Welcome to the Proteomics in Proximity podcast, where your co-hosts, Dale Yuzuki, Cindy Lawley, and Sarantis Chlamydus from Olink Proteomics talk about the intersection of proteomics with genomics for drug target discovery, the application of proteomics to reveal disease biomarkers, and current trends in using proteomics to unlock biological mechanisms. Here we have your hosts, Dale, Cindy, and Sarantis. Welcome to another episode of Proteomics in Proximity.
I'm your host, Dale Yusuki, with my co-hosts, Cindy and Sarantis. And this morning - Hey there - and, Sarantis, what are we going to be talking about today? Today, it's a great paper, actually. There's a great paper where it describes ways of using biomarkers - protein biomarkers - in early discovery of inflammatory bowel disease (IBD), and it is a great example of a crosstalk between gut microbioma and
immunohomeostasis. And it's really great because it combines a multiomics approach where they have investigated genetics and proteomics, and metabolomics, and also they have done sequencing of microbiota in order to identify this crosstalk. Cindy, what do we know about the DUOX2? What is the importance of this protein? So the name, DUOX, stands for Double Oxygen. So it's a
marker within the genome. And, in fact, this paper highlights some variants in the DUOX2 gene that code for something that will produce hydrogen peroxide, H2O2, in the epithelial lining. So in what is the apical layer, or the layer that's right next to where essentially the results of our food go by. And so that production has an effect on the
microbiome. So this idea that we might have a proteomic biomarker that is indicating something about the host microbiome interaction, that gut bacteria that we all are so curious about and has been the topic of so many interesting publications that we might have a protein biomarker that's indicating something about the dysbiosis or the shift in that microbiome that precedes inflammatory bowel disease, is super exciting. So that's what I see here. Inflammatory bowel disease is
devastating. I mean, I personally know three individuals who are affected. You may know others, right? Their digestion, their immune system, and it's really, really difficult because you talk about the interplay between the person, their gut microbiota, which is off. It is different than regular people and it's a complex organ, right? The gut microbiome is essentially, I mean, a lot of people talk about it as an
organ. This group, including a group out of Institute for Systems Biology, does a fair amount of work on gut microbiome, particularly in their wellness cohort. So, Cindy, why don't you tell us more about this wellness cohort? I understand this is Arivale. That's right. So Arivale, was a company that spun out of the Institute for Systems Biology. It was based in Seattle.
And the team focused on collecting data longitudinally for a bunch of people that just opted in for reconsenting, so opted in for updating their consent and recontact. And the great thing about that is they're coming in regularly. They're "well" when they start. And we'll be talking about some other papers that are coming out of this exciting group. But it offers an opportunity to look retrospectively at samples where individuals who were healthy actually will develop some diseases.
As we age, we tend to do that, unfortunately. So it's a powerful demonstration of a wellness cohort and having longitudinal sampling. Like I said, we'll double-click on that in this podcast I think quite a bit. This paper talks about over 2800 individuals from this cohort, but it was larger. Was it almost 5000, something like that? Yeah, that's right. And of course, over time, you're going to have people dropping out inevitably. And so this is a really
great sample number. And the longitudinal data that they have collected for this group of individuals is stunning. So they already had their whole genome sequence. And then what? Did they also sample their microbiome, like from feces? Yeah, exactly. But then they don't need to necessarily collect data on all the samples in all the modalities. What they've done is they've seen interesting things evolve in this cohort and then do specific, work within
a subset. And that's exactly what happened here. So a subset of the individuals from this 2800 subset of the Arivale cohort had these DUOX2 variants and those DUOX2 variants had some, I would say this sort of brings me back to this idea that if we had - what was the total number, Dale? You pulled it up for me right before the podcast. 300 and some odd individuals? It was like 357 individuals. With rare variants.
