Author contributions: W.W.: conception and design, data collection and analysis, final approval of manuscript; P.O., Y.E., and R.S.: data collection and analysis, final approval of manuscript; M.B.: conception and design, financial support, final approval of manuscript; L.E.: conception and design, financial support, manuscript writing and final approval.
First published online in STEM CELLS EXPRESS October 15, 2010.
Disclosure of potential conflicts of interest is found at the end of this article.
Differentiation of neural stem cells (NSCs) involves the activation of aerobic metabolism, which is dependent on mitochondrial function. Here, we show that the differentiation of NSCs involves robust increases in mitochondrial mass, mitochondrial DNA (mtDNA) copy number, and respiration capacity. The increased respiration activity renders mtDNA vulnerable to oxidative damage, and NSCs defective for the mitochondrial 8-oxoguanine DNA glycosylase (OGG1) function accumulate mtDNA damage during the differentiation. The accumulated mtDNA damages in ogg1−/− cells inhibit the normal maturation of mitochondria that is manifested by reduced cellular levels of mitochondrial encoded complex proteins (complex I [cI], cIII, and cIV) with normal levels of the nuclear encoded cII present. The specific cI activity and inner membrane organization of respiratory complexes are similar in wt and ogg1−/− cells, inferring that mtDNA damage manifests itself as diminished mitochondrial biogenesis rather than the generation of dysfunctional mitochondria. Aerobic metabolism increases during differentiation in wild-type cells and to a lesser extent in ogg1−/− cells, whereas anaerobic rates of metabolism are constant and similar in both cell types. Our results demonstrate that mtDNA integrity is essential for effective mitochondrial maturation during NSC differentiation. STEM CELLS 2010;28:2195–2204
Mitochondria are subcellular organelles with a heterogeneous composition that varies depending on tissue type . Specialization of cells in differing tissue, therefore, relies on the corresponding specialization of cellular mitochondria. The process by which mitochondrial components become specialized, known as mitochondrial maturation, is essential for stem cell differentiation . Neural cells have a high aerobic metabolic rate, which is higher than that in neural stem cells (NSCs) . In neurons, sufficient respiratory capacity provided by functional mitochondria is essential to restore resistance to excitotoxic effects of neurotransmitter activation . The heterogeneity of mitochondria in the brain is important for ensuring normal neuron-astrocyte interactions by balancing nitrogen metabolism . Differentiation of NSCs to neurons, astrocytes, or oligodendrocytes is controlled by several factors including oxygen availability and intracellular redox state [6–8]. Mitochondrial reactive oxygen species (ROS) serve important signaling roles in both oxygen sensing and redox balance [9, 10]. The generation of ROS is an inevitable byproduct of electron transport chain (ETC) activity; therefore, increased aerobic activity can potentially influence the differentiation process and render mitochondrial macromolecules vulnerable to oxidation damage.
Mitochondrial DNA (mtDNA) residing in the mitochondrial matrix is readily damaged by ROS. Damage and mutations to mtDNA have a critical effect on ETC activity because mtDNA encodes 13 essential subunits of ETC complex I [cI], cIII, cIV, and cV. cII (succinate dehydrogenase) is composed of nuclear encoded subunits. Damages to mtDNA potentially can interfere with transcription and replication. The mitochondrial transcription factor A (TFAM) is present in sufficient quantities to completely cover mtDNA, it plays an essential role in both transcription and replication of mtDNA, and it binds preferentially to damaged DNA [11, 12]. In the mitochondrion, mtDNA damage is removed by the base excision repair (BER), an innermembrane-associated multienzyme DNA repair cascade . The BER pathway is initiated by a DNA glycosylase, like the 8-oxoguanine DNA glycosylase (OGG1), which recognizes and removes several base lesions including 8-oxoguanine. The resulting abasic site is processed by the Ape1 endonuclease, which cleaves the phosphodiester bond on the 5′ side thereby producing a substrate for mtDNA polymerase γ. Its nucleotide incorporation prior to ligation by DNA ligase IIIα completes the repair pathway. Mice lacking OGG1 specifically accumulate 8-oxoguanine in mtDNA  but are otherwise normal. In contrast, deletion of any downstream repair gene in the pathway results in embryonic lethality. It is thought that substrate overlap between different mtDNA glycosylases may explain the apparent innocuous effect of BER defect and that the repair deficiency only manifests into phenotype during stress conditions.
