All tags Epigenetics Epigenetic Editing: Permanently Modulate Gene Expression webinar

Epigenetic Editing: Permanently Modulate Gene Expression webinar

Epigenetic Editing (EGE), the local rewriting of epigenetic marks, opens the door for many innovative approaches in gene expression reprogramming

Dr Marianne Rots, of the University Medical Center Groningen, introduces the current approaches to epigenetic editing, and discusses how it can ultimately help correct epigenetics-associated diseases.  

Webinar Topics:

  • Engineered DNA binding domains
  • Gene-targeted expression modulation
  • Epigenetic Editing


About the Presenter:

Dr Marianne Rots is currently the Professor of Molecular Epigenetics at the University Medical Center Groningen. She completed her undergraduate studies at the Faculty of Biology of the University of Amsterdam in the Netherlands and obtained her Ph.D. from the University Medical Center VU, Amsterdam in 2000.

Marianne was a postdoctoral fellow at the Gene Therapy Center, University of Birmingham in Alabama USA before moving to Groningen, the Netherlands. Since 2010 she has led a multidisciplinary team on understanding and exploiting epigenetic regulation mechanisms to further develop Epigenetic Editing as an exciting approach to permanently modulate a gene’s expression profile.

Webinar Transcript:

MR:    Thank you for inviting me to present this webinar on Epigenetic Editing. Epigenetic editing refers to the technology of actively overwriting epigenetic marks at any given genomic locus, and, as such, the technology promises a unique approach to permanently reprogramme gene expression levels.​​

When I started this technology in Groningen, the Netherlands in 2007 the approach faced much disbelief. To re-express sleeping genes, accessibility of these genes is a prerequisite and these genes were thought to be located in dense heterochromatin, therefore, re-expressing sleeping genes was thought of to be unrealistic. On the other hand, epigenetic marks were and are still not generally accepted to actually control gene expression, and therefore this would prevent us from reaching our goal of muting screaming genes. 

In this webinar I will show you that these two dogmas do not hold true. Actually, we can make use of the dynamic nature of epigenetics. Epigenetics provides the cell a way to remember its cell-type identity throughout cell division. If we can make use of this reversible nature and change genome function in a mitotically stable, heritable way as a simple definition of epigenetics promises, then this will provide us with a very unique research tool.

Beyond that, it would also have clinical relevance as many diseases and clinical phenotypes have been associated with epigenetic abnormalities; and certain journals arose to cover these progresses, like Clinical Epigenetics. So epigenetic marks refer to modifications to the DNA molecule itself, or to proteins where the DNA is wrapped around the histones. Epigenetic marks are referred to, for example, as H3K9 trimethylation when referring to histone modifications. Dependent on the type of histone, like histone 3; dependent on the position and type of amino acid, like lysine at position 9; and, dependent on the chemical modification: in this example, trimethylation. 

Modifications on the DNA molecule generally mainly refer to methylation on the cytosine in the context of CpGs. This small chemical modification was thought to be a very stable epigenetic mark, however, we know now that there are several other intermediates, including hydroxymethylcytosine. 

The epigenetic landscape is being maintained by epigenetic enzymes called writers, which write epigenetic marks; or erasers, which can remove these epigenetic marks. We have increasing lists of these writers and erasers and this insight into basic chromatin biology is exploding in these recent years. However, it's kind of difficult to address cause versus consequence of epigenetic marks with respect to, for example, gene expression in differentiated cells; and also, order of events are still highly unknown. 

Despite this, clinical epigeneticists initiated large-scale epigenome-wide association studies, and these studies have yielded putative diagnostic markers. However, again, to identify biological function of these epigenetic mutations, it's technically challenging. The identification of many epigenetic mutations, however, sparked the interest of pharmaceutical companies, and many inhibitors have now been designed for these epigenetic writers or erasers, and even readers resulting in epigenetic therapies. Some actually got FDA-approved, but many, many more are being tested in clinical trials which are ongoing, and which are expected to be initiated in the nearby future. However, inhibiting enzymes which at genome-wide might result in unintended genome-wide effects like the up or down regulation of unintended genes, and also maybe it might result in genome instability.

