In the RA synovial tissue, cells talk to each other. Imagine a cell that looks like a dendritic cell with  long processes, and there are lymphocytes around it. Other cells may be macrophages or activated lymphocytes. There are fibroblasts, and there's a blood vessel as well. And, a lot of these cells are in direct cell-cell communication. The abundant cell types in the synovial tissue are T lymphocytes, macrophage-like, or type A, synoviocytes, and the fibroblastic, or type B, synoviocytes. Other key cell populations, not as numerous but still important, are dendritic cells, B cells, the plasma cells that differentiate from B cells, mast cells, and osteoclasts.

 

Some of these cells are T lymphocytes, monocyte macrophage synoviocyte, fibroblastic synoviocyte, dendritic cell, endothelial cell, the lumen of the endothelium, chondrocyte, and articular cartilage. These different cell types in RA communicate with each other in 2 general ways. One way is by the receptors and ligands that are built into the membranes of these cells. So when dendritic cells, for instance, present antigen to T cells, there are a series of costimulatory receptors and ligands that engage, and this is called cognate cell-cell interactions in immunologic terms.

 

When T cells and monocytes come through the circulation and could potentially  egress into an inflammatory lesion, they have to stick to the luminal side of the endothelium first, again through a series of receptor-ligand interactions by means of adhesion molecules. So cognate cell-cell interactions are one way that cells communicate and alter their behavior. The other way is through the secreted mediators, which in the synovium are primarily cytokines.

 

For instance, IL-1 and TNF are coming out of this monocyte macrophage cell. And, scientists understand that these secreted mediators are the key communication signals between the cells that orchestrate the inflammatory process. The fibroblast at the edge of the pannus,erodes into the articular cartilage. About 10 to 12 years ago, scientists were using monoclonal antibodies to try to destroy individual cell subsets like T cells, or to try to block their function by virtue of binding to specific targets on the cell surface. And cell-directed therapy, for a variety of reasons, has not worked well in RA, unless the anti-B-cell antibody rituximab proves to be as good as very scant preliminary data claim to show. Suffice it to say that anti-cell surface molecule therapy hasn't worked well or been practical to use, but anticytokine therapy, that is, cutting the communication links between the cells by tying up the secreted mediators, has worked surprisingly well.

 

To define a cytokine, it's interesting to compare and contrast cytokines, the hormones of the immune and inflammatory system, with the conventional endocrine hormones of the endocrine system. As we review these different characteristics, we'll make that comparison. Cytokines are intercellular messenger molecules. Unlike conventional hormones, every cytokine is made by heterogeneity of cell types. With insulin, for instance, only the beta cells in the islets of the pancreas make insulin. Many cells of the body make tumor necrosis factor (TNF). Macrophages make the most, but lots of other cells make it as well. No cytokine is restricted to a single cell in its synthesis.

 

Cytokines are highly inducible, and so are all of the hormones of the endocrine system. But cytokines are even more inducible, many, many orders of magnitude more inducible. If you get a bolus of bacteria injected into you systemically, the amount of TNF, interleukin-1 (IL-1), and other cytokines that will be made very quickly is absolutely prodigious; it may go up 1000-, 10,000-fold within a few hours. The effects of cytokines are primarily local unless so much is made that you start to get systemic effects. The systemic effects are in proportion to the abundance. With endocrine hormones, the effects are usually at a distance from where the hormone is produced, and they're targeted by virtue of where the receptors for that hormone are expressed.

 

Cytokine receptors are usually fairly ubiquitous, with some exceptions, for example, interleukin-2 (IL-2) receptors mostly on T cells. Almost all cells of the body have TNF receptors. The receptors for cytokines are all on the cell surface and they're all proteins. All cytokines are proteins. This is different from conventional endocrine hormones to the extent that some of those are proteins and some of their receptors are proteins, but all steroid hormones go right into the cell. Steroid hormones aren't proteins, and the receptors are not on the cell surface.

 

Once a cytokine binds to its receptor on a target cell, it sets off a signaling cascade and alteration of gene transcription in the target cell. And that defines the biologic effects of that cytokine. Cytokine synthesis and effect is controlled at many levels: at the level of induction of the message for the cytokine, and once the protein is made, in some cases by endogenous soluble receptors, and in some cases by endogenous soluble receptor antagonists. Our therapies mimic some of these endogenous strategies.

