Defining the Evolution and Dynamics of CLL Transformation

Transcript:

Dr. Erin Parry: Thank you for the opportunity to be here today with you all, virtually. My name is Erin Parry, I’m at the Dana-Farber Cancer Institute, and I’m going to be talking about my CLL Global Research-funded project on defining evolution and dynamics of CLL transformation.

 

To start, what is transformation? Transformation is, on a basic level, a change in cancer cell type. Chronic lymphocytic leukemia, as you are all aware, is a typically indolent B-cell malignancy, although it can have a very variable disease course. And occasionally, for small subsets of patients, one of those CLL cells undergoes some dramatic changes that make it transform or become a different cancer cell type. And when this happens, this transformation is to a different kind of an aggressive lymphoma. We call this a Richter’s transformation, or sometimes even a Richter’s syndrome.

 

Why is this important for us to know, or understand, or study in the laboratory? Well, again, as many of you are aware, we have really made a lot of progress in CLL over the past decade or so. We have a number of really highly active targeted therapies, and others on the way, in clinical trials.

 

On the other hand, when we think of the transformation, this is a disease that typically has not responded as well to the therapies that we use for other types of regressive lymphoma. And therefore, there’s a lot of research that’s needed to understand why it’s different, and how we can better target it. And there’s a lot of, also, exciting clinical trials going on in this space, so that we’re hopefully starting to move the needle and make a lot of progress.

 

When I talk about this idea of transformation, I’m mostly talking about the type of transformation that’s most common, which is when CLL goes to a large B-cell lymphoma. But I do want to acknowledge that there are more rare types of transformation that I won’t be referring to today.

 

As we’ve made progress in understanding transformation, I like to show this slide because I think it highlights a lot of the work that’s been done as a research community, just in the last five years.

 

So, we’ve started to be able to unlock, layer by layer, different insights into what is going on in terms of transformation. Understanding things like the genetics or epigenetics, which is sort of the packing and organization of the DNA, outside of mutations or changes in the DNA. Immune microenvironment stuff, and also, understanding how we can study these things outside the body using model organisms or different functional pathways.

 

But why has it taken us so long to get here? We are supposedly in the age of genomic medicine, where we’re able to get sequencing in a rapid fashion. Many of you have even probably had next-generation sequencing from your physicians in the clinic. So, why do we know so little about transformation, given that these tools are readily at our disposal? And I think there are a couple unique features that made this a harder problem to crack until recently, and something that we’re still learning a lot about.

 

One of these has to do with the idea of sample acquisition. So, to really properly understand transformation, we really need serial samples to see how these cancer cells change over time. We can’t just have a single state; we need both states present. So, we really need samples from the CLL cells as well as the transformed cells, to understand what’s dynamically changing between the two. This relies on us having robust biobanks, and individuals with CLL and Richter’s transformation who are willing to share samples, and have these samples stored appropriately, so that they can be accessed by researchers.

 

In addition, when we think of biopsies that we get, they don’t tend to look as much like these simple cartoons that are on the slide right now. They tend to look more like this, which is that an average biopsy, which usually is a lymph node or bone marrow tissue, is usually a jumbled-up mixture of different cell types. There can be normal cells, normal healthy immune cells, admixed with CLL cells and Richter cells, all in the same biopsy.

 

So, if we’re looking at a single time point of a transformation, there are usually many cell types present. And we’re actually dealing with this issue of two different cancers, which complicate a lot of the traditional pipelines that are used for these types of analyses.

 

To highlight the first, I think, key finding, is collaboration and sample banking are really key to being able to study this problem. And we’ve taken the approach of using some of these annotated biobank samples, looking at normal samples, CLL samples, and Richter samples, in parallel with sequencing. And being able to apply this sort of detailed, long list of different computational tools, to basically get to the end result of being able to separate the CLL DNA changes and the Richter’s DNA changes out.

 

You’ll see that I refer to these, on this cartoon, as clones, and that is because cancer cells are really clones of each other. They look identical. And within a given cancer, we have these different subpopulations of different identical clones, as cancer itself is always constantly growing and changing.

 

So, what does this look like, and how can we then integrate this information? We do something that might be familiar to those of you that might do genealogy or know things about family ancestry. And that’s, we try to make a phylogenetic tree. So, we’re trying to put together these different cancer cell clones, and order them, and understand how they’re related to each other. We’re basically trying to reconstruct a family tree for an individual transformation event.

 

And so, you can see an example of this here, starting from something that is normal, to the earliest CLL cells that we find. And then, we can understand that some CLL cells, now progressing to the lighter green, ultimately are the ones that give rise to transformation, which is highlighted in the shades of pink and purple. While other CLL cells, or branches of the tree, sort of diverge, and are more like cousins to the transformation. They’re not directly related to the transformation but are evolving in a different direction.

