- Split View
-
Views
-
Cite
Cite
Gavin Hudson, Patrick F. Chinnery, Mitochondrial DNA polymerase-γ and human disease, Human Molecular Genetics, Volume 15, Issue suppl_2, 15 October 2006, Pages R244–R252, https://doi.org/10.1093/hmg/ddl233
- Share Icon Share
Abstract
The maintenance of mitochondrial DNA (mtDNA) is critically dependent upon polymerase-γ (pol-γ), encoded by the nuclear gene POLG. Over the last 5 years, it has become clear that mutations of POLG are a major cause of human disease. Secondary mtDNA defects characterize these disorders, with mtDNA depletion, multiple mtDNA deletions or multiple point mutations of mtDNA in clinically affected tissues. The secondary mtDNA defects cause cell and tissue-specific deficiencies of mitochondrial oxidative phosphorylation, leading to organ dysfunction and human disease. Functional genetic variants of POLG are present in up to ∼0.5% of the general population, and pathogenic mutations have been described in most exons of the gene. Clinically, POLG mutations can present from early neonatal life to late middle age, with a spectrum of phenotypes that includes common neurological disorders such as migraine, epilepsy and Parkinsonism. Transgenic mice and biochemical studies of recombinant mutated proteins are helping to unravel mechanisms of pathogenesis, and patterns are beginning to emerge relating genotype to phenotype.
INTRODUCTION
The production of adenosine triphosphate (ATP) by mitochondrial oxidative phosphorylation is critically dependent upon the structural integrity of the mitochondrial genome (mtDNA). MtDNA codes for 13 essential polypeptide subunits of the respiratory chain, which combine with over 70 nuclear subunits to form the final common pathway for energy metabolism (1–3). Pathogenic mutations of mtDNA lead to relative ATP depletion, cellular dysfunction and ultimately cell death. Until recently, it was thought that primary mtDNA mutations were the major cause of mitochondrial disease in humans (4,5), particularly those presenting in adult life (6). However, an emerging class of autosomal mitochondrial diseases account for a rapidly growing clinical group: disease due to mutations in the gene coding for the mtDNA polymerase-γ (POLG), which cause disease through a secondary effect on mtDNA.
This review focuses on recent advances in our understanding of inherited POLG mutations and human disease. An up-to-date mutation database can be found at http://dir-apps.niehs.nih.gov/polg/. In this article, we will initially refer to substitutions using conventional numbering from the codon initiating translation on the cDNA (NM 002693) and the predicted amino acid substitution. Subsequent reference to the same substitution will only refer to the amino acid change. Polymerase-γ (pol-γ) has been implicated in a number of non-inherited disorders, including nucleoside analogue toxicity (7), which are not the topic of this review.
POLYMERASE-γ AND THE MTDNA REPLISOME
Unlike nuclear DNA, which only replicates once during cell division, mtDNA is continuously recycled, independent of the cell cycle (see Fig. 1 for details) (8,9). MtDNA replication is achieved by a number of nuclear-encoded proteins (Fig. 1; Table 1), including the only DNA polymerase present in the mammalian mitochondria: pol-γ (10). In humans, mtDNA is copied by a 195 kDa heterotrimer consisting of a catalytic subunit (p140, coded by POLG on chromosome 15q25) and two identical accessory subunits (p55, coded by POLG2 on chromosome 17q) (11). The catalytic subunit (p140) possesses DNA polymerase (11), 3′→5′ exonuclease (12) and 5′dRP lyase activities (13), whereas the accessory subunit (p55) is a DNA binding factor that confers high processivity by increasing the affinity of the heterotrimer for template DNA (12). Pol-γ has a high base-substitution fidelity (<2×10−6 errors per nucleotide) and is relatively accurate over short repeat sequences, but longer homopolymeric tracts (>4 bp) lead to slippage during replication (14).
