Okay. Well, welcome. It's a pleasure to introduce our symposium. I'm Charlie Marmer. I'm the Chair of Psychiatry at NYU, and I direct a Center for Precision Medicine and Alcohol Use and PTSD, and I'm part of the collaboration and a co-author on the papers that will be presented today. I'm very delighted to introduce you briefly to three of my colleagues on this consortium. Now, this is a consortium primarily funded by the Department of Defense. It's been funded for more than 13 years. It started for Dietrich, Maryland, and at UCSF when I was a professor there, and it's now expanded to seven sites, including Harvard, including still UCSF, McLean Hospital, University of Memphis, the Institute for Systems Biology in Seattle, and several other sites. And the underlying goal of this project is to identify the molecular pathways and circuits involved in post-traumatic stress disorder and TBI in active-duty military and veterans and translate that into simple, practical laboratory tests that could be used for high-throughput screening. So if we could do it, which is very difficult, we would develop a blood test for high-throughput screening for PTSD, and we have some promising progress. But today you're going to hear more about some of the underlying metabolic pathway disruptions in PTSD. And the presentations today will be from three of my colleagues, all of whom have research appointments at the Walter Reed Army Institute for Research in Bethesda, Maryland at RARE, and they are Dr. Saeed Mouhi, Dr. Baruch Miskanah, who will present first, then Dr. Mouhi, and then Dr. Rutin Yang. And all are molecular biologists and computational biologists by training, and all hold positions as computational biologists in the RARE program focused on military readiness with a great interest in PTSD and TBI, but major depressive disorder as well. And they're going to be presenting on metabolic pathway disturbances in PTSD with a focus on mitochondrial dysfunction and related pathways. So I'm very pleased, and we'll introduce the first speaker, Dr. Miskanah. Let's start with his presentation. Okay, thank you. Thank you, Dr. Mouhi. So I'm Baruch, I'm a computational biologist. So my presentation today is mainly on one of our recent work on the relationship between metabolic syndrome and PTSD at two layers, at phenotypic level and as well as at genetic correlation level. Okay, so a disclaimer, so whatever, I'm not speaking on behalf of the Army or the U.S. government, so I have to say this, so this is a required slide. Okay, so as you know, PTSD often occurs with other comorbidities. One of these very common comorbidities of PTSD is metabolic syndrome, as well as its individual components. So why metabolic syndrome? So as you know, metabolic syndrome is a major risk factor for cardiovascular diseases, which is a leading cause of mortality in the U.S., but also metabolic syndrome, along with PTSD, especially metabolic syndrome, is a disqualifying condition for many military duties, directly impacting military readiness. So it's a very relevant topic for the Army, along with PTSD. Okay, so a lot of prior studies between PTSD and metabolic syndrome are mainly on observational studies or using epidemiological data showing higher incidence, higher prevalence of PTSD among PTSD-positive patients. However— Can everyone hear all right? Is that clear? Okay, yeah, okay, yeah, sorry. Yeah, thank you, sir. Okay, so we specifically want to explore three specific questions in this paper and in our talk today. The first one is, in addition to the increased incidence rate of metabolic syndrome on PTSD patients, what is the relationship between PTSD severity score and the level of metabolic dysregulation? Second question is, does this observed phenotypic correlation arise from underlying genetic, shared genetic background between metabolic syndrome and PTSD? The third, the last but not the least, is if we confirm this genetic correlation between metabolic syndrome and PTSD, is there any causal influences between the two, if there is any? Is PTSD causing higher incidence of metabolic syndrome, or is it the other way around? So we will explore these three questions in this paper. Okay, so Dr. Malmar already mentioned the CISBio cohort, CISBiology cohort. This is a well-characterized, well-funneled data set consisting of chronic PTSD patients from the U.S., from veterans. The second data set that we will use in our talk is, in addition to the individual data from CISBio cohort, we also gathered summary statistics from genome-wide association studies from 10 different papers. So large-scale genetic studies from some psychiatric phenotypes, PTSD, MDD, from PGC studies, from UK Biobank genetic studies. So we collected 10 genetic studies, and we make sure these genetic studies have enough heritability, h-squared. They explain enough of the phenotypic variants of these phenotypes. Okay, so as observed before, we see among PTSD patients, we see higher components of metabolic syndrome. For example, PTSD patients have higher BMI score compared to compared to healthy controls. So similarly, on the raw scores of PTSD severity score, CAHPS, we also see a significant Pearson correlation, as well as rank correlations. It's not shown here, but we clearly see, in addition to higher incidence, higher, let's say, for example, higher BMI score among PTSD cases, we also see a correlation with higher PTSD severity score being correlated to higher metabolic dysregulation. Okay, so as I mentioned earlier, so the next question we want to explore is, does this genetic correlation arise from some shared genetic background? Right, so the answer seems to be yes, at least for obesity-related components of metabolic syndrome. Oh, by the way, most of the slides are from a recent publication, so I will mention those so we don't have to take every note, a lot of notes, or pictures. Okay, so as we were mentioning, so we see a significant genetic correlation between PTSD diagnosis using those joint summary statistics and the components of metabolic syndrome. Similarly, we also did a similar genetic correlation estimate with depression summary statistics. As you can see, the genetic correlation between PTSD and metabolic syndrome components is very comparable to that of depression and metabolic syndrome. The correlation for PTSD being slightly higher, but it might be, but at least it's comparable. Okay, so we also did next, so those correlations are genome-wide genetic correlations. We wanted to see if there are any specific locations, genetic locations, that are responsible for this genome-wide genetic correlation. So in the next two talks, you will see more details on pathway genes. So in this analysis, what we did is we divided the whole genome into 2400 specific, separate genetic locations, and we found three specific locations where we see a higher contribution towards the overall genetic correlation between metabolic syndrome and PTSD. One notable specific location is a genomic location that has previously been reported to be responsible for the correlation between psychiatric illnesses and immune-related phenotypes. So there might be a three-way link between immune dysregulation, metabolic syndrome, and PTSD and MDD. Okay, so as I mentioned earlier, so the third question is to see, to test the causal relationship, the causal influence between metabolic syndrome and PTSD. So as I mentioned earlier, we saw a significant genetic correlation between PTSD and chronic PTSD, but not necessarily with recent-onset PTSD. So that led us to hypothesize