Heteroplasmic mutations of the mitochondrial genome cause paradoxical effects on mitochondrial functions

Abstract

Mitochondrial genome (mtDNA) mutation causes highly variable clinical features, and its pathogenesis is not fully understood. In this study, we analyzed the heteroplasmic mtDNA mutation C4936T (p.T156I) in ND2 of complex I and the homoplasmic mtDNA mutation A9181G (p.S219G) in ATPase 6 of complex V. Using cybrid technology, we found that in a high-glucose medium in which cultured cells mainly depend on anaerobic glycolysis for energy, the C4936T mutation inhibited cell growth by 50%. Oxygen consumption and reactive oxygen species production were also reduced by 60 and 75%, respectively. Because the subject also had conjunctiva carcinoma, we further tested whether the C4936T mutation was associated with tumor formation. In an anchorage-dependant growth test, we found that only cells with a high level of C4936T mutation formed colonies. In contrast, when the cells grew in a galactose medium in which cells were forced to generate ATP through oxidative phosphorylation, the C4936T mutation protected cells from apoptosis probably caused by the A9181G mutation. Our results suggest that the phenotype caused by mtDNA mutations may depend on the availability of the nutrients. This gene-environment interaction may contribute to the complexity of pathogenesis and clinical phenotypes caused by mtDNA mutation.—Zhang, C., Huang, V. H., Simon, M., Sharma, L. K., Fan, W., Haas, R., Wallace, D. C., Bai, Y., Huang, T. Heteroplasmic mutations of the mitochondrial genome cause paradoxical effects on mitochondrial functions.

Mitochondria are thought to be ancient bacterial symbionts that retain their own DNA-, RNA-, and protein-synthesizing apparatus. Each human cell contains hundreds of mitochondria and thousands of mitochondrial DNA (mtDNA) copies (1, 2). Nutrients in our diet are burned within the mitochondria with the oxygen that we breathe to generate energy in the form of ATP (3). As a by-product of energy production, the mitochondria also generate reactive oxygen species (ROS), which can damage mtDNA, as well as genomic DNA that resides in the nucleus (4, 5). In past decades, an increasing number of human disorders, including myopathy, cardiomyopathy, and neurological and endocrine disorders, as well as the aging process, have been attributed to mutations in the maternally inherited mtDNAs that may cause dysfunction of the electron transport chain (ETC) and ATP synthesis (2, 6, 7). A striking feature of mtDNA disorders is their clinical heterogeneity, ranging from single-organ involvement to severe multisystem diseases. The same mtDNA mutation or different mutations in the same mtDNA gene may result in very different clinical manifestations, whereas the same clinical phenotype may be caused by different mutations (2, 8). mtDNA mutations that cause severe diseases in humans are most often heteroplasmic (i.e., mutant and wild-type molecules coexist within cells), and the proportion of mutant molecules can vary between tissues and with age (8, 9). It was reported that the threshold for mutant mtDNAs to cause biochemical dysfunction may be ∼60% for deletions (8, 10) and up to 95% for tRNA mutations (8, 11). However, the factors modulating the pathogenic threshold are still unknown. The transmitochondrial cybrid model, introduced in 1989 by King and Attardi (12, 13), enabled us to study cellular consequences of mtDNA mutations in an identical nuclear background.

In this study, we used the cybrid model to investigate the functional effect of a rare heteroplasmic C to T mutation at nucleotide 4936 in the mitochondrially encoded NADH dehydrogenase 2 (MT-ND2) gene of a patient with typical neurological and autonomic mitochondrial disease. We also analyzed the underlying molecular mechanisms by studying oxygen consumption, ROS production, and apoptosis. Our data suggested that this heteroplasmic mtDNA mutation might play a paradoxical role in mitochondrial function, depending on the availability of nutrients.

MATERIALS AND METHODS

Clinical information

The proband (SP) is a 43-yr-old Caucasian man evaluated at the University of California–Irvine Medical Center. He had a very complicated clinical course and was diagnosed with mitochondrial disease. He was born to G1P1, a 27-yr-old woman. He reached all developmental milestones appropriately and was completely healthy until the age of 34, when he developed symptoms of coughing, conjunctivitis, night sweating, and photophobia after a trip to Bali. In a subsequent workup for infectious disease, he was found to be Epstein Barr virus-IgM-positive with low CD4⁺ cells. Results of the infectious disease workup were negative.

In the subsequent 2–3 yr, he developed systemic abnormalities in multiple organs. He complained of muscular contractions, fasciculation, soreness, and weakness after activity and exercise. Sleep apnea unresponsive to continuous positive airway pressure was diagnosed and attributed to upper airway obstruction. His magnetic resonance imaging scan showed that he had muscle atrophy and fat infiltration. The laboratory workup showed persistently elevated lactate, elevated creatinine kinase, and decreased plasma carnitine levels. Muscle biopsy showed a complex I deficiency without ragged red fibers. To evaluate the mitochondrial disease, his whole mtDNA was sequenced, and multiple mutations were found, including a 30–40% heteroplasmic 4936C>T mutation (T156I) in ND2 of complex I; two homoplasmic Leber's hereditary optic neuropathy (LHON) secondary mutations (4216T>C and 4917A->G) in ND1 of complex I; and a homoplasmic 9181A>G (p.S219G) mutation in ATPase 6 in complex V. Another genetic workup identified no mutation in the PYGM and CPT2 genes, but a mutation in the AMPD1 gene, which caused two variants (p.Q12X p.P48L). These two mutations were both inherited from his father. Because these two mutations are cosegregated, their contributions to clinical symptoms are considered to be less significant (14).

