The authors have declared that no competing interests exist.
Conceived and designed the experiments: EEM WS BK. Performed the experiments: EEM SMB. Analyzed the data: EEM SMB BK. Contributed reagents/materials/analysis tools: OS. Wrote the paper: EEM SMB BK. Provided technical support: JAM.
Current address: Royal Brompton and Harefield NHS Trust, London, United Kingdom
Epidemiological case-control studies have revealed associations between mitochondrial haplogroups and the onset and/or progression of various multifactorial diseases. For instance, mitochondrial haplogroup T was previously shown to be associated with vascular diseases, including coronary artery disease and diabetic retinopathy. In contrast, haplogroup H, the most frequent haplogroup in Europe, is often found to be more prevalent in healthy control subjects than in patient study groups. However, justifications for the assumption that haplogroups are functionally distinct are rare. Therefore, we attempted to compare differences in mitochondrial function between haplogroup H and T cybrids.
Mitochondrial haplogroup H and T cybrids were generated by fusion of HEK293 cells devoid of mitochondrial DNA with isolated thrombocytes of individuals with the respective haplogroups. These cybrid cells were analyzed for oxidative phosphorylation (OXPHOS) enzyme activities, mitochondrial DNA (mtDNA) copy number, growth rate and susceptibility to reactive oxygen species (ROS). We observed that haplogroup T cybrids have higher survival rate when challenged with hydrogen peroxide, indicating a higher capability to cope with oxidative stress.
The results of this study show that functional differences exist between HEK293 cybrid cells which differ in mitochondrial genomic background.
Mitochondria provide most of the energy in a cell by a process called oxidative phosphorylation (OXPHOS), where carbohydrates and fats are oxidized by oxygen to produce carbon dioxide, water and adenosine triphosphate (ATP)
Although the majority of mitochondrial proteins are encoded by nuclear DNA and imported into the mitochondria, these multi-functional organelles also contain their own DNA (mitochondrial DNA, mtDNA). Twenty-four genes of the mtDNA code for components of the mitochondrial translational machinery (2 ribosomal RNAs and 22 transfer RNAs) and 13 genes provide essential subunits of the energy-generating enzymes of the OXPHOS pathway, namely complex I, III, IV and V. Only succinate dehydrogenase (complex II) is completely composed of nuclear encoded subunits
Mitochondrial function declines with age, and both mtDNA alterations and oxidative damage accumulate. Oxidative damage is produced by reactive oxygen species (ROS), and most of the cellular ROS, such as superoxide, hydrogen peroxide and organic hydroperoxides, are generated in mitochondria from single electrons escaping the mitochondrial respiratory chain and reducing molecular oxygen
During evolution, the human population accumulated a high number of mtDNA base substitutions along radiating maternal lineages, where specific combinations of polymorphisms constitute what are referred to as mitochondrial haplogroups. Researchers have used and continue to use these population-specific polymorphisms to elucidate long-ago human migrations and human pre-history
However, mitochondrial haplogroups have also been shown to be associated with multifactorial diseases
Mitochondrial polymorphisms defining mitochondrial haplogroups or definite combinations of such polymorphisms could slightly alter OXPHOS coupling and performance as well as ROS production.
To exclude the influence of a nucleus–mtDNA genetic interaction, an excellent in vitro system to study functional differences of mtDNA variations is the use of cybrids. Transmitochondrial cybrids are produced by fusion of cells depleted of their mtDNA (ρ0 cells) with cells devoid of nuclear DNA (e.g. thrombocytes). This technique is very useful to discriminate the functional consequences of mtDNA variations from the nuclear background
Our aim was to elucidate functional differences between the European mitochondrial haplogroups H and T. The activity of mitochondrial OXPHOS enzymes, mtDNA copy number, proliferation capacity as well as susceptibility to oxidative stress were compared between haplogroup H- and T-specific cybrids.
