Depletion of mitochondrial DNA (mtDNA) appears to be an important cause of mitochondrial dysfunction in neonates and infants. We have identified another child in whom depletion of mtDNA was demonstrated in liver and serial skeletal muscle biopsies. A primary myoblast culture from the patient initially showed normal levels of mtDNA, but there was a progressive loss of mtDNA in later cell passages and clonal myoblast cell cultures, similar to that observed in the skeletal muscle tissue of the patient. Thus, these clonal myoblast cultures provide an in vitro model of the in vivo mtDNA dynamics. The levels of mitochondrial mRNAs for subunits I and II of cytochrome c oxidase declined with declining mtDNA levels, but the fall in mitochondrial transcript levels lagged behind that of the mtDNA levels. Levels of cytochrome c oxidase subunit I and II polypeptides, however, declined ahead of declining mtDNA levels. Immunocytochemistry showed that between individual cells of the clonal myoblast cultures, the expression of the mitochondrially encoded subunit I of cytochrome c oxidase was heterogeneous, suggesting variable levels of mtDNA. Transfer of patient mitochondria with residual mtDNA levels to control cells devoid of mtDNA ([rho]0 cells) led to restoration of mtDNA levels and, hence, suggests a nuclear involvement in the depletion.
Defects in oxidative phosphorylation are found in a clinically heterogeneous group of disorders. The molecular defects that underlie these disorders may arise from mutations of either the mitochondrial or the nuclear genomes or both. Mitochondrial DNA (mtDNA) is transmitted maternally and contains genes exclusively involved in the biosynthesis of enzyme complexes of the mitochondrial oxidative phosphorylation system (1 ). In adult patients, abnormalities of oxidative phosphorylation are often associated with rearrangements or point mutations in tRNA genes affecting a subpopulation of mitochondrial genomes, i.e. the mutations are usually heteroplasmic (2 ,3 ). The variable tissue involvement observed in these patients is generally ascribed to tissue-specific differences in their dependence on oxidative phosphorylation and the degree of heteroplasmy.
In 1991, Moraes and colleagues (4 ) identified a group of infants with marked depletion of mtDNA in association with defective oxidative phosphorylation. To date, this condition has been confirmed in 29 children (4 -15 ), suggesting that it may be an important cause of mitochondrial dysfunction in neonates and infants. Most patients present soon after birth with muscle weakness and hepatic failure or renal tubulopathy associated with a severe depletion of mtDNA (88-99%) in affected tissues at post-mortem. These infants usually die before 9 months of age. A milder variant has been described in some patients, with a slowly progressive mitochondrial encephalomyopathy starting in childhood, and this is associated with a less severe depletion of mtDNA in skeletal muscle (66-86%). Patients often have hypotonia and lactic acidosis, and show a deficiency of complex IV (cytochrome c oxidase) of the oxidative phosphorylation system and other enzyme complexes containing mtDNA-encoded subunits (4 -15 ).
The molecular defect responsible for mtDNA depletion syndrome has not been identified. Differential tissue involvement in related patients (4 ) and the positive family history in one case of neuromuscular disorders in the maternal lineage (12 ) suggest mtDNA heteroplasmy. However, sequence analysis of mtDNA replication origins did not demonstrate any potentially pathogenic mutation (4 ,12 ), and two families showed an apparent paternal transmission of the disease (4 ,6 ). All patients were born to clinically normal parents, and family histories were compatible with an autosomal recessive inheritance in all but one pedigree where the pattern of inheritance could be explained by an autosomal dominant transmission with incomplete penetrance (6 ). Whether recessive or dominant, the autosomal inheritance supports the involvement of a nuclear-encoded factor. Our group has demonstrated that mtDNA levels could be restored in cultured fibroblasts expressing mtDNA depletion by complementation with nuclear DNA from a control cell line (9 ), providing further evidence that the depletion is controlled by the nuclear genome.
In the present study, we examined myoblast cell cultures from another patient with mtDNA depletion syndrome and, as before, demonstrated nuclear involvement in the mtDNA depletion using complementation experiments. There was a progressive loss of mtDNA and, by following the expression of mitochondrially and nuclear-encoded subunits of cytochrome c oxidase, we have been able to dissect a sequence of molecular events during the course of the mtDNA depletion.
