Big New Genome Wide Association Study On Androgenetic Alopecia - Preprint

[Swoop]

@InBeforeTheCure,

Truthfully, I don't have a good enough understanding of genes on a deeper level and the methodology that goes with it (not even intermediate). But I'm reading books now, so hopefully I will progress to a better understanding of everything, so that I can discuss this with you. Give me a bit of time please and don't you dare to ever leave .

 

[InBeforeTheCure]

why only affect to certains hairs in scalp??

At the broadest level, probably the same reason you don't grow terminal hair all over your forehead. The details though have not really been studied at all. Along these lines, I noticed the GWAS identified a SNP in the Homeobox D cluster associated with baldness, and as you probably know homeobox genes are deeply involved in biological patterning. And there are many others involved in that besides just HOX genes.

As a side note, in Table S2 I can see that in addition to the genes I've mentioned already, there are some other interesting associations like RUNX1 (completing the set), TCF4 (another hit for the Wnt pathway), and BCL2 (which protects against apoptosis).

EDIT: The TCF4 here is actually NOT the TCF4 in the Wnt pathway -- that's encoded by the gene TCF7L2. Very tricky. This TCF4 is actually an E-protein and paralog of TCF12, and like TCF12 also forms heterodimers with TWIST1, so it belongs in the TWIST subnetwork.

 

[Admin]

Give me a bit of time please and don't you dare to ever leave

You are an invaluable resource here Swoop!

 

[Swoop]

You are an invaluable resource here Swoop!

 

[Swoop]

https://bmcdermatol.biomedcentral.com/articles/10.1186/s12895-017-0054-9

Fig. 2
STRING protein-protein interaction (PPI) query. PPIs of significantly correlated target genes a miR-24; b miR-106a; and c miR-31. Connecting lines represent confidence interactions according to the STRING database. The genes JUN, SFRP1, and FZD7 were targets of all three miRNAs and show a consistent PPI in combination with other miRNA-specific target proteins

 

[InBeforeTheCure]

https://bmcdermatol.biomedcentral.com/articles/10.1186/s12895-017-0054-9

Fig. 2
STRING protein-protein interaction (PPI) query. PPIs of significantly correlated target genes a miR-24; b miR-106a; and c miR-31. Connecting lines represent confidence interactions according to the STRING database. The genes JUN, SFRP1, and FZD7 were targets of all three miRNAs and show a consistent PPI in combination with other miRNA-specific target proteins

Also from the paper:

The third miRNA to show significant mRNA correlations in the present analyses, miR-106a, is reported to be upregulated in the balding, as compared to the non-balding, DPCs of males with male pattern baldness, which suggests that it may be implicated in male pattern baldness pathobiology [7]. Although none of the 53 identified target genes of miR-106a have yet been associated with male pattern baldness, two building blocks of the desmosome - Plakophilin 3 (PKP3) and Desmocollin 1 (DSC1), are reported to play a role in HF morphogenesis [35]. Pkp3 deficient mice develop an abnormal hair coat and secondary alopecia [36]. Although Dsc1 deficient mice show normal HF cycling and structures until the age of four weeks, they develop alopecia and HF degeneration in later life [37]. Moreover, one of the pathways identified in the present study was ‘WNT Signalling’, which is of key importance in terms of HF development and cycling [38, 39, 40, 41]. Interestingly, genetic evidence is available for the involvement of WNT signalling in male pattern baldness development. Heilmann et al. reported that a single nucleotide polymorphism (rs7349332) located intronically in WNT10A was associated with male pattern baldness risk (P ≤ 5 × 10-8) and resulted in reduced WNT10A expression in HFs of risk allele carriers [42]. The present analyses therefore provide strong support for the hypothesis that miR-106a contributes to male pattern baldness development via WNT signalling and that the regulation of cell-cell adhesion may be an important factor in male pattern baldness.

It's definitely easy to pull many Wnt-related genes from M.P.B loci. Also interesting that estrogen target genes in hair follicles are highly enriched for regulation of cell-cell adhesion. TGF-beta and phosphatidylinositol signaling -- categories enriched at M.P.B loci -- are heavily involved in cell-cell adhesion.

