Myelomeningocele exomes with canonical WNT/β-catenin pathway genes disrupted
Out of the seventeen genes that nominally associated with risk for myelomeningocele, eleven were from the canonical WNT pathway. A high proportion in the canonical pathway is surprising, given that most WNT signaling genes previously associated with myelomeningocele are from the noncanonical PCP pathway, whose role in neural tube closure is more established. When a WNT family protein binds FZD and low-density lipoprotein receptor-related protein 5 or 6 (LRP5/6), the canonical WNT signaling pathway is activated (Fig 5). The pathway involves multiple events that influence the level of β-catenin in the target cell. If present in sufficient amount, β-catenin translocates to the nucleus where it displaces transducin-like enhancer/TLE family proteins from binding T-cell factor/TCF proteins and proceeds to upregulate the expression of downstream gene targets (Steinhart & Angers, 2018).
As mentioned in the section on PCP, DVL2 associates with myelomeningocele in the EA population. In addition to their role in PCP, Dishevelled (Dvl/Dsh) proteins are important for canonical Wnt signaling, acting downstream of Wnt, Fzd, and Lrp5/6 (Klingensmith, Nusse, & Perrimon, 1994; Tamai et al., 2000). Dvl2 is already implicated in the cause of myelomeningocele, because two to three percent of Dvl2 double knockout mice display spina bifida phenotype (Hamblet et al., 2002). In humans, DVL2 is also differentially expressed in the caudal human neural tube during closure (Krupp et al., 2012).
DDB1 , which codes for UV damage-specific DNA-binding protein 1, associated with subjects in the EA population as well. DDB1 is used as an adaptor protein for the ubiquitin ligase protein, CUL4 (Angers et al., 2006). Other E3 ubiquitin ligases from this family (CUL1 and CUL3) target β-catenin and Disheveled for degradation (Higa & Zhang, 2007). Although more research is needed, it is possible that a nonfunctionalDDB1 gene might decrease β-catenin degradation, consequently increasing the β-catenin signal in WNT-targeted cells. DDB1 is particularly interesting, because the Fisher’s exact P value for its mutational burden analysis almost passed the strict Bonferroni correction threshold adjusted for the 189 genes compared in the populations of European ancestry (7.88E-4 versus 2.65E-4) andDDB1 is differentially expressed in the caudal human neural tube during closure (Krupp et al., 2012). However, DDB1 has not been previously associated with myelomeningocele.
As mentioned in the section on WNT trafficking, CPE associated with myelomeningocele in the MA population with one qualifying variant found in half of the relevant subjects. Whereas full length CPE interacts with the Wnt ligand, CPE ’s splice variant CPE-ΔN localizes to the nucleus, increases β-catenin expression, and induces expression of Wnt target genes (Skalka et al., 2013). Plausibly, loss of CPE-ΔN might indirectly lower gene expression driven by β-catenin. Though CPE has not been previously associated with myelomeningocele, CPE is differentially expressed in the caudal neural tube during human neural tube closure (Krupp et al., 2012).
PPP2R1A on chromosome 19 also associated with myelomeningocele in the MA population. Two of the three variants that created the association were at the same location in two of three MA subjects carrying the qualifying variants. The variant NC_000019.9:g.52725413G>T p.(Arg527Leu) was over 34 times more frequent in the MA population than in the gnomAD AMR exome control population and the variant NC_000019.9:g.52725413G>A (p.(Arg527His) was not reported in the gnomAD AMR exome controls.PPP2R1A codes for a conserved subunit of the heterotrimeric protein PP2A, a serine/threonine protein phosphatase (Mayer-Jaekel & Hemmings, 1994; Mumbly & Walter, 1993). Another subunit, B56, may direct PP2A to down-regulate the expression of Wnt/β-catenin effector genes by decreasing the amount of β-catenin in the cell (Seeling et al., 1999). It is possible this occurs by directly complexing with Axin (Ikeda, Kishida, Matsuura, Usui, & Kikuchi, 2000) and dephosphorylating part of the APC complex (Seeling et al., 1999). Given these understandings, loss of PPP2R1A function could lead to increased expression of β-catenin effector genes. PPP2R1A is another novel gene association with myelomeningocele.
TNRC6B codes for an Argonaut-associated RNA and shows an association with myelomeningocele in the MA population. The Argonaut molecule is a recognition motif-containing protein that participate in RNA interference (Meister et al., 2005). The TNRC6B RNA serves as one component of a RISC complex that inhibits NLK translation (K. Wang et al., 2013) and the NLK protein participates in the WNT/β-catenin pathway by causing the dissociation of the β-catenin complex from DNA (Ishitani et al., 2003). Importantly, TNRC6B is differentially expressed in closing caudal human neural tube (Krupp et al., 2012). So, it is possible that a nonfunctional TNRC6B transcript compromises the RISC complex and indirectly decreases β-catenin role in downstream gene expression in myelomeningocele subjects. Our study is the first to association TNRC6B and myelomeningocele.
SOSTDC1 on chromosome 7, which codes for sclerostin domain-containing protein 1, associated with myelomeningocele in the EA population. The variant NC_000007.13:g.16505280G>A p.(Ala5Val) was found in two of three EA subjects who possessed qualifying variants and both subjects were heterozygous for this variant. The NC_000007.13:g.16505280G>A p.(Ala5Val) variant was not present in any of the gnomAD NFE exome control subjects. In mouse tooth development, the homolog Sostdc1 serves as an inhibitor of Lrp5/6-mediated Wnt signaling (Ahn, Sanderson, Klein, & Krumlauf, 2010), but SOSTDC1 has not previously been associated with myelomeningocele.