Andy Sharrocks (Professor at The University of Manchester) joins us to discuss the epigenetics of esophageal adenocarcinoma, why studying epigenetics is useful for cancer research and the latest research coming out of his lab.
Please note the transcript has been edited for brevity and clarity
FLG: Hello and welcome to the latest “A Spotlight On” interview. Today I’m joined by Andy Sharrocks, and we’re going to be talking about cancer epigenetics. Andy, if you could please introduce yourself and tell everyone a little about what you do.
Andy Sharrocks: I’m Andy Sharrocks, a Professor of Molecular Biology at the University of Manchester, and my lab works on how cell signalling and chromatin structure impacts gene regulation. In the past, we’ve worked on a variety of systems ranging from yeast to stem cells. Recently, we’ve moved to studying cancer, particularly focusing on esophageal adenocarcinoma and the gene regulatory networks that are at play in that particular tumour type.
FLG: Why investigate the epigenetics of esophageal adenocarcinoma in particular? How does this disease differ from other forms of cancer?
Andy Sharrocks: Esophageal adenocarcinoma is a particularly deadly disease with high incidence and very poor survival, and part of the reason for this is because we don’t understand the molecular pathways involved as much as we do in many other cancers. One of the issues with esophageal adenocarcinoma is it is a highly mutated cancer type, and yet there are no recurrent mutations. P53 is highly commonly mutated, as in many cancers, but the mutation with the next highest incidence is probably the receptor tyrosine kinase up to 20 to 30%. After that, the usual common drivers aren’t there in this cancer.
Mutations like BRAF in melanoma, APC in colon cancer, or oncogenic fusions in leukaemia, they’re just not there, so getting targeted treatments becomes difficult. At the pathway level, you can see things slightly differently. Receptor tyrosine kinase pathways, for example, if you take different components and look at mutations within those, then you see much higher prevalences so maybe 60 to 65% of all tumours would have those mutations. But again, the common DNA mutations just aren’t there.
FLG: Focusing on epigenetics specifically, what insights and advantages could studying epigenetics have in our understanding of this cancer and other cancers?
Andy Sharrocks: So as I just explained, there are very few common DNA mutations. The next obvious thing to look at is epigenetics. DNA is not naked in a cell, it’s enclosed and encased in chromatin which controls the availability of regulatory elements that in turn control gene expression. So equally, epigenetic changes may cause a change in cell phenotype, which gives you a cancer phenotype. The other thing about epigenetics is it can also give you an idea of where the cells come from in the first place, the cellular origin of cancers. That’s something we try and pursue in our research, trying to understand the basic wiring of cancer cells and where that rewiring originates in the first place.
FLG: You spoke a little bit about chromatin. What is the importance of chromatin unravelling, and how do you go about investigating such a dynamic process that’s constantly changing? What are some of the challenges in exploring a process that’s changing?
Andy Sharrocks: As I just explained in the previous answer, the DNA is encapsulating chromatin and you need to then remodel the chromatin to get changes in gene expression. So you’re revealing enhancer elements, promoter elements and other regulatory elements in the DNA to allow access to the machine that turns on gene expression. Equally, you can do the opposite; you can close down chromatin and closing down chromatin will then shut down gene expression and that will typically shut down particular tumour suppressors that usually stop tumourigenesis. Shutting those down allows cancer to occur as opposed to opening up things, which tend to be oncogenic processes and the oncogenic proteins that are produced from these processes. You do mention dynamics, and studying dynamics is always very difficult. Chromatin is actually not that dynamic compared to other cell processes. The dynamics of chromatin occur in hours and days and weeks and months, so studying the dynamics of chromatin is easier than for some other processes.
FLG: Thank you. So you were speaking about the the genome level, the epigenome level, and then obviously the transcriptome level. How do you see recent advances in the rapidly developing field of multi-omics in terms of data integration and spatial technologies affecting your research and generally advancing our understanding of cancer?
Andy Sharrocks: I think the clue’s in the name with the ‘multi’ part of things. So if you want to understand any biological process, if you look at multiple facets of the same cell, you’re able to then uncover deeper insights into what’s going on in that cell That’s where multi-omics comes in, because you’re able to look at different aspects. For example, you can look at the chromatin level and the gene expression level, and you can integrate those together to understand how chromatin changes lead to gene expression. There’s a protein level downstream, and that’s what actually gives your cell its ultimate phenotype. Multi-omics can help to understand how that phenotype comes from gene expression profiles, so that’s me talking as somebody who understands gene expression.
