1.1. Optic Neuropathies

Optic neuropathies are among the most common causes of blindness in the working age population [1], with inherited forms (including Leber Hereditary Optic Neuropathy (LHON) [2] and Dominant Optic Atrophy (DOA)) affecting about 1 in 10,000 of the population [3,4,5]. In this review, we will present a brief introduction to these conditions and how the availability of powerful emerging techniques, including transcriptomics, are quickly revolutionizing both diagnosis and development of novel therapies with potential applications beyond the eye.

Vision is ultimately lost in both LHON and DOA as retinal ganglion cells (RGCs) die secondary to mitochondrial dysfunction [6]. This specific susceptibility of RGCs to such dysfunction is not completely understood. However, the relatively large metabolic demand for these specialized cells and their unique anatomy are thought to be important contributory factors. RGCs have long axonal segments which lack myelin throughout their intraocular course but gain a myelin sheath on exiting the eye beyond the lamina cribosa [7,8]. As the only nervous tissue visible in vivo and with increasingly sophisticated cell culture techniques [9,10], RGCs present a powerful system in which to interrogate mitochondrial dysfunction and the pathways that ultimately lead to cell loss and disease development. Such dysfunction has been implicated in major neurodegenerative diseases, such as Parkinson’s disease [11], Alzheimer’s disease [12], and other forms of dementia [13], but the polygenetic inheritance and environmental contribution to these common conditions are particularly challenging when investigating their pathogenesis.

As monogenetic conditions where mitochondrial function is disturbed, both LHON and DOA can mitigate some of these challenges and act as useful model diseases of more complex neurodegenerative disease processes. LHON is a primary mitochondrial DNA (mtDNA) disorder, with the majority of cases caused by one of three point mutations—namely, m.3460G>A in the MT-ND1 gene, m.11778G>A in the MT-ND4 gene, and m.14484T>C in the MT-ND6 gene—all of which encode for essential subunits of mitochondrial complex I [3,7,14,15,16]. LHON is additionally interesting due to its predilection to manifest in males and its marked incomplete penetrance—both of which could perhaps have origins in transcriptomic differences.

In comparison, DOA is nuclear-encoded mitochondrial optic neuropathy caused by mutations in the OPA1 gene (3q28-q29), which encodes for a multimeric dynamin GTPase protein located within the mitochondrial inner membrane. OPA1 subserves a number of functions, including the regulation of mitochondrial fusion, the stability of the mitochondrial respiratory chain complexes and mitochondrial biogenesis, the sequestration of pro-apoptotic cytochrome c within the mitochondrial cristae, and mitochondrial turnover (mitophagy) [4,17,18,19,20].

1.2. “-Omics” Technologies as Applied to Inherited Optic Neuropathies

Our understanding of mitochondrial biology and disease has advanced greatly over recent years, not least due to the development and maturation of “-omics” technologies. These can be defined as “high-throughput technologies capable of detecting differences in a multitude of molecular constituents in organisms [21]”, with those that represent the three strata of central biological dogma (genomics, transcriptomics, and proteomics) being prominent. These fields deal with the detection of differences in DNA sequences, gene transcription, and proteins within tissue. Additionally, particularly relevant to mitochondrial disease is the developing field of metabolomics (centered on the comparison of levels of products of metabolism) [22,23] and lipidomics [24].

Whilst these disciplines are linked by their comparative nature—experimental plans often involve the contrast of different conditions (e.g., control and “diseased” states, or between different cell types)—the emerging field of multi-omics (or vertical -omics) focuses on complimentary comparisons across domains (Figure 1). For example, highlighting changes replicated across the transcriptome, the proteome and metabolome will carry particular significance [24,25], and this approach is already being used in mitochondrial research [26]. As these technologies and their complementary bioinformatic analysis techniques develop, the power of “-omics” investigations is likely to increase.

