Mitochondrial DNA Variation of Leber’s Hereditary Optic Neuropathy in Western Siberia

Our data first represent the variety of Leber’s hereditary optic neuropathy (LHON) mutations in Western Siberia. LHON is a disorder caused by pathogenic mutations in the mitochondrial DNA (mtDNA), inherited maternally and presents mainly in young adults, predominantly males. Clinically, LHON manifests itself as painless central vision loss, resulting in early onset of disability. The epidemiology of LHON has not been fully investigated yet. In this study, we report 44 genetically unrelated families with LHON manifestation. We performed whole mtDNA genome sequencing and provided genealogical and molecular genetic data on mutations and haplogroup background of LHON patients. Known “primary” pathogenic mtDNA mutations (MITOMAP) were found in 32 families: m.11778G>A represents 53.10% (17/32), m.3460G>A—21.90% (7/32), m.14484T>C–18.75% (6/32), and rare m.10663T>C and m.3635G>A represent 6.25% (2/32). We describe potentially pathogenic m.4659G>A in one subject without known pathogenic mutations, and potentially pathogenic m.6261G>A, m.8412T>C, m.8551T>C, m.9444C>T, m.9921G>A, and m.15077G>A in families with known pathogenic mutations confirmed. We suppose these mutations could contribute to the pathogenesis of optic neuropathy development. Our results indicate that haplogroup affiliation and mutational spectrum of the Western Siberian LHON cohort substantially deviate from those of European populations.


Introduction
Leber's hereditary optic neuropathy (LHON) is a form of hereditary disorder caused by pathogenic mutations in mitochondrial DNA. These mutations are non-synonymous, and affect genes coding for different subunits of complex I of the mitochondrial respiratory chain. The occurrence of such kinds of mutations in mtDNA subunits leads to dysfunction of the electron transport, increased reactive oxygen species production, and defective ATP synthesis [1][2][3]. Retinal ganglion cells are highly susceptible to death during LHON progression, because of their high sensibility to disrupted ATP production and oxidative stress [4]. Therefore, LHON is usually painless, acute, or subacute, central visual loss of one or both eyes, which results in early onset of disability. The onset of Leber's hereditary optic neuropathy is relatively rare and has better visual prognosis. The peak age of onset of visual loss among Whole peripheral blood samples were collected from the donors in 10 mL Vacutech EDTA tubes. Total DNA was extracted from a buffy-coat layer using the SileksMagNA-G Blood DNA Isolation kit, according to the manufacturer's protocols. The complete sequencing procedure entailed PCR amplification of 22 overlapping mtDNA templates [24], which were sequenced in both directions with BigDye 3.1 terminator chemistry (PE Applied Biosystems, Foster City, CA, United States). The trace files were analyzed with Sequencher (version 4.5 GeneCode Corporation) software. To perform capillary electrophoresis on an ABI Prism 3130XL DNA Analyzer, we used core facilities of the "Genomika" Sequencing Center (SBRAS, Novosibirsk, Russian Federation). Variants were scored relative to the Reconstructed Sapiens Reference Sequence, RSRS [25]. MtDNA haplotypes were identified following the nomenclature suggested by the PhyloTree Build 17 [26]. Forty-two mitochondrial genomes obtained through this study were deposited in GenBank with accession numbers MN413201-MN413242. Two genomes, EU807741.1 and EU807742.1, had been deposited to the GenBank earlier [21].

Penetrance Analysis
We determined penetrance as the proportion of affected individuals from all maternally related family members using family pedigrees [19]. Values for both men and women were calculated separately.

Analysis of Pathogenicity for Non-Synonymous Mutations
To make sure that the revealed non-synonymous mtDNA mutations were not sequencing errors or to not point out to general population polymorphisms, and in turn to find out the disease-associated polymorphisms among those published earlier, we used several databases: MITOMAP [27]; mtDB (Human Mitochondrial Genome Database, containing 2704 human mitochondrial genomes) [28]; and HmtDB (Human Mitochondrial DataBase), which contains 32922 human mitochondrial genomes [29]. To assess the possible pathogenicity of these mutations, and to predict whether a protein sequence variation affects protein function, we used the following web applications: MutPred 1.2 [30], MutPred 2 [31], PolyPhen-2 [32], PROVEAN (Protein Variation Effect Analyzer) [33], and SIFT (Sorting Intolerant from Tolerant) [34]. All the sources are provided in the public domain.

