A decline in mitochondrial function has long been shown to exist in neurodegenerative disease. Whether this decline is a secondary consequence of other factors or whether it causes the eventual death of a cell is unknown. In this review, we will discuss some of the major evidence surrounding mitochondrial DNA mutations leading to mitochondrial dysfunction in neurodegenerative disease and discuss their possible role in neurodegeneration.
The vast majority of a cell's energy is produced by mitochondria in a process of OXPHOS (oxidative phosphorylation). Mitochondrial dysfunction can lead to reduced ATP production and increased production of reactive oxygen species. mtDNA (mitochondrial DNA) mutations have been widely described to increase in many tissues with aging as well as in neurodegenerative disease [1–4]. The effect that these mtDNA mutations have on the age-related decline in tissue function is uncertain, but one theory suggests that they cause increased oxidative damage and contribute to the aging process . Interestingly, one of the major risk factors for developing neurodegenerative disease is age; whether or not this relates to a decline in mitochondrial function is yet to be established.
Mitochondria contain their own DNA, located within the matrix, which is distinct and replicates independently of nuclear DNA. Apart from a small ∼1 kb non-coding region, mtDNA does not contain introns; therefore expression of the rest of the 16.5 kb molecule is crucial to maintain the OXPHOS process located on the inner membrane within mitochondria. There are multiple copies of mtDNA within each cell. Therefore, if a mutation does occur within the mitochondrial genome, it can exist among wild-type copies, a situation known as heteroplasmy. It is not until an mtDNA mutation reaches a certain threshold that a biochemical defect such as a respiratory chain deficiency is observed. Thresholds can vary depending on mutation and cell type, but are typically ∼60% for mtDNA deletions  and ∼85% for point mutations [7,8].
Pathogenic mutations such as large-scale deletions and point mutations of the mitochondrial genome can cause a wide range of classical mitochondrial disorders . Large-scale deletions generally cause sporadic disorders, such as PEO (progressive external ophthalmoplegia), PS (Pearson's syndrome) and KSS (Kearns–Sayre syndrome). Large-scale deletions are generally thought to be the result of polymerase stalling during replication, leading to deletions forming at direct repeats . Consequently, mutations in nuclear genes involved in mtDNA replication, such as the mitochondrial polymerase POLG1 or the helicase TWINKLE or ANT1 (adenine nucleotide translocator 1), have been associated with PEO [11–13]. Point mutations can result in both maternally inherited and sporadic diseases. Point mutations in protein-encoding genes can cause LHON (Leber's hereditary optic neuropathy) and specific point mutations in tRNA genes can result in MELAS (mitochondrial encephalomyopathy, lactic acidosis and stroke-like episodes) and MERRF (myoclonic epilepsy with ragged red fibres) . In mtDNA disease, the specific mtDNA mutation has exceeded the critical threshold of mutation:wild-type, and a histological hallmark of this phenomenon is observed by the presence of COX (cytochrome c oxidase)-deficient cells in a mosaic within the affected tissues . One of the interesting observations on mtDNA disease is that the tissues that are affected are usually those with a high energy demand such as muscle and brain . Therefore symptoms can often be similar to those suffering with more typical neurodegenerative diseases.
PD (Parkinson's disease)
Mitochondrial dysfunction in PD first came to light in 1983 when heroin drug users injected MPTP (1-methyl-4-phenyl-1,2,5,6-tetrahydropyridine) and developed striking Parkinsonian features . MPTP is a toxin that inhibits complex I of the respiratory chain. MPTP injected into animal models also developed Parkinsonian symptoms with selective loss of the dopaminergic neurons of the SN (substantia nigra) . A reduction in complex I has been reported in PD SN [16–19]. The underlying cause of the complex I deficiency in PD patients is unknown.
In 2004, mutations in POLG1 associated with PEO were also shown to co-segregate with Parkinsonian symptoms in these patients . Neuropathological examination showed loss of dopaminergic neurons, although without Lewy bodies, which are a characteristic hallmark of PD. Since then, there have been other reports of other autosomal dominant PEO patients with POLG1 mutations presenting with Parkinsonism, which highlights a potential link between the accumulation of mtDNA mutations and cell loss in the SN [21–24].
In PD, Lewy bodies have been shown to consist largely of α-synuclein along with a number of other proteins and also interestingly mitochondria [25,26]. Although there is no evidence as yet of a mitochondrial association or localization of α-synuclein, MPTP treatment in overexpressing α-synuclein mice produces abnormal and enlarged mitochondria . The damaging effects of MPTP in mice were shown to be absent from α-synuclein-null-containing mice; however, rotenone could still inhibit complex I . This suggests that the inhibition of complex I by rotenone may lie upstream of that of MPTP. Recent studies using double-mutant α-synuclein knock-in mice also produced abnormal mitochondria in neuronal somata together with reduced complex I activity in the SN .
If the cause of the PD is mitochondrial dysfunction, then transferring the mitochondria from affected cells into normal ones should transfer the defect. There have been a number of studies transferring the mitochondria from PD patients into mitochondria-less cells to identify whether there is evidence to suggest a direct role of mitochondria in the onset of PD. Studies have proved inconclusive; in one study, cybrids were created using mtDNA from platelets of patients with PD and were shown to form inclusions resembling Lewy bodies ; however, other studies have failed to show the cytoplasmic transmission of a mitochondrial functional defect [31,32]. Also, no specific mtDNA mutation has been identified as the cause of the transmission of abnormalities. The major problem with these studies is the use of platelets as the source of mitochondria, as the major mitochondrial defects will be located in the brain regions affected, such as the SN. Obviously technical difficulties prevent the transfer of the neuronal mitochondria.
