Shiv Pillai (Harvard) 3: IgG4-Related Disease: Collaboration Between B and T Cells

Shiv Pillai (Harvard) 3: IgG4-Related Disease: Collaboration Between B and T Cells


So, in this lecture, I’m going to take a disease that we know very little about — in fact, it’s a disease that was described less than two decades ago — and see how we can use immunology to try to understand this disease better. So I’m going to be talking mainly about IgG4-related disease, something you’ve probably not heard of before. But it’s a prototype for many other diseases. So, if you think about how we’ve understood disease over the centuries, we understood a long time ago that some diseases were Mendelian. And this… kings had children who had diseases, and people knew that there were diseases that were inherited. But it was only in the 20th century that we actually understood human genetics, and we could nail this, and that showed that there were diseases caused by single genes. Then, in the middle of the nineteenth century, as I mentioned in the first lecture, we started to understand that bugs can cause disease. So, we have infectious diseases. In the middle of the twentieth century, there’s a pretty good understanding that has emerged about cancer, and in tumors we have in a single cell the acquisition of multiple mutations that cause the cell to proliferate uncontrollably, and that is a general understanding of cancer. Finally, there’s everything else. There’s so many other diseases. All of us, everybody in the universe, by the time they die, is going to suffer from some complex disease. Maybe it’s an allergy, maybe it’s an autoimmune disease, maybe it’s atherosclerosis. There are lots of diseases. Maybe it’s autism, schizophrenia. These are what we call common diseases, or complex diseases, or multi-gene disorders. And we wave our hands a bit, because we don’t understand these very well. And beyond this there are a couple of tumors — so CLL, chronic lymphocytic leukemia, and medulloblastomas — where some patients — not all, a very small number — actually don’t have driver mutations. And so these diseases, these cancers, might be driven epigenetically. And they may fall into the same category as those complex diseases which we don’t understand well. So, what do we think about these complex diseases? Because the immune system is accessible, one can try to understand certain complex disorders which have an immune component. And the general thinking about autoimmune diseases, for instance, is that maybe there’s some genetic susceptibilities, so there’s some susceptibility. And we know from GWAS studies and so on that, yes, indeed it’s true, but that doesn’t explain everything. Maybe there’s a role for the environment. Maybe microbes, metabolites of microbes, play some role. Again, we don’t have a full mechanistic understanding, but this is probably true. These probably combine to allow lymphoid clones. We talked about clonal selection in the first lecture. You can imagine if there was a lymphoid clone which went through epigenetic changes, expanded, and became a rogue clone, and then drove a disease because either it broke tolerance or there was enhanced inflammation, maybe when you put all this together this would explain a complex disease. And our thinking today in the field is if we can understand a complex disease then we will actually have treatments that are curative. Today we can treat diseases, but we treat the symptoms rather than the causes. So, our hope was, can we find clones of expanded cells in an untreated human disease? And maybe use those clones to start to understand what the epigenetic alterations might be and how the disease comes about. And that’s what I’m going to try to explain in the context of IgG4-related disease. So, the big picture… big picture question for us is, what is the underlying basis of common diseases? And then, can we use human diseases to understand human immunology? Okay? Because we study the mouse in great detail. It’s easy to manipulate the mouse. But in humans maybe we can learn more by studying disease. So, this disease — IgG4-related disease — is hard for you to actually put it all together, but it is a multitude of diseases that were described some couple of hundred years ago that have now come together under a single umbrella and are called IgG4-related disease. And all IgG4-related disease is is a chronic inflammatory disease which fulfills certain criteria. And this is true for all the varieties of this disease. So, it’s a multi-organ disease, and the diagnosis requires histology. So, you have to look at a section. Patients usually present with a mass. And they’re misdiagnosed, typically, as having cancer. But when you take a biopsy, you see that this is not a tumor, but it’s inflammation, okay? And so the characteristic features are lots of lymphocytes and plasma cells infiltrating the tissue; there’s fibrosis — fibrosis just means scarring… the tissue is