 

[hilbert]

That's a complicated question which has barely been explored, so it will probably be many years before we have a satisfactory answer to it. I think an even more interesting question is why susceptibility to androgen-induced hair loss "switches on" in particular HFs at a particular age. Hamilton showed that while 20-year-old castrates injected with testosterone lose hair at the same rate as normal 20-year-olds, but castrates injected with testosterone in their 60s can lose all of their hair within months. It seems then that in these 60-year-old castrates, there is some latent androgen sensitivity which exists independent of exposure to androgens, but it needs that DHT trigger to set it off.

can this just be due to hf ar upregulation after so many years with no T?

 

[InBeforeTheCure]

can this just be due to hf ar upregulation after so many years with no T?

Why would that happen over so many years though? Lack of selective pressure against cells expressing high levels of AR which would otherwise die off if exposed to DHT?

 

[hilbert]

Why would that happen over so many years though? Lack of selective pressure against cells expressing high levels of AR which would otherwise die off if exposed to DHT?

you're right. very scary...
btw, was that study statistically relevant?

edit: is there a link to that study?

 

[Armando Jose]

Maybe this
https://www.ncbi.nlm.nih.gov/pubmed/13711016


or

http://www.sitri.it/hamilton/Hamilton.html

In 1942 a Dr. James B. Hamilton, a Yale anatomist, studied the castrated inmates. The story is that one day Hamilton noticed an identical twin of one of the inmates who came to visit his brother. The castrated inmate brother had a full head of hair. His twin, with the family jewels intact, was quite bald. This gave doctor Hamilton an idea. He experimented with the castrated brother by giving him testosterone. It is rumored that the poor inmate's voice got deeper, he developed acne, large muscles and a sex drive. Dr. Hamilton tells us that he became bald. His hair never grew back.
The "population" of institutionalized, castrated males provided Dr. Hamilton with a means of demonstrating the relationship between baldness and hormones. Testosterone was orally given to 104 castrated inmates; they were compared to 312 "normal" men. When given testosterone, the castrated inmates grew bald, if baldness was in their family history. There was a direct connection between the length of time that testosterone was given and the degree of baldness that occurred: the longer the treatment, the more baldness. Echoing Hippocrates Dr. Hamilton concluded "Men who failed to mature sexually did not become bald".

 

[Feelsbadman.jpg]

I



That's a complicated question which has barely been explored, so it will probably be many years before we have a satisfactory answer to it. I think an even more interesting question is why susceptibility to androgen-induced hair loss "switches on" in particular HFs at a particular age. Hamilton showed that while 20-year-old castrates injected with testosterone lose hair at the same rate as normal 20-year-olds, but castrates injected with testosterone in their 60s can lose all of their hair within months. It seems then that in these 60-year-old castrates, there is some latent androgen sensitivity which exists independent of exposure to androgens, but it needs that DHT trigger to set it off.

Maybe it's because 60 year olds produce more 5 alpha reductase in their hair follicles than 20 year olds.

Suppose the following is known to be true, 5 alpha reductase increases with age (including in hair follicle), even if castrated.

On a scale of 1 to 10, 10 being the highest, let's say the 20 year old has a rating of 2 for 5AR in hf and the 60 year old has a rating of 8. When T is added, Androgenetic Alopecia would commence much more rapidly in the 60 year old.

Exposure to large amounts of DHT will upregulate androgen receptors in a way that is conducive to the progression of Androgenetic Alopecia.

 

[InBeforeTheCure]

you're right. very scary...
btw, was that study statistically relevant?

edit: is there a link to that study?

This was an observation from Hamilton 1951.

But his reference for this is Hamilton 1942, and unless I'm missing something, "the eunuch who reaches the sixth decade of life" really is THE eunuch -- it's one guy Hamilton identifies as "R.J." He didn't lose "all of his hair" in a few months either, but only got to around a NW3.

So I was mistaken on this one, and clearly we shouldn't take this too seriously. Evidence for switch vs. cumulative then is:

1) A small sample of eunuchs, only ONE of which was in his fifties.
2) The observation that co-culture of DPCs and ORS cells from juvenile macacques are resistant to testosterone's effects.
3) Accelerated loss after quitting finasteride (maybe).

...which is suggestive but quite weak.

Maybe it's because 60 year olds produce more 5 alpha reductase in their hair follicles than 20 year olds.

Suppose the following is known to be true, 5 alpha reductase increases with age (including in hair follicle), even if castrated.

