Treatment of CoQ10 Deficient Fibroblasts with Ubiquinone, CoQ Analogs, and Vitamin C: Time- and Compound-Dependent Effects

Background Coenzyme Q10 (CoQ10) and its analogs are used therapeutically by virtue of their functions as electron carriers, antioxidant compounds, or both. However, published studies suggest that different ubiquinone analogs may produce divergent effects on oxidative phosphorylation and oxidative stress. Methodology/Principal Findings To test these concepts, we have evaluated the effects of CoQ10, coenzyme Q2 (CoQ2), idebenone, and vitamin C on bioenergetics and oxidative stress in human skin fibroblasts with primary CoQ10 deficiency. A final concentration of 5 µM of each compound was chosen to approximate the plasma concentration of CoQ10 of patients treated with oral ubiquinone. CoQ10 supplementation for one week but not for 24 hours doubled ATP levels and ATP/ADP ratio in CoQ10 deficient fibroblasts therein normalizing the bioenergetics status of the cells. Other compounds did not affect cellular bioenergetics. In COQ2 mutant fibroblasts, increased superoxide anion production and oxidative stress-induced cell death were normalized by all supplements. Conclusions/Significance These results indicate that: 1) pharmacokinetics of CoQ10 in reaching the mitochondrial respiratory chain is delayed; 2) short-tail ubiquinone analogs cannot replace CoQ10 in the mitochondrial respiratory chain under conditions of CoQ10 deficiency; and 3) oxidative stress and cell death can be counteracted by administration of lipophilic or hydrophilic antioxidants. The results of our in vitro experiments suggest that primary CoQ10 deficiencies should be treated with CoQ10 supplementation but not with short-tail ubiquinone analogs, such as idebenone or CoQ2. Complementary administration of antioxidants with high bioavailability should be considered if oxidative stress is present.


Introduction
Coenzyme Q 10 (CoQ 10 ; ubiquinone) and its analogs have been evaluated as antioxidant agents and enhancers of mitochondrial functions in patients with mitochondrial disorders and clinical trials of neurodegenerative diseases including Parkinson disease, amyotrophic lateral sclerosis, Huntington disease, Friedreich ataxia, and Alzheimer's disease with modest or no objective benefits [1][2][3][4][5][6]. The use of CoQ 10 therapy and its analogs in primary CoQ 10 deficiency, an autosomal recessive syndrome due to defects of ubiquinone biosynthesis, could provide valuable data to evaluate the effectiveness of these compounds in restoring respiratory chain activities and preventing oxidative stress. The disorder manifests clinically with four major phenotypes: 1) an encephalomyopathy with brain involvement and recurrent myoglobinuria [7]; 2) an infantile multisystem disorder with encephalopathy usually associated with nephropathy and variable involvement of other organs [8,9]; 3) ataxic syndrome with cerebellar atrophy [10,11]; and 4) an isolated myopathy [12,13].
Molecular defects in genes encoding CoQ 10 biosynthetic proteins have been reported in 18 patients. Four patients improved with CoQ 10 supplementation [9,14-17], five died before or during the treatment, and 9 had no definite response [14,15,[18][19][20][21][22]; it is therefore difficult to reach definitive conclusions about the effectiveness of CoQ 10 supplementation in primary CoQ 10 deficiencies. To better understand the pathogenesis of CoQ 10 deficiency, we have characterized the bioenergetics and oxidative stress in PDSS2 and COQ2 mutant fibroblasts, and have demonstrated that severe CoQ 10 deficiency caused marked defects of ATP synthesis without oxidative stress whereas milder CoQ 10 deficiency produced reactive oxygen species (ROS) and oxidation of proteins and lipids [23]. Here, we evaluate the in vitro effects of CoQ 10 supplementation on the bioenergetics and oxidative stress status of CoQ 10 deficient fibroblasts with mutations in PDSS2, COQ2, and COQ9 (Fig. 1). In addition, because CoQ 10 analogs and vitamin C are being used in clinical trials based on their antioxidant properties, we concurrently evaluated the effect of CoQ 2 , idebenone, and vitamin C.

