BH4 - GCH1 Question

[Valentijn]

My homozygous GCH1 risk SNPs
rs4411417(C;C)
rs752688(T;T)
rs8007267(T;T)

Prometheus groups these into gs224 if you have issues like me:
View attachment 10776

I'm confused ... the linked research only mentions one of those SNPs (the one which isn't even on a gene), and is otherwise looking at a completely different haplogroup.

There is some research involving those three, as part as a larger haplogroup: http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2716034/ which is itself looking somewhat statistically dodgy.

 

[shoponl]

@nandixon,

I have the gs224 genoset, but normal phe and tyr values on a plasma amino acid panel. I also have unexplained mild neurological symptoms: fatigue, especially late afternoon and post-exertion; transient brain fog; and tingling in extremities.

Over a decade of self-experimentation, I've ended up with a supplement stack that remarkably resembles Martin Pall's recommendations for NO/ONOO. So, after initially discounting gs224's systemic effects because of my normal phe and tyr, I'm now wondering if GCH1 might explain my symptoms simply due to its effects on NO or peroxynitrite.

This study suggests a possible relationship, for example, and some of the cardiac/GCH1 studies might as well:
http://www.ncbi.nlm.nih.gov/pubmed/17717598 (GCH1 C+243T = rs841)
"Common GCH1 variant C+243T in the 3′-untranslated region (3′-UTRs) predicted NO excretion, as well as autonomic traits: baroreceptor coupling, maximum pulse interval, and pulse interval variability, though not catecholamine secretion. In individuals with the most extreme BP values in the population, C+243T affected both diastolic and systolic BP, principally in females. In functional studies, C+243T decreased reporter expression in transfected 3′-UTRs plasmids. We conclude that human NO secretion traits are heritable, displaying joint genetic determination with autonomic activity by functional polymorphism at GCH1."

Have you run across anything that might support or discount a link between GCH1 and NO/ONOO?

~judi

 

[shoponl]

Judi, your Phe and Tyr might be normal because you have slow PAH and TH SNPs. Or you could be low in Iron. Just being low in BH4 does not always mean your downstream products will also be low.

@ppodhajski,
I never considered that. What a depressing thought.
No, wait-- I just remembered that other people with gs224 have also said their phe and tyr are within range, so it's not just me. Whew.

So the question still stands: could BH4 be high enough to facilitate phe-to-tyr, but still be low enough to cause excess peroxynitrite formation...? Threading a biochemical needle here.

Or maybe an easier question: what symptoms happen with NO/ONOO problems? How do people know?

 

[Gondwanaland]

NO/ONOO is a dead end in my opinion. I know why they were so interested in it but the issue is oxidative stress. If you are worried about peroxynitrite you need to get rid of the superoxides, and that is through the SOD2 gene which turns the superoxide into hydrogen peroxide.

My SOD2 SNPs are bad. I take manganese and eat high manganese foods for that.

But this goes back to why the superoxides are being created in the mitochondria in the first place. And that I blame on slow MAO and COMT SNPs.

That seems a pretty acurate description of what I am dealing with. My question is why it is so hard for me to cope with those issues that many people around me seem to have as well but can deal with them so much better than me.

My husband is very adversely affected by COMT issues (he is 2x +/+), while MAO-A issues are very hard on me (+/+).

 

[shoponl]

My SOD2 SNPs are bad. I take manganese and eat high manganese foods for that....But this goes back to why the superoxides are being created in the mitochondria in the first place. And that I blame on slow MAO and COMT SNPs.

@ppodhajski,
My SOD2 is also possibly a little impaired (4 +/+ out of 13 on LiveWello). I take 6mg manganese in a multi every day, do you take more than this?

**edit** So if superoxides are a problem, what are the symptoms? How do you know?

 

[aquariusgirl]

If your zinc is in the tank, SOD is screwed right? SOD is a zinc/Cu dependent enzyme. Heavy metals lower zinc.

 

[Gondwanaland]

I would like to link some info about the dangers of high Manganese.
http://forums.phoenixrising.me/inde...mutation-anyone-else.23147/page-5#post-460273
I supplemented with Mn for about 3 months, but then I got some high estrogen problems that could have been caused by excess Manganese.

But then again, I was feeling so much better during that time and will restart my low dose supplementation taking it just a few times weekly.

