Executive Interview with Yesim Gokmen-Polar, Emory University School of Medicine

[0:03] Good afternoon, everyone. Today, I'm joined by Yesim, an associate professor of Pathology and Laboratory Medicine at Emory School of Medicine. Her research focuses on breast cancer with an emphasis on biomarkers and the understanding of resistance mechanisms at Oxford Global's biomarkers and precision medicine Congress, Yesim is presenting exciting work using spatial transcriptomics to uncover how the tumour microenvironment drives resistance to CDK4/6 inhibitors in ER positive breast cancer.

 So Yesim, thank you for being with us today, and I guess the first question would be, could you share with the audience a little bit about what drew you to breast cancer research in the first place, and how your career path led you to your current work?

[0:56] Thanks for the invitation first. Yeah, I'm a molecular biologist by training, and during my postdoctoral research, I worked on a project where we were able to validate or preclinical findings using patient tumour samples. And this experience was pivotal. I it truly sparked my deep interest in translational cancer research. I love it, because we can take the discoveries from the lab and bring them closer to real world clinical applications that I believe that that real impact on the patient lives.

So why I chose breast cancer? Well, it was affecting so many women, and I can see in my daily life, or even, as you know from the studies, regardless of their age, that was, really was the main point. It is not like a rare cancer, or anyone can get it no matter what. So therefore I wanted to contribute to this space and make difference in patients lives.

[2:08] That's great. Thank you very much. And what are some of the key biomarkers you found so far that help explain why some ER positive breast cancer patients do respond to the CDK4/6inhibitors while others develop resistance?

[2:27] I would say like most of standard of care CDK4 inhibitors. Inhibitors are also centred around proliferating epithelial tumour cells, so they are like a class of targeted cancer therapies that interfere with the cell cycle machinery, so specifically blocking the transition from the G1 phase to S phase the cells are accumulated in the G1 phase, so this is take care of the exponentially proliferating tumour cells.

But from our studies, from our clinical models, we know that despite the fact that those cells are taken care of and then you see like 80% of decrease in cell proliferation. However, there are some sub populations, we call them slow cycling cells, or drug persistent drug tolerant, persisters. So those cells are not affected by these therapies, because they are not growing exponentially. Therefore we focus on those, on those cells, and also later, we can see in our patient samples too, we focused identifying markers beyond proliferation, and we see that particularly markers involved in shaping the tumour micro environment very important. We found markers that link to immunosuppression and also altered metabolic states. So mainly beyond the tumour itself, the micro environment is very important, and we assume that in the future, these targets could be in combination therapies help new knowledge strategies to improve outcomes in er plus with breast cancer.

[4:18] Thank you very much. And when conducting this particular study, were there any particular challenges or limitations that you had to overcome?

[4:32] Well, the major limitation in those studies is especially the sample size, because when we are going through the you know, special technologies, as you know, they are still very expensive, and in your studies, if you can later apply those special technologies to larger cohorts, or obviously, if it can be incorporated in the clinical sample setting. I think that's a true validation right now, we are still having those challenges. You know, the sample size is it's so costly that we can't increase or sample size, and also the analysis, maybe the data analysis could be more user friendly for people like, you know, like pathologists, or, you know, just other biologists, molecular biologist type of thing. And another limitation, I would say, the resolution issue. So in our studies, we use Visium, but at that time, Visium HD was not launched, so we use Visium V2 which, as you know, it doesn't have the resolution like HD. And so right now, of course, the HD provides you a close single cell capability, but the one that we use  has limited resolution.

[6:10] That's great. Thank you very much. And I guess your last point kind of comes on to my next question a bit. So we mentioned Visium HD, and how do platforms like Visium HD and Cell DIVE help you characterise the tumour microenvironment?

Well, definitely, platforms like vision and cell dive, they are very powerful tools to allow us to study the tumour micro environment at the spatial contact and the higher molecular resolution, and the most importantly, they offer different and complementary insights. So as you know, Visium provides spatial transcriptomic data and at the whole transcriptomic level, so it can be used as a discovery tool, and then you can identify gene expression patterns. And of course, their spatial context with it. On the other hand, Cell DIVE is also providing you a multiplex imaging platform, not at the level as a high like probably full proteome, but it is quite good, up to 60 targets. Or you can definitely focus on the targeted approaches, like, for example, what we are using immune panels, and which is important, especially in immuno suppression. Therefore, it also gives you single cell protein level analysis, and you can look at the sub cellular level. So they both give different advantages. And I think the best, I think the field, goes also to this, combining these platforms, both at the RNA protein level, and having a multi dimensional view to understand what is really important to target are the best approach.

