Presentation - The human microbiome lies at the heart of autoimmune disease

Type: Conference presentation
Presenter: Trevor Marshall, PhD
Conference: BIT's 2nd World DNA and Genome Day
Location: Dailan, China
Date: May 2011

Okay, so this is going to be a very different presentation because I'm going to focus on Homo sapiens, and I'm going to focus on the diseases of homo sapiens because that is what interests me most of all. I'm going to talk about metagenomics. This is a field that in just the last three, four years has gone from nothing to an explosive growth of knowledge. The early work that we saw in the, well it wasn't early work but, the earliest work was done in the marine environment, the other environments. What happens now is the libraries of genomes that have been built up allow us to very quickly using next generation sequencing technology, very quickly try and figure out exactly what organisms are commensals in with homo sapiens. This book has just been published by, edited by Karen Nelson who heads up the J. Craig Venter organization – Metagenomics of the Human Body – and it is a very good textbook of exactly what is known at the moment about the human microbiomeThe bacterial community in the human body. Many species in the microbiota contribute to the development of chronic disease.. The key thing is that we now know the human body contains an ecosystem of microbes. Thousands of species live in the cells and the tissues of the human body. Their genomes and proteomes interact with those of the host. So clinical chronic infection is no longer an illusory concept. Even though if you go to a clinical medicine conference, they'll say that persistent infection is unreal, it's very real. It's just that the microbes that cohabit in the microbiome generally are not separately cultivatable - just as in that last case you saw that huge list of different species and only one was easily cultivatable. But even more important than that, many of them don't even have their full genomes within the metagenome. They are sharing functions with other components in the microbiome.

So metagenomics got off to a very quick start with, with mice. And early work from Jeff Gordon's group at Washington University was published in Nature in December 2006. And it showed definitively that in mice, obese mice had more Firmicutes in their GI tract and lean mice had more Bacteroidetes in their GI tracts. And that you could transplant the microbiotaThe bacterial community which causes chronic diseases - one which almost certainly includes multiple species and bacterial forms. from one type of mice to another type of mice and the obesity trait would follow it. So NIH thought great, this is wonderful, let's pour a lot of money into the human microbiome, and figure out how to treat obesity and other diseases. However it turned out that murine obesity differs from human obesity.

This slide was presented by Claire Fraser-Liggett, University of Maryland, at the Human Microbiome Conference just last month. And it results from a study of Amish families that she did, a very concentrated, isolated community in America. And she was looking at lean, overweight, obese benign, and obese with metabolic syndrome. And when you look at it in terms of family, whether it's Rhuminococcaceae, Bacteroidaceae or whatever the family is, you can see that there is more variation from person to person – each vertical bar here is an individual person within the cohort – there is more variation from person to person than there is from phenotype to phenotype or disease to disease state. And, it doesn't matter, you can look at it by phylum and you see exactly the same result. No statistically determinant characteristics like we found in the mouse, or like Washington U. found in the mouse. Which is unfortunate.

Now EMBL in Europe has, is just in the middle of a big study, and they've gone even further. They have some Japanese human samples, some American human samples. This was published by Arumugam in Bioinformatics and you can see that if you look at the various phyla across here – the Japanese samples firstly are different from each other in the propensities, but also they are very different from the American samples. So there is intra-individual variation, which is quite high; and there is intra-regional variation, which is quite high. And it's at the level of the species, at the level of identifying the microbes that are present, really, at this point, human metagenomics has hit a brick wall. At this point the Human Microbiome Project knows what's there. Pretty well they know what's there, certainly on all the externally facing tissues. The next phase is to figure out what is it doing. And clearly they can't figure out what is this doing based just on the species. We have to figure out other ways of analysis.

Now this is also from Claire Fraser-Liggett's group at University of Maryland. This is a Manhattan Plot of the homology between the Fusobacter genome and the Human genome. And for more information on what we're talking about here, just read our chapter in the Metagenomics of the Human Body book because we point out the homology between genes in the microbiota and human genes is actually quite high. You can see some of the points in the Manhattan Plot here go well above the seven and a half line that Claire chose as being quite deterministic. You've got your chromosomes across the bottom obviously and log p up the left.

You see the difficulty with the human microbiome is its sheer complexity. This is the salivary microbiome which was produced by the group at Max Plank, published in Genome Research back two years ago. And this was sampled at twelve places around the planet, and multiple individuals at each location. Once again they found that there was more statistical variation between individuals than there was from location to location. But once again there was no real statistical inference that could be drawn, except that there was a heck of a lot of species that were identified in the saliva of healthy individuals. Things such as Yersinia. It's printed in fine print because there's not very much of it around. But Yersinia, of course, is the species that was responsible for the black death. I'm sure its not the exact pathogen, but a mutated pathogen, or a different species. Neisseria, Streptococcus, Prevotella, Staph's here somewhere. Anyway, most of the species that we recognize, -?, as being pathogens, are present, or represented in the healthy salivary microbiome.

