Interpreting the impact of noncoding structural variation in neurodevelopmental disorders

Abstract

The emergence of novel sequencing technologies has greatly improved the identification of structural variation, revealing that a human genome harbors tens of thousands of structural variants (SVs). Since these SVs primarily impact noncoding DNA sequences, the next challenge is one of interpretation, not least to improve our understanding of human disease etiology. However, this task is severely complicated by the intricacy of the gene regulatory landscapes embedded within these noncoding regions, their incomplete annotation, as well as their dependence on the three-dimensional (3D) conformation of the genome. Also in the context of neurodevelopmental disorders (NDDs), reports of putatively causal, noncoding SVs are accumulating and understanding their impact on transcriptional regulation is presenting itself as the next step toward improved genetic diagnosis.

INTRODUCTION

Structural variation (SV) represents the greatest source of genetic diversity in the human genome1,2,3. Copy-number variants (CNVs), such as deletions and duplications, as well as balanced genomic rearrangements, e.g., translocations and inversions, affect more base pairs than single-nucleotide variants (SNVs)1,2,3. CNVs, per definition, result in a gain or loss of DNA and can therefore affect gene dosage. Balanced rearrangements on the other hand, although not accompanied by dosage alterations, can have a significant impact on linear and 3D chromatin conformation.

Despite their large impact on genome structure, it remains a challenging task to comprehensively map all SVs in a human genome. Microarray technology, long the standard in clinical diagnostics to identify DNA gains and losses, neither allows the precise mapping of breakpoints, nor the detection of balanced rearrangements. The advent of next-generation sequencing technologies has improved SV discovery, although short-read genome sequencing (GS) approaches have trouble detecting SVs in repeat-rich regions. Therefore, the most comprehensive overview of human structural variation to date has been achieved through a combination of short-read GS and long-range sequencing technologies, identifying on average over 27,000 SVs per genome2. For an in-depth discussion of strategies and algorithms for SV detection we refer to other reviews4,5.

Given their size (per definition >50 bp), it is unsurprising that germline SVs can contribute greatly to congenital disease6. Both de novo and inherited SVs are frequently linked to the pathogenesis of neurodevelopmental disorders (NDDs)6,7,8,9,10. NDDs are a heterogeneous group of phenotypes in which normal development and functioning of the brain is disrupted, resulting in impairment of motor and behavior skills, speech, vision, hearing, cognition, etc. They include, among others, autism spectrum disorder (ASD), intellectual disability (ID), schizophrenia (SCZ), and developmental delay (DD). Moreover, these disorders are often syndromic, with patients also exhibiting a range of other, non-neurological comorbidities.

It has been estimated that gene dosage alterations caused by large CNVs are responsible for ~15% of NDD cases11. Apart from affecting the protein-coding portion of the genome, it has also been clearly established that SVs can cause disease through noncoding mechanisms, by altering the copy number or position of regulatory elements, or by reshuffling higher-order chromatin structures12. The overall contribution of such regulatory SV effects in disease etiology is still unclear. Yet, several experimentally validated examples, in particular in the context of limb development, have demonstrated their clinical importance and highlighted the diverse ways in which SVs can influence gene regulation12. As development of the brain, the most complex human organ, is tightly regulated, the impact of noncoding SVs should also be carefully considered in the context of neurodevelopmental disease. Although the literature contains multiple examples of noncoding SVs disrupting loci linked to NDD etiology, a comprehensive overview of these cases and the underlying, noncoding disease mechanisms is currently lacking. Therefore, this review aims to shed more light on the importance of noncoding SVs in NDD etiology by discussing (1) the noncoding functional elements involved in gene regulation during neurodevelopment, (2) the contribution of (noncoding) structural variation in NDDs, and (3) an extensive collection of reported NDD cases in which noncoding SVs appear to be at the root of the NDD phenotype.

GENE REGULATION IN NEURODEVELOPMENT

Development of the human brain is a highly regulated process in which genes must be switched on in the right place at the right time. Dynamic, spatiotemporal gene expression programs orchestrate all stages of neurodevelopment, including neural stem cell proliferation, neuronal differentiation, and, ultimately, the migration and integration of postmitotic neurons13. Errors in the regulation of any of these processes could result in aberrant development and give rise to NDDs.

The activity of noncoding functional elements regulates neurodevelopment

The regulatory machinery steering these transcriptional programs depends on the transcription of noncoding RNA (ncRNA) molecules, the activity of noncoding regulatory elements such as enhancers, and the 3D interaction between these noncoding regulatory sites and protein-coding target genes (Fig. 1).

Fig. 1: The gene regulatory landscape from a linear and three-dimensional perspective.
figure1

(a) Genetic locus, illustrating the regulatory function of multiple noncoding elements. The region is delimited by topologically associated domain (TAD) boundaries on either side, each consisting of a cluster of CTCF binding sites. A protein-coding sequence is flanked by a promoter and 5’ and 3’ untranslated region (UTR). An intergenically transcribed long noncoding RNA (lncRNA) performs its regulatory function by acting as a scaffold for the binding of transcription factors (TFs). The activity of multiple enhancer elements in the locus is tissue- and even cell-type-dependent. (b) Via a loop extrusion mechanism, anchored by the CTCF-bound TAD boundaries, the functional elements in the locus are brought into close physical proximity, allowing interaction between the promoter and active enhancers and the assembly of the transcriptional machinery.

Noncoding RNAs

Transcription of noncoding sequences by RNA polymerase II is widespread in the human genome14. Both the process of transcription itself, as well as the resulting ncRNA molecules, can have a regulatory function. Small (21–25 nt) microRNAs (miRNAs) are primarily associated with gene repression, among others, via binding to the 3’ untranslated region (3’UTR) of genes (Fig. 1)15. Several examples highlight their relevance to NDD etiology. For instance, miR-137 is thought to perform a regulatory role during neurodevelopment and has been associated with SCZ and other neuropsychiatric disorders through genome-wide association studies (GWAS)16. Also, the overexpression of miRNAs on chromosome 21 and ensuing haploinsufficiency of their target genes is thought to contribute to ID in Down syndrome patients17.

However, the largest and most diverse group of ncRNA molecules is that of the long noncoding RNAs (lncRNAs). These are per definition longer than 200 nucleotides and include transcripts overlapping other genes (sense or antisense), intergenic transcripts, as well as enhancer RNAs (eRNAs) transcribed from active enhancer elements18. LncRNAs perform their regulatory activity via different mechanisms, both in cis, at the site of transcription, and in trans. They can influence gene regulation by acting as scaffolds, mediating the formation and sequence-specific binding of regulatory protein complexes, or decoys, sequestering and therefore inactivating transcription factors (TFs) and miRNAs (Fig. 1)18. However, as regulatory elements (both promoters and enhancers) initiate bidirectional transcription, it is possible that many antisense lncRNAs and eRNAs are by-products of this process and do not fulfill sequence-specific functions19. In these cases, the act of transcription itself or underlying regulatory element may still contribute to gene regulation20.

Although the function of many lncRNAs remains elusive and their functionality in some cases uncertain, several observations have suggested that they play an important role during neurodevelopment. Expression profiling by the GENCODE consortium showed that many lncRNAs are tissue-specific, with the largest group (~40%) being expressed specifically in the brain21. There are multiple examples of lncRNAs fulfilling a specific regulatory function in all stages of neurodevelopment and in neuronal plasticity18. The Evf2 lncRNA, transcribed from a Dlx5/6 enhancer, recruits the transcription factors Dlx and Mecp2 to Dlx5/6 enhancer elements22. Moreover, it also influences chromatin topology to regulate Dlx5/6 enhancer interactions by binding both activated and repressed target genes on chromosome 6 and regulating cohesin positioning23. Pnky is involved in neocortical development by regulating neuronal differentiation24. Although being transcribed divergently from the Pou3f2 locus, it works via a trans mechanism. Dlx1as and Six3os both play a role in glial–neuronal lineage specification of neural stem cells25, while Paupar regulates olfactory bulb neurogenesis26. Kleaveland et al. even described a regulatory network in which a lncRNA (Cyrano), a circular RNA (Cdr1as), and two miRNAs (miR-7 and miR-671) work together to regulate neuronal activity27.