As discussed above, reports in the literature support a decline in mitochondrial function, such as complex I of the respiratory chain, in PD. But is this decline due to mutations of the mitochondrial genome? Studies trying to identify whether certain mitochondrial haplogroups confer susceptibility to developing PD have concluded that haplogroups JTUK are associated with some decrease in risk [33,34]. Some groups have searched for point mutations in the seven mitochondrial-encoded complex I genes in tissue samples from the frontal cortex of PD patients and age-matched controls. Low levels of several mtDNA mutations were detected on the MTND5 gene in PD compared with controls and the authors hypothesized that the cumulative effect of these low-level mtDNA mutations may cause PD [35,36]. However, the actual effect of multiple low-level mtDNA mutations remains unknown. More recently, we and others have found high levels of sporadic large-scale mtDNA deletions in SN neurons accumulating with age, with a further increase observed in PD compared with age-matched controls [1,2]. These two reports were the first to show high levels of mtDNA mutations associated with aging. What is interesting is that the high levels of mtDNA deletions detected in the PD patients mirror the high levels of mtDNA deletions seen in those PEO patients with high levels of multiple mtDNA deletions caused by a mutation in POLG1 and the onset of Parkinsonism. With the common link between these two disease groups being the high levels of mtDNA deletions in SN neurons, it is tempting to speculate that they may have some role in the eventual death of these neurons. These recent results are interesting and further investigation is required in this area to establish the actual role that these mtDNA deletions play in the neurodegeneration observed in PD.
AD (Alzheimer's disease)
Mitochondrial dysfunction is also well described in AD. A reduction in complex IV/COX, the terminal complex of the mitochondrial genome, has been described in brain regions in AD cases [37,38]. An increase in COX-deficient fibres in the hippocampus and choroid plexus has been detected in AD cases compared with controls . Cybrid work transferring AD mitochondria into mitochondria-less cell lines much like the PD cybrid work, remains inconclusive. Some studies have shown that the COX defect is transferred into the cybrids, while other groups fail to reproduce these results [32,40]. Again, no specific mtDNA mutation has been identified as the cause of the COX deficiency in the cybrid cell lines.
The underlying cause of the mitochondrial dysfunction in AD is still unknown. One study found evidence for increased mtDNA mutations in the control region in AD compared with controls , while a larger cohort study did not detect a difference . There are mixed reports of mitochondrial haplogroup association with AD, with a suggestion of excess of haplogroup J in AD patients from a French-Canadian population , whereas no such association has been observed in other studies ; these differences are most likely to be due to limitations in sample size.
A single report has described a patient with autosomal dominant PEO caused by a POLG1 mutation, who also developed Alzheimer's symptoms before the patient died at age 60 . The patient had an ApoE4/4 (apolipoprotein E4/4) genotype and neuropathological examination showed severe loss of dopaminergic neurons of the SN with no Lewy bodies, but the patient did have neurofibrillary tangles in the hippocampus, entorhinal cortex, amygdale and neuritic plaques, plus evidence of phosphorylated tau protein in the neocortex. The patient had frequent COX-deficient muscle fibres and multiple deletions in different brain regions, as expected from a POLG1 mutation patient. Although the development of AD symptoms could be a coincidence in this autosomal dominant PEO patient, the high levels of somatic mutations may have caused premature aging and this was reflected by the early onset of AD in this patient.
The evidence for mtDNA mutations in neurodegenerative disease discussed above largely relies on work performed on post-mortem tissue; therefore it is often difficult to prove actual causality of the mtDNA mutations in the disease. The development of transgenic mice with deficiencies in mitochondrial function came to the forefront in 2005, when mice were genetically engineered to have increased mtDNA mutations caused by a mutated POLG1 gene resulting in reduced proofreading capabilities . These mice appeared normal for the first 25 weeks of life, but then started to show symptoms of premature aging, with most mice dying before ∼60 weeks of age, much sooner than normal aging, which typically lasts for 2 years. If aging is the major factor in neurodegenerative disease, it would be interesting to study these mice in relation to the onset of such disorders. The same group also developed transgenic mice deficient in TFAM (mitochondrial transcription factor A) targeted to the dopaminergic neurons of the SN . TFAM is a protein necessary for mtDNA maintenance by regulating copy number by binding the mitochondrial genome and coating it. The transgenic mice developed selective neuronal loss in the SN and progressively developed key symptoms of PD. The mice had reduced mtDNA copy number, leading to an increased number of COX-deficient cells. Another study, where the authors developed transgenic mice overexpressing amyloid β-peptide, showed increased immunostaining of a 5 kb mtDNA deletion compared with wild-type mice in vascular walls . All these models highlight the importance of mitochondrial integrity for the normal functioning of cells. The POLG1 and the TFAM mice give much-needed support to the involvement of mtDNA mutations in cell function decline and the possible contribution to cell loss.
In conclusion, mtDNA mutations continue to feature prominently in studies searching for a decline in mitochondrial function in many neurodegenerative diseases. The presence of high levels of mtDNA deletions in SN neurons in PD patients along with the Parkinsonism observed in multiple mtDNA deletion patients is intriguing. In mtDNA disease, the specific mtDNA mutation is known to cause the disease and once the threshold for COX deficiency is reached, cell death is observed in the affected region. Future work will involve unravelling the connection between the accumulation of high levels of mtDNA mutations and neurodegeneration both in normal aging and in neurodegenerative disease.
Central Nervous System: A Focus Topic at Life Sciences 2007, held at SECC Glasgow, U.K., 9–12 July 2007. Edited by C. Dart (Liverpool, U.K.), M. Houslay (Glasgow, U.K.), M. Ludwig (Edinburgh, U.K.), R. Porter (Trinity College Dublin, Ireland) and J. Potts (Misouri-Columbia, U.S.A.).