12.4%. Yeah, if we were going to do a GWAS on 357 individuals and compare that to control samples, that's not a very powerful GWAS by our standards, in the genetic space. And so making the association between these variants and inflammatory bowel disease, or Crohn's disease, or this IBD phenotype, might have been very challenging. And so this is a novel discovery, this DUOX2
association with IBD. And it was made through an association that includes phenotypic data that is closer to the disease, which includes proteomics. That, as well as the microbiome. So the ability to have this multiomics as Sarantis highlighted and amplify the power to detect the relationship between genetics and disease is something I think is exciting. We end up with whole genome data, right? And we also have then all the clinical
laboratory measurements. You have metabolites, where apparently they measured some over 950 metabolites from these patients or these volunteers, healthy individuals, it looks like they used three Olink Target 96 panels for a total of 266 proteins. And then the microbiome, which is a really important part, they're doing 16S ribosomal RNA sequencing, which gives you this idea of what the constellation of bacterial species looks like just from sampling rsRNA. But then they do a
PheWAS. And maybe we can walk through that a little bit? What is a PheWAS again? I don't know. Sarantis, you want to talk about it, or cindy? I guess I'm not so familiar with a PheWAS, actually, to be honest. I mean, it's one of the few times I have seen that or come across that. But I checked a little bit in the literature. It has to do with phenotypic associations and how these are correlating with the changes that happens in the genome, for
example. And that's how phenotype is like the correlation. And the phenotype, their phenotype is IBD. So they have individuals with IBD from these 2800 people, and they're
What's associating out of all this multiomic data? You just say, "Well, Dale, there's how many different kinds of microbes in the stool?" We're looking at 266 proteins and 950 metabolites and all these other clinical measurements. But then they zero in on this one particular enzyme. Or is it an enzyme? This DUOX2? Yeah, it's an enzyme. Enzyme. And what is it doing? What is the function of DUOX2? Producing hydrogen peroxide actually kills bacteria. And that is adjusting the microbiome, it
appears, right? I like to use the language that we're unveiling here in this paper - these authors are unveiling - a little bit about the mechanism. Super exciting. But there could also be complexity that they're yet to uncover here. As far as that particular function, we have this idea of hydrogen peroxide, and we normally use hydrogen peroxide to clean up. It's antibacterial. It oxidizes. Exactly. What is it then? DUOX2 is normally doing what in the
body? I mean, DUOX2 in the context of these individuals is keeping a certain type of gram-negative bacteria away. We're talking about homeostasis at the end. Whatever happens in biology is homeostasis. It's like the equilibrium. And when you have the shifting of equilibrium, then you have all of these pathologies that's at the end, according to Greek philosophy also, I thought, let's talk about the equilibrium. And keeping the equilibrium at the end. That's pretty much what we have in our body.
Looking there, yeah, that's the thing. And looking a little bit at the rare mutations, they have mutations of frameships. Let's say mis-sense. We have different types of rare mutations in the DUOX2 gene that leads to loss of function. They have heterozygous individuals with loss-of-function allele. And, of course, this affects the level of DUOX2. And, as a consequence, they have seen this increase of IL-17C.
But just to repeat what Sarantis said in terms of homeostasis in the normal, healthy gut with a normal, functioning DUOX2 sort of enzyme right at the surface of the gut lining, you've got gram-negative bacteria held at bay because of the hydrogen peroxide. Apparently there's some kind of segmented filamentous bacteria that's also a healthy bacteria that's associated in that side of the gut that's facing the external environment.
But we have these individuals with these DUOX2 mutations, where DUOX2 is not doing its job correctly. This PheWAS, this genetic association. And then we have the gram-negative bacteria disrupting, right? No more segmented filamentous bacteria, gram-negative bacteria invade. And then we have this IL-17C being activated. And so now we're talking about no more homeostasis. It's dysbiosis. Dysbiosis. IBD, right? The Greek word of the day. There you go.