. | Gene . | Chromosome . | Product . | Size (kDa) . |
---|---|---|---|---|
mtDNA replication | POLG 1 | 15q25 | Catalytic unit | 140 |
POLG 2 | 17q23–24 | Accessory subunit | 55 | |
SSBP 1 | 7q34 | Single-stranded binding protein | 15 | |
PEO 1 | 10q23.3 | T7-primase/helicase (twinkle) | 77 | |
TOP1MT | 8q24.3 | Topoisomerase (DNA) I, mitochondrial | 67 | |
TOP3a | 17p12–11.2 | Topoisomerase (DNA) III-alpha | 112 | |
LIG3 | 17q11.2–12 | Ligase III, DNA, ATP-dependent | 96 | |
RNA transcription | POLR2E | 19q13.3 | Core RNA polymerase | 150 |
TFAM | 10.21 | Transcription factor A | 24 | |
TFB1M | 6q25.1 | Transcription factor B1 | 40 | |
RMRP | 9p21–p12 | Component of RNase MRP RNA | 275–285 nt | |
RNASEH1 | 2p25 | Ribonuclease H1 | 32 | |
mtDNA repair | UDG | 12q23–q24.1 | Uracil-DNA-glycosylase | 27.5 |
OGG1 | 3p26.2 | 8-Oxygaunine-DNA-glycosylase | 38 | |
NTHL1 | 16p13.3 | Nth Endonuclease III-like (Escherichia coli) | 34 | |
MUTYH | 1p34.3–p32.1 | mutY homologue (E. coli) | 60 | |
APEX2 | xp11.21 | APEX nuclease (apurinic/apyrimidinic endonuclease) 2 | 57 | |
ENDOG | 9q34.1 | Endonuclease G | 30 |
. | Gene . | Chromosome . | Product . | Size (kDa) . |
---|---|---|---|---|
mtDNA replication | POLG 1 | 15q25 | Catalytic unit | 140 |
POLG 2 | 17q23–24 | Accessory subunit | 55 | |
SSBP 1 | 7q34 | Single-stranded binding protein | 15 | |
PEO 1 | 10q23.3 | T7-primase/helicase (twinkle) | 77 | |
TOP1MT | 8q24.3 | Topoisomerase (DNA) I, mitochondrial | 67 | |
TOP3a | 17p12–11.2 | Topoisomerase (DNA) III-alpha | 112 | |
LIG3 | 17q11.2–12 | Ligase III, DNA, ATP-dependent | 96 | |
RNA transcription | POLR2E | 19q13.3 | Core RNA polymerase | 150 |
TFAM | 10.21 | Transcription factor A | 24 | |
TFB1M | 6q25.1 | Transcription factor B1 | 40 | |
RMRP | 9p21–p12 | Component of RNase MRP RNA | 275–285 nt | |
RNASEH1 | 2p25 | Ribonuclease H1 | 32 | |
mtDNA repair | UDG | 12q23–q24.1 | Uracil-DNA-glycosylase | 27.5 |
OGG1 | 3p26.2 | 8-Oxygaunine-DNA-glycosylase | 38 | |
NTHL1 | 16p13.3 | Nth Endonuclease III-like (Escherichia coli) | 34 | |
MUTYH | 1p34.3–p32.1 | mutY homologue (E. coli) | 60 | |
APEX2 | xp11.21 | APEX nuclease (apurinic/apyrimidinic endonuclease) 2 | 57 | |
ENDOG | 9q34.1 | Endonuclease G | 30 |
Adapted from Graziewicz et al. (87).