Moreover, epigenetic enzymes do affect non-chromatin targets. This might also result in unintended effects. So I reasoned if we could override epigenetic marks at any given location, then we could start to provide more insights into mechanisms, including cause versus consequence, and also order of events. With the long lists of epigenetic mutations, we might start to actually functionally validate the biological functions of these epigenetic mutations. Maybe in the future, we might have a novel avenue to circumstance some of these advantages of the current epigenetic therapies. 

So the concept of epigenetic editing is shown on this slide, we take catalytic domains of writers or erasers and we fuse these to engineer to DNA-binding domains, zinc finger proteins or the CRISPR cross approach, which I will explain in a minute. This example shows you a very simple schematic concept slide, so for this gene which is moderately expressed, and which is controlled by a few repressors, epigenetic marks, we can start to modulate its expression by designing DNA-binding domains to target this gene.

If we would fuse a writer of repressive epigenetic marks, then expression of a fusion protein like this will write the repressive epigenetic marks and thereby preventing RNA molecules from being formed. We can also swap the writer and fuse an eraser to the same DNA-binding domain, and now repressive epigenetic marks can be erased allowing more RNA molecules to be formed. Of course, this can also be reversed for other writers and other erasers.

This animation shows you a little visualisation of the concepts. So we know that epigenetic mutations can cause diseases by encoding inactive proteins. Next to genetic mutations, epigenetic mutations are associated with disease by inducing abnormal levels of gene expression. Actually, most diseases are associated with abnormal gene expression levels. Excitingly, these epigenetic mutations, like genetic mutations, are being copied with every cell division. Epigenetic editing can mimic these epigenetic mutations by either adding epigenetic repressive marks or removing, or activating marks. In that way we can study the biological relevance of, for example, identified epigenetic mutations.

The tools for epigenetic editing are minimally a gene-targeting DNA-binding domain fused to an epigenetic effector domain. We have a long list of epigenetic effector domains which have been validated to, indeed, write or erase epigenetic marks, as we reviewed in 2012. The majority of these studies, however, made use of exogenous DNA-binding domains and, therefore, these require like reporter plasmids or integrated transgene constructs to actually determine the effect. Alternatively, people may choose human or mammalian DNA-binding domains like MeCP2 or NFkB, but these bind known, but multiple genes. Actually, at that time there was only one study which one research group, which made use of engineered DNA-binding domains, in this case, to target their favourite gene of interest, VEGF-A.

The platform these people used is based on zinc finger proteins, and zinc finger proteins are very abundant mammalian transcription factors which can be engineered to bind whatever stretch of DNA you would be interested in. Interestingly, these engineered zinc finger proteins have been tested in various clinical trials now, and do not show any adverse effects.

Zinc fingers consist of individual modules over fingers, and every individual finger consists of two beta sheets and an alpha helix. Within this alpha helix there are three amino acids, which predominantly decide to which triplet of DNA this particular finger is binding to. There is a lexicon which states which stretch of amino acids you need to incorporate into this alpha helix to encounter the triplet you are interested in. Of course, targeting one finger would only give you three base pairs, and that doesn't give any specificity. So we and many others, fused six fingers together to target 18 base pairs, which by mathematics should give a unique genomic address. 

So zinc finger proteins are based on mammalian transcription factors have been tested in clinical trials are quite small, and not immunogenic. But recently we have two other platforms of designer protein possibilities, were introduced: the TALEs derived from plant pathogens, and the CRISPRs which is based on the bacterial defense system. All these three platforms have been mainly exploited to either introduce genetic mutations, or to maybe even replace and correct genetic mutations. This is done by fusing DNA scissors to the DNA-binding domains. Well, the ease and low cost and flexibility of the CRISPR approach has allowed nature to declare genome editing as a method to watch in 2014. This vital large size and a potential immunogenicity, we and others, indeed, are repurposing this great technology for epigenome editing as well. 

In this webinar I will give you some insights into the field of gene expression modulation of any given genomic locus, without changing the underlying DNA sequence. We can distinguish two classes of tools: one, is the artificial transcription factors abbreviated as ATFs. In this approach, engineered DNA-binding domains are fused to transcriptional effector domains, like VP64, which is a tetramer of VP16 - a viral transactivator - or a KRAB domain, which is derived from repressive transcription factors. Over two decades of research into this field, as shown, that actually any gene can be targeted for expression modulation. However, these studies also indicated that there are context-dependent effects, and I will show you some data from my lab as well in a minute.

