Scientist Stories: David Furman, Systems Immunology of Aging for Precision Medicine

I'm David Fuhrman. I am an immunologist by background turned data scientist. I've been leading the Stanford 1000 Immunomes project for now 15 years and recently moved to— well, recently, a couple of years ago now, moved to the Buck Institute for Research on Aging to lead an AI platform to better understand the molecular mechanisms of aging., and my field again is in the inflammatory space and immunological aging and how that defines non-communicable disease. Today I'm going to start sharing my screen and I'm going to talk about the work we do at Stanford and at the Buck Institute. And also a full disclosure, we have founded a company called Edifice Health to commercialize this novel metric that I'm going to talk about, which is a systemic chronic inflammation biomarker and has a number of implications for different areas. So the title of my talk is Systemic Immunology of Aging to Enable Precision Preventative Medicine. And again, I'm director of the 1000 Immunomes Project. This Study is continuing for the next 5 years with an additional funding from NIH that now is going to the Buck Institute, Stanford, and our collaborators in Technion University in Haifa, Israel. So I just wanted to kick off this presentation by showing you this plot where we see a tremendous variation in human rates of aging. So this is work we're doing with Joel Kramer at UCSF, where we have weighted diffusion imaging, a number of functional MRI done in the brain of relatively healthy humans and people also with mild cognitive impairment and some individuals with Alzheimer's disease. And we've been following this cohort for a number of years now. So we have longitudinal data. But you also have collected a number of molecular biomarkers from bodily fluids, especially CSF and blood. So what you see in this plot is the gray matter volume in the y-axis as a function of age. So the most important take-home message from this figure is that there are people, that appear to be young in their age, but they have a fasting aging phenotype, those in red here, and those that are protected also in the young cohort, looking absolutely healthy from a standpoint of brain aging. And that also occurs in the older adults. So we are using this variability, we're embracing this variability in human phenotypes to better understand how people age in their brain, in their cardiovascular system, and other physiological systems as well. So in the immune system, that's also true. We see tremendous variability here, and we've been able to map these slow aging phenotypes with data in centenarians and the accelerated phenotype of aging with multimorbidity, and I'm going to talk on that a little later during this meeting. So I would like to bring up this idea, and I think there's now a consensus in the field that systemic chronic inflammation and in general the immune system sits as a hub for aging. There are now over 200 publications showing that sustained inflammation in the body will trigger genomic instability, telomere attrition, epigenetic alterations, and all 9 hallmarks of aging. So what if we could detect this type of inflammation and treat, or at least prevent, these different diseases from happening by tackling systemic chronic inflammation? This will have a tremendous implication for the economy and for healthcare systems in general. We published this paper a couple of years ago pushing forward the idea that systemic chronic inflammation not only affects aging-related diseases, but also you can see inflammation in young adolescents and childhood in autism diseases that are, for example, accelerated metabolic disorders, type 2 diabetes in the young, and etc. So again, we have published in cardiovascular disease, immunosenescence. There is a lot of data that suggests and demonstrates that systemic chronic inflammation triggers all forms of cancer from transformation to metastasis and even depression. We have a paper that we published in 2014 showing in major depression disorder that systemic chronic inflammation is involved in this type of condition. So we're using an unbiased approach. We're using omics approaches, mostly inspired by the 1000 Genomes Project. And we founded the 1000 Immunomes Project with the idea to understand how the cells and cytokines in the body will interact with one another to ultimately yield a phenotype. So you can imagine that this could be the immune system in an n-dimensional immune space with proteins and cells interacting. And you could be really, really good at studying this type of cell or molecule or even peptide, but you're going to be missing what's happening in these different immunological states that I'm pointing with my cursor. So the more you study, the higher the chances to capture those states. And those states may very well be driven by the age of a person or a particular infection, for And based on that premise is that we founded the Stanford 1000 Immunomes Project, and we do systems immunology. And the way it works is we have set up the Human Immune Monitoring Center at Stanford. That's a technology center here that was founded in 2007. We have a pretty strong recruitment and clinical team and bioinformatics core. And with the right immunological expertise, the idea here is that we will be able to generate new models and hopefully new therapies