Keywords

1 The ‘Normal’ Ageing Brain

Like any organ, the human brain is susceptible to the endless march of time with the effects of ageing often manifesting as diverse pathologies. To understand the role of mitochondrial DNA (mtDNA) variation in brain ageing, we must first understand what brain ageing entails.

1.1 Physiological Changes

At the organ level, brain weight and volume decline at a rate of around 5 % per decade after 40 years of age [1], with rates increasing in individuals over 70 years [2]. Volume loss is associated with an increase in ventricular volume and cerebrospinal fluid (CSF) spaces. Longitudinal magnetic resonance imaging (MRI) and comparative cross-sectional studies have shown that age-related brain changes are regional [2, 3] with the hippocampus and frontal lobes most affected. Estimates indicate that, over an age range of 30–90 years, the cerebral cortex, hippocampus and cerebral white matter are subjected to volumetric losses of 14 %, 35 % and 26 %, respectively [4, 5]. The occipital cortex appears least affected.

The exact mechanism of reduction is up for debate, with studies indicating that grey matter shrinkage [6] and changing neuronal morphology [7], rather than the dogmatic theory of neuronal loss, are the principle components contributing to decreasing brain volume. Additional ageing hallmarks include dendritic sprouting [8, 9], a compensatory mechanism to maintain synapse number and cell number [10], a loss of dendritic arbour [11] and a decline in white matter [12].

The effects of ‘normal’ ageing are also seen at the cellular level. Lipofuscin, the pigmented residues of lysosomal digestion, accumulates in neurons; it is a hallmark of ageing and characteristic of a failure to eliminate the products of peroxidation-induced cell damage [13]. Neurofibrillary tangles and plaques, hallmarks of Alzheimer’s disease (AD), are identifiable in hippocampal, amygdala and entorhinal cortex neurons in healthy aged brains [14] although at subclinical levels. Amyloid β-peptide (Aβ)-containing plaques, another hallmark of AD, form in the grey matter of ‘normal’ aged brains, although present only in small numbers compared to the diseased brain. Advanced age increases the frequency of vacuoles, immunoreactive for neurofibrillary tangles (155K and 210K neurofilament polypeptides) [15], in the cell bodies of pyramidal and hippocampal cells leading to granulovacuolar degeneration [16]. Large (30 μm long) rod-shaped, paracrystalline structures, termed Hirano bodies, accumulate in hippocampal pyramidal cells [17]. Similar to granulovacuoles, these bodies appear to be composed of cytoskeletal proteins (including actin) [18] as well as the constituents of neurofibrillary tangles [19]. As we age, a reduction of oxygen through a reduction in blood supply causes cerebral microinfarcts (cellular necrosis). Typically associated with dementia, these microinfarcts, difficult to detect by conventional means (MRI), are surprisingly common in aged brains (typically >80 years) [20].

1.2 Genetic and Transcriptomic Changes

Genetic factors, such as apolipoprotein E haplotype (APOE), have been identified largely through the investigation of pathological ageing such as in AD [21]. However, these studies are mixed, with some indicating age-related defects in elderly ε4-haplotype carriers [22] and others showing no significant effect [23]. Studies in mice indicate that genetic loci (Hipp1a, Ch1 and Hipp5a, Ch5) can modulate hippocampal structure and volume during normal ageing (potentially mediated through the genes: retinoid X receptor gamma, Rxrg, and fibroblast growth factor receptor 3, Fgfr3, respectively) supporting the role of genetics in brain ageing [24].

At the transcriptomic level, work in mice indicates that the gene expression landscape of the brain, particularly genes involved in stress and inflammatory response, changes dramatically over time and can be modulated by dietary restriction [25]. This supports earlier work investigating the role of ageing in mice at the protein level [26]. In humans, a reduction in synaptic plasticity, modulated by a reduction in α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptor expression, is a hallmark of ageing [27]. In addition reductions of key brain transcripts, including N-methyl-D-aspartate (NMDA) receptors (required for learning and memory) [28], GABA-A receptor subunits (required for neurotransmission) [29] and genes involved in long-term potentiation (e.g. CALM1, calmodulin 1, required for calcium signalling) [30], begin at around 40 years of age.

