Listen in as Dr Ballestar discusses an overall perspective on the connections between DNA methylation and other epigenetic marks, as well as the interplay with transcription factors.
Esteban Ballestar obtained his PhD from the University of Valencia under the supervision of Luis Franco, specializing in chromatin and histone modifications. He then joined the group of Alan Wolffe at the National Institutes of Health (NIH) as a postdoctoral fellow, where he investigated associations between nuclear factors such as methyl-CpG binding proteins and methylated DNA.
From 2001 to 2008, Esteban Ballestar worked as a senior scientist at the Spanish National Cancer Research Center (CNIO), in association with Manel Esteller, working on epigenetic alterations in human cancer.
Esteban is currently a senior group leader at the Bellvitge Institute for Biomedical Research (PEBC-IDIBELL) in Barcelona working on epigenetic deregulation in the context of autoimmune diseases, focusing on DNA methylation changes and various differentiation models relevant to this group of disorders. Esteban is recipient of various national and international grants, has authored more than 90 publications, and has organized several international conferences on Epigenetics.
Hello, welcome to Abcam's webinar on DNA Methylation changes during cell differentiation. Today's principal speaker is Esteban Ballestar, a Senior Group Leader at the Bellvitge Institute for Biomedical Research in Barcelona. At the moment, he is working on epigenetic deregulation in the context of autoimmune diseases, focusing on DNA methylation changes and the various differentiation models relevant to this group of disorders. Esteban obtained his PhD from the University of Valencia. He then worked as a postdoctoral fellow in the laboratory of Alan Wolffe at the National Institute of Health, where he investigated associations between elements of the chromatin machinery and methylated DNA. From 2001 to 2008, Esteban worked as a senior scientist at the Spanish National Cancer Research Centre, in association with Manel Esteller, on epigenetic alterations in human cancer.
Joining Esteban today will be Miriam Ferrer, Project Manager for cellular assays at Abcam. Miriam completed her biology degree at the University of Barcelona, and has a PhD from the Vrije University in Amsterdam. After completing her PhD she joined the MRC Laboratory of Molecular Biology in Cambridge. I will now hand over to Esteban who will start this webinar.
EB: Thank you, Vicky and good morning, good afternoon, good evening to everyone who's attending; thank you for being here. I'm going to discuss different topics during my conference and the conference is going to be talking on different aspects on DNA methylation, genomic distribution and functional roles. I will also be discussing different mechanisms of DNA demethylation, particularly active DNA demethylation. I will also talk about the changes in DNA methylation in the context of differentiation processes. Finally, I will also talk about different targeting mechanisms of DNA methylation changes.
As you probably know, DNA methylation is the best studied epigenetic modification. This occurs by the introduction of methyl group in the 5' position of cytosines. In mammalians it's mainly in the context of CG or CpG sites. CpG methylation has different effects on gene transcription, depending on the location. I will discuss a little bit more on that topic later. This mechanism is used generally for genes and the inactive X-chromosome imprinted genes, and several tissue-specific genes also use DNA methylation.
This epigenetic modification is introduced by DNA methyltransferases, and regarding DNA demethylation this has been a controversial topic for many years. Not so much for passive demethylation, which is due to the inefficient maintenance of DNA methylation; and this is the basis of compounds, DNA methyltransferase inhibitors that lead to demethylation and this is just for therapeutic purposes. Regarding active DNA demethylation, which I will discuss in more detail later, the most accepted mechanism involves deoxidization of 5-methylcytosine into different oxidized forms: hydroxymethylcytosine, 5-formylcytosine and 5-carboxylcytosine.
When talking about CpG methylation or DNA methylation, it's important to talk about CpG islands. CpG islands are represented in about half of the genes, invertebrate genomes and these are short, approximately 1 kb length stretches of DNA with a high density of CpG sites. So CpG islands have been a major topic in the context of DNA methylation studies and they are, in general, demethylated. We know now that also they are adjacent regions to CpG islands, like CpG island shores, CpG island shelves that are upstream and downstream to CpG islands can also have an impact on gene expression.
