sign in sign up
<<
Home , Emma Whitelaw, Epigenome, Epigenomics, Human epigenome, Human Epigenome Project

Feature Human Project—Up and Running Jane Bradbury

pigenomics is one of the are switched on and off to produce a many ‘’ that is being person. And knowing more about the The Human Project Etalked about in the wake of may provide clues the Human . But to what goes wrong in and other … ‘provided the blueprint, what is an epigenome, and why have diseases. the Wellcome Trust Sanger Institute for life but the epigenome What Is ? (Hinxton, United Kingdom) and will tell us how this whole AG (Berlin, Germany) Most people have a fair idea of what recently announced the launch of the is meant by . They know that thing gets executed’. (HEP), characteristics such as eye colour are a fi ve-year undertaking during which specifi ed by the DNA sequence within DNA sites throughout their genome. But not everything is That ‘something else’ is chemical the human genome will be mapped? that simple. For example, genetically modifi cations of genes that are The HEP is the brainchild of identical twins can be very different. heritable from one cell generation immunogeneticist Stephan Beck of the ‘One might be normal, while the to the next and that affect gene Sanger Institute and Alexander Olek, other is autistic’, explains expression but do not alter the DNA chief executive offi cer of Epigenomics researcher David Allis (Rockefeller sequence. Epigenetic modifi cations can AG. The , University, New York, United States). affect the DNA itself or the explains Olek, ‘provided the blueprint ‘We can’t explain that on the basis that package the DNA into chromatin. for life, but the epigenome will tell us of pure genetics because the DNA is Developmental geneticist Wolf Reik how this whole thing gets executed’, identical. Something else must be at (The Babraham Institute, Cambridge, what determines when and where genes play’. United Kingdom) describes these modifi cations as ‘red and green traffi c lights that are superimposed on top of the genome to tell the genes whether Box 1. The Same but Different they should be active or inactive’. For Geneticist Emma Whitelaw (University of Sydney, Australia) studies phenomena in mice similar individual cells, this code on top of a to those seen in human monozygotic twins, whereby genetically identical individuals can look code helps to determine whether a cell or behave very differently. About a decade ago, she explains, ‘we noticed that in a litter of mice is a blood cell, a fat cell, or something that had all stably inherited a transgene at a specifi c locus, some mice expressed the transgene, else. And in the case of monozygotic but others didn’t’. This variable gene expressivity in genetically identical animals suggested that twins, unexpected differences may some kind of epigenetic mark had occurred differentially between individuals by chance. Similarly, result from chance variations in this identical mice carrying the agouti viable yellow coat colour gene can range in colour from yellow superimposed code (Box 1). through mottled to brown, depending on whether the gene is expressed or not. And while a The study of these modifi cations— in the axin gene called axin-fused produces mice with kinky tails, the degree of kinkiness what they are, how they are laid varies among genetically down, and the processes that they identical littermates (Figure 1). control—is the fi eld of research known For all these mice, Whitelaw as epigenetics. An epigenome is the has discovered that the description of these modifi cations mysterious epigenetic marks across the whole genome, but unlike responsible for variable the genome DNA sequence, each expressivity are inherited organism has multiple — between sexual generations. for example, in different cell types— So, in the case of the agouti viable yellow mice, yellow mice Copyright: © 2003 Jane Bradbury. This is an open- have more yellow offspring access article distributed under the terms of the than mottled or brown Creative Commons Attribution License, which offspring. Variable expressivity permits unrestricted use, distribution, and reproduction in any medium, provided the original may be a quicker way to deal work is properly cited. with environmental change than DNA mutation and, Abbreviations: HEP, Human Epigenome Project; MHC, DOI: 10.1371/journal/pbio.0000082.g001 major histocompatibility complex suggests Whitelaw, it may Figure 1. Spot the Difference: Same Genes but a Different Kink be that variable expressivity Jane Bradbury is a freelance science news in the Tail writer based in Cambridge, United Kingdom. E-mail: is involved in some way in (Photograph courtesy of Emma Whitelaw, University of [email protected] evolution. Sydney, Australia.) DOI: 10.1371/journal/pbio.0000082

