Pathway-focused genetic evaluation of immune and inflammation related genes with CFS

[Bob]

This is a CDC study co-authored by Dr Unger, M Rajeevan and others. It follows on from Rajeevan's (CDC) recent paper on the same subject.

(It came to my attention on Twitter via Linda Vansteenwinckel.)

Open Access.

Pathway-focused genetic evaluation of immune and inflammation related genes with chronic fatigue syndrome
Mangalathu S. Rajeevan, , Irina Dimulescu, Janna Murray, Virginia R. Falkenberg, Elizabeth R. Unger
Available online 24 June 2015
Human Immunology
doi:10.1016/j.humimm.2015.06.014
http://www.sciencedirect.com/science/article/pii/S0198885915001809

Abstract
Recent evidence suggests immune and inflammatory alterations are important in chronic fatigue syndrome (CFS). This study was done to explore the association of functionally important genetic variants in inflammation and immune pathways with CFS. Peripheral blood DNA was isolated from 50 CFS and 121 non-fatigued (NF) control participants in a population-based study. Genotyping was performed with the Affymetrix Immune and Inflammation Chip that covers 11 K single nucleotide polymorphisms (SNPs) following the manufacturer’s protocol. Genotyping accuracy for specific genes was validated by pyrosequencing. Golden Helix SVS software was used for genetic analysis. SNP functional annotation was done using SPOT and GenomePipe programs. CFS was associated with 32 functionally important SNPs: 11 missense variants, 4 synonymous variants, 11 untranslated regulatory region (UTR) variants and 6 intronic variants. Some of these SNPs were in genes within pathways related to complement cascade (SERPINA5, CFB, CFH, MASP1 and C6), chemokines (CXCL16, CCR4, CCL27), cytokine signaling (IL18, IL17B, IL2RB), and toll-like receptor signaling (TIRAP, IRAK4). Of particular interest is association of CFS with two missense variants in genes of complement activation, rs4151667 (L9H) in CFB and rs1061170 (Y402H) in CFH. A 5′ UTR polymorphism (rs11214105) in IL18 also associated with physical fatigue, body pain and score for CFS case defining symptoms. This study identified new associations of CFS with genetic variants in pathways including complement activation providing additional support for altered innate immune response in CFS. Additional studies are needed to validate the findings of this exploratory study.

 

[Bob]

Anyone who can read this entire paper, and report back on it, will get the Phoenix Rising special prize. (BTW, previously, the special prize was "The Bob" so please think carefully about whether you want one of these in your home!)

I can't read this paper at the moment, so I skipped to the 'discussion' section, which starts as follows:

4. Discussion
This study identified several previously unrecognized genetic associations with CFS that are worthy of further study for validation. If validated, these associations support the hypothesis that immune and inflammatory mechanism may be involved in CFS. Many of the SNPs associated with CFS were located in genes involving complement activation, chemokines and cytokines and toll-like receptor (TLR) signaling. The findings in the complement system are of a particular interest, as a role for complement activation in CFS has been suggested by a recent case report and prior gene expression and proteomic studies. In the case report, the patient remained chronically fatigued while the levels of complement split products were elevated but CFS symptoms resolved within two months of normalization of split products [33]. Gene expression studies indicated differential expression of complement protein MASP2 contributing to the C4a split product in CFS following exercise [14] and [15]. Complement proteins in the cerebrospinal fluid of subjects with CFS were elevated compared with healthy controls[34].

 

[Dolphin]

Empiric criteria study unfortunately:

Participants in the follow-up study of a population based surveillance of CFS in Georgia, USA (Georgia CFS surveillance study) were clinically evaluated as described in the baseline surveillance [27], and classified as CFS if they met the 1994 international research definition as previously described.

27. W.C. Reeves, J.F. Jones, E. Maloney, C. Heim, D.C. Hoaglin, R.S. Boneva, et al. Prevalence of chronic fatigue syndrome in metropolitan, urban, and rural Georgia Popul. Health Metr., 5 (2007), p. 5

The Reeves study found a prevalence of 2.54% of the population (i.e. very high. When they used the Fukuda criteria normally they got a prevalence of 0.235% to give you an idea of how many extra people the empiric criteria brings in).