 

This simple concept is that in some disease, such as RA in the synovial tissue, the balance between the endogenous anti-inflammatory mediators and the proinflammatory mediators is out of whack. There is too much of the proinflammatory cytokine around the TNF alpha, IL-1 beta, and others. The anti-inflammatory substances that are produced endogenously in RA include soluble TNF receptors and the IL-1 receptor antagonist, and some cytokines such as IL-10. But, it's not enough to counterbalance the other side unless you start loading up the anti-inflammatory side with exogenous partners, such as monoclonal antibodies and TNF, or exogenous soluble TNF receptors. And that would just shift this in the direction you want.

 

There are far more players on each side of this balance, including a range of proinflammatory cytokines and the anti-inflammatory ones including those not in the joint. IL-4 isn't there and science might like to give it; IL-10 is there, maybe not enough of it; and some other newer ones. Some of the key cytokines that have been described and studied in the synovial tissue in RA: To put it in perspective, there are probably about 150 cytokines that are well known and characterized in immunology and inflammation research, and of those, at least 40, maybe 50, have been characterized as being present in RA synovial tissue or fluid.

 

Gamma interferon is a cytokine that's been around for a while. It's made by T cells. It acts on all sorts of other cells, and its major effect is to induce expression of antigen-presenting molecules, such as Class II MHC molecules. Antigen-presenting cells chop up proteins into little peptides and stick them into those Class II MHC molecules and stick that on their cell surface; and the T cells come along and eventually there's going to be a match between a T cell with the right receptor and the correct antigen MHC displayed on the surface of the antigen-presenting cell.

 

Gamma interferon is the major cytokine that induces the expression of Class II MHC molecules and the upgrade regulation of Class I on those antigen-presenting cells. So it is a key step in the initiation of immune responses. There is gamma interferon in the joint in RA. The amounts you can measure are small, yet the biologic effects of it are prominent. In other words, there's lots of MHC on all the other cells there. Presumably there has been enough over time to have an important effect.

 

IL-1 and TNF are extremely important, dominant cytokines in the inflammatory cascade and RA. IL-6 is an interesting one because it's hard to be sure whether you want to block it or not. On the one hand, it's a proinflammatory cytokine; it's made by a variety of cells including macrophages , fibroblasts, even by T cells to some extent, and it acts on lots of different cells. It's one of the cytokines that acts on the liver to cause the change in protein synthesis that gives us our acute-phase response pattern. It acts on the chondrocyte, and makes the chondrocyte secrete matrix metalloproteases, matrix-degrading enzymes, into its surroundings. So IL-6 is proinflammatory and tissue destructive in that sense.

 

It also, at least in some systems, has feedback effects to shut off synthesis of TNF. So it has both proinflammatory and anti-inflammatory effects, and I'm not sure whether it's going to be a useful target for blocking RA. Maybe together with TNF inhibition, it'll be a good target. There are other diseases, such as Still's disease in which IL-6 levels seem to be very high. IL-6 may be one of the key cytokines causing the systemic manifestations in Still's disease.

 

IL-8 is a chemokine. The term chemokine is a contraction of chemotactic cytokine. So chemokines are mediators that attract other cells to come into sites of inflammation. There are many chemokines, several dozen actually, and the terminology is quite complex. Their pattern of binding to different receptors is also quite complex, because each chemokine binds to several receptors and each receptor has many chemokines that bind to it. There's this kind of grid-like overlapping pattern. Several of these are in the joint in RA, and IL-8 is a representative of that group that attracts T cells as well as neutrophils into the joint. It's made by synovial fibroblast primarily.

 

The ones that are called ILs are interleukins. There are 25 of those and it keeps increasing. There's an international naming committee that will name them. But there are a lot of other cytokines that don't get called interleukins, such as TNF, interferons, colony-stimulating factors, fibroblast growth factors, and most of the chemokines. When you get all of these together, it adds up to 150 or so cytokines. IL-10 is made by macrophages, fibroblasts, and T-cells. In RA, IL-10 is primarily produced by macrophages. It's in the joint, it acts on a variety of cells, and it turns off various kinds of immune responses.

