Page 17 - Read Online

Page 6 Donskov et al. J Transl Genet Genom 2021;5:136-62 https://dx.doi.org/10.20517/jtgg.2021.12

coregulators have been reported by the Nuclear Receptor Signaling Atlas consortium, but recent efforts have
both added to this list and significantly extended the known interactions between NRs and NR coregulators
[Supplementary Table 6] [48,49] . Hence, we composed a defined list of NTC encoding genes based on curated
databases and strictly validated protein-protein interactions (the NR gene set with 48 genes and the NR
coregulator gene set with 522 genes) and mapped the overlay of these lists with genes annotated to genome-
wide significantly associated (GWS) loci in PDs [7-11,70] . Consistently, ~15% of loci across diagnostic entities
harbored NTC encoding genes, except for ASD, where only three GWS loci were identified [Figure 1 and
Supplementary Table 7]. In addition, > 13% of all brain-expressed NTC genes reside in loci associated with
ADHD, BPD, MDD, SZ, or CD [Figure 1 and Supplementary Table 8]. Individually, this represents a
significant overrepresentation in MDD [chi-squared test (one-tailed), P = 0.012] and SZ [chi-squared test
(one-tailed), P = 0.002]. Notably, > 20% of NTC genes in GWS loci are associated with two or more PDs
(e.g., EP300 and ESR2). A similar overlap was seen for non-psychiatric traits whose biology is closely
interlinked with NR-mediated signaling [97-104] [Supplementary Table 7].

Genetic variants displaying GWS account for only the most significant, small fraction of the total
heritability of PDs. Hence, to further explore the genetic PD burden in the NTC, we employed a gene set
analysis approach based on the aggregated association of individual genetic markers within the NTC gene
[88]
set . Analyses using the most recently available GWAS summary statistics from each of the five PDs,
namely SZ, BPD, ASD, ADHD, and MDD [7-11,70] , as well as the currently largest CD GWAS , revealed a
[4]
significant association of the NR gene subset of the NTC to both MDD (P = 0.008) and BPD (P = 0.005),
while the NR coregulator subset and complete NTC gene set showed association to BPD (P = 0.003) and SZ
(P = 0.033) [Figure 2 and Supplementary Table 9]. While not taking into the account the significant genetic
overlap between PDs , these associations remained significant for MDD and BPD even after adjusting for
[5]
multiple testing by applying a conservative Bonferroni correction [Figure 2]. When we applied the same
approach to summary statistics from non-PD GWASs where NR-mediated signaling has been reported to
play a role [97-104] , a very significant association was seen for the NR coregulator subset in height and BMI, as
well as a moderate significant association of the NR coregulator subset to HF (Supplementary Figure 1; P =
0.002). For COVID-19 (positive vs. population), in which NR biology plays no obvious role, no association
was observed [Supplementary Figure 1].

[105]
Whereas common variants of small effect contribute to all PDs , particularly early onset disorders, such as
ASD, are enriched with RCVs . To assess the genetic burden of NTC RCVs in PDs, we focused on large (>
[106]
[77]
[78]
3000 cases and controls) WES studies that have been conducted in SZ and ASD . In these studies, PD-
associated RCVs were identified in SZ (a single gene) and ASD (102 genes). Strikingly, 19% of genes with
ASD-associated RCVs are NTC-encoding genes [Supplementary Table 8], representing a significant
overrepresentation [chi-squared test (one-tailed), P < 0.0001]. Furthermore, 32% of identified ASD-
associated NTC RCV-harboring genes reside in PD GWS loci (e.g., RORB and FOXP1), thus supporting the
pathoetiological relevance of these particular NTC genes within multi-gene GWS loci.

Patient epigenetic signature and brain transcriptomic profile support the implication of dysregulated
nuclear receptor-mediated signaling in mental illness
Complementing genome-wide studies of DNA sequence variation, studies of variation to the epigenome
have the potential to reveal biosignatures associated with disease-causing factors in mental illness .
[107]
Particularly, methylome-wide association studies (MWASs) have revealed hundreds of DNA methylation
changes associated with PDs and psychiatry-related traits [80-84,108-114] . The epigenome is dynamic, changes in
response to environmental as well as endogenous factors (e.g., hormonal transitions [116,117] and aging ),
[118]
[115]
[86]
and plays a crucial role in the orchestration of gene transcription in the developing human brain . Clinical
MWASs in brain tissues are rare and of small sample sizes. Hence, we assessed the burden of genes with