Immunoglobulin G structure and rheumatoid factor epitopes, 2019, Maibom-Thomsen

[Jonathan Edwards]

A nice piece of science

I feel bad constantly pointed out how rubbish scientific papers are. But here is what looks to me to be the most important paper in the field of rheumatoid arthritis for twenty years.

For a start I suspect most people can actually understand what the abstract says. It says something very simple.

Nobody was able to explain how people with RA could have rheumatoid factor antibodies to IgG in their plasma when the plasma was full of IgG. The antibodies should have clumped the IgG and cleared it away within minutes.

What this study suggests - and the authors are careful to provide their evidence without claiming too much - is that IgG is normally in the plasma in a 'closed' form in which the places where rheumatoid factor would bind are invisible. It is as if the IgG is like a printer cartridge without tearing off the yellow paper that covers the ink hole. The printer cannot find any ink.

But if the IgG binds to an antigen through its arms the arms open up away from the 'body' revealing the binding sites for rheumatoid factor. We knew something like this happens for IgM and complement and we ought to have known something like this had to be true for IgG but nobody did the experiment for sixty years.

The great thing about this idea is that it not only explains something we should have known was a problem but it explains a whole load of other things, including giving a helpful clue as to why we make rheumatoid factors that only bind to this hidden site.

It may turn out wrong but I would bet not.

PLoS One. 2019 Jun 14;14(6):e0217624. doi: 10.1371/journal.pone.0217624. eCollection 2019.
Immunoglobulin G structure and rheumatoid factor epitopes.
Maibom-Thomsen SL1, Trier NH1,2, Holm BE2, Hansen KB1,2, Rasmussen MI1, Chailyan A1, Marcatili P3, Højrup P1, Houen G1,2.
Author information
1
Department of Biochemistry and Molecular Biology, University of Southern Denmark, Odense, Denmark.
2
Department of Autoimmunology, Statens Serum Institut, Copenhagen, Denmark.
3
Department of Bioinformatics, Technical University of Denmark, Kongens Lyngby, Denmark.
Abstract
Antibodies are important for immunity and exist in several classes (IgM, IgD, IgA, IgG, IgE). They are composed of symmetric dimeric molecules with two antigen binding regions (Fab) and a constant part (Fc), usually depicted as Y-shaped molecules. Rheumatoid factors found in patients with rheumatoid arthritis are autoantibodies binding to IgG and paradoxically appear to circulate in blood alongside with their antigen (IgG) without reacting with it.

Here, it is shown that rheumatoid factors do not react with native IgG in solution, and that their epitopes only become accessible upon certain physico-chemical treatments (e.g. heat treatment at 57 °C), by physical adsorption on a hydrophobic surface or by antigen binding. Moreover, chemical cross-linking in combination with mass spectrometry showed that the native state of IgG is a compact (closed) form and that the Fab parts of IgG shield the Fc region and thereby control access of rheumatoid factors and presumably also some effector functions.

It can be inferred that antibody binding to pathogen surfaces induces a conformational change, which exposes the Fc part with its effector sites and rheumatoid factor epitopes. This has strong implications for understanding antibody structure and physiology and necessitates a conceptual reformulation of IgG models.

 

[Jonathan Edwards]

I should perhaps add that the crucial experiment here, with cross-linking of amino acids in a whole protein and then mass spectrometric analysis of chopped up protein, where the cross-linked fragments are carried through together as if a single chain, has probably been possible since about 1995 but it is a stroke of genius to do it this way. This is the sort of careful detailed chemistry that moved mountains in the 1970s and still can.

 

[Hoopoe]

In ME/CFS we still have to discover what it is that isn't working right.

I don't get why everyone isn't focused on finding out what it is in the blood that affects cells exposed to it. If we figure that out then we have something to work with.

The other thing we have to work with, PEM is hard to study. If there were no ethical concerns then maybe we could discover the best and most specific ways to trigger or prevent PEM and thereby understand it much better.

 

[Kitty]

I love those 'headslap' discoveries – they make scientists seem much more human! (I feel as if I'm from a different planet to people who can grasp scientific and mathematical logic.)