that the relationship between PTSD and metabolic syndrome might be from PTSD to metabolic syndrome. Obviously, we cannot conduct a randomized controlled trial, so the next best alternative is Mendelian randomization. For those of you who are not very familiar with genetic studies, so Mendelian randomization, a way of nature's randomized controlled trial, where the randomization happens at birth, right? So what we need is genetic variants that influence both the causal variable as well as the outcome variable. In this case, PTSD and metabolic syndrome phenotypes, and then we use those GWAS summary statistics to test this specific link, to test the specific test, right? So one relevant, so these are the required criteria for this, for selecting these SNPs, these genetic variants. One specific notable one is we need well-powered genetic studies to implement this, right? So we implemented this Mendelian randomization between PTSD and metabolic syndrome. So our test on the relationship of the effects of PTSD on metabolic syndrome did not result in statistically significant causal relationship. It might be due to the power of PTSD GWAS, so then we tested, we implemented a similar Mendelian randomization analysis between depression and metabolic syndrome. As you can see, two-way Mendelian randomization shows there is a causal, there might be a causal relationship from MDD to metabolic syndrome, but not necessarily the other way around. Okay, so just to summarize, we confirmed the phenotypic relationship between metabolic syndrome and its components with PTSD diagnosis, as well as severity. Using publicly available large-scale genetic studies, we also showed statistically significant genetic, shared genetic between PTSD diagnosis and metabolic syndrome as a whole diagnosis, but also its individual components. We also found some specific genetic locations that contribute to this overall genetic, shared genetic between the two. So our next work would be to replicate this on multi-ancestry genetic studies and explore specific pathways and genes that contribute to this overall genetic correlation. So our next two speakers will explore the last two points. So if you want to read more about today's talk, these are some recent publications on the topic from our consortium and our group. And of course, you can always email us and talk to us. And that is all. Thank you. Great, thanks. We have time for several questions before we go to the next thing. Maybe come to the mic, though, to ask your question, if you don't mind. Go back one slide. We're happy to send the references, too, if you just, if you want to email me, I'll make sure you get all the references. Great. Any questions from the first presentation? Wonderful. I think we can move right on to our second presentation. And just to, I'm delighted to introduce Saeed Mouhi, who's going to, also a computational biologist at RARE, and he will continue this theme of looking at metabolic pathways in PTSD and, to some extent, a major depressive disorder as well. But just to remind you that the cohorts we're studying, the cohorts that we generated in the consortium in this case were a study of 323 veterans with and without PTSD, Iraq and Afghanistan veterans who returned from service five to eight years after their deployment, and a second cohort, the Fort Campbell cohort, which is a prospective longitudinal study of 2,000 active duty military personnel deployed out of Fort Campbell, Kentucky from the 101st Airborne to Afghanistan. The index deployment was in February of 2014, and we continue to follow them now. We're in the fifth wave of data collection. So that's, yes, please, come to the mic so everyone can hear your question. The question I have is the sequence of both variables. Which comes first, the PTSD and the metabolic? And how long is the difference in timing from one to the other? Please. So how long between the two, it's not time-wise, but we, on the main randomization, we did two-way tests. We tested if PTSD causal influences metabolic syndrome, or MDD causal influences metabolic syndrome. We see a significant genetic p-value, but if we do the same test the other way around, we don't see a significant statistical significance. So to answer the question of time, I think maybe Dr. Mahmoud might be better, but we also have to answer other mediating effects, could be smoking, less activity, those kind of things. That we didn't test. We just simply tested MDD causal influences. It's been difficult to answer that question because most of the studies are cross-sectional, but we have two very important PTSD cohorts. There are others, but two very important prospective longitudinal cohorts. The one is the Fort Campbell cohort of 2,000 people. And what's interesting about our Fort Campbell cohort is we assess them, including in terms of their metabolic parameters, before deployment. So we actually have pre-exposure molecular pathway data on the subjects, which we're following now to 10 years from 2014. Soon it'll be 2024. There's also an important civilian cohort called the Aurora cohort that was developed out of Harvard by Sam McLean and Kerry Ressler. And that's also a large several thousand person, and that's a prospective longitudinal civilian cohort of people seen in emergency departments and followed over time. And in the current grant, I was recently fortunate to be awarded a new grant to compare the two cohorts. And so we will have data comparing the longitudinal course of PTSD, metabolic syndrome, and other biological features in both veterans and civilians. And then I hope we'll have an answer to your question at that point. If not, I'll have to refund the DOD a large amount of money. Yes, next question. With regard to your cohorts with the veterans, are you looking at the incidence of childhood trauma or pre-deployment trauma? Yes, we're looking at trauma in three frameworks, childhood trauma, trauma in military service but outside of the war zone, and trauma in military service in the war zone. And we're attempting to model all of those on the longitudinal outcomes. And unfortunately, among men and women who serve in high-intensity combat roles, and we first demonstrated this in the Vietnam War, and it's true again in Iraq and Afghanistan, there's a disproportionate levels of high adverse childhood events among men and women who volunteer for military service and end in high-intensity conflict situations as opposed to ending in combat support and service support and leadership positions. So those who actually see the heaviest level of combat have the most adverse childhood events, have the lowest level of educational attainment, have poor social support, and come to the military with significant vulnerabilities prior to combat exposure. And we showed that very clearly in a national study of the Vietnam generation and showed that they disproportionately developed PTSD and unfortunately died prematurely over the course of their lifetime from psychiatric and medical comorbidities. So it's a double vulnerability. Okay, I think we better... Okay, so this is a paper submitted to the Brain Behavior and Immunity. It is under review. What I'm going to do here is we have two cohorts as mentioned by Dr. Marwa, and we took the subset of those cohorts to do proteomics assays and epigenetic as well as specifically DNA methylation and the metabolomic assays. So now the question is, if we analyze these three data sets and check which pathways are activated and inhibited, do these three data sets converge to point to a specific pathway