And it's a pleasure to be here today to talk to you. So, I'm going to tell you today about a novel omic-space [sp] method that we've been developing in collaboration with Agilent Technologies for sensitively measuring DNA damage throughout the genome.
So, I've broken today's talk down into the following three sections. I'm going to tell you a little bit about the--first of all, tell you a little bit of the background to the development of this method. I'll move on to tell you a little bit about some of the tools that are that are available for analyzing DNA damage and repair mechanisms. And I'll include in here this new method that we've been working on.
And then I'll finish up with a sort of a personal view really on where I think the system-wide view of cellular response to genotoxins should be heading, which really incorporates integrating many of the omic-state data that we're going to be hearing about at this meeting.
Okay. So, a little bit about the background then. Why did we develop this method? So, as Nigel mentioned, my interest is in the DNA damage and, in particular, the mechanisms of DNA repair.
As I'm sure most of you will know that there are a wide range of DNA-damaging agents and a wide range of lesions that can be induced in the DNA molecule and, of course, a wide range of different DNA-repair mechanisms that have evolved in order to correct these lesions.
Now, the pathway that I'm particularly interested in is this pathway here, called the nucleotide excision repair pathway. And that's a pathway that takes care of lesions that get induced in the genome as a result of exposure to UV radiation, for example, or indeed chemical mutagens from the environment.
So, this is the--what we currently know about the biochemical mechanism of the nucleotide excision repair process. And the only thing I want you to really notice here is that it's a multistep, multiprotein pathway. And we've learned quite a bit about how this mechanism works. There are proteins that are involved at all these different stages of the process here, which work in conjunction to remove the damage that's been induced as a result of the exposure.
So, the other thing I wanted to mention was that, in this case, we--although we're studying UV, we're using UV here, the well-known genotoxin UV as a paradigm really for DNA damage. And the--what I'm going to tell you about can be applied to any of the kinds of DNA damage and indeed the other DNA pathways that are involved.
The other thing I wanted to mention on this slide was that, in fact, you can see that there's a wide variety of model organisms that have been used in order to sort of elucidate this pathway. And what I'm going to be telling you about has been developed in the simple eukaryote yeast.
But, the point is that, as I'll show you a little bit later on, in working with Agilent, we're developing these assays for use in human cells, which is where we think they'll have an important application in the area of genotoxicity testing.
So, the reason that we studied this pathway, of course, is that, when this mechanism goes wrong in human cells, you end up with a severe cancer-prone disorder known as xeroderma pigmentosum. And as a result of understanding the biochemical mechanisms in some detail, we've been able to understand the reasons behind a number of diseases that are associated with defective DNA repair, not just XP.
So, now, I'm going to tell you a little bit about some of the tools that are available to analyze the DNA repair process. And really that sort of detailed understanding that we've gleaned, if you like, over the years, has really come from new assays that have been developed, which have allowed us to tease apart the details, the mechanisms behind these processes.
So, if we go way back to the early days of DNA repair studies, we can see that we had these very simple cellular-based assays that measured the incorporation of radio-label into the DNA damage patch.
So, what we're looking at here is, as I say, a cell-based assay that uses BRDU, radio-labeled nucleotide analog. And you can see that, in undamaged cells, you can see the incorporation into cells that are dividing. So, this is called scheduled DNA synthesis in replicating cells.
If you take normal cells that have been irradiated, you can see that there are these little patches of what's known as unscheduled DNA synthesis or repair synthesis that is taking place.
If you take cells from one of those patients that I showed you earlier, those XP patients that are defective in nucleotide excision repair, you can see that they're capable of conducting the replication synthesis here, the scheduled DNA synthesis, but not this repair synthesis here anywhere else in the undividing cells.
Now, this is a very simple assay. It's a cell-based assay that you can see major defects in the repair processes, such as you have in the case of XP patients. But, it's not a very sensitive assay particularly.
Now, if we go to looking at some of the more recent DNA repair assays, here, we're looking at the induction of DNA damage CPDs induced by UV light at the level of the nucleotide. This is a sequence ladder that you see just along the side here. And by analyzing the repair process over a period of time, you can see the removal of these lesions over time, showing the repair of those damage.
