IMPACC longitudinal metabolomics and genomics analyses show that disruptions in one‑carbon/methionine metabolism together with MTHFR C677T genotype at hospital admission improve prediction of severe COVID‑19 and long COVID risk
0:00Welcome to Base by Base, the papercast that brings genomics to you wherever you are. Thanks for listening, and don't forget to follow and rate us in your podcast app. You know, for all the incredible progress we've made against SARS Kobe too.
0:11I mean, the vaccines, the antivirals. We are still dealing with this huge lingering public health crisis. Long CVID. It's just, it's a massive problem. It's not a marginal issue at all. When you look at patients who are hospitalized.
0:25Some studies are showing that up to half 50% of them are left with these persistent debilitating impairments. Yeah, physical, cognitive, even mental health issues long after the virus is gone. It's this enormous, long tail burden on, you know, people's lives in our healthcare systems.
0:41So that brings us to the, I guess, the $1000000 question for medicine right now. What if a doctor could look at a patient? Right? When they're admitted to the hospital. We're talking within the 1st 72 hours.
0:54And no, with real precision, who is at the highest risk? Not just for getting severely ill. Exactly, for acute severity, for mortality, and for developing those long-term problems. That would just, it would change everything.
1:08Triage, treatment, the whole ball game. And that's what this deep dive is all about. We're looking at a foundational molecular access that, well, it makes that kind of early prediction a real possibility.
1:18So it's not just one thing. No, it's a synergy. It's not just a single gene, and it's not just one marker in the blood. It's the combination of a very common genetic predisposition and a very specific, acute metabolic crash.
1:33A metabolic crash. You can think of it like a molecular crystal ball. It lets you do this precision risk assessment. The 2nd a patient walks through the door. We're going to unpack how these researchers proved what they call the 2 hit hypothesis.
1:45The 2 hit hypothesis. I like that. The idea that you have a pre-existing vulnerability, and it only really explodes when it's hit by an acute infection. That's the one. The really elegant concept, and I think we can use it as our guide through what is, you know, a pretty complex set of data.
2:01Absolutely. But before we get into the nuts and bolts, we have to pause and just acknowledge the massive effort that went into this. Yeah for sure. We're celebrating the incredible work of Boreana Petrova, Joanne Darier's, Namma Conorak, and really the entire Ampas Network.
2:17Their success came from leveraging this huge longitudinal cohort. And, I mean, this was the key, solving the puzzle of how to integrate these incredibly complex data sets. Genomics and metabolomics. Exactly.
2:30In a way that really hadn't been done effectively for COVID-19 before. And that integration is really at the heart of personalized medicine, isn't it? Especially since severe COVID, as we said, is still a major problem.
2:40We need tools that are way more sophisticated than just, you know, checking someone's age or oxygen levels. Right. So to really get this, we have to define the metabolic playing field. We're going to be talking about one carbon metabolism.
2:53Okay, what is that? Think of it like a cellular factory. It's the part of the cell that supplies all the essential raw materials, these one carbon groups, that are needed to build new things really fast.
3:03Like building blocks for DNA and RNA. The building blocks. Exactly. And Sarscovi too, like pretty much all viruses is a master thief. It hijacks the host's machinery. Specifically, this one carbon pathway, to crank out its own viral RNA. So if that host pathway is weak, the virus just has a field day.
3:22It is a field day, and the researchers, they didn't just look at the whole factory. They zeroed in on one very specific department. Okay. They focused on something called the methionine cycle. It's an absolutely critical part of one carbon metabolism, and its main job is to produce a molecule called S adenosyl methyanine. Or Sam. Or Sam.
3:41Sam is. It's often called the universal methyl donor. It sounds like a tiny labeling machine. That's a great way to put it. It provides these little chemical tags metal groups for almost all cellular methylation reactions, you know, tagging DNA, RNA, proteins.
3:57You compromise the cycle, you compromise basic cell function. Okay, so if the methane machinery is the system the virus is trying to hijack, that brings us right to the 1st hit in your hypothesis, the genetic part.
4:07It does. The key genetic player here is a gene called MTHFR. It sits at this critical intersection that connects the folate cycle to the methanine cycle. And the study looked at a really common variation in this gene, right?
4:22A very common one. The C677 polymorphism. Specifically, they were looking at people who are homozygous for the AA variant of this polymorphism. So how common is this? I mean, how many people have this subtle metabolic vulnerability?
4:37It's surprisingly widespread, which is why this is so important globally. This homozygous AA variant is present in, what, roughly 10 to 15% of the world's population? Wait a minute. So you're saying one out of 10 people in a hospital waiting room could already have this sort of metabolic handbrake pulled.