No.
Family Name

Number of
According to the family pedigrees, only 50% (22/44) of cases had a family history of vision loss in the maternal lineage in more than one generation, among which m.11778G>A represented 36% (8/22), m.3460G>A covered 23% (5/22), and the m.14484T>C was 14% (3/22), as well as those without primary mutations (LHON-like cases), which represented 18% (4/22). Rare mutation cases (m.10663T>C and m.3635G>A) were family-inherited. In total, 39% (17/44) of cases were sporadic, among which 13 were cases with only one affected person diagnosed, and four were cases with two affected persons in one generation. Among the sporadic cases, m.11778G>A represented 41% (7/17), m.3460G>A was 12% (2/17), m.14484T>C was 6% (1/17), and the LHON-like cases represented 41% (7/17) of the total. Summary information on penetrance is shown in Table 2. The penetrance is highly variable between separate families, even with the same primary mutation. The average penetrance among men was 32% (6-100%) and 12% among women (0-58%); these correlate with data previously published [4]. However, there are some families with higher penetrance among females than among males: L24, L26, and L28. In 12 families with clear-cut LHON phenotypes, no pathogenic mtDNA mutations were found. Analysis of the mtDNA revealed non-synonymous mutations: m.4766A>G, m.13105A>G, m.14002A>G, which have not been noted as associated with LHON or other diseases in the MITOMAP database. All pathogenicity prediction tools indicated low probability that the amino acid substitutions are disease-associated for these mutations. Mutation m.4659G>A has been previously reported as being associated with Parkinson's disease [35], as well as in an Australian LHON pedigree that was heteroplasmic for the m.14484T>C [36]. Polyphen-2 predicted the pathogenicity for this mutation as benign and MutPed 2 showed low probability score, but MutPred 1.2, PROVEAN, and SIFT determined this mutation as deleterious. The results are shown in Table 3. In several families with primary mutations (L01, L03, L12, L28, L30, L40, L43, and L50) we found out additional, non-synonymous mutations (Table 4). Mutations m.8875T>C, m.14582A>G, m.8400T>C, and m.4639T>C were neutral, and mutation m.9444C>T had a high probability of being pathogenic, according to data from all the pathogenicity prediction tools; for other mutations, we observed divergence of prediction results. Since prediction results for primary pathogenic mutations diverged too (see Table 3), novel non-synonymous nucleotide change was considered potentially pathogenic if it had extremely low frequency in the general population, and if it was predicted by at least three algorithms to have an effect on protein function. For mutations m.6261G>A and m.15468C>T, only PolyPhen2 predicted pathogenicity as probably damaging and possible damaging, respectively. However, mutation m.6261G>A had already been reported by Abu-Amero [37] in patients with optic neuropathy, and also as a somatic mutation associated with prostate cancer. Interestingly, the family (L01) with m.6261G>A and m.11778G>A has the same haplogroup, T2, as the case reported by Abu-Amero. Other mutations (m.8412T>C, m.8551T>C, m.9921G>A, m.15077G>A) were predicted as pathogenic by at least by three algorithms, but the first three of them have not been noted as associated with diseases in the MITOMAP database, and mutation m.15077G>A was reported as being associated with maternally-inherited isolated deafness [38].

Leber's Hereditary Optic Neuropathy Primary Mutations
Regarding our preliminary data, the frequencies of primary mutations are different from frequencies reported for Europe and Asia. The most prevalent m.11778G>A (~55%) is less common in Western Siberia than in Europe, at~69% [9]; as well as in China and Japan, at~90% [10,13]. The prevalence of m.3460G>A (~24%) is twice as much as in Europe, but m.14484T>C (~14%) do not deviate from those of other European populations [9].
In 12 families with LHON-like manifestation, no known pathogenic mtDNA mutation was found. However, there are other elaborations in which the clinical diagnosis cannot be confirmed by molecular genetic analysis [37,[39][40][41]. Definitively, mtDNA mutation-caused LHON is clinically indistinguishable from the other forms of optic neuropathy, such as a dominant optic neuropathy (DOA), especially when it is sporadic. Compared to LHON, DOA visual loss is detected between ages 4 and 6 in the majority of patients, and 58-84% of patients with DOA report visual impairment by age 11 [42]. In our 12 cases, the ages of onset were between 13-36 years. We will be studying these cases for the mutation spectrum of common pathogenic genes for DOA in future.

