Ginkgo Biloba Extract Ameliorates Oxidative Phosphorylation Performance and Rescues Aβ-Induced Failure

Background Energy deficiency and mitochondrial failure have been recognized as a prominent, early event in Alzheimer's disease (AD). Recently, we demonstrated that chronic exposure to amyloid-beta (Aβ) in human neuroblastoma cells over-expressing human wild-type amyloid precursor protein (APP) resulted in (i) activity changes of complexes III and IV of the oxidative phosphorylation system (OXPHOS) and in (ii) a drop of ATP levels which may finally instigate loss of synapses and neuronal cell death in AD. Therefore, the aim of the present study was to investigate whether standardized Ginkgo biloba extract LI 1370 (GBE) is able to rescue Aβ-induced defects in energy metabolism. Methodology/Principal Findings We used a high-resolution respiratory protocol to evaluate OXPHOS respiratory capacity under physiological condition in control (stably transfected with the empty vector) and APP cells after treatment with GBE. In addition, oxygen consumption of isolated mitochondria, activities of mitochondrial respiratory enzymes, ATP and reactive oxygen species (ROS) levels as well as mitochondrial membrane mass and mitochondrial DNA content were determined. We observed a general antioxidant effect of GBE leading to an increase of the coupling state of mitochondria as well as energy homeostasis and a reduction of ROS levels in control cells and in APP cells. GBE effect on OXPHOS was even preserved in mitochondria after isolation from treated cells. Moreover, these functional data were paralleled by an up-regulation of mitochondrial DNA. Improvement of the OXPHOS efficiency was stronger in APP cells than in control cells. In APP cells, the GBE-induced amelioration of oxygen consumption most likely arose from the modulation and respective normalization of the Aβ-induced disturbance in the activity of mitochondrial complexes III and IV restoring impaired ATP levels possibly through decreasing Aβ and oxidative stress level. Conclusions/Significance Although the underlying molecular mechanisms of the mode of action of GBE remain to be determined, our study clearly highlights the beneficial effect of GBE on the cellular OXPHOS performance and restoration of Aβ-induced mitochondrial dysfunction.


Introduction
Standardized Ginkgo biloba extract (GBE) derived from dried leaves of Ginkgo tree is a valuable therapeutic drug for the treatment of memory impairment and dementia including Alzheimer's disease (AD). A number of clinical studies repeatedly showed the improvement of cognitive symptoms in the elderly and in AD patients [1][2][3][4][5] but negative data have also been published. The discussion about the benefits of GBE for different indications has been revitalized after the publication of two major trials: (i) the study by McCarney and colleagues [6] demonstrating no evidence of effectiveness of GBE in mild to moderate dementia and (ii) the GEM study by DeKosky and colleagues [7] showing no favourable effect of GBE for the prevention of dementia onset in older people without or with only mild cognitive impairment. GBE exhibits a complex mode of action. Notably, multiple effects on mitochondrial function and on the apoptotic pathway that seems to be crucial for its beneficial effects in AD were reported: stabilization of mitochondrial membrane potential, improvement of energy metabolism, upregulation of anti-apoptotic Bcl-2 protein and down-regulation of pro-apoptotic Bax protein, inhibition of cytochrome c release, reduction of caspase 9 and caspase 3 activity after oxidative stress and reduction of apoptotic cell death [8][9][10][11][12][13][14]. However, evidence how GBE influences the mitochondrial oxidative phosphorylation system (OXPHOS) in neuronal cells is lacking.
AD is characterized by amyloid-beta (Ab)-containing plaques, neurofibrillary tangles, as well as synapse and neuron loss. Ab which represents with tau the main neuropathological hallmarks of AD is supposed to play a pivotal role in the pathogenesis of the disease. Within the Ab toxicity cascade, mitochondrial dysfunction and energy metabolism deficiencies have been recognized as earliest events [10,15] and have been correlated with impairments of cognitive abilities in AD clinical scenario [16]. The most consistent defect in mitochondrial electron transport system (ETS) enzymes in AD is the deficiency in cytochrome c oxidase (complex IV) activity in post-mortem brain tissues of AD patients [17][18][19] and APP transgenic mice [20], as well as in other tissues, such as platelets from AD patients and AD cybrid cells [21,22]. Moreover, abnormal mitochondrial dynamics may play a role in mitochondrial dysfunction in AD [23,24]. Interestingly, few years ago, the target of interest of Ab toxic species switched from its extracellular, fibrillar form to its soluble, oligomeric form [23,25] emphasising the important early role of mitochondria in AD pathogenic pathways [10,24,26,27].
In line with these findings, we recently reported that chronic exposure to Ab in human neuroblastoma cells (SH-SY5Y) overexpressing human wild-type APP (APP cells) resulted in an impairment of the respiratory capacity of OXPHOS and a drop in ATP generation by complex V which in turn may initiate cell death pathway [28].
Therefore, the aim of the present study was to investigate the potential effect of standardized GBE LI 1370 on Ab-induced mitochondrial dysfunction in APP cells. In addition to the measurement of oxygen consumption in whole cells as well as of isolated mitochondria, we examined activities of mitochondrial enzymes assembling the ETS. Based on previous studies [11,29,30] we decided to focus on the more important mitochondrial respiratory enzymes implicated in AD and aging, that are the complexes I, III and IV. Furthermore, we determined ATP and reactive oxygen species (ROS) levels as well as the mitochondrial membrane mass. Finally, we studied the mitochondrial DNA/nuclear DNA ratio using real-time PCR to determine changes at the genetic level.