We have examined myoblast cultures from a patient with mtDNA depletion syndrome. Initial studies (Morris et al., unpublished observations) revealed a 73% reduction of mtDNA levels in a skeletal muscle biopsy taken at 2 months of age and, at a post-mortem age of 7.5 months, a 91% reduction in muscle and an 89% reduction in liver compared with age-matched controls. Depleted levels of mtDNA in muscle of the patient are illustrated in Figure 1 A.
In parallel with the blotting experiments of cell culture extracts, the expression and localization of cytochrome c oxidase subunits was examined in individual cells of the clonal myoblast cell cultures by immunocytochemistry. Cultures were first stained with MitoTrackerT CMXRos-H2 to label the mitochondria, followed by immunostaining with monoclonal antibodies against subunits of cytochrome c oxidase and staining of the nuclei with bis-benzimide. In these experiments, control cultures showed a uniform expression of cytochrome c oxidase subunits in mitochondria of all cells. When cell passages with partly depleted levels of mtDNA were immunostained with antibodies against subunit I of cytochrome c oxidase, some cells showed normal levels of subunit I in mitochondria, whereas other cells had decreased levels or completely lacked this mitochondrially encoded subunit (Fig. 6 A). On the other hand, nuclear-encoded subunits IV and VIc were present in the mitochondria of all cells but the amount of immunoreactive material varied from cell to cell (Fig. 6 B and C). When partly depleted cell passages were stained for cytochrome c oxidase activity, a proportion of the cells appeared to have a relatively normal enzyme activity, whereas other cells lacked activity (not shown). In cells from passages completely devoid of mtDNA, neither cytochrome c oxidase activity nor subunit I could be detected, but subunits IV and VIc were still present in mitochondria of all cells, though at a lower level than in control cells (not shown).
Finally, we investigated whether residual mtDNA in the patient's myoblasts could replicate and restore mitochondrial expression in a normal nuclear environment. Myoblasts cultures from clone 7 with mtDNA levels at 20% of controls were enucleated, and these cytoplasts were fused with [rho]0A549.B2 cells which are devoid of mtDNA. From preceding Southern analyses, we knew that the myoblasts used for mitochondrial transfer would essentially lose their residual mtDNA within four passages and, like [rho]0A549.B2, would become auxotrophic for pyruvate and uridine (16 ). As both fusion partners depend on pyruvate and uridine for growth, complementation was examined by plating the fusion products in medium lacking these nutrients. Colonies were observed 10 days after fusion and ring-cloned. These presumptive transmitochondrial cybrid cell lines had the morphology and growth characteristics of the parental A549 cell line. No colonies were found in dishes plated with either enucleated myoblasts used as donor or [rho]0A549.B2 (not fused) in selective medium.
Genomic DNA was isolated from two cybrid cell lines, designated A12 and A17, and analysed together with the fusion partners and the parental A549 cell line. First, we verified the origin of the mtDNA in the cybrids by taking advantage of a known polymorphism present in mtDNA of cell line A549 (see Materials and Methods). PCR amplification of genomic DNA from both cybrids followed by incubation with the restriction enzyme AluI yielded a single fragment of ~164 bp, identical in size to the DNA fragment from the patient's myoblasts, whereas a DNA fragment amplified from A549 cells was ~142 bp (Fig. 7 A). This result confirms that the mtDNA in the cybrids was derived exclusively from the patient's cells and was not due to reversion of the [rho]0A549.B2 cell line to its original genotype.
We have characterized some of the molecular events in myoblast cultures from an infant with mtDNA depletion syndrome. The early primary myoblast culture displayed apparently normal levels of mtDNA despite the fact that the skeletal muscle biopsy had a 73% reduction of mtDNA levels compared with age-matched controls. However, later passages of the primary myoblast culture and clonal myoblast cell cultures did show a progressive mtDNA depletion. A similar progressive loss was observed in the patient's skeletal muscle tissue, with post-mortem muscle taken 5 months after the initial biopsy showing a 91% reduction of mtDNA levels. It is possible that the early primary myoblast culture did not show the 73% depletion observed in the muscle biopsy because these cells are usually dormant and are, thus, more likely to reflect the situation in the myoblast precursor population during embryogenesis (17 ). It has been suggested that mtDNA depletion may be caused by an error in the resumption of mtDNA replication after the replication arrest during oogenesis, fertilization and early embryological development, causing reduced amounts of mtDNA in individual stem cell populations (4 ). This is apparently not the case in our patient since the primary myoblast culture had normal levels of mtDNA and half of the myoblast clones contained normal levels of mtDNA by the time they were first analysed after cloning. To our knowledge, the clonal myoblast cell cultures reported here are the first that have been shown to mimic the events in the host patient and as such the in vitro cultures provide a useful model of mtDNA depletion syndrome.
