Page 14 - Read Online

Donskov et al. J Transl Genet Genom 2021;5:136-62 https://dx.doi.org/10.20517/jtgg.2021.12 Page 3

context [50,51] . The modes by which NR coregulators affect NR action are diverse and include direct
recruitment of transcriptional machinery as well as chromatin remodeling, histone modifications, and
chaperone activity [52,53] . The complexity of NR coregulator interactions is reflected in the palette of
[54]
pathologies associated with genetic variation to this group of transcriptional regulators . As has been
reported for NRs, LoF mutations in several NR coregulators lead to intellectual disability and mental health
problems [55-60] . Supporting an overall increased genetic risk load in NR transcriptional networks in mental
illness, genetic variation in loci harboring NR coregulators has been reported in a range of PDs (see
Supplementary Table 1 for a summary), and increased polygenic burden in retinoid and glucocorticoid
biogenesis and signaling pathways has recently been associated with schizophrenia (SZ) and depression,
respectively [18,61] . The importance of NR coregulator-mediated modulation of NR action has further been
[41]
demonstrated by molecular genetic studies in preclinical models , where genetic disruption to NR
coregulators generally results in behavioral impairments and neurobiological alterations with translational
relevance to PDs [55,62-69] . Collectively, ample evidence implicates dysregulated NR-mediated signaling in the
pathoetiology of mental illness, and it is thus conceivable that genetic vulnerability to NR-mediated
signaling, in combination with their ligand-associated risk factors, collectively shapes the risk and clinical
manifestation of PDs.

Here, we provide a comprehensive and systematic data-mining effort and functional genomic analysis of the
NTC in large-scale genetic and epigenetic data and present new evidence that supports dysregulated NR-
mediated signaling as a common and core molecular pathway in mental illness with significant diagnostic
and therapeutic potential in psychiatry.

METHODS
Gene set selection, filtering, and overlap analyses
NTC gene set
NTC gene set includes genes encoding NRs and NR coregulators in the human genome. A defined list of
NR coregulators was obtained by compiling curated entities from the now deprecated Nuclear Receptor
Signaling Atlas (NURSA; http://www.nursa.org), NRIDs containing NR coregulators with validated
biophysical NR interactions from a recent large-scale peptide array-based study , and minimal
[49]
endogenous modules of NR coregulators identified in a recent comprehensive IP/MS-based study of
endogenous human coregulator protein complex networks . The final list consisting of 48 NR encoding
[48]
genes and 522 NR coregulator-encoding genes can be viewed in Supplementary Table 2.

Genome-wide associated gene sets
For the analysis of overlap between NTC gene sets and genes in genome-wide significant (GWS) loci in
[70]
PDs, the following PGC/iPSYCH PD GWASs were assessed: SZ , bipolar disorder (BPD) , major
[10]
[10]
[8]
depressive disorder (MDD) , ASD , attention deficit/hyperactivity disorder (ADHD) , and cross-
[11]
[4]
disorder (CD) , which includes SZ, BPD, MDD, ASD, ADHD, anorexia nervosa, obsessive-compulsive
disorder, and Tourette syndrome. For the illustration of NTC genes (NTCs) among genes in GWS loci in SZ
[7]
[Figure 1], a smaller PGC GWAS with 108 GWS loci was used with the readability of the illustration in
mind. Additionally, the following non-PD GWASs were assessed: Alzheimer’s disease (AD) ; type 2
[71]
diabetes (T2D) , heart failure (HF) , body mass index (BMI) , height , and COVID-19 (positive vs.
[74]
[74]
[73]
[72]
population) downloaded from GRASP (see Supplementary Table 3 for details). PGC genotype data were
[75]
[76]
all processed using the PGC-developed Ricopili pipeline ; thus, to obtain comparable locus boundaries and
in turn GWS gene sets, summary statistics from non-PGC studies were similarly processed using Ricopili
with 1000 Genomes Project (Phase 3 v5a) as reference.