As these effector domains do not contain catalytic activity themselves, they are considered to have transient effects. So we reasoned, and others, that swapping of the effector domain with an epigenetic writer or eraser might indeed induce sustained effects. I will show you that, indeed, epigenetic marks can instruct gene expression, but whether this is sustained I will show you some examples, but also some considerations. Because the influence of the local chromatin environment indeed effects sustainability.

But, first, some more data on artificial transcription factors. So they can be used to repress any gene of interest, and this is an example that we targeted, the SOX2 gene, which is an undruggable protein of high importance in, for example, breast cancer. This study was performed in the laboratory of Pilar Blancafort who is one of the pioneers of engineering zinc finger proteins. The nice thing of the artificial transcription factor technology, is that you can use the same approach also to induce gene expression. If the gene is being expressed from its endogenous locus, you can express all splice isoforms, and maybe even in their natural ratios. 

We made use of the fact that you have one technology to address the function of a gene, to explore the unknown function of the epithelial cell adhesion molecule in ovarian cancer. To induce gene expression, it is a prerequisite that the heterochromatin does not present binding per se. Many studies, including some of ours, have shown that this indeed is the case. 

Actually, we did not encounter any gene which is localised in the heterochromatin, so an epigenetically silenced gene, which we could not target for upregulation. This is an example where we targeted EPB41L3: we designed two six finger zinc finger proteins to target this gene and fused the transcription or activator VP64 to it. Here you see the results for a panel of cervical cancer cell lines where the gene is hardly expressed by cells transduced by an antivirus, or cells transduced to express the zinc finger without an effector domain, let's say, for number 22. Now, 21ab seems to outperform, with respect to gene induction of EPB, 22; but this can be explained by differences in the expression level of the artificial transcription factor itself.

However, if we then went to a panel of breast cancer cell lines MDA-MB-231, it showed the same effects. So number 21 were better than number 22, but in the SKBR3 cells the opposite was shown. This could not be attributed to differences in gene expression, as number 21 was very inefficiently expressed in SKBR3. Actually, every gene can be re-expressed, but not every construct works in the same manner. 

Here are some more examples where the heterochromatin is accessible, but not to every construct. These are cells transduced to express an artificial transcription factor to upregulate TFPI2. In HeLa number 43 seems to do the trick, and this could also be confirmed in another cell line, and number 43 performed the best. All the others - we designed 11 - were far less robust in changing gene expression of this particular target gene. Another target gene, CCNA1, again, we designed six artificial transcription factors: one worked the best in HeLa cells, and this one could also do the trick in the three other cervical cancer cell lines. So it seems that, in general, if we design artificial transcription factors, we can induce gene expression from a silenced heterochromatin location. However, not all constructs are equally effective.

This is a more extreme example where we tried to upregulate a gene called C13ORF18. It did not have a function, but it was known or found out by collaborators of us who thought that C13ORF18 was hypermethylated in cervical cancer cells, but not in normal cervical ileal cells. So we set out to design artificial transcription factors to upregulate this gene, and to address its biological function. In HeLa cells we saw that number 1ab fused to feed the 64, indeed could upregulate the expression level of about 100 folds. Number 2, 3, 4 and 5 were not very effective in doing so. In another cervical cancer cell line where C13ORF18 was equally hypermethylated, number 1 was not effective at all, but now number 3 and number 5 could do the trick. 

But we wanted to investigate if epigenetic drugs could maybe synergise somehow with artificial transcription factors, and improve the effect. So now cells were treated with suboptimal doses of 5-aza or TSA. These treatments, if there was no artificial transcription factor, did not induce because it was suboptimal, it did not induce the expression of this gene. Now, if cells were transduced to also express the artificial transcription factor, then especially cold treatment with TSA could induce the effect. Interestingly, for numbers 2, 3, 4 and 5, which were not effective without epigenetic drugs, we could read similar induction levels as for number 1, which was effective without epigenetic drugs. 

These kinds of studies are very important to identify the parameters which are underlying efficiency of the artificial transcription factor approach. Actually, this doesn't only go for the zinc finger proteins, but it's known that also TALEs and CRISPRs are not acting in a very straightforward way. So we do need systematic research in order to have gene expression modulation as a robust technology. 

