance, however, has nothing to do with Lamarckism, as the similarity between the two ideas is only superficial, and can be deeply misleading. Yet, plenty of professional biologists keep falling for it, sometimes writing very confused and entirely unhelpful articles for the general public. That’s what happened recently in the usually excellent Aeon magazine, which published “On epigenetics: we need both Darwin’s and Lamarck’s theories,” by Washington State University’s Michael Skinner.

Do we really need both Darwin and Lamarck? Hello no, we don’t, and it’s time to stop repeating such ill-informed nonsense. To see why, let me tell you a bit more about epigenetic effects.

Let’s take the most commonly studied class of epigenetic phenomena: DNA methylation. In many organisms (though it is far from clear just how frequent this phenomenon is) DNA strands inside cells are “marked” by a number of simple chemical molecules, known as methyl groups, which act as signals for certain proteins to ignore a particular sequence of DNA, which thus is not transcribed into RNA. As a result, that gene does not produce a protein, and hence has no phenotypic effect, until the methyl group is removed and the cellular machinery receives the okay to proceed with transcription.

In other words, methyl groups, which are themselves placed and erased by proteins (in turn, obviously, encoded by genes), act as regulators or modulators of gene expression. The specific pattern of methylation varies from cell lineage to cell lineage during development, and it is a major mechanism through which living organisms generate different cell types (say, muscles vs bones) in different places and at different times. When methylation goes astray, for instance, a given cell may lose the brakes that preclude uncontrolled cell division, and a cancer is born.

Moreover — and here is the real kicker as far as evolution is concerned — in some (but by all means not all) organisms, epigenetic mechanisms have two additional properties: (i) they can be modified by environmental effects, and (ii) they are heritable. An example of (i) is the alteration in methylation patterns (and hence, remember, in gene expression) when the organism is exposed to an external stress, such as high temperatures, or starvation. The extent to which (ii) occurs is still debated, with many confirmed examples from plants and insects, and a few tantalizing ones from mammals. (But we also have plenty of instances of organisms where the methylation pattern is automatically reset every generation, which means that no alteration induced by the environment is actually transmitted to the next generation.)

Superficially, (i) and (ii) sound like Lamarckism: we have an internally generated response of the organism to an external change, as well as a mechanism of inheritance of the induced traits. But the analogy is, in fact, only superficial. First, because environmentally-triggered epigenetic changes are not inherently adaptive, that is, they are not necessarily good for the organism. Rather, they are just like standard mutations of DNA sequences (which can also be induced by external factors, such as temperature, or radiation). Indeed, biologists refer to changes in epigenetic factors as “epi-mutations,” in direct analogy with genetic mutations. Remember, by contrast, that Lamarck thought that living organisms are capable of responding positively to alterations in the external environment. The standard example is the elongation of the neck in giraffes, allegedly to counter the fact that the leaves they feed on got higher and higher up the trees. (This may not be the actual reason why giraffes have long necks, by the way.)

Second, what is inherited is not the phenotypic change — as Lamarck thought — but rather the methylation pattern (or whatever other epigenetic mechanism is at play, there are several). This is, again, directly analogous to genetic mutations, where what is inherited is the information encoded in the genes, not the phenotypic outcomes that result from the action of such genes.

In other words, epigenetic inheritance should be properly treated as an additional layer (or, more precisely, set of layers) of information that is sensitive to environmental changes and passed from one generation to another. Additional, that is, to the standard channel of genetic inheritance, but not qualitatively different from it. Eva Jablonka — who unfortunately herself has often flirted with the L-word — correctly labelled this phenomenon in her book with Marion Lamb, Evolution in Four Dimensions. Evolution is then made possible by four channels of inheritance, acting partially independently and partially in reciprocal feedback loops: genetic, epigenetic, behavioral and symbolic (i.e., language).

In agreement with what I just explained, many of Skinner’s own examples in the Aeon article are of maladaptive epigenetic changes. For instance, he writes:

“Findings published by my group in Science in 2005 showed the ability of environmental chemicals to promote inheritance of disease in rats through three generations, to great-grand offspring and beyond, in the absence of any continued exposures. … An example is when Graham Burdge and his team at the University of Southampton in the United Kingdom reported that excessive nutrition in rats created epigenetically induced metabolic abnormalities three generations out. … A number of studies have indicated that environmental stress can promote epigenetic alterations that are transmitted to and induce pathologies in subsequent generations. … A recent study by Gerlinde Metz and her colleagues at the University of Lethbridge in Canada demonstrated that restraining pregnant rats or, alternatively, forcing them to swim, produced epigenetic damage that put newborns at risk. … Several studies now support the role of environmental stress in promoting the epigenetic transgenerational inheritance of disease.”

Right, so none of these examples would be recognized by Lamarck as an instance of what he was talking about. What is going on is fascinating, but it is in a sense not that different from the standard genetic story: cells encode information for the development of the organism; this information is inherited, but can be altered by the environment; the alterations that are — by chance — adaptive are retained by natural selection, while the rest are eliminated from the population. This is pure Darwinian natural selection at work. Indeed, biologists have understood for some time now that the Darwinian theory is neutral with respect to the specific mechanisms of inheritance: natural selection works regardless of which underlying mechanisms generate the phenotype variation that is being selected, and regardless of how those same mechanisms assure that the pertinent information gets passed from one generation to the next.

So, no, we don’t need Darwinism and Lamarckism, we just need Darwinism, to which the MS added a lot of mechanisms and theoretical understanding back in the 1920s-40s, and to which the EES is adding even more now.

Skinner sees himself as nothing less than the proponent of a new theory of evolution, which he summarizes in a diagram (fig. 2 of his Aeon piece), and that he grandly refers to as the “Unified Theory of Evolution.” He seems blissfully unaware that lots of people have already worked out a new theory — the EES — and that Lamarckism just doesn’t enter into it.

Skinner even fancies himself as triggering a “paradigm shift” in evolutionary biology — referring to the famous account of how science makes progress, proposed back in the ‘60s by Thomas Kuhn (and largely rejected by modern philosophers of science, who Skinner, evidently, has not read). Indeed, I have published a detailed account showing why evolutionary biology has never actually seen, and likely will never see, anything resembling a paradigm shift.

I’m picking on Skinner not because he is the only one writing such things, but because his article is the most recent, highly visible, example of this sort of misunderstanding of evolution on the part of some scientists, a misunderstanding that in turns leads to misinformation of the public. A public, one might add, that is already skeptical and confused about evolution in particular and science in general. We scientists and science popularizers really ought — morally — to do much better than this.