scarred because of so much inflammation; and there’s a phenomenon called obliterative phlebitis, which is that the blood vessels are occluded because of inflammation. Okay? Now, this is the final kicker, that actually gives you the diagnosis, is that when you count the plasma cells… so, the pathologist looks at it and sees that most of the plasma cells, about 40% or more, make IgG4, which is one subtype of IgG… we have IgG1, 2, 3, and 4, okay? So, to look at the tissue… disease… if you looked at the tissue, you’d see… you see storiform fibrosis — that’s the pink stuff swirling around. You can see obliterative phlebitis, so this is the blood vessel that’s occluded with inflammation. It’s narrowed it down. And this is the fact that you have lots of IgG4. You’re looking at plasma cells, and most of them are straining for IgG4. So, you have lots of IgG4 plasma cells, lots of lymphocytes. Most of them are T cells. Okay? Now, IgG4 is a funky antibody. It actually can split into half, and it’s… this is called Fab-arm exchange. So, sometimes it’s not even bivalent. Why we have it we don’t fully understand. Okay? As an antibody. It does not fix complement, so it can’t actually cause destruction of microbes through complement fixation. It does not bind the Fc receptors, that Fc-gamma receptor of phagocytes. So, phagocytes cannot ingest a pathogen if it’s coated with IgG4. So, it’s not gonna cause inflammation. Pharma loves it because they can convert a therapeutic antibody into an IgG4. It’s not gonna cause inflammation. If they can mutate it to make sure it doesn’t fall apart. But why do we have it? So, what we think is that this is actually an antigen sink. Nature created it to suck off the antigen which might cause disease, but suck it off and not cause inflammation. Now, this is just a thought. We don’t think of it as an inflammatory antibody. Then why do we have a disease with so much IgG4? Okay? So, we assumed that IgG4-related disease, because of the fibrosis, would be driven by T cells and by helper T cells, which have CD4 on them, and we decided to search, in an unbiased manner, for T cells that would drive fibrosis. And we decided… this is a human disease; we need some criteria. So, we have certain criteria. And the criteria to look for these cells was, are these cells clonally expanded? Have they… in the disease, do they grow in number, and does… does a single clone become large? Do the cells infiltrate tissues? If they don’t infiltrate the tissues, they can’t be causal, in terms of the disease. Do they make products in the tissue which could drive fibrosis? And finally, when you treat them, do they decline? And we decided that if a cell met these criteria we could assume… you know, this is not an animal, model we can’t retest it in a different way… we could assume that these cells might be causal. Okay? So, on this flow plot all I want you to focus on is, if you look where my finger is right now, where I’m pointing to, that population that you see there refers to activated or effector T cells. These are CD4 T cells. If you look below in the control, there is no such population. Okay, so it’s only in the disease that we see these activated T cell. That’s… it’s… if you have a disease, you’re going to see activated T cells. If you don’t have disease, the T cells are going to remain unactivated. So, it makes sense. When we took these activated T cells from different individuals with the disease — completely different phenotypes — and then we looked for their gene expression, we’re now looking to see, are these cells expanded? Are there more of them? Are they clonal? What do they make? Okay? So, when we looked by gene expression, and we… the four different samples over here from four different patients, four different phenotypes, different organs, were very similar. And the expressed genes were somewhat unusual. These were CD4 T cells, but they expressed certain genes that are normally found in killer cells. So, these are helper cells making gene products found in killer cells. These have been described before, not in the context of disease. And these are now called CD4 CTLs. So, CD4 cells with a cytotoxic phenotype, okay? So, we had discovered that in this disease, the T cells that we found as activated effector T cells were CD4 T cells with this CD4 effector phenotype, which was CD4 CTL. Okay? If you looked in different patients, the numbers of these cells, of these CD4 CTLs, was increased in patients with disease. If you had active disease, you had lots of these cells. Okay? These cells made certain cytokines. They had some unusual cytokines. They also made IL-1, interferon gamma, and TGF-beta. This combination had not been seen before in other T cells, but this combination of cytokines was being made by these T cells, which can also make killing molecules, like CD8 T cells. Okay? This is just to show you that, at the protein level, they are making IL-4… IL-1-beta. So, IL-1 is a protein