Inherited epigenetic variation - revisiting soft inheritance
Eric J. Richards
Nature Reviews Genetics 7, 395-401 (May 2006)
In case you don't have access, from the conclusion:
Does the existence of inherited epigenetic alleles that are to some degree independent of genetic variation necessitate a modification of our models of evolutionary change? The neo-Darwinian concept of inheritance posits that the hereditary material is 'hard' and impervious to environmental influences. If the formation of epialleles is random and not initiated or guided by the environment, the generation of epigenetic variation could be equated with random genetic mutation without otherwise altering our current view of evolutionary mechanisms. On the other hand, if the physical or behavioural environment of the cell or organism influences epiallele formation, a mechanistic foundation for soft inheritance exists in which the environment could mould a malleable hereditary material.
A growing body of evidence indicates that epigenetic states can be influenced by the environment. For example, prolonged cold-temperature treatments in plants can lead to both chromatin37 and DNA methylation changes at specific genomic loci38. Treatment with DNA damaging agents that are used in traditional chemical mutagenesis protocols can also change epigenetic states39, 40, 41. Indeed, some of the best-studied inherited epialleles in plants were derived originally from chemical mutagenesis experiments42, 43. Some of the most striking examples of environmental modulation of epigenetic states are derived from the recent animal literature. One class of examples involves the alteration of DNA methylation through dietary regimes that alter single carbon metabolism in rodents. In the case of the mouse Avy epiallele (Table 1), dietary supplements (for example, folic acid and vitamin B12) that increase the abundance of the central methyl donor metabolite S-adenosylmethionine elevate DNA methylation of the upstream IAP element and suppress Agouti overexpression44, 45, 46. Another intriguing example of an environmentally induced, mitotically stable epiallele was recently described in rats: nurturing maternal behaviour leads to postnatal remodelling of the epigenetic state of the hippocampal glucocorticoid receptor gene (GR; also known as nuclear receptor subfamily 3, group C, member 1 (NR3C1)), creating a hypomethylated epiallele that persists into adulthood47. Mothers that show poor nurturing behaviour rear individuals with an alternative silent GR gene epiallele. The preceding examples show that epigenotypes can respond to an organism's physical, nutritional and even behavioural environment. Although these examples do not involve meiotic transmission of the environmentally induced epialleles, this is not always the case. A recent publication reported that treatment of gestating female rats with industrial chemicals that disrupt endocrine function can lead to male fertility defects in subsequent generations (F1 to F4), which are correlated with widespread alterations in DNA methylation48. This study resembles an earlier report demonstrating transgenerational effects on gene expression, DNA methylation and growth efficiency that is induced by nuclear transplantation in mice49.
All the elements are in place to allow a type of soft inheritance that is based on DNA methylation and chromatin-level silencing: the creation of alternative epigenetic alleles that are biased by environmental inputs; the stability and mitotic propagation of epialleles; absent or incomplete epiallele erasure; and meiotic transmission. There is no reason on mechanistic grounds to reject the possibility that environmentally induced or modified epialleles can be inherited. It might be more meaningful to ask why we are not constantly confronted with the inheritance of environmentally induced phenotypic variation. In the case of mammals, the answer probably lies in a reasonably comprehensive erasure of epigenetic marks and the early germ-soma divergence that ensures that epigenetic alterations in somatic lineages are not transmitted through the germ line. The germ-soma division formed the core of Weismann's rejection in the late nineteenth century of neo-Lamarckian inheritance. These considerations indicate that epigenetic inheritance is unlikely to mediate in mammals the most extreme form of soft inheritance that involves the transmission of adaptive acquired characters (Box 4). However, a less extreme form of soft inheritance is possible that might be based on the transmission of environmentally induced or influenced epialleles that are generated in the germ line. In this case, there is no reason to propose that these epialleles will have any adaptive significance, without resorting to the contortion of invoking a parallel induction of epigenetic changes in reproductive and somatic lineages. However, in organisms in which reproductive lineages or germ lines are derived from vegetative or somatic lineages late in development and epiallele erasure is less extensive, such as plants, both forms of soft inheritance could operate through epigenetic mechanisms.
Even if it is conceded that the molecular mechanisms are present to mediate soft inheritance through epigenetic mechanisms, the significance of such mechanisms must be questioned. The variation that has been shown to underlie the developmental and phenotypic differences between species occurs at the genetic rather than the epigenetic level. Among individuals in a population, however, epigenetic variation might have a significant role in controlling phenotypic variation50, 51. In addition, epigenetic variation might have a role as a bridge towards genetic end points by facilitating genetic assimilation of characters (for example, accelerated genetic decay of hypermethylated epialleles)52, 53, 54.
Future directions
There are many questions to address about the mechanistic aspects of epigenetic inheritance and the significance of inherited epialleles. Several continuing lines of enquiry will be important in resolving these questions. First, it is necessary to continue to flesh out our knowledge of the machinery that orchestrates epigenetic regulation through biochemical and genetic dissection approaches. Second, studies that parallel Johanssen's pure line experiments55 should continue to determine whether alternative epigenetic alleles can be detected or selected in inbred backgrounds with limited genetic variation56, 57, 58, 59. Third, the meiotic behaviour of epigenetic alleles that are created or manipulated by environmental regimes needs to be examined, starting with the better-understood epialleles. For example, do dietary supplements in mice that carry the Avy allele alter the epigenetic state of the epiallele in untreated progeny, as well as the somatic tissue of the individuals that are treated in utero44, 45, 46? Fourth, several epigenomics projects that are currently underway to chart the epigenetic landscape of large eukaryotic genomes will pinpoint new loci that are sensitive to epigenetic modification and variation60. This information will inform systematic efforts to monitor the stability, environmental sensitivity and meiotic behaviour of many epigenetic alleles in different experimental models. Finally, it will be necessary to extend these findings to natural populations to evaluate the role of inherited epigenetic variation in a real-world context.