Synovial fluid (synovia) contains two types of components: first a passively dialysed serum through the synovial membrane and second cells and substances which are actively secreted by the synovial tissue.

Therefore, any siginificant changes in the composition of the blood serum leads to changes of the joint fluid (like hyperuricaemia).

There are two types of known cells, which are normal: type A-cells (phagocytes) and type B-cells (fibroblast-like cells).

The mechanical properties (lubrication, viscoelasticity) are mainly derived from the actively secretet substances: hyaluronic acid, lubricin, proteinases, collagenases, prostaglandins.

To prepare artificial synovial fluid presupposes the definition of objectives: is it for mechanical reasons (e.g. to test wear in a simulator) or shall cells have an optimal environment as close as possible to a physiological situation...

I hope, this short answer is of any help, good luck!

Since you've put RA as one of your keywords, I thought I'd mention a recently discovered fact - the concentration of H2S (hydrogen sulphide) in the synovial fluid of patients with RA is 4x higher than the norm. (I have the paper where that is described).

Thanks a lot Jorn Kircher and Derek Townshend. i have few more questions in this aspects.....pls clarify. Is proteinases/collagenases have any role in giving the viscosity to the fluid. if this proteinases or collagenases or prostaglandins are there in SF, then there will be a cartilage degradation, what will happen to that degradation products (peptides or aggregans), is that will be dialysed out again?....i read nutritional supply to chondrocytes are from SF. if so what kind of substane is acting as nutritional subtsance here...?

Derek Townshend ....it will be very helpful, if you share the paper that you mentioned.

I have been looking at the available data on the role of H2S in various chronic inflammatory diseases. Generally, it appears to me that H2S is very much like its counterpart, NO (Nitric Oxide), in that: A] they are both gaseous neurotransmitters involved with inflammation, and that B] they both appear to have a narrow range of being beneficial, above-which they become problematic and pro-inflamatory.

The paper I mentioned is the 2011 one titled; "Hydrogen sulfide and inflammation: the good, the bad, the ugly and the promising" by Matthew WHITEMAN and Paul G WINYARD. (See pages 17 and 18) https://www.researchgate.net/publication/51828701_Hydrogen_sulfide_and_inflammation_the_good_the_bad_the_ugly_and_the_promising

Whether H2S in the synovial fluid is doing good, trying to do good and building up, or is a baddie that should be removed, still needs to be determined. Whatever the answer, WHITEMAN suggests that H2S "could represent a novel index of disease activity in the inflamed joint". I have yet to see if anyone has followed up on that avenue.

Article Hydrogen sulfide and inflammation: The good, the bad, the ug...

There is a lot of talk about supplements for Arthritis. These usually include Glucosamine sulphate, Chondroitin sulphate and Methylsulphonylmethan (MSM). I am fairly confident that some or all 3 of these increase the availability of H2S and may ultimately aggravate circumstances. Glucosamine and Chondroitin are definitely beneficial could theoretically be joined to a different anion, not a sulphate, for the same benefit. (This is not really my field but I think it is a question worth raising).

Dear Derek

Thanks for the link and for your efforts. There was lot of critics on usage of CS and GC supplements. Their MOA is still unclear. As you said, its really important to check the contribution of these supplements in H2S formation.

Dear Venkatachalam,

I wrote a book on synovial structure and function with Brain Henderson in 1987 (the only text of its type as far as i know). Since then I have written chapters in several of the major textbooks. But it is an extensive field. I think we need to know why you are interested to provide any help online. Collagenases and other proteases probably do nothing in synovial fluid unless cartilage is dead. Live cartilage appears to be almost totally resistant to digestion because of inhibitors and physical barriers. The viscosity is determined largely by water, hyaluronan and lubricin and SF is non-Newtonian in behaviour. As far as I can tell supplements such as chondroitin and glucosamine are of no value and it is hard to see the scientific basis for any useful action.

What is your objective? I think there may be a difficulty these days in that the dozen or so people who were interested in synovial fluid function in the 1980-2010 period are now mostly retired and there may not be recent reviews of quality. Unfortunately a lot of old ideas get perpetuated.

Dear Jonathan Edwards

Thank you very much for sharing your experience in synovial structures and function.

Ya as you mentioned, i didnt get any clear info about synovial fluid compositions and synovial functions in online.