or they come together, coalesce together, to point to a specific activated and inhibited pathway? That's the question we are trying to answer here. So first, I don't have any disclosure to do, and this is our work, not the military's, and that's the first thing. So as we said, we have two cohorts. So here in this data set, we used 300 of the veterans and 159 of the active duty participants, even though we have 2,000 as mentioned. So we... because, you know, doing a proteomic assay is expensive, so we have to take the subset of the active duty cohorts. So we did metabolomic, proteomic, and genome-wide DNA methylation on these two cohorts. And then, you know, we did all the necessary preprocessing specific to each data set to make sure that, you know, we account confounders, we adjusted for some of the confounders so that we can have, you know, somehow reliable outcome or analysis output. And also, each cohort, the controls of each cohort is compared to the cases of each cohort so that we don't introduce, you know, cohort, because there is a difference between the two cohorts, so we have to compare within the cohort. So, we don't introduce something because of cohort variability or cohort difference. So, we did all those necessary adjustments or co-variating and all the necessary steps so that what we are observing is, you know, more of what we intended to see. Okay? So, this is the overall flow of our analysis. As you can see, we have two main cohorts, and each one has sub-cohorts. Okay, let me show you. So, these two cohorts are the active duty, and these two cohorts are the veterans. So, why you have here two cohorts? Here in this one are sub-clinical, and these are, namely, these are with PCL value between 20 and 37, and these are more than 38 PCL values. So, these are sub-clinical. That's why we have two subsets. And these ones are the recall of the veterans, and these ones are the original recruits. So, that's why we have two different sub-cohorts for each. And then we make sure that we compare the controls of this one with the case, and likewise here, and also for each sub-cohort, so that we are not going to introduce demographic variability to our analysis output. Okay? So, as I said, we have three types of modalities, proteomic, metabolite, and DNA methylation. So, then, at the end, we did a relation between these three modalities, because we know that DNA methylation is upstream regulator of, you know, epigenetic markers, upstream regulators of both protein and metabolite, and then we also metabolites and proteins, they can have regulatory interaction as well as functional interaction. So, we use those biological interactions to integrate these three data sets. Okay? So, this is overall analysis and workflow. So, the main finding here is, you can see, this Venn diagram, this part of the Venn diagram shows what we found, the pathways we found in the samples we got from recent, I mean, participants with recent PTSD, that means the active duty. And this part of the Venn diagram shows the pathways related to chronic PTSD, which are from veterans. Okay. Now, mainly, these pathways are related to a lot of comorbidities, like inflammatory response, oxidative stress, and other, which I'm going to show you. And most of these pathways are related to homeostasis. So, of course, there are also shared pathways. So, there are some metabolites which are really behaving in the same direction in both the recent and chronic PTSD participants. So, you can see, I'll show you the details of each pathway in the next few slides. Okay. So, you can see here, these are the pathways mainly highlighted in the active duty participants. So, you can see this, the orange or the red ones are activated pathways. So, these are important in homeostasis. And these are activated in active duty participants, or in recent PTSD. Okay. And these pathways, you can see, they are changed in both active duty and veteran. But you can see in the veterans, these are the veteran quarters, they are inhibited, while they are activated in the active duty participants. You can see these are the active duty participants, these are the veterans. And these are also important in remodeling or maintenance and senescence. You can see there is a difference or a distinction between this set of pathways between the two quarters. And you can see here, also, pathways important in adaptive immunity, energy metabolism, and also polyunsaturated fatty acids, as shown here by the individual metabolites. Here, in active duty participants, these are activated while they are inhibited in veterans. That means there is sort of, you know, the system tries to regain homeostasis or going back to normality in activity while there is, you know, a suppression in chronic PTSD, which is already established in those participants. And likewise, you can see neurogenesis and related neural activities are inhibited in veterans while they are activated in active duty participants. On the other hand, I will show you next, neurodegeneration and apoptotic pathways are inhibited in active duty but activated in veteran quarters. That means there is more neurodegenerative pathways activated in chronic PTSD compared to the active duty participants. And also, we have DNA methylation, as I will show you next, that the other set of pathways are related to chronic inflammation and reactive oxygen species. You can see in veterans, they are activated while they are inhibited. Some of them are inhibited in FCC. So when you come to DNA methylation, let me show you, here also insulin resistance, metabolic disorder and cardiovascular problems are activated in veterans while most of them are inhibited in active duty. So these are some validation using ELISA assay for especially oxidative stress-related proteins. You can see there is a difference between veterans and active duty participants. And likewise, we also showed or we also saw a similar pattern with regard to DNA methylation or epigenetic markers. You can see, the cursor is not working for some reason, okay, here it is. So if you look here, these are the proteins, okay, this is a protein and these are the DNA methylation or the epigenetic markers, okay. And also, you can see these ones are the FCC and these two are the veterans. So you can see there is a different pattern between the two groups, both in terms of protein and also epigenetic markers. So you can see hypomethylated promoter CBC-Citus or decreased protein level in a CBC cohort. So these are related to these functions. On the other hand, hypomethylated ones are related to chronic inflammation and neurodegenerative pathways which are, you know, increased in terms of protein and decreased in epigenetic markers in chronic PTSD. So you can see both protein metabolite and epigenetic markers point to the same set of pathways whereby inflammation, oxidative stress, neurodegenerations were activated in veterans, while they are in, I mean, homeostasis-related pathways and neurogenesis and maintenance pathways are activated in recent PTSD, active duty participants. So in conclusion, you can see molecular signatures indicative of homeostasis and related pathways are activated in participants with recent PTSD, while those related to inflammation and neurodegeneration and metabolic disorder are activated in chronic participants in chronic PTSD. So what we think is, you can see molecular alterations in the recent PTSD seems to indicate some sort of recalibration or compensatory response, maybe. It may not be, but that is a speculation. So that's what we