Cellular bioenergetics is impaired in patients with chronic fatigue syndrome

Cara Tomas, Audrey Brown, Victoria Strassheim, Joanna Elson, Julia Newton, Philip Manning
Plos One 2017 oct


Abstract
Chronic fatigue syndrome (CFS) is a highly debilitating disease of unknown aetiology. Abnormalities in bioenergetic function have been cited as one possible cause for CFS. Preliminary studies were performed to investigate cellular bioenergetic abnormalities in CFS patients.

A series of assays were conducted using peripheral blood mononuclear cells (PBMCs) from CFS patients and healthy controls. These experiments investigated cellular patterns in oxidative phosphorylation (OXPHOS) and glycolysis. Results showed consistently lower measures of OXPHOS parameters in PBMCs taken from CFS patients compared with healthy controls.

Seven key parameters of OXPHOS were calculated: basal respiration, ATP production, proton leak, maximal respiration, reserve capacity, non-mitochondrial respiration, and coupling efficiency. While many of the parameters differed between the CFS and control cohorts, maximal respiration was determined to be the key parameter in mitochondrial function to differ between CFS and control PBMCs due to the consistency of its impairment in CFS patients found throughout the study (p≤0.003).

The lower maximal respiration in CFS PBMCs suggests that when the cells experience physiological stress they are less able to elevate their respiration rate to compensate for the increase in stress and are unable to fulfil cellular energy demands. The metabolic differences discovered highlight the inability of CFS patient PBMCs to fulfil cellular energetic demands both under basal conditions and when mitochondria are stressed during periods of high metabolic demand.

http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0186802

Edit: paragraph breakings mine

 

@Cheshire, thanks for posting, but would you mind taking that out of quotes? It's very hard to read as a block of small italics. (Don't know if you can break the paras up a bit, if that's not taking an inch and asking for a mile. )

Edit: Perhaps you could copy @Luther Blissett's kind formatting?

 

I'm just flicking through the pictures and one or two measures look interesting.

@Jonathan Edwards I wonder if there is something interesting here for the biobank in that they are plotting fresh and frozen samples separately?

 

I will need to read this paper a few more times before I say anything, but this test might be used as a possible confirmatory test for CFS, and more likely for ME. Its a research test though, so I am not sure if its going to be useful clinically. However it is in agreement with the CPET data, and may in fact be the long awaited alternative to CPET.

What I have been saying lately about CPET is that it cannot be used to diagnose ME, but it can be used to exclude an ME diagnosis except possibly in mild cases. I think this is the same. Normal mitochondrial flux under the conditions tested will probably mean the person does not have ME. The issue is not sensitivity, its specificity, again.

 

From what I understand, the research was done on PBMCs, which are certain types of blood cells. Basically it would include B cells, T cells, NK cells, and monocytes. So those are all cells involved in the immune system. The cells of the ME/CFS patients didn't respond nearly as well to stress as the control cells, and weren't helped by glucose as much.

The methodology looks pretty good, as far as I can tell. Corrections were made for making multiple comparisons. Patients met Fukuda criteria, which could use some improvement but also could've been worse. Sample sizes were pretty good, with 52 patients and 35 controls, though not particularly well-matched, probably due to recruiting controls through a university (mean age 6 years young and a bit too male in comparison). But they were able to eliminate age and sex as possible confounding factors in the results.

Their core hypothesis seems to be that ME/CFS is a hypometabolic disorder resulting from an acquired mitochondrial dysfunction:

In the discussion, they basically confirm our perception that our cells at least are already operating at their maximum capacity at baseline, and accordingly can't react normally when there's a need to temporarily compensate for higher demands:

They also looked at the cells under low-glucose conditions, where they'd be forced to engage in energy production via alternate means. The control cells were able to compensate, but not the ME/CFS cells, which may reflect our problems with crossing the aerobic threshold:

 

I suspect this might be a confirmatory test. If its not confirmed, you probably don't have ME. You might have a particular CFS subtype, that is a separate issue. Its not about diagnosing ME, its about removing non ME cases.

One of the huge difficulties we face is that we can reliably get good test sensitivity now, but not specificity. It takes a lot more research to assess specificity, and maybe even a mechanism is required.

If you focus on fresh samples, for basal respiration, maximal respiration, and reserve capacity, there is a clear separation between a cluster of four and the rest of the CFS patients, though this is less obvious with basal respiration. Most patients cluster together, and clearly are below the controls.