 

And so, we apply these kinds of approaches to every set of samples across an initial cohort of 50, and then in another validation set of 45, as well as an external cohort of 14, so that we were able to look at more than 100 different samples in our initial study. This allowed us to get some key insights into some of the different molecular or genetic alterations that were happening. These are the type of genetic alterations that are different, that set cancer cells aside from those of normal cells.

 

And we could find that some of these impacted some of the common pathways in CLL. Some, for example, like NOTCH mutations, were early CLL events. Others, such as those impacting DNA damage and some pathways of MEK activation, tended to happen in those slightly later-evolving CLL clones that were getting closer towards transformation.

 

And then, around transformation, there were a lot of new events occurring. Many of these involved in pathways that allowed the cells to escape from the immune system. Pathways that allowed the cells to suddenly start dividing more than they should and impacting those breaks that are on the cell cycle, and on cell division, typically, as well as things involving inflammation, and as I mentioned, epigenetics, are the things that aren’t encoded in the DNA, but are important in terms of regulating the DNA, and how it’s organized.

 

And this correlated, too, with what we call different regulatory pathways. So, moving from the level of DNA, to thinking of how those DNA genes are actually expressed. And this led us, again, to these similar pathways of inflammation and cell cycle, as well as some pathways of altered cell metabolism.

 

And this data was not only seen in our study, but also a couple other studies that were done around the same time. Lending a lot of robustness and providing us with a really unified understanding of some of the changes that are going on during this process. But there are still a number of open clinical questions. And as we strive to make things better with individuals with CLL and Richter’s constantly, here are a number of the areas that we’re trying to address with our current research.

 

First, what’s the diagnosis? If you know of anyone who’s undergone a biopsy for ruling out transformation, or ruling in transformation, it can be a complicated process. Not only do we rely on specialized scans to try to target the biopsy, and it’s not always straightforward to be able to get the tissue that we need. But there’s also a process of the pathologists, which are the doctors that look under the microscope, being able to look at the tissue and tell us that, what we are actually seeing.

 

And so, there’s a spectrum of the way that CLL even appears. And being able to diagnose it with infiltration of large cells, which is what we need for transformation, is not always straightforward. So, if we could get to a place to aid in that diagnosis, make it simple, more streamlined, and perhaps even non-invasive, I think this would be really impactful in the clinic.

 

In addition, thinking of who is going to get transformation. It’s only a small subset of our patients with CLL that ever experience a transformation event. And if we could predict who is most likely to have that happen, we could reassure the vast majority of patients that that is really unlikely for them, based on their individualized risk profile, and perhaps follow those patients differently that are at the most high risk.

 

And leading along those same lines, when does transformation happen? Is there a certain point at which a CLL cell gets committed along that pathway? And if that’s something that we could detect and intercept, or know when it’s going to happen, then we could perhaps intervene with therapies before people get sick. And perhaps divert these clones away from that pathway, so that we could actually prevent it from happening in the first place.

 

In order to start seeing how we could get to those places, the focus of my CLL Global Research, currently, is looking at many of these single cells and trying to understand patterns of genetic evolution better. We started by looking at samples as a whole, and now that our technologies are getting better, we’re looking with deeper and deeper lenses to be able to see things that we were missing in our initial studies. Because we couldn’t look with a big enough magnifying glass, so to speak.

 

So, we’re looking, now, at single-cell DNA and RNA sequencing approaches. So, we’re able to see the genetic alterations, or the gene expression changes, or the DNA packaging changes, within a single cell. And we can use this information to help us similarly reconstruct those family trees to understand cancer cell clonal evolution, and understand how, in some patients, that CLL cells could transform into a Richter’s transformation.

 

And so, we’re taking this approach of looking at serial samples. And so, you can see in this cartoon format, you see a cartoon of serial samples where you see different color cells. So, the changes of different cancer cell clones over time. And we’re capturing these different unique populations in serial samples, and then asking, “What are the unique populations? What are their properties, and how are they related?” In other words, can we make that family tree?

 

And so, in our initial attempts at this, this is sort of early data, we found a couple of really unique things that allowed us to proceed further. One is that because these cell changes are so different, because there’s a fundamental switch in cancer cell type, they’re very different. And so, when we conduct sequencing using RNA, which is the gene expression, the cells really separate.

 

So, you can see these colored blobs. But if you can look back, you almost see two different-colored blobs. One of those large blobs made up of different colors is CLL, and the other multicolored blob is actually Richter cells. So, they separate from each other, which allows us to distinguish them readily by their gene expression profile. We also saw that there were some populations that were kind of on the fence. These looked like a little bit of an in-between genetic step but also had in-between transcriptome profiles. This is pictured by Cluster 7 and Clusters 4 here, if you can see the numbers.

