DNA structural changes in cancer

Interview with Katie Reed, National Institutes of Health

Part of the show Individuality, and non-monogamy boosts chickadee cognition

The sequence of DNA letters in the genetic code spells out words - or genes - that act as instructions that guide cells on how to behave and what proteins to produce. The DNA itself is coiled up in a highly organised and compact way to enable cells to pack literally metres of it into a nucleus just a hundredth of a millimetre or so across. But just as you probably tend to keep the recipe books you rely on regularly within easy reach, and file away at the back of the cupboard the documents you refer to only once in a blue moon, cells also arrange their DNA so that the genes that are in regular use are right at the front of the drawer, together with regions of the DNA that can influence and control those genes. So, the question becomes: when a cell becomes malignant and starts to switch on genes it shouldn’t, does that arrangement change? Katie Reed, at NIH in Bethesda, Maryland, has discovered that it does…

Katie - So, as humans, if you stretched our DNA out from end to end, if you stretched it out straight, we'd have about six feet of DNA, and that has to get packaged into every single cell of our body. So every single cell has a copy of that DNA, and it's packed into a teeny tiny space, right? And what I found is really interesting is that the way that the DNA is packed up into cells is not random, because different cells need to use different parts of the DNA. For example, if you think about an immune cell, or a blood cell, or a muscle cell, they all need to use different parts of the DNA, and at different times, in order to, you know, be the cells that they are and function the way that they need to. And essentially, these different cell types will package their DNA differently so that different regions of the genome are accessible and available and to be used, and the parts that they don't need, are basically stowed away in the corner as like a sort of, like, do not use- 

Chris - Is this like in your sock drawer? You put those socks that you don't like so much at the back, or your underwear drawer, the ones that you don't like wearing at the back, and the stuff that is the kind of comfy pants for everyday use are at the front. Is it that the DNA gets organised, similarly?

Katie - Yes, that's a perfect analogy, yeah.

Chris - So what was the unknown?

Katie - So we know that different cells in the body have different ways that they package their DNA, but we know a lot less about how cells change their organisation of their DNA during diseases, and especially in cancer, we really know surprisingly little about how the genome changes during the disease. So the question that we wanted to seek out was when you look at cells that are undergoing progression into breast cancer, going from a normal healthy cell to a more aggressive breast cancer state, how is the genome sort of reorganised in that cell? How is the DNA restructured? And what impacts might that reorganisation have on what parts of the genome it's using?

Chris - So that organisation isn't immutable then? So if you've got a cell in a breast, for example, which is going to become cancerous, it can go from the organisation you see in a healthy breast cell into a completely different organisation, as though someone's ransacked the room, pulled all the contents out of the drawers, they're all over the room, so that different bits of the DNA can then start to be turned on, turned off, whatever the cancer wants to use.

Katie - Exactly. Exactly. That was sort of our hypothesis going in is that we would see these sort of structural changes because we know that gene expression changes a lot during cancer. Cancer cells grow more, they grow sort of out of control, and they use different genes than healthy cells, and that's part of what kind of drives the disease. And so our hypothesis was that these changes in gene expression and what genes are being used might be potentially accompanied by changes in how the genome is organised physically and in 3D space in the cell.

Chris - How did you explore that then? How did you actually look at that organisation? And how did you then answer the question, well, is that chicken or egg? Did the cancer do that and then it became cancerous? Or did that happen because the cell is cancerous and it's just very disorganised?

Katie - Right. Yeah. Great questions. Yeah, so our approach involved using this really cool cell-based model of breast cancer that mimics different stages of the disease. So we have three different cell lines. They're all called MCF10, and then A, T, or C. They all were originally derived from essentially a patient that didn't have breast cancer. And the first cells, the A cells, are not cancerous. If you inject them into mice, they don't form tumours. But the T line is sort of this precancerous state. It has an oncogene, a gene that causes cancer, more highly expressed in that cell line, and they cause sort of precancerous lesions in mice. And then we have the C line that is really aggressive. It's gone on to grow in mice, and if you inject that into mice, it causes, it always causes cancer and it metastasises into the lungs, so it's a very aggressive sort of state. What makes them cool is that because they're all originally derived from the same patient, they have basically the same genome starting out, and it sort of is the same genome that's sort of been changed and changed over time and is sort of been mutated to reach this breast cancer state. And in each of these three cell lines, we use this method called micro-seed that lets us essentially map the organisation of the genome and the three-dimensional structure of the DNA in the nucleus at a pretty fine scale resolution, and so this lets us see, you know, what parts of the genome are stowed away, what parts are being used. And within the parts that are being used, we actually get to see these like sort of loops, so interactions between genes and really, really distal, far away regions that might be actually acting on them. So a really cool thing about genome organisation is that regions as far as millions of base pairs away can functionally act on genes and change whether they're being turned on or off.

Chris - So do you mean that bits of DNA unwind and can bend themselves around and influence bits of the genome at a distance?

Katie - Yes, exactly. Yeah. So it's not so much about how many A, T, Cs, and Gs separate two parts of the genome, but if they're actually in close three-dimensional space, if they're near each other, if they're touching each other, they can sort of act on each other. And so we found with this micro-seed method, we basically were able to map the structure in non-cancerous, pre-cancerous, and metastatic cells. We also used genomic methods, genomic methods meaning we're looking across the entire genome. So we're not selecting in any specific way, we're not biassing for anything, we're just looking genome-wide and looking at how the structure changes. We also looked at gene expression and we looked to see what regions of the genome are basically active versus inactive.

Chris - And do you see consistent changes here, which map onto those different cell states, normal beginning to be unhealthy, frank cancer? And do those changes or those patterns explain why those cells behave the way they do?

Katie - Yes, yeah. So we do see really consistent and strong changes in the genes that are expressed at each of these stages. We also see changes in the regions that they're looped to, these really far regions that can be hundreds of thousands to millions of base pairs apart. When genes are connected to these distal regions, they tend to agree and help explain how those genes might be changing. So our data suggests that for about a quarter of the genes that are expressed differently between normal and cancer cells, for about a quarter of them, they seem to be influenced by these really, really far away regions. And for a couple of hundred genes, we actually saw that the structure itself is also changing. So changes in structure were definitely rarer. You know, it's not like every gene that's changing in cancer is due to a sort of restructuring event, but what we saw is that when you do see a restructuring event, it's very likely to also be at a region that's changing gene expression. So you asked about the chicken and the egg earlier about which one of these is causal, that's sort of the next step. We're not sure if the reorganisation is causing the change in gene expression or vice versa, but that's basically where we're going to be going next with this data, and we're really excited to dig into that more.