As might be expected given their role in neurodevelopment, lncRNAs have also been implicated in NDDs, among others through GWAS28. Also, they were found to be enriched in CNVs identified in ASD patients (Alinejad-Rokny et al., unpublished data) and showed differential expression patterns in blood and brain tissue samples from ASD and major depressive disorder (MDD) patients29,30,31. However, it must be stressed that these disease associations are not conclusive evidence that the implicated lncRNAs play a role in disease etiology.

Regulatory elements

Both proximal and distal noncoding regulatory sequences interact to fine tune protein expression levels. The former class, found adjacent to the protein-coding gene body, includes the promoter, which facilitates binding of RNA polymerase II and initiation of transcription, and the 3’UTR, harboring miRNA binding sites to mediate gene repression. Variation in the promoter as well as the 3’UTR of developmental genes has been linked to NDDs32,33. For example, several studies leveraged exome sequencing data to identify noncoding SNVs with a putative regulatory effect in 3’UTRs in patients with ASD, ID, and specific language impairment34,35.

Yet, the most abundant noncoding regulatory elements are enhancers, short DNA sequences that can activate gene expression by recruiting the transcriptional machinery (sequence-specific TFs, coactivators, and RNA polymerase II) to a target promoter in a stage- and tissue-specific manner14. Whether or not enhancers and other regulatory elements are active in certain tissues or at specific developmental stages is determined epigenetically, through DNA methylation and post-translational histone modifications (reviewed in36,37). Promoters are often regulated by multiple enhancers, which can display both overlapping and distinct spatiotemporal activity patterns. While enhancers with overlapping activities confer a level of redundancy that ensures a robust transcriptional output resistant to genetic variation38, those exhibiting differential activities determine the full spectrum of target gene activity. For example, tissue-specific enhancers are active in different subregions of the cerebral cortex, driving precise spatial expression of putative target genes during cortical development39,40. Song et al. even demonstrated that 20–40% of active regulatory elements that interact with the promoters of protein-coding genes are specific to particular neuronal subtypes, underlining their importance in cell-type specific regulation41.

Enhancers can be located exonic, intronic, or intergenic. They can act upon their target promoter from up to megabases away, even skipping intervening genes. These distal enhancers are brought into close physical proximity with their target promoters via chromatin looping (Fig. 1)42,43. Enhancer–promoter (E–P) loop formation has been strongly associated with gene activation during neurodevelopment. Throughout mouse neural development, for example, dynamic E–P interactions are formed at the time of gene activation and disappear upon gene repression44. Also during lineage commitment of human embryonic stem cells into early neural precursors, the rewiring of E–P contacts happens in conjunction with changes in chromatin state and target gene expression45. Although chromatin looping has now been widely accepted as the predominant mechanism underlying E–P interaction, recently several cases have been described in which this mechanism does not apply and E–P distance even increases upon gene activation, suggesting alternative modes of E–P communication might be in play46,47.

Variants affecting enhancer elements have been linked to several NDDs, which are now part of a rapidly expanding class of what are sometimes collectively termed “enhanceropathies”48. For example, enhancers active during cortical neurogenesis and in different neural cell types are enriched for common variants associated with cognitive function and psychiatric disorders41,49. Moreover, de novo variants identified in patients with NDDs were also found to be enriched in regulatory elements, including promoter regions50 and brain-specific enhancer elements51 (Vas et al., unpublished data). For a more detailed discussion on enhancer function in brain development and disease we refer to other reviews52,53,54.

3D chromatin conformation

As discussed earlier in the context of E–P looping, gene regulation cannot be interpreted on a linear scale. Indeed, the human genome is organized into a hierarchical 3D structure (reviewed in55). On the smallest level E–P loops facilitate communication between enhancers and their target promoters42,43. E–P interactions are mostly confined to topologically associated domains (TADs), delimited by CTCF-bound insulator elements (Fig. 1). These insulated domains spatially constrict interactions, limiting E–P communication across their boundaries. TADs are thought to be formed through a “loop extrusion” mechanism, in which cohesin extrudes a chromatin loop through its ring-shaped structure until it runs into convergent CTCF-bound sites. At the highest level, compartments group active (A compartments) or inactive (B compartments) TADs. TADs switch compartments (i.e., compartment switching) when they become activated or repressed, for example during differentiation56.

The functional importance of this organization during neurodevelopment is highlighted by two studies demonstrating a massive rewiring of 3D genomic structures during mouse44 and human57 neural differentiation, including changes in compartmentalization, an increase in TAD size and interaction strength and the formation (or pruning) of dynamic chromatin loops. Rajarajan et al. found that these dynamic interactions also include SCZ-associated sequences57. There is indeed ample evidence that such structural changes play a role in the etiology of NDDs. For example, variants affecting the architectural proteins CTCF, YY1, and STAG1 (a cohesin subunit) were found to cause ID58,59,60, while the multisystem disorder Cornelia de Lange syndrome is frequently associated with variants in the cohesin loading factor NIPBL and the cohesin subunits SMC1A and SMC361.

Accessible tools for interpreting variation in the noncoding genome

The past decade, several consortia have made considerable efforts to comprehensively map (ncRNA) expression, regulatory elements, as well as 3D genomic interactions, across a variety of cell types, tissues, and developmental stages. These large data sets are typically easily accessible through dedicated websites or genome browsers and consulting them should always be considered as a first step in assessing the functional consequence of noncoding variants (Table 1). For example, Middelkamp et al. devised a computational pipeline, based on a combination of phenotype association and publicly available chromatin organization data, to predict driver genes that were directly or indirectly affected by SVs in patients with multiple congenital anomalies and ID62. However, many functional elements remain to be discovered and/or have not been experimentally validated. These lacunae in the functional annotation of the noncoding genome, as well as the lack of experimental validation, complicate the interpretation of noncoding variation. Therefore, additional functional assays might be required to fill the gaps. Recently, experimental strategies to identify and validate regulatory elements were extensively reviewed by Gasperini et al.63.

Table 1 Publicly available resources for the functional interpretation of noncoding variants.

STRUCTURAL VARIATION IN NEURODEVELOPMENTAL DISORDERS

The genetic etiology underlying NDDs is heterogeneous, ranging from large chromosomal aberrations (SVs) to SNVs, affecting hundreds of genes64. Variants that have arisen de novo account for the majority of cases65. For example, ~60% of severe ID cases can be explained by de novo variants (both SNVs and SVs) in known ID genes66. However, rare inherited variation has also been shown to contribute to the pathogenesis of NDDs. Inherited SVs contribute in almost 4% of ASD cases9. In addition to these rare de novo and inherited variants, even common variants add to disease predisposition67.

Noncoding regulatory SVs are enriched in NDDs

As labs are shifting to GS, there has been a tremendous increase in the identification of variants within noncoding DNA sequences, both in healthy and diseased individuals. For example, only ~0.5% of SNVs identified through GS lie within coding exons1. Interpreting the functional effect of these noncoding variants represents an enormous challenge. Most noncoding SNVs and even small indels, unless they disrupt a crucial functional site (e.g., TF binding site [TFBS]), have a low probability of affecting the function of regulatory elements. Indeed, it has been estimated that only 0.15% of possible SNVs in active brain enhancers could cause NDDs51. Even so, it is estimated that de novo SNVs in putative regulatory elements could explain up to 5% of NDD cases6,51,68.

SVs affecting noncoding regions, on the other hand, remove or multiply kilobases of DNA sequence or even relocate entire sections of chromosomes and are therefore more likely than SNVs to have a biologically meaningful impact on gene regulation. Redin et al. demonstrated that 7% of balanced translocations in NDDs cause the disruption of TADs69. Also, SVs affecting regulatory elements seem to be depleted in the normal population, underscoring their potential to be deleterious9,70. Brandler et al. showed SV depletion in cis-regulatory elements (transcription start sites [TSSs], 3’UTRs, and fetal brain promoters) of loss-of-function intolerant genes9, while Han et al. found SVs, especially those disrupting CTCF sites, to occur at significantly lower frequencies than intergenic SVs70.

