The Protective Effect of Ubiquinone against the Amyloid Peptide in Endothelial Cells Is Isoprenoid Chain Length-Dependent

Vascular brain pathology constitutes a common feature in neurodegenerative diseases that could underlie their development. Indeed, vascular dysfunction acts synergistically with neurodegenerative changes to exacerbate the cognitive impairment found in Alzheimer’s disease. Different injuries such as hypertension, high glucose, atherosclerosis associated with oxidized low-density lipoprotein or inflammation induce NADPH oxidase activation, overproduction of reactive oxygen species, and apoptosis in endothelial cells. Since it has been shown that pretreatment of cultured endothelial cells with the lipophilic antioxidant coenzyme Q10 (CoQ10) displays a protective effect against the deleterious injuries caused by different agents, this study explores the cytoprotective role of different CoQs homologues against Aβ25–35-induced damage and demonstrates that only pretreatment with CoQ10 protects endothelial brain cells from Aβ25–35-induced damage. Herein, we show that CoQ10 constitutes the most effective ubiquinone in preventing NADPH oxidase activity and reducing both reactive oxygen species generation and the increase in free cytosolic Ca2+ induced by Aβ25–35, ultimately preventing apoptosis and necrosis. The specific cytoprotective effect of CoQ with a side chain of 10 isoprenoid units could be explained by the fact that CoQ10 is the only ubiquinone that significantly reduces the entry of Aβ25–35 into the mitochondria.


Introduction
Alzheimer's disease (AD) is characterized by the presence of senile plaques of betaamyloid and neurofibrillary tangles of hyperphosphorylated tau protein. The amyloid cascade hypothesis synthesizes the histopathological and genetic information and constitutes the central hypothesis that explains the onset and progression of the disease [1]. However, most patients who have received a diagnosis of sporadic AD also exhibit cerebrovascular disease [2]. Indeed, vascular brain pathology is associated with the rate of cognitive decline [3]. Longitudinal studies have suggested that high blood pressure in midlife is the most substantial vascular risk factor for both AD and vascular dementia (VaD) in later life [4], and it constitutes a major cause for the dysregulation of sustained blood flow [2], which increases the expression of hypoxia-inducible factor 1-α leading to upregulation of the beta-secretase (BACE)1 [5,6]. The generation of oxidative stress in hypertension is well defined and remains a crucial modulator of endothelial function and nitric oxide (NO) bioavailability [7]. In addition to hypertension, other risk factors for AD that increase oxidative stress are high glucose, atherosclerosis associated with oxidized low-density lipoprotein (ox-LDL), and inflammation [2]. In endothelial cells in vitro, these injuries induce overproduction of reactive oxygen species (ROS), apoptosis, The levels of mitochondrial O 2 − , total H 2 O 2, and free cytosolic Ca 2+ were measured as reported before [17]. Briefly, bEnd.3 cells were seeded in 96-well plates (10,000 cells/well) and pre-incubated for 24 h with vehicle (EtOH; controls), CoQ2, CoQ6, CoQ9, or CoQ10 (5 µM). Then, the medium was removed, and cells were incubated with Aβ [25][26][27][28][29][30][31][32][33][34][35] (5 µM) in fresh medium for 24 h. For the determination of O 2 − and H 2 O 2, the cells were incubated for 30 min with the probes MitoSOX (1 µM) and H2DCF-DA (2.5 µM), respectively. For free cytosolic Ca 2+ quantification, the cells were incubated 15 min with the probe Fluo-4 AM (1 µM). Following PBS wash, images were acquired on live cells using a Cytation 5 multiimaging system (20× objective). The percentage of mean intensity (relative fluorescent units, RFUs) vs. the control was determined using ImageJ 1.53 (n = 3; >50 cells/well).

