Scientist Stories: Biology Across Scales, Liver Metabolism from Systems Physiology to Cell Biology

Hello, everyone.

Greetings, and thank you for joining our workshop today. My name is Jennifer Lippincott-Schwartz, and I'm a senior group leader here at Chimilia. I want to welcome you to this workshop today, which is focused on integrating nutrition and metabolism across scales. Together with myself, I'm very excited to be working with Vamsi Motha and Rong Tian, who've helped organize this workshop, which is one of a series of workshops that will help Janelia define a new research area focused on 4-dimensional cellular physiology. And this is an area that we think is going to complement and expand on our existing research programs in mechanistic cognitive neuroscience molecular tools and imaging, and computation and theory. So the, the broad swath of workshops, which will include 10 altogether, are being organized by myself along with Ron Vale, who's our executive director here at Janelia, Nelson Spruceman, who's a senior director of scientific programs, and Janine Stephens, who's the director of scientific programs at Janelia. And I want to thank many others who've also provided helpful contributions in driving this process. So what is 4D-CP, as we call it? Well, we're articulating it as sort of the structure and function of different cell types and tissues, and our hope is to address how cells interact to establish organ function and multi-organ coordination for animal survival and reproduction. Now, this whole field of 4D-CP is a well-established one, but it seems to us at Janelia that we're at a point of time when major advances in this field could be accelerated by taking advantage of the unique research environment at Janelia and developing new approaches and datasets that will be enabling for researchers across the globe. So we feel that Janelia has an opportunity to contribute to this general effort in 4D cellular physiology by bringing together people with varied expertise and ideas, collaboration in a very focused, highly supportive environment. We envision uniting this endeavor not only with biologists, but also to bring on board physicists, doctors, chemists, engineers, computational scientists, theorists, as well as tool developers and tool users. In this new research direction, we'll be establishing a 15-year, 15-lab effort, which we think is going to be significant but modest, admittedly, from a global perspective. So in order to maximize our impact in this field, we've decided to hold several workshops like the one today that's focused on metabolism in order to learn about the state of the art in this area, as well as the conceptual and practical challenges that researchers face. So I'm very excited to be part of this effort, which I want to say Vongsouvanh Mutha and Rong Tian has sort of taken very beautiful control of in putting together really a wonderful group of people who are going to discuss various aspects of this field. So with that introduction in terms of Janelia's perspective on this workshop today, I want to now hand it over to Bongsi and Rong, who are going to tell you a little bit more about what to expect in the workshop as a participant. So, Vamsi.

Thanks, Jennifer. I want to begin by just welcoming all of you to the Nutrition and Metabolism Workshop. My name is Vamsi Mutha. I'm an HHMI investigator and a professor of systems biology at Harvard Medical School, and I direct a research lab that's based both at Mass General Hospital and the Broad Institute. It's been a lot of fun working with Jennifer and with Rong over the last couple of months in putting together this workshop. And I think I really, really want to thank all of you for participating because it is the middle of August, and so this is particularly precious time. So we really appreciate all the effort that you've put forward already and in the time that you're putting forward today and tomorrow as well. So myself. Um, we're really excited about today and tomorrow for a couple of reasons. I mean, first of all, we're just all so parched for any scientific interactions because of the pandemic, and so we're already opening with a credit because we all get to see each other, meet new people, and also reconnect with old friends from our fields. And so that's already a great starting point. But I think what's really special about this particular workshop is that By working together and by sharing, we actually have a chance to imagine the future of metabolism research. And this is something totally different than most traditional conferences where it's so backwards looking. But now we really get to look into the future and even help to try to shape the future. [INAUDIBLE] And at a very practical level, the hope is that some of the learnings from today and tomorrow may help to positively influence the 4D cell physiology program at Janeline HHMI. So those would sort of be the 3 broad goals from the— just a few, like, personal reflections on metabolism research. What I think is pretty amazing is that if you go back about 2 decades, about 20 years ago or so, I think what most people thought about when they hear metabolism research is either diabetes or obesity or inborn errors of metabolism. But over the last 20 years, on the one hand, there's brand new areas and facets of metabolism research that have sprouted up. So we have fields like immunometabolism and cancer metabolism that really didn't exist 20 years ago or so. On the other hand, if you go back 20 years ago or so, what we knew was that so many common and even degenerative diseases, some of the best ways to have an impact on those diseases was through diet and through exercise. 20 years later, we still know that diet and exercise are great ways of intervening on common degenerative diseases, but we still don't know too much about mechanistically how this is happening. And so it's a real privilege, of course, to be able to do science. And I think especially in the world of metabolism, There's so many opportunities right now to focus on the very basic— [NOISE] human disease. So these are some of the reasons that we're so excited about this workshop, and it's gonna be a little bit like no other conference in the past. Over the next 2 days, we're gonna have a total of 24 talks. There's gonna be a lot of discussion that's been pre-programmed over 7 sessions total. And we ask that if any of you have any questions, please type them into the Q&A box. And again, I want to thank everyone for participating today, and let's just have a lot of fun over the course of the next 2 days. So with that, now what I want to do is introduce the chair of the first session. That's Dr. Chris Newgard from Duke University.

Thanks, Bumsi. Let me— I'm sharing my screen, I hope. Can someone confirm that I am been successful?

Yes.

Yes.

You look good.

