This is another archival repost from the old blog — this one from January 2008. The post is part five in a series. The series so far can be found here.
In the first installment of Sunday Syndrome I used the example of Prader-Willi Syndrome. This week we’ll bring in a very different disorder, Angelman’s Syndrome (OMIM:#105830), whose surprising connection to Prader-Willi gives us some interesting insights into how inheritance, evolution and development work — insights whose application to medicine includes everything from understanding cancer to developing stem-cell therapies. Angelman’s syndrome is also known as “Happy Puppet syndrome”, because of its symptoms: severe mental retardation with limited language ability and a habit of laughing; and abnormal movement, a little bit like the jerky movements of the puppets in Thunderbirds. The interesting thing about Angelman’s is how it is inherited: the most common culprit is a deletion of bands 11-13 in the long arm of chromosome 15 (15q11-q13): the same region as Prader-Willi.
But Angelman’s and Prader-Willi, if they are deletions of exactly the same piece of DNA, should surely be one syndrome, combining the symptoms of each and always co-occurring? And yet we only ever see one or the other, never both. The clue to why this should be comes from a closer inspection of the inheritance pattern. We inherit two copies of each of our chromosomes: one from each parent (with the exception of the sex-chromosomes). A deletion on the paternally inherited chromosome 15 causes Prader-Willi, but never causes Angelman’s. A deletion on the maternally inherited version causes Angelman’s, but never causes Prader-Willi. Actually, it’s not quite that simple: “uniparental inheritance”, in which one parent contributes two copies of a chromosome, while the other parent contributes none (an improbable coincidence of errors, but with some documented cases). Two copies from the father and the child has Angelman’s, two copies from the mother and the child has Prader-Willi. Conclusion: an intact paternally inherited copy of chromosome 15 is required to prevent the child developing Prader-Willi, and an intact maternally inherited version is required to prevent the child developing Angelman’s.
It can’t just be a coincidence that the syndromes correlate so perfectly with the parent in every one of the thousands of documented cases, nor can it be some simple historical accident: during gametogenesis (sperm and egg production) homologous chromosomes pair up and sections are shuffled between them. Therefore, Y chromosome aside, no chromosome follows an all male or an all female line: the relevant section of the paternal chromosome 15 may be a part of your grandmother’s chromosome 15, for example. Something must be being done to the chromosome duringgametogenesis , something that males and females do differently, but which is ubiquitous throughout the species. The solution to this riddle is a marvelous quirk of inheritance called “genomic imprinting”. During gametogenesis chemical modifications are made to a selection of genes. These modifications do not change the gene sequence, but they do alter gene expression (time, place, quantity, and so on) — sometimes effectively switching a gene off altogether.
This raises the question: why should such a mechanism arise? It could be purely the result of neutral evolutionary mechanisms, of course, but this seems unlikely: a layer of complexity that carries risks such as Prader-Willi and Angelman’s suggests something more complicated is going on. Deducing how and why something evolved is difficult — even when one has fossils or comparative genomics with which to test hypotheses. Still, we can make hypotheses, and we can build models to test their plausibility — so long as we don’t put too much weight on mere hypotheses and models. One of the more plausible and interesting models proposed for imprinting is Tom Moore and David Haig’s 1991 “parental conflict” model. This model follows from the fact that mothers and fathers do not invest an equal amount of their resources in a child: the mother invests in the pregnancy, with a considerable risk of death, followed by protection of the child during vulnerable early life; the investment of the father depends on the species, but in all mammals it is less than the mother. An additional important fact is that mothers can be certain which children are their own, while it is possible for fathers to be deceived.
In such situations, mothers and fathers are predicted to evolve different strategies for maximising the survival of their genes. The mother’s strategy is to value her long term fertility above the survival of any one child. The father has less interest in his partner’s long-term fertility — after all, the relationship may not last — so he values a child’s survival above the ability of the mother to bear children in the future. These are not conscious strategies or values, of course, and for the purpose of this post, we are only interested in their physiological rather than behavioural manifestations. Moore and Haig’s model proposes that we should expect fathers to encourage maximum growth rate of the child in order to minimise the length of the most vulnerable stage of the child’s life, and to give the child a head-start against all those children of other fathers. The mother should fight back against this: rapid growth could damage her body and her chances of future offspring, or may put her other children in danger.
Whatever the reasons for imprinting having evolved, the process itself need not have been especially complicated: the mechanisms for heritable modification of gene expression are not used for imprinting alone, and are likely far more ancient than imprinting. The steps involved in setting up an imbalance in imprinting may be relatively small, and once initiated could canalise evolution into reinforcing the imbalance (and I may elaborate on that topic another day). Imprinting is just one of several examples of epigenetics: the heritable (either in terms of parent to child, or in terms of cell division within an individual) modification of gene expression. One of the most important mechanisms of epigenetics is chromatin modification. Chromatin is a protein structure associated with DNA, and which can control access to genes by tightly packing the DNA and preventing it associating with the transcription machinery. Some of the other roles of epigenetics include X-inactivation, the process by which the extra X chromosome is silenced in females, in order to equalise gene expression with males, who have a single X; the silencing of some of the junk DNA in the genome, such as pseudogenes and parasitic DNA acquired from viruses; and most interestingly, with the differentiation of cells into their specialised roles.