therapeutics for HLA-related diseases. So this is how the Stanford 1000 Immunomes data looks like. We conducted HLA-omics, transcriptomics, proteomics, cell enumeration by mass cytometry, and we stimulate cells in vivo, etc., etc. We've been publishing all this data over the years with more than $30 million from NIH funding over the past 10 years, and now we're continuing this project, as I mentioned, I mentioned before, with another $17.5 million that the NIH has provided us to continue this for the next 5 years. So I'm going to focus on multimorbidity for a second. And I think we all understand in the aging space multimorbidity is the number one priority for global health research. And that was published in the Academy of Medical Sciences in 2018. And the reason is If you cure one disease, right, if you focus on single diseases and single conditions, you only extend the lifespan of the population by 1 or 2 years, right? And so many important aspects of multimorbidity are listed in this slide here. Coordination of medical care, management of polypharmacy. We're having now a whole program of understanding how polypharmacy accelerates immunological aging, for example. The time spent per patient and the aggravation of one condition by symptoms or treatment of another, as I just mentioned before. And the annual cost in the US is estimated to be $3.5 trillion annually. So I talked about inflammation and the consequences of systemic chronic inflammation, but it is very important to notice that this type of inflammation is different from what we know about acute inflammation, right? In general, when we hear the word inflammation, we think of a red, swollen tissue, and that's a response to an infection. That's acute inflammation. It's short-term, it's high-grade, it causes tissue repair, and it should go back to baseline in 3 to 5 days. In case of systemic chronic inflammation, this is very different. These are different triggers.— is not infectious, is mostly the exposome, what we eat, how we treat our body, how we sleep, the levels of stress— is persistent. And I should say there are many subtypes of systemic chronic inflammation. We're actually redefining as I speak on a couple of papers that are now we're putting forward on subtypes of acute inflammation and subtypes of chronic inflammation as well. But the important part of this slide is to really to demonstrate that there has been canonical standard biomarkers for this type of inflammation. One of the most commonly used markers is hs-CRP, but this is a very poor predictor of cardiovascular disease. And since this is the only metric that existed so far, cardiologists have been using it, you know, left and right. But now people are— this have decided, healthcare providers, that this is not a really good predictor of of cardiovascular disease or aging, by that matter. And it's mostly an acute phase reactant, really, that raises with infection and vaccination. So we've been focusing on IL-1 beta. That was initial work that we did in 2015, 2017. And the reason is we've seen that many types of sterile inflammation— that are listed here, like cholesterol crystals, amyloid beta, calcium pyrophosphate, silica, etc., will trigger the inflammasome and will produce IL-1 beta. And so based on this premise is that we focus on inflammasome gene modules that are listed here on the left-hand side. And what this plot represents is the age association of those inflammasome gene modules. The inflammasomes are molecular scaffolds that enable production— much of IL-1 beta and IL-18, not a good thing to have in high levels. And as you see in red are upregulated gene expression inflammasome modules, in blue are downregulated. The most important two are here, module 78 and 62. Those are inflammasome gene modules. Again, they are elevated in older adults. And that is true across the board. If you look at 5 years of data, this is high in old Module 62, high in old, Module 78, regardless of the year of data where we looked into. Now, when we stratify patients based on whether or not they have constitutive expression of this inflammasome gene modules, we're able to differentiate, to stratify individuals based on their levels of hypertension. The inflammasome Module low and high status will separate people into those that are hypertensive versus those have normal blood pressure. We also found that there is a link to pulse wave velocity, so their tilt stiffness phenotype, and we validated that in a number of studies in animal models and also in vitro. Importantly, the inflammasome module low group had a very poor history of longevity. This is simply by asking individuals whether or not they have a family member that lived for 90 years and older. We're also following these individuals longitudinally. We are able to determine whether or not they died during the study. Some people died. What I'm showing you in here in E is the inflammasome module median expression for those who were still alive in 2016. Relatively low compared to those who died during the study.. And that's also true for module 78. So this differentiates individuals that healthy aging and unhealthy aging, and we have a strong link with longevity here. And these are the different cytokine biomarkers that we found. We were also able to identify a number of danger-associated molecular patterns by using metabolomics that we injected into the