Others have demonstrated a direct link between noncoding SVs and neurodevelopmental disease. De novo CNVs encompassing human accelerated regions with regulatory activity are implicated in up to 1.8% of ASD cases71. Also in the context of ASD, Turner et al. found a slight excess of de novo CNVs in putative regulatory elements within the vicinity of ASD genes, although this was not statistically significant due to a limited sample size10. Similarly, Monlong et al. detected an enrichment of noncoding CNVs near known epilepsy genes72, while Brandler et al. showed an enrichment of paternally inherited regulatory SVs in ASD cases9. Within the latter ASD cohort no de novo regulatory SVs were identified, illustrating another mode of NDD etiology and underscoring that screening of larger NDD cohorts will be required to uncover the full spectrum of noncoding SVs underlying NDDs.

Functional consequence of noncoding SVs

The pathogenicity of SVs affecting protein-coding genes can in most cases be explained, quite intuitively, by gene dosage effects. However, the effect of noncoding aberrations on gene expression is more difficult to predict. Still, it has been shown that noncoding SVs alter the expression of nearby genes with larger effect sizes than noncoding SNVs or indels73. Moreover, the expression of target genes is negatively correlated with the total sum of enhancer sequence affected by an SV (both for deletions and duplications)70. Both observations demonstrate a link between the number of affected nucleotides and the magnitude of the functional effect.

There are several extensively studied cases that have given insight into the different mechanisms through which noncoding SVs can impact gene regulation, 3D chromatin structure, and, eventually, influence gene expression and phenotypic outcome (reviewed in detail in12,43). These include both dosage and positional effects (i.e., they can change either the copy number or order of regulatory elements). CNVs that delete, duplicate, or amplify noncoding sequences potentially alter the dosage of ncRNA genes or regulatory elements, in turn affecting the expression of target genes. However, this effect on target gene expression might be difficult to predict. As gene expression is often the result of the combinatorial action of multiple enhancer elements with (partly) overlapping activity patterns, the degree of enhancer redundancy will influence the functional consequences. For example, the individual deletion of ten different enhancer elements at loci involved in limb development did not result in a limb phenotype, while the deletion of pairs of enhancers did38.

SVs that span TAD boundaries can, in addition to causing dosage effects, also influence 3D chromatin structure. The deletion of a TAD boundary facilitates the fusion of adjacent TADs, while a duplication involving a TAD boundary can result in the creation of a new chromatin domain or neo-TAD. Inversions and translocations reshuffle the TAD structure by repositioning boundaries. These alterations in the 3D chromatin structure can bring about the decoupling of a promoter from its cognate enhancers resulting in a regulatory loss of function, while at the same time the adoption of new enhancers with different spatiotemporal activities might lead to ectopic gene activation74. However, these mechanisms are not necessarily generalizable. For some loci the removal of TAD boundaries and ensuing TAD fusion does not appear to result in gene expression changes75.

SVs IMPACTING GENE REGULATION IN NDDs: SPECIFIC CASES

Because of the complexity of the gene regulatory landscape illustrated above, the medical interpretation of SVs in noncoding regions requires a case-by-case evaluation of the functional impact on gene expression and phenotypic outcome. To better understand the diverse ways in which noncoding SVs contribute to the etiology of NDDs, we have compiled an extensive list of loci harboring putatively causal, noncoding SVs in patients with NDDs (Table 2).

Table 2 Examples of noncoding structural variants (SVs) putatively causal in neurodevelopmental disorders (NDDs).

lncRNAs disrupted by SVs in NDDs

Although differential expression of lncRNAs has frequently been linked to NDDs29,30,31, it is often difficult to disentangle association from causation. Cases in which disruption of a specific lncRNA gene appears to be the causal disease mechanism are still limited in number and come with a varying degree of functional evidence. Therefore, it remains unsure whether this is a widespread phenomenon in the etiology of NDDs and disease in general.

By 1996, chromosomal aberrations affecting the lncRNA gene DGCR5 had been identified in patients with DiGeorge syndrome76. More recently, CNVs situated within the same 22q11.2 critical region were found to be associated with SCZ as well7. Although this region also harbors several protein-coding genes, based on coexpression analysis using brain transcriptome data from the PsychENCODE project, DGCR5 was identified as a hub regulator77. Moreover, DGCR5 knockdown and overexpression in induced pluripotent stem cell (iPSC)–derived human neural progenitors demonstrated that this lncRNA regulates the expression of several SCZ-associated genes77.

Very recently, differential expression analysis during mouse neuronal induction implicated the highly conserved lnc-NR2F1, transcribed divergently from the NR2F1 locus, in neurodevelopment78. Overlap of this lncRNA with a focal deletion found in ASD/ID patients and a newly identified, paternally inherited translocation confirmed its clinical relevance in NDD etiology (Fig. 2a). Both gain- and loss-of-function experiments in mouse have demonstrated that lnc-Nr2f1 promotes neuronal maturation pathways in a functionally distinct fashion from its neighboring gene Nr2f1, while chromatin association assays showed that it binds to neuronal targets in trans to exert that function78. Also in the context of ASD, PTCHD1-AS was found to be frequently affected by microdeletions in male patients79. This lncRNA lies upstream of the PTCHD1 gene, which is a transmembrane protein known to be involved in NDDs. iPSC-derived neurons of patients with PTCHD1-AS deletions showed decreased excitatory synaptic activity, although PTCHD1 expression does not appear to be affected. This in contrast to lncLRFN5-10, which does appear to regulate expression of its nearby gene LRFN5 and was also found to be affected by a microdeletion in an ASD patient80. Translocations and deletions disrupting LINC00299 have been identified in patients with DD and ID81,82. Although it has been demonstrated that LINC00299 expression increases during neural differentiation, its precise function is still unknown.

Fig. 2: Structural variants disrupt noncoding RNA genes in patients with neurodevelopmental disorders (NDDs).
figure2

Illustration of NDD cases in which the disruption of a noncoding RNA locus has been identified as the causal mechanism (four examples, ad). (Noncoding) genes are depicted as blue and gray boxes, red bars are patient deletions, arrows are translocation breakpoints. The epigenetically regulated MEG3 differentially methylated regions (DMRs) and Prader–Willi imprinting control (IC) region are represented by orange boxes.

RMST was also first associated with neurodevelopment through transcriptomic analyses. This lncRNA was found to be upregulated during neuronal differentiation and was shown to regulate this process through interaction with the SOX2 TF83,84. Recently, a de novo balanced translocation disrupting RMST was identified in a patient with Kallmann syndrome, a disorder caused by deficient development of gonadotropin-releasing hormone (GnRH) neurons and featured by abnormal sexual development and an impaired sense of smell (Fig. 2b)85. The translocation caused a reduction in RMST expression in patient neural crest cells (NCCs), the cells from which GnRH neurons originate, resulting in abnormal NCC morphological development85.

The dysregulation of lncRNAs within imprinted regions has also been associated with NDDs. For instance, microdeletions involving the differentially methylated regions (DMRs) of maternal origin upstream of the maternally expressed gene 3 (MEG3) lncRNA result in a loss of MEG expression and a phenotype resembling that of paternal uniparental disomy 14 (upd[14]pat) patients (growth retardation, DD, facial abnormalities, small bell-shaped thorax, abdominal defects, and polyhydramnios) (Fig. 2c)86. Finally, the deletion of the imprinted SNORD116 noncoding gene cluster on the paternal allele results in Prader–Willi syndrome (PWS) (Fig. 2d)87,88. SNORD116 is processed into 30 small nucleolar RNAs (snoRNAs). The production of these snoRNAs appears to be neurospecific89 and their deletion results in smaller neuronal cell bodies due to a decrease in nucleolar size88. However, the precise role of these snoRNAs in the nucleolus remains unsolved.