GBE did not modify cells morphology but decreased oxidative stress
As described in our previous study [28], transfection of SH-SY5Y cells with cDNAs (pCEP4 vector) containing the vector alone or the entire coding region of human APP (APP695) did not significantly change the general morphological aspect of cells (Fig. 1A). Similarly, vector control as well as APP cells conserved their neuroblast-like morphology with differentiated perikaria and short neurites after treatment with GBE (Fig. 1B). APP cells showed a slight trend to increased ROS levels compared to controls cells (Fig. 1C). Mitochondria-associated ROS levels decreased significantly in both cell lines after treatment with GBE (100 mg/ml) for 24 h (Fig. 1D).

GBE ameliorated OXPHOS capacity and restored Abinduced deficits
To investigate the protective effect of GBE against Ab toxicity at the mitochondrial level, we used a high-resolution respiratory protocol established lately by our group [28]. Thus, physiological substrate combinations (pyruvate, glutamate and malate as substrates for complex I and succinate for complex II) were used to obtain mitochondrial bioenergetics approaching the most of physiological states in whole cells ( Fig. 2A). We compared OXPHOS that is the whole ETS composed of the four mitochondrial enzymes (complex I to complex IV) and the F 1 F 0 ATP synthase of vital control and APP cells after pretreatment with GBE for 24 h. As already previously observed [28], APP cells exhibited a significant impairment of OXPHOS capacity (Fig. 2B). Of note, GBE was able to significantly ameliorate the global failure of mitochondrial respiration in APP cells and increased oxygen consumption in control cells as well (Fig. 2B). Importantly, GBE-treated control and APP cells presented comparable mitochondrial respiratory rate validating the protective role of GBE on stabilization and normalization of mitochondrial capacity, respectively (Fig. 2B).
To analyse the impact of GBE on metabolic states of mitochondrial respiration, two flux control ratios have been evaluated. First, the respiratory control ratio (RCR) which is an indicator of the coupling state of mitochondria was determined. State 3 is the rate of phosphorylating respiration in the presence of exogenous ADP, when mitochondria are actively making ATP, whereas state 4 is the rate of resting respiration, when all ADP has been consumed. State 4 is associated with proton leakage across the inner mitochondrial membrane, when mitochondria exhibit basal activity, i.e. they are respiring but not making ATP. Therefore, RCR represents the ADP-activated flux to measure coupled OXPHOS capacity (state3) divided by leak flux (state4). Whereas increased Ab levels led to a decrease of RCR in APP cells, treatment with GBE significantly increased the ratio in both cell types (Fig. 3A). Secondly, ROX/ETS yields an index of the magnitude of residual oxygen consumption relative to maximum oxygen consumption capacity. This ratio was decreased in both cell types after treatment with GBE with the strongest effect in APP cells (Fig. 3B). Both flux control ratio changes indicate an increase of the coupling state of mitochondria leading to a better efficiency of OXPHOS. Consistent with this result, we observed a rise of ATP levels which are closely correlated with a higher complex V activity, the final OXPHOS enzyme, in GBE-treated control and APP cells (Fig. 3C).
Taken together, respiratory analyses showed that GBE enhanced metabolic pathways by increasing the coupling state of OXPHOS promoting finally a rise of ATP synthesis in both cell types but with the strongest effects on APP cells. As a consequence, GBE-treated control and APP cells presented comparable mitochondrial respiratory system capacity suggesting a complete restoration of Ab-induced deficits in energy metabolism.