mice. We generated hypertensive mice that had very similar phenotype here and immune activation across the board. More recently, we have decided to take a more unbiased approach and not just focus on IL-1 beta, but go and use some more sophisticated approaches to define this concept of inflammaging. We use deep neural network to fit a model, basically to take inflammatory cytokines that were measured in 1,000 Immunomes Project and target— basically have as a target variable aging. So we're predicting age, and you're using this number of cytokines and chemokines and growth factors from blood to derive what we call inflammatory age. And this is the loss function or how the algorithm works. We use cytokines here. This is a deep neural network, a guided autoencoder. On the one hand, we want to be able to reconstruct cytokine data. And on the other hand, in parallel, we're actually predicting the age of a person. So this is the age of a person, right? And if you have the predicted age in the inflammatory space, that represents what we call inflammatory age. So this is how it looks, the results of those studies. Inflammatory age in the y-axis, calendar age in the x-axis. Seems exciting, but, you know, in real life, how important is it to predict the age of someone? You know your age. You might as well ask you for your ID. The important part is really whether or not these individuals in the upper side of the spectrum, right, are at risk for developing multimorbidity and other diseases, whereas these individuals down here may be protected. In other words, does the variability, the variance in inflammatory age for any given age bracket be predictive of disease or health phenotypes? And that's what we ask next. Can inflammatory age predict multimorbidity? And the answer is yes. So here I'm showing you age-adjusted inflammatory age in the y-axis as a function of a number of diseases in the x-axis. And the conditions that went into the model are listed in the left-hand side of the slide: cancer, cardiovascular disease, respiratory disorders, etc., etc. And we published this paper a few months ago. We were also able to get data from our collaborator in University of Bologna, Claudio Franceschi, where we measured in 19 centenarians their inflammatory age. And what we did here was to subtract inflammatory age minus chronological age. So that's the inflammatory age index. And we were able to see that there's, first of all, a huge huge variability in centenarians, right? So in order for you to get to be a centenarian, there are multiple ways— genetics for sure, maybe microbiome as well. But here we see that inflammatory age also plays a role, right? Based on some criticisms that we've had from our reviewers that centenarians— there may be a ceiling effect there— we decided to rank individuals for their IH. We also ranked their age, and what you see in this plot is that the frequency of the number of individuals with a low IH index is enriched for centenarians, whereas those with high IH index is enriched for people 50 and 80 years old, so the older controls. Because we had a number of stimulations that we did in the immune system, especially looking at the JAK-STAT pathway. So we take cells out of the body, we plate them into these petri dishes, and we start stimulating these ex vivo using IL-6, IL-10, IL-21, interferons, etc., etc., and we look at the phosphorylation of signaling pathways like this STAT pathway. We were able to correlate the inflammatory age age level with the different responses in these immune challenges in vitro. And so what you see from this slide is the result of a regression model where high inflammatory age correlates with poor immune responses across the board for B cells, different flavors of CD4 and CD8 T cells. And so we— this is basically the antiviral, antitumoral response that is compromised in individuals with high levels of inflammatory age. We still need to figure out why monocytes have the opposite behavior. So they're ready to shoot, they're ready to produce high levels of phosphorylated STAT and have that delta. And we know that they actually produce a number of different inflammatory stimuli. But overall, the antiviral and antitumor response is much reduced in individuals with high inflammatory age. So we also looked into frailty. So we— and the frailty score was based on a Rockwood frailty index modified version where we have time up and go, so functional measurements, but also a number of questions that we asked these individuals about their general autonomy and independence to move around, to do their daily activities. And what you see here is the predicted frailty predicted frailty score from data we gather in 2010 and the actual observed score from 2017. So we're able to predict who's going to become frail 7 years in advance. And this is in individuals 75 years and older. Again, because we have this multidimensional multi-omics approach, we were able to bring inflammatory age in the gene expression space. So how do we do this? Gene expression transcriptomics data, and we fit a model to predict inflammatory age. Very simple model. And what you can see here is we use almost 400 people to generate this surrogate marker of inflammatory age in the gene expression space. And much like we do with the image algorithm with my friend Shai Shinor here, where we took a cellular metric for immune aging and