So you can try and examine what's going on under light microscopy. This is an infected cell, an infected monocyte actually from a patient with Chronic Fatigue Syndrome. They tend to be very ill and its very easy to pick up the infected cells. But with light microscopy, you're fairly limited with what you can see. There's a sort of a cell layer – it looks as though the cytoplasm is somehow swollen and infected, and you can't do much.

Okay, so transmissionAn incident in which an infectious disease is transmitted. electron microscopy – that's the next step – and this is a study from the Wirostko group at Columbia University back in the 1980's. It's a juvenile rheumatoid arthritis lymphocyte – a lymphocyte from a patient with juvenile rheumatoid arthritis. And the Wirostko group studied a number of serious inflammatory autoimmune conditions and they found that there were inclusions which were staining for bacterial nuclei inside the phagocytes of the immune system - lymphocytes, neutrophils, monocytes, and macrophages. So you can see a little bit more there. But still what actually happens in the human body happens at the level of the cell. The transcription and translation and folding and the interference and everything that happens in order to produce the human proteome is at the level of the cell. And how do you study at the level of the cell? Even the electron microscope can't really get you the detail that you need to study specific metabolisms. You can florescent tag, there are lots of very good techniques for doing it, but ultimately all of them somehow alter the environment that they're pulled from.

This particular concept of the computational microscope was presented by Klaus Shulten, University of Illinois at Urbana-Champagne. Klaus really, and his group, were one of the pioneers in mathematical computer analysis of biological processes. And Klaus says that his microscope is made of Chemistry, Physics, Mathematics, software - in his case the NAND software that they developed at the University of Illinois - and of course, supercomputers. Big, big supercomputers - lots of computing power. Now I don't use NAND, I use GROMACS. I use a different sort of thing. What he is interested in doing is simulating complete metabolisms. And in this case you've got a slide produced by Klaus, a movie produced by Klaus which shows genetic activity of E. coli simulated. You've got the lacY receptors being activated by lactose on these simulated E. coli cells and then sometime later mRNA is produced within the cells and when that happens it pops out as a pink spot. You can see that the cells operate independently at different rates. Your dealing with stochastic processes obviously. But it gives you a way of visualizing what is happening that is additional to what you get with in-vitro and in-vivo techniques.

Now I haven't been doing that. I've been looking primarily at the level of the molecules. And in particular, the focus here – let me just pause this – what we've got here is a nuclear receptorIntracellular receptor proteins that bind to hydrophobic signal molecules (such as steroid and thyroid hormones) or intracellular metabolites and are thus activated to bind to specific DNA sequences which affects transcription. called the VDRThe Vitamin D Receptor. A nuclear receptor located throughout the body that plays a key role in the innate immune response.. And you've got the normal helices – you actually can't see it very well when it's not moving can you - anyway there's a ligand in the binding pocket. And this particular artifact – this helix, helix 12 – is what's known to be critical in binding to the cofactor, the DRIP205 cofactor, that allows dimerization of the VDR, and consequent transcription. But what I first showed was that this drug that acts as an agonist or a partial agonist in homo sapiens, behaved differently in the mouse. And the easiest way to do that is in fact to set up the forces, the van der Waals, the electrostatic forces, in a simulation system and see what happens. And in particular, the helix which is thrust laterally across the receptor is not bound properly to this tetrazole moiety. This tetrazole moiety on the rat in fact has two less hydrogen bonds and is just wandering around aimlessly whereas it's bonded quite strongly on homo spaiens. This drug doesn't do the same thing in the rat as it does in homo sapiens; which of course is news to the FDA because most of the approval for this drug was based on tests in mice. Now if you go a little bit further, what I've shown here is the same VDR, is the same ligand in the binding pocket. The picture was drawn at different times, so I've got different colors and stuff. But this actually shows exactly how this receptor molecule binds to the coactivator. The coactivator sits on the top across there. There are three hydrogen bonds from these two moeities, and there's a hydrogen bond from the moiety there that stabilize the DRIP205 on the receptor. And what actually happens you can see very easily under the computational microscope because what happens is that these two moeities are kept apart. You've got the double oxygens, the electron -?, the electron receptors, and they're kept apart when the ligand is in the binding pocket; but when that ligand is removed from the binding pocket, then the helices shift together and these bind firmly together and therefore can't engage the coactivator when the coactivator comes along – very very handy to visualize what's going on and figure it all out. And in statics, with more detail as to hydrogen bond positions in which residues are bound there's more details of the OlmesartanMedication taken regularly by patients on the Marshall Protocol for its ability to activate the Vitamin D Receptor. Also known by the trade name Benicar. as an agonist.

