Circulating microRNA expression signatures accurately discriminate [ME] from fibromyalgia and comorbid conditions 2023 Moreau et al

[Andy]

Abstract

Myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS), and fibromyalgia (FM) are two chronic complex diseases with overlapping symptoms affecting multiple systems and organs over time. Due to the absence of validated biomarkers and similarity in symptoms, both disorders are misdiagnosed, and the comorbidity of the two is often unrecognized. Our study aimed to investigate the expression profiles of 11 circulating miRNAs previously associated with ME/CFS pathogenesis in FM patients and individuals with a comorbid diagnosis of FM associated with ME/CFS (ME/CFS + FM), and matched sedentary healthy controls. Whether these 11 circulating miRNAs expression can differentiate between the two disorders was also examined.

Our results highlight differential circulating miRNAs expression signatures between ME/CFS, FM and ME/CFS + FM, which also correlate to symptom severity between ME/CFS and ME/CFS + FM groups. We provided a prediction model, by using a machine-learning approach based on 11 circulating miRNAs levels, which can be used to discriminate between patients suffering from ME/CFS, FM and ME/CFS + FM. These 11 miRNAs are proposed as potential biomarkers for discriminating ME/CFS from FM.

The results of this study demonstrate that ME/CFS and FM are two distinct illnesses, and we highlight the comorbidity between the two conditions. Proper diagnosis of patients suffering from ME/CFS, FM or ME/CFS + FM is crucial to elucidate the pathophysiology of both diseases, determine preventive measures, and establish more effective treatments.

Open access, https://www.nature.com/articles/s41598-023-28955-9

 

[cassava7]

Looking at Figure 2, it seems that a better title for this article would be “Circulating microRNA expression signatures accurately discriminate fibromyalgia from healthy controls and comorbid conditions”. In terms of statistical significance, patients with ME/CFS have the same relative miRNA expression as healthy controls except for (c) hsa-miR-127-3p (p < 0.01), (g) hsa-miR-374b-5p (p < 0.05) and possibly (d) hsa-miR-140-5p (p < 0.05).

 

[rvallee]

I don't think it's valid to separate them as distinct diseases. Most of the chronic illness diagnoses are just historical artifacts from medicine happening in siloes trying to force a concept that doesn't fit in the traditional definition of a disease that has to be a unique snowflake determined by a single pathophysiology. They should be viewed similarly to cancer in that depending on the location affected the impact and illness will be different but ultimately have similar mechanisms.

What's especially important is to move on from an immutable model. There are many attempts at making phenotypes in Long Covid and none of them account for large fluctuations, account for variation over time, so all they do is take single snapshots in time that basically miss most of the information.

If we take examples of long haulers, or any of us really, who would fit various criteria for ME, fibromyalgia, IBS, MCAS and other criteria, what does it even mean? Especially as it changes over time. The current model of medicine is unfit for this, because it makes assumptions about what constitutes a disease based on the first easy pickings. It's the lack of adaptation that gets to me, the absolute lack of learning from reality contradicting flimsy models, and yet the models win it every time. What a mess.

Usually this is framed as comorbidity, but the traditional way of independent diseases happening without any connection to one another, like having cancer and diabetes. But it's clear that it's a package deal where different parts happen for some and not for others. The whole thing is badly in need of a full reset, skipping the foundations never works out well when you try to build anything on top of it.

 

[Mij]

I haven't been able to read the entire paper.

Are the miRNA expressions in ME/CFS and FM patients tested before, during and after a post-exertional stress challenge?

 

[ME/CFS Skeptic]

So ME/CFS patients had higher levels of these microRNA's than healthy controls while FM patients had lower levels than healthy controls. Interesting result, just hope this wasn't due to the different selection process of the FM samples (they came from a different database).

 

[Jonathan Edwards]

I find this very hard to interpret but I suspect that the findings in figure 2 of high levels in ME and low in FM and middling in normals is just an artefact of them looking for the biggest differences between ME and FM - which are more or less bound to show that pattern or near enough.