Cellular biogenesis of mitochondria is mediated by a redox/AMP-sensitive activation of nuclear encoded respiration factors via the coactivator peroxisome proliferator-activated receptor gamma coactivator 1-alpha (PGC-1α), which also controls the expression of other genes involved in aerobic, mitochondrial metabolism, including TFAM . Given the potential impact of mitochondrial biogenesis during neural differentiation, we investigated the vulnerability of mtDNA to damage during differentiation, focusing on cells that are deficient in OGG1. It is expected that these cells accumulate oxidative damages more frequently than wild-type cells. By comparing the effects that mtDNA damage has on mitochondrial maturation, we aimed to characterize the way mtDNA integrity influences on mitochondrial biology and neural differentiation.
MATERIALS AND METHODS
C57BL/6 mice and C57BL/6 OGG1 knockout mice (ogg1−/−) were bred in-house. The generation of ogg1−/− mouse has been reported previously . Animals were housed in accordance with the laws and regulations controlling experimental procedures in Norway and the European Union's Directive 86/609/EEC.
Primary Cultures of Neural Stem/progenitor Cells
Primary NSCs were prepared from hippocampus of C57BL/6 and C57BL/6 OGG1 knockout mice at postnatal day 5 and propagated as previously described with modifications [17, 18]. Low passage NSCs (P2–P8) were used throughout all the experiments. Briefly, neurospheres were isolated from hippocampal dentate gyrus and cultured in neurosphere proliferation medium to maintain multipotential NSCs. Initial passages of cells were cultured for 7 days as floating neurospheres in proliferation medium consisting of serum-free Neurobasal-A medium with 2% B27 supplement, 20 ng/ml basic fibroblast growth factor, 10 ng/ml epidermal growth factor (EGF), 2 mM L-glutamine, and penicillin/streptomycine. Subsequently, neurospheres grown for 7 days were trypsinized, dissociated, and plated as single cells at density of 2.5 × 104 per centimeter square onto poly-L-lysine-coated plates or flasks in differentiation medium (proliferation medium without EGF) until further experiments. In our in vitro differentiation setup, the growth medium was not changed during the cultivation, and consequently cells started to die approximately 1 week after plating.
Cellular Respiration Assay
Oxygen consumption of trypsinized cells at 37°C was determined by high-resolution respirometry (Oroboros Oxygraph-2K, Innsbruck, Austria, http://www.oroboros.at). The integrated software (datlab 4.2) presents respiration as oxygen flux; pmol O2 per 106 cells per second. The stirrer speed was set to 750 rpm. Briefly, cells were trypsinized for 2 minutes and resuspended in 2 ml differentiation medium. The cells were transferred to chambers maintained at 37°C. Basal respiration was measured for 15 minutes and 2 μg/ml of ATP synthase inhibitor oligomycin was added to determine nonphosphorylating respiration. Mitochondria were thereafter uncoupled with potentially carbonyl cyanide p-(trifluoromethoxy)-(phenylhydrazone) (FCCP) at step titration of 0.5 μM each adding into the chambers till maximal respiration capacity was obtained.
For Seahorse metabolic profiling, cells were plated at 5 × 104 cells per well onto 24-well plate and analyzed in Seahorse XF24-3 analyzer (Seahorse Bioscience, Billerica, MA, http://www.seahorsebio.com/) after equilibration and calibration according to the manufacturer's instructions. The wells were coated with poly-L-lysine prior to plating and differentiation, similarly as above-mentioned. One hour prior to recording, medium was replaced by Seahorse Dulbecco's modified Eagle's medium containing 5 mM glucose, 2 mM pyruvate, and 2 mM glutamine.
Immunocytochemistry and Live Cell Imaging
Immunohistochemistry for neurons and astrocytes was performed as described previously . The cells were fixed with 4% paraformaldehyde, permeablized with 0.1% Triton X-100 in PBS. Following blocking for 30 minutes, the cells were incubated with monoclonal anti-neuron-specific β-III tubulin (Tuj-1, 1:200, R&D Systems, Minneapolis, MN, http://www.rndsystems.com/), rabbit polyclonal astrocyte-specific anti-glial fibrillary acidic protein (GFAP; 1:200, Sigma, St. Louis, MO, http://www.sigmaaldrich. com), or monoclonal anti-nestin (1:200, Chemicon, Temecula, CA, http://www.millipore.com/) at 4°C overnight. Then the cells were incubated with fluorescent anti-mouse secondary antibody (Alexa 594, 1:500, Molecular Probes, Eugene, OR, http://www. invitrogen.com). The nuclear dye 4′,6-diamidino-2-phenylindole (DAPI) at 1 μg/ml (Molecular Probes, Eugene, OR) was added to visualize all cells. To obtain the percentage of each cell type, 1,300–1,500 cells in 10 random fields were counted. Percentage of positive cells was calculated in relation to total cells, as detected by DAPI nuclear staining.