Also, the effect of the effector domains are not always very straightforward. In this example, which is in press in Molecular Oncology, we aim to down-regulate Nrf2. Nrf2 is a master switch off antioxidant genes. We engineered six different artificial transcription factors now fused to SKD, the super KRAB domain, which is supposedly a repressing domain. None of these had effect in one ovarian cancer cell line SKOV3 which we tested, but in A2780 we could down-regulate the expression using artificial transcription factor number 2. This number 2 was also effective in a non-cancerous ovarian epithelial cell line. Interestingly, number 5 induced the expression of Nrf2 expression, even though it was fused to SKD which is supposed to be a repressive marker. The same was found in the normal epithelial cells, but we looked a little more in detail and we saw that the association of the artificial transcription factors, to their intended target sites using a HA TECH, which is fused to the zinc finger by conventional chromatin immunoprecipitation was equally effective. 

Then we constructed the luciferase reporter plasmid to see if it was directly on the promoter area and, again, OX2-SKD down-regulated the promoter activity, whereas OX5-SKD upregulated the promoter activity. Interestingly, upregulation by SKD has also been observed by other groups, and these kinds of studies will identify much more about regulation mechanisms of epigenetic, or, otherwise, effector domains. 

Another aspect I need to address is specificity of binding. Here you see four engineered zinc finger proteins and they associate to their intended targets, ranging from 20-60% and we also have examples going up higher. However, ChIP-seq approach identified thousands of off-target binding events. Interestingly, promoter areas seems to be overrepresented in the areas where the reads map to as it was equally represented when regions of plus or minus, 100 base pairs from the TFS were taken into account, compared to regions plus or minus 10,000 base pairs from the TFS, so this gives an overrepresentation. 

These kind of ChIP-seq studies which are performed by us and have been reported also by others, are very important in understanding binding of engineered zinc finger proteins, and also in improving the specificity profile. In parallel, we and others are exploring other approaches to increase the specificity of the effect by, for example, co-targeting split enzymes. These are enzyme parts which you need to dimerize and upon dimerization they will do what they have to do. 

So two decades of artificial transcription factors clearly show that we can induce or repress any gene of interest, and the dogma that heterochromatin is inaccessible has been found to be not true. Zinc finger proteins did not show any adverse effects in clinical trials, and also TALEs and CRISPRs have now been exploited as artificial transcription factors, and have shown great effectivity. So maybe this approach would allow to realise the druggable genome concept, because at this point we can only drug less than 2% of our proteins being expressed. So capability of modulating gene expression of all protein coding, as well as non-coding genes, would open up new avenues. However, the artificial transcription factors do not contain catalytic activity by themselves, so they are presumed to have transient effects. 

To start to explore these kinds of effects, people and also we, made stable cell lines which express the artificial transcription factor upon doxycycline treatment. Here, you see the kinetics of a cell line engineered to express an artificial transcription factor to upregulate EPB. Upon treatment of two days with DOX, the artificial transcription factor expression indeed goes up; but upon removal of DOX the expression goes down. The target gene follows the same kinetics with an optimal of three days, but then, again, goes down to background levels. 

For the same gene, but now in a different cell line, shows that for this cell line it was more difficult to upregulate EPB using the artificial transcription factor. But a 5-aza could upregulate and upon removal of 5-aza, the upregulation was maintained. But if both 5-aza, as well as the artificial transcription factor were allowed to be expressed, then the induction of gene expression was improved. Again, this improved the effect was maintained over five days. 

This is another example where the effect of 5-aza was also investigated and shown to be a transient for this particular gene. So here you see a stable cell line expressing upon DOX treatment. A zinc finger protein fused to no effector domain, so by itself, or the same zinc finger protein but now fused to VP64 - so the artificial transcription factor. The gene is not expressed, there's no DOX and no 5-aza. Now, if we treat the cells to induce the zinc finger only, as well as 5-aza, gene expression goes up but it's not maintained. So upon removal and sub-culturing for seven or a total of 20 days, you see that the gene expression levels go back. If we go and treat the cells with the artificial transcription factor and 5-aza for three days, again, gene expression goes up, but now it stays when cells are being washed from the doxycycline treatment, and allowed to subculture for another seven days at an even state until 20 days. 