What do we know about DNA methylation in mammals? We know that most CpG islands are not methylated, at least under physiological conditions, when located at the transcription start sites of genes. We also know that CpG island methylation of the transcription start sites is associated with long-term silencing, particularly in the genes that I mentioned before, like X-chromosome deactivation genes, imprinted genes; all genes expressed predominantly in germ cells and some tissue-specific genes. We also know that CpG islands in gene bodies can be sometimes methylated in a tissue-specific manner. Also, that non-CpG island methylation, in other words, methylation that occurs in CpG sites that are not in a CpG island, it's more dynamic and more tissue-specific than CpG island methylation. Finally, we know that methylation blocks the start of transcription, but no elongation.
So there is different roles of DNA methylation, depending on the genomic location. This is just a few examples to illustrate these potential different roles. We know that the transcriptional machinery can prevent or protect certain active regions from becoming methylated by DNA methyltransferases. We also know that DNA methylation can actively silence promoters. Also, that methylation is used to maintain pericentromeric regions to silence repeat, and also in maintaining imprinted genes and transcriptional starters.
One of the main or most interesting topics in the context of the DNA methylation field, are the pathways to active DNA demethylation. Active DNA demethylation occurs in the absence of DNA replication, in the absence of cell division; and is relevant for many reasons. First, it has been very hard to identify real demethylase: an activity that actively removes in a single step methyl groups from cytosines. Also, constitutes an interesting topic, given their potential applications for therapies, for instance. We know that active DNA demethylation involves several steps, and that glycosylases and excision repair machinery is involved. It is less clear the steps - the initial steps of this process. One of the models proposes that activation-induced deaminase, the amino group next to the 5' position of the cytosine. So these four, which is thymine can be - then this modified base can be removed by the glycosylase and then this is going to be repaired by the base excision repair machinery. An alternative model proposes that the oxygenases oxidized the methyl group into different forms, which I will describe in the next slide. These oxidized forms of these bases are then removed by glycosylases, and then that a basic site is repaired by the base excision repair machinery.
So going into more detail into the TET mechanism, which is the most accepted currently. We know that, as I said before, that DNA methyltransferase is catalyzed the addition of the methyl group in a cytosine. So we have here the methylated cytosine, so TET proteins can oxidize in subsequent steps. This methylcytosine into 5-hydroxymethylcytosines, 5-formylcytosine or 5-carboxylcytosine, and then these oxidized forms can be removed by the thymine DNA glycosylase, which generate this basic site that can be filled with unmethylated cytosine. So at least DNA demethylation involves two steps, and at least three enzymatic activities: the proteins, thymine DNA glycosylase and the base excision repair machinery.
One of the most interesting models to investigate DNA methylation changes, their role and also their interplay with different factors involved in transcription factors, is the hematopoietic system. The hematopoietic system has been studied for many years, we know that hematopoietic stem cells during differentiation are going to eventually differentiate into progenitors. They come only from progenitors and they come on myeloid progenitors that will eventually differentiate in all the differentiated cells in the hematopoietic system. We know for years that different transcription factors participate in different stages during differentiation, and this has been very well-characterized in different molecular biology studies. Also, a lot of this information comes from the hematological malignancies, and the specific transcription factors participating specific stages of differentiation; and also, they are aspecific to the branch of differentiation. For instance, the transcription factor CEBP alpha is a specific of the myeloid branch, and factors like PAX5 or EBF1 are specifically lymphoid branch, particularly the B cell differentiation process. We have very detailed information on the participation of transcription factors during all these steps in hematopoietic differentiation.
So we now have a lot of information as well on DNA methylation profile in different steps of differentiation during hematopoietic differentiation, in both the lymphoid and the myeloid branch. So we have detailed profiles of methylation from different cell stages, different branches. We have here just a few examples to illustrate the detail of information that we have at this level. For instance, if we look at the promoter sequences of B-lymphocyte-specific genes like CD19, or myeloid-specific genes, particularly when encoding for the CEBP alpha transcription factor. So we have detailed maps of methylation at the base resolution, and we can see that these are specific to the cell type. We can look here, for instance, at the CD19 promoter, we can see that in B cells where this gene is active, you can see lower methylation levels, for instance, than if we compare these profiles in neutrophils that belong to the myeloid branch. Regarding the CEBP alpha gene, we can see here that the levels of methylation for this gene, which is myeloid-specific, are lower in neutrophils that are in the myeloid branch than when we compare the levels in B cells.
