Thesis Thesis: Investigating the Genetic and Immunological Aetiology of Myalgic Encephalomyelitis/Chronic Fatigue Syndrome 2022 Dibble

[Andy]

This thesis describes two investigations into the disease Myalgic Encephalomyelitis/Chronic Fatigue Syndrome (ME/CFS), specifically its genetic aetiology and immune system alterations. The first study investigated the genetic basis of ME/CFS using Genome-wide Association Studies (GWAS) by attempting to replicate and extend results previously found using UK Biobank cohort data. GWAS attempt to identify associations between DNA variants and phenotypes. This GWAS was novel, conducted on new phenotypes constructed by combining those in the most up-to-date UK Biobank data release. A new, previously unseen, genome-wide significant association was found on chromosome 6 for males with ME/CFS within the gene PDE10A. Further results were not genome-wide significant, but many were suggestive and hence independent replication may justify further research.

A previous analysis on the UK Biobank cohort had identified an indicative association in females between variants around the SLC25A15 gene at genome-wide significance. I adopted a hypothesis that the dietary protein intake of people with the CFS risk variants would be lower than those with the alternative alleles, due to potentially reduced production of mitochondrial ornithine transporter 1 (ORNT1). However, this association with dietary protein intake was not supported by UK Biobank data. Additionally, I investigated associations between the human leukocyte antigen (HLA) alleles and the ME/CFS phenotype using UK Biobank data. Associations between alleles within the HLA-C and -DQB1 genes had previously been found in a cohort of Norwegian people with ME/CFS, and my goal was to seek replication of these results in a larger dataset. None of the associations found in the UK Biobank proved to be genome-wide significant.

In my second study I investigated the use of T-cell clonal diversity as a potential biomarker for ME/CFS. This project used cells from CureME Biobank samples in collaboration with Systems Biology Laboratory (SBL). I developed a data analysis pipeline to analyse T-cell receptor (TCR) genomic DNA data based on the best practices currently used in the fields of immunology and mathematical biology. This approach used a mathematical notion of entropy as a measure for the diversity of TCR repertoires, in this way combining all of the most commonly used metrics in mathematical biology. When combined, these measures form a profile for each repertoire, a set of which can be sorted using a machine learning algorithm to partition the repertoires into subgroups.

My hypothesis was that the T-cell clonal expansion of people with ME/CFS would be greater than for healthy controls, and comparable to disease (multiple sclerosis) controls. Although this method was able to effectively classify TCR chains using simulated data, results from experimentally-derived data did not support the hypothesis, with the most effective classifications for both CD4+ and CD8+ cells failing to pass corrections for multiple hypothesis significance testing.

Open access, https://era.ed.ac.uk/bitstream/handle/1842/39763/DibbleJJ_2022.pdf

 

[Andy]

Joshua was first author of this publication, Genetic Risk Factors of ME/CFS: A Critical Review. Joshua J Dibble, Simon J McGrath, Chris P Ponting. 2020

 

[Simon M]

If I’ve understood right, using best practice mathematical and immunological methods, Joshua did find a single genetic difference that was significant in Uk Biobank:
for males with ME/CFS within the gene PDE10A.

He also helpfully ruled out a number of other findings. This is what our field desperately needs: replication attempts, even if results don’t replicate (here, GWAS and the original Mark Davis T cell Receptor findings).

Joshua is now a postdoctoral research at Harvard Medical School, and Massachusetts General Hospital
https://uk.linkedin.com/in/joshua-dibble-1114aa166 | LinkedIn

I believe he’s working as part of an ME/CFS group.

 

[Simon M]

Joshua was working in Chris Ponting’s group, and his PhD was jointly funded by Action for. ME and the office of the Scottish chief medical officer. Kudos to Joshua and all involved. (With a nod to the company that did a lot of sequencing of Tcell receptors genes, Systems Biology Laboratory,).

 

[ME/CFS Skeptic]

Looks like useful work. Hope he will continue to study ME/CFS.

 

[Hoopoe]

PDE10A is implicated in central nervous system diseases, such as Parkinsons and Huntingtons disease.

 

[Jonathan Edwards]

The abstract is soberingly impressive. I hope Joshua goes on to find the needle in the haystack.
In the meantime, this is useful work precisely because the findings were negative.

