Functional genomics and small molecules in mitochondrial neurodevelopmental disorders

Abstract

Mitochondria are critical for brain development and homeostasis. Therefore, pathogenic variation in the mitochondrial or nuclear genome which disrupts mitochondrial function frequently results in developmental disorders and neurodegeneration at the organismal level. Large-scale application of genome-wide technologies to individuals with mitochondrial diseases has dramatically accelerated identification of mitochondrial disease-gene associations in humans. Multi-omic and high-throughput studies involving transcriptomics, proteomics, metabolomics, and saturation genome editing are providing deeper insights into the functional consequence of mitochondrial genomic variation. Integration of deep phenotypic and genomic data through allelic series continues to uncover novel mitochondrial functions and permit mitochondrial gene function dissection on an unprecedented scale. Finally, mitochondrial disease-gene associations illuminate disease mechanisms and thereby direct therapeutic strategies involving small molecules and RNA-DNA therapeutics. This review summarizes progress in functional genomics and small molecule therapeutics in mitochondrial neurodevelopmental disorders.

Keywords: Functional genomics; Mitochondrial disease; Neurodevelopmental disorders; Small molecules; Therapeutics.

Fig. 1

Physiological functionsof mitochondria and examples of mitochondrial dysfunction in neurodevelopmental disorders. Physiologic functions of mitochondria are shown in the left column. Examples of mitochondrial dysfunction responsible for mitochondrial neurodevelopmental disorders are shown in the right column. Figure created using Biorender.com.

Fig. 2

Genomic approaches to mitochondrial neurodevelopmental disorders. A) Short-read sequencing (SRS, also known as next generation sequencing, NGS) is the most common technology used in genomic sequencing. In SRS, sequencing data consists of small (50–150 bp) sequences called reads. Short-read data must be computationally assembled to generate genomic sequences. Gaps in genome coverage (gray rectangles) in SRS can limit resolution of complex genomic regions (CGR) and structural variants (SV). In contrast, long-read sequencing technologies generate kilobase to megabase sequencing reads. Long sequencing reads provide better coverage without gaps across the genome. They improve SV and CGR resolution and played an integral role in the Telomere-to-Telomere (T2T) project. B) Exome sequencing (ES) involves the selective capture and sequencing of the exome. The exome is composed of all protein-coding genes and makes up ∼1 ​% of the genome. Capture regions (purple rectangles) often include flanking intronic regions since intronic genetic variants near splicing junctions often cause splicing dysfunction. Genome sequencing (GS) involves sequencing of the entire genome including introns, promoters, and intergenic regions including regulatory elements like enhancers and suppressors. C) The advantages and disadvantages of SRS, LRS, ES, and GS are shown. SVs ​= ​structural variants; CGRs ​= ​complex genomic regions; SNVs ​= ​single nucleotide variants; indels ​= ​small insertion and deletions. Figure created using Biorender.com.

Fig. 3

Transcriptomic approaches to mitochondrial neurodevelopmental disorders. RNA sequencing (RNA-seq) can identify gene expression outliers, splicing outliers, and allelic imbalance/dropout within cells, tissues, or other biospecimens. A) Normal gene expression within cells or tissues is established by performing RNA-seq on control samples (blue bars). Each bar represents gene expression levels within a sample. Gene expression within the sample (red bar) is found to be an outlier. The lower expression within the sample indicates that a genetic variant reducing gene expression may be present within the gene or gene regulatory elements. B) RNA splicing results in the removal of intronic sequences and the splicing of exons (colored rectangles). Abnormal splicing patterns (splicing outliers) may result from pathogenic splicing variants; these can be detected in comparison to control samples. The pathogenic splicing variant in the figure results in exon skipping in half of transcripts (yellow box). C) The relative expression of two alleles at a particular locus can be quantified by examining how many reads contain a single nucleotide variant or polymorphism versus the reference sequence. Pathogenic variation within a gene or gene regulatory element that reduce one allele's expression will result in allelic imbalance or dropout as shown in the figure. Figure created using Biorender.com.

Fig. 4

Proteomic approaches to mitochondrial neurodevelopmental disorders. Summary of a bottom-up proteomic study. Proteomics can be performed on biospecimens, tissue samples, or cell lines (sample selection). Samples are then enzymatically or chemically digested to generate short peptides. These peptides then undergo MS, LC-MS, or HPLC-MS. Raw data then undergoes peptide-spectrum matching to identify the sequence of each peptide. Individual peptides are then quantified, and peptide fragment sequences are compared with the sequences of known proteins; the quantified proteins are then biologically interpreted. Figure created using Biorender.com.

Fig. 5

Saturation Genome Editing and Multiplexed Assays of Variant Effect. Saturation Genome Editing (SGE) involves the introduction of all possible single nucleotide variants (SNVs) and indels into a gene or locus of interest (top half) into cell lines using CRISPR-Cas9. The experimental process is described in the bottom half of the figure. The cell line HAP1 is often used for genome editing since it contains a haploid genome. An SNV or indel library and CRISPR-Cas9 is introduced into cells to induce homology directed repair. Each cell acquires a single mutation. If the gene is essential for cellular survival, genetic variants resulting in a null allele will cause cell death. The cell population is subsequent sequenced by ES or GS. The dropout of variants within the population indicates that the variant is pathogenic. The degree of variant depletion is calculated and used to generate a functional score. Multiplexed assays of variant effect (MAVE) thus consist of the functional score of each variant assayed in an SGE study.

Fig. 6

Functional annotation of the human genome through global data aggregation. The integration and aggregation of individual-level genomic and phenotypic data derived from health records (top half) on a global scale is necessary to provide full functional annotation of the human genome. Such efforts are necessary to fully realize the promise of precision medicine through accurate prediction of mitochondrial dysfunction and mitochondrial neurodevelopmental disorders from an individual's genotype (bottom half). Figure created using Biorender.com.