I could be totally wrong on this one, which is cool. Happens often enough. And I won’t profess to be an expert in epigenetics, though I’m not completely ignorant of the field either. But as I was reviewing a book on epigenetics the other day, I stopped for a moment and wondered whether, what we think of as “epigenetic inheritance,” is that actually genetic inheritance in disguise?
Let me explain. The epigenome is an extroardinary discovery that completely changes the way we think about molecular biology and inheritance. Many forms of epigenetic inheritance are relegated to somatic cells within a single organism and are probably at least partially the causes of progressive aging, disease, and death. Some, however, appear to transcend the single organism and are transmitted to its offspring. This is called Transgenerational Epigenetic Inheritance. Lamarck is surely doing a jig in his grave over this one, even if his example of the elongation of a giraffe’s neck through stretching still sounds just as wrong today as it did back then.
Jean-Baptiste Lamarck. Image borrowed from here.
While most forms of acquired epigenetic changes incurred over a lifetime are reprogrammed during both gametogenesis (the development of sperm or egg) and within one cell cycle of the newly-formed zygote, a select number of epimutations are carried over to the offspring [1]. It’s still currently unknown by what means this carry-over is achieved, except that it does sometimes occur.
DNA is an awesome structure. Not only can it replicate itself relatively faithfully, but it provides a fairly contiguous scaffolding that’s inherited each new generation. Think of all the molecules that must dock with it, interact with it, move it, mutate it, reshape it, close it down, and open it up. These occur in predictable though dynamic fashions because of DNA’s relative stability.
The epigenome, on the other hand, does have moderately stable means for replication in terms of the various machinery which are used during DNA replication to transfer similar patterns from mother to daughter strand. However, it’s replication is still dependent upon the patterns that are already present, and in the case of newly-formed gametes or a zygote, those patterns are wiped clean, especially in the case of methylation, so as to promote totipotency. If reprogramming is ubiquitous, how could newly-acquired epigenetic patterns of the parent gamete be transferred to the offspring? How indeed.
Unlike DNA, the epigenome is not contiguous. Instead, it’s made up of many unrelated parts. And it’s dependent upon many products which arise from the DNA and comprise its components. In short, the epigenome houses information on how to interact with DNA and other nuclear molecules but it doesn’t store information on how to replicate itself as a whole.
But if the epigenome is constant and exhibits a certain predictability post-reprogramming, then the data for its replication must be stored somewhere. As it may come as no surprise, the first thing that pops into my head is DNA. We know its the closest thing cells have to a hard drive. But the problem with this theory though is that we also know that certain transgenerational epigenetic changes can be inherited and yet the underlying gene sequence remains the same [2].
But remember that DNA is contiguous and sequences both upstream and down affect the dynamic shape of a given gene region. Even extraordinarily distant sequences on other chromosomes can have the potential to act as enhancers. And the shape of DNA undoubtedly affects its epigenetic patterning, and vice versa. Ultimately, shape in biochemistry is everything. Shape equals function. You gotta have the right molecules in the right place and the right time.
So how could epigenomic changes be inherited from parent to offspring without changing the underlying sequence of the gene they’re directly affecting? One possibility I would entertain is by changing the gene’s shape through mutation somewhere else in the DNA which ultimately affects local epigenetic patterning. That can potentially be inherited and via a method we’re already familiar with.
I don’t know. Like I said, I’m no epigenetics expert. This was just an idea that occurred to me and thought I’d throw it out there. Please feel free to pipe up and throw any wrench into my works. Always eager to learn and refine ideas. 
