Neurodevelopmental copy-number variants: A roadmap to improving outcomes by uniting patient advocates, researchers, and clinicians for collective impact

Abstract

Copy-number variants and structural variants (CNVs/SVs) drive many neurodevelopmental-related disorders. While many neurodevelopmental-related CNVs/SVs give rise to complex phenotypes, the overlap in phenotypic presentation between independent CNVs can be extensive and provides a motivation for shared approaches. This confluence at the level of clinical phenotype implies convergence in at least some aspects of the underlying genomic mechanisms. With this perspective, our Commission on Novel Technologies for Neurodevelopmental CNVs asserts that the time has arrived to approach neurodevelopmental-related CNVs/SVs as a class of disorders that can be identified, investigated, and treated on the basis of shared mechanisms and/or pathways (e.g., molecular, neurological, or developmental). To identify common etiologic mechanisms among uncommon neurodevelopmental-related disorders and to potentially identify common therapies, it is paramount for teams of scientists, clinicians, and patients to unite their efforts. We bring forward novel, collaborative, and integrative strategies to translational CNV/SV research that engages diverse stakeholders to help expedite therapeutic outcomes. We articulate a clear vision for piloted roadmap strategies to reduce patient/caregiver burden and redundancies, increase efficiency, avoid siloed data, and accelerate translational discovery across CNV/SV-based syndromes.

Keywords: CNVs; biobank; community engagement; copy-number variants; genomic disorders; iPSCs; inclusion; infrastructure; long-read sequencing; neurodevelopment; neurological; patient centered; patient led; structural variants; systematic phenotyping; team science.

Figure 1

Infrastructure and roadmap to facilitate cross-syndrome studies (A) Shared infrastructure to support aggregation, access, and analysis of shared data through open access. (B) Research roadmap to translation for neurodevelopmental-related CNVs/SVs starting with three distinct yet similar CNVs/SVs (dup15q, 8p, and ring 14). Abbreviations include EBV (Epstein-Barr virus DNA), FCLs (fibroblast cell lines), and PBMCs (peripheral blood mononuclear cells).

Figure 2

Mechanistic models of pathogenic CNVs (A) Primary driver model: dosage sensitivity of a gene or genes encompassed within the CNV is the leading hypothesis underlying CNV pathogenicity. In the simplest scenario, altered dosage of a single gene may contribute to all or many phenotypes. For example, in 22q13 (Phelan-McDermid) and 15q11 (Angelman syndrome), the majority of defects seem to be due to haploinsufficiency of SHANK3 and UBE3A, respectively. Emerging data from a systematic approach testing constraints on haploinsufficiency and triplosensitivity across the genome suggests that the phenotypes associated with roughly 1/3 of recurrent CNVs are produced by a single primary driver gene. (B) Multiple driver model: one or more genes at a CNV locus are each responsible for discrete phenotypes. For example, in Williams syndrome (7q11.23Del), LIMK1 is proposed to be responsible for visuospatial deficits whereas ELN has been linked to cardiovascular phenotypes.^, Importantly, in both this paradigm and the primary driver model, restoration of expression levels of only one gene should be sufficient to ameliorate acute phenotypes. (C) Cis-interaction model: haploinsufficiency of multiple genes within a CNV locus may be required to produce a single given phenotype (“cis-interaction model”). This seems to be the case at 16p11.2, where multiple genes are involved in craniofacial abnormalities. (D) Trans-interaction model: a fourth possibility is trans-interaction. In one scenario, trans-interactions could imply that phenotypes associated with a CNV only emerge in specific genetic backgrounds, most likely because of polygenic risk load or the presence of secondary rare disruptive gene variants. This scenario is observed in some cases of inherited CNVs where the full phenotype is not expressed and can even go undetected in the parent carrier. In another scenario, the change in dosage or arrangement of a gene regulatory element within the CNV locus impacts the expression of genes outside the locus. In this case, the manifestation of phenotypes is not dependent on a change in the dosage of a protein-coding gene. This model may explain ring chromosome or complex inversions and deletions/duplications found on chromosome 8p. These four models are not mutually exclusive, and it is likely that complex interactions are a feature of many CNVs that show variable phenotypic expressivity.