Determination of Cellular Energy Phenotype
To analyze the cellular energy phenotype (OCR and ECAR), we used a Seahorse XFp Analyzer (Seahorse Biosciences, Agilent, Santa Clara, CA, USA), following the protocol previously set up by Divakaruni et al. [23]. bEnd.3 cells were seeded in XFp miniplates (7500 cells/well), pretreated for 24 h with vehicle (EtOH; controls), CoQ2, CoQ6, CoQ9, or CoQ10 (5 µM) and, after that, treated with Aβ 25-35 (5 µM) for 24 h. Then, cells were incubated in Seahorse XFp base medium without phenol red (103, 193-100 Seahorse Biosciences) for 60 min at 37 • C before loading into the Seahorse analyzer. Three basal OCR and ECAR measurements were obtained within the first 20 min. Then, cells were stressed with different mitochondrial inhibitors: oligomycin (1 µM), which inhibits ATP production by the mitochondria, and carbonyl cyanide-p -trifluoromethoxyphenylhydrazone (FCCP) (0.3 µM), which depolarizes the mitochondrial membrane. To analyze the data, we used the Seahorse XFp software Wave 2.6 (Seahorse Biosciences, Agilent, Santa Clara, CA, USA). The OCR levels are represented as pmol of oxygen per minute, and ECAR levels are represented as the rate of change in mpH per minute. Data are presented as the mean ± SEM for each time point normalized to the number of cells per well (n = 3) as previously described [24].

Determination of Cellular Mitochondrial Respiration
We used the Seahorse Mito Stress test to analyze mitochondrial respiration, following the protocol previously described [23]. Briefly, bEnd.3 cells were seeded in XFp > miniplates (7500 cells/well), pretreated for 24 h with vehicle (EtOH; controls) or CoQ10 (5 µM), and treated with Aβ 25-35 (5 µM) for 24 h. Then, cells were incubated in Seahorse XFp > base medium without phenol red (103, 193-100 Seahorse Biosciences) for 60 min at 37 • C before loading into the Seahorse analyzer. Three basal OCR measurements were obtained within the first 20 min. Then, different mitochondrial inhibitors were added [oligomycin, 1 µM; FCCP, 0.3 µM; antimycin A and rotenone, 1 µM]. The inhibitors led us to measure ATPlinked respiration, maximal respiration, and non-mitochondrial respiration, respectively. After different injections, three OCR values were automatically measured by the Seahorse XFp > software Wave 2.6. The OCR levels are represented as pmol of oxygen per minute. Data are presented as the mean ± SEM for each time point normalized to the number of cells per well (n = 3) as previously described [24].

Statistical Analysis
Data are expressed as the mean ± S.E.M. Statistical analyses were carried out with GraphPad Prism 8.0.2 (San Diego, CA, USA), using Student's t-test and one-way ANOVA followed by Dunnett's multiple comparison test. Two-way ANOVA (Dunnett's multiple comparison test) was used for the real-time internalization studies. Significant differences were considered at * p < 0.05; ** p < 0.01; *** p < 0.001.
Data are expressed as the mean ± S.E.M. Statistical analyses were carried out with GraphPad Prism 8.0.2 (San Diego, CA, USA), using Student's t-test and one-way ANOVA followed by Dunnett's multiple comparison test. Two-way ANOVA (Dunnett's multiple comparison test) was used for the real-time internalization studies. Significant differences were considered at * p < 0.05; ** p < 0.01; *** p < 0.001.

CoQ10 but Not Shorter Side Chain CoQ Prevents Aβ25-35-Induced Endothelial Cell Death
It was previously reported that cellular exposure to the amyloid fragment Aβ25-35 leads to endothelial cell toxicity by inducing cell apoptosis and necrosis [15,17], and our previous study demonstrated the cytoprotective role of CoQ10 in Aβ25-35-induced endothelial cell death [17]. Thereby, we explored the possible cytoprotective role of different CoQ homologues. First, we observed that the addition of those CoQ homologues (CoQ2, CoQ6, CoQ9, and CoQ10) for 24 h did not alter basal apoptotic and necrotic indexes in bEnd.3 cells, which were 4.3% and 1.5%, respectively ( Figure 1A,B). As expected, adding 5 µM Aβ25-35 for 24 h increased the percentage of apoptotic nuclei up to 14.2% ( Figure 1A), and the number of necrotic cells increased in a greater proportion from 1.5% to 9.1% ( Figure 1B). Data are presented as the mean ± SEM. Student's t-test was used to compare the differences between groups treated with the same ubiquinone and Dunnett's multiple comparison test to compare each Aβ25-35 (plus or minus ubiquinone) treatment with the control. ns: not significant, * p < 0.05, ** p < 0.01, *** p < 0.001, (n = 3).
Pretreatment with different CoQ homologues (CoQ2, CoQ6, CoQ9) did not ameliorate the Aβ 25-35 -induced apoptosis and basal necrosis of bEnd.3 cells (p < 0.001). However, in cells pretreated with CoQ10, the percentage of Aβ 25-35 -induced apoptotic and necrotic cells was significantly lower than with other ubiquinones, and, although the number of apoptotic nuclei and necrotic cells was significantly higher than that observed in the absence of Aβ 25-35 (p < 0.05), CoQ10 pretreatment was able to prevent the Aβ 25-35 -induced increase in apoptotic cells compared to the control ( Figure 1A,B).