On a scale of 1 to 10, 10 being the highest, let's say the 20 year old has a rating of 2 for 5AR in hf and the 60 year old has a rating of 8. When T is added, Androgenetic Alopecia would commence much more rapidly in the 60 year old.

It's possible, but then the question would be: Why is 5ar regulated differently with age?

Exposure to large amounts of DHT will upregulate androgen receptors in a way that is conducive to the progression of Androgenetic Alopecia.

Do you mean in the sense that DHT binding increases AR protein stability, so DHT should upregulate AR protein if all else remains equal?

Anyway, I think the answer to this is probably very complex. Since there have been some 287 SNPs associated with A.G.A which control perhaps ~100 different genes, and since a large number of genes change in expression with age, the number of possible interactions is immense.

 

[Swoop]

Good job on that Hamilton reference of literally THE "eunuch" lol.

It's definitely easy to pull many Wnt-related genes from M.P.B loci. Also interesting that estrogen target genes in hair follicles are highly enriched for regulation of cell-cell adhesion. TGF-beta and phosphatidylinositol signaling -- categories enriched at M.P.B loci -- are heavily involved in cell-cell adhesion.

Perhaps there are problems with cell migration and cell adhesion in Androgenetic Alopecia?

Jahoda talked about this in the following paper: http://replicel.com/wp-content/uploads/2015/02/Jahoda-comment-on-our-paper-JID-2003.pdf

Thus, McElwee’s paper, having raised the prospect of being able to augment follicle size by recruitment, is balanced by Tobin’s evidence of movement of dermal cells not only within the follicle, but outside to the dermis. In skin undergoing androgenetic alopecia, there is the possibility that the balance of migration is altered and incontinence of dermal sheath cells to the skin dermis leads to reduction in size of the dermal papilla, and in turn to miniaturization of the follicle structure. If this leakage is the result of signals from a dermal environment unique to this region of skin, then addition of cells by recruitment might only be postponing the inevitable.

I find it interesting that he mentions that it might be "unique to this region of skin", since experiments have been done where grafts have been transplanted to the forearm. I don't think he sees these observations as definitive?

 

[Swoop]

A genomic approach to susceptibility and pathogenesis leads to identifying potential novel therapeutic targets in androgenetic alopecia.

https://www.ncbi.nlm.nih.gov/pubmed/28263792

 

[InBeforeTheCure]

A genomic approach to susceptibility and pathogenesis leads to identifying potential novel therapeutic targets in androgenetic alopecia.

https://www.ncbi.nlm.nih.gov/pubmed/28263792

Hopefully they make all of their microarray data available. For now though, out of what they've published in their paper I noticed several differentially expressed genes that are also near GWAS loci for A.G.A.

PRKCA (PKC-alpha) is upregulated in bald scalp. PRKCA was a hit in this very recent GWAS meta-analysis.
IGF1R is upregulated in bald scalp.
MAP2K4 (upstream of JNK) is downregulated in bald scalp.
TGFB1 (TGF-beta1) is upregulated in bald scalp.
FZD7 is downregulated in bald scalp.
PIK3R1, which encodes the p85-alpha subunit of PI3K, is downregulated in bald scalp.
MAPT is upregulated in bald scalp.
BCL2 is upregulated in bald scalp and is one of their "hub genes". Paradoxically, it seems like overexpression of BCL2 can actually accelerate catagen onset in mice...?

 

[Feelsbadman.jpg]

It's possible, but then the question would be: Why is 5ar regulated differently with age?

I'm pretty sure it is well established that 5AR does increase w/ age but I don't have a reference. The reasons it does are various I'm sure however, one that stands out to me, is insulin. Insulin sensitivity decreases with age. Insulin has been shown to increase 5ar (again no reference).

 

[Swoop]

Among the top significant TFs associated with transcriptional regulation network (Supplementary Table 7) SP1, ESR1, GCR-alpha (NR3C1) and FOS were included in the Androgenetic Alopecia accepted. On the other hand, CREB1, c-Myc (MYC), Androgen receptor (AR), p53 (TP53), c-Jun (JUN), Oct ¾, RelA (RELA) and YY1 were among a set of TFs that are not found directly in the
dataset, but significantly connected to several objects in it (“hidden”).