1.3 Cognitive Changes

Cognitive performance declines with age, although the differentiation between ‘normal ageing,’ ‘non-pathological ageing’ and neurodegenerative disease confounds quantitative measurement. The most observable phenotype associated with ‘normal’ brain ageing is a loss of memory, with no significant loss of visuospatial, cognitive function or language [31]. Memory is classified as four distinct classes: episodic (memory of situation, e.g. a memory of a birthday party), semantic (memory of facts, e.g. Washington, DC, is the capital of the USA), procedural (memory of motor skills, e.g. how to walk) and working (transient memory, for new and stored information) [32]. Studies have investigated the effects of ageing on memory using neuropsychological testing and neuroimaging, demonstrating that brain ageing has a profound effect on both episodic and semantic memory.

Episodic memory, typically affected in AD, is thought to decline from middle age onwards, affecting both recall and recognition in elderly life [33]. Semantic memory also decreases from middle age and then declines rapidly during elderly life [33], driven in part by slower cognitive reaction times, lower attention levels and slower processing speeds in aged individuals [6, 34]. In addition to memory, ageing affects an individual’s orientation and attention. Orientation, or the awareness of one’s surroundings, has been shown to decline mildly with advancing age [35], although this is inconsistent between studies, possibly reflecting the interaction of other variables [36]. Attention, or our ability to focus our brain’s resources on a task, declines with age; moreover secondary phenotypes associated with advancing age [37], i.e. impaired hearing or vision, compound valid attention testing. Indeed, it is worth noting that differentiating these separate cognitive functions, which are intrinsically linked, makes data interpretation difficult [38],

1.4 The Drivers of Brain Ageing

Dopamine, a neurotransmitter involved in motor control, arousal, cognition and reward, declines from early adulthood at a rate of around 10 % per decade, subsequently affecting both cognitive and motor function in aged individuals [33]. Serotonin (a neurotransmitter) and brain-derived neurotrophic factor (BNDF, a neuronal growth factor) are both expressed to maintain synaptic plasticity and both decline with age. We also know that sex hormones, particularly oestrogen, can affect dopaminergic efficacy – purportedly relaying a protective effect in neurodegenerative disorders such as AD [39]. As we age, cerebrovascular efficiency falls, which in turn impairs our ability to metabolise glucose [33], although subsequent data indicates that impaired glucose metabolism may be an effect of cellular atrophy, rather than a cause [40]. Finally, it is becoming clearer that mitochondria play an integral role in brain ageing, from modulating calcium influx and homeostasis [41] to direct dysfunction and the increased proliferation of reactive oxygen species (ROS) [42].

2 Mitochondrial Function in the ‘Normal’ Ageing Brain

Even in the absence of a primary pathogenic defect, changes in mitochondrial function are intrinsically linked to ‘normal’ brain ageing, attributed principally to increased mitochondrial dysfunction, a direct loss of cellular energy or an increase in ROS production, which increases mitochondrial decay [43]. It is worth noting that, from a bioenergetic perspective, the brain is an extremely heterogeneous and dynamic organ; for example, there is a reciprocal relationship between neurons, which have aerobic terminals, and astrocytes, which are more dependent on glycolysis [44]. This complexity, which makes energetic interpretations difficult, must be considered when investigating mitochondrial function in the brain.

Aged animal brains show a significant decline in mitochondrial function, characterised as a loss of phosphorylation capacity, decreased mitochondrial membrane potential [45], decreased respiration, ATP synthesis and activation of the mitochondrial permeability transition pore. The aged brain shows a characteristic reduction in two key respiratory chain (RC) enzymes: NADH/ubiquinone oxidoreductase (complex I, CI) and cytochrome c oxidase (complex IV, CIV) [46]; however, the extent and localisation of the RC deficit are variable [47]. This reduction in RC activity appears to trigger a corresponding proliferation of mitochondria with age, presumably an attempt to attenuate a loss of cellular energy [48] and neuronal mtDNA [49].