 

IL-12 is an important one, and may be one you haven't heard of or thought about much. But this could be a good target as well. It's made by antigen-presenting cells, such as macrophages and dendritic cells. It acts on T cells to induce gamma interferon synthesis. IL-15 is also important, and may be a good target. It's made by the synovial macrophages and fibroblasts. It acts on T cells, and it's a T-cell growth factor. There isn't much IL-2 in the joint. IL-2 is the first described and best known T-cell growth factor, and it's an autocrine one. So T cells make it and T cells use it. And when it became clear there wasn't much IL-2 in the joint, the potential role of T cells in RA came into question. But that was before IL-15 was discovered, and IL-15 is kind of the IL-2 surrogate in the joint in RA. It even uses almost the same receptor. It uses a 3-subunit receptor, and 2 of those 3 subunits are identical to 2 of the subunits in the IL-2 receptor. There's lots of it in the joint, and it could be a good target.

 

IL-16 is made by macrophages and CD8 T cells, and it both activates and acts as a chemotactic factor for CD4 T cells. IL-17 is very interesting and potentially a very exciting target. It's made by T cells and it activates the synovial fibroblast to do bad things, to make tissue-destructive enzymes and other cytokines. There's lots of it in the joint. It's the kind of cytokine you think maybe you can block without doing too much harm to your host defenses.

 

IL-18 is a cytokine that has IL-1-like effects. It is present in the joint in RA and it seems to be even more important in Still's disease where it's found at very high levels. It's a tricky one to evaluate because there is a binding protein that is often made together with IL-18. When IL-18 is complexed to its binding protein, you can measure it but it's not biologically active. So levels of IL-18 may be very high but it's free IL-18 that really counts. It is potentially another interesting target in some diseases.

 

The cytokines to some extent have competing or even antagonistic actions, and this is true in many aspects but also in terms of how T cells differentiate and how the T-cell role in immune responses plays itself out. In an attempt to classify  and understand this, these cytokine patterns have been called Th1 and Th2. This is something that immunologists talk about a lot. It started first with the realization in the mouse that helper T cells were different in which cytokines they made. There's not just a single pattern. There are different patterns.

 

When a T cell is first activated, its function is not yet specialized and this cell is called Th0. At this stage, the T cell makes gamma interferon, IL-4, and usually IL-2, which is an autocrine growth factor. Then it chooses a way of differentiating, and which way it goes depends upon what kind of antigen it saw, what kind of antigen-presenting cell showed the T cell that antigen, and what were the costimulatory factors and ligands that were around when the antigen was being recognized. The T cell will either become a Th1 or a Th2 T cell.

 

The Th1 T cell is defined as one that makes gamma interferon but not IL-4. The Th2 T cell makes IL-4 but not gamma interferon. They make other cytokines as well but IL-4 and gamma are the ones that define what class they belong to. Many of the diseases that we call autoimmune or immune-mediated are considered to be Th1 diseases, such as RA because there is no IL-4 in the joint. You can't find it. There is gamma interferon and the biologic effects of gamma interferon can be seen.

 

Cellular immune responses that lead to T cells able to cytotoxically damage targets arise out of Th1 pattern of differentiation. Allergic responses such as asthma and atopic allergies arise from immune reactions that go through this Th2 pathway. One of the other Th2 cytokines is IL-5, which promotes development and release of eosinophils from the bone marrow. Think of Th2 with an allergic immune response as the opposite of Th1.

 

Then there are regulatory T-cell subsets. There are various ways of defining these but Th3 produces TGF-beta, which is a cytokine that tends to lead to tissue fibrosis and tends to downregulate inflammation. The Tr1 subset makes IL-10, and IL-10 downregulates Th1 responses and sometimes Th2 responses. It's a little complicated because the cells that make these 3 cytokines (IL-4, IL-10, TGF-beta), can sometimes be found coming out of the same cell. So these are not absolutely clean distinctions but they are the functional definitions of these particular labels.

 

This was first defined in the mouse. It's not as neat and clean in man as it is in the mouse, but it's still a useful paradigm. Schematically you have a Th1 cell activated by an antigen-presenting cell. In this example, a dendritic cell presents antigen and makes IL-12. The IL-12 is really the key to get this cell going. With the action of IL-12 on the Th1 cell, the Th1 cell will make gamma interferon. It will continue to make IL-2 and it also makes a little TNF. Th2 cells on the other hand make IL-4 and IL-5. Some of them make IL-10. The IL-4 and the IL-10 can inhibit the function of the Th1 cells.  Scientist would  like to get a little more Th2 immunity in RA. It would create perhaps a balance that could modulate the disease.