that needs to be processed in a structure called an inflammasome. But these cells are making a lot of it and processing it to the right size. So, IL-1 is normally made by macrophages and myeloid cells in general, but clearly these T cells make a lot of IL-1. Okay? So, we have these cells, we have these expanded CD4 T cells over here. And we want to look and see whether this population comes from 1 cell or from 10,000 cells or from 100 cells. We really want to know… I mean, is this clonally expanded, or is it just 10,000 different cells? So, we take maybe a few 10s… a few 10s of thousands of cells, we sort them out, and then we sequence the T cell receptor beta chain gene. So, T cell receptors have beta and alpha chains. They’ve gone through VDJ recombination. The variable domains are very different. So we sequence just the T cell receptor beta chain gene, and we sequence it a few million times, starting from let’s say 20,000 cells. So, we know we’re not gonna miss any cells, here. We’re gonna get every cell, and we’re gonna get it multiple times, and we’re going to be able to say, do I have 20,000 cells in this mixture or do I have 1 cell or do I have 10 cells? Okay? So, when you do this, what we find is that if you look… now, you look on the left there, it says IgG4-RD… we have these big blobs. So, we actually have big populations of clones of CD4 CTLs in the patients with the disease. If you look at healthy controls, you see little dots, multiple little dots. There’s no large clone. Nothing is expanded. But in the disease, we see big clones of these CD4 CTLs. We assumed that maybe we aren’t testing this aggressively enough. So we chose four patients in whom we knew the patients had this disease, but they also had a history of allergic disease, diseases like asthma. Now, you know allergic diseases are found in 25% of the population, so it was easy to find a bunch of patients who had both IgG4-related disease and who had allergic diseases like asthma. When we took their cells, now we were looking for two types of cells. Asthma is driven by cells called TH2 cells. Okay? So, TH2 cells have GATA-3 in them. So, from the same patient we could pull out CD4 CTLs and TH2 cells. And you’ll notice the TH2 cells aren’t clonal at all. This is the history of their allergic diseases. But the CD4 CTLs have big clones. Okay? So, we have big clonal expansions of CD4 CTLs in this active disease, and then we see the history of allergy in multiple small clones. So clearly, here we have a disease in which a certain population of effector cells — CD4 CTLs — are expanded, and these cells are clonally expanded, so that some antigen is probably driving their expansion. Okay? The next question was, do these CD4 CTLs actually infiltrate into the tissues? Are they just found in the blood? Or do they now go into the tissues? If they go into the tissues, it may be meaningful. So, we look in the tissues, and now we can do multicolor staining of tissues. And what you can see here… so, P11 and P25 refer to two patients. The control on the right doesn’t have disease. And what we can see… the stained cells that you’re seeing over there are stained CD4 CTLs in the tissue, okay? The quantitation below… the bars that are in red are referring to the CD4 CTLs. So, we quantitated in each patient… so, we looked at a bunch of patients… how many CD4… what CD4 cells do they have? Which ones are most abundant? The most abundant cell is the CD4 CTL. Okay? So, we have CD4 CTLs that are clonally expanded in the blood. We have them infiltrating the tissue. We can see them in the tissues. They’re… they’re the largest, most dominant cell in the tissue. But are they reactivated in the tissue? Do they make cytokines? Now, we’ve looked at all of the cytokines… IL-1-beta, TGF-beta, interferon gamma… I’m only going to show you the data from TGR-beta. And what you see here is… on the left-hand side we can see IgG4-related disease. And you can see that the patients have got lots of CD4 CTLs making TGF-beta. That means they’re reactivated; they’re making TGF-beta. On the right-hand side we have Sjogren’s syndrome, a different disease with lots of lymphocytes in there, lots of lymphocytes in the tissues, but they don’t have CD4 CTLs, they don’t have TGF-beta made by the CD4 CTLs either. So, what we had established by then is that CD4 CTLs are expanded in this disease. In fact, they’re expanded in other fibrotic diseases too, which I’m not going to go into today. And they’re clonally expanded, so they’re big clones of them that are there. To be sure that the clones actually go into the tissue, we’ve actually shown that these are also clonal in the tissue. And in the tissue, they are reactivated, because they’re making cytokines. So, something, some self-antigen or some other antigen is triggering them, and they’re making the cytokines, including IL-1-beta, which is known to drive fibrosis. So, in animal models, you can put IL-1-beta