A potential association of RNF219‐AS1 with ADHD: Evidence from categorical analysis of clinical phenotypes and from quantitative exploration of executive function and white matter microstructure endophenotypes

Abstract Aims Attention‐deficit/hyperactivity disorder (ADHD) is a neuropsychiatric disorder of substantial heritability, yet emerging evidence suggests that key risk variants might reside in the noncoding regions of the genome. Our study explored the association of lncRNAs (long noncoding RNAs) with ADHD as represented at three different phenotypic levels guided by the Research Domain Criteria (RDoC) framework: (i) ADHD caseness and symptom dimension, (ii) executive functions as functional endophenotype, and (iii) potential genetic influence on white matter architecture as brain structural endophenotype. Methods Genotype data of 107 tag single nucleotide polymorphisms (SNP) from 10 candidate lncRNAs were analyzed in 1040 children with ADHD and 630 controls of Chinese Han descent. Executive functions including inhibition and set‐shifting were assessed by STROOP and trail making tests, respectively. Imaging genetic analyses were performed in a subgroup of 33 children with ADHD and 55 controls using fractional anisotropy (FA). Results One SNP rs3908461 polymorphism in RNF219‐AS1 was found to be significantly associated with ADHD caseness: with C‐allele detected as the risk genotype in the allelic model (P = 8.607E‐05) and dominant genotypic model (P = 9.628E‐05). Nominal genotypic effects on inhibition (p = 0.020) and set‐shifting (p = 0.046) were detected. While no direct effect on ADHD core symptoms was detected, mediation analysis suggested that SNP rs3908461 potentially exerted an indirect effect through inhibition function [B = 0.21 (SE = 0.12), 95% CI = 0.02‐0.49]. Imaging genetic analyses detected significant associations between rs3908461 genotypes and FA values in corpus callosum, left superior longitudinal fasciculus, left posterior limb of internal capsule, left posterior thalamic radiate (include optic radiation), and the left anterior corona radiate (P FWE corrected < 0.05). Conclusion Our present study examined the potential roles of lncRNA in genetic etiological of ADHD and provided preliminary evidence in support of the potential RNF219‐AS1 involvement in the pathophysiology of ADHD in line with the RDoC framework.

gion; and we proposed to examine their potential roles in ADHD pathophysiology by adapting the RDoC approach. More specifically, in this study, we explored the association of lncRNAs with ADHD, which we chose to be represented at three different phenotypic levels as guided by the RDoC: (i) at the first level, ADHD caseness and symptom dimension, (ii) at the second, executive functions as functional endophenotype and (iii) finally, potential genetic influence on white matter architecture as brain structural endophenotype.
Indeed, it has been well-established that the polymorphisms and allelic variation of noncoding regions of the genome influences gene expression; and these variations-scattered throughout the whole genome-are more common than polymorphisms in the coding regions in influencing disorder risks. 7 More specifically, these variants play critical roles in the cis-regulation of nearby genes or trans-regulation of distant genes at both transcriptional and post-transcriptional levels. The risk-associated noncoding SNPs identified by GWAS studies include variants in promoters, enhancers, transcription factor (TF) binding sites, and other important regions associated with components of critical signaling pathways. 4 Moreover, specific molecular mechanisms have been revealed by systematic research for neurodevelopmental and neuropsychiatric disorders; and these include regulative loci (enriched in the noncoding regions) which could fine-tune the splicing process of transcription product and interact with other genes and proteins within the molecular networks in the pathophysiological pathways for ADHD (such as synaptic transmission, catecholamine metabolic process, G-protein signaling, coupled to cyclic nucleotide second messenger, learning or memory, and cell migration). 8 Therefore, noncoding SNPs may represent novel fruitful avenues to explore etiological factors in ADHD.
In the last decade, the Encyclopedia of DNA Elements (ENCODE) project has systematically sequenced the whole human genome.
Strikingly, only 2.94% of genomic sequences were classified as protein-coding genes (ie, GENCODE-annotated exons), with 97% of DNA sequences found in the noncoding regions. 9 Long noncoding RNAs (lncRNAs) is a specific class of RNA with the length more than 200 nucleotides and lacked protein-coding potential; and these sequences cover about 27% of non-protein-coding transcripts in the human genomic annotation. 10 The majority of lncRNAs were expressed uniquely in the brain with the spatial and temporal specificity; and their levels undergo dynamical changes during the development. 11 Critically, lncRNAs participated in the process of maintaining the differentiation potential of brain neural progenitor cells (NPCs), synaptogenesis and cell-specific differentiation all of which are critical in epigenetic regulation of gene expression 12 ; as such, lncRNAs have been proposed as candidate substrates in the pathophysiology of other neuropsychiatric disorders such as autism spectrum disorder (ASD), schizophrenia (SZ), major depressive disorder (MDD), and substance dependence.
Notably, the largest GWAS meta-analysis to date performed in ADHD has reported 12 genome-wide significant loci (p < 5E-08), which span several long noncoding RNA sequences, such as KDM4A-AS1,LINC02497,LINC02060,TMEM161B-AS1,LINC01288,LINC01572,MEF2C-AS1,and LINC00461. 13 Of particular interest is the intergenic lncRNA LINC00461, which showed the highest pleiotropic effects involved in five psychiatric traits: SZ, ADHD, depression, neuroticism, and anxiety disorder. 14 Interestingly, the index variant rs12661753 located in the long noncoding RNA gene STXBP5-AS1 was found to be associated Conclusion: Our present study examined the potential roles of lncRNA in genetic etiological of ADHD and provided preliminary evidence in support of the potential RNF219-AS1 involvement in the pathophysiology of ADHD in line with the RDoC framework.