Now why is this important, why was this interesting to me? Well, in homo sapiens, and only in homo sapiens – even in the higher primates this is not the case – but in homo sapiens and only in homo sapiens – one nuclear receptor, the VDR, expresses genes for TLR2A receptor which is expressed on the surface of certain cells and recognizes native or foreign substances and passes on appropriate signals to the cell and/or the nervous system. as well as Cathelicidin Family of antimicrobial peptides found primarily in immune cells and transcribed by the Vitamin D Receptor. and beta-Defensin An antimicrobial peptide found primarily in immune cells and transcribed by the Vitamin D Receptor. anti-microbial peptides – or antimicrobials, they break into peptides - all of which are essential to the intra-cellular innate immune defenses. In other words, the innate immune system of homo sapiens is key to a functional VDR. If the VDR is rendered inoperable or less capable by a pathogen, then TLR2, Cathelicidin – which is the primary endogenous antimicrobial for intra-cellular defense – are knocked out. So this is the obvious mechanism that a pathogen would use in order to evade the immune system. If they can knock out the VDR, then they can persist very very easily. And indeed that's exactly what happens.

If you look at human chronic diseases, then you'll find that the human body exhibits VDR dysfunction in at least the following diseases: depression, multiple sclerosis, arthritis, lupus, sarcoidosis, thyroiditis, diabetes, dementia, autism, schizophrenia, and tuberculosis. It's all documented by studies of others. My colleague, Greg Blaney from Canada put together a paper recently in the Annals of the New York Academy of Sciences that will give you more details. And just yesterday, when I looked at Reuters Health, I found that the World Health Organization has declared a chronic disease alert. Chronic disease is now the primary concern of WHO, more so than infectious diseases. But the funny thing is, WHO thinks these diseases are non-communicable. These diseases actually arise from the microbiome and whether they are communicable or not is not a done deal at this point.

If we take typical pathogens that we know cause human chronic disease, or are associated with human chronic disease – EBV – it knocks down the VDR very heavily, especially in lymphoblastoid cell lines such as in the bone marrow, peripheral blood cells to a somewhat lesser degree. And estrogen receptor beta is knocked down very heavily as well. That's not surprising because estrogen receptor beta is responsible for the VDR precursor protein. Similarly for Mycobacterium tuberculosis, that knocks down the VDR receptor as part of its survival mechanism. It additionally wants to knock out another gene called TACO that's been documented in Chinese Medical Journal 2003 amongst other citations. Borrelia burgdorferi is another one that very heavily downregulates VDR expression, 50 fold by live Borrelia in that particular study, which was published in PLOSpathogens. So what we have been working with is that the understanding that the catastrophic failure of the human metabolism which we see in chronic inflammatory disease, which at first glance appears to be diverse, actually is due a single underlying mechanism, which is the ubiquitous microbiota evolved to persist in the cytoplasm of nucleated cells by knocking out the VDR nuclear receptor, thus suppressing innate immune functions, incrementally during life, and interfering with gene expression, proteome-proteome interactions, with the consequence that chronic disease sets in. And you're really dealing with an imponderable number of pathogens, an imponderable number of interactions, and a semi-imponderable number of discrete syndromes that are lumped together under this, either under the chronic disease or the autoimmuneA condition or disease thought to arise from an overactive immune response of the body against substances and tissues normally present in the body banner.

And what we've managed to do in the last 8 years is using that drug I showed you earlier to reactivate the VDR, we've been able to work with our clinical collaborators worldwide on a whole range of chronic inflammatory diseases and show the disease processes can actually be reversed. I'm not going to go into how much they can be reversed because the state of health is a moving target in and of itself. But certainly the patients can be returned to the work force, returned to their family, and be very very happy people. And what this graph shows in some great detail is by an average 36 months, range from 18 to 53 in our cohort, 81% of the cohort reported reduced disease and symptoms. This is an observational retrospective, we went back retrospective to check the data as well as the observational reports during that period. It's a very long term process to recover, to work down the microbiota. The reason is very simple. The moment you start, the moment your immune system can start to recognize the microbiota, the microbiota that have been causing relatively mild symptoms, suddenly those symptoms become worse. They become significantly worse. The word is immunopathologyA temporary increase in disease symptoms experienced by Marshall Protocol patients that results from the release of cytokines and endotoxins as disease-causing bacteria are killed.. The immune system, while it's killing the microbes, actually is producing more cytokinesAny of various protein molecules secreted by cells of the immune system that serve to regulate the immune system., more auto or self-targeting antibodies, and making the symptoms worse. We just published a paper, “Immunostimulation in the era of the metagenome,” that deals with the difficulty in trying to target the microbes. Normally you'd know that we try and suppress the immune system so it ignores the microbes. And that's good in the short term, it returns the patient immediately to the community, but in the longer term, 3, 5, 10 years, it's very bad news because the microbiome continues to accumulate and the disease states continue. That's all I am going to say, since we're pretty short of time. Thank you.