I would like to study about the role of synovial fluid protein/other substances interaction with drug (in diseased condition like rheumatoid arthritis)....is there any barrier in intra articular NSAIDs efficasy?

Even i would like to know about the cartilage turnover.

please clarify this point "Collagenases and other proteases probably do nothing in synovial fluid unless cartilage is dead. Live cartilage appears to be almost totally resistant to digestion because of inhibitors and physical barriers"

How the collagenases are acting in cartilage degradation in rheumatoid arthritic conditions? What "unless cartilage is dead" mean...is tat chondrocyte death?

Dear Venkatachalam,

The composition and dynamics of synovial fluid are in some ways quite unlike any other human compartment and it is easy to draw false conclusions. In the 1980s and 1990s J Rodney Levick, Peter Simkin and I came to a consensus about roughly how it works, but only after we had worked through initial misconceptions. It is hard to explain the subject without actually sitting down with some examples in a lab and reading a large body of historical literature. On the other hand, in disease, 'synovial fluid' is pretty much a standard exudate mixed with some hyaluronan and most of the subtle dynamics of the normal fluid are irrelevant.

Studying synovial tissue and fluid twenty five years ago led us to understand why synovial joints are involved in diseases like RA, but to be honest, now that we understand that, the synovial fluid physiology is no longer of much relevance to drug study. The relevant drugs now are those that work on the immune process that eventually causes joint inflammation. If the joint is affected then you are already too late. The objective is to block the immune signals that come before anything happens in the joint. Nobody is really interested in NSAI these days. They have no lasting effect.

For a long time it has been popular dogma that collagenases and other proteases degrade cartilage in inflammation in joints. However, if you think about it this does not make sense. Many patients with RA have inflammation in joints for months without any loss of cartilage. Even if the collagenase in the fluid was 98% inhibited the cartilage should be all gone after this time. Uninhibited collagenase will destroy cartilage in an hour or so. So the collagenase must be effectively 100% inhibited. Yet during some episodes of inflammation cartilage can be destroyed in a few days. The only solution to this problem I could find after thinking about it for decades was that type II collagen in live cartilage is totally protected, both from collagenase attack, and in fact from mechanical wear, by a constant extrusion of hyaluronan and small proteoglycans from its surface - much in the way that the skin of fish and dolphins constantly extrudes mucins to lubricate and provide a barrier against attack. However, if the chondrocytes die there will be no extrusion of macromolecules so the surface collagen will be directly available to attack. When I first became interested in joint pathology I looked at hundreds of slides of RA joints in the Path library at St Bart's Hospital in London and one thing that struck me was that whenever there was cartilage invasion or resorption the underlying chondrocytes were dead (absent). So that made sense.

Chondrocyte death is very easily explained in RA because it occurs in septic arthritis and would be expected in any situation (such as both of these) when glucose levels fall to near zero because of metabolism by cells derived from inflamed synovium.

Very informative discussion. The qaulity of SF is really important, but what about Quantity? As a lubricant there should be suffient quantity. Smaller quantity will produce pain in joints. And yet it seems that measuring/ assesing quanitity in a joint, is difficult. Can you light on this point?

Dear John Kircher and Jonathan Edwards

The discussion was very informative and useful for me. Really many thanks for lightning lot of info about joint physiology and joint pathology.

Ya i accept that NSAIDs are in sideline now. But the DMARDs and Biologics are having serious side effects and risks equally to the efficasy.

Whats conditions/Phase of RA patient making the Doctors to worry lot?

i have still some doubt that, if we want to protect the cartilage degradation in RA or to reduce the joint inflammation, the biologics/DMARDs available are should be administered directly in to the synovial fluid rit? why systemic, is that route of administation will give cartilage protection, why not Intra articular?

Because , there is no way of any cell filtration(bcoz its passively diffused serum) into synovial compartment, the only way to get inflammatory signals and the substances of degradation is from synovial tissue and chondrocytes of cartilage.

Sorry if i would have made any mistake in my questions.

