Epigenetic Mechanisms, Quick & Dirty
Jablonka & Raz (2009) show us this elegant illustration of broad and narrow epigenetic transmission.
Epigenetic inheritance in the broad sense is the inheritance of developmental variations that do not stem from differences in the sequence of DNA...information transference that can take place through developmental interactions between mother and offspring..., through social learning..., and through symbolic communication.
We...define cellular epigenetic inheritance as the transmission from mother cell to daughter cell of variations that are not the result of differences in DNA base sequence and/or the present environment. Transmission can be through chromatin marks, through RNAs, through self-reconstructing three-dimensional structures, and through self-sustaining metabolic loops.
In the single-cell "bottleneck" variety of epigenetic inheritance (pathway a in the above diagram) Jablonka & Raz focus on...
The environment may induce epigenetic variation by directly affecting the germline or by affecting germ cells through the mediation of the soma, but, in either case, subsequent transmission is through the germline.
According to Jablonka & Raz (2009), there are 5 effects of epigenetic mechanisms & inheritance vis-a-vis evolution:
(i) evolutionary change occurring through selection of epigenetic variants, without involvement of genetic variation; (ii) evolutionary change in which an initial epigenetic modification guides the selection of correlated genetic variations; (iii) evolutionary change stemming from the direct effects of epigenetic variations and epigenetic control mechanisms on the generation of local and systemic epigenomic variations; (iv) evolutionary change resulting from the constraints and affordances that epigenetic inheritance imposes on development; and (v) evolutionary change that leads to new modes of epigenetic inheritance.
Siberian Silver Fox Experiments
The Siberian silver fox experiments are so cool, & often cite them as an example of gene linkage. Honestly, I was just BSing in suggesting that the curly tails, rounded nose, etc. were possibly linked on the same chromosome to tameness & recognized that there might be other factors involved. Lo & behold, a citation in Jablonka & Raz (2009) pointed us toward epigenetic studies to come out of that body of research.
It turns out that the coat spotting & non-spotting variation that we associate with domestication occurs too quickly to be pure mutation, though it behaves like a dominant & semi-dominant trait, & couldn't be explained by inbreeding because the inbreeding coefficient was too low (0.03). Instead, they believe
the stress of domestication and selection for tameness targeted genes with large effects in the neuro-hormonal system...and may have heritably reactivated some of them...This epigenetic interpretation, in terms of new epimutations rather than new mutations, explains the high rate of appearance and disappearance of some phenotypes, and support for this comes from the fact that at least two of the genes (Agouti and C-kit) that seem to be involved in the changes are known to have heritable epigenetic variants in mice...
One aspect of epigenetics that seems important here is the concept of canalization, introduced by Waddington several decades back (he also introduced the concept of epigenetics in general, which everyone rightly thought was Lamarkian & wrongly ignored--turns out he was on the money). Roughly, canalization means that some environmental perturbation pushes a phenotype into a canal or valley, whereafter selection pressures prevent the phenotype from returning to its previous state because the "climb" up the sides of the canal or out of the valley are too steep. Think of a marble on a tabletop that is essentially flat but has a valley to one side of it. Stochastic chance dictates that the marble can roll any which way, but if it happens to roll toward the valley, it gets stuck there & can only roll further in the valley. Or as this image illustrates, there are several possible environmental variations possible, but once a phenotype goes one way (plastically), it cannot go back.
So it seems to be with the silver foxes. Once an environmental condition pushes silver foxes (or wolves before them) one way (luring tame ones to their yummy debris & handouts) or another (spooking the nervous ones to run away), a cascade of epigenetic mechanisms pushes them further along. At that point, according to this model, tame ones cannot become anxious/aggressive & vice versa.
While cute silver foxes that you can cuddle with get all the press, the less publicized but equally fascinating is the aggressive foxes that want to rip your face off.
So what's going on with these aggressive foxes? According to Popova (2006), there are at least 16 genes that influence aggression, but most aggression behavior is influenced by just a few of those. A major player seems to be serotonin (5-HT). The 5-HT pathway in the brain suppresses aggression. 5-HT is not a gene though, it is a hormone; & genes code for proteins. So if there's a gene change, what is/are the gene(s)? It could be any gene that produces an enzyme involved in the essential mechanisms of the 5-HT system, which include synthesis/degradation, reuptake in the synaptic cleft, & density/sensitivity of receptors (for more background on 5-HT, I've written on this before here). As the figure below illustrates, there are enzymes that catalyze serotonin synthesis (TPH & decarboxylase of aromatic l-amino acids), two enzymes that help break serotonin down (MAO A & B), & an enzyme (SERT) that transports serotonin.