CoQ10 but Not Shorter Side Chain CoQ Prevents Aβ25-35-Induced Activation of NADPH Oxidase
Next, we investigated whether the different CoQ homologues could hamper Aβ25-35induced activation of NADPH oxidase. As shown in Figure 2, basal NADPH oxidase levels in bEnd.3 cells (19.4 ± 0.86 pg/mL) were not modified after incubation with the different CoQs (CoQ2, CoQ6, CoQ9, and CoQ10).

Figure 2.
Characterization of NADPH oxidase levels. bEnd.3 cells were pretreated with CoQ2, CoQ6, CoQ9, or CoQ10 for 24 h and then treated with Aβ25-35 for 24 h. NADPH oxidase was assessed using a NOX1 ELISA kit. Results are expressed as pg/10 5 cells compared with the standard. Data are presented as the mean ± SEM. Student's t-test was used to compare the differences between groups treated with the same ubiquinone and Dunnett's multiple comparison test to compare each Aβ25-35 (plus or minus ubiquinone) treatment with the control. ns: not significant, * p < 0.05, ** p < 0.01, *** p < 0.001 (n = 3).

CoQ10 Prevents the Aβ25-35-Mediated Increase in O2 − , H2O2, and Free Cytosolic Ca 2+ Levels in Endothelial Cells and Mitochondrial Calcium Depletion in Endothelial Cells
Then, ROS production was measured by fluorescence microscopy by quantifying the levels of mitochondrial superoxide anion (O2 − ) and intracellular H2O2, which were detected with the MitoSOX and H2DCF-DA probes, respectively.
As shown in Figure 3A, bEnd.3 cells treated with Aβ25-35 (5 µ M) for 24 h showed an increase in O2 − levels of 30% vs. untreated control cells. Treatment with the different CoQs in the absence of Aβ25-35 did not induce significant changes in O2 − , although CoQ6 generated a slight increase. When cells were treated with 5 µM Aβ25-35, only CoQ10 pretreatment prevented the increase in O2 − levels, showing similar levels compared to CoQ10-treated cells in the absence of Aβ25-35, but this preventive effect was not observed in cells treated with CoQ2 and CoQ9. In cells pretreated with CoQ6, although Aβ25-35 also increased O2 − levels, the difference was not significant (Student's test p < 0.015). Nonetheless, only the group > pretreated with CoQ10 showed O2 − levels with values The addition of Aβ 25-35 (5 µM) to the culture medium significantly stimulated the NADPH oxidase levels until values of 17.85 ± 0.55 pg/10 5 cells (p < 0.01) were reached and, similarly to apoptosis and necrosis, the addition of CoQ homologues with chains of 2, 6, and 9 isoprenes did not impede the Aβ 25-35 -induced activation of NADPH oxidase ( Figure 2). Only preincubation with CoQ10 protected bEnd.3 cells by abolishing the increase in Aβ 25-35 -induced NADPH oxidase levels (12.71 ± 1.48 pg/10 5 cells for CoQ10 and 13.44 ± 0.59 pg/10 5 cells for Aβ 25-35 plus CoQ10; Student s t-test and Dunnett s multiple comparison test).