The brain, a high-energy organ, is particularly vulnerable to oxidative damage, in part due to a lack of antioxidant enzymes and the increased abundance of polyunsaturated fats, transition metals and the high rate of oxygen consumption [50]. Markers of oxidative stress, carbonylation (protein oxidation), lipid oxidation and the oxidation of the mitochondrial genome [51] are increased with age; however, whether increased oxidative stress markers are a result of a loss of antioxidant enzyme activity or an increase in pro-stress factors is contentious [52]. The coexistence of mitochondrial dysfunction and increased oxidative stress in the brain likely acts as a trigger for apoptotic activation, yet the role of apoptosis in the ‘normal’ ageing brain is controversial, with studies indicating that mitochondrially mediated apoptosis is not a cardinal element of ‘normal’ brain ageing [53] and is rather restricted to disease [54].

2.1 Mitochondrial DNA in the ‘Normal’ Ageing Brain

The risk of mtDNA mutation is high, primarily due to its localisation near to the site of oxidative stress [55], and is not restricted by tissue type or disease [56]. The result is a spectrum of mtDNA variation, ranging from the ‘inherited’ population-level polymorphisms which classify mtDNA haplogroups [57] to the acquired or somatic variation that accumulates with age (either large-scale deletions or single base-pair substitutions).

Despite strong links to mitochondrial functionality [58, 59], there is little evidence for the role of inherited mtDNA variation in ‘normal’ brain ageing. However, there are several reports linking inherited variation to the progression of neurodegenerative disorders such as Parkinson’s and Alzheimer’s disease (discussed later).

Conversely, there is mounting evidence indicating that acquired mtDNA variation accumulates in the brain during ‘normal’ ageing and much work has focused on the aptly named ‘common deletion’ (a 5 Kb deletion with known disease associations [60]), with reports showing that age-related oxidative damage increases the levels of the common deletion in several regions of the ‘healthy’ brain, particularly in the cortex, cerebellum, putamen [61, 62] and more recently substantia nigra and basal ganglia [60], and appears to correlate with regional differences in energy demand [63]. In addition to the common deletion, several studies have reported an increase in random and variable mtDNA deletions, typically smaller in size (>50 bp), which can be differentially distributed amongst brain regions [62, 64]. Studies of acquired single nucleotide variation in the brain are sparse; however studies have shown that healthy individuals harbour an increase in the frequency of low-level heteroplasmic variation in cytochrome c oxidase subunit I (MTCO1), the core catalytic subunit of complex IV of the RC [65] and the non-coding, regulatory, D-loop region [66].

3 Mitochondrial DNA Mutation and Neurodegeneration

As discussed, mitochondria are the cellular powerhouses, supplying critical ATP through oxidative phosphorylation. Secondary roles in apoptosis, responses to cellular stress and modulation of reactive oxygen species make mitochondrial function a central component of almost all cellular survival (see Chap. 1). The central nervous system (CNS), a highly energy-demanding system, is dependent on efficient mitochondrial function. A disruption of this function, through mitochondrial DNA mutation, is in many cases associated with neurodegenerative disease (Fig. 3.1).

Fig. 3.1
figure 1

Cartoon of the human brain showing the complex neurodegenerative diseases associated with mtDNA variation

Full size image

3.1 Mitochondrial DNA Mutations

Mitochondrial DNA, a circular 16.5 Kb molecule encoding the 13 core RC polypeptides, has a significantly higher mutation rate compared to its nuclear counterpart (around 3× or 2.7 × 10−5 base pairs per generation) [67], due in part to an absence of histone and its proximity to the inner mitochondrial membrane – the site of ROS production. However, mitochondria do contain antioxidant and DNA repair enzymes, including oxoguanine glycosylase (OGG1, responsible for 8-oxoguanine base excision) and (MUTYH, involved in oxidative DNA damage repair); furthermore mitochondrial gene expression is adaptive and isolated from the nucleus, allowing the post-translational machinery to generate individual gene products in response to ROS through retrograde signalling [68]. This retrograde signalling stimulates an adaptive nuclear response to mtDNA impairment, which modulates the expression of over 40 nuclear genes [69].