My interpretation of this is we are possibly seeing either two or more subtypes, or the same subtypes on a good day. If you remove the clear cluster at the top then the very close cluster below is totally separable. This needs more investigation. It has no clinical use right now, but is very much a valid research question.

It would be nice to see a study looking at both PBMC results and biopsy results, such as muscle. The NIH appears to be doing this. If the two correlate we can then dispense with biopsy results and rely on blood.

Even better is if we can find a good correlation between PBMC results and repeat CPET testing.

 

Where's all the fanfare from the SMC for this?

 

Their deafening silence is an admission of guilt.

 

My first post here

Thoughts on: Cellular bioenergetics is impaired in patients with chronic fatigue syndrome​

My notes from reading the full paper. This looks like an important contribution to the growing field of research probing problems in producing energy in mecfs. These preliminary findings need replicating and showing that they are unique to mecfs. Some points have already made by others.

The big story
These results are impressive. Even with the very small control sample size (12), these are "strong data, robust to statistical correction". There's remarkable separation between patients and controls, with almost all patients below the lowest control: you don't see that too often.


(my adaption of figure 5D, maximal respiration in low glucose)

Corrected p values for the key measures, including basal respiration, ATP production, maximal respiration and reserve capacity, are <0.005.

So these findings suggest that (white blood) cells from mecfs patients struggle to produce energy as healthy cells can, and in particular can't ramp up energy production to meet increased demand. Which intuitively makes a lot of sense.

[Summary of the "Seahorse" techonology behind this study, below]​

Significant issues with this study (most are discussed by the authors)

Small sample: "The primary limitation of this study was sample size which resulted in underpowered analysis", the authors plan to seek larger samples in future. There are only 3 fresh control samples so these particular findings simply aren’t reliable.

Questionable controls: Perhaps the biggest issue is the lack of activity-matched controls and the lack of sick controls. @alex3619 mentions specificity, the ability to separate mecfs patients from others that are sick. This is crucial both in clinical practice and for research hoping to find out what's gone wrong in mecfs. That said, it's a widespread problem in the field. (A good time to mention the UK mecfs biobank that has both large sample size and sick (MS) as well as healthy controls?)

The authors say sedentary controls would be better. Activity could plausibly affect mitochondrial performance and the authors are considering measuring activity levels in future work.

Need to validate methodology: the authors note the need to validate these findings using other well-established measures of mitochondrial performance such as morphology and membrane potential.

Results almost too good? What makes this study so impressive is that the results are so striking, which also makes me nervous, particularly for a study using Fukuda criteria that are unlikely to identify a tightly-defined cohort. So I'd have expected a bit more variation in results unless, say, inactivity is a major factor driving these results. Maybe I'm being unfair.

Some wider questions about probing energy issues in mecfs

Which types of cells are affected? Some people have already commented that work this is on PBMCs (white blood cells) and the results could be down to one sub-group such as T-cells (I'd be surprised, given how big the overall effect is). The authors did note this, and it's true of any PBMC study - which includes much mecfs work.

Conflicting findings - what's really going on?
This study found that glycolysis is normal in mecfs, while oxidative phosphorylation is the issue, while others have found glycolysis is the issue e.g. Armstrong 2015 metabolomics work, the Fluge/Mella amino acid metabolism study and the unpublished Ron Davis metabolomics work. I think earlier work by Julia NEwton's group has suggested a problem in glycolysis, or at least in pyruvate dehydrogenase at the end of glycolysis.

I gather Fluge/Mella found opposite mitochondrial findings using Seahorse (does anyone have details?). @viggster said these findings are broadly in line with very early results with Seahorse in the NIH intramural study.

There’s also the issue of using serum or not, the fluid that bathes the cells themselves. Potentially it carries “factors" that could be causing the problems, even if the cells themselves are normal. So, I think the Fluge results on Seahorse were with serum, and the Ron Davies results too. These new results are serum-free. However, it would be easy enough to test with serum using this tech, and see what happens if you swapped over healthy and patient serum.

Meanwhile, the CPET 2-day exercise studies often find normal day-one results, which doesn't fit well with these results.

Which looks like a mess in places, or simply the signs of an early and promising field that needs more work (and more funding). I'm going with the latter for now and hope the researchers keep going with bigger and better studies. (Don't forget there are all those ready-to-go UK MEcfs biobank samples for large studies with sick as well as well as healthy controls.)

 

@Simon M Hi Simon - and what a post it is!

 

One thing I've been looking at recently (with help!) using very different data is to assess whether a particular feature is any good. That is does the data suggest a feature is IID (independently identically distributed). It feels like such tests would be good on this data as I think it would help determine whether say the control group seems to have a common pattern within it. If the data appears genuinely random then it feels like there is a greater chance of random differences.