 

[CRG]

The abstract is soberingly impressive. I hope Joshua goes on to find the needle in the haystack.
In the meantime, this is useful work precisely because the findings were negative.

"this is useful work precisely because the findings were negative" - would be my (only slightly facetious) suggestion for the S4ME forum motto !

 

[Kitty]

Oh, you can get a postdoc fellowship at a prestigious institution by doing diligent science and publishing what you find, instead of over-hyped, under-evidenced flummery.

Should someone tell psychology?

(Seriously: congratulations, Joshua, and thank you, and good luck.)

 

[RedFox]

What I understand about this gene, in plain English:
Men with ME/CFS were more likely to have variations in the gene PDE10A, which regulates the transmission of certain signals inside cells. There's a lot of it in part of the brain called the corpus striatum, which helps coordinate movement, is an important part of the brain's reward system, and participates in many functions. When this gene gets really broken, it causes limb and orofacial dyskinesia (involuntary movements) and a disorder called striatal degeneration. From what little I can find online, striatial degeneration is a rare brain disorder that causes hyperkinesia (again, involuntary movements or exaggerated voluntary ones) starting in infancy.

None of this sounds remotely similar to ME. This gene can cause severe movement disorders, and the brain region where it's used the most also plays a role in mental disorders.

 

[Hubris]

The abstract is soberingly impressive. I hope Joshua goes on to find the needle in the haystack.
In the meantime, this is useful work precisely because the findings were negative.

I have seen you mentioning this often - if all genetic studies (including decodeME) turn out completely negative, what kinds of pathology would that suggest?

 

[SNT Gatchaman]

Although this sounds obvious, I wonder whether the positive finding may turn out to be much more useful than the negatives.

None of this sounds remotely similar to ME. This gene can cause severe movement disorders, and the brain region where it's used the most also plays a role in mental disorders.

PDE10A has broad expression.
Relevant PubMed search — https://pubmed.ncbi.nlm.nih.gov/?filter=hum_ani.humans&sort=date&linkname=gene_pubmed&from_uid=10846

From Role of PDE10A in vascular smooth muscle cell hyperplasia and pathological vascular remodelling (Sep 2022, Cardiovascular Research) —

PDE10A is the single member of PDE10 protein family, and it catalyzes both cAMP and cGMP hydrolysis. The literature on PDE10A has been focused on psychiatric and neurological disorders because PDE10A26 expression is enriched in the striatal region of the brain under normal conditions.

There are three plausible hypotheses for the aetiology of ME/CFS as a result of this research.
[1, 2 ...]
Third, that ME/CFS is an endothelial disease, mediated through (among others) PDE10A. This would be consistent with the blood-pressure associated variant found by Warren et al., orthostatic intolerance symptoms, and PEM symptoms induced by cardiovascular testing. PDE10A protein has been found in some endothelial tissues, and hence may be relevant to the proposed blood-pressure pathway. However, this hypothesis relies upon the connection (through [Linkage Disequilibrium]) between two not yet replicated results, one of which was seen in males only. Further research would need to confirm the connections between PDE10A and ME/CFS, and PDE10A with endothelial cell dysfunction.

See also PDE-Mediated Cyclic Nucleotide Compartmentation in Vascular Smooth Muscle Cells: From Basic to a Clinical Perspective (2021, J Cardiovasc Dev Dis) which is a comprehensive review of the various PDEs (phosphodiesterases).

The main function of vascular smooth muscle is to regulate vascular tone. In this process, the signalling cascades regulate contraction and relaxation in response to various hormonal and hemodynamic stimuli. Therefore, SMCs are the cells responsible for the contractile property of blood vessels, which are involved in the regulation of blood pressure and blood flow distribution.

The cyclic nucleotides cAMP and cGMP are essential modulators of vascular function.

There are links to be investigated here between vascular smooth muscle cell action, nitric oxide signalling and blood flow control with impaired flow-mediated dilation in ME/CFS. This finding in males warrants replication and I'd like to see more investigation down this particular vascular pathway. Perhaps the female-predominant patient cohorts may have obscured related findings in studies to date.

In addition, PDE10A function has been associated with blood pressure, which is consistent with the symptoms of ME/CFS that overlap with POTS (see Section 1.5.4). Warren, et al. report that the SNP rs147212971 is protective of elevated diastolic blood pressure, and this SNP appears to be in perfect linkage disequilibrium with rs78375762 (D′ = 1), and hence is in LD with rs76346913.