Disruption of near cis-regulatory elements (promoter, 5’ & 3’ UTR) in NDDs

As near cis-regulatory elements, such as the promoter and 5’ and 3’ UTR, are typically small in size and directly flanking the coding sequence, they are inherently less likely to be affected by SVs that leave the protein-coding gene body intact. However, there are a few examples of larger genomic variants disrupting these near-cis sequences in the context of NDDs. Undoubtedly one of the most well-known examples is the CGG repeat expansion in the 5’ UTR of the FMR1 gene, resulting in DNA hypermethylation at the promoter, silencing FMR1 and giving rise to fragile X syndrome90. Interestingly, this repeat expansion also disrupts the TAD boundary adjacent to FMR1, decoupling FMR1 from putative downstream enhancers91. As the extent of the disruption correlates to FMR1 silencing, it is possible that part of the FMR1 loss is attributable to this 3D rearrangement. Repeat expansions in near cis-regulatory elements are a recurrent cause of NDDs. A GGC repeat expansion in the XYLT1 promoter of patients with Baratela–Scott syndrome results in hypermethylation of the first exon and reduced XYLT1 expression92. Repeat expansions also give rise to other forms of ID, including FRAXE (5’ UTR of AFF2), FRA2A (promoter of AFF3), and FRA12A (5’ UTR of DIP2B)90.

Other types of SVs have also been reported within near cis-regulatory sequences. For example, a deletion in the promoter region of GPR56, disrupting an RFX TFBS, leads to gyral malformations in a specific region of the cortex, resulting in speech delay, ID, and seizures52. Variants affecting the coding sequence of GPR56 typically result in polymicrogyria of the entire cortex. However, as the regulatory deletion is located within the promoter of only one of multiple alternative TSSs, it only eliminates GPR56 expression in the lateral neocortex, explaining the regionally restricted phenotype.

SVs disrupt long-range gene regulation and 3D chromatin structure in NDDs

Most noncoding SVs associated with NDDs are situated within the large stretches of intergenic space, affecting regulatory interactions between promoters and enhancers and/or altering the 3D chromatin conformation of the locus (see section above). Variants affecting the ZRS limb enhancer at the SHH locus are well known to cause limb malformations. However, translocations upstream of SHH, disrupting the interaction between the SHH promoter and SHH brain enhancers (SBE6, SBE4, SBE2, and SBE3), have been identified as a cause of holoprosencephaly (Fig. 3a)93,94.

Fig. 3: Noncoding structural variants disrupt gene regulation and 3D chromatin structure in neurodevelopmental disorders (NDDs).
figure3

Illustration of NDD cases in which the disruption of regulatory elements and/or 3D chromatin conformation has been identified as the causal mechanism (ten examples, aj). Blue triangles reflect the topologically associated domain (TAD) structure of the locus. Depicted are (noncoding) genes represented by blue (gene of interest) and gray (other gene) boxes, brain enhancer elements (green ovals), patient deletions (red bars), duplications (purple bars), insertions (red triangles), and translocation or inversion breakpoints (red arrows).

The FOXG1 TF has been associated with a congenital form of Rett syndrome, an NDD featured by severe DD, absence of speech, seizures, hypotonia, and stereotypic movements. The gene is located in a large, gene-poor TAD with a pronounced sub-TAD structure and multiple interaction loops, bringing the promoter into the proximity of several in vivo validated brain enhancers95,96. This regulatory structure is disrupted by translocations and deletions distal to FOXG1 in multiple patients with similar Rett-like features (Fig. 3b), likely caused by either enhancer deletion/translocation or a rewiring of interactions due to the deletion of TAD boundary elements69,97. SVs in the 5q14.3 region upstream of MEF2C result in a Rett-like syndrome as well (Fig. 3c)69,98. This upstream region harbors multiple enhancer elements that display in vivo neuronal activity during zebrafish development and form a physical interaction network with the MEF2C promoter in neuronal cells, suggesting that MEF2C transcriptional dysregulation by enhancer deletion or translocation lies at the root of the MEF2C-related phenotype98.

Loss-of-function variants in SATB2 or microdeletions in the 2q33.1 region affecting SATB2, typically give rise to SATB2-associated syndrome, an NDD characterized by ID and dysmorphic facial features99. Six patients with BCA breakpoints in the gene desert 3’ to SATB2 exhibited overlapping clinical features (Fig. 3d)69,100. Each of these breakpoints disrupts the long-range interactions between SATB2 and multiple putative enhancer elements, of which at least one (CRE2) drives SATB2-like craniofacial expression in zebrafish100. Interestingly, the activity of this element appears to be dependent on binding of the SOX9 TF, which has been associated with a craniofacial disorder with overlapping clinical features called Pierre Robin sequence (PRS), suggesting that SATB2 regulation might be primarily driven by SOX9. Translocations and microdeletions upstream of SOX9 have also been identified in patients with PRS, while duplications involving the TAD boundary give rise to Cooks syndrome and intra-TAD duplications cause sex reversal (Fig. 3e)101,102. The effects of these SVs on gene regulation and 3D conformation at the SOX9 locus have been discussed at length by others12,103. Duplications and deletions of the SOX3 TF cause ID and growth hormone deficiency. Intriguingly, multiple SVs have been identified in the gene desert surrounding SOX3 in patients with varying phenotypes without ID (Fig. 3f). A region 82 kb downstream of the gene is especially prone to insertions due to the presence of a human-specific short palindromic sequence. The distinct symptoms observed in these patients suggest the phenotypes might be caused by the introduction of tissue-specific enhancer elements driving ectopic SOX3 expression. A 170-kb fragment from chromosome 9 was inserted at this site in a patient with cleft palate and facial dysmorphism62. The insertion contained part of a superenhancer region with craniofacial activity, possibly altering SOX3 expression during craniofacial development and resulting in the patient’s phenotype. Insertions of other genomic fragments resulted in a severe hair overgrowth phenotype (hypertrichosis), drooping eyelids (ptosis), XX male sex reversal, and X-linked recessive hypoparathyroidism104,105. Additionally, an upstream microdeletion was identified in a patient with XX male sex reversal106.

Chromosomal aberrations affecting the DLX5/6 locus cause split hand/foot malformation 1 (SHFM1), often combined with ID, craniofacial defects, and hearing loss. Several of these SVs have breakpoints upstream of DLX5/6 and disrupt multiple tissue-specific enhancer elements, which regulate DLX5/6 expression in the forebrain, branchial arch, ear, and limb (Fig. 3g)107,108. Patients can even be classified into three phenotypic groups, correlating with the deletion of specific enhancer elements107,108,109. Variants disrupting the PITX2 TF are typically associated with Rieger syndrome, a developmental disorder characterized by ocular and craniofacial anomalies, with some patients also displaying neurological deficits110. Translocations and deletions affecting conserved enhancer elements (with brain, eye, and craniofacial activity) in the gene desert upstream of PITX2 result in a similar phenotype (Fig. 3h)69,111,112. Variants in the coding sequence of PAX6 give rise to the congenital eye malformation aniridia, as well as neurodevelopmental defects. Translocations in the downstream regulatory region were found to result in a similar phenotype52,113.

While many of these examples consist of deletions or translocations, duplications of noncoding sequences have also been associated with NDDs. Coding variants or small duplications in the ARX TF are a frequent cause of X-linked ID, epilepsy, and lissencephaly. Its specific expression in different regions of the forebrain is tightly controlled by ultraconserved enhancers downstream of the coding sequence, with partially overlapping spatial activity patterns39. Duplications encompassing these ARX forebrain enhancers cause a similar, although milder, phenotype (Fig. 3i)52.

Also in patients with complex genomic rearrangements noncoding SVs can contribute to the overall phenotype. For example, Middelkamp et al. identified four candidate driver genes (i.e., PHIP, COL12A1, BMP2, and TFAP2) in a patient with a complex rearrangement consisting of six breakpoint junctions and two deletions on three different chromosomes62. Each of the driver genes individually can only account for part of the phenotype (i.e., DD, autism, seizures, facial dysmorphism, growth delay, missing ribs, renal agenesis, and cryptorchidism), yet together they might explain the full phenotypic spectrum. PHIP and COL12A1 were directly affected by a deletion and have been associated with DD and facial dysmorphisms. In addition, BMP2 and TFAP2A appeared to be affected by a disruption of long-range interactions. Several breakpoints were identified upstream of BMP2, linked to short stature, facial dysmorphisms, and skeletal anomalies, and also the TFAP2A TAD was disrupted by a translocation. Recently, a de novo heterozygous inversion disrupting the TFAP2A TAD was also identified in a patient with branchiooculofacial syndrome (BOFS; branchial cleft, ocular anomalies, facial dysmorphisms) (Fig. 3j)114. The TFAP2A TF regulates neural crest development and its expression in NCCs is controlled by multiple enhancers. Laugsch et al. have demonstrated that the inversion separates TFAP2A from its NCC enhancers, leading to monoallelic expression and TFAP2A haploinsufficiency. Interestingly, no enhancer adoption occurs in this case, even though the inversion places these relocated enhancers within the spatial proximity of other genes.

