Elevated ATG13 in serum of pwME stimulates oxidative stress response in microglial cells , 2022, Gottschalk et al

[Hutan]

On the topic of whether it is normal or good/bad to have autophagy proteins in serum, the paper says:

One report found that plasma ATG5 level was elevated in Alzheimer’s disease (AD) patients (Cho et al., 2019b), while another observed lower ATG5 and mitophagy in serum from patients exhibiting mild cognitive impairment, vascular dementia, and Alzheimer’s disease (Castellazzi et al., 2019). Beclin-1 and LC3 are observed in the serum in patients with acute ischemic stroke (Li et al., 2015). Levels of autophagic markers might be associated with coronary total occlusion and childhood cerebral palsy (Demircan et al., 2018; Xu et al., 2017). Upregulation of α-syn was demonstrated in plasma of PD patients (Bougea et al., 2019).However, it is not known if the levels of these autophagy markers are also altered in ME/CFS.

There's no mention of ATG13 there. For ATG5, a study is referenced that found higher levels in Alzheimers, as well as a study that found lower levels of the same protein, also in Alzheimers. The paper on cerebral palsy found lower levels of ATG5 in cerebral palsy patients:

Furthermore, mean plasma ATG5 levels were lower in CP patients than in controls, and individuals carrying the AA genotype of rs6568431 that was positively associated with CP had lower plasma ATG5 levels (P < 0.05). This study demonstrated an association of an ATG5 gene variant and low level of ATG5 protein with CP, and stronger associations with severe clinical manifestations were identified.

So, clearly we can't proceed with a certain idea of 'having any autophagy proteins in your serum, or higher levels than average, is definitely bad'.

Another complication is that there are different types of each autophagy protein. Aside from whether they are phosphorylated or not, there are different genetic variants e.g.

Several GWASs for systemic lupus erythematosus (SLE) confirmed genetic associations between common variants in/near ATG5 and SLE, in Caucasians and Asians.

source

And for ATG13, the gene might be constant, but there are something called 'splice-variants' producing at least 5 different isoforms.

ATG13: Just a companion, or an executor of the autophagic program? 2014
Although vertebrate genomes have only one ATG13 gene, the alternative splicing of the ATG13 pre-mRNA results in the generation of several ATG13 splice variants. According to Jung et al. and the UniProt database, there is evidence for at least 5 different human ATG13 isoforms at the protein level (ref. 30 and UniProt entry O75143). The longest human isoform 1 (O75143-1) comprises 517 amino acids, while the shorter isoforms 2, 4, and 5 lack various exons in the middle and N-terminal region of ATG13. Interestingly, human isoform 3 (O75143-3) additionally lacks the last 75 C-terminal amino acids. Its mRNA instead makes up a part of the 3′ UTR and thus encodes for an alternative 15 amino acid sequence at the C terminus that completely differs from the other isoforms (Fig. 1B).30 Notably, ATG13 isoform 3 misses both the LIR motif, which is responsible for GABARAP interaction,32 as well as the ULK1 binding site. Isoform 3 is hence unable to interact with ULK1.30

I don't know if different isoforms can exist in a single human, but it raises the possibility of ATG13 having different forms that perform different functions, some possibly outside the cells. So, potentially, to understand what ATG13 might be doing in ME/CFS, if anything, we might need to know about the isoforms.

I'm wading out of my depth on this. But, potential complications to keep in mind.

 

[Hutan]

Back to the paper - Section 3.4 Evaluation of lysosomal function

So, they used a protease array to test for lysosomal proteins in the serum. The particular array covered, matrix metalloproteinases, cathepsins, kallikreins (a cool name, a bit like Kali, the goddess of destruction and change but also love - it occurs to me that a cell biologist could call their daughter Kallikrein; Cathepsin works too, for that matter. I'm heading off-topic, must be time for lunch.)

Anyway, this is short.
They found no difference in the 2 case control pairs with respect to lysosomal proteins. It isn't explained how finding lysosomal proteins in the serum, or not, tells us anything about how the lysosomes are functioning - because surely the lysosomal proteins need to be in the lysosomes, inside the cells, to work, or not work.

They did find a couple of significant differences between the male case control pair for non-lysosomal proteins that happened to be in the test array (MMP-7 and proteinase-3). But they didn't find that in the female case control pair, and they didn't do this analysis on a larger sample.