Figure 1. A schematic representing the processes of modeling and investigating inherited optic neuropathies using “-omics” methods. (A) Model diseases such as DOA and LHON can be powerful tools in which to investigate the effects of mitochondrial dysfunction. The study of patients with inherited optic neuropathies is often a two-way process: in one direction, the characterization of their phenotype and genotype allows the development of useful disease models (for example, cell -and animal-based). These provide an efficient environment in which to increase our understanding of underlying disease processes as well as a testing ground for novel therapies before their return in the opposite direction, back into patients. (B) Tissues from model systems can be investigated using used in multiple “-omics” techniques—in some cases, these can be performed simultaneously and analyzed in vertical “multi-omics” experiments (see text). ATACseq—Assay for Transposase Accessible Chromatin—a method for assessing which areas of the chromatin superstructure are open and so likely available for active in transcription, is an example of the rapidly expanding field of epigenomics. (C) Multiple transcriptomic techniques are available and have particular strength in different experimental situations. “Novel”: Novel techniques are emerging that will further the resolution of these techniquesm including those with abilities to sequence longer transcripts in one read and those that integrate temporal and spatial information regarding transcripts (see text). RNA-seq—RNA-sequencing; scRNA-seq—single-cell RNAseq; snRNAseq—single nucleus RNAseq. (D) Transcriptomic techniques gain power from the large quantity of data that they produce. This necessitates adequately designed bioinformatic pipelines that are tailored to the exact scientific question being asked in order to produce a list of candidate genes for further investigation back in model systems. (E) Model systems of diseases can include those based on cultured patient cells or be animals carrying pathogenic variants, leading to phenocopies of human disease. Samples from these models used in transcriptomic analysis can include tissue (such as retinas) or cell cultures. As many transcriptomic experiments compare expression between conditions (e.g., disease and control) to produce lists of differentially expressed genes, the further technical and functional validation of these can be performed back in model systems in preparation for therapeutic translation in patients (see text).

1.3. Transcriptomics

Specifically, “the transcriptome” refers to the RNA transcribed within a cell, or group of cells, often with a particular focus on mRNA (both coding and non-coding). Several methods to quantify mRNA have been developed (see below) and are applied to genetic eye diseases. For example, assessing transcribed features in a particular sample can be used to compare changes in gene expression either over time or between control and diseased states (such as optic neuropathies) [27]. Features showing differential expression may be implicated in the disease process, highlighting areas for further investigation to uncover aetiologic pathways, novel biomarkers, and therapeutic targets.

2.1. Applications of Transcriptomics in Optic Neuropathies

Several factors make transcriptomics a particularly suitable methodology for the investigation of optic neuropathies. Whilst the anatomy and physiology of the retina and optic nerve are relatively well understood compared with other areas of the central nervous system [28], our understanding of pathophysiology in these structures can be less comprehensive. This is particularly true for inherited optic neuropathies such as LHON and DOA, where the genomic determinant of the disease in many patients is readily identifiable as a single gene (monogenic disorder) [14,18,19]. However, less is known regarding how this translates into the clinical phenotype of RGC death as well as other as yet unexplained facets of these model diseases (such as the incomplete penetrance in LHON when the pathogenic mitochondrial DNA mutation is present in the homoplastic state in both affected patients and carriers). Thus, this presents an unmet need to identify the novel pathways and genes involved to which comparative transcriptomics is particularly suited. Whilst direct access to RGCs and patient tissues is limited, cellular and mouse model systems [7,29,30] have developed in recent years into powerful platforms with which to perform transcriptomic studies (and, more importantly, validate and investigate their findings). For example, in vivo neuro-retinal tissue can easily be visualized (if not directly sampled) at the cellular level with techniques such as optical coherence tomography (OCT), and there are well defined metrics of RGC function at the behavioral (acuity) [31], reflex (pupillary) [32], and electrophysiological levels [33]. To compliment this, induced pluripotent stem cell (iPSC) RGC models derived from LHON and DOA patient tissues have proved invaluable for molecular investigations [7].