Okay, great. So, today we're going to— the title of our session ostensibly was Liver, which is kind of a block-like title. So, instead we're going to be talking about the Liver and Metabolism from Systems Physiology to Cell Biology. And in an organizational meeting on Friday, we've swapped up the order a little bit. I'm going to lead off and I'll be followed by Christine. And we'll be talking about the more systems physiology types of issues related to metabolism, and then we'll move into single-cell and even subcellular aspects of metabolic control as we move to Heidi McBride and Daniel Feliciano. My own comments are going to be sort of a little history of the past 12 years of my life, tracing metabolomic signature of cardiometabolic disease from the liver to other organ systems, and what I really want to emphasize here is the importance of understanding the participation of multiple organs in metabolic homeostasis and progression of cardiometabolic diseases. So, when we think of a classical metabolic disease, of course, type 2 diabetes, but really when you think about control of something as seemingly simple as circulating blood glucose concentration, the first thing that you have to acknowledge is that this is controlled by multiple organ systems. So, the gut, the islet beta cell, the islet alpha cell, hepatic glucose disposal and production, neurotransmitters, insulin action in the periphery, in the muscle and adipose, reabsorption of glucose by the kidney, and sensitivity of the adipose to insulin, not only for control of glucose metabolism, but to suppress lipolysis. And it really takes a constellation of dysfunction in all of these tissues to contribute to the development of hyperglycemia. So, we really have to think about multiple organ systems. And the fourth dimension in, uh, Janelia's quest to understand 4D physiology is time. And there's a temporal aspect to the development of type 2 diabetes that's important, too. Insulin resistance is an early event, but then what we see is at first the beta cell responding to insulin resistance by an increase in insulin secretion, but gradually losing the capacity to secrete insulin, which heralds the onset of hyperglycemia and diabetes. So, we really have to think about systems physiology when we talk about metabolic homeostasis. Now, the other background comment that I'll make, and I'm sure this will be covered as well by other speakers here, is in terms of the toolbox. Metabolomics is a growing tool that we use for understanding metabolic regulation. But the point I would like to make here, which I know will be echoed in other talks in this workshop, is it's really no longer sufficient to take snapshots of the concentrations of a large array of metabolites, whether you're using targeted methods with internal standards, stable isotope internal standards, or various non-targeted platforms, you really need to link this to a dynamic measure, which you can get by using stable isotope-labeled fuels and tracing the heavy atoms to downstream metabolic products in the pathways in which the substrates engage. And I really think there's a big opportunity and almost a necessity to bring together the growing capacity to measure metabolites in a static fashion with metabolic flux. In my brief example here of things that have interested me certainly highlights that connection. All right. So, no one that knows me will be shocked that my example is one of branched-chain amino acids. Some years ago, we either discovered or rediscovered, you know, if you think about Phil Felig and George Cahill years ago, but we were using metabolomics to look at a very simple comparison of obese insulin-resistant to lean subjects, and we found that the metabolic signature that most strongly differentiated those two groups was not lipid metabolites, as you might expect, but rather branched-chain amino acids and related metabolites. And the particular ones that, that show up repeatedly in principal component factors when we do this kind of clinical survey of different cohorts is the three branched-chain amino acids, glutamate, which is the byproduct of branched-chain amino acid transamination, 3-carbon and 5-carbon acylcarnitines derived from oxidation of the carbon side chains of branched-chain amino acids. The branched-chain keto acids and the aromatic amino acids, which may seem odd but are probably in this signature because they're taken into cells by the same large neutral amino acid carrier that transport the branched-chain amino acids, LAT1. And what we found is that this group of metabolites are strongly associated with insulin resistance and type 2 diabetes. At baseline predict who will improve in obese subjects in insulin sensitivity brought about by weight loss through dietary behavioral intervention. At baseline, as shown by Tommy Wong and Rob Gersten in Boston using the Framingham cohort, these same metabolites at baseline predict a future risk of type 2 diabetes, and this same cluster of metabolites are extremely responsive to our most efficacious intervention methods. One example being bariatric surgery, as we studied with Blandine La Ferrere at Columbia. So every which way to Sunday, this metabolic signature associates with metabolic disease phenotypes. But what are the cellular and molecular mechanisms that drive the association of branched-chain amino acids with metabolic diseases? What's under the floorboards in terms of what's going on in regulation of metabolism? Are branched-chain amino acids actually involved in disease pathogenesis? Is there a causality feature to the increase in their levels? And is manipulation of branched-chain amino acid metabolism a viable target for treatment of metabolic disease? So this is about liver, and here's what, what we discovered about branched-chain amino acid metabolism in liver that relates to this whole program of disease pathology and pathogenesis. Years ago, Chris Lynch and others showed that the kinase that phosphorylates branched-chain ketoacid dehydrogenase, the rate-limiting enzyme of branched-chain amino acid metabolism, is increased, and the phosphatase that strips phosphate off is decreased in liver in obese insulin-resistant animals and in people. Okay? And so, that helps to explain immediately why branched-chain amino acids and branched-chain keto acids are elevated in the plasma in obesity and insulin resistance. But by using an unbiased phosphoproteomics approach, Paul Grimsrud and Philip White in our group were able to show that this kinase, BDK, also phosphorylates— a key core enzyme of de novo lipogenesis in the liver, ATP citrate lyase, and PPM1K dephosphorylates that enzyme. So this increase in kinase-phosphatase ratio that you get in obesity not only reduces the rate of branched-chain amino acid catabolism, but also phosphorylates and activates a key lipogenic enzyme to increase its activity and increase flux through de novo lipogenesis. And indeed, overexpression of BDK is sufficient to activate de novo lipogenesis, as we showed in animals with using metabolic flux of deuterated water to newly synthesize palmitate in the liver. And then with linking this to nutrition, my colleague Mark Herman at the Duke Molecular Physiology Institute has been studying for some, some years the role of high sucrose and high fructose diets on inducing a global transcription factor, ChREBP beta, which is an important transcription factor that activates expression of the core enzymes of de novo lipogenesis. What we discovered in concert with Mark is that ChREBP also upregulates the BDK kinase and downregulates PPM1K. So, when you eat— consume a high fructose diet, not only do you induce the core enzymes, the ChREBP target genes that encode the core de novo lipogenesis enzymes, by increasing BDK and suppressing PPM1K, you supercharge de novo lipogenesis by phosphorylating and activating ATP citrate lyase. And to the question, is there any therapeutic potential in this biology? Well, if we use a small molecule inhibitor of BDK created by David Chuang at UT Southwestern, or overexpress the PPM1K phosphatase, both of these maneuvers will cause reduced phosphorylation of BCKDH and increased consumption of branched-chain amino and branched-chain Keto acids, not only do you get a lowering of branched-chain amino and keto acid levels in the blood, but you also get an improvement of glucose tolerance and a 50% lowering of liver fat in an extremely lipogenic animal, the Zucker diabetic fatty rat. There does seem to be some therapeutic potential. Now, to the other question I posed a few minutes ago, Are there direct effects of branched-chain amino or keto acids on other tissues? And under these conditions of obesity and insulin resistance, when BCKDH is phosphorylated and inhibited and the circulating branched-chain keto and amino acids rise, we have recently obtained evidence that branched-chain keto acids activate protein synthesis, uh, to potentially trigger unhealthy cardiac hypertrophy, and structural remodeling. Specifically, Rob McGarra and Philip White, former fellows in our group and now the faculty members in our institute, used UC13 ketoacid, ketoacylvalerate, the ketoacid of valine, and showed that it's poorly metabolized to the level of the TCA cycle intermediates in the isolated perfused working heart. Instead, the main fate of KIV is reverse transamination to form valine. This pathway seems to be involved in activating protein synthesis and causing phosphoproteomic remodeling. There's a shift to reduced MAPK signaling, increased ERK signaling, and increased phosphorylation of 4ABP1 when you expose the heart to the levels of branched-chain keto acids found in obesity compared to those found in, in a lean state. So just the exposure of the heart to elevated branched-chain keto acids leads to a key regulatory event that activates protein synthesis. And of course, you can imagine if these metabolites are elevated chronically, that could lead to a situation of unhealthy cardiac hypertrophy and remodeling. Last but not least, I've studied the pancreatic islets for years and the mechanisms of fuel-stimulated insulin secretion using metabolic flux analysis. We recently found that, as in cancer metabolism, there is active reductive flux through the TCA cycle of glutamine via glutamate dehydrogenase and reverse TCA cycle flux. And we think the significance of this in the islet beta cell is that the citrate and isocitrate that you generate interacts with a cytosolic NADP-linked isocitrate dehydrogenase to make NADPH that activates a protein desumylase that triggers insulin granule exocytosis. Well, the fascinating thing here is that leucine is an allosteric activator of glutamate dehydrogenase. So, is this potentially an explanation for the early phase of insulin resistance, where there's a compensation of the beta cell to increase insulin output. Whereas if you drive that chronic increase in insulin secretion, perhaps by this mechanism—and we're investigating that now—you ultimately exhaust the beta cell, ultimately leading to insulin deficiency and onset of diabetes. And the last thing I'll say that's, I think, germane to the skills of Janelia I am very interested in the pancreatic islet. This is an organoid that you can pull out of the pancreas by enzymatic digestion and culture as a 3D complex in vitro. But the problem is that it's comprised of multiple hormone-producing cells, the glucagon-producing alpha cells, the insulin-producing beta cells, and the somatostatin-producing delta cells. Each of these cells respond differently to metabolic cues, and they also communicate with each other intensively by a paracrine mechanism. So, for example, glucagon and GLP-1 generated by the proglucagon gene in the alpha cell enables glucose-stimulated insulin secretion through the glucagon-like GLP-1 receptor and the glucagon receptor. And cyclic AMP in the beta cell. So, we really need to understand both the paracrine signaling and the differential responses of these cells to nutrients. My last slide, sort of raising issues that we can address in the discussion. Primary cell culture systems can be of limited utility due, for example, to horrible dedifferentiation that occurs in hepatocytes, or limitations for genetic engineering in a PrEP-like the isolated perfused working heart? What can we do to improve upon 3D organoid systems? Primary islets are a natural 3D model system, but it's difficult to decipher metabolic and cell signaling signatures of hormone-specific cell types. Could one potentially use generative protein engineering and imaging technologies to facilitate ID of metabolic and signaling circuits in the hormone-specific cells. And as I hope I've illustrated, there's a critical role in metabolic homeostasis in tissue crosstalk. And I would say that that's poorly replicated currently in the in vitro setting. Are there possibilities to do better in that regard? Okay, so with that, I'll stop and stop share, and I see, I think, no questions in the chat. Unfortunately, we blasted off, I think, in this session without telling everyone that the ideal place to put questions for the brief Q&A sessions that we have would be in the Q&A box. I think I've probably gone over slightly in my comments.

