Okay, so now we're going to consider the other epigenetic aberrations that are seen in cancer. We've spoken about DNA methylation, but what do we know about the aberrations that are found in terms of histone modifications, nuclear architecture, and noncoding RNAs? So, in terms of histone modifications, just like we know for DNA methylation, we know there are global alterations. We also know that there's some relationship between the histone modifications and DNA repair or genomic stability. So some histone modifications or some particular histone variants are found particularly at the centromeres. And centromeres are essential when a cell divides to make sure you get the right number of chromosomes at each pole and therefore the right number of chromosomes in each daughter cell, because this is the point of attachment to the mitotic spindle. So if these histone modifications or the histone variants that are involved with centromere function go wrong in cancer, then this is associated with increased genomic instability. So there's some relationship between these epigenetic modifications, the histone modifications that go awry, and genomic stability just like there was for DNA methylation. But we also know that histone modifications have some relationship to DNA repair. So this is because the way that DNA repair works in regions of heterochromatin is different to the way that it works of regions of euchromatin. So if you take this image here for example and this is an image you've seen before. We're looking at the histone variant called gamma H2AX in green. And gamma H2AX is just the phosphorylated version of H2AX, which is a variant of histone H2A you'll remember. And this is associated with regions that have double-strand breaks in the DNA in this case, post irradiation, which is an insult that's going to induce DNA damage. We also have the DNA which is counter stained in blue and what you'll be able to see is that the green is not really found in the very densely blue sections. In other words, the green which is marking these regions where gamma H2AX is bound is not really marking the densely heterochromatic regions. This is probably not because there isn't any DNA damage in the densely heterochromatic regions but rather it's not accessible to have Gamma H2AX binding. And so, it's been relatively recently that we've found out that this is because DNA repair occurs via a different mechanism in regions of heterochromatin. So, in regions that are densely heterochromatic, you have to first of all unwind the chromatin. Then you need to make sure you inhibit transcription in these regions that would normally be silent during its repair process, but now they're euchromatic. The repair enzymes have to go in and access the DNA and perform the repair. And then the region has to be re-compacted again because it was originally heterochromatic. Because this repair happens differently to how is does in euchromatin, first of all it's slower and it seems to be less efficient. In the context of cancer, we have many regions that didn't used to be heterochromatic, and now are heterochromatic, and their repair doesn't occur as efficiently. You can map where the mistakes occur, where the mutations occur commonly. And they seem to occur in these regions of heterochromatin in cancer. So there's some relationship, like there was with DNA methylation, between repair, or genomic instability and these epigenetic mistakes. So as I said, there are these global alterations and histone modifications, just like there were global alterations in DNA methylation. So it's been said that it's probably the repeats that are being analysed in this case, although it's likely to be also genome-wide. Now, this is a little bit of an odd histone picture. I've shown you histones And nucleosomes shown like this before. But what they're saying in this case that in the context of cancer, you have decreased, histone H4K16 acetylation. So acetylation of lysine residue 16 on histone H4. And this is normally an active mark. But you also have decreased methylation of histone 4 at lysine 20. And this is normally an inhibitory mark. So you have this consequence, genome-wide, in the repeats, which we know become active. You have both the loss of an active mark, but also the loss of a repressive mark. But somehow these changes seem to be associated with an overall activation. It's not yet clear why, when you lose one active and one silent, but also probably lose methylation so you're losing that inhibitory mark. It will necessarily result in activation of these repeats. And, but, in general these are some of the features that you see of histone modifications genome wide in cancer cells. If we think about those CPG islands that are hypermethylated in terms of DNA methylation. We know there's some collaboration between the histone modification changes that are seen here and, in fact, these DNA methylation changes. So, in a normal cell, if the gene was being expressed you would have H3K4 methylation, which is shown here, these little blue triplet of circles. But you also have heavier acetylation of histone residues found in the tails of histone H3 and H4 shown here as the green triangles. In the context of cancer, we know that these particular sites, the sites that were just looking at that are hypermethylated at the CpG islands, at the DNA level. Then the nucleosomes, the histones, making up the nucleosomes that are in these same regions, now have found, you find that this decreased acetylation, decreased methylation, decreased methylation of H3K4 which is normally active. But increased methylation of H3K9 and H3K27, these two inactive marks. So this seems to make sense that you've got an accumulation of epigenetically silencing marks. And so you have DNA methylation along