GBE improved oxygen consumption in isolated mitochondria from APP cells and led to an up-regulation of mitochondrial DNA
To confirm the improvement of mitochondrial performance observed in whole cells, we investigated oxygen consumption of mitochondria that were isolated from control and APP cells after 24 h treatment with GBE. Consistent with our results on whole cells, isolated mitochondria from APP cells still exhibited a significant impairment of OXPHOS capacity and reduced RCR compared to mitochondria isolated from control cells ( Fig. 4A and B). Again, GBE-induced enhancement of oxygen consumption was significantly present in both cell lines ( Fig. 4A) with the strongest improvement in APP cells ( Fig. 4A and B) suggesting a regulatory effect of GBE at the mitochondrial level even after its removal. Of note, this functional enhancement can be associated with an increase of the mitochondrial DNA/nuclear DNA ratio after treatment with GBE ( Fig. 4C) but without a change of the mitochondrial membrane mass (Fig. 4D).

GBE modulated the activity of respiratory complexes
To improve the mitochondrial respiratory capacity, GBE may act on one or several mitochondrial enzymes. To study this hypothesis, activities of individual respiratory complexes have been investigated. Complex I activity significantly increased after GBE treatment selectively in APP cells (Fig. 5A), whereas complex II activity was unaffected by GBE in both cell lines (data not shown). Complex III activity which was markedly increased in APP cells was normalized to the level of control cells after treatment with GBE (Fig. 5B). Complex IV activity which was decreased in APP cells, significantly increased in control and APP cells after treatment with GBE (Fig. 5C).

GBE partially decreased Ab secretion of APP cells
To delineate the relationship between Ab level and GBE, we further examined the levels of secreted Ab1-40 in control and APP cells. Figure 6 shows that APP cells secreted Ab to a 3-4 fold higher extent than control cells bearing the empty vector. Of note, treatment with GBE significantly decreased medium Ab levels in APP cells of about 20% (medium Ab1-40, APP cells: 100%64.8%, APP cells + GBE: 83.22%62.6%, control cells: 22.03 60.4%, control cells + GBE: 21.06%60.5%).