Cellular CoQ 10 levels after treatment with compounds for 24 hours
Fibroblasts from the four patients with different molecular defects in the CoQ 10 biosynthetic pathway used in this study showed significantly decreased levels of CoQ 10 relative to controls (P,0.001) (Fig. 2). When control and patients' cells were treated for 24 h with 5 mM of CoQ 10 , cellular levels of ubiquinone increased significantly in all cells (P,0.001), resulting in values 20-85-fold higher than in control cells (Fig. 2). When the cells were treated with idebenone, CoQ 2 , or vitamin C, cellular levels of CoQ 10 remained unchanged (Fig. 2).
The effect of compound administration on superoxide anion levels: 24 hours versus 1 week of treatment After incubation for 24 h in galactose medium supplemented with dialyzed FBS, P3 cells showed increased MitoSOX Red stain indicating elevated levels of superoxide anions (P,0.001) (Fig. 5A). Under the same culture conditions, the other three CoQ 10 deficient fibroblasts did not show any significant changes in MitoSOX stain. After 24 h of treatment with CoQ 10 , idebenone, CoQ 2 , or vitamin C, superoxide anion levels decreased significantly in P3 cells (P,0.01) (Fig. 5A). Cells from controls and from the three other patients did not show significant changes in MitoSOX stain after 24 hours of treatment with any of the four compounds; however, there was a trend towards increased superoxide anions in P4 cells treated with CoQ 2 for 24 h (Fig. 5A).
After one week of incubation in galactose medium plus dialyzed FBS, P2 and P3 cells showed increased MitoSOX Red staining (Fig. 5B). After 1 week of treatment with CoQ 10 , idebenone, CoQ 2 , or vitamin C, superoxide anion levels were decreased significantly in both P2 and P3 cells (P,0.001) (Fig. 5B).
The effect of compound administration on cell death: 24 hours versus 1 week treatment In adherent cells, cell death was significantly higher in untreated P2 and P3 cells than in control cells (P,0.05 and P,0.01, respectively) (Fig. 6A). Cell death was significantly reduced in both P2 and P3 treated for 24 h with CoQ 10 (P,0.01) or idebenone (P,0.05). Vitamin C also reduced cell death in P3 cells (P,0.05) (Fig. 6A). Control, P1, and P4 cells showed similar proportions of dead cells with and without treatment (Fig. 6A).
Because we observed abundant floating COQ2 mutant cells cultured in galactose medium with dialyzed FBS, we performed Trypan blue staining of P2 cells and found that 15% of the cells were dead after 24 h and 63% were dead after 1 week. Percentages of floating dead cells were significantly decreased after one week of treatment with CoQ 10 (36% dead cells), CoQ 2 (45% dead cells), or idebenone (43% dead cells).