Another neurological finding was autonomic dysfunction. The patient was found to have a significantly positive tilt test, hypokinetic circulatory status, increased heart rate during a cold pressor procedure, and unexplained lack of sweating on the dorsal side of the left foot. He was found to have increased gastrointestinal (GI) motility, which seemed to be modified by diet. Another workup showed a large spleen on a magnetic resonance imaging scan. His metabolic workup showed high triglycerides and high cholesterol. It seemed that he responded well to statins. He had a persistent decrease in vitamin D levels, which may be associated with his osteopenia. He was treated with large doses of vitamin D. He also had acidosis at only 50% of creatine depletion with exercise. His creatine depletion rate with exercise was significantly faster than that of normal healthy controls. He was diagnosed with a squamous cell carcinoma of the left eye conjunctiva. This is a very rare cancer, and whether its etiology was associated with his mitochondria dysfunction was not known.

The mother is affected. The clinical phenotypes are very similar. Besides muscular dysfunction, the mother is bedridden and is unable to walk, to dress herself, or to perform most daily activities. She has progressive multiple sclerosis. Many symptoms overlap between SP and his mother. Both of them have peripheral neuropathy, a history of muscular fasciculation, and increased GI mobility. She carries the same mutation in the mitochondrial genome.

The family history is shown in the pedigree. The proband has a 41-yr-old brother with a history of hypothyroidism, joint problems, GI hypermobility, and chest pain. His 36-yr-old sister has a history of depression. There is a strong family history for cancer. There is no consanguinity. The family is Ashkenazi Jewish. The medication history included coenzyme Q₁₀ and vitamin supplements, including high-dose B vitamins and vitamin D.

At the time of the physical examination SP was 43 yr old, and his height and weight were in the normal range. He had no dysmorphic features. His cranial nerves II–XII were grossly intact. He had significant fasciculation and hypertrophy of the calf muscle and first and second interrosseous interspace atrophy in both hands. However, he had normal reflexes. The rest of the physical examination was unremarkable.

Generation of cybrid cells

Skin fibroblast cells were established, and a cybrid cell line was generated. Fibroblast cells were also used to generate induced pluripotent stem cells. Transmitochondrial cybrid cells were generated as described previously (12, 15). In brief, enucleated fibroblasts carrying a heteroplasmic 4936 C>T alteration in the MT-ND2 gene were fused with the ρ⁰206 cell line, a derivative of 143B.TK⁻ cells, by polyethylene glycol-DMSO solution (Sigma-Aldrich, St. Louis, MO, USA). Transformants were isolated in selection medium comprising DMEM (without sodium pyruvate) supplemented with 5% dialyzed FBS (Invitrogen, Carlsbad, CA, USA) and 50 μg/ml of bromodeoxyuridine (BrdU; Sigma-Aldrich). After 4–6 wk of selection, BrdU was omitted from the medium.

mtDNA analysis

Total DNA was extracted from cells after overnight digestion at 55°C in 100 mM Tris-Cl (pH 8.0), 5 mM EDTA, 0.2% SDS, 200 mM NaCl, and 10 μg/ml proteinase K. DNA (50 ng) was used to amplify regions of interest using the following primers: hmt4810F, 5′-GAGTCCCAGAGGTTACCCAAG-3′; hmt5046R, 5′-CTGCTATTATTCATCCTATGTGGG-3′; hmt9070F, 5′-TCAACCATTAACCTTCCCTCT-3′; and hmt9300R, 5-CTAGGCCGGAGGTCATTAGG-3′. The PCR products were purified using Qiagen PCR purification columns (Qiagen, Valencia, CA, USA) and subsequently sequenced using both forward and reverse primers with BigDye Terminator Cycle Sequencing Kits (Applied Biosystems, Carlsbad, CA, USA). DNA sequence chromatography was verified manually using DNAStar Lasergene 8 (DNAStar, Madison, WI, USA).

Cell growth competition assay

To compare the growth rate of cybrid cells, a competition assay was performed. We equally mixed two cell lines with a known degree of mutation, plated a total of 100,000 cells in one 6-well plate, and grew them in either glucose or galactose medium. The cells were passaged several times when they reached confluence. Each time the cells were passaged, a portion was collected for mtDNA analysis as indicated above. The heteroplasmy of the mtDNA was determined by chromatogram.