 

And then, we could also find some populations that were really small. So, things that would be below the resolution of what we can find in a bulk sequencing sample, where we’re sort of limited, and it’s hard to pick things up that are really rare. And we could find these really rare populations when we’re looking at single cells. But there were initial limitations. And these first attempts at this, that were published a couple years ago, were limited by resolution.

 

We also had an issue with what we call data sparsity, meaning, our sensitivity in detecting things was low. So, if they were there, and we found them, that was great. But sometimes, they were there, and we just weren’t able to detect them. We had dropouts or missing values. And also, initial assays were single data levels. And remember that initial figure where I showed you integrating different data layers? Ideally, we’d be able to get multiple kinds of information from one of these single cells.

 

So, how did we improve upon this? We first looked with a combination of looking both at that gene expression RNA assay, but as well as looking at some of the DNA accessibility or packaging. And we integrated these data with this genetic tool called Numbat-multiome, that was developed by some of our computational biologist collaborators. We also did genome sequencing in parallel to make sure that we were on the right track.

 

And what we did, and how we were able to integrate these, was to take advantage of little variations that we all have in our DNA. These are things that we inherit from our parents and aren’t related to cancer. They’re normal benign variants that we all have, that are some of the things that make us different. And by looking at these different things, it allowed us to be able to detect how copy number aberration went wrong in the cancer cells compared to an individual’s normal cells. And so, allowed us to detect things and to reconstruct these trees. I’m going to show you an example of this now.

 

In each one of these different-colored circles is one of these cancer clones, or a stop on the tree. And you can see that the black lines that are now connecting these is an example of one of those trees. Where you can see that there’s an original clone that gives rise to a side branch that’s green, and then a side branch that’s blue, and that blue side branch gives rise to the transformation clone in pink.

 

And now, when you look at these really complicated plots with lots of dots that are next to them, these are actually showing, from one chromosome all the way to the end of your chromosomes, the copy number profiles. And so, starting at the top, there’s very little copy number. And as we move towards the bottom, we see more and more copy number, which is shown in green, in red, and in blue. And these copy number alterations, including complex ones, are ones that help define or mark these clones, and allow us to trace them. And then, we can, in turn, look at what is unique about these clones, and when they’re occurring.

 

And what we found is, all of these clones were pre-existing, for instance, before therapies start. But the Richter’s clone really emerged and newly developed from one of those pre-existing clones after therapy started. And you can see a little bit of how this looks on this map. This is another cartoon version of this trajectory, or evolution, or tree, occurring in real time.

 

And then, also, we are able to infer or determine which transcription factors. These are like master switches in the cell that help turn things on or turn things off. And in this particular case, we could see how sequential changes in these was really helping change that global signaling and behavior of the cell and helping it really reprogram into a different cell type.

 

This started giving us a lot more information. So, we’re on the right track. But how does DNA sequencing compare? I’m telling you about all these other data layers. But what if we go back to the DNA layer? And the bottom line is, we actually have more evolutionary stops on the tree. So, using DNA, and being able to do sequencing of the DNA level as opposed to those other assays, gives us a tree that has eight clones as opposed to four. So, even more steps in evolution, even more in between to see. But this is only DNA.

 

So, going back to that same question, how do we combine more data layers? How do we get DNA, and RNA, and things like that, in a single cell? And so, to do this, we’ve partnered with collaborators of ours at the Broad Institute, where they’ve developed new technology to be able to do just that. And so, we’ve designed customized panels and approaches that are sort of unique to being able to study CLL and transformation. And one example of this is on this slide, but we’re now able to get RNA and DNA in single cells.

 

So, this is just a single time point, for instance, where we’re able to pick up already normal cells in four different distinct clones of CLL. And now, we can link their clonal phenotype and clonal properties, so we can see each step or each change in DNA, how that reprograms the cell and allows it to behave differently. So, now, for the first time, we’re being able to really link cancer cell genetics to how they behave, and this is going to hopefully give us a great, deeper understanding into the process of transformation.

 

In conclusion, we’re working on these new tools and technologies, and being able to leverage them to better understand this process of longitudinal CLL evolution, including to transformation. And as part of our long-term goals, we really want to be able to work towards an improved diagnosis and recognition of transformation, to be able to intercept, and identify, and then even prevent transformation.

 

And ultimately, as a community, all of us researchers are really working hard to make sure that we can continually dedicate our efforts to improving outcomes and quality of life for individuals with CLL and transformation.

 

Thank you for your attention. I have lots of people to thank, but first and foremost, want to thank the CLL Global Research Foundation for their support, and for being part of this really wonderful community of scientists, and researchers, and physicians, all dedicated to improving things for CLL.