PLoS Biology | http://biology.plosjournals.org Volume 1 | Issue 3 | Page 316 that may change during its lifetime in enormous undertaking. ‘Knowing the DNA methylation sites throughout the response to environmental cues. methylation sequence’, says Bird, ‘will entire MHC region, the consortium be an essential backdrop to future set out to look at the 150 expressed DNA Methylation research on DNA methylation and its genes within this region, explains Beck. Methylated DNA was the fi rst biological effects’. Allis agrees that For each gene, the scientists chose epigenetic mark to be discovered. DNA methylation is an important part two areas to analyse for methylation, That this epigenetic modifi cation is of epigenetic marking, but ‘although each about 500 basepairs long. ‘One important is suggested by the waves mapping DNA methylation patterns is window was in what we thought was of demethylation and de novo (new) the logical and good place to start, by the region’, says Beck, methylation that occur at specifi c its lonesome it won’t explain the whole ‘and the other corresponded to the stages during the development of an epigenetic phenomenon’. most CpG-rich region in the gene’. animal from a fertilised egg. Soon Beck and Olek began thinking about The promoter regions are the places after , there is a massive mapping the human epigenome more where the elements that control gene active loss of methylation from the than fi ve years ago. Even as long ago expression are often located, and DNA paternal genome, explains Reik. At the as 1998, explains Olek, ‘it was pretty methylation occurs at residues same time, the maternal genome loses natural to be thinking about what within CpG motifs, hence the choice of some methyl groups through passive would come after the Human Genome CpG-rich regions. demethylation. ‘We think this process Project, and Stephan believed that The methylation status of more erases all the epigenetic memory in the it had to be methylation sequencing than 100, 000 sites was determined gametes [ and egg cells] except because that has the potential to tell during the three-year pilot study and for some special imprinted genes whose us about the hundreds of we the results analysed to show where expression depends on whether they really have’. there were methylation differences are of maternal or parental origin’, says In October 2000, the Human in the MHC between different Reik. Later in development, widespread Epigenome Consortium (the Sanger tissues (Figure 2). ‘We found major de novo DNA methylation occurs in a Institute, Epigenomics AG, and the methylation differences between loci process known as reprogramming. De Centre National de Génotypage in and between tissues’, says Beck, ‘and novo methylation rarely occurs after Evry, France) started a European we are particularly interested in what an early developmental event known Union-funded pilot project to map we call methylation variable positions as gastrulation, except in cancer cells the methylation sites within the major (MVPs), which we believe will advance where special unmethylated regions histocompatibility complex (MHC) our ability to understand and diagnose of the genome, known as CpG islands, region in seven different human human disease’. Olek echoes Beck’s often become methylated. tissues. Rather than look at all the satisfaction with the results to date. ‘We ‘People generally agree that DNA methylation is important for imprinting and reprogramming’, says Reik, ‘but not everyone agrees that it plays a role in DNA activation and inactivation during development’, a role fi rst proposed more than 25 years ago. ‘People are still hesitant about saying that DNA methylation is important in making sure the right genes are expressed in the right cells at the right time because until recently we have had relatively few good examples’, says DNA methylation expert Adrian Bird (The Wellcome Trust Centre for , Edinburgh, United Kingdom). Many people have reported correlations between methylation states and gene activity but, says Reik, ‘you can look at such correlations until the cows come home and it tells you nothing about causality’.

The HEP Pilot DOI: 10.1371/journal/pbio.0000082.g002 Even given the doubts about the Figure 2. Example of HEP Methylation Data for the BAT8 Gene, an MHC-Linked exact role of DNA methylation, Bird, Methyltransferase Allis, and Reik believe that mapping In this case, two regions marked in red were analysed in 31 samples (rows) from seven tissues (blocks of rows). The predicted structure of the gene is shown below with methylation patterns across the coloured boxes indicating the . As part of the launch of the HEP on October 7, human genome, which is what the 2003, these data and others, together with experimental details of the pilot study, were HEP plans to do, is a worthwhile albeit released at www.epigenome.org.