K E Y W O R D S
ADHD, inhibition, lncRNA, RNF219-AS1, white matter microstructure with ADHD symptom scores in another GWAS meta-analysis consisted of adult population-based and case-only cohorts of adults (p = 3.02E-7). The gene-based analysis showed that three lncRNAs (STXBP5-AS1, LINC01247, and LINC00534) were also associated with ADHD symptoms (p < 0.007). 15 One SNP rs4404327 which located in a neuron-specific lncRNA (BC200 RNA) was also found to be associated with ADHD in an Iranian population. 16 However, none of the above analyses have adopted the RDoC approach to probe beyond the ADHD clinical phenotypes, such as on executive function and brain structural integrity as potential endophenotypes more proximal to gene-molecule mechanisms.
ADHD is currently conceptualized as a neurodevelopmental disorder characterized by multi-dimensional symptoms, which represent the extreme ends of quantitative behavioral traits, continuous with the general population; and ADHD is therefore not the expression of a categorical or discrete disease state. 17 The pathogenesis of ADHD is considered as multi-factorial, and genetic variants might affect multiple dimensional traits such as clinical symptoms, neuropsychological performance, and brain structural/functional features. 18 Recently, advances in behavioral genetics and neuroimaging genetics have provided important theoretical perspectives for understanding the pathophysiological changes of ADHD, which can be mapped onto the biological pathways from gene to disease.
Here, we posit that the RDoC framework applied to ADHD at the cognition and neurocircuitry levels would provide additional advantages for research initiatives. 19 Executive function (EF) refers to task-oriented cognitive processes which require deliberate and effortful control that involves inhibition, working memory, monitoring, and execution. 20 Inhibition and set-shifting-two core processes of EF-are informative cognitive endophenotype of ADHD. [21][22][23][24] The brain tissue-expressed lncRNAs have been found to carry out important functions in brain development, maintenance of neural cell function, pluripotency, and neuronal differentiation, 25 which are relevant to executive functions. Further support for the critical roles of lncRNAs in neural integrity has been shown in animal model. In a rodent ADHD model, one recent study identified a large number of dysregulated expression of lncRNAs in the hippocampus of the brains of the affected rats, and these lncRNAs were involved in the biological pathway of brain developmental processes and neuronal function and maintenance. 26 Overall, different strands of emerging evidence from GWAS and animal studies converge to indicate that the risk variants of lncRNAs can detrimentally affect brain development, structure, function, and cognition; and these neural substrates represent potential candidate targets in the exploration of the underlying neuropathological mechanisms of ADHD.
To further elucidate the potential role of lncRNA in the etiology of ADHD, our study first conducted a categorical case-control association analyses on the clinical phenotype of ADHD in Chinese Han subjects. For the identified risk loci, quantitative trait loci (QTL) analyses were further conducted to explore their genetic effects on ADHD core symptoms and executive functions. Thereafter, imaging analyses in a subgroup were performed to investigate the potential underlying brain mechanisms of ADHD-related lncRNA variants.  These MRI data were partial from the study of Jin et al. 29   All subjects of our study were from the Han Chinese ADHD GWAS project. 31 We extracted the data for children samples (1040 children with ADHD and 630 healthy controls) for genetic association analysis. First, 501 SNP markers for the above 10 lncRNA genes were obtained from the Affymetrix 6.0 SNP array at Capital Bio Ltd.  (Table S1) were generated from these 328 SNPs for analysis according to a screening standard of tagging r 2 > 0.8. was tested through analysis of skewness and kurtosis, with acceptable scores between −1 and 1. were heritable 35 and maybe involved in the pathological pathways from ADHD-associated genes to the core symptoms. 36 Therefore, in this study, the association analyses of ADHD symptom scores and cognitive function scores were conducted by Pearson correlation analysis, and mediation analyses were further performed to explore the cognitive mediation effects between genotype and clinical symptom scores by using PROCESS. 37