 

[Gondwanaland]

This is too complex for my brain right now...

The family from my mother's side is prone to Parkinson's.

 

[shoponl]

ppodhajski,
I take Thorne's Citramins multimineral (plus extra magnesium and a large handful of other supps), and this seems to work well for me. AFAIK the multimineral hasn't given any negative effects at all.

Thanks for pointing me to SOD2. I tried taking 50mg manganese today and it wiped me out a little. I'll search the archives for more data on this one.


Gondwanaland,
Very curious to hear that manganese and estrogen are linked... I am trying to increase estrogen via BHRT but my E2 levels won't stay high. So, yet another reason to look into manganese! Thanks for the idea.

 

[shoponl]

Hi @Valentijn and @ppodhajski,

Prometheus groups these into gs224 if you have issues like me:

Just to clarify a point of possible confusion, gs224 is a grouping of minor alleles on GCH1. It is shown on 23andme as:
rs10483639 = (CC)
rs3783641 = (AA)
rs8007267 = (TT)

We affectionately refer to this as the Double-Cat genoset.

This is the SNPedia page for it:
http://www.snpedia.com/index.php/Gs224

And the Yahoo group:
https://groups.yahoo.com/neo/groups/GCH1discussions/info

And here are a couple of threads in 23andme:
https://www.23andme.com/you/community/thread/13937/
https://www.23andme.com/you/community/thread/15234/

 

[Valentijn]

Just to clarify a point of possible confusion, gs224 is a grouping of minor alleles on GCH1. It is shown on 23andme as:
rs10483639 = (CC)
rs3783641 = (AA)
rs8007267 = (TT)

So two completely different SNPs than ppod listed above ... that explains why I couldn't find his version in the cited study

Generally it looks like a decent study, but there is one major problem. While the XX (double-CAT) haplotype is associated with much lower average values, the OO and XO groups completely overlap it. In fact, their values go lower in the samples taken from 23 and 17 patients in Figure 1. People in all groups are going down to zero, and they're all clustered near the bottom. The OO and XO groups just have more spread ... most of it clustered a bit above the XX group, but some clustered below, and more high outliers.

Hence I really have to wonder if those results hold up in a replication study. The number of XX patients is quite low (as is to be expected) so that complicates things considerably. At any rate, I don't think too much can be assumed based on that study, when clusters of OO and XO patients have even lower values than every XX patient.

 

[shoponl]

cited study ...
Generally it looks like a decent study, but there is one major problem.

Confusion reigns (as per normal in my head)... to which study are you referring?

I see "View attachment 10776" in ppodhajski's quote you included on Monday, but I don't see the original post nor can I open the attachment.

 

[Valentijn]

Hence I really have to wonder if those results hold up in a replication study.

15 papers have cited the haplogroup study

http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3704723/ : the CAT (T-A-C in that paper) haplotype wasn't significant: but was looking at risk of a disease associated with BH4, rather than actual gene function.

Here's a much better paper by the same authors of the original CAT haplogroup paper - it's not listed at SNPedia, but does a far better job of getting a lot of data into one location and analyzing it in a useful manner: http://circ.ahajournals.org/content/124/17/1860.full.pdf html

Basically people without GHC1 problems have a normal response to certain inflammatory triggers and produced raised BH4 in response. But the people with "double CAT" haplotype are not producing extra BH4 when receiving a vaccination, unlike the OO majority. However their levels are normal in other circumstances.

So it looks like they have replicated the results in different groups, and gotten similar results, in addition to exploring the underlying mechanism, and probably explaining the early discrepancy with overlapping values. Very cool stuff

 

[Valentijn]

Confusion reigns (as per normal in my head)... to which study are you referring?

I see "View attachment 10776" in ppodhajski's quote you included on Monday, but I don't see the original post nor can I open the attachment.

It looks like he deleted it and/or his original post. His post which I'd quoted is now missing from the previous page

At any rate, it was the CAT paper at http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2699614/ I think. It only included one of the three SNPs which he was referring too, which is why the entire affair was so confusing.

 

[nandixon]

@Valentijn and @shoponl

Just a technical note here because I've been researching this gene a fair bit: I think the most up-to-date understanding of how GCH1 is likely working in terms of haplotypes, in a European population, is found in the first reference I gave in post #20 above, which is from 2014.