Chris, I'll ask you a question. Okay. Thank you. While people are coming into the chat box and typing, I think the field should really thank you for inspiring the branched-chain amino acid study in the past decades. And it's been really stimulating and very exciting for the field. Maybe you can take this opportunity to address the future, and especially for the branched-chain amino acid research to come out of the obesity and diabetes and go to other fields such as exercise and cancer. And also, for most of us are studying these metabolic pathways, we rely heavily on mouse models. I know some of the mechanisms that you discovered has species specificity. Maybe you can also address that. How do we do that in the future? Yeah.

Well, that's not a question. That's a soliloquy. But just very, very briefly, You're right. And let me elaborate on the comments that you made about species specificity. This ChRBP induction of BDK that I mentioned that we worked on with Mark Herman, the response element for ChRBP is present in the gene— in the genes for BDK in all mammalian species except the mouse, which is interesting. So We do find from time to time that the mouse isn't a great model system for metabolic disease, and I think that there are missing elements of circuitry, such as the one that I just described, that we have to be careful of when we think about the mouse as our only model of trying to model something complex like metabolic disease. And I think I'll stop there. I think, I think it's time, uh, actually to move on. Vamsi, you're shaking your head. Would you agree in terms of time?

I think that makes a lot of sense. And then we do have a, a group Q&A built in as well. That would make a lot of sense. Thanks, Chris.

Thank you. And I'm going to go ahead and introduce then Christine Desrogeurs from University of Montreal. And, uh, Christine is putting your slides up. Thank you.

So, okay.

So, hello everyone.

It's a real pleasure to be here today and participate to this special workshop. Thank you for the invitation. And I'll be mostly focusing on the role of liver and the lipidome. So we'll switch a bit in terms of subject. So, and I'll start as my first slide just to, with a short statement about my background so you better understand my perspective. I'm a biochemist by training, and I've always had a focus on my— of my research on metabolism with the underlying premise, which goes in the line of Chris Newgard, of metabolic alteration being an early biomarker and actor in disease development and target for therapy. Even though I'm in a liver session, I'm not a liver expert, but I did initially start my career studying the dynamics of glucose and ketone body metabolism in perfused heart using C-13-labeled isotopes. So, I did use all the toolbox that Chris was mentioning, and with mass spectrometry. Subsequently, as an independent researcher, I've transposed this technique to the working heart, mouse working heart. But more recently, I've been transposing my research in a more global approach, and in a more clinical setting through the application of metabolomics and lipidomics.— and more in a biomarker discovery approach. As director of the metabolomic platform at the Montreal Heart Institute, I've been collaborating with many researchers, and I have been working in many diseases beyond heart disease. So, since we were asked to be like asking high-level questions, I would say my first study is why would Janilia Working on 4D cell physiology, want to study metabolism. I think it's all in the statement 4D, you mentioned temporal dynamics, not snapshot. So, in relationship to this statement, I'm showing here my favorite illustration of the different level of regulation of cell function, which I've adapted from Joanne Kelleher publication, who's also participating in this workshop. This slide shows the interrelationship between all these levels of regulation with the, the omics disciplines that are used to study them. And among all these regulatory levels, the metabolome is definitely the most dynamic one. Metabolite levels are changing in seconds in response to environment or prevailing— the prevailing— and it's close to the phenotype. In addition, as was suggesting the slide with the orange line, metabolites will also feed back on the other different levels, for example, through covalent modification of protein, changing their function, but also changing gene expression, for example, through methylation. So for me, the second maybe high-level question is why study liver metabolism? I would say that liver is a metabolic hub, as you all heard from Chris in the previous presentation. It is a source of circulating energy, nutrient, signaling metabolites, but also toxic metabolites for other tissue, and its dysregulation leads to impaired body function. I don't need to tell you a lot about the founding principle. You all know from your first-year university textbook how liver is important for glucose homeostasis. Ketone body synthesis, ammonia detoxification. But what I would like to talk to you about is how the use of metabolomics and lipidomics have been changing a bit the way we look at the metabolism and how we have learned a lot about liver metabolism. You've already heard about the favorite molecule of Chris, branched-chain amino acid. I will talk more about specialized lipids. So it goes beyond what we know about fatty acids, cholesterol, and lipoprotein. And I'll focus as an example on ether phospholipids and plasmalogens. So why did I become interested in ether phospholipids and plasmalogens? This all comes from a study published in 2019, in which we applied a non-targeted lipidomic approach to try to better understand the metabolic perturbation partially responsible for the clinical manifestation in patients with an inherited mitochondrial disorder, which is Leigh syndrome French-Canadian variant, or LSFC. So, LSFC is due to a mutation in LRP-PRC, leucine-rich pentatrichopeptide-containing protein, which was identified by Bence Nitta, one of the organizers, and with whom we have collaborated in this study, and result in a defective oxidative phosphorylation, principally complex IV. LFC patients show a wide range of clinical manifestations including hepatic and neurological involvement, and they die very early in life. And specifically, we applied a targeted lipidomic method that we have developed and validated in plasma, which enables the measurement of about 21 lipid subclasses, to the cohort of the LSF-C patient and matched control, but also to a mouse model, which is a knockout, a liver-specific knockout of LRP-PRC. So this slide only shows— is the only one showing really some results, and it's only that you gather better appreciate the type of data we obtain using this approach. So, from the analysis of the plasma of these LSFC patients versus controls, we were able to detect about 3,000, more than 3,000 features, of which 146 passed our criteria, selected criteria for statistical significance. And this is shown in the volcano plot here. And the red line indicates the our selected threshold. So on your right are all— this is expressed as a patient-to-control ratio. So on your right are all these lipid features shown as dots that are increased in patients. On the left are those decreased in patients. And the color refers to those that we have identified shown here in this table. And the Gray dots are those that are yet to be identified. So as expected, we did see increased acyl carnitine, long-chain acyl carnitine, shown as the pink dot, and also increased triglycerides, the blue dot. But on the left, those that were decreased were mostly those phospholipid ethers and also plasmatogenes. And this was, for us, a plasma signature of disturbed peroxisomal lipid metabolism. And peroxisome is often a forgotten organelle in lipid metabolism. And for those interested, you can see here the box plot from those data. So to make a long story short, why we were so interested in going further in this discovery is that the ether phospholipid has been linked to neurodegenerative disease and cognition problems. And these are characteristics of our patients. So we, and they are options for therapy. So this slide, to make a long story short and to stay within my time limits, this slide summarizes, let's say, the majority of our results. So in general, both the patient and the mice, the mouse model, show the similar plasma lipidomic signature. And this slide shows in blue those that would arise from the mitochondria, which are the long-chain acyl carnitine. But I'll focus on those in green because additional finding, which I didn't show, supported the perturbation of the peroxisomal lipid metabolism, as we did find very long-chain acyl carnitine, which are product from the peroxisomal metabolism, but also lower docosahexaenoic acid and lower conjugated bile acid, in addition to those lower phospholipid ether phospholipid. Also, in the mouse model, we were able to look in the liver, and we did see many markers of perturbed peroxisomal, for example, biogenesis. However, one unexpected finding— liver is often seen as a supplier of ether phospholipid and plasmatogen, especially those containing docosahexaenoic acid.. But in the liver, those were increased. And this led us to think that there may be an export problem of these etherphospholipids to the circulation. However, this remains to be investigated, as is the impact of lower circulating etherphospholipids and plasmatogen. What is the impact on other tissue? And of interest to us, the brain. But to learn about this, we need new tools. And so this is my last slide. What are the challenges in my view? About tools. So I would say I've listed two points. I think we need to develop more approaches to do functional lipidomics at a more global level to assess the biological role and function of those specialized lipid subclasses that we discover with lipidomics. I've talked about etherphospholipids, but others would say, okay, there's ceramides, there's many other subclasses that are being beginning to be studied. It should be in liver, but of course, I think it's a global thing. We need to know more the interaction between liver and other tissue, and how these functions are modulated by factors like diet and disease. And in terms of dynamics, I think we should develop— where I think Janilia could work is develop a dynamic 3D imaging and mapping of this lipidome, really the 3D spatial localization in real time of specialized lipid subclasses at the cellular but also subcellular level, not only mitochondria, but peroxisomes, lipid droplets, exosomes, lipid-protein complex level in liver and other tissue. So I still have just a few seconds just to show an example of what can be done now with mapping for example, triglyceride distribution in human liver with nonalcoholic fatty liver disease by MALDI imaging mass spectrometry. This was done by my colleague Pierre Chaurant at the University of Montreal. And here we see when looking at section of liver from periportal or pericentral zone, which is part of the heterogeneity of the liver, that there is a concentration of these short-chain triglycerides. There is a different distribution of triglycerides in peripotency between these two zones. And this, I know, can be also done at the level of the brain. But as such, in my view, progress in this field would really greatly benefit from Janelia's multidisciplinary and collaborative approach to move forward. So thank you.