This high mutation rate and lack of effective mtDNA repair mechanisms lead to the formation of two classes of mtDNA variants: (1) inherited mtDNA variants, typically single base-pair exchanges that are present at varying population frequencies, and (2) acquired or somatic mtDNA variants, which can be either single base-pair exchanges or deletions of mtDNA.

3.2 Primary Mitochondrial Disorders and Neurodegeneration

Since the 1980s mitochondrial dysfunction has been a hallmark of metabolic disease [70] and subsequently over 300 mtDNA mutations have been associated with disease [71]. More recently, the phenotypic spectrum has broadened to include neurodegenerative disease [72], typically as a function of a primary mtDNA defect.

Leber’s hereditary optic neuropathy (LHON) is the commonest cause of maternally inherited blindness and a cardinal example of how mtDNA mutations cause neurodegeneration [73]. MtDNA mutations (m.3460G>A, m.11778G>A and m.14484T>C in ~90 % of cases) in NADH-dehydrogenase subunits (complex I) reduce cellular energy and increase oxidative stress and ROS production to cause a selective loss of retinal ganglion cells (RGC) in the retina whilst sparing the retinal pigmented epithelium and photoreceptor layer [74]. Atrophy of neuronal cell bodies and demyelination of the optic nerve are associated with impaired activity of the EAAT1 (excitatory amino acid transporter) glutamate transporter and increased mtDNA mutation-mediated ROS production triggering the apoptotic cascade [75]. RGC selectivity is attributed to high energy demand; however, atypical individuals may experience secondary neuronal loss.

Myoclonic epilepsy with ragged-red fibres (MERRF [76]) is typically attributed to m.8344A>G in the mitochondrial tRNA lysine gene (MT-TK) and presents as a progressive degeneration of the olivocerebellar pathway [77, 78], with severe neuronal loss in the inferior olivary complex, Purkinje cells and dentate nucleus. Biochemical studies indicate that m.8344A>G significantly impairs mitochondrial protein synthesis, hampering oxidative phosphorylation, decreasing ATP production and mitochondrial membrane potential and finally culminating in apoptotic cell death [79].

In addition, mtDNA mutations are associated with cerebellar or sensory ataxia [80], where a secondary loss of Purkinje or other cerebellar cells is seen in syndromes typically associated with a primary mtDNA defect, including Kearns-Sayre syndrome (KSS [81]), mitochondrial encephalopathy and stroke-like episodes (MELAS [82]) and neuropathy, ataxia and retinitis pigmentosa (NARP [83]).

3.3 Mitochondrial DNA Mutation in Complex Neurodegeneration

There is increasing evidence that an increase in oxidative stress and apoptosis is linked to the aetiopathogenesis of several age-related neurodegenerative diseases. Much research has focused on the role of mitochondrial impairment and oxidative stress in the onset and progression of the major ageing brain disorders, particularly Parkinson’s disease (PD) and AD; however, research has also linked mitochondrial mutations to Huntington’s disease (HD) and amyotrophic lateral sclerosis (ALS) (see Fig. 3.1).

3.3.1 Parkinson’s Disease

Parkinson’s disease is a degenerative disorder of the central nervous system, primarily affecting the motor system through a selective loss of dopaminergic neurons in the substantia nigra pars compacta. Early disease is typically movement related, manifesting as tremor, rigidity, slowness of movement initiation and gait instability, progressing to cognitive impairment in advanced stages.

The link between mitochondrial dysfunction and postmortem PD tissue is well recognised [84]. The substantia nigra pars compacta (SNpc), the most vulnerable tissue in PD patients, suffers an age-related reduction of complex I activity [85] and exposure to potent inhibitors of respiratory chain activity such as rotenone [86], 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) [87] and certain pesticides [88] manifest with neurological abnormalities similar to PD. The combined reduction in cellular energy availability and an increase in ROS may lower the threshold for neuronal apoptosis killing the energy-demanding dopaminergic neurons in the SNpc.