Another question that has become clear during the past years is that transcription factors and epigenetic enzymes are interconnected. Transcription factors are able to recruit different epigenetic enzymes, highlighting the role of these transcription factors to direct epigenetic changes, or epigenetic profiles to specific target sequences for those transcription factors. These epigenetic enzymes include DNA methyltransferases, or also the TET enzymes that are involved in active DNA demethylation.
What is the effect of site-specific factors like these transcription factors on DNA methylation? Well, there are different possibilities: we know that transcription factors like SP1 protect CpG islands from becoming methylated by DNA methyltransferases. We know that other factors are, on the contrary, able to recruit DNA methyltransferases and, therefore, direct DNA methylation to a specific site. We also have factors that are able to recruit TET proteins and, therefore, target active DNA demethylation. Also, in other cases, transcription factors can recruit other histone modifying enzymes that are going to have an indirect effect on the DNA methylation profiles.
What are the models that I'm going to discuss about today? We have been using different models in our lab; I'm going to discuss different models to illustrate that close interplay between transcription factors and the setting of DNA methylation changes. The first one is going to be a model, and it's not a real differentiation model, it's more of a model where we see the transformation of resting B cells into proliferative ones; and with the potential participation of activation-induced deaminase. The second model is an engineered differentiation model where it's possible to dissect the specific role of that given transcription factor, and its impact in the acquisition of DNA methylation changes. Finally, I will present an example of one of the complex differentiation systems that we're using in the lab, to see how different transcription factors can participate in the setting of DNA methylation changes.
The first model that I'm going to discuss is this model where resting B cells can be transferred with the Epstein-Barr virus, and this is going to lead to proliferation. So this is a classical model to generate lymphoblastoid cell lines in the labs, from valuable blood samples from patients. But these are approaches that actually takes place in humans, so 99% of the human population is actually infected by Epstein-Barr virus. This process in the primary infection and leads to a continuous B cell proliferation, and it has been associated with different disorders. Epstein-Barr virus is associated with different types of lymphoma with a particular impact in Burkitt lymphoma, and also in diffuse large-cell lymphoma; and also, it's associated with different autoimmune diseases.
So in this model: this model has been studied for many years, we perform high throughput screening of DNA methylation changes; and we compared the profiles of resting B cells versus the cells that are derived from them following this infection with the Epstein-Barr virus. When we analyzed DNA methylation changes we observed that a large majority of the changes occur in the direction of demethylation. We here have a selection of the genes with a fold-change larger than two, and around 250 genes underwent demethylation in this process. So that provides a very interesting model to investigate potential mechanisms of demethylation.
One of the possibilities is that activation-induced deaminase could have been involved in this process. We know that this enzyme, AID, is highly overexpressed following Epstein-Barr virus infection. I'm going to discuss this shortly, because we would involve a few slides, but basically what we saw is that in the topic introduction of AID in these cells didn't lead to demethylation. Also, that when we explored changes in 5-hydroxymethylcytosine trying to explore the alternative pathway to active demethylation, we couldn't see any changes associated with the transformation of resting B cells into proliferative ones. Also, just to illustrate some of the things that we perform in this project, is that demethylation changes occurs only when cells start to proliferate, suggesting that this demethylation is actually coupled with the acquisition of B cells to divide, to the DNA to replicate, so it would be more of a passive mechanism.
However, we explored the potential connection of transcription factors with the set of genes that became demethylated, as many of the genes that we identified were clearly genes relevant for B cell biology. What we observed when we performed transfac analysis of the genes that become demethylated, is a high enrichment for the transcription factors that are relevant to B cells. Also, for factors that are actually relevant for this transformation process, including subunits of the NFkB transcription factors and other factors that are already around for B cell biology.
We, for instance, focusing on the NFkB, we observed that around half of the genes that become demethylated are targets for NFkB, and this is a comparison of the list of genes that become demethylated and ChIP-seq data for NFkB. Also, we observed that close to half of the genes were genes that were actually bound by Pol II. In other words, genes that are actively transcribed in these cells. We can look at a detail of some of the genes where we can see that the CpG that becomes demethylated coincides very nicely with NFkB, ChIP-seq peaks, or Pol II peaks. You can see here or here, or here; this is just a few examples.
We also observed that from comparison of this data with our demethylation data, with ChIP-seq data for different transcription factors, we observed a very nice overlap between the genes that have become demethylated and targets for these transcription factors.