Thank you, Christine. Let me just say to everyone, I don't think this was clearly stated at the start of our— as we kicked off, these 10-minute talks have a 5-minute Q&A afterwards, and anyone, audience or panelists, can send in questions via the Q&A link at the bottom of your screen. As there has been nothing— let me just check the chat here, sorry. Okay, there, there is a question in the chat. The 20 years it seems to develop T2D may be too short, but could the seeds of insulin disturbance start in adolescence fueled by hormonal changes? I'm not sure that pertains exactly to— no, this, this talk about lipidomics. So, uh, let's, let's table that question and let me ask you, Christine. These complex lipids, what do— how do you view them in terms of their biological effects? Do you view them as signaling molecules? And if so, do you think there's a receptor biology that we understand incompletely? What's your take on how these things signal to cells?

Well, they can't— well, depending on the complex molecules, some of them can be signaling molecules. I'm thinking about, let's say, maybe pro-resolving lipid. But in terms of the plasmalogen, I think, for example, they are— they may influence the three-dimensional structure of protein. And what I'm thinking about is more like, say, for example, Maybe I have a better example with cardiolipin. When you think about mitochondrial function, there are complex interactions between protein and lipid that will determine what will be the respirosome and will make the cell do oxidative phosphorylation more efficiently or not. This is one example, but I think in the brain for the plasmalogen, They are part of the myelin structure, but we don't know exactly where in the brain and how they influence the neuronal activity.

But these are only examples. Here's a question from the floor from Isaac Harris. Accumulation of ether lipids has been linked to ferroptosis. Do you see any connections between ferroptosis and the neurodegenerative phenotypes observed with Leigh syndrome?

Yeah, that's an excellent question, and I've read that paper. This is very interesting. To be honest, I don't know for the Leigh syndrome, but it's certainly a possibility. But one thing I can tell you is that we have seen this low level of plasmalogen also in beta thalassemia patients, which are known to have like iron overload when they're treated. And these are the major changes in the signature of these patients. So certainly, let's say, I think I have only a small lab and I cannot do it all, but I really would like to share those, those observations.

And I think there's something there for sure. Okay, I think we'll take a last question. Thank you, from Maria Forni. Thank you for the amazing talk. Would it be possible to have image resolution to image modifications such as lipid oxidation as an arrow in the quiver of tissue lipidomics?

Yeah, I think, let's say, anything is possible would be useful to see the— well, plasmaproteins are often seen as being antioxidant and their decrease being a sign of having oxidative stress, but there are also other lipids that could be measured as a better sign. You can already look at, for example, for eodrexin nonenal adduct protein. These can be already looked at in diverse ways, but there may be others too.

All right, thank you, Christine. I think we're going to have to move on, so we're going to go now from sort of systems physiology view of the world that Christine and I have tried to present and dive down into individual cells. And our speaker is Heidi McBride from McGill University. Heidi, whenever you're ready.

Hi, let me share my screen. Here we go.

Let me go here.