In the clonal myoblast cell cultures, steady-state levels of mRNAs for mitochondrially encoded subunits of cytochrome c oxidase decreased with depleting levels of mtDNA, but this decline was delayed relative to the decline of mtDNA (Fig. 5 ). This indicates an increase in transcription of the residual mtDNA or an increase in transcript stability, and strongly suggests that the mitochondrial transcription machinery is not impaired in the cells. The stable steady-state levels of mRNA for the nuclear-encoded subunit IV of cytochrome c oxidase (Fig. 3 ) are in line with results in mouse cells rendered [rho]0 by treatment with ethidium bromide, in which mRNA levels for subunit VIc of cytochrome c oxidase were unaffected (18 ), and illustrates the absence of transcriptional coordination of nuclear and mitochondrial cytochrome c oxidase genes in these situations.
In contrast to mitochondrial transcript levels, steady-state levels of mtDNA-encoded subunits of cytochrome c oxidase decreased prior to the decrease in mtDNA levels in the myoblast clones and were already reduced to 40% of normal levels when mtDNA levels were only decreased to 74% and mitochondrial mRNA levels to 83% of normal levels (Fig. 5 ; clone 6, passage 10). Studies of cybrid cell lines with increasing levels of a heteroplasmic deletion of the mtDNA have shown that a defect in cytochrome c oxidase expression is only detectable when the proportion of normal mtDNA drops below 40% of total mtDNA (19 ). The difference in threshold levels may reflect the effects of mtDNA deletions and depletion on mitochondrial rRNA expression. In mitochondria, the number of rRNAs, and hence the number of ribosomes, appears to be a limiting factor for translation (20 ). Whereas heteroplasmic mtDNA deletions do not usually affect the mitochondrial rRNA genes, the copy-number of mitochondrial rRNA genes will decrease with depleting mtDNA levels, and might affect translation and, thus, account for the disproportionally low steady-state levels of mitochondrially encoded cytochrome c oxidase subunits.
Replication of mammalian mtDNA is initiated at the origin of the heavy (H)-strand (OH) which is located within the displacement (D)-loop of the mitochondrial genome. Replication priming for OH is an event dependent upon transcriptional promoter function. There are no known differences between the initiation of transcription at this promoter and the initiation of RNA primer synthesis for replication (21 ). Transcription of mtDNA appears to be unaffected both in the patient described here and in a patient described earlier by our group in which we demonstrated that remnant mtDNA was both transcribed and translated (22 ). This suggests that the depletion is not caused by a factor involved in RNA primer formation. Although the basic mechanism of mammalian mtDNA replication has been elucidated, knowledge of the regulatory factors is still limited and only a few protein factors have been purified (21 ). Furthermore, knowledge concerning maintenance of mammalian mtDNA in the cell is also scarce, making it difficult to suggest candidate genes that could be defective.
