Deep Sequencing of BCR Heavy Chain Repertoires in Myalgic Encephalomyelitis/Chronic Fatigue Syndrome, 2025, Ryback et al

I love when I have the energy to do this. The human body is so facsinating and complex.

I used to watch this francophone TV show as a kid “Il était une fois la vie”, basically following lives of red blood cells, immune cells, germs, salts, fats, bone marrow, mitochondria, etc. who were all characters inside the body. And the show was basically just kind of following what happened inside. I feel like I’m watching that show again when I read wikipedia pages about how the body works ahah.

 

I think the show you are talking about is called Once Upon a Time... Life in English and many kids watched it all over the world. I even posted about it earlier myself as I was also a kid who grew up with it. It was a great show for children

 

Yes ahahha. That’s so cool. I didn’t know it made it out of France/Belgium/Switzerland.

 

Thank you @Yann04 and @Wyva for sharing your memories, and the link! It looks interesting, I'll listen later.

Some questions/thoughts. I'm not specifically directing that anyone or even expecting answers!

Why would here be a skew towards IGHV3-30 gene usage? My understanding is VDJ recombination means a pretty random/even split usually. Do people think there’s an upstream (complement system or other feedback) or downstream (something within VDJ recombination itself) cause?

How much do we know about these VDJ gene areas and variation in the population? Would DecodeME or SequenceME potentially be able to tell us anything by looking for differences here between people with ME/CFS and healthy controls?

With all these potential combinations could there be some form of B cell receptor that itself is problematic in some way in ME/CFS? Is there are way of doing a more detailed survey of what the repertoire looks like (type or perhaps maturation status) in people with ME/CFS to try to spot this by comparing with healthy controls? Or more indirectly by doing some combinatorial analysis of patterns of possibilities based upon the genetic data of different groups? This study seems to have done some of this but I don’t understand how comprehensively.

 

Maybe going a bit off topic but I’ve been interested in how significant my response was to the covid-19 vaccines and often wondered if there is something to be learned here. Largely because of just the way I used to work troubleshooting problems in large complex computer systems. If you have something which reliably triggers a problem, try to understand more about why.

IGHV3-30 (amongst others) has popped up in covid-19 studies and here, although it seems to pop up in various places so maybe it isn’t significant.

But presumably as part of the immune system response to vaccination there would be clonal expansion once B cells are triggered? Do we know what B cell receptors are triggered or what genes these are sourced from? Are these the same for everyone or would slightly different receptor makeups be triggered depending upon our own repertoires and genetics and what ‘fits’? Could there be something here, either in genes used or in the somatic hypermutation phase?

I just wonder about how much the triggering events are because of general ‘load’ on the immune system or a specific bit of the pathways for us is not quite right, either always and waiting to be triggered or perhaps gets broken in some way.

And maybe we’re all the same or maybe different people with ME/CFS have different cells going wrong but in the same way. So maybe mine links to IGHV3-30 and someone else’s a different IGHV (or D or J) gene?

Or maybe this is all down/upstream from the regulatory stuff Jonathan talks about. The immune system is complicated isn’t it!

 

Well it's not uniformly distributed between genes in anyone, and there's a pattern of which genes are more common.

Here's another study on the B cell repertoire of people with hepatitis C just to illustrate that between the infected and healthy groups, as well as when comparing with this ME/CFS study, there's a pattern of specific genes being more frequent in everyone, with VH3-23 looking to be the most common, but with VH3-30 not far behind.

Biased IGH VDJ gene repertoire and clonal expansions in B cells of chronically hepatitis C virus–infected individuals (2018, Blood)


From this thread's study:


I wonder if some genes, over the course of evolution, like 3-23 and 3-30, have proven themselves to be more useful against the kinds of pathogens humans encounter, so are upregulated. And maybe there are signals in the body that can upregulate them further in preparation without the B cells actually seeing the antigen. Maybe something like the trained innate immunity being discussed on another thread sending signals to B cells saying "be prepared for this general class of pathogens by skewing your newly created B cell V genes appropriately".

 

You'd have to run it through an alignment algorithm. It looks like they used IgBlast, which I haven't used. They likely have a table of sequenced counts saved as a CSV--might be worthwhile to just email the corresponding author and see if they'd be willing to share with you. They'd probably just need to remove columns with identifying demographic info.

 

Thanks that is interesting. So a sort of fingerprint and then that varies with various diseases?

interesting thoughts on evolutionary selection over time and up regulation too.