CoQ10 Prevents the Aβ 25-35 -Mediated Increase in O 2 − , H 2 O 2, and Free Cytosolic Ca 2+ Levels in Endothelial Cells and Mitochondrial Calcium Depletion in Endothelial Cells
Then, ROS production was measured by fluorescence microscopy by quantifying the levels of mitochondrial superoxide anion (O 2 − ) and intracellular H 2 O 2, which were detected with the MitoSOX and H 2 DCF-DA probes, respectively.
As shown in Figure 3A Figure 3A).  Figure 3A). On the other hand, administration of 5 µM Aβ25-35 increased H2O2 levels by 60% vs. untreated controls. Incubation with some CoQs homologues slightly increased H2O2 levels, although there were no significant differences ( Figure 3B). Most of them prevented the increase in H2O2 levels induced by Aβ25-35, thus reflecting their antioxidant activity. However, H2O2 levels observed in Aβ25-35 plus CoQ2 or CoQ6 were significantly higher than those found in control, and only cells cultured with Aβ25-35 and pretreated with CoQ9 or CoQ10 had similar levels to the control (Dunnett´s multiple comparison test, Figure  3B). On the other hand, administration of 5 µM Aβ 25-35 increased H 2 O 2 levels by 60% vs. untreated controls. Incubation with some CoQs homologues slightly increased H 2 O 2 levels, although there were no significant differences ( Figure 3B). Most of them prevented the increase in H 2 O 2 levels induced by Aβ [25][26][27][28][29][30][31][32][33][34][35] , thus reflecting their antioxidant activity. However, H 2 O 2 levels observed in Aβ 25-35 plus CoQ2 or CoQ6 were significantly higher than those found in control, and only cells cultured with Aβ 25-35 and pretreated with CoQ9 or CoQ10 had similar levels to the control (Dunnett s multiple comparison test, Figure 3B).
The increase in Aβ-associated oxidative stress in AD is closely related to alterations at the mitochondrial level [25]. To discern the potential protective effect of CoQs on mitochondria, we first assessed mitochondrial status by measuring the levels of MitoTracker Red CMXRos, a red fluorescent dye that is concentrated inside mitochondria by their negative mitochondrial membrane potential (MMP). Aβ [25][26][27][28][29][30][31][32][33][34][35] addition resulted in a 27% reduction in MMP in bEnd.3 cells (Figure 3D; Student s t-test p < 0.001). Treatment with the different CoQs in the absence of Aβ 25-35 did not induce significant changes in MMP. However, only CoQ10 could prevent the Aβ 25-35 -induced decrease in MMP compared to the control (Student s t-test and Dunnett s multiple comparison test, Figure 3D).

The CoQ Isoprenoid Chain Length Determines the Protective Effect against Aβ 25-35 -Mediated Mitochondrial Bioenergetics Alterations
To further investigate mitochondrial damage, we evaluated two priority energetic pathways in cells through the oxygen consumption ratio (OCR) and extracellular acidification rate (ECAR) using Seahorse XFp > technology, following previously standardized protocols. Mitochondrial respiration was quantified by OCR under basal and energy demand conditions (stressed phenotype).
At the same time, we quantified the glycolytic function through ECAR. Aβ 25-35treatment triggered a 25-30% reduction in ECAR, thus suggesting an induction of a more quiescent phenotype in cells ( Figure 4B). Aβ 25-35 treatment significantly reduced the ECAR level versus the control with an intensity similar to OCR (24%, Figure 4B). Still, the different ubiquinones reversed this effect under all conditions, except CoQ2 under stressed conditions (Student's t-test and Dunnett's multiple comparison test p > 0.05, Figure 4B).
To further explore the protective effect of CoQ10, we performed a Seahorse Mito Stress assay to measure mitochondrial function in cells treated with Aβ 25-35 and/or CoQ10. For this purpose, different compounds were added sequentially to the cells (oligomycin, FCCP, and a mix of rotenone and antimycin A), which allowed the assessment of ATPlinked respiration, maximal respiration, and non-mitochondrial respiration, as well as the calculation of proton leak and spare respiratory capacity (Supplementary Figure S1A). Basal respiration results obtained were in line with those previously shown (Supplementary Figure S1B). Aβ 25-35 treatment significantly reduced ATP-linked respiration, maximal respiration, and non-mitochondrial respiration by 17.6%, 15%, and 18%, respectively. Pretreatment with CoQ10 prevented this increase, showing similar levels to control cells (Supplementary Figure S1B). No changes in spare capacity or proton leaking were observed (Supplementary Figure S1B).