Now, that's a very useful assay to increase the sensitivity and very high resolution. But, of course, using all these gel-based-type systems, you're only able to look at fairly short regions of the genome.
So, what we wanted to do was to be able to combine these two extremes, if you like, so that we could look at the genome overall, as we could with the cell, but look at high resolution, as we could with the gel-based assays. And so, that's why we developed this particular method.
And the way that we chose to tackle this was to use a location-analysis approach or ChIP on chip, as it's sometimes known, chromatin immuno-precipitation on arrays. And this is just a simple walkthrough, if you like, of the process.
The method was originally developed in order to identify the position of proteins of interest to find out where they bind within the genome. You extract the DNA as chromatin. And you immuno-precipitate your protein of interest. And then you recover the DNA. And you label it with a fluorescent tag. And then you hybridize that captured DNA on a DNA chip, which contains sequences that represent the locations of the genome right throughout the whole genome.
And that's--so, that's the wet lab part of the work, if you like. And then the second part of the analysis, of course, involves this so-called--the bioinformatic component, if you like, is where you take these arrays. And then you analyze them. This is not a trivial process at all.
There's a normalization process. And the data analysis eventually allows you to convert the readouts that you get from these chips to show you where the location of your protein of interest was in the genome.
Now, we adapted this instead of capturing proteins that are bound in chromatin to affinity capture the damage to the DNA itself. So, I'll just walk you through very quickly again so that you get the point.
Here's our UV-induced DNA damage. And now, we've coupled antibodies or other methods of affinity capture to a Dynabead in this case, which we can then introduce. And then we can use a magnet in this case, which will then allow us to isolate our damaged fraction from our undamaged fraction. And then we're able then to take our immuno-precipitated sample and our input sample. And just like Al, we use the Agilent's two-dye system for this purpose. We label with a psi-3 and psi-5.
I just mention at this point here, in case we get time at the end, this is a point where you can complete the chromatin IP part of the process. And you can validate the process using QPCR at this stage.
The next stage, of course, is to take it onto the microarrays. So, let's just look at the labeling reaction. Here we are. We have a red and a green signal. And then we take the damaged and the input sample, and we put it onto Agilent's microarray platform, which we found to be, as Al mentioned, very reliable and robust platform for this purpose. So, here are our arrays. And that will then eventually go off and be read. And you'll get an example of that in the next slide.
So, this is telling you a little bit more about the arrays and the bioinformatics component of the work. So, what we're looking at here is--oops, if I can show it--is each feature that's printed on the array here has multiple sequences of a defined region of the genome. And this is where you hybridize your sample that has--contains the input DNA and the damaged DNA.
This will then get read in Agilent's scanner. And you'll end up with an output, which gives you a readout like this, which will give you an indication of the level of damage that is present at each location within the genome.
This can then be analyzed and turned into numerical values and compared with the location of that damage in the genome. And you'll end up with a kind of a trace that looks something like this, which will show you the precise location of where that damage is in the genome.
And then, of course, if you repeat this process over a period of time, you can then get an estimation of the level of repair that takes place for that damage.
So, here's a quick scan of the actual data that comes out from this. And what we're looking at here is a small section of chromosome four in yeast. This represents around one two-thousandth of the total genome. And you can multiply that a couple hundred times if you're talking about the situation in human cells.
And so, you can see--this is the--if this thing works--this is the chromosomal location across the bottom here. And here we have the levels of damage at quite high resolution. And we can go higher than this actually right the way throughout the entire genome.
Now, one of the nice things about CPDs is they've been very well studied. And so, what we can--what we've done here is build an algorithm that gives you the theoretical distribution of CPDs in this case. And you can see that our actual pattern matches very closely. There are differences in different parts of the genome. And those are real differences based on the chromatin structure that exists in yeast.
So, then if we look what happens during the repair process, so now we can see that, if we look at one hour versus one hour after repair here instead of immediately after DNA damage, you can see this reduction in the signal that you can see, which is indicative of repair synthesis. And I'll show you a little bit more about that here in just a moment.