Despite a consistent link between mitochondrial proteomics and PD, the underlying genetic mechanisms remain complex and conflicting, due in part to the complex interactions of both inherited and acquired mtDNA mutations.

The role of inherited mtDNA variation, mediated primarily through mtDNA haplogroups, in the aetiopathology of PD is well studied, although the results are often conflicting and contentious.

Evolutionarily, inherited mtDNA mutations have created stable population subgroups termed ‘mitochondrial haplogroups,’ groups of phylogenetically similar mtDNA molecules, separated by a collection of stable and frequent polymorphic variants [57]. Many of these subdivisions occurred over 50,000 years ago, developing as humans migrated into new geographical regions. Over 95 % of Europeans belong to one of ten major haplogroups, H, J, T, U, K (a subgroup of U), M, I, V, W and X, with each haplogroup defined by specific mtDNA sequence variants within the population. The role of inherited mtDNA variants and their defined haplogroups has evolved; once thought to be benign population variants, mtDNA haplogroups are now known to directly affect mitochondrial function [58, 89] and have become associated with a broad spectrum of human disease [90–92]. In addition, there is evidence to suggest that inherited mtDNA may influence neuronal cytosolic pH and calcium regulation, which over time could influence neuronal function and viability [93].

In early work, the common European haplogroup J, driven largely by the presence of a single nucleotide substitution in MT-ND3 (m.10398G>A), was shown to be underrepresented in PD, suggestive of a protective effect, and is interesting given the purported effects of haplogroup J on ‘normal’ ageing, where it is overrepresented in healthy centenarians [94]. This is supported by a moderately powered study which found that a phylogenetically linked cluster containing haplogroup J (JTUK) reduced PD risk [95]. Smaller studies have reported a contradictory increase in PD risk with haplogroup J [96], but sample number and genetic heterogeneity, critically misunderstood in the highly mutable mtDNA [97], may account for this disparity. More recently, a much larger study and meta-analysis confirmed the role of haplogroup J/K in reducing the risk of PD [98], an effect which was replicated in a follow-up study later [99]. A meta-analysis of all available mtDNA-PD data was able to identify that the phylogenetically linked H/V haplogroups increased the risk of PD [99].

It is intriguing that ostensibly common, and once thought benign, variation in the mitochondrial genome could convey a direct effect on the pathoaetiology of PD. More likely, haplogroup J/K/H/V are genetic tags of further mtDNA variation not directly investigated in these studies. For example, haplogroup K contains the lowest frequency of nonsynonymous complex I gene variants (when compared to the other haplogroups), raising the possibility that natural selection against genetic variation of complex I reduces the risk of PD in haplogroup K individuals. Similarly, an association with haplogroup H, which is linked to infection survival, raises the possibility that natural selection has led to the emergence of variants which predispose to age-related neurodegenerative disease through antagonistic pleiotropy. Similar to HD, where an increased CAG(n) repeat length in the Huntingtin gene causes a neurodegenerative disorder, but is associated with a lower incidence of cancer, recent mutations in mtDNA may be tolerated in humans because they increase the chance of surviving early-life insults such as sepsis, but are pathogenic in later life through the increased generation of reactive oxygen species. In both of these instances, advances in genotype and sequencing technology will allow researchers to understand the role of inherited mtDNA variants, by enabling more and more sequences to be collected and analysed.

More speculative is the exact role of acquired or age-related somatic, mtDNA variants. Much early work focused on the role of mtDNA deletions, large and small lesions in the genome. These deletions typically remove large proportions of the heavily coded mtDNA molecule, dramatically affecting mitochondrial protein synthesis and replication, which due to their obvious pathogenic nature appear heteroplasmic (a mixed population of mtDNA molecules within a cell or tissue).