Discussion
In this study, we present for the first time clear evidence that under physiological conditions GBE improved metabolic energy pathways by increasing the coupling state of mitochondria. Importantly, the comparison of the mitochondrial energetic capacity in both cell types after treatment with GBE indicates a recovery of the disturbed bioenergetic homeostasis found in APP cells. Notably, GBE-enhancing effect on OXPHOS was still present in mitochondria after their removal from cells suggesting possible continuing, regulatory actions of GBE at the mitochondrial level. These results corroborate findings showing an increase of the coupling state of isolated mitochondria by GBE in ischemia models [31] and findings demonstrating that GBE ameliorated oxygen consumption of mitochondria from rat heart mainly driven by effects of GBE on complex I and III [29]. The capacity of mitochondria to re-phosphorylate ADP in state 3 is positively dependent on the degree of coupling. In accordance with this assumption, we also observed an enhancement of ATP levels in  both cell types after treatment with GBE. Moreover, the respiration rate in state 3 is determined by the activity of mitochondrial enzymes of the whole ETS and ATP-synthase [29,30]. Thus, the global improvement of mitochondria functionality may be explained by the action of GBE on one or several mitochondrial enzymes. The increased activity of complex III in APP cells that can be interpreted as a compensatory mechanism to rescue Ab-induced mitochondrial defects [29] normalized after treatment with GBE. This effect can be taken as a back to a physiological functionality, i.e. the over-activity of complex III in untreated APP cells was down-regulated in the presence of GBE which in turn up-regulated complex IV activity and ATP synthesis, both markedly reduced in our AD cell model. Surprisingly, complex I activity was increased only in APP cells after treatment with GBE while it was unchanged in control cells. APP cells did not exhibit an Ab-induced defect in complex I activity compared to control cells [29]. GBE by increasing the coupling mitochondrial state may improve complex I activity. Complex I as a main ROS producer may also be a key element in GBE efficacy. However, this can not explain why the GBE effect on complex I was only found in APP cells. One can speculate that a damaged ETS with changes in downstream complexes, such as complex III and IV in APP cells, might force the emergence of this effect. Our study is further in line with recent findings studying the effect of GBE after in vivo treatment on mitochondrial membrane potential in dissociated brain cells from aged mice after inhibiting OXPHOS using specific inhibitors of complexes [11]. Here, GBE showed protective effects on mitochondrial membrane potential at the complexes I, IV and V [11]. Even if the exact molecular mechanisms by which GBE stabilizes mitochondrial enzymes is not yet clear, it can be speculated that besides other regulatory effects on mitochondrial components: (i) GBE may act as an antioxidant scavenging superoxide anion generated by electron leakage which occurs mainly at complexes I and III. (ii) GBE may also scavenge nitric oxide leading to an increase of complex IV activity which is the main target of NO [32]. Consequently, the protection of mitochondrial complexes by GBE may not only lead to an increase of mitochondrial performance (coupling state) but also to a decrease of ROS production. Moreover, some authors proposed that GBE provides anti-apoptotic effects [8,9,33]. Thus, GBE significantly attenuated mitochondria-initiated apoptosis and decreased the activity of caspase 3, a key enzyme in the apoptosis cell signalling cascade, in neuroblastoma cells stably expressing an AD-associated double mutation in APP [12]. In addition, GBE inhibited formation of Ab-fibrils [12] either by a direct interaction with Ab [34,35] or by modulation of its precursor APP [36] and GBE flavonols were able to reduce Ab levels in vitro and in the brain from transgenic Alzheimer mice [37]. We also observed that GBE treatment partially decreased Ab generation in APP cells. Therefore, it might be possible that GBE restored impaired OXPHOS pathway through decreasing Ab level, since Ab interacts with mitochondria [27,29,31,38,39] and Ab levels determine the extent of mitochondrial dysfunction [38]. Considerable evidence exists that Ab can increase the production of ROS [40]. Consistent, we found a slight increase in ROS levels in APP cells compared to control cells suggesting that cellular antioxidant defences are still able to counterbalance excessive ROS production at least in part in our cell culture model.
Of note, we observed an up-regulation of mitochondrial DNA in cells treated with GBE, suggesting that GBE may act not only at a functional level of mitochondria but also at their genetic one. The number of mitochondrial DNA copies per mitochondrion as well as mitochondria per cell is exquisitely calibrated on cellular energy demand. Mitochondrial DNA (mtDNA) encodes important subunits of catalytic cores of mitochondrial respiratory complexes constituting OXPHOS. Consequently, in our study the increase of mtDNA associated with the increase of OXPHOS performance (coupling state) may facilitate the assembly of mitochondrial complexes into supercomplexes to optimize the respiratory activity. This idea is in agreement with the results of D'Aurelio and coworkers who revealed a correlation of improved mitochondrial respiration with the presence of supercomplexes [41]. Additionally, Tendi and colleagues [33] showed that GBE upregulated mitochondrial gene expression of a mtDNA encoded subunit of complex I (ND1) in PC12 cells which may perfectly correspond to the increase of complex I activity detected in our GBE-treated APP cells. Moreover, the transcript level of the antiapoptotic Bcl-2-like protein was found to be elevated, whereas the transcript level of the pro-apoptotic caspase 12 was decreased in PC12 cells after treatment with GBE [9,12]. Although further studies are necessary, we can speculate that a GBE treatment may not only enhance mtDNA as an early event after acute treatment, but may instigate mitochondria biogenesis as a long-term effect after chronic exposure [42]. In line with this idea, recent studies showed that GBE induced hippocampal neurogenesis, a very energy demanding process, in young and old transgenic AD mouse model (TgAPP/PS1) after receiving a diet supplemented with GBE for 1 month [43].
Oxidative stress and especially mitochondria-derived oxidative stress is thought to have an important role in the pathogenesis of AD, a disease that likely begins years, if not decades, before clinical onset of dementia. In line with this hypothesis, experimental data support the notion that antioxidants may prevent or cure pathological conditions of AD. With regard to beneficial effects of antioxidants on OXPHOS performance, only few data are available for other antioxidants than GBE. Thus, one study provided evidence that vitamin E which is considered the major lipophilic antioxidant in the human body has GBE-like beneficial effects at high doses (corresponding to a human daily intake of 1,300 mg vitamin E/day) on age-dependent decline of OXPHOS and enzyme activities I as well as IV in aging male mice brain [44]. But concerns have been raised about the long-term safety of vitamin E supplementation, particularly in such high doses, and no evidence is provided demonstrating similar effects of vitamin E on OXPHOS at lower and rather physiologically relevant doses. MitoQ10 (mitochondria-targeted coenzyme Q) has also antioxidant properties but only when electron flow within the respiratory chain or complex I proton pumping is already retarded [45], while, in contrast, MitoQ10 action in the intact respiratory chain led to increased ROS production [45]. This is also contrary to our results of GBE on OXPHOS performance and ROS levels under non-pathological conditions found in control cells. Several meaningful differences distinguish the mode of action of GBE from those other antioxidants. Due to different constituents in the extract such as flavonols, bilobalide, and ginkgolides, GBE seems to exhibit a unique spectrum of action probably via complementary effects of its constituents [11] on mitochondrial function including (i) direct radical scavenging properties, (ii) direct stabilization of mitochondria, (iii) indirect protection of mitochondria, and (iiii) up-regulation of mtDNA, leading to improved OXPHOS performance and energy homeostasis. Whereas a huge body of evidence from preclinical research support the notion that antioxidants protect against neurodegeneration, previous observational studies of antioxidants, including vitamin E and GBE, and dementia risk have yielded inconsistent results based on short-or long-term follow-up periods [7,46] raising the question of the  ''window of opportunity'' for successful prevention. Further studies should continue to assess points at which treatment might be most relevant to dementia risk [47].
In total, our findings and findings of others indicate substantial mitochondria modulating properties of GBE. We could clearly show that GBE improved OXPHOS performance and was able to restore Ab-induced mitochondria failure. Hereby, the increase in mitochondrial content might represent an important early event. Taking into account the increasing interest in mitochondrial stabilization as intervention strategy in AD [48], the regulating mechanisms of GBE on mitochondria function suggested by this study qualifies this drug as therapeutic candidate.