Based on our previous observation that patients with mitochondrial haplogroup T have a higher risk of developing CAD
Whole mtDNA sequencing of cybrids was performed and individual polymorphisms (compared to the Cambridge Reference Sequence) were used to construct a phylogenetic tree of sub-haplogroups, according to phylotree.org
The phylogenetic tree was constructed according to phylotree.org
Non-haplogroup-defining polymorphisms not present in all cybrids are listed in Supplementary
Because mtDNA encodes subunits of the OXPHOS system, variations of mtDNA could primarily affect the activity of OXPHOS enzymes. We determined possible functional differences between mitochondrial haplogroup H and T in HEK293 cybrids by measuring the enzymatic activities of OXPHOS enzymes and the tricarboxylic acid (TCA) cycle enzyme citrate syntase (CS) as a reference
There were no significant differences between HEK H and T cybrids in the enzymatic activities of CS or of OXPHOS complexes (
Haplogroup H |
Haplogroup T |
P-value |
|
n = 3 | n = 3 | ||
Citrate synthase (mUnits/mg protein) |
1089.2 (228.4) | 1005.8 (236.6) | 0.683 |
Complex I (mUnits/mg protein) | 17.7 (1.4) | 18.7 (2.4) | 0.562 |
Complex I (mUnits/mUnits CS) | 0.017 (0.004) | 0.020 (0.007) | 0.492 |
Complex I+III (mUnits/mg protein) | 92.0 (23.4) | 131.4 (26.5) | 0.125 |
Complex I+III (mUnits/mUnits CS) | 0.088 (0.033) | 0.133 (0.008) | 0.076 |
Complex II (mUnits/mg protein) |
243.7 (34.3) | 215.8 (43.9) | 0.434 |
Complex II (mUnits/mUnits CS) | 0.232 (0.037) | 0.244 (0.051) | 0.750 |
Complex II+III (mUnits/mg protein) | 431.8 (84.7) | 356.2 (105.3) | 0.388 |
Complex II+III (mUnits/mUnits CS) | 0.407 (0.010) | 0.393 (0.120) | 0.889 |
Complex III (mUnits/mg protein) | 451.9 (98.5) | 420.1 (169.5) | 0.792 |
Complex III (mUnits/mUnits CS) | 0.408 (0.095) | 0.454 (0.275) | 0.795 |
Complex IV (mUnits/mg protein) | 412.1 (55.1) | 316.5 (96.3) | 0.210 |
Complex IV (mUnits/mUnits CS) | 0.376 (0.085) | 0.349 (0.157) | 0.809 |
Complex V (mUnits/mg protein) | 195.1 (46.6) | 132.1 (33.6) | 0.130 |
Complex V (mUnits/mUnits CS) | 0.183 (0.021) | 0.149 (0.037) | 0.245 |
Values are given as mean ± standard deviation (SD).
P-value: Independent samples t-test.
Reported previously in
Enzymatic activity measurements were made on isolated mitochondria of cells grown in glucose medium with antibiotics, and on cells with five to 15 passages after the cybridization process.
Because mtDNA copy number is a reliable indicator of OXPHOS activity, we further analyzed mtDNA copy number in haplogroup-specific cybrid cells. Due to the fact that the inter-assay variation associated with the determination of mtDNA copy number is lower than the variation associated with the determination of OXPHOS enzyme activities, we hypothezised that true variation in mtDNA copy number has a better chance to impart subtle but significant differences among haplogroup-specific cybrids.
For determination of mtDNA copy number, the cells were cultivated in glucose or galactose medium. High concentrations of glucose in the cultivation medium frequently leads to inhibition of respiration (crabtree effect)
As expected, cultivation of cells in galactose medium resulted in a significantly higher mtDNA copy number compared to cells cultivated in medium supplemented with glucose (
(A) Comparison of all cybrid clones cultivated in glucose medium and galactose medium. (B) Comparison of HEK H and HEK T cybrids cultivated in glucose medium. (C) Comparison of HEK H and HEK T cybrids cultivated in galactose medium. Mean values of copy numbers are given; error bars: standard deviation; *p<0.05.
The proliferation capacity of cells mirrors their energy-producing capacity. Hence, we next investigated the proliferation rate of HEK H and T cybrid cells by two different methods.
Growth curves were determined by measuring the number of HEK H and HEK T cybrid cells using a CyQUANT® NF Cell Proliferation Assay Kit.
The mean growth rate of the cybrid cells was higher on days three to seven for HEK T compared to HEK H cybrids when cultivated in glucose medium, and significant p-values were obtained on days three and four (
The number of cells on the days given were normalized to the number of cells on day two and determined as growth rate. (A) Comparison of HEK H (n = 3; gray circles) and HEK T (n = 3; black squares) cybrids at days three to seven in glucose medium. (B) Comparison of HEK H (n = 3; gray circles) and HEK T (n = 3; black squares) cybrids at days three to seven in galactose medium. Mean values of growth rates are given; error bars: standard deviation; *p<0.05.