DNA content and cell cycle kinetic analysis

Cells were harvested with 0.25% trypsin-EDTA (Invitrogen) and resuspended in single-cell suspension in PBS + 0.1% BSA at 1–2 × 10⁶ cells/ml in a 15-ml polypropylene tube. Cold absolute ethanol (3 ml) was added dropwise while vortexing to prevent clumping and cell loss. The cells were fixed for at least 1 h at 4°C or stored in 70% ethanol at −20°C. Before nuclear staining and flow cytometric analysis, the cells were washed in PBS twice. One milliliter of propidium iodide (PI) staining solution (3.8 mM sodium citrate, 50 μg/ml PI, and 10 μg/ml RNase A) was used to stain the nuclei. The cells were analyzed immediately using a BD FACScan flow cytometer (Becton Dickinson, San Jose, CA, USA) with excitation at 488 nm. Forward and side scatter were used to gate the viable population of cells. A minimum of 20,000 events were collected for each sample. Data were analyzed using FlowJo 7.5 (Tree Star, Inc., Ashland, OR, USA). The Dean-Jett-Fox model was used to calculate the percentage of cells in the G₁, S, and G₂ phases.

Oxygen consumption measurement

The oxygen consumption rate was measured with a Seahorse XF24 Analyzer (Seahorse Biosciences, North Billerica, MA, USA). The experiment was performed using 100,000 cells/well according to the manufacturer's instructions. After the basal respiration was recorded, oligomycin was added to a final concentration of 5 μM to measure the uncoupled respiration, and then carbonyl cyanide 4-trifluoromethyoxyphenylhydrazone (FCCP) was added to a final concentration of 0.25 μM to measure the maximal respiration. Finally, mitochondrial ETC-dependent respiration was inhibited by the addition of rotenone and anti-mycin A at the final concentration of 1 μM.

ROS measurement

The mitochondrial and cytoplasmic levels of ROS were measured using MitoSOX red mitochondrial superoxide indicator and carboxyl-2′-7′-dichlorofluorescein-diacetate (H₂DCF-DA; Invitrogen), respectively, as described previously (16). For each analysis, 1 × 10⁶ cells were plated in 6-well culture plates. After addition of MitoSOX (2.5 μM) and carboxyl-H₂DCF-DA (20 μM), cells were incubated for 30 min at 37°C in the dark. The cells were then washed with PBS, harvested by trypsinization, resuspended in HBSS (Invitrogen), and analyzed immediately using a FACScan flow cytometer with excitation at 488 nm. Forward and side scatter were used to gate the viable population of cells. Carboxyl-H₂DCF-DA emits at 530 nm (FL-1 channel), whereas MitoSOX emits at 590 nm (FL-2 channel). A minimum of 20,000 events were collected for each sample. Data were analyzed using WinMDI version 2.9 (Scripps Research Institute, San Diego, CA USA; http://facs.scripps.edu/software.html).

Cell viability assays

Cybrid cells maintained in high-glucose (25 mM) medium were rinsed with PBS, harvested with 0.25% trypsin, and plated at a density of 5000 cells/cm² in 12-well plates using high-glucose medium. Live cells were counted by trypan blue exclusion at 24, 48, and 72 h after plating. For viability testing in galactose medium, the cybrid cells maintained in high-glucose (25 mM) medium were rinsed with PBS, harvested with 0.25% trypsin, and then plated at the density of 15,000 cells/cm² in 12-well plates in galactose medium (DMEM without glucose and sodium pyruvate, supplemented with 5 mM galactose, 5% FBS). Live cells were counted by trypan blue exclusion at 24, 48, and 72 h after plating.

Subcellular fractionation

After incubation in DMEM-galactose medium for the indicated time, cells were harvested and resuspended in 0.5 ml of 200 mM mannitol, 70 mM sucrose, 1 mM EGTA, 10 mM Hepes (pH 7.6), and 1× protease inhibitor cocktail (Roche, Indianapolis, IN, USA) and homogenized for 30 strokes with a Dounce homogenizer. This and the subsequent steps were performed at 4°C. Samples were centrifuged for 10 min at 500 g, and the resulting supernatant was centrifuged for 20 min at 10,000 g. The supernatant (cytosolic fraction) was stored at −80°C. The pellet (mitochondria-enriched fraction) was resuspended in 50 μl of PBS containing 1% Triton X-100, 0.5 mM EDTA, and 1× protease inhibitor cocktail and stored at −80°C. Protein concentration was measured using the Bradford assay (Bio-Rad Laboratories, Hercules, CA, USA).

Western blot analysis

Protein samples (25 μg) were separated on NuPAGE Novex 4–12% Bis-Tris polyacrylamide gel (Invitrogen) and blotted onto PVDF membranes after electrophoresis (Invitrogen). The membrane was blocked with 5% nonfat dry milk in PBS with 0.1% Tween 20 for 1 h at room temperature and incubated with primary antibodies against cytochrome c (1:500, BD Biosciences, San Jose, CA, USA), apoptosis-inducing factor (AIF; 1:100; Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA), and superoxide dismutase 2 (SOD2; 1:1000; MitoSciences, Eugene, OR, USA) at 4°C overnight. After several rinses in PBS, the membrane was incubated with peroxidase-conjugated goat secondary antibodies (1:10,000; Invitrogen) for 2 h at room temperature, and the signals were detected with the ECL Plus kit (GE Healthcare, Chalfont St. Giles, UK).

