Dr Mady Hornig, MD, Director of Translational Research at the Center for Infection and Immunity at Columbia University, made a brilliant presentation at the RME Sweden Severe ME conference in late October last year.
Dr Hornig went into more depth about their published cytokine work, as well as what they are working on and trying to achieve. The talk was jam-packed with great science and information. Dr Hornig talks about the crisis in funding, looking at gene expression and gene variants, screening for up to 1.7 million vertebrate viruses, metabolomics, looking at how the immune system and the microbiome could affect metabolism and the brain – and much, much more!
Here’s the full transcript of the talk, together with most of the slides. It is a long piece – but very much worth reading! Even if you can’t quite grasp the science, it is almost impossible to not notice the depth and quality of investigation and effort going on with Columbia CII’s research into ME/CFS. Thanks to RME Sweden it is available to be viewed on YouTube.
There is a paucity of research and funding going into ME/CFS, you can help Dr Lipkin and Dr Hornig, and their world class research into the illness at CII by donating directly
Disclaimer: This is an unofficial transcript, it has not been approved & does not perfectly match the talk in order to optimise reader experience.
Dealing with heterogeneity (sub-groups)
Thank you very much. That was really a master class that we just experienced. It was a really wonderful introduction to the concept that lead us to our research is really trying to find ways to parse out what we believe is perhaps some significant heterogeneity in ME/CFS, both on the continuum of severity of illness, but also perhaps the duration of illness in addition and perhaps in terms of different stimuli that are important in the disorder. We were really keen to try to find biomarkers that might help us to identify and parse out this heterogeneity.
We have been fortunate to have many researchers including Dr Petersen, and a host of United States researchers. I second that motion or third the motion that we should really have a global representation in our clinical sample sets and our patient population, not only so that we can get to answers faster but also to ensure that the answers that we’re finding in one country or one location, are generalisable to other areas, and there may be a multitude of stimuli, including viruses that could vary in their identity, but yet have similar responses across the world and we need to understand that more deeply.
Columbia’s ME/CFS cohorts and studies
We began our work looking for the XMRV/PMLV answer back in 2010 and set up a network of clinicians with 5 sites, 150 cases 150 controls, the blinded analysis and have the samples that are still banked that we’re able to use. Also fortunate through funding from the Chronic Fatigue Initiative funded by the Hutchens Family Foundation, to have collected and characterised 200 subjects along with matched controls and embarked on a very wide range of analysis, including looking for a pathogen, trying to define some of the immune signatures that may be associated with different subsets, metabolomics and proteomics.
And more recently have completed a 50 case, 50 control subset analysis, where we are looking at the microbiome as well as a longitudinal immune analysis, a year to a year and a half after the first sampling.
I want to mention just for a moment the importance of the matched controls when one is looking for the possibility that an infectious agent may be involved in kicking off this process. And so we match on season of the year, we match on the time of day of blood sampling because circadian rhythms can alter these phenomenon. And we also match on geography as well as some of the more classical matching variables like sex and age and you know socioeconomic factors, race – ethnicity. I think it’s important to note that because it’s really critical I think for the comparison because one can be fooled if you collect all of the controls during the flu season for example and you have cases collected in other times of the year, you can have very different results both in terms of the pathogens that you might find as well as the immune disturbances.
We’ve worked also with Dr Montoya’s group doing pathogen discovery, spinal fluid study with Dr Petersen, I will show you some of our immune results there as well. Looking at some more unusual phenomena, some patients who have more, you know, strange or different types of presentations or a high rate of development of cancers particularly Lymphoma’s and trying to understand whether the stimuli may be different there.
And now we are just embarking on a new collection of samples that will allow us to tap the microbiome at multiple time points as well as the plasma samples at multiple time points throughout a year, utilising the clinician network. Dr Petersen as well as a number of other physicians who were involved in these other studies. We’re hoping that this will be foundation for establishment of a center of excellence in ME research that will hopefully ultimately have a global component.