4:54Precisely. And the effect is pretty big. Carrying this version of the gene reduces the enzymes activity down to only about 30% of normal. It creates this backlog in the whole system, which leads to higher levels of a marker called homocystine.
5:06It basically sets up a compromised metabolic state before you even get sick. This is the 1st hit. That's the 1st hit. Wow. Okay, that really puts the vulnerability in context. So how do they prove this connection in actual COVID patients?
5:19The methodology sounds like a huge data headache. It was an immense undertaking. They use the NIH and NIA supported ImpactC cohort. This is like the gold standard, a perspective longitudinal study. So they followed patients over time.
5:34Yes. 11,164 unvaccinated patients who were hospitalized across the U.S. during the 2020 and 2021 surges. The whole idea was to track how a patient's initial biology dictated their health journey. And the real magic was in combining the different types of data.
5:50That's it. They brought together 2 different scientific lenses. First, genomics. They genotyped every single patient for their MTHFR allegal status. That's a fixed piece of data, right? It never changes.
6:00Then second, a metabolomics. They did these incredibly detailed plasma analyses at different times, but they started with that crucial visit one within 72 hours of admission. That 72 hour windows just terrifyingly fast.
6:12It means this metabolic crash is kicking off before a lot of the classic clinical signs even get bad. And that's what makes a prediction so valuable. Absolutely. The whole study really hinged on integrating data from 2 different, very specialized metabolomics platforms, and then using some pretty sophisticated stats likelihood ratio analysis to test the predictive power of combining that fixed genetic piece with the acute metabolic stress.
6:38And they define their outcomes very clearly. Oh, yeah. Disease severity was split into 5 trajectory groups, TG1 to 5, with TG5 being fatal. And long COVID was categorized based on really detailed patient reported outcomes.
6:52Okay, so let's get to the findings. What did the data show in those critical 1st 3 days? What was that 2nd hit they saw in the blood? The early signal was just, it was unmistakable. The metabolomics data showed very specific disruptions in the methane cycle that correlated directly with how severe the disease would eventually become.
7:10And this was clear right from visit one. Visit one. So which molecules were, you know, screaming the loudest? Two in particular were really strong predictors of a bad outcome. They saw significant increases in methane, sulfoxide, and SAH, which stands for S.
7:27Adena, sulhomocystine. Both of these metabolized just tracked perfectly. The higher the levels, the more likely a patient was to end up in the most severe trajectory groups. Can you break down what those increases actually mean for a cell that's under attack?
7:41Sure. Mythian insulfoxide is like a textbook marker for severe oxidative stress. It tells you the cell is basically drowning in free radicals from the infection. Okay, so massive damage. Massive damage.
7:53And SAH is a direct indicator of stress on the methane cycle itself. It means the system is jammed up and can't clear its own waste products efficiently. And on the flip side. On the flip side, the amino acid serene, another vital building block for this pathway was consistently down.
8:09It was decreased by almost a twofold factor in the most severe patients. So the whole supply chain was just collapsing, right at the start. The cellular supply chain was collapsing. Yes. Now this brings up a really interesting point.
8:20You mentioned the MTHFR genetic status alone wasn't a slam dunk predictor. The P value was 0.077, which is, you know, it's a trend, but not statistically conclusive. So why didn't they just discard the gene and say, okay, it's all about the metabolites.
8:36What made them stick with it? That is a brilliant question, and it really shows the insight of these scientists. They knew, from a biological standpoint, that this pathway was essential. They knew the link between MTHFR and the Methianing cycle was fundamental.
8:51So the biology pointed them in that direction. Yes. A P value of .077 doesn't mean there's no effect. It just means that in isolation, The gene doesn't guarantee a bad outcome. Because your environment, your diet, inflammation, the acute stress of the infection.
9:06It all plays a huge role. They suspected the gene was like the unexploded bomb, and the infection was the trigger. So they didn't scrap the idea, they proved the synergy. They proved the synergy. And here is the core result, the finding that really shifts things.
9:18The combined predictive model. When they integrated the MTHFRAA allele status, with those acute metabolite levels, SAH, methionine, and methionine sulfoxide, the ability to predict mortality, just it absolutely skyrocketed.
9:35Can you give us the hard numbers on that? How do we know the combination was so much better? Yeah, so they used a statistical measure called the Akike information criterion or AIC. Basically, a lower AIC score means a better, more powerful predictive model.
9:49lower is better. Lower is better. A model with just the clinical data had an AIC of around 791. When they added the gene and the metabolites, the AIC dropped dramatically to about 739. Wow, that's a big drop.