Immunohistochemistry for cI was performed on cells fixed in ice-cold MeOH for 5 minutes. After blocking for 15 minutes, the cells were incubated with cI monoclonal antibody (39 kDa subunit, 1:400, Molecular Probes, Eugene, OR). Then the cells were incubated with fluorescent anti-mouse secondary antibody (Alexa 488, 1:300, Molecular Probes, Eugene, OR). Images were retrieved with a Zeiss LSM 510 META confocal microscope, with a ×63 objective, using an Ar488 laser line.
For Mitochondrial membrane potential (MMP) assessments, adherent cells were incubated in cultivation medium with 25 nM tetramethylrhodamine (TMRM) for 15 minutes at 37°C, and images were retrieved with a Zeiss LSM 510 META confocal microscope, with a ×20 objective.
Quantitative Real-Time Polymerase Chain Reaction
Total DNA containing nuclear DNA and mtDNA was isolated using the QiagenAmp Blood and Tissue Kit following manufacturer's protocols with minor modifications (Qiagen, Hilden, Germany, http://www.qiagen.com). Total RNA from cultured cells was isolated using the RNeasy Kit (Qiagen) following manufacturer's protocols with modifications. The isolated total DNA and RNA were quantified using the ND-1000 spectrophotometer (Nano-drop Technologies, Saveen & Werner AB, Sweden, http://www.nanodrop.com). Quantitative real-time polymerase chain reaction (PCR) was performed with a 7900HT Fast Real-Time PCR System (Applied Biosystems) using the Power SYBR green PCR Master mix (AB Applied Biosystems, Cheshire, U.K., http://www.appliedbiosystems.com). PCR was carried out with 30 ng of total DNA in a total volume of 20 μl containing 2X Power SYBR Green PCR master Mix, 0.5 μM of each primer. The real-time PCR was run with the standard cycling conditions preset in ABI PRISM 7000: 50°C for 2 minutes, 95°C for 10 minutes, then 95°C for 15 seconds, and 60°C for 1 minute for 40 cycles. A relative standard curves were established by plotting the threshold value versus the log of serial dilutions of total DNA added to the reaction according to the protocol described in Bulletin #2 (Applied Biosystems, CA). Relative mtDNA copy numbers were calculated based on the standard curve and the ratio of the amount of mtDNA versus 18S for each sample. The values were normalized to the amount of mtDNA of undifferentiated cells and converted into fold changes according to the manufacturer's guidelines.
For mitochondrial RNA (mtRNA) levels, the total RNA (0.3–0.5 μg) was reverse transcribed into first strand cDNA using High Capacity cDNA reverse transcription kit (Applied Biosystems). Quantitative real-time PCR was performed and relative mtRNA levels were calculated based on the standard curve and the ratio of the amount of mtRNA versus glyceraldehyde 3-phosphate dehydrogenase (GAPDH) for each sample as described earlier. Oligonucleotide primers used were: mtDNA and mtRNA (see primers 117 bp product for mtDNA damage detection below); 18S, forward 5-tagagggacaagtggcgttc, reverse 5-cgctgagccagtcagtgt ; GAPDH, forward 5-tcgtcccgtagacaaaatggt, reverse 5-cgcccaatacggccaaa .
mtDNA Damage Detection
The following two distinct methods were used to evaluate mtDNA damage:
(a)Quantitative PCR-Based amplification of a large fragment of mtDNA .