Well, first, um, Thank you so much to the organizers, to Bamsi and Rong and Jennifer for the kind invitation to join this workshop, which is really, really exciting and challenging. So I'm going to talk to you today about how my challenge is to mechanistically integrate organelle behavior into metabolism and signaling across scales, which is basically the topic of this whole thing. So my lab works really on the fundamental mechanisms of mitochondrial dynamics, which is mitochondrial fusion and fission and how they bud vesicles and all of this kind of stuff, historically in kind of a reductionist approach. But, you know, the last 10 years have really, you know, shown everyone that the motivation of mitochondrial dynamics is primarily through these contact sites with other organelles, with the ER shown here in purple and peroxisomes in yellow. And the objective is really to try to integrate the metabolic flux and the function of these organelles and understand how the dynamics, the positioning, the regulation of the cell biology can impact be impacted by metabolism and impact metabolism in the other way. But that's a real big challenge. So to dissect the mechanisms of dynamic metabolism, we have to think about, well, what metabolites are we going to study, which signaling pathways. And, you know, a challenge in the contact site world is you may have grown up in mitochondrial dynamics, but now you have to know about endosomes and peroxisomes and ER. And so there's lots of growth challenges there. And then, so what model systems are best to do this? To allow you to also move across scales. So this is a liver topic, and I'm going to talk about how liver is really a highly informative and very important model system in cell biology. As you've just heard from the previous talks, you know, they're central for generating bile acids, which are huge signaling molecules across the body, and obviously in the gut. They're central iron regulators, and they control metabolism, as you've just heard very well. And for a cell biologist, they're The cells within the liver are highly polarized, which makes it really fun to look at. So we can study at all different scales, at the intercellular scale. We can— there's lots of work in the generation of functional organoids. I mean, as Chris said, primary hepatocytes are very problematic. You can also look at interorgan, and as you heard, inter-system scale. So I think liver is an important— is a very interesting and exciting place to start. So how did I get into that? So in this very short talk, I'm going to give you a snapshot of the current challenges in my lab that pushed me into the cell biology of liver metabolism, which is something I never expected to be doing. So I work on this protein called MAPL, which is mitochondrial anchored protein ligase. We made a knockout mouse. So MAPL is a mitochondrial and peroxisomal membrane protein with a piece of it in the intermembrane space and a cytosolic exposed ring finger domain responsible for SUMO and ubiquitin ligation. It's also called MUL1 in the literature. So, you know, it's localized to mitochondria peroxisomes. We've worked on it because it stimulates DRP1 and stabilizes ER contacts and cell death. We and others have shown roles in inflammation because it stimulates NLRP3 and RIG-I, and there's another bunch of people working on ubiquitination activity of MAPL working with Parkin in mitophagy. So coming into the project, our expectation was that MAPL sits in the membrane and acts as something to stabilize and coordinate signaling complexes in response to external cues to drive things like inflammation or cell death. So we make a mouse, and it was a surprising phenotype to us. So Vanessa Goyon in my lab has been working on this for a number of years, and the first thing is the mice were lean. So the gray, the gray line here is the MAPL knockout mice, and all these have been done in males and females. They're lean mice, and if you give them a high-fat diet, they don't gain any weight like the wild-type mice do. And later on, we realized that basically 100% of the mice spontaneously developed hepatocellular carcinoma. So this— I'm like, okay, so what's that? So how do we find the mechanism? So with BioID experiments and a whole lot of stuff I'm not going to show— this is a very short, short story— MAPL regulates bile acid metabolism, which was probably a nightmare for me. So it stimulates this transporter called ABCD3, which is a peroxisomal membrane protein responsible for the translocation of bile acid precursors into the peroxisome. Peroxisome, where the bile acids are then further modified to be released from the peroxisome and into the bile canaliculi. So we show that it interacts with this protein very strongly. It also stimulates it. This is from liver extracts. Stimulated ABCD3 is here, and if you knock out MAPL, you lose that, along with DRP1 stimulation. And physiologically, if you look at bile acids in the mice, we had a significant elevation of bile acids in the plasma and in the liver. And the transcriptomics of the liver told us that the enzymes controlling bile acid synthesis were greatly altered. So there's a downregulation of bile acid synthesis enzymes and an upregulation of detoxifying pathways. So when I come into this field, by mistake, I look at where are these enzymes. And of course, they're in the ER, the mitochondria, and the peroxisome. And there are two different bile pathways to generate it, a canonical and an alternative. And so this sets up a system for a cell biologist like me to think about contact site biology. How can— well, this must be part of our story. So you look, and so we think that mitochondria, ER, and peroxisomes, they act together as a functional triad, which is not probably news to anyone. And so this is an electron micrograph of mitochondria. There's the ER and the peroxisomes. They form these nice units inside the liver. Those units are maintained in our knockout mice, but the peroxisomes, as you can see here, are very enlarged, and we're trying to figure out all different elements of cell biology, but we are stuck because there's a huge frustration in the inability to monitor bile flux at contact sites. In fact, there's almost no studies on, on the cell biology of bile acids. Mostly it's considered to be regulated at the transcriptional level by retrograde signaling from the gut to the liver, etc., etc. So there's a big opening here for how to do this. So we were looking at primary hepatocytes, which I understand are somewhat problematic, And this is done by Aurel Besspatin, a postdoc in my lab, and it was helped with an undergrad, Maria Graham. And while we were looking for mitochondrial contact site interactions, most hepatocytes look like this. This is from a wild-type mouse, but we noticed that, or they noticed, that many of the cells, at least half of them, would have highly polarized peroxisomes, as you can see in these three examples. So if you zoom into a couple of these places, you can see they almost form like these rivers of peroxisomes coming towards the plasma membrane. And so it makes me ask a whole bunch of questions. So are peroxisomes known to be polarized? Is this going to become the apical bile canaliculi? Are there contact sites with the plasma membrane? Meaning the final step of bile acid release, is it just released into the cytoplasm and floats around, or do you have a cell biology underlying this, which is completely unknown? And what are the contributions of peroxisome biogenesis, for example, to the development of the bile canaliculi? And why do we know so little about paroxysmal behavior in vivo, which is what Christine was just saying? So these are just open questions. I have no answers, and we're continuing— Orel's very excited and interested in these kinds of pathways, and we need to plug this into more developmental systems to have a look. So last point, why do the mice develop cancer? So there's been a movement in the literature that high levels of bile acids can activate a protein called YAP, which is a growth protein, regulates cell size, it regulates growth, is a signal transduction molecule, and that this can promote liver carcinogenesis in a study done by David Moore back in 2013. Marina O'Zerial has been doing some beautiful work lately with liver architecture, has shown that bile can activate YAP during liver regeneration. So we looked in our mice and the YAP and the MAPL knockout hepatocytes looking at YAP1 translocation, from the cytoplasm to the nucleus, you clearly see there's translocation. It's quantified here. So if that's happening, even in the primary hepatocytes, we also see the same thing in histology sections of the mice. So one of the downstream consequences of that is in the transcriptome of these mice, we had seen a high elevation of a YAP transcriptional target, cyclin D1, which is a known oncogenic driver. So even in young mice, although the cancers don't appear until the mice are older, at 2 months old, we still see a great increase in the proliferating hepatocytes seen here. And by Western blot in these 3 different pairs of young mice, you see elevated Cyclin D1, which is liver-specific. This is a whole-body MAPL knockout, but we only see Cyclin D1 elevated in liver and not the other tissues, consistent with this bile-YAP system. So that's what I wanted to show you. This, of course, we have a lot of other data in these mice, but What I want to tell you is MAPL, we've learned that they— its function at peroxisomes is to play a major role in bile acid flux. So, which leads me to ask, what is the cell biology behind bile acid synthesis and secretion? And we hope that this model we have will offer a new system of liver perturbation and cancer development. But obviously, as you've seen from the other talks, I've highly simplified this discussion. Those same kinds of contact sites and other contact sites are going to play roles in fatty acid and glucose metabolism in liver, which have not really been studied well. At the kind of level of regulated cell biology, the dynamics that change during the feed-fast cycle. And, you know, iron, I'm very interested in iron. We don't know how iron gets from the endosome into the mitochondria, if those are contact sites. There's lots of movement there. And it's a unique place where we should be studying peroxisome biogenesis turnover and signaling, because where they have a very clear role in the generation of bile acid and ether lipids, etc. As you've just heard. So with that, I'm going to stop. I think that— I hope I've convinced you that liver cell biology is really ripe for discovery, especially with all of the different models of stem cells, etc., that we can make organoids and, and look at these things. And I, I have so many questions from starting into this. So the urgent need, I think, for cell biologists over the next decade would be, you know, to develop some type of metabolic probes for bile and iron, and other things, to spatially map in complex cell systems and tissues how these things work, to apply the already existing beautiful Janelia high-resolution imaging systems and expand them across scales, and establish and share protocols for robust organoid and co-culture experiments that are meaningful to human physiology, the impact of zonation, which I know you're going to hear about in the next talk by Daniel, etc. So much work has been done here in the last 5 years even, so it's an exploding area of of biological systems. So lastly, one of the things I've learned from my journey through the liver is that I actually know nothing. So it's really important to collaborate. You need— so if Janelia will continue to act as a hub to encourage collaborative research across disciplines, you know, I came in here knowing nothing about bile or nothing about the liver, and suddenly my mouse has a phenotype, and here we are. So with that, I'll just thank the people who did the work. The mice was developed by Vanessa, and Rudy's done a lot of the assimilation, and Orel has been really spearheading the YAP signaling later file steps in here and the potential role of peroxisomes in the liver, and all of my collaborators that I wouldn't be able to do anything without.