Penetrance
We suppose that relatively reduced incidence of LHON in Western Siberia is associated with incomplete penetrance and diagnostic difficulties of atypical (e.g., late-onset) and combined (e.g., multiple sclerosis) forms of LHON. Patients with LHON could have also been wrongly diagnosed as suffering from toxic amblyopia, tobacco-alcohol amblyopia, or optic neuritis [41]. On the other hand, the problem is that patients do not know their family history. Molecular testing for LHON is not routinely performed in patients with optic atrophy in Russian Federation. Identification and registration of unaffected carriers plays an important role for prevention of disease manifestation. For example, there is strong evidence that smoking is associated with an increased risk of visual failure among LHON carriers-93% penetrance of vision loss in male smokers versus 66% in male non-smokers [17].
Our observations highlight the importance of molecular genetic examination for unaffected carriers. The presence of pathogenic mutations should be tested not only for probands, but for all relatives in the maternal line, Since the proportion of sporadic cases is about 40% according to our published data [4].

Potentially Pathogenic Mutation m.4659G>A
We found m.4659G>A in one subject without any known primary mutations (L51). This sequence change is located at codon 64 in the functional domain of the ND2 gene and changes an alanine-a hydrophobic amino acid-into threonine-a neutral amino acid. Mutation m.4659G>A has been reported as being associated with LHON in an Australian pedigree that also had heteroplasmic mutations m.14484T>C and m.5460G>A [36]. This family had 10 maternally-related descendants, five of whom had vision loss. Unfortunately, our patient does not know his family history, and we could not confirm maternal inheritance for this mutation. However, m.4659G>A has very low frequency in the general population (0.0011-0.00161), and has a high probability of being pathogenic (Table 3).

Additional Non-Synonymous Mutations Revealed in LHON Cases
The phenomenon of co-existence of two pathogenic mutations in one family has already been described. The first case included m.4659G>A, m.5460G>A, and m.14484T>C in an Australian LHON pedigree, described above [36]. In the second case, a Polish family harboring two primary LHON mutations m.3460G>A and m.11778G>A occurred in a haplogroup H background [43]. In the third case, a family harbored two primary LHON mutations, m.11778G>A and m.14484T>C, and both mutations had a synergistic pathogenic effect on protein function, as well as a higher degree of heteroplasmy of the m.14484T>C, correlated with an earlier age at onset [44]. Finally, the fourth example is a unique double-mutant ND4L with two concurrent mutations (m.10609T>C and m.10663T>C) in an Arab pedigree from Kuwait [45].
We reported mutations m.6261G>A, m.8412T>C, m.8551T>C, m.9444C>T, m.9921G>A, and m.15077G>A, which could be potentially pathogenic because of their low frequency in the general population, and high probability of pathogenicity according to data from different prediction tools. Two of them, m.6261G>A and m.15077G>A, have already been reported in subjects with optic neuropathy and maternally-inherited isolated deafness, respectively. However, we suppose that additional non-synonymous mutations could either have a synergistically pathogenic or a protective effect. To demonstrate the full significance of novel mutations, a respiratory chain assay would need to be performed. An example is the study [38], where cybrids with m.15077G>A showed normal activities for mitochondrial electron chain enzymatic complexes.
It is known that the clinical impact of mDNA mutations may be modulated by mitochondrial haplogroup background. For example, Hudson et al. performed a multicenter study of 3613 subjects from 159 different families, and showed that the risk of visual failure is greater when m.11778G>A or m.14484T>C mutations are present in specific subgroups of haplogroup J; the same as the m.3460G>A mutation is present in haplogroup K, and the risk of visual failure is significantly lower when m.11778G>A occurs in haplogroup H [50]. Romero et al. supposed that haplogroup D has a protective effect in carriers of LHON mutations. His hypothesis was based on the fact that there was a markedly decreased frequency of haplogroup D in Chilean subjects with LHON, as haplogroup D is one of the most common in the Chilean population [12]. Also, other experimental research serves as proof that cybrids and fibroblasts bearing LHON mutations have different response to neurotoxic agents, depending on haplogroup background [51].
It has been suggested that at the end of the last glaciation, phylogenetically more ancient mutations could have provided their carriers with adaptive advantages during the human population expansion. Today, those mutations contribute to the saving and expression of weakly pathogenic LHON mutations, which appear randomly in different region-specific genetic backgrounds [52,53]. The theory could be tested by further searching of pathogenic, LHON-causing mutations in relation to specific mtDNA backgrounds (phylogenetically ancient set of mutations).
New data collected from future studies regarding mtDNA variations of LHON in Western Siberia might be used to develop a LHON system registry in the Russian Federation. We intend to conduct consecutive experimental research, including the parameters of the pathogenicity of each novel substitution.

Limitations of the Study
The main limitation of the data presented is the absence of physiological tests as proof with respect to the pathogenicity of novel mtDNA substitutions. Additional tests should be done, such as oxygen consumption, ATP and ROS measuring, and electron microscopy study (for example, [54]). In addition, development of the newest editing systems [55] could give us more reliable instruments to test pathogenicity.