Soft agar assay

A total of 20,000 cells were seeded in 60-mm dishes (4 dishes/line), previously prepared with a lower layer of 0.4% solid agar in DMEM with 5% FBS and overlaid with a 0.27% soft agar in the same medium. Cells were allowed to grow into colonies. The overlaid medium was changed every third day. Images were taken after 2–3 wk using an Alpha Innotech instrument (Alpha Innotech Corp., San Leandro, CA, USA), and colonies were scored by AlphaEaseFC software (Alpha Innotech). Colonies showing a minimum pixel area of >2 were scored positive (16).

Apoptosis assay

Whole DNA was extracted from cultured cells, and DNA fragmentation was examined using 1% agarose gel electrophoresis. For in situ nuclear staining, cybrid cells maintained in high-glucose (25 mM) medium were rinsed with PBS, harvested with 0.25% trypsin, and plated at the density of 15,000 cells/cm² in 4-well chamber slides (Nalge Nunc International, Rochester, NY, USA) coated with poly-d-lysine (BD Biosciences) the day before. After 24 h, the cells were rinsed in PBS and incubated with 10 μg/ml Hoechst 33342 (Invitrogen) and 25 μg/ml PI at room temperature for 30 min to stain the nuclei. The cells were then fixed in 4% paraformaldehyde and imaged with a Zeiss fluorescence microscope (Carl Zeiss, Oberkochen, Germany).

RESULTS

Our subjects exhibited multiple neurological and autonomic problems that are typical for mitochondrial diseases, including elevated baseline lactate levels in blood and urine, a persistently elevated creatine kinase level, muscle fasciculations of the calves, numbness, paresthesias, and myoclonic jerks of the arms. Full mtDNA sequencing revealed a number of alterations in the patient's mtDNA, including a heteroplasmic change from cytosine to thymine at nucleotide 4936 in the MT-ND2 gene, two homoplasmic LHON ancillary mutations (4216 T>C and 4917 A>G), a change from adenine to guanine at nucleotide 9181 in the ATPase 6 gene, and 31 additional polymorphisms (Table 1).

Table 1.

Alterations in the patient's mtDNA

Nucleotide no.	Nucleotide change	Amino acid change	Gene	Population frequency (%)	Conservation level	Comments
41	C>T	—	HV	0.1	Not conserved	Pm*
73	A>G	—	HV2	83.4		Pm*; control region
150	C>T	—	HV2, OH	13.2		Pm*; control region
263	A>G	—	HV2	99.7	Not conserved	Pm*; most common nucleotide is G; A is a rare polymorphism in the reference sequence
303	insC	—	OH, CSB2		Not conserved	Pm*
311	insC	—	OH, CSB2		Not conserved	Pm*; insC most common form; a deletion is a rare polymorphism in reference sequence
709	G>A	—	RNR1	16.4	P1 (−Ndr)	Pm*
750	A>G	—	RNR1	99.2	G–P1	Pm*; most common nucleotide is G; A is a rare polymorphism in the reference sequence
1438	A>G	—	RNR1	96.9	G–P1	Pm*; most common nucleotide is G; A is a rare polymorphism in the reference sequence
1888	G>A	—	RNR2	5.3	P1	Pm*
2706	A>G	—	RNR2	80.5	G–P1 (−Ndr)	Pm*; most common nucleotide is G
4216	T>C	Y304H	ND1	9	P1	Pm*; missense; also LHON secondary
4769	A>G	M100M	ND2	98.9	Silent	Pm*; most common nucleotide is G; A is a rare polymorphism in the reference sequence
4936	C>T	T156I	ND2	0.1	P1	Heteroplasmic; missense
4917	A>G	N150D	ND2	4.8	P1	Pm*; missense; also LHON secondary
7028	C>T	A375A	COI	81.3	Silent	Pm*; most common nucleotide is T
8697	G>A	M57M	ATP6	4.7	Silent	Pm*
8860	A>G	T112A	ATP6	99.8	Ala–P1	Ala is most common form; Thr is a rare polymorphism in the reference sequence (rCRS)
9181	A>G	S219G	ATP6	0.04	P3	Highly conserved: likely to be deleterious
10463	T>C	—	MTTR-Arg	4.7	P2	Pm*
11251	A>G	L164L	ND4	8.7	Silent	Pm*
11719	G>A	G320G	ND4	77.7	Silent	Pm*; most common nucleotide is A
11812	A>G	L351L	ND4	3.3	Silent	Pm*
13368	G>A	G344G	ND5	4.9	Silent	Pm*
14233	A>G/T>C	D147D	ND6	3.4	Silent	Pm*
14766	C>T	T7I	CYTB	77.4	Ile–P1	Pm*; Ile most common form
14905	G>A	M53M	CYTB	5.1	Silent	Pm*
15326	A>G	T194A	CYTB	99.4	Ala–P1	Ala is most common form; Thr is rare polymorphism in the reference sequence (rCRS)
15452	C>A	L236I	CYTB	8.7	P2	Pm*; missense
15607	A>G	K287K	CYTB	5.5	Silent	Pm*
15928	G>A	—	MTTR-Thr	4.9	P1	Pm*
16126	T>C	—	HV1	8.9		Pm*; control region; fast evolving site
16153	G>A	—	HV1	0.6		Pm*; control region; fast-evolving site
16294	C>T	—	HV1	5.7		Pm*; control region; fast-evolving site
16519	T>C	—	D loop	59.7	Not conserved	Pm*