PLoS Biology | http://biology.plosjournals.org Volume 1 | Issue 3 | Page 317 can see that there is hidden treasure in way’. Recently, a link between DNA ‘methylation has gone a bit wonky in the data that needs mining, and that methylation and various respects. In particular, CpG treasure is obviously going to get much was uncovered, a fi nding that Reik islands, regions of the genome that bigger when we sequence the whole says has ‘galvanised the fi eld’. In are normally nonmethylated, get thing’, he says. Neurospora and Arabidopsis, he explains, methylated, and in some cases this there is now good evidence that seems to shut off tumour suppressor On to the Full HEP: A Public/Private histone modifi cation puts down some expression’. Loss of tumour suppressor Initiative sort of mark that is read by a series function can remove the controls that That endeavour—the mapping of of binding proteins, including the normally restrict cell division, leading DNA methylation sites in all 30,000 enzymes that methylate DNA. How the human genes in around 200 samples— positional specifi city of the fi rst mark is is now underway. ‘These will be very determined remained a mystery until carefully chosen’, says Olek, ‘with the the end of 2002, when the fi rst reports Recently, a link between help of additional academic centres appeared suggesting that in Arabidopsis who we hope will join our consortium’. small might be involved. The DNA methylation and Sample preparation and amplifi cation small RNAs are thought to bind to histone methylation was will be done at Epigenomics AG, the ‘complementary’ sequences in DNA Sanger Institute will do the sequencing and somehow target the region for uncovered, a fi nding and raw data analysis, and then the data epigenetic modifi cation. Whether will be jointly evaluated and mined, ‘a similar mechanisms exist in animals that [Wolf] Reik says has huge collaborative experiment’, says remains to be seen. ‘galvanised the fi eld’. Olek. To answer this and other epigenetic The Wellcome Trust and puzzles, scientists are organising Epigenomics AG are jointly funding themselves into collaborative networks phase I of the HEP, thus avoiding a to maximise progress. In Europe, for to the unrestrained cell growth that is rerun of the situation that occurred in example, (Research a characteristic of cancer. There are the Human Genome Project, in which Institute of Molecular Pathology, also alterations in histone a commercial company and a public Vienna, Austria) is in negotiation patterns in cancer, and cataloguing effort ended up competing with each with the European Commission over these epigenetic changes in different other. An agreement has been drawn the establishment of a Network of tumours may provide a way to tag up that ensures that both partners Excellence in epigenetic research. to help clinicians to choose the benefi t from the collaboration, and This ‘Epigenome Network’, which best therapeutic regimen for individual all the data generated will be publicly will focus on epigenetic mechanisms, patients. Epigenetic changes may also available through the Internet in ‘will allow us to build a long-lasting provide scope for new therapeutic accordance with the consortium’s data platform to bring together colleagues approaches, says Allis. If a gene is release policy (www.epigenome.org and working on epigenetics in Europe inappropriately switched off during www.sanger.ac.uk/epigenome). and make this area of research very cancer development, he explains, strong’, says Jenuwein. Negotiations ‘maybe if we could de-silence it, we with the European Commission for could reverse tumour development’. an anticipated €12 million of funding Bird agrees that reversing epigenetic The mapping of DNA over fi ve years should be completed marks in this way is an experiment methylation sites in all soon, and, if all goes well, the Network that has to be done. But, he asks, should start work in Spring 2004. ‘is epigenetics the Achilles heel of 30,000 human genes in The Network will not itself fund any cancer, or are there too many genetic epigenome mapping projects like changes that are hard-wired to be able around 200 samples is now HEP, says Jenuwein, but will build to reverse it simply by interfering with underway. strong links with scientists involved in epigenetics?’ mapping the ‘epigenetic landscape’, A better understanding of the pattern of DNA and epigenetics might also improve the histone modifi cations. effi ciency of cloning animals from Histone Modifi cations and somatic cells (reproductive cloning). Epigenetic Landscaping Just an Academic Exercise? It is very hard to clone animals from DNA methylation is not the only Epigenetics is a fascinating somatic cells and, says Bird, ‘the epigenetic mark. There are also a phenomenon, but why are funding more you look at the surviving cloned staggering number of acetylation and bodies keen to fund expensive animals in molecular terms, the methylation marks on the — collaborative enterprises like the odder they look’. Reik suggests that the proteins that bind DNA to form HEP and the Epigenome Network? this could indicate that epigenetic chromatin. Histone modifi cations One hope is that an understanding reprogramming—the process needed alter the chromatin structure and thus of how cells execute their genomes to turn a somatic cell back into a regulate , says Allis, normally will provide important clues totipotent cell—is very ineffi cient. ‘We ‘and somehow the cell exploits each about what goes wrong in diseases have some evidence that there is a and every one of these in a meaningful such as cancer, in which, says Bird, link between the success of epigenetic

PLoS Biology | http://biology.plosjournals.org Volume 1 | Issue 3 | Page 318 reprogramming and that of cloning, is an exciting area of research. A far DNA Methylation The DNA Methylation Society homepage provides and we really have to understand cry from the situation 10 years ago, more details on DNA methylation at http:⁄⁄ the reprogramming process to get says Reik, ‘when the fi eld had the dnamethsoc.server101.com. reproductive cloning to work’, says smell of something esoteric that no Reik. ‘At the moment the whole one really understood and whose process is a black box’. Because of signifi cance was not clear. Now it is one Recent Reviews on Epigenetics Fischle W, Wang Y, Allis CD (2003) Histone and this, many scientists argue that human of the mainstream exciting areas of chromatin cross-talk. Curr Opin Cell Biol 15: reproductive cloning, leaving aside the post-genomic biology’, he concludes, 172–183. ethical concerns about it, simply isn’t and one which will hopefully yield to Jaenisch R, Bird A (2003) Epigenetic regulation of gene expression: How the genome integrates a safe option at present, a concern the might of large consortia like the intrinsic and environmental signals. Nat Genet refl ected in the recent call by 60 Epigenome Network and the HEP. Ⅵ (Suppl) 33: 245–254. Reik W, Santos F, Dean W (2003) Mammalian scientifi c academies for a ban on its epigenomics: Reprogramming the genome for development. Where to Find Out More development and therapy. Theriogenology 59: Epigenetics then, both at the level of The Human Epigenome Project 21–32. Information on the project, publications, and up- In addition, there is a collection of articles on unravelling mechanisms and mapping to-date data releases are posted at http:⁄⁄ many aspects of epigenetics in Ann N Y Acad epigenomes and at the applied level, www.epigenome.org. Sci 2003: 983.

PLoS Biology | http://biology.plosjournals.org Volume 1 | Issue 3 | Page 319