Even in related patients with mtDNA depletion syndrome, different tissues may be involved (4 ). Our patient showed depletion in liver and skeletal muscle, whereas, in the patient's elder sister, mtDNA levels were decreased in liver but not in skeletal muscle (Morris et al., unpublished observations). Despite cloning of the primary myoblast culture of the patient, cells were still heterogeneous with respect to the expression of mtDNA-encoded polypeptides (Fig. 6 ), suggesting variable levels of mtDNA in the cells. These findings are corroborated by other investigators who demonstrated heterogeneity for mitochondrially encoded polypeptides among individual muscle fibres of patients with mtDNA depletion syndrome (5 ). In all 16 clonal myoblast cell cultures that were established from our patient, mtDNA levels gradually dropped with increasing cell passages but the starting point of the depletion varied. While these observations may suggest a drift of normal and replication-deficient mtDNA populations, the transfer of the patient's mitochondria to [rho]0 control cells suggested a nuclear control of the depletion. However, these results with the cybrids do not fully exclude a mtDNA mutation. For instance, a mutation in a cis-acting replication element on the mtDNA leading to a decrease in affinity for its cognate nuclear-encoded factor may compromise mtDNA replication if the trans-acting factor is present at low intramitochondrial concentrations (below the Kd for the mutated element). The trans-acting factor may be present at higher levels (still above the Kd for the mutated element) in other cell types, like A549, allowing near normal replication of mtDNA in these cells. If this mtDNA mutation is heteroplasmic, drift of normal and mutated mtDNA could explain the apparent tissue and developmental specificity of the mtDNA depletion and the remarkable heterogeneity of the myoblast clones, but would still not explain why all clonal myoblast cultures lose their mtDNA eventually.
Because a mtDNA mutation cannot fully explain our observations and because transfer of patient mitochondria to [rho]0 A549.B2 cells led to restoration of mtDNA levels, we favour the possibility of a nuclear involvement in the depletion of mtDNA in our patient. The depletion could be the result of a nuclear-mitochondrial incompatibility, due partly to a paternal nuclear polymorphism of a factor involved in mtDNA replication and partly to a maternal mtDNA polymorphism which come together in the offspring. Although amplification and sequencing of the origins of mtDNA replication from several patients did not reveal any abnormalities except for assumed neutral polymorphisms (4 ,12 ), these analyses do not exclude the possibility that these polymorphisms or other mtDNA mutations elsewhere could interact with a specific nuclear genotype.
Alternatively, we can speculate that no (heteroplasmic) mutation is present in mtDNA of the patient but that mtDNA replication is arrested due to a nuclear-encoded trans-acting factor. The tissue-specific and neonatal expression of symptoms suggest that this trans-acting factor is tissue specifically and developmentally regulated, and is possibly only expressed after a certain number of cell divisions. Hence, this factor may be involved in the fine tuning of mtDNA levels to the demands of the cell, and may not be one of the core enzymes of mtDNA replication, such as DNA polymerase [gamma] or mitochondrial single-stranded-binding protein (21 ). For example, a nuclear mutation may increase the affinity of the trans-acting factor for mtDNA, thereby inhibiting subsequent dissociation and, thus, blocking further steps of the mtDNA replication process. The development of the cellular model described in the present work opens the way for complementation studies with human cDNA expression libraries to define factor(s) underlying mtDNA depletion syndrome.
Clinical details of the patient will be described in detail elsewhere. In brief, he was the third child of healthy, unrelated parents and was born after an uncomplicated pregnancy. He presented at 8 weeks with hypotonia, poor visual fixation and variable lactic acidaemia. Liver biopsy was normal apart from patchy cytochrome c oxidase staining, whilst muscle biopsy showed only a mild excess of lipid droplets. He became more hypotonic and unresponsive, and died of progressive liver failure at the age of 7.5 months. Post-mortem examination of the liver showed destruction of normal architecture with collapse and fibrosis producing a micronodular cirrhosis. Occasional regenerating nodules and regenerating acinar areas were present which showed cytochrome c oxidase activity, in contrast to the cholestatic areas which were lacking cytochrome c oxidase activity. Post-mortem histochemical examination of skeletal muscle revealed no increase in fat or cytochrome c oxidase deficiency. The eldest sister of the patient is clinically normal, but two other sisters died of a similar illness at 5 and 3 months of age.
A muscle biopsy of the patient was obtained at 2 months of age. Control muscle biopsies were acquired from children aged <2 years undergoing orthopaedic surgery and from infants dying of acute asphyxia. All muscle biopsies were obtained from quadriceps femoris after informed parental consent. Primary myoblast cultures were established as described (23 ) and clonal myoblast cell cultures were obtained by plating a mass myoblast culture at low density followed by ring cloning. In some experiments, the human myoblast control cell strain Hs51Mu (European Collection of Animal Cell Cultures, Salisbury, UK) was employed. The cell line [rho]0A549.B2 (a kind gift of Dr I.J. Holt) is a derivative of the human lung carcinoma cell line A549, and was depleted of mtDNA by long-term exposure to ethidium bromide (24 ). The [rho]0 (mtDNA-less) genotype was confirmed by PCR amplification. Cultures were checked periodically for exclusion of mycoplasma infection.