. | Gene . | Chromosome . | Product . | Size (kDa) . |
---|---|---|---|---|
mtDNA replication | POLG 1 | 15q25 | Catalytic unit | 140 |
POLG 2 | 17q23–24 | Accessory subunit | 55 | |
SSBP 1 | 7q34 | Single-stranded binding protein | 15 | |
PEO 1 | 10q23.3 | T7-primase/helicase (twinkle) | 77 | |
TOP1MT | 8q24.3 | Topoisomerase (DNA) I, mitochondrial | 67 | |
TOP3a | 17p12–11.2 | Topoisomerase (DNA) III-alpha | 112 | |
LIG3 | 17q11.2–12 | Ligase III, DNA, ATP-dependent | 96 | |
RNA transcription | POLR2E | 19q13.3 | Core RNA polymerase | 150 |
TFAM | 10.21 | Transcription factor A | 24 | |
TFB1M | 6q25.1 | Transcription factor B1 | 40 | |
RMRP | 9p21–p12 | Component of RNase MRP RNA | 275–285 nt | |
RNASEH1 | 2p25 | Ribonuclease H1 | 32 | |
mtDNA repair | UDG | 12q23–q24.1 | Uracil-DNA-glycosylase | 27.5 |
OGG1 | 3p26.2 | 8-Oxygaunine-DNA-glycosylase | 38 | |
NTHL1 | 16p13.3 | Nth Endonuclease III-like (Escherichia coli) | 34 | |
MUTYH | 1p34.3–p32.1 | mutY homologue (E. coli) | 60 | |
APEX2 | xp11.21 | APEX nuclease (apurinic/apyrimidinic endonuclease) 2 | 57 | |
ENDOG | 9q34.1 | Endonuclease G | 30 |
. | Gene . | Chromosome . | Product . | Size (kDa) . |
---|---|---|---|---|
mtDNA replication | POLG 1 | 15q25 | Catalytic unit | 140 |
POLG 2 | 17q23–24 | Accessory subunit | 55 | |
SSBP 1 | 7q34 | Single-stranded binding protein | 15 | |
PEO 1 | 10q23.3 | T7-primase/helicase (twinkle) | 77 | |
TOP1MT | 8q24.3 | Topoisomerase (DNA) I, mitochondrial | 67 | |
TOP3a | 17p12–11.2 | Topoisomerase (DNA) III-alpha | 112 | |
LIG3 | 17q11.2–12 | Ligase III, DNA, ATP-dependent | 96 | |
RNA transcription | POLR2E | 19q13.3 | Core RNA polymerase | 150 |
TFAM | 10.21 | Transcription factor A | 24 | |
TFB1M | 6q25.1 | Transcription factor B1 | 40 | |
RMRP | 9p21–p12 | Component of RNase MRP RNA | 275–285 nt | |
RNASEH1 | 2p25 | Ribonuclease H1 | 32 | |
mtDNA repair | UDG | 12q23–q24.1 | Uracil-DNA-glycosylase | 27.5 |
OGG1 | 3p26.2 | 8-Oxygaunine-DNA-glycosylase | 38 | |
NTHL1 | 16p13.3 | Nth Endonuclease III-like (Escherichia coli) | 34 | |
MUTYH | 1p34.3–p32.1 | mutY homologue (E. coli) | 60 | |
APEX2 | xp11.21 | APEX nuclease (apurinic/apyrimidinic endonuclease) 2 | 57 | |
ENDOG | 9q34.1 | Endonuclease G | 30 |
Adapted from Graziewicz et al. (87).
Other components of the mtDNA replisome include the mitochondrial single-stranded binding protein (mtSSB), which probably stabilizes single-stranded regions of DNA at replication forks, enhancing pol-γ activity (15). Twinkle is a 5′→3′ helicase which unwinds double-stranded mtDNA at the replication fork, facilitating mtDNA synthesis (16). Twinkle may also regulate mtDNA copy number within cells (17) and act as an mtDNA primase (18). Both RNase MRP and endonuclease G may modify the D-loop RNA primer that precedes heavy strand replication (19). A small proportion of cellular RNase H1 localizes to mitochondria and might be involved in the processing of Okazaki fragments in strand symmetric replication (Fig. 1). Several topoisomerases have been identified in humans, including mitochondrial topoisomerase 1 (TOPOImt) (20) and IIIα (hTOP3α) (21) regulating supercoiling of mtDNA, and DNA ligase III is also present in the mitochondria (22). Finally, the close relationship between the mitochondrial transcription factor A (mtTFA) and mtDNA levels (23) suggests that mtTFA binds to mtDNA as a chaperone, protecting against oxidant damage (24).
Factors regulating the expression of the replisome are poorly understood. Nuclear respiratory factor-1 (NRF-1) is a transcription factor that regulates the expression of many mitochondrial proteins (25). NRF-1 expression is related to cellular ATP levels and binds to promoter regions of POLG, POLG2 and mtTFA (25).