Cell culture and treatment with GBE
In the present study, human neuroblastoma SH-SY5Y cells stably expressing vector alone (pCEP4, control cells) or the entire coding region of human wild-type APP (APP695, APP cells) were used as described previously [49]. APP cells secreted Ab levels within pg/mL range (around 150 pg/mL and 35 pg/mg protein respectively Ab1-40 compared to approximately 50 pg/mL and 10 pg/mg protein respectively Ab1-40 secreted by vector control cells) [28]. Stably transfected cell clones were selected with hygromycin [49]. Cells were grown at 37uC in DMEM medium supplemented with 10% calf serum, 2 mM L-glutamine, and 0.3 mg/mL hygromycin. Cells were pre-treated with standardized Ginkgo biloba extract LI 1370 (Vifor SA, Switzerland) for 24 h before analyzing the cells. A dose response for GBE was tested by treating the cells with different GBE concentrations starting from 0.001 mg/ml up to 0.5 mg/ml in distinct assays. The optimum concentration of GBE yielding maximum beneficial effects (with regard to reduction of ROS levels, enhancement of RCR, and increase in ATP levels) in APP cells was 100 mg/ml (see figures S1, S2, and S3, respectively), but significant beneficial effects could be detected at concentration as low as 10 mg/ml (see ATP results, figure S3) and 50 mg/ml (see ROS and ATP results, figures S1 and S3). A similar dose response was found for control cells (see figures S1, S2, and S3, respectively). Based on these findings and in agreement with other in vitro studies [11,50], a standard concentration of 100 mg/ml of GBE was used in the following assays. In further pre-experiments, we excluded effects of the vehicle.