with these two inactive marks H3K9 and H3K27. What's quite interesting but we don't really understand yet, is that the genes that you find this hypermethylation along with, targeting by H3K9 and H3K27, were actually targets of, H3K27 methylation because of the complex PRC2 in embryonic stem cells. So the question is being raised, do we, at some earlier point in development, become primed for silencing because of what happens at this extremely early embryonic stage. And a jury is still out but, at the moment, we have this correlation which is fairly tantalising that there may be some establishment of a priming signal way back in embryonic development. So just like with DNA methylation, we know that aberrations in histone modifications can occur genome-wide. And just like with DNA methylation aberrations, we know that they can be tumour-type specific alterations. I'm not going to go through all of these because actually we don't know as much as we did about DNA methylation. But just like those DNA methylation changes those tumour-specific DNA methylation changes can be used for diagnosis or prognosis. Similarly the tumour-type specific histone modifications can be as well. However histone modifications are not as simple to assess in the lab and so they're not as useful, quite as useful as biomarkers. So I wanted to briefly mention histone variants. This is extremely recent data that’s come out in the literature. So it's always been thought that perhaps you wouldn't have mutations in histones particularly commonly found or maybe they may not might not have a functional consequence. And this is thought because you have a lot of copies of the histone genes, you don't just have one version of histone H3. You have many copies so, what would the affect be of just mutating one of them? Well, very recent evidence has shown that you would indeed find mutations in histones, core histones, and histone variants. In this case, only in childhood high grade glioma. You don't see it in adult glioma, this adult type of brain cancer. And not if that tumour is low grade. In other words, it's not as aggressive. So these mutations that are found are found in the canonical H3, in other words the normal histone H3 that's found in nucleosomes. And this is called H3.1 and also in histone variant H3.3. Interestingly, these mutations appear to affect whether or not histone will become post translationally modified. So, whether it will end up having a histone mark. So one of them is a direct effect. You can have lysine 27 on histone H3 or 3.3 is mutated to methianine. Methianine's not longer able to be acetylated or methlyated, and so you won't have that histone modification happening. It's possible though that this methianine actually could mimic methylation. It might look a little bit like methylation to the cell, we don't yet know. But this certainly chemically is a possibility. In terms of H3.1 you also find a mutation at residue 34. Now 34 isn't a residue that's normally post-translationally modified. But, K36 is, and appears by mutating 34, it alters the ability to methylate downstream at residue number 36. So by mutating residue 34, you can influence what's happening to residues downstream. So while these mutations have only just been reported and we don't understand everything about them. It suggests that not only will you have alterations to the histone modifications that are found in cancer. But alternations to the underlying histone genes, meaning that those modifications may or may not be able to be laid down. I just want to make one brief mention at the end of this lecture on histone modifications about the contribution of the environment to these altered histone modifications. We spent much of week five thinking about how the environment can influence our epigenetic makeup. So in this case, let's think about carcinogens. So, carcinogens are defined as something that will actually alter or increase the incidence of cancer. But carcinogens can either be mutagenic or not. So, they'll be non-mutagenic. In the case of mutagens like tobacco smoke, for example, we know that how they work. They cause an increased incidence of mutations and therefore you're more likely to get cancer. But how do carcinogens work that don't cause mutations? How do they act? Well there are some non-mutagenic, carcinogenic heavy metals. Okay, so there are heavy metals that are found in the environment for example. And these appear to alter the activity of histone modifier enzymes. In other words, those enzymes that lay down histone methylation, histone acetalation - pr remove each of these. And so its possible that these non, mutagenic, but carcinogenic heavy metals may have an action epigenetically because they'll be altering our ability to lay down these particular histone modification marks. So it might be that the environment can alter our epigenotype and therefore promote tumourigenesis. Now we also spent a lot of time in week five thinking about sensitive periods. Well, this is in an adult, this isn't exactly a sensitive period. But what you've got to remember is to have a consequence for cancer, perhaps just one cell needs to be affected. So while most cells may be able to very well maintain their particular epigenetic state. If even one cell is adversely affected and you then have downstream consequences and additional hits like the Knudsen hypothesis. This can result in a tumour and so perhaps it's not really defined as a sensitive period, but you don't need to be very sensitive for one cell to react. So in the next stage you'll start to think about gene by gene epigenetic alteration and how they can spread spread out to much larger genetic intervals. And so have large megabase regions that are all epigentically abnormal in comparison to normal cells.