Cells were cultured at 37oC in a humidified atmosphere of 8% CO2 in air. Myoblasts were grown in Dulbecco's modified Eagle's medium that included 25 mM glucose and 4 mM l-glutamine, and was supplemented with 20% batch-tested fetal bovine serum, 2% detoxified chick embryo extract, 2 [mu]M uridine, 1 [mu]M sodium pyruvate, 50 U/ml of penicillin and 50 [mu]g/ml of streptomycin. Cell line [rho]0A549.B2 was maintained in the above medium lacking chick embryo extract and supplemented with 10% instead of 20% fetal bovine serum. Selective growth medium was used where indicated and was the same as for [rho]0A549.B2 cells but lacking uridine and sodium pyruvate, and 10% dialysed fetal bovine serum was used.
For molecular biological analyses, ~5*106 myoblasts were harvested by trypsinization and washed once with phosphate-buffered saline (PBS). The cell pellet was resuspended in PBS on ice and divided between three Eppendorf tubes (1.5*106, 3*106 and 0.5*106 cells, respectively). Cell suspensions were centrifuged at 3000 g for 3 min and cell pellets were used for the extraction of DNA (1.5*106 cells), RNA (3*106 cells) and protein (0.5*106 cells).
Total genomic DNA was extracted with the NucleonT I kit (Scotlab, Strathclyde, UK). Similar amounts (2-4 [mu]g) of DNA were digested with PvuII, electrophoresed through 0.8% agarose gels and blotted onto HybondT-N membrane (Amersham, Buckinghamshire, UK) as recommended by the supplier. Blots were hybridized with two probes simultaneously: a cloned 5.8 kb EcoRI fragment of the gene encoding 18S rRNA [courtesy of Dr I.J. Holt; (25 )] and the entire mtDNA isolated from human placenta (26 ). Probes were labelled with fluorescein using the ECLT Random Prime Labelling System (Amersham). Blots were pre-hybridized, hybridized and developed according to the Fluorescein Gene ImagesT detection system (Amersham).
For transcript analysis, total cellular RNA was isolated from cell pellets with an adaptation of the method described by Chomczynski and Sacchi (27 ). To remove contaminating DNA, RNA was precipitated selectively in 0.8 M LiCl twice and recovered by ethanol precipitation (28 ). The integrity of the RNA preparations was verified visually by electrophoresis through an agarose gel containing ethidium bromide. For Northern blotting, RNA was denatured with glyoxal and dimethyl sulphoxide and samples (3 [mu]g) were subject to 1.4% agarose gel electrophoresis in 10 mM sodium phosphate buffer (pH 6.8), and then transferred to HybondT-N filters in 3 M NaCl, 0.3 M sodium citrate (pH 7.0) and fixed with UV (28 ). Blots subsequently were hybridized with a cloned cDNA fragment encoding cytochrome c oxidase subunit IV [courtesy of Dr M.I. Lomax; (29 )], a cloned XbaI fragment of mtDNA containing the complete gene of cytochrome c oxidase subunit II and 74 bp of subunit I (30 ), and the 18S rRNA probe (see above). Probes were labelled with [[alpha]-32P]dCTP using the Rediprime DNA Labelling System (Amersham). Blots were pre-hybridized and hybridized with the probes in buffers recommended by the supplier containing 50% formaldehyde. Hybridizations were performed at 50oC for the cytochrome c oxidase subunit IV probe and at 60oC for the other probes. Filters were exposed to Kodak BioMax MS film.