So to just simplify the story, basically, although this model represents a nice model of passive demethylation, even under these conditions we see a coupling between the genes that become demethylated and transcription factors. In this case, probably because transcription factors that are bound to these classically active sites are preventing the position of DNA methylation in these sites by DNA methyltransferases.
I'm going to move to a second model; the second model that I mentioned earlier is an engineer model where we induced differentiation by overexpressing that transcription factor. We have several models of this: in the first model of reprogramming or transdifferentiation, was the overexpression of MyoD and fibroblasts, which is able to direct the differentiation of these fibroblasts into differentiated muscle cells. I think the most dramatic example of reprogramming models is the generation of induced pluripotent stem cells, where a cocktail of four transcription factors is able to reprogram fibroblasts into this IPS.
The model that I am going to discuss today is a model that was developed by Thomas Graf's laboratory, where he managed to differentiate or transdifferentiate B lymphocytes into macrophages by overexpressing the C/EBP alpha transcription factor. If you remember, C/EBP alpha is this myeloid-specific transcription factor, and this only transcription factor is able to orchestrate this differentiation across branches in the hematopoietic system, so B cells become functional macrophages.
This model has been optimized, and in collaboration with Thomas Graf's group we were able to prove that this process is highly efficient, and transdifferentiation of B cells into macrophages can be completed in two days. This process becomes irreversible after four or five days, and these reprogrammed macrophages that come from the overexpression of C/EBP alpha in B cells, have all the functional properties of a macrophage. They migrate, they have inflammatory response, they display phagocytic activity, as we can see here with this staining of bacteria that they are able to phagocyte.
This process involves large changes in gene expression, and we can see here genes that are silenced in B cells and following overexpression of C/EBP alpha, they are overexpressed in macrophages, and other genes that are active in B cells become silent.
One of the questions was whether C/EBP alpha, by its own means, was able to induce epigenetic changes and orchestrate changes, epigenetic changes in these cells to become also at the epigenetic level macrophages. An interesting thing from the DNA methylation point of view, was that transdifferentiation associates with an overexpression of TET2, one of these dehydroxygenases that oxidize 5-methylcytosine. We know by using chromatin immunoprecipitation assays, that C/EBP alpha binds and activates the promoter of TET2, as we can see here.
We also saw that TET2 has a relevance in the acquisition of myeloid identity, as we saw that downregulation of TET2 by using Short Hairpin RNAs, downregulation was able to impair the acquisition of expression for myeloid-specific genes, therefore, reinforcing the role of TET2 in this acquisition of myeloid identity.
Surprisingly, when we look at the content of 5-methylcytosine in myeloid-specific genes, we have an interesting finding. So if we look at, for instance, myeloid-specific genes like C/EBP alpha, we can see here the profile of methylation, the promoter near the transcription start site, high levels of methylation in the CpG island at this TSS. This is heavily methylated in pre-B cells, it is unmethylated in control macrophages and following this transdifferentiation process, these cells retain the methylation profile of a pre-B cell. You can see here a couple of examples.
When we look at lymphoid-specific genes like the promoters of the EBF1 or PAX5 genes, these are unmethylated in pre-B cells. These are heavily methylated in macrophages, but when we look at reprogrammed macrophages they still retain the methylation profile of pre-B cells. We were able to see not only in these examples, but also in all the myeloid genes that we looked at. I'm not going to show the data corresponding to a high throughput analysis, but we were unable to see changes in 5-methylcytosine in these reprogrammed macrophages.
This model gives us a perspective on the relevance of different components of the DNA methylation machinery. We see changes in TET2, we see changes in 5-hydroxymethylcytosine, but we don't see changes in 5-methylcytosine, so this gives us a perspective and also provides information about perhaps the inability of C/EBP alpha to complete all the changes that are necessary to these cells, to become epigenetically indistinguishable from these B cells, despite the completion of the functional role of these cells.
This is going to move us to our third model, so these are a third sort of group of models are monocyte-related differentiation processes that we are using in our lab for epigenetic studies. We are performing different differentiation processes starting from monocytes obtained, or isolated from peripheral blood from humans. So if we isolate monocytes from peripheral blood by using different cocktails of cytokines, we can differentiate them into dendritic cells, into macrophages that can be later activated with different factors. We can differentiate monocytes also into osteoclasts, and these processes are interesting for us to investigate the targeting and the mechanism of active demethylation; and I will explain to you why.
