From Genotype to Phenotype in Leber Hereditary Optic Neuropathy: Still More Questions than Answers

In 1988, the first point mutation in the mitochondrial DNA (mtDNA) to be associated with a human disease was reported in pedigrees with Leber hereditary optic neuropathy (LHON) (1). An explosion of clinical, genetic, and biochemical studies followed (2–5). The advent of molecular diagnosis for LHON allowed us to more easily make a definitive diagnosis of LHON in patients without a family history of visual loss and expanded the clinical phenotype of the disorder to include some unusual features in individuals and in pedigrees. In the current edition of the Journal of Neuro-Ophthalmology, the article by Howell et al. (6) highlights how observational clinical and genetic studies in LHON continue to help us sort through the various theories on pathogenesis and the determinants of expression of this disease. Three point mutations in the mtDNA, the primary LHON mutations, are believed to cause 90% to 95% of cases of LHON worldwide. They are located at mtDNA nucleotide positions 11778 (69% of cases), 3460 (13% of cases), and 14484 (14% of cases) (2–5) (Fig. 1). Screening for LHON in a patient with visual loss should begin with the three primary mutations. In those primary mutation-negative patients in whom suspicion remains high, testing for the other mtDNA mutations associated with LHON is probably warranted, especially for those mutations deemed likely causal in a few previous pedigrees (2–5,7) (Fig. 1). Alternatively, because the majority of these other mtDNA mutations reside in genes encoding subunits of complex I, complete sequencing of complex I, perhaps beginning with the so-called “hot spot” ND6 gene (8), may also be considered. Finally, sequencing the entire mitochondrial genome is possible, although labor intensive. This should be performed only in those cases of high suspicion and interpreted by someone versed in the complexities of mitochondrial genetics (9).Fig. 1.: Mitochondrial genome showing the point mutations associated with Leber hereditary optic neuropathy (LHON). Over 90% of all cases of LHON are associated with the three primary mutations located inside the genome, and the other mutations are shown outside the genome. These other mutations vary markedly in their prevalence, degree of evolutionary conservation of the encoded amino acids altered, and frequency among controls. Mutations marked * may be primary, but they each account for only one or a few pedigrees worldwide. Mutations marked ** are primary mutations associated with LHON and dystonia. (Adapted with permission [7], courtesy of Marie Lott).How do we get from a point mutation in the mtDNA of all the cells of an individual to a sudden, selective loss of optic nerve function primarily in males in their early adulthood? Some hints as to the mechanisms of the disease are provided by clinical and genetic studies performed over the past decade. Does the specific mtDNA mutation dictate particular clinical features? There are pedigrees with “LHON plus” in which certain mtDNA mutations seem to result in specific disease patterns of LHON-like visual loss combined with other neurologic syndromes, such as dystonia (4) (Fig. 1). However, among the other LHON pedigrees, few significant clinical differences have been demonstrated to date among those patients with the three primary mutations, those with other mtDNA mutations, and those that are genetically unspecified. The only major exception is the far greater spontaneous recovery rate among patients with the 14484 mutation, as compared with those affected individuals with the 11778 mutation and even the 3460 mutation (3). Does heteroplasmy (the coexistence of mutant and normal mtDNA) within individuals in a pedigree play a role in phenotypic expression of LHON? Although there are exceptions, the mutant mtDNA content does seem to correlate positively with disease expression (10). However, once a person becomes symptomatic, there does not appear to be any clinical difference in expression. Furthermore, heteroplasmy is found in only a minority of LHON pedigrees and, therefore, not in the majority of unaffected maternal relatives worldwide. Although it is theoretically possible that unrecognized heteroplasmy exists in the optic nerves of these individuals despite homoplasmy in their accessibly tested tissues, this remains highly unlikely. Although some investigators have claimed that having multiple LHON-associated mtDNA mutations may be necessary for visual loss, case reports of unaffected individuals who even harbor two primary mutations make this claim improbable (3,6,11). Similarly, although the underlying mtDNA haplotype may influence the presence, penetrance, or expression of a mtDNA