| Functional annotations of significant SNPs
The functional information for significant SNPs was retrieved from the public bioinformatics resources. We detected the effects of positive SNPs on gene expression levels by carrying out expression quantitative trait loci (eQTL) analysis based on BRAINEAC database (https://capri ca.genet ics.kcl.ac.uk/BRAIN EAC/) generated from the UK Brain Expression Consortium (UKBEC).

| Imaging genetic analysis
Given the findings from the eQTL analysis and our previous findings that white matters alterations existed in many brain regions of ADHD children compared to health children, 38    University). To further explore whether the white matter alterations have an effect on the cognitive functions, the mean FA values within each significant cluster matching to these atlases were extracted and entered into partial correlation analyses with the cognitive functional scores, after adjusting for age, gender, and IQ. The analyses were performed separately in the ADHD and control groups.

| RE SULTS
Of 1670 participants (1040 ADHD children and 630 controls) examined in this study, there was no significant difference of age between the two groups, but the IQ of children with ADHD was lower than the controls, with higher percentage of male children in the ADHD group. The same patterns were also observed in the imaging genetic analyses, and the details were summarized in Table 1. The mediation analysis for rs3908461 and core symptom with cognitive function scores as the mediator were performed. First, we explored the correlations between the STROOP test ("color interference time" and "word interference time")/Trail making test ("shifting time") and ADHD core symptoms (Table S2). The positive correlations of the word interference time with hyperactive/ impulsive (r = 0.13, P = 4.460E-4) and ADHD total scores (r = 0.12, p = 0.001) were detected, likewise, the set-shifting time and hyperactive/impulsive scores (r = 0.08, p = 0.024). After adjusting for the gender, age, and IQ, only weak correlation between the word interference time and the ADHD total scores remained significant (r = 0.08, p = 0.022) ( Figure 1A). Then, mediation analysis was carried to explore the potential mediation effect of inhibition on rs3908461 genotype and ADHD total symptoms. The result showed that the direct path from genotype to total scores did not reach statistical TA B L E 2 Association between candidate SNPs and ADHD  Figure 1B): this means that the mechanism of SNP rs3908461 effects on ADHD total symptom was accounted by the indirect effect of inhibition control as the mediator.

| eQTL analysis for rs3908461
The eQTL data from the BRAINEAC database showed that the rs3908461 genotype affected RNF219 mRNA expression in the region of intralobular white matter (p = 0.006, Figure 2A), indicating that the minor C-allele can significantly increase the gene expression level compared to T allele. The UK Brain Expression Cohort was based on Caucasian subjects. When checking the allele frequency in Human Genome Diversity Project, we found that the minor allele in Chinese Han subjects was T, but not C ( Figure 2B).

| Genotype, white matter microstructure, and inhibition function
To increase statistical power, we examined the effect of C-genotype rather than ADHD status on white matter structure. First, the effect of C-genotype on white matter microstructures was detected in the

| DISCUSS ION
There are three key findings in our study, and the associations of lncRNAs with ADHD were detected at three different Phenotype Genotype (n)  based on ADHD GWAS has found that RNF219-AS1 was also associated with ADHD as genetically regulated gene. 40 Although the exact function of OBI1-AS1 is still unclear, some studies suggest that natural antisense RNA hybridize with the endogenous mRNA to play the regulatory role. 41 The host gene OBI1 has been found to be involved in the ubiquitin signaling pathway which regulated the DNA replication and effected cell growth and transformation. 42

