So, in this lecture and the subsequent lecture I'd like to now consider how the methylation at these imprint control regions actually brings about that mono-allelic parent of origin specific gene expression or the imprinted gene expression. So, the most important thing to remember is that while I've said in other circumstances that DNA methylation is synonymous with silencing this is not necessarily the case for imprint control regions. So because imprint control regions have different mechanisms of action, and they are not necessarily promoters, then methylation at an imprint control region does not necessarily mean that the surrounding genes will be silenced or inactive. And so now, we need to consider that, instead of this, there are different mechanisms dependent on the classes that you're looking at. And methylation of an imprint control region will tell you about whether its maternally or paternally imprinted, but not about the expression of the genes within that cluster unless you know how that cluster works. So in this context we are going to think of three different clusters. And we are going to go through first of all the Kcnq1 cluster, which is located on human chromosome 11. This is controlled by a long non coding RNA. We'll then go through the H19 or Igf2 cluster also found closely linked on human chromosome 11 but this controlled in a different way by enhancer blocking. And each of these two clusters are disrupted in an imprinting disorder known as Beckwith Wiedemann syndrome. So we'll talk about Beckwith Wiedemann Syndrome and also after describing these two clusters. Then in the next lecture we'll start to think about the Snrpn cluster which is on human chromosome seven and is also controlled by a long non-coding RNA. Disruption to the Snrpn cluster can lead to 2 different imprinted disorders, Angelman Syndrome and Prader–Willi Syndrome and this depends on which parental allele is disrupted. So lets start with the Kcnq1 cluster. So here's a picture of the Kcnq1 cluster. I've actually simplified it a little bit although obviously theres a lot of detail there. I'm going to try and use the same general notation for each of the imprinted clusters we talk about. So, I'll take some time to talk through these now. I'm showing the maternal allele, and the paternal allele, because we know each of those have differential expressions, so they're different from one-another. Genes that are expressed from the maternal allele are shown in pink Genes that are expressed from the paternal allele is shown in blue. And the imprint control region is shown as an oval. So if it's methylated then it will be coloured in. Like I've showed you for CPG dinucleotides when we think about their methylation. And if it's unmethylated, it would be an open oval. Genes that aren't expressed from that parental allele are shown in grey. So if a gene is imprinted in its expression, it'll be pink on one chromosome and grey on the other or blue on one chromosome and grey on the other. The genes that are not coloured in at all and instead are shown in white, are genes that still show biallelic expression. In other words, they're expressed from both parental alleles. They're not imprinted, but they happen to be found within this imprinted cluster. So obviously their expression is not being controlled by the imprint control region. I'm also showing long noncoding RNAs as these transcripts like this, with the arrows that are outside of the genes. And this will be the same notation that I'll use. And although I'm not tending to name most of the genes in the cluster, I'll just name the occasional gene that's useful to hear about. For example here, Cdkn1c or Kcnq1. I'm also not showing the directionality in which the genes are expressed. Remember they could be expressed from either strand. So, either in this direction. Or in this direction. I'm not going to these details because usually not important for the control of the region but where it is important to know which direction they're transcribed in I'm showing you internal to the gene itself. Say for example the Kcnq1 which is transcribed in the right to left orientation so for the reserves strand of DNA. So in this case if we think about the Kcnq1 cluster, we know that there's the imprint control region which sits internal to the Kcnq1 gene. When this Kcnq1 imprint control cluster is unmethylated, as it is in the paternal allele, it results in expression of this long noncoding RNA called Kcnq1ot1, which stands for the Kcnq1 overlying transcript 1. Now interestingly, this overlying transcript is transcribed antisense to the Kcnq1 gene, which is transcribed on the negative strand, and this one is transcribed in this direction. So when this Kcnq1 imprint control region is methylated as it is in the maternal chromosome, then this results in silencing of that Kcnq1ot1, overlying transcript 1. So this long noncoding RNA. And then we end up with expression of most genes within this region as found in the maternal allele. However, on the pattern allele, when the long non coding RNA is expressed. We know that most of the genes, then, in this region, or all of the genes that are imprinted are then silenced. So the way that this happens is actually a little bit similar to the way that it happens with Xist. So, the Kcnq1ot1 long non coding RNA binds to G9A, which