There are two TPH genes, & it is the 2nd one (TPH2), expressed in the brain & responsible for the central nervous system, that effects 5-HT & seems to be responsible for aggressive behavior.
Silver foxes displaying friendly responses to human contact were shown to have higher 5-HT and 5-HIAA levels, and higher TPH activity in the midbrain and hypothalamus in comparison to nonselected wild-type silver foxes bred in captivity. Importantly, the changes were found in the midbrain representing the area of main location of TPH2-synthesizing cell bodies. These findings were interpreted as an indication of an increased activity of the brain 5-HT system in the tame animals and, subsequently, a decreased activity of this system in highly aggressive animals.
MAO A has a higher affinity for 5-HT & is considered the principle enzyme in breaking down serotonin. When MAO A is disrupted in mice, they get more aggressive. Deletion of SERT (the transporter that allows 5-HT molecules not taken up by post-synaptic receptors to be recycled & reused) in knockout mice also produces aggressive behavior. Finally, there are 14 different subtypes of 5-HT post-synaptic receptors. Genetically low aggression has been associated with increased expression of specific subtypes of these receptors in the midbrain & specific densities & function in specific regions of the brain. These likely function to suppress aggressive behavior.
The figure below depicts this as essentially two pathways, which we can compare analogously to the Jablonka & Raz depiction of the narrow "bottleneck" pathway, albeit via two cells (or genes). I think.
If any one of these mechanisms or either of these pathways influences aggression, they will interact with the environment to mutually reinforce themselves & push the marble down toward the other pathway too. In other words, if the stress of domestication bumped the marble off the plane, having even only a slightly higher tendency of aggression relative to tameness will result in amplification of the entire aggression pathway, even if the environmental conditions of captivity are thereafter removed (i.e., the animal is released). What still remains to be clarified is how the initial brain changes occur & the roles of other mechanisms in the system.
Zane Thayer & Chris Kuzawa review data that offer a clue. They point out that "psychosocial stress contributes to the social gradient in health" (2011:799). This is well-established by now, but the mechanisms are interesting. In two studies particularly relevant to our question of how the stress of domestication may influence aggression in silver foxes, childhood abuse was associated w/ methylation differences at the GR (glucocorticoid) locus in the hippocampus & the serotonin transporter protein (SLC6A4) locus. Another study found that maternal depression during pregnancy predicted stress reactivity & methylation of the GR locus in buccal cells of their infants 3 months after birth. Methylation is the addition of a methyl group to a substrate or substitution of an atom by a methyl group. This can take place in DNA or proteins. In DNA, it can result in the change of an amino acid base, thus a change in the genetic code resulting in production of different or altered proteins. In proteins, it results in regulatory changes in the protein functions, so methylation can have wide-ranging effects. The methylation of the GR receptor locus may affect things like the binding of glucocorticoids to the receptor. Glucocorticoids are best known for their role in stress response, but, relevant to this discussion, they are also operative in memory consolidation and learning, as contextual fear conditioning, among many other functions. The SLC6A4 serotonin transporter protein terminates action of serotonin in the synapse & recycles it, which is a key function in mood stabilization. Low serotonin is associated with high fear response.
So the psychosocial stress of domestication in some silver foxes could result in methylation of glucocorticoid & serotonin receptors, directly influences fear/aggression response in pups, that persists throughout their lives. It can also influence depression in mothers that is passed on in the receptor activity of their pups. We can link this with Larry Schell's model of risk-focusing. Replace "SES" with "personality." Fear/aggression in mothers increases the risk for fear/aggression in descendants, as tameness in mothers increases the risk of tameness in descendants.
Although the model above suggests gene line changes, the broad epigenetic view suggests that some of these influences may not influence the germ yet still persist over multiple generations because of the influence that maternal disposition has on offspring gene expression. Or, as Thayer & Kuzawa note
...Genes are regulated by biological "memories" of experiences acquired earlier in our own lives, and even by recent predecessors... (2011:801)