A more sensitive method of comparing the growth capability of different cells is a competitive mix experiment. The prolonged cultivation time and direct competition of cybrids within the same culture flask allows a direct comparison of growth rates. Each HEK H cybrid clone was co-cultivated with each HEK T cybrid clone at a 1∶1 mixture of cells. After 10, 20 and 30 days of co-culture, DNA was isolated and the proportion of each genotype was analyzed using TaqMan quantitative real-time PCR (qPCR) with probes specific for haplogroup H (7028C) or all other haplogroups (7028T, in our case indicative for haplogroup T).
After 10, 20 and 30 days in glucose medium, a trend toward haplogroup T as the dominant genotype was observed (
After 10, 20 and 30 days (d10, d20, d30) of co-culture, isolated DNA of the cell mixtures was analyzed using TaqMan qPCR. ΔCt values were calculated by subtraction of the mean Ct value of the FAM signal (probe recognizing haplogroup T) from the mean Ct value of the VIC signal (probe recognizing haplogroup H). ΔΔCt values were calculated by subtraction of the mean ΔCt values of the original cell mixtures (n = 36; day zero) from the mean ΔCt values of all co-cultures at days 10, 20 or 30 (n = 36; except for d30 in galactose: n = 35). Dominance of haplogroup H results in a negative ΔΔCt value and is presented as gray bars, whereas dominance of haplogroup T results in a positive ΔΔCt value and is presented as black bars. (A) ΔΔCt values of competitive co-cultures cultivated in glucose medium, at days 10, 20 and 30. (B) ΔΔCt values of competitive co-cultures cultivated in galactose medium, at days 10, 20 and 30. Mean ΔΔCt values are given; error bars: standard deviation.
Subtle differences between haplogroup-specific cybrids might only become apparent under stress conditions. Therefore, we challenged the cybrids with different concentrations of hydrogen peroxide (H2O2) and measured susceptibility to ROS by determination of cell viability.
Cells grown in galactose medium were more sensitive to H2O2 treatment than cells grown in glucose medium (
Cell survival was measured 24 hours after H2O2 treatment and calculated as a percentage of the ratio between treated and untreated cells (% cell survival). (A) Comparison of HEK H (n = 3; gray bars) and HEK T (n = 3; black bars) cybrids at 250 µM to 475 µM H2O2 in glucose medium without serum and without sodium pyruvate. (B) Comparison of HEK H (n = 3; gray bars) and HEK T (n = 3; black bars) cybrids at 100 µM to 250 µM H2O2 in galactose medium without serum and without sodium pyruvate. Mean values of % cell survival are given; error bars: standard deviation; *p<0.05.
Cell survival of HEK T cybrids was higher than that of HEK H cybrids. Statistically significant differences were found at all four concentrations of H2O2 in glucose medium as well as in the two highest concentrations (200 µM, 250 µM) of H2O2 in galactose medium (
Previous studies showing mitochondrial haplogroup T to be a risk factor for CAD and diabetic retinopathy, as well as other literature, motivated us to elucidate functional differences between haplogroups H and T. Haplogroup T was found to be a risk factor for developing peripheral neuropathy during antiretroviral therapy
Some attempts have already been made to determine differences between haplogroup H and other haplogroups in cybrid cells. Carelli et al. compared the European haplogroups H, T and J in cybrids derived from the osteosarcoma cell line 143B.TK-. There were no significant differences of oxygen consumption, inhibition of cellular respiration by rotenone, and of complex IV enzymatic activity
In another study, Caucasian haplogroups H and T cybrids generated from a human lung carcinoma cell line (A549.B2) also did not show functionally important bioenergetic differences
In contrast to haplogroup cybrid comparisons performed by other laboratories
Because there was no certain sequence variation consistently detectable in either HEK H or HEK T cybrids we hypothesize that the haplogroups themselves are responsible for the discriminative performance observed in the present study (Supplementary
Because in transmitochondrial cybrids the nuclear genetic background is excluded in all cases, effects of individual specific nuclear-mtDNA interactions are excluded. However, we cannot exclude that in mtDNA so far unknown epigenetic alterations can occur during disease progression, which are also still present in the cybrid lines.