So thank you very much for your time. Thank you, Heidi. That was really cool. A reminder again that questions from anyone listening can be submitted via the Q&A. I have a quick one. So surely, the ability of bile acids to regulate YAP1 wasn't created in nature to cause liver cancer. What is its healthful contribution? It's analogous to branched-chain keto acids weren't created to cause unhealthy cardiac hypertrophy. That response is probably there as a healthful response to exercise.

What's the analogy in your system? Well, I think what's emerging, especially— certainly not from my work, I mean, we're trying to follow people— but Marino Zerial and others would say it's in regeneration of liver, right? So if you start to have cold stasis, if there's some problem in the bile ducts or something where you have an, you know, a pathology, that this can initiate repair and regeneration. So it does for sure would have some sort of functional role. People are looking at— there's so many ways into— yeah. Obviously, it's a mechanical stress situation because that the bile is offering, but there's probably much more to the story than that, and I think there's a lot to tease out there.

Okay, Maria Forni again. Thanks for a fascinating talk. The liver changes its function during embryonic development. Is this followed by alterations in peroxisomal dynamics?

I mean, yeah, I have no idea. I mean, you know, when we, when we see this peroxisomal polarization, Aurel and I were looking through the literature, well, mostly Aurel, and there's like a couple of studies in the '80s and maybe one in the early 2000s where it was hinted at that there was some sort of functional polarization of peroxisomes to come towards the bile canaliculi, but that they weren't really followed up, and it's really an understudied area. And I mean, as a person who doesn't— I know it's not a belief system, but I mean, the idea of free diffusion of metabolites is kind of out the window. Window, right? So are you going to make this soap, this emulsifying bile, and then let it float around your hepatocyte? I mean, there must be some targeted mechanism to get it to that apical surface, and it's a really unknown thing, so I don't know.

Okay, and does MAPL regulate processes like cell cycle or autophagy?

So, yes, we think a lot of that can be indirect. We have, we're looking into it. So there's, the only cancer we see in the knockout mouse is this liver cancer, which we think is really indirect from a lot of the biochemistry that we're seeing. It's probably not a profound cell cycle thing. However, I will say that early studies that we had looking at Senp5, there's a SUMO protease that translocates to mitochondria at the G2 to M transition of the cell cycle to de-SUMOlate things. And you need that, we think, to disengage the contact site so you can go through mitosis and other things. We're still working on that aspect of where you have simulation dynamic changes, we think, in regulating contact sites for, for cell cycle transition. But there's, there's so many surprises that came out of this mouse, it's, it's been a little bit overwhelming. But yeah, great question.

Okay, last question. Well, wait, and here came some more. Okay, does MAPL also affect intracellular calcium waves in hepatic tissue?

We haven't looked at hepatocytes, but we published some years ago with Julien Prudhomme, who was a previous postdoc of mine, in the cell death pathways, that when MAPL stimulates DRP1, it stabilizes these— the oligomer of DRP1 and the contact sites, allowing increased calcium flux. So if you lose the stimulation, you lose that calcium wave. So it's not essential for division, but it seems to be a regulator that can modify the fission events, the contact site events that are that are occurring different, different times, different cell events.

And last one, um, are the mitochondria fully functional in these settings? And I presume settings meaning with manipulation, manipulation of MAPL.

Yeah, so we've done, uh, a lot of work on just regular basal respiration using Clark electrode, old-fashioned kind of stuff, and there's— they're totally fine, they respire well, um, general metabolism is okay. And, and there's a long story to go out. We should be releasing releasing this manuscript on bioRxiv really soon in the next week or two, so you can read a lot more about what you'll see there if you're interested. Thank you so much.

Terrific, Heidi. Thank you. And the last speaker in this session is a chosen one. This is Daniel Feliciano from Janelia, from the Lee Lab.

And please, fire when ready. All right.

Can you hear me?

Yes.

Perfect. All right.

Can you see the presentation? Yep, looks good.

All right. So, I thank you, Chris, for the introduction, and thank you for the coordinators of these workshops for allowing me to share my view of how we can expand the toolbox to to study liver biology across scales. This is very important because, as you know, our bodies are built as efficient machines, and this is because we have different levels of organizations that regulate our biology. This includes organ systems, organs, tissue, and the cells that build these tissues. If you go into the subcellular level, we see that we have different organelles, which are in charge of maintaining the cells alive by working different tasks. But these organelles can also perceive how the extracellular space change, and they actually adapt to this. These are 3 examples of 3 organelles that have been studied for many years. And what has been shown in vitro is that depending on the extracellular, environment and the nutrients that these cells perceive, the organelles change in their position, numbers, and, and their morphologies, adapting these organelles to then allow enzymes to perform specific functions that allow them to survive in these environments. Well, these are organelle signatures, and as you can see, tell us something about the metabolic state of the cell. But the important thing is that now we can go into tissues and try to address whether these signatures also operate the same way. If you go into the liver, we see that it is formed of the functional unit, the lobule. The lobule is made of a central vein that has 6 corners, with a portal side. These portal sides contain the hepatic arteries and the portal veins. We can break down this lobule into smaller pieces, or the hepatic acinus. When you look at the hepatic acinus, you see that blood flows from the portal side to the central veins carrying oxygen, hormones, and nutrients. And this unidirectional flow of blood generates different microenvironments, a gradient of these different components that makes hepatocytes within this acinus to operate in different ways. So if you look at different zones within the hepatic acinus and you interrogate what is the morphology of these organelles, you can see that they are different. In pericentral hepatocytes, they are elongated and tubulated mitochondria, while in the portal hepatocytes you can find fragmented large mitochondria. So what this suggests is that these microenvironments that exist within the intact tissue of the liver determine the baseline or the basal activity of these cells by modulating the signatures of the organelles. This is important because cells have more than just mitochondria. They have all these different organelles that have been linked with metabolic diseases. But the importance of these organelles, it doesn't only, uh, is because of these metabolic diseases, but because they are being linked also to their neurodegenerative and cardiovascular disease. So we can say that there is in reality a multi-level interdependence. Where the architecture of the organ established a microenvironment or microenvironments that pretty much set the baseline of how these cells are going to respond to external cues. Depending on what are the cues that they detect, these organelles will shape their morphology and adapt to these cues, feeding back into the organ, allowing it to operate the way it's supposed to. So we can say that metabolic flexibility is encoded in organelle adaptive responses. Unfortunately, we don't know much about these because of lack of multiscale analysis approaches where we can look at multiple scales at the same time to address or interrogate how organelle dynamics function in different environments. And also, we don't have the tools available to to look at multiple organelles at the same time and in intact tissues. So today I'm going to talk to you about two main areas that I started working on to try to address these gaps, and one is to use different ways of studying the liver using a multi-scale approach, and the other one is to generate new tools to study multiple organelles simultaneously. The first part of the presentation will address the use of tissue organelle scale analysis, which basically does is to separate the tissue into different bins or zones. And doing it this way, computationally, we can ask specific questions depending on the zone where these hepatocytes reside. We can use this approach and also interrogate changes in organelle dynamics in different nutritional regimens like extended fasting or starvation in mice or Western diet, which is equivalent to nutrient overload. So as you can see, we can image entire tissues and see how the organelles are distributed. But also we have the cellular and subcellular resolution to look at independent cells within these different zones. So by implementing these tissue organelle scale analyses, we can now assess the periportal to pericentral axis. And doing so, we found that different nutrient regimens impact organelle distribution across the hepatic acinus. In addition, we can use these different variables that we are using to assess the different zones within the liver acinus and combine them into a multiple component analysis to generate organelle signatures. And we found that different metabolic states are represented by specific multi-organelle signatures, as you can see. The second area that I would like to talk about today is the generation of tools to multiplex organelle studies. So for that, I started building these plasmids, polycistronic plasmids that allow the expression of multiple organelles simultaneously. These are two versions of the plasmids, and when you combine them together, you could inject them using hydrodynamic tail injection, which basically delivers those plasmids into the liver, and you can study multiple organelles at the same time. This is just a zoom, high-resolution image of one of those hepatocytes within the liver of a mouse that was injected through hydrodynamic injection, and as you can see, we can now look at multiple organelles in the same cell or multiple organelles within the entire hepatic acinus and ask questions about their morphology, their numbers, and different features that they possess, and also compare these in, in conditions where nutrients are deprived or nutrient overload, as I showed you before. The other direction that I'm taking here at Janelia is trying to develop new transgenic tools because the injections of these into mouse is one way, but we can also do it looking at the entire organism or comparing organs between each other. And zebrafish is a great model for this. So right now we are generating in the Li lab two lines that have— that express different subsets of organelles that if we cross these two, we can look at up to 8 subcellular structures at the same time.