To me it is odd that the mixed group does not have a wider standard error - more variation. I would have expected that mixing a process that put something high with a process that put it low would produce more scatter than either process alone.

I am sceptical that if you need to measure 11 things to make a distinction that any of them are very interesting. In other diseases you may need three or even four measurements to distinguish different diseases (for instance calcium, phosphate and alkaline phosphatase in the metabolic bone disorders). But mostly you only need one measure or maybe two to distinguish disease from normal. If you need 11 measures I doubt they are telling us anything direct about mechanism. And as I have said before, if tests like this tell us nothing about mechanism they have no advantage over the original gold standard of clinical diagnosis.

 

[FMMM1]

If GWAS identifies sub-groups then I wonder if repeating this miRNA study, on those sub-groups, would yield something useful.

 

[Tilly]

I find this very hard to interpret but I suspect that the findings in figure 2 of high levels in ME and low in FM and middling in normals is just an artefact of them looking for the biggest differences between ME and FM - which are more or less bound to show that pattern or near enough.

To me it is odd that the mixed group does not have a wider standard error - more variation. I would have expected that mixing a process that put something high with a process that put it low would produce more scatter than either process alone.

I am sceptical that if you need to measure 11 things to make a distinction that any of them are very interesting. In other diseases you may need three or even four measurements to distinguish different diseases (for instance calcium, phosphate and alkaline phosphatase in the metabolic bone disorders). But mostly you only need one measure or maybe two to distinguish disease from normal. If you need 11 measures I doubt they are telling us anything direct about mechanism. And as I have said before, if tests like this tell us nothing about mechanism they have no advantage over the original gold standard of clinical diagnosis.

They are saying only 3 "For the first time, to our knowledge, we provided evidence showing that miRNA expression levels miR-127-3p, miR-140-5p and miR-374b-5p could be potential biomarkers for ME/CFS and FM illnesses.

However like cancer and other diseases the circulating dysregulation can point to a part of a disease that has common elements of the same type of disease, but need the rest of the miR to pin point the exact issue and for ME the pin point should always be on PEM. They found that "has-miR-6819-3p is associated PEM and also with symptom severity", so do we have it or not? We need studies large enough and sensitive enough to detect PEM. So one part of the complex story that is ME and PEM and that part alone, I would like studied. We have too many children not being recognised as needing rest, needing tube feeding etc.

Does this stand alone and could it be used as a control like insulin detection? we simply do not know but it could! Could this be sensitive enough and what exactly does it communicate to the rest of the systems and circulating miRNA target treatment or in this case maybe rest, like cancers ascertain certain breast can be detected with 19 of them to target personalised care, so they can target treatment?

It is too early to know, but it is too early to dismiss because we have only just started to take these expressions seriously. I believe the earliest look at this in ME was in 2003, where could we be now if research money was put in. We must not dismiss this study! We need to make sure researchers get interested and take this forward, remember the study of TNf-a and four days of up regulation after a hospital visit, that was in 2004. We have a lot to learn and we are very slow at looking and very fast at dismissing.

 

[Simon M]

Striking results

I don't think I've ever seen a set of results as striking as those in figure 2, so I am interested in the study. Here are some comments on it:

Pros

Prespecified miRNAs: they didn't just hook out a select set of nice 11 micro-RNA results from a much larger dataset, instead they had been selected on the basis of a previous study of ME/CFS. That makes them more interesting.

There is a clear and logical pattern of high expression for ME/CFS, low for fibromyalgia, and ME/CFS cases with comorbid fibromyalgia someone somewhere in between.

There are clear differences between
- ME/CFS and fibromyalgia.
- ME/CFS without FM and ME/CFS with FM.
- FM and controls.

The biomarker models achieved remarkable separation between groups in most cases. And they were developed by splitting the samples into a training set and a separate test set to generate results (good practice).