transgen… transgenically into a pancreas, and you’ll get pancreatic fibrosis. Okay? So, we said, okay, maybe this is the cause of disease. Why do they get IgG4? They have so much IgG4. What is causing the IgG4 class switch? Now, I have to walk you through a little bit about T-B collaboration to explain this. So normally, if you have a protein antigen, T cells can be activated by dendritic cells. Okay? So we see there on… on one end of the slide, you see the activation of T cells. The activated T cells then migrate towards the B cell zone, so they change their expression of the chemokine receptors, and they move towards the B cell zone. The selected B cell that sees the same protein antigen — but is now seeing some bump on that protein, and is recognizing it through the B cell receptor — it also gets activated, slightly, and it migrates towards the T cell zone. So, it comes towards the T cell zone, and they meet. So, the first interaction, labeled 1, is where you have T cells activating B cells. The activated B cells then expand slightly — they form a little focus — and then they decide to return the favor and then they activate previously activated T cells to differentiate, to express a lot of a chemokine receptor called CXCR5, which allows them to migrate into the B cell follicle, into the B cell zone shown over here. And those T cells are called T follicular helper cells. And those T follicular helper cells orchestrate the formation of a Darwinian structure called a germinal center, where we have genetic diversification, natural selection, and survival of the fittest B cells. So, what happens in a germinal center is illustrated here. So, we have… in the dark zone, we have B cells that are proliferating at a very rapid rate — every 3-4 hours — and as they proliferate… if you remember the six fingers I talked about in the first lecture… those six fingers, the genes encoding those fingers, those… those CDRs, are being mutated. So, the B cell that’s proliferating in the dark cell zone, in the dark zone, as it proliferates it’s mutating the fingers of the antibody genes so that each progeny B cell will have a mutated receptor. Okay? In the light zone, as the B cells emerge into the light zone, the T follicular helper cells help select the highest-affinity B cell, which can capture the original antigen most efficiently. And then they… they bless this highest-affinity B cell and allow it to differentiate and become immortal, as a long-lived memory cell or a plasma cell. So, T follicular helper cells actually do two things. One is they select the highest-affinity B cells. And they also mediate a process called isotype switching, where they allow the B cell to switch from IgM to either IgG or IgE, maybe IgG4 or IgA, and so on. So, the T follicular helper cell helps the B cell both to select the highest-affinity B cell coming out of somatic hypermutation and also to allow class switching to the isotype that is most useful in the context of a given infection. Okay? So, what we found is a type of T follicular helper cell in IgG4-related disease which expresses CXCR5, which is the marker for T follicular helper cells. It also expresses CD4. So, you’ll see CXCR5 in purple. You’ll see IL-4 in green. And you’ll see CD4 in blue… sorry, CD4 is in red. Okay? So, you have these cells expressing CD4, IL-4, and CXCR5. And so we have T cells, which have CXCR5 — they’re T follicular helpers — but which specifically make IL-4. And they are highly expanded in IgG4-related disease… seen on the top left. And they’re not expanded in other lymph nodes or in the tonsils, in normal tonsils. Okay? So, in this disease we have a huge expansion. It goes through about 50%… if you look at the bottom panel over there, the bars in red reflect to the fraction of the T follicular helper cells in the diseased tissue which are making IL-4. So, the vast majority of these cells, in many patients, are making IL-4. So, this is a subset of T follicular helper cells that makes IL-4, which we believe is driving class switching to IgG4. Okay? So, we can do a thing called cytokine capture. So, we can capture these T follicular helper cells based on the cytokines they make. And you can see from this RNAseq that the T follicular helper cell that makes IL-4 is very different from the T follicular helper cell that does not. It’s also very different from the non-T follicular helper cell which secretes IL-4 — that’s called a TH2 cell. Okay? And this is from the tonsils, just to show you these are different populations, that there are subsets of T follicular helper cells, and so in the humans, for the first time, we’re showing it as a subset that can separated out in disease, which makes a lot of IL-4, and which is likely to be driving the class switch to IgG4. When we look at a bunch of patients and we quantitate these cells, we can see a correlation of these cells — this is the subset making IL-4 — of T follicular helper cells, which correlates