Core symptoms (ADHD RS-IV)
The ubiquitin-proteasome system, regulating numerous cellular processes in eukaryotes including cell cycle and protein quality control, plays a pivotal role in the central nervous system and has linked to various psychiatry disorders such as SZ, 43 ASD 44 and depression. 45 The dynamic change of the ubiquitination-associated proteasome, from ubiquitination to deubiquitination, could affect the myelin proteins in cellular trafficking and the oligodendrocytes differentiation and are implicated in the demyelinating disease of central nervous system (CNS) such as multiple sclerosis. 46 In addition, the ubiquitin signal could interact with autophagy and then cause the mitochondria damage and degradation. 47 Hwang et al found that mitochondrial DNA haplogroups polymorphism was associated with ADHD children in Korean population. 48 As such, how rs3908461 could affect ADHD expression through the specific molecular pathways deserve further exploration. While a recently published ADHD GWAS analyses identified two SNPs of OBI1-AS1, rs2243638, and rs9574218 as ADHD-related risk variants, 13 the SNP rs3908461 identified in our sample nevertheless is independent from these two SNPs. The SNP rs3908461 did not show any linkage disequilibrium (LD) with those two SNPs neither in CHB nor CEU population ( Figure S1). The evidence from eQTL analysis in brain also supported

F I G U R E 1
The correlation between ADHD total scores and word interference time in the ADHD group. A, The increased word interference time in STROOP test (indicated the poor inhibition function) was associated with the increases in ADHD total scores. B, The full mediation model of word interference time (inhibition) on the relationship between genotype and ADHD total scores (symptoms) that SNP rs3908461 was functionally distinct from rs2243638 and rs9574218, in particular that rs3908461 could influence the RNF219 expression in white matter while rs2243638 and rs9574218 in medulla (http://www.brain eac.org/).
In our study, quantitative analyses detected nominal genetic effects of RNF219-AS1 rs3908461 on inhibition function deficits among ADHD subjects, with TT genotype (as protective factor) carriers performing better. Although no significant association was found for ADHD core symptoms, our mediation analysis found a significant indirect effect of RNF219-AS1 rs3908461 on ADHD total symptom through executive inhibition as the mediator. Previous studies have found that children with ADHD exhibited a range of F I G U R E 2 A, eQTL analysis for rs3908461 on RNF219 transcriptional expression in human brain tissues based on the BRAINEAC database. B, Worldwide diversity of rs3908461 allele frequencies in Human Genome Diversity Project (https://genome.ucsc.edu/trash/ hgc). CRBL, cerebella cortex; TCTX, temporal cortex; OCTX, occipital cortex (specifically, primary visual cortex); FCTX, frontal cortex; HIPP, hippocampus; PUTM, putamen; MEDU, medulla (specifically, inferior olivary nucleus); WHMT, intralobular white matter; SNIG, substantia nigra; THAL, thalamus executive deficits, especially the inhibition function. 49 Family-based behavioral genetics studies conducted in either ADHD 50 or general population 35 have suggested that inhibitory function-as an important executive function refers to the ability to control decisionmaking, behaviors, and emotional impulse-is significantly heritable; and it may be considered a valid endophenotype of ADHD. Executive inhibition is also closely associated with other ADHD traits such as inattention and hyperactivity-impulsivity; and shares substantial genetic risks. Twin studies showed that inhibitory defects in early childhood as a potential genetic risk factor predict later ADHD behavioral problems. 51 To our knowledge, our study is the first to provide preliminary empirical support for the influence of lncRNA RNF219-AS1 on ADHD symptoms through inhibition function as the mediator. However, the detected association between rs3908461 and the inhibition in our study did not survive Bonferroni correction, and our findings must be regarded with caution and as preliminary, awaiting further replication.
The eQTL analyses based on the UK Brain Expression Cohort of Caucasian subjects showed that minor C-allele increased the while not for the exprID t3518971 which was derived from the mean over 13 probesets. Such phenomenon has been found in some other studies. 53,54 The precise reason of the inconsistency between the overall gene expression and the individual exon-specific probesets was not unclear based on the available data. One possible explanation is that the potential influence of genetic variants on the gene expression might be restricted to some specific transcripts, which might involve the complex process of gene splicing. 55 More neurobiological studies were needed to explain. For our present study, the eQTL result was set as suggestive evidence for the following imaging genetic analysis to explore whether candidate SNPs could indeed affect the white matter features. The comprehensive relationship from "genetic variants" via "gene expression" to "brain structural alteration" could not be established at present. Future studies involving mRNA expression data of our own samples might promote such investigation.
