Metabolomic Evidence for Peroxisomal Dysfunction in Myalgic Encephalomyelitis/Chronic Fatigue Syndrome, 2022, Levine,Hornig,Lipkin et al

I clearly remember that Ron Davis some years ago said they were seeing high levels of lipids in cells in one of the experiments they did. Maybe in connection with cells being lighter?

Accumulation of lipids is something that can happen when the peroxisomes are not working well.
 
Jaybee00 said:
Yup. Sleep time.
Click to expand...
Yes, sorry, I did get rather carried away last night. I had a coffee later than I should have, and then it was a nice change to see a study that might actually explain something after the bad Australian knyurenine paper. I liked that the findings with this paper seemed to be in line with other reports, and that the findings in the female cohort were not markedly different from the total cohort (I think for things like cell membrane lipids, it's a bit of a red flag when cohort subsets have findings that are all over the place).

But, waking up this morning, I feel a bit embarrassed. An issue is that these metabolites are in the plasma, so, assuming these results do hold up as a reliable thing in ME/CFS, it's still hard to know what that means for us. We really need to know what is going on in cell membranes and inside cells.
 
Jaybee00 said:
Acetyl-CoA is an enzyme. Acetylcholine is neurotransmitter.

View attachment 17695
Click to expand...

Hutan said:
Whoops. No doubt there are other errors above. Past my bedtime.
Click to expand...
Coenzyme ;)

I'm not a fan of comparing whole classes of compounds between groups (the metabolic cluster figure @Hutan shared) as it might be specific compounds within the groups that are important and differences between whole classes of lipids etc could be due to something so simple as different fat intakes.

Look forward to read it later when my head is a bit less woolly :)
 
There are earlier studies which point to fats in the circulation as markers, and possible effectors of the disease. In 2003 Yoshitsugi Hokama’s discovery, showing that a lipid in the sera of patients cross-reacted to the antibody used to detect ciguatera toxin [23] shot like a bolt of light across the sky. Later, in 2009, it was determined that the identity of the antigen was phospholipids associated with cardiolipin from the mitochondria [22]. Unfortunately, further studies did not take place. Assuming this result was correct, this leaves the question: what are these phospholipids doing at increased concentrations in the blood? A likely consequence of Che and Lipkin’s study is that there may be increased, toxic, oxidized, lipids in the blood of patients with ME, which could, in some measure, also vindicate this early study.
Click to expand...
 
I have a low allele frequency mutation in the gene for the first step in the synthesis of plasmalogens. It could be a coincidence but it's also suspicious.

Also a mutation in a choline transporter that is involved in the synthesis of phospholipids and acetylcholine.

Affordable whole genome sequencing seems like a very powerful tool, if only we can improve our ability to interpret all the mutations that one can find with it. We need more data.
 
Last edited:
Jaybee00 said:
If anyone has questions on this study, I can forward to my contact at Columbia. Probably better to forward as a batch versus one by one.
Click to expand...

I have questions:

Do they think that plasmalogens could be a useful drug target? I noticed that there is some interest in treating other diseases with plasmalogen replacement.

Do they think carnitine supplement could be an effective treatment?

Do they have any thoughts on what could be causing this kind of problem?
 
I have read in two different articles that beta-adrenergic signalling is linked to peroxisome activity. People with a genetic disorder that impacts peroxisome formation have more beta-adrenergic receptors on their cells. In cell cultures, beta-adrenergic stimulation also reduces peroxisome gene expression.

So in a very superficial way, it looks like the peroxisome dysfunction might be a result of excessive beta-adrenergic signalling. But there is little information on this available. This stood out to me because a problem with beta-adrenergic signalling (due to autoimmunity) is one of the popular hypotheses on ME/CFS
 
Review article on peroxisome protein importing (paywalled) —

Import and quality control of peroxisomal proteins (2023)
Rudowitz, Markus; Erdmann, Ralf

Peroxisomes are involved in a multitude of metabolic and catabolic pathways, as well as the innate immune system. Their dysfunction is linked to severe peroxisome-specific diseases, as well as cancer and neurodegenerative diseases. To ensure the ability of peroxisomes to fulfill their many roles in the organism, more than 100 different proteins are post-translationally imported into the peroxisomal membrane and matrix, and their functionality must be closely monitored.

In this Review, we briefly discuss the import of peroxisomal membrane proteins, and we emphasize an updated view of both classical and alternative peroxisomal matrix protein import pathways. We highlight different quality control pathways that ensure the degradation of dysfunctional peroxisomal proteins. Finally, we compare peroxisomal matrix protein import with other systems that transport folded proteins across membranes, in particular the twinarginine translocation (Tat) system and the nuclear pore.

Link | PDF (Journal of Cell Science)
 
I'm revisiting some metabolomics papers and thought I'd share some thoughts about this one.