You can see in Fig. 1C of Figure 1 of that article, which I show below, that there appear to be 2 large haploblocks, the first of which can be defined by rs10483639, rs841 and rs3783641; the second of which by rs10137071 and rs8007267.

23andMe no longer tests for rs841 in its latest test (v4), which is a little unfortunate because I think this may well be the most functional common SNP for the first haploblock (and maybe the whole gene). But since it's in tight linkage disequilibrium with rs10483639 and rs3783641 those should work okay as proxies.

23andMe doesn't test for rs10137071 (on either v3 or v4), so that leaves rs8007267.

So from the latest research, as limited by 23andMe, we're left with exactly the same 3 SNPs that comprise what Promethease/SNPedia calls "gs224" (rs10483639; rs3783641; rs8007267).

About rs841, it's actually located in the 3'-UTR region of the gene, so it doesn't affect the function of the resulting protein/enzyme (i.e., GTP cyclohydrolase 1). However, from everything I've read, it does seems to be very significantly affecting how well the gene is expressed (post-transcriptionally) and therefore how much of the enzyme is actually made (and thus ultimately how much BH4 is made and whether the NOS enzymes properly produce nitric oxide versus peroxynitrite). This particular 3'-UTR SNP seems subject to affects from a great variety of different regulators, like cytokines, blood sugar, oxidative stress, AMPK, etc.

(Rs841 is probably the best single proxy not only for the first haploblock, but for the entire gene, and alone might give 90% or more specificity for a person's genotype for GCH1. But it's a moot point now that it's no longer available, and the other two SNPs for the first haploblock in gs224 seem to mostly represent it.)

Just to mention also that a Japanese study looked specifically to see whether there might be a correlation between rs841 and "CFS." They came to the conclusion that there wasn't one (except for, allegedly, "harm avoidance"), but I haven't seen the full text yet to see how the cohort was selected and all the statistics. (Association of monoamine-synthesizing genes with the depression tendency and personality in chronic fatigue syndrome patients)

Regardless, I do think that rs841 (rs10483639/rs3783641), while not likely a cause of ME/CFS, is very likely to exacerbate/make worse any disease that involves significant oxidative stress, and that includes ME/CFS.

 

[Valentijn]

@nandixon - The linkage disequilabrium between rs841 and rs10483639 is only very strong for one European CEU group. It differs a bit in other CEU groups, and has even bigger differences for other ethnic groups. So I'm not sure it's a suitable substitute.

rs10137071 and rs8007267 are quite different for all available groups. So that's a pretty poor substitute.

 

[shoponl]

@nandixon - The linkage disequilabrium between rs841 and rs10483639 is only very strong for one European CEU group. ...rs10137071 and rs8007267 are quite different for all available groups. So that's a pretty poor substitute.

So far, of the five people in the gs224/gs223 Yahoo group whom we know have data for rs841, all have the homozygous minor (AA) rs841 alleles, as well as homozygous minor alleles for:

rs10131232 (AA)
rs2878168 (AA)
rs3783641 (AA) = in gs224 genoset
rs3783642 (CC)
rs441417 (CC)
rs4411417 (CC)
rs7147286 (AA)
rs7522688 (TT)
rs8017210 (AA)
rs841 (AA)
rs9671371 (TT)
rs10483639 (CC) = in gs224 genoset; 4/5 confirmed CC, one missing data

Three out of five were homozygous minor (TT), two were hetero (CT) at:
rs8007267 = in gs224 genoset

 

[Valentijn]

So far, of the five people in the gs224/gs223 Yahoo group whom we know have data for rs841, all have the homozygous minor (AA) rs841 alleles, as well as homozygous minor alleles for:

That's an incredibly small sample set to be drawing any conclusions from. On dbSNP there are many samples numbering over 100 which can be directly compared between the two SNPs.

 

[nandixon]

This large 2008 study gives good support for the "gs224" haplotype (i.e., rs10483639/rs3783641/rs8007267) being in strong linkage disequilibrium with rs841 in European ethnicities. So gs224 should be a good proxy for rs841, and vice versa, for our purposes. (This is in agreement with @shoponl's small 5-person analysis above, and is somewhat important since 23andMe no longer tests for rs841, as I mentioned in an earlier post.)

