Mitochondrial optic neuropathies - disease mechanisms and therapeutic strategies

Abstract

Leber hereditary optic neuropathy (LHON) and autosomal-dominant optic atrophy (DOA) are the two most common inherited optic neuropathies in the general population. Both disorders share striking pathological similarities, marked by the selective loss of retinal ganglion cells (RGCs) and the early involvement of the papillomacular bundle. Three mitochondrial DNA (mtDNA) point mutations; m.3460G>A, m.11778G>A, and m.14484T>C account for over 90% of LHON cases, and in DOA, the majority of affected families harbour mutations in the OPA1 gene, which codes for a mitochondrial inner membrane protein. Optic nerve degeneration in LHON and DOA is therefore due to disturbed mitochondrial function and a predominantly complex I respiratory chain defect has been identified using both in vitro and in vivo biochemical assays. However, the trigger for RGC loss is much more complex than a simple bioenergetic crisis and other important disease mechanisms have emerged relating to mitochondrial network dynamics, mtDNA maintenance, axonal transport, and the involvement of the cytoskeleton in maintaining a differential mitochondrial gradient at sites such as the lamina cribosa. The downstream consequences of these mitochondrial disturbances are likely to be influenced by the local cellular milieu. The vulnerability of RGCs in LHON and DOA could derive not only from tissue-specific, genetically-determined biological factors, but also from an increased susceptibility to exogenous influences such as light exposure, smoking, and pharmacological agents with putative mitochondrial toxic effects. Our concept of inherited mitochondrial optic neuropathies has evolved over the past decade, with the observation that patients with LHON and DOA can manifest a much broader phenotypic spectrum than pure optic nerve involvement. Interestingly, these phenotypes are sometimes clinically indistinguishable from other neurodegenerative disorders such as Charcot-Marie-Tooth disease, hereditary spastic paraplegia, and multiple sclerosis, where mitochondrial dysfunction is also thought to be an important pathophysiological player. A number of vertebrate and invertebrate disease models has recently been established to circumvent the lack of human tissues, and these have already provided considerable insight by allowing direct RGC experimentation. The ultimate goal is to translate these research advances into clinical practice and new treatment strategies are currently being investigated to improve the visual prognosis for patients with mitochondrial optic neuropathies.

Fig. 1

The mitochondrial respiratory chain and oxidative phosphorylation. Reproduced with permission from Nijtmans et al. (2004).

Fig. 2

The human mitochondrial genome. Protein coding (yellow), rRNA (red), and tRNA (purple) genes are depicted on the heavy (H-, outer) and light (L-, inner) strands. The 22 tRNAs are indicated by their cognate amino acid letter code and the 2 rRNAs by their sedimentation coefficients (12S and 16S). The origins of mtDNA replication and the direction of synthesis are denoted by O_H for the H-strand, and O_L for the L-strand.

Fig. 3

Mitochondrial and nuclear-encoded subunits of the mitochondrial respiratory chain complexes.

Fig. 4

Skeletal muscle sections illustrating the characteristic histochemical features of mitochondrial dysfunction: (A) Ragged-red muscle fibre identified using the modified Gomori trichome stain. The red component of the staining mixture is selectively sequestered by mitochondria, which have accumulated in the subsarcolemmal region, giving the fibre an irregular red outline, (B) Serial section of the same muscle fibre after SDH staining. This is a more specific assay for detecting the subsarcolemmal accumulation of mitochondria, SDH being a specific marker for complex II activity, (C) Abnormal COX–SDH histochemistry from a patient with chronic progressive external ophthalmoplegia (CPEO) due to a single 5 kb mtDNA deletion, showing normal COX-positive (Brown) and energy deficient, COX-negative (Blue) muscle fibres.

Fig. 5

The minimum prevalence of inherited optic nerve disorders in the North of England.

Fig. 6

Summary of linkage studies investigating the existence of putative LHON nuclear modifiers on the X-chromosome. A nonparametric LOD score (NPL) >2 is indicative of significant linkage, and these chromosomal areas possibly harbour susceptibility loci which influence the risk of visual loss among LHON carriers. The different studies are colour coded: red (Hudson et al., 2005), blue (Shankar et al., 2008), and black (Ji et al., 2010). Reproduced with permission from Ji et al. (2010).

Fig. 7

The complex interaction of genetic, hormonal, and environmental factors in the pathophysiology of LHON.

Fig. 8

The optic disc appearance of a patient with a confirmed pathogenic OPA1 mutation showing pallor of the neuro-retinal rim, which is more marked temporally. The bottom panels illustrate the pattern of retinal nerve fibre layer (RNFL) thinning seen in patients with OPA1 mutations, with relative sparing of the nasal peripapillary quadrant. The RNFL profile for each eye is superimposed on the normal distribution percentiles, and compared with each other (Bottom left panel). Various measurement parameters are automatically generated by the analysis software including sectorial RNFL thickness for each individual quadrant and clock hour, and an overall value for the average RNFL thickness (Bottom middle panel). The normal distribution indices are colour-coded: (i) red <1%, (ii) yellow 1–5%, (iii) green 5–95%, and (iv) white >95% (Bottom right panel).

Fig. 9

Age-related decrease in average RNFL thickness, consistent with progressive retinal ganglion cell loss among OPA1 mutational carriers (n = 40). Spearman rank correlation coefficient = −0.2419, P = 0.0307 (PWYM, unpublished data).

Fig. 10

Optic nerve sections stained for myelin and mitochondrial COX activity: (A) Sudan black staining revealing the presence of myelin posterior to the lamina cribosa, (B) marked differential COX activity in the pre- and post-lamina cribosa segments, with intense COX staining in transverse sections taken from the pre-laminar region (C), and significantly lower levels of COX activity in the pos-tlaminar region (D). Reproduced with permission from Andrews et al. (1999a).

Fig. 11

Histology, mitochondrial histochemistry and immunohistochemistry (IHC) performed on serial longitudinal optic nerve sections: (A) haematoxylin and eosin, (B) Van Gieson preparation for connective tissue fibres (Red), with the arrow pointing towards the lamina cribosa, (C) Weigert iron haematoxylin preparation for myelin (Dark blue), (D) COX activity, with a darker stain (Brown) evident in the unmyelinated pre-lamina cribosa segment of the optic nerve, (E) IHC for COX subunit IV revealing a pattern consistent with the level of mitochondrial enzyme activity, (F ) IHC for Na_v 1.1 showing a greater concentration of these specific voltage gated Na⁺ channels in the pre-laminar region, (G) IHC for Na_v 1.2 demonstrating a uniformly pale labelling pattern in both pre- and post-lamina cribosa areas, (H) IHC for Na_v 1.6 with a strong staining reaction observed in the unmyelinated pre-laminar optic nerve, and (I) control optic nerve section with the primary antibody omitted. Reproduced with permission from Barron et al. (2004).

Fig. 12

The importance of the cytoskeleton in maintaining the differential concentration of mitochondria in the pre- and post-lamina cribosa segments of the optic nerve. The left panel shows a longitudinal section of a human optic nerve stained sequentially for COX and SDH. The pre-lamina, unmyelinated segment has a much darker COX staining consistent with the higher concentration of mitochondria. The right panel is a schematic representation of the cytoskeletal–mitochondrial interactions which facilitate the transport, distribution, and localisation of mitochondria to different areas of the optic nerve.