Discussion
Over the last 4 years, studies of CoQ 10 deficient patients have focused on two major issues: identifying the molecular genetic basis of ubiquinone deficiencies and characterizing pathogenic mechanisms [24]. We identified the two first mutations in genes encoding proteins required for CoQ 10 biosynthesis, COQ2 [17] and PDSS2 [20], in patients with primary CoQ 10 deficiency. Our studies of cultured fibroblasts from these patients revealed that both CoQ 10 biosynthetic disorders caused bioenergetic defects, but only COQ2 mutant fibroblasts showed increased ROS production and signs of oxidative stress [23]. In the present study, we have evaluated the in vitro efficacy of CoQ 10 supplementation in normalizing the bioenergetic status and the oxidative balance in fibroblasts of CoQ 10 deficient patients harboring mutations in PDSS2, COQ2, and COQ9 [14,17,18,20,23]. We chose a final concentration of 5 mM CoQ 10 because this is the approximate concentration reached in the plasma of patients treated with oral supplementation of CoQ 10 [3,6,[25][26][27]. We have also evaluated the efficacy of short-tail ubiquinone analogs (idebenone and CoQ 2 ), which are less lipophilic, and vitamin C, a well-known hydrophilic antioxidant. It is noteworthy that 5 mM concentrations of ubiquinone analogs exceed the EC 50 of idebenone (,0.4-0.5 mM) and CoQ 2 (,0.03 mM) in cultured fibroblasts under oxidative stress [28,29].
After 24 h of 5 mM CoQ 10 supplementation, cellular CoQ 10 levels increased dramatically. The magnitudes of the increases were independent of the molecular defects of the cells since both control and patients' fibroblasts showed similar degrees of CoQ 10 uptake. However, after 24 h of CoQ 10 treatment, none of the cell lines showed significant improvement in ATP levels or in ATP/ ADP ratios, which are markers of respiratory chain function. In contrast to our results, López-Martín and colleagues [30] noted normalization of mitochondrial complexes I+III and II+III activities in COQ2 mutant fibroblasts after 24 h of 10 mM CoQ 10 supplementation. Paradoxically, however, the same authors using the same COQ2 mutant cells and other genetically undefined CoQ 10 deficient fibroblasts found that the activities of complex II+III activities increased only slightly and remained below control values after 72 hours of 100 mM CoQ 10 supplementation [31]. This discrepancy may be due to the fact that in the second study, but not in the first, respiratory chain enzyme activities were normalized to the activity of citrate synthase [31], a marker of mitochondrial mass [32]. In another study, embryonic cells from COQ7 knockout mice treated with 25 mM of watersoluble CoQ 10 for 5 h only showed slight increase of ATP [33]. The negligible effects of CoQ 10 supplementation on the bioenergetic status of CoQ 10 deficient cells can be explained by the strong lipophilic nature of CoQ 10 , which accumulates in membranes reaching saturated concentrations [34,35]. Thus, exogenous CoQ 10 is mainly distributed in lysosomes, endoplasmic reticulum, and plasma membrane and only a small proportion (,11%) reaches the mitochondria [36,37]. In addition, much of the CoQ 10 in mitochondria is likely to be trapped in the outer membrane, not available to the respiratory chain, which is located in the mitochondrial inner membrane [38].
Because 5 mM CoQ 10 did not increase the ATP levels after 24 h of supplementation, we tried two other strategies: 1) increasing the duration of CoQ 10 supplementation to 1 week; and 2) using lesslipophilic ubiquinone analogs. In marked contrast to treatment for 24 h of CoQ 10 , incubation of ubiquinone-deficient fibroblasts for 1 week with 5 mM CoQ 10 increased ATP levels and ATP/ADP ratios significantly, indicating normalization of the bioenergetic status. Similar to these results, yeast coq mutants show inefficient uptake of exogenous CoQ 6 to the mitochondrial inner membrane, which is reflected in a low succinate cytochrome c reductase activity after 2-15 mM CoQ 6 supplementation for 48 h [37,39]. The complete rescues of growth of the yeast coq mutants' supplemented with 15 mM CoQ 6 were only possible after 6-8 days [37,39,40]. These results suggest that the pharmacokinetic constraints of CoQ 10 in reaching the mitochondrial respiratory chain are key limiting factors in determining its efficacy in CoQ 10 deficient patients. Furthermore, an in vivo study of PDSS2 mutant mice treated for 4 months with 100 mg/kg body weight (b.w.)/day or 200 mg/kg b.w./day of water-soluble CoQ 10 showed that only the highest dose of CoQ 10 improved the renal function in these CoQ deficient animals [41]. Thus, the dose of CoQ 10 is another important factor influencing the effectiveness of CoQ 10 supplementation in ubiquinone deficient patients. Nevertheless, the gold standard to evaluate the effectiveness of CoQ 10 treatment in ubiquinone deficiency is in vivo measurements of mitochondrial function, bioenergetics, and oxidative stress.