To explore more fully the potential functional mechanisms of OBI1, we investigated the 11 key proteins which interacted with OBI1 gene in the STRING Interaction Network (https://versi on11. strin g-db.org/) ( Figure 2) and found two of them associated with the degradation of myelin debris, the generation of new oligodendrocytes (MYD88), 56 immune responses in CNS, and demyelinating disease activity (TOB1). 57 As such, our FA findings of white matter ous study were related to impaired inhibition only in the children F I G U R E 3 A, Regional differences between two genotypes of rs3908461 on fractional anisotropy (FA) maps. B, Higher FA was associated with rs3908461 risk allele (C) dosage in the combined sample (n = 88) of ADHD participants (n = 33) and controls (n = 55) in three white matter regions. C, The correlation between FA and inhibition function in the ADHD group (the red lines) and the control group (the blue lines) separately in the three brain white matter tracts with ADHD in left superior longitudinal fasciculus. The superior longitudinal fasciculus (SLF) is a major frontoparietal white matter tract which connects the parietal and temporal brain regions with the frontal lobes 58 and potentially plays an important role in complex cognitive processes including inhibitory control and set-shifting. 59 These are in line with findings from other studies.
Wolfers et al also found that decreased FA values were associated with the poor attention in the right superior longitudinal fasciculus in adults with ADHD. 60 The decreased FA might be an enduring trait of ADHD existing both in children and adults. 61 Urger et al found that decreased FA values in the left SLF was also correlated with poor attention and language in health children and adolescents. 59 It is possible that differential organization of white matter and the interactions between different brain regions may contribute to the functional diversity. 62 Thus, the integrated analysis of genotype, behavior-executive function and structural connectivity may be an effective approach to unravel the etiology of ADHD.
However, our current findings that risk-allele carriers (C-allele) were linked with increased FA values were not in line with our expectation.
There could be three explanations. First functional changes could be more accurately determined by task-based functional MRI studies 63 rather than only white matter microstructure (FA values). Analyses of multi-modal imaging data, including other brain structural features such as sulcal pits 64 cerebral perfusion and functional brain connectivity, 65 might help to elucidate the true direction of functional changes.
Second, this counter-intuitive finding, however, could represent a compensatory mechanism, similar to another study which found ADHD patients showed increased impulsivity and lower ventral striatal activity, yet the risk genotype carriers showed higher ventral striatal activity, 66 which has been suggested to be a potential compensatory mechanism.
As such, the increased FA values might be a secondary phenomenon in the disease progresses as a compensation for the decreased integrity of white matter upon the occurrence of ADHD. Finally, the increased FA values may reflect delayed neural pruning in our ADHD participants.
Future replication studies are needed to confirm our findings or to unravel potential mechanisms.
Several limitations need be considered. First, for the functional endophenotype analysis, we only detected a marginal relationship between genetic variation and inhibitory function, and the result did not survive correction for the multiple comparisons. The correlation between inhibition and symptom score was also relatively weak.
Our results should therefore be interpreted with caution. Second, not be able to identify the genetic risk for childhood-onset conditions. 67 It would be appealing to enhance the inclusion of children for the research in psychiatry genetics with special attention to those of non-European ancestry. 67 We only recruited children samples in our present study that future studies are needed to verify whether the results are retained in the other age-groups. Fourth, our present study was to investigate the association between lncRNAs genetic variants and ADHD by a post hoc analyses involving additional measurement. Replication in independent cohorts will verify and promote the interpretation of genetic effects of RNF219-AS1 in the etiology of ADHD. Lastly, we only focused on the effects of variations in the DNA polymorphisms of lncRNAs to detect the correlation with ADHD. We could not sequence the actual lncRNAs, which would necessitate accessing brain tissues and therefore would not be practically feasible in human studies. Moreover, we could not explore the molecular function of RNF219-AS1 with ADHD. Therefore, functional mechanisms could only be inferred. As an alternative strategy, in the future, we plan to analyze comprehensively the expression patterns of lncRNAs transcriptome of ADHD patients using high-throughput sequencing technology to explore more fully the molecular mechanism by which lncRNAs influence ADHD etiology.

| CON CLUS IONS
Our study explored the potential involvement of lncRNA RNF219-

CO N FLI C T O F I NTE R E S T
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

AUTH O R CO NTR I B UTI O N S
Guanghui Fu, Lu Liu, and Qiujin Qian contributed conception and design of the study; Haimei Li and Guanghui Fu organized the database; Guanghui Fu, Lu Liu, Wai Chen, Yufeng Wang, and Qiujin Qian performed the statistical analysis, interpreted the results, and wrote sections of the manuscript. All authors contributed to manuscript revision, read, and approved the submitted version.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.