PATHOGENIC MUTATIONS OF POLG
Autosomal dominant syndromes
The first pathogenic mutations in POLG were identified in families with autosomal dominant chronic progressive external ophthalmoplegia (adPEO, MIM 157640) (26). Affected individuals had multiple different secondary deletions of mtDNA in skeletal muscle causing a mosaic cytochrome c oxidase (COX) defect (complex IV of the mitochondrial respiratory chain) (27). A high incidence of psychiatric disease, a Parkinsonian syndrome, and primary gonadal failure have also been documented in some families transmitting dominant POLG mutations (28,29).
Autosomal recessive syndromes
Compound heterozygous POLG mutations were also identified in patients with sporadic and recessive PEO (26). Some cases have a profound peripheral neuropathy and ataxia, similar to the previously described SANDO syndrome (sensory ataxic neuropathy with dysarthria and ophthalmoparesis) (30). Recessive POLG mutations also present with adult-onset ataxia without ophthalmoplegia [also called mitochondrial recessive ataxia syndrome (MIRAS)] (31,32), which appears to be common in Scandinavia because of a founder effect (32). Parkinsonism and peripheral neuropathy have recently been described as the presenting feature of recessive POLG mutations (33).
Recessive POLG mutations are also a major cause of the Alpers–Huttenlocher syndrome (MIM 203700), which is a severe hepatoencephalopathy with intractable seizures and visual failure which presents in early childhood and is associated with depletion (loss) of mtDNA in affected tissues (33–38).
MOLECULAR PATHOPHYSIOLOGY
Biochemical consequences of POLG substitutions
Recombinant proteins have been generated for a limited number of POLG mutations and studied biochemically. The dominant mutations c.2828G→A (R943H) and c.2864A→G (Y955C) alter amino acid side chains that interact with dinucleotide triphosphates and have <1% of wild-type polymerase activity (39), whereas the c.2768G→A (G923D) and c.2869G→C (A957S) have <50% of wild-type polymerase activity (39). G923D, R943H and A957S show a slightly increased DNA binding capacity, in keeping with the proposed dominant negative mechanism, where the mutated polymerase competes with the wild-type polymerase for the mtDNA template (39).
c.1399G→A (A467T) has been described in homozygous or heterozygous individuals with autosomal recessive Alpers–Huttenlocher syndrome or late-onset ataxia (32,35,40,41). Recombinant A467T protein has <5% of the wild-type polymerase activity (42) and impairs interaction between the catalytic (p140) and accessory (p55) subunits (43).
Some recessive POLG mutations are predicted to truncate the pol-γ protein. Two children homozygous for a predicted stop codon mutation c.2617G→T (E873X) had a normal birth and early infancy (34). Some full-length transcript must, therefore, have been expressed in these individuals, possibly through mRNA editing or because some ribosomes read through the stop codon (34). Shorter mRNA species generated by stop codon mutations are removed by nonsense mediated decay (44).
Secondary mtDNA defects due to POLG mutations
A high frequency of mtDNA point mutations in the non-coding mtDNA control region in muscle from POLG patients (45) raised the possibility that the point mutations are instrumental in the formation of mtDNA deletions. However, point mutations are equally present in both deleted and non-deleted molecules (46), do not cluster around the deletion breakpoints (47) and are generally not found in COX negative muscle fibres (47), questioning their role in pathogenesis. The preferential localization of deletion breakpoints around homopolymeric tracts points towards replication stalling as a mechanism (46), but homopolymeric tracts are common in mtDNA (48), and a much larger study is required before firm conclusions can be drawn.
It is intriguing that the heterozygous POLG point mutations that cause point mutations and deletions of mtDNA can also cause mtDNA depletion when there are two mutated alleles. Moreover, the same POLG mutations can cause an early onset encephalomyopathy with severe mtDNA depletion or late-onset PEO with ataxia [for example, A467T homozygotes (35,40,41)]. The reasons for this are not fully understood.