see in these three datasets. Thank you. Terrific. We are open for questions. Yeah, please come to the mic. If I understand correctly, there are different patterns of mentalization for current active military folks with trauma, not just general population. Those that have experienced traumatic events, are they the ones who are without exposure to trauma? Correct. Those that have the active duty were deployed to war, all saw combat at various levels in 2014. Some have been deployed subsequently as well. So these are both cohorts, the active duty and the veteran, were heavily exposed to trauma during military service. And when we use the definition of PTSD, they've been screened in for having symptoms of post traumatic stress disorder. Say that again, what's the question? When we use the PTSD qualification, are they the subset of those exposed to the traumatic event that develop PTSD? All of these PTSD diagnoses are war zone related or military service related PTSD. We did not include in these analysis, childhood trauma related PTSD. And so finally, the difference between chronic versus acute, if you can tell us a little bit more about that and then also- Strictly speaking, this is not acute and chronic. You can think of this as PTSD in the early phase of its development over the first, for the most part, 90 to 180 days following deployment compared to PTSD five to eight years after the index deployment. I don't mean to be jumping in, but I'm responsible for that part of the study. So if those were good decisions or bad, they're my decisions. Just wanted to commend on the thoroughness of the work that they did. I did have a question. There has been some research on telomerase activity and the length of telomeres in relation to stress, chronic stress. I was just wondering if there was any intention to look at telomerase activity and length in these combatants? I just know from the paper, so telomerase shortening is known and there are a number of publications, but we didn't look at that here. Perhaps time for one more question if there is. Maybe not. I'll just make one comment that probably everyone here is aware of. There's a lot of complexity to develop reliable and replicable findings on these very deep genomic, proteomic, and metabolomic features in a psychiatric disorder. The history of our field is studying five proteins like cortisol or some other protein and saying, well, maybe we can understand PTSD if we just understand cortisol levels. The answer is no, you can't understand PTSD because the human body doesn't work that way. Cortisol is one of hundreds of relevant outputs from stress. That's the first problem. The second problem is that up till now, we've been limited for the most part in psychiatry to study molecular features in peripheral blood. The problem with that is that epigenetic features, proteomics, and metabolomics are organ-specific, tissue-specific, and cell-specific. The reason that we are lagging behind oncology in precision medicine and psychiatry is in oncology, they do the multiomic analysis in single cells in the cancer cell. We look in the peripheral blood, which turns out to be enormously noisy and complex and includes DNA, RNA, proteins, and metabolites from hugely diverse sources. So what we're moving this work to and all of these studies are going to be replicated in the next few years in neuronally-derived exosomes, which we hope will actually mirror what's going on in the brain in the molecular pathways as opposed to the periphery. So more to be said on that later. Our final presentation today is from Routine Yang, and Routine will be continuing this journey of discovery into the molecular signatures of PTSD and metabolic syndrome. Thank you, Dr. Mala. So today I will talk about the epigenetic and metabolic interplay in the active duty. So you see this the third time, a standard disclaimer. So our study, like Dr. Mala has said about this activity cohort from four combo, and we collected this data in three different times, pre-deployment and immediately after post-deployment, and three to six months post-deployment. And recently we have a new phase in the six years post-deployment called phase four, phase five. And this is a systems biology approach. So we collect all levels of molecular data, and also different clinical assessment. And so our aim is to try to identify distinct molecular signatures and uncover important biological pathways, and also try to see the stratification and subtyping of PTSD population. And this study I will focus on the metabolite and DNA methylation. So we collect 591 metabolites and also 985 lipids from the metabolite. I will focus on only this part. And we also have a DNA methylation used 450,000 methylation sites. So this study will see how these two molecular datas interact with each other. And we have almost 1,400 samples from three different phase. And here are some statistics. And we match the data between the DNA methylation and metabolites. So we still have almost 1,300 samples in both data types. And we find that we have different levels of severity of PTSD symptoms. So first, we have the PTSD. We used a PTSD checklist over 30 to define a PTSD. And then we had defined a substratio. A lot of the active duty do not have very, very high level of self-report PCL score. We have defined that substratio. They have symptoms, but they're not necessarily PTSD. And the most important things we find that in the control, so-called the lower PCL scores. We find this actually in our early study in 2021. We find that the PSQI, the sleep disorder score, or the depression score, PHQ, or that general anxiety score, GAD, is kind of a risk predictor for developed PTSD in the later phase. And also, there are some other risk factors we all know. They can be intensive combat exposure. They can have many different previous deployments. They can also have TBI, or they have alcoholic syndromes, the other things. So we can have that risk factor. So we can define a class called the low-resilient group. They have a lower PCL score, but they have all kinds of other comorbidities and risk factors. And they are, of course, the last stage is the high-resilient group. They're everything low. So we really check this low-resilient group and high-resilient group. We find that from the pre-deployment to the post-deployment, within like six months, we find a lot of the low-resilient group actually become PTSD. And so for the high-resilient group, they remain clear. And even we do that in the six-year post-deployment, we find that from that six months to six years, low-resilient group, previous low-resilient group. And we recall some of that, like a 25% of all the patients we collect in the post-deployment phase 3. And now we have phase 5, we see that a lot of them become PTSD, really. So this is the kind of correlation map. We see that, by definition, this low-resilient group, they have a low PCL score, but they can have some other symptoms in other categories. But the high-resilient groups is kind of pretty consistent for everything. Everything's low. So, first of all, we will try to see the epigenetics and metabolites. So, what's the association? The, here is some mathematic formulas. I try to confuse you. But just, this is a very large-scale correlation matrix. It will take a very long time to compute each of the CpG island, like 450,000 CpG island and 591 metabolites, the relationship between each other. So, there's a huge data. And we will adjust for some of the confounders, like some of that, like cell compositions in DNA