Early studies failed to find a direct association between mtDNA deletion levels in the brains of PD patients, showing no significant increase when compared to age-matched healthy control tissue [100–102]. Advances in mtDNA isolation technology, enabling capture from single cells, and improvements in the PD diagnostic efficacy improved deletion detection. Studies taking advantage of these technologies show that the frequency of mtDNA deletions in individual neurons in the SNpc of PD patients clonally accumulates with age, increasing after 65 years old [103], and is associated with a decrease in cytochrome c oxidase activity [104]. Both of these reports suggest that the SNpc is particularly vulnerable to free radical-mediated mtDNA damage and that mtDNA deletions may play a role in the pathoaetiology of PD.

The identification of mtDNA deletions in non-PD-related disease [105] and healthy individuals [106] questions the exact contribution of these lesions to the development of PD. In addition, the observation of mtDNA deletions in newborn brains [107] raises the question of whether the clonal expansion mtDNA deletions are the cause of, or are driven by, SNpc failure.

Studies of acquired mtDNA point mutations, typically heteroplasmic, in PD are extremely limited and are often focused on idiopathic cases with no evidence of mitochondrial dysfunction. Studies have identified heteroplasmic variation in PD cases [108, 109]; however, without supporting biochemical or functional data, these variants and their role in the development of PD are difficult to interpret.

3.3.2 Alzheimer’s Disease

Alzheimer’s disease is a progressive loss of neurons and synapses in the cerebral cortex and subcortical regions and is the commonest cause of dementia in the elderly [110]. The disease is typically associated with plaques or neurofibrillary tangles (aggregates of hyper-phosphorylated tau protein).

Like PD, there are several lines of evidence linking mitochondrial dysfunction and oxidative damage in the aetiopathogenesis of AD. Animal experiments using mice which overexpress precursor amyloid protein (APP) demonstrated mitochondrial dysfunction and ATP production through a reduction in mitochondrial protein synthesis [111] or an interaction with mitochondrial matrix proteins [112]. Mitochondrial dysfunction, primarily a reduction in neuronal cell energy, promotes tau phosphorylation, an indication that mitochondrial vitality and oxidative stress are linked to the neurofibrillary tangles seen in AD [113].

Unlike PD, the mitochondrial genetics of AD remain nebulous. Studies have shown that inherited mtDNA variants, again mediated through mtDNA haplogroups, may affect the pathoaetiology of AD; however, like PD these studies are conflicting. Reports show that haplogroups H, U, K, T, I, W and X (seven of the ten typical European haplogroups) associate with an increased risk of developing AD, with effect sizes ranging from OR>2 [114–120]. If correct, the only haplogroups not associated with AD are haplogroups J and V, which is seemingly unlikely, especially given the reciprocal relationship between mtDNA haplogroups.

It is possible that geographical variation could affect these results, with the relative contribution of specific mtDNA variants varying in different ethnic groups; however, this would imply that the overall effect of inherited mtDNA on AD is small, or restricted to very rare pathogenic variants, and would certainly require significantly larger studies to fully investigate. A more recent and comprehensive study supports this hypothesis, investigating the role of inherited mtDNA variants in a large AD cohort concluding that there is no clear role for inherited mtDNA variation in the pathoaetiology of AD [121]. Further, when mtDNA from AD patients is transferred to Rho0 cell lines (cells devoid of native mtDNA), a biochemical defect is observed, suggesting an effect rare inheritable mtDNA abnormalities [122]; however, to date, studies have failed to find an overrepresentation of inherited mtDNA mutations in AD patients [123].

Despite strong evidence that amyloid beta (Aβ) enters mitochondria and increases ROS production [124], likely increasing mtDNA mutation formation, few studies have estimated the contribution of acquired mtDNA variation in AD. The role of the ‘common’ mtDNA deletion, a recurring pathogenic ~5 Kb mtDNA deletion, [125] was assessed in AD, but no significant association was found [126]. The same authors did identify a significantly increased acquired mtDNA mutational burden, two to threefold higher than controls, in the parietal gyrus, hippocampus and cerebellum of AD patient brains – consistent with the hypothesis of oxidative-induced cell death seen in AD [49] and supported by a more recent study [65]. Additionally, it has been hypothesised that somatic mutations in the mtDNA control region (displacement or D-loop) impair mitochondrial transcription and translation in AD brains [127].