Dear Venkat,

All stages of RA should worry doctors a lot. An adequate treatment nowadays is one that maintains the patient with no joint inflammation. If there is no inflammation in the joints there is no point in putting drugs into the joint because there is nothing going on there. The focus of therapy has moved to treating the cause - which is the immune disturbance in bone marrow and lymph node/spleen. If successful, no cells or products from this disturbance causes trouble in joints. Over 100 joints can be affected so local administration of drugs into joints never did make much sense except as a way of dealing with the very worst problems from time to time. That should all be history now, with the exception of the immediate initial clinical presentation, which can probably be dealt with using corticosteroid.

I missed Ulhas's question. We do not have ways of proving exactly how lubrication works but I think the likely explanation would mean that there can never be no fluid, unless the cartilage is dead. The reason is that chondrocytes are constantly producing hyaluronan and proteoglycan at a low level. These products are lost from the cartilage surface by a constant 'ooze'. The molecules are so enormous that conventional analysis in terms of solid and liquid breaks down. The molecules 'hanging out' of the cartilage will be bound to enough water to allow the combination of lubricin and hyaluronan to confer their specialised non-Newtonian lubricating effects. So in a sense a joint with living cartilage can never be 'dry'. During prolonged standing any free fluid will be expelled but water trapped in 'hanging out' hyaluronan will stay where it is and since there is no shear the hanging out hyaluronan may actually build up.

The thickness of synovial fluid in normal joints is probably about 40 microns at rest without load over much of the surfaces but goes up to several millimetres. It is almost certainly the minimum amount that does not cause significant vacuum effects on movement. There is probably no mechanism available for making the volume less so there is probably never a clinical problem associated with inadequate amounts. The problems occur when volume increases in effusion. Movement is limited and so may be cartilage nutrition because of the increased diffusion gap.

Pain probably comes not from poor lubrication but other secondary effects on tissue irritation or internal bone pressures when structure is disrupted.

Dear Jonathan! I’ve got extremely interested in your statement "Studying synovial tissue and fluid twenty five years ago led us to understand why synovial joints are involved in diseases like RA ... ". As for your following saying - I completely agree with it. For the last 10 years, we were studying the relatives of the RA patients. 10% of them developed RA during our observation. And we see that the various immune and inflammatory indexes are similar to those in the RA patients. Why, in your opinion, the process is toggled to the joints?

Micro-traumas? Accomplices in joint tissues (SNPs of some genes or somatic mutations of the factors the indifferent to the functioning of cells in other tissues and critical for the joints)?

In 1979 I set out to try to find out why synovium was the target for RA. I thought mechanical stress must be a factor but it was too crude an explanation. So we studied the specialised properties of synovial tissue cells. By 1987 the use of monoclonal antibodies to pull out gene products was in full swing and we were able to identify two special features of synovial cells. The intimal macrophages belong to a small subset of tissue macrophages that express CD16 - found only in synovium, alveoli, pleura/other serosae, sclera, bone marrow, secondary lymphoid tissue, salivary gland and subcutaneous tissue only at sites of pressure. Which, as you know are exactly the sites of synovial and extra-articular features of RA. The localisation of CD16 in skin to pressure points is quite remarkable - and we suspect mechanical stress does have a role in inducing this gene product. The second finding was that synovial intimal fibroblasts express much higher levels of VCAM-1 and DAF than stromal cels in other tissues, except bone marrow (nurse cells) and lymphoid tissue (FDC).

During 1996-2000 we built an explanation based on these findings and published in a number of journals and chapters (see on Pubmed). The proposal is that the critical inflammatory pathway in RA is ligation of tissue macrophage CD16 by very small immune complexes with TNF production. This would be predicted to occur at all the CD16 sites. However, VCAM-1 and DAF are crucial for B lymphocyte survival (as in nurse cell and FDC function) so in synovium B lineage cells can colonise the tissue in response to TNF and survive there to form ectopic lymphoid tissue (synovial fibroblasts also more readily express CD21 of the FDC isoform on stimulation). As a result complex forming antibodies are manufactured in situ in the joint.

The test of this idea was to see if B cell depletion would produce improvement in association with a fall in autoantibody levels. We tried this with rituximab in 1998 and it turned out to work very nicely, although there is still much more to learn.

I am no expert but am curious to know your opinion on the new paper describing the correlation between RA and the gut bacteria, Prevotella copri.

