Wednesday, March 18, 2015

Epigenomics: the germ of an idea

The 19 February issue of Nature has many papers reporting a major human epigenomics project (other Nature journals have other papers from this project).  We cannot summarize all of them here by any means, and haven't read all the papers, which range far and wide.  But the growing awareness of and, more significantly, serious attention being paid to DNA usage rather than sequence modification as an important aspect of biological function is worth discussing.

A recent article in Aeon by Guthman and Mansfield discusses the conceptual way in which epigenetics, shows that organisms (including humans) are not closed vessels, whose nature is strictly determined by its carefully sequestered DNA sequence, the individual's insides isolated from their environment and its effects.  We knew, of course, about injury, infection, and lifestyle effects.  But what is becoming apparent is that there probably are far wider ranging effects on our cells that are inherited when they divide and hence become functional parts of our genomics.

Epigenetic mechanisms; Wikipedia

The DNA in our genomes contains codes for the amino acid sequence of proteins among other things. But in a given cell, a given protein code is only used when other non-coding sequence elements, called regulatory elements, near the coding part are properly activated.  This is known as gene expression, or gene usage, and is based on specific proteins binding these regulatory elements.  That in turn depends on the environment of the cell within the body.  During our lifetimes, from conception onward, some aspects of our internal physiology alter a cell's gene usage to suit current circumstances.  This can be due to internal feedback, such as during hunger or sleep and so on.  But for a bit of coding DNA to be expressed, or not, depends on the attachment, or not, of particular chemicals to its regulatory elements.  This modification is called epigenetic, because it affects how the DNA is used without changing its nucleotide sequence.

Some epigenetic changes, and their resulting effects on gene expression, seem to be responses to environmental exposures of whatever sort.  These then can affect 'normal' traits but also may lead individuals to have traits ranging beyond normal to pathogenic.  The longer one is exposed, or the earlier, the more likely a complex of epigenetic changes may, like the simultaneous accumulation of mitotically inherited somatic DNA mutations, affect traits.  That is, one might expect epigenetic traits to show an accelerating age-of-onset distribution.  Of course, this would mainly apply to steady exposures throughout life, or to responses a person is programmed in gestation to make to environmental experiences.

It is somewhat fashionable to refer to epigenetic changes as 'Lamarckian'.  In the early1800s, Jean Baptiste de Lamarck suggested that evolution's slow path was driven by its life-ways.  As organisms strove through their behavior to survive and reproduce (this need not be 'conscious'), that behavior modified the organism's transmitted material (genes, of course, were not known), and this led to a continual adaptive track over eons of time.  Lamarck is accused of mysticism because he suggested that striving influences one's inheritance, but he didn't really mean long-term goal-directedness, just trying to get along in one's environment.  Still the organism was more 'proactive', or had more 'agency', to use modern catchwords, than in Darwin's more blind mechanical process of natural selection by which randomly arising variation was screened for fitness.

In a fundamental way, epigenetics does not necessarily imply any sort of Lamarckism.  If life-ways modify DNA expression in cells and that is transmitted when those cells divide, then gene expression patterns acquired by experience are inherited--but one would expect this to be only somatically: exercise might affect epigenetic changes in, say, muscle or heart or lung cells.  Diet might change intestinal cell behavior.  When those cells divide, the epigenetic marking may be transmitted. In that sense, from the point of view of gene usage, we are not what we inherit, except to the extent to which our DNA sequence affects the epigenetic-marking mechanisms of a given environment.

This sort of leakage between a person's bodily integrity and the environment may be very important in terms of who we are during our lifetimes, and the diseases we get or resist.  This could be called 'Lamarckian' at the somatic (body) cell level, so long as one didn't attribute any sort of mental striving to it.  Of course, brain cells might be modified by such striving.  But that's a kind of loose or careless talk that doesn't reflect an accurate understanding of Lamarcikan evolution, which this is not--and yet....

Is it Lamarckian?
For a Lamarckian sort of evolution, the effect of the environment must be imprinted on the genome either before fertilization or during gestation so that the change is transmitted to the next generation, in the germline (to be incorporated in sperm or egg cells).  This is why much interest attaches to epigenetic changes that may occur to a fetus because of a mother's lifestyle during pregnancy.  This is important because then the fetus may carry the change in all its cells, including its own germline cells, which could make it heritable by its offspring.