were able to map that to gene expression and then use the Framingham Heart Study to predict all-cause mortality. We do the same here, and these are the different genes that went into the model and the coefficients, and we're able to predict all-cause mortality with a p-value of 0.02 after adjusting for a number of risk factors such as age, gender, smoking, cholesterol levels, etc., etc. So, what are the different biomarkers and mechanisms of inflammatory age in driving cardiovascular aging? Because that's our stronger phenotype here, or at least one of the strongest phenotypes. What you see from this slide is deconstruction of inflammatory age. In other words, the Jacobian, the first partial derivative of inflammatory age. And cytokines like IL-6 is not coming up as being significant. So if you're looking at age-related cytokines, I think looking at TNF or IL-6 or CRP is probably not appropriate because there's been a lot of confusion in the field. I think after adjusting for diseases, then the correlation between IL-6 and aging is gone. There's a lot of, I guess, conflicting literature. And what we see is that acute biomarkers are not coming up as being significant. And then there's an overwhelming contribution of this chemokine. This is a small chemokine called CXCL9. And so there are protective factors here that are downregulated. There are detrimental factors. I'm going to show you some examples on how CXCL9 will actually act upon the endothelium We're able to model this in iPS-derived endothelial cells and age them in vitro, and they produce a huge level of CXCL9. When we knock down CXCL9— this is beta-Gal staining— we were able to reverse cellular senescence in the endothelium. We're also able to reverse some phenotypes like production of tube structures, in other words, angiogenesis. And we are able to do this not only in iPS but also in animal models where we incubate aortic tissue with increasing concentrations of CXCL9, and the whole function of that tissue is compromised. Now, in the same cultures, we analyze RNA sequencing data from this aging endothelium in vitro, and we're able to see elevated inflammation as early as passage 8 after culturing these cells in vitro and decreased proliferation. Much like what we see in senescent cells after knocking down that single gene, CXCL9, we have almost a complete reversal of the signature, uh, both when it comes to inflammatory, um, pathways and proliferation pathways as well. So to demonstrate that this phenotype is related to stiffness, because we had data in humans that we can predict arterial stiffness and left ventricular hypertrophy, remodeling ratio in the heart. This is something I didn't show just not to overwhelm the audience here, but these are the cultures of the iPS endothelial cells after passage 8. You see an upregulation of CAM molecules, metallopeptidases, and collagens, an arterial stiffness phenotype that can be reversed by knocking down the CXCL9 gene. So more broadly, and now jumping on the work we're doing in the company I founded, we take public domain, and this is using the DrugBank, FoodDB, LINCS-1000, the NIH consortia, and also comparative toxicogenomics databases. We build a number of interaction networks, so the whole interactome, 1.6 billion interactions. And then we use our in-house data center to ask the algorithm, given a particular inflammatory age fingerprint, give me the combination of compounds that will normalize that signature. If you have very high levels of an undesirable cytokine, you want to be able to drive that down. For example, CXCL9 or eotaxin. If there are protective cytokines that are downregulated, like gamma interferon, for example, or TRAIL, you want to be able to bump them up. And by doing so, we identify about 260 inflammatory age modulators, and we have now a number of double-blinded placebo-controlled trials ongoing. This is the result of just one of them where we show that our, our immunohistochemical compounds can decrease the level of inflammatory age, and that correlates with a decrease decrease in this protein E1. I can't really discuss what this protein is, but you can see this beautiful correlation between the decrease in protein E1, which is— which was the one that we predicted will go down, and inflammatory age. In the placebo group, we have no such signal. It's all noise here. So just to summarize, we have been using systems immunology in translational medicine, and the idea here is that we do to the 3 Ps— phenotyping, predicting, preventing. And we have a number of programs now in brain aging and infertility. We've been monitoring 150 infertile women where we have particular phenotypes of aging there that we're able to associate with noncommunicable disease. We're applying this to space medicine in collaboration with NASA. We have a story in cardiovascular disease, in frailty, in metabolic function, as I mentioned before, and has application also for immuno-oncology. We have a study that deals with PD-1 and PD-L1 inhibition, and we're able to predict who's going to be responding to immune checkpoint inhibitors. We're also having now starting a project with AMD, so age-related macular degeneration and diabetes. Diabetic retinopathy and hearing loss as well. With that, I'd like to thank many people from my group and other institutions that have collaborated with this work, and I'm happy to take questions.