methylation, because like Dr. Mama said, we have whole blood. They have multiple cell types in that. So, we will adjust for those things. And adjust for gender, adjust for other things. And eventually, we find that there will be 95 metabolites have strong, significant CpG island association. And their p-value is lower than 10 of the minus 12. And that's kind of a control for the multi-comparison. So, here is, you may not see all these things, but actually, there's a large tables. I only pick up some of them. Some of that, very important. You'll see that it's AHRR. And some of that, like a C1q, like a company passwords. Yeah, so actually, there's a long list with 65 metabolites. And they have a correlation with different DNA methylation. So, if we carefully look to all these metabolites and DNA methylation correlations, we can lay out some of the key network. We lay out this one by one. And all of a sudden, I realize the central of, the central of this pathway is this, called AHRR. This L-hydrocarbon receptor repressor. And so, a lot of this is related to this pathway, called AHRR pathway. And this pathway, the main function is moderate mitochondria and mediate oxidative stress. And then, of course, that will lead to the inflammation. And actually, in a lot of the study, so this AHRR is a famous, famous genes. They have been found in the PTSD. And this is, after adjusting for the smoking status, the methylation of the AHRR is actually slower in the PTSD-positive than all the health controls. And this is, have multiple publications. Our publication also identify these genes. And also, this gene is not only shown up in PTSD. These genes actually is the, it's kind of responsive factor to the environmental damage. So, like a smoking, alcohol intake, or like a very high coffee, coffee consumptions and exposure to some other chemicals. And they will all kind of change this AHRR. And this AHRR actually is a kind of protection function for the oxidative stress. And this has been published in sales and probably in many place. So, the take-home message is that we see that AHRR pathway in related to the oxidative stress in the DNA methylation. So then, we go back to the metabolite, or what we do the metabolite differential analysis. We try to see what's the most important metabolites between PTSD and control. There's another confusing mathematics formula. But the most important thing, we try to compare PTSD versus the resilient, high-resilient group and PTSD versus low-resilient groups. So, we try to see all kind of differential metabolites between the different categories. So, the first things we check is the PTSD versus high-resilient group. And of course, we know that if we compare the PTSD, the substratio with the high-resilient group, is kind of, we find that they have exactly the same regulation pattern, regulatory pattern, just a little bit lower. We see that all the colors are same. All the regulatory, the four-chain regulatory pattern are exactly the same between the PTSD versus the high-resilient group, all the substratio versus the high-resilient group. And the PTSD have higher, greater significance of association than the substratio. So, that make a lot of sense. In here, we find that there's something called a glycolytic ratio. We defined that before in our veteran study. And this one is also up-regulated between PTSD versus, compared to the health control in the veteran data. So, in active duty, we see exactly the same thing. In these parts. So, this is mostly in the post-deployment. And we, next, we will see just this glycolytic ratio. We find that in both the, this is related to the mitochondrial dysfunction. And this is validated in our veteran cohort and active duty cohort. And so, if we consider any phase altogether, and we will see a little bit more genes here. But still, they are in the same directions. And we have other things come up. And mostly, it's also mitochondrial pathways. And then we started to see those are more difficult to separate group, like a low-resilient group. And we have much less differential expressed metabolites. But one of that is significant. It's called the GABA ratio. This is also one of our defined ratio in our previous publication in the veteran. And we find that this one is also down-regulated. And it's consistent in both directions, and also same as in our veteran cohort. So, let's look at the whole picture. And although we have many, many metabolites here, but look at the whole picture. We find that a lot of these so-called, like a glycolytic ratio, there's lactate, pyruvate, and those things. And also, these GABA ratios. And a lot of this is related to the mitochondrial defunctions. And this will cause the oxidative stress. And oxidative stress will kind of lead to the inflammation. And this is same story we find in DNA methylation. We also find this oxidative stress and also the mitochondrial defunctions. So, here maybe have more beautiful figures. And we see that this is the markers of that mitochondria. And that lead to the oxidation. A lot of the PTSD has elevated oxidative stress. And we also see a lot of the inflammation markers in the PTSD. So, all this just linked together, consistent to the literature and consistent to a different type of cohorts. And consistent of different type of medical datas. And so, this is our take-home message. We see there's a co-AHR pathway. There's a central between the metabolite and epigenetics. And we see that the glycolytic ratio and GABA ratio to validate PTSD signatures. And we find that there were three basic components. Inflammation, oxidative stress, and mitochondrial defunctions. And we have a lot of people to thank. From the UCSF, from the NYU, and from the rare teams and PTSD consult teams. We have big collaborations for all this work. And thank you very much. Thank you. We're open for questions. While you're thinking about questions, I just wanted to make one brief comment. When we started our journey together, now with seven major academic departments of psychiatry nationwide, and the Army's elite molecular biology lab, and given the privilege and the resources, now over $100 million worth of research to do the first ever very deep molecular dive to PTSD and its comorbidities. Our assumptions at that time, mine, Owen Walkowitz's from UCSF, Kerry Ressler's from Harvard, and others, our assumption was that the major molecular pathways that would declare themselves in PTSD would be of three types. The HPA axis, the adrenergic axis, and the renin-angiotensin axis, based on previous studies. And what we have found over more than a decade is those pathways are relevant, but as the psychopathology of PTSD evolves, inflammatory pathways, oxidative stress pathways, and mitochondrial pathways are much more dominant in the multi-omic molecular signatures of PTSD than the ones we had originally hypothesized. I think that's very interesting and perhaps very important as we think about novel treatments and prevention. So that's just by way of a very broad overview of what we found. And the most surprising thing, I think you would say, routine, is how the mitochondrial pathways, particularly those from the Krebs cycle, and those that represent aerobic and anaerobic metabolism, have been very important findings, particularly the metabolite citrate, pyruvate, and lactate, which have appeared in all of our studies across multiple hormones as important features of PTSD. And if you would have asked me 13 years ago, would we be looking more closely at the mitochondria than the HP axis, I would have