Referencing Genome-wide association analysis identifies novel blood pressure loci and offers biological insights into cardiovascular risk (2017, Nature Genetics)

I'd be grateful if those with expertise in GWAS etc could comment further. If I'm following correctly rs76346913 is the SNP on chromosome 6 identified in males in this thesis. rs147212971 is previously shown to be protective against elevated DBP. Does this "perfect linkage disequilibrium" mean that rs76346913 is at the same site — but multi-allelic — so maybe could have the opposite effect on smooth muscle cells by a different substitution?

This is a multi-allelic site, which complicates the interpretation. There are two possibilities at this site: a C → T nucleotide substitution, or a G deletion (CG → C). The deletion is rare (∼ 0.0090 in the non-Finnish European population), and therefore this variant would have been filtered during the PLINK analyses. Hence, only the more common variant is considered. In the UK Biobank genotyping, the two alleles at this position are coded as reference (T) and alternate (C), with T being the minor and risk allele for ME/CFS.

 

[Chris Ponting]

If I’ve understood right, using best practice mathematical and immunological methods, Joshua did find a single genetic difference that was significant in Uk Biobank:
for males with ME/CFS within the gene PDE10A.

He also helpfully ruled out a number of other findings. This is what our field desperately needs: replication attempts, even if results don’t replicate (here, GWAS and the original Mark Davis T cell Receptor findings).

Joshua is now a postdoctoral research at Harvard Medical School, and Massachusetts General Hospital
https://uk.linkedin.com/in/joshua-dibble-1114aa166 | LinkedIn

I believe he’s working as part of an ME/CFS group.

Unfortunately, when we asked 23andMe for their help (using their date), the male PDE10A signal did not replicate. So we need to look again once we have DNA data from DecodeME. (Also for completeness, note that the FinnGen paper finds no genome-wide significant associations for N=7563 cases for the phenotype: "malaise and fatigue" (60% female). This is equivalent to R53 ICD10 rather than, e.g. G93.3. https://www.nature.com/articles/s41586-022-05473-8.)

 

[Simon M]

Crossed with Chris P

I'd be grateful if those with expertise in GWAS etc could comment further. If I'm following correctly rs76346913 is the SNP on chromosome 6 identified in males in this thesis. rs147212971 is previously shown to be protective against elevated DBP. Does this "perfect linkage disequilibrium" mean that rs76346913 is at the same site — but multi-allelic — so maybe could have the opposite effect on smooth muscle cells by a different substitution?

I don't claim expertise, but do know that the SNP identified is rarely the active, relevant DNA difference. Usually, it is simply a tag near the actual site of action (and sometimes even thousands of kb away). IT is not always easy linking SNP to critical genetic change (Joshua may have done other work to link it to the relevant gene, but it is not an exact science).

Thanks for a very informative post.

 

[leokitten]

Does anyone know the functional consequence of the PDE10A SNP found in ME males?

 

[Tom Kindlon]

In my second study I investigated the use of T-cell clonal diversity as a potential biomarker for ME/CFS. This project used cells from CureME Biobank samples in collaboration with Systems Biology Laboratory (SBL). I developed a data analysis pipeline to analyse T-cell receptor (TCR) genomic DNA data based on the best practices currently used in the fields of immunology and mathematical biology. This approach used a mathematical notion of entropy as a measure for the diversity of TCR repertoires, in this way combining all of the most commonly used metrics in mathematical biology. When combined, these measures form a profile for each repertoire, a set of which can be sorted using a machine learning algorithm to partition the repertoires into subgroups.

My hypothesis was that the T-cell clonal expansion of people with ME/CFS would be greater than for healthy controls, and comparable to disease (multiple sclerosis) controls. Although this method was able to effectively classify TCR chains using simulated data, results from experimentally-derived data did not support the hypothesis, with the most effective classifications for both CD4+ and CD8+ cells failing to pass corrections for multiple hypothesis significance testing.

Open access, https://era.ed.ac.uk/bitstream/handle/1842/39763/DibbleJJ_2022.pdf

See pre-print here:
https://www.s4me.info/threads/compa...sus-controls-2023-dibble-ponting-et-al.34314/