I kind of heard like an echo. Is that—

I think it's gone, Daniel.

Keep going and we'll let you know. All right. So, so now using these new genetic tools, right, we can study both zebrafish and mouse and implement the tissue organelle level analysis that I showed you at the beginning to ask questions about organelle dynamics within these tissues. But not only for the liver, we can also use it for other organs. To end this talk, I want to share with you my view on how I envision 4D physiology can help fill up the gaps in our knowledge in liver physiology. And this is by tackling specific areas. The first is to build new tools. These tools should allow us to measure metabolites within different organs or the bloodstream, or even allow us to detect the organelle's activity or the activity of the enzymes within these organelles. Also building microscopes, the new microscope modalities that allow us to look at whole organs at high resolution to have these tissue-to-organelle approach, and new genetic tools for organelle multiplexing, as I mentioned. The other area will be a multiscale approach, where, as all the other speakers mentioned, we can tackle this by looking at interorgan communication. This is liver with other organs or liver with systems like the nervous system, tissue to organelle cell analysis, or even organelle scales, because all of these are necessary for the function of the organ itself. Finally, the dynamic processes that we can study in the liver are development, as it was mentioned in one of the talks, regeneration, the circadian cycles, both the light and dark cycles, and feeding-fasting cycles. And disease models. And I will stop with this.

Thank you. All right, we're supposed to— thank you, Daniel. We're supposed to have 15 to 20 minutes for a general discussion, so I'm just going to throw one question at you. All right, the imaging experiments that you have been discussing, can you do this in live cells to see how organelles respond to perturbation in real time? Yes. Yes. Okay, good. That was a succinct answer. Thank you.

So I am going to— Chris, there's one question from Ashley Salmonson in the chat.

You want to grab that one?

That's what I just did. Oh, you did.

Sorry about that. Yep. Yes, okay. I don't look like I'm very observant, but I kind of am. So, okay, so we've come now to the discussion session of session 1, and this is— these are too many points to discuss, obviously, but this is ideas for discussion that each of us in this session came up with. And two things to say. One is this phase of each of the sessions is open to all of the panelists. That means all of the speakers and all of the session chairs are able to ask questions during this time or contribute to the discussion either by the chat or by coming on live. Now, as you gather your thoughts, I think what we'll do is ask each of our discussants to choose one of their bullets for 2 to 3 minutes to expand upon the concept. And we'll go in reverse chronological order here. And Daniel, if you don't mind, either talk about metabolic zonation or circadian cycles as an additional important thing to consider in the context of, you know, 4D metabolic physiology.

Yes. So I can actually use both topics at the same time because they are important, both of them, and this is going to be brief. So there are many studies in liver biology that when they report their, their studies, they don't say what zone they're looking at. And depending on the zone they're looking at, the biology might be different, right? That's one of the things that I feel that needs to be improved, right, in different studies that look at the liver. The other aspect is the timeframe, right? If you look at a liver at noon, it might be different than when you look at it at midnight. So also reporting the time in the day, right, not only studying this in a circadian cycle way, but reporting that these organelle distribution are different, and maybe the response of the liver of different cues are going to be different depending on when in the day you're looking at. So keeping in mind these two aspects of liver biology would be very important for, you know, expanding our knowledge in liver physiology.

So Heidi, you also mentioned zonation. What about one of your other bullets?

Would you like to elaborate on one of them, please? Well, I think what I— the point I was trying to make generally is just the challenges in, as a cell biologist working on mitochondrial dynamics, you know, trying to understand what it means, the signaling. It's hard for us to move into complex systems because, you know, it's just easier to knock stuff out in a COS-7 cell, right? And you can map basic machineries in a simple system, and once you want to go into an organoid or a human iPSC-derived something or other, it just takes so much longer, and for basic discovery, it's a huge challenge, right? So, but those tools are changing, and so part of it is, is the communication between the kinds of questions. When I look at a liver cell, it's different from people who've been studying bile and iron metabolism in liver for a long time. They think of it in a different way, right? So it's It's part of it is just understanding stuff. And I'd like to ask Daniel, I mean, how do you see overlaying what you're doing with metabolism, with the outputs?

Yes. So one of the things that I try to do when I was doing these type of studies is I actually look around the clock how the morphologies of these organelles were. But I was actually taking the animal and I was taking a blood sample to try to look at how the blood or the components in the blood, like glucagon, insulin, were at that specific time point. Because that way you can correlate what is in the blood, what you see in the liver, and try to make sense of if they are regulating each other. Right? So this is one way. The other way is obviously developing the sensors that allows you to to detect, if possible, metabolites within the liver itself. And then you can ask questions about, okay, so my organelle signature looks like this, and these are the metabolites present inside the cell. Are these regulating each other? You know, there are multiple ways, but I think right now we just need to develop more of these tools to ask these difficult questions.

Okay, thank you. Christine, would you like to elaborate on one of your points?

There's someone who has raised his hand.

Maybe it's a question to Danielle. Ashley Sommers.

Yeah, thank you.

I just have a quick question in kind of response to what Danielle was just saying, and this was the reason for my question of whether you can do this in live cells. Is, um, the— what is the timing that it takes to see these changes in organelle morphology? And so if we think about metabolism and things happening on a scale of seconds to minutes, are you able to see those changes in a live cell image, uh, let's say if you flow nutrients over the slide or something like that?