Phase contrast microscopy and morphological analysis
For the morphological analysis, cells were seeded at a density of 4610 4 cells/mL on coverslips previously coated with 0.05 mg/mL collagen. Phase contrast pictures were taken from living neuroblastoma cells using a Zeiss Axiolab microscope with a 406/1,2 W Korr objective equipped with a digital camera Zeiss AxioCam MRc.

Preparation of isolated mitochondria for the determination of the activities of the single complexes I-IV
Cells were incubated for 15 min in an ice-cold lysis buffer (75 mM NaCl, 1 mM NaH 2 PO 4 , 8 mM Na 2 HPO 4 , 250 mM sucrose, 1 mM Pefabloc, 0.05% digitonine, complete protease inhibitor mixture tabletsH (Roche Diagnostics)). Then, cells were homogenized with a glass homogenizer (10 strokes at 400 rpm and 5 strokes at 700 rpm), and the resulting homogenate was centrifuged at 800 g for 10 min at 4uC to remove nuclei and tissue particles. The supernatant 1 (S1) was saved and the pellet resuspended in the lysis buffer. The homogenization step as well as the low-speed centrifugation step was repeated. The supernatant 2 (S2) was saved and added to S1. Combined mitochondriaenriched supernatants (S1+S2) were centrifuged at 20,000 g for 15 min at 4uC to obtain the mitochondrial fraction. The mitochondrial pellet was resuspended in PBS and stored at 4uC until use, followed by determination of protein content [51].

Complex I activity
A total of 240 mg of isolated mitochondria was solubilized in n-dodecyl b-D-maltoside (20%). NADH-ubiquinone oxidoreductase activity was measured at 30uC in a buffer containing 2 mM Na + /MOPS, 50 mM NaCl, and 2 mM KCN, pH 7.2, using 100 mM n-decylubiquinone (DBQ) and 100 mM NADH as substrates and 5 mM rotenone as inhibitor [20,28,52,53]. Oxidation rate of NADH were recorded with a Shimadzu Multi-Spec-1501 diode array spectrophotometer (e 340-400 nm = 6.1 mM 21 .cm 21 ). Complex I activity was normalized to citrate synthase activity to take into account variations in the amount of mitochondrial and non-mitochondrial protein contamination between the samples [28,54] and was given as CI/CS ratio.

Complex III activity
The oxidation of 50 mM decylubiquinol obtained by complex III was determined using cytochrome c as an electron acceptor as described previously [28,55]. Briefly, decylubiquinol was prepared by dissolving decylubiquinone (10 mM) in ethanol acidified to pH 2. The quinone was reduced with excess solid sodium borohydride. Decylubiquinol was extracted into diethylether: cyclohexane (2:1, v/v) and evaporated to dryness under nitrogen gas, dissolved in ethanol acidified to pH 2. The assay was carried out in a medium containing 35 mM KH 2 PO 4 , 5 mM MgCl 2 , 2 mM KCN (pH 7.2), supplemented with 2.5 mg/mL BSA, 15 mM cytochrome c, 0,6 mM n-dodecyl b-D-maltoside and 5 mg/mL rotenone. The reaction was started with 10 mg of mitochondrial protein and the enzyme activity was measured at 550 nm. The extinction coefficient used for cytochrome c was 18.5 mM 21 .cm 21 . Complex III activity was normalized to citrate synthase activity to take into account variations in the amount of mitochondrial and non-mitochondrial protein contamination between the samples [28,54] and was given as CIII/CS ratio.