said, you need to come to my psychedelic psychiatry program, people. But that is the wonderful world of science. And what I love about the work that this team is doing is that they take a very unbiased approach and look very broadly, rather than looking in a more hypothesis-guided way. That's how we've discovered these other pathways as being so important. The same, I think, could be said true of major depression, which also began by focusing heavily on looking at monomergic pathways and is now looking very carefully at oxidative stress, mitochondrial, and inflammatory pathways. So I think we're learning something very interesting as we delve deeper into a genome-wide understanding of molecular pathways and so forth. Questions for Rutan? Yes, please. Come to the microphone if you don't mind. I'm not sure about phrasing this question. It's been a while since I was in learning about citrate, pyruvate, and lactate. That sounds really interesting, that correlation that you found that you definitely weren't looking for. And I think what I'm looking for is more of a overview of how do citrate, pyruvate, and lactate compare in the normal versus PTSD functioning? How does that relate to metabolic stress, oxidative function, and inflammation? And how does that correlate to things we know about, like exercise? That isn't too broad of a question. Yeah, this is a very complicated story. First of all, we don't know this is the PTSD cause of all this change, or is previous experience cause this change then lead to PTSD? So we need to answer this question first. And also, a lot of the biomarker research is we find association. So we don't necessarily fully understand their function, but we know they are correlated. And we can use that as a diagnostic biomarker. And we can use that screening purpose. And maybe we can use that for the treatment purpose. So in the treatment purpose, possibly, you have invented many ways, exercise, or a lot of other ways, to reverse this oxidative stress and reverse this mitochondria. So that's really, I think, it's really possible. Should I start answering your questions? I'm not sure if I'm going to have a clear question on that, but I'll try. About 10 years ago, I went to Herbert Benson's MindBody year, day-long program with Jon Kabat-Zinn at Harvard. And they had talked about, then, the beginning of discovering that using the relaxation response on mindfulness had an effect on the Krebs cycle and on the mitochondria. And there was a difference between those that were regular practitioners of meditation and those that weren't. Do you think, one day, that puzzle piece will fit with what you're discovering? Yeah, I think really, really possible. The current pathway, we keep seeing that in a lot of other studies, DNA methylation and subtyping. We all publish papers based on that pathway, even the PTSD treatment. And we also see those kind of mitochondrial functions between the responders and non-responders. We also see those kind of pathways. So I think it could be a very, very important piece as we can do. Social support. You know, we in our cohorts are pretty deeply phenotyped, especially the veteran, or there's a third cohort that we didn't present to you today, which we call the Cone Veteran Center cohort, which is nearly 1,000 veterans, also five to eight years after deployment, all Iraq and Afghanistan veterans. They're the most deeply phenotyped of all these cohorts. On that cohort, which we're now analyzing, and we're now just doing the multi-omic analysis, we have incredibly detailed histories. We have full, skid diagnoses, psychiatric disorders. We have the caps for psychiatric disorders. We have early childhood environment. We have lifetime stressors. We have multiple cognitive measures of attention, concentration, working memory, verbal intelligence. And we also have structural and resting state imaging on almost all of them. So that is our most deeply phenotyped cohort. The idea is we'll take the best findings from these studies and then attempt to replicate them in that cohort, which we've set aside and never looked at for these questions. So it will be an unbiased, truly independent external validation. We also have very large amounts of whole blood EDTA plasma, PAX gene DNA, and PAX gene RNA biorepository for that cohort, so that we will be able to do full-on neuronally-derived exosome work as well as peripheral blood work. So a lot more results will be coming in the next few years, and we're hoping to see which of these pathways will be most important as we understand PTSD and its co-relationships. Question. I noticed, I guess I should say, I didn't notice this, so my apologies if it was included in the presentation, I just didn't see it, but earlier I saw that quinolytic acid was potentially a biomarker for suicide in a previous presentation today, and I'm wondering if you noticed any association between quinolytic acid and PTSD? HVA, you're asking about? What's that? No association between? Quinolytic acid and PTSD. And so then tyrosine pathway? The tyrosine pathway? We have interrogated that. I don't think it's been exceptionally important at this point, but much more is to come. In the current study that we're doing now, we have five cohorts, three military and two civilian. We have a total of 3,000 PTSD cases and 12,000 controls, and we will be looking at all of these pathways in greater detail, and if we can, in exosomes as well. So we're not sure. Yes. Can you repeat the question? It's a wonderful question. We should actually, yeah, the question is about ice water baths, about cold plunges and their effects on our mitochondrial health. I think there's enough anecdotal evidence to suggest that we should, but it would be wonderful to actually study it in the way that we've been studying, if someone will give us money, we'll be happy to study it. We're quite promiscuous in terms of our scientific interests and it's only limited by the resources that we have, but I think it's very, very interesting. I think back, I don't know, what your experience was of studying, yeah, or just your experience of studying biochemistry as a student or as a medical student, and I remember particularly feeling after my last biochemistry exam in medical school, I had the following thought. This is fantastic. I'll never have to think about the Krebs cycle again. And I was very sober, though, because I had a friend here at UCSF, Mike Weiner, who's a very famous brain imager. He headed the ADNI project for Alzheimer's disease neuroimaging. He's one of the world's most accomplished dementia imaging researchers. But Mike, by training, was a nephrologist and metabolism expert. And he actually did a fellowship with Hans Krebs at Oxford. And he told me about it. And I found it. My arrogant disregard for this incredible body of work and its relevance in medicine was very humbling. The work that Hans Krebs did was just absolutely spectacular and will turn out to be so important for the understanding of psychiatric research, something most people don't know. Hans Krebs was Jewish. He grew up in pre-war Germany. And when he was in medical school, it was right before World War II. And he was working on the urea cycle, which we know is the Krebs urea cycle. Perhaps one of the most important biochemical pathways in the human body for protein metabolism. And he was working on the cycle. He had a PhD in biochemistry before going to medical school. So he was already doing advanced medical research. And he asked his supervisor, who was a member of the Nazi party, if he could