In contrast to CoQ 10 , short-tail ubiquinone analogs and vitamin C failed to increase ATP levels and ATP/ADP ratio in patients' cells after both 24 h and 1 week treatment. These results are in agreement with previous studies showing that hydrophilic ubiquinone analogs (CoQ 2 and idebenone) or mitochondrialtargeted ubiquinone (MitoQ) are less efficient than hydrophobic ubiquinone analogs in enhancing energy production by the mitochondrial respiratory chain [42][43][44], since the specific effects  . CoQ 10 biosynthesis pathway. CoQ 10 is composed of a benzoquinone and a decaprenyl side chain. PDSS2 is the second subunit of decaprenyl diphosphate synthase, a heterotretameric enzyme that catalyzes the formation of the decaprenyl side chain. COQ2 or parahydroxybenzoate (PHB)-polyprenyl transferase catalyzes the condensation reaction of PHB and decaprenyl diphosphate. The function of COQ9 is still unknown. The mutant fibroblasts used in this study harbor mutations in COQ9, p.244R.X (P1) [18], COQ2, p.197R.H, p.228N.S (P2) [14] and p.297Y.C (P3) [17], and PDSS2, p.322Q.X, p.382S.L (P4) [20]. doi:10.1371/journal.pone.0011897.g001 of ubiquinones depend on their interaction with hydrophilic and hydrophobic (or physiological) binding sites [42,43]. Accordingly, hydrophobic and hydrophilic ubiquinone analogs are not interchangeable. Experience with two patients with CoQ 10 deficiency and cerebellar ataxia due to ADCK3/CABC1 mutations is relevant. Both patients were treated with CoQ 10 (5-10 mg/kg/ b.w.) [21,45]. One patient, who also had exercise intolerance and hyperlactatemia, improved after 3 months of treatment whereas the other, who had only ataxia, did not benefit from 3 years of therapy. The ataxia did not improve in either patient and for this reason CoQ 10 was replaced with low-dose idebenone in both patients (5-10 mg/kg/b.w.). However, their symptoms worsened, prompting reinstitution of CoQ 10 therapy [21,45]. Based on our in vitro results showing improvements of bioenergetics by ubiquinone in CoQ 10 deficient cells, we postulate that CoQ 10 therapy improved the first patient's exercise intolerance and may have stabilized symptoms in the second patient because of enhanced ATP production. The lack of improvement of cerebellar symptoms may be due to poor transfer of CoQ 10 across the blood brain barrier, irreversible structural alterations in the cerebellum, or both factors [36], especially considering the low dose used (5-10 mg/kg/b.w.). Our finding that high doses of idebenone in vitro failed to increase ATP levels in CoQ 10 deficient fibroblasts or the low-dose used for therapy may explain ineffectiveness of idebenone therapy compared to CoQ 10 treatment in both patients.
Because oxidative stress is another important pathogenic factor in CoQ 10 deficiencies, we evaluated also the effect of CoQ 10 and three other antioxidants on superoxide anion production. After both 24 h and after 1 week of treatment, we noted that all four compounds significantly reduced MitoSox staining in P3 cells (the cell line with the most prominent MitoSox fluorescence). Reduction of superoxide anion levels correlated with decreased cell death, suggesting that ROS generation and oxidative damage are the main causes of death in COQ2 mutant fibroblasts. This notion is supported by our observation that P3 COQ2 fibroblast line had the highest levels of superoxide anion production and protein oxidation and showed the most robust features of apoptosis and the highest proportion of cell death [46]. In contrast to COQ2 mutant fibroblasts, COQ9 and PDSS2 mutant fibroblast do not have increased ROS, a difference that may be related to the disparate CoQ-dependent flow of electrons in the mitochondrial respiratory chain in the different mutant cell lines. As we recently reported, 10-15% residual CoQ 10 are not associated with significant ROS production, whereas 30-50% residual CoQ 10 is accompanied by increased ROS production and cell death [23,46]. Interestingly, there was a trend towards increased ROS production in P4 cells supplemented with CoQ 2 for 24 h and in P1 cells supplemented with idebenone for 1 week. In contrast, idebenone decreased ROS levels in the other 3 cell lines after one week of treatment. Our data suggest that these compounds may have pro-oxidant and anti-oxidant properties under certain in vitro conditions [47] and confirm in cultured fibroblasts observations previously reported in isolated mitochondria [48].
In summary, our in vitro study tackles important issues regarding treatment of CoQ 10 deficiency and, more in general, of mitochondrial diseases. First, the prolonged pharmacokinetics of CoQ 10 to reach the mitochondrial respiratory chain is critical to the restoration of energy status of human CoQ 10 deficient cells. This may explain the delayed clinical response of CoQ 10 deficiency patients to oral supplementation with CoQ 10 [16] and suggest that high doses of CoQ 10 are needed. Second, short tail ubiquinone analogs cannot substitute for CoQ 10 in the mitochondrial respiratory chain of human CoQ 10 deficient cells revealing the importance of the decaprenyl tail. Third, oxidative stress and cell death can be attenuated by the administration of lipophilic and hydrophilic antioxidants.
Thus, our results confirm that in CoQ 10 deficient patients: (i) early treatment based on early diagnosis is critical to maximize the efficacy of ubiquinone supplementation; (ii) short-tail ubiquinone analogs (i.e. idebenone or CoQ 2 ) are not suitable substitutes for CoQ 10 ; and (III) complementary administration of antioxidants with high bioavailability may be helpful in CoQ 10 deficiency patients. Further studies on the pathogenesis of CoQ 10 deficiency in patients with different molecular defects and in animal models will lead to more rational and improved therapies.