Thank you so much for listening. Wow, fantastic. Thank you so, so much. This was great. I think you, if you don't mind leaving this slide on for just a second, because I think the first question from Steve Fowkes is related to one of your slides, and perhaps, Steve, you'd like to start up.

Yes, you had a slide that I didn't get the title of it written down, but it had to do with white blood cell profiles that had shifted to the left and the right based on your quantification of some kind of I can't describe what measurement you were using, but can you talk about any toll-like receptor involvements with those shifts?

So we didn't specifically stimulate it or signal our assays through TLRs, but we know that the microbiome may be interacting interacting with TLRs in the gut, and that apparently has a consequence in driving at least gut inflammation. And we've seen that in animals. So we don't study that in humans because it's very tricky in vivo, obviously. But we do see TLR4, for example, and 9, surprisingly, being a player in driving inflammation in the gut in these animals. But from the human data, we don't have a direct assessment of TLRs.

All right, lovely. I think Nir had a question and he may have to hop off. So Nir, if you're still here, maybe you want to jump in.

David, you can take off the slides so we can see each other better. And I just want to— and, you know, David, we're going to be collaborators, and I really think your work, all this kind of work, is important. I want to make more of a general observation, and for all of you to know, every one of us in biology of aging has his favorite hallmark, okay? And what we tend to do is to take this hallmark and put it in the center, okay? And in fact, all those hallmarks are connected to each other. You treat one, you treat some of the others, right? So we have to be careful not to say in this case inflammation drives aging. Inflammation can be a sign for a bad biology of aging in all other organs, okay? It's just an association. And a more important question will be Let me start with not your example. mTOR, right? We all know about mTOR. mTOR increases— the expression of mTOR increases with aging. If you give rapamycin, you rejuvenate. Okay, but isn't the real question why mTOR increases with aging? Isn't the real question what is the mechanism behind increasing inflammation in aging that we want to know rather than to know the hallmarks that are important treatable, targeted, but we really don't know to do the hierarchy there of which is more important, which is more causative. That's absolutely well taken, Nir. Thank you for that observation. And I think the studies that we're planning hopefully will drive us towards an understanding of the contribution of these different aspects, right? Because we're doing a number of multi-omics, including microbiome, including metabolomics, including all these different omics. And hopefully we will be able to see how at least the genes related to exceptional longevity will interact with these different biomarkers in order to drive a certain phenotype. Absolutely agree that things may be more complex than just, you know, systemic chronic inflammation is the cause of aging. I agree on that. Thank you, Nir. Thank you.

Thank you. Great job. Great job.

Thanks.

Okay, lovely. Next one up, we have Terry. Right. Lovely talk. Thank you very much for it. And I've been following the work really closely. It's so fascinating and important. My question was a very simple one. What about SUPAR, the soluble urokinase plasminogen activator receptor? If I say the whole name. So in epidemiological studies, it's been really outperforming more traditional indicators such as CRP and IL-6 in terms of its status as an outcome of early life factors that should predict chronic systemic inflammation and as a predictor of aging as well. So I've seen quite a lot of papers out of Denmark about that, and it seems really interesting.

It is absolutely interesting, and I would love to see if we could incorporate that in our studies. We haven't, unfortunately. We have not measured súpar. My hunch here is that, one, there's going to be a correlation between súpar levels and our biomarkers, and two, incorporating both ought to improve the predictability of certain phenotypes, including aging per se. So we're thinking more of an integrated view, right, rather than comparing notes. I think it's important to bring all these biomarkers together, right, and be able to improve the predictability and translation ability of the work we're doing. But we have not measured Super in our studies. So unfortunately, I don't have specific data or answer for you at the moment.

So we'd like to implement your inflammaging measure in our epidemiological cohorts, and we've gone pretty far with SUPPAR.

So if you would like to make an exchange of information, we could do that offline. Absolutely. Absolutely. Let's connect offline.

Absolutely. Thank you, Terry, for that. All right. Lovely.

Next one up, we have Sergei. Okay, hello everyone. Actually, my question is really similar because my question is about correlation between BNP, brain natriuretic peptide, with systemic inflammation. Because it's already known that this substance, BNP, is a pretty good biomarker for acute short conditions. Definitely as predictor of high risk. What's about these correlations? It should be interesting because it's definitely known that this biomarker is better than, for example, already mentioned high sensitivity CRB. Yeah, I think, Sergi, you were muted for a moment, but I get the gist. So we, again, we have not measured the biomarkers that you're mentioning. If you have samples, you know, and maybe plasma or even serum samples from your cohorts, or you would be interested to know how these compare, I would be more than happy to introduce you to the Edifice team and see if we can explore a collaboration with you either to improve the predictability or even to compare the accuracy in predicting these different conditions, especially specifically when it comes to the brain. Right.