And then we have the same cells that are being affected by these gram-negative bacteria producing IL-17C in the system, in the body that can be picked up. In mice, this DUOX2 deficiency and this IL-17 elevation are associated with this expansion of proteobacteria, sorry, pathogenic bacteria. That was pretty fascinating, right? Where they're able to take the same system and draw these conclusions even down to invertebrates. I'm like, wait, we're talking about DUO function in
invertebrates. Primordial, they say, right? So you've got your primordial DUOX2 producing hydrogen peroxide. You've got, in mice, this DUOX2 deficiency increasing IL-17C levels in the intestine and proteobacteria elevation. And then you've got in IBD patients this evidence. I love this approach to see multiple lines of evidence in different systems, especially in such a historic pathway. Historic meaning going back to invertebrates in evolution. Evolution, exactly. In C. elegans,
in worms. And here, even this whole idea of, we've got this mechanism that they're able to look at via - what did they use - knockout mice? Yeah. And so these mice were deficient in DUOX2. They can go ahead and do a lot of experimentation on those tissues. And then there is something called a colonoid. And I'm like, colonoid? What's a colonoid? Sarantis, what is a colonoid? Again, I'm taking the difficult questions now. I try to be creative also in the way that I'm answering. Organoids. A lot of
buzzwords, different things. They called organoids, I'm guessing, that what they mean here is just a culture. They take a tissue and they culture this tissue, like primary cell culture, in a way, but keeping the integrity of the tissue. And they can do measurements of the proteome based on this tissue, I'm guessing this is pretty much what they are doing there. I see. And this is what? Colonoid. But mini colons
from primary tissue. Mini tissues in vitro or something, or 3D tissue culture or something. To take then that finding in mice and then being able to translate it to humans, to say that they believe that these sort of deficient individuals. And I thought one of the interesting figures was just how rare these variants were, but yet it was in certain populations it was like the odds ratio was really high. Was it Ashkenazi Jews
as a population? Yeah, that's right. And then there was another cohort that it seemed like these odds ratio estimates were pretty high because they were enriched. Right. These populations have a lot of these DUOX2 mutations in the population. I just thought that was really interesting, where you can have then maybe a genetic test where somebody's susceptible. I will say, though, the IL-17C levels were high in other subjects that didn't have that DUOX2 variant enrichment. Didn't have
that DUOX2. So you have the option, a possibility suggested here, that you could do a genetic test to see about those rare variants. But in fact, the IL-17C levels were elevated. There was a common response, sort of a cascade, that suggested this. I don't know if they use the word inflammatory, but in my mind that's what I
picture. And I think this is where I think the excitement around proteomics is, but also with the DUOX2, just going back to the DUOX2 variants, this is kind of how I think about proteomics in the context of genomics. If you have a DUOX2 rare variant that suggests an increased IL-17C level - Dale, we'll just pick you - and I don't you see how
I end up in the good category? And then you have an exposure to a protein level over your lifetime, perhaps that I don't, that exposure internally in our body is a way that I sometimes think about the relevance of how we make these connections between genetics and protein levels. But anyway, that's just a way of thinking about it that I think. No, that was great. And then I will add also the macrobiome factor here because they have done a really crazy experiment and they really like it in mice.
They use antibiotics to eliminate actually the negative, the gram- negative and they see that IL-17C levels, they drop down again. That means there's really close talk with microbiome. One of the applications of this paper is: could fecal microbiota transplantation, FMT, or antibiotic treatment for people with this dysbiosis where the gram-negative bacteria are where they shouldn't be is pretty, how do I say, remarkable.