Ethics Statement
Skin biopsies, to generate cultured fibroblasts, were performed after obtaining written informed parental consents under study protocols approved by the Institutional Review Boards of the Columbia University Medical Center, University of Padova, and Charité University Hospital as well as Local Research Ethics committees of Great Ormond Street Hospital for Children NHS Trust and Institute of Child Health. The informed consents and clinical studies were in compliance with the principles expressed in the Declaration of Helsinki.
Regular FBS contains 85-245 mg glucose/ml and dialyzed FBS less than 5 mg glucose/ml. Because our experiments required glucose-free media supplemented with galactose to force energy production through oxidative phosphorylation rather than glycol-ysis [23,49,50], we used 10% dialyzed FBS. This was particularly important for the 24 h treatment experiments since cells can use the glucose contained in 10% regular FBS during the first 24-48 h.
Idebenone, CoQ 2 , CoQ 10 , and vitamin C were dissolved in dialyzed fetal bovine serum prior to addition to culture medium.
All cells culture reagents were obtained from Invitrogen Corp. (Carlsbad, CA, USA). Patients and control cell lines at passage 7-10 were grown to replicate every experiment at least 3 times.

Cellular CoQ 10 levels
One week after treatment, fibroblast pellets were collected and washed 4 times with phosphate buffered saline (PBS). After the fourth wash, no CoQ 10 was detected in the PBS supernatant. CoQ 10 from fibroblasts was extracted in a hexane:ethanol mixture [20]. The lipid component of the extract was separated by HPLC on a reverse phase SymmetryH C18 3.5 mm, 4.66150 mm column (Waters), using a mobile phase consisting of methanol, ethanol, 2propanol, acetic acid (500:500:15:15) and 50 mM sodium acetate at a flow rate of 0.9 ml/min. The electrochemical detector consisted of an ESA Coulochem II with the following setting: guard cell (upstream of the injector) at +900 mV, conditioning cell at -600 mV (downstream of the column), followed by the  analytical cell at +500 mV. CoQ 10 concentration was estimated by comparison of the peak area with those of standard solutions of known concentration [20].

Mitochondrial bioenergetics assessment
Fibroblasts cultured under glucose-free media supplemented with galactose possess ATP and ADP pools that are predominantly dependent on the oxidative phosphorylation [23,49]; therefore, using these culture conditions, ATP levels and ATP/ADP ratio are reliable markers of oxidative phosphorylation and correlate with the rates of ATP synthesis in CoQ 10 deficient cells, as reported [23]. To determine the adenine nucleotides levels, cells were grown in 15-cm-diameter plates to 90% confluency and the adherent cells were collected in ice-cold PBS using a scraper. After centrifugation at 3,000 g for 3 min at 4uC, pellets were suspended in 200 ml of ice-cold 0.5 M perchloric acid, vortexed for 30 s, and centrifuged at 11,000 g for 10 min at 4uC. Pellets were stored at 280uC for protein measurement. Adenine nucleotides were measured in the supernatants injected into an Alliance HPLC (Waters Corporation, Milford, MA, USA) with an Alltima C18 NUC reverse-phase column (Alltech Associates, Deerfield, IL, USA) [23]. Adenine nucleotide levels were expressed in nmol/mg protein. Because of poor growth of P3 cells, the 1 week experiments were performed only in P1, P2 and P4 fibroblasts.

Oxidative stress analyses
Previously, we demonstrated that oxidative stress can be detected by MitoSOX Red assay, lipid peroxidation assay, and carbonyl groups detection in CoQ 10 deficient cells due to COQ2 mutation [23]; and results of the three assays were highly concordant [23]. Here, to evaluate the effect of the proposed treatments in those and others CoQ 10 deficient cells, we estimated mitochondrial matrix oxidant levels in control and mutant fibroblasts using MitoSOX Red, a fluorochrome specific for mitochondrial matrix reactive oxygen species (ROS) burden (Molecular Probes, Invitrogen Corp., Carlsbad, CA, USA) as described [23,51]. Approximately 1610 6 adherent cells (90% confluence) were trypsinized, incubated with 5 mM MitoSox for 20 min at 37uC in the dark, washed twice with PBS and resuspended in 500 ml of PBS. Cytofluorometric analysis was performed using a FACSCalibur cell analyzer equipped with a 488 nm At-kt laser and fluorescence was measured using the FL2 channel. Fluorescence was background subtracted and normalized to the signal of control values (F P /F C ). Data were acquired using Cell Pro Quest and analyzed using Flowjo software (Becton Dickinson, NJ, USA). Before to start with the set of experiments for the cytofluorometric analysis, the MitoSOX staining was evaluated and monitored by fluorescence microscopy (IX70 inverted system microscope, Olympus, Tokyo, Japan). Additionally, a control cell line treated with 0.5 mM antimycin A for 15 h before the MitoSOX staining was used as a positive control for increased ROS production [23].

Oxidative stress-induced death cells studies
Measurement of cell viability was performed with the Trypan blue exclusion method. Numbers of living and dead adherent cells were determined by light microscopy using a hemocytometer. We also determined the proportion dead non-adherent P2 cells. Healthy nuclei from viable cells appeared round and phase bright, whereas nuclei from dead or dying cells appeared blue and irregularly shaped. All cells were counted and results were expressed as percentage of dead cells relative to total cells [52].

Statistical analysis
Control data are expressed as mean 6 SD of 5 different samples in triplicate experiments. Patients' data are expressed as mean 6 SD of triplicate experiments. One-way analysis of variance (ANOVA) followed by Student-Newman-Keuls multiple comparisons test was used. A P value of less than 0.05 was considered statistically significant.