Yes, um, so changes in morphology and numbers you can see them in scale of hours and sometimes minutes, like 20 minutes, you can see fragmentation of mitochondria pretty quickly. I think the problem that I see right now into trying to image the entire, let's say, hepatic acinus in live animals using intravital microscopy is how fast you can do that, right? Because getting a snapshot of the cell or that position in the acinus will give you a good idea of how things are within 30 minutes, right? But if we are trying to ask questions of very dynamic processes that happens within minutes or seconds, that is not possible unless we— design new microscopy modalities that could capture all that area at the same time, right, with the high resolution to really see these morphology changes.

Yeah. So there was a question in the Q&A: can the imaging techniques be used to track macromolecules such as glycogen? That's interesting because glycogen has different subcellular addresses depending on where you are in the fasted to fed transition.

Hello?

You're asking Daniel that question, right? Oh yeah, yeah. Well, either, either Heidi or David, uh, it's about the imaging techniques.

Yeah, no, glycogen, um, I think is, is, uh one of those molecules that, kind of tricky to label, I think, in live animals.

So my answer is I really don't know how. I mean, I think some people are moving towards, you know, this rapid organelle IP metabolomic combination, right, where you can— but again, in tissues that's a different, probably a different story, but if you have a labeled organelle And I'm still— it's still not great, but you can do metabolomics on whether you're getting flux in and out. I've seen a number of people doing that kind of approach, starting with David Sabatini, obviously, where you can look at amino acid levels in different places and other metabolites using that. But it's still a smash them up, right? You're losing spatial identity in that context. But it's a big challenge, and I'm not sure you know, how— what the expectation should be. So here's one that you follow iron really directly at contact sites, you know, it's that kind of thing.

But here's an intriguing question, which I think gets to your topic of metabolic zonation, Daniel. Heterogeneity in hepatocyte metabolism is guided by oxygen levels. How can one account for that, or simulate that in the context of liver 4D metabolism.

That's a really great question. Yeah, I mean, I know studies that they have tried to reproduce the, you know, the oxygen concentration in tissue— not tissue culture, primary hepatocytes— that their medium and the oxygen is similar to either the pericentral or periportal. But I feel that the contact with the other cells that surrounds them also gives them cues for the sonation to occur. Like you mentioned oxygen as a factor for sonation, but glucagon, for example, is also important. And there are cells within these hepatic acinus that use glucagon more than others. By using it, let's say portal cells use this glucagon, then the pericentral cells receive less of that when the blood pass. So what I'm trying to say with that is that I still feel that you need to do this in the tissue if you want to really get deep into the more complex biology of, you know, organisms.

So let's come back to some other questions that are coming in, but Christina, I want to give you a chance.

Did you want to elaborate on either of your bullets? Yes, well, I think my bullet will integrate themselves to the other bullets and maybe to what Heidi is suggesting or Danielle. And I think with what Danielle is saying, like, you could have one level where you would do imaging of lipids, they don't vary that rapidly, not as much as intermediary metabolites. So, you do have lipid environment around at contact sites, and we know around also the respirasome in the mitochondria, and these can— the type of lipid that you have around can really influence, let's say, fusion, fission, and I've learned this working with Eric Schubridge and analyzing his transgenic cell where they knock down different genes. So I think it's something that can be added. Let's say what they're doing now is 2D imaging of lipids. And they do this in tissue slides. So you do keep the 2D localization, there is spatial resolution. At this time, some people are starting to do like single cell, but not with the imaging, but it's coming there. And I know they're tackling those results and making like a 3D reconstruction. So I think this can really add to the information in terms of how, for example, nutrition, we know that the composition of this lipid will be influenced by disease and will be influenced by nutrition. We all know about the impact of, for example, omega-3 diet, which can be beneficial. For example, plant phylogen are very enriched in those omega-3. Also the impact of oxidative stress. In a way, I think I'm adding one level to what you're already doing, Daniel, probably. I would think that it would have information And it could be done also in the brain, because I think different— the lipids will be different depending on the brain region, and because all these complex lipids have different roles. So yeah, that would be my point. And we know much less about— although it's like an opening field since the past 10 years, because it has been lagging on what we call metabolomics, which are more more the small molecule. And what we have learned from lipidomic with untargeted, it's more— it has lag in time. It's more over the, let's say, past 5 years that we've learned more about it. So yeah, I don't know what you think, Daniel, if you see this as possible. Oh, for sure.

Yeah, it would need developing techniques, of course. I was actually also thinking on doing a tissue-to-organ scale analysis, right? Getting those organ signatures but using the same tissue and implementing techniques like MERFISH, for example. So in that same tissue, you have the signature of the organelle, so you have an idea how the metabolic state is. Then on top of that, you have the transcriptomic, right, the information of what proteins. So the same way you can put some sort of measurement of lipids or the metabolites that are in that same cell along the acinus.

One thing I can say is that with the actual techniques that we have, at least we can, and others have even better one, but we can even measure the signature of lipids in mitochondrial membrane that are isolated. For example, Eric Shubridge sent us the outer mitochondrial membrane or something like that. And we've been analysing depending on which gene has been modified. And we've analysed this in C. elegans and bacteria, Pili and things like that. But this is still like taking, let's say, doing a homogenate extraction, but you lose the spatial localization, but at least you learn about the signature of this specific organelle. Are subagranular.

Okay, before our time is up, I want to take 1 minute to do some shameless lobbying for a project area that I think is actually ripe for consideration for this kind of 4D project, and I come back to my second point, which is about pancreatic islets, because when you go in and remove the pancreatic islets from the whole pancreas, they represent only 1% of the mass of the whole pancreas, and you put them in tissue culture, they are functional 3D spheroid balls of cells that work in many ways like they do in situ. So there's the paracrine communication networks are there, the ability of the glucagon-producing cells to respond uniquely to some metabolites like alanine is a glucagon but not an insulin secretagogue. The glucose-stimulated insulin secretion response, the incretin-stimulated or enhanced stimulation of insulin by glucose, all of those things are there. But as I, as I tried to say in my earlier remarks, when you have that collection of cells in the islet spheroid, and you are, for example, administering a metabolic stimulus, you don't know in real time how the different cell types are responding. So, could one use Janelia's protein engineering capabilities, for example, to create proteins that sense specific metabolites inside the cell, or that allow us to see activation of signaling circuits like GLP-1 activating the GLP-1 receptor in the beta cell, and using imaging technologies on top of these protein engineering strategies to really decipher across the cell types of the islet how there's an orchestration of the function of the alpha, the beta, the delta cells to create an end result of a change in islet function. That really excites me and hopefully excites somebody else. And a related question that came in on that topic from Amy Elliott is— Amy says, there are several exciting combinatorial studies outlined here, many of which are facilitated by imaging. What strategies are being developed to handle the multiscale large datasets and appropriate algorithms to create computational models from different experiments? And I think that's a very relevant question. So any further discussion on these points? Because we are going to have to wrap this up here in a minute or two.

Any other burning questions that anyone would like to ask? Well, maybe if someone wants to address that last question, it could be Vamsi himself.

Yeah, go ahead, Vamsi.

So I do think that an important challenge for the future is developing theory for trying to integrate across scales. I mean, I think one thing that's fantastic about this last session is the multi-scale nature of all of the problems. And I think that's, that's going to be a problem that I think is going to get revisited throughout the next 2 days.

All right, on, on that note, your humble host yields the floor and Vamsi, take back over please. Thank you to all of our panelists in this session. Thanks everybody. It was great. Great.

I learned a lot. Thank you.

Thank you, everyone.

Thank you. Great, great start to the workshop. I think at this point we will go ahead and break until— we'll take a 7-minute break and resume at 12:40 PM Eastern.

Great.