The new Nov 2013 paper by Jose U Scher, et al., is titled, "Expansion of intestinal Prevotella copri correlates with enhanced susceptibility to arthritis"

http://elife.elifesciences.org/content/2/e01202

https://www.researchgate.net/researcher/34494854_Jose_U_Scher/

https://www.researchgate.net/profile/Andrew_Sczesnak/publications/

Jonathan, thanks for the detailed answer. Of course, the puzzle is not fully formed, but thanks to you part of the picture has become more complete.

It seems that our discussion smoothly flows from the problem of the synovial fluid into to the problem of RA. I’m also interested in this article about the correlation between RA and Prevotella . However, the attempts to link RA with the certain infectious agent are plentiful and endless. Think of a whole pile of the publications on the relationship between RA and Proteus. We showed (unpublished data) the increased bacterial colonization (Staph. epidermidis in particular) in the RA patients and their relatives. It seems to me that the problem lies not in the specific infection, but in the excessive burden of various trivial infections.

I think we may be wandering off topic a bit! I see no need to invoke any infectious trigger for RA. The process can be explained by a vicious cycle set up when a chance combination of antibody species generated by random mutation (as always) allows feedback mechanisms to be subverted. That was the model that led to the rituximab study. The best account may be in our Nature Reviews Immunology review, or our 1999 Immunology paper. I was puzzled by the Prevotella study. It is hard to see how it could make sense in terms of what we know of the epidemiology but I am prepared to believe that things are more complex and that might be some sort of role for gut bacteria. It is just that it seems to me that we have an adequate explanation without invoking an external trigger. That idea was to my mind just a hang over from the days when rheumatologists treated rheumatic fever.

I like the felicitous definition of Meda F., et al (2011): RA patient is an individual with "bad genes and bad luck”. The genetic and environmental factors are believed to play equal roles in RA development (40-60% - SNPs, shared epitopes+TNFalpha+PTPN22 genes being the leading trio; thus, the environment is quite an equal partner) . So RA, in my opinion, not so much nosological unit, but a condition to which the various "curve, deaf and devious paths" (Strugatsky, "Definetely Maybe") can lead. In some cases it may be an infection in the other - trauma, time pressure or happy marriage ...

Ah, but Marina you leave out the most important factor of all. Ask an epidemiologist and she will tell you that there are three types of factor (maybe more subtypes but 3 important ones): genetic, environmental and internal stochastic. Cancer is very dependent on the internal stochastic aspect - which is random mutation. And for B lymphocytes this mutation is truly random in the sense of independent of any external factors because it is driven by the random effect of activation induced deaminase on the immunoglobulin genes. So bad luck is the right answer in an important sense. The main environmental factor in RA is smoking, which is not bad luck really. But the main stochastic factor is almost certainly the chance generation of a group of B cell clones that can form a vicious circle of autoantibody production. The action of AID is the same as in any phsyiological Ig rearrangement and probably also in most of the translocations that cause lymphoma.

In fact I suspect we cannot split cause by percentage because the mathematical relation between the different factors is geometric, not additive, but I would guess that if we are apportioning blame it is about half genes, a third stochastic and only a tiny bit or one sixth environmental and if people stopped smoking it would not even be that!

Why do rheumatologists always ignore the stochastic part? The oncologists have always been aware of it!!

The interesting idea, really. Never thought from this point of view. Is the role of the stochastic factors in the RA provocation realized only in the form of random mutations in immunoglobulin genes? Or it is only an example of the random events in RA? All the same, B lymphocytes are undoubtedly important, but not the only and not the basic link in the pathogenesis of RA.

I must say, this idea (annexed to the the RA) seems to be debatable, in any case, contrary to many well-known positions.

Perhaps we should debate? Why do you say B cells are not the 'basic link' in RA, Marina? What else could be? And what would these 'well-known positions' be? Hopefully not the T cell theories that arose as a misinterpretation of early histological studies. They should have died from starvation of any supportive evidence by now! Nobody ever found evidence for an abnormal T cell in RA, or a response to specific T cell targeting, but evidence for abnormal B cells goes back to 1948 and is used in routine diagnosis every day of the week and specific B cell targeting works well.

Ok, Jonathan! The challenge is accepted. The decent answer will not be slow in coming.

Think I'm joining this conversation a bit late but I did my PhD with Prof JR Levick who was, in my opinion, an understated legend in this field and has published a lot of very useful information about synovial fluid composition and fluid dynamics across the synovium from experiment on rabbit joints in vivo. Unfortunately a lot of the full texts are unavailable online.