Cons
– In the original findings, these 11 micro-RNAs were different between ME/CFS and healthy controls, but only three were significantly different here, which is hardly replication.
- Expression of all miRNAs in fibromyalgia is very low. My concern is that the sample preparation was different for fibromyalgia-only patients, and that might have affected the results.

The team drew blood for the ME/CFS (+/- FM) and healthy controls and extracted plasma from that. For fibromyalgia, they took plasma from a biobank. It would've been reassuring to see that other miRNAs not implicated in ME/CFS show normal values for biobank samples.

- Correlations with symptoms are probably just by chance. They made a lot of uncorrected comparisons (299) of miRNAs vs symptoms and found 15 significant ones, no more than would be expected by chance.

There were 299 comparisons, so with P < 0.05, we would expect 15 significant findings, which is exactly what we got. (11 miRNAs x 13 symptoms groups x 2 [ME +/- FM] = 286 comparisons + 13 for fibromyalgia [1 pain symptom] = 299). Using P < 0.05, we would expect 5% – 15 of 299 – cases to be significant just by chance. so we should assume the correlations with symptoms are chance ones.

More if you are interested.

Biomarker issues. It's much easier to get clean separation between groups if you use a lot of 'biomarkers' and this group used 11 (one or a few is normal). Also, the 'test' groups all contained fewer than 10 subjects, which is very low.

The discussion does pick out 3 miRNAs that are higher controls for ME/CFS and lower than controls for fibromyalgia. As the authors suggest, these might be useful as biomarkers to distinguish ME/CFS from fibromyalgia, but that would need independent replication. I'm surprised they didn't attempt it on the current dataset.

I am unconvinced by the biological speculation.
In biology, many factors (in this case, regulatory micro-RNAs) affect many other things (in this case, gene regulation), but it's not that hard to construct a plausible story of how factor X affects disease Y. As the study notes, further research would be needed to demonstrate that the differences in micro RNA levels do indeed cause biological changes relevant to the disease. Even then, it wouldn't be clear if this was a cause of the illness or an effect.

Though it would still be quite something.

I'll be very interested to see what this research group do next on micro-RNAs.

 

[Hutan]

Yes, the fact that there is hardly any variation in the fibromyalgia cohort (everyone in the cohort has values for the eleven miRNAs that are close to zero) makes it very likely that that is an artefact of sampling error (treatment or storage). I'd be willing to bet on it. If that's right, then the study doesn't tell us anything about whether fibromyalgia and ME/CFS are different. I'm with Rvallee in thinking that a lot of fibromyalgia is probably the same as or similar to a lot of ME/CFS.

I'm surprised the authors felt so sure that they could claim that ME/CFS and fibromyalgia are distinct diseases in the abstract.

The results of this study demonstrate that ME/CFS and FM are two distinct illnesses

The three miRNA where there were significant differences between ME/CFS and healthy controls are interesting.

 

[FMMM1]

The three miRNA where there were significant differences between ME/CFS and healthy controls are interesting.

Apologies but could someone specify the "three miRNA where there were significant differences between ME/CFS and healthy controls"

Anyone know their function?

Thanks

 

[SNT Gatchaman]

The three miRNA where there were significant differences between ME/CFS and healthy controls are interesting.

Apologies but could someone specify the "three miRNA where there were significant differences between ME/CFS and healthy controls

Anyone know their function?"

I think it was the three mentioned by cassava7:

In terms of statistical significance, patients with ME/CFS have the same relative miRNA expression as healthy controls except for (c) hsa-miR-127-3p (p < 0.01), (g) hsa-miR-374b-5p (p < 0.05) and possibly (d) hsa-miR-140-5p (p < 0.05).

Links to miRBase.org —

hsa-miR-127-3p
hsa-miR-374b-5p
hsa-miR-140-5p

Links to RNACentral.org —

hsa-miR-127-3p
hsa-miR-374b-5p
hsa-miR-140-5p

Both sites link to publications relevant to these mIRs.