Finally, in some cases functional evidence for a noncoding disease mechanism is still limited. For example, translocations 3’ to the BCL11B TF gene were found in patients with DD, speech impairment, and ID115. Expression was reduced by 50% in patient cells, suggesting BCL11B haploinsufficiency due to relocation of regulatory elements. Redin et al. identified a patient, displaying epilepsy and DD, with a translocation affecting the SLC2A1 TAD and decreasing SLC2A1 expression in patient cells69. The translocation disrupts the interaction between SLC2A1, a gene associated with the seizure disorder GLUT1 deficiency syndrome, and several putative enhancer elements. Duplications of the 7q36 region, either including or just upstream of VIPR2, result in upregulation of VIPR2 and cause schizophrenia in patients52. Interestingly, it seems like the overexpression pattern cannot only be explained by an increase in gene dosage, suggesting that these duplications affect VIPR2 regulation as well. A de novo duplication 300 kb upstream of NR2F2, a gene associated with ASD and ID, duplicates a human accelerated region that has been shown to interact with the NR2F2 promoter, possibly exerting a regulatory function during neural development71. Intronic CNVs can also affect gene regulation. A 14-kb inherited, intronic deletion in the DSCAM gene has been identified in an autism patient10. The deletion removes at least nine enhancer elements driving expression in the central nervous system (CNS). Recently, Melo et al. identified a translocation disrupting the CTNNA2 TAD in a patient with ID and DD116. Although homozygous variants in this gene cause cortical dysplasia and other brain malformations, a noncoding disease mechanism has not yet been investigated. Furthermore, putatively disease-causing SVs have been found disrupting the noncoding regions surrounding RAP1A (Kabuki syndrome), PPP3CA (epilepsy and ID), RAC1 (ID), PAFAH1B1 (lissencephaly), ALX4 (Potocki–Shaffer syndrome), FOXP2 (speech and language disorder), and TGFB2 (~ID)62,117.

CONCLUSION AND PERSPECTIVES

The cases discussed above clearly demonstrate the importance of considering noncoding effects when interpreting SVs in the context of NDDs. Studying these cases and experimentally investigating gene regulation within these loci could greatly improve our understanding of the noncoding disease mechanisms at play, ultimately benefiting the medical interpretation of structural variation. For many, however, the underlying noncoding disease mechanism has not yet been fully resolved. This is especially true for the SVs putatively disturbing long-range gene regulation and/or 3D chromatin structure. In many of these cases a disruption of communication between the promoter and its cognate enhancer sequences, either due to enhancer deletion or relocation, is thought to be the causal mechanism. However, (part of) the effect may also be caused by the acquisition of new interactions as a result of TAD fusion or reshuffling. Further experimental validation will be needed to ascertain how these different mechanisms contribute to the disease phenotype.

There are some loci for which the effect of noncoding SVs has been extensively studied, especially in the context of limb malformations12. These studies have demonstrated that the effects are locus-dependent and are therefore difficult to predict. This clearly exemplifies the complexity of the gene regulatory landscape and our imperfect understanding of the role, determinants, and necessity of the 3D chromatin structure. Also hampering the functional assessment of structural variation is the incomplete annotation of the noncoding genome on a tissue-specific level. Although several large studies have predicted the presence of functional elements throughout the genome for a variety of tissues and cell types, these predictions are based on biochemical properties (e.g., TF binding, open chromatin, histone modifications) and do not guarantee that these sequences perform a regulatory function in vivo. As a result, most of these putative functional elements still require experimental validation, be it via high-throughput screening assays (e.g., ChIP/ATAC-STARR-seq118,119 or CRISPR(i) screening120) or in focused studies.

A more complete annotation of the noncoding genome and a better understanding of different noncoding disease mechanisms will also improve our ability to predict the transcriptional and phenotypic consequences of newly identified, noncoding SVs. This is not only true for the large de novo SVs discussed here, but also for (combinations of) inherited variants with smaller individual effects. Moreover, while this review focused on structural variation, an enormous challenge is looming ahead to interpret the millions of noncoding SNVs identified in patient genomes as well. In contrast to the relatively large SVs, most of these single-nucleotide changes are likely to have no (or very small) functional effect(s), rendering the prediction of variant effects, prioritization, and validation possibly even more crucial.

A multiomics approach has been proposed to unify variant detection and interpretation, by combining information on a genomic, epigenomic, transcriptomic, and even functional genomic level121. This could be achieved, for example, by simultaneously implementing GS for variant identification, Hi-C analysis of 3D chromatin structure, RNA-seq profiling of transcriptional activity, and high-throughput assays for the functional validation of putative regulatory elements. Such an approach should ultimately aid in closing the gap in the genetic diagnosis of NDD patients, while at the same time improving our understanding of gene regulatory mechanisms.

References

  1. 1.

    The 1000 Genomes Project Consortium. A global reference for human genetic variation. Nature. 2015;526:68–74.

    Google Scholar 

  2. 2.

    Chaisson MJP, et al. Multi-platform discovery of haplotype-resolved structural variation in human genomes. Nat Commun. 2019;10:1784.

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  3. 3.

    Collins RL, et al. A structural variation reference for medical and population genetics. Nature. 2020;581:444–451.

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  4. 4.

    Ho SS, Urban AE, Mills RE. Structural variation in the sequencing era. Nat Rev Genet. 2020;21:171–189.

    PubMed  Article  CAS  PubMed Central  Google Scholar 

  5. 5.

    De Coster W, Van Broeckhoven C. Newest methods for detecting structural variations. Trends Biotechnol. 2019;37:973–982.

    PubMed  Article  CAS  PubMed Central  Google Scholar 

  6. 6.

    Wilfert AB, Sulovari A, Turner TN, Coe BP, Eichler EE. Recurrent de novo mutations in neurodevelopmental disorders: properties and clinical implications. Genome Med. 2017;9:101.

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  7. 7.

    Marshall CR, et al. Contribution of copy number variants to schizophrenia from a genome-wide study of 41,321 subjects. Nat Genet. 2017;49:27–35.

    PubMed  Article  CAS  PubMed Central  Google Scholar 

  8. 8.

    Coe BP, et al. Refining analyses of copy number variation identifies specific genes associated with developmental delay. Nat Genet. 2014;46:1063–1071.

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  9. 9.

    Brandler WM, et al. Paternally inherited cis-regulatory structural variants are associated with autism. Science. 2018;360:327–331.

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  10. 10.

    Turner TN, et al. Genome sequencing of autism-affected families reveals disruption of putative noncoding regulatory DNA. Am J Hum Genet. 2016;98:58–74.

    PubMed  Article  CAS  PubMed Central  Google Scholar 

  11. 11.

    Kaminsky EB, et al. An evidence-based approach to establish the functional and clinical significance of copy number variants in intellectual and developmental disabilities. Genet Med. 2011;13:777–784.

    PubMed  PubMed Central  Article  Google Scholar 

  12. 12.

    Spielmann M, Lupiáñez DG, Mundlos S. Structural variation in the 3D genome. Nat Rev Genet. 2018;19:453–467.

    PubMed  Article  CAS  PubMed Central  Google Scholar 

  13. 13.

    Nord AS, Pattabiraman K, Visel A, Rubenstein JLR. Genomic perspectives of transcriptional regulation in forebrain development. Neuron. 2015;85:27–47.

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  14. 14.

    ENCODE. An integrated encyclopedia of DNA elements in the human genome. Nature. 2012;489:57–74.

    Article  CAS  Google Scholar 

  15. 15.

    Iwakawa H, oki, Tomari Y. The functions of microRNAs: mRNA decay and translational repression. Trends Cell Biol. 2015;25:651–665.

    PubMed  Article  CAS  PubMed Central  Google Scholar 

  16. 16.