point mutation (12), this is unlikely to be a major factor. Nuclear-encoded factors modifying mtDNA expression, mtDNA products, or mitochondrial metabolism may be necessary for phenotypic expression of LHON (13). Although most studies have not been able to confirm X-linkage as an explanation for the male predominance of the disease, the X-linked hypothesis may still be viable. (4) Mitochondrial energy production decreases with age, and the timing of visual loss in patients at risk for LHON may reflect the threshold at which already reduced mitochondrial function deteriorates to a critical level (14). Immunologic factors have also been proposed, especially to explain the association of LHON with multiple sclerosis (15), but conclusive evidence is lacking. Finally, environmental factors, both internal and external, may play a role. Systemic illnesses, nutritional deficiencies, medications or toxins that stress or directly inhibit mitochondrial metabolism could conceivably initiate or increase phenotypic expression of the disorder. However, widespread nutritional deficiency in Cuba did not appear to increase the expression of LHON in one large 11778-positive pedigree (16). Similarly, although anecdotal reports suggest a possible role for tobacco and excessive alcohol use as precipitants of visual loss, other studies, including one large case-control study of sibships (17), failed to confirm this. Indeed, despite the extraordinary advances in our understanding of mitochondrial genetics and disease over the past decade, the underlying pathophysiological mechanisms linking the specific gene defect with the clinical manifestations of LHON remain unknown (18–20). Theories must reconcile how multiple different mtDNA mutations located in different genes encoding different proteins result in an essentially identical clinical phenotype that is expressed only in the optic nerve and with sudden and bilateral involvement. The obvious place to begin in theorizing about the pathogenesis of LHON is at oxidative phosphorylation and the generation of ATP. The three primary LHON mutations are found in genes that code for protein subunits of complex I of the respiratory chain. However, even demonstration of a definite deficiency of complex I activity and/or ATP synthesis as measured by enzyme assay has proved problematic (18), especially with the 11778 mutation, the most common and severe mutation in terms of visual prognosis. Furthermore, although the central nervous system is very reliant on mitochondrial ATP, studies of the bioenergetics of vision do not indicate that ganglion cell function is markedly dependent upon mitochondrial energy production (18). Indeed, it is the photoreceptor layer of the retina, rather than the retinal ganglion cell layer, that is richest in enzymes for oxidative metabolism (18). These observations have led some to speculate that the pathogenesis in LHON is not via direct effects on oxidative phosphorylation but rather via more indirect mechanisms. Complex I is the site of NADH reoxidation, and it may be that maintenance of the proper NAD/NADH redox balance is particularly necessary for normal maintenance and function of the ganglion cells or their axons, especially for axonal transport (18,21). Transporter defects involving channels and pumps particularly reliant on mitochondrial ATP production could have specific tissue effects (20). Another avenue of investigation involves free radical damage (19,22–24). Inhibition of the respiratory chain causes increased generation of free radicals. The respiratory chain may, in turn, be inhibited by oxidative damage. This self-amplifying cycle of oxidative damage and respiratory chain dysfunction could cause direct or indirect damage to vulnerable tissues such as the ganglion cell or its axon. The different mtDNA mutations may cause different amounts of free radical production at the level of complex I, and this increased production of reactive oxygen species may occur particularly in a neuronal cell environment (24,25). What are the subsequent mechanisms by which respiratory chain dysfunction or free radical production result in irreversible damage to the ganglion cells and their axons? Mitochondria play a major role in the induction of apoptosis (26). The collapse of the mitochondrial transmembrane potential and opening of the permeability transition pore leads to the release of proapoptotic proteins, including cytochrome c, and resultant cell death (19,27). Many physiologic and pathologic stimuli can directly trigger opening of the mitochondrial transition pore. Multiple consequences of mitochondrial dysfunction–collapse of the transmembrane potential, uncoupling of the respiratory chain, hyperproduction of superoxide anions, disruption of