They see the empiric criteria (Reeves et al., 2005) as an operationalisation of the Fukuda criteria which is why they call it a Fukuda study (they do this on lots of their papers).

 

[Bob]

Empiric criteria study unfortunately

Darn it, I thought the CDC had quietly abandoned the Reeves criteria.

 

[Denise]

I thought the CDC had quietly abandoned the Reeves criteria.

I wish they had abandoned it.

 

[Dolphin]

I thought the CDC had quietly abandoned the Reeves criteria.

They probably won't use it for the newer cohorts.. But this is on the Georgia sample.

The study they published a couple of weeks ago http://forums.phoenixrising.me/inde...-from-pathway-focused-candidate-markers.38076 explicitly mentioned using the empiric criteria:

Subject recruitment, clinical evaluation, laboratory tests, and their classification were described previously [25]...

25. Reeves WC, Wagner D, Nisenbaum R, Jones JF, Gurbaxani B, Solomon L, et al. Chronic
fatigue syndrome—a clinically empirical approach to its definition and study. BMC Med.2005;3:19.

That study also involved Mangalathu S. Rajeevan

 

[Valentijn]

It's a very good sign that most of the relevant SNPs listed are missense mutations or in the UTR regions. I haven't looked at the paper yet, but 23andMe does test for both the SNPs listed in the abstract: rs4151667 (minor allele "A", frequency 3%), and rs1061170 (minor allele "C", frequency 27%). Both the V3 and V4 chip test for those.

 

[Dolphin]

Note how they say:

The clinical evaluation included physical examination, laboratory screening and Structured Clinical Interview for DSM-IV (SCID) to identify exclusionary conditions and completion of the Multidimensional Fatigue Inventory (MFI), the SF-36 Health Survey (SF-36), and the CDC Symptom Inventory (SI).

The MFI, SF-36 and SI are what they use for the empiric criteria.

 

[Valentijn]

For rs4151667 on the CFB gene, 4 out of 31 (12.9%) ME patients who I have data for are heterozygous ("AT"), compared to 1 out of 31 (3.2%) ethnically matched controls.

For rs1061170 on the CFH gene, 14/54 (25.9%) alleles are "C", compared with 19/60 (31.7%) in ethnically matched controls. The ME patients have a lot of no-call results on that one.

 

[nandixon]

The Abstract gives a third SNP as well, rs11214105, located in the 5′ UTR of the IL18 gene. The minor allele for that SNP is "A" (frequency about 27-29% in European ethnicity populations).

Assuming the minor allele is the risk allele for the 3 SNPs given, I'm homozygous negative for all three:

rs4151667 TT -/-

rs1061170 TT -/-

rs11214105 GG -/-

I'm probably in the majority subset of ME/CFS phenotypes, with PEM in the form of delayed severe post-exertional fatigue and additionally orthostatic intolerance that requires Florinef (and lots of other classic ME/CFS problems like IBS, etc.), and have had the disease for more than 17 years, so I'm a bit doubtful as to the study findings - relative to those particular SNPs, anyway. (Still have to read the full paper.)

 

[Bob]

This paper is too complex for me to read and understand much of it at the moment, and too specialised for me to interpret it, but it does at least seem to suggest that the CDC are now taking ME/CFS seriously. The study used an unhelpful cohort, but that was probably for convenience, as they had already defined and identified the patients. They do seem to have retired the Reeves criteria now, when creating new cohorts.

The paper includes detailed and specific discussions of the immune system all the way through it, and discusses (e.g.) specific genes, various types of immune cells, and cytokine polymorphisms in relation to specific symptoms in CFS.

For example:

Prior studies have reported a few polymorphisms in cytokines to be associated with CFS[23] and [25]. These previously reported SNPs were not included in the Affymetrix Human Immune and Inflammation Chip we used, however we identified additional cytokine polymorphisms, particularly rs11214105 in IL18, a pleotropic cytokine that enhances perforin mediated T-cell and NK-cell cytotoxicity. IL18 has been reported to contribute to the development and pathogenesis of infectious and neuro-inflammatory diseases with immune and cognitive dyfunctions [43] and [44]. Animal studies linkedIL18 to diseases with sex-specific prevalence and complex, regional and sex-specific parental effects in the brain [45]. We found this IL18 polymorphism to be associated with CFS and quantitative measures of CFS symptoms: body pain, physical fatigue, symptom summary scores and number of CFS symptoms. Our results suggest additional study ofIL18 as a candidate gene in CFS pathophysiology

In conclusion the paper says:

In conclusion, this study identified several new genetic associations of CFS with variants in the complement activation, chemokine’s, cytokines and toll-like receptor signaling. Associations in these pathways provide additional support for a role of altered innate immune response in CFS. Additional studies in larger patient cohorts are needed to validate the findings of this exploratory study.