    Mahmoudi E, Cairns MJ. MiR-137: An important player in neural development and neoplastic transformation. Mol Psychiatry. 2017;22:44–55.

    PubMed  Article  CAS  PubMed Central  Google Scholar 

  17. 17.

    Elton TS, Sansom SE, Martin MM. Trisomy-21 gene dosage over-expression of miRNAs results in the haploinsufficiency of specific target proteins. RNA Biol. 2010;7:540–547.

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  18. 18.

    Briggs JA, Wolvetang EJ, Mattick JS, Rinn JL, Barry G. Mechanisms of long noncoding RNAs in mammalian nervous system development, plasticity, disease, and evolution. Neuron. 2015;88:861–877.

    PubMed  Article  CAS  PubMed Central  Google Scholar 

  19. 19.

    Li W, Notani D, Rosenfeld MG. Enhancers as noncoding RNA transcription units: recent insights and future perspectives. Nat Rev Genet. 2016;17:207–223.

    PubMed  Article  CAS  PubMed Central  Google Scholar 

  20. 20.

    Engreitz JM, et al. Local regulation of gene expression by lncRNA promoters, transcription and splicing. Nature. 2016;539:452–455.

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  21. 21.

    Derrien T, et al. The GENCODE v7 catalog of human long noncoding RNAs: analysis of their gene structure, evolution, and expression. Genome Res. 2012;22:1775–1789.

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  22. 22.

    Bond AM, et al. Balanced gene regulation by an embryonic brain ncRNA is critical for adult hippocampal GABA circuitry. Nat Neurosci. 2009;12:1020–1027.

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  23. 23.

    Cajigas I, et al. The Evf2 ultraconserved enhancer lncRNA functionally and spatially organizes megabase distant genes in the developing forebrain. Mol Cell. 2018;71:956–972.e9.

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  24. 24.

    Andersen RE, et al. The long noncoding RNA Pnky is a trans-acting regulator of cortical development In vivo. Dev Cell. 2019;49:632–642.e7.

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  25. 25.

    Ramos AD, et al. Integration of genome-wide approaches identifies lncRNAs of adult neural stem cells and their progeny in vivo. Cell Stem Cell. 2013;12:616–628.

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  26. 26.

    Pavlaki I. et al. The long non‐coding RNA Paupar promotes KAP 1‐dependent chromatin changes and regulates olfactory bulb neurogenesis. EMBO J. 2018;37:e98219.

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  27. 27.

    Kleaveland B, Shi CY, Stefano J, Bartel DP. A network of noncoding regulatory RNAs acts in the mammalian brain. Cell. 2018;174:350–362.e17.

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  28. 28.

    D’haene E, et al. Identification of long noncoding RNAs involved in neuronal development and intellectual disability. Sci Rep. 2016;6:28396.

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  29. 29.

    Ziats MN, Rennert OM. Aberrant expression of long noncoding RNAs in autistic brain. J Mol Neurosci. 2013;49:589–593.

    PubMed  Article  CAS  PubMed Central  Google Scholar 

  30. 30.

    Wang Y, et al. Genome-wide differential expression of synaptic long noncoding RNAs in autism spectrum disorder. Transl Psychiatry. 2015;5:e660.

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  31. 31.

    Liu Z, et al. Microarray profiling and co-expression network analysis of circulating lncRNAs and mRNAs associated with major depressive disorder. PLoS One. 2014;9:e93388.

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  32. 32.

    Zhang L, et al. A promoter variant in ZNF804A decreasing its expression increases the risk of autism spectrum disorder in the Han Chinese population. Transl Psychiatry. 2019;9:31.

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  33. 33.

    Coutinho AM, et al. MECP2 coding sequence and 3′UTR variation in 172 unrelated autistic patients. Am J Med Genet B. 2007;144B:475–483.

    Article  CAS  Google Scholar 

  34. 34.

    Devanna P, et al. Next-gen sequencing identifies noncoding variation disrupting miRNA-binding sites in neurological disorders. Mol Psychiatry. 2018;23:1375–1384.

    PubMed  Article  CAS  PubMed Central  Google Scholar 

  35. 35.

    Devanna P, van de Vorst M, Pfundt R, Gilissen C, Vernes SC. Genome-wide investigation of an ID cohort reveals de novo 3′UTR variants affecting gene expression. Hum Genet. 2018;137:717–721.

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  36. 36.

    Bannister AJ, Kouzarides T. Regulation of chromatin by histone modifications. Cell Res. 2011;21:381–395.

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  37. 37.

    Greenberg MVC, Bourc’his D. The diverse roles of DNA methylation in mammalian development and disease. Nat Rev Mol Cell Biol. 2019;20:590–607.

    PubMed  Article  CAS  Google Scholar 

  38. 38.

    Osterwalder M, et al. Enhancer redundancy provides phenotypic robustness in mammalian development. Nature. 2018;554:239–243.

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  39. 39.

    Visel A, et al. A high-resolution enhancer atlas of the developing telencephalon. Cell. 2013;152:895–908.

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  40. 40.

    Pattabiraman K, et al. Transcriptional regulation of enhancers active in protodomains of the developing cerebral cortex. Neuron. 2014;82:989–1003.

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  41. 41.

    Song M, et al. Mapping cis-regulatory chromatin contacts in neural cells links neuropsychiatric disorder risk variants to target genes. Nat Genet. 2019;51:1252–1262.

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  42. 42.

    Schoenfelder S, Fraser P. Long-range enhancer–promoter contacts in gene expression control. Nat Rev Genet. 2019;20:437–455.

    PubMed  Article  CAS  PubMed Central  Google Scholar 

  43. 43.

    Robson MI, Ringel AR, Mundlos S. Regulatory landscaping: how enhancer-promoter communication is sculpted in 3D. Mol Cell. 2019;74:1110–1122.

    PubMed  Article  CAS  PubMed Central  Google Scholar 

  44. 44.

    Bonev B, et al. Multiscale 3D genome rewiring during mouse neural development. Cell. 2017;171:557–572.e24.

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  45. 45.

    Freire–Pritchett P, et al. Global reorganisation of cis-regulatory units upon lineage commitment of human embryonic stem cells. eLife. 2017;6:e21926.

    PubMed  PubMed Central  Article  Google Scholar 

  46. 46.

    Benabdallah NS, et al. Decreased enhancer-promoter proximity accompanying enhancer activation. Mol Cell. 2019;76:473–484.e7.

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  47. 47.

    Alexander JM, et al. Live-cell imaging reveals enhancer-dependent Sox2 transcription in the absence of enhancer proximity. eLife. 2019;8:e41769.

    PubMed  PubMed Central  Article  Google Scholar 

  48. 48.

    Smith E, Shilatifard A. Enhancer biology and enhanceropathies. Nat Struct Mol Biol. 2014;21:210–219.

    PubMed  Article  CAS  PubMed Central  Google Scholar 

  49. 49.

    de la Torre-Ubieta L, et al. The dynamic landscape of open chromatin during human cortical neurogenesis. Cell. 2018;172:289–304.e18.

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  50. 50.

    An J-Y, et al. Genome-wide de novo risk score implicates promoter variation in autism spectrum disorder. Science. 2018;362:eaat6576.

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  51. 51.

    Short PJ, et al. De novo mutations in regulatory elements in neurodevelopmental disorders. Nature. 2018;555:611–616.

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  52. 52.

    Perenthaler E, Yousefi S, Niggl E, Barakat TS. Beyond the exome: the noncoding genome and enhancers in neurodevelopmental disorders and malformations of cortical development. Front Cell Neurosci. 2019;13:352.

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  53. 53.

    Carullo NVN, Day JJ. Genomic enhancers in brain health and disease. Genes. 2019;10:43.

    PubMed Central  Article  CAS  Google Scholar 

  54. 54.

    Nord AS, West AE. Neurobiological functions of transcriptional enhancers. Nat Neurosci. 2020;23:5–14.

    PubMed  Article  CAS  PubMed Central  Google Scholar 

  55. 55.

    Bonev B, Cavalli G. Organization and function of the 3D genome. Nat Rev Genet 2016;17:661–678.