Transgenic mice
Mice with a homozygous mutation in the second exonuclease domain of pol-γ (D257A) developed increased levels of both mtDNA point mutations and mtDNA deletions when compared with controls, associated with a reduced life span and classical ageing phenotypes including hair loss, kyphosis, osteoporosis and cardiac enlargement (49). The accumulation of secondary mtDNA point mutations was linear during life and not associated with the increased production of oxygen-free radical species (50). These observations confirm the potential importance of mitochondrial mechanisms in ageing, but cast doubt on the ‘vicious circle’ hypothesis, whereby the mutations lead to respiratory chain dysfunction, which further compromises the mitochondrial genome. In the second model, a 2 bp substitution corresponding to positions 1054 and 1055 of mouse POLG cDNA was introduced, resulting in a conserved amino acid substitution in the exonuclease domain (D257A) (51). This also led to the accumulation of somatic mtDNA mutations during life, associated with the induction of apoptotic markers, particularly in rapidly dividing tissues (51).
PHENOTYPIC CONSEQUENCES OF POLG SUBSTITUTIONS
Over 70 different substitutions have been described in POLG in over 200 published cases (June 2006), including missense mutations, nonsense mutations and splice-site variants (Fig. 2). Although there is no direct relationship between genotypes, some patterns are beginning to emerge.
To date, the vast majority of dominant mutations directly affect the polymerase domain of the protein, although there are exceptions. A467T alters the linker region of the protein and segregated with autosomal dominant ptosis in an Austrian family (42). However, because mutation also affects processivity of the enzyme, a categorical subdivision of the gene is probably an over simplification (43). This is supported by our recent observation of another heterozygous linker-region mutation segregating with adPEO (52).
In contrast, the vast majority of patients with recessive disease have one heterozygous mutation in the linker region and one in the polymerase region (Fig. 2A). A467T is present at high frequencies in some European populations (Fig. 2C), acting as a reservoir for recessive disease. In two series, ∼60% of childhood onset cases harboured A467T (35,40), but only ∼20% of the adult-onset recessive cases (40). Intriguingly, compound heterozygotes with A467T have a significantly shorter survival than homozygotes for either A467T or c.2243G→C (W748S) (41). This supports a relative dominant negative effect for these alleles, possibly through a quaternary interaction between catalytic subunits in different heterotrimers (41).
There are now numerous case reports describing acute deterioration in both cerebral and hepatic functions shortly after exposure to valproate, both in children and in adults (36,40,41,53). The underlying mechanisms are not clear at present. Sodium valproate is known to inhibit mitochondrial fatty acid oxidation (54), but its effect on mtDNA replication has not been investigated.
Non-specific headache or migraine may precede ataxia or PEO by many years, and PEO may be a late feature if at all (32,40,41). Myoclonus and epilepsy may be the presenting features in both children and adults, including focal motor seizures and status epilepticus (41). In addition to the fluctuating encephalopathy in adults (41), stroke-like episodes have been described in children (40). Hyperkinetic movement disorders (32) have been documented in addition to Parkinsonism (28), and early onset dementia is well recognized (40,41). The psychiatric features also include severe depression and psychosis (32). Dysphagia, diabetes and deafness are well recognized, and muscle pain or isolated myopathy may be the presenting feature (40) (Table 2). A key clinical point is that at least some patients do not have evidence of mitochondrial dysfunction in muscle and may not have multiple deletions of mtDNA in muscle (31,53). This places great emphasis on clinical intuition in deciding which patients to screen for POLG mutations.