So I would suggest that this is a sincere piece of work, doing a serious and sincere investigation into the biomedical pathogenesis of ME/CFS. As far as my limited understanding allows me to interpret it.

 

[medfeb]

The problem with this study is that since it used the Empirical definition, its difficult to know what patient population the CDC is actually studying.

As Dolphin points out, the Empirical definition increased prevalence 10-fold over earlier CDC estimates and 6-fold over Jason's estimate. The IOM noted that the empirical definition "resulted in a biased sample with overrepresentation of individuals with depression and posttraumatic stress disorder (PTSD)."

Given that, I'd think its hard to say much about the relevance of these findings to ME.

 

[SOC]

The problem with this study is that since it used the Empirical definition, its difficult to know what patient population the CDC is actually studying.

As Dolphin points out, the Empirical definition increased prevalence 10-fold over earlier CDC estimates and 6-fold over Jason's estimate. The IOM noted that the empirical definition "resulted in a biased sample with overrepresentation of individuals with depression and posttraumatic stress disorder (PTSD)."

Given that, I'd think its hard to say much about the relevance of these findings to ME.

Agreed. They have found some interesting genetic correlations, but how specific they are to ME/CFS is unknown. The correlation may be more to chronic fatigue or neurochemical disturbance than to ME/CFS. That doesn't mean it isn't useful or show that there are abnormalities, just that the conclusion may not be specific to ME/CFS.

If the Empiric collected nine non-ME chronic fatigue patients for every ME/CFS patient, it's also possible that the correlation applies more to non-ME conditions and that any actual genetic correlations to ME/CFS are lost in the noise.

It will be interesting to see how any follow-up research using a better definition alters the results.

 

[JaimeS]

Wow, @Bob, you are fast! I got wind of this in my email just now.

 

[JaimeS]

Umm:

rs4151667 - TT
rs1061170 - CT

I don't appear to have data for rs11214105.

....we identified additional cytokine polymorphisms, particularly rs11214105 in IL18, a pleotropic cytokine that enhances perforin mediated T-cell and NK-cell cytotoxicity. IL18 has been reported to contribute to the development and pathogenesis of infectious and neuro-inflammatory diseases with immune and cognitive dyfunctions

....but....but... darn it!

-J

 

[Snow Leopard]

I'm not excited about this paper at all. It won't be replicated and it seems like the generation of a paper for the sake of it...

 

[Bob]

It won't be replicated and it seems like the generation of a paper for the sake of it...

I agree that it probably won't be replicated because of the infancy and complexity of the science and because of the cohort used. They need to be looking towards defining subsets, rather than comparing all-inclusive populations of fatigued patients with normal controls. But I'm not sure about your second point; Rajeevan seems to have a particular interest in this line of enquiry in ME/CFS, and the discussions in the paper are very detailed and specific. And I'd rather they were pursuing these lines of enquiry than an investigation of sleep patterns or the effects of employment on the quality of life of people with chronic fatigue. Not that I have enough expertise to really understand how robust or meaningful this study is. This science is very much in its infancy, so perhaps no point in getting excited by it yet.

 

[Snow Leopard]

I agree that it probably won't be replicated because of the infancy and complexity of the science and because of the cohort used.

It won't be replicated as the findings are not significant.

 

[Bob]

It won't be replicated as the findings are not significant.

Could you expanded on that at all? I haven't analysed the data.

 

[Snow Leopard]

Could you expanded on that at all? I haven't analysed the data.

It's a limitation of these types of studies in general. Correction for multiple associations needs to be done not just for a few, but all possible, to eliminate bias. None of the P values for SNPs were <0.0000045

In terms of meaningfullness of the results, unless there is a strong effect size, the results usually don't mean much.