Briefly, each 30 ng of total DNA was used in two PCR reactions, one amplifying a short 117-bp product of mtDNA and the other amplifying a long fragment of 10-kb product of mtDNA. The amount of PCR products from the 10-kb amplification reflects the quality of the template mtDNA, whereas the 117-bp amplification serves as an internal control for template concentration. The primer sequences were: 10-kb product, forward 5-gccagcctgacccatagccataatat, reverse 5-gagagattttatgggtgtaatgcgg; 117-bp product, forward 5-cccagctactaccatcattcaagt, reverse 5-gatggtttgggagattggttgatg . PCR reaction was run with the following conditions: 10-kb PCR, denaturation at 94°C for 15 seconds, annealing, and extension at 68°C for 12 minutes for 20 amplifying cycles, followed by a final 10 minutes extension; 117-bp PCR, 94°C for 30 seconds, 60°C for 45 seconds, 72°C for 45 seconds for 20 cycles, and followed by final extension at 72°C for 10 minutes. The PCR reaction mixture was separated on an agarose gel stained with SYBR Safe DNA gel stain (Invitrogen, Carlsbad, CA) and PCR product quantified by densitometric analysis. mtDNA integrity was determined by calculating the ratios of amount of 10 kb products to that of 117 bp products and normalized to the average ratio obtained from nondifferentiated wt cells, which was set to 100%.
(b)mtDNA damage-inhibition of restriction enzyme cleavage.
The method is modified from that described by Vermulst et al.  and based on the ability of damages to inhibit restriction enzyme cleavage. Real-time PCR is subsequently used to quantify the amount of noncleavaged DNA template after restriction enzyme digestion. See Supporting Information for details.
Isolation of Supercomplexes, Protein Gel Electrophoresis, and Western Analysis
Isolation of supercomplexes was carried out as described [23, 24]. Briefly, NSCs or differentiated cells were harvested (differentiated cells were trypsinized) and cell pellets were frozen at −70°C. When all the samples were collected, the pellets were thawed on ice and the mitoplasts were isolated with digitonin solubilization (3 mg/ml in 200–400 μl total volume for 10 minutes on ice). Respiratory supercomplexes were extracted from the mitoplasts by a second digitonin solubilization (6 mg/mg protein, final concentration of 2% digitonin, for 30 minutes). Protein concentration was determined by Bradford Reagent Dye (BioRad), and 5–10 μg was separated on a 3%–12% NativePAGE Bis-Tris Gel (Invitrogen) (30V/40 minutes, 90 V/5 hours).
Total cell lysate (10–30 μg) was separated on 10% NuPAGE (Invitrogen), and gels were blotted onto poly(vinylidene difluoride) membrane (PVDF membrane; MilliPore, Billerica, MA) followed by blocking and probing against specific subunits of the respiratory complexes.
Western analysis was performed as described previously . Briefly, separated supercomplexes and the separated subunits from blue native gel electrophoresis (BN-PAGE) and sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) electrophoresis were blotted to a PVDF membrane using a semidry blotting procedure. Following blotting, the membrane was blocked with 5% dry milk in PBS/0.1% Tween 20. Primary antibodies used were: A21344 α39 kDa (0.5 μg/ml; cI), A11142 α70 kDa (0.1 μg/ml; cII), A21362 core I (0.1 μg/ml; cIII), A6403 subunit I (1 μg/ml; cIV), and A21350 α subunit (0.5 μg/ml, CV). All antibodies were purchased from Molecular Probes (Invitrogen, Carlsbad, CA). Secondary antibody was alkaline phosphatase (AP)-conjugated goat anti-mouse IgG (sc-2047; Santa Cruz Biotechnology, Santa Cruz, CA). Fluorescent signals from enhanced chemifluorescence (ECF) substrate (GE Healthcare, Chalfont St Giles, U.K.) were detected by KODAK Fluorescent Imager and analyzed by Kodak Molecular Imaging software.
Biochemical NADH Oxidoreductase Assay
The cI activity assay was performed basically as described by Janssen et al. . A minimum of 5 × 106 cells were homogenized by 20 strokes with pestle B (clearance 0.0005–0.0025 inch; from Kimble-Kontes). Protein amounts were determined by Bradford assay, and mitochondrial fractions frozen at −70°C for later complex assessment. Immediately before analyses, the mitochondria were thawed at 25°C and diluted to 0.20–1.5 μg/μl and 20 μl was added to a the reaction mixture (25 mM KHPO4 pH 7.8, 3.5 g/l bovine serum albumin (BSA), 60 μM 2,6-dichloroindophenol, 70 μM decylubiquinone, and 1 μM Antimycin A). The sample was preincubated at 37°C for 3 minutes before the reaction was started by addition of 200 μM NADH. A parallel reaction containing 1 μM rotenone served as blank, and absorbance (600 nm) was sampled every 30 seconds for 10 minutes. Enzymatic activity was determined from the initial rate constant, calculated from assumption of first-order reaction.