 

Why blockade of TNF or IL-1 makes so much sense: TNF is at the apex of the proinflammatory cytokine cascade in RA. It's important to remember that it triggers production of many other cytokines, especially IL-1, so TNF engenders synthesis of IL-1. It facilitates activation of T cells. Actually IL-1 does that but TNF also helps, so these can be among the costimulatory signals for T cells. It causes the endothelium near the inflammatory foci in the joint to express adhesion molecules and, therefore, T cells can come into the synovial tissue, and neutrophils into the synovial fluid.

 

TNF and IL-1 stimulate fibroblasts and macrophages to release the proteases, the tissue-destructive enzymes. And there is very important recent information that development of osteoclasts, the bone chewing cells, is greatly enhanced by TNF and IL-1 together with the rank ligand (RANKL) molecule; also known as osteoclast differentiation factor (ODF). TNF, IL1, and RANKL cause osteoclasts to form right in the synovial pannus and to activate the osteoclasts down in the bone. The bone is outflanked and eroded.

 

Why neutralizing TNF could be so powerful: We'll have a monocyte that's making TNF, and the TNF can feed right back on that monocyte and cause the monocyte to make IL-1. Monocytes are the cells that differentiate into osteoclasts, and TNF and IL-1 together in the presence of this RANKL, here called ODF, drive a fraction of the monocytes to become osteoclasts.

 

Meanwhile,TNF has caused the endothelium nearby to become sticky, that is, to express adhesion molecules that will be ligands for receptors on the T cells and other cells that are floating by in the circulation. This becomes a sticky endothelium and some of the cells will roll along it, and eventually stick and be capable of migrating into the lesion. The reason they migrate into the lesion is because the TNF and IL-1 have caused the synovial fibroblast to release IL-8 and other chemotactic cytokines. Now the cells that are stuck will go in and begin to build up the mass of inflammatory cells.

 

At the same time, the synovial fibroblast under the influence of the TNF and IL-1 is making matrix metalloproteases, prostaglandin E2, and IL-6. These enzymes chew up the cartilage. The IL-6 activates chondrocytes to release even more matrix metalloproteases and it chews up the cartilage from the inside. So much of what's going on, both with inflammation and with tissue destruction of cartilage and bone, is orchestrated by TNF, so that taking this away really would make sense.

 

TNF is synthesized as a trimer. It's primarily made by macrophages, and also by activated T cells. It's at first on the cell surface and then it's released and diffuses around. It binds to TNF receptors on any of a variety of target cells, setting up a signaling cascade in that cell.  The TNF trimer, which is initially on the surface of the TNF-synthesizing cell, is released from the surface by a specific enzyme -- there's TNF-alpha converting enzyme or cleaving enzyme -- and the biologic activity of TNF is primarily after it's released. So this enzyme itself could be a target for pharmacologic inhibitors designed to affect TNF action and, in fact, some are being developed.

 

TNF receptors come in "2 flavors," the p75 and the p55. These are not subunits of the same receptors. They are distinct receptors yet most cells have both. It is not clear why both are needed. They are transmembrane proteins like all cytokine receptors. The big portion is the extracellular region where TNF is bound. There is the transmembrane region, and the cytoplasmic tails are very important because that's where the signaling occurs.

 

Signal transduction and adapter molecules attach themselves to these cytoplasmic tails and that's how the signals get brought into the nucleus. TNF itself is part of a multigene family and TNF receptors are also part of a multigene family, including other molecules such as Fas and the Fas ligand. Those are molecules involved in apoptosis or programmed cell death. These TNF family receptors have prototypic cysteine-rich amino acid domains. They signal through the cytoplasmic tails. Two flavors of TNF receptor and naturally occurring soluble TNF receptors occur. Some of those receptors get cleaved off of the cell surface, but they circulate as monomers rather than dimers and that probably limits their potency and affinity.

 

How the TNF receptor signals: It's a very interesting story because there are 2 arms. One arm has proteins, called death-domain proteins, leading to activation  of the caspase enzymes which leads to apoptosis or programmed cell death. This has to do with the actual name of TNF, tumor necrosis factor; it's a cytokine that can actually kill cells.  How does it kill cells? By activating apoptosis. Fortunately, it's not our only host defense against tumors. In fact, only a small fraction of tumors are killed this way. The reason that it doesn't generally lead to apoptosis is because there's a second and dominant signaling pathway that goes through a series of kinases, here called MEKK, JNKK, and so forth, leading to activation of a transcription factor called nuclear-factor kappa B (NF-B). And this is a key transcription factor for activating genes involved in inflammation. Almost all of the genes involved in the inflammatory response have binding sites in their promoters for NF-B. That's why there's so many genes turned on when TNF binds to its receptor. When both signaling pathways are hooked up in a cell, the activating one always dominates over the apoptotic one.