have one day a week during his internship to finish work on the urea cycle. And his supervisor said, if you're so smart that you have time to do it, and he knew he was Jewish, of course, that if you're so smart to have time to work on biochemical pathways as an intern, that I'm going to assign you two medical awards as the doctor instead of one medical award. And during that year, he finished the work on the urea cycle, for which he was subsequently awarded the Nobel Prize. And I tell that story to all of my residents when they aggravate me about the 80-week work, the 80-hour work. So I sort of remember back in medical school, I was running around with some research project and having people fill out their Holmes-Ray scale about stresses and looking at correlations with other medical illnesses and so on. Then we move forward, and we have PTSD and all of that. And then we have the ACEs scales for looking at the life traumas that are really associated with tremendous medical illness, really odd diseases and terribly difficult to explain things that people come in with. They've had that much trauma. And then we have kind of the domestic one version, I think, is what they're using for most quality measures when they want you to do that. But I'm curious if you're dealing with these lifelong histories, and you're correlating, and you have so many pieces of data. And I'm thinking that in my experience, these people, at least the ones that start out with childhood trauma, their whole life keeps compounding the traumas. They get into relationships that are amazingly horrible, or sometimes they get out, but then their kid dies of an overdose. And then just all kinds of things just keep happening. And I'm thinking, what is that going to do to the data here? You're going to get repeated impulses, or I don't know what you'd call them, but episodes over and over. And I'm just wondering if you, I'm sure you've thought about it, and maybe it's just too much noise to manage. But what are your thoughts about what that's going to do to this flow? The first thing to say is it's not too much noise. I would have thought so when I was in training. But the reason, in our subjects, it may not be clear because we're going very quickly today. We have over 2 million molecular features on each subject. And we're able to manage that level of complexity, even with a few thousand subjects. I was taught when I was in training, when I studied statistics, that you had to have 10 subjects for each variable of interest. And now with the explosion of molecular biology and psychosocial psychiatry, we have enormous complexity. The first thing is we have computational approaches which are sufficiently robust, including machine learning, including random forests, support vector machine learning, and other approaches, and new AI approaches which are more advanced than that, which allow us to manage all of that complexity. So the first thing is, if we have the data, we can model it. And we can model the longitudinal consequences of the kinds of lifelong adversity you're talking about. Second thing to say is, unfortunately, in the populations I've studied, which are active duty military personnel, veterans, and police officers, their lives are loaded with adversity. Not all of them, of course. They have different pathways. One has to be careful not to simplify it. I was quite fortunate to work with Commissioner Ray Kelly of the New York Police Department when he was commissioner. He invited us to do studies on the NYPD, including these longitudinal studies from the academy through to police service. And Commissioner Kelly was a successful police officer, of course. He was a Marine in the Marine Reserves. He also happened to have a law degree. And he had a master's in public administration from the Kennedy School at Harvard. And he didn't come from an abusive family. And he was educated at the highest levels. And he didn't spend his time on the streets of New York in violent interactions, because he was identified early in his career as brilliant, gifted, and able to administrate. And he was quickly moved off the street. So what happens in the real world of military personnel and first responders, the more difficult your life was, the more aces you were exposed to, the more losses you suffered, the poorer your community, the higher the levels of interpersonal violence in your community, and so on, the more likely you end up stuck in a high-intensity exposure role. So all of these things are very important. And unfortunately, they're correlated. And also, when you return from war or you retire as a police officer and firefighter, it matters a lot what the quality of your life is at that point in time. So your points are extremely important. Those variables are very important. We believe those variables profoundly affect the molecular pathways we've been discussing today. And we believe we can manage the complexity computationally. Wow. By the way, I also have a deep learning project with Google. Try to use the state-of-the-art artificial intelligence to find out the PTSD biomarker. That's a very large scale of model. And I hope that will kind of migrate some of the noise here. Thank you for your wonderful presentations and remarkable data. It's really a lot of food for thought. And I haven't thought about the Krebs cycle either. But I work in the eating disorders and am interested in the role that trauma and PTSD play. In terms of the mitochondrial deficits showing, I'm wondering if that relates possibly to metabolic rate or a decrease in metabolic rate. And there is some literature that a decreased metabolic rate, of course, that's associated with obesity. And you've linked PTSD to obesity. And there are many studies that do link PTSD to obesity, as well as to binge eating and eating disorders. Binge eating disorder, bulimia nervosa, and so on. But one of the mechanisms of obesity, in addition to increased intake, and I'm wondering if you have data on eating behavior at all. On BMR, is lamentable. What's that? It's a wonderful question. I think it's very important. But do you have the behavioral data on binge eating at all in this group? Do we have binge eating data? Yeah. I don't think so. We know with the whole research in food addiction and ultra-processed foods, and diet definitely contributes to inflammation. And there is a story there that may interface in terms of the obesity. I think we have data to suggest, more anecdotal than systematic, that war fighters, police officers, and civilians with PTSD shift towards poor eating habits and shift towards comfort foods. And so I think there is a very important relationship, as you know from your work, between chronic severe stress, metabolic activity, and food choice. And I think that, and also, by the way, it's interesting. What's very interesting to us is when we study active duty military personnel, even those with PTSD in the military, they're in a very regimented, structured life. And they're required to run five miles so many days a week and work out. And they eat the military rations. When they separate from the military, those with PTSD often stop exercising, gain weight rapidly, use too much alcohol, some drug. And I'm sure that eating disorders or eating, poor eating is part of that. You've got their baseline BMIs. And then you have their follow-up BMIs. My guess is that the PTSD folks are going to be, have gained a lot more weight. Yes. And that they've been engaging in binge eating. And they may also have a resultant lowered metabolic rate based