 

FGF21 has popped up repeatedly, including in a recent LC study I did trying to predict post-hospitalization physical function. My guess is that it is a marker of energy deprivation (tied to Sirtuin signaling) released by vascular epithelial cells.

 

I did a bit of single-cell RNA-seq and TCR-seq in a very small LC study--on a single-cell level, the VH gene RNA levels did correlate to the TCR sequence of that particular cell. You might have detected some non-utilized VH genes, but the most abundant were the utilized ones.

But if you're doing bulk RNA-seq and haven't sorted cell types very finely, I think a lot of that goes out the window. As you and others have already mentioned, it's more likely to be due to a difference in cell type frequencies than VH gene utilization. Theoretically, you could somewhat address this through deconvolution by looking at the frequency of celltype-specific marker genes in your RNA-seq. But if the relevant subtypes don't really have strong RNA-seq signatures (as often happens when looking at T-cell subtypes, for example), this might not be possible.

Notably, we didn't see any hyperclonality related to LC, but it was a very small study.

 

I'm only just learning this stuff and don't really know how consistent these "fingerprints" are for different diseases. But in the case of an infection like HCV or COVID, a lot of the pattern should be influenced by B cells with certain gene segments binding to those specific viruses, then making lots of copies of themselves (clonal expansion), but that seems to be different from what's going on in this ME/CFS study where they didn't see any clonal expansion.

 

I can only agree that the last few posts are asking the right questions.

I have a thought about what this stuff might mean but it will be about month before I make it public.

In the very simplest terms the idea invokes what in philosophy might be called the 'screwdriver function'. Some philosophers put a lot of weight on the 'function' of something- to the extent that it might determine why we see the colour red (I don't agree). But although screwdrivers are supposed to have the function of driving screws, in my house they are more often used to open cans of paint.

Biology makes use of the can of paint function all over the place. Why not? If it works it is fit to survive. The immune system uses it for adhesion molecules. Although we learn that Ig gene recombination is 'random' there are lots of check points before cells get into the circulation that select by 'function', so it is not surprising that there is skewing. There may well some inherent skewing in recombination formal I know. (Mike Neuberger and others showed that sometimes Ig species randomness is more like a pseudorandom number generator that very nearly allocates options evenly.) Vh genes like Vh4-34 seem to have acquired some sort of paint can function that skews this - strangely to allow clones to survive to a certain point but not then affinity mature. Vh3-30 seems, from what I have seen recently, to be another 'usual suspect' that can buck the random system. People may have developed ideas on why but I am not up on that yet.

People have spent a lot of time looking for high affinity autoantibodies but it might be that antibodies are contributing to abnormal mechanisms because they are subverting the rules by using functions that at other times are useful but in this case are out of place. I think Audrey would agree that the Vh30-3 skewing is just a non-specific sign of some shift in control. But that does nutmeat it isn't an important clue.

 

I’m sure we all look forward to hearing more on your thoughts!

And I really like the reminder that nature finds ways of doing things, which we often miss trying to look for logical or efficient paths or patterns.

I like Hutan’s idea of people with higher levels being more ‘protected’. I also do wonder if it’s very unintuitive and severe is actually a sign of our bodies working ‘better’.

As I mentioned before:

So imagine the thing elsewhere which is trying to fix or clear up whatever is wrong is itself the cause of the symptoms we see in ME/CFS. While the thing that is going wrong is actually non-symptomatic for the most part. Say, something with another gene or protein folding step, something which means some other b cells malfunction perhaps (maybe just one type) so they are filtered out in the checkpoint systems while the unaffected IGHV3-30s increase as a percentage.

I’m imaging milder people’s bodies just doing an okay job of keeping on top of the clear up, but the skew appearing and maybe the body isn’t very good at fixing things, just okayish. Then for some reason (chance, genetic predisposition, infection or other immune trigger) it decides ‘hey, let’s sort this out’ and then does a really good job of cleanup and gets the repertoire back into balance, but at the cost of health elsewhere. So it can look similar to auto immune but it’s not really… it’s something a bit different perhaps. Does this make sense?

I think it’s maybe what Jonathan was saying earlier….

So question. If IGHV3-30 is over represented for mild/moderate, what genes are we seeing under represented? Could this be the real culprit? Could it also vary depending upon patient so we wouldn’t see a single tell in statistical analysis?

 

To revisit the discussion of pathways, triggers, complexity, and concerns over disease categorisation I thought I’d give an example from software development.

In a simple software bug, someone performing certain steps will cause a particular problem to occur. This is reliable, repeatable. The only difficulty is getting a good description of the problem and those steps to reproduce the problem and then we can go about fixing it.