CoQ10 but Not Other Ubiquinones Delayed and Decreased β-amyloid Incorporation into Mitochondria in Endothelial Cells
We next tested the effect of CoQ on the incorporation of the fluorescent Aβ25-35 peptide into bEnd.3 cells and, specifically, its accumulation into mitochondria, using confocal microscopy on living cells. The increase in the fluorescent Aβ25-35 signal was observed 2 h after peptide administration, and the intensity of the signal progressively increased during the next 10 h of incubation ( Figure 5A). Preincubation of bEnd.3 cells with 5 µ M CoQ2, CoQ6, and CoQ9 did not alter the amount of fluorescent Aβ25-35 detected in the cells.

CoQ10 but Not Other Ubiquinones Delayed and Decreased β-amyloid Incorporation into Mitochondria in Endothelial Cells
We next tested the effect of CoQ on the incorporation of the fluorescent Aβ 25-35 peptide into bEnd.3 cells and, specifically, its accumulation into mitochondria, using confocal microscopy on living cells. The increase in the fluorescent Aβ 25-35 signal was observed 2 h after peptide administration, and the intensity of the signal progressively increased during the next 10 h of incubation ( Figure 5A). Preincubation of bEnd.3 cells with 5 µM CoQ2, CoQ6, and CoQ9 did not alter the amount of fluorescent Aβ 25-35 detected in the cells.
However, CoQ10 pretreatment resulted in delayed internalization, which was observed after 6 h of Aβ 25-35 administration and was sustained until the end of the experiment. This significant reduction in uptake reached a decrease of 63.7% compared with control cells after 10 h of incubation (Dunnett's multiple comparison test p < 0.001, Figure 5A, orange line). Mitochondrial damage was evaluated parallel at 2 h, 6 h, and 10 h after Aβ 25-35 administration.
However, CoQ10 pretreatment resulted in a delayed internalization, which was observed after 6 h of Aβ 25-35 administration and was sustained until the end of the experiment. This significant reduction in uptake reached a decrease of 63.7% compared with control cells after 10 h of incubation (Dunnett's multiple comparison test p < 0.001, Figure 5A  However, CoQ10 pretreatment resulted in delayed internalization, which was observed after 6 h of Aβ25-35 administration and was sustained until the end of the experiment. This significant reduction in uptake reached a decrease of 63.7% compared with control cells after 10 h of incubation (Dunnett's multiple comparison test p < 0.001, Figure 5A, orange line). Mitochondrial damage was evaluated parallel at 2 h, 6 h, and 10 h after Aβ25-35 administration.
However, CoQ10 pretreatment resulted in a delayed internalization, which was observed after 6 h of Aβ25-35 administration and was sustained until the end of the experiment. This significant reduction in uptake reached a decrease of 63.7% compared with control cells after 10 h of incubation (Dunnett's multiple comparison test p < 0.001, Figure 5A, orange line). Mitochondrial damage was evaluated in parallel at 2 h, 6 h, and 10 h after Aβ25-35 administration. We observed that, after 2 h, the Mitotracker CMXRos signal was higher in cells pretreated with the different CoQs compared to Aβ 25-35 treated cells (Dunnett's multiple comparison test p < 0.01; Figure 5B), indicating that mitochondrial damage was one of the first cellular alterations induced by Aβ [25][26][27][28][29][30][31][32][33][34][35] . After 6 h and 10 h of incubation, no differences were observed in cells pretreated with CoQ2 and CoQ9 compared to non-pretreated cells, thus indicating that their preventive effect observed at 2 h was not maintained for longer durations. Nonetheless, CoQ10-pretreated cells preserved the mitochondrial staining and, therefore, the MMP, suggesting that CoQ10 protects against Aβ 25-35 -induced mitochondrial dysfunction. It is interesting to note that, although no differences were found for CoQ6 at 6 h, it also showed a protective effect after 10 h of incubation. (Figure 5B).