Having said that the clusters seem quite tight with the energy production graph which gives a good feel with the separations.

 

I think one of the big issues that may come up in this type of testing is the dynamic nature of the disease. So after rest I wonder if effects are not there until after some level of exertion. However, going to give blood can be quite an exertion depending where people have to go, how much help they have in getting there, and how far the car park is away. I wish they would measure activity measures prior to take samples and look at whether this is a factor in results.

One of the good thing about the bio bank is that they are sending people out to take sample from house bound patients (I think).

 

I suspect part of the problem here is that the sample for controls (N= 12) is so small, too small really. I think statistical theory says you need at least this number, probably a few more to have a chance of a representative sample. I think you would know better than me.

Thanks! It will be all downhill from here.

 

Yes that doesn't seem enough to really tell.

I do feel someone should do a massive replication project to look at all the interesting results with many more people and controls to see what is reliable.

 

Thanks for your analysis Simon, I was looking forward to it.

Also, welcome to the forum!

I share your concern over this issue. On the one hand I do think we've started focusing on the right areas of research so I would expect to see results that contrast quite strongly with healthy controls. On the other hand, because of the limitations of diagnostic criteria you might expect some overlap instead. It could be due to inactivity as you suggest. Or just a common picture in various fatigue states.

 

The technology behind the new study finding (above) that cells from mecfs patients can't ramp up energy production​

Here's a bit more about the Seahorse technology and what it measures, using info from the Seahorse site. First off, it's looking at the body's two main energy pathways, glycolysis and respiration, which you can see in these diagrams (from Seahorse videos).

First 55 seconds of this video is good on the underlying biology (sorry, these videos won't embed here).

• Glycolysis produces modest amounts of ATP, the cells' main fuel as well as pyruvate, the molecule that feeds mitochondria. Glycolysis produces protons (H+ or acid, effectively) that acidify the fluid around cells.
  
• Oxidative phosphorylation (or respiration) is the process that takes place in mitochondria, the real fuel factories that burn fuel efficiently, consuming oxygen and foodstuffs and pumping out ATP.

Seahorse measures glycolysis through the protons (acid) and oxidative phosphorylation through the drop in oxygen levels. Its secret is tiny probes that are lowered to trap cells in a micro-chamber of a droplet and it measures activity only in this micro-chamber.



But this micro-chamber is only a tiny amount of the fluid in each compartment. After each measurement, the probes are raised, everything mixes together and it can then take a new measurement. This ability to take repeated measures over time is key.

This study measures a bunch of different aspects of oxidative phosphorylation, which it does by injecting different compounds to probe what's really going on. Think of it as an exercise test for mitochondria, revealing what it's doing and how much capacity it has to ramp up activity. The central measures in the study are:

1. Basal respiration, the tickover rate of activity
2. ATP production, which is simply basal respiration less a bit of inefficiency loss (proton leak)
3. Maximal respiration: what the mitochondria can theoretically do using a drug to put the foot to the floor.
4. Reserve capacity: simply maximal respiration less basal respiration (peak capacity less what it's doing, or headroom to increase activity).

Fig 1 from the paper shows an example of how this testing works but I think my version is a bit simpler (the names at the top are the compounds added to put mitochondria through their paces).


The study looked at glycolysis too, which uses a slightly different detailed methodology, but that found nothing of interest, so I will spare you the details.

 

The authors specifically said that was something they were considering.

 

I've not had a chance to read the paper yet but will hopefully this week. It looks very interesting as does the seahorse testing thing.

 

Maureen Hanson talked about her Project "cellular metabolism of immune cells in ME/CFS" at the Discovery Fórum :

Data from T cells (immune regulatory cells) isolated from 20 ME/CFS patients and 20 controls indicate that T cells in the former group are using significantly less of their respiratory capacity. The group is looking for additional quantities of T cells to further explore whether T cell subtypes can be differentiated.

I think this is the Project :
http://neuroimmune.cornell.edu/research/metabolism/

Ohhh my first post here

 

Nice post

Do you know if she also presented data for B-cells and NK cells? What’s really interesting about this work is that her group are using Seahorse technology too, so should be directly comparable.

The finding of using less of maximal respiratory capacity is potentially slightly different. Here they found that maximal respiratory capacity was lower in cells from patients; i’m not sure they found patients were using less of their capacity, as Hanson reports. But others, above, have wondered if T-sells specifically with the issue, possibly that’s exactly what Hanson has found.