    PubMed  Article  CAS  PubMed Central  Google Scholar 

  56. 56.

    Dixon JR, et al. Chromatin architecture reorganization during stem cell differentiation. Nature. 2015;518:331–336.

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  57. 57.

    Rajarajan P. et al. Neuron-specific signatures in the chromosomal connectome associated with schizophrenia risk. Science. 2018;362:eaat4311.

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  58. 58.

    Gregor A, et al. De novo mutations in the genome organizer CTCF cause intellectual disability. Am J Hum Genet. 2013;93:124–131.

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  59. 59.

    Gabriele M, et al. YY1 haploinsufficiency causes an intellectual disability syndrome featuring transcriptional and chromatin dysfunction. Am J Hum Genet. 2017;100:907–925.

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  60. 60.

    Lehalle D, et al. STAG1 mutations cause a novel cohesinopathy characterised by unspecific syndromic intellectual disability. J Med Genet. 2017;54:479–488.

    PubMed  Article  PubMed Central  Google Scholar 

  61. 61.

    Liu J, Krantz I. Cornelia de Lange syndrome, cohesin, and beyond. Clin Genet. 2009;76:303–314.

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  62. 62.

    Middelkamp S, et al. Prioritization of genes driving congenital phenotypes of patients with de novo genomic structural variants. Genome Med. 2019;11:79.

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  63. 63.

    Gasperini M, Tome JM, Shendure J. Towards a comprehensive catalogue of validated and target-linked human enhancers. Nat Rev Genet. 2020;21:292–310.

    PubMed  Article  CAS  PubMed Central  Google Scholar 

  64. 64.

    Hu WF, Chahrour MH, Walsh CA. The diverse genetic landscape of neurodevelopmental disorders. Annu Rev Genomic Hum Genet 2014;15:195–213.

    Article  CAS  Google Scholar 

  65. 65.

    Vissers LELM, Gilissen C, Veltman JA. Genetic studies in intellectual disability and related disorders. Nat Rev Genet 2016;17:9–18.

    PubMed  Article  CAS  PubMed Central  Google Scholar 

  66. 66.

    Gilissen C, et al. Genome sequencing identifies major causes of severe intellectual disability. Nature. 2014;511:344–347.

    PubMed  Article  CAS  PubMed Central  Google Scholar 

  67. 67.

    Gaugler T, et al. Most genetic risk for autism resides with common variation. Nat Genet. 2014;46:881–885.

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  68. 68.

    Turner TN, Eichler EE. The role of de novo noncoding regulatory mutations in neurodevelopmental disorders. Trends Neurosci. 2019;42:115–127.

    PubMed  Article  CAS  PubMed Central  Google Scholar 

  69. 69.

    Redin C, et al. The genomic landscape of balanced cytogenetic abnormalities associated with human congenital anomalies. Nat Genet. 2017;49:36–45.

    PubMed  Article  CAS  PubMed Central  Google Scholar 

  70. 70.

    Han L, et al. Functional annotation of rare structural variation in the human brain. Nat Commun. 2020;11:2990.

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  71. 71.

    Doan RN, et al. Mutations in human accelerated regions disrupt cognition and social behavior. Cell. 2016;167:341–354.e12.

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  72. 72.

    Monlong J, et al. Global characterization of copy number variants in epilepsy patients from whole genome sequencing. PLoS Genet. 2018;14:e1007285.

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  73. 73.

    Chiang C, et al. The impact of structural variation on human gene expression. Nat Genet. 2017;49:692–699.

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  74. 74.

    Lupiáñez DG, et al. Disruptions of topological chromatin domains cause pathogenic rewiring of gene-enhancer interactions. Cell. 2015;161:1012–1025.

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  75. 75.

    Despang A, et al. Functional dissection of the Sox9–Kcnj2 locus identifies nonessential and instructive roles of TAD architecture. Nat Genet. 2019;51:1263–1271.

    PubMed  Article  CAS  PubMed Central  Google Scholar 

  76. 76.

    Sutherland HF, et al.Identification of a novel transcript disrupted by a balanced translocation associated with DiGeorge syndrome. Am J Hum Genet. 1996;59:23–31.

    PubMed  PubMed Central  CAS  Google Scholar 

  77. 77.

    Meng Q, et al. The DGCR5 long noncoding RNA may regulate expression of several schizophrenia-related genes. Sci Transl Med. 2018;10:eaat6912.

    PubMed  PubMed Central  Article  Google Scholar 

  78. 78.

    Ang CE, et al. The novel lncRNA lnc-NR2F1 is pro-neurogenic and mutated in human neurodevelopmental disorders. eLife. 2019;8:e41770.

    PubMed  PubMed Central  Article  Google Scholar 

  79. 79.

    Ross PJ. et al. Synaptic dysfunction in human neurons with autism-associated deletions in PTCHD1-AS. Biol Psychiatry. 2020;87:139–149.

    PubMed  Article  CAS  PubMed Central  Google Scholar 

  80. 80.

    Cappuccio G, et al. Microdeletion of pseudogene chr14.232.a affects LRFN5 expression in cells of a patient with autism spectrum disorder. Eur J Hum Genet. 2019;27:1475–1480.

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  81. 81.

    Talkowski ME. et al.Disruption of a large intergenic noncoding RNA in subjects with neurodevelopmental disabilities. Am J Hum Genet. 2012;91:1128–1134.

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  82. 82.

    Dornelles-Wawruk H, et al. A balanced reciprocal translocation t(2;9)(p25;q13) disrupting the LINC00299 gene in a patient with intellectual disability. Mol Syndromol. 2019;10:234–238.

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  83. 83.

    Ng S-Y, Johnson R, Stanton LW. Human long noncoding RNAs promote pluripotency and neuronal differentiation by association with chromatin modifiers and transcription factors. EMBO J. 2012;31:522–533.

    PubMed  Article  CAS  PubMed Central  Google Scholar 

  84. 84.

    Ng S-Y, Bogu GK, Soh BS, Stanton LW. The long noncoding RNA RMST interacts with SOX2 to regulate neurogenesis. Mol Cell. 2013;51:349–359.

    PubMed  Article  CAS  PubMed Central  Google Scholar 

  85. 85.

    Stamou M, et al. A balanced translocation in Kallmann syndrome implicates a long noncoding RNA, RMST, as a GnRH neuronal regulator. J Clin Endocrinol Metab. 2020;105:e231–e244.

    Article  Google Scholar 

  86. 86.

    Kagami M, et al. Deletions and epimutations affecting the human 14q32.2 imprinted region in individuals with paternal and maternal upd(14)-like phenotypes. Nat Genet. 2008;40:237–242.

    PubMed  Article  CAS  PubMed Central  Google Scholar 

  87. 87.

    Bieth E, et al. Highly restricted deletion of the SNORD116 region is implicated in Prader–Willi syndrome. Eur J Hum Genet. 2015;23:252–255.

    PubMed  Article  CAS  PubMed Central  Google Scholar 

  88. 88.

    Burnett LC, et al. Loss of the imprinted, noncoding Snord116 gene cluster in the interval deleted in the Prader Willi syndrome results in murine neuronal and endocrine pancreatic developmental phenotypes. Hum Mol Genet. 2017;26:4606–4616.

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  89. 89.

    Coulson RL, et al. Prader–Willi locus Snord116 RNA processing requires an active endogenous allele and neuron-specific splicing by Rbfox3/NeuN. Hum Mol Genet. 2018;27:4051–4060.

    PubMed  PubMed Central  CAS  Google Scholar 

  90. 90.

    Poeta L, Drongitis D, Verrillo L, Miano MG. Dna hypermethylation and unstable repeat diseases: A paradigm of transcriptional silencing to decipher the basis of pathogenic mechanisms. Genes (Basel). 2020;11:1–18.

    Article  CAS  Google Scholar 

  91. 91.

    Sun JH, et al. Disease-associated short tandem repeats co-localize with chromatin domain boundaries. Cell. 2018;175:224–238.e15.

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  92. 92.

    LaCroix AJ, et al. GGC repeat expansion and exon 1 methylation of XYLT1 is a common pathogenic variant in Baratela–Scott syndrome. Am J Hum Genet. 2019;104:35–44.