mitochondrial biogenesis, outflow of matrix calcium and glutathione, and release of soluble intermembrane proteins–probably indirectly activate apoptosis or necrosis (27). Alternatively, excitotoxic death can be initiated directly by reduced mitochondrial energy production, which leads to activation of the NMDA-type glutamate receptors (19). Free radicals are key components in some excitotoxicity pathways and may also trigger apoptotic cell death directly. The next step in theorizing how any of these proposed pathophysiologic processes result in selective damage to the optic nerve is fraught with even more uncertainty. Various mechanisms have been proposed, but supportive evidence is scarce. Based on the abnormal appearance of the papillary and peripapillary vasculature in many LHON patients, a primary vascular process with subsequent ischemia has been proposed (28). However, vascular changes may be secondary, especially given the intact vascular endothelium as demonstrated by the lack of fluorescein leakage and the absence of fundus abnormalities in many patients. A further hypothesis is that the acute angle turn made by optic nerve axons in the prelaminar region creates a “chokepoint” region of impaired or labile axoplasmic transport (18). Small-caliber, constantly firing P-cells may be particularly vulnerable (29). Histochemical studies of the optic nerve in animals have shown a high degree of mitochondrial respiratory activity within this unmyelinated, prelaminar portion of the optic nerve, suggesting a particularly high requirement for mitochondrial function in this region (30–32). Of great interest in this regard is the recent identification of the causative gene product in autosomal dominant optic atrophy (another inherited disorder that selectively involves the optic nerve) as a nuclear-encoded protein destined for the mitochondria and important in the formation and maintenance of the mitochondrial network (33). Similarly, a recent genetically induced mouse model of complex I deficiency shows the histopathologic features of optic nerve degeneration and demyelination (34). These studies support the theory of a selective vulnerability of the retinal ganglion cell and its axon to defects in mitochondrial function. The two independent pedigrees described in the article by Howell et al (6) both harbor two primary LHON mutations, the 14484 homoplasmic and the 11778 heteroplasmic. This genetic “double whammy” has been reported in only two previous families (3,11). Yet even harboring two primary mutations, these pedigrees exhibit no exceptional presentation of visual loss, including penetrance, the role of heteroplasmy, timing, severity, and even eventual recovery. Although one branch of the Baltimore pedigree has an extremely high incidence of a fatal infantile encephalopathy, the absence of this manifestation in other branches of the family and in the other pedigrees previously reported makes it unlikely to be a direct manifestation of the combination of the two primary LHON mutations. This branch of the Baltimore family is reminiscent of the Australian pedigree with maternally related LHON-like visual loss, movement disorders, and episodic infantile encephalopathy who harbor both the 14484 mutation and an ND1 mutation at mtDNA position 4160, the latter presumably responsible for the additional neurologic manifestations (35). However, sequencing of the entire mtDNA of the Baltimore family, including those neurologically affected individuals, failed to reveal any variance in mtDNA to explain these additional manifestations. Although the influence of a nuclear DNA factor has been proposed, that is a somewhat unsatisfactory explanation for this highly penetrant phenotype in a single generation. How can we reconcile the observations of Howell et al. (6) with the various proposed theories of LHON pathogenesis? Clearly, having a primary mtDNA mutation associated with LHON, or even two, is not sufficient for the expression of the disease. Even if the two mutations combine to cause a more severe biochemical defect in mitochondrial respiratory chain activity (11), this is not translated into increased clinical severity, therefore suggesting a more indirect pathogenesis of optic nerve dysfunction. The mtDNA mutation, or mutations, may set the stage for visual loss, but unknown internal or external precipitants determine its occurrence. Further elucidation of these triggers of the pathologic cascade in susceptible individuals will require more genetic, biochemical, physiological, and pathologic studies. Development of mouse models for LHON and other mitochondrial diseases should help focus our efforts toward directed therapies for both prevention and recovery (34,36). [See also Biousse et al. (37) in this issue.]