The viscosity of the fluid is largely accounted for by hyaluronan (HA), facilitated by its interactions with lubricin making it an excellent hydrodynamic lubricant, and to a lesser extent boundary lubricant (only in association with lubricin).

HA in the synovium is very important in the retention of synovial fluid within the cavity - indeed removal of HA from the synovium will increase the permeability of the lining 5-fold in rabbit joint (Coleman et al 1998). In addition HA in the synovial fluid (3.62 ug/ul, rabbits) forms a concentration polarisation layer against the synovium at a concentration >0.9-1.3 ug/ul (rabbits) which in results in a rise in osmotic pressure at the synovial surface, reducing fluid outflow. Therefore I would ensure you have an adequate concentration of high molecular weight hyaluronan (>3x10^6 Da) in your artificial fluid if you want to preserve 'normal' fluid dynamics.

I will try to find some of the older papers which may be relevant to your initial question about composition of synovial fluid.

Yes, I quite agree, Karen. Rod Levick and I had a lot of fun working out how synovial fluid really worked in the late 1980s and early 90s. Rod went on to provide the most comprehensive account there is, while I wandered off into immunology. Now that we are both retired hopefully someone like yourself can keep the work going!

I did think you must know of him Jonathan given you were working in the same field over the same period. I will remember you to him when I next visit. I would love to return to this work as I had a fantastic PhD, unfortunately Rod retired the year I submitted so I have had to change field (neuroscience) but maybe one of the funders will take a chance on me and fund a HA-joint related project so I can return to this field in the future :)

OK so I have dug out a chapter Rod wrote in a book which has two tables relating to human synovial fluid composition which may be useful to answer your original question, Venkatachalam. I have photographed them and attached them below. If you want a copy of the entire chapter I can send you one but the scanner at work is broken so it is probably quicker for me to have your address and mail it. The reference for these tables is listed below

Levick JR (1983) Synovial fluid dynamics: the regulation of volume and pressure, In: Studies of Joint Disease (eds. Maroudas, A. and Holborow, E.J.) pp153-240 Pitman Books London

Serum proteins in normal and rheumatoid synovial fluids from human knees

Dear Karen,

Our group's interest in hyaluronan has been continued by Andy Pitsillides at the Royal Veterinary College, particularly on the embryological side. The remaining puzzle I see is whether there is truly an 'oozing lubrication' mechanism and whether oozing is crucial to cartilage resistance to proteases. The problem has always been devising a convincing model system. If you do not know Andy you might be interested to compare notes. I think there is one more significant piece of the jigsaw to fill in and you and he might be best placed to tackle it.

Yes I know Andy well and I have been in contact since leaving Rods lab. I expect I will catch up with him at the upcoming BSMB meeting in Bristol so I will be sure to research and discuss this aspect when I do. Thank you for bringing this puzzle to my attention :)

Hi Jonathan

I have been thinking about your idea - I read your previous answer to Ulhas and as I understand it your point is that the 'ooze' at the surface of the cartilage is preventing inflammatory mediators and enzymes from direct access to the cartilage surface - hence protecting them from degradation? I imagine that this 'ooze' would have a higher concentration of both HA and lubricin than synovial fluid further in the joint and therefore have a similar composition as the concentration polarisation layer Rod described at the surface of the synovium? Given the absence of significant fluid flow through this cartilage boundary you think it is likely that the HA in the ooze, maybe in conjunction with lubricin, is bound directly to the cartilage surface like a glycocaylx - if so do you have any idea if this is an active secretion from healthy chondrocytes hence 'hanging out' of cartilage or derived from HA in the synovial fluid and potentially bound to the cartilage by HA binding proteins like CD44 (I know Andy has shown expression of CD44 on the articular surface)?

The reason I ask is directionality - if it is coming from the chondrocytes then when effusion occurs there would be an opposing force (net outflow) against extrusion of HA into the cavity which may explain the slow break-up of the 'ooze', conversely if the ooze is primarily made up of HA actively bound to its surface there might be more resistance to the breakdown although I expect both would breakdown at a certain point.