3.3.3 Huntington’s Disease

Huntington’s disease is a comparatively early-onset neurodegenerative disorder (physical symptoms can begin at any age from infancy to old age, but usually begin between 35 and 44 years of age) and is caused by autosomal dominant repeat expansions in the Huntington gene. Early pathology begins in the striatum but rapidly progresses to other brain regions.

Mitochondrial dysfunction appears to play an important role in the progression of HD, characterised by metabolic deficits and a decrease in core mitochondrial enzymes [128]. In addition, mutant huntingtin protein (Htt) interacts with the mitochondrial membrane, detrimentally affecting calcium metabolism [129], and Htt cytotoxicity triggers mitochondrial transcriptional changes and initiates a cascade of energy failure ultimately leading to cell death [130]. At the mtDNA level, increased oxidative stress resulting from Htt toxicity induces mtDNA lesion formation and a reduction in mtDNA copy number, which in turn triggers the vicious circle of ROS-mediated mtDNA damage [130]; however, this appears more as a consequence rather than a cause of disease progression. HD patients have higher frequencies of mtDNA deletions when compared to age-matched controls; however, much of this is based on measurements in peripheral blood [131, 132], with only one study investigating cortical tissue [133].

Studies of inherited mtDNA variation are conflicted. Early work investigating the role of mtDNA haplogroups in HD concluded that they do not affect disease development or progression [134]; however, this study was relatively small compared to typical genetic association studies. Contrary to this, a more recent study demonstrated that HD patients with a haplogroup H (the commonest European haplogroup, ~40–45 % population) background showed significantly higher ATP concentrations than non-H HD patients [135]. This is puzzling, given the synonymous nature of the haplogroup H defining mtDNA variant (m.7028C, Ala375Ala in MTCOI), and likely reflects either a statistical power inadequacy or haplogroup tagging or rare mtDNA variants, which only analysis of whole mtDNA genome sequencing could identify.

3.3.4 Amyotrophic Lateral Sclerosis

Amyotrophic lateral sclerosis (ALS, often termed Lou Gehrig’s disease in the USA, Charcot disease or motor-neurone disease in the UK) is a typically sporadic disease (~95 % of cases), causing progressive loss of neurons in the motor cortex and spinal cord. Clinical hallmarks of ALS include muscle stiffness and twitching, gradually progressing to atrophy [136].

Although the underlying mechanism is not fully understood, reports link nDNA mutations in superoxide dismutase 1 (SOD1), one of three superoxide dismutases responsible for destroying free superoxide radicals, to mitochondrial dysfunction in ALS patients [137]. Mutant SOD1 inhibits voltage-dependent anion selective protein channel 1 (VDAC1) [138], reducing energy production and driving an increase in oxidative damage in transgenic mice motor neurons [139]. ALS patients were found to harbour significantly higher levels (>30×) of the mtDNA common deletion, compared to matched controls, in the motor cortex [140], and spinal cord [141]. It is likely, given the reported SOD1-mediated reduction in mtDNA repair enzymes in ALS patients [142], that mtDNA lesions and point mutations are proliferating; although this has yet to be fully investigated.

3.3.5 Friedreich’s Ataxia

Friedreich’s ataxia (FA) is the commonest form of hereditary ataxia, pathophysiologically characterised as a reduction of the mitochondrial protein frataxin in the peripheral sensory nerves, dorsal root ganglia, posterior columns, spinocerebellar and corticospinal tracts of the spinal cord, gracile and cuneate nuclei, dorsal nuclei of Clarke and dentate nucleus [143]. This deficiency causes an accumulation of iron in mitochondria and a reciprocal depletion in neuronal cytosol, which in turn leads to respiratory chain dysfunction and an increase in oxidative stress [144]. At the mtDNA level, frataxin-deficient yeast hybrid cell lines show a progressive loss of mtDNA and a reciprocal rise in mtDNA lesion formation [145, 146].