    PubMed  Article  CAS  PubMed Central  Google Scholar 

  93. 93.

    Roessler E, et al. Cytogenetic rearrangements involving the loss of the Sonic Hedgehog gene at 7q36 cause holoprosencephaly. Hum Genet. 1997;100:172–181.

    PubMed  Article  CAS  PubMed Central  Google Scholar 

  94. 94.

    Benabdallah NS, et al. SBE6: a novel long-range enhancer involved in driving sonic hedgehog expression in neural progenitor cells. Open Biol. 2016;6:160197.

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  95. 95.

    Visel A, Minovitsky S, Dubchak I, Pennacchio LA. VISTA Enhancer Browser—a database of tissue-specific human enhancers. Nucleic Acids Res. 2007;35:D88–D92.

    PubMed  Article  CAS  PubMed Central  Google Scholar 

  96. 96.

    Wang Y, et al. The 3D Genome Browser: a web-based browser for visualizing 3D genome organization and long-range chromatin interactions. Genome Biol. 2018;19:151.

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  97. 97.

    Mehrjouy MM, et al. Regulatory variants of FOXG1 in the context of its topological domain organisation. Eur J Hum Genet. 2018;26:186–196.

    PubMed  Article  CAS  PubMed Central  Google Scholar 

  98. 98.

    D’haene E, et al. A neuronal enhancer network upstream of MEF2C is compromised in patients with Rett-like characteristics. Hum Mol Genet. 2019;28:818–827.

    PubMed  Article  CAS  PubMed Central  Google Scholar 

  99. 99.

    FitzPatrick DR, et al. Identification of SATB2 as the cleft palate gene on 2q32-q33. Hum Mol Genet. 2003;12:2491–2501.

    PubMed  Article  CAS  PubMed Central  Google Scholar 

  100. 100.

    Rainger JK, et al. Disruption of SATB2 or its long-range cis-regulation by SOX9 causes a syndromic form of Pierre Robin sequence. Hum Mol Genet. 2014;23:2569–2579.

    PubMed  Article  CAS  PubMed Central  Google Scholar 

  101. 101.

    Gordon CT, et al. Identification of novel craniofacial regulatory domains located far upstream of SOX9 and disrupted in Pierre Robin sequence. Hum Mutat. 2014;35:1011–1020.

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  102. 102.

    Franke M, et al. Formation of new chromatin domains determines pathogenicity of genomic duplications. Nature. 2016;538:265–269.

    PubMed  Article  CAS  PubMed Central  Google Scholar 

  103. 103.

    Ibrahim DM, Mundlos S. Three-dimensional chromatin in disease: what holds us together and what drives us apart? Curr Opin Cell Biol. 2020;64:1–9.

    PubMed  Article  CAS  PubMed Central  Google Scholar 

  104. 104.

    Bunyan DJ, et al. X-linked dominant congenital ptosis cosegregating with an interstitial insertion of a chromosome 1p21.3 fragment into a quasipalindromic sequence in Xq27.1. Open J Genet. 2014;04:415–425.

    Article  Google Scholar 

  105. 105.

    Haines B, et al. Interchromosomal insertional translocation at Xq26.3 alters SOX3 expression in an individual with XX male sex reversal. J Clin Endocrinol Metab. 2015;100:E815–E820.

    PubMed  Article  CAS  PubMed Central  Google Scholar 

  106. 106.

    Sutton E, et al. Identification of SOX3 as an XX male sex reversal gene in mice and humans. J Clin Invest. 2011;121:328–341.

    PubMed  Article  CAS  PubMed Central  Google Scholar 

  107. 107.

    Birnbaum RY, et al. Functional characterization of tissue-specific enhancers in the DLX5/6 locus. Hum Mol Genet. 2012;21:4930–4938.

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  108. 108.

    Johnson KR, et al. Deletion of a long-range Dlx5 enhancer disrupts inner ear development in mice. Genetics. 2018;208:1165–1179.

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  109. 109.

    Rasmussen MB, et al. Phenotypic subregions within the split-hand/foot malformation 1 locus. Hum Genet. 2016;135:345–357.

    PubMed  Article  PubMed Central  Google Scholar 

  110. 110.

    Idrees F, et al. A novel homeobox mutation in the PITX2 gene in a family with Axenfeld-Rieger syndrome associated with brain, ocular, and dental phenotypes. Am J Med Genet B 2006;141B:184–191.

    Article  CAS  Google Scholar 

  111. 111.

    Trembath DG, et al. Analysis of two translocation breakpoints and identification of a negative regulatory element in patients with Rieger’s syndrome. Birth Defects Res A Clin Mol Teratol. 2004;70:82–91.

    PubMed  Article  CAS  PubMed Central  Google Scholar 

  112. 112.

    Protas ME, et al. Mutations of conserved noncoding elements of PITX2 in patients with ocular dysgenesis and developmental glaucoma. Hum Mol Genet. 2017;26:3630–3638.

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  113. 113.

    Kleinjan DA, et al. Long-range downstream enhancers are essential for Pax6 expression. Dev Biol. 2006;299:563–581.

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  114. 114.

    Laugsch M, et al. Modeling the pathological long-range regulatory effects of human structural variation with patient-specific hiPSCs. Cell Stem Cell. 2019;24:736–752.e12.

    PubMed  Article  CAS  PubMed Central  Google Scholar 

  115. 115.

    Lessel D, et al. BCL11B mutations in patients affected by a neurodevelopmental disorder with reduced type 2 innate lymphoid cells. Brain. 2018;141:2299–2311.

    PubMed  PubMed Central  Article  Google Scholar 

  116. 116.

    Melo US, et al.Hi-C identifies complex genomic rearrangements and TAD-shuffling in developmental diseases. Am J Hum Genet. 2020;106:872–884.

    PubMed  Article  CAS  PubMed Central  Google Scholar 

  117. 117.

    Feuk L, Marshall CR, Wintle RF, Scherer SW. Structural variants: changing the landscape of chromosomes and design of disease studies. Hum Mol Genet. 2006;15:R57–R66.

    PubMed  Article  CAS  PubMed Central  Google Scholar 

  118. 118.

    Barakat TS, et al. Functional dissection of the enhancer repertoire in human embryonic stem cells. Cell Stem Cell. 2018;23:276–288.e8.

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  119. 119.

    Wang X. et al. High-resolution genome-wide functional dissection of transcriptional regulatory regions and nucleotides in human. Nat Commun. 2018;9:5380.

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  120. 120.

    Gasperini M, et al. A genome-wide framework for mapping gene regulation via cellular genetic screens. Cell. 2019;176:377–390.e19.

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  121. 121.

    Hasin Y, Seldin M, Lusis A. Multi-omics approaches to disease. Genome Biol. 2017;18:1–15.

    Article  CAS  Google Scholar 

  122. 122.

    Andersson R, et al. An atlas of active enhancers across human cell types and tissues. Nature. 2014;507:455–461.

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  123. 123.

    Kundaje A, et al. Integrative analysis of 111 reference human epigenomes. Nature. 2015;518:317–330.

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  124. 124.

    Li M, et al. Integrative functional genomic analysis of human brain development and neuropsychiatric risks. Science. 2018;362:eaat7615.

    PubMed  PubMed Central  Article  CAS  Google Scholar 

Download references

Acknowledgements

This work was supported by the Research Foundation Flanders (FWO) under grant G044615N and 1520518N. In addition, E.D. is also supported by a doctoral fellowship of the FWO Research Fund and was previously supported by a fellowship of the Marguerite-Marie Delacroix Fund. We thank our colleagues Björn Menten, Annelies Dheedene, and Griet De Clercq for their valuable feedback on the manuscript.

Author information

Affiliations

Authors

Corresponding author

Correspondence to Sarah Vergult PhD.

Ethics declarations

Disclosure

The authors declare no conflicts of interest.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

D’haene, E., Vergult, S. Interpreting the impact of noncoding structural variation in neurodevelopmental disorders. Genet Med (2020). https://doi.org/10.1038/s41436-020-00974-1

Download citation

Key words

  • neurodevelopmental disorders
  • noncoding variation
  • structural variation
  • gene regulation
  • chromatin conformation

Search