Fluorescent Cytometry Analyses
Cells were detached by trypsinization fixed with 70% ethanol. The cells were washed with PBS and incubated with 20 nM nonylacridine orange (Invitrogen, Carlsbad, CA) for 20 minutes at 37°C. Flow cytometry analysis was performed with excitation 488 nm and emission 550 nm.
All results were performed with at least three independent preparations of NSC from hippocampal dentate gyrus, and presented as average with standard deviation or error. Student's t test was employed to calculate significant, assuming unpaired, equal variance. *p < .05; **p < .01.
Strong Mitochondrial Biogenesis During Differentiation of NSCs In Vitro
Primary NSCs were isolated from hippocampus (dentate gyrus, P5), and expanded in vitro. These cells readily differentiate into neurons and astrocytes on growth factor removal. Similarly to what was reported, more than 90% of cells expressed Nestin, a precursor filament protein for progenitor neuronal cells at day 1 (D1) of differentiation. The nestin-positive cells were reduced to less than 20% after D4 (Wang et al., unpublished observations). Tuj-1-positive neurons were identified as early as 24 hours post plating, whereas astrocytes developed slower and expressed GFAP after D2. At D5, the cells were recognized as 34% of neurons and 15% of astrocytes (Fig. 1A). To study mitochondrial biogenesis, we evaluated mitochondrial parameters in the differentiating cells. First, cI was identified in both nondifferentiated cells (D0) and in cells 3 days postdifferentiation (D3; Fig. 1B). The cI staining was concentrated around nucleus in D0, whereas much more distributed throughout the neurites in D3. This indicates that distribution of ETC does not coincide with the initial extension of neurites but occurs later. MMP was assessed by TMRM uptake, and TMRM staining correlated to the pattern of cI in that fluorescent intensity was more concentrated in D0 compared with D3 (Fig. 1B). In correlation with the expansion of mitochondrial network, mtDNA increased strongly during the first 3 days of differentiation, finally reaching fourfold elevation in copy number after 5 days (Fig. 1C). The elevation coincided with an augmented ability of the cells to bind the fluorescent dye nonylacridine orange, which is a quantitative marker of cardiolipin in the mitochondrial inner membrane, demonstrating that mitochondrial mass increased during differentiation (Fig. 1D). The functionality of the mitochondria was monitored by respiratory analyzes after carefully detaching cells from the plate and transferring them into an Oxygraph-2K (Fig. 1E). Respiration characteristics were evaluated as initially described by Yadava et al.  (Fig. 1F), where cellular respiration was measured first in the “resting state” in the absence of any additives. Oligomycin, an inhibitor of ATP synthase, was then added to inhibit oxidative phosphorylation, to unravel phosphorylation-coupled respiration. Next, the mitochondrial uncoupler FCCP was titrated to saturation to induce maximal mitochondrial respiration. Maximal respiration is an estimate of the respiration capacity of the mitochondria in vivo. It was found that the respiration capacity increased in a near linear mode during the first 5 days of differentiation (Fig. 1E). The ratio of phosphorylation-coupled respiration to respiration capacity remained constant, while ratio of resting respiration to respiration capacity increased slightly during the differentiation (Supporting Information Fig. S1). Rotenone is a cI inhibitor, and rotenone addition nearly abolished respiration, inferring that the respiration was mainly cI dependent in these conditions (Fig. 1F). However, when cells were permeabilized and assessed for succinate oxidation, the robust increase in respiration verified the presence of functional succinate dehydrogenase/cII as well (Fig. 1F).
OGG1 Is Required for Functional Mitochondrial Maturation
The increased respiration during differentiation is believed to mediate elevated intramitochondrial levels of ROS, which potentially damage mtDNA. To test this hypothesis, we examined the integrity of mtDNA in OGG1-deficient cells, which have reduced ability to repair mtDNA. The integrity of mtDNA was assessed by the quantitative PCR method developed by van Houten and coworkers  and is based on that mtDNA damage statistically lead to a selective disability to amplify a large product over a short product. When the method was applied on differentiated cells, it was found that mtDNA integrity became severely reduced during differentiation, and to a significantly larger extent in ogg1−/− cells than in wt cells (Fig. 2A). There was no significant difference between mtDNA integrity in nondifferentiated cells. We employed a separate method that is based on the ability of mtDNA damage to inhibit restriction enzyme cleavage to verify the mtDNA damage load during differentiation (Supporting Information Fig. S2). Thus, differentiation indeed exerts a damaging effect on mtDNA and functional mtDNA repair is essential to counteract damage accumulation.
