Semir Beyaz, PhD - Multi-omic Analysis of Human Endometrium and Endometriosis

Well, thank you so much. It's great to be here again. I first met the Endometriosis Foundation of America almost three years ago at a conference like this around that time. And it really changed my scientific trajectory. And I will show you a little bit about what we have been doing and what we are thinking and how we can work together to make an impact for the stabilitating condition. So last May, we embarked on a new journey and with Dr. Setskin and the foundation, we established a center, a unique experiment that hasn't been done at the scale that we are attempting to do at a basic science institution like Cold Spring Harbor Laboratory. So Coldspring Harbor Laboratory was established around 120 years ago. It's where initial work on genetics was done. The first hybrid corn was developed there. So if you like popcorn or corn, that's because of the genetics research was done in 1920s.

And then it became really a foundational piece in molecular biology where DNA and RNA and splicing and many of those fundamental molecular principles were discovered. So I came to Cold Spring Harbor Laboratory to study how genetics and environment get together and influence tissue health. And so fate and serendipity led me to study biology of endometrium. And I was fortunate to over a random barbecue that I was invited to, I met Sarper, who is the IT guy over there. And he told me, "What do you work on? " And I said like this and that. And also endometrium and endometriosis. He said, "Wait a second, what did you say?" I said, "Endometriosis." He said, "You have to talk to Dr. Stetskin." So I had a Zoom call with Dr. Stetskin and then we hit it off right away. He's one of the most inspirational people I've met in my whole life, committed for the cause.

And you all know him very well. And I was fortunate that I was able to convince the leadership of Cold Spring Harbor Laboratory to join in that vision that a basic science institution can commit their resources and foundational work to tackle endometriosis because it's not a fringe disease, as you all know more than I do. It is a fundamental disease that affects hundreds of millions of women worldwide. So that's what we are hoping to do, but we cannot do it alone. So we are very excited to partner with everybody here. Today, I'm going to tell you a little bit about how we are thinking about that problem, and it has a lot of parallels for the discussions that were done today. But as I told you, in my lab, we are really passionate to decode the complex relationship with the genetics and the environmental exposure.

So the environment, as you heard in the talks before, nutrition, microbiome, pollution, all of these things can impact. Also, stress, sleep, many things can be considered as an environmental factor, but genes times environment is how our tissues adapt either a healthy physiology or a disease state such as endometriosis. So historically, what my focus was studying to the deepest molecular mechanisms, how stem cells in our tissues become fit, how they regenerate when they need to regenerate an organ like the gut, like the skin or endometrium. So those organs have this process called homostatic regeneration. This is three to four days for the gut. This is every month for endometrium. And so the fitness of the stem cell decides how this tissue is going to maintain physiology. And if the physiology is there, then you have fertility, then you have healing, then you can recover from perturbations that you overcome.

But if the fitness is less or perturbed or altered due to either genes or environment, then you can have diseases like cancer, endometriosis, or scarring. And so stem cells are not doing that by themselves. We have, in our research and others, have shown that stem cells really sense the environment. They talk to immune cells, they talk to microbes. And so we are looking into intersection of these multicellular communication networks by diving into the molecular and cellular mechanisms. And I was so happy to hear a lot of talks from neurobiology today because neurons, immune cells, stem cells, those are pillars of a tissue, a tissue like endometrium, tissue like gut, but also stromal cells, endothelial cells. These are the components that you cannot isolate and pinpoint, hey, this is the cause of the disease. And this is, I think, like an emerging theme from this conference today that it should involve all of these components together so that you can decode how to promote healing or inhibit scarring.

And that will be my main focus. How do we decode that? How can we use information from human biology to then figure out actionable ways that we can develop either diagnostic tools or therapeutic tools or hopefully one day preventative tools? So I don't need to tell this to you guys. I think you know this better than me, but we really need to have a better understanding of human endometrial met cellular and molecular level. Mouse models are great. I know good friends here, they do amazing job with mouse models, but human endometrium is something that you cannot fully recapitulate in mice. And what we are hoping to do is learn from the human and understand the biological principles in human, which is not an easy thing because humans are genetically heterogeneous and the exposure complexity makes it almost impossible to figure out what those cellular and molecular features that you're identifying.

But if you learn from humans and test in mice in physiological conditions, then maybe you can go back and forth and do impactful discoveries. And I would be remiss not to mention all this amazing work that has been done so far in that field. And one of the pioneers in the field is in the audience, Elsie, at least here, not Elsie, sorry. At Jacksonab, they've done really great work. And also another person I've met at the end of our meeting is also doing really good work in defining these human cell states and both for normal endometrium and also endometriosis so that we can go back and forth to figure out what's the physiological cell state and what's the pathological cell state. And so we started going into this now almost around five years ago where we were able to receive some funding from Chan Zuckerberg Initiative.

Their initiative was focused on building human cell atlas. So let's map all the cells in a human body and probably not to your surprise, but to my surprise, there was no endometrium related proposal. So we said, okay, let's do something with respect to endometrium. And one thing that we wanted to highlight is that can we generate a representative endometrial cell outlast? Because in genomics, unfortunately, patients who are represented are mostly who can afford to go to a research hospital. So there is a disparity how many people are represented in human genetics and genomics studies. So we wanted to, this is from a AI generated summary, summary of our project. So we wanted to capture human genetic diversity, meaning that you have African, Asian, Hispanic, white, so different genetic backgrounds, and then look at both the gene expression program that can define a cell, but also chromatin program that can really give this kind of like from the hardwired, the real software for the cell.