I think it's potentially quite interesting. It's got a decently large cohort especially by ME research standards 101 pwME, 91 controls, seem fairly well matched in most areas, recruited from different sites, CCC criteria.

Statistics

They use fairly complicated statistics in this paper to determine significant metabolites. They use a Bayesian linear regression model where they also account for other features like age, ethnicity and so on.

Hutan said:
View attachment 17694
This is for the whole ME/CFS sample versus all the controls.



This sort of sounds good, but it's quite a bit of fiddling with the data. For example, I wonder how much race/ethnicity or even geography actually influenced metabolite levels, and if the adjustment process might have happened to reduce variability and make results more significant?

The ME/CFS cohort had more acetominophen (paracetamol) than the controls; the ME/CFS women had more acyclovir (anti-viral medication) and alprazolam (Xanax - a sedative used to treat anxiety and panic disorders :() than the control women. That's a bit of a worry on two counts - the different drug profiles could be having all sorts of effects, and it's a damn shame some of the women either feel the need for the Xanax or get put on it. I can't see how many people were on the drugs - the log transformation makes it hard to know how important the difference between ME/CFS and controls actually is (the data might be in supplementary tables).
Click to expand...

I think their mixed effects linear model is a good thing as it allows you to see the variation in the data that is only coming from the difference between the ME and HC groups with the variation due to other features stripped away. I didn't see them report anywhere on whether there was an effect with age, ethnicity, size and so on. That would have been interesting to see. Of course you can't say for sure that they didn't try analyzing in different ways too to try and find the most compelling results.

Instead of simply comparing a p value they also have cut offs from a bayesian analysis of the data. Their criteria for significance is an FDR adjusted p < 0.15, a Bayes factor > 3 and 95% HDI not overlapping with 0. It's quite a complicated way to delineate interesting metabolites.

What the Bayes factor effectively means is 'how many times better does the alternative hypothesis fit the data than the null hypothesis' after a Bayesian analysis. A Bayes factor of 3 is generally interpreted as providing moderate evidence for your alternative hypothesis (that ME and HC have different levels of a given metabolite). The HDI is basically saying that there is a 95% chance that there is a difference between the groups (I've oversimplified a little bit there).

I'm not experienced enough in Bayesian statistics to know whether using all three of these metrics is really more powerful than just using a p-value, or whether practically it's basically saying the same thing 2 or 3 different ways. The cynic in me would say they've used an adjusted p-value cut off of 0.15 instead of 0.05 because they barely saw anything at the 0.05 thereshold. Then again, you can see for example in these two metabolites that they have similar adjusted p values (second from the right) but vastly different Bayes factors (right most column):

upload_2024-3-3_11-53-4.png

(A bayes factor of 178 would be interpreted as extreme evidence in favour of a difference between HC and ME for this metabolite).
 
Last edited:
Biology

Hoopoe said:
I clearly remember that Ron Davis some years ago said they were seeing high levels of lipids in cells in one of the experiments they did. Maybe in connection with cells being lighter?

Accumulation of lipids is something that can happen when the peroxisomes are not working well.
Click to expand...

DMissa said:
Interesting. The lipidomics I presented at IACFSME 2023 showed broadly elevated levels of lipids in lymphoblastoid cell lines generated from pwME.
Click to expand...

This sounds interesting. It's interesting that all the reductions in membrane lipids they find are unsaturated (containing double bonds rather than completely saturated with hydrogens). That is matched by an increase of unsaturated (but not saturated) triglycerides. Triglycerides are the storage form of fatty acids and are stored in lipid droplets that are visible inside of the cell. So if as the authors argue there is a peroxisomal problem, and that affects synthesis of membrane lipids containing unsaturated fatty acids, maybe that would follow that you would have an accumulation of unsaturated triglycerides and lipid droplets.

Comparison to Naviaux's ME/CFS metabolites paper

As the authors say they replicate a lot of the same findings found in pwME from Naviaux et al 2016. The strongest signal in that paper was a reduction in the levels of various membrane lipids in particular sphingolipids (ceramides) and phospholipids. They don't however replicate Naviaux's findings of lower purine metabolites (uric acid, xanthine), or glutamate related metabolites (arginine, pyrroline 5-carboxylate). I think these findings are also potentially interesting, especially the glutamate ones (for some other reasons from other papers).

Midnattsol said:
A recurring problem is that the different metabolomics studies look at different lipids, so it is difficult to say. Beyond phospholipids, there are also the sphingolipids that can influence cell membranes and cell signalling (of these there are a few ceramides that have garnered some attention as they keep coming up in regards to metabolic syndrome, cardiovascular disease and diabetes. Then again, the reason they keep coming up is also because they are actively included in studies based on previous studies, while other ceramides or lipids are not included).