Complex IV activity
Cytochrome c oxidase activity was determined in intact isolated mitochondria (100 mg) using the Cytochrome c Oxidase Assay Kit (Sigma-Aldrich GmbH, Buchs, Switzerland). The colorimetric assay was based on the observation that a decrease in absorbance at 550 nm of ferrocytochrome c was caused by its oxidation to ferricytochrome c by cytochrome c oxidase. The Cytochrome c Oxidase Assay was performed as described previously [28,53,56]. Complex IV activity was normalized to citrate synthase activity to take into account variations in the amount of mitochondrial and non-mitochondrial protein contamination between the samples [28,54] and was given as CIV/CS ratio.

Reactive oxygen species (ROS) levels
Cells were plated the day before at a density of 1610 5 cells/well in a 48 well plate. The mitochondria-associated ROS levels were determined using the fluorescent dye dihydrorhodamine (DHR) at a concentration of 10 mM. The cells were loaded for 15 min with the dye. After washing twice with HBSS, the fluorescence was determined with the Victor2 reader multiplate (PerkinElmer Life Sciences) at 485 nm (excitation)/535 nm (emission).

Oxygen consumption of vital cells
Mitochondrial oxygen consumption was measured at 37uC using an Oroboros Oxygraph-2k system. In contrast to the respiratory protocol measuring oxygen consumption of isolated mitochondria in a rather artificial experimental environment, this protocol enables the determination of oxygen consumption in whole, vital cells under the most physiological conditions. Five millions of cells were added to 2 mL of a mitochondrial respiration medium containing 0.5 mM EGTA, 3 mM MgCl 2 , 60 mM K-lactobionate, 20 mM taurine, 10 mM KH 2 P0 4 , 20 mM HEPES, 110 mM sucrose, 1g/l BSA (pH 7.1). To measure state 4 ( = state 2) of complex I, 5 mM pyruvate and 2 mM malate were added and cells were permeabilised with 15 mg/mL digitonin. Afterwards, 2 mM ADP was added to measure state 3, and in order to increase the respiratory capacity, 10 mM glutamate was added. To study the effect of convergent complex I+II electron input on respiration, 10 mM of succinate was added [50]. The integrity of mitochondrial membrane was checked through the addition of 10 mM cytochrome c. After determining coupled respiration, 0.4 mM FCCP (Carbonyl cyanide p-(trifluoromethoxy) phenyl-hydrazone) was added and respiration was measured in the absence of a proton gradient. In order to inhibit complex I and III activities 0.5 mM rotenone and 2.5 mM antimycine A, respectively were added. Then, 2 mM ascorbate and 0.5 mM TMPD were added to have access to complex IV activity. Finally, 100 mM sodium azide was added to inhibit the mitochondrial respiration. Control and APP cells untreated or treated with GBE were measured in parallel pairs using the same conditions (crossover design).

Oxygen consumption of isolated mitochondria
The respiratory protocol used for isolated mitochondria was the same than this for whole cells except for the following points: (i) isolated mitochondria (0.5 mg) were added to 2 mL of a mitochondrial respiration medium containing 65 mM sucrose, 10 mM potassium phosphate, 10 mM Tris-HCl, 10 mM MgSO 4 , and 2 mM EDTA (pH 7.0), (ii) digitonin was not added and (iii) 0.05 mM of FCCP was sufficient to obtain the maximal mitochondrial respiration [53].

Determination of ATP levels
Cells were plated one day before at a density of 2.5610 4 cells/ well in a white 96 well plate. The assay kit is based on the bioluminescent measurement of ATP (ViaLightTM HT, Cambrex Bio Science, Lonza, Rockland, USA). The bioluminescent method utilizes the enzyme luciferase, which catalyses the formation of light from ATP and luciferin. The emitted light was linearly related to ATP concentration and was measured using a luminometer [48,53].