So the artificial transcription factors can help to maintain gene expression induction, but why not fuse epigenetic writers or erasers to induce gene expression, or repress gene expression in a more permanent way. At the moment, we have a dozen of publications showing that indeed epigenetic editing results in gene expression modulation. But the first paper I already introduced, was already published back in 2002 where this research group targeted gene 9A, a repressive histone modification; and, indeed, showed that gene repression was associated with writing this gene. I will show you in a minute some of our data where we validated this approach for an overexpressed oncogene Her2/neu. 

Later that year, Nature and Nature Biotechnology also showed that writing published papers where researchers wrote repressive histone marks on genes, or even enhancers which did induce gene repression. This paper actually showed that in vivo histone modifications can interfere with behaviour with respect to addiction.

Before that, we actually already tried to target DNA methylation and showed four different endogenous genes that the induction of DNA methylation does result in gene repression. Excitingly, also removing DNA methylation can induce gene activation, and this has been shown by Fedulov by targeting DNA repair enzymes using engineered zinc finger proteins, or by using TALEs or zinc fingers fused to Tet1. I will show you some data of our published work where we compared Tet1 with Tet2 and Tet3 and showed that, indeed, DNA demethylation can induce gene activation. Very recently, Pilar Blancafort published on the maintenance of induced DNA methylation in an in vivo model, and I will get back to that. 

First, some of our published data, quickly I will walk you through the target of histone modifications resulting in gene repression, targeted DNA demethylation resulting in gene activation. I will also show you some unpublished data where we showed that this is a feasible approach for hypermethylated genes. Then I will move to the very exciting aspects of sustainability of the rewritten epigenetic landscape. I will show you our data on the targeted DNA methylation resulting in gene repression and targeted histone modifications; but in this case to induce gene expression. 

First, the study on inducing repressive histone marks. We took advantage of a zinc finger protein which was engineered by Carlos Barbas - who sadly died far too young last year - which was designed to target the Her2/neu gene. This is the position where the zinc finger binds to, and this is the region we referred to as region C for where we did the ChIP assay to show that, indeed, targeting of a zinc finger fused to a G9a could write the intended mark, being H3K9 dimethylation on to its intended area, and also on a neighbouring area, region B, we could find association of the mark to this gene. We could also show association of H3K9 dimethylation 1kb upstream, but whether this actually indicates that they are spreading and that would need more thorough investigation.

The other constructs like zinc finger-only or the zinc finger fused to the repressive domain SKD, did not induce; it represses histone mark. So does the induction of the histone mark, be it quite inefficiently in this total cell population, indeed, instruct gene repression? We tested the breast cancer cell line and an ovarian cancer cell line, and in the breast cancer cell line some level of repression was observed not as effective as the artificial transcription factor. But in the ovarian cancer cell line, equal repression was observed for targeting G9a, the epigenetic editor, or SKD an artificial transcription factor. We also targeted SU(VAR) which is an enzyme writing H3K9 trimethylation, and this also resulted in some level of gene repression. 

So in order to exclude that this level of gene repression is actually due to the written histone modifications, we set out to construct a catalytically mutant form of G9a. Now, on the y-axis you can see the amount of protein being formed, the zinc finger targeting G9a to the Her2 now locus results at about 50% inhibition of protein level. Whereas the mutant G9a did not affect protein levels at all. So this does show, like the Snowden paper that histone marks by themselves are instructive of gene repression.

Can we also re-express silence genes? For this purpose we, again, took advantage of a Barbas zinc finger protein, but now targeting the ICAM-1 promoter, which is hypermethylated in ovarian cancer. This ICAM-1 promoter, referred to as CD54, binds the genomic locus at a position which contains two CpGs. We'll refer to these as CpG number 10 and number 11. As effector domains we fused VP64, the tetramer of VP16, and either member of the Tet family. Note that these Tet catalytic domains, so not full-length, but only the catalytic domain still is quite large compared to the VP64. So we had to sort the cells, because it's quite difficult to transduce cells to express these large constructs. Now, for the zinc finger-only, CD54 no effector domain, or the zinc finger fused to the transcriptional activator, the artificial transcription factor. There was no need to sort the cells, and we compared to the untreated cells. For the percentage of methylation first on the DNA-binding site CpG number 10 and number 11. We saw that the fact that the protein was binding to this region already induced a decrease in percentage of methylation, going from 80% to 60% for the zinc finger-only.