So, here is then the rates of DNA repair that differ at different regions right the way throughout the genome, from very low rates of repair here to much faster rates of repair here. And this is just to show you that what we're looking at here is indeed repair.
So, if we knock out a gene that's important for nucleotide excision repair, as [unintelligible] those XP patients, for example, you can see that this is the flat line that you get when you have a defective repair process. So, this indicates that this is real repair that we're measuring here.
Okay. So, hopefully--so, this method, by the way, was published last year some time I think and is--I think Nigel has some copies of that. That's available if anybody would like to look at that in more detail.
So, as I mentioned at the start, we've been developing this method with Agilent. We have a knowledge-transfer partnership back in the U.K. for doing this. And of course, the aim here is to develop a method for use in human cells. And we're quite some way down the line with that at the moment.
The aim here is to develop an in vitro alternative to genotoxicity assays and to get a--to develop a much better understanding of the mechanisms of toxicity.
So, just a couple of bits of data on the human work that we've done so far, so we've done the affinity capture of CPDs in this instance. And we've done--we've got a couple of clinical guys in the lab who are working on other types of DNA-damaging agents, mainly chemotherapeutic agents. In this case, we're looking at oxaliplatin and cisplatin damage. Those projects are to try to get at issues of stratified medicines or understanding why individuals respond well to drugs or not respond well to drugs and how they do it in terms of peripheral neuropathy and those kinds of clinical questions.
So, we've moved on a little bit from there. And this is some of the traces that we've got actually for the cisplatin damage here, working now in human cells. We're a little bit behind on the UV work, haven't quite got that working just yet, but we're making good progress I think in that area.
So, in the last few slides, I wanted to just tell you a little bit about my--how I think, you know, this sort of technique and these kind of technologies can be used in the context of getting a system-wide view of the cellular response to genotoxins.
And what I'll tell you about a little bit is some of our own work, where we've been starting to examine the--get a genomic view of how the DNA repair factors themselves respond to genotoxins; take a look at some of the epigenetic changes that occur that are necessary for the DNA repair process to take place; but then, in addition, I think it's going to be very useful for doing the kinds of things that Al's been telling us about earlier today, looking at the transcriptomic data and indeed metabolomic data and indeed proteomic data as well, for example.
So, I think by combining all of these techniques and being able to analyze the data in total is where you'll sort of be able to leap above the sort of signal-to-noise problem to really understand how the cell is responding to genotoxins.
Okay. So, what I was just briefly going to tell you a little bit about is our DNA repair work to show you how we've used this--the techniques that I've just told you about to, in addition to looking at the damage and the repair of that damage to look to see how the DNA repair complex themselves are responding to the genotoxins.
So, this is a complex of proteins that I'm not going to tell you too much about, only to say that they're involved in the initiation of repair of UV-induced DNA damage. This is--the ABF1 component is approaching a DNA-binding protein that binds to a consensus sequence in this cell and is found at thousands of sites throughout the genome.
The RAD16 component here is a--what's known as a swy-sniff type protein, which is--actually behave as a DNA translocator. We can--it has ATPase motors that can allow it to move along the DNA.
So, I'm going to tell you a little bit about what we've learned about how this complex of proteins does in response to the genotoxin UV.
So, what we've done here, we've looked at the induction of --sorry, the binding of ABF1 and the RAD16 components of those--of that protein complex. And this is the pattern pre-UV. And we've looked obviously post-UV. And I'll show you a little bit about that data.
So, what we then did was we analyzed the data by identifying the ABF1 binding sites and then forming this so-called composite plot. So, what we've done here is we've taken each of these little individual peaks. And we piled the data up and analyzed the data in the context of the position of the ABF1 binding sites, as I said, found thousands of sites throughout the genome. And then we can develop a sort of a trend line here which will give even an average of what's happening to these sites throughout the genome.
And there are many more ABF1 binding sites than there are RAD16 proteins in the cell. And so, what we've done is we've analyzed the sites at which the RAD16 and the ABF1 and the RAD7 are co-localized. So, in other words, with our GGNA [sp], our complex is binding in the genome.