And it's the way that you turn the genes on and off. And the measurement that we do here is chromatin accessibility. If your chromatin is accessible, your genes are most likely be utilized and either at that time are being utilized or poised to be utilized at some point. And so this is really important when we are trying to understand cell states that can explain us how a cell is producing certain proteins or mRNA molecules and they are facilitating healing or associated with physiological states and versus in a pathological condition, why do they not function well? And in the context of endometriosis, our paradigm is to look at healing versus scarring. And so we spend a lot of time actually to optimize the reproducible method where you can get tissue from endometrium, bank it snap freeze so that you don't get batch effect, isolate the nuclei and collect both accessible chromatin and mRNA.

And it took us like two and a half years to optimize that to become reproducible. But once we were able to optimize, this is the multi-institutional team that helped us. We have really good collaboration with New York Genome Center and Northwell Health. So thank you, Northwell physicians for your support. Probably some of those specimens came to your operation room. And originally Brian was a technician who really worked very well with the New York Genome team to establish that. We now shared this protocol with the community. So if anyone of you want to utilize it, please let me know. Very happy to help out with that. And so Reese, who's in the audience, joined us last year to tackle all these ambitious projects that we have in mind to tackle that. She's right over there, and this is her picture. So Reese came to us from Berniger Ingelheim from a pharmaceutical company.

She was doing her PhD there. And I was so fortunate that I met Risa at another conference at Keystone stem cell biology meeting, and I told her about this idea that, "Hey, we'll do something big about endometriosis. Do you want to come?" So she declined all of her fancy offers and she came to Cold Spring Harbor Laboratory and now she's helping us to push this to finish line of very soon. So what I'm showing you here, we did genetic sequencing to determine not just self-reported ancestry, but genetic ancestry. And this is important. I'm not going to go into the social construct of genetic versus race, but it's important to have an inclusive and representative map so that we can know that what we have in hand actually looks like a human. If you just have one, then it is like you're doing all of your experiments in bulb C mouse and not using BLACK6 miles, which is not very rigorous.

So our hope is to use this genetic heterogeneity to begin with is to create an inclusive atlas. And down the road, are there any specific genetic lesions or loci that may or may not associate with certain cellular states? That's our next goal. But for now, we are not comparing any ancestry with one another. We are just building an inclusive atlas that encompass all these different ancestries. And so we did some, this is like a clustering that you do when you get the multidimensional data. So you have all of your usual suspect cells, epithelial cells, endothelial cells. We couldn't capture that many neurons, unfortunately from that, but what we were able to do now is also integrate to different demographic and physiological states. So our 62 patients were from premenopausal and post-menopausal women, and the pre-menopausal ones were from different cycle stages. Unfortunately, we are not powered enough to go into these specific categories and make cross comparisons.

Hopefully, in the next few years, we will be able to get to that granularity, but this is just to show you, we are able to define endometrial cell states based on what genes they produce and what kind of genes have accessible chromatin, meaning that what are the transcriptional programs that are utilizing that hardwire and running that software to make the cells the way that they operate. And so for your relevance, what kind of use that dataset can have, you can basically link up from this amazing resource that exists in endometriosis through genome-wide association studies and say that, okay, what are those cells that I'm now identifying from this RNA and chromatin accessibility-based analysis imply some suspects, right? So in this analysis, we were able to see that mesenchymal cells, parasites, smooth muscle cells, stromal cells, and endothelial cells. I'm sorry for the resolution. It didn't look like that on my computer.

But this highlighted that the gene expression program and chromatin accessibility program that we define may be associated with some of those GWAS based studies in those compartments for endometriosis. And so this is correlation. How do you test the causality? You can look at those loci and you can develop models and methods such as like in organoids, we are trying to do that to test the significance of these association. The next thing that we are able to do now is, let's say you define the cells. You define, let's say, mesenchymal cell or epithelial cell that has potential to be a cell of origin for endometriosis. Then you can ask, okay, what drives that cell state? So from the gene expression itself, you go to a thing called transcription factor. Those are the factors that drive cell identity, and you can predict transcription factors that drive those cell identities, and then you can go to human-based organoid systems or develop some mouse models and then assess causality of these predicted factors, whether or not they are actually driving that cell state.

And when you define a cell state that is relevant for healing or scarring program, and if you define the factor that is responsible for this process, then you have a new therapeutic target. So cancer biology, neurobiology utilized that approach very well over the past 30, 40 years. Oh, okay. Yes, I hear it. Yeah, I see it. And for us, we would love to get endometriosis lesions and grow as organoids, but cellularity is very limiting. So we started using normal endometrial organoids, changed their microenvironment, changed their stiffness and the media conditions, and we pushed them towards endometriosis like states. So here, this is a single cell analysis looking at these human endometrial organoids grown in two different conditions. When we grow them in endometriosis-like conditions, you can see that they upregulate markers that associate with invasion inflammation. So those are the cell states that emerge by changing just the environment.

And so now we are using this system to test causality of potential factors, or maybe if you have some favorite drug target, we can test those in those organoid conditions. We want to keep the genetics and the environment in mind while we do that. I didn't talk to you about microbiome today, but in my lab, we are really passionate in understanding how microbiome, both in the gut and in other parts of the body can influence communication between stem cells, immune cells, nerves, and other cells to facilitate physiology in normal setting, and then how in pathology like endometriosis things go wrong. And with that, I have one last second and I'm done. Thank you so much.