A crisis in funding: need recognition of this as a disease worthy of consideration
I would like to emphasise however that the reason that we continue to be concerned and really devastated by the crisis in funding is that although we were very happy to have NIH funding, (National Institute of Neurologic Disorders and Stroke), again we are just starting this collection – it only covers for half of the sample collection for microbiome immune and pathogen discovery banking of samples. It has zero dollars for actually running any of these assays and indeed we had to add half a million dollars in additional funds from other sources in order to be even able to complete that collection, so it’s a very important scenario that we need to address and get this problem of funding and recognition of the biologic nature of this as a disease that’s worthy of consideration at the NIH level and other high level research funding agencies.
We are used to using an approach that helps us to try to connect, in a chronic disease state, try to connect pathogens as well as host responses to environmental factors to try to understand these disorders.
And our laboratory has really embarked on producing many of these techniques, there’s Mass Tag PCR, can do multiplex assays to look for the agent with high throughput sequencing and also we just recently published, Ian Lipkin and the team at the rest of the lab at our center have published about a new technique that will allow us to identify with sensitivity better than the standard molecular techniques – any virus that has ever been found to be in a vertebrate animal – so it’s 1.7 million agents that are tapped through this one reaction and relatively low cost, so we’re very excited about that and are hoping that will increase our yield.
And we’re also doing the microbiome analysis which is including both the bacteria in the gut as well as in oral pharyngeal areas, where many patients are starting with a sore throat, swollen glands in the head and neck, also the fungal agents through the mycobiome analysis.
Looking at gene expression and gene variants
In addition we are also are used to, in trying to figure out what’s going with a chronic disorder, to really add in the host response and do that through a variety of techniques. Looking both at the RNA level, through RNA Seq or microRNA approaches, to look at the abundance of the transcripts produced by various genes. We have data that are about to come out in that regard, and these may tell us something not only about which genes are expressed, but can also be informative as to what types of gene variants may exist in the population.
We know many immune response genes may be a link in other disorders to vulnerability to infectious agents or to have an untoward or prolonged course after an infection. So we are very eager to look there to see both in terms of gene expression but also trying to detect the variants in the genes that may be involved, which could help us for earlier detection.
I’ll be speaking to you about our immune marker analysis. We’re still analysing the longitudinal data but we’re also beginning to use immune profiling to look for antibodies that may be to pieces of viruses or bacteria or fungi that lead to an autoimmune type of response.
And recently in Norway as well as other groups have found that there is some suggestion that there may be autoantibodies to receptors that are involved in the adrenergic system, the beta adrenergic system, in control of blood pressure and pulse and autonomic nervous system responses. They’ve been just recently detected in ME, and indeed as Dr Petersen alluded to, you know Rituximab is a B cell depleting agent is also used not only in cancer situations but also in autoimmune diseases, and whether you know it’s safety and efficacy need to be really determined in the clinical trials but does look promising, and suggests that there could be an autoimmune subset of disease and we need to try to understand who that subset is and whether they have genetic factors perhaps, immune response genes that are involved, as well as certain pathogens that may kick the process off.
Screening for up to 1.7 million vertebrate viruses
We also are using a phage approach for anti-pathogen antibodies, again using an approach that will allow us to look for an antibody to any virus, any vertebrate virus that has ever been deposited in the data base, so that’s saying 1.7 million agents that have been identified up till now.
So that will give us a history if somebody comes to us and they’ve already – it was a hit and run event perhaps, or the viral agent perhaps has gone to a part of the, you know an organ system and is more remote, so brain perhaps or in immune glands, if a virus has retreated is no longer circulating in the blood or in blood cells we may still at least be able to find the historical trace that that agent had at one time been present and lead to a reaction in the individual, and then try to see whether that really is associated with the onset and or persistent of disease.