System . | Feature . | Notes/references . |
---|---|---|
Psychiatric | Depression | (28) |
Psychosis | (32,40) | |
Dementia | (40,53) | |
Seizures | Myoclonus | Common in children (40) and adults with ataxia (32,41,53) |
Focal motor seizures | (41) | |
Generalized seizures | (31,32,40) | |
Status epilepticus | (41) | |
Extrapyramidal | Parkinsonism | Responds to levodopa (28,58) |
Chorea | (32) | |
Cerebellum | Ataxia | (31,32,40,53) |
‘Cerebrovascular’ | Migraine | May precede other features by many years (32,41) |
Stroke-like episodes | Usually asymptomatic in children, diagnosed on imaging (40) | |
Special sensory | Sensorineural deafness | (32,40,89,90) |
Retinopathy | (28,32,90) | |
Cataract | (28) | |
Muscle | Ptosis and external ophthalmoplegia | May be isolated ptosis (42) |
Proximal myopathy | Distal myopathy reported (40) | |
Exercise intolerance | (28,40) | |
Peripheral nerve | Sensory neuronopathy/ganglionopathy | Corresponds to SANDO (30). Profound sensory ataxia |
Axonal sensorimotor neuropathy | (33,40) | |
Endocrine | Diabetes | (40) |
Primary ovarian failure | (28,32) | |
Primary testicular failure | (89) | |
Gastrointestinal | Liver failure | Spontaneous or precipitated by sodium valproate in children (34,36,40), also in adults with ataxia (26,41) |
Gastrointestinal dysmotility | (53,89) | |
Cardiac | Cardiomyopathy | (40,53) |
System . | Feature . | Notes/references . |
---|---|---|
Psychiatric | Depression | (28) |
Psychosis | (32,40) | |
Dementia | (40,53) | |
Seizures | Myoclonus | Common in children (40) and adults with ataxia (32,41,53) |
Focal motor seizures | (41) | |
Generalized seizures | (31,32,40) | |
Status epilepticus | (41) | |
Extrapyramidal | Parkinsonism | Responds to levodopa (28,58) |
Chorea | (32) | |
Cerebellum | Ataxia | (31,32,40,53) |
‘Cerebrovascular’ | Migraine | May precede other features by many years (32,41) |
Stroke-like episodes | Usually asymptomatic in children, diagnosed on imaging (40) | |
Special sensory | Sensorineural deafness | (32,40,89,90) |
Retinopathy | (28,32,90) | |
Cataract | (28) | |
Muscle | Ptosis and external ophthalmoplegia | May be isolated ptosis (42) |
Proximal myopathy | Distal myopathy reported (40) | |
Exercise intolerance | (28,40) | |
Peripheral nerve | Sensory neuronopathy/ganglionopathy | Corresponds to SANDO (30). Profound sensory ataxia |
Axonal sensorimotor neuropathy | (33,40) | |
Endocrine | Diabetes | (40) |
Primary ovarian failure | (28,32) | |
Primary testicular failure | (89) | |
Gastrointestinal | Liver failure | Spontaneous or precipitated by sodium valproate in children (34,36,40), also in adults with ataxia (26,41) |
Gastrointestinal dysmotility | (53,89) | |
Cardiac | Cardiomyopathy | (40,53) |
System . | Feature . | Notes/references . |
---|---|---|
Psychiatric | Depression | (28) |
Psychosis | (32,40) | |
Dementia | (40,53) | |
Seizures | Myoclonus | Common in children (40) and adults with ataxia (32,41,53) |
Focal motor seizures | (41) | |
Generalized seizures | (31,32,40) | |
Status epilepticus | (41) | |
Extrapyramidal | Parkinsonism | Responds to levodopa (28,58) |
Chorea | (32) | |
Cerebellum | Ataxia | (31,32,40,53) |
‘Cerebrovascular’ | Migraine | May precede other features by many years (32,41) |
Stroke-like episodes | Usually asymptomatic in children, diagnosed on imaging (40) | |
Special sensory | Sensorineural deafness | (32,40,89,90) |
Retinopathy | (28,32,90) | |
Cataract | (28) | |
Muscle | Ptosis and external ophthalmoplegia | May be isolated ptosis (42) |
Proximal myopathy | Distal myopathy reported (40) | |
Exercise intolerance | (28,40) | |
Peripheral nerve | Sensory neuronopathy/ganglionopathy | Corresponds to SANDO (30). Profound sensory ataxia |
Axonal sensorimotor neuropathy | (33,40) | |
Endocrine | Diabetes | (40) |
Primary ovarian failure | (28,32) | |
Primary testicular failure | (89) | |
Gastrointestinal | Liver failure | Spontaneous or precipitated by sodium valproate in children (34,36,40), also in adults with ataxia (26,41) |
Gastrointestinal dysmotility | (53,89) | |
Cardiac | Cardiomyopathy | (40,53) |
System . | Feature . | Notes/references . |
---|---|---|
Psychiatric | Depression | (28) |
Psychosis | (32,40) | |
Dementia | (40,53) | |
Seizures | Myoclonus | Common in children (40) and adults with ataxia (32,41,53) |
Focal motor seizures | (41) | |
Generalized seizures | (31,32,40) | |