@InBeforeTheCure, you not surprised with the proto-oncogenes coming up again?

http://www.sciencedirect.com/science/article/pii/S0040816616301720

 

[Armando Jose]

Regarding common baldness the Hamilton's studies are like Mendel's with colors and flowers

New insight in this issue
http://www.nature.com/articles/ncomms14694

Meta-analysis identifies novel risk loci and yields systematic insights into the biology of male-pattern baldness

 

[InBeforeTheCure]

@Feelsbadman.jpg I searched and found references to 5ar contributing to insulin resistance but not the other way around...?

Among the top significant TFs associated with transcriptional regulation network (Supplementary Table 7) SP1, ESR1, GCR-alpha (NR3C1) and FOS were included in the Androgenetic Alopecia accepted. On the other hand, CREB1, c-Myc (MYC), Androgen receptor (AR), p53 (TP53), c-Jun (JUN), Oct ¾, RelA (RELA) and YY1 were among a set of TFs that are not found directly in the
dataset, but significantly connected to several objects in it (“hidden”).

@InBeforeTheCure, you not surprised with the proto-oncogenes coming up again?

http://www.sciencedirect.com/science/article/pii/S0040816616301720

c-Myc activity induces epidermal and sebocyte differentiation in HFSCs at the expense of hair follicle differentiation (see for example Arnold and Watt, 2001, and Frye et al., who suggest that c-Myc activation may promote epidermal/sebocyte differentiation by decreasing expression of focal adhesion related genes). The interesting thing in the paper you've posted is that beta-catenin apparently induces c-Myc and promotes epidermal differentiation, but beta-catenin is generally pro-HF differentiation so it's a bit surprising.

There was also this transcriptomic study on different populations of stem cells in telogen epidermis in mice. I ran the signature genes for each spatial cluster through TFacts, and it found that one of the epidermal clusters (Cluster I) was highly enriched for c-Myc target genes:

Cluster I - MYC (p = 0, e = 0)
Cluster II - GLI2 (p = 0, e = 0), CTNNB1 (p = 4.4e-4, e = 1.76e-2), TP53 (p = 5.6e-4, e = 2.24e-2)
Cluster III - GLI1 (p = 4.9e-4, e = 1.764e-2)
Cluster IV - CTNNB1 (p = 5.1e-4, e = 1.683e-2)
Cluster V - USF2 (p = 5e-4, e = 1.7e-2), USF1 (p = 9.8e-4, e = 3.332e-2)
Cluster VI - SMAD5 (p = 1.4e-4, e = 5.18e-3), CTNNB1 (p = 8.3e-4, e = 3.071e-2)
Cluster VII - CTNNB1 (p = 0, e = 0), LEF1 (p = 5e-5, e = 2.1e-3)
Cluster VIII - SP1 (p = 3.2e-4, e = 9.28e-3), CTNNB1 (p = 1.02e-3, e = 2.958e-2)

Proto-oncogenes and tumor suppressors by nature are involved in cell proliferation, differentiation, apoptosis, senescence, and so on so it's not really surprising that they would be involved when there are known alterations to these processes in A.G.A.

Regarding common baldness the Hamilton's studies are like Mendel's with colors and flowers

New insight in this issue
http://www.nature.com/articles/ncomms14694

Meta-analysis identifies novel risk loci and yields systematic insights into the biology of male-pattern baldness

Yeah, a lot of the gene regions identified in that study were also found by Hagenaars et al. The latter found more regions due to larger sample size, but there are still some new ones in this meta-analysis. After combining them, I think that makes ~115 genetic regions associated with A.G.A so far.

 

[Swoop]

https://www.ncbi.nlm.nih.gov/pubmed/28403520

This monocentric study reveals that genes encoding mast cell granule enzymes, inflammatory and immunoglobulin-associated immune mediators were significantly over-expressed in Androgenetic Alopecia. In contrast, under-expressed genes appear to be associated with the Wnt/β-catenin and BMP/TGF-β signaling pathways. Although involvement of these pathways in hair follicle regeneration is well-described, functional interpretation of the transcriptomic data highlights different events that account for their inhibition. In particular, one of these events depends on the dysregulated expression of proopiomelanocortin (POMC), as confirmed by RT-qPCR and immunohistochemistry. In addition, lower expression of CYP27B1 in Androgenetic Alopecia subjects supports the notion that changes in vitamin D metabolism contributes to hair loss.