was observed inside the cells (Supplementary
durations. Nonetheless, CoQ10-pretreated cells preserved the mitochondrial staining and, therefore, the MMP, suggesting that CoQ10 protects against Aβ25-35-induced mitochondrial dysfunction. It is interesting to note that, although no differences were found for CoQ6 at 6 h, it also showed a protective effect after 10 h of incubation. ( Figure  5B).
After 24 h of incubation, whereas Aβ25-35 uptake in cells pretreated with CoQ2, CoQ6, and CoQ9 was similar to that of non-pretreated cells, CoQ10-pretreated cells showed a significant reduction of 53% compared to Aβ25-35 treated cells (Dunnett´s multiple comparison test p < 0.001; Figure 6A). CoQ10 also significantly reduced the internalization of Aβ25-35 compared to the other CoQ homologues ( Figure 6A). The z-stack projections confirmed that fluorescent Aβ25-35 was observed inside the cells (Supplementary Figure  S2).  The detailed images also show how Aβ 25-35 surrounds the mitochondria, implying that the interaction could occur at a mitochondrial membrane level ( Figure 6C, yellow arrows). Taken together, these results demonstrate that CoQ10 is the ubiquinone homologue that more efficiently delays and reduces both the uptake of Aβ 25-35 into bEnd.3 cells and their trafficking and incorporation into mitochondria.
NADPH oxidase exerts a crucial role in the brain due to its involvement in the generation of oxidative stress triggered by Aβ-toxicity in endothelial, glial cells, and neurons, which constitutes the cause of neuronal death [26,27]. NADPH oxidase is the major source of free radicals in blood vessels and is responsible for the cerebrovascular dysregulation induced by Aβ [28,29]. NADPH oxidase activation induces ROS production in endothelial cells and stimulates downstream signaling pathways, such as ERK1/2 and p38 MAPK, which ultimately trigger neurovascular dysfunction and neuronal death [8,10,11,16,29,30]. This effect can be attenuated in the presence of NADPH oxidase inhibitors or by adding antioxidants and ROS scavengers [26,[30][31][32]. Aβ [25][26][27][28][29][30][31][32][33][34][35] stimulated NADPH oxidase activity in bEnd.3 cells by 40%. A variety of stimuli, such as angiotensin II, oxLDL, high glucose, high salt, or Aβ, have been shown to activate NADPH oxidase, and it has been previously demonstrated that CoQ10 inhibited the activation of NADPH oxidase induced by most of them [8,11,27,29,33]. CoQ10 prevented oxLDL-and angiotensin II-induced up-regulation of p22 phox and Nox2 and impaired the assembly of NADPH oxidase in HUVEC cells [8,11]. In the present study, we report for the first time that CoQ10, but no other ubiquinone homologue, inhibits Aβ 25-35 -induced activation of NADPH oxidase in endothelial bEnd.3 cells, thereby reverting the overproduction of ROS and mitochondrial calcium dysregulation, leading to necrosis and apoptosis [17]. Although in this study we did not quantify the expression of the subunits, NADPH oxidase activity was associated with the upregulation of p22 phox mRNA [31]. NADPH oxidase is a plasma membrane enzyme, and its activation in endothelial cells involves the translocation of cytosolic oxidases to the plasma membrane and the phosphorylation of p47phox and Rac [32]. This activation induced by angiotensin II is strongly inhibited by CoQ10 [11]. Exogenous CoQ accumulates in the mitochondria, but significant amounts are also incorporated as a very light endo-lysosomal fraction in associated mitochondrial membranes, and endoplasmic reticulum and further in the plasma membrane [33]. Thus, the CoQ10 action on the Aβ-induced increase in the NADPH oxidase activity can be a cytosolic effect.
Remarkably, the other CoQ homologues, especially CoQ9, whose side chain contains only one less isoprenoid and constitutes the predominant form in mouse cells, could not prevent Aβ 25-35 -induced NADPH oxidase activation or reduce Aβ 25-35 -induced apoptosis and necrosis. Nonetheless, they partially reduced ROS levels, exhibiting the antioxidant activity of the different CoQ homologues, though the most efficient in preventing the Aβ 25-35 -induced excess of ROS was CoQ10. Due to its deeper location in the lipid bilayer, CoQ10 has been proposed to be a better antioxidant than other CoQs homologues since this distribution reduces autoxidation of ubisemiquinone and, thereby, ROS generation [34]. The higher activation of NADPH oxidase might also contribute to the lower reduction in ROS levels found with the other CoQ homologues. Furthermore, the extracellular addition of CoQ considerably increased its concentration in the different cell compartments, mainly in mitochondria [33]. CoQ homologue uptake by cells seems to be independent of the side-chain length [33,35,36]. Thus, modifying the CoQ9/CoQ10 ratio naturally present in the cell may explain why some homologues are less efficient in reducing Aβ 25-35 -induced generation of ROS since at considerably higher concentrations than the physiological level, cardiac submitochondrial particles reconstituted with CoQ9 exhibited higher rates of O 2 − generation than those obtained with CoQ10 [34]. Similarly, human cells supplemented with CoQ6 underwent a mitochondrial impairment, not observed with CoQ10 supplementation [35].