 

[RuthT]

So - relates to the *something in the plasma’ that Ron Davis got from his basic test of swapping healthy/MECFS muscles and plasma. It was the plasma that showed *something*.

The big question is *what*

Ed Corrected spelling of ‘Davis’ and punctuation

 

[FMMM1]

So - relates to the *something in the plasma’ that Ron Davis got from his basic test of swapping healthy/MECFS muscles and plasma. It was the plasma that showed *something*.

The big question is *what*

Ed Corrected spelling of ‘Davis’ and punctuation

Bhupesh Prusty found something in the blood caused increased mitochondrial fragmentation [NIH Conference April 2019]. So maybe that gives a clue to what something in the blood might be and what it's target receptor might be.

E.g. this paper seems to link "adrenergic receptor signalling" to increased mitochondrial fragmentation [https://www.ahajournals.org/doi/full/10.1161/circresaha.117.310725]. So could the something in the blood be acting on the adrenergic receptor?

I'd like to see research to identify the something in the blood, it may be a way to make a significant breakthrough in this disease.

I think Bhupesh is drafting a paper - so maybe that will provide more insight.

 

[FMMM1]

I assume this is not relevant to the ME Community - at this point.

 

[Jonathan Edwards]

I assume this is not relevant to the ME Community - at this point.

The relevance to the ME community is that hidden away in the sea of meaningless pseudo-research in biomedicine it seems there are still islands of the sort of creative, tight work that was the norm fifty years ago. There are biochemists out there who can do the necessary experiments.

But I think it requires two things. One is technical scientific skill and perseverance. The other is having a broad enough perspective to generate a sensible hypothesis. What worries me about a lot of current ME research is that it takes ideas that sound OK on the surface but which with a bit of background knowledge begin to look pretty implausible. There is nothing surprising about that because it happens throughout biomedicine, but it would be good to have some islands of clear thinking.

 

[alex3619]

It may turn out wrong but I would bet not.

It sounds plausible to me. In any case this exposure of sites when configuration changes has been speculated in a range of antibody targets. We should indeed be investigating this more. Not only might a configuration change expose the target, but a lack of change may allow antibody making B cells to not be eliminated during early development.

 

[Snow Leopard]

The reason why no one did the experiment is because only researchers interested in "how people with RA could have rheumatoid factor antibodies to IgG in their plasma when the plasma was full of IgG" would be thinking about it. The sensitivity of the RF test for is quite low, 60-70% which suggests that the clearing out that Jonathan proposes does often happen.

 

[Jonathan Edwards]

The sensitivity of the RF test for is quite low, 60-70% which suggests that the clearing out that Jonathan proposes does often happen.

That is conceivable but in practice about 80% of people with clinical RA have RF. The negative ones have a limited disease without extra-articular pathology that has always seemed likely to be of different aetiology. RF-negative patients show no response to rituximab, which tends to confirm that they do not have the same disease. Psoriatic arthropathy can be indistinguishable from seronegative 'RA' and is almost certainly T cell mediated so it is not very surprising that in the past two unrelated diseases were conflated under 'RA'.

If people with seronegative RA had 'invisible' RF that were constantly cleared then it would be hard to see why they should get any arthritis or pericarditis, since these appear to be antibody dependent in RF positive individuals.

The reason why no one did the experiment is because only researchers interested in "how people with RA could have rheumatoid factor antibodies to IgG in their plasma when the plasma was full of IgG" would be thinking about it.

Yes, but hundreds of rheumatologists have been thinking about this since 1950. There is an interesting symposium from around 1960 where all the RA academics discussed the problem in detail and in fact proposed that Fc epitopes were curled up - long before Rodney Porter even sequenced IgG and very long before anyone had a tertiary structure. In the 1980s in Ivan Roitt's department the paradox was often discussed. In 1995 I looked into techniques like fluorescence correlation spectroscopy to try to analyse the form IgG existed in in RA plasma. I could have done the mass spec study these people did but it never occurred to me.