[8:09] Perfect. Thank you very much. And could you share an example of how

spatial data has shifted your understanding compared to, let's say, traditional bulk profiling?

[8:23] Yes, one clear example is, for example, the in on the context of immune markers, especially in ER positive breast tumours. So ER positive breast tumours are traditionally considered non immunogenic cancers because they are mainly like immune cold cancers, because you don't see the like in the triple negative breast cancer, the tumour infiltrating lymphocyte cells, and therefore these immune checkpoints are not very much useful in those cancers. However, what we see that we can see that it is not the tumour tissue itself, but it is the specific niches around the tumour. The tumour micro environment that we call is very important in providing an immune suppressive zone or niche. So therefore, with usually you don't see this kind of differences if you only look at the tumour tissue, but with the spatial context, you can look at around the neighbourhood. And then it is not only expression. Whether it's absent or not, it is definitely where it's located, special arrangement and proximity this to the tumour sub populations. So these are important factors that we can now analyze and go deeper about the tumour micro environment and incorporate those markers into or understanding the biology.

[10:04] Thank you very much. And I believe from our last conversation, you mentioned that you have received a grant to study immune models in triple negative breast cancer. Could you tell us a little bit more about this project?

[10:19] Yes, we received our one in collaboration with the GE Healthcare. The focus is to assess the risk of lethality in triple negative breast cancer. And then for that, we proposed, we will develop a novel multimodal score, risk score that integrates the radiomic analysis from of tumours from mammogram images and Potomac, that is the H&E analysis coupled with the generative adversarial network and AI tools. And then a third aspect is the spatial profiling of immune markers using cell dive. And then the idea is to incorporate all these models. It's a multi model integration, and to understand the risk of lethality, mainly the recurrence or metastasis within five years for triple negative breast cancer and develop this model score for this purpose.

[11:34] Thank you. Definitely sounds like an exciting project. And, yeah, sorry, feel free.

[11:42] No, it is. I think when we are talking of integrating multi omics, I think this is giving a different layer besides multi omics, RNA protein level, you can also integrate other modalities, like radiomics, Potomac, and then incorporate all these spatial profiling, then it will be more powerful to understand what is really driving the recurrence of metastasis.

[12:10] Very cool. Hopefully we get to hear about it, maybe not this conference, but an upcoming conference. It would be great to hear the progress of that project. And I guess a broader question, Yesim, looking to the to the future, how do you envision integrating spatial transcriptomics and protein profiling into clinical decision making?

[12:35] Well, definitely integrating these spatial technologies, both transcriptomics and protein [unclear] into clinical decision making would revolutionise the precision medicine, because we will understand better the neighbourhood’s molecular and standard architecture of the disease. So love to see they can be impact the not only diagnosis, not only targeted therapy selection, but also predicting the treatment response and resistance, saying that, of course, in order to have this, we have to address the several challenges that I briefly also mentioned, like to see their incorporation into precision medicine, the major issue is the cost and scalability, so we need to improve better models or algorithms that can make it more user friendly and that can also decrease the cost, and that helps to the integration of the clinical setting. And of course, we deal with the regulatory approval too, but I think the cost and scalability data integration and that will also help its approval once they see that they are more applied in the clinical setting.

[14:06] Perfect. Thank you very much. And final question from me is, we are very excited to be hosting you at Oxford Global's biomarkers and precision medicine Congress this year. And what would be the key message you would like to leave the audience with regarding your program?

[14:28] Well, I think we are not just looking at biomarkers anymore. We are looking at their context, their neighbours, and their role in the biological neighbourhood. So I think that's the new frontier of personalised medicine. So neighbourhood is very important. And if we can integrate these findings to the clinical setting as much as possible, that would impact a huge in the clinic setting, that that's what we are hoping for.

[15:07] Sounds great. Thank you very much Yesim, and I think that's all the questions from me. So thank you very much for sharing your insights and your work with us today, and we very much look forward to seeing you at the event.

[15:22] Thank you for having me and see you in the meeting. Thank you.