To think this is a potential sort of mechanism, by which, again, knowing the mechanism suggests therapeutics, and even though the complexity of DUOX2, and then the IL-17C, and then there's a whole bacterial story, and then the immune system ... There's T-helper cells. CCR6, the chemo attractant for lymphocytes, there's FGF23 There's just a cascading mucosal immunity, intestinal mucosal immunity that has a protein signature that I think can be shifted and has that potential to
provide value. Well, we should probably highlight the Crohn's and Colitis Foundation, actually, in this discussion. Right. So Crohn's and Colitis Foundation, of course, has developed a subset of proteins that are useful in helping identify in a pediatric cohort which of those kids is likely to develop complications from their diagnosis of
IBD. So, again, a longitudinal cohort that they followed, and they came up with this signature that helps them have the potential to score and insert a pause between taking out a kid's belly, which seems like a pretty good use of a proteomic signature. So, that's an important story. We've got several webinars from the team there. Yes. One of the interesting take homes, getting back to the IL-17C story ...
I'm reading from, right before the discussion, high IL-17C in carriers of DUOX2 loss-of-function variants is not only a potential biomarker for disturbed gut microbe immune homeostasis, but appears to reflect an early stage of IBD pathogenogenesis. So here now we're talking about a biomarker that could be an early predictive marker of disease. And you think, wow, by studying genetics, by studying phenotype, by looking at multiomics
we come up with a plasma biomarker. And you're doing it simultaneously, or they're doing it simultaneously. So I don't know if you remember the Alnylum story, where hereditary amyloidosis has this diagnosis that's based upon a Gait test, like your walking test in your doctor's office. So that is sort of a difficult thing to diagnose, but you can have the genetic test very and know you've got a predisposition for it. You just don't know if it will ever penetrate and if you'll ever actually
be diagnosed with it. It took time for them to identify a protein biomarker that had promise for that diagnosis. Here they're doing both at the same time. So here you've got a potential for a genetic test, but maybe these people never develop IBD. And so you can do the genetic test, and then you know who you will have to monitor over time and again. This is all research use only, right? But we're talking about the potential for the future. What would be the clinical
utility of something like this? And that seems like -- A comment on that. Probably more to a philosophical point-of-view, it's like, at the end, having a genetic test will help for a prediction because there are so many rare, let's say, mutations, that at the end, you will never know the real levels of your protein. Because at the end, what people care about is the proper levels of DUOX2, independent of the mutation or not.
That means probably you will need the protein biomarker, the IL-17C, and all the cascades that follows to be more sure and more concrete in what we are seeing. That's the way that they see, because so many rare mutations, difficult to predict the levels of the protein. I think that's really important to have a plasma biomarker to follow at this point. Yeah. And everybody's not going to use our 3K, our Explore panel, right? We're talking about something that's developed for the clinic that's very
specific. That would need to go through regulatory approval to get into clinical utility. But yeah, very exciting. And you think this could be ... well, in the commentary, the researchers saying patients and physicians and scientists are looking for how to unlock this microbiome host and then immune system as like a Holy Grail within IBD. And I thought that was so interesting, that here it is. They're looking at so
many different variables. When you think about whole genome data and all the genetics that could be evolved, and then you think about all the different other measures that they did, and then to settle on a model, a gene, a particular plasma-based biomarker, and then looking at how they all interact, it's just really interesting. And down to the types of microbiota that are being affected. It's so cool. I mean, it's a
big data story, right? It's how big data and a cohort that's collected and consented longitudinally - I know I've already said this - can drive mechanistic discovery to help define disease biomarkers. And I think having a biomarker is great, but having it where you actually kind of have a sense of the, like these multiple lines of evidence - a mouse model - having multiple lines of evidence and some mechanistic understanding of it, makes it so much ... I'm much more comforted in
seeing it implement in the clinic. And I would hope it has the potential to move to the clinic a little more quickly when we actually have that mechanistic insight. Sorry, go ahead. I was going to say that after reading a paper like this, because it was pretty dense, it's pretty intense, right? It was, I don't know, maybe 12 or 14 pages of heavy duty reading and lots and lots of immunology, which can be difficult to grasp. We had
mouse models in terms of dual knockouts. We have all these multiomics, and yet it's almost as if, man, this is like the final word, right? This is one conclusive particular avenue. And then it makes me think, wow, it was only 250 or so proteins in the plasma [that were measured]. What if they did an Explore 3K on it, right? And then I think, wow, I don't know. Cindy, maybe you can shed light on this. Does the UK Biobank collect, the microbiome samples from individuals?