The reduced integrity of mtDNA in ogg1−/− cells is likely to influence on the maturation of mitochondria. To address this point, we first compared the mtDNA copy numbers in differentiated wt and ogg1−/− cells (D3; Fig. 2B). The mtDNA copy numbers in nondifferentiated cells (D0) were similar for the two cell types. Although mtDNA concentration was increased in ogg1−/− cells, the level was strongly attenuated compared with wt cells. To evaluate mitochondrial transcription, we quantified mtRNA of the ND6 gene by quantitative PCR. mtRNA increased strongly during the first 3 days in wild type, but the increase was significantly decreased in ogg1−/− cells (Fig. 3C). In parallel with these observations, respiration capacity in differentiated ogg1−/− cells was correspondingly reduced (Fig. 2D). In nondifferentiated cells, however, mtDNA copy numbers, mtRNA levels, and respiration capacities were similar in wt and ogg1−/− (Fig. 2B–2D). Thus, mtDNA integrity is important for normal mtDNA synthesis, which is an early event in mitochondrial biogenesis. The attenuated respiration capacity in ogg1−/− cells correlates with the reduced mtRNA and mtDNA levels.
To address the correlation between mtDNA integrity and mitochondrial biogenesis, we manipulated mtDNA integrity by administration of the antioxidant N-acetylcysteine (NAC), a precursor of glutathione, or buthionine sulfoxide (BSO), an inhibitor of glutathione synthase. NAC treatment improved mtDNA integrity, whereas BSO exerted the opposite effect (Supporting Information Fig. S3). Interestingly, the treatments induced corresponding alterations on mtDNA copy numbers (Supporting Information Fig. S4). Inducing mtDNA damage by BSO reduced mtDNA copy number, whereas NAC exposure reduced mtDNA damage and stimulated mtDNA synthesis. The effects were observed in wt and ogg1−/− cells although the increased mtDNA copy number in ogg1−/− after NAC administration was not statistically significant. This may be explained by the insufficient protection of mtDNA in the ogg1−/− cells (Supporting Information Fig. S3). Thus, these experiments demonstrate the correlation between mtDNA integrity and mtDNA copy number.
Divergent Mitochondrial ETC Development During Neural Differentiation
The mitochondrial maturation is expected to be the result of an underlying balanced biogenesis in mtDNA copy number, mitochondrial mass, and ETC function. As shown in Figure 1F, respiration was mainly cI dependent. Thus, it appears that impaired mitochondrial maturation in ogg1−/− cells is caused by alterations in ETC function, involving cI level and activity. As shown in Figure 2C, the level of ND6 transcript increased more in wt than in ogg1−/− cells, inferring that cI levels could be the rate limiting for mitochondrial maturation. To follow up on this point, we examined the ETC function by three distinct approaches. First, levels of ETC subunits were quantified by Western analyses using antibodies against distinct subunits of either of the complexes. The levels of cI, cIII, and cIV increased during differentiation, whereas cII levels remained constant (Fig. 3A). The average increase at cI, cIII, and cIV levels at D3 was twofold to threefold relative to D0. The same complexes also increased in ogg1−/− cells, but the elevation was attenuated by average 1.5-fold relative to D0 (Fig. 3A). Second, we analyzed cI activity in mitochondrial fractions of cells. It was found that cI activity in D0 cells was as prominent as in D3 (Fig. 3B). Furthermore, there were no differences in cI activity between wt and ogg1−/− cells, neither as NSCs nor as differentiated cells.