So I hope that answers at least partially your question, sorry. You muted yourself again, but I think you're fine. Yeah. Yeah. Okay, good. Okay. Let me know in case anything else pops up. I'm now asking a naive question from the chat, which is, do traditional treatments for acute inflammation, like anti-inflammatory medication, also work to mitigate chronic inflammation or its downstream effects?

No, the answer is not at all. Those anti-inflammatory molecules will target acute inflammation, and will most likely ruin your microbiome. And so, as you're, you know, we know ibuprofen and all, uh, most of the different anti-inflammatory medications will wipe out your protective microbiome, and that will cause disruption in the gut, um, lining. So the mucin level will probably be much affected, and then you have you have all sorts of translocation of antigens into the periphery, and that also, I mean, only increases inflammation rather than decreasing it in the long run.

All right, thank you.

Next one up, we have Steve. Yes, I think it's brilliant that you're focusing on the subtleties of inflammation, and one of the aspects of it that I'm considering in terms of aging is this issue— you know, as people hit older and older ages, that their regulatory and homeostatic systems may be failing due to focal infections and opportunistic infections and maybe even exposure to things like mycotoxins. And I'm wondering if you have or are using your technology to assess the underlying profiles that drive humoral immunity from innate immunity in specifically, let's say, toll-like receptors, that if you have viral fragments in people who have viral disease, you may be able to pick up a pattern that would be able to distinguish them from bacterial infections or environmental exposures or mycotoxins?

Yeah, unfortunately, the answer is no. We don't have the data to be able to separate things out. And I think you need a large sample size anyway. We only have 1,000 people that we followed over the years. That will require, I think, a much larger sample size to be able to distinguish those details. Now, what I can tell you is that we have now a new application that we're submitting to NIH to study organ-specific acceleration of aging phenotypes, specifically looking at inflammation biomarkers by measuring 7,000+ proteins in people with HIV, where we expect we'll have an acceleration of immune aging and cardiovascular aging, as opposed to HCV, where, you know, we'll have probably a more pronounced acceleration of aging of the liver, and people with SARS-CoV-2, for example, where we have have that more organ-specific acceleration of aging in their lungs. So we now have even results from the COVID-19 studies. In about 800 people, we're measuring 7,000 proteins, and we see this accelerated aging phenotype in their blood even after 6 months months recovery, but we don't know if this relates to TLR.

Again, we don't have the data for that. Sorry. So are you tracking these people over time?

Yes.

Or are you—

is there any utility in, let's say, a snapshot? There is. There is. For the most part, those with accelerated those showing protective phenotypes that are maybe in the upper side of the quadrant longitudinally stayed there. And those that are more accelerated phenotype longitudinally also stayed there. So those signatures of aging in long haulers are actually stable for almost a year after they've cleared the virus and they have gone on of symptoms. So this is kind of a—

it's going to be, I think, taken by reporters pretty broadly, this story. Lee, you get a thumbs up. Okay, next one up we have Tinka and then Ravi. Hi, I have a question regarding the slide that you showed that that there is a higher activity of monocytes with aging and there's decreased activity of other T cells and B cells with aging. Is it the reason for that possible that is because adaptive immunity activity decreases with aging and innate immunity activity increases with aging? So, is it possible that because of that, since monocytes are part of innate immunity, that they have higher activity with aging than T and B cells? That are part of adaptive immunity have decreased activity with aging.

Yeah, and I think what we see here is exactly what you're saying. Inflammatory age is in the middle, right, of this high inflammatory burden in the innate immune system with hyperactivation and a decrease immune response in the adaptive arm, I think that's where exactly inflammatory age will sit, right? We've seen that there is a hyperphosphorylated form of STAT1 in CD8 T cells, and that is because these individuals have high levels of circulating cytokines that almost what's happening is it exhausts the resources in the phosphorylated form of the CD— in STAT1 CD8 T cells. So what really matters there is the delta of the response, not the absolute phosphorylation levels. So if you have monocytes that are pumping up cytokines, that will maximize the resources in CD8 T cells, and you won't have have that delta anymore. And so the cells, when you take them and you do RNA-seq, for example, you see almost no signal. They don't— they not only don't respond to these cytokines in vivo or in vitro, rather, but they also don't produce any downstream gene expression. So it's completely dysfunctional.

I agree with you on that, Tinka.

Thank you. Thank you.

Next one up, Ravi.