methylates H3K9. So it's an H3K9 methyltransferase and to PRC2, just like Xist does, which methylates H3K27. So both of these are repressive histone marks. And Kcnq1ot1, then leads to spreading of heterochromatin throughout the region, and silencing of expression of the genes. So Kcnq1ot1's mechanism of action is somewhat similar to that for Xist. And so you’ll notice if we just think about methylation of the imprint control region, that its unmethylated state is actually associated with silencing of all of these genes, but its activation of this important long noncoding RNA. So I'd just like you to point out this single gene that I'd like you to remember in this cluster. The Cdkn1c gene. This gene, which is just a couple of genes away from this imprint control region, is maternally expressed. So it's shown in pink. But the function of this gene is to restrict growth, because it's a cyclin-dependent kinase inhibitor. So, normally it's involved in restricting the growth of a cell and inhibiting the growth, and so it's often called a tumour suppressor. So now let's think about the closely linked cluster which is the H19/Igf2 cluster of imprinted genes. So here we have again the maternal chromosome and the paternal chromosome. We have the imprinted control region here. in the middle, and we know that it's unmethylated on the maternal allele, and methylated on the paternal allele, and this means it's paternally imprinted, seeing as the DNA methylation imprint is on the paternal chromosome. This operates through a very different mechanism than the Kcnq1 cluster. So here we know that there is a long noncoding RNA that's being produced and this is called H19. It's being produced just from the maternal allele, but this long noncoding RNA has nothing to do with the establishment of imprinting at this particular cluster. Rather H19 is actually has a role in being a reservoir for a particular micro RNA. So it's role as a long non coding RNA is not bind epi genetic modifier as Kcnq1ot1 or Xist role was. So when this imprint control region is unmethylated as a maternal allele. It is bound by a protein called CTCF, and CTCF is an insulator protein. What that means is in this case its able to insulate. CTCF's binding insulate Igf2 from the downstream enhances that here are shown as green stars. Because it's insulating Igf2 these enhancers are now free to act on H19 and enhance H19 expression on the maternal allele so this is thought to happen because of chromatin looping. So whats thought is the preferred loop is between the enhancers and IGF two. However, when CTCF blocks this by it's binding, now the secondary preference is occurring. And the enhancers loop to H19, to enhance H19 expression. On the paternal allele when this imprint control region is methlyated CTCF can now no longer bind. And if CTCF isn't binding and there is no insulator action then these enhancers are indeed free to act on Igf2 and promote Igf2's expression from the paternal allele. And therefore, Igf2 is expressed only from the paternal allele and not the maternal allele. So this is how we end up with imprinted expression of Igf2. But the reason that H19 is now no longer active on the paternal allele is because of heterochromatin spreading, or DNA methylation spreading, in this case, which I mentioned last week. So this DNA methylation that's found at the imprint control region can spread downstream into the H19 promoter. And when it becomes methylated, it is silenced because it's a CPG island promoter. And this, in this case, this methylation is indeed synonymous with gene silencing. So you can see that the way that this imprint control region works by enhancer blocking is very different from the way that the one works for the Kcnq1 cluster. So these two clusters, which are closely linked, can be disrupted in a disorder known as Beckwith Wiedemann syndrome. So Beckwith Wiedemann syndrome is an imprinted disorder. And it can result from a large number of different abnormalities that are found in this particular region of chromosome 11. Chromosome 11q15.5 and normally its by some abrogation to this whole region which is about one megabase in size. So these disruptions lead to problems with both Kcnq1 cluster expression and the Igf2/H19 cluster expression. But perhaps the most common in the ones that I'll explain, are disruptions to Cdkn1c expression and disruption to Igf2 expression. So I mentioned Cdkn1c as a tumour suppressor, it restricts growth. However, what Igf2 does is it promotes growth, it's an oncogene. So in this very simplified version at the bottom of these two clusters and their linkage, what happens is that you have effectively appearance of two paternal like alleles and loss of the maternal like allele. So this then means if you lose this allele and you end up having two alleles that look like this one then you get too much Igf2. And you'll get no Cdkn1c expression. So that means you're not growth promoting a lot but you're not inhibiting growth at all at least based on these two genes. So, how does this happen then? Well, it can happen because of a number of different abnormalities. There can be mutations or deletions within this region. And they call it loss of imprinting, and so you'll have both alleles behaving like the paternal allele. You can have something that's known as uniparental