So, first, as I said before, monocytes can differentiate in vitro to macrophages, dendritic cells, osteoclasts, also myeloid-derived suppressor cells. This can be achieved by using different cocktails of cytokines. Why are these processes interesting to us? Well, these processes are relevant to some of the clinical manifestations of autoimmune or autoinflammatory diseases, including, for instance, the destruction of the joints in rheumatoid arthritis by osteoclasts, and in cancer.
So these are models that are relevant to some of the diseases that we're interested for. Also, these processes have been studied for many years and we have a lot of information on the sets of transcription factors that participate in these processes. Also, these are post-mitotic differentiation processes, so DNA methylation changes, particularly active DNA demethylation, if it occurs, can be only explained if it's not related to DNA replication; or, in other words, if it's an active process.
Also, we know that monocytes and other myeloid cell types, as we just saw in the previous model, express high levels of TET proteins and, therefore, provide useful models to explore targeting and active DNA demethylation mechanisms.
One of the first models that we've started to work with in the lab in this context is differentiation of monocytes to osteoclasts. This process can start from monocytes, but also with different myeloid precursors and is induced by inflammatory and inflammatory context, particularly the presence of RANK ligand and M-CSF. So this initiates the differentiation process into osteoclast precursors. We have information on the transcription factors that are involved in this process like P1, NFkB. Also, the osteoclast-specific transcription factor NFATc1, and this is going to lead eventually to the fusion of these cells into these multinuclear polycarions that are the osteoclasts, that are large proteins that are able to degrade bone. As I said before, this process is relevant, for instance, in the destruction of the joints in rheumatoid arthritis, or in bone metastases.
We are able to recapitulate this process very efficiently in the labs, so this is just an example where we have monocytes and matching osteoclasts. You can see here these mononuclear cells, and how osteoclasts display a large number of nuclei and a big vacuole. So this process is very efficient, we can see here the TRAP staining which demonstrates the activity of these cells. Also, deactivation of osteoclast-specific genes, like carboxyanhydrides, to cathepsin K and MP9 etc. So these genes become activated during this process, and other monocyte-specific genes become repressed.
So in this process that can be achieved at a very high yield in the lab, we explore DNA methylation changes. This is the result of a comparison of monocytes and matching osteoclasts where we see that a high number of genes became demethylated with several thousand genes. At the same time, a number of genes in the same range, several thousand genes become hypermethylated.
In particular, it was interesting to see the categories, the functional categories that were enriched in the demethylated set where we saw that many of the CpG sites that had become demethylated are occurring in genes that are relevant for osteoclast differentiation, the organization of the membrane ruffle of these cells, or in different processes that are relevant to this process.
These changes that we were able to explore with this high throughput technology, were also tested and confirmed by using bisulfite sequencing. This is the result for several genes like cathepsin K, or ACP5, etc., that are demethylated and these validations using bisulfite sequencing confirm the results obtained in the high throughput screening. Also, this monocyte-specific gene, which is hypermethylated from this high throughput screening was confirmed by bisulphite to be hypermethylated in osteoclasts.
So changes occur - changes in methylation occur in different regions, they occur in the body of the genes, they are occurring also in promoters, in intergenic regions, etcetera. One of the interesting things to see how this DNA methylation changes relate with expression changes. We cross DNA methylation with expression data, we see all kinds of relationships, we see genes that follow this canonical sort of rule, which is that genes that are hypermethylated are silenced and genes that are left methylated are overexpressed. But, as you can see here, we see all kinds of relationships.
We saw that the majority of the genes that display this inverse relationship was an inverse relationship between methylation and expression, occur at the first axon near the transcription start site. We’ll focus on that just to simplify the story. We saw that overexpression of genes that become demethylated occurs very rapidly, and it already occurs within the first five days. Genes that become silenced also display these changes very rapidly in the first five days of this three-week process.
Another interesting aspect is the direction of these changes, and the relationship between this direction and the expression changes dynamics. For instance, if we look at the genes that become demethylated, as we can see here, genes like ACP5, cathepsin K, etc. So these genes become demethylated very rapidly, as we can see here, and these changes in DNA methylation, this loss of methylation is concomitant to the acquisition of expression. So it occurs, more or less, at the same timescale, as you can see here. So loss of DNA methylation is concomitant to gain of expression.