The mitochondria is a target for NADPH oxidase-generated ROS, which can induce the opening of the mitochondrial permeability transition pore (mPTP), which causes MMP collapse and ultimately apoptosis [37]. The Aβ 25-35 -induced increase in cytosolic free calcium was also impaired by most of the ubiquinones, but only CoQ10 could prevent the loss of MMP, which would further impair mitochondrial function and apoptosis.
CoQ2, CoQ6, and CoQ10 could only prevent the MMP loss after two hours of Aβ 25-35 administration, thus suggesting an initial protective effect that cannot be maintained over time. The mPTP presents a ubiquinone-binding site that regulates its aperture, and the interaction of quinones induces mPTP conformational changes that are ligand-specific [38]. Different quinone structural features are required for binding and for stabilizing the pore in the closed conformation. It has been reported that, whereas CoQ5 was ineffective, CoQ10 inhibited its opening and, therefore, prevented apoptosis [38,39]. The effect of CoQ10 in inhibiting mitochondrial depolarization is independent of its antioxidant capacity [39]. Indeed, extracellular incubation with CoQ10 increases its concentration in all cell membranes, and we have described that increasing CoQ10 plasma membrane levels in vivo by dietary supplementation entirely abolishes the aging-related increase in liver plasma membrane neutral sphingomyelinase activity, thereby inhibiting the apoptotic pathway [40].
The mitochondrial damage manifested in this study also entails decreased oxygen consumption, as shown by Seahorse XFp > technology, which allows a real-time and highly accurate assessment of metabolic alterations in cells. This harmful phenotype generated by Aβ [25][26][27][28][29][30][31][32][33][34][35] was entirely prevented by pretreatment with CoQ10 and partially rescued by other shorter chain CoQs. To date, the effect of Aβ on OCR and ECAR is not entirely clear. A recent study reported that exposure of Aβ for 7 h induces an increase in OCR, accompanied by mitochondrial damage, increased ROS, and electron transport disruption in cerebrovascular endothelial cells [41]. However, in the present study, we observed a decrease in OCR and ECAR induced by Aβ 25-35. bEnd. 3 cells may show a biphasic response of mitochondrial oxygen consumption upon Aβ 25-35 administration by exhibiting increased oxygen consumption after short exposure times and a reduction upon more prolonged exposure (24 h), as obtained in our study and previous others [42][43][44][45][46]. This decrease correlates with the alteration of MMP and the increase in the ROS level.
It has been previously shown how the direct and indirect (through ROS increase) interaction of Aβ with mitochondria disrupts the function of different components of the electron transport chain, thus affecting cellular respiration [47,48]. Here, we report that Aβ [25][26][27][28][29][30][31][32][33][34][35] induces an alteration of the whole mitochondrial respiratory chain, decreasing basal respiration, ATP production, and maximal respiration. These effects were prevented by the pretreatment with CoQ10, confirming the results reported by Sadli et al., in neuroblastoma cells [45]. The protective effect of CoQ10 on different respiration parameters confirmed the role of CoQ10 at the mitochondrial level. Pretreatment with CoQ10 also prevented the deleterious effect of Aβ 25-35 on non-mitochondrial respiration (enzymatic oxygen consumption), thus implying a dual effect at both mitochondrial and cytosolic levels. Pretreatment with CoQ9 and CoQ6 prevented the decrease in OCR and ECAR levels generated by Aβ 25-35 but did not hamper MMP alteration. This may be explained since exposure to these CoQs slightly reduced the interaction of Aβ 25-35 with mitochondria, but not as markedly as CoQ10, a potent inhibitor of mPTP opening.