The patient's mitochondrial DNA belongs to the haplogroup T, subclass 2e (T2e). Variants are expressed relative to the revised Cambridge Sequence (rCS). P1: conserved in humans with the oldest mitochondrial haplogroups (L0 Uganda, L1, L2, L3, L3 Yoruba) + Neanderthal. P2: P1 + conserved in apes and old world monkeys (chimp, orangutan, gorilla, gibbon, rhesus). P3: P2 + conserved in 12 nonsimian mammals (tarsier, cow, dog, cat, rabbit, guinea pig, mouse, elephant, horse, armadillo, opossum, platypus). P4: P3 + conserved in 5 nonmammalian vertebrates [chicken, lizard, frog (Xenopus), and two species of fish (medaka, stickleback)].

Pm*, polymorphism.

The A9181G ATPase 6 mutation (p.S219G) per se is most likely pathogenic. First, it is very rare in the normal population. In the MITOMAP database (http://www.mitomap.org) with 2701 full mitochondria genome sequences, only one of those sequences carried this variant. In addition, this position is highly conserved. In our clinical database for individuals who were suspected to have mitochondria disease, we found that several individuals carry this mutation. The clinical phenotypes can be very diverse, including developmental delay, encephalopathy, seizure disorder, delayed gastric empty, lactate acidosis, failure to thrive, and seizure disorder. Some of these phenotypes overlap with that of our subject. The 4936 C>T alteration in the MT-ND2 gene results in an amino acid substitution at position 156 from threonine to isoleucine that is also highly conserved through the great apes. We hypothesize that this alteration is possibly deleterious or modifies the clinical phenotypes because it is heteroplasmic in the patient, and thus we focused our studies on this mutation. Because the mitochondrial genome replicates and transmits independently of the nuclear genome, it is challenging to study the pathogenesis of mtDNA mutations. Cybrid technology can be used to study the pathogenesis of mitochondrial genome mutations. This technology is particularly useful to test pathogenesis for heteroplasmic mutations of the mitochondrial genome. We first established several cybrid cell lines containing heteroplasmic mutations at this position by transferring mitochondria from the patient's fibroblasts into the mtDNA-deficient ρ⁰206 cells. We chose the cybrid cell lines with 0% mutant (CI-6), 50% mutant (CI-3), and 100% mutant (CI-30) mtDNA as representatives for wild-type, heteroplasmic, and homoplasmic variations, respectively, for further analysis.

Cell growth studies

We first analyzed the growth rate of these cybrid cells and found that cells with 100% altered mtDNA (CI-30) grew slower than cells with 50% heteroplasmic (CI-3) or 0% mutant mtDNA (CI-6) in regular medium with high glucose content (Fig. 1A). We then formulated a competition assay to compare the growth rate between cells with 100 and 0% mutant mtDNA by plating equal numbers of each line in the same culture dish, and analyzed the degree of heteroplasmy of mtDNA in the mixed population at each passage. Interestingly, the 4936 C>T variation load diminished gradually in the mixed population after 6 passages (∼3-4 wk), confirming that CI-30 cells grew slower than CI-6 cells, and had a growth disadvantage during growth competition (Fig. 1B). To test the effect of the 4936 C>T mutation on the cell cycle, we examined DNA content. We found no significant difference in cell cycle progress in any of the cybrid cell lines. All cybrid cells exhibited similar cell cycle kinetics, with ∼55% of cells at G₁ phase, 27% at S phase, and 16% at G₂/M phase (Fig. 1C), suggesting that the 4936 C>T mutation mainly affects cell death. To test whether the mtDNA C4936T mutation increases sensitivity to oxidative stress in glucose culture medium, we performed an apoptosis assay. Sensitivity of cells to oxidative stress was detected by treating cultured cells with 100 mM tert-butyl hydrogen peroxide. We then stained with annexin V. Cells were visualized with a fluorescence confocal microscope. Our results confirmed that the mtDNA C4936T mutation increases sensitivity to oxidative stress and thereby increases apoptosis in glucose culture medium (data not shown).

Figure 1.

A) Cybrid cells with various loads of 4936 C>T variant in the MT-ND2 gene were incubated for the times indicated in regular glucose medium, and the total numbers of live cells were counted as described in Materials and Methods. Each data point is the mean ± se of 3 determinations. B) DNA sequencing chromatographs of mitochondrial DNA in a mixed population of cybrid cells carrying either wild-type (cytosine, blue) or altered (thymine, red) nucleotide at position 4936 (indicated by arrowheads) at different passages. C) DNA content analysis of cybrid cells growing in glucose medium. The Dean-Jett-Fox model was used to calculate the percentage of cells in the G₁, S, and G2 phases.