disomy. Uniparental disomy means that chromosome, in this case chromosome 11 is inherited in two copies from the father so that would be uniparental paternal disomy, as is shown for Beckwith Wiedemann syndrome. Although uniparental disomy could be also from maternal chromosome but that's not associated with this syndrome. So you get two copies of the chromosome from one of your parents and none from the other. And finally the way that this can happen is because of epigenetic disruption, the underlying DNA can have no particular mutations or translocations or disruption, but loss of imprinting can be because of disruption those DNA methylation imprints. And this actually very very rare. So if you have no Cdkn1c expression and up regulation of Igf2 in other words in it's overall you're growth promoting, what's then is the phenotype of these patients? Well, as you might expect they, they display both foetal, and post-natal overgrowth. And this is related to the expression of these two genes although others as well. They also display macroglossia, or in other words a large tongue. That's something that's really important to remember is that it have predisposed to embryonic or childhood tumours but interestingly not to adult tumours. So the most common one that's discussed in this regard is something known as Wilms tumour which is found in the kidney. So this loss of imprinting and over-expression of growth promoting genes or loss of those tumour suppress genes because of the loss of imprinting is actually a common feature not just of Beckwith Wiedemann syndrome patients. But in fact in general in cancer and so many imprinted genes are involved in controlling growth. And loss of imprinting then is actually a hallmark of cancer and we'll come back to this in week six when we talk about cancer epigenetics. So the final thing I'd like to talk about with regard Beckwith Wiedemann is how it looks like in terms of it's inheritance so in this case we're going to look at a pedigree. So for those of you who haven't seen many pedigree's before the squares are males, the circles are females and the carriers have a dot in the middle as shown here. The affected or in other words the patients are here filled with red, but the people that are wild types are normal and also they are not carriers, are open and white. The most important thing to remember for Beckwith Wiedemann
in terms of its inheritance is that it needs to be transmitted from the mother so it requires maternal transmission. From a carrier or a symptomatic mother to the offspring. So you can not have an affected offspring coming from the father. So I'm going to look at the pedigree. What I want you to take away from it is that this parent of origin specific. Inheritance rather than sex specific inheritance so first of all we have a carrier father and he produces 2 carrier children and 2 wild type children and this will be because only 1 of his 2 alleles carries the mutation. So half his children inherit it. But the father, a carrier father, either this top one or this one shown here, can never produce an affected offspring. Because we only see Beckwith Wiedemann syndrome following maternal transmission. Now that's because if you have a mutation which makes the paternal chromosome look like. A paternal chromosome then off-course there is no problem. But, it doesn't matter that we have the expression like the paternal chromosome. You should have expression like paternal chromosome. The problem comes when you should have expression from the maternal chromosome, but you have a mutation that makes it appear like a paternal chromosome. So, in this case we have a carrier in mother. And she's' produced two affected children. One of each gender, and two wild type children. Now again, only one of her chromosomes, only one of her chromosome 11's will have had that mutation, so the children that inherent the wild type copy will be fine, but the children that inherit the mutant copy now display the disorder and that's because they are unable to set up expression in a maternal specific way, and instead it looks like a paternal chromosome like the one that they inherit from their dad so you have a loss of Cdkn1c and overdose of Igf2. Similar to the maternal carrier, a maternal patient will produce offspring at that 50% that are affected and 50% that are wild type. Whereas again, if we have a male patient, a father that is affected, he will only produce carrier children. Okay, so these hallmarks are that you have transmission from the carrier or an affected mother to her children. But fathers, even if they're affected, only produce carrier children. And this hall mark of Beckwith Wiedemann syndrome and other imprinted disorders that are maternally transmitted. So in summary of this lecture then we know imprint control regions have different modes of action and this controls the parent of origin specific monoallelic gene expression. We know that loss of DNA methylation at imprint control regions is common in cancer and this is a point we'll come back to in later weeks. And finally, that aberrant expression of imprinted genes can result in specific diseases. And we just went through Beckwith Wiedemann syndrome. In the next lecture, we'll think about the separate cluster. The Snrpn cluster, and its two imprinted disorders that relate from its disruption.