Aβ is a highly hydrophobic peptide capable of channel formation in lipid bilayers, a property that has been proposed as part of their pathogenic mechanism because it disrupts intracellular calcium regulation and MMP [49]. We previously reported that CoQ10 protection from Aβ-induced deleterious effects in endothelial cells also involves these mechanisms, reducing Aβ trafficking and accumulation into the mitochondria, impeding both mPTP opening and cytochrome c release [17]. In the present work, we describe that only CoQ10 among the ubiquinone homologues tested could reduce the uptake of Aβ 25-35 into both the cell and the mitochondria and that this effect correlates with the maintenance of MMP. To form a pore into the membrane, Aβ aggregates spontaneously into extended fibrils and precipitates out of solution, and then, dozens of these peptides are inserted in lipid bilayers forming a channel [49]. It has been reported that CoQ10 inhibits the formation of Aβ fibrils and their extension, and it destabilizes preformed fibrillar Aβ [50]. To our knowledge, there are no studies that address the interaction of Aβ with other CoQ homologues. CoQ10 addition to the culture medium of the cells increases the cellular concentration of CoQ10 more than 1000 times compared to the control [36], so one possible mechanism by which CoQ10 decreases Aβ incorporation may be by reducing its membrane insertion or by destabilizing the fibril formed during its insertion into cell membranes. It has been proposed that decreased lipophilicity might reduce Aβ and/or fibrillar Aβ binding affinity and exhibits anti-amyloidogenic and fibril-destabilizing effects in vitro [50]; thereby, the three-dimensional conformation of CoQ10 could be more efficient than other homologues to impede the association between Aβ molecules that leads to fibrillar Aβ formation. Furthermore, the influence of CoQ10 on the permeability and mechanical properties of phospholipid membranes is similar to that of cholesterol [51]. The increase in cholesterol in membranes inhibits Aβ channel insertion in liposomal and cellular membranes, and oppositely, cholesterol depletion facilitates peptide insertion in membranes and pore formation [49].
It is difficult to explain why CoQ10 exhibits a protective effect on endothelial cells from Aβ 25-35 -induced damage while other homologues do not, even though the major form in mouse cells is CoQ9. However, this intriguing difference has been described in other processes such as Mumps virus infection, which causes the death of 30-35% of cultured rat neurons and decreases the concentration of CoQ10 and CoQ9. It was demonstrated that supplementation of the medium with 5 µM CoQ9 did not prevent this neuronal death, whereas CoQ10 supplementation was shown to be protective and increased the number of surviving neurons to 90% in the dorsal root ganglion and cerebral cortex [52]. Additionally, longevity correlates positively with CoQ10 levels and negatively with CoQ9 levels in different mammal species, and it has been suggested that CoQ10 may be a better antioxidant than CoQ9 due to its deeper localization in the lipid bilayer, which reduces autoxidation of ubisemiquinone and thereby ROS generation [34]. We previously demonstrated that CoQ10H 2 and ascorbate reduce cerebral hypoxia and prevent BACE1 expression and Aβ deposition while reducing the thickness of the cerebrovascular basement membrane during early asymptomatic stages of AD in the 3xTg-AD mice model of AD [18]. Additionally, CoQ10H 2 reduced chronic inflammation and improved cerebral vasoreactivity and endothelial dysfunction in patients with mild cognitive impairment when the CoQ10H 2 concentration reached a concentration of 5 µg/mL in plasma [19]. AD is a neurodegenerative disease whose main risk factor is age. The CoQ10/CoQ9 ratio declines during aging, and this process appears to be attenuated by calorie restriction, which increases CoQ10 biosynthesis [53]. Calorie restriction also attenuates AD-type brain amyloidosis in the 3xTg-AD mouse model and monkeys by reducing the Aβ content in the brain [54,55]. Another dietary intervention that favors the synthesis of CoQ10 over CoQ9, thus altering the ratio between CoQ homologues, also ameliorated the risk of many age-related disorders [56].

Conclusions
The results presented in this work demonstrate that only CoQ10 among tested ubiquinones protects endothelial cells from Aβ 25-35 -induced damage. CoQ10 was the only ubiquinone that inhibited Aβ 25-35 -mediated activation of NADPH oxidase, maintained MMP, and reduced Aβ 25-35 cellular and mitochondrial uptake, thus suggesting that this protective effect is not only mediated by its antioxidant capacity.