Cell respiration studies

We also examined the respiration of the CI-6, CI-3, and CI-30 cells and found that the basal oxygen consumption rates of CI-3 and CI-30 cells were ∼40–50% of the level of the CI-6 cells (Fig. 2A). When the chemical uncoupler FCCP was added to completely uncouple the mitochondrial respiration and raise the respiratory capacity to the maximum level, oxygen consumption in the CI-3 and CI-30 cells remained at ∼40–50% of that of the CI-6 cells, suggesting that the NADH dehydrogenase (complex I) in those cells was not as efficient at pumping electrons into the ETC as that in the CI-6 cells. No significant difference was detected with the oxygen consumption rate when the whole ETC was inhibited by rotenone (complex I inhibitor) and anti-mycin A (complex III inhibitor; Fig. 2A).

Figure 2.

A) Oxygen consumption rate of cybrid cells incubated in glucose medium was measured in the resting stage and in the presence of oligomycin, the uncoupler FCCP, and rotenone with anti-mycin A. Each data point is the mean ± se of 4 determinations. B) Levels of ROS at the mitochondrial and intracellular compartments were assessed using the red fluorescence mitochondrial superoxide indicator MitoSOX and carboxyl-H₂DCF-DA, respectively. C) Cells (20,000) were seeded on 60-mm dishes containing 0.4% solid agar. After 3 wk, colonies were stained with p-iodonitrotetrazolium violet. Images were taken using an Alpha Innotech instrument. D) Numbers of colonies were scored with AlphaEaseFC software. Colonies showing a minimum area > 2 pixels were scored positive. Mut, mutant.

ROS production

Because the cybrid cells exhibited marked differences in oxygen consumption rate, we further measured the level of mitochondrial and intracellular ROS using the mitochondrial superoxide indicator, MitoSOX, and the intracellular superoxide indicator, carboxyl-H₂DCF-DA, respectively. Surprisingly, a lower level of mitochondrial superoxide was observed in the CI-30 cells compared with that of CI-3 and CI-6 cells (Fig. 2B), suggesting that the 4936 C>T mutation blocks the inflow of electrons in NADH dehydrogenase (complex I) before reaching ROS production and the electron transport chain. Consistently, the intracellular ROS level in CI-30 cells was also lower than that in the CI-3 and CI-6 cells (Fig. 2B). Taken together, these results indicate that the cells with the homoplasmic 4936 C>T mutation had overall lower levels of intracellular and mitochondrial ROS than cells with heteroplasmic or wild-type mtDNA.

Oncogenesis

To investigate the role of the 4936 C>T mutation in MT-ND2 on tumor formation, an anchorage-dependent growth test was performed. A total of 20,000 cells were seeded in 0.4% soft agar medium and allowed to form colonies for 3 wk. To our surprise, only the homoplasmic CI-30 cells could form colonies in the soft agar assay (Fig. 2C). After 3 wk, ∼150 colonies were formed in the dish seeded with CI-30 cells. However, no colony was found in the dishes seeded with the heteroplasmic CI-3 or wild-type CI-6 cells. This result indicated that the homoplasmic MT-ND2 mutation might have an enhancing effect on cell growth and/or survival under nutrient or hypoxic stress or that CI-30 cells have phenotypic changes such as loss of contact inhibition and anchorage independence, which improve survival.

Galactose culture

A well-established carbon source stress test is to replace glucose with galactose in the culture medium and force the cells to generate ATP through oxidative phosphorylation (17, 18). We plated the cybrid cells in galactose medium. To our surprise, we found that cells with the 4936 C>T mutation, CI-30 and CI-3, were more resistant to carbon stress than the wild-type MT-ND2 cells, CI-6. CI-6 cells started to die within 24 h after being plated in galactose medium; after 72 h, only a few of the CI-6 cells were still alive. The number of CI-3 cells remained stable within 24 h and only a fraction gradually died after 72 h. On the contrary, the CI-30 cells managed to proliferate and survive for at least 48 h and only started to die after prolonged culture (72 h) in medium with galactose as the only carbon source (Fig. 3A).

Figure 3.

A) Cybrid cells were transferred into medium containing galactose in place of glucose for the times indicated, and the total numbers of live cells were counted. Each data point is the mean ± se of 3 replicates. B) DNA sequencing chromatographs of mitochondrial DNA in a mixed population of cybrid cells carrying either wild-type (cytosine, blue) or altered (thymine, red) nucleotide at position 4936 (indicated by arrowheads) at different passages in the galactose medium. C–E) Images of CI-6 (C), CI-3 (D), and CI-30 (E) cybrid cells taken 24 h after culture in glucose-free galactose medium. Clockwise from top left: phase contrast; nuclei staining with the membrane-impermeable PI dye; nuclei staining with Hoechst 33342; and merged images.