Cellular protein, including mitochondrial membrane proteins, was extracted from cell pellets in 200 [mu]l of PBS containing 1.5% n-dodecyl-[beta]-d-maltoside and a cocktail of protease inhibitors (31 ). Protein contents of solubilized fractions were determined with the bicinchoninic acid kit (Pierce, Rockford, IL). Similar amounts of protein (~8 [mu]g) were dissociated for 20 min at 37oC in the presence of 4% SDS, 2% [beta]-mercaptoethanol, and subsequently resolved on 15% polyacrylamide, 5.5 M urea mini-gels and electro-blotted onto ImmobilonT-P poly(vinylidene difluoride) membranes (Millipore, Bedford, MA) as described (31 ). Blots were developed with monoclonal antibodies against cytochrome c oxidase subunits [Molecular Probes, Eugene, OR; (32 )] and a monoclonal against porin (31HL; Calbiochem-Novabiochem, Nottingham, UK) according to Capaldi et al. (31 ), except that immunoreactive material was visualized by chemiluminescence (RenaissanceT, DuPont NEN, Boston, MA) after incubation with goat anti-mouse IgG horseradish peroxidase conjugate.
Exposures of blots to films were chosen such that the signals were always within the linear range. Signals were measured by densitometry and corrected for background noise. In order to correct for possible quantitative variations among the samples, the mtDNA signals were normalized relative to the 18S rRNA gene signals, the mRNA signals to the 18S rRNA signals, and the signals of cytochrome c oxidase subunits to the signal of porin. To be able to evaluate the signals on the blots, serial dilutions with control material were run on parallel gels, blotted and developed simultaneously with the blots with patient material, and signals of the patient material were quantified by comparison with those of the dilution series. All experiments shown were analysed on two independent blots; duplicate values did not differ more than 5% from each other.
Myoblasts, seeded on glass coverslips, were cultured for 45 min in medium containing 2 [mu]M MitoTrackerT CMXRos-H2 (Molecular Probes) followed by culturing in plain medium for 30 min. Cells on coverslips were washed subsequently in PBS, fixed with 4% paraformaldehyde in PBS for 20 min, washed, permeabilized for 15 min in methanol at -20oC and washed again. Protein binding sites were saturated for 30 min with 10% normal goat serum in PBS at 37oC in a humidified atmosphere, followed by incubation for 45 min with monoclonal antibodies against cytochrome c oxidase subunits [Molecular Probes; (32 )] and 10 [mu]g of bis-benzimide per ml of PBS at 37oC (humidified). Finally, coverslips were washed and incubated for 45 min with goat anti-mouse IgG-FITC in PBS at 37oC (humidified), before being washed once more and mounted onto glass slides in Citifluor-glycerol-PBS solution. Fluorescence was inspected with a Zeiss Axiophot photomicroscope equipped with a 40* Plan-Neofluar lens. Photographs were taken on Kodak EktachromeT Panther P1600 film.
Cytochemical staining for cytochrome c oxidase activity with 3,3' diaminobenzidine (DAB) was adapted from a technique described previously (33 ). For this purpose, cells, seeded on coverslips, were rinsed with PBS and briefly air-dried before the incubation with the DAB-cytochrome c-catalase mixture.
The transfer of patient mitochondria to cell line [rho]0A549.B2 was performed essentially as described (9 ). To confirm the origin of the mtDNA in the cybrids, we took advantage of a known DNA polymorphism present in mtDNA of the parental A549 cell line (Dr I.J. Holt, unpublished observations). The CA repeat at position 513-523 of the mtDNA Cambridge sequence (34 ) is one CA dinucleotide shorter in A549 mtDNA. We performed restriction fragment length polymorphism (RFLP) analysis for this micro-deletion after PCR amplification (28 ) using as forward primer an oligodeoxynucleotide encompassing the Cambridge sequence 371-390 and as reverse complementary primer 5'-GGTTAGCAGCGGTGTGTG
Monoclonal antibodies against cytochrome c oxidase were developed in the laboratory of Dr R.A. Capaldi. This research was supported by the Medical Research Council.
Human Molecular Genetics
Pages
Introduction
Results
Steady-state levels of mtDNA, and cytochrome c oxidase mRNAs and subunits
Cytochrome c oxidase immunocytochemistry
Nuclear complementation experiments
Discussion
Materials And Methods
Clinical history
Tissues, cell cultures and media
Southern, Northern and Western blot analyses
Cytochemical staining
Enucleation, cell fusion and RFLP analysis
Acknowledgements
References
REFERENCES
This page is maintained by OUP admin. Last updated Mon May 12 18:10:04 BST 1997. Part of the OUP Journals World Wide Web service. Copyright Oxford University Press, 1996