So, not much we can conclude from this part of the study I think, about lysosome function or much else, except that some proteinases have good names.

 

[Hutan]

3.5 The effect of ATG13 in serum on microglia - ROS
No one could accuse these researchers of not covering enough ground in a paper. I think this finding alone would have been worth a paper; I hope the finding is not buried under all the rest.

So, in this study, they put human microglia cells (microglia being the immune cells in the central nervous system) into media (I'm imagining growing media like the agar jelly in a Petri dish) made with ME/CFS serum or control serum. They then measured reactive oxygen species using a fluorescent marker from 30 mins to 2 hours.

The results look convincing, but there was a very small sample size.
Figure 4b - results from one male case-control pair.



Figure 4D - left a female control, right the ME/CFS case



They then neutralised the ATG13 in the serum of the two case-control pairs, and repeated the experiment, finding much less indication of ROS. They went on to repeat the experiment with 6 more case-control pairs (we don't know how they chose them) - results in Figure 4H. The result seems pretty clear, although of course, still a small sample size - serum from ME/CFS cases caused more fluorescence indicative of ROS, but the fluorescence reduced to levels seen in the controls when atg13 was neutralised.



The authors rightly ask if ATG13 inside the microglia might be having the same effect - i.e. creating reactive oxygen species.

Interestingly, an atg13 siRNA (Cat # 122699; human; ThermoFisher) that mediates knock - down of atg13 gene expression (Fig.4I) evoked significant ROS production in microglial cells (Fig. 4J), suggesting that the intracellular ATG13 actually mitigates microglial ROS production. Similar siRNA knock-down of other autophagy -related proteins such as ATG5 (PMID: 34336554) and Beclin (24452380) also induced ROS productions in microglia.

That sentence takes some unpicking. siRNA are small interfering RNA! So, ATG13 siRNA prevents atg13 gene expression, but itself causes ROS production, as do other siRNA. I'm not sure that we can necessarily conclude from that that ATG13 inside the cell mitigates microglial ROS production. I'm going to park that one.

 

[Hutan]

3.5 The effect of ATG13 in serum on microglia - NO
They found something similar with nitric oxide. It looks like they found that nitric oxide produced by microglial cells was increased by exposure to the serum from the initial male case control pair and 8 other case control pairs (I don't know what happened to the initial female case control pair, or the others.) For the initial male case control pair, they applied a range of dilutions and found that NO production was increased with increasing concentrations in a dose-dependent manner. Serum that had had the atg13 neutralised didn't have that effect.

5A - production of nitrite (?) by microglia treated with the serum of initial male case-control pair, at various concentrations
5B - production of nitrite by microglia treated with the serum from heathy controls, me/cfs patients, me/cfs patients with the atg13 neutralised



Again, that's a remarkable result.

Upregulation of NO could be due to the direct stimulation of iNOS enzyme in HMC3 microglial cells. To evaluate that, Microglial cells were stimulated with 25 μM of L-NIL, a pharmacological inhibitor of iNOS, followed by the stimulation with 1:2 dilution of ME serum. Interestingly, ME serum was found to be ineffective towards NO production once iNOS is inactivated in microglia

Wikipedia on the iNOS enzyme:

Nitric oxide synthases (EC 1.14.13.39) (NOSs) are a family of enzymes catalyzing the production of nitric oxide (NO) from L-arginine. NO is an important cellular signaling molecule. It helps modulate vascular tone, insulin secretion, airway tone, and peristalsis, and is involved in angiogenesis and neural development. It may function as a retrograde neurotransmitter. Nitric oxide is mediated in mammals by the calcium-calmodulin controlled isoenzymes eNOS (endothelial NOS) and nNOS (neuronal NOS). The inducible isoform, iNOS, involved in immune response, binds calmodulin at physiologically relevant concentrations, and produces NO as an immune defense mechanism, as NO is a free radical with an unpaired electron. It is the proximate cause of septic shock and may function in autoimmune disease.

This therefore seems to be a big deal if it holds up - ME/CFS serum causing microglia, and perhaps other immune cells to produce (ROS and) nitric oxide. Production of NO, with its signalling role, could really set the cat among the pigeons.

It will be interesting to look at the findings we have about NO from other papers, to see if the idea is supported.