Metabolomics (chemical fingerprints left by metabolism both in the body, and in the gut microbiome)
We’re also using metabolomics as a tap, not only for looking at the host machinery but also as a representation of the gut microbiome because many of the metabolites that are put into the blood stream by the gut microbiome shape your immune system. 60% of your immune cells traffic through your intestines, are exposed to products as they traffic through your intestines – exposed to products of bacteria, these metabolites of bacteria that shapes the immune system. It can increase TH17 cells which are pro-inflammatory cells and then lead to an increase of IL -17 as well as a decrease in the T regulatory cells which may help to suppress certain immune responses.
Analysing 51 different cytokines
The approach that we’ve used is a 51-plex for the studies that I’ll be showing you today – it’s a magnetic bead-based immunoassay done on a luminex platform, has fluorophores that are on the bead and then a secondary fluorophore to help to quantitate. So there is the identification of the cytokine or chemokine as well as the quantitation. And we’ve identified, we’ve put together a panel that makes sense to us, not only in terms of what we know about viral infection, but also with respect to the whole host of agents that, you know bacterial agents that could also play a role, but also innate immune molecules that are known to have a role in brain responses. Your hypothalamic-pituitary-adrenal HPA axis has cytokine receptors and so for IL-1 beta IL-6 and TNF alpha, so we’ve tapped into those in building this assay.
We know that there is a signalling that occurs within the brain, the toll-3 receptor that Dr Petersen was describing. This is amped up by Ampligen, and can lead to activation of these innate immune signalling pathways that lead to the production of inflammatory cytokines including IL-1 beta as well as a host of other downstream cytokines and chemokines, with additional effects thereafter. Type 1 interferons IFN alpha and beta, and then also viral type 2 interferon as well, and these all play a role in brain physiology as well.
Why else might we want to look at circulating cytokines if we’re looking at a disorder that is involved with cognitive difficulties, obviously there are muscle difficulties in the overall bodily and systemic fatigue, but we’re very keen to try to understand the range of cytokines that may have a role in sympathetic nervous system activity, the overall regulation of the autonomic nervous system response. Again, the antibodies, autoantibodies have recently been detected to beta 2 adrenergic receptors and we think that perhaps this dysregulation of blood pressure and pulse or the improper regulation along with orthostasis may be regulated in part by improper cytokine increases or decreases over time and we need to understand that more. In building our assay we were trying to include these cytokines so we could understand this better.
There are also Allergy-related cytokines, IL-4, IL-13, IL-17A, IL-10, Eotaxin because we know there are many pathways through which the immune cells are interacting with various cytokines and leading to both eosinophilic inflammation as well as histamine production and a variety of other agents that can alter not only fever responses like prostaglandins but also can regulate blood pressure and other vascular responses including histamine.
So, in our combined analysis of the NIH study population as well as our Chronic Fatigue Initiative study population, we found that there were some very intriguing immune signatures that were present early in the course of illness, but that also, and I think the title’s a little bit misleading, as we’re focusing on the ability to perhaps have something that would identify patients early on after onset of disease, because diagnosis is often so delayed, but we also found some differences from healthy controls in individuals who had long duration ME/CFS.
So we defined this as we found our cut off as 3 years or less for the short duration or recent onset individuals, and long duration being greater than 3 years, and we found that duration of illness was more important than severity of illness in identifying these differences in our population of ME/CFS in contrasting them from healthy controls.
And what we found was quite interesting, as we had in the short duration subjects we had an increase in proinflammatory cytokines, IL-1 beta, IL-6, IL-12p40, 1L-17A, these are the short duration here versus the long, and the controls, and interferon gamma, and in many cases the long duration were different than the short duration subjects, and also there was a difference from control subjects as well.