3.3.6 Hereditary Spastic Paraplegia

Hereditary spastic paraplegia (HSP) is not a single entity, rather a group of neurodegenerative disorders phenotypically characterised by progressive lower extremity weakness and spasticity (progressive stiffness and contraction) [147]. Mitochondrial dysfunction is linked to two forms of HSP: SPG13, the mitochondrial chaperonin Hsp60 [148], and SPG7, a nuclear-encoded mitochondrial metalloprotease protein which increases ROS [149].

Investigations into the effects of inherited mtDNA mutation on HSP are limited. An mtDNA mutation in MTATP6 (m.9176T>C) was identified in a large HSP pedigree [150] and a mutation in the mitochondrial 12 s rRNA (m.1432A>G) was identified in an Amish family who demonstrated abnormal mRNA maturation [151]. A single study investigating the role of mtDNA haplogroups appears negative, but is largely inconclusive due to low sample size and genetic heterogeneity [152].

3.3.7 Multiple Sclerosis

Multiple sclerosis (MS) is one of the commonest neuroinflammatory diseases in the world [153] and is phenotypically characterised by autonomic, visual, motor and sensory deficits [154]. Until relatively recently, MS pathology was restricted to the formation of lesions or plaques and the destruction of myelin in the spinal cord; however, improvements in imaging technology confirm that neurodegeneration begins in the grey and white matter of MS patients [155, 156].

Mitochondrial dysfunction appears a common component of MS. N-Acetylaspartate and N-acetylaspartylglutamate (NAA), which correspond to MS relapses, are markers for neuronal integrity produced by mitochondria and indicate mitochondrial functionality [157]. RC chain function, particularly complex I, is reduced in MS tissue and complexes I and III are reduced in non-lesional MS motor cortex tissue [158, 159]. Further, there are established links between inherited mtDNA mutation and MS, with female primary LHON (discussed previously) mutation carriers presenting with an MS-like phenotype – termed ‘Harding’s disease’ [160]. In support, m.3460G>A (a primary LHON mutation) was identified in a small cohort of neuromyelitis optica cases, where spinal nerve degeneration is a predominant feature [161], indicating that inherited mtDNA variants could modulate MS pathogenesis. However, subsequent studies focusing on the role of LHON mutations (m.3460G>A, m.11778G>A and m.14484T>C) refute a link to MS [162, 163].

Early attempts to link mtDNA haplogroups to MS indicate that European super-haplogroup ‘UK’ increases MS risk, with a reciprocal risk reduction for individuals on a super-haplogroup ‘JT’ background [164, 165]. These associations were confirmed in a much larger investigation into inherited mtDNA mutations in late-onset disease, with haplogroup U increasing MS risk and haplogroup J/K reducing MS risk [99].

Attempts to identify somatic or acquired mtDNA mutations, focusing on the identification of mtDNA deletions in normal-appearing grey and white matter or paraspinal muscle, failed to find a significant difference between MS cases and controls [166, 167].

4 Conclusion

The modulation of mitochondrial function is a key component to neurodegeneration, both during ‘normal’ ageing and disease (Fig. 3.2). Changes in mitochondrial function as we age and when disease manifests are often mediated by mitochondrial DNA variation, which can be either inherited (typically single base-pair exchanges) or acquired as we age (either single base-pair exchanges or mtDNA deletions).

Fig. 3.2
figure 2

Summary of the complex interaction between mitochondrial dysfunction, mtDNA mutation and neurodegeneration

Full size image

Despite strong evidence of mitochondrial involvement in several neurodegenerative diseases, the ubiquity of mitochondria makes identifying disease-causing mtDNA mutations, and how they interact with disease-specific pathologies, difficult. Nevertheless, if we are to fully understand if this phenomenon is the cause or correlation, it is important that studies of mtDNA in neurodegenerative disease continue.

As the evidence demonstrates, mitochondrial function, mediated through genomic flux, can play a role in neuronal loss, making mitochondria a useful target for therapeutic strategies in an increasingly aged population.