 

[Hutan]

3.6 How might the ATG13 in the serum make microglia produce ROS and Nitric oxide?
Off-the shelf ATG13 (i.e. standard stuff, purchased, not from ME/CFS patients) reacts with a type of receptor called RAGE on the microglia to produce ROS and NO. RAGE= Receptor for Advanced Glycation of End Products. The authors found that ATG13 bound to biotinylated* RAGE in a dose-dependent manner (i.e. the more ATG13, the more RAGE binding). Apparently, it is known that 'activation of RAGE', presumably anything binding to that receptor, produces ROS and NO. (I wonder what normally binds to RAGE - edit - advanced glycated end products?)

*(What is biotinylated? Wikipedia says we can probably just ignore it - "the added biotin is unlikely to disturb the natural function of the molecule". I think it might be part of the protein production process.)

So, then the authors worked with the ME/CFS sera. This statement was unhelpful:

Surprisingly, three different ME/CFS patient serums displayed a strong binding affinity towards biotinylated- RAGE protein. ME/CFS serums were serially diluted up to 10-fold (Fig. 6B) and added onto ATG13 antibody-coated plates. Strong binding was observed at 0- to 2- fold dilution (Fig. 6C). However, the binding is markedly inhibited with ME∆ATG13 serum samples (n=3)

Three different ME/CFS serums displayed strong binding affinity - but how many serums did they try? 12, or just the three? Figure 6B suggests that they just tried 3 samples.



They found that when ME serum had the ATG13 neutralised, there wasn't binding.



"An in-silico homology modeling study" - that's computerised modelling of the structures. They found that it is structurally possible for the ATG13 to be binding to the RAGE.

An immunoblot analysis (yeah, I don't know) and "probing with a pan-phosphoserine antibody" (yeah...) led the authors to conclude that at least some of the ATG13 in the ME/CFS serum is phosphorylated (unlike the control serum it is implied). And the computer modelling of the structures suggested that phosphorylated ATG13 could bind well to RAGE.

Next, (and you have to give the authors credit, they have chased down leads), they tried to neutralise RAGE, to see if ME/CFS serum still caused microglia to produce ROS and NO. RAGE neutralising antibody is fortunately something that you can buy.

7b - binding of ATG13 decreased as the concentration of RAGE neutralising antibody increased.

Blocking RAGE in microglia (HMC3∆RAGE) significantly attenuated the ME/CFS serum-induced ROS production (Fig. 7C), which we confirmed in a time-sensitive study

(Including the quote because 'blocking RAGE in microglia' is quite a vivid image, one that might make the psychosomatic people feel they have something to work with.) But anyway, the authors report that neutralising RAGE resulted in much less ROS production when microglia were exposed to ME/CFS serum and no induction of iNOS expression.

7c - the caption in the paper isn't clear at all. The label says it's a ROS assay
top left is microglia +control serum -> not much ROS
top right is microglia + ME serum - > lots of ROS
bottom left is microglia + neutralised RAGE +ME serum -> not much ROS
bottom right is microglia + IgG + ME serum -> a lot of ROS

I really don't know what is going on in this image. There's no mention of an IgG treatment - I'm not sure if that is the same as the RAGE neutralising antibody. There needs to be more explanation somewhere, including whether this was just for one ME/CFS person's serum. (Edit - SNT Gatchaman's suggestion below that the IgG was a control, expected to be inactive, and therefore have no moderating effect on the ROS-inducing effects of ME/CFS serum, makes sense.)



7d - also isn't clear. There are images labelled with CD11 - that's Integrin aM - I don't think the text explains what that is doing there. Here are the images for iNOS production, with, from left to right, control serum, ME/CFS serum, and ME/CFS serum and neutralised RAGE.

 

[Hutan]

I haven't gone through the discussion, but will stop here.

There's certainly very interesting stuff, but I get the feeling that, as the study and paper writing went on, everything just got more hurried and less organised. Basic things like how many samples there were for each analysis, and how they were chosen have been left out. Captions are muddled and very unclear. I'd suggest 'less is more' in terms of a paper's scope.

Still, good on the researchers for what looks to have been a lot of work. edit - and, I should say, possibly, a lot of important work.

Please do correct my misunderstandings.

It will be interesting to poke into this more, and see what eventually comes of it. Top of the S4ME list of threads tagged with nitric oxide is
Decreased NO production in endothelial cells exposed to plasma from ME/CFS patients, Bertinat et al (2022)
That's decreased, not increased, as in this study. Which is a little deflating. The authors mention that study in the discussion. Maybe microglia are different, or something.