Apoptosis

When equal numbers of CI-6 and CI-30 cells were seeded in the same culture dish in galactose medium and subjected to a competition assay, copies of the wild-type mtDNA decreased after 8 d (2 passages), suggesting that these cells had been taken over by mutant (CI-30) cells in the competition (Fig. 3B). Microscopic analysis of the cells 24 h after culturing in the galactose medium showed that most of the CI-6 cells exhibited typical apoptotic or necrotic morphology (Fig. 3C–E). Furthermore, in situ staining of the cells with membrane-impermeable PI revealed that the majority of the CI-6 cells could be stained by PI, suggesting that they had lost their membrane integrity due to apoptosis or necrosis. On the contrary, most of the CI-3 and CI-30 cells looked healthy and only a few of them were stained positive by PI after 24 h in the galactose medium (Fig. 3C–E).

Further analysis by agarose gel electrophoresis revealed evident DNA fragmentation in the CI-6 cells after 24 h of culture in galactose medium (Fig. 4A). No DNA fragmentation was observed in CI-3 and CI-30 cells incubated in the galactose medium. We focused our attention on the release of two mitochondrial apoptogenic factors, cytochrome c and AIF, in these cybrid cells after culturing them in the galactose medium for 24 h. Both cytochrome c and AIF were present in the cytosolic fractions isolated from homogenates of CI-6 cells (Fig. 4B). On the other hand, little cytochrome c or AIF was released from the mitochondria of CI-30 cells. We also determined cytochrome c and AIF levels in the mitochondrial fractions. After 24 h, there was less cytochrome c and AIF in the mitochondrial fraction of CI-6 cells than in that of CI-3 or CI-30 cells. Taken together, these results indicate that CI-6 cells are more prone to undergo apoptosis under nutrient stress in the galactose medium.

Figure 4.

A) Nuclear DNA was isolated from cells incubated in galactose medium for 24 h. Then 4 μg of DNA was separated in 1% agarose gel and stained with ethidium bromide. M, 1 kb plus DNA ladder. B) Release of cytochrome c and AIF in the cytosolic fraction in the CI-6 cybrid cell lines with wild-type nucleotide at position 4936. Protein from the cytosolic and mitochondrial fractions (20 μg) was separated by SDS-PAGE, and Western blotting was performed with specific antibodies against cytochrome c, AIF, and mitochondrial matrix protein SOD2.

DISCUSSION

We studied mtDNA alterations found in a patient with multiple neurological and autonomic problems that are typical of mitochondrial disease. Several alterations in the mtDNA are likely to be deleterious, although none have been proven to be pathogenic mutations. A change from cytosine to thymine at nt 4936 in the MT-ND2 gene was detected as a heteroplasmic change, which suggests a putative deleterious mutation (8, 9). This variant results in a change from threonine to isoleucine at aa 156 in the complex I subunit. This variant is rare (0.1% frequency), found only 3 times among 2700 apparently healthy individuals of different ethnicities, and has been reported to belong to haplogroup M1b1 (identified by changes 4936 C>T, 13111 T>C, and 15247 C>G; ref. 19). The amino acid at this position is conserved through the great apes, although it is not well conserved in lower species.

The heteroplasmic nature of the 4936 C>T nucleotide alteration in the MT-ND2 gene suggests that it may well contribute to the clinical phenotype. Transmitochondrial cybrid cells carrying the heteroplasmic as well as homoplasmic mtDNA 4936 C>T variant were generated to study the relationship among the accumulation of mtDNA mutation, mitochondrial dysfunction associated with this mutation, and cellular dysfunction. With an increasing load of the mutant MT-ND2 mtDNA, the endogenous respiration declined progressively, resulting in a reduced ATP synthesis rate and restricted cell growth compared with those of cells with wild-type mtDNA. At the same time, the intracellular and mitochondrial ROS levels decreased inversely, suggesting that the dysfunctional NADH dehydrogenase (complex I) was not efficient at transferring electrons into the ETC, thus reducing the ROS generated. The defect may be present in the pathway of electron transfer before reaching ROS production. Because the three-dimensional structure of complex I is not available, we decided to test whether this variant affected complex 1 assembly and performed a PAGE Blue experiment using anti-complex 1 antibody. Our results show that this mutation does not affect complex 1 assembly significantly (data not shown).

In our soft agar assay to test the tumorigenesis ability of the transmitochondrial cybrid cells, only cells with the homoplasmic mtDNA 4936 C>T variant formed colonies, indicating that the homoplasmic MT-ND2 mutation might promote several phenotypic changes, such as loss of contact inhibition and anchorage independence, which facilitate cellular transformation under low oxygen tension and limiting nutrient concentration. Interestingly, in our endeavor to generate induced pluripotent stem (iPS) cells from the patient's skin fibroblasts, we could not obtain iPS cells with wild-type or low heteroplasmy mtDNA 4936 C>T content. Most of the established iPS lines had homoplasmic or near homoplasmic 4936 C>T nucleotide alteration in the MT-ND2 gene (data not shown). All these results suggested that this particular MT-ND2 mutation may have an enhancing effect on cell growth and/or colonial expansion ability. To determine whether homoplasmic 4936 C>T nucleotide alteration in the MT-ND2 is indeed associated with tumor formation, it is important to do more experiments in vivo in the future.