This latter function, sometimes described as “cellular memory”, brings us back to our tale of development. Development is to a large extent about cells taking specialised jobs in different tissues, organs and systems. To do this, cells remodel their genome by changing the pattern of repression by chromatin in response to external signals (such as the concentration gradients discussed at the end of the last entry). Differentiation by chromatin remodeling of the genome allows different cells to produce unique sets of metabolic enzymes, receptors and cellular structures, and to respond differently to events outside of the cell. Not all such modifications are permanent, however, and chromatin remodeling goes on throughout the life of the cell — in response to external events, or in order to control cell division, for example — as a contributor to short- and medium-term control of gene expression. It is tempting to think of chromatin as being in control of genes because of this role it plays. Yet the opposite is just as true: the way in which chromatin remodeling responds to signals is just as much a product of genes, and ultimately evolution.
Just as the genome is vulnerable to damage, leading to diseases, so the epigenome is vulnerable to damage. Cancers, for example, occur when there is a loss of regulation of the cell cycle. We talk of tumour suppressor genes (which slow down the cell cycle) and oncogenes (which speed up the cell cycle) picking up a mutation that jams them in an always-on or always-off state. But the genes themselves need not be mutated: abnormal control of transcription will have the same outcome, and we commonly see cancers in whichepigenetic abnormalities lead to an over- or under-expression of important genes. Additionally, our knowledge of epigenetics is enabling the development of cancer drugs which target those genes which have become jammed, whether by epigenetic or genetic mutation. Epigenetics has another important medical application: induced pluripotent stem cells (IPS), which were the big science news of November, when the first human IPS cells were produced. The production of IPS cells essentially involves the de-differentiation of an adult cell into a mimic of early embryonic stem cells, which have the potential to develop into any other adult cell. An important part of this process is to reverse all of those changes in chromatin structure that have been made during development. This, it turns out, can be done by switching on a few specialised genes which have already evolved to do that job.
With such a diverse range of important applications epigenetics is set to be a newsworthy field. This leads on to my final point about epigenetics. As the youthful epigenetics begins to break through to the mainstream it has begun also to pick up quacks and cranks. The religious-in-crisis, afraid of the advances that are explaining consciousness, have cited epigenetics as the source of free will. Faith healers and quacks jump on epigenetics as the mechanism by which one can alter gene expression with “mind energy” (which also has something to do with quantum physics, I’m told). Others cite imprinting as a mechanism for the inheritance of acquired characteristics. Epigenetics is important and exciting, but it is not the revolution that these groups have been so desperately seeking. Epigenetics does not break up the rules of biology, and it does not mean that any old failed ideas suddenly become true; it does not support the claims of alternative medicine practitioners or eastern mysticists any more than quantum mechanics does (that is: not at all). Epigenetics fits comfortably in mainstream biology: it is no less a product of evolution than any other aspect of our heredity; and in common with the rest of molecular biology, it is a product of and a servant to genetics, not just a master over it.
Note: I tried a variety of titles for this post, and didn’t like any of them. A prize† for anyone who can come up with a good title.
† The author reserves the right not to award any prizes.
- ^ Moore, T. and Haig , D. 1991. Genomic imprinting in mammalian development: a parental tug-of-war. Trends Genet 7(2), pp. 45-49. doi
- ^ See e.g., Kim TY, Bang YJ, Robertson KD (2006) Histone deacetylase inhibitors for cancer therapy. Epigenetics 1(1):14-23.
- ^ Nimet Maherali, Rupa Sridharan, Wei Xie, Jochen Utikal, Sarah Eminli, Katrin Arnold, Matthias Stadtfeld, Robin Yachechko, Jason Tchieu, Rudolf Jaenisch, Kathrin Plath, and Konrad Hochedlinger (2007) Directly Reprogrammed Fibroblasts Show Global Epigenetic Remodeling and Widespread Tissue Contribution. Cell Stem Cell 1:55-70. doi
- ^ I also spotted Dawson Church’s book The Genie in Your Genes: Epigenetic Medicine and the New Biology of Intention, which looks like it is making similar claims, though I’ve only seen the blurb so I can’t really comment on that.
- ^ Eva Jablonka and Marion J. Lamb are supposedly proponents of this, but my only knowledge of their work is second hand, so I can’t be sure that I’m not getting the garbled accounts of those trying to misuse their work.