We always maintained that the source of synovial HA was almost entirely derived from the fibroblast-like synoviocytes since the dissociation constant of HA in the cartilage matrix is so high and unbound HA would not pass easily through the tiny pores of the cartilage matrix but I guess it is entirely possible that the surface chondrocytes are secreting a glycocaylx like Andy sees in culture and this is having the protective effect?

I guess as you have pointed out the answers will only come when someone can derive a means of identifying and measure the 'ooze' on the cartilage - I will get my thinking cap on as to how this may be possible.

Oh, the joys of arthrotribology! I am not sure that much will change with an effusion. The increase in intra-articular hydrostatic pressure would be modest in most cases. And since convective water flux within cartilage matrix will be so constrained by association with macromolecules I am fairly sure that what we used to call the coffee grounds effect (in intima) would not operate at the cartilage surface. Rather than being like trying to squeeze a hyaluronan solution through a 0.22 millipore filter (what made me think of the coffee grounds in the first place) it would be a bit like trying to squeeze the solution into a brick wall I think - or maybe like the effect of water in a bucket on the rate of escape of air from a car tyre with small puncture (not much). One gets into the strange no man's land between hydrodynamics and solid colloid colligative properties that we used to discuss with people like Jill Urban and George Nuki.

The oozing theory may not be consistent with the rates of macromolecular flux, although the more the matrix HA is bound the more it should hang in there, in proportion to the slower it comes out - it cancels out I think. Maybe somebody needs to do atomic force microscopy on fresh live cartilage surface under vigorous washing conditions. Andy and I have puzzled over this for years, but one day it ought to be possible to think of the right experimental system. I think it is quite important to basic clinical care because it would give us an idea why it might be urgent to use TNF inhibition under certain circumstance (like within hours, to prevent surface chondrocyte death from hypoglycaemia) and not others. At the moment I don't think rheumatologists have a clue about this. At the moment it is routine for patients to wait 3 months for a TNF inhibitor to be approved, but as you know, not many biological processes wait 3 months - certainly not enzymes.

Hello to my old friend Joe Edwards as well as the other contributors to this dialogue. Some of you may be interested in a fairly recent paper I did with John Bassett entitled "Pathways of microvascular permeability in the synovium of normal and diseased human knees" (J Rheumatology 2011; 38:2635-2642. In it, we used a classical 3-pore model and gel filtration chromatography of the entire serum and synovial proteomes together with my earlier kinetic data to assess the radii of fenestrae as well as "small" and "large" pore sizes and the mean flows through these apertures. The data fit well with this model (r=0.992 in 17 normal knees and 0.980 in 16 synovial effusions). Those interested in the proteins of synovial fluid and how they got there may want to take a look at it.

Peter Simkin

And hello Peter, glad to hear you are still hard at it. I have to say I have been largely cheering from the terraces since Andy took over the HA work in London.

Thank you Jonathan, Venkat, Marina, Derek, Karen & Peter for a very interesting and informative discussion.

My interest in this subject is formulating a medicine to improve synovial fluid quality and quantity. I am technologist, with a focus to build a bridge between pure science and its application for better life.

Because of SF deficiency or change in quality, the person experiences pain and difficult movements. How to reduce pain and improve movements, with a safer medicine? I always think of safety of the medicine, and that is why I have moved to herbal sources. After discussion with practicing doctors, I designed a formulation. And surprisingly with some modifications it is working very well. Many patients have expressed their satisfaction.

So my interest increased. But how and why it happens, is the question for me. So I thought that if somebody measures quantity of SF before and after the treatment, I would be happy. And I prefer, it should be done, without disturbing the synovial capsule.

So there is need of some technique for the assessment. With my limited sources, it is not possible. But I am trying to understand the phenomenon. So I had opened a separate question in the forum, about assessment of SF quantity. But did not get good response except two answers, of which one is from Jonathan.

Now my next project is about cartilage management. Difficult but no harm in trying.

If any institute, can do it, as a research project, I will offer my full cooperation.

I will be in London on 24th & 25th June 2014. I would be happy to meet Jonathan and K , if they wish and have time.

synovial fluid composition can be found in large number of papers; old and new.

Dont know about artificial synovial fluid; but , viscosupplementation is a treatment option for osteoarthritis in that direction. Synvisc or hylan GF 20 is one such agent made from rooster coomb.not artificial though.