The cybrid model is based on the repopulation of an immortalized, usually cancer-derived, cell line previously devoid of its original mtDNA but replaced with patient-derived mutant mitochondria (12, 13, 18). Cancer cells maintain a high glycolytic rate even in the presence of oxygen, a phenomenon first described >70 yr ago and known historically as the Warburg effect (20). Indeed, glycolysis is the major ATP-generating pathway of cultured cells grown in glucose-rich medium; only a minimal amount of pyruvate is oxidized to CO₂ and water in the mitochondria to generate ATP through oxidative phosphorylation (21). Galactose has 6 carbons like glucose and differs from glucose only in the stereochemistry of 1 carbon, C4. The enzymes of carbohydrate metabolism are specific enough that galactose must be changed to glucose before it can enter glycolysis. To convert galactose to glucose, it is first phosphorylated by galactokinase to produce galactose-1-phosphate. Galactose is then exchanged with the glucose group in UDP-glucose to create UDP-galactose and releases glucose-1-phosphate. An epimerase enzyme changes the stereochemistry of C4 in UDP-galactose, creating UDP-glucose. In the next round of the transfer reaction, this glucose is released as glucose-1-phosphate. Once released, glucose-1-phosphate is converted to glucose-6-phosphate and can enter glycolysis to generate energy. In glucose-free galactose medium, cultured cells are forced to use galactose, and the restricted flow of galactose to glucose-6-phosphate determines the formation of very little lactate because pyruvate, which is formed at a much slower rate, is further oxidized within the mitochondria (17, 22). Thus, cultured cells with defective mitochondrial metabolism often die quickly when cultured in glucose-free galactose medium (17, 22). Growth impairment of transmitochondrial cybrid cells carrying the LHON pathogenic mutations has been reported when cells were incubated in a medium containing galactose in place of glucose (15, 18, 23).

In the present study, we surprisingly found that cybrid cells with the wild-type nucleotide at position 4936 in the MT-ND2 gene died much faster when the cells were incubated in glucose-free galactose medium than those cells with heteroplasmic or homoplasmic nucleotide alteration. Two LHON ancillary alterations in the mitochondrially encoded NADH dehydrogenase 1 (MT-ND1) and MT-ND2 genes (4216 T>C and 4917 A>G) are present in the patient's mtDNA. However, these two nucleotides are not well conserved through evolution and are present in the population at frequencies of ∼9 and 5%, respectively. They are also included among the polymorphisms of the T haplogroup (MITOMAP). They have been identified as secondary mutations that may alter the severity of LHON when they occur with one of the primary LHON mutations (24, 25). Because no primary LHON mutations were detected in this patient, it is unlikely that these two alterations are deleterious in this patient.

A rare homoplasmic missense variant was found in the ATPase 6 gene (9181 A>G) in the patient's mtDNA. This variant results in a change from serine to glycine at aa 219, which is highly conserved in mammalian species as well as in chicken and fish mitochondrial DNA, suggesting that it is important for the function of the ATPase 6 protein. This variant is very rare, being reported only once among ∼2700 mtDNA sequences from healthy individuals from multiple ethnic groups (frequency 0.04%; ref. 26). To our knowledge, it has not been detected in individuals with mitochondrial disease. This missense variant is likely to be deleterious, but the magnitude of its contribution to the clinical presentation is uncertain. If this variant was heteroplasmic at a level of 80–90%, its likelihood of being a major contributor to the clinical picture would increase dramatically. Initially deemed counterintuitive, the fact that cells with the wild-type cytosine at position 4936 in the MT-ND2 gene died faster in galactose medium than cybrid cells with heteroplasmic or homoplasmic alterations suggests that the 9181 A>G variant is most likely deleterious, because all cybrid cells in our study had this variant in the ATPase 6 gene. For cells with the wild-type cytosine at 4936, the NADH dehydrogenase (complex I) efficiently stripped electrons from the substrate, transferred them into the ETC, and pumped protons into the mitochondrial intermembrane space, even though the defective ATP synthase should have impeded the smooth backflow of protons into the mitochondrial matrix to increase the intramitochondrial electron pool. We hypothesize that this excess of electrons together with additional electrons leaked from the ETC will generate excessive ROS, leading to a proapoptosis effect (5). In cybrid cells having an altered thymine at 4936, the slow inflow of electrons into the ETC caused by defective NADH dehydrogenase (complex I) keeps the intramitochondrial electron pool in check, resulting in lower intracellular and mitochondrial ROS levels and a counterbalancing antiapoptosis effect.

Most of other sequence variants observed in the patient's mtDNA are ancient alterations that are thousands or hundreds of thousands of years old. Although they are not causative for this patient's disease, the possibility that they may contribute to an increased risk of certain conditions in a multifactorial manner cannot be ruled out. Our results suggest that a particular mutation of mtDNA may contribute to the clinical phenotype modulated by the background of mtDNA and the availability of nutrients. The consequences of mtDNA variants could be tissue-specific, contributing to the diversity of clinical symptoms in mitochondrial diseases.