Hi. I was continuing the discussion on the sort of adaptive immune system. I was wondering if you had taken any look at the actual adaptive immune repertoires, like things like CD4/CD8 ratio, naive versus memory T cells, particular clonotypes or antigen-specific clonotypes. Yes. The answer is yes. We've looked at all those. So we have CBC, right, complete blood count. We have CyTOF done on— those individuals. We have also TCR repertoires for some, not all 1,000. That will be a big undertaking. But we see that inflammatory age predicts lower naive CD8 T cells, not better than age, but better than cytomegalovirus. We also see a shift in the CD4 to CD8 ratio. Ratio. So all those hallmarks of immune aging are correlated with inflammatory age. I didn't have time to go over those because of the short, you know, time that they gave me, 20 to 25 minutes. I tried to keep it short, but we do have hallmarks of immunological aging being correlated with inflammatory age.

Thank you. Yeah. All right. So we I would like to ask you a few more general questions, if you don't mind. And again, this is an invitation for anyone to stop me by asking more questions. But at the beginning, you mentioned as a disclaimer that you're currently spinning out a company based on this, or at least are in the process thereof. I would love to know a little bit more of just, you know, what's going on on that end. How far are you? What are your next steps? Are you seeking investment?

What's going on on that entrepreneurial side? Right. So So I would say probably 4 years ago when we have analyzed all these results and we were getting so exciting predictions of multimorbidity and everything else, we sat down with my co-founder Mark Davis and said, we have to bring this to the masses, commercialize this. It's too good to keep it in the academia. And so we founded the company a couple of years ago. You know, early stage. We secured our seed round about a year ago, and now 3 or 4 months ago, our Series A round, and we're completely oversubscribed. And so to the extent that we had to open an A+ round, and with almost $20 million in the bank, and we kind of trying to just push the investors for a B round next year. So at the moment, we have a fully developed test because going from academia to industry standards is a leap, right? It's something not straightforward. There are some challenges. Obviously, we want to be CLIA certified, and we're about to get CLIA certification to be able to take commercial and clinical samples. So the test is up and running, has all these different industry standards, less than 5% error. We have secured our A+ financing round. And we're now embarking in a large, large study. And we're recruiting as I speak. Where we're testing 10 different combinations of compounds for 10 different immunotypes. I didn't go too far in my presentation there, but it may be that Stephen is 45 inflammatory age years old, and I'm also 45, but the way he gets there is different from, from, from the way I I get there. He may have protein A being very high. I may have eotaxin or lack of interferon. So the interventions will be personalized. And that's where the individual-level interventions come into place. So we are now clustering individuals based on these different immunotypes. There are young-like immunotypes and common-type immunotypes and a variety of different ones within those. And the interventions are now being tested in this double-blinded placebo control study in 1,000 individuals. And the results are going to be released by the end of April, I think. And in addition to measuring inflammatory age and protein deviations, we're measuring a slew of different clinical phenotypes, from arterial stiffness to hearing capacity to even wrinkle severity. We're going wild with the different claims here, but we'll see what pans out. I think having these surrogate markers of aging is absolutely critical. And so the first outcome will be moving the needle in inflammatory age. Second will be to try to see how we move the protein deviations and bring them to optimal state. And then third will be all the clinical endpoints that we're studying. With that, we will be able to add a number of claims to our existing interventions that we already have a bunch, um, and launch by Q2, Q3 of next year. This is a DTC, it's a direct-to-consumer, uh, with some ties to functional medicine physicians, um, and wellness clinics and corporate wellness et cetera.

Okay, fantastic. And you say you're actively recruiting.

What would people like, what conditions would they have to fulfill and how would they go about signing up? Yes. So the, our website is www.edificehealth.com and you'll find at that website ways to enroll. This is a study that we're doing in the Bay Area mostly. Mostly there's a site in Atlanta, Georgia, but mostly people from the Bay Area, specifically in Marin County, because our doctors are actually from the Marin County area. So thank you for pasting that in the chat box, Mika. So anyone can go to this website and find ways to to enroll in this study if they're local here in the Bay Area. And I think it's free or even paid to participants where we get a blood draw, a number of different measurements from their functional physiological, different physiological systems. And then we give them an intervention based on their immunotype phenotype, and this is a 6-month duration study with 4 different visits. And in each visit, we evaluate microbiome, all the omics possible, inflammatory age, and all these different clinical endpoints that I just mentioned.