But a lot of bugs are more complicated. Particularly as software gets more complex and has multiple threads or networked systems (so multiple bits of code running and multiple inputs or things happening simultaneously).

You may for instance get lots of different reports of problems or crashes in different places when doing different things. To the people using the software these may look different, they may all be cases of something bad happening but there may be little pattern beyond that. There may be hundreds of ‘this crashes’ reports that look otherwise unrelated.

To us when we eventually isolate and fix the problem, it may be that many of these bugs which occurred in different ways with differing results were all down to a few or even just one problem. I’m hoping the latter happens with ME/CFS.

How do we do this? We gather lots of reports, we try to look for patterns, we try to reproduce the issue and hopefully we can analyse the actual systems to get telemetry or crash dumps to understand the state of the system when the problem occurs.

We may find more commonality immediately, maybe there are bits of code that pop up time and again in call-stacks (information on what code was running when the crash occurred), this could allow us to group and identify different crashes as being the same problem (or at least categorise them into a smaller number of problems).

Maybe not, maybe the crashes are all over the place in different bits of code, but eventually we find another bit of code somewhere else has been trashing bits of memory and overwriting things, so whatever happened to be stored there at the time breaks when it is accessed/run.

I see commonality here with a complex biological system and the approach of looking at DNA (our source code) but also the need to understand more about the state of the running system when things go wrong, and this is where things like this study add information.

I’m not sure if this helps people understand anything but hopefully it’s at least an interesting comparison.

 

Okay the last of my collection of posts from recent weeks and reading! Thanks for everyone’s indulgence. Sorry it’s been more of a brain dump than a discussion.

Big questions for me around the findings seem to be…

We don’t seem to know where in the system or B cell lifecycle this change or skew in gene usage originates. Or if there are there external or environmental factors. Is that correct?

From what Audrey says in the paper and comments from @Jonathan Edwards and @DMissa we need another study here. Something like:

Longitudinal study into gene usage and isotype patterns of B cells of differing maturation in ME/CFS

This would involve subtyping B cells (e.g naive cells, memory cells, or antibody-producing cells) for comparison. So we can

(1) see if the IGHV3-30 (or any other gene usage) pattern is clearer in a particular B cell subtype and where in the lifecycle/maturation it shows up

(2) look at ratios of different isotypes to see if the IgM/IgG ratio finding (or that of other isotypes) can be replicated and again if so in which cells it (first) shows up.

Possibly worth looking at Light as well as Heavy chain too although maybe that’s just asking too much? Could his help get a fuller picture of gene usage though?

He study would need a decent number of mild, moderate and severe patients. If this was a fresh longitudinal study then gathering info on activity, illness, symptoms etc as is done in other studies.

Adding a common vaccination many of us have in a year like the Flu vaccine could be an interesting challenge too. Although this all adds more complexity.

Would a study like this seem reasonable/feasible to people more in the know?

Is there anything that can be done to subtype retrospectively with the available data?

CureME and the ME/CFS Biobank do have some people with multiple samples (and there’s some participant crossover with the longitudinal HHV-6 study @wigglethemouse mentions which could be interesting) but I’m not sure if this would be enough people.

An entirely new longitudinal sample collection of severe people especially seems to often be a barrier for cost/feasibility of projects…

Then the usual: Who can we get to do it? How can we help it happen? I’m offering my blood and support, although it’s already in the Biobank

 

Thanks, yep, I contacted her and she plotted it and I posted later in the thread. I later reread the study and realized I totally missed they already checked basically the same thing I was wondering, so that was unnecessary. They looked at the IgM of only the participants with ME/CFS with really high VH3-30 vs HC and it wasn't significant.


There's one more thing I've thought of that might be interesting to check in this study's data. Sato et al found that the CDR3s of those with high VH3-30 were skewed toward length 18 (black bars in left-hand chart below, gray is HC). Might mean something if it's 18 here too.

Edit: As far as I can tell, Sato et al only reports CDR3 length in cohort 1. Cohort 2 also had high VH3-30, so I would think they'd want to check if the length replicates as well.

 

You said this on the 25th of March. Are you still thinking of making your ideas public on that kind of timescale?

And I seem to remember you said you theory would also indicate a potential existing treatment. Does that still look like a possibility?

thanks

 

As always, things have got a bit more complicated, but only because they are more interesting. And this has become a collaborative venture.

It is still not a month yet but being realistic I think this is going to take a little bit longer than that. Not that long though.

 

Thank you, that’s still encouraging.

 

I am working on some paragraphs now. And Jo should be round for tea tomorrow!