It has been found that ETC complexes assemble into supercomplexes, and there are reports suggesting that these heterocomplexes are necessary for efficient respiration . In our experimental model, we examined the supercomplex organization as a third approach to search for differences in ETC function that could contribute to the increased respiration capacity in differentiated cells. Supercomplexes were isolated from mitoplasts from D0 and D5 cells in a stepwise digitonin-based solubilization procedure as described  and separated by BN-PAGE. We identified supercomplexes in all cells (Fig. 3C), and detected cI, cIII, and cIV by a two-dimensional BN-PAGE/SDS-PAGE method (data not shown). It was found that the supercomplex assembly of cI, cIII, and cIV was similar in D0 and D5 and homogenous in wt and ogg1−/− cells. The BN-PAGE/Western analyses in Figure 3C are semiquantitative, and there was a possibility that differences in supercomplex levels would not be identified because of low resolution due to experimental artifacts such as epitope masking/G-250 interference. By titration experiments, we determined the accuracy to be approximately ±20%–25% (data not shown). Thus, the fourfold difference in respiration capacity (Fig. 1E) cannot be related to a similar increase in supercomplex levels that is masked in the BN-PAGE experiments (Fig. 3C). Thus, in combination, these results demonstrate that the cellular mitochondrial ETC subunits in general increase during differentiation, except for the cII subunit. However, the overall cI activity and supercomplex organization of cI per mitochondrion are similar during differentiation.
Aerobic Metabolism Increases, Whereas Anaerobic Metabolism Remains Constant During Differentiation
The elevated respiration activity implies that aerobic metabolism increases during differentiation. To sort out whether the elevation reflects a general activation of metabolism or alternatively that the aerobic metabolism is specifically stimulated, the differentiating cells were examined by simultaneous real-time analysis of oxygen consumption rate (OCR) and extracellular acidification rate (ECAR) in a Seahorse extracellular flux analyzer. Although OCR reflects respiration activity, ECAR is an estimate of the glycolytic activity. wt D3 cells had a higher respiration capacity than ogg1−/− D3 cells (Fig. 4A), in agreement with the results obtained from the oxygraph-2K measurements with trypsinized cells. Because of technical limitations, we could not analyze suspension NSCs, but at 4 hours after differentiation initiation, there was no significant difference between wt and ogg1−/− cells (∼D0). Both resting respiration and respiration capacity (identified by FCCP-induced uncoupling; Fig. 4A) were increased in D3 cells compared with D0, but ogg1−/− cells displayed only 60% OCR values of wt cells. On the other hand, there was no significant divergence in ECAR activities between wt and ogg1−/− cells, even after 4 hours, 3 days, or 5 days (Fig. 4B and data not shown). Only after 5 days, FCCP-induced ECAR values in wt cells were significantly higher than in ogg1−/− cells (Supporting Information Fig. S5). These results demonstrate that OGG1 deficiency and/or mtDNA integrity influences on aerobic metabolism, whereas glycolytic flux remains constant in the differentiating cells and is independent of OGG1 function.
The results from this study demonstrate that differentiation of NSCs involves maturation of mitochondria, and the integrity of mtDNA plays an essential role in the process. The transition from the stem cell state into neurons and astrocytes involves activation of aerobic metabolism, whereas glycolysis rate remains constant. The aerobic activation is mirrored by increased mitochondrial capacity with respect to mitochondrial mass, mtDNA copy number, ETC function, and respiration capacity. The differentiation process exerts damaging insults on mtDNA, which if unrepaired, results in impaired mitochondrial maturation.
Although mitochondrial bioenergetics during neural differentiation to our knowledge has been an unexplored field, results from other types of progenitor cells suggest that differentiation generally involves increased aerobic activity. This includes upregulation of mitochondrial mass (cardiolipin level), mtDNA copy number, ETC capacity, and activity of NADH-generating enzymes including those involved in fatty acid oxidation [2, 3, 28–31]. Our data on neural mitochondria correlate with these observations. The stable level of cII subunits in contrast to the subunits of the other complexes implies that the composition of mitochondria from differentiated cells differ from that of NSCs. It is known that the cI activity differ between cells, tissue, and in different subregions of the brain [32, 33]. The various complexes mature independently during postnatal development in brain mitochondria. In addition, the ETC establishment is different in synaptosomal versus nonsynaptosomal mitochondria [34, 35]. Although cII was similar at D0, D3, and D5 in this in vitro model, cII is shown to increase strongly after birth, stimulated by NO . The heterogeneity in mitochondrial complexes are therefore established in the early differentiation and later developmental stage, for instance during the migration or on integration in intercellular circuits in vivo. Our data imply that mtDNA integrity is important for the initial development of cI, cIII, and cIV but not cII, probably reflecting that cII is the only complex that is not encoded by mtDNA.