 

There are actually 2 forms of IL-1. IL-1 beta is the main form in the joint. They have similar biologic activities. And there is this natural IL-1 receptor antagonist, which looks very much like IL-1. There is no TNF receptor antagonist so this system works a little differently. What happens when IL-1 binds to the type I IL-1 receptor? It causes a conformational change in the IL-1 receptor, which allows this accessory protein to join the complex. And you have a 3-protein complex -- the IL-1, its receptor, and this other molecule -- and these two cytoplasmic tails together lead to signaling of the IL-1 signal. There's also a type II receptor, which is just a decoy receptor and is not functionally active. The accessory protein is key.

 

Like TNF, IL-1 is a key proinflammatory cytokine in RA. It triggers production of other cytokines such as TNF. It also induces endothelial adhesion molecules, stimulates collagenase and stromelysin, which is another enzyme that degrades connective tissue, and stimulates osteoclast differentiation, all very similar activities in TNF. But the difference is that there is this receptor antagonist lurking that is in balance with IL-1 in terms of the actions of IL-1.

 

The receptor antagonist has been adapted to a new medication, anakinra or Kineret. This is recombinant anti-inflammatory protein that is almost identical to normal human IL-1 receptor antagonist; a single amino acid, methionine, has been added at the N-terminus which stabilizes it. Its biologic activity is identical to endogenous IL-1Ra. Its problems include the fact that its half-life is fairly short and you need to have a lot of it around to effectively compete against IL-1, because binding of IL-1 to only small numbers of IL-1 receptors on target cells is enough to activate the cell.

IL-1ra Does Not Signal: First of all, when the receptor antagonist binds, the IL-1 receptor folds around it, but in a way that this accessory protein cannot join into this complex. Therefore, you can't get a signaling complex and this receptor does not function. That looks like it should work well, but the problem is you essentially have to do that to almost all the receptors on the cell to really turn off the effect of IL-1.

 

IL-1 Inhibition by Receptor Antagonist-different scenerios: 1st, you have the case where there's IL-1 around; it's binding to some of the receptors and activating them;, and those cells signal. Think of it as an all or none thing. Either the cell responds to IL-1 or it doesn't.  Secondly,you have IL-1 receptor antagonist around. Most of the receptors are bound by the antagonist, but a few of them are bound by IL-1. Even if it's 80% occupied by the antagonist, this is still enough to signal and the cell responds to IL-1, ignoring the IL-1 receptor antagonist. You basically have to get almost all of them occupied by the antagonist to effectively compete against IL-1. It's about a 100 to 1 level of antagonist to IL-1 that's needed for this to be effective. That's why, in addition to its short life pharmacologically, it has to be injected every day when patients take it, and it has to be given in such high doses. It's because of what's going on at the receptor level.

 

How about future targets? There are many cytokines to choose from. Some we'd like to block, such as IL-12 and 15, and others we might like to use, such as IL-4 and IL-10.  If you think about ideal therapy, you want something that is effective for symptoms. As another surrogate marker, you also want a drug that effects radiographs; x-ray progression and clinical progression being 2, perhaps, separate ways that RA manifests itself. I'm bringing that up because it's an issue. How tightly correlated clinical disease and radiographic disease are continues to be a debate, and we've seen disconnect there. In fact, there are maybe disconnects between clinical response, radiographic response, HAQ, and functional response, let alone laboratory profiling. Those 4 areas may not all move at quite the same pace.

 

That may provide opportunities for different approaches treatment-wise -- you want a drug that relieves symptoms and inhibits disease progression, not just clinically but radiographically, with some measure of functional improvement also. The drug should be efficacious and that efficacy should be sustainable.

 

One of the criticisms of our previous drugs is that people didn't take them for very long because they didn't think they worked. They should be safe and well tolerated -- another criticism of some of our drugs. They should be relatively cheap with relatively simple monitoring. This raises the question of whether we believe we're going to get away with a single drug, a magic bullet, for RA, or are we going to have to use combinations of drugs?