Differences between haplogroup-specific cybrids have been observed, for instance in a recent study by Gómez-Durán et al., who compared haplogroup H with haplogroup Uk in cybrids derived from 143B.TK- cells
In a study of Moreno-Loshuertos et al.
In a similar way, HEK T cybrids of the present study might possess similar OXPHOS capacity (no differences in OXPHOS enzymatic activities), but lower respiration capacity per molecule of mtDNA, compared to HEK H cybrids. A compensatory mechanism, through ROS-induced up-regulation of mtDNA, adopted to overcome a slightly less efficient OXPHOS may be compatible with less growth in galactose medium. Moreover, a higher amount of antioxidative enzymes may be the reason for the HEK T cybrids to be more successful in buffering exogenous ROS exposure.
Mitochondrial haplogroup-specific differences were also analyzed in vivo. Martínez-Redondo et al. analyzed maximal oxygen uptake (VO2max), mitochondrial oxidative damage (mtOD), and mtDNA haplogroups in 81 healthy Spanish men. VO2max was significantly lower in haplogroup J compared to haplogroup H individuals. When mtOD in skeletal muscle was assessed, oxidative damage was found to be significantly higher in haplogroup H individuals (p = 0.04) and there was a positive correlation between mtOD and VO2max (p = 0.01)
Previous literature hints toward a lower uncoupling and higher ATP production through OXPHOS in mitochondrial haplogroup H. For instance, comparison of mitochondrial haplogroups H and T in sperm cells showed a significant reduction of complex IV activity in T sperm compared to H sperm (p = 0.0184), indicating a lower OXPHOS performance of haplogroup T
In conclusion, we were able to show that mitochondrial haplogroups H and T are functionally different in our model system. However, the functional differences between mitochondrial haplogroups and their consequences are far from being fully elucidated.
The generation of cybrids used in the present study has been described previously
According to Gómez-Durán et al.
Sequence analysis of the mtDNA was performed from two overlapping PCR fragments (107 to 8561; 7401 to 276) generated by long range PCR
Confluent cells were harvested, washed with phosphate-buffered saline (PBS), and mitochondria were isolated according to Bentlage et al.
Cybrid cells were cultivated in glucose or galactose medium. DNA was isolated three times for each cybrid clone. The cell pellet was washed in PBS and subsequently resuspended in proteinase K-containing buffer [2 mg/ml proteinase K (Roche, Basel, Switzerland) in 1×Reaction Buffer B used for Hot Fire Polymerase (Solis Biodyne, Tartu, Estonia)]. After incubation for at least one hour at 60°C, proteinase K was inhibited by incubation at 95°C for 10 minutes.
MtDNA content was determined by qPCR using SYBR Green SuperMix for iQ (VWR International, Radnor, Pennsylvania, USA). Two mtDNA fragments and two nuclear DNA fragments were amplified using 0.2 µM of primers, 1×SYBR Green Supermix for iQ, 1 µl of DNA (diluted 1∶10 to 1∶40) in a total volume of 10 µl. Thermal cycling conditions were: 95°C for 1 minute; 40 cycles at 96°C for 15 seconds, 63°C for 40 seconds and 72°C for 10 seconds; and finally 95°C for 1 minute, 55°C for 1 minute and a 0.5°C increase per cycle (80×5 seconds) from 55°C to 95°C for the generation of a melting curve. Primer sequences are listed in Supplementary
Over a period of six days the number of cybrid cells was measured by CyQUANT® NF Cell Proliferation Assay Kit (Invitrogen by Life technologies, Carlsbad, California, USA).
Cells were seeded at 1000 cells/ml in glucose medium and at 2500 cells/ml in galactose medium (200 µl per well) in a black 96-well plate. Cell number measurements were performed at 48, 72, 96, 120, 144 and 168 hours after seeding of the cells. We determined growth rates from day three to day seven, as there was no proliferation of cells observed before day three.
Measurements were carried out according to the manufacturer’s protocol using 50 µl dye solution and an incubation time of 60 minutes in the dark (37°C, 5% CO2). Fluorescence was measured (excitation: 490 nm, emission: 510–570 nm) on a GloMax®-Multi Microplate Multimode Reader (Promega, Madison, Wisconsin, USA).