on these mitochondrial data. I think it would be very interesting to look at this. If you're interested, we're happy to welcome you to collaborate. Yeah, absolutely. Great data, thank you. So just let's talk about it. Yeah, and also in the active duty, according to our record, we see they consume a lot of energy drink in the morning. So a lot of monsters, red bulls, and yeah. Presumably, these are all males. Mostly. Well, not all, but the majority, about 80%, 85%. Yeah, a lot of people just early morning can start to have two monsters. Maybe they started to do exercise and they drink a lot of energy drink. Alcohol, monster drinks, nicotine, sedentary lifestyle, cravings for carbohydrates. These are actually part of a network, so to speak, which you see a lot of. And they profoundly affect these pathways, of course. So some of it is stress. Some of the effects of chronic severe stress are mediated through those changes in behavior, which then further compounds the underlying metabolic disturbance. On top of which, this data suggests there may be some shared genetic vulnerability to both PTSD and metabolic syndrome. Yes, as there is in the eating disorder. Right, so it's a very heartbreaking thing in life. We used to say the rich get richer and the poor stay poor. We're trying as a society to structurally address that, which is difficult but very critically important, right? But when it comes to health, the rich get richer and the poor health get poorer. And that's for the reasons you're suggesting. So it's very, very difficult. And then you add to that access to health care, education about health care, education about lifestyle, management of stress, exposure to lifetime stress. What we see is a virtuosic cycle in those that are privileged, and we see a malignant cycle in those that are vulnerable. And that is at the root of much of health inequity, I think, unfortunately. And that has to be structurally addressed also. My turn. So the first talk, when you were talking about the relationship between metabolic syndrome and PTSD, you mentioned kind of like three genes and two of them you named, and then you said there was another one related to immune function. So my question is, have you seen any correlations with HLA? Like, we know that some HLA phenotypes are more predisposed to autoimmune disease, like may have ATP and all these other things. Like, have you seen any relationship between this mitochondrial stuff and the inflammation stuff and whatever, and all this proteomics and everything you're doing? Like, is any of it related to HLA? Yeah. Yeah. I also have a project of HLA and T cell receptors. We basically scan every single T cell receptors of the 400... 200 PTSD veteran and 200 health control, and also 200 active duty and 200 health control. And we do see that there are some HLA correlated to PTSD, but it's different from active duty and veteran. And the same thing, we find some of the T cell receptor rearrangement is enhanced in PTSD. And we use that, we can define our PTSD biomarker based on the T cell receptors. So it can get like a 70, 80% accuracy based on those. Of course, there will be a lot of other things to change their T cell receptors. And recently we also invested the B cell receptors antibodies, and we have another project where we find that previous, based on the DNA methylation, we define two biotypes. One is more severe PTSD, one is less severe PTSD. And we find that they are totally different in the antibody level. So there are some of the immune-related antibodies pretty high in the severe group. So we are going to, we are collaborating with the NIH and try to identify what exactly that antibodies, and that could be a treatment purpose, yeah. You said antibodies are elevated? Yeah, antibodies. Do you know which type of antibodies? We don't know yet, but we know there are significant difference. But we are on the process to identify which is which. Another one of our findings, and we published this, we're happy, if anyone emails me, I'm happy to send any of the relevant papers. You can find me at the NYU website. We did a study with the UCSF group, with Owen Walkowitz and Cindy Mellon. And they asked a simple question. Is there an effect of PTSD on natural killer cell activity? And they found fundamentally that in young, mostly males, young, late 20s and early 30s, premature senescence of natural killer cells. And that you can, as people age, there are markers of natural killer cell health. And there's a specific shift from immunocompetent to less immunocompetent. Natural killer cells has a normal function of aging. And what we found in the PTSD cases is they were showing premature senescence of natural killer cells. Yeah, also we- Which, by the way, may be important for the higher rates of cancer, which we've seen lifetime in service members with PTSD. Of course, they smoke more. They've been exposed to hazards in the war zone. So it's not, obviously, causality is complicated. But one factor may be immuno-incompetence. Yeah, we also use the DNA methodation to estimate the biological age, so-called the DNA epigenetic age. And so we find that PTSD patients did have a higher biological age than the healthy control. Which, of course, goes with the- Mm? Tim? I mean, can you ask me? It's a significant difference, yeah. But what you just said goes exactly with the adverse childhood experiences. You know, the termination, these people really die young. My patients, 40s, you know, almost across the board. If you, you know, it's not a huge number, but still. But what I was going to say, it was very interesting that you brought up the eating disorder, you know, and, of course, the metabolic things that change. Because the person who did the- came up with the adverse childhood experiences data, was a bariatric surgeon, of all things. And, you know, but kept talking to his patients who were failures. These were the ones who had the surgery, and while most of them wait, and so on, the ones that would come back and they, you know, lost a little at the beginning, but then they climb right back up. And that's when he began asking them, you know, what's going on? And so on, and learned. And that's where he- how that sort of idea was developed. And ran it with a larger, you know, numbers of people to see if there's a real correlation, besides just what he found in the clinic. But it's very interesting that it was right there with, you know, the bariatric surgeon and the eating, overeating, and so forth. Come to the mic, and it's 5.15, and I'm epigenetically programmed for my OCD. My OCD to kick in. So, one of my great frustrations as chair at NYU is I ran rounds at 11 o'clock, and my faculty come at 11.03, and it's immutable, and I believe it's epigenetic. At least they do come, and they do seem to enjoy it. No, I can write that up, yeah. You get the last question. Okay, so, mine is Charles.Marmar, M-A-R, M-A-R, at NYU Langone, one word, N-Y-U-L-A-N-G-O-N-E dot org. And I have everybody else's, and so I'll share them. Anything you send, I'll distribute to the team. So, our consortium has published 104 papers on this collective work. Dear friends, this is unusual devotion, the love of learning. So, we thank you, and we're very excited about this work. It's very difficult, but we're making progress, and we look forward to coming back and continuing the conversation.

PTSD

metabolic disorders

molecular pathways

TBI

military veterans

genetic correlations

metabolic syndrome

homeostasis

inflammation

epigenetics

proteomics

metabolomics

DNA methylation