There were also the so called counter-regulatory or anti-inflammatory Th2 type cytokines that were increased in the short duration subjects as well. IL-4, IL-13, Eotaxin as well as IL-1 receptor antagonist and I’ll note in a moment that the IL-1 receptor antagonist there’s a mimic of it if you will in anakinra which is an agent that was mentioned by Dr Petersen. And so some of the, you know questions that we’ve had, is that if the short duration subjects have an increase in IL-1 receptor antagonist, perhaps that you don’t want to add more to that mix, but perhaps you would want to in the long duration subjects if there’s this decrease, even though it’s not significant, that perhaps amping up their cytokine response either through Ampligen or through direct administration of IL-1 receptor antagonist could potentially be of benefit. Similarly for short duration subjects, agents, antibodies to IL-17A or interferon gamma could potentially be used to dampen down the response to these cytokines. Perhaps it could potentially be helpful. There are many commercial antibodies, I think there is four or five of them now against IL-17A, monoclonal antibodies that are used in a wide range of autoimmune diseases.
So this gave us this opportunity to really think about the possibility, of course it needs replication, and we also want to understand what happens to individuals over time, so we’re waiting for our longitudinal immune analysis to see over the course of a year to year and a half what happens to cytokines over time. But the opportunity to perhaps use a blood test that might enable physicians who are primarily using clinical measures, maybe they can use orthostatic intolerance as a contributor to diagnosis but we really need as much as we can to help in this process, so we’re very eager to understand that, and to see whether this holds out over time.
Evidence of the immune response is suppressed in patients who have been ill over three years
Short duration subjects have a very limited regulation, there’s a network diagram, all of the cytokines that are included have been those that have passed the multiple comparisons tests. Only CD40 Ligand is in red here connecting to IL-12p40, so that’s the only down-regulatory mechanism that’s present early on in disease. Whereas later on in disease it’s a wild, you know dysregulation, with many red lines, which means that there are down-regulatory pathways there, and it’s more intensive regulation than one sees in the controls although there is also down-regulation here as well. So we believe that maybe understanding these regulatory networks may help us to understand why there’s this apparent immune exhaustion that occurs in the late onset of disease.
It doesn’t seem to only be, of course people who are sick for longer or you know get older, and so have to control for age, and so differences in age did not account for these differences alone, there were subsets in many different age groups with these various cytokines, distinguishing between short and long duration subsets.
Very strong link between interferon-gamma levels and duration of illness
What was most intriguing was that although the interferon gamma levels were small, there’s 105 fold increased risk of being a short duration subject, being early during your disease for every unit increase in interferon gamma, and highly significant. Almost never does one see and odds ratio of that magnitude, but also some increase with IL-12p40 proinflammatory cytokine as well, but some had a decrease association.
We found that in a study using some very presciently collected samples from subjects with ME, from Dr Petersen’s collection which is probably one of the largest and longest standing collections, we also looked at these cytokines using the same panel in spinal fluid of ME subjects and compared them to subjects with MS as well as no disease controls. And I just want to mention for a moment that the duration of illness there was a wide range, there were some that were earlier on but the ME subjects were on average 8 years out from their onset of illness.
Exceptionally low level of IL-6 in ME/CFS patients, may be linked to memory problems
And consistent with what we found with the long duration subjects in the blood we found decreases in many of the same cytokines, almost non-existent IL-6. IL-6 intriguingly is extraordinarily important in memory models, IL-6 knockout mice don’t have the capacity to lay down a memory, so their longer term memory, taking something from short term memory and commemorating it in longer term memory is impaired. So, we wonder whether in the central nervous system if this is really so diminished whether that may be associated with those symptom complexes, and we need to do further studies to understand that.
So similarly these were all reduced as compared to the no disease controls, so when you looked at the correlation between time since onset of illness and the cytokine levels within the ME subjects, you did see that the ones who were shorter duration had slightly higher levels and then it started to decrease over time but we didn’t have as many subjects who were early in the onset of their illness.
IL-1 receptor- antagonist really quite low here again, again could an agent, you know like Anakinra which is a IL-1 receptor antagonist recombinant, would that be of benefit if you could get it into the central nervous system, you know that’s another challenge if you’re taking it peripherally you don’t necessarily get it right into the central nervous system.
Evidence the immune system is dampened down in the brain too (spinal fluid sample)
Again dysregulation, differential regulation in Chronic Fatigue Syndrome in the central nervous system, with IL-1 receptor antagonist being a key regulator of down-regulation in the central nervous system, whereas there’s no particular down-regulator in the no disease controls.