Not to my knowledge. I haven't seen any studies on it, which I would expect. I know they're doing metabolomics on samples in the UK Biobank, but my understanding is it's on plasma. Don't quote me on that. But I just haven't seen any data come out on the microbiome on those patients. Because that data set, that 16S ribosomal data is super important, because that's one-third of the whole story here in terms of how that microbiome is interacting with the host. Any
final comments on this? I mean, this is such a cool paper, right? Even though it was published in 2021. This association with inflammatory bowel disease and its practical application, and a wonderful multiomic story. I mean, it'll be really interesting to follow.
Yeah, it's a great story. And so I'm thinking in the show notes, we can maybe even put some of these proteins in our Insight app and provide a link to the ability to browse through those, because I think looking at these pathways might be really interesting. And just to remind folks around our Insight app, we've had a podcast about this before, but essentially it's a browser where you can very quickly what I love about it is that you can actually convert gene names to
UniProt IDs. And with UniProt IDs, there's only one UniProt ID per protein, whereas we've got multiple gene names for them. So that's really handy. But also being able to just look at the pathways and see which proteins we have in our panels, that you can get at that pathway through multiple proteins versus the ones that we don't have in our panels. And that also helps people to identify the ones
they want us to put in the panels. And we have a mechanism by which you can request proteins to be included in our future product development efforts. And so we've talked about how we're trying to cover the proteome. We've got 3000 [proteins that our platforms measure] today, but certainly our R&D team is working hard on covering more. Yeah, thank you for bringing that up, Cindy. Insight.olink.com is how to access this free resource. There's some really great
tools inside it. Even if people wanted to browse publications by biomarker, right? Very good point. Wouldn't that be interesting to punch in IL17-C and see what other publications come up? Which, by the way, is a low, abundant protein, right? So if you look in our validation data, also available on our website that you can freely download, you'll see that IL-17C is in the dilution category of one-to- one. So it's neat, right?
You're not adding any dilution factor to it to manage how much reagent might be needed to count when our NGS readout or our qPCR readout is used. And so that one-to-one ratio suggests that it's in that area that we've talked about before, where Olink has really shined a light that makes it much easier to see these low, abundant proteins than some traditional methods that we, of course, have already made lots of great discoveries using mass spectrometry, for
example. But this low abundant area, I think the number of publications we have in our 1100-plus publication database, many of those are focused on those one-to-one neat proteins that are in that low, abundant range. Speaking of show notes, I'll be sure to include Dr. Uh, Hurtada-Lorenzo's talk from the Crohn's and Colitis Foundation. This is where they were looking at pediatric Crohn's and colitis.
And it's a remarkable story in terms of was it maybe 70 or 80 real-time PCR markers that they had developed from biopsy tissue, but then they went to plasma and they found a much smaller signature that's much more practical. Instead of using biopsy tissue and real-time PCR, here it is. They're able to look at circulating biomarkers. Some really exciting work.
And that team has created, of course, it's a nonprofit Crohn's and Colitis Foundation, but they've created a ventures arm for that entity in order to bring investment in to take things like this to the clinic. And they're funding the majority of Crohn's and Colitis research in the world.
In fact, I was at a meeting in South Africa, in Cape Town a couple of weeks ago, and there was a poster on IBD, and I talked to the authors, and in fact, they were also funded by the Crohn's and Colitis Foundation. So it's just a small world in some of these disease areas where the movers and shakers are really making a difference. Well, thank you for joining us this afternoon or this morning, wherever you may be, and, Cindy and Sarantis, till next time, so long.
Thank you for listening to the Proteomics in Proximity podcast brought to you by Olink Proteomics. To contact the hosts or for further information, simply email info@olink.com.