Inflammation doesn't seem to play much causative role in A.G.A if at all, but Wnt signaling clearly does (as suggested by many Wnt-related genes affecting A.G.A risk, and the absence of inflammation-related genes affecting A.G.A risk). And genetic variants affecting hair follicles' response to PGD2 also have no effect on A.G.A risk, so that's probably not a good candidate either. Also, while we can infer that inhibition of Wnt signaling is a major causative factor in A.G.A risk, we don't know whether differential regulation of CXXC5 in particular is a cause or a downstream redundant effect.
Having said that, what can we say about regulation of CXXC5 in hair follicles?
In the dermal papilla in vivo, Cxxc5 mRNA is highest in cells where Wnt signaling activity is the highest (
Yang et al., 2017). This is consistent with previous reports that Cxxc5 is a negative feedback regulator of Wnt signaling.
From top to bottom, the DP clusters are Cluster 1 (C1), Cluster 3 (C3), Cluster 2 (C2), and Cluster 4 (C4). Expression of Lef1, a Wnt target gene, is a good marker for Wnt pathway activity. Lef1 expression is highest in Cluster 2, as are Wnt target genes Fgf7, Fgf10, and Alx4. Cxxc5 mRNA expression is also highest in Cluster 2.
View attachment 72333 View attachment 72334
CLUSTER 2 includes...
Genes, Fold change, FDR
Lef1, 2.42, 6.72E-08
Cxxc5, 2.17, 2.54E-03
But this latest study found that Cxxc5
protein is
negatively correlated with Wnt signaling activity. And in the human hair follicle to the left, Cxxc5 protein was most highly expressed near the bottom (i.e. Cluster 4) DPCs, where Wnt pathway activity is weakest.
View attachment 72335
But in their in vitro experiment, they found that treatment of DPCs with Wnt3a strongly induced Cxxc5 protein expression - this seemingly contradicts in vivo data showing an inverse correlation. It does, however, agree with in vivo single-cell transcriptome data showing a positive correlation between Wnt signaling activity and Cxxc5 mRNA.
View attachment 72336
And here again, another in vivo observation where they used valproic acid, which as expected stabilized beta-catenin but did
not induce Cxxc5 protein expression.
View attachment 72337
Here's a figure from the supplemental info showing expression of beta-catenin and Cxxc5 protein through the hair cycle:
View attachment 72338
Neither beta-catenin or Cxxc5 are expressed in telogen. Beta-catenin protein accumulation comes first, but Cxxc5 protein is still low. Later, Cxxc5 is strongly expressed and beta-catenin starts to fade. Catagen soon follows, beta-catenin fades further, and Cxxc5 eventually fades as well. By the next telogen, both beta-catenin and Cxxc5 are gone once again.
So...
In vitro: Wnt -> beta-catenin stabilization -> Cxxc5 mRNA -> Cxxc5 protein -| Wnt pathway activity (negative feedback)
In vivo, anagen: Wnt -> beta-catenin stabilization -> Cxxc5 mRNA -?- no Cxxc5 protein (permissive for long-lasting, amplified Wnt signaling)
In vivo, catagen: Cxxc5 protein is expressed at high levels and inhibits the Wnt pathway
What could explain this? Maybe there's some secreted factor that's present in vivo, but not in vitro, that disrupts Cxxc5 protein stability? There's actually an analogous system that also works on the Wnt pathway, which very briefly goes like this:
Wnt -> Fzd, Lrp5/6 -> beta-catenin -> Znrf3, Rnf43 -| Fzd
Znrf3 and Rnf43 are transmembrane E3 ubiquitin ligases that promote the endocytosis and proteasomal degradation of Frizzled (the Wnt receptor). Znrf3 and Rnf43 are both themselves Wnt target genes (much like Cxxc5), and therefore act as negative feedback regulators of Wnt signaling (also like Cxxc5). There's a family of secreted growth factors called R-spondins (RSPO for short). R-spondin forms a complex with its receptor Lgr4 or Lgr5 and with Znrf3/Rnf43, and when this happens, it's the Rspo/Lgr/Znrf43 or Rnf43 complex that undergoes proteasomal degradation. As a result, Wnt signaling persists much longer, and is much stronger.
View attachment 72339
So I'm wondering if there's a secreted protein analogous to R-spondin -- sort of like an endogenous PTD-DBM -- that inhibits Cxxc5 protein stability. This protein, if it exists, would have to be expressed in anagen, fall in late anagen prior to onset of catagen, and might also be downregulated in A.G.A hair follicles. Any ideas? If anyone knows of such a protein, or another possible mechanism that might regulate Cxxc5 protein stability, then we can discuss how that mechanism itself is regulated.