Okay, before I go to my final question, we have one more from Marina here in the chat. Marina, do you want to unmute? Yeah, hello. I'm just coming from a completely different field, from neuroscience. In our case, for example, the amyloid deposition will start much earlier than the actual Alzheimer's disease symptoms occur. My question was, at what age does the inflammastatin become predictive of multimorbidity?

Maybe I was inattentive during your talk. Thank you for that question, Marina. The power of our study does not allow us to establish that kinetics, but we are now with Joel Kramer with 1,300 participants followed for about 10 years, up to 10 years. We will be able— I don't have an answer for you yet, but we will be able to tell at which point in time cytokines are predictive of phenotype, at which point in time gray matter volume is predictive of cognitive impairment. So we're actually putting all that together as I speak. I don't have an answer for you yet, but expect an answer for me by early Q1 of next year. I actually— my neuroscientist working in my group in this thing is actually here, Nikola Markov.

But yeah, we'll have an answer for you probably in the next 6 months. Okay, Nicola will be a fellow fellow fellow, so you have lots of time to talk. Okay, then, you know, this is my kind of template question that I always ask at the end, nearing the end of it, which is like, do you have an idea of where this could be in like 5, 7 years down the road if everything goes well? And then really to finish up, what can people in this group what can people in this group do to help your work move along? So you already mentioned, you know, perhaps like signing up if they're interested. But are there any other bits that you're currently seeking collaborators on or just like info on? But yeah, if you want to take us on a bit of a walk, like where could this be excitingly? And then also what can people here in this group help you with in case they decide to do so?

Yeah, that's a great question or set of questions, Alison. I would say that in 5 to 10 years, at least special aspect of what we're doing, which I think is the one that moves fastest, will probably be democratizing the test. And I hope to see 1 billion people taking the test and improving their health. And this is absolutely critical. If we don't get there, you know, the health condition of the population is declining so bad that I think we have to tackle it from the inside, and systemic chronic inflammation is one aspect of this. I would like to see in 5-10 years, a billion people having taken the test and improving their health. From a commercial side of things, the entrepreneurs here can reach out to Edifice to see if there's room for Series B funding early next year. We'll probably be pitching VCs early, as early as February of next year. If there's interest there, reach out to the CEO of the company. I can put you in contact with them and investor relations and all that part. That's, that's obviously not me dealing with that. Stuff, and, and I prefer to keep it that way. Um, and, uh, from a purely scientific standpoint, I would be glad to see, uh, that scientists here in the, in the audience can collaborate with us, right? Um, on the many projects that I mentioned— infertility, um, especially in infertility, because I have, uh, you know, putting a lot of effort in that stuff. Space medicine,, with the three aspects being, you know, low gravity, stress levels from being in this enclosed capsule for a long time, and also cosmic radiation that will cause accelerated aging phenotypes. If you have samples you would like to study in your own research doing maybe metabolic disorders or, you know, obesity or whatnot, I would say reach out because it's really interesting that our test is applicable to other conditions.

Fantastic. Okay. This was pretty succinct, and we are almost at the hour. So I just want to thank you on behalf of everyone in the group. Thanks a lot. I already got lots of requests for email address. So let's see. But thank you so, so much. Perhaps, you know, we'll have to invite you back in again to check on all of the data that the people here keep on asking about. So thank you so much. Thank you so, so much for joining. I really, really, really appreciate it. I shared your longevity challenge with the group, and I want to take, I guess, the last 30 seconds or so just to remind people that we have Vision Weekend coming up, which is our annual member gathering. And for the first time, it's at the Internet Archive, at a rocket company, in a ship in San Francisco, and at a castle in France at the same time. We'll have lots of speakers that you know and cherish a lot, including our fellow and mentors from our biotech track here, both in the US and in France. We integrate that with our molecular machines track, so better control of matter in the molecular machine space. And our Feynman Prize winners will join us both in the US and in France. And we have our decentralized computing track mixing privacy-preserving computing and cryptography and AI. And then finally, a newer energy space bucket for all of the technologies that people in the community are interested in, but that don't fit into one of our technical programs yet. All right. If you're interested in that, I just shared the website here in the chat and there's applications for income support tickets so you can apply for a free ticket if you want to join. All right. I hope to see you there. Thanks a lot for staying one minute longer. And thank you, David. I hope it wasn't the last time that we had you here. And I'll see many of you again, I think, already next week.

So thanks for joining everyone.

Bye-bye. Thank you for having me. Bye-bye.