The same was observed, similar data, for targeting VP64. Now, in this sorted population of cells treated to express these large fusion constructs, we did not see a tremendous effect on the DNA-binding domain of these zinc finger fusions. This indicates that on a per cell basis, the expression of these fusion proteins are really low; and the same goes for CpG site number 11 which is also within the zinc finger binding site. 

What about the target CpGs; CpG number 12, 13 and 14? Now you can see that the zinc finger by itself does not affect DNA methylation anymore, whereas the artificial transcription factor which induces expression of the gene, so it's recruiting a lot of other players, still removes or reduces the amount of methylation. But now the ineffectively-expressed fusions of the zinc fingers to Tet1, and especially to Tet2 could remove DNA methylation to more or less the same effects as VP64. Tet3 was incapable of lowering methylation degrees and Tet1 was somewhat less effective than Tet2.

So, again, we wanted to make sure that the effect of DNA demethylation was actually due to the catalytic activity of Tet1 and Tet2. So we made catalytic mutants and we targeted those to the ICAM promoter by fusing these to the same zinc finger proteins. Again, you see that there is a little effect on the CpG in this zinc finger-binding site, so in this example CpG number 11. Again, Tet1 and Tet2 could lower the degree of methylation on the target CpGs, and this example of CpG number 13, the same as for the VP64 construct, but the mutants did not induce any lowering of the methylation degree. So the effect on methylation of targeting Tet1 and Tet2 actually is due to the catalytic domain. Now, the exciting question is, does it result in gene repression, and it does. Here you see the artificial transcription factor, it induces ICAM expression to about 500-fold, very effective. Using the same protocol with Tet2 targeting induces gene expression two-fold, which was not observed for the catalytic mutant, so it's not because of recruiting secondary players.

This re-expression of silenced genes has also been confirmed for other genes. Here we see C13ORF18 where the previously shown artificial transcription factors induce gene expression, antivirus does not. A combination maybe outperforms the single agents by themselves, but now number 3 fused to Tet2 and number 5 fused to Tet2, and also induced gene expression. This was obtained by an improved retroviral transduction protocol, so we could reach effective gene expression levels. It was not observed for zinc fingers which did not contain any effector domains, and indeed methylation profiling using pyrosequencing would confirm that DNA lowering of methylation has taken place. 

This was also shown for CCNA-1, message RNA goes up compared to the empty vector NTFBi2 message RNA goes up, compared to the empty vector. It is not as effective as for VP64, but this is quite localised effects and if we can combine several epigenetic editors to actually rewrite the epigenetic landscape, then we might obtain a technology to permanently re-express silenced genes. 

Targeting histone modifications, repressive histone modifications has been shown by us and others to resolving gene repression, targeting demethylation of the DNA and, indeed, induce gene activation. But what about sustainability? Pilar Blancafort published a few weeks ago a very intriguing paper in Oncogene, where she showed stable oncogenic silencing by targeting DNA methylation. I will show you some data in a non-cancerous context that, again, DNA methylation results in gene repression, and that it is mitotically stable. I will also show you some data on targeting activating histone modifications, and that these indeed are also instructive of gene re-expression. I will show you some data on the sustainability of this approach. 

So for this purpose, again, we engineered stably expressing cell lines, which upon DOX treatment will induce the epigenetic editor, in this case, zinc finger proteins fused to MSss1 which is a bacterial DNA transferase. Upon removal of DOX, expression of the editors goes back down. 

If we are harvest the cells after two days' treatment of TGF data, which is an inducer of the target gene PLOD2, then we see indeed induction of gene expression, the normal expected ten-fold for cells, cultured cells, the anti-vector cells. Also, for cells which vaguely expressed the zinc finger fused to a double catalytically dead MSss1 mutant. Now, upon expression of the zinc fingers 7 and zinc finger 8 fused to MSss1, we can prevent the induction to about 50% of PLOD2 levels. This was associated with targeted methylation of about 45% for both constructs.

Now, what happens after another eight days of culturing, then we see that the gene repression, or prevention of induction is maintained for one construct, but not for number 8. The DNA methylation profile was mitotically stable for number 8, and even got induced for zinc finger protein number 7 fused to MSss1. We have been capable of improving the repression of induction, and we are currently looking into the mechanisms to achieve sustained repression of this gene. 