Status epilepticus | (41) | |
Extrapyramidal | Parkinsonism | Responds to levodopa (28,58) |
Chorea | (32) | |
Cerebellum | Ataxia | (31,32,40,53) |
‘Cerebrovascular’ | Migraine | May precede other features by many years (32,41) |
Stroke-like episodes | Usually asymptomatic in children, diagnosed on imaging (40) | |
Special sensory | Sensorineural deafness | (32,40,89,90) |
Retinopathy | (28,32,90) | |
Cataract | (28) | |
Muscle | Ptosis and external ophthalmoplegia | May be isolated ptosis (42) |
Proximal myopathy | Distal myopathy reported (40) | |
Exercise intolerance | (28,40) | |
Peripheral nerve | Sensory neuronopathy/ganglionopathy | Corresponds to SANDO (30). Profound sensory ataxia |
Axonal sensorimotor neuropathy | (33,40) | |
Endocrine | Diabetes | (40) |
Primary ovarian failure | (28,32) | |
Primary testicular failure | (89) | |
Gastrointestinal | Liver failure | Spontaneous or precipitated by sodium valproate in children (34,36,40), also in adults with ataxia (26,41) |
Gastrointestinal dysmotility | (53,89) | |
Cardiac | Cardiomyopathy | (40,53) |
MULTIPLE SUBSTITUTIONS: PHENOTYPIC MODIFIERS?
In one series, ∼30% of the patients had more than one substitution in POLG (40). Although some of these variants are well established single nucleotide polymorphisms (Fig. 2C, http://www.genome.utah.edu/genesnps/), for others it can be less clear [for example, c.752C→T (T251I) and c.1760C→T (P587L) discussed in (40)]. c.3428A→G (E1143G) is present in ∼4% of controls (Fig. 2C) and alters a highly conserved residue (55). The frequency of E1143G in patients with sporadic PEO is greater than would be expected by chance (55), and segregation analysis in families with other primary pathogenic POLG mutations suggests that E1143G interacts with other pathogenic alleles to modify the disease phenotype (40,56). On the contrary, E1143G homozygotes have been detected in healthy control subjects (40). Further clinical and functional studies are needed to resolve this important issue.
Given the intimate relationship between the various proteins involved in mtDNA replication (Fig. 1), the possibility of mutations in two genes with interacting gene products is an attractive hypothesis. In keeping with this, a novel mutation in the PEO1 gene (c.1031G→A, R334Q, formerly C10Orf2) and a novel mutation in POLG were identified in a man with PEO (57). We also detected a novel heterozygous substitution in POLG c.1282G→C (L428V) and a novel PEO1 mutation c.2038T→C (S680T) (Hudson and Chinnery, unpublished data). However, the same PEO1 mutation that was identified in the published family has also been identified in a family with adPEO in the absence of a POLG mutation, demonstrating that R334Q is capable of causing adPEO on its own (40). The majority of patients with confirmed pathogenic mutations in POLG do not have mutations in PEO1 (40), questioning the role of digenic inheritance in PEO.
The identification of multiple mutations in the same individual reaffirms the polymorphic nature of POLG and stresses the importance of screening large ethnically matched control groups before reporting an apparent disease association. This is particularly important for large-scale screening studies searching for POLG mutations in common conditions such as idiopathic Parkinson's disease. A novel heterozygous substitution (c.2492A→G, Y831C) was detected in two siblings with Parkinsonism in a family with PEO (58) and not detected in 130 control subjects, but we recently detected the same substitution in control subjects in the UK, questioning its role as a dominant mutation with reduced penetrance (Tiangyou et al. submitted for publication).
Given emerging evidence that the carrier frequency of established recessive mutations varies throughout Europe (Fig. 2C), the choice of control panels must be done with great care. Recurrence of the same mutation in different families, the absence of heterozygous asymptomatic carriers of putative dominant mutations and the absence of homozygous healthy controls would clearly support a pathogenic role. Ideally, functional studies should be carried out, either in vitro or in animal models, before reaching a firm conclusion—but this is rarely possible in a diagnostic setting. This makes the screening of appropriate control subjects and a single mutation database all the more important.