Just to add my two cents regarding the one of the previous comments (between Jonathan and Derek ) about the possible role of infectious agents in triggering RA, in light of the recent paper Science Translational Medicine paper describing the role of neutrophil NETosis as a potential source of citrullinated proteins (that could be the antigen that the auto-antibodies in RA are binding too), perhaps there is a feedback cycle where an infection could trigger neutrophil NETosis, releasing citrullinated proteins that are then bound by ACPA, which then in turn can cause more NETosis by binding to and activating more neutrophils, resulting in more auto-antigen... etc etc. This sort of a feedback loop seems to explain the strong B-cell/antibody dependence of RA, as well as evidence that infectious agents seem to also be linked (increased RA flares after dental work, relationship between gingivitis and RA, gut bacteria, etc). I would appreciate insight from experts in the field as I am relatively new to the RA field.

Dear Jonathan,

The trouble with trigger theories is that they ought to be triggering all the time - every time someone with DR4 has a boil or a root abscess they should get RA. And the epidemiology really does not hold up. People may get more joint pain after dental problems but across the board it does not add up in solid statistical terms. Lots of people have gingivitis. We all have gut bacteria!

For some reason people have been obsessed with looking for infective triggers for RA for over a century. There is this myth about that causation is either genetic or environmental. But when Stastny discovered the genetic link to DR4 he made a point of emphasising that the rest of the causation was statistically stochastic (random). That would fit with random generation of loops within the B cell repertoire through Ig gene mutation, not with an infective trigger. It is extraordinary how people will cling on to ideas when the maths says they must be wrong.

So yes, loops involving endogenous citrullination are very likely to be relevant, but they are just as relevant with no hint of an infective trigger - the loop structure is internal. It is only interesting if it can drive itself once a trigger has gone anyway so why do we need a trigger at all? What is needed is some function for citrulline in B cell survival signalling. (It is not enough just to generate more antigen, citrullination goes on in the body all the time anyway.) Various people think they may know what that is but it is still up for grabs.

Thanks for the information Jonathan. I am reminded of a lecture that I saw once many years ago about platelets (nothing to do with RA or inflammation) and the speaker was describing experiments where they injected small, platelet sized polystyrene beads into a mouse (i.v.). The speaker stopped his presentation to talk about how it was quite striking that when they opened the mice up, they found that a strikingly high percentage of the beads ended up in the joints of the mice. It was just an off hand comment at the time, but it always stuck with me. Now that I am working on RA I can't help but wonder if the same would be true in humans. Given that DNA and proteins from bacteria have been found in the joints of many people (healthy and diseased), and that the human joint is lined with macrophages, and the afore-mentioned links between infection and RA, it is tempting to speculate that the joints are one of the sites that bacteria/small particles are cleared from the circulation. Maybe there is a threshold effect where if too many bacteria end up in the joint, they are able to trigger neutrophil activation, and in DR4 patients this leads to ACPA due to the citrullinated protein. Yes, citrullination happens everywhere, but maybe there is a citrullinated antigen that is specific to NETosis. I agree that stochastic processes are likely involved in the initial generation of the auto-antibody, but it doesn't seem to explain why RA goes through remissions and flares. Maybe only a small percentage of patients with gingivitis end up with enough bacteria in the blood (due to flossing, injury in the mouth, etc) to trigger the flare? This could also be true for gut bacteria (yes, we all have gut bacteria, but maybe some people have injury (ulcer, food allergy, etc) that allows enough bacteria to enter the blood stream. In this hypothesis, a boil or root abscess would only trigger RA if enough bacteria entered the blood stream

I am curious as to what you think the auto-antigen is? If it is something common that is present all of the time, you would expect that the auto-antibody titers (e.g. ACPA) would be sky high in RA patients all of the time, being constantly boosted by endogenous exposure. Alternately, you might expect a tolerance effect due to the chronic exposure (such as seen in hemophillia patients with antibodies against FVIII, after high dose, daily exposure to recombinant FVIII, the majority of patients establish immune tolerance and their antibody titers disappear). The kinetics of ACPA titers in RA reminds me more of an intermittent exposure model (such as anti-AAV antibodies, titers can remain high for many years after initial exposure, and then the titers can increase again after subsequent exposure) rather than a constant exposure model.