So what about targeting activating histone marks? Again, we set out to make stable cell lines, in this case to target a methylated silence gene, to target activating histone modifiers to induce re-expression. In this case, treatment with doxycycline to express the artificial transcription factor induced the gene to be expressed, that this was a transient approach. So the expression went back to background after another seven days of culturing. Now, if we have the expression of an activating histone modifier, the expression goes up more or less to the same extent as the artificial transcription factor; but, again, it went back down upon removal of the DOX treatment. The mutant could not increase, obviously, the gene expression level that this transient effect caused us to hypothesise that maybe the hypermethylated circumstances did not actually allow maintenance of the activating gene. So we set out to use the same approach. This is to show that the histone modifications, indeed, were written, but, again, were not stable.

So we set out to target an unmethlyated, yet silenced gene and now upon induction of the artificial transcription factor the expression goes up. Again, it was transient, so it went back down upon removal of DOX. Expression of the activating histone modifier increased expression far less effective than VP64, but upon removal of the editor the effects got reinforced. So not only being maintained, but actually being reinforced and this is very promising to us that dependent on the micro chromatin, so the local epigenetic environment we can induce sustained gene re-expression. 

So to sum up epigenetic editing, the rewriting of epigenetic marks at any genomic locus, it is feasible that with respect to gene expression modulation it is effective. So this is a powerful research tool to study epigenetic molecular mechanisms, and maybe in the future this can be a biomedical tool where we can envision a cure for incurable diseases. However, we do need to unravel rules which underline binding effects and maintenance for all three available platforms: zinc fingers, TALEs and CRISPRs. So despite these exciting potentials there are pitfalls with respect to chromatin-context dependency, we don't know too much yet. We do know that heterochromatin does not prevent access per se. Also, with respect to instructiveness of epigenetic marks, all studies need to be controlled by targeting inactive catalytic mutants. But the studies out there already indicate that certain epigenetic marks can enforce gene expression changes.

So can we make this into a hit and run approach? Well, that depends on the choice of the effector domains, which should be guided by the local chromatin environment, but effects can be sustained and may even be reinforced. Of course, there are other approaches which can realise the druggable genome concept, and sleeping genes can be awake by transferring of cDNA. But these approaches are limited by the size and the isoform choice, and also the expression is uncontrollable. With epigenetic editing, we might induce natural expression control, thereby mimic nature more closely. The mute screaming genes, siRNA, is an extensively explored technology; however, to circumvent the transient nature, we would require potentially harmful integrations of the expression itself. With epigenetic editing, we can aim to induce sustained silencing, thereby developing a hit and run approach.

This leaves me to acknowledge some essential collaborators, and local and national collaborators, but, more importantly, my team who is now actively collaborating to get sustained gene expression modulation by using epigenetic editing. Obviously, funding agencies are very important. 

Some final words, we start recruiting PhD students on a European Union Consortium project where breast cancer epigenetics is central. It's called EpiPredict and please keep this website insight, if you would be interested. Also, we are initiating an EU cost action where European chemical epigenetic research groups are invited to join. Lastly, I'm proud to announce the first impact factor of the Official Journal of the Clinical Epigenetics Society (IF 6.2). So thank you for your attention and I will guide you back to Kalina. I will be back in five to ten minutes to answer any of your questions.

K:    Thank you very much Marianne for such an interesting talk. Hello everyone! I'm going to tell you about the epigenetics resources and products that we have available for epigenetics research. All our epigenetics resources can be found in one central location on the website. Here we have grouped everything together to make it easier for you to access what you need. I will briefly show you examples of the types of information you can find on the page, but also encourage you to explore the rest of the resources at

Firstly, on our main epigenetics page you will be able to find useful protocols, tips and troubleshooting guides related to your epigenetics work. From here you can also register and attend free webinars featuring the latest research and technology talks. You can also watch all our previously recorded webinars and videos by visiting our webinar library at Another useful resource are the posters and interactive pathways, which you can view and download for free at

Additionally, here on the epigenetics main page you will see that each month we summarise the latest published papers in epigenetics so that it is not so time-consuming to keep up-to-date with the scientific literature. You can also read inspiring interviews with top epigenetic scientists.