And IL-1 beta being a decreased factor for risk of ME, whereas Eotaxin an allergy associated eosinophil drawing chemokine, when that is increased there’s a higher risk of ME as compared to the controls.
Looking at how the immune system and the microbiome could affect metabolism and the brain
I just want to conclude on some of our thoughts about how we can pull together the gut microbiome with the immune system and the metabolome and to try to understand the system more comprehensively, and perhaps allow us an opportunity to think about other means of addressing the immune disturbances, in addition to or you know as an alternative to immunomodulatory drugs.
Because we think that, you know probiotics may also be able to serve as an immunomodulatory agent of sorts. We know that in our diet we have a variety of amino acids in particular L-tryptophan is one that is particularly of interest because it is the precursor of serotonin, which of course is involved in a whole host of nerve vegetative functions, you know, your sleep, your eating, your sex drive as well as contributing to some of your vigilance and mood regulation, but also melatonin which is a very potent immuno-active agent, but of course also regulates circadian rhythm and sleep cycles.
But the tryptophan can be taken away from the serotonin synthesis and melatonin synthesis by being shunted down this kynurenine pathway and were now looking at these metabolites through metabolomic analysis and find disturbances in kynurenine in ME that look suggestive and interferon gamma and TNF-alpha induce the enzyme that takes it down this pathway.
In addition, glucocorticoids, product of your hypothalamus pituitary adrenal axis, reactive oxidative stress species reducing agents activate another enzyme called Tryptophan 2,3-dioxygenase and this goes down and these have affects both in brain and in the white blood cells, and there are a variety of agents that are considered to be regulators of IDO including T Gondii, Herpes viruses, Chlamydia and a variety of extracellular bacteria as well.
The pathway, again, it’s important both in memory function, immune function but it occurs in the brain as well as in the white blood cell, and it’s a necessary process and so if you don’t get to the end of this process, you won’t have good memory processes.
IL-6 is a participant here along with NAD+ which will alter the synaptic activity, modulate synaptic activity in hippocampus and cortex. But a variety [of these] affect glutamate receptors with some which can increase excitotoxicity, and some which can be protective. But again it’s taking it away from serotonin and melatonin synthesis and both viral infection and stressors can set this process off. In your white blood cell it’s the source of autoimmunity when you go down this process so you shift from TH1 to TH2 type of phenotype in your T cells and you are leading to an autoimmune vulnerable state. So we’re very keen to try to understand this and to put it together with the cytokines that are associated with these processes.
And just in our very preliminary analysis this isn’t dissecting according to ME but rather in the entire population we do see a relationship. If you can’t find serotonin in the blood, again so it’s suggesting that there’s a movement of tryptophan away from serotonin synthesis and activation of this pathway. So for individuals who have non-detectable serotonin in their blood they have higher levels of cytokines that are known to activate this pathway, IL-1 beta, TNF alpha, IL-12p40, IL-17F.
As well as chemokines that are associated with interferon gamma and are induced by interferon gamma so, CCL-2 , CCL-3, CCL-5, Eotaxin etc, and again so this hasn’t been said but these relationships hold in individuals who are healthy controls as well, we want to understand whether there are some subsets that we can tease out here.
And this work, you know, takes many, many many individuals and in particular the astute characterisation of the clinicians in our network, Dr Petersen as well as our other clinicians. Really, so important to any quality that we might hope to have in our research, begins with the appropriate characterisation, the appropriate controls and matching of subjects – as well as the ongoing persistence despite the difficulties in funding and keeping on with this very challenging disease – to try to make inroads. Thanks very much. END
This excellent research needs our help with donations. Please consider making a donation here and help them to help us!
Huge thanks to RME Sweden for putting their Severe ME conference talks on YouTube where you can view other speakers such as Dr Petersen and Dr Leonard Jason. Here is Dr Hornig’s presentation.