10:03It's a huge drop. And it proves that the combination gives you so much more predictive power than either factor on its own. It's not just addition. It's a synergistic effect. They multiply each other's value.
10:12And this wasn't just about predicting who would survive the initial illness. It also helped forecast that long-term risk. That's the huge clinical application. This integrated genetic metabolic factor was also highly predictive of long COVID.
10:25Patients who carried that MTHFR, A-A-A-L, they showed a much greater disruption in their Methianine cycle metabolites at Visit one. If they were one of the ones who later developed chronic deficits compared to those who recovered fully.
10:39So the pre-existing vulnerability just amplified the acute damage, and that damage stuck around. It led to lasting damage. That's the 2 hit hypothesis in action. Okay, let's circle back to the mechanism.
10:52Why does messing with this methioning metabolism cause so many problems, both during and after the infection. It has to be tied to the immune system. It absolutely is. We know the cycle is critical for proper T cell function.
11:06Right, the cells you need to clear the virus. Exactly. They're your cellular cavalry. If your T cells can't work properly because their underlying metabolic engine is broken, the virus just has more time to do damage.
11:16And like we said with methylian sulfoxide, It's also a sign of massive oxidative stress. And there's another fascinating theory, long COVID specifically, right? There is a really interesting one. Methionine metabolism is known to regulate the latency of other persistent viruses like Epstein bar virus or EVV.
11:35Which a lot of people carry. A huge number of people carry it. And there's a strong hypothesis that the reactivation of these latent viruses could be contributing to long COVID symptoms. A compromise methionine system might be the very thing that allows those sleeping viruses to wake up.
11:51This all moves the conversation firmly into what we can actually do about it. If we can get this data within 72 hours, How does that change patient care? The actionability is the endgame. I mean, since these markers are so informative at visit one, you could use them to stratify patients immediately.
12:07Imagine someone comes in and instead of waiting for them to get worse, you run these 2 tests. A genetic test and a metabolite panel. And based on that, you immediately move them from a standard ward to a unit with much closer monitoring, maybe more specialized care.
12:22That's the power of this kind of precision risk assessment. And what would a personalized treatment look like for someone you identify as high risk? Well, it opens the door to targeted nutritional or even pharmacological interventions.
12:36If you know the me signing cycle is struggling, maybe you could start aggressive nutritional support with co-factors like high dose folates, or certain amino acids, to try and shore up that system and fight back against the crash the virus is causing.
12:51We do have to talk about the limitations though, as the authors did. This was an observational study. Correct. So we can't definitively prove causality. We know this perturbed metabolism correlates with getting sicker, but we don't know for sure if it causes it.
13:05It's a bit of a chicken and egg question right now. And practically speaking, these tests aren't available everywhere. That's the other big one. The specialized metabolomics platforms they used in this research.
13:15They're just not standard in most hospital labs yet. So given those equipment constraints, what's the most practical next step for the research community? I think the focus has to be on translating this insight into more accessible tools.
13:27Yes, we need prospective studies to validate everything. But critically, I think researchers should investigate using a very cheap, very accessible marker. Homocystine levels. Ah, because it's directly related to MTHFR.
13:41Exactly. Given its strong link to MTHFR activity and the whole methanine pathway, it could serve as a really useful proxy clinical marker, you know, for that initial risk stratification, while the more sophisticated platforms are rolled out.
13:53So when we boil it all down, what's the ultimate take-home message for you, the listener? It seems to be that predicting severe COVID and long COVID? It means we have to move beyond just looking at one thing at a time.
14:04Yes, that's it. The study provides just crucial evidence that the synergy is everything. It's the combination of a patient's MTHFR genetic predisposition, that pre-existing vulnerability, and their acute metabolic status right when they get to the hospital.
14:18A two-pronged biomarker. A powerful two-prong biomarker that allows for a kind of personalized risk stratification we haven't had before, right when it matters most, which really leaves us with a provocative question for the future of individualized care, doesn't it?
14:34What does this mean for actually developing targeted nutritional or pharmacological interventions that could head off these metabolic disruptions in the high risk people we can now identify? This episode was based on an open access article under the CCBY 4.0 license.
14:51You can find a direct link to the paper and the license in our episode description. If you enjoyed this, follow or subscribe in your podcast app and leave a 5 star rating. If you'd like to support our work, use the donation link in the description.
15:03Now, stay with us for an original track, create especially for this episode, and inspired by the article you've just heard about. Thanks for listening and join us next time as we explore more science, base by base.