And so, this is what happens. So, this--here's our ABF1 peak in the absence of damage. And here's our trend line. And this is what happens to that ABF1 plot in response to UV. So, you can see it's not very much of a change in terms of the binding of the ABF1 protein after UV damage.
But, if we look now at the RAD16 component of that complex, we can see that it's bound here in the absence of damage. And then we get this redistribution, which is looking a bit skewed from where I'm standing here. But, it's a redistribution away from that ABF1 binding site to other regions.
Now, if we plot the data in a slightly different way, this time, instead of just putting ABF1 binding sites piled up on one another, we now do that in the context of gene structures. Here, we have our promoter region of the gene. This is dying on me. And here's our--the open reading frame and then the downstream region.
What you can see is that these ABF1 binding sites are found predominantly in the promoter regions of the gene. You see the loss of occupancy in response to the UV damage and a redistribution of that 16 component in response to damage.
So, what I'm trying to get across here is that you can see how the repair factors respond to these genotoxins, get a sense of how the process is really organized in the cell.
The last couple of slides, I'm just going to show you correlating some of the epigenetic changes that occur in response to the genotoxin. So, this is a quick slide showing all the various different types of post-translational modifications of histones.
And one that we know is very important for the response to UV damage is this acetyl group here. An acetylation on histone H3 at lysine nine and lysine 14 is very important for the response to DNA damage. And it's actually controlled by that complex of proteins that I was showing you here.
So, what we've been able to show previously is that this complex of proteins here controls the acetylation status in the chromatin structure in response to UV damage actually by regulating the occupancy of this protein here called--which is called GCN5. It's a histone acetyl transferase that puts acetylation groups here, acetyl groups onto the histone tails here.
So--and let me just show you then how that works out. So, here, we're measuring histone H3 acetylation status, both pre-UV and then post-UV. And so, we can see that--how that works out.
So, this is the pre-UV status of histone acetylation, again, plotting the data in the context of those ABF1 binding sites. And so, in response to UV, you can see this increase in spreading of the acetylation signal in response to the genotoxin.
And indeed, as far as the GCN5 is concerned here, remember that's the histone acetyl transferase that I was telling you about. So, here's the pre-UV--the pre-UV status is here. And then this is the post-UV increase in GCN5 occupancy. And then it comes back down to normal here with time.
So, what I'm trying to show you really is how, by analyzing these datasets, you can get a really mechanist--a real mechanistic sort of understanding of how the cell's responding to these kinds of genotoxins.
And so, what you can see in the basic model here is that our complex is here bound at ABF1 binding sites in the genome here. And it keeps--and the undamaged state, you keep the hat away from the nucleosome. But, then in response to damage, you drive changes which allow the translocation of this particular part of the complex, which changes the chromatin structure, which allows this histone acetyl transferase group to come on and acetylate the genome and facilitate the repair process.
So, hopefully, you get a sense that it's possible to really get a mechanistic understanding of how the UV is causing these changes in the cell.
So, on the basis of that, we have a collaboration with GSK, who are interested in looking at how their HDAC inhibitors are involved in changing the epigenetic status of their--in the cell here.
So, an HDAC is an enzyme that is capable of limiting the amount of acetylation. And an HDAC inhibitor is one that blocks that. And that changes the way that the genes are regulated in the cell.
And this is a big concern at the minute for these companies that are developing these types of drugs because it's becoming clear that HDACs are actually--have a very important role in the maintenance of genome stability. They have a role in DNA repair, both directly and indirectly. And they also may play a role in chromosome segregation. And so, there's an important need to analyze the effect of these types of drugs and their effect on genotoxicity.
So, just in the final slide, I think the--what I've hopefully given you some idea of is that, by integrating the various types of omics data that exists, both looking at DNA repair data, DNA repair factor binding, genomic, epigenetic changes, and transcriptomic responses, will give us a much more detailed understanding of the mechanisms of genotoxicity. I'll stop at that point. Thanks very much.