Interestingly, and also in agreement with observations in other models by others, we see that genes that become more methylated, display this gain of methylation after the gene expression has been lost, so we see that these monocyte-specific genes that become hypermethylated, this gain of methylation occurs after gene expression is lost. This also provides a perspective on the different role of methylation, depending on the direction, gain or loss.
As I said before, this differentiation model is interesting, because it allows us to explore active DNA demethylations since this differentiation process, of course, in the absence of DNA replication. So there are several methods to measure 5-hydroxymethylcytosines, or other oxidized cytosine forms, and we used two methods here, you can just see here. For instance, when we look at the content of 5-hydroxymethylcytosine in genes that become demethylated, we were able to see changes, we were able to see an increase followed by a decrease in 5-hydroxymethylcytosine. In other cases, we just saw a decrease, and this has been confirmed by an alternative method, where we can see this transient increase in this 5-hydroxymethylcytosine, which occurs at the same time that unmethylated cytosines are increasing. This data suggests that changes in DNA demethylation are occurring through these active mechanisms that I described earlier involving oxidization by TET proteins.
So before going into the next things that we explore, we also investigated the potential relationship between genes that become less or more methylated and transcription factors. For that we used TRANSFAC enrichment analysis, and we looked at the binding sites of transcription factors that are associated with genes that become demethylated or more methylated. For genes that become demethylated, we observe by using this TRANSFAC analysis that they are enriched in sites for AP-1 binding sites, like this AP-1 family includes transcription factors like Jun or Fos. We also saw an enrichment for different subunits of the NFkB transcription factor, and we also saw enrichment, a little bit less, but we saw enrichment for PU.1.
For genes that become more methylated, we observed an enrichment for a different set of transcription factors. Well, not completely different, because we saw enrichment for PU.1 and other ETS family transcription factors, but we didn't observe an enrichment, for instance, for NFkB or AP-1. So we saw a clear difference between genes that become demethylated that are enriching these three binding motifs, and those that become hypermethylated are only enriched in PU.1 and other different sets of transcription factors. This is just the summary of this analysis, and we see hypermethylated AP-1 family, NFkB, PU.1, hypermethylated PU.1.
This TRANSFAC analysis fact gave us a very nice picture on the potential involvement of transcription factors in this process. These are transcription factors that had been previously described to be involved in monocyte or osteoclast differentiation, and by just looking at the enrichment of binding motifs in genes that become less or more methylated, we actually complete that same picture. So all these highlighted in blue are those that are involved in loss of methylation, and the ones that are highlighted in red are associated with genes that have become more methylated.
Our question was to see whether some of these transcription factors are actually involved in recruiting elements of the DNA methylation/demethylation machinery. So this is just the example of a couple of transcription factors, the p65 subunit of NFkB and PU.1. This is just an immunoprecipitation experiment where we see in this western blot, we see p65 and PU.1 at zero, two and four days after differentiation. So when we performed this immunoprecipitation experiment, we saw an enrichment of DNMT3b in association with PU.1, but not with p65. That was in agreement with the data regarding those genes that become more methylated. The surprise came when we saw that TET2, which is likely to be involved in demethylation, is also associated with PU.1 and not with p65. We also tested other transcription factors like some of the ones belonging to the AP-1 family, but in this study we were able to see exclusively association between these enzymes and PU.1. These are the results obtained by doing the complementary immunoprecipitation experiment, where we see that TET2 interacts with PU.1 and also DNMT3b.
These are just chromatin immunoprecipitation experiments where we were able to see that PU.1 associates with those genes that become demethylated, like these three examples here. We see an increase, in some cases, in others it seems to be - PU.1 seems to be already associated under the initial conditions. But we saw an increase in the recruitment of TET2 at two days, and this was concomitant with a decrease in DNMT3b for those genes that become demethylated.
So this is a ChIP-seq experiment where we saw that PU.1 seems to be widely associated with both genes that become demethylated, and more methylated in this differentiation process. So sort of suggesting general involvement of PU.1 in the targeting of these methylation changes. So our next step was to test whether the downregulation of PU.1 is able to impact changes in DNA methylation, expression and also the recruitment of DNMT3b and TET2. So these are siRNA experiments with PU.1 in this differentiation to osteoclasts, and at two, four and six days we can see here efficient downregulation of PU.1.