 

[lansbergen]

Overview for lay people

https://www.andrew.cmu.edu/course/03-410/Lec03/lec03.html

Lecture 3: Immunoglobulin Structure and Function

 

[FMMM1]

That is conceivable but in practice about 80% of people with clinical RA have RF. The negative ones have a limited disease without extra-articular pathology that has always seemed likely to be of different aetiology. RF-negative patients show no response to rituximab, which tends to confirm that they do not have the same disease. Psoriatic arthropathy can be indistinguishable from seronegative 'RA' and is almost certainly T cell mediated so it is not very surprising that in the past two unrelated diseases were conflated under 'RA'.

If people with seronegative RA had 'invisible' RF that were constantly cleared then it would be hard to see why they should get any arthritis or pericarditis, since these appear to be antibody dependent in RF positive individuals.



Yes, but hundreds of rheumatologists have been thinking about this since 1950. There is an interesting symposium from around 1960 where all the RA academics discussed the problem in detail and in fact proposed that Fc epitopes were curled up - long before Rodney Porter even sequenced IgG and very long before anyone had a tertiary structure. In the 1980s in Ivan Roitt's department the paradox was often discussed. In 1995 I looked into techniques like fluorescence correlation spectroscopy to try to analyse the form IgG existed in in RA plasma. I could have done the mass spec study these people did but it never occurred to me.

I assume that the other form of MS doesn't look like ME!

I don't know anything about using Mass Spectrometry to gain information about the structure of a protein; however, I wouldn't have expected that it would provide information on protein folding.

 

[Jonathan Edwards]

I assume that the other form of MS doesn't look like ME!

I don't know anything about using Mass Spectrometry to gain information about the structure of a protein; however, I wouldn't have expected that it would provide information on protein folding.

I cannot decipher the first comment!

But in terms of mass spec - exactly! why should it provide information on folding.
But they did something clever that apparently people have been doing for a while for other proteins.

You take native plasma protein and mix it with a cross linking agent that will bind at both ends to exposed residues of something like leucine (I forget which one). You then chop up the protein with the selective enzymes normally used for MALDI-TOF mass spec to generate peptides that are identified by their specific molecular weights (if I remember rightly).

You then discover that you have a peptide that let us say includes amino acids A+2T+M+R+B+S+G+ one cross linking molecule. That means your fragment is composed of two peptides that are not in sequence in the original protein but are close together on the protein surface. Let us say maybe MATR and TSGB. You then look up the protein sequence to see where the peptides come in the primary and secondary structures. If MATR comes from heavy chain CH3 region and TSGB from VH at the opposite end of the heavy chain then you are entitled to infer that the heavy chain is bent back so that VH lies in contact with CH3 with MATR nestling against TSGB.

The accepted tertiary structure of IgG has a flexible hinge half way along the heavy chain so that VH can go where it likes and it was assumed that it floated about at random. The cross linking agent might perhaps glue together a randomly floating VH to a CH3 on the odd occasion but my understanding of the paper is that the cross link occurred in two specific ways - which both required VH to be completely bent back on CH3, but with two slightly different options.

We have a resin model of IgG we made in collaboration with Brian Sutton, who crystallised a RF Fab bound to IgG Fc and from there worked out a complete tertiary structure for IgG. But it was done after first cutting the hinges so would not have picked up any further folding back. The interesting thing is that when we took the model and bent the Fab arms (CH1 +VH + light chain) back on to CH2/CH3 the two fitted together as if they had been made to. The prominent arginine sticking out the side of CH3 fitted beautifully into a deep pocket on the Fab.

 

[FMMM1]

I cannot decipher the first comment!

But in terms of mass spec - exactly! why should it provide information on folding.
But they did something clever that apparently people have been doing for a while for other proteins.

You take native plasma protein and mix it with a cross linking agent that will bind at both ends to exposed residues of something like leucine (I forget which one). You then chop up the protein with the selective enzymes normally used for MALDI-TOF mass spec to generate peptides that are identified by their specific molecular weights (if I remember rightly).

You then discover that you have a peptide that let us say includes amino acids A+2T+M+R+B+S+G+ one cross linking molecule. That means your fragment is composed of two peptides that are not in sequence in the original protein but are close together on the protein surface. Let us say maybe MATR and TSGB. You then look up the protein sequence to see where the peptides come in the primary and secondary structures. If MATR comes from heavy chain CH3 region and TSGB from VH at the opposite end of the heavy chain then you are entitled to infer that the heavy chain is bent back so that VH lies in contact with CH3 with MATR nestling against TSGB.