Also, lipids are easily influenced by diet, which these studies rarely control for adequately.
Click to expand...

I agree that diet might be the most obvious source of variance. In this paper the participants are sampled after overnight fasting. The Naviaux paper only required a 3 hour fast before sampling. I also think that differing levels of exercise might affect the results. The patients may have had to travel to be sampled which might have involved exertion.

Another possibility is sample storage. Metabolites have vastly different concentrations and degradation rates, and they can also degrade into other metabolites. Glutamine can spontaneously deaminate to glutamate for example. So the way the samples have been stored and for how long matters.

Finally, they see see an increase in levels of alpha-ketoglutarate, the TCA cycle metabolite and 2-oxo acid version of glutamate. This is the opposite of what other metabolomics sources report. Wüst's team (Appelman et al 2024) recently reported a reduction in glutamate and alpha-ketoglutarate in the plasma from their muscle study for example. I wonder if changes in these molecules could be an exercise related effect.
 
Mij said:
I don't know if this means anything, but my alpha-ketoglutarate (urine) was at 1 (ref range 5-80) during 2 day post-exercise PEM.
Click to expand...

Had a look in hanson urine study and it is lower relative to controls (0.7 of controls) but not significant p~0.3. Armstrong has also done some urine stuff but doesn't seem to have reported it.
 
chillier said:
This sounds interesting. It's interesting that all the reductions in membrane lipids they find are unsaturated (containing double bonds rather than completely saturated with hydrogens). That is matched by an increase of unsaturated (but not saturated) triglycerides. Triglycerides are the storage form of fatty acids and are stored in lipid droplets that are visible inside of the cell. So if as the authors argue there is a peroxisomal problem, and that affects synthesis of membrane lipids containing unsaturated fatty acids, maybe that would follow that you would have an accumulation of unsaturated triglycerides and lipid droplets.
Click to expand...
Given the role of saturated fatty acids in membranes, such as their importance in lipid rafts, I think it's reasonable they will be more preserved in this area.

chillier said:
I agree that diet might be the most obvious source of variance. In this paper the participants are sampled after overnight fasting. The Naviaux paper only required a 3 hour fast before sampling. I also think that differing levels of exercise might affect the results. The patients may have had to travel to be sampled which might have involved exertion.
Click to expand...
Due to the storage property of fatty acids, both the short and long term diet can influence fatty acids composition in the blood as the fatty acids are released into the bloodstream when the body needs energy. Not suggesting you're saying standardising the participants' fasting time and exertion levels would solve matters, just writing it out since not everyone might be aware.
 
This paper has been extended by Che, Lipkin and Fiehn here:

https://www.s4me.info/threads/bayes...ydges-che-lipkin-and-fiehn.27730/#post-496038

They bundle the results of their paper with Nagy-Szakal's paper and Naviaux's paper in a sort of Bayesian meta-analysis and the peroxisome pops out even more.

fulltext: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10535181/

We also found that only unsaturated triacylglycerides were upregulated in ME/CFS patients, but not saturated triacylglycerides. Such a finding points to a specific biochemical mechanism instead of simple explanations such as differences in the number or type of lipid-carrying lipoprotein particles. A further indication for this mechanism was a profound downregulation of unsaturated phospholipid ethers and plasmalogens (Figure 6 and Figure 7), which are exclusively produced by peroxisomes and which, hence, might be involved in the etiology of the disease [17]. Importantly, the results of the Bayesian analysis further strengthened the biological interpretation of peroxisome damage as an important factor underlying ME/CFS: very long chain polyunsaturated triacylglycerides were found at increased levels in ME/CFS subjects (Figure 6), pointing to a lack of oxidation in peroxisomes that exclusively perform this reaction (not mitochondria).

Lipids and membranes could be the bridge to the Hwang paper on endoplasmic reticulum stress and wasf3. Could the ER be perpetually stressed because a lipid imbalance in the membrane means it runs at reduced efficiency?
 
Peroxisomes help make plasmalogens. A plasmalogen shortage is one of the findings by Lipkin. Also one of the theories in this hypothesis paper from Canada: https://www.s4me.info/threads/chron...sis-addressing-the-overlap-2023-chaves.34083/

Plasmalogens are an intriguing idea to pursue because plasmalogen replacement therapy is established and it's a non-toxic supplement so collecting some anecdotal data should be straightforward.

I regret that the supplement itself has an animal-linked name that makes it sound suspicious rather than a scientific name that makes it sound serious, nevertheless a team of scientists from a good university have found it effective in raising blood levels of plasmalogens:

Shark liver oil supplementation enriches endogenous plasmalogens and reduces markers of dyslipidemia and inflammation

https://pubmed.ncbi.nlm.nih.gov/34146594/
 
Last edited:
You must log in or register to reply here.
Back
Top Bottom