Clinical genetics of functionally mild non-coding GTP cyclohydrolase 1 (GCH1) polymorphisms modulating pain and cardiovascular risk

(The full text can be downloaded from here.)

From p.198 of the transcribed pdf file:

To systematize the genetics of non-functional loss GCH1 variants, we analyzed whether the GCH1 ‘‘pain-protective’’ haplotype [i.e., gs224] and the c.*243C>T SNP [i.e., rs841], both associated with distinct but also partly overlapping clinical symptoms, are carried by the same subjects, i.e., represent the same genetic modulator of GCH1. For this, from 891 healthy unrelated Caucasians (cohort 1, Frankfurt) screened previously for the GCH1 haplotype [13] (informed consent and Ethics approval obtained), all 23 homozygous carriers, and a random sample of 60 heterozygous and 79 non-carriers were analyzed for the c.*243C>T SNP (for assay details, see Appendix A). In addition, from 347 patients with coronary artery disease (cohort 2, Oxford) of the cohort analyzed in a GCH1 cardiovascular association context [12], all 7 homozygous carriers of the pain-protective haplotype and a randomly drawn sample of 48 heterozygous and of 120 noncarriers were analyzed for the c.*243C>T SNP.

The three GCH1 SNPs analyzed to reliably diagnose the pain protective haplotype[13] and the GCH1 c.*243C>T SNP were found to be in strong linkage disequilibrium (values of D'>=89 and r2>=75 for the four GCH1 SNPs in cohort 1, and values of D'>=87 and r2>=58 for the four GCH1 SNPs in cohort 2; for details about the genetics statistics, see Appendix A). This agreed with the localization of all four SNPs within one haploblock [Note: rs8007267 may actually be in a second haploblock] according to a solid spine of linkage disequilibrium analysis implemented in the Haploview software [14]. On a subjects’ level, non-carriers, heterozygous or homozygous carriers of the variant GCH1 c.*243T allele were in almost all cases also non-carriers, heterozygous or homozygous carriers of the GCH1 pain-protective haplotype, respectively (x2 test: p<0.001 for both cohorts). The high agreement between test allocations as noncarriers, heterozygous or homozygous carriers of either the haplotype or the c.*243T allele was also reflected in a value of Cohen’s k [15] of 0.82 in cohort1, and 0.74 in cohort 2, which indicates very good or good agreement, respectively[16]. Moreover, the test sensitivity [17] to detect the pain-protective GCH1 haplotype by assessing the GCH1 c.*243C>T SNP was 91.7 and 100% for heterozygous and homozygous carriers, respectively, with respective test specificities [17] of 88.2 and 97.1% in cohort 1.In cohort 2, the respective values of test sensitivity were 85.4 and 100% for heterozygous and homozygous carriers, respectively, with test specificities of 88.2 and 95.2%, respectively.

 

[nandixon]

In practical terms, it's going to happen frequently that the genotypes (i.e., the homozygous or heterozygous states) for the 3 SNPs found within gs224 will not always all be identical. (And from reading many GCH1 papers, I see that the linkage disequilibrium between rs841 and rs8007267 actually may vary considerably in different European groups. Note that these are at opposite ends of the GCH1 gene.)

So to determine the importance of each SNP within gs224 relative to its ability to predict the genotype for rs841 (for those people who don't have 23andMe results for that SNP), the following ranking may be best, I think:

rs10483639>rs3783641>>>rs8007267

Here are some examples:

If you have European ethnicity and have either of the following two results for gs224 (with the SNPs in the order given above), you are very likely homozygous (AA, +/+) for rs841:

1. CC, AA, TT
2. CC, AA, CT

If you have either of the following two results you are very likely heterozygous (AG, +/-) for rs841:

4. CG, AT, CT
5. CG, AT, CC

If you have either of the following two results you probably have the non-risk result (GG, -/-) for rs841:

6. GG, TT, CC
7. GG, TT, CT

Other combinations may be less clear, I think. For example, for:

CC, AT, CT

even though the two heterozygous results in this gs224 combination outnumber the one homozygous result, I think there's actually at least an equal chance of being homozygous for rs841 rather than heterozygous.

(There shouldn't be many, if any, results for the SNPs in gs224 that mix +/+ and -/- results together, I don't think.)