The accepted tertiary structure of IgG has a flexible hinge half way along the heavy chain so that VH can go where it likes and it was assumed that it floated about at random. The cross linking agent might perhaps glue together a randomly floating VH to a CH3 on the odd occasion but my understanding of the paper is that the cross link occurred in two specific ways - which both required VH to be completely bent back on CH3, but with two slightly different options.

We have a resin model of IgG we made in collaboration with Brian Sutton, who crystallised a RF Fab bound to IgG Fc and from there worked out a complete tertiary structure for IgG. But it was done after first cutting the hinges so would not have picked up any further folding back. The interesting thing is that when we took the model and bent the Fab arms (CH1 +VH + light chain) back on to CH2/CH3 the two fitted together as if they had been made to. The prominent arginine sticking out the side of CH3 fitted beautifully into a deep pocket on the Fab.

Wow - you certainly understand this -- and if I understand correctly, you were very close to solving it.

What I meant was there is a form of MS which does not respond to rituximab. As you highlight, this suggests that it is not B-cell autoimmunity. I wonder if symptomatically it looks like ME i.e. fatigue? Also, does the MRI indicate demyelination or some other evidence of inflammation/activation of microglia?

From memory there are two forms of NMDAR encephalitis; one responds to antiepileptic drugs and the other to rituximab (B-cell autoimmunity - Angela Vincent's Laboratory?).

Thank you for your insight.

 

[Jonathan Edwards]

What I meant was there is a form of MS which does not respond to rituximab. As you highlight, this suggests that it is not B-cell autoimmunity. I wonder if symptomatically it looks like ME i.e. fatigue? Also, does the MRI indicate demyelination or some other evidence of inflammation/activation of microglia?

What I have heard is that progressive MS does not respond, although I think there may be some response to one of the newer B cell depleters. The pharmacodynamics for MS are quite different than for RA because the problem is entry of B cells into the CNS and survival there with generation of plasma cells in a compartment that does not admit rituximab except if inflamed. My guess is that rituximab works in MS by removing B cells from circulation that are needed to prime a new relapse. In progressive disease there are no remissions and relapses, the disease just bats on inside the CNS regardless. But I don't think we have as clear a grasp of what is going on as in RA. There might also be a T cell form of 'MS' but even primary progressive disease has Its in CSF as far as I know.

I don't think any forms of MS really look like ME at all. There are specific localising neurological signs relating to areas of demyelination in all types of MS. ME does not have localising neurological signs.

 

[FMMM1]

What I have heard is that progressive MS does not respond, although I think there may be some response to one of the newer B cell depleters. The pharmacodynamics for MS are quite different than for RA because the problem is entry of B cells into the CNS and survival there with generation of plasma cells in a compartment that does not admit rituximab except if inflamed. My guess is that rituximab works in MS by removing B cells from circulation that are needed to prime a new relapse. In progressive disease there are no remissions and relapses, the disease just bats on inside the CNS regardless. But I don't think we have as clear a grasp of what is going on as in RA. There might also be a T cell form of 'MS' but even primary progressive disease has Its in CSF as far as I know.

I don't think any forms of MS really look like ME at all. There are specific localising neurological signs relating to areas of demyelination in all types of MS. ME does not have localising neurological signs.

Thank you very much for your reply.

 

[alex3619]

ME does not have localising neurological signs.

My understanding is that ME lesions have been shown to be transitory. We have them, and sometimes many, but their locations are nto stable. This was observed in the 80s, with I think a radiologist Cheney was using saying something about how our brains looked just like HIV dementia brains. I am not aware than any follow-up study was done.

What does seem to be consistent, at least in some patients, is deficits in brain metabolism found by using a radio-isotope tag and scan. Back when I was a mild patient my brain metabolism was way down on normal. As a moderate-severe patient its probably worse, but I have not been tested in many years.