Estimation of mitochondrial membrane mass
Cells were plated the day before at a density of 1610 5 cells/well in a 48 well plate. Mitochondrial membrane mass was measured using the cell-permeable mitochondria-selective dye MitoTracker Green FM (Molecular Probes, Invitrogen, Lucerne, Switzerland) (400 nM, 15 min). This probe accumulates in active mitochondria independently of mitochondrial membrane potential and then reacts with accessible thiol groups of proteins and peptides [57]. Fluorescence was determined using a Fluoroskan Ascent FL multiplate reader (Labsystems) at 490 nm (excitation) /516 nm (emission).
Quantitative real-time PCR for determination of mitochondrial DNA/nuclear DNA ratio Samples consisted of confluent grown SH-SY5Y cells (APP or vector, untreated or treated for 24 h with GBE 100 mg/ml). Total DNA was extracted with the Qiagen FlexiGen Kit (Qiagen, Basel, Switzerland) according to the manufacturer's protocol. To determine the mitochondrial DNA (mtDNA) content in relation to nuclear DNA (nDNA) for each sample [58,59,60] a standard curve was generated via extracted DNA from pooled SH-SY5Y cell populations. For this purpose, mitochondrial and genomic genes were amplified using the same primers that were used for the real-time PCR (see below) and templates were purified with the Qiagen QIAquick PCR Purification Kit (Qiagen, Basel, Switzerland). The absorption at 260 nm was measured to determine the gene copy number and produce serial standard dilutions.
Primers for the RT Q-PCR analysis of mitochondrial DNA (mtDNA)-tRNAleu [61] were mtF3212 (59CACCAAGAACAG GGTTTGT39) and mtR3319 (59TGGCCATGGGTATGTT GTTAA39) [62], those for nuclear DNA (nDNA), human polymerase c accessory subunit gene (ASPOLG) were ASPG3F 59GAGCTGTTGACGGAAAGGAG39 and ASPG4R (59CAG AAGAGAATCCCGGCTAAG39) [63]. For each DNA extract, the nDNA and mtDNA gene respectively were quantified separately in triplicates and/or duplicates by real-time quantitative PCR in the presence of SYBR-green with the use of the Applied Biosystems StepOne TM cycler (Applied-Biosystems). PCR reactions contained 1x Power SYBR Green Master Mix (ABI P/N 4367659, Rotkreuz, Switzerland), 1 mM of each primer and 1 ng of DNA extract. The PCR amplification consisted of a single denaturation-enzyme-activation step of 10 min at 95uC, followed by 40cycles of 15 sec denaturation at 95uC and 60 sec of annealing/extension at 60uC. Fluorescence was measured at the end of each extension step. A standard curve of corresponding mtDNA and nDNA template equivalents with defined copy numbers were included in each run to quantify the content of mtDNA and nDNA in each sample. Data thus obtained were analyzed by using the cycle threshold (C T ) of each amplification reaction relating it to its respective standard curve. Results from the quantitative PCR were expressed as the ratio of the mean mitochondrial DNA to the mean nuclear DNA content.

Detection of Ab levels
Detection of Ab1-40 secretion was performed by using sandwich enzyme-linked immunosorbent assay (Invitrogen, Basel, Switzerland) according to the manufacturer's protocol. Secreted Ab1-40 levels within the medium (Ab1-40 pg/ml) were normalized to mg protein/ml. Values of intracellular Ab1-40 as well as medium and intracellular Ab1-42 levels were below detection limit and are therefore not shown.

Statistical Analysis
Data were presented as mean 6 S.E.M. For statistical comparisons, unpaired and paired Student's t-test, respectively, or Two-way ANOVA was used. P values less than 0.05 were considered statistically significant. Figure S1 GBE decreased reactive oxygen species (ROS) levels in a dose response manner. Mitochondria-associated ROS levels measured after incubation with DHR (DHR fluorescence units/ 1610 5 cells normalized to the respective untreated control and APP cells). A) Control cells showed significantly decreased ROS levels after treatment with GBE (0.001-0.5 mg/ml; 24 h) significantly for concentrations between 0.05-0.5 mg/ml. B) GBE-treated APP cells (0.001-0.5 mg/ml; 24 h) exhibited significantly reduced ROS levels for the same concentration range 0.05-0.5 mg/ml compared to control cells. Values represent the means 6 S.E., GBE treatment effect, paired student9s t-test, number of pairs n = 11: #, p,0.05, ##, p,0.01; GBE treated versus corresponding untreated control and APP cells.