 

[SNT Gatchaman]

I really don't know what is going on in this image. There's no mention of an IgG treatment - I'm not sure if that is teh same as the RAGE neutralising antibody. There needs to be more explanation somewhere, including whether this was just for one ME/CFS serum.

As I read it —

(HMC3 is "human microglial clone 3 cell line")

Top left image is HMC3 + HC serum
Top right image is HMC3 + ME serum
Bottom left image is HMC3+RAGE neutralising antibody + ME serum
Bottom right image has HMC3+(some other - not RAGE neutralising) IgG + ME serum

So I think the bottom right image is a control, just proving that the IgG has to be specific anti-RAGE, not just any old IgG. As you also said they seem to have dotted all their i's and crossed their t's.

 

[SNT Gatchaman]

It will be interesting to poke into this more, and see what eventually comes of it. Top of the S4ME list of threads tagged with nitric oxide is
Decreased NO production in endothelial cells exposed to plasma from ME/CFS patients, Bertinat et al (2022)
That's decreased, not increased, as in this study. Which is a little deflating. The authors mention that study in the discussion. Maybe microglia are different, or something.

Yes the authors talk about a "double-edged sword". Apart from different effects with intracellular and extracellular ATG13, I think this is principally a reference to beneficial and deleterious effects of NO, as they state that increased NO in microglia would lead to increased neuroinflammation and neurodegeneration (uh-oh). On the upside, NO from eNOS helps vascular control (notably defective in ME/LC).

I think NO acts quite locally, e.g. inhaled NO in ventilated ICU patients only really affects the pulmonary vasculature, being rapidly scavenged by Hb in red cells and inactivated. I imagine this explains why NO could have quite different signalling effects with different tissues, with less likelihood for cross-talk.

So, in their study (phosphorylated) extracellular ATG13 upregulates iNOS (inducible isoform, in immune cells) -> increased NO in microglial cells.
While, the umbilical vein study showed ME serum downregulates eNOS (endothelial cell isoform) -> decreased NO.

Sounds like the next step would be to combine findings/techniques and see if RAGE-blockade of umbilical endothelial cells prevented the observed decrease in NO.

 

[Midnattsol]

And for ATG13, the gene might be constant, but there are something called 'splice-variants' producing at least 5 different isoforms.

I don't know if different isoforms can exist in a single human, but it raises the possibility of ATG13 having different forms that perform different functions, some possibly outside the cells. So, potentially, to understand what ATG13 might be doing in ME/CFS, if anything, we might need to know about the isoforms.

Different splice variants are found in single humans, and splice variants can also vary by tissue type.

 

[CRG]

I found this study helpful in understanding what seems to be the territory being explored by the Gottschalk paper:

Altered serum levels of autophagy proteins Beclin-1 and mTOR in patients with exudative age-related macular degeneration

although I'm unsure how it relates directly to Gottschalk.

Serum level of BECLIN-1 are reported:

"In patients with AMD, serum levels of Beclin-1 were significantly lower than in controls (median; interquartile range (IQR): 0.100 ng/ml; 0.100 – 0.100 versus 1.123 ng/ml; 0.100 – 3.452, p = 0.0033), while mTOR serum levels showed no significant difference between the groups (median; IQR: 4.377 ng/ml; 1.107 – 7.518 versus 3.608 ng/ml; 1.602 – 6.357, p = 0.4522) (Fig. 1). In the post-hoc analysis, following stratification by age of AMD patients and controls together, participants older than 70 years had lower Beclin-1 levels than those aged 70 years or less (p = 0.0444). This difference was mostly noted in patients with AMD. Moreover, in the AMD group, 80% of patients older than 70 years demonstrated lower serum Beclin-1 levels, as compared with only 52.6% of patients with cataract of the same age (p = 0.0024) (Fig. 2a)."

 

[Hutan]

I'm not sure that the Gottschalk paper particularly joined the dots between mitochondrial dysfunction and the ATG13 in the serum. (They did do a lot of dot joining from the serum to the microglia though.)

