The chromosome 17q21.31 inversion of the MAPT (microtubule-associated protein Tau) locus represents one of the most structurally complex and evolutionarily dynamic regions of the genome . The distinct clades of haplotypes (H1 and H2) represent the direct and inverted orientation of the inversion, each with different functional impacts. Specific H1 haplotypes are associated with neurodegenerative disorders such as progressive supranuclear palsy  and Parkinson’s disease , whereas the H2 haplotype is associated with recurrent microdeletions resulting in the 17q21.31 microdeletion syndrome [7, 8]. Neurodegenerative diseases associated with the H1 haplotype exhibit aggregation of hyperphosphorylated protein Tau in neuronal cell bodies [2, 12].
Gene expression differences have been described for MAPT, but there has been no systematic approach to study the effect of inversion status on expression of the other genes at this locus. We used principal component analysis to identify inversion haplotypes at chromosome 17q21.31, and observed that the effect of inversion status is not limited to MAPT expression levels, but also affects several other genes in the 17q21.31 region. In line with literature, we found increased expression of MAPT to be associated with the number of H1 alleles in brain. However, we only observed this in frontal cortex and cerebellum, suggesting that regulation of this gene may differ between brain regions. A previous study identified a specific sequence variant in MAPT (htSNP167/rs242557) in the 17q21.31 region on the H1 haplotype regulating the expression of MAPT in neuronal and non-neuronal cell lines . In this study we focused on the effect of the entire inversion on gene expression with use of a robust principal component analysis strategy. Genotype data of this particular SNP was available in the brain dataset but was not significantly associated with gene expression values.
Importantly, we observed that genes other than MAPT are functionally regulated by the inversion haplotypes as well and may therefore be of importance in diseases associated with the inversion region.
The expression of CRHR1 (corticotrophin releasing hormone receptor 1) is significantly decreased in the H1/H1 haplotype and that of PLEKHM1 (pleckstrin homology domain containing, family M (with RUN domain) member 1) increased. These associations were found in cerebellum only. The PLEKHM1 gene is involved in osteopertrosis by affecting vesicular transport and therefore osteoclast-osteoblast cross-talk . Currently there is no functional data available on MGC57346 (hypothetical protein LOC401884) that was found to be differentially expressed due to inversion status in whole blood. MGC57346 consists of 4 exons of which 3 are shared with the long isoform of CRHR1 (Figure 1). It is therefore possible that these are different splice forms of a single gene, which would suggest that the concurrent association findings with inversion status represent a single event. The fact that directionality of MGC57346 expression in blood and CRHR1 expression in cerebellum is the same in this study supports this view. The CRHR gene is a critical part of the hypothalamic-pituitary-adrenal (HPA) axis that mediates stress response and has been implicated in the pathophysiology of stress-related psychiatric disorders. Of the two receptors in this system (CRHR1 and CRHR2), overactivity of CRHR1 in anxiety and depression has been a consistent finding in animal studies . In human, there is evidence for an interaction of CRHR1 function and stressful life events on vulnerability to depression and alcoholism through regulation of HPA-axis and possibly additional interaction with serotonin transporter loci . In addition, multiple sclerosis (MS) has been associated to HPA-axis activity, specifically genetic variants in CRHR1. We find increased expression of CRHR1 associated with H2 configuration, suggesting that CRHR1 activity and/or stress response might also be altered in or contributing to the H2 related phenotypes such as developmental delay and learning disability.
There is no functional data available for leucine rich repeat containing 37, member A4 (LRRC37A4) expressed in whole blood and LRRC37A (leucine rich repeat containing 37A) expressed in brain. For both genes we observed a significant association with inversion status, however, with opposite effects. It is important to note that the LRRC37 gene family is located at either inversion breakpoints and is therefore likely to be affected by copy number variation that are associated with 17q21.31 inversion status  (Figure 1). Of the LRRC37 family, member A4 (LRRC37A4) in particular has been shown to be most variable in copy number . For these reasons, the association between inversion status and gene expression levels of these genes could be entirely due to differences in copy number linked to H1 and H2 haplotypes.
A recent study finds a strong association between germline hypomethylation and genomic instability, describing that DNA methylation deserts are highly enriched for structural rearrangements . The authors report that rare CNVs that are associated with several neuropsychiatric disorders are significantly linked with local hypomethylation. In fact, germline hypomethylation seems to play a more important role in chromosomal rearrangement than the presence of segmental duplications. Future studies should address whether inversion status of the 17q21.31 region can be linked to (large-scale) changes in epigenetic tags.