Equally, if you want to phenotype cells or want to get diagnostics, they can come from any particular molecular aspect or a combination of them. You mentioned spatial technologies and these are particularly important in cancer, where you have a heterogeneous set of cells that are interacting in space. We need to understand where they are, how they interact with each other and the different cell types that are in there. Again, multi-omics can give insights into this in combination with the spatial technologies which are now arising.
FLG: You mentioned stem cells, and you also investigate how stem cells differentiate and some of the transcriptional drivers involved in this process. Could you walk us through your work in this area and how insights into this process are important for your research in terms of the transcriptional side as well as in epigenetics?
Andy Sharrocks: So if you think about stem cells, they’re very primitive cells. They then go through a series of steps and differentiate into terminally differentiated cells. Our work in stem cells has focused on trying to understand those transitions; how you go from one cell type to another. Typically, in the past, people have looked at the start point, your stem cell, and the end point, your differentiated cell. You’re trying to work out how you got from A to B by understanding the transcriptomes in both chromatin states. Where we’ve been looking is the intermediate states in between. When you mentioned dynamics before, that’s where we look at what happens to the chromatin, the gene expression profiles, as you transition from primitive embryonic stem cell through to a more differentiated state. The beauty of studying that system is you can do it in a synchronous manner. You can start off with a synchronous population, and you can watch them through time as you go to the endpoint. By doing that, we’ve been able to uncover transcription factors which define the two endpoints, but also the central transition point. In mouse embryonic stem cells, for example, Zic3 is identified as a transcription factor which is important in the transition states, but not in either of the two end states.
In terms of how this relates to cancer, you can think about cancer cells as having undergone the opposite process. A cancer cell is effectively an undifferentiated cell which has come from some differentiated or partially differentiated cell in the adult, so effectively it’s a process in reverse. Studying that is effectively the same, but we’re looking at how you potentially go from a differentiated cell to a different differentiated cell. That process then goes one step further, where this cancer cell changes properties to a metastatic state which, again, you can see as a progression from one cell state to another. So both processes are pretty analogous.
FLG: Obviously, these transitions happen over time, and recent advances in temporal analysis have allowed researchers to track and analyse these biological events as they occur over time. How do you see these new technologies and advances helping your work as well as other investigations into how stem cells differentiate and how cancer cells develop?
Andy Sharrocks: Well, I mentioned how upon looking at things temporally, understanding how things go from A to B into this intermediate state, especially if you’re looking at signalling events. In esophageal cancer, for example, we have the receptor tyrosine kinase signalling pathways upregulated and they’re obviously signalling through to chromatin and gene expression. One of the problems with looking at signalling is it’s dynamic, again we are looking at endpoints compared to startpoints. There are new technologies like Live-seq at that allows us to look at gene expression changes in real time. You’re able to extract part of cytoplasm from individual cells and analyse that, and it’s quite an exciting technology because you’re able to sample the same cell multiple times. At the moment, most single cell technologies are not able to do that and we have to take snapshots. You can do that through multi-omics, in that you can use it to connect things together in single cells. However, I think the new technologies where you’re able to look at what’s going on in a single cell over time are particularly exciting. You can’t do that with chromatin changes at the moment, and I can’t envisage a way of currently doing it. But if we could do that, it’d be really exciting.
FLG: Thank you. Going back to epigenetics and esophageal adenocarcinoma, do you anticipate epigenetic insights into this cancer and other cancers to translate into patients in the future? Will these studies help cancer patients in the clinic in the near future or do you anticipate that future being far off?
Andy Sharrocks: Well, there’s multiple answers to that question. What epigenetics can give you is an insight into tumours which other types of analysis don’t particularly give you. While transcriptomics, DNA mutation analysis and typical genomics, for example, can give you one answer, epigenomics can reveal new things. In a way, that’s giving you new pathways, new therapeutic targets and potentially new diagnostic targets. There are techniques which have been developed where we can look at circulating tumour DNA from patients, and for example, and you can then observe fragmentation patterns. From that, you can infer the chromatin state of the patient. This can be done so often that you can begin to understand the pathways that are changed in patients. Using this as a diagnostic tool, you can (A) tell if the patients have cancer, so it’s a non-invasive biological tool from blood and (B) you can then begin to understand what those changes might be. So yes, there are potential future uses of epigenetics and revealing new things, but there is also the potential for new diagnostic approaches through non-invasive blood sampling like liquid biopsies.
FLG: You mentioned diagnostics, but do you think these epigenetic insights could also be used to inform preventive medicine treatments?