It's possible that the ATG13 is getting into the serum for a reason unrelated to autophagy. For example, we know that ATP is important in energy production in the cell, but when it's outside the cell, it's a signalling molecule to deal with threats. Maybe the ATG13 is like that - maybe there isn't really a problem with autophagy, I'm not sure the evidence for there being a problem with autophagy in ME/CFS is yet bullet-proof, even if it is attractive.

There's this - I think this sub complex formed by ATG13 and FIP200 to deal with certain viruses is inside the cell, but maybe the extracellular ATG13 with its ability to activate microglia has more to do with this anti-viral function.
HSBP1 Is a Novel Interactor of FIP200 and ATG13 That Promotes Autophagy Initiation and Picornavirus Replication, 2021

In particular, we examined the impact of the depletion of each component of the ATG proteome on the replication of 6 viruses from 6 different virus families. With this approach, we also identified an anti-viral role of the subcomplex formed by ATG13 and FIP200, outside the context of the ULK kinase complex (Mauthe et al., 2016). This anti-viral role is specific for picornaviruses, a large virus family of non-enveloped, small (~30 nm in diameter) viruses with a positive-stranded RNA genome, which cause diseases in humans and animals (Mauthe et al., 2016). Picornaviruses are classified into different genera based on a complex set of rules (ICTV taxonomy) (Zell et al., 2017). Examples of genera are enteroviruses, which include coxsackievirus B3 (CVB3) and enterovirus 71 (EV71), and cardioviruses, a member of which is the encephalomyocarditis virus (EMCV) (Zell et al., 2017).

 

[Robert 1973]

Some questions and answers on Twitter.

Edit to add: looking forward to reading some write ups of this paper as most of it is beyond me.

 

[Amw66]

I haven't read the thread here or Twitter as not a good day here, but hope to catch up later.

A thought - are all the subjects male ?

From previous metabolinics etc there is evidence of females being different re metabolism ( tending to use protein as fuel) which may subtly alter signalling and processes

 

[LarsSG]

There are different subgroups of patients used for different experiments, but overall they have 8 female patients and 3 male patients (plus one male patient with small fibre neuropathy but not ME).

 

[Amw66]

There are different subgroups of patients used for different experiments, but overall they have 8 female patients and 3 male patients (plus one male patient with small fibre neuropathy but not ME).

Thank you.

 

[Midnattsol]

The female metabolome might also be influenced by the menstrual cycle, just to make it even more complicated.

 

[LarsSG]

I asked Avik Roy about the protein aggregation and he does think it is related to the LC microclots findings.



This was a pretty strong finding, with a clear difference between the 7 patients and 7 controls in the rate of protein aggregation.

 

[Ravn]

He mentions further study in a mouse model.

Maybe this is related?

Simmaron researchers are working to create two ME/CFS mouse models: a transgenic and a chemical-induced model. As we’ll see simply the process of creating an ME/CFS mouse model has resulted in significant potential insights.

https://www.simmaronresearch.com/blog/seeing-mecfs-creating-the-first-mouse-models-for-the-disease

 

[LarsSG]

That's interesting. Looks like they are basically doing what they did in serum, as described in the paper, in mice now:

"They gave the mice a compound , an mTOR activator and autophagy inhibitor that inactivates a protein called ATG13 which activates two mytokines called IL-6 and RANTES. Although, IL-6 is released by the muscles during exercise and is believed to enhance energy production, together with RANTES it can cause molecular changes in muscle tissue causing fatigue.
[...]
Interestingly, the mice displayed a dramatic gender split with the female mice much more apt to display fatigue – a sign given the gender split in ME/CFS that they were on the right track. Electromyography (EMG) muscle tests revealed that bicep muscles in the mice given the autophagy inhibitor quickly became fatigued, and their grip strength declined. Likewise, a treadmill test indicated they were less active and displayed increased fatigue."

And it sounds like they will try to do the same genetically too and test some drugs on them.

Will be interesting to see their mouse data. It sounds like not all the mice are showing fatigue, so it doesn't sound like a home run, but maybe it is a piece of the puzzle.

 

[LarsSG]

Slides from Dr. Roy's presentation today at IACFS/ME about their work on a mouse model for ME with a drug that acts on mTOR -> ATG13 -> autophagy impairment.

It definitely does something that looks a bit like PEM, if I'm understanding the slides correctly, and much more effectively in female mice.

Did anyone catch the presentation?

[Edit: Thanks to @wigglethemouse for asking Dr. Roy to share these on Twitter (see here).]