Here, we find that mtDNA damage results in impaired cellular mitochondrial function. However, the specific function of purified mitochondria is apparently normal in ogg1−/− cells, and the mitochondrial dysfunctional phenotype of ogg1−/− cells is therefore due to lower mitochondrial content rather than dysfunctional mitochondria. Hence, cellular mitochondrial dysfunction cannot be identified unless mitochondrial parameters are correlated to nuclear markers. The dispensability of mtDNA integrity for ETC capacity was described by Stuart et al.  supporting the conclusions in this work. Despite the apparent synergy between increased respiration and ETC capacity, it can be deduced that the ETC capacity exceeds that of respiration capacity in our in vitro differentiation model. Although respiration capacity reached 200 pmol O2 (400 pmol O molecules) per second per mill cell (Fig. 1E), the cI activity corresponds to about 5 nmol dichloroindophenol per second per milligram (equivalent of 5 nmol NADH per second per milligram) extracted mitochondrial protein (Fig. 3B). Typically, we routinely obtain 0.2 mg mitochondrial protein from one mill cells, which imply that there is an excess of ETC capacity over respiration capacity. If we take into account the low yield of mitochondria for ETC assessment (5%–10%; data not shown), the exact difference may be as much as hundred folds. We conclude that ETC is present in strong excess with respect to mitochondrial capacity and respiration capacity is limited by production of reducing equivalents, such as NADH, even though both respiration capacity and ETC levels are dependent on mtDNA integrity. The apparent redundancy of ETC indicates that the in vivo ETC activity is below 10%, or alternatively, that the majority of ETC complexes reside in an inactive state.
Hildrestrand et al. previously showed that nuclear OGG1 was strongly upregulated in NSCs compared with differentiated cells . It is still not known to what extent mitochondrial repair is strengthened in NSCs, however, in view of mtDNA integrity in the ogg1−/− cells, it is unlikely that increased repair capacity in NSCs is the explanation for the low-damage frequency. The kinetics of mtDNA damage formation in ogg1−/− cells suggest that it is formed along with the increased respiration activity (compare Fig. 1E and Fig. 2A), which may be accelerated in the high-oxygen pressure in this in vitro differentiation model.
The mtDNA copy number at D3 was threefold compared with D0, close to the increase in mtRNA. The OGG1 deficiency resulted in reduced levels of mtDNA and mtRNA (between 15% and 30%). By the qPCR method, the relative amplification of 10-kb product from ogg1−/− mtDNA was approximately 75% relative to wt mtDNA, which corresponds to a calculated lesion frequency of 0.25 (25%), assuming Poisson distribution. Thus, the specific reductions of mtDNA and respiration capacity in ogg1−/− cells correlate with the calculated numbers of lesions, whereas mtRNA levels are reduced to a lesser extent. mtDNA replication and transcription require the mitochondrial transcription factor TFAM, which binds to damaged mtDNA . This binding potentially interferes with replication and transcription and could also be important for repair of mtDNA, that is, a complex between TFAM and damaged mtDNA includes p53, which stimulates mitochondrial repair [37, 38]. As inferred from the specific increase in aerobic metabolism over glycolytic activity and the limited NADH production discussed earlier, our data also imply that cellular NADH production is influenced by mtDNA integrity. Relevant to this are the observations that TFAM resides in the vicinity of important metabolic proteins in the mitochondrial nucleoid . It may be speculated that the integrity of mtDNA repair secondarily facilitates regulation of nucleoid proteins harboring metabolic function, which could influence on NADH production.
In summary, our in vitro differentiation experiments demonstrate that functional maturation of mitochondria during differentiation of NSCs is dependent on the integrity of mtDNA. Importantly, mtDNA in nondifferentiated cells were unaffected by repair deficiency, which is motivating for future therapeutic considerations as amplification of NSCs in in vitro increases the risk of oxidative stress. The apparent robustness of NSCs is supported by a recent report demonstrating that NSCs isolated from older animals retained normal ability to differentiate into neurons and astrocytes in a defined differentiation setup . On the other hand, neurodegenerative disorders and an aging brain might provide elevated ROS environments that induce mtDNA damage during differentiation in vivo, resulting in dysfunctional mitochondrial maturation, and inefficient replacement of brain cells. In this scenario, one possible strategy might be to limit ROS production during recovery periods sufficiently long to promote normal mitochondrial maturation.
We are indebted to Rajikala Suganthan for the isolation of primary neurospheres, to Dr. David Kunke for technical assistance, and to Dr. Adam Robertson for critical reading of the article. This work was supported by grants from Norwegian Research Council (NevroNor and Norwegian stem cell program).
DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
The authors indicate no potential conflicts of interest.