with IgG4 levels in the serum, negatively correlates with IgM levels in the serum, but does not correlate with anything else. So, it is extremely likely… IgG4 is not made by mice… it’s only made by humans, so it’s extremely likely that we have learned from patients that T follicular helper cells that make IL-4 — which have a novel transcription factor which we’ve described, which is not the one that’s found in TH2 cells — this unique population found expanded in disease is likely to be the driver of the IgG4 class switch. Okay? The broader question is, does B cell depletion actually reduce CD4 CTLs? Now, in this disease, we found a lot of activated B cells. We found activated B cells in the tissues. We found activated B cells in the tissues that kiss CD4 CTLs, and we’ve quantitated them. Okay, so we actually think CD4 CTLs — the cells I first described as being the drivers of this disease, the likely drivers — are actually being nurtured by activated B cells. Okay? I don’t have time to go into all of this, but I’ll touch on some features. So, here is the description of activated B cells, one type of activated B cells called plasmablasts. So, plasmablasts, you’ll see, is this population that you can see in patients below — not found in the controls above — highly expanded in patients in their blood. Okay? Also found in tissues. When you treat patients with Rituxan, which is an antibody to CD20, which depletes B cells — and this has been seen in more than one disease, now — a disease caused by T cells is improved by depleting B cells. Okay? So, in this case, as in multiple sclerosis, another disease in which we think T cells cause the disease, but depleting B cells improves the disease, dramatically. So, what do you… what do we see here? We see, clinically, the patients get much better with Rituxan. We’re depleting B cells; patients are getting better. The IgG4 levels change a bit, but not great… not… not a big difference. B cell depletion is perfect. You can see the B cells go way down, over here. Okay? Look at CD4 CTLs, we can follow clones. The clones start to decline. So we know that in this disease treating the patient, depleting B cells, presumably depleting activated B cells in the tissue as well, causes a reduction in CD4 CTLs, improvement of symptoms. So, some of the take-home messages from this lecture. So, this is… we have shown that CD4 CTLs can cause inflammatory disease, but CD4 CTLs have been shown before to be important in viral infections. And Susan Swain, in her review, has put this down as being a new subset of T cells. So, when you think of CD4 T cell subsets, we’d usually call them TH1, TH2, TH17, Tregs, and so on, but we should include T follicular helper cells, which is now pretty accepted nomenclature, as well as CD4 CTLs, as a separate subset. The other take-home message is that when we think of the pathogenesis of fibrotic diseases, there are some diseases like asthma which are driven by TH2 cells. That’s fine. But in diseases like IgG4-related disease and scleroderma, or systemic sclerosis, which I haven’t had time to tell you about, we know that CD4 CTLs are the most abundant T cell in the tissues, that they are in contact with activated B cells, so our model for the disease is that antigen is being presented by the activated B cell, which because it’s somatically mutated can capture antigen very efficiently, perhaps; is nurtured by the activated B cell; the CD4 CTL makes cytokines; it can make killing molecules; and it can drive this process, which we call inflammatory fibrosis. Okay? The other broad message is, at the same time as which… whatever is causing the disease is driving the formation of CD4 CTLs, it’s also driving T follicular helper cells that make IL-4, and this is driving the switch towards IgG4. Okay? So, I’ve talked to you today about IgG4-related disease, trying to give you a sense of how you can identify a cell that might be causal in a disease, and how we can learn something about human immunology from studying patients. And this… the work that I’ve talked about was done by many people. I want to highlight a few. So, Hamid Mattoo and Vinay Mahajan started the project on IgG4-related disease in the lab. Takashi Maehara did a lot of the tissue studies. A lot of the more recent studies on TFH have been done by Corey Perugino. Hugues Allard-Chamard has been looking mainly at the B cells in IgG4-related disease. And all of this work was done with collaborators. Our key clinical collaborator has been John Stone at MGH, who is the clinician who sees IgG4-related disease, and… probably, in America, this is the biggest collection of patients with this disease that we have. Seiji Nakamura in Japan has given us many tissues to look at. And I have a number of other collaborators that I haven’t had time to talk about their work or to tell you about today.

3 Comments

Leave a Reply

Your email address will not be published. Required fields are marked *