 

[FMMM1]

My understanding is that ME lesions have been shown to be transitory. We have them, and sometimes many, but their locations are nto stable. This was observed in the 80s, with I think a radiologist Cheney was using saying something about how our brains looked just like HIV dementia brains. I am not aware than any follow-up study was done.

What does seem to be consistent, at least in some patients, is deficits in brain metabolism found by using a radio-isotope tag and scan. Back when I was a mild patient my brain metabolism was way down on normal. As a moderate-severe patient its probably worse, but I have not been tested in many years.

If you check out Jarred Younger's video link presentation in Sweden (last month), he discusses some recent experiments in brain imaging MRI etc. I assume that they've found further evidence of activation of microglia. He also discusses PET scans using a tracer (ligand); I think the idea is to repeat a previous study with one of the new more specific tracers. Again, I assume that most of this work uses MRI.

It seems like MS and ME are distinguishable i.e. since MS results in demyelination (detectable on MRI) and ME does not.

 

[FMMM1]

A nice piece of science

I feel bad constantly pointed out how rubbish scientific papers are. But here is what looks to me to be the most important paper in the field of rheumatoid arthritis for twenty years.

For a start I suspect most people can actually understand what the abstract says. It says something very simple.

Nobody was able to explain how people with RA could have rheumatoid factor antibodies to IgG in their plasma when the plasma was full of IgG. The antibodies should have clumped the IgG and cleared it away within minutes.

What this study suggests - and the authors are careful to provide their evidence without claiming too much - is that IgG is normally in the plasma in a 'closed' form in which the places where rheumatoid factor would bind are invisible. It is as if the IgG is like a printer cartridge without tearing off the yellow paper that covers the ink hole. The printer cannot find any ink.

But if the IgG binds to an antigen through its arms the arms open up away from the 'body' revealing the binding sites for rheumatoid factor. We knew something like this happens for IgM and complement and we ought to have known something like this had to be true for IgG but nobody did the experiment for sixty years.

The great thing about this idea is that it not only explains something we should have known was a problem but it explains a whole load of other things, including giving a helpful clue as to why we make rheumatoid factors that only bind to this hidden site.

It may turn out wrong but I would bet not.

PLoS One. 2019 Jun 14;14(6):e0217624. doi: 10.1371/journal.pone.0217624. eCollection 2019.
Immunoglobulin G structure and rheumatoid factor epitopes.
Maibom-Thomsen SL1, Trier NH1,2, Holm BE2, Hansen KB1,2, Rasmussen MI1, Chailyan A1, Marcatili P3, Højrup P1, Houen G1,2.
Author information
1
Department of Biochemistry and Molecular Biology, University of Southern Denmark, Odense, Denmark.
2
Department of Autoimmunology, Statens Serum Institut, Copenhagen, Denmark.
3
Department of Bioinformatics, Technical University of Denmark, Kongens Lyngby, Denmark.
Abstract
Antibodies are important for immunity and exist in several classes (IgM, IgD, IgA, IgG, IgE). They are composed of symmetric dimeric molecules with two antigen binding regions (Fab) and a constant part (Fc), usually depicted as Y-shaped molecules. Rheumatoid factors found in patients with rheumatoid arthritis are autoantibodies binding to IgG and paradoxically appear to circulate in blood alongside with their antigen (IgG) without reacting with it.

Here, it is shown that rheumatoid factors do not react with native IgG in solution, and that their epitopes only become accessible upon certain physico-chemical treatments (e.g. heat treatment at 57 °C), by physical adsorption on a hydrophobic surface or by antigen binding. Moreover, chemical cross-linking in combination with mass spectrometry showed that the native state of IgG is a compact (closed) form and that the Fab parts of IgG shield the Fc region and thereby control access of rheumatoid factors and presumably also some effector functions.

It can be inferred that antibody binding to pathogen surfaces induces a conformational change, which exposes the Fc part with its effector sites and rheumatoid factor epitopes. This has strong implications for understanding antibody structure and physiology and necessitates a conceptual reformulation of IgG models.

Please excuse further questions. Once the autoantibody can bind (IgG changes shape) then I assume this promotes the inflammatory process.