So when we tested DNA methylation changes following downregulation of PU.1, we saw that downregulation of PU.1 had an impact both in genes that become demethylated. We see here partial impairment of DNA demethylation for some examples here, and we can also see that PU.1 seems to be impairing downregulation of PU.1, impairing their position of methylation at genes that become more methylated. This occurs not only at the DNA methylation level, but this impact also occurs at the gene expression level, as we can see here: four genes have become less methylated and more methylated. These are just controls. This is also the same experiment where we tested downregulation of PU.1 and we can see that downregulation of PU.1, which is here, demonstrated also at its binding at hypomethylated and hypermethylated genes, PU.1. So we can see that the downregulation of PU.1 has also an impact in the recruitment of TET2, or DNMT3b in hypomethylated genes and hypermethylated genes. So this gives us to our temporary model where we see that PU.1 is able to recruit both TET2 and DNMT3b in genes that become less methylated, and in genes that become more methylated.
This is taking us to the last slide where I'm going to just go through the general conclusions. I indicate the roles of DNA methylation and this is a key epigenetic mechanism that drives and stabilizes cell-fate positions. It is highly coupled with transcription factor mediated regulation. We know now that CpG methylation has different effects on gene transcription, depending on the location. But the direction of changes in DNA methylation gain or loss, is related in a different way with gene expression changes. Also, we know that the analysis of DNA methylation changes in a differentiation process identified genes, but also pathways and mechanisms that can be relevant to modulate these processes. DNA methylation changes in monocyte-related differentiation processes, recapitulate a set of transcription factors involved in these processes; and helps to determine novel targets to modulate their efficiency. One of the interesting characteristics of these models, is that we can explore active mechanisms involved in TET2, and we have identified PU.1 as a transcription factor targeting both active DNA demethylation and de novo deposition of DNA methylation. Finally, we are - just to give a hint of what we're doing in the lab, we are doing ongoing studies where we are trying to dissect the specific contribution of different transcription factors. Most importantly, upstream signaling pathways to DNA methylation changes in different groups of genes during this differentiation model.
This is just the last slide acknowledging all the people in my group that participate in these models. I want to just mention Lorenzo de la Rica who has been working on the osteoclast differentiation model. Javier Rodriguez who has been working on the transdifferentiation model, in collaboration with Thomas Graf. Also, another Henar Hernando who has been working in the Epstein-Barr virus model. Now, I will hand you over to Miriam and in five, ten minutes I will be back to answer your questions.
MF: Thank you, Esteban, for such an interesting and comprehensive presentation. I just want to remind you that you can keep submitting your questions to Esteban to the Q&A panel located at the right hand side of your screen.
Now, I would like to take this opportunity to talk to you about some Abcam products that might be interesting for your research. We have just launched a high-sensitivity ChIP kit that has been specifically designed to be used when there is a limited amount of sample available to perform ChIP, for example, when working with patient material, transgenic mice tissue or stem cells. With a high-sensitivity ChIP kit you can obtain enrichment of your sequence of interest by using as little as 2,000 cells, or half mg of tissue per reaction. You can get your results in only five hours, so the experiment can be done in one working day, and the eluted DNA can be used straightaway in sequencing, microarray or qPCR.
Our high-sensitivity ChIP kit is part of our range of high sensitivity products designed for when there is a limited amount of starting material. If you want to create a DNA library with a limited amount of DNA, our high-sensitivity DNA library preparation kit for Illumina sequencing allows you to create a DNA library from only 0.2 ng of DNA.
Our ChIP-seq high-sensitivity kit combines the benefit of the two kits I've just mentioned. This kit is designed to carry out a successful ChIP-seq starting directly from as little as 105 cells.
Bisulfite sequencing is the use of bisulfite treatment to determine the pattern of methylation. To help you to take your methylation pattern studies further, we now offer a couple of products to prepare post-bisulfite DNA libraries for Illumina sequencing. With our post-bisulfite DNA library preparation kit, ab185906, you can prepare a library using pre-treated DNA in only five hours. If you haven't treated your DNA, then we recommend our bisulphite high-sensitivity kit, which has all the reagents to perform the bisulfite conversion, followed by DNA library preparation steps. The library is modified and ready in only six hours.