Andy Sharrocks: Again, that’s possible. One of the beauties of doing epigenetics of the sort that we do is we use open chromatin accessibility assays. What that does is reveal not just individual genes, but also programs. So if you get a transcriptional regulatory program regulated by a particular transcription factor, it gives you an insight into how to intervene in that pathway. Obviously, targeting transcription factors isn’t as easy as targeting signalling pathways, but people are doing it which means it can be done. You just have to think a little bit out of the box to be able to do it.
FLG: Thank you. What novel research coming out of your lab are you most excited about and what should we be keeping an eye out for in the near future?
Andy Sharrocks: Something we’re doing at the moment, which is something I won’t necessarily talk about at the Tri-Omics Summit, is looking at transcriptional repression in cancer. It’s a less studied element of cancer, probably because it’s just more difficult to study. We’re looking at this in different ways, but one of the things we’ve started to look at is the role of phase separation in transcriptional repression, and then applying that to cancer. At the moment we’re currently not doing this in cancer cells, but we’re working up methodologies to look at repression and how phase separation might do that. In the future, this could be a completely novel way of treating cancer patients. Rather than trying to affect binary molecular interactions, if you can split apart phase separated condensates which are involved in key processes in cancer cells, it could open up new therapeutic angles as well as a field for new therapies.
FLG: That’s really interesting. I haven’t heard much about that at all. Could you explain in a little more detail what phase separation is?
Andy Sharrocks: So normally you think of proteins and different molecules interacting through the lock and key hypothesis where you’ve got one thing binding to another, and then another, and another as you build up complexes and assemblies. Phase separation is something slightly different, as you have multiple low affinity interactions which bring things together to hold them in a more liquid environment. It’s not a solid, but it’s more of a continuum of things driven by low affinity interactions as opposed to the high affinity lock and key hypothesis. These interactions are driving compartmentalization of various things in the cell, leading to the formation of granules and the nucleolus, for example. But at a smaller level, things like enhancer regions where allthe transcriptional regulatory apparatus is in there. And then similarly, repression where you’re shutting down gene expression. You bring things together in these localised conglomerates, these phase separated areas, to control gene expression.
FLG: Did you anticipate something like alpha fold or other proteomics tools coming into play for that field of research?
Andy Sharrocks: Probably not. So, the issue with the issue with alpha fold is it deals with globular domains. What alpha fold can’t deal with at the moment is the non-folded domains, because we don’t know what the structures are and these non-structured domains are what drive phase separation. Because of this, we need new insights into what these disordered regions are doing in proteins, and this can’t be readily predicted through comparing across different structures. What alpha fold does is it takes a structure and looks for something which may map onto that structure. In many ways, it’s the antithesis to Alpha fold. It’s the new frontier to understanding what these other domains and proteins do, which we’ve conveniently ignored over the last few decades.
FLG: Thank you, that’s really interesting. I just wanted to finish off with some questions about The Tri-Omics Summit. You’re going to be speaking at the Tri-Omics Summit in London; why are you looking forward to the event? What do you think people could get from attending the summit?
Andy Sharrocks: For me, the three strands are just to map onto where science is going and where my lab is going at the moment. You’ve got the multi-omics side of it, which we’ve discussed quite a bit in this interview, and multi-omics is important in understanding biological processes. You also have spatial technologies which are now emerging at an increasing pace, and keeping up to date with what those are and how they’re being applied is really important.
When you do single cell work and disaggregate tissues, it’s really important to understand how things come back together in situ, especially in cancer where you’ve got these really complex environments. This is something that we now integrate into our workflows. The other element is obviously the genomics of cancer, which people have looked at extensively in the past. We have to build on these with more functional ‘omic’ methodologies on top of that, and coming to the conference is going to give us an idea of all of those three things. Unfortunately, I can’t be in two places or strands at once and I’ll have catch up on one of them after the conference has ended on-demand.
FLG: Thank you. So, what will attendees look forward to learning from your talk?
Andy Sharrocks: It’s going to have two different elements to it. The first part is going to be about understanding the basis to cancer and where it comes from. I mentioned before about development going in reverse, and I’ll present some epigenetic data of open chromatin accessibility which enables us to cross-compare what happens in tumour evolution to human embryonic development. I’ll show you how the two things interrelate and how tumorigenesis looks like it’s going backwards into the embryonic camps, going back to a primitive state. In the second part, I’ll explain how you can go from this accessible chromatin opening to give you some insights into tumour biology, integrating RNA sequencing data on top of that to understand how regulatory elements work and how that then tells you new things about what goes on in esophageal cancer.
FLG: Thank you very much.
Watch more interviews in our series here.