Pol-γ CAG REPEAT AND MALE INFERTILITY
In humans, there is a CAG(n) repeat sequence in exon 2 of POLG coding for a polyglutamine tract (Fig. 2). About 88–96% of Eurasian alleles have 10 CAG repeats (59). Two studies reported a difference in allele distributions between normal controls and males with male infertility (60,61), but this has not been replicated by others (62–64). Deleting the CAG repeat has no effect on respiratory chain activity in vitro (65), and it remains to be seen whether genetic variation in the repeat region has any functional consequences in vivo.
RELATED DISEASE GENES AND NEW CANDIDATES
Dominant PEO
Dominant PEO with multiple deletions can also be due to mutations in the gene coding for the mtDNA helicase Twinkle (PEO1) (66) and the adenine nucleotide translocase (ANT1) (67). Convincing recessive mutations in these genes have not been described, although sporadic cases have been identified (55).
A single heterozygous substitution in POLG2 impaired interaction between the accessory subunit and the catalytic subunit and compromised interaction between the heterotrimer and a DNA template in vitro, leading to incomplete stimulation of DNA synthesis and replication stalling, potentially generating mtDNA deletions in vivo (68). It was not possible to demonstrate a dominant negative effect of the mutant protein in vitro, pointing towards haploinsufficiency as a mechanism, possibly through heterodimerization (68). POLG2 mutations appear to be rare, accounting for ∼1% of adult onset cases of PEO with multiple deletions (68).
Recessive disorders with multiple mtDNA deletions
Recessive mutations in the gene coding for thymidine phosphorylase (TP) cause mitochondrial neurogastrointestinal encephalomyopathy (69,70) and secondary mtDNA defects including depletion, multiple deletions and point mutations (71,72).
MtDNA depletion
MtDNA depletion has also been observed in a number of recessive disorders of nucleoside metabolism, including thymidine kinase deficiency (TK2) (73) and deoxyguanosine kinase deficiency (DGUOK) (74). These two disorders present in childhood with hypotonia because of myopathy, with hepatic involvement in patients with DGUOK mutations (75). Mutations in the ADP-forming succinyl-CoA sythase (SUCLA2) were also described in a family with autosomal recessive encephalopathy with mtDNA depletion (76). In addition, mutations in MPV17, which codes for a protein which localizes to the inner mitochondrial membrane, also cause mtDNA depletion and infantile hepatic encephalomyopathy (77).
Despite the massive international effort to identify disease genes involved in mtDNA maintenance, a large proportion of patients with secondary mtDNA defects still elude diagnosis. There is a growing list of proteins known to be involved in mtDNA replication and repair in humans and other organisms (Table 1), forming an attractive list of candidate genes for these disorders.
CONCLUSION
The first pathogenic mutations in POLG were reported only 5 years ago in patients with a ‘classical’ mitochondrial phenotype: autosomal dominant PEO. The breadth of clinical phenotypes that have subsequently been linked to this gene is remarkable, from severe childhood encephalopathy to late-onset mild ptosis. Crude estimates indicate that POLG mutations will be a major cause of human mitochondrial disease, possibly more prevalent than primary mtDNA mutations in adults (A.M. Schaefer, unpublished data).
From the clinical perspective, mitochondrial disorders are generally considered to be difficult to recognize and diagnose. Paradoxically, the work on POLG in the last 5 years has made diagnosis even more difficult, overlapping with common neurological traits such as Parkinsonism and epilepsy. Although patients with POLG mutations currently seem to fall into discrete categories, the boundaries are starting to merge, pointing towards an overlapping spectrum of disease with multiple substitutions with functional consequences in vitro that may interact and modify the clinical phenotype. The public health implications of POLG mutation are potentially massive. Fortunately, a number of groups have developed a means of assessing the function of novel genetic variants [for example (39,42,78)], and two mouse models have been generated (50,51). Hopefully, these resources will provide the necessary tools to develop treatments in the near future.
ACKNOWLEDGEMENTS
P.F.C. is a Wellcome Trust Senior Fellow in Clinical Science. P.F.C. also receives funding from Ataxia (UK), the Alzheimer’s Research Trust, the Association Française contre les Myopathies, the United Mitochondrial Diseases Foundation and the EU FP program EUmitocombat and MITOCIRCLE.
Conflict of Interest statement. None declared.