The median of the fluorescence units of eight wells was calculated for each clone. Measurements of 72, 96, 120, 144, 168 hours were normalized to the value at 48 hours (determined as growth rate).
Each HEK H cybrid cell line was co-cultured in an initial 1∶1 ratio with each HEK T cybrid cell line either in glucose or galactose medium, resulting in a total of 36 co-culture combinations. After 10, 20 and 30 days of co-culture, the proportion of each genotype was determined by qPCR using TaqMan methods and probes specific for 7028C (haplogroup H) and 7028 T (non haplogroup H)
The cells were seeded at 2500 cells/ml (glucose medium) or at 10,000 cells/ml (galactose medium) in 12-well plates (1 ml per well; 2 wells per co-culture). The medium was changed on days two, six, 12, 16, 22, 26 or additionally if necessary. Before reaching confluency, on days 10 and 20, co-cultures were trypsinized and sub-cultured. DNA was isolated on days 10, 20 and 30.
The genotype proportion in the mix experiments was determined by qPCR, using TaqMan® Gene Expression Master Mix and TaqMan reagents (Applied Biosystems by Life Technologies, Carlsbad, California, USA). Primer and probe sequences are listed in Supplementary
The PCR mixture contained 0.9 µM primers, 0.05 µM 7028C probe (VIC-labelled), 0.1 µM 7028 T probe (FAM-labelled), 1×TaqMan® Gene Expression Master Mix and 1 µl of DNA (diluted 1∶10 to 1∶100) in a total volume of 15 µl. Thermal cycling conditions were: 95°C for 8.5 minutes and 40 cycles at 95°C for 15 seconds and 65°C for 1 minute.
The proportion of a specific haplogroup in the mixture was calculated with the following formula: ΔCt = Ct (VIC)−Ct (FAM). Positive ΔCt values stand for a higher ratio of haplogroup T in the sample and negative ΔCt values reveal a higher fraction of haplogroup H. ΔΔCt values represent ΔCt values at days 10, 20 or 30 minus the initial ΔCt values (day zero; only one mixture).
In cases where only the signal of one probe was detected, as this haplogroup has probably displaced the other, the Ct value of the second probe was set to 31, in order to be able to calculate a shift. This value was determined as the highest detectable Ct value was 31.
As two wells were seeded per co-culture, the mean of both wells was calculated and used as the Ct value of the respective mixture in the statistical analysis.
Cells were seeded in 96-well plates at 105 cells/ml in glucose medium (200 µl per well). Twenty-four hours after seeding, the medium was changed. For ROS susceptibility experiments, medium without serum and without sodium pyruvate was used in order to avoid potential antioxidative components in the medium
Twenty-four hours after the change of medium, the cells were treated with different concentrations of H2O2. For each concentration of H2O2, eight wells were treated per clone. Cell survival was analyzed after 24 hours using resazurin (resazurin sodium salt, Sigma-Aldrich, St. Louis, Missouri, USA). Resazurin is reduced in living cells to the pink fluorescent dye resorufin
The median of the fluorescence units of eight blank wells (only medium with dye) was subtracted from the median of the fluorescence units of the eight wells per clone. Cell survival at a certain concentration of H2O2 was calculated as the percentage of the fluorescence units of the treated cells in relation to non-treated cells.
Normal distribution was checked by the Kolmogorov-Smirnov test. Differences between mean enzymatic activities, mtDNA copy number, growth rates and cell survival upon challenge with H2O2 were statistically evaluated using an independent samples t-test (for normally distributed variables) or a non-parametric Mann-Whitney U-test (for not normally distributes variables). In the competitive mix experiments, variables were analyzed using a dependent samples t-test (for normally distributed variables) or a Wilcoxon signed-rank test (for not normally distributed variables). A p-value<0.05 was considered statistically significant. All analyses were performed using PASW Statistics 18.0 (SPSS GmbH).
(PDF)
(PDF)
(PDF)
(PDF)
We thank Thomas Verwanger for constructive discussions as well as Waltraud Eder for support in statistical analysis. Moreover, we thank Eduardo Ruiz-Pesini and Aurora Gómez-Durán for providing their protocol and primer/probe sequences for the competitive mix experiment.