We also have a range of MeDIP and hMeDIP kits for the testing of specific cytosine modifications. Immunoprecipitation of 5-methylcytosine or hydroxymethylcytosine residues can be ready in less than five hours.
To finish this section, I just want to highlight other kits we have available for studying DNA methylation. We have available kits for detecting DNMT activity; some of the enzymes that Esteban has mentioned during his talk. These are colorimetric ELISA-like assays that can detect the activity of a specific enzyme using only 0.5 µg of nuclear extracts in only four hours. Lastly, just to highlight one of our products to detect a lesser-known cytosine modification: 5-formylcytosine. This product can detect as low as 1 pg of 5-formylcytosine in less than five hours. Thank you very much for your attention. Without further delay, I will pass the microphone back to Esteban who is ready to answer some of your questions.
EB: Thank you, Miriam. I am going to read some of the questions that I have received from the attendees. For instance, I have this question that says: Do you know how the transcription factors associate with a methyl group and, if so, do you foresee a method to regulate methylation processes, and possibly change diseases? So my answer is, okay, we know that the methyl group can have an effect on, a direct effect on the binding of different factors. For instance, we know that the methyl group can allow the recruitment of methyl CpG binding domain proteins. Also, that can have a negative effect on the binding of certain transcription factors, particularly if the binding site has a CG dinucleotide in the motif. But regarding the generality of transcription factors, I don't think that the methyl group has a direct role in the association of transcription factors. Transcription factors that have their targets and most of them are insensitive to the methyl group, so I think it's an independent thing.
I am going to link that question with another one, so I have been asked also: Can we consider 5-hydroxymethylcytosine an epigenetic mark on its own merit? That's a very interesting question, and we have an example in one of the results that we had in collaboration with Thomas Graf. Where we see that hydroxymethylcytosine can have, or when we knocked out the two we are able to see an effect on gene expression, but, at the same time, in the same model we didn't see changes in the recovery of unmethylated sites. So that suggests that 5-hydroxymethylcytosine could have a direct effect on gene expression. In fact, in vitro studies have shown that methyl CpG binding proteins that I just mentioned early, have a reduced binding affinity for 5-hydroxymethylcytosine versus 5-methylcytosine.
Let me see. I have seen here that someone is asking me about the method that we used to detect 5-hydroxymethylcytosine. Well, I think we just had a very nice interaction by Miriam, and some of the methods that can be used to measure 5-hydroxymethylcytosines are the ones that Abcam is offering. But, I mean, you can also use currently some published methods to do it in a sort of bisulfite-modified version of the protocol. But most of the CS1 are just the kits that were just offered.
Regarding another question that I saw, which is the tool that we used to identify the transcription factor binding sites. I mentioned it during the presentation, but I can repeat it now, so it's TRANSFAC. I will send to the attendee the response, but it's T-R-A-N-S-F-A-C. So TRANSFAC is for us is the best tool to identify this enrichment in transcription factors binding motifs.
I think I'm just going to go to my last question, so I had someone asking me whether I could elaborate further my conclusions on the lack of DNA methylation changes during the CEBP alpha mediated transdifferentiation; and whether these results can be extrapolated to other reprogramming models?
Okay, so this result, I think it's very specific to that transdifferentiation model, so I think the conclusion is that in that particular model we did not see any changes that give us a perspective on the ability of CEBP alpha to not recruit some elements of the DNA methylation machinery; but that cannot be generalized to all transdifferentiation reprograming models. I think it just provides a perspective on the different hierarchy between different epigenetic marks, but I think it's not a general, not necessarily a general trend. So I think each transdifferentiation reprogramming model has to be studied independently. So that was my last question, and thank you everyone.
Thank you Esteban and Miriam for presenting today. Unfortunately, we were not able to answer all the questions received. For those whose questions were not answered, our scientific support team will contact you shortly with a response to your question. If you have any questions about what has been discussed in this webinar or have any technical enquiries, our scientific support team will be very happy to help you and they can be contacted at firstname.lastname@example.org. We hope you have found this webinar informative and useful to your work, and we look forward to welcoming you to another webinar in the future. Thank you again for attending, and good luck with your research!