Furthermore, we keep a calendar of all global epigenetics conferences, meetings, seminars and workshops which you might want to attend. This calendar contains links and information for each event to help you choose and plan.

Now, a little bit about the products that we have available for epigenetics research. We have an extensive range of products to help you with your work, such as ChIP grade antibodies to histone modifications and transcription factors. Kits and assays, for example, high sensitivity ChIP kits, assays for DNA methylation and DNA demethylation, as well as DNA modification detection kits. In addition to that, we have other reagents to complete all of your experiments. 

In line with the latest technology, last year we launched a mouse monoclonal antibody to CRISPR-Cas9. We have validated this antibody in-house in western blot, as shown on the left hand side of the slide. However, in order to demonstrate the usefulness of our products to you, we think it is important to show you how they work in the hands of other scientists. Here, on the right hand side you can see an IHC image from a customer's review showing you how they used this antibody, and how it performed in their experiment. 

Finally, in our pipeline, using our RabMAb technology, we are developing CRISPR-Cas9 rabbit monoclonal antibody, which we intend to be suitable for more applications. Additionally, we will be producing HRP directly conjugated CRISPR-Cas9 RabMAb antibody for multicolour experiments. Thank you all very much for listening, and now back to Marianne who is ready to answer your questions.

MB:    Thank you, Kalina. In view of time, I've made a very quick, short list of the questions which came in. One of the questions proposed by one of the attendees, is that the tools are really big molecules, can we use them for living animals? Indeed, we can and some of the publications which I've shown on one of these slides, actually used living animals to induce gene expression modulation. We and others have also targeted G9a to induce the repressive histone mark, so it's actually also dependent on the effector domain you choose. So when I pointed out that the size was quite large, this was for the Tet family members and these really are one of the largest effector domains we have tested. So that is about the top limit of what we can do, and that also poses the problems for the TALEs and the CRISPRs, because these DNA-binding domains actually are much larger than the zinc finger proteins. Even more hampering accessibility to mainly heterochromatin environments. So that will be one question.

Another question is, do you think targeting enhancer regions could increase stability of gene transcription or modulation by epigenetic editing? Yeah, that's a very interesting question, especially since enhancer epigenetics seems to be more predictive of the actual gene expression level, so enhancers might be more instructive. On the other hand, maybe the promoter area is more served as a lock for gene expression, so if you want to repress a gene it might be a wise decision to target these. So, to be honest, the field is just too young to actually already answer these kinds of questions, we just have to do it and to see what is the most effective way of doing so? With respect to analysis, I have to mention that Mendenhall already published the targeting of enhancers for repression, so please take a look through that very interesting paper. 

Let me see, another question, yeah, obviously, specificity is a problem, so if we are thinking about clinical trials then we really have to find more insights into the binding profiles. One way to go about its specificity is what people do currently for targeted nucleases where the enzyme needs to dimerise. Though we and others are looking into splitting epigenetic writers or erasers, which only act upon fusion of the two halves of the enzymes and this will increase the specificity. 

I'm not sure in view of time; I guess I'll do one more. Is it likely that the level of gene expression and also the choice of alleles are determining the outcome in disease? Would it be possible to achieve mono-allelic targeting? That's an interesting question. I think if snips allow a differential targeting, then this might be explored because single base pairs changes do affect binding efficacy of either platform. Also, the differential epigenetic context can be exploited to make use of this difference in epigenetics between the two different ileals, and then rewrite the epigenetic circumstances. How to have appropriate level of gene expression that is considered to be healthy to resolve diseased state, I think that is the advantage of epigenetic editing that we are not actually enforcing a gene to be expressed like with artificial transcription factors, or CDNA approaches; but we are mainly opening up the regulatory areas to allow gene expression to occur.

This might also be a problem though, because if those signalling pathways area also disturbed, then it will not function. But, again, we have to try and find out and see. So thank you very much for attending this webinar, I will answer the many remaining questions by email and also you can request the PDF of this presentation. Thanks again.

Thank you, Marianne, and thank you Kalina for presenting today. Unfortunately, like Marianne said, we were not able to answer all your questions received. For those questions that were not answered, we will contact you shortly with a response to the 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 We hope you have found this webinar informative, and useful to your work. We look forward to welcoming you to another webinar in the future, and thank you again for attending and good luck with your research!

Sign up