Q: What is the mass of the microbiome in a typical human being?

A: Well that too was discussed in Vancouver. And the original NIH estimate was 90% of human cells are likely bacterial cells. The current estimate is 99%. So even that is a moving target at this point as discovery continues forward. There's more money that the Human Microbiome Project is basically looking at all the external cavities and documenting exactly what's there, multiple centers, but at this point it's far more than anybody ever imagined.

Q: How much of it gets into the blood, I mean do these peptides, proteins that you mentioned….

A: Nobody knows, but certainly, you saw it, the cell that was disintegrating, so something gets into the blood. And in people who are very ill, their innate immune system is very weak, so the macrophages aren't going to clean it out very aggressively, so something gets into the blood. So what happens when it's in the blood, do we measure it when we do our genome scans? Very interesting questions. We raise those in our chapter in the book Metagenomics of the Human Body. Just what are we measuring when we measure sick people?

Q: What was the name of that drug again?

A: Olmesartan. Olmesartan Medoxomil is the drug. It's a sartan, an angiotensin receptor blocker. Its primary target is the angiotensin nuclear receptor, but unique amongst all the small molecules that are currently available, it has a significant effect on a number of other receptors; the key one in this case being VDR. And at normal doses, well at reasonable doses 4 to 6 times higher than you'd use for hypertension, you get the action on the VDR in a dose-dependent manner to turn on the immune system and create the immunopathology to attack the microbiome.

Q: Why is the human VDR so different from the other primates?

A: Well it's because of the transcriptome. There's very little data about the transcriptome available at this point. But you know one of the receptors that there was a transcriptome produced for, at least a partial transcriptome, was the VDR. And so there are a number of key P450 enzymes. The beta defensins are mentioned. It's also a primary transcriptor of the alpha defensins as well. But the beta defensins in the GI tract – very dependent upon the VDR and so they're very present in paneth cells, for example – TLR2, and the key one's cathelicidin because cathelicidin's a very long protein and it's broken up presumably by RNA interference, nobody really knows now, but into a whole lot of antimicrobial peptidesBody’s naturally produced broad-spectrum antibacterials which target pathogens. which are even effective against fungi. They go after viruses, bacteria, and fungi. And they're the primary intracellular defense of homo sapiens. Now even in the great apes, cathelicidin is not transcribed by the VDR, transcribed by something else. So when you look at the microbiome of the great apes, it's different from man. And in fact that's being done. There's been some vaginal microbiome work done on the great apes.

Q: This may be a question but, the other one that strikes me as being very important is… the microbiome is absolutely essential for our good health too, right?

A: Do we know that for sure?

Q: I mean I'm just saying presumably, if it's there …

Q: Well, let me give you one data point. When I gave the keynote at the World Gene Congress down at Foshan in 2008, I showed a slide which had Lactobacillus on it, so called “friendly bacteria.” It highly increases the expression of a gene called ACE, angiotensin converting enzyme, which is right at the heart of many inflammatory diseases. So okay, Lactobacillus may be beneficial, but it may, and is likely to also be not necessarily totally beneficial. Whenever you've got such a huge genome, as any of the bacterial genomes, huge – a thousand odd genes – but you know what I'm saying, a very large number of genes, the likelihood that they're all going to be innocuous in the human proteome of 100,000-plus proteins is very small indeed. So I don't actually agree with the concept of friendly bacteria.

Q: My actual question was… is there any hints as to what kind of antibiotics we could use to address this issue?

A: The body does a really good job. I mean all we do is use the Olmesartan to let the body start figuring out what …

Q: you said in your initial remarks, when the body doesn't, then what could we do? You presumably have to use antibiotics at some point, right?

Q: Well we have. I mean that cohort, this cohort did use antibiotics for at least part of the recovery process but eventually they just go back to the Olmesartan, to the activating drug. No the immune system is pretty good at figuring it all out. I mean these endogenous antimicrobials don't target everything equally, you know. So that's the answer I'd like to give. Yes, we use antibiotics, obviously bacteriostatic antibiotics like minocycline, the tetracyclines, the ribosome attacking antibiotics, but they're not essential to recovery.