Epigenetic changes can be actively imposed or erased, so inheritance of these sorts seems generally to be less permanent than changes in DNA sequence.  The latter can be created or erased, by mutation, but that is a much slower process that only happens to individuals and is far less likely to be erased (except by natural selection if harmful).  Epigenetic changes, even in utero, may be common if many or most mothers are exposed to the same environment or behavior.

Even so there are reports, that clearly need to be made rigorous and not exaggerated by scientists or the media, of multi-generational persistence of epigenetic changes that arose in adults.   The effects of famine in Holland during World War II seem to be an example of heritable epigenetic changes as a result of environmental conditions, effects that are still being seen in grandchildren of women who had insufficient calories during pregnancy.  The effect crossed the placenta and altered the fetus' gene usage.  If such a multigenerational effect prove true, the evolutionary, or trans-generational effect must be marked not just on the person's body cells, but on the germline, too.  For example, if exercise-induced changes in muscle-cell gene usage occur, how could such specific change also be imposed on that person's sperm or egg lineages?

There should be a heavy burden of proof on anyone arguing, as some do, that epigenetic changes due to post-natal behavioral or environmental experiences that do not seem genetic (that is, are not encoded in the DNA sequence) are inherited.  If something dietary stimulates epigenetic change, that may affect all cells.  But (to cite one recent claim) epigenetic response to specific odorants by nasal epithelial cells have no obvious way to be imprinted on the exposed animal's sperm cells.  Yet one study reports pairing a pain experience with the expression of a specific odorant receptor in adult males, and that same expression change was seen in his offspring, and then in the grand-offspring, but without further conditioning.  How does the sperm-cell lineage in the rodent's body 'know' to mark a particular receptor gene out of the many hundreds in its genome?  We need a mechanism if we're to believe the report!  That doesn't mean there isn't one, of course, but it seems highly questionable, because what might it be?

Let's put it this way:  If such histologically 'local' (tissue-specific) experience-targeted changes really can be imposed on the germ line, and that can truly persist for multiple generations, then we really may have something different!

There are many reasons why Lamarckian inheritance was rejected after Darwin's and Mendel's work were united a century or so ago.  Lamarck wasn't being foolish, but he did seem to miss the important point that biological traits arose via random mutation, rather than being directly applied on patrimony by experience.  Natural selection screening existing variation is an easier mechanism to understand. Lamarckian inheritance doesn't necessarily lead to species' extinctions, because they can just modify what they do based on their conditions, whereas in Darwinian evolution, most lineages go extinct (that isn't formally necessary, but it seems to be true, and the generic explanation is that the species was too specialized or simply didn't experience adaptive mutations when environments changed).

Still, we can't be dogmatic about accepting Darwin and rejecting anything else.  Epigenetic mechanisms do seem to provide some nuance to our understanding of inheritance.  Chasing down instances and mechanisms at the cell level is a hot fad right now, and it will probably die down to its useful aspects as most fads do.  But doubt is important when it comes to claims of changes not directly related to germline cells to be imposed on those cells would require much more internal cellular communication than we currently understand. On the other hand, life is to a great extent all about signaling among cells, and maybe there is something going on we haven't understood.  Time will tell but meanwhile circumspection is called for.


Hollis said...

Thanks for the interesting discussion. As I read the post I was composing a Comment in my head but then you said it - last paragraph: "Still, we can't be dogmatic about accepting Darwin and rejecting anything else."

It sounds like in your world epigenetics is overrated to ill effect. In mine, any alternative to a simple strictly Mendelian model is blown off as rare at most -- also to ill effect. So I was going to argue that this focus on epigenetics is good, though obviously a matter of perspective. :-)

Ken Weiss said...

I agree. The tendency to dogmatism, including both the temptation to try to hard to be causing a 'paradigm shift' and the denigration of any non-orthodox ideas, is ever-present. I think in fact that none of these things, if they are clearly demonstrated, will cause such a shift. They will show (perhaps) that Lamarck wasn't so far wrong, given his time, and Darwin not quite so right (given his time--and pangenesis is close to Lamarckian inheritance), as generally credited.

Inheritance of experience was part of Hippocrates since it seems so obviously true. The idea of random mutation blindly screened has its obvious appeal and ease of understanding (no mechanism need be known to invoke Darwinian processes).

But life has many surprises remaining, I think. However, I see nothing in this that will shake our basic understanding of evolution, even if they were to give us really new mechanisms of causation and inheritance.

However, of course, we have to wait for the dust to settle to see how much of epigenetics is very different from what is easily understood. If it's a lot, then we will have learned a lot, which is good. If it's a bust, we'll also have learned a lot, which is good.